Anal. Chem. 2008, 80, 7778–7785
Direct Probe-Atmospheric Pressure Chemical Ionization Mass Spectrometry of Cross-Linked Copolymers and Copolymer Blends Sara E. Whitson,† Gabor Erdodi,‡ Joseph P. Kennedy,‡ Robert P. Lattimer,§ and Chrys Wesdemiotis*,† Department of Chemistry and Department of Polymer Science, The University of Akron, Akron, Ohio 44325, and Lubrizol Advanced Materials, Inc., 9911 Brecksville Road, Cleveland, Ohio 44141 Complex copolymers are heated to slowly increasing temperatures on a direct probe (DP) inside the plasma of the atmospheric pressure chemical ionization (APCI) source of a quadrupole ion trap. Slow heating allows for temporal separation of the thermal degradation products according to the stabilities of the bonds being cleaved. The products released from the DP are identified in situ by APCI mass spectrometry and tandem mass spectrometry. DP-APCI experiments on amphiphilic copolymers provide conclusive information about the nature of the hydrophobic and hydrophilic components present and can readily distinguish between copolymers with different comonomer compositions as well as between cross-linked copolymers and copolymer blends with similar physical properties. The dependence of DP-APCI mass spectra on temperature additionally reveals information about the thermal stability of the different domains within a copolymer. Electrospray ionization (ESI)1 and matrix-assisted laser desorption/ionization (MALDI)2,3 have made it possible to form intact gas-phase ions from most classes of synthetic polymers, enabling mass spectrometry (MS) analyses on such macromolecules.4-20 MS experiments provide the masses of the individual
oligomers contained in a polymer, from which important compositional and structural information about the polymer can be deduced, for example, its (co)monomer composition, end groups, compositional heterogeneity, and molecular weight distribution. Still, numerous synthetic polymers designed for important industrial or biomedical applications cannot be analyzed by MS, because they are too large or too polar to be dissolved (for ESI) or desorbed (for MALDI), or they are unable to form gas-phase ions due to the lack of functional groups that can attract and bind a charged particle, such as a proton or metal cation. Fortunately, most large or unionizable polymers become amenable to MS analysis by pyrolysis, i.e., thermal degradation.9,21-30 Depending on the thermal stability of the polymer and the temperature used (typically within 150-1000 °C), pyrolysis leads either to small fragments that can be ionized by electron ionization, chemical ionization (CI), field ionization, or photoionization, or to higher-mass pyrolyzates that can be subjected to ESI, MALDI, or both. The resulting mass spectra unveil the composition of the polymer sample. Molecular weight information is lost, but information about the thermal stability and degradation pathways of the polymer is gained. Rapid heating of the sample to a high temperature, via hightemperature Curie-point or resistive heating flash pyrolysis,
* To whom correspondence should be addressed. E-mail: wesdemiotis@ uakron.edu. Phone: 330-972-7699. Fax: 330-972-6085. † Department of Chemistry, The University of Akron. ‡ Department of Polymer Science, the University of Akron. § Lubrizol Advanced Materials, Inc. (1) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64–71. (2) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T. Rapid Commun. Mass Spectrom. 1988, 2, 151–153. (3) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299–2301. (4) Scrivens, J. H. Adv. Mass Spectrom. 1995, 13, 447–464. (5) Wu, K. J.; Odom, R. W. Anal. Chem. 1998, 70, 456A–461A. (6) Nielen, M. W. F. Mass Spectrom. Rev. 1999, 18, 309–344. (7) Scrivens, J. H.; Jackson, A. T. Int. J. Mass Spectrom. 2000, 200, 261–276. (8) Wesdemiotis, C.; Arnould, M. A.; Lee, Y.; Quirk, R. P. Polym. Prepr. 2000, 41 (1), 629–630. (9) Hanton, S. D. Chem. Rev. 2001, 101, 527–569. (10) Montaudo, G., Lattimer, R. P., Eds. Mass Spectrometry of Polymers; CRC Press: Boca Raton, FL, 2002. (11) McEwen, C. N.; Peacock, P. M. Anal. Chem. 2002, 74, 2743–2748. (12) Pasch, H.; Schrepp, W. MALDI-TOF Mass Spectrometry of Synthetic Polymers; Springer: Berlin, 2003. (13) McEwen, C. N. Adv. Mass Spectrom. 2004, 16, 215–227. (14) Arnould, M. A.; Polce, M. J.; Quirk, R. P.; Wesdemiotis, C. Int. J. Mass Spectrom. 2004, 238, 245–255.
