FEBRUARY, 1967 THE ANALYST Vol. 92, No. 1091 Techniques in Gas Chromatography Part I. Choice of Solid Supports A Review * BY J. F. PALFRAMAN AND E. A. WALKER (Ministvy of Technology, Labovatory of the Government Chemist, Cornwall House, Stamford Stveet, London, S.E. 1) SUMMARY OF CONTENTS Introduction Diatomaceous supports Types of diatomaceous supports Modified diatomaceous supports Treatment of diatomaceous supports Polar support de-activators Fluorocarbon supports Specialist supports Graphitised carbon Boehmite “Tide’ ’ Glass beads Sterrasters Vermiculite Porous polymer beads (a) Porapak ( b ) Chromosorb 102 (c) Polypak Conclusion Classification of supports THIS review of solid supports has been written as part of a general review of techniques of gas chromatography, which is intended to cover the period since the last general review in the Analyst by Rose1 in 1959.As the period has been one of intense activity in this field (over 7000 papers have been published on various facets of gas chromatography) it would clearly be a difficult task to cover the subject adequately in a single paper, and we therefore intend to produce a series of short reviews, by which means it should be possible to maintain the subject matter at a current level. The column has justifiably been described as the “heart of the gas chromatograph” and in its efficient functioning the solid support plays a vital r61e. However suitable a particular stationary phase may be for a given separation it will accomplish little if insufficient care is paid to the selection, gradation and treatment of the support, and as this is the first step in column preparation it is a good point at which to start this review.The effects of particle size and particle-size distribution were reviewed by Rose,l and apart from reiterating the importance of working with the minimum practicable distribution of particle size, it is not proposed to enter into any discussion of this topic, although some reference will be made to it in the discussion of preparative techniques at a later date. Since 1959 a considerable volume of literature has, however, been published on the treatment and choice of the so-called “inert” solid support. That the support could have a significant influence on such an important parameter as the retention time was acknowledged at the earliest stages of development.2 One has, in fact, only to realise that gas - solid chromatography constitutes a wide field of endeavour to appreciate that the support surface can have significant effects in this respect.Taking an extreme example, we have found that the tendency of carbon dioxide to tail when * Reprints of this paper will be available shortly. For details see Summaries in advertisement pages. 7172 PALFRAMAN AND WALKER: TECHNIQUES IN GAS CHROMATOGRAPHY [Analyst, Vol. 92 chromatographed on silica gel is considerably reduced by the addition of a 1 per cent. coating of SE30 as a stationary phase. The retention characteristics, on the other hand, are not very different from those on the bare silica-gel surface, and are certainly uncharacteristic of a 1 per cent.SE30 column on a normal “inert” type of support. Adsorption and its effect on peak symmetry still frequently remain limiting factors both in the interpretation of retention data, and quantitative evaluation. I t is not surprising, therefore, that considerable effort has been expended in the last few years in the quest for the elusive ideal solid support. DIATOMACEOU s SUPPORTS TYPES OF DIATOMACEOUS SUPPORTS- Ideally the function of the support is simply that of a retentive sponge, capable of holding immobile a relatively large volume of liquid, at the same time exposing a large inert surface area to ensure rapid attainment of equilibrium between solute and solvent. Although many materials have been tried with varying degrees of success, nothing yet has been found to be more satisfactory, in general application, than diatomaceous earths, which have been in use since the early days of James and Martin.3 Natural diatomite (Kieselguhr), which is somewhat fragile, is treated by calcining with a flux of sodium carbonate to 1600” C.This causes the material to fuse and the silica to be converted into crystalline cristobalite, the original grey diatomite becoming white in colour because of the conversion of iron oxide to a colourless complex sodium iron silicate. This flux-calcined material is marketed by Johns Manville as Celite or Chromosorb W. In a variation of the above process, the calcination takes place without the addition of a flux. This causes the mineral impurities to form complex oxides imparting a pink colour to the material, and this is marketed as firebrick or Chromo- sorb P.While neither Chromosorb P nor W may be regarded as ideal, they each fulfil supplementary functions that cover a wide range of applications. Chromosorb P has a greater density than W because of the destruction of the diatom structure in the calcination process, it is less fragile than W and is capable of holding a larger volume of liquid phase before becoming too sticky to flow freely. However, when used with polar solutes severe tailing is often encountered. This arises from adsorptive centres in the support which give rise to non-linear isotherms, an effect which is much less pronounced with the white supports. In a comparison of the pink and white supports Baker, Lee and Wall4 found that the properties varied with respect to surface area and pore-size distribution.The difference in pore size was considered by Otten~tein,~ who reported that the white support had pore sizes of about 9 p while the pink support had a smaller pore size (about 2.0 p). I t was shown that this difference explained the variation in column behaviour of the two supports, Chromosorb W holding the liquid phase in large pools and Chromosorb P in small ones. The column efficiency being controlled by the mass transfer between the liquid and