Liquid Chromatography at the Critical Condition for Polyisoprene

Adnan Murad Bhayo , Rubina Abdul-Karim , Syed Ghulam Musharraf , Muhammad Imran Malik ... Muhammad Imran Malik , Sanghoon Lee , Taihyun Chang...
1 downloads 0 Views 84KB Size
Anal. Chem. 2001, 73, 3884-3889

Liquid Chromatography at the Critical Condition for Polyisoprene Using a Single Solvent Wonmok Lee, Soojin Park, and Taihyun Chang*

Department of Chemistry and Center for Integrated Molecular Systems, Pohang University of Science and Technology, Pohang, 790-784, Korea

Liquid chromatography at the chromatographic critical condition has drawn much attention as an attractive characterization method of block copolymers since it has been proposed that a part of a polymer chain becomes “chromatographically invisible” at this condition, which would permit the characterization of individual blocks. A critical condition for a polymer species has been commonly established by use of mixed-solvent systems. It is not easy, however, to reproduce the critical condition since the retention of polymers depends very sensitively on the solvent composition and purity. Furthermore, the preferential sorption of a component in a mixed solvent may cause an additional problem. Therefore, the use of a single solvent is highly desirable to improve the reproducibility as well as the repeatability. In this study, a singlesolvent critical condition for polyisoprene was established with 1,4-dioxane and C18 bonded silica as the mobile and stationary phases, respectively. At this condition, the “chromatographic invisibility” of polystyrene-polyisoprene diblock copolymers was critically examined and it was found that a rigorous chromatographic invisibility was not achieved and the retention of the block copolymers was affected by the length of the blocks under the critical condition. Some other chromatographic applications using the single-solvent system are also reported. Liquid chromatography (LC) is a powerful tool for the fractionation of polydisperse and/or chemically heterogeneous polymers. Among a variety of LC methods, size exclusion chromatography (SEC) has been the most frequently employed for the analysis of molecular weight distribution (MWD) of polymers.1-3 On the other hand, nonexclusion chromatography, which will be called henceforth interaction chromatography (IC), has been also applied to characterize MWD4-14 as well as chemical * Corresponding author: (tel) +82-54-279-2109’ (fax) +82-54-279-3399; (e-mail) [email protected]. (1) Yau, W. W.; Kirkland, J. J.; Bly, D. D. Modern Size-Exclusion Liquid Chromatography, Practice of Gel Permeation and Gel Filtration Chromatograph; John Wiley & Sons: New York, 1979. (2) Snyder, L. R.; Kirkland, J. J. Introduction to Modern Liquid Chromatography, 2nd ed.; Wiley-Interscience: New York, 1979. (3) Balke, S. T. In Modern Methods of Polymer Characterization; Barth, H. G., Mays, J. W., Eds.; John Wiley & Sons: New York, 1991. (4) Armstrong, D. W.; Bui, K. H. Anal. Chem. 1982, 54, 706-708. (5) Snyder, L. R.; Stadalius, M. A.; Quarry, M. A. Anal. Chem. 1983, 55, 1413A1430A.

3884 Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

heterogeneity15,16 of various polymers. These two different modes in LC separation of polymers can be distinguished by the thermodynamics involved in the chromatographic process. The capacity factor k′ for the solute retention in LC is represented by the following well-known relationship.2,17

ln k′ ) -∆H°/RT + ∆S°/R + ln Φ

(1)

