Ionization Mass Spectrometry of

Desorption/Ionization Mass Spectrometry of Discrete Mass Poly(butylene glutarate) Oligomers. John B. Williams, Toby M. Chapman*, and David M. Herc...
0 downloads 0 Views 171KB Size
Download PDF
Anal. Chem. 2003, 75, 3092-3100

Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry of Discrete Mass Poly(butylene glutarate) Oligomers John B. Williams,† Toby M. Chapman,*,† and David M. Hercules‡

Departments of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260 and Vanderbilt University, Nashville, Tennessee 37235

The mass dependency of matrix-assisted laser desorption/ ionization time-of-flight mass spectrometry (MALDI-TOFMS) response has been studied using equimolar mixtures of synthetic discrete mass poly(butylene glutarate) (PBG) oligomers of known structure having degrees of polymerization of 8, 16, 32, and 64. Mass discrimination observed was attributed to choice of matrix and detector saturation caused by higher laser intensity and inclusion of matrix ions in the MALDI spectra. Optimization of sample preparation and instrumental parameters provided uniform response over the mass ranged spanned by these four oligomers. The oligomer mixture was shown to serve as a model of more complex polymer distributions in the mass range 780-6000 Da, and application of the discrete mass oligomers as internal and calibration standards was demonstrated. Inclusion of PBG discrete mass oligomers as an internal standard in a quasi-equimolar mixture with polydispersed poly(butylene adipate) (PBA) indicated that some diminution of response occurred during the analysis of this mixture of materials. Reasons for differences in the corrected molecular weight averages of the polydispersed PBA obtained from measurements using MALDI and GPC were studied using individual discrete mass oligomers as calibration standards for GPC. The data indicated that differences in hydrodynamic volumes of PBG oligomers and PEG standards at similar masses resulted in an overestimation by GPC of the molecular weight averages of the PBA distribution. Over the past decade, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS)1-5 has become an important technique for simultaneous determination of structural information such as repeat unit and end group masses, molecular weight averages, and purity of synthetic * Corresponding author. E-mail: [email protected]. † University of Pittsburgh. ‡ Vanderbilt University. (1) Tanaka, K.; Waki, H.; Ido, Y.; Yoshido, Y.; Yoshido, T. Rapid Commun. Mass Spectrom. 1988, 2, 151. (2) Karas, M.; Hillenkamp, K. Anal. Chem. 1988, 60, 2299. Bahr, U.; Deppe, A.; Karas, M.; Hillenkamp, K.; Giessman, U. Anal. Chem. 1992, 64, 2866. (3) Hillenkamp, F.; Karas, M.; Beavis, R. C.; Chait, B. T. Anal. Chem. 1991, 63, 1193. (4) Beavis, R. C.; Chait, B. T. Anal. Chem. 1990, 62, 1836. (5) Creel, H. S. Trends Polym. Sci. 1993, 1, 336.

3092 Analytical Chemistry, Vol. 75, No. 13, July 1, 2003

polymers.6-26 The soft ionization technique, in combination with time-of-flight mass spectrometry, allows the desorption and mass measurement of intact oligomers of large, nonvolatile polymers (up to 106 Da)26 without the necessity of chemical modification or reduction in molecular weight prior to analysis. The continuing development of MALDI methods has expanded its analytical application to a wide range of chemically dissimilar synthetic materials25 including homopolymers, copolymers, dendrimers, resins, surfactants, and fullerenes. In contrast to relative techniques of polymer characterization, such as viscometry and gel permeation chromatography (GPC), absolute molecular weights and molecular weight distributions can be determined from MALDI spectra. However, molecular weight values provided by MALDI-TOF-MS spectra have been shown to agree with results (6) Danis, P.; Karr, D.; Mayer, F.; Holle, A.; Watson, C. Org. Mass Spectrom. 1992, 27, 843. (7) Danis, P.; Karr, D. E.; Westmoreland, D. G.; Piton, M. C.; Christie, D. I.; Clay, P. A.; Kable, S. H.; Gilbert, R. G. Macromolecules 1993, 26, 6684. (8) Danis, P.; Karr, D. Org. Mass Spectrom. 1993, 28, 923. (9) Bu ¨ rger, M.; Mo`ller, H.; Seebach, D.; Bo ¨rnsen, O.; Scha¨r, M.; Widmer, H. M. Macromolecules 1993, 26, 4783. (10) Montaudo, G.; Montaudo, M. S.; Puglisi, C.; Samperi, F. Macromolecules 1995, 28, 4562. (11) Cottrell, J. S.; Koerner, M.; Gerhards R. Rapid Commun. Mass Spectrom. 1995, 9, 1562. (12) Danis, P. O.; Karr, D. E.; Simonsick, W. J.; Wu, D. T. Macromolecules 1995, 28, 1229. (13) Pasch, H.; Gores, F. Polymer 1995, 10, 1999. (14) Danis, P. O.; Karr, D. E.; Xiong, Y.; Owens, K. G. Rapid Commun. Mass Spectrom. 1996, 10, 862. (15) Wilczek-Vera, G.; Danis, P. O.; Eisenberg, A. Macromolecules 1996, 29, 4036. (16) Williams, J. B.; Gusev, A. I.; Hercules, D. M. Macromolecules 1996, 29, 8144. (17) Chaudhary, A. K.; Critchley, G.; Abderhammane, D.; Beckman, E. J.; Russell, A. J. Macromolecules 1996, 29, 2213. (18) Belu, A. M.; DeSimone, J. M.; Linton, R. E.; Lange, G. W.; Friedman, R. M. J. Am. Soc. Mass Spectrom. 1996, 7, 11. (19) Scamporrino, E.; Vitalini, D.; Mineo, P. Macromolecules 1996, 29, 5520. (20) Williams, J. B.; Gusev, A.; Hercules, D. M. Macromolecules 1997, 30, 3781. (21) Laine, O.; O ¨ sterholm, H.; Ja¨rvinen, H.; Wickstro ¨m, K.; Vainiotalo, P. Rapid Commun. Mass Spectrom. 2000, 14, 482 (22) Ganachaud, F.; Monteiro, M. J.; Gilbert, R. G.; Dourges, M.; Thang, S. H.; Rizzardo, E. Macromolecules 2000, 33, 6738. (23) Mehl, J. T.; Murgasova, R.; Dong, X.; Hercules, D. M.; Nefzger, H. Anal. Chem. 2000, 72, 2490. (24) Blais, J. C.; Tessier, M.; Bolbach, G.; Remaud, B.; Rozes, L.; Guittard, J.; Brunot, A.; Mare´chal, E.; Tabet, J. C. Int. J. Mass Spectrom. Ion Processes 1995, 144, 131. (25) Neilen, M. W. F. Mass Spectrom. Rev. 1999, 18, 309. (26) Schriemer, D. C.; Li, L. Anal. Chem. 1996, 68, 2721. 10.1021/ac030061q CCC: $25.00

