Scanning Microscopy

Volume 5 Number 2 Article 3

4-20-1991

Restricted Energy Transfer in Laser Desorption of High Molecular Weight Biomolecules

Akos Vertes University of Antwerp, Belgium, [email protected]

Renaat Gijbels University of Antwerp, Belgium

Follow this and additional works at: https://digitalcommons.usu.edu/microscopy

Part of the Biology Commons

Recommended Citation Vertes, Akos and Gijbels, Renaat (1991) "Restricted Energy Transfer in Laser Desorption of High Molecular Weight Biomolecules," Scanning Microscopy: Vol. 5 : No. 2 , Article 3. Available at: https://digitalcommons.usu.edu/microscopy/vol5/iss2/3

This Article is brought to you for free and open access by the Western Dairy Center at DigitalCommons@USU. It has been accepted for inclusion in Scanning Microscopy by an authorized administrator of DigitalCommons@USU. For more information, please contact [email protected]. Scanning Microscopy, Vol. 5, No. 2, 1991 (Pages 317-328) 0891-7035/91$3.00+ .00 Scanning Microscopy International, Chicago (AMF O'Hare), IL 60666 USA

RESTRICTED ENERGY TRANSFER IN LASER DESORPTION OF HIGH MOLECULAR WEIGHT BIOMOLECULES

Akos Vertes* and Renaat Gijb els

Departm ent of Chemistry, University of Antwerp (U.I.A.), Universiteitsplein 1, B-2610 Wilrijk (Belgium)

(Received for publication November 15, 1990, and in revised form April 20, 1991)

Abstract Introdu ction Producing ions from large molecules is of distin­ With the growing importan ce of biomedical investi­ guished importance in mass spectrometry. In our present gations in organic analysis the emphasis has been shift­ study we survey different laser desorption methods in ing to the detection and structure determination of ever view of their virtues and drawbacks in volatilization and larg er and more complex molecules. Gas phase analyt­ ion generation. Laser induced thermal desorption and ical methods in general - and mass spectrometry was matrix assisted laser desorption are assessed with spe­ not an exception - exhibited increasing handicap with cial emphasis to the recent breakthrough in the field the growing molecular masses, because of the involatility (m/z > 100,000 ions produced by matrix assisted laser and instability of these materials. desorption). Efforts to understand and describe laser de­ The early answer to this challenge came in the sev­ sorption and ionization are also reported. We emphasize ent ies, in the form of the so called "soft" ionization tech­ the role of restricted energy transfer pathways as a pos­ niques. Field desorption, chemical ionization, plasma de­ sible explanation to the volatilization of non-degraded sorption, secondary ion, electrohydrodynamic and laser large molecules. desorption ion sources were developed to cope with non­ volatile compounds. A thorough overview of these meth­ ods with respect to their high mass capabilities was pub­ lished by Daves in 1979. Th ere seemed to be two dis­ tinct strategies to follow: (a) optimization of volatiliza­ tion conditions in terms of sample dispersion and heating rate and (b) direct ionization of molecules from a surface by exte rnal electrostatic field, particle bombardment or laser radiation . At this stage of development, however, analytical possibilities were limited to the mass range m/z < 10,000. During the eighties four techniques were able to overcome this limit. 252 Cf plasma desorption mass spectrometry (Sundqvist and Macfarlane, 1985) became available in the m/z < 45,000 mass range. Fast atom bombardment - an offspring of secondary ionization with spec ial sample preparation - reached the m/z ~ 24,000 region (Barber and Green, 1987). The two fore­ runners, however, became the electrospray ionization Key words: laser desorption, laser ionization, mass spec­ (Fenn et al., 1990) and matrix assisted la.5er desorption trometry, biomolecules, peptides, proteins, nucleotides, (Karas et al., 1989a) with m/z = 133,000 and m/z = nucleosides, saccharides, energy transfer, phase transi­ 250,000 high mass records , respectively. Th e latest de­ tion, disintegration, fragmentation, ion formation. velopments indicate capabilities for both techniques even *Address for correspondence: beyond these limits (Nohmi and Fenn, 1990; Williams A. Vertes and Nelson, 1990). Department of Chemistry, University of Antwerp On the other hand, one should not overemphasize (U.I.A.) , Universiteitsplein 1, B-2610 Wilrijk, Belgium the value of molecular weight determination by these Phone: (32)(3) 820.23.82, Fax : (32)(3) 820.23.76 methods . Several other, long establis hed techniques Electronic mail : [email protected] .be are available in the same or even in broader molecu­ lar weight range. Ultracentrifugation, light scattering,

317 A. Vertes and R. Gijbels gel permeation chromatography and gel electrophoresis existing ideas and models which try to explain laser de­ cover the molecular weight region from about 1,000 to sorption of large molecules. The concepts range from about 5,000,000 (Siggia, 1968). Th e cost and availability rapid heating (Beuhler et al. 1974) to mechanical ap­ of th e necessary instrum entation compares favorably to proaches (Williams, 1990), to restricted energy transfer the mass spectrometric equipment. When assessing the (Vertes, 1990) and to expansion cooling of the laser gen­ different molecular weight determination methods it is erated plume (Vertes et al., 1989b, Nelson et al. , 1989). necessary to keep in mind that mass spectrometry pro­ Las er indu ced thermal desorption vides directly mass to charge ratio data whereas the sep­ aration techniques measure particle mobilities and may The idea of laser indu ced thermal desorption incorporate systematic errors. (LITD) stems from early studies of the influ ence of laser An important aspect of the analytical value of the radiation on small molecul es adsorbed on solid surfa ces. mass spectrometric methods is their remarkably high If the substrate absorbs the laser radiation , it heats up sensitivities and low detection limits. Recent report s on a time-scale comparable to the lase r puls e leng th. The on electrospray ionization put sensitivities in the low resulting temp erature rise leads to the detachment of the pmol/ µL and detection limits in the low fmol region for adsorbed mol ecules. several 556 < m/z < 66,000 peptides (Van Berke! et al., It was realized soon that Q-switch ed laser pulses 1990). Both sensitivity and detection limits, however, ( durations are on th e ns time-scale) may lead to sub­ are highly compound dependent. The detection limit thermal velocity distribution s of the desorbed particl es increases with increasing molecular weight, because the (Wedler and Ruhmann, 1982). Furth er st udies suggested ion signal is distributed over more charged states (Loo that the desorption proce ss is adiabatic, espec ially in the et al., 1990). case of monol ayers (Simpson and Hardy , 1986). Similar , or even bette r results were obtained with The possibility to desorb cold molecul es from hot laser desorption . The sensitivity was in the low pmol/ µL, surfaces seemed very attractive and triggered st udi es on whereas the detection limits in the sub-pmol (Karas et larg er molecul es. Th e conflict of volatilization versus al., 1990) (10,000 < m/z < 100,000 peptides), or in the thermal !ability of these compounds was thought to be sub-fmol region (Hahn et al., 1987) (for porphyrin deriva­ overcome (van der Peyl et al., 1982a). Calculated sur ­ tives). face temperatur es showed strong correlation with th e Naturally, these numb ers should be evaluated with generated ion curr ent s. To impl ement the met hod for the achievements of other techniqu es in mind . As an higher mass compound s, however , was apparently not example, we quote here some results on the analytical completely successful: m/z ~ 10,000 seemed to be an up ­ performance of open tubular liquid chromatography. As per limit of ion produ ct ion by LITD (Ijam es and Wilkins , littl e as 4 fmol prot ein was hydrolysed and deriv at ized 1988). resulting in approximately 25 nL volume analyte solu­ Clearly , the evolution of substrat e surfac e tempera­ tion. Quantitativ e determination of 14 an1ino acid con­ ture plays a decisive role in th e course of events in LITD . st ituents was carried out with about 5 % error (Oates Tempo ral and spat ial temperature distributions, T(r,t) , and Jorgenson, 1990). Chromatographic and immunoa s­ in the sub strat e can be described by the heat conduction say analysis are serious competitors in terms of sens itiv ­ equations: ity , with usually much lower cost of instrum ent at ion and consequently with better availability. Adv entur es to volatilize large mol ecules and de­ C(T) EJT(r, t) - divJ(r, t) + l (r , t), (1) tect minute quantiti es with the aid of lasers and mass at spectrometry will be the topic of this revi ew. Laser desorption mass spectro metry of nonvolatile organic molecules was surveyed by Shibanov (Shibanov 1986), J (r, t) = K (T )VT (r, t) (2) but vigorous development in the field since then has changed the landscape considerably. where I(r,t) is the source term describing the laser heat ­ We will focus our attention on what the different ing and J(r,t) is th e heat flow. C(T) and K(T) are th e laser desorption and ionization methods have to offer in specific heat and the heat conductivity, respectively, gen­ excess of the capabilities of the more widespread analyt­ erally exhibiting considerable temperature depend ence. ical tools. Special emphasis will be given to the mass The source term can be expressed in terms of laser irra­ range and mass accuracy, to detection limits and sensi­ diance , I0 (r ,t), corrected for surface reflection: tivity, to the freedom from matrix interferences and to the availability of structural information. Two types of I(r, t) = alo(r, t)exp[-a z] (3) experiments will be distinguished in separate sections: laser induced thermal desorption and matrix assisted where the light propagates along the z axis perpendic­ laser desorption. ular to the substrate surface. a is the light absorption In the last section we enlist and comment on the coefficient of the substrate at th e laser wavelength.

