Separation of branched polystyrene by comprehensive two-dimensional liquid chromatography

Journal of Chromatography A, 1103 (2006) 235–242

Separation of branched polystyrene by comprehensive two-dimensional liquid chromatography Kyuhyun Im a , Youngtak Kim a , Taihyun Chang a,∗ , Kwanyoung Lee b , Namsun Choi b a

Department of Chemistry and Polymer Research Institute, Pohang University of Science and Technology, Pohang 790-784, South Korea b Korea Kumho Petrochemical Co. Ltd., Kumho Petrochemical R&D Center, P.O. Box 64, Yuseong, Daejeon 305-600, South Korea Received 20 June 2005; received in revised form 1 November 2005; accepted 7 November 2005 Available online 6 December 2005

Abstract Branched polystyrenes (PS) featuring a bivariate distribution in the molecular weight and in the number of branches were characterized by comprehensive two-dimensional liquid chromatography (2D-LC). The branched PS were prepared by anionic polymerization using n-butyl Li as an initiator and a subsequent linking reaction with p-(chlorodimethylsilyl)styrene (CDMSS). The n-butyl Li initiator yields polystyryl anions with broad molecular weight distribution (MWD) and the linking reaction with CDMSS yields branched PS with different number of branches. For the first dimension (1st-D) separation, reversed-phase temperature gradient interaction chromatography (RP-TGIC) was employed to separate the branched polymer according to mainly the molecular weight. In the second dimension (2nd-D) separation, the effluents from the RP-TGIC separation are subjected to liquid chromatography at chromatographic critical conditions (LCCC), in which the separation was carried out at the critical condition of linear homo-PS to separate the branched PS in terms of the number of branches. The 2D-LC resolution of RP-TGIC × LCCC combination worked better than the common LCCC × size-exclusion chromatography (SEC) configuration due to the higher resolution of RP-TGIC in molecular weight than SEC. Furthermore, by virtue of using the same eluent in RP-TGIC and LCCC (only the column temperature is different), RP-TGIC × LCCC separation is free from possible ‘break through’ and large system peak problems. This type of 2D-LC separation could be utilized efficiently for the analysis of branched polymers with branching units distinguishable by LC separation. © 2005 Elsevier B.V. All rights reserved. Keywords: Two-dimensional liquid chromatography; Branched polymer; Molecular weight distribution; Branch number distribution; Functionality; Temperature gradient interaction chromatography; Liquid chromatography at the critical condition; MALDI-TOF MS

1. Introduction Liquid chromatography (LC) is a powerful tool for the molecular characterization of polymers that often have multivariate distribution in molecular characteristics such as molecular weight, chain architecture, chemical composition, and functionality [1–8]. The LC separation of polymers can be largely divided into three different modes: the size exclusion mode, the critical condition mode, and the interaction mode. Size-exclusion chromatography (SEC) is the most widely employed technique in the molecular characterization of polymers due to its high speed, facility to use, and wide applicability [1]. Owing to the separation mechanism of size exclusion, however, SEC has an intrinsic limitation. SEC cannot distinguish polymers of the same hydro-

∗

Corresponding author. Tel.: +82 54 279 2109; fax: +82 54 279 3399. E-mail address: [email protected] (T. Chang).

0021-9673/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2005.11.050

dynamic size but with differences in various molecular characteristics such as composition, chain architecture, microstructure, functionality and so on. Nonetheless, due to the lack of better methods, SEC has been the most popular method for the polymer analysis of various heterogeneities and long chain branching has been characterized by SEC separation combined with light scattering and/or viscometry detection [9,10]. The average number of branches can be estimated according to the method of Zimm and Stockmayer from the molecular weight and the size of the polymers in the SEC fraction determined by light scattering and viscometry detection [11]. The analysis scheme is based on the assumption that the fractions of the SEC effluents are homogeneous in molecular weight and branch number, which is certainly not correct. Unlike SEC, interaction chromatography utilizes the interactions between polymer segments and the stationary phase, and the molecular weight resolution of interaction chromatography is not much affected by the chain architecture [12]. Interaction

