ANALYST, SEPTEMBER 1993, VOL. 118 1111 Spectroscopic Probes for Hydrogen Bonding, Extraction Impregnation and Reaction in Supercritical Fluids" Andrew 1. Cooper, Steven M. Howdle,t Catherine Hughes, Margaret Jobling, Sergei G. Kazarian,* Martyn Poliakofft and Lindsey A. Shepherd Department of Chemistry, University of Nottingham, Nottingham, UK NG7 2RD Keith P. Johnston Department of Chemical Engineering, University of Texas, Austin, TX 78712, USA Spectroscopy is used for monitoring a number of processes relevant to solution, extraction and impregnation in supercritical C02 (scC02). Examples include: a combined infrared (IR) and ultraviolet study of the interaction between para-hydroquinone (HQ) and tributyl phosphate in scC02, which reveals hydrogen bonding, detected by the characteristicv(0-H) IR bands; IR measurement of the solubility of c ~ M n ( C 0 ) ~ (Cp = q5-C5H5) in scC02 as a function of temperature and pressure; an investigation of the uniformity of supercritical impregnation of CpMn(C0)3 into 4 mm diameter pellets of polyethylene (PE) using Fourier- transform infrared (FTIR) microscopy and FTIR depth profiling by photoacoustic detection; and an IR study of the photochemical reaction of c ~ M n ( C 0 ) ~ with N2 with PE film.Keywords: Supercritical fluid; hydrogen bonding; impregnation; polyethylene; photochemical reaction The role of supercritical fluids in analytical chemistry is the focus of increasing interest not only because of the inherent potential of supercritical fluidsl-3 themselves but also because they represent an environmentally more acceptable alterna- tive to many of the solvents currently in use in the analytical laboratory.Most of the work in this area has concentrated, quite rightly, on applications of supercritical fluids, particu- larly in the areas of extraction and use of modifiers to enhance the solubility of individual solutes.2 By contrast, there has been relatively little development of spectroscopic techniques to probe and monitor these processes in situ. It is the purpose of this paper to describe new applications of infrared (IR) spectroscopy to a range of problems associated with the use of supercritical fluids in an analytical context. These applications have largely evolved from the study of chemical reactions in supercritical fluids at Nottingham ,4-8 which required, inter a h , the development of versatile IR and ultraviolet (UV) cells for monitoring the progress of the reactions.9-11 Our results are divided into three sections: (i) a combined IR and UV study to establish the role of hydrogen bonding in the action of tributyl phosphate (TBP) as a modifier for the solubilization of hydroquinone in super- critical C02 (scC02); (ii) the use of IR techniques for monitoring the impregnation and extraction of organometallic compounds from polyethylene (PE); and (iii) a study of photochemical reactions of organometallics within polymers to explore the potential of these compounds for in situ derivatization and spectroscopic labelling of polymer addi- tives.Experimental Spectroscopy The UV spectra were obtained on a Perkin-Elmer Lambda 5 spectrometer with Epson data station.Transmission IR spectra were recorded on a Nicolet Model 730 interferometer and 680D data system (16K data collection, 32K transform * Based on a lecture presented at the Current and Future Applications of Supercritical Fluid Extraction meeting of the Western Region of the Analytical Division of The Royal Society of Chemistry, Cardiff, UK, October 2, 1992. t To whom correspondence should be addressed. * Permanent address: Institute of Spectroscopy, Russian Academy of Sciences, 142092 Troitzk, Moscow Region, Russia. points, 2 cm-1 resolution). Fourier-transform infrared ( R I R ) microscopy was carried out at 8 cm-1 resolution with a NicPlan IR microscope with SpectraTech automatic trans- lation stage, using a Nicolet Model 730 interferometer.Photoacoustic measurements were