Ion Desorption Efficiency and Internal Energy Transfer in Carbon

May 18, 2009 - In a SALDI-MS experiment with a N2 laser (at 337 nm) used as the excitation light ... The Journal of Physical Chemistry C 2016 120 (36)...
0 downloads 0 Views 3MB Size
Download PDF
Anal. Chem. 2009, 81, 4720–4729

Ion Desorption Efficiency and Internal Energy Transfer in Carbon-Based Surface-Assisted Laser Desorption/Ionization Mass Spectrometry: Desorption Mechanism(s) and the Design of SALDI Substrates Ho-Wai Tang, Kwan-Ming Ng,* Wei Lu, and Chi-Ming Che* Department of Chemistry and Open Laboratory of Chemical Biology of the Institute of Molecular Technology for Drug Discovery and Synthesis, The University of Hong Kong, Pokfulam Road, Hong Kong SAR Ion desorption efficiency and internal energy transfer were probed and correlated in carbon-based surface-assisted laser desorption/ionization mass spectrometry (SALDIMS) using benzylpyridinium (BP) salt as the thermometer chemical. In a SALDI-MS experiment with a N2 laser (at 337 nm) used as the excitation light source and with multiwalled carbon nanotubes (CNT), buckminsterfullerene (C60), nanoporous graphitic carbon (PGC), non-porous graphite particles (G), highly oriented pyrolytic graphite (HOPG), or nanodiamonds (ND) as the SALDI substrate, both the desorption efficiency in terms of ion intensity of BP and the extent of internal energy transfer to the ions are dependent on the type and size of the carbon substrates. The desorption efficiency (CNT ∼ C60 > PGC > G > HOPG > ND) in general exhibits an opposite trend to the extent of internal energy transfer (CNT < C60 ∼ PGC < G ∼ HOPG < ND), suggesting that increasing the extent of internal energy transfer in the SALDI process may not enhance the ion desorption efficiency. This phenomenon cannot be explained by a thermal desorption mechanism, and a non-thermal desorption mechanism is proposed to be involved in the SALDI process. The morphological change of the substrates after the laser irradiation and the high initial velocities of BP ions (1100-1400 ms-1) desorbed from the various carbon substrates suggest that phase transition/destruction of substrates is involved in the desorption process. Weaker bonding/interaction and/or a lower melting point of the carbon substrates favor the phase transition/destruction of the SALDI substrates upon laser irradiation, consequently affecting the ion desorption efficiency. Laser desorption/ionization mass spectrometry (LDI-MS) has emerged as an important tool for rapid and sensitive analysis of biomolecules and, hence, for the development of proteomics and * To whom correspondence should be addressed. Fax: (852) 2857 1586 (K.-M.N); (852) 2857 1586 (C.-M.C.). E-mail: [email protected] (K.-M.N); [email protected] (C.-M.C.).