236

K. Im et al. / J. Chromatogr. A 1103 (2006) 235–242

chromatography was successfully employed to characterize a few model branched polymers such as star shaped polymers [12,13], H-shaped polymers [14], mikto-arm star block copolymers [15,16], and hyper-branched polymers [17]. These model branched polymers were prepared by linking precursor polymers with a narrow distribution in molecular weight. Therefore, the number of branches is proportional to the molecular weight and interaction chromatography was able to separate them in terms of the branch number far better than SEC. This is due to the fact that TGIC separates not with respect to hydrodynamic size but with respect to molecular weight. However, most of the branched polymers in practical use have distributions in both branch molecular weight and number of branches. Accordingly, for the analysis of the polymers having a bivariate distribution, a new method must be devised to obtain the information on branch number distribution as well as molecular weight distribution (MWD). For such an analysis, two-dimensional liquid chromatography (2D-LC) is a natural choice if each LC separation can separate the polymers according to one type of distribution independent of the other. The type of LC for the first and the second dimension separations needs to be chosen to maximize the resolution of the bivariate distribution. The best configuration of a 2D-LC separation is to find chromatographic methods exclusively sensitive to one of the two molecular characteristics while suppressing the effect of the other so that each separation becomes orthogonal to each other [18]. 2D-LC has been used for many years to characterize synthetic polymers, biomolecules and complex mixtures [5,19–31]. In the comprehensive 2D-LC setup, a switching valve equipped with two loops was used in a symmetrical configuration. While one loop is being filled with the 1st-D effluent, the fraction that was previously collected in the other loop is analyzed in the 2ndD separation. The collection time of each fraction in the 1st-D separation has to be equal to or longer than the analysis time in the 2nd-D separation. As a consequence, the analysis time in the 2nd-D separation and the loop volume together determine the flow rate for the 1st-D separation. The total analysis time is essentially the product of the analysis time of the 2nd-D separation and the number of fractions collected from the 1st-D separation. Therefore, to reduce the total analysis time and to keep the 1st-D separation efficiency from the aggravation due to the slow elution, it is imperative to keep the 2nd-D separation fast. The common configuration of 2D-LC in the analysis of copolymers or functional polymers has been interaction chromatography or LC at critical conditions (LCCC) as the 1st-D to separate in terms of the molecular characteristics other than the molecular weight, mainly the composition or functionality, and SEC as the 2nd-D to separate according to the molecular weight [4,25,28,30,32]. SEC has been a preferred choice for the 2nd-D separation of polymers since SEC is a universal technique to separate polymers according to the molecular size. The SEC retention is seriously affected by molecular characteristics other than molecular weight, but once homogeneity of the sample is achieved after the 1st-D separation, the SEC retention is more meaningfully correlated with the molecular weight. Furthermore, with the development of new type of

