Anal. Chem. 2003, 75, 5602-5607
Laser Desorption Combined with Hyperthermal Surface Ionization Time-of-Flight Mass Spectrometry Christian Weickhardt*
Department of Computer Science, Mathematics and Natural Sciences, Leipzig University of Applied Sciences, Trufanow Strasse 6, 04105 Leipzig, Germany Lars Draack
Department of Physical and Analytical Chemistry, Brandenburg University of Technology, Erich-Weinert-Strasse 1, 03044 Cottbus, Germany Aviv Amirav
School of Chemistry, Sackler Faculty of Exact Sciences, Tel Aviv University, Tel Aviv 69978, Israel
A setup combining laser desorption of nonvolatile molecules and their aerodynamic acceleration in a supersonic molecular beam followed by hyperthermal surface ionization in a reflectron time-of-flight mass spectrometer is described. While laser desorption performs the intact transfer of the analyte molecules into the gas phase, hyperthermal surface ionization opens up the possibility to efficiently ionize even larger molecules with a small and potentially controlled degree of fragmentation. Being an ionization technique, which is particularly effective for aromatic and heterocyclic compounds, the selectivity can further be increased by tuning the kinetic energy to which the molecules are accelerated in the supersonic beam. The results obtained for several polycyclic aromatic hydrocarbons and biochemical substances show that sufficient acceleration can be achieved even for molecules with a molecular weight above 5000 amu and that HSI preserves its advantageous features even for thermally labile large molecules such as insulin.
to be powerful tools in this context. Although completely different in respect to their underlying processes, a common aspect that these two techniques share is the fact that the steps of desorption and ion formation are unified and carried out simultaneously in one step. The unification of desorption and ionization processes sets rather strict limits to the ability to tune and optimize the features of each of them individually. This can be of particular concern when it is desirable to achieve a certain degree of selectivity or fragmentation in the ionization step. Therefore, the separation of desorption and ionization steps in space and time though increasing the complexity of the setup results in mass spectrometric techniques that offer a higher degree of flexibility. Currently, this approach is followed, for example, by coupling laser desorption at atmospheric pressure to chemical ionization.6 In fact, laser desorption time-of-flight mass spectrometry also started its way with time-separated processes of laser desorption and subsequent laser ionization of the desorbed species.7-16 As demonstrated in numerous applications, this combination offers an outstanding degree of selectivity.8-17
With the growing importance of the life sciences, the need for analytical methods and instrumentation in the biochemical, pharmaceutical, and medical fields has experienced a dramatic increase. As far as mass spectrometry is concerned, its successful application to the analysis of biologically or medically relevant substances required the development of special ionization techniques that allowed the intact transfer of large thermally labile molecules into the gas phase and their ionization with a small degree of fragmentation. In particular, matrix-assisted laser desorption/ionization (MALDI)1,2 and electrospray3-5 turned out
(5) Gaskell, S. J. J. Mass Spectrom. 1997, 32, 677-688. (6) Coon, J. J.; McHale, K. J.; Harrison, W. W. Rapid Commun. Mass Spectrom. 2002, 16, 681-685. (7) Weyssenhoff, H. v.; Selzle, H. L.; Schlag, E. W. Z. Naturforsch. 1985, 40A, 674-676. (8) Tabet, J. C.; Cotter, R. J. Anal. Chem. 1984, 56, 1662-1667. (9) Grotemeyer, J.; Boesl, U.; Walter, K.; Schlag, E. W. Org. Mass. Spectrom. 1986, 21, 645-653. (10) Boesl, U.; Grotemeyer, J.; Walter, K.; Schlag, E. W. Anal. Instrum. 1987, 16, 151-171. (11) Engelke, F.; Hahn, J. H.; Henke, W.; Zare, R. N. Anal. Chem. 1987, 59, 909-912. (12) Lubman, D. M. Mass Spectrom. Rev. 1988, 7, 535-554. (13) Arrowsmith, P.; deVries, M. S.; Hunziker, H. E.; Wendt, H. R. Appl. Phys. 1988, B46, 165-173. (14) Zenobi, R.; Philippoz, J.-M.; Buseck, P. R.; Zare, R. N. Science 1989, 246, 1026-1029. (15) Meijer, G.; deVries, M. S.; Hunziker, H. E.; Wendt, H. R. Appl. Phys. 1990, B51, 395-403. (16) Kinsel, G. R.; Linder, J.; Grotemeyer, J. J. Phys. Chem. 1992, 96, 31573162 and 3162-3166.
