543
J. Sep. Sci. 2015, 38, 543–549
ˇ Dusan Berek Eva Macova´ Polymer Institute, Slovak Academy of Sciences, Bratislava, Slovakia Received September 11, 2014 Revised November 24, 2014 Accepted November 24, 2014
Research Article
Liquid chromatography under limiting conditions of desorption 6: Separation of a four-component polymer blend Baseline separation was achieved of a model four-component polymer blend of polystyrenepoly(methyl methacrylate)-poly(ethylene oxide)-poly(2-vinyl pyridine) in a single chromatographic run with help of the unconventional method of liquid chromatography under limiting conditions of desorption. Narrow barriers of liquids were employed, which selectively decelerated elution of particular kinds of macromolecules. Bare silica gel was the column packing, and the eluent was a mixture of dimethylformamide/tetrahydrofuran/toluene 30:50:20 w/w/w. Barrier compositions were neat toluene, B#1, neat tetrahydrofuran, B#2, and dimethylformamide/tetrahydrofuran/toluene 15:55:30, B#3. Minor blend constituents (1%) could be identified, as well. The result represents a step toward the separation and molecular characterization of triblock-copolymers, many of which are expected to contain besides both parent homopolymers also the diblock chains and thus they are in fact fourcomponent polymer blends. Keywords: Entropy-enthalpy combination / Four-component blends / Liquid chromatography / Minor constituents / Synthetic polymers DOI 10.1002/jssc.201400992
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction The comprehensive molecular characterization of multicomponent polymers, also called complex polymer systems, is an important analytical and technological challenge because numerous synthetic industrial polymeric materials are meaningful blends of distinct kinds of macromolecules [1, 2]. Actually, also many copolymers are complex polymer systems. Typical examples represent block copolymers [3] that as a rule contain their parent homopolymers. Evidently, the constituents of complex polymer systems must be mutually separated to allow their comprehensive molecular characterization in terms of molar mass, chemical structure, and physical architecture. The task is especially demanding when the complex polymer system in question contains macromolecular constituents at a low concentration. Correspondence: Dr. Dusan Berek, Polymer Institute, Slovak Academy of Sciences, Dubravska cesta 9, 84541 Bratislava, Slovakia E-mail: [email protected] Fax: +421-0-2-54775923
Abbreviations: EG-LC, eluent gradient liquid chromatography; LC-CC, liquid chromatography under critical conditions of enthalpic interactions; LC-LCD, liquid chromatography under limiting conditions of desorption; PEO, poly(ethylene oxide); PMMA, poly(methyl methacrylate); PS, polystyrene; S2D-LC, sequenced two-dimensional liquid chromatography; THF, tetrahydrofuran C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
As known, SEC (or gel permeation chromatography) [4–7], does not enable separation of macromolecules of different composition or architecture that possess similar size in solution. Moreover, both low sample capacity and limited detector sensitivity of SEC prevent identification and further processing of minor blend constituents even if their hydrodynamic volumes differ sufficiently. Coupled methods of LC [8] were developed to discriminate macromolecules of distinct compositions or architecture independently of their molar mass. They combine entropic and enthalpic contributions to retention of macromolecules within the chromatographic column packed with porous particles. The best known are liquid chromatography under critical conditions of enthalpic interactions (LC-CC) [9–11], and eluent gradient liquid chromatography (EG-LC) [12, 13]. LCCC was also employed at elevated temperature for separation of polyolefins [11], however the method enables efficient separation of only binary polymer systems. Both LC-CC and EG-LC suffer from limited sample recovery [14–20]. A more recent approach to separation of complex polymer systems is liquid chromatography under limiting conditions of desorption (LC-LCD) [8]. LC-LCD was able to fully discriminate and identify constituents of binary and ternary polymer blends, even those present at the concentration less than 1% [21]. Especially attractive proved direct separation of both parent homopolymers from the diblock copolymers with help of LC-LCD in a single step [22–26]. www.jss-journal.com
544
D. Berek and E. Macova´
In present work, baseline LC-LCD separation is demonstrated of a model polymer blend comprising four chemically distinct kinds of macromolecules including minor macromolecular constituents.