made at 8 cm-1 on a Bio-Rad Model 896 step-scan FTIR interferometer equipped with a demodulator accessory for phase modulation experi- ments and fitted with an MTEC photoacoustic detector. Materials The following reagents were all used without further purifica- tion: c ~ M n ( C 0 ) ~ (Cp = q5-C5H5) (Strem), TBP (BDH), para-hydroquinone (HQ; BDH), C02 (Air Products SFC grade), N2 (Air Products) and low density PE pellets (BP Research) were used without further purification. The PE film was formed from powdered, low-density PE (Aldrich) using a constant thickness melt press Specac (Model 15620). 12 Super- critical impregnation was carried out as described in detail elsewhere ,11712 by immersing PE pellets overnight in a near-saturated solution of CpMn(C0)3 in scC02 at about 40 "C and 2000 psi pressure (145 psi = 1 MPa) with a modified Nupro TF Series in-line filter, 5 ym pore size, as the pressure vessel.12 Pellets were sectioned with a Leitz sledge microtome.Results and Discussion Hydrogen Bonding Hydrogen bonding has been intensively studied for many years and the process is well understood in both gas and condensed phases. 13 Supercritical fluids offer the unique opportunity to observe the effect on hydrogen bonding of varying the density of the medium without altering the concentrations of either proton donor or acceptor.14.15 More importantly, hydrogen bonding appears to be a major factor in the action of many, if not the majority, of the modifiers currently used to enhance the solubility of particular com- ponents in supercritical extraction.A better understanding of hydrogen bonding in supercritical fluids should, therefore , lead eventually to a more rational approach to the design and choice of modifiers for individual applications. There have been a number of studies of supercritical hydrogen bonding,14Js particularly by Yee et aZ.14 The usual approach has been to use IR spectroscopy to detect the characteristic shift in the v(0-H) vibration of the proton1112 ANALYST, SEPTEMBER 1993, VOL. 118 2.9 2.4 Q) 2 1.9 e 8 2 1.4 0.9 0.4 _. . 195 215 235 255 275 295 315 335 Wavelengthlnm f(O-H---O) HQ 3400 3200 3000 2800 Wavenumbedcm-1 Fig. 1 (a) UV and ( b ) IR absorption spectra showing the effect of increasing amounts of TBP on the solubility of HQ in scC02.The IR spectra cover the v(C-H) region of TBP itself and the ~ ( 0 - H ) bands of the hydrogen-bonded HQ-TPB complex. Each trace (UV or IR) was recorded with a separate solution with a different concentration of TBP but the IR and UV spectra for the same concentration of TBP were recorded from the same solution in the same cell (2 ml volume and optical pathlength 5 mm). In each case, a measured amount of TBP was syringed into the cell, containing excess solid HQ. The cell was then pressurized with scC02 to 3700 psi (c 25 MPa). The traces in the UV spectra (a) correspond to additions of 4, 14 and 24 pl of TBP and those in the IR spectra (b) to 0 , 4 , 14,24 and 34 pl of TBP donor, which occurs on formation of a hydrogen bond.One problem is that, in scC02, the absorptions of the scC02 itself obscure the ~ ( 0 - H ) bands of the free proton donor or, with deuteriated donors, the ~ ( 0 - D ) of the D-bonded complex.14 Clearly, this limitation can be overcome by switching to fluids that do not absorb in this region of the spectrum ( e . g . , C2H614 or SF615*16) but this option is not open if the aim is to study bonding in scC02 under realistic analytical conditions. Here, we illustrate a different approach, the combined use of UV and IR spectroscopy. Tributyl phosphate is well known for its use in a wide range of extraction processes in conventional solvents and for its ability to form hydrogen bonds.17 Recently, Lemert and Johnston18 reported how TBP could be used as a far more effective modifier than CH30H in scCO2 to increase the solubility of HQ by over two orders of magnitude.para- Hydroquinone has a higher melting point and lower