848
Anal. Chem. 1985, 57,648-651
(9) Buydens, L.; Massart, D. L. Anal. Chem. 1981, 53, 1990-1993. (10) Buser, H. R.; Rappe, C. Chemosphere 1979, 3 , 157-174. (11) Van Den Dooi, H.; Kratz, P. D. J . Chromatogr. 1963, 7 7 , 463-471. (12) "BAS User's Guide: Statistics, 1982 Edltlon"; SAS Institute Inc.:
Gary, NC, 1982. (13) Albro, P. W.; Haseman, J. K.; Clemmer, T. A.; Corbett, B. J. J . Chromafogr. 1977, 136,147-153. (14) Mosteller, F.; Tukey, J. "Data Analysis and Regression";Addison-Weslev: New York. 1977. (15) Banerjee, A. Acta Crystallogr., Sect. 8 1973, 629, 2070-2074. (16) McKinney, J. D.; Aibero, P. w.; COX, R. H.; Hass, J. R.; Waiters, D. B. In "The Pestlcide Chemist and Modern Toxlcology"; Bandal, S. K.,
Marco, G. J., Goibert, L., Leng, M. L., Eds.; American Chemical SocieWashington, DC, 1981; ACS Symposlum Series No. 160, pp 439-460. (17) Mazer, T.; Hileman, F. D.; Noble, R. W.; Brooks, J. J. Anal. Chem. ty:
1983, 55, 1648. (18) Haken, J. K.; Korhonen, I. 0. 0. J . Chromatogr. 1983, 285, 323-327.
RECEIVED for review July 27, 1984. Resubmitted December 7 , 1984. Accepted December 7, 1984.
Determination of Volatile Organics in Sediment at Nanogram-per-Gram Concentrations by Gas Chromatography T h a k o r A. Amin and Rajinder Singh Narang*
Division of Environmental Sciences, Wadsworth Center for Laboratories and Research. N e w Y o r k State Department of Health, Albany, N e w York 12201
Volatile compounds were stripped from sediment samples and absorbed on cartridges fllled with Porapak N. The compounds were eluted from the cartrldges wlth methanol. The eluate was assayed for varlous organic oompounds by gas chromatography wlth electrontapture and photolonlzation detectlon. A detectlon llmlt of 7 ng/g for each photolonlzatlon-active and 1 ng/g for electron-capturing compound was achleved. Time study experiments showed that, whlle untreated samples should be analyzed wlthln 7 days after collection, samples stored In methanol could be held for up to 90 days without slgnlflcant loss of the volatlle compounds,
Volatile organics in sediment can be analyzed by two general approaches: (i) diluting the sample with water and carrying out the purge-and-trap procedure (1)used for water analysis or (ii) extracting organics from the sediment by vacuum into a trap cooled with liquid nitrogen, adding interference-free water, and carrying out a normal trap-and-purge procedure (2). Invariably sediments, and especially those containing clays, contain significant amounts of humus and other materials that, when purged after dilution with water, tend to foam and make stripping very difficult. Vacuum stripping involves special hardware and multiple steps and is slow to carry out. We describe a method in which closed-loop stripping is combined with steam distillation of the volatile organics onto an absorbent (Porapak N) in a single step. The volatile compounds are eluted from the absorbent with methanol. The eluate is analyzed by gas chromatography with electron-capture (EC) or photoionization (PI) detectors. EXPERIMENTAL S E C T I O N Equipment and Materials. The following equipment is required test tubes, 1.8 X 15 cm, with serum-bottle-typetops; small serum bottles, 5-10 mL; assorted Swagelok fittings; Pyrex tubes, 6 mm X 25 cm; a Metalbellows pump (M-41);Sharon MA); a block heater; Hycar septums, 18 mm, with aluminum caps and crimper; and four pieces of stainless-steel tubing, one 0.16 X 45 cm, one 0.16 X 15 cm with one end sharpened into a needle (needle A), one 0.16 X 45 cm with both ends sharpened (needle B), and one 0.16 X 60 cm with one end sharpened (needle C). Interference-free methanol and Porapak N (80-100 mesh) cartridges were prepared as previously described ( 3 , 4 ) . Assembly of Apparatus. The purging apparatus is assembled as shown in Figure 1. Whenever necessary Swagelok fittings are used to make connections. One end of the absorbent tube is attached to the outlet of the pump, while the other is attached
