Thermal Analysis. Fundamentals and Applications to Material Characterization. Proceedings of the International Seminar: Thermal Analysis and Rheology, Ferrol, 30 June-4 July 2003

Thermal Analysis. Fundamentals and Applications to Material Characterization Proceedings of the International Seminar: ...

Author: Artiaga Diaz

196 downloads 629 Views 7MB 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

Thermal Analysis. Fundamentals and Applications to Material Characterization Proceedings of the International Seminar: Thermal Analysis and Rheology. Ferrol, Spain, 30 Juny – 4 July 2003

Ramón Artiaga Díaz (ed.)

A Coruña 2005

Universidade da Coruña Servizo de Publicacións

Thermal Analysis. Fundamentals and Applications to Material Characterization. Edición a cargo de Ramón Artiaga Díaz. A Coruña. Universidade da Coruña, Servizo de Publicacións. 2005. xiv +288 páxinas. 17 x 24 cm. Cursos_Congresos_Simposios nº 80. Índice: pp. ix. Depósito legal: C-268-2006. ISBN: 84-9749100-9.

Edición Universidade da Coruña, Servizo de Publicacións http://www. udc.es/publicaciones Coa colaboración de TA Instrumentes © Universidade da Coruña

Distribución Galicia:

CONSORCIO EDITORIAL GALEGO. Estrada da Estación 70-A, 36818, A Portela. Redondela (Pontevedra). Tel. 986 405 051. Fax: 986 404 935. Correo electrónico: [email protected]

España:

BREOGÁN. C/ Lanuza, 11. 28022, Madrid. Tel. 91 725 90 72. Fax: 91 713 06 31. Correo electrónico: [email protected]. Web: http://www.breogan.org

Deseño de cuberta: Julia Núñez Calo Imprime:

NINO-Centro de Impresión Digital. C/ Rosalía de Castro, 58 15702 Santiago de Compostela

Reservados todos os dereitos. Nin a totalidade nin parte deste libro pode reproducirse ou transmitirse por ningún procedemento electrónico ou mecánico, incluindo fotocopia, gravación magnética ou calquera almacenamento de información e sistema de recuperación, sen o permiso previo e por escrito das persoas titulares do copyright.

In memoriam

Prof. Lisardo Nuñez Regueira passed away on September 1st 2005, when this book was in press. He was a close colleague and friend. He promoted the seminar leading to the edition of this book.

This page intentionally left blank

To María and Rocío


ix

Contents Foreword Acknowledgements

xi xiii

1. Fundamentals of TGA and SDT Weibing Xu, Sen Li, Nathan Whitely and Wei-Ping Pan

1

2. An Introduction to the Techniques of Differential Scanning Calorimetry (DSC) and Modulated DSC® Leonard C. Thomas

9

3. Thermal Analysis in Thermoset Characterization R. Bruce Prime

27

4. Characterization of Pharmaceutical Materials by Thermal Analysis Leonard C. Thomas

47

5. The Application of Thermal Analysis in the Study of Metallic Materials Angel Varela and Ana García

87

6. Thermal Analysis of Inorganic Materials José L. Mier

99

7. Characterization of Coal by Thermal Analysis Methods Sen Li, Nathan Whitely, Weibing Xu and Wei-Ping Pan

111

8. Characterisation of Polymer Materials Using FT-IR And DSC Techniques Pere Pagès

121

9. Characterization of Polymeric Materials by Thermal Analysis, Spectroscopy and Microscopic Techniques Nathan Whitely, Weibing Xu, Sen Li and Wei-Ping Pan

141

10. Energy Evaluation of Materials by Bomb Calorimetry José A. Rodríguez and Jorge Proupín

155

11. Introduction to the Viscoelastic Response in Polymers María L. Cerrada

167

12. Fundamentals of DMA Ramón Artiaga and Ana García

183

13. Dynamic Mechanical Analysis of Thermosetting Materials R. Bruce Prime

207

14. Fundamentals and Applications of DEA Lisardo Núñez, Carlos Gracia-Fernández, Silvia Gómez-Barreiro

225

15. Dielectric Analysis. Experimental Silvia Gómez-Barreiro, Carlos Gracia-Fernández, Lisardo Núñez Regueira

245

16. Statistical Applications to Thermal Analysis Ricardo Cao and Salvador Naya

265


xi

FOREWORD The idea for this book arose in the summer of 2003, during the seminar on Thermal Analysis and Rheology that took place in Ferrol. Some of the lecturers and attendees agreed that it would be helpful to have a book dealing with the techniques and applications of thermal analysis, but following a similar approach to one taken in the seminar. Such a text would be helpful for beginners and experienced practitioners who just wanted to get an accurate insight and put what they learned into practice. This book provides an overview of thermal analysis techniques. It focuses on the basic principles and looks at their application to polymers, pharmaceuticals, coals, metals and other inorganic materials. The text was conceived as a reference book and practical guide for material researchers, engineers and technologists who use thermal analysis. It also provides an academic approach for university students. The expertise of the contributors spans several fields, including industrial R&D on polymers, instrument development and research on materials characterization. A more academic approach is given by teaching staff from the Thermal Analysis research groups of the Universities of Santiago de Compostela and Coruña, who had been involved in organising the seminar mentioned earlier. The techniques covered in this book are DSC, M-DSC, TGA, Simultaneous DTA-TGA, Bomb Calorimetry, DEA and DMA. The contents were classified according to specific topics related to different areas of interest within thermal analysis. Therefore, apart from the chapters dedicated to the fundamentals of the different techniques, there were others devoted to specific applications: thermosets, pharmaceuticals, metals and inorganic materials, coal, evaluation of the power content of materials and viscoelastic behaviour of polymes. The chapter authored by P. Pages from the Universitat Politecnica de Catalunya exemplifies an application to material characterization where thermal analysis techniques, among others, play an important role. A final chapter was included, emphasizing how important it is to consider the mathematical treatment of thermal analysis data. This chapter introduces smoothing/fitting techniques and pattern recognition.


xiii ACKNOWLEDGEMENTS Many thanks are sent to TA Instruments, Universidade da Coruña, Xunta de Galicia and Aginsu S. L. They kindly supported the seminar whose proceedings were the starting point of this book. My special thanks go to Sergio Ruiz from TA Instruments for also supporting the Seminar and encouraging the production of this book. I am grateful to all the contributors, especially Bruce Prime, Wei-Ping Pan and Leonard Thomas, for making an additional effort of coming from the USA to participate in the Seminar. Finally, I wish to extend my thanks to Ana Demitroff for her revision of the English in some parts of the book.


Fundamentals of TGA and SDT Weibing Xu, Sen Li, Nathan Whitely and Wei-Ping Pan Thermal Analysis Laboratory, Materials Characterization Center, Western Kentucky University, Bowling Green, KY 42101 [email protected] Thermal analysis is one of the most useful methods of analysis in collecting both physical and chemical information. Probably the most used technique of thermal analysis is thermogravimetric analysis (TGA). TGA is used in all types of applications, providing information about the bonding of components within the sample. TGA can become an even stronger analytical technique when coupled with other thermal analysis techniques such as differential scanning calorimetry (DSC) and spectroscopic techniques such as Fourier transform infrared spectroscopy (FTIR) and mass spectrometry (MS). TGA measures the absolute amount and rate of change in weight of a sample as either functions of time or temperature in a controlled environment. TGA has a wide range of properties that can be measured such as thermal stability, oxidative stability, effects of different atmospheres, moisture and volatile content, and sometimes the composition of multi-component systems. TGA determines if and how different components within a material are bonded differently. When TGA is coupled with DSC or differential thermal analysis (DTA), the mode of analysis is called simultaneous DSC-TGA (or DTA) (SDT). SDT measures the amount and rate of change in weight, but also measures the heat flow of the sample as a conventional DSC does. SDT measures the same properties as TGA, but extends the list to include heats of reactions, melts, and boiling points. The three most important signals that TGA collects while it is analyzing the sample are weight, rate of weight change—differential thermogravimetry, and temperature. A differential thermogravimetry curve—DTG—is generated as the first derivative of the weight with respect to temperature or time. The DTG curve can be used to provide both qualitative and quantitative information about the sample. Qualitative modes of analysis include fingerprinting a material and distinguishing between two or more overlapping reactions. Quantitative modes include peak height and temperature at maximum weight loss measurements. The most important aspect of TGA operation is the validity of measurements made. Confidence in data collection can be achieved through regular calibration. For TGA, both mass and temperature calibrations must be performed. Most instrument and software packages possess a relatively automated mass calibration procedure in which the user places certified calibration weight onto the instruments sample platform. Temperature calibrations are performed by the calculation of the Curie point of standard metals. The Curie point of a material is the temperature at which the material loses its magnetic susceptibilities. To perform the Curie point temperature calibration, a strong magnet must be place below or on top of the furnace the cause an initial weight gain or loss at room temperature. Figures 1and 2 show the experimental apparatus for both vertical and horizontal instrumental configurations.

