ARTICLE pubs.acs.org/ac
2D-LC Characterization of Comb-Shaped Polymers Using Isotope Effect Seonyoung Ahn,† Kyuhyun Im,† Taihyun Chang,*,† Pierre Chambon,‡ and Christine M. Fernyhough‡ †
Department of Chemistry and Division of Advanced Materials Science, Pohang University of Science and Technology (POSTECH), Pohang, 790-784, Republic of Korea ‡ Department of Chemistry, University of Sheffield, Sheffield, S3 7HF, United Kingdom ABSTRACT: A rigorous molecular characterization of combshaped polystyrene (PS) was carried out taking advantage of its molecular structure, a normal hydrogenous backbone, and deuterated side chains. Normal phase LC (NPLC) can separate the comb PS species well according to their molecular weight. Nonetheless, it cannot distinguish the backbone from the side chains and the differently structured polymers having a similar molecular weight, e.g, a single backbone comb and a coupled backbone comb with fewer side chains. In contrast to NPLC, the hydrogenous polymer is retained longer than the deuterated counterpart in reversed phase LC (RPLC). When the isotope sensitivity of RPLC is taken advantage of, the comb PS is cross fractionated by NPLC and RPLC, and a two-dimensional mapping with respect to the backbone chain length and the number of branches is fully established.
C
hain branching in macromolecules influences the rheological and mechanical properties of polymeric materials significantly, and model branched polymers (e.g., star-shaped or comb-shaped polymers) have been studied extensively to understand their single chain properties and rheological behavior.1,2 Most model branched polymers are prepared by anionic polymerization because it remains the best available polymerization method to produce a narrow molecular weight distribution (MWD) and can be used to introduce functionality for subsequent reactions such as introduction of branches. Despite the use of the best synthetic method available and the time-consuming post-fractionation, it is difficult to obtain structurally well-defined polymers of high purity. Furthermore, the purity of such branched polymers has not been confirmed unambiguously due to the lack of precise characterization methods. The most widely adopted method to characterize branched polymers is based on size exclusion chromatography (SEC), which separates the polymer molecules by an entropy-driven exclusion equilibrium. However, the separation based on chain volume in solution cannot separate a mixture of branched polymers having different molecular weights (MW) and chain architectures effectively.3,4 An alternative fractionation method for macromolecules is interaction chromatography (IC) which has been used increasingly in recent years. Unlike SEC, IC can separate polymers according to MW (not chain size) for homopolymers. In addition, IC can separate copolymers according to their chemical composition since it utilizes enthalpic interaction of polymers with the stationary phase.3,5,6 Furthermore, the resolution of IC is much higher than SEC due to lower band broadening of IC.711 In this study, we report on a rigorous characterization of a comb-shaped polystyrene (PS), where deuterated PS (d-PS) r 2011 American Chemical Society
branches are attached to a hydrogenous PS (h-PS) backbone. Such comb-shaped polymers resemble the structure of branched polymers with 3-way branching points in commercially important polymers and have been used extensively as model polymers having long chain branches. The comb-shaped polymer was prepared by grafting d-PS anions to a chloromethylated h-PS backbone.12 While both branch and backbone polymers are prepared by anionic polymerization and have narrow MWD, chloromethylation of the backbone PS has to be statistical and the number of branches is not uniform. In addition, side reactions occurring during the anionic polymerization and postreactions yield various byproducts. It is not possible to characterize this complex polymer precisely by conventional SEC based analyses. In this study, two-dimensional liquid chromatography (2D-LC) was employed for rigorous characterization: one LC is more sensitive to MW and the other LC is more sensitive to the deuterium content. The LC separation according to the isotope content is relatively new in polymer characterization, but it has been utilized for the separation of small molecules for a long time.1315 When these two IC methods were combined, it was possible to map the distribution of a complex mixture of combshaped polymers with unprecedented precision.