vapour pressure than the ortho-isomer; the consequence of intermol- ecular hydrogen bonding. It was, therefore, postulated18 that the effect of TBP as a modifier is to hydrogen bond to the OH groups of HQ. This proposition has now been tested spectro- scopically. The strategy has been to use UV absorption to detect the total amount of HQ in solution and IR spectroscopy not only to establish the amount of TBP in solution but also to detect the presence of hydrogen bonding via the shifted ~ ( 0 - H ) bands. O=P /OBU -0Bu OBu H O G O H HQ \ TBP "0-0-H ---O=P(OBu)3 1 Fig. 1 shows the IR and UV spectra of a series of solutions in scC02 containing increasing amounts of TBP and in each case saturated with HQ.The UV spectra, [Fig. l(a)], clearly confirm that the concentration of HQ in scC02 increases with an increasing concentration of TBP. At the same time the IR spectra, [Fig. l(b)] indicate increasing concentrations of hydrogen-bonded species in solution, thus supporting the postulation that, in this case, the role of TBP involves hydrogen bonding. Unfortunately, IR is not a precise enough technique to distinguish easily between the bands caused by hydrogen bonding to one or to both of the OH groups in HQ, i.e., structures 1 and 2, or indeed to identify species containing more than one molecule of HQ. However, mathematical modelling18 of the effects of TBP on the solubility of HQ favoured the formation of an HQS(TBP)~ adduct (2).Cur- rently, we are applying similar techniques to simpler systems to quantify the role of the supercritical solvent in hydrogen bonding,16 using (CF3)3COH as the proton donor; (CF3)3COH has the advantage that it does not self-associate to any significant extent, thus simplifying the systems even further. Impregnation of Polymers In a supercritical extraction experiment, it is usually much easier to measure the amount of material, extracted into the fluid, which is initially 'clean', than it is to measure the amount of guest material remaining in the host matrix. Impregnation is the inverse of extraction and the converse is true of the ease of measurement. In impregnation experiments, it is much easier to measure spectroscopically how much material has pene- trated the host matrix because, at the start of the experiment, the host matrix does not normally contain any of the guest material to be impregnated. Transition metal carbonyl compounds are particularly suited to impregnation experiments.They are relatively volatile, strongly hydrophobic compounds with very intense IR absorptions, due to v(C-0) vibrations, the wavenumbers of which are highly sensitive to the environment.11 Thus it is possible to distinguish spectroscopically between a compound dissolved in scC02 and the same compound impregnated into PE. The way in which this property can be exploited for monitoring the impregnation and extraction of CpMn( CO), in thin 4 0 0 pm PE film has recently been described.11 In this paper, the work is concentrated on investigating more bulky samples, specifically, near spherical pellets about 4 mm in diameter.Such pellets are of considerable relevance to current models of supercritical extraction, including the Leeds 'Hot- Ball' model. 19320 Briefly, the impregnation experimentllJ2 involves immers- ing the polymer pellet in a solution of c ~ M n ( C 0 ) ~ in scC02. Impregnation is allowed to occur, the pressure is vented and the pellet removed. This study has had two aims: (i) to establish the uniformity of the impregnation of the pellets; and (ii) to investigate the solid residue left on the surface of the polymer, when the supercritical solution is depressurized. There have been relatively few studies21 on the solubility of carbonyl compounds in scC02 and no data were available for c ~ M n ( C 0 ) ~ .Therefore, a brief study of solubility as a function of temperature and pressure using the circulating system illustrated in Fig. 2 was carried out, by monitoring