Research article Received: 28 February 2014
Revised: 9 July 2014
Accepted: 29 July 2014
Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI 10.1002/jms.3452
Addition–elimination versus Tishchenko reaction in the gas phase Beata Wilenska,a† Pawel Swidera and Witold Danikiewicza* Gas phase reactions of the substituted phenide ions with methyl formate have been studied. It was found that the results of these reactions depend mainly on the basicity of the phenide ion, which is related to the presence of the electron-accepting or electrondonating substituents in the benzene ring. It was shown that the phenide ions substituted with electron-withdrawing groups react with methyl formate in the gas phase in a two-step reaction. The first step that proceeds according to the typical addition– elimination mechanism results in the formation of the anion of the respective benzaldehyde derivative with the negative charge located either in the aldehyde group (acyl anion) or in the benzene ring (phenide anion) in position ortho to an aldehyde moiety. In the second step, the preliminary-formed anion reacts with the second molecule of methyl formate yielding formally product of the second addition–elimination reaction. Theoretical calculations as well as collision induced dissociation spectra of the model compounds suggest that this reaction proceeds according to the Tishchenko reaction mechanism yielding the respective phthalide anion. According to our knowledge, this is the first example of the Tishchenko-type reaction in the gas phase. Copyright © 2014 John Wiley & Sons, Ltd. Keywords: ion-molecule reactions; phenide ions; addition-elimination reaction; Tishchenko reaction; DFT calculations
Introduction
J. Mass Spectrom. (2014), 49, 1247–1253
Experimental The aromatic carboxylic acids and 5-cyanophthalide were purchased from Aldrich and used without further purification. 6-Cyanophthalide
* Correspondence to: Witold Danikiewicz, Institute of Organic Chemistry, Polish Academy of Sciences, ul. Kasprzaka 44/52, Warsaw, Poland. E-mail: witold. [email protected] †
Present address: Faculty of Chemistry, Warsaw University, ul. Pasteura 1, 02-093, Warsaw, Poland
a Institute of Organic Chemistry, Polish Academy of Sciences, ul. Kasprzaka 44/52, Warsaw, Poland
Copyright © 2014 John Wiley & Sons, Ltd.
1247
The addition–elimination reaction to the carbonyl group and the Tishchenko reaction are very common and useful reactions in the organic chemistry. A typical addition–elimination reaction runs through two steps: at first, an attack of nucleophile to the carbonyl group of carboxylic acid derivatives occurs and then in the second step, an elimination of an anionic molecule takes place (Scheme 1). The first step of the addition–elimination reaction in condensed phase is reversible and the second one limits the reaction rate. The nucleophiles are generated in situ just before the process in the reaction of bases with appropriate precursors. Hence, the addition–elimination reaction is catalyzed by basic compounds, especially amines.[1] In the gas phase, the addition–elimination reaction has been studied extensively. Most of the researchers investigated reactions of inorganic anions such as hydroxyl anion with carbonyl compounds and focused on the formation of a tetrahedral transition state, similar to the condense phase.[2–5] There are also few reports about gas phase addition–elimination reaction of aromatic nucleophiles with carbonyl compounds. The Danikiewicz group investigated reactions of (halo)nitrophenide ions with alkyl acrylates, where the addition– elimination reaction was observed as one of the main processes.[6,7] In the gas phase, this reaction is supposed to proceed along the same mechanism as in the condense phase, but after the elimination step, a proton transfer reaction occurs. Hence, the final product is in an anionic form. Moreover, in contrast to the liquid phase, the nucleophiles are generated in the gas phase by simple decarboxylation of the appropriate carboxylic acid anions during collision induced dissociation (CID) in the electrospray ion source of mass spectrometer.[8] The carbanions formed in this process, with exception of simple alkyl anions, are stable and do not undergo further fragmentation. The Tishchenko reaction is also a catalytic process in the liquid phase and involves the disproportionation of two aldehyde molecules to generate an appropriate ester.[9] Traditional Tishchenko
reaction is carried out in the presence of a Lewis acid, but it is also possible to conduct the reaction using a nucleophilic catalyst (nucleophile-induced conversion through a Cannizzaro– Tishchenko-type mechanism) (Scheme 2). These two approaches to Tishchenko reaction have a common feature – a hydride shift from one aldehyde molecule to another is observed. This reaction can proceed also intramolecularly yielding the corresponding lactone.[10] So far, any variant of the gas phase Tishchenko reaction was not described in the literature. In our laboratory, the reactions of aromatic carbanions substituted with electron-withdrawing or electron-donating groups with methyl formate were investigated. During these experiments, the formation of the appropriate aldehyde anion as a product of gas phase addition–elimination reaction was observed. Interestingly, this aldehyde anion underwent further reaction with another molecule of methyl formate creating a new adduct, which can be transformed to the product of addition–elimination reaction or undergo intramolecular gas phase Tishchenko reaction yielding lactone. In this paper, we present the results of our mass spectrometric and computational study concerning these reactions.
B. Wilenska, P. Swider and W. Danikiewicz
Scheme 1. Mechanism of addition–elimination reaction to the carbonyl group of carboxylic acid derivatives.
H H
H H
Scheme 2. Mechanisms of the condensed phase Tishchenko (a) and Cannizzaro–Tishchenko (b) reactions.
was synthesized according to the procedure described in the literature.[11,12] Fresh stock solutions of carboxylic acids were prepared by dissolution in methanol (HPLC grade, Merck) and infused to the electrospray ion source at the rate of 10 μl/min using Harvard Apparatus Pump 11 syringe pump. All gas phase reactions were performed using an API 365 triple quadrupole mass spectrometer (Applied Biosystems) equipped with a TurboIonSpray source. All spectra were recorded in a negative ion mode standard electrospray ion source. Source parameters were as follows: capillary voltage, 4.5 kV; nebulizer gas, 12; and curtain gas, 8 (arbitrary units, N2 in both cases). Typical DP (Declustering Potential) values were between 20 and 60 V and FP (Focusing Potential) values between 180 and 300 V. To enable the use of different collision gasses, the collision gas inlet was modified to allow introduction of the selected collision gas independently from the curtain gas. For the CID experiments, nitrogen was used as the collision gas. The ion molecule reactions in the collision cell were studied using appropriate gaseous reagents. In the case of liquid reagents (methyl formate), a stream of nitrogen was flushed over the surface of the appropriate liquid in a small flask. The pressure of collision gas was set at 3 (arbitrary units). Collision cell parameters were set as follows: entrance potential – 10 V, exit potential 15 V. The collision energy during the reaction was set at 5 eV but during CID experiments was adjusted manually in the range of 5–25 eV to achieve the best results.
