Critical Conditions for Liquid Chromatography of Statistical

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article

Critical Conditions for Liquid Chromatography of statistical copolymers: Functionality Type and Composition Distribution Characterization by UP-LCCC/ESI-MS Ruben Epping, Ulrich Panne, and Jana Falkenhagen Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b04064 • Publication Date (Web): 28 Dec 2016 Downloaded from http://pubs.acs.org on January 3, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Critical Conditions for Liquid Chromatography of statistical copolymers: Functionality Type and Composition Distribution Characterization by UP-LCCC/ESI-MS Ruben Epping, Ulrich Panne and Jana Falkenhagen* Bundesanstalt für Materialforschung und –prüfung (BAM), Richard-Willstätter-Straße 11, 12489 Berlin; Germany; *E-Mail: [email protected], Phone: +493081041632; Fax no: +493081041137 ABSTRACT: Statistical ethylene oxide (EO) and propylene oxide (PO) copolymers of different monomer compositions and different average molar masses additionally containing two kinds of end groups (FTD) were investigated by Ultra High Pressure Liquid Chromatography under Critical Conditions (UP-LCCC) combined with Electrospray Ionization Time-of Flight Mass Spectrometry (ESI-TOF-MS). Theoretical predictions of the existence of a Critical Adsorption Point (CPA) for statistical copolymers with a given chemical and sequence distribution1, could be studied and confirmed. A fundamentally new approach to determine these critical conditions in a copolymer, alongside the inevitable Chemical Composition Distribution (CCD), with mass spectrometric detection is described. The shift of the critical eluent composition with the monomer composition of the polymers was determined. Due to the broad molar mass Distribution (MMD) and the presumed existence of different end group functionalities as well as Monomer Sequence Distribution (MSD), gradient separation only by CCD was not possible. Therefore, isocratic separation conditions at the CPA of definite CCD fractions were developed. Although the various present distributions partly superimposed the separation process the goal of separation by end group functionality was still achieved on the basis of the additional dimension of ESI-TOF-MS. The existence of HO-H besides the desired AllylO-H end group functionalities was confirmed and their amount estimated. Furthermore, indications for a MSD were found by UPLC/MS/MS measurements. This approach offers for the first time the possibility to obtain a fingerprint of a broad distributed statistical copolymer including MMD, FTD, CCD and MSD.

Aliphatic polyethers of ethylene oxide (EO) and propylene oxide (PO) are among the most widely used polymer classes. Copolymers of EO and PO have many different applications due to their amphiphilic character of the hydrophilic EO and hydrophobic PO. Especially the biocompatibility and the wide availability are major benefits for universal applicability. Commercial production reaches several million tons per year.2 Starting materials are the respective alkenes, which are oxidized to form epoxides. The ring opening polymerization (ROP) is most commonly base-initiated. The standard method to produce poly(ethylene glycols) (PEG) is the controlled addition of EO to water or alcohols as initiators in a polar aprotic solvent like THF in the presence of alkaline (mostly NaOH) catalysts. The result of the living anionic polymerization is a poisson-distribution in molar mass.2,3 A main obstacle of the polymerization of PO is the proton abstraction from the methyl group in a chain transfer reaction. This results in a limitation of the oxyanionic polymerization to lower weight PPG’s and an unsaturated allyl end group for the newly formed chain.4 The polymerization rate of PO is considerably lower than that of EO. This is due to the formation of secondary hydroxyl groups in PPG formation and primary hydroxyl groups in PEG formation. Primary alkoxides are much more reactive than secondary. The formation of primary hydroxyl groups in the reaction of PO is negligible. These formations can be influenced by various factors such as type of initiator, solvent or temperature.5 For example higher alkalines or counterion complexation with crown ethers decrease the isomerisa-

tion of PO to allyl-alcohols and lead to higher molecular weight polymers.2 Block copolymers of EO and PO can be obtained by the polymerization of one kind of monomer to a homopolymer followed by the alkoxylation of one or both alcohol ends of the homopolymer. Statistical copolymers are formed by the simultaneous polymerization of EO and PO. Due to the fact that the methyl proton abstraction and thereby the formation of allyl end groups does not occur in the polymerization of EO, the statistical copolymer will likely consist of a distribution in end group functionality. The polymerization will result in AllylO-(EO)nco(PO)m-H and HO-(EO)nco(PO)m-H copolymers. Scheme 1 shows the initiation, polymerization and side reactions of the formation of PEG and PPG. Scheme 1. Reactions in the formation of EO-POcopolymers.

