formation of carbamate anions by the gas-phase reaction of ... · formation of carbamate anions by...

B American Society for Mass Spectrometry, 2016DOI: 10.1007/s13361-016-1357-5

J. Am. Soc. Mass Spectrom. (2016) 27:927Y939

RESEARCH ARTICLE

Formation of Carbamate Anions by the Gas-phase Reactionof Anilide Ions with CO2

Chongming Liu, Upul Nishshanka, Athula B. AttygalleCenter for Mass Spectrometry, Department of Chemistry, Chemical Biology, and Biomedical Engineering, Stevens Institute ofTechnology, Hoboken, NJ 07030, USA

Abstract. The anilide anion (m/z 92) generated directly from aniline, or indirectly as afragmentation product of deprotonated acetanilide, captures CO2 readily to form thecarbamate anion (m/z 136) in the collision cell, when CO2 is used as the collision gasin a tandem-quadrupole mass spectrometer. The gas-phase affinity of the anilide ionto CO2 is significantly higher than that of the phenoxide anion (m/z 93), which adds toCO2 only very sluggishly. Our results suggest that the efficacy of CO2 capturedepends on the natural charge density on the nitrogen atom, and relative nucleophi-licity of the anilide anion. Generally, conjugate bases generated from aniline deriva-tives with proton affinities (PA) less than 350 kcal/mol do not tend to add CO2 to formgaseous carbamate ions. For example, the anion generated from p-methoxyaniline

(PA = 367 kcal/mol) reacts significantly faster than that obtained from p-nitroaniline (PA = 343 kcal/mol). Althoughdeprotonated p-aminobenzoic acid adds very poorly because the negative charge is now located primarily on thecarboxylate group, it reacts more efficiently with CO2 if the carboxyl group is esterified. Moreover, mixture of CO2

and He as the collision gas was found to afford more efficient adduct formation than CO2 alone, or as mixturesmade with nitrogen or argon, because helium acts as an effective Bcooling^ gas and reduces the internal energyof reactant ions.Keywords: Nucleophilic addition, CO2 addition, Ion molecule reactions, Anilides, Negative ion

Received: 22 November 2015/Revised: 18 January 2016/Accepted: 22 January 2016/Published Online: 25 February 2016

Introduction

Methods to utilize carbon dioxide as a reagent has becomean important topic of research because increased levels of

CO2 in the atmosphere is known to cause many environmentalproblems [1, 2]. Thus, novel chemical reactions that consumeCO2 as a reactant are widely sought [3, 4]. One such reaction isthe addition of CO2 to carbanions [5, 6].

In general, the CO2 addition reaction has also been provento be useful as a structural probe for gaseous anions [7–16]. Forexample, Bienkowski and Danikiewicz [7] showed that thephenyl anion acts as a nucleophile and reacts with CO2 in thegas phase to generate the benzoate anion. Subsequently, the

reaction has been successfully used to derive structural infor-mation of carbanions generated by decarboxylation of isomerichydroxy- and sulfo-benzoic acids [8, 17]. Previously, we re-ported that even the phenoxide anion (m/z 93) adds to CO2 to asmall extent [9].

Under aqueous solution-phase conditions, NaHCO3 reactswith organic amines to form carbamates [4]. The gas-phasefragmentation mechanism of anions generated from carba-mates under electrospray ionization conditions has been inves-tigated by tandem mass spectrometry [4]. Although the gas-phase ion/molecule reactions of carbanions with CO2 havebeen widely studied, using the collision cell of tandem quad-rupole mass spectrometers as a reaction chamber [6], the effi-cacy of such an addition reaction for structural studies ofanilide ions has not been exploited. Danikiewicz and Zimnickanoted previously that gas-phase anilide ions do add to CO2 [6];however, the reaction has not been explored in depth. In ourexplorations on gaseous ion/molecule reactions, we also ob-served that gas-phase anilide anions generated under negative-mode ESI conditions undergo facile addition to CO2 in thecollision cell. Herein, we describe details of our explorations on

Electronic supplementary material The online version of this article (doi:10.1007/s13361-016-1357-5) contains supplementary material, which is availableto authorized users.

Correspondence to: Athula B. Attygalle; e-mail: [email protected]

http://crossmark.crossref.org/dialog/?doi=10.1007/s13361-016-1357-5&domain=pdf

http://dx.doi.org/10.1007/s13361-016-1357-5

http://dx.doi.org/10.1007/s13361-016-1357-5

gaseous anilide ions as a nitrogen-centered nucleophile thatadds to CO2. The efficacy of the reaction was evaluated withmany derivatives synthesized from substituted anilines, and theconclusions were rationalized by DFT calculations.

ExperimentalMaterials

Analytical grade phenol, aniline, acetanilide, p-fluoroaniline,p-chloroaniline, p-bromoaniline, p-methoxyaniline, p-nitroaniline, ethyl 4-aminobenzoate, p-aminobenzoic acid,benzenesulfonamide, p-toluenesulfonamide, and other reagents(purities >99%) were obtained from Sigma-Aldrich ChemicalCo. (St Louis, MO, USA) and used as purchased. Anilinesample solutions (1000 μg/mL and 100 μg/mL) and solutionsof other compounds (100 μg/mL) were prepared in acetonitrile/water (50:50, v/v).

