Importance of Solubility in the Sample Preparation of Poly (ethylene

Anal. Chem. 2005, 77, 750-756

Importance of Solubility in the Sample Preparation of Poly(ethylene terephthalate) for MALDI TOFMS Andrew J. Hoteling,†,‡ Thomas H. Mourey,† and Kevin G. Owens*,‡

Research & Development Laboratories, Eastman Kodak Company, Rochester, New York 14650-2132, and Department of Chemistry, Drexel University, Philadelphia, Pennsylvania 19104-2875

The role of solubility in the sample preparation process for matrix-assisted laser desorption/ionization (MALDI) mass spectrometry is demonstrated for oligomeric and medium molar mass poly(ethylene terephthalate) (PET). For low molar mass oligomers (PET-1), minor discrimination effects were observed when the sample was not completely in solution. MALDI spectra of medium molar mass PET, representative of the entire molar mass distribution, were obtained only when a good solvent for PET was used, such as 1,1,1,3,3,3-hexafluoro-2-propanol (commonly referred to as HFIP), as the sample preparation solvent and dithranol as the matrix. The azeotropic composition of 70:30 CH2Cl2/HFIP better solubilizes the more nonpolar matrixes, which enables more latitude in selecting sample preparation conditions than pure HFIP. Segregation effects were observed when the azeotrope mixture was diluted with tetrahydrofuran, resulting in large molar mass distribution discrimination effects in the MALDI spectra. Dilution with CH2Cl2 resulted in a significant decrease in the overall signal intensity for the entire polymer distribution. With each attempt to dilute the azeotrope, the sample after solvent evaporation was visibly heterogeneous, which resulted in shot-to-shot variability. Both examples demonstrate the importance of constant solvent composition during solvent evaporation. The compatibility of matrix and polymer was explored using relative HPLC retention times. Consistent with previous work in our laboratories, it was found that the matrix/polymer combination that has the closest match of retention time resulted in the best MALDI signal intensity. Over the past decade, matrix-assisted laser desorption/ionization (MALDI) TOFMS has evolved as a useful tool for synthetic polymer characterization.1-3 However, there are few publications that report on the use of MALDI with polyesters4-8 in general or with poly(ethylene terephthalate) (PET)9-13 specifically. The existing publications have focused primarily on the very low molecular mass oligomers of PET, where successful MALDI analyses were accomplished with the matrixes 2,5-dihydroxybenzoic acid (DHB), * To whom correspondence should be addressed. Tel: 215-895-2621. Fax: 215-895-1265. E-mail: [email protected]. † Eastman Kodak Co. ‡ Drexel University. (1) Rader, H. J.; Schrepp, W. Acta Polym. 1998, 49, 272-293. (2) Nielen, M. W. F. Mass Spectrom. Rev. 1999, 18, 309-344. (3) Hanton, S. D. Chem. Rev. 2001, 101, 527-569.

