Block Lengths and Block Sequence of Linear Triblock and Glycerol

Peran Terrier, William Buchmann,* Bernard Desmazie` res, and Jeanine Tortajada. Laboratoire Analyse et Modélisation pour la Biologie et l'Environneme...
0 downloads 0 Views 293KB Size
Anal. Chem. 2006, 78, 1801-1806

Block Lengths and Block Sequence of Linear Triblock and Glycerol Derivative Diblock Copolyethers by Electrospray Ionization-Collision-Induced Dissociation Mass Spectrometry Peran Terrier, William Buchmann,* Bernard Desmazie`res, and Jeanine Tortajada

Laboratoire Analyse et Mode´ lisation pour la Biologie et l’Environnement, Universite´ d’Evry-Val d’Essonne, CNRS UMR 8587, Baˆ t. Maupertuis, Bd. F. Mitterrand, 91025 Evry Cedex, France

Chemical properties of ethylene oxide (EO) and propylene oxide (PO) block copolymers are strongly dependent on their sequence. Useful information about copolymer sequence can be obtained by tandem mass spectrometry (MS/MS). In this work, collision-induced dissociation (CID) of ammonium adducts of various linear triblock and glycerol derivative diblock copolyethers produced by electrospray ionization was studied under low-energy conditions. At first, homopolymers MS/MS spectra enabled us to identify the nature of the product ions and to suggest decomposition pathways. Then, it was shown that copolyethers with the same composition in each repeat unit but with inversed block sequences (i.e., PEO-b-PPOb-PEO vs PPO-b-PEO-b-PPO and gPEO-b-PPO vs gPPOb-PEO) can be easily distinguished with characteristic fragment ions. In the case of linear copolymers, CID spectra gave pertinent information about block lengths. Block copolymers of ethylene oxide (EO) and propylene oxide (PO) are widely used in the chemical industry as nonionic surfactants. Some of them have been found to be efficient vectors for drug and gene delivery.1 Global composition in each repeat unit, nature of end groups, and block sequence and lengths determine properties of these copolymers. Efficient analytical methods are thus required to control these features. Mass spectrometry, with the introduction and the development of soft ionization methods such as electrospray ionization (ESI) by Yamashita and Fenn,2 and matrix-assisted laser desorption/ ionization (MALDI) by Tanaka et al.3 and by Karas and Hillenkamp,4 has been shown to be a very powerful tool for polymer analysis.5-7 MALDI-time-of-flight (TOF) applied to copolymer * Corresponding author. E-mail: [email protected]. (1) Kabanov, A. V.; Batrakova, E. V.; Alakhov, V. Y. J. Controlled Release 2002, 82, 189-212. (2) Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4451-4459. (3) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, T. Rapid Commun. Mass Spectrom. 1988, 2, 151-153. (4) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-2301. (5) Montaudo, G.; Lattimer, R. P. In Mass Spectrometry of Polymers; Montaudo, G., Lattimer, R. P., Eds.; CRC Press: Boca Raton, FL, 2002. (6) Pasch, H.; Shrepp, W. In MALDI-TOF Mass Spectrometry of synthetic polymers; Barth, H. G., Pash, H., Eds.; Springer-Verlag: Berlin, Germany, 2003. 10.1021/ac051308h CCC: $33.50 Published on Web 02/07/2006

