The Peroxymonocarbonate Anion HCO4− as an Effective Oxidant in the Gas Phase: A Mass Spectrometric and Theoretical Study on the Reaction with SO2

The peroxymonocarbonate anion, HCO4−, the covalent adduct between the carbon dioxide and hydrogen peroxide anion, effectively reacts with SO2 in the gas phase following three oxidative routes. Mass spectrometric and electronic structure calculations show that sulphur dioxide is oxidised through a common intermediate to the hydrogen sulphate anion, sulphur trioxide, and sulphur trioxide anion as primary products through formal HO2−, oxygen atom, and oxygen ion transfers. The hydrogen sulphite anion is also formed as a secondary product from the oxygen atom transfer path. The uncommon nucleophilic behaviour of HCO4− is disclosed by the Lewis acidic properties of SO2, an amphiphilic molecule that forms intermediates with characteristic and diagnostic geometries with peroxymonocarbonate.

In contrast to the wealth of studies performed in solution, in the gas phase, to the best of our knowledge, only one study has reported the reactivity of an isobaric electrostatic ionic complex [HO 2 ···CO 2 ] − towards atmospheric constituents [24], suggesting the possible occurrence of a bound species, the peroxymonocarbonate ion.
In the gas phase, the peroxymonocarbonate anion has been known since 1977, when it was observed in a flame ionisation detector [25,26] upon clustering reactions involving free oxygen radicals with CO 2 or carbonate derivatives, and more recently in air plasma [27,28], either as a free species or as a negative core ion of large water clusters [29]. The lack of gasphase reactivity studies concerning HCO 4 − ions prompted us to investigate the reactions of this anionic peracid in a joint mass spectrometric and theoretical approach. Over the years, this approach has provided a wealth of knowledge on the intrinsic reactivity of isolated ions towards neutral species in a pristine environment, i.e., in the absence of the masking effects of solvents or counter-ions typically operating in bulk environments [30][31][32]. We therefore undertook an investigation of the oxidative behaviour of HCO 4 − towards sulphur dioxide by ion trap mass spectrometry, exploiting an in-house-modified experimental setup to perform ion-molecule reactions [33]. This experimental approach allows one to precisely ascertain the identity of the reactive (oxidant) species, as well as the relationship between the reagent (oxidant), intermediates and the products involved [30,31]. In the last few years, we have thoroughly investigated the reactivity of sulphur dioxide, one of the most produced commodity chemicals worldwide [34] and, at the same time, a harmful pollutant [35], with singly [36,37] and doubly charged [38][39][40] metal oxide anions and non-metal anionic oxides [41][42][43][44]. The very rich chemistry observed is the result of the amphoteric nature of SO 2 , whose sulphur atom can either behave as an electron acceptor or as an electron donor. The nucleophilic ability is due to the lone pair on sulphur in a high-lying σ-based HOMO, whereas the electrophilic properties are due to the low-lying π-symmetry LUMO of sulphur dioxide [45]. Interestingly, the donor-acceptor behaviour of SO 2 is reflected in the diagnostic geometrical features of non-metal [46] and metal-SO 2 complexes [45,47].
We found that SO 2 was effectively oxidised by HCO 4 − to hydrogen sulphate, sulphur trioxide, and SO 3 − , together with carbon dioxide, hydrogen carbonate, and the hydrogen carbonate radical. Hydrogen sulphite was also observed as a result of a secondary reaction promoted by HCO 3 − [48]. Our study adds a new contribution to the gas-phase chemistry of both HCO 4 − and sulphur dioxide, highlighting the nucleophilic oxidative properties of peroxymonocarbonate, which are disclosed thanks to the electrophilic character of sulphur dioxide. A brief description of the generation of the anionic peroxide HCO 4 − by electrospray ionisation (ESI) in our experimental set-up is also given, whereas experiments on sulphur-containing molecules of atmospheric, environmental, and biological interest are ongoing in our laboratory.
The first approach is advantageous as sodium percarbonate is known to be a stable and easy-to-handle, commercially available compound, providing at once proper amounts of HCO 3 − and H 2 O 2 in water. The peroxymonocarbonate ion was indeed formed in Na 2 CO 3 ·1.5 H 2 O 2 solution and gently transferred to the gas-phase environment for structural and reactivity investigations.