* Corresponding author. e-mail:
[email protected]. (1) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-2301. (2) Hillenkamp, F.; Karas, M. Int. J. Mass Spectrom. 2000, 200, 71-77. (3) Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4451-4459. (4) Whitehouse, C. M.; Dreyer, R. N.; Yamashita, M.; Fenn, J. B. Anal. Chem. 1985, 57, 675-679.
5602 Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
10.1021/ac0302197 CCC: $25.00
© 2003 American Chemical Society Published on Web 09/12/2003
However, while the employment of a second (maybe even tunable) laser for ionization increases the complexity and the costs of the complete system significantly, low-cost standard ionization techniques such as electron ionization are of limited use when larger and particularly labile molecules are to be investigated due to their limited ionization efficiency and high degree of fragmentation. Such significant fragmentation with frequent absence of the molecular ion is especially extensive with large molecules that are not cooled in supersonic molecular beams, and it makes the interpretation of the obtained mass spectra difficult or impossible. A promising alternative is the use of vacuum ultraviolet (VUV) photons for ionization. When it is carried out on large molecules at or above room temperature it leads to mass spectra similar to those with EI. However, fragmentation can be almost completely suppressed when the analyte molecules are jet cooled before VUV ionization.18,19 Actually, the degree of fragmentation can be controlled over a wide range by the internal temperature. Furthermore, the technique offers some selectivity by choice of a VUV photon energy that is sufficient to ionize the substances of interest while leaving the others with higher ionization energy unaffected. However, the generation of pulsed, intense VUV radiation also depends on laser as well as nonlinear optical techniques or on special lamps. An ionization process that exhibits features similar to VUV ionization but does not require such complex equipment is hyperthermal surface ionization (HSI).20-22 HSI has proven to be an ionization technique offering a number of advantageous features for mass spectrometry. During the last years it was extensively studied from its mechanisms23 through its various analytical applications.22,24,25 HSI is based on the electron transfer between a molecule and a surface, facilitated by the image potential that exists when an ion is scattered from the surface. In HSI, the kinetic energy of a molecule accelerated by a supersonic expansion is used to effectively bridge the gap between the surface work function and the molecular ionization potential. Since the electron is transferred from a finite molecule into an infinite bulk via its surface, the HSI process, in contrast to photoionization or electron ionization, is anticipated to be irreversible and without molecular size increased reneutralization.22 Among the advantageous features HSI offers for analytical mass spectrometry are its high ionization yield (up to 10%) particularly for aromatic and heterocyclic compounds, variable and adjustable fragmentation pattern, and a certain degree of selectivity based on the different ionization energies of the compounds present in the sample. For perdeuterated anthracene and pyrene, a minimum detected amount of 400-500 ag was demonstrated using fast GC/MS.26,27 (17) Zenobi, R. Chimia 2001, 55, 773-777. (18) deVries, M. S.; Hunziker, H. E. J. Photochem. Photobiol., A 1997, 106, 3136. (19) Nir, E.; Hunziker, H. E.; deVries, M. S. Anal. Chem. 1999, 71, 1674-1678. (20) Danon, A.; Amirav, A. J. Chem. Phys. 1987, 86, 4708-4709. (21) Danon, A.; Kolodney, E.; Amirav, A. Surf. Sci. 1988, 193, 132-152. (22) Danon, A.; Amirav, A. J. Phys. Chem. 1989, 93, 5546-5562. (23) Amirav, A. Org. Mass Spectrom. 1991, 26, 1-17. (24) Danon, A.; Amirav, A. Int. J. Mass Spectrom. Ion Processes 1990, 96, 139167. (25) Danon, A.; Amirav, A. Int. J. Mass Spectrom. Ion Processes 1993, 125, 6374. (26) Dagan, S.; Amirav, A. Int. J. Mass Spectrom. Ion Processes 1994, 133, 187210.
