Catalysis: Volume 7 [Reprint 2021 ed.] 9783112576564, 9783112576557

Citation preview

Catalysis Science and Technology Volume 7

EDITORS:

(Melbourne/Australien) PROF. D R . M . BOUDART (Stanford/USA)

PROF. DR. J. R. ANDERSON

CONTRIBUTORS: B. E . KOEL, J . V . SANDERS, G . A . SOMORJAI, S. A . TOPHAM

CATALYSIS-

Science and Technology

Volume 7 With 94 Figures

Akademie-Verlag Berlin 1986

Die Originalausgabe erscheint im Springer-Verlag Berlin • Heidelberg New York • Tokyo

Vertrieb ausschließlich für alle Staaten mit Ausnahme der sozialistischen Länder: Springer-Verlag Berlin • Heidelberg New York • Tokyo

Vertrieb für die sozialistischen Länder: Akademie-Verlag Berlin

Erschienen im Akademie-Verlag Berlin, DDR-1086 Berlin, Leipziger Straße 3—4 Alle Rechte vorbehalten © Springer-Verlag Berlin • Heidelberg 1985 Lizenznummer: 200 • 100/549/85 Printed in the German Democratic Republic Gesamtherstellung: VEB Druckerei „Thomas Müntzer", 5820 Bad Langensalza Umschlaggestaltung: Eckhard Steiner LSV 1215 Bestellnummer: 763 573 2 (3071/7) 13200

Editorial

Our series of books on Catalysis: Science and Technology is by now more than half finished. Its purpose has been to collect authoritative and, if possible, definitive chapters on the main areas of contemporary pure and applied catalysis. Its style is not that of an Advances series, nor is it meant to be a collection of up-to-date reviews. If the chapters and the volumes were following each other in a neat, logical order, our series might be considered as trying to emulate the original Handbuch der Katalyse, pioneered by Professor G . - M . Schwab in the 1940's, or be a new version of Catalysis, the series edited by Professor P. H. Emmett in the 1950's. As a matter of expediency, to avoid the delays involved in assembling a complete volume of related chapters, we decided at the outset to publish the chapters as received f r o m our authors. W e submit that, by the time our series is complete, our main objectives will have been met. We are most thankful to all our contributors for their co-operation. The Science and the Technology of Catalysis will prosper as a result of their hard work.

General Preface to Series

In one form or another catalytic science reaches across almost the entire field of reaction chemistry, while catalytic technology is a cornerstone of much of modern chemical industry. The field of catalysis is now so wide a.nd detailed, and its ramifications are so numerous, that the production of a thorough treatment of the entire subject is well beyond the capability of any single author. Nevertheless, the need is obvious for a comprehensive reference work on catalysis which is thoroughly up-to-date, and which covers the subject in depth at both a scientific and at a technological level. In these circumstances, a multi-author approach, despite its wellknown drawbacks, seems to be the only one available. In general terms, the scope of Catalysis: Science and. Technology is limited to topics which are, to some extent at least, relevant to industrial processes. The whole of heterogeneous catalysis falls within its scope, but only biocatalytic process which have significance outside of biology are included. Ancillary subjects such as surface science, materials properties, and other fields of catalysis are given adequate treatment, but not to the extent of obscuring the central theme. Catalysis: Science and Technology thus has a rather different emphasis from normal review publications in the field of catalysis: here we concentrate more on important established material, although at the same time providing a systematic presentation of relevant data. The opportunity is also taken, where possible, to relate specific details of a particular topic in catalysis to established principles in chemistry, physics, and engineering, and to place some of the more important features into a historical perspective.

VIII

General Preface to Series

Because the field of catalysis is one where current activity is enormous and because various topics in catalysis reach a degree of maturity at different points in time, it is not expedient to impose a preconceived ordered structure upon Catalysis: Science and Technology with each volume devoted to a particular subject area. Instead, each topic is dealt with when it is most appropriate to do so. It will be sufficient if the entire subject has been properly covered by the time the last volume in the series appears. Nevertheless, the Editors will try to organize the subject matter so as to minimize unnecessary duplication between chapters, and to impose a reasonable uniformity of style and approach. Ultimately, these aspects of the presentation of this work must remain the responsibility of the Editors, rather than of individual authors. The Editors would like to take this opportunity to give their sincere thanks to all the authors whose labors make this reference work possible. However, we all stand in debt to the numerous scientists and engineers whose efforts have built the discipline of catalysts into what it is today: we can do no more than dedicate these volumes to them.

Preface

The development of a commercially successful process for the catalytic synthesis of ammonia was a scientific as well as a technical triumph. Its implications were considerable. It demonstrated the power of a combination of innovative technology and engineering together with basic chemical science, and it introduced ideas and techniques into catalytic science and process engineering which are still with us today. In a real sense, this process changed the face of industrial chemistry and process technology. Of course, the key step in the direct synthesis of ammonia was the development of an efficient catalyst, and the historical account given by Dr. S. A. Topham in the first chapter of this volume shows how this was successfully accomplished, and how this was combined with the successful solution of other daunting technical problems to make the overall process possible. The microstructure of a catalyst is an important feature which determines its behaviour, and the electron microscope is one of the most important instrumental methods by means of which structural and microstructural information can be obtained. Nevertheless, the electron-optical processes of image formation are complex, but need to be properly understood if image interpretation is to be done reliably. In the second chapter of this volume, Dr. J. V. Sanders addresses the entire field of the application of electron microscopic methods to the examination of catalysts. This provides both practical examples of what can be done, together with a discussion of the theory which enables the various image-forming techniques to be understood, and the images themselves reliably interpreted. Heterogeneous catalysis involves the breaking and

X making of chemical bonds in species adsorbed at the catalyst surface. Although many practical catalysts are chemically and structurally complex, it has always been extremely valuable to use as a reference, structural data obtained from the study of well-characterized, singlecrystal surfaces. This is the topic dealt with in chapter 3 by Dr. B. E. Koel and Professor G. A. Somorjai. Their account provides many stimulating insights and reference points by means of which the behaviour of practical catalysts may be illuminated.

Preface

Contents

Chapter 1 The History of the Catalytic Synthesis of Ammonia (S. A. Topham) Chapter 2 The Electron Microscopy of Catalysts (J. V. Sanders) Chapter 3 Surface Structural

1 51

Chemistry

(.B. E. Koel and G. A. Somorjai)

. . . .

159

Subject Index

219

Author Index Volumes 1-6

223

List of Contributors

Dr. B. E. Koel Cooperative Inst, for Research in Environmental Sciences and Dept. of Chemistry University of Colorado 80309, USA Dr. J. V. Sanders CSIRO Div. of Materials Science Catalysis and Surface Science Lab. University of Melbourne Parkville, Victoria 3052, Australia Professor G. A. Somorjai Materials and Molecular Research Division Lawrence Berkeley Laboratory Dept. of Chemistry University of California, Berkeley Berkeley, CA 94720, USA Dr. Susan A. Topham I.C.I. Agricultural Division Billingham Cleveland TS23 ILE, U.K.

The History of the Catalytic Synthesis of Ammonia Susan A. Topham I.C.I. Agricultural Division Billingham Cleveland TS23 1LE U.K.

Contents 1. Driving Forces for the Fixation of Atmospheric Nitrogen

. . .

2

2. Haber's Work on the Synthesis of A m m o n i a A. Initial Involvement of Fritz H a b e r B. Early Attempts at the Synthesis of A m m o n i a C. Investigations on the Position of the A m m o n i a Equilibrium 1. Work of H a b e r and Van O o r d t at Atmospheric Pressure 2. W o r k of Nernst at Elevated Pressure 3. W o r k of H a b e r and Le Rossignol at Atmospheric Pressure 4. W o r k of H a b e r and Le Rossignol at Elevated Pressure D. Attempts at a Technical Exploitation of A m m o n i a Synthei.s

4 4 7 9 9 11 12 13 14

3. Bosch's Work on the Fixation of Nitrogen A. Education of Bosch B. Early Career of Bosch at B A S F C. Visit to Karlsruhe D. Further Work Carried Out by H a b e r E. Initial Progress at Ludwigshafen 1. The Problems Ahead a) The search for the catalyst b) Development of the high pressure a p p a r a t u s c) The supply of pure gases F. Decision to go Full-Scale: F u r t h e r Problems to be Overcome G. The First A m m o n i a Plant at O p p a u H. The Refined Plant at O p p a u

16 16 18 22 23 24 24 26 31 34 35 37 38

4. Influence of the First World War on the Haber Process

42

5. Development Outside G e r m a n y

44

6. Developments in A m m o n i a Technology

46

7. Epilogue

49

8. References

49

2

Chapter 1: Susan A. Topham

1. Driving Forces for the Fixation of Atmospheric Nitrogen In September 1898, in his presidential address to the British Association for the Advancement of Science, Sir William Crookes [1] warned that "England and all civilised nations stand in deadly peril of not having enough to eat.'"

He had arrived at this disturbing conclusion after a critical examination of the area of arable land available for wheat growing, and the rate of increase of the wheat-eating population, which showed that "our wheat producing it"

soil is totally

unequal to the strain put

upon

Having brought attention to the possibility of future famine, however, he then pointed the way to a possible solution to the problem, using the following prophetic words: "It is the chemist who must come to the rescue of the threatened communities. It is through the laboratory that starvation may ultimately be turned into plenty. Before we are in the grip of actual dearth the chemist will step in and postpone the day of famine to so distant a period that we and our sons and grandsons may legitimately live without undue solicitude for the future.''''

What evidence was there for Crookes' optimism? It was recognised, at that time, that certain materials improved the yields of agricultural crops. J 3hn Bennet Lawes had carried out extensive experimentation on his estate at Rothamstead, during the mid nineteenth century, investigating the requirements of various crops, and, with Sir Henry Gilbert, had demonstrated that the application of nitrogenous fertiliser to wheat crops greatly improved the yield per acre of soil. Thus, wheat supplies could be ensured, provided a reliable source of nitrogenous fertiliser was available. At that time the principal sources of such 'fixed' nitrogen compounds were threefold. Ammonium sulphate was produced during the distillation of coal, but amounts were not great. Guano deposits, off the coast of South America, had also been worked, but were coming close to exhaustion. The major source was situated in the rainless districts of Chile, where large nitrate beds had assumed vast commercial importance. It was considered unlikely, at that time, that any similar such nitrate deposits would be discovered elsewhere in the world, because, owing to its great solubility, sodium nitrate can only accumulate in areas of little rainfall. Contemporary estimates of the reserves of Chilean nitrate indicated that they would reach exhaustion 50 years from the date of Crookes' address if the usage was maintained at the prevailing level, but would run out much sooner if the fertiliser application rate were to increase commensurately with the rate of increase of population growth. Thus, while the application of nitrogenous fertiliser to the soil could provide a temporary respite to the problem of food supply, there appeared to be no escape from the ultimate fate the exhaustion of nitrate reserves would bring. There was, however, one ray of hope amidst

The History of the Catalytic Synthesis o f A m m o n i a

3

all this gloom: although the natural reserves of fixed nitrogen were finite, the reserves of free atmospheric nitrogen are vast. If a method could be found of fixing this atmospheric nitrogen into a form more useful as fertiliser, the problem of food supply would be solved. This led Crookes to declare that "r/ze fixation of atmospheric nitrogen is one the great discoveries awaiting the ingenuity of chemists." Up until 1900, numerous attempts had been made to fix atmospheric nitrogen. Some of these attempts had been successful to a degree, but none had converted more than a small amount, and at a cost largely in excess of the corresponding market value of fixed nitrogen. In 1892, Crookes himself had demonstrated the fixation of atmospheric nitrogen in a lecture entitled "The Flame of Burning Nitrogen", in which, by the passage of a strong induction current between terminals "the air takes fire and continues to burn with a powerful flame, producing nitrous and nitric acids." The first atmospheric nitrogen fixation process to be successfully developed commercially, the Birkeland-Eyde process, was based on this observation by Crookes. The direct combination of nitrogen and oxygen of the air was effected by means of an electric arc; the very high temperatures required to form nitric oxide (greater than 3300 K) could be attained only by its use, and thus vast quantities of energy were needed. A yield of only 2 per cent nitric oxide in the effluent gas was achievable and the large excess of air which remained uncombined had also to be heated up, thus only 3-4 per cent of the electrical energy consumed was actually used to unite nitrogen and oxygen. Consequently, the process was only viable in places where electric power was cheap, the first fullscale plant being set up in Norway in 1905. The Badische Anilin and Soda Fabrik of Germany were also interested in the electric arc method and had been studying the reaction since 1897. In 1904 they published a patent which claimed to produce a steadier arc than that of the the Norwegians, and in 1906 a joint Norwegian/German venture was set up using BASF's furnace. BASF withdrew five years later, being more interested in another method of nitrogen fixation which will be described later. An alternative process of nitrogen fixation had been based on the observation by Frank and Caro in 1898 that a mixture of barium cyanide and barium cyanamide could be obtained when barium carbide, was heated, at high temperature, in a nitrogen atmosphere. It was later discovered that calcium cyanamide could similarly be obtained by heating calcium carbide with nitrogen at 1500 K, and the cyanamide compound thus formed could be used directly as a fertiliser. Ammonia could also be obtained from the cyanamide by hydrolysis. One advantage of this process over the arc process was that, once the reaction of calcium carbide with nitrogen had begun, the electric power could be disconnected, the exothermic reaction being selfsustaining. The initial production of calcium carbide, by heating calcium oxide with carbon, did, however, require temperatures of 2300 K. The overall power requirement was only a quarter of that of the arc process, and considerable development of the cyanamide process occurred in Europe, successful plant operation being achieved by 1910.

4

Chapter 1 : Susan A. T o p h a m

Although both the arc and cyanamide processes did assume some commercial importance, they were eventually to be displaced by another method of nitrogen fixation, which did not suffer the disadvantage of requiring extremely high temperatures. This new process achieved commercial viability in 1913, and was destined to become one of the most significant developments in industrial chemistry. This process was the catalytic synthesis of ammonia from its elements, pioneered in Germany by Fritz Haber and his research group at Karlsruhe and developed to industrial scale by Carl Bosch at the Badische Anilin and Soda Fabrik in Ludwigshafen. In this process, the combination of nitrogen with hydrogen was effected at temperatures of around 820 K, by means of an iron catalyst, and since the reaction was exothermic, once started it did not require further heat input. U p until 1912, general opinion seemed to agree that the future expansion of the fixed nitrogen industry lay with the further development of the arc and cyanamide processes, the 'Haber process' merely warranting occasional mention in the literature of the time. In fact, an extensive programme of work had been carried out by BASF in Germany, with the ultimate aim of perfecting the Haber process, but the details of this work, for obvious commercial reasons, had been kept largely secret. Events of the next few years were to have an important effect on the further development of the process. Although the commercial production of fertiliser was undoubtedly the primary aim behind all efforts to fix atmospheric nitrogen, there arose, in addition, another driving force. Nitrates were essential for the manufacture of conventional explosives of military value, and an adequate supply was essential to any type of military programme. At the outbreak of the First World War, the major source of nitrates was still in Chile, and it became obvious that any belligerent country that was effectively cut off from these supplies would be at a grave disadvantage. During the war, the German authorities supported the development of a nitrogen fixation programme, with the aim of ensuring independence from Chilean nitrates. Both the cyanamide and Haber processes received government support, and further development of the Ostwald process, which effected the oxidation of ammonia to nitric oxide, which could subsequently be converted to nitric acid, was also encouraged. It was reported at the time that it was highly likely that the war would have ended in 1916, but for the existence of the Haber process.

2. Haber's Work on the Synthesis of Ammonia A. Initial Involvement of Fritz Haber [2, 3] Fritz Haber first became interested in the fixation of atmospheric nitrogen when visiting the United States in 1902 on a fact-finding mission on behalf of the Bunsen Society to observe the technical and educational progress of that country. He visited the establishment of the Atmospheric Products Company, at Niagara Falls, at which the arc process was being operated.

T h e History of the Catalytic Synthesis of A m m o n i a

5

Plate 1. Fritz Haber. (Reproduced with permission f r o m "Geschichte der A m m o n i a k s y n t h e s e " , A. Mittasch, (1951), Verlag Chemie, facing p 65)

He reported that, in spite of the importance of the undertaking, the degree of development was disappointing, the installation being more of an experimental apparatus than an industrial operation. On his return, Haber carried out his own investigations, supported by BASF, on the formation of nitric oxide and was granted German and British patents in 1907, for "the production of compounds of nitrogen and oxygen by subjecting such air or mixture to the action of an electric arc". Although this work showed some promise in commercial application it was to be far outweighed in in importance by his later accomplishments. The beginnings of his involvement in the attempts to synthesise ammonia from hydrogen and nitrogen occurred in 1903, when the Margulies Brothers of the Austrian Chemical Works in Vienna asked him to assist them in their work. They had been engaged in attempting to synthesise ammonia by the action of hydrogen on

6

Chapter 1 : Susan A. Topham

the nitrides of calcium, lithium and magnesium, but had only obtained trivial amounts up to that point. Haber was then 35 years old and was a professor at the Technische Hochschule in Karlsruhe. He had achieved certain recognition up to that time, as evidenced by his selection as representative of the Bunsen Society to visit the U.S.A., but he was essentially still an ambitious man who was anxious for success. His earlier career had. been a curious mixture of failure and disappointment. He was born into a prosperous family with a chemical tradition, his father, Siegfried, being owner of an established pigments and dyestufifs merchandising company in Breslau, which was one of Germany's largest importers of natural indigo. His mother, Paula, did not survive the birth of her son, who was therefore brought up by a series of relatives until his father's second marriage. His childhood was, nevertheless, a happy one, and he was encouraged in his school studies becoming well-acquainted with literature, classics, drama and philosophy, although by tradition little science was taught at that time. When almost eighteen he began his studies at the University of Berlin, where he developed an interest in chemistry which was fuelled by various factors. Ability in chemistry would be of obvious assistance in his father's business. Also, at that time Germany was the centre of world expertise in chemistry, and to be a German chemist was to be a member of a prestigious profession. Haber studied at Berlin under Professors Helmholz and Hoffmann, the former codiscoverer of the first law of thermodynamics, the latter discoverer of aniline and founder of the coal-tar industry. Following the tradition of the time, after a year at Berlin Haber moved on to the University of Heidelberg, where he studied under Professor Bunsen. After Heidelberg, Haber returned home to Breslau in order to serve his compulsory year of military training, which passed enjoyably, especially the academic lectures on philosophy, and in 1889 he enrolled at the Charlottenburg Technische Hochschule, the largest engineering college in Germany. It was here that he carried out his first research work, under the guidance of Karl Liebermann, one of the cosynthesisers of alazarin dye. In 1891 he received his doctorate for work related to piperonal, but did not distinguish himself in his final exam. Having spent some time with organic synthesis, Haber's next goal was to study with Wilhelm Ostwald at Leipzig, but his hopes were dashed when Ostwald turned him down. Consequently he sought work in the chemical industry but had to go to Budapest to find it. This position, in a distillery, lasted a year but was totally without distinction as were subsequent appointments in a fertiliser factory and a textile company, which involved purely clerical work. In 1892 he returned to the study of chemistry, this time in Switzerland at the Federal Polytechnic School in Zurich, where he studied physical science and technology under Professor Lunge, and then, aged 24, he returned to work for his father, whilst waiting for a suitable employment opportunity in the chemical industry. The association between father and son, however, was not successful, frequent arguments and disagreements being common, and a parting of the ways was inevitable. This took place in the summer of 1892. The two had disagreed

T h e H i s t o r y of t h e C a t a l y t i c S y n t h e s i s of A m m o n i a

7

about the importing of a chemical to treat a cholera epidemic. Siegfried had eventually accepted his son's judgement, but later, when it proved disastrous, he ordered his son out, declaring, " G o to a university. You don't belong in business!" [2], Haber had reached a low spot in his career. His academic achievements had not been great, his industrial record still worse and his association with his father a disaster. A further unsatisfactory period at the University of Jena as assistant in Professor Knorr's laboratory followed before he finally achieved a satisfactory position at the Karlsruhe Technische Hochschule, as assistant in the Department of Chemical and Fuel Technology, under Professors Bunte and Engler. Here, under the guidance of Bunte, Haber carried out some work of substance. His first two years were taken up by a study of the pyrolysis of hydrocarbons, searching for a general rule for such decompositions; this work was one of the first investigations of the cracking process now used in the oil industry. He published his results in a volume entitled "Experimental Studies on the Decomposition and Combustion of Hydrocarbons" in 1896, the book was received well and he was accepted onto the faculty at Karlsruhe. Here, under the influence and direction of Hans Luggin, who had studied with Arrhenius, Haber developed a keen interest in physical chemistry, a discipline so far unknown to him. Investigations in the field of electrochemistry culminated in the publication, in 1898, of his book, "Outline of Technical Electrochemistry on a Theoretical Basis", which was unusual at the time in that it brought together theory and practice as a unified whole. Although not received with great enthusiasm by the German Electrochemical Society, the book was considered of great importance by other prominent physical chemists of the day and was a major reason for Haber's promotion to associate professor at Karlsruhe. Over the next few years, Haber devoted much time to the study of electrochemistry and gained prominence in the German Electrochemical Society, so that when approached by the Margulies brothers in 1902 he was already of some fame among physical chemists. The following year he published his famous book, "The Thermodynamics of Technical Gas Reactions", which was to have a profound influence on physical chemistry. At Karlsruhe, Haber had acquired a reputation as an adept raconteur, with a particular liking for the classics. He had a prodigious capacity for work, often continuing late on into the night, completely immersed in a particular problem to the exclusion of all else. Above all, he was renowned for his ability as a critic, never failing to bring attention to errors made by others. B. Early Attempts at the Synthesis of Ammonia Haber agreed to help the Margulies Brothers, and in involving himself with the preparation of ammonia, had undertaken a task which had occupied many investigators for well over a century. Ammonia was first prepared in 1774, by Priestley [4], and since the first comprehension of its composition

8

Chapter 1 : Susan A. T o p h a m

by Berthollet [5] in 1784, attempts to synthesise it from its elements had been numerous. In 1795, Hildebrandt [6] carried out the first attempts at a synthetic preparation, carrying out systematic and extensive experiments, but achieved no success. In 1823, Dobereiner [7] was the first to claim a successful catalytic combination of nitrogen and hydrogen. During numerous experiments on the combustion of gases, using platinum as catalyst, he reported that during the combination of hydrogen with the oxygen in the air "sufficient of the oxygen was not available to satisfy all the hydrogen, so the excess reacts with nitrogen to form ammonia." His conclusion was obviously incorrect, but as the experimental method was not explained, the source of his error cannot be ascertained. In 1839, whilst undertaking experiments on the catalytic synthesis of ammonia, Kuhlmann [8] proposed the 'status nascendi' theory. He first tried to combine free hydrogen and nitrogen by means of platinum sponge or platinum black, but did not succeed. Further experiments were made with 'nascent' nitrogen, by passing nitric oxide and hydrogen over platinum sponge, and this time he claimed success. He made the following generalisation: "When hydrogen or hydrocarbons come into combination, under favourable conditions, with 'nascent' nitrogen, ammonia is formed". He also believed that, in the process of rusting, nitrogen of the air was bound with nascent hydrogen to form ammonia. There were notable critics of Kuhlmann's theory, but no less an authority than Liebig [9] supported Kuhlmann's views. In 1844, he reported "We know of no case where combination of elemental hydrogen and nitrogen could be accomplished. The free state is a hindrance to the combination, however, once the elements are 'fixed' the reaction may go in this direction. What combines by this reduction with hydrogen is not normal nitrogen gas, but nitrogen gas in the 'nascent' state". Kuhlmann's ideas prevailed for many years, and exerted considerable influence on both academic and technical work. Numerous attempts were made to synthesise ammonia from hydrogen and nitrogen, by first changing the experimental conditions in such a way as to render one or other of the reacting species 'nascent'. In particular, easily-accessible means were sought by which hydrogen could be obtained in the nascent state and subsequently made to react using cheap, practicable reactions. No success was obtained, however, and the problem of the catalytic synthesis of ammonia from its elements began to appear hopeless. Although the synthesis reaction had not been successfully achieved, it had long been known that the decomposition reaction could readily be effected, and it was the study of the decomposition of ammonia which gave clues to the reasons for the failure of attempts at synthesis. St-Claire Deville [10] showed that when ammonia was exposed to an electric spark in a closed tube, there always remained a small quantity undecomposed. In 1884, Ramsey and Young [11] studied the decomposition of ammonia in a hot porcelain tube. Again, the most significant result of this investigation was that, even at very high temperatures (1050 K), the decomposition was not complete. The possibility of the existence of an equilibrium position had been demonstrated.

9

T h e H i s t o r y of t h e C a t a l y t i c Synthesis of A m m o n i a

The question of an equilibrium position was taken up by various investigators, some, notably, Ostwald [12], accepting its existence, others, such as Perman and Atkinson [13], claiming that their experimental results gave no evidence for such an equilibrium. At the turn of the century, however, the concept of chemical equilibrium, and the effects upon it of temperature and pressure, became clearer, due to the work of scientists such as Le Chatelier. The French chemist turned his attention to the position of the equilibrium between hydrogen, nitrogen and ammonia, at elevated pressure, where he had reasoned that the equilibrium concentration of ammonia should be more favourable. Unfortunately, this exercise terminated in a violent explosion, which killed one of his assistants; the project was abandoned, leaving only an obscure French patent [14], filed in the name of one of Le Chatelier's coworkers, to show for this effort.

C. Investigations on the Position of the Ammonia Equilibrium 1. Work of Haber and Van Oordt at Atmospheric

Pressure

[15]

The aim of the work carried out by Haber, in collaboration with the Austrians, was to achieve the continuous synthesis-of ammonia by the simultaneous formation and reduction of a metallic nitride by the mixed gases, the metal phase acting as a catalyst. Haber decided first to determine the position of the equilibrium between hydrogen, nitrogen and ammonia as this would obviously govern the maximum yield of ammonia attainable. The first paper published by Haber on the formation of ammonia from its elements appeared in 1905, in collaboration with Gabriel van Oordt. In a series of what Haber later referred to as "fairly rough experiments", the position of the ammonia equilibrium in the region of 1300 K was determined. At a temperature of 1293 K, using pure iron as catalyst, the equilibrium position could easily be arrived at, from both the ammonia side and the hydrogen/nitrogen side. The results of the single experiments varied between 0.0052% and 0.012% ammonia, and originally Haber regarded the upper figure as the more likely value. The same results could be obtained with nickel, calcium and manganese as catalysts. From the determination of the position of equilibrium at a single temperature, pressure and mixture of hydrogen and nitrogen, it was possible to predict, on the basis of theory, the approximate percentages of ammonia to be expected for any set of conditions. At that time, thermodynamics was becoming more and more useful to researchers concerned with equilibria. Van't Hoff, Gibbs and Helmholz had demonstrated the relation between heat of reaction, temperature and equilibrium constant, the application of which Haber was to elucidate in his famous book "Thermodynamics of Technical Gas Reactions" which appeared that same year. Van't H o f f s generalisation of the Clausius-Clapeyron equation allowed the calculation of the effect of temperature change on equilibrium from thermal data, and

10

Chapter 1 : Susan A. T o p h a m

permitted the determination of the equilibrium constant at any set of conditions, provided it was already known for one set of conditions i.e. d M

_

AH_

d T ~ RT2 where K is the equilibrium constant T is the absolute temperature R is the gas constant AH is the heat of reaction Haber used this equation to extrapolate the data obtained at 1300 K down to lower temperatures. Table 1 shows the results of these calculations taking the total pressure of the three gases as one atmosphere and the proportions of the hydrogen and nitrogen as stoichiometric; the table is based on the assumption that the upper value of 0.012% ammonia at 1293 K was the more likely figure. Haber concluded, from these figures, that the slight variations observed in temperature over the series of experiments were trivial compared with the error in the determination of the ammonia concentration, since the equilibrium varied only slowly with temperature. The more important conclusion, however, was his opinion that "from the beginning of red heat onwards, no catalyst can produce more than a trace of ammonia from the most favourable gas mixture if the experiment is conducted at normal pressure. Even at highly elevated pressures, the position of the equilibrium always remains very unfavourable. If one wishes to make ammonia successfully with a catalyst at normal pressure, one must not allow the temperature to go above 520 K, The uncertainty of the value used for the specific heat of ammonia, obtained by extrapolation, is, for the time being unavoidable, but it is inferred that the conclusions drawn will not essentially change, if a more precise figure for this quantity is ascertained in the future." It now became clear why many earlier investigators, who had generally worked at normal pressures and elevated temperatures, had failed in their attempts to synthesise ammonia. At this point, in 1905, Haber regarded the further pursuit of the subject of the catalytic synthesis of ammonia from hydrogen and nitrogen as hopeless. Table 1. Equilibrium data predicted by Haber and van Oordt. Based on their experimental results obtained at 1293 K and atmospheric pressure Composition of equilibrium mixture temperature/K

vol%H2

vol % N 2

vol % N H 3

300 600 900 1200 1293

1.12 68.46 74.84 75 75

0.37 22.82 24.95 25 25

98.51 8.72 0.21 0.024 0.012

The History of the Catalytic Synthesis of Ammonia

11

In the work with van Oordt, the preparation of ammonia from its elements had been demonstrated successfully, and the limiations to a synthetic route had been identified. At that time, Haber regarded the discovery of a catalyst which could bring the gases to equilibrium quickly at 570 K as unlikely, and also the commercial exploitation of a gas reaction at red heat under pressure as impossible, so the question of a technical-scale synthesis of ammonia was left dormant by Haber for the next three years. 2. Work of Nernst at Elevated Pressure [16]

Meanwhile, in 1906, Haber was obliged to consider a redetermination of the ammonia equilibrium to be necessary, because Professor Walther Nernst, of the University of Berlin, in the progress of his investigations on his 'Heat Theorem', had arrived at an approximation formula which could predict the position of equilibrium from the values of the heat of reaction and the so-called 'chemical constants'. As stated previously, Van't H o f f s generalisation of the Clausius Clapeyron equation allowed the calculation of the effect of temperature change on equilibrium from thermal data, but failed to give values of the equilibrium constant unless one value was already known experimentally for one set of conditions. Thus the integration constant had to be known before values of the equilibrium constant could be generated. This obstacle was overcome when Nernst arrived at the Third Law of Thermodynamics, which was then known as the Nernst Heat Theorem. With this concept, it was possible to calculate the chemical equilibrium for a given system from a few physical constants. Nernst used the new theorem to predict the positions of equilibria in various reactions, and, in many cases, found good agreement between his predictions and experimental results. This was not so in the case of the ammonia equilibrium. According to the Heat Theorem, Haber's results were incorrect. Nernst and his students set about establishing their own experimental data, conducting experiments at elevated pressures (50-75 atmospheres), and so becoming the first experimenters to synthesise ammonia at pressure. They obtained data which agreed with Nernst's predictions, but not with Haber's figures. Table 2. Nernst's data for the ammonia equilibrium at normal pressure. Extrapolated from the experimental results obtained at elevated pressure temperature/K

% N H 3 at equilibrium (observed)

% N H 3 at equilibrium (predicted by heat theorem)

958 1082 1109 1149 1193 1273 1313

0.0178 0.0087 0.0072 0.0055 0.0043 0,0032 0.0026

0.0196 0.0082 0.00702 0.00561 0.00448 0.00308 0.00261

(c./: Haber and Van O o r d t ' s figure of 0.0120% at 1293 K)

Chapter 1 : Susan A. T o p h a m

12

A comparison of the two sets of data is shown in Table 2. Nernst communicated the results of his experiments to Haber, by letter in 1906. The paper containing these results [16] was not due to appear until the meeting of the Bunsen Society in the summer of the following year. Meanwhile, Haber, in collaboration with Robert Le Rossignol, an extremely able experimentalist from England, who had studied in the laboratory of Sir William Ramsey, repeated the earlier determinations at normal pressure "paying more attention to detail than before". 3. Work of Haber and Le Rossignol at Atmospheric Pressure [17] The new experiments of Haber and Le Rossignol were carried out using the earlier method of Haber and van Oordt, but employing a greater number of catalysts, various different reactor materials, a refined method for the determination of ammonia, and an improved apparatus; in particular, the measurements covered a temperature range between 970 and 1270 K. Much better reproducibility was observed in the later experiments. Nernst's equilibrium ammonia figures had been about a quarter of those of Haber and van Oordt. The later experiments of Haber and Le Rossignol showed that the equilibrium values lay at the lower limit of the band originally quoted by Haber and van Oordt, and not at the upper limit as they had originally thought. The higher figures were later ascribed to a particular property of the catalyst in the fresh state. Haber and Le Rossignol's newer figures, obtained at atmospheric pressure, were then as shown in Table 3, together with the calculated figures of Nernst using physical data available at the time. The discrepancy between the Berlin and Karlsruhe results was now somewhat smaller, Haber's figures being now only 3/2 times those of Nernst. Although Haber had been obliged to modify his earlier views on the exact position of the ammonia equilibrium, he was now convinced that the experimental results obtained in collaboration with Le Rossignol more accurately reflected the true equilibria positions than did those of Nernst. He considered that, since Nernst had not studied the decomposition of ammonia in addition .to the synthesis, it was questionable whether the equilibrium position had been achieved. Haber did, however, concede that the calculations carried out by Nernst would predict what Haber considered Table 3. Haber and Le Rossignol's experimental data for the a m m o n i a equilibrium at n o r m a l pressure temperature/K

% N H 3 at equilibrium

973 1023 1123 1203 1273

0.0221 0.0152 0.0091 0.0065 0.0048

U-.f. Nernst's figure of 0.0032% at 1273 K )

The History of the Catalytic Synthesis o f A m m o n i a

13

to be the correct figures if the figure used for the specific heat of ammonia at high temperature was modified slightly. Since this parameter was not known accurately, it was not an unreasonable statement to make. The disagreement between the two scientists was aired in public at the meeting of the Bunsen Society in the summer of 1907. Nernst gave a paper [16] outlining his experimental results on the position of the ammonia equilibrium and their theoretical treatment according to his Heat Theorem. This was a more detailed version of the note he had sent to Haber the previous year. If Nernst's paper was not well accepted, his Heat Theorem might lose credibility. Haber attended the meeting armed with his experimental data, realising that his reputation was at stake. At the end of Nernst's presentation there was a lively exchange between the two scientists, Haber asserting that Nernst could not be sure he had arrived at the equilibrium position in the absence of decomposition studies, and Nernst claiming that Haber should carry out his experiments at higher pressure where the amounts of ammonia attainable were more amenable to precise measurement. The views of Nernst prevailed, and Haber took this as a personal affront. On returning to Karlsruhe he immediately set up further work on the ammonia equilibrium, this time using a pressure of 30 bar, again in collaboration with Le Rossignol. Haber was fortunate also to have an extremely able mechanic called Kirchenbauer who did the bulk of the construction work on the apparatus. 4. Work of Haber and Le Rossignol at Elevated Pressure [18] The results of the high pressure work were published in 1908. The essential conclusion was that the results obtained at high pressure fully confirmed those obtained by the same authors at normal pressure, and that Nernst's criticism was not valid. Thus Nernst's figures were still lower than those of Haber. In addition, Nernst's data fell more quickly with ascending temperature than did those of Haber. At 1273 K Haber's figures were half as much again as those of Nernst, at 973 K only a quarter as great. Haber ascribed these differences to a number of possible causes: the position of equilibrium was not reached by Nernst, his method of measurement of temperature was subject to inaccuracy and there was a possibility of absorbing ammonia in parts of his apparatus. In conclusion to this work, Haber stated that the determination of the equilibrium at high pressure and at 973, 1074 and 1174 K yielded results which confirmed the results obtained at atmospheric pressure, and which still departed considerably from the reports of Nernst on the subject. Later, when more reliable figures became available for the specific heat of ammonia at elevated temperatures, Nernst's calculations were seen to correspond with Haber's experimental results. Thus, at the end of 1907, work by Haber and Nernst had achieved great significance, and had shown why the earlier experimenters had not succeeded in their attempts to make ammonia from the elements. Because of the thermodynamic limitations of equilibrium, at normal pressures only small amounts of ammonia were possible, except at temperatures of 570 K and under, where known catalysts failed to work. Even at elevated pressures,

14

Chapter 1 : Susan A. T o p h a m

only small amounts of ammonia could be obtained at the temperatures at which catalysts were known to work ( > 8 1 0 K). It was clear, therefore, that the transition to the most feasible high pressure would be advantageous, if the idea of a technical exploitation of the reaction was to become at all practical. D. Attempts at a Technical Exploitation of Ammonia Synthesis [19] Three years previously, Haber had considered a technical exploitation of ammonia synthesis as impossible, since in the absence of catalysts which would work at low temperatures, the reaction would have to be carried out at elevated pressures and red heat. After having undertaken the higher pressure work with Le Rossignol, Haber's views on the subject changed. He now was of the opinion that, even though only a small part of the hydrogen/nitrogen mixture could be converted to ammonia, a technical process might still be feasible, provided a) even higher pressures than those so far employed could be used, b) the high pressure gas was circulated over the catalyst, and c) the ammonia so synthesised was separated out at high pressure before returning the gas to the catalyst. In this way, the total quantity of ammonia that could be produced per unit time, (later called the 'space-time-yield'), could be technically profitable, particularly if new catalytic materials could be discovered which were significantly more active than those so far known. Haber's work was now directed towards the development of these ideas. At the same time, his association with the Margulies Brothers ceased. Instead, on the advice of Professor Karl Engler in Karlsruhe, who was an adviser to the board of BASF, he approached the firm for support for his work on the synthesis of ammonia from the elements. On 15 February 1908, Haber visited BASF in Ludwigshafen. The firm had already supported his studies on the oxidation of nitrogen, and Haber hoped for further financial assistance with his work on the catalytic synthesis of ammonia. The head of the patents department, Professor August Bernthsen, expressed some doubts as to the feasibility of Haber's ideas, but the general director, Heinrich von Brunck, was enthusiastic and an agreement was drawn up between Haber and BASF. Haber made the results of his work available to the works management of BASF, and the personnel in the Nitrogen Section were entrusted with their further development at Ludwigshafen. It must be remembered that it required the strongest belief by Haber that his ideas would come to fruition, since at that time only quite inefficient catalysts were known. It is interesting to note that Professor Nernst had also approached industry with the idea of exploiting his high pressure equilibrium studies further with a view to a technical synthesis. The director of the chemical firm GriesheimElektron, Dr. Bernhard Lepsius, who was an acknowledged specialist in inorganic chemistry, informed Nernst that the practical difficulties involved

The History of the Catalytic Synthesis of A m m o n i a

15

in working at high pressures and temperatures were too great. Nernst abandoned any hope of finding a technical use for his results. Haber set about the quest for more efficient catalysts. Iron had been the original catalyst used, and later it was found that manganese was slightly more active than iron, and cerium more active still, allowing the use of lower temperatures. Meanwhile, on the assumption that a better catalyst would be found, he developed the idea of the circulation of the high pressure gas mix-

C R S P

Compressor Reactor Separator Circulation pump

Figure 1. Circulation scheme for a m m o n i a synthesis p r o p o s e d by H a b e r

ture, and the removal from it of the synthesised emmonia without flash evaporation of the high pressure gases. In addition, in order to minimise not only the work of compression, but also the heat requirement, he proposed that the hot exit gas from the catalyst chamber should be used to heat the incoming cold gas mixture; he was familiar with this technique from the commercial apparatus developed by the firm Linde for the liquefaction of gases. The scheme was completed by continually replacing the gas used up to form ammonia by fresh, unreacted gas. The plan was the basis of a patent application by BASF [20], dated October 12th 1908, in which was claimed "a process for the synthetic manufacture of ammonia from the elements, whereby an appropriate mixture of nitrogen and hydrogen was subjected continually to the synthesis of ammonia by means of heated catalysts, and afterwards the ammonia was removed by cooling or absorption, under high pressure, and the heat of the ammonia-containing gas was transferred to the ammonia-free gas mix". A schematic diagram of the set up is shown in Figure 1. At this time, the ideas expressed in the patent were purely theoretical, and BASF were not convinced of the practical feasibility of the proposed process, as it was believed that no steel vessel could contain a continual pressure of 100-200 atmospheres at red heat. Undaunted by the doubters, Haber and his students worked on, seeking a method of reducing the temperature required to 670-770 K. The problem to be overcome was that, as one reduced the temperature, the rate of reaction became too small. Nevertheless, after a short time, an investigation of the 'higher homologues' of the known catalysts in groups six, seven and eight of the periodic table gave the hoped-for breakthrough.

