Free and immobilized matrix molecules: impairing ionization by quenching secondary ion formation in laser desorption MS

Research Article

Received: 18 March 2011 Revised: 24 June 2011 Accepted: 27 June 2011 Published online in Wiley Online Library

(wileyonlinelibrary.com) DOI 10.1002/jms.1965

884

Free and immobilized matrix molecules:impairing ionization by quenching secondaryion formation in laser desorption MSAndreas Schinkovitz,a*,† Ghislain Tsague Kenfack,b Eric Levillain,b

Marylène Dias,b Jean-Jacques Helesbeux,a Séverine Derbré,a

Denis Séraphina and Pascal Richommea

Within the last 25years, matrix-assisted laser desorption ionization (MALDI) has become a powerful analytical tool in massspectrometry (MS). While the method has been successfully applied to characterize large organic molecules such as proteins,sugars and polymers, its utilization for small molecules (≤ 600Da) is significantly impaired by the coformation of matrix ions.Reducing or eliminating matrix-related signals has been subject of many studies. Some of which propose the enhancement ofso-called matrix suppression effects, while others suggest the replacement of matrix molecules by materials such asmicroporous silicon. Alternatively, the immobilization of matrix molecules by utilizing them as self-assembled monolayers(SAMs) has been discussed. In continuation of this research, the current manuscript focuses on the elucidation of ionformation processes occurring on the surface of light absorbing SAMs. Ion yields obtained by free and immobilized matrixmolecules as well as those generated by matrix-free gold film-assisted laser desorption ionization (GF-LDI) were compared.Experiments showed that the formation of strong analyte signals essentially required the presence of free matrix molecules,while the immobilization of the latter severely impaired ionization. The observed effect inversely correlated with the surfacecoverage of SAMs determined by cyclic voltammetry (CV). Based on these findings, the MS signal generated on lightabsorbing SAMs could be used supplementary to CV for determining the surface coverage of light absorbing SAMs. Copyright© 2011 John Wiley & Sons, Ltd.

Supporting information can be found in the online version of this article.

Keywords: MALDI; Ionization; Matrix Immobilization; Self-assembled Monolayers; Laser Desorption Ionization MS; DIAMS

* Correspondence to: Andreas Schinkovitz, University of Vienna, Department ofPharmacognosy, Althanstraße 14, A-1090, Vienna, Austria.E-mail: [email protected]

a Université d’Angers, EA 921 SONAS, IFR 149 QUASAV, 16 bd Daviers, 49100,Angers, France

b Université d’Angers, MOLTECH-Anjou, UMR-6200 CNRS, 2 bd Lavoisier, 49045,Angers, France

† Current address: University of Vienna, Department of Pharmacognosy,Althanstraße 14, A-1090, Vienna, Austria.

INTRODUCTION

Ion formation in matrix-assisted laser desorption mass spec-trometry (MALDI-MS) has been subject of intense research, andvarious mechanisms of ionization have been proposed.[1,2] Inprinciple, the latter can be divided into two groups: Processespredominantly taking place immediately after laser irradiation(0.005–0.02ms, primary ion formation) followed by those occurringat a later stage in the MALDI plume (0.02–5ms, secondary ionformation).[1] It is important to mention that there is no clearseparation but a gradient transition between these two events.Nevertheless, for both processes, ionization is based on theinteraction of matrix and analyte molecules after laser irradiation,inducing the formation of their respective ions. Most commonlyutilized matrices such as 2,5-dihydroxy benzoic acid (DHB) or a-cyano-4-hydroxycinnamic acid (CHCA), but also many othersdisplay their molecular ions in the mass range of 100 to 600m/z.Consequently, the detection of small molecules is severelyimpaired by the coformation of matrix ions in that region.Successful attempts to reduce or eliminate matrix noise by

lowering the amount of applied laser energy and modifying theanalyte to matrix mixing ratio have been reported.[1,3,4] Accordingto these studies, matrix-to-analyte ratios of ≤100 together with lowlevels of applied laser energy induce so-called matrix suppressioneffects (MSE). Unfortunately, the latter are essentially accompaniedby analyte suppression effects, as both mechanisms share a

J. Mass. Spectrom. 2011, 46, 884–890

common underlying principle.[1] This is particularly of relevancewhen several analytes are present in the same sample. Conse-quently, MSE may yield cleaner spectra, but signal intensities ofcertain analytes might be impaired or even entirely suppressed.