’ EXPERIMENTAL SECTION Materials. The comb-shaped PS was prepared by grafting d-PS anions to a partially chloromethylated h-PS backbone. Both backbone and branch PS were prepared separately by anionic Received: March 6, 2011 Accepted: April 21, 2011 Published: April 21, 2011 4237
dx.doi.org/10.1021/ac2005907 | Anal. Chem. 2011, 83, 4237–4242
Analytical Chemistry
ARTICLE
Table 1. Molecular Weight Characteristics of the CombShaped PS and Its Precursorsa backbone
branch
Mnb
b
Mw/Mn
Mn
190 000
1.05
83 000
comb b
Mw/Mn
Mn
1.02
540 000
Mw/Mn average nbranch 1.1
4.2
a
The nbranch is estimated from the number averaged molecular weight determined by SEC measurements. b Number average molecular weight determined by SEC/LC measurements.
polymerization using high vacuum techniques. The chloromethylation of h-PS backbone was carried out by the reaction of h-PS with in situ prepared chloromethyl methyl ether. Details of the synthetic procedure were reported previously.12 The comb and precursor PS samples were characterized by SEC/light scattering detection as summarized in Table 1. SEC Analysis. Two mixed bed columns (Polymer Lab, PL gel mixed C, 300 7.8 mm i.d.) were used at a column temperature of 40 °C. SEC chromatograms were recorded with a triple detector (light scattering (LS), refractive index (RI), and viscosity) system (Viscotek TDA 300) and a UV absorption detector (TSP, UV100). The solvent was tetrahydrofuran (THF; Samchun, HPLC grade) at a flow rate of 0.8 mL/min. Polymer samples for the SEC analysis were dissolved in THF at a concentration of ∼0.5 mg/mL, and the injection volume was 100 μL. HPLC Analysis. For the normal phase (NP) temperature gradient interaction chromatography (TGIC) analysis, a bare silica column (Nucleosil, 5 μm, 500 Å pore, 250 4.6 mm) and a mixed eluent (52/48, v/v) of iso-octane (J.T. Baker, HPLC grade) and THF (Samchun, HPLC grade) was used (0.5 mL/ min). For the reversed phase (RP)-TGIC analysis, a C18 bonded silica column (Nueleosil C18, 7 μm, 500 Å pore, 150 4.6 mm) was used. The mobile phase was a mixture of CH2Cl2/CH3CN (57/43, v/v, Samchun, HPLC grade) at a flow rate of 0.5 mL/min. The temperature of the column was controlled by circulating fluid from a programmable bath/circulator (ThermoHaake, C25P) through a homemade column jacket. The sample solutions (∼5 mg/mL) were prepared by dissolving the polymers in a small volume of the corresponding eluent, and the injection volume was 100 μL. The chromatograms were recorded with a LS detector (Wyatt MiniDawn), a RI detector (Shodex RI-101), and a UV absorption detector (TSP, UV 2000) operating at a wavelength of 260 nm for online determination of the absolute MW of polymers. 2D-LC Analysis. For the first-D NP-TGIC analysis, the same condition as for the 1D-NP-TGIC analysis was used except for the flow rate and column dimension. The flow rate of the first-D NP-TGIC was set at 0.05 mL/min in order to synchronize with the second-D reversed phase LC (RPLC) separation, and a smaller size column (50 4.6 mm) was used to reduce the void volume sweeping time and, thus, total analysis time. The reduction of the column size did not result in a significant deterioration of the peak resolution. In the second-D solvent gradient RPLC separation, a C18 bonded silica column (Nucleosil C18, 7 μm, 500 Å pore, 150 4.6 mm) was used. For the solvent gradient run, two mixtures of CH2Cl2/CH3CN (Samchun, HPLC grade) differing in their composition (A: 57/43, B: 59/41 in v/v) were used at a flow rate of 1.5 mL. The solvent composition started and was maintained at 100% A for 2 min after the sample injection. The solvent composition was then changed linearly
Figure 1. Schematic diagram of the comprehensive NP-TGIC RPLC 2D-LC experiment setup used in this study.