4720

Analytical Chemistry, Vol. 81, No. 12, June 15, 2009

metabolomics.1-3 The key feature of LDI-MS is the sufficient and controllable transfer of laser energy to vaporize/ionize samples without inducing extensive molecular fragmentation.4-7 A major advancement in this technology has been the incorporation of photoabsorbing media for indirect transfer of laser energy to analyte molecules. This has made the laser desorption/ionization process a relatively controllable event and eventually led to the most popular LDI technique, matrix-assisted laser desorption/ ionization (MALDI).7 Although MALDI is well-known for its capability of generating intact macromolecular ions, the presence of serious matrix interference in the low-mass region makes it less usefully applicable to small-molecule analysis. An alternative approach termed matrix-free or surface-assisted laser desorption/ ionization mass spectrometry (SALDI-MS) adopting non-volatile materials for indirect laser energy transfer has been attracting attention for applications to small-molecule analysis in a soft and relatively non-perturbing manner.8,9 The type, form, and size of SALDI substrates are the critical parameters affecting the analytical performance of SALDI-MS in terms of ion generation efficiency. Despite the availability of numerous materials, the majority of SALDI substrates can be grouped into three types, i.e., silicon-based,10-13 carbon-based,14-18 and metal particle-based substrates.19-22 Nevertheless, the proper type of material does not guarantee good SALDI (1) Chaurand, P.; Luetzenkirchen, F.; Spengler, B. J. Am. Soc. Mass Spectrom. 1999, 10, 91–103. (2) Rubakhin, S. S.; Jurchen, J. C.; Monroe, E. B.; Sweedler, J. V. Drug Discovery Today 2005, 10, 823–837. (3) Reyzer, M. L.; Caprioli, R. M. Curr. Opin. Chem. Biol. 2007, 11, 29–35. (4) Fenner, N. C.; Daly, N. R. Rev. Sci. Instrum. 1966, 37, 1068–1070. (5) Zakett, D.; Schoen, A. E.; Cooks, R. G.; Hemberger, P. H. J. Am. Chem. Soc. 1981, 103, 1295–1297. (6) Karas, M.; Bachmann, D.; Hillenkamp, F. Anal. Chem. 1985, 57, 2935– 2939. (7) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299–2301. (8) Peterson, D. S. Mass Spectrom. Rev. 2007, 26, 19–34. (9) Dattelbaum, A. M.; Iyer, S. Expert Rev. Proteomics 2006, 3, 153–161. (10) Wei, J.; Buriak, J. M.; Siuzdak, G. Nature 1999, 399, 243–246. (11) Cuiffi, J. D.; Hayes, D. J.; Fonash, S. J.; Brown, K. N.; Jones, A. D. Anal. Chem. 2001, 73, 1292–1295. (12) Chen, Y.; Vertes, A. Anal. Chem. 2006, 78, 5835–5844. (13) Wen, X.; Dagan, S.; Wysocki, V. H. Anal. Chem. 2007, 79, 434–444. (14) Sunner, J.; Dratz, E.; Chen, Y.-C. Anal. Chem. 1995, 67, 4335–4342. (15) Dale, M. J.; Konchenmuss, R.; Zenobi, R. Anal. Chem. 1996, 68, 3321– 3329. 10.1021/ac8026367 CCC: $40.75  2009 American Chemical Society Published on Web 05/18/2009

performance. Porous silicon surface23,24 and porous tungsten titanium oxides (MTTO)25 exhibit higher ion generation efficiency than their non-porous counterparts, revealing the importance of the substrate surface area. The performance of SALDI-MS is also affected by substrate surface properties, as revealed by a previous work that the SALDI activity is suppressed when amorphous silicon with a rough surface is hydrogenated.26 In another study, the sensitivity of desorption/ ionization on silicon mass spectrometry (DIOS-MS) could be enhanced to the yoctomole level by increasing the surface hydrophobicity.27 The size of the substrates can also affect the SALDI performance. A TiN nanoparticles/glycerol mixture as the substrate desorbs ions with a wider mass range (up to 10 kDa) than microparticles, and the use of bulk TiN with glycerol does not exhibit any SALDI activity.19 In addition, several types of nanowire (ZnO, SiC, SnO2, and GaN) have been shown to have different desorption/ionization laser thresholds, and their ion generation efficiencies can vary by 10-fold upon changing ZnO to GaN.28 Although the discrepancies in SALDI-MS performance have been shown to be highly dependent on the different parameters of the substrates used (e.g., the type, form, and size), the understanding of the desorption/ionization mechanism(s) of SALDI process remains limited. Determination of the extent of internal energy transfer to ions generated by SALDI can provide critical insight into the mechanism(s) of the SALDI process. The mechanism(s) of the SALDI process is believed to be different from that of the MALDI process.8,9,15,19,23,29 A rapid laser-induced heating of substrates, leading to the desorption/ionization of analytes, is widely regarded as the SALDI mechanism. Although most mechanistic studies have focused on the effects of the different forms or size of silicon substrates on the desorption/ionization efficiency,30,31 internal energy transfer in the silicon-based SALDI process has only been recently probed by Vertes and co-workers.32,33 The effect of different SALDI substrates on the extent of the internal energy (16) Hu, L.; Xu, S.; Pan, C.; Yuan, C.; Zou, H.; Jiang, G. Environ. Sci. Technol. 2005, 39, 8442–8447. (17) Cha, S.; Yeung, E. S. Anal. Chem. 2007, 79, 2373–2385. (18) Najam-ul-Haq, M.; Rainer, M.; Szabo´, Z.; Vallant, R.; Huck, C. W.; Bonn, G. K. J. Biochem. Biophys. Methods 2007, 70, 319–328. (19) Schu ¨ renberg, M.; Dreisewerd, K.; Hillenkamp, F. Anal. Chem. 1999, 71, 221–229. (20) Kinumi, T.; Saisu, T.; Takayama, M.; Niwa, H. J. Mass Spectrom. 2000, 35, 417–422. (21) Su, C.-L.; Tseng, W.-L. Anal. Chem. 2007, 79, 1626–1633. (22) Wada, Y.; Yanagishita, T.; Masuda, H. Anal. Chem. 2007, 79, 9122–9127. (23) Alimpiev, S.; Nikiforov, S.; Karavanskii, V.; Minton, T.; Sunner, J. J. Chem. Phys. 2001, 115, 1891–1901. (24) Shen, Z.; Thomas, J. J.; Averbuj, C.; Broo, K. M.; Engelhard, M.; Crowell, J. E.; Finn, M. G.; Siuzdak, G. Anal. Chem. 2001, 73, 612–619. (25) Yuan, M.; Shan, Z.; Tian, B.; Tu, B.; Yang, P.; Zhao, D. Microporous Mesoporous Mater. 2005, 78, 37–41. (26) Alimpiev, S.; Grechnikov, A.; Sunner, J.; Karavanskii, V.; Simanovsky, Y.; Zhabin, S.; Nikiforov, S. J. Chem. Phys. 2008, 128, 014711. (27) Trauger, S. A.; Go, E. P.; Shen, Z.; Apon, J. V.; Compton, B. J.; Bouvier, E. S. P.; Finn, M. G.; Siuzdak, G. Anal. Chem. 2004, 76, 4484–4489. (28) Kang, M.-J.; Pyun, J.-C.; Lee, J.-C.; Choi, Y.-J.; Park, J.-H.; Park, J.-G.; Lee, J.-G.; Choi, H.-J. Rapid Commun. Mass Spectrom. 2005, 19, 3166–3170. (29) Dreisewerd, K. Chem. Rev. 2003, 103, 395–425. (30) Kruse, R. A.; Li, X.; Bohn, P. W.; Sweedler, J. V. Anal. Chem. 2001, 73, 3639–3645. (31) Lewis, W. G.; Shen, Z.; Finn, M. G.; Siuzdak, G. Int. J. Mass Spectrom. 2003, 226, 107–116.