Alternatively, ‘surface-assisted laser desorption ionization’(SALDI) MS and related methods represent an entirely matrix-free approach to laser desorption ionization.[5–7] SALDI replacesmatrix molecules by nanostructured materials such as germani-um, carbon and particularly silicon.[5–7] The method has beensuccessfully applied for analyzing a large variety of compoundsincluding benzodiazepines, various xenobiotics, peptides andproteins.[7] It has been shown that MS results highly depend onthe specific structure of the utilized nanostructured surfaces,[5,7]

and further on the way how these surfaces are produced.[5]

Copyright © 2011 John Wiley & Sons, Ltd.

Energy exchange study

Substance P for example could be ionized on microporoussilicon, which was fabricated in SF6 and water but not if thesurface was produced in air.[5] Nevertheless, under optimizedconditions, SALDI may facilitate the detection of selectedanalytes in the low femtomole range.[5]

Over the last years, another matrix-free method namelydesorption/ionization on self-assembled monolayers (DIAMS)has been developed.[8–10] Unlike SALDI, DIAMS replaces classicmatrix molecules with light absorbing self-assembled monolayers(SAMs), which are covalently bound to a gold surface. This is avery distinct difference to classic MALDI, where matrix moleculesare never covalently attached to the sample plate. In order tofacilitate optimal energy exchange in DIAMS, the utilized SAMsare equipped with a chromophore exhibiting an absorptionmaximum in the range of the emission wavelength of a standardnitrogen laser (337nm) commonly utilized in laser desorptionMS. In principle, the light absorbing monolayer is supposed toact like a classic matrix by taking up laser energy andsubsequently transferring part of the latter to the sample inorder to aid ionization processes. Unlike classic matrices, SAMsdo not get desorbed from the gold surface, and consequently nomatrix ions should be present in the plume yielding spectrasolely exhibiting analyte signals.

Hitherto DIAMS analysis were successfully applied for thedetection of fatty acids, glycerides and salicylic acid.[9,10] One reportdirectly compared signal intensities of triglycerides obtained byDIAMS and by gold film-assisted desorption ionization (GF-LDI).[9]

Out of eight observed [M-H]+ ions, five displayed stronger signalsin GF-LDI, while two showed better results in DIAMS. For onesample, nodifference between the twomethods couldbe detected.It shall be mentioned that this study solely compared mean valuesof analyte signal intensities without evaluating whether theobserved differences were statistically significant or not. In con-tinuation of this work, the current manuscript takes a closer look onion yields generated by MALDI, DIAMS and GF-LDI. A key questionof this study is, whether matrix molecules behave differentlydepending on their being free or covalently bond to a gold surface.Further, provided there is a distinct difference, could the latter bequantified and are there any supplemental applications for lightabsorbing SAMs in combination with laser desorption MS?

Figure 1. Chemical structures.

J. Mass. Spectrom. 2011, 46, 884–890 Copyright © 2011 John

The realization of the project demanded careful planning andpreceding considerations. In the first place, it was necessary toselect a molecule, which could be used as a classic MALDI matrixbut also produced stable SAMs. Furthermore, as many MALDImatrices often show specific preferences to certain compounds,potential analytes needed to show sufficient interaction with theselected matrix/SAMs molecule.

All SAMs, which were so far used for DIAMS experiments shareone common structural feature: the bithiophenic chromophore,exhibiting an absorptionmaximumaround 340nm.[8–11] This is veryclose to the emission wavelength of a standard nitrogen laser(l=337nm) commonly used in MALDI-TOF instruments. Thebithiophenic moiety also allows studying the formation of SAMsas well as their stability by cyclic voltammetry (CV). Hitherto, fromany reported bithiophenic SAM, 8-(1-(3-((5’-(methylthio)-[2,2’-bithiophen]-5-yl)thio)propyl)-1H-1,2,3-triazol-4-yl)octane-1-thiol(1) (Fig. 1) exhibited the highest surface coverage on goldsurfaces.[11] Another preceding study revealed that a precursor of1 showed highly selective ionization of alkaloids when utilized asclassic MALDI matrix. Particularly, claviculine: (12aS)-(9CI)-2,3,12,12a-tetrahydro-9-methoxy-1-methyl-1H-[1]benzoxepino[2,3,4-ij]isoquinoline-6,8-diol (CAS: 87035-67-4) (2), quinidine: 6’-Methoxy-(9S)-cinchonan-9-ol (CAS: 56-54-2) (3) and yohimbine: 17a-hydroxy-yohimban-16a-carboxylic acid methyl ester (CAS: 146-48-5)(4) displayedmost intensemolecular ions. On the other hand, littleto no interaction was observed with compounds of differentchemical origin. Based on these findings, matrix molecule 1,together with analytes 2–4, represented ideal candidates for thestudy and were used for subsequent experiments.

