Anal. Chem. 2005, 77, 5364-5369
Requirements for Laser-Induced Desorption/ Ionization on Submicrometer Structures Shoji Okuno,† Ryuichi Arakawa,†,‡ Kazumasa Okamoto,§ Yoshinori Matsui,§ Shu Seki,§ Takahiro Kozawa,§ Seiichi Tagawa,§ and Yoshinao Wada*,†,|
Wada Project Laboratory, Japan Science and Technology Agency, Innovation Plaza Osaka, 3-1-10 Technostage, Izumi, Osaka, Japan, Department of Applied Chemistry, Kansai University, 3-3-35 Yamatecho, Suita, Osaka 564-8680, Japan, The Institute of Scientific and Industrial Research, Osaka University, 8-1 Mihogaoka, Ibaraki, Osaka 567-0047, and Osaka Medical Center and Research Institute for Maternal and Child Health, 840 Murodo-cho, Izumi, Osaka 594-1101, Japan
Laser-induced and matrix-free desorption/ionization on various submicrometer structures was investigated. First, to examine the effect of surface roughness on ionization, a silicon wafer or stainless steel was scratched with sandpaper. The fluences of a 337-nm nitrogen laser, required for ionization of synthetic polymers and reserpine, were markedly reduced on the scratched stainless steel or silicon as compared to the corresponding untreated surface. Next, arrays of submicrometer grooves, which had been lithographically fabricated on a silicon wafer, yielded protonated angiotensin, and the morphologic orientation demonstrated the positive relation between the laser and groove directions for promoting ionization. The fabricated structure also suggested the submicrometer, but not smaller, or nanometer, structures to be a key factor in direct desorption/ionization on rough surfaces. Finally, submicrometer porous structures of alumina or polyethylene yielded intense molecular ion signals of angiotensin and insulin, in response to direct UV irradiation, when the surface was coated with Au or Pt. The coating provided the additional advantage of prolonged activity for a porous alumina chip, exceeding a month even when the chip was left in the open air. These results indicate that laser-induced desorption/ionization of organic compounds can be implemented on submicrometer structures with an Au- or Pt-coated surface irrespective of the basal materials. Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS)1-4 allows analysis of the intact structures of biological macromolecules5-7 and synthetic polymers8-11 and is currently * To whom correspondence should be addressed. Phone: +81-725-56-1220. Fax: +81-725-57-3021. E-mail:
[email protected]. † Japan Science and Technology Agency. ‡ Kansai University. § Osaka University. | Osaka Medical Center and Research Institute for Maternal and Child Health. (1) Karas, M.; Bachmann, D.; Hillenkamp, F. Anal. Chem. 1985, 57, 2935-9. (2) Karas, M.; Bachmann, D.; Hillenkamp, F. Int. J. Mass Spectrom. 1987, 78, 53-68. (3) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-301. (4) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T. Rapid Commun. Mass Spectrom. 1988, 2, 151-3. (5) Chance, D. H. Chem. Rev. 2001, 101, 445-77. (6) Aebersold, R.; Goodlett, D. R. Chem. Rev. 2001, 101, 269-95.
