The Mastery and Uses of Fire in Antiquity
The Mastery and Uses of Fire in Antiquity J.E. REHDER
McGill-Queen's Unive...
207 downloads
1531 Views
19MB 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
The Mastery and Uses of Fire in Antiquity
The Mastery and Uses of Fire in Antiquity J.E. REHDER
McGill-Queen's University Press Montreal & Kingston· London· Ithaca
© McGill-Queen's University Press
2000
ISBN 0-7735-2067-8
Legal deposit quarter 2000 Bibliotheque nationale du Quebec Printed in Canada on acid-free paper This book has been published with the help of a grant from the Humanities and Social Sciences Federation of Canada, using funds provided by the Social Sciences and Humanities Research Council of Canada. McGill-Queen's University Press acknowledges the financial support of the Government of Canada through the Book Publishing Industry Development Program (BPIDP) for its activities. We also acknowledge the support of the Canada Council for the Arts for our publishing program.
Canadian Cataloguing in Publication Data Rehder, ].E. The mastery and uses of fire in antiquity Includes bibliographical references and index. ISBN 0-7735-2067-8 I. Pyrometallurgy - History. furnaces - History. 1. Tide.
2.
Ceramics - History. 3. Metallurgical
TN688·5·R442000
All figures were drawn by the author. This book was typeset by Typo Litho Composition Inc. in 10/12 Sabon.
To Nonnie, wife and best friend
Contents
Acknowledgments Foreword Preface
Xl
xiii
xvii
Introduction
3
1
The Nature of Heat and the Management of Its Temperature 9
2
How Furnaces Work
13
3 The Properties and Combustion of Biomass 4 Furnace Configurations for Biomass Fuel
25 38
5 Products Made in Antiquity in Biomass Fuelled Furnaces 46 6 The Manufacture and Properties of Charcoal 7 Combustion in Beds of Lump Charcoal 8 Combustion Air Supply for Charcoal
55
63 74
9 Furnace Configurations for Charcoal Fuel
84
10
The Reduction of Metals and the Functions of Slags
11
The Smelting of Copper
I2
The Smelting, Forging, and Properties of Iron
I I
3 122
101
Vlll
Contents
13 Fuel Consumption by Pyrotechnology in Antiquity 14 Fuel Su ppl y and Deforestation
153
15 Artifacts from the Operation of Furnaces
160
Appendices I
Combustion in Fuel Beds of Charcoal
2
Pressure Drop in Tuyeres and Fucl Beds and Power Required 175
3 Natural Draft in Fuel Beds
167
180
4 A Furnace to Reliably Make a Bloom of Iron Glossary References
195 199
189
145
Tables and Figures
TABLES I
Pro-Forma Heat Balance
2
Proximate Analyses, Carbon Contents, and Heat Contents of Selected Biomass Fuels on Oven-Dry Basis 29
20
3 Biomass Fuel Consumption by Furnace Product
152
4 Effects of Air Supply Rate on Temperature and on Its Extent in a Coke Fuel Bed 17 I FIGURES I
Effect of Furnace Size on Space Velocity Necessary to Maintain Temperature 16
2
Effects of Space Velocity and Fuel Reactivity on the Height of the High Temperature Zone 18
3 Heat Loss Rate as a Function of Wall Thickness and Inner Face Temperature 23 4 Effects of Moisture Content and Excess Air on Flame Temperature of Biomass Fuel 28 5 Reverberatory or Air Furnace
43
6 Gas Composition and Temperature in a Charcoal Fuel Bed 65
x Tables and Figures 7 Approximate Plumes from Tuyeres in Fuel Beds
68
8 Temperature Distribution in a Fuel Bed with Single Tuyere 86 9 Iron Ore Reduction in a Small Bowl Furnace
87
IO
Carbon Content of Iron Blooms versus FeO Content of Associated Slag 126
I I
General Arrangement of Experimental Bloomery furnace I91
Acknowledgments
It was on the gentle but persistent encouragement of Professor Emeritus Ursula M. Franklin that the writing of this book was started. I have taken advantage of her practised teacher's eye to try to keep the writing style reasonably simple and interesting, while being informative. The subject matter has many facets, and it is all too easy to drift into a byway that can confuse the issue. Her continued interest and support in both the subject matter and its transmission have been invaluable. Dr Martha Goodway of the Smithsonian Institution and Professor Bruce Trigger of McGill University read an early draft and made useful comments. The support of Professor Alex MacLean and the late Professor Alan Miller of the Department of Metallurgy and Materials Science in the University of Toronto is appreciated, especially in making available space and the services of a technician for the construction and operation of a bloomery furnace to make several blooms of iron. In particular I appreciate the excellent work done by my editor, Maureen Garvie, in smoothing out the wrinkles and style changes in the manuscript, occasioned by its having been written over a period of several years, and the ease of working with coordinating editor Joan McGilvray.
Foreword
It is a great privilege for me to write this foreword to the book by my friend and colleague J.E. Rehder. The book is an important and quite unique work, and I would like to illustrate its great potential and at the same time give the reader an indication of the philosophy behind the work. As a contribution to the reconstruction of the past, Rehder's book is essentially a companion, a knowledgeable friend to have by one's side while studying ancient objects and thinking about the accomplishments of those who made them. This companion, like its author, is both a scholar and a practitioner, steeped in the complexities of the field but waiting to be asked, trying not to overwhelm the inquirer with unessential details - yet at the same time also not permitting illusions of quick comprehension or an easy glossing of the intricacies of technical processes. Like any real teacher and friend, the book makes it easy to come back to a particular question as more and deeper knowledge is needed. The appendices serve this process by providing additional information without intimidating the initial enquirer, and it is here that the basic philosophy of Rehder's approach to the study of ancient materials and processes becomes apparent. Two central considerations inform this approach - evident not only in this book but also in his many scientific papers in this field. The first relates to artifacts as primary historical evidence, and the second focuses on the requirements for scientific and technical rigour when interpreting ancient processes and products.
XIV
Foreword
First, the "reading" of ancient artifacts, based on their microstructures and compositions, provides essential sources of historical evidence. There is, however, an important proviso here: those "reading" and interpreting technical evidence must know the "grammar" of technology, i.e., the relevant laws of physics and chemistry. And in this proviso we find the second strand of Rehder's approach - to give to the archaeologist and historian technically and scientifically correct and consistent guidance. Today's engineers may express the processing variables in terms of symbols and equations, while the ancient artisans transmitted the knowledge of their hands and minds in the language of their craft and culture; yet the underlying reality is the same. It is on the bases of these premises that Rehder gives those who do not have an adequate pertinent scientific or technical background the tools to assess the basic technical processes of antiquity. The book begins with an outline of the natures of temperature and of heat, acknowledging that the combustion of biomass was essentially the sole source of manageable heat in antiquity. The reader is then quickly introduced to the modes of heat conduction and loss, and to the realization that if materials are subjected to higher than ambient temperature, problems of containment arise - and furnaces and their precursors appear. At the same time it is made clear that the combustion of biomass does not only result in the production of heat but the chemistry of burning changes the composition of the products of combustion, often resulting in complex composition and temperature profiles in the furnace. These in turn must be taken into account in the design of furnaces and the tasks to which they are to be put. Rehder looks at charcoal, the synthetic fuel used particularly for metallurgy throughout antiquity and into the early twentieth century A.D. In a series of quite unique compilations on the making of charcoal, he provides information on its technology in a concise and accessible form not available anywhere else, and the appendices guide the more technically sophisticated to some of the roots of the issue addressed. Having laid this thorough foundation, Rehder turns his attention first to the furnace designs appropriate to charcoal fuel and then the tasks to which such furnaces arc put. Here we find ourselves in the heartland of the development of the smelting and uses of metals, the important role of slags, illuminating details on the metallurgy of copper, and the complexity of the production and manipulation of iron. The reader will find these chapters powerful and totally convincing and will realize that it is the author's thorough understanding of fur-
xv Foreword nace and high temperature reactions - in terms of chemistry, physics, and spacial arrangements - that brings about this seamless understanding. The cobwebs of antiquarian jargon and all hints of magical knowledge or long-forgotten processes known to the ancients vanish in the light of the clear, consistent rationale offered here. The book's final chapters of fuel consumption and deforestation in antiquity round out the work in the same spirit of unbending integrity, so that by taking into account the growth rates of forest, some conclusions on causes of deforestation in antiquity are arrived at that are quite different from those in the current literature. Ursula M. Franklin
Preface
The text which follows considers the subject of pyrotechnology from the perspective of the unchanging rules of physics and chemistry, and my fifty years of experience in industrial pyrotechnology in operations, research, and consulting, with pu blication of more than one hundred papers on the subject. The book is an expansion of a series of papers on ancient pyrotechnology that I have been publishing occasionally in archaeological journals since I986, as a senior research associate in the Department of Metallurgy and Materials Science in the University of Toronto. It also uses data from my industrial research and development papers published some time ago and contains some of my unpublished experimental work on replicated ancient furnaces. The approach I have taken is from a furnace operator's point of view, so that the writing style is that of the natural sciences rather than that of the social sciences. This can create problems in understanding across cultures because of differences both in accustomed jargon and in habits of style, but these arc perennial difficulties, and a conscious effort has been made to avoid jargon or to explain it when unavoidable. While development of furnace practices through time is necessarily involved, this book is not intended as a history of pyrotechnology. Many of my examples of ancient pyrotechnology are taken from the archaeological literatures of the east end of the Mediterranean basin where the evidence is over the longest term, in most abundance, and most easily available. Furnaces operate on the same physical and chemical bases everywhere, and I refer to, for example, South America or Africa only for interesting or unusual practices.
xviii Preface Two existing books may seem to be on the same subject as this one, but in fact are not. One is The Beginning of the Use of Metals and Alloys, edited by R. Maddin (M.l.T. Press, I988), a collection of thirty papers from a conference in Zhengzhuo, China, in I986. Their contents deal almost entirely with evidences of early metallurgy, and only two papers describe iron smelting furnaces, each with insufficient technical detail to permit analysis of their operation. The second is the excellent book by P.T. Craddock, Early Metal Mining and Production (I995), which discusses the mining of ores and the production of a wide range of metals in useful historical detail. However, only about one-third of that book concerns matters in which there is some overlap with the material here, and its author has an understandable tendency to take the accustomed archaeological consensual approach, which is not always reliable on testable matters of natural science. Moreover, like the Maddin book it discusses only the branch of pyrotechnology that deals with the smelting of metals, while the present one includes other important products of ancient pyrotechnology such as ceramics, lime, and glass. Pottery, lime, and the two metals copper and iron were made in increasingly large quantities throughout antiquity, in time effectively on an industrial scale. The organization of this book, in addition to reflecting the nature or kind of fuel used, also is divided according to the products made. These were fired clay, lime from limestone, metals from the reduction of ores, and to a lesser extent, glass from sand. This list is short and simple, but it subdivides into many kinds of products used in quantities that increased throughout antiquity. The complexity of their production technologies increased exponentially from clay to lime to metals, and the ability finally to smelt metals from their ores took millennia to become a reasonably reliable practice. This increase in technical complexity accounts for the considerable amount of space given here to the smelting and uses of copper (for making bronze) and of iron. These were the two metals in major use in antiquity, iron having a wider range of useful mechanical properties than copper and its alloys. Moreover, the orcs of iron arc much more widely distributed geographically and on the average are richer than are those of copper. However, the metallurgy of iron is even more complex than those of copper and bronze, and the ability to smelt iron that could be hot-forged to a bar or sheet of strength and ductility superior to those of bronze required successive considerable advances in furnace technology through nearly two millennia. For these reasons, chapter I2, on the smelting and properties of iron, is the longest in the book. Other notes on organization: The first three appendices are technical in nature, intended for serious students of furnace technology, and can
XIX
Preface
be ignored by others. Also, illustration is limited to line drawings and necessary graphs. While a valid criticism might be made that there are not enough illustrations of ancient furnaces and their associated equipment, there is a basic reason for this lack. The book covers a very long time span - about IO,OOO years - during which the technology of the uses of fire was slowly developed through simple trial and error. Thousands of different furnace shapes and kinds were tried, by definition mostly unsuccessful, and remnants of these attempts fill the extensive published archaeological record. There is then the question of selection, since this book is not intended as a history of furnace architecture. Reconstructions have been made, but they necessarily involve speculation and it is one of the purposes of this book to decrease the extent of such guesswork by supplying the technology of furnace mechanisms. Thus only a moderate number of functional graphs and a few sketches have been included.