318 Restrict ed Energy Transfer in Laser Desorption

Th e solution of the linear problem (K and C con­ ionize the sample. Mass analysis was carried out by a stant) for general laser intensity distributions and for time-of-flight mass spectrometer. the special case of a Gaussian beam was found by Lax Achieving softer ionization conditions for a broader in 1977. Th e non-linear problem of temperature depen­ class of materials was a remarkable advance in itself. dent heat conductivity was also solved analytically (Lax, Further exciting developments came with the introduc­ 1978). tion of about 20 keV post-accel eration at the end of Experim ental results became also available to trace the flight tube. The purpose of this modification was surface temperature rise as the consequence of a laser to increase the detection sensitivity for large mass ions. puls e. Resistance change of a vapor deposited platinum These particles otherwise hit the conversion dynode of stripe on an insulating substrate provided tim e resolved the electron multiplier detector with relatively low ve­ information on surface heatin g (Zenobi et al. , 1988). locity resulting in poor conversion. Laser desorption of Comparison of the measured temperatures with analyti­ large (m/z > 10,000) (Karas and Hillenkamp , 1988) and cal solutions of Eqs. (1-3) in the linear, one-dimensional very larg e (m/z > 100,000) (Tanaka et al., 1988, Karas et case and with numerical solutions in the nonlin ea r case al., 1989a) ions generated distinguished inter est among concluded in favor of the latter (Philippoz et al., 1989). mass spectroscopists who have been struggling with the Both experiment and calculations indicated ex­ volatilization problem already for more than a decade. 8 tremely high heating rates , generally in excess of 10 Due to the extended mass range , several important K/s. In the study of Deckert and George , 1987, surface classes of molecules (proteins , certain polymers) became decomposition reaction and desor ption were consid ered available for ma ss spect rom et ric analysis and invest iga­ to be competing first order processes. It was possible to tion. A typical example of ,8-D-galactosidas e MALD show the existence of a crossover point of the product spectrum is shown in Fig. 1. Bett er preparation proce­ yield curves indicating the takeover of desorption. dure and the introdu ct ion of new matrices yielded even Analytical application of LITD excel in particu­ higher upper mass limits reaching m/ z ~ 250,000 Dalton larly low detection limits. In the case of protoporphyrin (Karas et al., 1989b). IX dimethyl ester, subfemtomol quantities of the ana­ lyt e provided det ecta ble signals (S/N=2) (Hahn et al., n-O-GALACTOSIOASE 1987). Th e mol ecules were desorbed by a CO2 laser .,., 4 pulse, and subsequently ioni zed by a frequency quadru­ ( L 30) C pled Nd-YAG laser. Ion detection was carried out by a ::, time-of-flight mass spectrometer. Lin ear dependence of -... 3 the parent ion signal int ensity with adso rb ate concentra­ '- tion was found over a rang e of five orders of magnitude '- 2 M, 111 no (Zare et al., 1988) offering remarkable quantitation pos­ .,> sibi lit ies. -.. C .,.. Matr ix assisted laser desorption C

Another possibility to deliver the laser energy to the samp le is direct light absor ption. In this scheme the sam­ --- ple has to hav e high absorption coefficient at the laser ------wavelength. For commo nly used ultra violet lasers this --- condit ion is not met for a larg e numb er of important bio­ molecules. Th erefore, the idea of mixing these samples with a good absorber came as a possible enhanc ement of the method (Karas et al., 1987) . Ind eed, resonantly absorb ing sub sta nces exhibit ed about an order of mag­ nitude lower threshold irradianc es ( the lowest irradian ce necessary for ion generation). Th e lower the irradianc e &ONO is, the softer the ionization met hod can be considered. '°"° Based on this observation, the matrix assisted laser Figure 1. High mass region matrix assisted laser de­ desorption (M ALD) met hod was introduced, esse ntially sorption (MALD) ma ss spectrum of ,8-D-galactosidase as a special samp le pr eparatio n technique. In the early enzym e. Signal averag ing improv ed the signal to noise version of the expe riment a dilut e solution of the ana­ ratio (30 shots) (Karas et al., 1989b). lyt e ( ~ 10- 5 M) was mixed with an equal amount of a 5 x 10- 2 M aqueous solution of nicotinic acid. A droplet of this mixture was air dried on a metallic substrate and These findings had a considerable impact on the introduced into the mass spectrometer. Mod erat e irra­ pace of laser desorption investigations. In Fig. 2. we 6 8 2 dian ce (10 - 10 W /cm ) frequency quadrupl ed Nd-YAG show, how the number of submitted articles related to laser pul ses (266 nm, 10 ns) were used to desorb and MALD increased during the eighties. Clearly , the num-

319 A. Vertes and R. Gijbels ber of art icles shows exponentia l growth after the dis­ broadened even further by the experiments on MALD covery of MALD. using infr ared radiation of a Q-switched Er-YAG laser (>-=2.94 µm) (Overberg et al. , 1990) and of a CO 2 laser (>-=10.6 µm) (Hillenk amp, 1990). Th e relative ly high 40 * price tag on laser equipment made these findings espe­ cia l'.y rema rkab le. As a consequence, a number of groups 35 eqmpp ed already with any of the indicated lasers can j oin the studies . ~30 Th e other direction of extend in g the method's pos­ 0 sibiliti es was the sea rch for approp riate and better m a­ 025 trices. In a thorough study Beav is and Cha.it tested (_) abo ut fifty different matrix materials (Beavis and Chait ~20 1990c). Using frequency quadrup led Nd-YAG lase; ::J (>-=266 nm) the best results in peptide volat ilizat ion 0..15 an_d ionization could be obtained by app lying ni cot ini c 4- acid , pyrazynoic acid, van illic acid, ferruli c acid, sinap- 010 1mc acid, caffeic acid and certain other cinnam ic ac id :::::i:t::: derivatives. Nitrobenzy l alcohol on a fibrous paper sub ­ 5 strate turned out to be a useful matr ix as well (Zhao et al. , 1990 ). In addition to the matrices used in the UV 84 85 86 87 88 89 90 9 1 experiments ca rb oxilic acids, glycero l and urea proved Year to be also useful in the infr ared expe rim ents. Low an­ alyte specificity and moderate volatility in the vacuum system are among the desirable features expected from Figure 2. The number of pub lications related to matrix prospective matrices. assisted lase r desorption in different years. (The year of manuscript submiss ion was conside red.) * - only the Exp_eriments aimed at the better understanding of first half of the year. the physica l bac kground of the MA LD phenomena were launc hed at several lab orato ries. Ini tial kin et ic ene rgy distributions of MALD generated ions were measured A growing number of groups joined the investiga­ utilizing pulsed ion extract ion in a Wiley-McLaren type tions (Beavis and Cha.it, 1989b; Nelson et al., 1989; time of flight instrument (Speng ler and Cotter, 1990). Salehpour et al. , 1989; Ens et al., 1990; Frey and Departing ion energies showed about 1 eV mean value. Holle, 1990; Hettich and Buchanan , 1990; Nuways ir and The question of substrate participation was rev isited in a Wilkins, 1990; Spengler and Cotter , 1990; Vertes et al. , study, where the laser and the ion optics were situated on 1990a; Zhao et al., 1990). Detection limits ranging from opposite sides of the suspended sample crystals (Vertes 5 pmol to 50 fmol and mass accuracy of about 0.1 % et al. , 1990a). Successfu l matrix assisted experiments on were establ ished (Karas et al. , 1989c). Both charac­ different low mass peptides in this transmission geometry terist ics, however, showed strong samp le dependence. suppo rted the idea of negligible substrate participation Beavis and Chait introduced cinnam ic acid derivatives in the case of a strong ly absorbing matrix. as new matrix materials showing sim ilar or superior fea­ tures to nicotinic acid (Beavis and Chait, 1989a). Also Several groups were ab le to establish laser irradi ­ negative ion spectra of proteins were observed (Beavis ance thresholds in ion generation (Hedin et al., 1990; and Chait, 1989b; Salehpour et al., 1989). High ma ss Ens et al., 1990). This phenomenon has been reported glyco-proteins and hydroph obi c proteins not access ibl e many times ea rlier in the genera l cont ext of laser ioni za­ for plasma desorption mass spectrometry showed enco ur­ tion. Alth ough the abso lu te numbers for the irr ad ian ce ag ing response in MALD expe rim ents (Salehpour et al., vary from study to study or even from samp le to sam­ 1990). ple, the values are generally situated in the vicinit y of 2 severa l tim es 106 W/cm . The uncertainties are mainly Mass accuracy of t he MALD method was squeezed due to p oo rly defined light intens ity measurements and to ± 0.01 % by using int ern al calibration of the ma ss to uncontroll ed variations in sa mpl e morphology. Usin g sca le (Beav is and Chait, 1990a). In most cases, the re­ singl e ion cou nting the irr ad ian ce threshold of insulin ion quired sa mpl e amount (0.5 to 1 µL of samp le solution generat ion and a narrow range irr adiance dependence of cont aining severa l pmol or severa l hundr ed fmol ana­ ion yield were determined (En s et al., 1990) (see Fig . 3). lyt e) com par ed favorably with the samp le require ments With the analysis of pulse height distributions they also of other mass spectroscop ic techniques. inferred that MALD is the result of a collective effect. Frequ ency tripled Nd-YAG la ser (>.=355 nm) was Preliminary results were repor ted on th e metast a­ successfully tried to demon st rate th at MALD is feasible ble decay of MALD generat ed ions. 354.5 nm la ser de­ with more affordable nitrogen lasers (>.=337 nm) (Beav is sorption of insulin molecules resu lt ed in molecular ion s an d Chait, 1989c). Recently, possibilities have been decaying in the field free flight tube, loosin g probably a