transfer and its correlation with the desorption/ionization efficiency remain to be elucidated. Analytical application of carbon-based SALDI-MS was first reported by Sunner et al. in 1995,14 and the technique has subsequently been applied to the analysis of environmental, biological, and herbal samples16,18 and also to mass spectrometric imaging.17 Carbon exists in different allotropic forms and offers diverse types of substrate materials with different physiochemical properties, such as thermal conductivity, specific heat capacity, and melting point.34,35 Investigating the relationship between desorption/ionization efficiency and the extent of internal energy transfer with different carbon substrates is expected to provide useful information on the fundamental parameters governing the desorption/ionization process. In this investigation, we adopted benzylpyridinium (BP) salts as the “thermometer chemical” to investigate the effects of various carbon substrates on the ion desorption efficiency of carbon-based SALDI-MS and to probe the extent of internal energy transfer in the desorption process based on the survival yield measurement.36-39 The different carbon substrates examined include non-porous graphite particles (particle size HOPG > ND) were attributable to the performance of the various carbon substrates. A plot of the ion intensity of desorbed BP (i.e., a summation of the intensities of parent ions and fragment ions) against laser fluence is shown in Figure 3. The laser fluence threshold for the observable desorption of BP ions (corresponding to the ion intensity of BP at about 200 counts accumulated in a combined spectrum of 150 scans) is also largely dependent on the type and size of the carbon substrates. In this work, ND was found to require a higher laser fluence threshold (at least 64 mJ cm-2) for the desorption of BP ions than HOPG (∼57 mJ cm-2). In contrast, a lower laser fluence threshold of approximately 32 mJ cm-2 was required in the case of G (in the form of microsized particles with a larger surface area than HOPG). The presence of a nanoporous structure (∼30-100 nm) in the microsized graphite particles (i.e., PGC) further reduced the laser fluence threshold, this result is comparable to that of the nanostructured CNT at ∼25 mJ cm-2. Previous studies on silicon-based SALDI also revealed that nanosized silicon substrates,13,33 such as silicon nanowire and nanoparticles, efficiently desorb analytes at a lower laser fluence than bulk silicon substrates. In contrast, the nanosized ND with efficient UV absorption and a large surface area requires the highest laser fluence threshold for the observable ion desorption, whereas C60 in the form of microsized crystals

Figure 3. Effect of laser fluence on the total intensity of benzylpyridinium (BP) ions desorbed from various carbon substrates. The total ion intensities are the summation of intensities of BP ion at m/z 170 and the fragment ion at m/z 91.