750 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

trans-indoleacrylic acid (IAA), 2-(4-hydroxyphenylazo)benzoic acid (HABA), 2′,4′,6′-trihydroxyacetophenone (THAP), and 1,8,9-anthracentetriol (dithranol). The difficulty in analyzing higher molar mass distributions of PET has been described. Solubility has been demonstrated to be important in MALDI sample preparation. The solvent used in sample preparation has been identified as a potential cause of molar mass discrimination effects. It is generally thought best to use the same solvent to dissolve the matrix, the analyte (polymer), and the cationization agent (if needed)14 and that sample segregation can occur when mixed solvents are used.15 The effects of using mixed solvents in the MALDI analysis of PS and PMMA polymers was investigated by Li and co-workers.16 Cyclic and linear PET oligomers with molar masses of less than ∼1000 u have limited solubility in common organic solvents such as THF and dichloromethane; however, higher molar mass PET polymers are insoluble in these solvents. Solvents that are useful for higher molar mass polymers include strong hydrogenbonding solvents, such as m-cresol and o-chlorophenol; mixtures of halogenated hydrocarbons with phenol or nitrobenzene; and strong organic acids, such as dichloroacetic and triflouroacetic acids.17 Many of the phenolic solvents are useful only at elevated temperatures, and the high boiling points or corrosivity of these solvents limits their use for MALDI sample preparation. Of the (4) Blais, J. C.; Tessier, M.; Bolbach, G.; Remaud, B.; Rozes, L.; Guittard, J.; Brunot, A.; Marechal, E.; Tabet, J. C. Int. J. Mass Spectrom. Ion Processes 1995, 144, 131-138. (5) Wachsen, O.; Reichert, K. H.; Kruger, R. P.; Much, H.; Schultz, G. Polym. Degrad. Stab. 1997, 55, 255-231. (6) Williams, J. B.; Gusev, A. L.; Hercules, D. M. Macromolecules 1997, 30, 3781-3787. (7) Laine, O.; Vainiotalo, P.; Osterholm, H.; Jarvinen, H. Eur. J. Mass Spectrom. 2001, 7, 15-23. (8) Murgasova, R.; Brantley, E.; Hercules, D. M.; Nefzger, H. Macromolecules 2002, 35, 8338-8345. (9) Weidner, S.; Kuehn, G.; Friedrich, J.; Schroeder, H. Rapid Commun. Mass Spectrom. 1996, 10, 40-46. (10) Jackson, A. T.; Yates, H. T.; MacDonald, W. A.; Scrivens, J. H.; Critchley, G.; Brown, J.; Deery, M. J.; Jennings, K. R.; Brookes, C. J. Am. Soc. Mass Spectrom. 1997, 8, 132-139. (11) Weidner, S.; Kuehn, G.; Werthmann, B.; Schroeder, H.; Just, U.; Borowski, R.; Decker, R.; Schwarz, B.; Schmuecking, I.; Seifert, I. J. Polym. Sci., Part A: Polym. Chem. 1997, 35, 2183-2192. (12) Weidner, S.; Kuhn, G.; Decker, R.; Roessner, D.; Friedrich, J. J. Polym. Sci., Part A: Polym. Chem. 1998, 36, 1639-1648. (13) Gidden, J.; Wyttenbach, T.; Batka, J. J.; Weis, P.; Jackson, A. T.; Scrivens, J. H.; Bowers, M. T. J. Am. Chem. Soc. 1999, 121, 1421-1422. (14) Danis, P. O.; Karr, D. E. Org. Mass Spectrom. 1993, 28, 923-925. (15) Chen, H.; Guo, B. Anal. Chem. 1997, 69, 4399-4404. (16) Yalcin, T.; Dai, Y.; Li, L. J. Am. Soc. Mass Spectrom. 1998, 9, 1303-1310. (17) Brandrup, J.; Immergut, E. H.; Grulke, E. A.; Abe, A.; Bloch, D. R. Polymer Handbook, 4th ed.; John Wiley and Sons: New York, 2003. 10.1021/ac048525n CCC: $30.25

