Anal. Chem. 1998, 70, 4094-4099
All-Hydrocarbon Liquid Crystalline Polysiloxane Polymer as Stationary Phase in Gas Chromatography Capillary Column for Separation of Isomeric Compounds of Polynuclear Aromatic Hydrocarbons Wei-Shan Lee and Guo-Ping Chang-Chien*
Department of Chemical Engineering, Cheng-Shiu College of Technology and Commerce, Kaohsiung, Taiwan 833, R.O.C
Two new and high-purity all-hydrocarbon side-chain liquid crystalline polysiloxane polymers were synthesized by grafting all-hydrocarbon liquid crystal monomers onto a polymethylhydrosiloxane backbone. The two polysiloxane polymers show both smectic B and E mesophases which were characterized by (differential scanning calorimetry and X-ray analysis. As stationary phases, these liquid crystalline polysiloxane polymers were coated on the inner surface of a capillary column (i.d. ) 0.32 mm, film thickness df ≈ 0.25 µm) using the static coating method. The capillary column was installed on a GC/MS instrument. We used a standard commercial mixture of 21 species of polynuclear aromatic hydrocarbons (purchased from Supelco Co. and Merck Co.) to test the chromatographic behavior of the coated stationary phase. Test results of the isomeric pair compounds show a better separation resolution than identical tests using the commercial HP-5 capillary column, which is a standard and state-of-the-art analytical tool for the chromatographic resolution of PAHs. The applications of side-chain liquid crystalline polymers (SCLCPs) have received much attention recently.1-5 Applications such as use in stationary phases for high-resolution gas chromatography have been reported.6-12 A significant advancement in this respect was achieved by coupling mesomorphic monomers (1) Blackwood, K. M. Science 1996, 273, 909-912. (2) Andersson, H.; Sahlen, F.; Trollsas, M.; Gedde, U. W.; Hult, A. J. Macromol. Sci.-Pure Appl. Chem. 1996, A3310, 1427-1436. (3) Nakamura, K.; Kikuchi, H.; Kajiyama, T. Polym. J. 1994, 26, 1090-1092. (4) Hwang, J.C.; Kikuchi, H.; Kajiyama, T. Polym. J. 1995, 27, 292-299. (5) Hwang, J. C.; Fuwa, Y.; Moritake, H.; Gu, H.; Ozaki, M.; Yoshino, K. Jpn. J. Appl. Phys., Part 2: Lett. 1995, 34, L560-L562. (6) Jones, B. A.; Bradshaw, J. S.; Nishioka, M.; Lee, M. L. J. Org. Chem. 1984, 49, 4947-4951. (7) Markides, K. E.; Nishioka, M.; Tarbet, B. J.; Bradshaw, J. S.; Lee, M. L. Anal. Chem. 1985, 57, 1296-1299. (8) Gorczynska, K.; Kreczmer, T.; Ciecierskastoklosa, D.; Utnik, A. J. Chromatogr. 1990, 509, 53-57. (9) Fu, R.; Jing, P.; Gu, J.; Huang, Z.; Chen, Y. Anal. Chem. 1993, 65, 21312144. (10) Mazur, J.; Witkiewicz, Z.; Dabrowski, R. J. Chromatogr. 1992, 600, 123127. (11) Sojak, L.; Ostrovsky, I.; Kubinec, R.; Kraus, G.; Kraus, A. J. Chromatogr. 1992, 609, 283-288. (12) Betts, J. T. J. Chromatogr. 1991, 588, 231-237.