2

WEIBING XU, SEN LI, NATHAN WHITELY AND WEI-PING PAN

Figure 1. Horizontal temperature calibration configuration

Figure 2. Vertical temperature calibration configuration A small lab jack can be used to adjust the magnet’s distance from the sample such that a 2-3% weight gain or weight loss occurs once the magnet is positions above or below the sample. Figure 3 shows the Curie point determination for nickel and alumel. Note that the Curie point is denoted as the offset.

Figure 3. Curie point determination for vertical TGA SDT also has an alternate method for temperature calibration. The melting points of standard materials can be determined by the onset of the endotherms and compared to the theoretical melt temperature. A good exercise for both TGA and SDT is to perform multiple analyses of calcium oxalate monohydrate. By performing such

FUNDAMENTALS OF TGA AND SDT

3

an analysis the performance and precision of both you and the instrument can be measured. An overlay of five calcium oxalate experiments is shown in Figure 4.

Figure 4. Performance testing using calcium oxalate monohydrate Although calcium oxalate monohydrate is not typically a standard material, it does hold good utility in intra-laboratory analysis. The weight change and peak temperature can be inputted into a spreadsheet program to check your instrument and operators performance. The accuracy of the instrument can be used to assess your instruments long-time performance, and help single out a damaged component of the instrument. The baseline can also be quite usual in quantifying your instrument’s performance and sensitivity. Small weight losses become increasingly difficult to measure if the instrument’s baseline is large compared to that of the instrument. TGA is the foremost analysis technique in determining quantitative properties of the original sample. A polyethylene (PE) sample filled with CaCO3 was analyzed as shown in Figure 5.

Figure 5. TGA curve of polyethylene sample filled with calcium carbonate By knowing the degradation reaction of CaCO3, the initial percentage of CaCO3 in the PE can be calculated. At approximately 550ºC the PE is completely decomposed; thus, the weight loss occurring at approximately 650ºC is due to the decomposition of CaCO3. The weight loss is a direct result of the evolution of CO2 gas. The residue is the remaining CaO that fails to decompose. From the weight change and the residue,

4


the stoichiometric relationships can be used to determine a percentage of CaCO3 that exists in the original PE sample. Calculating the initial percentage of CaCO3 from the weight change is more accurate than calculating it from the residue. Most polymers contain fillers; hence, the residue is a combination of CaO and these fillers making this calculation less accurate. TGA and SDT can also be used to demonstrate the important of reaction atmosphere. Calcium oxalate monohydrate was analyzed under the same experimental conditions except the purge gas. The sample was analyzed in air, CO2, and nitrogen of equal flow rates. Figure 6 illustrates Le Chatilier’s principle.

Figure 6. Le Chatilier’s principle shown using TGA Because the degradation of CaC2O4 produces CO2, the reaction is inhibited when it occurs in a CO2-saturated atmosphere. Figure 7 shows the heatflow data collected with the SDT.

Figure 7. DSC Curve of calcium oxalate monohydrate in multiple atmospheres The CaC2O4 oxidizes in air as shown by the endotherm at approximately 500ºC while in nitrogen and CO2 oxidation does not occur but rather pyrolysis. Hi-Resolution TGA is useful to separate overlapping weight losses. HiResolution TGA exposes the sample to an isotherm once a weight loss is detected. The isotherm allows the weight loss occurring at the lower temperature to complete before the second weight loss begins. Figure 8 shows that as the resolution increases, the two weight losses are more separate and defined.


5

Figure 8. TGA curves at multiple hi-resolution settings Quality control testing often exposes a product to a particular atmosphere for very extended periods of time which can be costly and time consuming. TGA in conjunction with kinetics software can be used to decrease the time and money spent on tedious lifetime testing procedures. A sample is analyzed over the same temperature range using at least four different heating rates. Software is then used to generate numerous plots that can predict the product’s performance over time. The activation energy, rate constant, and other kinetics related information can be provided as seen in Figure 9.

Figure 9. Log[heating rate] curve at multiple conversions Figure 10 shows the lifetime of the sample at varying isotherms.

6


Figure 10. Lifetime plot for polymer sample Although this lifetime plot may not eliminate the need for lengthy quality control testing, it may help predict poorly performing products at an earlier stage in the production process. Two automotive belts composed of alkylated chlorosulfonated polyethylene (ACSM) were tested using TGA to identify the reason why one belt performed at 10% of a normal functioning belt. The belts were analyzed under the exact heating rates under an air atmosphere. The belts each showed the typical degradations profile of a rubber sample as seen in Figure 11.

Figure 11. TG and DTG Curve of Passed and Failed Belt Sample in Air Oil was decomposed first followed by the decomposition of the polymer, and finally the carbon black combusted with the oxygen in the air. This analysis showed that both the oil and polymer portions of the rubber were not the cause of the bad belt’s failure. The decomposition of the bad belt was approximately 20ºC lower than the good belt. Figure 12 shows that the bad belt was composed of carbon black 1, which has the lower decomposition temperature.


7

Figure 12. TG and DTG Curves of Carbon Black Components of Failed Belt Sample TGA and SDT can used be in nearly any application to gather information. TGA and SDT provide a method of analysis that is fast and easy to operate, but provides precise and accurate results. In situations where TGA and SDT cannot be used to study a system directly, TGA and SDT can provide estimations that help alleviate some of the difficulty in using more complicated analysis methods.

Acknowledgements Many thanks to Len Thomas of TA Instruments who allowed use of his short course presentation given at Western Kentucky University.


An Introduction to the Techniques of Differential Scanning Calorimetry (DSC) and Modulated DSC® Leonard C. Thomas TA Instruments [email protected]

1. Introduction Differential Scanning Calorimetry or DSC is one of a series of analytical techniques called thermal analysis. These techniques can be used to characterize the physical properties of a wide variety of materials and how they change with temperature. The most frequently used techniques and the properties measured include:

Differential Scanning Calorimetry (DSC) – Heat Flow Thermogravimetric Analysis (TGA) – Weight Change Thermomechanical Analysis (TMA) – Dimensional Change Dynamic Mechanical Analysis (DMA) – Modulus (Stiffness) Rheology – Viscosity (Flow)

DSC is the most important analytical tool because all transitions (change in physical properties) involve heat flow. Endothermic transitions such as melting absorb energy (J/g = Joules/gram) while exothermic transitions such as crystallization release energy. Other advantages of DSC include the ability to use small samples (1-10 mg), analyze solids and liquids, and use short test times (10-30 minutes). Sample preparation is easy and most commercial DSCs offer auto samplers and auto analysis. The utility of DSC for characterizing a wide range of materials can be seen in Table 1 which lists the types of most frequently made measurements (properties) on those materials. Table 1. Frequent DSC measurements TP

TS

EL

CH

PE GL ME

Bio

Glass Transition Temperature (Tg)

Glass Transition Size (ΔCp)

Melting Temperature ™

Crystallization Temperature (Tc)

Crystallinity (J/g not %)

Heat Capacity (J/g°C)

Oxidative Stability (Temp or Time)

Texturing (process) Temperature (°C)

Curing/Degree of Cure (%)

Polymorphic Transitions

Denaturation/Gelatinization TP = Thermoplastics TS = Thermosets EL = Elastomers

CH = Chemical/Drugs PE = Petroleum GL = Glasses

ME = Metals Bio = Proteins/Starches

10

LEONARD C. THOMAS

Heat flow is always the result of a temperature difference between two objects or two points in a single body. With DSC, the difference in heat flow rate between a sample and an inert reference is measured as both are heated or cooled (scanning) in a controlled temperature environment. The temperature difference (ΔT) between the sample and reference changes each time the sample goes through an exothermic or endothermic transition. This ΔT causes a proportional difference in heat flow rate (ΔQ). The magnitude of the heat flow rate is also a function of the thermal resistance (R) and is expressed by:

ΔQ =

ΔT R

In order to obtain the highest level of performance (sensitivity, resolution, single-run measurement of heat capacity, straight baseline, etc.), modern DSC instruments such as the TA Instruments Tzero™ technology also account for heat flow within the components of the DSC. These components have the ability to store heat (thermal capacitance) and transfer heat (thermal resistance). Instead of the simplified equation above, Tzero™ technology uses the following equation in order to more accurately measure the sample heat flow by separating it from heat flow within components of the DSC.