320 Restricted En ergy Tr ansfer in Laser Desorpt ion relati_vely small _neutra l ent ity (Beav is and Chait, 1990e). Another attempt to combine MA LD with FT MS con­ Beav is and C hait also concluded that meta stab le decay is cluded that th e ni cot ini c acid matrix material sub limed an indi cation of high int erna l (vibrat ional) temperatures from the prob e tip within 15 minutes und er the usual of the desorbed ions. vacuum of FTMS measurements (lo-s torr) (Nuway sira and Wilkins , 1990). Possible alternatives for matrix ma­ ter ial were tested with encourag ing results in the case of sinap inic acid and sucrose. Rece ntly , a different scheme of la ser volatilization h as been introdu ced (Nelson et al., 1989, Becker et al., 1990 ). Thin films of the aqueous analyte soluti on were 10° frozen onto cooled metal prob e tips. Th e laser radiation ab late d the ice film. Th e volatilized material was eith er "O collected on a solid surface and subsequently ana lyzed Q) 10·1 using gel electro ph ores is or directly analyzed by time >= of flight ma ss analysis with or with out postionization. Q) > Electrophoresis results on DNA digest showed very high ...... mass part icles (m/z ~ 6 MDa) in the plume, whe reas ro TOF detection revea led m/z ~ 18,500 ions (Ne lson et Q) al. , 1990a). Th ese results were rationali zed in terms of a: sudd en heating of the metal surface fnllowed by fast heat trans fer to the ice layer and by exp losive boiling (Nelson ~ I and William s, 19906 ). I 6 Applications of MA LD MS are just beginning to 1 ~2x 10 W/cm2 mushroom. The most imp ortant ach ievements are ex­ I pected on areas where the uniqu e features of the metho d 1o ·· 6 r 7 10 10 can be utilized. The un para lleled mass acc ur acy in the high mass region made it po ssible to rev ise the cali­ brat ion curves of certain com mercial gel electrop hore­ Power Density (W/cm2) si~ molecular weight mark er kit s (K rat zin et al., 1989). High mas s capab iliti es are exp loited in the st udi es of F igure 3. Irradiance dependence of the insulin molecu­ monoclona l antibod ies (Siege l et al.,1990a. and Siegel lar ion yie ld clearly exhibit thresho ld beha vior (Ens et et a.I., 19906 ) and in the invest igation of large enzy me al. , 1990). The different symbols cor respond to different mol ecules (Hillenkamp and Karas, 1990). An ot her ben­ detector openings. eficial feature was emph asized in an ex periment where the mass spectrum of commercia l bovine milk was taken (Bea.vis and Cha.it, 19906 ). This exerci se intended to D esorb ed neutra ls and molecular ions were investi­ demonstrat e the potential to ana lyze prot eins in unpu­ gated using tunable UV laser light for desorption , super­ rified biological fluid s, which are comp lex mixtures of son ic j et cool ing and resonance enhanced multi-photon organic and inorg anic const ituents (Bea.vis and Chait, ioni zat ion (REMPI ) (Fr ey and Holl e, 1990). It was in­ 1990d). Phy siological sa lt concentrat ions - usua lly pro­ structive to compare the effect of matr ix assistance and hibitive for MS methods - were not detrim ental on MALD spectra (Karas et al., 1990). superson ic coo ling on the fragm entation pattern s. With­ out coo ling and without matrix no molecul ar ion signa ls Oth er groups of bio-organic compounds such as were obse rv ed for gramicidin D (Mw= l 881), whereas the oligonucl eot ides and car bohydr ates also exhibit spect ral responses , although for the time being only in the lower introduction of either jet cooling or matrix assista nce mass region (m/z < 32,500; m/z < 25,000 and m/z < alone was sufficient to suppress the frag mentatio n. Pr e­ limin a ry est im ates of neutral velocity distributions in­ 4,000 respectiv ely) (Karas et al., 1990). dicate about 300 K translational temperature, ind epe n­ Fast development is expected in the near future in dent of lase r wavelength (>, = 10.6 µm or 266 nm). the diversificat ion of applications , althou gh, the inertia of stan dard bioch emical approach is imm ense . Molecular Pil ot expe rim ents to utili ze MALD in Fouri er­ weight dete rminati on of large biomolecules is routinely transform ma ss spectrometers (FTMS ) have been de­ done as part of the gel electro phor esis separatio n step. scrib ed (Hett ich and Buchan an , 1990). Low ma ss ions It is difficult to comp ete in pri ce an d in ease of opera­ (m/ z < 2000) were exhibiting strong matrix en hance­ tion with th e conventional met hod , especially, since the ment und er FTMS conditions, but efforts to detect introduction of pre-ca.st gels considerably shortened the mol ec ul ar ion s in the high er ma ss region were un success­ an alysi s time, as well. MALD will hav e to offer sub­ ful. The origin of this deficiency was linked to the de­ st antial superiority in certain features, in order to get tection limitation s of the Fourier-transform instrument. widespread recog nition. Th e price of the instrumenta.-