© 2003 American Chemical Society Published on Web 05/17/2003

obtained from conventional methods (usually GPC) only for polymers with narrow polydispersities (Mw/Mn < 1.2).27 Many studies27-46 have indicated that MALDI-TOF-MS spectra of polymers of higher polydispersities systematically diminish the relative contribution from components having higher degrees of polymerization, thus providing erroneously low molecular weight averages. However, overestimation of higher mass components of polymer distributions by MALDI has also been reported.47 Recent efforts to improve MALDI detection of the higher mass region of polymer distributions have been focused on coupling of GPC with MALDI.48-58 In these experiments, the column eluent is fractionally or continuously deposited with or onto a MALDI matrix for sequential mass spectrometric analysis. Separations by GPC of the higher mass region of a polydispersed material into more narrow distributions have been shown to improve MALDI detection of these oligomers and further indicate MALDI may provide erroneously low molecular weight averages for more polydispersed materials. To study mass dependency of the MALDI response, uniform oligomers of poly(methyl methacrylate)59,60 and polystyrene61 have been isolated using preparative supercritical fluid chromatography (27) McEwen, C.; Jackson, C.; Larson, B. S. Int. J. Mass Spectrom. Ion Processes 1997, 160, 387. (28) Lehrle, R. S.; Sarson, D. S. Rapid Commun. Mass Spectrom. 1995, 9, 91. (29) Hagelin, G.; Arukwe, J. J.; Kasparkove, V.; Nordbø, S. Rapid Commun. Mass Spectrom. 1998, 12, 25. (30) Farmer, T. B.; Caprioli, M. J. Mass Spectrom. 1995, 30, 1245. (31) Montaudo, G.; Montaudo, M. S.; Puglisi, C.; Samperi, F. Rapid Commun. Mass Spectrom. 1995, 9, 453. (32) Jackson, C.; Larson, B.; McEwen, C. Anal. Chem. 1996, 68, 1303. (33) Just, U.; Holzbauer, H.; Resch, M. J. Chromatogr., A 1994, 667, 354. (34) Axelsson, J.; Scrivner, E.; Haddleton, D. M.; Derrick, P. J. Macromolecules 1996, 29, 8875. (35) Martin, K.; Spickermann, J.; R×c2der, H. J.; Mo`llen, K. Rapid Commun. Mass Spectrom. 1996, 10, 1471. (36) Schriemer, D. C.; Li, L. Anal. Chem. 1997, 69, 4169. (37) Schriemer, D. C.; Li, L. Anal. Chem. 1997, 69, 4176. (38) Chen, H.; Guo, B. Anal. Chem. 1997, 69, 4399. (39) Rashidzadeh, H.; Guo, B. Anal. Chem. 1998, 70, 131. (40) Byrd, H. C. M.; McEwen, C. N. Anal. Chem. 2000, 72, 4568. (41) Marie, A.; Fournier, F.; Tabet. J. C. Anal. Chem. 2000, 72, 5106. (42) Tang, X.; Dreifuss, P. A.; Vertes, A. R Rapid Commun. Mass Spectrom. 1995, 9, 1141. (43) Rashidzadeh, H.; Wang, Y.; Guo, B. Rapid Commun. Mass Spectrom. 2000, 14, 493. (44) Whittal, R. M.; Schriemer, D. C.; Li, L. Anal. Chem. 1997, 69, 2734. (45) Chen, R.; Zhang, N.; Tseng, A. M.; Li, L. Rapid Commun. Mass Spectrom. 2000, 14, 2175. (46) Montaudo, G.; Garozzo, D.; Montaudo, M. S.; Puglisi, C.; Samperi, F. Macromolecules 1995, 28, 7983. (47) Hanton, S. D.; Clark, P. A. C.; Owens, K. G. J. Am. Soc. Mass Spectrom. 1999, 10, 104. (48) Montaudo, G.; Montaudo, M. S.; Puglisi, C.; Samperi, F. Rapid Commun. Mass Spectrom. 1995, 9, 1158. (49) Fei, X.; Murray, K. K. Anal. Chem. 1996, 68, 3555. (50) Montaudo, G.; Scamporrino, E.; Vitalini, D.; Mineo, P. Rapid Commun. Mass Spectrom. 1996, 10, 1551. (51) Nielen, M. W. F.; Malucha, S. Rapid Commun. Mass Spectrom. 1997, 11, 1194. (52) Pasch, H.; Rode, K. J. Chromatogr., A 1995, 699, 21. (53) Kassis, C. E.; DeSimone, J. M.; Linton, R. W.; Remsen, E. E.; Lange, G. W.; Friedman, R. M. Rapid Commun. Mass Spectrom. 1997, 11, 1134. (54) Nielen, M. W. F. Anal. Chem. 1998, 70, 1563. (55) Puglisi, C.; Samperi, F.; Carroccio, S.; Montaudo, G. Rapid Commun. Mass Spectrom. 1999, 13, 2260. (56) Puglisi, C.; Samperi, F.; Carroccio, S.; Montaudo, G. Rapid Commun. Mass Spectrom. 1999, 13, 2268. (57) Hanton, S. D.; Liu, X. M. Anal. Chem. 2000, 72, 4550. (58) Carroccio, S.; Rizzarelli, P.; Puglisi, C. Rapid Commun. Mass Spectrom. 2000, 14, 1513.