Figure 4. Survival yield of benzylpyridinium (BP) ions desorbed from various carbon substrates in a range of laser fluences.

utilizes a lower laser fluence threshold. These results demonstrate that the larger surface area of carbon substrates is not the only reason for the higher ion desorption efficiency. Internal Energy Transfer in Carbon-Based SALDI. To rationalize the effect of different carbon substrates on the ion desorption efficiency, the extent of internal energy transfer to BP ions generated by SALDI was determined from the survival yield (SY) of the BP ions. A plot of the SY against the laser fluence is illustrated in Figure 4. The SY is largely dependent on the type and size of the carbon substrates, but relatively less sensitive to the variation of laser fluence, except for CNT and C60 exhibiting a slight decrease SY (∼5%) as the laser fluence increased from 25 to 40 mJ cm-2. The average SY of ions desorbed from ND in the form of nanosized particles had the lowest value of 34.6% ± 3.8%, which corresponds to the highest extent of internal energy transfer, whereas CNT,

another type of nanostructured carbon substrate, exhibited the highest SY at 96.4% ± 2.3%. G and HOPG are on microsize scale, with both belonging to the same type of graphite substrates, and exhibited similar and intermediate SY at the values of 65.8% ± 1.3% and 66.0% ± 3.6%, respectively. However, the presence of a nanoporous structure in the graphite particles (i.e., PGC) significantly increased the SY to 92.5% ± 1.4%. Moreover, microsized crystals of C60 showed a relatively high SY, at 88.7% ± 2.4%. The energy-dependent dissociation rate coefficient curve of BP ions derived from the RRKM formalism is given in Supporting Information (Figure S4). With the dissociation time frame of 150 ns, the experimentally measured dissociation rate coefficients of BP ions desorbed from the various carbon substrates were determined to be 2.45 × 105 s-1 (CNT), 5.18 × 105 s-1 (PGC), 8.04 × 105 s-1 (C60), 2.79 × 106 s-1 (G), 2.78 × 106 s-1 (HOPG), and 7.10 × 106 s-1 (ND). By correlating the dissociation rate coefficients with the energy-dependent dissociation rate coefficient curve, the average internal energies of BP ions desorbed from the different carbon substrates were obtained and are in the order of ND (5.86 ± 0.04 eV) > G (5.51 ± 0.02 eV) ∼ HOPG (5.50 ± 0.05 eV) > C60 (5.09 ± 0.07 eV) ∼ PGC (4.96 ± 0.05 eV) > CNT (4.71 ± 0.15 eV). A similar order of the internal energy transfer was also observed with other thermometer ions, including 4F-BP and 4Cl-BP; the results are given in Table S1 (Supporting Information). The thermal desorption mechanism could partly account for the different extent of internal energy transfer to ions generated from the carbon-based SALDI using the different carbon substrates. For instance, the higher internal energy transfer from ND compared with that from G/HOPG may be attributed to the smaller specific heat capacity (0.68 vs 0.71 J g-1 K-1) and thermal conductivity (0.12 vs 19.5 W cm-1 K-1) of ND versus G/HOPG,46-48 resulting in higher temperature upon the laserinduced heating. However, although PGC, G, and HOPG belong to the same type of graphite materials, the presence of the nanoporous structure in PGC could lead to significantly lower internal energy transfer than that in the non-porous graphite (G and HOPG). In addition, the nanostructured CNT also showed a lower internal energy transfer when compared with the other carbon substrates. In fact, laser-induced heating would be expected to be more efficient in the carbon nanosubstrates than their bulk counterparts due to the thermal confinement effect. However, the lower internal energy transfer of nanostructured carbon substrates (CNT and PGC) implied that additional process(es) could be involved in dissipating the energy. It is worth noting that the desorption efficiency in terms of the ion intensity of BP exhibits in general a trend which is opposite to the extent of internal energy transfer. For instance, the ion intensity of BP desorbed from ND is at least 20 times lower than that from HOPG at a similar laser fluence, whereas the internal energy transfer from ND is at least 0.3 eV higher than that from HOPG. The results showed that increase in (46) CRC Handbook of Chemistry & Physics, 88th ed. [Online]; Lide, D. R., Ed.; CRC Press/Taylor and Francis: Boca Raton, FL, 2008. (47) Moelle, C.; Werner, M.; Szu ¨ cs, F.; Wittorf, D.; Sellschopp, M.; von Borany, J.; Fecht, H.-J.; Johnston, C. Diamond Relat. Mater. 1998, 7, 499–503. (48) Angadi, M. A.; Watanabe, T.; Bodapati, A.; Xiao, X.; Auciello, O.; Carlisle, J. A.; Eastman, J. A.; Keblinski, P.; Schelling, P. K.; Phillpot, S. R. J. Appl. Phys. 2006, 99, 114301.