1248
Scheme 3. Investigated carbanions and their proton affinities (PA, kcal/mol) in the gas phase.
wileyonlinelibrary.com/journal/jms
Figure 1. Spectra recorded in the negative ion mode for the reaction of (a) 4-metoxyphenide anion, (b) 2,4-dichlorophenide anion, and (c) 4-cyanophenide anion with methyl formate in the gas phase. See Table 1 for explanation of A1 and A2 symbols.
Copyright © 2014 John Wiley & Sons, Ltd.
J. Mass Spectrom. (2014), 49, 1247–1253
Gas phase Tishchenko reaction All quantum chemical calculations described in this work were performed using the Gaussian 09 software package.[13] Starting geometries for geometry optimization were created using the GaussView 5.0 program. Reaction potential energy profiles and molecular properties were computed applying density functional theory (DFT) methods at the B3LYP/6-311+G(3df,2p)//B3LYP/631G(d) level. Proton affinities (PAs) of the anions at 298.15 K were calculated as the enthalpies of the reaction: AH = A + H+ using the equation 298 298 ¼ H298 PA ¼ ΔH298 r A þ HHþ HAH H298 ¼ E 0 þ Htherm where Htherm is the thermal correction to enthalpy The enthalpy of the proton at 298.15 K was set to 1.48 kcal/mol, i.e. 5/2RT.
Results and discussion During our study, we investigated aromatic carbanions with a variety of substituents, which were obtained in the reaction of decarboxylation of appropriate carboxylic acid anions. For comparison, one aliphatic carbanion, i.e. benzyl anion, has been added. On the basis of their PA in the gas phase, we decided to divide them into three groups: (1) strongly basic carbanions with PA value Table 1. Aromatic carbanions with their PA value (kcal/mol) and the results of their reaction with methyl formate in the gas phase Carbanion
p-MeOPh Ph p-ClPh p-CF3Ph o-CF3Ph p-CNPh o-CNPh PhCH2 2,4-diClPh
PA (kcal/mol) A1 403 401 394 390 387 385 384 381 381
[A1–MeOH]
A2
[A2–MeOH]
++ + + ++ +
+ ++ + + + ++
+ + + + ++ +
+ + ++ ++ +
A1 , primary adduct anion, [A1–MeOH] , anion of the methanol elimination product from the primary adduct, A2 , secondary adduct anion, [A2–MeOH] , anion of the methanol elimination product from the secondary adduct.
higher than 400 kcal/mol, (2) carbanions with moderate basicity (PA values in the range between 394 and 384 kcal/mol), and (3) carbanions with lower basicity (PA value below 384 kcal/mol), (Scheme 3). Most of the aromatic carbanions, except for the phenide anion, undergo the addition–elimination reaction in the gas phase. In the group of strongly basic carbanions only for the p-methoxyphenide ion in the reaction with methyl formate, the product of addition–elimination reaction was observed [Fig. 1(a)]. Similarly, when the reaction of carbanions with low basicity was investigated, the product of addition– elimination reaction was recorded in the spectra (Table 1). Moreover, for the reaction of 2,4-dichlorophenide anion, the tetrahedral adduct A1 was observed in the spectrum [Fig. 1 (b)]. In the case of moderately basic carbanions, besides the products and adducts of addition–elimination reaction, the products and adducts of the second addition–elimination reaction were also recorded in the spectra [Fig. 1(c)]. It was noted that in all recorded spectra, intense peaks with m/z 63 and 75 are observed. The first one corresponds to the protonbound dimer of the methoxide anion with methanol. Such product can be formed from methyl formate in a Riveros-type reaction.[14] The m/z 75 ion is formed in the reaction of the methoxide ion with carbon dioxide. It was observed in all gas phase reactions of methyl formate with anions studied in our lab. Traces of CO2 are present in the nebulizer and curtain gasses, and the reaction of methoxide ion with CO2 appears to be extremely effective because even ppm concentration of CO2 is sufficient. An additional proof for the composition of m/z 75 ion was obtained by addition of 13CO2 to the curtain and collision gas (in API 365 this is the same gas) and observation of m/z 76 ion. Also, the replacement of methyl formate with ethyl homologue resulting in the formation of m/z 89 ion confirms the proposed formula.
Table 2. Proton affinity (in kcal/mol) of anion G1 and G2 calculated by various quantum chemical methods Anion G1 (acyl) Δ G2 (ortho)
B3LYPa
M05-2Xb
G3MP2
G4MP2
369.0 2.3 371.3
368.7 1.2 369.9
373.5 2 371.5
373.8 2.6 371.2
a
B3LYP/6-311+g(3df,2p)//B3LYP/6-31g(d). M05-2X/6-311+g(3df,2p)//M05-2X/6-31g(d).
b
J. Mass Spectrom. (2014), 49, 1247–1253
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms
1249
Scheme 4. The mechanism of the gas phase addition–elimination reaction of p-cyanophenide anion with methyl formate.