1.1 LC-LCD LC-LCD is based on an original approach to mutual separation of macromolecules with different chemical structure or physical architecture, irrespective of their molar mass. It makes use of the action of liquid barriers within the LC column. Though the principle of LC-LCD operation was described more than ten years ago [8], the method remains a “terra incognita,” which is so far almost undisclosed in practice. As yet only one research group outside of this laboratory applied LC-LCD for separation of products of their syntheses [25, 26]. Therefore the selected details of the LC-LCD procedure are briefly elucidated in present work. The aim is to facilitate launching the method by potential users. The LC-LCD column is packed with porous particles. The transport rate of the low-molecular substances along column is low because they permeate practically all pores of the packing. At the same time, the partially or fully pore-excluded macromolecules tend to elute rapidly. The appropriately chosen low-molecular-weight substance promotes enthalpic interactions of macromolecules to be separated within the chromatographic column namely their adsorption, enthalpic partition or phase separation. In the case of LC-LCD, adsorption of macromolecules within the column packing is employed. Obviously the LC-LCD column packing is to be adsorptive. The eluent is composed of at least two liquids of distinct polarities. One of them suppresses adsorption of all polymeric sample constituents within the column packing, it is their desorli. With advantage, another eluent component promotes adsorption of all but one sample constituents, it is their adsorli. Multicomponent mobile phases can also be employed to allow dissolution and fine tuning of adsorption of particular kinds of macromolecules. Desorli action prevails in the LC-LCD eluent so that if the concerned multicomponent polymer sample is dissolved and injected in the mobile phase, all its constituents elute from the column freely in the exclusion mode. A narrow zone of a mixed solvent with increased concentration of adsorli is injected into the LC-LCD column in front of a two-component polymer sample. The composition of the liquid zone is adjusted to selectively promote retention of the more adsorptive sample constituent. It acts as a slowly eluting “impermeable liquid barrier,” which decelerates adsorbing macromolecules and forces them to elute behind the barrier with the velocity that corresponds to its low elution rate. At the same time, the nonadsorbed macromolecules break-through the above zone and elute unhindered in the SEC mode. In this way, macromolecules, which exhibit distinct adsorptivity are rapidly and efficiently mutually separated. In the case of the complex polymer systems that contain more than two, for example n constituents, n distinct barriers are to be employed, while C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Sep. Sci. 2015, 38, 543–549
each of them decelerates just one sample constituent and letthrough the others. Evidently, n–1 barriers suffice if one kind of macromolecular chains present in the sample is completely nonadsorptive considering particular column packing. Typical representatives of the three-component complex polymer systems are the diblock copolymers that contain their parent homopolymers. Two distinct barriers are employed to separate such complex polymer systems provided elution of one homopolymer is not affected by adsorption [22–24]. Our previous studies indicate several attractive features of LC-LCD: (a) Macromolecules, which are decelerated by a barrier elute exclusively in dependence on their chemical structure or physical architecture, practically nonaffected by their molar mass. This means that molar mass of polymers eluted behind the barriers does not affect their discrimination. (b) LC-LCD separation is fairly independent of the eluent composition [24] provided the SEC elution is attained of all sample constituents in absence of barrier. In many cases solely the barrier composition governs polymer deceleration. Thanks to its robustness, repeatability of the LC-LCD separations is generally high. (c) Identification of appropriate experimental conditions is relatively easy. The same eluent is applied in the course of barrier composition optimization. Often, just a few experiments suffice for adjusting appropriate barrier composition. The latter two advantageous properties contribute to the experimental feasibility of LC-LCD. (d) Peaks of macromolecules eluted behind the barrier are well narrow because a considerable focusing process takes place in the course of polymer transport. The retained macromolecules are accumulated on the barrier edge and their peaks are compressed. This phenomenon allows comprehensive 2D molecular characterization of numerous complex polymer systems. Entire LC-LCD fractions can be one-by-one forwarded into the online SEC column for assessment of molar mass average and dispersity of contained macromolecules. The band broadening effect of the SEC eluted sample is to be considered. The resulting method is denoted as sequenced 2D polymer liquid chromatography, (S2D-LC) [27]. As known, most 2D polymer HPLC procedures necessitate slow separation in the first-dimension column followed with the high-speed SEC. In S2D-LC, both LC-LCD and SEC separations proceed independently so that there is no need to mutually adjust their velocity. Moreover, the LC-LCD separation is repeated for each sample constituent with the injected concentration just required for the optimum SEC characterization. (e) LC-LCD retention volumes are largely independent of the injected sample concentration and the sample capacity of method is very high in terms of injected both sample volume and concentration. This enables not only www.jss-journal.com
J. Sep. Sci. 2015, 38, 543–549
identification but with help of S2D-LC also molecular characterization of minor macromolecular constituents of complex polymer systems. (f)
LC-LCD method is fast and the baseline separation of an ordinary two-constituent polymer blend can be accomplished even within 1 min. This is important not only for the exploratory surveys but also for the comprehensive analyses of multicomponent polymer systems by S2D-LC.
(g) Compared to other coupled methods of polymer HPLC such as LC-CC and EG-LC, LC-LCD exhibits improved sample recovery. (h) Retention volumes of polymers eluted behind barriers can be manipulated with help of time delays between the barrier and sample injections. The longer period of time, in which a polymer is eluted in the SEC mode before its impact with the barrier, the lower its LC-LCD retention volume because the shorter the path of macromolecules behind barrier. LC-LCD also exhibits certain limitations: (i) Separated macromolecules are to be soluble in solvents of different polarity. This is necessary for identification of appropriate desorli and adsorli solvents. To overcome this restriction, multicomponent eluents can be employed. (ii) Column packing must exhibit appropriate adsorptivity. Bare silica gel satisfies most requirements. However, it is difficult to identify appropriate desorli for the highly polar polymers such as poly(N-vinyl pyrrolidone) or poly (4-vinyl pyridine), which are strongly adsorbed on bare silica gel. Less adsorptive column packing such as surface-modified silica gel is to be used in that case. (iii) Sample constituents must exhibit appropriately different adsorptivities. This may be an issue with the low-polarity polymers. To mutually separate macromolecules with low adsorptivity, a distinct retention mechanism is to be utilized instead of adsorption, namely phase separation due to reduced solubility with a nonsolvent barrier [28] or enthalpic partition (absorption) on the nonpolar column packing such as C18 silica bonded phase with a poor solvent barrier [29]. All of the above attributes of LCLCD are to be considered when selecting appropriate separation system.