(SFC) and have served as models for polymer distributions in the 300-5000-Da mass range. The results indicated that the relative intensities of peaks calculated from the MALDI spectra of equimolar mixtures of these oligomers do not vary significantly with mass under optimized experimental conditions. Uniform oligomers of nylon-6 having degrees of polymerization of 4, 6, 8, and 12 have been prepared by stepwise synthesis62 and were utilized to study fragmentation processes in time-of-flight secondary ion mass spectrometry (TOF-SIMS),62 applied as models for the analysis of nylon-6 polymers using electrospray ionization mass spectrometry and MALDI,63 and studied using MALDI-collisioninduced dissociation to obtain structural information and investigate fragmentation mechanisms.64 The purpose of the present work was to study and quantify discrepancies observed in the MALDI-TOF-MS mass determinations of synthetic polymers. An equimolar mixture of discrete molecular mass oligomers of perfectly alternating copolymers of 1,4-butanediol and glutaric acid (poly(butylene glutarate)) having degrees of polymerization of 8, 16, 32, and 64, spanning a mass range of 760-6000 amu, has been prepared for application as a model of more complex distributions produced by standard polymerization techniques. This model was used to investigate the effects of matrix, laser intensity, and detector saturation on the degree of mass dependency of response observed during MALDI analysis. Determination of the relative influence of these parameters allowed optimization of experimental conditions resulting in significant reduction in the mass dependency of the MALDI response. The utility of this equimolar mixture as an internal standard for the examination of mass dependency of response in polydispersed synthetic materials, and as a calibration standard for MALDI and GPC, is demonstrated. This report is the first work detailing the application of synthetic discrete mass oligomers as a quantitative model for mass spectrometric analysis of macromolecular systems. EXPERIMENTAL SECTION Instrumentation. MALDI-TOF-MS spectra were acquired using a PE Biosystems Voyager System STR 4087 mass spectrometer (Perkin-Elmer Biosystems) equipped with a N2 laser emitting at 337 nm and a chevron microchannel plate detector (Carnegie Mellon University). All spectra were acquired in the positive ion, reflectron mode with an acceleration voltage of 25 000 V, and an extraction delay time of 225 ns. Spectra were acquired over the mass range of 500-7000 Da unless otherwise indicated. Typically, 200-256 laser shot spectra were summed to render a composite spectrum with 5-10 composite spectra averaged to provide integrated peak area data utilized for quantification. These data were baseline corrected and tabulated using Perspective Biosystems software and then processed using Microsoft Excel. Linear regression analysis was used to determine slopes and (59) Larsen, B. S.; Simonsick, W. J.; McEwen, C. N. J. Am. Soc. Mass Spectrom. 1996, 7, 287. (60) Sakurada, N.; Fukuo. T.; Arakawa, R.; Ute, K.; Hatada, K. Rapid Commun. Mass Spectrom. 1998, 12, 1895. (61) Shimada, K.; Lusenkova, M. A.; Sato, K.; Saito, T.; Matsuyama, S.; Nakahara, H.; Kinugasa, S. Rapid Commun. Mass Spectrom. 2001, 15, 277. (62) Reddy, S. S.; Dong, X.; Murgasova, R.; Gusev, A. I.; Hercules, D. M. Macromolecules 1999, 32, 1367. (63) Shan, L.; Murgasova, R.; Hercules, D. M.; Houalla, M. J. Mass Spectrom. 2001, 36, 140. (64) Murgasova, R.; Hercules, D. M. J. Mass Spectrom. 2001, 36, 1098.