from 100% A to 100% B over 4 min followed by a quick change back to 100% A and maintained at 100% A for 2 min to flush out the column of which the void volume is ∼2.4 mL. Therefore, it took 8 min for a second-D RPLC run while the effective separation time is 6 min. For the NP-TGIC RPLC 2D-LC analysis, three HPLC pumps (Pump 1: Waters 515 HPLC pump, Pump 2: Bischoff compact pump 2250, and Pump 3: Spectra System P4000) and two UV detectors (TSP, UV100) were used. The two LC systems were connected via two electronically controlled 10-port 2-position valves (Alltech, SelectPro) equipped with two 100 μL storage loops and with two short diol bonded silica trapping columns (Nucleosil diol, 7 μm, 100 Å, 50 4.6 mm). The experimenta1 setup is schematically shown in Figure 1. It is similar to the one reported previously16 but at a different trapping condition and with a different elution mode for the second-LC separation. To trap the polymers from the first-D LC effluent, CH3CN with 4% iso-octane was used. The mixed solvent did not cause a break-through problem in the second-LC separation; furthermore, full desorption took place by the second-D LC eluent.
’ RESULTS AND DISCUSSION SEC Separation of Comb-Shaped Polymers. The combshaped polymer and its precursors (h-PS backbone and d-PS branch) were first characterized by SEC as displayed in Figure 2. The SEC elution is not affected significantly by the isotope content of polymers. The MW of the polymer species for each peak can be estimated from the relative intensity of the concentration detector (RI) signal (Δn) to the light scattering (LS) detector (R90) signal (LS signal is proportional to the product of concentration and MW). In the SEC chromatograms of both branch and backbone PS, a small amount of coupled products (labeled with solid arrows showing 2 MW) is apparent eluting earlier than the narrow peak of the main products. Both polystyryllithium backbone and the branches can undergo coupling reactions in the event of exposure to air, hence the use of high vacuum and careful experimental techniques in order to keep this to a minimum.17 The coupled branches of d-PS are incapable of further reactions since they have lost their anionic reactivity and are left as a side product together with the unreacted branches 4238
dx.doi.org/10.1021/ac2005907 |Anal. Chem. 2011, 83, 4237–4242
Analytical Chemistry
Figure 2. SEC chromatograms of the two PS precursors (backbone and branch) and the PS comb recorded by a differential refractive index (solid line) and light scattering (dashed line) detectors. Separation condition: two mixed bed columns (Polymer Laboratories., PL mixed C), THF eluent at a column temperature of 40 °C.
of d-PS. In addition, the coupled h-PS backbone can, however, participate in the subsequent chloromethylation and grafting reactions. Furthermore, the chloromethylation can also crosslink the h-PS backbone to yield more highly coupled backbone h-PS to some extent. The fronting of the coupled backbone PS peak at tE ≈ 13 min (marked with dashed arrow) revealed a possibility of a 3 backbone product, but it is hard to confirm unambiguously due to the low intensity and the overlap with the 2 MW peak. The PS comb contains a trace amount of the PS precursors resolved as separate peaks, but the major product of PS comb shows a unimodal elution peak. It is apparent that the SEC separation cannot resolve the polymer species having coupled backbone but yields a broad unimodal peak due to the nonuniform number of branches. From the SEC results alone, it is difficult to obtain more details beyond the average branch numbers determined by the average MW. Since the high MW species formed from the coupled backbone can affect the rheological properties of the branched polymer significantly, it is important to resolve the comb polymer species containing a coupled backbone. TGIC Separation of Comb-Shaped Polymers. Figure 3a shows an NP-TGIC chromatogram of the comb PS. It separates the comb-shaped polymers according to MW; i.e., species with different number of branches are resolved. In the previous study, the number of branches in the comb PS was found to follow the Poisson distribution reflecting that the chloromethylation reaction proceeded randomly.12 The peak MW increases as integral multiples of branch MW, but the peaks start to overlap each other as the number of branches increases due to the finite MWD of the branch and backbone. The contribution of coupled backbone species also needs to be taken into account since a single backbone comb and coupled backbone comb with a fewer number of branches can have similar MW. Figure 3b displays the RP-TGIC chromatograms of the PS comb recorded by a LS and a UV absorption detector. Unlike the well-resolved NP-TGIC chromatogram, RP-TGIC shows only three peaks. It was surprising to see a totally different RPLC
ARTICLE
Figure 3. (a) NP-TGIC and (b) RP-TGIC chromatograms of PS comb recorded by UV (solid line) and LS (dashed line) detectors. MW determined by LS detection is also plotted. The column temperature is shown in the top abscissa. Condition (a) Nucleosil, bare silica, 500 Å pore, 5 μm particle, 250 4.6 mm i.d.; iso-octane/THF (52/48, v/v) at a flow rate of 0.5 mL/min. (b) Nucleosil, C18, 500 Å pore, 7 μm particle, 150 4.6 mm i.d.; CH2Cl2/CH3CN (57/43, v/v) at a flow rate of 0.5 mL/min.