Table I. Analytes chloroform l,l,I-trichloroethane trichloroethylene carbon tetrachloride bromodichloromethane 1,1,24richloroethane tetrachloroethylene 1,2-dibromoethane
bromoform 1,1,2,2-tetrachloroethane benzene toluene chlorobenzene o,p-xylene o,p-chlorotoluene fluorobenzene"
Used as an internal standard. to the blunt end of needle A. The needle end of A is pushed through the septum of the small vial so that the end is near the bottom of the septum. One end of needle B is pierced through the septum of the small vial so that the end touches the bottom of the vial. The other end of B is pushed through the septum of the sample tube so that the end is about 2 cm below the septum. The needle end of needle C is pierced through the septum of the sample tube so that the end is as near the sample as possible. The blunt end of C is attached to the inlet of the pump. Standards. Standard solutions of each analyte (Table I) were prepared in interference-free methanol at 1000 pg/mL. Appropriate mixed standards and dilutions were then made as required. Spiking of Sediments. Injection (Method SS-1). Dry sediment (clay, 5 g) was placed in a serum tube, which was then sealed with an aluminum cap lined with a Hycar septum. The tube was inverted, and 10-20 p L of methanol containing the appropriate volumes of analytes was injected into the body of the clay. The tube was shaken manually for 10 min. Addition to Spiked Water (Method SS-2). Cold interference-free water (1.5 mL) was poured into a serum tube and spiked with 10-20 p L of methanol containing the appropriate volumes of analytes. Clay (5 g) was added, and the tube was sealed immediately with an aluminum cap lined with a Hycar septum. Purging and Elution. Without Additional Water (Method PE-I). The sample was spiked with 1 pg of fluorobenzene, and the tube was placed in the heating block and kept at 120 O C . The pump was started, and after 30 min of purging the pump was stopped. The sampling cartridge was removed and placed on a stand so that the packed end was near the 1-mL mark of a 5-mL graduated tube. It was then eluted with methanol, and the volume of the eluate was recorded. With Additional Water (Method PE-2). The sample was placed in the test tube and sealed with an aluminum cap lined with a Hycar septum for processing in the apparatus shown in Figure 1. Interference-freewater (8-10 mL) was injected through the septum via a 10-mL glass syringe equipped with a fine needle. The tube was tapped a few times to make a slurry. It was then spiked with 1pg of fluorobenzene. The pump was started, and
0003-2700/85/0357-0648$0 1.50/0 0 1985 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 57, NO. 3, MARCH 1985
9
/P
.-
649
Table 11. Recoveries (%) after Various Methods of Spiking and P u r g i n g
-\i,,
1/16" STAINLESS
F - A S T E E L TUBING WITH
compound
methods (spiking/purging) SS-I/PE-l SS-l/PE-2 SS-2/PE-1
81 Jr 4 chloroform 84 k 6 l,l,l-trichloroethane 80 zI=1 2 carbon tetrachloride 83 Jr 4 trichloroethylene 86 k 7 bromodichloromethane 77 Jr 3 1,1,2-trichloroethane 91 Jr 2 tetrachloroethylene 84 k 1 2 dibromochloromethane 71 k 2 bromoform 73 Jr 3 l,2-dibromoethane 1,1,2,2-tetrachloroethane 60 k 5 111 h 7 fluorobenzeneb
96 f 4
101 f 3 102 f 3 84 f 6 87 f 5 91 f 4 88 f 4 83 f 7 73 f 5 81 f 6 82 f 6 96 f 5
a Recovery values are mean f standard deviation. standard. 'ND Not done.
Flgure 1. Assembled stripping apparatus.