MATERIAL AND METHODS

Preparation of gold plates and electrodes

Gold plates and electrodes utilized for DIAMS and electrochem-ical experiments were prepared by physical vapor depositionaccording to a previously described procedure.[9,12,13] Both,electrodes and DIAMS plates, consisted of a borosilicate glassplate, which was covered by a layer of chrome (5nm) and gold(35nm). Plates and electrodes were stored in an oxygen-free

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jms

885

A. Schinkovitz et al.

886

atmosphere {dry, oxygen-free argon [(≤1ppm)]}. Experimentswere carried out under normal atmospheric conditions.

Synthesis, monolayer formation and stability evaluation

The synthesis of 1[11] is sketched in the scheme 1.Monolayer formation was carried out using freshly prepared gold

plates (DIAMS) and electrodes, which were dipped for 30min into a2mmolar dichloromentane (DCM) solution of1. The incubationwasperformed at room temperature. Plates and electrodes were thenrinsed several times with HPLC-grade DCM and immediately usedfor DIAMS or CV experiments. For quality control, the surfacecoverage of immobilized 1 on gold electrodes was monitored byCV. This experiment was performed at the beginning and the endof any DIAMS experiment.

Cyclic voltammetry

Electrochemical experiments were carried out on a BioLogic SP 150potentiostat (Claix, France) using a three-electrode cell equippedwith a platinum-plate counter electrode and a silver quasi-referenceelectrode. Experiments were performed in a 0.1 molar solution oftetrabutylammoniumhexafluorophosphate prepared inDCM. Fig. 5Bdisplays a cyclic voltammogram of 1 obtained at a scan rate of 500mV/s. Surface coverage (Γ ) was calculated according to: Γ ¼ Q

J�S. Inthis equation, Q indicates the area under the curve of the oxidationpeak (mC), S represents the surface of the electrode (0.2cm2) and Jdisplays the Faraday constant (96485C/mol). The absolute potentialwas verified by a ferrocene solution, and previously observed valueswere recalculated according to the ferrocene/ferricinium couple.

Mass spectrometry and instrument settings

All experiments were carried out on a Bruker Biflex III time offlight (TOF) mass spectrometer (Bruker Daltonik, Bremen,Germany) equipped with a 337nm pulsed nitrogen laser (modelVSL-337i, Laser Sciences Inc., Boston, MA). Spectra were acquired

Scheme 1. (I) 2,2 Bithiophene+n-butyllithium (nBuLi)+sulfur (�78�C, 1h).[10

iodide, DMF/MeOH;[10,13,14] (III): Tosylated azides Ts–(CH2)3–N3+CsOH in DMdec-1-yne, sodium ascorbate and CuSO4 in tert-butanol/H2O; (V): Alkaline hy

Copyright © 2011 Johnwileyonlinelibrary.com/journal/jms

within a mass range of 20–2000m/z. Acceleration voltage was setto 19kV, pulse ion extraction was 200ns and laser frequency was5ns. If not stated differently, laser energy was set to 30% (30.8mJ).

MS-sample preparation

Plates utilized for the DIAMS and GF-LDI experiments were gluedonto a stainless steel MALDI sample plate (Bruker Daltonik,Bremen, Germany), using commercial cyanoacrylate-based super-glue. Stock solutions of analytes (2–4) were prepared inMeOH at aconcentration of 2.58mmol/l. Subsequently, stock solutions werediluted 1:9 with equimolar mixtures of either MeOH/H2O, MeOH/DCM or DCM. These working solutions were carefully depositedonto DIAMS and GF-LDI plates (0.7ml per sample spot).

MALDI: The DCM solution of 1, previously used for incubatingDIAMS plates and electrodes, was further utilized as matrixsolution for MALDI experiments. To avoid any covalent bindingof thiol groups to the metal surface, experiments were carriedout on a standard MALDI steel plate instead of a gold plate. Forthat, 0.7ml of the sample-matrix solution was spotted onto theMALDI steel plate. It shall be mentioned that the formation ofthiol-based SAMs on steel has been previously reported, butspecial preconditioning of the steel surface is essentiallyrequired.[18,19] The latter includes mechanical and chemicalpolishing (EtOH, acetone, nitric acid) as well as glow dischargegas plasma treatment and an incubation time of 48h. As no suchpreconditioning measures were performed and the samplesolution dried within approximately 1s, the presence ofconsiderable S–Fe bindings was highly unlikely. A summary ofthe sample preparation is displayed in Fig. 2.