In eq 1, ∆H° and ∆S° are the standard enthalpy and entropy changes associated with the distribution of solute molecules from the mobile phase to the stationary phase. Φ is the volume ratio of the stationary phase to the mobile phase. In SEC mode, the size exclusion mechanism is governed by the loss of the conformational entropy of polymer chains when they are placed into the small pore space (stationary phase in SEC). Therefore, ∆S is the main thermodynamic parameter for the SEC fractionation of macromolecules and it is a common practice to minimize ∆H by employing a thermodynamically good solvent for the solutes. In contrast, enthalpic interaction between the solutes and the stationary phase (∆H) plays a major role in the IC mode of separation. When porous column packing materials are employed in the IC mode separation, the size exclusion effect on polymeric solutes always exists due to the large size of polymer molecules. Since it is common in IC to employ a porous stationary phase in order to increase the surface area to interact with the solute molecules, both interaction and size exclusion mechanism work together in the LC separation of polymers.15,18 The chromato(6) Stadalius, M. A.; Quarry, M. A.; Mourey, T. H.; Snyder, L. R. J. Chromatogr. 1986, 358, 17-37. (7) Lochmu ¨ ller, C. H.; McGranaghan, M. B. Anal. Chem. 1989, 61, 24492455. (8) Northrop, D. M.; Martire, D. E.; Scott, R. P. W. Anal. Chem. 1992, 64, 16-21. (9) Shalliker, R. A.; Kavanagh, P. E.; Russell, I. M.; Hawthorne, D. G. Chromatographia 1992, 33, 427-433. (10) Lochmu ¨ ller, C. H.; Moebus, M. A.; Liu, Q. C.; Jiang, C.; Elomaa, M. J. Chromatogr. Sci. 1996, 34, 69-76. (11) Lee, H. C.; Chang, T. Polymer 1996, 37, 5747-5749. (12) Lee, W.; Lee, H. C.; Kim, S. B.; Chang, T. Macromolecules 1998, 31, 344348. (13) Chang, T.; Lee, H. C.; Lee, W.; Park, S.; Ko, C. Macromol. Chem. Phys. 1999, 200, 2188-2204. (14) Lee, W.; Lee, H. C.; Park, T.; Chang, T.; Chae, K. H. Macromol. Chem. Phys. 2000, 201, 320-325. (15) Glo ¨ckner, G. Gradient HPLC of Copolymers and Chromatographic CrossFractionation; Springer-Verlag: Berlin, 1992. (16) Pasch, H.; Trathnigg, B. HPLC of Polymers; Springer-Verlag: Berlin, 1997. (17) Dorsey, J. G.; Cooper, W. T. Anal. Chem. 1994, 66, 857A. (18) Berek, D. Prog. Polym. Sci. 2000, 25, 873-908. 10.1021/ac010072o CCC: $20.00

