by
Separation of Some Low Molecular Weight Fluorocarbons Gas Chromatography SIR: I t is the purpose of this note to report a liquid phase that yields excellent results for the separation of certain low molecular weight fluorocarbons. The efficacy of the material, CHF CHC02CH2(CF2CF2)3H, is superior to liquid phases previously reported in the literature and is somewhat difficult to rationalize. As an adjunct to gas liquid chromatographic (GLC) separation, separations were also carried out by gas solid chromatography (GSC) on a silica gel column. Reed ( 7 ) evaluated several liquid phases for the GLC separation of some fluorocarbons (C5 to Co). He concluded that fluorocarbon and chlorofluorocarbon media give better resolution of fluorocarbon mixtures than hydrocarbon media after evaluating di(2-ethylhexyl) sebacate, C1(CF2CFC1)3CF2COOCsHb, Kel-F No. 90 grease, perfluorokerosine, and perfluorotributylamine. Campbell and Gudzinowicz (3) reported the separation of some C1 to C, fluorocarbons on S o . 3 Kel-F oil columns of various lengths; one was augmented with a short column containing diisodecylphthalate. Unfortunately, Kel-F oil did not constitute a completely satisfactory liquid phase for fluorocarbon separation. EXPERIMENTAL
Apparatus and Reagents. The gas chromatographic apparatus consisted of a Beckman GC-2A gas chromatograph to monitor the column effluent and a 20-foot coiled column maintained a t 0' C. The column was constructed from '/,-inch 0.d. copper tubing and was packed with packing material consisting of 45- to 60-mesh Chromosorb W and CHFCHCOZCH?(CF2CF2)3H(courtesy of E. I. du Pont de Kemours & Co., Wilmington, Del.) in the weight ratio of 70 grams of Chromosorb W to 30 grams of the liauid substrate. The packing material was prepared in the usual manner by dissolving CH2=CHC02CH2(CF2C F A H in ethyl ether. The carrier gas was helium, and the flow rate measured a t room temperature was 95 cc. per minute, For GSC, the apparatus was similar to that described by Greene, Moberg, and Wilson ( 4 ) . Silica gel (Davison Chemical Co., Division of W. R. Grace Corp., Baltimore, Md.) was ground to 50- to 80-mesh and packed in a 10-foot, 1/4-inch 0.d. copper tube. Fluorocarbons were obtained from various sources. CF4, C3Fs, and a mixture of cis- and trans-C4Fg-2 were purchased from the Matheson Co., East Rutherford, N. J. Cyclo-CZa (Freon-318) was purchased from E. I. 928
ANALYTICAL CHEMISTRY
5
06-
E
I Lu n 2 0
04'
0
2
4
6
5
I3
I2
I6
I4
I8
I
T W E (min.'
Figure 1.
Gas liquid chromatogram of fluorocarbons
du Pont de Kemours & Co., Wilmington, Del, CzFa, C2F4, %CQF~,and iso-CZ8 were obtained by the pyrolysis of cyclo-Cz*. Cyclo-CP6 was prepared by the photolysis of C2Fd using a low pressure mercury arc ( I ) . Procedure. The retention volumes of CF4, CJ?,, and cyclo-CcF, were obtained in the usual manner from the pure samples. The retention volumes of cis- and trans-CJi'~k2 were determined after the individual components of the mixture had been identified by infra-
red spectrometry. The retention volume data for C2Fe, C,B, and n-CaFa were obtained from the analysis of the pyrolysis products of cyclo-C4Fs between the temperature interval of 400' to 1000° C. The cyclo-C4F~was pyrolyzed by passing it through a l / ~ inch 0.d. stainless steel tube, 16 inches in length, a t a flow rate of 1.0 cc. per minute. At 400' C., only two peaks were observed which correspond to the equilibrium (6) Cyclo-CIFs
~
2C2F4
2 N
L.
Lz
C
15
30
45
T ' W E Im n:
Figure 2.
Gas solid chromatogram of fluorocarbons
60
At B5U' C., a single additional peak was generated and was identified as n-C3Fe (2). Verification was obtained from the analysis of the pyrolysis products of Teflon, heated a t 650' C. (3, 6), which produced the identical peaks. 4 t 800' C., cis- and trans-CaF~-2and C3Fs were identified, as was .SO-CJ?~,by the method described by Campbell and Gudsinowicz ( 3 ) . Between 700" to 800' C., the yield of C2F4 decreased sharply while the adjscent peak, which was assigned to C2F&( 2 ) , appeared and increased in size. DISCUSSION
A gas liquid chrom:itogram of a comples, gaseous, fluorocarbon mixture is shown in Figure 1. The individual components of cis- and trans-C4F8-2 niisture were identified by infrared
spectrometry. This c i s form gave characteristic absorption bands a t 5.78, 7.42, 9.0, 10.5, and 13.83 microns, while the trans form gave absorption bands at 7.74, 11.27, and 14.6 microns. Cyclo-CaFe and n-CaFs were not resolved. The small peak which emerges immediately before n-CaF8 was assigned to C2F2; it is difficult to conceive of another assignment. The small peak preceding iso-CdFs is probably octaBuoro-l-butene (C4F8-1). A gas solid chromatogram of a similar gas mixture is shown in Figure 2. The temperature was programmed from room temperature to 180' C. and the separation was completed in 60 minutes. The cis and trans isomers of C4F8-2 were not resolved, but n- and cyclicC,F6 could be resolved.