4094 Analytical Chemistry, Vol. 70, No. 19, October 1, 1998
to a stable polysiloxane polymer. It has generally been thought that nematic phases give improved resolution over smectic phases in gas chromatography due to greater diffusion in the former and, thus, higher efficiency. However, improved selectivity has been achieved by using smectic phases.6 Polynuclear aromatic hydrocarbons (PAHs) are a group of chemicals with two or more fused benzene rings. The U.S. Environmental Protection Agency lists PAHs as priority pollutants, since some are carcinogens or mutagens, and thus analysis of PAHs in environmental samples has become an important topic. Current analytical techniques for PAHs utilize chromatographic methods (eg., HPLC or GC/MS). The capillary columns installed in HPLC and GC/MS instruments and recommended by the capillary column maker are the HP-5 column (cross-linked 5% phenyl methyl silicone, Hewlett-Packard), the PTE-5 QTM column (Supelco Co.), and the DB-5.625 and DB-5 columns (J & W Scientific Co., Ltd.). Some PAHs, such as families 5-6, 1011, 12-13, 14-15-16, and 18-19 (Figure 3, below), are isomeric compounds and cannot be thoroughly resolved by these commercial GC/MS capillary columns. Some of these isomeric compounds display different electropholic properties13 and thus have different impacts on the human body despite their similar molecular structures. Accordingly, accurate and high-resolution identification of PAHs in real atmospheric samples has become crucial. In this study, two new liquid crystalline polysiloxane polymers, PS4DBT2 and PS4DBT3, are synthesized.14 The side-chain molecule is characteristic of an all-hydrocarbon structure which does not contain a reactive group (eg. -COO-, -CO-NH2-, -CO-). With these polymers as stationary phases in capillary columns, the possibility of reaction between the stationary phase and the compounds injected is slim. We have tested these polymers as stationary phases in a capillary column during GC/ MS chromatography of PAHs and compared the selectivity of the SCLCP stationary phase columns with a commercial capillary column (HP-5). As far as we know, this is the first example of side-chain liquid crystalline polysiloxane polymers containing nonpolar and nonreactive side-chain groups. (13) Ebert, L. B. Polynuclear Aromatic Compounds; American Chemical Society: Washington, DC, 1988. (14) Chang-Chien, G. P. J. Polym. Sci. Part A: Polym. Chem. In press. S0003-2700(98)00276-5 CCC: $15.00
© 1998 American Chemical Society Published on Web 09/03/1998
Table 1. Thermal Transition Temperature (°C), Mesophase Type, and Thermodynamics Data for the Polysiloxane Polymers PS4DBT2 and PS4DBT3 polymer
heating scans and cooling scans
PS4DBT2
SE 119.0 °C(1.7/4.2) SBa 133.7 °C(6.7/16.5) I Ib 123.9 °C(6.1/15.6) SB 117 °C(2.3/5.3) SE SE 124.5 °C(0.5/1.3) SB 143.9 °C(8.2/19.6) I I 133.9 °C(7.9/19.7) SB 128 °C(0.9/1.6) SE a
PS4DBT3
a S , smectic B phase; S , smectic E phase; I, isotropic phase. B E Thermodynamic data (∆H/∆S) are shown in parentheses as (kJ/ mol repeat unit and J/mol repeat unit‚K. b
Table 2. X-ray Diffraction Data for the Liquid Crystalline Polysiloxane Polymers PS4DBT2 and PS4DBT3 polymers PS4DBT2 PS4DBT3
EXPERIMENTAL SECTION Materials. The stationary phase materials used in this study are polysiloxane polymers PS4DBT2 and PS4DBT3. Their syntheses will be pubilished.14 The mesophase type and transition temperatures are listed in Table 1. These two polysiloxane polymers have average molecular weights of 10 500 and 11 000, respectively, with a polydispersity of 2.8 as measured by gel permeation chromatography (GPC). An un-deactivated fused silica capillary column (i.d. ) 320 mm) was obtained from HewlettPackard. Technique. The properties of the liquid crystalline polysiloxane polymers were determined by using a differential scanning calorimeter (Perkin-Elmer DSC-7), a polarized optical microscope (Olympus BH-2 with a LINKAM THMS 600 hot stage), and an X-ray diffractometer (Siemens D5000 apparatus).14 A laboratorymade device was used to prepare the capillary column. Capillary Column Preparation. Preparation procedures for the two