5364 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005
indispensable in life science research because of other unique merits including its high sensitivity, tolerance to salts, applicability to complex mixtures, and throughput. Although the fundamental processes of ion generation and desorption are still poorly understood, the roles of the chemical matrix in MALDI are suggested to be isolation of analyte molecules, absorption of photon energy and its transfer to analytes, and charge-transfer reactions in the gas phase.12 In a typical MALDI analysis using a UV laser, small organic acid matrixes with strong absorbance at the laser wavelength are mixed with analytes. Matrix selection and optimization of the sample preparation protocol are the important steps in analysis but are still empirical procedures.13,14 Recently, a porous silicon surface has been used for ionization, and this method is called desorption/ionization on silicon (DIOS).15 Porous silicon is derived from oxidative degradation of the silicon surface, and the resulting porosity or roughness is presumed to be necessary for desorption/ionization.16-18 In DIOSMS, the analytes in solution are directly deposited on the silicon surface without a chemical matrix. Thus, DIOS is often termed “matrix-free” LDI to highlight its advantages including lack of lowmass interference due to matrix ions.15,16,19-22 However, the enhanced surface area and altered electric and thermal properties (7) Rappsilber, J.; Moniatte, M.; Nielsen, M. I. L.; Podtelejnikov, A. V.; Mann, M. Int. J. Mass Spectrom. 2003, 226, 223-37. (8) Montaud, G., Lattimer, R. P., Eds. Mass Spectrometry of Polymers; CRC Press: Boca Raton, FL, 2001. (9) Hanton, S. D. Chem. Rev. 2001, 101, 527-69. (10) Macha, S. F.; Limbach, P. A. Curr. Opin. Solid State Mater. Sci. 2002, 6, 213-20. (11) Murgasova, R. M.; Hercules, D. M. Int. J. Mass Spectrom. 2003, 226, 15162. (12) Gluckmann, M.; Karas, M. J. Mass Spectrom. 1999, 34, 467-77. (13) Vorm, O.; Roepstorff, P.; Mann, M. Anal. Chem. 1994, 66, 3281-7. (14) Dai, Y.; Whittal, R. M.; Li, L. Anal. Chem. 1996, 68, 2494-500. (15) Wei, J.; Buriak, J. M.; Siuzdak, G. Nature 1999, 399, 243-6. (16) Shen, Z.; Thomas, J. J.; Averbuj, C.; Broo, K. M.; Engelhard, M.; Crowell, J. E.; Finn, M. G.; Siuzdak, G. Anal. Chem. 2001, 73, 612-9. (17) Alimpiev, S.; Nikiforov, S.; Karavanskii, V.; Minton, T.; Sunner, J. J. Chem. Phys. 2001, 115, 1891-901. (18) Kruse, R. A.; Li, X.; Bohn, P. W.; Sweedler, J. V. Anal. Chem. 2001, 73, 3639-45. (19) Lewis, W. G.; Shen, Z.; Finn, M. G.; Siuzdak, G. Int. J. Mass Spectrom. 2003, 226, 107-16. (20) Trauger, S.; Go E. P.; Shen, Z.; Apon, J. V.; Compton, B.; Bouvier, E. S. P.; Finn, M. G.; Siuzdak, G. Anal. Chem. 2004, 76, 4484-9. (21) Arakawa, R.; Shimomae, Y.; Morikawa, H.; Ohara, K.; Okuno, S. J. Mass Spectrom. 2004, 18, 961-5. (22) Okuno, S.; Wada, Y.; Arakawa, R. Int. J. Mass Spectrom. 2005, 241, 43-8. 10.1021/ac050504l CCC: $30.25
© 2005 American Chemical Society Published on Web 06/28/2005
of porous silicon can act as a matrix for desorption/ionization, meaning that DIOS would be a form of so-called surface-assisted laser desorption/ionization (SALDI) that utilizes inorganic matrixes such as metals or graphitized carbon.4,23 Based on this broader view, various particles or substrates with porous and nonporous surface morphologies have been evaluated to date regarding properties facilitating laser-induced desorption/ionization. For example, Zhang et al. used porous silicon powder and porous silica gel as inorganic matrixes for MALDI and found the photoluminescence of porous silicon to contribute little to desorption/ionization performance.24 This was consistent with results obtained using porous substrates with varying luminescence properties.18 Moreover, in a study utilizing etched silicon and graphite surfaces, Alimpiev et al. demonstrated submicrometer surface porosity, irrespective of the substrate material, to be a key factor promoting desorption/ionization of peptides.17 They suggested surface porosity, roughness, or both to enhance ion signals