B. Wilenska, P. Swider and W. Danikiewicz The formation of products in the second addition–elimination reaction in the group of aromatic carbanions with moderate basicity is probably associated with basicity of carbanions in the gas phase, which is responsible for their stability under these reaction conditions. The basicity is closely linked to the presence of the electron-withdrawing groups. Their presence as an ortho or para substituent in the aromatic ring has a significant influence on stability, not only of aromatic carbanions but also of benzaldehyde anions formed in the elimination reaction of methanol molecule from adducts in the gas phase [Fig. 1(c)]. Confirmation of this effect can be the fact that in the case of aromatic carbanions with electron-withdrawing groups, such as CN, CF3, and Cl, adducts and products of the second addition–elimination reaction are observed in the recorded spectra (Table 1). These products
Figure 2. Energy diagram of reaction presented in Scheme 4 (B3LYP/6-311 +g(3df,2p)//B3LYP/6-31g(d)).
are not observed in the case of 2,4-dichlorophenide anion [Fig. 1(b)]. The most likely rationalization for this observation is that two chlorine atoms show high stabilizing effect of the carbanion [A1–MeOH] making it less reactive. The mechanism of gas phase addition–elimination reaction is very similar to this observed in a liquid phase; however, in the elimination step, a methanol molecule is eliminated, instead of methoxyl anion, which is eliminated in the condensed phase reaction. Moreover, during the gas phase reaction (what is opposite to the reaction in the liquid phase) two kinds of products are possible, which differ from each other by location of a negative charge in the molecule after the elimination of methanol molecule from the adduct. Because of its high basicity in the gas phase, the methoxyl anion, which is eliminated from the adduct, takes off the most acidic proton from benzaldehyde derivative. Hence, two final products of the addition–elimination reaction are possible: an acyl anion (G1 in Scheme 4) and the substituted phenide anion (G2 in Scheme 4). To distinguish between these two anions, quantum chemical calculations have been performed. Proton affinity calculations for the addition–elimination reaction of p-cyanophenide anion with methyl formate in the gas phase confirmed that two anionic products: G1 and G2 in Scheme 4 can be formed. The PA values for these two anions are very similar, and the difference between them is approximately 2 kcal/mol (the results of PA calculation are presented in Table 2). It is worth noting that the PA of G2 anion calculated by DFT methods is higher than for G1 anion, but using quantum chemistry composite methods (G3MP2, G4MP2), which are supposed to be more accurate, the reverse result was obtained (Table 2). On the basis of these calculations, it is not possible to establish unequivocally which anion is mainly generated. It is very likely that both anions are formed in the addition–elimination reaction. Moreover, there are
Scheme 5. Proposed mechanism of the addition–elimination reaction of G1 anion with methyl formate.
1250
Scheme 6. Proposed mechanism of the addition–elimination reaction of G2 anion with methyl formate.
wileyonlinelibrary.com/journal/jms
Copyright © 2014 John Wiley & Sons, Ltd.
J. Mass Spectrom. (2014), 49, 1247–1253
Gas phase Tishchenko reaction no kinetic and thermodynamic obstacles, which could adversely affect the formation of G1 and G2 anions (Fig. 2). Because of the formation of two possible products in the gas phase addition–elimination reaction (structures G1 and G2 in Scheme 4), there are two different ways of the second addition– elimination reaction, in which different products are obtained
(Schemes 5 and 6). The G1 anion is supposed to react with another molecule of methyl formate producing the adduct I (Scheme 5). The reaction energy diagram corresponding to this reaction pathway is depicted in Fig. 3. We were not able to model the direct elimination of methanol from anion I. In the course of the calculation, we observed that during the departure of the methoxy group from the adduct, this group is moved to the neighboring carbonyl group. The resulting anion J undergoes the elimination of methanol
Figure 3. Energy diagram of reaction presented in Scheme 5 (B3LYP/6-311 +g(3df,2p)//B3LYP/6-31g(d)).
Scheme 7. Conversion of Q1 and Q2 anions into 5-cyanophthalide (U1) and 6-cyanophthalide (U2) anions.
Figure 4. Energy diagrams of reaction presented in Scheme 6 and 7 (B3LYP/6-311+g(3df,2p)//B3LYP/6-31g(d)).
1251
Scheme 8. Proposed mechanism of the Tishchenko reaction in the gas phase.
J. Mass Spectrom. (2014), 49, 1247–1253
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms
B. Wilenska, P. Swider and W. Danikiewicz yielding finally anion L. The energies of all transition states are lower than the energy of the reactants so the formation of anion L is feasible. An alternative structure for anion L, i.e. with the negative charge located in the ortho position of the benzene ring (analogously to G2) has to be rejected because its PA is about 19 kcal/mol higher than the PA of the anion L. Different reaction products responsible for the m/z 158 peak [Fig. 1(c)] should be formed in the reaction of anion G2 with methyl formate. We expected that analogous to the reaction of the starting
cyanophenide ion formylation of the benzene ring via addition– elimination reaction should take place (Scheme 6). The DFT calculations show that such process is energetically possible (Fig. 4). The energies of all intermediates and transition states are located below the energy of reactants or – in the case of TS10 and TS11 – are only slightly higher. Searching for the optimal geometries of anions Q1 and Q2, we observed an unexpected ‘in silico reaction’ in which these two species were converted into much more stable anions U1
Figure 5. Energy diagrams of Tishchenko reaction presented in Scheme 8 (B3LYP/6-311+g(3df,2p)//B3LYP/6-31g(d)).
1252
Figure 6. The fragmentation spectra of (a) anionic product formed during the gas phase reaction in an electrospray ion source, (b) 6-cyanophthalide anion, and (c) 5-cyanophthalide anion. All experiments were performed in the collision cell with 15-eV collision energy. Because of the presence of acidic protons in 6-cyanophthalide and 5-cyanophthalide molecules, both anions were generated directly in the electrospray ion source in the negative mode. Ions with m/z 146 and 148 are formed in the reactions of m/z 130 fragment ion with oxygen and water, respectively.
wileyonlinelibrary.com/journal/jms
Copyright © 2014 John Wiley & Sons, Ltd.