© 2006 American Chemical Society

analysis has been reported by several authors.5-8 This technique allows one to determine the nature of the repeat units and the end groups and to estimate the average molecular weights, the polymerization degrees, and the polydispersity indices. ESI-MS has been less studied due to the presence of several charge states. On the other hand, tandem mass spectrometry (MS/MS) of polymers7 and particularly polyethers and copolyethers9-25 has been reported a few times. An extensive fundamental work about ways of gas-phase fragmentation of polyethers has been reported by Lattimer et al.9-13 These authors studied collision-induced dissociation (CID) of poly(ethylene oxide) (PEO) ions generated by fast atom bombardment (FAB). Different types of fragment ions from cationized PEO precursors were observed according to the nature of the cation and the strength of translational energy. At low-energy CID, they proposed that protonated and lithiated (7) Montaudo, M. S. Mass Spectrom. Rev. 2002, 21, 108-144. (8) Terrier, P.; Buchmann, W.; Cheguillaume, G.; Desmazie`res, B.; Tortajada, J. Anal. Chem. 2005, 77, 3292-3300. (9) Lattimer, R. P. Int. J. Mass Spectrom. Ion Processes 1992, 116, 23-26. (10) Lattimer, R. P.; Mu ¨ nster, H.; Budzikiewicz, H. Int. J. Mass Spectrom. Ion Processes 1989, 90, 119-129. (11) Lattimer, R. P. J. Am. Soc. Mass Spectrom. 1992, 3, 225-234. (12) Selby, T. L.; Wesdemiotis, C.; Lattimer, R. P. J. Am. Soc. Mass Spectrom. 1994, 5, 1081-1092. (13) Lattimer, R. P. J. Am. Soc. Mass Spectrom. 1994, 5, 1072-1080. (14) Chen, R.; Li, L. J. Am. Soc. Mass Spectrom. 2001, 12, 832-839. (15) Chen, R.; Tseng, A. M.; Uhing, M.; Li, L. J. Am. Soc. Mass Spectrom. 2001, 12, 55-60. (16) Chen, R.; Yu, X.; Li, L. J. Am. Soc. Mass Spectrom. 2002, 13, 888-897. (17) Stolarzewicz, A.; Neugebauer, D.; Silberring, J. Rapid Commun. Mass Spectrom. 1999, 13, 2469-2473. (18) Hanton, S. D.; Parees, D. M.; Owens, K. G. J. Mass Spectrom. 2004, 238, 257-264. (19) Pastor, S. J.; Wilkins, C. L. Int. J. Mass Spectrom. Ion Processes 1998, 175, 81-92. (20) Yalcin, T.; Gabryelski, W.; Li, L. Anal. Chem. 2000, 72, 3847-3852. (21) Jackson, A. T.; Yates, H. T.; Scrivens, J. H.; Critchley, G.; Brown, J.; Green, M. R.; Bateman, R. H. Rapid Commun. Mass Spectrom. 1996, 10, 16681674. (22) Bottrill, A. R.; Giannakopulos, A. E.; Waterson, C.; Haddleton, D. M.; Lee, K. S.; Derrick, P. J. Anal. Chem. 1999, 71, 3637-3641. (23) 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. (24) Przybilla, L.; Francke, V.; Ra¨der, H. J.; Mu ¨ llen, K. Macromolecules 2001, 34, 4401-4405. (25) Cerda, B. A.; Horn, D. M.; Breuker, K.; McLafferty, F. W. J. Am. Chem. Soc. 2002, 124, 9287-9291.