The Mastery and Uses of Fire in Antiquity
Introduction
The material fabrics of nearly all settled civilizations have by and large consisted of things that exist only because of pyrotechnology - the generation, control, and application of heat, which at sufficient temperature can alter the properties and compositions of all materials. When the materials are of the earth itself, since antiquity the resulting products have formed a large part of the material bases of human wellbeing. The list of such products is today extensive, as a little thought will suggest: steel beams, reinforced concrete floors, brick walls, glass windows, metal and plastic pipes for water, gas, and sewage, copper wires for electricity and communication, kitchen ware - as well as trains, automobiles, airplanes, and so on. Such accomplishments are the result of some ten thousand years of development of the intentional use of fire for other than warmth and food. The early archaeological evidence is slim and scattered, but it noticeably increases towards the end of the most recent ice age. The firing of clay into pottery was apparently the first major product of pyrotechnology, but early evidence of another considerable application was the extensive use of lime plaster at Cayonu Tepesi in 6500 B.C., where a house floor was found to be made of 4.5 cubic metres of lime plaster. The decomposition of limestone into quicklime requires heat at close to I,OOOoC; however, reduction of iron ore to iron metal not only requires heat but involves complex and invisible chemical changes at different temperatures, and this took another five thousand years to master. By the close of antiquity, defined here as the collapse of the Roman hegemony before the middle of the first millennium A.D., a plateau of
4
Mastery and Uses of Fire in Antiquity
development of the uses of fire had been reached. This was exemplified by the following quotation from Pliny the Elder in his Natural History of 77 A.D.: "Fire takes sand and melts it into glass. Minerals are smelted to produce copper. Fire produces iron and tempers it, purifies gold, and burns limestone to make mortar that binds blocks together in buildings." Pliny overlooked, probably from its familiarity, the firing of clay into pottery, which probably was the earliest widely used product of high temperature heat. When our modern extensive abilities in the uses of heat arc considered, arguably the growth of skill in such an important and complex technology could be considered as the longest continuous intellectual endeavour undertaken by humankind. Indeed there is justification for considering that the extent and the complexity of uses of heat is a measure of the level or quality of a civilization. This position has been noted by Mumford (I946), White (I962), and Kef (I967). Caasen and Girifalco (I986) have created a chart showing that a linear relationship exists between per capita consumption of heat energy and personal income for twenty-eight countries around the world, from a Pacific island to the United States. The practice of pyrotechnology evidently was basic to the civilizations of antiquity, and so is of proportionate importance to archaeology and to anthropology. However, as far as I know, the subject matter and its importance receive only minor attention in the teaching of archaeology, and there also seems to be no publication that discusses in useful detail how pyrotechnology was practised in antiquity, using the limited knowledge and sources of heat then available. Such information is essential to both the tracing of its development through its artifacts and to the understanding and replication of ancient practices, and it is the objective of this book to supply it on the basis of modern knowledge of the subject. Curiously, the word "pyrotechnology" is in few dictionaries, though its entomology is obvious. Apparently its earliest appearance was in Italian as the title of the famous work of Biringuccio published in 1540, Pirotechnia, which described in useful technical detail the contemporary uses of heat to alter materials. In modern times the late Theodore A. Wertime used the word in this sense in his many well-known publications on ancient pyrotechnology, as have others, and it is now common in archaeology and anthropology. The history of the development of pyrotechnology following the end of the last icc age was outlined in a collection of papers edited by Theodore and Steven Wertime (I982). Yet almost without exception, this and other modern studies of ancient pyrotechnology have dealt with the artifacts that are the products of furnaces and, to a much lesser
5 Introduction
extent, with the functioning of the furnaces themselves. Considerable attempts have also been made to replicate the operation of furnaces recovered by archaeology and anthropology, but with limited success, since little has been published in the archaeological literature on the basic principles on which such furnaces work. The appropriate conditions are today well known from modern industrial experience, and it is the purpose of this book to add to the resources of archaeology and the other disciplines that deal with past civilizations a detailed discussion of how pyrotechnology was practised throughout antiquity. This discussion makes clear what kinds of artifacts need to be collected, and in what detail, to supply the necessary data base for detailed analysis of particular furnace operations, permit such analysis, and guide the construction and operation of replica furnaces. THE STUDY OF PYROTECHNOLOGY AS PRACTISED IN ANTIQUITY
Details of the first long stage of development of pyrotechnology during the Pleistocene are largely speculative, with archaeological evidence spread very thinly through time. No written record exists until about 3000 B.C. at Sumer, but even then, throughout most of antiquity, not only was there a low levcI of general literacy but the pervasive opinion among the literate and the powerful was that matters dealing with use of the hands for other than fighting were not worth recording. Xenophon in 400 B.C. noted, "What are called the mechanical arts carry a stigma and are rightly dishonoured in our cities - it is not legal for a citizen to ply a mechanical trade." The question of the social position of craftsmen in antiquity is complex and has been discussed in detail by, for example, Burnford (I972), and its connection with technical innovation in antiquity has been treated by Finley (1985). As an example of minimal information, iron is mentioned in literatures from the cuneiform of the Hittites in the second millennium B.C. to the Latin of Plutarch in the second century A.D., but very little was written about details of any of the technical or manufacturing practices involved. Our knowledge today of what was made, by what procedures, and for what purposes therefore depends almost entirely on interpretation based on modern scientific and engineering knowledge of the artifacts recovered by archaeology during about the past century and a half. An excellent book on artifacts in general published in I964 by Hodges has been reprinted several times. However, the objectives here are limited to those artifacts that have been produced by the influence of heat, and to the methods by which this was done.
6
Mastery and Uses of Fire in Antiquity
MATERIALS
A general classification of materials necessary or convenient to civilized human life in various ways is that they are either organic or inorganic in composition. Organic materials have their origin in some kind of life, such as wood, plant fibres, foods, leather, and horn. Composed largely of carbon, oxygen, and hydrogen, they are generally combustible but start to decompose when their temperature is raised above about 250°C. Most varieties also resist the destructive effects of the environment poorly over time. Inorganic materials such as clays, stones (some, like limestone, being consolidated skeletons of previously living matter), sand, and minerals of various metals are generally harder and stronger than organic materials. When heated to elevated temperatures, they change in composition and properties to create new materials that are useful to humans. Lime mortar, for example, is used to join building stone and to make plaster walls, floors, and ceilings. Fire-hardened clay can be made into pottery, and minerals can be sources of metals for tools and weapons. Inorganic materials' resistance to environmental effects over time varies but is better on average than that of organics, so they tend to predominate as artifacts. The practice of pyrotechnology in antiquity discussed here deals almost entirely with the effects of heat on such inorganic materials. Throughout antiquity the necessary heat was generated by the combustion of organic materials, and only two fuels were used. The primary one was some form of biomass, almost entirely organic in composition: vegetable or other organic matter living or recently dead such as wood, shrubs, vine prunings, straw, and dung. On the artifactual evidence fossil biomass such as coal, oil, and natural gas was very little used. Theophrastus (370-288 B.C.) makes mention in his History of Stones of the use of lignite by metalworkers in Italy and Greece but includes no useful detail on how, what for, and to what extent it was used. Coal was used in Roman Britain but only for space heating (Birley I977), and in China a few centuries later (just outside our time frame here) for pyrotechnology (Needham I958). The other fuel was charcoal, made by heating biomass out of contact with air to above a temperature of about 250°C. It then decomposes to form charcoal, which is largely carbon, and a gas that is combustible but in antiquity was wasted to atmosphere. Charcoal burns in a very different pattern from that of biomass, and this difference is sufficiently large to require different furnace structures for the combustion of biomass or of charcoal. It also gives abilities to reach different maximum temperatures and to make different chemical changes in materials.
7
Introduction
The heat developed by the combustion of each fuel was therefore used for quite different purposes. That from biomass was typically for making pottery, quicklime, and glass; and from charcoal for smelting ores. For these reasons the discussions in this book of furnaces and the uses of heat are divided on the basis of the fuel used. DIFFERENCES BETWEEN ANCIENT AND MODERN PRACTICES
Pyrotechnology was conducted in antiquity differently from today for several reasons. First, heat was generated almost entirely by combustion of some form of biomass and to a lesser extent, charcoal. The enormous amount of heat used for today's extensive pyrotechnology is still largely from combustion but of one of the fossil fuels. A second difference involved the supply of the all-important air necessary for combustion. Throughout antiquity this was limited to bellows or blowpipes driven almost entirely by the power of human muscle; today the necessary power is generated by engines themselves driven by combustion. A third difference followed from the low population density in antiquity; lower demand for product could thus be supplied by smaller furnaces. The fourth and intractable difference was complete lack of a tradition of science throughout antiquity. All exploration of what could be done and how to improve on it was by trial, error, keen observation, and a good folk memory that became embedded in mythology. Some statements made in the course of this book about the practice of pyrotechnology in antiquity, particularly in metallurgy, will be found to differ from ideas found in the literature of archaeology until quite recently. However, these statements arc easily verifiable, and in every case are the results of either modern published industrial research and experience, which has been little noticed in the archaeological literature, or of associated specific experimental work published by myself, or both. The more important issues dealt with in these pages include demonstrations that maximum furnace temperatures attainable in antiquity with biomass fuel could be about IAOOoC, and with charcoal, a little over I,6oooc, sufficiently so that temperature limitation could be necessary; that carburization of iron while being heated directly in a forge fire does not occur; that combustion air is not preheated to a useful extent in long tuyeres, and in any case preheat is not necessary to make high carbon steel in charcoal fuelled furnaces; that bellows air supply can generate temperatures in a charcoal fuel bed about 400°C
8
Mastery and Uses of Fire in Antiquity
higher than is possible with a blowpipe, making possible the smelting of iron, and is many times more efficient in the use of human muscle power to move air; and that the practice of pyrotechnology may not have been the major factor in the deforestation that was noted in the ancient literature.