J. Mass Spectrom. (2014), 49, 1247–1253
Gas phase Tishchenko reaction
J. Mass Spectrom. (2014), 49, 1247–1253
Conclusions It was shown that the phenide ions substituted with electronwithdrawing groups react with methyl formate in the gas phase in a two-step reaction. The first step that proceeds according to the typical addition–elimination mechanism results in the formation of the anion of the respective benzaldehyde derivative. Quantum chemical calculations show that the negative charge can be located either in the aldehyde group (acyl anion) or in the benzene ring (phenide anion) in position ortho to an aldehyde moiety. In the second step, the preliminary-formed anion reacts with the second molecule of methyl formate yielding formally product of the second addition–elimination reaction. In the reaction of an acyl anion, the derivative of phenylglioxal can be formed. The reaction of the ortho-formylphenide anion can lead either to the anion of phthalic aldehyde or – according to the Tishchenko reaction mechanism – to the respective phthalide anion. Theoretical calculations as well as CID spectra of the model compounds suggest that the formation of the latter product is most likely. This is the most thermodynamically stable anion, and on the reaction path leading to its formation, there are no high activation energy barriers. These results show that it is very likely that we observed the first example of the Tishchenko-type reaction in the gas phase. Acknowledgements This work was financed from the Polish Ministry of Science and Higher Education grant no. N N204 155536.
References [1] M. L. Bender. Chem. Rev. 1960, 60, 53. [2] H. van der Wel, N. M. M. Nibbering. Recl. Trav. Chimiques Pays-Bas 1988, 107, 479. [3] J. Haib, D. Stahl. Org. Mass Spectrom. 1992, 27, 377. [4] J. L. Wilbur, J. I. Brauman. J. Am. Chem. Soc. 1994, 116, 5839. [5] O. I. Asubiojo, J. I. Brauman. J. Am. Chem. Soc. 1979, 101, 3715. [6] M. Zimnicka, B. Wileńska, O. Sekiguchi, E. Uggerud, W. Danikiewicz. J. Mass Spectrom. 2012, 47, 425. [7] M. Zimnicka, O. Sekiguchi, E. Uggerud, W. Danikiewicz. Int. J. Mass Spectrom. 2012, 316–318, 76. [8] T. Bieńkowski, W. Danikiewicz. Rapid Commun. Mass Spectrom. 2003, 17, 697. [9] W. Tishchenko. J. Russ. Phys. Chem. 1906, 38, 355. [10] J. I. Uenishi, S. Masuda, S. Wakabayashi, Tetrahedron Lett. 1991, 32, 5097. [11] A. R. Katritzky, F.-B. Ji, W.-Q. Fan, P. Beretta, M. Bertoldi. J. Heterocycl. Chem. 1992, 29, 1519. [12] W. Borsche, K. Diacont, H. Hanau. Ber. Deutsch. Chem. Ges. (A and B Series) 1934, 67, 675. [13] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Wallingford CT, 2009. [14] N. M. M. Nibbering. In The Encyclopedia of Mass Spectrometry, vol. 4, M. L. Gross, R. M. Caprioli (Eds). Elsevier: New York, 2005, pp. 475.
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms
1253
and U2 corresponding, respectively, to 5-cyanophthalide and 6-cyanophthalide anions (Scheme 7). Performed calculations showed that U1 anion is 23.8 kcal/mol more stable than Q1 (Fig. 4). In the case of Q2 and U2 anions, this difference is 31.8 kcal/mol. Search for the transition states of these two reactions showed however that the activation energy barrier of both of them is quite high and is equal to 18.6 kcal/mol for Q1 to U1 transformation and 20.8 kcal/mol for the second reaction. These results show that direct transformation of ions Q into U is possible but should be rather slow. Looking for another pathways leading to the U1 or U2 ions from the adduct of methyl formate to anion G2, we considered a variant of the Tishchenko reaction (Scheme 8). It turned out that the migration of hydride anion from the adduct N2 (N and N2 adducts differ in conformation) and consequent elimination of methanol can proceed with very small energy barrier. In the first step of the bicyclic product formation, the Cannizzaro disproportionation reaction occurred within the adduct yielding ion R. Subsequently, an intramolecular attack of the anion, in which a negative charge is localized at the oxygen atom, to the carbonyl group occurs and the new adduct S is formed. The elimination of methanol molecule from the adduct R is the final step of the gas phase version of Tishchenko reaction, which ends with phthalide anion U2. It has to be noted that this mechanism leads solely to U2 anion and it is not possible to rationalize the formation of the U1 isomeric ion via Tishchenko reaction mechanism. Obtained products and transition states are located below the energy of reactants, and additionally, the process is highly exothermic (Fig. 5). Without doubts, this reaction should proceed much faster than the reactions leading to the anions L, Q1, and Q2. With high degree of certainty, it can be concluded that the anion U2 obtained in Tishchenko reaction (observed in mass spectrum of reaction as signal at m/z 158) is the major/final product of the reaction of anion G2 with methyl formate. To get at least partial proof for the theoretical predictions, we compared CID spectrum of the product formed during the gas phase reaction of p-cyanophenide ion with methyl formate with CID spectra of 5-cyanophthalide and 6-cyanophthalide anions. The results are shown in Fig. 6. Unfortunately, we were not able to record the CID spectrum of the anion L because the parent compound is not available commercially and difficult to synthesize. The main fragmentation mechanism of all studied anions consists of the consecutive elimination of two carbon monoxide molecules from the precursor ions. It is not possible to distinguish unequivocally between 5-cyanophthalide and 6-cyanophthalide anions; however, the relative intensity ratios suggest that the structure of 6-cyanophthalide (U2) is more likely. This is in agreement with the theoretical predictions. On the basis of the CID experiments, we cannot exclude the possibility of the formation of anion L with phenylglioxal structure because it is very likely that its fragmentation pattern will be quite similar. However, according to the results of the calculations, this anion is about 26 kcal/mol less stable than the anion U2 so its formation seems to be unlikely. So far, this kind of processes involving aromatic carbanions, in which the addition–elimination and Tishchenko reactions are occurring at the same time, were not observed in the liquid phase. This fact shows how reaction conditions have a huge influence on its course. Probably, the absence of solvation effects in the gas phase allows formation of ions during the first addition–elimination reaction have enough energy to undergo subsequent reactions as it was confirmed by quantum chemical calculations.
