Materials Processing In Magnetic Fields

Materials Processing I n Magnetic Fields This page intentionally left blank Materials Processing In Magnetic Fields...

Author: Hans J. Schneider-Muntau | Hitoshi Wada

241 downloads 1898 Views 20MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Materials Processing I n Magnetic Fields

This page intentionally left blank

Materials Processing In Magnetic Fields Proceedings of the International Workshop on Materials Analysis and Processing in Magnetic Fields 17 - 19 March 2004

Tallahassee, Florida

editors

Hans J Schneider-Muntau Florida State University, USA

Hitoshi Wada National institute for Materials Science, Japan

N E W JERSEY * L O N D O N

v -

World Scientific

SINGAPORE * B E l J l N G * S H A N G H A I * HONG KONG

-

TAIPEI * CHENNAI

Published by

World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224

USA ofice: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 U K ofice: 57 Shelton Street, Covent Garden, London WC2H 9HE

British Library Cataloguing-in-PublicationData A catalogue record for this book is available from the British Library.

MATERIALS PROCESSING IN MAGNETIC FIELDS Proceedings of the International Workshop copyright 0 2005 by World Scientific Publishing Co. Re. Ltd.

All righis reserved. This book, or parts ihereof; may not be reproduced in any form or by any means, elecironic or mechanical, includingphotocopying, recording or any information storage and retrieval sysiem now known or to be invented, wiihoui written permission from the Publisher.

Forphotocopying of material in this volume, please pay acopying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.

ISBN 981-256-372-5

Printed in Singapore by Mainland Press

Editors’ Preface Processing in magnetic fields is a rapidly expanding research area with a wide range of promising applications in materials science, development and design. Industry now offers a variety of superconducting magnets specifically designed for this purpose and equipped with cryocoolers that eliminate the need for cryogenic fluids. Numerous research centers dedicated to materials research and processing in magnetic fields have been created around the world. This book is the result of an international by-invitation-only workshop that has been organized to review the most recent activities in this field. Over 50 scientists participated and 39 papers were selected for inclusion in this book. Magnetic fields are at the origin of many effects in materials, of which the following are a few examples. The high fields available now allow us to levitate diamagnetic matter with the advantage of containerless processing of materials, or control gravity on earth from several -g to +g. The magnetization induced in dia- and paramagnetic matter is strong enough to change the structure and characteristics of materials, typically as a collective phenomenon. Magnetic anisotropy can be used for aligning fibers, polymers, and carbon nanotubes, resulting in matrix systems with superior quality. The magnetic field has an impact on texturing of materials during a phase transition, in both liquid-to-solid and solid-to-solid state transitions. Grain boundary migration and mobility changes have been demonstrated in Bi and Zn crystals, giving us a perspective for texture development in metals. The damping effect of magnetic fields on conductive liquids is exploited for improved crystal growth quality. Further research will investigate the use of static fields for convection and texture control, possibly at reduced gravity levels. Field geometry and the application of rotational fields are additional optimization parameters. Of growing interest are the effects of magnetic fields in biology and their beneficial applications. Magnetic fields can help manipulate cells and cellular processes, such as cell divisions. Magnetic microspheres can be guided within the body for drug delivery and tumor treatment. Another future application is to mix nanomagnetic particles with biological blood components for treatment, blood cell separation or as a marker. Other applications of magnetic field processing are magnetic separation, and processing of crystalline fibers through alignment. Additional research areas were presented and discussed during the workshop and are contained in this book. The International Workshop on Materials Analysis and Processing in Magnetic Fields has been jointly organized and sponsored by the Tsukuba Magnet Laboratory of the National Institute for Materials Science in Tsukuba, Japan, and the National High Magnetic Field Laboratory of the Florida State University in Tallahassee, Florida. It was held at the NHMFL in Tallahassee on March 17-19,2004.

Hans J. Schneider-Muntau Hitoshi Wada V


Table of Contents v

EDITORS’ PREFACE

Texturing and Phase Transitions APPLICATION OF HIGH MAGNETIC FIELDS IN MATERIALS PROCESSING S. ASAI

3

HIGH MAGNETIC FIELDS EFFECTS ON SOLID STATE TRANSFORMATIONSAT HIGH TEMPERATURE E. BEAUGNON, E GAUCHERAND

11

ENHANCEMENT OF MATERIAL PROPERTIES BY MAGNETIC FIELD ASSISTED PHASE TRANSFORMATION B.Z. CUI, K. HAN, H. GARMESTANI, H.J. SCHNEIDER-MUNTAU, J.H. SU, J.R LIU

19

EXPERIMENTAL INVESTIGATIONOF THE CRYSTALLIZATION OF BHF IN HIGH MAGNETIC FIELDS W ERTEL-INGRISCH, K. HARTMANN, X. WANG, D. HULSENBERG

29

VARIATION OF PHASE TRANSFORMATIONTEMPERATURE IN FE-C ALLOYS IN A HIGH MAGNETIC FIELD X.J. HAO, H. OHTSUKA, H. WADA

41

MARTENSITIC TRANSFORMATIONIN SOME FERROUS ALLOYS UNDER MAGNETIC FIELD T. KAKESHITA

48

EXPLORING ULTRA-HIGH MAGNETIC FIELD PROCESSING OF MATERIALS FOR DEVELOPING CUSTOMIZED MICROSTRUCTURESAND ENHANCED PERFORMANCE G.M. LUDTKA, R.A. JARAMILLO, R.A. KISNER, J.B. WILGEN, G. MACKIEWICZ-LUDTKA, D.M. NICHOLSON, T.R. WATKINS, I? KALU, R.D. ENGLAND

55

vii

viii

FUNDAMENTALSAND APPLICATIONS OF GRAIN BOUNDARY DYNAMICS IN HIGH MAGNETIC FIELDS D.A. MOLODOV

66

ENHANCEMENT OF TEXTURE AND CRITICAL CURRENT DENSITY IN Bi,Sr2CalCu208SUPERCONDUCTING TAPES THROUGH MAGNETIC FIELD PROCESSING P.Vl?S.S. SASTRZ U.l? TROCIEWIlZ, H. MAEDA, J. SCHWAR'IZ

80

APPLICATION OF HIGH MAGNETIC FIELD TO TEXTURE MODIFICATION IN ZINC ALLOY A.D. SHEIKH-ALI, D.A. MOLODOY H. GARMESTANI

91

TEXTURING FROM LIQUID TO SOLID STATE BY ALIGNING ANISOTROPIC MAGNETIC NUCLEI IN HIGH FIELDS R.E TOURNIER

102

Chemical and Physical Processes REFRACTIVE INDICES OF WATER AND AQUEOUS ELECTROLYTE 115 SOLUTIONS UNDER HIGH MAGNETIC FIELDS H. HOSODA, H. MORI, N. SOGOSHI, S. NAKABAYASHI SYNTHESIS OF CARBON MATERIALS BY THE IMPOSITION OF A HIGH MAGNETIC FIELD M.-G. SUNG, K. SASSA, A. GEDANKEN, K. IWAI, S. ASAI

124

HIGHLY EXCITED MOLECULES IN MAGNETIC FIELDS K. TAKAZAWA

136

APPLICATION OF HIGH MAGNETIC FIELD TO CHEMICAL AND PHYSICAL PROCESSES I: TANIMOTO, W DUAN

141

MAGNETO-CHEMICAL SYSTEMS UNDER STRONG MAGNETIC FIELDS: FUNDAMENTALSAND APPLICATIONS M. YAMAGUCHI, I. YAMAMOTO

147

ix

Control of Liquids APPLICATIONS OF AC AND DC MAGNETIC FIELDS IN METALLURGICALAND CRYSTAL GROWTH PROCESSES A. CRAMER, S. ECKERT, V GALINDO, J. PRIEDE, G . GERBETH

157

ELECTROMAGNETIC PROCESSING OF MATERIALS: FROM THE CONCEPTS TO INDUSTRIALAPPLICATIONS I:DELANNOY

169

SEMICONDUCTOR CRYSTAL GROWTH IN STATIC AND ROTATING MAGNETIC FIELDS M.I?VOLZ

178

Magnetic Separation REMOVAL SYSTEM OF ARSENIC FROM GEOTHERMAL WATER BY MAGNETIC SEPARATION TECHNOLOGY WITH A SUPERCONDUCTINGMAGNET H. OKADA, K. MITSUHASHI, 7: OHARA, H. WADA, I: KUDOH, H. NAKAZAWA, A. CHIBA

197

MAGNETICALLY ENHANCED SOLID-LIQUID SEPARATION C.M. REX K. KELLER, B. FUCHS

206

NEW APPLICATIONS OF MAGNETIC SEPARATION USING SUPERCONDUCTINGMAGNETS AND COLLOID CHEMICAL PROCESSES S. TAKEDA, S.-J. YU, A. NAKAHIRA, I: IZUMI, S. NISHIJIMA, 7: WATANABE

220

Biological Applications NANOMAGNETICS IN BIOTECHNOLOGY C.-J. CHEN, I: HAIK, J. CHATTERJEE

229

X

STRONG MAGNETIC FTELD INDUCED CHANGES OF GENE EXPRESSION IN ARABIDOPSIS A.-L. PAUL, R.J. FERL, B. KLINGENBERG, J.S. BROOKS, A.N. MORGAN, J. YOWAK, M.W MEISEL

238

NEW APPLICATIONS OF MAGNETIC FIELD TO HUMAN FRIENDLY MATERIALS AND HUMAN SUPPORTIVE SYSTEMS S.TAKEDA, U. HAFELI, M.TONOIKE, I:IZUMI, K. EMA, S.NISHIJMA

243

MAGNETIC ORIENTATION IN BIOLOGY VIRUS STRUCTURE - BLOOD CLOT ASSEMBLY - CELL GUIDANCE J. TORBET

249

MANIPULATING CELLS WITH STATIC MAGNETIC FIELDS J.M. VALLES, JR., K. GUEVORKIAN

257

Diamagnetic Effects EFFECTS OF MAGNETIC FIELDS ON FEEBLE MAGNETIC MATERIALS N. HIROTA, H. UETAKE, T TAKAYAMA, H. NAKAMURA, M. KURASHIGE, S.HARA, Z SAITO, I:IKEZOE, T ANDO, H. WADA, K. KITAZAWA

269

APPLICATION OF MAGNETIC LEVITATION TO PROCESSING OF DIAMAGNETIC MATERIALS I. MOGI, K. TAKAHASHI, S.AWAJI, K. WATANABE, M. MOTOKAWA

278

PROTEIN CRYSTAL GROWTH IN LOW GRAVITY PROVIDED BY A NEW TYPE OF SUPERCONDUCTINGMAGNET N.I. WAKAYAMA,D.C. YIN, Z TANIMOTO, M. FUJIWARA, K. HARATA. H. WADA

285

xi

Magnetic Anisotropy and Alignment ALIGNMENT OF SINGLE WALL CARBON NANOTUBES UNDER HIGH MAGNETIC FIELDS UTILIZING A SELFiiORGANIZING OF EPOXY MATRIX M.S. AL-HAIK, H. GARMESTANI, D.S. LI, M.Y:HUSSAINI, K. DAHMEN, R. TANNENBAUM

295

ENHANCEMENT OF NANO-MECHANICALPROPERTIES OF AN EPOXY PROCESSED UNDER HIGH MAGNETIC FIELDS M.S. AL-HAIK, H. GARMESTANI, D. LI, M.I:HUSSAINI, R. TANNENBAUM, K. DAHMEN

303

PROCESSING OF POLYMERS USING MAGNETIC FIELDS E.P DOUGLAS

310

PROCESSING OF POLYMERIC MATERIALS UNDER MAGNETIC FIELDS Z KIMURA,M. YAMATO

32 1

MAGNETIC FIELD CONTROL OF STRUCTURESAND PROPERTIES 330 OF DIAMAGNETIC MOLECULAR ASSEMBLIES I. OTSUKA, Z TAKAHASHI, K YAGUCHI, H. ABE, S. OZEKl MAGNETIC ALIGNMENT AND CRYSTALLIZATIONBEHAVIOR OF ISOTACTIC POLYSTYREm M. YAMATO, Z KIMURA

337

Other Topics INDUSTRIAL APPLICATIONS OF MAGNETIC RESONANCE M.J. HENNESSY

347

GENERATION OF UNIFORM MAGNETIC FORCE FIELDS 0. OZAKI, S. MATSUMOTO, Z KIYOSHI, H. WADA

352

xii

EFFECT OF MAGNETIC FIELDS ON EXPLOSIVE WELDING OF METALS AND EXPLOSIVE COMPACTION OF POWDERS G.A. SHVETSOV VI. MALI, YU.L. BASHKATOK A.G. ANISIMOV A.D. MATROSOV T.S. TESLENKO

360

AUTHOR INDEX

37 1

Texturing and Phase Transitions


APPLICATION OF HIGH MAGNETIC FIELDS IN MATERIALS PROCESSING S . ASAI Dept. of Materials Processing Engineering, Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku,Nagoya, 464-8603. Japan The history of Electromagnetic Processing of Materials (EPM)is described. The application of a high magnetic field in EPM is classified and then the two topics of quantitative evaluation of phase transformation and texture alignment of ceramics are introduced, which are our recent endeavors relating EPM to a high magnetic field.

1. Introduction In the metal industry, electrical energy has long been used as heat energy due to its cleanliness, high controllability and high energy density. Technologies using electric energy were developed rather early and progressed without a background of sufficient scientific understanding. Good examples of this- are electromagnetic levitation and electromagnetic mixing that were invented in 1923 and 1932, respectively. To bridge the gap between technology and scientific understanding, magnetohydrodynamics, which had been established by Alfven in 1942, was first intioduced in 1982 at the IUTAM Conference titled “The Application of Magnetohydrodynamics to Metallurgy”, held in Cambridge, England. The conference may have introduced many people to the field of Electromagnetic Processing of Materials (EPM), though the term “EPM” was first formally used at the initial Symposium of EPM held in Nagoya, Japan in 1994. EPM research has been hitherto devoted to the economics of mass production and nanotechnology in relation to high quality materials. Today, EPM involves both Lorentz and magnetic forces relating to high magnetic fields. Here, the application of a high magnetic field in EPM is classified and the two topics of quantitative evaluation of phase transformation and texture alignment of ceramics are introduced. These are our recent endeavors relating EPM to a high magnetic field. 2.

Application of a High Magnetic Field in EPM

The technology relating to crystal orientation, structure alignment and spin chemistry has emerged in EPM thanks to the development of superconducting technology. Now, a high magnetic field utilizing a large space is available even 3

4 in small-scale laboratories. Table 1 indicates the utilization of a high magnetic field in EPM. A high magnetic field allows induction of crystal orientation, i.e. structural alignment, even in non-magnetic materials. There are four necessary conditions for crystal orientation under the imposition of a magnetic field. The first is .that unit crystal cells of the material have a magnetic anisotropy. The second is that the magnetization energy should be larger than the thermal energy. The third is that materials should exist in the weak constraint medium in which a particle can rotate under a feeble magnetization force. The fourth is that each particle should be dispersed in the medium as a single crystal. Table 1 Utilization of a high static magnetic field in EPM Lorentz Force-----------Appearance of small electric current effect

-

MassTransport- - - - - - - - - - -Eliminatbn of inclusionsand surface defect

( x M B - V)B

Magnetizatior

I

(structualalignrnent phase transfomtion

Spin Chemistry----------Intermolecular cross-linking reaction

The possibility of magnetic transportation and magnetic rotation was examined under several processes, such as solidification [ 1-41, electro-deposition [ 5 ] , vapor-deposition [6-81, and solid phase reaction [9]. The application of a high magnetic field has now been proven to be a promising method in EPM. Strengthening carbon fibers by imposing a high magnetic field is an example of spin chemistry. Carbon fibers produced from PAN (polyacrylonitrile) as a precursor are generally subjected to the three heat treatment processes of stabilization and carbonization, followed by graphitization. The carbon fibers produced from stabilized fibers in a magnetic field showed higher tensile strength than those produced without a magnetic field [9]. The fibers processed in a magnetic field have a larger crystallite size than those treated in no magnetic field. An intermolecular cross-linking reaction model [ 101 describes the crystallite size increase due to the imposition of a high

5

magnetic field. Regarding the spin chemistry [ll-161 upon which the background of this model stems, one can say that this study is the first attempt to link EPM with spin chemistry. The development of materials processing based on the spin chemistry is considered to be a promising area in EPM.

3. Quantitative Evaluation of Phase Transformation A new apparatus has been developed that can continuously measure the magnetic force during phase transformation. The magnetic susceptibility is calculated from the magnetic force obtained by using the apparatus, then the transitional solid fraction during the solidifying and melting processes is evaluated from the magnetic susceptibility. The magnetic susceptibility was measured using the Gouy method [17,18] and is evaluated by measuring a magnetic force F,.

Figure 1. Calculation of solid fraction.

The solid fraction in a solid-liquid mixed phase can be calculated from the observed magnetic susceptibility, as follows. From Figure 1, which displays the relationship between magnetic susceptibility and temperature, the magnetic susceptibilities of solid and liquid phases can both be expressed with good approximation by linear functions of the temperature around the melting point. Specifically, the magnetic susceptibilities in the single solid and liquid phases are given by Equations ( 2 ) and (3).

6

xms=C,lT+Cs2

(3)

The magnetic susceptibility of a liquid and solid mixture is given by Equation

(4).

xm=

f lxmlf

f sxms

(4)

In addition, Equation (5) holds that

f1ffs=l

(5)

Then

f, = x m xm

-xld

-Xd

Once the magnetic susceptibility and temperature of a mixture are measured, the solid fraction f, can be derived from Equations (2), (3) and (6). The relation between the solid fraction and temperature in the cooling process is shown in Figure 2. The solid phase of about 20mass%precipitated until the moment that the re-coalescence had finished and the temperature had recovered to the melting point. The liquid phase of about 20mass%remained even after the temperature descended below the melting point, i.e. about 2Omass% of melt was supercooled. 1 Q9 Q8

3 e Q7 .IQ6 z

E

Q5 Q4

z* a3

Q2 Q1 0

2 4 0 2 5 0 m 2 7 0 2 8 0 2 9 0 Temperature&) ( Sample:Bi, under Ar atmosphere )

Figure 2. The relation between temperature and solid fraction (cooling).

7

The method developed here can be applied to the direct observation of various phase transformation phenomena in solid, liquid and gas phases, therefore we hope it will lead to a better understanding of various phase transformations and reactions.

Rotatingof a crucible

Figure 3. Schematic view of the experimental apparatus used for rotating a crucible under a magnetic field.

4. A Novel Process to Fabricate Highly Textured Ceramics in a High Magnetic Field A novel process, where a specimen is rotated during a slip casting under a high magnetic field, has been proposed in order to fabricate highly textured ceramics. The usefulness of the newly proposed process has been confirmed in Si3N4 ceramics. Figure 3 shows the schematic of the experimental apparatus using rotation. In order to examine the effect of rotation, green samples were prepared in a magnetic field and without rotation. For comparison, a sample was also prepared with no magnetic field. After drying, the green samples were embedded in a 60wt%Si3N4 +40 wt% BN powder bed in a graphite crucible and heated to 1800°C for 1.5 hours in an N2 atmosphere, with no magnetic field. Figure 4 schematically shows the functions of the magnetic field and the rotation. In the substance in which the magnetic susceptibility in the a, b axis is higher than that in the c-axis, xc c X,,b, a one-directional crystal orientation can not be obtained in a slip casting under a high magnetic field, as the free choice of crystal orientation exists in the a, b axis. When the magnetic field is imposed on the suspension, the c-axis of the particles can align perpendicular to the magnetic field. The condition where the specimen is rotated in the magnetic field is equivalent to the case where the specimen is fixed and the magnetic field

8

Rotating of a crucible

Figure 4. Schematic view showing the functions of the magnetic field and rotation of a crucible in a magnetic field.

Figure 5. SEM micrographs of specimens made of a-SijN4 powder with p-SijN4 seeds: (a), (b) without magnetic field; (c), (d) with magnetic field under crucible rotation.

is rotated. In this case, the c-axis of the particles will be perpendicular to the plane in which the magnetic field rotates. Thus, the c-axis of the particles aligns parallel to the direction of gravity. Figure 5 shows the SEM micrograph of the polished surfaces of the specimen. Also, 0- Si3N4rod grains appear randomly distributed in the specimen prepared without exposure to a magnetic field (Figure 5. a and b). When the specimens are prepared by rotation under a

9

magnetic field, a highly textured material can be obtained as shown in Figure 5(c) and (d).

5.

Conclusion

The history of Electromagnetic Processing of Materials (EPM) has been described. The application of a high magnetic field in EPM has been classified and listed, and the two topics of quantitative evaluation of phase transformation and texture alignment of ceramics have been introduced, which are our recent endeavors relating EPM to a high magnetic field.

Acknowledgement This work was supported in part by the 21st Century COE Program “Nature-Guided Materials Processing” of the Ministry of Education, Culture, Sports, Science and Technology.

References 1. Morikawa, H., Sassa, K., and Asai, S., Muter. Trans., JZM, 39, (1998) pp. 814-818. 2. Yasuda, H., Tokieda, K., and Ohnaka, I., Muter. Trans., JIM, 41, (2000) pp. 1005-1021. 3. Legrand, B.A., Chateigner, D., Pemer de la Bathie, R., and Tournier, R., Journal of Magnetism and Magnetic Materials, 173, (1997) pp. 20-28. 4. Noudem, J.G., Beille, J., Bourgault, D., Chateigner, D., and Tournier, R., Physica C, 264, (1996) pp. 325-330. 5. Taniguchi, T., Sassa, K. and Asai, S., Muter. Trans., JIM, 41, (2000) pp. 981-984. 6. Mitani, S., Bai, H.L., Wang, Z.J., Fujimori, H., and Motokawa, M., The 3rd International Symposium on Electromagnetic Processing of Materials, Japan, ISIJ, (2000) pp. 630-634. 7. Tahashi, M., Sassa, K., Hirabayashi, I., and Asai, S., Muter. Trans., JIM, 41, (2000) pp. 985-990. 8. Awaji, S., Watanabe, K., Ma, Y., and Motokawa, M., Physica B, 294-295, (2001) pp. 482-485. 9. Ito, M., Sassa, K, Doyama, M., Yamada, S., and Asai, S., TANSO, 191, (2000), pp. 37-41. lO.Sung, M.G., Sassa, K., Ogawa, H., Tanimoto, Y., and Asai, S., Muter. Trans. 43 (2001), pp. 2087-2091. ll.Tanimoto, Y., Hayashi, H., Nagakura, S., Sakuragi, H., and Tokumaru, K., Chem. Phys. Lett. 41 (1976) p. 267. 12.Hatta, N., Chem. (1976) p. 547.

10

13.Shulte1-1, K., Staerk, H., Weller, A., Werner, H.J., and Nickel, B., Z. Physik. Chem. NF 101( 1976) p. 37 1. 14.Michel-Beyerle, M.E., Haberkorn, R., Bube, W., Steffens, E., Schroder, H., Neusser, H.J., Schlag, E.W., and Seidlitz, H., Chem. Phys. 17 (1976) p. 139. 15.Ihara, I., Kato, M., Kanamori, I., Nakamura, K., Shimada, E., and Watanabe, T., Symposium on New Magneto-science 2002, Proceedings of the 6" Meeting, Nov. 2002, Tsukuba, Japan, pp. 298-302. 16.Asai, S.,Koumoto, K., Matsushita, Y., Yashima, E., Morinaga, M., Takeda, K., Iritani, E., Tagawa, T., Tanahashi, M., Miyazawa, K., Science and Technology of Advanced Materials, 4 (2003) to be published. 17.Iguchi, Y., Experimental chemical course 9, Maruzen Ltd. (1991) pp. 439-450. 18.Suzuki, N., Metal data book, Japan Metal Institute, Maruzen Ltd. (1974) pp. 10, 18.

HIGH MAGNETIC FIELDS EFFECTS ON SOLID STATE TRANSFORMATIONS AT HIGH TEMPERATURE E. BEAUGNON, F. GAUCHERAND CNRSKRETA-LdC, Grenoble, France Annealing in a high magnetic field of asquenched Co-B near-eutectic alloy promotes solid-state anisotropic growth of ferromagnetic Co particles along the magnetic field. Competition between surface energy and demagnetizing energy or effect of the magnetic torque and creeping of the matrix can both qualitatively explain the observed alignment. However, both mechanisms are in contradiction with the non-saturation of the phenomenon in fields up to 16 Tesla. The effect of the magnetic force on the diffusion near a ferromagneticparticle is discussed.

1. Introduction In recent years, high magnetic fields have been widely used in material processing experiments, particularly in magnetic texturing from a solidifying melt where the residual magnetic anisotropy of solidification nuclei allows their magnetic alignment in the liquid phase [1,2]. The aim of this work is to study strong magnetic field effect on texturing, but during solid-state transformations. In this case, no fluid phase can allow the free rotation of anisotropic particles. However, it is expected that slow diffusion processes can also lead to anisotropic textures. Magnetic field effect on diffusivity has been tested in a few experiments. Youdelis et al. [3], from experimental results in the Al-Cu system, discussed the inhibition of diffusion perpendicular to an applied field due to Lorentz forces on diffusion-transported ions and electrons. In the A1-Cu system, the Cu diffusion was reduced by 25 % in a magnetic field of 3 Tesla. In contrast, Nakajima et al. [4] could not observe any significant effect on the diffusion of Ni in Ti, in fields up to 4 Tesla. In the austenite to ferrite transformations in steels, Xu et al. [5] measured a magnetic-field-dependent parabolic growth rate, but, depending on the temperature, the magnetic field effect could be either an inhibition or an enhancement of the diffusivity. Strong texturing effects have also already been observed in clean Bi bicrystals, where the growth of the grain with the axis of larger magnetic susceptibility is promoted at the expense of the other grain [6]. In steels, Ohtsuka et a1 observed a shape alignment of large ferrite grains after the austenite to ferrite transformations in high magnetic field. 11

12 In this study, the magnetic texturing of Co precipitates in a Co-B eutectic is studied. The high Curie temperature of the Co phase allows annealing temperatures close to the melting point to promote diffusion, but still in the ferromagnetic state to enhance the magnetic field effects on the precipitates. 2.

Experiments

Cylindrical samples of Co-B near-eutectic alloys, containing 18.5 B at% [7], were prepared by induction melting in a cold crucible and cast in a cold copper mold. As seen by SEM images, the as-quenched alloys exhibit a fine, submicron, lamellar eutectic structure. Parallel lamellas are organized in domains, but no net orientation of the many domains could be observed. Samples were then annealed at 900 "C for 65 h in 0 Tesla, 7 Tesla and 16 Tesla. Polished surfaces cut in a plane along the field direction were observed by optical microscopy. After annealing, the microstructure strongly differed from that of the as-quenched alloy: Cobalt ferromagnetic particles coalesced to form large (several microns) globular particles with random shapes dispersed in an homogeneous matrix which is assumed to be the paramagnetic Co2B phase. Images were analyzed using a PC version of the free software NlHimage (Psion Image from Psion Corp.). In this analysis, each Co particle is fitted as an ellipsoid; the ellipsoid size and orientation versus applied field is then measured on several thousands of particles for statistical analysis of the magnetic orientation.

3. Results on Magnetic Alignment The particle orientation distribution after annealing is presented in Figure la, b and c. The number of particles is plotted versus the angle between the particle long axis and the magnetic field (when applied) or the vertical direction (at 900). In zero magnetic field, although no orientation could be observed on the ascast lamellar eutectic, a large anisotropy exists in the annealed samples where more particles are aligned near the horizontal plane. This result suggests that some texture already existed prior to the heat' treatment: the strong radial temperature gradient induced in the cylindrical sample during the rapid solidification in the cold mold promoted a radial growth of the eutectic. In 7 Tesla, the radial structure still remains but a large alignment parallel to the field is observed. The alignment effect is even more obvious in the sample submitted to 16 Tesla, where almost all of the radial alignment is erased and only the magnetic alignment at 90" prevails.

13

0

30

60

90

120

150

180

Argyle Figure la. Particle counts versus angle for a zero magnetic field annealing.

0

30

60

90

120

Angle

150

la0

0

30

60

90

120

150

1BO

Angle

Figure Ib (left) and Ic (right). Magnetic field effect on the particle angle distribution after annealing.

An order parameter has been defined as the number of particles whose long axis is closer to the field direction divided by the number of particles whose long axis is closer to the perpendicular direction. A random distribution would give 1 and, for an axial distribution along the field direction, a value larger than 1. In zero field, the value is 0.7, revealing the radial texture due to fast solidification of the cast alloy. In 7 Tesla, it reaches 1.2 and rises to 1.95 with 16 Tesla: no saturation of the phenomenon is observed above a few Tesla. Several models can qualitatively account for the magnetic texturing in the solid state and are discussed in view of this non-saturation. 4.

Competition Between Surface Energy and Demagnetizing Energy

The minimization of the interfacial energy of the Co particle in the surrounding matrix, obtained for a spherical shape, competes with the demagnetizing energy, which is a minimum for an elongated shape along the field. For an ellipsoid along the field direction, the equilibrium shape is defined by the minimization of

14 the total energy E as a function of cL=c/a, the shape factor of the ellipsoid, for a constant volume V. The total energy E is given by: E = oS+'/z~nM2Vwith S the particle surface, M its magnetization, (T the surface energy and n the demagnetizing factor. For a given average radius R, and an elongated ellipsoid, n and S are given by: 1 a*-1

J

argch(a) - 1 JZ

The demagnetizing factor n varies from 1 for a flat shape to 0 for a needle. Below the saturation, the magnetization M in an applied field Ba is governed by the demagnetizing field with hM=Ba/n. h M is about 1 Tesla at high temperature, so that the saturation is at least reached from Ba = 1 T, whatever the n value. The surface energy is unknown, but typical values are in the range of 0.1 - 1 J/m2 for metals. The observed average radius of the particle is in the range of 4 - 10 pm. Several examples of the total energy are given Figure 2. Jhm3

4 p m 0.1 J/m2

Jhi?

4 p m 0.5 Jjm'

6MoW

\

625000 M)oo00

575000

y.Il_. 525000

2

4

6

8

i

O

8 p r n 0.5 J/mz

Jhi? 31GQW 330000 320000 310000

moo 290000

I

2-

6

8

10

Figure 2: sum of the demagnetizing energy and total surface energy for different average radius R and surface energy o.

15

In each case, it is found that the Co particle equilibrium shape is noticeably elongated along the field. However, the saturation magnetization is already reached at 1 Tesla, so that no more effect should have been observed in 16 Tesla as compared to 7 Tesla. In addition, this simple model does not account for particles that are already elongated in random directions as observed in the samples. 5. Magnetic Torque As most of the particles have an anisotropic shape, a magnetic torque is developed in the magnetic field. It can then be expected that, for a long experiment time, the slow creeping of the surrounding matrix will allow the rotation of the anisotropic Co particles. In a low external field, Ba, below a threshold field, Ba*, where the magnetization saturates, M is governed by the demagnetizing field, so that the inner field (Ba - k n M ) is zero, with n the demagnetizing factor along the long axis of the particle. Let I$ be the angle between the applied field and the major axis of the particle, then the magnetization, M, is set at a constant angle 0 with Ha so that 0 < 4, with:

Ba < Ba* = p,Msl,/a d p,,M = B

and tan(6) = tan(cp)E is constant

When M saturates to Ms, the demagnetizing field can no longer cancel the external field. The magnetization is no longer locked at a constant angle but rotates to align with Ha, so that 0 is now given by:

1-3n Ms sin(cp - 0) = -sin(26)4 Ha As Ba is increased, and because 0 goes to zero, the magnetic torque

r = (MAB)V

has a finite limiting value r(Ba=-):

r(Ba = -) = poM s 2 Vysin(2cp) For any n values and for different angles I$, it is then found that the limiting torque value is rapidly approached below a few Tesla, thus the experimental high field dependence of the magnetic orientation is not taken into account in this torque model.

16

6. Curvature Effects and Local Magnetic Forces The preceding mechanisms involve a particle transformation (shape factor or angle versus field) at a constant volume. In the actual annealing experiment, the particles are growing from a sub-micron lamellar eutectic to a globular distribution with a typical particle size of a few microns. The mechanism by which particles coalesce is the Ostwald ripening where large particles are growing at the expense of the smaller, which dissolve. The driving force of the diffusion from small to large particles is the shift of the concentration equilibrium in the matrix at the interface with the particle, due to curvature effects. For a mean curvature (Ur), this concentration shift is given by the Gibbs-Thomson equation:

with C the concentration and AC the concentration shift of the Co ions in the surrounding matrix at the interface with the ferromagnetic Co particles. y is the interfacial energy and 52 the molar volume. Near a ferromagnetic particle, the local magnetic field is distorted with a maximum value at the vertical poles (along the external field direction) and a minimum value at the equatorial plane. Strong local magnetic gradients exert forces on the diffusing Co ions in the surrounding matrix (see Figure 3), but the thermal energy kT almost counterbalances the effect. The equilibrium concentration profile then obeys a Boltzman distribution and is proportional to x , ~is the magnetic susceptibility per single Co e x p ( ~ , ~ B ~ / 2 k kwhere T) paramagnetic ion.

.-... -.-...< \ * , * , , , , , , - - . < . . 9 , , , , . , , , . - -

C

--..,,.l.I.1\......~ _ _ , , , , l l l l L . \ , , . . - -

.-*.,,,,...,..,..--Figure 3: schematic magram of the magnetic forces near a sphencal particle (left) and local effect on the Co ion concentration in the surroundmgmatnx (right).

A first order development leads to the concentration variation AC that then follows:

17

-_AC C

(x.1

Ba)AB kT

1 0

- ( x d Ba)AB 1 0

RT

xmol

with Ba the applied field, the molar susceptibility and AB the local field variation. Near the magnetic poles of the particles, a small radius rtipcan be stabilized by the magnetic force that increases the local concentration in the matrix and prevents dissolution. On the other hand, near the equatorial plane, the magnetic force expels ions except in the case of a very elongated particle where the local distortion field is nearly zero. A stable configuration, considering both Gibbs-Thomson and magnetic effect, is then a particle elongated along the field direction, with a radius rtip at the poles for which both effects are equal. With the S.I. susceptibility of the ions in the matrix and Ms the magnetization of the particle, the radius r,ipis then given by:

x

x

For y = 0.1 J/m2, Ba = 16 Tesla, p&ls = 1.1 Tesla, and = 7.5 (a value extrapolated from the susceptibility measured in the liquid state [8]), one finds rtip= 1.5 pm, a quite realistic value for the particles actually observed. 7.

Conclusion

The application of a high static magnetic field (up to 16 Tesla) on the high temperature solid-state annealing of a Co-B quenched eutectic alloy leads to an anisotropic microstructure where ferromagnetic Co particles are statistically aligned along the field direction. Two models were proposed to explain this result: - the particles elongate along the field to minimize their demagnetizing field; - the particles rotate to align their longer axis along the field. In both cases, transformations occur by solid-state diffusion, and the magnetic field effect should saturate with the maximum magnetization of the ferromagnetic Co particles. It is experimentally demonstrated that, up to 16 Tesla, no saturation is measured, in contradiction with both models. A new model is proposed, based on the strong local magnetic forces around a ferromagnetic particle that polarize the diffusion and modify the Ostwald ripening classical scheme. In this model, growth near the magnetic poles is promoted and local magnetic attraction prevents the dissolution of the tip of the particle so that a final, stable elongated shape along the field can be obtained.

18 Further investigations are required to develop this new model, but the first estimations of the magnetic force effects are consistent with experimental results.

References 1. 2. 3. 4. 5. 6. 7. 8.

P. de Rango et al, Nature, (1991), p. 349. E. Beaugnon et al, Journal de Physique I, 3 , 2 (1993). Youdelis et al., Canadian Journal ofphysics, 48, (1970) and 42 (1964). Nakajima et al, Trans. JIM, 26, 1 (1985). Xu et al., Trans. MRS Japan, 25,2 (2000). Molodov et al., Sriptu Muterialla, 37, 8 (1997). S. Omori et al., Trans. JIM, 17 (1976). F. Gaucherand, PhD thesis, Joseph Fourier University (2001).

ENHANCEMENT OF MATERIAL PROPERTIES BY MAGNETIC FIELD ASSISTED PHASE TRANSFORMATION B.Z. CUI

'**, K. HAN ',H. GARMESTANI

H.J. SCHNEIDER-MUNTAU J.H. SU J.P.LIU *

'National High Magnetic Field Laboratory, Florida State University, Tallahassee, FL 32310 'Department of Physics, Universig of Texas at Arlington, Arlington, TX 76019 3School of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, GA 30332

Using magnetic field assisted phase transformation, an enhancement of the exchange coupling and hard magnetic properties of melt-spun Nd~FellBla-Fe-typenanocomposites was achieved by optimizing their nanosmcture and morphology. Compared with the Nddr5.6DylFeuMolBs sample annealed without a magnetic field, the magnetic annealing results in a noticeable improvement in the coercivity i&, the remanence 4nM, and energy product (BH),. for Nd2.&'rs.~DylFe~MolB6 alloys. (BH)mu at 50 K was enhanced by 43.7% after magnetic annealing in a 19 T field. The kink in the demagnetization curve disappeared and, additionally, a much better squareness of the demagnetizationcurves was observed in the magnetically annealed samples.

1. Introduction

Nanocomposite magnets have drawn extensive attention due to their great commercial potential and broad applications in nano-electromechanical systems, automatic control engineering, micro-mechanical devices, magnetic imaging, magnetic fluids, biomagnetic sensors, nanomedicine, catalysts and other applications. Nanocomposite magnets have extremely high theoretical energy products (BH),, of up to 100 MGOe. However, there is still a big discrepancy between the theoretical and experimental (BH),, [ 1-51. Some challenges still remain in the improvement of the naonostructure morphology and crystal texture, such as refinement of grain size, homogeneity of the grain size distribution, optimization of grain configuration, and hard nanograin alignment. A breakthrough in (BH),, could be achieved by preparing homogenized and textured nanostructures to take advantage of the magnetic anisotropy of hard phases and exchange coupling between the soft and hard nanograins [4,5].The present work reports an approach to achieve nanostructural optimization and magnetic field-induced crystal texture through magnetic field assisted or controlled phase transformation of Nd2Fe14B/cl-Fe-type melt-spun Nd2.J'r5.,JIy1Fes4MolB6 nanocomposites. 19

20

2. Experimental Procedure Nd2.~r5,6Dy1Feg4MolB6 ribbons were prepared by melt spinning with a molybdenum wheel speed of 35 d s . The ribbons were annealed at 690" C for 20 min with and without an in-plane field of 19 T. The phase components, nanostructured morphology and crystallization behavior of the samples were studied by x-ray diffraction (XRD) using Cu Ka radiation, a JEOL-2010 transmission electron microscopy (TEM) and a PerkinElmer DSC7 differential scanning calorimeter (DSC) at a heating rate of 20" C h i n . The average grain sizes of a-Fe were deduced from Scherrer's method of XRD. The magnetic properties were measured by a Quantum Design SQUID magnetometer in fields up to 6.5 T. The ribbon plane was placed parallel to the magnetic field direction. No demagnetization correction was done for these ribbon samples.

3. Results and Discussions As indicated by XRD investigations, the as-spun Nd2.4Pr5.~y1Feg4Mo sample is a composite in which a small amount of a-Fe and (Nd,Pr,Dy)2(Fe,Mo)l4B (2: 14:1) nanocrystallines are embedded in an amorphous matrix. DSC study of the as-spun Nd2.4Pr5.6Dy1Fe84MolB6 alloy without a magnetic field shows that nucleation temperatures T,, of the a-Fe and 2:14:1 phases are 525" C and 590" C, respectively. Annealing at 690" C for 20 min leads to the formation of a mixture of nanostructured a-Fe and 2:14:1 phases. Figure 1 shows XRD patterns of the Nd2.4Pr5.6Dy1Feg4M01B6 ribbons annealed at 690°C for 20 min without and with a 19 T in-plane field. Compared with the sample annealed with a 19 T field, it can be seen that there are fewer XRD peaks of the 2: 14: 1 phase in the sample annealed without a field i.e., some peaks with relatively low intensity even disappear. The average grain sizes of aFe, calculated from the a-Fe (110) diffraction peak using the Scherrer formula are 17 nm and 20 nm, respectively, for the samples annealed with and without a 19 T in-plane field. The error bars for the average grain sizes are 10%. Figure 2 shows the TEM bright field images and corresponding selected area diffraction of the two samples shown in Figure 1. It can be seen that magnetic annealing introduces somewhat finer, less angular (fewer sharp edges) and more homogeneously distributed soft and hard nanograins.

21

I

25

- I1

1199 TT

0

30

35

40

45

0-2:14:1 t - a-Fe

50

2 8 (degree) Figure 1. XRD patterns of N ~ ~ ~ P ~ s ~ D ~samples ~ F ~ & annealed o I B at~ 690" C for 20 min without and with a 19 T in-plane field.

Figure 2(a). E M bright field image and corresponding selected area diffraction of Nd~Srs.6DylFe&lolB6sample annealed at 690" C for 20 min without a magnetic field.

Figure 2(b). TEM bright field image and Corresponding selected area diffraction of Ndz S r 5 6DylFes&O& sample annealed at 690" C for 20 min with a 19 T in-plane field.

22

Crystallization of amorphous solids is generally considered as a nucleation and growth process of the crystalline phases. In the nucleation theory the steady state homogeneous nucleation rate I,, is given by [6]

I,, = b*exp(-Q/RT)*exp(-AGJRT)

(1)

where b is a pre-exponential factor, Q is the activation energy for the transfer of atoms across the surface of the nucleus, which is approximately equal to the diffusion activation energy, and AG, is the free energy required to form a nucleus of the critical size. AG, can be written as

AG= ~ A?/(AG:).

(2)

where AG, is the Gibbs free energy difference between the crystal and the matrix amorphous phase, ie. AG, = Gcrysa- G m o r p ~cu s0, A is a coefficient and y is the interfacial energy of the crystaVamorphous interface. In the present work, Gcvsa can be the Gibbs free energy G,-F~or G2.4:. for the a-Fe and 2:14:1 phases, respectively. The amorphous matrix is magnetically soft with Curie temperature T, of 310" C and coercivity of 0.6 kOe at room temperature, which is paramagnetic during the crystallization process. During the crystallization process and subsequent isothermal anneal at 690"C, the newly formed nanocrystallines are a ferromagnetic a-Fe phase (T, = 771" C) and a paramagnetic 2:14:1 phase (T, = 325" C). When a magnetic field is applied to the system, the Gibbs free energy Ga-~e of the a-Fe nanocrystallines decreases due to the addition of the magnetostatic energy. As the new phase a-Fe is still a ferromagnetic phase and the matrix is paramagnetic, the extra Gibbs free energy difference AGHintroduced by a magnetic field H is [7]

AG" = -@M(T).H

- 1/2*~*H'+ E , * ( & ~ ~ H ) * H * B

(3)

The first term, -kAM(T)*H, represents the energy due to the magnetostatic effect, where AM(T) is the difference in magnetization between the a-qe and amorphous matrix at certain temperature. The terms - 1 / 2 * ~ * H and &,*(dw/dH)*H*B represent the energies due to the high field susceptibility and forced volume magnetostriction effects, respectively, where x is the high field susceptibility in the matrix, E, the volume change associated with the phase transformation, o the force volume magnetostriction, and B the matrix bulk modulus. It follows that the presence of H increases the absolute value of AG,, resulting in an increase of the driving force for crystallization and further the nucleation rate I,, of a-Fe. The above analysis also indicates that the higher the

23

magnetic field, the greater is the I,, for a-Fe. Therefore, magnetic annealing promotes crystallization of the amorphous matrix and allows more nucleation centers to form and grow the soft a-Fe phase during the crystallization process. It was reported that the nucleation rate of ferromagnetic ferrite was remarkably accelerated by three times under the influence of a 10 T field in the Fe-C alloy. However, the growth rate is nearly the same with and without a 10 T field, so the magnetic field has little effect on the atomic jump frequency and the activation energy for the crystal growth of ferrite [S]. Compared to annealing without the magnetic field, magnetic annealing leads to reduced a-Fe grain sizes and more uniform distribution of the grains for both the hard and soft phases, as is shown by the above-mentioned experimental results. Similar results were also observed in the melt-spun Nd2Fe14B/Fe3B-typenanocomposites [9]. Alternately, the Gibbs free energy G2:.k1 of the 2:14:1 phase should also decrease slightly, so the nucleation rate I,, of 2:14:1 phase should increase marginally in the presence of a high magnetic field. But these changes are much smaller than those of a-Fe as the 2:14:1 phase is in a paramagnetic state during the crystallization process. However, from the shape and numbers of the XRD peaks of the 2:14:1 phase (Fig. I), it is clear that magnetic annealing also promotes crystallization of the amorphous matrix to form the 2: 14:1 phase as the magnetic field is as high as 19 T. Figure 3 gives the dependence of the intrinsic coercivity iH,, the maximum magnetic energy product (BH),, and the remanence 4nMr on temperatures for Nd2.$r5.&~1Fe84Mo$6 annealed at 690" C for 20 min without and with a field of 19 T. Figure 4 shows the representative room temperature demagnetization curves of the two samples of Figure 3. It is observed from Figures 3 and 4 that, compared with the sample annealed without a magnetic field, there is a noticeable improvement in i&, (BH), and 4nM, for magnetically annealed Nd2.$r5.&y1Fe84MolB6 alloys. Especially, (BH)max at 50 K and 300 K is enhanced by 43.7% and 35.7%, respectively, after magnetic annealing in a field of 19 T. The kink in the demagnetization curve disappears and, in addition, a much better squareness of the demagnetization curves is observed with magnetic annealing.

24

--e no

field

-19T 9.0 50

LOO

150

200

250

300

T 6)

Figure 3. Dependence of iK, (BH),,, and 4nM, on temperatures for Nd2.4Pr5.6DylFe~Mo~B6 annealed at 690" C for 20 min without and with a 19 T field.

-0-

690°C, 20 min, 19 T

+69OoC,20 min, no fielc

10 n v1

8 E 5

e

d

0

-6

-4

-2

0

2

4

H (kOe) Figure 4. Room temperature demagnetization curves of the two samples of Figure 3.

25 The exchange coupling interactions between soft and hard phases were evaluated by using 6M (Henkel) plots, in which 6M = w(H)- [l - 2m(H)1 [lo]; where a(€€) is M,(H)/M,, the reduced isothermal remanence, and m(H) is &(H)/M,, the reduced'demagnetization remanence. Figure 5 shows 6M plots as a function of the applied field for the two samples of Figure 3. An initially positive 6M is observed in both annealed samples with and without a 19 T field, confirming the existence of ferromagnetic exchange coupling interactions between the 2:14:1 and a-Fe nanograins. The magnitude of the 6M peak of the magnetically annealed sample is higher than that of the sample annealed without a magnetic field, demonstrating stronger exchange coupling interactions between the soft and hard nanograins in the magnetically annealed sample. We further estimate the exchange coupling in the two samples of Figure 3 by spinreorientation temperature changes of the 2:14:1 phase. Figure 6 shows the thermomagnetic curves for the same two samples as shown in Figure 3. As can be seen in Figure 6, the spin-reorientation temperatures T,, of the hard phase 2:14:1 in the samples annealed with and without a field of 19 T are 55 K and 61 K, respectively. The inter-gain exchange coupling is known to lead to a rotation of the magnetization in neighboring grains away from the local easy direction and to hinder spin reorientation. The stronger the inter-grains exchange coupling interactions between the nanograins, the lower is T,, [11,12]. In this experiment, a lower Tsr in the sample annealed with a 19 T field convincingly demonstrates that magnetic annealing leads to a stronger inter-grain exchange coupling. The improvement of the hard magnetic properties of the magnetically annealed sample results mainly from the magnetic-field-induced enhanced exchange coupling, which is due to optimization of the nanostructured morphology by magnetic annealing, such as reduced grain sizes of a-Fe and the more uniform distribution of the grains for both the hard and soft phases. Additionally, magnetic annealing improves the crystal quality of the hard 2:14:1 phase (Figure l), which is also favorable for the magnetic hardening for nanocomposite magnets.

26 I

1

0.2

E

I-\

0.0

-

-.-0--8=.== /o-

0

5

15

10

20

25

H (kOe) Figure 5. 6M plots as a function of the applied field for the two samples of Figure 3.

0

50

100

150

200

250

300

T (K) Figure 6. Thennomagnetic curves in an applied field of 0.1 T for the two samples of Figure 3.

It is known that a significant improvement in (BH),, is expected from successfully fabricated anisotropic nanocomposites with crystallographic textures of hard phases [4,5]. Magnetic annealing in a high field is a promising method for manufacturing textured naomagnetic materials. If the matrix phase is a liquid, the new phase can move freely and may align along the magnetic field due to its magnetocrystalline anisotropy. In this case, to align a paramagnetic or ferromagnetic particle, the driving force (magnetic torque) due to magnetocrystalline anisotropy of the new phases in a paramagnetic or

27 ferromagnetic state could still be larger at high magnetic fields than those due to the shape anisotropy and the thermal disordering effects [ 131. In these cases, a hard phase-textured nanostructure has lower energy than a random structure in a magnetic field. The magnetically induced crystallographic alignment of the new magnetic phases (such as 2:14:1 and fct FePt etc.) driven by a magnetocrystalline anisotropy can, therefore, still be successfully achieved during the magnetic annealing if the applied magnetic field strength is high enough [ 131. If the matrix phase is a solid, the magnetic field must overcome the interface energy term in order to align the new phase as well as the energy associated with the shape anisotropy and the thermal disordering effects. The alignment of the new phase can be achieved by changing the habit plane of the matrix and the new phase in the presence of a high magnetic field. Such a change of orientation is difficult but possible, especially if the new phase is still in the ferromagnetic state during the magnetic field assisted solid-state phase transformation. Further research is underway toward achieving a breakthrough in (BH),, by preparation of homogeneously nanostructured and, especially, highly textured nanocomposites.

4. Conclusions For Nd2Fe14Bla-Fe-type melt-spun Nd2.&5.&~~Fe8~Mo$6 nanocomposites, magnetic field assisted phase transformation from amorphous matrix to a composite of 2: 14: 1 and a-Fe nanocrystallines results in an optimization of the nanostructure through promotion of the crystallization of the amorphous matrix and the generation of more nucleation sites. The optimization of the nanostructure morphology leads to an enhanced exchange coupling and a noticeable improvement in the hard magnetic properties for the Nd2,J’r~.6Dy1Fes4Mo~B6 samples. Acknowledgments This work was supported by DARPA through AEtO under grant DAAD19-03-10038. Dr. B. Z. Cui thanks Prof. N. Dalal at the Department of Chemistry at Florida State University for the partial financial support. References 1. Manaf, R.A. Buckley, H.A. Davies, J., Magn. Magn. Mater., 128, (1993) p. 302. 2. Cui, B.Z., Sun, X.K., Xiong, L.Y., Liu, W., Zhang, Z.D., Yang, Z.Q., Wang, A.M. and Deng, J.N., J. Muter. Res. 16, (2001) p. 709.