Analytical Chemistry, Vol. 78, No. 6, March 15, 2006 1801

PEO mainly dissociate via charge-initiated decompositions and to a lesser extent via charge-remote 1,4-H elimination. Rearrangement with loss of repeat units was mentioned.9-11 Almost no fragmentation was observed from sodium or potassium adducts.10 At high-energy CID, protonated PEO undergoes charge-induced decompositions, whereas metalated PEO mainly undergoes chargeremote 1,4-H elimination.12 Similar results were found for poly(propylene oxide) (PPO).10,11 Chen and Li compared Lattimer’s FAB-CID experiments with their own ESI-CID experiments for PEO methyl ether with proton, ammonium, lithium, and silver adducts.14 Lithium and silver adducts gave both satisfactory MS and MS/MS spectra. Ammonium adducts underwent a little insource dissociation and dissociated easily under low-energy CID by a presumed charge-remote dissociation from the protonated oligomer. Protonated PEO ions underwent large in-source dissociations. MS/MS experiments have been carried out on various polyethers in order to get structural information. First, they permitted the determination of end groups of homopolyethers. Several combinations of ion sources and dissociation techniques have been used for this purpose: ESI-low-energy CID (Li et al.15,16 and Stolarzewicz et al.17), MALDI-PSD (Hanton et al.18), MALDI-SORICID (Pastor and Wilkins19), ESI-high-energy CID (Li et al.20), MALDI-high-energy CID (Jackson et al.,21 Derrick et al.22). Tandem mass spectrometry can be applied to copolyethers too. Lattimer obtained structural information from an ethylene and propylene oxide diblock using FAB-low-energy CID experiments.13 Jackson et al., from ESI-low-energy CID experiments, distinguished random from diblock copolyethers, identified end groups, and determined the sequence of diblock oligomers bearing a fatty chain end.23 Ra¨der et al. demonstrated that MALDI-PSD can allow the determination of the exact block lengths of an ethylene oxide and phenylene ethynylene diblock.24 McLafferty et al. showed that rearrangement in CID experiments could lead to an underestimation of block lengths and applied the electron capture dissociation method in order to sequence diblock and triblock with short external blocks copolyethers.25 Nevertheless, the sequence of triblock copolyethers or branched diblock copolyethers remains very difficult to determine. In this work, tandem mass spectrometry was evaluated as a method to determine block lengths and block sequence of various copolyethers. First, we briefly described gas-phase dissociation of PEO and PPO homopolymers, and then we studied dissociation of linear triblock copolymers (PEO-block-PPO-block-PEO and PPOblock-PEO-block-PPO) and glycerol derivative diblock copolymers (glycerol ethoxylate-block-propoxylate and glycerol propoxylateblock-ethoxylate). It will be shown from the results that oligomers with the same composition in each repeat unit but with inversed sequence can be distinguished and that information about block lengths can be obtained. EXPERIMENTAL SECTION Materials. Hydroxyl-terminated homopolymers PEO Mn ) 1000 g‚mol-1, PPO Mn ) 1200 g‚mol-1, and linear triblock copolymers (PEO-b-PPO-b-PEO Mn ) 1900 g‚mol-1 and PPO-bPEO-b-PPO Mn ) 2000 g‚mol-1, both containing 50% w/w PEO) were obtained from Sigma-Aldrich (Saint Quentin-Fallavier, France). To confirm the triblock structure of these copolymers, 13C NMR, 1H NMR, and 2D 1H NMR experiments were carried out. As 1802

Analytical Chemistry, Vol. 78, No. 6, March 15, 2006

Figure 1. ESI-CID mass spectrum of the EO20 ammonium adduct. Collision energy, 35 eV.

Chart 1. Structure of a Linear Triblock Copolymer (PEO-block-PPO-block-PEO) and a Glycerol Derivative Diblock Copolymer (Glycerol PEO-block-PPO)

expected, in both cases, only one type of end group was observed. Hydroxyl-terminated glycerol derivative diblock copolymers (glycerol ethoxylate-block-propoxylate, gPEO-b-PPO, glycerol propoxylate-block-ethoxylate, gPPO-b-PEO, both with Mn ) 1000 g‚mol-1 and containing 60% w/w PEO) were given by Henkel-Nopco (St Fargeau, France). An example of both kinds of copolymers is shown in Chart 1. Solutions of polymer (with a concentration between 10-4 and 10-3 M according to its nature) in H2O/CH3OH 50/50 (v/v) were prepared prior to mass analysis. D2O/CH3OD 50/50 (v/v) was used instead of H2O/CH3OH during H/D exchange experiments. No salt was added. ESI-MS/MS. All MS/MS experiments were carried out using the quadrupole-TOF mass spectrometer Q-Star Pulsar equipped with an electrospray ionization source (Applied Biosystems, Foster City, CA). Polymer solutions were injected at either 2 or 5 µL/ min according to the nature of the polymer. Capillary voltage was 5.5 kV; declustering and focusing potential were respectively 50 and 225 V. Collision energy was between 35 and 71 eV. It was adjusted in order to obtain a parent ion and a most intense product ion with similar intensities. N2 was used as collision gas with a 3 psi pressure. RESULTS AND DISCUSSION Gas-Phase Dissociation of PEO and PPO. Prior to the study of copolymers, CID spectra of PEO and PPO homopolymers were recorded under low-energy conditions. The MS/MS spec-

