Anal. Chem. 2002, 74, 3911-3914
Sample Deposition Device for Off-Line Combination of Supercritical Fluid Chromatography and Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry Josef Planeta, Pavel R ˇ ehulka, and Josef Chmelı´k*
Institute of Analytical Chemistry, Academy of Sciences of the Czech Republic, Vever 97, 611 42 Brno, Czech Republic
A new sample deposition device for off-line SFC-MALDI combination of supercritical fluid chromatography and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry was assembled. This device was successfully applied to the detailed characterization of synthetic silicone oils. SFC was used to separate samples of silicone oils on micropacked capillary columns and to determine their molecular mass distribution. The separated fractions for the identification studies were obtained from SFC runs at defined time intervals. Using the constructed deposition device, these fractions were sprayed directly from the restrictor on the target probe covered with a proper matrix. MALDI-TOF MS was used for the identification of individual oligomers in the separated fractions and also in the unfractionated sample. The determined molecular mass distributions based on supercritical fluid chromatography with flame ionization detector, MALDI-TOF MS, and combined SFC-MALDI measurements were compared and the results were in a good agreement. The sample deposition device is based on a common plotter unit, complemented by a microcontroller PIC16C84. The unit is connected by an RS-232 interface to a PC with the main control software running under MS Windows. The new sample deposition device made the off-line combination SFC-MALDI simpler, faster, and more sensitive. Characterization of polymers is not a trivial task. After solving chemical composition, determination of polymer size distribution is a next integral step. For evaluation of the accuracy of molecular mass distributions of polymers determined by a given technique, it is also advantageous to compare the obtained molecular mass distributions with results measured by other methods. Combinations of MS with separation techniques (e.g., steric exclusion chromatography, HPLC, supercritical fluid chromatography (SFC)) are often used for characterization of polymers.1 Several mass spectrometric techniques can be used for this purpose in the case of SFC. An on-line coupling of an SF chromatograph to an MS detector seems to be straightforward, and the on-line SFC-MS was successfully used.2-5 However, some mass spectrometers have limitations2 in determination of high (1) Tomer, K. B. Chem. Rev. 2001, 101, 297-328. 10.1021/ac020085h CCC: $22.00 Published on Web 07/04/2002
© 2002 American Chemical Society
molecular mass oligomers. An off-line MS analysis of fractions separated by SFC can be performed in various ways. In the case of fractions containing shorter oligomers, GC/MS is a suitable and sensitive technique.6 Higher oligomer fractions require other MS techniques for mass determination. One of these MS techniques, matrix-assisted laser desorption/ ionization time-of-flight mass spectrometry (MALDI-TOF MS), is often used for characterization of polymers. Although the molecular mass and in some cases even chemical composition of particular oligomers7 can be determined by MALDI-TOF MS, determination of the accurate molecular mass distribution of synthetic polymers is difficult to achieve8,9 because of mass discrimination caused for several reasons (e.g., sample properties, sample preparation, instrumental factors10-12). Off-line combinations of MALDI-TOF MS and liquid chromatography can be performed in various ways: transfer (pipetting or spraying13) of aliquots of separated fractions onto classical MALDI targets, various membranes (e.g., Teflon and polyethylene14), or porous frits.15 Similar ways can be used for off-line combination SFCMALDI. The off-line combination SFC-MALDI based on pipetting of aliquots of separated fractions (trapped in a proper solvent) on the MALDI target covered with a proper matrix was found useful to determine the molecular mass distributions of silicone oils.6,16 (2) Combs, M. T.; Ashraf-Khorassani, M.; Taylor, L. T. J. Chromatogr., A 1997, 785, 85-100. (3) Just, U.; Mellor, F.; Keidel, F. J. Chromatogr., A 1994, 683, 105-113. (4) Pinkston, J. D.; Chester, T. L. Anal. Chem. 1995, 67, A650-A656. (5) Sjo ¨berg, P. J. R.; Markides, K. E. J. Chromatogr., A 1999, 855, 317-327. (6) Chmelı´k, J.; Konecˇny´, P.; Planeta, J.; Zdra´hal, Z.; Vejrosta, J.; Chmelı´k, J. HRC-J. High. Resolut. Chromatogr. 2000, 23, 502-506. (7) Montaudo, G.; Montaudo, M. S. Anal. Chem. 1994, 66, 4366-4369. (8) Axelsson, J.; Scrivener, E.; Handleton, D. M.; Derrick, P. J. Macromolecules 1996, 29, 8875-8882. (9) Nielen, M. W. F. Mass Spectrom. Rev. 1999, 18, 309-344. (10) Vorm, O.; Roepstorff, P.; Mann, M. Anal. Chem. 1994, 66, 3281-3287. (11) McEwen, C. N.; Jackson, C.; Larsen, B. S. Int. J. Mass Spectrom. Ion Processes 1997, 160, 387-394. (12) Danis, P. O.; Karr, D. E. Org. Mass Spectrom. 1993, 28, 923-925. (13) Esser, E.; Keil, C.; Bruan, D.; Montag, P.; Pasch, H. Polymer 2000, 41, 4039-4046. (14) Stevenson, T. I.; Loo, J. A.; Greis, K. D. Anal. Biochem. 1996, 262, 99109. (15) Zhan, Q.; Gusev, A.; Hercules, D. M. Rapid Commun. Mass Spectrom. 1999, 13, 2278-2283. (16) Chmelı´k, J.; Planeta, J.; Rˇ ehulka, P.; Chmelı´k, J. J. Mass Spectrom. 2001, 36, 760-770.