columns, SEC separation can be carried out fast in a few minutes [33]. In this work, we demonstrate the use of a comprehensive 2DLC method to separate branched polystyrenes (PS) according to both molecular weight and branch number. 2. Experimental 2.1. Preparation of branched polystyrenes The branched PS were prepared by linking polystyryl anion precursors with 4-chlorodimethylsilylstyrene (CDMSS) [34]. CDMSS was prepared following the literature procedure [35] and stored at −25 ◦ C in a dry box filled with Ar gas. nButyl Li (2.0 M solution in cyclohexane), cyclohexane (anhydrous, 99.5%), and styrene (99%) were purchased from Aldrich. Tetrahydrofuran (THF) and methanol were purchased from Merck (HPLC grade). Styrene was distilled from calcium hydride under reduced pressure. THF and cyclohexane were distilled from sodiobenzophenone. Styrene, THF, cyclohexane, and methanol were transferred into a dry box filled with argon gas and purified again by passing through a column filled with activated neutral alumina (Merck, 70–230 mesh) that was baked at 250 ◦ C for 24 h in vacuo. Purified styrene was stored at −25 ◦ C in the dry box. Polystyryl anions were prepared by adding n-butyl Li 2.55 mL (5.09 mmol) to a solution of cyclohexane 105 mL, styrene 10.5 mL and THF 284 ␮L under Ar atmosphere at 25 ◦ C. The polymerization mixture was stirred for 30 min at 25 ◦ C and three portions of 38 mL solution were transferred to three flasks. One of the separated solutions of polystyryl lithium was terminated with methanol to obtain the precursor PS and the other two portions were treated with CDMSS at the CDMSS/living end molar ratios of 0.49 (162.5 ␮L, BS49), and 0.65 (216.6 ␮L, BS65) at 25 ◦ C. After allowing the reaction to proceed for 3 h, the reaction mixtures were terminated by adding a small amount of methanol. The PS was precipitated by dropping the reaction mixture into a large excess of methanol and the precipitated PS was collected by filtration and dried at 60 ◦ C under vacuum. The scheme of the coupling reaction with CDMSS is shown in Fig. 1. 2.2. SEC analysis For the SEC analysis, two mixed bed columns (Polymer Labs., PLgel Mixed-C, 300 mm × 7.5 mm) were used at a column temperature of 40 ◦ C. SEC chromatograms were detected with a multi-angle laser light scattering (MALLS, Wyatt, miniDAWN) at a wavelength of 690 nm and a refractive index detector (Wyatt, Optilab DSP) using THF (Samchun, HPLC grade) as the eluent at a flow rate of 0.8 mL/min. Polymer samples for the SEC analysis were dissolved in THF at a concentration of ca. 1.0 mg/mL and the injection volume was 100 ␮L. 2.3. RP-TGIC and LCCC analysis For the RP-TGIC separation, a C18 bonded silica column ˚ 150 mm × 4.6 mm) was used. The mobile (Kromasil C18, 100 A,

K. Im et al. / J. Chromatogr. A 1103 (2006) 235–242

237

Fig. 1. Reaction scheme and possible products from the linking reaction of polystyryl anions with CDMSS. (There are other possible isomeric structures [34].)

phase was a CH2 Cl2 /CH3 CN mixture (53/47, v/v, Samchun, HPLC grade) delivered by a Bischoff HPLC compact pump at a flow rate of 0.5 mL/min. The same system was used for the LCCC separation. The critical condition for polystyrene was established at 53.3 ◦ C. The injection samples were prepared by dissolving the polymers in the eluent. The temperature of the separation columns was controlled by circulating fluid from a programmable bath/circulator (ThermoHaake, C25P) through a homemade column jacket. All chromatograms were obtained with a UV absorption detector (TSP, UV 100) operating at a wavelength of 260 nm. 2.4. RP-TGIC × LCCC 2D-LC analysis For the RP-TGIC × LCCC 2D-LC analysis, the same conditions as in the 1D-LC analysis were used except for the flow rate. The flow rate of the 1st-D RP-TGIC was set low at 0.02 mL/min to synchronize with the 2nd-D LCCC separation operated at a flow rate of 1.2 mL/min. The schematic diagram of the comprehensive 2D-LC instrument is shown in Fig. 2. The key component of the system is an electronically controlled 2-position, 10-port switching valve (Alltech, SelectPro) that enables con-

tinuous, alternate sampling of the 1st-D column effluent and injection to the 2nd-D LC column through two equivalent sample loops. In one of the two positions, the effluent from the 1st-D column fills one of the two 100 ␮L sample loops. At the same time, the content of the other 100 ␮L loop from the previous cycle is sampled onto the second column. When the valve is switched to the second position, the effluent from the 1st column fills up the other loop, while the loop filled during the previous cycle is sampled onto the second column. The valve is switched back and forth every 5 min. A UV detector (TSP, UV 100) is inserted between 1st-D column and the 10-port switching valve to obtain the chromatogram of the 1st-D separation. 2.5. LCCC × SEC 2D-LC analysis For the 1st-D LCCC separation, the same experimental condition as the 2nd-LCCC in RP-TGIC × LCCC 2D-LC analysis was used except for the flow rate. The flow rate of the 1st-D was set low at 0.02 mL/min to synchronize with the 2nd-D SEC separation. For the 2nd-D SEC separation, a high speed SEC column (PSS SDV linear M, 50 mm × 20 mm) was employed. Eluent was THF (Samchun, HPLC grade) at a flow rate of 4.0 mL/min. 2.6. MALDI-TOF MS analysis

Fig. 2. Schematic diagram of the 2D-LC apparatus used in this study.