the purging was carried out for 30 min at room temperature. The cartridge was eluted as in Method PE-1. Time Study. Using Aqueous Spiking Solution (Method TS-1). Sediment (clay, 5 g) was spiked by Method SS-2 with selected analytes from Table I at 0.2 pg/g. The sample tubes were tben frozen at -5 "C till ready for analysis. Method PE-1 was used for purging. Using Excessive Methanolic Spiking Solution (Method TS-2). Sediment (clay, 5 g) was spiked by Method SS-1 with selected analytes from Table I at 0.2 pg/g in 1mL of methanol. The spiked samples were frozen at -5 "C till ready for analysis. Method PE-2 was used for purging. Transfer of Samples. Large- Volume Transfer (Method TR-1). Chilled clay (25 g) was added to 7.5 mL of water spiked with 1 pg/mL of various analytes and kept over ice. After the clay was mixed for 15-30 s with a spatula, 6.5-g aliquots were weighed out immediately in sampling tubes and sealed with aluminum caps lined with Hycar septums. These were purged by Method PE-1. Individual-Aliquot Transfer (Method TR-2). Aliquots of sediment (clay, 5 g) were spiked by Method SS-2, quantitatively transferred into other sample tubes, and purged by Method PE-1. Gas Chromatography. A PI detector equipped with a 10.2-eV lamp and electrometer (Method P-101, HNU System4 was mounted on a Hewlett-Packard 5840 gas chromatograph. The signal from the electrometer was fed into the build-in data System of the chromatograph for recording and integrating the chromatographic peaks. The PI active compounds were separated on a 2.7 m X 2 mm column packed with Supelcoport (100-120 mesh) coated with 4% OV-11 and 6% SP2100. The oven temperature was held at 40 OC for 5 min and then programmed to 70 "C at a rate of 3 "C/min. Other general conditions of analysis were as follows: N2 carrier gas, 30 mL/min; detector, 255 "C; injection port, 190 O C . The eleetron-capturing compounds were separated on a 3.6 m X 2 mm column packed with AWDMCS Chromosorb W (100-120 mesh) coated with 15% SF-96 and 6% OV-225. The oven temperature was either (i) held at 60 "C for 10 min, programmed to 80 OC at a rate of 3 "C/min and held at the final temperature for 10 rnin or (ii) held constant at 65-70 "C. Other conditions of analysis were as follows: detector, 300 "C, injection port, 190 "C, carrier gas (Ar/Me 955) flow, 27 mL/min. Both PI-active and electron-capturing compounds were also separated using a fused silica (60 m X 0.25 mm) DB-5 capillary column with 1 pm coating (purchased from J & W Scientific, Rancho Cardova, CA). A splitless mode of injection was used. He was used as the carrier gas with a linear velocity of 60 cm/s. Other conditions of analysis were as follows: for PI-active compounds, injector 200 "C, detector 220 "C, makeup gas He (8 mL/min); oven temperature programmed from 50 to 170 "C at rate of 4 "C/min with a hold of 15 min at initial temperature and 14 rnin at final temperature; for electron-capturing compounds,
78 f 2 70 f 3 70 f 1 70 f 1 78f 1 84 f 2 69 f 1 79 f 1 79 f 1
ND' ND' 85 f 1
Internal
injector 200 "C, detector 300 "C, makeup gas Ar/Me (955; 45 mL/min), oven temperature programmed from 50 to 260 "C at a rate of 8 OC/min with a hold of 4 rnin at initial temperature and 10 min at final temperature.