Analytical Chemistry, Vol. 81, No. 12, June 15, 2009

4725

the extent of internal energy transfer does not necessarily increase the ion desorption efficiency. A possible reason could be due to the different binding affinities of the BP ions toward the various carbon substrates, resulting in different partitioning of internal energy to internal modes of BP ions and, thus, leading to different extent of fragmentation. Another reason could be the involvement of a non-thermal desorption mechanism in the SALDI process, which allows the desorption of ions from a SALDI substrate through a lower energy transfer process. In fact, a recent study of DIOS-MS found that surface restructuring of porous silicon upon laser irradiation is involved in the desorption/ionization process.49 Another study using a thin metal film on porous alumina as the SALDI substrate revealed that the melting point of the metal affects the ion generation efficiency.22 Previous studies showed that the laser fluence threshold for damaging graphite is approximately 4 times lower than that for diamond,50 and the ablation rate for HOPG is about 6 times faster than that for a diamond thin film at a similar laser fluence.51 These phenomena could be attributed to the lesser energy required for dislocating the graphite layer structure than that for breaking the giant covalent linkage in diamond. The higher susceptibility of HOPG to laser-induced damage might account for its better ion desorption efficiency than ND. In fact, examination on the surface of HOPG by scanning electron microscope revealed that clusters of spherical particles (∼10 nm diameter) were formed after N2 laser irradiation, as shown in Figure 5a. We suggest that, in addition to thermal desorption process, the process of phase transition/destruction of the substrate is involved in the desorption mechanism(s) of the carbon-based SALDI. The melting point of CNT (∼1600-3200 K) is significantly lower than that of G (3800-4762 K).46,52,60 In addition, C60 crystals sublime at ∼1000 K.53 A theoretical study showed that the melting point of graphite particles would decrease with its size from the bulk to nanoscale.54 Taken together, these results suggest that PGC with a nanoporous structure has a lower melting point than that of the non-porous graphite (G). Carbon substrates with weaker bonding/interaction (e.g., sp2-hybridized carbon in graphite with interlayer π-π interaction and C60 molecules with intermolecular π-π interaction vs sp3hybridized carbon in diamond with strong three-dimensional carbon-carbon network) and lower melting point (e.g., nanostructured CNT/PGC with lower melting points vs microsized G/HOPG with higher melting points) favor the process of phase transition/destruction of SALDI substrates upon the laser irradiation, thus leading to higher ion desorption efficiency. The process of phase transition/destruction of SALDI substrates dissipates the laser energy and, thus, results in decreased internal energy transfer. Laser vaporization/sputtering of carbon substrates has long been reported in the literature.55 Besides HOPG (Figure 5a), (49) Northen, T. R.; Woo, H.-K.; Northen, M. T.; Nordstro ¨m, A.; Uritboonthail, W.; Turner, K. L.; Siuzdak, G. J. Am. Soc. Mass Spectrom. 2007, 18, 1945– 1949. (50) Reitze, D. H.; Ahn, H.; Downer, M. C. Phys. Rev. B 1992, 45, 2677–2693. (51) Windholz, R.; Molian, P. A. J. Mater. Sci. 1997, 32, 4295–4301. (52) Sun, C. Q. Prog. Solid State Chem. 2007, 35, 1–159. (53) Rausch, H.; Braun, T. Chem. Phys. Lett. 2001, 350, 15–18. (54) Yang, C. C.; Li, S. J. Phys. Chem. C 2008, 112, 1423–1426. (55) Krajnovich, D. J. J. Chem. Phys. 1995, 102, 726–743.

