Analyst, October, 1967, Vol. 92, $$. 639-641 639 The Collection of Fractions Separated by Gas - Liquid Chromatography Part II.* The Direct Transfer of the Fraction from the Trap to a Silver Chloride Infrared Cell or a Nuclear Magnetic Resonance Spherical Microcell BY I. A. FOWLIS AND D. WELT1 (Unilevev Research Laboratory, Colworth House, Shavnbrook, Bedfod) Gas-chromatographic fractions are trapped in column packing. They are then eluted directly into silver chloride infrared microcells or nuclear magnetic resonance spherical microcells immersed in liquid nitrogen. A gas flow through the heated trap is caused by the condensation of the argon in the cells. A closed gaseous system is used. THE introduction of commercial silver chloride infrared cells has simplified the technique described by Howlett and Weltil for recovering samples trapped from a gas chromatograph for examination by infrared spectroscopy.The fractions are trapped, as before, in tubes packed with column-packing material, but they are now eluted directly into silver chloride cells cooled in liquid nitrogen. The examination of small amounts (about 50 pg) is easier and more reliable. Alternatively, the fractions are eluted directly into cooled nuclear magnetic resonance spherical microcells. EXPERIMENTAL PROCEDURE- Infrared-The apparatus (Fig. 1 (a)), which consists of a 100-ml round-bottomed flask containing desiccant (5 g of silica gel), a vacuum gauge (0 to 30 inches of mercury) and a stopcock, is evacuated to a pressure of 25 inches of mercury and then re-filled with argon.Although vacuum grease is used in the stopcock and on the flask connection, no grease should be used on the trap connections. A Research and Industrial Instrument Co. silver chloride infrared cell is fitted with a silicon rubber cap, and a trap containing 5 per cent. Apiezon L. on Celite column packing is pushed tightly into a hole bored in the cap. A heater is placed round the trap, which is then attached to the evacuated system. When the trap temperature reaches 200" C, a Dewar flask containing liquid nitrogen is raised up to the cell until the latter is almost completely immersed. Condensation of the argon and air in the cell produces a flow of gas down through the trap. No interfering hydrocarbon impurities will be produced if the trap has been pre- treated at 200" C.The chromatographic fraction is eluted from the trap and re-condensed in the infrared cell. The vacuum in the system reaches equilibrium at a pressure of between 15 and 20 inches of mercury and is allowed to remain at this level for 5 minutes, the liquid nitrogen level being maintained the whole time. The heater is then switched off and the liquid nitrogen allowed to evaporate. The liquid argon evaporates slowly until atmospheric pressure is again attained, the solute remaining on the cold wall of the cell. This technique works equally well with air or argon, but the use of argon reduces the risk of oxidation by liquid oxygen. * For details of Part 1 of this series, see reference list, p. 641.640 FOWLIS AND WELTI: COLLECTION OF FRACTIONS SEPARATED [Anahst, Vol.92 The infrared cells of 0-025-mm path-length hold 1 pl of liquid and the 0-01-mm cells, about 0.2 p1. As the former are easier to fill than the latter, it is preferable to use the 0.025-mm cells and fill them with an infrared-transmitting solvent to avoid the interference fringes caused by a partly empty cell. However, if the sample is small and needs to be examined as a liquid, or is immiscible with infrared solvents, then it is preferable to use a 0-01-mm cell. A = 100-mi flask E = Silver chloride infrared cell B = Vacuum gauge (0 t o 30 F = Nuclear magnetic resonance C = Trap G = Liquid nitrogen D = Heater inches of mercury) micros p he re Fig. 1. Apparatus ( a ) , for recovery into silver chloride infrared cell; ( b ) , for recovery into nuclear magnetic resonance spherical microcell If, after removing the trap from the rubber cap an infrared solvent can be used, the solute is washed down with 1 p1 of the solvent by using a Hamilton syringe.A B5 stopper is fitted into the cap and the cell placed into its holder. The holder and cell are fitted into a polystyrene moulding and spun in a centrifuge for 2 minutes. If no solvent is used, the cap is stoppered and the cell at once spun in a centrifuge. The cell and holder are then placed into the optimum position of the light beam of the spectrometer. Solvent absorption in 0-01 and 0.025-mm cells is generally so weak that no reference solvent cell is needed. Bands like those of carbon tetrachloride a t wavenumbers 700 to 800 cm-1 are shown strongly, but no more information is lost than with a balanced cell system.The cell stoppers must be such as to prevent evaporation of the sample in the light beam. The cells are emptied by inversion in a centrifuge spinning at 4000 r.p.m. for several minutes. Nuclear magnetic resonance-The apparatus is the same as that used for infrared cells. The spherical microcell is attached by a piece of silicone rubber tubing to the underside of a B5 socket that has been drawn out into a tube of the same diameter as the stem of the microcell (Fig. 1 ( b ) ) . The two pieces of glass are made to touch to reduce the amount of exposed rubber to a minimum.October, 19671 BY GAS - LIQUID CHROMATOGRAPHY. PART I1 641 The experimental procedure is similar to that used for the infrared cell, except that with the cell of smaller volume less gas is condensed and, therefore, the vacuum has not exceeded 15 inches of mercury.The cell is not spun in a centrifuge, the sample being washed down into the sphere with 50 p1 of carbon tetrachloride containing a reference standard. The sample is then examined in the way described by Frost, Hall, Green and Leane.2 W c W n 0 3 E I I I I I I I I . I -7- I 650 Wavenumber, cm-1 Fig. 2. Infrared spectra of ethyl hexanoate after collections from gas chro- matograph with 10 : 1 splitter: curve A, 0.2 p1; curve B, 0.1 pl; curve C, 0.05 p1. Silver chloride,cell, path-length, 0.025 mm RESULTS AND DISCUSSION Infrared spectra obtained with a Unicam SP200 from 0.2, 0.1 and 0.05-pl samples of ethyl hexanoate are shown in Fig.2. Recoveries calculated on the basis of the carbonyl band intensity are nearly 100 per cent. for samples injected directly into a trap and about 95 per cent. for samples collected as fractions from a gas chromatograph. Recoveries of 0.2-pl samples of eugenol, either injected directly into a trap or collected from a gas chromatograph, were both 90 per cent. The elution of high boiling alcohols or phenols, such as eugenol, from traps packed with Apiezon L. on Celite is more difficult because of band tailing. It is therefore important that the traps are thoroughly stripped before being used again. Nuclear magnetic resonance spectra of trapped and normal samples of 0-2 p1 of ethyl- benzene have been obtained as an accumulation of 274 scans on a Northern Instrument NS 544 computer of average transcients, with 5 per cent. trifluoroacetic acid as the triggering material on a Perkin-Elmer R10 60m/c spectrometer. The recovery efficiency was about 95 to 100 per cent. The authors thank Mr. D. Frost for his advice on nuclear magnetic resonance spectro- metry. REFERENCES 1. 2. Howlett, M. D. D., and Welti, D., Analyst, 1966, 91, 291. Frost, D., Hall, G. E., Green, M., and Leane, J. B., Chem. & Ind., 1967, 116. NOTE-Reference 1 is to Part I of this series. Received March 31st, 1967