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

a.) Initiation of the anionic living polymerisation. b.) Alternative initiation reaction resulting in an allyl end group. M = alkaline c.) Ring opening anionic polymerization to poly ethylene glycols. d.) Ring opening anionic polymerization to poly propylene glycols with chain transfer reaction resulting in allyl end group.

The properties of these copolymers depend strongly on the distributions of Molar mass (MMD), Chemical Composition (CCD), Functionality (FTD), Monomer Sequence (MSD) and possible other factors like monomer/ homopolymer content. Usually the average molar mass and the monomer composition are given. These data are relatively easy accessible. The comprehensive characterization of these products with their numerous distributions to produce polymers with specific targeted properties is time-consuming and requires hyphenated analytical techniques. Liquid chromatographic methods are among the most frequently used techniques to obtain the desired information.6 The mechanisms of chromatographic separations are determined by the interaction parameter c.7 C describes the interaction of a structural unit with the stationary phase. It depends on the mobile and stationary phase composition as well as the temperature. It is independent of the pore size, overall pore volume and interstitial volume.8 In most cases a change of mobile phase composition is used to alter the value of c and thereby the retention mode. In Size Exclusion Chromatography (SEC) the interaction of the analytes with the stationary phase is ideally only of entropic nature. The interaction is characterized by conformation changes of the polymer chains while entering into pores of the stationary phase. Enthalpic interaction should not occur. The interaction parameter c is thereby negative. This chromatographic mode is commonly used to separate polymer chains according to their size in solution. However, the size of polymers in solution is also influenced by all the other possible polymer distributions present, which superimpose the separation. Thus, the determination of the average molar mass and molar mass distribution for copolymers remains a challenge. In Liquid Adsorption Chromatography (LAC) the interaction parameter is positive. Enthalpic interactions with the stationary phase dominate the separation. But different distributions in copolymers also superimpose the separation of one another. Besides separation by CCD or FTD the separation in LAC also strongly depends on the MW. The retention time is related to M1/2. To apply this method to higher molar mass polymers a gradient can be used to squeeze retention times. Between the SEC and LAC separation mode the Critical Point of Adsorption (CPA) is located where c equals zero.9,10 At this point the enthalpic and entropic interactions for a specific monomer compensate each other. Therefore the separation at this point should not depend on the chain length/ molar mass of the polymer and only other interactions, for example end group functionality influence the retention time.11 These conditions exist for a definite eluent composition, stationary phase and temperature. The isocratic separation at these conditions is called Liquid Chromatography at Critical Conditions (LCCC).12 This chromatographic mode can also be used for the separation of copolymers. At the critical conditions for one of the monomers of a block copolymer, separation according to the length of the other block can be obtained.13,14,15 Unfortunately this use of the LCCC mode can only be applied for block

Page 2 of 9

copolymers. Due to the different sequence structure statistical copolymers cannot be characterized the same way. Brun proposed that also statistical copolymers possess a CPA. The chemical correlation segment, characterizing the randomness of the copolymer can act as a hypothetical homopolymer with a single value for c.1 This is true as long as the statistical copolymer does not comprise an additional CCD. For a statistical copolymer with a CCD each compositionally homogeneous fraction has its own CPA. If the molar mass of a copolymer is high enough Brun showed that in a gradient separation each compositional fraction elutes at its own critical eluent composition.16 Wang et al confirmed this theory by Monte Carlo simulations.17 She also found that separation by CCD can only be successful if the sequence distribution (blockiness of the copolymer) is narrow. Copolymers with the same chemical composition but different degree of blockiness have slightly different CPA’s. Blockier copolymers have always a greater retention (higher value of c) than alternating structured copolymers. Block copolymers do not possess a CPA. Only if the mean length of chain segments connecting the attractive walls of the pore exceeds the chemical correlation segment, characterizing the randomness of the copolymer, the copolymer possesses a CPA. If a polymer sample has more than one distribution, which is always the case with a copolymer, a possible solution for the separation could be the use of 2D-Chromatography.18,9 The polymer is separated according to one distribution in the first dimension and according to another distribution in the second dimension.19 The difficulty is to find orthogonal separation modes for both dimensions, because in many cases the separation partly superimposes each other. Block copolymers of EO and PO for example can be separated according to the length of one block in the first and according to the length of the other block in the second dimension.15 Homopolymers can be separated by MWD in one and FTD in the other dimension. In other cases, an orthogonal separation mode is not possible. For example, there is no separation mode that can distinguish between CCD and End Group Functionality Distribution. For lower and medium polymerization degrees of statistical copolymers it is also impossible to be separated by MMD and CCD in different dimensions. In cases like these, the detection with mass spectrometry can offer an additional non-chromatographic dimension to the analysis. ESI-MS is well suited for the online coupling with UPLC.20,21 In general ESI-MS enables the analysis of polar analytes. To analyze less polar molecules, dissolved in less polar solvents, the addition of an inorganic salt solution in a more polar solvent between the LC-column and the ion source through a mixing tee is necessary. Usually alkaline salts with sodium or potassium are used.22 Depending on the polarity and the molar mass of the analytes, ESI-MS spectra show multiply charged ions. Multiply charged ions can enable the analysis of very high molar mass analytes with the limited m/z range of a quadrupol or TOF analyzer but can also hinder the analysis due to the additional complexity of the mass spectra. A suppression of multiply charging to a degree, by the addition of the alkaline salts, is possible. Mass spectrometry of polymers can provide information on the MWD, end groups and the CCD. It is though impossible to discriminate between block and statistical copolymers by MS alone. The use of MS/MS may lead to insights of the sequence distribution. Quantification with mass spectrometry is difficult due to different ioniza-


