Isotopic Effect in the Separation of Polystyrene by Normal Phase and

Jan 19, 2010 - Phone: +82-54-279-2109. ... It was found that d-PS was retained slightly longer than h-PS in NPLC (bare silica column and tetrahydrofur...
0 downloads 5 Views 672KB Size
Download PDF
Anal. Chem. 2010, 82, 1509–1514

Isotopic Effect in the Separation of Polystyrene by Normal Phase and Reversed Phase Liquid Chromatography Youngtak Kim, Seonyoung Ahn, and Taihyun Chang* Department of Chemistry and Division of Advanced Materials Science, Pohang University of Science and Technology (POSTECH), Pohang, 790-784, Korea The retention behavior of deuterated polystyrene (d-PS) and normal hydrogenous polystyrene (h-PS) was investigated in normal phase (NP) and reversed phase (RP) liquid chromatography (LC). It was found that d-PS was retained slightly longer than h-PS in NPLC (bare silica column and tetrahydrofuran (THF)/n-hexane mixed mobile phase) while in RPLC (C18-bonded silica column and THF/methanol mixed mobile phase) h-PS was retained longer than d-PS. The retention behaviors are interpreted in terms of the thermodynamic variables associated with LC separations. The different retention behavior of deuterated polymers from that of hydrogenous polymers could allow separation of partially deuterated polymers according to the deuterium content, such as for copolymers in LC characterization. Deuterated polymers have been widely used for many decades to study their physical properties.1,2 Physicochemical properties of deuterated molecules similar to those of the normal hydrogenous counterparts make deuterium labeling a versatile technique due to the large abundance of hydrogen atoms in organic compounds, high contrast in relative mass and neutron crosssection, etc. Despite the wide use of deuterium labeling for various physical measurements under the assumption that physicochemical properties between normal polymer and deuterated counterpart are indistinguishable, it is well established that there exists a small but clear difference between the two. For example, deuterium substitution has been reported to affect the theta temperature3 as well as chain conformation in solution,4 critical temperature of the polymeric blend,5 and nonideal mixing of deuterated and hydrogenous polymers.6 The chromatographic differentiation of deuterated molecules from their hydrogenous counterparts was also realized a long time

ago for small molecules.7 In liquid chromatography (LC), separation according to the isotopic content of small isotopic molecules has been carried out mostly by reversed phase (RP) LC systems because better separations can be accomplished in a shorter analysis time. Good RPLC separation of several isotopic compounds has been reported by a number of authors.8-11 Even the separation of racemates based on isotopic chirality has also been reported.12,13 Recently Valleix et al. carried out an extensive study on the isotope effects on the separation of deuterium- and tritiumlabeled compounds from unlabeled ones using RPLC separation.14 They also reported that the use of deuterated solvents improved the resolution but did not change the selectivity. All the results in the literature reported that deuterated compounds are retained for less time during RPLC separation, and the retention difference is generally ascribed to the lower polarizability of the deuteriumcontaining bond relative to the hydrogen counterpart. Although it is well established that different isotopes of a chemical species can be resolved by chromatography, few studies have been performed on the chromatographic separation of deuterium-labeled polymers. It is likely because of the predominant use of size exclusion chromatography for polymer analysis, which is insensitive to the deuterium content. Perny et al. reported that there is a significant isotope effect in the RPLC retention of polybutadiene.15 Like small molecules, deuterated polymers are retained less than their hydrogenous counterparts in RPLC separation. Kayillo et al. reported on the deuterium isotope effect in the RPLC retention of polystyrene (PS) oligomer.16,17 They found the same trend as that observed for polybutadiene: the deuterated PS was retained less. In this study, we examined the retention of deuterated PS (d-PS) relative to hydrogenous PS (h(7) (8) (9) (10) (11) (12)