Mass Spectrometry

An Applied Biosystems Sciex API-3000 (Concord, ON,Canada) triple quadrupole mass spectrometer equippedwith an electrospray ion source was used to obtain allCID mass spectra. Samples were infused to the ESIsource as acetonitrile-water-ammonia (50:50:0.01) solu-tions at a rate of 10 μL/min. For tandem mass spectro-metric experiments, the regular gas supply lines of theSciex API-3000 instrument were modified. Under stan-dard conditions, the nitrogen gas delivered from theexternal supply (at 60 psi) is divided to supply thecurtain gas and the collision gas (SupplementaryFigure 1a). To introduce CO2 or any other gas as thecollision gas, a T-junction and two toggle valves wereadded at the entrance to the collision cell [9] (Supple-mentary Figure 1b). To mix an additional gas, another T-junction was introduced and the gas flows were con-trolled by needle valves (Supplementary Figure 1c).The standard Sciex API-3000 instrument does not pro-vide a separate gauge to measure directly the pressure inthe collision cell. However, the analyzer housing pres-sure, which is about 0.6 × 10–5 Torr, increases to about2.1–3.1 × 10–5 Torr, when the collision gas engulfs thecell, because some gas escapes through the pinholes ofthe collision cell. Thus, the overall analyzer housingpressure can be used as an indirect measure of thecollision gas pressure.

Product-ion spectra were recorded from a mass-isolatedcomposite mixture of m/z 92 and 93 anions generated byelectrospraying a 200:1 (w/w) mixture of aniline and phenolin acetonitrile/water/ammonia (50:50:0.01). The first quadru-pole of the Sciex API-3000mass spectrometer was set to a low-resolution window to mass-isolate and pass bothm/z 92 and 93ions simultaneously. Spectra were acquired at a laboratory-frame collision energy setting of 5 eV, using CO2 or N2 asthe collision gas.

To investigate the effect of gases admixed to CO2 to theoverall reaction yield, first the analyzer-housing pressure obtain-ed using nitrogen as the collision gas was noted (1.6 × 10–5 Torr).Then CO2 was admixed to nitrogen by opening the needle valve(Supplementary Figure 1c) and new pressure was noted(3.1 × 10–5 Torr), and a product-ion spectrum was recorded fromm/z 92 ion generated from aniline. Subsequently, the nitrogen andCO2 supply to the collision cell was shut off, and helium wasintroduced in a controlled manner to achieve the same pressurereading as that previously achieved with nitrogen alone(1.6 × 10–5 Torr). After readjusting the CO2 supply to deliver apressure reading of 3.1 × 10–5 Torr, a product-ion spectrum wasrecorded. Analogous experiments were performed solely withCO2 as the collision gas (2.1 × 10–5 Torr), or by admixing argonwith CO2 (3.1 × 10–5 Torr).

The source temperature was held at 100 °C. The ion sprayvoltage was set to –4.5 kV, and the declustering potential washeld at –30 or –70 V to enforce low-energy or high-energy in-source fragmentation conditions. The focusing potential,entrancing potential, and exiting potential were set at –200 V,–10 V, and –5 V, respectively. The collision energy was setbetween –5 eV and ––50 eV, to achieve low- or high-energycollision-induced dissociation in the collision cell.

LC-MS

A sample of acetanilide in acetonitrile/water (100 μg/mL) wasinjected via a 20-μL loop onto a RP-C18 column (SunFire,Ireland, 3.5 μm, 4.6 × 150 mm), which was eluted isocraticallyusing a mixture of methanol:water (90:10, v/v) at a flow rate of1.000 mL/min. A total-ion chromatogram (m/z 20–200) wasrecorded by acquiring product ion spectra of the m/z 92 iongenerated in the ESI source at a declustering potential of –70V,from acetanilide dissolved in acetonitrile/water using N2 as thecollision gas at collision energy setting of 5 eV. Similarly, anintensity versus time plot was generated by a multiple reactionmonitoring (MRM) experiment using CO2 as the collision gas(overall housing pressure 2.5 × 10–5 Torr) monitoring thetransition of m/z 92 → 136, for the in source generated m/z92 ion at a declustering potential of –70 V.

Computational Methods

Geometry optimizations of chemical structures were carriedout by the density functional theory method B3LYP with a 6-311++G(2d, 2p) basis set for all atoms. All calculations wereperformed using Gaussian 09 software package [18, 19]. Toverify that the optimized structures correspond to theminimum-energy states on the respective potential energy sur-faces, frequency analysis was done at the same level. Gibbsfree energy values for CO2 addition reactions were calculatedat standard thermochemical conditions (temperature 298.150K, and pressure 1.00000 atm). Gibbs free energy values wereadjusted by adding the zero-point vibrational energy (unit:kcal/mol). Charge densities on atoms were computed byBnatural population analysis.^

928 C. Liu et al.: Gas Phase Reaction of Anilide Anions with CO2

Results and DiscussionAromatic amines are widely used as synthetic raw materials orintermediates in chemical industry [20]. For mass spectromet-ric investigations of aromatic amines, the positive-ion-generating ionization mode is the preferred method becausethe amino group undergoes facile protonation under ESI con-ditions. Although it is not customary to record negative-ionmass spectra of aromatic amines, gaseous anions can be gen-erated under mass spectrometric conditions even from aniline(Supplementary Figure 2a) albeit its gas-phase acidity is rela-tively low. On the other hand, more acidic amines such asprocaine (N,N-diethylaminoethyl 4-aminobenzoate), and ben-zocaine (ethyl 4-aminobenzoate) undergo more facile depro-tonation and generate abundant gaseous anions under negative-

ion ESI conditions [21]. Although molecular and electronicstructures and vibrational frequencies of anilide andsubstituted-anilide ions had been theoretically evaluated [22],and anilide-ions have been employed as nucleophiles to formaromatic substitution products [23], to our knowledge no massspectrometry-based structural work has been carried out previ-ously to determine the efficacy of anilide ions as reagents fornucleophilic addition reactions.