The selectivity as well as the degree of molecular fragmentation observed in HSI can be tuned by varying the kinetic energy of the analyte molecules by means of the carrier gas type, pressure, and nozzle temperature. In HSI, the molecular kinetic energy can control the degree of fragmentation from practically none to almost complete and extensive dissociation as, for example, demonstrated with cholesterol.28 Actually, HSI combines both a charge transfer and a collision-induced dissociation step in a single scattering event. Originally HSI was developed as a continuous ionization method based on continuous supersonic molecular beams (SMBs). Consequently, it was coupled to scanning mass spectrometers, especially quadrupole instruments. Recently29 we reported on the coupling of HSI and our time-of-flight (TOF) mass spectrometer. A different combination of HSI and TOF-MS was reported by Davis and co-workers.30 This combination enables the coupling of the advantages of time-of-flight mass spectrometry such as simultaneous detection of practically an unlimited mass range with high transmission and good mass resolution to those of HSI. As timeof-flight mass spectrometers work in a noncontinuous mode, they are preferably combined with pulsed ionization techniques. Therefore, we carried out HSI in a pulsed molecular beam and developed an optimized ion source geometry for this purpose.29 To analyze nonvolatile substances with this technique, a desorption step has to be added, in which the analyte molecules are introduced into and coaccelerated with the supersonic beam. While continuous desorption can be achieved by heating, a pulsed technique such as laser desorption increases the duty cycle and therefore the sensitivity when combined with a pulsed SMB. Furthermore, being a very soft desorption technique, laser desorption allows the intact vaporization of even thermally labile compounds, a fact that cannot be overestimated for the analysis of medically and biologically relevant substances. Separation of the desorption and the ionization steps increases the possibility to tune and optimize each of them separately according to the sample to be analyzed and the information to be obtained. In this work, we report on the addition of a laser desorption device to our HSI-TOF-instrument and the results achieved with it for a few polycyclic aromatic hydrocarbons and biomolecules. EXPERIMENTAL SETUP The HSI-TOF-mass spectrometer was described in detail previously.29 A schematic drawing of it is shown in Figure 1. The carrier gas forming the supersonic molecular beam is expanded through a pulsed nozzle with an inner diameter of 200 µm (General Valve Corp., Fairland, NJ) that can be heated to temperatures up to 170 °C. The opening time of the nozzle was adjusted to 100 µs. Either hydrogen, helium, neon, or argon at stagnation pressures up to 6 bar was used as carrier gas. For the introduction of nonvolatile analyte molecules into the supersonic expansion, these substances were laser desorbed on the low-pressure side of the pulsed nozzle by means of the fourth (27) Amirav, A.; Gordin, A.; Tzanani, N. Rapid Commun. Mass. Spectrom. 2001, 15, 811-820. (28) Dagan, S.; Amirav, A. J. Am. Soc. Mass Spectrom. 1993, 4, 869-873. (29) Weickhardt, C.; Draack, L.; Grotemeyer, J. Eur. J. Mass Spectrom. 2000, 6, 319-323. (30) Davis, S. C.; Makarov, A. A.; Hughes, J. D. Rapid Commun. Mass. Spectrom. 1999, 13, 247-250.
Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
5603
Figure 1. Instrumental setup for laser desorption hyperthermal surface ionization in a reflectron time-of-flight mass spectrometer. Figure 3. Temporal course of events during an LD-HSI-TOF measurement. Table 1
Figure 2. Pulsed nozzle with attached laser desorption device for jet entrainment of nonvolatile compounds.
harmonic of a pulsed Nd:YAG laser (wavelength 266 nm, pulse energy 0.3-0.7 mJ, pulse width ∼5 ns) focused down to a spot of 1-mm diameter. As can be seen from the schematic drawing in Figure 2, the axis of the round sample with a diameter of 2 mm was located 2 mm downstream from the nozzle and its surface was placed at a distance of 1 mm from the SMB axis. The sample substances were pressed to thin pallets and glued onto the tip of a manipulator stick by means of which they could be placed in the desorption position. The analyte was desorbed into a mixing channel with an inner diameter of 2 mm and a length of 10 mm where it was accelerated by collisions with the carrier gas and entered the vacuum chamber through a conical end. Further downstream, the central part of the SMB passes through a skimmer with an aperture of 2 mm. It then collides under 45° with the surface where HSI is aimed to occur, a 1-cm2 rhenium foil with a thickness of 25 µm prepared and oxidized by a procedure developed by Dagan and Amirav.26 The foil is mounted in a stainless steel holder where it is resistively heated and flushed with oxygen throughout the measurement. The oxygen partial pressure in the ion source chamber was established at 10-4 mbar. By collisions with the HSI surface, analyte molecules are ionized, deflected by ∼90° in respect to the SMB, and pulse fed into the acceleration region of the time-of-flight mass spectrometer. The latter is achieved by means of a pulsed positive voltage with an amplitude of 5-15 V applied to the HSI surface for ∼20 µs. As already discussed in previous work, the 45° angle between the SMB and the HSI surface is not optimal with respect to the ionization efficiency, but it reduces problems related to initial velocity components of the ions perpendicular to the acceleration field. The acceleration voltages are applied as soon as the first acceleration region is filled with ions. In this way, it is possible to establish a reasonable trade-off between the number of ions 5604 Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
substance
empirical formula
molecular mass (amu)
CAS number
fluorene anthracene phenanthrene fluoranthene pyrene chrysene decacyclene melanostatin β-carotene insulin
C13H10 C14H10 C14H10 C16H10 C16H10 C18H12 C36H18 C13H24N4O3 C40H56 C254H377N65O75S6
166.22 178.23 178.23 202.26 202.26 228.29 450.54 284.4 536.9 5733
86-73-7 120-12-7 85-01-8 206-44-0 129-00-0 218-01-9 191-48-0 2002-44-0 7235-40-7 11070-73-8
transferred into the mass spectrometer and the kinetic energy spread caused by the spatial distribution of the ion cloud inside the acceleration field, which affects the mass resolution in a negative way. The acceleration electrodes form a conventional WileyMcLaren-type ion source31 whose voltages are applied in a pulsed mode by means of two high-voltage transistor switches. Typical ion source voltages were 1150 V applied to the repeller electrode and 800 V to the intermediate electrode. Figure 3 shows the succession of events occurring when this setup is used. The reflectron time-of-flight mass spectrometer is a home-built instrument with a total flight path of ∼2.5 m, differentially pumped by two turbomolecular pumps and equipped with a two-stage reflector for second-order energy focusing. The signals from the single-stage microsphere plate detector are recorded by a digital oscilloscope from where data are transferred to a PC computer system for storage and processing. The analyte substances studied by laser desorption HSI-TOF mass spectrometry in this work (the polycyclic aromatic hydrocarbons antracene, phenanthrene, pyrene, fluorene, fluoranthene, chrysene, and decacyclene, the tripeptide melanostatin, the provitamine β-carotine, and the hormone bovine insulin) were purchased from Sigma-Aldrich and used without further purification. All substances studied in this work are listed in Table 1. RESULTS AND DISCUSSION For seven polycyclic aromatic hydrocarbons within the mass range from 166 to 450 amu, the LD-HSI efficiency was studied (31) Wiley, W. C.; McLaren, I. H. Rev. Sci. Instrum. 1955, 26, 1150-1157.