28

3. Jurczyk, M., Collocott, S.J., Dunlop, J.B. and Gwan, P.B., J . Phys. D: Appl. Phys. 29, (1996) p. 2284. 4. Skornski, R. and Coey, J.M.D., Phys. Rev. B, 48, (1993) p.15812. 5. Fischer, R., Schrefl, T., Kronrniiller, H. and Fiddler, J., J. Mugn. Mugn. Muter. 150, (1995) p. 329. 6. Lu, K., Muter. Sci. & Eng. R, 16, (1996) p. 161. 7. T. Kakeshita, T. Saburi, and K. Shimizu, Muter. Sci. & Eng. A , 21 (1996) pp. 273-275. 8. Ohtsuka, H., Hao, X.J., and Wada, H., International Workshop on Materials Analysis and Processing in Magnetic Fields Tallahassee, FL, March 2004, to be published. 9. Zhao, T.M., Hao, Y.Y., Xu, X.R., Yang, Y.S. and Hu, Z.Q., J . A p p l . P h y s . 85, (1999) p.518. 10.Henke1, O., Phys. Stat. Sol., 7 (1964) p. 919. ll.Kou, X.C., Dahlgren, M., Grossinger, R. and Wiesinger, G., Jour. Appl. Phys. 81, (1997) p. 4428. 12.Cui, B.Z., Sun, X.K., Xiong, L.Y., Tang, S.T., Zhang, X.X., Liu, W., Geng, D.Y. and Zhang, Z.D., J. Alloys and Compounds 340, (2002) p. 242. 13.Courtois, P., de la Bgthie, R.P., and Tournier, R., J. Mugn. Mugn. Muter. 153, (1996) p. 224.

EXPERIMENTAL INVESTIGATION OF THE CRYSTALLIZATION OF BHF IN HIGH MAGNETIC FIELDS

w. ERTEL-INGRISCH’, K. HARTMAN”, x.WANG’, D. WLSENBERG’ ‘Junior Research Group “Electromagnetic Processing of Materials” 2Departmentfor Glass and Ceramic Technology Technische Universitat of Ilmenau, Ilmenau, Germany Ba-Hexafemte (BHF) powder is a hard magnetic material of wide t e c h c a l application. High-end applications require a very homogeneous nanometer-scale single domain BHF powder with optimum magnetic properties that are mainly controlled by the conditions (temperature, time) during its formation. One method to synthesize BHF nanocrystalline powder with satisfying magnetic properties is the glass crystallizationtechnique [ l ] - [3]. This method starts from melts prepared in the BaO-Fe~03-B~O3-system.which are homogenized, and quenched to glass flakes applying a double-roller rapid quenching technique. Obtained glass flakes are processed in a subsequent tempering process crystallizing very homogeneous nanometer-size BHF powder. A wide variety of analytical techniques is applied to monitor chemical composition (electron microprobe analysis, XRD), structural and physical properties of intermediate products (flakes) and obtained powders @EM, TEM, DTA, XRD, vibrating sample magnetometry). Precise control of the parameters and optimization of the process leads to BHF powders with a maximum coercivity of up to 400 W m . Application of high magnetic fields during the crystallization process of BHF is expected to result in a net increase of its magnetic properties. Therefore, experiments in a cryogen-free magnet (CFM) of up to 5 tesla magnetic flux density equipped with a hightemperature oven to precisely control process parameters will be performed to investigate the influence of a strong magnetic field on its crystallizationprocess. Of special interest is the complex question whether magnetic fields can help control and improve the single domain crystal structure by influencing nucleation, domain growth and orientation of growing domains towards each other. Initial experimental results indicate a fundamental change in the crystallizationprocess resulting in further optimized magnetic properties of BHF powders when crystallizationis performed in a strong magnetic field.

1. Introduction Barium-Hexaferrite (BHF) powder is one of the most important hard magnetic materials used in industry. Its technical applications range from basic use as permanent magnet in daily life to high-end technological applications as analog and digital high-density magnetic recording media, in electronic computation devices, and generally in electric drive technology. Especially for high-end applications, very homogeneous sub-micrometerscale single domain BHF powder is required that can be produced by applying a glass crystallization technique (GCT [ I]-[3]). However, starting from the 29

30

ternary system BaO-Fe~03-Bz03, or in technical respect, BaC03 and FezO3, compositions with mole fractions of Fez03 higher than 20% tend to crystallize spontaneously into numerous phases of vastly varying magnetic properties. Therefore, precise knowledge of the influence of the various parameters controlling the crystallization of BHF powders and their precise control during production is essential to guarantee both right chemical composition and structure as well as optimum magnetic properties of the final product. Magnetic properties are, of course, controlled on one hand by the corresponding chemical starting composition to obtain the intended chemical phases and their magnetic properties. However, the right thermal treatment has a decisive influence on the initiation of any crystallization due to the formation of crystallization nuclei. Nucleation is crucial for all subsequent processes, and decisive for the size and homogeneity of the final product. During an initial step, as many seeds should be created as possible while growth and crystallization must be avoided - a prerequisite which can be obtained only hypothetically. During the subsequent tempering process, BHF crystals start to grow on top of originally formed crystallization seeds. Since BHF belongs to the hexagonal system, with its c-axis easily magnetized, honeycomb-shaped crystals grow. The goal is to obtain crystals of small, uniform size corresponding to a single magnetic domain to avoid losses in the net magnetic properties of the powder. Optimum magnetic properties were reported [2] for BHF powders with crystal sizes of less than 1 pm in diameter, a homogeneous size distribution, and a narrow aspect ratio (diameter versus thickness) of approximately 3: 1. Compared to theoretical considerations with the aim to increase both remanence and coercitivity, the maximum energy density ((B€€)& obtainable from BHF is, however, about twice as high as technically obtained [3]. Scientific investigations for further improvement of BHF powders are urgently needed. Furthermore, there are several scientifically interesting questions still unanswered: Can crystallization processes like that of BHF be influenced, or even controlled, applying strong magnetic fields? Can we support and control the direction of crystallization by magnetically induced transport processes? Can we control the formation of crystallization seeds using magnetic fields? Can we use uncompensated spin moments of heavy 3d-metals (e.g., Fe, Co, Ni, Mn, Cr, V) above the Curie temperature in connection with strong magnetic fields to initiate nucleation and subsequent crystallization? Can nucleation in general be controlled by magnetic fields? We, therefore, chose to investigate crystallization behaviour of BHF under the influence of a strong magnetic field (up to 5 T) inside a superconducting,

31

cryogen-free magnet (CFM) system, equipped with a precisely controllable high-temperature furnace inside its warm bore. Experiments can be performed starting from 0 to 5 T with the magnetic field facing upward or downward, while the furnace allows maximum temperatures of up to 1600°C over extended periods of time. The temperature profile over the area of maximum temperature ("hot spot") is very constant, with a maximum variation in temperature of set point +/- 5°C. This results in a very homogeneous temperature regime over an experimentally accessible range of 5 x 8 cm (diameter versus height). The included ramping programs and gas-mixing equipment guarantees precise control of all process parameters such as temperature, time and oxygen fugacity (Fe2'/Fe3'-ratio).

2. Experiments Without Magnetic Field 2.1. Glass Crystallization Technique The GCT traces back to investigations by Kubo et al. (1980) [4]and starts, for our purposes, from melts prepared in the BaO-FezO3-B~O3-system as investigated by [2], [5] and [6].Since melts prepared in this system contain Fe3+, these melts are black, and tend to crystallize instantaneously above Fe203concentrations of more than 20 mole-%. Since crystallization should take place under controlled conditions of temperature and time to enable us to control chemical composition, crystallized phase, and simultaneously crystal size, nucleation and subsequent crystal growth must be avoided during homogenisation of the primary melt and subsequent quench to an amorphous glass. To avoid crystallization during quench, quench rates larger than lo5 Ks-' are necessary. These high quench rates can be obtained using a double-roller rapid quenching apparatus as described below. The GCT consists of basically two fusion processes: In the first step, starting components are vigorously mixed together and homogenized on a roller bench. This homogenized oxide mixture is then fused to obtain a primary melt, which is quenched onto a steel plate internally cooled by refrigerated water ( 10°C). In a subsequent fusion process, obtained glass chips from the first fusion step are added and fused inside a vertical Pt container equipped with a tiny hole in the thin lower tip. Through this tiny hole, liquid drops of melt trickle directly onto a double-roller rapid quenching apparatus made of 2 stainless-steel cylinders of about 100 mm diameter, which are internally water cooled and counter-rotating at speeds of 300 rpm or more. The gap between these two cylinders is less than 0.1 mm and the melt drop is squeezed through this gap.

32

The heat is transferred onto the steel surfaces and the melt drop is quenched instantaneously to a black, glassy flake. The double-roller rapid quenching technique supplies quench rates of up to lo5 Ks-’. Consequently, all glass flakes showed no indication of any crystallization. This finding was confirmed by XRD measurements on powdered glass flakes. Starting from a pristine, amorphous glassy state, these glass flakes withstood a tempering process under temperature conditions optimum for the growth of BHF crystals. Experiments were performed between 600 and 900°C for a period of 2 hours and longer. At 800°C and for run durations of 2 hours, homogeneous nanometer-size BHF crystals were obtained. Crystals are separated from the B203-matrix by chemical treatment with diluted acetic acid and recovered by centrifugation. With this technique, BHF powders of far less than 500 nm diameter, a narrow crystal size distribution, a crystal aspect ratio of 3:1, and a maximum coercivity of up to 400 kA/m ([2,5,6]) were obtained. Examples of BHF crystals are shown in Figure 1. Powders produced by the method described above are used as reference materials (“zero point”) for all subsequent investigations performed inside magnetic fields. By direct comparison of results obtained under conditions without any magnetic field present with results obtained by crystallization inside the CFM, the potential influence of a magnetic field on the crystallization of BHF powders can be directly studied and demonstrated.

Figure 1: REM image of BHF crystals obtained from a starting composition of 0.40 Ba0-0.27 Fe203-0.33 BzO,, tempered at 800°C for 2 hours.

33

2.2. Analytical Investigations Presently, six compositions of the system Ba0-FezO3-BzO3 have been investigated, and BHF powders were produced using the GCT. Since at least Bz03represents a volatile species in our system, melt and product compositions were checked throughout the entire GCT process by electron microprobe analysis (EMP). This allows precise knowledge and monitoring of potential changes in chemistry during the entire process. Each step of the sample preparation was accompanied by extensive studies of chemical and physical properties. Composition of the glass chips obtained after initial fusion of oxide mixture was checked by electron microprobe analysis for their major element composition. DTA investigations were performed to determine optimum conditions for the formation of BHF in successive tempering experiments. Flakes obtained from the second fusion step and rapid quenching were checked for their amorphous state by XRD and REM measurements prior to any further use in tempering experiments. After tempering, powdered flakes were analysed for their crystalline phases by XRD. A wide variety of analytical investigations such as density determinations, REM, optical microscopy, and magnetic property measurements were routinely performed on all powders. Based on these data and the precise record of the experimental conditions existing during their formation, parameters controlling the BHF crystallization can be evaluated, and optimum conditions for the formation of high-end grade BHF powder can be determined.

2.3. Results

EMP confirmed the major composition of the mixtures. XRD investigations showed patterns identical with pure and perfectly shaped BHF crystals. BHF powders obtained and analysed by REM revealed in consequence as well hexagonal platelets of single-crystalline and single-domain BHF of far less than 500 nm diameter, with a narrow aspect ratio (diameter versus thickness) of about 3:l. This is in perfect agreement with results obtained by [1,2,5,6,7,8,9]. DTA investigations showed that BHF crystals can be obtained when annealing (tempering) flakes above 650°C. However, perfect crystal morphology can be obtained between 780 to 800°C. Higher temperatures lead to degradation of the hexagonal shape of the BHF crystals due to solid-body reactions and beginning fusion. Increasing crystal perfection with increasing annealing temperature up to a maximum temperature is visible in the determined magnetic properties. Results are shown in Figure 2 for the determination of the coercivity Hc. Measurements

34

were performed at the IPHT (Jena) applying a vibrating sample magnetometer (static measurement), measuring the B-H behavior of powdered samples, assuming spherical particle size and Stoner-Wohlfahrt behavior without performing a demagnetization correction. The coercivity & increases from 600°C up to 850 or 900°C depending on the chemical starting composition of

400 n

EI

-!!

xu

z

3s u a2

3

-.-

+35 BaO, 35 Fe a03,30 B )03 40 BaO, 25 Fe *O,,35 BIO,

I

I

I

I

-

350 - - 40 BaO, 33 Fe ,O,, 27 B aO, 300 - - 0 - - 44 BaO, 15 Fe,O,, 41 BaO, 250 200 150

-

100

-

50 01

I

I

I

I

1

I

I

I

550 600 650 700 750 800 850 900 950

Temperature [“C] Figure 2. Coercivity (in kA/m) of BHF powders versus annealing temperature (in “C)after 2 hours of annealing. The grey triangle represents a repetition of the same composition with an extended tempering duration of 48 hours. Chemical composition of the starting mixtures (in mole-%) as indicated.

the oxide mixture from which the BHF powder was obtained. The “saddle” shaped dependence at lower temperatures (< 750°C) is most likely due to the change of the major phase of Ba-ferrite present (monoferrite more stable at lower temperatures, hexaferrite formed above 650°C). Similar behaviour was observed in [7] and explained as change in size dependence of the particles from superparamagnetic to stable single domain ones [4,10]. Chemical composition of the initial oxide mixture has a decisive influence on the coercivity &. Compositions with too low Fez03 content of 15 mole-% (open diamonds) do not supply the necessary amount of Fe to form the active magnetic phases. In consequence, coercivity & stays low and nearly constant. A starting oxide mixture with 25 mole-% Fez03 results in an already excellent coecivity &, while higher amounts (33 or 35 mole-%) lead to generally higher coercivities below 800°C. A maximum coercivity of about 350 kA/m is obtained for the

35 25 mole-% composition for annealing temperatures of 850°C while compositions with higher initial Fe20s-content have not yet reached their maximum. Increasing the annealing time from 2 hours to 48 hours for the same composition and annealing temperature increases the coercivity from 186 W m to 260 kA/m (compare Figure 2: grey triangle) most likely due to an increased crystal perfection and orientation of the magnetic domains. The determination of the remanence MR showed nearly identical behavior in respect to the annealing temperature. The remanence MR increases between 600°C and 900°C from about 30 mT to a maximum value of 230 mT obtained at around 800°C with no significant dependence on chemical starting composition. While performing aspect ratio determinations, BHF powders were deposited on glass slides held above a BHF solution kept inside a beaker in an ultrasonic bath. Consequently, particles were deposited such that aspect ratio measurements could be more easily performed. A REM picture of such a deposited BHF sample is given in Figure 3. As a consequence of this procedure, BHF crystals start to rearrange themselves in a chain-like configuration along their crystallographic c-axis. This facilitates the determination of aspect ratios with a higher precision than possible from images as shown in Figure 1. On the other hand, this image is a good example of what crystallization experiments inside strong magnetic fields are hoped to yield: the control of seed formation and crystal growth. After hitiation, crystals growth should result in the formation of single domain crystals of nearly identical size and shape. Crystal perfection should be high, and magnetic domains should be oriented parallel to each other by the magnetic field, resulting in improved magnetic material properties both on the nanometer as well as on the macroscopic scale. Materials and results obtained under magnetic field-free conditions will be used as magnetic field-free (MFF) reference material for all future materials synthesized inside magnetic fields. Comparison of these materials will allow the precise evaluation of the influence of magnetic fields on the crystallization process.

36

Figure 3: REM image of BHF powder deposited applying ultrasonic techniques. BHF powder was obtained from a starting oxide mixture with 0.40 Ba0-0.27 Fez03-0.33€3203, tempered at 800°C for 2 hours.

3. Experiments In High Magnetic Fields 3.1. Ctyogen-Free Magnet (CFM) System

Observed macroscopic magnetic properties of BHF are not exclusively controlled by the magnetic properties of the crystals and their single domains but as well by the relative orientation towards each other, and some other physical properties (e.g., density or packing). A high magnetic field applied during the entire crystallization process (initiation and formation of crystallization seeds, crystallization with altered crystal growth and transport properties) should result in a net increase of the magnetic properties (e.g., coercivity, remanence, energy density). Therefore, experiments in a cryogen-free magnet (CFM) of up to 5 T magnetic flux density equipped with a high-temperature oven to control precisely process parameters like temperature and tempering conditions will be performed to investigate whether and how strong magnetic field can help to increase the magnetic potential of BHF nano-crystalline powders especially in respect to an optimization of magnetic properties and their industrial application. As mentioned earlier, an increase of (BH),, of a factor of 2 should theoretically be possible [3]. For our experiments we use a CFM system built by CRYOGENICS (London, UK) as shown in Figure 4a and b. It can supply a maximum field strength of 5 T in its warm bore of 300 mm diameter. The CFM is mounted on a tiltable stand, and can be rotated into a vertical orientation of the warm bore axis

37 suitable for the installation of a high temperature furnace (HTF). Determination of the supplied magnetic flux density between 1 and 5 T leads to a 3-D calibration of the magnetic field inside the warm bore, and indicated that it is constant within 0.05 T over the dimensions of the Pt crucibles (4 cm diameter and 6 cm height) used for future crystallization experiments. These dimensions allow for about 200 g of melt in a single experiment. This mass is more than sufficient to guarantee an as wide range of analytical investigations of the obtained crystallized glass samples as possible. 3.2. The High-Temperature-Furnace (HTF) Based on this “3-D map” of the magnetic field distribution inside the warm bore of the CFM, an HTF was designed and ordered by XENON Advanced Heating (Freiberg, Germany). It is a specially designed, vertical muffle tube furnace that runs at peak temperatures of 1600°C over extended periods of time, while cooled at its exterior to guarantee the 40°C maximum temperature of the inner warm bore surface of the CFM. The built-in heating system guarantees a flat maximum temperature profile inside the furnace (“hot zone”), and a nearly constant temperature condition over the experimental charge kept in the Pt crucible of 6 cm height and 4 cm width.

Figure 4a. CFM (cryogen-free magnet) system on tilt stand power supply, temperature readout system, and computer for automatic control in background.

38

Figure 4b. CFM rotated to vertical position and HTF inserted into warm bore.

Both hot zone and area of maximum, constant magnetic field strength are designed to overlap in the center of the experimental region. The HTF uses a EUROTHERM controller that allows high accuracy and reproducibility of experimental conditions for experiments with ramp-up and ramp-down procedures as well as for those at constant tempering temperatures. The installed gas mixing setups will allow future experiments at constant oxygen fugacities to additionally investigate the influence of the Fe2+/Fe3+ratio on the crystallization behavior and involved phases of the melts investigated. 3.3. Crystallization Experiments at High Temperatures Inside High Magnetic Fields Serious problems with the reliability of the heating system forced us to cancel the initiation of crystallization experiments under controlled temperature conditions and to remodel the entire built-in heating system. The remodeling is still in progress. We, therefore, decided to start some qualitative experiments. Since controlled temperature conditions are not available, we fused a BHF starting mixture inside a standard GERO high temperature furnace at 1400°C. We then transferred the melt contained in a Pt crucible with lid after 2 hours of fusion into our CFM standing at 1 T. Outside of the magnetic field, the melt convected vigorously. The damping effect of the magnetic field on the convective flow

39

while transferring the melt into the magnetic field was impressive. During this cooling process, temperature conditions inside the sample were monitored over 1 hour, decreasing from 1400°C to 200°C. Parallel to this experiment a batch of BHF powder of identical composition was allowed to cool naturally outside the magnetic field under identical temperature conditions. After 1 hour, both samples, still at roughly 200"C, were quenched to room temperature by running tap water over the Pt crucible walls. Both samples were then inserted into the CFM and tested for their magnetic behavior. BHF that was crystallized outside the magnetic field showed only small magnetic properties and could easily be moved in and out of the CFM standing at 1 T. The sample crystallized inside this field, however, showed a strong paramagnetic behavior when inserted into the warm bore of the CFM. Fast movements were damped significantly in direct comparison with the sample crystallized outside the magnetic field. Attractive forces of the magnetic field were significantly larger in comparison to the sample crystallized without magnetic field. Unfortunately, further quantitative investigations cannot be presumed until the HTF is delivered and experiments are repeated under precisely known temperature conditions. However, results of these preliminary experiments have already demonstrated the potential of crystallization processes performed inside magnetic fields and the effect on the magnetic properties of materials created in this way. Acknowledgements This study was financially supported by the Thiiringer Ministerium fiir Wissenschaft, Forschung und Kunst (TMWFK). Many thanks go to the Bayerisches Geoinstitut (University of Bayreuth) for performing the electron microprobe analyses, and to Dr. Muller from the Institut fiir Physikalische Hochtechnologie (IPHT, Jena, Germany) for helping with the vibrating sample magnetometer measurements and their interpretations. References 1. Watanabe, K., Hoshi, K., Crystallisation kinetics of fine barium hexaferrite, BaFe12019,particles in a glass matrix. Phys. Chem. Glasses 40(2), (1999), pp. 75-78. 2. Knauf, O., Nutzung groOer Abkiihlungsgeschwindigkeiten zum Amorphisieren spontan kristallisierender oxidischer Schmelzen, dargelegt am System BaO-Fe203-B203. Dissertation, Technische Universitat of Ilmenau, 1988. 3. Heck, C., Magnetische Werkstoffe und ihre technische Anwendung. A. Hitthig Verlag, Heidelberg (1975), 280 pages.

40

4.

5. 6. 7.

8. 9.

10.

Kubo, O., Ido, T., Inomato, T., Yohoyoma, H., German patent # DE 304 1960, Tokyo Shibaura Denki, 1980/81. Hiilsenberg, D., Knauf, 0.. Hamann, B., Glass Crystallization Technique for Ultrafine Ceramic Powders. AGM Meeting of the German Ceramic Society, Weimar (1993), Germany. Hulsenberg, D., Knauf, O., Hamann, B., Glass crystallization Technique for Ultrafine Ceramic Powders. DKG 71(11-12), (1994), pp. 707-711. Gornert, P., Sinn, E., Schiippel, W., Pfeiffer, H., Rosler, M., Schubert, Th., Jurisch, M., Sellger, R., Structural and Magnetic Properties of BaFelz. &oXTixOl9 Powders Prepared by the Glass Crystallization Method, ZEEE Transactions on Magnetics, 26 (January 1990), pp. 12-14. Rosler, M., Gornert, P., Sinn, E., Structural and Magnetic Properties of BaFerrite Fine Particles Grown by Glass Crystallization. Z.Phys. D - Atoms, Molecules and Clusters, 19, (1991), pp. 279-281. Gornert, P., Schiippel, W., Sinn, E., Schumacher, F., Hempel, K.A., Turilli, G., Paoluzi, A., Rosler, M., Comparative measurements of the effective anisotropy field Ha for Barium Ferrites. J. Magnet. Magnet. Mat. 114, (1992), pp. 193-201. Shirk, B.T., Ruessem, W.R., Magnetic Properties of Barium Ferrite Formed by Crystallization of a Glass, J. Am. Cerum. SOC. 53(4), (1970), pp. 192196.

VARIATION OF PHASE TRANSFORMATION TEMPERATURE IN FE-C ALLOYS IN A HIGH MAGNETIC FIELD X.J. HAO, H. OHTSUKA, H. WADA Tsukuba Magnet Laboratory, National Institute for Materials Science, Tsukuba, Ibaraki, 305-0003 Japan A magnetic field can affect the transformation temperature and microstructure if a transformed phase has a different susceptibility from the parent phase. Fe-C alloy is an ideal system to show the magnetic field effect since, in this system, austenite (FCC stmcture) is a paramagnetic phase and ferrite (bcc structure) is a ferromagnetic phase below 770 "C. In this paper, phase transformation temperature in Fe-C alloys in a magnetic field was measured from a cooling curve. It was found that the transformation temperature for pure Fe from austenite to ferrite has a linear relationship with magnetic field strength, increasing about 0.8 "C per tesla. For eutectoid transformation in Fe-0.8C alloy, similar relationship exists; the transformation temperature increases about 1.5 "C per tesla. The measured interaction energy between magnetic field and ferrite is larger than that calculated from molecular field theory. An elongated and aligned microstructure by ferrite transformation in a high magnetic field was found in a Fe-0.4C alloy, but was not found in pure Fe and Fe-0.8C alloy.

1. Introduction The expectation is that an external high magnetic field affects solidsolid phase transformation behaviors and transformed structures, and possibly improves the mechanical and magnetic properties of materials. In fact, the structural alignment in solidsolid transformations in high magnetic fields has been reported for ferrite transformation [ 1-51 and reverse transformation [6,7] by continuous cooling or isothermal holding in austenite and ferrite dual phase region. Thermodynamic calculation OR an equilibrium phase diagram of Fe-C binary system proposed that a high magnetic field increases the austenite(y)/ferrite(a) equilibrium temperature, carbon solubility in c1 phase and eutectoid carbon content [8,9]. Kakeshita et a1 [lo] have investigated the effects of magnetic field on martensite transformation temperature, but there are few experimental data to confirm the effect of magnetic field on ferrite transformation temperature. In this paper, we report on our experiments on the effect of a magnetic field on phase transformation temperature and microstructure in Fe-C alloys and compare them with the results of theoretical calculations.

41

42

2. Experiments The alloys used in this study were Fe-0.8C alloy prepared by vacuum induction melting and high purity Fe (99.99%). After hot rolling, specimens were machined to 5mmx 5mm xlmm and then set in a vacuum furnace, which was installed in a helium-free-type superconducting magnet with a bore size of @ 100 mm. A magnetic field perpendicular to the 5mmx5mm specimen's surface was increased to 10 T in 27 minutes before austenitization and kept constant during austenitization and subsequent cooling, then decreased to 0 T. Specimens were fixed at the center of the magnetic field and the magnetic force on the specimen is negligible. Specimens were austenitized at 1000 "C for 15 minutes and cooled to 600 "Cat a cooling rate of 10 "Umin. The specimen temperature was measured by a thermocouple contacted with the specimen and recorded by a digital recorder. Since the temperature controller has some delay to control the sample temperature to programmed set value if the sample temperature changes suddenly, it is possible to find a peak in the cooling curve. This peak determined the phase transformation temperature. Microstructure observation was performed on the plane parallel to the direction of magnetic field by optical microscope after polishing and 3% Nital etching. 10-

(b) 8\

Time

T(0)=906.Z°C

-

0

2

4 6 8 1 Magnetic field strength, HIT

0

Figure 1. Cooling curves (a) of pure Fe in magnetic field after austenitizationat 1000 "C for 15 min, and the femte transformation temperature increasing with magnetic field strength (b).

43

Time

Figure 2. Cooling curves (a) of Fe-0.8C alloy in magnetic field after austenitization at 1000 "C for 15min, and the pearlite transformationtemperature increasing with magnetic field strength (b).

3. Results and Discussions Figure l(a) shows the cooling curve segments of pure Fe during transformation in magnetic fields. Without a magnetic field, the transformation temperature of pure Fe from austenite to ferrite is about 906.2 T , which is about 6 "C lower than the equilibrium temperature (912 "C). This difference is known as supercooling and it provides the chemical driving force for transformation. The transformation temperature increases gradually with increasing magnetic field strength. With a maximum magnetic field of 10 T, the transformation temperature is about 914.9 'C, which is 8.7 "C higher than with no magnetic field. The increased temperature AT, (=T(H)-T(O), the transformation temperature difference between field and no field) was plotted against magnetic field strength in Figure l(b). It shows that the transformation temperature increases linearly with magnetic field strength. The transformation temperature increases about 0.8 "C for an increasing magnetic field strength of 1 T. Figure 2(a) shows the cooling curve segments of Fe-0.8C alloy during transformation in magnetic fields. The measured eutectoid transformation temperature is 704.9 "C without applying the magnetic field. The equilibrium eutectoid transformation temperature is 727 "C. A larger supercooling (about 22 "C) results compared to ferrite transformation in pure Fe. Eutectoid transformation needs element redistribution in ferrite and cementite, and the transformation temperature is relatively lower. With magnetic field, the transformation temperature increases. With a magnetic field of 10 T, the temperature increase to about 720.1 "C, which is about 15 'C higher than that without a magnetic field. Figure 2(b) shows the relationship between AT and magnetic field strength. They have a linear relationship, same as the ferrite

44 transformation. The AT is about 1.5 "C per tesla of increasing magnetic field strength.

I

1.0-

0.0 0

"

", ,

, I , ,

200

,,,...'I

I I , .

400

,, ,,, 0 , .

. ..

--. ,,.'.I,,

600

800

, , , lTTF

1000

Temperature. "C

Figure 3. Variation of magnetization of pure Fe with temperature in an external magnetic field calculated from molecular field theory. & is the spontaneous magnetization of Fe at 0 K. The square symbols are calculated from measuring the transformationtemperature change (see text).

It is surprising that AT for pure Fe is more than half of that for eutectoid transformation if we remember that ferrite is in a paramagnetic state at the transformation temperature, whereas ferrite is in a ferromagnetic state for eutectoid transformation. Figure 3 shows the magnetization of pure Fe without or with an external field of 10 T calculated by molecular field theory. Magnetic moment exists even above Curie temperature and the field-induced magnetization at ferrite transformation temperature in pure Fe, 915 "C, is about 0.07w(&=1.74 X lo6A h , saturation magnetization of pure Fe at 0 K). Below T,, the magnetization is composed of spontaneous magnetization (M,f) in zero field and field-induced magnetization (Mfi). At eutectoid transformation temperature in Fe-0.8C alloy, 720"C, total magnetization was calculated to be 0.42&, which includes a field induced magnetization of 0.07 &. The change of Gibbs free energy of one mole pure Fe in an external field H is,

where V, is the volume of one mole Fe and is the permeability of vacuum. The free energy of ferrite in a magnetic field of 10 T decreases about 4.3 J/mol at 915 "C and 47.5 J/mol at 720 "C calculated from this equation using above

45

magnetization data. The free energy change in magnetic field at 915 "C is more than one order smaller than that at 720 "C. If we consider that the magnetic energy has a similar effect on phase transformation as that of chemical driving force, it is possible to measure the magnetic energy from the transformation temperature. Figure 4 shows the chemical driving force changes with temperature in pure Fe and Fe-0.8C alloy calculated by Thermo-Calc. The chemical driving force of ferrite transformation for pure Fe is 4.8 J/mol at 906.2 "C and -2.6 J/mol at 914.9 "C. So the magnetic field provides a free energy of 7.4 J/mol to compensate the chemical driving force (4.8-(-2.6)=7.4 J/mol). Similarly, the magnetic energy for eutectoid transformation was calculated to be about 71 J/mol by measuring AT. It should be mentioned that only the ferrite phase is considered during calculation, though pearlite consists of ferrite and cementite, in which cementite is paramagnetic and has a volume fraction of about 13%. aaa

1

"I

;

.$ 30

n 880

890

900

910

Temperature. "C

920

930

0 680

690

700

710

720

730

Temperature. "C

Figure 4. Chemical driving force changes with temperature for ferrite transformation in pure Fe (a) and pearlite transformation in Fe-0.8C alloy (b). The magnetic energy of the transformed phases in a magnetic field of 10 T is shown in the figure.

The calculated magnetic energy based on molecular field theory was found to be smaller than the value calculated from transformation temperature. There may be two reasons for this. First, it is possible that the magnetization in the high magnetic field calculated by molecular field theory is lower than the actual magnetization. The square symbol shown in Figure 3 is the magnetization calculated from transformation temperature. In fact, some research has already proved that the magnetization near Curie temperature, calculated by molecular theory, is smaller than experimental data. Another possibility is that a high magnetic field not only changes the thermodynamic properties of phase transformation but also affects the kinetic properties, such as nucleation sites,

46

interfacial migration and atom diffusion. Some experiments already show this kind of effect, however, there are presently no conclusive results. Though the transformation temperature in pure Fe is about 145 "C higher than the Curie temperature, a high magnetic field of 10 T can shift it as much as 8°C. Thermodynamic calculation results shown in Figure 4 indicate that the driving force for ferrite transformation of pure Fe increases much more slowly with decreasing transformation temperature than that of eutectoid transformation, so a weak magnetization and then a small magnetic energy can induce large transformation temperature shifting. In a system, it was suggested that if the slopes of the Gibbs free energy curve of two phases is close, it is possible to control the phase transformation by an external magnetic field even if both phases are not in a ferromagnetic state. An elongated and aligned structure was found in the Fe-0.4C alloy by ferrite transformation during slow cooling or isothermal transformation in a high magnetic field, and it was suggested that this was due to the demagnetization field developed in ferrite [2,3]. However, this type of structure was not found through microstructural observation in pure Fe and Fe-0.8C alloy. For pure Fe, this can be attributed to the fact that the magnetization at transformation temperature is too small, and the demagnetization field is too weak. For eutectoid transformation in Fe-0.8C alloy, the lamellar morphology and orientation relationship between pearlite and parent austenite phase possibly account for the lack of the abovementioned structure. 4.

Conclusions

The effects of high magnetic field on the phase transformation temperature and microstructure in Fe-C alloys were investigated. It was found that the transformation temperature for pure Fe from austenite to ferrite has linear relationship with magnetic field strength, increasing about 0.8 "C per tesla. For eutectoid transformation in Fe-0.8C alloy, a similar relationship exists; the transformation temperature increases about 1.5 "C per tesla. The measured interaction energy between magnetic field and ferrite is larger than that calculated from molecular field theory. An elongated and aligned microstructure by ferrite transformation in a high magnetic field was found in Fe-0.4C alloy, but was not found in pure Fe and Fe-0.8C alloy.

References 1. Ohtsuka, H., Xu, Ya, and Wada, H., Materials Transactions, JIM, 41, (2000) pp. 907-910.

47

2. Hao, X.J., Ohtsuka, H., DE Rango, P., and Wada, H., Muter. Trans., 44, (2003) pp. 21 1-213. 3. Hao, X.J., Ohtsuka, H., and Wada, H., Muter. Trans., 44,(2003) pp. 25322536. 4. DE Rango, P., Hao, X.J., Ohtsuka, H., and Wada, H., Trans. Muter. Res. SOC. Japan, 28, (2003) pp. 225-226. 5. Shimotomai, M., Maruta, K., Mine, K., and Matsui, M., Actu Muter., 51, (2003) pp. 2921-2932. 6. Hao, X.J., Ohtsuka, H., and Wada, H., Trans. Muter. Res. SOC. Japan, 28, (2003) pp. 223-224. 7. Ohtsuka, H., Hao, X.J., and Wada, H., Muter. Trans., 44, (2003) pp. 25292531. 8. Choi, J-K., Ohtsuka, H., Xu, Y., and Choo, W-Y., Scriptu. Muter., 43, (2000) pp. 221-226. 9. Enomoto, M., Guo, H., Tazuke, Y., Abe, Y.R., and Shimotomai, M., Metull. Muter. Trans., 32A, (2001) pp. 445-453. lO.Kakeshita, T., Kuroiwa, K., Shimizu, K., Ikeda, T., Yamagishi, A. and Date, M., Muter. Trans. JIM, 34, (1993) pp. 423-428.

MARTENSITIC TRANSFORMATIONIN SOME FERROUS ALLOYS UNDER MAGNETIC FIELD T.KAKESHITA Department of Materials Science and Engineering, Graduate School of Engineering, Osaka University, 2-1, Yamada-oka,Suita, Osaka 565-0871, Japan Martensitic transformations of ferromagnetic materials are extensively influenced by a magnetic field. The transformationtemperature changes, because the free energy between parent and martensite phases changes under magnetic field. In a special case, the martensite phase is induced only while a magnetic field is applied, and is transformed back to the parent phase when the magnetic field is removed. The martensite plate formed under magnetic field tends to align to the field direction. Furthermore, in some ferromagnetic shape memory alloys, a giant magnetic field-induced strain of more than 1 % appears due to rearrangement of martensite variants.

1. Introduction Martensitic transformations are generally influenced by external fields [ 1,2], of which the magnetic field is an example [3,4]. In fact, many researchers [3-91 have examined effects of magnetic fields on martensitic transformations. We also examined them systematically [4-81 and found many interesting phenomena. In this paper, we discuss our studies on the effects of magnetic fields on martensitic transformations in some ferrous alloys; (i) effect of magnetic field on the martensitic transformation start temperature, M,,and the validity of the equation proposed by our group to evaluate the relation between M, and magnetic field; (ii) Magnetoelastic martensitic transformation (maretensites are induced only while a magnetic field is applied and are transformed back to the parent phase when the magnetic field is removed) in an ausaged Fe-Ni-Co-Ti shape memory alloy; (iii) morphology of martensite formed under magnetic field; (iv) a giant magnetic field-induced strain due to rearrangement of variants in the martensite state of Fe-Pd and Fe,Pt ferromagnetic shape memory alloys exhibiting a thermoelastic martensitic transformation.

2. Experiment Specimens used were Fe-3 1.7at.%Ni, Fe-24at.%Pt and Fe-3 1.9Ni-9.8Co-4.1Ti(at.%) polycrystals, and Fe-3 1.6Ni, Fe-3 1.2at.%Pd, and Fe3Pt single crystals. They were prepared using a high frequency induction

48

49 furnace or by arc melting. An ordering heat treatment was made in the Fe-24at.%Pt and its degree of order was 0.8. Single crystals of Fe-31.2at.%Pd and Fe3Pt were grown by a floating zone method. The Fe-31.2at.%Pd was solution treated at 1373 K followed by quenching into iced water. Fe3Pt was solution treated at 1373 K followed by ordering treatment at 873 K. High field magnetization measurements were performed at Research Center for Materials Science at Extreme Conditions, Osaka University, using a pulsed magnetic field with a maximum strength of about 31 MA/m. Magnetic field-induced strain was measured by a sensitive three terminal capacitance method, where specimens were mounted in a parallel-plate capacitance cell.

3. Results and Discussion 3.1. Effect of Magnetic Field on Martensitic Transformation Temperature The martensite phase is usually induced thermally by cooling a specimen. In some alloys, the transformation temperature Ms changes to Ms’ under a magnetic field Hc. The solid circles of Figure 2 shows AM, (= M,’- M,) as a function of magnetic field for the Fe-31.7at.%Ni (a) and Fe-24at.%Pt (b) examined by pulsed magnetic field experiments. It is known from the figures that the shift of M, increases with increasing magnetic field. We proposed the following equation [4] to estimate the relation between the critical magnetic field and the transformation start temperature:

AG(Ms)-AG(Ms’) =-AM(Ms’). Hc-(ll2)*x& .Hc2 + €0 .(awl

Hc. B

(I)

where AG(Ms) and AG(Ms’) represent the difference in Gibbs chemical free energy between the parent and martensite phases at M, and M,’ temperatures, respectively, hM(Ms’) the difference in spontaneous magnetization between the parent and martensitic states at Ms’,yhf the high magnetic field susceptibility in the parent phase, 6 the volume change associated with martensitic transformation, w the parent forced volume magnetostriction and B the parent bulk modulus. The first, second and third terms on the right-hand side of Eq. (1) represent the energies due to the magnetostatic, high field susceptibility and forced volume magnetostriction effects, respectively. Based on the equation, H, vs. M,’relations have been thermodynamically calculated for the present alloys by using the physical quantities involved in the equation, obtained by referring to the previous studies and by measurment in the present study [4,5]. The calculated results are shown in Figure 1, where the dotted lines indicated with M.S.E., H.F.E., F.M.E. and (M.S.E.+H.F.E.+F.M.E.) represent the H, vs. M,’

50

relations calculated for the magnetostatic, high field susceptibility, forced volume magnetostriction and their total effects, respectively. The calculated relations (M.S.E.+H.F.E.+F.M.E.) agree well with the experimental ones for both of the alloys.

60

Fe-31.7al%Ni

Magnetic Field. ti IMAm-‘

Figure 1. Calculated and measured shift of Ms as a function of magnetic field for Invar Fe-3 1.7at% Ni(a) and Invar Fe-24.0at.%R.(b).

3.2. Magnetoelastic Martensitic Transformation Alloys exhibiting a thermoelastic martensitic transformation exhibit pseudoelastic behavior due to the stress-induced martensitic and its reverse transformation upon loading and unloading. Analogous to this behavior we expected that an ausaged Fe-31.9Ni-9.8Co-4.1Ti (at.%) shape memory alloy [ 151 should show field-induced martensitic and its reverse transformation upon applying and removing magnetic field. The M,,A, and Af of the alloy used are 127, 60 and 159 K, respectively. The latent heat of the transformation is about 334.4 J/mol. The difference in spontaneous magnetization between the parent and martensite phases is about 0.3 pB/atom at M,. We applied a pulsed high magnetic field to the specimen at a temperature above Af, 163 K( d T (= T-M,)= 36 K,T > Af). As seen in the M(r)-H(r) curve of Figure2(a), there is no hysteresis of the magnetization when the maximum strength is 22.22 MNm, meaning that the martensite phase is not induced. When a higher field is applied, the rate of increase of magnetization against magnetic field changes at H, = 23.08 MNm, as indicated with an arrow. When the magnetic field is removed, the increased magnetization returns to the initial

51

value at about Hf=5.56 MA/m indicated with another arrow. This means that martensitic transformation is induced at H, and its reverse transformation is completed at Hf. These observations show that the magnetoelastic martensitic transformation is certainly realized in the ausaged Fe-Ni-Co-Ti alloy, and such behavior is always realized at temperatures above Af.

5.56MAIm

lo

10

Tr163KfT > A f )

20

Magnetic Field (MA/m)

Figure 2. M-Hcurves of an ausaged Fe-Ni-Co-Ti alloy at 163 K, which is just above the reverse transformation finish temperature Af.

3.3. Morphology of Magnetic Fieldinduced Martensite When martensite is induced thermally from the parent phase, many crystallographic domains (variants) form nearly equivalently. Alternatively, when it is induced by magnetic field, specific variants tend to grow preferentially. Figure 3 shows the microstructure of the magnetic field-induced martensite of an Fe-31.6at.%Ni alloy single crystal. The field is applied along [110] direction, which corresponds to the horizontal direction in Figure 3. It is clearly seen in Figure 3 that several martensite plates grow nearly parallel to the field direction and run through the single crystal from one end to its other. This preferential growth in the field direction is also observed by applying the magnetic field along [loo] and [ l l l ] . The reason for the lengthwise growth under a magnetic field is not clear, but a shape magnetic anisotropy effect seems to play an important role.

52

Figure 3. An optical micrograph of magnetic field-induced m e n s i t e in Fe-31.6at.%Nialloy single crystal. The magnetic field direction is horizontal and is parallel to the [I101 direction.

3.4. Giant Magnetostriction in Ferromagnetic Shape Memory Alloys The large strain appearing in shape memory alloys is caused by the rearrangement of variants under external stress. In some ferromagnetic shape memory alloys, the rearrangement of variants can be induced by a magnetic field, resulting in a giant magnetic field-induced strain [ 12-14]. Typical examples exhibiting such behavior are Ni-Mn-Ga, Fe-Pd and Fe3Pt. Each of them transforms from a cubic structure to a tetragonal structure and its tetragonality ( c h ) is slightly smaller than unity. As a result, there are three lattice correspondence variants. The easy axis of magnetization is the a-axis for Fe-Pd alloy, and is the c-axis for Ni-Mn-Ga and Fe3Pt. These easy axes correspond to one of the p directions of the parent phase. In the following, we show some results of Fe-31.2at.%Pd and Fe3Pt [15-161. I

.

I

I

I

I

I

I

3.0 h

E 2.0

0.0 0.0

1.o 1.5 Magnetic Field, H/(MAlrn)

0.5

Figure 4 Magnetic field-induced strain of Fe-31.2Pd alloy at 77 K. Measurement was made by applying a magnetic field along [OOI] after cooling down without a magnetic field.

53 The martensitic transformation temperature of Fe-3 1.2at.%Pd is about 230 K. A single crystal of Fe-31.2Pd was cooled down to 77 K under zero magnetic field, then a magnetic field was applied along [OOI] direction. In this process a large field-induced strain (expansion) of about 3% appeared as shown in Figure 4. We observed the rearrangement of variants through an optical microscope under a magnetic field. In association with the rearrangement of variants, the magnetization curve shows a large hysteresis whose area is nearly the same as the energy dissipation evaluated by a stress-strain curve. The uniaxial magnetocrystalline anisotropy constant K,, is about -350 W/m3 at 77 K. The large value of lKul and small value of twinning shear will be the primary reason for the twinning plane movement under magnetic field. The martensitic transformation temperature of Fe,Pt (degree of order 0.8) is about 85 K. A single crystal of Fe3Pt was cooled down to 4.2 K under zero magnetic field, then a magnetic field was applied along [OOl] direction. A large field-induced strain (contraction) of about 2.3% appeared in this process. A part of the strain (about 0.6%) recovers in the field removing process and it repeatedly appears in the subsequent field applying processes. This reversible strain is nearly three times as large as that of TERFENOL-D (Tbl.xDyxFez)[17], which is well known as a magnetostrictive material. 5.

Summary

We have shown that magnetic field influences extremely martensitic transformations: the martensitic transformation temperature is controlled by the magnetic field; in an Fe-Ni-Co-Ti alloy, martensites are induced only while a magnetic field is applied and are transformed back to the parent phase when the magnetic field is removed; the morphology of the martensite can be controlled by magnetic field; a giant magnetic field-induced strain can be induced in some ferromagnetic shape memory alloys. From these results, it is expected that magnetic field is effective for the development of new structural and smart materials as well as basic research for martensitic transformation.

Acknowledgements This work is supported by the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT), through MEXT Special Coordination Funds for Promoting Science and Technology (Nanospintronics Design and Realization, NDR), and the 21st COE Program (Center for Excellence for Advanced Structural and Functional Materials Design).

54

References 1. Patel, J.R. and Cohen, M., Actu Metull., 1 (1953) pp. 531-38. 2. Otsuka, K., Sakamoto, H. and Shimizu K., Actu Metull., 24 (1976) pp. 585-601. 3. Sadovsky, V.D., Smirnov, L.V., Fokina, Ye., Malinen, P.A. and Soroskin, I.P., Fiz. Met. Metulloved., 27 (1967) pp. 918-39. 4. Kakeshita, T., Shimizu, K., Funada, S. and Date, M., Actu Metull., 33 (1985) pp. 1381-89. 5. Kakeshita, T., Shimizu, K., Funada, S. and Date, M., Trans. JIM, 25 (1984) pp. 837-44. 6. Kakeshita, T., Shimizu, K., Maki, T., Tamura, I., Kijima, S . andDate, M., So: Metull., 19 (1985) pp. 973-76. 7. Kakeshita, T., Yoshimura, Y., Shimizu, K., Endo, S., Akahama, Y. and Fujita, F.E., Trans. JIM, 29 (1988) pp. 781-89. 8. Kakeshita, T., Shimizu, K., Nakamichi, S., Tanaka, R., Endo, S . and Ono, F., Trans. JIM, 32 (1992) pp. 1-6. 9. Chernenko, V.A., Babii, O.M., Kokorin, V.V., Lotkov, A.I. and Grishkov, V.N., Phys. Met. Metullogruphy, 8 (1996) pp. 549-552. lO.Kakeshita, T., Shimizu, K., Maki, T., Tamura, I., Kijima, S . and Date, M., Scr. Metull., 19 (1985) pp. 973-976. 11.Ullakk0, K., Huang, J.K., Kantner, C., OHandley, R.C. and Kokorin, V.V., Appl. P h y ~Lett., . 69 (1996) pp. 1966-1968. 12.Tickle, R. and James, R.D., J. Mug. Mug. Muter. 195 (1999) pp. 627-738. 13.Murray, S.J., Marioni, M., Allen, S.M., O'Handley, R.C. and Lograsso, T.A., Appl. Phys. Lett., 77 (2000) pp. 886-888. 14.James R.D. and Wuttig M., Phil. Mug. A77 (1998) pp. 1273-1299. 15.Kakeshita, T., Takeuchi, T., Fukuda, T., Tsujiguchi, M., Oshima, R. and Muto, S., Appl. P h p . Lett. 77 (2000) pp. 1502-1504. 16.Sakamot0, T., Fukuda, T., Kakeshita, T., Takeuchi, T. and Kishio, K., J. Appl. Phys. 93 (2003), pp. 8647-8649. 17.Clark, A.E., Teter, J.P., McMaster, O.D., J. Appl. Phys. 63 (1988) pp. 3910-3912.

EXPLORING ULTRA-HIGH MAGNETIC FIELD PROCESSING OF MATERIALS FOR DEVELOPING CUSTOMIZED MICROSTRUCTURES AND ENHANCED PERFORMANCE G.M. LUDTKA', R.A. JARAMILLO', R.A. KISNER', J.B. WILGEN', G. MACKIEWICZ-LUDTKA', D.M. NICHOLSON', T.R. WATKINS', P. KALU', AND R.D. ENGLAND3 'Oak Ridge National Laboratory, Oak Ridge, Thf 37830 2 National High Magnetic Field Laboratory, FAMU-FSU, Tallahassee, FL 32310 3Cummins Inc, Columbus, Indiana 42701 This paper will highlight results from research investigating magnetic field effects in phase equilibrium and transformation kinetics in several ferrous materials. In this work, a Fe-15Ni binary, 1045 steel, 52100 steel and a high strength bainitic steel were exposed to various thermal histories both with and without a magnetic field. Temperature and hardness measurements, metallography and X-ray analysis indicate a significant shift in phase transformation kinetics and resulting microstructure. Shifts in phase solubility and constituent volume fractions were measured for the Fe-15Ni alloy when treated in the presence of a magnetic field. The magnetic field reduced retained austenite during rapid quench and accelerated austenite decomposition during isothermal transformation in 52100 steel. Temperature data shows that the transformation temperature during continuous cooling is increased by approximately 3"CIT for 1045 steel. The application of a magnetic field during continuous cooling of a high strength bainitic steel has produced a microstructure not possible without the additional thermodynamic driving force associated with a 30 T magnetic field. These results validate thermodynamic arguments for the magnetic field effect and indicate significant potential for developing novel microstructures and, therefore, properties in steel and other ferromagnetic materials.