To start with, we pursued to find out whether the anilide ionwould add to CO2 in a manner similar to that reported forphenoxide and phenide ions [8, 9]. Our preliminary resultsindicated that the anilide ion undergoes facile addition toCO2. For example, when the m/z 92 ion generated from anaqueous solution of aniline by electrospray ionizationwas massisolated and transferred to the collision cell of a tandem

Table 1. Relative Gibbs Free Energies (in kcal/mol) of Species 1-4, with Respect to That of 1, and Their Reaction Energies with CO2

Species 1 2 3 4Relative Gibbs free energy 0 29.61 37.35 39.27Reactions 1 + CO2 → 5 2 + CO2 → 6 3 + CO2 → 7 4 + CO2 → 8Relative Gibbs free energy –27.06 –54.47 –54.43 –55.70

Figure 1. Product ion spectra recorded from amixture ofm/z 92 andm/z 93 anions generated by ESI from a 200:1 (w/w) mixture ofaniline and phenol in acetonitrile/water. Spectra were acquired on a Sciex API-3000 mass spectrometer, at a laboratory-framecollision energy setting of 5 eV, using CO2 (a) or N2 (b) as the collision gas. The first quadrupole was set to a low-resolution window tomass-isolate and pass both m/z 92 and 93 ions

C. Liu et al.: Gas Phase Reaction of Anilide Anions with CO2 929

quadrupolar mass spectrometer containing CO2 as the collisiongas, the spectrum recorded showed a peak at m/z 136 for theadduct (Supplementary Figure 2d). A similar peak was notobserved when N2 was used as the collision gas (Supplemen-tary Figure 2a).

To find out the efficacy of the CO2 addition under compet-itive reaction conditions, we prepared a mixture of aniline andphenol at a molar ratio of 200:1 in acetonitrile/water/ammoniawith the intent of generating two mass spectrometric peaks ofnear-equal intensity at m/z 92 and 93 for the anilide and

phenoxide ions, respectively (a high molar ratio of anilinehad to be used because its gas-phase acidity is much lowerthan that of phenol). When the mass isolation window was setto m/z 92.2 and the resolution of the first quadrupole wasselected to be low (the Sciex API-3000 instrument allows onlythree resolution settings: low, unit, or high), the spectrumrecorded from the aniline and phenol mixture showed twopeaks at m/z 92 and 93 to indicate that both ions were trans-mitted to the collision cell. Under these conditions, no adductpeaks were observed when N2 was used as the collision gas

Table 2. Structures of Species 1–8 (atom labels; blue = nitrogen; red = oxygen; gray = carbon)

(1) (2) (3) (4)

(5) (6) (7) (8)

Scheme 1. A schematic illustrating relative Gibbs free energies of precursor and product ions associated with direct carboxylationof the anilide ion, or indirect addition after the rearrangement to ortho-aminophenide ion (2)


(Figure 1b). On the other hand, an intense peak signifying theadduct formation was observed atm/z 136 when CO2 was usedas the collision gas (Figure 1a; note that the peak at m/z 137 isprimarily due the contribution of the 13C-isotopologue of theanilide ion to the ions corresponding to the peak at m/z 93). Aratio calculation after peak integration showed that the relativepeak area of the m/z 137 peak was about 3.8 when normalizedto that of the m/z 136 peak. Evidently, the anilide ion demon-strates a strong affinity to CO2 compared with the extremelymediocre affinity shown by the phenoxide ion [9].

To comprehend the intense affinity of the anilide anion to CO2,a detailed quantum chemical study was performed by B3LYPdensity functional method using 6-311++G(2d,2p) basis set[19]. Our computational results support the conclusion ofVakula et al [22] that the deprotonation occurs primarily atthe NH2 group. The Gibbs free energies calculated for four ofthe hypothetical structures expected from deprotonated aniline(m/z 92, 1–4) disclosed that the anilide structure (1) is theenergetically most favorable form (Tables 1 and 2). The neg-ative charge on the anilide ion is accumulated primarily on thenitrogen atom because computations assigned a formal charge

Scheme 2. Proposed pathway for the formation of the CO2 adduct from the m/z 92 anilide anion

Figure 2. Negative-ion product ion spectra recorded from them/z 134 ion generated by ESI from acetanilide in acetonitrile/water.Spectra were acquired on a Sciex API-3000 mass spectrometer, using CO2 (a) or N2 (b) as the collision gas at a laboratory-framecollision energy setting of 5 eV


of –0.89 e to the nitrogen atom (Supplementary Table 1). Of thethree aminophenide structures (2, 3, and 4), the ortho isomer isthe preferred form over the meta and para isomers.