16

Chapter 1 : Susan A. T o p h a m

In March 1909 it was discovered that the precious metal osmium was a surprisingly good catalyst. Haber had obtained some rare metals (e.g. Pt, Os, W, U) from the German Gaslight Society for whom he was a consultant. The sample of osmium was a powder, which was used for osmium lamps, and which was prepared in a secret process which yielded a very fine, amorphous product; this sample was much more active than an earlier sample of osmium prepared by reduction with hydrogen at high temperature. Using this new catalyst, high reaction rates could be obtained at temperatures around 820 K, and so at 175 atmospheres pressure, exit concentrations of ammonia as high as 8 % could be obtained. Haber wrote to BASF immediately, informing them of his discovery, but was not pleased with the answer he received. BASF were still sceptical, doubting the practicalities of working at high pressures and temperatures of about 870 K. Engler intervened on Haber's behalf, writing a personal letter to von Brunck, and this resulted in a visit to Karlsruhe by von Brunck and two colleagues, during which heated argument ensued concerning the feasibility of working at high pressure. Von Brunck informed Haber that, on his return to Ludwigshafen, he would seek the advice of the chemist Carl Bosch on the issue, as Bosch had had experience of the iron and steel industry. Bosch's answer was "I believe it can work. I know the capability of the steel industry well enough. It should be risked!". Von Brunck laid much weight on this view. Haber's work was supported energetically, and Haber and Le Rossignol concentrated on the development of a small laboratory apparatus in which it would be possible to circulate gases at high pressure Thus, at this point in early 1909, Haber had succeeded in interesting one of the largest chemical companies in Germany in his ideas, and was about to discover whether or not these ideas might be exploited on a practical basis. The BASF chemist Bosch had obviously had a large influence on the decision to continue supporting Haber, and was destined to play an even greater role in the further development of the synthesis of ammonia. He was then 35 years old, six years younger than Haber, and had worked for BASF for almost ten years, gaining a reputation for both thoroughness and innovation. Since he was to become an important figure in the events of the next few years, an investigation of his background, achievements and capabilities is warranted.

3. Bosch's Work on the Fixation of Nitrogen A. Education of Bosch [21] Carl Bosch was born on August 27th 1874, in Cologne. His family owned a wholesale business dealing in piping materials for the supply of gas and water, and his uncle, Robert Bosch, was to become the founder of the famous Bosch magneto factory in Stuttgart. As a child, Carl showed a keen interest in natural history, and later, as a teenager, he began carrying out chemical experiments at home. At school he showed promise in the sciences, and

I he History of the C a t a l y t i c Synthesis of A m m o n i a

17

Plate 2. Carl Bosch. (Reproduced with permission f r o m " I m Banne der Chemie, Carl Bosch, Leben und W e r k " , K. H o l d e r m a n n , (1953), Econ-Verlag, facing p 16)

decided that he wanted to become a chemist but, before taking up his studies seriously, he spent a year at an iron and steel works, the Marien Works at Kotzenan, where he became familiar with mechanical apparatus. This was on the advice of his father, who considered that knowledge of a more practical application of chemistry would be of benefit to his son. After spending a successful time at the foundry, he began his studies, in 1894, at the Technische Hochschule at Charlottenburg, as a student of foundry technology and machines, but also spent a great deal of time attending lectures in chemistry, from such teachers as Professors Knorr, Rudorf, Liebermann, Witt and Vogel. He learnt much inorganic and organic chemistry and also some analysis, and spent three terms in the inorganic laboratory; all these studies in chemistry were accomplished in addition to his set

18

Chapter 1 : Susan A. Topham

work on foundry technology. During one of the long vacations, he obtained further practical experience in the Krupp foundry at Neuwied. Bosch was critical of the teaching of engineering at Charlottenburg, considering it to be far too empirical in nature, and expressed the desire to study pure science (although he later appreciated the engineering experience he had gained.) In the summer of 1896, he began further studies at Leipzig University, working towards a doctorate in the field of organic chemistry under the supervision of Professor Wislencius. Having received his doctorate on 24 May 1898, he became assistant to the professor, a position he held until the following Easter. At the time Bosch was at Leipzig, German universities were important centres of scientific discovery; many new concepts were being developed within the spheres of chemistry and physics, by such names as Arrhenius and Nernst, and at Leipzig Ostwald was at the forefront of the teaching of the relatively new discipline of physical chemistry. One •important aspect of Ostwald's work was the first systematic series of experiments in the field of catalysis. As at Charlottenburg, Bosch again took a particular interest in pure chemistry, but also found time to attend many other lectures on subjects such as mineralogy and microscopy, in addition to his compulsory lectures. He also became known for his extraordinary dexterity in the construction of laboratory apparatus, astonishing his colleagues with a hand-made vacuum pump which he had put together in three days so that he might be able to record the spectrum of helium. On finishing his studies at Leipzig, Bosch wished to pursue an academic cariier, but his father sought to dissuade him. A visit by Bosch senior to his son in Leipzig resulted in Carl deciding to seek employment in the chemical industry. He made a job application to the Badische Anilin and Soda Fabrik in Ludwigshafen, a company well-known for its notable achievements in the area of dyestuff chemistry and for its willingness to embrace new technology. BASF was at that time enjoying the benefits of a great success, as two years previously the synthesis of the dyestuff indigo had been perfected at Ludwigshafen. In addition, the BASF chemists Rudolf Knietsch was engaged in the technical development of the sulphuric acid contact process, which was destined to become highly-successful. Bosch hoped that a position at such an establishment would offer scope for utilising his scientific education to the full. He took up his appointment as chemist at BASF on April 15th 1899.

B. Early Career of Bosch at BASF [21] The conditions existing within German industry at that time were very paternalistic, with little freedom, a twelve hour day and few holidays. The normal induction procedure for new chemists, on entering the company, involved a period in the main laboratory, where the individual's aptitudes were explored. Bosch began with the preparation of an intermediate for azo dyestuffs: only a few weeks were necessary to establish his practical proficiency. This preparation marked his first encounter with catalysis. The fol-

T h e H i s t o r y of t h e C a t a l y t i c S y n t h e s i s of A m m o n i a

19

lowing year, Knietsch gave him an exercise which was to have a profound meaning in later life. Professor Ostwald, in Leipzig, believed he had found a catalytic method for the synthesis of ammonia from its elements, a process which, as has already been mentioned, had occupied many chemists for decades. Ostwald had communicated these results to BASF, hoping for financial support to continue his work. BASF showed considerable interest, and gave Bosch the task of verifying Ostwald's findings, which suggested that up to 6 % ammonia could be obtained at normal pressure, irrespective of whether ammonia was synthesised from the elements or decomposed into its constituents. After numerous experiments, however, Bosch was obliged to inform the BASF directorate that he could not repeat Ostwald's results. This failure exasperated Ostwald, who admonished the BASF board for setting an inexperienced chemist to the task. Bosch was, however, convinced that he was correct, and defended himself strongly, despite being in a rather difficult situation, his competence as a chemist being challenged by a famous professor of physical chemistry. Bosch had observed that, on using the iron wire in the manner described by Ostwald, an appreciable amount of ammonia was produced, but that this initial rate of production ceased after a short time. After spending some time consulting the literature, familiarising himself with the properties of iron and its compounds, Bosch was able to explain the differences between his and Ostwald's results. On heating with ammonia, iron forms a nitride which, by treatment at elevated temperature with hydrogen, forms ammonia. Since Ostwald had used the same sample of iron wire for the decomposition and synthesis steps, the source of the discrepancy could be explained. In the decomposition step, iron nitride had been formed which then decomposed, yielding ammonia, in the synthesis step. The fact that Ostwald had also obtained a higher level of ammonia than Bosch in the decomposition step could be explained by the fact that the gases had been passed over the the catalyst too quickly so that the true equilibrium position had not been achieved. Ostwald was greatly disappointed when the true source of the ammonia had been demonstrated, but by this exercise, Bosch had proved himself to be a thorough experimenter. Over the next few years, Bosch was to gain a reputation as an innovative chemist with both a practical approach and a considerable knowledge of fundamental science; he was often seen at the bench, or in the workshop,' carrying out the required practical work himself, which went somewhat against the normal working practices then existing in the great firm. His superiors, Knietsch and von Brunck, realised ¡that the best way to utilise such a strong, singleminded character was to give him as much freedom as possible. Soon, he was given his own laboratory, a workshop, a technician, a chemist and space in an adjacent building in which to carry out largerscale experiments. He also obtained accommodation in one of the factory houses, an old water tower that had been converted for the purpose of housing young chemists and engineers; he thus settled into working in the chemical industry.

20

Chapter I : Susan A. T o p h a m

Since first being involved in the problem of nitrogen fixation, via the work of Ostwald, Bosch had been considering ways in which he might attempt to solve it. Von Brunck was convinced it was an important question, and felt that its solution was potentially of greater value to BASF than had been the successful synthesis of indigo, so he initiated a considerable effort on the arc process. Two members of the BASF staff, Schonherr and Hessberger, had developed an electric arc furnace at about the same time as the Norwegians. Bosch soon joined this work, with the task of improving the conversion of the gas from the electric arc to nitrates, and was allocated a dextrous glassblower, Kranz, as an assistant. As has already been noted, the fixation of nitrogen by this route was very expensive, since much of the energy was lost, and, unlike Norway, Germany did not possess cheap water power. Thus, for the process to be viable, the combination of nitrogen with another species had to be effected at a lower temperature, so a method was sought which did not require an electric arc. BASF was apparently well-placed to attempt the solution of this problem. Through their association with the sulphuric acid work, Knietsch and Dr. Eugen Sapper, Bosch's immediate superiors, possessed considerable experience in dealing with catalysts and large quantities of gas, and also with the building of new apparatus. Bosch had already shown himself a capable and thorough experimenter. A start was made on the work in the spring of 1902. Bosch began by following up all the possible lines of research to be found in the literature, and became interested in the cyanamide process, wherein nitrogen reacted with a glowing mixture of barium carbonate and carbon to yield a mixture of barium cyanide and cyanamide, which was easily converted to ammonia. Bosch considered that this reaction was worth further investigation, and requested a physical chemist to assist him in his work. Professor Bodenstein, in Leipzig, recommended Dr. Alwin Mittasch, a school teacher who, in addition to his school duties, had studied chemistry under Ostwald and Bodenstein, and thus had received an extensive grounding in physical chemistry and catalysis. Mittasch, who stood only five feet tall, joined BASF in March 1904, spent the customary probation period in the main laboratory successfully, and was assigned to the investigation of the interaction of nitrogen with barium carbonate and carbon. Bosch was a demanding supervisor, requiring that everything that could possibly be measured was measured; this was extremely valuable to Mittasch, an inexperienced newcomer to industrial research, and during this period he became familiar with the use of various instruments for measuring flow rates, temperatures and the composition of gases. The nitrogen used in this reaction had to be extremely pure. Bosch considered various methods for the preparation of pure nitrogen, including the removal of the oxygen from air by means of glowing copper, and the removal, by the use of potash solutions, of carbon dioxide from gas streams after effecting the combustion of various gases in air. In addition to the problems associated with the preparation of pure gases, new apparatus, capable of withstanding extremely high temperatures, had to be devised and constructed. Despite the immense amount of work carried out, the difficulties

The History of the Catalytic Synthesis of A m m o n i a

21

Plate 3. Alwin Mittasch. P h o t o : Deutsches Museum Miinchen

remained apparently insurmountable, and by 1907 still no satisfactory method had been devised whereby nitrogen could be 'fixed' at temperatures below 1770 K. Many individual successes had been achieved, and various patents were generated from this work, but the fact remained that the fixation of nitrogen still relied on the use of extremely high temperatures. In the summer of 1907, a large plant was built for the preparation of barium cyanide. From this plant, 5000 kg of cyanide mass could be produced in a day, and from this, 300-350 kg of ammonia was obtainable, rather than the 400-500 kg predicted from the laboratory-scale experiments. The cost of the plant was unexpectedly high, and many difficulties were encountered due to the failure of ceramic materials, despite extensive investigation of such materials beforehand in the laboratory. After the production of about 90,000 kg of cyanide mass had been effected, at enormous cost, the decision was made to shut the plant down in June of the following

22

Chapter 1 : Susan A. Topham

year. Bosch had not been working exclusively on this route for nitrogen fixation. He and Mittasch had been investigating other nitriding reactions, using titanium, silicon and aluminium, and were not relying solely upon the success of the barium cyanide venture. Gradually, however, it began to appear that the experiments with the nitrides of various elements had been taken to such a point that the aim of fixing nitrogen without the aid of an electric arc was unattainable. It was at this time that Haber approached BASF with the results of his work on the fixation of nitrogen by means of its reaction with hydrogen over a catalyst. The firm had already sponsored some investigations by Haber on the efficiency of the arc process, but it was the catalytic synthesis of ammonia which was now to prove decisive in BASF's attempts to solve the nitrogen problem. C. Visit to Karlsruhe [21, 22|

During 1908 and 1909, Haber and Le Rossignol had worked on, developing a small laboratory apparatus in which it was possible to circulate gases at high pressure, and in July 1909, representatives from BASF were invited to inspect it. On July 2nd, Bosch visited Karlsruhe, accompanied by Mittasch and Kranz, the technician, and since some members of the BASF management still did not believe Haber's results, this demonstration was of vital importance to Haber and Le Rossignol. Unfortunately, however, things did not proceed smoothly. The apparatus sprang a leak, and the demonstration had to be interrupted for a few hours whilst repairs were made. Bosch had watched the proceedings so far. calmly and patiently, but now decided he could not

Plate 4. Haber and Le Rossignol's recirculatory system for the preparation of ammonia. (Reproduced with permission from F. Haber, Z. Elek., 16, 244, (1910), Verlag Chemie) 1 Catalytic vessel with heat exchanger, 2 Condenser, 3 High pressure circulation pump, 4 Fresh gas feed, 5 Ammonia outlet, 6 Sample line

The History of the Catalytic Synthesis o f A m m o n i a

23

wait until the repair had been carried out, as he had pressing business in Ludwigshafen. He outlined to Kranz means whereby repeated failure of the particular seal could be avoided, and deemed it unnecessary to wait to see how the apparatus worked, as he could imagine it well enough. To the great disappointment of Haber and Le Rossignol, he and Kranz returned to BASF. Only Mittasch remained, and he became the first member of the firm to witness the appearance of ammonia in the receiver of the small apparatus. Haber's small apparatus consisted of an upright furnace of about 750 mm in height, provided with a heat exchanger and an electrically-powered upper preheat section. Beneath the heating coil there was a shelved insert of sieve platelets on which osmium powder was sprinkled, and underneath the catalyst space was the heat exchanger, through which the incoming cold gas was preheated. The temperature at the upper end of the catalytic section was about 1170 K, and at the lower end 870 K. The high pressure furnace formed part of a recycle apparatus in which a circulation pump continually fed compressed gas, and a separation tube, maintained at 233 K, condensed the ammonia. Under a pressure of 185 atmospheres, this apparatus yielded 90 grams of ammonia per hour at a concentration in the effluent gas of 2.8 volume percent, using a charge of catalyst of 98 grams of osmium. On March 18th 1910, at a meeting of the Scientific Society at Karlsruhe, Haber gave a lecture, illustrated by experiment, to demonstrate the new technique for the fixation of nitrogen. The accompanying photograph of the original recirculatory system (Plate 4) is taken from the account of the lecture given in the literature [23]. The lecture excited great interest, accounts even being given in the popular press, and Haber achieved considerable fame. In 1909, a thorough account [24] of the work carried out by Haber and Le Rossignol on this system was submitted, in manuscript form, to BASF, for their approval to publish. It contained a detailed account of the theoretical and practical aspects of the synthesis and separation of ammonia. Agreement to publish was granted, in principle, after the removal of some comerciallyvaluable information, but the actual publication of the paper was suppressed until 1913. In the paper, Haber expressed the hope that "the way forward will be opened for a new industry". D. Further Work Carried Out by Haber

Between 1910 and 1912, Haber and his students carried out further thermodynamic and kinetic studies related to ammonia synthesis. Redeterminations of the precise position of the equilibrium at various temperatures and pressures were carried out, as were measurements of the heat of formation of ammonia as a function of temperature. The activities of several catalysts (uranium, osmium, molybdenum and tungsten) were determined, and efforts were made to identify the factors influencing the rate of reaction; seven papers were later published (1914-15) [25-31] summarising all of this work. The appointment of Haber to the directorship of the Kaiser Wilhelm Institute in Berlin, in 1912, signalled the end of his involvement in work on ammonia synthesis.

24

Chapter 1 : Susan A. T o p h a m

It is interesting that Haber made a settlement with BASF which allowed him a flat sum of one pfennig for every kilogram of ammonia that might be sold [2], He had considered that this arrangement was likely to be more favourable than a simple percentage of sales, as the price of ammonia was likely to fall as production rates increased. Le Rossignol and Kirchenbauer became co-beneficiaries of Haber's arrangement. E. Initial Progress at Ludwigshafen [21) 1. The Problems Ahead Only a few weeks after von Brunck's visit to Karlsruhe, Haber had successfully demonstrated that ammonia could be produced from his small experimental apparatus. Von Brunck now put a great deal of effort into developing Haber's ideas further at BASF so that a large-scale manufacturing process at Ludwigshafen might become a reality. Characteristically, he moved rapidly, and on a large scale. Bosch was given extensive powers to progress the project. He was the ideal man for the job, being a chemist with experience of iron and steel technology, and great enthusiasm for the construction of apparatus. He was broad in approach, possessed an extensive fundamental knowledge of science, and in addition he had acquired valuable experience in dealing with large quantities of gas during eight years' acquaintance with the nitrogen problem. Bosch faced three problems, all of which had to be solved before the construction of a factory could be contemplated. First of all, it was clear that osmium was not suitable as a basis for a technical synthesis, because of its high cost, scarcity and tendency to oxidise to osmium tetroxide. The whole world's supply, obtainable as an impurity in crude platinum, only amounted to about 100 kilograms. Also, Haber's later catalyst, uranium, a discovery made after the successful demonstration of the recirculatory apparatus, was rare and sensitive to moisture. Thus, it was necessary to seek new catalysts which were cheaper and easier to handle. The second problem facing Bosch concerned the construction of apparatus. The technology of the time could furnish no examples of vessels which were required to contain high pressure gases at high temperature. Most chemical reactions of technical importance occurred at normal pressures, and the one example of a process operating at high pressure, the Linde process for the liquefaction of the air, required the containment of 200 atmospheres at a maximum temperature not exceeding that of ambient air. The copper apparatus used here, with its soft soldering, was not suitable for temperatures up to 870 K, so the design and construction of apparatus was obviously an exercise of vital importance to the success of the enterprise. Thirdly, it was necessary to supply the raw materials, nitrogen and hydrogen, in sufficient quantity and purity for the full-scale unit. The contemporary processes for the production of hydrogen were all to prove too expensive and to yield gas too impure for use in the catalytic reaction, so the provision of a suitable process for hydrogen production was also of great importance. Work aimed at the solution of all three problems was set underway simultaneously in the summer of 1909, and as the enormity of the project

T h e H i s t o r y of t h e C a t a l y t i c S y n t h e s i s of A m m o n i a

25

became apparent, many experts considered that Bosch had undertaken a task that was impossible to achieve. Even among his colleagues, scepticism was rife, opinions being expressed that the enterprise would eventually be the ruination of the firm. Bosch was not concerned by these views, and immediately set about the task of setting up a workshop and the gathering together of a team of coworkers suitable for the work ahead. He required workers who could work precisely and reliably, and who were capable of tackling new and unusual concepts quickly and without prejudice, but as he sought them out he met resistance. The normal working practices of the firm were wellstructured, with separate departments being responsible for individual areas of work, but Bosch preferred to supervise all work personally, testing out and proving all processes himself. This alienated some of the established engineers, who declined to join him in the venture, so he was obliged to recruit younger engineers, and Franz Lappe, who had studied with Professor Bach at Stuttgart, specialising in metal work, joined the project. Lappe typified the commitment and optimism which Bosch was seeking. Bosch put great reliance on independent thought, reliability and positive attitude. He considered a fundamental chemical education to be the prerequisite for success in all areas to be investigated, setting less value on specialism, which he believed to lead to a narrow approach to problem solving. Chemists were assigned preferentially to engineering exercises. His own experiences at Charlottenburg had led him to develop a distrust of the education of engineers, which he believed provided them with insufficient chemical grounding. However, an engineer who had a fundamental command of physics or thermodynamics was just as welcome to the team as a chemist. By these criteria, Bosch surrounded himself with capable coworkers, and embarked on the daunting tasks with a small, inexpensive workshop consisting of one large and one small lathe, a drilling machine, a turner, a fitter and an assistant. Showing a characteristic distaste for beauracracy, he combed the entire factory for ideas which might be useful in the work. Again, this was unusual at BASF, which was not organised to accomodate such independence, but Bosch had the backing of the directorate, which had, by then, allocated a million marks for the ammonia synthesis work, and awaited a positive result to justify such expense. Bosch made Mittasch responsible for the catalytic work, and took charge of the problems related to the high pressure apparatus and hydrogen supply himself. The two men were opposite in character and approach. Bosch saw things on a large scale, was impulsive and generated a continual stream of ideas. He worked rapidly, and made few notes, preferring to rely mainly on his memory. He was quick to assess the possible practical application of an idea, and made decisions equally speedily. In contrast, Mittasch favoured the systematic, logical approach, making reasoned deductions from careful observations, and keeping extensive, accurate notes. Despite these differences in approach, both men shared a total commitment to precise experimental work. Like Haber, Bosch frequently worked late into the night, and had no time for those who did not share his dedication to the tasks in hand.

26

Chapter 1 : Susan A. Topham

a) The Search for the catalyst [21, 22, 32] The first exercise was to repeat Haber's results. The board had already taken the precaution of obtaining options on the world's supply of osmium, and it was now necessary to obtain pressure vessels and ancillary equipment for the construction of catalyst testing equipment. A skilled fitter was despatched to Karlsruhe to become familiar with the construction of Haber's apparatus, and in the summer of 1909 a high pressure rig, based upon Haber's heated catalytic vessel, was made ready in the phthalic acid laboratory. Initially there were two or three catalytic vessels capable of taking a few grams of catalyst, and a unit was constructed to provide the hydrogen/nitrogen mixture. A combustion furnace took in hydrogen/air mixtures, and produced nitrogen, containing hydrogen, to which was added some electrolytic hydrogen in order to achieve stoichiometric proportions. A heated palladium/ asbestos tube was used to remove the last traces of oxygen from the gas mixture. A compressor was supplied, which allowed pressures of up to 200 atmospheres to be obtained, and the compressed gas was stored in a high pressure battery, from which it could be distributed to test units within the building, via a high pressure manifold. The first attempts failed to achieve the ammonia concentrations obtained by Haber; erratic results were obtained, and a large amount of osmium was lost as volatile osmium tetroxide. More osmium was obtained from the options taken up, and, finally, Haber's results were replicated. Mittasch had divided the work into two areas; attempts at improving the activity of iron were to be carried out in conjunction with Dr. Hans Wolf, while, with Dr. Georg Stern, he would investigate all other elements. What was the reason for the belief of Mittasch and Bosch that success could be achieved with iron? Nernst and Haber had used iron in their studies, but had not achieved results that gave good reason for optimism. Iron did possess obvious fundamental advantages over osmium and uranium, being cheap and abundant, and it was known that iron easily effected the decomposition of ammonia to hydrogen and nitrogen, at red heat. It was, therefore, considered not unreasonable to expect it to be able to catalyse the formation of ammonia, and so it seemed a profitable line of research to follow. Mittasch, in his earlier work at BASF on nitrogen fixation, had investigated the optimal conditions for the formation of many nitrides, and also ammonia. From this work arose the idea of multicomponent catalysts. If it was assumed that a labile iron nitride was an intermediate in the catalytic synthesis of ammonia, then any addition to the iron which favoured the formation of such a nitride might be advantageous: such additives were termed 'flux promoters'. So Mittasch and his coworkers began to test the activity of all available types of iron and iron compounds and ores, experiments being carried out on the materials as supplied, as well as after the addition of various other compounds. In particular, various catalysts were made on the 'flux promoter' principle, chlorides, sulphates and fluorides of the alkali metals and alkaline earths being added to the iron compounds. The initial results of these experiments were disappointing, and the degree

The History of the Catalytic Synthesis of A m m o n i a

27

of success did not seem commensurate with the efforts expended. Despite many variations in the preparative conditions of the iron, and on the nature of the additives, often no ammonia was obtained, and in other cases only small amounts were observed; no degree of uniformity was observed in the results. Bosch was, at that time, aware of a patent from Schering concerning the effect of alkali on the activity of nickel catalysts for the catalytic dehydrogenation of borneol to camphor, and he believed that this might shed some light on the various effects occurring with the iron catalysts, but the results obtained thus far showed that alkalies had only a very small effect on the activity of iron. During the course of these experiments, a particular form of iron oxide, 'crocus martis', obtained from Merck, yielded a favourable result, but again this was found to be very variable. Haber's apparatus was proving too complicated to enable sufficient rapidity to be achieved in the testing of samples, as it was necessary to dismantle the apparatus completely every time a catalyst change was carried out. Stern devised a much simpler arrangement, consisting of a steel tube of about half a meter in length with internal electric heating and air cooling. It had a small, removable catalyst container which could take about two grams of materials, and which was easily accessible by the turning of handles; two dozen of these vessels were constructed, enabling the rapid evaluation of many hundreds of catalysts. On November 6th, a much more favourable result was obtained with, a sample of Swedish magnetite from Gallivare, which had stood on the laboratory shelf for many years. Under the experimental conditions, 3 % by volume of ammonia was achieved in the exit gas, compared with less than 1 % in the best of the previous experiments. The activity of this material was not improved by the addition of alkali nor alkaline earths. The particular efficacy of this iron compound was ascribed, initially, to its physical state rather than to the additives it contained; it was particularly dense, but also granular and crumbly, and offered much more finely-divided metal area than had the other catalysts previously tried. Thus, the 'compact-porous' nature was considered important, and a patent [33] was submitted in November 1909, claiming the discovery of "catalytic materials for the synthesis of ammonia from its elements, in that the catalytic material has a compact yet porous form". The catalysts appeared to be more effective the higher the specific weight, which seemed to contradict the expected effect of metals on supports; iron catalysts consisting of finely-divided powders, and as thin coatings on inert supports, were far less active than the denser forms. It soon became apparent that it was impossible for this extreme view of a compact, porous mass to persist, as it was discovered that numerous other compact-porous forms of iron, including other magnetite samples, were not as active. It was concluded that, while structural factors might well be important, the composition of the material must be the predominant factor influencing its activity. Despite the variability of the results, by the beginning of 1910 there had been a sufficient number of clear examples of specific promoting effects to enable a patent application [34] to be made relating to the use of promoted catalytic material for the synthesis of ammonia.

28

Chapter 1 : Susan A. T o p h a m

The catalytic work had taken a decisive turn, and, by January 1910, Mittasch and his team had discovered an iron catalyst, containing small amounts of impurities, which was capable of producing as much ammonia as previously obtainable with osmium and uranium. On hearing of the success with the iron catalysts, Haber sent the following letter to the BASF board, on January 17th 1910 [21, 22], "It makes me extraordinarily happy that Herr Bosch and his coworkers have made this great step forward in the area of ammonia synthesis. I congratulate you. It is remarkable how matter continually reveals new facets. Iron, with which Oswald first worked, and which we have studied a hundred times in the pure state, now works in the impure state. I recognise that every trail must be followed to the end. Again, I congratulate Bosch, and hope to learn more news from him tomorrow."

Haber was very interested to see how the success had been achieved, and arranged to visit Ludwigshafen. The news that the professor was to visit caused much excitement amongst the team; he was conducted around the workshop in which the catalytic vessels were made, visited the laboratory, and also saw progress being made on the construction of larger-scale, high pressure facilities. These experiments were to be conducted in a reinforced, explosionproof concrete bunker situated at the extreme northernmost tip of the works. The original patent application relating to promoted catalysts marked the beginning of an extensive programme of systematic experiments. Attempts were made to reproduce the activity of the highly-active magnetite sample, but, although many synthetic magnetites were prepared, none was as active as the Swedish variety. It was clear, from the literature of the time, that natural magnetites could contain various impurities. Experiments were carried out to mimic the exact chemical composition of the Gallivare magnetite, by preparing melts consisting of iron oxidé and various other oxides, and good results were obtained. It was obvious, therefore, that the chemical composition of the magnetite was very important to its catalytic activity, so the experimental programme shifted from the synthesis of magnetites of definite composition, to the systematic study of the influence of all possible additives on iron. To this end, it was necessary that the following conditions be fulfilled. Firstly, the catalyst samples should be prepared from iron of the highest purity obtainable, and two such varieties of iron were obtained from Kahlbaums in Berlin, the purer form being twice as expensive as silver. Secondly, special methods had to be developed for the preparation of intimate mixtures of catalyst and additive; fusion of the metallic iron and additive under pure oxygen was used initially, and later other methods were employed. Finally, the synthesis gas used in the experiments had to be of the highest-possible purity, and, to this end, electrolytically-produced hydrogen and oxygen-free nitrogen were supplied. Many samples of iron catalysts, containing a single additive, were prepared and investigated for activity. The initial results of these investigations revealed that iron alone was only slightly catalytically active, but could be improved (promoted) or worsened (poisoned) by specific additives. By March 21st

T h e H i s t o r y of t h e C a t a l y t i c S y n t h e s i s of A m m o n i a

29

1910, it had been realised that it was probably more important to exclude harmful additives than to include favourable ones. Further systematic experiments were carried out on the promotion and poisoning of iron (and also cobalt and nickel) using the test units simultaneously, arranged in a battery; in this way, fifty to a hundred materials could be tested per week. U p until May 1910, the catalytic samples had all been produced using the melt method. Experiments were then conducted in which additives were incorporated into the iron by wet methods, e.g. by coprecipitation of mixed metal nitrates with pure ammonia, followed by evaporation and heating to dryness in a quartz tube, or by solution of the iron and additive in nitric acid with subsequent evaporation and calcination. This nitrate method was capable of various modifications; iron powder could be impregnated directly with the additive in the form of a nitrate solution, and then calcined, or waterinsoluble oxide could be stirred into molten iron nitrate. The various preparative methods gave essentially the same results in terms of catalytic activity, although some differences did exist which were attributed to a more or less perfect distribution of the additive in the iron. The wet methods allowed the production of certain catalytic materials which were not possible by the fusion method. These systematic experiments led to the following broad conclusions. i

ii

iii iv v

vi

The catalytic activity of iron is improved by the incorporation of finelydivided metal oxides, particularly those oxides which, by virtue of their high melting points and resistance to reduction, present an unchanged structure in the iron. Examples of such promoters are alumina and magnesia, and, to a lesser extent, calcium oxide and barium oxide. Only very small amounts are usually required for the effect to become apparent. The activity of iron is adversely affected by certain substances, particularly sulphur, phosphorus, arsenic and chlorine. Levels of as low as 0.01 % can be harmful, and if these rise to 0.1 %, can completely inhibit the catalytic activity. This was a particularly important discovery, as it explained why so many failures had been observed with iron compounds in the past, since most of the iron preparations used contained sulphur as a result of their method of manufacture. The addition of certain metals and oxides to iron catalysts proved to be of no effect, e.g. copper, silver and titanium dioxide. In the case of the simultaneous presence of promotors and poisons, a limited compensation effect might be observed, but the overall activity is always less than in the absence of the poison. Further improvements can be achieved by producing catalysts from mixtures of more than two components. For example, the combination of iron and alumina can be favourably influenced by small amounts of potash. Enhanced activity is observed for combinations of iron with other catalytically-active metals such as cobalt, molybdenum, tungsten and uranium. This effect was termed 'mutual activation'.

30

Chapter I : Susan A. Topliam

vii Metals which catalyse the ammonia synthesis reaction (osmium, ruthenium, uranium, molybdenum) can also be promoted by added substances, but do not all show improvements of the same magnitude as iron. The type and amount of promoter has to be varied from one metal to another. Typical examples of such combinations are osmium-potassium oxide and nickel-molybdenum. viii A support effect, such as observed in the classical case of platinised asbestos, is of negligible importance with ammonia catalysts, (except with osmium and ruthenium). ix It is necessary to use as pure source of iron as possible in order to avoid poisons. High temperatures are useful in removing volatile poisons, but a single heat treatment is not usually sufficient. x It is particularly important that all the gases used are pure and kept in a pure state. Sulphur-free hydrogen should be employed, and all oxygen compounds should be removed from the gas stream. These results were published in a series of patents [35]. Further discoveries were then made concerning the effect of sintering. The use of similar materials of the same chemical composition, treated under the same reduction conditions, could produce catalysts of differing catalytic activities, depending on whether the preparation involved high or low temperatures. Low reduction temperatures were found to be advantageous, and it was inadvisable to go beyond temperatures which corresponded to the operating conditions. The extent of sintering was found to vary from metal to metal, but could be significantly different depending on the purity of the metal. After the initial discovery of the active iron catalyst, further work was carried out to determine whether better catalysts could be found, and to develop a formulation which would be utilisable in the large-scale operations planned at BASF. Ultimately, it was envisiged that the catalyst would also be sold externally. Mittasch and his coworkers tested over 20,000 catalysts, and, in an elegant and throrough fashion, produced a comprehensive picture of the relative activities of various metals for the synthesis of ammonia, and also of the specific effects of promoters on these metals. The promoter studies were particularly interesting. Mittasch discovered that no sharp demarcation existed between the promoting and indifferent components, nor between the indifferent and the harmful ones. The important factors were the proportions of additive used and the structures of the solid products obtained; comparison of the effects of additives on various catalysts revealed the high specificity of promoter action. A particular promoter did not necessarily act in the same sense and to the same extent on different metals. During the course of these systematic investigations, it became obvious that the iron-alumina-potash combination was the most suitable catalyst for use in the large-scale synthesis of ammonia; despite extensive work, and the gathering together of a great deal of fundamental knowledge, the promoted iron catalyst was not surpassed by any other catalyst.