4726

Analytical Chemistry, Vol. 81, No. 12, June 15, 2009

destruction of CNT (Figure 5b) after laser irradiation has also been observed. No observable changes were noted on the surface of any of the particle-based materials, including ND, G, PGC, and C60, probably because their surfaces are too rough for observable changes. In fact, phase vaporization of the graphite-based substrates (i.e., G, HOPG, and PGC) upon laser irradiation has also been revealed by the presence of a series of carbon cluster ions (differing by 12.0 Da) in the background mass spectra recorded in the negative mode, as shown in Figure S5 in the Supporting Information. In addition, the ion peak of C60 at m/z 720 was recorded in the background mass spectrum of C60 (Supporting Information Figure S5). However, no obvious carbon cluster ion peak was observed in the background mass spectra of ND and CNT (results not shown) upon the laser irradiation, which may be due to insufficient laser energy for the ionization of the stable ND and CNT materials sputtered from the substrate surface. Another possible reason accounting for the different ion desorption efficiencies and the extent of internal energy transfer to the BP ions is the effect of a charge-transfer process/ interaction between the carbon substrates and the BP ions. Lower ionization potential of the carbon substrate (Table 1) would favor electron transfer to the BP ions, thus suppressing the ion intensity. However, in correlating the ionization potentials of the carbon substrates to the intensities of BP ions, and to the extent of internal energy transfer, we found that the order of ionization energies of the carbon substrates (i.e., ND (6.9-8.07 eV) ∼ C60 (7.6 eV) > CNT (5.3 eV) > G (4.39 eV)) does not show any correlation to the order of the intensities of BP ions (i.e., CNT ∼ C60 > G > ND) and to the extent of internal energy transfer to the BP ions (ND > G > C60 > CNT). This result reveals that the charge-transfer process is not significant in affecting the ion desorption efficiency and the extent of internal energy transfer to the BP ions desorbed from the carbon substrates. Both the effects of organic solvent (toluene for dissolution of C60) and substrate deposition method (sample-on-the-top method vs premixing method) on the ion desorption efficiency and the extent of internal energy transfer for different carbon substrates, including CNT, C60, PGC, G, and ND, have also been investigated. We found that toluene for dissolving C60 substrate could slightly enhance the ion desorption efficiency by less than 25% but showed negligible (56) Karas, M.; Bahr, U.; Fournier, I.; Glu ¨ ckmann, M.; Pfenninger, A. Int. J. Mass Spectrom. 2003, 226, 239–248. (57) Yi, W.; Lu, L.; Zhang, D.-L.; Pan, Z.-W.; Xie, S. S. Phys. Rev. B 1999, 59, R9015–R9018. (58) Allen, K.; Hellman, F. Phys. Rev. B 1999, 60, 11765–11772. (59) Yu, R. C.; Tea, N.; Salamon, M. B.; Lorents, D.; Malhotra, R. Phys. Rev. Lett. 1992, 68, 2050–2053. (60) Begtrup, G. E.; Ray, K. G.; Kessler, B. M.; Yuzvinsky, T. D.; Garcia, H.; Zettl, A. Phys. Rev. Lett. 2007, 99, 155901. (61) Shiraishi, M.; Ata, M. Carbon 2001, 39, 1913–1917. (62) Kazaoui, S.; Minami, N.; Matsuda, N.; Kataura, H.; Achiba, Y. Appl. Phys. Lett. 2001, 78, 3433–3435. (63) Lichtenberger, D. L.; Nebesny, K. W.; Ray, C. D.; Huffman, D. R.; Lamb, L. D. Chem. Phys. Lett. 1991, 176, 203–208. (64) Kuroda, H. Nature 1964, 201, 1214–1215. (65) Lenzke, K.; Landt, L.; Hoener, M.; Thomas, H.; Dahl, J. E.; Liu, S. G.; Carlson, R. M. K.; Möller, T.; Bostedt, C. J. Chem. Phys. 2007, 127, 084320. (66) Sque, S. J.; Jones, R.; Goss, J. P.; Briddon, P. R.; Öberg, S. J. Phys.: Condens. Matter 2005, 17, L21–L26.