capillary columns coated with polymers PS4DBT2 and PS4DBT3 are similar. Taking PS4DBT2 as an example, a fused silica capillary tube (not deactivated) with a 0.32-mm i.d. (HewlettPackard, Avondale, PA) was used. The capillary column was washed with 20 cm3 of methylene chloride before coating. The stationary phase, 20 mg of polysiloxane polymer PS4DBT2, was dissolved in 10 cm3 of methylene chloride, which was degassed before use. The coating solution was filtered before use by a syringe filter (PTFE, pore size 0.2 µm). The filtered solution was placed in a screw-cap septum bottle and was forced through the capillary by nitrogen gas. After filling, the column was sealed at one end with a microflame gas torch. The column was then subjected to a vacuum and the solvent evaporated, completing static coating. The sealed end of the column was then cut off, and cross-linking was performed by injection of vapor-phase azotert-butane (Lancaster, England) with a nitrogen carrier at room temperature for 30 min and a flow rate of 10 cm3/min. Both ends of the column were then sealed, and the column was heated from 50 to 140 °C at 20 °C/min and maintained at 140 °C for 1 h. Both ends of the column were reopened, and the column was cleaned with 10 cm3 of methylene dichloride with a nitrogen carrier and then dried with pure nitrogen. Finally, the column was installed
temp (°C)
spacingsb (Å)
120 115 125 30
22.3(s), 10.6(vs), 7.0(s), 4.7(vs), 4.3(s)a 22.3(s), 10.6(vs), 7.0(s), 4.7(vs), 4.3(s),a 3.4(w) 22.1(s), 10.9(vs), 7.3(s), 4.6(vs), 4.4(s)a 22.1(s), 10.8(vs), 7.3(s), 4.6(vs), 4.4(s),a 3.3(w)
a A right shoulder reflection peak beside the peak of 4.6 Å. b The spacings are given in angstroms, and the letter in parentheses after the value represents the intensity. Abbreviations: vs, very strong; s, strong; and w, weak.
on a gas chromatography apparatus and conditioned at 140 °C for 12 h by a continuous nitrogen stream at 1.2 cm3/min. The column prepared for chromatogram tests was 30 m long, with an internal diameter of 0.32 mm. The gum film of the coated SCLCP was 0.25 µm thick. Column Evaluation. The solute standards were obtained commercially from Supelco Co. (610 M Pash standard mixture). A Hewlett-Packard model 5890 series II gas chromatograph equipped with a 5972 series mass-selective detector and a 6890 series injector with an autosampling controller were used for column evaluation. Helium was used as the carrier gas. RESULTS AND DISCUSSION Polymers PS4DBT2 and PS4DBT3 have similar molecular structures, except for the side-chain terminal group being either an ethyl or a propyl group, respectively. The thermal behavior of polymers PS3DBT2 and PS4DBT3 was detailed in a previous study.14 Table 1 shows the thermodynamic data of PS4DBT2 and PS4DBT3. The melting point and glass transition temperature cannot be observed from the DSC thermograms. The polymers appear as brittle powders at room temperature. They distinctly display two broad and partially overlapped mesophase transition temperatures. However, GPC traces of the polymers show only a monomodal molecular weight distribution, characterized by a polydispersity of about 2.8. In the X-ray diffraction data from Table 2, polymer PS4DBT3 shows three reflection peaks at the layer spacings 22.1, 10.9, and 7.2 Å at 130 °C. The two reflection peaks at 2θ ) 19.3° and 20.4° correspond to intermolecular spacings of 4.6 and 4.4 Å, respectively, which are indicative of the smectic B mesophase. When the temperature is cooled to 125 °C, another diffraction peak appears at 26.9°, which corresponds to an intermolecular spacing of 3.3 Å. The ratio of 4.6/3.3 is apAnalytical Chemistry, Vol. 70, No. 19, October 1, 1998
4095
Table 3. Effect of Mobile Phase Linear Velocities (u) on the Number of Theoretical Plates (N) N u (cm/s)
0.25 µm anthracene
0.25 µm pyrene
0.5 µm anthracene
0.5 µm pyrene
48.0 67.3 84.6 115.0 141.0 164.0 185.0
5075 6674 6530 5374 4849 3978 2683
6243 7877 6780 3943 3800 2921 2077
2055 2945 3204 2883 2645 2254 2018
3023 3446 3509 2493 2212 1852 1746
Figure 2. Relationship of the mobile phase linear velocities and the plates heights (mm/theoretical plate).