For the MALDI-TOF MS experiments, a Bruker Reflex III mass spectrometer was used. The spectrometer is equipped with a nitrogen laser (λ = 337 nm), pulsed ion extraction, and a reflector. Polymer solutions were prepared in HPLC grade THF at a concentration of 5 mg/mL. The matrix, 1,8-dihydroxy-9(10H)anthracenone (dithranol, Aldrich, 97%), was dissolved in THF at a concentration of 20 mg/mL. A 5 ␮L aliquot of the polymer solution was mixed with 50 ␮L of the matrix solution and 1.5 ␮L of a silver trifluoroacetate (AgTFA, Aldrich, 98%) solution (1 mg/mL in THF), respectively. A 0.5 ␮L portion of the final solution was deposited onto a sample target plate and allowed to dry in air at room temperature.

K. Im et al. / J. Chromatogr. A 1103 (2006) 235–242

238

Fig. 3. SEC chromatograms of the PS precursor and the two branched PS prepared by linking PS precursor anions with CDMSS at different molar ratios of CDMSS to n-butyl Li (BS45: 0.49, BS65: 0.65). Separation condition: two mixed bed columns (Polymer Labs., PL mixed C), THF eluent at column temperature of 40 ◦ C.

3. Results and discussion Fig. 3 displays the SEC chromatograms of the PS precursor and the two branched PS, BS49 and BS65. The broad elution peak of the linear PS precursor as well as the non-overlapping chromatograms obtained with RI and light scattering detectors reflect the broad MWD and the Mw /Mn value was determined as 1.13. It is due to the slow initiation speed of n-butyl Li initiator [36]. Both branched PS show a unimodal elution peak. While the branched polymers prepared by the reaction with CDMSS elute earlier than PS precursor indicating that their molecular sizes are larger than the precursor PS, it is impossible to find more details on the molecular characteristics of the branched polymer from the SEC results alone. The characterization results of the precursor PS and the branched PS are listed in Table 1. Fig. 4 displays RP-TGIC chromatograms of the PS precursor and the branched PS. A small sharp peak appearing at tR ≈ 3 min in all three chromatograms is the injection solvent peak. Unlike the well-resolved RP-TGIC chromatograms for the branched polymers prepared with the PS precursor anions initiated by sec-butyl Li [17], the branched polymers are not well-resolved according to the number of branches. It is an expected result since the PS precursors initiated by n-butyl anions have a broad MWD as seen from its Mw /Mn value of 1.13 while those initiated by sec-butyl anion has a very narrow Poisson distribution in molecular weight [37,38]. Therefore, the overlap in the molecular weight distribution is severe for the branched PS initiated

Fig. 4. RP-TGIC chromatograms of the PS precursor, BS49 and BS65. ˚ 150 mm × 4.6 mm). Eluent: Column: C18 bonded silica (Kromasil, 100 A, CH2 Cl2 /CH3 CN (53/47, v/v) at a flow rate of 0.5 mL/min. Temperature program is shown in the plot.

by n-butyl Li, which results in poorly resolved RP-TGIC chromatograms. The MALDI-TOF mass spectra of BS49 and BS65 are displayed in Fig. 5. The mass spectra show that the branched PS exhibits very broad MWD. The intensity of the peaks decreases rapidly as the molecular weight increases. It is due to the fact that the mass spectrum reflects the number distribution of the polymers as well as the mass discrimination phenomenon in MALDITOF MS analysis [39]. The low molecular weight region below

Table 1 Characteristics of PS precursor and branched PSa Sample

CDMSS/n-butyl Li

Mw (kg/mol)

Mn (kg/mol)

Mw /Mn

Precursor BS49 BS65

– 0.49 0.65

3.9 8.5 13.6

3.5 5.9 8.9

1.13 1.44 1.53

a

Measured by SEC/MALLS.