I
The Nature of Heat and the Management of Its Temperature
Heat has three singular and interesting properties that affect how it can be generated and used. The first and most important is its intensity or temperature, which measures its ability to affect materials. The second is that heat flows instantly from a higher to a lower temperature. The third is that heat cannot be confined, since there is no known material that does not conduct heat to some extent. From the instant it is generated, heat leaks everywhere and constantly, but since different materials have different conductivities, some control can be achieved over the rate of flow or loss of heat. It is because rates of flow are involved that a time factor enters all discussion of the generation and use of heat. When heat is generated by combustion of a fuel, its quantity and temperature can be easily and directly controlled by the rate of air supply. Then ease of generation and control, and incessant but controllable loss, are the key factors that make possible and then govern the generation, control, and application of heat for human benefit. Because heat escapes so readily, a container for its generation is necessary, which should obviously be made of a material with suitably high softening temperature. The material should also have low thermal conductivity to slow the rate of heat loss to surroundings, which both saves fuel and can increase the temperature of combustion. Such a container is called a furnace. The earliest evidence of furnace use is apparently early post-Ice Age, but it existed certainly by 7000 B.C. at Catal Huyuk in the Near East. Enclosure of fire was a major advance in pyrotechnology; furnace development then began to accelerate into
10
Mastery and Uses of Fire in Antiquity
a wide variety of shapes, sizes, and internal designs, governed by the kind of fuel used and the kind of change to be made in the treated material. THE MANAGEMENT OF TEMPERATURE
The chemical compositions of all materials are sensitive to heat, and the effectiveness of heat in changing their chemistry and properties increases exponentially with temperature. There are also pronounced threshold effects of temperature, in which a particular change in a material will not take place until some minimum temperature is reached, no matter what quantity of heat is supplied. Obvious examples are that water will not boil until its temperature reaches 1000e and a metal will not melt until its specific melting point is reached. The temperature of heat and the ability to control it are therefore of primary importance to pyrotechnology. CREATION AND CONTROL OF TEMPERATURE
In the combustion of a fuel, the maximum temperature of the heat generated can be calculated from the thermochemistry of the combustion reaction, and is called the adiabatic flame temperature or AFT. It varies with the composition of the particular fuel, such as biomass versus charcoal, and with the source of combustion oxygen, such as human breath through a blowpipe versus ambient air from a bellows as source of oxygen. AFT is calculated on the basis of zero heat loss, but in practice there is always some heat loss so in a real furnace the AFT of the fuel can be only approached. MAXIMUM TEMPERATURE IN A FURNACE
When using a particular fuel and source of combustion oxygen, the rate of heat loss from a given furnace is one major factor in determining the actual maximum temperature attainable in it. The other factor is the rate of heat generation, since if this is less than the rate of heat loss, furnace temperature will decrease. An analogy is trying to fill with water a bucket with a hole in it; if water leaks out faster than it is added, it can never be filled. The maximum actual temperature in a given furnace will be less than the AFT of the fuel by the same percentage that the rate of heat loss is of the rate of heat generation. For example, with a heat generation rate of 1,000 MJ (megajoules) per hour and a heat loss rate of 100 MJ per hour, the heat loss rate is IOOir,ooo or 10 per cent of the
II
Heat and Temperature Management
heat generation rate. Then with a fuel AFT of 2,000°C, maximum furnace temperature will be 2,000 - (0.10 X 2,000) = I,800 o c. HEAT LOSSES
Heat losses from furnaces are by conduction through containing walls and then by radiation and convection from the exterior to surrounding air. Heat is also lost in flue gases. In practice, then, the management of temperature consists of selecting a fuel, which defines the AFT and maximum possible temperature, and then burning it in an enclosure so that heat is generated at a rate sufficiently exceeding the rate of heat loss from the furnace used to develop a desired fraction of the AFT of the fuel. As will be shown in chapters 3 and 7, the rate of heat generation is controlled directly by the rate of supply of combustion air, and the rate of heat loss depends largely on the furnace size, wall thickness, and materials of its construction. QUANTITY OF HEAT
In today's international SI units, the standard quantity of heat is a called a joule. As this is a very small unit equal to one watt per second or to 0.239 gm calories in the old CGS system, energy units of practical size require multipliers. These are thousands of joules, or kilojoules (KJ), for example, oak wood having a potential heat energy of 18.7 KJ per kg; millions of joules or megajoules (MJ), such as a ceramic pot at a temperature of 1,000°C containing 0.84 MJ per kg of heat energy; and thousands of millions of joules or gigajoules (GJ), such as 4.0 GJ of heat per hour being generated in the firebox of a large pottery kiln. MEASUREMENT OF TEMPERATURE IN ANTIQUITY
The temperature of heat is measurable by its effects on materials, such as the expansion of mercury in a thermometer, or the electrical voltage generated at the junction of two dissimilar metals (a thermocouple), or radiation in the visible or ultraviolet spectrum which can be measured by optical instruments. Since the control of and the ability to reproduce a given temperature is essential to being able to repeat a successful heat process, a note is in order as to how temperature was measured without such instruments in antiquity. Evidently this was, and today still to some extent is, done simply by noticing the colour of the object being heated, since there is a direct and quantitative relationship between visible colour and temperature. A practised eye can (in my and others' ex-
12
Mastery and Uses of Fire in Antiquity
perience) in this way estimate a temperature and recognize its reproduction to within at most 20°C of the temperature measured by a modern pyrometer. This is close enough for most practical purposes. Colours represent specific wavelengths of visible light that are functions of temperature and are difficult to describe in words; but, for example, the lowest temperature that can be seen with the naked eye in a darkened room is about 550°C, visible as a very faint dark-red glow. A bright red is 850-9 50°C, a yellowish red is 1,050-1,1 50°C, and a white colour is above about 1,500oC. Higher temperatures are simply more intensely white and can be estimated only by use of a darkening glass. Ranges are given here because of personal differences in perception or in naming of colours. The measurement of the temperatures of gases is a complex subject, and their temperature can be estimated visually only if they contain enough finely divided particles of matter, such as ash particles or unburned carbon, that radiate energy at wavelengths visible to the human eye. Then, even with modern thermocouples and radiation pyrometers, accurate measurement is difficult due to the complexity of flow patterns at different temperatures in a combustion system, caused by poor mlxmg. As will be noted in any combustion engineering handbook, the reasonably accurate determination of the temperature of gas within a furnace, or of average temperature of flue gases, requires many measuring points and a statistical analysis of their readings. For this reason measurements made on modern replicas of ancient furnaces are largely meaningless if taken at single points. An example is given in chapter 7 of how temperature measurements of a gas flow in a tuyere, which were seriously in error due to faulty technique, led to conclusions about process that were seriously mistaken.
2
How Furnaces Work
HEAT LOSSES
As a general custom the word "furnace" implies that temperatures above about 250°C are involved, as this is the level above which most organic materials start to decompose and form a char. Enclosures of fire used for lower temperatures are mostly for food preparation, which is incipient decomposition, and are called ovens. Two common inorganic materials whose properties can be usefully altered by such moderate temperatures are gypsum (discussed in chapter 5) and flakeable stones such as flint and chert (not discussed here). A kiln is a high temperature furnace that was used through antiquity for firing ceramics and making lime from limestone. A primary objective of a furnace is to be able to provide an enclosed space in which an atmosphere of controlled temperature and composition can be created and maintained. A secondary objective is to use as little fuel per hour as possible. This depends mostly on the heat loss rate from the furnace, which must therefore be kept as low as possible. This becomes a question of shape, size, and materials of construction. Since heat starts to flow from the instant that it is created, and since all materials conduct heat at rates characteristic of the material, heat is continuously lost through the furnace walls. There are in addition two other ways in which heat is "lost" from a furnace. One is in the form of the gaseous products of combustion of the fuel, which must escape from the furnace and carry amounts of heat that depend on the conditions of combustion. The other is the heat absorbed by the objects in
14
Mastery and Uses of Fire in Antiquity
the furnace that are being heated, ranging from an ore to be smelted into metal to pottery to be hardened. This last "loss" is of course the objective in operating the furnace, but it must be taken into account in order to understand and so control the furnace's operation. The first consideration in designing a furnace is the maximum temperature to be reached, and this affects the choice of fuels. Then the furnace must be made of material that not only resists the softening effect of heat at that temperature but can slow the rate of loss of heat to surroundings. The shape of the furnace must also assist the uniformity of transfer of heat and products of combustion, from burning fuel to the objects to be heated, and the addition of unheated objects and removal of heated ones must be convenient. BATCH VERSUS CONTINUOUS OPERATION
From a functional viewpoint, furnaces can be operated in either of two ways. One is to treat material in batches, by heating it to some desired temperature, holding at temperature if necessary, cooling both furnace and contents to ambient temperature, removing the processed material, refilling the furnace with fresh material, and repeating the heating cycle. The firing of pottery in a kiln is an example. The other method of operation is to maintain part or all of the interior of the furnace at some desired temperature, passing the material to be heated continuously through the furnace. An example is a shaft furnace with combustion and high temperature near its base, and lower temperatures at higher levels, for smelting ores of metals. These two procedures produce large differences in the amount and rate of heat loss. In the batch method, the furnace structure is unavoidably heated along with its contents, and all of this heat is wasted to open air on cooling the furnace for emptying and refilling. Since the weight of the furnace structures of batch furnaces can be several times that of the contents of a single filling, the heat loss per unit weight of material heated is large, and the thermal efficiency of use of the fuel is correspondingly low. In antiquity this was between about .5 to 4 per cent, which resulted in high fuel consumption per unit weight of product. In the continuous method, the empty furnace body is preheated by combustion to "fill" the furnace walls with heat, so that when material mixed with fuel is added and passed through the hot furnace, the heat loss rate through the walls is stable and temperature is under better control. The ratio of furnace weight to that of total product becomes low, and thermal efficiency is then much increased. In antiquity this was 10 to 25 per cent, depending on the size of the furnace, and fuel consumption was accordingly lower.