Figure 5. Scanning electron micrographs of (a) clusters of spherical particles formed on the surface of HOPG after N2 laser irradiation (laser fluence ) 109 mJ cm-2) and (b) destruction of multiwalled carbon nanotubes (CNT) after N2 laser irradiation (laser fluence ) 43 mJ cm-2). The carbon substrates were irradiated with the N2 laser at a frequency of 10 Hz for 1000 laser pulses.

effect on the internal energy transfer (Supporting Information Figure S6). In a comparison between the sample-on-thetop method and the premixing method (Supporting Information Figure S7), the latter generally showed a higher ion desorption efficiency. The result could be attributed to the larger interaction/mixing surface area between the carbon substrates and the BP analytes in the premixing process. Although higher ion desorption efficiency could be obtained from the premixing method, the order of ion desorption efficiency with the carbon

substrates followed the same order of CNT ∼ C60 > PGC > G > ND. This is consistent with the results obtained from the sample-on-the-top method. In addition, the different deposition methods showed insignificant effect on the order of the extent of internal energy transfer (i.e., ND > G > C60 ∼ PGC > CNT). Thus, both the type and size of the carbon substrate are more important in governing the ion desorption efficiency and the internal energy transfer in the SALDI process than the substrate deposition method. Analytical Chemistry, Vol. 81, No. 12, June 15, 2009

4727

Table 1. Selected Physical and Physiochemical Properties of Multiwalled Carbon Nanotubes (CNT), Buckminsterfullerene (C60), Graphite (G),a and Nanodiamonds (ND) CNT -1

-1

specific heat capacity (J g K ) at ∼300 K thermal conductivity (W cm-1 K-1) at ∼300 K melting point (K) laser fluence threshold for damaging carbon substrates by using a Nd:YAG laser (λ ) 620 nm, pulse width ) 90 fs) (J cm-2) ionization potential (eV)

b

C60 c

G d

ND e

0.45 0.25b 1600-3200h,i l

0.7 0.004f 1000 (sublimation)j l

0.71 19.5d 3800-4762d,h 0.13m

0.68 0.12g ∼5000 (at 8.5 GPa)k 0.63m

5.3n

7.6o

4.39p

6.9-8.07q

a Graphite particles (G): highly oriented pyrolytic graphite (HOPG) and nanoporous graphitic carbon (PGC) are graphite-based materials. In general, G and HOPG would have very similar physical and physiochemical properties. PGC would possibly have a lower melting point than that of G/HOPG due to the presence of nanoporous structure (see text and ref 54). b Refer to ref 57. c Ref 58. d Ref 46. e Ref 47. f Ref 59. g Ref 48. h Ref 52. i Ref 60. j Ref 53. k Ref 54. l No literature value was found for CNT and C60 at the same experimental conditions as reported for graphite (ref 50) and diamond (ref 50) for comparison. m Refer to ref 50. n Assuming the difference (i.e., 0.35 eV) between the work function (5.05 eV) and the ionization potential (5.4 eV) of a single-walled carbon nanotube (refs 61 and 62) is similar to that of the multiwalled carbon nanotube, the ionization potential of CNT is estimated to be 5.3 eV based on the work function (4.95 eV) of CNT (ref 61). o Refer to ref 63. p Ref 64. q The ionization potential of ND is estimated to be in the range of 6.9-8.07 eV. The upper bound value of the ionization potential of ND is assumed to be the ionization potential (8.07 eV) of a diamondoid member with five diamond cages (i.e., [1,(2,3)4]-pentamantane) (ref 65). Assuming the magnitude of the underestimation (3.0 eV) of the calculated ionization potential (4.57 eV) of C60 (ref 66) from its experimental value (7.6 eV) (ref 63) is applicable to the calculated ionization potential (3.90 eV) (ref 66) of a diamond sample obtained at the same level of theory, the lower bound value of the ionization potential of ND is estimated to be 6.9 eV.