ΔQ =

dTs dΔT ΔT §1 1· - Cs + ΔTo ¨ - ¸ + (Cr - Cs) Rr dt dt © Rs Rr ¹

Where:

ΔQ

= QSample – QReference

ΔT Rr

= Principle Heat Flow Term

§1 1· ΔTo ¨ - ¸ = Imbalance in Sensor Resistance Term © R s Rr ¹

(Cr - Cs)

Cs

dΔT dt

dTs dt

= Imbalance in Sensor Capacitance Term

= Imbalance in Heating Rate During Transition Term

AN INTRODUCTION TO THE TECHNIQUES OF DIFFERENTIAL SCANNING CALORIMETRY (DSC) AND MODULATED DSC® 11

Figure 1 shows a cross-section of a Tzero™ DSC cell. The cell is the actual measuring chamber which would be located within an environmentally controlled cabinet complete with electronics and a cooling system. The cabinet is interfaced to a computer controller which uses software to perform experiments and analyze the resulting data. Silver Base for Cell Lid #2

Silver Base for Cell Lid #1

Chambers for Temperature Conditioning of Purge Gas Measuring Chamber

Tzero™ Sensor

Furnace

54 Nickel Cooling Rods

Cooling Flange

Figure 1. TzeroTM DSC cell schematic

Figure 2 shows the major components of a modern DSC system.

Figure 2. Modern DSC system

A DSC system usually contains a DSC module, which can have numerous options such as Autosampler, Autolid, and MDSC®, a Refrigerated or LN2 Cooling System and a computer-based controller

12

LEONARD C. THOMAS

2. Typical DSC Measurements

exo

Figure 3 illustrates the most common type of report obtained from a DSC experiment, a plot of endothermic (heat absorbed) and exothermic (heat released) heat flows as a function of the sample's temperature. No one sample would contain all of the transitions shown in Figure 3. It simply is an illustration of the most common types of transitions.

Glass Transition

Crystallization

Melting

Oxidation or Decomposition

endo

HEAT FLOW

Crosslinking (Cure)

TEMPERATURE

Figure 3. Transitions in a DSC curve

A brief definition/description of commonly measured transitions is provided below: Glass Transition

A change in the physical properties of an amorphous material. One of the properties to change is heat capacity, which can be measured by DSC as an endothermic shift in the heat flow baseline as sample temperature is increased. Glass Transition Temperature (Tg)

The temperature, usually a range, over which the properties of an amorphous material change. Below Tg, materials exhibit a glassy, rigid structure; while above Tg, they are rubbery and flexible.


Figure 4. Glass transition analysis

Crystallization

The conversion of amorphous structure into crystalline structure. Most crystalline polymers have both types of structure and are usually referred to as semicrystalline. Crystallization is normally seen during cooling from a temperature above the melting point, but it can also occur during heating. In this case, it is often called “Cold Crystallization.” Melting

The conversion of crystalline structure to a viscous amorphous structure. Melting occurs over a temperature range in polymers due to the molecular weight distribution and range in crystal sizes and defects within the crystal. With chemicals, the melting range increases and moves to lower temperatures as the impurity level in the sample increases. Although amorphous materials can flow at higher temperatures, due to decreasing viscosity, they do not melt.

14

LEONARD C. THOMAS

Fiure 5. Effect of heating rate on crystallization and melting of quench-cooled PET Cure

A chemical reaction within a material that increases the crosslink density of the material and reduces molecular mobility. The term is usually used to define the reaction that takes place in thermosetting polymers (epoxies, phenolics, etc.). The glass transition temperature of the sample increases as the degree of cure increases. -0.04 -0.08

155.93°C

First Tg

Residual Cure

-0.12 -0.16

Second

Tg

102.64°C 20.38J/g

H

-0.20

t Fl -0.24 (W/ )

0

50

100 150 200 Temperature (°C)

250

300

Figure 6. Thermosets.. Comparision of first and second heat


Decomposition

The breaking of chemical bonds due to heat or a chemical reaction such as oxidation. Partial decomposition usually results in a decrease in the glass transition and melting temperatures. Other, but less common, measurements are as follows: Oxidative Stability

A measure of the ability of a material to resist a chemical reaction between components of the material and oxygen. Tests are normally performed at elevated temperatures to reduce test time to less than one hour. Pressure DSC is the preferred technique for characterizing samples with volatile components. Reaction Kinetics

Uses mathematical models to analyze the shape and temperature of reaction exotherms to determine kinetic parameters such as activation energy. Purity (Figure 7)

Determines the purity of high purity (> 98%) metals and chemicals. The DSC method of determining purity is based solely on the shape of the melting curve as compared to the shape of the curve for a pure melting material. The higher the impurity level, the lower the melting temperature and the broader the melting range.

Figure 7. Melting temperature decreases and melting range increases as the level of impurity increases

16

LEONARD C. THOMAS

Specific Heat Capacity (Figure 8)

The quantity of heat required to raise the temperature of one gram of a material by 1°K. It is measured in DSC by comparing the endothermic heat flow of an unknown with the endothermic heat flow of a standard material at a specific heating rate. A baseline scan is typically done first, which means that three separate experiments need to be performed. Modern instruments such as Tzero™ DSC and advanced techniques such as Modulated DSC® can measure specific heat capacity in a single run. S am ple: PE T; Q uenched S ize: 16.0000 m g Method: Heat@ 20 C om m ent: Heat@ 20

DSC

File: C:...\Crystallinity\RIqP ETcycle20.001

6

600

275.00°C 530.8J/g

400 Heat Capacity (Single Run) 135.54°C 0.7311J/g

2

200

[ ––––– · ] Integral (J/g)

H eat C apacity (J/g/°C )

4

Running Integral

0

-2 0

50

100

150

200

Temperature (°C)

0 300

250

Universal V3.8A TA Instrum ents

Figure 8. Engineers often need heat capacity information 3. Use of DSC in R&D, Analytical and QC/QA Laboratories DSC in the R&D Laboratory

The goal of R&D is to create new or improved products that provide financial benefit to the corporation. Although we often think that new products improve profitability through increased market share and sales, an equally good way is to lower the cost of manufacturing a product without adversely affecting its end-use performance. The most successful new products are ones that provide the consumer with a better performing product that also has a lower manufacturing cost. Regardless of which approach you are taking, DSC is a valuable tool that will help you achieve program goals. In developing a new product, there are at least three major elements in the development process that need to be considered. 1. Product Formulation and Costs In selecting materials to be used in a product, it is necessary to consider their: a. Cost per unit of product. b. Ability to meet mechanical and any chemical resistance requirements.


c. d. e.

Aging characteristics: do properties change with time? Manufacturing costs and sensitivity to normal process variations. Environmental impact: are any of the components volatile, and how do you dispose of the product when it is no longer needed?

DSC can measure changes in structure that result from formulation, processing or aging. Before subjecting new materials or formulations to more extensive testing, which often requires more elaborate sample preparation or long-term oven testing, run a quick check with the DSC to see if the formulation is even a candidate for a product. It is sometimes possible to use cheaper materials in a product through the use of additives such as fillers, plasticizers and antioxidants. Although DSC could not measure a change in color, it can determine the relative effectiveness of antioxidant concentrations and measurer the effect of plasticizer or filler level on transition temperatures. 2. End-Use Performance Obviously, the product needs to perform the task that the consumer wishes to satisfy. This means that the product must have the physical and/or chemical properties required for its specific end-use. Those properties, and how they are measured or predicted from DSC data are as follows: •

Mechanical Strength In order to determine if a new or modified material has sufficient mechanical stability, it is important to define the temperature range over which the product could be used. Once that is known, DSC can identify the temperatures at which transitions or phase changes occur, which is also the temperature where physical properties can change by orders of magnitude. For example, if a rubber O-ring is to be used as a gasket between two rigid parts, it needs to remain flexible during end use. If it is rigid, it cannot fill the space between the two rigid parts as they move. By measuring the Glass Transition temperature of the material, it is possible to determine the temperature at which the properties change from rigid to flexible.

•

Mechanical Stability Materials can be amorphous (non-crystalline), crystalline or a mixture of both. Amorphous materials tend to creep or flow over time as stress is applied to them. However, they tend to have better impact properties than crystalline materials. Although crystalline materials are usually more brittle, they have higher modulus or mechanical strength. DSC can measure the relative crystallinity of materials. In order to optimize both mechanical strength and stability, it is often necessary to use blends of two or more polymers or to precisely control processing to achieve an optimum level of amorphous or crystalline structure.

18

LEONARD C. THOMAS

•

Visual Characteristics Color, color uniformity, surface gloss, etc., are affected by the components and the distribution of the components used to make the product. But, they are also affected by thermal or oxidative degradation as well as relative crystallinity. Since sample sizes can be as small as a milligram or as large as tens of milligrams, it is often possible to sample the product to determine if a particular material is susceptible to thermal or oxidative degradation, or tends to undergo phase separation or has surface properties that differ from bulk properties.

3. Processability and Cost Although some materials have high strength as well as high thermal and oxidative stability, it may be necessary to heat them to higher temperatures in order to process them and this costs money. DSC is very useful in helping to determine the suitability of a particular formulation based on the following: a. b. c. d. e.

Maximum processing temperature required. Process time, temperature cooling rate, and total energy requirements. Potential for thermal or oxidative degradation during processing. Environmental issues due to volatilization of components. Product variability due to normal processing variability.