The History of the Catalytic Synthesis o f A m m o n i a

31

b) Development of the high pressure apparatus [21, 36] At the same time that the catalytic studies were showing the first signs of success, an experimental unit was being constructed which, it was hoped, would enable scale up of the laboratory experiments by a factor of a thousand. The small-scale catalytic reactors operated with a few grams of catalyst whereas the vessel proposed for the larger scale work would contain a few kilograms. At the beginning of 1910, construction started on a vessel consisting of a steel tube of about 1 meter in length, 70 mm in diameter, and wall thickness of 30 mm, which was to be connected to a circulation pump and an ammonia separator. It was essentially the same set-up as that used by Haber, but somewhat larger; it was completed in early spring. The vessel was heated electrically, using external nickel windings, and it had an internal heat exchanger. A small circulation pump enabled a flow rate of 50 litres per hour of compressed gas to be achieved. A chiller unit, containing a toluene-carbon dioxide mixture, was provided for the separation of the ammonia. A battery of compressed hydrogen/nitrogen gas mixture provided the fresh gas. Two catalytic vessels of the new design were enclosed in an explosion-proof concrete bunker, which lay a considerable distance from all other work, and was shielded by walls of black iron plate, since Bosch had already learnt of the dangers of fire and jets of flame caused by selfignition of escaping jets of high pressure hydrogen. Inlet and exit connections and all services were led through the concrete walls to the rest of the plant. The planned experimental conditions were 100 atmospheres and 870 K, and the first catalyst used in this large scale apparatus was the Swedish magnetite, which had performed so well in the recent laboratory investigations. Experiments with this larger vessel came to a rapid end, however; it operated for only 80 hours, then burst. Had the catalyst been osmium, instead of the newly-discovered iron catalyst, the whole world's supply of the precious metal would have been lost! The performance of the steel tubes used as catalytic vessels had been much worse than would have been expected had they contained air at the same temperature and pressure. Investigation of the failed tube showed that it had swollen, and that the inner wall had completely lost its tensile strength, apparently by some change in the material of its construction. The alteration was stepwise, until finally the undamaged part had become so thin that the internal pressure had caused it to rupture. The immediate conclusion was that it had suffered chemical attack. At first, hydrogen was considered harmless, but nitrogen was suspected, as it was known from the literature that iron and ammonia could form iron nitride, a silvery-lustrous, brittle compound; however, chemical investigation revealed no trace of nitrogen in the brittle material. Bosch investigated the failure of the steel personally, in a thorough fashion. He prepared ground sections, and examined them by means of a metallographic etching process, a technique then almost unknown in the chemical industry, but known to him because of his earlier experience in the iron and steel industry. This investigation revealed a zone of light colouration at the surface of the steel which had come into contact with the high pressure gases.

32

Chapter 1 : Susan A. T o p h a m

Carbon-containing steel had been chosen for the material of construction of the pressure vessel because of it mechanical strength; it is usually dark in colour because of the carbon-containing perlite interspersed in a matrix of pure iron. Bosch observed that the perlite had disappeared from the lightened inner portion of the tube, and the structure of the steel had been broken down by fissuring. Decarbonisation of the steel had also occurred, but it was obvious that no pure, soft iron had been formed because the altered material was hard and brittle; it appeared that hydrogen had combined with the iron to give a brittle alloy. In addition to this embrittlement, the steel had been weakened further as a result of the formation, during the decarbonisation process, of methane occluded under high pressure within the steel: in association with mechanical strain, this contributed to the disintegration of the structure. In the case of rapid depressurisation of the surrounding gas atmosphere, the occluded gas could form bubble-shaped bulges in the material, particularly if this occurred at high temperature. Thus it had been determined that the diminution of the mechanical resistance of steel, by the action of hydrogen under high pressure and temperature, was due to two processes, i.e. decarbonisation, and dissolution of hydrogen in iron. A series of experiments was performed, with the aim of overcoming this newly-encountered difficulty. Although the first synthetic ammonia was made in the experimental unit on May 18th, no real sense of success was felt, since the vessels always failed after a short time on line. During these first months, most of the time was taken up with repairs, a run lasting between one and two days being the best that coitld be hoped for. In order to maintain the temperature inside the oven at 870 K, local overheating of the vessel walls (up to 1070 K) was unavoidable, since external heating was being used. Also, despite the fact that the synthesis of ammonia from its elements was an exothermic reaction, and it should have been possible to maintain the temperature of the vessel, once the reaction was underway, without additional input of heat, autothermal operation was not possible in these early stages. There was a conflict of requirements. The length of time to failure of the vessel was increased, by means of careful avoidance of overheating, but, as laboratory experiments confirmed, carbon steel at high pressure and temperature was always transformed by hydrogen in this manner. It was merely a question of how long it took. Repeated experiments showed that, even using the best-available catalysts which permitted operation of around 670-720 K, the vessels lasted only a matter of days. Various measures were considered as means to overcome the problem. Replacement of the external method of heating by internal heating, together with the use of a strong internal insulation and external cooling was proposed, but at first proved to be unsuitable and uneconomic. The insulation material, under high pressure conditions in the presence of hydrogen, had a tendency to break down, and even that which remained intact had a high heat conductance in the presence of high pressure hydrogen. It appeared, at that time, that the idea of using steel as the material of construction of the pressure vessel would have to be abandoned, and a further series of systematic experiments was conducted to investigate all possible metals and materials

T h e H i s t o r y of t h e C a t a l y t i c S y n t h e s i s of A m m o n i a

33

which could be used as alternatives for the construction of the vessel. However, no other metal was comparable with steel in strength and ability to contain high pressures of hundreds of atmospheres, and the samples tested either disintegrated or allowed hydrogen to diffuse through very easily. The position appeared hopeless, and even Bosch himself began to believe that a solution to the problem of the pressure vessel would not be found. Meanwhile, an idea developed for a general solution to the production of such an apparatus. It was accepted that the decarbonisation of perlite and the concomitant formation of brittle iron hydride was unavoidable, so means were sought by which this alteration in structure could be accommodated and rendered harmless. The functions of the wall of the vessel were twofold: it took the pressure, of the highly-compressed gases, and it provided a gastight seal. Bosch attempted to separate the two functions of the vessel wall, utilising two different materials of construction for the two separate duties. The first such vessel consisted of a steel tube lined on the inside by a tube of fine silver of 2-3 mm wall thickness. Unfortunately, owing to the differing rates of thermal expansion of the two materials, the silver tube buckled on heating, and on cooling an end was torn off. There had been great hopes for the lined tube, and the failure was a severe setback for Bosch and his team. The failed apparatus was examined thoroughly by Bosch, who carried out a microscopical investigation of samples removed from the vessel. One Saturday morning at the beginning of February 1911, work was progressing on the problem, with alternatives such as the use of a bronze inner lining being discussed as a last resort, when Bosch suddenly arrived at the solution. He announced, in dramatic fashion, "Use soft iron!". Since soft iron contains only a very small amount of carbon it cannot be damaged by decarbonisation. Although the soft iron would suffer chemical change by reaction with hydrogen, a brittle hydride being produced, once this had occurred the inner lining could suffer no further alteration, and, although embrittled, would still retain its ability to contain the gases, provided it was enclosed in a pressure-resistant shell. Soon after this concept arose, Bosch had another idea which contributed the final piece of technology leading to the solution of the problem of the apparatus; apparently this idea came to him on the way to his office one morning. He proposed that the outer steel tube should be bored with small holes, without affecting its mechanical strength. The hydrogen, which diffused through the inner soft iron lining, could then escape before having the opportunity to build up any significant pressure. In this way, the outer vessel could be protected from hydrogen attack, and should, therefore, maintain its mechanical integrity. This idea arose by inspiration, and Bosch immediately asked the patents department to lodge an application. It was this advance which allowed the technical exploitation of Haber's process to become a reality. A new vessel was constructed, in accordance with the new criteria, consisting of a perforated steel outer shell with a soft iron lining; once the soft lining had become brittle due to formation of hydride, it could suffer no further distortion or failure, because it lay snugly against the outer tube.

34

Chapter 1 : Susan A. Topham

The inner tube contained the gas (except for a small amount of hydrogen which diffused through the wall), and the outer tube withstood the high pressure ; in practice, losses of hydrogen through the perforated outer wall were minimal. The inner tube bore the chemical attack and protected the outer pressure vessel from corrosion. The whole development of the technical process depended, more or less, on the solution to this problem. The first such double tube was put into operation on 5th March 1911; it continued beyond normal expectations, and after it had been on line until the end of April 1911, von Brunck was ready to put the whole of the financial might of BASF behind the development of the large scale synthesis of ammonia. Von Brunck had already had the experience of staking large sums of money on one project — the synthesis of indigo. The participation in the Norwegian saltpeter undertaking was dissolved, the majority of BASF's interest in it being sold, and various other projects were also given up at that time, in order to concentrate on the synthesis of ammonia. c) The supply of pure gases [21, 22] The third major problem which Bosch and his team had been addressing was the production of the reactant gases in sufficient quantity and of sufficient purity. The catalytic experiments had demonstrated the importance of the use of gases of the highest purity, and the cost of the hydrogen now became of vital importance to the profitability of a large-scale operation. Since the summer of 1910, after the initial success of the catalytic studies, the facilities of the redundant barium cyanide plant had been allocated to the ammonia synthesis project. Work began on the construction of a pilot plant, which was to be used to investigate all aspects of the process. Supplies of hydrogen and nitrogen were already available on the barium cyanide plant; the source of hydrogen was the nearby chlorine works, where pure hydrogen was generated electrolytically, and a process for the production of nitrogen had been developed, in which electrolytic hydrogen was burnt in air. The pilot plant was soon established, and by the beginning of 1911 it could be regarded as a small-scale manufacturing plant, being capable of producing about 25 kg of ammonia a day. By this time the original idea of using refrigeration to effect the separation of ammonia had been replaced by the injection of water to form an ammoniacal solution. Further improvements were made to the catalytic pressure vessel; gas heating was substituted for the electric heating coils, and was found to be easier to control, although continual input of heat was still required to maintain the reactor temperatures. A large step forward was made when, in mid 1911, by improvements in the insulation and heat recovery, a vessel containing 5 kg of catalyst was maintained for the first time at high temperature by its own heat of reaction. As improvements were made, the production capacity of the pilot plant rose steadily, reaching 100 kg a day in July 1911. The main source of hydrogen used thus far had been electrolytic hydrogen, which, although eminently suitable for experimental work because of its

The History of the Catalytic Synthesis o f A m m o n i a

35

high purity, was obviously not a viable proposition for use on a large scale, on account of its very high cost. So Bosch and his coworkers had been investigating alternative methods for the production of hydrogen. A cheap, readily-available source of hydrogen could be obtained using the well-known water gas process, in which steam was blown over white hot coke to give a mixture of equal volumes of hydrogen and carbon monoxide. In order to obtain pure hydrogen, it was necessary successfully to remove not only the carbon monoxide, but also the sulphur-containing compounds, with which water gas was always contaminated. The Linde refrigeration process, used to obtain nitrogen from the air, appeared to be a suitable method for removing, by liquefaction, most of the impurities from the water gas, and further work was carried out to investigate the utility of the process. An alternative process for generating purer hydrogen was considered, in which steam was decomposed over red hot iron; at first it appeared promising, but it was later rejected as the regeneration of the oxidised iron mass was found to require too much reduction gas, and the iron gradually sintered becoming less efficient in effecting the reaction. Thus, the production of the two reactant gases, in sufficient purity, for use on the large-scale plant, was to be by refrigeration; nitrogen would be obtained from the air, and hydrogen would be won from water gas. Refrigeration did not, however, remove all traces of carbon monoxide and sulphur compounds from the gas stream, so a further purification step was necessary; Bosch hoped to use a method which had been employed in the manufacture of formic acid salts, namely the absorption of impurities in hot caustic soda solution under pressure. F. Decision to Go Full-Scale: Further Problems to be Overcome [21] Meanwhile,-in the summer of 1911, the board of BASF was making important decisions concerning the future of the ammonia synthesis project. In September of that year, land was acquired in the neighbourhood of Oppau, on the northern edge of the Ludwigshafen works; drawings were begun for a large ammonia factory, and the plans were ready in only a few weeks. The building plans were submitted in November. Although the decision to build the factory had been made, many difficulties still remained to be overcome. The technology was completely new, and still carried with it considerable risks. It was obvious that, since the project relied on the proper integration of many separate parts, the smallest of interruptions in one part could affect the operation and profitability of the whole plant. Bosch undertook the most fundamental investigation and proving of every single plant item, and, in most cases, it was necessary to build completely new apparatus for a particular duty. For example, it was essential that the compressors used should neither allow the valuable and flammable synthesis gas to leak out, nor allow any air to be drawn in at the suction side. The specialist compressor firms at that could not meet these two conditions, so Bosch set about solving the problem himself. In particular, it was impossible to hold a seal against rotary shafts, hence, in order to eliminate loss of gas from the circulation pumps, the blower

36

Chapter 1 : Susan A. T o p h a m

was installed in the gas line so that only the inlet and exit lines remained to be sealed; the whole system was then completely enclosed in a strongwalled steel tube. As the bearing was inaccessible, it was necessary for the motor to run for many months at a time in an aggressive gas atmosphere; the arrangement was known as a 'mole pump', and although it astounded the machine contractors, Siemens, it worked. Further problems which remained centred around seals, flange materials, armatures and instruments etc. Bosch required that all pieces of machinery, apparatus and events be investigated with the utmost care and precision; he later credited his ultimate success to this rigorous and fundamental way of working. On taking delivery of a piece of machinery, he was not content with proving its reliability, but demanded that it be taken apart and every piece characterised. A special laboratory was provided for these measurements. The end of 1911 was an exciting time at Ludwigshafen. Scarcely a couple of days went by at a time without an interruption. Numerous fires occurred, and often the roar of high pressure gas flames could be heard, as a result of a rupture of a main or vessel. Fortunately, no serious accidents occurred. One of the most difficult problems was found to be the bringing back on line of the catalytic vessel, after it had been allowed to cool for maintenance; a thick-walled steel vessel, containing catalyst and heat exchanger, had to be raised to a temperature of 770 K in a reasonable time. The first doublewalled vessel had been provided with an external means of heating only. The vessel was enclosed by a case which contained hot gases, and the necessary heat input through the thick shell resulted in overheating, which became even worse when larger-sized vessels were used, thus increasing the danger to the mechanical strength of the steel during start-up. It became clear that an internal means of heating was required, and many different methods were tried during 1911, without conspicuous success. Bosch considered the idea of using inverted flames in the interior of the vessel: oxygen or air under high pressure would be burnt in the hydrogen-rich synthesis gas i.e. an oxyhydrogen flame inside the vessel was proposed. This idea did not fill his colleagues with enthusiasm, particularly when they learnt that they would have to supervise continually the burning of the flames, by observation with a sight glass; at normal pressures the flame was hardly visible, but it was established that the flame became much more visible at higher pressure, becoming bright blue. It was at this time that von Brunck, possibly worried by the dangers inherent in this form of internal heating, and impelled to examine things thoroughly, decided to make a tour of inspection of the works. Three days later, on December 4th 1911, he died, completely unexpectedly. At that time BASF formed, with Bayer and A G F A , a group of three large chemical concerns united in a mutual partnership (Interessengemeinschaft), with the aim of protecting each other's interests. With the death of von Brunck, it was the turn of Carl Duisberg of Bayer to head the partnership, and Bosch and his team were anxious to see whether the new leadership would back the ammonia venture with as much enthusiasm as had von Brunck. Later in December, it was decided that the project to construct the ammonia

The History of the Catalytic Synthesis o f A m m o n i a

37

factory should be expanded to twice its original size, and that Bosch should be responsible for procuring all the equipment. More land was acquired in the vicinity of Oppau, and test boring began. Installation of essential services commenced on May 7th 1912. Bosch began the task of ordering all the machines and equipment, and his work now took on a different aspect; istead of being concerned solely with technical matters, he now had the construction of the plant to worry about. The first problem was the sandy soil, which was not solid enough to support the heavy machines, which came loose from their foundations. The scale of the exercise was awesome, with large quantities of flammable gas requiring compressing and conducting around the site safely and with as little loss as possible. There were still numerous sceptics expressing doubt as to the final success of the venture, but in autumn 1912 Bosch considered that all the necessary preparative work had been carried out successfully, and reported so to the Board, which was completely convinced of his eventual success. In September 1913, two years after the building work had begun, the Oppau works were put into operation. G. The First Ammonia Plant at Oppau [21, 22] The first factory was designed for a production of 30 tonnes per day of ammonia. Hydrogen, obtained from water gas by refrigeration, was passed through caustic soda solution at 473 K and 200 atmospheres pressure to remove small amounts of residual carbon monoxide, and then mixed with nitrogen from the liquefaction of air; before reaching the synthesis vessels it was passed through a heated guard vessel, a preliminary catalytic vessel which removed the final traces of catalyst poisons. The synthesis vessels themselves were 300 mm in diameter, contained 300 kg catalyst and yielded 3-5 tonnes of ammonia a day. After a few months on line, it became apparent that the Linde refrigeration process for the production of hydrogen was unsuitable for use on such a scale; an alternative method of production of hydrogen was thus required, and work on a replacement source was undertaken in 1914. A member of Bosch's team, Dr. Wilhelm Wild, who had been a student of Nernst, developed a catalytic reaction between water gas and steam. The water gas mixture, consisting of hydrogen and carbon monoxide, on passage with steam over a heated bed of iron oxide, yielded carbon dioxide and additional hydrogen:

CO + H 2 O ?± co 2 + H 2 The process was exothermic and produced excess heat which could be used to generate some of the necessary steam; carbon dioxide could easily be removed from the gaseous products by scrubbing with water. The new catalytic process had the added advantage that the carbon dioxide so formed could be utilised in the subsequent manufacture of fertilisers. Bosch became convinced of the value of this method, and devoted a great deal of effort to its development.

38

Chapter 1: Susan A. Topham

The use of the 'water gas shift reaction' permitted the purification of a hydrogen/nitrogen mixture. Instead of using the Linde process for the provision of nitrogen, it was obtained by first reacting air with red hot coke to give 'producer gas', (a mixture of 60 % nitrogen and 40 % carbon monoxide), which could then be mixed with the water gas, and the carbon monoxide resulting from both sources could be converted to carbon dioxide by the catalytic reaction described above; the effluent gas from this reactor would then consist of a mixture of hydrogen, nitrogen, carbon dioxide and a small amount of carbon monoxide. The carbon dioxide could then be removed satisfactorily by dissolving in water in washing towers. The method so far employed for the removal of residual carbon monoxide, namely by means of caustic soda solution, proved troublesome, however, as the vessels containing the alkali became badly corroded. An alternative method for the removal of carbon monoxide was investigated, which used copper liquor as the absorbent, following the method for the analytical determination of carbon monoxide. Again, the solution was found to be very corrosive, and could not be contained in iron apparatus, so it seemed that more expensive metals would have to be used. There was a danger that the whole ammonia synthesis project might falter on account of the failure to contain the copper solution, with the consequent additional cost of replacement material. Bosch set the young chemist Carl Kranch the task of finding an additive to the copper solution which would give protection against corrosion. Kranch set up a large series of corrosion tests, then went off on his summer vacation. On his return, he observed that only one sample had resisted corrosion, one which had been immersed in copper liquor mixed with a large quantity of ammonia; by the simple expedient of adding ammonia to the solution, the corrosion had been arrested, and this method of carbon monoxide removal went on to be used on the large-scale. The remaining problems with the final purification step for the synthesis gas had been solved, and the gas leaving the copper-liquor gas-scrubbing towers should be pure hydrogen and nitrogen, in the volumetric ratio of three to one. The ratio was maintained precisely by the addition of nitrogen which had been obtained from a Linde unit. H. The Refined Plant at Oppau [21, 22, 37] The new catalytic 'shift' process for the removal of carbon monoxide from the reactant gases was introduced into the large scale plant in February 1915, and the Linde hydrogen facility became redundant later that year. A further improvement to the plant was made at the beginning of 1915, when a building was established dedicated to the synthesis reactors. In April, the first large-size vessel was installed; it was 800 mm in internal diameter, and 12 meters long, and was designed for a yield of 20 tonnes of ammonia per day. A photograph of one of the vessels is shown in Plate 5; the synthesis building is also shown in Plate 6. The yield from the plant rose steadily throughout the next year, reaching 200 tonnes per day in June 1916, and peaking at 230 tonnes per day in

T h e H i s t o r y of t h e C a t a l y t i c S y n t h e s i s of A m m o n i a

39

Plate 5. Large catalytic vessel for ammonia synthesis of Oppau plant. (Reproduced with permission from "Geschichte der Ammoniaksynthese", A. Mittasch, (1951), Verlag Chemie, facing p 128)

December 1917. Thereafter, a steady decrease in output was observed, due to damage caused by air attack on Oppau. The main features of the refined plant were as follows. Water gas and producer gas were manufactured, alternately, in large generators, and the product gas streams so produced were led into gas holders, where they were allowed to^nix and freed from dust by spraying with water. The mixed gas was then saturated with steam, before passing through the first catalytic vessels ('shift' reactors) where the majority of the carbon monoxide was converted to carbon dioxide and more hydrogen was generated. The exit gas was a mixture of hydrogen, nitrogen and carbon dioxide with 2 % carbon monoxide. The hot gas then entered a heat exchanger where the heat evolved in the reactors was used to generate steam and saturate the incoming cold gas with water vapour, all with a saving of energy. Compression of the gas to 25 atmospheres was then effected in the first three stages of a compressor, after which the gas entered, at pressure, towers where the carbon dioxide was removed by solution in water. The carbon dioxide-containing solution was pumped to towers and then flash evaporated through turbines which were

40

Chapter 1 : Susan A. T o p h a m

Plate 6. A m m o n i a synthesis area of O p p a u plant. (Reproduced with permission f r o m "Geschichte der A m m o n i a k s y n t h e s e " , A. Mittasch, (1951), Verlag Chemie, facing p 96)

mounted on the same axle as the pumps, thus recovering half of the energy of compression. The gas, which by now was almost completely free of carbon dioxide (0.5-1.0% carbon dioxide, 1 - 2 % carbon monoxide), was compressed to 200 atmospheres in the fourth and fifth stages of the compressor before passing into the carbon monoxide removal towers, where the carbon monoxide was absorbed in ammoniacal copper liquor; the copper liquor was regenerated by evacuation, and then returned to the absorption towers. The synthesis gas now contained about 77 % hydrogen and 23 % nitrogen, with only very small quantities of catalyst poisons. A stoichiometric ratio of 3:1 was obtained by adding nitrogen from a refrigeration unit. The final purification stage was then reached, in which the gas was passed through a vessel of 300 mm internal diameter and 8 meters high, heated to 623 K, containing a catalyst capable of effecting the conversion of the last traces of carbon monoxide to methane, and oxygen (from the Linde nitrogen) to water; it also removed final traces of sulphur. The gas was then introduced into the circulation system, just before the ammonia absorption stage, so that the final residues of carbon dioxide were removed, before the

T h e H i s t o r y of t h e C a t a l y t i c Synthesis of A m m o n i a

41

Figure 2. The BASF Ammonia Synthesis Plant at Ludwigshafen. (Reproduced with permission from "Im Banne der Chemie, Carl Bosch, Leben und Werk", K. Holdermann, (1953), Econ-Verlag, facing p 81) 1 Production of raw gas, 2 Storage of raw gas, 3 Catalytic reactor for production of hydrogen, and removal of carbon monoxide, 4 Compressor (for production of 25 atmospheres), 5 Removal of carbon dioxide, by means of scrubbing with water, 6 Pump for high pressure water, 7 Compressor (for production of 250 atmospheres), 8 Removal of residual carbon monoxide by means of scrubbing with ammoniacal copper liquor, 9 Pump for high pressure copper liquor, 10 Ammonia synthesis reactors, 11 Condenser and evaporator, 12 Separator, 13 Circulation pump, 14 Gasholder for gaseous ammonia

42

Chapter 1: Susan A. Topham

fully-purified gas was finally introduced into the synthesis reactors. These v 12 meter high, double-walled vessels were half-filled with about 2 tonnes of catalyst, the other half being occupied by a heat exchanger. The effluent gases then went via a circulation pump capable of dealing with 600 cubic meters of compressed gas per hour, to the ammonia absorber, where it was absorbed under pressure by a water spray. The ammonia solution was collected in large, high pressure vessels, let down into lower pressure holders and then pumped, as a 20-30% solution in water, into supply vessels. A simplified schematic diagram of the Oppau plant is shown in Figure 2.

4. Influence of the First World War on the Haber Process The influence of the first world war on the development of the Haber process has been the subject of much debate. Some reports in the scientific journals dating from the years immediately following the war alleged that war was not declared by Germany until the success of the Haber process had been assured [38, 39], and also that the Ostwald process for the production of nitric acid from ammonia was held back from commercial exploitation by the German authorities as a strategic asset [40], Claims were also made that the Haber process was developed with government aid, finance being made available for extensive and secret research [41, 42], The available evidence [43, 21], however, reveals a rather different story. Before the declaration of war, Germany had relied on Chilean nitrates to make up half her annual requirement of fixed nitrogen. These supplies were cut off in August 1914, however, and as the statistics of German nitrate imports for the previous few years reveal, no attempt had been made to build up a strategic reserve. It was apparent that the German authorities had not fully appreciated the implications of a sea blockade, as, in September that year, Professors Haber and Emil Fischer had deemed it necessary to draw the attention of the War Ministry to the situation. Soon afterwards, Fischer submitted a report to the authorities, detailing the required action to be taken to ensure that stocks of nitric acid would be available for explosives manufacture. A very small Ostwald ammonia oxidation unit existed in Gerthe, in the Ruhr, but its capacity was slight and it relied on the use of the precious metal platinum as a catalyst. The government decided to support the development of the Ostwald process, and Bosch, at BASF, became involved in this work. He had already been summoned to the War Ministry in September 1914, and had been alarmed at the lack of awareness of the importance of fixed nitrogen to the success of military operations. By working night and day, Bosch, Mittasch and Lappe developed a cheaper catalyst, based on iron-manganese-bismuth oxide, and an improved reactor, for the Ostwald process, and, by May 1915, this new unit at BASF was producing nitric acid, using raw material ammonia produced by the Haber process. The oxidation unit at Gerthe was enlarged, and further Ostwald units built elsewhere, all of which used ammonia obtained as a by-product from coke ovens. This usage of ammonia for military purposes meant that

The History of the Catalytic Synthesis of A m m o n i a

43

less was available for agricultural use, and the government decided to give financial support to operators of the cyanamide process, since the product obtained could be used directly as a fertiliser. It was decided, late in 1914, that although the Oppau works had been designed for the production of the fertiliser ammonium sulphate, agricultural fertiliser manufacture would cease for the duration of the war, ammonia liquor and sodium nitrate becoming the more important products for military purposes. Despite the Oppau plant's increased production during 1914 and 1915, it became apparent to the authorities, towards the end of 1915, that more capacity was required. Bosch was again summoned to the War Ministry, and the proposal was put forward that BASF should enlarge their facility for the production of ammonia. All the development costs of the Haber process so far had been borne by BASF and its partners; BASF had paid for all the original research and pilot plant work itself, and the Oppau plant had been financed jointly by the Interessengemeinschaft of BASF, Bayer and AGFA, purely as a commercial venture. The enlargement of the synthetic ammonia facility was too great a financial burden for BASF, so the government decided to transfer its financial support from the cyanamide process to the Haber process, and a loan was arranged in April 1916. Since Oppau had already suffered some damage by air attack, it was decided not to enlarge the existing plant, but to build another plant at a lessvulnerable site. The small town of Leuna, in the east of Germany, was chosen. It had the advantages of being on the main railway line from Frankfurt (am Oder) to Berlin, and having readily-available supplies of water and brown coal nearby. Despite numerous problems during construction, including extremely bad weather and shortage of materials, the first ammonia was produced at the new plant on April 27th 1917, only eleven months after building work had started. The design of the new plant incorporated all the improvements made to the Oppau plant, and its capacity, when complete, was 250 tonnes of ammonia per day. On completion, plans were put forward for the enlargement of capacity to over 400 tonnes per day, and by the end of the war the Oppau and Leuna works were capable of producing, respectively, 50,000 and 130,000 tonnes of fixed nitrogen a year. Though they never achieved full capacity because of shortages and interruptions due to the war, they did provide, in 1918, half of Germany's requirement for fixed nitrogen. This statistic is especially impressive when compared with the contribution of the Oppau plant at the beginning of the war; the first Haber plant had the capacity then to supply about 3 % of Germany's needs, although it never actually produced at full capacity. Thus, whilst it is extreme to suggest that the Haber process was a decisive factor in the declaration of war, there is no doubt that the war had a profound influence on the development of the Haber process, to the extent that it did ultimately play a significant role in the later years of hostilities. The scale of operation increased from a small, and still somewhat experimental, unit in 1913, to an extensive facility based on two sites, and capable of producing twenty-five times as much, by 1918. Without this capacity, the war would almost certainly have ended at an earlier date.

44

Chapter 1 : Susan A. T o p h a m

5. Developments Outside Germany [37, 44, 45] During the war, other nations did not suffer shortages of fixed nitrogen to the same extent as did Germany. Britain was able to produce large quantities of by-product ammonia from her town's gas industry, sufficient to supply her allies' as well as her own requirements for nitrates of military value, thanks to an indirect process, developed by Brunner Mond and Company, for the production of ammonium nitrates from ammonium compounds and Chilean nitrates. Nevertheless, extensive work was carried out in many countries, with the aim of producing synthetic nitrogen compounds. In Britain, the Munitions Inventions Department set up a team of research workers to investigate methods for the synthesis of ammonia and its subsequent oxidation; in charge of this work was H. C. Greenwood, who had studied with Haber at Karlsruhe. A team with similar aims existed in France, but by the end of the war neither had progressed anywhere near a viable process. In the U.S.A., the government had supported work on the synthesis of ammonia by the Haber and Cyanamide processes, but it was a commercial organisation that made the first progress in this area. A small, experimental catalytic ammonia unit had been developed by the General Chemical Company in Long Island, using a process similar to that of BASF, but using a different catalyst (a cobalt-sodium catalyst was used). Interest in the Haber process had first been aroused in the U.S.A. when Professor Bernthsen of BASF had given a paper [46] there in 1912. After the entry of the U.S.A. into the war, the fixation of nitrogen came under military control,, and two plants were planned, one to be based on the General Chemical process, the other to be a cyanamide plant; the economics of the Haber process, as compared with other nitrogen fixation methods, were not known outside Germany at that time, so the cyanamide process was still to be reckoned with. The catalytic plant was built at Muscle Shoals, Alabama, but was not successful, producing only a small quantity of ammonia; the process was not sufficiently well-developed when taken to full-scale, and the catalyst was found to die off very rapidly. It was later developed to commercial scale, with a much more active catalyst which lasted far longer than the original one, and the first successful plant was commissioned in Syracuse, New York, in 1921. Meanwhile, a research facility, the Fixed Nitrogen Research Laboratory, was established by the authorities, so the foundations for the development of an American ammonia industry had been laid. Although the team supported by the French government had not made much progress by 1918, an advance had been made by the French engineer Georges Claude. He developed a modified Haber process, which operated at much higher pressure (1000 atmospheres), making the separation of the ammonia product much easier. A higher conversion per pass was also obtained. After successful trials of this process in Paris in 1919, Grande Paroisse erected a full-scale plant in the 1920s. Also, in 1919, the French Minister of Industrial Reconstruction signed a convention with BASF for assistance with regard to the technical details for the economic working of the Haber process, enabling the Kuhlmann company to exploit the BASF

The H i s t o r y of t h e C a t a l y t i c S y n t h e s i s of A m m o n i a

45

patents obtained by the French War Minister under the terms of the Peace Treaty. The British government had intended to build a Haber synthetic ammonia plant, and had bought a site at Billingham, in the north-east of England, for this purpose. At the end of the war, government support ceased, and the project was put up for sale. The Brunner Mond company considered taking it over, but, like the Americans, were unsure of the economics of the Haber process; at that time, it was not known whether fixed nitrogen from this process could compete with Chilean nitrate on cost. Directors of Brunner Mond wished to inspect the BASF plant, and in April 1919 embarked on a mission to Oppau [42, 45], hoping to obtain such detailed information as would enable the design, construction and operation of a similar plant in Britain. BASF were extremely unhappy at this prospect, and informed their employees that if they divulged any information they would be dismissed; as the plant was in the French sector, BASF also informed the French of their threat to close the plant, making 10,000 workers idle and leaving the French authorities to,cope with the situation. The British mission received no cooperation. The plant was shut down during the visit, dials and gauges being blacked out or removed, ladders and stairs taken away, and maker's names removed from machines. All employees stopped work immediately the British entered the buildings. No sketches were permitted on the premises, so these had to be made afterwards from memory. Despite all these problems, a detailed report was written, but it was stolen by the resourceful and determined expedient of sawing through the bottom of the guarded railway wagon in which it had been stored. Fortunately, one member of the mission, Captain Cowap, had kept a set of duplicate drawings, and the report could be rewritten on returning to England. After carfully considering the economic viability of the process, Brunner Mond decided to acquire the Billingham site, and also obtained the rights granted to BASF under British Patent, which had been confiscated during the war. A plant was planned to produce 20 tonnes per day of ammonia and 10 tonnes of nitric acid, and the new company formed to progress the plan, Synthetic Nitrates and Ammonia (later to be one of the founders of Imperial Chemical Industries), began work on the biggest project so far undertaken by any British chemical firm. Research work was initiated on all aspects of the process, and problems similar to those encountered by BASF were met and overcome. Eventually, and with the help of two engineers who had worked at both BASF plants, coalbased ammonia plants were established in Britain, using the same basic technology as that developed in Germany. The Haber process, in several variations, became established in many countries during the 1920s. The process represented a significant technical advance over the fixation of nitrogen by the earlier methods. The energy requirement of the new catalytic process per tonne of fixed nitrogen was considerably less than for alternative processes, and so further development of the electric arc and cyanamide technologies ceased.

46

Chapter 1: Susan A. Topham

6. Developments in Ammonia Technology [47, 48] Between the 1920s and the 1950s, the major changes taking place in the synthetic ammonia industry were related to size. The technology upon which the plants were based changed very little from that originally developed by BASF, but the world capacity for the production of ammonia increased. A greater number of plants were built, and existing units were enlarged, generally by adding parallel streams of similar equipment. Since the 1960s, however, ammonia production has undergone a major change, as a result of improvements in technology. Instead of multiple parallel streams with numerous interconnections to maximise use of equipment, the new plants were based on the concept of large, reliable single streams, involving lower capital and operating costs. In addition, instead of regarding the plant as consisting of essentially unrelated, but linked, individual units, integration of the whole plant was effected in order to make the most effective and efficient use of all its parts; energy generated as a result of one part of the process was recovered, to be used elsewhere in another part of the plant. This concept of large, integrated single stream plants became a reality because of the coming together of several individual advances in technology. Plants continued to be based on coal, with hydrogen generated via the water gas reaction, until the 1950s, when, in the U.S.A., plants based on natural gas feedstocks were commissioned. The use of the methane steam reforming process, which originated from work carried out at BASF in the 1920s and which was further developed by I.C.I, in the 1930s, permitted significant reductions in the capital cost of plant. Also, in 1962, a process developed by I.C.I, at Billingham made possible the production of synthesis gas from light naphthas, whereby the plant simplifications made possible by the natural gas steam-reforming process became available at sites which did not have a ready supply of natural gas. Further simplification resulted from the replacement of the water-wash method for the removal of carbon dioxide by a process using aqueous solutions of potassium carbonate or monoethanolamine, which were much more efficient for this purpose. Development of a low temperature water gas shift catalyst, and the installation of a methanator, enabled the elimination of the copper liquor stage for the removal of carbon monoxide, while advances in pressure vessel technology enabled larger and larger synthesis converters to be used. The most important advance in the engineering aspects of the plant, however, was the development of large centrifugal compressors which could assume the duties previously carried out by reciprocating machines. Although less efficient, the centrifugal compressors cost less, occupied less space and were more reliable than th$ older machines; less maintenance-was required and, more significantly, the centrifugal machines could be driven by steam generated elsewhere on the plant. The concept of integrated plant design was pioneered by the Kellogg Company, and the first three 900 tonne per day plants in the world were built at I.C.I, in Billingham. Plants such as these rapidly rendered the older, smaller units obsolete. The attached flowsheets (Figure 3), show a comparison of the old coal-based process and the new process based on natural gas.

Coal based route to Ammonia (Old technology) Cdal • Air

COKE OVENS \ • »WATER GAS GENERATORS •

Steam

t

SULPHUR REMOVAL BOXES T WATER GAS S H I F T REACTORS (HI«« t.n.p.r.tur.1

Power Station Electricity

L P COMPRESSORS (Rotar» • IP COMPRESSORS

(Reciprocati»«)

WATER WASH CO, REMOVAL TOWERS HP COMPRESSORS

-

CO,

(Reciprocating)

•

COPPER LIQUOR WASH TOWERS -

•

AMMONIA S Y N T H E S I S LOOPS

T

CO (& C 0 2 )

(Hater procoaa)

AMMONIA

Modern route to Ammonia Naphtha (now converted to natural gas) *

Power steam generation integral with process

DESULPHURISATION REACTOR • PRIMARY R E F O R M E R * •

Air

- Steam

- » SECONDARY REFORMER WATER GAS S H I F T REACTOR

(High

temperato..)