Fig. 5. MALDI-TOF MS spectra of BS49 and BS65. In the insets, magnified spectra of BS49 and BS65 are shown (matrix:dithranol, salt: silver trifluoroacetate).

K. Im et al. / J. Chromatogr. A 1103 (2006) 235–242

1500 Da is cut off to show the high molecular weight region more clearly. A precise analysis of this kind of polymers with broad MWD is not yet possible with MALDI–MS analysis only. However, some useful information on the polymers can be extracted from the mass spectra. The magnified spectra of the same low molecular weight regions are shown in the inset. Both BS49 and BS65 show nearly identical spectra except for the relative intensity of the peaks. At least four different groups of the mass peaks are clearly identified, each of which repeats at a mass interval of styrene unit (104 Da). Among the four peaks labeled with molecular weight, the highest intensity peak at m/z = 2144.9 corresponds to the precursor PS terminated with a hydrogen: the calculated molecular weight is 104.15 × 19 (19 styrene units) + 57.12 (n-butyl initiator) + 107.87 (Ag+ ion) + 1.01 (terminal H) = 2144.8. The next peak of m/z = 2154.9 corresponds to the two-branch species whose molecular weight is consistent with 104.15 × 17 (17 styrene units) + 57.12 × 2 (two n-butyl initiators) + 107.87 (Ag+ ion) + 161.30 (CDMSS) + 1.01 (terminal H) = 2155.0. The third peak (the smallest among the four peaks) at m/z = 2164.9 is due to the three-branch species containing two CDMSS and three n-butyl moieties. The m/z = 2200.8 of the last peak is consistent with the molecular weight of the PS precursor with DP = 18 terminated with CDMSS. Similarly, overlapped peaks of differently branched PS extend over wide molecular weight range. Since, the precursor PS anions have a broad MWD, branched PS with different number of branches can have similar molecular weight. It explains why the resolution of the RP-TGIC chromatogram of the branched PS prepared by using n-butyl Li initiator is poorer than the branched PS prepared by using sec-butyl Li initiator. High-resolution of the RP-TGIC separation with respect to molecular weight alone is not good enough to resolve the branched PS prepared by using n-butyl Li initiator according to the number of branches due to the broad MWD of the precursor PS anions. In the previous work with the similarly branched PS prepared by using sec-butyl Li initiator, RP-TGIC separation with respect to the molecular weight was able to separate the different structures according to the number of branches since the molecular weight is practically proportional to the branch numbers due to the narrow MWD of the branch chains [17]. For a type of polymers with bivariate distribution in both molecular weight and branch number, however, we need to find an LC method to separate the branched polymers in terms of the number of branches also. In the previous work, we found that it was possible to resolve branched PS in terms of the branch number by LCCC utilizing the fact that each branch unit contains additional functionality [17]. The CDMSS branching unit and the sec-butyl initiator are the additional groups to the repeating styrene units, of which the number is proportional to the number of branches except for the polymers terminated with CDMSS. The amount of CDMSS terminated species increases as the added amount of CDMSS relative to the living PS anions increases as can be seen in Fig. 5. The CDMSS terminated species has one more CDMSS moiety than hydrogen terminated ones. However, the amount of CDMSS terminated species is small in these samples and we do not take them into account in the forthcoming discussion.

239

Fig. 6. (A) RPLC chromatograms of six PS standards (Mw : 2.5, 7.0, 15.6, 30.9, 62, 113 kg/mol) at different temperatures to establish the critical condition. (B) LCCC chromatogram of PS precursor, BS49 and BS65 at the critical condition of ˚ 150 mm × 4.6 mm). PS (53.3 ◦ C). Column: C18 bonded silica (Kromasil, 100 A, Eluent: CH2 Cl2 /CH3 CN = 53/47 (v/v) at a flow rate of 0.5 mL/min.