(15) Peacock, P. M.; McEwen, C. N. Anal. Chem. 2004, 76, 3417–3428. (16) Montaudo, G.; Samperi, F.; Montaudo, M. S.; Carroccio, S.; Puglisi, C. Eur. J. Mass Spectrom. 2005, 11, 1–14. (17) Montaudo, G.; Samperi, F.; Montaudo, M. S. Prog. Polym. Sci. 2006, 31, 277–357. (18) Peacock, P. M.; McEwen, C. N. Anal. Chem. 2006, 78, 3957–3964. (19) Quirk, R. P.; Ocampo, M.; Polce, M. J.; Wesdemiotis, C. Macromolecules 2007, 40, 2352–2360. (20) Li, L. In MALDI MS. A Practical Guide to Instrumentation, Methods and Applications; Hillenkamp, F., Peter-Katalinic´, J., Eds. Wiley-VCH: Weinheim, 2007. (21) Wampler, T. P.; Levy, E. J. Analyst 1986, 111, 1065–1067. (22) Montaudo, G.; Puglisi, C. In Developments in Polymer Degradation-7; Grassie, N., Ed.; Elsevier Applied Science: London, 1987. (23) Lattimer, R. P. Rubber Chem. Technol. 1995, 68, 783–793. (24) Lattimer, R. P.; Polce, M. J.; Wesdemiotis, C. J. Anal. Appl. Pyrolysis 1998, 48, 1–15. (25) Carroccio, S.; Puglisi, C.; Montaudo, G. Macromol. Chem. Phys. 1999, 200, 2345–2355. (26) Lattimer, R. P. J. Anal. Appl. Pyrolysis 2000, 56, 61–78. (27) Lattimer, R. P. J. Anal. Appl. Pyrolysis 2001, 57, 57–76. (28) Tsuge, S.; Ohtani, H. In Mass Spectrometry of Polymers; Montaudo, G., Lattimer, R. P., Eds.; CRC Press: Boca Raton, FL, 2002; pp 113-147. (29) Montaudo, G.; Puglisi, C. In Mass Spectrometry of Polymers; Montaudo, G., Lattimer, R. P., Eds.; CRC Press: Boca Raton, FL, 2002; pp 181-236. (30) Lattimer, R. P. J. Anal. Appl. Pyrolysis 2003, 68-69, 3–14.
7778
Analytical Chemistry, Vol. 80, No. 20, October 15, 2008
10.1021/ac801198g CCC: $40.75 2008 American Chemical Society Published on Web 09/12/2008
degrades the sample to small molecules that can be separated and identified by GC/MS.28 Fast heating promotes secondary (consecutive) pyrolysis reactions, ultimately yielding thermally stable, nonreactive, compounds of lower molecular weight.21,22 Gradual heating, on the other hand, permits the observation of the primary pyrolysis products, formed in the early stages of thermal decomposition, which often have higher molecular weights, but may be prone to further decomposition.21,22 Gradual heating allows for temporal separation of the thermally desorbed components, with additives vaporizing first,23 followed by the thermal degradation products.24-27,29,30 If the polymer contains segments with different bond stabilities, pyrolyzate composition may change with temperature, helping to establish precise decomposition mechanisms.24-27,29,30 Gradual heating is ideally performed in situ, within a soft ionization source, without a GC, which might retain or destroy polar pyrolyzates. Since a mixture is ionized, a soft and sensitive ionization method, like CI, is preferable to minimize fragmentation and facilitate identification of the thermal degradation products; the loss of chromatographic separation is compensated for by tandem mass spectrometry (MS/ MS) experiments.26,27,30 Pyrolysis involving gradual heating and subsequent MS and MS/MS characterization of the pyrolyzates is a powerful method for the determination of composition, microstructure, and additives of industrial polymers, especially in unknown samples. Such studies have so far employed mostly sector tandem mass spectrometers, which are increasingly phased out by more robust and less costly time-of-flight- and trap-based instrumentation. Here, we report a simple modification of a commercial quadrupole ion trap to permit in situ pyrolysis of synthetic polymers inside an atmospheric pressure chemical ionization (APCI) ion source. The polymers are slowly heated on a direct probe (DP), and the thermally desorbed products and thermal decomposition products are identified by the resulting DP-APCI mass and tandem mass spectra. A similar technique, also using APCI to ionize solids desorbed from a direct probe, was recently introduced by McEwen and coworkers and named atmospheric