The goal is to produce the best product in the shortest amount of time and at the lowest cost. DSC provides many of the answers required to meet that challenge. Using DSC to determine the effect of a nucleating agent on crystallization time and temperature would be one such example of a possible cost reduction. DSC in the Analytical Laboratory

Many companies use a central laboratory to meet the analytical needs of the company. Although there are pros and cons of having a single, central laboratory, it usually results in a better-trained and staffed laboratory that is a useful resource to materials scientists throughout the company. Support can be provided to R&D, Process Control, Quality Control/Assurance and even Marketing. A sales force loves to have proof statements about why their product is better, and the Analytical Lab can provide them through competitive product analyses. DSC in the Quality Control/Assurance Laboratory

In order to provide consistent product at the lowest possible cost, it is necessary to monitor physical properties of both incoming raw materials as well as outgoing finished product. This way, if a problem occurs, it can be traced back to the material supplier or to the manufacturing process. The procedures used in the R&D laboratory to develop the optimum product can just as easily be used in the QC/QA laboratory. For high value-in-use products, vendor certification through DSC analyses provides a fast and reliable way to help insure product quality.


4. Introduction to Modulated DSC® (MDSC®)

Although DSC has been an extremely useful analytical technique for over forty (40) years, it has natural limitations with which most thermal analysts are somewhat familiar. These include: •

Baseline straightness limits sensitivity to detect weak transitions.

•

Sensitivity increases with higher heating rates but resolution decreases as heating rate increases.

•

Most modern materials are mixtures of plastics, fillers and additives which have overlapping transitions that are difficult to interpret.

•

Most engineering plastics are semi-crystalline which increase in crystallinity while being heated in the DSC. Because DSC data usually makes the detection of the changing crystallinity difficult to detect, DSC crystallinity values can often be wrong by 50% or more.

•

The measurement of specific heat capacity by DSC is often slow and laborious.

MDSC® overcomes these natural limitations of DSC as will be illustrated. However, MDSC® also has a limitation which results in it being a much slower technique (5-10 times). Therefore, the best approach for characterizing new materials is to always start with DSC and then switch to the MDSC® mode if one or more of its advantages are needed. 5. Operating Principle of MDSC®

With traditional DSC, a linear temperature ramp (heat/cool) is applied as a function of time. The resulting heat flow is a function of the rate of temperature change, absolute temperature, sample mass and the specific heat of the sample.

dH dT = Cp + f (T, t) dt dt Where: dH dt

= Heat Flow Rate (mW or W/g)

Cp

= Sample Specific Heat (J/g°C) x Sample Mass (g)

dT dt f (T, t)

= Heating Rate = Heat Flow Rate due to Kinetic Processes (mW or W/g)

MDSC® uses two simultaneous heating rates, a linear ramp that provides the same information as traditional DSC plus a sinusoidal ramp superimposed on the linear

20

LEONARD C. THOMAS

ramp that provides information about the sample's heat capacity. Figure 9 shows how temperature changes as a function of time in an MDSC® experiment, and Figure 10 provides the time-based derivatives (°C/min) which are the applied heating rates. Although it is beyond the scope of this paper, the applied rates can be selected to provide cooling during the modulation or have heat-only conditions. MDSC® does not require cooling during modulation but does use the change in heating rate to calculate the sample's heat capacity. 62

62

(Heat-Iso) M odulate +/- 0.42 °C every 40 seconds R am p 4.00 °C/m in to 290.00 °C Modulated Temperature

57.0

Amplitude

60

56.5

56.0

56.0

55.5

55.5

55.0

55.0

Modulated Temperature (°C)

Tem perature (°C )

Temperature (°C)

56.5

58

58

54.5 13.70

13.75

13.80

13.85

13.90

13.95

14.00

54.5 14.05

Time (min)

56

56

Average Temperature 54

54

N ote that tem perature is not decreasing during M odulation i.e. no cooling 52 13.0

13.5

14.0

52 15.0

14.5

Time (min)

Figure 9. MDSC average & modulated temperature

10

10

Period

Note That Heating Rate is Never Negative (no cooling)

8

6

6

Average Heating Rate

4

4

2

2

Modulated Heating Rate

0 13.0

13.5

14.0

14.5

0 15.0

Time (min)

Figure 10. MDSC average & modulated heating rate

Deriv. Modulated Temperature (°C/min)

Deriv. Temperature (°C/min)

8

M odulated Tem perature (°C )

57.0

60


The result of the sinusoidal heating rate is a sinusoidal heat flow as shown in Figure 11. The modulated heat flow signal (MHF) is measured during the experiment and used to calculate the signals used by MDSC® for analysis of material properties. With traditional DSC, there is only one heat flow signal (Total) which is the sum of all heat flows. With MDSC®, there are three primary signals: the Total, Reversing and Nonreversing.

dH dT = Cp + f (T, t) dt dt Total = Reversing + Nonreversing

Figure 11. MDSC raw data signals These three signals are shown in Figure 12 which is a quench-cooled sample of PET. The Total signal is calculated from the average value of the MHF signal while the Reversing signal is calculated from the ratio of the amplitudes of the MHF and modulated heating rate (MHR). The Nonreversing signal is simply the Total minus the Reversing heat flow. All averages and amplitudes are calculated using Fournier transform analysis.

Total = Avg. MHF Reversing =

Amp MHF x Avg. Heat Rate Amp MHR

Nonreversing = Total – Reversing

22

LEONARD C. THOMAS

Figure 12. Calculated MDSC heat flow signals

6. Applications Advantages of MDSC®

As previously stated, MDSC® overcomes the many natural limitations of standard DSC to provide superior sensitivity, resolution and separation of overlapping transitions. These benefits have been well documented in hundreds of papers since the commercialization of MDSC® in 1992. Therefore, only a few of the newer applications will be illustrated here.

Complex Polymer Blends

There is little doubt that blends of semi-crystalline and amorphous polymers are very difficult to characterize by traditional DSC. The reason is due to multiple glass transitions and often several crystallization peaks that occur while heating the sample in the DSC. Figure 13 shows DSC data on a common engineering plastic, Xenoy®, a product of the General Electric company. Xenoy® is a blend of Polybutylene Terephthalate (PBT) and Polycarbonate (PC). Except for the melting peak near 225°C, the results are very difficult to interpret.


Figure 13. DSC of complex polymer blend Figure 14 shows the same material run with MDSC®. Now it is relatively easy to measure the glass transitions of the PBT and PC and interpret the exothermic peaks near 60 and 150°C in the Nonreversing signal. Once the sample is heated above each of the glass transition temperatures, there is a step-increase in molecular mobility. This increase in mobility allows more of the amorphous PBT to crystallize.

Figure 14. MDSC of complex polymer blend

24

LEONARD C. THOMAS

Analysis of Polymer Crystallinity

Although DSC has been used for more than forty years to measure polymer crystallinity, results are often wrong by 50% or more. The reason is due to the sample's increasing crystallinity as it is being heated in the DSC and the difficulty in identifying the true heat capacity baseline in the data. Figure 15 shows DSC data on a sample of Nylon 6/6. Most DSC users would assume that baseline is best selected as shown in the blue curve. This yields a crystallinity value of approximately 50 J/g as compared to 29 J/g in the same data with the green curve. With standard DSC, it is very difficult to judge which is correct.

Figure 15. DSC @ 10 ºC/min on Nylon 6/6; Where is the baseline? Figure 16 shows MDSC® results on the same Nylon 6/6 and shows the actual crystallinity to be only about 24 J/g.

Figure 16. MDSC of nylon 6/6


Figure 17 shows MDSC® results on a sample which is a mixture of quenchcooled PET and PC. Since it was quench-cooled in liquid nitrogen, the crystallinity of the sample is known to be approximately 0 J/g. The Total curve, which is typical of the data from traditional DSC, is impossible to analyze correctly because the glass transition of the PC is not even visible. The sum of the melting and crystallization processes seen in the MDSC® Reversing and Nonreversing signals provides the expected crystallinity, and the PC glass transition is easily seen in the Reversing signal. Sam ple: Q uenched PET and P C Size: 13.6000 m g DSC Method: MD SC .318/40@ 3 Com m ent: MDS C 0.318/40@ 3; P E T13.60/P C 10.40/A l film 0.96m g

File: C:\TA\D ata\Len\C rystallinity\qPET-PC .002

3 57% PET ; 43% PC MD SC Signals

H eat Flow (m W )

95.13J/g 110.00°C

270.00°C

-1

-3 155.00°C 270.00°C

1

0

-1

-2

[ ––––– · ] R ev H eat F low (m W )

2

[ –– –– – ] N onrev H eat F low (m W )

1

-3

93.60J/g

-5

-4

Initial C rystallinity = 93.6 + (-95.1) = -1.5 J/g

-5 50 Exo Up

100

150

200

250

Temperature (°C)

300 Universal V3.8A TA Instrum ents

Figure 17. MSDC of 57/43 % PET/PC mixture

7. Summary

The combination of DSC and MDSC® provides an extremely useful analytical tool for the characterization of polymer structure and detection of changes (transitions) in that structure. Whereas DSC is a faster, easier to use technique, MDSC® offers advantages in sensitivity, resolution and separation of overlapping transitions.


