Introductory Lecture BY COURTENAY S. G. PHILLIPS Merton College University of Oxford Oxford OX1 4JD Received 6th January 1981 This meeting is primarily concerned with the interface between physical chemistry and chromatography. Looked at from one side we shall be interested in the manner in which physicochemical techniques and ideas may be used to increase the develop- ment and our understanding of chromatographic methods. Thus the paper by Carley Moroney and Roberts shows how the nature of chromatographic supports can be revealed by appropriate application of photoelectron spectroscopy. The papers by Scott and by Scott and Simpson tell us a great deal about the complexity of liquid-solid surfaces but although these results are of great intrinsic physico- chemical interest the main thrust of their work has been the understanding of h.p.1.c.separations. Again the paper by Knox Kaliszan and Kennedy on enthalpic exclu- sion chromatography may be taken as a neat example of the use of physicochemical logic to extend the boundaries of chromatography as an analytical tool. On the other side and this is to be the main thesis of my own comments the chromatographic methods which have been primarily developed for separation and analysis can be used as simple and remarkably effective techniques for the investigation of a variety of problems of physicochemical interest. T have naturally enquired why I should have been asked to open this meeting. Perhaps it was because I provide a fragile link with the 1949 Discussion of the Faraday Society on Chromatographic Analysis at which I presented my second scientific paper.Perhaps it was because my tutor's tutor was a pupil of Faraday. Perhaps it was be- cause I did not have the courage as so many of you have had to submit a paper for this meeting. However it has been suggested to me that as I have been more fortu- nate than most in benefiting in my research from the physicochemical potential of chromatographic methods I might draw on this experience to illustrate the general field which will be more amply and forcibly brought out in the papers and discussion which are to follow. I can do this with a clearer conscience since the use of chromato- graphic methods to study physicochemical phenomena has been so clearly described and reviewed in the recent book by Conder and Young.' And if I remain open to the accusation of blowing too much my own penny whistle I am at least less likely to run the risk of missing or misrepresenting others.Perhaps I may also sound the antiphon to the remarks of my late lamented mathematical colleague Dodgson about jam to- morrow and jam yesterday but never jam today.* EQUILIBRIA We are today familiar with the use of say electrons neutrons and of a variety of photons in the presence or absence of magnetic and other fields as probes of molecular structure. But molecules themselves are also extremely delicate probes. Moreover these probes can lead directly to the most relevant information for real physico- INTRODUCTORY LECTURE chemical problems.Chromatography of course is often the most practical way of using these molecular probes for it provides us with a range of precise yet rapid methods for the measurement of molecular interactions. The existence of such systems as the Kovkts index illustrates the way in which we regularly use column stationary phases to " feel " the nature of sample molecules. Iremember for example that many years ago we were interested in the identification of new inorganic species such as the simple and mixed hydrides of germanium and silicon of which there turn out to be a very large number indeed only four of which had hitherto been ~haracterised.~ The patterns of retention times which closely followed those of the analogous hydrocarbons enabled us to determine the structures of all these new molecules and to confirm them by a number of auxiliary chromatographic techniques.These included chlorination and separation of the SiCl, GeCl and HCl peaks to determine the stoichiometry use of molecular sieves to distinguish the straight-chain isomers and the combination of a mass-density balance with a thermal- conductivity cell to measure their molecular weights. (In order however to put this into proper perspective it should be mentioned that when we had succeeded in iden- tifying all these new species chromatographically most of my inorganic colleagues were quick to point out that they knew all along that they must be there!) But all this is merely putting chromatography on something of the same level as a tool for molecular structure as infrared spectroscopy was at the end of the last century.The real potential of chromatography in this area has yet to be properly tapped and it is therefore particularly valuable to have this bolder concept explored at this meeting in the most stimulating paper by Kiselev and Poshkus in which they demonstrate how surfaces may be used to reveal remarkable details of the structure of simple mole- cules. The reverse of this the investigation of surface structure is also very neatly illus- trated in the paper by von Rybinski Albrecht and Findenegg which specifically in- vestigates the interaction of adsorbed molecules with one another. In comparison they conclude that the study of the adsorption of single species is only of limited in- terest.However even here it must be noted that the rapidity of chromatographic methods (of which there are of course several e.g. elution on a plateau elution by critical point frontal displacement and thermal desorptionl) can often be of value. Thus Dr. K. F. Scott in my laboratory has recently developed methods for the rapid and automatic generation of adsorption isotherms using a standard chromatograph and a simple mini-computer. These methods are particularly useful to us as we often wish to probe a catalytic surface with a variety of molecules of different shape or chemical character (thus metal sites may be distinguished by hydrogen, exchangeable hydrogen by