© 2005 American Chemical Society Published on Web 01/05/2005

available room-temperature solvents, fluorinated alcohols such as 2,2,2-triflouroethanol and 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) are popular choices for chromatography,18 and the latter is most widely usedsdespite being expensive and a severe eye irritant. HFIP is easily evaporated (bp ) 58 °C) and can be diluted with dichloromethane to form a minimum boiling azeotrope (bp ) 36 °C) at a ratio of 70:30 (v/v) dichloromethane/HFIP, which is a better solvent for PET at room temperature than HFIP.19,20 The azeotropic composition is particularly attractive as a solvent for MALDI because it maintains 70:30 (v/v) composition during evaporation, which would minimize fractional precipitation of components. In addition to solvent-related issues, the compatibility of analyte (polymer) with matrix is an important consideration. The proper choice of matrix is necessary in order for the analyte to be homogeneously incorporated within the matrix after solvent evaporation. We refer to this as a “solid solution”, where the matrix is analogous to a solvent and the analyte molecules are “solvated” by the matrix molecules in the solid state. Hanton and Owens observed, using matrix-enhanced secondary ion mass spectrometry as a probe,21,22 that the solid solution is best achieved by matching the relative polarity of the matrix and analyte. Recent work within our group expanded on the idea of matching matrix and analyte polarity using reversed-phase gradient HPLC retention times as a guide.23 Solid sample preparation has also been reported as an alternative approach for analytes that are difficult to dissolve.24,25 In this paper, the importance of considering both solution- and solid-phase solubility in MALDI sample preparation is demonstrated in the development of conditions suitable for the analysis of PET samples of low and medium molecular mass. Because of the large difference in sample preparation protocols between the conventional “solution-phase” and solid sample preparation methodologies, no solid sample preparation results will be presented here. EXPERIMENTAL SECTION Materials. Samples of PET were synthesized at Eastman Kodak Co. Sample PET-1 is a low molar mass sample (Mn ) 410, Mw ) 627 measured by size-exclusion chromatography, SEC), and the sample PET-2 is a medium molar mass sample (Mn ) 6280, Mw ) 17500, measured by SEC). Matrix materials dithranol, IAA, and DHB and the cationization reagent, sodium trifluoroacetate (NaTFA), were purchased from Aldrich Chemical Co. (Milwaukee, WI). Prior to use, the large excess of sodium, potassium, or both was removed from IAA and DHB by recrystallization. All other materials were used without further purification. HFIP, purchased (18) Mourey, T. H.; Bryan, T. G. J. Chromatogr., A 2002, 694, 169-178. (19) Overton, J. R.; Browning, H. L. Org. Coat. Appl. Polym. Sci. Processes 1983, 48. (20) Overton, J. R.; Browning, H. L. In Size Exclusion Chromatography; Provder, T., Ed.; ACS Symposium Series 245; American Chemical Society: Washington, DC, 1984; Vol. 245, p 219. (21) Hanton, S. D.; Owens, K. G. Proc. 46th ASMS Conf. Mass Spectrom. Allied Topics, 1998; p 1185. (22) Hanton, S. D.; Cornelio Clark, P. A.; Owens, K. G. J. Am. Soc. Mass Spectrom. 1999, 10, 104-111. (23) Hoteling, A. J.; Erb, W. J.; Tyson, R. J.; Owens, K. G. Anal. Chem. 2004, 77, 5157-5164. (24) Trimpin, S.; Rouhanipour, A.; Az, R.; Rader, H. J.; Mullen, K. Rapid Commun. Mass Spectrom. 2001, 15, 1364-1373. (25) Skelton, R.; Dubois, F.; Zenobi, R. Anal. Chem. 2000, 72, 1707-1710.

from Lancaster, Inc. (Pelham, NH), was purified by fractional distillation over 3A molecular sieves. Sample Preparation. PET-1: Depending on the solvent conditions under study, the reagents were either prepared in THF (unstabilized) or a mixture of CH2Cl2/HFIP (70:30 v/v). The polymer was prepared at a concentration of 0.1 mg/mL, the matrix solution (dithranol) was prepared at a concentration of 5 mg/ mL, and the cationization reagent, NaTFA, was prepared at 0.1 mg/mL. The samples were prepared by mixing the polymer solution with dithranol solution and the NaTFA solution at a volume ratio of 1:1:1. A volume of 0.5 µL of the mixture was deposited onto a sample plate and allowed to air-dry. For the experiments using THF as the solvent, the polymer was either heated below the boiling point of THF for 30 min (solution) or dispersed in THF at room temperature. PET-2: Depending on the solvent conditions being studied, the reagents were prepared in THF (unstabilized), HFIP, or the mixture of CH2Cl2/HFIP (70:30, v/v). The polymer was prepared at a concentration of 1 mg/mL, the matrix solution (dithranol) was prepared at a concentration of 5 mg/mL, and the cationization reagent, NaTFA, was prepared at 0.1 mg/mL. The samples were prepared by mixing the polymer solution with dithranol solution and NaTFA solution at a volume ratio of 1:1:1. A 0.5-µL aliquot of the mixture was deposited onto a sample plate and allowed to air-dry. MALDI Time-of-Flight (TOF) Instrumentation. MALDITOFMS experiments were carried out using a TofSpec2E laser TOF mass spectrometer (Micromass, Inc.), equipped with dual microchannel plate detectors for linear and reflectron modes and a nitrogen laser (337 nm). Positive-ion mode was used for all analyses, with an accelerating voltage of 25 kV for linear mode and 20 kV for reflectron mode. Spectra were acquired using delayed extraction mode with a 500-ns delay time. The delayed extraction pulse voltage was optimized for resolution based on the mass range of the individual polymer distributions. HPLC Conditions. Reversed-phase gradient HPLC analyses were performed using an HP1090 liquid chromatograph (Agilent Technologies, Palo Alto, CA), equipped with a diode array detector (DAD) and a Sedex model 55 (Sedere, Alfortville Cedex) evaporative light-scattering detector (ELSD). A YMC ODS-AQ (3.0 × 100 mm; S-5 120-Å particles) column was used for all analyses. A binary gradient was used, with the A phase being 0.1 M ammonium acetate buffer (pH 4.65) and the B phase being THF. The gradient conditions started at 10% B, were ramped to 100% B in 10 min, and were held at 100% B for 1 min. For the DAD detector, the data were collected at 254 nm (full diode array data collected) with a bandwidth of 4 nm. The ELSD was run at a temperature of 42 °C, with a nitrogen purge pressure of 2.3 bar and a gain (PMT) of 12. The injection volume was 2 µL. RESULTS AND DISCUSSION Low Molar Mass PET (PET-1). PET oligomers formed from the ester exchange reaction of dimethyl terephthalate with excess ethylene glycol have the expected structure shown below and a mass-average molecular mass of 627.

Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

751

Figure 1. MALDI TOF mass spectra of PET-1 using dithranol as the matrix and NaTFA as the cationization reagent. The spectra demonstrate the effects of solubility with THF as the solvent, where (a) the sample was heated for 30 min below the boiling point of THF with stirring (sample in solution) and (b) the sample was suspended at room temperature (sample not completely in solution).

This sample was dissolved in THF with heating (below the boiling point of THF). The reflectron-mode positive-ion MALDI TOF mass spectrum of this sample solution, using dithranol as the matrix and NaTFA as the cationization reagent (both dissolved in THF), is displayed in Figure 1a. The spectrum contains one series of equispaced peaks with the correct mass difference of the repeat unit of the PET polymer (i.e., 192 u). After subtraction of the mass of sodium and the mass of n repeat units (n ) degree of polymerization), the residual mass (62 u) is consistent with the expected end groups depicted in the structure above. PET-1 has limited solubility in THF, and it required heating to completely dissolve the sample. The sample does not dissolve in THF completely at room temperature and, at best, forms a suspension. To investigate the effect of analyte solution-phase solubility, the room-temperature suspension of the oligomeric sample was mixed with dithranol and NaTFA and was analyzed by MALDI (Figure 1b). The high molar mass region of each spectrum in Figure 1 is expanded to show the detail of the signalto-noise ratio (S/N). Note that the S/N of the higher molar mass peaks (e.g., m/z 2006, 2198, 2390, 2582, 2774) in the spectrum of the sample prepared from the THF suspension is significantly lower than in the spectrum of the sample completely solubilized by heating. This sample was also prepared using the azeotropic mixture of CH2Cl2/HFIP (70:30, v/v), which rapidly and completely dissolves PET at room temperature.18 In Figure 2, the MALDI spectrum obtained from the sample preparation using the azeotrope is compared to the spectrum obtained from heated THF. Note that the range of the distribution in each spectrum is comparable, which suggests that this sample was in solution in the heated THF. The matrix cluster peaks in the ranges of m/z 450 and 675 are less abundant in Figure 2b, making it easier to analyze the low molar mass oligomers. Medium Molar Mass PET (PET-2). Condensation polymerization of the ester exchange sample (PET-1) results in PET2, with a mass-average molecular mass of 17 500. This sample is largely insoluble in THF and other common organic solvents. 752 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

Figure 2. MALDI TOF mass spectra of PET-1 comparing two different solvent conditions used to prepare the sample: (a) THF heated to below the boiling point and (b) the azeotrope mixture of CH2Cl2/HFIP (70:30).