chromatogram from that of normal phase LC (NPLC) since both RPLC and NPLC are expected to yield similar separation according to MW for a homopolymer unless the polymer contains certain functionalities that interact strongly with one of the stationary phases.18 Furthermore, MW determined from the relative signal intensity of LS and UV detector signals reveals that the RPLC retention is not dictated by polymer MW only. In general, IC retention increases with MW since the interaction strength of a polymer molecule with the stationary phase increases with MW as in the NPLC separation shown in Figure 3a.19 The MW dependence of the RPLC retention, however, shows an unusual curvature. The peculiar phenomenon was found to be due to the chromatographic selectivity of isotopes in RPLC.2022 d-PS interacts with the RP stationary phase less strongly than h-PS, and the unreacted branches elute early at tE ≈ 4 min. The comb polymers containing a single backbone elute later at tE ≈ 611 min as the main peak in which the grafted d-PS branches seem to reduce the retention time of the comb-shaped polymer as the number of grafted d-PS branches (thus MW) increases. After the main peak, MW increases abruptly toward the last peak eluting at tE ≈ 1218 min. The MW then decreases again with increasing elution time in a similar way to the main peak. The last peak appears to be the comb species with coupled backbone. Online 2D-LC Separation of Comb-Shaped Polymers. To characterize the comb-shaped PS more rigorously, we carried out 2D-LC separation by combining NP-TGIC (separating the PS comb mainly by MW regardless of the isotope content) and the RPLC (separating the comb PS mainly by the isotope composition). In a 2D-LC separation, two independent LC separations are combined and several points need to be considered for the optimization of the analysis.2325 First of all, the second-D LC separation needs to be executed at a high repeating rate to complete the analysis in reasonable time. The 2D-LC chromatogram is essentially an overlay of multiple second-D LC chromatograms, and the 2D-LC analysis time is the product of the second-D LC analysis time and the number of the second-D LC chromatograms taken. Since the NP-TGIC separation of the comb PS exhibits better resolved peaks than the RP-TGIC separation as shown in Figure 3, it is desirable not to lose the 4239
dx.doi.org/10.1021/ac2005907 |Anal. Chem. 2011, 83, 4237–4242
Analytical Chemistry
Figure 4. RP-TGIC chromatograms of comb PS dissolved in CH2Cl2/ CH3CN (solid line) and THF/iso-octane (dashed line) mixture, respectively. C18 column (Nucleosil, 500 Å, 7 μm particle, 150 4.6 mm); Eluent: CH2Cl2/CH3CN (57/43, v/v) at a flow rate of 0.5 mL/min.