Therefore, the LCCC separation according to the number of branches was realized at the chromatographic critical condition of homo-PS. Although the branched PS in this work has a broad MWD, it would be still possible to resolve the branched PS if the contribution of PS molecular weight is successfully suppressed under the critical condition. Fig. 6 displays the results of our search for the critical condition of homo-PS (A) and the LCCC chromatograms of the branched PS at the critical condition (B). Fig. 6A shows the RPLC chromatograms of a mixture of six PS standards at different temperatures near the critical condition. The critical condition is established at the column temperature of ∼53 ◦ C at which all PS standards are co-eluted. Fig. 6B displays the LCCC chromatograms of PS precursor, BS49 and BS65 taken at 53.3 ◦ C. It is easily noticeable that the elution peak of the PS precursor becomes very sharp at the critical condition despite its broad MWD since the separation according to the PS molecular weight is suppressed. The branched polymers are also resolved reasonably well with respect to the number of branches as expected. In 2D-LC analysis of complex polymers with bivariate distribution, SEC has been a typical choice to characterize MWD although SEC does not separate the polymer according to the molecular weight but molecular size. This limitation of SEC becomes less serious in the 2D-LC analysis since the injection sample to the 2nd-D SEC is homogeneous in the molecular characteristic other than MWD after the 1st-D fractionation.

240

K. Im et al. / J. Chromatogr. A 1103 (2006) 235–242

Fig. 7. Contour plots of LCCC × SEC 2D-LC chromatograms of BS49 (A) ˚ and BS65 (B). 1st-D LCCC: C18 bonded silica column (Kromasil, 100 A, 150 mm × 4.6 mm); CH2 Cl2 /CH3 CN = 53/47 (v/v) eluent at a flow rate of 0.02 mL/min. 2nd-D SEC: PSS SDV column (5 ␮m, 50 mm × 20 mm), THF eluent at a flow rate of 4.0 mL/min. The numbers indicate the number of branches.

Therefore, the 2D-LC coupling of LCCC with SEC has been employed frequently for the systems having bivariate distribution in functionality type as well as in molecular weight. In Fig. 7, the contour plots of LCCC × SEC 2D-LC chromatograms of BS49 (A) and BS65 (B) are shown. The 1st-D separation was carried out by LCCC to separate the branched PS according to the branch number and the effluent from the LCCC separation was subjected to the 2nd-D SEC separation. The differently branched PS are resolved as separate peaks and the contour plot shows that the molecular size increases (SEC retention decreases) as the number of branch increases (LCCC retention increases). However, the resolution is not good mainly because SEC separation of the branched polymers is not efficient as already shown in Fig. 3. To separate the branched PS with a better resolution according to both molecular weight and number of branches, we combined RP-TGIC (separating the branched PS mainly by molecular weight) and LCCC (separating the branched PS mainly by the number of branches). Since, the LCCC separation can be repeated more rapidly than RP-TGIC by virtue of its isother-

Fig. 8. Contour plots of RP-TGIC × LCCC 2D-LC chromatograms of BS49 (A) and BS65 (B). 1st-D RP-TGIC: C18 bonded silica column (Kromasil, ˚ 150 mm × 4.6 mm); CH2 Cl2 /CH3 CN = 53/47 (v/v) eluent at a flow rate 100 A, ˚ of 0.02 mL/min. 2nd-D LCCC: C18 bonded silica column (Kromasil, 100 A, 150 mm × 4.6 mm); CH2 Cl2 /CH3 CN = 53/47 (v/v) eluent at a flow rate of 1.2 mL/min. The numbers indicate the number of branches.

mal elution, RP-TGIC was employed for the 1st-D separation. Although the RP-TGIC can be practiced fast for individual run [40], the cooling time of the column is too long to be effective for the 2nd-D LC separation. Fig. 8 displays the contour plots of the RP-TGIC × LCCC 2D-LC separation of BS 49 (A) and BS 65 (B). Comparing with the LCCC × SEC 2D-LC chromatograms shown in Fig. 7, the apparent resolution of RP-TGIC × LCCC configuration is better than LCCC × SEC. Fig. 9 shows the reconstructed 1st-D RP-TGIC chromatogram (A) and 2nd-D LCCC chromatogram (B) for BS49 from Fig. 8A. They were obtained by integrating the 2DLC chromatogram intensity along the corresponding separation axes. The 1st-D RP-TGIC chromatogram (A) does not match well with the RP-TGIC chromatogram in Fig. 4 since TGIC retention depends on the flow rate and the temperature program that are different in the two TGIC separation [40]. On the other hand, the isothermal LCCC chromatogram in Fig. 9B matches well with the LCCC chromatogram shown in Fig. 6B despite the