the concentration of dissolved c ~ M n ( C 0 ) ~ by FTIR. The results are summarized in Fig. 3. Like other solutes,' the solubility of CpMn(C0)3 in scC02 increases sharply with pressure and decreases with temperature at a constant pressure under the conditions used. Thus, although impregna- tion would be expected to proceed more effectively at higher temperatures, the concentration of CpMn(CO)3 in super- critical solution will be lower if the pressure is not increased. In the present experiments, therefore, a temperature close to the critical point of C02 was used.ANALYST, SEPTEMBER 1993, VOL.118 6 5 - 4 - 0 e B 3 - 2 0: 2 - 1 - 1113 (a) Once the pellet is impregnated, the strategy has been to microtome thin sections from the pellet and then to scan the section using an IR microscope with a 100 pm spot-size to measure the distribution of the impregnated carbonyl com- pound. Fig. 4 shows the result of one such scan across the diameter of a section, which indicates that the impregnation is remarkably uniform. Although, the concentration of scco, Fig. 2 Schematic diagram of the system used for monitoring the solubility of CpMn(C0)3 in scC02. The components are marked as follows; P, Micropump Model 180SC ultrahigh pressure circulating pump; IR, cell of 1 mm pathlength for IR detection of CpMn CO ; and R, Nupro in-line filter to act as a reservoir for solid CpMn[CO{i.All components, except the cell, were placed in a heated box; the cell itself and connecting pi ework were heated with heating tape. Additional scCO2 was afded as appropriate from a Lee Scientific Model 501 syringe pump. The whole system was allowed to equilibrate for 30-40 min before each measurement Pressure/psi 2.0 1.8 1.6 1.4 1.2 1 .o 20 30 40 50 60 70 Tern peratu rePC Fig. 3 IR spectra obtained with the apparatus illustrated in Fig. 2, showing the qualitative effects of (a) pressure and (b) temperature on the solubility of CpMn(C0)3 in scC02. (a) Effect of pressure at 32 "C. The ordinate scale refers to the absorbance of the al v(C-0) band of CpMn(C0)3 and the maximum absorbance value, about 6, corre- sponds to an approximate concentration of 2 x mol 1-l.(b) Effect of temperature at a constant pressure of 2000 psi (e 13.8 MPa). Note that in ( a ) , the data are presented as if IR absorbance can be measured accurately to a value of >6. In reality, these high absorbance values were obtained by a two stage process: (i) while the v(C-0) bands were in the absorbance range 2-3, the absorbance values were measured for the first overtone and combination bands about 4000 cm-1, which are inherently much weaker than the fundamental vibrations, and their intensities relative to the fundamen- tal bands were established; (ii) when the intensity of the v(C-0) bands exceeded an absorbance of 3, the absorbance values of the overtones could be used to calculate the 'absorbance' of the fundamental bands c ~ M n ( C 0 ) ~ is higher close to the outside of the pellet, overall it varies by less than 30%. Two dimensional scans, such as that illustrated in Fig.5 , provide very similar results and the high symmetry of the contours in Fig. 6 suggest that sectioning of the pellet with the microtome causes minimal distortion of the pellet and has little effect on the distribution of c ~ M n ( C 0 ) ~ . As the overall procedure (impregnation, sectioning, spec- troscopy) was relatively drawn out, these results do not show the distribution of c ~ M n ( C 0 ) ~ within the pellet at the moment immediately following the end of impregnation. Recent experiments in our laboratory22 have suggested that molecules the size of CpMn(C0)3 are relatively mobile, at least over short distances, in PE at room temperature over a period of minutes.Nevertheless, the fact that the present experiments revealed a slightly higher concentration of C ~ M n ( c 0 ) ~ close to the outside of the pellet suggests that bulk diffusion of c ~ M n ( C 0 ) ~ within the PE must be slow