pressure solids analysis probe (ASAP).31 ASAP has been employed to characterize volatile and semivolatile liquid or solid materials, desorbed from a glass capillary inserted into the heated nitrogen (desolvation gas) stream exiting an ESI or APCI source.31,32 In this mode, ASAP closely resembles the direct analysis in real time method,33 in which analytes are desorbed by energetically excited species produced in a gas stream. ASAP has also been used to detect the products from flash pyrolysis.34 In DP-APCI, we emphasize the advantages of combining slow heating of complex copolymers, to cause gradual degradation of their constituents according to their intrinsic thermal stabilities, with in situ ionization of the volatilized products at atmospheric pressure. This capability can easily be adapted to most modern mass spectrometers equipped with an atmospheric pressure ionization interface. DP-APCI and ASAP (31) McEwen, C. N.; McKay, R. G.; Larsen, B. S. Anal. Chem. 2005, 77, 7826– 7831. (32) McEwen, C.; Gutteridge, S. J. Am. Soc. Mass Spectrom. 2007, 18, 1274– 1278. (33) Cody, R. B.; Larame´e, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297– 2302. (34) McEwen, C. N. Abstracts of Papers, Pittcon Conference and Expo 2007, February 25-March 2, 2007, McCormick Place, Chicago, IL, Abstract 2602.
involve thermal desorption or thermal degradation of the analyte sample. In contrast, the related methods of desorption electrospray ionization35 and desorption atmospheric pressure chemical ionization35,36 involve desorption of the analyte(s) by a stream of droplets. EXPERIMENTAL SECTION All experiments were performed on a Bruker Esquire-LC quadrupole ion trap (Bruker Daltonics, Billerica, MA) equipped with an APCI source. The glass window of the source housing was removed to allow the insertion of a Chemical Data Systems Pyroprobe 100 (CDS Analytical, Inc. Oxford, PA), which was used to create thermal desorption and pyrolysis products. Total ion intensity was maximized by spraying solvent through the APCI probe at a flow rate of 0.4 mL/h with a Cole Parmer 74900 series syringe pump (Vernon Hills, IL) and by using a nebulizing gas (N2, 25 psi) flowing through a heated ceramic tube (425 °C), which surrounded the APCI nebulizing needle, and a desolvation gas (N2, 5 L/min, 300 °C).37 The exit of the ceramic tube was located ∼5 cm above the APCI corona discharge needle. The solvent flow rate was lower than that employed in conventional APCI experiments, whereas the nebulizing and desolvation (“drying”) gas conditions were identical to those used in APCI. A 99:1 mixture of methanol and water (v/v) that contained 1% (v/v) glacial acetic acid (EMD Chemicals, Gibbstown, NJ) served as the spraying solvent for the copolymers examined in this study; the acetic acid helped to increase the extent of protonation. Using 1 mg/mL solutions of sodium or potassium trifluoroacetate (Sigma Aldrich, Milwaukee, WI) in 99:1 methanol/water (v/v), on the other hand, helped to detect sodiated ions by increasing their intensities or by causing Na+/K+ exchange, respectively. The spectra shown are those acquired from acidified solutions. The copolymers investigated by DP-APCI were amphiphilic conetworks (APCNs), synthesized by the Kennedy group (vide infra).38-42 A small piece (2-3 mm2) of an APCN sample was softened by mild heating (∼100 °C) and attached to the platinum ribbon at the tip of a Pyroprobe, which was inserted into the APCI source and positioned ∼2 cm below the corona discharge needle and angled toward the capillary entrance to the ion trap. The probe temperature was gradually ramped from ∼100 to ∼700 °C at a rate of 30 °C/min, as mass spectra were acquired continuously. Ions were accumulated for 50 ms and scanned, and 50 scans were averaged per spectrum. Spectra were acquired continuously and averaged over 1-min intervals to determine the distributions of species desorbed from the sample with increasing temperature. The flow of the nebulizing gas exiting the heated ceramic tube (see above) was directed at the discharge needle (∼5 cm below the ceramic tube) and Pyroprobe tip (∼2 cm below the discharge (35) Taka´ts, Z.; Cotte-Rodriguez, I.; Talaty, N.; Chen, H.; Cooks, R. G. Chem. Commun. 