Attempts to prepare a MALDI sample using an undissolved suspension of sample in THF, mixed with matrix and cationization reagent dissolved in THF, resulted in no polymer-related signals in the MALDI spectrum (vida infra). This emphasizes the importance of analyte solubility on a molecular level in the sample solvent. Sample PET-2 is, however, completely soluble in HFIP. Figure 3a shows the MALDI spectrum obtained for PET-2 using HFIP as the solvent for the sample, matrix, and cationization reagent. For higher resolution to assist interpretation, a reflectron MALDI TOF spectrum was acquired using the same conditions (Figure 3b). The spectrum contains one major series of equispaced peaks and a few minor series. The adjacent peaks in each series have the expected repeat unit difference of 192 u. Structures that are consistent with the series present in the spectra are presented in Figure 4. The masses of the oligomers present in the major series (labeled series a) are consistent with the expected end groups and cationized with Na+. The minor series labeled b in the spectrum is 44 u higher in mass than the a series, which is consistent with the presence of a diethylene glycol unit (a minor contaminant in ethylene glycol). The diethylene glycol unit could be incorporated either within the sequence of the polymer or as an end group. The series labeled c in the spectrum is consistent with the formation of a cyclic species. These structures are consistent with those reported previously by Weidner and coworkers.9 When compared with the overall distribution obtained from SEC, the distribution observed in the MALDI spectrum (Figure 3) shows some discrimination of the high molar mass end of the distribution, as expected for samples with broad polydispersity.26 The SEC chromatogram of PET-2 showed a distribution that extended beyond 20 000. Mixed Solvent Effects. Many matrix materials have limited solubility in HFIP. A potential solution to this problem is to dissolve the PET in HFIP and the matrix and cationization reagent (26) Mourey, T. H.; Hoteling, A. J.; Owens, K. G.; Balke, S. T. J. Appl. Polym. Sci. Accepted.

Figure 3. MALDI TOF mass spectrum of PET-2 using HFIP as the solvent, dithranol as the matrix, and NaTFA as the cationization reagent: (a) linear mode spectrum and (b) expanded region of the reflectron mode spectrum.

in a different solvent (e.g., THF or CH2Cl2). It is generally believed, however, that using mixed solvents in MALDI sample preparation (especially for polymers) should be avoided. Chin and Guo15 and Li and co-workers16 demonstrated that segregation can occur as a consequence of the change in the solvent composition during the evaporation process, resulting in discrimination effects. The medium molar mass PET (PET-2 sample), as well as other limited solubility polymers, could suffer from the same effects when different solvents are used for analyte and matrix (and cationization reagent), resulting in a mixed solvent composition. The azeotrope mixture of CH2Cl2/HFIP (70:30) has the potential to address this solubility problem because it is a better solvent for some matrixes, it is an improved room-temperature solvent for PET, and it maintains a constant solvent composition during

evaporation, which should serve to minimize segregation effects. A qualitative comparison of the solubility of a range of matrixes in HFIP and in the azeotrope is presented in Table 1. The order of the matrixes is arranged according to their relative polarity (RA being the most nonpolar), based on previous work in our laboratory.23 Note that the more nonpolar matrixes are more soluble in the azeotrope than in pure HFIP; however, the more polar matrixes seem to be less soluble in the azeotrope. Figure 5a displays the MALDI TOF mass spectrum of PET-2 obtained using dithranol and NaTFA. All three components were dissolved in the azeotrope mixture prior to deposition. The distribution of peaks, as well as the series present, closely matches those seen using pure HFIP as the solvent (see Figure 3). To evaluate the effects of deviating from the azeotrope composition, Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

753

Figure 4. Structures consistent with the different series present in the MALDI TOF mass spectra of PET-2 shown in Figure 3. Table 1. Comparison of the Solubility of Different Matrixes in Two Different Solvent Systems Tested at 20 mg/mL matrix

HFIP

RA dithranol THAP HABA IAA CHCA DHB

very slightly soluble soluble very slightly soluble nearly soluble very slightly soluble slightly soluble slightly soluble

70:30 (v/v) CH2Cl2/HFIP soluble soluble very slightly soluble soluble very slightly soluble very slightly soluble very slightly soluble

various amounts of THF were introduced into the sample preparation. The spectrum displayed in Figure 5b was obtained from the preparation where PET-2, matrix, and NaTFA were dissolved in the azeotrope, and THF was added for a final ratio of 12:1 (azeotrope/THF). The spectrum in Figure 5c was obtained from the preparation where the analyte and cationization reagent were dissolved in the azeotrope, the matrix (dithranol) was dissolved in THF, and the components were mixed 1:1:1, resulting in a final ratio of 2:1 (azeotrope/THF). Figure 5 demonstrates that the addition of a polymer nonsolvent (THF) to the sample preparation has a dramatic affect on the range of the molar mass distribution represented in the MALDI spectrum. The molar mass discrimination in Figure 5a is representative of the known issues with samples with broad polydispersity, as mentioned above, while segregation effects from the sample preparation conditions additionally influence the discrimination in Figure 5b and c. Here, segregation appears to be molecular mass dependent, with the high molar mass oligomers precipitating before the matrix crystallizes, while the low molar mass oligomers remain in solution longer, allowing them to become incorporated within the matrix. The PET polymer has better solubility in the azeotrope than THF, and the matrix has better solubility in THF than the azeotrope. Note that the boiling point of the azeotrope is 36 °C (compared 754 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