Gibbs free energy calculations showed that the addition ofCO2 directly to anilide anion is an exergonic reaction(Scheme 1), whereas the direct addiction of CO2 to a phenoxide

Scheme 3. Proposed fragmentation and addition pathways to generate an m/z 136 ion from deprotonated acetanilide

Figure 3. A chromatogram obtained by acquiring product ion spectra (m/z 20–200) of the m/z 92 ion generated by in-sourcefragmentation of them/z 134 ion obtained at the declustering potential setting of 70 eV, from a solution of acetanilide in acetonitrile/water. The m/z 92 ion was mass-isolated by Q1 and subjected to fragmentation using N2 as the collision gas at a collision energysetting of 5 eV (a). An intensity versus time plot generated by monitoring the MRM transition of m/z 92 → 136, using CO2 as thecollision gas at a collision energy setting of 5 eV (b). Both experiments were conducted by eluting a RP-C18 column (5 μm, 4.6 ×150 mm) isocratically using a mixture of methanol:water (90:10, v/v) at a flow rate of 1.000 mL/min


anion is an endergonic reaction [9]. Moreover, the bond dis-tance between Ph–NH and CO2 moieties in the carbamatestructure (5) was found to be 1.475 Å. In other words, in theoptimized carbamate structure (5), the two constituent entitiesare held sufficiently close to indicate that they are held by acovalent bond. Although the CO2 addition to ortho (2), meta(3), or para (4) aminophenide ions is even more exergonic(Table 1), for the formation of the respective precursor ion fromthe anilide ion, the precursor ion must undergo a priorrearrangment. For example, for the formation of the ortho-aminophenide ion (2), a proton must be transferred to thenegatively charged nitrogen atom via a four-member transitionstate (TS 1-2; Scheme 1). In the cyclicTS 1-2 structure, the H–N and H–C distances were calculated to be 1.348 and 1.479 Å,respectively. However, an energy barrier of 55.26 kcal/molrenders this transformation prohibitive (Scheme 1). Thus, thedirect carboxylation of the anilide ion, instead of indirectaddition after a rearrangement to ortho-aminophenide ion(Scheme 2), is considered to be the more plausible mechanism.

The dramatic affinity of the anilide anion to CO2 was furtherdemonstrated by an MS/MS experiment conducted with them/z 134 ion generated from acetanilide. The anion generated bydirect deprotonation of acetanilide is not sufficiently

nucleophilic to add to CO2 (because a peak was not observed atm/z 178; Figure 2a). However, the m/z 134 ion, once activatedunder tandemmass spectrometric conditions, loses elements ofketene to form an ion of m/z 92, which underwent a rapidreaction with CO2 to form an adduct of m/z 136 when CO2

was used as the collision gas (Figure 2b, Scheme 3). Asexpected, peak atm/z 136 was not observed when N2 was usedas the collision gas (Figure 2a).

Moreover, if a compound generates fragment ions that showspecific affinity to CO2, this addition reaction can be employedunder MRM LC-MS conditions to specifically detect the pres-ence of such analytes in a mixture. For example, a chromato-gram obtained by recording product-ion spectra of the in-source-generated m/z 92 anion from a sample of acetanilideshowed a chromatographic peak at 1.88 min when N2 was usedas the collision gas (Figure 3a). A subsequent chromatogramgenerated by an MRM protocol monitoring the m/z 92 → 136transition produced a chromatographic peak at the same reten-tion time confirming the adduct formation when CO2 was usedas the collision gas (Figure 3b).

To determine the effect of ring-substituents on thecarbamate-ion formation reaction, we exposed anions generat-ed from aniline, p-methoxyaniline, or p-nitroaniline to either

Figure 4. Plot of ln(int(M – H + CO2)–/int(M – H)

–) versus the σp– substituent constants [from reference 24] for p-substituted anilines

Scheme 4. Proposed pathway for the formation of the CO2 adducts from para-substituted anilide anions


nitrogen or CO2 as the collision gas. The spectra recordedunder these conditions using N2 as the collision gas showedno peaks for adducts (Supplementary Figure 2a–c). In contrast,the spectrum recorded from the m/z 122 ion generated from p-

methoxyaniline showed an intense peak at m/z 166 (Supple-mentary Figure 2e) for the carbamate ion. A comparison withthe data recorded from the anilide ion as a reference (Supple-mentary Figure 2d) confirmed that the presence of a para-

Figure 5. Negative-ion product ion spectrum recorded from the m/z 136 ion generated by ESI from p-aminobenzoic acid inacetonitrile/water (a) and that from them/z 164 ion generated from ethyl 4-aminobenzoate in acetonitrile/water (b), using CO2 as thecollision gas

Table 3. Proton Affinities (PA) Reporteda for Most Stable Conjugate Bases of Some para-Substituted Anilines, Hammett Valuesb for Substituent Groups, andCalculated Natural Charges on the Anilide-Nitrogen atomc

R group PA of conjugate base (kcal/mol)a Hammett value (σp–)b of substituent Natural charge on the N atomc

OCH3 367 –0.26 –0.913H 366 0 –0.893F 364 –0.03 –0.904Cl 360 0.19 –0.881Br N/A 0.25 –0.876NO2 343 1.27 –0.777COOH 343 0.77 –0.826e

COOC2H5 N/A (–COOCH3: 353)d 0.75 –0.816

a –ΔrH0 data obtained for the reaction: [M – H]– + H+ → M, from NIST [25].