* To whom correspondence should be addressed. Phone: +82-54-279-2109. Fax: +82-54-279-3399. E-mail: [email protected]. (1) Cotton, J. P.; Nierlich, M.; Boue, F.; Daoud, M.; Farnoux, B.; Jannink, G.; Duplessix, R.; Picot, C. J. Chem. Phys. 1976, 65, 1101–1108. (2) Higgins, J. S.; Benoit, H. C. Polymers and Neutron Scattering; Oxford University Press, Cary, NC, 1997. (3) Strazielle, C.; Benoit, H. Macromolecules 1975, 8, 203–205. (4) Wang, X.; Xu, Z.; Wan, Y.; Huang, T.; Pispas, S.; Mays, J. W.; Wu, C. Macromolecules 1997, 30, 7202–7205. (5) Lin, J. L.; Roe, R. J. Macromolecules 1987, 20, 2168–2173. (6) Bates, F. S.; Wignall, G. D. Macromolecules 1986, 19, 932–934. 10.1021/ac902622t  2010 American Chemical Society Published on Web 01/19/2010

(13) (14) (15) (16) (17)

Filer, C. N. J. Labelled Compd. Radiopharm. 1999, 42, 169–197. Falconer, W. E.; Cvetanovic, R. J. Anal. Chem. 1962, 34, 1064–1066. Tanaka, N.; Thornton, E. R. J. Am. Chem. Soc. 1976, 98, 1617–1619. Tanaka, N.; Thornton, E. R. J. Am. Chem. Soc. 1977, 99, 7300–7307. Turowski, M.; Yamakawa, N.; Meller, J.; Kimata, K.; Ikegami, T.; Hosoya, K.; Tanaka, N.; Thornton, E. R. J. Am. Chem. Soc. 2003, 125, 13836–13849. Kimata, K.; Kobayashi, M.; Hosoya, K.; Araki, T.; Tanaka, N. J. Am. Chem. Soc. 1996, 118, 759–762. Kimata, K.; Hosoya, K.; Araki, T.; Tanaka, N. Anal. Chem. 1997, 69, 2610– 2612. Valleix, A.; Carrat, S.; Caussignac, C.; Leonce, E.; Tchapla, A. J. Chromatogr., A 2006, 1116, 109–126. Perny, S.; Allgaier, J.; Cho, D.; Lee, W.; Chang, T. Macromolecules 2001, 34, 5408–5415. Kayillo, S.; Gray, M. J.; Shalliker, R. A.; Dennis, G. R. J. Chromatogr., A 2005, 1073, 83–86. Kayillo, S.; Shalliker, R. A.; Dennis, G. R. Macromol. Chem. Phys. 2005, 206, 2013–2017.

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1509

PS) in normal phase (NP) LC as well as in RPLC. If the effect of deuterium labeling is different in NPLC and RPLC retention, it would provide a unique tool for the separation of partially deuterated polymers according to deuterium content and total molecular weight, just like in copolymer characterization. For the HPLC analysis of oligomers, the analysis method needs to be reviewed briefly. To describe the chromatographic retention of a solute, a dimensionless retention factor, k, is often used which is defined as follows. k)

VR - Vo Vo

(1)

VR and Vo are the retention volume of solute and the total holdup volume of the column, respectively. A chromatographic separation process is a consecutive equilibrium process of solutes between the stationary and the mobile phases. The equilibrium distribution constant (K) of solutes between the stationary and the mobile phases is related to the standard Gibbs free energy (∆G°) change, enthalpy (∆H°), and entropy change (∆S°) associated with the transfer of the solutes from the mobile phase to the stationary phase. ∆G° ) ∆H° - T∆S° ) -RT ln K

(2)

The retention factor, k, is related to K as follows. k ) Kφ

(3)

where φ is the volume ratio of the stationary phase to mobile phase. Combining equations (eq 2) and (eq 3), the thermodynamic variables can be determined from the temperature-dependent retention factor via the following van’t Hoff equation. ln k ) -

∆S° ∆H° + + ln φ RT R

(4)

From the slope and intercept of the van’t Hoff plot (ln k vs 1/T), the standard enthalpy change (∆H°/R) and the entropy change (∆S°/R + ln φ ) ∆S*/R) can be obtained, respectively. The hold-up volume of a column, Vo, in the calculation of retention factor (eq 1) represents the elution volume of an unretained molecule, and the elution peak of the injection solvent is often taken as a marker to measure Vo. If the elution peak position of the injection solvent was used as Vo, however, the plot of eq 1 does not follow Martin’s rule well (eq 5),18 and it often shows a significant deviation toward low DP.19-21 Martin’s rule is based on the additivity of the interaction Gibbs free energy of a solute molecule, and, for a polymeric solute, it leads to the form of ln k ) A + Bn

