Nobel Lecture, December 11, 1912
The Method of Direct Hydrogenation by Catalysis
I find it most moving to be speaking to you within these precincts, immortalized as they are by so many glories, and I should find it difficult to overcome my emotion if I were not aware of the kindness with which the Royal Swedish Academy welcomes those whom it has judged worthy of its world-famous rewards.
I have to talk to you on the general method of direct hydrogenation by catalysis, which I established some dozen years ago in collaboration with Senderens and which I have continued to develop and generalize with the assistance of my devoted pupils, Mailhe and Murat.
I have already on several occasions – in publications and lectures – given a detailed description of how this method is used in practice and I shall not inflict upon you an exhaustive technical account of the procedures, or the long list of the wide variety of processes which can be carried out with it.
I prefer to describe how I came upon this new method, what its principal applications are at the present time and how its intimate mechanism can be explained theoretically.
The earlier methods of hydrogenation were almost all based on the use of what is known as nascent hydrogen, i.e. chemical systems which, when used alone, in the absence of a reducible substance, release hydrogen as well as heat energy. In the presence of transformable substances such systems will transmit to them a portion of the hydrogen supplied.
These hydrogenating systems are well known. Some proceed in an alkaline medium, such as sodium or an amalgam of sodium in contact with water or with an alcohol (classical method of transforming aldehydes or acetones into corresponding alcohols; Bouveault’s method whereby the ester of a fatty acid can be changed into the alcohol corresponding to this acid).
Others, still more widely used, proceed in an acid medium, such as zinc, iron or tin reacting with dilute sulphuric, hydrochloric or acetic acid (method of preparing aniline from nitrobenzene, etc.). A still more powerful agent belongs to this group, namely a concentrated solution of hydroiodic acid, the use of which we owe to Berthelot. When heated to a high temperature in a sealed tube this solution brings about intense splitting of the hydroacid, the hydrogen of which may become fixed to an organic substance. However, the use of sealed tubes containing hydroiodic acid is very dangerous for the chemist, since the pressure of the free hydrogen liberated in the tube may exceed 100 atmospheres, and terrible accidents due to bursting of the tubes have been all too frequent. The use of these tubes also has another disadvantage, this time of a purely chemical nature; the presence of concentrated hydroiodic acid often causes isomerization of the reducible substance. Thus Berthelot, using benzene, produced not cyclohexane, which he had expected, but an isomer, methylcyclopentane.
The direct use of gaseous hydrogen in hydrogenating processes had been quite exceptional and was confined to a small number of cases where the catalytic properties of finely divided platinum, sponge or black, had been used, properties which had proved so valuable in direct oxidation reactions. These few applications were the only ones discovered during the whole of the 19th century.
In 1838 Kuhlmann transformed nitrogen oxide into ammonia in contact with platinum sponge; in 1852 Corenwinder succeeded in the same way in combining iodine with hydrogen.
In 1863 Debus transformed hydrocyanic acid into methylamine with platinum black; however, this reaction is fugitive and soon ceases because cyanidation of the metal rapidly destroys its capacity to induce the reaction.
In 1874 De Wilde found that by placing platinum black in a test-tube containing a suitable mixture of hydrogen and acetylene he could induce combination in the cold and produce ethylene or ethane, depending on the proportions of the initial substances.
Such was the very rudimentary stage reached in direct hydrogenation when, more than fifteen years ago, I entered this field.
Mond, Langer and Quincke had just isolated nickel carbonyl by the action of carbon monoxide on powdered nickel, which had been prepared by reduction of its oxide; iron had given an analogous compound. Thus the incomplete carbon monoxide molecule was being added to metals.
It occurred to me that it might be possible to add metals to other incomplete gaseous’molecules, such as nitrogen oxide, in the same way. In collaboration with Senderens I tried to make this gas react with various metals which had been recently reduced from their oxides, but in no case did I observe anything other than oxidation of the metal. We were more fortunate with nitrogen peroxide, and in about 1894 we succeeded in inducing regular fixation to copper, cobalt, and nickel, which thus yielded true nitro metals.