on the timescale of the experiment (i.e., over a period of days). Experiments with finely powdered KBr have shown" that c ~ M n ( C 0 ) ~ deposited from scCO2 solution onto the surface has an IR spectrum with relatively sharp v(C-0) bands, a spectrum quite distinct from the very broad bands normally observed with bulk solid CPM~(CO)~, possibly due to the amorphous nature of the deposited material.Although the IR spectrum of this deposited CpMn( CO), is also significantly different from that of the compound impregnated into PE, it is a challenging task to use IR spectroscopy for the characteriza- tion of the deposits of CpMn(C0)3 on the surface of the 4 mm PE pellets. Direct transmission spectroscopy is not practic- 2100 2000 1900 Wave n u m be r/cm - 1 Fig. 4 Results of the FTIR microscopic investigation (100 pm spot size) to establish the distribution of CpMn(COh, impregnated into a PE pellet. (a) Schematic view of the pellet showing a section (about 50 pm thick) and the path scanned across it by the FTIR microscope. (b) Computer-generated display of the distribution of CpMn(C0h across the pellet. The ordinate axis shows the ratio of the absorbance the al band of CpMn(C0)3 to the absorbance of a weak band of PE itself (2017 cm-I).By using the ratio of bands rather than the absorbance, any effects that might be caused b non-uniformity in the thickness of the section can be eliminated. (cf A spectrum showing the v(C-0) bands of impregnated CpMn(C0)3 with the al mode arrowed and absorptions of PE removed1114 1.8 Fig. 5 Two-dimensional FTIR microscope scan of a microtomed section obtained by exactly the same method as that described for Fig. 4 3.0 E E R 0 1.8 xlm m 0 Fig. 6 FTIR contour plot corresponding to the data illustrated in a perspective view in Fig. 5. The contours give the values of the absorbance ratio (as defined in Fig. 4), which increase in the direction of the arrow and are plotted from (A) 6.3 to (G) 8.1, at intervals of 0.3.Note in particular, that the contours are essentially circular. If the microtome were distorting the pellet severely during sectioning, one would expect a corresponding distortion of the contours able, because the thickness of the surface layer is undoubtedly small compared with the dimensions of the pellet. Equally, the layer is thin compared with the diameter of the spot viewed through the IR microscope. Although diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS) has been highly successful in making such measurements on impreg- nated PE powder,” we have found DRIFT’S difficult to use with these pellets because the surface gives rise to sufficient specular reflection to degrade the spectrum. The solution to these problems has been to use FTIR with photoacoustic detection, because this technique has a rela- tively shallow penetration depth and, furthermore, offers the possibility of depth profiling.23-28 Photoacoustic detectors rely on absorption of IR radiation by the sample to heat a carrier gas, causing a pressure rise which is detected with a sensitive microphone.Depth profiling of a sample can be achieved in two ways, by varying the scan velocity or by so-called ‘step’ scanning. ANALYST. SEPTEMBER 993, VQL. 118 2060 1920 Wavenumber/cm-l Fig. 7 FTIR photoacoustic depth profiling of an intact impregnated PE pellet by means of interferometer mirror velocity. In this case, mirror velocity is defined by the frequency at which the data points are recorded.Spectrum A was obtained with the highest frequency, 20 kHz, corresponding to the shallowest depth of sampling. The other spectra were recorded at: B, 800 Hz; C, 200 Hz; and D, 50 Hz, corresponding to increasing depths of penetration. The bands are labelled as follows: S, surface deposited solid CpMn(C0)3; M, molecular CpMn(C0)3, impregnated into the bulk of the solid PE. Note that this particular sample was washed with heptane prior to recording the spectrum in an attempt to reduce the thickness