via a variety of features: resupplying the surface with analyte after a laser pulse, providing a local environment for the laser-induced field that facilitates physical separation of preformed analyte ions from their counterions, and decreasing heat capacity and heat conductivity for effective sublimation of analytes by laser irradiation as well as absorbing laser light.17 As to pore size, silica gels with 10- and 30-nm-diameter pores, or porous silicon with a pore diameter exceeding 10 nm, yielded intense signals as compared to those with smaller pores.18,24 More recently, direct laser desorption/ionization on carbon nanotubes or silicon nanowires has been reported.25,26 In the present study, we investigated various structures and modifications, focusing on surface morphology, to delineate the requirements for a sample target for laser-induced desorption/ionization of synthetic polymers and biomolecules. EXPERIMENTAL SECTION Materials. Tetrahydrofuran (THF), methanol, ethanol, acetonitrile, 2-propanol, 46% hydrofluoric acid, sodium iodide, trifluoroacetic acid (TFA), insulin (human), reserpine, poly(propylene glycol) (PPG, average Mr ∼700), and 2,5-dihydroxybenzoic acid (DHB) were purchased from Wako Pure Chemicals (Osaka, Japan), and R-cyano-4-hydroxycinnamic acid (CHCA) and poly(1,4-butylene adipate) (PBA, average Mr ∼1000) were from SigmaAldrich (Milwaukee, WI). Angiotensin I was obtained from Bachem AG (Bubendorf, Switzerland). Octylphenol polyethoxylate (Triton X-100, average Mr ∼600) was from ICN Biomedicals (Aurora, OH). DIOS chips were purchased from Waters (Milford, MA). Silicon (100) wafers of n-type and F ) 0.008-0.02 Ω‚cm resistivity were from Sumitomo Mitsubishi Silicon Corp. (Tokyo, Japan). Metal-coated porous alumina (Alumi Surface Technologies, Ibaraki, Japan) was prepared by anodizing an aluminum sheet followed by spattering of Au or Pt at a surface thickness of 50 (23) Zumbuhl, S.; Knochenmuss, R.; Wulfert, S.; Dubois, F.; Dale, M. J.; Zenobi, R. Anal. Chem. 1998, 70, 707-15. (24) Zhang, Q.; Zou, H.; Guo, Z.; Zhang, Q.; Chen, X.; Ni, J. Rapid Commun. Mass Spectrom. 2001, 15, 217-23. (25) Go, E. P.; Apon, J. V.; Luo, G.; Saghatelian, A.; Daniels, R. H.; Sahi, V.; Dubrow, R.; Cravatt, B. F.; Vertes, A.; Siuzdak, G. Anal. Chem. 2005, 77, 1641-6. (26) Ren, S. F.; Zhang, L.; Cheng, Z. H.; Guo, Y. L. J. Am. Soc. Mass Spectrom. 2005, 16, 333-9.
nm. Glass fiber-containing porous polyethylene (Nippon Sheet Glass Corp,. Tokyo, Japan) was surface-coated with Pt/Pd in 20nm thickness by spattering. Fabrication of Groove Arrays on a Silicon Surface. Arrays of submicrometer grooves were fabricated on a silicon wafer as follows. ZEP520 (Nihon Zeon, Tokyo, Japan) resists were coated uniformly using standard spin-coating techniques. The resists on silicon substrates were exposed using a 75-kV electron beam with the exposure system ELS-7700 (Elionix, Tokyo, Japan). Exposed samples were developed in ZED-N50 (Nihon Zeon). The resist patterns were transferred to Ni by a conventional liftoff method. The silicon substrates, with the Ni mold (mask), were etched using a reactive ion etching machine, RIE-10NR (Samco, Kyoto, Japan), with CF4-Ar gas. Then, the Ni mask was removed by soaking in diluted HNO3. Just before the mass spectrometric measurements, the fabricated silicon surfaces were refreshed by soaking in 5% HF solution for 2 min, washed in water, methanol, 2-propanol, and, finally, ethanol, and then dried in vacuo. Scanning Electron Microscopy (SEM). The SEM images of porous alumina were provided by Alumi Surface Technologies. Other SEM images were acquired using JSM-5310 (JEOL, Tokyo, Japan) or JSM-6700F (JEOL). Sample Preparation. Reserpine and peptide samples were dissolved in a solution of 0.1% TFA and 50% acetonitrile at 50 µM, and polymers were dissolved at 0.3 mg/mL in THF with 1 mg/ mL NaI. An aliquot of each sample solution, 0.5 µL for scratched silicon and stainless surfaces, 0.1 µL for a grooved silicon surface, and 0.2 µL for metal-coated porous alumina and porous polyethylene, was deposited on the sample target and dried