Data collection and processing

Before launching analytical MS experiment, a sample-free area ofeach DIAMS plate was exposed to increasing amounts of laserlight until degradation of the monolayer was observed. This wasgenerally indicated by the appearance of a signal at 221m/z. For

,13]; (II): Methyl iodide (1 equivalent) and caesium hydroxide (CsOH) methylF, H2O; (IV): Azide-Alkyne cycloaddition [11,15–17] using 10-(methylsulfinyl)drolysis using CsOH in DMF/MeOH.[11]

J. Mass. Spectrom. 2011, 46, 884–890Wiley & Sons, Ltd.

Energy exchange study

subsequent DIAMs experiments, the laser energy was thenadjusted below the threshold where first signs of degradationcould be detected.

For each sample, 30 spectra, acquired from 30 differentlocations, evenly distributed on the sample deposition area, wererecorded and averaged. Each spectrum consisted of thecollective signal of 30 laser shots. Data were processed usingBruker Daltonics Flex Analysis 2.0 software. Total signalintensities were summed and averaged. Statistical calculationswere performed using SigmaPlot11 (Systat Software Inc.Chicago, USA). Results from MALDI, DIAMS and GF-LDI werecompared in pairs by t-test evaluation. In case the Shapiro–Wilknormality test failed, the Mann–Whitney rank sum test wasperformed instead. Standard deviations (STD) are either dis-played as error bars in figures or expressed as percent of themean in the text.

RESULTS AND DISCUSSION

A general observation previously made for DIAMS was confirmedat an early stage of the experiments. Any compound that couldbe detected by DIAMS could also be observed by GF-LDI,[9]

indicating that self-ionization of analytes is essentially involvedinto both processes.

As previously outlined, preceding experiments have identifieda precursor molecule of 1, as an excellent MALDI matrix foralkaloids. Despite its more complex structure, these propertieswere preserved in compound 1. The latter strongly promoted theionization of compound 2–4, when being utilized as a classicMALDI matrix. Consequently, a similar picture was expectedwhen 1 was immobilized on a gold surface, but, to our surprise,the exact opposite was the case. Ionization of compound 2–4

Figure 2. Experimental setup.

J. Mass. Spectrom. 2011, 46, 884–890 Copyright © 2011 John

887

was severely impaired, yielding MS signals weaker than thoseobtained by GF-LDI (Figs 3 and 4). In comparison to analyte 2 and3, 4 showed the most prominent decrease in signal intensity, butalso for 2 and 3, the decay was very pronounced. With oneexception (compound 2: MALDI vs. GF-LDI), any observeddifference between MALDI and GF-LDI, MALDI and DIAMS aswell as DIAMS and GF-LDI was significant (p: <0.001–0.005).These results suggest that the immobilization of 1 redirected theenergy exchange between 1 and the analytes, inducing a lightprotective and ionization diminishing effect.

At that point, it shall be mentioned that observed ion yields formatrices and analytes may significantly vary depending onwhether MALDI experiments are performed on steel or goldsurfaces. The issue has been discussed controversially as quitedifferent results were reported for DHB.[22,23] One studysuggested that DHB should exhibit a stronger surface-relatedsignal enhancement on steel than it does on gold.[23] This isexplained by the observation that the LUMO-derived orbitals ofDHB fall below the Fermi level if the sample is deposited on steeland irradiated with a 355nm laser. Consequently, two (steel),instead of three photon ionization (gold), should occur, yieldinghigher amounts of DHB ions. Additionally, increased reflectivityon steel (50%) compared to gold (35%) should further promotehigher ion yields on steel. On the other hand, a different studyinvestigating photo electron emission and cross captureionization on both metals concluded, that higher positive ionyields should be observed on gold rather than on steel.[22] It hasbeen observed that photo electron emission,[22,24] and conse-quently negative ion yields of DHB were greatly decreased (byabout a factor of 2) when going from a stainless steel to a goldsurface.[22] This is explained by a shift of the maxima of theelectron energy distribution from about 1eV on steel to 0.55eVon gold. For DHB, this is a shift away from the maximum ofenergy cross section located around 1eV. Subsequently, thepresence of a reduced number of negative ions should limit theneutralization of positive ions in the plume, yielding strongeranalyte signals in the positive mode. These considerations wereconfirmed by experimental data showing a twofold increased [M+H]+ signal of bradikinine acquired on gold in comparison tosteel.[22] It shall be mentioned that any of the outlined processesexclusively apply to very thin sample layers [22,23] and may varywith matrix choice.[23] In accordance to the latter, another studydid not find any significant difference in analyte and CHCA ionyields generated on gold and steel surfaces.[21]