Page 3 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


tion, transmission, and detection probabilities depending among others on MW or functional groups. In this work we analyzed statistical EO-PO copolymers with different average molar masses and different monomer compositions by UP-LCCC/ESI-TOF-MS. The main question was if it is possible to distinguish between the AllylO(EO)nco(PO)m-H and HO-(EO)nco(PO)m-H copolymers in the mixture. If this is the case the next step should be the roughly estimation of the amount of different end group species. An additional open question is the detection of possibly existent MSD.

EXPERIMENTAL SECTION Materials. Statistical copolymers of EO and PO were obtained from a commercial supplier. The notations used in this paper and the mean molar mass and monomer composition given by the supplier are listed in Table 1. Samples 1 to 5 were synthesized intentionally to yield AllylO-H end groups by using allyl-alcohol as starter. Sample 6 were synthesized to yield diol copolymers. The first four samples were the samples of interest for this work. The last two were derived as reference samples for the two end group possibilities. Table 1. Notation of the Copolymer samples. Notation

mean molar mass [g/mol]

monomer composition [mol% EO]

1

1800

50

2

550

90

3

2300

25

4

1000

75

5

750

75

6

750

75

The mean molar masses and monomer compositions were calculated by the applied amounts of monomers and reaction starters. Chromatography. For the chromatographic separation a UPLC system (Waters GmbH) was used. The system was equipped with one ACQUITY UPLC BEH C18 Column (130Å, 1.7 µm particle size, 2.1 mm x 50 mm) and one ACQUITY UPLC BEH C18 VanGuard Pre-column (130Å, 1.7 µm, 2.1 mm x 5 mm). The autosampler and column were operated at 25 °C. The experiments were performed with a mobile phase system of water and tetrahydrofuran (THF). An amount of 0.2% formic acid was added to the mobile phase. The flow rate was 0.2 mL/min. For the isocratic measurements the samples were dissolved in a solvent mixture corresponding to the mobile phase composition. Two microliters of the samples with a concentration of 0.1 mg/mL were injected. UPLC/MS grade solvents (TH Geyer, Sigma-Aldrich) were used for chromatography. Mass Spectrometry. A Q-TOF Ultima ESI-TOF mass spectrometer (Micromass) running at 3 kV capillary voltage, at a source temperature of 120 °C and a desolvation temperature of 350 °C was used for all measurements. The mass spectrometer was operating in the positive ion mode. A solution of 0.05 mg/L sodium trifluoroacetate was added to the eluent from the chromatography via a mixing tee in a flow of 3µL/min to improve ionization conditions. Tandem mass spectrometry (MS/MS) experiments using collision induced

dissociation (CID) were carried out with collision energy between 50 and 70 eV. Argon was used as collision gas.

RESULTS AND DISCUSSION To analyze the statistical copolymers, first the chromatographic mode had to be selected to achieve the best results. The main goal was to get information on the functionality distribution of the samples 1 to 4. These samples were expected to provide AllylO-H end groups by using allyl-alcohol as a starter. Due to water impurities, which can also function as a starter, or side reactions like chain transfer or deprotonation/ reprotonation reactions, the formation of HO-H end groups is possible. The two end group possibilities are shown in Chart 1. Chart 1. The two possible end group combinations of the samples. EO and PO groups are statistical distributed. Top: AllylO-H, bottom HO-H end groups.