Figure 4. Time-of-flight mass spectra of pyrene, fluoranthene, chrysene, and decacyclene obtained after their laser desorption into either a helium (left-hand side) or neon (right-hand side) expanding supersonic free jet and subsequent hyperthermal surface ionization.
using either helium, neon, or argon as carrier gas for the supersonic beam. For each substance, the desorption laser intensity had to be optimized individually. It turned out that laser desorption is efficient only within a narrow intensity range while the velocities of the desorbed molecules are small enough to allow their coupling into the supersonic jet. Furthermore, high-energy laser pulses may enrich the expanding gas too much with desorbed species, thus leading to their reduced acceleration by the carrier gas. In Figure 4, the LD-HSI-TOF spectra of pyrene, fluoranthene, chrysene, and decacyclene obtained by using either helium or neon at a pulsed nozzle stagnation pressure of 5.5 bar are compared. It can be seen that for the lighter molecules helium with its higher jet velocity gives rise to a higher ionization efficiency. Only for the case of decacyclene with a molecular mass around 450 amu the heavier is neon superior. This observation can be understood in terms of the velocity slip effect.32 After (32) Kolodney, E.; Amirav, A. Chem. Phys. 1983, 82, 269-283.
leaving the nozzle, the particle density of the carrier gas rapidly decreases due to the divergence of the beam. Therefore, considerable collision rates with embedded analyte molecules are only achieved within a distance of a few nozzle diameters after which the acceleration is completed. As a result, the velocity of the analyte molecules is always smaller than that of the carrier gas and depends on the number of collisions they underwent with the carrier gas and on the momentum transferred in each collision. While the first factor depends on the nozzle geometry and the carrier gas stagnation pressure prior to expansion, the latter is completely governed by the laws of elastic two-body collisions, which state a maximum for the momentum transfer as the masses for the collision partners become equal. On the other hand, for a given stagnation pressure, lighter gases result in higher beam velocities. From these facts it becomes evident that for a certain molecular mass an optimum carrier gas has to be found that offers the most efficient acceleration. For our setup, it could be clearly observed that heavier molecules require carrier gases with higher Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
5605
molecular weight, demonstrating that the effect of larger momentum transfer outdoes the disadvantage of a smaller gas velocity. None of the PAHs investigated exhibited fragmentation under the conditions applied, and only their molecular ion was observed. The inset in the spectrum of fluoranthene shows the molecular ion region with the 13C isotope. The mass resolution calculated according to the 50% valley definition is only ∼500. Several factors contribute to this rather poor value compared to resolutions of above 20 000 obtained with other reflectrons. The one considered most important in our experiment is the broad distribution of initial ion velocities due to the surface collision process. This source of flight time errors cannot be compensated by the reflector. Also techniques such as “delayed extraction”33 are of little use in this case because the ions are not formed within a small volume or at a surface inside the acceleration field but rather the complete acceleration region is filled with ions showing the full distribution of initial velocities at every point. Furthermore, as a large portion of the first acceleration region is filled with ions prior to switching on the voltages, the distribution of ion kinetic energies in the field free drift region is also very broad (typically 30% of the average ion kinetic energy). As the reflector can only compensate for a 10% kinetic energy difference, this results in an additional broadening of the time-of-flight signal. Of course, these observations raise the question whether a time-of-flight instrument is the right mass spectrometer to couple with the laser desorption-HSI technique. The starting conditions in which the ions are created are completely different from cases where TOF-MS proved to be powerful, e.g., in combination with MALDI, where they are formed within a thin layer and within a short period of time. It is our main interest to further develop LD-HSI-MS toward an analytical tool for toxicological, biochemical, and pharmaceutical applications. As a first step toward this goal, we investigated the LD-HSI behavior of three