2 Materials and methods 2.1 Materials The task of present study was to demonstrate the separation potential of LC-LCD, rather than quantitative molecular characterization of discriminated model sample constituents. Four polymers exhibiting distinct polarity and adsorptivity on bare silica gel were chosen. The sample of C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Liquid Chromatography
545
polystyrene (PS) with mass average molar mass Mw 97 kg/mol and with dispersity index (PDI) 1.08 was from Pressure Chemicals, USA. PS with Mw 4 kg/mol with PDI 2.2 was prepared in this Institute by conventional radical polymerization with the AIBN initiator at 80⬚C. Poly(methyl methacrylate) (PMMA) with Mw 65.5 kg/mol and PDI 1.45 was from Rohm, Germany. Poly(ethylene oxide) (PEO) with Mw 50 kg/mol and PDI 2.6 was provided by Novaky Chemical Works, Slovakia. Poly(2-vinyl pyridine) (P2VP) with Mw 68 kg/mol and PDI 1.6 was synthesized in Institut Sadron, France by anionic polymerization. The molar masses of all polymers were provided by suppliers. The SEC molar mass determination of polar PEO and P2VP employing standard polystyrene/divinylbenzene column packings is subject to some uncertainty [30]. Therefore, a polar column GRAM linear (300 × 8) mm from PSS, Germany was used with a mixed eluent of DMF/tetrahydrofuran (THF) 50:50 w/w for all polymers under study. The polystyrene-equivalent molar masses and dispersities of all polymers reasonably agreed with the values provided with suppliers. Appreciable low-molar mass tail was observed on the chromatogram of P2VP. DMF, THF, and toluene were employed as eluent and barrier components. DMF was from Scharlau, Spain, while THF and toluene were from Central Chem Slovakia. All solvents were of analytical grade. THF was distilled immediately before use and stabilized with 0.2 g/L of 2,6-di-tert-butyl-4methyl phenol. All compositions of mobile phases and barriers are given in weight percent. Due to the preferential adsorption effect, the composition of mixed solvent layer on the surface of column packing may differ from the composition of bulk eluent. All polymers under study were well soluble in polar solvents DMF and THF, which tend to strongly interact with bare silica gel and to adsorb on its surface. DMF and THF are solvents for all polymer employed so that precipitation of sample constituents on the surface of column packing could be excluded. The cononsolvency phenomenon, when a mixture of solvents for particular polymer becomes its nonsolvent [31] could be ruled out because the interaction between DMF, THF, and toluene is moderate.
2.2 Instrument and procedures A common HPLC instrument was employed. It was provided with one additional six-port two-way manual injector from Knauer, Germany with a loop of 1000 L. The latter valve was situated between the sample injector and the column and it was used for application of barriers. A valve from Rheodyne, USA, Model 7125 with a loop of 20 L was used to inject sample solutions into the LC-LCD column. The pumping system was HPLC Model 64 from Knauer. The linear velocity of eluent was 1 mL/min in all experiments. The evaporative light scattering detector Model 1000 from PL-Agilent UK/USA was employed for monitoring sample concentration in the column effluent from both GRAM and silica gel columns. www.jss-journal.com
546
D. Berek and E. Macova´
Temperature of nebulizer and evaporator in the detector was set at 40 and 80⬚C, respectively. The gas flow rate was 1.0 mL/min. The detector did not respond to particular solvents at this temperature of evaporator. However, it is known that the evaporative light scattering detector response is not very linear and it may also depend on the composition of liquid, which surrounds macromolecules. These occurrences must be taken into account in the course of quantification of results. Both SEC and LC-LCD measurements were done at ambient temperature between 22 and 24⬚C. As stated above, the SEC column was GRAM from PSS. Its excluded molar mass lies above 1000 kg/mol. The LC-LCD column (300 × 7.5 mm) was home-packed with bare silica gel Kromasil 60A from Eka Chemicals, Sweden. Its particle size was 10 m and its effective pore diameter was 6 nm. The excluded molar mass of PS in THF lies at about 50 kg/mol. Approximate injected concentrations of polymers ci varied between 0.5 and 30 mg/mL. The actual approximate ci values are specified in the figure captions. Different time delays between injections of barriers and sample were employed. The moment of injection of the first barrier is designated on the time scale “zero.” The injection times of further barriers are given in minutes. Sample injection time is indicated as last. For example, the code 0–2–5–9.5 would mean that barrier #2 was injected 2 min after barrier #1, barrier #3 was introduced 3 min after barrier #2, and it was followed with sample in 4.5 min. De facto, the sample was injected 9.5 min after barrier #1.The data collection was started at the moment of sample introduction. Chromatographic data were collected and processed with help of software Clarity from DataApex, Czech Republic.