Figure 1. Scanning electron micrography of the cross-sectional view of the (a) PS4DBT2, (b) PS4DBT3, and (c) HP-5 columns (×20 000).
proximately the square root of 2, consistent with the transformation of a smectic B (hexagonal) phase to a smectic E (orthorhombic) phase. The thickness of the gum film was initially estimated by the equation of Bouche and Verzele15 as follows:
C (g/cm3) ) 2000F (g/cm3)df (µm)/r (µm) where r is the radius of the fused silica column in this study (r ) 160 µm), df is the desired thickness of the polymer gum on the inner surface of the fused silica column (df ) 0.20 µm), and F is the density of the polymer gum (about 0.8 g/cm3). The given data (r, df, and F) are put in the Bouche and Verzele equation, and the calculated concentration C is 2.0 mg/cm3. We put 10 cm3 of polymer solution with 20 mg of polymer in 10 cm3 of dichloromethane. The thickness of the coating film has been (15) Bouche, J. and Verzele, M. J. Gas Chromatogr. 1968, 6, 501.
4096 Analytical Chemistry, Vol. 70, No. 19, October 1, 1998
further confirmed by the scanning electronic microscopy (SEM) technique. The SEM pictures are shown in Figure 1. The measured thickness obtained from SEM is 0.25 µm, close to the thickness calculated by the equation of Bouche and Verzele. The surface of the coating film of the PS4DBT2 and PS4DBT3 columns appears visually to be rougher than that of the HP-5 column. It is known that an increase in the film thickness generally increases the peak width (reduces column efficiency) and increases the solute’s retention times. It may also increase the resolution value. We have worked on three columns, whose films were 0.25, 0.5, and 1.0 µm thick, respectively. They were all 10 m in length, and their performance was tested on GC by using anthracene and pyrene as solutes. However, increasing the film thickness increases the fastest of the retention times. We do not show the test results for the 1.0-µm column due to too long a retention time, more than 300 min at the condition of 140 °C isothermal and 4 cm3/min He carrier gas flow. Table 3 shows the calculated theoretical plates for the 0.25- µm and 0.50-µm columns. Figure 2 shows the relation of the mobile phase linear velocities (cm/s) to plate heights (mm/theoretical plate). From of Figure 2, the highest efficiency of the column occurs at a linear velocity around 67 cm/s. The retention times of the 0.25-µm column with anthracene and pyrene are 4 and 20 min, respectively, at the condition of 140 °C isothermal and 4 cm3/min He carrier gas flow. However, for the 0.5-µm column, elution takes 54.2 and 71.6 min, respectively. Therefore, the thickness of the coating film profoundly affects the retention time of the solute. Apparently, the diffusion separation mechanisms for the solutes in a liquid crystalline stationary phase should have a greater influence on resolution than other separation factors, such as vapor pressure and polarity.
Table 4. Calculated Resolution Values (Rs) for Anthracene and Pyrene at Various Mobile Phase Linear Velocities (Pyrene as Basis)
at the condition of 140 °C isothermal and various carrier gas flows. The resolution value for anthracene and pyrene is calculated from the following equation:
Rs ) (1/4)(N)1/2((R - 1)/R)(K′/(K′ - 1))
Rs u (cm/s)
0.25 µm
0.5 µm
48.0 67.3 84.6 115.0 141.0 164.0 185.0
6.7 6.7 7.8 7.2 6.6 6.8 6.7
6.2 6.5 7.1 6.7 5.8 6.3 5.6
Table 4 shows the resolution values of the 0.25- µm and 0.50µm columns for the isomeric compounds anthracene and pyrene
where N is the number of theoretical plates, defined as 5.545(tR/ Wh)2, tR is the retention time and the Wh is the width at half-height, R is the separation factor, defined as (tR2 - 1)/(tR1 - 1), and K′ is the retention factor, defined as (tR2 - 1). As can be seen in Table 4, the resolution value of the 0.25-µm column is a little larger than that of the 0.5-µm column with pyrene as the calculated basis. In general, increasing the thickness of the coating film may increase the resolution value for commercial columns. However, in this study, the results apparently differ from the results of commercial columns.
Figure 3. Molecular structures of the 21 PAHs in the standard mixture. Compounds indicated with an asterisk were purchased from E. Merck Co.
