Study of ionization process of matrix molecules in matrix-assisted laser desorption ionization

Chemical Physics 419 (2013) 37–43

Contents lists available at SciVerse ScienceDirect

Chemical Physics journal homepage: www.elsevier.com/locate/chemphys

Study of ionization process of matrix molecules in matrix-assisted laser desorption ionization Kazumasa Murakami, Asami Sato, Kenro Hashimoto, Tatsuya Fujino ⇑ Department of Chemistry, Graduate School of Science and Engineering, Tokyo Metropolitan University, 1-1 Minami-Osawa, Hachioji-shi, Tokyo 192-0397, Japan

a r t i c l e

i n f o

Article history: Available online 20 December 2012 Keywords: Intramolecular proton transfer Matrix-assisted laser desorption ionization Mass spectrometry Polar analyte

a b s t r a c t Proton transfer and adduction reaction of matrix molecules in matrix-assisted laser desorption ionization were studied. By using 2,4,6-trihydroxyacetophenone (THAP), 2,5-dihydroxybenzoic acid (DHBA), and their related compounds in which the position of a hydroxyl group is different, it was clarified that a hydroxyl group forming an intramolecular hydrogen bond is related to the ionization of matrix molecules. Intramolecular proton transfer in the electronic excited state of the matrix and subsequent proton adduction from a surrounding solvent to the charge-separated matrix are the initial steps for the ionization of matrix molecules. Nanosecond pump–probe NIR–UV mass spectrometry confirmed that the existence of analyte molecules having large dipole moment in their structures is necessary for the stabilization of [matrix + H]+ in the electronic ground state. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction Matrix-assisted laser desorption ionization (MALDI) is one of the useful ‘‘soft ionization’’ methods as it does not decompose analyte molecules during the ionization process. MALDI combined with time-of-flight (TOF) mass spectrometry is widely used as a powerful tool to study biological macromolecules, such as DNA, proteins, and peptides, because it enables observation of analyte ions in terms of their molecular weights [1–3]. MALDI consists of two important processes: ionization and desorption. Since the discovery of MALDI, several studies aimed at fully understanding the ionization process have been carried out. In the photochemical ionization model proposed by Ehring et al., the multi-photon ionization of matrix molecules and the production of [m]+ are considered to be the initial process [4]. Karas et al. proposed a cluster ionization mechanism in which higher clusters of analyte are desorbed during laser irradiation. The higher clusters are decomposed in the gas phase to produce analyte ions [5–7]. This model is similar to the mechanism in electrospray ionization (ESI). Meanwhile, Chang et al. reported a pseudo proton transfer process that occurred between a matrix and an analyte during crystallization. In the pseudo proton transfer process, a proton in the matrix is dominantly shared with the basic site of the analyte [8]. Although the three models offered new knowledge of the ionization process in MALDI, they did not explain why typical MALDI matrices, such as CHCA (a-cyano-4-hydroxycinnamic acid), THAP (2,4,6-trihydroxy-

⇑ Corresponding author. Fax: +81 42 677 2525. E-mail address: [email protected] (T. Fujino). 0301-0104/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chemphys.2012.12.014

acetophenone), and DHBA (2,5-dihydroxybenzoic acid), are suitable for many measurements and widely used. It is necessary therefore to understand the reaction mechanism after the electronic excitation of matrix molecules. Recently, Wan and co-workers argued initial reactions of typical MALDI matrices based on their chemical properties. They concluded that the ionization of matrix molecules may occur via multiple pathways including photoionization ([m]+) in the electronic excited state and thermal ionization in the ground state, depending on their chemical properties such as absorption cross section and excited state lifetime [9]. In addition to the ionization process, the mechanism of the desorption process has remained unclear. In this regard, we recently proposed a possible desorption process based on exciton migration in organic crystals [10]. In this work, we tried to clarify the ionization process in MALDI, focusing on the reaction mechanism after the electronic excitation and their molecular structures of matrix molecules. By using THAP, DHBA, and their related compounds as MALDI matrices, we found that the existence of a hydroxyl group (ortho-OH) in the matrix structure, which typically formed an intramolecular hydrogen bond with a carboxyl group, was important to produce ions of both protonated matrix and protonated analyte. The intramolecular proton transfer from ortho-OH to carboxyl group (CO) in the electronic excited state and the subsequent proton adduction to orthoO (C–O) from a solvent were the initial process of protonation in MALDI. The protonated matrix was stabilized in the electronic ground state by nonvolatile organic compounds having large dipole moment, and this stabilization was confirmed by nanosecond time-resolved mass spectrometry with an NIR pump and a UV probe.