© 2001 American Chemical Society Published on Web 07/07/2001

graphic critical condition is a unique phenomenon for polymeric solutes at which the entropic size exclusion effect of the polymer solutes is precisely compensated by the enthalpic interaction effect. Liquid chromatography at the critical condition (LCCC) has been reasonably well established experimentally as well as theoretically, and the technique is regarded as an attractive tool for the characterization of telechelic polymers and block copolymers.16,19-24 At the critical condition of a homopolymer, retention of the polymer becomes independent of molecular weight. If the polymer solutes have an additional difference in chemical structure other than the portion under its critical condition, the retention of the polymer is supposed to be solely governed by the extra part such as end groups in telechelic polymers,24-26 the other blocks in a block copolymer,27-29 or tacticity differences.30, 31 With a given stationary phase, a critical condition for a polymer of interest is achieved by choosing an appropriate mobile phase and adjusting the column temperature. So far, mixed-eluent systems have been commonly employed for the purpose.16 By adjusting the composition of a mixed-solvent system, the enthalpy of interaction between the solute and the stationary phase is controlled and the fine-tuning for a precise critical condition is done by adjusting the temperature of the column. It is not easy, however, to reproduce the critical condition since the retention of polymers is very sensitive to the solvent composition. A small variation in the solvent composition or a little moisture contamination of the solvents affects the critical condition significantly.23,32 Therefore, the use of a single solvent is desirable to improve not only the reproducibility among different laboratories but also the repeatability in the same laboratory. In this paper, we present the first report of a single-solvent chromatographic critical condition to our knowledge. In addition, we found that the solvent condition is suitable for the separation of the polymer by temperature gradient interaction chromatography (TGIC), which has been established as a very efficient method for the separation of macromolecules in terms of molecular weight.11-13 EXPERIMENTAL SECTION HPLC Apparatus. The HPLC system consists of a solvent delivery pump (LDC, CM 3200), a six-port sample injector (19) Gorshkov, A. V.; Evreinov, V. V.; Entelis, S. G. Zh. Fiz. Khim. 1985, 59, 958. (20) Gorshkov, A. V.; H., M.; Becker, H.; Pasch, H.; Evreinov, V. V.; Entelis, S. G. J. Chromatogr. 1990, 523, 91. (21) Gorbunov, A. A.; Skvortsov, A. M. Adv. Colloid Interface Sci. 1995, 62, 31. (22) Guttman, C. M.; Di Marzio, E. A.; Douglas, J. F. Macromolecules 1996, 29, 5723-5733. (23) Berek, D. Macromol. Symp. 1996, 110, 33-56. (24) Yun, H.; Olesik, S. V.; Marti, E. H. Anal. Chem. 1998, 70, 3298-3303. (25) Evreinov, V. V.; Gorshkov, A. V.; Prudskova, T. N.; Guryanova, V. V.; Pavlov, A. V.; Malkin, A. Y.; Entelis, S. G. Polym. Bull. 1985, 14, 131-136. (26) Skvortsov, A. M.; Zhulina, Y. B.; Gorbunov, A. A. Polym. Sci. USSR 1980, 22, 908-918. (27) Pasch, H.; Brinkmann, C.; Gallot, Y. Polymer 1993, 34, 4100-4104. (28) Lee, H.; Lee, W.; Chang, T. Macromolecules 1999, 32, 4143-4146. (29) Lee, W.; Cho, D. Y.; Chang, T. Y.; Hanley, K. J.; Lodge, T. P. Macromolecules 2001, 34, 2353-2358. (30) Berek, D.; Janco, M.; Hatada, K.; Kitayama, T.; Fujimoto, N. Polym. J. 1997, 29, 1029-1033. (31) Kitayama, T.; Janco, M.; Ute, K.; Niimi, R.; Hatada, K.; Berek, D. Anal. Chem. 2000, 72, 1518-1522. (32) Philipsen, H. J. A.; Klumperman, B.; van Herk, A. M.; German, A. L. J. Chromatogr., A 1996, 727, 13-25.

(Rheodyne, 7125) equipped with a 20-µL injection loop, a variablewavelength UV/visible absorption detector (TSP, Spectra 100) and/or a refractive index (RI) detector (Shodex RI-71) and/or an evaporative light-scattering detector (ELSD) (PL-EMD 950). For the LCCC and simultaneous SEC-TGIC experiments, three reversed-phase columns (Nucleosil C18, 100, 300, and 500 Å, 4.6 mm i.d. × 250 mm length) were serially connected. For the TGIC experiments, one column (Nucleosil C18, 100 Å, 4.6 mm i.d. × 250 mm length) was used. The temperature of the separation column was controlled by circulating fluid from a programmable bath/circulator (Neslab, RTE-111) through a homemade column jacket. A low-angle laser light scattering (LALLS) detector (LDC, KMX-6, λ ) 632.8 nm) was connected before RI detector in the TGIC and SEC-TGIC experiments for on-line determination of the absolute molecular weight of polymers. When the LALLS detector was used, a 100-µL injection loop was used to ensure sufficient scattering intensity. Materials. HPLC grade 1,4-dioxane was purchased from Fischer Scientific Inc. and further purified by fractional distillation in contact with CaH2 (95%, Aldrich) to remove residual moisture. Dried 1,4-dioxane was then degassed by He bubbling for 30 min before use in the chromatography analysis. Polyisoprene (PI) samples were synthesized using an anionic polymerization apparatus under Ar atmosphere, and cyclohexane (99%, ACS reagent, Aldrich) was used as the polymerization solvent. Details of the polymerization apparatus were reported previously.33 The molecular weight distribution of PI samples was analyzed by SEC with two polystyrene (PS) gel columns (Polymer Lab. PL-mixed C, 8 mm i.d. × 300 mm length). In SEC, the eluent was THF (HPLC grade, Aldrich) and the column temperature was kept at 35 °C by use of a column oven (Eppendorf, TC-50). The weight average molecular weight (Mw) and the molecular weight polydispersity (Mw/Mn) were calculated by the universal calibration method34 with PS standards (Polymer Lab.) using MarkHouwink constants (PS, [η] ) 9.8 × 10-3 M0.74 mL/g; PI, [η] ) 19.5 × 10-3 M0.73 mL/g at 35 °C) measured in our laboratory.14 The calibration curve was obtained by fitting the experimental data to a third-order polynomial. Microstructure of PI was characterized by 1H NMR (Bruker, DPX-300).35 The characterization results are summarized in Table 1. Two PS/PI diblock copolymers (PS-b-PI) differing in the PI block length (SI-1 and SI-2) and the precursor PS were prepared by a single batch anionic polymerization. A PS block was polymerized first and then PI block by adding the corresponding monomers sequentially, which ensure a constant PS block length. Prior to each addition of isoprene monomers, appropriate amounts of the PS precursor and SI-1 were taken out from the reactor at the end of each polymerization step. Characterization by SECLALLS and 1H NMR revealed that 6.0 and 21.4 kg/mol PI blocks were attached to 12.0 kg/mol PS precursor for SI-1 and SI-2, respectively.(Table 2). RESULTS AND DISCUSSION Critical Condition of PI with 1,4-Dioxane as the Eluent. At a chromatographic critical condition, the entropic size exclusion (33) Kwon, K.; Lee, W.; Cho, D.; Chang, T. Korea Polym. J. 1999, 7, 321-324. (34) Grubisic, Z.; Rempp, P.; Benoit, H. J. Polym. Sci. B, Polym. Lett. 1967, 5, 753. (35) Tanaka, Y.; Takeuchi, Y.; Kobayashi, M.; Tadokoro, H. J. Polym. Sci. Part A-2 1971, 9, 43.

Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

3885

Table 1. Characterization of PS and PI Used in This Study

sample

microstructure of PIa/ trans-1,4: cis-1,4:vinyl

PI-1 PI-2 PI-3 PI-4 PI-5 PS-1 PS-2 PS-3 PS-4 PS-5

27:66:7 24:69:6 18:77:5 20:75:5 23:71:6

SEC calibrationb Mw(× 103)

Mw/Mn

2.7 10.4 24.5 53.0 208 2890 433 135 14.0 2.5

1.08 1.08 1.18 1.03 1.04 1.09 1.04 1.04 1.06 1.12

TGIC-LSc

SEC/TGIC-LSd

Mw(× 103)

Mw/Mn

Mw (× 103)

Mw/Mn

10.0 25.3 57.1 199

1.03 1.07 1.01 1.01

10.1 24.8 54.2 188

1.02 1.08 1.01 1.02

432 133 14.2

1.01 1.01 1.01

a Measured by 1H NMR. b From SEC-universal calibration analysis. c From TGIC-light-scattering detection analysis shown in Figure 4. d From simultaneous SEC/TGIC-light-scattering detection analysis shown in Figure 5.

Table 2. Characterization of PS-b-PI and PS Precursor

PS precursor SI-1 SI-2

PS wt %

MPS - MPI (kg/mol)

Mw (kg/mol)/Mw/Mn of PS block by LCCC of PI at 47.6 °C at 47.7 °C at 47.8 °C

100 67 36

12.0-0 12.0-6.0 12.0-21.4

12.0/1.06 10.4/1.07 9.8/1.07

11.9/1.06 10.4/1.07 9.9/1.07

12.1/1.06 10.4/1.06 9.6/1.07

effect cancels the enthalpic interaction effect. Therefore, a critical condition eluent is generally not a good solvent for a given polymer.13,36 At the same time, interaction strength has to be not too strong. Otherwise, the polymer chains are adsorbed to the stationary phase too strongly to elute in a reasonable time. When we use C18 bonded silica reversed-phase columns, we empirically observed that θ solvents for a polymer often fit this requirement very well.13 1,4-Dioxane, a θ solvent for polyisoprene, has a solubility parameter similar to the mixture of CH2Cl2/CH3CN (80/ 20, v/v) which was used for a critical condition of PI with a C18 bonded silica reversed-phase column.29 The Hildebrand solubility parameter, δ, is a good measure of the solvent power for polymeric solutes.32,36 The δ value of 1,4-dioxane is 20.5 MPa1/2 37 while that of the 80/20 (v/v) mixture of CH2Cl2/CH3CN is 20.7 MPa1/2, which was determined by the following relationship.38