15
How Furnaces Work
The one major qualification of the use of continuous processing is that the material to be heated must be in small pieces relative to the size of the furnace, in order to be able to flow with the lump fuel down the furnace interior. Preheating is usually done with miscellaneous biomass fuel since charcoal has a considerable labour content, as will become evident in chapter 6. There is of course little to be gained by preheating a batch furnace such as a pottery kiln. TEMPERATURES ATTAINABLE
The maximum actual temperature attainable in a furnace (discussed in chapter I) was demonstrated experimentally some years ago. A wellinsulated shaft furnace 410 mm inside diameter (I.D.) was used, containing only coke fuel. It was supplied with combustion air at the rate of 120 m 3 per m 2 of internal cross-section per minute, considerably above that used in antiquity. The AFT of the fuel was 2,070oC and the measured maximum temperature of the fuel bed was 2,020 o C or 97.6 per cent of the AFT (Draper et al. 1979). It mnst be noted, however, that there was nothing but fuel in the furnace, and if material had been added to be melted or smelted, the heat absorbed by it would be a "loss" and so would have decreased the maximum furnace temperature by a predictable amount. The AFT of biomass fuel is more variable than that of charcoal, as will be shown in chapters 3 and 7; but it can be as high as I,6oo°c and the temperature in a kiln can then reach I,400oC or 87 per cent of the AFT. The lower percentage of AFT than in the Draper experiment is due to the higher heat loss rate in typical kilns. EFFECTS OF FURNACE SIZE AND SHAPE
Since heat is generated by combustion in a volume of space, and heat losses to and through walls arc through enclosing surfaces, the heat loss rate increases with the ratio of surface area to volume enclosed. As any container becomes smaller or more complex in shape, the ratio of its surface area to volume increases, and so as furnace size is decreased a larger proportion of the heat generated is lost to and through walls. This results in decreased thermal efficiency, and also means that heat input rate must be higher in a smaller furnace to achieve a given internal temperature. This effect rapidly becomes seriously limiting as furnaces become quite small, as is shown in a pro-forma way in figure I for shaft furnaces. In this example wall thicknesses were taken as 10 per cent of their inside diameter, and an inner face temperature of I,200 o C was to be maintained after five hours operation. To maintain I,200°C in a fur-
16
Mastery and Uses of Fire in Antiquity
50
'"
10
O~----r-----~--~~--~----~ :zoo 400 o I, (JOt) 600 furrtA-t!t!- l D. w",
eoo
Figure I Effects of Furnace Size on Space Velocity (Heat Input Rate) Necessary to Nlaintain Temperature
nace with less than about 100 mm I.D., the necessary rate of flow of combustion air could blow the fuel out of the furnace. Throughout antiquity, this size effect was an important consideration in furnaces, which were small, particularly in early pyrotechnology, because of smaller population density and thus limited objectives in quantities of material to be heated. The general tendency in antiquity of furnaces to be rounded in shape (a round furnace or circular domed kiln will have lower heat loss rates than the same volumes enclosed by a square or cube) is evidence that the effects of size and shape were recognized. A mechanical effect of the shape of a furnace is how it directs and controls the flow of hot products of combustion through the contents of the furnace. In the case of a kiln filled with pottery, the arrangement of the ware can thus affect the uniformity of temperature reached in various parts of the kiln. In the case of a charcoal-fuelled bowl or shaft furnace, the way in which the plume of products of combustion issues from the nose of a tuyere into the fuel bed depends on the height and the lateral shape of the fuel bed and the placement of tuyeres. WALL EFFECT
An important effect of the inside diameter of a bowl or shaft furnace fuelled with charcoal is the indirect effect that it has on the optimum
17
How Furnaces Work
lump size of the fuel. When lump material lies in the form of a bed against a uniform flat or curved surface, its void fraction (inter-lump volume as a fraction of total bed volume) adjacent to the surface or wall is considerably higher than within the bed, for purely geometrical reasons. Gases thus flow more readily up the wall-bed interface and partly escape reaction with the bed. The phenomenon is known in today's chemical industry as the "wall effect." In a circular container, it increases with the ratio of diameter of lumps to diameter of container, and becomes a serious factor when the average lump size is more than 8 to 10 per cent of the container diameter; a ratio of about 8 per cent has become accepted as the practical desired limit. This means that for best operation of a charcoal fuelled furnace, average fuel lump diameter should be proportionately changed as the inside diameter of the furnace is changed. SPACE VELOCITY
Rate of supply of heat is clearly an essential matter, and in the case of the combustion of charcoal fuel, it will be shown in chapter 6 and appendix I that one cubic metre of air at ambient temperature, burning charcoal by passing vertically through a contained bed that is more than 200 to 250 mm deep, will generate typically about 2. I MJ of heat energy, varying moderately with the carbon content of the charcoal. There is therefore a direct connection between volume of combustion air and the quantity of heat generated. This is not, however, the case when the fuel bed is of biomass, because of the structure of such a bed and the composition of biomass, as will be discussed in chapter 3. A furnace containing a charcoal fuel bed has an average internal cross-sectional area measured in m\ and the rate of air supply is measured as m} per minute. This is a velocity in m per minute in the empty furnace at ambient temperature, and is called "space velocity," or less often, "specific air rate." It is also a heat supply rate in MJ per m 2 per minute, as will be noted in chapter 7. In practice, of course, the actual gas velocity in the interstices of a fuel bed is higher due to the void space between lumps and to the large expansion of the hot products of combustion. However, space velocity has become a widely adopted convention developed in the chemical industry for fluid flow in packed beds, and it has been employed for many years by myself and others as a useful measure of rate of heat energy supply in charcoal and coke-fuelled shaft furnaces, such as cupolas and blast furnaces. In a given furnace, as the space vcIocity (i.e., the rate of heat supply) is increased, the volume of fuel bed at high temperature is increased by
r8
Mastery and Uses of Fire in Antiquity
~ ~
.
U
2S'tJ
I:>
~ ~
,
,.1 CJ¥
8-
(,"
~oo
....tl ~
'" .....
Although the relationship between volume of combustion air and the quantity of heat developed when burning biomass is not a uniform one, certainly a higher rate of supply of combustion air in general produces heat at a greater rate, and so in a kiln of particular construction and heat loss rate a higher air supply rate can generate furnace temperatures that are a higher proportion of the AFT of the fuel. However, as furnace temperature increases, the power of the natural draft then increases, which further increases air flow rate, so there is then a runaway condition. As much as 90 per cent of the AFT of the fuel could be thus developed in a well-proportioned kiln, and with an AFT of, for example, I,600 o c, such as could be possessed by a well-dried hardwood, the furnace temperature could soften or mclt some of the materials of
36
Mastery and Uses of Fire in Antiquity
which a kiln is constructed and possibly cause the kiln roof to collapse - a disaster. There are examples of such over-heated kilns in the archaeologicalliterature, and these are evidence of the high temperatures that can be developed in simple biomass-fuelled kilns. In practice the maximum air rate was often limited by chance or ignorance, such as by too small a firebox or its opening, and/or too short a height from air entry to smoke-hole. DUNG, PEAT, AND STRAW AS FUELS
Dung is biomass that has been processed by a digestive system, and the most common varieties are from herded herbivorous animals such as cattle, sheep, pigs, and llamas. As produced, the chemical composition is similar to that of the feedstock but may contain more or less water, VM, and ash, depending on the species of animal and associated feedstock involved; the fibre structure is less organized than that of the feedstock. Drying is important to increase the heat content and the AFT of dung when it is burned, and on a completely dry basis its heat content varies with the species of animal but not widely, from about 17 to 18 MJ/kg (Winterhalter et al. 1974). This is about 10 per cent less than the heat contents of grasses and low resin content woods. Air-dried dung is an important fuel for warmth and the preparation of food in tropical and sub-tropical countries even today, particularly where fresh or fossil biomass is scarce or expensive. As a fuel for generating high temperature heat, dung dried to similar water content could satisfactorily replace or supplement wood in natural draft kilns and reverberatory furnaces, since it not only can produce a flame but is already in a particulate form that makes a fuel bed of moderate void space. This would decrease excess air and so give higher flame temperature (chapter 4). However, as a metallurgical fuel for bowl or shaft furnaces, dried dung would act like wood and be converted to charcoal in the upper part of the fuel bed, which is then burned as it sinks to air entry level, as described in chapter 7. Peat, which is vegetable matter that has started to decompose out of contact with air and which in time can become lignite and eventually coal, has high moisture content as found. It often has little directional structure and contains VM, and so can be similar in properties, heat content, and combustion characteristics to dung. Archaeological evidence of its use for metallurgy is slim, though it would be a good kiln fuel if thoroughly dried. Straw and chaff are common agricultural residues and burn readily. Their advantage is a usually low moisture content and easy drying; disadvantages are a quite high ash content, heating value about 17 per
37
Biomass Combustion
cent below that of wood, and in particular, small size and very low bulk density. The latter limits the rate of heat generation possible since as combustion air rate is increased, its higher velocity can entrain larger amounts of fuel that can escape in the flue gases before being fully burned. But in practice these fuels can be successfully used, and even today straw is burned in kilns for firing clay building brick.
4 Furnace Configurations for Biomass Fuel
A fuel bed of burning biomass generates heat both within the bed and as a flame outside the bed. Thus to transfer heat to one or more objects, they can be immersed in the fuel bed before its ignition, or contained in an enclosure to be heated by the flame from biomass burned in a separate, but connected, firebox. Each has its advantages, but the latter arrangement can be under much better control and is more efficient in the use of fuel. It became dominant in the Near East apparently by about the sixth millennium B.C. OBJECTS MIXED WITH FUEL
The simplest procedure to heat objects is to mix them with the fuel as a heap on the ground before ignition as a bonfire. Heat is transferred to the objects largely by radiation and conduction, but there are large losses of heat as flame and radiation to surroundings, so fuel consumption is high. Also since access of combustion air is uneven due to variable wind and to bed structure, there is uneven temperature distribution, particularly in hollow objects such as clay pots; this can result in their cracking due to uneven thermal expansion. If the heap is covered with turf and potsherds before fuel ignition, heat loss is considerably decreased, and access of air, and resulting temperature, can be not only more even but controllable by openings in the base of the cover. This arrangement is still in some scattered use and was the progenitor of the separation of the generation and application of heat.
39
Biomass Fuelled Furnaces
WORK SEPARATE FROM FUEL
An important and considerable step in the ability to control temperature, increase its maximum level, and decrease fuel consumption was taken when the combustion of fuel was separated from the work to be heated. Since objects placed in a work chamber by themselves have much lower resistance to the passage of gases than when the spaces between them are filled with fuel, much higher rates of gas and therefore heat flow can be created by natural draft. This considerably increases the maximum furnace temperature attainable. An important further result is the ability to develop and more efficiently use the hot luminous flame from the combustion of the VM in the biomass fuel. There are many ways of joining one or more fireboxes to a work chamber, with three basic objectives. These are obtaining as complete combustion of fuel as possible, providing adequate length of combustion space to fully develop the combustion of the VM flame, and giving good distribution of hot products of combustion in the chamber holding the work. One widely used variation from a simple work chamber with one or more attached fireboxes at the same level was used at least as early as the middle of the six millennium B.C. at Yarim Tepe (Merpert and Munchaev 1973), and is still in use today. This was to place the work chamber over the firebox with a perforated floor between. A smokehole or flue in the top of the work chamber created natural draft, which drew combustion air through openings in the lower wall of the firebox and then hot products of combustion through the perforated floor of the work chamber to heat the work in it. This construction is thermally efficient and can increase the uniformity of distribution of temperature. To this end there was considerable variation through time in the geometry of the supports for the floor of the work chamber, to give necessary support while interfering as little as possible with uniform distribution of air. A commonly used variation was to build a permanent combustion chamber with a flat perforated roof on which ware to be heated was placed, and over which was built a domed cover with vent on top, which was demolished again after firing was completed. Another variation was to place the firebox at a side or end of the work chamber but at a lower level, and to make several horizontal loosely covered channels across the floor of the work chamber, connecting the exit from the firebox to a common flue up the end wall of the kiln. The floor thus became a radiant distributor of the flow of hot products of combustion, as a conductive and radiant heater. In later antiquity there was also considerable use in kilns of walls that had verti-
40
Mastery and Uses of Fire in Antiquity
cal clay pipes set into them to carry the products of combustion, thus separating the gases from the objects to be heated. This was extended to the use of ceramic plates with knobs on them as spacers, fastened to the wall to create a continuous open space for products of combustion (Brodribb 1978). Kilns inherently use large quantities of fuel because of their low thermal efficiency, and variations in geometry of kilns to decrease fuel consumption andlor increase uniformity of product were endlessly explored by trial and error. Illustrations of some of the very wide variety actually used can be seen in, for example, Majidzadeh (1975), Barnard and Tamotsu (1975), and Alizadeb (1985), to mention only three of many collections. Classification of kilns by geometry is difficult because of the great number of small and large variations in the archaeological record, but a system based on how the hot gases from the firebox are distributed to and among the work would seem to directly use a basic factor. Temperatures Attainable The important details of kiln design and operation that lead to increased maximum temperature are as follows: sufficiently high ratio of volumes (or floor areas) of firebox to those of work chamber, and of areas of flue opening and total air entry; choice of kind of biomass for flame length suited to the length of path of combustion in the work chamber; greatest possible dryness of the biomass; and adequate internal height from the floor of the firebox to the outer edge of the smoke outlet or flue opening, which directly controls the draft pressure available. A kiln will reach an equilibrium maximum temperature that depends on these factors, and to the extent that they can be determined on an excavated kiln, we can begin to retrieve its operation. In kilns of moderate size such as 10 to 15m2 work floor area, a firebox floor area of about one third this volume can be sufficient to develop kiln temperatures of more than 1,200°C. In a kiln of 16 m 2 work floor area excavated at Ayia Triada from fourteenth century B.C. in Crete, chemical analysis of droplets of wall refractory (Levi and Laviosa 1986) showed it to be very similar to that of a slag that had its free-flowing temperature measured experimentally as 1,250°C (Gale et al. 1985; also sec chapter 10). Since the temperature of the kiln atmosphere must have been higher than this because of heat loss through the wall, its temperature would have been at least 1,300oC. There is also the evidence of ancient collapsed kilns mentioned on page 36. However, if a kiln is small, of less than about one m 2 floor
41
Biomass Fuelled Furnaces
area, its heat loss rate becomes sufficiently high that the firebox area may need to approach that of the work chamber to achieve high temperature.