The second option is also convenient since, by injecting pure water into the ESI source, line saturation effects and the need for frequent cleaning operations are avoided. Furthermore, the application of in-source electric fields to assess the ionisation process lead to the formation of different oxidising agents [49] that react with O 2 or CO 2 typically found in the ambient air or dissolved in water in different forms (CO 2(aq) , HCO 3 − ). The ESI approach is also a gentler ionisation method than atmospheric pressure chemical ionisation (APCI), which permits the application of a negative corona discharge in room air to obtain the HCO 4 − ion [27][28][29]50]. According to these studies, we also succeeded in generating the peroxymonocarbonate ion through the APCI of ambient air, but the electrospray ionisation of water proved to be the best approach to assess the stable ionic signal of the HCO 4 − species, which is a prerequisite for kinetic measurements. Interestingly, to the best of our knowledge, the formation of the HCO 4 − ion under these conditions has never been emphasised by other authors, since the studies analysing the ESI speciation of pure water mainly investigate the aggregation phenomena of H + (H 2 O) n and OH − (H 2 O) n cluster ions [51].
Focusing on the gas-phase formation pathways of the HCO 4 − ion, two different reactions were proposed, namely the three-body association reactions of (1) CO 3 − + OH • + M and (2) HO 2 − + CO 2 + M [25][26][27][28][29]. In our experimental conditions, the ESI-(-) mass spectrum of pure water acquired in the low 20-150 mass range shows (Figure 1 [52], is present only in a very low amount, as also already reported in the corona discharge ionisation systems by Ninomiya et al. [28], together with other ionic species also present in the ESI-(-) mass spectrum, but not involved in the formation of the HCO 4 − ion.
The second option is also convenient since, by injecting pure water into the ESI source, line saturation effects and the need for frequent cleaning operations are avoided. Furthermore, the application of in-source electric fields to assess the ionisation process lead to the formation of different oxidising agents [49] that react with O2 or CO2 typically found in the ambient air or dissolved in water in different forms (CO2(aq), HCO3 − ). The ESI approach is also a gentler ionisation method than atmospheric pressure chemical ionisation (APCI), which permits the application of a negative corona discharge in room air to obtain the HCO4 − ion [27][28][29]50]. According to these studies, we also succeeded in generating the peroxymonocarbonate ion through the APCI of ambient air, but the electrospray ionisation of water proved to be the best approach to assess the stable ionic signal of the HCO4 -species, which is a prerequisite for kinetic measurements. Interestingly, to the best of our knowledge, the formation of the HCO4 -ion under these conditions has never been emphasised by other authors, since the studies analysing the ESI speciation of pure water mainly investigate the aggregation phenomena of H + (H2O)n and OH − (H2O)n cluster ions [51].
Focusing on the gas-phase formation pathways of the HCO4 − ion, two different reactions were proposed, namely the three-body association reactions of (1) CO3 − + OH • + M and (2) HO2 − + CO2 + M [25][26][27][28][29]. In our experimental conditions, the ESI-(-) mass spectrum of pure water acquired in the low 20-150 mass range shows (Figure 1), among others, intense ionic signals at m/z 32 (O2 •− ), at m/z 50 (the hydrated cluster ion [O2•H2O] •− ), and at m/z 60 corresponding to the CO3 − ion. The HO2 − species at m/z 33, commonly obtained through the discharge ionisation of H2O/O2 mixtures [52], is present only in a very low amount, as also already reported in the corona discharge ionisation systems by Ninomiya et al. [28], together with other ionic species also present in the ESI-(-) mass spectrum, but not involved in the formation of the HCO4 − ion. To evaluate the possible formation pathway occurring under ESI conditions, we carried out two different experiments. In the first, we infused aliquots of water (previously decarbonated) enriched with Na2 13 CO3 or with NaH 13 CO3 as a source of 13 CO3 − ion. As a result, the HCO4 − ion at m/z 77 always largely predominates over the labelled H 13 CO4 − species