2005, 1950–1952. (36) Cotte-Rodriguez, I.; Hernandez-Soto, H.; Chen, H.; Cooks, R. G. Anal. Chem. 2008, 80, 1512–1519. (37) Esquire-LC Operations Manual, version 3.1; Bruker Daltonik GmbH, 1999. (38) Erdodi, G.; Kennedy, J. P. J. Polym. Sci., A: Polym. Chem. 2007, 45, 295– 307. (39) Kang, J.; Erdodi, G.; Yalcin, B.; Cakmak, M.; Kennedy, J. P. Polym. Prepr. 2008, 49 (1), 824–825. (40) Kang, J.; Erdodi, G.; Kennedy, J. P. J. Polym. Sci., A: Polym. Chem. 2007, 45, 4276–4283. (41) Halford, B. Chem. Eng. News 2008, 86, 46–47. (42) Erdodi, G.; Kennedy, J. P. Manuscript in preparation.
Analytical Chemistry, Vol. 80, No. 20, October 15, 2008
7779
Figure 1. (a) Graft copolymer (PDMAAm-g-PDMS) and cross-linked APCN resulting after reaction of the graft with pentamethylcyclopentasiloxane. The hydrophilic PDMAAm domains are shown in blue and the hydrophobic PDMS domains in red. The cross-linking regions are shown in green. (b) Polyurethane (Elast-Eon) synthesized from methylene diisocyanate (MDI), dihydroxy PDMS, and 1,4-butanediol. The TP-APCNs investigated were blends of this polyurethane and the PDMAAm-g-PDMS graft shown in (a).
needle). Because of this layered arrangement, the heat provided to the sample by the nebulizing gas is minimal, and the actual pyrolysis temperature is practically equal to the Pyroprobe temperature. MS/MS experiments were conducted on select ions at a probe temperature that maximized their intensity. The precursor ion of interest was isolated in the trap and subjected to collisionally activated dissociation inside the trap by resonance excitation with a radio frequency (rf) field. The excitation time and rf amplitude were set to 50 ms and 0.60 V, respectively. Fifty scans were averaged per spectrum, and a chromatogram of such spectra was collected over 1 min and averaged again. RESULTS AND DISCUSSION Amphiphilic Conetworks Studied. The DP-APCI method was tested with APCNs composed of hydrophilic poly(N,Ndimethyl acrylamide) (PDMAAm) and hydrophobic poly(dimethylsiloxane) (PDMS) domains, cross-linked with a poly(methylhydrosiloxane) (PMHS),38 as well as with thermoplastic APCNs consisting of blends of amphiphilic block copolymers.42 The synthesis of these systems is briefly summarized in order to facilitate discussion of the acquired DP-APCI mass spectra; a detailed description of the synthetic procedure is reported elsewhere.38,42 The silane CH2dC(CH3)COOsCH2CH2CH2sSi(CH3)2sOs Si(CH3)2sH, an end-functionalizing agent (symbolized by MASiH), was prepared by hydrosilation of allyl methacrylate with tetramethyldisiloxane in the presence of Karstedt catalyst. Before purifying the product by distillation, triphenylphosphine was added to the charge to prevent oxidation of the Si-H groups to radicals, 7780
Analytical Chemistry, Vol. 80, No. 20, October 15, 2008
which would cause polymerization of MA-SiH. The asymmetrictelechelic macromonomer MA-PDMS-V, i.e., a poly(dimethylsiloxane) chain with methacrylate and vinyl end groups, was prepared subsequently by hydrosilation of commercially available V-PDMS-V with 1 equiv of MA-SiH. Symmetric-telechelic MAPDMS-MA is coproduced, while some of the V-PDMS-V starting material stays unreacted. This three-component mixture was subjected to free-radical polymerization with dimethyl acrylamide (DMAAm). DMAAm terpolymerizes with the MA-PDMS-V and MA-PDMS-MA macromonomers to produce a high molecular weight graft of a PDMAAm backbone carrying PDMS-V branches (Figure 1a). This graft is slightly cross-linked due to the presence of MA-PDMS-MA in the reaction mixture. Since vinyl groups do not copolymerize with