Compound 1 is a relatively new matrix molecule, and itsspecific characterization is still in progress. At the current stage,no studies investigating its specific behaviour on different metalsurfaces have been performed. Nevertheless, steel, and not gold,was purposely chosen for MALDI experiments in the currentstudy. As mentioned earlier, the main objective was to minimizeany covalent binding between free thiols and the sample carrierwhich could be best accomplished by using an ‘untreated’MALDI steel plate.

Independent from the potential effect of different metalsurfaces on ion yields observed in MALDI experiments, thereduced ionization capacity of DIAMS in comparison to GF-LDIrepresents a separate observation and was further studied. In asubsequent experiment, the monolayer present on DIAMS plateswas artificially degraded by ultrasound treatment. Four DIAMSplates, exhibiting 100, 81, 39 and 0% of the initial SAMs coverage,were then loaded with an equal amount of analyte 2 andanalyzed by laser desorption MS. As expected, the observed ion

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jms

Figure 4. MS spectra of 4 obtained by MALDI, GF-LDI and DIAMS. Theionization of compound 4 yielded [M-H]+: 353.036, which, on first site,appears quite unusual for a positively charged ion. On the other hand,the observation of [M-H]+ has been reported for several MALDIexperiments[2,20,21], and its formation is associated with photochemicalinteractions between matrix and analyte molecules.[2,20]

Figure 3. Comparison of ion yields: Claviculine: Group 1 (MALDI), 4 (GF-LDI), 7 (DIAMS); Quinidine: Group 2 (MALDI), 5 (GF-LDI), 8 (DIAMS); Yohimbine:Group 3 (MALDI), 6 (GF-LDI), 9 (DIAMS); Except for group 1 to 4, all observed differences were statistically significant. Calculated p values from t-tests:Group 1 to 4 p: 0.065, 1 to 7 p: 0.002, 4 to 7 p: <0.001; group 2 to 5 p: 0.002, 2 to 8 p: 0.002, group 5 to 8 p: 0.005; group 3 to 6 p: <0.001, 3 to 9 p:<0.001, group 6 to 9 p: <0.001.

A. Schinkovitz et al.

888

yield of 2 constantly increased in inverse correlation to thesurface coverage of immobilized 1 (Fig. 5).This is particularly worth mentioning as laser desorption MS

may provide a supplemental method to CV for the generalassessment of the surface coverage of SAMs carrying an UV

Figure 5. A: MS signal response of 2 in correlation to SAMs with varying su

Copyright © 2011 Johnwileyonlinelibrary.com/journal/jms

chromophore. Undoubtedly, CV is a powerful analytical tech-nique, but it requires the presence of at least one redox center(e.g. sulfur) in the SAM’s structure. Further, every CV cycle puts a tollon the monolayer, potentially breaking the covalent Au–S bond.This is particularly true for low scanning rates (≤100mV/s).Consequently, CV as such may impair the quality of SAMs andnot reflect their original condition. Figure 6 displays the inversecorrelation of CV and MS signals. Although the two curves showsome differences, their general appearance is quite similar,allowing a good general estimation of SAMs’ coverage by eithermethod. Therefore current data present a good starting pointfor further studies of the phenomenon and potential technicalapplications.

Results from the degradation experiment as well as thoseobserved for 2–4 by GF-LDI and DIAMS raise further questionsabout the energy transfer on immobilized 1. Contrary to classicMALDI, the impact of laser irradiation is no longer reflected byMS signals of the analytes or 1. In other words: How can thesuppression of ionization be explained and to where is the laserenergy redirected?

A possible explanation is provided by the concept of light-to-current conversion. The latter is associated with light absorbingSAMs, which show the same general build-up as immobilized 1.Any reported light-to-current converting SAM consists of achromophore, an aliphatic chain of varying length, and an anchorgroup attaching the molecule to a metal surface.[25–27] Accordingto these reports, light-to-current conversion represents a

rface coverage. B:Cyclic voltammogram of artificially degraded SAMs of 1.