RESULTS AND DISCUSSION Spiking and Purging. The sediment can be spiked either by placing it in a closed container and then injecting the required amount of various analytes in methanol (Method SS-1)or by adding it to water spiked with the required analytes (Method $S-2). The sediment can then be purged without (Method PE-1) or with (Method PE-2) additional water. The recoveries of various analytes after various combinations of these methods showed no significant differences (Table 11). Practical limitations, however, make some methods better than others. T o achieve a good spike it is important that the sediment be coated uniformly with the various analytes. When the spiking is done with a few microliters of methanol or any other solvent, this uniform distribution is difficult to achieve. Water, on the other hand, is invariably present in significant amounts in any sediment sample. A more uniform spike can be achieved by using enough aqueous solution of various analytes to wet the whole sediment. We observed that 1.5 mL of water was required to wet completely 5 g of sediment (clay), and this method of spiking was used for all of the following studies. Organics can be purged with or without addition of water. When water is added for purging, it is possible to strip the organics at room temperature. Such purging, however, usually causes samples to foam: this is difficult to control, and the emulsions formed are carried through the system, invariably plugging it. In contrast, when a small amount of water (1.5 mL) has been used for spiking or when the sample is analyzed as received, purging must then be carried out a t a higher temperature (120 "C-but no emulsions are formed; there is no need to prepare interference-free water; and at the end of the run the exact weight of the dry sediment can be calculated. This method was used in the following studies. T o determine the effect of storage of sediment samples on the recovery of volatile compounds, clay samples (5 g) were spiked with each analyte at 0.2 pg/g by Method PE-2. These were stored a t -5 "C and analyzed over a span of 1-60 days. The results (Table 111) show that samples can be stored up to 7 days without significant loss of any volatile. Between 14 and 60 days of storage, losses up to 50% occur. However, when samples were stored with small amounts,of methanol (1mL), losses were negligible over a period of 90 days (Table IV). Transfer of samples between containers entailed a considerable loss of volatile compounds regardless of whether the sample was aliquoted from a large sample or was merely
650
ANALYTICAL CHEMISTRY, VOL. 57, NO. 3, MARCH 1985
Table 111. Time Study of Recovery (%) after Spiking by Method 89-2 and Purging by PE-1" % recovery
compound
day 1
day 2
day 3
day4
day 7
day 14
day 28
day35
day 50
day 60
chloroform l,l,l-tiichloroethane carbon tetrachloride trichloroethylene bromodichloromethane 1,1,2-trichloroethane tetrachloroethylene dibromochloromethane bromoform benzene toluene chlorobenzene p-xylene o-xylene o-chlorotoluene p-chlorotoluene fluorobenzeneb
78f2 70 i 3 70 f 1 70fl 78 f 1 84 f 2 69fl 79 f 1 7911 72f5 7511 68f3 7011 68i2 65i1 69f3 85f1
75f10 70 f 11 66 f 12 67f6 78 f 3 85 f 3 69f3 86 f 1 79f3 73f8 76f3 70f2 72f2 72f2 70f4 72f4 92f4
69f5 67 f 6 62 f 11 63f4 72 f 4 77 f 4 71f16 71 f 3 72f2 71f4 69f5 64f2 65f2 66f3 63f2 66f1 87f3
77f4 71 f 1 68 f 2 63f3 74 f 6 81 f 4 63f3 74 f 5 70f2 74i3 66f3 64f5 64i4 65f4 63f4 67f6 87f4
75f3 75 f 4 71 f 4 68f1 72 f 2 77 f 2 65f5 72 f 1 65f2 71f6 68f4 63f2 66f2 64f2 61f2 62fl 89f4
49f5 42 f 15 46 f 7 41f7 48 f 8 52 f 9 42f7 47 f 9 45f5 49f8 45f7 42f8 42f8 41f8 40f6 44i10 85f4
25f3 42 f 3 46 f 2 35f3 43 f 2 65 f 14 39f1 47 f 1 42f3 40f4 37f1 34f2 33f2 34f3 32f4 36f3 80fl
49f3 37 f 9 46 f 5 38f4 47 f 5 56 f 2 42f3 48 f 4 4613 46f5 41f5 38f4 38i3 37f3 42f4 4454 86f2
46f14 46 121 44 i 13 37i12 43 f 11 46 i 6 38f8 43 f 8 39i6 41113 37i6 3117 32i6 32i6 34i9 37i11 9Oi4
40f4 36 f 5 40 f 7 31f4 40 f 4 43 f 16 37f11 34 f 8 41f5 28f2 33f5 29f4 33f9 32f7 29f4 32f2 80f8
'Recovery values are mean f standard deviation. ( n = 3). *Internal standard. Table IV. Time Study % Recovery of Various Organics from Spiked Clay" % recovery
a All
compound
day 1
day 7
day 14
day 30
day 90