3 Results and discussion The SEC chromatogram of the model four-constituent blends PS + PMMA + PEO + P2VP with approximately equal relative concentrations obtained with the column GRAM linear and eluent DMF/THF 50:50 is depicted in Supporting Information Fig. S1. PS with highest molar mass was partially separated from the other three blend constituents but its full discrimination from the rest of sample was not attained. Other sample constituents were not separated at all. Addition of second and further similar columns would increase extent of separation but it is hardly possible to expect the baseline discrimination of polymer constituents needed for determination of their relative concentration in sample and assessment of their molar mass. Presence of PS added below about 5% to any other polymer under study remained practically unnoticed (chromatograms not shown). The results clearly manifest the limitation of SEC in separation and molecular characterization of multicomponent complex polymer systems. It is evident that coupled methods of polymer HPLC are to be employed to solve the task. LC-LCD represents a typical example. As outlined above, the important prerequisite of any LCLCD separation experiment is identification of appropriate C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Sep. Sci. 2015, 38, 543–549
mobile phase and barrier(s). The mobile phase must be chromatographically strong enough to suppress adsorption of all sample constituents within LC-LCD column and to secure their free, exclusion controlled elution. Our introductory experiments confirmed high adsorptivity of both PEO and P2VP on bare silica gel. The latter polymers were fully retained in the column from tetrahydrofuran, which otherwise acts as a desorli for many medium-polarity polymers considering bare silica gel. Therefore a chromatographically very strong solvent dimethylformamide, DMF was added to the LC-LCD mobile phase. On the other hand, the liquid barriers have to efficiently and selectively promote adsorption of all sample constituents, except for the least adsorptive polymer, which was PS in present case. To fulfill this condition, a chromatographically weak solvent, toluene was chosen. Toluene is an adsorli for PMMA, PEO, and P2VP and a desorli for PS with respect to bare silica gel. Consequently, ternary mixtures of DMF, THF, and toluene were employed as both mobile phase and barrier B#3, and neat toluene and THF as B#1 and B#2, respectively. However, toluene is a rather poor solvent for PEO and it is a nonsolvent for P2VP. Therefore, the solubility of the latter polymers in eluent and especially in barriers with increased concentration of toluene was carefully controlled by means of independent static experiments. A nonsolvent barrier, which is employed in LC under limiting conditions of insolubility [28] would prevent further processing of fractions with help of S2D-LC. It was shown that in the LC-LCD separation of particular polymer pair, moderate variations of mobile phase composition did not play important role [24]. However, it was anticipated that for separation of a complex polymer system with four constituents some fine tuning of mobile phase could be useful. Mobile phases with composition containing 20 to 30% of DMF, 40 to 50% of THF, and 20 to 30% of toluene were tested. The SEC elution of all polymers under study in particular eluent was checked. Eventually, an eluent composition of DMF/THF/toluene 30:50:20 was chosen, in which the sample recovery was high. Chromatograms are shown of all four polymers in above mixed eluent with Kromasil 60A in Fig. 1. PS and PMMA polymers under study were practically fully excluded from the column and produced well narrow peaks. Such situation is welcome in the LC-LCD separations of multicomponent polymer systems because also the peaks of polymers decelerated by a barrier are to some degree broadened if a low molar mass fraction is present. PEO and especially P2VP contain fractions with lower molar mass, the “tails.” For comparison, also chromatogram of PS with molar mass 4 kg/mol is displayed in Fig. 1A. This polymer gives a broader peak compared to PS 97 kg/mol because it is subject to a regular SEC separation. It is to be stressed that the broad peak of the blend constituent, which is not decelerated by the adsorption-promoting barrier, may consume rather large portion of eluent volume available. As a result, the space is markedly reduced on the LC-LCD chromatogram, which is in hand for polymers eluted behind barrier(s). This is why the column packings with narrow effective diameter of pores and large pore volume are welcome in LC-LCD. Unfortunately www.jss-journal.com
J. Sep. Sci. 2015, 38, 543–549
Figure 1. Chromatograms of PS (A), PMMA (B), and PEO (C) at ci = 6 mg/mL, as well as P2VP (D) at ci = 10 mg/mL obtained without barriers with column Kromasil 60A with eluent DMF/THF/toluene 30:50:20. The column worked in the SEC mode. For explanation see the text.