J. Mass. Spectrom. 2011, 46, 884–890Wiley & Sons, Ltd.

Figure 7. MS signals of compound 5 and 6 obtained by GF-LDI andDIAMS at a laser energy of 30% (30.8mJ) (5) and 20% (22.8mJ) (6). MALDIdata not displayed in the graph were as follows: Ion yields of compound5 exceed the detector cut-off limit at 38 250 units. The mean value forcompound 6 was 11777.5 (STD: 29.37%).

Figure 6. Correlation between CV and MS signals.

Energy exchange study

889

general observation without being associated to a specificchromophore or metal surface. Various – and actually verydifferent – molecules attached to diverse metal surfaces haveshown the same effect. Although no experimental data for light-to-current conversion was collected within the current study, theconcept presents a most reasonable explanation for the loss ofion yields and shall be followed up.

Analyzing the current results from a different perspective, anadditional explanation is provided by basic principles of MALDI-MS. As noted, earlier ion formation in laser desorption MS can bedivided into two processes, those occurring within the matrix–analyte mixture (primary ionization) and those taking place in theMALDI plume (secondary ionization).[1,2] In the state of immobi-lization, 1 can exclusively contribute to primary ionization, whilesecondary processes entirely depend on the self-ionization ofanalytes. Primary ionization predominantly occurs within the first0.005–0.02ms after laser irradiation, while secondary ionizationtakes place within 0.02–5ms. The comparison of the time scale ofboth events reveals a proportion of 1:250, which is very un-favourable for DIAMS. Furthermore, even certain mechanisms ofprimary ionization such as cluster formation or processesproposed in the polar fluid and the pneumatic assistance modeldo also require free matrix molecules.[1]

In the reviewing process the interesting question, whether theabsence or presence of a triazole ring in the monolayer mighthave significantly impacted the observed results, was raised.Therefore, an additional experiment utilizing 10-[[5’-(methylthio)[2,2’-bithiophen]-5-yl]thio]-decan-1-thiol (CAS: 928822-73-5), thepioneering molecule of DIAMS, was performed. In the process,the molecule was analyzed exactly the same way as compound1, showing basically the same results as reported for the latter.These findings indicate that the triazole ring had no substantialimpact on observed ion yields. Like for 1, highest ion yields wereobserved in the MALDI experiment, followed by GF-LDI andeventually DIAMS. Differences between MALDI versus GF-LDI, GF-LDI versus DIAMS and MALDI versus DIAMS were all statisticallysignificant (p≤0.001 to 0.002). A graphic summary of theseresults is provided in the supplemental section of the paper.

Apart from ionization, desorption represents an equallyimportant process for the detection of ions in MALDI or DIAMS.In the current context, the question whether solely ionization oralso desorption is impaired by immobilized 1 is evident.Therefore, a final experiment compared the MS signal responsesobtained for precharged molecules such as tetrabutylammoniumhexafluorophosphate (5) (CAS: 3109-63-5) and berberine: 5,6-Dihydro-9,10-dimethoxybenzo[g]-1,3-benzodioxolo[5,6-a]quinoli-zinium chloride (6) (CAS: 2086–831). Results are displayed in

J. Mass. Spectrom. 2011, 46, 884–890 Copyright © 2011 John

Fig. 7. Contrary to uncharged 2–4, ion yields observed for 5 and6 did not exhibit a significant difference between DIAMS and GF-LDI. Completing the picture, these findings show that ionizationbut not desorption seems to be impaired by the immobilizationof 1, and further that desorption seems not to require free matrixmolecules.

Eventually, presented data raise the question whether SAMscould still aid laser desorption ionization. Most recently, Nayakand Knapp have reported the detection of peptides in the lowpicomole range by laser desorption MS using methyl terminatedSAMs on nanostructured gold surfaces.[28] Undoubtedly, nanos-tructures are capable of promoting ionization, and SAMs maycertainly aid this process. In the cited report, the effect is linkedto the increased surface area of nanostructured gold due to thepresence of methyl terminated SAMs. Considering that SAMswithout chromophore were used, the work of Nayak and Knappdoes not conflict with results presented in the current study.Furthermore, it can be seen as a confirmation of the essentialrole of the chromophore for the light protective, energyredirecting effect of immobilized 1.