Revised: 9 July 2014
Accepted: 29 July 2014
Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI 10.1002/jms.3452
Addition–elimination versus Tishchenko reaction in the gas phase Beata Wilenska,a† Pawel Swidera and Witold Danikiewicza* Gas phase reactions of the substituted phenide ions with methyl formate have been studied. It was found that the results of these reactions depend mainly on the basicity of the phenide ion, which is related to the presence of the electron-accepting or electrondonating substituents in the benzene ring. It was shown that the phenide ions substituted with electron-withdrawing groups react with methyl formate in the gas phase in a two-step reaction. The first step that proceeds according to the typical addition– elimination mechanism results in the formation of the anion of the respective benzaldehyde derivative with the negative charge located either in the aldehyde group (acyl anion) or in the benzene ring (phenide anion) in position ortho to an aldehyde moiety. In the second step, the preliminary-formed anion reacts with the second molecule of methyl formate yielding formally product of the second addition–elimination reaction. Theoretical calculations as well as collision induced dissociation spectra of the model compounds suggest that this reaction proceeds according to the Tishchenko reaction mechanism yielding the respective phthalide anion. According to our knowledge, this is the first example of the Tishchenko-type reaction in the gas phase. Copyright © 2014 John Wiley & Sons, Ltd. Keywords: ion-molecule reactions; phenide ions; addition-elimination reaction; Tishchenko reaction; DFT calculations
Introduction
J. Mass Spectrom. (2014), 49, 1247–1253
Experimental The aromatic carboxylic acids and 5-cyanophthalide were purchased from Aldrich and used without further purification. 6-Cyanophthalide
* Correspondence to: Witold Danikiewicz, Institute of Organic Chemistry, Polish Academy of Sciences, ul. Kasprzaka 44/52, Warsaw, Poland. E-mail: witold. [email protected] †
Present address: Faculty of Chemistry, Warsaw University, ul. Pasteura 1, 02-093, Warsaw, Poland
a Institute of Organic Chemistry, Polish Academy of Sciences, ul. Kasprzaka 44/52, Warsaw, Poland
Copyright © 2014 John Wiley & Sons, Ltd.
1247
The addition–elimination reaction to the carbonyl group and the Tishchenko reaction are very common and useful reactions in the organic chemistry. A typical addition–elimination reaction runs through two steps: at first, an attack of nucleophile to the carbonyl group of carboxylic acid derivatives occurs and then in the second step, an elimination of an anionic molecule takes place (Scheme 1). The first step of the addition–elimination reaction in condensed phase is reversible and the second one limits the reaction rate. The nucleophiles are generated in situ just before the process in the reaction of bases with appropriate precursors. Hence, the addition–elimination reaction is catalyzed by basic compounds, especially amines.[1] In the gas phase, the addition–elimination reaction has been studied extensively. Most of the researchers investigated reactions of inorganic anions such as hydroxyl anion with carbonyl compounds and focused on the formation of a tetrahedral transition state, similar to the condense phase.[2–5] There are also few reports about gas phase addition–elimination reaction of aromatic nucleophiles with carbonyl compounds. The Danikiewicz group investigated reactions of (halo)nitrophenide ions with alkyl acrylates, where the addition– elimination reaction was observed as one of the main processes.[6,7] In the gas phase, this reaction is supposed to proceed along the same mechanism as in the condense phase, but after the elimination step, a proton transfer reaction occurs. Hence, the final product is in an anionic form. Moreover, in contrast to the liquid phase, the nucleophiles are generated in the gas phase by simple decarboxylation of the appropriate carboxylic acid anions during collision induced dissociation (CID) in the electrospray ion source of mass spectrometer.[8] The carbanions formed in this process, with exception of simple alkyl anions, are stable and do not undergo further fragmentation. The Tishchenko reaction is also a catalytic process in the liquid phase and involves the disproportionation of two aldehyde molecules to generate an appropriate ester.[9] Traditional Tishchenko
reaction is carried out in the presence of a Lewis acid, but it is also possible to conduct the reaction using a nucleophilic catalyst (nucleophile-induced conversion through a Cannizzaro– Tishchenko-type mechanism) (Scheme 2). These two approaches to Tishchenko reaction have a common feature – a hydride shift from one aldehyde molecule to another is observed. This reaction can proceed also intramolecularly yielding the corresponding lactone.[10] So far, any variant of the gas phase Tishchenko reaction was not described in the literature. In our laboratory, the reactions of aromatic carbanions substituted with electron-withdrawing or electron-donating groups with methyl formate were investigated. During these experiments, the formation of the appropriate aldehyde anion as a product of gas phase addition–elimination reaction was observed. Interestingly, this aldehyde anion underwent further reaction with another molecule of methyl formate creating a new adduct, which can be transformed to the product of addition–elimination reaction or undergo intramolecular gas phase Tishchenko reaction yielding lactone. In this paper, we present the results of our mass spectrometric and computational study concerning these reactions.
B. Wilenska, P. Swider and W. Danikiewicz
Scheme 1. Mechanism of addition–elimination reaction to the carbonyl group of carboxylic acid derivatives.
H H
H H
Scheme 2. Mechanisms of the condensed phase Tishchenko (a) and Cannizzaro–Tishchenko (b) reactions.
was synthesized according to the procedure described in the literature.[11,12] Fresh stock solutions of carboxylic acids were prepared by dissolution in methanol (HPLC grade, Merck) and infused to the electrospray ion source at the rate of 10 μl/min using Harvard Apparatus Pump 11 syringe pump. All gas phase reactions were performed using an API 365 triple quadrupole mass spectrometer (Applied Biosystems) equipped with a TurboIonSpray source. All spectra were recorded in a negative ion mode standard electrospray ion source. Source parameters were as follows: capillary voltage, 4.5 kV; nebulizer gas, 12; and curtain gas, 8 (arbitrary units, N2 in both cases). Typical DP (Declustering Potential) values were between 20 and 60 V and FP (Focusing Potential) values between 180 and 300 V. To enable the use of different collision gasses, the collision gas inlet was modified to allow introduction of the selected collision gas independently from the curtain gas. For the CID experiments, nitrogen was used as the collision gas. The ion molecule reactions in the collision cell were studied using appropriate gaseous reagents. In the case of liquid reagents (methyl formate), a stream of nitrogen was flushed over the surface of the appropriate liquid in a small flask. The pressure of collision gas was set at 3 (arbitrary units). Collision cell parameters were set as follows: entrance potential – 10 V, exit potential 15 V. The collision energy during the reaction was set at 5 eV but during CID experiments was adjusted manually in the range of 5–25 eV to achieve the best results.