such materials are hardly available. Alternatively, further barrier can be employed, which would also decelerate the least adsorptive sample constituents to produce its focused peak practically without molar mass and PDI effect. In this case, advantage is sacrificed of work with the LC-LCD system, in which one sample constituent is not retained by any barrier. The next step in the LC-LCD system selection included identification of appropriate barriers. The trial-and-error approach was employed. In fact, the necessary experiments were simple and fast. Eluent composition was kept constant and the stock solutions of individual polymers dissolved in eluent were injected behind barriers of various compositions. The barriers were sought that would decelerate the more polar sample constituent but let through the others. The introductory experiments with Kromasil 60 and barriers of distinct compositions indicated increasing adsorptivity in the following sequence: PS
J. Sep. Sci. 2015, 38, 543–549
ˇ Dusan Berek Eva Macova´ Polymer Institute, Slovak Academy of Sciences, Bratislava, Slovakia Received September 11, 2014 Revised November 24, 2014 Accepted November 24, 2014
Research Article
Liquid chromatography under limiting conditions of desorption 6: Separation of a four-component polymer blend Baseline separation was achieved of a model four-component polymer blend of polystyrenepoly(methyl methacrylate)-poly(ethylene oxide)-poly(2-vinyl pyridine) in a single chromatographic run with help of the unconventional method of liquid chromatography under limiting conditions of desorption. Narrow barriers of liquids were employed, which selectively decelerated elution of particular kinds of macromolecules. Bare silica gel was the column packing, and the eluent was a mixture of dimethylformamide/tetrahydrofuran/toluene 30:50:20 w/w/w. Barrier compositions were neat toluene, B#1, neat tetrahydrofuran, B#2, and dimethylformamide/tetrahydrofuran/toluene 15:55:30, B#3. Minor blend constituents (1%) could be identified, as well. The result represents a step toward the separation and molecular characterization of triblock-copolymers, many of which are expected to contain besides both parent homopolymers also the diblock chains and thus they are in fact fourcomponent polymer blends. Keywords: Entropy-enthalpy combination / Four-component blends / Liquid chromatography / Minor constituents / Synthetic polymers DOI 10.1002/jssc.201400992
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction The comprehensive molecular characterization of multicomponent polymers, also called complex polymer systems, is an important analytical and technological challenge because numerous synthetic industrial polymeric materials are meaningful blends of distinct kinds of macromolecules [1, 2]. Actually, also many copolymers are complex polymer systems. Typical examples represent block copolymers [3] that as a rule contain their parent homopolymers. Evidently, the constituents of complex polymer systems must be mutually separated to allow their comprehensive molecular characterization in terms of molar mass, chemical structure, and physical architecture. The task is especially demanding when the complex polymer system in question contains macromolecular constituents at a low concentration. Correspondence: Dr. Dusan Berek, Polymer Institute, Slovak Academy of Sciences, Dubravska cesta 9, 84541 Bratislava, Slovakia E-mail: [email protected] Fax: +421-0-2-54775923
Abbreviations: EG-LC, eluent gradient liquid chromatography; LC-CC, liquid chromatography under critical conditions of enthalpic interactions; LC-LCD, liquid chromatography under limiting conditions of desorption; PEO, poly(ethylene oxide); PMMA, poly(methyl methacrylate); PS, polystyrene; S2D-LC, sequenced two-dimensional liquid chromatography; THF, tetrahydrofuran C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
As known, SEC (or gel permeation chromatography) [4–7], does not enable separation of macromolecules of different composition or architecture that possess similar size in solution. Moreover, both low sample capacity and limited detector sensitivity of SEC prevent identification and further processing of minor blend constituents even if their hydrodynamic volumes differ sufficiently. Coupled methods of LC [8] were developed to discriminate macromolecules of distinct compositions or architecture independently of their molar mass. They combine entropic and enthalpic contributions to retention of macromolecules within the chromatographic column packed with porous particles. The best known are liquid chromatography under critical conditions of enthalpic interactions (LC-CC) [9–11], and eluent gradient liquid chromatography (EG-LC) [12, 13]. LCCC was also employed at elevated temperature for separation of polyolefins [11], however the method enables efficient separation of only binary polymer systems. Both LC-CC and EG-LC suffer from limited sample recovery [14–20]. A more recent approach to separation of complex polymer systems is liquid chromatography under limiting conditions of desorption (LC-LCD) [8]. LC-LCD was able to fully discriminate and identify constituents of binary and ternary polymer blends, even those present at the concentration less than 1% [21]. Especially attractive proved direct separation of both parent homopolymers from the diblock copolymers with help of LC-LCD in a single step [22–26]. www.jss-journal.com
544
D. Berek and E. Macova´
In present work, baseline LC-LCD separation is demonstrated of a model polymer blend comprising four chemically distinct kinds of macromolecules including minor macromolecular constituents.