at m/z 78, even if the H 13 CO3 − ion represents the base peak, thus excluding the CO3 − + OH ⸱ + M reaction as a possible HCO4 − formation pathway. To evaluate the possible formation pathway occurring under ESI conditions, we carried out two different experiments. In the first, we infused aliquots of water (previously decarbonated) enriched with Na 2 13 CO 3 or with NaH 13 CO 3 as a source of 13 CO 3 − ion. As a result, the HCO 4 − ion at m/z 77 always largely predominates over the labelled H 13 CO 4 − species at m/z 78, even if the H 13 CO 3 − ion represents the base peak, thus excluding the CO 3 − + OH · + M reaction as a possible HCO 4 − formation pathway. In the second experiment, we analysed an aliquot of water previously degassed to eliminate the dissolved CO 2 and subsequently placed in a closed vessel saturated with 13 CO 2 (P 13 CO 2 = 700 torr). Although the occurrence of the 13 CO 2(g) 13 CO 2(aq) H 13 CO 3 − (aq) equilibrium in water is unavoidable [53], a strong increase in the H 13 CO 4 − ion compared to its 12 C isotopologue was observed, highlighting the occurrence of the HO 2 − + CO 2 + M reaction in the ESI generation of the peroxymonocarbonate ion. According to this hypothesis, the neutral CO 2 , rather than CO 3 − or HCO 3 − ions, is supposed to be the oxidation target, as demonstrated by the absence of an increase in the HCO 4 − ionic signal when water is enriched with bicarbonate or carbonate species, as previously described.
Prior to investigating its gas-phase reactivity (see next paragraph), the HCO 4 − species was mass-selected and characterised by collision-induced dissociation (CID) experiments. The HCO 4 − ion at m/z 77 fragments by releasing a hydroxyl radical counterpart (OH • ; 17 Da) [28], leading to the formation of a carbonate radical anion (CO 3 •− ) at m/z 60 ( Figure 1, upper inset), which accounts for the higher EA of CO 3 • with respect to HO • [54]. Likewise, the fragmentation of the 13 C-isotopologue H 13 CO 4 − at m/z 78 gives rise to the corresponding 13 CO 3 •− species at m/z 61 ( Figure 1, lower inset).

Mass Spectrometric Results
Regardless of the preparation method (if formed by electrospray ionisation of a percarbonate solution, of water, or of a saturated solution of 13 CO 2 ), the peroxymonocarbonate HCO 4 − and its heavy 13 C-form predictably show superimposable reactivity towards sulphur dioxide. They were mass-selected and exposed to SO 2 in the ion trap cell to monitor the formation of ionic products as a function of the activation time. A typical product ion spectrum is reported in Figure 2, showing the formation of ionic signals at m/z 97, 81, 80, and 61, respectively, attributed to HSO 4 − , HSO 3 − , SO 3 − , and HCO 3 − . Ions at m/z 97 and 81 clearly contain one sulphur atom, as witnessed by the characteristic signature of the 34  eliminate the dissolved CO2 and subsequently placed in a closed vessel saturated with 13 CO2 (P 13 CO2 = 700 torr). Although the occurrence of the 13 CO2(g) ⇄ 13 CO2(aq) ⇄ H 13 CO3 − (aq) equilibrium in water is unavoidable [53], a strong increase in the H 13 CO4 − ion compared to its 12 C isotopologue was observed, highlighting the occurrence of the HO2 − + CO2 + M reaction in the ESI generation of the peroxymonocarbonate ion. According to this hypothesis, the neutral CO2, rather than CO3 − or HCO3 − ions, is supposed to be the oxidation target, as demonstrated by the absence of an increase in the HCO4 − ionic signal when water is enriched with bicarbonate or carbonate species, as previously described. Interestingly, our results parallel the evidence of the intermediacy of CO2 in bicarbonate/hydrogen peroxide solutions, where HCO4 − is formed through CO2 perhydration or base-catalysed perhydration [6,8,11].
Prior to investigating its gas-phase reactivity (see next paragraph), the HCO4 − species was mass-selected and characterised by collision-induced dissociation (CID) experiments. The HCO4 − ion at m/z 77 fragments by releasing a hydroxyl radical counterpart (OH • ; 17 Da) [28], leading to the formation of a carbonate radical anion (CO3 •− ) at m/z 60 ( Figure 1, upper inset), which accounts for the higher EA of CO3 • with respect to HO • [54]. Likewise, the fragmentation of the 13 C-isotopologue H 13 CO4 − at m/z 78 gives rise to the corresponding 13 CO3 •− species at m/z 61 ( Figure 1, lower inset).