the gas phase, the larger pools of liquid required longer transit times for the solute, and peak broadening was a result. Saha and Giddings6 attempted to correlate the C, terms of the Van Deemter equation with the distribution of pore size, by comparing experimental results with theoretical values.The results obtained for Chromosorb W were satisfactory but those of Chromosorb P were about fifty times too low. Several descriptive technical bulletins are issued by Johns Manville (London) describing the properties and uses of these materials. MODIFIED DIATOMACEOUS SUPPORTS- The realisation of the critical r61e played by the support in gas - liquid chromatography has led manufacturers to prepare and market materials made especially for this purpose. Johns Manville recently introduced Chromosorb G, which is also a diatomite, especially processed for use in gas - liquid chromatography, for which high efficiencies and greatly reduced surface activity are claimed. In our own experience, confirmed by personal com- munications from other workers, it is the most inactive support of its type available on the market, but it is often difficult to obtain good efficiencies by using this material. In a survey of comparative column efficiencies of some common solid supports Saha and Giddings’ showed decreasing potential efficiency through Chromosorb P, Chromosorb W, Gas-Chrom S to a minimum in Chromosorb G.Its physical characteristics differ somewhat from conventional supports, and this factor must be taken into consideration in column preparation. For example, it has a density of about two and a half times that of Chromosorb W, which meansFebruary, 19671 PART I. COMPARISON OF PHYSICAL Free fall density, g per ml . . Packed density, g per ml .. Surface area, m2 per gm . . Surface area, m2 per ml . . pH . . . . . . . . CHOICE OF SOLID SUPPORTS TABLE I PROPERTIES OF 60 TO 80-MESH CHROMOSORBS* Teflon, A P w G 40 to 60 mesh 0.40 0.38 0.18 0.47 0.42 0.48 0.47 0.24 0.58 0.49 2.7 4.0 1.0 0.5 7.8 1.3 1.88 0.29 0.39 - 7 - 1 6.5 8.5 8.5 - 73 TABLE I1 Silica . . . . . . . . Aluminium oxide . . . . Iron(II1) oxide . . .. Titanium dioxide . . . . Calcium oxide . . . . Magnesium oxide . . . . Sodium monoxide + potas- sium monoxide . . . . Moisture . . . . . . TYPICAL CHEMICAL ANALYSES OF CHROMOSORBS~ White Pink r A \ --7--- Non-acid washed, Acid washed, Non-acid washed, Acid washed, per cent. per cent. per cent. per cent. 90.6 91.6 88.9 90.0 4-4 4.1 4.0 3.6 1.6 1-4 1.6 1.4 0.2 0.2 0.2 0.2 0.6 0.4 0.6 0.4 0.6 0.5 0.6 0.5 1-0 0.9 3.6 3.2 0.3 0.3 0.3 0-3 TABLE III* COMPARISON OF WEIGHTS OF LIQUID PHASE IN ~OO-ML OF EACH SUPPORT AT VARIOUS LIQUID LOADINGS Chromosorb I 3 G w P Weight of support, g .. . . . . 58.0 24.0 47.0 Liquid phase a t 1 per cent., g . . 0.58 0.24 0.47 Liquid phase at 2 per cent., g . . 1.18 0.48 0.95 Liquid phase a t 3 per cent., g . . 1.79 0.74 1.45 Liquid phase a t 4 per cent., g . . 2.42 1.00 1.95 Liquid phase at 5 per cent., g . . 3.05 1.26 2.47 TABLE IV* COMPARISON OF SUPPORT FRIABILITY Breakdown, per cent. Mechanical shaking of 30 t o 60-mesh material G IT' I; -60 mesh in 5 minutes . . . . 1.6 19.4 12.0 -60 mesh in 10 minutes . . . . 8.6 53.4 27-6 -60 mesh in 15 minutes . . . . 12.4 75.8 46.0 * Johns Manville technical bulletin. t Ottcnstein.26 that a 5 per cent.w/w loading of liquid phase would be equivalent to a conventional 12 per cent. column. Chromosorb G is particularly useful when low loaded columns are required, as, for example, in steroid and pesticide analysis, where columns of the order of 1 per cent. w/w are necessary in order to minimise retention times. However, because of its low surface area, loadings greater than about 5 per cent. are not recommended. In contrast with this type of support, Chromosorb A, another flux calcined diatomite, was developed for preparative chromatography, where high liquid-phase loadings are required. Like Chromosorb P it has a large surface area, thus allowing a maximum loading of 25 per cent. (compared with 30 per cent. for P) and good mechanical properties, but in its low adsorptive characteristic it is closer to Chromosorb W or G.A summary of the properties of these materials is given in Tables I to IV.74 PALFRAMAN AND WALKER: TECHNIQUES I N GAS CHROMATOGRAPHY [Analyst, Vol. 92 TREATMENT OF DIATOMACEOUS SUPPORTS- Limitations in the reliability of the support imposed by the effects of solute adsorption have prompted a number of studies to be made directed at the elimination of the “active centres.” It is generally agreed that these consist of -OH groups associated with silicon, aluminium and iron,* and for this reason many workers have pre-treated the supports by prolonged extraction with various strengths of hydrochloric acid, which is said to leach out surface alumina and iron. For use in the analysis of amines James and Martin,g among others, have treated the neutralised acid-washed support with alkali; however, this treatment is queried by Fales and Pisano,lo who observe that tailing would persist if the compounds being analysed contain -OH or -NH groups.A detailed description of these techniques is given by Burchfield and Storrs.ll As an alternative to the removal of active centres by acid treatment, Omerod and Scott12 sought to cover them by coating the support with silver. The method is expensive and carries the risk of reactions taking place between themetal coating and the solute; it does, however, give surfaces of high thermal stability and the coating may be carried out in situ. Although treatment with acid can be effective in removing active -OH groups associated with surface impurities such as iron and aluminium, it does nothing to mitigate the effects of the silanol (-%OH) groups which still remain.Alkali treatment, as already menti~ned,~ may be used where basic solutes have to be resolved, but for many other compounds there remains an inherent risk of decomposition by residual alkali. Saturation of