of the surface layer. Given the restricted wavenumber region of the spectra, the ‘photoacoustic response’ of the ordinate scale of the spectra can be reasonably equated to absorbance. Altering the mirror velocity of the FTIR interferometer alters the modulation frequency of the IR radiation incident on the sample.In general, the more rapid the modulation, the shallower the depth of sample from which the photoacoustic signal can originate; fast mirror velocities yield information only from the surface region, whereas lower’velocities provide the sum of the spectra from the surface and from the layers beneath it.23-25 Fig. 7 shows a set of photoacoustic spectra of an impregnated pellet , recorded with different mirror velo- cities. These spectra confirm the existence of the surface layer but not its thickness. There is a clear transition from the bands of solid CPM~(CO)~, arrowed, to the superimposed spectra of solid and impregnated c ~ M n ( C 0 ) ~ as the mirror velocity is reduced. Step-scanning is a relatively new feature in commercial FTIR instrumentation.2c28 Although the detailed mechanics of the process are intricate,2628 the principle of operation is straightforward.In a conventional interferometer, the IR signal is not measured continuously but at regularly spaced points along the path of the moving mirror, points determined by the calibrating Heme laser. In a step-scan instrument, the moving mirror appears to spend a relatively long time stationary at one measurement position and then jumps almost instan- taneously to the next position.26-28 This behaviour permits additional modulation of the IR beam at a frequency that is fast compared with the time spent by the mirror at each measurement point. When this double modulation is applied to photoacoustic experiments, the detector can be linked to a phase-sensitive amplifier and, as described elsewhere ,2628 varying the phase of detection relative to the phase of modulation allows spectra to be obtained from different depths within the sample, the larger the phase angle the greater the depth. Fig.8 illustrates the effect of varying the phase angle in phase modulated photoacoustic spectra of a PE pellet impregnated with c ~ M n ( C 0 ) ~ . The spectra show a good transition from the spectrum of pure solid CpMn(C0)3 to that of pure impregnated CPM~(CO)~, in a manner that the spectra in Fig. 7 do not. However, the technique does not haveANALYST, SEPTEMBER 1993, VOL. 118 1115 an inherent scale of depth so, again, the precise thickness of the surface layer (probably only a few pm) cannot be established from these spectra. It is therefore, clear that step-scan photoacoustic measurements of this type have considerable promise for probing such samples but without extensive calibration the results are only semi-quantitative.Photochemical Reactions Within the Polymer Currently, there is considerable interest in the use of supercritical fluid extraction (SFE) for identification and quantification of polymer additives.2.29 Unfortunately, some classes of additives, e.g., those with amine functions, are often almost insoluble in scC02 because of reaction with the COz itself. * However, transition metal carbonyl complexes, e.g., CPM~(CO)~, undergo facile photochemical reactions with many such compounds.30 Work is currently underway to explore the feasibility of using carbonyl complexes for reactive extraction of additives, exploiting the carbonyl not only to t v) c 0 n 2 8 0 3 .