321 A. Vertes and R. Gijbels tion will still be prohibitive for many laboratories but la ser generated plum e expans ion has cont ribut ed to the centralized handling of th e tough problems in spec ial­ insight of laser solid interaction (Vertes et al. , 1988a , ized fa ciliti es seems feas ible. Vertes et a.I., 1990b) , especially in the neighborhood of th e pla sma ignition threshold (Vertes et a.I., 1989a) . Propo sed mechani sms and mod els Shock wave developm ent und er moderate and high ir­ ra.diance conditions was demonstrated and kinetic en­ Mechanisms for the desorption of larg e molecules by ergy distributions of desorbed and ablated ions were fast ion excitation hav e been revi ewed recently (John­ successfully accounted for (Vertes et al., 1989c). A re­ son , 1987). Beca use of similariti es between fast ion de­ fined version of the one-dimensional model was able to sorption and lase r desorption of biomolecules and be­ handle phase transitions, surface reces sion and heat con­ cause the former has a longer hi sto ry th ere are plenty duction processes in the solid, and electro n- neutral in­ of fertili zing ideas to be transferred. However, here we verse Br emsstrahlung light absorption, multipl e ioniza ­ choose a differ ent ap proach. tion and radiation cooling in the plume (Bala zs et a.I., There are two major questions to answer in con­ 1990). The adaptation of this model to typical conditions nect ion with the lase r indu ced volatilization of larg e of MALD experiments provided encouraging preliminary molecul es . results (Vertes et a.I., 1991). A frequen cy quadrupl ed a/ What is the natur e of the laser induced pro cess lead­ Nd-YAG laser pul se acted on a matrix surface at 107 ing to the transition from solid to gas phase? W /cm 2 irra.dia.nce. As a result , the surface temperature b/ How can large molecules escape fragm entation in an was temporarily elevated somewhat above the sublima­ environment abruptly energized by the laser pulse? tion temp erat ure rel eas ing , therefore, a certain a.mount Thes e questions are equally releva nt both in LITD of vaporized matrix mat erial. The rema rkabl e feat ure of and in MALD experiments. The rough options to an­ this vapor cloud was that th e expansion cooling overcom­ swer question a/ are to invoke thermal or electronic pensated the laser heating, and th e plum e temperature process es, but closer insp ect ion of the problem has led dropped below 200 K . In this way the plume acted as to sugg est ing several different mechanism falling within a jet-cooling device, possibly contributing to the la.ck of these categories. Question b/ is specifically emphasized fragmentation in the case of the embedded molecules . in the MALD situation where the matrix ion s may un ­ Thi s mod el is able to account for the suitability of ma­ dergo extensive fragmentation whereas the embedded trices with low sublimation temperatures, but the expla­ large mol ecul es desorb intact. nation of matrix fragmentation is not straightforward. Extremely fast energy depo sition - as it is the Measuring las er induced thermal decomposition case with laser heating - generates a stro ngly non­ pro cesses in th erm ally ext remely la.bile sub sta n ces with ­ equilibrium population of energy levels in the solid . out and in the presence of nicotini c acid matrix indi ­ Ther e are num ero us different regimes for the relaxation cated the participation of a cooling mechani sm if the of this situ at ion. We enlist here certain possibilities matrix was present (C la.ereboudt et a.I., 1991). Arylt­ which involve volat ilizatio n and have shown up in the riph enylpho sp honium halid e guest molecules were used exp lanation of related phenomena. as "mol ecul ar thermometers" in typical MALD expe ri­ ( a/i} Spinodal decomposition. It is known from th er­ ments. Th e ma ss sp ect ra revea led no thermal decompo­ modynamics that condensed phases can only be super­ sition and low int ern al energi es of the guest molecules heated up to a point where homogeneous vapor nucle­ and therefore supported the cool plume model. ation beco mes dominant over het eroge neous nucleation ( evaporation) and the whole pha se is suddenly trans­ ( a/iii} Phonon avalanche. Fain and Lin treated formed into vapor (von Alleman , 1987). A rough esti­ UV lase r induc ed non-s elect ive desorption theoretically mate of the corresponding temperature is 0.9 x Tc where (Fa.in and Lin, 1989a , Fa.in and Lin , 1989b ). In their Tc is the critical temp erature . This ph enom enon - mod el the a.dsorba.tes (CH2h on Al2O 3 , Ag or Al) are sometimes called phase explosion or sp inodal decompo­ exc it ed to higher electronic states. Th e exc itation en­ sition - was referred to in the explanation of fast atom ergy is then converted into internal vibrations of the bombardment (FAB) ionization experiments (Sunner et adsorba.tes and subsequently to phonons. They showed al., 1988). Recently , preliminary molecular dynamics that a system of excited anharmonic oscillators can be­ simulations have shown the feasibility of the spinodal de­ come unstable under the influence of some external force composition mechanism in particle and laser desorption field. If the energy dissipation from th e phonon modes experiments (Shiea and Sunner, 1990). Molecular dy­ is slower than their energy gain the number of phonons namics simulation of the mod el is capable of visualiz­ shows exponential divergence or in other words a phonon ing strongly non-equilibrium phase transitions induced avalanche can be observed . The ma.in characteristics of by high energy projectiles in two dimensional Lennard­ the model are the existence of a surface coverage de­ Jones fluids. Extension of the calculations to three di­ pendent threshold laser fluence for the avalanche and mensions and to laser excitation in MALD experiments the mo lecular non-selectivity of the above-threshold de­ a.re expected. sorption. On the other hand the below-threshold reg ime of their model predicted selective desorption . Similar ( a/ii} Cool plume. Hydrodynamic description of treatment of IR laser induced desorption exhibited nei-

322 Restricted Energy Transf er in Laser Desorption ther molecular selectivity nor threshold behavior and no From our point of view these investigations should phonon avalanche was reported (Fain et al., 1989). have a slightly different emphasis. First of all, in LITD ( a/iv) De1Jorption induced by electronic tran1Jition1J and in MALD the energy is clumped into the solid sub­ {DIET). Low energy electrons or UV photons may ex­ strate and into the matr ix, not into the large molecu le. cite the surface adsorbate complex to a repulsive anti­ Ther efore, a reverse flow of energy is generated and bonding electronic state. A possible relaxation of this should be studied. Second , the amount of deposited excitation is the departure of the adsorbate from the energy is deliberately set to levels where desorption surface. This mechanism of desorption has been discov­ and/or phase disintegration occur, therefore, highly an­ ered and rediscovered several times (Menzel and Gomer, harmonic displacements are induced. Nevertheless, we 1964 , Antoniewicz, 1980). A relatively new development cited the studies above, becaus e they indicate the pos­ in the utili:tation of the DIET model was its application sibility of time resolved energy transfer expe riments of to bulk etching of organic polymers by far UV radiation similar nature. Scarce reports of vibrational tempera­ (Garrison and Srinivasan, 1984, Garrison and Srinivasan ture measurements of molecules after laser evapo ration 1985). In the framework of this DIET model it was pos~ from a cryogenic matrix show vibrational cooling down sible to rationalize polymer ablation without melting or to T=170±30 K ( aniline in CO 2 matrix , Nd-YAG laser, 8 2 any other thermal effect. In the irradiated volume of 3x10 W/cm ) (Elokhin et al., 1990) . Under different the polymer the monomer units were thought to be in­ conditions, similar temperatures were also indicated by stantaneously excited to the repul sive state and leave the cool plume model (Vertes, Irinyi and Gijbels, work in the bulk of the solid with coher ent motion. Concerning progress). the applicability of this model to :tvIALD, the recently (b/ii) De1Jorption V/J. fragmentation in LITD. There demonstrated possibility of IR MALD seems to be a clear is a long history of invest igating desorption and fragmen­ contraindication. tation kin et ics at different heating rat es (Beuhl er et al., It is st raightforward from the arrangement of LITD 1974, Deckert and George, 1987) . Th e microscopic dy­ and MALD exper iments that laser energy deposition oc­ namics of energy transfer from a rapidly heated surface curs mostly to anything but the larg e molecules. In th e to the adsorbate species has been treated by stochastic case of LITD the solid substrate acts as an absorber trajectory modeling on a computer (Lucchese and Tully, wherea in MALD expe rim ent s th e matrix molecules are 1984, Lim and Tully, 1986) by classical molecular dy­ to be energized. The answer to question b/ is relat ed namics simulation (Holme and Levine, 1989) and by the to this separat ion of energy deposition from the larg e evaluation of the survival probability of adsorbates with­ molecules. out degradation (Muck erman and Uzer, 1989, Zare and Levine, 1987). All these studies arriv at similar con­ (b/i) Vibrational energy transfer experiment/J. In­ clusions: if the heating of the surface is rapid eno ugh tramolecular and int ermole cular vibrational energy the desorbing spec ies may hav e considerably lower tem­ transfer in condensed phases has been studied by picosec­ perature. In the case of frequency mismatch between ond laser spectroscop ic method s (Se ilmeier and Kai ser, the physisorption bond and the chem ical bonds of the 1988) . Th ese investigations are aimed at the understand­ adsorbat e an energy transfer bottleneck is formed for ing of vibrational energy redi st ribution and equilibration the surface phonon s pump readily only the adsor~tion after ultra short light pulse excitation of 10 ps duration bond (Zare and Levine, 1987). Th e bottl eneck model in and less. Larg e organic molecul es ( usually dyes) dis­ its original form was suitable to explain LITD but not solved in sma ll molecul e organic liquid s xhibit the fol­ MALD exper iments. lowin g features: If excited above 1000 cm- 1 they often redistribute (b/iii) Popcorn model. Another approach, stem­ the excess ene rgy over the vibrational manifold of ming from high energy particl e indu ced desorption was the electron ic ground state with a relaxation time based on the rapid vibrational excitation of the surface sma ller than 1 ps. It has also been shown that sample layer including the embedded large molecules the transient internal temperature of the excited (W illiam s and Sundqvist, 1987). According to this pic­ molecule is a meaningful term. ture, vibrational excitation is accompanied by mechani­ cal expan sion on the picosecond timescale. This expan­ Int ermol ecular energy transfer is also very fast and sion detac hes particles from the surface layer s by pushing has a relaxation time of about 10 ps, largely depen­ the expa nding molecules against the substrate and gener­ dent on th e degree of excitation and on the solvent ating reaction momentum. This model suggests an inter­ molecules. esting way of converting internal excitation of molecules Excitation to high vibrational states of the S 0 to their translation. It is worthwhile to consider this ground state may exhibit fast relaxation to the bot­ mechani sm to explain MALD , however , one can expect tom of the first electronically excited state (S ). Vi­ 1 difficulties in accounting for matrix fragmentation and brational redistribution of the energy in So and in in describing the role of the matrix . S1 has relaxation times in the order of 0.5 ps . . {b/iv) Homogeneous bottleneck model. This descrip- Similar studies were carried out on surface adsorbates too (Heilweil et al., 1989). tion has been an extension of th e bottleneck model in­ troduced for LITD to the MALD situation (Vertes et