Moissan had just found, in his calcium carbide, an easy means of preparing pure acetylene, which has a very incomplete molecule, and he thought that this substance might, like nitrogen peroxide, become fixed to metals and yield addition products directly. But his attempts to bring this about in 1896 with Moureu produced a totally different, though very interesting, result. Acetylene, when brought into contact in the cold with freshly reduced nickel, cobalt or iron or even with platinum black, immediately decomposes with intense incandescence. Deposition of large quantities of carbon occurs together with evolution of gas – which Moissan thought to be hydrogen – and formation of liquid products, which appeared to him to be benzene accompanied by other aromatic hydrocarbons.
He explained this curious reaction as being due to physical condensation of the acetylene in the pores of the metal, this condensation producing locally a temperature sufficiently high to break down part of the acetylene into hydrogen and carbon and to transform the rest into benzene and other hydrocarbons, as in the famous experiment of Berthelot.
A more thorough examination would have changed this opinion; the condensed liquids are in reality very different from those yielded by simple polymerization of acetylene at dull-red heat; the gas evolved is not hydrogen but contains large proportions of ethane.
Moissan, whose earlier work and inclinations had kept him well away from the field of organic chemistry, thought no more about this reaction, which he considered merely as a nice demonstration experiment. For me, on the other hand, it suggested great possibilities.
My own ideas on the mechanism of catalytic phenomena were very different from those at one time commonly held, ideas which I no doubt owed to the influence of the illustrious teacher who had guided my first steps in chemistry nearly twenty years before – I refer, of course, to Berthelot.
I thought and I still think (I shall be returning to this point later on) that the decisive cause of the catalytic activity of porous platinum is not a simple process of physical condensation producing a local rise in temperature but that it is a real chemical combination of the surface of the metal with the surrounding gas.
I attributed the decomposition of the acetylene in the experiment of Moissan and Moureu to an affinity of the metal, either for the acetylene itself or for the constituents of the latter, carbon or hydrogen, which it is apparently capable of loosening from the endothermic molecule of this hydrocarbon.
Having made certain that Moissan was not thinking of continuing the study of this reaction. I took it up myself, and first of all together with Senderens I made a similar test on ethylene.
When a stream of ethylene is directed on to nickel, cobalt or iron, which has been freshly reduced and kept in the region of 300°C, intense incandescence of the metal with deposition of large quantities of carbon due to breakdown of the ethylene occurs. However, the gas which leaves the apparatus is not hydrogen but consists mainly of ethane.
This latter could arise only from hydrogenation of undecomposed ethylene, and no doubt this hydrogenation has been induced by the metal. In fact, if a mixture of ethylene and hydrogen is directed on to a column of reduced nickel, the ethylene is changed into ethane and the same metal can be used indefinitely to bring about the same transformation (June 1897).
Thus, nickel appeared to us to possess a remarkable capacity to hydrogenate ethylene without itself being visibly modified, i.e. by acting as a catalyst.
During 1898 I suffered a cruel bereavement which for many months made it quite impossible for me to do any useful work.
In 1899 we succeeded in establishing that reduced nickel possesses the same hydrogenating activity towards acetylene, which, under similar conditions, is transformed into ethane, and the following year, in 1900, we found that reduced cobalt, iron and copper, as well as powdered platinum, possess a similar, though less vigorous, property.
This catalytic hydrogenating power of nickel appeared to me so perfect that I then thought of using it for a major reaction, in which the various known hydrogenating agents had shown themselves ineffective, i.e. the hydrogenation of benzene.
Decisive success came at the end of 1900 when, with Senderens, I found that benzene can be totally changed into cyclohexane in contact with nickel at a temperature of about 180°C. After that I was absolutely confident of the general nature of the method, the principle of which we stated at the beginning of 1901:
“Vapour of the substance together with an excess of hydrogen is directed on to freshly reduced nickel held at a suitable temperature (generally between 150 and 200°C).”
This very simple process- which requires a minimum of equipment, is quite safe and usually needs but little supervision- then seemed to be applicable to a whole series of important cases, namely the transformation of unsaturated ethylenic or acetylenic hydrocarbons into saturated hydrocarbons, and the transformation of nitro derivatives into arnines.