δmixture ) φCH2Cl2δCH2Cl2 + φCH3CNδCH3CN

(2)

where φ is the volume fraction of the corresponding component in the mixed eluent. In Figure 1 are displayed five chromatograms of PI samples of four different molecular weights taken at different temperatures using 1,4-dioxane as the eluent. Varying the column temperature, the retention mode of PI changes between SEC and IC mode via the critical condition at 47.7 °C. The high sensitivity of the retention of the PI samples to the temperature variation indicates that the temperature of both mobile and stationary phases has to be maintained precisely for accurate establishment of a critical condition. In this study, not only the column temperature but also (36) Baran, K.; Laugier, S.; Cramail, H. Macromol. Chem. Phys. 1999, 200, 20742079. (37) Grulke, E. A. In Polymer Handbook; Brandrup, J., Immergut, E. H., Grulke, E. A., Eds.; John Wiley & Sons: New York, 1999; pp VII-675. (38) Snyder, L. R. J. Chromatogr. 1974, 92, 223.

3886 Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

Figure 1. Temperature effect on the retention of four PI samples (Mw ) 2.7, 10.4, 53.0, and 208 kg/mol). 1,4-Dioxane was the eluent and three reversed-phase columns (Nucleosil C18, 100, 300, and 500 Å, 250 mm × 4.6 mm i.d.) were used.

the eluent temperature was controlled before it reached the injector by passing through a long stainless tubing (Alltech, 1 m × 0.5 mm i.d.) immersed in a water bath/circulator. When mixed eluents are used as a mobile phase, the solute retention also depends on the eluent composition. Only a fraction of a percent in composition would make a noticeable difference in the solute retention and this is why fine-tuning of the retention is commonly done by adjusting the temperature. In Figure 1, the critical temperature of PI is found at 47.7 °C. The θ temperature in 1,4dioxane is reported as 31.2 °C for 96% cis PI and 47.7 °C for trans

Figure 2. Plot of log M vs retention time (tR) for five PSs and four PIs at various temperatures. Inset shows an enlarged plot for three temperatures near the critical condition. Other experimental condition is the same as in Figure 1.

PI, respectively.39 Since four PIs used in this study have 70 ( 2% cis content, the θ temperature would be a little higher than 31.2 °C but certainly lower than 47.7 °C, which ensures good solubility of the polymer samples. The temperature-dependent elution behavior of PIs was examined more critically, and the results are displayed in Figure 2, which shows a plot of log M versus elution time (peak position) for the PI and PS samples at various temperatures. As shown in an enlarged plot in the inset, the critical condition of PI is found at 47.7 °C at a precision of 0.1 °C. We found that it is not worth it to pursue a more precise critical condition since the deviation of the critical condition was of the order of 0.1 °C when a different batch of 1,4-dioxane was used or the experiment was carried out in a different day. But it is a significant improvement in reproducibility compared with mixed-solvent systems. Figure 2 also shows a clear difference in the temperature dependency of the retention behavior of PS and PI. The elution time of the PS samples remained practically unchanged with the variation of the column temperature while the elution of PI samples spans between SEC and IC regime. This indicates that PS samples are retained exclusively by the SEC mechanism since 1,4-dioxane is a good solvent for PS. Separation by the SEC mechanism is governed by the entropic effect and the retention is insensitive to the temperature variation. Analysis of PS Block of PS-b-PI under the Critical Condition of PI. For an AB diblock copolymer, the elution behavior of the diblock copolymer would be solely determined by the size of the B block according to the proposition that A block of the AB diblock copolymer becomes “chromatographically invisible” at the chromatographic critical condition of the A block. Since the proposition was first made by Russian scientists,40,41 based upon the early theoretical works of Casassa on the size exclusion process from the pore42 and of DiMarzio and Rubin on (39) Elias, H.-G. In Polymer Handbook; Brandrup, J., Immergut, E. H., Grulke, E. A., Eds.; John Wiley & Sons: New York, 1999; pp VII-291. (40) Belenkii, B. G.; Gankina, E. S.; Tennikov, M. B.; Vilenchik, L. Z. Dokl. Akad. Nauk. SSSR 1976, 231, 1147. (41) Skvortsov, A. M.; Gorbunov, A. A. Vysokomol. Soedin. 1979, A21, 339.