b Hammett values for substituent groups from reference [24].c Geometry optimizations of chemical structures and natural charges on nitrogen atom were calculated by the density functional theory method B3LYP with a 6-311++G(2d, 2p) basis set for all atoms, using Gaussian 09 [18, 19].d PA data is not available on NIST for ethyl 4-aminobenzoate; the data of methyl 4-aminobenzoate was taken instead.e The natural charge on nitrogen atomwas calculated for the p-aminobenzoate. In this case, the deprotonation of p-aminobenzoic acid takes place on the COOH group.


methoxy group facilitates the formation of the CO2 adduct. Onthe other hand, the spectrum recorded from the m/z 137 ionfrom a sample of p-nitroaniline exhibited only a minute peak atm/z 181 for the adduct (in fact, the m/z 170–190 region of thespectrum had to be magnified 20 times to render the peakvisible; Supplementary Figure 2f). Clearly, the para-nitrogroup significantly attenuates the adduct formation.

For the addition reaction, the gaseous anilide ion acts as anucleophile and donates an electron pair to the CO2 molecule.In other words, the anilide anion acts as a Lewis base in the ion-

molecule addition reaction. We envisaged that electron-donating or electron-withdrawing properties of the substituentgroups should correlate well with observed relative peak inten-sity differences of the adduct ion. To evaluate this hypothesis, asemi-logarithmic plot of the ratio of the relative peak area of theadduct ion [(M – H) + CO2]

– to the precursor ion [(M – H)]–

was plotted against the Hammett substituent constant valuesdescribed by Hansch [24] (Figure 4, Scheme 4). The Hammettconstant (σ) provides a quantitative evaluation of the totalelectronic effects, consisting of resonance and field inductiveeffects, of a group bound to a benzene ring. A positive σ valueindicates an electron-withdrawing group; a negative value anelectron-donating group. The plot indicated that electron-donating groups, such as –OCH3, at the para-position of thephenyl ring facilitate the CO2 addition process, whereas,electron-withdrawing groups, such as –NO2, significantly at-tenuated the addition reaction. The σp

– values were selectedbecause the resonance contribution of the para-substituentgroup was expected to be the primary parameter that definesthe overall charge density on the nitrogen atom, which in turndetermines the extent of the nucleophilic addition reaction. Inother words, higher charge densities on the nitrogen atom

Table 4. Reported Proton Affinities (PA)a of Most Stable Conjugate Bases ofSome Aromatic Compounds

Aromatic compound PA of conjugate base (kcal/mol)a

Acetanilide 348Benzenesulfonamide 333Benzoic acid 340Phenol 349Benzene 401

a –ΔrH0 data obtained for the reaction: [M – H]– + H+ → M, from NIST [25].

Figure 6. Negative-ion product ion spectra recorded from them/z 156 and 170 ion generated by ESI from benzenesulfonamide (a),(b) and p-toluenesulfonamide (c), (d), respectively. Ions were generated from acetonitrile/water solutions and subjected to dissoci-ation using N2 (a), (c) or CO2 (b), (d) as the collision gas at the collision energy setting of 25 eV


should favor the addition reaction. Although the correlation isnot as good as that observed with the σp

– values, the chargedensities computed by natural population analysis also showedthat the probability of charge to be present on the anilidenitrogen atom is higher for compounds bearing an electron-donating group at the para position (Supplementary Tables 1,10).

Although Hammett values afford some predictability(Figure 4 shows data from six compounds; Supplementary

Figures 2, 3), they are not good universal indicators to predictthe degree of carbamate formation reaction. For example, thepresence of the COOH group (σp

– = 0.77) at the para positionappears to predict a more facile addition to CO2 than for thenitro substituent (σp

– = 1.27). However, in reality, deprotonatedp-nitroaniline is a better nucleophile than p-aminobenzoateanion. This is not a surprise because the deprotonation of p-aminobenzoic acid takes place primarily at the COOH group.Consequently, the product ion spectrum recorded from the

Scheme 5. Proposed formation of the CO2 adduct of the anilide (m/z 92), or p-methylanilide (m/z 106), generated frombenzenesulfonamide, or p-toluenesulfonamide, respectively

Scheme6. Proposed fragmentation pathways of deprotonated ethyl 4-aminobenzoate and formation of the CO2 adduct ofm/z 180


mass-isolated m/z 136 ion derived from p-aminobenzoic acid,using CO2 as the collision gas, shows only a small adduct peakatm/z 180 because the negative charge is located essentially onthe carboxylate group (Figure 5a). However, if the COOHgroup is esterified, the charge is then located primarily on thenitrogen atom. For example, the mass-isolated m/z 164 ionderived from ethyl 4-aminobenzoate, in which the charge isprimarily located on the NH group, undergoes a very facileaddition of CO2 to generate an adduct peak at m/z 208(Figure 5b). Undoubtedly, a high charge density of the nitrogenatom is vital for facile addition of CO2 to the anilide ion.

Apparently, the Hammett values are useful only to predictthe efficacy of anilides with substituent groups that do notcompete with the NH2 group for deprotonation. Even for thecarboxyethyl group, the Hammett value does not rationalize theobserved experimental results (Table 3 and Figure 5b).