the silanol groups can also be accomplished by including in the stationary phase a small proportion of a functionally active compound capable of hydrogen bonding with these surface groups. Here again, the technique imposes limitations particularly for low loaded columns when the amount required may constitute a significant proportion of the total phase. An alternative method, which has been used with great success, is the modification of the surface hydroxyl groups by silanisation.A procedure for this, involving the use of dimethyldichlorosilane, was reported by Horning et aZ.13914 and a detailed study of the technique was carried out by Bohemen, Langer, Perrett and P~rne1l.l~ The dimethyl- dichlorosilane (DMCS) is assumed to react with the hydroxyl groups on the silica surface as follows- I I I I I I \ / / \ -Si-0-Si- + SiCI,(CH,),--+ -Si-0-53- + 2HC1. OH OH I 0 I 0 Si CH, CH, As the two adjacent -OH groups are required for complete reaction it is unlikely that all of the reactive sites are removed. A single -OH group might leave an undesirable Si-C1 linkage as follows- I I 1 I -Si-O-Si- + SjCI2(CH,), + -Si-O-Si-O-SiCl(CH,), + HC1. I AH I I As successful silanisation by means of DMCS has usually been followed by removal of hydrochloric acid by methanol it appears that the Si-C1 reacts with the methanol to form a methoxy compound- OCH, I 1 I I / I I I I \ -Si-0-Si-OSiCl(CH,), + CH,OH + -Si-0-Si-0-Si-CH, + HC1.CH, An alternative reagent, hexamethyldisilazane (HMDS) was used by Boheman, Langer, Perrett and Purnell15 to modify a pink support; this reacts quantitatively with single -OH groups forming a non-reactive linkage- I I I I 1 I I 1 -5-0-Si- + Si,(CH,),NH --+ -Si--0--Si- + NH,. OSi(CH,), OSi(CH,), OH OHFebruary, 19671 PART I. CHOICE OF SOLID SUPPORTS 75 In another paper Perrett and PurnelP compared HMDS-treated white and pink supports, and concluded that the surface of each was reduced by this treatment. As some reactive sites still remain, Littlewoods suggested that not all adsorption is associated with -OH groups.In another critical study of silanisation in which Chromosorb W was treated with various reagents, Kirklandl’ concluded that acid washing was the first essential in the production of a silanised material of minimum activity. A master batch of this support was, therefore, prepared and separate portions reacted with each of the silanising agents. From these supports 20 per cent. Apiezon L columns were prepared and used to separate the same polar test mixture. Although the separations gave similar retention times and efficiencies for each constituent of the test mixture there were pronounced differences in peak shape. Thus the peak asymmetry factorls for the DMCS-treated support was found to be the least, and that of TMCS (trimethylchlorosilane) was greatest, while severe tailing was found when an un- treated support was used.It is interesting to note that the surface area of the DMCS-treated support was considerably less than that of the other two. Kirkland’s views also received support from Fales and Pisano,lo who expressed preference for the DMCS treatment. Pro- cedures for silanisation have also been described by Supina, Kruppa and Henly,lg Purnel120 and elsewhere.21 In the last publication it was emphasised that the common test, flotation on water, for silane-treated supports is unsatisfactory as it qualifies many unsuitable supports. I t was also claimed that peak tailing is not an infallible guide with which to test supports; a more reliable method was to perform quantitative determinations of a known standard to check adsorption losses.Another approach, the co-polymerisation of hexamethyldisilazane on firebrick by treatment with fifty megarads of gamma irradiation, was discussed by Urone and Parcher,22 who claimed a “higher degree of paraffinic character” to the support than previously obtained. The preparation has been described of A e r ~ p a k , ~ ~ a “highly efficient’’ inert support based on silanised Chromosorb W, in which great emphasis has been given to the removal of “fines” at all stages of preparation. This is claimed to account for the improved efficiency and reduced tailing. In the process of silanisation, the character of the surface is changed from hydrophilic to hydrophobic, making it particularly effective for the retention of non-polar and moderately polar phases.However, the resulting production of a support of low surface energy makes it difficult to coat with polar phases because of poor wetting of the surface,17 and in an attempt to overcome this, Vanden-Heuvel, Gardiner and H ~ r n i n g ~ ~ used a coating of poly(viny1 pyrrolidone) which, they claim, does not act as a liquid phase itself but as a satisfactory support de-activator for use with polar phases. It is particularly suitable when using columns with a low percentage of liquid phase. This coating technique has been used successfully in the Government Laboratory for preparing polar columns for steroid analysis, and a similar technique involving the use of an epikote resin is used for pesticide analysis.Experience has shown that, to prepare a satisfactory column by this technique, it is essential to dry the support thoroughly by heating at 200” C for 4 hours before coating it with poly- (vinyl pyrrolidone) . Occasionally slight differences in retention times have been observed when this technique is used, but it is not clear whether this is due to the poly(viny1 pyrrolidone) itself, or the reduced effect of the support. The maximum temperature of operation of a poly(viny1 pyrro1idone)-treated support is 220” C. POLAR SUPPORT DE-ACTIVATORS- The main forces contributing to adsorption are considered to be weak van der Waal’s forces and stronger hydrogen bonding. Scholtz and Brandt26 suggested that all liquid phases neutralise the weak van der Waal’s forces, but that liquid phases capable of hydrogen bonding are required to de-activate the polar adsorption