deuterium-exchange chr~matography,~ while one atom 4.g. Ni in Cu may be identified by Co adsorption).The probing of metals with atomic hydrogen is suggested to us in the paper to be presented by Clifford Gray Mason and Waddicor. I may perhaps give two rather different examples of the way in which such surface studies have assisted us in our investigations of heterogeneous catalysis. In the first we have found that very small changes in overall stoichiometry e.g.in Bi&fOO, can produce very profound changes in catalytic activity and selectivity.6 These result from substantial changes in surface composition which may be shown up for example by p.e.s. but also very dramatically by a complete change in the whole pattern of reten- tion times. Thus on passing from a Bi-rich to a Mo-rich surface of Bi,MoO the relative retention of benzene and hexane changes by a factor of 16.My second ex- ample is taken from work on the catalytic dehydration of alcohols in which the two alternative mechanisms of cis and trans elimination could be distinguished on a model basis by the fact that the transition-state species could differ by one CH group in their C. S. G. PHILLIPS interaction with the s~rface.~ The difference in the retention times of two hydro- carbons on this catalyst thus provides a quantitative estimate of the difference in the activation energies. The delicacy of chromatographic methods means that they can be of particular value in probing weak interactions. The determination of activity coefficients is admirably reviewed for us in the paper by Letcher. On the other hand the nature of solution itself still seems to be a matter of hot dispute which the papers of McCann Purnell and Wellington and of Tiley will provide an opportunity for supporters of regularity or of micropartitioning to show their colours.It is perhaps worth pointing out that molecular interactions can often be highlighted by removing them from solu- tion and thus from the competition of solvent molecules and exposing them on a sur-face. Thus olefins are retarded some 2 or 3 carbon atoms by solution of Ag ions but by more than 50 when AgN03 is adsorbed on a suitable support. Similarly we have found good quantitative evidence for olefin complexes of cadmium when CdF2 is deposited on a surface but not from cadmium-containing solutions.' It seems to me that the most exciting applications of chromatography and molecular probes may well be in the biological field.Thus some years ago we were able to demonstrate that desoxycholic acid in a gas-chromatographic column liquid illustrated some of the characteristics of the bile.9 The paper of Sebille Thuaud and Tillement now opens up for us a fascinating window on this new world of application. KINETICS Since the chromatogram depends on both thermodynamic and kinetic factors it can in turn be used to derive information about diffusion processes and reaction kine- tics as well as equilibria. Indeed chromatographic apparatus may often be con- venient even if no chromatography occurs as is elegantly demonstrated in the paper by Wakeham. The most common method makes use of the plate height.We have at this meet- ing examples of this in the measurement of the diffusion of macromolecules in solu- tion in the paper by Dawkins and Yeadon and in the gaseous diffusion of atomic hydrogen in the paper by Clifford Gray Mason and Waddicor in which the chroma- tography may perhaps be said to have occurred as something of an afterthought. When two competing mass-transfer processes with rather different time-scales can be distinguished the simple van Deemter approach is no longer sufficient and the chromatographic peaks begin to lose their symmetry. The theory of this was worked out by Giddings" in 1963 who also showed that the relative position of the peak maximum would be a function of flow rate for at slow flow rates molecules would spend a higher proportion of their time on the more slowly exchanging sites.We have recently" observed what appears to be a nice example of this phenomenon on a carbonaceous surface laid down on an oxide support as a result of the disproportiona- tion reaction of butadiene. Thus at 300 "Cthis gives normal symmetric peaks for both n-hexane and a branched isomer such as 2,2-dimethylbutane. However at 200 "C while the n-hexane remains normal the 2,2-dimethylbutane peak becomes ab- normally broad asymmetric and with its peak maximum relative to n-hexane sensitive to flow rate. The results fit quite well to a two-site adsorption model in which de- sorption from the slower sites has a half-life of ca. 10 s and extremely well if a few molecules are allowed to explore a third type of site with a longer half-life.The diffi- culty seems to arise when one tries to find a model for these sites which provides slow adsorption for a branched but not for a straight-chain hydrocarbon and the best we have to suggest so far is that there may be an activated rearrangement of the surface INTRODUCTORY LECTURE to better accommodate the more branched molecule. In the presence of even slower adsorption-desorption processes the tail of the peak may become so long and so attenu-ated that it can easily be missed. In such a case the slow process can be shown up most easily by means of stopped-flow chromatography. For example if one takes a column of activated AIL03at say 200 "C and after injection of n-hexane stops this for about half an hour at some point in the column this allows a certain amount of the n-hexane to diffuse slowly into what we presume are fine pores or cavities of an ink-bottle shape with an entry port requiring activation energy.When the gas flow is re- started the bulk of the n-hexane is removed but the desorption of the remainder may be followed in rather minute detail and over some considerable time by periodically stopping the flow allowing some to diffuse out and measuring this as a sharp peak with the detector at a suitable higher sensitivity." The study of the kinetics of chemical reactions may often be considerably assisted by judicious linking of the reactor and the analytical chromatographic column as in the microreactor or sample-vacancy technique^.'