acrylates under free-radical conditions, the V-PDMS-V component did not react at this stage, but remained as an admixture (cf. Figure 1a). In the final step, the PDMAAmg-PDMS graft and V-PDMS-V were cross-linked with a cyclic poly(methylhydrosiloxane) using Karstedt’s catalyst and traces of moisture to form the APCN depicted in Figure 1a. Using this procedure, APCNs with varying proportions of PDMS, PDMAAm, and cross-linking segments can be synthesized by changing the molar amounts of DMAAm, V-PDMS-V, and polyhydrosiloxane, respectively. The described APCNs are chemically cross-linked and have the biological, permeability, and mechanical properties desired for an immunoisolatory membrane.40,41 Most recently, physically cross-linked APCNs with similar properties have been developed by Kennedy and co-workers; these consist of a blend of thermoplastic (TP) elastomers.42 The TP-APCNs examined by DP-APCI
Table 1. Weight Percent Compositions of Copolymer Networks Studied
APCN-20 APCN-60 TP-APCN-42 TP-APCN-55
PDMAAm
PDMS
PMHSa
PUb
20 60 42 55
72 34 50 27
8 6 0 0
0 0 8 18
a The cyclic cross-linker pentamethylcyclopentasiloxane (cf. Figure 1a) was used for the conetworks analyzed. b Percent PU hard segment.
were blends of the commercially available polyurethane (PU) Elast-Eon with a PDMAAm-g-PDMS graft copolymer, prepared as described above (Figure 1a). Elast-Eon is produced from methylenediphenyl isocyanate (MDI), dihydroxy-terminated poly(dimethylsiloxane) (HO-PDMS-OH), and 1,4-butanediol,43 cf. Figure 1b. In a typical synthesis, the macrodiol HO-PDMS-OH is end-capped with excess MDI and the resulting prepolymer is chain-extended with butanediol.43 PUs contain hard and soft segments; in Elast-Eon, the MDI units linked by the chain extender constitute the hard and the PDMS chains constitute the soft segments. Four conetwork membranes were examined, two conventional APCNs and two TP-APCNs. The corresponding compositions are listed in Table 1. Conventional and thermoplastic APCNs will be discussed separately before interclass comparisons are made. Conventional APCNs. Chemically cross-linked APCN-20 started giving observable products when the pyrolysis probe reached ∼250 °C. Under the slow pyrolysis conditions used (vide supra), PDMAAm oligomers dominate up to ∼350 °C, as attested in Figure 2a by the DP-APCI mass spectrum obtained at 320 °C. Six different PDMAAm series with the expected 99-Da repeat unit are detected in the m/z 100-500 range; these have been marked A-F. The intensity of m/z 253 in distribution B was high enough to permit measurement of the corresponding MS/MS spectrum, which is depicted in Figure 3. The fragments formed agree well with the structure of a protonated trimer that has lost HN(CH3)2 from one side chain and carries no nominal end groups (i.e., the combined end groups have the same composition as the repeat unit), cf. inset of Figure 3. Hence, series B is assigned to such DMAAm oligomers. Intact PDMAAm oligomers without nominal end groups are also observed, cf. series A in Figure 2a and Table 2. These appear as [M + H]+ ions, except for the trimer, which is coproduced in protonated (mainly) and sodiated forms (m/z 298 and 320, respectively). If the solution sprayed through the APCI source is doped with potassium trifluoroacetate, 320 shifts to m/z 336 (Na+/ K+ exchange), confirming the given assignment. Series C-F, all of which are less abundant than A-B, correspond to PDMAAm oligomers with different end groups, as listed in Table 2. One of these series (C) represents [M + H]+ ions, and the remaining ones (D-F) represent [M + H HN(CH3)2]+ fragments. The series carry saturated or olefinic chain ends, depending on which bond in the PDMAAm segments was cleaved upon thermal degradation and how the incipient radical reacted further. (43) Gunatillake, P. A.; Martin, D. J.; Meijs, G. F.; McCarthy, S. J.; Adhikari, R. Aust. J. Chem. 2003, 56, 545–557.