1. Introduction Magnetic field processing is proving to be an innovative and revolutionary research focus area that is creating the basis for an entirely new research opportunity for materials and materials process development. This approach has both scientific and industrial relevance with significant energy savings and environmental benefit ramifications and represents a major step towards achieving "materials by design" goals. Our experimental and modeling research efforts are clearly demonstrating that phase stability (conventional phase diagrams) can be dramatically altered through the application of an ultrahigh magnetic field. The ability to selectively control microstructural stability and alter transformation kinetics through appropriate selection of the magnetic field strength is being shown to provide a very robust mechanism to develop and

55

56

tailor enhanced microstructures with superior properties through a more efficient processing technology for a broad spectrum of material applications. The results of ongoing research conducted collaboratively utilizing the resources of the Oak Ridge National Laboratory and the National High Magnetic Field Laboratory will be presented in this paper. Specifically, the use of magnetic field processing to modify phase equilibria and kinetics in several ferromagnetic materials will be demonstrated and discussed. Results from proof-of-principle experiments will be presented for a Fe-15Ni alloy to show that phase stability and solubility can be altered through application of a high magnetic field. Similarly, the influence of high field magnetic processing on austenite decomposition in several steels will be presented. These alloys will include a hypoeutectoid 1045 plain carbon steel, a 52100 hypereutectoid gear steel, and a recently developed high strength bainite steel [ 11. A key component of the research is the ability to rapidly heat and cool the specimen inside the bore of the magnet. The experimental apparatus employs induction heating and a gas purge/quench system to provide temperature control during experiments. Details of the apparatus are provided elsewhere [2] and will not be reviewed here. In the following sections, only results from our research will be presented. Additional details and information related to the experimental system, procedures and sample analysis will appear in forthcoming papers.

2. Modifying Phase Equilibria in Fe-15Ni A Fe-15Ni (atomic percent) alloy was employed to demonstrate the influence of magnetic field processing on phase equilibria in the a+y region of the Fe-Ni phase diagram. The experimental procedure was to simply hold the material at 500°C for 4 hours both with and without a 29 T magnetic field. The samples were quenched to room temperature and X-ray analysis was performed to measure phase volume fractions and associated chemistries [3]. The results of X-ray analysis are summarized in Table I. These results show that the 29 T magnetic field significantly alters solubility and volume fractions for the BCC (a) and FCC ( y ) phases compared to the no field condition. Of significant interest is that the amount of FCC phase was reduced by 50% and Ni solubility of the FCC phase increased (from 0.28 to 0.4 atomic fraction Ni). Also, the Ni content for the BCC phase was increased by the presence of a magnetic field and the BCC volume fraction increased.

57 Table 1. Quantitative X-ray Diffraction Analyses Results Volume Fraction Ni in a (BCC phase) Ni in y (FCC phase)

I

No Field BCC: 70% FCC: 30% 9 at. % 28 at. %

29 T Field BCC: 85% FCC: 15% 11at. % 40 at. %

The implication of X-ray results are illustrated in Figure 1. The figure is the two-phase region of the Fe-Ni phase diagram showing Ni solubility over a range of temperatures. The horizontal, double-headed arrow identifies the experiment temperature and the diamond marker indicates bulk nickel content. The dashed lines approximate the shift in Ni solubility for each phase with a 29 T magnetic field. The increase in Ni solubility for each phase shifts phase fractions predicted by the lever rule such that an increase in BCC (a) fraction is observed. To supplement these results, modeling efforts where undertaken to predict the influence of magnetic fields on microstructure stability. This modeling was accomplished by combining the thermodynamic model of Chuang [4]with local density calculations [5]to predict Ni solubilities and phase fractions with a 29 T magnetic field. These results are published elsewhere [3].

c

WT. Ye Ni Figure 1. Fe-Ni phase diagram [6] illustrating the shift of phase boundaries under the imposition o f a 29 T magnetic field. The dashed lines approximate the position of boundaries (solvus lines) associated with a 29 T magnetic field.

58

3. Austenite Decomposition in SAE 52100 Steel This section deals with experiments conducted on a 52100 steel to evaluate the effect of a magnetic field on retained austenite during and after rapid quench and isothermal transformation at 740 "C. Samples were austenitized at either 850 "C or 900 "C for 20 minutes followed by a rapid quench to room temperature with magnetic fields of 10, 17 and 30 T applied a) during the quench or b) at room temperature after quenching. Retained austenite values were measured via X-ray diffraction and data for all conditions are presented in Figure 2. These data show that the amount of retained austenite is decreased with a magnetic field applied and that this is also true when the field is applied subsequent to the quench.

%

40

4 35 W

30

.z 25

B 20 01

4

15 a c: 10

3 5 2

0 0

5

10

15

20

25

30

Figure 2. Retained austenite in a 52100 alloy steel for different applied field strengths. The magnetic field is applied either during (hollow symbols) or post (filled symbols) quench.

The second experiment using 52100 steel was designed to evaluate the influence of a magnetic field on isothermal transformation behavior. After austenitization at 950 "C, the sample was cooled to 740 "C and held for 5 minutes. The magnetic field was applied during the 740 "C hold and subsequent quench. Figure 3 shows the microstructures obtained for the 52100 alloy for the conditions of without (Figure 3a) and with (Figure 3b) a 30 T magnetic field. Although the no-field sample shows some evidence of isothermal transformation at prior austenite grain boundaries, the microstructure is predominately martensitic. However, when the 30 T magnetic field is applied,

59

the resulting microstructure is very fine pearlite. Clearly, the presence of the magnetic field, even for very short times, dramatically accelerates austenite decomposition.

, . \

(b)

03-2507-04

52100-D

03-250946

52100-016

S?45pm

Figure 3. 52100 microstmctures obtained after holding at 740 "C for 5 minutes without and with a 30 T magnetic field. The no-field condition (a) is dominantly martensite (and retained austenite) with some grain boundary and spheroidal carbide present; the 30 T condition (b) is very fine pearlite.

60

4.

Transformation Kinetics in 1045 Steel

Experiments performed with 1045 steel investigated the influence of a magnetic field on austenite decomposition during continuous cooling. Figure 4 plots specimen temperature during cooling at three different cooling rates with and without a 30 T magnetic field. For each plot, thermal recalescence due to the release of latent heat associated with austenite decomposition is observed. However, a constant and significant shift in the recalescence is observed for the samples exposed to a 30 T magnetic field. Analysis of the transient cooling rates indicates an average temperature shift of -80 "C. In each case, the volume fraction of proeutectoid ferrite in the final microstructure increased by approximately 0.3.

'iunc h e c ) Figure 4. Cooling curves for various cooling rates with (solid l i e ) and without (dashed l i e ) a 30 T magnetic field for the 1045 steel specimens. Specimens held at 850 "C for 5 minutes prior to cooling.

Additional experiments were run on the 1045 steel to quantify the shift in transformation temperature as a function of magnetic field strength. In these experiments, a constant cooling rate was used and the magnetic field strength was held at 0, 10, 20, and 30 T. Temperature histories for these experiments are plotted in Figure 5. This figure shows a monotonic increase in transformation temperature as the magnetic field strength is increased. Using the instantaneous cooling rate, shifts in the transformation temperature as a function of magnetic field strength were determined. An estimate of the apparent Gibbs free energy change due to the presence of a magnetic field, AGB, can be determined. Starting with the fundamental

61 relationship between the change in free energy and temperature for the y-a transformation, Equation (1) is derived assuming 1) that the enthalpies of the parent and product phases are nearly equal and 2) the entropy is not affected by the magnetic field.

Here, is the enthalpy change and TE is the equilibrium temperature. By assuming the magnetic field has an additive contribution to AG"", this contribution can be estimated from the apparent decrease in undercooling as shown in Equation (2). my'"

A G =~-(TO

- T)

TE

(P-T)is the shift in transformation temperature from the no-field case (P).Applying values from the literature for AW'" of 4,200 J h o l [7] and

Here,

T = A,-l = 720 "C [S],a AGB for each magnetic field strength was estimated. The resultssof these calculations are summarized in Figure 6. 850 1045 Steel

750 800

700 650

-

600

-

550

-

5 0 0 ~ ' . .1 " . . . "2 . . . " .3 . . " " .4 ' Time (sec) Figure 5. Continuous cooling temperature data for the 1045 low carbon steel alloy for different magnetic field strengths.

62

400

300 D

100

0

Figure 6. Plot showing transformation temperature shifts and AGB as a function of magnetic field strength.

The figure plots measured values for AT and the associated AGB for each magnetic field strength along with a linear fit of the data. The fit indicates that the transformation temperature is shifted 3 "C/T by the magnetic field and the change in free energy is 12.6 JlmoYT. 5.

Unique Microstructure in a Hard Bainitic Steel

Slow cooling experiments were performed using a steel chemistry designed to produce very strong bainite when isothermally transformed at the proper temperature. The purpose of the experiments conducted on the "super bainite" steel was to determine if the application of a magnetic field during continuous cooling could accelerate austenite decomposition. The procedure called for samples to be heated to 1000 "C, held for 3 minutes and cooled at 1 "CISto ambient temperature. The magnetic field was applied during the 1000 "C hold and remained during cooling for the magnetically processed condition. Figure 7 shows optical metallography performed on the specimens for (a) the no field condition and (b) the sample cooled with a 30 T magnetic field. The no field condition has a fully martensitic microstructure, even for this relatively slow 1 "Cls cooling rate. In contrast, the microstructure for the sample cooled with the magnetic field is substantially different. Subsequent hardness measurements and TEM analysis have identified the microstructure as a fine pearlite with an interlamellar spacing on the order of 0.2 microns. The no field

63

04-0101-03

SP11C-1 30TSC

a 1Opm

7(4

04-0100-03

SPllC-1 NFSC

39pTi Opm

7(b) Figure 7. Microstructures of the “super bainite” steel formed at a cooling rate of 1 ‘CIS: (a) “no field” condition showing a martensitic microstructure; (b) 30 T condition showing a very fine pearlite.

microstructure transformed at -200 “C while temperature measurements indicate that with a 30 T field, the transformation occurred at -700 “C. These results

64

supplement previous data indicating accelerated austenite decomposition and show that the final microstructure can be completely altered by an external magnetic field.

6. Summary and Conclusions The results of this research demonstrate the influence that magnetic field processing can have on the resulting microstructure in a broad spectrum of ferromagnetic alloys. This shift in thermodynamics was observed as shifts in phase transformation kinetics and in the final microstructure. The Fe-15Ni results show that the thermodynamics of the system are significantly altered by a magnetic field as evidenced by changes in the volume fractions and solid solubilities of the a and y phases at 502 "C. The Ni solubility in austenite increased from 28 at. % to 40 at. % and the austenite volume fraction decreased from 30 % to 15 % due to a 30 T magnetic field. These results validate the thermodynamic argument that equilibrium between ferromagnetic and paramagnetic phases is altered by an external magnetic field. Various continuous cooling and isothermal hold experiments were conducted with several steels representing a wide range of compositions. Conclusions that can be summarized from the experiments are that magnetic field processing can: 1) reduce retained austenite in the final microstructure; 2 ) increase transformation temperatures and accelerate austenite decomposition during continuous cooling; 3) accelerate austenite decomposition during isothermal transformation; and 4) influence the equilibrium and metastable phase decomposition transformation sequences. Results from experiments with the 1045 steel indicate that the transformation temperature shift during continuous cooling due to magnetic field is approximately 3 "C/T and the associated shift in the Gibbs free energy is 12.6 JlmoVT. These results tap the significant promise of magnetic field processing to open new avenues in ferrous metallurgy to develop custom alloys with novel microstructures and enhanced properties and performance. The concept of 3dimensional phase diagrams that literally adds a new dimension to materials research development will provide an important tool for advancing alloy and process development in the ferrous metals industry.

65

Acknowledgements Dr. Sudarsanam Suresh Babu is thanked for his input and for providing the bainitic steel specimens. This research could not have been accomplished without access to high magnetic field facilities through the NSF and State of Florida supported National High Magnetic Field Laboratory User Program at Florida State University in Tallahassee, Florida. The authors wish to acknowledge Dr. Bruce Brandt, Ms. Merry Ann Johnson and the staff at the NHMFL for their efforts and support. T h ~ sresearch was sponsored by the U.S. Department of Energy, under contract DE-AC05-000R22725 with UT-Battelle, LLC through the Industrial Materials for the Future and Laboratory Directed Research and Development Programs.

References 1. Garcia-Mateo, C., Cabellero, F.G., Bhadeshia H.K.D.H., ISIJ International, 43 (2003), pp. 1821-1825. 2. Ludtka, G.M., J~amillo,R.A., Kisner, R.A., Nicholson, D.M., Wilgen, J.B., Mackiewicz-Ludtka, G., Kalu, P.K., Scripta Muter., 51 (2004), pp. 171-174. 3. Nicholson, D.M., Kisner, R.A., Ludtka, G.M., Sparks, C.J., G., Petit, L., Roger Jaramillo, R.A., G. Mackiewicz-Ludtka, G., Wilgen, J.B., Sheikh-Ali, A., Kalu, P.K., J. Appl. Physics, in press (accepted for publication). 4. Chuang, Y-Y., Chang, Y.A., Schmid, R., Lin, J-C., Metall. Muter. Trans., 17A (1986) p. 1361. 5 . Wang, Y., Stocks, G.M., Shelton, W.A., Nicholson, D.M., Phys. Rev. Lett., 75 (1995), p. 2867. 6. Zhang, J., Williams, D.B., Goldstein, J.I., Metall. Muter. Trans., 25A, (1994), p. 1639. 7. Bammann, D.J., et al., 2"d Int'l Conference on Quenching and the Control of Distortion, (1996), p. 367. 8. Andrews, K.W., J. Iron Steel Inst., 203 (1965), pp. 721-727.

FUNDAMENTALS AND APPLICATIONS OF GRAIN BOUNDARY DYNAMICS IN HIGH MAGNETIC FIELDS D.A. MOLODOV Institute of Physical Metallurgy and Metal Physics, Aachen University, 52056 Aachen, GERMANY

The latest research on dynamics of grain boundaries in non-magnetic materials in high magnetic fields is reviewed. A control of grain boundaxy migration means control of microstructure evolution, which is a key for the design of materials with desired properties. Grain boundary motion can be affected by a magnetic field if the anisotropy of the magnetic susceptibility generates a gradtent of the magnetic free energy. In contrast to curvature driven boundary motion, a magnetic driving force also acts on planar boundaries so that the motion of crystallographicallywell-defined boundaries can be investigated and the true grain boundary mobility can be determined. The results of migration measurements obtained on bismuth and zinc bicrystals are addressed. Selective growth of new grains in locally deformed zinc single crystals driven by a magnetic force is reported as well. Finally, implications for materials processing, particularly the effect of magnetic fields on texture development in hexagonal metals, are discussed.

1. Introduction Since recrystallization proceeds by generation and migration of grain boundaries during heat treatment of deformed structures, grain boundary motion is the dominating process of microstructure formation during annealing of cold worked materials. The development of the microstructure during grain growth is caused both by the change of average grain size and, what is of concern for crystallographic texture evolution, by the change of the grains’ orientation and misorientation distribution in the grain structure. The latter circumstance is due to the large difference in the properties of grain boundaries with different crystallographic parameters, primarily in the boundary mobility. A control of grain boundary migration means control of microstructure evolution, critical for the design of materials with tailored properties. Grain boundary motion can be affected by a magnetic field owing to a driving force induced by the susceptibility difference due to crystal and shape magnetic anisotropy. Since the investigation by Smoluchowski and Turner (1949) [l] of the preferred orientation produced by annealing an iron-cobalt alloy in a magnetic field and the work of Goetz (1950) [2], who observed the orienting effect of a magnetic field on bismuth during its solidification, the 66

67

application of high magnetic fields has been established as one of the advantageous technologies in materials processing. Magnetic fields have been reported to have a substantial effect on texture and microstructure evolution in ferromagnetic materials. Boothby el al. [3] found polycrystalline MnBi to recrystallize when being subjected to a magnetic field at a temperature below the Curie point. A magnetic annealing effect has been later observed by Chen and Stutius [4] in a MnBi single crystal. Marticainen and Lindroos [5] have studied the effect of a magnetic field on recrystallization in iron. They observed selective nucleation upon annealing in a magnetic field resulted in the retardation of the recrystallization process. Watanabe et al. [6] also reported the retardation of recrystallization and grain growth in iron-cobalt alloy under the annealing in a magnetic field. Later Masahashi et al. [7] found out that a magnetic field affects the development of recrystallization texture in kon-3.25% silicon and favors the Occurrence of low energy grain boundaries in the recrystallized microstructure. Quite recently Bacaltchuk et al. [8] observed that an application of a magnetic field affected texture development in silicon steel by increasing the strength of the Goss component and decreasing the intensity of the gamma fiber. The influence of a magnetic field on sintering of iron powder has been studied by Tsurekawa et al. [9], who observed a significant enhancement of grain growth by application of a magnetic field resulting in an increase of the compact density. Recently, Harada el al. [ 101 reported an effect of annealing in a magnetic field on microstructure development in electrodeposited nanocrystalline nickel. The microstructure evolution in the course of recrystallization and grain growth can be affected by a high magnetic field not only in ferromagnetic materials, but also in non-magnetic ones, such as paramagnetic and diamagnetic. The effect of a magnetic field on metallurgical phenomena in these materials has been conventionally considered as negligible and, correspondingly, much fewer data concerning this issue can be found in the literature. An application of a magnetic field for crystal alignment in A1-Cu and Cd-Zn alloys during solidification has been reported by Mikelson and Karklin [ 113. Recently, Sugiyama et al. [ 121 have experimentally shown that the orientation of crystals in zinc and bismuth-tin alloy can be controlled by reheating up to a solid-liquid zone in the presence of a magnetic field. Two efforts were previously made to apply the magnetic field method to study grain boundary kinetics in non-magnetic materials. The orienting effect of a magnetic field on Bi during grain growth was investigated by Mullins [13]. Fraser et al. [ 141 showed that a grain, which was generated by recrystallization

68

following a local deformation of a bismuth single-crystal, grew and consumed the initial single-crystal under the action of a magnetic force. However, no specific boundary motion was investigated. Moreover, the magnetic field in the experiments of Mullins and Fraser et al. was always imposed on the growing grains, i.e. a magnetic force was applied to boundaries that were already moving under the force exerted by their surface tension and curvature. In experiments to which we will refer in the following sections, grain boundary motion was measured by applying a magnetic field to specially prepared bicrystals, i.e. to systems that were in equilibrium without an applied magnetic field [15,16].

2. Magnetic Driving Force in Non-Magnetic Materials

A force, p, on a grain boundary of area, S, arises if the boundary displacement, d l , perpendicular to the boundary element, S, results in a decrease of the Gibbs free energy, dG:

where, V is the volume swept by the boundary element during its displacement. The gain of free energy can result from a change of the grain boundary energy due to a reduction of the grain boundary area in case of curved capillary driven boundaries, or from a volume free energy difference across the boundary e.g. by the anisotropy of free energy density in an external field. If the volume density of the magnetic free energy, W, in a crystal induced by a uniform magnetic field is independent of the crystal shape and size, then the magnetic force, p, acting on the boundary separating two adjacent crystals having different susceptibilities is given by [ 131

where x1and ~0 are the magnetic susceptibilities of neighboring crystals 1 and 2, respectively, along the magnetic field, H. This driving force depends only on the strength of the magnetic field and its orientation with respect to the two neighboring crystals. The direction of p remains the same when the sense of the field is reversed. According to elementary crystallography, the magnetic susceptibility of a single crystal can be written as X = X i + AX COS* 6, where AX is the difference in susceptibilities parallel X I I and perpendicular X i to the principal (or c) axis. Here 6 is the angle between the c axis and the magnetic field. Substituting this expression into Equation (2) leads to

69

(3) A maximum magnetic force P,

1

= -AxH2

2

(4)

occurs when the angles between the field and the c axis in both neighboring grains are 81= 0 and 82= 90". For, p = 0 the grain orientations need not be identical but only 81= 6.The sign of P depends on the magnetic anisotropy of a material (AX) and the asymmetry of the spatial orientation of both neighboring grains with respect to the magnetic field direction.

3. Absolute Grain Boundary Mobility in Bismuth Bicrystals Since the magnetic driving force is induced by an external field and does not depend on boundary properties, the method provides an opportunity to investigate the motion of a planar grain boundary with well-defined structure. The motion of such boundaries was measured by applying a strong magnetic field to specially prepared bismuth bicrystals [ 15,161. Bismuth is a most suitable material for a model experiment to measure grain boundary migration driven by the magnetic force, since it possesses the largest magnetic anisotropy with different susceptibilities parallel and perpendi~ ; 252"C, AX = 0 . 2 3 ~ 1 0 cular to the trigonal axis (at 22"C, AX = 0 . 5 3 ~ 1 0 at [17]. As shown by Kapitza [18], the susceptibility of bismuth does not depend on H up to 2x107A/m, and the energies associated with magnetostriction at 0.8x107A/m are less than 1% of the magnetic free energies [19] and thus, can be neglected. The experiments [15,16] were carried out on specially grown bicrystals of high purity (99.999%) bismuth. Symmetrical and asymmetrical pure tilt boundaries with 90" misorientation were examined. The misorientation angle between trigonal axes in both crystals of the bicrystal was chosen to be 90" in order to gain the maximum possible magnetic driving force (Eq. (2)). Field strength applied ranged between 0 . 8 0 ~ 1 0and ~ 1 . 5 9 ~ 1 0A/m. ~

70

Figure 1. Gram boundary displacement (from A to B) m a bismuth bicrystal under a magnetic dnving force of H=1.63.107 A/m after annealmg for 180 s at 252°C [15].

a

1' I

H c

Figure 2. Movement of the same grain boundary in opposite Qrection m a bicrystal Qfferently oriented with regard to the magnetic field.

The experiments unambiguously confirmed that grain boundaries in bismuth bicrystals actually move under the action of a magnetic driving force. To prove that boundary motion was caused exclusively by the magnetic driving force, the experimental procedures were varied. First, the specimen was annealed without a magnetic field to probe the effect of the orientation dependent surface energy. In these experiments the grain boundary did not move, proving that the surface energy did not substantially contribute to the driving force. Second, a specimen was mounted on a holder such that the c axis ()of crystal 1 was directed parallel to the field (Figure 2a). The axis in crystal 2 in this case was perpendicular to the field, and the grain boundary moved in the direction of the latter crystal due to its higher magnetic

71

free energy density. Third, a specimen was mounted in a position where the axis < 1 1 b in crystal 2 was close to the field direction and the corresponding axis in crystal 1 was perpendicular to the field. The direction of boundary motion in this case was opposite compared to the previous case, i.e., from crystal 2 toward crystal 1 (Figure 2b). This result provides unambiguous evidence that the grain boundaries in our bicrystals moved due to the magnetic driving force only. In addition, some bicrystals were annealed in a magnetic field in both positions which could be attained by specimen rotation - and boundary motion in opposite directions was observed in the same specimen dependent on its position with regard to the direction of the magnetic field.

0.3

p [kJlms]

Figure 3. Dependence of the velocity of a 90" symmetrical tilt boundary in bismuth on the magnetic driving force at 250°C [16].

The freedom to change the magnitude of the driving force for boundary migration by exposing the samples to magnetic fields of different strength yields the unique opportunity to change the driving force on a specific grain boundary and thus, to obtain the driving force dependence of grain boundary velocity (Figure 3). This dependence is one of the most fundamental relations of the theory of grain boundary motion and is repeatedly discussed in literature [20,21]. Some authors question the linearity of the velocity-driving force relationship [22], but there is overwhelming experimental evidence that the boundary velocity does depend linearly on the driving force [21]. The measurements of boundary migration in different magnetic fields confirm that the boundary velocity changes linearly with the driving force v = mp (Figure 3). Therefore, from current results, the absolute value of the grain boundary mobility m = v/p can be immediately extracted, and the dependence of grain

72

boundary mobility on temperature and on the specific grain boundary character can be determined. The temperature dependence of the mobility m = rno exp(-Q/kT) of a symmetrical and an asymmetrical 90" tilt grain boundary revealed that the migration parameters (activation enthalpy Q and mobility pre-exponential factor mo) for the symmetrical boundary (Q=0.51 eV, mo=0.67 m4/J.s) drastically differed from the migration parameters for the asymmetrical boundary (Qll=3.38 eV, rnol~=2.04x1024 m4/J.s and Q1=3.79 eV, rno~=1.10x1028 m4/J.s, refer to the orientation of the c-axis with regard to where the symbols 1) and I the boundary plane normal, as referred to below). Consequently, the symmetrical boundary has a much higher mobility than the asymmetrical boundary in the entire investigated temperature range up to the melting point of bismuth (Figure 4)but particularly at low temperatures. Apparently, the inclination of the tilt boundary in Bi has a very strong influence on boundary mobility.

10'

:

u)

2

E E

Y

:

1.8

1.85

1.9

1.95

2

2.05

2.1

l O 3 K [K-11

Figure 4. Temperature dependence of mobility of 90" symmetrical ( 0 )and asymmetrical (A, I boundaries ) in bismuth-bicrystals. Trigonal axis in the growing grain is parallel (A) or perpendicular).( to the growth direction [16].

The most surprising feature is that in contrast to the symmetrical boundary, for an asymmetrical tilt boundary the measured mobility was found to be distinctly different for motion in opposite directions (Figure 5). For the chosen crystallography of the bicrystals the boundary was less mobile when the c (411>) axis in the growing grain was perpendicular to the direction of motion ( m i = 8.2~10.' m4/J.s at 252°C) but faster, when the trigonal c axis in the growing grain was close to the direction of motion (mil = 1 . 3 ~ 1 0 m4/J.s -~ at 252°C). There are several potential reasons for this anisotropy. First, there is an essential difference in the distance between the crystallographic planes on each

73

side of the boundary. An estimate shows that this factor may change the velocity of grain boundary motion, however this difference is unlikely to affect the velocity of boundary motion by more than 20%, which is distinctly less than the observed effect [15]. Second, because boundary motion in Bi-bicrystals may be influenced by impurity drag, the difference in the diffusivity of impurities in two opposite directions in the anisotropic structure of Bi should be taken into account. Finally, as shown recently, the motion of a grain boundary in a magnetic field can be considered as motion of a conductor in a magnetic field or, more strictly, as the motion of a region with a conductivity different from that of the surrounding matrix in a magnetic field. Such motion causes an electromotive force and, consequently, an additional dissipation of energy in a magnetic field [23]. This dissipation means a decrease of the effective driving force for grain boundary motion. With respect to the proportionality between grain boundary migration rate and the applied driving force, this appears as decreased grain boundary mobility. This effect, however, depends on orientation of the magnetic field with regard to the crystal axes and, therefore, on the direction of motion. Consequently, this causes a different apparent mobility for boundary motion in opposite directions. The only exception is, of course, a symmetrical tilt boundary. In this case, the boundary plane represents a mirror plane to both adjacent crystals, thus the relative orientation of the magnetic field to the crystal axes is identical on both sides of the boundary. The calculated difference of the mobility in bismuth is of the order of 1% [23], while the experimentally observed effects are of the order of 50%. Therefore, the observed mobility asymmetry for asymmetric tilt boundaries in Bi must be attributed to other reasons, and is definitely not caused by induced magnetic moments during grain boundary motion in a magnetic field. In any event, if this asymmetry of grain boundary mobility also holds for other metals, it will have a serious impact on our understanding of grain boundary motion, since the mobility of a grain boundary is commonly conceived as not dependent on its direction of motion.

74 5

0

0

in growing grain perpendicular to direction of motion 4 1 1> in growing grain parallel to direction of motion

100

200

300

400

t [SI

Figure 5. Displacement (normalized by driving force) - time diagrams for 90" was observed to be quite mobile [26,27]. In the experiment [26] during annealing in a field of A/m at 390"C, the boundary experiences a simultaneous action of two different driving forces for boundary migration (Figure 6): a magnetic driving force which moves the boundary in the direction of grain A and a curvature driving force which acts in the direction of grain B on the left half of the specimen and in the direction of grain A on its right half. Therefore, the effective driving force for boundary motion on the left specimen side is pefl= pm- pc and on its right half pen = pm + pc. The action of the capillary driving forces in the opposite directions at the boundary ends combined with the magnetic driving force reorients the boundary decreasing its length, while the magnetic force acting in the direction of grain A is mostly responsible for the boundary displacement in the longitudinal direction (Figure 6). It is easy to see that the middle of the boundary does not experience an action of the capillary driving force, since the capillary forces acting in the opposite directions compensate each other at that point. Therefore, in the experiment with a magnetic field the center of the boundary is moving under the action of the magnetic force, pm, only and its displacement can be used for measuring the absolute grain boundary mobility. The mobility of the investigated 89" < 1070 > tilt boundary in a Zn bicrystal was found to be m=5.1.10-9 m4/J.s. For a comparison, the reduced mobility of the curved 86" < 1070 > tilt boundary A=m.a at 673 K in a Zn bicrystal was measured to be AF=3.2.10-' m2/s [28] yielding an absolute mobility of =7.0.10-' m4/J.s [27]. It is worth noting that the mobility of 89" < 1070 > tilt grain boundaries with the same misorientation but different boundary orientations was found to be distinctly different depending on the degree of boundary asymmetry, namely, the boundary with larger inclination with respect to symmetrical boundary position moves slower [27].

rnrK

5. Magnetic-Field-Induced Selective Grain Growth in Zinc The application of a magnetic field to a locally deformed zinc single crystal resulted in growth of a new specifically oriented grain, i.e., in the generation and motion of an individual grain boundary [29]. For this, one of the sample surfaces was deformed by a single macro Vickers hardness indentation. A successive heat treatment resulted in a local recrystallization of the deformed volume only. Then a magnetic field was applied parallel to the principal axis of this single

76

crystal. Due to the magnetic anisotropy of zinc, a magnetic driving force acted on all grain boundaries between the host matrix and those new grains whose principal axis was not parallel to the field. However, usually only one boundary, whose product of mobility and driving force was maximal, swept the sample. Grains having their c-axis perpendicular to the field direction experience the maximum magnetic driving force. In fact, selectively grown grains in this experiment showed a strong preferential orientation. This is illustrated by an inverse stereographic projection of the field direction in these grains (Figure 7a) and by the frequency distribution of the angles between c-axes in new grains and the field direction (Figure 7b) [29]. 0.04,

100

. ,

.

,

.

,

.

,

. ,

m

06

U

2

1.0

08

6 003

--

-

P a

002 04

2

0 01

02

nn

O W

inclination [deg]

a

0

15

30

45

60

75

90

inclination [deg]

b

Figure 7. (a) Inverse stereographic projection of the magnetic field direction in new grains after magnetic annealing; (b) relative frequency of inclination angles shown in Figure 7a and normalized magnetic driving force (right axis) for the grain growth as a function of the inclination angle. Open circles denote a random inclination angle distribution [29].

One must note that the character distribution of the boundaries between new grains and the host matrix was also found to be far from random. Possible reasons of such behaviour are a misorientation dependence of the boundary mobility or a dependence of the nuclei orientation on the deformation mode.

6. Control of Crystallographic Texture by a High Magnetic Field The observation of a magnetically caused selective grain growth clearly indicates the possibility of affecting the microstructure development in magnetically anisotropic non-magnetic materials by means of annealing in a magnetic field. This was recently unambiguously proved in experiments with zinc-1.1% aluminum [30] and titanium [31]. Zinc-1.1% aluminum alloy sheet samples after 99% reduction by rolling were annealed (55 min at 390°C) in a high magnetic field of 32 T [30]. The specimens were oriented differently with respect to the field. For oRe set of specimens the rolling direction (RD) was parallel to the field direction, and other

77

specimens were tilted by +19" and -19" to the field direction around the transverse direction. As seen from the {0002) pole figure after cold rolling (Figure 8a), two components with the principal c-axes are tilted by 15-20" from normal to the rolling direction around the transverse direction. Annealing such specimens without a magnetic field slightly changes intensities of the texture components but retains the original type of the pole figure (Figure 8b). In contrast, annealing in a magnetic field drastically changes the texture dependent on the specimen orientation with respect to the field direction. After magnetic annealing of specimens tilted by + 19" the B-component totally disappears, while the intensity of the A-component rises (Figure 8c). In contrast, in the case of specimens tilted by - 19" the A-component disappears completely and the Bcomponent becomes much more intense (Figure 8d). It is easy to see that the caxes of differently oriented specimens are tilted by three different angles to the field and the magnetic annealing results in an increase of the texture peak, which corresponds to grains with the c-axis perpendicular to the field.

a

b

C

d

Figure 8. [ 00021 pole figures of 99%-rolled zinc-I .1 % aluminum sheet samples (a) before and (b) after annealing at 390°C without a magnetic field, (c) in a magnetic field of 32 T tilted by +19" to the field around the TD and (d) tilted by -19" to the field around the TD [30].

A similar effect was observed in an investigation of the effect of a magnetic field on texture development in titanium [31]. The annealing of titanium sheet samples after an 82% reduction by rolling at 750°C in a magnetic field of 19.4 T for sheet normals being either parallel or perpendicular to the magnetic field direction does not change the final texture (Figure 9). In contrast, the same heat treatment in the same magnetic field results in a distinct difference between usually symmetrical texture peaks when the sample is tilted by +30° or -30" to the field direction around the rolling direction leading to a configuration where the c-axis of grains corresponding to one texture component is aligned normal to the field direction.

78 (0002)

{0002)

RD

RD

TD

Intensity levels: 1.0 1.5 2.0 2.6 ( m a 2.70)

TD

Intensity levels: 1.0 1.5 2.0 2.5 (max4.99) 3.5 4.5 4.9

TD

Intensity levels: 1.0 1.5 2.0 2.5 (max5.68) 3.5 4.5 5.6

Figure 9. (0002) pole figures and pole density along TD (a) for 82%rolled titanium sheet samples; (b) after annealing at 750°C;(c) after annealing in the field of 19.4 T at 750°C with TD tilted by 30" to the field direction [31].

Apparently, the observed behavior in both diamagnetic zinc and paramagnetic titanium is due to an additional driving force for grain growth arising in the magnetic field by the anisotropy of the magnetic susceptibility of these materials [30,31]. The results obtained demonstrate that the texture in zinc and titanium can be effectively changed by means of annealing in a high magnetic field and present a new method to control the development of crystallographic texture in magnetically anisotropic non-magnetic materials. Acknowledgements The author expresses his gratitude to the Deutsche Forschungsgemeinschaft for financial support through grant MO 848/4-1, as well as to the National High Magnetic Field Laboratory for using the facilities and financial support under the Visiting Scientist Program. References 1. 2. 3. 4.

Smoluchowski R. and Turner R. W., J. Appl. Phys., 20 (1949), p. 745. Goetz A., Phys. Rev., 35 (1950), p. 193. Boothby O.L., Wenny D.H., Thomas E.E., J. Appl. Phys., 29 (1958), p. 353. Chen T., Stutius W.E., J. Appl. Phys, 45 (1974), p. 4622.

79

5. Marticainen H.O., Lindroos V.K., S cu d . J. Metallurgy, 10 (1981), p. 3. 6. Watanabe T., Suzuki Y., Tanii S., Oikawa H., Phil. Mug. Letters, 62 (1990), p. 9. 7. Masahashi N., Matsuo M., Watanabe K., J. Mat. Research, 13 (1998), p. 457. 8. Bacaltchuk C.M.B., Castello-Blanco G.A., Ebrahimi M., Garmestani H., Rollett A.D., Scriptu Muteriuliu, 48 (2003), p. 1343. 9. Tsurekawa S., Harada K., Sasaki T., Matsuzaki T., Watanabe T., Mater. Trans. JIM, 41 (2000), p. 991. 10.Harada K., Tsurekawa S., Watanabe T., Palumbo G., Scriptu Muteriuliu, 49 (2003), p. 367. 11.Mikelson A.E., Karklin Ya.Kh., J. Cryst. Growth, 52 (198 l), p. 524. 12.Sugiyma T., Tahashi M., Sassa K., Asai S . , ZSZJZnt.,43 (2003), p. 855. 13.Mullins W.W.,Actu Metull., 4 (1956), p. 421. 14.Fraser M.J., Gold R.E. and Mullins W.W., Actu Metull., 9 (1961), p. 960. 15.Molodov D.A., Gottstein G., Heringhaus F., Shvindlerman L.S., Scriptu Muter., 37 (1997), p. 1207. 16.Molodov D.A., Gottstein G., Heringhaus F., Shvindlerman L.S., Actu Muter., 46 (1998), p. 5627. 17.Goetz A. and Focke A., Phys. Rev., 45 (1934), p. 170. 18.KapitzaP., Proc. Roy. SOC.(London), A131 (1931), p. 224. 19.Kapitza P., Proc. Roy. SOC.(London), A135 (1932), p. 556. 20.Burke J.E., Turnbull D., in Progress in Metal Physics, ed. B. Chalmers (Pergamon Press, New York, 1952). 21.Gottstein G., Shvindlerman L.S., Scriptu Metull. Muter., 27 (1992), p. 1521. 22.Rath B.B., Hu H., in The Nature and Behaviour of Grain Boundaries, ed. H. Hu (Plenum Press, New York, 1972), p. 405. 23.Gottstein G., Molodov D.A., Rabkin E., Shvindlerman L.S., Snapiro I., ZnterfQce Sci., 10 (2002), p. 279. 24.Huning F., Molodov D.A., unpublished work (1999). 25.Konijnenberg P.J., Molodov D.A., Gottstein G., Shvindlerman L. S., 2000 NHMFL Annual Research Review, p. 266. 26.Sheikh-Ali A.D., Molodov D.A., Garmestani H., Appl. Phys. k t t e r s , 82 (2003), p. 3005. 27.Sheikh-Ali A.D., Molodov D.A., Garmestani H., Scriptu Muter., 48 (2003), p. 483. 28.Sursaeva V.G., Andreeva A.V., Kopetskii Ch.V., Shvindlerman L.S., Phys. Met. Metull., 41 (1976), p. 4103. 29.Konijnenberg P.J., Molodov D.A., Shvindlerman L.S., Gottstein G., 2002 NHMFL Annual Research Review, p. 266. 30. Sheikh-Ali A.D., Molodov D.A., Garmestani H., Scriptu Muter., 46 (2002), p. 857. 31.Molodov D.A., Sheikh-Ali A.D., Actu Muter., 52 (2004), p. 4377.

ENHANCEMENT OF TEXTURE AND CRITICAL CURRENT DENSITY IN BizSrzCalCuzOS SUPERCONDUCTING TAPES THROUGH MAGNETIC FIELD PROCESSING P.V.P.S.S. SASTRY,U.P.TROCIEWITZ,H. MAEDA, J. SCHWARTZ National High Magnetic Field Laboratory and Centerfor Advanced Power Systems, Florida State University, Tallahassee, Florida, USA Ag-sheathed Bi-2212 tapes with varying core thickness were heat treated in flowing 0 2 in magnetic fields up to 15 T. A uniform, high-degree of texture is achieved throughout the thickness of the tapes when heat treated in a magnetic field, whereas a large portion of non-textured region exists without magnetic field. The critical current density J, of the tapes increases with increasing field strength due to enhancement in texture. The self field critical current I, > 1000 A is achieved for tapes with a core thickness of 180 pm. For tapes with larger core thicknesses, however, I, decreases due to inhomogeneous melting. The magnetic field is more effective in enhancing texture in the early stages of crystal growth.

1. Introduction

Among the many high temperature superconductors, Ag-clad Bi2Sr2CalCuzOs (Bi-22 12) has great potential for high field magnet applications, and particularly for NMR applications [l-41. Fabrication of long lengths of multifilament Ag/Bi2212 tapes has been developed with the use of the power-in-tube technique [561. Heat treatment of Bi-2212 tapes involves a partial melt process to achieve textured microstructures. High degree of anisotropy in the superconducting properties makes it necessary to have texture in Bi-2212 tapes in order to achieve high critical current density and to mitigate the effect of magnetic field on critical current density. The partial melt processing, aided by the superconductor-silver interface, produces high degree of texture in Bi-22 12 tapes if the individual filaments are thin. The fnaximum thickness of individual filaments is 10 pm to achieve texture [6]. For practical applications, however, it is necessary to increase the tape thickness for realizing large critical currents and to minimize the volume fraction of expensive silver. Increase in superconducting core thickness invariably lowers critical current density due to the lack of high degree of texture in the middle of the filaments. If the grain alignment in the center of the core is improved, enhancements in J, and I, are expected. Magnetic field processing ( M l T ) exploits the anisotropy in the paramagnetic susceptibility of the Bi-2212 crystal structure to align the grains [7-91. The magnetic field creates a torque on the crystallites, and during the melt processing stage, when 80

81 there is significant amount of liquid phase, aligns the crystallites with the c-axis parallel to the magnetic field. As the solidification and grain growth continues during the cooling stage, the process results in a well-textured microstructure with enhanced superconducting properties [ 101. Magnetic field processing has been used to enhance texture in bulk materials, thick films and metal-clad tapes of superconductors [ 10-201. It is relatively easier to achieve texture enhancement in thick bulk samples by MFP because bulk samples usually have a poor degree of texture when processed without magnetic field. Previous studies on h4FP reported in the literature have been carried out to prove the principle that the structural anisotropy of HTS materials can be used to enhance texture through MFP. Most studies reported have used magnetic fields less than 10 T, and the sample sizes were limited due to small sizes of room temperature bores available in typical superconducting magnets. The studies initiated at the National High Magnetic Field Laboratory (NHMFL) on MFP are focused on understanding the potential benefits of MFP in enhancing texture and critical current density of practical AgBi-22 12 superconducting tapes. Studies were carried out on 5 cm long tape samples to ensure reliable critical current density measurements. The samples were cut from 20-30 m long batches to check for reproducibility of the results. The core thickness of the tapes was varied between 80 and 700 pm to understand the effectiveness of MFP in enhancing texture in very thick AgBi-2212 tapes. The Bi-2212 tapes with a core thickness of 180 pm have been investigated in detail to understand the effect of magnetic field strength and duration during the heat treatments. There are several water-cooled, resistive magnets with varying field strengths and bore sizes available for MFP at the NHMFL. The maximum field strength of the magnets varies from 20 T to 45 T, and the available bore sizes vary from 32 mm to 195 m q in diameter. There are also a 3 T and a 6 T superconducting magnet used for MFF' with room temperature bores of 153 mm diameter. The studies reported in this article are carried out in the 20 T, 195 mm large bore resistive magnet [21]. The furnace used for the heat treatments was designed with the heating elements arranged around the bore parallel to the length of the furnace to minimize induced currents during the ramping up and down of the magnetic field. The furnace is powered by a DC power supply and the temperature is controlled by a PID-controller. The furnace has a 60 mm bore and a temperature gradient of about 10 "C across a ceramic sample holder, which provides slots for simultaneous heat treatment of 12 samples. By careful planning of the set temperature, it is possible to heat treat 6 sets of samples

82

(2 samples per set), each set at a different temperature within the 10 O C window. This kind of sample arrangement is very useful for heat treatment of Ag/Bi-2212 as the properties of the material are extremely sensitive to heat-treatment temperature. A quartz tube insert is used to maintain oxygen flow during the heat treatment. Figure 1 depicts a schematic of the experimental setup, a view of the magnet and the furnace, and the sample holder used for the heat treatments. 2. Experimental Procedure Ag-sheathed Bi-2212 monocore tapes with core thickness between 80 to 700 pm were prepared by the powder-in-tube method. Two batches of 180 pm thick tapes with core widths of 2.1 mm and 3.1 mm were fabricated. The 3.1 mm wide tape is referred to as the wide tape in the following sections. 40 mm long samples were set horizontally in the vertical tube furnace, which is installed in the 19 T, 195 mm bore resistive magnet and heat treated in applied vertical magnetic fields Ha up to 15 T. Heat treatments were carried out by the partial melt process in flowing 0 2 according to the temperature profile shown in Figure2. After holding at the temperature T, for 12 min, the samples were cooled rapidly by 10 "Cat a rate of 48 "C/h, then slowly down to 840 "Cat a rate of 6 ' C h , where the samples were held for 5 h. The total heat treatment time was about 22 h. Several heat treatment variations were evaluated before settling with the schedule described above. This heat treatment schedule is similar to the

Figure 1. A schematic and top view of the magnet and furnace assembly used for the MFP studies.

83

I

Time

Figure 2. Temperature profile used for magnetic field processing of Bi-2212 tapes.

process used by others [22] and is suitable for heat treating one batch of tapes per day and fits into the typical 8 h magnet time schedules of the NHMFL. A constant magnetic field was applied for 0.5-6 h starting at T,. One tape sample was arranged, with the broad surface perpendicular to the applied magnetic field direction, in each of the twelve rectangular shaped horizontal slots of the sample holder. The holder has a vertical cylindrical opening in the center to ensure continuous and uniform O2 flow; current density of Bi-2212 tapes depends strongly on the flow of 0 2 gas, which is affected by magnetic field due to the paramagnetic nature of 02.The sample holder was placed in a temperature gradient, so that each sample was heat treated at different T,,, at the same time. The variation in temperature from the bottom to the top of the sample holder is about 10 "C.

3. Results and Discussion Figure 3 shows representative SEM micrographs of a polished and etched longitudinal section of the tapes with a core thickness of 180 pm heat treated at the optimum temperature T, in magnetic fields of 0 and 10 T. In the tapes heat treated in a magnetic field of 10 T, a uniform high texture is found to exist throughout the thickness. In the tape heat treated without magnetic field, on the other hand, a large portion of non-textured area exists in the interior of the tape. A clear enhancement of texture is seen in the MFP samples. Furthermore, outgrowth of some crystals into the silver matrix can be seen in the tapes heat treated without field, whereas no such outgrowths are noticeable in the tape heat treated with field applied. This observation suggests that the magnetic field during MFP would also be beneficial in suppressing bridging of filaments and

84

"..

.,

"

--I

..

."

.. ". ,.

~

._"" .. .__,. "

........,,

..~" ....

Figure 3. SEM micrographs of the polished longitudinal sections of the tapes with a core thickness of 180 lm , heat treated at the optimum temperature T,,, in magnetic fields H. of (a) 0 T, (b) 10 T. Ha was applied vertically.

1000

5

800

a. The electrostatic contribution stabilizes the particle free energy above T,,,. The melting temperature T, of nuclei is obtained for AGv= a:

k=(T,-T,,,)/T,

= aV/A& with CI = 2yS&

(6)

R, can be determined from R,=R, by using Eq. (2). The surface energy ysl is often known and is of the order of 0.1-0.3 Joule/m2 for materials having a melting temperature of about 1000°C [2]. The solidification in the presence of nuclei is homogenous and takes place at T = T, as the critical radius R, = -2ysl/(AG,-a) for solidification becomes smaller than the nuclei radius R, = 2ysl/(cr). R, tends to 2ysl/AGv for large undercooling rates, which is in agreement with the classical theory of crystal nucleation [2].

8. Discussion of the Orders of Magnitude The critical radius of three materials has been determined using Eq. (2). A typical value of R, = R,= 404 10-I0mcorresponding to a SmZCol7type alloy was chosen. The experimental phenomenon could be well described by the addition of an electrostatic contribution to the particle Gibbs free energy if the orders of magnitude of the various parameters would be realistic. A contact potential Acp of 0.1 Volt corresponds to a chemical potential difference AW of 0.1 eV. The contact potential between two different metals can be much larger. Using Eq. (4), it is possible to determine:

ysl is assumed to be equal to 0.27 J/m2. Various values of p are used in Table 2. p is calculated as % for charged colloids [12]. The screening charges could be more easily distributed in the liquid to maximize their mutual distances. In metals, the electrons are also free to move. The lattice existence and surface imperfections prevent charges from maximizing their distances. Values of f3 are

108 Table 2. Orders of magnitude for a material having R, = R,= 40410-'0 m: p is a numerical coefficient reducing the electrostatic interaction in the double layer of conduction electrons; R, i s the particle critical radius above the melting temperature; a is the energy per unit of volume deduced from R, and ysl through Eq. (6) and which is subtracted from the Gibbs free energy in Eq. (5) to explain the existence of nuclei above T,; Ar is the solid-liquid interface thickness; A ~ isJ the expected contact potential between the solid and its melt in function of 0; AqVAr is calculated using Eq. (7); the charge number Z, carried by a particle having the radius R, is calculated using Eq. (3); the surface charge is deduced from Eq. (8).

chosen between 1 and 0.1 in Table 2. The charge number Z, is calculated by using Eq. (3) and the charge per surface unit is deduced:

Z= ,

0.95'10'8R2(3-1Rys~1R. (5 = 1.2'10-' (3-'nys11nUrn'

(8)

Z, increases from 805 to 2550, Acp varies from 0.071 to 0.225 volts when Ar= lo-'' m. The necessary difference of work functions appears as being reasonable for all fl values. The charge per surface unit is equal to 62.10-4and 199.10-4C/m2 for fl = 1 and 0.1, corresponding respectively to one charge per (51 8)' and per (28 These calculations show that the supplementary term needed in the Gibbs free energy to stabilize nuclei above T, could be produced by an electrostatic double layer induced by a small difference in the work functions of a solid and its melt. It is also possible to predict a good value of the critical temperature T, of the nuclei disappearance by using Eq. (6); T,-T, = 7.7 K assuming [email protected] AHf = ys~'V"NA". Turnbull suggested this approximation. AHf , ySl, V and NA are the gram-atomic fusion heat, the gram-atomic interfacial energy, the gramatomic volume and the Avogadro number, respectively [2].