^ There are also many occasions when the chromatographic column may itself be used as the reactor.We have an example at this meeting in the paper by Clifford Gray Mason and Waddicor. Infor-mation on the chemical reaction occurring within the chromatographic column can be obtained simply from the loss of reactant or as in diffusion from measurements of peak spreading. However a further feature of the chromatographic process can use- fully be called into play namely the ability to move molecules at will and under precise control through the reactor-chromatographic bed. Reactants individual products and indeed even impurities may all thus be conveniently separated and distinguished from each other particularly by the use of stopped-flow chr~matography.~ Potential inhibitors or cocatalysts may be caused to meet with the reactant at predetermined times and positions in the column.A number of different reactants may be studied simultaneously but at different positions within the reactor. Unwanted isomers may be continuously brought into the most catalytically active part of the column so that a reaction may be driven way beyond its thermodynamic limit for example the al-most complete conversion of n-hexane into 2,2-dimethylbutane.14 In my group we have found that the chromatographic methods can provide a simple means for the rapid investigation of heterogeneous catalytic reactions so rapid in fact that it has become necessary to use on-line computers to keep upwith the flow of data.At the same time catalytic surfaces are being investigated by a variety of molecular probes as I have already discussed. However these probes are not limited to the thermodynamic and diffusional background of the reactions studied. The use of the reactions themselves either with different molecules in the same reaction or with different reactions can also prove to be extremely revealing. Thus by compar- ing the hydrogenolysis of 2,3-dimethylbutane and 2-methylbutane we have been able to distinguish between a surface and a desorption-resorption mechani~m,~~ while Dr. K. F. Scott has been able to use stopped-flow chromatography in systems where the reactant is virtually stationary to measure the distribution across the surface of reac-tion sites of different activity and the manner in which these may be selectively poisoned.16 CHROMATOGR A PHY AND PHY S I CAL CHEMISTRY I have this morning tried to communicate and the rest of this meeting will illumi- nate some of the many ways in which chromatography may be used to unravel prob- lems of real interest in physical chemistry.Indeed in every chromatogram Nature is revealing to us an astonishing wealth of information if only we had like the boy and C. S. G. PHILLIPS the emperor's clothes the eyes and the will to see. Why then one must ask are chromatographic methods so little used apart from their traditional and well-recog- nised function in chemical analysis ? In part this may be due to the complex and often bizarre nature of chromato- graphic stationary phases and operating procedures and from the multiplicity of physicochemical processes which seem to be involved in any chromatographic experi- ment.But as has now been shown by a number of different investigations there is in general good agreement between the physicochemical data obtained by chromato- graphy and by other more traditional methods.' In part this may be due to the curious belief that there must always be something of empiricism and black art in any analytical method. I hope very much that now that so much gas and more recently so much liquid has flowed through high-perfor- mance chromatographic columns for mainly analytical purposes the faith of the Faraday Division in once again casting its benevolent mantle upon us may do some- thing to give chromatography its physicochemical respectability.Finally and once more upon a personal note and as the shadows of life begin to lengthen and policemen and chromatographers get so remarkably young I would like to emphasise the power rapidity and above all the simplicity of the chromato- graphic methods. When I started research some 35 years ago there seemed to be plenty of things well worth investigating with relatively simple apparatus. Today my younger colleagues do not seem to find themselves so fortunately situated and so often are made to feel that nothing worth while can be done without elaborate and expensive equipment. They might be well advised to consider taking up some of the many opportunities which this meeting is about to discuss.J. R. Conder and C. L. Young Physicochemical Measurements by Gas Chromatography (John Wiley London 1979). [See also R. J. Laub and R. L. Pecsok Physicochemical Applications of Gas Chromatography (John Wiley New York 1978)l. L. Carroll Alice through the Looking Glass chap. 5. C. S. G. Phillips P. Powell J. A. Semlyen and P. L. Timms 2. Anal. Chem. 1963 197 202. K. F. Scott and C. S. G. Phillips J. Chem. SOC.,Faraday Trans. 1 1980 76 683. K. F. Scott and C. S. G. Phillips J. Catal. 1978 51 131. A. G. Mitchell P. M. Lyne K. F. Scott and C. S. G. Phillips J. Chem. Soc. Faraday Trans. I 1981 77 in press. 'R. M. Lane B. C. Lane and C. S. G. Phillips J. Catal. 1970 18 281. I. Hadzistelios F. Lawton and C. S. G. Phillips J. Chem. SOC.,Dalton Trans. 1973 2159. A. 0.S. Maczek and C. S. G. Phillips J. Chromatogr. 1967 29 7. lo J. C. Giddings Anal. Chem, 1963 35 1999. G. J. S. Vint and C. S. G. Phillips unpublished work. l2 C. S. G. Phillips M. J. Walker C. R. McIlwrick and P. A. Rosser J. Chromatogr. Sci. 1970 8 401. l3 C. S. G. Phillips and C. R. McIlwrick Anal. Chem. 1973 45 782. l4 C. M. A. Badger J. A. Harris K. F. Scott M. J. Walker and C. S. G. Phillips J. Chromatogr. 1976 126 11. N. D. Perkins and C. S. G. Phillips J. Catal. 1980 66 248. l6 K. F. Scott J. Chem. SOC.,Faraday Trans. 1 1980 76 2065.