CONCLUSION

The current manuscript has compared ion yields of six differentcompounds generated by methods of MALDI, GF-LDI and DIAMS.Analytes 1–6 could be successfully detected by any of the threemethods, but observed signal intensities significantly differedand were highly depending on the nature of the compound andthe applied method. Contrary to the free matrix molecule,immobilized 1 exhibited a light protective, ionization diminishingeffect on compounds 2–4 yielding lower signal intensities thanobtained by GF-LDI. Furthermore, an inverse correlation betweenthe MS signal response of analyte 2 and the surface coverage ofSAM 1 was observed, suggesting that laser desorption MS couldbe used supplementary to CV for the quality assessment of lightabsorbing SAMs. In addition, the utilization of the latter asselective light protecting agents could be an interesting field offuture research. Compared to GF-LDI, ion yields originating fromprecharged molecules were not impaired by the presence ofimmobilized 1. This proposes that predominantly ionization, andnot desorption, was impacted by the energy redirecting effectpresent on light absorbing SAMs.

It has to be kept in mind that ionization processes are verycomplex, and their investigation on SAMs is still at an early stage.

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jms

A. Schinkovitz et al.

890

The current manuscript has elucidated some aspects of theseprocesses, but certainly further studies are needed to gaindeeper inside into the matter. Alternative to substituting matrixmolecules, light absorbing SAMs could be used for laserdesorption MS in a completely different way. FunctionalizedSAMs capable of actively releasing certain molecules (e.g. matrixmolecules) upon direct laser irradiation may offer mostinteresting alternatives to classic MALDI. That way, the numberof free matrix molecules could be sufficient to aid ionization, butstill low enough to limit, the formation of matrix noise. Suchphotocleavable SAMs do already exist,[29] but have not beenapplied in MS and therefore represent a most interesting subjectfor future research.

Supporting information

Supporting information can be found in the online version of thisarticle.

REFERENCES[1] R. Knochenmuss. Ion formation mechnisms in UV-MALDI. Analyst

2006, 131, 966.[2] R. Zenobi, R. Knochenmuss. Ion Formation in Maldi Mass

Spectrometry. Mass Spectrom. Rev. 1998, 17, 337.[3] R. Knochenmuss, F. Dubois, M. J. Dale, R. Zenobi. The Matrix

Suppression Effect and Ionization Mechanisms in Matrix-assistedLaser Desorption Ionization. Rapid Commun. Mass Spectrom. 1996,10, 871.

[4] G. McCombie, R. Knochenmuss. Small-Molecule MALDI Using theMatrix Suppression Effect To Reduce or Eliminate Matrix BackgroundInterferences. Anal. Chem. 2004, 76, 4990.

[5] Y. Chen, A. Vertes. Adjustable Fragmentation in Laser Desorption/Ionization from Laser-Induced Silicon Microcolumn Arrays. Anal.Chem. 2006, 78, 5835.

[6] K. P. Law. Laser desorption/ionization mass spectrometry onnanostructured semiconductor substrates: DIOSTM and Quick-MassTM. Int. J. Mass. Spectrom. 2010, 290, 72.

[7] K. P. Law, J. R. Larkin. Recent advances in SALDI-MS techniques andtheir chemical and bioanalytical applications. Anal. Bioanal. Chem.2011, 399, 2597.

[8] M. Bounichou, O. Aleveque, M. Dias, L. Sanguinet, E. Levillian, D.Rondeau. Self-assembled monolayer-assisted mass spectrometry. J.Mater. Chem. 2009, 19, 8032.

[9] M. Bounichou, L. Sanguinet, K. Elouarzaki, O. Aleveque, M. Dias, E.Levillian, D. Rondeau. Evaluation of new matrix-free laser/desorptionmethod through statistic studies: comparison of the DIAMS(desorption/ionization on self assembled monolayer surfact) methoswith MALDI and TGFA-LDI techniques. J. Mass Spectrom. 2008, 43,1618.

[10] L. Sanguinet, O. Aleveque, P. Blanchard, M. Dias, E. Levillian, D.Rondeau. Desorption/ionization on self-assembled monolayer sur-faces (DIAMS). J. Mass Spectrom. 2006, 41, 830.

[11] G. T. Kenfack, A. Schinkovitz, S. Babua, K. Elouarzaki, M. Dias, S.Derbré, J. Helesbeux, P. Richomme, E. Levillain, D. Séraphin.Triazolobithiophene light absorbing self-assembled monolayers:synthesis and mass spectrometry application. New J. Chem. 2011,in review

Copyright © 2011 Johnwileyonlinelibrary.com/journal/jms

[12] O. Aleveque, in Immobilisation de dérivés aminoxyles. Relationsstructure/proprietés et structure/réactivé vis-a-vis de l’électrocata-lyse, Phd thesis: l’Ecole Doctorale 3MPL (ED 500): Matière, Molècules,Matériaux en Pays de la Loire (Universty of Angers, 2011).