sites. The hydrogen-bonding sites are of two types; the first arises from silanol (Si-OH) groups, where the support is the proton donor in the hydrogen bond, and the second from siloxane (Si-0-Si) groups, where the support acts as the proton acceptor.Ottenstein26 points out that the siloxane group is much more effective in forming a hydrogen bond than the silane oxygen, and that it is the strength of the hydrogen bond that determines the extent of solute adsorption. Therefore, compounds that form strong hydrogen bonds, e.g., water, alcohols and amines, cause severe tailing, whereas ketones, esters, etc., with less tendency to form hydrogen bonds, cause little tailing.Scholz and Brandt25 saturated the silanol group by adding small amounts of Armeen S.D.,76 PALFRAMAN AND WALKER: TECHNIQUES I N GAS CHROMATOGRAPHY [Afizalyst, vol. 92 a long-chained fatty amine, to a non-polar liquid phase for the analysis of amines. Similarly, James and Martin3 eliminated tailing in the separation of fatty acids by incorporating a small amount of stearic acid in the silicone fluidused as a liquid phase. A ~ e r i l ~ ~ discussed the use of corrosion inhibitors in reducing tailing in packed and capillary columns, and among such compounds investigated were dicarboxylic acids (anionic), long-chain amines (cationic) and esterified polyglycols (non-ionic). He found that the use of these inhibitors allowed polar materials to be analysed on non-polar columns without loss of resolution due to tailing.By this procedure the retention times of the non-polar solutes are increased and those of polar solutes decreased. FLUOROCARBON SUPPORTS Of all supports, fluorine polymers are considered to be the most inert, but it is worth noting that although the properties required of an ideal support suggest the use of a totally inert substance, it must still have sufficient surface energy to be capable of holding the liquid phase. A detailed study of the fluoro-polymers as support materials has been made by Kirkland,17 ,28 in which he compares the properties of Kel-F (polymer of chlorotrifluoro- ethylene) , Fluoropak-80 and Teflon-6 (the last two forms of tetrafluoroethylene).The use of these substances as supports requires a more careful consideration of handling techniques than when the conventional diatomaceous ones are used. The choice of liquid phase for the fluorocarbons can be critical, as surface energies are much lower than those of diatomites and they are not so readily “wetted”; they are also fragile and need cooling below the transition point of 19” C before handling. A summary of the optimum conditions for preparation of Teflon columns is given1’- Optimum percentage Optimum carrier gas Support of liquid phase velocity, cm per second Teflon 6 . . . . . . .. 15 to 20 4 to 5 Fluoropak-80 . . . . . . 2 to 5 2 Kel-F . . .. . . . . 15 t o 20 10 t o 15 Under optimum conditions Kirkland17 claimed the following HETP values obtained when n-butanol was chromatographed- Support HETP, mm Surfacc area, m2 per gni Fluoropak-80 . ... .. 3.6 1.3 Kel-F (50 to 80 mesh) . . . . 2.6 2.2 Teflon 6 (full range) . . .. 2.3 10.9 Teflon 6 (40 to 60 mesh). . .. 1.7 10-5 1.1 - Chromosorb W . . . . . . From these values it seems that properly prepared Teflon columns have efficiencies similar to those prepared from Chromosorb W. However, other workers have reached different conclusions. Sawyer and Barr29 in a comparative study of a number of support materials (including Fluoropak-80, glass beads, Chromosorb W and carborundum) , found that the fluorocarbon support produced “very poor plate heights.” In a review of solid supports, Ottenstein26 cited several authors who had reported upon the use of Teflon, most of whom had com- mented upon the low efficiencies produced when this material was used as a support. It is, however, inert and has been considered useful in the analysis of aqueous samples.30 Japanese workers, Onaka and 0kamot0,~l have used diatomaceous-earth supports coated with PTFE, and found them useful for the separation of polar materials.Kirkland32 com- pared the performance of these coated supports with Teflon-6. From his survey it appears that the unmodified Teflon-6 produces peaks with less tailing than the PTFE-coated diato- maceous-earth supports, and he suggested that this arises from incomplete coating of the support surface. The coated supports, however, have the better handling properties and unlike the Teflon-6 do not have to be handled at temperatures below 19” C during column preparation.Teflon-6 generally produced columns of higher efficiencies, but unlike the coated support did not produce a linear retention - volume plot with change of liquid-phase loading, due no doubt to its low surface energy. SPECIAL~ST SUPPORTS GRAPHITISED CARBON- Since the time Eggertsen, Knight and Groenning~~~ used Pelletex (a modified carbon black) to separate c6 and C, hydrocarbons, an increasing number of workers have used aFebruary, 19671 PART I. CHOICE OF SOLID SUPPORTS 77 vai-iety of modified adsorbents both as supports and partitioning media. A butoxy-modified silica gel was used by Kirkland17 to separate the components of a Phillips’ hydrocarbon mixture. Perhaps the most interesting advance has been the use of thermally graphitised carbon black. The most serious objection to adsorption chromatography has been the asymmetry of the peaks obtained as a result of the heterogeneous nature of the surface. Graphitised carbon black, which is a thermal carbon black heated to 3000” C, has a moderately homogeneous surface which is also, within limits, reproducible. It, therefore, does not suffer from the same limitations as the more common adsorbents such as silica gel.Its nature and use have been discussed in detail by K i ~ e l e v , ~ ~ $ ~ ~ who also demonstrated a number of useful separations. Spheron-S coated on a polythene moulding powder (100 to 120 mesh) has been used analytically by Pope36 and B r ~ d a