(5)

Recently it was suggested that the column hold-up volume can be determined based on Martin’s rule, i.e., not to take the Vo in eq 1 value from an independent measurement such as elution volume of the injection solvent but to take it as an adjustable parameter (Vo*) to make the plot of eq 5 linear for a polymer.22,23 A simple mathematical derivation converts eq 5 to eq 6, which is a form to obtain Vo* easily from the experimental data. Vn ) Vo* + γ(Vn - Vn-1) ) Vo* + γ∆Vn

(6)

where Vn and Vn-1 are the retention volumes of n-mer and (n 1)-mer, respectively. Vo* is an adjustable column hold-up volume. In an earlier study, the retention behavior of low MWPS in RPLC and NPLC was investigated.24 It was found that if Vo* is used as the column hold-up volume, not only the retention factor but also the thermodynamic variables calculated from the temperature dependence on the retention factor show an excellent linear relationship with DP, conforming to Martin’s rule. EXPERIMENTAL SECTION Low-MW perdeuterated PS (d-PS) was purchased from Polymer Source Inc. (Mn ) 850), and normal PS (h-PS) was homemade (Mn ) 890). Both were prepared by anionic polymerization initiated with nondeuterated sec-butyl lithium. The details of the anionic polymerization apparatus and procedure were reported previously.25,26 The MW and the end group of the PS samples were confirmed by MALDI-TOF mass spectrometry (Bruker REFLEX III). For NPLC analysis, a silica column (Nucleosil, 5 µm, 50 Å pore, 250 mm × 4.6 mm) and a mixture (90/10, v/v) of n-hexane and THF (Samchun, HPLC grade) were used as stationary and mobile phases, respectively. For RPLC separation, C18-bonded silica column (Luna C18, Phenomenex, 5 µm, 100 Å pore, 150 mm × 4.6 mm) and a mixture (90/10, v/v) of methanol and THF (Samchun, HPLC grade) were used as stationary and mobile phases, respectively. Injection samples were prepared by dissolving the polymers in the corresponding eluent. The temperature of the column was controlled by circulating fluid from a programmable bath/circulator (Thermohaake, C25P) through a homemade column jacket. The chromatogram was recorded with a UV absorption detector (TSP, UV100) operated at a wavelength of 260 nm. RESULTS AND DISCUSSION Retention Behavior of d-PS and h-PS in NPLC. Figure 1A shows NPLC chromatograms of d-PS and h-PS obtained at a column temperature of 10 °C. All oligomer species are well resolved. The degree of polymerization (DP ) n) for the oligomer peaks was identified by MALDI-TOF MS, and the 10-mer peaks are labeled in the chromatograms. As can be seen in Figure 1A, d-PS is retained slightly longer than h-PS under the NPLC

where n is the degree of polymerization (DP). (18) Martin, A. J. P. Biochem. Soc. Symp. 1950, 3, 4–20. (19) Philipsen, H. J. A.; Claessens, H. A.; Lind, H.; Klumperman, B.; German, A. L. J. Chromatogr., A 1997, 790, 101–116. (20) Kayillo, S.; Dennis, G. R.; Wormell, P.; Shalliker, R. A. J. Chromatogr., A 2002, 967, 173–181. (21) Skvortsov, A.; Trathnigg, B. J. Chromatogr., A 2003, 1015, 31–42.

1510

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

(22) Trathnigg, B.; Jamelnik, O.; Skvortsov, A. J. Chromatogr., A 2006, 1128, 39–44. (23) Trathnigg, B.; Skvortsov, A. J. Chromatogr., A 2006, 1127, 117–125. (24) Kim, Y.; Ahn, S.; Chang, T. Anal. Chem. 2009, 81, 5902–5909. (25) Kwon, K.; Lee, W.; Cho, D.; Chang, T. Korea Polym. J. 1999, 7, 321–324. (26) Lee, W.; Cho, D.; Chang, T.; Hanley, K. J.; Lodge, T. P. Macromolecules 2001, 34, 2353–2358.