Scheme 1. Possible Mechanisms Involved in the Gas-Phase Fragmentation of Protonated PEO

trum in Figure 1 shows the fragmentation of a PEO chain. The precursor ion (m/z 916.6) is the ammonium adduct of the hydroxyl-terminated 20-mer of ethylene oxide ((EO)20NH4+), selected among the most intense ions of the MS spectrum (not shown). The intense ion at m/z 899.6 is the protonated oligomer ((EO)20H+) resulting from a loss of ammonia. All the other ions are fragments of (EO)20H+. The most intense of them are in the lower half of the mass range. Several series of fragment ions can be distinguished depending on their end groups. We suggest that

Figure 2. ESI-CID mass spectra of ammonium adduct of triblock copolymers (a) PO7.5EO15PO7.5 and (b) EO7.5PO15EO7.5. Collision energy, 55 eV.

different kinds of mechanisms are involved in the formation of these ions. First, a charge-induced dissociation from MH+ as described by Lattimer et al.9,10 (Scheme 1a) generates fragment ions with a hydroxyl function as an end group and a carbonium as the other (labeled OH,C+(EO)n). The most abundant series apparently results from this mechanism, but a part of it can correspond to internal fragments produced by a more complex fragmentation process. As a piece of evidence, in the MS/MS spectrum of deuterated ammonium adduct of EO20 (data not shown), a part of this series has lost both deuterium atoms from end groups. This can be explained by H/D exchanges only, as described by Lattimer for FAB-generated EO10D+.9 However, experiments with triblock copolyethers (described below) give evidence of the loss of both terminal parts, leading to internal OH,C+(EO)n fragment ions. Although dissociations are produced under low-energy conditions, charge-remote dissociations from OH,C+(EO)n seem to occur. Sixmembered intermediates are likely to be involved. Two examples of six-center mechanisms are given on Scheme 1b and c. The main series (OH,C+(EO)n) may partially result from a mechanism such as the one described in Scheme 1b. The less intense series, which consists of oligomers with carbonium and aldehyde functions as end groups, labeled CHO,C+(EO)n, may result from the same mechanism with the charge on the other end or from a mechanism such as the one described in Scheme 1c. Some minority ions would be the other products of these charge-remote dissociations. For example, a series of fragments with carbonium and ethylene as end groups (denoted ),C+(EO)n, not labeled in Figure 1) can be produced by the mechanism 1c with the charge on the other end of the chain. It must be noted that the ions of the main series can result from both charge-induced and chargeremote dissociation, but it is difficult to evaluate the extent of ions coming from each way. The last proposed mechanism is a six-center charge-induced dissociation from MH+ as described by Lattimer for Li adducts12 (Scheme 1d). It can produce the secondary series, which consists of hydroxyl-terminated fragment ions and is labeled OH,OH(EO) H+. n Analytical Chemistry, Vol. 78, No. 6, March 15, 2006

1803

Figure 3. ESI-CID mass spectra of ammonium adduct of triblock copolymers (a) PO8.5EO17PO8.5, (b) EO8.5PO17EO8.5, (c) PO10EO17PO10, (d) EO8.5PO20EO8.5, (e) PO8.5EO20PO8.5, (f) EO10PO17EO10, (g) PO10EO20PO10, and (h) EO10PO20EO10. Collision energy: 60 eV for (a) and (b), 65 eV for (c-f), and 71 eV for (g) and (h).