1.1 LC-LCD LC-LCD is based on an original approach to mutual separation of macromolecules with different chemical structure or physical architecture, irrespective of their molar mass. It makes use of the action of liquid barriers within the LC column. Though the principle of LC-LCD operation was described more than ten years ago [8], the method remains a “terra incognita,” which is so far almost undisclosed in practice. As yet only one research group outside of this laboratory applied LC-LCD for separation of products of their syntheses [25, 26]. Therefore the selected details of the LC-LCD procedure are briefly elucidated in present work. The aim is to facilitate launching the method by potential users. The LC-LCD column is packed with porous particles. The transport rate of the low-molecular substances along column is low because they permeate practically all pores of the packing. At the same time, the partially or fully pore-excluded macromolecules tend to elute rapidly. The appropriately chosen low-molecular-weight substance promotes enthalpic interactions of macromolecules to be separated within the chromatographic column namely their adsorption, enthalpic partition or phase separation. In the case of LC-LCD, adsorption of macromolecules within the column packing is employed. Obviously the LC-LCD column packing is to be adsorptive. The eluent is composed of at least two liquids of distinct polarities. One of them suppresses adsorption of all polymeric sample constituents within the column packing, it is their desorli. With advantage, another eluent component promotes adsorption of all but one sample constituents, it is their adsorli. Multicomponent mobile phases can also be employed to allow dissolution and fine tuning of adsorption of particular kinds of macromolecules. Desorli action prevails in the LC-LCD eluent so that if the concerned multicomponent polymer sample is dissolved and injected in the mobile phase, all its constituents elute from the column freely in the exclusion mode. A narrow zone of a mixed solvent with increased concentration of adsorli is injected into the LC-LCD column in front of a two-component polymer sample. The composition of the liquid zone is adjusted to selectively promote retention of the more adsorptive sample constituent. It acts as a slowly eluting “impermeable liquid barrier,” which decelerates adsorbing macromolecules and forces them to elute behind the barrier with the velocity that corresponds to its low elution rate. At the same time, the nonadsorbed macromolecules break-through the above zone and elute unhindered in the SEC mode. In this way, macromolecules, which exhibit distinct adsorptivity are rapidly and efficiently mutually separated. In the case of the complex polymer systems that contain more than two, for example n constituents, n distinct barriers are to be employed, while C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Sep. Sci. 2015, 38, 543–549
each of them decelerates just one sample constituent and letthrough the others. Evidently, n–1 barriers suffice if one kind of macromolecular chains present in the sample is completely nonadsorptive considering particular column packing. Typical representatives of the three-component complex polymer systems are the diblock copolymers that contain their parent homopolymers. Two distinct barriers are employed to separate such complex polymer systems provided elution of one homopolymer is not affected by adsorption [22–24]. Our previous studies indicate several attractive features of LC-LCD: (a) Macromolecules, which are decelerated by a barrier elute exclusively in dependence on their chemical structure or physical architecture, practically nonaffected by their molar mass. This means that molar mass of polymers eluted behind the barriers does not affect their discrimination. (b) LC-LCD separation is fairly independent of the eluent composition [24] provided the SEC elution is attained of all sample constituents in absence of barrier. In many cases solely the barrier composition governs polymer deceleration. Thanks to its robustness, repeatability of the LC-LCD separations is generally high. (c) Identification of appropriate experimental conditions is relatively easy. The same eluent is applied in the course of barrier composition optimization. Often, just a few experiments suffice for adjusting appropriate barrier composition. The latter two advantageous properties contribute to the experimental feasibility of LC-LCD. (d) Peaks of macromolecules eluted behind the barrier are well narrow because a considerable focusing process takes place in the course of polymer transport. The retained macromolecules are accumulated on the barrier edge and their peaks are compressed. This phenomenon allows comprehensive 2D molecular characterization of numerous complex polymer systems. Entire LC-LCD fractions can be one-by-one forwarded into the online SEC column for assessment of molar mass average and dispersity of contained macromolecules. The band broadening effect of the SEC eluted sample is to be considered. The resulting method is denoted as sequenced 2D polymer liquid chromatography, (S2D-LC) [27]. As known, most 2D polymer HPLC procedures necessitate slow separation in the first-dimension column followed with the high-speed SEC. In S2D-LC, both LC-LCD and SEC separations proceed independently so that there is no need to mutually adjust their velocity. Moreover, the LC-LCD separation is repeated for each sample constituent with the injected concentration just required for the optimum SEC characterization. (e) LC-LCD retention volumes are largely independent of the injected sample concentration and the sample capacity of method is very high in terms of injected both sample volume and concentration. This enables not only www.jss-journal.com
J. Sep. Sci. 2015, 38, 543–549
identification but with help of S2D-LC also molecular characterization of minor macromolecular constituents of complex polymer systems. (f)
LC-LCD method is fast and the baseline separation of an ordinary two-constituent polymer blend can be accomplished even within 1 min. This is important not only for the exploratory surveys but also for the comprehensive analyses of multicomponent polymer systems by S2D-LC.
(g) Compared to other coupled methods of polymer HPLC such as LC-CC and EG-LC, LC-LCD exhibits improved sample recovery. (h) Retention volumes of polymers eluted behind barriers can be manipulated with help of time delays between the barrier and sample injections. The longer period of time, in which a polymer is eluted in the SEC mode before its impact with the barrier, the lower its LC-LCD retention volume because the shorter the path of macromolecules behind barrier. LC-LCD also exhibits certain limitations: (i) Separated macromolecules are to be soluble in solvents of different polarity. This is necessary for identification of appropriate desorli and adsorli solvents. To overcome this restriction, multicomponent eluents can be employed. (ii) Column packing must exhibit appropriate adsorptivity. Bare silica gel satisfies most requirements. However, it is difficult to identify appropriate desorli for the highly polar polymers such as poly(N-vinyl pyrrolidone) or poly (4-vinyl pyridine), which are strongly adsorbed on bare silica gel. Less adsorptive column packing such as surface-modified silica gel is to be used in that case. (iii) Sample constituents must exhibit appropriately different adsorptivities. This may be an issue with the low-polarity polymers. To mutually separate macromolecules with low adsorptivity, a distinct retention mechanism is to be utilized instead of adsorption, namely phase separation due to reduced solubility with a nonsolvent barrier [28] or enthalpic partition (absorption) on the nonpolar column packing such as C18 silica bonded phase with a poor solvent barrier [29]. All of the above attributes of LCLCD are to be considered when selecting appropriate separation system.