Mass Spectrometric Results
Regardless of the preparation method (if formed by electrospray ionisation of a percarbonate solution, of water, or of a saturated solution of 13 CO2), the peroxymonocarbonate HCO4 − and its heavy 13 C-form predictably show superimposable reactivity towards sulphur dioxide. They were mass-selected and exposed to SO2 in the ion trap cell to monitor the formation of ionic products as a function of the activation time. A typical product ion spectrum is reported in Figure 2, showing the formation of ionic signals at m/z 97, 81, 80, and 61, respectively, attributed to HSO4 − , HSO3 − , SO3 − , and HCO3 − . Ions at m/z 97 and 81 clearly contain one sulphur atom, as witnessed by the characteristic signature of the 34   The kinetic profile of the reaction (Figure 3) shows that products are formed through parallel and consecutive reactions, and that the precursor ion conversion is fast and efficient, with a decay constant k dec at 298 K of 7.2 × 10 −10 cm 3 molecule −1 s −1 and efficiency (k/k coll ) of 54.8% (Table 1). Three parallel reaction pathways can be identified in the following: Molecules 2023, 28, 132 5 of 14 (k/kcoll) of 54.8% (Table 1). Three parallel reaction pathways can be identified in the following:  The main reaction channel 1 consists of a formal SO2/CO2 switching, leading to the formation of the bisulphate HSO4 − product ion at m/z 97, accounting for a branching ratio of 64.5% (k1 = 4.6 × 10 −10 cm 3 molecule −1 s −1 ) and efficiency of 35.1% (Table 1). From another point of view, this reaction can also be considered as a formal hydroperoxide anion HOO − transfer. Pathway 2 gives rise to the HCO3 − ion at m/z 61 with the release of SO3 as the neutral counterpart through an oxygen atom transfer (OAT) reaction to SO2 occurring with a rate constant of k2 = 2.5 × 10 −10 cm 3 molecule −1 s −1 (eff. = 19.1%) and a branching ratio of 34.8%. Finally, the very minor path 3, accounting for less than 1% of the products, leads to SO3 − at m/z 80, with a k3 = 5.0 × 10 −12 cm 3 molecule −1 s −1 (reaction 3, eff. = 0.4%) through oxygen ion transfer (OIT). The preferential formation of products from reaction 2 with  14.0 c >100 a ±30%; b k/k coll ; c see ref. [48].
The main reaction channel 1 consists of a formal SO 2 /CO 2 switching, leading to the formation of the bisulphate HSO 4 − product ion at m/z 97, accounting for a branching ratio of 64.5% (k 1 = 4.6 × 10 −10 cm 3 molecule −1 s −1 ) and efficiency of 35.1% (Table 1). From another point of view, this reaction can also be considered as a formal hydroperoxide anion HOO − transfer. Pathway 2 gives rise to the HCO 3 − ion at m/z 61 with the release of SO 3 as the neutral counterpart through an oxygen atom transfer (OAT) reaction to SO 2 occurring with a rate constant of k 2 = 2.5 × 10 −10 cm 3 molecule −1 s −1 (eff. = 19.1%) and a branching ratio of 34.8%. Finally, the very minor path 3, accounting for less than 1% of the products, leads to SO 3 − at m/z 80, with a k 3 = 5.0 × 10 −12 cm 3 molecule −1 s −1 (reaction 3, eff. = 0.4%) through oxygen ion transfer (OIT). The preferential formation of products from reaction 2 with respect to those of reaction 3 accounts for the considerably higher electron affinity of the bicarbonate radical (EA HCO 3 = 3.68 eV) [55] with respect to SO 3 (EA SO 3 = 1.90 eV) [54].
The HCO 3 − ion formed by reaction 2 is known to be reactive towards SO 2 to form HSO 3 − [48], and indeed it is rapidly depleted in a consecutive reaction forming the bisulphite anion at m/z 81 and carbon dioxide (Equation (4)), as illustrated in Figure 2b.