Temperature Distribution in the Work Chamber Combustion of the VM and the development of flame have been discussed in chapter 3, but the distribution of temperature in a kiln is also important. When a kiln is filled with pottery and firing is started, the combustion air is at first drawn in slowly because of the low temperature of the interior of the kiln. But it will increase as kiln temperature rises (see appendix 3), and control of air rate, i.e., combustion rate, is usually necessary to restrict the rate of increase of temperature to avoid cracking of some ware by uneven thermal expansion. Then as kiln temperature increases further to where the ware is less likely to crack, the VM flame can be developed further through the work chamber, by increasing the air supply rate or changing to a biomass fuel that gives a longer flame. The hot products of combustion decrease in temperature as they flow because of the heat being absorbed by the ware and the kiln walls, so throughout most of the firing cycle the temperature of the ware nearest the firebox will be well above that of ware near the flue. The latter will reach desired temperature as the ware and the kiln structure as a whole "fill up" with heat and temperature; but the ware ncar the firebox will then have been at temperature for some time and can become over-fired if care is not taken in the rate of temperature lllcrease. This problem of achieving reasonably even temperature distribution in the ware throughout a kiln can never be solved completely, but the variation can be decreased by judicious placement of objects, placing less sensitive or more massive pieces near the firebox, and the erection of permanent partitions within the kiln to deflect and guide the flow of hot gases, leading to modern terminology of "up-draft" and "downdraft" kilns. The use of clay pipes along kiln walls to carry the hot products of combustion to the flue(s) noted above would much improve uniformity of temperature in the kiln. Although every change in the direction of flow of a gas increases its pressure drop and therefore decreases its rate of flow with a given draft pressure, in practice there is often a self-compensating feature. This is that while the need for guidance of the hot gas flow increases with the size of the work chamber, the increase in size usually includes an increase in the height of the roof, so that the vent or flue in the top cre-
42
Mastery and Uses of Fire in Antiquity
ates stronger draft; heat loss rate is also lower as kiln size increases. However, in smaller kilns it can be difficult to adequately control gas flow path without decreasing net draft and therefore air flow rate, and so decreasing maximum temperature.
Smelting of Ores and Melting Metal in a Kiln It is now well established in the archaeological literature that in Neolithic times pottery kilns fuelled with biomass were able to develop temperatures in the order of I,200°C, sufficient to both smelt copper ores and to remelt copper and bronze. It was noted in chapter 3 that the atmosphere in a biomass-fuelled kiln is very oxidizing and cannot directly smelt an oxide ore to its metal, but as will be discussed in chapter 10, if a copper oxide ore, a flux, and a source of carbon are pulverized, mixed, and placed in a crucible, the mixture is then unaffected by the atmosphere in the kiln and when temperatures of 1,100 to I,200 o C are reached, will form molten copper under a layer of molten slag.
Reverberatory or "Air" Furnace A rectangular plan of kiln with the firebox at one end, a shallow hearth in the middle, and the flue at the other end was eventually developed into an effective and simple furnace for smelting and melting metals, and for melting glass. As shown in figure 5, the sidewalls are low in height, the low arched roof becoming very hot on its interior and radiating heat downward onto the hearth - which is why this shape of kiln or furnace is called a reverberatory furnace. The low roof also decreases the cross-sectional area of the furnace interior, which elongates the flame from the firebox to increase longitudinal heat distribution. The length of the hearth depends on the length of flame of the biomass used as fuel. In size, such furnaces ranged from as little as one metre for glass-melting in England in the seventeenth century to six or seven metres in China in the sixteenth century. The hearth of the furnace is formed as a shallow elongated dish in a bed of rammed silica or beach sand, which is retained by the side walls and a low brick wall at each end. The top of the firewall next to the firebox should be only a small distance above the level of the molten contents of the hearth, to permit the flame to sweep the contents as soon as possible. The top of the bridgewall at the flue end is usually a little higher than the firewall since its height determines the cross-sectional open area above it, which should be less than that above the firewall to help retain heat within the furnace.
43
Biomass Fuelled Furnaces
Fue!
"nl
Illy
'ca.I" Ifemovabk>
~/.y p/"9\..L~"==r~""",, .:~~;~;:[(j?:y~~;:
I ,/',0:" , I
C"'f'".~;tr ~o()l:>kJ
Figure 5 Reverberatory or "Air" Furnace
Material to be melted can be added through a temporary opening in the roof or, before lighting the fire, by removing two or three rows of roof bricks and then replacing them. An opening that can be shut tight (a leak will short-circuit the draft) is usually placed in one sidewall just above liquid level, to permit periodic observation of the contents and to manipulate them as necessary with a bar. A small peep-hole near the base of the stack or flue just above bridgewallievel permits viewing the state of combustion and the evenness of its lateral spread. The temperature of the flame decreases as it flows through the furnace, transfering heat to the material on the hearth and to furnace walls, but it must be high enough at the flue end to continue to transfer heat to melted material on the hearth and not cool it. Flue gas temperature is therefore high, and since the strength of draft increases with temperature, a strong draft can be created by a short flue or stack. The rate of heat generation in the firebox is a direct function of the rate of combustion air supply, and this can be controlled by a movable damper over the firebox feed-hole. Another arrangement that gives smoother operation, since it permits regular and frequent additions of fuel, is to leave the firebox entry partly open most of the time, and to regulate the rate of air flow by a damper across an opening made at the base of the stack, as shown in figure 5 (top). This opening short-circuits the draft suction, and so regulates the rate of gas flow through the furnace. The peep-hole nearby permits frequent observation and adjustment. As with kilns in general,
44
Mastery and Uses of Fire in Antiquity
the ratio of firebox area to hearth area is important. The balancing of their proportions and the ratios of the areas of openings have been discussed for modern practices, which are not greatly different, by Rehder (1953). Temperatures of 1,400oC in the metal, which are ample for melting bronze and cast iron, were regularly attained in Europe in the nineteenth century with wood fuel and natural draft. ATMOSPHERE COMPOSITION CONTROL
The effects of the furnace atmosphere can be avoided by enclosing the ware in a separate container of low porosity. This can be as simple as a covered clay pot or "saggar," which can both keep atmosphere out and provide support for delicate pieces that might deform under their own weight at high temperature. A more extensive step is to construct a double-walled or "muffle" furnace, the inner chamber being in effect a very large inverted pot. In both saggars and muffles, the atmosphere within can be made reducing even at high temperature by enclosing a small amount of any form of dry biomass. The kilns mentioned above with closed channels in floors and up walls were muffle furnaces. FUEL ECONOMY
The hardness and strength of fired clay increases, and its porosity decreases, as firing temperature is increased; and clearly temperatures adequately high to make low porosity or waterproof pottery from many clays have been possible from very early times with biomass fuel and natural draft. However, examining structures of pottery from most of the Near East suggests that much of it was not fired above about 900°C. Also, it was a common practice to coat the insides of amphorae for wine with pitch, so there must have been porosity needing sealing. This has supported speculation that high firing temperatures were somehow not possible at the time, but since it is clear from the discussion above that they were attainable, the reason for lower temperatures was more likely to have been the cost and/or availability of fuel. AN EGYPTIAN KILN OF ROMAN DESIGN
Another instructive example comes from Roman Egypt, in a papyrus dated A.D. 243 (Cockle 1979). The treatise concerns the lease of a fully equipped factory for making pottery, with a capacity of fifty thousand jars per year for the wine trade. Operation was through the winter and spring of the year and finished by June. The fuel used was
45
Biomass Fuelled Furnaces
apparently straw and chaff. The jars held nineteen litres each and were lined with pitch, forty-seven kilograms of pitch being required for each thousand jars. Assuming the operation of the factory to have been for about seven months of the year or thirty weeks, and that wastage was IO per cent, I,830 jars per week were fired. If the weight of a jar is estimated as nine kg, and the firing temperature was 700°C, giving a heat content of the clay of 1.04 MJ/kg, then the heat necessary to fire one jar was 9.4 MJ. The potential heat content of straw and chaff is about 13.5 MJI kg, and if it is burned at a thermal efficiency of 2.0 per cent (the kiln is moderately large but combustion of such lightweight fuel would be incomplete), the quantity of fuel required per jar would be thirty-five kg. One week's production of 1,830 jars would require sixty-four tonnes of fuel, which at an estimated bulk density of I50 kg/m3 would occupy a volume of 430 m 3 • For the thirty-week season nearly I3,000 m 3 of fuel would be required; and the logistics of gathering and transporting this volume of fuel during the season would have been considerable.
