Atmospheric Vinyl Alcohol to Acetaldehyde Tautomerization Revisited

Sep 18, 2015 - Recently, the atmospheric fate of vinyl alcohol (VA) has received considerable attention, as it has been proposed(1) and later supporte...
0 downloads 10 Views 2MB Size
Letter pubs.acs.org/JPCL

Atmospheric Vinyl Alcohol to Acetaldehyde Tautomerization Revisited Jozef Peeters,*,† Vinh Son Nguyen,† and Jean-François Müller‡ †

Department of Chemistry, University of Leuven, B-3001 Heverlee, Belgium Belgian Institute for Space Aeronomy, Avenue Circulaire 3, B-1180 Brussels, Belgium



S Supporting Information *

ABSTRACT: The atmospheric oxidation of vinyl alcohol (VA) produced by photoisomerization of acetaldehyde (AA) is thought to be a source of formic acid (FA). Nevertheless, a recent theoretical study predicted a high rate coefficient k1(298 K) of ≈10−14 cm3 molecule−1 s−1 for the FA-catalyzed tautomerization reaction 1 of VA back into AA, which suggests that FA buffers its own production from VA. However, the unusually high frequency factor implied by that study prompted us to reinvestigate reaction 1. On the basis of a high-level ab initio potential energy profile, we first established that transition state theory is applicable, and derived a k1(298 K) of only ≈2 × 10−20 cm3 molecule−1 s−1, concluding that the reaction is negligible. Instead, we propose and rationalize another important VA sink: its uptake by aqueous aerosol and cloud droplets followed by fast liquid-phase tautomerization to AA; global modeling puts the average lifetime by this sink at a few hours, similar to oxidation by OH. to be as high as k1 = 1.30 × 10−14 cm3 molecule−1 s−1 at 298 K, which would result in a short atmospheric VA lifetime of less than 1 h at an average HC(O)OH mixing ratio5 of a few parts per billion by volume (ppbv). Note that in that study, the reaction was considered as “chemically activated” and the kinetics were evaluated using RRKM- and master-equation approaches, applying a zero-curvature tunneling correction (ZCT) assuming an asymmetric Eckart barrier. In another, more recent theoretical study of the tautomerization of VA catalyzed by HC(O)OH as well as by inorganic oxo-acids, Karton12 computed enthalpy barrier heights ΔH‡ and transition-state free energies ΔG‡ at high levels of theory, but did not evaluate rate coefficients; it was concluded that the investigated acids can efficiently catalyze the tautomerization, specifically mentioning that inorganic second-row acids H2SO4, HClO4 and H3PO4 lead to submerged energy barriers, but remaining unclear about the expected efficiency of HC(O)OH. The high theoretical rate coefficient k1(298 K) reported by da Silva11 has been adopted, among others, in the very recent modeling study on missing HC(O)OH sources by Millet et al.5 The resulting large impact of (R1) on the modeled HC(O)OH led Millet et al. to conclude that the efficiency of acetaldehyde phototautomerization as a source of atmospheric HC(O)OH must be inversely related to the abundance of HC(O)OH itself and that therefore HC(O)OH in fact buffers its own production by this mechanism.5

R

ecently, the atmospheric fate of vinyl alcohol (VA) has received considerable attention, as it has been proposed1 and later supported by theoretical calculations2 that its oxidation by hydroxyl radicals (OH) produces formic acid, a major source of cloud and rainwater acidity in remote environments. Current state-of-the-art models underestimate the atmospheric abundances of formic acid by at least a factor of 2,3−5 suggesting, most likely, one or more large, missing HC(O)OH sources, presumably through the oxidation of other volatile organic compounds, such as vinyl alcohol. Besides being a known combustion byproduct,6 VA has recently been shown to be formed by the photoisomerization of acetaldehyde (AA),7,8 through keto−enol tautomerization of the excited AA with quantum yield estimated at 21 ± 4%.7 As acetaldehyde is a very common atmospheric carbonyl with a global budget exceeding 200 Tg per year,9 the source strength of VA through this mechanism should be considerable5 and the potential for formic acid production from VA oxidation therefore substantial (of the order of 15 Tg/year globally), if VA did not have any other significant sink. Whereas the uncatalyzed reverse tautomerization of gasphase VA back into AA is believed to be negligible,10 catalysis by gas-phase carboxylic acids was proposed to greatly enhance tautomerization by da Silva,11 who theoretically predicted the bimolecular rate coefficient of the formic acid-catalyzed tautomerization of VA,