Figure 5. Series of MALDI TOFMS spectra of PET-2, obtained by varying the amount of THF added to the azeotrope (CH2Cl2/HFIP). Dithranol is used as the matrix and NaTFA as the cationization reagent: (a) CH2Cl2/HFIP (70:30) (azeotrope), (b) (CH2Cl2/HFIP 70: 30)/THF (12:1 volume ratio), and (c) (CH2Cl2/HFIP 70:30)/THF (2:1 volume ratio).

to 66 °C for THF). If we assume that the addition of THF to the azeotrope does not form a new azeotrope (i.e., has a different boiling point), it would be expected that, during the solvent evaporation step of the sample preparation, the azeotrope would evaporate before the THF. If this were the case, it would be expected that, as the azeotrope evaporates, the polymer would precipitate out of solution while the matrix would stay in solution longer in the remaining THF-rich solvent. The appearance of the sample spots after solvent evaporation supports the idea that segregation is occurring. To demonstrate this segregation, optical microscopy images of the dried sample spots from two sets of sample preparation conditions are presented in Figure 6. The conditions used for the image in Figure 6a were

Figure 6. Optical microscopy images of dry-drop MALDI sample surfaces of PET-2 using dithranol as the matrix and NaTFA as the cationization reagent. Demonstration of the effect of choice of solvent: (a) azeotrope mixture of CH2Cl2/HFIP (70:30) and (b) (CH2Cl2/HFIP 70:30)/THF (12:1 volume ratio).

Figure 7. MALDI TOFMS spectra of PET-2 using dithranol as the matrix and NaTFA as the cationization reagent. The spectra were obtained using the solvent compositions (CH2Cl2/HFIP 70:30)/CH2Cl2 (2:1 volume ratio), which results in a final ratio of 80:20 CH2Cl2/ HFIP: (a) spectra taken from the center region of the sample and (b) spectra taken from the edge region.

the same conditions used to produce the spectrum displayed in Figure 5a, and likewise, Figure 6b corresponds with Figure 5b. Note that, with the azeotrope, the sample surface is uniform, and with the addition of THF, there is obvious visible sample segregation occurring. As additional evidence of sample segregation, MALDI analysis of the sample preparation depicted in Figure 6b suffers from greater shot-to-shot variability. The data in Figures 5 and 6 described above demonstrate a differential segregation affect on the molar mass distribution. A different effect is observed when the azeotrope mixture was varied, by changing the amount of CH2Cl2 (i.e., the mix ratio), as shown in Figure 7. With the diluted sample preparation, the sample spot after solvent evaporation appeared as a ring of large crystals with smaller crystals in the center. The MALDI spectrum displayed in Figure 7a was obtained from the small crystals in the center of the spot, and the spectrum in Figure 7b was obtained from the larger edge crystals. Note that the distributions represented in the spectra are very similar to that from the azeotrope (see Figure 5a). This indicates that the entire polymer distribution is segregating from the matrix. Further dilution of the HFIP results in no measurable MALDI signal for the polymer. The difference in the