5 Products Made in Antiquity in Biomass Fuelled Furnaces
The categories of products in antiquity were simple - ceramics, lime, glass, and the smelting and melting of metals - but each of these was divided into many kinds of products. Fired clay, for example, was used for pottery and other containers, building bricks both plain and glazed, roofing tiles, mosaic tesserae, and small statuary. CLAY PRODUCTS
Through antiquity fired clay was a widely used product of high temperature heat. Our discussion of furnaces so far has taken as examples mostly kilns that have been used for firing pottery, a practice that has continued into the present day. The hardness and strength of fired clay pottery increases, and its porosity decreases, as firing temperature is increased. However, as noted in chapter 4, there is considerable evidence that the majority of pottery used in antiquity was not fired to temperatures much over 900°C, possibly as practical compromises between serviceability, the properties of a local clay, and economy of fuel and of kiln time. Yet stoneware requiring temperatures of 1,200 to 1,300oC was made in some quantity. PLASTER OF PARIS
Two materials, plaster and lime, are made from thermally decomposed stones that when pulverized and mixed with water can make a smooth, adherent mass that hardens with time. The material we call Plaster of
47
Products of Biomass Fuelled Furnaces
Paris is made very simply by heating the common mineral gypsum (a hydrated calcium sulphate) to between 128 and 163°e, which drives off half its water of crystallization. If this material is re-hydrated by moistening with water, it will harden into a plaster. However, if in the original heating the temperature of the gypsum has exceeded 163 °e, all of the combined water has been lost and with it the ability of the material to be re-hydrated. If a temperature of 128°e has not been reached, no water of crystallization is lost and no re-hydration is possible. The temperatures involved are those of a food-cooking oven, and must be carefully controlled to within this narrow range. Plaster from gypsum was and still is widely used for making a smooth coating on a wall or ceiling and can easily be moulded into decorative shapes. When hardened, it is finely porous and has moderate hardness but will not withstand elevated temperature. Its porosity when hardened makes it a useful mould material into which a clay slip can be poured, which then forms a leathery skin of a thickness that increases with time, as water is absorbed from the slip by the plaster. When the thickness of skin is what is desired, the remaining slip can be poured out and the leather-hard clay object taken out to be dried and fired. LIME
Lime, which is calcium oxide and is also called "quick-lime," is made by thermally decomposing limestone, which is calcium carbonate of various degrees of purity. When lime is hydrated to calcium hydroxide by the addition of water, it can be made into a smooth plaster that hardens, first by evaporation of water and then by absorption of carbon dioxide from the atmosphere, to re-create calcium carbonate. Since this is the composition of the limestone from which the lime was made, hydrated lime can make a plaster that is much stronger than Plaster of Paris. It can be used as a binder of sand to make a strong mortar. When mixed with crushed pozzolan, a variety of lightweight volcanic silicoaluminate, hydrated lime reacts with it to form a strong concrete that will set under water; this process was extensively used by the Romans. Considerable quantities of lime were used in the Near East from an early date. Excavations at Asikli Huyuk have shown a lime plaster floor in a dwelling dated to 7000 B.e. that contained 1,800 kg of lime. At Jericho a millennium later, many houses had lime plaster floors and walls that would have required the calcining of an average 450 kg of limestone per house (Gourdain and Kingery 1975). A minimum temperature of about 9000e is necessary to decompose calcium carbonate to calcium oxide and carbon dioxide, and the rate of decomposition increases with higher temperature. If during the firing
48
Mastery and Uses of Fire in Antiquity
of the limestone its temperature rises much over about 1,100°C, the lime made increases in density and becomes more difficult to hydrate, so good temperature control is necessary in the kiln. Since the atmosphere in the kiln has negligible effect on the decomposition, biomass fuel is satisfactory. Limestone is preferably broken into small lumps before heating, as decomposition occurs at the rate that heat penetrates from the surface of a lump; small lumps thus decompose more rapidly. These can be heated loose on the floor of a kiln or in a crucible, or can fill a specially shaped kiln, as will be described below. As heat penetrates a lump of limestone, the progress of its decomposition inward from the surface of the lump alters the appearance of the stone. Completeness of reaction can be followed by periodically extracting and examining a broken lump. The decomposition of calcium carbonate requires a considerable amount of heat. It takes 0.96 MJ per kg to increase its temperature to 1,OOOoC, and then 1.75 MJ per kg is absorbed by the decomposition reaction. However, limestone is seldom pure calcium carbonate, averaging possibly 85 per cent. In antiquity an average of probably not more than So per cent of the calcium carbonate was decomposed, some of it remaining in the centres of large lumps. The heat of decomposition would therefore be 1.75 x 0.S5 x o.So = 1.19 MJ per kg of limestone, which added to the heat necessary to raise it to decomposition temperature (its enthalpy) would be 1.19 + 0.96 = 2.15 MJ per kg of limestone. Lime (calcium oxide) is 56 per cent calcium carbonate, so the lime made would be 0.85 x 0.80 x 0.56 = 0.38 kg per kg of limestone, and the heat necessary per kg of lime made would be 2.15Io.3S = 5.7 MJ. This is 42 per cent more than the heat necessary to heat and reduce iron oxide to one kg of iron as a bloom, and more than eight times as much heat as is necessary to fire one kg of pottery to Soooc. Clay and limestone were both fired in biomass-fuelled kilns which have low thermal efficiencies, so lime-making would have been a major consumer of fuel through antiquity. Limestone in small lumps can be decomposed slowly in a windblown wood fire, or in an ordinary pottery kiln heated to 9001,OOOoC, but it can be more efficiently done in a vertical kiln, which is geometrically a shaft furnace. In a rare example of recording details of technical matters in antiquity, Cato in early Roman times (165 B.C.) described how to build such a kiln to make lime: Build the lime-kiln ten feet across, twenty feet from top to bottom, sloping the sides in to a width of three feet at the top. If you burn with only one door, make a pit inside large enough to hold the ashes, so that it will not be necessary to clear
49
Products of Biomass Fuelled Furnaces
them out. Be careful in the construction of the kiln; see that the grate covers the entire bottom of the kiln. If you burn with two doors there will be no need of a pit; when it becomes necessary to take om the ashes, clear through one door while the fire is in the other. Be careful to keep the fire burning constantly, and do not let it die down at night or at any other time. Charge the kiln only with good stone, as white and uniform as possible. In building the kiln, let the throat run straight down. When you have dug deep enough, make a bed for the kiln so as to give it the greatest possible depth and the least exposure to the wind. If you lack a spot for building a kiln of sufficient depth, run up the top with brick, or face the top on the outside with field stone set in mortar. When it is fired, if the Harne comes out at any point but the circular top, stop the orifice with mortar. Keep the wind, and especially the south wind, from reaching the door. The calcining of the stones at the top will show that the whole has calcined; also, the calcined stones at the bottom will settle, and the flame will be less smoky when it comes out.
Such a kiln is a shaft in the sense that it is considerably taller than wide, though this one was distinctly conical in vertical interior profile. The twenty-foot (6.I metre) height would give a strong natural draft, and the narrowing of the shaft towards the top was probably intended to give increased uniformity of heating by decreasing the weight of stone to be heated per vertical metre. Over-burning of the lime near the grate would be unavoidable. The kiln may be estimated to have contained about 36,000 kg of lump limestone, and assuming it to be of the quality estimated above for heat content necessary, the lime made at each firing would be I3,700 kg or I3.7 tannes. Modern practice would be to place holes for air access around the base of such a shaft, and fill it with a mixture of fuel and lumps of limestone in a ratio of about ten to one by weight. After ignition at the base, the rate of combustion of fuel is controlled by shutters over the air-holes. Combustion of fuel allows the mixture to sink, and when fully burned lime appears at an air-hole, a layer of it is extracted with rakes, and fresh fuel and limestone are added at the top. In this way a continuous operation is achieved with less fuel consumption and more evenly burned lime. Since at the time of Cato's writing the Romans had a widespread empire and good communications, and evidently were not aware of the advantages of continuous operation, it seems unlikely to have been used elsewhere in antiquity. GLASS-MAKING
Glass-making appeared in the Near East after copper smelting was well developed and could have been a development from experience
50
Mastery and Uses of Fire in Antiquity
with metallurgical slags. These usually contain considerable silica derived from the gangue of the ore, and this contributes the quality called "glassiness," described below. Slags are usually opaque because of their content of other oxides such as those of iron, manganese, and other metals, but on occasion can be quite transparent, especially in thin section. There seems to be no archaeological record of steps by which clear glass could have been derived from semi-transparent slag, and the first all-glass vessels were not made until the end of the sixteenth or the early fifteenth century B. C. (Oppenheimer et al. 1970). However, faience had been known for a long time, made from clean silica sand mixed with a little soda or potash, which acted as a binder of grains of sand when heated to moderate temperature. If the mass was heated to a higher temperature, a plastic mass of clear or coloured glass would form. The addition of small amounts of oxides of iron, manganese, copper, and cobalt can create a wide range of colours in glass, with tin oxide producing a white opacity. As long as the silica in a completely fused mixture is more than about 45 per cent, the solidified and cooled melt has the rigidity of a crystal with the random atomic structure of a liquid, so that it is amorphous in structure. This may be taken as a definition of the "glassy" state. Silica is the most common glassifier, though there are others such as boron and phosphorus oxides. The naturally occurring obsidian is a glass that is solidified lava of suitable composition and cooling rate. As a corollary to its disordered molecular state, glass has no distinct melting point, and when its temperature is increased, it simply decreases in viscosity from an extremely high value at room temperature to a pourable liquid at sufficiently elevated temperature. Modern convention is that the "working point" of a glass, where it is manipulable somewhat like warm toffee, is at a viscosity of 1,000 pascal-seconds (Pa.s), and the "melting point," where it can be poured out of a container like a heavy oil, is at 10 Pa.s. Viscosity is the basic property in the handling and forming of glass, and the temperature at which a particular viscosity is reached can vary by 500°C or more with change in composition of the glass. Typical practical working points or temperatures would be in the order of 1,000 to I,200°C. The viscosity-temperature curves for glasses can be quite steep, and this creates a major problem in melting and blending the raw materials to make a clear glass. Until high temperatures are reached, internal mixing is very slow, and inhomogenieties not only can show at the surface of the finished glass but can distort light passing through it. Excessive temperature is not wanted since the container of the materials
51
Products of Biomass Fuelled Furnaces
is itself composed of metal oxides and can dissolve and spoil the glass or let glass leak out. The problem was solved in most cases throughout antiquity by melting the raw materials at least twice, the first melt being cooled or quenched in water and then pulverized, thoroughly mixed, and remelted (Oppenheimer loco sit.). A second difficulty in melting glass is that the mixture of granular solids contains about 50 per cent void space, decreasing its thermal conductivity by about half, so that it is slow to absorb heat by conduction from a hot crucible wall. However, the rough surface of a granular mix is a good receptor of radiant heat, so use of shallow saucer-shaped crucibles is more effective. Such saucers were indeed used in ancient Mesopotamia (Oppenheimer loco cit.) but so were more normally shaped crucibles. The kinds of furnaces used for glass-making seem to be remarkably poorly reported both in written literature and in archeological work, attention tending to be on the glass produced. My impression is that the furnace was usually a round or sometimes rectangular biomass-fuelled kiln with a low roof, in which radiation from the VM flame was the important factor. There were various ways of forming the hot plastic mass into useful objects, such as simple manipulation against gravity of a viscous" gob" taken from a crucible on the end of an iron rod; moulding and pressing a gob in a shaped mould; casting into moulds like metal; and blowing with an iron pipe. All were used in antiquity, although the making of blown glass was late. A kiln would be the versatile furnace for all of these methods, including heating crucibles and moulds, but it is puzzling that apparently little use was made of crucibles in beds of charcoal with bellows air supply. Various arrangements of the interior of a kiln were used for glassmaking. Crucibles were of different shapes to increase the rate and uniformity of heating. Apparently a common arrangement was to place crucibles around the interior walls of a circular or rectangular kiln. This was described in a later time period by Theophilus in 1100 A.D. and by Biringuccio in 1540 A.D. It is little different from a pottery kiln design with a low roof. The contents of each crucible had to be accessible through its own hole in the wall of the kiln for the removal of glass for forming when it was in satisfactory condition. Removal was done at intervals to form objects, so the crucible had to be kept up to temperature and the glass at suitable viscosity for considerable periods of time. Another property of glass that made a modified domed pottery kiln an excellent furnace for its manufacture is that unless it is slowly cooled from the temperature at which it can be shaped, thermal stresses build up within the material as its viscosity increases to high levels.