carbon tetrachloride l,2-dibromoethane 1,1,2,2-tetrachloroethane benzene toluene chlorobenzene p-xylene o-xylene o-chlorotoluene p-chlorotoluene fluorobenzene
94 f 4 87 f 6 78 f 8 99 f 2 95 f 9 76 f 6 70 f 5 74 f 5 57 f 8 57 f 2 88 f 9
76 f 8 76 f 6 77 f 10 86 f 10 95 f 8 75 f 8 79 f 7 69 f 6 58 f 3 66 f 2 80 f 11
83 f 6 75 f 5 76 f 3 88 f 9 96 f 6 86 f 1 85 f 1 82 f 10 61 f 8 61 f 8 86 f 4
103 f 8 84 Et 3 72 i 5 96 f 14 111 f 7 86 f 3 88 f 3 78 f 2 62 f 5 59 f 7 86 f 7
93 f 7 71 f 2 81 f 6 84 i 7 86 f 7 72 f 6 70 f 6 59 f 7 40 i 2 28 f 1 95 f 3
analvsis were done in tridicate. Method "PE-2" of purging was used. Table VI. Recoverya (%) in Spiked Clayb
Table V. Losses on Transfer by Two Methods"
compound
method compound benzene toluene chlorobenzene o-xylene p-xylene o-chlorotoluene p-chlorotoluene chloroform l,l,l-trichloroethane carbon tetrachloride trichloroethylene bromodichloromethane 1,1,2-trichloroethane tetrachloroethylene l,2-dibromoethane 1,1,2,2-tetrachloroethane
bromoform fluorobenzene
TR-1
TR-2
11 f 1 13 k 2 18 f 3
37 f 4 36 f 3 34 f 5 31 f 5 37 f 5 35 f 7 35 f 7 33 f 6 22 f 5 14 f 4 34 f 6 36 f 2 45 f 3 29 f 4 45 f 5 42 f 6 53 f 5 88 f 6
20 f 2 21 f 6 22 f 2 24 f 3 13 f 2 8 f l 7f1 15 f 1 17 f 3 30 i 4 20 f 2 29 f 3 36 f 5 33 4 80 f 4
*
'Recoverv values are mean f standard deviation (n = 3). transferred from one container to another (Table V). We have shown that volatiles in sediments can be analyzed by closed-loop stripping with Porapak N as adsorbent. This method does not require any special equipment and is easy to carry out. By use of the criteria of signal to noise ratio of 5:l for a 15-g sample, a detection limit of 7 ng/g for PI-active and 1 ng/g for most electron-capturing compounds can be
chloroform l,l,l-trichloroethane carbon tetrachloride trichloroethylene bromodichloromethane 1,1,2-trichloroethane tetrachloroethylene dibromochloromethane bromoform benzene to1uene chlorobenzene p-xylene o-xylene o-chlorotoluene p-chlorotoluene 1,1,2,2-tetrachloroethane
fluorobenzene
recovery, %
recovery,C %
98 f 3 86 f 3 73 f 4 102 f 4 94 f 6 97 f 7 87 f 5 92 f 6 80 f 1 122 f 4 102 f 4 84 f 1 77 f 2 79 f 1 83 f 3 80 f 3 87 f 0 90 f 6
114 f 7 107 f 12 89 f 12 94 f 17 90 f 60 ND 136 f 25 74 f 25 69 f 4 NDd ND ND ND ND ND ND 52 f 8 83 f 8
Recovery values are mean f standard deviation. 15 g of clay spiked with 0.1 pg of each analyte in 4.5 mL of water using method SS-2 of spiking and PE-1 of purging. "Spiking and purging were done as above but with 0.015 pg of each analyte. dND = not detected. achieved (Table VI). T o achieve analytical accuracy the methanol eluate can be subjected to multiple analyses on various columns or detectors. To keep handling of the samples to a minimum, the same container should be used for sampling and purging. If a sample must be stored for an extended time,
Anal. Chem. 1985, 57, 651-658
methanol should be added to keep the loss of volatiles to a minimum. Registry No. Chloroform, 67-66-3; l,l,l-trichloroethane,7155-6; trichloroethylene, 79-01-6; carbon tetrachloride, 56-23-5; bromodichloromethane, 75-27-4; 1,1,2-trichloroethane, 79-00-5; tetrachloroethylene, 127-18-4; 1,2-dibromoethane, 106-93-4; 79-34-5; benzene, bromoform, 75-25-2; 1,1,2,2-tetrachloroethane, 71-43-2; toluene, 108-88-3; chlorobenzene, 108-90-7;p-xylene, 106-42-3;o-xylene, 95-47-13;p-chlorotoluene, 106-43-4;o-chlorotoluene, 95-49-8.
65 1
LITERATURE CITED (1) May, W. F.; Chester, S. N.; Craln, S. P.; Gump, B. H.; Hertz, H. S.; Enagonio, D. P.; Dyszel, S. M. J . Chromatogr. Sci. 1975, 13, 535-540. (2) Hiat, M. H. Anal. Chem. 1981, 53, 1541-1543. (3) Van Tassel, S.; Amalfatino, N.; Narang, R. S. Anal. Chem. 1981, 53, 2 130-2 135. (4) Bush, B.; Narang, R. S. Anal. Chem. 1980, 52, 2076-2079.
RECEIVED for review July 16, 1984. Accepted November 9, 1984.