This process occurs at a collision rate, with a measured rate constant k 4 of 1.40 × 10 −9 cm 3 molecule −1 s −1 (Table 1), in good agreement with previous data [48]. Spectra relative to the reaction of the H 13 CO 4 − ion are reported in Figure S1 (Supplementary Materials). The reaction of HCO 4 − + SO 2 → HSO 4 − + CO 2 described in Equation (1) is an irreversible process, since no displacement of SO 2 by CO 2 occurred when the HSO 4 − ion was reacted with CO 2 , not even with the high pressure of carbon dioxide or extension of the activation time up to the maximum value of 10 s. Similar irreversible switching reactions were also observed in the gas phase by reacting carbonate cluster ions with SO 2 and resulting in the complete conversion of carbonate into sulphite ions [43].

Theoretical Calculations
The oxidation routes of SO 2 promoted by HCO 4 − were investigated using density functional calculations. Optimised geometries of relevant stationary points and transition states on the B3LYP/aug-cc-pV(T+d)Z potential energy surface for the processes are reported in Figure 4, while charge density distributions of relevant minima are shown in Figure 5. Complete geometrical parameters of minima and saddle points are shown in Figure S2 (Supplementary Materials). Enthalpy changes (kcal mol −1 ) at 298.15 K computed at the B3LYP/aug-cc-pV(T+d)Z and CCSD(T)/aug-cc-pV(T+d)Z levels of theory are detailed in Table 2. In the following, only the CCSD(T) values will be considered.       HCO 4 − features a five-membered-ring intramolecular hydrogen bond between the carbonyl oxygen and the hydrogen atom bound to the peroxidic oxygen [28]. The initial attack of the peroxymonocarbonate anion involves this oxygen atom interacting with the incoming SO 2 and leads to an encounter complex (MIN1) stabilised by a binding energy of 12.7 kcal mol −1 . The two moieties, at a distance of 2.37 Å, almost conserve their original geometric features, with SO 2 binding to the peroxyl oxygen in a bent coordination mode: the angle between the peroxyl O-S bond vector and the SO 2 plane is around 100 • , with the dihedral OSOO angle equal to 102.1 • . With a negative small activation barrier, MIN1 converts to a slightly more stable MIN2, in which the proton is now bound to the carbonyl oxygen. The saddle point TS12 is highly stabilised by the sharing of the proton between the two oxygen atoms. In MIN2, the peroxidic O-S bond shortens to 2.04 Å, developing a small negative charge on SO 2 (−0.27 e − ) and still binding with a bent geometry (angle between peroxyl O-S bond vector and the SO 2 plane is 100 • , with the OSOO angle equal to 100.1 • ). In MIN2, an incipient SO 3 moiety can be recognised and the charge density distribution on the two entities, SO 3 and HCO 3 , is − 0.83 e − on the former and −0.17 e − on the latter, therefore defining a sulphur trioxide anion and a bicarbonate radical. Proceeding along the reaction coordinate, MIN2 easily converts into MIN3, overcoming a barrier of only 5.7 kcal mol −1 . In TS23, the easy formation of an S-O bond and the breaking of the O-O bond lead to a pyramidal SO 3 and an HCO 3 moiety. From this saddle point, a second oxygen atom is transferred from the carbonate to the sulphur trioxide moiety, forming an incipient sulphate in MIN3. This minimum is highly stabilised by 88.3 kcal mol −1 ( Table 2) with respect to the entrance asymptote, thanks to the formation of a covalent S-O bond in the SO 3 moiety and to the strong interaction of the sulphur atom with the oxygen of the bicarbonate, preluding at the formation of a sulphate species. MIN3 is further stabilised by an intramolecular hydrogen bond, forming a six-membered ring. The charge density distribution of this stationary point is almost equally distributed between the SO 3 (−0.47 e − ) and HCO 3 (−0.53 e − ) groups; nonetheless, in MIN3, one can also consider a (distorted) tetrahedral