1248
Scheme 3. Investigated carbanions and their proton affinities (PA, kcal/mol) in the gas phase.
wileyonlinelibrary.com/journal/jms
Figure 1. Spectra recorded in the negative ion mode for the reaction of (a) 4-metoxyphenide anion, (b) 2,4-dichlorophenide anion, and (c) 4-cyanophenide anion with methyl formate in the gas phase. See Table 1 for explanation of A1 and A2 symbols.
Copyright © 2014 John Wiley & Sons, Ltd.
J. Mass Spectrom. (2014), 49, 1247–1253
Gas phase Tishchenko reaction All quantum chemical calculations described in this work were performed using the Gaussian 09 software package.[13] Starting geometries for geometry optimization were created using the GaussView 5.0 program. Reaction potential energy profiles and molecular properties were computed applying density functional theory (DFT) methods at the B3LYP/6-311+G(3df,2p)//B3LYP/631G(d) level. Proton affinities (PAs) of the anions at 298.15 K were calculated as the enthalpies of the reaction: AH = A + H+ using the equation 298 298 ¼ H298 PA ¼ ΔH298 r A þ HHþ HAH H298 ¼ E 0 þ Htherm where Htherm is the thermal correction to enthalpy The enthalpy of the proton at 298.15 K was set to 1.48 kcal/mol, i.e. 5/2RT.
Results and discussion During our study, we investigated aromatic carbanions with a variety of substituents, which were obtained in the reaction of decarboxylation of appropriate carboxylic acid anions. For comparison, one aliphatic carbanion, i.e. benzyl anion, has been added. On the basis of their PA in the gas phase, we decided to divide them into three groups: (1) strongly basic carbanions with PA value Table 1. Aromatic carbanions with their PA value (kcal/mol) and the results of their reaction with methyl formate in the gas phase Carbanion
p-MeOPh Ph p-ClPh p-CF3Ph o-CF3Ph p-CNPh o-CNPh PhCH2 2,4-diClPh
PA (kcal/mol) A1 403 401 394 390 387 385 384 381 381
[A1–MeOH]
A2
[A2–MeOH]
++ + + ++ +
+ ++ + + + ++
+ + + + ++ +
+ + ++ ++ +
A1 , primary adduct anion, [A1–MeOH] , anion of the methanol elimination product from the primary adduct, A2 , secondary adduct anion, [A2–MeOH] , anion of the methanol elimination product from the secondary adduct.
higher than 400 kcal/mol, (2) carbanions with moderate basicity (PA values in the range between 394 and 384 kcal/mol), and (3) carbanions with lower basicity (PA value below 384 kcal/mol), (Scheme 3). Most of the aromatic carbanions, except for the phenide anion, undergo the addition–elimination reaction in the gas phase. In the group of strongly basic carbanions only for the p-methoxyphenide ion in the reaction with methyl formate, the product of addition–elimination reaction was observed [Fig. 1(a)]. Similarly, when the reaction of carbanions with low basicity was investigated, the product of addition– elimination reaction was recorded in the spectra (Table 1). Moreover, for the reaction of 2,4-dichlorophenide anion, the tetrahedral adduct A1 was observed in the spectrum [Fig. 1 (b)]. In the case of moderately basic carbanions, besides the products and adducts of addition–elimination reaction, the products and adducts of the second addition–elimination reaction were also recorded in the spectra [Fig. 1(c)]. It was noted that in all recorded spectra, intense peaks with m/z 63 and 75 are observed. The first one corresponds to the protonbound dimer of the methoxide anion with methanol. Such product can be formed from methyl formate in a Riveros-type reaction.[14] The m/z 75 ion is formed in the reaction of the methoxide ion with carbon dioxide. It was observed in all gas phase reactions of methyl formate with anions studied in our lab. Traces of CO2 are present in the nebulizer and curtain gasses, and the reaction of methoxide ion with CO2 appears to be extremely effective because even ppm concentration of CO2 is sufficient. An additional proof for the composition of m/z 75 ion was obtained by addition of 13CO2 to the curtain and collision gas (in API 365 this is the same gas) and observation of m/z 76 ion. Also, the replacement of methyl formate with ethyl homologue resulting in the formation of m/z 89 ion confirms the proposed formula.
Table 2. Proton affinity (in kcal/mol) of anion G1 and G2 calculated by various quantum chemical methods Anion G1 (acyl) Δ G2 (ortho)
B3LYPa
M05-2Xb
G3MP2
G4MP2
369.0 2.3 371.3
368.7 1.2 369.9
373.5 2 371.5
373.8 2.6 371.2
a
B3LYP/6-311+g(3df,2p)//B3LYP/6-31g(d). M05-2X/6-311+g(3df,2p)//M05-2X/6-31g(d).
b
J. Mass Spectrom. (2014), 49, 1247–1253
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms
1249
Scheme 4. The mechanism of the gas phase addition–elimination reaction of p-cyanophenide anion with methyl formate.