WATER GAS S H I F T REACTOR IL«. tempwature)

- CO,

C0 2 REMOVAL • METHANATION REACTOR COMPRESSOR •

(Rotary)

AMMONIA S Y N T H E S I S LOOP

T

(Hat»,

p,oc...>

AMMONIA Figure 3. Comparison of old and modern routes to ammonia at Billingham

48

Chapter 1 : Susan A. Topham

Whereas the older plants required a power station for the provision of electricity and steam, this is no longer necessary, as the steam required to drive the compressors in the newer process is generated on the plant itself. 10000 r

Electric arc p r o c e s s

Single stream ammonia plants

Figure 4. Efficiency of nitrogen fixation 1900

1920

1940

Year

1960

1980

The accompanying graph (Figure 4) shows the cost of fixing a tonne of nitrogen as a function of time. The electric arc process required a colossal 6800 therms for every tonne of nitrogen fixed. The cyanamide process was still more than four times less efficient than a modern ammonia plant, demanding 1800 therms per tonne of ammonia produced. The revolution brought about by the advent of single stream plants is apparent as a step change in efficiency on the graph. Since then, the curve has continued to approach the minimum practically achievable, but in the 1970s this has been accompanied by a gradual rise in the capital cost of equipment. In 1982, a new, energy-saving process was announced by I.C.I. Agricultural Division [49], which permits an even closer approach to the minimum practical expenditure of energy to form a tonne of ammonia. This new process allows efficiency gains together with reduced capital cost, and uses an improved synthesis catalyst, which permits the use of lower pressures in the synthesis loop (80 atmospheres). Although seventy years have elapsed since the first catalytic ammonia plant made ammonia, it is still possible to make significant improvements to the process.

The History of the Catalytic Synthesis of A m m o n i a

49

7. Epilogue The preceding account has shown that Sir William Crookes' prophesy, that the chemist would ultimately solve the problem of providing a replacement for the finite natural reserves of fertiliser, has been fulfilled. Haber had originally thought the technical exploitation of such a process unlikely, based on his original (and over-optimistic!) data on the position of the ammonia equilibrium. When prompted to reconsider his experimental results, by the criticisms of Nernst, who actually made the first synthetic ammonia at pressure, Haber's interest in the technical synthesis was reawakened, and he then carried the ideas through to a successful prototype stage. It was thus Haber who laid the foundations for the technical synthesis of ammonia from its elements. The development of the ideas of Haber to fruition on the large scale was, however, due to Bosch, and his achievements in the space of just a few years cannot be overestimated. Bosch considered, with hindsight, that the solving of the problem of the high pressure catalytic vessel was the vital stage in this work and remarked how surprising it was how, "after months and years with a problem, overnight the brain can provide, subconsciously, the final correlation". In 1919, Haber received the Nobel Prize for his work related to the catalytic synthesis of ammonia. The same distinction was afforded to Bosch in 1931, for his contributions in the development of high pressure technology. A

cknowledgements

The author would like to thank the following for help with research for this chapter:'Dr. Martyn V. Twigg, and the staff of the Information Centre at I.C.I. Agricultural Division (in particular, Mrs. Sandra Brady). Thanks are also due to Professor Dennis Dowden for reading the manuscript and making numerous helpful comments, and to Mr. Philip I. Newbold for preparing the figures.

8. References 1. Crookes, W . : Report of the 68th meeting of the British Association for the Advancement of Science, Bristol, 1898. L o n d o n : J o h n Murray 1898. pp. 3 - 3 8 2. G o r a n , M . : The story of Fritz Haber. University of O k l a h o m a Press. 1967 3. Coates, J. E.: The H a b e r memorial lecture. J. Chem. Soc, 1939, 1642 4. Priestley, J . : Experimental observations on different kinds of air. 1774 5. Berthollet, C. L.: Ann. de Chim. 67, 218 (1808) 6. Hildebrandt, G . F . : Crells A n n . 1, 303 (1795) 7. Dobereiner, J . : J. f. Chem. und Phys. 38, 3215 (1823) 8. K u h l m a n n . F . : Lieb. Ann. 29, 272(1839) 9. Liebig, J . : Chem. Briefe. Heidelberg 6 (10), 86 (1844) 10. Sainte-Claire Deville, H . : C o m p t e s R e n d u s 60, 317 (1865) 11. Ramsey, W., Young, C . : J. Chem. Soc. L o n d o n 45, 88 (1884) 12. Ostwald, W . : Grundlinien der anorganischen chemie. 1900. p. 345 13. Perman, E. P., Atkinson, G . A. S.: Proc. Roy. Soc. 74, 110 (1904): 76, 167 (1905)

50

Chapter 1: Susan A. Topham

14. E. P. 313,950 (1903), in the name of M. Klotz 15. Haber, F., van Oordt, G.: Z. Anorg. Chem. 43, 111 (1904): 44, 341 (1905): 47, 42 (1905) 16. Nernst, W„ Jost, F.: Z. Elek. 13, 521 (1907): Jost, F.: Z. Anorg. Chem. 57, 414 (1908) 17. Haber, F., Le Rossignol, R.: Ber. Buns. Phys. Chem. 40, 2144 (1907) 18. Haber, F., Le Rossignol, R.: Z. Elek. 14, 181 (1908): 14, 513 (1908) 19. Haber, F.: Fünf Vorträge aus den Jahren 1920-23. Berlin: Springer 1924. p. 1. (Nobel Prize Lecture) 20. D R P 235,421 (1908) 21. Holdermann, K.: Im Banne der Chemie, Carl Bosch, Leben und Werk. Düsseldorf: EconVerlag 1953 22. Mittasch, A.: Geschichte der Ammoniaksynthese. Weinheim: Verlag Chemie 1951 23. Haber, F.: Z. Elek. 16, 244 (1910) 24. Haber, F., Le Rossignol, R.: Z. Elek. 19, 53 (1913) 25. Haber, F.: Z. Elek. 20, 597 (1914) 26. Haber, F., Tamaru, S„ Ponnaz, C.: Z. Elek. 21, 89 (1915) 27. Haber, F., Maschke, A.: Z. Elek. 21, 128 (1915) 28. Haber, F., Tamaru, S.: Z. Elek. 21, 191 (1915) 29. Haber, F., Tamaru, S., Oeholm, L. W.: Z. Elek. 21, 206 (1915) 30. Haber, F., Tamaru, S.: Z. Elek. 21, 228 (1915) 31. Haber, F., Greenwood, H. C.: Z. Elek. 21, 241 (1915) 32. Mittasch, A.: Early studies of multicomponent catalysts. In: Advances in catalysis. 2, 81 (1950) 33. This patent was withdrawn on 12th August 1910 34. D R P 249,447 and supplementaries 258,146 and 262,823 (1910) 35. A list of these patents can be found in Ullmann's Encyclopaedia der Technischen Chemie, 2nd Ed., Berlin, Wien: Urban and Schwarzenberg 1928. pp. 420-6 36. Bosch, C.: Die Chem. Fabrik 12, 127 (1933). (Nobel Prize Lecture) 37. Partington, J. R., Parker, L. H.: The nitrogen industry. London: Constable 1922 38. McConnell, R. E.: J. Ind. Eng. Chem. 12 (9), 837 (1919) 39. Jones, C. H.: Chem. Met. Eng. 22 (23), 1071 (1920) 40. Jones, G.: Quart. J. Econ. May 1920, 391 41. Lamb, A. B.: Chem. Met. Eng. 22 (21), 977 (1920) 42. Parke, V. E.: Billingham. The first ten years.: Billingham Press (1957) 43. Haber, L. F.: The chemical industry 1900-30. Oxford: Clarendon Press 1971 44. Zanetti, J. E., Garvan, F. R.: The significance of nitrogen. New York: The Chemical Foundation 1932 45. Reader, W. J.: Imperial Chemical Industries, a history, Vol. 1. The forerunners 1870-1926. London: Oxford Univ. Press 1970 46. Bernthsen, A.: Z. für Angew. Chem. 1, 10 (1913) 47. Lyon, S. D.: Tenth Brotherton memorial lecture, 30th April 1975. Billingham Press 48. Slack, A. V., James, G. R. (eds.): Ammonia. New York: Marcel Dekker Inc. 1973 49. Livingstone, J. G., Pinto, A.: Chem. Eng. Prog. 79 (5), 62 (1983)

Chapter 2

The Electron Microscopy of Catalysts J. V. Sanders

C.S.I.R.O. Division of Materials Science Parkville, Victoria Australia 3052

Contents 1. Introduction A. The Scope

53 56

2. Instruments A. Electron Microscope Components 1. Electron Guns 2. Lenses 3. Vacuum Systems 4. Elemental Analysis Equipment B. Construction of the Instruments 1. Scanning Electron Microscopes (SEM) 2. Transmission Electron Microscopes (TEM) a) In-situ experiments 3. Scanning Transmission Electron Microscopes (STEM) C. Operation 1. Modes of Operation 2. Electron Diffraction a) Powders b) Single crystals D. Performance 1. SEM a) Point analysis and X-ray mapping in SEM b) Scanning Auger analysis 2. TEM 3. STEM E. Optical Diffraction

57 59 59 60 60 60 61 61 62 63 64 66 67 67 67 68 68 69 69 70 71 71 72

3. Specimens A. SEM 1. Typical Specimens 2. Specimen Preparation B. TEM 1. Typical Specimens 2. Specimen Preparation a) Powders b) Pellets c) Replicas d) Model systems

73 73 73 73 74 -74 75 75 76 76 77

52

4.

5.

6.

7.

Chapter 2: J. V. Sanders e) Size determination of dispersed metallic particles 0 Solids C. STEM D. Specimen Damage Interpretation of Images A. SEM 1. Resolution 2. Quantitative Uses B. T E M 1. Contrast Mechanisms a) Adsorption b) Diffraction contrast c) Orientation effect d) Thickness effect e) Dark-field observations f) Phase contrast images g) Fourier images C. STEM D. Elemental Analysis E. Image Processing Atoms, Particles, Steps A. Single Atoms B. Clusters 1. Metallic atoms 2. Non metallic Clusters C. Particle Shape 1. Resolution of Projected Shape 2. Thickness . : 3. Multiply-Twinned Particles (MTP) D. Surface Steps 1. T E M , Dark-field, Forbidden Reflexion 2. T E M , Bright-field 3. Topographic Images 4. Phase Contrast 5. Decoration 6. Surface Steps: Grazing Incidence Applications to Catalysts A. SEM Applications B. T E M Applications 1. Particles a) Particle size b) Sintering c) Redispersion d) Particle movement 2. Carbonaceous Deposits a) Coke structures b) Coke formation on catalysts 3. Supports a) Silica b) Aluminas c) Zeolites 4. Catalyst Characterization C. Surface Structure D. Applications of STEM References

77 78 78 78 79 80 83 85 85 85 85 86 88 89 90 94 98 98 99 101 102 102 104 104 107 108 108 109 Ill 112 112 113 113 114 114 115 117 118 121 121 121 122 124 125 127 127 130 133 133 133 135 141 145 146 148

The Electron M i c r o s c o p y o f Catalysts

53

1. Introduction Many practical catalysts have a complex geometric structure with a range of hierarchies of porosity to enable reacting gases or fluids to reach as much of the surface of the catalyst as possible, and of course for the products to diffuse out and away from the catalyst. The catalysing surface frequently consists of a complex chemical mixture of different phases produced by an empirically evolved chemical process which may incorporate activators, promoters, and so on. The active component is frequently in the form of very small crystals dispersed on the large surface area of a spongy supporting component of the catalyst. The behaviour and misbehaviour of this type of catalyst can depend upon the structure and composition of the active component as well as the morphology of its supporting medium. In order to understand the catalyst's behaviour it may be necessary to examine its structure at a range of magnifications from a few hundreds to a few millions of times. This will allow us to determine the form and distribution of the active components, the nature of the porosity, the distribution and nature of a coke or poisoning deposit, and so on. This can be achieved by the use of electron microscopy in a variety of forms. Research chemists and industrial chemists have only recently started to use these techniques. In the past they have probably been inhibited from making full use of them because of a lack of experience or appreciation of the various different ways in which an electron microscope can be used. Furthermore, whilst it is relatively easy to produce an image of a specimen in an electron microscope, there can be many difficulties in the interpretation of blacks and whites in the image. Intuitive interpretation can be very misleading, and the appearance of a micrograph can be quite sensitive to the instrumental settings under which the image was recorded. The theoretical background for the understanding of image formation and the various ways of using an electron microscope are in literature not usually covered by chemists, and anyway it tends to be scattered through a variety of journals. One of the aims of this article is to present a simplified versions of how catalytic chemists can make use of electron microscopy in its various froms so that they can assess if and when it might provide useful information about their catalysts, and to make them aware of how the techniques can be used not only for producing images, but also in the diffraction mode for identifying phases and obtaining structural information about crystals, and in the analytic mode to examine elemental composition and distribution. Whilst the main emphasis is placed on the usefulness of the technique, at the same time the difficulties and limitations are considered in some detail. Whilst measurements of such properties as surface area, porosity, or particle size give a mean value or a distribution of values for a parameter representing this property, it is often necessary to have a physical picture or model of the catalyst structure in order to understand and explain its behaviour. The easiest means of examining the structural detail of a catalyst beyond what is visible directly by eye is by optical microscopy. The technique is

54

Chapter 2: J. V. Sanders

simple and its operation naturally understood, but the range of magnification is limited to less than about 2000X because of the limit in resolution imposed by the wavelength of light used to illuminate the specimen [1], At the higher magnifications within the optical range the depth of focus and working distance are very small, and become inhibiting when rough surfaces are to be examined. An extension of magnification by several orders of magnitude up to about 100,000 is obtained with a scanning electron microscope (SEM). Current instruments of this type are convenient and quick to use, and they have provided a natural extension to optical microscopy. Apart from the extended range of magnification available they have additional advantages of much better depth of focus than optical microscopy, and their images have an inherent three-dimensional appearance which produces realism to surfaces with a rough or porous texture. Furthermore, the instruments normally accept macroscopic-sized specimens which can be easily handled, and their preparation is usually simple and quick. However, even the best conventional current instruments of this type have at the best a resolution limit of about 5 nm, and are often much worse. Beyond this it is necessary to use a transmission electron microscope (TEM) which has become the work-horse of high resolution microscopy. The instruments are expensive, the preparation of specimens can be tedious, and the interpretation of the electron micrographs is frequently complicated; nevertheless the TEM has become an essential component of catalysis research. Over the last 30 year period in which good instruments have been commercially available their performance has been continually improved. Current instruments are routinely used at final magnifications of millions of times (for example a primary magnification of 500,000X, followed by optical enlargement up to 10X) to produce micrographs with detail of at least atomic dimensions. As will be shown later, this does not mean that structures can be routinely imaged at atomic resolution. The addition of scanning facilities to a TEM produces a scanning transmission electron microscope (STEM) which has some peculiar advantages, but the conventional commercial instruments have spatial resolution which is much less in this mode than in direct transmission. The addition of a field emitting electron gun to a STEM instrument improves the resolution to close to that of a TEM, but at a considerable cost. Such instruments are commercially available, and are being evaluated in a number of laboratories interested in catalytic materials. A comparison of the various techniques of electron and optical microscopy is given in Table 1. This article is concerned with the use of all these ways of using fast electrons to obtain structural information about catalysts and surfaces used in catalytic investigations. The various techniques are frequently complementary, and a wise person uses them all to obtain the best idea of the structure of a catalyst. As a bonus, it is frequently possible, given the appropriate equipment, to obtain a rough elemental analysis of the specimen either by an energy analysis of the X-rays excited from the specimen by

The Electron Microscopy of Catalysts

£ w H i/J oi X

D. _« . oo 5. c 10 2 H d o •tua ~x E -o£ o = in "O Ì« W I-

O —I E -a ^ Q. 15 X UJ 3 V%»

10nm Figure 41. Lattice image of a crystal of the zeolite ZSM-5/11. Arrows in a show the positions of faults, and the optical diffraction pattern f r o m this image in b contains streaks similar to those in Figure 40, confirming that intergrowth faults exist in the crystal

to follow the behaviour of the metallic component. Frequently it is found as clusters within the zeolite, with the cluster size controlled by the dimensions of the pore in which it forms. In this way uniformly sized particles of Pt of 0.6-1.3 nm diameter have been produced in the largest cavities in Y zeolite (1.3 nm), by a mild treatment (573 K in 0 2 , then 573 K in H 2 ). However, a more drastic treatment (873 K in 0 2 , then 573 K in H 2 ) produces larger particles (0.6-2.0 nm) known by TEM still to be within the zeolite. Those greater than 1.3 nm in diameter must be formed at the expense of damage to the aluminosilicate framework [406], In L zeolite (one set of channels, 0.71 nm aperture), containing Pt exchanged into the zeolite, reduction produces a variety of different types of particle [407], They are: (i) 10-60 nm particles outside the channels. (ii) 1-25 nm particles and 1.5 x 4 - 7 nm cylinders both inside and outside the channels. (iii) unstable particles, 0.6-0.8 nm within the pores. An examination of a series of transition metals exchanged into zeolites and then treated, showed that most of the particles were less than 1.3 nm in Na Y zeolite, with a geometric arrangement and separation consistent with the geometry of the pores. There were a few larger crystals, around which the zeolite crystal was locally damaged [408], Particles of Ni on or within a zeolite can be distinguished, and are seen [409] to be angular when on the surface of zeolite X, but round when within the zeolite. Phase contrast fringes from the zeolite terminate quite close to the included metallic particles, showing that any destruction of the aluminosilicate framework is very localised. Exner et al. [409] have made some suggestions about how the Ni may diffuse through the crystal and break framework bonds as it aggregates.

The Electron Microscopy of Catalysts

141

The structure of particles of Pt and Pd in faujasite-type zeolites has been reviewed by Gallezot in 1979 [410], 4. Catalyst

Characterization

Many catalysts are complex mixtures, and it is usually a routine matter to select the various components in a mixture of crystals, take their electron diffraction patterns, and so identify the phases as has been done, for example by Sharma et al. [412] for the V 2 O s — K 2 S 0 4 system. In other examples given below, TEM has been used to try to understand the relationship between the structure of the catalyst at an atomic level, and its catalytic behaviour. For example, graphimets, formed by the reduction of salts of Fe, Co, Cu, Ni or Pt which have been intercalated into graphite, are found to consist of small particles of the metal or its oxides on an external surface of the graphite [413] (Figure 42). The metallic crystals are far too small (1-10 nm in size) for single crystal electron diffraction patterns to be obtained from individual particles, but the lattice fringes ( ~ 0 . 2 nm) obtained in T E M high resolution images can be interpreted with the assistance of optical

Figure 42. TEM of Pt/graphimet: a low magnification, showing that the metallic particles tend to segregate to steps (arrows) on the graphite surface, b high magnification shows that there are fringes through the dark particles, probably from (111) planes, indicating that they are crystalline even though they contain only about 10 lattice planes (Courtesy D. Smith)

142

Chapter 2: J. V. Sanders

Figure 43. Appearance ot'a typical C o M o S , / ' / - A I 2 0 3 catalyst after use for coal h y d r o g e n a t i o n : a lower magnification shows the p o r o u s n a t u r e of a typical fragment, b higher magnification image contains black lines f r o m individual S — M o — S sheets lying parallel to the electron beam, a n d patches of fine (0.27 nm) fringes f r o m the (100) spacing of M o S 2

transforms. In HDS catalysts the active component is MoS 2 together with Co or Ni in more than trace amounts. Although Co 9 S 8 has been detected in X-ray powder patterns, it is claimed [414] that the activity promoted by the Co comes from a close association of Co and MoS 2 , rather than synergetically from coexisting Co 9 S 8 and MoS 2 . High resolution TEM images of these catalysts (Figure 43) contain many single black lines at optimum defocus, which behave, with defocus, as is expected for a single S—Mo—S layer viewed edge-on. Sometimes "books" of up to about five sheets occur (Figure 44) (e.g. on Si0 2 rather than A1 2 0 3 ). Micrographs contain no evidence of the Co or Ni promoter [415]. It is thought that the promoter is at least necessary to produce or maintain the high degree of dispersion of the MoS 2 , and without it the MoS 2 would form into thicker "books" of molybdenite. It is possible that it also provides an electronic component by modifying the semi-conductor properties of the MoS 2 [416].

rhe Electron Microscopy of Catalysts

143

Figure 44. Laboratory prepared HDS catalysts: a Ni — MOS 2 /-/-A1 2 0 3 , where the MoS 2 mostly forms single or double layers, b Ni—MoS 2 /Si0 2 , in which the MoS 2 forms books of up to about five S—Mo—S layers, with a spacing of about 0.6 nm

HDS catalysts made with graphite as a support have been found in laboratory tests to have "outstanding performance" [417], In this case TEM shows that the MoS 2 lies parallel to the substrate, so that images contain lattice images of both graphite [(100) spacing 0.213 nm] and MoS 2 [0.2737 nm]. The latter set gives an arc in an optical transform of the image (Figure 45), showing that the MoS 2 is flat but randomly oriented about [001] on the graphite. On other supports (Ti0 2 , Z r 0 2 ) the MoS 2 forms a compact single or double layer skin on the surface of the support, which appears as an intense black line (or lines) surrounding the support crystals, in underfocus

144

Chapter 2: J. V. Sanders

Figure 45. a Electron micrograph of a M o S 2 / g r a p h i t e catalyst, b Optical t r a n s f o r m of the micrograph, showing that the M o S 2 is disordered (arcs in the transform) on the graphite surface (lattice fringes give spots in the transform)

images. Interpretation of many of these images is aided by the use of optical diffraction and filtering techniques. In order to understand the source of catalyst behaviour, experiments often involve the determination of the relative activity and selectivity of a group of related catalysts, dividing them into "good" and " b a d " categories, and then examining the catalysts for structural differences to determine the desirable features. With a tilting stage in the TEM, it is a rather simple matter to obtain zone-axis diffraction patterns giving the orientation of individual

The Electron Microscopy of Catalysts

145

crystals (providing they are more than about 0.1 ¡am in size); from the shape of the crystal and a knowledge of its crystal structure, one can then work out the structure of the exposed surfaces. Defects in the bulk of the crystal can be seen in lattice images. This type of work has been carried out on W—Mo oxide catalysts, used for the selective production of acrylonitrile or acrolein from propene. It was found that the selectivity of the oxidation by W 0 3 depended upon the morphology of the crystals; also, whilst the selectivity depended upon the surface arrangement, but not upon the bulk structure, the activity did depend upon the bulk structure [418], In this system it has been suggested that crystallographic shear (CS)-planar faults in the structure by which nonstoichiometric composition is accommodated in the crystals — could influence the selectivity of the oxide catalyst [419, 420], Experiments showed that selectivity increased when shear planes were present in Mo—W oxides, and it was considered that the effect could be associated with either the structure, an electronic property, or availability of 0 2 vacancies at the CS plane. Catalysts used commercially (Te—Mo oxides) for the same process have been examined in a high voltage TEM where they can be watched while being heated in an oxidising/reducing atmosphere [421], at typical operating temperatures. It was found that in an atmosphere of C 3 H 6 : 0 2 at 673-723 K the catalyst consisted mainly of T e 2 M o 0 7 and /?-TeMo 5 0 1 6 , neither of which contained CS planes. Whilst allowing that anion vacancies may play an important role in the catalytic reaction at the surface, Gai et al. [421] do not believe that the presence of shear planes is necessary for catalytic selectivity in this system. Crystal defects in catalysts have also been found in lattice images of L a P 0 4 crystals; however, in this case no specific activity was associated with them, but rather with the morphology of the surfaces of the crystals. It was found that "good" L a P 0 4 catalysts for the hydrogenation of aromatic chlorides contained mainly acicular crystals exposing preferentially (hko) surfaces, in which there are chains of La and P 0 4 ions. "Bad" catalysts exposed a greater proportion of other surfaces [422], C. Surface Structure Idealised systems using surfaces of single crystals have provided basic information on the structure of the surfaces and adsorbed overlayers by LEED, the geometric arrangement of adsorbed molecules of increasing size and complexity by EELS, the strength and nature of the adsorbed species by TPD, and the catalytic behaviour [Review, 423], Electron microscopy has contributed rather less, mainly because the surface of the specimen is likely to become contaminated while it is being examined under the conditions existing in routinely used TEMs [424], However, as discussed in Section 5.D, under special conditions surface structures can be resolved by a variety of techniques. The most dramatic results have come from achieving UHV conditions around a specimen observed by grazing incidence microscopy. This technique

146

Chapter 2: J. V. Sanders

was popular in the early 1950s, but was not subsequently developed until recently, probably in part because of the drastic foreshortening in the image, and also because contamination of the specimen obscured the finer details of the surface structure [425], When vacuua of the order of 10~ 6 -10~ 7 Pa are achieved around the specimen, within the T E M , detailed structure in steps on surfaces can be seen [426], Using a special stage in their T E M , Osakabe et al. [425] were able to follow the morphology of the cleaning-up of a silicon surface, the movement of steps on the clean surface, the relaxation of the surface to the 7 x 7 structure and its subsequent transformation to a 1 x 1 arrangement above 1083 K, and then the deposition of very thin ( < 0 . 1 nm) gold deposits on this surface. Grazing incidence observations have also been made in a STEM on very small crystals of MgO and NiO, in which surface steps were resolved. This system provides the additional facility of electron energy loss spectroscopy, by which small parts of the crystal can be examined, and energy losses from the excitation of surface states can be detected [427], Surface structure has also been resolved at the edges of crystals in direct T E M images at very high resolution {e.g. MgO, [428]; graphite [429]). Resolution at an atomic level on gold crystals has recently been achieved at 500 keV [430], and images show steps at the edges of crystals, and reconstruction of a (110) surface which confirms the new model proposed by Bonzel and Ferrer [431], Interpretation of such images, particularly at the edges of crystals, must be made carefully, and to be convincing should (as in the case quoted above) be accompanied by images computed from models and a good knowledge of the defect of focus. In the micrograph reproduced in Figure 19, channels through the structure appear as white dots and the atoms are black. However, a small change of defocus (to —97.5 nm) reverses the contrast, so that columns of atoms appear as white dots [430]. Transmission images through thin sheets of graphite show surface steps, sometimes in the form of closed loops which expand with exposure to the electron beam, which are thought to be of atomic height [429], In this case the steps are visible because of phase-contrast conditions in the bright-field image. D. Applications of STEM Much of the recent work on STEM systems has been concerned with demonstrating the value of the technique and establishing the breadth of applicability of micro-electron diffraction, X-ray elemental analysis and electron energy-loss spectroscopy at a high spatial resolution [Review Cowley, 432]. These uses will become more widely applied to catalytic problems as instruments become more generally available for routine une. An example of the use of microdiffraction is provided by Pennycook [433] who established that all the individual particles which he examined in a R u / S i 0 2 catalyst were single crystal, h.c.p., i.e. the conventional crystal structure. The epitaxial relationships between metallic particles and their support have been investigated by obtaining microdiffraction patterns from individual crystals, only about

I lie E l e c t r o n M i c r o s c o p y o f C a t a l y s t s

147

Figure 46. Example of the use of microdiffraction facilities in S T E M to show that in this case a metallic particle (2 nm) is epitaxed on the support (catalyst, P d / A l 2 0 3 ) . (1) and (2) are bright field images with different beam convergences. (3) is the microdiffraction pattern f r o m the Pd particle arrowed in (1). (4), (5) and (6) are dark field images f r o m the single beams in (3) marked (4), (5) and (6) respectively. Epitaxy is shown only for the (100) spot in (5) ( R e p r o d u c e d with permission f r o m Dexpert, H . ; F r e u n d , E.; Lesage, E.; Lynch, J. P.: Studies in Surface Science in Catalysis 11, 53 (1982))

1-2 nm in size, on y-Al 2 0 3 [434], An example of microdiffraction and darkfield images from a variety of diffracted beams is shown in Figure 46. The analytic facilities in a STEM have been used to examine Pt—Ru/yA1 2 0 3 catalysts, and it was found (Figure 47) that in some preparations the individual metallic particles (1-2 nm) were alloys, whereas, in other preparations, the Ru and Pt were separated [435], The extent of the dispersion of NiO on silica-alumina supports of varying Si/Al ratio has been found to increase as the alumina concentration increased [436]. Microanalysis has also been carried out on zeolite preparations, and it has been shown that for ZSM-5 and ZSM-11 the Al/Si ratio, whilst reasonably constant within individual crystals, varied by a factor of up to 3 between crystals. In ultra-thin sections, the core of an aggregate could be A1 rich by up to four times [437], The shape of individual Pt catalyst particles in the nanometer size range have been determined to about atomic resolution, and the scattered intensity distribution printed out in a simple form for individual particles [438].

148

C h a p t e r 2: .1. V. S u n d e r s

Figure 47. Results of elemental analysis of X - r a y s f r o m f o u r d i f f e r e n t metallic particles on a P t — P d / A l 2 0 3 s u p p o r t s h o w i n g t h a t their c o m p o s i t i o n varies. S o m e a r e f o u n d to be p u r e Pt, o t h e r s a r e alloys. T h e size of t h e metallic particles is 5 - 0 . 5 n m , a n d they h a v e been individually analysed. ( R e p r o d u c e d with permission f r o m D e x p e r t , H . ; F r e u n d , E . ; Lesage, E . ; L y n c h , J. P . : Studies in S u r f a c e Science in Catalysis 11, 53 (1982))

The technique holds great promise, but it is not clear how the apparent behaviour of the specimen may be influenced by the very high electron density in the smallest probes used. It has been shown, for example, that holes the size of the electron probe ( < 1 nm) can be drilled in crystals of diamond by a stationary probe [439], Applications of the STEM technique to the study of the movement of individual atoms and clusters have already been considered in Section 5.

7. References 1. A b b e , E.: Archiv, f. M i k r o s k o p i s c h e A n a t . 9, 4 1 3 (1873) 2. Heidenreich, R. D . : F u n d a m e n t a l s of t r a n s m i s s i o n electron microscopy. N e w Y o r k : I nterscience P u b l i s h e r s 1964 3. Hirsch, P. B., Howie, A., N i c h o l s o n , R. B., Pashley, D. W., W h e l a n , M. J. W . : Electron microscopy of thin crystals. L o n d o n : B u t t e r w o r t h s 1965 4. Cowley, J. M . : D i f f r a c t i o n physics. A m s t e r d a m , O x f o r d : N o r t h - H o l l a n d Publishing C o . 1975 5. H u m p h r i e s , C. J . : R e p o r t s o n t h e Progress in Physics 42, 1825 (1979) 6. Spence, J. C. H . : E x p e r i m e n t a l high-resolution electron m i c r o s c o p y . O x f o r d : C l a r e n d o n Press 1981 7. Treacy, M . M . J., H o w i e , A . : J. C a t a l . 63, 265 (1980)

The Electron Microscopy of Catalysts

149

8. Kishimoto, S.: J. Phys. Chem. 77, 1719 (1973) 9. Craido, J. M „ Herrera, E. J., Trillo, J. M.: In: Catalysis I. (Hightower, J. W. ed.). A m s t e r d a m : N o r t h - H o l l a n d 1973, p. 541 10. Baird, T . : Catalysis 5, 172 (1981) 11. Howie, A.: Characterization of catalysts. (Thomas, J. M., Lambert, R. M. eds.). Chichester, U K : J o h n Wiley & Sons 1980, p. 89 12. Chen, J. R„ G o m e r , R.: Surface Sei. 79, 413 (1979) 13. Thurstons, R. E., Walls, J. M . : Field-Ion Microscope and Related Techniques. Birmingh a m : Warwick Publishing Co. 1980 14. Muller, E. W.: In: M e t h o d s of surface analysis. (Czanderna, A. W., ed.). A m s t e r d a m : Elsevier Scientific Publishing Co. 1975 15. Block, J. H . : In: Electronic structures a n d reactivity of metal surfaces. ( D e r o u a n e , E. G., Lucas, A. A., eds.). New Y o r k : Plenum Press 1976, p. 485 16. Jones, J. P.: Chemistry and Physics of Solids and their Surfaces 8, 18 (1980) 17. Boyes, E. D . : In: Electron Microscopy a n d Analysis 1981. (Goringe, M. J., ed.). London. Conference Series N u m b e r 61. Institute of Physics, L o n d o n 1981, p. 27 18. Humphreys, C. J „ Spence, J. C. H . : Optik 58, 125 (1981) 19. Lauer, R.: Advances in Optical and Electron Microscopy. (Barrer, R., Cosslett, V. E., eds.). 8, 137(1982) 20. Kasper, E.: Advances in Optical and Electron Microscopy. (Barrer, R., Cosslett, V. E., eds.). 8, 207(1982) 21. Baker, R. T. K „ Harris, P. S.: J. Phys. E: Sei. Instr. 5, 793 (1972) 22. Mills, J. C., Moodie, A. F . : In: 8th International Congress on Electron Microscopy. (Sanders, J. V., Goodchild, D. J., eds.). C a n b e r r a : Aust. Academy of Science 1, 1974, p. 182 23. Barna, A., Barna, P. D., Pocza, J. F.: Vacuum 17, 219 (1967) 24. Valle, R., Genty, B., M a r r a u d , A., Cadoz, J.: In: Electron Microscopy and Analysis 1981, Institute of Physics Conference Series No. 61. Bristol a n d L o n d o n , The Institute of Physics 1981, p. 35 25. Poppa, J.: J. Vacuum Sei. a n d Tech. 2, 42 (1965) 26. Takayamagi, K., Yagi, K., Kobayashi, K., H o n j o , G . : J. Phys. E.: Sei. Instr. 11, 441 (1978) 27. P o p p a , H „ H e i n e m a n n , K . : Optik 56, 183 (1980) 28. Parsons, D. F . : In: Physical aspects of electron microscopy and microbeam analysis. (Siegel, B. M., Beaman, D. R „ eds.). New Y o r k : J o h n Wiley & Sons 1975, p. 267 29. Chung, T. T., Dash, J., O'Brien, R. J.: 9th International Congress on Electron Microscopy 1. (Sturgess, J. M., ed.). T o r o n t o , Microscopical Society, C a n a d a 1978 30. Heinemann, K., Rao, B„ Douglass, D. L.: Oxid. Metals 9, 379 (1975) 31. Rhoades, B. L.: 9th International Congress on Electron Microscopy 1. (Sturgess, J. M., ed.). T o r o n t o , Microscopical Society, C a n a d a 1978, p. 70 32. Fryer, J. R.: The chemical applications of transmission electron microscopy. L o n d o n : Academic Press 1979 33. Gai, P. L., Goringe, M. J.: Proc. Ann. Meeting Electron Microscopical Soc. A m e r . 39th. 1981, p. 68 34. Pashley, D. W . : Advances in Physics 14, 327 (1965) 35. Swann, P. R . : 9th International Congress on Electron Microscopy 1. (Sturgess, J. M., ed.). T o r o n t o , Microscopical Society, C a n a d a 1978 36. Fujita, H „ K o m a t s u , M . : Jap. J. Appl. Phys. 15, 2221 (1976) 37. Robinson, V. N. E.: In: Scanning electron microscopy. (Johari, O., ed.). C h i c a g o : I.I.T. Research Institute 1975, p. 51 38. Crewe, A. V.: Chemica Scripta 14, 17 (1979) 39. Ulaut, M . : Proceedings E M S A . (Bailey, G. W., ed.). Baton R o u g e : Claitors Publ. Div. 1979 40. Butler, J. H . : I n : Proc. A n n . Meeting Electron Microscopical Soc. A m e r . 39th, 1981, p. 136 41. Philips supertwin lens in the EM420 is quoted as having C s = 1.2 m m . 1983 42. Cowley, J. M „ Spence, J. C. H . : Ultramicroscopy 3, 433 (1979)