Received: August 14, 2015 Accepted: September 18, 2015

CH 2CHOH + HC(O)OH → CH3CHO + HC(O)OH (1) © XXXX American Chemical Society

4005

DOI: 10.1021/acs.jpclett.5b01787 J. Phys. Chem. Lett. 2015, 6, 4005−4011

Letter

The Journal of Physical Chemistry Letters

Figure 1. Adiabatic potential energy surface of the tautomerization of vinyl alcohol (VA) to acetaldehyde (AA) catalyzed by formic acid (FA). Adiabatic energies E (at 0 K): VA + FA, TS, and AA + FA at 3-point extrapolated CBS-UCCSD(T)//UQCISD/6-311++G(d,p) level (see text); reactant complex RC and product complex PC at level UCCSD(T)/aVTZ//UQCISD/6-311++G(d,p); ZPVE from UM06-2x/6-311++G(3df,2p) harmonic vibrational frequencies scaled by a factor of 0.957.

work (see below), but the following description will mention the relative enthalpy data ΔH(298 K) reported by da Silva11 and by Karton,12 which our E results generally agree with taking into account that the relative TS energy E0‡ = ΔH‡(298 K) + 1.0 kcal mol−1 for this case.12 Da Silva11 and Karton12 concur that the first step is the barrier-less formation of a prereaction complex (RC) between the reactants HC(O)OH and CH2CHOH, stabilized for 10.411 or 7.912 kcal mol−1 (ΔH298 K) by hydrogen bonding. The reactant complex RC faces a barrier of 16.011 or 14.212 kcal mol−1 (enthalpy barrier, 298 K) to convert by a concerted double H-shift through a tight transition state TS (relative ΔH‡298 of +5.611 or 6.312 kcal mol−1) into an H-bonded product complex (PC) between HC(O)OH and CH3−CHO lying 20.311 or 17.712 kcal mol−1 (ΔH298 K) below the reactants. The H-bonded PC breaks up quickly into the separated end products, with relative ΔH298 K of −9.511 or −9.812 kcal mol−1. The molecular mechanism of the critical double H-shift, described in more detail by da Silva,11 is illustrated by the structures shown on Figure 1. However, of paramount importance for the overall kinetics, product formation through the rigid, high-lying TS is only a very minor fate of the reactant complex RC that will much more rapidly revert to the lowerlying original reactants, through a loose, variational bottleneck. Obviously, given its energetic and entropic advantage, redissociation of RC into the reactants is many orders of magnitude faster than its forward reaction through TS to the products, even accounting for the tunneling factor of order 10 at 298 K (see below). As a result, the formation and redissociation of the reactant complex will be quickly quasiequilibrated, in all conditions. Then, all structures up to and including the rate controlling transition state will be in thermal equilibrium with the reactants,13 such that the prerequisite for applying straightforward thermal transition state theory (TST)

However, there are serious grounds to question such a high k1 value.11 The reported11 enthalpy ΔH‡298 = +5.6 kcal mol−1 of the rate limiting transition state (TS) relative to the reactants, and the corresponding TS energy E0‡ = +6.6 kcal mol−1 (using E0‡ or ΔH‡(0 K) = ΔH‡(298 K) + 1.0 kcal mol−1, computed by Karton;12 see below), implies a low Boltzmann fraction exp(−E0‡/kBT) of only ≈10−5 at room temperature, such that the reported k1 value would require an excessively high bimolecular frequency factor A(T) of order 10−9 cm3 molecule−1 s−1. Allowing for a tunneling factor κ(T) of order 10 (see below), this is about 5 orders of magnitude higher than the expected A(T) of order 10−15 to 10−14 cm3 molecule−1 s−1 for a reaction between polyatomic molecules through a tight TS involving a concerted double H-shiftas characterized by both da Silva11 and Karton.12 That the TS is indeed highly rigid and the A(T) factor accordingly very low can be inferred from the strongly negative entropy of activation as derived from the enthalpy of activation, ΔH‡(298 K) = 6.26 kcal mol−1, reported by Karton and the Gibbs energy of activation, ΔG‡(298 K) = 18.95 kcal mol−1 (standard state 1 atm), listed in the Supporting Information of his paper.12 The strongly negative ΔS‡(298 K) of −42.6 cal mol−1 K−1 (standard state 1 atm) retrieved in this way implies an A(T) (without tunneling correction) even below 10−15 cm3 molecule−1 s−1, clearly incommensurate with the high k1 predicted by da Silva.11 The prime objective of the present work is therefore to theoretically re-evaluate the rate coefficient of reaction (R1). In order to assess whether the reaction should be treated either as a thermal or as a chemically activated process, the detailed overall reaction mechanism should first be examined. The potential energy surface (PES) based on the studies of da Silva and Karton as well as this work is schematized in Figure 1. The structures and the relative adiabatic potential energy data at 0 K, denoted by E, shown on the figure are taken from this 4006