Figure 1. (A) NPLC chromatograms of d-PS and h-PS obtained at 10 °C. DP of the oligomers in each peak was identified by MALDITOF MS, and the 10-mer peak was labeled. Column: Nucleosil bare silica, 50 Å pore, 250 mm × 4.6 mm. Eluent: THF/n-hexane ) 10/90 (v/v) at a flow rate of 0.5 mL/min. (B) Plots of ln k vs n for NPLC separation of d-PS (0) and h-PS (9).

separation condition. This retention difference between d-PS and h-PS in the NPLC separation is more clearly contrasted in the ln k vs n plot displayed in Figure 1B. To make the plot, the retention factor (k) was calculated using Vo* determined from eq 6 for Vo in eq 1.22–24 The Vo* values were reported earlier24 and are identical for d-PS and h-PS. As can be seen in Figure 1B for d-PS (0) and h-PS (9), when Vo* is used to calculate k, the ln k vs n plots show a nearly perfect linear relationship conforming to Martin’s rule, i.e., the retention of a polymer or homologue series increases exponentially with the number of repeating units or DP (n).24 The ln k vs n plots clearly show that d-PS is retained slightly longer than h-PS. Figure 2 displays the temperature dependence of the NPLC retention of low-MW d-PS (top) and h-PS (bottom) obtained under the same experimental condition as in Figure 1 except for the column temperature. Both d-PS and h-PS chromatograms show a decreasing trend of retention with increasing column temperature, indicating an exothermic adsorption process of PS oligomers to the bare silica stationary phase. Figure 3 displays the van’t Hoff plots of d-PS (top) and h-PS (bottom) oligomers. All van’t Hoff plots exhibit good linearity over the whole DP range. From the slope and intercept of van’t Hoff

Figure 2. NPLC chromatograms of d-PS (top) and h-PS (bottom) obtained at four different column temperatures. DP of the oligomer peaks was identified by MALDI-TOF MS, and the 10-mer peaks are labeled. Column: Nucleosil bare silica, 50 Å pore, 250 mm × 4.6 mm. Eluent: THF/n-hexane ) 10/90 (v/v) at a flow rate of 0.5 mL/min.

plot, the enthalpy change (∆H°) and the entropy change (∆S*) related to the chromatographic equilibrium process can be obtained according to eq 4, respectively. ∆S* is used in place of ∆S° because the phase ratio of stationary to the mobile phase (φ) cannot be determined unambiguously. One interesting thing to note in these van’t Hoff plots is that the slope of the plots changes the sign at DP ∼ 4. This means that the solute transfer process from the mobile to stationary phase in the NPLC separation changes from an exothermic process to an endothermic process as DP decreases. In Figure 4, the dependence of ∆H° /R and ∆S*/R on DP is plotted for d-PS (0) and h-PS (9). Both thermodynamic parameters exhibit a good linear relationship with DP. Thermodynamic variables of d-PS in the NPLC separation are more negative than h-PS, indicating that the transfer process of d-PS from the mobile to the NP stationary phase is more exothermic than that of h-PS. The thermodynamic parameters change the sign from negative to positive as DP decreases. It reflects the unfavorable, endothermic contribution of the end group in the solute transfer process to the stationary phase.24 The end groups of these samples are the initiator moiety of the anionic polymerization: sec-butyl group Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1511

Figure 3. van’t Hoff plots of d-PS (top) and h-PS (bottom) prepared from the NPLC chromatograms shown in Figure 2.