s k y , ~ ~ who compared Sterling FT 2700, a graphitised carbon, with the more conventional supports, Chromosorb W, Haloport F and Anakrom.When analysing a mixture of polar compounds (comprising water, alcohol, ketones and amines) on a non-polar phase (Dow Corning, Silicone Oil 200) he found that although adsorption was greater when using the graphitised carbon, it was superior in the prevention of peak asym- metry. Brodasky concluded that the advantages of using graphitised carbon as a support material included the elimination of chemical pre-treatment to prevent tailing, even with low loadings of non-polar stationary phases. High efficiencies were also obtainable, and there is little problem from temperature limitations, the principal disadvantage being the necessity of careful size grading of the coated support. Halasz and H o r ~ a t h ~ ~ reported the use of graphitised carbon as a porous layer on glass beads, with increased resolution in spite of shorter re ten tion times.BOEHMITE- Another versatile adsorbent, finding increasing use in a similar fashion, is Baymal (a fine alumina boehmite), which Kirkland17932,39 has used both for gas - solid and gas - liquid chromatography. This material consists of small crystalline fibrils of colloidal alumina, about 1000 A long and 50 A in diameter, with a high specific surface area of about 275 m2 per g. Because of its high positive surface charge, boehmite has a unique property, vix., ease of deposition from aqueous solutions on to a variety of surfa~es.3~,40~~~ This enables it to anchor to such negatively charged surfaces as carboxylic acids, colloidal silica, proteins, organic chelating agents and anionic surfactants.It has, therefore, great potential as a material for the preparation of “custom” surfaces for specialised separations, as, for example, capillary columns of a highly selective nature. The surfaces prepared from fibrillar boehmite can be further modified by conventional liquid phases or by prior treatment with acetate ions. Kirkland17 used boehmite, modified with stearic acid, to separate a mixture of low boiling fluorocarbons. “TIDE”- A commercial household detergent “Tide” has been used by a number of workers both as a support and a partitioning material. Its use was first reported by Gohlke and M~Lafferty~~ as a general-purpose material, and later by Desty and Harb0u1-n~~ for the separation of hydro- carbon mixtures.Decora and Dinneen44 obtained a porous, inert residue from “Tide” by grinding, sieving and extracting the organic material with 30” to 60” C light petroleum in a Soxhlet apparatus. This was then coated with a silicone oil and used as column packing in the separation of a number of pyridines, which was accomplished without producing tailing peaks. However, in later work45 on more basic nitrogenous samples, a modification (the residue was reacted with potassium hydroxide) had to be made to overcome the tailing. B e n ~ ~ ~ studied the adsorption characteristics of several supports used in gas chromatography including “Tide,” and found that while the elution of hydrocarbons was in the order of the boiling-points, the elution of ketones, acetates and alcohols was not.He was unable to predict the order of elution of oxygenated compounds when using “Tide” as support. Pocaro and Johnston4’ used unmodified “Tide” (although perfume and water had been removed by drying at 110” C) to separate 2-methyl-1-butanol from 3-methyl-1-butanol; previous attempts in which both polar and non-polar stationary phases were used had been unsuccessful in resolving these materials. GLASS BEADS- Glass beads have been used intermittently for a number of years as an alternative support. Callear and Cvetanovits4* used them as long ago as 1955 to reduce tailing, but although78 PALFRAMAN AND WALKER: TECHNIQUES IN GAS CHROMATOGRAPHY [Analyst, Vol. 92 the beads are thought by many to be relatively inactive, Littlew~od~~ has suggested that this is not universally acceptable.One of the main advantages of glass-bead columns is speed of analysis, arising from the low volume of liquid phase involved and improved per- meability due to evenness of packing which is made possible by the use of homogeneous, uniformly sized spherical particles. These factors also made glass beads ideal for theoretical studies on column p e r f o r m a n ~ e . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ They are, on the other hand, usually much less efficient than conventionally packed columns. Several authorsa 955 have suggested that this is caused by the formation of liquid pools at the contact points between the beads and that it is, therefore, necessary to distribute the liquid thinly and evenly over the surface.An attempt to overcome this by mechanical etching of the beads was U ~ S U C C ~ S S ~ U ~ , ~ ~ but a study of the effect of chemical etching56 with hydrogen fluoride showed some improvement in efficiency, In 1960, Golay5’ suggested that the contact effect could be overcome by coating the beads with a porous inactive layer, a system that would retain the advantages of a solid non-permeable core, and more recently this approach has been given closer attention with some success. Dewar and Maier54 prepared a packing by coating the beads with a stationary phase containing very fine particles of “Super Floss” (a diatomaceous earth) relying upon surface tension for adherence to the beads. HETP values of 0-8 mm were obtained, as compared with 0.6 to 2.0 mm for columns prepared from diatomaceous-earth packings and 4.5 mm for unmodified beads.As already mentioned, graphitised carbon black has been used3* as a coating for micro glass beads giving “porous-layer” glass beads,” the coating of fine particles apparently adhering by van der Waal’s forces. In comparison with conventional columns a %fold reduction in analysis time is claimed. Kirkland59 made a similar approach by modifying glass beads with a thin layer of diatomaceous earth using Baymal (a fibrillar boehmite) as an anchoring agent. With columns prepared from these beads, a HETP value of 0.8 mm was obtained, in agreement with values also reported by D e ~ a r , ~ ~ presumably arising from the homogeneity of liquid-phase deposition over the surface which improved the liquid mass- transfer effects. The modified beads showed no tendency to give asymmetric peaks when chromatographing polar solutes on non-polar solvents.In another the same author prepared glass beads modified by silica gel, again with Raymal as adherent. With this technique he was able to prepare a support of desired thickness and surface area retaining the good packing characteristics of the glass-bead nucleus. The increase in surface area after successive treatment with the Baymal- silica sol is indicated- Untreated . . . . . . . . . . . . 