9. Conclusion Several demonstrations have been made that magnetic processing is very efficient to texture materials 'from their melts. The anisotropy energy of particles having the same volume is proportional to the square of the magnetic field. High magnetic fields would be able to orient materials still having a very weak anisotropy near the melting point T,. Such processing works perfectly under the

109

condition that the overheating rate remains smaller than a critical value. A too large overheating destroys the particle alignment and produces a liquid undercooling in some alloys. Some magnetic properties must be used to determine the alignment degree after processing. The susceptibility anisotropy near T = T, must be measured in order to calculate the particle size aligned in a magnetic field. These phenomena are described by assuming that nuclei survive above T, and are aligned in a magnetic field. They disappear above a critical overheating temperature. These observations are in contradiction with the classical theory of crystal growth nucleation as it does not predict that particles can exist above T,. An electrostatic contribution related to the possible difference between the chemical potentials or the Fermi levels of a solid and its melt could be added to the particle Gibbs free energy in order to stabilize particles in the solid state above T,. A new critical radius governs the nucleus size above T, up to a critical temperature. The orders of magnitude of the work function differences and of the contact potential are sufficient to create an electrostatic term sufficiently high in the particle Gibbs free energy to stabilize nuclei above T,. Only one electron charge per (51 A)’ or (28 A)’ is needed in the example of a SmzCo17 type alloy. The critical radius for nucleation of the crystal growth below T, tends to the classical value when the undercoooling rate increases. Contradictions with the classical theory appear around and above T, when the electrostatic term is dominating the Gibbs free energy. The approximations used to evaluate the electrostatic interaction have knocked the rough edges off the problem of overheated nuclei. Precise calculations are now needed to determine the true effects of electron correlations if they do exist. Measurements of the contact potential of the work functions could allow direct determination of the charge number when the surface energy ysl is known. The critical radius and the critical overheating can be determined by magnetic processing. The electrostatic contribution to the free energy could be deduced not only from the magnetic critical radius but also from the critical overheating when it is acting as well as the reduction coefficient p due to electron correlations. Magnetic processing is also very efficient in solids that behave differently from metals and have an anisotropic structure, such as fibres and crystals suspended in a liquid or crystallised from a solution, crystalline polymers undergoing crystallization from melts, and organic materials. The orientation appears sometimes to be possible for large overheating above the melting temperature [13-151. This type of material could have a dielectric constant 5

110

larger than 1. In this case, Q, must be replaced by ~ 4This . change decreases the electrostatic interaction. The charge number will be increased in such materials. In addition, the Gibbs free energy must take the differences between the specific heats of the solid and of the liquid into account. Other considerations of the double-layer must be used. Acknowledgements Discussions with E. Beaugnon and L. Porcar have helped the author to clarify this paper. References 1. De Rango, P., Lees, M., Lejay, P., Sulpice, A., Tournier, R., Ingold, M., Germi, P. and Pernet, M., Texturing of magnetic materials at high temperature by solidification in a magnetic field, Nature, 349, (1991) p. 770. 2. Kelton, F., Crystal nucleation in liquids and glasses, Solid. State Phys. 45 (1991) pp. 75-177; Vinet, B., Magnusson, L., Fredriksson, H., Desr6, P.J., Correlations between surface and interface energies with respect to crystal nucleation. Journal of Colloids and Interface Science, 255 (2002), p. 363. 3. Tournier, R.F., Beaugnon, E. and Pavard, S., Texturing by solidification in a magnetic field, The 4th Pacific Rim Int. Conf. on Advanced Materials and Processing (PRICM4) Ed. S. Hanada, Z. Zhong, S., Nam, W. and Wright, R.N., The Japan Institute of Metals, (2001) pp. 285-288. 4. Legrand, B., PhD thesis, Texturation par solidification sous champ magnCtique d’alliages samarium-cobalt. Application i 1’Claboration d’aimants massifs, Universit6 Joseph Fourier, Grenoble, (1996) unpublished. 5. Toumier, R., Beaugnon, E., Beille, J., Bonino, O., Bourgault, D., Chaud, X., Courtois, P., Legrand, B., Michaud, B., Noudem, J., Pemer de la Biithie, R., Pavard, S., De Rango, P., Sulpice, A., Rakotoarison, S., Soubeyroux, J.L. and Villard, C., The magnetic field as a tool for the transformation of the matter, Symposium on New Magneto-Science, Ed. Japan Science and Technology Corporation, 1-1-56 Shibashimo, Kawaguchi, Saitama, 333-0848 Japan, (1998) pp. 58-71. 6. Lees, M.R., Bourgault, D., De Rango, P., Lejay, P., Sulpice, A. and Tournier, R., A study of the use of a magnetic field to control the microstructure of the high-temperature superconducting oxide, Phil. Mag. B, 65 (1992) pp. 13951404. 7. Noudem, J.G., Beille, J., Bourgault, D., Chateigner, D. and Toumier, R., Bulk textured Bi-Pb-Sr-Ca-Cu-0 (2223) ceramics by solidification in a magnetic field, Physica C 264, (1996) pp. 325-330.

111

8. Tournier, R. Pavard, S., Bourgault, D. and Villard, C., Bulk Bi 2212 texturing by solidification in a magnetic field and hot forging, Int. Symp. On Supercond. XTI (ISS 99 Proceedings) Morioka, Japan, (1999) pp. 527-529. 9. Pavard, S., Villard, C., Bourgault, D. and Toumier, R., Effect of adding MgO to bulk Bi2212 melt textured in a high magnetic field, Supercond. Sci. Technol. 11(1998) pp. 1359-1366. lO.Bruzek, C.-E., Lallouet, N., Flahaut, E., Bourgault, D., Saugrain, J.M., Allais, A., Arsa, S., Bock, J., Ehrenberg, J., Wesolowski, D.E. and Rikel, M.O., High-performance Bi22 12/Ag tape produced at Nexans, Eucas 2003 Proceedings, to appear. ll.legrand, B.A., Chateigner, D., Pemer de la Bathie, R. and Tournier, R.F., Orientation by solidification in a magnetic field, a new process to texture SmCo compounds used as permanent magnets, J. Mugn. Mugn. Muter. 173 (1997) pp. 20-28. 12.Messina, R., Holm, C. and Kremer, K., Strong interactions in spherical colloidal systems, Phys. Rev. E, 64 (2001) pp. 021405-1-021405-14. 13.Kawai, T. and Kimura, T., Magnetic orientation of isotactic polypropylene, Polymer 41 (2000) pp. 155-159. 14.Kimura, T., Kawai, T. and Sakamoto, Y., Magnetic orientation of poly(ethy1ene terephthalate), Polymer 41 (2000) pp. 809-812. EFujiwara, M., Tokunaga, R., Chidiwa, T. and Tanimoto, Y., J. Phys. Chem. B 102 (1998) p. 3417.


Chemical and Physical Processes


REFRACTIVE INDICES OF WATER AND AQUEOUS ELECTROLYTE SOLUTIONS UNDER HIGH MAGNETIC FIELDS H. HOSODA, H. MORI, N. SOGOSHI AND S. NAKABAYASHI Department of Chemistry, Faculty of Science, Saitama University, Shimo-Ohkubo255, Sakura-Ku, Saitamu, Japan The refractive index of water was measured at atmospheric pressure under magnetic fields up to 10 T and found to increase by -0.1 % with increasing magnetic field strength. In contrast, the refractive index of saturated aqueous electrolyte solutions decreased under increasing magnetic field. No change of the optical property of n-hexane caused by a magnetic field was found. A possible explanation is that the lifetime of the hydrogen bond is prolonged due to the electron delocalization of a water dimer under a magnetic field.

1. Introduction Water is the ubiquitous liquid on earth, and indispensable to life and the environment. Water has many peculiar properties, including a large heat of vaporization, high boiling and melting temperatures, and high solubility for charged and polar molecules [l]. The distinctive features of liquid water are mainly due to the 3-D hydrogen bonding network. Recent works on water have been extended to its dynamical structure studied by ultra-fast laser techniques [2,3], theoretical studies [4-61, and the relaxation dynamics of the interfacial water near the protein surface based on neutron scattering [7]. Other recent topics are collected in the literature [S]. While a wealth of studies on water and solutions by light absorption and scattering experiments have been reported, here we concentrate on the refractive index of water. Recently, several convenient commercial sensing devices based on the surface plasmon resonance (SPR) have been developed [9-121 which can determine the refractive indices of liquid samples with very high sensitivity (Adnc Homola et al. [9] review the recent development of the SPR sensor and its application especially to chemical and biological sensors. It is also promising and important for the applications to study the basic properties of water, itself, using this sensitive device. In this paper, the effect of a high magnetic field on the refractive indices of water and several aqueous electrolyte solutions is studied in order to obtain insight into the static structure of water. 115

116

2. Experimental The refractive index (n) of water was measured by two methods, namely SPR and PSD, as shown in Figure 1 (not to scale), as a function of the magnetic field. All the measurements were carried out at ambient pressure and temperature of 25.0 "C stabilized within +1 "C. The refractive indices of aqueous electrolyte solutions were measured by the SPR method. Figure l(a) shows a setup using an SPR sensor (SpreetaTM PTSPRlA170100, Texas Instruments Inc.) which is based on the resonance between the evanescent wave and the surface plasmon [lo-121. The sensor consists of a light emitting diode (AIGaAs, 840nm), a molded epoxy waveguide, a sensing area (50nm gold film), and a photodiode array. The resolution of the refractive index is 5 ~ 1 0 . ~ Due . to the rapid damping of the evanescent wave, this device is sensitive to the thin layer of the analyte material at the vicinity of the interface. The typical thickness for a water-gold interface is 400 nm [ 121. A recommended calibration procedure was performed to adjust the refractive index of ultra-pure water at 25.0 "C to 1.333000. The literature value of the refractive index of water at 840 nm and 25.0 "C is 1.32796 [ 131 and thus

Figure 1. Two experimental setups for measuring the refractive indices of water and aqueous solutions (not to scale). Both measurements were made under magnetic fields up to 10 T using a superconducting magnet. (a) A schematic diagram using a commercial surfact plasmon resonance (SPR) sensor. The sensor is comprised of a light emitting diode, a molded epoxy waveguide, a sensing area, and a photodiode array. (b) A setup using a He-Ne laser, a quartz cell, and a position sensitive detector (PSD).

117

the refractive index measured by the SPR method in this study contains a slight constant difference from the literature. T h i s difference, however, should not affect our considerations and conclusions. The sensor was placed at the magnetic center of a superconducting magnet (JMT-1OT150, Japan Super Conductor Technology Inc.), which generates a magnetic field up to 10 T. The sensor was positioned so that the gold film is perpendicular to the magnetic field. Another setup using a He-Ne laser (633 nm) and a position sensitive detector (PSD) obtained from Hamamatsu Photonics K.K. was used to measure the refractive index of bulk water and elucidate any surface effects for the SPR measurement. A quartz sample cell with the outer dimensions of 12.5 x 12.5 x 45.0 mm3 and a quartz wall thickness of 1.25 mm was placed at the magnetic center with a tilting angle (8)between the cell and the laser beam. The laser beam passing through the cell is deflected, and the displacement of the optical path is detected by the PSD. The refractive index of water is obtained from the displacement, 8, and the refractive index of water at 633 nm is 1.33158 [13]. The small difference of the optical path due to the quartz cell, itself, is also accounted for. By varying the tilting angle B (70, 73, 75, and 77 degrees), the consistency of the refractive index obtained was checked and the experimental error (Anln) was estimated to be < 1 . 5 ~ 1 0 - ~ . Ultra-pure water was prepared by a commercial water purification system (Direct-Q 5, Millipore Corp., 18.2 MQ cm). NaCl and NiC12 (Wako Pure Chemical Industries, Ltd. GR grade) were used without further purification. The refractive index of n-hexane (Wako Pure Chemical Industries, Ltd. 99.5%) was also measured without further purification.

3. Results and Discussion The dependence of the refractive index of pure water on the magnetic flux density ( B ) is shown in Figure 2. The refractive indices measured by the SPR setup (triangles, nspR> and the PSD setup (circles, n p s ~ )show increases by 1 . 8 ~ 1 0(0.14 ~ %) and 1.3~10”(0.09 %) at 10 T, respectively, from those in the absence of the magnetic field. The increase of ~ S P Ris slightly larger than that of nPsD.The origin of this discrepancy is unclear at this time, even though it could be attributed to the magnetic effect on water at the vicinity of the interface, or the difference of the dielectric constant between the wavelengths of 840 nm (SPR) and 633 nm (PSD). The temperature derivative of the refractive index of deg at room temperature in the visible region around 600 nm water is l x

-

-’

118 [ 131. In the current study, the increase of ~ S P Rand n p s ~at 10 T exceeds the temperature fluctuation effects. Thus, the refractive indices of both of the vicinity of the interface and the bulk are increased by the magnetic field effect.

1.3350 I

1

1.3330

1.3320

0

2

4

6

8

ID

B/T Figure 2. Refractive index of pure water plotted as a function of the magnetic field obtained by the position sensitive detector (PSD,circles) setup and the surface plasmon sensor (SPR, triangles) setup.

A possible explanation for the increase of the refractive index of water is that the hydrogen bond is stabilized under the magnetic field. From a classical electromagnetic point of view, diamagnetism is explained by the anti-parallel magnetization of a molecule to the external magnetic field by electromagnetic induction. It is well known that Pauling explained the diamagnetism of aromatic hydrocarbons using a molecular-size ring current model [14]. Since diamagnetism of a molecule depends on the extent of electron distribution, the electron delocalization of hydrogen-bonded molecules should increase its diamagnetism. Therefore, the hydrogen bond should become more stable in a magnetic field. Iwasaka et. a1.[15] found that the frequency of the higher harmonic vibrations of water shifts toward the longer wavelength under 14 T. In comparison to the spectra of water at higher pressures, they suggested the enhancement of hydrogen bonds under high magnetic fields. The enhancement of hydrogen bond strength should lead to the change of the electronic absorption, which affects the refractive index in the near infrared region. According to the electronic spectra of ice in the vacuum ultraviolet region [16],

119

the increase in the absorption of the first electronic excited state of the crystalline hexagonal ice from the one of amorphous ice was observed. The increase in the absorption due to the formation of hydrogen bonds should cause the increase of the refractive index via the Kramers-Kronig relation. The present observation strongly indicates that the lifetime of hydrogen bonds is prolonged. 1.3350 I

'I water x

al

hexane

T

Figure 3. Refractive index of n-hexane plotted against the magnetic field (dots) measured by the SPR method. The refractive index of water measured by the SPR method is reproduced for clarity (triangles).

In Figure 3, the refractive index of n-hexane (nhexme) is plotted against B (dots) measured by the SPR method. The result of pure water (nwakr)by the SPR method is reproduced for clarity (triangles). While nwater gradually increases with B, nhexme does not change up to 10 T. This difference supports the hypothesis that hydrogen bonds are stabilized in magnetic fields. Moreover, the refractive index of ethanol in a magnetic field (unpublished data) was measured and shows little dependence on B. This may mean that although ethanol forms hydrogen bonds, they are not stabilized significantly in a magnetic field since the number of hydrogen bonds per molecule for ethanol is less than that for water.

120

1.3670

1.3660

n 0

1.34201 ?.

1.3350

1.3340

-

I

I 1

13330 0

2

4

6

8

1

0

8I T Figure 4. Refractive indices of aqueous solutions of 2.5 M NiC12 (filled squares), 5.0 M NaCl (unfilled squares), 0.40 M NiC12 (filled circles), and 0.50 M NaCl (unfilled circles) plotted against the magnetic field, measured by the SPR method. The refractive index of pure water measured by the SPR method is again reproduced for clarity (triangles)

Figure 4 shows the dependences of the refractive indices of aqueous electrolyte solutions on B as measured by the SPR method. Each mark represents the following electrolyte solutions: NaCl solutions at concentrations of 5.0M (open squares) and 0.50M (open circles), and NiC12 solutions at concentrations of 2.5 M (closed squares) and 0.40 M (closed circles). The refractive index of pure water (nWater) is again reproduced for clarity (triangles). The figure indicates that (1) the refractive indices of electrolyte solutions increase with the increase of its concentration in the absence of a magnetic field,

121 and ( 2 ) the slope of the n-B curves is positive at lower concentrations and negative at higher ones. Without the magnetic field, the variation of the refractive indices of electrolyte solutions from that of pure water increases in sequence of 0.50 M NaCl (0.009) < 0.40 M NiC12 (0.011) < 5.0 M NaCl (0.034) < 2.5 M NiC12 (0.035). It is intriguing that the increase of the refractive index is dominated by [Cl-] at higher concentrations. The color of the aqueous NiClz solution is green due to the formation of [Ni(H20),#’ [17] and therefore its refractive index is expected to be dependent on [Ni”] due to the Kramers-Kronig relations. The results, however, show the refractive index depends more strongly on [Cl-] than on [Ni”] and “a’] at higher concentrations. The solutions of 2.5 M NiC12 and 5.0M NaCl are near the saturation; NaCVaq; 5.6M, NiClz/aq; 6.0M. At this concentration, the solution is dominated by ionic atmospheres and the extent of ion-pair formation becomes large [18]. The refractive index at 840 nm seems to correlate to Cl-, possibly due to the change or the appearance of the electronic states of C1- perturbed by ion-pair formation or complexation with cations. It should be noted that, in the case of a concentrated NiClz solution, the formation of [Ni(H20)&1]’ is reported [ 191. The changes of the refractive indices of the solutions at 1 0 T from the solutions without the magnetic field are -0.0006 (2.5 M NiC12), -0.0010 (5.0 M NaCl), 0.0016 (0.40 M NiClZ), and 0.0014 (0.50 M NaCl), respectively, as is shown in Figure4. The slope of the n-B curves of these electrolyte solutions also seems to be dependent on [CT] rather than “a’] or [Ni”] at higher concentrations. Although the paramagnetism of Ni2+ due to unpaired 3d electrons is important, the refractive index under 10 T shows little difference between the paramagnetic species (Ni”) and the diamagnetic species (Na’). Therefore, the n-B curves obtained should originate from Cl-. Although the complex magneto-optical behavior of electrolyte solutions cannot be explained easily, we can presume that two species are responsible for the different n-B curves, e.g., the one dominant at higher concentrations, and the other at lower concentrations. Possible candidates for the higher concentrations are a hydrate complex or ion-pair containing more than one chloride ions, since the absolute value of n and the negative dependence of n on B is dominated by [Cl-] at higher concentrations for both of the NiClz and NaCl solutions. Also, there are no distinct differences between the two cations. The origin of the n-B curve at lower concentrations should be water, itself. According to the threezone model by Frank and Wen [20] ions hydrated by water (A zone) are further surrounded by a weakly interacting “destructured” region (B zone), where water

122

molecules are neither oriented to the core ion nor hydrogen-bonded to each other. Around them is bulk water (C zone) in which water molecules are structured by hydrogen bonds. Thus, the n-B curve at lower concentrations is considered to be the one for bulk water (C zone) superimposed by another for the ionic species (A or B zone). For the saturated solutions, the n-B curve should be purely dominated by the hydrated ions or the ion-pairs. 1.370

I

I

x 1.360 0

'0 C

-

=g 1.350

s

1.3450

rt

2 1.340 rl

Y 1.330

I

0

1. 1

1

3

I

2

a

0.50 I 3

, 4

,

, 1. 5

Concentration of NaCl I M Figure 5. Refractive index of the aqueous NaCl solution measured by the SPR method plotted against the concentration the absence (squares) and the presence (circles) of the 10 T magnetic field. The magnified view ranging from 0.5 M - 10 M is shown in the inset.

In Figure 5, the refractive index of aqueous NaCl ( ~ N ~ c Isolution ) is plotted ~ ) (circles) and without (squares) the 10 T against its concentrations ( c N ~ c with magnetic field, as measured by the SPR method. The increases rapidly up to -0.5 M, above which it increases linearly in both of the absence and the CI at OT and at 1 0 T it presence of the magnetic field. The ~ N ~ C I - C N ~curves crosses at 0.75 M (see the inset of Figure 5). The figure clearly shows the occurrence of the trade-off between the two sources of the different magnetooptical behavior at concentrations of 0.5-0.75 M. 4.

Conclusion

The refractive indices of water and aqueous electrolyte solutions were measured. The refractive index of pure water under 10 T increases by -0.1 % from the measurement done without the magnetic field. It is proposed that the hydrogen bond of water is stabilized under the magnetic field. Therefore, the optical

123

properties of ultraviolet absorption and refractive index should increase. Aqueous electrolyte solutions at higher concentrations show the decrease of the refractive index under 10 T, which is possibly explained by the formation of a hydrate complex or an ion-pair containing more than one chloride ion.

References 1. Eisenberg, D., Kauzmann, W. The structure and the properties of Water; Oxford University Press: Oxford, 1969., Japanese translation, MisuzuShobou: Tokyo, 1983. 2. Yeremenko, S., Pshenichnikov, M.S., Wiersma, D.A., Chem. Phys. Lett. 369 (2003), pp. 107-113. 3. Winkler, K., Lindner, J., Biirsing, H., Vohringer, P., J. Chem. Phys. 113 (2000), pp. 4674-4682. 4. Bour, P., Chem Phys. Lett. 365 (2002), pp. 82-88. 5. in het Panhuis, M., Popelier, P.L.A., Munn, R.W., Angytin, J.G., J. Chem. Phys. 114 (2001), pp. 7951-7961. 6. Tschumper, G.S., Leininger, M.L., Hoffman, B.C., Valeev, E.F., Schaefer, H.F., III, Quack, M., J. Chem. Phys. 116 (2002), pp. 690-701. 7. Dellerue, S., Bellissent-Funel, M. -C., Chem. Phys. 258 (2000), pp. 315-325. 8. The special issue for water. Chem. Phys., 258 (2000), pp. 2-3. 9. Homola, J., Yee, S.S., Gauglitz, G., Sensors and Actuators B 54 (1999), pp. 3-15. lO.Melendez, J., Cam, R., Bartholomew, D.U., Kukanskis, K., Elkind, J., Yee, S.,Furlong, C., Woodbury, R., Sensors and Actuators B 35-36 (1996), pp. 2 12-216. ll.MelCndez, J., Carr, R., Bartholomew, D., Taneja, H., Yee, S., Jung, C., Furlong, C., Sensors and Actuators B 38-39 (1997), pp. 375-379. 12.Elkind, J.L., Stimpson, D.I., Strong, A.A., Bartholomew, D.U., Melendez, J.L., Sensors andActuators B 54 (1999), pg. 182-190. 13.Handbook of Chemistry and Physics 83 ed. by Lide, D.R., CRC Press: Florida, 2002. 14.Pauling,L., J. Chem. Phys 4 (1936), pp. 673-679. 15.Iwasaka, M., Ueno, S., J. Appl. Phys. 83 (1998), pp. 6459-6461. 16.Kobayashi, K., J. Phys. Chem. 87 (1983), pp, 4317-4321. 17.Lever, A.B.P., Inorganic Electronic Spectroscopy; Elsevier: Amsterdam, 1968; p 334-335, and references cited therein. 18.Harned, H.S., Owen, B.B., ed. The Physical Chemistry of Electrolytic Solutions, Reinhold New York, 1950, pp. 42-45 and references cited therein. 19.Magini, M., Paschina, G., Piccaluga, G., J. Chem. Phys. 76 (1982), pp. 11161121. 20.Tanaka, N., Ohtani, H., Tamamushi, R., ed. Ions and Molecules in Solution; Elsevier: Amsterdam, 1983; W.-Y. Wen, “Hydration of some solution in Aqueous Solutions” pp. 45-59.

SYNTHESIS OF CARBON MATERIALS BY THE IMPOSITION OF A HIGH MAGNETIC FIELD M.-G SUNG', K. SASSA', A. GEDANKEN3,K. IWAI', S. ASAI' 'Venture Business Laboratory, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan 'Department of Materials Processing Engineering, Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan 'Department of Chemistry, Bar-llan University, Ramat-Can, 52900, Israel Carbon materials produced from organic materials as a precursor are generally subjected to three heat treatment processes of stabilization and carbonization followed by graphitization. A high magnetic field was imposed in the thermal decomposition reactions of carbon materials. Thermal decomposition reactions such as solid-phase, liquid-phase and gas-phase reactions have been used. It is found that the magnetic field imposed along with the heat treatment controls the magnetic shape and increases the tensile strength of carbon materials formed by thermal decomposition reactions. The mechanism of magnetic shape control and increase of the tensile strength due to the imposition of a high magnetic field is discussed based on an intermolecular cross-linking reaction model in which the radical pair theory is modified by accounting for the magnetic field.

1. Introduction Recently, high magnetic fields have become available rather easily due to the development of superconducting magnet technologies. Research has been conducted in various disciplines, as well as in materials science, in which a high magnetic field has been applied. The results show that a high magnetic field may change basic physical or chemical properties not only in ferromagnetic substances but also in paramagnetic and diamagnetic materials ignored hitherto.

124

125

Carbon is considered to be an advanced material and possesses several distinctive properties that cannot be obtained in metals, ceramics and polymers. It has been utilized in a wide variety of industries such as construction, sports, medicine and more. Carbon material has mainly been produced through thermal decomposition reactions from organic materials by controlling parameters such as temperature, atmosphere, and pressure. A number of researchers have begun to study Fig.1 Schematic view of experimentalapparatus and fibers setting configuration in carbonization materials processing by using a process. high magnetic field, which has been mainly applied on materials in a liquid state. However, the application of a high magnetic field to processing carbon fibers, in which many internal chemical reactions occur in the solid or liquid state, has barely been studied. In this presentation, we introduce some experimental results obtained in carbon materials by applying a high magnetic field as follows: 1. While the demand for high strength materials is increasing due to the increase of human activities, carbon fibers having a specific tensile strength and high temperature characteristics are observed with great interest by the aerospace industry and others. Here we report the discovery of a new technique, applying a high magnetic field to increase the tensile strength of carbon fibers. This technique results in decreasing defects in the carbon fibers. The application of a high magnetic field could lead to a regular structure with low defect density by suppressing rapid chemical reactions that induce defects. 2. Mesophase spherules produced from a pitch and a mesityoene have been used as the precursor of mesocarbon microbeeds (MCMB), which are adopted for use in a lithium ion secondary battery. In order to improve the efficiency of the battery, it is important to control the crystal structure of mesophase spherules. The application of a high magnetic field in the carbonization of pitch strongly promotes the size of mesophase spherules and their size uniformity.

126

2. Carbon Fibers

2.1. Experiment A schematic view of the experimental apparatus and its setting for the configuration of fibers are shown in Figure 1. The fibers were installed so they could be rolled from a lower to an upper robin and a tension of 8 N was applied on a fiber tow of 12,000 pieces. A motor was used to wind fibers to the upper section and a brake was set in the lower section. The atmosphere was controlled by flowing Ar gas (purity of 99.9999%) from the upper to the lower 2 c d v section in the reaction & tube. The heat treatment .$ .-x rn temperature was kept at 1773 K and a magnetic field of 0 or 5 T was applied, Two types Of 1000 1400 1800 ld00 ' 1400 fibers were used: Raman shift (cm') Rau~anShift (CM') stabilized fibers (A A series B series series) and high Figure 2. First-order Raman spectra for A and B series. temperature heat-treated A: Oxidized fibers B: Heat treated fibers in 673-773 K after oxidization fibers (B series) made from the stabilized fibers by heat- treating at 673-773 K to reduce any structural defects. Fifty fiber pieces were randomly selected as tensile test specimens in each series. In order to obtain a Raman spectrum, a carbon fiber was mounted on a glass slide and its axis was positioned on a table to be parallel to the polarization direction of an incident beam. The laser power was approximately 1.5 mW with a wavelength of 514.5 nm. The exposure time required for spectrum acquisition was 5 sec. Y

4

J

'

lhl

2.2. Results and Discussion Figures 2 and 3 show an example of Raman spectra and the intensity ratio R (=I&Il580 and Idk), respectively. In general, the G-line near 1580 cm-' appears due to the vibration of a graphite cell in high graphitization materials such as HOPG (Highly oriented pyrolytic graphite) and carbon fibers [l-41, and the D-line seen near 1355 cm-' indicates a slightly graphitized state that is going

127

to disappear in higher graphitization temperatures [2-91. The D'-line appeared near 1620 cm-', which is detected as a shoulder of the G-line. It is attributed to a small crystallite size and sstructural&'Order [5-81.

. 0

&

4.5

r,

A series

B series

Figure 3. Average tensile strength of A and B series. A: Oxidized fibers B: Heat treated fibers in 673-773 K after oxidization

In order to simulate the distribution of the Raman spectrum, the original data were synthesized to produce three different curves. Figure 2 shows the effect of the magnetic field on the intensity of the D'-line. The intensity of the D'-line of the carbon fibers processed under a magnetic field of 5 T is suppressed compared to that of the fibers processed without a magnetic field in both the A and B series. The intensity ratio R, defined as the ratio of the area under the D-line curve to the area under the G-line curve, was used for the structural analysis of carbon [lo]. The average tensile strength of fibers obtained in the carbonization process $ 2.7 of the A and B series is .shown in Figure 3. For the .-2 2.5 A series, the strength of 3 2.3 fibers processed under a 2.1 magnetic field of 5 T is A series B series 4320 MPa, while that of the fibers processed without a Figure 4. Crystallite sizeLa of A and B series. magnetic field is 3990 m a . A: Oxidized fibers B: Heat treated fibers 673-773 K after oxidization For the B series, the strength of the fibers produced under a magnetic field of 5 T is 3330 MPa and that without a magnetic field is 4110 MPa. Comparing the tensile strengths of the A and B series processed without a magnetic field, the B series is stronger than A series by 120 MPa. That is, the higher strength in the B series seems to be attributed to the heat treatment of 673-773 K during the first carbonization stage. The tensile strength of the A series processed under a magnetic field of 5 T is v)

4

":

128

higher by about 210 MPa than that of the B series 4200treated without a magnetic field. This indicates that the 3800 application of a magnetic 8 field in the A series r= 3400 produced from stabilized fibers plays a role in reducing the structural t+ 3000 defects of carbon fibers 2.2 2.4 2.6 2.8 3.0 compared to the B series. Crystallite sizednm Thus, the application of a Figure 5. Relationship between crystallite size La and tensile magnetic field during the strength. first carbonization stage is more likely to reduce defects in fibers. The crystallite size La, which is closely related to the intensity ratio R [l-3,6], can be calculated by using the empirical equation, La =4.35/R (nm) [11,12]. Figure 4 shows the calculated crystallite size for the A and B series processed with and without a magnetic field. The values of L, for fibers processed under a magnetic field of 5 T are higher than those processed without a magnetic field in both A and B. The relationship between the crystallite size L, and the tensile strength of fibers is shown in Figure 5 The tensile strength increases with increasing La in the A series, but is inversely proportional to L, in the B series. As far as A series is -c/ \ c > ,c/ N \ c 9 the N /--

$ g

5

8

concerned, the direct CH I relationship between ' \:/'? the tensile strength and L, agrees with the result in a previous report [13], but the result obtained in the case of B series requires a further study. us Now, let introduce a radical pair mechanism to explain

i

OH

I

i ;

I

I

, CH, + H,O

,CH,.

CH

4

4

--....t........' .CH

A

1

,cHc'H'

.CH

I

I bN/"" (a)

c\

/CH,

/ C b N @)

.

CH

I

'

\c>

\c/

I ,C",CH,

+H@

C

I

C 'CH.H'CH.

I

I

/C\

% ,

N

(C)

Fig.6Schematic view on intermolecular cross-linking mechanism in carbonizationprocess of stabilized carbon fibers.

129

the increase in the crystallite size L, caused by the application of a magnetic field. An ideal intermolecular cross-linking reaction mechanism in a dehydration reaction in the carbonization process is proposed, as shown in Figure 6, in which (a) indicates the representative structure of oxidized fibers with ROH and R H molecules. Due to heat treatment, they are converted into free radicals R and R by passing through the homolysis as shown in (b). The chains of free radicals are cross-linked by a collision process as shown in (c). When these free radicals collide randomly, the ratio of generated singlet radical pairs to triplet radical pairs is 1:3 [14,15]. Then, an intersystem crossing (ISC) occurs through an electron-nuclear hyperfine mechanism [ 151. In the case with no magnetic field, the ISC of triplet radical pairs takes precedence over an escape through the loop (B) as shown in Figure 7. This mechanism rapidly increases the ratio of the singlet radical pairs, which quickly takes the place of a random recombination reaction to terminate polymerization. This mechanism leads to increased defects in a graphite crystal plane and to a reduced L.Alternately, when a high magnetic field is applied, the electron-nuclear hyperfine mechanism does not work so that the loop (A) of the ISC indicated in Figure 7 is suppressed. The high magnetic field increases the ratio of the triplet radicals compared with the case of no magnetic field. The triplet radical pairs hardly recombine due to the difference of their spin-states so that they escape from the loop (B) and become free radicals as shown in Figure 7, and the ratio of escaped radicals increases compared to that of the case with no magnetic field. Thus, the application of a high magnetic field reduces the recombination reaction rate due to the decrease of the ratio of singlet to triplet radicals. It plays a role in forming better regular graphite crystal planes by propagating a polymerization. This mechanism implies that the imposition of a magnetic field is favorable for increasing La and explains the experimental results given in Figure 4.

136.

ROH

Heat

(Re .OH)

(R'*.H)

Heat

C- R ' H

-Heat Propagation of polymerization

escape: polymer radical

1

[

Q-.R),

.R' ), y R * . .R' )

I Recombination(R-R, R-R' , R'-R' ) (Termination of polymerization)

3

[

'(R. OR), 3(R* .R'),'(R'*.R')

]

(A)

Intersystem crossing

Figure 7. Intermolecular cross-linking mechanism in carbonization process of stabilized carboi fibers.

3. Mesophase Pitch

3.1. Experiment Coal tar pitch with a molecular weight less than 300, a softening point of 353-358 K, and a fixed carbon ratio of 5557% was thermally decomposed under a magnetic field. A schematic view of the experimental apparatus and the setting configuration of samples of the coal tar pitch are shown in Figure 8. The temperature of the samples Aroutleti II was raised at the rate of Water outlet 0.05 Wsec from room temperature to 703 K and kept for 7.2 ks at 703 K in a magnetic field of either 0 T or 8 T. The atmosphere in a reaction tube was protected by flowing Ar gas (purity of 99.9999%), from the lower to the upper section, to prevent Thermocouple oxidation of the samples. After 9 ml quinoline Fig.8 Schematic view of experimental apparatus and sample setting configuration in carbonization process of pitch. was added to the

131

heat-treated sample of 0.5 g, it was kept for one month to separate the mesophase spherules from pitch matrix at room temperature. Then the treated sample solution was filtered using a 0.1 pm nitrocellulose filter. The separated mesophase spherules were observed with a scanning electron microscope (SEM), and the specimen of 1 mg added KBr of 150 mg was analyzed using Fourier transform infrared spectroscopy (FT-IR: PERKIN ELMER PARAGON1000 spectrophotometer).

3.2. Results and Discussion Microstructures of the OT 8T separated mesophase spherules are shown in Figure 9. In the case of no magnetic field, the diameters of the mesophase spherules are 1-10 pm as shown in Figure 9(a). In the case of a 10 T magnetic field, the diameters of the mesophase spherules are Figure 9. SEM micrographs of each sample kept for 7.2 ks at 703 K with and without magnetic field. around 10 pm as shown in Figure 9(b). Thus, the application of the high magnetic field increases the diameter of the mesophase spherules. By comparing Figure 9(a) with (b), it can be understood that the application of the high magnetic field increases the uniformity in diameter and easily separates the mesophase spherules from the pitch matrix when one observes their dispersed appearance as a set of clumps containing spherules and matrix in the case of no magnetic field. FT-IR spectra for the raw material and the spherules produced with and without the magnetic field are shown in Figure 10. According to the deflection of the hydrogen atoms with respect to the plane defined by the six carbon atoms in the aromatic ring, the various possible carbon-hydrogen vibrations are differentiated into three types: one stretching and two bending vibrations. The C-H stretching vibrations defined in this plane appear between 3100 and 3000 cm-'. In this region, all vinyl-bound hydrogen atoms absorb weak intensity. Hence, they are not specific to aromatics. Because of hydrogen bond absorptions, these bands are often hard to find [ 161.The carbon-hydrogen bending vibrations

132

occur as in-plane bending and out-of plane bending vibrations, which usually exist in the fingerprint region of 1300-1000cm-' and 900-700 cm-l, respectively. The in-plane bending vibrations are much less consistent than the out-of plane bending vibrations. The in-plane bending vibrational frequency depends on the number of C-H groups and the substituent groups. Out-of-plane bending is much more consistent and intense in the infrared region. Out-of-plane bending is not particularly influenced by substituents, but the frequency for the bending vibrations depends on the number of adjacent hydrogens [17]. Hence, the sets of strong absorptions relating to the 900-700 cm-' region, which result from C-H out-of-plane bending vibration can often determine the degree of substitution of aromatic rings [18]. For spectra of the raw material in Figure 10, absorptions in the vicinity of 1600 cm-', 1440 cm-' and 900-700 cm-' indicate that the raw material is mainly comprised of aromatic rings with elements such as carbon, hydrogen and substituents. From spectra in the 900-700 cm-' region, we could presume that the raw material contains various complicated substitutive structures with one, two, three, four and five neighboring hydrogen atoms. Especially, 750 cm-' of all 2000 1600 1200 800 400 spectra in the 900-700 cm-' Wavenumbers, Y 1cm-I region is the strongest and with Figure 10. IT-IR spectra of pitch corresponds 1 ,Zdisubstituted, which has four neighboring hydrogen atoms. The IR spectra of OT-S and 8T-S obtained by separating micro-spherules from matrix also have a difference, which appears in 750 cm-'. However, on the contrary, the band of OT-S in 750 cm-' is smaller than that of 8T-S as shown in Figure 10. That is, from the degree of the substitute of C-H bond, FT-IR results of OT-S and 8T-S imply that a chemical reaction generating micro-spherules from matrix by heat-treating raw material is dramatically changed by the applicaton of a high magnetic field. These results might qualitatively explain the increase in size of mesophase spherules and their size uniformity by introducing a radical pair mechanism in a magnetic field.

3

133

4. Spherules by Using Mesitylene

pipe at the center of the furnace under a static magnetic field of 10 T, generated using a helium-free superconducting magnet. The S wagelock was placed at the point of

OT

1OT

v

I

-

4.2. Results

The representative SEM image of carbon materials heat treated under no magnetic field shows the spherical morphology of the carbon spherules, with an average diameter of 2.5 pm as shown Figure ll(a). A perfect spherical shape and smooth surface can be seen at the SEM picture. The as-prepared carbon spherules are solid and not hollow, confirmed by the TEM image of the horizontal cross-section of carbon spherules. The SEM image of carbon materials heat treated under a magnetic field (Figure 1 l(b)) shows sausages with

134

diameters ranging from 2 pm to 4 pm and lengths between 10 pm and 32 pm. In addition to the sausages, carbon spherules (average diameter 2.5 pm) are also observed. When observed at low magnification, more than 90% of the sausages are not converted into carbon spherules under the magnetic field and only 10% are converted into 2.5 pm carbon spherules. A perfect sausage shape and smooth surface can be seen in the SEM picture.

5.

Conclusions

A high magnetic field was applied in the thermal decomposition reaction of carbon materials. The effect of the magnetic field was examined by using SEM, FT-IR and X P S . The formation mechanism of carbon materials has been theoretically studied. The results obtained in this study is summarized as follows: 1. Application of a magnetic field on carbon fibers improves the tensile strength. 2. The application of a high magnetic field in the carbonization of pitch strongly promotes the size of mesophase spherules and their size uniformity. 3. The results obtained using IT-IR and X P S could prove the change of chekcal reactions by application of a high magnetic field. 4. The formation mechanism of carbon materials has been explained using the radical pair theory in a magnetic field.

Acknowledgements We wish to thank Professor Tanimoto of Hiroshima University for some useful discussions. This research was supported by the Monbukagakushou (Ministry of Education, Culture, Sports, Science and Technology of Japan) Grant-in-Aid for Creative Scientific Research (No. 13852013).

References 1. 2. 3. 4. 5. 6. 7. 8. 9.

Nemanich, R.J., Solin, S.A., Phys. Rev. B 20 (1979), p. 392. Tuinstra, F., Koenig, J., J. Chem. Phys. 53 (1970), p. 1126. Tuinstra, F., Koenig, J., J. Compos. Muter. 4 (1970), p. 492. Nakamizo, M., Kammereck, R., Walker, P.L., Carbon 12 (1974), p. 259. Tsu, R., Gonzalez, J.H., Hernandez, C.I., Solid State Commun. 27 (1978), p. 507. Chieu, T.C., Dresselhaus, M.S., Endo, M., Phys. Rev. B 26 (1982), p. 5867. Menarch, T.P., Cooney, R.P., Johnson, R.A., Carbon 22 (1984), p. 39. Katagiri, G, Ishida, H., Ishitani, A., Carbon 26 (1988), p. 565. Fitzer, E., Gantner, E., Rozploch, F., Steiner, D., High Temp. High Press 19

135 (1987), p. 537. lO.Endo, M., Hakamada, K., Kim, C., Miyazawa, N., Kasai, T., Tanso No.183 (1998), p. 156 [in Japanese]. ll.Knight, D.S., White, W.B., J. Muter: Res. 4 (1989), p. 385. 12.Gruber, T., Zerda, T.W., Gerspacher, M., Carbon 32 (1994), p. 1377. 13.Sung, M.G, Sassa, K., Inoue, K., Ogawa, H., Doyama, M., Yamada, S., Asai, S., Tanso N0.200 (2001), p. 255 [in Japanese]. 14.Steiner, U.E., Ulrich, T., Chem. Rev. 89 (1989), p. 51. 15.Turr0, N.J., Kraeutler, B., Acc. Chem. Res. 13 (1980), p. 369. 16.Gunzler, H, Gremlich, H-U., IR Spectroscopy, Weinheim: Wiley-VCH, 2002, pp. 171-278. 17.Workman, J.J., Handbook of Organic Compounds Vol.l, San Diego: Academic Press, 2001, pp. 211-228. lS.Harwood, L.M, Claridge, T.D.W., Introduction to Organic Spectroscopy, Oxford: Oxford University Press, 1997, pp. 22-32.

HIGHLY EXCITED MOLECULES IN MAGNETIC FIELDS K.TAKAZAWA Tsukuba Magnet Laboratory, National Institute for Materials Science 3-13 Sakura. Tsukuba 305-0003Japan Two-color resonanceenhanced multiphoton ionization (REMPI) spectra via a single Zeeman sublevel in the A state of gaseous nitric oxide (NO) were measured in magnetic fields, B, ranging from 0 5 B S 10 T to observe the magnetic field effects on electronic states near the ionizationpotential (I€'). It was found that new resonance appears above the I€'for B 2 4 T. By using a semi-classical calculation, this resonance was assigned to the quasi-Landau resonance, which was observed for the first time in molecules.

1. Introduction Recent progress in magnet technology enables the use of magnetic fields, B, for molecular spectroscopy that exceed 10 T [1,2]. Relative to the binding energy of an electron in a molecule, however, such high magnetic fields still represent only a small perturbation of the orbital motion of the electron. The Zeeman shift caused by a 10 T magnetic field is on the order of meV, even for NO, which is a paramagnetic molecule. This Zeeman shift is only less than 0.1% of the binding energy of the electron in the ground state molecule. Therefore, the orbital motion of the electron is not significantly influenced by magnetic fields that can be generated in a laboratory. The Rydberg state of a molecule is an atomic-like excited state, in which one electron is excited to a high-n orbital (n: principal quantum number). Thus, similar to a hydrogen atom, the energy of the Rydberg state can be expressed as

E =IP - R / ( I I - ~ ) ~ where IP is the ionization potential, R is the Rydberg constant, and 6 is the quantum defect, respectively. Rydberg states, therefore, are composed of a series of states, which converge to IP at n=m. High-n Rydberg states in a magnetic field are of particular interest because of the low binding energy of the electron. The Coulomb force that attracts the Rydberg electron to the core (positively charged molecular ion) is proportional to r i 3 , and the energy difference between adjacent n levels is proportional to ri3. In the n=28 Rydberg state, the Coulomb force on the Rydberg electron is of the same order as the Lorenz force from a 10 T magnetic field. Therefore, a 10 T magnetic 136

137

field is not a small perturbation for the Rydberg state; the orbital motion of the electron is strongly influenced by the Lorenz force and the levels with different n are strongly mixed (n-mixing) because of the small energy gap. Two-color resonance-enhanced multi-photon ionization (REMPI) spectroscopy is useful for studying high-n Rydberg states of molecules. In this technique, the ground state (the X state) molecule is excited to a specific rotational level in the first excited state (the A state) by a first laser. A second laser is used to excite the molecule in the A state to a high-n Rydberg state. We developed a setup for two-color REMPI spectroscopy in magnetic fields and measured spectra of the high-n Rydberg state of NO for 0 5 B 5 10 T. 2.

Experiment

The setup for measuring two-color REMPI spectra is shown in Figure 1. A chamber containing a pair of electrodes inside was attached to the superconducting magnet. The NO pressure in the chamber was maintained at 80 mTorr. Two dye lasers were pumped by using split outputs from a Xe-C1 excimer laser. The outputs of the dye lasers were frequency-doubled and coaxially focused Excimer Laser

I

SHG

Focusing lens

I

Pulse generator

, + 1 5 v n

Superconducting magnet Personal computer

Boxcar integrator

Current amplilier

Figure 1. Setup for measuring two-color REMPI spectra in magnetic fields. ER: Electrode. BS: Beam splitter.

138

in the middle of the electrodes. The first laser, v l , was used to excite molecules to a specific Zeeman sublevel in the A state. The second laser, v2, was scanned in the energy region of the transition from the A state to the high-n Rydberg states. Laser v2 was polarized perpendicular to the magnetic filed to observe AM=fl transitions. 5 ps after laser excitation, a +I5 V voltage was applied to the electrode to detect ions produced from the Rydberg molecule. A 5 ps delay was used to avoid electric field effects on molecules at laser excitation. The ion signal was amplified by using a current amplifier, integrated with a boxcar integrator, and recorded with a computer.

3. Results and Discussion Figure 2 shows two-color REMPI spectra measured after exciting the -1, J=2.5, M=-1.5 (v: vibrational quantum number) Zeeman sublevel in the A2C' state by

0-

. x I-

=

2-

4.

D 0

c u)

2

68-

ia

l

'

76950

'

'

~

I

77000

~

f~

iI

* ~

~

.

I

,

I

I

I

I

I

77050 77100 77150 Wavenumber / cm-i

I

I

I

I

77200

I

L

I

I

I

I

I

77250

Figure 2. Two-color REMPI spectra of NO measured via the -1, J=2.5, M=-l.5 level in the A state for 0 S B S 10 T. The broken line at 77069 cm-' shows the ionization potential, IP,&=I).

,

139

using laser v1 for 0 S B I 10 T. Levels with M=-2.5 and M=-0.5 are observable due to the AM= & 1 selection rule. For B=O, an increase in the ion signal at 77,050 cm-' (corresponding to the IPv(v=1))were observed, and below the ionization potential the Rydberg states converging to the IP,(v=l) were also observed. With increasing magnetic field, Rydberg states show complicated level splitting, reflecting the large Zeeman shift of the high-n Rydberg states. However, the most dramatic feature of the spectra is an appearance of periodic modulations of the signal intensity above the IP for B 1 4 T. It is well known that the Schrodinger equation for a Rydberg atom in a magnetic field is non-separable and, as a result, cannot be solved analytically. Many theoretical attempts have been made to obtain the solution of this problem because a Rydberg atom in a magnetic field is one of the best systems with which to study quantum chaos [3,4]. The motion of a Rydberg electron in a magnetic field is classically chaotic when B and n satisfy certain conditions. Although calculations successfully reproduce energy levels of simple atoms, such as hydrogen, it is difficult to calculate energy levels of Rydberg molecules because of the complicated structure of the energy levels due to rotation and vibration. To analyze the spectra of NO, therefore, we used a semi-classical calculation that was developed to calculate resonance structures of high-n Rydberg atoms in a magnetic field [5-71. Using cylindrical coordinates, the potential for the Rydberg electron in a magnetic field is given by:

V(p, Z)=(m2 - 1/4)/2p2 - (p2+Z2)-in- mB/2 + B2 p2/8.

(2)

Because the electron is excited near the IP, the electron can move freely in the Z direction, although the motion in the X-Y plane is bound by the magnetic field. Therefore, the motion in the Z direction does not contribute significantly to the resonance structure, and can therefore be neglected. Applying the Bohr-Sommerfeld quantization condition to the motion in the X-Y plane leads to:

Pl

I

[E-V(p, O)lindp = (n,+1/2)7~,

(3)

P2

where E is the energy of the electron, n, is a reduced quantum number given by n,=(n-6)-rn-l, and p, and p2 are the inner and outer classical turning points, respectively. The energy was calculated numerically as a function of n. The positions of the peaks appearing above the IP were well reproduced by the

140

calculation [8,9]. After the first observation in Ba atoms by Garton and Tomkins [lo], a resonance around the IP has been observed in other atoms such as Na and Rb. The resonance is known as quasi-Landau resonance and can be explained semi-classically as periodic, two-dimensional motion of the Rydberg electron in the X-Y plane driven by both Coulomb and Lorenz forces. Therefore, the resonance observed above the IPis assigned to the quasi-Landau resonance of NO, which was observed in molecules for the first time. Acknowledgments

I wish to thank Dr. Hitoshi Wada for valuable advice. References 1. Takazawa, K., Abe, H., Electronic spectra of gaseous nitric oxide in magnetic fields up to 10 T, J. Chem. Phys. 110 (1999) pp. 9492-9499. 2. Takazawa, K., Abe, H., Wada, H., Zeeman electronic spectra of gaseous NO in very high magnetic fields up to 25 T, Chem. Phys. Lett. 329 (2000) pp. 405-4 11. 3. T.F. Gallagher, “Rydberg Atoms”, Cambridge university press, 1944. 4. Ruder, H., Wunner, G., Herold, H., Geyer, F., Atoms in Strong Magnetic Fields, Springer-Verlag, Berlin, 1994. 5. Edmonds, A.R., The Theory of the Quadratic Zeeman effect, J. Phys. (Paris) 31 (1970) pp. C4-71424-74. 6. Starace, A.F., Quasi-Landau Spectrum of a Hydrogen-Like Atom in a High Magnetic Field, J. Phys. B 6 (1973) pp. 585-590. 7. Economou, N.P., Freeman, R.R., Liao, P.F., Diamagnetic structure of Rb in intense magnetic fields, Phys. Rev. A 18 (1978) pp. 2506-2509. 8. Takazawa, K., Abe, H., Landau level of gaseous nitric oxide studied by two-color multiphoton ionization spectroscopy, J. Chem. Phys. 110 (1999) pp. 11682-11684. 9. Takazawa, K., Magnetic field effect on highly excited states near ionization potential of nitric oxide, Sci. Tech. A h . Mat., 4 (2003) pp. 253-260. lO.Garton, W.R.S., Tomkins, F.S., Diamagnetic Zeeman Effect and Magnetic Configuration Mixing in Long Spectral Series of Bd*,Astrophys. J. 158 (1969) p. 839.

APPLICATION OF HIGH MAGNETIC FIELD TO CHEMICALAND PHYSICAL PROCESSES Y. TANIMOTO, W. DUAN Institutefor Molecular Science, Okazaki, Japan We have researched the effects of high vertical magnetic field (15 T, 1500 T'm-') on (1) laser-induced convection of benzene solution of a photochromic compound and (2) magnesium silicate membrane tube formation reaction. In (I), the speed of convection of the solution is retarded initially and then accelerated in a magnetic field gradient of -1500 T'm-', whereas it is accelerated initially and then retarded in a field gradient of +1200 T'ni'. The results are interpreted in terms of magnetic force on the solution. In (2), the membrane tubes grow helically in a magnetic field of 15 T, whereas they grow straight upward in zero field. In situ observation revealed that sodium silicate aqueous solution undergoes circular convection in the magnetic field only during the reaction. The results are interpreted in terms of the Lorentz force on the solution flowing out of the tubes.

1. Introduction We have researched the effects of high vertical magnetic field on chemical and physical processes using a new superconducting magnet (JASTEC, LH15T40). It is compact in size (about $800 mm x height 1800 mm) and generated a constant high magnetic field (15 T) and high magnetic fieldxfield gradient (1500 T'm-') in a (40 mm vertical room temperature bore tube for two years. Here, we shall present results of studies done while using this magnet.