Completely confident of success I undertook the generalization of this method, another advantage of which seemed to be that in most cases it yielded only the desired main product, without by-products or isomerization, and consequently gave extremely high yields.
During the period 1901 to 1905, together with Senderens, I showed that nickel is very suitable for the direct hydrogenation of nitriles into amines and, no less important, of aldehydes and acetones into corresponding alcohols. Carbon monoxide and carbon dioxide are both changed immediately into methane, which can therefore be synthesized with the greatest ease.
Like benzene, the homologous aromatic hydrocarbons, toluene, xylenes, etc. fix directly six hydrogen atoms to produce the corresponding cyclic compounds; phenol is transformed into cyclohexanol, aniline into cyclohexylamine. Naphthalene and acenaphthene can also be usefully hydrogenated with nickel.
Tetravalent terpenes, such as limonene, take four atoms of hydrogen whilst divalent terpenes (pinene, camphene) can fix only two, in accordance with the predictions made in the excellent work of Wallach.
From 1904 to 1911 together with my pupil Mailhe I continued to develop the method, applying it with success to the direct hydrogenation of the aromatic halogen derivatives, which return to hydrocarbons; to the hydrogenation of ally1 products and of unsaturated acids, which turn into saturated molecules without altering their functions; to the hydrogenation of oximes, amides, isocyanic esters, carbylamines, diketones and quinones, and to that of cresols, xylenols, etc.
Hydrogenation of the aromatic nucleus in the diphenols, pyrogallol and finally in benzylamine was a more delicate matter, although here too we succeeded in producing the corresponding cyclohexane compounds.
Some gaps still remained, and in collaboration with my assistant Murat I set out less than a year ago to fill these in. On several occasions, despite very great care, I had failed to bring about direct hydrogenation of benzoic acid and its esters. I could not bring myself to accept this ineffectiveness of the method, and by controlling its application in minute detail we succeeded, a few months ago, in producing hexahydrobenzoic esters very efficiently in this way.
Recently, too, we have successfully returned to the hydrogenation of diphenyl, and where Eykman was able to obtain only phenylcyclohexane we obtained normal production of dicyclohexyl.
As soon as it had been described, however, our method caught the imagination of many research workers both in France and abroad, and their efforts contributed towards an extension of its use. It would take too long to enumerate their individual works, but I must at least mention the names of Darzens, Brunel, Godchot, Leroux, Breteau, Willstätter, Padoa and others, as well as those of my pupils Gaudion and Mignonac.
Two fundamental conditions are essential for the success of the method, namely purity of the substances and choice of a suitable temperature.
Powdered nickel, which, as catalyst, is the most important agent in the reaction, is comparable in every way with a ferment and, as in the case of the living organisms which constitute ferments, infinitesimal doses of certain substances are sufficient to attenuate and even suppress altogether their functional activity. Traces of sulphur, bromine or iodine are true poisons for the metal ferment, which reduced nickel in fact is, and every care must be taken to ensure that these are not present in the metal itself, in the hydrogen (as a result of insufficient purification), or in the substance subjected to hydrogenation.
Benzene from which every trace of thiophene has not been eliminated cannot be transformed into cyclohexane.
A little bromine vapour which escaped in a large laboratory where there was some phenol in an unstoppered flask was sufficient to render this phenol totally incapable of hydrogenation.
The temperature of the metal should also be carefully regulated since each hydrogenation process takes place profitably only within a well-defined temperature range.
Enormous differences have, moreover, been observed in the facility with which the various hydrogenation processes are effected, some being easy, others very delicate.
As typical easy hydrogenation processes I might mention those of the ethylenic hydrocarbons or of the nitro derivatives. These processes can be effected by more or less any nickel, even if somewhat poisoned, and they take place within very wide temperature ranges.
Hydrogenations are much more difficult in the case of the aromatic nucleus, of phenol and still more of benzene, and this difficulty reaches its maximum in the case of diphenols, pyrogallol and above all benzoic esters. Here it is essential to use high-purity nickel, which has been prepared at a not too high temperature, and to keep the reaction temperature within strictly controlled limits. The use of heating apparatus fitted with temperature regulators is strongly recommended here.