Analytical Chemistry, Vol. 74, No. 15, August 1, 2002 3911
Figure 1. A new sample deposition device for off-line combination of SFC-MALDI.
They were separated by SFC into several fractions with baseline resolved peaks, and GC/MS and MALDI-TOF MS were used for identification of particular peaks. The SFC fractions taken at defined time intervals were trapped in acetone during the SFC runs, when the end of the restrictor was pulled out from the flame ionization detector (FID) and inserted into a trapping vial.6 However, this procedure leads to certain sample dilution. Therefore, a new sample deposition device was assembled, which makes the transfer of samples separated by SFC to a MALDI target more efficient. In this work, we used direct spraying of the separated fractions from the restrictor on the target covered with a proper matrix, which increased the sensitivity of the procedure. EXPERIMENTAL SECTION Materials. R-Cyano-4-hydroxycinnamic acid, acetone, and ethanol were obtained from Sigma. Lukoil M50 (sample of silicone oil) was provided by the manufacturer Lucebnı´ za´vody (Kolı´n, Czech Republic). The ODS sorbent (Biospher Si C18E) was purchased from Labio a.s. (Prague, Czech Republic). SFC. The Varian 3700 gas chromatograph was arranged as an SFC unit. An air-actuated Valco valve ACI4 with a 60-nL inner loop was used for sample injection. For faster switching of the valve, helium at 0.5 MPa was used as the driving gas. The output of the valve was connected by a fused-silica capillary (85-µm i.d./ 30-cm length) with a micropacked capillary column, made from fused-silica capillary (320-µm i.d./15-cm length) and packed with 5-µm ODS particles. The column was prepared by the technology of packing in supercritical carbon dioxide.17 The end of the column was connected to an integral restrictor. Connections were realized by the ZU1T Valco union and PEEK tubing (0.01-in. i.d., redrilled from one side to the outer diameter of the column). The integral restrictor was made from 85-µm-i.d fused-silica capillary. One end was fused in the flame and then polished on a ceramic disk until the required flow rate was reached. This restrictor exhibited good peak resolution and a spikeless baseline.18 A sample of silicone oil in acetone (10 mg/mL solution) was separated using pure CO2 as mobile phase with a linear pressure gradient 8-36 MPa/30 min at 80 °C. (17) Tong, D.; Bartle, K. D.; Clifford, A. A. J. Microcolumn Sep. 1994, 6, 249255. (18) Guthrie, E. J.; Schwartz; H. E. J. Chromatogr. Sci. 1986, 24, 236-241.
3912 Analytical Chemistry, Vol. 74, No. 15, August 1, 2002
Figure 2. Detail of the restrictor holder: (1) restrictor holder (which replaced a pen), (2) MALDI target probe with spots covered with the matrix, and (3) plotter unit.
Figure 3. Schematic drawing of the restrictor holder: (1) aluminum body of the holder, (2) PEEK tubing 1/16-in. o.d., 0.010-in. i.d., (3) integral restrictor, (4) outlet for CO2, and (5) silicone rubber seal.