Analytical Chemistry, Vol. 75, No. 13, July 1, 2003

3093

Figure 1. MALDI-TOF-MS spectrum of a polydispersed and predominately dihydroxyl-terminated PBA (Mn ) 2000) acquired using IAA as the matrix. The MALDI spectrum exhibits oligomers of the copolymer distribution having degrees of polymerization of 4-68. The higher intensity peaks, separated by the molecular weight of the adipic acid/butanediol copolymer repeat units (200 Da), represent (M + Na)+ of each oligomer. The inset presents the GPC chromatogram acquired from the same polymer.

intercepts from the experimental data. Hypothesis testing was utilized to statistically verify differences in the slopes of lines presented in Figures 4 and 8. GPC measurements were performed in THF at 35 °C with a flow rate of 0.30 mL/min maintained by a Waters 590 HPLC pump. The instrument was equipped with a Phenomenex Phenogel 5-µm, 500-Å column, a Waters 410 differential refractometer for detection, and a Waters 745 data module. Calibration was performed using poly(ethylene glycol) (PEG) standards (Polyscience) of molecular weights 960, 1500, 5000, and 9000 (1.0 mg/mL in HPLC grade THF). Sample Preparation. Synthesis of discrete molecular mass poly(butylene glutarate) (PBG) oligomers has been detailed previously.65 Oligomers of degree of polymerization of 8, 16, 32, and 64 were repurified by precipitation from CH2Cl2 with diethyl ether, then filtered, washed, and dried in a vacuum for 48 h. The 64-mer was recrystallized from hot ethanol, filtered, and washed before drying. Differential scanning calorimetry (DSC) pans were sonicated for 2 h in acetone, rinsed, and dried. This procedure was repeated using CH2Cl2, and the pans were dried in an oven for 24 h. An equimolar mixture of oligomers with degrees of polymerization of 8, 16, 32, and 64 was prepared by weighing a predetermined amount of each oligomer in individual preweighed DSC pans. The DSC pans containing the oligomers were placed in a vial and stirred in acetone for 4 h, then removed, and rinsed with acetone to wash any remaining material into the mixture vial. MALDI matrixes used were trans-3-indoleacrylic acid (IAA), 4-(2-hydroxyazophenyl) benzoic acid (HABA), 2,5-dihydroxyben(65) Williams, J. B.; Chapman, T. M.; Hercules, D. M. Macromolecules, in press.

3094 Analytical Chemistry, Vol. 75, No. 13, July 1, 2003

zoic acid (DHB), and 1,8,9-anthracenetriol (dithranol) (Aldrich Chemical Co.). Matrix solutions were prepared by mixing 40 mg of IAA in 1 mL of dry acetone or 20 mg of HABA, DHB, or dithranol in 1 mL of dry tetrahydrofuran (THF; Fisher Scientific). Typically, 2-5 µL of the PBG oligomer mixture (0.5-1.0 mg/ mL) in acetone or tetrahydrofuran was mixed with an equal volume of matrix solution. Since sodium cationized peaks appeared in the MALDI analysis of the experimental oligomers prepared using lithium, potassium, and cesium, sodium was chosen as the ionizing agent to simplify spectra. Sodium salt additions were made to all matrix/analyte preparations using 0.1-0.2 µL of a saturated NaCl solution in the matrix solvent. These resulting mixtures were sonicated before deposition (2-5 µL) onto a gold-plated sample plate. Sample depositions using IAA, HABA, DHB, and dithranol were prepared using the dried droplet method. Poly(butylene adipate) (PBA) (Mn ) 2000) was obtained from the Bayer Corp. and was prepared for MALDI analysis by mixing an equal volume of a 1 mg/mL PBA in acetone with an IAA matrix solution (acetone) followed by the addition of 0.2 µL of a saturated NaCl solution. GPC analysis was performed using 2 mg/mL PBA in THF. The experimental responses of PBA oligomers were determined by measurement of the peak areas of all individual ion species (oligomer isotopes and cationization modes) detected in MALDI spectra. The number-average (Mn) and weight-average (Mw) molecular weight and polydispersities (Mw/Mn) were then calculated using the above tabulated measurements. RESULTS AND DISCUSSION Polydispersed Materials. Figure 1 combines the MALDITOF mass spectrum and GPC chromatogram of the distribution

Figure 2. MALDI-TOF-MS spectrum of an equimolar mixture of discrete mass PBG oligomers having degrees of polymerization of 8, 16, 32, and 64 using IAA as the matrix. The most intense peaks represent (M + Na)+ of each discrete mass oligomer.