Thermal Analysis in Thermoset Characterization R. Bruce Prime IBM (Retired) / Consultant [email protected] Thermosetting polymers are unique. Unlike thermoplastic polymers, chemical reactions are involved in their use. As a result of these reactions the materials cross-link and become “set”, i.e. they can no longer flow or dissolve. Cure most often is thermally activated, hence the term “thermoset”, but cross-linking materials whose cure is light activated are also considered to be thermosets. Some thermosetting adhesives cross-link by a dual cure mechanism, that is by either heat or light activation. Prime [1] is a general reference for this article. In this paper the distinguishing characteristics of thermosetting materials will be described, followed by a detailed discription of thermal analysis of the cure process, some brief comments on properties of cured thermosets and concluding with a discussion of kinetics including a recent case study. Note that Dynamic Mechanical Analysis of Thermosets is treated in a separate paper. Uncured thermosets are mixtures of small reactive molecules, often monomers. They may contain additives such as particles or fibers to enhance physical properties or reduce cost. Adhesives are probably the most common application of thermosets but they are also found in aerospace, electronic, medical and dental, and recreational materials. The most common thermoset is epoxy, and the most common epoxy resin is the diglycidyl ether of bisphenol-A:

Epoxide group or oxirane ring

The number of repeat units n is usually 0-6, 0 being the monomer. Epoxy resins can cross-link with themselves, referred to as homopolymerization, an example of which is anionic polymerization promoted by imidizoles. But it is more common to cross-link epoxies with a co-reactant such as a diamine. For cross-linking to occur at least one of the reactants must be trifunctional or higher. Epoxy resins are typically difunctional, reacting through the oxirane group, although in some cases reaction can occur through the -OH group. Diamines are four-functional where each amine hydrogen can react. The heat of reaction ΔHrxn for epoxy-amine is ~25½ kcal/mole (~106 kJ/mol) and the activation energy for cure can vary from 10 to 25 kcal/mole (40-100 kJ/mol) depending on the particular epoxy and curing agent. Two distinct phenomena are characteristic of thermoset curing, gelation and vitrification. Gelation is the transformation from a liquid to an elastic gel or rubber and it will always occur in a thermoset. Gelation is abrupt and irreversible and the gel point can be defined as the instant at which the molecular weight becomes infinite [2]. A thermoset is no longer processable above the gel point and therefore gelation defines the

28

R. BRUCE PRIME

upper limit of the work life. For a “Five Minute Epoxy” the five minutes refers to the gel point at room temperature (RT). For example, after the two parts are mixed the user must form an adhesive joint within five minutes before the material becomes rubbery. However, completion of cure requires a much longer time at RT but may be shortened by increasing the temperature. The degree of conversion at the gel point αgel is constant for a given thermoset, independent of cure temperature, i.e. gelation is iso-conversional. Therefore the time to gel versus temperature can be used to measure the activation energy for cure. Gelation does not affect the rate of cure and therfore is not detected directly by DSC but only indirectly if αgel is known. Gelation is detected directly by rheolgy and DMA and because it is a specific point along the reaction path it is determined by the chemical reaction and therefore independent of frequency. Vitrification is a completely distinct phenomenon that may or may not occur during cure depending on the cure temperature relative to the Tg for full cure. Vitrification is the glass transition due to reaction and occurs when the increasing Tg becomes equal to the cure temperature, i.e. when Tg = Tcure. Vitrification can occur anywhere during the reaction to form either an ungelled glass or a gelled glass. It can be avoided by curing at or above Tg∞, the glass transition temperature for the fully cured network. Unlike gelation, vitrification is reversible by heating. Also unlike gelation it causes a shift from chemical control to diffusion control and a dramatic slowing of the reaction. Vitrification is detected by TMDSC and DMA as a frequency-dependent transition. Thermoanalytical techniques include differential scanning calorimetry (DSC), rheology, dynamic mechanical analysis (DMA), thermal mechanical analysis (TMA) and thermogravimetric analysis (TGA). DSC measures heat flow into a material (endothermic) or out of a material (exothermic). Thermoset cure is exothermic. DSC applications include measurement of Tg, conversion α from the area under the exotherm, the reaction rate dα/dt and the heat capacity Cp. Gelation cannot be detected by DSC but vitrification can be measured by modulated-temperature DSC (MTDSC). Rheology measures the complex viscosity in steady or oscillatory shear. In oscillatory shear the advance of cure can be monitored through the gel point and both gelation and the onset of vitrification can be detected. DMA measures the complex modulus and compliance in several oscillatory modes. Gelation and vitrification can be detected, and the cure reaction can be monitored beyond the gel point in the absence of vitrification. Tg, secondary transitions below Tg, creep and stress relaxation can also be measured. TMA measures linear dimensional changes with time or temperature, sometimes under high loading. Measurements include linear coefficient of thermal expansion (CTE), Tg, creep and relaxation of stresses. TGA measures mass flow, primarily in terms of weight loss. Measurements include filler content for inert fillers; weight loss due to cure, e.g. loss of water for condensation reactions; outgassing; moisture sorption and desorption; and thermal and thermo-oxidative stability.

THERMAL ANALYSIS IN THERMOSET CHARACTERIZATION

29

Figure 1. Schematic, two-dimensional representation of thermoset cure. For simplicity difunctional and trifunctional co-reactants are considered. Cure starts with A-stage monomers (a); proceeds via simultaneous linear growth and branching to a B-stage material below the gel point (b); continues with formation of a gelled but incompletely cross-linked network (c); and ends with the fully cured, C-stage thermoset (d). From Ref. 1. Cure is illustrated schematically in Fig. 1 for a material with co-reactive monomers such as an epoxy-diamine system. For simplicity the reaction of a difunctional monomer with a trifunctional monomer is considered. Reaction in the early stages of cure {(a) to (b) in Fig. 1} produces larger and branched molecules and reduces the total number of molecules. Macroscopically the thermoset can be characterized by an increase in its viscosity η (see Fig. 2 below). As the reaction proceeds {(b) to (c) in Fig. 1}, the increase in molecular weight accelerates and all the chains become linked together at the gel point into a network of infinite molecular weight. The gel point coincides with the first appearance of an equilibrium (or time-independent) modulus as shown in Fig. 2. Reaction continues beyond the gel point {(c) to (d) in Fig. 1} to complete the network formation. Macroscopically physical properties such as modulus build to levels characteristic of a fully developed network. Fig. 3 shows DSC after isothermal cure at 160°C at various stages of cure for a typical epoxy-amine from uncured to fully cured. Note the residual exotherm decreasing and the Tg increasing in step with cure time. DSC at 10°C/min. From Ref. 4.

30

R. BRUCE PRIME

0

Steady State Properties

Conversion

η0

100%

Ge

,

Newtonian liquid

Network at the gel point

Hookean solid

Figure 2. Macroscopic development of rheological and mechanical properties during network formation, illustrating the approach to infinite viscosity and the first appearance of an equilibrium modulus at the gel point. From Ref. 3.

Figure 3.


31

Fig. 4 shows conversion-time curves for the same epoxy for cure temperatures from 100° to 180°C [4]. Note that the curves are parallel during the first part of cure. Epoxy-Amine Cure DGEBA-PACM-20 (1:1)

Conversion (α)

Wisanrakkit and Gillham, J.Appl.Poly.Sci. 41, 2885 (1990)

Time (minutes)

Figure 4.

Epoxy Cure

Tg (°C)

5.4°C/% Conversion 90 - 100%


DSC Fractional Conversion

(1 - α) In(Tg0) + In(Tg) = (1 - α) +

ΔCp∞ α In(Tg∞) ΔCp0 ΔCp∞ Venditti and Gillham, α J.Appl.Poyl.Sci. 64, 3, (1997) ΔCp0

Figure 5. Fig. 5 shows the Tg - conversion relationship for the same epoxy fitted to the DiBenedetto equation [5]. Note that as cure progresses Tg becomes an increasingly more sensitive measure of cure relative to the residual exotherm. From Ref. 4. Also shown is the Venditti-Gillham equation [6] relating Tg and conversion.