SO 4 and a bent HCO 2 moiety. Therefore, two competing paths leading to the final products can be envisaged from this point. The first is the dissociation of MIN3 into HCO 3 − + SO 3 or into HCO 3 + SO 3 − , which is driven by the difference in AE between HCO 3 and SO 3 . Accordingly, the exit path into HCO 3 − + SO 3 experimentally observed (reaction 2) largely prevails over the product couple HCO 3 + SO 3 − , accounting for the much higher electron affinity of the bicarbonate radical (EA HCO 3 • = 3.68 eV) [55] with respect to SO 3 (EA SO 3 = 1.90 eV [54], 2.22 eV [56]). The difference in the EA values of SO 3 is due to the uncertainty attached to the heat of formation of SO 3 − [54,56]. Accordingly, the enthalpy of reaction 3 is slightly positive (see also Table 2), which accounts for the low branching observed. A quite high dissociation energy of MIN3 (59.4 kcal mol −1 , Table 2) is required both to break the hydrogen bond and the S-O bond, and to change the SO 3 geometry from pyramidal to trigonal planar, to eventually release sulphur trioxide and the bicarbonate anion. This dissociation energy is, nevertheless, provided by the exothermicity of the process, calculated to be 28.9 kcal mol −1 (Table 2). Alternatively, MIN3 can isomerise with only 5.3 kcal mol −1 to the isoenergetic MIN4, a loose ion-neutral complex between the hydrogen sulphate anion and CO 2 , from which only 8.1 kcal mol −1 is required to release the products. The products HSO 4 − /CO 2 are formed in a very exothermic process, calculated to be 80.9 kcal mol −1 ( Table 2). Regarding the competition between the two product couples, HSO 4 − /CO 2 and HCO 3 − /SO 3 , the experiments revealed a branching ratio of 3:1. Accordingly, the most stable HSO 4 − /CO 2 products are also formed through the faster path, occurring through a reaction coordinate developing well under the entrance asymptote, whereas HCO 3 − /SO 3 , although forming from direct dissociation, requires a quite large amount of energy to attain the geometry of the products.

Discussion
Sulphur dioxide is effectively oxidised in the gas phase by the peroxymonocarbonate ion to hydrogen sulphate, sulphur trioxide, and, in a very minor amount, to SO 3 − . The reactions proceed through formal HOO − , oxygen atom, and oxygen ion transfers, respectively. In the absence of solvent molecules, the reaction is found to proceed along a PES initially characterised by the formation of two encounter complexes (MIN1 and MIN2, Figure 4), in which the two reactants, SO 2 and HCO 4 − , maintain their geometric features and charge distribution. Nevertheless, the geometry of these adducts can be taken as a diagnostic of the bonding mode of sulphur dioxide, i.e., whether SO 2 behaves as a donor or acceptor molecule. In fact, in analogy with known complexes, metallic [46,47] or non-metallic [47], when SO 2 acts as a Lewis acid, it forms complexes with bent geometries, while linear geometries characterise SO 2 , which behaves as a Lewis base.
In agreement, the coordination of HCO 4 − to SO 2 in MIN1 and MIN2 results in a bent geometry of the complexes (see Figure S3), the angle between the O-S bond vector and the plane of SO 2 moiety being around 100 • in both structures, thereby qualifying SO 2 as a Lewis acid, accepting the lone pair of the oxygen atom in the OH group through its lowest unoccupied molecular orbital, and the interaction as HCO 4 − → SO 2 . The binding energies of the two minima, 12.7 and 15.7 kcal mol −1 (Figure 4, Table 2), are also in line with those calculated for complexes with pure donors, such as ammonia or methyl-substituted N-heterocyclic carbene, ranging from 10 kcal mol −1 (NH 3 ·SO 2 ) to 18 kcal mol −1 (NHC·SO 2 ) [47]. Therefore, the Lewis acidic character of SO 2 discloses the nucleophilic nature of HCO 4 − , which has rarely been observed in solution studies [20][21][22] when not questioned [14].