- 4- 0 .c L 4- 'J 171" 0" 1 u 9" 1 18" B B 1 2060 1920 2060 1920 vlcm-1 Fig.8 FTIR photoacoustic depth profiling of an intact impregnated PE pellet using step-scan phase modulation. Spectra were recorded as the phase angle was varied from 0" to 180" in steps of 9" and a representative set of spectra are shown. From 45" to about 90" the bands in the spectra are those of solid CpMn(C0)3 and from 90" the arrowed bands of CpMn(C0)3 impregnated into the bulk of the PE gradually increase so that by 18" all trace of the surface bands have disappeared. In principle, the surface species might be expected to appear at a phase angle of 0 O but the naoture of the equipment appears to generate a phase shift of about 45 .As in Fig. 7, 'photoacoustic response' can be reasonably equated to absorbance sequester the reactive functional group of the additive but also to provide an excellent spectroscopic label. Carbonyl com- pounds have intense absorptions in both IR and UV and, therefore, should greatly increase the sensitivity of spectro- scopic detection of the complexed additives. For example, (C6H3Me3)Cr(C0)3 has an injected minimum detectable quantity of only 20 pg for capillary supercritical fluid chromatography-FTIR .31 As the first step in this investigation, the photochemical reactions of carbonyl compounds in pure PE film were examined. Fig. 9 shows spectra obtained during and after the UV irradiation of CPM~(CO)~, previously impregnated into low-density PE film, under a high pressure of N2 gas, as the added reactant.It can be seen that there are in fact two products, not only the expected dinitrogen complex32 C P M ~ ( C O ) ~ N ~ but also a complex formed by coordination of the c ~ M n ( C 0 ) ~ moiety to the pendant olefinic C=C bonds, which occur at random intervals down the polymer chain.2>33 This olefinic product can also be formed in the absence of N2 and appears not to be easily extractable from the polymer by subsequent SFE with scCO2. There are two reasons why this experiment is of relevance to reactive extraction. Firstly, it demonstrates that there is sufficient mobility within the polymer matrix for photochemical reactions of CpMn(C0)3 to occur.Secondly, it shows that, although reaction between carbonyl and polymer would clearly consume some of the carbonyl compound, such products will be unextractable so may well not interfere with the overall SFE process. Conclusions The experiments described in this paper have shown the IR and, to a lesser extent, UV spectroscopy can provide valuable in situ monitoring of processes of importance to the exploita- tion of supercritical fluids in analytical chemistry. Transition metal carbonyl complexes, a range of compounds not nor- mally associated with this area of chemistry, have the 4.4 3.3 al c m + 2.2 s 2 B B (a) 1 .I 0 2065 1983 1901 Wavenumberlcm-1 Fig. 9 IR spectra illustrating the photochemical reaction of CpMn(C0)3 with N2 in low density PE film, 500 pm thickness.The film was first impregnated with CpMn(C0)3 at 3000 psi (c 20.7 MPa). The C02 was vented, excess solid CpMn(C0)3 removed and the cell refilled with N2 at 3000 psi. (a) Spectra recorded over a 60 min period of UV irradiation, showing the decay of the bands, B, of im regnated CpMn CO and the growth of bands N, of CpMn(CO)2(N2fand P, of CpMn[CO]2(q2-polymer)1116 ANALYST, SEPTEMBER 1993, VOL. 118 spectroscopic properties needed for probing the impregnation and extraction of polymers. Supercritical fluid extraction is still a largely empirical technique. We believe that, in the near future, spectroscopic investigations will provide a new approach for studying supercritical extraction, an approach which will complement the methods in use today and will help to reach a better understanding of SFE.We thank SERC (Grant No. GR/G0823), the Royal Society, the Nuffield Foundation, Nicolet Instruments Ltd. and BP International Ltd. for support. We are particularly grateful to Dr. S. F. Parker and Dr. C . Baker for their help with the FTIR microscopy and Dr. A. Grady for his assistance with the photoacoustic measurements. We thank Dr. I. G. Anderson, J. A. Banister, Dr. G. Davidson, J. G. Gamble, Dr. T. J. Jenkins, Dr. M. A. Healy, T. Lynch, K. Stanley and Professor J. J. Turner for their help and advice. 