32

R. BRUCE PRIME

Epoxy-Amine Cure DGEBA-PACM-20 (1:1) Tg∞ = 178°C


Figure 6. Fig. 6 shows the Tg - time curves for the same epoxy system [4]. Note the similarity to the conversion – time curves. The arrows indicate vitrification. Dynamic mechanical analysis (DMA) measures the complex modulus and compliance as a function of temperature, time and frequency where, for example, •

storage modulus (E', G') which isҏ aҏ measure of stress stored in the sample as mechanical energy • loss modulus (E", G") which isҏ a measure of the stress dissipated as heat • tan į (E"/E' = G"/G') which isҏ the phase lag between stress and strain Properties measured include storage and loss modulus, storage and loss compliance, tan δ, Tg, secondary transitions below Tg, gelation and vitrification and reaction beyond the gel point. DMA of thermosets is covered in a subsequent paper.Gelation is the first appearance of a cross-linked network. It is the irreversible transformation of a liquid to a gel or rubber and it is accompanied by a small increase in the storage modulus. A distinction may be drawn between molecular or chemical gelation (the phenomenon) and macroscopic gelation (its consequence). Chemical gelation as defined by Flory is the approach to infinite molecular weight. It is an isoconversional point (αgel) that is observable as the first detection of insoluble, crosslinked gel in a reacting mixture (sol). Chemical gelation is also defined as the point where tan δ becomes frequency independent [2]. Macroscopic gelation may be observed as the approach to infinite viscosity, the first evidence of an equilibrium modulus, the G' = G" crossover in a rheology measurement, or as a loss peak in fiber and mesh supported systems. Vitrification is distinct from gelation. It is glass formation due to Tg increasing from below Tcure to above Tcure as a result of reaction. It only occurs when Tcure < Tg∞ and begins when Tg = Tcure (the definition of vitrification). Vitrification is reversible by heating: liquid or gel ⇔ glass. It causes a dramatic slowing of rate of cure as a result


33

of a shift in the reaction from chemical control to diffusion control. Vitrification is mechanically observable as a large increase in modulus and frequency dependent loss peak (note that vitrification occurs at shorter times with increasing frequency, i.e. it is not iso-conversional. This phenomenon is illustrated in the companion paper on Dynamic Mechanical Analysis of Thermosetting Materials. It is also observable by MTDSC as a step decrease in heat capacity, as demonstrated in Fig. 7 below during the slow heating of an acid anhydride cured epoxy [7]. 1 = heat, 2 = cool). Note that the onset of vitrification at ~100°C results in a diminished rate of cure under diffusion control until cure is complete at ~140°C.

Figure 7. Fig. 8 shows the non-isothermal cure of the same epoxy-anhydride at three heating rates as well as for the fully cured material. Only at the fastest heating rate does cure proceed to completion without vitrification. From Ref. 7, courtesey M. Reading. 1 = 0.2, 2 = 0.4 and 3 = 0.7°C/min. -0.2

2.1 -1 -1

1 1.7

4 -0.1

3

exo

1.3

2 1

-0.0 25

75

heat capacity /Jg K

non-reversing HF / Wg

-1

3 2

125

temperature / °C

Figure 8.

0.9 175

34

R. BRUCE PRIME

Thermomechanical analysis (TMA) measures linear dimensional changes in a material with temperature, time or applied load. Tg and the linear coefficient of thermal expansion (CTE) may be measured as well as irreversible expansion or contraction due to relaxation of stresses on heating through the glass transition. Creep or time-dependent strain under load may also be determined. Fig. 9 shows classical TMA in the expansion mode on heating [1]. CTE (α) may be measured from the slope of the TMA curve below and above Tg or it may be read directly from the derivative DTMA curve. CTE in the rubbery state (α2) is typically ~3x that in the glassy state (α1). Note the similarity of the DTMA curve to heat capacity through the Tg interval. Also note that obtaining a “textbook” curve such as this usually requires a preheat to just above Tg to relieve any residual stress.

Bair, Boyle, Ryan, Taylor and Tighe, SPE [Proc. Ann. Tech. Conf.] 33, 362 (1987)

Figure 9.

Figure 10.


35

Fig. 10 shows the thermal expansion and stress relief of a transfer molded integrated circuit (IC) device [8]. Note that the IC device is small enough to fit into the TMA sample chamber. On the first run notice the accelerating expansion as the temperature approaches Tg (~160°C). This is due to the relief of molded in and other residual stresses. From the second run CTE values may be calculated as well as irreversible dimensional change due to the relaxation of stresses. Thermogravimetric analysis (TGA) measures mass flow, ΔW, out of a material (volatility, degradation) as a function of temperature, time and atmosphere. Properties measured include evaporation of volatile components due to outgassing and cure, filler content for inert fillers (carbon/graphite contents can be estimated from nitrogenfollowed-by-air pyrolyses), thermal and thermo-oxidative stability, and degradative weight loss. Fig. 11 shows two-step isothermal TGAs in dry N2 for a UV cured acrylic coating cured at three doses designated “High”, "Typical”, and “Low” [9]. Note how weight loss at 150°C tracks cure dose suggesting that uncured acrylic monomer contributes to outgassing.

Best and Prime, Proc. SPIE - Int. Soc. Opt. Eng. 1774, 169 (1992)

Figure 11. Fig. 12 compares both room temperature and elevated temperature volatility of three acrylic coatings cured with the same “Typical” dose [9]. Coating 1 is from the previous slide. Note that Coating 2 exhibits the greatest RT weight loss but the lowest weight loss at 150°C. The authors attributed the high RT weight loss to greater water sorption capacity for this coating. Similar coatings are used for optical storage compact discs. The same authors showed that uncured acrylate monomers in these coatings will hydrolyze to form acrylic acid which is corrosive to the recording materials. Similar outgassing is also harmful inside hard disk drives. Fig. 13 begins the discussion of kinetics, especially cure kinetics but degradation and aging kinetics may be treated in the same manner [1]. The methodology described here may be characterized as model-free kinetics where the activation energy E is constant. The assumption of a single or overall activation energy applies when the only effect of temperature is to speed up or slow down the reaction. This assumption applies well to most thermoset systems, a notable exception being those that exhibit multiple DSC exotherms. As illustrated below, when E is constant conversion-time curves (or Tg–time curves through the Tg-conversion relationship) will be parallel on a ln(time)

36

R. BRUCE PRIME

plot, allowing construction of master cure or aging curves via time-temperature superposition.

Best and Prime, Proc. SPIE - Int. Soc. Opt. Eng. 1774, 169 (1992)

Figure 12. The shift factor aT is described by the Arrhenius equation where t is the time to constant conversion or constant Tg. Master curves are useful for succinctly summarizing all of the kinetic data and for predicting behavior at times and temperatures that may be of interest. It is recommended that behavior be predicted within the range of temperatures measured but estimates outside these limits can often be useful. Fig. 14. shows the same Tg – ln(time) curves for epoxy-amine cure shown earlier in Fig. 6 [4]. Note the parallel nature of the curves prior to vitrification which is demarcated with arrows. Master Curve T 1 > T2 Conversion (Tg)

aT

ln time aT = t2 / t1 = exp

ln reduced time E(T1 - T2) RT1T2

Figure 13.

t = isoconversional time


37

Epoxy-Amine Cure DGEBA-PACM-20 (1:1) Tg∞ = 178°C


Figure 14. Below in Fig. 15 is the master curve from shifting the above data along the ln(time) axis using the measured activation energy [4]. This curve clearly shows the reaction under chemical control (solid line) as well as the shift to diffusion control following vitrification. Epoxy-Amine Cure DGEBA / PACM-20 (1:1) Tg∞ = 178°C E = 15.2 kcal/mol, Tref = 140°C


at 140°C

Figure 15. Fig. 16 shows DSC conversion – time data for a low modulus adhesive with Tg close to room temperature [10]. One application of this adhesive is to produce bonded joints with low residual stress.

38

R. BRUCE PRIME

120 Conversion (%)

100 -20°C

80

24

60

100 150

40 20 0 1.E-01

1.E+01

1.E+03

1.E+05

Time (minutes)

Figure 16. The data above were shifted along the ln(time) axis by varying the activation energy E in an Excel spreadsheet to create master curve shown in Fig. 17 [10]. In this case E was determined from the best fit of the data. The reference temperature was chosen to be the maximum oven temperature allowed by the process, 120°C. Cure can be seen to be complete in 10 minutes at 120°C. 120

Conversion (%)

100 -20°C

80

24 60

100 150

40 20 0 0

2

4

6

8

10

12

Time at 120°C (minutes)

Figure 17. Fig. 18 shows the same master curve at a reference temperature of 180°C [10]. The question prompting this curve was “What temperature will be required for cure to be complete in 30 seconds?”.

39


120

Conversion (%)

100 -20°C

80

24 60

100 150

40 20 0 0

0.1

0.2

0.3

0.4

0.5

0.6

Time at 180°C (minutes)

Figure 18. This paper will conclude with presentation of an actual case study. The full paper [11] will be presented at the NATAS Conference in Albuquerque, NM, September 2003. The subject is a fast curing, two-component polyurethane. Parts are made by mixing the components in-line and rapidly processing and curing. The objective of this study was to determine the kinetic equation for cure for input into process modeling software. Below in Fig. 19 is a typical time-temperature profile for cure of the parts.