Among the extremely numerous results already obtained from the method of direct hydrogenation by catalysis I shall confine myself to mentioning a few particularly interesting cases.
(1) Direct hydrogenation of acetylene with nickel provided us, Senderens and me, with some very curious results, which differ a great deal with the conditions under which they are obtained.
In the cold, in the presence of an excess of hydrogen, acetylene changes into ethane only. But if the temperature of the metal is raised to more than 200°C, the metal causes the molecule to split up into CH groups, hydrogenation of which then yields, beside free methane, methylene and methyl groups, which unite in various ways, giving a certain quantity of a liquid whose fluorescence, odour, density and chemical composition are those of Pennsylvanian petroleum.
By bringing acetylene alone into contact with nickel we get incandescence, as observed by Moissan, which breaks part of it down into carbon and hydrogen and changes the rest into aromatic hydrocarbons, benzene and its homologues; these last can, in turn, under the action of the nickel situated at a greater distance from them, be reduced by the hydrogen arising from the splitting and be transformed into cyclic hydrocarbons, and this transformation may be total if an excess of hydrogen is allowed to react beyond the point of incandescence. A liquid which in its properties and compositions is entirely similar to Baku petroleum is then obtained by condensation.
By the same procedure, but with acetylene mixed with a certain proportion of hydrogen, an intermediate product, consisting of both straight-chain and cycle hydrocarbons, resembling Rumanian petroleum, is obtained.
If the column of metal intended for hydrogenation of the products of incandescence is raised to a temperature of more than 300°C, then a proportion of the aromatic hydrocarbons is retained and the liquid obtained is analogous to Galician petroleum.
From these syntheses I have deduced a general theory on the natural production of mineral oils in the interior of the Earth; it confirms the earlier views of Mendeleeff and Berthelot and it has the advantage over other theories that it easily explains the diversity of the products while attributing to them a more or less similar origin.
(2) By direct hydrogenation vapours of liquid fatty acids (oleic acid) can be transformed into solid acids (stearic acid); and it has been found that this reaction, which is very easy to bring about, can be used to advantage with the oils themselves, the catalyst metal being held in suspension in the presence of gaseous hydrogen, which changes them into solid fats. This is at present the basis of a great industry in Britain and Germany.
(3) Continuous transformation of nitrobenzene into aniline can be induced conveniently by catalysis; here, copper is to be preferred to nickel because it never brings about hydrogenation of the aniline formed and is not very sensitive to accidental impurities.
(4) Direct hydrogenation of ordinary acetone produces, with a very high yield and without any incidental disturbance, isopropyl alcohol, the cost of which is thus tremendously reduced. The classical method of hydrogenation by means of sodium amalgam and water gave only a very mediocre yield.
(5) The hydrogenating action exerted by nickel on carbon monoxide can be used on a large scale for the production of methane or of gaseous mixtures rich in methane. By using water gas prepared at a suitable temperature, and with elimination of carbon dioxide, we can obtain at very low cost a gas of high calorific power and, because it no longer contains any trace of carbon monoxide, completely non-toxic.
(6) The most valuable advantage of the method of hydrogenation is indisputably the fact that it enables the aromatic nucleus to be hydrogenated, since it is thus a very easy matter to prepare the whole series of cyclohexane compounds, which hitherto could be obtained only with great difficulty by expensive and complicated methods. This is the starting-point of an infinite number of syntheses of great theoretical and even practical interest, for with these processes it is possible to produce many artificial perfumes.
Cyclohexanol and paramethylcyclohexanol, which are produced so easily by direct hydrogenation of phenol and paracresol, seem destined to play a useful role in the synthesis of rubber – a problem at present engaging so many research workers.
How can we explain hydrogenation by catalysis? I assume that hydrogen acts upon the metal by very rapidly producing a compound on its surface. The hydride thus produced is readily and rapidly dissociated, and if it is placed in the presence of substances capable of using hydrogen it gives it up to them, at the same time regenerating the metal, which again produces the same effect, and so on.