at room temperature. Mass Spectrometry. Mass spectra were acquired using a Voyager-DE Pro time-of-flight mass spectrometer (Applied Biosystems, Foster City, CA) with a pulsed nitrogen laser (337 nm) in positive ion mode. The laser shot angle to the sample target surface was set at 45°, and the laser repetition rate was 3 Hz. A chip of the fabricated sample target was taped onto a stainless steel stage, prepared in-house, to offset the thickness of the chip. In all cases, mass spectra from 100 laser shots were accumulated. The levels of the minimum laser fluence needed to generate the molecular ion signals with a signal-to-noise (S/N) ratio of more than 50 in the vicinity were recorded in arbitrary units. The S/N value of the ion was obtained by a signal-to-noise ratio calculator equipped in the data system of the instrument. The mass spectra were all obtained on the same day and under the same experimental conditions from more than three different sample spots. RESULTS AND DISCUSSION Irregular Roughness of Silicon and Stainless Steel Surfaces. The silicon surface was scratched with No. 400 sandpaper, and the resulting surface roughness was examined for its capability in UV laser-induced desorption/ionization of deposited analytes. In an earlier study, this type of processing was undertaken as a control setup for infrared laser desorption/ionization, but protic solvents such as methanol, ethanol, or water worked as a sample matrix, such that surface roughness or porosity of the sample target had no effect on ionization.27 As illustrated by the SEM (27) Bhattacharya, S. H.; Raiford, T. J.; Murray, K. K. Anal. Chem. 2002, 74, 2228-31.
Analytical Chemistry, Vol. 77, No. 16, August 15, 2005
5365
Figure 1. SEM images of manufactured surfaces: (a) silicon wafer scratched with No. 400 sand paper (surface), (b) stainless steel plate scratched with No. 400 sandpaper (surface), (c) submicrometer groove arrays on silicon (cross section), and (d) Au-coated porous alumina (cross section). Table 1. Minimum Laser Power for Detecting Molecular Ions in Reflectron Mode by Direct UV Laser Desorption/Ionization (Arbitrary Units) Triton-X stainless steel unscratched scratched with sandpaper silicon unscratched scratched with sandpaper groove arrays DIOS porous alumina noncoated Au coated MALDI DHB matrix CHCA matrix a
PPG
PBA
reserpine angiotensin I
1400 900
1350 1650 850 1100
1050
1700 850
1450 1900 750 850
750
a
850 1150
800 900 1150 1250
800 850
1150 1000
1450 950
1500 1450 950 1000
1650 850
950
1050 700
1000 1150 700 800
650 1100
1000 600
Detectable in linear mode.
image in Figure 1a, the scratched silicon displayed submicrometer surface irregularities. Hydrophilic polymers, Triton X-100 and PPG, and a hydrophobic polymer, PBA, were directly deposited on this target, and the laser powers required for generating ions of these compounds in reflectron mode were recorded (Table 1). A mass spectrum composed of Na-adducted ions was obtained from the scratched surface with the laser fluence at a level required for MALDI, while these ions were generated from the smooth surface only with higher laser power and were accompanied by a number of unknown signals (Figure 2 and Table 1). The scratched surface enabled us to observe the protonated molecules of reserpine and angiotensin in the mass spectra of reflectron and linear modes, respectively (Figure 3 and Table 1). However, the angiotensin ion was not found in the reflectron mode mass spectrum. Subsequently, the same processing was applied to a stainless steel surface (Figure 1b). Previously, a similar roughness was created on a MALDI sample target for the purpose of making fine matrix/analyte crystals.28 In the present study, however, we 5366 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005
Figure 2. Direct UV laser-induced desorption/ionization mass spectra acquired on scratched or smooth silicon surfaces: (a) Triton X-100 (scratched), (b) Triton X-100 (smooth), (c) PBA (scratched), and (d) PBA (smooth). Each 0.5-µL aliquot of 0.3 mg/mL sample solution was deposited on the surface. Reflectron mode.