K. Murakami et al. / Chemical Physics 419 (2013) 37–43

We first examined the excitation conditions to clarify whether the cationic or the charge-neutral matrix was the reactant. An excitation laser pulse was typically focused to 0.2 mm diameter. In the case of 266 nm excitation, photon numbers of the excitation pulse were estimated to be 1.07  1020, 2.14  1020, 3.00  1020, and 4.28  1020 photons/m2 when pulse energies of 2.5, 5, 7, and 10 lJ were used, respectively. For the calculation of excitation rate, we used the absorption coefficient of matrix in solution (THAP in acetonitrile, 6120 mol1 dm3 cm1) at 266 nm because we could not determine the absorption coefficient of matrix in the crystals prepared for MALDI-MS measurements. Considering the time profile of the used laser, the excitation rate was estimated to be 2.5  108, 5.0  108, 7.0  108, and 9.3  108 s1 for the 2.5, 5, 7, and 10 lJ excitation, respectively. Therefore, it was calculated that THAP could be photoexcited every 4, 2, 1.4, and 1.1 ns. However, as the pulse width of the used laser was 1 ns, the multi-photon excitation of THAP during one pulse was unlikely at these pulse energies. The absorption coefficient of THAP at 337 nm was less than that at 266 nm, and was  2130 mol1 dm3 cm1. The pulse width of N2 laser in a commercial MALDI system was 2.5 ns. Therefore, the excitation velocities were lower than those in the case of 266 nm excitation, indicating that the multi-photon excitation of THAP was more unlikely than the 266 nm excitation. In the mass spectrometry in the following sections, we treated only the results measured in the above-mentioned one-photon excitation condition. We did not argue the results obtainable in the high-power excitation condition by which multi-photon excitation of THAP could occur. Therefore, we considered that the ionization process in MALDI was a chemical process rather than ionization by multi-photon excitation (photoinization) in the low-power excitation condition.

(a) THAP only (negative)

15 10 5.0

503 [3THAP-H]-

0.0 (b) THAP only (positive) 15 10 5.0 0.0 (c) SubP/THAP (negative)

5.0

167 [THAP-H]-

0.0 (d) SubP/THAP (positive) 169 [THAP+H]+

5.0

1349 [SubP+H]+

0.0 0

3.2. Mass spectra measured in positive and negative ion modes We carried out mass spectrometry of matrix and analyte/matrix systems in the negative and positive ion modes. Mass spectra were

335 [2THAP-H]-

359 [2THAP+Na]+

3.1. Excitation conditions

191 [THAP+Na]+ 337 [2THAP+H]+

3. Results and discussion

167 [THAP-H]-

THAP (m/z = 168) and other compounds used as the matrix were purchased (Wako Chemical). Model peptides, substance P (SubP) and angiotensin II (Ang), were used as analyte molecules. To measure mass spectra of SubP with THAP, for example, THAP was dissolved in a mixture of acetonitrile and water (7:3 in volume), and a 59.5 mM solution was obtained. SubP was dissolved in a mixture of acetonitrile and water (7:3 in volume), and a 74.2 lM solution was obtained. One microliter each of the matrix solution and the analyte solution was pipetted onto a stainless steel plate (matrix 59.5 nmol, analyte 74.2 pmol), left in air for a few minutes to evaporate the solvent, and then analyzed with a commercial (MALDI micro MX, Waters; 337 nm, 2.5 ns) or homemade mass spectrometer as described previously[11]. Briefly, matrix-analyte crystals on the stainless steel plate were put inside a vacuum chamber set at 5  105 Pa. The fundamental laser output from a Nd:YAG laser (Rayture Systems, 1064 nm, 10 Hz, 1 ns) and the fourth harmonics generation (266 nm) were used as excitation light. For nanosecond pump–probe experiments, NIR pump (1064 nm) and time-delayed UV probe (266 nm) pulses were collinearly focused onto the sample. The produced ions were accelerated with an electric field of 1.7 kV/cm and detected with a linear TOF tube and an MCP detector.