Moreover, predictions based on Hammett values are lessmeaningful for deprotonated N-substituted amines. We envis-aged that proton affinity values of corresponding conjugatebases would be more effective to predict the reactivity ofdeprotonated amines toward the CO2 adduct formation. Toevaluate this hypothesis, the proton affinity values reportedfor a variety of anions were surveyed from the data availablein NIST chemistry webbook [25]. For example, the protonaffinities of phenide and anilide anions are 401 and 366 kcal/mol, respectively, and both anions do act as good nucleophilesin gas phase. In contrast, phenoxide and benzoate anions,which bear proton affinity values of 349 and 340 kcal/mol,respectively, do not add well to CO2 (Table 4). In other words,conjugate bases (A–) generated from compounds of low gas-phase acidities act as better nucleophiles for the CO2 additionreaction. The proton affinity values of para-substituted anilineswere also included in the survey (Table 3). When there is anelectron-donating group on the para position, the basicity ofthe substituted anilide anion is higher than that of unsubstitutedanilide anion. It appears that conjugate bases with protonaffinities less than about 350 kcal/mol do not add well to forma gaseous carbamate ion. For example, the proton affinity valueof deprotonated p-nitroaniline is 343 kcal/mol, and it adds onlypoorly to CO2 (Supplementary Figure 2f).

The product-ion spectra recorded from a sample ofbenzenesulfonamide or p-toluenesulfonamide showed peaksof extremely minute intensity for direct addition when CO2

was used as the collision gas at a collision energy setting of5 eV [PA of deprotonated benzenesulfonamide = 333 kcal/mol;(Supplementary Figure 4b, d)]. However, when the collisionenergy was increased to 25 eV, two new intense peaks ap-peared at m/z 136 and 150, respectively, in the spectra ofbenzenesulfonamide and p-toluenesulfonamide (Figure 6). Itis well known that once activated, deprotonated sulfonamideseliminate SO2 by an intramolecular rearrangement [26–29].The anilide anions generated as products by this rearrangementprocess bear higher proton affinities than their precursors (PAof deprotonated aniline = 366 and that of deprotonatedbenzenesulfonamide = 333 kcal/mol). Once SO2 is eliminated,the product ions bear less internal energy and undergo imme-diate addition to CO2 in the collision cell (Figure 6, Scheme 5).

Furthermore, the product ion spectrum recorded using N2 asthe collision gas from deprotonated ethyl 4-aminobenzoateshowed two distinct fragment-ion peaks at m/z 135 and 136,attributed to ethyl radical and ethylene losses, respectively [21](Supplementary Figure 5a). When the collision gas waschanged to CO2, two new peaks appeared in the spectrum(Supplementary Figure 5b). In addition to the expected peakfor the addition of CO2 to the m/z 164 ion, a small but signif-icant peak also appeared at m/z 180. Presumably, the m/z 136ion formed by the ethylene loss adds CO2 to form the m/z 180ion (Scheme 6). Interestingly, the distonic radical anion formedby the ethyl radical loss is not sufficiently nucleophilic to add toCO2 because a peak was not observed at m/z 179 (Supplemen-tary Figure 5b). It was conclusive that the distonic radical aniondoes not react with CO2 because no peak was observed at m/z179 even when the in-source generated m/z 135 ion was mass-isolated by the first quadrupole, and exposed to CO2 in thecollision cell (Supplementary Figure 5c). According to computa-tions, the natural charge and the Mulliken spin density on thenitrogen atom of the expected distonic radical anion were –0.61 eand 0.54, respectively (Supplementary Table 18). In other words,in the proposed structure of the m/z 135 ion, the charge center islocated on the carboxyl group rendering the ion less nucleophilic.

Scheme 7. Proposed fragmentation pathways of the CO2 adduct of p-methoxyaniline


The product ion spectrum recorded using CO2 as the colli-sion gas from themass-isolatedm/z 107 radical anion generatedfrom p-methoxyaniline showed an adduct-ion peak at m/z 151(Supplementary Figure 6c, Scheme 7). According to computa-tions, the natural charge on the nitrogen atom (–0.74 e) was alittle higher than that on the nitrogen atom of the m/z 135 ionfrom deprotonated ethyl 4-aminobenzoate (–0.61 e). Moreover,the spin density on the nitrogen atom was 0.42 compared withthat on the nitrogen atom of the m/z 135 ion from deprotonatedethyl 4-aminobenzoate (0.54) (Supplementary Table 11). Incontrast, the charge center of the m/z 107 distonic ion isprimarily on the nitrogen atom (–0.74 e); thus, the additionreaction is expected to take place to a small extent (Supple-mentary Figure 6c). These experimental results further confirmthat a high charge density of the nitrogen atom is a prerequisitefor facile addition of CO2 to the anilide ion.

As mentioned previously, the addition of CO2 to anilideanion is an exergonic reaction. Computations showed that thepresence of an electron donating substituent increases theexergonicity (Table 5). When the exergonicity was less thanca. 10 kcal/mol, a peak was not observed for the product underthe experimental conditions used.