B. Wilenska, P. Swider and W. Danikiewicz The formation of products in the second addition–elimination reaction in the group of aromatic carbanions with moderate basicity is probably associated with basicity of carbanions in the gas phase, which is responsible for their stability under these reaction conditions. The basicity is closely linked to the presence of the electron-withdrawing groups. Their presence as an ortho or para substituent in the aromatic ring has a significant influence on stability, not only of aromatic carbanions but also of benzaldehyde anions formed in the elimination reaction of methanol molecule from adducts in the gas phase [Fig. 1(c)]. Confirmation of this effect can be the fact that in the case of aromatic carbanions with electron-withdrawing groups, such as CN, CF3, and Cl, adducts and products of the second addition–elimination reaction are observed in the recorded spectra (Table 1). These products
Figure 2. Energy diagram of reaction presented in Scheme 4 (B3LYP/6-311 +g(3df,2p)//B3LYP/6-31g(d)).
are not observed in the case of 2,4-dichlorophenide anion [Fig. 1(b)]. The most likely rationalization for this observation is that two chlorine atoms show high stabilizing effect of the carbanion [A1–MeOH] making it less reactive. The mechanism of gas phase addition–elimination reaction is very similar to this observed in a liquid phase; however, in the elimination step, a methanol molecule is eliminated, instead of methoxyl anion, which is eliminated in the condensed phase reaction. Moreover, during the gas phase reaction (what is opposite to the reaction in the liquid phase) two kinds of products are possible, which differ from each other by location of a negative charge in the molecule after the elimination of methanol molecule from the adduct. Because of its high basicity in the gas phase, the methoxyl anion, which is eliminated from the adduct, takes off the most acidic proton from benzaldehyde derivative. Hence, two final products of the addition–elimination reaction are possible: an acyl anion (G1 in Scheme 4) and the substituted phenide anion (G2 in Scheme 4). To distinguish between these two anions, quantum chemical calculations have been performed. Proton affinity calculations for the addition–elimination reaction of p-cyanophenide anion with methyl formate in the gas phase confirmed that two anionic products: G1 and G2 in Scheme 4 can be formed. The PA values for these two anions are very similar, and the difference between them is approximately 2 kcal/mol (the results of PA calculation are presented in Table 2). It is worth noting that the PA of G2 anion calculated by DFT methods is higher than for G1 anion, but using quantum chemistry composite methods (G3MP2, G4MP2), which are supposed to be more accurate, the reverse result was obtained (Table 2). On the basis of these calculations, it is not possible to establish unequivocally which anion is mainly generated. It is very likely that both anions are formed in the addition–elimination reaction. Moreover, there are
Scheme 5. Proposed mechanism of the addition–elimination reaction of G1 anion with methyl formate.
1250
Scheme 6. Proposed mechanism of the addition–elimination reaction of G2 anion with methyl formate.
wileyonlinelibrary.com/journal/jms
Copyright © 2014 John Wiley & Sons, Ltd.
J. Mass Spectrom. (2014), 49, 1247–1253
Gas phase Tishchenko reaction no kinetic and thermodynamic obstacles, which could adversely affect the formation of G1 and G2 anions (Fig. 2). Because of the formation of two possible products in the gas phase addition–elimination reaction (structures G1 and G2 in Scheme 4), there are two different ways of the second addition– elimination reaction, in which different products are obtained
(Schemes 5 and 6). The G1 anion is supposed to react with another molecule of methyl formate producing the adduct I (Scheme 5). The reaction energy diagram corresponding to this reaction pathway is depicted in Fig. 3. We were not able to model the direct elimination of methanol from anion I. In the course of the calculation, we observed that during the departure of the methoxy group from the adduct, this group is moved to the neighboring carbonyl group. The resulting anion J undergoes the elimination of methanol
Figure 3. Energy diagram of reaction presented in Scheme 5 (B3LYP/6-311 +g(3df,2p)//B3LYP/6-31g(d)).
Scheme 7. Conversion of Q1 and Q2 anions into 5-cyanophthalide (U1) and 6-cyanophthalide (U2) anions.
Figure 4. Energy diagrams of reaction presented in Scheme 6 and 7 (B3LYP/6-311+g(3df,2p)//B3LYP/6-31g(d)).
1251
Scheme 8. Proposed mechanism of the Tishchenko reaction in the gas phase.
J. Mass Spectrom. (2014), 49, 1247–1253
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms
B. Wilenska, P. Swider and W. Danikiewicz yielding finally anion L. The energies of all transition states are lower than the energy of the reactants so the formation of anion L is feasible. An alternative structure for anion L, i.e. with the negative charge located in the ortho position of the benzene ring (analogously to G2) has to be rejected because its PA is about 19 kcal/mol higher than the PA of the anion L. Different reaction products responsible for the m/z 158 peak [Fig. 1(c)] should be formed in the reaction of anion G2 with methyl formate. We expected that analogous to the reaction of the starting
cyanophenide ion formylation of the benzene ring via addition– elimination reaction should take place (Scheme 6). The DFT calculations show that such process is energetically possible (Fig. 4). The energies of all intermediates and transition states are located below the energy of reactants or – in the case of TS10 and TS11 – are only slightly higher. Searching for the optimal geometries of anions Q1 and Q2, we observed an unexpected ‘in silico reaction’ in which these two species were converted into much more stable anions U1
Figure 5. Energy diagrams of Tishchenko reaction presented in Scheme 8 (B3LYP/6-311+g(3df,2p)//B3LYP/6-31g(d)).
1252
Figure 6. The fragmentation spectra of (a) anionic product formed during the gas phase reaction in an electrospray ion source, (b) 6-cyanophthalide anion, and (c) 5-cyanophthalide anion. All experiments were performed in the collision cell with 15-eV collision energy. Because of the presence of acidic protons in 6-cyanophthalide and 5-cyanophthalide molecules, both anions were generated directly in the electrospray ion source in the negative mode. Ions with m/z 146 and 148 are formed in the reactions of m/z 130 fragment ion with oxygen and water, respectively.
wileyonlinelibrary.com/journal/jms
Copyright © 2014 John Wiley & Sons, Ltd.