Mean Initial Velocity of BP Ions. Previous studies on the fundamental aspects of MALDI process suggested that a high translational energy of organic matrix ions formed upon laser irradiation, as determined from a DE experiment in linear TOF, could be the result of a phase transition/explosion process.29,56 Here, the mean initial velocities of BP ions desorbed over a range of laser fluences were determined using the DE method, and the results are summarized in the Supporting Information (Table S2). The data reveal that the mean initial velocities of the desorbed BP ions are relatively less sensitive to the type and size of the carbon substrates, as well as the variation of laser fluence. The mean initial velocities of BP ions desorbed from the carbon substrates are comparable [1191 ± 51 m s-1 (CNT) ∼ 1210 ± 14 m s-1 (PGC) ∼ 1286 ± 32 m s-1 (C60) ∼ 1280 ± 37 m s-1 (G) ∼ 1209 ± 79 m s-1 (HOPG)], except for ND having a relatively high value at 1386 ± 56 m s-1. The values are also comparable to or even greater than that (1298 ± 38 m s-1) of BP ions desorbed from CHCA. The high translational energy of BP ions desorbed from the carbon substrates also suggests that a phase transition/destruction process is involved in the desorption mechanism(s) of the carbon-based SALDI. CONCLUSION We have shown that different carbon substrates, including CNT, C60, PGC, G, HOPG, and ND, have significantly different ion desorption efficiencies and extent of internal energy transfer to ions generated in the SALDI process. Using BP salt as the thermometer chemical, both the desorption efficiency and the extent of internal energy transfer are dependent on the type and size of the carbon substrate. The desorption efficiencies in terms of ion intensity, measured at a similar laser fluence, are in the order CNT ∼ C60 > PGC > G > HOPG > ND. This trend can be partially accounted for by the increasing surface area of PGC > G > HOPG. Internal energy transfer in the carbon-based SALDI was determined based on the survival yield of the BP ions. With the use of RRKM formalism, the average 4728

Analytical Chemistry, Vol. 81, No. 12, June 15, 2009

internal energies of the ions desorbed over a range of laser fluences (∼23-123 mJ cm-2) were derived and found to be relatively less sensitive to the variation in laser fluence and are in the order ND (5.86 eV) > G (5.51 eV) ∼ HOPG (5.50 eV) > C60 (5.09 eV) ∼ PGC (4.96 eV) > CNT (4.71 eV). The order of ion desorption efficiency displays a generally opposite trend to that of the extent of internal energy transfer to ions generated by the SALDI. This phenomenon is not accounted for by a thermal desorption mechanism and implies that a non-thermal desorption mechanism is involved in the SALDI process. Both the high mean initial velocities of the desorbed BP ions and the morphological change of the carbon substrates after N2 laser irradiation suggest that phase transition/destruction of the carbon substrate is involved in the SALDI process. The BP ions generated in the carbonbased SALDI process had mean initial velocities in the range of 1100-1400 m s-1, comparable to that of ∼1300 m s-1 of the BP ions formed in MALDI process mainly involving a phase transition/explosion process. The inverse relationship between the ion desorption efficiency and the internal energy transfer reveals that the increase in the extent of internal energy transfer in the SALDI process may not be able to enhance the ion desorption efficiency. Other factors, such as the weaker bonding/interaction and/or lower melting point of the SALDI substrate, favoring the ease of phase transition/destruction of the substrate upon laser irradiation, could play more important role(s) in enhancing the ion desorption efficiency. In the present study, the investigation of the relationship between the ion desorption efficiency and the extent of internal energy transfer in the carbon-based SALDI process has revealed certain important criteria for the design of SALDI substrates. The ease of phase transition/destruction of SALDI substrates upon laser irradiation is critical for the development of SALDI substrates with better desorption efficiency, and hence higher detection sensitivity, for SALDI-MS analysis.

ACKNOWLEDGMENT We acknowledge Professor Chun-Wai Tsang and Professor Kwok-Yin Wong of the Hong Kong Polytechnic University for access to the Waters micro MX MALDI-TOF mass spectrometric system. This project is supported by the Area of Excellence Scheme (AoE/P-10/01) administered by the University Grant Council (Hong Kong SAR, China). We thank Ms. Amy S. L. Wong and Mr. Frankie Y. F. Chan of the Electron Microscope Unit of The University of Hong Kong for technical assistance in SEM measurements. We acknowledge Dr. Charlie Liu for providing the porous graphitic carbon. We also thank

Dr. Daniel Kenny of Waters Corporation for providing assistance to derive the flight time of desorbed ions from the m/z value of ion peaks recorded in MALDI-TOF mass spectra. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review December 15, 2008. Accepted April 7, 2009. AC8026367

Analytical Chemistry, Vol. 81, No. 12, June 15, 2009

4729