Table 1. Molecular Weight Averages and Polydispersity Indexes Determined by MALDI and GPC Analysis of the PBA Distribution Shown in Figure 1 technique

Mn

Mw

Mw/Mn

MALDI-TOF-MS GPC

2254 2830

3094 4810

1.37 1.70

of PBA (Mn ) 2000) and is presented to illustrate the difficulties in accurately determining molecular weight averages and oligomer distributions of more polydispersed materials. The MALDI spectrum of this polymer indicated an asymmetric distribution having an apex at 1300 Da and provided a series of sodium cationized peaks (repeat unit mass 200 Da) of diminishing intensity (molar concentration) extending to 7000 Da. The GPC trace of the same material indicated a more symmetrical distribution at higher masses (shorter retention times). The molecular weight averages, determined by MALDI and GPC for this material, along with the polydispersity index (Mw/Mn), are given in Table 1. It is seen that molecular weight values measured by MALDI are low in comparison to those calculated from GPC, which was calibrated using PEG standards. These results are consistent with reports27-46 that MALDI may systematically diminish the contributions (intensity) of components having higher degrees of polymerization, thus providing erroneously low molecular weight values for polymers having higher polydispersities (Mw/Mn > 1.2).27 The ability to evaluate the accuracy of the application of MALDI-TOF-MS for the determination of molecular weight averages of polydispersed materials, such as the PBA shown in Figure 1, has been impeded by a lack of suitable mass standards or models. Biological polymers, such as insulin and cytochrome c, have been well characterized but are structurally diverse and have chemical properties, such as solubility and functionality, that differ significantly from those of most synthetic counterparts. Thus, blends of biological polymers of known composition would not serve as adequate models to study the mass-dependent effects observed during the analysis of synthetic polymers. Discrete

oligomers of poly(methyl methacrylate)59,60 and polystyrene61 have been isolated using SFC but have not been extensively studied. Since the compositions of highly polydisperse materials are not accurately known, multicomponent blends of homopolymers having narrow polydispersities and differing molecular averages, such as commercially available polymer standards, have been used to mimic these polymers.36-40 However, components of these blends yield varying molecular weight averages, dependent on analysis technique, and thus, the composition of the blend cannot be accurately determined. In addition, blends of these materials produce multimodal distributions that are atypical of most polydispersed materials. Discrete Mass Oligomers. Figure 2 presents the MALDI spectrum of an equimolar mixture of PBG oligomers of precisely controlled degree of polymerization (8, 16, 32, 64), mass, and structure (hydroxylcarboxylic acid termination). The mixture served as model to study the extent of mass discrimination over the breadth of polymer distributions such as that of PBA shown in Figure 1. The PBG spectrum, acquired under optimized experimental conditions, is characterized by intense peaks representing sodium cationized oligomers (M + Na)+, lower intensity peaks from potassium cationized oligomers (M + K)+, and weaker peaks from sodium cationized oligomers of the sodium salt of the PBG acid end groups (M + 2Na - H)+ and was acquired over the mass range 500-7000 Da using IAA as the matrix. A list of ions that may be formed during the MALDI analysis of the PBG mixture is presented in Table 2. The 500-800-Da region of the spectrum exhibited low-intensity peaks resulting from the MALDI matrix (IAA). PBG ions resulting from fragmentation or matrix adduct attachment were not detected under normal experimental conditions. However, at higher laser intensities or higher analyte concentrations (>2 mg/mL), a distribution of peaks separated by the mass of the repeat unit, resulting from fragmentation, was observed. The matrix-to-analyte weight ratio was found to be optimal in the 40:1-80:1 range. Cyclic oligomers, corresponding to [M - H2O + Na]+, were not observed in the MALDI-TOF-MS, TOF-SIMS, and LSI-MS Analytical Chemistry, Vol. 75, No. 13, July 1, 2003

3095

Table 2. Ions That May Be Formed during MALDI-TOF-MS Analysis of PBG Mixtures oligomer

cation

13C

PBG 8-mer C36H58O17

Na+

0

785.4 32-mer (contd)

Na+ Na+

1 2

Na+ K+ K+ K+ K+ Na+ Na+ Na+

3 0 1 2 3 0 1 2

786.4 787.4 sodium salt 788.4 801.4 802.4 803.4 804.4 807.4 808.4 809.4 PBG 64-mer C288H450O129 810.4

sodium salt

PBG 16-mer C72H114O33

sodium salt

PBG 32-mer C144H226O65

m/z

oligomer

13C

m/z

K+

5

3038.4

K+ Na+

6 1

3039.4 3040.4

Na+ Na+ Na+ Na+ Na+

2 3 4 5 6

3041.4 3042.4 3043.4 3044.4 3045.4

Na+

0

5995.9

Na+ Na+ Na+

1 2 3

5996.9 5997.9 5998.9

4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9

5999.9 6000.9 6001.9 6002.9 6003.9 6004.9 6111.9 6112.9 6113.9 6114.9 6115.9 6116.9 6117.9 6118.9 6119.9 6120.9

0 1 2 3 4 5 6 7 8 9

6117.9 6118.9 6119.9 6120.9 6121.9 6122.9 6123.9 6124.9 6125.9 6126.9

cation

Na+

3

Na+

0

1529.7

Na+ Na+ Na+ Na+ K+ K+ K+ K+ K+ Na+ Na+ Na+ Na+ Na+

1 2 3 4 0 1 2 3 4 0 1 2 3 4

1530.7 1531.7 1532.7 1533.7 1545.7 1546.7 1547.7 1548.7 1549.7 1551.7 1552.7 1553.7 1554.7 1555.7