0.02 Methanol - potassium hydroxide cleaned . . . . 0.03 After 1 Baymal- silica treatment . . . . .. 0.05 After 3 Baymal- silica treatments . . . . .. 0.28 Unfortunately, it also exhibited a greater tendency to tail. Condition of the glass beads, 60 to 80 mesh Surface area, m2 per g After 5 Baymal - silica treatments .. .. .. 0.47 STERRASTERS- As a possible alternative support to glass beads the naturally occurring sterrastcrs, obtained from marine sponges, were studied by Webb, Smith and Wells.58 They are mainly spherical in shape with an average diameter of 54.9 p, which is about the optimum size, as calculated, for glass beads.53 The entire surface of the sterrasters is covered with a fine, elevated texture which, acting as a natural etching, causes the liquid phase to be evenly distributed in a thin film. From this material columns with HETP values of 0.6 mm were obtained. As a result of their hard impervious core, sterrasters have the relative inertness to polar solutes characteristic of glass beads, and show little tendency to produce tailing peaks with higher alcohols.VERMICULITE- Another naturally occurring material, vermiculite, has been examined by McKinney.60 This is a hydrated magnesium iron silicate, composed of platelets, about 9.3 A thick, of tetrahedral sheets bonded to a central bivalent sheet of Mg2+ ions. An interesting property of this material is its ability to adsorb water and carboxylic acids which penetrate between the platelets, suggesting its use in the analysis of aqueous samples. However, adsorption is greater and efficiencies lower than with the diatomaceous supports.February, 19671 PART I. CHOICE OF SOLID SUPPORTS 79 POROUS POLYMER BEADS- First described by Lloyd and Alfrey6l962 these materials were used by Moore63 to prepare columns for separating polymers by “Gel-permeation Chromatography.” The bead polymer is synthesised by “suspension p~lymerisation”~~ and by drying the styrene divinyl benzene beads they still retain their wet structure.A detailed description of these beads and their properties when used as a support material in gas chromatography is given by Hollis,G5@ and a summary has also been published.23 Porapak-Five basic materials are marketed under the name of Porapak (supplied by Waters Associates Inc., Massachussetts, U.S.A.) (P, Q, R, S and T) differing in the degree of cross linking of the basic styrene with ethyl vinyl benzene, and some remarkable separations have been achieved by the use of these materials. It appears that the separation of the solute components by direct partition from the gas phase throughout the solid polymer is not just a surface effect as with conventional supports.As adsorption sites are non-existent, highly polar materials are eluted without tailing, and as no liquid phase is normally used, column bleed is not a problem even when temperature programming (the beads are thermally stable to about 250” C). Compounds containing hydroxyl groups (e.g., water, alcohols and glycols) are usually a problem in conventional gas-chromatographic analysis because of the tailing produced by adsorption of these polar compounds by the support. The porous polymer beads are, however, non-adsorptive and hydroxylated materials are eluted with symmetrical peaks, an interesting point being the early elution of water and certain glycols, while less polar materials are retarded.In the Government Laboratory, porous polymer beads have been successfully used to separate components from aqueous alcoholic solutions and promising results are being achieved in the separation of glycols and in the determination of their trace impurities. It is expected that information on this subject will be published shortly. The beads are supplied in the same mesh size as the diatomic supports and columns are prepared from them in a similar manner, although, in this case, packing is facilitated by the more rigid structure of the beads. These columns are capable of producing efficiencies of up to 800 plates per foot, but are easily overloaded and sample sizes of about 0.2 p1 per component are recommended.However, the non-sorptive nature of the beads allows for a rapid recovery of base-line should the column be overloaded, as in the analysis of trace constituents. Chromosorb 102-Prepared from Rohm and Haas Amberlite XAD-2, this material is a resin of high surface area, size graded for use as a packing as a support in gas-liquid chromatography. Polypak-This is a porous polymer developed by F. & M. Scientific Ltd. Its properties and separations are similar to Porapak. Its properties are similar to those of the porous polymer beads. CONCLUSION An “ideal” support must combine apparently conflicting qualities, for it is evident that what may be ideal for one circumstance may be unacceptable in another. Nevertheless, considerable advances have been made in the development of the conventional types of solid supports. In the technology of diatomaceous earths, development is probably approaching a limit, but undoubtedly there still remains room for further advances in the use of porous layer glass beads.Perhaps the ideal is to be found in the development of the various types of solid phases such as the cross-linked polystyrene. This might undoubtedly lead to a vast field of development. Classification of Supports DIATOMACEOUS SUPPORTS THE WHITE DIATOMACEOUS SUPPORTS These are derived from flux-calcined diatomite, and are friable but relatively inert and useful for the analysis of polar samples. They are commercially available in several forms. UNMODIFIED DIATOMACEOUS SUPPORTS- in a number of mesh sizes by J.J. (King’s Lynn). Celite 545-Manuf actured by Johns Manville, and supplied in this country already sieved80 PALFRAMAN AND WALKER: TECHNIQUES I N GAS CHROMATOGRAPHY [Analyst, Vol. 92 Chromosorb W (N.A . W.)