the adsorption of polymer chain to the stationary phase,43 there have been many reports on the analysis of individual block size of block copolymers by the LCCC method.20,22,24,25,27-29,44-47 The applications of LCCC for the characterization of block copolymers made so far can be divided into two categories. One is to elute the block copolymers in the IC elution regime (thus eluting after the system peak) in terms of the length of the block(s) to be analyzed. This is realized if the “visible” block is more interactive with the stationary phase than the “invisible” block. Since the elution condition needs to be fixed at the critical condition for the block made “invisible”, the applications are limited to end group analysis or to rather short block lengths due to the exponential dependence of the retention on the chain length (Martin’s rule).20,25,28,47,48 The other separation mode is to elute the block copolymers in the size exclusion regime (eluting before the system peak) in terms of the block length of the block(s) to be analyzed if the “visible” block is less interactive with the stationary phase. The majority of the LCCC applications to the characterization of block copolymers has been made in the latter mode, and the molecular weight distribution of the block of interest is determined by the standard calibration method commonly used in size exclusion chromatography.22,24,27,29,44-46 Although most of the reports are in strong favor of the feasibility of the individual block analysis, we found a small but systematic deviation from the ideal chromatographic invisibility at the critical condition in a recent study with the PS-b-PI system.29 One difficulty associated with the LCCC analysis is the use of mixed-solvent systems to establish a critical condition. Since individual components in a mixed solvent can adsorb differently onto polymer chains and stationary phase, and furthermore, the preferential sorption may depend on the molecular weight of the polymer chains, it is not simple to draw a conclusion on the issue of chromatographic invisibility.49 Therefore, the establishment of single-solvent critical conditions is important not only to improve the reproducibility of the LCCC technique but also to shed light on the issue of the chromatographic invisibility. Under the critical condition of PI illustrated in Figures 1 and 2, we performed the block size analysis of PS-b-PI. We also did the same analysis under slightly off-critical conditions. Figure 3a displays the elution peaks of four different PI samples at different temperatures near the critical condition. The peak amplitudes are adjusted for visual aid. One can easily find that the elution sequence is reversed across the critical condition (47.7 °C) at which four PIs were eluted at the same elution time. Under each temperature, the PS block size of PS-b-PI diblock copolymers was analyzed for two block copolymers, SI-1 and SI-2 differing in the (42) Casassa, E. F. J. Polym. Sci., Part B, Polym. Lett. 1967, 5, 773. (43) DiMarzio, E. A.; Rubin, R. J. J. Chem. Phys. 1971, 55, 4318-4336. (44) Gorbunov, A. A.; Skvortsov, A. M. Vysokomol. Soedin., Ser. A 1988, 30, 895. (45) Zimina, T. M.; Kever, J. J.; Melenevskaya, E. Y.; Fell, A. F. J. Chromatogr. 1992, 593, 233-241. (46) Falkenhagen, J.; Much, H.; Stauf, W.; Muller, A. H. E. Macromolecules 2000, 33, 3687-3693. (47) Lee, H.; Chang, T.; Lee, D.; Shim, M. S.; Ji, H.; Nonidez, W. K.; Mays, J. W. Anal. Chem. 2001, 73, 1726-1732. (48) Pasch, H.; Brinkmann, C.; Much, H.; Just, U. J. Chromatogr. 1992, 623, 315. (49) Berek, D.; Janco, M.; Meira, G. R. J. Polym. Sci. Part A. Polym. Chem. 1998, 36, 1363-1371.

Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

3887

Figure 4. TGIC chromatogram of five PIs using 1,4-dioxane and a reversed-phase column (Nucleosil C18, 100 Å) as the mobile and the stationary phases, respectively. Both RI (solid line) and LS (dashdot) detector responses are shown. Temperature program applied to the column is shown at the top abscissa.

Figure 3. (a) Overlapped chromatograms of four PIs obtained at three different temperatures near the critical condition of PI. Amplitudes of the elution peaks are rescaled for visual aid. (b) Chromatograms of SI-1, SI-2, and the PS precursor at the same temperatures as in (a). Calibration curves are also shown in the plot.

PI block length as well as the PS precursor of the block copolymers. Figure 3b displays the chromatograms of three samples obtained at three different temperatures together with the calibration curves made with four PS standards by fitting the experimental data to a third-order polynomial. For the detection of the eluted samples, the wavelength of UV detector was set at 235 nm at which both PI and PS absorb the light. Since the PS precursor, SI-1, and SI-2 were made in the same batch by sequential addition of the styrene and isoprene monomers, the PS block length should be precisely the same for the three samples. At the critical condition of PI (47.7 °C), however, the elution peaks of the three samples are not overlapped precisely but show a clear trend that the retention increases as the PI block length increases. This indicates that the attached PI block is not completely “invisible” but affects the retention of the block copolymer at the apparent critical condition of PI. The molecular weights (Mw) and polydispersity (Mw/Mn) of the PS part of the three samples calculated by the calibration curve are listed in Table 2. The Mw of the PS precursor was exactly the same as the SEC-LS result while Mw of PS blocks in SI-1 and SI-2 were underestimated by more than 10% of the true values, which is far beyond the repeatability of these measurements. This phenomenon is consistent with the results obtained with a mixedsolvent system.29 Furthermore, it is interesting to note that almost the same deviation of block copolymer elution was observed: The elution time of the precursor and the two block copolymers remains almost unchanged at three different temperatures around the critical temperature. This result indicates that the conclusion 3888 Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

of the chromatographic invisibility of a block at the critical condition for block copolymers is not true at least for the PS-b-PI system examined in this study. Rigorous examination on a variety of polymer/eluent systems is called for to verify whether our observation from PS-b-PI is a general behavior in LCCC analysis of block copolymers. In any event, the use of a single solvent improved the reproducibility of the critical condition a lot and the rigorous test of chromatographic invisibility for the block copolymer analysis cast serious doubt on its validity. Temperature Gradient Interaction Chromatography of PI. As shown in Figures 1 and 2, at a temperature lower than the critical point, the retention of PI increases rapidly with molecular weight. TGIC utilizes the temperature dependence of the retention for the separation of polymers according to the molecular weight.13 By changing the column temperature during the chromatographic run, five PIs of different molecular weights were completely separated according to their molecular weights as shown in Figure 4. As shown at the top abscissa, the temperature of the column was raised by a series of two linear ramps from 24 to 42 and to 46 °C. A single-solvent system also allowed us to employ light scattering (LS) detection without complication of the preferential sorption problem.12,50 In Figure 4, the TGIC run was monitored by double detection of LS and RI. The baselines of both chromatograms are very stable, indicating that the temperature variation during the elution causes little refractive index drift when the eluent reaches the detectors, which is crucial to obtain absolute molecular weights from the LS detection.12 Relative intensities of the elution peaks recorded by LS and RI detectors directly reflect the molecular weight differences of eluted samples. A small peak that appears around tR of 35 min corresponds to the coupling product of the molecular weight twice as large as (∼416 kg/mol) the major product. Polymers prepared by anionic polymerization often (50) Lee, H. C.; Chang, T.; Harville, S.; Mays, J. W. Macromolecules 1998, 31, 690-694.