found direct or matrix-free desorption/ionization of synthetic polymers and reserpine to be enhanced on the sandpaperscratched stainless steel surface, and the levels of laser fluence were comparable to those required for MALDI (Table 1). However, more laser power was required for ionization compared with the scratched silicon surface, and the molecular ion of angiotensin was not observed even in linear mode measurements. The difference in the performance between these materials is due to the blunt irregularities of stainless surface or to their physicochemical property, and the former morphologic effect is probably dominant considering the following results on the metal-coated porous structures. Our results indicate that the roughness produced by submicrometer irregularities supports direct UV laser-induced desorption/ionization, irrespective of the materials onto which a rough structure is introduced. This observation is in part consistent with the results of Alimpiev et al.,17 who investigated the conditions for SALDI using etched graphite as well as porous silicon, although carbon and silicon belong to the same group in the periodic table. The scratching gives expansion of the surface area and probably renders refraction of laser light, enhancing desorption/ionization of these analytes. Although the scratched surface on a sample target was practicable for MS of synthetic polymers, ionization of peptides was obviously insufficient and required the ordered or more dense morphologies as follows. Lined and Oriented Structure on Silicon. The regular submicrometer structures were created lithographically in arrays of 150-nm-width grooves, at 150-nm intervals, and at depths varying from 100 to 300 nm (Figure 1c). The laser shot angle to the sample target was fixed at 45°. (28) Giannakopulos, A. E.; Bashir, S.; Derrick, P. J. Eur. Mass Spectrom. 1998, 4, 127-31.
Figure 3. Direct UV laser-induced desorption/ionization mass spectrum of angiotensin I deposited on a silicon surface scratched with No. 400 sandpaper. An aliquot of 0.5 µL of 50 µM sample solution was deposited on the surface. Linear mode.
Figure 4. Direct UV laser-induced desorption/ionization mass spectrum of angiotensin I deposited on submicrometer grooves on a silicon surface. The groove depth was 300 nm. An aliquot of 0.1 µL of 50 µM sample solution was deposited on the surface. Reflectron mode.
Interestingly, the protonated angiotensin could be generated by direct UV laser irradiation (Figure 4). Moreover, the directionality of the structures enabled us to investigate the geometric features of the ionization properties. First, efficiency was dependent on the vertical dimension of the groove; a depth of 200 nm generated more intense signals than the 100-nm depth at a fixed laser fluence, and the efficiencies at depths of 200 and 300 nm were similar. Next, the effect of the relative direction, at right angles versus parallel, between grooves and laser light was examined using a 300-nm-depth target (Chart 1). The right angle was effective in ionization compared to parallel; i.e., an 1150 versus a 1450 minimum laser fluence, and the mass spectrum obtained by the parallel geometry was noisy. The difference cannot be explained by the total surface area, but may be associated with the reflection of laser light at the frontal wall. The surfaces of DIOS chips have different scales, micro (50 nm), of porous structures.29-31 As shown in Figure 1c, the grooves in the present (29) Cullis, A. G.; Canham, L. T.; Calcott, P. D. J. Appl. Phys. Rev. 1997, 82, 909-66. (30) Bisi, O.; Ossicini, S.; Pavesi, L. Surf. Sci. Rep. 2000, 38, 1-12. (31) Foll, H.; Christophersen, M.; Hasse, G. Mater. Sci. Eng. R 2002, 39, 93141.
Chart 1. Directions of Laser Irradiation Relative to Grooves on the Silicon Surface
study apparently have few mesosized and probably few microsized structures, suggesting the macrosized structure to be essential for desorption/ionization on nanofabricated structures. We carried out these experiments only with a 337-nm laser. Any possible correlation between the laser wavelength or polarity and the surface dimensions thus remains to be determined. Porous Alumina with Metal-Coated Surface. In an earlier study, Alimpiev et al. were able to ionize organic compounds including peptides on an etched carbon surface with UV laser Analytical Chemistry, Vol. 77, No. 16, August 15, 2005
5367
Figure 5. Direct UV laser-induced desorption/ionization mass spectra of angiotensin I and insulin deposited on Au-coated porous alumina: (a) angiotensin I in reflectron mode; (b) human insulin in linear mode. Each 0.2 µL aliquot of 50 µM sample solution was deposited on the surface.