S/N shown in Figure 7 demonstrates the shot-to-shot variability that occurs with this type of heterogeneous sample. Matrix/Analyte: Solid Solution Effects. Previous work in our laboratory used reversed-phase gradient HPLC to measure the relative polarities of a range of MALDI matrixes.23 With regard to signal intensity (as measured by S/N), it was demonstrated that a close match of matrix and analyte HPLC retention times resulted in the best MALDI spectra. It was proposed that the HPLC retention times for analytes (e.g., polymers) of interest could be used to help guide in matrix selection. Here, the same HPLC method was used to measure the sample PET-1, and the retention times obtained were used to compare with a range of matrixes that were attempted with this polymer system. The HPLC chromatogram of PET-1 is displayed in Figure 8a. Note that the higher molecular mass oligomers are likely underrepresented in the chromatogram, which is due to limited solubility in THF; these would elute at longer retention time if soluble in the eluent. For comparison, the HPLC chromatograms of dithranol, IAA, and DHB are displayed in Figures 8b-d, respectively. The retention time of dithranol is the closest match with PET-1 and would also be similarly expected to be a good match with PET-2. To investigate the effects of matching the polarity of the matrix and polymer using the HPLC retention times as a guide, MALDI spectra were acquired for PET-2 using dithranol, IAA, and DHB as matrixes. NaTFA was used as the cationization reagent, and the azeotrope mixture of CH2Cl2/HFIP (70:30) was used as the solvent in all analyses. The MALDI spectra obtained using dithranol and IAA as the matrix are compared in Figure 9. The third matrix, DHB, had limited solubility in both HFIP and the azeotropic mixture of CH2Cl2/HFIP (70:30), and attempts to obtain MALDI spectra failed to produce any polymer signals. The sample prepared using IAA as the matrix showed noticeable segregation upon solvent evaporation, and spectra of PET could only be obtained from the edge crystals. Also, there were large matrixNa clusters that dominated the low-mass region of the spectrum (labeled with * in Figure 9b). The spectrum obtained using dithranol as the matrix displayed significantly higher intensity polymer signals, with no interfering matrix-Na cluster peaks. As mentioned above, the sample prepared using dithranol as the matrix formed a fairly homogeneous sample spot with small matrix crystals. Note that this matrix/polymer combination had the Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

755

Figure 9. MALDI TOF MS spectra of PET-2 using CH2Cl2/HFIP (70:30) as the solvent and NaTFA as the cationization reagent, comparing the matrixes (a) dithranol and (b) IAA. Peaks labeled with * are matrix-Na clusters.

Figure 8. Reversed-phase gradient HPLC chromatograms: (a) PET-1 (THF-soluble portion) and the matrixes (b) dithranol, (c) IAA, and (d) DHB.

closest match of HPLC retention times, resulting in the most homogeneous “solid solution” and best MALDI spectrum with regard to S/N. Also, the matrix/polymer combination that had the largest difference in HPLC retention time resulted in segregation and difficulty in obtaining MALDI spectra.

mixture of CH2Cl2/HFIP increases the solubility of some matrixes. When the correct azeotrope mixture ratio is used, the mixture behaves as one solvent during the solvent evaporation process, which helped to avoid sample segregation. When a small amount of THF is added to the azeotrope mixture, severe segregation occurs, resulting in large molar mass distribution discrimination effects in the MALDI spectra beyond that mentioned above. It appears that the higher molar mass oligomers precipitate out before the matrix has a chance to crystallize and for the higher molar mass oligomers to be incorporated within the matrix crystals. When the azeotrope mixture is diluted with CH2Cl2, a different type of discrimination occurs. In this case, the entire molecular mass distribution is affected, and the result is a significant decrease in the overall signal intensity for the polymer. Both attempts to dilute the azeotrope composition resulted in greatly increased shot-to-shot variability. The compatibility of matrix and polymer was explored using relative HPLC retention times. Consistent with previous work in our laboratories, it was found that the matrix/polymer combination that has the closest match of retention times resulted in the best MALDI signal intensity. ACKNOWLEDGMENT

CONCLUSION The analysis of PET is an excellent example of the role of analyte and matrix solubility in the sample preparation process for MALDI MS. The sample of low molar mass oligomers (PET1) displayed minor discrimination effects when the sample was not completely in solution, compared to when the sample was completely in solution. Spectra with fairly representative polymer distributions were obtained for the medium molar mass sample (PET-2) using HFIP as the solvent and dithranol as the matrix. As expected from the broad polydispersity, there was some discrimination at the high molar mass end when the distribution was compared with the SEC distribution. The use of the azeotrope 756 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

We thank William Nichols (Eastman Kodak Co.) for helpful discussions, Lou Harasta (Eastman Kodak Co.) for the PET samples, and Kim Le and Trevor Bryan (Eastman Kodak Co.) for assistance with size-exclusion experiments and handling of HFIP. A.J.H. gratefully acknowledges financial support from Eastman Kodak Co.

Received for review October 5, 2004. Accepted December 8, 2004. AC048525N