In solution, the generally accepted mechanism of the oxidation promoted by peroxymonocarbonate follows that of electrophilic peroxides [5]. The reaction occurs via a solvent-aided oxygen transfer mechanism, which involves a nucleophilic attack by the substrate, Nu, at the electrophilic oxygen of HCO 4 − with the formation of the oxidised substrate, NuO, and the bicarbonate ion. An example is given by the electrophilic oxidation of sulphides, substrates with clear nucleophilic properties, which selectively leads to sulfoxides [9]. Secondary oxidation to sulfones is due to O 2 •− , generated upon the decomposition of HCO 4 − , or to H 2 O 2 , used to activate NaHCO 3 , rather than the peroxymonocarbonate itself [14]. Much uncertainty regarding the nucleophilic behaviour of HCO 4 − still exists, also regarding the precise nature of the oxidant. Accordingly, the oxidative degradation of organophosphorus compounds, such as Paraoxon, in the NaHCO 3 -activated H 2 O 2 solutions appears to be attributable to HO 2 − or HO − and not to HCO 4 − [14]. This problem is easily circumvented in the gas phase, where it is possible to select a precise oxidant, i.e., HCO 4 − , and study its reactivity towards a selected substrate, i.e., SO 2 . The actual oxidation step of the reaction occurs only in MIN3, in which one peroxidic oxygen atom has been fully transferred to sulphur dioxide, and the second one is still shared by the sulphur and carbon atoms. Accordingly, MIN3 can directly dissociate into HCO 3 − + SO 3 or into HCO 3 + SO 3 − , with a branching that reflects the thermodynamic stability of the products, or it can evolve, through a small barrier, to a loose ion-neutral complex, eventually leading to the separate products HSO 4 − + CO 2 . Hence, the driving force of the major, highly exothermic, oxidation process is the liberation of the stable carbon dioxide molecule. This circumstance can explain the different reactivity observed in the reaction of the strong nucleophile hydroperoxide, HOO − , which reacts with SO 2 at collision rate, forming SO 3 − and only a very small quantity of an adduct of formula HOOSO 2 − and of unspecified structure [57].
Finally, the oxidation pathways described here show different characteristics compared to the SO 2 oxidation promoted by other anionic systems, such as oxyhalogenated ions (XO n − X = Cl, Br, I; n = 1,2) [42] or transition metal oxide anions (MO 2 − , M = Co, Ni, Cu, Zn; CrO 4 − ) [37], whose common feature is that oxidation already occurs within the first highly exothermic intermediate, leading to free or bound SO 3 0/− . The same oxygen transfer process is not observed in the case of HCO 4 − , in which an intramolecular hydrogen bond engages the carbonyl oxygen in a stable 5-term cycle directing the reactivity towards the peroxidic OH that binds to SO 2 . Oxidation is only observed when the peroxidic oxygen is no longer bound to the hydrogen atom and can be eventually transferred to SO 2 .

Mass Spectrometry
Mass spectrometric investigations were performed by using an LTQ-XL linear ion trap mass spectrometer (ThermoFisher Scientific, Waltham, MA, USA) equipped with an electrospray ionisation (ESI) source. This instrument was previously adapted to realise ion-molecule reaction (IMR) experiments [33,[36][37][38][39][40][41][42][43][44]. Accordingly, the gaseous neutral reactant can be introduced into the ion trap through a deactivated fused silica capillary entering the vacuum chamber from a 6.25 mm hole placed in the backside of the mass spectrometer. A Granville-Phillips Series 370 Stabil Ion Vacuum Gauge was used to measure the pressure of the neutral gas, which was kept constant by a metering valve. Owing to the position of the Pirani gauge compared to the ion trap, the actual gas pressure was obtained after reading calibrations [58]. The peroxymonocarbonate ion (HCO 4 − ) was formed in solution by dissolving sodium percarbonate (Na 2 CO 3 ·1.5 H 2 O 2 ) in H 2 O/CH 3 CN (1:1, v/v) to a millimolar concentration, injected into the ESI source of the instrument at a flow rate of 5 µL min −1 . Alternatively, the HCO 4 − ion was obtained in the gas phase by directly subjecting pure water to the electrospray ionisation process. Nitrogen was used as a sheath and auxiliary gas at a flow rate of 11 and 2 arbitrary units (a. u.), respectively, a. u.