biomolecules: the tripeptide melanostatin (M ) 284 amu), the provitamin β-carotene (M ) 537 amu), and the hormone bovine insulin (M ) 5733 amu). Their LD-HSI-TOF mass spectra are shown in Figure 5 and the insets show their mass spectra zoomed around the molecular ion TOF. Due to the limited mass resolution for β-carotene, the isotopic distribution is not resolved but the width of the envelope matches the theoretical one. While melanostatin gave the best results when helium is used as carrier gas, β-carotene was more efficiently ionized with neon and insulin with argon as could already be expected based on the need to better accelerate the lighter compounds. In all cases, the molecular ion was the only significant signal observed. Due to vacuum restrictions, we were unable to further increase the stagnation pressure of the carrier gas in order to achieve higher kinetic energies of the analyte molecules, which could induce fragmentation. For bovine insulin, the heaviest molecule investigated in this work, argon under a stagnation pressure of 6 bar had to be used for acceleration in order to observe ionization. Also, here the molecular ion signal is the only significant peak in the spectrum. The low signal intensity in this mass spectrum reflects the reduced HSI efficiency, probably due to a combination of lower acceleration efficiency with argon, lower laser desorption efficiency, and poorer insulin entrainment in the pulsed argon jet using our current (33) Vestal, M. L.; Juhasz, P.; Martin, S. A. Rapid. Commun. Mass Spectrom. 1995, 9, 1044-1050.
5606 Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
Figure 5. Time-of-flight mass spectrum of melanostatin, β-carotene, and bovine insulin obtained after laser desorption, acceleration in a supersonic molecular beam, and subsequent hyperthermal surface ionization. While for melanostatin helium and for β-carotene neon was used as carrier gas, argon gave the best results for bovine insulin. Each spectrum represents a summed average of 100 individual single-shot spectra.
nozzle design, combined with reduced TOF-MS ion optics and ion detection efficiency. Thus, further optimization and investigation is needed with insulin and other large molecules. However, comparing the signal intensities in the spectra shown and assuming desorption and detector efficiency to be independent of the molecular mass to a first approximation, one can estimate that the HSI efficiency for insulin is comparable or even higher than for the other substances investigated here. We believe that HSI of large molecules should be particularly effective since their ionization potential should be reduced with increasing molecular size and probably the ionization potential of insulin itself is already around or below the 6.4-eV work function of rhenium oxide. Thus, the kinetic energy of large molecules serves mostly to ensure their desorption as ions from the surface and not for the ionization process itself. On the other hand, the kinetic energy required for the collision-induced dissociation of large molecules increases with
the molecular size due to their increased vibrational heat capacity, and thus, HSI fragmentation of large molecules is unlikely due to their insufficient and inefficient acceleration. CONCLUSION This work successfully demonstrated the application of hyperthermal surface ionization for the mass spectrometry of large, laser-desorbed and jet-accelerated molecules. The results obtained for polycyclic aromatic hydrocarbons and biomolecules with molecular weights up to 5733 amu confirm that HSI is a soft ionization method producing mainly intact molecular ions. Unfortunately, due to vacuum restrictions, it was not possible to study a possible enhancement of fragmentation by increased jet velocities in these experiments. Further investigations in the optimization of the jet acceleration of the laser-desorbed neutrals and the ionization behavior of other biomolecules are underway. However, at the current stage of the
experiments, the combination of laser desorption and hyperthermal surface ionization already promises to be a versatile tool for biological and biochemical mass spectrometry. It requires only minimal sample preparation, is flexible due to the set of adjustable experimental parameters, and is almost universally applicable. ACKNOWLEDGMENT The authors thank Prof. J. Grotemeyer for the loan of instrumentation. This work was supported by a grant from the German Israeli Foundation for Scientific Research and Development G.I.F. which is gratefully acknowledged.
Received for review May 27, 2003. Accepted May 29, 2003. AC0302197
Analytical Chemistry, Vol. 75, No. 20, October 15, 2003
5607