2 Materials and methods 2.1 Materials The task of present study was to demonstrate the separation potential of LC-LCD, rather than quantitative molecular characterization of discriminated model sample constituents. Four polymers exhibiting distinct polarity and adsorptivity on bare silica gel were chosen. The sample of C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Liquid Chromatography
545
polystyrene (PS) with mass average molar mass Mw 97 kg/mol and with dispersity index (PDI) 1.08 was from Pressure Chemicals, USA. PS with Mw 4 kg/mol with PDI 2.2 was prepared in this Institute by conventional radical polymerization with the AIBN initiator at 80⬚C. Poly(methyl methacrylate) (PMMA) with Mw 65.5 kg/mol and PDI 1.45 was from Rohm, Germany. Poly(ethylene oxide) (PEO) with Mw 50 kg/mol and PDI 2.6 was provided by Novaky Chemical Works, Slovakia. Poly(2-vinyl pyridine) (P2VP) with Mw 68 kg/mol and PDI 1.6 was synthesized in Institut Sadron, France by anionic polymerization. The molar masses of all polymers were provided by suppliers. The SEC molar mass determination of polar PEO and P2VP employing standard polystyrene/divinylbenzene column packings is subject to some uncertainty [30]. Therefore, a polar column GRAM linear (300 × 8) mm from PSS, Germany was used with a mixed eluent of DMF/tetrahydrofuran (THF) 50:50 w/w for all polymers under study. The polystyrene-equivalent molar masses and dispersities of all polymers reasonably agreed with the values provided with suppliers. Appreciable low-molar mass tail was observed on the chromatogram of P2VP. DMF, THF, and toluene were employed as eluent and barrier components. DMF was from Scharlau, Spain, while THF and toluene were from Central Chem Slovakia. All solvents were of analytical grade. THF was distilled immediately before use and stabilized with 0.2 g/L of 2,6-di-tert-butyl-4methyl phenol. All compositions of mobile phases and barriers are given in weight percent. Due to the preferential adsorption effect, the composition of mixed solvent layer on the surface of column packing may differ from the composition of bulk eluent. All polymers under study were well soluble in polar solvents DMF and THF, which tend to strongly interact with bare silica gel and to adsorb on its surface. DMF and THF are solvents for all polymer employed so that precipitation of sample constituents on the surface of column packing could be excluded. The cononsolvency phenomenon, when a mixture of solvents for particular polymer becomes its nonsolvent [31] could be ruled out because the interaction between DMF, THF, and toluene is moderate.
2.2 Instrument and procedures A common HPLC instrument was employed. It was provided with one additional six-port two-way manual injector from Knauer, Germany with a loop of 1000 L. The latter valve was situated between the sample injector and the column and it was used for application of barriers. A valve from Rheodyne, USA, Model 7125 with a loop of 20 L was used to inject sample solutions into the LC-LCD column. The pumping system was HPLC Model 64 from Knauer. The linear velocity of eluent was 1 mL/min in all experiments. The evaporative light scattering detector Model 1000 from PL-Agilent UK/USA was employed for monitoring sample concentration in the column effluent from both GRAM and silica gel columns. www.jss-journal.com
546
D. Berek and E. Macova´
Temperature of nebulizer and evaporator in the detector was set at 40 and 80⬚C, respectively. The gas flow rate was 1.0 mL/min. The detector did not respond to particular solvents at this temperature of evaporator. However, it is known that the evaporative light scattering detector response is not very linear and it may also depend on the composition of liquid, which surrounds macromolecules. These occurrences must be taken into account in the course of quantification of results. Both SEC and LC-LCD measurements were done at ambient temperature between 22 and 24⬚C. As stated above, the SEC column was GRAM from PSS. Its excluded molar mass lies above 1000 kg/mol. The LC-LCD column (300 × 7.5 mm) was home-packed with bare silica gel Kromasil 60A from Eka Chemicals, Sweden. Its particle size was 10 m and its effective pore diameter was 6 nm. The excluded molar mass of PS in THF lies at about 50 kg/mol. Approximate injected concentrations of polymers ci varied between 0.5 and 30 mg/mL. The actual approximate ci values are specified in the figure captions. Different time delays between injections of barriers and sample were employed. The moment of injection of the first barrier is designated on the time scale “zero.” The injection times of further barriers are given in minutes. Sample injection time is indicated as last. For example, the code 0–2–5–9.5 would mean that barrier #2 was injected 2 min after barrier #1, barrier #3 was introduced 3 min after barrier #2, and it was followed with sample in 4.5 min. De facto, the sample was injected 9.5 min after barrier #1.The data collection was started at the moment of sample introduction. Chromatographic data were collected and processed with help of software Clarity from DataApex, Czech Republic.