52
Mastery and Uses of Fire in Antiquity
These stresses can be strong enough to shatter the object before it is cool enough to handle, or to make the cooled object fly into pieces from the smallest external stress. The necessary slow cooling is called "annealing," and covers a range of temperature from a level of about 700 to 400°C (depending on the glass composition) to room temperature. This essential annealing can be done in a compartment or portion of the upper part of the kiln where the products of combustion on their way to the flue are too cool to soften glass appreciably but are hot enough to anneal it. Since the remnants of most of the kilns used in antiquity are no longer complete enough to determine their full original structures (particularly their height and therefore available draft and attainable temperature), it may be difficult to distinguish between pottery kilns and glassmaking furnaces, and so it is difficult to follow their independent developments. However, one of the objectives in this book is to point out the technical boundaries within which developments necessarily took place, and not their chronology. Glass is today melted directly on a hearth in a reverberatory furnace as was described above for melting metals, but to my knowledge this was not done in antiquity. Useful data on ancient glasshouse operations are scarce, but records are available from sixteenth-century England when glass was still being made with wood as fuel (Kenyon 1967). These furnaces were apparently operated continuously for a week, making 2,000 kg of glass, which represents a daily output of about 128 litres, the equivalent of a 23 cm cube of glass. This required 19 cords of wood per week, which at 1,600 kg per cord of air-dry wood is 15 kg of wood per kg of glass. If the heat content of molten glass is taken as 1. 5 MJ Ikg, and of air-dry wood as 14.0 MJ/kg, the thermal efficiency of use was 0.7 per cent. The volume of glass made per day indicates a small furnace with inherently high heat loss rate and low thermal efficiency, and so may have relevance to ancient practices. It would certainly result in glass being a relatively expensive material. MELTING AND SMELTING METAL WITH BIOMASS FUEL
Chapter 4 noted that the melting and smelting of metal can readily be done in a crucible in a pottery kiln, but that the elongated, low-roofed kiln called a reverberatory furnace is much better suited when quantities of metal larger than 40 or 50 kg are required. A description of how such a furnace can be built was also given to show its simplicity. When a considerable quantity of metal is necessary such as for the casting of a series of ingots or statuary or ritual vessels, it could be
53
Products of Biomass Fuelled Furnaces
made in several or many crucibles placed on the flat hearth of a kiln for melting. But their removal when filled with molten metal takes time, temperature loss in the metal would then be considerable, and composition of the poured object variable. Melting it all as one mass gives the important advantages of uniform composition and temperature of the whole mass of metal as poured into the mould. So in general, for pouring large objects in bronze or cast iron, the reverberatory furnace is much simpler, safer, of lower cost, and under better control. The furnace would be built specifically for the large casting to be made and placed close enough to the mould and at a suitable level for the molten metal to be transferred by a short trough or "launder" directly to the entrance to the mould. The material to be melted could be scrap metal, or the mixtures of pulverized ore and carbon to make directly reduced metal as described in chapter 10. The building of the furnace would likely be at less cost than the preparation of the mould; there would be no difficulty in delivering molten metal to the mould at nearly its maximum temperature in the furnace; and more than one large casting could be made by preparing a new mould. A furnace with a hearth, for example, one metre wide and three metres long, containing a 200 mm depth of molten bronze, would contain about 5,000 kg that could be poured at one time. It seems very likely that sections of impressive monuments such as the Colossos of Rhodos, built in 280 B.C. and reportedly 32 metres tall (Sarton 1959), would have been cast from such a furnace. I am of the opinion that air furnaces were used in antiquity from an earlier time and more extensively than is generally believed, because of their simplicity, lack of need for blowing equipment, and easy adaptation from kiln practice. However, archeological evidence is sparse, partly because archaeologists have been unaware of the possibilities, and partly because after two or more millennia the oblong ridges of burned earth or clay that could be the remains of such furnaces can very easily be mistaken for those of pottery kilns. Melting in such a furnace has a valuable metallurgical effect on the composition of cast iron in particular. The hot gases above the metal are quite oxidizing because of the presence of oxygen in excess air, and the metal surface is large relative to its volume. Cast iron contains 2 to 4 or more per cent of carbon and as will be noted in chapter 12, becomes higher in strength as its carbon content decreases. Therefore when cast iron from a shaft furnace (which is usually high in carbon content), and pieces of scrap iron to be recovered are remelted together in an air furnace to accumulate a large quantity of molten iron, the strength of the iron can be considerably increased because of the removal of carbon by oxidation. Carried far enough, this process can make steel, and then would be called" fining."
54
Mastery and Uses of Fire in Antiquity
Similarly, in remelting copper or bronze that have too much iron in them from poor smelting practice (see chapter II), the iron is readily oxidized preferentially, to be absorbed as an iron silicate slag by a silica flux. Some of the ancient furnace remains shown in the Chinese archaeological literature and termed "blast furnaces" seem much more likely be the remains of reverberatory furnaces. TEMPERATURES IN BIOMASS FUELLED FURNACES IN ANTIQUITY
I showed above that the combustion of biomass in simple furnaces can easily produce the temperatures necessary for much of the pyrotechnology practised in antiquity, and that not infrequently temperature had to be prevented from rising too high for the purpose intended. It is interesting to note in passing that while the maximum temperature necessary for most of the purposes noted is in the order of I,200 o C, the maximum temperatures that occur in nature (with the exception of lightning) are in volcanoes and are about this temperature. These natural temperatures are not as high as generally thought, many careful measurements having shown that the temperature of molten lava varies from about 800°C to a maximum of I,225°C, the latter being basalts (Williams and McBirney 1979), which can be similar in chemical composition to many smelting slags.
6 The Manufacture and Properties of Charcoal
When biomass is burned with ample access of air, a series of reactions and decompositions takes place that were described in chapter 3. If, however, combustion is stopped before it is complete, as for example by rain or by excluding air by a cover of earth, some charcoal usually remains. Charcoal has very different physical and chemical properties from those of the biomass from which it is made, and for reasons to be given in chapter 7, it was universally used in antiquity for the smelting of ores of metals. It was therefore necessary to be able to manufacture charcoal from biomass with some control of process, and the method developed by an early but unknown date was to burn an enclosed heap of biomass slowly with a very limited supply of air. The heat generated pyrolized (decomposed) some of the unburned biomass to charcoal, and the air supply was limited to burn as little as possible of the charcoal made. MANUFACTURE
When biomass is heated out of contact with air, the products are a solid residue (charcoal) and gaseous volatile matter (VM) in proportions that are sensitive to temperature. As temperature is increased, charcoal starts to form above about 22SoC and the fixed carbon (FC) content of the charcoal increases approximately in proportion to the increase in temperature. For example, when wood is heated to soooc, the charcoal may contain 80 to 8 S per cent FC, and at 900°C, 90 to 9 S per cent FC, the charcoal becoming stronger, harder, and less chemically reactive as
56
Mastery and Uses of Fire in Antiquity
the final temperature of its formation increases. The yield of charcoal as a percentage of the weight of the original air-dry biomass also decreases at higher temperatures due to loss of VM and is moderate even in modern externally heated kilns. Depending on the species of biomass, its water content, and final heating temperature, this would be a maximum of 35 to (at most) 50 per cent. However, throughout antiquity yield was much lower. Charcoal was made in antiquity by the heap method, or its inversion, the pit method. In the heap method, wood is piled vertically into a low cone or mound, with an opening left at the centre as a flue. The mound is covered completely by turf except at the flue, and openings are left in the cover around its base to admit combustion air. The wood is ignited by a brand thrown down the flue, and the rate of combustion is controlled to a low level by adjusting covers over the air entry holes. When the heap is considered on the basis of experience to be completely converted to charcoal, a time of one to two weeks, all air entries are closed and the whole left to cool for several days. If the mound is opened while still too warm, the charcoal will burst into flame because of its high chemical reactivity. The method is simple, and many detailed descriptions and illustrations have been published that show the variety of ways in which it can be conducted (e.g., Overman 1854 and Theophrastus 280 B.C.). Yields are inherently low since part of the biomass is burned to generate the necessary heat, some of the charcoal is unavoidably burned, and not all of the biomass in a heap may be sufficiently well carbonized for use in metallurgy and must be discarded. Yield throughout antiquity would vary depending on skill and raw material, with 10 to 15 per cent possibly being an average. The species of biomass used to make charcoal has a strong effect on both the strength of the charcoal and on its chemical reactivity. For example, a hardwood such as birch will make a lump charcoal of good strength and hardness, while a softwood such as pine under the same conditions will make lump charcoal of lower but still useful strength and hardness, but of higher reactivity. Some species of biomass make charcoal so soft that it is not coherent and appears as a powder. Recorded conversations about the smelting of bloomery iron in Africa (Cline 1937) show that for each practitioner trees of a particular local variety were the only ones that his experience showed to make good charcoal. Harder and denser charcoal is the preferred product of a heap, since it results in charcoal of larger average lump size after the abrasions of digging it out of the heap, transporting it to the site of use, and measuring it into a furnace. Size is an important factor in charcoal's efficiency as a smelting fuel for reasons discussed in chapter 2 and again in chapter 7.
57
Charcoal Manufacture and Properties
Modern experienced manufacturers of charcoal (Overman r854, Constantine 1975) have noted that small-sized wood, less than about 70 mm diameter, makes harder, denser, and less reactive charcoal than will wood split from the trunk of the same tree. The reason lies in the difference in the coarseness of the "ray" structure of the wood, which means that younger wood, and branches of a tree, make better charcoal than does wood from its trunk. This advantage of smaller wood for charcoal making has three implications. One is that it seems likely that throughout antiquity most of the charcoal made was from smaller size and therefore younger wood. This is of some importance to modern c- 14 dating of charcoal. The date measured is that of the cessation of growth of the wood, and often some years are added to the measured date to allow for growth of the wood. However some of these additions have been unreasonably large, by assumption that the charcoal was made from a century-old tree. Addition of ten or twenty years reflecting charcoaling of limbs and coppices would be a more practical average. Another is that large trees would be felled only when the trunk was necessary for some structural use, or more often as clearance for farming. The felling of a large-diameter tree in antiquity was a major undertaking because of the brittleness of early stone or flint axes, and the relatively low hardness of the bronze axes available later. The third implication is that the preferred use of smaller wood meant that when a large tree was felled for making lumber or ship masts, the branches would be preserved and used for charcoal making, resulting in excellent use of the tree as a whole and in that sense decreasing the rate of deforestation. PHYSICAL PROPERTIES
In heap charcoal making, temperature and time at temperature will differ from the bottom to the top of the heap, partly because of the skill applied in control of the combustion, and partly because temperature is highest at the air entries at the base of the heap. There will thus be a range of properties of the charcoal in different parts of the heap, and selection is necessarily made while taking out the finished charcoal. As noted above, the hardness of charcoal is significant because of its effect on final size distribution and on its reactivity. There are clear visual clues to charcoal quality. A sixteenth-century A.D. writer, Biringuccio ([r540] 1942), notes for example, that if charcoal is "rightly baked it is thick and strong, and when struck with another piece is as resonant as glass." A nineteenth-century comment (Overman r854) was that the variation in a freshly opened carbonized pile can be recognized by the relative appearance and hardness of the
58
Mastery and Uses of Fire in Antiquity
charcoal, and that on digging out the pile, the charcoal that rings when struck and breaks with a shiny, conchoidal fracture is the best. These criteria were as easily recognized eight or more millennia ago as today. Judging from several nineteenth-century reports on charcoal making when some knowledge of modern chemistry had become available, it would appear that the desirable minimum FC content of smelting charcoal was about 80 per cent. The negative effects of low Fe content are discussed in appendix 1. The average quality of charcoal made in antiquity was probably very good, because of the direct and reliable connections between the visible appearance and the hardness of charcoal and its effectiveness as a smelting fuel. The operator of a smelting furnace undoubtedly examined the charcoal closely before accepting it for use, and if it was unsatisfactory, it would have relegated to some minor use such as cooking or warming fires. The charcoaller would have had to adjust his practices accordingly. The description by Theophrastus in the third century B.C. of how to make and to recognize good charcoal matches those of the nineteeth century A.D. Each lump of charcoal contains considerable very fine porosity as manufactured, resulting from the cell structure of wood, so it has large internal surface. It can therefore absorb water rapidly, so much so that it is always kept under cover to keep it as dryas possible. Absorption of water increases the weight of a piece of chacoal but does not change its volume since it is in its pores; so a basketful of lump charcoal contains about the same weight of FC wet or dry. Wet charcoal absorbs heat by evaporation of its water in the upper part of a burning fuel bed, which can actually be an advantage since it can decrease consumption of charcoal, as will be explained in chapter 7. Charcoal as a mass or fuel bed has an inter-lump void fraction usually in the range of 0.35 to 0.50 depending on the shapes and the size distribution of the lumps; these in turn depend to a considerable extent on the species and physical size of the wood used. The bulk density is a function of the specific gravity of the charcoal itself, which includes its internal porosity, its water content, and the void fraction of the lumps as a packed bed. This was a fortunate circumstance throughout antiquity since there is no archaeological evidence of use of weight to measure charcoal or ore into a furnace, volume being invariably used. In modern practice, bulk density is customarily taken to he that of dry, or nearly so, charcoal. The bulk density of hardwood charcoal is typically about 280 kg per m 3 and of softwood charcoal about 175 kg per m 3 • If the heat of combustion of charcoal is taken as 28 MJ/kg, then the volumetric heat content of a fuel bed of hardwood charcoal may be taken as 7.84 GJ per
59
Charcoal Manufacture and Properties
m 3, and of softwood charcoal as 4.90 GJ per m 3. It will be noticed that these heat contents per unit volume average about twice those of fuel beds composed of wood or woody materials given in chapter 3, and they are less variable. A higher intensity of combustion can therefore be achieved with a charcoal fuel bed, and higher temperatures can be developed as a greater percentage of a higher AFT of the fuel. The bulk density of charcoal is important also because of a hidden effect on the operation of ancient smelting furnaces. A matter of first importance in burdening or charging a smelting furnace is the ratio by weight of carbon in the charcoal to that of the ore. However, as noted above, both charcoal and ore were measured into furnaces by volume, as some number of baskets or barrows. Variability in the bulk density of charcoal meant that the weight of carbon in a basketful, and so the ratio of charcoal to ore, could vary without the change being visible. This created an undetectable source of uncertainty in the temperature of smelting and is another reason why individual smiths adhered to using charcoal made in a particular kind of heap, from wood from particular trees. This was in order to achieve, without realizing why, some uniformity of bulk density of charcoal to obtain better repeatability of smelting results. The crushing strength of a piece of hardwood charcoal is considerably lower than that of coke made from coal, while that of softwood charcoal is lower than that of hardwood. In a furnace smelting an ore (which has several times the bulk density of charcoal), the average bulk density of a complete charge of ore plus charcoal determines the pressure on the charcoal in the hearth. However, with the high ratios of charcoal to ore used throughout antiquity, the average bulk density of the furnace contents remained well below the strength of the charcoal to support the burden, even in furnaces several metres tall. This point needs emphasizing, since well into the second millennium A.D. blast furnace operators believed that the use of charcoal fuel limited the height that furnaces could be built. This view was mistaken, as experience with large and very tall furnaces has shown. Charcoal continued to be used for smelting both iron and copper in the West until early in the eighteenth century A.D., then was slowly replaced by coke for reasons of cost and availability, not strength. However, in North America, charcoal was low in cost and ample in supply, and by late in the nineteenth century, charcoal-fuelled furnaces twenty-four metres high with hearth diameters of 2.5 metres were still in production (Sweetser 1908). During the same period the ratio of ore to charcoal was increased for economy, which increased the bulk density of the burden to 400 - 500 kg 1m 3 • At no time was there difficulty due to the fuel, unless it was admittedly badly made.