~0.37 L min −1 . Other source parameters were set as follows: spray voltage 3.2 kV, capillary temperature 275 • C, tube lens 15 V, and capillary voltage 10 V. Once optimised, the HCO 4 − reactant ion was isolated into the ion trap and exposed to gaseous sulphur dioxide (SO 2 ). Typical SO 2 pressures ranged between 1.1 × 10 −7 Torr and 6.5 × 10 −7 Torr, with uncertainty of ±30%. During the kinetic experiments, the normalised collision energy was set to zero, whereas the activation Q value was tuned to optimise the stable trapping fields for each ion. The ionic signals of the HCO 4 − reactant and reaction products were monitored over time as a function of the SO 2 concentration, and an average of 10 scans was acquired. Xcalibur 2.0.6 software was used to record and process all the mass spectra. Other selected neutrals, such as C 2 H 4 , 13 CO 2 , C 18 O 2 , CO, and O 2 , proved unreactive with HCO 4 − . The reaction of HCO 4 − ion with gaseous SO 2 can be assimilated to a pseudo-first-order process due to the excess of neutral reactant compared to the precursor ion into the ion trap. Nonlinear least squares regression was performed by using the DynaFit4 software package [59], which simultaneously fit reactant and product concentrations vs. time. The reaction mechanism was verified by fitting the experimental results of the kinetic analysis with the mathematical model proposed for the postulated reaction, which simulated the time progress of the reaction based on the unimolecular rate constants obtained from the software calculations. The pseudo-first-order constants (s −1 ) were divided into the concentrations of neutral reagent gas to derive the bimolecular rate constants k (cm 3 molecule −1 s −1 ), whereas the branching ratios of the different reaction channels were obtained from the formation constants of the direct products. According to the average dipole orientation (ADO) theory [60], the efficiency of the process was calculated as the ratio of the bimolecular rate constant k to the collision rate constant (k coll ). Approximately 15 independent measurements were performed on different days over a sixfold neutral pressure range to guarantee the linearity of the k values. The standard deviation in the absolute rate constant was typically <10%, although a conservative error of 30% was given owing to the uncertainties affecting the measurement of the neutral pressure. The ionic reactant and products were also characterised by collision-induced dissociation (CID) experiments performed by furnishing energy to mass-selected ions in the presence of helium background gas (P He ca. 3 × 10 −3 Torr). Normalised collision energies were tuned in a 20-40% range as a function of the isolated species and applied with a standard activation time of 30 ms. Ions were mass-selected with a window of 1 m/z and a Q value optimised to ensure stable trapping fields for all the ionic species under investigation. The 13 C-labelled experiments were performed by preparing and mass-analysing millimolar solutions of Na 2 13 CO 3 or NaH 13 CO 3 salt. Analogous studies were also carried out on 1.5 mL of ultrasound-degassed water placed in a double bottle-neck flask that was filled with 13 CO 2 until a pressure of 700 torr.

Computational Details
The potential energy surface of the system [HCO 4 − ···SO 2 ] was investigated localising the lowest stationary points at the B3LYP [61,62] level of theory in conjunction with the correlation consistent valence polarised set aug-cc-pVTZ [63][64][65], augmented with a tight d function with exponent 2.457 for the sulphur atom [66], to correct for the core polarisation effects [67]. This basis set will be denoted aug-cc-pV(T+d)Z. At the same level of theory, we computed the harmonic vibrational frequencies in order to check the nature of the stationary points, i.e., minimum if all the frequencies are real, saddle point if there is one, and only one for the imaginary frequency. The energy of all the stationary points was computed at the higher level of calculation CCSD(T) [68][69][70] using the same basis set, aug-cc-pV(T+d)Z. Both the B3LYP and the CCSD(T) energies were corrected to 298.15 K by adding the zero point energy and the thermal corrections computed using the scaled harmonic vibrational frequencies evaluated at the B3LYP/aug-cc-pV(T+d)Z level. All calculations were performed using Gaussian 09 [71], while the analysis of the vibrational frequencies was performed using Molekel [72,73].

Conclusions
Sulphur dioxide is efficiently oxidised in the gas phase by the peroxymonocarbonate ion, HCO 4 − , through the formal transfer of HO 2 − , oxygen atom, and oxygen ion, to the hydrogen sulphate anion, sulphur trioxide, and sulphur trioxide anion, respectively. Secondary reactions due to the formation of the hydrogen carbonate anion lead to the formation of hydrogen sulphite. The three oxidative pathways have been characterised using a joint approach of mass spectrometry and theoretical calculations. Contrary to the common behaviour observed in solution, in the gas phase, the peroxymonocarbonate exhibits nucleophilic properties, which are unravelled by the Lewis acid character of sulphur dioxide.