Synthesis and Properties of High-Temperature Mesomorphic Polysiloxane Solvents: Biphenyl- and Terphenyl-Based Nematic Systems M. A. Apfel, € Finkelmann,’ I. G . M. Janini? R. J. Laub,* B.-H. Liihmann,’ A. Price, W. L. Roberts, T. J. Shaw, and C. A. Smith Department of Chemistry, S u n Diego State University, S u n Diego, California 92182
The synthesis and characterization of a variety of mesomorphic (Iiquid-crystalline) sidechain polysiloxane (MEPSIL) solvents, said to be useful as gas chromatographic statlonary phases, are descrlbed and discussed. The synthetlc scheme Is based upon the hydrosilatlon reaction that occurs when 4-(aliyioxy)benzoyl esters are contacted with poiy(methy1siloxane) in the presence of a dicyciopentadienyiplatinum( I I ) catalyst, while product characterlzatlonis carried out by IR, NMR, GC, DSC, elemental analysis, and direct-observatlon hot-stage iight-polarized microscopy. Selectivity of the MEPSIL phases is shown to differ very substantially from those exhibited by all other common GC solvents. The MEPSIL found overall to be the best suited as a GC statlonary n) and a neliquid exhlbits a meltlng point of 139 OC (k maticAsotropic (n -,I) transitlon temperature of 319 OC, while the practical operating limlts of the material span 150 to in excess of 300 OC.
-
The development and optimization of chromatographic techniques for maximal resolution in minimal time of analysis, coupled with maximum sample throughput, continue to represent a substantial problem when dealing with analytes comprised of structurally related isomers. Among the most notably difficult of these are polycyclic aromatic hydrocarbons (PAH) which are, at the very least, ubiquitous environmental pollutants produced by the incomplete combustion of organic matter. Moreover, some of the most efficient carcinogens known at present are members of, or products derived directly from, this class of compounds. In addition, the carcinogenic and mutagenic properties of these materials are frequently found to be isomer specific. Major efforts have therefore been directed in chromatography at the separation of these species (1). ‘Present address: Institut fur Physikalische Chemie der Technischen Universitat Clausthal, D-3392 Clausthal-Zellerfeld,West Germany.
Present address: Department of Chemistry, University of Kuwait, Kuwait, Kuwait. 0003-2700/85/0357-0651$01.50/0
Reverse-phase packings and selective detection in column-liquid (2) and supercritical fluid (3) chromatographic techniques have been utilized for the separation of PAH isomers in a number of cases; however, the present state of technology in each of these methods precludes the high column efficiency necessary for analysis of complex mixtures comprised of large numbers of components. In gas-liquid chromatography (GLC), studies of the separation of PAH have been concerned mainly with the fabrication of open-tubular columns of ever-higher efficiency with “slightly polar” phases such as SE-52 and SE-54. And, while some degree of success has certainly attended these efforts ( 4 , 5 ) ,it is nonetheless fair to say that the resolution of isomers such as anthracene/phenanthrene, benz[a]anthracene/triphenylene/ chrysene, and benzo[ a]pyrene/ benzo [e ]pyrene/ perylene/benzofluorenes is not entirely satisfactory, particularly when quantitation of these materials is attempted at parts-per-billion levels in natural or industrial matrices. In contrast, it has been recognized for some time that liquid-crystalline (mesomorphic) stationary phases (the nematic state in particular) provide enhanced GLC separations of many isomeric species, including PAH, that are difficult or impossible to resolve by other traditiona1 chromatographic techniques (1,6-13). Unlike most other common GLC solvents that provide separations based upon solute vapor pressure and/or differential solubility arising from specific energetic interactions, nematic stationary phases yield separations based upon differences in solute molecular shape or, in the instance of cholesteric phases, solute chirality. Simply put, rodlike mesomorphs exist in the ,bulk-liquid state in well-ordered domains such that the more rodlike the solute isomer, the more readily it is accommodated into the nematic matrix and, hence, the longer it is retained. Thus, mesomorphic solvents provide selectivity on an entirely different basis from all other GLC stationary phases. However, the liquid-crystal phases reported to date suffer from column bleed as a result of volatility a t high temperatures, poor column efficiency as a result of slow kinetics of mass transfer, and restricted useful nematic ranges. Even so, Laub and his co-workers (14-18)have argued that a suitable member representative of this class of materials must be 0 1985 American Chemical Society