measured with the commercial MALDI system with the excitation power slightly above the threshold for observing ions (337 nm, 5 lJ). Fig. 1(a) shows the mass spectrum of matrix (THAP only) in the negative ion mode; a prominent peak of [THAP-H] was clearly observed at m/z = 167. Strong peaks of the THAP dimer, [2THAP-H], and the THAP trimer, [3THAP-H], were observed as well. In the positive ion mode shown in Fig. 1(b), the peak of such dimers as [2THAP + Na]+ was also observed. However, peaks from the THAP higher clusters almost disappeared when the analyte molecules were mixed. Fig. 1(c) and (d) show the mass spectra of SubP/THAP obtained in the negative and positive ion modes, respectively. In both spectra, peaks from the THAP higher clusters were not observed. In addition, the decrease of the peak intensity of the THAP monomer was clearly observed in both spectra. Particularly for the mass spectrum observed in the negative ion mode, the peak intensity of [THAP-H] was remarkably decreased and no longer comparable to that observed for the [THAP + H]+ monomer in the positive ion mode. In this series of experiments, the conditions, such as the amounts of analyte and THAP and the excitation power, were the same. Therefore, it could be easily understood that the difference in the mass spectra observed without and with an analyte (SubP) was a result of the difference in the ionization mechanism. From the spectra obtained without the analyte molecules in which (1) intense peaks of the matrix higher clusters were observed, and (2) peak intensities in negative and positive ion modes were almost comparable, it became clear that the ionization in this case was completed only in THAPs; THAP solid dissociated into [THAP + H]+ and [THAP-H] ions with the irradiation of excitation laser pulses. We named this ionization process the ‘‘dimer mechanism’’ for further discussion. This result is practically the same as that reported in the recent paper [9], although we

169 [THAP+H]+

2. Experimental

intensity (a.u.)

38

200

400

600

800

1000

1200

1400

mass (m/z) Fig. 1. Mass spectra of THAP solid in (a) negative and (b) positive ion modes. MALDI mass spectra of SubP/THAP in (c) negative and (d) positive ion modes.

39

K. Murakami et al. / Chemical Physics 419 (2013) 37–43

are not sure that above dissociation occurs whether in the electronically excited state or thermally in the ground state. In the case with an analyte (SubP), however, a different ionization process that was caused by chemical reactions was suggested. To understand the detailed ionization mechanism for the analyte/matrix system, we carried out the following experiments.

3.3. Hydroxyl group with intramolecular hydrogen bonding Fig. 2 shows the MALDI mass spectra of a model peptide (SubP) by using THAP, DHBA, and their related compounds as matrices. Mass spectra were measured with the commercial MALDI system (337 nm, 7 lJ). When the typical organic matrix, THAP, was used, peaks from the protonated matrix (m/z = 169) as well as the protonated analyte (m/z = 1349) were clearly observed, as shown in Fig. 2(a). In order to identify the functional group related to the proton adduction reaction, monohydroxyacetophenones in which the position of the OH group was different were used as matrices, and MALDI mass measurements of SubP were carried out. Peaks from the protonated matrix and the protonated analyte were also observed at m/z = 137 and 1349, respectively, when orthohydroxyacetophenone (oHAP) was used (Fig. 2(b)); however, the peak intensities were remarkably reduced compared to those of THAP. The peak from SubP could not be observed when metaHAP (Fig. 2(c)) or para-HAP (Fig. 2(d)) were used. The peak from each matrix could not be observed in both spectra. Protonation of the analyte could be achieved by proton transfer from the protonated matrix, and the mass peak of the protonated analyte would not be observed unless ions of the protonated matrix were produced. From these results, it was clarified that the ortho-OH group was important to observe ions of matrix and analyte molecules. We also carried out mass spectrometry of SubP by using DHBA and its related compounds as matrices. Fig. 2(e) shows the mass spectrum of SubP obtained with DHBA. Peaks from the protonated matrix (m/z = 155) as well as the protonated analyte (m/z = 1349) were observed. When monohydroxybenzoic acids were used, ortho-HBA gave peaks of the protonated matrix and the protonated analyte, as shown in Fig. 2(f). However, those peaks were not observed when meta-HBA (Fig. 2(g)) or para-HBA (Fig. 2(h)) was used, as in the case of acetophenones. Taken together, the results