Statistical copolymers were previously investigated by gradient elution liquid chromatography. According to investigations by Brun,16 a separation into fractions with different CPA’s is possible. In our experiments the samples did not show this separation. We used a solvent system of water and THF with a C18 stationary phase. The gradient was run from 5 to 95% THF in 5 minutes. The results showed that the separation was superimposed by molar mass and end group functionality. This is due to the relatively low average molar masses. To be eluted at their correspondent critical eluent compositions, the molar masses would have to be much higher, so that the polymer chains are excluded from the pores of the stationary phase. Therefore we decided for the isocratic elution at the CPA. The respective CPA for the stated chemical composition by the manufacturer was used for each sample. To determine the critical conditions an ESI-MS as a detector was used. The method previously reported23 for this was improved. Different from the analysis of homopolymers, where it is easy to determine critical conditions from just a few chromatographic runs of a single sample, the analysis of a copolymer pose a challenge. The main obstacle is the CCD. If the CPA for a certain ratio of EO and PO is to be determined, the dispersity in chemical composition in the sample can superimpose the identification of the CPA. In Figure 1 the Total Ion Chromatogramm (TIC) and three mass spectra at different times in the peak are shown exemplarily for the sample 4. The Chromatogramm was recorded at 26% THF which represents the critical eluent composition for a 3:1 EO/PO ratio in copolymer composition (see determination of actual critical conditions later). Nevertheless, the maximum of the peak distribution is shifted towards lower molar masses with increasing retention time. This would indicate an elution in the SEC mode in homopolymer analysis. However, because the analyzed sample was a copolymer and did not only contain 3:1 copolymer, the distribution in copolymer composition complicated the analysis.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1. Total Ion Chromatogramm and mass spectra at three different elution times for sample 4 at the critical condition for the respected overall 3:1 EO/PO ratio.

To determine the critical conditions, the contemplation of selected ion chromatograms (SIC’s) was necessary. In Figure 2 a. three SIC’s with the same chemical composition but different polymerization degrees are pictured. Here it becomes clear that they (and all other polymerizations degrees with the same chemical composition) have the same retention time and therefore the selected eluent composition and other experiment parameters represent the critical conditions for them. The other series of SIC’s in Figure 2 b, c and d, which show the increase of one or both monomer amounts in the copolymer to different chemical compositions, did not show retention under LCCC conditions. This could not be seen in Figure 1. Note that in the SIC’s (and some of the other chromatograms following) one or two additional small peaks of low intensity at the same mass trace are visible. A possible explanation is the separation of fractions with different monomer sequence distributions (MSD). Only if the chemical correlation segment does not exceed the pore diameter the copolymer chains belong to the same CPA fraction. This observation will be discussed later in this article.

Page 4 of 9

the relatively low polydispersity and high mass increment of the chemical composition. In Figure 3 the mass traces for both end group possibilities AllylO-H and HO-H are shown. As expected all figures show an elution under critical conditions for both end group combinations, which additionally verifies the chromatographic mode. The more polar HO-H copolymers elute before the AllylO-H copolymers due to the nonpolar C18 stationary phase used. The higher the EO amount, and thereby the higher the polarity of the copolymer (see Table 1), the better is the separation. That is because the more polar the polymer back bone, the more polar is the eluent composition at the respected CPA. By definition an unfunctionalized polymer chain would elute at the system dead time (here: 1.49 minutes) when eluted under its critical conditions. For the samples in Figure 3 that means, the retention of the polymers is only influenced by the end group functionalities. That further means, the more polar the critical eluent composition for the different samples is, the higher is the retention time for the less polar AllylO-H chains. Due to the similarity in polarity to the backbone the retention time of the HO-H polymers is only slightly influenced. Judging by the peak area significant amounts of HO-H copolymer is present in the samples. Although the end group separation is still superimposed by the CCD (not by the MMD anymore) and therefore no separated peaks could be detected in the overall (TIC) chromatogram, with the additional dimension of ESI-MS the presence of HO-H end group copolymers could be confirmed. The analyses could be carried out in just less than 4 minutes each.