high resolution in the NP-TGIC separation. Since a fast secondD separation has to be done at the cost of resolution to some extent, NP-TGIC was employed for the first-D separation. For the second-D RPLC separation, the solvent gradient elution method was employed instead of the temperature gradient elution since the column temperature cannot be changed as rapidly as the solvent composition. Second, it is necessary to consider the solvent compatibility of the first-D LC with the second-D LC since the eluent of the firstD LC separation becomes the injection solvent for the second-D LC separation. An injection solvent different from the eluent may cause a variety of artifacts in LC analyses. The effluent of the firstD LC is a mixture of iso-octane/THF (52/48, v/v), which is a strong solvent for the second-D RPLC. If PS comb dissolved in iso-octane/THF (52/48 v/v) solvent is injected into the RPLC system, we cannot obtain the desired results. Figure 4 shows the RP-TGIC chromatograms of the PS comb dissolved in two different injection solvent mixtures; iso-octane/THF (52/48 v/v) and CH2Cl2/CH3CN (57/43 v/v). While the comb PS dissolved in the CH2Cl2/CH3CN mixture shows a similar elution behavior to a normal RP-TGIC separation, a large portion of the polymer sample elutes together with the injection solvent when iso-octane/ THF (52/48 v/v) is used as the injection solvent. This is known as a “solvent plug” or “break through” effect, a type of artifact frequently observed in comprehensive 2D-LC analyses when the solvent strength of the injection solvent is stronger than that of the eluent.16,2628 To solve the problem, we used a trapping system similar to that described by Im et al. as shown in Figure 1.16 The trapping system consists of two switching valves and two trap columns. Two switching valves enable continuous and alternate sampling of the first-D effluent and injection to the second-D LC column through two trap columns.2932 The function of the trap column is to fully adsorb the polymer sample from the first-D LC effluent using a solvent which is compatible with the second-D LC. To examine the sorption/desorption efficiency, the PS sample dissolved in CH3CN with 4% iso-octane was injected via the trapping column to the RPLC. Figure 5 displays the RPLC chromatograms for single injection (dashed line) and 4 injection (solid line). In the 4 injections, 100 μL samples were transferred four times to the trapping column every 2 min. The intensity of a 4 injection chromatogram is precisely 4 times as high as that of the single injection (dotted line) which confirms
ARTICLE
Figure 5. RPLC chromatograms of PS comb when the sample is injected via a trap column: Single injection (dashed line) and 4 injection (solid line). The dotted line is the intensity of the single injection chromatogram multiplied by four to compare with the multiple injection chromatogram. Column: C18 column (Nucleosil, 500 Å, 7 μm particle, 150 4.6 mm); Gradient: CH2Cl2/CH3CN (57/43, v/v) to CH2Cl2/CH3CN (59/41, v/v) in 6 min at a flow rate of 1.5 mL/min.
Figure 6. Contour plots of NP-TGIC RPLC 2D-LC chromatograms of comb PS. first-D NP-TGIC: Bare silica column (Nucleosil, 500 Å, 5 μm particle, 50 4.6 mm); Eluent: iso-octane/THF (52/48, v/v) at a flow rate of 0.5 mL/min. second-D RPLC: C18 column (Nucleosil, 500 Å, 7 μm particle, 150 4.6 mm); Gradient: CH2Cl2/CH3CN (57/43, v/v) to CH2Cl2/CH3CN (59/41, v/v) in 6 min at a flow rate of 1.5 mL/min.
good sample recovery and full sorption/desorption. The analysis time of second-D LC for separation and restabilization of the column is 8 min. Therefore, during a second-D RPLC separation which takes 8 min, the first-D NP-TGIC separation elutes 400 μL at the flow rate of 0.05 mL/min. The storage of the first-D NPTGIC effluent in the 100 μL loop and the transfer to the trapping column were conducted four times during a second-D RPLC separation. In this manner, all the effluent from the first-D separation was analyzed by the second-D separation. Figure 6 displays the NP-TGIC RPLC 2D-LC chromatogram of the comb-shaped PS. In this two-dimensional mapping, a lot more details of the comb-shaped PS are revealed. Unreacted d-PS branches, eluting as a sharp peak in the RPLC separation shown in Figure 3b, is further resolved by NP-TGIC into two peaks corresponding to single (1) and coupled branches (10 ). The faint peak (2) is the single backbone without any branches. The major peak group (3) corresponds to the comb-shaped PS having different numbers of branches. As the number of branches increases, NP-TGIC retention increases while RPLC retention decreases as already shown in the 1D separation (Figure 3). 4240
dx.doi.org/10.1021/ac2005907 |Anal. Chem. 2011, 83, 4237–4242
Analytical Chemistry
ARTICLE
Enhancement of the precision in their molecular characterization would render such partially deuterated polymers more valuable. Not only deuteration but also any type of functionality with discernible interaction contrast from the rest of the polymer chains can also play a similar role for their precise molecular characterization.