irradiation.17 In MALDI, powdered porous silicon or porous silica gel is an effective inorganic matrix for ionization of small molecules, indicating the pore structure to apparently be essential in transferring the laser energy to analytes for ionization.24 Taking these and our prior results into account, we employed porous alumina created by anodization. Unfortunately, strong laser fluence was required for the ionization of synthetic polymers, and intact molecular ions of angiotensin could not be obtained (Table 1). DIOS chips are often modified with organic groups to prevent surface oxidation. Such modification raises the ionization threshold, but the deposition of platinum on this modified DIOS surface can regenerate the performance (our unpublished results). These observations suggested electric conductivity of the surface to be required for non-silicon porous substances to be effective. After the porous alumina surface was coated with gold, angiotensin and insulin could be ionized on this surface (Figure 5). The pore diameter and depth of this substance were 200-300 and 500700 nm, respectively (Figure 1d). Platinum was similarly effective as a coating metal. Subsequently, the Pt/Pd-coated porous polyethylene with pores of ∼200 nm in diameter was also successful in generating protonated molecules of angiotensin and insulin, although the mass spectra were noisy compared with porous alumina (data not shown). These results indicate the surface electric conductivity as well as the submicrometer structure to be necessary for direct laser-induced desorption/ 5368
Analytical Chemistry, Vol. 77, No. 16, August 15, 2005
ionization. It is also noteworthy that coating the surface with these noble metals conferred the additional advantage of prolonging shelf life for a porous alumina chip. Ionization performance was unchanged after more than 30 days of exposure to room air. Our results for the porous alumina surface coated with Au or Pt may provide insights into, or even arguments against, the results of previous studies. It is well known that porous silicon emits photoluminescence under UV light irradiation, and a number of mechanisms including dangling bonds on the silicon surface, surface molecules such as Siloxene, nanometer-sized structural defects, and quantum-size effects have been suggested.29-31 However, the porous alumina with Au coating is not photoactive. This was quite consistent with the DIOS observations in previous reports, in which the conditions for making a DIOS chip are milder than those for preparing photoluminescent porous silicon materials, and no correlation was observed between the DIOS-MS signal and luminescence.16,18 DIOS is largely based on the physical properties of porous silicon being different from those of luminescence. Kruse et al. reported that porotic structures of GaAs or GaN were ineffective as compared with porous silicon and ascribed the difference to a combination of three factors, surface area, the absorption cross section at the irradiation wavelength, and thermal conductivity, that can effectively couple the deposited energy to the pore bound analytes.18 Although we still cannot exclude the
possibility that a different mechanism is involved in the desorption/ionization on a porous semiconductor and that the current Au- or Pt-coating may play some unknown roles in desorption/ ionization, electric conductivity is apparently a key factor in porous or submicrometer structures. CONCLUSIONS The laser fluence required for desoprtion/ionization of synthetic polymers and reserpine on stainless steel as well as silicon was greatly reduced, when the surface was scratched with sandpaper. Lined submicrometer grooves that had been created on the silicon surface allowed desorption/ionization of angiotensin with 337-nm N2 laser irradiation. This morphology demonstrated for the first time that performance depends on the angle between laser irradiation and the groove, also suggesting submicrometer but not smaller structures to be a key factor in the laser-induced desorption/ionization on rough surfaces. The porous structures of alumina or polyethylene promoted ionization of angiotensin and insulin molecules, when the surface was coated with metals. The coating may be required to render the substance electrically
conductive and provides the additional advantage that the porous alumina chip had prolonged activity, exceeding a month even when the chip was left in the open air. These results indicate that laser-induced desorption/ionization of organic compounds can be implemented on submicrometer structures with a metal-coated surface irrespective of the basal materials. ACKNOWLEDGMENT This work was supported in part by a Grant-in-Aid for the Nanotechnology Support Project from the Ministry of Education, Science, Sports, and Culture of Japan. The authors thank Prof. Ochi of Kansai University for portions of our SEM measurements. Helpful suggestions regarding the porous alumina were given by Dr. Ikemoto of Mitsubishi Gas Chemical.
Received for review March 24, 2005. Accepted June 3, 2005. AC050504L
Analytical Chemistry, Vol. 77, No. 16, August 15, 2005
5369