150

Chapter 2: J. V. Sanders

43. Lynch, J. P., Lesage, E., Dexpert, H., F r e u n d , E.: In: Institute of Physics Conference Series 61. (Goringe, M. J., ed.). L o n d o n : Institute of Physics, L o n d o n 1981, p. 67 44. Agar, A. W., Alderson, R. H., Choscoe, D . : Principles and practice of electron microscope operation. I n : Practical methods of electron microscopy. (Glauert, A. M., ed.). Amsterd a m : N o r t h - H o l l a n d 1974 45. Andrews, K. W., Dyson, D. J., Keown, S. R.: In: Interpretation of electron diffraction patterns. L o n d o n : Hilger and Watts 1967, p. 17 46. Beeston, B. E. P.: In: Electron diffraction a n d optical diffraction techniques. (Glauert, A. M., ed.). A m s t e r d a m : N o r t h - H o l l a n d 1973, p. 1 47. G a r d , J. A.: In: Electron microscopy in mineralogy. (Wenk, H.-R., ed.). Berlin: Springer-Verlag 1976, p. 53 48. Riecke, W. D . : In: Electron microscopy in materials science I. (Valdré, U., Ruedl, E., eds.). Brussels: Commission for E u r o p e a n Committees 1975, p. 81 49. Jaeger, H. G . : Private communication (1981) 50. Pease, R. F. W „ Nixon, W. C.: J. Sei. Instruments 42, 81 (1965) 51. Pease, R. F. W . : 4th Annual S E M Symposium. (Johari, O., ed.). Chicago: I.I.T. Research Institute 1971, p. 11 52. Ledebo, L.-A.: In: Electron microscopy and analysis. (Goringe, M. J., ed.). Bristol: Institute of Physics Conference Series 61, 1981, p. 513 53. Lifshin, E.: I n : Scanning electron microscopy. (Wells, O. C., ed.). New Y o r k : M c G r a w Hill 1974, p. 243 54. Newbury, D. E.: In: Scanning electron microscopy. ( O ' H a r e , A. M. F., ed.). Chicago: I.I.T.R.I. 1979, p. 1 55. Joshi, A., Davis, L. E., Palmberg, P. W . : In: M e t h o d s of surface analysis. (Czanderna, A. W., ed.). A m s t e r d a m : Elsevier Scientific Publications 1975, p. 159 56. M a c D o n a l d , N. C.: Appl. Phys. Letters 16, 76(1970) 57. M a c D o n a l d , N. C., Waldrop, J. R . : Appl. Phys. Letters 19, 315 (1971) 58. Holloway, P. H., Battacharya, R. S.: Surface and Interface Analysis 3, 118 (1981) 59. Wells, O. C.: In: Scanning electron microscopy. New Y o r k : McGraw-Hill 1974, p. 87 60. Venables, J. A., Janssen, A. P., Harland, C. J., Joyce, B. A . : Phil. Mag. 34, 445 (1976) 61. Alford, N. A., Barrie, A., D r u m m o n d , I. W., Herd, Q. C.: Surface and Interface Analysis 1, 36(1979) 62. Stein, D. F., Joshi, A . : Annual Reviews of Materials Science 11, 485 (1981) 63. Mathieu, H. J., Landolt, D.: Surface and Interface Analysis 3, 153 (1981) 64. Bursill, L. A., Barry, J. C „ Hutchinson, J. L.: Optik 65, 271 (1983) 65. Moodie, A. F . : private communication (1983) 66. Cosslett, V. E „ C a m p s , R. A., Saxton, W. O., Smith, D. J., Nixon, W. C., A h m e d , H., Catto, C. J. D., Cleaver, J. R. A., Smith, K. C. A., Timbs, A. E., Turner, P. W., Ross, P. M . : N a t u r e 281, 49 (1979) 67. Jefferson, D. A., T h o m a s , J. M „ Smith, D. J., Camps, R. A., Catto, C. J. D., Cleaver, J. R. A . : N a t u r e 281, 51 (1979) 68. Horiuchi, S.: Chemica Scripta 14, 75 (1978) 69. Crewe, A. V.: Chemica Scripta 14, 17 (1978-79) 70. Doig, P., Lonsdale, D., Flewitt, P. E. J . : Phil. Mag. A41, 761 (1980) 71. Oppalzer, H., Knauer, U . : In: Scanning electron microscopy 1, 1979 (Johari, O., ed.). Chicago: SEM Inc. 1979, p. I l l 72. S c h u m a n , H „ Somlyo, A. P.: Proc. N a t . Acad. Sei. U S A 73, 1193 (1976) 73. Dexpert, H . : private communication 74. A d a m s o n - S h a r p e , K. M., Ottensmeyer, F. P.: J. Microscopy 122, 309 (1981) 75. H o m e , R. W., M a r k h a m , R . : In: Electron diffraction and optical diffraction techniques. (Glauert, A. M., ed.). A m s t e r d a m , L o n d o n : N o r t h - H o l l a n d Publishing Co. 1973, p. 327 76. Salmon, E. D „ Rosier, D . : J. Microscopy 123, 239 (1981) 77. Agar, A. W „ Keown, S. R . : Proc. Roy. Microscopical Soc. G . B. 13, 147 (1978) 78. Erickson, H. P.: In: Advances in optical and electron microscopy. (Barrer, R., Cosslett, V. E.. eds.). L o n d o n , New Y o r k : Academic Press 5, p. 163 (1973)

The Electron Microscopy of Catalysts

151

79. Lipson, H., Taylor, C. A . : Fourier transforms and X-ray diffraction. L o n d o n : G. Bell and Sons 1958 80. Iijima, S., Anstis, G. R . : Proc. Ann. Meeting Electron Microscopical Soc. Amer. 34th, (Bailey, G. W „ ed.). Florida 1976, p. 476 81. Klug, A., De Rosier, D. J.: N a t u r e 212, 29 (1966) 82. Erickson, H. P., Klug, A . : Phil. Trans. Roy. Soc. 261, 105 (1971) 83. Klug, A . : Chemica Scripta 14, 245 (1979) 84. Cross, P. M.: In: The use of the scanning electron microscope. (Hearl, J. W. S., Sparrow, J. T., Cross, P. M., eds.). O x f o r d , P e r g a m o n Press 1972, p. 75 85. Lyman, C. E.: Microbeam Analysis 16, 261 (1981) 86. Kay, D. H . : Techniques for electron microscopy. O x f o r d : Blackwell Scientific Publications 1965 87. G o o d h e w , P. J.: In: Practical m e t h o d s in electron microscopy. (Glauert, A. M., ed.). A m s t e r d a m : N o r t h - H o l l a n d 1973 88. Baird, T . : Catalysis 5, 179 (1982) 89. Freel, J.: J. Catal. 25, 139 (1972) 90. F u k a m i , A., Adachi, K., K a t o h , M.: J. Electron Microscopy (Japan) 21, 99 (1972) 91. Hoelke, C. W . : Micron 5, 307 (1975) 92. Ohtsuki, M „ Isaacson, M. S., Crewe, A. V.: In: Scanning electron microscopy. (Johari, O., ed.). Chicago: SEM Inc. 1979 93. Muller, M „ Koller, T . : Optik 35, 287 (1972) 94. M i h a m a , K., T a n a k a , N . : J. Electron Microscopy (Japan) 25, 65 (1976) 95. Iijima, S.: Micron 8, 41 (1977) 96. Baumeister, W „ H a h n , M.: Micron 7, 371 (1976) 97. Glauert, A. M . : Fixation, dehydration and embedding of biological specimens. Amsterd a m : N o r t h - H o l l a n d 1975 98. Dalmai-Imelik, G., Leclerq, C., Martin, I.: J. Microscopie 20, 123 (1974) 99. Sprys, J. W., Bartosiewicz, L., M c C u n e , R „ Plummer, H . K . : J. Catal. 39, 91 (1975) 100. Baker, R. T. K., Prestridge, E. B„ Garten, R. L.: J. Catal. 56, 390 (1979) 101. Glassi, H „ K r a m e r , R., Hayek, K . : J. Catal. 63, 167(1980) 102. Chu, Y. F., Ruckenstein, E.: J. Catal. 55, 281 (1978) 103. Chen, M „ Schmidt, L. D . : J. Catal. 55, 348 (1978) 104. Hamilton, J. F . : In: G r o w t h and properties of metallic clusters. (Burdon, J., ed.). Holland: Elsevier Scientific Publishing Co. 1980, p. 289 105. Tighe, N. J.: In: Electron microscopy in mineralogy. (Wenke, H. R., ed.). Berlin: Springer Verlag 1976, p. 144 106. Bursill, L. A., Lodge, Eliz. A., T h o m a s , J. M . : N a t u r e 286 111 (1980) 107. Bursill, L. A., T h o m a s , J. M „ R a o , K. J.: N a t u r e 289, 157 (1981) 108. Hirsch, E. H . : N a t u r e 293, 759 (1981) 109. Smith, A. E., Eyring, LeR. : Ultramicroscopy 8, 65 (1982) 110. Cowley, J. M . : Ultramicroscopy 8, 1 (1982) 111. Cowley, J. M . : Appi. Phys. Letters 15, 58 (1969) 112. Dekker, A. J . : Solid State Physics 6, 251 (1958) 113. Lane, G. S. : In: The use of scanning electron microscopy. (Hearle, J. W. S., Sparrow, J. T., Cross, P. M., eds.). O x f o r d : Pergamon Press 1972, p. 219 114. Wells, O. C.: In: Scanning electron microscopy. (Wells, O. C., ed.). New Y o r k : McGraw-Hill 1974, p. 129 115. Pawley, J. B.: In: Proc. Scanning Electron Microscopy Symposium. (Johari, O., ed.). Chicago: I.I.T.R.I. 1972, p. 153 116. Pfferkorn, G . E., G r u n t e r , H., Pfautsch, M . : In: Proc. Scanning Electron Microscopy Symposium. (Johari, O., ed.). Chicago: I.I.T.R.I. 1972, p. 147 117. Broers, A. N . : Ultramicroscopy 8, 137 (1982) 118. Robinson, V. N. E.: J. Phys. E.: Sei. Instrum. 7, 650 (1974) 119. George, E. P., Robinson, V. N. E.: In: Scanning Electron Microscopy 1977. (Johari, O., 1 ed.). Chicago: I.I.T.R.I. 1977, p. 63 120. Wells, O. C., Broers, A. N „ Bremer, C. G . : Appi. Phys. Letters 23, 353 (1973)

152

Chapter 2: J. V. Sanders

121. Scanning electron microscopy and X-ray microanalysis. (Goldstein, J. I., Newbury, D. E., Echlin, P., Joy, D. C., Fiori, C„ Lifshin, E„ eds.). New Y o r k : Plenum 1981 122. Stephani, D.: J. Vac. Sci. Technol. 16, 1739 (1979) 123. M o r t o n , R. R. A . : In: Stereology and Quantitative Metallography. P a . : A.S.T.M.504, 1972, p. 81 124. Shaw, M. P., Köhler, V. L„ Ralph, B.: J. Microscopy 128, 251 (1982) 125. Humphries, C. J . : Rep. Prog. Phys. 42, 1825 (1979) 126. Lynch, D. F., O'Keefe, M. A.: Acta Cryst. A28, 536 (1972) 127. Ishizuka, K., Uyeda, N . : Acta Cryst. A33, 740 (1977) 128. Self, P. G., O'Keefe, M. A., Buseck, P. R „ Spargo, A. E. C.: Ultramicroscopy 11, 35 (1983) 129. Allpress, J. G „ Sanders, J. V.: J. Appi. Cryst. 6, 165 (1973) 130. O ' K e e f e , M. A., Fryer, J. R., Smith, D. J.: Acta Cryst. A39, 839 (1983) 131. Van Dyck, D . : J. Microscopy 132, 31 (1983) 132. O ' K e e f e , M. A.: Acta Cryst. A29, 389 (1973) 133. Metherell, A. J. F.: In: Electron microscopy for materials science. (Valdré, U., Ruedl, E., eds.). Brussels: Commissioner for European Communities 1975, p. 496 134. Y a c a m a n , M. J., O c a n a . T.: phys. stai. sol. 42, 571 (1972) 135. H e a d , A. K., Humble, P., Clareborough, L. M „ M o r t o n , A. J., F o r e w a r d , C. T . : C o m p u t e d electron micrographs and defect identification. A m s t e r d a m : N o r t h - H o l l a n d 1973 136. H e i n e m a n n , K., P o p p a , H . : Appi. Phys. Lett. 20, 122 (1972) 137. Freeman, L. A., Howie, A., Treacy, M. M. J.: J. Microscopy 111, 165 (1977) 138. K r a k o w , W „ H o w l a n d , L. A.: Ultramicroscopy 2, 53 (1976) 139. Cowley, J. M . : In: Proc. Ann. Meeting Electron Microscopical Soc. Amer. 31st, (Arceneaux, C. J., ed.). New Orleans, La 1973 140. Cowley, J. M.: Acta Cryst. A29, 529 and 537 (1973) 141. Rudec, M. L„ Howie, A . : Phil. Mag. 25, 1001 (4972) 142. Berry, M. V., Doyle, P. A . : J. Phys. C.: Solid State Phys. 6, L6 (1973) 143. Iijima, S., O'Keefe, M. A . : In: Proc. A n n . Meeting Electron Microscopical Soc. A m e r . 35th, (Bailey, G. W., ed.). Boston, Mass. 1977, p. 190 144. Cowley, J. M.. Massover, W. H „ J a p , Bing. K . : Optik 40, 42 (1974) 145. Y a c a m a n . M. J., Dominguez, J. M.: J. Catal. 67, 475 (1981) 146. Cowley, J. M „ Moodie, A. F . : Proc. Phys. Soc. 76, 378 (1960) 147. Olsen, A., Spence, J. C. H . : Phil. Mag. A43, 945 (1981) 148. O ' K e e f e , M. A., Sanders, J. V.: Optik 46, 421 (1976) 149. G o o d m a n , P., Moodie, A. F . : Acta Cryst. A30, 280 (1974) 150. Self, P. G., O ' K e e f e , M. A., Busek, P. R., Spargo, A. E. C. : Ultramicrobcopy 11, 35 (1983) 151. Scherzer, O . : J. Appi. Phys. 20, 20 (1949) 152. Cowley, J. M „ Fields, P. M . : Acta Cryst. A35, 28 (1979) 153. Cowley, J. M., Iijima, S.: In: Electron microscopy in mineralogy. (Wenk, H.-R., ed.). Berlin, Heidelberg, New Y o r k : Springer-Verlag 1976, p. 123 154. Allpress, J. G., Hewat, E. A., Moodie, A. F., Sanders, J. V.: Acta Cryst. A28, 528 (1972) 155. Lynch, D. F., O'Keefe, M. A . : Acta Cryst. A28, 536 (1972) 156. Ishizuka, K „ Uyeda, N . : Acta Cryst. A33, 740(1977) 157. H a s h i m o t o , H „ Tokai, Y., Y o k o t a , Y „ E n d o h , H., F u k u d a , E.: Jap. J. Appi. Phys. 19, LI (1980) 158. Hiraga, K., Shindo, D., Hirayabayashi, M., Terasaki, O., W a t a n a b e , D . : Acta Cryst. B36, 2550(1980) 159. Marks, L. D . : Surface Sci. 139, 281 (1984) 160. Cowley, J. M . : Appi. Phys. Letters 15, 58 (1969) 161. Crewe, A. V., Wall, J.: Optik 30, 461 (1970) 162. Cowley, J. M „ Spence, J. C. H . : Ultramicroscopy 6, 359 (1981) 163. Cowley, J.: Ultramicroscopy 7, 181 (1981) 164. Z h u , J., Cowley, J. M . : Acta Cryst. A38. 718 (1982)

T h e E l e c t r o n M i c r o s c o p y of C a t a l y s t s 165. 166. 167. 168. 169. 170.

171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 198. 199. 200. 201. 202. 203. 204.

153

R o v , R. A . , M e s s i e r , R „ C o w l e y , J. M . : T h i n Solid F i l m s 79, 207 (1981) Cliff, G . , L o r i m e r , G . W . : J. M i c r o s c o p y 103, 203 (1975) D e l a n n a y , F . : C a t a l . Rev.-Sci. E n g . 22, 141 (1980) C h a m p n e s s , P. E., Cliff, G „ L o r i m e r , G . W . : U l t r a m i c r o s c o p y 8, 121 (1982) G o l d s t e i n , J. I., C o s t l e y , J. L., L o r i m e r , G . W . , R e e d , S. J. B.: I n : S c a n n i n g e l e c t r o n m i c r o s c o p y . ( J o h a r i , O . , ed.). C h i c a g o : 1.1.T.R.I. 1977, p. 315 D e x p e r t , H . , F r e u n d , E., L y n c h , J . P., P e n n y c o o k , S. J . : I n : E l e c t r o n M i c r o s c o p y a n d A n a l y s i s 1981. Inst. P h y s . C o n f . Ser. 61. ( G o r i n g e , M . J., ed.). Bristol a n d L o n d o n : I n s t i t u t e of Physics 1981, p. 209 O t t e n s m e y e r , F. P., A n d r e w , J. W . , B a z e t t - J o n e s , D . P., C h a n , A . S. K., H e w i t t , J . : J. M i c r o s c o p y 109, 259 (1977) A e b i , U . , F o w l e r , W . E „ S m i t h , P. R . : U l t r a m i c r o s c o p y 8, 191 (1982) H a s h i m o t o , H . , E n d o h , H . , T a k a i , Y., T o m i o k a , H . , Y o k o t a , Y . : C h e m i c a S c r i p t a 14, 25 (1978-79) K l u g , A . : C h e m i c a S c r i p t a 14, 245 ( 1 9 7 8 - 7 9 ) K i r k l a n d , E . J., Siegel, B. M . , U y e d a , N . , F u j i y o s h i , Y . : U l t r a m i c r o s c o p y 5, 4 7 9 (1980) K a n a y a , K., B a b a , N . , T a k a m i y a , K . : M i c r o n 12, 353 (1981) Boyes, E. D . , M u g g r i d g e , B. J., G o r i n g e , M . J . : J. M i c r o s c o p y 127, 321 (1982) H a w k e s , P. W . : ed. I m a g e p r o c e s s i n g a n d c o m p u t e r - a i d e d design in e l e c t r o n o p t i c s . A c a d e m i c Press 1973 Misell, D . L . : I m a g e analysis, e n h a n c e m e n t a n d i n t e r p r e t a t i o n . A m s t e r d a m : N o r t h - H o l land 1978 S a x t o n , W . O . : J. M i c r o s c . S p e c t r o s c o p . E l e c t r o n 5, 665 (1980) H o w i e , A . , M a r k s , L. D., P e n n y c o o k , S. J . : U l t r a m i c r o s c o p y 8, 163 (1982) C r e w e , A . V., W a l l , J., L a n g m o r e , J. P . : Science 168, 1338 (1970) H a s h i m o t o , H . , K u m a o , A . , H i ñ o , K . , Y o t s u m o t o , H . , O n o , A . : J a p a n J. A p p l . P h y s . 10, 1115(1971) F o r m a n e k , H „ M u l l e r , M . , H a h n , M . H . , K o l l e r , T h : N a t u r w i s s e n s c h a f t e n 58, 339 (1971) E i s e n h a n d l e r , C. B., Siegel, B. M . : J. A p p l . P h y s . 37, 1613 (1966) H a l l , C. R „ H i n e s , R. L . : Phil. M a g . 21, 1175 (1970) H a s h i m o t o , H . , K u m a o , A . , H i ñ o , K., E n d o h , H . , Y o t s u m o t o , H . , O n o , A . : J. E l e c t r o n M i c r o s c o p y 22, 123 (1973) I i j i m a , S . : O p t i k 48, 193 (1977) M i h a m a , K., S h i m a , S., U y e d a , R . : I n : E l e c t r o n M i c r o s c o p y 1974. ( S a n d e r s , J. V., G o o d c h i l d , D . J., eds.). C a n b e r r a : A u s t r a l i a n A c a d e m y of Science 1974, p. 262 F r e e m a n , L. A . , H o w i e , A., T r e a c y , M . M . J . : J. M i c r o s c o p y 111, 165 (1977) T r e a c y , M . M. J., H o w i e , A., W i l s o n , C. J . : Phil. M a g . 38, 569 (1978) C r e w e , A . L . : C h e m i c a Scripta 14, 17 ( 1 9 7 8 - 7 9 ) H a i n f e l d , J. F., W a l l , J. S.: I n : S c a n n i n g E l e c t r o n M i c r o s c o p y 1980. ( J o h a r i , O., ed.). C h i c a g o : S E M Inc. 1980, p. 107 U l a n t , M . : I n : E M S A 79. (Bailey, G . W . , ed.). B a t o n R o u g e : C l a i t o n s P u b l . Div. 1979, p. 470 W a l l , J. S „ H a i n f e l d , J. F., B a r t l e t t , P. A . , Singer, S. J . : U l t r a m i c r o s c o p y 8, 397 (1982) L a n g m o r e , J. P . : I n : Principles a n d t e c h n i q u e s of e l e c t r o n m i c r o s c o p y . ( H a y a t , M . A., ed.). N e w Y o r k , V a n N o s t r a n d R e i n h o l d C o . 1977 F l y n n e , P. C „ W a n k e , S. E „ T u r n e r , P. S.: J. C a t a l . 33, 233 (1974) B a r r i e , J. C., Bursill, L. A . : I n : S e v e n t h A u s t r a l i a n C o n f e r e n c e o n E l e c t r o n M i c r o s c o p y . C a n b e r r a : C a n b e r r a P u b l . a n d P r i n t . C o . 1982, p. 135 P r e s t r i d g e , E. B „ Y a t e s , D. J. C . : N a t u r e 234, 345 (1971) Y a c a m a n , M . J . : I n : G r o w t h a n d p r o p e r t i e s of m e t a l l i c clusters. ( B u r d e n , J., ed.). T h e N e t h e r l a n d s : Elsevier Sei. P u b l . C o . 1980, p . 2 2 5 S m i t h , D . J., M a r k s , L. D . : Phil. M a g . A44, 735 (1981) M i l l w a r d . G . R . : J. C a t a l . 64, 381 (1980) N e w c o m b , S. B., S t o b b s , W. B.. Little, J. A . : p h y s . stat. sol. A76, 191 (1983) N a g a t a , F., M a t s u d a , T.. K o m o d a , T . : J a p a n J. A p p l . P h y s . 14, 1815 ( 1 9 7 5 ) -

154

C h a p t e r 2: J. V. Sanders

205. 206. 207. 208. 209. 210. 211. 212.

Jaeger, H. J.: unpublished results (1983) Schwanke, J., Allard, L. F „ Deeba, M „ Gates, B. C.: J. Catal. 84, 27 (1983) Millward, R. G . ; Jefferson, D. A., T h o m a s , J. M : J. Microscopy 113, 1 (1978) Jefferson, D. A., Millward, G. R., T h o m a s , J. M.: Acta Cryst. A32, 823 (1976) Lynch, D. F „ Sanders, J. V.: unpublished results (1984) Von Borries, B„ Kausche, G. A.: Kolloid Z. 90, 132 (1940) Y a c a m a n , M. J., O c a n a , Z. T.: phys. stat. sol. (a) 42, 571 (1977) Goringe, M. J . : In: Electron microscopy for materials science. (Valdré, U., Ruedl, E., eds.). Commission for E u r o p e a n Communities 1975 Clareborough, L. M „ Humble, P., Loretto, M. H . : Acta Cryst. 15, 1007(1967) Hayashi, T „ O h n o , T „ Yatsuya, S., Uyeda, R.: J a p . J. Appl. Phys. 16, 705 (1977) Allpress, J. G., Sanders, J. V.: Aust. J. Phys. 23, 23 (1970) Allpress, J. G „ Sanders, J. V.: Surface Sci. 7, 1 (1967) Ogawa, S„ Ino, S.: J. Crystal G r o w t h 13/14, 48 (1972) Ogawa, S., Ino, S.: In: Advances in epitaxy and endotaxy. (Schneider, H. G., R u t h , V., eds.). Leipzig 1971 Uyeda, R.: J. Crystal G r o w t h 24/25, 69 (1974) Avery, N. R „ Sanders, J. V.: J. Catal. 18, 129 (1970) Roy, R. A., Messier, R „ Cowley, J. M . : Thin Solid Films 79, 207 (1981) Larroque, P., Brieu, M . : Acta Cryst. A34, 853 (1978) K i m o t o , K . : N i p p o n Kessho Seicho Gakkaishi 6, 88 (1979) Solliard, C . : Surface Sci. 106, 58 (1981) Smith, D. J., White, D „ Baird, T „ Fryer, J. R.: J. Catal. 81, 107 (1983) Foger, K . : Catalysis Sci. Tech. 6, 227 (1984) Burton, J. J . : Cat. Rev.-Sci. Eng. 9, 209 (1974) Gillet, M.: Surface Sci. 67, 139 (1977) K o , E. I., Madix, R. J . : Adv. Mech. Phys. Surf. 1, 153 (1981) Cherns, D . : Phil. Mag. 30, 549 (1974) K r a k o w , W., Ast, D. G . : Surface Sci. 58, 485 (1976) Hiñes, R. L.: Thin Solid Films 35, 229 (1976) Takayanagi, K . : Ultrastructure 8, 147 (1982) K l a u r a , M „ Bethge, H . : Ultramicroscopy 11, 125 (1983) Cullis, A. G., Maher, D. M . : Ultramicroscopy 1, 97 (1975) G o m e z , A., Harnández, P., José-Yacamán, M . : Surface and Interface Analysis 4, 120 (1982) Lynch, D . F., M o o d i e , A. F., O'Keefe, M . A . : I n : Electron Microscopy 1974. (Sanders, J. V., Goodchild, D. J. eds.). C a n b e r r a : Australian Academy of Science 1974, p. 222 Cowley, J. M., Moodie, A. F . : Proc. Phys. Soc. 76, 378 (1960) Moodie, A. F., Warble, C. E.: In: Sintering and Catalysis, Materials Science Research 10. (Kucznski, G. C., ed.). New Y o r k : Plenum 1975, p. 1 Iijima, S.: Optik 47, 437 (1977) Allpress, J. G., Sanders, J. V.: J. Appl. Cryst. 6, 165 (1973) Moodie, A. F „ Warble, C. E.: J. Crystal G r o w t h 10, 26 (1971) Marks, L. D „ Smith, D. J.: N a t u r e 303, 316 (1983) Marks, L. D., Smith, D. J.: Phys. Rev. Lett. 51, 1000 (1983) Y o u n g , . R . T „ Wong, C.: J. Catal. 85, 154 (1984) Hojlund Neilsen, P. E„ Cowley, J. M . : Surface Sci. 54, 340 (1976) Osakabe, N., Tanishiro, Y., Yagi, K., H o n j o , G . : Surface Sci. 102, 424 (1981) S h u m a n , H . : Ultramicroscopy 2, 361 (1977) Osakabe, N., Tanishiro, Y „ Yagi, K.., H o n j o , G . : Surface Sci. 97, 393 (1980) Cowley, J. M . : In: Proc. Ann. Meeting of Electron Microscopical Soc. Amer. 39th (Bailey, G . W „ ed.). Atlanta 1981, p. 212 . Hsu, T „ Cowley, J. M . : Ultramicroscopy 11, 239 (1983) Turner, P. S„ Cowley, J. M . : Ultramicroscopy 6, 125 (1981) Ultramicroscopy 1 1 , 9 1 - 2 2 3 ( 1 9 8 3 ) Baker, R. T. K., Prestridge, E. B„ G a r t e n , R. L.: J. Catal. 56, 390 (1979) Aral, M „ Ishikawa, T., Nishiyama, Y . : J. Phys. Chem. 86, 577 (1982)

213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225. 226. 227. 228. 229. 230. 231. 232. 233. 234. 235. 236. 237. 238. 239. 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250. 251. 252. 253. 254. 255.

The Electron Microscopy of Catalysts

155

256. Ambrose, B. K . : J. Phys. E.: Sei. Inst. 9, 382 (1976) 257. Jaeger, H. J., Sanders, J. V.: unpublished information 258. Inui, T., Morinaga, N., Ishihara, T „ Kanie, T „ Takegami, Y . : J. Catal. 79, 176 (1983) 259. K n a p t o n , A. G . : Platinum Metals Review 22, 131 (1978) 260. Flytzani-Stephanopolous, M „ Wong, S„ Schmidt, L. D . : J. Catal. 49, 54 (1977) 261. Gault, A. G., Masel, R. 1.: J. Catal. 73, 294 (1982) 262. Amariglio, A., Amariglio, H . : J. Catal. 68, 86 (1981) 263. Cooper, B. J., T r i m m , D . L.: J. Catal. 62, 35 (1980) 264. Murray, R. T . : I n : Characterisation of catalysts. (Thomas, J. M., Lambert, R. M., eds.). J o h n Wiley and Sons, 1980, p. 105 265. Klein, J. C „ Hercules, D. M.: J. Catal. 82, 424 (1983) 266. Rao, V. U. S., Szirmac, A., Fisher, R. M.: J. Catal. 62, 44 (1980) 267. Sanders, J. V., Spink, J. A., Pollack, S. S.: J. Appl. Catal. 5, 65 (1983) . 268. Davis, S. M., Zaera, F., Somorjai, G. A.: J. Catal. 77, 439 (1982) 269. Whyte, T. E. Jr.: Catalysis Reviews 8, 117 (1973) 270. Pope, D „ Smith, W. L., Eastlake, M. J., Moss, R. L.: J. Catal. 22, 72 (1971) 271. Takasa, Y., A k i m u r a , T., K a s a h a r a , K., Matsuda, Y., Miura, H., T o y o s h i m a , I.: J. A m e r . Chem. Soc. 104, 5249(1982) 272. Paryjczak, T „ Syzmura, J. A . : React. Kinet. Catal. Lett. 13, 345 (1980) 273. Moss, R. L „ Pope, D „ Davis, B. J., Edwards, D. H . : J. Catal. 58, 206 (1979) 274. Cocco, G „ Schiffini, L., Strukul, G . : J. Catal. 65, 348 (1980) 275. Gallizot, P.: Catal. Rev.-Sci. Eng. 20, 121 (1979) 276. Bond, G. C., Hierl, G . : J. Catal. 61, 348 (1980) 277. Moss, R. L„ Pope, D „ Davis, B. J.: J. Catal. 61, 57 (1980) 278. Gault, F. G., Garin, F., Maire, G . : In: G r o w t h a n d properties of metallic clusters. (Bourdon, J., ed.). Netherlands: Elsevier Scientific Publ. Co. 1980 279. Naccache, C., K a u f h e r r , N., D u f a u x , M., Pandiera, J., Imelik, B.: In: Molecular sieves II. (Katz'er, J. R „ ed.). W a s h i n g t o n : A.C.S. 1977, p. 538 280. Haaland, D „ Williams, F. L.: J. Catal. 76, 450 (1982) 281. White, D., Baird, T., Fryer, J. R., Freeman, L. A., Smith D. J.: J. Catal. 81, 119 (1983) 282. Kramer, R., Zuegg, H . : J. Catal. 85, 530 (1984) 283. Cullis, C. F., Willatt, B. M . : J. Catal. 83, 267 (1983) 284. Fuentes, S., Figueras, F.: J. Catal. 61, 443 (1980) 285. T h o r n t o n , E. W., Knözinger, H „ Tesche, B„ R a f a l k o , J. J., Gates, B. G . : J. Catal. 62, 117 (1980) 286. G r a y d o n , W . F., Langan, M. D . : J. Catal. 69, 180 (1981) 287. Ellestad, O. H., Naccache, C.: In: Perspectives in catalysis. (Larsson, R., ed.). Sweden: C. W . K. Gleerup 1981, p. 95 288. McVicker, G . B„ Baker, R. T. K „ G a r t e n , R. L„ Kugler, E. L.: J. Catal. 65, 207 (1980) 289. Lee, P. C„ Meisel, D . : J. Catal. 70, 160 (1981) 290. Galvagno, S., Schwank, J., Parravano, G . : J. Catal. 61, 223 (1980) 291. Prestridge, E. B., Via, G . H., Sinfeldt, J. H . : J. Catal. 50, 115 (1977) 292. Bartholomew, C. H „ Pannell, R. B„ Fowler, R. W . : J. Catal. 79, 34(1983) 293. Smith, J. S., Thrower, P. A., Vannice, M. A.: J. Catal. 68, 270 (1981) 294. Mustard, D. G., Bartholomew, C. H . : J. Catal. 67, 186 (1981) 295. K u o , H. K., Ganesan, P., De Angelis, R. J.: J. Catal. 64, 303 (1980) 296. Scott, B. A., Plecenick, R. M „ Cargill, G. S. III, M c G u i r e , T. R „ Herd, S. R.: Inorg. Chem. 19, 1252(1980) 297. Jiang, X-Z., Hayden, T. F., Dumesic, J. A.: J. Catal. 83, 168 (1983) 298. C h u , Y. F., Ruckenstein, E.: J. Catal. 55, 28 (1978) 299. Chen, M., Schmidt, L. D . : J. Catal. 55, 348 (1978) 300. Tartachuk, B. J., Dumesic, J. A . : J. Catal. 70, 308 (1981) 301. Bartholomew, C. H „ Sorensen, W. L.: J. Catal. 81, 131 (1983) 302. Glassl, H., K r a m e r , R., Hayek, K . : J. Catal. 68, 388 (1981) 303. Harris, P. J. F., Boyes, E. D., Cairns, J. A.: J. Catal. 82, 127 (1983) 304. Baker, R. T. K „ Prestridge, E. B„ G a r t e n , R. L.: J. Catal. 56, 390 (1979)

156

C h a p t e r 2: J. V. Sanders

305. 306. 307. 308. 309. 310. 311. 312. 313.

Baker, R. T. K.. Prestridge, E. B„ G a r t e n , R. L.: J. Catal. 59, 293 (1979) Smith, D. J., White, D „ Baird, T., Fryer, J. R.: J. Catal. 81, 107 (1983) Clark, R. W „ Tien, J. K „ Wynblatt, P.: J. Catal. 61, 15 (1980) Chen, J. J., Ruckenstein, E.: J. Catal. 69, 254 (1981) W a n k e , S. E„ Flynn, P. C : Catal. Rev.-Sci. Eng. 12, 93 (1975) Schmidt, L. D., Wang, T., Vacquez, A . : Ultramicroscopy 8, 175 (1982) Foger, K., Jaeger, H.: J. Catal. 70, 53 (1981) Jaeger, H „ Foger, K.: Micron 13, 267 (1982) White, D., Baird, T., Fryer, J. R., Freeman, L. A., Smith, D. J., Day, M. : J. Catal. 81, 119 (1983) Ruckenstein, E., C h u , Y. F.: J. Catal. 59, 109 (1979) T a t a r c h u k , B. J., Chludzinski, J. J., Sherwood, R. D., Dumensic, J. A., Baker, R. T. K . : J. Catal. 70, 433 (1981) Wang, T., Schmidt, L. D . : J. Catal. 66, 301 (1980) Ruckenstein, E„ Chen, J. J.: J. Catal. 70, 233 (1981) H e i n e m a n n , K., Osaka, T., P o p p a , H . : Ultramicroscopy 12, 9 (1983) Stulga, J. E„ Wynblatt, P., Tien, J. K . : J. Catal. 62, 59 (1980) Chen, M „ Schmidt, L. D . : J. Catal. 55, 348 (1978) Baker, R. T. K.: J. Catal. 63, 523 (1980) Clark, R. W „ Wynblatt, P., Tien, J. K.: Acta Met. 30, 139 (1982) Chen, M., Wang, T., Schmidt, L. D . : J. Catal. 60, 356 (1979) W a n g , T., Schmidt, L. D . : J. Catal. 70, 187 (1981) Baker, R. T. K „ Prestridge, E. B„ G a r t e n , R. L.: J. Catal. 59, 293 (1979) Baker, R. T. K., Prestridge, E. B„ Murrell, L. L.: J. Catal. 79, 348 (1983) Derouane, E. G., Chendzinski, J. J., Baker, R. T. K . : J. Catal. 85, 187 (1984) Mériaudeau, P., Ellestad, O. H., D u f a u x , M., Naccache, C.: J. Catal. 75, 243 (1982) Y a o , H. C „ Sieg, M., Plummer, H . K. j n r . : J. Catal. 59, 365 (1979) K o o p m a n , P. G. J . ; Kieboom, A. P. G., van Bekkum, H . : J. Catal. 69, 172(1981) Wang, T., Schmidt, L. D . : J. Catal. 71, 411 (1981) Foger, K., Jaeger, H . : J. Catal. 92, 64 (1985) H o n j o , G., Yagi, K . : I n : Current Topics in Materials Science. (Kaldis, E., ed.). N o r t h Holland Publishing Co. 6, 1980, p. 197 H o n j o , G., Takayanagi, K „ Kobayashi, K., Yagi, K . : phys. stat. sol. 55, 353 (1979) Metois, J. J., Heinemann, K., P o p p a , H . : Phil. Mag. 35, 1413 (1977) Heinemann, K „ P o p p a , H . : Thin Solid Films 33, 237 (1976) Jaeger, H., Mercer, P. D „ Sherwood, R. G . : Surface Sei. 13, 349 (1969) Wasserman, E. F., Hines, R. L.: J. Appl. Phys. 38, 196 (1967) Baker, R. T. K., Thomas, C„ T h o m a s , R. B.: J. Catal. 38, 510 (1975) McVicker, G. B., G a r t e n , R. L„ Baker, R. T. K. : J. Catal. 54, 129 (1978) Gai, P. L., Goringe, M. J.: Proc. Ann. Meeting Electron Microscopical Soc. A m e r . 39th, 68 (1981) Baker, R. T. K., Harris, P. S.: In: Chemistry and Physics of C a r b o n . Vol. 14. New Y o r k : Dekker 1978, p. 83

314. 315. 316. 317. 318. 319. 320. 321. 322. 323. 324. 325. 326. 327. 328. 329. 330. 331. 332. 333. 334. 335. 336. 337. 338. 339. 340. 341. 342. 343. 344. 345. 346. 347. 348. 349. 350. 351. 352. 353.