Figure 3. Selected ion chromatograms of specific polymerization degrees and fixed chemical composition of Sample 1 at 54% THF, Sample 2 at 18% THF, Sample 3 at 77% THF and Sample 4 at 26% THF. Both AllylO-H and HO-H end group mass traces are pictured. Figure 2. Reconstructed ion chromatograms of specific polymerization grades and chemical composition of sample 4 for the critical eluent composition of 26% THF. a.) with the same CCD (EO/PO ration 3:1); b.) with increasing number of EO; c.) with increasing number of PO; d.) with increasing EO and PO but different CCD.

The critical conditions were determined for all samples for the assumed chemical composition. SIC’s at the CPA’s for different polymerization degrees for all samples are shown in Figure 3. For sample 2 only two SIC’s could be created due to

Additionally, the two samples 5 and 6 with the same overall monomer composition and average molar mass were analyzed. Sample 5 was synthesized with the allyl-alcohol starter (like the samples above) and sample 6 was synthesized without the allyl-alcohol starter and indented to yield only HO-H end groups. In Figure 4 the results can be seen. In sample 5, just like in the other samples, AllylO-H and HO-H copolymers are present. The results for sample 6 do not show any AllylOH copolymers. This indicates that the side reactions resulting in AllylO-H copolymers seen in Scheme 1 did not take place


Page 5 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


in the formation of this copolymer. It seems that the side reaction could have been avoided by the manufacturer. Several approaches, like the “active monomer strategy” or the use of coordinative catalysts are conceivable for this.

are also accounted for the calculations. The spectra consist mainly of Na+-adduct signals with a very small amount of H2O+/NH4+-adducts. The signal distances show a typical copolymer pattern with distances of m/z = 14 resulting from the monomer masses of m/z = 44 (EO) and m/z = 58 (PO).

Figure 4. Selected ion chromatograms of specific polymerization degrees and fixed chemical composition of sample 5 (left) and sample 6 (right). Both AllylO-H and HO-H end group mass traces are pictured.

In Figure 5 the critical eluent compositions are plotted against the copolymer composition. Additionally, to the samples above the critical conditions for EO and PO homopolymer were determined. The curve shows that the analysis under critical conditions of any possible chemical composition EOPO copolymer is possible with the chosen chromatographic system. The shape of the curve does not show a linear dependence. That means the adjustment of critical conditions is more difficult towards both extreme chemical compositions of homopolymers. Critical conditions for any given chemical composition can be calculated through interpolation.

Figure 5. Critical eluent composition vs Copolymer composition for statistical EO-PO copolymers in the eluent system THF/Water on a C18 stationary phase at 25°C.

To get a complete picture of the chemical composition distribution 2D-plots of the samples were calculated. To obtain this, first the mass spectra for each sample were summed up over the entire polymer peak of the respective sample. Hereby the whole SIC of every ion species is taken into account. The summed mass spectra are shown in Figure 6. It is obvious that clear spectra with no impurities or other interfering signals could be gained for the samples. For the higher average mass samples, the amount of double and even higher charged species increases significantly. However multiply charged ions

Figure 6. Mass spectra of the analyzed samples summed over the entire sample peak eluted under the respected critical conditions.

Such highly resolved and high quality mass spectra could only be obtained by the chosen chromatographic mode and conditions. Without chromatography or even in the SEC mode mass spectra displayed poorer resolution and disturbing background peaks, which would deteriorate the processing. In Figure 7 the 2D-plots for the AllylO-H end group functionalities and plots for the SIC peak areas of both end group possibilities, following a respective chemical composition trace of the samples, are shown. For each sample the chemical composition trace of the distributor given composition was used. However, the relative amounts calculated are only true under the assumption, that the ionization efficiency is constant or at least quite similar over the entire mass range and FTD. Given the relatively low average masses and low dispersities this is a fair assumption for these samples. In the representation it can be seen that for all samples the intended AllylO-H copolymers indeed reassemble the average copolymer composition given by the distributor (white lines). The differences in the CCD are visible. Only for sample 2 the average composition seems to be a little off the expected value. For this sample it was not possible to issue the peak area graph for both end group options. Due to the high mass increment of the 9:1 trace and relatively low polydispersity only two measuring points existed, which were too few to integrate the derived curves. For the HO-H species, integrating the SIC peak areas along just one mass trace was the only way to get the amount of them present. 2D-plots as for the AllylO-H species (not shown) showed massive false findings for this end group composition. The mass resolution was not high enough to distinguish between hypothetical isomers of HO-H species with higher PO and lower EO content to the AllylO-H copolymers actually present. By integrating the SIC peak areas of one trace only these false findings could be excluded by also taking the retention time into account.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 9

Figure 7. Number of EO units vs. number of PO units 2D-plots for AllylO-H end groups (top) and SIC peak areas vs. m/z graphs following one chemical composition trace (white line, bottom) for sample 1 to 4.