Na+

0

3018.4

Na+ Na+ Na+ Na+ Na+ Na+ K+ K+ K+ K+ K+ K+ K+ K+ K+ K+

Na+ Na+ Na+ Na+ Na+ Na+ K+ K+ K+ K+

1 2 3 4 5 6 1 1 3 4

3019.4 3020.4 sodium salt 3021.4 3022.4 3023.4 3024.4 3034.4 3035.4 3036.4 3037.4

Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+ Na+

spectra of individual or mixed PBG discrete mass oligomers. Cyclic oligomers have been detected in MALDI,20,21 SIMS,66 and fast atom bombardment67 analysis of aliphatic polyesters synthesized using polycondensation techniques. Since all couplings of discrete mass PBG oligomers were made under orthogonally protected conditions,65 i.e., one functional site per molecule, intramolecular couplings should not have occurred. This would indicate that the observation of cyclic oligomers in oligomers from spectra using the above techniques resulted from intramolecular terminal group couplings during synthesis, rather than sample preparation or gas-phase reactions under mass spectrometric analytical conditions. Isotopic Distribution and Resolution. It is seen in Figure 2 that the relative intensities (heights) of the peaks representing the highest isotopic abundance of each discrete mass oligomer decrease with increasing mass. However, accurate determination (66) Kim, Y. L.; Hercules, D. M. Macromolecules 1994, 27, 7855. (67) Ballistreri, A.; Garozzo, D.; Giuffrida, M.; Montaudo, G. Anal. Chem. 1987, 59, 2024.

3096 Analytical Chemistry, Vol. 75, No. 13, July 1, 2003

of the relative contribution of each of the discrete mass oligomers requires summation of the intensities of all isotopic peaks of all species (cation attachment or salt formation) detected for each oligomer (Table 2). The necessity for this approach is due to the profound change in the isotopic distribution of these structurally similar molecules over the experimental mass range (∼5250 Da). The theoretical isotopic distributions of the experimental PBG oligomers, cationized with Na+, are shown in Figure 3 at an instrumental resolution (M/∆M) ) 6000. The isotopic distribution of the 8-mer consists of four statistically significant isotopic compositions containing from zero to three 13C atoms per molecule (1.11% abundance) with small contributions from 2H (0.02%), 17O (0.04%), and 18O (0.20%). In contrast, the full isotopic composition of the PBG 64-mer is represented by a distribution of 10 statistically significant isotopic structures with the most probable structure having 3 13C atoms in the oligomer chain. Accurate determinations of relative intensities from MALDI spectra of equimolar mixtures of the model PBG were further complicated by instrument resolution (typically M/∆M ) 2000-3000 (fwhm)). An instrumental resolution of ∼5000 is necessary to adequately separate the isotopic envelope of the 64-mer. Although a resolution of M/∆M ) 5000-6000 was attainable using the PE Biosystems Voyager instrument, optimization of experimental parameters (resolution and peak shape) did not yield uniform results over the experimental mass range. More typically, optimization of all instrumental parameters allowed a resolution of M/∆M ) 20003000, which was inadequate for definitive identification of all isotopes expected for the 64-mer. Therefore, peak area was chosen for quantification of response to provide minimization of errors resulting from diminution of peak height due to peak broadening and the inability to separate the individual isotopic peaks at higher masses. Quantification of Mass Dependency of Response. Matrix Effects. Figure 4 presents the averaged mass-dependent response calculated from 10 MALDI spectra of the PBG equimolar mixture acquired using optimized experimental conditions with IAA as the matrix. The response values (Rn) are derived by summing of the peak areas of all representative ions (discussed above) detected for each oligomer having a discrete degree of polymerization and then normalizing to the summed intensity (R8-mer) of the all 8-mer ions. The results shown in Figure 4 indicate that, within experimental error, there is no observed mass dependency of the response during MALDI analysis of the PBG equimolar mixture (mass range 785-6050 Da). Mass dependency of response can result from many factors including sample preparation (matrix, matrix-to-analyte ratio, cationization salt, sample deposition),36,38-43 laser intensity,35 massdependent desorption/ionization,36 ion focusing and transmission,37 mass-dependent detector response,40 detector saturation,27,40 and signal-to-noise limitations.40 Thus, optimization of experimental parameters is crucial for the accurate determination of polymer distribution. Figure 5 provides the relative MALDI responses for the PBG 16-, 32-, and 64-mers provided by four matrixes previously employed for the analysis of aliphatic polyesters.25 The normalized intensities of the individual oligomers, calculated from spectra acquired using DHB, HABA, and dithranol as the matrixes, decrease with increasing oligomer mass, and the magnitude of

Figure 3. Theoretical isotopic distributions of the sodium cationized PBG oligomers shown in Figure 2 at resolution M/∆M ) 6000.

Figure 4. Mass dependency of response of an equimolar mixture of PBG discrete mass oligomers in the 500-7000-Da mass range using IAA as the matrix.