-Manufactured by Johns Manville by size grading Celite, and supplied by Perkin-Elmer in a variety of mesh sizes. Gas-Chrom S-Prepared by Applied Science, from Celatom (a Celite type of diatomite manufactured by Eagle Picher Co. from deposits in Nevada). Available in a wide variety of mesh sizes. Gas-Chrom CL-Prepared by Applied Science from Celite. Anakrom U-Prepared by Analab Inc., and distributed in this country by Gas Chromato- graphy Ltd. This support is specially size graded in a 10-mesh range cut claimed to give higher efficiencies. ACID-WASHED DIATOMACEOUS SUPPORTS- These include the previous supports, which have been modified by various acid-washing procedures to remove some of the adsorptive sites. Celite (A.W.)-Washed with hydrochloric acid. Chromosorb W (A. W.)-Washed with hydrochloric acid. Gas-Chrom (A)-Celatom acid washed. Gas-Chrom (CL.A)-Celite acid washed. Anakrom (A)-Acid washed. Size graded. ACID AND ALCOHOLIC-BASE WASHED DIATOMACEOUS SUPPORTS- This technique was developed to reduce tailing when chromatographing basic materials (e.g., amines) . The acid-washed supports are dried, treated with alcoholic potassium hydroxide and washed with methanol. Gas-Chrom P-Celatom, acid and alcoholic-base washed. Anakrom AB-Acid and alcoholic-base washed Anakrom U. Gas-Chroulz CL.P.-Celite, acid and base washed.Diatomite C-Celite prepared by J.J. (can be obtained acid or alkali washed). SILANISED DIATOMACEOUS SUPPORTS- These are recommended for very polar materials and for low loaded columns. supports are incompatible with certain highly polar liquid phases. with dimethyldichlorosilane (DMCS) . For further reduction of active centres many supports are available in a silanised form. Silanised Chromosorb W (A. W.-DMCS)-The acid-washed Chromosorb W that has been treated Gas-Chrom 2-Celatom acid washed and DMCS treated. Gas-Chrom Q-Celatom acid and base washed, and DMCS treated. Gas-Chrom (CL.H)-Celite treated with hexamethyldisilazane. Gas-Chrom (CL.S)-Acid washed and DMCS treated Celite. Anakrom (A .S.)-Siliconised, acid-washed Anakrom U. Anakrom (ABS)-Acid and alcoholic-base washed, and vacuum silanised L4nakrom U.The residual hydrochloric acid is removed by methanol washing. It is inert, suitable for steroid analysis and is available in a wide variety of 10-range-mesh cuts. An inert support prepared especially for steroid analysis. THE PINK DIATOMACEOUS SUPPORTS These are derived from calcined diatomite, without flux, and they have better handling They give greater efficiencies, but are less inert characteristics than the white supports. and not suitable for polar solutes. UNMODIFIED PINK DIATOMACEOUS SUPPORTS- Chromosorb P-Prepared by Johns Manville from firebrick, suitable for hydrocarbon separations. Gas-Chrom R-Prepared by Applied Science from Johns Manville Sil-o-Cel C22 firebrick (smaller surface area and less reactive than the pink supports prepared from calcined diatomite).Anakrom P-Prepared from calcined diatomaceous earth. Available in 10-range mesh cuts. Diatomite S- J. J’s. prepared firebrick.February, 19671 PART I. CHOICE OF SOLID SUPPORTS 81 ACID-WASHED PINK DIATOMACEOUS SUPPORTS- are, however, still unsuitable for polar materials. These are less active than the untreated firebrick, due to removal of surface iron. They Chromosorb P ( A . W.)-Acid-washed Chromosorb P. Gas-Chrom (R. A .)-Acid-washed Gas-Chrom R. Gas-Chrom (R.P.)-Acid and alcoholic-base washed Gas-Chrom R. Anakrom (PA .)-Acid-washed Anakrom P. SILANISED PINK DIATOMACEOUS SUPPORTS- These are the least active of the pink supports but still not as inert as the best white supports. They will make efficient columns for semi-polar materials but are not satisfactory for low loaded columns.Chromosorb P (HMDS)-Hexamethyldisilazane-treated Chromosorb P. Chromosorb P (A. W.-DMCS)-Acid washed and dimethyldichlorosilane-treated Chromo- Gas-Chrom R.2.-Acid-washed and silanised C22 firebrick. Wide variety of mesh cuts sorb P. available. Limited range of mesh cuts available. MISCELLANEOUS DIATOMACEOUS SUPPORTS Chromosorb G-This is specially prepared by Johns Manville as a support for gas chromatography. As it is very inert, robust and dense, it is particularly useful for columns of less than 5 per cent. weight loadings of liquid phase. (Because of its high density this is equivalent to about 10 per cent. with usual supports.) It is available in three mesh sizes 45 to 60, 60 to 80, 100 to 120 and modified as: Chromosorb G, untreated; Chromosorb G (A.W.), acid washed; and Chromosorb G (A.W., DMCS), acid washed and silanised.J.J’s. M-This is similar in property to Chromosorb G, and is available in acid washed and DMCS and HMDS-treated forms. Chromosorb A-This is a flux-calcined diatomite manufactured by Johns Manville especially for preparative columns. It is similar in appearance to Chromosorb P but less reactive. Its high surface area allows it to take up to 25 per cent. liquid loadings. High performance Chromosorbs G and W-These have been developed for steroid analysis and are available in A.W. and DMCS-treated forms. The surface is claimed to be highly inert and capable of HETP values of 0.45 mm. FLUOROCARBON SUPPORTS These supports are very inert, and useful for aqueous or other highly polar samples, but they are very difficult to handle and often produce very inefficient columns.Cool them to 0” C before use; they are fragile. Chromosorb T-This is prepared by Johns Manville from Teflon G and is available in mesh sizes 30 to 60 and 40 to 60. FZuoroport T-This is supplied by Applied Science. It has a wide mesh range, and is derived from Teflon 6. KeZ-F 6051-Chlorotrifluoroethylene manufactured by the Minnesota Mining and Manufacturing Co. Flztoropak-80-Fluorocarbon support produced by the Fluorocarbon Co. Tee-Six-Processed by Analabs from Teflon 6. It is available in a large variety of REFERENCES mesh sizes down to 160 to 170 mesh-cut in 10-mesh-cut ranges. 