showed that the existence of an ortho-OH group was an important requirement for use as a MALDI matrix. In a previous study of quantum chemical calculations, it was reported that the orthoOH groups in THAP and DHBA formed intramolecular hydrogen bonds with carboxyl groups [12]. Therefore, it was considered that the ortho-OH group forming an intramolecular hydrogen bond was related to the proton transfer reaction in MALDI. In organic molecules having both proton donor and acceptor groups in close proximity, intramolecular hydrogen bonds (IHBs) are generally formed in the electronic ground state. It is widely known that these molecules induce the fast rearrangement of molecular structures, referred to as the electronically excited state intramolecular proton transfer (ESIPT), which is due to the intramolecular charge redistribution and the deformation of the aromatic ring by photon absorption [13–15]. In a previous study of methyl salicylate in the gas phase, it was shown that methyl salicylate that forms IHB between a carboxyl group and an adjacent ortho-OH group in the ground state is photoexcited to the electronic excited state (A⁄) and then undergoes ultrafast change from A⁄ to product tautomer T⁄, which is the proton-transferred form from the ortho-OH group to the carboxyl group[13]. Felker et al., through fluorescence lifetime measurements, reported that the activation energy for T⁄ A⁄ transition should be higher than 1300 cm1 [16]. The T⁄ tautomer relaxed to T in the ground state and then underwent barrier free transition to A. The potential-energy diagram of methyl salicylate is illustrated in Fig. 3(a). The intramolecular proton transfer reaction has been reported for many molecules having IHBs in their structures. Therefore, it is easily assumed that MALDI matrices undergo such ESIPT reactions through photoexcitation as major MALDI matrices, such as THAP, DHBA, CHCA, and sinapinic acid (SA), have intramolecular hydrogen bonds.

3.4. Protonation of matrix molecules Next, the origin of the proton to form the protonated matrix was studied. A THAP matrix was recrystallized three times from deuterated methanol (MeOD) to exchange OH groups with OD. The thus-treated matrix, THAP(D), and the analyte, SubP, were dissolved in a mixture of MeOD and D2O (7:3 in volume) separately,

15 10

(e) SubP/DHBA

191[THAP+Na]+

30 1349[SubP+H]+ HO

5

20

OH (THAP)

O

1349[SubP+H]+

155[DHBA+H]+

169[THAP+H]+

(a) SubP/THAP

HO (DHBA) O

10

OH

intensity (a.u.)

OH OH

0 (b) SubP/oHAP 137[oHAP+H]+

5

(oHAP)

O

(c) SubP/mHAP

5

O OH OH

(g) SubP/mHBA OH

0 (d) SubP/pHAP

OH (pHAP)

O

5

0

400

800

mass (m/z)

1200

OH (pHBA)

(h) SubP/pHBA O OH

0

0

(mHBA)

O

(mHAP)

0

1349[SubP+H]+

HO

(oHBA)

139[oHBA+H]+

OH

O

5

(f) SubP/oHBA

5 0

0 5

0 1349[SubP+H]+

HO

0

400

800

1200

mass (m/z)

Fig. 2. MALDI mass spectrum of a model peptide, (a) Substance P (SubP) by using THAP as matrix. MALDI mass spectra of SubP with (b) oHAP, (c) mHAP, and (d) pHAP. MALDI mass spectrum of (e) SubP by using DHBA as matrix. MALDI mass spectra of SubP with (f) oHBA, (g) mHBA, and (h) pHBA.

40

K. Murakami et al. / Chemical Physics 419 (2013) 37–43

(b) stabilized matrix molecules

(a) methyl salicylate HO

(c) NIR-UV pump-probe mass measurement

O

O

HO OCH3

ESIPT

A*, S1

,

T*, S1

energy

energy

S1

,

energy

OCH3

,

S1

S1

,

,

UV 266 nm

S0

T, S0

NIR 1064 nm

A, S0

S0

S0

displacement

displacement

,

S0

displacement

Fig. 3. Schematic potential diagram of (a) methyl salicylate, (b) stabilization of charge-separated matrix in the electronic ground state, and (c) excitation diagram of NIR–UV pump–probe mass measurement in the SubP/THAP system.

O

intensity (a.u.)

-O

OH

D(1)O

OH

D(1)+O

UV OH

D+

OH

D(2)O D(1)+O

many hydrogen atoms in SubP were already exchanged to deuterium atoms in the solvent prior to the ionization and desorption process.