Baker, R . T. K . : Catal. Rev.-Sci. Eng. 19, 161 (1979) Keep, C. W „ Wells, M.: J. Catal. 66, 451 (1980) Baker, R. T. K „ Sherwood, R. D . : J. Catal. 70, 198 (1981) Baker, R. T. K „ Sherwood, R. D „ Derouane, E. G . : J. Catal. 75, 382 (1982) Simoens, A. J., Derouane, E. G., Baker, R. T. K . : J. Catal. 75, 175 (1982) Derouane, E. G., Baker, R. T. K., Dumesic, J. A., Sherwood, R. D . : J. Catal. 69, 101 (1981) Baker, R. T. K „ Sherwood, R. D „ Dumesic, J. A.: J. Catal. 62, 221 (1980) Baker, R. T. K „ Sherwood, R. D., Dumesic, J. A.: J. Catal. 66, 56 (1980) Baker, R. T. K.., Sherwood, R. D . : J. Catal. 61, 378 (1980) Coates, D. J., Evans, J. W., Cabrera. A. L„ Somerjai, G. J., Heineman, H. : J. Catal. 80. 215 (1983) Yang, R. T., Wong, C.: J. Chem. Phys. 75, 4471 (1981)

The Electron Microscopy of Catalyst-,

157

354. Takayanagi, K... Yagi, K „ Kobayashi, K.., H o n j o , G . : J. Phys. E.: Sei. Inst. U , 441 (1978) 355. Howie, A . : J. Microscopy 129, 239 (1983) 356. Smith. D. J., Stobbs, W. M „ Saxtcfn, W. O . : Phil. Mag. B43, 907 (1981) 357. Bursill, L. A., Mallinson, L. G „ Elliot, S. R „ T h o m a s , J. M . : J. Phys. Chem. 85, 3004 (1981) 358. Rudec, M. L.: phys. stat. sol. B46, K1 (1971) 359. Rudec, M. L„ Howie, A.: Phil. Mag. 25, 1001 (1972) 360. Berry, M . V., Doyle, P. A . : J. Phys. C. Solid State Phys. 6, L6 (1973) 361. Oberlin, A . : C a r b o n 17, 7 (1979) 362. Fryer, J. R.: C a r b o n 19,431 (1981) 363. Sanders, J. V., Spink, J. A., Pollack, S. S.: Applied Catal. 5, 65 (1983) 364. Haldeman, R. G „ Botty. M. C.: J. Phys. Chem. 63, 489 (1959) 365. Marsh, H „ C r a w f o r d , D „ O ' G r a d y , T . M „ Wennerberg, A . : C a r b o n 20, 419 (1982) 366. Iijima, S.: Chemica Scripta 14, 117 (1978-79) 367. Ban, L. L„ C r a w f o r d , D „ Marsh, H . : J. Appl. Cryst. 8, 415 (1975) 368. Rouzaud, J. N „ Oberlin, A., Beny-Bassez, C.: Thin Solid Films 105, 75 (1983) 369. Jefferson, D. A., Millward, G . R., T h o m a s , J. M . : Acta Cryst. A32, 823 (1976) 370. Thomas, J. M „ Millward, G. R., Schlögl, R. F., Boehm, H. P.: Mater. Res. Bull. 15, 671 (1980) 371. Beer, M., Carpenter, R. W., Eyring, LeR., L y m a n , C. E., Thomas, J. M.: Chem. & Eng. News 59, 40(1981) 372. Millward, G. R., Jefferson, D. A., Thomas, J. M . : J. Microscopy 113, 1 (1978) 373. Smith, P. P. K., Buseck, P. R.: Science 212, 322 (1981) 374. Heimann, R. B„ Kleiman, J., Salanzky, N. M . : N a t u r e 306, 164 (1983) 375. Hayatsu, R., Scott, R. G „ Studier, M. H „ Lewis, R. S„ Anders, E . : Science 209, 1515 (1980) 376. Whittaker, A. G., Watts. E. J., Lewis, R. S., Anders, E.: Science 209, 1512 (1980) 377. Whittaker, A. G . : C a r b o n 17, 21 (1979) 378. Smith, P. P. K., Buseck, P. R.: Science 212, 322 (1981) 379. Smith, P. P. K „ Buseck, P. R.: Science 216, 984 (1982) 380. Davis. W. R „ Slawson, R. J., Rigby, G. R.: N a t u r e 171, 756 (1953) 381. Baird, T „ Fryer, J. R „ G r a n t , B.: Carbon 12, 591 (1974) 382. Audier, M...Oberlin, A.. Coulon, M.: J. Crystal G r o w t h 55. 549 (1981) 383. Baker, R. T. K „ Harris, P. S.. Thomas. B„ Waite, R. J . : J. Catal. 30, 86 (1973) 384. Ruston, W. R „ Warzec, M „ H e n n a n t , J., Waty, J.: C a r b o n 7, 47 (1969) 385. Baker, R. T. K.. Alonzo, J. R.. Dumesic, J. A., Yates, D. J. C.: J. Catal. 77, 74 (1982) 386. Baker, R. T. K..: Cat. Rev.-Sci Eng. 19, 161 (1979) 387. Baker, R. T. K „ Chludzinski, jnr., J. J . : J. Catal. 64, 464 (1980) 388. Jackson, S. D „ T h o m p s o n . S. J., Webb, G . : J. Catal. 70, 249 (1981) 389. Kawasumi, S„ Egashira, M „ Katsuki, H.: J. Catal. 68, 237 (1981) 390. Audier, M., Oberlin, A., Oberlin, M., Coulon, M.. Bonnetain, L.: C a r b o n 19, 217 (1981) 391. 392. 393. 394. 395. 396. 397. 398. 399. 400. 401. 402.

Sacco, j n r . A., Thacker, P., Chang, T. N „ Chiang, A. T. S.: J. Catal. 85, 224 (1984) H a r v a r d , D. C„ Wilson, R.: J. Colloid & Interface Sei. 57, 276 (1976) Gaskell, P. H „ Mistry, A. B.: Phil. Mag. A39. 245 (1979) Wilson. S. J.: J. Solid State Chem. 30, 247 (1979) Wilson, S. J.: Mineral. Mag. 43, 301 (1979) Spitler, C. A., Pollack, S. S.: J. Catal. 69, 241 (1981) Lippens, B. C„ de Boer, J. H . : Acta Cryst. 17, 1312 (1964) Soled, S.: J. Catal. 81, 252 (1983) Dominguez. E. J. M „ Acosta, N. D. R„ Schifter, S. I.: J. Catal. 83, 480 (1983) Bursill, L. A.. Lodge, E. A., T h o m a s , J. M.: Nature 286, 111 (1980) Bursill, L. A., Thomas, J. M . : J. Phys. Chem. 85, 3007 (1981) Thomas, J. M., Ramdas, S., Millward, G. R., Klinowski, J., Audier, M., GonzalezColbet, J.. Fyfe, C. A . : J. Solid State Chem. 45, 368 (1982)

158

Chapter 2: J. V. Sanders

403. Audier, M , Thomas, J. M., Klinkowski, J.. Jefferson, D. A., Bursill, L. A.: J. Phys. Chem. 86, 581 (1982) 404. Thomas, J. M., Millward, G. R . : J. Chem. Soc. Chem. C o m m u n . 24, 1380 (1982) 405. Millward, G. R.. Ramdas, M„ Thomas, J. M., Barlow, M. T.: J. Chem. Soc. Faraday Trans. II 79, 1075 (1983) 406. Gallezot, P., Alarcon-Diaz, A., Dalmon, J. A., Renouprez, A. J., Imelik, B.: J. Catal. 39, 334(1975) 407. Beshoukhanova, C., Guidot, J., Barthomeuf, D.: J. Chem. Soc. Faraday Trans. I 77, 1595 (1981) 408. Schmidt, H.: In: Metal microstructures in zeolites. (Jacobs, P. A., Jaeger, N. I., Jiru, P., Schulz-Eklott, G. eds.). Amsterdam: Elsevier Scientific Publ. Co. 1982, p. 191 409. Exner, D., Jaeger, N „ Nowak, R., Schriibers, H., Schulz-Eklott, G . : J. Catal. 74, 188 (1982) 410. Gallezot, P.: Catal. Rev.-Sci. Eng. 20, 121 (1979) 411. Terasaki, O., Thomas, J. M „ Ramdas, S.: Chem. C o m m u n . 216 (1984) 412. Sharma, R. K „ Rai, K. N „ Srivastava, R. D.: J. Catal. 63, 271 (1980) 413. Smith, D. J., Fisher, R. M „ Freeman, L. A.: J. Catal. 72, 51 (1981) 414. Candia, R „ Clausen, B. S„ Tops0e, H . : J. Catal. 77, 564 (1982) 415. Sanders, J. V., Pratt, K. C.: J. Catal. 67, 331 (1981) 416. Pratt, K. C„ Sanders, J. V.: J. Catal. 66, 82 (1980) 417. Duchet, J. C„ van Oers, E. M„ de Beer, V. H. J.. Prins, R.: J. Catal. 80, 386 (1983) 418. Figueras, F., Figlarz, M., Portefaix, J. L., Forissier, M., Gerand, B., Guenot, J . : J. Catal. 71,389 (1981) 419. Barber, S., Booth, J., Pyke, D. R „ Reid, R „ Tilley, R. J. D . : J. Catal. 77, 180 (1982) 420. De Rossi, S., Iguchi, E„ Schiavello, M „ Tilley, R. J. D.: J. Catal. 61, 256 (1980) 421. Gai, P. L„ Boyes, E. D „ Bart, J. C. J.: Phil. Mag. A45, 531 (1982) 422. Cowley, J. M „ Wheatley, J. C „ Kehl, W. L.: J. Catal. 56, 185 (1979) 423. Somorjai, G. A.: Catal. Rev. 7, 87 (1972) 424. Hajlund, P. E„ Cowley, J. M.: Surface Sei. 54, 340 (1976) 425. Oäakabe, N., Tanishiro, Y.. agi, K „ H o n j o , G.: Surface Sei. 97, 393 (1980) 426. Osakabe, N., Tanishiro, Y „ Yagi, K „ Honjo, G.: Surface Sei. 102, 424 (1981) 427. Cowley, J. M.: Surface Sei. 114, 587 (1982) 428. Moodie, A. F., Warble, C. E.: J. Crystal Growth 10, 26 (1971) 429. Iijima, S.: Optik 47, 437 (1977) 430. Marks, L. D „ Smith, D. J.: Nature 303. 316 (1983) 431. Bonzel, H. P., Ferrer, S.: Surface Sei. 118, L263 (1982) 432. Cowley. J. M.: J. Microscopy 129, 253 (1983) 433. Pennycook, S. J.: J. Microscopy 124, 15 (1981) 434. Dexpert, H., Freund, E., Lesage, E., Lynch, J. P.: Studies in Surface Science in Catalysis 11, 53 (1982) 435. Blanchard, G., Charcosset, H., Dexpert, H., Freund, E., Leclercq, C., Martino, G.: J. Catal. 70, 168 (1981) 436. Houalla, M., Delannay, F., Delmon, B.: J. Chem. Soc. Faraday Trans. I, 76, 1766 (1980) 437. Auroux, A., Dexpert, H., Leclercq, C„ Vedrine, J.: Applied Catal. 6, 95 (1983) 438. Turkevich, J., Ban, L. L., Wall, J. H.: In: Perspectives in catalysis. (Larsson, R., ed.). Sweden: C W K Gleerup 1981, p. 59 439. Cowley. J. M.: private communication (1984)

Chapter 3

Surface Structural Chemistry B. E. Koel1 and G. A.

Somorjai2

1 Cooperative Institute for Research in Environmental Sciences and Department of Chemistry, University of Colorado Boulder, Colorado 80309, USA 2 Materials and Molecular Research Division . Lawrence Berkeley Laboratory, Department of Chemistry University of California, Berkeley, California 94720, USA

Contents 1. Introduction: The Atomic and Electronic Structure of Surfaces 2. The Low Energy Electron Diffraction Technique for Atomic and Molecular Surface Structure Determination 3. High Resolution Electron Energy Loss Spectroscopy (HREELS) as a Probe of Surface Structure 4. Additional Techniques for the Determination of Surface Structure 5. Structure of Solid Surfaces A. The Atomic Structure of Clean Surfaces 1. The Atomic Structure of Unreconstructed Low Miller Index Planes of Transition Metal Surfaces 2. The Atomic Structure of High Miller Index Stepped and Kinked Surfaces . . . B. The Structure of Adsorbed Atoms on Solid Surfaces 1. Non-Metal Adsorption 2. Adsorption and Growth of Layers of Metals on Surfaces of Other Metals . . C. Surface Structure of Molecules on Solid Surfaces 1. Structure of Adsorbed CO on the Rh(l 11) Surface 2. Structure of the Adsorbed Benzene Monolayer on Rh(l 11) 3. The Temperature Dependent Character of the Surface Chemical Bond: The Adsorption and Thermal Decomposition of Alkenes on R h ( l l l ) and P t ( l l l ) Surfaces 6. Future Directions in Surface Structural Determinations 7. Implications to Catalysis: The Structure Sensitivity of the Surface Chemical Bond . . 8. References

160 161

172 177 177 180 181 182 182 185 185 187 194 201 208 209 214

1. Introduction: The Atomic and Electronic Structure of Surfaces The atomic geometry and electronic structure of surface atoms are responsible for the chemical and electronic properties of surfaces. The atomic and electronic structures are rarely separable. However, the experimental techniques used to study these two structural features are different, and therefore they are often separately investigated and discussed. In this review we shall give an overview of what we know about the atomic structure of surfaces and adsorbed monolayers, the methods used to obtain this information, and point

160

Chapter 3 : B. E. Koel, G . A. Somorjai

out the importance and relation o f surface structure to chemical bonding and heterogeneous catalysis. Over the past 15 years, there has been a major revolution in the field o f surface chemistry that has permitted the atomic scale scrutiny o f surface monolayers. The low energy electron diffraction ( L E E D ) technique was developed which enables one to determine the location o f ordered layers o f surface atoms and of molecules adsorbed on surfaces. High resolution electron energy loss spectroscopy ( H R E E L S ) which was also developed over the past ten years can yield vibrational spectra o f adsorbed atoms and molecules on surfaces. These two techniques have been used most extensively for studies o f the surface structure o f single crystal substrates and adsorbed monolayers on these surfaces, and provide information about a large and rapidly increasing number of systems. We have chosen to rely mainly on these two techniques in our studies and the emphasis o f this review will primarily be on data from these two methods. However, we will also mention many other promising techniques for surface atomic structure analysis that are available or are being developed. Since detailed investigations o f surface structure have mainly used single-crystal substrates under ultrahigh vacuum ( U H V ) conditions, this review will focus on these systems. We will point out the application of results from these fundamental studies to the understanding of some of the elementary processes in heterogeneous catalysis. In the next two Sections (2 and 3), we discuss briefly the basic principles and methods o f L E E D and H R E E L S for surface structural analysis. Section 4 considers several other methods for studying surface structure. The main part of this review, Section 5, is an assessment o f our understanding o f the surface structure of clean surfaces, atoms adsorbed on surfaces, and molecules adsorbed on surfaces, as determined primarily by L E E D and H R E E L S . In Section 5.C which considers molecules adsorbed on solid surfaces, we discuss several case studies that illustrate the application of surface structural analysis. The chemisorption o f CO is discussed because of its involvement in important catalytic reactions and as a prototype o f more complex systems, clearly exhibiting various modes o f molecular bonding to surfaces and bond strength variations due to the nature o f the substrate. A L E E D and H R E E L S study o f benzene chemisorption on R h ( l l l ) illustrates the utility o f a combined techniques approach in surface studies. The important area o f hydrocarbon reactions is dealt with and elementary chemical transformations are illustrated in alkene adsorption and decomposition on transition metal surfaces. Section 6 contains some future directions in surface structural analysis that we expect to have a major impact on our understanding o f surface structural chemistry. Finally, in Section 7, we discuss the structure sensitivity o f the surface chemical bond and its implications to catalysis. In our discussion o f surface structure, we will often refer to the periodic geometry o f the substrate and of the adsorbed monolayer. The surface unit cell is the basic structural unit in the description o f the ordering o f surfaces. Often when adsorbates form ordered structures or when reconstruction o f the substrate atoms occur, the unit cells o f those structures is different from the unit cell of the substrate. When this unit cell is larger than that of the

Surface Structural Chemistry

161

substrate, the surface lattice is called a superlattice. It is necessary therefore to have a notation that allows the unique characterization of the surface or adsórbate lattice relative to the substrate lattice. Two c o m m o n notations are used: the matrix notation and the W o o d notation [1], In matrix notation, the unit cell basis vectors (a l5 a 2 ) of the substrate surface lattice are related to those of the adsórbate (b l5 b 2 ) by a matrix M :

The Matrix M uniquely characterizes the relationship between the unit cells. The Wood notation, in which the relationship between the unit cells is somewhat more transparent, can be used when the angles between the pairs of basis vectors are the same for the adsorbate and substrate, i.e., when the angle between bj and b2 is the same as the angle between and a 2 . Then the unit cell relationship is given by, in general, c or p (v x w) R a . Here v and w are the elongation factors of the basis vectors: v =

|b,| la,I

,

w =

|b2| 1*21

The angle of rotation between the lattices, i.e., the angle between ai and b,, is a. The suffix Ra is omitted when a = 0. The prefixes " c " and " p " mean "centered" and "primitive", respectively, with centered denoting the case where a lattice point is added in the center of the primitive unit cell. The prefix p is optional, and often omitted. The two notations for simple unit cells are easily related. For example, the W o o d notation for an overlayer unit cell identical to that of the substrate is p(l x 1) or (1 x 1), while in matrix . In another slightly more complicated case, the W o o d notation is c(2 x 2) = ( j / 2 x ]/2) R45° or in matrix notation

2. The Low Energy Electron Diffraction Technique for Atomic and Molecular Surface Structure Determination Most of the experiments that are aimed to determine surface structure use a single crystal surface of about 1 centimeter in diameter, placed in an ultra-high vacuum system which is equipped for a variety of surface science techniques. Foremost among them is low energy electron diffraction ( L E E D ) and Auger electron spectroscopy (AES) that determine the atomic structure and composition, respectively, of the surface layers. The surface that is to be studied has to be suitably prepared by ion sputtering and/or chemical treatments to remove surface impurities and then the surface must be

162

C h a p t e r 3: B. E. Koel. G. A. Somorjai

annealed to move atoms into their equilibrium position and to minimize surface structural damage. Figure 1 shows a typical geometry that is utilized in low energy diffraction and other single crystal surface studies. Electron spectroscopic techniques give information about surfaces due to the high inelastic scattering cross-section of electrons. A "universal curve" for the inelastic mean free path in solids shows that between 10 and 500 eV electron kinetic energy, the mean free path is of the order of 0.4 to 2.0 nm. Figure 2 shows the number of back scattered electrons as a function of their energy when a 2,000 eV electron beam strikes the surface. At 2,000 eV (Region I) there is an elastic peak due to nearly elastically scattered electrons that have lost only small amounts of energy. At higher resolution, this energy region can provide information about atomic vibrations that are in the range of 0 to 0.4 eV. Region II shows inelastically scattered electrons which have caused electronic excitations, along with bulk or surface plasmon excitations. Higher energy losses (Region III) are due to ionizing excitations of electrons and these provide information about the surface composition by identifying the atoms the electrons came from. At very low energies (Region IV), there is a large secondary electron emission background that is due to multiple inelastic scattering that often results in the emission of several electrons of lower energy upon the incidence of one electron of higher energy. In LEED, the elastically backscattered (diffracted) electron fraction (Region I in Figure 2) is used to study the structure of surfaces and adsorbates.

Figure 1. Single-crystal metal sample m o u n t e d in an ultra-high vacuum ( U H V) c h a m b e r prepared for surface studies

163

Surface Structural Chemistry

E 0 = 2000 eV

0

50

- -i-

n-

E 0 -50

Eo E—HIK

Energy/eV

Figure 2. Energy distribution of backscattered electrons. Plot is of the number of scattered electrons, N(E), as a function of their kinetic energy, E

In LEED, a collimated beam of electrons of well-defined (but variable) energy is diffracted by a crystal surface. The electrons are scattered mainly by the individual atom cores of the surface and produce wave interferences that depend strongly on the relative atomic positions of the surface under examination, because of the quantum-mechanical wave nature of electrons. The de Broglie wavelength of electrons, X, is given by the formula X (in nm) = j / l • 5 / E , where E is measured in eV. In the energy range of 10 to 500 eV, the wavelength varies from 0.39 to 0.055 nm, comparable to interatomic distances. Thus, the elastically scattered electrons can diffract to provide information about the periodic surface structure. Figure 3 shows the scheme of the L E E D experiment. A monoenergetic beam of electrons in the range of 10 to 500 eV is incident on a single crystal. Roughly 1 to 5 percent of the incoming electrons are elastically scattered. A retarding field analyzer separates this fraction, which is then post-accelerated onto a fluorescent screen where the intensity is displayed and may be photographed. If the crystal surface is well-ordered, a diffraction pattern consisting of bright, well-defined spots will be observed on the screen. The sharpness and overall intensity of the spots is related to the degree of order on the surface. When the surface is less ordered, the diffraction beams broaden and become less intense while some diffuse background intensity appears between the beams. The electron beam source commonly used has a coherence width of about 10 nm. This means that sharp diffraction features are obtained only if the regions of well-ordered atoms ("domains") are of the order of (10 nm) 2 or larger. Diffraction from smaller size domains gives rise to beam broadening and finally to the disappearance of recognizable beams from a disordered surface. The diffraction pattern from the (111) face of a platinum single crystal is shown in Figure 4. The brightness and sharp definition of the diffraction

164

Chapter 3: B. E. Koel. G. A. Somorjai

FLUORESCENT SCREEN I T " " -

TWO DIMENSIONAL CRYSTAL LATTICE (MAGNIFIED)

DIFFRACTION SPOT

Figure 3. Scheme of the LEED experiment

beams and the weak intensity of the diffuse background clearly indicates a well-ordered surface. One may distinguish between "two-dimensional" LEED and "threedimensional" LEED. In two-dimensional LEED one observes only the shape of the diffraction pattern (as seen and easily photographed on a fluorescent screen) [2, 3], The bright spots appearing in this pattern correspond to the points of the two-dimensional reciprocal lattice belonging to the repetitive crystalline surface structure, i.e., they are a (reciprocal) map of the surface periodicities. Therefore, they give information about the size and orientation of the surface unit cell; this is important information, since the presence of, for example, reconstruction-induced and overlayer-induced superlattices is made immediately visible. This information also includes the presence or absence of regular steps in the surface [4, 5], The background in the diffraction pattern contains information about the nature of any disorder present on the surface [6], As in the analogous case of X-ray crystallography, the two-dimensional LEED pattern in itself does not allow one to predict the internal geometry of the unit cell (although good guesses can sometimes be obtained); that requires an analysis of the intensities of diffraction. Nevertheless, two-dimensional LEED already can give a very good idea of essential features of the surface geometry, in addition to those mentioned before. Thus, one may follow the variation of the diffraction pattern as a function of exposure to foreign atoms: it is often possible to obtain semiquantitative values for the coverage, for the attractive and/or repulsive interactions between adsorbates [7], for some details of island formation [6], etc. The variation of the diffraction pattern with changing surface temperature also provides information about these interactions (in particular at an order/ disorder transition) [6], while the variation with electron energy is sensitive to quantities such as surface roughness perpendicular to the surface and step heights [4, 5],

Surface Structural Chemistry

165

Figure 4. L E E D pattern from a Pt( 111 ) crystal surface at 51 eV (upper left), 63.5 eV (upper right), 160 eV (lower left), and 181 eV (lower right) incident electron energy and normal incidence. With increasing energy the diffraction spots converge towards the specular reflection spot, here hidden by the sample

In three-dimensional LEED, the two-dimensional pattern is supplemented by the intensities of the diffraction spots (thereby focusing the attention on the periodic part of the surface structure, i.e., the ordered regions) to investigate the three-dimensional internal structure of the unit cell. This is most readily carried out by considering the variation of the spot intensities as a function of electron energy and/or incidence direction. The pictures in Figure 4 were taken at different incident electron energies. As the electron energies increase, the de Broglie wavelength decreases, bringing in higher order diffraction beams into the view of the fluorescent screen. If the intensity of each diffraction beam is monitored as a function of electron energy, an intensity versus electron voltage curve, or I-V curve, is obtained as shown in Figure 5. The fluctuations of the diffraction beam intensities

Chapter 3: B. E. Koel, G. A. Somorjai

166 Clean Pt (111) 6 = 4° 4>=0 T= 300 K

Figure 5. Experimental intensity versus electron energy (I-V) curves for electron diffraction from a P t ( l l l ) surface. Beams are identified by different labels (h, k) representing reciprocal lattice vectors parallel to the surface. Here the angle of incidence was 4° from the surface normal 80 120 Energy/eV

160

clearly indicate that diffraction is not two dimensional. The beam as it penetrates the surface undergoes diffraction from the successive layers, providing 3-dimensional diffraction. As a result, the structure of not only the surface layer of atoms but also the positions of atoms in the second and third layers are determined by LEED. The extreme surface sensitivity of the technique is due to the high elastic as well as inelastic scattering cross sections of the electrons as compared to x-rays. Because of the high scattering cross sections, a large fraction of incident electrons are backscattered in the first two or three layers at the surface. This surface sensitivity, of course, is exceedingly important in surface structural analysis. However, as a consequence, multiple scattering of the electrons cannot be neglected, i.e., there is a significant probability that an electron scattered once will be scattered again before exiting the surface region. Thus, the structure analysis must include multiple scattering of electrons, and in fact, this multiple scattering is very sensitive to the precise locations of atoms and molecules in the surface.

Surface Structural Chemistry

167

It is necessary to theoretically simulate the electron diffraction in order to extract the atomic positions from the experimental data. This simulation normally must include the multiple scattering o f the electrons in the surface region, resulting in so-called "dynamical" calculations [8, 9]. A suitable scattering potential, calculated from first principles, is used for this purpose. A multiple scattering calculation presupposes given atomic positions. Consequently, the simulation must be repeated for a variety of a priori plausible surface configurations. For each configuration, the theoretical diffraction intensities are then compared with the experimental data. The best agreement in this comparison occurs for the correct configuration. Refinements o f atomic positions can be carried out as desired, usually with the aid o f computed reliability factors (R-factors) that remove the subjectivity o f visual evaluation which is inevitable when many comparisons must be made.' L E E D has developed over the past ten years into a relatively well established technique for surface structure determination and has been the most productive technique used to analyze atomic positions, bond lengths and bond angles at surfaces [2, 8-12], The largest number o f results concern clean, flat (low Miller Index) single-crystal surfaces and atomic adsorbates on them. These have established the technique on a sound and reliable footing and have served as the necessary base for the more recent studies of adsorbed molecules. Overall, over 150 detailed structures have been determined with L E E D so far, of which about 10 involve molecules adsorbed at metal surfaces. In addition, hundreds of ordered L E E D patterns have been observed and used to understand the two-dimensional periodicity of solid surfaces. Still, L E E D has some limitations. A chemical identification of the surface atoms is not possible by L E E D alone. Also, for a L E E D structural analysis, it is desirable to first obtain a well-ordered arrangement of the surface. This means studies can be carried out only on single-crystal substrates. Furthermore, atomic and molecular adsorbates preferably should also give an ordered surface structure for L E E D analysis. Electron beam damage o f molecular adsorbates is currently often a problem, but new developments in the L E E D experimental method should reduce this limitation. Also, hydrogen can only be detected in unusual circumstances. Another limitation also concerns the cost of computing which can become large for certain types o f structures.

3. High Resolution Electron Energy Loss Spectroscopy (HREELS) as a Probe of Surface Structure H R E E L S has undergone an explosive development in the last ten years due to its ability to extract important structural information about atomic and molecular species adsorbed at surfaces [13, 14], and has been applied to a large number ( ~ 2 5 0 ) of adsorption systems [15]. By a suitable monochromatization of incident electrons of energy 2 to 10 eV and energy analysis of the scattered electrons, small energy losses due to vibrational excitations of surface

168

Chapter 3: B. E. Koel, G. A. Somorjai

atoms and molecules are detectable with an energy resolution of 2.5 to 10 meV (20 to 80 wavenumbers; 1 meV = 8.065 c m - 1 ) . This monochromatization and analysis is achieved by using an electrostatic deflection spectrometer, typically using 127° cylindrical or hemispherical sectors. A spectrometer used in the author's laboratory, which is similar to that used commonly [16], is shown in Figure 6. Thermal electrons from a hot tungsten filament are focussed with an Einsel lens onto the monochromator entrance slit. After exiting the monochromator, the monoenergetic electron beam is focussed on the sample by additional lenses. The sample beam current is 10~ 9 -10 - 1 1 A. The electrons that are back-reflected from the sample surface are focussed on the analyzer entrance slit and energy analyzed to produce an electron energy loss (vibrational) spectrum. A channeltron electron multiplier with pulse-counting electronics is used to detect the scattered electrons. For specular reflection, typical elastically scattered intensities are 10 4 -10 6 counts per second, while inelastic channels have 1—104 counts per second. Energy losses of scattered electrons can be measured over a large range, typically 15 to 500 meV (120^4000 cm" 1 ) and higher. Electrons that are inelastically scattered in the specular direction have undergone a long-range interaction with surface vibrational modes that is similar to the interaction experienced by photons in reflection infrared spectroscopy at surfaces [17, 18]. This interaction is called (dynamic) dipole scattering and involves only those vibrational modes that have a long wavelength in the direction parallel to the surface (these are small-wavevector VIEWING

ULTRA HIGH VACUUM CHAMBER

PORT

,1 IN.,

Figure 6. Schematic diagram of the HREELS spectrometer used in our studies. The energy dispersive elements are cylindrical sector analyzers

Surface Structural Chemistry

169

modes that therefore can only impart momentum to cause a small deviation of the electrons away from specular reflection). The dipole scattering mechanism produces inelastic scattering that is sharply peaked near the specular beam [19-21], The angle of displacement of this intensity from the specular direction is a function of ha>0/2E, where co0 is the frequency of the vibration and Ex the impact energy, and for common experimental conditions is 0.1 to 5°. Note that the large-angle scattering (from the incidence direction to the specular direction) implicit in specular reflection is due mainly to a LEED-like diffraction by the surface (especially the substrate), which usually causes no detectable loss of kinetic energy. A specular HREELS spectrum thus exhibits loss peaks at those energies that correspond to the vibrational frequencies of the molecular (or atomic) species in their adsorbed state on the surface. This allows the ready identification of the adsorbed species by comparison with known frequencies in other circumstances, as in gasphase molecules and in particular organometallic clusters. Phonons of the substrate can also be detected in this manner [22]; their frequencies generally fall below those of interest in adsorbed molecules. In dipole scattering from metal substrates, the surface dipole selection rule states that only vibrational modes with a dynamic dipole component perpendicular to the surface can be excited. The physical basis of the selection rule is shown in Figure 7. Any dipole perpendicular to the surface generates an image force in the solid that enhances the strength of the dipole. As a result, it absorbs more energy and can be exited vibrationally quite strongly. Any dipole that is parallel to the surface has an image dipole that tends to cancel it. Therefore, the dipole scattering for dynamic dipoles oriented

«"V • pn -(frr) p .. s -¿rp.. Figure 7. Physical basis for the dipole selection rule for metal surface. A point charge above the surface induces an opposite charge at the image point below the metal surface, as shown on the left. The same argument holds for the interaction between a dipole and a metal, which is shown in the center and on the right. The relationship of the potential ((P\\) is small

170

Chapter 3: B. E. Koel, G. A. Somorjai

parallel to the surface is weak. Of course, we are concerned with the symmetry of the vibrational mode and not the perpendicular or parallel motion of the atoms involved, i.e., there are vibrational modes with a symmetry that generates no dynamic dipole moment perpendicular to the surface even though the atoms move normal to the surface, and vice-versa. The selection rule allows the intensity of energy loss peaks in specular HREELS spectra to be used to determine the symmetry of the adsorbed species and the adsorption site, and to indicate the alignment of an adsorbed molecule with respect to the surface plane. However, this selection rule is sometimes difficult to apply, since the magnitude of the dynamic dipole moment normal to the surface may be small. In these situations, the weak dipole scattering lobe may be obscured by the presence of impact scattering. These difficulties in applying the dipole selection rule can complicate the determination of the symmetry of the surface complex, especially in the case of adsorbed hydrocarbons. Reflection well away from the specular direction (greater than about 5°) occurs by so-called impact scattering, which is a short-range interaction with short-wavelength surface vibrations; in the limit it becomes the inelastic scattering of an electron by just one atom of the surface. All vibrational modes

—wnoouH — W(100)*D

6-2.0 —

0-2.0

05 •— 8-0.9 0

50

100 150 Energy loss/meV

200

Figure 8. HREELS of H and D atoms adsorbed on W(100) [23], The elastic peak is shown at left on the plot of scattered electron intensity versus the loss energy (in meV). The H coverage varies from 0 = 0.4 to 2.0 (saturation), exhibiting a change in adsorption site, while the D spectrum is shown at 0 = 2.0 only. Reproduced with permission from ref. [23]

Surface Structural Chemistry

171

in principle shbuld be detectable in off-specular HREELS data, except for certain directions not allowed by symmetry. This data is a very useful complement to the specularly measured data. The physical basis of impact scattering is still being investigated, while the transition between impact scattering and dipole scattering is essentially unexplored. New effects may thus still be discovered that can open up unexpected ways of obtaining new information about adsorbed species. Due to the high electron inelastic cross section, very weak scatterers such as hydrogen can be detected on single-crystal metal surfaces by HREELS. Figure 8 shows the spectrum obtained when hydrogen and deuterium are adsorbed on the W (100) surface [23], The complexity of the vibration spectra indicates that hydrogen is located in various sites, with various metal hydrogen stretching frequencies on the metal surface. This high sensitivity also makes adsorbed hydrocarbons relatively easier to study than currently possible by other vibrational techniques. For strong scatters, e.g. adsorbed CO, HREELS can be used to study concentrations of 0.1 % of a monolayer. Several other advantages of HREELS can be listed, in addition to the large frequency range and high sensitivity mentioned above. Both disordered and optically rough surfaces can be studied. It does not require long-range ordering of the surface, thereby giving access to the very important low coverage limit of adsorption where adsorbate-adsorbate interactions are negligible. Few techniques can handle as well as HREELS the spectral complications due to several different coadsorbed species. Finally, due to the low incident beam energies and beam currents, HREELS is, a non-destructive technique which can be used to probe even the structure of weakly adsorbed molecules or molecules especially susceptible to damage during analysis using other techniques. There are two main disadvantages of HREELS. First, the assignment of vibrational modes to individual loss peaks may not be unique due to frequency shifts as a result of bonding, especially with the relatively poor instrumental resolution (usually used) as compared to optical spectroscopies. The poor resolution limits somewhat the use of isotopic substitution and makes the analysis of closely spaced vibrational modes difficult to carry out. At present, the resolution (full width at half maximum of the elastically scattered peak) in HREELS is limited practically to ~ 5 0 c m ~ 1 (see Section 5.C) and studies have often been carried out at resolutions of 80 to 160 c m - 1 , with peak assignments made more accurately, within 10 c m - 1 . However, developments in spectrometer design, along with construction of a quiet, ultra-stable HREELS power supply [24], have recently allowed spectra to be obtained from Rh(l 11) with 20 c m - 1 resolution. The second major drawback is that the maximum pressure under which spectra can be obtained is about 5 x l 0 - 5 torr due to electron-gas collisions inside'the spectrometer. Thus, surfaces during high pressure catalytic reactions and chemisorption at the solid-liquid interface can not be directly studied. Nevertheless, the combination of a high pressure cell inside a vacuum system which has an HREELS spectrometer is helping to bridge this gap [25].