Experiments were conducted to determine if CO2 admixedanother inert gas furnishes more or less efficient carbamate ionformation in the collision cell. For this purpose, product-ion spectrawere recorded under negative-ion conditions using the m/z 92 iongenerated from aniline as the precursor ion. The partial pressure ofCO2 was adjusted to the same value for each experiment. Spectrarecorded using CO2, or CO2/He, CO2/N2, and CO2/Ar mixtures asthe collision gas are shown in Supplementary Figure 7, whichillustrate that CO2/He provides a more efficient mixture. Presum-ably, helium acts as cooling gas and lowers the kinetic energy of

Table 5. Gibbs Free Energy Calculated for CO2 Addition Reactions of Some Deprotonated para-Substituted Anilines and Other Aromatic Compounds Containingan NH Group

Compound Calculated Gibbs free energy for the CO2

addition reaction (kcal/mol)

CO2 adduct observed or notb

R = OCH3 -28.77 Yes, intenseR = H -27.06 Yes, intenseR = F -27.24 Yes, intenseR = Cl -24.71 Yes, moderateR = Br -24.11 Yes, moderateR = COOC2H5 -18.24 Yes, moderateR = NO2 -13.37 Yes, but very weakR = COOH -17.70a Yes, but very weak

R = O.

(derived from p-methoxyaniline)

-13.28 Yes, but very weak

R = COO.

(derived from ethyl 4-aminobenzoate)

-1.08 No

Aromatic compounds containing an NH groupBenzenesulfonamide -7.93 Yes, but extremely weakAcetanilide -5.30 No

a The deprotonation of p-aminobenzoic acid primarily takes place on the COOH group. However, the Gibbs free energy was calculated for the anions, thedeprotonation of which takes place on the NH2 group.b Red color denotes an adduct peak is expected. Green color denotes an adduct peak is not expected.


the anilide ions and promotes the addition reaction when amixtureof CO2 and He is used as the collision gas.

ConclusionsWe have demonstrated that gaseous anilide anions add readilyto CO2 and form carbamate ions. Computations showed that itis an exergonic reaction (ΔG ca. –27 kcal/mol). The electrondensity on the nitrogen atom of the anilide ion determines theefficacy of the reaction. The presence of electron-donatinggroups attached to the ring increases the electron density onthe nitrogen and facilitates the addition process. In practice, apeak for the adduct was not observed if the calculated Gibbsfree energy release for the reaction was less than ca.10 kcal/mol.

AcknowledgmentsThe authors thank Bristol-Myers Squibb (Princeton, NJ) for thedonation of the Sciex API 3000 tandem quadrupole massspectrometer, and John Ressler for technical support.

References1. Tlili, A., Blondiaux, E., Frogneux, X., Cantat, C.: Reductive functionalization of

CO2 with amines: an entry to formamide, formamidine, and methylaminederivatives. Green Chem. 17, 157–168 (2015)

2. Halmann, M.: Chemical Fixation of CO2. CRC, Boca Raton (1993)3. Leung, D.Y.C., Caramanna, G., Maroto-Valer, M.M.: An overview of

current status of carbon dioxide capture and storage technologies. Renew.Sust. Energ. Rev. 39, 426–443 (2014)

4. Jackson, P., Fisher, K.J., Attalla, M.I.: Tandem mass spectrometry mea-surement of the collision products of carbamate anions derived from CO2

capture sorbents: paving the way for accurate quantitation. J. Am. Soc.Mass Spectrom. 22, 1420–1431 (2011)

5. Bierbaum, V.M., DePuy, C.H., Shapiro, R.H.: Gas phase reactions ofanions with nitrous oxide and carbon dioxide. J. Am. Chem. Soc. 99,5800–5802 (1977)

6. Danikiewicz, W., Zimnicka, M.: Negative ion gas-phase chemistry ofarenes. Mass Spectrom. Rev. 35, 123–146 (2016)

7. Bienkowski, T., Danikiewicz, W.: Generation and reactions of substitutedphenide anions in an electrospray triple quadrupole mass spectrometer.Rapid Commun. Mass Spectrom. 17, 697–705 (2003)

8. Kumar, M.K., Sateesh, B., Prabhakar, S., Sastry, G.N., Vairamani, M.:Generation of regiospecific carbanions under electrospray ionization con-ditions and their selectivity in ion-molecule reactions with CO2. RapidCommun. Mass Spectrom. 20, 987–993 (2006)

9. Liu, C., Zhang, Y., Attygalle, A.B.: Selectivity of gas-phase ion/moleculereaction of carbon dioxide with phenide ions. J. Mass Spectrom. 49, 692–699 (2014)

10. DePuy, C.H., Grabowski, J.J., Bierbaum, V.M.: Chemical reactions ofanions in the gas phase. Science 218, 955–960 (1982)

11. DePuy, C.H., Bierbaum, V.M., Robinson,M.S., Davico, G.E., Gareyev, R.:The gas-phase reactions of the allenyl anion with CS2, COS, and CO2.Tetrahedron 53, 9847–9856 (1997)

12. Seburg, R.A., Hill, B.T., Jesinger, R.A., Squires, R.R.: The phenylacrbyneanion. J. Am. Chem. Soc. 121, 6310–6311 (1999)

13. Okada, S., Abe, Y., Yamabe, S.: Potential surfaces of Gibbs free energies ofgas-phase NH2

– + N2O and NH2– + CO2 reactions. J. Phys. Chem. 99,

16877–16882 (1995)14. Bierbaum, V.M., Grabowski, J.J., DePuy, C.H.: Gas-phase synthesis and

reactions of nitrogen- and sulfur-containing anions. J. Chem. Phys. 88,1389–1393 (1984)