Analytical Chemistry, Vol. 70, No. 19, October 1, 1998
4097
Table 5. Values of the Peak Resolution (Rs), Separation Factor (r), Retention Factor (K′), and Theoretical Plates (N) for Isomeric Compounds (a) 5(PA)-6(ANT), (b) 10(BaA)-11(Chr), (c) 12(BbF)-13(BkF), (d) 14(BeP)-15(BaP), (e) 15(BaP)-16(Per), and (f) 18(DbA)-19(BbC) of the Polynuclear Aromatic Hydrocarbons Using the PS4DBT2, PS4DBT3, and HP-5 Columns (a) 5(PA)-6(ANT) PS4DBT2 PS4DBT3 (90 °C) (90 °C) Rs (base ANT) R (base ANT) K′ (base ANT) N (base ANT)
5.57 ( 0.15 1.46 ( 0.02 26.6 ( 0.24 5361 ( 50
5.41 ( 0.14 1.42 ( 0.01 24.8 ( 0.25 5783 ( 39
(b) 10(BaA)-11(Chr) PS4DBT2 PS4DBT3 (140 °C) (140 °C) Rs (base Chr) R (base Chr) K′ (base Chr) N (base Chr)
2.00 ( 0.08 1.082 ( 0.02 72.5 ( 0.14 11 500 ( 60
1.83 ( 0.13 1.076 ( 0.02 69.5 ( 0.27 11 100 ( 52
(c) 12(BbF)-13(BkF) PS4DBT2 PS4DBT3 (180 °C) (180 °C) Rs (base BkF) R (base BkF) K′ (base BkF) N (base BkF)
0.73 ( 0.07 1.022 ( 0.01 26.1 ( 0.20 19 620 ( 70
0.74 ( 0.14 1.023 ( 0.04 24.3 ( 0.30 18 650 ( 49
(d) 14(BeP)-15(BaP) PS4DBT2 PS4DBT3 (170 °C) (170 °C) Rs (base BaP) R (base BaP) K′ (base BaP) N (base BaP) Figure 4. Chromatograms of the 21 PAHs in the standard mixture. (a) PS4DBT2 column, temperature programmed from 50 °C to 80 °C at 10 °C/min and to 260 °C at 3 °C/min; splitless injection. (b) PS4DBT3 column, temperature programmed from 50 °C to 80 °C at 10 °C/min and to 260 °C at 3 °C/min; splitless injection. (c) HP-5 commercial capillary column, temperature programmed from 50 °C to 120 °C at 20 °C/min and to 290 °C at 3 °C/min; splitless injection.
Figure 3 shows the 21 species in the PAH mixtures purchased from Supelco Co. and Merck Co. The compound families 5-6 (molecular weight 178), 10-11 (molecular weight 228), 1213 (molecular weight 252), 14-15-16 (molecular weight 252), and 18-19 (molecular weight 278) are isomers. They were used to test the comparative chromatographic behavior of the PS4DBT2 and PS4DBT3 capillary columns and the HP-5 commercial capillary column (crossed-linked 5% phenyl methyl silicone, HewlettPackard Co.) with programmed temperature. Peaks were identified by comparison with the retention times of peaks generated by pure single-PAH samples. In Figure 4, we find the chromatographic peaks of the isomeric families (compounds 5-6, 1011, 12-13, 14-15-16, and 18-19) overlapping to some degree. Comparing panels a and b of Figure 4 with panel c, we find the PS4DBT2 and PS4DBT3 columns showing better resolution for the above isomeric compounds. It is worth noting that 4098 Analytical Chemistry, Vol. 70, No. 19, October 1, 1998
2.66 ( 0.06 1.089 ( 0.02 55.96 ( 0.24 17 682 ( 56
2.73 ( 0.11 1.086 ( 0.03 54.3 ( 0.24 19 831 ( 50
(e) 15(BaP)-16(Per) PS4DBT2 PS4DBT3 (170 °C) (170 °C) Rs (base Per) R (base Per) K′ (base Per) N (base Per)
2.62 ( 0.04 1.104 ( 0.01 61.8 ( 0.25 12 916 ( 58
2.32 ( 0.12 1.086 ( 0.02 59.3 ( 0.26 14 261 ( 60
(f) 18(DbA)-19(BbC) PS4DBT2 PS4DBT3 (170 °C) (170 °C) Rs (base BbC) R (base BbC) K′ (base BbC) N (base BbC)
2.50 ( 0.04 1.086 ( 0.02 41.2 ( 0.23 16 785 ( 60
2.97 ( 0.14 1.097 ( 0.04 38.7 ( 0.24 18 951 ( 62
HP-5 (130 °C) 2.48 ( 0.05 1.055 ( 0.01 19.88 ( 0.22 40 813 ( 65
HP-5 (160 °C) 1.47 ( 0.07 1.029 ( 0.02 53.8 ( 0.24 45 380 ( 69
HP-5 (180 °C) 0.54 ( 0.06 1.025 ( 0.01 41.2 ( 0.24 18 924 ( 50
HP-5 (200 °C) 0.68 ( 0.06 1.025 ( 0.02 48.2 ( 0.30 17 638 ( 52
HP-5 (200 °C) 0.98 ( 0.05 1.03 ( 0.02 49.5 ( 0.24 18 924 ( 72