The distinction which I have made between several kinds of activity of nickel would suggest that there are several stages of combination. For instance, well-prepared and pure nickel might produce the perhydride NiH2 which is capable of hydrogenating benzene. On the other hand, nickel prepared at too high a temperature or containing some impurities, would give only a poorer hydride, such as
Ni — H
Ni — H
which would be incapable of reacting with benzene, though active in respect of ethylenic hydrocarbons or nitro derivatives.
If this theory is correct, it can be deduced that nickel and the other active metals (copper, cobalt, iron, platinum) must be capable not only of fixing free hydrogen, but must also take hydrogen from such substances as are capable of giving it up, and that consequently they must be dehydrogenation catalysts.
This is in fact the case. At a temperature of between 250° and 300°C powdered copper very conveniently brings about dehydrogenation of primary alcohols into aldehydes and of secondary alcohols into acetones, and this is a very useful practical method of effecting this transformation.
Thus the concept of a temporary compound prompted me to use finely divided metals, first as hydrogenation catalysts and then as dehydrogenation catalysts.
For several years I have been carrying out, together with Mailhe, research inspired by the same idea.
Some ten years ago Ipatief had pointed out that alumina readily splits alcohols into ethylenic hydrocarbons and water by catalysis. I have established that this property is also possessed by other oxides, especially blue tungsten oxide and thorium oxide. This latter can be used indefinitely for dehydrating an alcohol; when it has become fouled through prolonged use and its activity has thereby been reduced, all that is necessary is to oxidize it to restore all its original catalytic power.
Here again I turn to the temporary formation of a compound for explanation. I liken the part played by thoria in respect of the alcohols to that of sulphuric acid in the classical mechanism, known as Williamson’s method, for the production of diethyl ether or ethylene. I assume that a kind of alcohol thorinate, an ester comparable with the acid alcohol sulphate, is produced at the surface of the oxide and from this I deduce that this thorinate must, like the sulphate, react with various substances and thus induce, as a catalyst, various reactions with alcohols.
My conjectures have proved correct. When in contact with thoria, and without any apparent modification in this latter substance, the vapours of the alcohols react directly with hydrogen sulphide to give thiols, with ammonia to give primary amines, with primary amines to give secondary amines, with phenols to give mixed oxides and with fatty acids to give esters of these acids; these reactions may find important industrial applications.
The further I go the more confident I become in the potentialities of this hypothesis of a temporary compound formed by the catalyst with one of the gaseous elements of the system, whilst on the other hand I see no other plausible explanation of the very marked specific character of different catalysts of similar physical make-up.
At the same temperature of 350°C amyl alcohol vapour yields valeric aldehyde and hydrogen with copper, and amylene and water with thoria. True, in these cases the direction taken by the catalysis might be connected with the metallic or non-metallic nature of the catalyst, determining whether or not electromotive contact forces are to intervene; but such a differentiation could not be invoked in many cases. Thus, I have recently found, with Mailhe, that formic acid vapour breaks down at 250°C only into carbon dioxide and hydrogen in the presence of zinc oxide, and only into water and carbon monoxide in the presence of titanium oxide. Here the two catalyst oxides have no physical dissimilarity, and such a complete reversal of the phenomenon can only be explained by the intervention of special chemical affinities operating at the surface of these catalysts.
For the past fifteen years this idea of mine on the mechanism of catalysis has never left me, and it is to the inferences drawn from it that I owe all my useful results. I shall retain it until such time as I find it irreconcilable with observed and well-established facts.
Theories cannot claim to be indestructible. They are only the plough which the ploughman uses to draw his furrow and which he has every right to discard for another one, of improved design, after the harvest. To be this ploughman, to see my labours result in the furtherance of scientific progress, was the height of my ambition, and now the Swedish Academy of Sciences has come, at this harvest, to add the most brilliant of crowns. Permit me once again to express to the Academy my deepest gratitude.
Their work and discoveries range from the Earth’s climate and our sense of touch to efforts to safeguard freedom of expression.
See them all presented here.