and the terminal hydrogen. The nonpolar saturated hydrocarbon end group must not be an attractive moiety to the polar silica stationary phase. The plots of d-PS and h-PS converge to a common intercept as expected because both PS samples contain the same end groups. The thermodynamic variables involved in the NPLC retention per repeating unit of d-PS and h-PS are calculated from the slopes of plots in Figure 4 and summarized in Table 1. d-PS shows a ∼5% more negative ∆H°/R value, and it is apparent that the longer retention of d-PS from the mobile to NP stationary phase is due to its more favorable energetic interaction with the polar stationary phase than h-PS. The butyl end group exhibits a highly endothermic contribution. This is not surprising because the saturated alkyl end group should not be attracted to the polar surface of bare silica packing materials. The mobile phase condition was established in such a way for the styryl repeating unit and the butyl end group to exhibit different signs of ∆H° with the bare silica stationary phase. The end group contributions in the d-PS and h-PS retention are the same as expected because they are the sec-butyl group with identical isotopic composition, which supports the validity of our thermodynamic analysis. Retention Behavior of d-PS and h-PS in RPLC. Figure 5 displays RPLC chromatograms of d-PS and h-PS at 10 °C (A) and 1512

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

Figure 4. DP dependence of ∆H°/R (top) and ∆S*/R (bottom) for NPLC separation of d-PS (0) and h-PS (9) obtained from the slope and intercept of the van’t Hoff plots in Figure 3. Table 1. Comparison of Thermodynamic Variables in HPLC Separations of d-PS and h-PS ∆H°/R (K)

NPLC d-PS h-PS RPLC d-PS h-PS

repeating units

end group

-135.4 ± 0.5 -129.0 ± 0.5 -88.0 ± 0.9 -96.3 ± 0.5

464 ± 5 460 ± 4 -185 ± 8 -192 ± 5

∆S*/R repeating units -0.306 -0.290 -0.095 -0.114

± ± ± ±

0.001 0.001 0.001 0.001

end group -0.62 -0.60 -0.35 -0.35

± ± ± ±

0.01 0.01 0.01 0.01

the ln k vs n plot (B). The chromatograms of both PS samples are resolved well into individual oligomers, and Vo* determined from eq 6 is used in the calculation of k. Again the Vo* values of d-PS and h-PS were the same, and it is identical with the value reported earlier.24 Contrary to the NPLC separation, h-PS is retained longer than d-PS. The ln k vs n plots again show excellent linearity conforming to Martin’s rule. Figure 6 shows the temperature dependence of the retention of d-PS and h-PS in the RPLC separation. Similarly to the NPLC result in Figure 2, their retentions decrease as the temperature is raised, indicating that the solute transfer from mobile to RP stationary phase is an exothermic process. In Figure 7, the van’t Hoff plots of d-PS and h-PS with different DP are plotted. The plots for both oligomers show good linearity

Figure 5. (A) RPLC chromatograms of d-PS and h-PS obtained at 10 °C. DP of the oligomer peaks was identified by MALDI-TOF MS, and the 10-mer peak was labeled. Column: Luna C18, 100 Å pore, 150 mm × 4.6 mm. Eluent: THF/methanol ) 10/90 (v/v) at a flow rate of 0.5 mL/min. (B) Plots of ln k vs n for RPLC separation of d-PS (0) and h-PS (9).

over the whole range of DP. Unlike the NPLC separation, the slope of the van’t Hoff plot does not change the sign in the RPLC separation. From the slope and intercept of the plots, thermodynamic variables associated with the separation process are determined. Figure 8 displays the dependence of the thermodynamic parameters on DP for d-PS and h-PS. Contrary to the NPLC separation, h-PS shows more negative thermodynamic parameters than d-PS. The thermodynamic parameters per repeating units and the end group of d-PS and h-PS in the RPLC separation are summarized in Table 1. The ∆H°/R of h-PS is ∼10% more negative than d-PS, which is the main reason of the longer retention of h-PS than d-PS. Kayillo et al. reported earlier on the thermodynamic contrast between the retention of d-PS and h-PS under similar RPLC separation conditions (C18 silica column, THF/methanol ) 7/93).16,17 They also observed longer retention of h-PS than d-PS. However, the plots such as Figure 8 showed a significant curvature toward the lower DP. In the results, the difference of ∆H° between d-PS and h-PS repeating units was not a constant but changes with DP, and the group contribution could not be determined. We found a similar problem in the earlier study if we used the injection solvent peak as the column hold-up volume.24 By virtue

Figure 6. RPLC chromatograms of d-PS (top) and h-PS (bottom) obtained at four different column temperatures. DP of the oligomer peaks was identified by MALDI-TOF MS and the 10-mer peaks are labeled. Column: Luna C18, 100 Å pore, 150 mm × 4.6 mm. Eluent: THF/methanol ) 10/90 (v/v) at a flow rate of 0.5 mL/min.