323 A. Vertes and R. Gijb els al. , 1990c, Vert es and Levin e, 1990). Th e laser energy ers (P . Willi ams, M. Karas and U. Bahr ) con tribut ed is adsorb ed by th e matr ix in th e same ph ase where th e sub st anti ally to th e clari ty of the presentat ion t hrough large mol ecules are. As a consequ ence, th e energy tr ans­ th eir com ments and qu est ions. fer is from th e m atri x molecules towar ds th e emb edd ed References large molecules and th e bott leneck is p osed by the cou ­ plin g hydroge n bond s be tween them. Tw o types of ph ase Ant oniewicz PR . ( 1980). Model for elect ron- and disint egrat ion mechani sm s were considered: sublimati on ph ot on -stimul ated desorpt ion . Ph ys. Rev . B 21, 3811- and mechani cal frag ment at ion. A sim ple comp eti tive ki­ 3815. net ic mod el was able to rep rodu ce several expe rim ental Bal azs L, Gijb els R , Vertes A. (1990). Expansion finding s: th e existence of a lase r irr adiance thr eshold and of lase r generat ed plum es nea r th e pl asm a form at ion it s est im ate d value, th e suit abili ty of mat ri ces with low thr eshold. An al. Chem . 63, 314-320. phase tr ansition te mp erat ure, the low volum e concentra ­ Barb er M, Gr een BN. (1987). Th e analysis of sm all tion requir emen t for the large mo lecules , the need for a prote ins in the m olecul ar weight range 10- 24 kD a by short lase r pul se, i.e. for fast hea tin g rates . It was also m agneti c secto r mass sp ectromet ry. R ap id Com m . Mass pr edi cted th at sampl e cooling would ext end th e po ssibil­ Sp ectro m. l , 80-83 . iti es of th e meth od . A generalizat ion of the idea involved Beav is RC, Chait BT . (1989a). Cinnam ic ac id in MALD suggests that new meth ods can be successful , de rivat ives as mat ri ces for ultrav iolet laser deso rpt ion if th e lib erat ion of large molecules by ph ase disinteg ra­ m ass sp ectro met ry of prote ins. Rapi d Comm. Mass tion pr ece des their destru cti on by frag mentat ion (Vert es, Spec tro m . ;i, 432-435. 1990). Beavis RC, Ch ait BT . (1989b ). Facto rs affect ing the Th ere are un answered quest ions in all the models ult rav iolet lase r deso rpti on of prot eins . Rap id Com m . Mass Sp ectrom . ;i, 233-237. discussed abo ve. Th ese models do not acco un t for ion form ati on , - a pr erequi sit e for m ass analysis and de­ Beav is RC, Chait BT. (1989c). Mat rix-ass isted tec tion in m ass spectrom et ry. Fur ther shortco ming of lase r-d esorp tion ma ss sp ect rom etr y using 355 nm ra di a­ th ese m echani sm s th at neith er of th em explain th e larg e tion . Ra pid Comm. Mass Spectro m . ;i, 436-439 . differences in efficiency b etween matri ces with simil ar Beav is RC , Ch ait BT . ( 1990a). High-acc uracy ph ase tr ansiti on te mp erat ur es and optical ch aracte ris­ mo lecul ar mass dete rmin atio n of prote in s usin g mat rix­ tics (Beav is and Ch ait 1990c) . assisted laser desorpt ion m ass spectro met ry. Anal. Chem. 62, 1836-1840. Conclu sions Beavis RC , Chait BT . (1990b ). Lase r desorp tion of prot ein s in unpurifi ed bi ological fluid s: a new too l for It is ap pare nt from the sur vey of lase r desorp tion lit­ biologica l resea rch . In : P rocee dings of the 38t h ASMS erat ur e, espec ially since th e fer tilizing discovery of new Conference on Mass Sp ectro metry and Allied Topi cs. sa mpl e prepa rat ion tec hniques that the met hod has great Tu cson , Ari zona. Jun e 3- 8. Capr ioli R M . (ed) , 26-2 7. potent ial for the ana lysis of large b iomo lecules. Th e Beav is R C, Chait BT . (199 0c). Th e ro le of m at rices vir tues of lase r desorp tion put it into the forefront of or­ and wave lengt h in the m at rix assiste d laser deso rpti on ga nic m ass sp ectromet ric tec hni ques, toget her with elec­ of prote ins. In : P rocee din gs of the 38t h ASMS Confer­ trosp ray ionizat ion . At present, it is not poss ible to es­ ence on Mass Sp ectro metry and Allied Topics. Tucson, tab lish super iority of one of these ion sources , rat her we Ari zona. Jun e 3- 8. Ca pri oli RM. (ed), 152- 153. h ave to regard them as comp leme ntary poss ibilities of Beav is RC , Chait BT . (1990d). Rapid , sensitive analys is. anal ysis of pro te in mi xt ur es by mass spectro met ry. Pro c. Th ere are several suggeste d mechanisms and a num ­ Nat l. Acad. Sci. USA 81, 6873-6877. ber of different models for the desc rip tion of physica l Bea vis RC, Ch ait BT. (1990e). Invest igat ion of th e phenomena involved in laser deso rp tion . Th e volum e ma tri x ass iste d laser desorp tion of prote ins using a tim e­ and th e acc ura cy of exp erim ent al findin gs is not enough of-flight m ass spect rom eter. In: Meth ods and mecha­ at th e mom ent to ex clude or dispro ve any of them with nism s for pro du cing ions from large molecules. Sta ndin g certaint y. However , a general featur e of all th e concepts KG , En s W. (eds), Pl enum . New York. can b e ext rac ted already . Successful lase r desorp t ion of Becker C, Ju sin ski LE , Moro L. (1990). Infr ared int act larg e mol ecules seem s to show the repet itive pat­ laser-induc ed desorp tion of neutral orga ni c comp oun ds tern of res trict ed energy tr ansfer whi ch pr events exces­ from fro zen aqu eous solution followed by singl e-photon sive energy flow into th e desorbing sp ecies. ionization . Int . J . Mass Spectrom. Ion Pro cesses 95, Acknowledgement Rl -R4 . Beuhl er RJ , Fl anigan E , Gr eene LJ , Friedman L. The resear ch herein was partl y supp orte d by th e (1974). Prot on tran sfer mass sp ect rom etr y of p eptid es. Belgian Na tion aal Fond s voor Wetenschapp elijk Ond er­ A rapid heating techniqu e for und eri vati sed peptid es zoek. Th e author s would like to expr ess th eir gr atitud e cont aining arginin e. J . Am . Ch em . Soc. 96, 3990-3999 . to Elsevier Science Publish ers B.V. , th e copyright hold er Chi arelli MP , Gro ss ML . (1987). Mechani sms of of Fig . 1. Court esy of F . Hillenkamp (Fi g. 1) and K.G . ca ti oni za ti on of sucrose by sodium in laser desorption : St anding (Fig. 3) is also ackn owledged . Several review-