Figure 5. Matrix-dependent response of an equimolar mixture of PBG oligomers. The MALDI matrixes were IAA ([), DHB (9), dithranol (2), and HABA (×).

the change is observed to vary from matrix to matrix. The diminution of response in the mass range 780-1500 Da was significant and is seen to plateau at masses above 3000 Da. The least mass-dependent decrease of response was observed for IAA; thus, this matrix was utilized in subsequent experiments. The MALDI spectra acquired using dithranol and HABA exhibited more intense peaks, relative to IAA, from matrix cluster ions in the 500-850-Da mass region. Inclusion of matrix peaks

in the MALDI-TOF-MS experiment has been shown to result in detector saturation,37 which may cause diminution of peak intensities at higher masses. In the time-of-flight mass spectrometric experiment, lower mass ions are first to reach the detector and a finite time interval may be required for detector recovery. If the recovery is incomplete, the signal from ions having longer flight times will be diminished in intensity. This effect could be especially problematic in the analysis of polydispersed synthetic Analytical Chemistry, Vol. 75, No. 13, July 1, 2003

3097

Figure 6. Influence of laser intensity on mass-dependent response. The laser intensity was set at 1.36τ ([), 1.54τ (9), 1.72τ (2), and 1.90τ (×), and 2.08τ (O), where τ is the threshold intensity required for observation of all copolymer molecular ions (M + Na)+.

materials, where a large number of ions, separated by relatively small flight times, are continuously arriving at the detector. Laser Intensity. The contribution of higher mass components of synthetic polymers measured by MALDI analysis has been shown to both increase27 and decrease34,61 with increasing laser intensity. The relative dependence of response of the PBG oligomers with changes in laser intensity is shown in Figure 6. The experiment was performed using IAA as the matrix and a low-mass threshold of 700 Da, thus minimizing the influence of matrix ion detection. The unit of laser intensity τ is defined as the threshold intensity required for detection of ions from all oligomers in the equimolar mixture. It is seen that with the laser set at 1.36τ, there was essentially no discrimination over the experimental mass range (R64-mer/R8-mer ) 1.05 ( 0.06). Use of higher laser intensities progressively decreased the intensities of all higher mass oligomers relative to the 8-mer. Mass spectra

taken at higher laser intensities (>1.36τ) exhibited higher total ion currents (TIC) and increases in matrix ion background in the lower mass region of the recorded spectra. Fragmentation of PBG oligomers was not observed over the range of laser intensities presented in this report. However, when laser intensities above 2.08τ were applied for analysis, fragments of PBG oligomers separated by the mass of the repeat unit (186 Da) were detected. The results shown in Figure 6 also indicated that mass discrimination increased with increasing TIC, further supporting the belief that detector saturation is a significant factor affecting of mass dependency of response of oligomers in the MALDI-TOF-MS analysis of polydispersed materials. Standard Applications. Effects of Polymer Distributions. To further examine mass discrimination in polydispersed systems, a quasi-equimolar mixture of polydispersed PBA 2000 (Figure 1) and the PBG equimolar mixture (Figure 2) was analyzed over the mass range 500-7000 using IAA as the matrix (Figure 7). The mixture of the two copolymers is deemed quasi-equimolar due to the analytical difficulties encountered in determination of the accurate molecular weight of polydispersed PBA. The PBA polyester and PBG model copolymers are similar in molecular structure, differing by only one methylene group in the repeat unit and the predominately di-1,4-butanediol termination of PBA. The spectrum of the PBA/PBG mixture exhibits peaks from both components, and the distributions of each moiety are similar to those seen in Figures 1 and 2, respectively. The calculation of molecular weight averages from the PBA oligomers shown in Figure 7 yielded Mn ) 2231 and Mw ) 2869 (Mw/Mn ) 1.29), values lower than those obtained from the unmixed PBA analysis (Figure 1 and Table 1). Comparison of the PBA distributions in Figures 1 and 7 reveals some diminution of response from PBA oligomers in the mass range 5500-7000 Da. This loss of signal would lower molecular weight averages and may have resulted either from detector saturation caused by the presence of higher intensity signals from the PBG 32 and 64-mer or from small

Figure 7. MALDI-TOF-MS spectrum of a quasi-equimolar mixture of polydispersed PBA (Figure 1) and equimolar PBG oligomers (Figure 2) using IAA as the matrix. The PBG (M + Na)+ ions are identified. 3098 Analytical Chemistry, Vol. 75, No. 13, July 1, 2003

Figure 8. Relative responses of equimolar PBG oligomers in the PBA/PBG mixture analysis shown in Figure 7.

changes in the matrix-to-analyte ratio of the PBA/PBG mixture relative to that utilized for the PBA analysis (Figure 1). Quantification of the PBG distribution shown in Figure 7 yielded a plot of normalized response (Rn) versus mass (Figure 8) that indicated some diminution in the response of higher mass components relative to that derived from the PBG mixture spectra (Figure 4) taken under similar experimental conditions. The decrease in PBG response with increasing mass (R64mer/R8mer ) 0.95 ( 0.07) is most likely due to detector saturation resulting from the high number of PBA and PBG ions striking the detector at shorter flight times. However, the inclusion of the equimolar PBG mixture allows the determination that the asymmetric distribution of PBA oligomers is largely the result of polycondensate synthesis. It is seen that the PBA molecular weight values calculated from the MALDI spectra shown in Figures 1 and 7 are significantly lower than the values obtained from GPC (Table 1). To investigate reasons for this discrepancy, GPC chromatograms were obtained for unmixed PBG oligomers and PEG calibration standards. The individual chromatograms for the PBG oligomers are shown in Figure 9. The chromatographs are characterized by symmetrical peaks with retention times increasing with decreasing degree of polymerization. Figure 10 presents the plot of log MW versus retention time of the PBG oligomers (9) and PEG standards (b). The molecular weights provided for the PBG oligomers represent the apex peak of the isotopic distribution while PEG molecular weights are those quoted by the manufacturer. Clearly the two plots (calibration curves) differ in both slope and position on the retention time axis indicating that, at molecular weights below 5000, PBG oligomers of a given mass elute at shorter retention times than PEG distributions of comparable masses. A possible explanation is that this discrepancy is due to differences in hydrodynamic volume at given masses of PBG and PEG chains in the lower mass range. The relationship between hydrodynamic volume and mass is given by the equation Vh ) [η]M/K, where [η] is the intrinsic viscosity, M is the molecular weight, and K is a proportionality constant.68 Differences in hydrodynamic volume of polymer chains at equal mass arise from variations in the affinities of the solvent (THF) for structurally dissimilar polymers and result from variations in parameters [η], in K, or in both. (68) Yau, W.; Kirkland, J. J.; Bly, D. D. Modern Size Exclusion Chromatography; John Wiley and Sons: New York, 1979.