OH

OH

3.5. Stabilization of protonated matrix

171 172

10 170

173 1360

5 169

174 175

0

166

170

174

178 1350

1355

1360

1365

1370

mass (m/z) Fig. 4. MALDI mass spectrum of SubP. Matrix THAP was recrystallized from deuterated methanol to exchange oOH to oOD and hydrated H2O to D2O. Crystals for MALDI measurement were prepared from a mixture of MeOD and D2O (7:3 in volume).

and one microliter each of the matrix solution and the analyte solution was pipetted onto a stainless steel plate (matrix 59.5 nmol, analyte 74.2 pmol) to evaporate the solvent. Then, MALDI mass spectrometry was carried out using the same apparatus and conditions as those for Figs. 1 and 2, and the results are shown in Fig. 4. A prominent peak was observed for the matrix at m/ z = 171, whereas matrix-related peaks were also observed at m/ z = 170, 172, 173, 174. However, only the very small peak was observed at m/z = 169. Judging from the facts that molecular weight of THAP is 168 Da and that THAP molecules are surrounded by MeOD and D2O (there are no proton sources around THAP), the origin of the peak was considered to be [THAP(D) + D]+, indicating that the matrix received a deuteron from the surroundings. Considering that the strongest peak for matrix molecules was at m/ z = 171, it can be assumed that only one OH group in THAP, probably ortho-OH whose surrounding environment should differ from two other free OH groups, was mainly exchanged by recrystallization with MeOD to produce THAP(D). The deuteron transfer reaction at the matrix molecule after the photoexcitation is illustrated in the inset of Fig. 4, by assuming ortho-OH in THAP was substituted by recrystallization. On photoexcitation, a deuteron at the ortho-OD group (D(1) in the inset) underwent intramolecular deuteron transfer to the carboxyl group. Then, a deuteron (D(2) in the inset) from solvent adducted to the chargeseparated matrix molecule for stabilization. For the analyte, the peak at m/z = 1360 was observed with the highest intensity, whereas the exact mass of SubP was 1348 Da. This meant that

As mentioned above, molecules having IHB in their structures underwent ESIPT. After the T⁄ tautomer was formed, it relaxed to T in the ground state and then underwent barrier less transition to A (Fig. 3(a)). In conventional MALDI, matrices give ions of [matrix + H]+ in the ground state. This means that the T tautomer can exist even in the ground state and a proton in the surroundings can adduct to the charge-separated matrix molecule. Therefore, it is necessary to understand what can stabilize T and the protonated matrix in the electronic ground state, as illustrated in Fig. 3(b). First, we studied whether water molecules used as solvent stabilized the ionization of matrix molecules. 3.5.1. Water molecules The tautomer with the intramolecular proton transfer (IPT) often appears by its growing ionic character in the ground state by mixing with the charge separated S1 state whose energy is lowered by solvation. The geometries and energetics of [THAP + H]+ and THAP are shown in Fig.5, which were calculated at the B3LYP/631 + G(d, p) level by using Gaussian09 program [17]. The optimized structure of THAP in the S1 state is also included, which corresponds to A⁄ in Fig. 3(a). It was computed by the time-dependent density functional theory (TD B3LYP/6-31 + G(d, p)). In the most stable form of [THAP + H]+ (PA), the IPT occurs in THAP by switching the direction of the hydrogen bond, and the ortho-O is protonated, in harmony with the experimental detection of both [THAP + H]+ and [oHAP + H]+. These results naturally indicate that the proton source is one of the surrounding water molecules, which donates its OH bond to the ortho-O. The most stable structures of hydrated THAP, THAP(H2O)n (n = 1–4), are also presented in Fig.5, while other high-energy isomers are in the supporting materials. In the two most stable 1:1 complexes (1A and 1B), a water molecule is bound to oxygen in ortho-OH or in carbonyl group from different sides, respectively. The binding energies of the water to THAP were 5.9–6.1 kcal/mol. The second water molecule tends to bind with the water molecule in 1A or 1B via hydrogen bond. A water dimer is bound to THAP in 2A–2D, respectively. Similarly, the most stable forms for n = 3 and 4, respectively, can be regarded as the structures where a hydrogen-bonded water trimer or tetramer is bound to THAP. Taking into account the structure of [THAP + H]+, the isomers with the la-