Temperature (°C)

200 150 100 50 0 0

1

2

3

4

5

Time (minutes)

Figure 19. Fig. 20 shows the time-temperature profile together with the desired output, the development of conversion along the profile. Following is the path taken to arrive at this endpoint.

40

Temp (°C)

R. BRUCE PRIME

200 180 160 140 120 100 80 60 40 20 0

100 90 80 70 60 50 40 30 20 10 0 0

1

2

3

4

5

Time (minutes)

Figure Temp (°C)20.

S a m p le : 2 0 0 8 4 1 E x p 8 S u p e r M ix S ize : 5 .2 2 0 0 m g M e th o d : C u re ra m p /T g ra m p C o m m e n t: c u re ra m p 1 0 °C /m in , T g ra m p

Conversion

F ile : C :\T A \D a ta \D S C \2 0 0 2 \2 0 0 8 4 1 \2 0 0 8 4 1 .0 0 9 i O p e ra to r: c rm R u n D a te : 1 7 -M a y -0 2 1 4 :4 8

DSC

0 .4

0 .3

cu re ram p , im m e dia te (n o iso the rm a l) @ 1 0 °C /m in S td D SC

8 8 .6 6 °C

Heat Flow (W/g)

0 .2

0 .1

0 .0

-0 .1

3 6 .0 9 °C 1 9 8 .1 J /g

-0 .2

-0 .3 -1 0 0 E xo U p

-5 0

0

50

100

1 50

200

2 50 U niv ers a l V 3.0 G T A Ins tr um ents

Tem perature (°C )

Figure 21. Fig. 21 shows the DSC at 10°C/min of the uncured two-component polyurethane. Note the onset of the cure exotherm near 30°C which necessitated rapid mixing and sample preparation and chilling of the DSC prior to measurement. While a small secondary exotherm was noted near 200°C it was decided to ignore this because it was small and possibly due to errors in mixing. We can now state the objective, which was to experimentally evaluate the parameters of the kinetic equation {E, f(α) and A} where dα/dt = k f(α), k = A exp[-E/RT]

(1)

41


and the step-by-step strategy to get there: 1. 2. 3. 4.

determine E from multiple heating rate measurements develop the conversion-time master curve from isothermal measurements determine f(Į) and k from the shape of the master cure curve determine A from the Eq. 1 as the only remaining unknown

The results of Step 1 are shown in the Ozawa plot of peak temperature versus heating rate in Fig. 22. Corrections applied to the raw data yielded an activation energy E of 14.4 kcal/mole. See General Reference for procedural details.

E ~ -R / 1.052 x Slope = 14.1 kcal/mole

Heating Rate (°/m)

100

Serie1 10

Exponencial (Serie1)

y = 9E+09e-7476,9x R2 = 0,9999 1 0,0026

0,0027

0,0028

0,0029

1 / Peak Temp (K)

Figure 22. In Step 2 conversion – time data were obtained by curing samples for various times in the DSC followed by 10°C/min scans to measure the residual exotherm. Conversion was measured from the residual exotherms and the heat of reaction ΔHrxn measured as the average of scans at 5, 10 and 20°C/min on the uncured thermoset. Scans on partially cured thermoset also gave Tg from which the Tg – conversion relationship was constructed. Shown in Fig. 23 is the unshifted conversion – time data.

42

Conversion (%)

R. BRUCE PRIME

100 90 80 70 60 50 40 30 20 10 0

30°C 45 60 80

1

10 ln (time) (minutes)

100

Figure 23. These data were shifted to a reference temperature of 80°C by means of Eq. 2, using the activation energy measured in Step 1. Note that 80°C ≡ 353K. T80°C = tT [E(T-80) / R(T+273)(353)]

(2)

Conversion (%)

The resulting master curve is shown in Fig. 24. 100 90 80 70 60 50 40 30 20 10 0

30°C 45 60 80

0

50 100 Time @ 80°C (minutes)

150

Figure 24. Note that the highest conversion on this master curve is ~90%. To be truly representative values closer to 100% are needed. The difficulty in achieving this with isothermal cures is the interference of vitrification. Since vitrification does not occur in the actual process because the profile temperature quickly rises to above Tg∞ = 104°C, it must also be avoided in the modeling. To accomplish this two data points were obtained from DSC cure profiles which simulated the process. The goal was to achieve one conversion just below 90% to overlap with the above master curve results and the other between 95 and 100%. To accomplish this DSC profiles were designed which would give equivalent isothermal times at 80°C (EIt80°C) of ~35 and ~125 minutes, where EIt is computed by summing along the time-temperature profile as indicated in Eq. 3 below.

43


EIt80°C = ¦t80°C = ¦ti [E(T-80) / R(T+273)(353)]

(3)

130° Profile: EIt = 36 minutes at 80°C

Temperature (°C)

160° Profile: EIt = 124 minutes at 80°C 180 160 140 120 100 80 60 40 20 0 0

2

4

6

8

10

Time (minutes)

Figure 25. Fig. 25 shows the DSC profiles together with their respective EIt computations. Samples were cured according to these profiles and their conversions determined from 10°C/min scans. These results were added to the master curve to give a Version 2 master curve, shown below.

Conversion (%)

100

30°C

80

45

60

60

40

80

20

130° Profile 160° Profile

0 0

50

100

150

Time @ 80°C (minutes)

Figure 26. In Step 3 data from the Version 2 master curve are analyzed to determine f(α). The data appear to have a general nth order shape. The cure is clearly not autocatalytic, missing the characteristic inflection. The general form of the nth order equation is dα/dt = k f(α) = k (1-α)n

(4)

where k is the rate constant and n is the order of the reaction. The data were fit to 1st and 2nd order forms of the nth equation shown below

44

R. BRUCE PRIME

1st order, n=1: -ln(1-α) = kt

(5)

2nd order, n=2: 1/(1-α) = 1 + kt

(6)

The data exhibited a very poor fit to the 1st order equation but an excellent fit to the second order equation, Eq. 6, with a high correlation coefficient as shown in Fig. 27. From the above 2nd order equation the slope, 0.235, is the rate constant at 80°C. y = 0,2348x + 1 R2 = 0,9818

35

1/(1 - alpha)

30 25 20

Serie1

15

Lineal (Serie1)

10 5 0 0

50

100

150


Figure 27. Step 4 may now be addressed. With E, f(α) and k80°C known the pre-exponential factor can be computed from the Arrhenius equation, Eq. 1, as 2.88E-10, providing a complete mathematical description of cure. From this kinetic equation the development of conversion along the profile shown initially was determined. The EIt80°C for the profile was computed to be 115 minutes from which a conversion of 96.5% may be estimated. The kinetic equation for cure also allows the computation of the master curve as shown in Fig. 28 below.

Conversion (%)

100 30°C

80

45 60

60

80

40

130° Profile 160° Profile

20

Kinetic Eqn

0 0

50

100


Figure 28.

150


45

References 1.

R. B. Prime, Chapter 6 “Thermosets” in Thermal Characterization of Polymeric Materials (E. A. Turi, ed.) Academic Press, San Diego (1997). 2. H. H. Winter, Polym. Eng. Sci. 27, 1698 (1987). 3. H. H. Winter, et al. in Techniques in Rheological Measurement (A. A. Collyer, ed.) Chapman and Hall, London (1997). 4. G. Wisanrakkit and J. K. Gillham, J. Appl. Poly. Sci. 42, 2453 (1991). 5. A. T. DiBenedetto, J. Polym. Sci., Par B: Polym. Phys. 25, 1949 (1987). 6. R. A. Venditti and J. K. Gillham, J. Appl. Polym. Sci. 64, 3 (1997). 7. B. Van Mele et al., Thermochim. Acta 266, 209 (1996). 8. H. E. Bair, D. J. Boyle, A. L. Young, and K. G. Steiner, Soc. Plast. Eng. [Proc. Annu. Tech. Conf.] 33, 262 (1987). 9. M. A. Best and R. B. Prime, Proceed. SPIE Int. Soc. Opt. Eng. 1774, 169 (1992). 10. R. B. Prime, unpublished data. 11. R. B. Prime, C. Michalski and C. M. Neag, Proceed. 31st NATAS Conference, Albuquerque, NM, (2003).


Characterization of Pharmaceutical Materials by Thermal Analysis Leonard C. Thomas TA Instruments 109 Lukens Drive, New Castle, DE 19720, U.S.A. [email protected] 1. Introduction Thermal analysis has been an extremely important analytical tool within the pharmaceutical industry for more than forty (40) years. Although the technique could easily be classified as mature, recent advances in Differential Scanning Calorimetry (DSC) have generated renewed enthusiasm in thermal analysis from pharmaceutical scientists. These new developments include Tzero DSC™ and Modulated DSC (MDSC®) which provide significantly improved performance in critical areas such as sensitivity, resolution and separation of complex transitions. This paper will illustrate the use of these improved DSC technologies on characterization of a wide variety of pharmaceutical materials including amorphous and crystalline drugs, drug delivery systems such as tablets and biodegradable polymer microspheres, proteins and frozen solutions used for freeze-drying. 2. Recent Advances in DSC Technology A brief description of the new technologies is provided to help explain how the improved performance is obtained over tradition DSC instrumentation. 2.1.