K. Im et al. / J. Chromatogr. A 1103 (2006) 235–242

Fig. 9. (A) Reconstructed 1st-D RP-TGIC chromatogram and (B) reconstructed 2nd-D LCCC chromatogram of BS49 from Fig. 8A. They were obtained by integrating the 2D-LC chromatogram intensity along the corresponding separation axes.

different flow rate in the two separations, which indicates that the comprehensive 2D-LC separation was carried out successfully. An additional advantage of the TGIC × LCCC combination over other 2D-LC combinations is in the use of common eluent for the two LC separations. The 2D-LC using the same solvent for both dimensions is completely free from the prob-

241

lems often encountered in 2D-LC such as the “breakthrough” or “solvent-plug” effect at the 2nd-D separation and the possible solvent incompatibility between the two separation modes [41]. Furthermore, the intensity of injection solvent peak is greatly reduced in the 2nd-D LC separation. As can be seen in Fig. 7, the injection solvent peak appears as a deep trench near VR ≈ 13–14 mL. The magnitude of the solvent peak can be seen in Fig. 10, in which the 2nd-LC chromatograms of LCCC × SEC (A) and RP-TGIC × LCCC (B) 2D-LC separation of BS49 are displayed. Both 2nd-LC chromatograms were taken near the maximum intensity position of the two-branch species eluting at VR ≈ 2.2 mL and VR ≈ 9 mL in the 1st-D LCCC (Fig. 7A) and the 1st-D RP-TGIC (Fig. 8A) separations, respectively. The solvent peak in the 2nd-D SEC chromatogram is intense since the 1st-D LCCC effluent is the solution in the CH2 Cl2 /CH3 CN mixture while the solvent peak in the 2nd-D LCCC appearing at VR ≈ 1.8 mL is hardly visible since the 1st-D RP-TGIC effluent is in the same solvent as eluent of the 2nd-D LCCC separation. Acknowledgment This study was supported by Korea Research Foundation (Basic Research Program, R02-2004-000-10115-0 and the BK21 program). References

Fig. 10. 2nd-LC chromatograms of the LCCC × SEC (A) and the RPTGIC × LCCC (B) 2D-LC separation of BS49 showing the magnitude of the injection solvent peak (marked with arrows). Both 2nd-LC chromatograms were taken near the maximum intensity position of the two-branch species at VR ≈ 2.2 mL (A) and VR ≈ 9 mL (B) of the 1st-D LCCC (Fig. 7A) and the 1st-D RP-TGIC separation (Fig. 8A), respectively.