324 Restricted Energy Transfer in Laser Desorption a study by Fourier transform mass spectrometry. Int. J. studies of surface adsorbates. Annu. Rev. Phys. Chem. Mass Spectrom. Ion Processes 78, 37-52. 40, 143-171. Claereboudt J , Claeys M, Gijbels R, Vertes A. Heine n HJ. (1981). On ion formation in laser de­ (1991). Aryltriphenylphosphonium halid es as "mo lec­ sorption mass spectrometry with LAMMA . Int. J. Mas s ular thermometers" in matrix assisted laser desorption. Spectrom. Ion Phy s. 38, 309-322. In: Proceedings of the 39th ASMS Conference on Mass Hettich RL, Buchanan MV . (1990). Examination Spectrometry and Allied Topics. Nashville, Tennessee. of sample substrate effects in laser desorption FTMS May 19- 24. Paper in press. studies of biological compounds . In : Proc eedings of the Daves Jr. GD. (1979). Mass spectrometry of in­ 38th ASMS Conference on Mass Spectrometry and Al­ volatile and thermally unstable molecules. Mass Spec­ lied Topics. Tucson, Arizona. June 3-8. Caprioli RM. trom . 12, 359-365. (ed), 156-157. Deckert AA, George SM. (1987). Heating rates re­ Hillenkamp F. (1990). (UV)-laser desorption/ ioni­ quired for laser indu ced thermal desorption stud ies of zation mass spectrometry of larg e biomolecules: tech­ surface reaction kinetics. Surf. Sci. Letters 182, 215. niqu es, results, promises and limit ations. In : Proceed­ Elokhin VA, Krutchinsky AN, Ryabov SE. (1990). ing s of the 38th ASMS Conference on Mass Spectrometry Vibrational cooling of an ilin e molecules by laser evapo­ and Allied Topics . Tucson , Arizona. June 3- 8. Caprioli ration from a cryogenic matrix. Chem. Phys. Lette rs RM. (ed), 8-9. 170, 193-196. Hillenkamp F, Karas M. (1990). Mass spectrome­ Ens W, Mao Y, Mayer F, Standing KG. (1990). Ma­ try of peptides and proteins by matrix assisted UV laser trix assisted UV laser desorption with single-ion count­ desorption/ionization (LDI-MS). In: Methods in enzy­ ing. In: Proceedings of the 38th ASMS Conference on mology, Vol. 193, Mass Spectrometry. McCloskey JA . Mass Spectrometry and Allied Topics. Tucson, Arizona. (ed), Academic Press. New York. 280-294. June 3- 8. Caprioli RM. (ed), 24-25. Holme TA, Levine RD. (1989) . Computational Fain B, Lin SH. (1989a). Avalanche phenomenon in studies of rapid laser indu ced desorption: a microscopic UV laser-induced desorption. Chem. Phys. Letters 157, mechanism for select ivity. Surf. Sci. 216, 587-614. 233-238. Ijames CF, Wilkins CL. (1988). First demonstration Fain B, Lin SH. ( 1989b). Laser-induced exp losive of high resolution laser desorption mass spectrometry of desorption. J. Chem. Phys. 91, 2726-2734. high mass organ ic ions. J. Am. Chem. Soc. llQ., 2687- Fain B , Lin SH, Gortel ZW. (1989). Theoretical 2688. aspects of laser induced desorption. Surf. Sci. 213, 531- Johnson RE. (1987). Mechanisms for the desorption 555. of large organic molecules. Int. J. Mass Spectrom. Ion Fenn JB, Mann M, Meng CK, Wong SF, White­ Processes 78, 357-392. house CM. (1990). Electrospray ionization-principles Karas M, Bachmann D, Bahr U, Hillenkamp F. and practice. Mass Spectrom. Rev. l;l_,37- 70. (1987). Matrix-assisted ultraviolet laser desorption of Frey R, Holle A. (1990). Laser desorption of neu­ non-volatile compounds. Int . J. Mass Spectrom. Ion trals and molecular ions . A compar ison of mass spectra Processes 78, 53-68. and mechanisms. In: Proc eedings of the 38th ASMS Karas M, Bahr U, Hillenkamp F. (1989a). UV laser Conference on Mass Spectrometry and Allied Topics. matrix desorption / ionization mass spectrometry of pro­ Tucson, Arizona. Jun e 3- 8. Caprioli RM. ( ed), 212-213. teins in the 100 000 dalton range. Int . J. Mass Spectrom. Garrison BJ, Srinivasan R. ( 1984). Microscopic Ion Processes 92, 231-242. model for the ablative photodecomposition of polymers Karas M, Bahr U, Ingendoh A, Hillenkamp F. by farultraviolet radiation (193 nm). Appl. Phys. Let­ (19896 ). Laser desorption / ionization mass spectrome­ ters 44 , 849-851. try of proteins of mass 100 000 to 250 000 dalton. Angew. Garrison BJ, Srinivasan R. (1985). Laser ab lation of Chem. Int. Ed., Engl. 2.a, 760-761. organic polymers: Microscopic models for photochemical Karas M, Bahr U, Ingendoh A, Stahl B, Strupat K, and thermal processes. J. Appl. Phys. 57, 2909-2914. Nordhoff E, Hillenkamp F. (1990). Principl es and appli­ Hahn JH , Zenobi R, Zare RN. (1987). Subfemto­ cation of matrix-assisted UV-laser desorption/ionization mole quantitation of molecular adsorbates by two-step mass spectromet ry. In: Proce ed ings of the 2nd Interna­ laser mass spectrometry. J. Am. Chem. Soc. 109, 2842- tional Symposium on Applied Ma ss Spectrometry in the 2843. Health Sciences , Barc elona, Spain. Gelpi E. ( ed). Hedi n A, \1/estman A, Hak ansson P, Sundqvist Karas M, Hillenkamp F . (1988). Laser desorption BUR. (1990). Laser desorption mass spectromet ry - ionization of proteins with molecular masses exceeding some technical and mechanistic aspects. In: Methods 10000 daltons. Anal. Chem. 60, 2299-2301. and mechanisms for producing ions from larg e molecules. Karas M, Ingendoh A, Bahr U, Hillenkamp F. Standing KG, Ens W. (eds), Pl enum. New York. (1989c). Ultraviolet-laser desorption/ioni zatio n mass Heilweil EJ, Casassa MP, Cavanagh RR, Stephen­ spectrometry of femtomo lar amounts of large proteins. son JC. (1989). Pico second vibrational energy transfer Biom ed. Environ. Mass Spectrom . 18, 841-843.