contains such a side product by from oxygen-involved termination step. To measure the absolute Mw and the MWD of the PIs, individual samples were dissolved in 1,4-dioxane at a known concentration, and chromatograms were obtained from separate TGIC runs. For the calculation of absolute Mw, the specific refractive index increment (dn/dc) of PI in 1,4-dioxane was determined to be 0.104 mL/g by use of a differential refractometer (LDC, KMX-16). The Mw and MWD values are summarized in Table 1. The Mw values are in good agreement with that determined by SEC, but MWD values are much smaller due to the higher resolution of TGIC as reported earlier.11,13,14,51 This confirms that 1,4-dioxane is an excellent single-solvent eluent for TGIC of PI. Simultaneous SEC and TGIC of PI/PS Mixture. Binary polymer mixtures can be separated efficiently at the critical condition of one component.52 If the eluent at the critical condition of a component is a good solvent for the other component, the latter elutes at the SEC regime and the characterization of the individual component is possible. This type of separation can be done more efficiently if the temperature is varied during the elution so that one component is separated by the TGIC mechanism while the other component is separated by the SEC mechanism simultaneously.14,53 As shown in Figure 2, the retention of PS is nearly independent of the column temperature in 1,4dioxane solvent. Figure 5a shows a chromatogram where the simultaneous SEC separation for PS and TGIC separation for PI is established. Again the double detection of LS and RI allows us on-line determination of the absolute Mw of the polymers without difficulty. The temperature gradient program was applied in a series of linear ramps as shown in the upper abscissa. Prior to the solvent elution peak, five PSs were separated in decreasing order of their molecular weights (SEC mechanism) while five PIs were separated by TGIC mode after the solvent peak in increasing order of their molecular weight. The LS detector signal of PS samples is much larger than those of PI samples since the light-scattering intensity is proportional to (dn/ dc)2 and dn/dc of PS (0.181 mL/g) is 1.7 times as large as that of PI in 1,4-dioxane. Because the separation window in SEC regime is relatively narrow, the portion of the chromatogram before the solvent elution peak is enlarged for better visualization in Figure 5b. With the raw data of LS and RI in Figure 5, and the measured dn/dc values of PS and PI in 1,4-dioxane, Mw of individual species were calculated and listed in Table 1. Thus determined Mw matches well with the SEC characterization. In summary, a chromatographic critical condition using a single solvent was reported for the first time: PI with C18 bonded silica and 1,4-dioxane as stationary and mobile phases, respectively. Use of a single solvent provided excellent reproducibility in chromatographic retention, which is very sensitive to the solvent composition when a mixed eluent system is employed. θ Solvents having a solubility parameter similar to an already known mixed-solvent system seem to be good candidates for a single-solvent system (51) Lee, W.; Lee, H.; Cha, J.; Chang, T.; Hanley, K. J.; Lodge, T. P. Macromolecules 2000, 33, 5111-5115. (52) Pasch, H. Polymer 1993, 34, 4095-4099. (53) Lee, H. C.; Chang, T. Macromolecules 1996, 29, 7294-7296.

Figure 5. (a) Separation of PSs and PIs by simultaneous SEC/ TGIC using 1,4-dioxane as an eluent. (b) Enlarged chromatogram of (a) showing the separation of PSs by SEC mode. Experimental condition is the same as in Figure 1 except for the temperature gradient program (shown at the top abscissa) and the loop volume (100 µL).

to establish a critical condition. At the column temperature of 47.7 °C, PI samples elute at the same retention volume independent of molecular weight. Under the critical condition of PI, molecular weight analysis of PS block of PS-b-PI block copolymers was carried out. A small but systematic deviation from the expectation of chromatographic invisibility was observed in the retention behavior of the block copolymers. The single-solvent system also worked well for the TGIC fractionation of PI and for the simultaneous SEC/TGIC fractionation of PS/PI mixtures. The determination of absolute molecular weights by on-line lightscattering LS detection is facilitated using a single-solvent system. ACKNOWLEDGMENT This study was supported by KOSEF (Center for Integrated Molecular Systems) and KRF (BK21 Program). We also acknowledge POSTECH for the instrumental support. Received for review January 17, 2001. Accepted May 24, 2001. AC010072O

Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

3889