[1] W.W. Yau, J.J. Kirkland, D.D. Bly, Modern Size-Exclusion Liquid Chromatography, Practice of Gel Permeation and Gel Filtration Chromatograph, Wiley, New York, 1979. [2] G. Gl¨ockner, Gradient HPLC of Copolymers and Chromatographic Cross-Fractionation, Springer, Berlin, 1992. [3] B. Trathnigg, Prog. Polym. Sci. 20 (1995) 615. [4] H. Pasch, B. Trathnigg, HPLC of Polymers, Springer, Berlin, 1997. [5] D. Berek, Prog. Polym. Sci. 25 (2000) 873. [6] T. Chang, Adv. Polym. Sci. 163 (2003) 1. [7] T. Macko, D. Hunkeler, Adv. Polym. Sci. 163 (2003) 61. [8] T. Chang, J. Polym. Sci. Polym. Phys. Ed. 43 (2005) 1591. [9] S. Podzimek, Am. Lab. 34 (2002) 38. [10] C. Degoulet, T. Nicolai, D. Durand, J.P. Busnel, Macromolecules 28 (1995) 6819. [11] B.H. Zimm, W.H. Stockmayer, J. Chem. Phys. 17 (1949) 1301. [12] H.C. Lee, T. Chang, S. Harville, J.W. Mays, Macromolecules 31 (1998) 690. [13] H.C. Lee, W. Lee, T. Chang, J.S. Yoon, D.J. Frater, J.W. Mays, Macromolecules 31 (1998) 4114. [14] S. Perny, J. Allgaier, D. Cho, W. Lee, T. Chang, Macromolecules 34 (2001) 5408. [15] D. Cho, S. Park, T. Chang, A. Avgeropoulos, N. Hadjichristidis, Eur. Polym. J. 39 (2003) 2155. [16] S. Park, D. Cho, K. Im, T. Chang, D. Uhrig, J.W. Mays, Macromolecules 36 (2003) 5834. [17] K. Im, S. Park, D. Cho, T. Chang, K. Lee, N. Choi, Anal. Chem. 76 (2004) 2638. [18] C.J. Venkatramani, J. Xu, J.B. Phillips, Anal. Chem. 68 (1996) 1486. [19] F. Erni, R.W. Frei, J. Chromatogr. 149 (1978) 561. [20] M.M. Bushey, J.W. Jorgenson, Anal. Chem. 62 (1990) 161. [21] P.K. Kilz, H. Much, G. Schulz, Hyphenated and Multidimensional Techniques (Advances in Chemistry Series, vol. 247), American Chemical Society, Washington, DC, 1995 (Chapter 17). [22] R.E. Murphy, M.R. Schure, J.P. Foley, Anal. Chem. 70 (1998) 4353. [23] R.E. Murphy, M.R. Schure, J.P. Foley, Anal. Chem. 70 (1998) 1585. [24] J. Adrian, D. Braun, H. Pasch, LC–GC Int. 11 (1998) 32.

242

K. Im et al. / J. Chromatogr. A 1103 (2006) 235–242

[25] J. Adrian, E. Esser, G. Hellmann, H. Pasch, Polymer 41 (2000) 2439. [26] M. Janco, T. Hirano, T. Kitayama, K. Hatada, D. Berek, Macromolecules 33 (2000) 1710. [27] D. Berek, Macromol. Symp. 174 (2001) 413. [28] B. Trathnigg, C. Rappel, R. Raml, A. Gorbunov, J. Chromatogr. A 953 (2002) 89. [29] A.P. Sweeney, R.A. Shalliker, J. Chromatogr. A 968 (2002) 41. [30] A. van der Horst, P.J. Schoenmakers, J. Chromatogr. A 1000 (2003) 693. [31] X. Jiang, A. van der Horst, V. Lima, P.J. Schoenmakers, J. Chromatogr. A 1076 (2005) 51. [32] J. Gerber, W. Radke, e-Polymers (2005), No. 045. [33] H. Pasch, P. Kilz, Macromol. Rapid Commun. 24 (2003) 104.

[34] D.M. Knauss, H.A. Al-Muallem, T. Huang, D.T. Wu, Macromolecules 33 (2000) 3557. [35] Y. Kawakami, Y. Miki, T. Tsuda, R.A.N. Muthy, Y. Yamashita, Polym. J. 14 (1982) 913. [36] H.L. Hsieh, R.P. Quirk, Anionic Polymerization; Principles and Practical Applications, Marcel Dekker, New York, 1996. [37] W. Lee, H. Lee, J. Cha, T. Chang, K.J. Hanley, T.P. Lodge, Macromolecules 33 (2000) 5111. [38] J. Ryu, K. Im, W. Yu, J. Park, T. Chang, K. Lee, N. Choi, Macromolecules 37 (2004) 8805. [39] M.S. Montaudo, Mass Spectrom. Rev. 21 (2002) 108. [40] J. Ryu, S. Park, T. Chang, J. Chromatogr. A 1075 (2005) 145. [41] X. Jiang, A. van der Horst, P.J. Schoenmakers, J. Chromatogr. A 982 (2002) 55.