32S A. Vert es an d R. Gijb els

Krat zin HD, Wiltfang J , I-.::aras M, Neuhoff V, His­ biomolecules. Rapid Comm. Mass Spectrom. ±, 293- chmann N. (1989). Gas- ph ase sequ encing aft er elec­ 296. troblotting on polyvinylid ene difluorid e membranes as­ Philippo z JM , Zenobi R , Zare RN . (1989). Pul sed signs correct mol ecular weights to myoglobin molecular heat ing of surfa ces: compa rison betwee n num er ica l sim ­ weight markers. An al. Bioche m. 183, 1-8. ulation, analytical mod els, and expe rim ent s. Chem. Lax M. (1977). Temp erat ur e rise induced by a laser Ph ys. Lette rs 158, 12-17. b eam. J. Appl. Phy s. 48, 3919-3924. Sal ehpour M, P erera I, Kjellb erg J , Hedin A , Is­ Lax M. (1978). Temp erat ur e rise indu ced by a laser lamian MA, Hakan sson P , Sundq vist BUR. ( 1989). beam II. The nonlin ea r case. Appl. Phys . Lette rs .a;i, Laser-induc ed desorption of proteins. Rapid Comm. 786-788. Mass Sp ectro m . J, 259-263. Lim C, Tully JC. (1986) . Molecular dynamics Salehpour M, P erera IK , Kj ellberg J , Hed in A , Is­ of non eq uilibrium infr equent event s: La ser-indu ced de­ lamian MA , Hak ansson P, Sundqvist BUR , Ro epsto rff sorption from surfac es. J . Chem. Phys . 85, 7423-7433. P , Mann M, Klarskov K (1990). UV-laser induced de­ Loo JA, Edmonds CG, Smith RD , Lac ey MP, sorp tion m ass spectro met ry of hydroph ob ic- and glyco­ Keough T. (1990). Comparison of electrospray ioni za­ proteins. In : Ion formation from organi c solid s (IFOS tion and plasma desorption mass spectra of pept ides and V). Hed in A, Sundqvi st BUR , Benn ingh oven A. (eds), prot eins. Biom ed. Environ. Mass Spectrom. 19, 286- Wil ey. Chichester. 131-135. 294. Seilm eier A , Kai ser W. (1988). Ultr as hort in - Lu cchese RR , Tully JC. (1984). Las er indu ced ther­ tramolecular and inte rmol ecular vibrational energy mal desorption from surfa ces. J. Chem. Phy s. 81, 6313- transfer of polyat omic mol ecule s in liquid s. In : Ultra ­ 6319. short la ser pul ses and applications. Kaiser W . (ed), Menzel D , Gomer R. (1964) Desorp tion from meta l Spring er . Berlin. 279-317. surfa ces by low-energy electrons. J. Chem. Phy s. 41, Shibanov AN. (1986). La ser desorp tion mass spe c­ 3311-3328. trometry of non volatil e organic molecules. In: La ser Muckerman JT , Uzer T. ( 1989). A theoretical in ­ analyti ca l spectroc hemist ry. Leto kh ov VS. (ed ), Ad am vestigation of micro scop ic dyn amic s and time sca les of Hilger. Bri sto l. 353-404. predesorption processes. J . Chem. Ph ys. 90, 1968- 1973. Shiea J , Sum1er J. (1990). Spinodal decompositions Nelson RW , Rainbow MJ, Lohr DE , Williams P. in mol ecular dynamics simul at ion s of bulk desorption. (1989). Volatilization of high molecular weight DNA by In : Pro cee ding s of the 38th ASMS Conference on Mass pulsed laser ablation of froz en aqueous solution. Scien ce Sp ectro met ry and Allied Topi cs. Tu cson , Arizona. Jun e 21.6., 1585-1587. 3- 8. Caprioli RM. (ed ), 166-167. Nelson RW, Thom as RM , Willi ams P. (1990a ). Siegel MM , Holland er IJ , Karas M, In gendo h A, Hil­ Time-of-flight ma ss spectrometry of nucl eic acids by lenkamp F . (1990a ). Matrix assiste d UV laser deso rpti on lase r ab lat ion and ionizatio n from a frozen aq ueous ma­ mass spectra of monoclonal antibodies. I: Analysis of trix. Rapid Comm. Mass Spectrom. ±, 348-351. oligosacc harid e content of monoclona l antibod ies . In : Nelson RW, Williams P. (1990b ). Mechanisms in Proceedings of the 38t h ASMS Conference on Mass Spec­ laser ablation ma ss spect rom et ry from frozen aqu eous trometry and Allied Topic s. Tu cson , Arizona. Jun e 3- 8. solutions. In: Pro ceeding s of the 38th ASMS Confer­ Caprioli RM. (ed), 158-159. ence on Mass Spectromet ry and Allied Topi cs. Tucson , Siegel MM, Holland er IJ , J ames JP , Haman n Ph , Ari zona. June 3- 8. Caprioli RM . (ed), 168-169. Hinman L, Karas M, Ingend oh A, Hillenkamp F . Nohmi T, Fenn JB. (1990). High ma ss ion s - how (1990b ). Matrix assisted UV laser desorption mass spec­ high is high? In: Pro ceeding s of the 38t h ASMS Confer­ tra of monoclonal antibodi es. II: Analysis of the loading ence on Mass Sp ectrom et ry and Allied Topic s. Tucson , of anticancer drugs conjugat ed to monoclonal antibod ­ Arizona. June 3-8 . Caprioli RM. (ed), 10-11. ies. In: Pro ceeding s of th e 38t h ASMS Conference on Nuwaysir LM , Wilkins ChL. (1990 ). Laser de­ Mass Sp ectro met ry and Allied Top ics. Tu cson, Ari zona. Jun e 3- 8. Caprioli RM . (ed), 160-161. sorption in an FTMS : investigations of sampl e prepa­ ration and wavelength depend ence . In: Proc eedings of Siggia S. ( 1968) . Surv ey of analytic al chemistry . Mc the 38th ASMS Confer ence on Mass Spectrometry and Graw-Hill. New York. 141-175. Allied Topics. Tucson , Arizona . June 3- 8. Caprioli RM. Simpson CJSM, Hardy JP . ( 1986) La ser-indu ced (ed) , 844-845. th ermal desorption of a tom s and mol ecules phy sisorb ed Oates MD, Jorgenson JW. ( 1990). Quantitative onto a metal surfac e. Ch em. Ph ys . Lette rs 130, 175-180. amino acid analysis of subnanogram levels of protein by Spengl er B, Cott er RJ. (1990). Ultraviol et laser open tubular liquid chromatography. Anal. Chem. 62, desorption / ionization mass sp ect rom et ry of prot eins 1577-1580. above 100 000 dalton s by pul sed ion extract ion time-of ­ Overberg , Karas M, Bahr U, Kaufmann R, Hil­ flight analysis. Anal. Chem. 62, 793-796. lenkamp F . (1990). Matrix-assisted infrared-laser (2.94 Sriniva san R , Brar en B, Dre yfus RW. (1987). Ultra ­ µm) desorption/ionization mass spectrometry of large violet la ser ablation of polyimid e films. J . Appl. Phy s. fil., 372-376 .