DOI: 10.1021/acs.jpclett.5b01787 J. Phys. Chem. Lett. 2015, 6, 4005−4011

Letter

The Journal of Physical Chemistry Letters

Table 1. Relative Adiabatic Potential Energies Ea of All Structures (See Figure 1) in the Tautomerization of Vinyl Alcohol (VA) to Acetaldehyde (AA) catalyzed by Formic Acid (FA), in kcal mol−1 methods structures

UM06-2x/6-311++ G(3df,2p)

UCCSD(T)/ aVDZb

UCCSD(T)/ aVTZb

UCCSD(T)/ aVQZb

CBS-c UCCSD(T)

VA + FA RC TS PC AA + FA

0.0 −9.25 3.99 −17.50 −9.15

0.0 −8.67 5.65 −20.16 −11.47

0.0 −8.32 6.03 −18.95 −10.48

0.0

0.0

6.82

7.34 (7.15)

−10.05

−9.79 (−9.87)

a

Relative energies E include zero-point vibrational energies (see Table S1) obtained from UM06-2x/6-311++G(3df,2p) harmonic vibrational frequencies scaled by a factor of 0.957. bSingle-point UCCSD(T)/aug-cc-pVnZ energies (with n = D, T, and Q) based on UQCISD/6-311++G(d,p) optimized geometries. c“Complete Basis Set” UCCSD(T) energies, by extrapolation of the UCCSD(T) values for the aug-cc-pVnZ basis sets, with n = D, T and Q; in parentheses: energies obtained by 2-point extrapolation using only the n = T and Q data.

conventional TST expression for a direct thermal reaction A + B → D + E through a TS with relative energy E0‡. As a general remark, it is readily rationalized that (i) (quasi-)equilibration A + B ⇆ RC, as a prerequisite for the validity of TST, will be satisfied when the overall rate coefficient is much smaller than that for the formation of the prereaction complex, and (ii) the subsequent step RC → Products can only result in steady-state populations of RC at all vibration energies smaller than the thermal equilibrium populations, such that there is no question of superthermal chemical activation, and that TST theory will in principle always give an upper limit of the overall rate coefficient. The onlyand often minoreffect of the pre- and postreaction complexes is that the tunneling factor κ(T) is in fact that for the step RC → PC (see Figure 1 and eq 3), i.e., through a barrier with forward and reverse heights ETS − ERC and ETS − EPC, respectively. Essential for the evaluation of the rate coefficient are therefore the relative TS energy E0‡ and the partition function ratio QTS/QA × QB. Estimation of κ(T) following the asymmetric Eckart-barrier ZCT tunneling approximation, as da Silva,11 requires the imaginary frequency of the TS, and, second, also the energies of RC and PC that determine the heights of the barrier (see above). A consistent set of energy data and vibration−rotation parameters needed to implement eq 4presently not availablewas obtained in this work at levels of theory ensuring sufficient precision for our goal, i.e., without aiming for highest-accuracy energy data as obtained by Karton at the W1-F12 level,12 which we will use in any case to implement an alternative TST formulation, below. The geometries of all stationary points on the PES (see Figure 1) were first optimized at the UM06-2X/6-311G++(3df,2p) level of DFT theory,19 abbreviated below as M06-2X, the minima being characterized by all real vibration frequencies and the TS by one imaginary frequency. The M06-2X harmonic vibrational frequencies, scaled by 0.957,20 and corresponding zero-point vibration energies (ZPVE), as well as the M06-2X rotational constants, were used to derive the harmonic-oscillator vibration and rigid-rotor rotation partition functions for calculating the TST rate coefficient. The geometries of the lowest-energy conformers of all structures were then reoptimized using the Quadratic Configuration Interaction method, UQCISD/6311G(d,p),21 and refined single-point energies were computed using the high-level UCCSD(T)/aug-cc-pVTZ method.22−24 The coupled-cluster energies obtained in this way, and including the ZPVE at M06-2X level, are denoted further as CC//QC energies. The crucial relative energy of the TS, as well as that of the products, was then extrapolated to the complete basis set limit (CBS)25 using the UCCSD(T) energies for the aug-cc-pVDZ, -pVTZ and -pVQZ basis sets. The alternative 2-