2. Magnetic Field Effect on Laser-Induced Convection of Benzene Solution of a Photochromic Compound A photochromic compound, cis- 1,2-dicyano-1,2-bis (2,4,5-trimethyl-3-thienyl) ethene (CMTE) undergoes a photo-induced isomerization reaction as shown in Figure 1 [l]. This solution is used to follow the laser-irradiation induced convection of the solution since, upon UV-irradiation, the colorless CMTE solution becomes red due to the formation of CMTE photo-isomer [2].

141

142

Colmlm fnrm.4

ColorcdformB Abr. 520 nm

Abr. 370 nm

Figure 1 . Photo-inducedreaction of CIvlTE.

A benzene solution of CMTE in a quartz cell ( 1 x 1 ~ 4cm) is irradiated with a XeCl pulse laser (308 nm) from the bottom, and the movement of the colored solution is observed from the side using a CCD camera. Figure 2 shows the movement of the peak of the density of the colored solution in magnetic fields. The movement of the solution will be divided into two stages. The first stage is its removal from the bottom of the vessel. The second is its motion in bulk solution after this removal. At zero field, the colored solution rises vertically from the bottom surface of the vessel at about 5 s after laser excitation. At -1500 T'm-', the colored solution starts to move at about 9 s after laser excitation and then it moves rapidly upward. At +1200 T'm-', it moves quickly, rises slowly, and finally returns to the lower region of the solution. These interesting results will be explained in terms of the magnetic force on the CMTE solutions. 250 h

3

9 200

2 E

150

E

2 100 8 50

a Time (s) Figure 2. Movement of colored solution after laser excitation (250 pixel = ca. 10 mm).

143 Through laser irradiation, the photon energy is partly used to initiate the isomerization reaction and the rest is used to heat the solution. So it is expected that, using laser irradiation, density and magnetic susceptibility of the solution would be changed. The first step in the movement of the colored solution is its move from the bottom surface of the vessel. In order for this to occur, the bulk solution adjacent to the colored solution must replace the volume occupied by the colored solution. At -1500 T'rn-', the upward magnetic force almost balances with gravity, therefore the force (solution pressure) to replace the colorless solution is too small to push the colored solution upward immediately after laser irradiation. At +1200 T'm-', the force is downward and the sum of gravity and magnetic force is roughly estimated to be 1.8 times larger than that of gravity. The force in the solution to push out the colored solution is strong enough to remove it immediately. This is why the removal rate of the colored solution is different in two cases. The movement of the colored solution in the bulk solution can be explained by the magnetic buoyancy, Fmb, of the colored solution, given by the following equation:

where, pA and pB are the densities of bulk and colored solution, respectively, g is gravity, VB is the volume of colored solution, f i and a are the magnetic susceptibility of bulk and colored solution, respectively, is magnetic permeability of vacuum, B is the magnetic flux density and dB/& is its gradient to the z direction, which is vertical. Once the colored solution moves from the bottom surface and floats in the bulk solution, Equation 1 determines the speed of the upward movement. Since heated solution is low in density, the solution moves upward regardless of the presence of a magnetic field. During this motion, the colored solution is cooled by the surrounding bulk solution. If this is the sole process, the colored solution will be stopped at the upper part of the solution and it is hard to explain why the solution moves down at +1200 T'm-'. This unique phenomenon will be explained only by accounting for the change of the magnetic susceptibility. CMTE isomerizes by laser irradiation to the colored form. Thus, if &0 a

60

80

100

120

Electrophoresis voltage, V N Figure 3. Fractional change in velocityfvs. applied voltage Vin the magnetically aligned gels for the cyclic and linear DNAs.

The velocity v, or v, of the cyclic DNA is much larger than that of the linear DNA in spite of the same size. This suggests that the velocity depends strongly on the structure of the DNA itself. Figure 3 illustrates that the two DNAs have different dependences of the electrophoretic velocity on the applied voltage. Here, we evaluated the magnetic effect by the fractional change in velocity, which was defined byf= ( v , - v ~ ) l v As ~ . the applied voltage is increased, the linear DNA shows a monotonic increase in f; in contrast to this, the cyclic DNA has a peak in f. The effective size of the cyclic DNA is considered to be 857 bp of the circular diameter. This size is nearly equal to the gel mesh. Thus, the cyclic DNA receives a considerable influence to its velocity from the gel structure. We then selected two types of DNA that are difficult to separate by electrophoresis in the usual random gel because they have the same velocity. One was the liner DNA of Molecular Ruler (1000 bp) and the other is the cyclic DNA of pUC18 (2690 bp). We measured the electophoretic distance d in the aligned and random gels for the two DNAs when the duration was 100 min., the applied voltage was 70 V and the temperature was 5 "C. We observed one overlapped band with d = 27.5 mm for both DNAs in the random gel, but two separated bands with d = 30.0 for the cyclic DNA and d = 28.0 for the linear DNA in the aligned gel. This result leads to a new method of high-resolution electrophoresis that can separate DNAs with different structures.

154

Acknowledgements This work was supported by Grant-in-Aid for Scientific Research (B) (No. 10018046) and Grant-in-Aid for Scientific Research for Priority Areas (No. 15085204, Area 767) from MEXT of Japan. References 1. Yamamoto, I., Ishikawa, K., Mizusaki, S., Shimazu, Y., Yamaguchi, M.,

2.

3.

4.

5.

Ishikawa, F., Goto, T., and Takamasu, T., Jpn. J. Appl. Phys. 41 (2002) pp. 416-424. Yamaguchi, M., Yamamoto, I., Ishikawa, F., Goto, T., and Miura, S . , J. Alloy and Comp. 253-254 (1997) pp. 191-194. Yamaguchi, M., and Yamamoto, I., Dynamic Spin Chemistry, ed by: S. Nagakura, H. Hayashi and T. Azumi, pp.131-151, Kodansha-Wiley, Tokyo (1998). Yamamoto, I., Deguchi, N., Yamaguchi, M., Shimazu, Y., Ishikawa, F., and Miura, S., Physica B 246-247, (1998) pp. 408-41 1. Matsumoto, Y., Yatnamoto, I., Yamaguchi, M., Shimazu, Y., and Ishikawa, F., Jpn. J. A&. Phys. 36, (1997) L1397-L1399.

Control of Liquids


APPLICATIONS OF AC AND DC MAGNETIC FIELDS IN METALLURGICAL AND CRYSTAL GROWTH PROCESSES A. CRAMER, S. ECKERT, V. GALINDO, J. PRIEDE AND G. GERBETH Forschungszentrum Rossendor- Dresden, Germany Nowadays, magnetic fields are widely utilized in metallurgical and crystal growth applications on an industrial scale. This paper focuses on laboratory studies using liquid metals with a melting point up to 300 O C to model the realistic processes. Based on four examples it is shown that this physical modelig has a great potential in both, revealing an insight into the basic phenomena involved, and the optimization of the technological details. Direct melt extraction of fibers and the aluminum investment casting have been chosen representatively for the damping of flow by a DC magnetic field. Stabilization, i.e. the retarding of instabilities, is demonstrated with electromagnetic levitation. We are considering the floating-zone technique as an example kom the large field of crystal growth

1. Introduction Magnetic fields provide an attractive medium for controlling flow in disciplines such as metallurgy and crystal growth. Owing to the industrial relevant metallic and semiconductor melts being hot and aggressive, any installation of an electromagnetic system requires considerable effort. There will always be a compromise between effectuality and affordable costs. The design of the magnets primarily relies on numerical studies of both the electromagnetical and the fluid mechanical parts of the problem. Whereas the calculations for the first are highly trustworthy because of the linearity of Maxwell's equations, this is not true for the numerical simulation of the flow. The use of turbulence models to cope with the non-linear equations of motion absolutely demands experimental validation. Mere water experiments are performed in many cases. If there are any temperature gradients, free surfaces, or two-phase flows involved, the characteristic non-dimensional parameters such as Reynolds, Grashof, and Prandtl numbers cannot be met. All modeling by water experiments is meaningless for the flows exposed to electromagnetic fields. Often, the results of non-validated numerical simulations are more or less improper and, in turn, the performance of the expensive magnets deployed in an industrial installation falls far behind expectations. A variety of liquid metals exist in the reduced temperature range up to approximately 300 O C , commencing with eutectic InGaSn and mercury that are both liquid at room temperature. They all have the typical high electrical and 157

158

thermal conductivity, high surface tension and density, and low Prandtl number like those metals used in real technical processes. Moderate variations in the combination of the physical properties between the different low melting point model fluids, together with the freedom to adapt the geometrical parameters, facilitates to reproduce almost any industrial process with a sufficient matching of the similarity criteria [ 11. Several measuring techniques exist to determine local fluid velocities for this class of liquid metals [2]. Moreover, the upcoming ultrasonic Doppler velocimetry (UDV) is capable of acquiring spatio-temporal information in terms of a complete profile of the velocity component along the ultrasonic beam. The availability of precise experimental data allows for a detailed study of the impact of magnetic fields on a volume of electrically conducting media. As a result, a deeper insight into the physical mechanisms can be expected which, in turn, enables optimization of the field efficiency. Following this approach several magnetic systems have been designed, some of which are up-scaled and installed at industrial facilities. Subsequently four examples are described that we believe to be a representative selection for this strategy that may be termed a tailor-made flow control. In the next section we examine the extraction of metallic fibers from an open crucible. The sole difference between this and the process of melt spinning is the supply of the liquid metal to the extracting substrate, which is a swiftly rotating wheel in both cases. The advantage of the melt extraction process is its simplicity, since no precisely adjusted liquid metal jet must impinge onto the wheel. Due to the high circumferential speed, non-stationary deflections of the melt pool's surface are created. These deflections are the major drawback to this method. The solution appears to be to damp turbulent motion using a DC magnetic field. As we shall see, it is the topology of the field via prescribed flux guidance rather than solely the amplitude of the field that significantly increases the stability of the process. Section 3 is devoted to an aluminum investment casting process. The request was to reduce the high flow velocities arising in the early stage of the pouring procedure. If the speed of the liquid metal is too high, oxides and impurities may be entrapped in the bulk of the melt. Again, a DC magnetic field was utilized to damp the flow. An optimized solution for the strength and the geometry of the field, and its position at the caster was investigated. The instabilities of electromagnetically levitated samples are described in section 4. In real applications, high frequency magnetic fields are used to levitate metal and to melt as a means for containerless processing. For specific

159 working parameters, which are often unavoidable, the still solid sample begins to oscillate and rotate. Steady magnetic fields have proven to stabilize the levitation if they comply with certain conditions, one of which is that they must not be homogeneous and isotropic. In that case, stabilization cannot be achieved no matter how strong is the field. Finally, the example in section 5 is concerned with floating-zone crystal growth technology. The single-phase high frequency inductor that is usually deployed creates the typical double-vortex convection. For several substances to grow single-crystalline, the shape of the phase boundary determined by the heat transport properties of that flow is disadvantageous. Adding a second passive coil wired to a tunable resonance network allows changing the flow structure completely. The temperature distribution at the solidification front can be adjusted to amend the concave boundary to the desired convex one.

fiber 7

Figure 1. Principle of the direct melt extraction process

2. Direct Melt Extraction of Metallic Fibers High-grade porous metallic substrates may be manufactured from fibers with diameters in the range of several tens of pm. 'Melt extraction, which belongs to the near net-shape casting processes, is a promising technology for the production of such fibers. A high-speed wheel is brought in contact with the surface of an inductively heated pool. Due to its high circumferential speed, it tears out material directly from the surface. The melt is quenched at the watercooled wheel, then solidifies and shrinks, and is flung away in the end by centrifugal forces (Figure 1).

160

Several sources generating fluid motion are present in such a system. On the one hand, electromagnetic heating creates temperature differences within the liquid metal pool, giving rise to buoyancy and possibly thermocapillary convection. On the other hand, the alternating magnetic field directly has a A fourth stimng effect. mechanism is the stress supplied by the wheel at the surface. A more detailed description can be found in [3]. The resulting bulk flow turbulence and the wavy Figure 2. Calculated distribution of the flux density for motion on the surface give rise to the iron-ring equipped chill wheel. At the extractionzone (bottom) the iron is almost saturated. strong and time dependent surface deformations that, in turn, cause non-stationary conditions of the contact zone between the extraction wheel and the melt. Many attempts have been made to overcome these limitations by submerging mechanical parts, but they then suffer from corrosion or cracks and do not work reliably. The expected benefit from the impact of stabilizing measures was to attain higher revolution rates of the extraction wheel, and thereby decrease the mean fiber diameter as well as the width of the diameter distribution. A non-invasive control mechanism can be provided by a damping DC magnetic field. Reduced temperature model experiments (using SnPb at 250 OC and InGaSn at room temperature) have been performed. Over wide parameter ranges of the Reynolds number according to the alternating magnetic field, and the Hartmann number describing the strength of the DC field, the damping characteristic has been quantified in an inductively stirred InGaSn melt. It was found that, with moderate field strength of a few hundred mT, all velocity fluctuations can be completely damped and the liquid metal surface becomes as smooth as glass. This stabilization of the melt surface has also been demonstrated in practice by the installation of a solenoid at a working extraction facility under rough industrial conditions [3]. But it clearly turned out that a significant decrease of the fiber diameter could not be achieved. Further increasing the magnetic induction inhibited the extraction completely.

161

In a second series of model experiments, fibers have been extracted from a resistance-wire heated SnPb-melt. As soon as the extraction started, small ripples appeared in front of the chill wheel, which so far could not be resolved within the large surface deformation created by induction heating. With the help of a high-speed video camera, it could be proven that the wavy motion is coherent with the meniscus fluctuations (Figure 1). Additionally, the recordings revealed why the increase of the wheel speed does not diminish the fiber diameter. While fibers are extracted, the filling level of the liquid metal in the crucible decreases despite a constant lift. When the difference in height between the fluid level and the wheel exceeds a certain value, the meniscus is no longer able to keep the contact between the metal surface and the chill wheel. During the contactless period the level raises until it touches the wheel again. Astonishingly, this almost periodic procedure repeats at high frequencies well above inertia of the naked eye. Reasoning from order of magnitude estimations, the only possible driving force for that instability is either thermo- or chemocapillarity or both, the latter then is due to oxidation of the metal surface. It is easily calculated that the magnetic field required to calm down the capilIary instability is roughly an order of magnitude stronger than that required for the electromagnetically driven one. Unfortunately, the virtual viscosity of the melt increases along with the strength of the applied field, leading to a broadening of the liquid layer that is advected along with the wheel. As this boundary layer thickness is another determining parameter, the resulting fiber diameter increases.

Figure 3. Diameter histogram of fibers extracted without (dark)and with local stabilization.

A solution to the problem would be a strong magnetic field localized at the meniscus region and decaying rapidly with increasing distance. Using

162

ferromagnetic material for the outer part of the wheel with respect to the radius concentrates the magnetic flux density at the tip of the extraction wheel's edge that is the only part in contact with the melt. By numerical optimization we accomplished magnetic saturation of the iron for a field strength of only 70 mT, applied to the entire fluid volume. The numerical result is shown in Figure 2. From the histogram in Figure 3 it can be seen that both the mean diameter and the width of the distribution are significantly shifted to smaller values when applying the local stabilization. For a more detailed description and additional results we refer to [4]. 3.

Aluminum Investment Casting

Reduction of too high flow velocities is a persistent objective in investment casting processes. Particularly during the early stage of the filling, the related high rate of turbulence in the flow should entail the transport of impurities, oxides, or gas bubbles from the walls and the free surface into the bulk of the melt. Consequently, deterioration of the mechanical properties of the casting products is likely to be expected. Whenever breaking of fluid flow of electrically conducting media is a topic, DC magnetic fields are the first choice for a non-invasive control mechanism. However, the challenge in the pouring process lies in the avoidance of an accumulated generation of vortices inside the pouring channel rather than in dampening the mean velocity. In order to find an optimized solution for the strength, the geometry, and the position of the magnetic field at the casting unit, Figure 4. Calculated velocity field of a liquid aluminum the velocity field mussty be pouring flow. Left: without magnetic field; right: determined. This task was transverse field of 0.5 T. Velocities are given in d s . fulfilled by a combination of numerical simulation of the flow field and isothermal physical modeling of the casting process. We performed the numerical calculations using the commercial finite element code FIDAP (FLUENT Inc.) that we extended by a term for the

163 electromagnetic Lorentz force, and an additional equation to solve for the electric potential. UDV was applied for the local velocity measurements, and the flow rate was determined independently employing the contactless transmitter technique [2]. The use of eutectic InGaSn (Tmelt = l0T) permitted to set up an easy to build and flexible perspex model, which basically is a U-bend where the vertical funnel is connected via a horizontal channel to a cylinder being a dummy for the mold. Figure 4 depicts the numerical simulation, the result of which states that the highest velocities are found at the conjunction between the down sprue and the horizontal channel. For the channel flow, it is well known that transverse magnetic fields provide the most efficient damping for which the optimal position obviously must be sought in the bottom region of the casting model. Experimental results compare well with the calculated ones, thus validating the numerical model. In order to quantify the performance characteristics of the magnetic field, the dependence of the flow rate on the induction and the position was studied both experimentally and numerically. The zone of highest velocities, covering the entire bend between the funnel and the horizontal channel, was the optimal location of the magnet. At that position and for a field strength of 0.5 T, the braking effect on the flow in the mid-plane of the considered channel is obvious, as can be seen in Figure 4. Two experimental results that demonstrate impressively the effect of the DC field are shown in Figure 5. On the left side it can be seen that the velocity peaks at the 2.5 1.25

2.0

1.00

7 . 7

3 3

Y

2

lS

,G

1.0

0.50

s

$

0.75

0.25

0.5

0.00

0.0

0

5

10

time [s]

15

20

time [s]

Figure 5 . UDV velocity measurements for various values of the magnetic induction. Left: in the vertical funnel at 50 mm distance from the bottom; right: in the horizontal channel, a bubble comes past the position of the ultrasonic beam.

very beginning of the pouring process are completely suppressed. Further, for the smallest field strength of 0.25 T, the intensity of velocity fluctuations is significantly decreased. On the right hand side of Figure 5, a typical signal for a bubble passing the measuring volume demonstrates the capability of the UDV

164

method to track such events. From both the UDV signals and the visual observations through the transparent perspex walls of the channels, we see that the amount of gas inclusions is significantly reduced when the flow is exposed to the DC field. A magnetic system with parameters comparable to the model’s nondimensional parameters was manufactured and installed at an industrial facility where the feasibility of the method has been demonstrated in a real aluminum casting process. As a further improvement, we are currently investigateing the action of a linear traveling field, which allows for an initial breaking and a final pumping, thus enabling achievement of a constant flow rate during the entire casting process. 4.

Levitation

Electromagnetic levitation is a well-known technique for containerless processing of metals and alloys both in the solid and molten states. This technique basically consists of one or two coils with a few windings (Figure 6) fed by an alternating current of some 100 kHz in frequency. The alternating magnetic field induces eddy currents in the electrically conducting sample that provide the contactless heating as well as the levitation, the latter being established by the repulsive interaction between the applied field and the induced currents. Besides processing, electromagnetic levitation is widely used in material research to measure thermophysical properties of molten metals. This technique has an essential drawback. The sample begins to rotate andor oscillate as sketched in Figure 6. These instabilities may grow to such an extent that the sample escapes form the coil volume. Even if the amplitude of the objectionable movements stays constricted, a serious disturbance of the process and its measurement has to be considered. We analyzed the reasons for the occurrence of spontaneous rotational or oscillatory instabilities [5, 61. The instability depends Figure 6. Scheme of the coils together with a solely on the frequency of the levitated sphere. The possible spatial alignment of oscillatory and rotational instabilities is sketched

applied field rather than on its strength. The characteristic nondimensional parameter is determined by the frequency m = ~ o w R *with the sample radius R, magnetic permeability p, electrical conductivity o of the

165

sample, and frequency o of the current feeding the levitation coil. For the case studies of the two basic configurations, i.e. the uniform field typically used for heating, and the linear field representative for a positioning field, the theoretical results can be summarized as follows. A levitated sphere may show spontaneous rotations or oscillations if the non-dimensional field frequency exceeds a critical value m,. The values are mC= 11.609 both for the rotational instability in a uniform field and the oscillatory instability in a linear magnetic field, and m, = 27.682 for a rotational instability in a linear field. In all cases, a certain frequency m, can be found such that the growth rate of the instability has a maximum. Those values are a, = 18.073 both for the rotational instability in a uniform field and the oscillatory instability in a linear magnetic field, and m, = 47.196 for a rotational instability in a linear field. There is no oscillatory instability in a uniform alternating magnetic field. In axi-symmetric levitating fields, the rotation axis is perpendicular to the axis of symmetry of the field that is usually vertical (Figure 6). The spontaneous oscillations occur in vertical and radial directions, whereas the frequency of vertical oscillations is twice that of the radial ones. The simplest way to avoid such instabilities would be to keep the nondimensional frequency below the threshold m,. This is often difficult, as it would require working with field frequencies or samples that were too low or small. If instabilities are basically unavoidable, DC magnetic fields represent a powerful tool for an efficient, active damping of all such instabilities. There is no need to work with strong stabilizing DC fields. The theoretical predictions [5, 61 show that a magnetic field strength of about 1/4 of the levitation field amplitude (typically in the order of several tens of mT) is sufficient. However, the DC field direction is crucial. A field along the direction of the levitating field is expected to damp rotations but not oscillations. A DC field perpendicular to the levitating field direction damps oscillations, but not rotations. The rotation axis adapts to the given DC field direction and no rotational damping occurs anymore, independent of the DC field strength. Hence, sophisticated field geometries are necessary for an overall stabilization of the sample. This instability behaviour has been demonstrated in model experiments using solid Al or Mg spheres [7]. Two realizations have been achieved for DC field stabilization. An electrotechnical approach was first used by superimposing a DC current to the levitating coils, thus providing a cusp-type DC magnetic field. For complete stabilization, an additional horizontal magnetic field was added. The second approach consisted of using permanent magnets arranged on a common ring in order to produce a strongly non-uniform field distribution. A

166

full stabilization of the levitated samples was reached with both approaches [7]. The above considerations address instabilities of the sample as a whole; independent whether solid or liquid. For a molten droplet, the internal fluid flow and its stability can become important so they should also be taken into account

PI. 5.

Floating-zone Crystal Growth

In crystal growth technology, contactless methods of flow control are imperative to avoid any contamination of the melt. Historically, steady magnetic fields were used first to damp oscillatory flows. A superior flow control with regard to the 2 desired heat and mass transport I properties of the growth process 0 can be achieved by a suitable I 2 combination of flow damping 3 (dc fields) and flow driving (ac 4 fields). As examples, we refer to 5 the development of such a 6 7 combined magnetic field system for the industrial Czochralski Figure 7. Scheme of the floating-zone process and the growth of silicon [9], and to the leading flow structure in the molten p a . application of a rotating magnetic field in the vertical gradient freeze growth of gallium-arsenide [ 101. In this paper we dwell on another instance of a tailored magnetic flow control which, initially, seems to be a minor modification of the floating-zone an arrangement but has Figure 8. Scheme of the modified floating-zone extensive influence on the flow, arrangement and the leading flow structure in the molten thus on the growth process. The Part. usual float-zone system consists of a single coil fed by a current of several 1 O O k H z in order to melt the polycrystalline material that re-solidifies as a single crystal when the rod is moved vertically (Figure 7). The typical flow structure in the molten part consists of a toroidal double vortex that obviously supports a concave shape of the solid-liquid interface. From experience, it is commonly believed that this

167 concave contour promotes polycrystalline growth. Thus, the apparent request was to modify the convective pattern so that the solid-liquid phase boundary becomes convex in shape. Such a change of the flow structure can be obtained by additional AC magnetic fields. An efficient and easily implemented solution requires adding a second coil (Figure 8), which does not need an auxiliary power supply. Instead, a tunable resonance circuit consisting of a capacitor and a resistor is wired to the coil, and the current in this secondary circuit is induced by the primary coil. Thus, a two-phase traveling magnetic field is created that drives the melt from the primary to the secondary coil. The relative strength of this superimposed traveling field is determined by the phase shift that can be adjusted by the impedance of the resonance network. If this relative strength is sufficient, it changes the flow structure to a single vortex. On the right side of Figure 8 a numerical solution graphs the flow lines of the modified convective structure that obviously supports a convex shape at the upper solid-liquid phase boundary. A full numerical simulation of both the electromagnetic and the hydrodynamic problem has been made for a selection of process parameters such as the vertical distance between the coils, and the capacitance and resistance of the resonance circuit, which can be constructed in practice. Subsequently, the results were implemented at a float-zone facility. The first investigations using Ni as a model substance clearly demonstrate the theoretically predicted tendencies concerning the influence of this two-coil arrangement on the shape of the phase boundary ill]. Acknowledgement Financial support from ”Deutsche Forschungsgemeinschaft” in frame of the Collaborative Research Centre SFE3 609 is gratefully acknowledged. References 1. Cramer, A., Eckert, S . , Galindo, V., Gerbeth, G., Willers, B., Witke, W., Liquid metal model experiments on casting and solidification processes, Proc. Int. Symp. “Liquid metal processing and casting”, Eds. Lee, P.D., Mitchell, A., Bellot, J.-P., Jardy, A., Nancy, France (Sept., 2003), pp. 333343. 2. Eckert, S., Gerbeth, G., Gundrum, T., Stefani, F., Witke, W., New approaches to determine the velocity field in metallic melts, 4” Int. Conf. on “Electromagnetic processing of Materials” (EPM 2003), Proc. PL13, Lyon, France (Oct., 2003). 3. Cramer, A., Bojarevics, A., Gerbeth, G., Gelfgat, Yu., Stabilization of the melt extraction process with a magnetic field, Proc. Int. Symp. “Fluid flow phenomena in materials processing”, 128* ann. meet., Eds. N. El-Kaddah,

168

D.G.C. Robertson, S.T. Johansen, V.R. Voller, ISBN 0-87339-424-0, San Diego, USA (April, 1999), pp. 237-244. 4. Cramer, A., Gerbeth, G., Bojarevics, A., Gelfgat, Yu.,Stabilizin the direct melt extraction of intermetallic fibers by magnetic fields, Proc. 3‘9 Int. Conf. on “Electromagnetic processing of Materials” (EPM 2000), Nagoya, Japan (April, 2000), pp. 147- 152. 5. Priede, J., Gerbeth, G., Spin-up instability of electromagnetically levitated spherical bodies, IEEE Trans. on Magnetics, 36,2000, pp. 349-353. 6. Priede, J., Gerbeth, G., Oscillatory instability of electromagnetically levitated solid bodies, IEEE Trans on Magnetics, 36,2000, pp 354-357. 7. Priede, J., Gerbeth, G., Mkelsons, A., Gelfgat, Yu., Instabilities of electromagnetically levitated bodies and its prevention, Proc. 31dInt. Conf. on “Electromagnetic processing of Materials” (EPM 2000), Nagoya, Japan (April, 2000), pp. 352-357. 8. Shatrov, V., Priede, J., Gerbeth, G., Three-dimensional linear stability analysis of the flow in a liquid spherical droplet driven by an alternating magnetic field, Phys. Fluids, 15 (3), 2003, pp. 668-678. 9. Galindo, G., Gerbeth, G., von Ammon, W., Tomzig, E., Virbulis, J., Crystal growth melt flow control by means of magnetic fields, Energy Conversion and Management, 43,2003, pp. 309-316. lO.Patzold, O., Grants, I., Wunderwald, U., Jenkner, K., Croll, A., Gerbeth, G., Effect of a rotating magnetic field on the heat flow in vertical gradient freeze growth of GaAs, J. Crystal Growth, 245 (3-4), 2002, pp. 237-246. ll.Priede, J., Gerbeth, G., Hermann, R., Filip, O., Behr, G., Two-phase induction melting with tailored flow control, 4” Int. Conf. on “Electromagnetic processing of Materials” (EPM 2003), Proc. P2 13, Lyon, France (Oct., 2003)

ELECTROMAGNETIC PROCESSING OF MATERIALS: FROM THE CONCEPTS TO INDUSTRIAL APPLICATIONS Y. DELANNOY CNRS - Laboratory EPM lformerly Madylam), Grenoble, France. Electromagnetic fields are used for material processing in various industrial devices, such as induction furnaces, electromagneticbrakes and stirrers in metallurgy, inductive plasma torches to elaborate silica for optical fibres or electromagnetic flow control systems in crystal growth. New developments are needed whenever the coupling of physical phenomena is the key point of the process. Three examples are presented among the research activities of the EPM laboratory in Electromagnetic Processing of Materials: Electromagnetic continuous casting of steel slabs, plasma purification of silicon, electromagnetic stirring of solidifying alloys. Some scientific open questions important for such processes are presented.

1. Introduction In the context of elaboration of materials, electromagnetic processes are used to melt a material up to a very high temperature, and without the liquid contacting any solid material other than its own solid phase. By controlling the fluid flow, electromagnetic actuators are also useful for purification of materials, mixing (alloying) at high temperature, and separating inclusions such as solid oxides or bubbles. When used during solidification, electromagnetic fields can help control the structure (grain size, columnar to equiaxial transition) or the segregation (separation between alloy components). Such processes can be the key point to elaborate a new material, or simply to improve existing processes. Electromagnetic processing of materials (EPM) is based on the effects of electrical currents flowing inside the material. It is thus reserved for electrically conducting materials: high conductivity (o-1068-'m-' for metals), or weakly conducting gases (plasmas) or oxides (melted glass). The current can be induced by an alternating magnetic field, (coil with alternating current, AC), or from the interaction of a fluid flow with a steady magnetic field, produced by permanent magnets or electromagnets with direct current (DC).

169

170

Figure 1. Current induction from an AC magnetic field (a), or from a flow in a DC field (b)

In a typical induction case (Figure la), a single coil surrounds the material in which a current density j is induced in the opposite direction compared to the coil current. This current in the load produces a heating density j2/o (Joule effect). The induced magnetic field BACis directed along the coil axis and penetrates in the material inside a skin depth, depending on the frequency and on the electrical conductivity of the material. The mean resulting Lorentz force, F=jxB is generally normal to the material surface, but it can be tangential in other systems having several coils and traveling or rotating fields. In every case, the rotational part of the force field (variation of F sketched in Figure la) promotes a fluid flow (stirring effect), whereas the irrotational part is responsible for a repulsion (magnetic pressure). DC fields are used generally to brake a fluid flow of velocity V (Figure lb), using the electromotive force VxB due to the magnetic field B x from external magnets or electromagnets. The Lorentz force is, indeed, braking the fluid when the current flows in the direction of the electromotive force (such as in the center of the Hartmann flow sketched in Figure lb), but can accelerate it in regions of reversed electrical currents (near the vertical walls in Figure lb). This reversal is due to the voltage gradient, a potential created to close the current loops (Figure l b includes a voltage color map). The resulting effect is a bidimensionnalisation of the flow (it becomes uniform along the magnetic lines) and, generally, a net braking.

2. Industrial Examples A number of electromagnetic devices are used in industrial processes. Joule heating is certainly the most widely used effect of electromagnetic fields. It is the basis of induction furnaces or the direct heating of metal pieces, which is used for hardening processes [ 13. The stirring action of AC fields is also used in

171

many stirrers in metallurgy. Thousands of devices using rotating low frequency fields are used every day in continuous casting of metal billets or other long products, and linear motors are becoming popular in the steel industry for slab casters [2], [3]. A slab casting machine (flat products) is sketched in Figure 2a with two commonly used stirrer positions (in-mold for flow control, facing the final solidification zone for grain refinement). Alternative positions are inside the rolls, or just below the mold.

Figure 2. Stirrer in slab casting (a), plasma torch for optical fibers (b), MCZ system (c).

The heating and stining effects are used together in plasma torches, where a gas flow (partially ionized) is heated by induction up to typically 12000 K. Hundreds of torches are used in optical fiber factories for silica synthesis and deposition (Figure 2b). The synthesis reaction (SiC14+02+Si02+2Cl2) is activated by high temperature gas flow, which also ensures the absence of water vapor (that would lead to absorption of infrared light). Heat transfer towards the silica cylinder ("preform") ensures, in one pass, the vitrification of previously deposited silica soot, and a new deposition. However, the efficiency of the process is strongly dependant on the boundary layer separation along the preform, so the plasma flow must be carefully controlled. Hundreds of crystal growth facilities, using the Czochralski process (CZ) or the Floating Zone process (FZ), are strategic in that they provide wafers for the electronic industry. Those standard processes use the Joule effect for direct (FZ) or indirect (CZ) induction melting, the latest using a carbon crucible as susceptor [4]. Flow control systems are sometimes added, using DC fields [4] or rotating AC fields [5], in order to control the convection and thus the distribution of dopants and pollutants. The most popular configuration in industry is sketched in Figure 2c: opposite DC coils are added around a classical Czochralski machine in order to produce a cusp magnetic field in the melt. The resulting process is called Magnetic-Czochralski (MCZ).

172

3. Examples of Process Development The EPM laboratory specializes in the development of electromagnetic processes for materials, and in some associated research (plasma, solidification, magnetohydrodynamics). As do other teams, it now investigates processes for which a single electromagnetic field effect is insufficient, but where the interaction between electromagnetic and/or non-electromagnetic phenomena is important. Electromagnetic shaping of a free surface can thus interact with solidification, melting and/or dissolution, resulting in applications such as direct casting (the mold is replaced by a coil), production of amorphous metals in levitation, or dissolution in a superheated metal in semi-levitation. The problem is then to control the stability of the free surface [6]. Interaction between stirring and mass transfer controls purification processes, or segregation during the solidification of an alloy, as will be shown below. Similarly, electromagnetic forces can control solid particle transport, thus providing separation processes [7]. Li

Figure 3. Electomagnetic slab casting: principle (a), mercury model (b)

As part of a Japanese project and in cooperation with Usinor, a shaping system has been developed for continuous slab casters used in the steel industry [8].In this example, two actuators are combined. First, an AC field is used to repulse the free surface from the cold mold (Figure 3a); this is known to improve the surface quality of the cast metal. The magnetic pressure works together with heating (helping to remove the solidification hooks, see Figure 3a), and stirring, which has a negative effect on surface stability. Therefore, a DC field is added to damp the free surface (braking effect), and all of the effects are studied together in a lab scale mercury model (Figure 3b), where velocities and level fluctuations are measured. Figure 3b shows repulsion

173

and agitation with the AC field, and damping when using DC. In the real steel caster, these effects interact with solidification and inclusion transport (oxide, slag or bubbles). The positive effect of the system was proven in Japan on bench casting runs using steel. This 4-year project resulted in a patented process that is available for industrial applications.

A second example of process development concerns the purification of metallurgical-grade silicon to solar-grade silicon, used to build photovoltaic cells. An inductive plasma torch is used to provide very reactive gases at the surface of a silicon melt, which is heated and stirred by a second induction system (Figure 4a). Incidentally, the magnetic pressure on the liquid can be high

0Yo

5%

10%

15%

20%

25%

Figure 4. Plasma purification of silicon (a), results from [9] on the kinetics of reaction (b)

30%

174

enough to reach semi-levitation (no contact with the cold crucible walls), making it possible to overheat the silicon and to raise the reacting surface above the crucible top. A 3-year European project [9] evaluated and proved the feasibility of the process in terms of the kinetics of boron purification (Figure 4b) and also in terms of solar cell efficiency. This purification process now enters the industrialization phase, where the key points are to optimize the kinetics (in order to reduce the cost) and to adapt the process to various kinds of metallurgical silicon. There is scientific collaboration between chemists to study the interaction of pollutants, plasma modeling to study the effect of plasma on reactive gases (and therefore on the reaction), and EPM to study the induction melting, stirring and shaping of silicon. A consortium has been built with industries interested in various stages of the project in order to develop and build an industrial pilot program. A third example of process development is the control of segregation in concentrated alloys. It is presently in its first phase (laboratory studies) in the frame of the MICAST program (Microgravity Casting) of the European Space Agency. The segregation (concentration inhomogeneities) that appear during segregation, are responsible for the structure of the alloy at microscale (dendrites), mesoscale (channels called "freckles", Figure 5a), or macroscale (purification effect by reject of solute). Freckles are due to convection effects in the mushy zone, maintaining liquid zones during solidification by a chimney effect. Our goal is to use appropriate magnetic fields to remove those freckles by homogenizing the mushy zone. The stirring interacts with the solutal convection flow in the (porous) mushy zone, with solidification and mass transport, and eventually with therrnoconvection if the joule heating is significant.

Figure 5. Freckles in experimental ingot (a) and calculated concentrations (b) from [lo]

As shown in Figure 5, the model is able to reproduce qualitatively the phenomenon observed with natural convection only. A validation experiment

175

has also been run in a quasi-2-D box, and the model predicts fairly well the freckles together with a macro-segregation. It was also shown numerically that a rotating magnetic field creates a secondary flow in the mushy zone and makes the randomly distributed freckles disappear, in profit of a central channel due to that secondary flow [lo]. Electromagnetic effects are thus important, but the actuator needs to be carefully optimized to a homogeneous situation.

4. Perspectives and Open Questions Developing a new process sometimes makes a fundamental question appear because some phenomena are not well understood and cannot be modeled in a predictive manner. This is the case for free surface stability and turbulence in magnetic fields. Nevertheless, those phenomena are often of secondary importance compared to the key electromagnetic effects.

Figure 6. Fiber coating system with free surface effects (a) and 2-Dturbulent simulations (b)

The control and stability of free surfaces is crucially important in levitation, semi-levitation or direct casting systems where it controls the thermal losses, or in coating processes where it simply controls the existence of the coating film. Figure 6a shows a fiber coating process from a semi-levitating liquid metal blob. In that case, the shape of the free surface controls the height of the blob and its temperature, thus the immersion length of the fiber and the quantity of metal entrained. The stability of the surface-tension-controlled meniscus at the fiber exit is important to get a homogeneous coating thickness. The free surface shape and stability is a subject for fundamental research [ll], to be applied to the global blob in the segmented cold crucible, or to the meniscus at the fiber exit.

176

Turbulence in magnetic fields has very special transport properties because the magnetic field produces anisotropy of turbulent vortices. At the limit, one obtains two-dimensional turbulence, which is known for its inverse cascade of energy, producing large structures from small scales. Figure 6b presents a simulation of this inverse cascade, starting from a shear layer induced by a tangential Lorentz force induced in the external part of the dish, whereas the central part experiences electromagnetic braking [ 121. However, the situation is much more complicated in industrial processes and not yet well understood. The processes that will be developed in the future of EPM must recognize such fundamental research in the frame of multidisciplinary teams. In parallel, applied research is needed to develop those processes for industrial applications, and research-industry consortiums must be used to investigate all interaction between electromagnetic phenomena as well as to face technical problems.

References EPMO3 means 4" Int. Conf. on Electromagnetic Processing of Materials, Lyon, France, 14-17 Oct. 2003. 1. Lupi, S., Modeling for research and industrial development in induction heating, EPM03. 2. Takeuchi E., EPM in continuous casting and its extensive prospect, EPM03. 3. Kunstreich S., EMS for continuous casting in the steel industry, a very successful EPM application, EPM03. 4. Li, B.Q., Solidification Processing of Materials in Magnetic Fields JOM-e, 50, Feb. 1998. 5. Dold, P., Czochralski growth of doped germanium with an applied rotating magnetic field, Cryst. Res. Technol. 38, (7-8), (2003), pp. 659 - 668. 6. Bojarevics, V., Pericleous, K., AC and DC magnetic levitation: Melting, Fluid Flow and Oscillations, EPM03. 7. El Kaddah, N., The role and use of electromagnetic fields in materials processing, EPM03. 8. Gardin, P., Dumont, B., Anderhuber, M., Galpin, J.M., Delannoy, Y., Gagnoud, A., Hamburger, J., CC Clectromagnttique de brames: dCveloppement de modkles numkriques de la configuration AC+DC en lingotibre, Revue de mitallurgie-Cahiers d'informations techniques, 98( 11) Nov. 2001. 9. Delannoy, Y., Alemany, C., Li, K-I., Proulx, P., Trassy, C., Plasma refining process to provide solar grade silicon, Solar Energy Materials and Solar Cells (ISSN 0927-0248) 72, 1-4, (2002), pp. 69-75. lO&Quillet,G., Ciobanas, A., Lehrnann, P., Delannoy, Y., Medina, M., Fautrelle, Y., Modelling of the meso-segregations in a binary alloy under the influence

177

of a forced convection, 10th Int. Conf. on Modelling of Casting, Welding and Advanced Solidification Processes, Destin (Florida, USA), May 25-30 2003, ISBN 0-87339-555-7. ll.Sneyd, A., Etay, J., Fautrelle, Y., The starfish experiment: some theoretical considerations and some experiments, 5th International Conference on "Fundamental and applied MHD" (PAMIR) Ramatuelle, France, September 16-20, 2002. 12.Delannoy, Y., Pascal, B., Alboussikre, T., Uspenski, V., Moreau, R., Quasitwo-dimensional turbulence in MHD shear flows: the MATUR experiment and simulations, in Transfer phenomena in magnetohydrodynamics and electroconducting flows, ISBN 0792355326, (1999), pp. 93- 106.

SEMICONDUCTOR CRYSTAL GROWTH IN STATIC AND ROTATING MAGNETIC FIELDS M.P. VOLZ NASA Marshall Space Flight Center, Huntsville, AL, USA Convective transport in the melt has a major influence on the resulting material properties of semiconductorcrystals. Magnetic fields can be used to control this convection. A static magnetic field establishes Lorentz forces that tend to reduce the convective intensity in the melt. In contrast, a rotating magnetic field induces a controlled flow that can alter or dominate the existing flow. The effectiveness of both static and rotating magnetic fields can be enhanced in a microgravity environment, where the driving force for buoyancy convection is greatly reduced. A review of semiconductor crystal growth experiments in static and rotating magnetic fields is presented and the relative advantages and limitations of these techniques is discussed. Efforts to combine the use of magnetic fields and microgravity are also described.

1. Introduction Convection plays a major role in determining the quality of semiconductors grown from the melt. Convection influences heat transfer and affects the rate of solidification. It can also modify the transport of species that affects the dopant distribution or degree of alloy homogeneity. Periodic convective flows resulting from either thermal or solutal buoyancy can lead to undesirable dopant distributions or striations in the grown crystals. Inhomogeneous dopant distributions may result in a degradation of the material properties required for electronic circuits or optoelectronic devices. One approach to the problem of convection is to simply try and reduce it as much as possible. A static magnetic field can be used in this regard, as it interacts with the electrically conducting melt and acts as a brake on fluid motion. Growth in a microgravity environment is another approach to reduce natural convection. Convection from buoyancy is proportional to the gravitational acceleration, and this can be reduced by six orders of magnitude in an earth-orbiting spacecraft. An alternative approach to the problem of convection is to impose a controlled convection. Controlled convection can be induced by means of a rotating magnetic field (RMF). Controlled convection can both impede the onset of unstable flow and be used to tailor the heat and mass transport in the crystal growth system. This paper will review selected crystal growth experiments conducted in static and rotating magnetic fields. The potential benefits of combining magnetic fields and microgravity will also be 178

179

described. Finally, the relative advantages and limitations of these techniques will be discussed.

2. Static Magnetic Fields Applying a static magnetic field to an electrically conducting melt results in a Lorentz force term in the Navier-Stokes equations that govern the melt flow. The Lorentz force is anisotropic but acts to reduce the motion in the melt. Over the past several decades, a number of studies have examined the damping effect that a magnetic field has on melt motion [l-51. Reviews have been written concerning the application of magnetic fields to Czochralski [6],Float-zone [7] and Bridgman [8] crystal growth processes. In this paper we are primarily concerned with semiconductor crystal growth by the Bridgman method. A commonly used measure of convection in a melt growth system is the non-dimensional thermal Rayleigh number gmL4 RaT = -, VK

where g is the acceleration due to gravity, p i s the volumetric thermal expansion coefficient, VT is the characteristic temperature gradient, L is the characteristic dimension, v is the kinematic viscosity and K is the thermal diffusivity. Whereas Ra7 determines the degree of convection resulting from thermal gradients, Ras determines the degree of convection resulting from solutal concentration gradients and is given by Ra, =-

gWL4

VD ’

vc

where yis the solutal density coefficient, is the concentration gradient, and D is the mass diffusion coefficient. RUTmust be considered in any melt growth process and Ras must be considered for the growth of non-dilute alloys. In some semiconductor growth systems, Ras > RUT, and solutal gradients can drive a larger degree of convection than thermal gradients. The effectiveness of a magnetic field in damping convection in a dilute alloy, such as gallium-doped germanium, can be ascertained by examining the axial distribution of dopant in the grown crystal. The two limiting cases of convective mixing are termed “complete mixing” and “diffusion-limited” growth. The ratio of dopant concentration along the crystal interface to the dopant concentration in the melt is quantified by the segregation coefficient k.

180

For k < 1, there will be more dopant in the melt than in the adjacent solid. In the complete mixing case, strong circulation in the melt assures that the dopant concentration in the melt is uniform at each instant of time. As growth progresses, the dopant concentration in the melt also increases. In the diffusionlimited case, there is no convection and mass transport only proceeds by diffusion. Rejected dopant cannot diffuse far from the interface, and a massdiffusion boundary layer will be established adjacent to the interface. This layer will have a characteristic thickness of DlV, where D is the diffusion coefficient of the dopant in the semiconductor melt and V is the growth rate. After an initial transient region at the beginning of growth, the dopant concentration in the solid will equal the dopant concentration in the melt prior to growth. There will be a final transient of concentration in the solid when the diffusion boundary layer is less than or equal to the height of the remaining melt. Figure 1 is a plot of the gallium concentration versus crystal length for a gallium-doped germanium crystal grown by the vertical Bridgman method. The growth rate was 8 p d s and the sample diameter was 8 mm. Further experimental details of the processing conditions have been reported previously [9]. The gallium concentration was obtained from resistivity measurements using the 4-point probe technique. The experimental data are compared to the complete mixing model [ 101. Clearly, there is sufficient mixing in the melt to prevent the buildup of a mass diffusion boundary layer at the solidification front. Figure 2 is a plot of gallium concentration versus crystal length for a gallium-doped germanium crystal grown under the same conditions

OTesla Data ----Complete Mixing Model

c

,

0 ._

1

P

c

m C

0

2

4

6 8 Position (cm)

1

0

Figure 1 . Gallium concentration profile in a germanium crystal grown at 0 T.

1

2

181

1.2 I r 3 I Y I

'

I

I

. I

4

5Tesla Data Diffusion M ode1

0

Position (em) Figure 2. Gallium concentration profile in a germanium crystal grown at 5 T.

as the sample shown in Figure 1, with the exception that a 5 T field, collinear with the growing ingot, was applied during growth. The 5 T field reduces convection sufficiently to allow the buildup of a gallium-rich layer ahead of the interface. The solid line is the initial transient behavior predicted from a diffusion model [ I l l . The increase in gallium concentration near the end of growth results when the height of the remaining melt becomes less than the mass diffusion boundary layer. The results are consistent with Matthiesen et al. [12], who reported the diffusion-controlled growth of Ge:Ga in a field of 3 T. Both observing and predicting the effects of static magnetic fields on nondilute alloys is less straightforward than for the dilute alloys. In a non-dilute alloy, such as HgCdTe or GeSi, solute concentration variations in the melt couple with the melt hydrodynamics and must be considered along with purely thermal effects. During the growth of an alloy, the rejected solute is typically more dense than the bulk melt which leads to stable solute stratification near the growing interface. Such growth will result in an axial concentration profile similar to that obtained in diffusion-controlled growth, but where the diffusion coefficient is replaced with an effective diffusion coefficient. The effective diffusion coefficient is determined, in part, by the degree of solute stratification. A static field will, in general, reduce convection resulting from thermal and solutal gradients. But the reduction of convection near the interface can lead to the buildup of a steeper solute gradient than otherwise would be present. This may result in constitutional supercooling and polycrystalline growth. Grains that nucleate ahead of the interface may be lighter than the surrounding melt and

182

float to the top as has been reported for HgCdTe grown in a 5 T field [ 131. There are several key factors that must be considered when assessing the ability of a magnetic field to reduce the convective intensity in the melt. These include the thermophysical properties of the semiconductor (particularly the electrical conductivity), the growth rate and the orientation of the magnetic field with respect to the growth direction. The axial and radial temperature gradients in the melt are also important and are determined by the heat flux, the crystal, melt, and container thermal properties, the ingot size, and the latent heat of solidification. There are several reports of the effect of a magnetic field on segregation in semiconductor alloys grown by the vertical Bridgman method. The axial segregation was slightly reduced and the radial segregation increased during the growth of InGaSb in a transverse magnetic field of 0.4 T [ 141. A transverse field of 0.2 T altered the radial symmetry of the solid-liquid interface in HgCdTe [ 151. When the field was applied, the interface evolved from a radially symmetric concave shape to an off-center concave to a tilted plane, and again to an off-center concave interface. This effect was attributed to the breaking of the axial flow symmetry by the transverse field. The radial segregation was decreased but the axial compositional profile was unaffected during the growth of HgMnTe in a vertical field of 0.3 T [16]. Similar results were found in HgCdTe grown in a vertical field of 5 T [13]. The axial compositional profile was essentially unchanged from the 0 T results but the radial segregation approached the diffusion-limited regime. Finally, the axial composition profile of GeSi grown in a 5 T vertical field was indicative of diffusion-controlled growth [9]. A corresponding sample grown in 0 T had an axial compositional profile indicative of complete mixing. The thermal and solutal Rayleigh numbers are proportional to the acceleration of gravity. Therefore, a significant reduction in the driving force for convection can be achieved by simply processing in an earth-orbiting spacecraft. The mean acceleration attainable in a near-earth orbit is of the order of go. This value is determined by the atmospheric drag and gravity field gradient seen by spacecraft. A number of semiconductor crystal growth experiments have been conducted in a microgravity environment and, as early as the Apollo-Soyuz mission. It was shown that Ge:Ga could be grown under diffusion-controlled conditions [ 171. There are, however, limitations to microgravity processing that depend, in part, on the dimensions of the sample and the heat transfer conditions of the growth system. It is not usually possible to control the direction of the residual acceleration vector with respect to the growth axis. This is of concern

183

because the compositional uniformity is very sensitive to the orientation of the residual acceleration vector [ 181. It has been calculated that transverse accelerations less than lo-’ go can produce significant solute redistribution in typical systems with a characteristic dimension of about 1 cm [ 191. The effect of alignment was demonstrated experimentally on a 1994 Spacelab mission [20]. HgCdTe grown in a stable condition with the axial acceleration vector in the direction of the solid to the liquid showed a radially uniform composition profile. In contrast, the radial composition was asymmetric for HgCdTe grown in an unstable condition with the axial acceleration vector in the direction of the liquid to the solid. On an orbiting spacecraft, there also exist non-steady acceleration components over a large frequency range, which are typically orders of magnitude larger than the mean zero-frequency acceleration. Such periodic accelerations have been termed “g-jitter” and have typical amplitudes and lo-’ go. The effect of a single pulse on convection can extend between for times on the order of 1000 seconds [18]. Combining the effects of microgravity and static magnetic fields has been seriously considered [21-231. A relatively modest field of 0.2 T can affect convection driven by both the residual zero-frequenc y acceleration and by gjitters. In such a case, Ma and Walker [24] found that buoyant convection driven by axial and transverse spikes of acceleration decayed to one percent of their initial magnitudes in 3 and 14 seconds, respectively. Without a magnetic field, the detrimental melt motion induced by such spikes can last for several minutes. In light of the potential benefits of a static magnetic field in microgravity, the Marshall Space Flight Center began a program to develop magnet requirements and build flight prototype crystal growth furnaces with magnetic damping capability. Development of a permanent magnet system was done in collaboration with‘h4icrogravity Systems Inc., and the National High Magnetic Field Laboratory constructed an electromagnet with a Bitter coil design. Some properties of these magnets are given in Table 1. The permanent magnet consisted of an assembly of neodymium iron (NdFe) oxide magnets. An advantage of a permanent magnet for space flight operations is that it requires no power to operate. A disadvantage is that it cannot be turned off. This can present a potential safety hazard, as a crew would have to transport and install it onboard. It may also be difficult or impossible to remove the magnet from the furnace while onboard the spacecraft. Thus, it would not be possible to compare experiments done in the magnet to 0 T results. The NHMFL electromagnet consisted of four split coils electrically in series and hydraulically in parallel. A steel vessel was used to reduce the fringe field and was nickel-plated and Teflon

184

coated to minimize interaction with the water-cooling system. Figure 3 shows the magnet assembled with the Bridgman furnace. During operation, the sample inside the furnace and the magnet are in a fixed position and the furnace translates. A disadvantage of an electromagnet is that its operation requires power, a limited resource on a spacecraft. The maximum power expected to be available to the magnet is 3 kW, for which a central field of 0.14 T is obtained. Advantages of the electromagnet are that the field can not only be turned off but also varied both between and during experiments. Table 1. Properties of Potential Magnets for Microgravity Operations PERMANENT MAGNET BITTER MAGNET 79kg 51kg Magnet Mass 140mm ID 184mm 180mm OD 324mm 260mm Length 239mm 0.08 T Maximum Field 0.14T@3kW

Figure 3. Electromagnet built by the NHMFL for microgravity operations integrated with the furnace assembly.