41

K. Murakami et al. / Chemical Physics 419 (2013) 37–43

PA 0.0

0D 32.0

2B 5.1

3C 10.4

PB 3.7

0A* S1(ππ*) 74.7

2C 7.2

3D 10.9

PC 6.3

0A 0.0

1A 0.0

2D 8.2

4A 0.0

0B 15.7

1B 0.2

3A 0.0

4B 6.0

0C 18.7

2A 0.0

3B 10.3

4C 6.4

Fig. 5. Optimized structures of [THAP + H]+, and THAP(H2O)n (n = 0–4). Isomers of the protonated species are labeled by PA–PC. The values under those structures are relative energies from PA in kcal/mol. The labels of the form nA–nD such as 0A and 0B are used for hydrated THAP with n water molecules. 0A⁄ is the structure of free THAP in S1 (pp⁄) state. The relative energies (kcal/mol) of the isomers with the same n are also given under each structure.

bels nA and nB (n = 1–4), in which the oxygen in the ortho-OH group is hydrated, are the most probable candidates for the initial geometries of the protonation reaction of THAP. Those nA and nB isomers can be regarded as the hydrated structures of A in Fig. 3(a), because the IHB is kept in all of them. Although we made many attempts to locate the minima with IPT form, all optimization calculations finally reached the known structures without the proton transfer. We also scanned the potential energy curve along the IPT coordinate for 3A and 3B, and found a single minimum at 3A and 3B, respectively. In addition, we searched for the charge-separated structures by optimizing the [THAP + H]+(H2O)n 1[OH] -type geometries, but all calculations converged to the minima with the usual IHB. These results altogether suggest that the ground-state potential surface remains of Fig. 3(a) type with up to several water molecules, where there is no local minimum at the tautomer. Another possible reason for the stabilization of the tautomer with the IPT and the protonated species is the presence of the polarized molecule in the vicinity of the THAP. In fact, we have found the formation of [THAP + H]+ with only one water molecule on the polar model zeolite surface[18]. No discovery of the IPT isomer by the calculations suggests the coexisting polar molecule, namely, the analyte itself, plays an essential role to produce [THAP + H]+. If this hypothesis is true, the [THAP + H]+ should be scarcely observed with the non-polar analyte, which leads us to the further study described below.

3.5.2. Nonvolatile compounds We recognize that MALDI is used in many research fields as nonvolatile compounds such as biological samples, including proteins and peptides, can be ionized by MALDI. Nonvolatile compounds hardly evaporate because of their large dipole moment. In fact, the dipole moment for the model peptide, SubP, was roughly estimated to be 4.1 Debye by the B3LYP/6-31G method. Therefore, we next considered whether polar analytes stabilized the formation of [matrix + H]+ in the ground state. To examine the effect of analyte on the matrix, nanosecond time-resolved pump–probe MALDI mass spectrometry was performed. MALDI mass measurements were carried out by using our home-made mass spectrometer. The fundamental output of the nanosecond Nd:YAG laser (1064 nm; NIR; 15 lJ) was used as pump, which was expected to generate the tautomer (T) form in the ground state. Then, the fourth harmonic generation (266 nm; 5 lJ) was used as the probe for measurement. NIR pulses were not resonant to the electronic absorption band of the matrix (THAP). However, the NIR pulses were considered to be resonant to the combination band, including the 2nd overtone of OH stretching vibration. We also expected that the excitation energy of NIR pulses (9398.5 cm1) would be higher than the energy difference between enol (A) and keto (T); therefore, T A conversion in the ground state was expected (Fig. 3(c)). The nanosecond time-resolved mass spectra observed for the peptide, Ang, are shown in Fig. 6(a) and (b). In Fig. 6(a) at the delay time of 2 ns, the UV pulse

42

K. Murakami et al. / Chemical Physics 419 (2013) 37–43

(a) Ang/THAP -2ns 2 1 0 (b) Ang/THAP 4ns 169 [THAP+H]+

intensity (a.u.)

3 2 1

1047 [Ang+H]+

0 (c) Tet/THAP -2ns 1 0 (d) Tet/THAP 4ns 1 0 0

200

400

600

800

1000

1200

mass (m/z) (e) 6

[Ang+H]+

in Ang/THAP

peak area (a.u.)