J. Mass Spectrom. (2014), 49, 1247–1253
Gas phase Tishchenko reaction
J. Mass Spectrom. (2014), 49, 1247–1253
Conclusions It was shown that the phenide ions substituted with electronwithdrawing groups react with methyl formate in the gas phase in a two-step reaction. The first step that proceeds according to the typical addition–elimination mechanism results in the formation of the anion of the respective benzaldehyde derivative. Quantum chemical calculations show that the negative charge can be located either in the aldehyde group (acyl anion) or in the benzene ring (phenide anion) in position ortho to an aldehyde moiety. In the second step, the preliminary-formed anion reacts with the second molecule of methyl formate yielding formally product of the second addition–elimination reaction. In the reaction of an acyl anion, the derivative of phenylglioxal can be formed. The reaction of the ortho-formylphenide anion can lead either to the anion of phthalic aldehyde or – according to the Tishchenko reaction mechanism – to the respective phthalide anion. Theoretical calculations as well as CID spectra of the model compounds suggest that the formation of the latter product is most likely. This is the most thermodynamically stable anion, and on the reaction path leading to its formation, there are no high activation energy barriers. These results show that it is very likely that we observed the first example of the Tishchenko-type reaction in the gas phase. Acknowledgements This work was financed from the Polish Ministry of Science and Higher Education grant no. N N204 155536.
References [1] M. L. Bender. Chem. Rev. 1960, 60, 53. [2] H. van der Wel, N. M. M. Nibbering. Recl. Trav. Chimiques Pays-Bas 1988, 107, 479. [3] J. Haib, D. Stahl. Org. Mass Spectrom. 1992, 27, 377. [4] J. L. Wilbur, J. I. Brauman. J. Am. Chem. Soc. 1994, 116, 5839. [5] O. I. Asubiojo, J. I. Brauman. J. Am. Chem. Soc. 1979, 101, 3715. [6] M. Zimnicka, B. Wileńska, O. Sekiguchi, E. Uggerud, W. Danikiewicz. J. Mass Spectrom. 2012, 47, 425. [7] M. Zimnicka, O. Sekiguchi, E. Uggerud, W. Danikiewicz. Int. J. Mass Spectrom. 2012, 316–318, 76. [8] T. Bieńkowski, W. Danikiewicz. Rapid Commun. Mass Spectrom. 2003, 17, 697. [9] W. Tishchenko. J. Russ. Phys. Chem. 1906, 38, 355. [10] J. I. Uenishi, S. Masuda, S. Wakabayashi, Tetrahedron Lett. 1991, 32, 5097. [11] A. R. Katritzky, F.-B. Ji, W.-Q. Fan, P. Beretta, M. Bertoldi. J. Heterocycl. Chem. 1992, 29, 1519. [12] W. Borsche, K. Diacont, H. Hanau. Ber. Deutsch. Chem. Ges. (A and B Series) 1934, 67, 675. [13] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Wallingford CT, 2009. [14] N. M. M. Nibbering. In The Encyclopedia of Mass Spectrometry, vol. 4, M. L. Gross, R. M. Caprioli (Eds). Elsevier: New York, 2005, pp. 475.
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jms
1253
and U2 corresponding, respectively, to 5-cyanophthalide and 6-cyanophthalide anions (Scheme 7). Performed calculations showed that U1 anion is 23.8 kcal/mol more stable than Q1 (Fig. 4). In the case of Q2 and U2 anions, this difference is 31.8 kcal/mol. Search for the transition states of these two reactions showed however that the activation energy barrier of both of them is quite high and is equal to 18.6 kcal/mol for Q1 to U1 transformation and 20.8 kcal/mol for the second reaction. These results show that direct transformation of ions Q into U is possible but should be rather slow. Looking for another pathways leading to the U1 or U2 ions from the adduct of methyl formate to anion G2, we considered a variant of the Tishchenko reaction (Scheme 8). It turned out that the migration of hydride anion from the adduct N2 (N and N2 adducts differ in conformation) and consequent elimination of methanol can proceed with very small energy barrier. In the first step of the bicyclic product formation, the Cannizzaro disproportionation reaction occurred within the adduct yielding ion R. Subsequently, an intramolecular attack of the anion, in which a negative charge is localized at the oxygen atom, to the carbonyl group occurs and the new adduct S is formed. The elimination of methanol molecule from the adduct R is the final step of the gas phase version of Tishchenko reaction, which ends with phthalide anion U2. It has to be noted that this mechanism leads solely to U2 anion and it is not possible to rationalize the formation of the U1 isomeric ion via Tishchenko reaction mechanism. Obtained products and transition states are located below the energy of reactants, and additionally, the process is highly exothermic (Fig. 5). Without doubts, this reaction should proceed much faster than the reactions leading to the anions L, Q1, and Q2. With high degree of certainty, it can be concluded that the anion U2 obtained in Tishchenko reaction (observed in mass spectrum of reaction as signal at m/z 158) is the major/final product of the reaction of anion G2 with methyl formate. To get at least partial proof for the theoretical predictions, we compared CID spectrum of the product formed during the gas phase reaction of p-cyanophenide ion with methyl formate with CID spectra of 5-cyanophthalide and 6-cyanophthalide anions. The results are shown in Fig. 6. Unfortunately, we were not able to record the CID spectrum of the anion L because the parent compound is not available commercially and difficult to synthesize. The main fragmentation mechanism of all studied anions consists of the consecutive elimination of two carbon monoxide molecules from the precursor ions. It is not possible to distinguish unequivocally between 5-cyanophthalide and 6-cyanophthalide anions; however, the relative intensity ratios suggest that the structure of 6-cyanophthalide (U2) is more likely. This is in agreement with the theoretical predictions. On the basis of the CID experiments, we cannot exclude the possibility of the formation of anion L with phenylglioxal structure because it is very likely that its fragmentation pattern will be quite similar. However, according to the results of the calculations, this anion is about 26 kcal/mol less stable than the anion U2 so its formation seems to be unlikely. So far, this kind of processes involving aromatic carbanions, in which the addition–elimination and Tishchenko reactions are occurring at the same time, were not observed in the liquid phase. This fact shows how reaction conditions have a huge influence on its course. Probably, the absence of solvation effects in the gas phase allows formation of ions during the first addition–elimination reaction have enough energy to undergo subsequent reactions as it was confirmed by quantum chemical calculations.