’ AUTHOR INFORMATION Corresponding Author Figure 7. Distribution of branches in the comb-shaped PS.
The peak group (4) corresponds to the comb-shaped PS with a coupled backbone. Since MW of the h-PS portion (backbone) is large, it is retained longer in RPLC while its NPLC retention is overlapped with the single backbone combs having similar MW (i.e., a larger number of branches) and not resolved in 1D NPTGIC separation. The coupled backbone combs show the same trend as the single backbone combs, i.e., NP-TGIC retention increases while RPLC retention decreases as the number of branches increases. Another feature to note is the faint peak group (5) eluting after (4) in the RPLC separation. It is apparent that it can be resolved neither from the other comb groups by 1D-NPLC separation nor from the coupled backbone combs in the 1D-RPLC separation. Nonetheless, it is clearly resolved by virtue of the 2D-LC separation. Judging from its high NP-TGIC retention and the resolved subpeaks in the group, it is identified as the triply coupled backbone species. The triple backbone species can be formed as a side reaction product of chloromethylation. The presence of the triply coupled backbone can be recognized from the fronting of the coupled product peak (dashed arrow in Figure 2), but it is practically infeasible to identify it by SEC alone. Branch Number Distribution of Comb-Shaped Polymers. The branch number distribution of the comb-shaped polymer could be obtained from the intensity of the 2D-LC chromatogram, and it is displayed in Figure 7. The distribution shown in Figure 7 could be obtained by integrating the intensity of 2D-LC chromatogram (Figure 6), but in order to obtain the branching distribution more precisely, off-line 2D-LC analysis was performed. The comb-shaped PS was fractionated by RPLC to separate polymers containing different backbone species and subsequently subjected to NP-TGIC separation to obtain the branch distribution. The detailed procedure of NP-TGIC separation and branch analysis is identical to the method reported earlier.12 The analysis of the polymer species of triple backbone was not attempted since the amount is too small for a reliable analysis. In summary, we characterized a model comb-shaped polymer in detail using 2D-LC separation. The combination of high resolution with respect to MW in NPLC and the sensitivity to isotope content in RPLC allowed a successful characterization of the PS comb including all side products formed during the synthesis. This work demonstrated the power of 2D-LC in the characterization of complex polymers. Deuterated polymers might seem to be an exotic case at a glance, but deuterated polymers are used very widely in polymer research due to the large neutron cross section of deuterium. In fact, this polymer was not prepared for the molecular characterization purpose but for rheological studies combined with neutron scattering.
*Tel: þ82-54-279-2109. Fax: þ82-54-279-3399. E-mail:
[email protected].