Figure 9. GPC chromatograms of individual discrete mass PBG oligomers.

Application of the PEG calibration curve to determine PBG molecular weights from retention times, as shown for the MW ) 762, 1508, and 2996 oligomers in Figure 10, produced erroneously high molecular weight value (Mw ) 1100, 2200, and 3600) for the 8-, 16-, and 32-mers, respectively. The PBA molecular weight averages presented earlier were based on calibration using the same PEG standards. Overestimation of the molecular weights of components in 700-5000 region of the distribution by GPC could result in erroneously high GPC molecular weight averages for the PBA polymer. Application as Calibration Standards. Discrete mass PBG oligomers and mixtures, having precisely defined structures and known molecular weights, may serve as mass calibration standards for mass spectrometry and other macromolecular techniques such as GPC and viscometry. As seen in Figures 2 and 3 and Table 2, the MALDI response of the experimental PBG mixture can provide ∼40 peaks of accurately known mass over a mass range spanning 5200 Da for calibration. GPC calibration is usually performed using polymer molecular weight standards having narrow, but finite, distributions having unknown compositions. The application of monodispersed materials of known structure provides the ability to accurately calibrate GPC. The GPC traces shown in Figure 9 represent standards having accurately defined isotopic distributions spanning only a 4-10-Da mass range, as opposed to commercial polymer mass standards, which have distributions spanning several repeat units. Thus, these materials are applicable as standards for GPC analysis of materials having similar solvent affinities in the 700-6000-Da mass range. To illustrate this point, the polydispersed PBA 2000 (Figure 1) was analyzed using GPC calibrated with the PBG discrete mass standards (Figure 9). The molecular weight averages were found to be Mn ) 2290 and Mw ) 3624 (Mw/Mn ) 1.58), Analytical Chemistry, Vol. 75, No. 13, July 1, 2003

3099

Figure 10. Comparison of GPC calibration curves based on PEG standards (b) and PBG discrete mass oligomers (9).

values that are in more reasonable agreement with the MALDI values presented in Table 1. CONCLUSIONS An equimolar mixture of discrete mass PBG oligomers of precisely defined structure, degree of polymerization, and molecular weight was studied using MALDI-TOF-MS and GPC. Mass spectra of PBG oligomers exhibited peaks from cationized intact copolymer oligomers of the targeted degree of polymerization. The absence of fragmentation, cyclization, or heterotermination in the mass spectra of individual or mixtures of PBG oligomers indicated that the presence of these ions in the mass spectra of polyester distributions was the result of polycondensate synthesis. The extent and possible causes of mass-dependent response observed in the MALDI analysis of polydispersed materials were studied. The results demonstrated that mass dependency of response can be reduced by choice of matrix, laser power, and experimental parameters, such as threshold mass, that minimize detector saturation. Under optimized experimental conditions, the responses of components of the equimolar PBG oligomer mixture were found to be uniform over the mass range 780-6000 Da. Application of discrete mass PBG oligomers to study the saturation effects resulting from detection of high numbers of ions in polydispersed materials and as mass calibration standards for

3100

Analytical Chemistry, Vol. 75, No. 13, July 1, 2003

polymer analysis was demonstrated. Analysis of a quasi-equimolar mixture of PBA and PBG oligomers indicated that the asymmetrical distribution of PBA 2000 is the result of synthesis. These molecular averages calculated from MALDI PBA spectra showed more reasonable agreement with results from GPC analysis of PBA using discrete mass PBG oligomers as the calibration standards. The high mass accuracy and monodispersity of these materials should allow more accurate calibration of GPC and soft ionization mass spectrometric experiments. ACKNOWLEDGMENT This research was supported by the Department of Chemistry, University of Pittsburgh and by NSF Grant CHE-9985864. The authors thank Dr. Kasi Somayajula of the University of Pittsburgh Mass Spectrometry Laboratory for his technical assistance and access to his facilities. The authors additionally thank Dr. Isabella Verdinelli (Statistics Department) and Dr. Mark Bier (Center for Molecular Analysis (NSF Grant CHE-9808188)) at Carnegie Mellon University for their assistance.

Received for review February 11, 2003. Accepted March 18, 2003. AC030061Q