Modulated DSC®

An MDSC® experiment is performed on the same instrument as used for traditional DSC measurements. The difference between the two techniques is in the temperature profile applied to the sample and the deconvolution (separation) of the resulting heat flow signal into several components. Instead of the simple linear temperature change used by DSC, MDSC® uses two simultaneous heating rates; an average or underlying rate similar to DSC plus a sinusoidal or modulated heating rate. The average rate provides information equivalent to traditional DSC while the modulated heating rate provides unique information about the sample’s heat capacity. Figure 1 shows how temperature changes with time in an MDSC® experiment.

48

LEONARD C. THOMAS

Figure 1. As stated previously, the reason for applying simultaneous heating rates is to create additional information about the heat capacity or structure of the material. A brief examination of the equation used to describe the heat flow signal from a DSC or MDSC® experiment shows the benefit of the dual heating rates.

Where:

dH dT = Cp + f (T, t) dt dt

dH dt

is the Total heat flow due to the underlying or linear heating rate

Cp

is the Heat Capacity Component of the Total heat flow and is calculated from just the heat flow that responds to the modulated heating rate

dT dt

is the measured heating rate which has both an average (linear) and amplitude (modulated) component

f (T,t)

is the Kinetic Component of the Total heat flow and is calculated from the difference between the Total and Heat Capacity component.

Cp

dT dt

is the Reversing Heat Flow Component of the Total Heat Flow

CHARACTERIZATION OF PHARMACEUTICAL MATERIALS BY THERMAL ANALYSIS

49

Traditional DSC provides a single signal which is the sum of all thermal events occurring within the temperature range of the experiment. This often makes it difficult to interpret data or detect small transitions. MDSC® has a significant advantage over traditional DSC in that it measures both the Heat Capacity Component and the Total, and obtains the Kinetic Component from the difference. Separation of complex transitions into specific components greatly improves interpretation of results. In general, MDSC® provides the following advantages over traditional DSC.

increased sensitivity increased resolution separation of complex transitions more accurate measurement of crystallinity in semicrystalline materials direct measurement of heat capacity, either while programming temperature or holding it isothermal

Several of these benefits are illustrated on pharmaceutical materials later in this paper. 2.2.

Tzero DSC™ Technology

Until the recent introduction of this new approach to measuring absolute values of heat flow, DSC technology had not changed in a significant way since its commercialization in the mid-1960s. That technology was based on a single differential measurement and used either a heat-flux or power-compensation approach. The performance limitations of those early technologies could be seen in baselines that were neither flat nor very reproducible, and in peak-widths (resolution) that were much greater than expected from the melting of pure metals. Tzero DSC™ technology provides very significant improvements in baseline performance, sensitivity and temperature resolution of transitions. The improved performance of Tzero DSC™ results from a cell design which produces two simultaneous differential measurements and provides for the ability to calibrate the thermal resistance and heat capacitance of the individual sensors as a function of temperature. By knowing the actual thermal characteristics of a specific cell, any imbalance in capacitance or resistance of the sensors can be accounted for in the calculation of the absolute heat flow signal. The improved resolution of Tzero DSC™ technology is seen in Figure 2, a comparison of an indium melt with traditional and Tzero DSC™ technologies.

50

LEONARD C. THOMAS

Figure 2.

Tzero DSC™ provides a much more accurate heat flow signal than was previously available. Because of this, heat capacity values can be measured in a single run versus the three runs required for traditional DSC. This kind of performance results directly from the much more complete calibration of the physical components of the system and use of those calibration values in calculating the heat flow due to just the sample. Calibration factors are measured versus temperature and then continuously applied to the four-term heat flow equation used to calculate the sample heat flow.

q=−

§ 1 ǻT 1 + ǻT0 ¨¨ Rr © Rr Rs

· dT dΔT ¸¸ + (C r − C s ) s − C r dĲ dĲ ¹

Where: q

ΔT

= sample heat flow = qs - qr = temperature difference between sample and reference


ΔTo thermocouple

51

= temperature difference between sample and Tzero located between sample and reference sensors

−

ǻT Rr

= principle heat flow

§ 1 1 · = term to account for any imbalances in thermal Resistance ǻT ¨ - ¸ between 0 ¨© R r R s ¸¹ sample and reference sensors

R

(C r − C s )

dTs = term to account for any imbalance in heat Capacitance between the sensors dĲ

C

− Cr

= thermal Resistance of sample or reference sensors

dΔ T dĲ

= heat Capacitance of sample or reference sensors

= term to account for the difference in heating rates between the sensors (T4) and between the sample pans (T4P) during a transition in the sample

The advantages of Tzero DSC™ technology are illustrated on pharmaceutical materials in the applications section of this paper. In general, benefits fall into five areas:

flat and reproducible baselines higher sensitivity higher resolution single run measurement of heat capacity higher heating rate MDSC®

3. Pharmaceutical Applications

Because all transitions in materials involve the flow of heat (into the sample during endothermic events and from the sample for exothermic events), DSC is the universal detector for measuring a wide variety of transitions in pharmaceutical materials. This paper will focus on some of the most common measurements and illustrate the superior performance of Tzero DSC™ and MDSC® technology. These applications include measurement of:

52

LEONARD C. THOMAS

•

Amorphous Structure − Glass transition − Detection of amorphous material in semi-crystalline compounds

•

Crystallinity − Melting and crystallization − Purity − Polymorphs

•

Drug/Excipient Interaction

•

Protein Denaturation

•

Freeze Drying

•

Miscellaneous / Complementary Thermal Analysis Techniques

4. Amorphous Structure

The physical properties of amorphous structure are quite different from crystalline structure. Major differences include dissolution rate (faster bioavailability), storage stability and hygroscopicity, the tendency to absorb moisture or other solvents. It is, therefore, important to know if a drug or drug delivery system has an amorphous component. The most common DSC measurement of amorphous structure is the measurement of the glass transition. It is important to know both the size of the transition in heat flow or heat capacity units and the temperature (Tg) at which it occurs. The size of the transition provides quantitative information about the amount of amorphous structure in the sample, and the temperature identifies the point where there is a dramatic change in physical properties. Below the glass transition temperature there is limited molecular mobility while above there is high mobility that results in much lower viscosity and potentially much greater chemical interaction between components. Because of this, there is a general desire to store samples at least 40°C below their glass transition temperature. Since amorphous materials are often hygroscopic and because small amounts of moisture or solvent act to plasticize (lower Tg) the sample, it is important to measure the actual Tg of drug formulations as well as to control their volatile content. Figure 3 shows the glass transitions of an amorphous sucrose sample that was exposed to lab air (approx. 50% RH) for about thirty minutes. The first heat shows the midpoint of the glass transition centered near –28.70°C while the second heat to 100°C shows that it has increased by nearly 40°C to 11.8°C. Even this sample still has several percent moisture since the glass transition of completely dry sucrose is nearly 70°C.


53

Figure 3.

Figure 4 is an MDSC® experiment that shows how the size of the glass transition increases with increasing amounts of amorphous structure. The sample of Polyethylene Terepthalate (PET) was first quench cooled to produce a 100% amorphous structure then cooled at slower-and-slower rates to produce increasing amounts of crystalline structure. Even with a cooling rate of 0.2°C from above the melting temperature, the material retains a large amorphous component. To quantify the percentage of amorphous phase, the size of the glass transition (0.14 J/g°C) is divided by the size of the glass transition for a 100% amorphous sample (0.35 J/g°C).

% Amorphous Phase =

0.14 x 100 = 40% 0.35

Although this is a good approximation of the amorphous content of the sample, the actual content is probably slightly higher. Amorphous material that is sometimes trapped within crystalline lattices, often called the rigid amorphous phase, does not contribute to the step change in heat capacity at the glass transition and is therefore undetected.

54

LEONARD C. THOMAS

Figure 4.

One of the most difficult measurements for DSC is the detection of small amounts (99%) crystalline according to x-ray diffraction results, the crystal structure is also lost as the water evaporates. The resulting amorphous material crystallizes near 122°C and melts at 174°C. Since most endothermic transitions that can be confused with melting are kinetic events (evaporation, decomposition, and enthalpic recovery at Tg), it is relatively easy to distinguish between melting and these other transitions. This is done by changing the heating rate over the range of 1 to 20°C/min. The onset of a true melting peak will shift

58

LEONARD C. THOMAS

very little (

Thermal Analysis. Fundamentals and Applications to Material Characterization. Proceedings of the International Seminar: Thermal Analysis and Rheology, Ferrol, 30 June-4 July 2003

Thermal analysis of polymers: fundamentals and applications

Introduction to Thermal Analysis: Techniques and Applications