172

Chapter 3: B. E. Koel, G. A. Somorjui

4. Additional Techniques for the Determination of Surface Structure We will now discuss several other techniques that are useful for surface atomic structure analysis. Several of these have had only minor importance compared to LEED in determining surface structure to date, but their future appears to be bright. Also, the independent verification of surface structures by several techniques leads to increased confidence in the soundness of the results. The techniques which have been commonly used as structural methods in surface science can be grouped into several classes, as shown in Table 1. Techniques using diffraction and ion scattering are directly sensitive to atomic positions, and have been used widely in studying solid surfaces and adsorbed monolayers. Other techniques that measure vibrational structure, electronic structure, or the angular distribution of desorbed ions are indirectly sensitive to atomic positions by providing information on symmetry, general molecular configuration, and bond angles. These techniques have provided little structural information about clean surfaces, but have been extremely valuable for the study of atomic and molecular adsorbates. Electron microscopic techniques can directly image atomic structure in selected cases, but few surface science-type studies have been made. The techniques that use electrons as probes must be employed under vacuum conditions, but have the sensitivity to study fractional monolayers of atoms at single crystal surfaces ( ~ 1 0 1 3 atoms cm""2). The optical techniques have the large advantage of utility under atmospheric or higher pressures, but usually suffer from sensitivity problems so that well-defined single-crystal surfaces are often not studied. Development of existing and new techniques to bridge these gaps is being aggressively pursued. LEED is easily the most used diffraction technique for structural analysis, as discussed in Section 2. Two other electron diffraction techniques [26-28] differ from LEED in the range of electron energies used: Reflection highenergy electron diffraction (RHEED) uses 1-10 keV electrons and Mediumenergy electron diffraction (MEED) bridges the gap between LEED and RHEED. Multiple scattering of the electrons occurs at these energies, as in LEED. M E E D takes advantage of the larger amount of information in the I-V curves at energies up to 1000 eV, but Debye-Waller effects can require cold temperatures for the experiments. R H E E D can be used to probe to a depth of 2-10 nm and give information on structure in the near surface region. Several techniques have been developed that are based on Angle-resolved photoelectron spectroscopy (ARPES). These methods take advantage of the diffraction of the outgoing photoelectron when atoms in the solid surface are photoionized. The physics of these methods is similar to LEED, but in these cases the elctron source is internal to the sample. Angle-resolved ultraviolet photoelectron spectroscopy (ARUPS) and Angle-resolved X-ray photoelectron spectroscopy (ARXPS) have been used successfully [29, 30]

Surface Structural Chemistry Table 1. List of m a j o r techniques that are used to study surface structural chemistry Diffraction Techniques L E E D — Low-energy electron diffraction M E E D — Medium-energy electron diffraction R H E E D — Reflection high-energy electron diffraction A R P E S — Angle-resolved photoelectron spectroscopy A t o m diffraction N e u t r o n elastic diffraction Ion Scattering Techniques H E I S — High energy ion scattering M E I S — M e d i u m energy ion scattering LEI S — Low energy ion scattering Vibrational Spectroscopies H R E E L S — High resolution electron energy loss spectroscopy ITAS — Infrared transmission-absorption spectroscopy I R A S — Infrared reflection-absorption spectroscopy R a m a n scattering S E R S — Surface enhanced R a m a n scattering IETS — Inelastic electron tunneling spectroscopy NIS — N e u t r o n inelastic scattering P A S — Photoacoustic spectroscopy Ion Desorption Techniques E S D — Electron stimulated desorption E S D I A D — Electron stimulated desorption ion angular distribution P S D — P h o t o n stimulated desorption S I M S — Secondary ion mass spectroscopy Electronic Structure Spectroscopies U P S — Ultraviolet photoelectron spectroscopy XPS — X-ray photoelectron spectroscopy A E S — Auger electron spectroscopy I N S — Ion neutralization spectroscopy SPIES — Surface Penning ionization electron spectroscopy Techniques Sensitive to Absorption Coefficient Modulation S E X A F S — Surface extended x-ray absorption fine structure X A N E S — X-ray absorption near-edge structure E A P F S — Extended appearance potential fine structure Electron-Optical Techniques SEM — Scanning electron microscopy T E M — Transmission electron microscopy S T E M — Scanning-transmission electron microscopy Other Techniques T D S — Thermal desorption spectroscopy Work function measurements

173

174

Chapter 3: B. E. Koel, G. A. Somorjai

to determine atomic and molecular symmetries, and also geometries at surfaces when combined with dynamical calculations, but their utility has been limited somewhat by larger computational efforts than in L E E D (at least at lower energies) and uncertainty in final state relaxation energies. Polarization-dependent A R U P S (PARUPS) has also been used. Normal photoelectron diffraction (NPD) and Angle-resolved photoelectron diffraction fine structure (ARPEFS) methods [31, 32] are potentially as powerful as LEED, but only a few studies of this kind have been made. Atomic and molecular beams also readily diffract from surfaces [33, 34], For example, He atoms with thermal energies of 20 meV have a de Broglie wavelength of 0.1 nm. Helium is the particle most commonly used due to its low mass and its chemical inertness, but Ne, H 2 , HD, H, and D have also been used. These techniques have extremely high surface sensitivity and are non-destructive. In addition to the atomic positions, atom scattering gives additional information on the electronic charge distribution at surfaces. The angular distribution and intensity of ions scattered from a surface in channeling and blocking experiments give information on surface structure [35]. Several ion scattering spectroscopies differ only in the incident kinetic energies of the ions used: High-energy (HEIS) [36] with 0.4 to 2 MeV ions, Medium-energy (MEIS) [35] with 0.1 to 0.4 MeV ions, and Low-energy ion scattering (LEIS) [37] with ions of less than 400 keV energy. Depending on the energy and incidence direction the depth resolution can vary from one monolayer to 30 nm. At higher energies a binary collision model for the ion scattering is adequate (Rutherford backscattering) and a quantitative evaluation of the chemical composition of the surface can be made. At lower energies where the depth resolution is better, the main sources of error in the structural analysis are due to uncertainty in the ion-atom scattering potential and multiple scattering effects. There are many methods sensitive to the vibrations of surface atoms [38^1]. All of these methods indirectly give information on the atomic structure of surfaces through adsorption site symmetries, bond orders and general molecular configurations, as does H R E E L S (Section 3). Figure 9 compares the vibrational spectra obtained for CO adsorbed on dispersed rhodium particles on alumina by three techniques: (a) H R E E L S [42], (b) Infrared transmission-absorption spectroscopy (ITAS) [43] and (c) Inelastic electron tunneling spectroscopy (IETS) [44], It is clear that all of these vibrational techniques provide complementary information about the structure of adsorbed molecules. Techniques that take advantage of the absorption of infrared radiation by characteristic vibrations at surfaces include Infrared reflection-absorption spectroscopy (IRAS), and Infrared transmission-absorption spectroscopy (ITAS) [41], Each of these techniques have somewhat different advantages and disadvantages, but several broad generalizations can be made. Work on single-crystal surfaces is difficult except for studies of vibrational modes with large dynamic dipole moments, e.g., the C—O stretching mode in adsorbed CO, and only a couple of studies of hydrocarbons adsorbed on single crystal metal surfaces have been made. The accessible range of vibrational energies

Surface Structural Chemistry

175

Figure 9. Vibrational spectra taken by three different techniques for CO adsorbed on Rh supported on alumina. In the infrared spectra [43], the high resolution possible with optical techniques is evident. The inelastic electron tunneling spectrum [44] shows the downshift in the C—O stretching vibrations characteristic of this technique and relatively strong low frequency modes. The HREELS spectrum [42] shows the C—O stretching frequencies as a broad envelope of those observed in the infrared spectrum

is usually limited so that the interesting region of metal-atom stretching and bending modes usually cannot be studied. The resolution attainable is very high (0.1 cm - 1 ) so that instrumental broadening of the vibrational lines can be made negligible. This allows studies of the lineshapes, and also the detection of adsorbed species with only slightly different adsorption geometries. Importantly, the use of photons enables one to carry out studies on surfaces under high gas pressures or in the presence of liquids. In a new related development, the first observation of thermally emitted infrared radiation from both metal-carbon and C—O vibrational modes of CO adsorbed on Ni (100) has been made [45], Surface enhanced Raman scattering (SERS) [46] has been very usefully applied, especially to studying vibrations of molecules adsorbed at surfaces of electrodes in solution. Limitations on the nature of the substrate that can be used (mainly "roughened" Ag, Cu, Au) and the uncertainty on the details of the scattering mechanism have prevented broad applicability of SERS as

176

Chapter 3: B. E. Koel, G. A. Somorjai

a structural probe. Through careful and sensitive detection schemes, nonenhanced Raman scattering has been observed from pyridine on a A g ( l l l ) single-crystal surface [47], and similar studies should be widely applicable to other systems. Also, Raman scattering from near-surface layers ( ~ 2 0 n m ) can be used to observe phonon modes o f oxides and compounds, which often fingerprint the identity o f these layers. Other vibrational techniques include Neutron inelastic scattering (NIS) [48], Inelastic electron tunneling spectroscopy ( I E T S ) [49], and Photoacoustic spectroscopy (PAS) [50], None o f these techniques require vacuum, but they can only be employed for studies o f relatively high surface area materials. N I S can only observe vibrations o f H or D atoms, but can examine optically opaque samples. The scattering intensities give useful information since the atomic scattering cross-sections are known. I E T S has been used to study many large organic molecules adsorbed at surfaces, but suffers mainly from problems associated with the possible influence o f the metal counter-electrode. Ion desorption induced by bombardment o f the surface by electrons (Electron stimulated desorption, E S D ) [51], photons (Photon stimulated desorption, P S D ) [52], or ions (Secondary ion mass spectroscopy, S I M S ) [53] can also be monitored to give structural information. The ion desorption thresholds (for the ion yield versus incident excitation energy) in E S D and P S D can often be related to electronic levels o f surface atoms and used to determine the nature of the local atomic environment o f the bonding site, i.e., the identity o f the atoms to which the species was originally bound. Especially useful is the E S D ion angular distribution ( E S D I A D ) technique for determining molecular structure at surfaces [54]. In this technique, desorbed ions produce spots on a fluorescent screen, and the spot distribution can be used to directly determine bond angles in molecular species. There are several methods that measure the electronic structure of atoms and molecules at surfaces, and thus are indirectly sensitive to atomic structure. Ultraviolet (UPS) and X-ray photoelectron spectroscopy ( X P S ) have been used extensively in surface analysis and can give qualitative information about surface structure [55, 56], The valence electronic density o f states can be measured and energy level shifts can be used to determine the atoms involved in chemisorption bonds. In addition, chemical shifts in core level binding energies measured in X P S can often be used to distinguish between atoms in the adsorbed state, atoms incorporated within the first layer, and atoms which have penetrated several layers to form compounds. Chemical shifts and lineshape changes in Auger electron spectroscopy ( A E S ) have been shown to also give valuable structural information [57, 58]. Two electronic spectroscopies give information on the density-of-states distribution from the outer part o f the solid-vacuum interface: Ion neutralization spectroscopy (INS) [59] and Surface Penning ionization electron spectroscopy ( S P I E S ) [60, 61] or Metastable deexcitation spectroscopy ( M D S ) [62], By using synchrotron radiation and monitoring the total electron yield, Auger electron yield or ion yield, one can measure modulations in the photoabsorption cross-section for surface atoms [63, 64], analogous to E X A F S

Surface Structural Chemistry

177

[65], Surface extended X-ray absorption fine structure (SEXAFS) is a technique based on this observation and is a powerful source of information about the local environment of selected atoms on surfaces with or without longrange order. Use of this technique has allowed the determination of adsorption sites and bond lengths of fractional monolayers of atoms. Another technique, X-ray absorption near edge structure (XANES), also called Nearedge X-ray absorption fine structure (NEXAFS) [66], uses the yield structure within the first 50 eV of the absorption edge. The fingerprint of this region has been shown to be sensitive to the unoccupied electronic density-of-states and coordination symmetry of surface species. Extended appearance potential fine structure (EAPFS) [67], also analogous to EXAFS, probes the short-range order of a particular element. EAPFS does not require synchrotron radiation (only an electron gun and LEED retarding analyzer) and can be used to study surface atoms in monolayer concentrations. Using electron optical techniques, Transmission (TEM), Scanning (SEM) and Scanning transmission electron microscopy (STEM) can be used for direct imaging of the structure of solid surfaces [68-70], TEM and STEM have allowed resolution of individual atoms. These techniques are usually limited by electron-atom cross sections to heavy atoms on light substrates and to operation in relatively poor vacuums ( > 10~ 6 Pa) with high magnetic fields. Adsorbed molecules cannot be studied due to electron beam damage. Also, in transmission modes it is difficult to separate the effects of two surfaces. However, the potential for electron microscopy to study the atomic structure of surfaces is great. Two other techniques give indirect information about the atomic geometry of adsorbed monolayers on solid surfaces. Thermal desorption spectroscopy (TDS) [71] can be used to detect different bonding states of adsorbates by measuring the heat of desorption from these states. The relative populations of the bonding states, and sometimes the absolute coverage, can be found by integrating the spectra. Work function measurements [72] detect changes in charge distribution at the surface. Even though the work function change does not relate simply to the adsorption geometry, measurements can often indicate the general bonding configuration and direction of charge transfer between adsorbate and substrate atoms.

5. Structure of Solid Sufaces A. The Atomic Structure of Clean Surfaces The structure and bonding of an adsorbed species is greatly influenced by the structure of the substrate. In order to explore the structural sensitivity of chemical bonding and to obtain structural information on adsorbates, we must know the atomic structure of clean surfaces prior to adsorption. It is also important to know whether the presence of the adsorbate substantially alters the geometric structure of the substrate. Over the past ten

178

Chapter 3: B. E. Koel, G. A. Somorjai

years a good picture has emerged of the details of atomic structure of many surfaces of metals and semiconductors. Two major phenomena are found: bond length relaxation and reconstruction. Relaxation causes a contraction in the distance between the first and the second layers of atoms at the surface; such relaxations sometimes extend to deeper layers. The interlayer distance between the 1st and 2nd layers may contract up to 15 % with respect to interlayer distances in the bulk material. The more open the surface, that is the lower the surface density of atoms, the larger is the relaxation. The precise location of atoms in the first layer does not noticeably change parallel to the surface, only their location in the direction perpendicular to the surface shows alterations. This phenomenon can be understood if we assume that the surface is an intermediate between the diatomic gas phase molecule of the same atomic number and atoms in the bulk. The diatomic molecules have much smaller atomic distances than bulk atoms that have very large coordination numbers, namely 8 to 12 nearest neighbors. In the surface, because of the anisotropy of location and the reduced number of nearest neighbors, there is a contraction of the top interlayer distance. Reconstruction of the surface occurs when the forces on the surface atoms in the solid are very large and the atoms are forced to move to new atomic locations in order to minimize their surface energy. In this case, the atoms seek new locations in both perpendicular and parallel directions to the surface, which results in new surface structures. LEED diffraction patterns are observed that are very different from what is expected from the projection of the bulk unit cell to the studied surface. The diffraction pattern from a Pt (100) surface is shown in the upper left panel in Figure 10. The LEED pattern and structure that one would expect from the projection of the bulk unit cell is shown on the right and is a square unit mesh. The approximate structure of the clean reconstructed surface is shown in the lower left panel. While the LEED pattern was published in 1965, a solution of the surface structure wats reported only in 1981 [73]. The surface platinum atoms are reconstructed into an hexagonal configuration; the coincidence of atomic positions in this reconstructed hexagonal top layer and the unreconstructed second layer gives rise to the complicated diffraction pattern that is shown in Figure 10. The variation in the number of nearest neighbors forces the surface atoms into an undulating configuration. Since buckling increases the total energy, the atoms move into positions that minimize the surface undulation. The precise location of atoms in this reconstructed surface is governed by a delicate balance of forces. Upon adsorption of even small amounts (several percent of a monolayer) of a chemisorbed species such as carbon monoxide or hydrocarbon molecules, the atoms in this reconstructed surface snap back to the equilibrium position that they would have in the bulk: a square unit mesh appears, as shown in the upper right panel of Figure 10. On desorption of these molecules, the clean surface shows the reconstructed surface structure again. It appears that the (100) crystal faces of gold, platinum, and iridium all show the formation of large superlattices, e.g. (5 x 1) or (5 x 20) reconstruc-

Surface Structural Chemistry

179

Figure 10. Left: Diffraction pattern and model of the surface structure for the ( 5 x 1 ) surface reconstruction of the Pt(100) crystal face. Right: Diffraction pattern and surface structure that might be expected for the Pt(100) surface assuming simple termination of the bulk lattice

tions [73], The (110) faces of Au, Pt, and Ir often exhibit (n x 1) (with n = 2,3, 4) reconstructions [74], The "missing row" model best explains several of these systems. In this model, small facets of the (111) face are built. The driving force is the lower free energy of the (111) face. The tungsten and molybdenum (100) crystal surfaces also exhibit reconstruction that have been reviewed recently [74a], Reconstructions are generally observed on semiconductor surfaces, often with several different metastable reconstructions observed for the same compound. A model of the surface structure of the reconstructed Si (100) surface [75] is shown in Figure 11. In this case, one may consider the silicon surface atoms as existing in dimers with troughs in between. The contraction actually permeates at least three layers and so the effect of sur-

Chapter 3: B. E. Koel, G. A. Somorjii

180

-0.007nm 0.009nm 0.005nm

0.008nm Si ( 1 0 0 ) ideal

Si

(100)

p (2x1)

Figure 11. Top and side views of ideal bulk-like Si(100) at the left, and the Si(lOO) p(2 x 1) reconstruction. Layer-spacing contractions and intralayer atomic displacements relative to the bulk structure are given. Shading differentiates surface layers

face reconstruction is deeper than just the top surface layer. The reconstructions in semiconductors are throught to be due to rehybridization of the orbitals of the surface atoms. Several recent articles cover this exciting area of semiconductor surface structure [76-79]. The advent of increased computing power is currently revolutionizing our ability to understand the microscopic details of complicated reconstructions.

1. Atomic Structure of Unreconstructed Low Miller Index Planes of Transition Metal Surfaces One can generally observe very small contractions ( 1 ^ %) of the bond lengths between the surface atoms and the second-layer atoms for the relatively open faces, such as bcc (100), fee (110), bcc (111), and fee (311). This does not result in a reconstruction of the surface layer. The effect of adsorbates on such relaxed surfaces is to restore the bond lengths to their bulk values, or sometimes even to lengthen them. Contraction or relaxation of atoms at open crystal surfaces is due to the reduction of the positive surface free energy if the surface becomes less rough on the atomic scale. Also, with fewer neighbors the two-body repulsion energy is smaller, allowing greater atomic overlap at shorter bond lengths.

181

Surface Structural Chemistry

2. The Atomic Structure of High Miller Index Stepped and Kinked

Surfaces

When crystals are cut along high Miller index directions, the surfaces often assume sterned ordered, configurations.

(Ill) Plane (Hexagonally Close-packed)

P t (III)

(755)

Plane

(100) Plane

P t ( S ) - [ 6 (II l)x(IOO)]-*-*Pt ( 7 5 5 )

(lll)"Terrace" Planes

Figure 12. LEED patterns (left) and surface structures (right) of a flat, b stepped, and c kinked platinum surfaces

182

Chapter 3: B. E. Koel, G. A. Somorjai

These periodic steps in the surface produce recognizable diffraction features, permitting the determination of the height and orientation of the step as well as the terrace width. The orientation of the steps and terraces that are stable correspond normally to those of the highest density atomic planes: (111), (100) and (110) for the fee and the bcc crystals. By changing the angle of the cut, the terrace width and the step density can be altered. By cutting crystals in such a way that even the steps have high Miller indices, one obtains kinked surfaces. Figure 12 shows structures and diffraction patterns obtained for clean platinum surfaces with (111) terraces and high densities of steps and kinks. The splitting of the diffraction beams is characteristic of the new periodicities introduced by the periodic arrangements of atomic steps on the surface. Atoms at kink sites have even lower numbers of nearest neighbors than atoms in stepped positions. The heats of adsorption of atoms and molecules at these different sites are likely to be different. As a result, their chemical activities in various rearrangement or dissociation reactions at these sites are different. It is therefore very important to study the effect of changing atomic structure on the location, bonding, and of atoms and molecules on solid surfaces. The ordered, one-atom height step and periodic terrace configuration appears to be the stable surface structure for many high Miller-index surfaces of metals. Upon heating to near the melting point the steps disorder but reorder again when annealed at lower temperatures. In the presence of a monolayer of oxygen, carbon, or sulfur, many stepped surfaces undergo restructuring. The step height and terrace width may double or faceting may take place whereby the step orientation becomes more prominent than the terrace orientation, giving rise to new diffraction features that are detectable by LEED. The driving force for this surface restructuring in the presence of adsorbates appears to be the difference in chemical bonding of adsorbates to the different crystal faces of the metal which alters the relative surface free energies of the crystal faces. Surfaces that have the lowest free energies when clean become less stable than other crystal faces when covered with adsorbates.

B. The Structure of Adsorbed Atoms on Solid Surfaces 1. Non-Metal

Adsorption

The various high symmetry adsorption sites on solid surfaces with low Miller indices are shown in Figure 13. Most atoms whose adsorption and surface structure have been studied by LEED prefer these sites with highest symmetry. It appears that the atoms generally occupy positions with the largest number of metal nearest neighbors and this allows the greatest binding energy between adsorbate and substrate atoms. Figure 14 shows the interatomic distances that were obtained from the surface structures along with the range of

Surface Structural Chemistry

(d)bcc(100): hollow site

183

(e)fcc(110): center long-

(f) hep (0001) : underlayer

and short-bridge sites

Figure 13. T o p and side views (in top and bottom sketches of each panel) of adosprtion geometries on various metal surfaces. Adsorbates are drawn shaded. Dotted lines represent clean surface atomic positions.

interatomic distances that are indicated from X-ray or electron diffraction studies of bulk compounds or gas phase molecules. It appears that the bonding as judged by the interatomic distance for surface atoms falls in the range of bonding found for compounds in the solid state or in the gas phase. This in most cases indicates covalent bonding. Thus, the surface bonding is not qualitatively different from that found in other phases. In the right side of

184

Chapter 3: B. E. Koel, G. A. Somorjai

Figure 14, the ionic character of the bond is shown, as judged by the work function change that accompanies adsorption. The ionic character is very small indeed. It appears that this is an additional confirmation of the covalent bond character of these surface phases.

system

bond length (A) 2.0

3.0

charge transfer ( % electr.) -8 -4 0 4 8

H / N i (III) Na/Al (100) Na/Ni (100) Si/Mo (100) N/Mo (100) N/Cu (100) 0/W (110) O/Mo (100) O/Co (100) O/Fe (100) O/Ni ( I I I ) O/Ni (100) O/Ni (110) S/Fe S/Ni S/Ni S/Ni

(100) (III) (100) (110)

Se/Ni (100) Se/Ag (100) Te /Ni (100) Te/Cu (100) C l / A g (100) I/Ag (III) C d / T i (0001)

Figure 14. (Left) Comparison of adsorption bond lengths at surfaces (arrows show uncertainty) with equivalent bond lengths in molecules and bulk compounds (blocks extending over range of value found in standard tables). (Right) Induced charge transfer for adsorption

In some cases adsorption results in surface reconstruction. For example, when oxygen adsorbs on the Fe(100) surface [80] or sulfur adsorbs on the Fe (110) surface [81], the surface layer consists of both adatom and iron atoms in the same plane as a precursor to the formation of iron compounds. Reconstruction occurs when the adsorbate-substrate bond is stronger than the bonds between substrate atoms. It is likely that oxidation or other compound formation is accompanied by surface reconstruction. Future studies will certainly explore the role of reconstruction on the initial state of bulk compound formation. Under some conditions a small atom assumes

Surface Structural Chemistry

185

a position under the surface. The systems N/Ti(0001) [82], 0 / A l ( l l l ) [83], and H/Pd(110) [84] illustrate this point. 2. Adsorption and Growth of Layers of Metals on Surfaces of Other Metals At low coverages, the adsorbate-substrate interaction is dominant and when one metal is deposited on another metal, it is usually found by LEED studies that the surface structure of the deposited metal follows the periodicity of the substrate metal. For example, when gold is deposited on the Pt(100) surface, the gold atoms locate with the periodicity provided by the platinum atoms [85], This forces the gold atoms into a different interatomic distance than in its own lattice. That in turn may change not only its geometric structure, but also its electronic properties. For this and many other systems, the forces that control epitaxy, the strong interaction between adsorbate and substrate, seem to predominate and control the atomic surface structure. The relative importance of adsorbate-adsorbate interactions increases at higher coverages and can be dominant especially for large radius metallic adatoms (e.g., K, Rb, and Cs). Thus, at higher coverages, the adsorbate may continue to follow the substrate periodicity, or form coincidence structures, or new periodicities that are unrelated to the substrate periodicity. For example, alkali adatoms tend to form incomensurate hep layers on any metal substrate. Still, when gold is condensed in multilayers over platinum surfaces, the gold interatomic distance remains controlled by platinum for the first ten layers of gold [85]. Thus, the effect of the substrate that controls the structure of the adsorbed metal is felt during the growth of the thin metal film. The dominance of epitaxy in metal-metal interactions provides an opportunity to deposit metal monolayers and thin films with interesting atomic and electronic structures. This is an area of fruitful research for the near future.

C. Surface Structure of Molecules on Solid Surfaces A large number and wide variety of ordered monolayers of adsorbed molecules have been observed by LEED and studied by many other techniques [86]. Still, very few adsorbed molecular structures have been analyzed by LEED surface crystallography or other techniques to yield accurate atomic positions and bond lengths. Associatively adsorbed CO is the only diatomic molecule studied in this manner to date, and the adsorption geometry of CO on several metal surfaces has been .determined by LEED crystallography. These are shown in Table 2, in which we list the results for those CO adsorption systems that have been analyzed by both HREELS and LEED. In these cases, the CO molecules are found to stand perpendicularly to the surface in either top sites or bridge sites (hollow sites on clean metal surfaces are rarely occupied by CO). In addition, almost all of the above systems have been studied by other

186

C h a p t e r 3: B. E. K o e l , G . A .

serves as a confirmation of the postulated correspondence in HREELS between adsorption site and vibrational frequency range for CO adsorbed on different metal surfaces. This confirmation is thereby extended to other than the fcc(100) substrate face, for which it was established previously with CO on Ni, Cu, and Pd(100). A summary of these results is included in Table 2. Note that the frequency v c _ Q for the R h ( l l l ) + ( j / 3 + j/3) R30° structure is closer to the frequency range associated with a bridge-bonded CO molecule than that for CO on Ni or Cu(100). Such confirmations of the expected sites provide an important calibration of the vibrational techniques in the sense that the knowledge of the CO adsorption site at one coverage or on one crystal face can be used to determine, without the help of further LEED intensity analyses, the adsorption site on other substrate faces, at other coverages, or in disordered states. LEED analysis of the (2 x 2) structure of CO on Rh(l 11) at 3/4 monolayer coverage has in turn confirmed the HREELS prediction that both bridge sites and top sites are occupied in that dense structure. The structure of R h ( l l l ) + ( 2 x 2 ) 3CO is shown in Figure 18, as determined by LEED. This was a more complicated analysis, because three molecules fit in each unit cell and there were consequently more structural parameters to fit the experiment, a situation that LEED practitioners are only now learning to handle. A general surface arrangement for this case might assume a hexagonal lattice of molecules (due to the dense packing), all oriented perpendicularly to the surface. However, this choice forces the atop-site molecules off the atop sites by 0.078 nm, which may not be the most favorable bonding geometry. The LEED intensity analysis indicates that, while the CO molecular axes are indeed essentially perpendicular to the surface (within about 10°) the atop-site molecules appear to move closer to the atop sites than illustrated (by about 0.025 nm), but not all the way because of steric hindrance. These "near-atop" molecules have a Rh—C bond length of 0.194 ± 0.007 nm (compared with 0.195 ± 0.01 nm in the atop-only (j/3 x j/3) R30° structure) with a Rh—C—O bond angle of 164 ± 10°, while the C—O bond length is 0.115 ± 0.01 nm (compared with 0.107 ± 0.01 nm in the atop-only structure). The bridge site molecules have a larger Rh—C bond length of

Surface Structural Chemistry

193

Figure 18. Structure of R h ( l l l ) + ( 2 x 2 ) 3CO. The upper figure presents a side-view of the surface and the lower figure gives a top-view. The large circles represent Rh atoms (dotted— out of plane, full—in plane), and the small circles are either carbon or oxygen atoms (dotted— hexagonal mesh, full—measured positions). The five structural parameters that were varied in the LEED analysis are illustrated in the upper left corner. Reproduced with permission from ref. [98]

0.203 ± 0.01 nm, with again a C—O bond length of 0.115 ± 0.01 nm. These values are in good agreement with corresponding values found in rhodium carbonyl clusters [99], where atop site and bridge site molecules have Rh—C bond lengths of 0.182-0.192 nm and 0.200-0.208 nm, respectively, and C—O bond lengths of 0.109-0.117 nm and 0.114-0.117 nm, respectively. In conclusion, by combining TDS, HREELS, and LEED analyses we can present a fairly complete picture of CO chemisorption on R h ( l l l ) . At very low exposures a single species is present on the surface located in an atop site (v R h _ c = 468 c m - 1 , v c _ Q — 2016 cm" 1 ). As the coverage increases, the bonding to the surface becomes weaker and the TDS peak maximum shifts to lower temperatures [94, 100, 101]. This process continues until after approximately 0.5 L exposure where a (j/3 x | / 3 ) R30° LEED pattern is seen

194

Chapter 3: B. E. Koel, G. A. Somorjai

and all of the adsorbed CO molecules are in atop sites linearly bonded to individual rhodium atoms, with a Rh—C bond length of 0.915 ± 0.01 nm and a C—O bond length of 0.107 + 0.01 nm. Above this coverage, a second C—O stretching vibration corresponding to a bridge-bonded species is observed ( v R h - c = 3 8 0 c m _ 1 ' v c - o = 1 8 5 5 c m _ 1 ) - A "split" ( 2 x 2 ) L E E D pattern is seen indicating a loosely packed overlayer of adsorbate molecules. This overlayer structure compresses upon further CO exposure. Throughout this intermediate coverage regime there is a mixed layer of atop and bridge-bonded CO species, and we see a continuous growth of all H R E E L S peaks. Two peaks are also visible in the T D S spectra with a bridge-bonded CO having a 4 kcal/mole lower binding energy to the surface than the species located in the atop site. With a background pressure of ~ 1 . 3 x l 0 ~ 4 Pa CO at 300 K, a (2 x 2) L E E D pattern forms whose unit cell consists of three CO molecules, two atop and one bridged, in reasonable agreement with the 2:1 peak intensity ratio found in the H R E E L S spectra. L E E D indicates that all CO molecules are still oriented nearly perpendicular to the surface in this dense (2 x 2) structure, with Rh—C bond lengths of 0.194 + 0.01 nm and 0.203 ± 0.01 nm and CO bond lengths of 0.115 ± 0.01 nm and 0.115 ± 0.01 nm for near-atop and bridge site molecules, respectively.

2. Structure of the Adsorbed Benzene Monolayer on Rh(111) Four different L E E D patterns have been observed for benzene adsorption /3 1 on R h ( l l l ) at 240-395 K [102-106]. Most information is for I { = c (2 ]/3 x 4 ) rect pattern (the "rect" notation indicates a rectangular unit cell with sides 2 |/3 and 4 times the substrate surface lattice constant) and a 3 0\ Q

I = (3 x 3) pattern. The L E E D patterns and the geometry of the

adsorbed monolayer for these structures are shown in Figure 19 and Figure 20, respectively / 3 2\ [103], The other observed patterns were)V2 2 Jj = (2 1/T x 3) rect and I ^ J = ( ] / 7 x 7 ) R 1 9 . 1 ° . The sizes of the four corresponding unit cells are 8, 9, 6 and 7, respectively, in terms of the number of surface Rh atoms included. The unit cells of size 7, 8, and 9 are compatible with known Van der Waals dimensions of flat-lying benzene molecules, assuming one molecule per cell; the (2 j / 3 x 3) rect unit cell contains two very crowded flatlying molecules. The benzene molecules are known to lie flat from H R E E L S data [102, 106], The c(2 [/3 x 4) rect structure is stable up to about 370 K. At higher temperatures an irreversible order-order phase transition occurs to form the (3 x 3) structure. The (3 x 3) structure is stable to about 395 K, where it irreversibly disorders (just prior to the H 2 desorption peak in TDS). There are indications that the (3 x 3) structure might be stabilized by car-

Surface Structural Chemistry

195

Figure 19. Rh(111)

^ C 6 H„

a LEED pattern at normal incidence at beam energy 50 eV: diffraction photograph at left; schematic diagram at right showing three unit cells in reciprocal space, corresponding to three domain orientations. b Real-space unit cell corresponding to the observed diffraction pattern, exhibiting the

' \unit Vi 3 ; cell and the centered rectangular c(2|/3 x 4) rect cell for one domain orientation, c and d Two possible models for benzene adsorption, differing by the azimuthal orientation of flat-lying molecules. The molecules are drawn as lines connecting C and H nuclei. The closest intermolecular distances are shown between H nuclei. Reproduced with permission from ref. [103]

^•10

196

Chapter 3: B. E. Koel, G. A. Somorjai

X X A X J. J^J- A.J.,.. 1 . Figure 20. Rh(l 11) + ( 3 x 3 ) C 6 H 6 a Photograph of LEED pattern at normal incidence at beam energy 50 eV. b Real-space unit cell corresponding to the observed pattern, c and d Two possible models for benzene adsorption, analogous to Figures 19c and 19d. Reproduced with permission from ref. [103]

bonaceous fragments resulting from partial benzene decomposition at the higher temperatures or from CO coadsorption. TDS, using a 10 K/sec linear heating rate, shows two H 2 desorption states from a saturated surface: a peak at 413 K, due to decomposition of molecular benzene, and a broad state extending to about 700 K, due to subsequent dehydrogenation of the remaining hydrocarbon fragments. In addition, a small amount ( ^ 20 %) of molecular benzene desorption occurs prior to 415 K. The existence of commensurate overlayer structures and the high desorption temperature of benzene on R h ( l l l ) indicate strong metal-carbon bonding, which in the flat-lying geometry would involve the re-orbitals of the benzene ring. Strong bonding to the metal could distort the molecules: e.g., C—C bond length expansions and C—H bond bending away from the surface might be expected in analogy with acetylene and ethylene adsorption on transition metals and with benzene structures in organometallic clusters. However, HREELS which will be discussed later, shows that this molecular distortion, if any, preserves a high symmetry of type C3v(f7d) for both the c(2 l / 3 x 4) rect and (3 x 3) structures [102, 106]. By comparing measured and calculated LEED I-V curves, the detailed position of the adsorbed benzene molecules in the c(2 l/3 x 4) rect and (3 x 3) structures have been analyzed [104, 105], For both of these structures, about 960 structural models were investigated, differing in metal-molecule interlayer spacing, adsorption site, azimuthal orientation of the molecules about their six-fold axis, buckling, and planar distortions. In the c(2 j/3 x 4) rect case, as shown in Figure 21, LEED calculations find that benzene is centered over a hep hollow adsorption site (over a Rh atom in the second metal

Surface Structural Chemistry

197

Figure 21. Optimal structure found for R h ( l l l ) + c(2|/3 x 4) rect C 6 H 6 (H positions are assumed), including Van der Waals radii of 0.18 and 0.12 nm for C and H, respectively. A unit cell is outlined in the bottom panel. The right-hand benzene molecule shows the preferred in-plane distortion (C—C bond lengths of 0.125 and 0.16 nm). The side view in the top panel includes possible CH bending away from the surface. Reproduced with permission from ref. [105]

layer); each of the three metal atoms around the hollow site would be bonded to two carbon atoms equally distant at about 0.235 ± 0.005 nm. This bonding corresponds to a planar (possibly distorted, as in Figure 21) C 6 ring with a metal-molecule layer spacing of 0.215 nm, similar to corresponding values in organometallic clusters containing aromatic rings. The symmetry of the adsorption site is C3v((rd). In this symmetry group, the symmetry planes of the R h ( l l l ) substrate bisect the dihedral angles between the H atoms of the benzene ring. In the (3 x 3) case, no structural model has so far given satisfactory agreement between theory and experiment. HREELS spectra for specular scattering are shown in Figure 22 [102], These were taken following benzene adsorption on Rh(l 11) at 300 K to give a well-ordered c(2|/3 x 4) rect structure in LEED. The isotopic shifts observed for the spectra of C 6 H 6 and C 6 D 6 shown in Figures 22(A) and 22(B), respectively, allow for the identification of the losses at 345, 550, and 1420 c m - 1 as two Rh—C and one C—C vibration frequencies, and those at 810, 1130 and 3000 c m - 1 as C—H vibration frequencies. Strong bonding between the molecularly adsorbed benzene and metal occurs, as evidenced by the adsorption induced shifts of the C—H bending mode (from 673 cm" 1 in the gas phase to 3000 cm - 1 ), but substantial rehydridization to an in the gas phase to 30000 cm" 1 ), but substantial rehydridization to an adsorbed cyclohexane-like species does not occur. From specular HREELS spectra, using the surface dipole selection rule and comparing infrared spectra of gas and liquid phase benzene, one can immediately conclude that

198

Chapter 3: B. E. Koel. G. A. Somorjai

Figure 22. Vibrational spectra obtained by HREELS in the specular direction for saturation coverage of benzene chemisorbed on R h ( l l l ) at 300 K which gives a c(2 ]/3 x 4) rect LEED pattern: (A) C 6 H 6 (B) C 6 D 6 . The incident beam energy was 3.5 eV. Reproduced with permission from ref. [102] 1000 2000 Energy loss/cnr 1

the benzene molecule is adsorbed with the ring plane essentially parallel to the Rh(l 11) surface in the c(21/3 x 4) structure, in agreement with LEED molecular-packing arguments. The most intense feature in Figure 22(A) is the 810 c m " 1 loss corresponding to the out-of-plane C—H bending mode. The inplane vibrational modes have very small intensities: 1130 c m - 1 , C—H bend; 1420 c m " 1 , C—C bend; and 3000 cm" 1 , C—H stretch. Structural information regarding the adsorption geometry and the symmetry of the adsorbed complex can be determined by comparing the number, frequency, and intensity of the dipole-active modes with the correlation table of the point group for the gas phase molecule. Adsorption of benzene with a symmetry group lower than C 3v can be ruled out, due to the small number of vibrational modes observed on-specular in Figure 22. Further refinement of the symmetry of the adsorbed complex is more difficult, since it has been observed that for adsorbed hydrocarbons the impact and dipole scattering in specular spectra are often of similar intensity. Thus, observation of a loss peak in the specular spectrum does not necessarily mean that the mode is dipole active.

199

Surface Structural Chemistry

C 5 H 6 /Rh (111) 15° off Specular

3000 U20

V.

x300 810 Specular ^ ¿,8 cm

1000

2000 Energy loss/crrr 1

Figure 23. HREELS spectra obtained for specular and 15° off-specular scattering angles. The R h ( l l l ) surface was saturated with benzene (C 6 H 6 ) to produce the c(2|/3 x 4) rect structure. Reproduced with permission from ref. [102] 3000

Representative specular and off-specular spectra for benzene adsorbed to give the c(2 | / 3 x 4) rect structure are shown in Figure 23. The off-specular spectrum was taken after a 7.5° rotation of the R h ( l l l ) surface normal towards the analyzer, which corresponds to 15° off-specular scattering. This rotation caused a decrease in the elastic peak intensity by a factor of 170. The losses at 350 and 810 c m - 1 were reduced by a factor of 10-15, while the other losses decreased in intensity by factors of 1.5-4. In addition, loss peaks can be identified in the off-specular spectra at 780, 880, 990 and 1320 cm" 1 . Except for the 350 and 8 1 0 c m " 1 losses, the impact contribution to the observed intensity in specular scattering is substantial, and this makes the assignment of dipole-active peaks difficult. However, after a detailed angledependent study [106], we believe that all of the losses observed in the specular spectra have a non-zero dipole-active contribution. The observation of the dipole-active peak at 1130cm" 1 (in-plane C—H bend, v10 in the free molecule) leads to the conclusion that the adsorption site symmetry is C 3v (ff d ). [106, 107] This result confirms the symmetry assignment from dynamic LEED calculations.