185

3. Rotating Magnetic Fields A rotating magnetic field (Rh4F) is used to induce controlled stirring in an electrically conducting melt. The field can be generated by placing a liquid metal inside an induction motor, with the liquid replacing the rotor. In practice, the RMF can often be approximated as having only horizontal components which are perpendicular to the liquid metal column and which rotate with frequency w. An RMF may be described by [25]

B(r, p,t ) = B0[sin(p - @e, - cos(p- w)e,],

(3)

where Bo is the magnetic field amplitude and (r, p, z ) are standard cylindrical coordinates. A schematic diagram showing the applied magnetic field and the induced flow is shown in Figure 4. The precise nature of the induced flows depend on the geometry of the system, the boundary conditions, the Lorentz force distribution, and whether the flow is in the laminar or turbulent state. In general, an RMF applied in the manner described above will cause a primary azimuthal swirling flow. If the cylinder is truncated at one or both ends, a secondary meridional flow will result. The primary azimuthal flow induces a centripetal acceleration that is balanced by a radial pressure gradient [26]. The axial variation in pressure will drive a secondary meridional flow in the (r, Z ) plane.

Figure 4. Applied magnetic field and basic flow induced by an Rh4F.

Although RMF's have been used in the continuous-casting industry for several decades, their application to crystal growth is more recent. The increasing use of the technique for this application is driven by the potential benefits that RMFs can have on the melt crystallization process. These include homogenization of the melt temperature and concentration distribution [27],

186

control of the liquid-solid interface [28], and a reduction in growth-related defects [29, 301. Mixing of the melt prior to solidification is particularly important for space flight applications. There, gravity-induced mixing is unavailable and resource limitations make other mixing mechanisms less tenable. A RMF forcibly stirs the melt and creates a thin shear layer at the solidification front. Varying the strength and frequency of the applied field can control this layer. Secondary meridional flow can enhance mass transport and allow for faster growth rates than are possible with diffusion-controlled growth. There have been a number of published works that study the flow driven by a RMF in cylindrical geometries that are pertinent to bulk semiconductor crystal growth. Dold and Benz [31] give a recent overview. The nondimensional magnetic Taylor number scales the force that drives the convection induced by a RMF and is given as Tm=-

Bt R'ow

2pv' '

(4)

where R is the cylinder radius, CT is the electrical conductivity and p is the density. The magnetic Taylor number is related to the magnetic rotational Reynolds number and Hartmann number by Tm = Ha2Re, where r

and where is the permeability of free space. The factor 4112 in the Hartmann number originates from the time-averaged RMF. The benefits of applying an RMF during crystal growth can be compromised if the induced flow in the melt becomes unstable. Unstable or time-dependent flow can lead to striations in the grown crystal that will degrade the electrical properties. Even in a system of uniform temperature, an RMF will induce time-dependent oscillatory flow at a critical magnetic Taylor number Tm".A number of authors have sought to calculate this critical value and to determine the nature of the flow at the onset of convection. Grants and Gerbeth [32] reviewed these studies and determined that Tm" = 1.6 x lo5 in a cylinder with an aspect ratio (heighvdiameter) of unity. In any real crystal growth process, temperature gradients will be present and it is necessary to understand how an RMF will affect buoyancy-driven convection and stability. For example, in the zone melting or traveling heater

187

method (THM) process, the melt temperature has a maximum at a circumference near the middle of the melt, and the melt temperature decreases in both axial directions from this circumference to the solidification temperature at both the crystal-melt interface and the feed-rod-melt interface. Thus, roughly the top half of the melt is unstably stratified. In the Bridgman crystal growth process, radial temperature gradients are essentially unavoidable. An experimental study was undertaken at the Marshall Space Flight Center (MSFC) to elucidate the effect that an Rh4F has on buoyancy-induced convection. An experimental cylindrical cell with adiabatic sidewalls was filled with liquid gallium and placed inside an RMF operating at 60 Hz. The liquid gallium has electrical and thermophysical properties similar to semiconductor melts at growth temperatures. The geometry was such that the variations in the magnetic field strength were less than 4% over the vertical and horizontal dimensions of the cell. This was to ensure that no significant z-component of the Lorentz force would exist and that the experimental results would be more readily comparable to theoretical predictions. The temperatures of the top and bottom of the cell were controlled by fluid recirculating from constant temperature baths. Thin (1.5 mm) copper disks at the top and bottom of the cell helped to produce radially uniform temperatures at these surfaces. Thermistors were immersed in the gallium a few millimeters from the cell wall and were used to detect the onset of thermal oscillations. Further experimental details can be found in Volz and Mazuruk [33]. Figure 5 is a plot of the critical Rayleigh number for the transition to unsteady flow versus Hartmann number. The magnetic field value corresponding to the Hartmann number is shown on the upper horizontal axis. The critical Rayleigh number was measured for Ha = 0 for comparison to previous work. These measurements were made by slowly increasing or decreasing the temperature difference between the top and bottom of the cell and observing temperature changes at the thermistors. The measured value of Ra" was 3800, which agreed quite well with numerical modeling results [34] and theoretical predictions 1351. The flow pattern observed at Ra" was a single nonaxisymmetric roll, as expected for a cylinder with an aspect ratio of 1. As the RMF is turned on and Ha increases, RaCincreases and reaches a maximum value which is approximately 10 times larger than its value for Ha = 0. When Ha reaches a value of about 0.95, the RMF itself induces unsteady flow. It is clear that an RMF can greatly enhance the region of steady flow and suppresses the onset of convection induced by buoyancy. It is of note that the field strength

188

required for flow stabilization in the system is in the range of only a few millitesla. B (mT)

6 10'

0

0.4

0.3

1.2

1.6

2

2.4

2.8

Unsteady Flocrv I

..**

**

m..

SeadyFlow Ilo4

.*

1

Hartmann rdumber

Figure 5. Stability diagram for a liquid gallium cell subjected to an RMF.

Experimental crystal growth results utilizing RMFs encompass the Bridgman, float-zone, Czochralski, and travelling heater methods. Dold and Benz [36] grew gallium-doped germanium by the top-seeded vertical Bridgman technique. Thus, unsteady flow was always present, driven by thermal buoyancy. Application of an RMF reduced the magnitude and increased the frequency of striations. The solid-liquid interface also became flatter. For Ge:Ga crystals grown by the normal (hotter on top) Bridgman method [36], the interface also became flatter. In the float-zone technique, the surface-tensiondriven (Marangoni) convection is usually above the critical Marangoni number resulting in time-dependent flow. The main result for float-zone is that a RMF has reduced the intensity of striations in Si [37] and in GaAs [38] when the strength of the magnetic field was increased sufficiently to dominate the flow. Dold et al. [37] also reported that the radial dopant distribution became more symmetric when a RMF was applied. A review of the application of RMFs during Czochralski growth was given by Spitzer [39]. In an early experimental result, Briickner and Schwerdtfeger [40]used an RMF to induce stirring instead of mechanical rotation. The specially modified Czochralski apparatus was used to grow crystals of copper, germanium and silicon. The authors concluded that the electromagnetic stirring was far more elegant than conventional mechanical

189

stimng. Also, by changing the electrical layout of the system, it could also produce more complicated stirring patterns that might be advantageous. An RMF has been applied during the THh4 growth of HgCdTe [41] and GaP [42]. It was also used on the PHOTON 7 and PHOTON 8 microgravity missions [29, 301 for the growth of CdTe and CdTeSe crystals. On the microgravity missions, an RMF with a magnetic field strength of 2 mT and a frequency of 400 Hi was applied. The RMF was turned off partway through the crystal growth in order to compare samples grown with and without the RMF applied. The authors state that the magnetic field strength and frequency were chosen in order to produce a laminar fluid flow but no time-dependent convection [43]. The properties of the grown crystals clearly demonstrate the advantageous of the forced convection induced by the RMF. For the CdTe crystals, the portion grown with the RMF applied had a more homogenous distribution of p products and resistivity. For the CdTeSe crystals, the portion grown in the RMF had increased resistivity, fewer deep level defects, and were better suited as high energy radiation detectors. A series of Ge:Ga crystal growth experiments have been conducted at MSFC using the vertical Bridgman crystal growth method. The crystals were grown using the thermally stable configuration (hotter on top) and 60 Hz RMFs of various magnetic field strengths were applied. The crystals were grown in pyrolitic boron nitride ampoules on single crystal Ge seeds. They were 12 mm in diameter and approximately 80 mm long. One goal of the experiments was to determine Tmcas a function of aspect ratio. Barz et al. [44] predicted that as the aspect ratio increased, Tm" would decrease. They determined that solid layers adjacent to the melt have a stabilizing influence on the flow and their influence becomes smaller for higher aspect ratios. Figure 6 shows part of an etched surface of an axial crystal slice. The gallium concentration of the starting material was 6.8 x 10l8 atoms/cm3, the crystal orientation was [loo], furnace translation velocity was 5 mm/hr, and the growth direction was from left to right in the figure. The applied magnetic field was 3.9 mT during this portion of the growth, which corresponds to Tm = 2.6 x lo4. On the left side of the figure, Tm > Tmc.The striation spacing corresponds to flow oscillations with a period of 20 seconds, and this period increases as Tm approaches Tm".On the right side of the figure, the striations vanish because Tm < Tm" as a result of the decreased aspect ratio. The point here is that, in addition to thermal gradients and magnet intensity, the geometry must also be considered when selecting parameters to achieve the benefits of an RMF without triggering deleterious instabilities.

190

Figure 6 . Transition from unstable to stable flow as the aspect ratio is decreased. The growth direction is from the left to the right.

4.

Conclusions

Static magnetic fields do dampen convective flows and can lead to diffusionlimited growth and reduced defect formation. Yet the initial enthusiasm for the technique has been tempered as its limitations have become better understood. Magnetic field strengths on the order of a few tesla are usually required, and they increase as the sample dimensions increase. For some combinations of material systems and dimensions, currently available magnetic field strengths are insufficient [3]. A strong static field is not always beneficial. In some circumstances, it may increase radial segregation and the increasing solute gradient in the melt at the interface can lead to constitutional supercooling and polycrystalline growth. In addition, there can exist deleterious thermoelectric magnetohydrodynamic effects at very high field strengths [45]. RMFs offer certain advantages when compared to static magnetic fields. Most remarkable, perhaps, is that magnetic field strengths of only a few millitesla are required, in comparison to a few tesla for static fields. RMFs also offer greater control over mass transport, as the frequency and strength of the field can be adjusted to provide the optimal flow. But the application of an RMF during the Bridgman crystal growth process will not result in the axial composition profile indicative of diffusion-controlled growth that is often desired. There is also an upper limit to the maximum RMF field strength that can be applied as the RMF, itself, will induce instabilities at Tm".In summary, an understanding of the effects of static and rotating magnetic fields on heat and mass transport is essential in the choice

191

of which field is most beneficial for a given semiconductor alloy system and crystal growth process.

Acknowledgments The author is indebted to scientific collaborations with the following individuals: S.D. Cobb, D.C. Gillies, S.L. Lehoczky, K. Mazuruk, C.-H. Su and F.R. Szofran at MSFC, A. Croll and P. Dold at the University of Freiburg, J.S. Walker at the University of Illinois, S. Motakef at Cape Simulations, and M. Schweizer, currently at Schott Glas. This work was supported by the Office of Biological and Physical Research of the National Aeronautics and Space Administration.

References 1. Utech, H.P. and Flemings, M.C., Elimination of Solute Banding in Indium Antimonide Crystals by Growth in a Magnetic Field, Journal of Applied Physics 37 (1966), pp. 2021-2024. 2. Oreper, G.M. and Szekely, J., The Effect of a Magnetic Field on Transport Phenomena in a Bridgman-Stockbarger Crystal Growth, Journal of Crystal Growth 67 (1984), pp. 405-419. 3. Motakef, S., Magnetic Field Elimination of Convective Interference with Segregation During Vertical-Bridgman Growth pf Doped Semiconductors, Journal of Crystal Growth 104 (1990), pp. 833-850. 4. Series, R.W. and Hurle, D.T.J., The Use of Magnetic Fields in Semiconductor Crystal Growth, Journal of Crystal Growth 103 (1991), pp. 305-328. 5. Walker, J.S. and Ma, N., Convective Mass Transport During Bulk Growth of Semiconductor Crystals with Steady Magnetic Fields, Annual Review of Heat Transfer 12, Begell House, (2001), pp. 223-263. 6. Hurle, D.T.J. and Series, R.W., Use of a Magnetic Field in Melt Growth, Handbook of Crystal Growth 2a, North-Holland (1994), pp. 261-285. 7. CroII, A. and Benz, K.W., Static Magnetic Fields in Semiconductor FloatingZone Growth, Progress in Crystal Growth and Characterization of Materials 38 (1999), pp. 133-159. 8. Garandet, J.P. and Alboussiere, T., Bridgman Growth: Modelling and Experiments, Progress in Crystal Growth and Characterization of Materials 38 (1999), pp. 73- 132. 9. Volz, M.P., Szofran, F.R., Watring, D.A., Gillies, D.C., Su, C.-H.. and Lehoczky, S.L., Magnetic Damping of Convective Flows During Semiconductor Crystal Growth, High Magnetic Fields: Applications, Generation, Materials, World Scientific, (1997), pp. 57-7 1. lO.Pfann, W.G., Zone Melting, Wiley, (1966).

192 1LSmith, V.G., Tiller, W., and Rutter, J.W., A Mathematical Analysis of Solute Redistribution During Solidification, Canadian Journal of Physics 33 (1955), pp. 723-745. 12.Matthiesen, D.H., Wargo, M.J., Motakef, S., Carlson, D.J., Nakos, J.S., and Witt, A.F., Dopant Segregation during Vertical Bridgman-Stockbarger Growth with Melt Stabilization by Strong Axial Magnetic Fields, Journal of Crystal Growth 85 (1987). pp. 557-560. 13.Watring, D.A. and Lehoczky, S.L., Magnetohydrodynamic Damping of Convection during Vertical Bridgman-Stockbarger Growth of HgCdTe, Journal of Crystal Growth 167 (1996), pp. 478-487. 14.Sen, S . , Lxfever, R.A., and Wilcox, W.R., Influence of Magnetic Field on Vertical Bridgman-Stockbarger Growth of In,Gal.,Sb, Journal of Crystal Growth 43 (1978), pp. 526-530. 15.Su, C.-H., Lehoczky, S.L., and Szofran, F.R., Directional Solidification of HgCdTe and HgZnTe in a Transverse Magnetic Field, Journal of Crystal Growth 109 (1991), pp. 392-400. 16.Becla, P., Han, J.C., and Motakef, S., Application of Strong Vertical Magnetic Fields to Growth of 11-VI Pseudo-Binary Alloys: HgMnTe, Journal of Crystal Growth 121 (1992), pp. 394-398. 17.Witt, A.F., Gatos, H.C., Lichtensteiger, M., and Hermann, C.J., Crystal Growth and Segregation under Zero Gravity, Journal of the Electrochemical Society 125 (1978), pp. 1832-1840. 18.Alexander, J.I.D., Ouazzani, J., and Rosenberger, F., Analysis of the Low Gravity Tolerance of Bridgman-Stockbarger Crystal Growth: I. Steady and Impulse Accelerations, Journal of Crystal Growth 97 (1989), pp. 285-302. 19.Naumann, R.J., Modeling Flows and Solute Redistribution Resulting from Small Transverse Accelerations in Bridgman Growth, Journal of Crystal Growth 142 (1994), pp. 253-267. 20.Gillies, D.C., Lehoczky, S.L., Szofran, F.R., Watring, D.A., Alexander, H.A., and Jerman, G.A., Effect of Residual Acceleration during Microgravity Directional Solidification of Mercury Cadmium Telluride on the USMP-2 Mission, Journal of Crystal Growth 174 (1997), pp. 101-107. 21.Naumann, R.J., Desirable Limits of Acceleration Forces in a Space-Based Materials Processing Facility, NASA Conference Publication 3088 (1986), pp. 4.1- 4.18. 22.Hurle, D.T.J., The Complementary Roles of Magnetic Fields and Microgravity in Controlling Convection in Crystal Growth and in Materials Processing, Microgravity News from ESA 7 (1994), pp. 2-3. 23.Muller, G. and Friedrich, J., The Influence of Steady and Alternating Magnetic Fields in Crystal Growth and Alloy Solidification: Industrial Importance, Current Industrial R&D Topics, Links to Micro-Gravity Research, Proceedings of the 2nd European Symposium on the Utilisation of the International Space Station, ESA SP-433 (1999), pp. 309-3 14.

193 24.Ma, N. and Walker, J.S., Magnetic Damping of Buoyant Convection during Semiconductor Crystal Growth in Microgravity: Spikes of Residual Acceleration, Physics of Fluids 9, (1997), pp. 1182-1187. 25.Volz, M.P. and Mazuruk, K., Flow Transitions in a Rotating Magnetic Field, Experiments in Fluids 20, (1996), pp. 454-459. 26.Davidson, P.A., and Hunt, J.C.R., Swirling Flow in a Liquid-Metal Column Generated by a Rotating Magnetic Field, Journal of Fluid Mechanics 185 (1987), pp. 67-106. 27.Gelfgat, Y.M., Electromagnetic Field Applications in the Processes of Single Crystal Growth under Microgravity, 45th Congress of the International Astronautical Federation, Jerusalem, Israel IAF-94-J.4.251 (1994). 28. Senchenkov, A.S., Friazinov and Zabelina, M.P., Mathematical Modelling of Convection during Crystal Growth by the THM,Proceedings of the First International Symposium on Hydromechanics and Heamass Transfer in Microgravity, Perm-Moscow, Gordon and Breach (199 1), pp. 455-459. 29. Salk, M., Fiederle, M., Benz, K.W., Senchenkov, A.S., Egorov, A.V., and Matioukhin, D.G., CdTe and CdTeo.$eo,l Crystals Grown by the Travelling Heater Method using a Rotating Magnetic Field, Journal of Crystal Growth 138, (1994), pp. 161-167. 30.Fiederle, M., Eiche, C., Joerger, W., Salk, M., Senchenkov, AS., Egorov, A.V., Ebling, D.G., and Benz, K.W., Radiation Detector Properties of CdTeo,9S~,l:C1 Crystals Grown under Microgravity in a Rotating Magnetic Field, Journal of Crystal Growth 166 (1996), pp. 256-260. 31.Dold, P. and Benz, K.W., Rotating Magnetic Fields: Fluid Flow and Crystal Growth Applications, Progress in Crystal Growth and Characterization of Materials 38, (1999) pp. 7-38. 32. Grants, I. and Gerbeth, G., Stability of Axially Symmetric Flow Driven by a Rotating Magnetic Field in a Cylindrical Cavity, Journal of Fluid Mechanics 431 (2001), pp. 407-426. 33.Volz, M.P. and Mazuruk, K., An Experimental Study of the Influence of a Rotating Magnetic Field on Rayleigh-BCnard Convection,” Journal of Fluid Mechanics 444 (2001), pp. 79-98. 34. Neumann, G., Three-Dimensional Numerical Simulation of BuoyancyDriven Convection in Vertical Cylinders Heated from Below, Journal of Fluid Mechanics 214, (1990), pp. 559-578. 35.Buel1, J.C. and Catton, I., The Effect of Wall Conduction on the Stability of a Fluid in a Right Circular Cylinder Heated from Below, Journal of Heat Transfer 105 (1983), pp.255-260. 36.Dold, P. and Benz, K.W., Modification of Fluid Flow and Heat Transport in Vertical Bridgman Configurations by Rotating Magnetic Fields, Crystal Research and Technology 32 (1997), pp. 51-60.

194

37.Dold, P., Croll, A., Lichtensteiger, M., Kaiser, Th., and Benz, K.W., Hoating Zone Growth of Silicon in Magnetic Fields: IV. Rotating Magnetic Fields,” Journal of Crystal Growth 231, (2001), pp. 95-106. 38.Fischer, B., Friedrich, J., Weimann, H., and Muller, G., The Use of TimeDependent Magnetic Fields for Control of Convective Flows in Melt Growth Configurations, Journal of Crystal Growth 198/199 (1999), pp. 170-175. 39. Spitzer, K.-H., Application of Rotating Magnetic Fields in Czochralski Crystal Growth, Progress in Crystal Growth and Characterization of Materials 38 (1999), pp. 39-58. 40.Bruckner, F.-U. and Schwerdtfeger, K., Single Crystal Growth with the Czochralski Method Involving Rotational Electromagnetic Stirring of the Melt, Journal of Crystal Growth 139 (1994), pp. 351-356. 41.Senchenkov, A.S., Barmin, I.V., Tomson, A.S., and Krapukhin, V.V., Seedless THM Growth of Cd,Hgl-,Te (x =: 0.2) Single Crystals Within Rotating Magnetic Field, Journal of Crystal Growth 197 (1999), pp. 552556. 42.Inatomi, Y., Takada, A., and Kuribayashi, K., Morphological Change of Semiconductor Growth Interface from Solution in a Magnetic Field, Journal of Crystal Growth 198/199 (1999), pp. 176-181. 43.Barmin, I.V., Egerov, A.V., Filatov, I.G., Senchenkov, A.S., Benz, K.W., Lexow, B., Salk, M., Hofmann, P., Sickinger, P., Gelfgat, Y.M., Sorkin, M.Z., and Matioukhin, D.G., CdTe Crystal Growth by THM with a Rotating Magnetic Field”, Proceedings of the VIIIth European Symposium on Materials and Fluid Sciences in Microgravity, Brussels, Belgium, (1992), pp.3 15-320. 44.Barz, R.U., Gerbeth, G., Wunderwald, U., Buhrig, E., and Gelfgat, Y.M., Modelling of the Isothermal Melt Flow due to Rotating Magnetic Fields in Crystal Growth, Journal of Crystal Growth 180, (1997), pp. 410-421. 45.Khine, Y.Y. and Walker, J.S., Thermoelectric Magnetohydrodynamic Effects during Bridgman Semiconductor Crystal Growth with a Uniform Axial Magnetic Field: Large Hartmann Number Solution, in Transfer Phenomena in Magnetohydrodyanamics and Electroconducting Flows, Kluwer Academic Publishers (1999), pp. 269-278.

Magnetic Separation


REMOVAL SYSTEM OF ARSENIC FROM GEOTHERMAL WATER BY MAGNETIC SEPARATION TECHNOLOGY WITH A SUPERCONDUCTINGMAGNET H. OKADA, K. MITSUHASHI', T. OHARA*,H. WADA' Y. KUDOH', H. NAKAZAWA', A. CHIBA' Iwate Industrial Promotion Center 3-35-2, lioka-shinden, Morioka, 020-0852, Japan 'Department of Pure and Applied Sciences, Graduate School of Pure and Applied Sciences, University of Tsukuba, 1-1-1, Tennoudai, Tsukuba, lbaraki, 305-8573, Japan 'Tsukuba Magnet Laboratory, National Institutefor Materials Science. 3-13 Sakura, Tsukuba, Ibaraki, 305-0003, Japan 'Faculty of Engineering, Iwate University, 3-4-5, Ueda, Morioka, 020-8551, Japan We have developed an arsenic removal system from geothermal water by High Gradient Magnetic Separation (HGMS), thus using a superconducting magnet to supply hot water for public use. We reduced arsenic to approximately 0.02mgL (less than the effluence standard of O.lmg/L and slightly larger than the environmental standard of 0.OlmgL in Japan) by using an experimental plant, and we investigated the optimum HGMS system using a superconducting magnet. The plant consists of a pretreatment system that adds extra magnetization to arsenic by chemical reaction, and a HGMS device that uses a superconducting magnet that extracts magnetized arsenic from geothermal water. In this paper we present the experimental results.

1. Introduction Geothermal power is a promising candidate as a future energy source but thermal energy from wells is currently not used effectively [ 11. Most wells spout vapor and geothermal water and, after separating the vapor and hot water, the vapor is supplied to a generator and the hot water (>lOO°C) is returned underground. There are two geothermal power plants at Kakkonda and Matsukawa in the Iwate prefecture, Japan. The Kakkonda geothermal power plant has 30 and 50MW generators and returns 3000 t/h of geothermal water at 140°C underground through reinjection wells. The remainder of the geothermal water (500 t/h) is supplied to a heat-exchanging plant to heat river water (50 t/h) for greenhouses, an indoor swimming pool, and heating systems used for public facilities. This suggests that a direct supply of geothermal water increases usable amounts of hot water and uses for this water. Geothermal water from wells in Kakkonda contains 3.4 mg/L of arsenic, which exceeds both the environmental standard of 0.01 m g L and effluent standard of 0.1 mg/L in Japan. Arsenic concentration must be less than 0.1 mg/L 197

198

for public use water. We are developing an arsenic removal system using magnetic separation. Because arsenic forms weakly magnetized matter in geothermal water, it is difficult to remove arsenic by conventional magnetic separation. As a result, it is necessary to attach extra magnetization to arsenic before using magnetic separation and/or develop new magnetic separation techniques to remove weakly magnetized matter. Our study had two parts: pretreatment to attach extra magnetization and subsequent magnetic separation. We used iron (III) hydroxide (Fe(OH)3) to increase the magnetization of arsenic and we used high-gradient magnetic separation (HGMS) [l-31 to remove the magnetized arsenic. We present our recent works on removal efficiency of the system.

2. Experiments 2.1. Geothermal Water in Kakkonda We built a small experimental plant in Kakkonda in the winter of 2000. The geothermal water is supplied from the Kakkonda geothermal power plant, 2 km from our experimental plant. The geothermal water supplied from the power plant has a temperature and pressure of 14OOC and 2.5 kgf/cm2, respectively. Geothermal water was released into and stored in a tank located in the entrance of our experimental plant. Characteristics of the stored geothermal water used for our experiments are found in Table 1. As the geothermal water in Kakkonda contains 3.4 mg/L of arsenic, the system must be able to remove around 99% of the arsenic in order to pass the effluent standard. Table I Characteristics of geothermal water in Kakkonda used for experiments.

PH Temperature Pressure Concentration of As

8.2 98°C 1 atm. 3.4 m e n

2.2. Pretreatment Method In a recent paper, we presented our previous study on pretreatment methods to enhance magnetization of arsenic [2,3]. According to the study, we expect that the coprecipitation method of Fe(II1) hydroxide is the most influential to make suitable magnetized flocks for magnetic separation.

199

Geothermal

Iron (111)

I

Magnetic separation

I

Figure 1. Diagram of the pretreatment process.

Figure 1 shows a pretreatment process. Usually, arsenic forms arsenic trioxide (AsO;-) in geothermal water. Because arsenic trioxide was difficult to separate, our first step was to change the arsenic trioxide to arsenic acid (As021 by adding hydrogen peroxide to the geothermal water. Adding iron (111) sulfate (Fe2(S04),) and neutralizing it to pH=4 made flocks of iron (III) hydroxide (Fe(OH),) that adsorb the arsenic compound. Magnetic separation extracted the flocks-adsorbing arsenic from the geothermal water. The treated geothermal water had a pH of 4 and was hotter than 90°C. The magnetization curve of the flocks is shown in Figure 2. The flocks consisted of iron hydroxide and other matters and are desiccated to measure the magnetization. The flocks’ magnetic susceptibility was around lo-, corresponding to typical values of paramagnetic materials.

200

Magnetic F i e l d ( T )

Figure 2. Magnetization curve of flocks at room temperature. The sample was desiccated

2.3. High Gradient Magnetic Separator The HGMS system has a superconducting magnet applying a magnetic field to meshes woven out of thin ferromagnetic wires that are inserted in the flow 121. The wires are magnetized with an externally applied magnetic field, causing a sharp change of magnetic field around each wire. Subsequently, a strong magnetic force forms around the wires. The force is strong enough to extract and capture flocks (fine paramagnetic particles) in the flow to the wires [2]. If we turn off the superconducting magnet, the magnetic force disappears and captured materials are released from the wires. An ordinary magnetic separation system has a washer to rinse meshes when capture ability decreases due to build-up of captured materials on wire surfaces. The ordinary washers rinse materials captured on meshes when the magnetic field is turned off.

20 1

Figure 3. A photograph of the HGMS reciprocal filter set in the 10 T magnet.

We made a reciprocal HGMS filter [4], as shown in Figure 3. The filter is placed in the room temperature bore of the superconducting magnet. We used a cryo-cooled Nb-Ti and Nb& superconducting magnet manufactured by Japan Superconductor Technology Inc., which can generate a magnetic field to 10 T. In separation experiments, we applied the filter up to 2 T. The diameter of a room temperature bore is 10 cm. The filter is divided into two vertical sections with the same structure. The filter diameter is 50 mm and each division is 60 cm long. The structure of the filter is shown in Figure 4. Each section has 1400 meshes that are woven out of 100 pm diameter ferromagnetic wires and work individually as a HGMS filter. Geothermal water is supplied from a lower valve and emerges from the upper valve of each section. To rinse, we flushed air from the upper valve to the lower valve after moving the filter to extract the section from the magnet. The procedure is shown in Figure 4. We verified that the magnetic separation system repeatedly recovered the initial capture ability after rinsing the build-up from wires.

202

Drain

f-

Purified water

--+ Purified water

moving up and down CaDture

From the settling

From the settling tank Flush

Rinse

Drain

Figure 4. Schematic explanation of capture and rinsing process of the reciprocal filter. We repeated the procedure for filtering. Arrows show the direction of geothermal water and flushing air.

Figure 5 is a photograph of the latest experimental plant in Kakkonda in which the pretreatment process and subsequent magnetic separation are accomplished. In the first tank, located in the center of the photograph, geothermal water was supplied from the power plant, released into 1 atm, and stored. Main chemical reactions occurred in the first and the second tank next to the first one and behind the superconducting magnet. The treated geothermal water was pumped from the tank to the filter at a given rate. We collected water samples before and after the superconducting magnet to examine the magnetic separation efficiency.

203

Figure 5. Photograph of the experimental setup in Kakkonda. There is a superconducting magnet and reciprocal filter on one side. Three tanks on the other side are used for pretreatment. White arrows show the water flow.

3. Experimental Results Figure 6 compares arsenic concentrations before and after magnetic separation (decreased by the pretreatment and further decreased by magnetic separation). This means that large flocks and most of the arsenic were deposited in the settling tank and magnetic separation removed tine flocks. As the flow velocity in the filter decreased, efficiency of the magnetic separation increased. Arsenic concentration in a few samples was less than 0.01 m g L So far, the results show that the final arsenic concentration can be reduced below the standard for discharge of 0.1 m g L and close to the environmental standard of 0.01 m g L We will proceed with experiments to determine optimal conditions to pass the environmental standard codes.

204

-

1 to

\

I

v

m

._ L

w a

-

.-

0.1

L

L

m

L w

m

c

.z 4-

0.01

L

c c 0 0 v1

0.001

u

o

As concentration before filtering (mg/L)

Figure 6. Comparing As concentration before and after filtering at 2 T. “ X shows concentration at 12.6 c d s of the flow velocity, 0 at 8.4 c d s and 0 at 4.2 c d s . The diagonal solid line shows the boundary of efficiency of the magnetic filter.

4.

Summary

We established that arsenic can be removed from geothermal water by a magnetic separation system using a reciprocal filter and superconducting magnet. We demonstrated that our system always reduces arsenic from 3.4 mg/L to below the effluent standard and purifies a large amount of water at high speed. The reciprocal filter rinses the deposit without loss of time, but a very strong magnetic force acts on the filter and the reaction acts on the superconducting magnet. According to our study, an actual plant needs 30 cm 0 filters to supply 30 t/h of hot water. Magnetic force is a main problem for a larger magnetic separator. We continue to study decontamination of the geothermal water by adopting magnetic separation as a plan for practical plants.

Acknowledgment This work is a part of the Collaboration of Regional Entities for the advancement of Technological Excellence in Iwate. We are grateful to the staffs of the Iwate prefectural office and Geothermal Engineering Co. Ltd. for their kind help.

205

References 1. Umeno, J. and Iwanaga, T., “A Study on the Abatement Technology of the Harmful Chemical Components in Geothermal Hot Water,” Proceedings of the 20th New Zealand Geothermal Workshop 1988, pp. 209-213. 2. Okada, H., Tada, T., Chiba, A., Mitsuhashi, K., Ohara, T., and Wada, H., “High gradient magnetic separation for weakly magnetized fine particles,” to be published in IEEE Trans. Appl. Supercond, 3 . Chiba, A., Okada, H., Tada, T., Kudoh, H., Nakazawa, H., Mitsuhashi, K., Ohara, T., and Wada, H., “Removal of arsenic from geothermal water by high gradient magnetic separation,” to be published in IEEE Trans. Appl. Supercond. 4. Yan, L.G., Song, S.S., Yi, C.L., Ye, Z.X., Nan, H.L., Tu, G.B., Li, X.S., Li, X.Y., Zheng, J.X., “Laboratory Test of an Industrial Superconducting Magnetic Separator for Kaolin Clay Purification,” IEEE Trans. Mug. 30, (July 1994), pp. 2499-2502.

MAGNETICALLY ENHANCED SOLID-LIQUID SEPARATION

’,

C.M. REY K.KELLER’, B.FUCHS3 DuPont, Chestnut Run Plaza, Wilmington, DE 19880-0708 DuPont, Experimental Station, Wilmington, DE 19880-0304 Universitat Karlsruhe, Institut f i r Mechanische Verfahrenstechnik und Mechanik 76128 Karlsruhe, Germany I

DuPont is developing an entirely new method of solid-liquid filtration involving the use of magnetic fields and magnetic field gradients. The new hybrid process, entitled Magnetically Enhanced Solid-Liquid Separation (MESLS), is designed to improve the de-watering kinetics and reduce the residual moisture content of solid particulates mechanically separated from liquid slurries. Gravitation, pressure, temperature, centrifugation, and fluid dynamics have dictated traditional solid-liquid separation for the past 50 years. The introduction of an external field (is. the magnetic field) offers the promise to manipulate particle behavior in an entirely new manner, which leads to increased process efficiency. Traditional solid-liquid separation typically consists of two primary steps. The first is a mechanical step in which the solid particulate is separated from the liquid using e.g. gas pressure through a filter membrane, centrifugation, etc. The second step is a thermal drying process, which is required due to imperfect mechanical separation. The thermal drying process is over 100-200 times less energy efficient than the mechanical step. Since enormous volumes of materials are processed each year, more efficient mechanical solid-liquid separations can be leveraged into dramatic reductions in overall energy consumption by reducing downstream drying requirements have a tremendous impact on energy consumption. Using DuPont’s MESU process, initial test results showed four very important effects of the magnetic field on the solid-liquid filtration process: 1) reduction of the time to reach gas breakthrough, 2) less loss of solid into the filtrate, 3) reduction of the (solids) residual moisture content, and 4) acceleration of the de-watering kinetics. These test results and their potential impact on future commercial solid-liquid filtration is discussed. New applications can be found in mining, chemical and bioprocesses.

1. Introduction Solid-liquid separation is a classical filtration operation that is used in nearly every industrial solid process. Traditional solid-liquid separation typically consists of two primary steps. The first is a mechanical step in which the solid particulate is separated from the liquid using e.g. gas pressure through a filter membrane, centrifugation, sedimentation, etc. The second step is a thermal drying process, which is required due to imperfect mechanical separation. The thermal drying process is over 100-200 times less energy efficient than the mechanical step. This two-step process has been refined over years of development, and the existing processes are on a very high technical level. In 206

207

order to extend the field of applications for mechanical separation processes, synergetic effects and selectivity have to be exploited. Some well-investigated hybrid processes include: a) steam pressure centrifugation and b) electro-press filtration. Steam pressure centrifugation combines thermal effects with two driving forces for the separation: a pressure force and a mass force. The synergetic effects result in a significant improvement of the de-watering compared to conventional centrifugation or pressure filtration processes [ 11. Another hybrid process is the electro-press filtration. In addition to the traditional mechanical and hydrodynamic forces, the external electric field helps to accelerate the de-watering kinetics. This is especially advantageous for biological products such as poly-saccharides [2,3]. The new Magnetically Enhanced Solid-Liquid Separation (MESLS) shows many similar analogies to the above mentioned electro-press filtration. The new MESLS process, however, has some major advantages over electro-press filtration in that no electrolysis and no additional heating of the suspension occur. The magnetic force, driven by a magnetic field gradient (VB), can significantly influence the particle movement during the filtration process. The magnetic driving force can be seen as an additional degree of freedom for the mechanical separation process. In addition, experimental results have also shown that a homogeneous magnetic field (i.e. no magnetic force) can also have a positive impact on the solid-liquid separation process. The improvements that result from the application of the homogeneous magnetic field are due to structure changes within the suspension or the porous structure of the filter cake. Several authors [4,5] have investigated these structure changes or agglomeration effects within a suspension of magnetic particles. But until now, no one has used these phenomena in order to improve conventional solid-liquid-separation techniques like the cake-filtration process. This work will show how conventional cake filtration can be improved and influenced by a magnetic field in four specific areas: a) reduction of the time to reach gas breakthrough, b) less loss of solids into the filtrate, c) reduction of the (solids) residual moisture (RM) content, and d) acceleration of the de-watering kinetics.

2. Overview To gain a better understanding of the potential of this new and innovative process, a brief overview of the conventional cake-filtration process is presented. A more detailed description can be found elsewhere in the literature

208

(see for example [6] and many other publications from the Institute of Mechanical Process Engineering in Karlsruhe, Germany).

2.1. Conventional Cake-filtration The classical cake-filtration process uses pressure forces in order to dewater and separate a solid-liquid mixture. When the solid-liquid system is subjected to a pressure driven force, the solid particulates are held back by the filter media, which results in the building-up of a filter cake (Figure 1). The filtration effect is not only due to the filter media, but it is also due to the particles that build bridges on top of the pores of the filter media. If a product specific volume concentration of solids is not exceeded, the bridge building cannot take place anymore and the particles flow through the filter media and end up in the filtrate. Loss of product through the filter media is especially deleterious for high value products (e.g. bio-products) and, therefore, must be avoided. Once the particle bridges have been established, the filter-cake height begins to grow. As the filter-cake height increases, the filter-cake resistance correspondingly increases. This means that the flow resistance for the liquid phase through the porous system is also increasing with filtration time.

Figure 1: A simple schematic of the cake-filtrationprocess.

Simple models have been developed that theoretically describe the mechanism for cake building. The models that have been developed are derived from the Darcy equation for the one phase flow through a porous system. The classical cake building expression is given by the linear equation:

209

where t: filtration time,

VL:

filtrate volume, qL: viscosity of liquid phase,

4: specific cake resistance, K: concentration parameter, Ap: pressure difference,

A: filter area, R,: filter media resistance.

To simplify the expression in Eq. (1) we define the two variables (a) and (b) such that:

and

Substitution of Equations (2) and (3) into Equation (1) results in the simplified linear equation given by: t _ - a .V, + b

(4)

VL

In Equation (4), the slope (a) is directly proportional to the specific cake resistance and the y-axis intercept (b) is directly proportional to the filter media resistance. From Equation (4) a simple cake-filtration experiment can now be designed where the filtrate mass is registered online. It delivers the tlVL versus VL diagram, qualitatively shown in Figure 2. A decrease of the slope is directly related to faster cake building kinetics.

Figure 2: A qualitative VVLversus VLdiagram, where the slope (a) is proportional to the specific cake resistance (rc) and the y-axis intercept (b) is proportional to the filter media resistance (Rm).

210

B

Figure 3: The cake building phase with superposed magnetic field compared to the conventional process from left to right, a) zero applied B-field, b) homogeneous B-field, c) B-field gradient.

2.2, Magnetic Field Enhanced Cake-filtration The primary difference between the conventional cake-filtration process and the new h4ESLS process is either the additional magnetic force acting on the solid particulates due to the presence of a magnetic field gradient or the structure changes caused by changes in the particle-particle interactions, which were observed in presence of a homogeneous magnetic field (B-field). Using MESLS, the cake building can be entirely avoided or the structure of the built cake is different compared to the conventional filter-cake (Figure 3). Experimental results show that the independent application of either a magnetic field gradient or a homogeneous magnetic field can result in a positive impact on the dewatering process.

3. Products Four different products of varying degrees of magnetic susceptibility (xdwere studied in this investigation (see Table 1): 1) titanium dioxide (TiOz), 2) hematite (Fez03 FE 601), 3) natural hematite (Fez03 10),and 4)magnetite (Fe30r EO lo). Pertinent physical properties of these products (e.g. density, particle size, and magnetic susceptibility) are given in Table 1. All experiments were performed with de-ionized water as the liquid phase.

+

+

21 1 Table 1: Product properties

4. Experimental Apparatus

The experimental configuration consisted of a modified CUNO filter rigidly mounted at the center of a non-magnetic plastic frame. The plastic frame supporting the CUNO filter was then rigidly mounted within the internal bore of a High Temperature Superconducting (HTS) solenoidal magnet. The CUNO filter could then be placed at various locations along the magnet's central axis. Thus, if the CUNO filter were placed at the axial center of the magnet, the filter and its corresponding product would experience a homogeneous B-field. Likewise, if the CUNO filter were placed at the far ends of the axis of the magnet, the filter and its product would experience a magnetic field gradient. The HTS magnet had a warm bore inner diameter of 20 cm, a height of 30 cm, with a maximum central magnetic field of 3.0 T. Computer modelling was used to calculate the magnetic field distribution of the HTS solenoid coil and hence determined parameters, such as the degree of B-field homogeneity or strength of the B-field gradient. The HTS magnet itself was designed, fabricated, and tested, by DuPont Superconductivity under the financial support of the Department of Energy Superconductivity Partnership Initiative Program.

' See for example www.electricity.doe.gov

212

Figure 4:High-temperature superconducting magnet (DuPont)

The modified filter unit (see Figure 5) that was used for the solid-liquid filtration experiments reported here was originally purchased from the CUNO Filter Systems Company, located in Meriden, CT. DuPont then internally modified the CUNO filter for the MESLS investigation. The filter specifications are listed in Table 2. Table 2: Key Parameters of the CUNO filter unit

213

Figure 5 : The modified CUNO filter cell

5.

Experimental Results

5.1. Changes in Filter-Cake Structure Initial experiments performed on the (ferromagnetic) magnetite (EO 10) product showed that even a relatively small homogeneous B-field could significantly alter the filter-cake structure. For example, at applied B-fields of > 0.03 T, a filter-cake build-up could not be observed due to the destruction of the cake structure itself. This made it impossible to study the fundamentals of the cake building process in the presence of a B-field. Thus, even though the dramatic change in filter-cake structure for the Fe304 product could eventually be exploited to improve the de-watering process [7], a filter-cake build-up was necessary to study the fundamental influence of the B-field on this solid-liquid separation process. Therefore, (non-ferromagnetic) products with lower magnetic susceptibility were introduced.

0.05 T

0.03 T 0.0 T Figure 6: Cake structures of Fe304 (EO 10) at different magnetic flux densities, Ap = 1 6 bar.

214

5.2. Reduction of Solid Breakthrough into the Filtrate Often at the beginning of a cake-filtration process, the filtrate is not clear since a significant portion of the solid particles pass through the filter media. In most industrial mineral separation processes, this unwanted phenomena is tolerated because the value (i.e. cost per unit mass) of the mineral product is relatively low. The economics of the situation is drastically different when processing high value products, where loss of product in the filtrate is unacceptable. For example, for the processing of “functionalized magnetic beads”, which can cost upwards of 100,000 $/kg and carry target products like proteins, DNA plasmids, cells, etc. any loss of product would be economically disastrous. Experiments indicate that for the MESLS process, application of an external B-field results in a reduced loss of product to the filtrate. Shown in Figure 7 is a plot of the (mean) mass of solid in the filtrate for two conditions of applied Bfield for the product Fe304. The mean mass value/error bars in Figure 7 are the calculated averagektandard deviation of the mean (OJ of four separate experimental runs, respectively. Results show an 80% reduction in the solid filtrate mass or solid breakthrough. The reduction is primarily due to agglomeration of the particles or agglomerates, which are now too large to pass through the filter media. The Fe304product used for these experiments is in the same particle size range as many of the mentioned commercial “magnetic beads” now gaining popularity in high value bio-separations. It is possible that the new MESLS technology may improve the efficiency and economics of several bio-separation processes.

215

m

0.14

0 12

Y

Er .E

0.1

0.08

2 0.06 0.04 0.02

0 B = 0.0 T

B = 0.03 T

Figure 7: Mass of solid in filtrate,Fe304(EO lo), Ap = 1.6 bar.

5.3. Reduction of the Residual Moisture The residual moisture (RM) is a term that is used to quantify the amount of liquid that remains trapped inside the filter cake after the de-watering equilibrium is achieved. The lower the RM, the more effective the mechanical treatment. The more effective the mechanical treatment, the lower the energy that will be consumed in the overall solid-liquid separation process. RM is experimentally determined by measuring the mass of liquid trapped, relative to the mass of the remaining wet filter cake. Figure 8 shows a plot of the mean residual moisture content for three conditions of applied external B-field for the paramagnetic product 1 0 . The mean RM value/error bars in Figure 8 are the calculated averageh, of three separate experimental runs, respectively. Results indicate that the relative reduction in RM is nearly 17% for an applied B-field of 1.9 T and that further increases in the applied B-field do not further decrease the RM. Reductions in RM content during the mechanical separation step can have an enormous impact on decreasing the energy consumption, hence increasing the energy efficiency of the solid-liquid separation process. The subsequent step of thermal drying is nearly 100-200 times less energy efficient than the initial mechanical process.

216 40

I

16%

35 30

10 5

0

B 0.0 T

B=1.9T

B

2.73

r

Figure 8: Residual moisture (RM) of Fe203(FE601), Ap = 2.8 bar,

5.4. Acceleration of the De-watering Kinetics The de-watering kinetics can be subdivided into three separate categories: 1) the cake building kinetics, 2) the kinetics of the flow through the porous system, and 3) the kinetics of the under-saturation of the filter cake. For the evaluation of the cake building kinetics the t/VL versus VL diagram mentioned above can be used to quantitatively analyze the results. Recall from Eq. (4)that the slope of the ~ Nversus L VL curve is proportional to the specific cake resistance. In Figure 9, the influence of the B-field on the cake building process can be observed. Shown in Figure 9 is the comparison of the mean value of t/VL versus VL for experiments with and without a superposed B-field. The mean value of ~ / V Lwas calculated from four separate experimental runs. As the B-field is superposed, the slope of the t/VL curve correspondingly decreases. A decreasing slope implies faster cake building kinetics. An applied B-field of 0.8 T (the maximum used in this portion of the study) shows a decrease in the cake resistance of -25%. The primary reason for this improvement is the physical structure change in the filter cake as it builds in the presence of the external B-field. This structure change will also have a positive effect on the liquid flow through the porous medium and the de-watering process. It is important to note that in traditional cake filtration data analysis, a more common plot is that of the ~ / V L versus VL for increasing values of Ap. The traditional Ap plots show qualitatively similar behavior as the applied B-field. That is, as Ap increases the slope correspondingly decreases, which implies faster cake building kinetics. Using the data from Figure 9, a direct comparison between the kinetics of the

217

MELS process (B # 0 T, Apl = constant) and the traditional cake building process (B = 0 T, Ap2 = constant) can now be made. An applied B-field of 0.8 T at a Ap of 0.8 bar has equivalent kinetic behavior (i.e. cake resistance) as a Ap = 1.2 bar in zero applied B-field. Thus, the MESLS process reduces the required Ap by 50 %, leading to a more energy efficient solid-liquid separation.'

-

5

0

10

15

20

v km'l Figure 9: Cake building kinetics for iron oxide 10, Ap = 0.8 bar.

5.5. Acceleration of the Gas-breakthrough The data and results presented thus far were taken on products that have strongly 70

,

............................................................................................

....................................................

0 B = 0.0T

B = 2.73 T

Figure 10: The gas breakthrough time (ta) for TiOz, Ap = 3.3 bar.

The efficiency increase is dependent upon the entire solid-liquid separation system including the type of magnet used to generate the external B-field (e.g. HTS, normal resistive, permanent, etc.).