[13] M. Bounichou, in: LA METHODE DIAMS :Desorption/Ionization onself-Assembled Monolayer Surface Une nouvelle technique dedésorption ionisation laser sans matrice pour la Spectrométrie deMasse, PhD thesis: l’Ecole Doctorale 3MPL (ED 500): Matière,Molècules, Matériaux en Pays de la Loire (Universtiy of Angers, 2010).

[14] P. Blanchard, B. Jousselme, P. Frère, R. J. 3- and 3,4-Bis(2-cyanoethylsulfanyl) thiophenes as building blocks for functionalizedthiophene-based p-conjugated systems. J. Org. Chem. 2002, 67,3961.

[15] F. Himo, T. Lovell, R. Hilgraf, V. Rostovtsev, L. Noodleman, K. B.Sharpless, V. V. Fokin. J. Am. Chem. Soc. 2005, 127, 210.

[16] V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B. Sharpless. Angew.Chem. Int. Ed. 2002, 41, 2596.

[17] C. W. Tornøe, C. Christensen, M. Meldal. J. Org. Chem. 2002, 67, 3057.[18] A. Mahapatro, D. M. Johnson, N. P. Devang, M. D. Feldman, A. A.

Ayon, C. M. Agrawal. Surface Modification of Functional Self-Assembled Monolayers on 316L Stainless Steel via Lipase Catalysis.Langmuir 2006, 22, 901.

[19] A. Mahapatro, D. M. Johnson, D. N. Patel, M. D. Feldman, A. A. Ayon,C. M. Agrawal. The use of alkanethiol self-assembled monolayers on316L stainless steel for coronary artery stent nanomedicineapplications: an oxidative and in vitro stability study. Nanomedicine2006, 3, 182.

[20] H. Ehring, M. Karas, F. Hillenkamp. Role of photoionization andphotochemistry in ionization processes of organic molecules andrelevance for matrix-assisted laser desorption/ionization massspectrometry. Org. Mass. Spectrom. 1992, 27, 427.

[21] A. Kassler, E. Pittennauer, N. Doerr, G. Allmaier. Electrosprayionization and atmospheric pressure matrix-assisted laser desorp-tion/ionization mass spectrometry of antioxidants aplied tolubricants. Rapid Commun. Mass Spectrom. 2009, 23, 3917.

[22] M. Dashtiev, V. Frankevich, R. Zenobi. Kinetic Energy of FreeElectrons Affects MALDI Positive Ion Yield via Capture Cross-Section.J. Phys. Chem. A 2006, 110, 926.

[23] G. McCombie, R. Knochenmuss. Enhanced MALDI ionizationEfficiency at the Metal-Matrix Interface: Practical and MechanisticConsequences of Sample Thickness and Preparation Method. J. Am.Soc. Mass Spectrom. 2006, 17, 737.

[24] V. Frankevich, J. Zhang, S. D. Friess, M. Dashtiev, R. Zenobi. Role ofElectrons in Laser Desorption/Ionization Mass Spectrometry. Anal.Chem. 2003, 75, 6063.

[25] C. J. Faulkner, S. Lees, P. N. Ciesielski, D. E. Cliffel, G. K. Jennings.Rapid Assembly of Photosystem I Monolayers on Gold Electrodes.Langmuir 2008, 24, 8409.

[26] Y. Li, Y. Li, H. Liu, S. Wang, N. Wang, J. Zhuang, X. Li, X. He, D. Zhu.Self-assembled monolayers of porphyrin-perylenetetracarboxylicdiimide-[60] fullerene on indium tin oxide electrodes: enhancementof light harvesting in the visible light region. Nanotechnology 2005,16, (9), 1899.

[27] Y. Li, Y. Liu, N. Wang, Y. Li, H. Liu, F. Lu, J. Zhuang, D. Zhu. Self-assembled monolayers of C60-perylenetetracarboxylic diimide-C60triad on indium tin oxide surface. Carbon 2005, 43, (9), 1968.

[28] R. Nayak, D. R. Knapp. Marix-Free LDI Mass Spectrometry PlatformUsing Patterned Nanostructured Gold Thin Films. Anal. Chem. 2010,82, (18), 7772.

[29] K. N. Plunkett, A. Mohraz, R. T. Haasch, J. A. Lewis, J. S. Moore. Light-Regulated Electrostatic Interactions in Colloidal Suspensions. J. Am.Chem. Soc. 2005, 127, (42), 14574.

J. Mass. Spectrom. 2011, 46, 884–890Wiley & Sons, Ltd.