’ ACKNOWLEDGMENT T.C. acknowledges the support from NRF via NRL (R0A2007-000-20125-0), SRC (R11-2008-052-03002), and WCU (R31-2008-000-10059-0) programs. C.M.F acknowledges the EPSRC for funding via the Microscale Polymer Processing 2 grant. ’ REFERENCES (1) Graessley, W. W. Polymeric Liquids and Networks: Dynamics and Rheology; Garland Science: New York, 2008. (2) McLeish, T. C. B. Adv. Phys. 2002, 51, 1379–1527. (3) Chang, T. Adv. Polym. Sci. 2003, 163, 1–60. (4) Mori, S.; Barth, H. G. Size Exclusion Chromatography; SpringerVerlag: New York, 1999. (5) Pasch, H.; Trathnigg, B. HPLC of Polymers; Springer-Verlag: Berlin, 1997. (6) Ryu, J.; Chang, T. Anal. Chem. 2005, 77, 6347–6352. (7) Lee, W.; Lee, H.; Cha, J.; Chang, T.; Hanley, K. J.; Lodge, T. P. Macromolecules 2000, 33, 5111–5115. (8) Park, S.; Cho, D.; Ryu, J.; Kwon, K.; Lee, W.; Chang, T. Macromolecules 2002, 35, 5974–5979. (9) Im, K.; Park, S.; Cho, D.; Chang, T.; Lee, K.; Choi, N. Anal. Chem. 2004, 76, 2638–2642. (10) Park, S.; Park, I.; Chang, T.; Ryu, C. Y. J. Am. Chem. Soc. 2004, 126, 8906–8907. (11) Park, H. W.; Jung, J.; Chang, T. Macromol. Res. 2009, 17, 365–377. (12) Chambon, P.; Fernyhough, C. M.; Im, K.; Chang, T.; Das, C.; Embery, J.; McLeish, T. C. B.; Read, D. J. Macromolecules 2008, 41, 5869–5875. (13) Falconer, W. E.; Cvetanovic, R. J. Anal. Chem. 1962, 34, 1064–1066. (14) Tanaka, N.; Thornton, E. R. J. Am. Chem. Soc. 1976, 98, 1617–1619. (15) 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. (16) Im, K.; Park, H.-W.; Kim, Y.; Chung, B.; Ree, M.; Chang, T. Anal. Chem. 2007, 79, 1067–1072. (17) Park, S.; Cho, D.; Ryu, J.; Kwon, K.; Chang, T.; Park, J. J. Chromatogr., A 2002, 958, 183–189. (18) Lee, W.; Cho, D.; Chun, B. O.; Chang, T.; Ree, M. J. Chromatogr., A 2001, 910, 51–60. (19) Chang, T. J. Polym. Sci. Polym. Phys. 2005, 43, 1591–1607. (20) Perny, S.; Allgaier, J.; Cho, D.; Lee, W.; Chang, T. Macromolecules 2001, 34, 5408–5415. (21) Kayillo, S.; Gray, M. J.; Shalliker, R. A.; Dennis, G. R. J. Chromatogr., A 2005, 1073, 83–86. (22) Kim, Y.; Ahn, S.; Chang, T. Anal. Chem. 2010, 82, 1509–1514. (23) Dugo, P.; Cacciola, F.; Kumm, T.; Dugo, G.; Mondello, L. J. Chromatogr., A 2008, 1184, 353–368. (24) Berek, D. Anal. Bioanal. Chem. 2010, 396, 421–441. 4241
dx.doi.org/10.1021/ac2005907 |Anal. Chem. 2011, 83, 4237–4242
Analytical Chemistry
ARTICLE
(25) Vivo-Truyols, G.; van der Wal, S.; Schoenmakers, P. J. Anal. Chem. 2010, 82, 8525–8536. (26) Jiang, X.; van der Horst, A.; Schoenmakers, P. J. J. Chromatogr., A 2002, 982, 55–68. (27) Guiochon, G.; Marchetti, N.; Mriziq, K.; Shalliker, R. A. J. Chromatogr., A 2008, 1189, 109–168. (28) Shalliker, R. A.; Gray, M. J. Adv. Chromatogr. 2006, 44, 177– 236. (29) Janco, M.; Berek, D.; Prudskova, T. Polymer 1995, 36, 3295– 3299. (30) Nguyen, S. H.; Berek, D.; Chiantore, O. Polymer 1998, 39, 5127–5132. (31) Sweeney, A. P.; Shalliker, R. A. J. Chromatogr., A 2002, 968, 41–52. (32) Trathnigg, B.; Rappel, C.; Raml, R.; Gorbunov, A. J. Chromatogr., A 2002, 953, 89–99.
4242
dx.doi.org/10.1021/ac2005907 |Anal. Chem. 2011, 83, 4237–4242