218

paramagnetic to ferromagnetic properties (see Table 1). hfESLS technology may also have a positive benefit for materials with weak magnetic properties. For example, Figure 10 shows preliminary data taken on weakly paramagnetic TiOz (xm-10-7),presenting the (mean) gas breakthrough time (tb) for two different applied B-field conditions. The mean tb value/error bars in Figure 10 are the calculated average& of four separate experimental runs, respectively. These preliminary results indicate that the gas breakthrough time occurs earlier in TiOz subjected to an applied B-field. It is believed that this effect is due to changes in the particle-particle interaction of TiOz, but further experimental validation is required. 6. Summary DuPont is developing an entirely new method of solid-liquid filtration involving the use of magnetic fields and magnetic field gradients. The new hybrid process (MESLS) is designed to improve the de-watering kinetics and reduce the residual moisture content of solid particulates mechanically separated from liquid slurries. An experimental investigation was performed to evaluate the effect of the magnetic field on the solid-liquid separation processes. The experimental configuration consisted of a modified CUNO filter rigidly mounted within the internal bore of an HTS magnet, designed by DuPont Superconductivity. Four different products with varying degrees of magnetic susceptibility, ranging from ferromagnetic to weakly paramagnetic, were investigated. Results showed that the application of the magnetic field had a positive benefit for all four products studied in this investigation. One of the products, paramagnetic Fez03 (FE 601), when processed in the presence of an applied B-field of 1.9 T showed a relative reduction of 17 % in the residual moisture content when compared with the same product processed in zero applied B-field. When considering the volume of material processed worldwide, MESLS technology could have a significant impact on energy consumption. DuPont is also investigating the use of MESLS technology in other traditional solid-liquid separation processes e.g. centrifugation, gravitatiodsedimentation, magnetic beads, etc. New experimental results and a theoretical model describing the influence of the B-field on the solid-liquid separation process will be presented at a later date. MESLS technology may be applicable to mineral, chemical, and biological processing.

219

Acknowledgments This work has been financially supported by the U.S. Department of Energy Superconductivity Partnership Initiative under contract number DE FC3699GO- 10287.

References 1. Fuchs, B., Peuker, U., Stahl, W., Steam Penetration and Plug-flow during Steam Assisted Centrifugal Dewatering, Fluid4Separation Journal 14, (2002), pp. 177-183. 2. Weber, K., Untersuchungen zum Einfluss eines elektrischen Feldes auf die kuchenbildende Pressfiltration, Dissertation, Universitat Karlsruhe, (2002). 3. Hofmann, R., Improvement of Dead-end Filtration of Biopolymers with Pressure Electrofiltration, Chemical Engineering Science, 58, (2003), pp. 3847-3858. 4. Tsouris, C., Scott, T.C., Flocculation of Paramagnetic Particles in a Magnetic Field; Journal of Colloid and Interface Science, 171, (1995), pp. 319-330. 5. Svoboda, J., Zofka, J., Magnetic Flocculation in Secondary Minimum, Journal of Colloid and Interface Science, 94, (1983), pp. 37-44. 6. Anlauf, H., Entfeuchtung von Filterkuchen bei der Vakuum-, Druck- und DrucWakuum-Filtration, Dissertation, Universitat Karlsruhe, ( 1986). 7. Watson J.L. and Gardner P.L., Multi-force Dewatering for Magnetic Waste Materials, Minerals Engineering, 8,1/2 (1995) pp. 191-200.

NEW APPLICATIONS OF MAGNETIC SEPARATIONUSING SUPERCONDUCTINGMAGNETS AND COLLOID CHEMICAL PROCESSES

s.TAKEDA’, s.-J.YU’,A. NAKAHIRA~,Y.IZUMI’, s. NISHIJIMA’, T. WATANABE3 ‘Graduate School of Engineering, Osaka University Yamadaoka 2-1, Suita-city, Osaka, 565-0871 JAPAN ’Faculty of Engineering, Kyoto Institute of Technology, Goshokaido-cho, Matsugasaki, Sakyo-ku, Kyoto, 606-8585 JAPAN ’Graduate School of Engineering, Tokyo Metropolitan University, 1-1 Minami-Osawa, Hachioji, Tokyo, 192-0397JAPAN High gradient magnetic separation (HGMS) can be a promising new environmental purification technique as it produces no contaminants, such as flocculants, and could possibly treat large amounts of waste water within a short time frame. A colloid chemical process for magnetic seeding can allow us to rapidly recover a large quantity of adsorbate and to strongly magnetize individual particles in order to improve the recovery efficiency of magnetic separation. In this paper, we will report on the fundamental study of the magnetic seeding process and purification processes using HGMS, and also on studies of applications of the water treatment system for actual factories. Emphasized is a report on a system constructed for water treatment from a paper-manufacturing factory.

1. Introduction

In Japan, new water treatment systems using magnetic force have been actively developed since 1999. These new water treatment systems have many excellent features: low environmental load to the Earth, waste recycling, compactness, fast and reliable treatment. These excellent features are produced by magnetic force using a superconducting magnet. This paper describes the pre-treatment processes for magnetic separation called the “magnetic seeding process” and water treatment systems using magnetic force, both of which have been developed by the Osaka University group. The magnetic separation process had mainly been applied to separation of ferromagnetic materials such as magnetite (Fe304). However, when weak magnetic particles (paramagnetic andor diamagnetic) are to be separated with high efficiency, a magnetic seeding process is required. The idea of employing magnetic seeding, which is usually performed by using magnetite particles assisted by a flocculant to separate the weak magnetic particles, is relatively new [ 1,2] and not yet well developed. In order to magnetically seed the materials for 220

221

separation or recovery, ferromagnetic particles, such as magnetite, are adhered to their surface. So far, the magnetite particles generally used for magnetic seeding have relatively small specific surface areas, requiring recovery of a large volume of particles. In order to recover materials such as organic dyes in wastewater using an environmentally benign and highly efficient process, it is necessary to develop a new technique for diminishing the use of flocculant and for minimizing the volume of magnetic particles added. In general, small magnetic particles having large specific surface areas can effectively adhere to large amounts of adsorbate, implying that smaller magnetic particles are capable of adsorbing larger amounts of materials. But smaller particles, on the contrary, have less magnetization, depending on the particle size. This means that they must acquire a stronger magnetic force in spite of their already high adsorption capacity. From the viewpoint of material chemistry, Okamoto already indicated that one characteristic of iron oxide colloid is a strong scavenging action, which is responsible for high specific surface area, high adsorption capacity, and strong adhesion force [3]. This indication also suggests that using magnetic colloid containing iron oxide may enable us to magnetically seed organic dyes without flocculant. In practice, however, using a magnetic colloid in magnetic separation was historically not possible as the technical environment of HGMS was not yet fully developed and a high magnetic field could not easily be produced. In this paper, we review a new magnetic seeding process that we have developed for paramagnetic and diamagnetic materials, so far called the “colloid chemical process”. Furthermore, we will also show a practical application of the colloid chemical process for magnetic seeding to wastewater from a paper-manufacturing factory. This has been successful in purifying the wastewater by using a superconducting, high gradient magnet. In general, manufacturing paper produces a large amount of wastewater, so the factory should purchase the same amount of clean water. Thus, the cost of water issues becomes so high that the factory finds it necessary to introduce an advanced technology for purifying its wastewater. Since the factory treating used paper for recycling is usually located in the city (as this is where used paper is gathered), its wastewater treatment is subjected to severe conditions in order to purify it sufficiently for drainage into the city sewage system.

222

2.

Colloid Chemical Process for “Magnetic Seeding” and Magnetic Separation Test

2.1. Experimental Procedure 2.1.1. Materials Orange I1 was used as a test material for magnetic seeding. The Orange I1 used during this work was purchased from Merck. Distilled water was also used in the experiment. The magnetic colloids were produced by precipitation in aqueous solution. In a typical procedure, 0.1 moVdm3 of ferrous sulfate was dissolved in 100 cm3 of distilled water. After the ferrous sulfate was completely dissolved, 5 moVdm3 KOH aqueous solution was added to the prepared ferrous sulfate solution at a molar ratio of 1:2 for Fe:OH. Ferrous hydroxide (green rust) particles were formed just after the mixing. Subsequently, the mixture solution with suspended Fe(Om2 precipitate was then stirred for 30 min in order to oxidize Fe(0H)z precipitates to Fe304 colloid particles in air at room temperature (25 f 3 “C). Commercially available magnetite particles supplied by Toda Kogyo Cop., for comparison with magnetic colloid particles, was also used. 2.1.2. Magnetic Seeding Procedure In order to recover dissolved Orange I1 molecules, commercially available magnetite particles and prepared magnetic colloids were used to magnetically seed the molecule. The magnetic seeding procedure adopted in this work was as follows. 100 cm3 of Orange I1 aqueous solution was placed into a 140 cm3 polystyrene vessel. The concentration of Orange I1 solution was fixed at 1 . 0 ~ 1 0moVdm3. ~ When magnetite particles were mixed into the Orange I1 solution, the net concentration of iron oxide was changed from 60 mg/lOO cm3 to 5500 mg/lOO cm3. When magnetic colloid was mixed into the Orange II solution, the net concentration of iron oxide particles was adjusted to the same mass based on dry weight as when magnetite particles were added. The calculated volume of magnetic colloid was drawn out and mixed with the Orange I1 solution. For this procedure, in order to maintain the Orange II concentration in solution independent of the added magnetic colloid, distilled water was added to the colloid solution when the small amount of colloid was added. After mixing magnetic colloid particles or magnetite particles with the Orange I1 solution, the mixed solution was stirred for 2 min and subsequently

223

used for recovery and adsorption tests. The stirring time was fixed at 2 min because magnetic separation should be required to occur in the shortest time possible.

2.1.3. Magnetic Separation Test The high gradient magnetic separation (HGMS) apparatus is shown schematically in Figure 1. The magnet (Japan Magnet Technology, Inc.) has a bore diameter of 10 cm and length of 120 cm. The actual HGMS column (referred to as canister) has a 4 cm inner diameter polyethylene tube 30 cm in length. Fresh steel wool is gently packed to a bed length of 20 cm. The stainless steel wool has a 27 pm diameter and weight of 6.5 g. The Orange I1 / magnetite solution was passed through the HGMS filter at a rate of 2000 cm3/min and captured on the stainless steel wool. All of the separation tests were performed under a 2 T magnetic field. After separation, absorbance (W-240, Shimadzu, Ltd, Japan) at 485 nm through the filtrates is used to determine the recovery fraction. 2

Inlet

Outlet Figure 1. Schematic of the high gradient magnetic separator. 1-Superconducting magnet, 2-Refrigerator,3-Filter canister, 4-Magnetic particle-dye molecule, 5-Stainless steel wool.

2.2. Result of Magnetic Separation Test Recovery tests of Orange I1 using magnetic colloid and commercially available magnetite particles were performed. Figure 2 shows the recovery of Orange II molecules from an aqueous solution as a function of the added amount of seeding particles. Most of the adsorbed Orange II molecules were recovered

224

when passed through the stainless steel wool filter in the superconducting magnet. For the magnetic colloid, the removal efficiency of Orange I1 molecules abruptly increased as the amount of magnetic colloid added increased. Alternately, Orange I1 molecules were removed at only about 8% at 100 mg/100 cm3 of the added amount of magnetite particles. The recovery rate in the case using magnetite particles is very low compared to the case using a magnetic colloid. In order to remove much more of the Orange I1 molecules, more magnetite particles were added to the Orange I1 solution. The recovery increased as the added amount of magnetic particles increased. When the added amount of magnetite particles reached 5500 mg/100 cm3, the recovery rate reached 91%. Therefore, it is necessary to add a large amount of magnetite particles to treat an enormous volume of dyed wastewater. 100

80 --

h

U

--

magnetite

60

2

' 0

>

4

40

20

0

0

0.4 0.8 1.2 1.6 Seeding concentration(g/l OOcm3)

2

Figure 2. Recovery of Orange I1 from aqueous solution as a function of added amounts of magnetic seeding material.

3.

A Practical Application Study of Magnetic Seeding to Wastewater from the Paper Manufacturing Factory

Figure 3 shows a schematic diagram of the magnetic separation system for purification of wastewater from the paper factory. The system mainly consists of a mixing tank (magnetic seeding tank), a first separation tank, and a

225

superconducting magnet chamber. Floating magnetic flocks composed of magnetite particles and organic polymers such as pulp and dye are captured by magnetic force in the superconducting magnet chamber. Some magnetic flocks undergo gravity sedimentation at the first separation tank, which helps to reduce the volume of magnetic flocks traversing the magnet chamber. This test plant of 500 todday was constructed in the actual paper manufacturing process and a purification test of the wastewater was performed. A colloid chemical process for magnetic seeding of the organic pulp and dyes were successfully performed, therefore the COD value of drainage of less than 20 ppm was obtained after magnetic separation. ,____..____...___.....

-___--____

_I...-___.___...____-.-.-------

Figure 3. Schematic view of the magnetic separation system for paper factory wastewater purification.

Acknowledgements This work was partly supported by a Grant-in-Aid of “Research for the Future Program” the Japan Society for the Promotion of Science. The financial support from a Grant-in-Aid of NED0 is also greatly acknowledged. References 1. Takeda, S., Yu, S., and Tari, I., Proceedings of the Meeting of New Magneto-Science, TML, Ann. Rep. Suppl. 11, 183 (2001). 2. Takeda, S., Yu, S . , Tari, I., Nishijima, S., Nakahira, A., Bull. Chem. SOC. J u ~ u76, ~ , No.5, (2003), pp. 1087-1091. 3. Okamoto, S. and Okamoto, I., Yogyo-Kyokui-Shi, 85 (1977), p. 518.


Biological Applications


NANOMAGNETICS IN BIOTECHNOLOGY C.-J. CHEN, Y. HAIK, J. CHATTERJEE Center for Nunomagnetics and Biotechnology College of Engineering, Tallahassee, FL 32310, USA Applications of nanomagnetic particles in cell separation and magnetic hyperthermia are presented here. In general, biological cells are weakly paramagnetic or diamagnetic. Therefore, to enhance the magnetic susceptibility of the biological cells to interact with an applied magnetic field, nano- to micron-sized magnetic particles are tagged to the biological cells. Nano-size magnetic particles are produced in-house for various biomagnetic applications.

1. Introduction

Biological molecules and cells are considered either paramagnetic or diamagnetic materials. Diamagnetic materials pull away from a strong magnetic field. Levitation of biological materials, such as blood, was demonstrated in a high magnetic field. It has been found that human red blood cells have the characteristics of a paramagnetic fluid when deoxygenated (in veins) and diamagnetic when oxygenated (in arteries) [ 1,2]. The biological elements, whether diamagnetic or paramagnetic, generally have a weak magnetic susceptibility on the order of In order to manipulate biological cells, such as for separation of cellular components, magnetization of the biological cells is first enhanced by tagging the cellular or molecular components with magnetic material [3,4,5,6,7]. Labeled encapsulated magnetic particles attached to the biological cells are often used in clinical applications. These encapsulated magnetic particles are spherical in shape and generally are referred to as magnetic microspheres. In using magnetic microspheres for medical applications, the surface of the magnetic microspheres is treated so that if a target such as red cells, polymer or protein is present, the magnetic microspheres will hybridize to it. Bulte et al. [3] used monocrystalline iron oxide nanoparticles as contrast agents for magnetoimmunodetection of inflammations. Jordan et al. [4] showed that magnetic particles could be used to develop localized thermal heating. Thermal energy is generated by an invasive, applied electromagnetic energy on the magnetic beads. Lubbe et al. [S] used magnetic particles for site-specific drug delivery. Several investigators [6,7] have sorted white cells using magnetic microspheres. Zborowski et al. [6] developed a rapid cell sorting technique for human 229

230

lymphocytes. Lymphocytes are obtained by using centrifugal techniques then given a magnetic label. Such labeling is used to target specific lymphocytes, such as monocytes or T helper cells, and is performed by magnetic sorting. Assenmacher et al. [7] showed that a large number of specific lymphocyte targets could be isolated using magnetic nanoparticles. Hafeli [8] demonstrated that magnetic microspheres coated with biodegradable polymer showed no signs of toxicity. Experimental protocol involved injecting rats with 10 mM of magnetic microspheres. Lubbe et al. [5] and Kuznetsov et al. [9] injected up to 700 mg of magnetic microspheres for each kg of animal weight and showed no signs of toxicity in the rats. Many studies have been initiated to study the effect of magnetic fields on the human body. Studies in our center have shown that magnetic fields affect the growth of cancer cells. Several investigators reported the effect of high magnetic field on the cellular components of biological cells. Higashi et al. (1993) [lo] studied the orientation of the erythrocytes in strong magnetic fields (from 1 to 8 tesla) and reported that the erythrocytes are found to orient with their disk plane parallel to the magnetic field. Motta et al. [ll] studied the effect of magnetic fields on human whole blood light absorption. Their results showed an orientational effect on blood cells due to the applied magnetic field. The orientational effect due to the magnetic field was found to enhance the blood viscosity [ 121.

2. Nano-sized Magnetic Particles The production of specialized magnetic microspheres for various medical applications has become a major research thrust in recent years. In ow center, we have developed magnetic particles on the order of 2-10 nm in diameter. At these small sizes, the particles are considered biodegradable and can be used in medical applications. When the particles are attached to biological cells, the cells behave as magnetic particles. In medical applications, the magnetic particles are attached to the cells and then an applied magnetic field is used to manipulate cellular components. The magnetic microspheres are prepared by encapsulating magnetic particles in a biodegradable protein, and the protein surface is coated with a suitable surfactant that couples with the cellular component. This section outlines the preparation of the magnetic microspheres. The magnetic material to be encapsulated in the protein must be nano-sized for biodegradability and toxicity reasons, and also to not interfere with the functions of the cellular components. Nano-sized iron oxides are prepared in our laboratory as follows:

231

Co-precipitation of ferrous chloride and ferric chloride by sodium hydroxide: Ferrous chloride and ferric chloride are mixed in a molar ratio of 1:2 in deionized water at a concentration of 0.1 M iron ion. The solution is used immediately after preparation. A highly concentrated solution of sodium chloride (10M) is added for co-precipitation with continuous stirring. Peptization with nitric acid: The solution with the precipitate is stirred at high speed for 1 hour at 20°C then heated at 90°C for 1 hour with continuous stirring. The ultrafine magnetic particles are peptized by nitric acid (2 M). Sonication: The iron oxide dispersion is then sonicated for 10 minutes at 90°C at 50% amplitude. The precipitate is then washed repeatedly with deionized water, filtered and dried in a vacuum. Figure 1 shows a TEM monograph of the magnetic particles obtained by the above procedure. The average diameter of the magnetic particles is about 4 - 10 nm.

Figure 1: TEM of iron oxides

3.

Synthesis oi Albumin Magnetic Microspheres

Synthesis of albumin magnetic microspheres in the submicron range using nanosized magnetic particles has been accomplished. Human Serum Albumin (HSA) and iron oxide (-30% of the weight of HSA) is dissolved in distilled water and finally added to cottonseed oil containing 0.2 ml of sorbitan sesquioleate. The mixture is shaken vigorously and then sonicated (Cole Parmer Ultrasonic homogenizer) for three 30 seconds intervals at 60% amplitude. The

232 sonication process took place at 4°C using an ice-water bath. This primary emulsion was then added drop-wise to cottonseed oil heated at 130°C and stirred at 1500 rpm. The entire addition was done in 10 minutes and then the mixture was kept at that temperature and stirred at the same speed for another 15 minutes. The heat-stabilized microspheres were then cooled and extracted with diethyl ether and centrifuged. After being washed three times with diethyl ether, the dispersion of microspheres in ether was filtered successively using nylon filter membranes. The microspheres were then dried in air and stored in a vacuum desiccator. Figure 2 shows a SEM monograph of micron size magnetic microspheres synthesized in our laboratory.

Figure 2: SEM of magnetic microspheres

Figure 3: Red blood cell attached to albumin magnetic microspheres

The synthesized magnetic particles can be further modiEied to attach to cellular components. An application that was developed for Therakos, Inc. in

233

ow center is the separation of red blood cells from whole blood. For this application, magnetic microspheres were modified with avidin and biotinylated lectin. They can be attached to the red cell membrane as shown in Figure 3. When magnetic particles are attached to the cellular component, an external magnetic field can be used to isolate the attached cellular components from the blood. 4. Cell Separation Separation of cellular components using magnetic technology is one of the fastest growing applications of bio-magnetic fluids. In this technology, a magnetic force is imposed on cellular components to isolate the cells from the heterogeneous biofluid. Since the magnetic force is a function of the magnetic susceptibility, the magnetic field and the magnetic field gradient, attaching a magnetic particle to the cellular membrane enhances the magnetic susceptibility of the cellular components. When an applied magnetic field is imposed on the heterogeneous solution that contains the tagged cells, the cells will be attracted toward the higher magnetic field. We have developed a magnetic separation device to isolate red blood cells (RBC) from whole blood [ 151. This is done to facilitate photopheresis treatment for white cells. In the photopheresis treatment, an ultraviolet (W) lightactivated drug is administered to the patient. The drug attaches to the DNA of the white cells and requires a dose of UV light to be activated. In order for the UV light to penetrate through the whole blood, the red blood cells must be isolated. Thus, magnetic microspheres similar to the one shown in Figures 3 and 4 were synthesized. The new magnetic separation device was developed for Therakos, Inc. and has improved separation efficiency from 50% to 95%. Other research groups such as Liberti et a1 [17] and Zborowski et a1 [18] have reported the use of magnetic particles on the order of 150 nm to isolate breast cancer cells from peripheral blood.

234

Figure 4. SEM of attached RBC to magnetic particle

5. Magnetic Hyperthermia Many cancers contain a significant fraction of hypoxic (poorly oxygenated) cells that are much more resistant to radiation therapy than euoxic (well oxygenated) cells. This constitutes a major obstacle to successful radiotherapy. To overcome this problem, Hofer [16] developed a treatment technique where radiation is used in conjunction with mild hyperthermia (heat) and chemical agents (metronidazole, misonidazole) that enhance the radiation response of cancers. Heat alone (41SoC) enhanced radiation-induced cell death among hypoxic cancer cells by a factor of 1.7, misonidazole alone (0.5 mg/g body weight) by a factor of 2.2. These findings show that heat and misonidazole are effective in potentiating radiation death in hypoxic cancers, but neither mode of radiosensitization is as effective as full tumor oxygenation (oxygen enhancement ratio of 2.9). When heat and misonidazole are used in combination with each other, they produce synergistic potentiation effects that far exceed the action of oxygen. Figure 5 shows that the radiation dose can be substantially reduced if cancer cells can be heated. For hypoxic cancer cells subjected to the two agents during irradiation, the response was enhanced by a factor of 4.3, that is, radiosensitization by combination treatment was such that the hypoxic cancer cells actually became more radiosensitive than euoxic cells [ 161. Other scientists working on a variety of different tumor systems confirmed these findings. Thus, there is no longer any doubt that this effect exists. Equally important, the effect is selective for hypoxic cancers, that is, normal body tissues are not sensitized. In spite of the enormous magnitude of the sensitization effect, two improvements are required before this treatment can be used in clinical radiotherapy. First, we need better methods for selective tumor heating. The studies described above were performed on mice that were subjected to whole-

235 body heating in a water bath. This is not optimal for clinical application as whole-body heating limits the heat dose administered to the cancer. Combination therapy could be significantly improved by heating tumors selectively with magnetic particles subjected to external AC magnetic fields. The second improvement would be the development of a more selective method of drug delivery to tumors. Metronidazole and misonidazole are well tolerated in mice, but clinical use of these agents in human patients can result in severe peripheral neuropathy. One promising, potential solution to this obstacle is to encapsulate the drugs in heat-sensitive liposomes. Ideally, the liposomes should also contain magnetic particles, so they can be guided to the tumor by a magnetic field. After the liposomes reach the interior of the tumor, they could be heated by an external AC magnetic field. This would cause the liposomes to burst and selectively release the radiosensitizing agent in the interior of the tumor. Such a system may also facilitate selective heating of local tumors. If successful, this procedure could offer an elegant method to realize the promise of combination radiosensitization in the clinical radiotherapy of cancers. Hyperthermia with magnetic particles combined with drug therapy has the potential to be very effective. When magnetic particles are placed in an appropriate oscillating magnetic field, they reach a higher temperature than that of their surroundings. With the use of magnetic delivery devices, magnetic particles can be delivered to the site of the cancer tumor and activate the radiation treatment. The research conducted in our center includes optimization of the magnetic material mixture. It is shown that heat can be generated from hysteresis losses in the magnetic materials. One important requirement is that magnetic materials lose their magnetic properties above the Curie temperature to protect the cells. The Curie temperature for most magnetic materials used for hyperthermia is around 50°C. In our center, we are investigating the use of a combination of drugs and magnetic hyperthermia to produce the required heat in the tumor area to reach a temperature of 42.5"C.

236

Hyperthermia+Oxidant +Radiation 1

0.01 HoJer CI d

0I

5

10

15

x)

25

RADIATION DOSE (Gy)

Figure 5 . Cancer cell surviving fraction vs. radiation dose

6. Conclusions This paper presents some recent developments in nanomagnetics and biotechnolog y applications at the Center for Nanomagnetics and Biotechnology. It is shown that the use of encapsulated magnetic materials attached to cellular components enhances the magnetic susceptibility of the biological cells. Production of the nano-magnetic particles that are able to attach to cellular components in the cells is presented. Applications in the use of nano-magnetic particles for cell separation and magnetic hyperthermia are presented and discussed.

References 1. Shaylgin A.N., Norina S.B. and Kondorsky, E.I., “Behavior of Erythrocytes in High Gradient Magnetic Field,” Journal of Magnetism and Magnetic Materials, 31 (1983) pp. 555-556. 2. Haik, Y., Chen, C.J. and Pai, V.M., Development of Biomagnetic Fluid Dynamics. Proceedings of the Ninth International Symposium on Transport Phenomena in Thermal-Fluids Engineering, Ed. S.H. Winoto, Y.T. Chew and N.E. Wijeysundera, Singapore, (1996) pp. 121-127. 3 Bulte J.M.W., Brooks, R.A., Moskowitz, B., Verkuyl, J., Herynek, V., Brocke, S., Bryant, H., Katsanis, E., Frank, J.A., ”Antiferromagnetic

237

Nanoparticles as Contrast Agents For Magnetoimmundetection of Inflammation-Specific Adhesion Molecule." Proceedings of the 2nd International Conference on the Scientific and Clinical Applications of Magnetic Carriers, Cleveland-Ohio, (1998) p. 17. 4 Jordan, A., Scholz, R. Wust, P., Faehling, R. and Flix, R., "Magnetic Fluid Hyperthermia, Current Status and Future Perspectives." Proceedings of the 2nd International Conference on the Scientific and Clinical Applications of Magnetic Carriers, Cleveland-Ohio, (1998) p. 29. 5. Lubbe, A.S., Bergemann, C., Clure, D., Brock J., "Physiological and Technical Aspects of Magnetic Drug Targeting for Clinical Applications." Proceedings of the 2nd International Conference on the Scientific and Clinical Applications of Magnetic Carriers, Cleveland-Ohio, (1998) p. 37. 6. Zobrowski, M., Moore, L.R., Chalmers, J., "Rapid Cell Separation by Magnetic Flow Sorting." Proceedings of the 2nd International Conference on the Scientific and Clinical Applications of Magnetic Carriers, ClevelandOhio, (1998) p. 71. 7. Assesnmacher, M., Miltenyi, S. and Schmitz, J., "High Gradient Magnetic Sepration in Biomedical Applications." Proceedings of the 2nd International Conference on the Scientific and Clinical Applications of Magnetic Carriers, Cleveland-Ohio, (1998) p. 68. 8. Hafeli, U. and Pauer, G., "In Vivo and in Vitro Toxcity of Magnetic Microspheres." Proceedings of the 2nd International Conference on the Scientific and Clinical Applications of Magnetic Carriers, Cleveland-Ohio, (1998) p. 42. 9. Kuznetsov, A.A., Filippov, V.I., Gerlivanov, V.G., Dorbinsky, E.K. and Malashin, S.I., "New Ferro-Carbon Sorbent for Magnetically Guided Transport of Anti Cancer Drugs." Proceedings of the 2nd International Conference on the Scientific and Clinical Applications of Magnetic Carriers, Cleveland-Ohio, (1998) p. 17. lO.Higashi, T., Yamagishi, T., Takeuchi, A., Kawaguchi, N., Sagawa, S., Onishi, S. and Date, M., "Orientation of erythrocytes in a Strong Static Magnetic Fields," Journal of Blood, 82(4), (1993) pp. 1328-1334. ll.Motta, M., Haik, Y. and Chen, C.J., "Effect of High Magnetic Field on Human Blood," presented at the First Latin American Conference on Biomedical Applications, Mazatlan, Mexico, Nov. 1998. 12.Haik, Y., Pai, V. and Chen, C.J., "Biomagnetic Fluid Dynamics," Chapter 10, W. Shyy, editor, Cambridge University Press, In Print. 15.Haik, Y., Pai, V.M., and Chen, C.J., "Development of Magnetic Device for Cell Separation," Journal of Magnetism and Magnetic Materials, 194, (1999) pp. 254-261. 16.Hofer, K. "Hyperthermia and Cancer," Proceedings of the 4" International Conference on the Scientific and Clinical Applications of Magnetic Carriers, Tallahassee, Florida, (2002) pp. 78-80.

STRONG MAGNETIC FIELD INDUCED CHANGES OF GENE EXPRESSION IN ARABIDOPSIS A.-L. PAUL', R.J. FERL', B. KLINGENBERG', J.S. BROOKS3,A.N. MORGAN4, J. YOWTAK4, M.W. MEISEL4 'Department of Horticultural Sciences and 7he Biotechnology Program, University of Florida, Gainesville, FL 3261 1-0690, USA 'Department of Statistics, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL 3261 1-0339, USA 3Department of Physics and National High Magnetic Field Laboratory, Florida State University, Tallahassee, FL 32306-4005, USA 4Department of Physics and National High Magnetic Field Laboratory. University of Florida, Gainesville, FL 3261 1-8440, USA We review o w studies of the biological impact of magnetic field strengths of up to 30 T on transgenic arabidopsis plants engineered with a stress response gene consisting of the alcohol dehydrogenase (Adh) gene promoter driving the p-glucuronidase (GUS) gene reporter. Field strengths in excess of 15 T induce expression of the Adh/GUS transgene in the roots and leaves. Microarray analyses indicate that such field strengths have a far reaching effect on the genome. Wide spread induction of stress-related genes and transcription factors, and a depression of genes associated with cell wall metabolism are prominent examples.

1. Motivation

Since earth-based, low-gravity environments are restricted to durations of less than several seconds [ 11, we have investigated the possibility of using magnetic levitation as a reduced gravity environment appropriate for long-duration (i.e. up to several hours) studies of plant gene expression [2,3]. The possibility of using magnetic levitation to mimic a reduced gravity environment has been explored in a variety of systems [3-71. During our preliminary studies involving transgenic arabidopsis, we discovered that the plants were stressed by the presence of a strong, static, non-gradient magnetic field [2]. These preliminary, qualitative observations have now been corroborated by a series of systematic, quantitative studies [8]. The possibility that strong, static (non-gradient) magnetic fields might have an influence on biological processes has been discussed for many years [9-111, including a recent report that implicates high magnetic fields in alterations of the cleavage plane during cell division [ 121. Nevertheless, the common viewpoint is that presently achievable static magnetic fields do not have a lasting effect on biological systems [ 111. Indeed, magnetic resonance imaging (MRI), utilizing 238

239

static magnetic fields up to 12 T, is a powerful tool for non-invasive in vivo imaging at the molecular level [13-151. The demands for more precise in vivo imaging have driven the field strengths progressively higher (approaching 20 T) [ 161, yet information regarding the biological impact of exposing metabolically active cells to fields of this magnitude is limited. We have recently reported about the effect of high magnetic fields on the gene expression profile of the plant Arabidopsis (Arabidopsis thaliana) [8].

2. The Plants

Our research efforts employ transgenic Arabidopsis that had been engineered with a gene reporter shown to be induced by a spaceflight environment (TAGES - Transgenic Arabidopsis Gene Expression System) [17]. The TAGES Arabidopsis plants are engineered with the GUS (P-glucuronidase) reporter gene driven by the alcohol dehydrogenase (Adh) gene promoter [17]. The Adh promoter responds to a variety of environmental stresses (e.g. hypoxia, cold, abcissic acid) [18], which in turn initiates transcription of the GUS reporter gene. GUS expression can be monitored qualitatively, by histochemical staining, and quantitatively, by biochemical assays of the gene product. The magnetically levitated plants showed evidence of reporter gene activation, however, the control plants placed in the static magnetic field (19 T) showed similar patterns of transgene expression [2]. These observations lead to the design of additional experiments using transgenic plants as biomonitors of the effects of high magnetic fields in metabolically active tissues. The evaluation of global changes in the Arabidopsis genome in response to exposure to high magnetic fields was facilitated with the Affymetrix@ Genechip@ arrays. These arrays allow for the survey of over 8000 genes at a time and were used for genome-wide characterization of the effects of exposing Arabidopsis plants to a field strength of 21 T. The resulting differential patterns of gene expression from the array data were then used to guide quantitative analyses of gene expression with Real-Time, quantitative RT PCR. This approach is an effective means of characterizing an abiotic stress response [ 191. 3. Results and Discussion Of the 8000 genes surveyed, there were 112 genes that were differentially expressed to a degree greater than 2.5 fold over the control, Figure 1. Many genes associated with a variety of stress responses were induced (heat, cold, drought, touch), as were genes encoding proteins that are involved with ion

240

transport functions (chloride, sulfate, ammonium). The down regulated set included a number of genes involved in cell wall biosynthesis (e.g. Xtr7). Genes that encode transcription factors populate a final, large category (e.g. Athb12). Additional experiments are under development to address the mechanism by which high magnetic fields induce changes in gene expression patterns in Arabidopsis.

Figure 1. Levels of differentiated gene expression between.the control and treatment at 21 T. The bars represent fold-change differences for representative genes from the Affymetrix" AtHl array, and genes are clustered with regard to metabolic function.

In conclusion, magnetic fields above 15 T induce gene expression responses in Arabidopsis plants, and a detailed presentation of our work is given elsewhere [8].These data provide evidence for the perturbation of metabolic processes in the presence of strong magnetic fields and may be useful for guiding future research aimed at determining safe exposure standards for living organisms.

Acknowledgements , This work was supported, in part, by the National Science Foundation and the State of Florida through support and operation of the National High Magnetic Field Laboratory (NHMFL), the NHMFL In-House Research Program, and the NHMFL Research Experience for Undergraduates. We acknowledge useful conversations with S . J. Hagen, T. H. Mareci, and A. S. Edison.

241

References 1. “NASA Reduced-Gravity Carriers for Experiment Operations”, Microgravity Research Program Office, NASA, Marshall Space Flight Center, http://microgravity.nasa.gov/NASA-Carrier-User-Guide.pdf. 2. Stalcup T.F., Reavis J.A., Brooks J.S., Paul A.-L., Ferl R.J., Meisel M.W. 1999. Transgenic arabidopsis plants as monitors of low gravity and magnetic field effects. In: Z. Fisk, L. Gor’kov, R. Schrieffer, editors. Physical Phenomena in High Magnetic Fields - III. Singapore: World-Scientific, pp. 659-662. 3. Brooks J.S., Reavis J.A., Medwood R.A., Stalcup T.F., Meisel M.W., Steinberg E., Arnowitz L., Stover C.C., Perenboom J.A.A.J. 2000. New opportunities in science, materials, and biological systems in the low-gravity (magnetic levitation) environment. Appl. Phys. S7(9), pp. 6194-6199. 4. Beaugnon E., Toumier R. 1991. Levitation of organic materials. Nature 349, p. 470. 5. Valles J.M., Jr., Lin K.,Denegre J.M., Mowry K.L. 1997. Stable magnetic field gradient levitation of Xenopus laevis: toward low-gravity simulation. Biophys. J. 73(2), pp. 1130-1133. 6. Geim A.K., Simon M.D., Boamfa M.I., Heflinger L.O. 1999. Magnetic levitation at your fingertips. Nature 400, pp. 323-324. 7. Valles J.M., Jr., Guevorkian K. 2002. Low gravity on earth by magnetic levitation of biological material. J. Gravit. Physiol. 9(1), pp. 11-14. 8. Paul, A.-L., Ferl, R.J., Klingenberg, B., Brooks, J.S., Morgan, A.N., Yowtak, J., Meisel, M.W., prepnnt. 9. Maret G., Dransfeld K. 1985. Biomolecules and polymers in high steady magnetic fields. In: F. Herlach, editor. Strong and Ultrastrong Magnetic Fields and Their Applications - Topics in Applied Physics. Berlin: SpnngerVerlag, pp. 143-204. 10.Maret G. 1990. Recent biophysical studies in high magnetic fields. Physica B 164(1-2), pp. 205-212. 1l.Schenck J.F. 1998. MR safety at high magnetic fields. Magn. Reson. Imaging Clin. N. Am. 6(4), pp. 715-730. 12.Denegre J.M., Valles J.M., Jr., Lin K., Jordan W.B., Mowry K.L. 1998. Cleavage planes in frog eggs are altered by strong magnetic fields. Proc. Natl. Acad. Sci. USA 95(25), pp. 14729-14732. 13.Ichikawa T., Hogemann D., Saeki Y., Tyminski E., Terada K., Weissleder R., Chiocca E.A., Basilion J.P. 2002. MRI of transgene expression: correlation to therapeutic gene expression. Neoplasia 4(6), pp. 523-530. 14.Louie A.Y., Huber M.M., Ahrens E.T., Rothbacher U., Moats R., Jacobs R.E., Fraser S.E., Meade T.J. 2000. In vivo visualization of gene expression using magnetic resonance imaging. Nut. Biotechnol. 18(3), pp. 32 1-325.

242

15.Weissleder R., Moore A., Mahmood U., Bhorade R., Benveniste H., Chiocca E.A., Basilion J.P. 2000. In vivo magnetic resonance imaging of transgene expression. Nut. Med. 6(3), pp. 351-355. 16.Lin Y., Ahn S., Murali N., Brey W., Bowers C.R., Warren W.S. 2000. Highresolution, > I GHz Nh4R in unstable magnetic fields. Phys. Rev. Lett. 85( 17), pp. 3732-3735. 17.Paul A.-L., Daugherty C.J., Bihn E.A., Chapman D.K., Norwood K.L., Ferl R.J. 2001. Transgene expression patterns indicate that spaceflight affects stress signal perception and transduction in arabidopsis. Plant Physiol. 126(2), pp. 613-621. 18.Dolferus R., Jacobs M., Peacock W.J., Dennis E.S. 1994. Differential interactions of promoter elements in stress responses of the Arabidopsis Adh gene. Plant Physiol. 105(4), pp. 1075-1087. 19.Paul A.-L., Schuerger A.C., Popp M.P., Richards J.T., Manak M.S., Ferl R. J. 2004. Hypobaric biology: Arabidopsis gene expression at low atmospheric pressure. Plant Physiol. 134(l), pp. 215-223.

NEW APPLICATIONS OF MAGNETIC FIELD TO HUMAN FRIENDLY MATERIALSAND HUMAN SUPPORTIVE SYSTEMS S.TAKEDA', U. HAFELI', M. TONOIKE3,Y. IZUMI', K. EMA', S.NISHIJIMA' 'Graduate School of Engineering, Osaka University Yamadaoka 2-1, Suita-city, Osaka, 565-0871 JAPAN 'Radiation Oncology Department, The Cleveland Clinic Foundation 9500 Euclid Ave. T28.Cleveland, OH 441 95 USA 'Life Electronics Laboratory, AlST Kansai, 1-8-31 Midorigaoka, lkeda-city, Osaka, 563-8.577 JAPAN Developments of characterization techniques of magnetophoretic mobility were performed to explore clinical and biochemical applications in a study on human-friendly materials. Conceptually, magnetic drug delivery using particulate carriers is very efficient for delivering a drug to a localized disease site. Very high concentrations of chemotherapeutic or radiological agents can be achieved near the target site, for example in a tumor, without any toxic effects to normal surrounding tissue. Here, a feasibility study of magnetic microparticles capable of adsorbing bioactive compounds was made in order to confirm the capability of in vivo usage of the magnetic microparticles. In another application for human supportive system, magnetic properties were utilized to detectlcharacterize pain, comfofi, and other feelings utilizing the SQUID system. All applications developed here were related to development of human friendly materials and human supportive system utilizing a magnetic field.

1. Introduction

Our research group, working on the development of human-friendly materials and human supportive systems using magnetic force, has been actively promoting the creation of new materials and systems through a strong collaboration with scientists and engineers in materials science, colloid chemistry, physics, medical science and biotechnology. Our goaI is to improve QOL (Quality of Life) with the assistance of magnets and/or magnetic materials. Our conceptual framework of QOL investigation was defined as a study on the development of tools improving life and the environment through magnets and/or magnetic materials. In contributing to improve QOL, some trials have already begun; those are 1) developing a superconducting magnetic separation system using a new magnetic colloid that can adsorb a wide variety of soluble inorganic/organic materials in wastewater, 2) developing a variety of magnetic microparticles for magnetic targeted carriers and their characterization instrument, and 3) developing an evaluation system, which could measure satisfaction levels (this program has just begun). 243

244

2. Results of the Human Supportive System I (Magnetic Separation System for Water Treatment) There are many methods for cleaning wastewater depending on its origin and composition. An activated sludge process and a coagulation treatment are popular methods of wastewater purification. The former method can change the soluble organic materials into sludge containing the decomposed materials, and the latter makes organic materials flock and sediment with flocculants. These methods have many problems including long reaction times, large physical plant requirements, and large space requirements for mud disposal. A superconducting magnetic separation system provide wastewater purification with none of the abovementioned problems.

Figure 1. Superconducting magnet used in the water treatment system. (500 t/day).

Figure 2. Sample water from 500 t/day plant A: drainage from a paper mill B: mixture in the reaction tank C: water after separation tank

Figure 1 shows the superconducting magnet used for the wastewater purification system from the paper factory Cfordetails please see the other paper of these Proceedings). The obtained results are shown in Figure 2. C showed very clear water, which corresponds to the sample after magnetic separation using the magnet shown in Figure 1.

3. Results of the Human Friendly Materials (Magnetic Targeted Carriers) There is a great deal of interest in investigating new synthesis routes, controlling the size and the morphology, and understanding the overall behavior of magnetic microspheres (or magnetic targeted carriers). Magnetic microspheres are being studied in particular for their current and future applications in biology and medicine, including magnetic transport of anti-cancer drugs, magnetic cell separation, and magnetic resonance imaging contrast enhancement [ 11. In this field and especially for in vivo applications, the main challenges currently

245 consist in (i) tailoring their surface in order to functionalize and/or develop strong interactions with specific bioactive compounds (drug, protein, dye etc.) and (ii) controlling the size, morphology and "magnetic responsiveness (amount of velocity for a given magnetic field and field gradient)". In general, bioactive compounds have paramagnetism and/or diamagnetism, so a magnetic seeding process is needed when paramagnetic or weakly magnetic compounds are to be delivered by a magnet. The idea to employ magnetic seeding, which is usually performed by using magnetic particles assisted by a flocculant to deliver/separate weak magnetic particles, is relatively new [2,3] and not yet well-developed. The magnetite particles used in the usual process of magnetic seeding so far have a relatively small specific surface area, which requires the assistance of a flocculant or some chemical additives to combine magnetic particles with bioactive compounds. In order to combine magnetic microspheres with weakly magnetic compounds very simply, the "colloid chemical process for magnetic seeding (CCPMS)" was developed recently by the authors [5]. In this section, we show an application of this process. • albumin

0.35

1 lysozyme

0.3

a

,0

A y-globulin

0.25

0.2

1 8 0.15

T3

.

• O

• * •

m •

•

*.

"

*' A

•

•

i

0.1

£

0.05 n

.

•i

•• i

i

• !

10

l_J

12

14

Figure 3. Variation of protein adsorption of magnetite particles at different solution pH.

Figure 3 gives typical results, showing the capability of magnetic microspheres adsorbing proteins. In considering the clinical applications of magnetic microspheres combined with some bioactive compounds, an all-inclusive test that allows us to choose the most appropriate magnetic microsphere for a certain application would be advantageous. When a conventional technique such as measurement of magnetic susceptibility is employed, it only gives an approximate indication of magnetic 'responsiveness' as magnetic microspheres

246

not only span a large range of sizes, but are also made from many different matrix materials incorporating different types and amounts of magnetic compounds [5]. An apparatus for measuring magnetophoretic mobility of the prepared magnetic microspheres is shown in Figure 4. This apparatus can directly measure the magnetic ‘responsiveness’.

Syringe Capillary

-

Magnet

Microscope lens’!

\

Vial with magnetic microspheres

Digital system for imaqe acquisition

rcp

0-

Figure 4. Schematic of the microscopy table for magnetic microsphere measurement.

4.

Results of the Human Supportive System I1 Magneto-Encephalography (MEG)

A history of modern technology for non-invasive human measurements begins with the discovery of X-rays by Roentgen in 1895. The next important historical milestone was the invention of computer tomography (CT) using X-rays by EMI Company in 1971. This is the origin of various graphic-imaging technologies for medical diagnoses. X-ray imaging technology visualizes internal organs in the human body using differences in absorption and permeability to the interaction of X-rays with bones and tissues of the human body. Since X-ray imaging, physical and chemical technologies have been applied in the fields of medical electronics, for example: functional magnetic resonance imaging ( M I ) using nuclear magnetic resonance technology and positron emission tomography (PET) using a positron created from a radioisotope with an extremely short

247

decay rate, etc. Thus, internal problems can be analyzed with perfect resolution because physical energy can be externally introduced to the human body. On the contrary, electrical measurements of the inner body, such as electro-encephalography (EEG), electrocardiogram (ECG), and so on, cannot resolve perfectly. They are called ill-posed inverse problems because these electrical phenomena are signals existing inside the human body. Magneto-encephalography (MEG) is also the same electrical measurement as is the EEG for a human brain, but was measured by using a superconducting quantum interference device (SQUID). Though the response of an EEG is dependent on each individual electric resistance of each tissue in the human body, MEG is not dependent on electric resistance. Therefore, MEG can better localize sources than can EEG Functional MRI and PET are suitable for imaging blood flow and metabolic changes in the brain, but not for time resolution. The best advantage of MEG imaging is millisecond time resolution for brain analysis. The first step of the study is to evaluate the extent of ‘satisfactory’, which may become a measure of QOL improvement, by studying the human senses. In this section, we give results on olfactory sense experiments. Historically, the human olfactory sense was measured objectively by using olfactory evoked potentials. Though a few studies of the olfactory evoked potentials have been attempted in order to obtain responding peaks for various odorants, the characteristics of these peaks are still unclear and debatable. Several papers on olfactory evoked potentials were reported in humans by averaging electroencephalography following odorant stimulation. Although olfactory-related potentials evoked by odorant pulse stimuli have been recorded from the human scalp, the neuromagnetic fields of olfactory cognitive response have not yet been measured. In our study so far, we investigated olfactory event-related potentials and the neuromagnetic fields evoked by odorant pulses synchronized with the subject’s respiration using a whole-cortex, 122-channel SQUID neuromagnetometer (Neuromag- 122, TM), and analyzed the olfactory MEG responses using an olfactory oddball paradigm. The initial goal of this study is to measure the human olfactory sense using MEG experiments and to analyze human olfactory perception and cognition non-invasively. An oddball paradigm has sometimes been used in visual and hearing experiments in psychological tests, however it has never been used in olfactory experiments. We used two odors in this experiment, one pleasant and the other unpleasant. In general, a targeting odor in the two odors is usually given at the lower stimulating rates randomly more than another, non-targeting

248

odor. In this oddball experiment, a subject counted only odors stimulated by the targeting odor. We call this targeting odor a “rare olfactory stimulant”, the other a “frequent olfactory stimulant.” We found the MEG responding peak first. Then we found a later component, which was suggested to be a recognition factor of the targeting odor only found in the present olfactory oddball task. The later component was not found in non-targeting odor response. In this paper, we discuss the nature between perception and recognition of the odor sensed. We found the perception of odor in the first experiment using one odor, and the recognition of odor in the second oddball experiment using two odors in the human olfactory system. The results of the present two MEG experiments suggested that we recognize the nature and contents of odorants in different regions of our brain after the odor is perceived.

Acknowledgment We would like to express our appreciation to the financial support by a Grant-in Aid of “Research for the Future Program” the Japan Society for the Promotion of Science.

References 1. Scientific and Clinical Applications of Magnetic Microspheres (1997), ed by: U. Hafeli, W. Schtt, J. Teller and M. Zborowski, Plenum Press, New York. 2. Takeda, S., Yu, S.J., Tari, I., Proceedings of the 5‘h Meeting of New Magneto-Science, TML, Ann. Rep. Suppl .II, 183, (2001). 3. Takeda, S., Suemoto, H., Tari, I., Proceedings of the 5” Meeting of New Magneto-Science, TML, Ann. Rep. Suppl. I, 98, (2001). 4. Jiang, Y., Miller, M.E., Hansen, M.E., et al., “Fractionation and size analysis of magnetic particles using FFF and SPLITT technologies”, Journal of Magnetism and Magnetic Materials, 53 (1999), pp. 194-199. 5. Arshady, R., “Microspheres, Microcapsules and Liposomes: Magneto-and Radiopharmaceuticals”, Vol. 3, 1st ed. Citus Books, London, 2001.

MAGNETIC ORIENTATION IN BIOLOGY: VIRUS STRUCTURE - BLOOD CLOT ASSEMBLY CELL GUIDANCE .ITORBET . University of Pennsylvania, Dept. of Cell & Developmental Biology Philadelphia 19104-6058 Present address: IBCP, CNRS UMR 5086, Lyon Cedex 7, France

Our childhood games with permanent magnets leave us with the impression that matter, in general, does not respond to a magnetic field. In reality, virtually everything is subjected to minute forces of attraction, repulsion or orientation. Strong fields combined with better understanding allow us to exploit these effects to tackle biological problems. In particular, the very weak diamagnetic anisotropy associated with individual molecules can give rise to high orientation of well organized structures such as crystals, liquidcrystals, semi-rigid polymers and individual cells. High orientation is often accompanied by better data and superior properties. In some circumstances, such as in crystallization, the orientating torque might induce effects over and above simple orientation. Magnetic field orientation has a number of advantages over other orienting techniques. Drawing or spinning produce fibers and can alter structure or cause damage while template methods invariable work only over a short range. The application of an electric field can cause heating and electrophoresis. In contrast, a magnetic field acts at a distance allowing uniform orientation in bulk and the creation of composites with components having different orientations. The contribution that magnetic orientation has made to a range of biological topics is illustrated by briefly describing a number of examples. For example, it has been a boon to x-ray studies of some non-crystalline filamentous complexes (e.g. fibrin, actin, microtubules, bacterial flagella and filamentous viruses) and is being vigorously exploited in NMR. The blood-clot polymer, fibrin, forms highly oriented gels when polymerized in a strong field and a number of its properties have been elucidated as a result. Magnetically oriented scaffolds of collagen, the major connective tissue protein, and fibrin are being used to study cell contact guidance. Oriented biomaterials might eventually be incorporated into specialized wound dressings, for example, to direct nerve repair.