60
Mastery and Uses of Fire in Antiquity
CHEMICAL PROPERTIES
One of the most important properties of charcoal is its chemical reactivity, which is among the highest of all forms of carbon. Usually a compromise was made between reactivity and strength and size of charcoal without realizing it, since of course the concept of reactivity was not known and mixtures of softwood and hardwood charcoal were often intentionally used for convenience or economy. When high chemical reactivity is combined with the high fuel surface area that is due to desirable lump size being relatively small, as was noted in chapter 2, the total reactive power is high. The result is that in fuel beds of charcoal more than about 10 to IS fuel lump diameters thick, all of the oxygen in the combustion air is consumed in passing through. Thus not only is excess air as described for biomass fuel in chapter 3 then not possible but the products of combustion above this level are then a mixture of nitrogen and carbon monoxide that is strongly reducing in character. This makes charcoal by far the most effective fuel for smelting oxide orcs of metals, considerably more so, in fact, than the less reactive coke made from coal. The VM content of oven-dry charcoal is the inverse of its FC content since ash content is usually low and is a measure of the success of its removal during manufacture, being typically in the order of 5 to IS per cent. When charcoal is charged to a furnace and sinks with the burden as it is consumed at a tuyere, it descends through rising temperatures, and most of the contained VM is driven off in the upper part of the furnace. Although VM is a strongly reducing gas, it is removed in a temperature range too low to be very effective in reduction of orcs. The heat content of charcoal kept reasonably dry and burned completely to CO 2 will be in the range of 27.0 to 30.0 MJ/kg (Tillman et al. I98r) and the AFT will be in the range of r,820 to 2,OOOoC depending on the FC content and reactivity of the charcoal (Rehder 1990). However, it must be noted that while the chemical and physical properties of charcoal are the reasons why it is so effective as a fuel for smelting ores, as a heat source alone it is grossly wasteful. On the assumption that in antiquity the yield of charcoal is taken as 15 per cent or 0.15 kg per kg of air-dry wood, the heat content of this wood at 25 per cent moisture can be taken as 15.0 MJ per kg; charcoal used for smelting is burned nearly completely to carbon monoxide (as will be described in chapter 7 and appendix I), with a heat production of 8.3 MJ per kg for charcoal containing 80 per cent FC. The heat produced by the charcoal made from one kg of air-dry wood is therefore 0.15 x 8.3 = 1.24 MJ, which is a loss of 13.8 MJ or 92 per cent of the
61
Charcoal Manufacture and Properties
heat content of the original wood. Some heat is recovered from the carbon monoxide through partial oxidation by ore which decreases the loss somewhat, but this is balanced by the fact that wood suitable for charcoal-making is only about 60 per cent of a tree, and most of a tree can be burned directly as biomass, so the waste from a whole tree is larger. Charcoal has an ash content derived from that of the initial wood and varies with species from about 0.25 to as much as 5.0 per cent. However, common temperate climate species such as pine, oak, beech, and ash have moderate ash contents in the order of 0.3 5 per cent. If the wood used for charcoal-making contains, for example, 87 per cent VM, 12.6 per cent FC, and 0.35 per cent ash on a dry basis, then the ash content of the FC in the resulting charcoal would be 0.35/12.6 = 2.8 per cent. But in the heap practice used in antiquity, due to chemical reactions during carbonization, the recovery of charcoal is variable so the ash content of the charcoal would be on the average about half this or in the order of I. 5 per cent. Actual analyses vary with the time-temperature cycle in charcoaling but are about this size. Because, as has been shown, lump size distribution, bulk density, FC content, and reactivity of charcoal all can have major effects on the operation of smelting furnaces, it is important not only to preserve the lump charcoal found in the archaeology of smelting contexts but, wherever possible, to measure these properties. TRANSPORT OF FUELS
Materials were transported extensively by ship in antiquity (Casson 1959), but in the long list of items given by Casson, wood and charcoal are not included and so were apparently infrequent cargoes. A principal reason may have been the low bulk density, particularly of charcoal, and so a relatively low value per unit of volume: the cost of maintaining and manning a ship must be recovered in terms of the value of the volume of its hold per unit of time. Biomass fuel has airdry bulk density of from about 400 kg per cubic metre for cord-wood, to 200 or less for bundles of brush, and charcoal has typically about 200 kg per cubic metre. Since (as will be shown) the efficiency of use of the energy in the fuel was low, the bulk density of delivered fuel in terms of heat content was in the order of only 10 kg per cubic metre, which made fuel, either as biomass or as charcoal, expensive to transport very far. As late as the eighteenth century in England the maximum economic distance for transport overland of charcoal to a blast furnace was about five miles. Above this, it was cheaper to move the furnace.
62
Mastery and Uses of Fire in Antiquity
COKE
Certain kinds of coal can have most of their VM removed to create a high FC residue called coke using exactly the same methods as for making charcoal from biomass. The earliest evidence of such usc, however, is not until about I,OOO A.D. in China, and several hundred years later in the West, both outside the time frame of this book.
7 Combustion in Beds of Lump Charcoal
The combustion of fuel beds of the various forms of lump carbon low in VM, such as charcoal, coke, and anthracite coal, has been thoroughly studied both in the laboratory and industrially. Basically only two simple chemical reactions are involved, which occur in sequence and at quite different rates. The result is that in the same fuel bed, oxidizing gas at high temperature and reducing gas at lower temperature can be found at different levels. These can be changed as the lump size and air supply rate are changed, giving remarkable versatility in the use of lump fuels, particularly for the smelting, melting, and refining of metals. These fuels differ in their chemical reactivities, which affect their rates of reaction and so of heat and gas distribution; and as noted in chapter 6, charcoal has particularly high reactivity. But since coke and anthracite coal have little evidence of use in antiquity, discussion of their combustion is irrelevant here, except for a note in appendix I concerning problems that can arise in trying to reproduce ancient forging practices using coke as fuel. THE PHYSICS AND CHEMISTRY OF COMBUSTION OF CHARCOAL
The combustion of charcoal (for more quantitative detail, see appendix I) is essentially the combustion of carbon in two stages, the first reaction being between carbon and the oxygen of air. This takes place very rapidly to produce carbon dioxide and a large quantity of heat that
64 Mastery and Uses of Fire in Antiquity increases the gas temperature. On further passage of the resulting very hot mixture of carbon dioxide and nitrogen through the fuel bed, the carbon dioxide is reduced to carbon monoxide in a second reaction. This also consumes charcoal but proceeds much more slowly, and absorbs a moderate quantity of heat that decreases the temperature of the gas. The result, if the fuel bed is deep enough and air is admitted at near its base, is the complete gasification of the original carbon to a mixture of carbon monoxide and the nitrogen of the initial air. There can be no "excess" air due to the lower void space and high reactivity. Since the chemical reactions are sequential in time and space, the chemical nature of the products of combustion in a bed of charcoal fuel change from very hot and oxidizing close to air entry to very reducing and somewhat cooler at some distance from it. With charcoal as fuel, the maximum carbon dioxide content of the gas developed in the initial reaction with charcoal is in practice 10 to 12 per cent, depending on the source of the charcoal - considerably less than the 20.9 per cent theoretically possible. At the air entry rates used typically in antiquity, maximum gas temperature and carbon dioxide content are reached within milliseconds at a distance of two to three fuel lump diameters from the point of air entry. With usual average lump sizes, this is typically about 30 to 60 mm from air entry. However, the complete reduction of this carbon dioxide to carbon monoxide as it passes further through the fuel bed occurs at a much slower reaction rate, and so requires an appreciably further distance. This can be from as little as six to as much as fifteen fuel lump diameters from air entry, again depending on the source of the charcoal. The increased gas temperature in the initial stage of combustion depends on the amount of carbon dioxide formed, and at the carbon dioxide contents of about 10 to 12 per cent typical of charcoal, this creates an AFT of 1,880 to 2,000oC, with r,940oc as a practical average. It is usually possible to find with a thermocouple a spot near a tuyere nose in the bed of an operating charcoal furnace that will show a temperature as high as 1,800°c. This value is moderately lower than the AFT of the fuel due to heat losses, as discussed in chapter I and further in appendix 1. The heat absorbed by the ensuing reduction of the carbon dioxide to carbon monoxide decreases gas temperature to about 1,000°C, below which the reduction of carbon dioxide effectively ceases for thermodynamic reasons. Figure 6 shows experimental results from combustion of a charcoal fuel bed (Hiles and Mott 1944) that reflect the effects described above; they are directly applicable to furnace conditions in antiquity because the space velocity of 13m per minute that was used happened to be in the same order of size as that used in antiquity. The charcoal had an
65
Combustion in Lump Charcoal Beds CO COt. CO/C02
%
40
30
n7TIf. C 2,000
't.o
/,000
/0 COil
o
O~~~--~~-r~--r-~~~ () ~ 4/0
b
Fu..t.! Lt.