To approximate the amount of HO-H species present in the samples the peak areas in the SIC peak areas vs. m/z plots were integrated. The percentages of HO-H to AllylO-H end groups were calculated. With this we found a diol content of about 19% in sample 1, 16% in sample 3 and 6%in sample 4. The reference samples 5 and 6 were also analyzed in the same way (not shown). The HO-H content of sample 5 was calculated to about 12%. In sample 6, which were intentionally synthesized to yield HO-H end groups no AllylO-H end groups could be detected (see before). With the aid of these calculated 2D-plots it was possible to analyze the CCD of the EO-PO-copolymers with AllylO-H end groups although no overall baseline separation is possible for these polymers. The FTD and CCD would superimpose each other in all known separation processes so far. Due to the additional dimension of mass spectrometry, information could be obtained that could not have been accessed with any other detection technique. Based on that, we are further working on a better way to quantify the amount of different end group functionalities. The program MassChrom2D24, which was used to produce the 2D-plots, is only capable of producing normalized intensities, as seen in the figures above. By integration of selected ion chromatograms by hand, we showed that the amount of HO-H polymers in comparison to AllylO-H polymers is in the range of 6-19% for these samples. To quantify the different species accurately it is further needed to investigate the different ionization efficiencies of AllylO-H and HO-H copolymers as well as account for mass isomers in the data set by appropriate data processing. With that taken into account, it will be possible to obtain the amount of polymer end group fractions over the whole CCD range. Another possible distribution in copolymer samples is the monomer sequence distribution (MSD). Due to the relatively low molar mass of the samples and principally different reactivities of the two monomers it is unlikely that all polymer

chains with the same chemical composition and molar mass (polymerization degree) also exhibit the same monomer sequence. The microstructure (as well as the chemical composition) of a single polymer chain depends on the monomer feed, which changes with the conversion of the polymer. The resulting copolymer is of non-ergodic nature. It consists of ergodic fractions which can differ in chemical composition and monomer sequence.1 Since it is easy to distinguish between different chemical compositions with a mass spectrometric detection, for this section only the possibility of different monomer sequences is considered. Different monomer sequences (of a peculiar ergodic fraction with the same chemical composition) exhibit the same CPA as long as the chemical correlation segment does not exceed the pore diameter.1 If the chemical correlation segment exceeds the pore diameter the retention will increase and depend on the amount and length of alternating inner blocks. Figure 2 and Figure 3 show two or three peak maxima for many of the selected ion chromatograms (SIC’s). It is conceivable that these consist of different ergodic fractions varying in the monomer sequence. The main peak is most likely to consist of all monomer sequences with inner blocks shorter than the pore diameter. The second and third peak could possibly consist of monomer sequences with inner blocks exceeding the pore diameter. If some inner blocks of a statistical copolymer exceed the pore diameter, the retention at the CPA of the truly statistical analogue depends on the length and amount of these inner blocks.17 To investigate this, UPLC/ESI-MS/MS measurements were carried out. The typical fragmentation patterns of EO-POcopolymers are well known.25,26,27,28 The possible fragments are shown in Scheme 2. If the unknown sequence with the two possible monomers at any position of the polymer chain is added to the already large number of possible cleavage product ions, it becomes obvious that the resulting MS/MS-spectra can become especially complicated with many feasible mass isomers.27,29


Page 7 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 8. MS/MS-spectra of the three peaks from parent ion m/z 841.8; Left: full spectra; Right: enhanced spectra of the first and second peak.

Therefore the analyses focus on the most abundant series and only specific m/z regions of the spectra. Exemplarily the UPLC/MS/MS measurement of m/z 841.7 from sample 4 will be discussed. The SIC of this ion can be seen in Figure 3 Sample 4, red line. Figure 8 displays the MS/MS-spectra of the three peaks in that mass trace. Additionally a section of the first and second peak is enhanced. Scheme 2. Possible fragmentation scheme of EO/POcopolymers for the formation of the A’, A’’, B’, B’’, C’ and C’’ series ions. R Na

O

H O

n

MSMS

R

R

A x'

Na

O

O

Na

H

HO

R

R

B x'

O

Na

O

O

Na

H y

R R

R Na

By''