1 2 3 4 5 6 7 8 9 10 11 12 References McHugh, M. A., and Krukonis, V. J., Supercritical Fluid Extraction, Butterworth, Boston, 1986. Analytical Supercritical Chromatography and Extraction, eds. Lee, M. L., and Markides, K.E., Chromatography Conferences, Provo, UT, 1990. Vigdergauz, M. S., Lobachev, A. L., Lobacheva, I. V., and Platonov, I. A., Russian Chem. Rev., 1992, 61,267. Howdle, S. M., and Poliakoff, M., J. Chem. SOC. Chem. Commun., 1989,1099. Howdle, S . M., Grebenik, P., Perutz, R. N., and Poliakoff, M., J. Chem. SOC. Chem. Commun., 1989, 1517. Howdle, S . M., Poliakoff, M., and Healy, M. A., J. Am. Chem. SOC., 1990, 112, 4804. Jobling, M., Howdle, S. M., Healy, M. A., and Poliakoff, M., J. Chem. SOC., Chem. Commun., 1990, 1278. Howdle, S. M., Jobling, M., George, M. W., and Poliakoff, M., Proceedings of the Second International Symposium on Super- critical Fluids (Boston), ed. McHugh, M. A., Johns Hopkins University, MD, 1991, p. 189. Poliakoff, M., Howdle, S. M., Healy, M.A., and Whalley, J. M., Proceedings of the International Symposium on Super- critical Fluids, ed. Perrut, M., SocietC Franc. de Chimie, 1988, p. 967. Howdle, S. M., Jobling, M., and Poliakoff, M., Supercritical Fluid Technology; Theoretical and Applied Approaches in Analytical Chemistry, eds. Bright, F. V., and McNally, M. E., ACS Symposium Series, 1992, 488, 121. Cooper, A. I . , Howdle, S. M., and Ramsay, J. M., J. Polymer Sci. Polymer Phys., in the press. Jobling, M., Ph.D. Thesis, University of Nottingham, UK, 1992. 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 The Hydrogen Bond-Recent Developments in Theory and Experiment, eds. Schuster, P., Zundel, G., and Sandorfy, C., North Holland, Amsterdam, 1976. Yee, G. G., Fulton, J. L., and Smith, R.D., J. Phys. Chem., 1992,96, 6172, and references therein. Gupta, R. B., Combes, J. E., and Johnston, K. P., J. Phys. Chem., 1993,97,707. Kazarian, S. G., Gupta, R. B., Johnston, K. P., and Poliakoff, M., unpublished work. Ferraro, J. R., and Peppard, D. F., J. Phys. Chem., 1961, 65, 539. Lemert, R. M., and Johnston, K. P., Ind. Eng. Chem. Res., 1991,30, 1222. Bartle, K. D., Boddington, T., Clifford, A. A., Cotton, N. J., and Dowle, C. J., Anal. Chem., 1991, 63,2371. Bartle, K. D., Clifford, A. A., and Cotton, N. J., Analyst, submitted for publication. Warzinski, R. P., and Holder, G. D., Proceedings of the Second International Symposium on Supercritical Fluids (Boston), ed. McHugh, M. A., Johns Hopkins University, MD, 1991, p. 161. Cooper, A. I., Kazarian, S. G., and Poliakoff, M., Chem. Phys. Lett., 1993, 206, 175. Graham, J. A., Grim, W. M., 111, and Fateley, W. G., in Fourier Transform Infrared Spectroscopy: Industrial and Labor- atory Chemical Analysis, eds. Ferraro, R., and Krishnan, K., Academic Press, New York. vol. 4. 1990. Urban, M. W., Polym. Mater. Sci. Eng., 1991, 64, 31. Yang, C. Q., Appl. Spectrosc., 1991,45, 102. Crocombe, R. A., Curbelo, R., Leonardi, J., and Johnson, D. B., in Eighth International Conference on Fourier Transform Spectroscopy, eds. Heise, H. M., Korte, E. H., and Seisler, H. W., Proc. SPIE Int. SOC. Opt. Eng., 1992, 1575, 189. Crocombe, R. A., Compton, S. V., and Leonardi, J., in Eighth International Conference on Fourier Transform Spectroscopy, eds. Heise, H. M., Korte, E. H., and Seisler, H. W., Proc. SPIE Int. SOC. Opt. Eng., 1992, 1575, 193. Lerner, B., Nicolet FTIR Technical Note, TN-9253, Nicolet Instruments, Madison, WI, 1992. Ashraf-Khorassani, M., Boyer, D. S., Cross, K., Levy, J. M., and Houck, R. K., Proceedings of the Second International Symposium on Supercritical Fluids (Boston), ed. McHugh, M. A., Johns Hopkins University, 1991, MD, p. 219. Geoffroy, G. L., and Wrighton, M. S., Organometallic Photo- chemistry, Academic Press, New York, 1979. Jenkins, T. J., Kaplan, M., Davidson, G., Healy, M. A., and Poliakoff, M., J. Chromatogr., 1992, 626, 53. Sellmann, D., Angew. Chem. Int. Ed. Engl., 1971, 10,919. Cooper, A. I., Clark, M., Jobling, M., Howdle, S. M., and Poliakoff, M., unpublished work. Paper 31012711 Received March 4, 1993 Accepted May 5, 1993