of using Vo* as the column hold-up volume in this study, both thermodynamic parameters for d-PS and h-PS exhibit an excellent linear relationship with DP like in the case of NPLC in Figure 4. Furthermore, the plots of d-PS and h-PS converge to one point at n ) 0, which is supposed to be the case in a correct analysis because the two samples possess the same end group. The ∆H°/R value of the end group is negative and much larger than the value of the repeating unit. Unlike the bare silica surface in the NPLC separation, the sec-butyl end group should be partitioned in the RPLC stationary phase (octadecyl group) more favorably than the styryl repeating unit. The thermodynamic analysis result obtained in this study strongly supports the utility of Vo* for the column hold-up volume. So far, a number of studies have confirmed that the use of Vo* makes the ln k vs n plot straight, conforming to Martin’s rule.22,23 It was also confirmed that the use of Vo* makes the ∆H° and ∆S* vs n plot straight.24 Nonetheless, these studies do not present convincing evidence for the validity of Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

1513

Figure 8. DP dependence of ∆H°/R (top) and ∆S*/R (bottom) for RPLC separation of d-PS (0) and h-PS (9) obtained from the slope and intercept of the van’t Hoff plots in Figure 7. Figure 7. van’t Hoff plots of d-PS (top) and h-PS (bottom) constructed from the RPLC chromatograms shown in Figure. 6.

using Vo* as the column hold-up volume because the determination of Vo* by eq 6 stems from the assumption that Martin’s rule is valid. In this study, taking advantage of the slightly different retentions of d-PS and h-PS, we were able to confirm that the thermodynamic parameters for the sec-butyl end group are identical for both d-PS and h-PS, which strongly supports the validity of using Vo* as the column hold-up volume. It is often difficult to determine the thermodynamic parameters for the end group unambiguously because of the curvature of the plot. Straightening of the plot using Vo* as the column hold-up volume not only enables the unambiguous determination of the end group contribution but also confirmation of the identical end group contribution for h-PS and d-PS. (27) Lee, H.; Chang, T.; Lee, D.; Shim, M. S.; Ji, H.; Nonidez, W. K.; Mays, J. W. Anal. Chem. 2001, 73, 1726–1732. (28) Park, S.; Cho, D.; Ryu, J.; Kwon, K.; Lee, W.; Chang, T. Macromolecules 2002, 35, 5974–5979. (29) Im, K.; Park, H. W.; Kim, Y.; Chung, B. H.; Ree, M.; Chang, T. Anal. Chem. 2007, 79, 1067–1072. (30) Park, S.; Ryu, D. Y.; Kim, J. K.; Ree, M.; Chang, T. Polymer 2008, 49, 2170–2175. (31) Im, K.; Park, H. W.; Kim, Y.; Chang, T. Macromol. Res. 2008, 16, 544– 548. (32) Park, H. W.; Jung, J.; Chang, T. Macromol. Res. 2009, 17, 365–377.

1514

Analytical Chemistry, Vol. 82, No. 4, February 15, 2010

In summary, we carried out a thermodynamic study to examine the isotopic effect in NPLC and RPLC separation of h-PS and d-PS. The isotope effect is opposite in the two different stationary phases. In NPLC, the chromatographic retention of d-PS is slightly larger than that of h-PS while, in RPLC, h-PS is retained longer than d-PS. Deuterium labeling of specific locations, branches, or blocks of a polymer is a very useful technique in polymer science. The different retention behavior of d-PS and h-PS in NPLC and RPLC separation should allow HPLC separation of partially deuterated polymers such as copolymers27–32 and could allow characterization of partially deuterated polymers more precisely. In addition, the thermodynamic analysis results in this study strongly support the utility of Vo* for the column hold-up volume. ACKNOWLEDGMENT This study was supported by NRF via NRL (R0A-2007-00020125-0), SRC (R11-2008-052-03002) and WCU (R31-2008-00010059-0) programs.

Received for review November 16, 2009. Accepted December 26, 2009. AC902622T