O

x R

C x'

Ay''

H O y

x

O

O O x

R

Na

O

H O

O

Cy''

y

R

R = H for EO and R = CH 3 for PO

The parent ion and its proposed structure are marked in grey. The first and most abundant dissociation is that of the Allyl alcohol initiator, leaving fragments with the H end group. In the first and third peak the loss of Allyl (C15’’) and AllylO (B15’’) are observable (marked in green). In the second peak only the loss of AllylO is present. The enhancement shows the spectra of the first and second peak in the m/z area between 500 and 790. Unfortunately the intensities for the third peak were two low to further interpret. For the first two peaks the most intense series B’’ and C’’ are labeled. These series also contain fragment ions with a remaining H end group. For both series the loss of just EO units, (-C2H4 for

every Cy-1’’ followed by -O for By-1’’) is marked in red. If just one or more PO units (-C3H6 for Cy-1’’ followed by -O for By1’’) are split off at any position, the signals are labeled in blue. All fragment ions with the AllylO end group remaining (B’ and C’) are labeled in purple, undistinguished of losses of EO’s or PO’s. Small amounts of the A series are labeled in pink. It becomes clear from these spectra, that for the second peak the first four monomer units after the AllylO initiator can only be EO units. Only after the fourth monomer a secession of a PO unit could occur. For the first peak the secession of a PO unit is possible at any position of the chain. The relatively lower intensities of PO dissociations are attributed to the overall 3:1 ratio of EO:PO. This is in good consensus with the absence of the C15’’ fragment in the second peak. It is possible that this fragment only occurs when the AllylO end group is connected to a PO unit. Dissociations from the other end of the copolymer are not visible in the second peak spectra at all. For the first peak, dissociations of both EO and PO (labeled purple) are present. This may be due to an accumulation of PO units at this end of the polymer chain for the species in the second peak. It is conceivable that this may lead to a stronger bonding in this part of the polymer, which then could not be broken in the MSMS experiment. Lower m/z areas of the spectra could not accurately be interpreted, due to the increase of possible mass isomers. With every dissociation the number of possible fragment ions, attributed to the various fragment series multiplied by the loss of different monomers in every position, increases. This leads to possible mass isomers for most of the lower m/z peaks. Only the overall assimilation of the spectra for both peaks can be noted. At lower m/z regions hardly any difference in the spectra can be seen. From this experiment, structures for the polymer chains in the first and second peak can be proposed. In Chart 2 possible polymer sequences for the two peaks are pictured. The parts shown in brackets are assumed to be truly statistical. Therefore polymer chains in the first peak are totally statistical at any monomer position. Polymers of the second peak consist of an inner block of four EO units bounded to the Allyl-alcohol initiator, the rest of the polymer chain is still statistical. This inner block could explain the higher retention time of these



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

polymer chains (see before). The same tendency could be observed for other polymerization grades and copolymer compositions in other samples. Although an analysis of the third peak was not possible it is likely that the blockiness of the copolymer sequence in these species further increases. Chart 2. Proposed structures for the first (1.6-1.9 min, bottom) and second (2.2-2.5 min, top) peak of m/z 841.8. Monomer units in brackets are presumed to be distributed randomly.

CONCLUSION The characterization of copolymers is becoming more and more complicated with increasing number and kind of distributions present in the samples. In most cases the separation by one distribution is superimposed by another. Even with 2DChromatography the separation might not be possible due to the lack of orthogonal separation processes for the different distributions. In this case, the mass spectrometric detection as an additional, none chromatographic dimension is an enhanced approach. We could demonstrate the existence and show a way to determine the critical conditions for chemical composition fractions of statistical copolymers. For the analysis of random EO-PO-copolymers we used the LCCC separation coupled with an ESI-TOF MS as detection technique to gain insight in the FTD, CCD, MSD and MMD of the samples. We could verify the existence of different end group fractions and approximate their amount in the samples. A theoretical approach to calculate both end group fraction intensities over the whole CCD and quantify their amounts was described and will be further developed. UPLC/MS/MS measurements showed, that the LCCC mode of chromatography can be especially valuable for the analysis of the MSD. With the described method the analysis time for each sample was less than four minutes. In the future this may lead to a way to quickly analyze complicated polymer samples without the need of complete chromatographic separations. Especially in combination with technologies like MSE, enhanced data processing might be able to compensate for lack of chromatographic separation.