The MS/MS spectrum of a propylene oxide oligomer gives similar results (data not shown). Nevertheless, minor variations in relative intensities of each series of product ions suggest that the mechanisms involved in dissociation are in proportion different from that for PEO. Gas-Phase Dissociation of Linear Triblocks. Ammonium adducts of hydroxyl-terminated EO/PO linear triblock oligomers with the same composition in each repeat unit, but with inverse supposed block sequence, were selected among ions produced by ESI (MS spectra not shown) and dissociated under the same CID conditions. This experiment has been performed with five pairs of oligomers: EO15/PO15 (PO7.5EO15PO7.5 vs EO7.5PO15EO7.5) (Figure 2a,b), EO17/PO17 (Figure 3a,b), EO17/PO20 (Figure 3c,3d), EO20/PO17 (Figure 3e,f), and EO20/PO20 (Figure 3g,h). In Figure 1804 Analytical Chemistry, Vol. 78, No. 6, March 15, 2006

2, the main series are labeled. In Figure 3, the relative intensities of each series are similar and only their names are given. In each spectrum, we can observe the parent ion MNH4+, an intermediary MH+, and several series of fragment ions. These product ions are supposed to be the result of similar fragmentation mechanisms that are previously described in the case of homopolymers. For the assignment of the series, every kind of fragment observed with homopolymers (even the very minority ones) with every possible combination of numbers of each repeat unit has been considered. At first, under the same collision activation, the parent ion/ product ions ratios are similar for both PO-EO-PO and EOPO-EO spectra. However, PO-EO-PO fragments cover a larger mass range than EO-PO-EO fragments.

Figure 4. ESI-CID mass spectra of ammonium adduct of glycerol derivative diblock copolymers (a) gEO11PO6 and (b) gPO6EO11. Collision energy, 35 eV.

For each pair, the two triblock oligomers can be distinguished by the most intense fragment series. They consist of ions separated by 44 atomic mass units (amu) i.e., one EO repeat unit in the case of PO-EO-PO (see Figures 2a and 3a,c,e,g) and by 58; i.e., one PO repeat unit in the case of EO-PO-EO (see Figures 2b and 3b,d,f,h). These ions respectively labeled OH,C+(EO)2fn and OH,C+(PO)1fn are fragments of the internal block (which is the longest block) with a hydroxyl function as an end group and a carbonium as the other. It is worth noting that, assuming its triblock sequence, MH+ has lost both its external blocks to produce these ions. Therefore, as descibed for PEO, MH+ has undergone at least two successive dissociations (possibly a charge-induced dissociation followed by a charge-remote one). For the heaviest ions, a contribution of isobaric series (such as OH,OH(PO) (EO) 5 0fn in PO-EO-PO spectra) is possible, but these fragments are likely to be minority because this kind of fragment ion is of minor importance with homopolymers and because of the supposed sequence of the oligomer. Moreover, a major difference between the two spectra is the presence, in each POEO-PO spectrum only, of a series that very likely consists of an intact internal block alone and a part of the external blocks: OH,C+(EO)15,17or20(PO)0fn (see Figures 2a, 3a, c, e, g). In all PO-EO-PO spectra, the main series (OH,C+(EO)2fn) shows a maximum number of repeat unit equal to the size of the internal block. The series that consists of fragments of the external blocks (OH,C+(PO)1fn) shows a maximum number of repeat units smaller than the average size of the external block. In EO-PO-EO spectra, the maximum number of repeat units of