1. 2. 3. 4. 5.6. 7. 8. 9. Rose, B. A., Analyst, 1959, 84, 574. Graig, B. M., in Noebels, H. J., Wall, R. F., and Brenner, N., Editors, “Gas Chromatography,” James, A. T., and Martin, A. J . P., Biochem. J., 1952, 50, 679. Baker, W. J., Lee, E. H., and Wall, R. F., in Noebels, H. J., Wall, R. F., and Brenner, N., Editors, Ottenstein, D. M., “Progress in Industrial Gas Chromatography,” Volume 1, Plenum Press, New Saha, N. C., Giddings, J. C., Analyt. Chem., 1965, 37, 822. -, -- , Ibid., 1965, 37, 830. Littlewood, A. B., “Gas Chromatography,” Academic Press, New York and London, 1962, p. 213. James, A. T., Martin, A. J. P., and Smith, G., Biochem. J., 1952, 52, 238. Academic Press Inc., New York and London, 1961, p. 37. o p . cit., p. 21. York, 1961, p. 51.PALFRAMAN AND WALKER Fales, H.M., and Pisano, J. J., in Szymanski, H., Editor, “Biomedical Applicationsof GasChromato- Rurchfield, H. P., and Storrs, E. E., “Biochemical Applications of Gas Chromatography,” Academic Ormerod, E. C., and Scott, R. P. W., J . Chromat., 1959, 2, 65, Horning, E. C., Moscatelli, E. A., and Sweely, C. C., Chem. & Ind., 1959, 751. Horning, E. C., Vanden-Heuvel, W. J . A., and Creech, B. G., in Glick, D., Editor, “Methods of Biochemical Analysis,” Volume XI, Interscience Publishing Co. Ltd., New York, 1963, p. 80. Bohemen, J., Langer, S. H., Perrett, R. H. and Purnell, J. H., J . Chem. Soc., 1960, p. 2444. Perrett, R. H., and Purnell, J . H., J . Chromat., 1962, 7, 455. Kirkland, J. J ., in Fowler, L., Editor, “Gas Chromatography,” Academic Press, New York and Staszewski, R., and JanAk, J., Colln Czech.Chem. Commun., 1962, 27, 532. Suppina, W. R., Kruppa, R. F., and Henly, R. S., J . Amer. Oil Chem. SOC., 1965, 42, 459A. Purnell, H., “Gas Chromatography,” John Wiley and Sons Inc., New York and London, 1962, Applied Science Newsletter, 1966, 7, No. 2. Urone, P., and Parcher, J . F., J . Gas Chromat., 1965, 3, 35. Varian Aerugraph Research Notes, Spring 1966. Vanden-Heuvel, W. J. A., Gardiner, W. L., and Homing, E. C., Analyt. Chem., 1963, 35, 1745. Scholz, R. G., and Brandt, W. W., in Noebels, H. J., Wall, R. F., and Brenner, N., Editors, 09. Ottenstein, D. M., J . Gas Chromat., 1963, 1, 11. Averill, W., in Noebels, H. J., Wall, R. F., and Brenner, N., Editors, op. cit., p. 1. Kirkland, J. J., Analyt. Chem., 1963, 35, 2003.Sawyer, D. T., and Barr, J. K., Ibid., 1962, 34, 1518. Landault, C., and Gurochen, G., J . Chromat., 1962, 9, 133. Onaka, T., and Okamoto, T., Chem. Pharm. Bull., Tokyo, 1962, No. 10, 757. Kirkland, J. J., in Goldup, A., Editor, “Gas Chromatography 1964,” Institute of Petroleum, Eggertsen, F. T., Knight, H. S., and Groennings, F., Analyt. Chem., 1956, 28, 303. Kiselev, A. V., Q. Reu. Chem. SOC., 1961, 15, 99. -, in Goldup, A., Editor, op. cit., p. 238. Pope, C. G., Analyt. Chem., 1963, 35, 654. Brodasky, T. F., Ibid., 1964, 36, 1604. HalBsz, I., and Horvath, C., Ibid., 1964, 36, 1178. Kirkland, J. J., Ibid., 1963, 35, 1295. Bugosh, J., U.S. Patent 2,915,475, December, 1959. “Baymal Colloidal Alumina,” E. I. Du Pont de Nemours and Co., Bulletin A 18211, 1961. Gohlke, R. S., and McLafferty, F. W., Paper presented a t the American Chemical Society Meeting, Desty, D. H., and Harbourn, C. L. A., Analyt. Chem., 1959, 31, 1965. Decora, A. W., and Dinneen, G. U., Ibid., 1960, 32, 164. - -- , in Noebels, H. J., Wall, R. F., and Brenner, N., Editors, op. cit., p. 33. Bens, E. M., APzalyt. Chem., 1961, 33, 178. Porcaro, P. J., and Johnston, V. D., Ibid., 1961, 33, 361. Callear, A. B., and Cvetanovits, R. J., Can. J . Chew, 1955, 33, 1256. Littlewood, A. B., “Gas Chromatography,” Academic Press, New York, 1962, p. 210 Hawkes, S. J., Russell, C. P., and Giddings, J . C., Analyt. Chem., 1965, 37, 1523. Littlewood, A. B., in Desty, D. H., Editor, “Gas Chromatography 1958,” Butterworths Scientific Publications, London, 1958, p. 23. Frederick, D. H., Miranda, B. I., and Cooke, W. D., in Brenner, N., Callen, J. E., and Weiss, M. D., Editors, “Gas Chromatography,” Academic Press lnc., New York and London, 1962, p. 27. --- , Analyt. Chena., 1962, 34, 1521. Dewir, R. ’A., and Maier, V. E., J , Chromat., 1963, 11, 295. Giddings, J. L., Analyt. Chem., 1962, 34, 458. Ohline, R. W., and Jojola, R., Ibid., 1964, 36, 1681. Golay, M. J. E., in Scott, R. P. W., Editor, “Gas Chromatography 1960,” Butterworths Publishing Co. Ltd., Edinburgh and London, 1960, p. 139. Webb, J. J,., Smith, V. E., and Wells, H. W., J . Gas Chromat., 1965, 384. Kirkland, J. J., Analyt. Chew., 1965, 37, 1458. McKinney, R. W., J . Gas Chromat., 1965, 388. Lloyd, W. G., and Alfrey, T., Paper presented at the 130th American Chemical Society Meeting, graphy,” Plenum Press, New York, 1964, p. 41. Press Inc., New York and London, 1962, p. 41. London, 1963, p. 77. p. 236. cit., p.7. London, 1965, p. 146. Dallas, Texas, April, 1956. Division of Polymer Chemistry, St. Louis, March, 1961. , , J . Polym. Sci., 1962, 62, 301. -- Moore, J. C., Ibid., 1964, 2, 835. Billmeyer, F. W., jun., “Textbook of Polymer Science,” Interscience, New York, 1962, p. 341. Hollis, 0. L., Analyt. Chew., 1966, 38, 309. Hollis, 0. L., and Hayes, W. V., J . Gas Chromat., 1966, 235. Received August 25th, 1966 82 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66.