HP-5 (220 °C) 1.09 ( 0.04 1.019 ( 0.01 1.56 ( 0.25 52 212 ( 54
optimal chromatographic conditions for the HP-5 column as opposed to the PS4DBT2 and PS4DBT3 columns are different. Chromatography for each column was performed at the optimum conditions for each column. The retention times of the PS4DBT2 and PS4DBT3 columns are a little shorter than that of the HP-5 column, and the PS4DBT2 and PS4DBT3 columns operate at a lower temperature. Table 5 shows the calculated peak resolution (Rs), separation factor (R), retention factor (K′), and theoretical
Figure 5. Relationship between the natural logarithm of separation factor and reciprocal column temperature: (a) anthracene (6) vs phenanthrene (5), (b) chrysene (11) vs benz[a]anthracene (10), (c) benzo[k]fluoranthene (13) vs benzo[b]fluoranthene (12), (d) benzo[a]pyrene (15) vs benzo[e]pyrene (14), (e) perylene (16) vs benzo[a]pyrene (15), and (f) benzo[b]chrysene (19) vs dibenz[a,h]anthracene (18) on PS4DBT2 column, PS4DBT3 column, and HP-5 commercial column.
plates (N) of the isomeric PAH pairs (a) 5-6, (b) 10-11, (c) 12-13, (d) 14-15, (e) 15-16, and (f) 18-19 which result from use of the HP-5, PS4DBT2, and PPS4DBT3 columns. In the inspection of Table 5, we find that the PS4DBT2 and PS4DBT3 capillary columns show higher peak resolution (Rs) and separation factor (R) than the HP-5 column. The theoretical plates (N) of the PS4DBT2 and PS4DBT3 columns are lower than those of the HP-5 column. This may be caused by the lower film-forming ability of the liquid crystalline polysiloxane polymer or the empty capillary column not being deactivated before the coating procedure. Figure 5 shows the relationship of the separation factors to the reciprocal column temperature for (a) anthracene (6) vs phenanthrene (5), (b) chrysene (11) vs benz[a]anthracene (10), (c) benzo[k]fluoranthene (13) vs benzo[b]fluoranthene (12), (d) benzo[a]pyrene (15) vs benzo[e]pyrene (14), (e) perylene (16) vs benzo[a]pyrene (15), and (f) benzo[b]chrysene (19) vs dibenz[a,h]anthracene (18) for the PS4DBT2, PS4DBT3, and HP-5 capillary columns. As observed in Figure 5, the PS4DBT2 and PS4DBT3 columns show better resolution than the commercial capillary column. In addition, the temperature ranges of chromatographic separation ability are all wider than those for the commercial HP-5 capillary column. The new phases give an
interesting and possibly useful orthogonal selectivity and high enough efficiency. They show superior resolution for isomeric families that the HP-5 column cannot resolve and thus may be used in conjunction with the HP-5 for high-resolution analysis of isomeric PAHs. Despite the lack of full and confirmed data to support the premise of nonreaction between the stationary phase and the separated solutes, we have used the PS4DBT2 and PS4DBT3 columns more than 500 times, with temperatures reaching 260 °C, and the laboratory-made capillary columns still show no decomposition and bleeding. In light of these results and the molecular structures, it would seem that the all-hydrocarbon polysiloxane polymers, lacking reactive functional groups, do not react with the tested injected compounds. ACKNOWLEDGMENT This work was kindly supported by the National Science Council of the Republic of China (NSC-87-2221-E-230-001).
Received for review March 10, 1998. Accepted July 15, 1998. AC980276D
Analytical Chemistry, Vol. 70, No. 19, October 1, 1998
4099