the OH,C+(PO)1fn series is smaller than the size of the internal block while the one of the OH,C+(EO)1fn is sometimes slightly higher than the average size of the external block. Globally, these maximum numbers of repeat units are in good agreement with a triblock structure. These results show that MS/MS can give information about block lengths even if PO block lengths are underestimated. However, these values have to be considered with care because of the possible contribution of isobaric fragment ions. On the other hand, in the case of PO-EO-PO, the abovementioned series OH,C+(EO)15,17,or20(PO)1fn allows the identification of the exact length of the internal block. Indeed, the EO internal block always remained intact in this series. Unfortunately, the EO-PO-EO spectrum does not supply this information. The presence of intact internal block for PO-EO-PO in contrast with EO-PO-EO and the underestimation of PO block lengths in contrast with EO suggest that dissociations occur preferentially in PO blocks whatever the sequence of copolymer. The weakness of PO blocks compared to EO blocks could explain why the distinction between both sequences is possible. Gas-Phase Dissociation of Glycerol Derivative Diblocks. In a similar manner as previously, two hydroxyl-terminated EO/ PO glycerol derivative diblock oligomers with the same composition in each repeat unit, but with supposed inverse sequence (gEO11PO6 vs gPO6EO11), were selected from the ESI-produced ions (MS spectra not shown) and dissociated under the same CID conditions (Figure 4a and b). As with linear oligomers, the spectra contain the parent ion MNH4+, an intermediary MH+, and several series of fragment ions. In contrast to linear copolymers, the sequence of glycerol derivative copolyethers is not easily inferred from the spectra. Indeed, overlapping and difficulty to assign ions are more frequent with glycerol derivatives. For example, the main series in both spectra could be entirely or partially assigned to OH,C+(EO)n, OH,CHO(PO)2(EO)nH+, gOH,CHO,CHO(EO)nH+, and gOH,OH,C+(PO)1(EO)n (only the more likely assignments are given). Several kinds of fragments are likely to make up this series. In this case, the determination of the block lengths and the nature of end group is very difficult. In both cases, the main series is constituted of fragment ions separated by 44 amu, i.e., one EO repeat unit. As with linear oligomers, main series ions come from the longest block. These ions do not allow any distinction between the two copolymers. On the other hand, in Figure 4a, a series that is absent from Figure 4b can be observed. This series is likely to consist of intact glycerol linked to internal EO blocks with a part of external PO blocks (gOH,OH,C+EO11PO1f6). It is characteristic of gEO11PO6: no equivalent series is observed in the gPO6EO11 spectrum (gOH,OH,C+PO6EO1f11 would be between m/z 467 and 907). The analogy with linear copolymer spectra is obvious: in both cases, internal EO blocks remain intact whereas internal PO blocks are dissociated. Moreover, each oligomer has a characteristic series: fragment ions separated by 44 amu (one EO repeat unit) for gEO-PO and by 58 amu (one PO repeat unit) for gPO-EO. These ions are likely to be glycerols linked to a part of internal blocks: gOH,),C+(EO)1f8 for gEO-PO and gOH,),C+(PO)1f5 for gPO-EO. To conclude, tandem mass spectrometry experiments permit the unambiguous distinction of linear or glycerol derivative Analytical Chemistry, Vol. 78, No. 6, March 15, 2006

1805

copolyethers with the same composition in repeat units but with different block sequences. Information on block lengths can be obtained for linear triblock copolymers. Its accuracy depends on the nature of the repeat units (EO or PO) and on their relative location along the chain (internal or external). In our case, perfectly accurate block length information is provided for internal EO blocks, whereas the length of internal PO blocks is underestimated. In the same way, average length of external EO block is slightly overestimated, whereas that of PO blocks is underestimated. The weakness of PO blocks compared to the EO blocks seems to direct the fragmentation of the copolymer chain. These

1806

Analytical Chemistry, Vol. 78, No. 6, March 15, 2006

experiments enable us to hope for sequence determination of other complex copolymers. ACKNOWLEDGMENT Financial support from the Association Franc¸ aise contre les Myopathies (AFM) is gratefully acknowledged.We thank Renaud Perrin, Laboratoire Mate´riaux Polyme`res aux Interfaces, Universite´ d’Evry, for NMR studies. Received for review July 22, 2005. Accepted December 28, 2005. AC051308H