[THAP+H]+ in Ang/THAP 5

[Tet+H]+

4

[THAP+H]+ in Tet/THAP

in Tet/THAP

3 2 1 0 -4

-2

0

2

4

6

delay (ns) Fig. 6. Nanosecond time-resolved mass spectra of Ang/THAP at the delay times of (a) 2 ns and (b) 4 ns. Nanosecond time-resolved mass spectra of Tet/THAP at the delay times of (c) 2 ns and (d) 4 ns. (e) Delay time dependence of the peak area of protonated Ang (d) and THAP (N) in Ang/THAP measurement, and protonated Tet (s) and THAP (4) in Tet/THAP measurement.

reached the sample before the NIR pulse. At the delay time of 4 ns, the increase of the peak intensity of the protonated analyte became clear (Fig. 6(b)). To confirm the stabilization of protonated matrix molecules in the ground state, we also carried out nanosecond NIR–UV pump–probe mass measurements with a non-polar analyte, tetracosane (Tet; C24H50). Fig. 6(c) and (d) show the results of time-resolved mass spectra of Tet at 2 and 4 ns, respectively. In both spectra, only matrix-related peaks were observed. The delay time dependences of the peak area of protonated Ang (d) and THAP (N) in Ang/THAP measurement, and protonated Tet (s) and THAP (4) in Tet/THAP measurement, are plotted in Fig. 6(e). For the time dependence of protonated Tet, the peak intensity at m/z = 339 where the peak of protonated Tet was expected was plotted, because we could not see any signals of protonated Tet in the mass spectra. The solid curves in Fig. 6(e) are the results of fitting analysis. It was observed for the Ang/THAP system that the peak area of the protonated analyte increased with increasing delay time. Furthermore, an enhancement of the peak of protonated THAP was clearly observed by irradiation of NIR pulses. In this series of experiments, the molar amount of Ang was the same; therefore, the peak intensity of the protonated analyte would be proportional to the product of (1) the probability of proton adduction from the protonated matrix to the analyte and (2)

the amount of protonated matrix. Of course, it could be considered that the proton adduction probability would be the same for all experiments. Therefore, the increase of the peak intensity of protonated Ang would mean an increase in the amount of protonated matrix by the irradiation of NIR pulses. As was mentioned above, the NIR pulses were not resonant to the electronic absorption of the matrix. Considering that the protonated analyte was produced from the protonated matrix, this result clearly showed that the production of [matrix + H]+ in the electronic ground state was enhanced by the NIR pulses. On the other hand, the peak of protonated Tet was not observed and no enhancement was shown by the irradiation of NIR pulses. The important finding in Fig. 6(e) was the peak intensity of protonated THAP in the Tet/THAP system. The peak of protonated THAP was observed even in the negative delay time; however, no enhancement was shown by the irradiation of NIR pulses. Therefore, it was clarified that the non-polar analyte, Tet, could not stabilize [matrix + H]+ in the electronic ground state, and NIR pulses could not generate [matrix + H]+ as the charge-separated THAP underwent barrierless relaxation to its initial structure. From our results, it became clear that the peak of protonated THAP in the Tet/THAP system in the negative delay time came from the ‘‘dimer mechanism’’ mentioned in Section 3.1, in which a dimer or a trimer in solid THAP dissociated to produce protonated and deprotonated THAP at the same time, and it does not correlate with the protonation of analyte molecules. In this regard, it was understood in MALDI that the production of the protonated matrix involved the combination of the ‘‘dimer mechanism’’ and the above ‘‘analyte support mechanism,’’ whereas the production of the protonated analyte involved the ‘‘analyte support mechanism.’’ From the fitting analysis of the Ang/THAP system, the rise time that indicated the occurrence of the proton adduction and desorption process was determined to be 1.5 ns. Considering that the cross correlation of NIR and UV pulses is  1.5 ns, such event occurs in the early time period rather than in this time scale, although this reaction time can be changed by altering the proton affinity of the analyte and the kind of matrix. Finally, we make a comment on total mechanism of MALDI combining with our recent reports on desorption process [10]. Our proposal at the present stage is as follows: (1) Matrix molecules first absorb photons from excitation laser since analytes have typically no absorption coefficient in UV–vis region. (2) Generated exciton moves in matrix crystals and then localized in the matrix molecule which is adjacent to an analyte, since such matrix molecule is stabilized by the large dipole moment of the analyte and works as a trapping site of generated exciton. We think localization of exciton at the matrix molecule adjacent to the analyte occurs efficiently if the distance between excitation position and the stabilized matrix is within exciton migration distance, namely ‘‘sweet spot’’ called in conventional MALDI. In the case of tetracene-doped anthracene crystals, we observed extremely strong fluorescence from the S1 tetracene and mass peak of tetracene ions, although amount of tetracene is 1/100 of anthracene and anthracene molecules first absorb photons from the excitation laser [10]. Of course, the excitons can be trapped at matrix molecules in defect sites in general. Matrix crystals in matrix-analyte mixture are far from single crystals; therefore, it contains many defect sites. Matrix molecules in which excitons are localized then dissociate to produce [matrix + H]+ and [matrix-H] as we described as ‘‘dimer mechanism’’, or deactivate with producing excess vibrational energy in the electronic ground state. (3) In the matrix molecule which is adjacent to the analyte and electronically excited by exciton localization, ESIPT reaction occurs and then the matrix molecule receives a proton from the surrounding solvent with the help of analyte (‘‘analyte support mechanism’’). (4) Because of strong anharmonisity of OH stretching vibration of [matrix + H]+,