MALDI-TOF MS. Measurements were performed with a Kompact MALDI 4 (Shimadzu-Kratos, Manchester, U.K.) equipped with a nitrogen laser (337 nm) in a linear mode with pulsed extraction. R-cyano-4-hydroxycinnamic acid was used as a matrix without further purification. Matrix was dissolved in acetone (concentration of 5 mg/mL). Silicone oil was analyzed either as an unfractionated sample (concentration of 1 mg/mL) or as fractions resulting from the preceding SFC fractionations. A thin-layer method10 was used as a sample preparation for MALDI measurements. The aliquots (0.5 µL) of the matrix solutions were applied on the surface of the probe and allowed to
Figure 4. Separation of Lukoil by SFC-FID and the time periods of four fractions used for off-line combination SFC-MALDI. Sample: 10 mg/mL solution of Lukoil M50 in acetone. Injection: 60-nL inner loop, time injection 0.1 s. Column: i.d. 320 µm, length 15 cm, packed with 5-µm ODS particles. Oven: 80 °C. Mobile phase: CO2. Restrictor: integral. Detector: FID (40 °C). Pressure program: 8 MPa (5min), 8-36 MPa (30 min) with linear gradient. The numbers above the peaks correspond to the number of Si atoms in the oligomers.
dry at room temperature. Subsequently, the sample solution in acetone was applied (0.5 µL) and dried at room temperature. The SFC fractions were sprayed directly from the restrictor on the probe covered with R-cyano-4-hydroxycinnamic acid prepared by the same procedure as described above. SFC-MALDI Sample Deposition Device. A commercial plotter XY 4140 (Laboratornı´ Prˇ´ıstroje, Prague) was rearranged as an SFC-MALDI deposition device (Figure 1). A single-chip microcontroller PIC16C84 and RS232/TTL signal convertor MAX232A were added to the electronics of the plotter. Also, a MALDI target holder was mounted on the plotter and an integral restrictor holder was placed instead of a pen (Figure 2). The unit is connected by an RS-232 interface with a PC, where the main control software runs under the MS Windows operating system. The software especially developed for this device allows manual or programmable repositioning of the restrictor upon the spots and up/down movement. A silicone rubber tube is used as a seal between the target surface and the holder of the restrictor (Figure 3). The slightly greater inner diameter of the tube than the length of the spot prevents spot-to-spot contamination. The distance between the end of the restrictor and the target was optimized. The optimum distance of 2.5-3 mm gave the highest signal of MALDI-TOF MS analysis. After the optimization, the observed diameter of the zone of trapped analytes on the spot was ∼1 mm. RESULTS AND DISCUSSION Separation of the silicone oil sample was first performed by SFC-FID (Figure 4). The end of the restrictor was placed in the FID detector, which was heated to 40 °C. Because the SFC-MALDI sample deposition device does not provide restrictor heating, it was also necessary to avoid heating of the FID. As pointed out before,19 the flow rate through the restrictor depends on its temperature. It was observed that the low heating temperature slightly increased the noise in the chromatogram in comparison to previously published results.6,16 Then, the restrictor was pulled out from the FID and placed into the holder in the SFC-MALDI deposition device. A MALDI (19) Berger, T. A.; Toney, C. J. Chromatogr. 1989, 465, 157-167.
Figure 5. MALDI-TOF spectrum of 0.5 µL of Lukoil M50 solution (concentration 1 mg/mL in acetone) using R-cyano-4-hydroxycinnamic acid (concentration 5 mg/mL in acetone) as a matrix. The numbers above the peaks correspond to the number of Si atoms contained in the oligomers.
target with a prepared dried matrix layer on the probe was inserted in the deposition device, and a new SFC analysis under the same conditions was started. Four fractions at defined time intervals (see Figure 4) were sprayed directly from the restrictor on the target probe covered with the matrix. The same silicone oil sample was characterized by MALDITOF MS (Figure 5). From the multipeak self-calibration7 it followed that the sample was a poly(dimethylsiloxane) (PDMS) with fully methylated end groups, which corresponded to the producer’s data. The measured MALDI-TOF data were used for calculation of the polymer size distribution in terms of Mn, Mw, and polydispersity (Table 1). The analysis of four SFC fractions was performed by MALDI-TOF MS (Figure 6). The first and the last peak in fraction 1 were identified as Si11 and Si15, respectively. The first and the last peaks in fraction 2 were identified as Si15 and Si29, respectively. The first and the last peaks in fraction 3 were identified as Si29 and Si49, respectively. The first and the last peaks in fraction 4 were identified as Si49 and Si77, respectively. The minor peaks observed in fractions 3 (for oligomers of