326 Restricted Energy Transfer in Laser Desorption

Sundqvist B, Macfarlane RD. (1985). 252 Cf plasma Vert es A, Juha sz P, Jani P, Czitrovszky A. (19886). desorption mass spectrometry. Mass Spectrom. Rev. 1, Kinetic energy dist ribution of ions generated by laser 421-460. ionization sources. Int. J. Mass Spect rom . Ion Processes Sunner J, Ikonomou MG, Kebarle P. (1988). SIMS ~. 45-70. spectra of alcohols and the phase explosion model of de­ Vertes A, Levine RD . (1990). Sublimation ver­ sorption ionization. Int. J. Mass Spectrom. Ion Pro­ sus fragmentation in matrix-assisted laser desorption. cesses .81,221-237. Chem. Phys. Letters 171, 284-290. Tanaka K , Waki H, Ido Y , Akita S, Yoshida Y, Viswanadham SK , Hercules DM , Schreiber EM, Yoshida T . (1988). Prot ein and polymer analyses up Weller RR, Giam CS. (1988). Ion-molecule reactions to m/z 100 000 by laser ioni zation time-of-flight mass in the negative ion laser mass spectra of aromatic nitro spectrometry. Rapid Comm. Mass Spectrom. 2., 151- compounds. Anal. Chem. 60, 2346-2353 . 153. von Allmen M. (1987). Laser-b ean 1 int eractions Van Berke! GJ , Glish GL , McLuckey SA. (1990). with materials, physical principl es and applications. Electros pray ioniz at ion combined with ion trap mass Spring er. Berlin. spectro metry. Anal. Chem. 62, 1284-1295. Wedler G, Ruhmann H. ( 1982). Laser indu ced ther­ van der Peyl GJQ , Haverkamp J , Kist ema ker PG. mal desorption of carbon monoxid e from Fe( ll0) sur ­ (1982a ). Thermal aspects of laser desorption mass spec­ faces. Surfac e Sci. ill , 464-486. trometry. Int. J. Mass Spectrom. Ion Phy s. 42, 125-141. Willia.ms P. (1990). Mechanical approaches to van der Peyl GJQ , Isa K , Haverkamp J , Kiste maker volatili zat ion of larg e biomolecules. In: Ion formation PG. (1982b ). Cationization in laser desorption mass from organic solids (IFOS V). Hedin A, Sundqvist BUR , spectro metry. Nucl. Instr. and Meth. 198, 125-130. Benninghoven A. (eds), Wil ey. Chichester. 131-135. Vertes A. (1990). Lase r desorption of large Williams P, Nelson RW. (1990). Time offlight mass molecules: mechanisms and models. In: Methods and spectro metry of DNA by pulsed laser ablation of frozen mechanisms for producing ions from large molecules. aqueous solution s. In : Pro ceed ings of the 38th ASMS Standing KG , Ens W. (eds), Plenum. New York. Conference on Mass Spectrometry and Allied Topics. Vertes A, Bal azs L, Gijbels R. (1990a). Matrix­ Tu cson, Arizona. June 3- 8. Caprioli RM. (ed), 22-23. assisted laser desorption of peptides in transmission ge­ Willi ams P, Sundqvi st B. (1987) Mechanism of sp ut ­ ometry. Rapid Comm. Mass Spectrom. 1, 263-266. tering of lar ge biomolecules by imp act of highly ion izing Vertes A, Bal azs L, Gijbels R. (1991). Expansion particles. Phys. Rev. Let ters~ . 1031-1034 . cooling in matrix assisted laser desorption ( a hydro­ Zhao S, Somayajula KV , Sharkey AG , Hercules DM. dynamics study). In: Pro ceed ing s of the 39th ASMS (1990). Laser desorption mass spectrometry of protein Conference on Mass Spectrometry and Allied Topi cs. mixture s. In : Pro ceeding s of the 38th ASMS Confer­ Nashville, Tennessee. May 19- 24. Pap er in press. ence on Mass Sp ectromet ry and Allied Topics. Tucson , Vertes A, DeWolf M, Juh asz P, Gijbels R. (1989a). Arizona. Jun e 3- 8. Caprioli RM. (ed), 154-155. Threshold condition s of plasma ign ition in laser ioni za­ Za.re RN, Hahn JH , Zenobi R. (1988). Mass spec­ tion mass spectrometry of solids. Anal. Chem. 61, 1029- tromet ry of molecular adsorbates using laser desorption 1035. / laser multiphoton ionizat ion. Bull. Chem. Soc. Jpn . Vertes A, Gijb els R, Ad ams F. ( 1990b ). Diagnost ics 61, 87-92. and modeling of plasma processes in ion sources. Mass Zar e RN , Levi ne RD . (1987). Mechanism for bond­ Spectrom. Rev. fr, 71-113. select ive processes in lase r desorption. Chem. Phys . Vertes A, Gijbels R, Levine RD. (1990c). Homo ­ Lette rs 136, 593-599. geneous bottleneck model of matrix-assisted ultraviolet Zenobi R, Hahn JH , Zar e RN. (1988). Surfac e tem­ laser desorption of large molecules. Rapid Comm. Mass peratu re meas ur ement of dielectric mat erials heate d by Sp ectro m . 1, 228-233. pulsed las er radiation . Chem. Phys. Lette rs 150, 361- Vertes A, Juha sz P, Bal azs L, Gijbels R. ( 1989b). 365. Targ et heating , pla sma formation , and expansion pro­ cesses during laser ioni zatio n. Microbeam Analysis 1989, Discuss ion with Reviewers 273-276. M. Karas and U. Bahr: Major limitations are given for Vertes A, Juhasz P, De Wolf M, Gijbels R. (1988a). all the models if direct conclusions for the matrix laser The role of energy deposition proc esses in the under­ desorption-ioni zat ion (LDI) proc ess are to be drawn. standing of laser mi croprobe analysis mechanisms . Scan­ These are mainly : All models in principal deal only with ning Microsc. 1988; Il:1853-1877. the process of setting free neutral molecules. With re­ Vertes A, Juha sz P, DeWolf M, Gijbels R. (1989c). gard to MS - which has to rely on ions - - this is only Hydrodynamic mod eling of laser plasma ionization pro­ one, though important part of the coin. Equally impor­ cesses. Int. J. Mas s Spectrom. Ion Proc esses 94, 63-85. tant is the ionization process, it gives the prerequisite Vertes A, Juhasz P, Gijb els R. (1989d). Ion- for mass spectrometric detection. And with respect to molecule reaction rates in laser-generated plumes. Fre­ th e matrix-LDI process this is still not the whole story, senius' Z. Anal. Chem. 334, 682.

327 A. Vertes and R. Gijbels as the formation of a uniform solid solution of the an­ bottleneck model we hoped to demonstrate that 10 ns ir­ alyte molecules within the crysta lline matrix is another radiation time is short enough to avoid sul,stantia l heat­ decisive prerequisite for a successful matrix-LDI. ing of the guest molecules during their desorption period P. Williams: You fail to distinguish between volati liza­ (F ig. 2. in Vertes et al. , Hl90c). We also showed that a tion and ionization. Thus matrices which don't produce 50 ns laser pulse results in somewhat wanner desorbing protein mass spectra may st ill efficient ly ab late and ej ect molecules but may lead to extens ive therma l degradation neutral protein molecules. Ionization is a separate topic, in the solid phase (Fig. 3. of the same article). In the and using mass spectra l resu lt s alone to draw conclusions light of more recent results demonstrating that even 200 about t he process of molecule ejection can be misleading. ns Er-YAG lase r pulses produce intact ions ofl arge guest Authors: Vve fully agree with both comments. Ind eed, molecules (Overberg et al., 1990) the predicted 50 ns up­ ionization is a separate and very imp ortant topic . The per limit seems short , indeed. We shall note, however, reason we did not deal with this problem is the alm ost that the comb in ation of an energy transfer bottleneck in comp lete lack of reliable experiments and theoret ical the solid phase and expansion cooling in the gas phase models addressing the ionization step separate ly. We may st ill expla in the success of longer pulses. even tr ied to emphas ize the limit ed scope of this tutor ial P. Williams: Recent results indicate that MALD works in the title speak ing only about laser desorption. To the best at low irr adiances where only a fraction of the ma­ best of our knowledge there are, however, a few st ud ­ trix molecules in the irradi ated volume absorb a photon. ies reporting on laser ionization mechanisms (Heinen, These may then dissociate, wh ich is an endotherm ic pro­ 1981 , van der Peyl et al., 1982b, Chiarelli and Gro ss cess, and the surrounding matrix molecules may act to 1987 and Viswanadham et al., 1988). Pr elimin ary calcu­ shield the analyte by absorbing and dissipating the resid­ lat ions of ion-molecule reaction rates in laser gene ra ted ual energy of the dissociation products. In such a case plumes (Vertes et al., 1989d, Vertes 1990) and of the is the bottleneck model even necessary? effect of symmetrica l charge transfer processes are also Authors: Th e bottleneck mode l may not be necessary, ava ilabl e (Vert es et al., 1988b) in the lit erature. but it seems feasible. We are not conv in ced, however, M. Karas and U. Bahr : Th e discussion of the polymer about the feasibility of the scenario suggested by the re­ 6 ab lation was - the knowledge of the referees - neve r viewer. Taking the typical threshold irradiance as 5 x 10 to 2 4 exp licitly done within the "DIET -mode l". In any case W / cm and the typical absorpt ion coefficient as 4 x 10 most of the ea rli er considerations seem to be obso lete cm - i (nicotinic acid) the number density of the absorbed now, exper iments and discussion in the meantime hav e 266 nm UV photons in a r=l0 ns laser pulse is about 21 3 gone much furth er, many contradictions have ruled raI 0 /hv = 3x 10 cm- . This value is almost half of the too 21 3 out the ea rly model [especially the finding that the ma­ number density of solid nicotinic acid: 7x10 cm- . In jority of the ablated material is sti ll polymer ( ca. 2500 other words , almost every second nicotinic acid molecule absorbs a laser photon in contrast to the hypothesis of Da) rather than molecular). Now the process is described the reviewer. rather as a macroscopic one (shock waves, subsurface boiling and volcano-like ejection of material). Authors: Polym er ab lation is an intensively studied area and involves different mechanisms depending on the laser and on the nature of the polymer. For exan1p le, - in cont rast to the statement of the reviewer - in a quite recent report on ultraviolet laser ablation of polyimide films we find support for almost complete fragmenta­ tion: "Th e ablation pro cess must involve many photons per monomer unit to account for the production of pre­ dominantly sma ll ( < 4 atoms) products and the ejection of these fragments at supersonic velocities." (Srini vasan et al ., 1987). Furthermore, we are not interested in the validit y of the DIET-model for the polymer ab lation sit­ uation but suggest this mechanism for further conside r­ at ion with respect to MALD stud ies .

P. \Villi ams: How fast sho uld "fast heating" really be to allow eject ion to prevail over fragmentation? If intramo lecular relaxation occurs on a picosecond timescale, then, even with a bottleneck, are irradi at ion times 4-5 orders of magnitude longer (10 - 50 ns) really shor t enoug h to preclude inte rm olecu lar energy transfer? Auth ors: In the first presentation of our homo geneous

328