K. Murakami et al. / Chemical Physics 419 (2013) 37–43

a proton easily detaches from [matrix + H]+ and adducts to the analyte to produce [analyte + H]+ ion. (5) Finally, protonated analyte desorbs from matrix-analyte crystals by vibrational excess energy produced by electronic deactivation of photoexcited matrix molecules and dissipated into the lattice phonon modes in the matrix crystals as we proved in the previous paper [10]. Enough excess energy stored in the matrix-analyte system leads to multi-quantum excitation of dissociative modes between matrix and analyte to break the bond and to obtain the kinetic energy to desorb the analyte from the matrix-analyte crystals. Acknowledgment T. F. and K. H. acknowledge a Grant-in-Aid for Scientific Research on Priority Area (477) from MEXT. T. F. acknowledges a Grant-in-Aid for Scientific Research (C) (No. 24550030) from JSPS. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chemphys. 2012.12.014.

43

References [1] K. Tanaka, H. Waki, Y. Ido, S. Akita, Y. Yoshida, T. Yoshida, Rapid Commun. Mass Spectrom. 2 (1988) 151. [2] F. Hillenkamp, M. Karas, R.C. Beavis, B.T. Chait, Anal. Chem. 63 (1991) 1193A. [3] F. Hillenkamp, M. Karas, Int. J. Mass. Spectrom. 200 (2000) 71. [4] H. Ehring, M. Karas, F. Hillenkamp, Org. Mass Spectrom. 27 (1992) 427. [5] M. Karas, M. Gluckmann, J. Schafer, J. Mass Spectrom. 35 (2000) 1. [6] R. Kruger, A. Pfenninger, I. Fournier, M. Gluckmann, M. Karas, Anal. Chem. 73 (2001) 5812. [7] M. Karas, R. Kruger, Chem. Rev. 103 (2003) 427. [8] W.C. Chang, L.C.L. Huang, Y.-S. Wang, W.-P. Peng, H.C. Chang, N.Y. Hsu, W.B. Yang, C.H. Chen, Anal. Chim. Acta 582 (2007) 1. [9] Y.-H. Lai, C.-C. Wang, C.W. Chen, B.-H. Liu, S.H. Lin, Y.T. Lee, Y.-S. Wang, J. Phys. Chem. B 116 (2012) 9635. [10] Y. Minegishi, D. Morimoto, J. Matsumoto, H. Shiromaru, K. Hashimoto, T. Fujino, J. Phys. Chem. C 116 (2012) 3059. [11] S. Yamaguchi, T. Fujita, T. Fujino, T. Korenaga, Anal. Aci. 24 (2008) 1497. [12] J. Zhang, T.-K. Ha, R. Knochenmuss, R. Zenobi, J. Phys. Chem. A 106 (2002) 6615. [13] A. Douhal, F. Lahmani, A.H. Zewail, Chem. Phys. 207 (1996) 477. [14] S. Nagaoka, U. Nagashima, Chem. Phys. 206 (1996) 353. [15] S. Nagaoka, A. Nakamura, U. Nagashima, J. Photochem. Photobiol. A 154 (2002) 23. [16] P.M. Felker, W.R. Lambert, A.H. Zewail, J. Chem. Phys. 77 (1982) 1603. [17] M.J. Frisch et al., Gaussian 09, Revision B.01, Gaussian Inc., Wallingford CT, 2010. [18] Y. Komori, H. Shima, T. Fujino, J.N. Kondo, K. Hashimoto, T. Korenaga, J. Phys. Chem. C 114 (2010) 1593.