Page 8 of 9

REFERENCES (1) Brun, Y. J. Liq. Chromatogr. Relat. Technol. 1999, 22, 30273065. (2) Herzberger, J.; Niederer, K.; Pohlit, H.; Seiwert, J.; Worm, M.; Wurm, F. R.; Frey, H. Chem. Rev. 2016, 116, 2170-2243. (3) Santacesaria, E.; Tesser, R.; Di Serio, M. Ind. Eng. Chem. Res. 2005, 44, 9482-9489. (4) Cox, M. F.; Weerasooriya, U. J. Surfactants Deterg. 1999, 2, 5968. (5) Santacesaria, E.; Felippone, F.; Iengo, P.; Vairo, G.; Di Serio, M. Ind. Eng. Chem. Res. 1996, 35, 3848-3853. (6) Pasch, H. Adv. Polym. Sci. 1997, 128, 1-43. (7) Gorbunov, A. A.; Ya Solovyova, L.; Skvortsov, A. M. Polymer 1998, 39, 697-702. (8) Trathnigg, B.; Malik, M. I.; Cuong, N. V.; Skvortsov, A. M. J. Chromatogr. A 2008, 1207, 122-129. (9) Chang, T. Adv. Polym. Sci. 2003, 163, 1-60. (10) Brun, Y.; Foster, P. J. Sep. Sci. 2010, 33, 3501-3510. (11) Gorshlov, A. V.; Much, H.; Becker, H.; Pasch, H.; V., E. V.; G., E. F. J. Chromatogr. 1990, 523, 91-102. (12) Philipsen, H. J. A. J. Chromatogr. A 2004, 1037, 329-350. (13) Trathnigg, B. Polymer 2005, 46, 9211-9223. (14) Falkenhagen, J.; Much, H.; Stauf, W.; Müller, A. H. G. Macromolecules 2000, 33, 3687-3693. (15) Trathnigg, B.; Gorbunov, A. Macromol. Symp. 2006, 237, 18-27. (16) Brun, Y. J. Liq. Chromatogr. Relat. Technol. 1999, 22, 30673090. (17) Zhu, Y.; Ziebarth, J.; Macko, T.; Wang, Y. Macromolecules 2010, 43, 5888-5895. (18) Uliyanchenko, E.; van der Wal, S.; Schoenmakers, P. J. Polym. Chem. 2012, 3, 2313. (19) Uliyanchenko, E.; Cools, P. J.; van der Wal, S.; Schoenmakers, P. J. Anal. Chem. 2012, 84, 7802-7809. (20) Gruendling, T.; Weidner, S.; Falkenhagen, J.; Barner-Kowollik, C. Polym. Chem. 2010, 1, 599. (21) Murgasova, R.; Hercules, D. M. Anal. Bioanal. Chem. 2002, 373, 481-489. (22) Gruendling, T.; Guilhaus, M.; Barner-Kowollik, C. Anal. Chem. 2008, 80, 6915-6927. (23) Falkenhagen, J.; Weidner, S. Anal. Chem. 2009, 81, 282-287. (24) Weidner, S. M.; Falkenhagen, J.; Maltsev, S.; Sauerland, V.; Rinken, M. Rapid Commun. Mass Spectrom. 2007, 21, 2750-2758. (25) Lattimer, R. P. J. Am. Soc. Mass Spectrom. 1994, 5, 1072-1080. (26) Lattimer, R. P. J. Am. Soc. Mass Spectrom. 1992, 3, 225-234. (27) Terrier, P.; Buchmann, W.; Desmazieres, B.; Tortajada, J. Anal. Chem. 2006, 78, 1801-1806. (28) Jackson, A. T.; Scrivens, J. H.; Williams, J. P.; Baker, E. S.; Gidden, J.; Bowers, M. T. Int. J. Mass Spectrom. 2004, 238, 287-297. (29) Altuntas, E.; Schubert, U. S. Anal. Chim. Acta 2014, 808, 56-69.

ASSOCIATED CONTENT Supporting Information Table of the m/z values used to create the SIC’s for the presumed copolymer structures in this article.

AUTHOR INFORMATION Corresponding Author *E-Mail: [email protected], Phone: +493081041632

ACKNOWLEDGMENTS We gratefully acknowledge Clariant Produkte (Deutschland) GmbH (Dr. V. Feldmann, Dr. W. Böse and J. Kapfinger) for continues support with provision of test samples.


Page 9 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


For Table of Contents Only:


Critical Conditions for Liquid Chromatography of Statistical

Recommend Documents