1. Introduction Most molecules exhibit diamagnetic anisotropy - by far the weakest form of magnetic anisotropy. Nevertheless, this property is at the origin of the magnetic orientation of numerous biological complexes and synthetic polymers. The diamagnetic anisotropy, Ax, for objects having an axis of rotational symmetry, is given by the difference between the diamagnetic susceptibilities parallel and perpendicular to the axis of symmetry, Ax=x,,-xI.The degree of orientation in an applied field, H, is a function of Am2/kT (analogous to induced electric dipole orientation [ 1,2]). The magnetic anisotropies of macromolecules and even large filamentous viruses are so small that there is little prospect of significant 249

250 2

orientation in the absence of ordering interactions, IAxlH e k T . However, when N molecules behave in an ordered, cooperative fashion their individual magnetic anisotropies can additionally give an ensemble anisotropy of up to NxAx. For better than 80% maximum orientation, NlAxIHZ>17kT. Thus, the size of the unit ~ 2 . 5 ~ 1 0Daltons ~ undergoing orientation must be in excess of 7 ~ 1 0and (H=lO T) respectively for the filamentous viruses and fibrin polymer discussed below. Fortunately, ordered cooperative arrays exist in many guises such as in crystals, liquid crystals, macromolecular sheets, and gels. Conditions extant during the orientation process should favor cooperation without stifling rotation. For this reason viscous gels are unlikely to respond favorably during field exposure unless the transformation to the viscous state occurs in the field. For this reason, it is usually advisable to induce controlled assembly in the field. Samples that are naturally mechanically stable such as gels or rendered so by, for example, solvent removal can remain oriented indefinitely once withdrawn from the field. Orientation can also be "frozen in" by controlled gelation of an otherwise passive polymer. For some samples, the modest fields produced by permanent magnets or conventional electromagnets suffice to produce significant alignment, but on the whole stronger fields are necessary. A positive A x gives rise to orientation parallel to the field, whereas all orientations perpendicular to the field are possible when A x is negative. In the latter case, uniaxial orientation is nevertheless possible when additional forces, such as surface constraining forces, are present. A negative Ax can actually be used to advantage, for example, to form hollow tubes of circumferentially wound polymer [3]. Double orientation effects are possible with diamagnetically biaxial objects. In crystalline platelets of dipalmitoyl phosphatidylcholine both the hydrocarbon tails and the head groups orient perpendicular to the field because both regions have a negative Ax. The possible influence of this effect on lipid bilayer behavior seems to have been overlooked. Diamagnetism is an induced effect. There is no permanent moment so oriented samples have no preferred polarity (head-up and head-down are energetically identical). However, unidirectional orientation has been achieved with the purple membrane of Halobacterium halobium by briefly applying an electric field that, acting through the permanent electric dipole moment of the membranes, polarizes the system while the magnetic field maintains this condition until the system is stabilized by the gelation of a polymer [4]. Worcester [5] was the first to point out that the diamagnetic anisotropy of proteins largely depends on relative orientations of the peptide bonds (particularly in the form of aligned a-helices and to a lesser extent P-sheets) and

251

aromatic residues. Thus it is not easy to simply predict the sense of orientation except for specific cases such as collagen or for structures containing a high proportion of relatively aligned a-helices or P-sheets [5]. The diamagnetic anisotropy of nucleic acids is dominated by the negative AX of the polycyclic aromatic ring of the bases. Linear double stranded DNA thus tends to orient perpendicular to the applied magnetic field [6,7] although this is not necessarily true for supercoiled DNA [8]. Similarly, the sugar rings in polysaccharides have a negative A x so cellulose orients perpendicularly [9]. There are numerous examples of lipid orientation in the form of vesicles, discs, lipoprotein complexes, and crystalline platelets [10-12,381. Lipid bilayers usually preferentially orient perpendicular to the applied field direction, both in the fluid L a and gel phases because the fatty acyl chains have a negative A x and are preferentially extended perpendicular to the bilayer surface. The addition of detergents can improve the average orientation by breaking closed vesicles into discoid fragments. Some naturally occuning biomembranes orient with the perpendicular to the membrane surface running parallel to the field because the positive AX of the proteins dominates [5]. The channel forming hydrophobic peptide gramicidin can also flip the bilayer orientation when present at high enough concentration. In this case, the peptide bonds and aromatic groups probably have a concordant influence. Bilayer orientation has also been successfully altered by addition of aromatic ring amphiphiles, which have a A x opposite in sign and about an order of magnitude greater than that of phospholipids. However, the concentration of additive required is disappointingly high. The three broad topics outlined below illustrate the wide-ranging contribution magnetic Orientation has made to the study of biological systems. 2.

Structure

The degree of alignment of non-crystalline fibrous and membrane complexes is to their study by diffraction techniques as the quality of a crystal is to crystallography. Low-resolution neutron diffraction studies of retinal rod outer segments were the f i s t to show that structural studies could profit from magnetic alignment [ 131. Subsequent experiments with filamentous bacteriophages Pfl and fd, which undergo complete magnetic alignment in the liquid-crystalline state, demonstrated that this contribution could be decisive [14,15].The highresolution structure of these viruses is known thanks to the application of x-ray diffraction [161, neutron diffraction [171, solid-state NMR [18] and other techniques to magnetically oriented samples.

252

Diffraction studies of a diverse range of samples of biological origin have now benefited from magnetic orientation including bacterial flagellum, pamyloide protein, scrapie prion, fibrin, microtubules, lipid vesicles, light meromyosin, and some polysaccharides. Nh4R is, not surprisingly, the area in which magnetic orientation has made the most significant contribution. The spectra of oriented samples are characterized by single line resonances or simple multiplets. As these lines are narrow compared to those of a powder pattern, the effect of orientation on the spectra is dramatic. Oriented samples provide structural information in the form of measurements of the angles between bonds and chemical groups and the direction of sample orientation. The foremost application of solid-state NMR is in the study of lipid bilayers to which membrane proteins and other molecules are bound [ 191. The determination of the orientations of residues and domains of these proteins relative to the lipid bilayer requires macroscopically oriented samples. The small degree of molecular alignment induced by the field [20] or caused by the presence of a magnetically oriented liquid crystal [21] is also being used to improve the accuracy of structure determination using NMR. The possible beneficial effects of a magnetic field on protein crystallization are discussed elsewhere in this volume.

3. Biopolymer Assembly-Disassembly A number of biopolymers including actin, collagen, fibrin, microtubules and sickle cell hemoglobin have been highly aligned by assembling them in a strong magnetic field. Dynamic processes, such as assembly and disassembly of biopolymers in different conditions, can be investigated using magnetic Orientation jointly with another physical technique as illustrated by magnetic birefringence studies of fibrin [22-241. The enzyme thrombin liberates small fibrinopeptides from the blood plasma protein fibrinogen creating fibrin monomers, which assemble into double-stranded fibrin protofibrils; long protofibrils then zip together laterally forming a 3-dimensional gel. A number of proteins with vital haemostatic, lytic and wound repair functions are sequestered within the fibrin fibers. When fibrin assembly is provoked in a strong magnetic field the resulting gels are composed of fibrin fibers perfectly aligned parallel to the field direction. The extent of orientation is fixed at about the time of gelation, which occurs at about 15% fibrin formation. The entire assembly-lysis cycle can be readily studied with magnetically induced birefringence, as the signal is simply proportional to the fibrin concentration [23] even in blood plasma [24]. Fibrin is only a minor component by weight in plasma, but orients significantly

253

while other plasma components are rendered “invisible” because they either orient little or have stabile orientation [38]. Using this approach it has been shown that the thrombin generation pathway has a significant impact on clot structure, assembly kinetics and thrombolytic properties [24]. These studies show that it is possible to investigate various properties of a component in a complex system by exploiting differences in magnetic orientability. 8 6

An 4

2

0

0

50 (Magnetic Field)

100 (tesla)2

150

Figure 1. An is the magnetically induced birefringence (multiplied by 10’) measured from a suspension of free-swimming E. coli bacteria (MM103). Saturation in An indicates near complete orientation.

4. Biomaterials and Cells

Collagen has been incorporated into a number of health and cosmetic products and a fibrin based biological glue or sealant is used to stop bleeding and promote tissue repair. Collagen and fibrin both participate in wound repair and are oriented when assembled in a strong magnetic field [22-25,36,38]. Fibrin orients along the field whereas collagen orients perpendicular to the field. However, near uniaxial orientation is obtained by exploiting surface constraining effects. These samples can be dried and rewetted without loss of either orientation or capacity to support cell growth. Many cell types, including fibroblasts [26], keratinocytes, osteoblasts [29], neurites [27], endothelial cells undergoing angiogenesis [31], Schwann [30] and smooth muscle cells, align when cultured on magnetically oriented collagen or fibrin. As both the cells and the culture medium are oriented, the interplay between them can more readily be observed. Thus, it is anticipated that new

254

questions can be addressed regarding cell contact guidance, substrate remodeling and extracellular matrix deposition (ECM). Oriented bio-scaffolds might also have use in tissue engineering by acting as a template to direct and organize cell growth and ECM deposition. Indeed, it has been demonstrated that nerve guide tubes containing magnetically aligned collagen improve peripheral nerve regeneration in mice [28]. Some cells or cellular components can also be oriented in a magnetic field either in suspension or when cultured on unoriented substrates. The former category includes retinal rod outer segments, chloroplast membranes [40], blood cells [37], bull sperm [35] and the bacteria Escherichia coli (Figure l), while osteoblasts [29] and smooth muscle [32,33] and Schwann [30] cells align when cultured in a strong magnetic field. Pollen tubes also follow the field direction [341. A well-developed supramolecular organization of mutually aligned components is essential for the magnetic torque to have an effect. This condition is satisfied by the cytoskeletal components in nerve (microtubules and neurofilaments) and muscle (actomyosin) cells. However, the cytoskeleton is not the only potential source of magnetic anisotropy. The layers of membranes in chloroplast [40] and retinal rod outer segments [ 131 give rise to high orientation in modest fields. Also the polysaccharide and other components of plant and bacteria walls [39] can also make a significant contribution to the anisotropic. Acknowledgements This work was partly supported by NIH grant HL30954. I would like to thank Prof. J. W. Weisel for his support and encouragement. References

1. O’Konski, et al., Electric properties of macromolecules. IV. J. Phys. Chem. 63 (1959), pp. 1558-65. 2. Shah, M.J., Electric birefringence of bentonite 11. An extension of saturation birefringence theory J. Phys. Chem. 67 (1963), pp. 2215-19. 3. Tranquillo R.T., et al, Magnetically-oriented tissue-equivalent tubes: application to a circumferentially-orientedmedia-equivalent Biomaterials 17 (1996), pp. 349-57. 4. Dtr, A., et al, Orientation of purple membrane in combined electric and magnetic fields, FEBS Letters 377 (1995), pp. 419-20. 5. Worcester, D.L., Structural origins of diamagnetic anisotropy in proteins, Proc. Nat. Acad. Sci. USA, 75 (1978), pp. 5475-77. 6. Senechal, E., Maret, G., Dransfeld K., Long-range order of nucleic acids in aqueous solutions, Int. J. Macromol. 2 (1980), p. 256

255

7. Maret, G. Weill, G., Magnetic birefringence study of the electrostatic and intrinsic persistence length of DNA Biopolymers 22 (1983), pp. 2727-44 8. Torbet J., Solution behaviour of DNA studied with magnetically induced birefringence, In Methods in Enzymology "DNA Structures" edit. D.M.J. Lilley and J.E. Dahlberg (1992) 211, pp. 518-32. 9. Sugiyama, J. Chanzy H. Maret, G., Orientation of cellulose microcrystals by strong magnetic fields, Macromolecules 25 (1992), pp. 4232-34. lO.Rosenblatt, C., Yager, P., Schoen, P.E., Orientation of lipid tubules by a magnetic field, Biophys. J. 52 (1987), pp. 295-01. 1l.Prosser, R.S., Hwang, J.S., Regitze, R.V., Magnetically aligned phospholipid bilayers with positive ordering: a new model membrane system, Biophysics J . 74, pp. 2405-18. 12.Sakurai, I. et al., Magneto-orientation of lecithin crystals, Proc. Natl. Acad. Sci. USA 77 (1980), pp. 7232-36. 13.Saibi1, H., Chabre, M., Worcester, D.L., Neutron diffraction studies of retinal rod outer segment membranes, Nature 262 (1976), pp. 266-70. 14.Torbet, J., Maret,.G., High field magnetic birefringence study of the structure of rodlike phages Pfl and fd in solution, Biopolymers 20 (1981), pp. 265769. 15.Torbet, J., Maret, G., Fibres of highly oriented Pfl bacteriophage produced in a strong magnetic field, J. Mol. Biol. 134 (1979), pp. 843-45. 16.Nave, C., et al., Macromolecular structural transitions in Pfl filamentous bacterial virus, Nature 281 (1979), pp. 232-34. 17.Nambudripad, R., Stark, W., Makowski, L. Neutron diffraction studies of the structure of filamentous bacteriophage Pfl, J. Mol. Biol. 220 (1991), pp. 35979. 18.Zeri, A.C. et al., Structure of the coat protein in fd filamentous bacteriophage particles determined by solid-state NMR spectroscopy, Proc. Nut. Acad. Sci. USA 100 (2003), pp. 6458-63. 19.Marassi, F.M., Opella, S.J. Simultaneous assignment and structure determination of a membrane protein from NMR orientational restraints, Protein Sci. 12 (2002), pp. 403-11. 20.Tjandra N. et al., Use of dipolar 'H"N and 'HI3C couplings in the structure determination of magnetically oriented macromolecules in solution, Nature Structural Biology 4 (1997) p. 732. 21.Tjandra N., Bax, A., Direct measurement of distances and angles in biomolecules by N M R in a dilute liquid crystalline medium, Science 278 (1997), pp. 1111-14. 22.Torbet, J., Freyssinet, J.-M. and Hudry-Clergeon, Oriented fibrin gels formed by polymerisation in a strong magnetic field G, Nature 289 (1981), pp. 9193* 23.FreyssineC J.-M., et al, Fibrinogen and fibrin structure and fibrin formation using magnetic orientation, Proc. Nut. Acad. Sci. USA 80 (1983), pp. 161620.

256

24.Torbet, J., The thrombin activation pathway modulates the assembly, structure and lysis of human plasma clots in vitro, Throm. and Huemos. 73 (1995), pp. 785-92. 25.Torbet, J., Ronziere, M.C., Magnetic alignment of collagen during selfassembly, Biochem. J . 219 (1984), pp. 1057-59. 26.Guid0, S., Tranquillo R.T., A method for the systematic & quantitative study of cell contact guidance in oriented collagen gels, J. Cell Sci. 105 (1993), pp. 3 17-31. 27.Dubey, N., Letourneau, P.C., Tranquillo, R.T., Guided neurite elongation and Schwann cell invasion into magnetically aligned collagen in simulated peripheral nerve regeneration, Exp. Neurol, 158(2), (1999) pp. 338-50. 28.Ceballos D., et al., Magnetically aligned collagen gel filling a collagen nerve guide improves peripheral nerve regeneration, Exp. Neurol, 158(2), (1999) pp. 290-300. 29.Kotani, H. et al., Magnetic orientation of collagen and bone mixture, J. Appl. Physics 87 (2000), pp. 6191-93. 30.Eguchi, Y., Ogiue-Ikeda, M., Ueno, S. Control of orientation of rat Schwann cells using an 8-T static magnetic field, Neuroscience Letters 351 (2003), pp. 130-32. 3 l.Torbet, J., Tranqui, L., Bureau, C., Magnetic processing of biological and synthetic polymers and crystals, The 3d International Symposium on Electromagnetic Processing of Materials, (2000) pp. 133-37. 32.Iwasaka, M., Miyakoshi, J., Ueno, S., Magnetic field effects on assembly pattern of smooth muscle cells, In Vitro Cell. Dev. Bio1.-Animal 39 (2003), pp. 120-23. 33.W. Arnold, R. Steele and H. Mueller On the Magnetic Asymmetry of Muscle Fibers, Proc. Nat. Acud. Sci. USA 44 (1958), pp. 1-4. 34.D. Sperber, K. Dransfeld, G. Maret and M.H. Oriented growth of pollen tubes in strong magnetic fields, WeisenseelNuturewissenchaften 68 (198 1), pp. 40-41. 35.Emura, R., et al., Orientation of bull sperms in static magnetic fields, Biolectromugnetics 22 (2001), pp. 60-5. 36.Murthy, N.S. Liquid crystallinity in collagen solutions and magnetic orientation of collagen fibrils, Biopolymers 23 (1984), pp. 1261-1267. 37.Higashi, T., Ashida, N., Takeuchi, T. Orientation of blood cells in static magnetic field, Physicu B 237 (1997), pp. 616-20. 38.Torbet, J., Magnetic birefringence study of fibrin formation & chylomicron behaviour in human blood plasma, In 'Biophysical Effects of Steady Magnetic Fields' Springer-Verlag Pub. (1986) p. 23. 39.Torbet, J., Norton, M.Y., Structure of the cell wall of Staph. aureus with neutron scattering and magnetic birefringence, Febs Letters 147 (1982), pp. 201-06. 40.Geacintov, N.E., et al., Magnetic field induced orientation of photosynthetic systems, Biochem. Biophys. Actu. 267 (1972), pp. 65-79.

MANIPULATING CELLS WITH STATIC MAGNETIC FIELDS* J.M. VALLES, JR. and K. GUEVORKIAN Department of Physics, Brown University, Providence, Rhode Island 02912, USA We review our investigations of the use of static magnetic fields, B , for manipulating cells and cellular processes. We describe how B fields modify the cell division pattern of frog embryos and consequently can be used to probe the pattern determinants. We also observe that magnetic fields modify the swimming behavior of Paramecium Caudatum. We describe these modifications and their potential application to investigations of their swimming behavior.

1. Introduction The materials in cells are primarily diamagnetic, thus they respond weakly to magnetic fields. Nevertheless, magnetic fields are available that are sufficiently intense to align biopolymers, such as microtubules, and biomolecular assemblies, such as cell membranes [l]. In addition, common organic materials can be magnetically levitated [ 2 ] .We are exploiting these weak intrinsic responses to manipulate cellular processes. Our effort is one of a growing number that apply static magnetic fields to diamagnetic systems to address fundamental scientific issues. Other examples include the materials science of protein crystal growth [3], biopolymerization [4],the gravitational sensitivity of plant growth [5,6],and fundamental fluid dynamics [7]. Here, we describe the application of static magnetic fields to manipulate two systems, frog embryos, Xenopus laevis, and the single cell protozoa, Paramecium Caudatum. We have found that magnetic torques can alter the cell division geometry in frog embryos and we used this response to investigate the endogenous factors that control the nominal geometry. In work supported by NASA, we applied static magnetic force fields on swimming Paramecium to vary their effective buoyancy and thus, gravity response. Unexpectedly, we found that Paramecium experiences a net torque in magnetic fields that aligns their swimming direction.

has been performed in collaboration with Professor Kimberly Mowry, Dr.James Denegre, Sarah Wasserman, Caterina Schweidenback,Jill Edwardson, Kevin Lin, Juliet Liu, and Carl Quindel * This work

25 7

258

2.

The Principles

Consider a diamagnetic object, such as a cell, placed above the center of a current carrying solenoid oriented with its axis vertical (see Figure 1). The inhomogeneous magnetic field, B, and gravity exert forces on the cell so that the total force per volume experienced by the cell is f = -pg+XBB’. Here, p is the cell density, g is the gravitational acceleration and 2 is the diamagnetic susceptibility. As indicated in Figures l(a) and (b), it is possible to balance these two forces to produce a levitated state; a simulated low gravity state. For typical biological materials, levitation requires BB’ = 1450 T2 m-’. Figure l(b) shows the magnetic force and total force profiles for a solenoid magnet. The arrows indicate the direction and magnitude of the total force acting on the diamagnetic objec:tin different position along the solenoid axis.

Figure 1. (a) Schematic of a diamagnetic object experiencing two forces in opposite directions;fg is the gravitational force per volume mdfS is the magnetic force per volume. (b) The magnetic force and the total force cf) profiles; the arrows represent the total force direction.

For cells in solutions, we must account for buoyant forces, as well. In this case, the total force per volume acting on the cell is given by:

The first term on the right side of Eq. (1) is the buoyant force (subscripts c and s represent the cell and the solution, respectively) and the second term is the total magnetic force. Note that xC is the volume average of the magnetic susceptibilities of the constituents of the cell. To measure the extent by which the magnetic force changes the buoyancy of the cell, we define the effective

259

buoyancy as the ratio of the net force on the cell with magnetic field to the net force without, i.e,

(2)

6

According to Eq. ( 2 ) , we can make b,ff/b Yellow Si(51), Pb(40), Ca(6), K(l) 3.208 ................................................................................................................................................................................................................................................................< Blue Si(52), Pb(38), Ca(6). K(l), Co(O.4) 3.171 ................................................................................................................................................................................................................................................................1 Black

Si(46), Pb(34), Mn(ll), Ca(6), K(l)

3.199

x

[ x 10-61 -13.3 -9.27 -2.20 317

f-,

Figure 1. Magneto-Archimedes separation of colored glass particles. (a) at zero magnetic field, (b) at 8.0 T in the center of the field, (c) 9.5 T,(d) 12.0 T.

272

Colored glass particles in a manganese chloride solution were separated to demonstrate the magneto-Archimedes separation technique. Glass is usually colored by doping with impurities. Therefore, glasses with different colors should differ in density and in magnetic susceptibility, so they should be able to be separated using the magneto-Archimedes technique. In our experiment, glass particles with four different colors, red, blue, yellow and black, were used (SATAKE GLASS CO., LTD.). The components of these materials, examined by an energy dispersive X-ray spectroscopy analyzer (JEOL, JSX-3220), are shown in Table 1. The percentage of each element was estimated by the bulk fundamental parameter method. Their densities and magnetic susceptibilities were measured using a pycnometer and a SQUID (QUANTUM DESIGN, MPMS-5), respectively, and the results are also shown in Table 1. Glass fragments were riddled by a sieve with 1 x 1 mm2 mesh in advance, and smaller particles were used. Each of the glass particle aggregates of about 0.2 grams were put into a cylinder filled with a 6.0 wt% MnC12 aqueous solution (x = +81.4 x lo", p = 1.052 x lo3 kg/m3), then put in an ultrasonic bath for about a minute in order to detach some air bubbles adhering to the particle surface. The cylinder was set so that the position of the bottom of it was at the maximum point of IB,dB/dzl.The behavior of the colored glass particles was observed as the magnetic field increased. Figure 1 shows the photographs of magnetoArchimedes separation of colored glass particles taken under several magnetic fields. At the beginning, under zero magnetic field, all of the colored glass particles are in the bottom of the cylinder as shown in Figure l(a). As the magnetic field increased, glass particles were gradually floated in the solution. When the magnetic field in the center exceeded 8 T, red colored particles were floated (Figure l(b)). Yellow and blue particles floated in succession between 9.0 to 9.5 T (Figure l(c)). Even at 12 T, black colored glass was not levitated as expected. The values of B,dB/dzneeded to levitate red, yellow, and blue particles were 205, 298, and 325 T2/m, respectively. These values were good agreement with those calculated. Comparing Figures 1 (c) and (d), the distance between glass flocks was different, i.e., the distance was larger in (c) than in (d). This is due to the difference of spatial distribution of B.dB/dz.Around the maximum point of B.dB/dz,z = 113 mm, B.dB/dzvaries gently compared with the higher region. Thus, the resolution of magneto-Archimedes separation becomes high in this area. In this experiment, the order of the required absolute value of B,dB/dzcorresponds to that of magnetic susceptibilities. Judging from this result, we can say that the equilibrium position is mainly determined by the magnetic susceptibility, in this case. In some other experiment, such as a

273

separation of KCl and NaCl powder mixture [4],the equilibrium position mainly results from the difference in density rather than that in magnetic susceptibility. In magneto-Archimedes separation, one can choose which factor should be emphasized, the difference of magnetic susceptibilities or the difference of densities by selecting proper surroundings. The colored glass used in the previous experiment was made for arts and crafts use. The clear art glass contained more impurities than glass manufactured for practical use. This caused the relatively large differences in magnetic susceptibilities and densities among the samples and simplified their magnetoArchimedes separation. The application for Figure 2. Magneto-Archimedes separation of practical glasses. The materials recycling is intriguing if glass liquid medium used in this manufactured for practical use such as bottles, experiment is 5.27 wt% MnClz tableware or glass windows can be color aqueous solution. separated by the magneto-Archimedes technique, as separation by color is already required for present recycling programs. We prepared several kinds of glassware used in daily life and evaluated whether color separation is possible by the magneto-Archimedes technique. Almost all of the sample glasses used in this experiment had similar densities. However, owing to the slight differences of their magnetic susceptibilities, it was confirmed, as shown in Figure 2, that color separation of practical glassware was possible with the aid of the magneto-Archimedes technique. MagnetoArchimedes separation seems to be one beneficial method for materials recycling.

3. Interactions Among Magnetic Dipoles Induced in Feeble Magnetic Substances under High Magnetic Fields Recently, effects of magnetic fields on feeble magnetic substances have been vigorously investigated and various phenomena have been discovered. These effects are mainly based upon magnetic forces, and magnetic forces can be expressed as interactions between feeble magnetic substances and gradient fields. Conversely, interactions among feeble magnetic substances under magnetic fields have been neglected, so far. In ferromagnetic substances, and the interactions through their magnetic dipoles can be observed clearly [5,6], energy of the interactions of two magnetic dipoles is expressed as

274

where p,, is the permeability of vacuum, ma and mb are the magnetic dipoles, and r and r are the vector between two dipoles and its distance, respectively. On the other hand, for feeble magnetic substances, magnetic dipoles are induced only under magnetic fields and their values are extremely small. Therefore, the energy of the interactions is too small for casual observation. However, through elaborate experiments using high magnetic fields of several teslas, we confirmed that such interactions could be observed visually, even in feeble magnetic substances. It was then confirmed that utilizing the interactions could control alignments or structures of feeble magnetic substances. Here, we report the observations of the induced magnetic dipole interactions and basic research of the structure control of feeble magnetic substances.

Figure 4. Chain-like alignments of glass beads in 40 wt% MnC12aq.The magnetic field was directed from left to right, and its intensity was 2.5 T at the field center.

First, we observed induced magnetic dipole interaction between two objects. In this study, the magnet was placed vertically and the field direction was parallel to that of Figure 3. Repulsive interaction of gravity. Palladium (paramagnetic, volume ualladium cvlinders under a magnetic susceptibility y , = 7.78 X [in SI magnetic field of 6 T. units]) rods were used as samples in this experiment. The rods were 1.0 mm in diameter and 5.0 mm in height. Sample rods were held side by side, with some spacing (- 1.0 mm), by polyester fibers in the bore of the magnet. The area where the magnetic field is almost flat in the horizontal direction was selected as the samples’ position in order to avoid horizontal magnetic force effects. From this configuration, the magnetic field was increased gradually and the distance between samples was observed. Figure 3 shows the result.

275

The upper figure shows the initial state, and the lower shows the state with a 6 T magnetic field. From this result, we see that the distance between samples was increased about 0.4mm by the application of the 6 T magnetic field. This could be because the samples repelled each other through induced magnetic dipoles. Quantitative analysis was then performed. From the experimental result, the forces derived from the dipole interaction were estimated to N. However, this value be 8.5 X of the force can also be calculated. The magnetic field around the samples was spatially distorted due to the induced magnetic dipoles, and a computer calculated the distribution. Figure 5. Triangle-lattice alignments of gold balls in 40 wt% MnC12aq. From the result, the value of magnetic The magnetic field direction was forces acting in the horizontal perpendicular to this space, and its direction was found to be 2.4X intensity was 4.9 T. The lower figure is a close-up view. N. The experimental result and the calculated value were in substantial agreement. From these results, we see that magnetic dipole interactions can be observed even in feeble magnetic substances by controlling experimental conditions carefully, and we also confirmed that such interactions can be enhanced by considering the magnetoArchimedes effect. Subsequent experiments used many feeble magnetic particles. It is known that magnetic dipole interactions lead dispersed particles to some ordered alignments. However, only systems containing ferromagnetic substances have been considered so far, and such applications have been restricted to a few materials. The applications of this effect to systems of feeble magnetic substances would make it possible to control structures of various materials, which would be useful in material processing. Therefore, experiments to examine such possible applications were performed.

276

First, alignments parallel to magnetic fields were observed. The sample particles used in this experiment were glass beads (-0.8 mm@, which are diamagnetic and have volume magnetic susceptibility of -1.8 X 10.’ [in SI units]. Manganese dichloride aqueous solution of 40 wt% was used as a medium in consideration of magneto-Archimedes effect. In this experiment, the same magnet used previously was set horizontally. The glass beads and MnC12aq were put in a glass cell, which was inserted into the bore of the magnet. One of the cell edges was fixed at the center of the field, and the glass beads were initially gathered at that side. From this configuration, the magnetic field was increased gradually. Then, magnetic forces acted on the glass beads and they moved to the other side of the cell to avoid the higher field region. These processes were observed from the bottom of the cell with a CCD camera. The result of these experiments is shown in Figure 4. In this figure, the magnetic field was applied parallel to the space directed from left to right, and its intensity was 2.5 T. As seen in this figure, the glass beads aligned in chain-like formations parallel to the applied field as they moved away from the center of the field. These formations were derived from the attractive interactions among magnetic dipoles induced in the glass beads. Alignments perpendicular to the field were then observed. In this experiment, the magnet was placed vertically and a petri dish, containing the sample particles and surrounding medium, was placed in its bore. The sample particles used in this case were gold balls 1.0 mm in diameter; 40 wt% MnClzaq was used as the surrounding medium. The gold balls were positioned 149 mm above the field center, where magnetic fields were only slightly larger (-0.2 %) at the wall side than at the middle of the bore. From this configuration, a magnetic field was applied and the two-dimensional alignments of the balls observed from above. To clearly observe the forces derived from dipole interactions in the horizontal direction, the magnetic field intensity was adjusted and the apparent weights of the gold balls were set to zero, utilizing vertical magnetic forces. That is, magneto-Archimedes levitation was applied [2]. Figure 5 shows the magneto-Archimedes levitation state of gold balls observed from above. During levitation, the gold balls gathered to the middle of the bore, influenced by slight radial magnetic forces. However, the gathering balls were not closely packed and formed triangle-lattice alignments with some spacing. This formation of the lattices was caused by the repulsive interactions among each magnetic dipole induced in the gold balls. Alignments perpendicular to the field can be controlled by repulsive interactions among induced magnetic dipoles.

277

Therefore, formations of ordered alignments can be obtained not only in the systems containing ferromagnetic substances, but also in the systems of feeble magnetic substances by controlling experimental conditions properly. 4.

Conclusions

The magnetic separation of feeble magnetic particles or grains demonstrated here could be readily achieved with the aid of the magneto-Archimedes principle by using an ordinary superconducting magnet. This separation technique results from the difference of density and magnetic susceptibility of materials, and should enable one to quickly separate a powder mixture composed of more than one type of substance into clusters. This method is sensitive to both the density and magnetic susceptibility of materials. This technique will be beneficial in separating feeble magnetic materials. A practical application of this is the color separation and subsequent recycling of consumer glass using the magneto-Archimedes technique. Interactions among magnetic dipoles induced in feeble magnetic substances were observed in this study. Usually, such interactions are too small to be observed and have been so far neglected. However, by controlling experimental conditions carefully, we succeeded in observing the interactions. Furthermore, by applying the interactions to many-particle systems, some ordered alignments, such as chain-like and triangle-lattice alignments, were observed. These results show that structures of systems of feeble magnetic substances can be controlled by this interaction. Therefore, these phenomena suggest new applications of magnetic fields to various fields such as material processing. References 1 Beaugnon, E., and Tournier, R., Nature, 349, (1991) p. 470. 2 Ikezoe, Y., Hirota, N., Nakagawa, J., and Kitazawa, K., Nature 393 (1998) pp. 749-750. 3 Ikezoe, Y., et al., Trans. Mat. Res. SOC. of Jpn., 25 (2000) pp. 77-80. 4 Ikezoe, Y., et al., Proceedings of International Symposium on New MagnetoScience '99 Omiya, Japan, 2000, pp. 400-404. 5 de Gennes, P.G, Pincus, P.A., Phys. Kondens. Materie 11 (1970) p.189. 6 Skjeltorp, A.T., Phys. Rev. Lett. 5 1 (1983) p. 2306.

APPLICATION OF MAGNETIC LEVITATION TO PROCESSING OF DIAMAGNETIC MATERIALS I. MOGI, K. TAKAHASHI, S.AWAJI, K. WATANABE, M. MOTOKAWA Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan Magnetic levitation of diamagnetic materials under a gradient of strong magnetic fields was applied to a crystal growth solution and while melting levitated materials. Heterogeneous nucleation was considerably suppressed in W C 1 crystal growth in a levitating droplet, and dendrites grew along the droplet surface. In the crystal growth solution using a container, the W C l crystal was levitated in the solution with no contact between the walls and the solution surface. Containerless melting and solidification of paraffin was performed with the magnetic levitation furnace, and Marangoni convection was observed in the melt droplet. Furthermore, heat transfer behavior in water under gradient magnetic fields is reported.

1. Introduction Diamagnetic materials receive a repulsive force in a magnetic field gradient. When this magnetic force balances with gravity, it is possible to levitate these materials [l-51. Such a balance holds for each molecule constituting the materials, hence magnetic levitation provides a condition almost equivalent to microgravity. Its application to material synthesis enables many novel techniques such as containerless crystal growth. The containerless technique provides a clean, contamination free environment. Suppressing uncontrollable heterogeneous nucleation, the liquid can easily realize a supercooled or supersaturated state. The first application of magnetic levitation to crystal growth was the solidification of a water droplet [ 6 ] .As the first topic of this paper, we show the dendritic crystal growth of ammonium chloride in a levitating aqueous solution [7]. The containerless technique has been applied to melt growth of levitating materials. Levitating melting glasses was performed without a crucible by using a hybrid magnet and a CO2 laser. This apparatus is known as a “magnetic levitation furnace” [8].The second topic here is Marangoni convection in a melt droplet of paraffin in the magnetic levitation furnace. One advantage of the microgravity condition is the repression of thermal convection that must be responsible for crystal quality. However, the thermal behavior of diamagnetic liquids has not yet been studied under the condition of magnetic levitation. The last topic of this paper is control of thermal convection in water using gradient magnetic fields. 278

279 2.

Experiments

We used hybrid magnets producing up to 23 T or 28 T for magnetic levitation. The hybrid magnet consists of an inner water-cooled magnet and an outer superconducting magnet. When a material is inserted to the magnet bore, the force F acting on the unit mass of the material is given as

where x is the magnetic susceptibility per the unit mass of the material, A the vacuum permeability, B a magnetic flux density at the distance z from the center of the magnet along the vertical direction and g the gravity constant. The potential energy U per unit mass of the material at the position z is

u = - (1/2#&)x B 2 + gz + c , where C is a constant representing the potential energy at z = 0. The potential energy at the position holding F = 0 must increase in all directions so that the material levitates stably. An aqueous NH&1 solution was saturated at 12OC (25.4 wt%) and placed on the vertical axis of the hybrid magnet which generated a field at the magnet center of B, = 18.1 T. The magnetic susceptibilities of NH4Cl and water are 8 . 6 2 ~ 1 0 . ~[9] and -9.07~10-~m3 kg-' [lo], respectively; thereby the susceptibility of the saturated solution is estimated to be -8.95~1O-~rn~ kg-'. The potential energy curve of the solution has a minimum at z = 0.078 m in the vertical direction, at which the curve has a minimum at the horizontal center, thus the solution stably levitates in the magnet. The aqueous solution was introduced into the magnet bore through a silicon tube by an injector after the magnetic field reached 18.1 T, and a droplet of the solution levitated at z = 78 nun on the tip of a glass capillary. The temperature within the magnet bore was controlled by circulating thermoregulated water. Growth of the N h C 1 crystal was performed by decreasing the temperature, and the droplet side view was observed with a micro-CCD camera placed in the magnet bore. The magnetic levitation furnace was used in the melting experiment of levitating paraffin. The CO2 laser was placed near the hybrid magnet and the laser light was introduced into the magnet from the top of the bore, irradiating the upper side of the levitating sample.

280

3.

Crystal Growth in a Levitating Solution Droplet

Figure 1 shows the crystal growth of W C l in a levitating droplet (7 rmn diameter) of the solution. The temperature is decreased from 25OC through the saturation temperature (12OC) down to 4.8"C with a rate of -0.3OC min-'. The supersaturated state survives down to 4.8"C, and then a small, cross-shaped nucleus appears at a certain place in the upper hemisphere. It moves downward to the bottom of the droplet (Figure l(a)) and grows into a dendrite (Figure l(b)). The dendritic growth appears two-dimensional along the surface of the droplet, as shown in Figure l(c).

Figure 1. Crystal growth of NI-bCl in a levitating droplet (7 mm diameter) of aqueous solution in a hybrid magnet generating 18.1 T at the magnet center. The temperature within the magnet bore is 4.8"C: (a) A cross-shaped nucleus appears and is moving to the bottom of the droplet: (b) The nucleus is growing into a dendrite around the bottom at 5 min after (a): (c) The dendrite is growing along the surface of the droplet at 6 min after (a).

This result shows two characteristics of containerless crystal growth: nucleation and growth directions. In the case of crystal growth in a container, heterogeneous nucleation is induced by container walls and is uncontrollable. The above result demonstrates that the containerless crystal growth in the magnetic levitation condition eliminates such heterogeneous nucleation and reduces the number of growing crystals in the solution. On the other hand, the droplet surface considerably affects the growth direction of the crystal. Once the crystal sediments to the bottom of the droplet, the dendrite grows along the surface. This is partly due to the evaporation of the water and partly due to peculiarity of a liquid-gas interface, suggesting that mass transport at the interface would be quite larger than that in the bulk solution.

281

Why does the W C 1 crystal sediment in the levitating solution? Taking into account the magnetic force, gravity and buoyancy, the force Fcvs acting on the unit volume of the crystal is given as

-13.2~10-~ and xwl= where x i s magnetic susceptibility per unit volume (xcvs= -10.2~10-~) and p is density (pcvs= 1 . 5 3 ~ 1 0and ~ psol=1 . 1 3 ~ 1 0kg ~ m-5. The suffixes “crys” and “sol” represent the crystal and the solution, respectively. The force Fcrysis composed of magnetic and gravitational terms. The former is upward force, and the latter is downward. When the solution is levitating, the gravitational term is larger than the magnetic term, thus the crystal sediments to the bottom. 4.

Wall-Contact-Free Crystal Growth in a Solution

Here we consider the particular state that Fcrys= 0 in Eq. (3). Balancing the magnetic and gravitational terms means that the crystal does not sediment to the bottom nor does it float on the solution, implying that the crystal is “levitating” in the solution. Even in the case of crystal solution growth in a container, this allows wall-contact-free crystal growth. We have tried to conduct such an experiment. The m C 1 aqueous solution was saturated at 20°C (27.1 wt%), and a quartz cell (10xlOX30 mm3) was used as a container. When the hybrid magnet generates B, = 19.1 T, the position where Fcrys= 0 is estimated as z = 0.076 m. Thereby, the bottom of the cell was placed at z = 0.074 m in the magnet bore. A seed crystal was inserted to the solution, and the temperature was kept at 22°C. The result is shown in Figure 2. The seed crystal stays at the bottom of the cell below B, = 19 T (Figure 2(a)). When the magnetic field reaches B, = 19.1 T, the seed crystal moves to the horizontal center where the horizontal potential energy is minimum, and it levitates at z = 0.076 m in the solution (Figure 2(b)). When decreasing the temperature from 22°C to 17”C, a dendrite develops from the levitating seed without contact with the walls (Figure 2(c)). This result demonstrates that the magnetic force allows control of the crystal position in the solution.

282

(4

&I

(4

Figure 2. Levitating crystal growth of W C l from the aqueous solution in a quartz cell placed in a hybrid magnet: (a) A seed crystal lies at the bottom of the cell at 18.9 T (at the magnet center): (b) The seed crystal moves to the horizontal middle position and is levitating when the magnet generated 19.1 T at the center: (c) Decreasing temperature from 22°C tol7"C, a dendrite grows from the levitating seed.

5. Magnetic Levitation Furnace

Magnetic levitation also allows containerless melting and solidification of diamagnetic materials. Figure 3 shows the results for paraffin. A paraffin cube is levitated in the water-cooled magnet with B, = 14.1 T (Figure 3(a)). A COz laser light melts it to a liquid droplet (Figure 3(b)). After turning off the laser, the droplet is solidified into a sphere (Figure 3(c)). In the liquid state, surface convection rotating around the horizontal axis was observed by adding carbon powder as a marker. This is not gravitational convection but Marangoni convection, as the gravitational force is cancelled by the magnetic force. The laser irradiation at the top of the droplet causes a thermal gradient, inducing surface tension differential between the top and bottom. Such a surface tension gradient causes Marangoni convection in the levitating droplet.

(4

(b>

(c>

Figure 3. Containerless melting and solidification of paraffin in the magnetic levitation furnace: (a) A levitating paraffin cube in the 14.1 T field at the magnet center: (b) The melt droplet of paraffin during the COTlaser irradiation: (c) The solidified sphere of paraffin after turning off the laser.

From the viewpaint of materials processing, magnetic orientation is one of the advantages for the use of high magnetic fields. Kimura et al. reported that

283

the magnetic orientation of paraffin occurs in a 9 T field [ll]. In the above experiment, however, Marangoni convection prevents the magnetic orientation. The magnetic susceptibility of the paraffin sphere was measured, and it was isotropic. To accomplish magnetic orientation in the melt growth of levitating materials, we improved the laser furnace with more homogeneous irradiation of YAG laser light and obtained an oriented paraffin sphere [ 121.

6. Control of Thermal Convection in Water We succeeded in visually observing heat transfer in water placed in the hybrid magnet, using a liquid crystal sheet with thermochromism, as shown in Figure 4 [13]. The upward thermal convection around a heater was drastically suppressed by the magnetic force, however, the residual upward convection remained even under the levitation condition with the gradient field B(dB/dz) = -1360 T’m-’ (Figure 4(a)). The temperature dependence of the magnetic susceptibility of water is responsible for the convection behavior in the gradient field. A thermal conduction state without the convection was realized in the stronger gradient field B(dB/dz) = -2880 p m - ’ at temperatures 35-40°C (Figure 4(b)). Furthermore, it is surprising that the downward magnetic convection was observed in the same gradient field at higher temperatures 4550°C (Figure 4(c)).

Figure 4. Heat transfer in water (a) under the magnetic levitation condition B(dB/dz) = -1360 T’m-’ in a temperature range 35-40°C, @) B(dB/dz)= -2880 T2m-’at 35-40”C, and (c) B(dB/dz)= -2880 T2m” at 4045°C.

Acknowledgments The experiments in the hybrid magnets were carried out at the High Field Laboratory for Superconducting Materials, IMR Tohoku University. This work was partially supported by Grant-in-Aid for Scientific Research on Priority Area ‘‘Innovative utilization of strong magnetic fields” (Area 767) from MEXT of Japan.

284

References 1. Beaugnon E., Tournier, R., Levitation of organic materials. Nature, 349 (199 I), p. 470. 2. Weilert, M.A., et al., Magnetic levitation and noncoalescence of liquid helium. Phys. Rev. Lett.,77 (1996), pp. 4840-4843. 3. Valles, J.M., et al., Stable magnetic field gradient levitation of xenopus iaevis: Toward low-gravity simulation. Biophys. J., 73 (1997), pp. 11301133. 4. Berry, M.V., Geim, A.K., Of flying frogs and levitrons. Eur. J. Phys., 18 (1997), pp. 307-313. 5 . Motokawa, M., et al., Magnetic levitation experiments in Tohoku University. Physics B, 256-258 (1998), pp. 618-620. 6. Tagami, M., et al., Solidification of levitating water in a gradient strong magnetic fields. J. Crystal Growth, 203 (1999), pp. 594-598. 7. Hamai, M., et al., Crystal growth of ammonium chloride in magnetic levitation conditions. J. Crystal Growth, 209 (2000), pp. 1013-1017. 8. Kitamura, N., et al.,’ Containerless melting of glass by magnetic levitation method. Jpn. J. Appl. Phys., 39 (2000), pp. L324-L326. 9. Lynch, C.T., (Ed.) Handbook of Materials Science vol. I, (CRC PRESS, USA, 1974), p. 214. IO.Cini, R., Torrini, M., Temperature dependence of the magnetic susceptibility of water. J. Chem. Phys., 49 (1968), pp. 2826-2830. 11.Kimura, T., et al., Magnetic-field induced orientation of paraffin. Chem. Lett. (1999), pp. 1057-1058. 12.Takahashi, K., Orientation effect in melting of paraffin under the magnetic levitation condition, J. Appl. Mag. SOC.Jpn., 27 (2003) pp. 1125-1129. 13.Mogi, I., et al., Control of thermal convection in water by strong gradient magnetic fields, Jpn. J. Appl. Phys., 42 (2003) pp. L715-L717.

PROTEIN CRYSTAL GROWTH IN LOW GRAVITY PROVIDED BY A NEW TYPE OF SUPERCONDUCTINGMAGNET N.I. WAKAYAMA', D.C. YIN', Y. TANIMOTO', M. FUJIWARA*, K. HARATA3, H. WADA' 'National Institute for Materials Science, Tsukuba, Japan 'Institute for Molecular Science, Okazaki, Japan 3National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan As one of the best candidates for simulating the microgravity conditions in space, a low

gravity environment provided by applying an upward magnetic force has been considered to grow protein crystals. We grew protein crystals (orthorhombic lysozyme) for the first time at pseudo-microgravity.This study showed that pseudo-microgravityimproves the crystal quality effectively and reproducibly.

1. Introduction

Determining the structure of protein molecules is crucial. The most powerful technique for determining protein structure is X-ray crystallography. However, producing protein crystals of adequate size and quality is often the "bottleneck" for three-dimensional X-ray structure analysis of protein molecules. Recent crystallization experiments conducted in space have indicated that crystals grown in microgravity may be larger and yield diffraction data of significantly higher resolution than the best crystals grown on Earth [l]. An obvious difference between the space- and Earth-based experiments is the magnitude of gravity and buoyancy. One method to damp natural convection is the use of magnetic fields. The Lorenz force has been widely used to control natural convection in electrically high-conducting fluids such as melts of semiconductors and irons. However, this technique cannot be applied to electrically low- or non-conducting fluids. The electric conductivity of a typical protein aqueous solution is several ohm-'m-' while that of molten Si is about 4xlO6ohm-'m-'. There have been few methods to damp natural convection in electrically low-conducting fluids such as protein aqueous solutions on Earth. A new method to control effective gravity uses a vertical magnetic (Kelvin) force provided by a superconducting magnet, and is applicable to the control of natural convection in electrically low- and non-conducting fluids such as water, organic solvents, and typical protein aqueous solutions, etc. [2,3]. Low gravity environments obtained by the above method have been considered as one of the 285

286

best candidates for substituting or simulating the microgravity conditions in space [4-61 in order to grow protein crystals. A convenient type of superconducting magnet can now generate microgravity environments for more than several weeks [7,8]. For the first time, we grew protein crystals in pseudomicrogravity obtained by a superconducting magnet, and compared the crystal quality with those of crystals formed outside the magnet. The present study will suggest a new general means of growing high quality protein crystals.

2. Method to Control Effective Gravity Generally, a unit volume of substance in one-dimensional magnetic field gradient experiences the force:

F, = ~ O X(dH/dy) H = POp Xg H(dH/dy)

(1)

where y is a site coordinate, ,UO is the absolute magnetic permeability of vacuum, H is magnetic field strength, and x i s volume magnetic susceptibility, which is ). corresponds the product of density @) and mass magnetic susceptibility (xs to density in gravitational force.

x

Table 1. Volume magnetic susceptibility (3, mass magnetic susceptibility (xp ), and density @) of diamagnetic materials at room temperature.

Table 1 shows the list of x, xe,and p of some diamagnetic materials at room temperature. Most materials are diamagnetic, and they experience a weak repulsive force along the steepest gradient of the magnetic field strength since xK

Materials Processing In Magnetic Fields

Magnetic fields

Frontiers in Magnetic Materials

Magnetic Materials

High Magnetic Fields

Reconnection of magnetic fields

Thermoelectric Power in Nanostructured Materials: Strong Magnetic Fields (Springer Series in Materials Science)

Introduction to Magnetic Materials

Magnetic fields in the early universe

Introduction to magnetic materials

Magnetism and Magnetic Materials

Low Magnetic Fields in Anisotropic Superconductors

Magnetism, Nanosized Magnetic Materials

Handbook of Magnetic Materials

Galactic and intergalactic magnetic fields

Advance image processing in magnetic resonance imaging

Magnetic Materials: Fundamentals and Applications

Nanoscale Magnetic Materials and Applications

MICROWAVE PROCESSING OF MATERIALS

Solar Flare Magnetic Fields and Plasmas

Electric and Magnetic Fields 3rd edition (1956)

On the Rapid generalization of magnetic fields

High Magnetic Fields: Science And Technology

Laser materials processing

Materials Processing Defects (Studies in Applied Mechanics)

Advances in Laser Materials Processing Technology

Fluid Bed Technology in Materials Processing

Fluid Bed Technology in Materials Processing

Handbook of Magnetic Materials, Volume 14 (Handbook of Magnetic Materials Series)

Physics of Magnetism and Magnetic Materials

Materials Processing In Magnetic Fields

Magnetic fields

Frontiers in Magnetic Materials

Magnetic Materials

High Magnetic Fields

Reconnection of magnetic fields

Thermoelectric Power in Nanostructured Materials: Strong Magnetic Fields (Springer Series in Materials Science)

Introduction to Magnetic Materials

Magnetic fields in the early universe

Introduction to magnetic materials

Magnetism and Magnetic Materials

Low Magnetic Fields in Anisotropic Superconductors

Magnetism, Nanosized Magnetic Materials

Handbook of Magnetic Materials

Galactic and intergalactic magnetic fields

Advance image processing in magnetic resonance imaging

Magnetic Materials: Fundamentals and Applications

Nanoscale Magnetic Materials and Applications

MICROWAVE PROCESSING OF MATERIALS

Solar Flare Magnetic Fields and Plasmas

Electric and Magnetic Fields 3rd edition (1956)

On the Rapid generalization of magnetic fields

High Magnetic Fields: Science And Technology

Laser materials processing

Materials Processing Defects (Studies in Applied Mechanics)

Advances in Laser Materials Processing Technology

Fluid Bed Technology in Materials Processing

Fluid Bed Technology in Materials Processing

Handbook of Magnetic Materials, Volume 14 (Handbook of Magnetic Materials Series)

Physics of Magnetism and Magnetic Materials

Recommend Documents