3 Results and discussion The SEC chromatogram of the model four-constituent blends PS + PMMA + PEO + P2VP with approximately equal relative concentrations obtained with the column GRAM linear and eluent DMF/THF 50:50 is depicted in Supporting Information Fig. S1. PS with highest molar mass was partially separated from the other three blend constituents but its full discrimination from the rest of sample was not attained. Other sample constituents were not separated at all. Addition of second and further similar columns would increase extent of separation but it is hardly possible to expect the baseline discrimination of polymer constituents needed for determination of their relative concentration in sample and assessment of their molar mass. Presence of PS added below about 5% to any other polymer under study remained practically unnoticed (chromatograms not shown). The results clearly manifest the limitation of SEC in separation and molecular characterization of multicomponent complex polymer systems. It is evident that coupled methods of polymer HPLC are to be employed to solve the task. LC-LCD represents a typical example. As outlined above, the important prerequisite of any LCLCD separation experiment is identification of appropriate C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
J. Sep. Sci. 2015, 38, 543–549
mobile phase and barrier(s). The mobile phase must be chromatographically strong enough to suppress adsorption of all sample constituents within LC-LCD column and to secure their free, exclusion controlled elution. Our introductory experiments confirmed high adsorptivity of both PEO and P2VP on bare silica gel. The latter polymers were fully retained in the column from tetrahydrofuran, which otherwise acts as a desorli for many medium-polarity polymers considering bare silica gel. Therefore a chromatographically very strong solvent dimethylformamide, DMF was added to the LC-LCD mobile phase. On the other hand, the liquid barriers have to efficiently and selectively promote adsorption of all sample constituents, except for the least adsorptive polymer, which was PS in present case. To fulfill this condition, a chromatographically weak solvent, toluene was chosen. Toluene is an adsorli for PMMA, PEO, and P2VP and a desorli for PS with respect to bare silica gel. Consequently, ternary mixtures of DMF, THF, and toluene were employed as both mobile phase and barrier B#3, and neat toluene and THF as B#1 and B#2, respectively. However, toluene is a rather poor solvent for PEO and it is a nonsolvent for P2VP. Therefore, the solubility of the latter polymers in eluent and especially in barriers with increased concentration of toluene was carefully controlled by means of independent static experiments. A nonsolvent barrier, which is employed in LC under limiting conditions of insolubility [28] would prevent further processing of fractions with help of S2D-LC. It was shown that in the LC-LCD separation of particular polymer pair, moderate variations of mobile phase composition did not play important role [24]. However, it was anticipated that for separation of a complex polymer system with four constituents some fine tuning of mobile phase could be useful. Mobile phases with composition containing 20 to 30% of DMF, 40 to 50% of THF, and 20 to 30% of toluene were tested. The SEC elution of all polymers under study in particular eluent was checked. Eventually, an eluent composition of DMF/THF/toluene 30:50:20 was chosen, in which the sample recovery was high. Chromatograms are shown of all four polymers in above mixed eluent with Kromasil 60A in Fig. 1. PS and PMMA polymers under study were practically fully excluded from the column and produced well narrow peaks. Such situation is welcome in the LC-LCD separations of multicomponent polymer systems because also the peaks of polymers decelerated by a barrier are to some degree broadened if a low molar mass fraction is present. PEO and especially P2VP contain fractions with lower molar mass, the “tails.” For comparison, also chromatogram of PS with molar mass 4 kg/mol is displayed in Fig. 1A. This polymer gives a broader peak compared to PS 97 kg/mol because it is subject to a regular SEC separation. It is to be stressed that the broad peak of the blend constituent, which is not decelerated by the adsorption-promoting barrier, may consume rather large portion of eluent volume available. As a result, the space is markedly reduced on the LC-LCD chromatogram, which is in hand for polymers eluted behind barrier(s). This is why the column packings with narrow effective diameter of pores and large pore volume are welcome in LC-LCD. Unfortunately www.jss-journal.com
J. Sep. Sci. 2015, 38, 543–549
Figure 1. Chromatograms of PS (A), PMMA (B), and PEO (C) at ci = 6 mg/mL, as well as P2VP (D) at ci = 10 mg/mL obtained without barriers with column Kromasil 60A with eluent DMF/THF/toluene 30:50:20. The column worked in the SEC mode. For explanation see the text.
such materials are hardly available. Alternatively, further barrier can be employed, which would also decelerate the least adsorptive sample constituents to produce its focused peak practically without molar mass and PDI effect. In this case, advantage is sacrificed of work with the LC-LCD system, in which one sample constituent is not retained by any barrier. The next step in the LC-LCD system selection included identification of appropriate barriers. The trial-and-error approach was employed. In fact, the necessary experiments were simple and fast. Eluent composition was kept constant and the stock solutions of individual polymers dissolved in eluent were injected behind barriers of various compositions. The barriers were sought that would decelerate the more polar sample constituent but let through the others. The introductory experiments with Kromasil 60 and barriers of distinct compositions indicated increasing adsorptivity in the following sequence: PS