15. DePuy, C.H.: Fragmentation of organic anions induced by exothermicaddition reactions. Org. Mass Spectrom. 20, 556–559 (1985)

16. Gronert, S., O’Hair, R.A.J., Prodnuk, S., Sulzle, D., Damrauer, R., DePuy,C.H.: Gas-phase chemistry of the silaformyl anion, HSiO–. J. Am. Chem.Soc. 112, 997–1003 (1990)

17. Kumar, M.K., Prabhakar, S., Kumar, M.R., Vairamani, M.: Auxiliaryapproach to evaluate the isomeric decarboxylated anions from 2-, 3-, and4-sulfobenzoates in the gas phase by using ion-molecule reactions withcarbon dioxide in the collision cell. Rapid Commun. Mass Spectrom. 20,1045–1048 (2006)

18. Frisch, M.J., Trucks, G.W., Schlegel, H.B., Scuseria, G.E., Robb, M.A.,Cheeseman, J.R., Scalmani, G., Barone, V., Mennucci, B., Petersson, G.A.,Nakatsuji, H., Caricato,M., Li, X., Hratchian, H.P., Izmaylov, A.F., Bloino,J., Zheng, G., Sonnenberg, J.L., Hada, M., Ehara, M., Toyota, K., Fukuda,R., Hasegawa, J., Ishida,M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H.,Vreven, T., Montgomery Jr., J.A., Peralta, J.E., Ogliaro, F., Bearpark, M.,Heyd, J.J., Brothers, E., Kudin, K.N., Staroverov, V.N., Keith, T.,Kobayashi, R., Normand, J., Raghavachari, K., Rendell, A., Burant, J.C.,Iyengar, S.S., Tomasi, J., Cossi, M., Rega, N., Millam, J.M., Klene, M.,Knox, J.E., Cross, J.B., Bakken, V., Adamo, C., Jaramillo, J., Gomperts, R.,Stratmann, R.E., Yazyev, O., Austin, A.J., Cammi, R., Pomelli, C.,Ochterski, J.W., Martin, R.L., Morokuma, K., Zakrzewski, V.G., Voth,G.A., Salvador, P., Dannenberg, J.J., Dapprich, S., Daniels, O., Farkas,A.D., Foresman, J.B., Ortiz, J.V., Cioslowski, J., Fox, D.J.: Gaussian 09,Revision D.01. Gaussian, Inc, Wallingford (2013)

19. Becke, A.D.: Density-functional exchange-energy approximation with cor-rect asymptotic behavior. J. Chem. Phys. 98, 5648–5652 (1993)

20. Kouris, C.S., Northcott, J.: Aniline and its derivatives. In: Kirk, R.E.,Othmer, D.F. (eds.) Encyclopedia of Chemical Technology, Vol. II, 2ndedn, p. 411. Wiley, New York (1963)

21. Xia, H., Zhang, Y., Pavlov, J., Jariwala, F.B., Attygalle, A.B.:Competitive homolytic and heterolytic decomposition pathways ofgas-phase negative-ions generated from aminobenzoate esters. J.Mass Spectrom. (2016). doi:10.1002/jms.3740

22. Vakula, N.I., Kuramshina, G.M., Makhmutova, S.F., Pentin, Y.A.: DFTtheoretical studies of anions of aniline and its several derivatives. Struct.Chem. 22, 345–356 (2011)

23. Gronert, S., Garver, J.M., Nichols, C.M., Worker, B.B., Bierbaum,V.M.: Dehalogenation of arenes via SN2 reactions at bromine:competition with nucleophilic aromatic substitution. J. Org. Chem.79, 11020–11028 (2014)

24. Hansch, C., Leo, A., Taft, R.W.: A survey of Hammett substituentconstants and resonance and field parameters. Chem. Rev. 91, 165–195 (1991)

25. Bartmess, J.E.: Negative ion energetics data. In: Linstrom, P.J., Mallard,W.G. (eds.) NIST Chemistry WebBook, NIST Standard Reference Data-base Number 69. National Institute of Standards and Technology, Gaithers-burg MD, 20899. Available at: http://webbook.nist.gov. Accessed 23 Oct2015

26. Hibbs, J.A., Jariwala, F.B., Weisbecker, C.S., Attygalle, A.B.: Gas-phasefragmentations of anions derived from N-phenyl benzenesulfonamides. J.Am. Soc. Mass Spectrom. 24, 1280–1287 (2013)

27. Wang, Z., Hop, C.E.C.A., Kim, M.-S., Huskey, S.-E.W., Baillie, T.A.,Guan, Z.: The unanticipated loss of SO2 from sulfonamides in collision-induced dissociation. Rapid Commun. Mass Spectrom. 17, 81–86 (2003)

28. Hu, N., Liu, P., Jiang, K., Zhou, Y., Pan, Y.: Mechanism study of SO2

elimination from sulfonamides by negative electrospray ionization massspectrometry. Rapid Commun. Mass Spectrom. 22, 2715–2722 (2008)

29. Xiang, Z.: Mechanism of SO2 elimination from the aromatic sulfonamideanions: a theoretical study. Comp. Theor. Chem. 991, 74–81 (2012)


http://dx.doi.org/10.1002/jms.3740

http://webbook.nist.gov/

formation of carbamate anions by the gas-phase reaction of ... · formation of carbamate anions by...

Documents