LITERATURE CITED
(1) Atkinson, B.,
2684.
J. Chem.
SOC. 1952,
(2) Atkinson, B., Atkinson, V. A., Ibid., 1957,2086.
(3) Campbell, R. H., Gudzinowics, B. J., ANAL.CHEM.33, 842 (1961). ( 4 ) Greene, S.A,, Moberg, M. L., Wilson, E. M., Ibid., 28, 1369 (1956). (5) Larcher, J. R., Tompkin, G. W., Park, J. D., J. Am. Chem. SOC.74, 1693 (1952). (6) Lewis, E. E., NayIor, >I. -4., Ibid.,69, 1968 (1947). (7) Reed T. M., 111, . ~ N A L . CHEX 30, 221 (1958).
STANLEY A GREENE FRANCIS M. WACHI
Materials Sciences Laboratory Aerospace Corp. P. 0. Box 95085 LOBAngeles 45, Calif.
The Pyrolytic Graphite Electrode as an Indicating Electrode for Pote nt io imet ric Titrations SIR: Preliminary investigation has been made of the pclssibility of using pyrolytic graphite as an indicator electrode. Although therz have been many instances of the use 0' various types of graphite and carbon electrodes ( I ) , the only reported use cf pyrolytic graphite as an electrode has been the work of Laitinen and Rhodes (4) in fused salts. Since pyrolytic graph ltcl is a relatively nen- material that has unique properties, the further investigation of its analytical applications is necessary. Information on the history, ,properties, and applications of pyrolytic graphite can be found in commercial publications and elsewhere ( 2 , 3 , 5 ) . Pyrolytic graphite k a polycrystalline form of carbon that l-as a high degree of orientation. The material is produced by vapor phase deposition. The result of the deposition is the growth of a material that has a metallic-type behavior in the plane of deposition and a ceramic-type behavior in the direction perpcndicular to the plane of deposition. The properties that rcnder pyrolytic graphite of interest as an indicator electrode are a high degree of impermeability, freedom from triipped gases and metallic contaminant:,, and chemical inertnew The particular pyrolytic graphite that \yas used in this experimental work was obtitined from High Temperature Materials, Inc., Boston, ;\insy. Potentiometric titmtions were performed as a means of investigation. The potentiometric apparatus consisted of a Beckman model 11-2 p H meter, a Beckman calomel e.ectrode, Model S o . 39170 with a fiber tip, and a Beck-
man glass electrode, Model S o . 40498. The calomel electrode was used as a reference electrode for pH and redox titrations; the glass electrode as a reference for redox titrations. A syringe-type buret that was obtained from the Micro-Metric Instrument Co., Cleveland, Ohio, was used for the addition of titrant. The buret was equipped with either a 1-m1.-per-inch or a 3-m1.per-inch syringe. Stirring of the solutions was accomplished by use of a magnetic stirrer and stirring bar. The indicating electrode was fabricated by sealing a '/4-inch cube of pyrolytic graphite into the flared end of a borosilicate glass tube, 7 mm. in diameter, with epoxy cement. Only the face of the "a" plane of the pyrolytic graphite
-6 bV.1
-600
,
, -400
I
1
-200
t a
203
433
600
I
I -10
..../'.2
,
1'4
Figure 1. Effect of H f ion concentration on potential developed b y pyrolytic graphite electrode vs. calomel electrode
was exposed to the solution. Two milliliters of Hg were placed inside the tube to make contact only with the other slightly-conducting face of the graphite. Care must be taken to seal the graphih so that the Hg comes into contact only with the top surface of the graphite. If the Hg can penetrate the epoxy on the side of the cube, the layers of the graphite may be split with a resultant increase in resistance. A Cu wire was immersed in the Hg and connected to the pH meter. The electrode should be carefully examined to make sure that there is no pathway to physical contact of either the mercury or the copper wire with either the potential indicating surface or the solution in the titration cell. An H-type glass cell that was fitted with a fritted disk and an agar plug in the cross-arm was used as a titration vessel for nonaqueous titrations. When the pyrolytic-graphite electrode and a calomel reference electrode were placed in solutions of differing pH, a variation in potential was noted. Figure 1 shows the linear variation of potential with the change in H + concentration. The potential varied 205 mv. in going from 1.ON to 1 X lo-" HC1 solutions. As the OH- concentration increased, the potential became more negative, -15 mv. with l X 10-3,v KaOH to -150 mv. with 1 X lo-" KaOH. Acid-base titrations were performed with the same electrode system as an end point indicator. I n Figure 2 the titration curve of 0.089N NaOH with 0.093N HCl is shown. The curves are reproducible with an abrupt potential shift of 200 mv. at theendpoint. The reverse of this titration, in which 1 ml. VOL 35,
NO. 7, JUNE 1963
929