The PDMAAm pyrolyzates can be accounted for by homolytic cleavages within the cross-linked PDMAAm chains, followed by typical radical site reactions, viz. H rearrangement (backbiting), bond scission in β position to the unpaired electron, and H atom abstraction from a nearby location (see Schemes S1-S2 in Supporting Information). Such reactions have been shown to occur during the thermal degradation of acrylic polymers and are discussed in detail in several published experimental and computational studies.30,44,45 Due to the high basicity of the dimethylamide group, CON(CH3)2,46 the PDMAAm oligomers released during thermal degradation into the APCI source ionize easily to [M + H]+ ions. Further, since the proton affinity of the CON(CH3)2 moiety is ∼150 kJ/mol higher than the proton affinity of the protonation reagent, CH3OH, the [M + H]+ ions are formed with sufficient internal energy to undergo extensively HN(CH3)2 loss, explaining the dominance of [M + H - HN(CH3)2]+ series in the DP-APCI mass spectra. Very few PDMS containing ions are observed at low pyrolysis temperature. As the temperature of the Pyroprobe is increased, however, PDMS distributions begin to appear with considerable abundances and dominate the DP-APCI spectra above 400 °C, as exemplified in Figure 2b. Six PDMS series, labeled G-L (74-Da repeat unit), could be identified and their likely compositions are summarized in Table 2. MS/MS spectra could not be measured, even for the most abundant PDMS ions, presumably because of higher dissociation thresholds as compared to PDMAAm ions. The m/z values of the PDMS distributions are consistent with cyclic degradation products ionized by proton (G) or methyl ion (I) attachment, and subsequent dissociation (H) or ion-molecule reactions (J, K, L) of the initially formed protonated species. These assignments are supported by the insensitivity of relative abundances on the K+ or Na+ content of the APCI solvent. PDMS pyrolysis below ∼700 °C has been found to proceed via Si-O bond translocations through ringlike transition states, not by homolytic bond cleavages;47-49 such reactions would create cyclic pyrolyzates, as indeed observed (cf. Scheme S3). The temperature dependence of DP-APCI mass spectra is sensitive to the conetwork composition and can be used to distinguish conetworks with different proportions of hydrophobic vs hydrophilic segments. For example, APCN-60, which has a 6 times higher PDMAAm/PDMS weight ratio than APCN-20, leads to DP-APCI mass spectra that do not change as dramatically with temperature as those measured for APCN-20. In agreement with the much higher PDMAAm content of APCN-60 (Table 1), its dominant decomposition products are PDMAAm oligomers (series A-F) even above 400 °C; in contrast, APCN-20 mainly produces PDMS oligomers at such temperatures, as attested in Figures 4a and 2b by the spectra measured at 470 °C. The difference in the observed product distributions from the two conetworks at 470 °C is definitely beyond the experimental reproducibility of relative abundances (±