200

Chapter 3: B. E. Koel, G. A. Somorjai

Ordered structures of adsorbed benzene have been observed on several metal surfaces: it is significant that they are all compatible with flat-lying benzene molecules, as we shall now show. The area of the benzene molecule in projection on its ring plane can be roughly estimated as that of the smallest rectangle that encloses it, using the Van der Waals ra ii of 0.12 nm: 0.5 nm 2 . In the following cases the observed LEED pattern is consistent with one molecule per unit cell (the unit cell area A is given for comparison). 2 Ni(100) c(4 x 4 ) - C 6 H 6 , [107] A = 0.4960 nm Rh(l 11) c(2 l / 3 x 4) rect-C 6 H 6 , A - 0.4976 nm 2 , Rh(l 11) (3 x 3 ) - C 6 H 6 , A = 0.5598 nm 2 , Ir(l 11) (3 x 3 ) - C 6 H 6 , [108] A = 0.5766 nm 2 , Pd(100) c(4 x 4)—C 6 H 6 , [93] A = 0.6006 nm 2 , N i ( l l l ) ( 2 ] / 3 x 2 ] / 3 ) R30°—C 6 H 6 , [107] A = 0.6443 nm 2 ,

/ VY

rc/nm

a/deg

0.125

0.065

131.3

0.134

0.074

128.1

0.208

0.135

0.073

128.1

0.150

0.200

0.139

0.061

127.0

Rh(111)+(2x2)CCH 3

0.145

0.203

0.134

0.069

130.2

H3C-CH3

0.154

0.077

109.5

H2C =CH2

0.133

0.068

122.3

HC = CH

0.120

0.060

180.0

C/nm

m/nm

rM/nm

CO 3 (CO) 9 CCH 3

0.153

0.190

H 3 R U 3 (CO) 9 CCH 3

0.151

0.208

H3OS3(CO)9CCH3

0.151

Pt (111)+(2X2)CCH 3

Figure 24. A model of the ethylidyne surface species. A comparison is made between the bond angles and distances found for this structure by LEED and those for corresponding organometallic compounds. r c = carbon covalent radius; rM = bulk metal atomic radius. Reproduced with permission from ref. [110]

201

Surface Structural Chemistry

On Pt(l 11) two benzene patterns have been observed that may be explained in terms of two flat-lying molecules per unit cell (half the unit cell area is therefore given here): P t ( l l l ) ( 2 | / 3 x 4 ) rect-2C 6 H 6 , [110] A/2 = 0.5316 nm, Pt(l 11) (2 j / 3 x 5) rect-2 C 6 H 6 , [110] A/2 = 0.6645 nm 2 . It is of interest that no well-ordered incommensurate benzene structures have been reported in the literature or observed in our work on the various metal surfaces: this implies that the substrate-benzene interactions are strong compared with the benzene-benzene interactions. 3. The Temperature Dependent Character of the Surface Chemical Bond: The Adsorption and Thermal Decomposition of Alkenes on Rh(lll) and Pt(lll) Surfaces Molecular adsorption of ethylene occurs at low temperatures on metal surfaces, at less than 240 K on Rh(l 11) and 280 K on Pt(l 11). The adsorbed molecules give no ordered structures, i.e., no well-defined LEED patterns, but have been shown by UPS and H REELS to be bonded parallel to the metal surface and significantly rehydridized compared to the gas phase. Approximately sp3 hybridization of the carbon atom results, while strong di- \

co B-400 o V)

A

"O

a

•

o

2 200

^ A

x

\

ó

e

-

B 0

Se Ti Y Zr La Hf

o 1

V Cr Mn Fé Co Ni Nb Mo Te Ru Rh Pd Ta W Re Os Ir Pt

Is

i

1

1

Cu Zn Ga Ge o Ag Cd In Sn • Au Hg TI Pb A

Figure 35. Heats of adsorption of C O on polycrystalline transitionmetal surfaces

Surface Structural Chemistry

Se Y La

Ti Zr Hf

V

Nb

To

Cr Mo W

213

Mn Tc Re

Fe Ru Os

Co Rh Ir

Ni Pd Pt

Cu Ag Au

Zn Cd Hg

Ga In Tl

Ge Sn Pb

O • A

Figure 36. Heats of adsorption of hydrogen on single-crystal surfaces of transition metals

The physical picture that emerges from these surface studies is one of the predominance of surface structure-sensitive, localized bonding. An atom may adsorb on a high symmetry three-fold, bridge or two-fold, or on an atop or one-fold site. In each of these sites, the bonding strength may be different from that in other sites. Of course, in the presence of atomic steps and kinks, there are even more sites with different structures that may further change the local chemical bond. Thus, the localized bonding that involves an adsórbate atom or molecule and the nearest neighbor surface atoms indicates cluster-like surface chemical bonding, that describes well the structural and chemical characteristics of the surface adsorbate-substrate systems. Because

214

Chapter 3: B. E. Koel, G. A. Somorjai

Figure 37. Heat of adsorption of hydrogen on polycrystalline transitionmetal surfaces. Data f r o m : D. O. Hayward and B. M. W. Trapnell, Chemisorption, Butterworths, London, 1964; J. K. Anderson (ed.), Chemisorption and Reactions on Metallic Films, Academic, New York, 1971; I. Toyoshima and G. A. Somorjai, Catal. Rev. — Sci. Eng. 19 (1), 105 (1979)

of the structural richness of each surface, the nature of the surface chemical bond reflects the same diversity and complexity. Over the past ten years, there have been great advances in our understanding of the nature of the surface chemical bond and the structure of adsorbed atoms and molecules on surfaces. We have briefly reviewed some results of studies, mostly by LEED and HREELS. As these and other techniques become more widely applied, the increased availability of experimental data will further accelerate the rate of development of surface chemistry and its applied subfields, catalysis among them. A cknowledgement

We wish to thank Dr. M. A. van Hove for his many valuable discussions. One of us (BEK) would like to acknowledge support by the Miller Research Institute, University of California, Berkeley. This work was supported by the Director, Office of Energy Research, Office of Basic Energy Sciences, Materials Sciences Division of the U.S. Department of Energy under Contract DE-AC03-76SF00098.

8. References 1. Wood, E. A.: J. Appl. Phys. 35, 1306 (1964) 2. Somorjai, G. A., Van Hove, M. A.: Adsorbed Monolayers on Solid Surfaces, Structure and Bonding, Vol. 38. Berlin, Heidelberg, New York: Springer 1979 3. Ertl, G., Küppers, J.: Low Energy Electrons and Surface Chemistry. Weinheim: Verlag Chemie 1974 4. Heugier, M.: Surf. Sci. 19, 159 (1970)

Surface Structural Chemistry

215

5. Laramore, G. E., Houston, J. E., Park, R. L.: J. Vac. Sei. Technol. 10, 196 (1973) 6. Jagodzinski, H„ Moritz. W„ Wolf, D.: Surf. Sei. 77, 233 (1978) 7. Ertl, G., Schillinger, D.: J. Chem. Phys. 66, 2569 (1977) 8. Pendry, J. B.: Low Energy Electron Diffraction. New York —London: Academic Press 1974 9. Van Hove, M. A., Tong, S. Y.: Surface Crystallography by LEED. In: Springer Ser. Chem. Phys. 2. Berlin, Heidelberg, New York: Springer 1979 10. Tong, S. Y.: A Review of Surface Crystallography by Low-Energy Electron Diffraction. In: Electron Diffraction 1927-1977. Dobson, P. J., Pendry, J. B., Humphreys, C. J. (eds.). London: The Institute of Physics, Bristol 1978, pp. 270-280 11. Van Hove, M. A.: Surface Crystallography and Bonding. In: The Nature of the Surface Chemical Bond. Rhodin, T. N., Ertl, G. (eds.). Amsterdam—New York—Oxford: North Holland 1979, pp. 275-312 12. Heinz, K., Müller, K.: LEED Intensities-Experimental Progress and New Possibilities of Surface Structure Determination. In: Structural Studies of Surfaces, Springer Tracts in Modern Physics, Vol. 91. Berlin, Heidelberg, New York: Springer Verlag 1982, pp. 1-53 13. Ibach, H., Mills, D. L.: Electron Energy Loss Spectroscopy and Surface Vibrations. New York: Academic Press 1982 14. Sexton, B.: Appl. Phys. A26, 1 (1981) 15. Thiry, P. A.: J. Electron Spectrosc. Relat. Phenom. 30, 261 (1983) 16. Froitzheim, H., Ibach, H., Lehwald, S.: Rev. Sei. Instrum. 46, 1325 (1975) 17. Evans, E„ Mills, D. L.: Phys. Rev. B5, 4126 (1972) 18. Ibach, H.: Surf. Sei. 66, 56 (1977) 19. Evans, E„ Mills, D. L.: Phys. Rev. B5, 4126 (1971) 20. Newns, D. M.: Phys. Lett. 60A, 461 (1977) 21. Persson, B. N. J.: Solid State Commun. 24, 573 (1977) 22. Ibach, H., Bruchmann, D.: Phys. Rev. Lett. 44, 36 (1980) 23. Froitzheim, H„ Ibach, H „ Lehwald, S.: Phys. Rev. Lett. 36, 1549 (1976) 24. Katz, J. E., Davies, P. W., Crowell, J. E., Somorjai, G. A.: Rev. Sei. Instrum. 53, 785 (1982) 25. Cabrera, A. L., Spencer, N. D., Kozak, E., Davies, P. W., Somorjai, G. A.: Rev. Sei. Instrum. 53, 20(1982) 26. Menadue, J. F.: Acta Cryst. A28, 1 (1972) 27. Maserd, N „ Kinniburgh, C. G „ Pendry, J. B.: J. Phys. C10, 1 (1977) 28. Aberdam, D.: Electron Diffraction in the Medium-Energy Range. In: Electron Diffraction 1927-1977. Dobson, P. J., Pendry, J. B., Humphreys, C. J. (eds.). London: The Institute of Physics, Bristol 1978, pp. 239-253 29. Smith, N. V.: Angular Dependent Photoemission. In: Photoemission in Solids I, Topics in Applied Physics, Vol. 26. Cardona, M., Ley, L. (eds.). Berlin, Heidelberg, New Y o r k : Springer 1978, p. 237 30. Feuerbacher, B., Fitton, B., Willis, R. F. (eds.): Photoemission and the Electronic Properties of Surfaces. New York: John Wiley & Sons 1978 31. Rosenblatt, D. H., Kevan, S. D., Tobin, J. G., Davis, R. F., Mason, M. G., Denley, D. R., Shirley, D. A., Huang, Y„ Tong, S. Y.: Phys. Rev. B26, 1812 (1982) 32. Barton, J. J., Bahr, C. C„ Hussain, Z„ Robey, S-W., Tobin, J. G., Klebanoff, L. E„ Shirley, D. A.: Phys. Rev. Lett., 51, 272 (1983) 33. Wilch, H . : Atomic and Molecular Scattering from Surfaces — Elastic Scattering. In: Topics in Surface Chemistry. Kay, E., Bagus, P. S. (eds.). New York: Plenum 1978, p. 135 34. Engel, T., Rieder, K. H.: Structural Studies of Surfaces with Atomic and Molecular Beam Diffraction. In: Structural Studies of Surfaces, Springer Tracts in Modern Physics, Vol. 91. Berlin, Heidelberg, New York: Springer-Verlag 1982, pp. 55-180 35. Savis, F. W., Van der Veen, J. F.: Analysis of Surface Structure and Composition by Ion-Scattering Spectroscopy. In: Proc. 7th Intern. Vac. Congr. & 3rd Intern. Conf. Solid Surfaces, Vienna, 1977, p. 2503 36. Feldman, L. C., Kauffman, R. L., Silverman, P. J., Zuhr, R. A., Barrett, J. H.: Phys. Rev. Lett. 39, 38 (1978)

216

C h a p t e r 3: B. E. Koel, G. A. Somorjai

37. Heiland, W., Taglauer, E.: Surf. Sei. 68, 96 (1977) 38. Bell, A. T., Hair, M. L. (eds.): Vibrational Spectroscopies for Adsorbed Species. Washington, D . C . : American Chemical Society 1980 39. Haller, G . T.: Catal. Rev. Sei. Eng. 23, 477 (1981) 40. C a u d a n o , R., Gilles, J. M., Lucas, A. A. (eds.): Vibrations at Surfaces. New Y o r k : Plenum 1982 41. Brundle, C. R., Morawitz, H . : Vibrations at Surfaces. A m s t e r d a m — O x f o r d —New Y o r k : Elsevier 1983 42. Dubois, L. H „ H a n s m a , P. K „ Somorjai, G. A . : Appl. Surf. Sei. 6, 173 (1980) 43. Yates, J. T., Jr., D u n c a n , T. M., Worley, S. D „ Vaughn, R. W . : J. Chem. Phys. 70, 1219 (1979) 44. Kroeker, R. M „ Kasha, W. C „ H a n s m a , P. K . : J. Catal. 61, 87 (1980) 45. Chiang, S., Tobin, R. G., Richards, P. L.: J. Electron Spectrosc. Relat. P h e n o m . 29, 113 (1983) 46. Chang, R. K., F u r t a k , T. E. (eds.): Surface Enhanced R a m a n Scattering. New Y o r k : Plenum 1981 47. C a m p i o n , A . : J. Electron Spectrosc. Relat. P h e n o m . 29, 397 (1983) 48. Hall, P. G., Wright, C. J.: Chemical Physics of Solids and Their Surfaces 7, 89 (1978) 49. Kroeker, R. M „ H a n s m a , P. K.: Catal. Rev.-Sci. Eng. 23, 553 (1981) 50. Low, M. J. D „ Parodi, G. A.: Applied Spectrosc. 34, 76(1980) 51. Madey, T. E., Yates, J. T „ Jr.: Surf. Sei. 63, 203 (1977) 52. Knotek, M. L„ Jones, V. D „ Rehn, V.: Phys. Rev. Lett. 43, 300 (1979) 53. Winograd, N., Garrison, B. J . : Acc. Chem. Res. 13, 406 (1980) 54. Madey, T. E., Netzer, F. P., H o u s t o n , J. E., H a n s o n , D. M., Stockbauer, R.: The Determination of Molecular Structure at Surfaces Using Angle Resolved Electron and P h o t o n Stimulated Desorption. In: Proceedings of DIET-I W o r k s h o p , Williamsburg, 1982. New Y o r k : Springer Verlag (1983) 55. Bradshaw, A. M., C e d e r b a u m , L. S., Domcke, W.: Ultraviolet Photoelectron Spectroscopy of Gases Adsorbed on Metal Surfaces. Structure and Bonding, Vol. 24. Berlin, Heidelberg, New Y o r k : Springer 1975, pp. 133-169 56. Gustafsson, T., Plummer, E. W . : Valence Photoemission f r o m Adsorbates. In: Photoemission and the Electronic Properties of Surfaces. Feuerbacher, B., Fitton, B., Willis, R. F. (eds.). L o n d o n : Wiley 1978, p. 353-380 57. Rye, R. R„ Madey, T. E „ Houston, J. E., Holloway, P. H . : J. Chem. Phys. 69, 1504 (1978) 58. Fuggle, J. C.: High Resolution Auger Spectroscopy of Solids and Surfaces. In: Electron Spectroscopy: Theory, Techniques and Applications, Vol. 4. Brundle, C. R., Baker, A. D. (eds.). New Y o r k : Academic Press (1981), pp. 8 5 - 1 5 2 59. Hagstrum, H. D . : Studies of Adsorbate Electronic Structure Using Ion Neutralization and Photoemission Spectroscopies. In: Electron and Ion Spectroscopies of Solids. Fiermans, L., Vennik, J., Dekeyser, W. (eds.). New Y o r k : Plenum 1978, pp. 273-323 60. J o h n s o n , P. D „ Delchar, T. A.: Surf. Sei. 77, 400 (1978) 61. C o n r a d , H., Ertl, G., Küppers, J., Wang, S. W „ Gerard, K., H a b e r l a n d , H . : Phys. Rev. Lett. 42, 1682 (1979) 62. Boiziau, C.: Scanning Electron Microsc. 1982, 949 73. Citrin, P. H „ Eisenberger, P.. Hewitt, R. C.: J. Vac. Sei. Technol. 15, 449 (1978) 64. Stohr, J., Jaeger, R., Brennan, S.: Surf. Sei. 117, 503 (1982) 65. Stern, E. A., Sayers, D. E „ Lytle, F. W.: Phys. Rev. B l l , 4836 (1975) 66. Stohr, J., Jaeger, R.: Phys. Rev. B26, 4111 (1982) 67. den Boer, M. L„ Einstein, T. L., Elam, W. T., Park, R. L., Roelofs, L. D . : J. Vac. Sei. Technol. 17, 59(1980) 68. Cowley, J. M . : Diffraction Physics. A m s t e r d a m : N o r t h Holland 1975 69. Isaacson, M. S., Langmore, J., Parker, N. W „ K o p f , D., Utlaut, M . : Ultramicroscopy 1. 359(1976)

Surface Structural Chemistry

217

70. Yagi, K., Takajanagi, K., Kobayashi, K., Osakabe, N., Tanishiro, Y., Honjo, G.: Electr 'i Micrsocopy, 1978, Vol. I (Proc. 9th Intern. Congr. on Ed. Micr., Toronto, 1978). Sturgess, J. M. (ed.). Toronto: Microscopical Society of Canada 1978 71. Yates, J. T., Jr.: The Thermal Desorption of Adsorbed Species. In: Methods of Experimental Physics, Vol. 22, Park, R. L. (ed.). New Y o r k : Academic Press (1985), pp. 4 2 5 ^ 6 4 72. Holzl, J., Schulte, F. K., Wagner, H.: Work Function of Metals. In: Springer Tracts Mod. Phys. 85, 1 (1979) 73. Van Hove, M. A., Koestner, R. J., Stair, P. C., Biberian, J. P., Kesmodel, L. L., Bartos, I., Somorjai, G. A.: Surf. Sci. 103, 189 (1981); 103, 218 (1981) 74. Adams, D. L., Nielson, H. B., Van Hove, M. A., Ignatiev, A.: Surf. Sci. 104, 47 (1981) 74a. King, D. A.: Physica Scripta T4, 34 (1983) 75. Tong, S. Y„ Maldonado, A. L.: Surf. Sci. 78, 459 (1978) 76. Duke, C. B.: Appl. Surf. Sci. 11/12, 1 (1982) 77. Chadi, D. J.: Phys. Rev. Lett. 41, 1062 (1978) 78. Northrup, J. E., Cohen, M. L.: Phys. Rev. Lett. 49, 1349 (1982) 79. Eastman, D. E.: J. Vac. Sci. Technol. 17, 492 (1980) 80. Legg, K. O., Jona, F., Jepsen, D. W., Marcus, P. M.: Phys. Rev. B16, 5271 (1977) 81. Shih, H. D„ Jona, F„ Jepsen, D. W„ Marcus, P. M. : Phys. Rev. Lett. 46, 731 (1981) 82. Shih, H. D., Jona, F., Jepsen, D. W„ Marcus, P. M.: Surf. Sci. 60, 445 (1976) 83. Strong, R. L., Firey, B., de Wette, F. W„ Erskine, J. L.: J. Electron Spectrosc. Relat. Phenom. 29, 187 (1983) 84. Behm, R. J., Penka, V., Cattania, M.-G., Kristmann, K., Ertl, G.: J. Chem. Phys. 78, 7486 (1983) 85. Sachtler, J. W. A., Van Hove, M. A., Biberian, J. P., Somorjai, G. A.: Surf. Sci. 110, 19 (1981) 86. Somorjai, G. A.: Chemistry in Two Dimensions, Surfaces. Ithaca and London: Cornell University Press (1981) 87. Yates, J. T., Jr., Madey, T. E.. Campuzano, J. C. In: The Physics and Chemistry of Solid Surfaces and Heterogeneous Catalysis. King, D. A., Woodruff, D. P. (eds.), in press 88. Kesmodel, L. L„ Baetzold, R., Somorjai, G. A.: Surf.. Sci. 66, 299 (1977) 89. Casalone, G., Cattania, M. G., Simonetta, M.: Surf. Sci. 103, L121 (1981) 90. Casalone, G., Cattania, M. G., Merati, F„ Simonetta, M.: Surf. Sci. 120, 171 (1982) 91. Felter, T., Weinberg, W. H.: Surf. Sci. 103, 265 (1981) 92. Lloyd, D. R„ Quinn, C. M„ Richardson, N. V.: Angle-Resolved Ultraviolet Electron Spectroscopy of Clean Surfaces and Surfaces with Adsorbed Layers. In: Surface and Defect Properties of Solids, Vol. VI. Roberts, M. W., Thomas, J. M. (ed.). London: Chemical Society (1976), pp. 179-217 93. Nyberg, G. L„ Richardson, N. V.: Surf. Sci. 85, 335 (1979) 94. Dubois, L. H„ Somorjai, G. A.: Surf. Sci. 91, 514 (1980) 95. Crowell, J. E., Somorjai, G. A.: Appl. Surf. Sci. 19, 73 (1984) 96. Yang, A. C„ Garland, C. W.: J. Phys. Chem. 61, 1504 (1957) 97. Koestner, R. J., Van Hove, M. A., Somorjai, G. A.: Surf. Sci. 107, 439 (1981) 98. Van Hove, M. A., Koestner, R. J., Frost, J. C., Somorjai, G. A.: Phys. Rev. Lett. 50, 903 (1983) 99. Chini, P., Longoni, V., Albano, V. G.: Adv. Organomet. Chem. 14, 285 (1976) 100. Castner, D. G „ Sexton, B. A., Somorjai, G. A.: Surf. Sci. 71, 519 (1978) 101. Thiel, P. A., Williams, E. D„ Yates, J. T., Jr., Weinberg, W. H.: Surf. Sci. 84, 54 (1979) 102. Koel, B. E., Somorjai, G. A.: J. Electron Spectrosc. Relat. Phenom. 29, 287 (1983) 103. Lin, R. F., Koestner, R. J., Van Hove, M. A., Somorjai, G. A.: Surf. Sci., 134, 161 (1983) 104. Van Hove, M. A., Lin, R. F „ Koestner, R. J., Koel, B. E., Mate, M., Crowell, J. E„ Somorjai, G. A.: LEED and HREELS Studies of Benzene Adsorbed on Rh(l 11). Proceed. Interdisc. Surf. Sci. Conf. (ISSC-6) April, 1983. University of Warwick, UK, in press 105. Van Hove, M. A., Lin, R. F„ Somorjai, G. A.: Phys. Rev. Lett., 51, 778 (1983)

218

Chapter 3: B. E. Koel, G. A. Somorjai

106. Koel, B. E., Crowell, J. E„ Mate, C. M „ Somorjai, G. A.: J. Phys. Chem. 88, 1988 (1984) 107. Lehwald, S., Ibach, H „ Demuth, J. E.: Surf. Sei. 78, 577 (1978) 108. Nieuwenhuys, B. E., Hagen, D. I., Rovida, G., Somorjai, G. A.: Surf. Sei. 59, 155 (1976) 109. Stair, P. C„ Somorjai, G. A.: J. Chem. Phys. 67, 4361 (1977) 110. Koestner, R. J., Van Hove, M. A., Somorjai, G. A.: J. Phys. Chem. 87, 203 (1983) 111. Kesmodel, L. L., Dubois, L. H., Somorjai, G. A.: J. Chem. Phys. 70, 2180 (1979) 112. Koestner, R. J., Van Hove, M. A., Somorjai, G. A.: Surf. Sei. 121, 321 (1982) 113. Skinner, P., Howard, M. W„ Oxton, I. A., Kettle, S. F. A., Powell, D. B., Sheppard, N.: J. Chem. Soc., Faraday Trans. 2, 77, 1203 (1981) 114. Dubois, L. H „ Castner, D. G., Somorjai, G. A.: J. Chem. Phys. 72, 5234 (1980) 115. Koel, B. E„ Crowell, J. E„ Bent, B. E„ Mate, C. M „ Somorjai, G. A.: J. Phys. Chem., to be published 116. Ibach, H., Mills, D. L.: Electron Energy Loss Spectroscopy and Surface Vibrations. New York: Academic Press (1982), p. 326 117. Gates, J. A., Kesmodel, L. L.: Surf. Sei. 124, 68 (1983) 118. Minot, C., Van Hove, M. A., Somorjai, G. A.: Surf. Sei. 127, 441 (1982) 119. Avery, N., et al.: to be published and private communication

Subject Index

Acetylene, adsorbed, structure 187 Adsorption, metal on metal 185 Alkenes, adsorbed, structure 203 —, —, vibrational spectra 203 Alloy oxidation 125 Alumina, structure by electron microscopy 133 Ammonia, first synthesis 8 —, oxidation 42 Ammonia decomposition 8 Ammonia equilibrium, atmospheric pressure 9-11, 12-13 Ammonia synthesis, coal feedstock 47 —, development in Britain 44 —, development in France 44 —, early catalysts 15 —, early development 18 —, early technical problems 24-30 —, effect of impurities 29 —, first commercial plant 37 —, first practical demonstration 22 —, high pressure apparatus 31-33 —, influence of First World War 42 —, integrated plant 46-4 —, Ludwigshafen plant 41 —, naphta feedstock 46-4 —, natural gas feedstock 46 —, Oppau plant 38-40 —, promoted iron catalyst 30 —, pure gas supply 34 —, recirculatory system 22 —, scale-up 35-37 —, search for a practical catalyst 26-30 —, technical 14—16 Anatase, structure by electron microscopy 136 Atomic beams, surface diffraction 174 Auger analysis, coked catalyst 120

—, scanning 70 Auger electron spectroscopy

161

Back scattered electrons 62 —, energy distribution 163 —, de Broglie wavelength 163 Benzene, adsorb on Rh(l 11) 194-201 Birkland-Eyde process 3 Bond lengths, adsorbed CO 187 —, adsorbed atoms 184 Bosch, Carl, biography 16-25 Brightness, electron microscope sources 59 Butenes, adsorbed, vibrational spectra 204 Carbon, filaments 131 —, graphitized 129 Carbon monoxide, adsorbed, bond lengths 186, 192 —, —, structure 188-194 —, —, vibration frequencies 186 —, —, vibrational spectra 175 —, adsorption site 186 Carbonaceous residues 207 Catalysts, for ammonia synthesis 26-30 Charge transfer, adsorbed atoms 184 Charging, of specimens 74 Clusters, resolution 104-108 Cobalt sulfide 142 Coke, on Fischer-Tropsch catalyst 131 —, on hydrodesulfurization catalyst 129 —, structure, by electron microscopy 127-133 Coked catalyst, Auger analysis 120 —, electron microscopy 120 Contamination 79 Contrast, in transmission electron microscopy 85 Contrast-transfer-function 96

220

Subject Index

Controlled atmosphere electron microscopy 118

Damage, to specimens 78 D a r k field imaging 89 —, out of focus 93 D a r k field methods 90-94, 106 Decoration, steps 114 Defect of focus 56 Defocus, optimum 105 Desorption energy, carbon monoxide on Pt 210 Diffraction, optical 72 Diffraction contrast 56, 86 —, orientation effect 88 Dipole scattering 169 Electron absorption 85 Electron diffraction 67-68 Electron energy spread 68 Electron guns 59 Electron microscope, sources 59, 60 —, use modes 66 Electron microscope sources, resolution 60 Electron microscopy, in-situ experiments 63 —, schematics 58, 59 —, structure of surfaces 177 Electron scattering 56 —, in specimens 68 Elemental analysis 99-101 Energy spread, electron microscope sources 59 Ethylidyne surface species, structure 201-202 — , vibrational frequencies 201 Extinction frings 88 Ferrierite, electron microscopy Fixation, nitrogen 2 Fourier images 98

137

Gas purification, for ammonia synthesis 37-38 Gold clusters 106 Gold particles 109 Gold surface, reconstruction 178 Graphimets 141 Graphite, formation on catalysts 132 —.oxidation 113 Graphite-metal reaction 126 Graphitization 129 Grazing incidence illumination 115 Guano 2 Haber, Fritz, biography 4 - 7 Heat of adsorption, carbon monoxide on metals 211-212 — , hydrogen on metals 213-214

High pressure apparatus, for ammonia synthesis 31-33 —, effect of hydrogen 32 High resolution electron energy loss spectroscopy 167-171 —, experiment 168 - , H, D on W (100) 170 —, mechanism 169 —, resolution 171 —, selection rule 170 Hollow-cone illumination 91, 107 Hydrocarbon residues, structure on Pt (III) 207 Hydrodesulfurization catalysts 107 —, coked 129 —, structure, by electron microscopy 142-144 Icosahedron 111 Image formation 56 —, scanning transmission electron microscopy 98 Image processing 101 Imaging, scanning electron microscopy 80 —, thickness effect 89 —, transmission electron microscopy 85 et seq. Impact scattering 170 Incoherent illumination 106 Inelastic electron tunneling spectroscopy 174 Intercalation 130 Ion desorption methods 176 Ion scattering, angular distribution 174 —, from surfaces 174 Iridium surface, reconstruction 178 L a n t h a n u m phosphate 145 Lattice imaging 80, 94, 110 Lenses, for electron microscope 60 Low energy electron diffration 161-167 - , benzene on R h ( l 11) 195-198 —, C O o n R h ( l l l ) 189-194 - , P t ( l l l ) 166 —, three-dimensional 165 —, two-dimensional 164 Matrix notation, superlattice 161 Mean free path, electrons in solids 162 Metal-graphite reaction 126 Metallic clusters, C O vibration frequencies 186 —, in zeolites, by electron microscopy Metallic particle size, by electron microscopy 122 Metallic particles, movement, by electron microscopy 125

221

Subject Index

—, scanning transmission electron microscopy 72 —, single atoms 102-104 —, transmission electron microscopy 71 Rutherford backscattering 174

—, redispersion 124 —, sintering, by electron microscopy 123 Microanalysis, by scanning transmission electron microscopy 147 Microdiffraction, by scanning transmission electron microscopy 146 Microscopies, comparison 55 Molecular beams, surface diffraction 174 Molybdenum sulfide 108, 112, 142-144 Movement, metallic particles, by electron microscopy 125 Multiply twinned particles 111 Nascent state theory 8 Neutron inelastic scattering 176 Nitric acid, syntehesis 42 Nitrogen fixation 2 —, arc process 20 —, by barium cyanamide 20 —, by calcium cyanamide 3 Nitrogen fixation, efficiency of processes —, by electric arc 3 Nitrogenous fertilizer 2

48

Optical diffraction 72 Optical microscopy 53 Particle shape, by electron microscopy 108-111 Particle size, in catalysts, by electron microscopy 121-125 Particle thickness, by electron microscopy 109 Particles, multiply twinned 111 Pellets 76 Phase contrast 56, 94-98, 106, 114 Photoacoustic spectroscopy 176 Platinum-rhodium gauze, electron microscopy 120 Platinum surface, reconstruction 178 Polyolefin polymer, electron microscopy 120 Projected shape 108 Pure gas supply, for ammonia synthesis 34 Raman scattering, surface enhanced 175 Reconstruction, in adsorption 185 - , P t ( 1 0 0 ) 179 Redispersion, of metallic particles 124 Reconstruction, Si (100) 180 —, surface 178 Relaxation, surface bonds 178, 180 Replicas 76 Resolution, clusters 104-108 —, in scanning electron microscopy 69, 83

Scanning electron microscope 54, 61 Scanning electron microscopy, image formation 80 —, resolution 69, 83 —, specimens 73 Scanning transmission electron microscope 54, 64-65 Scanning transmission electron microscopy, image formation 98 —, resolution 72 —, single atom resolution 103 Secondary electrons 61 Secondary emission 162 Size, electron microscope sources 59 Size distribution, determination 77 Silica, structure by electron microscopy 133 Silicon, steps 116 Single atoms, resolution 102-104 Sintering, in catalysts, by electron microscopy 122-123 Specimen, charging 74 —, scanning electron microscopy 73 —, transmission electron microscopy 74-77 Specimen damage 78 Step decoration 114 Stepped-kinked surfaces, structure 182 Stepped surfaces 180-182 Steps, by grazing incidence illumination 115 Steps on silicon 116 Superlatice 161 Surface extended X-ray absorption fine structure 177 Surface sites, on low index surfaces 183 Surface steps, by electron microscopy 112-117 Surface structure, techniques 173 —, by transmission electron microscopy 146 Surface topography 76 —, from scanning electron microscopy 81 Techniques, for surface structure 173 Tellurium molybdate 145 Thermal desorption spectra, from adsorbed alkenes 205 —, carbon monoxide on Pt 210 —, hydrogen on Pt 209 Thickness effect, in imaging 89 Transmission electron microscope 54, 62

Subject Index

222 Transmission electron microscopy, image formation 85 et seq. —, resolution 71 —, specimens 74-77 Ultraviolet photoelectron spectroscopy

Water gas shift reaction, for gas purification 37 Wood notation, superlattice 161 176

Vacuum systems, for electron microscope 60 Vibration frequencies, adsorbed CO 186 —, CO in metallic clusters 186 —, surface ethylidyne 201 —, adsorbed alkenes 203 Vibrational spectra, adsorbed CO 175 —, benzene on Rh(l 11) 198-199 —, carbon monoxide on R h ( l l l ) 188 —, deuterated ethylidyne 206 Vitreous carbon 130

X-ray X-ray —, in —, in X-ray

absorption near edge structure 177 analysis 99-101 electron microscope 60 scanning electron microscopy 69 photoelectron spectroscopy 176

Zeolite L, lattice image 138 Zeolite ZSM-5, scanning electron microscopy 119 Zeolites ZSM-5/11, electron microscopy 139 —, lattice image 140 Zeolites, structure by electron microscopy 136-140

Author Index Volume 1-7

Aika, K. see Ozaki, A.: Vol. 1, p. 87 Boehm, H.-P., Knozinger, H.: Nature and Estimation of Functional Groups on Solid Surfaces. Vol. 4, p. 39 Boreskov, G. K.: Catalytic Activation of Dioxygen. Vol. 3,0. 39 Butt, J. B.: Catalyst Deactivation and Regeneration. Vol. 6, p. 1 Donath, E. E.: History of Catalysis in Coal Liquefaction. Vol. 3, p. 1 Dry, M. E.: The Fischer-Tropsch Synthesis. Vol. 1, p. 159 Ertl, G.: Kinetics of Chemical Processes on Well-defined Surfaces. Vol. 4, p. 209 Foger, K.: Dispersed Metal Catalysts. Vol. 6, p. 227 Froment, G. F., Hosten, L. H.: Catalytic Kinetics: Modelling. Vol. 2, p. 97 Gallezot, P.: X-Ray Techniques in Catalysis. Vol. 5, p. 221 Garin, F. F. see Maire, G. L. C. Vol. 6, p. 161 Giannini, U. see Pasquon, I. Vol. 6, p. 65 Haber, J. : Crystallography of Catalyst Types. Vol. 2, p. 13 Heinemann, H. : A Brief History of Industrial Catalysis. Vol. 1, p. 1 Hosten, L. H. see Froment, G. F. Vol. 1, p. 97 Knor, Z.: Chemisorption of Dihydrogen. Vol. 3, p. 231 Knozinger, H. see Boehm, H.-P. Vol. 4, p. 39 Koel, B. E., Somorjai, G. A.: Surface Structural Chemistry Vol. 7, p. 159 Lecloux, A. J. : Texture of Catalysts. Vol. 2, p. 171 Maire, G. L. C., Garin, F. G . : Metal Catalysed Skeletal Reactions of Hydrocarbons. Vol. 6, p. 161 Morrison, S. R.: Chemisorption on Nonmetallic Surfaces. Vol. 3, p. 199 Ozaki, A., Aika, K.: Catalytic Activation of Dinitrogen. Vol. 1, p. 87 Pasquon, /., Giannini, U.: Catalytic Olefin Polymerization. Vol. 6, p. 65 Peri, J. B.: Infrared Spectroscopy in Catalytic Research. Vol. 5, p. 171 Rostrup-Nielsen, J.: Catalytic Steam Reforming. Vol. 5, p. 1 Rylander, P. N.: Catalytic Processes in Organic Conversions. Vol. 4, p. 1 Sanders, J. V. : The Electron Microscopy of Catalysts Vol. 7, p. 51 Schwab, G.-M.: History of Concepts in Catalysis. Vol. 2, p. 1 Sinfelt, J. H. : Catalytic Reforming of Hydrocarbons. Vol. 1, p. 257 Somorjai, G. A. see Koel, B. E. Vol. 7, p. 159 Tanabe, K. : Solid Acid and Base Catalysts. Vol. 2, p. 231 Taylor, K. C. : Automobile Catalytic Converters. Vol. 5, p. 119 Topham, S. A.: The History of the Catalytic Synthesis of Ammonia Vol. 7, p. 1 Turner, J..C. R.: An Introduction to the Theory of Catalytic Reactors. Vol. 1, p. 43 Vannice, M. A.'. Catalytic Activation of Carbon Monoxide on Metal Surfaces. Vol. 3, p. 139