is fully satisfied, though with a proviso for the tunneling factor, as shown and clarified by Alvarez-Idaboy et al.14,15 and now generally applied for reactions through a prereaction complex with TS energy well above the reactant level (see, for example, Hermans et al.16) or slightly below (Mansergas and Anglada et al.17). Recently, TST theory was shown to agree reasonably with experiment even for a well-submerged, tight TS (Smith et al.18). The above shows that reaction 1 does not have to be treated as “chemically activated” and there is no need to use the more involved RRKM- and master-equation approaches. In fact, the term “chemically activated reaction” is generally used only when a reactant does not require thermal energy to overcome a TS, and clearly does not apply here as this TS lies about 7 kcal mol−1 above the (thermal) reactants. Still, anticipating the sharp disagreement between the RRKM-ME k1 results of da Silva11 and the TST-evaluated k1 from this work, below, and given the general importance of this kinetics issue, it is useful to summarize here the general proof for the validity of TST theory for a system A + B ⇆ RC → PC → D + E in which redissociation of the complex RC into A + B is much faster than its forward reaction to the products. For such a case, the RC will quickly attain quasi-perfect thermodynamic equilibrium with the reactants and hence also a Boltzmann energy distribution in detailed balance with the thermal reactants, such that the RC concentration can be expressed as [RC]eq = [A][B]Keq = [A][B](Q RC/Q A × Q B) × exp[−(ERC − E A + B) /kBT ] A

B

(2) RC

in which Q , Q , and Q are the respective total partition functions, and ERC − EA+B is the energy difference (at 0 K) between complex RC and the separated reactants. The rate of the thermal product-forming reaction RC → D + E can be found from TST theory: Rate = [RC]eq × κ(T ) × (kBT /h) × (Q TS/Q RC) × exp[−(ETS − ERC)/kBT ]

(3)

with similar meanings as above for ETS − ERC and QTS. Substitution of eq 2 in 3 leads to the overall rate coefficient k1(T) ≡ Rate/([A][B]): k1(T ) = κ(T ) × (kBT /h)(Q TS/Q A × Q B) exp[− (ETS − E A + B)/kBT ]

(4)

in which E − E is the effective energy barrier for the overall reaction, denoted above as E0‡. Equation 4 indeed becomes the TS

A+B

4007

DOI: 10.1021/acs.jpclett.5b01787 J. Phys. Chem. Lett. 2015, 6, 4005−4011

Letter

The Journal of Physical Chemistry Letters

state 1 atm) retrieved from Karton’s data12 (see above). The TST prefactor ATST ≡ κ(T) × (kBT/h) × QTS/(QA × QB) turns out to be 4.51 × 10−15 cm3 molecule−1 s−1 at 298 K, of the expected low magnitude. Using the 3-point extrapolated CBS-CC//QC result E0‡ = 7.34 kcal mol−1, TST expression (4) yields the rate coefficient at 298 K: k1(298 K) = 1.86 × 10−20 cm3 molecule−1 s−1. Clearly, k1 is much smaller than the rate coefficient for the barrierless formation of RC, meaning that A + B ⇆ RC is quasi-fully equilibrated and TST theory does indeed apply. The k1(T) results for the 280−320 K range in Table 2 can be fitted within ±3% by the Arrhenius expression:

point extrapolation using only the latter two energies, gives 0.1−0.2 kcal mol−1 lower results. The adiabatic potential energies, denoted by E, and always including the ZPVE, are given in Table 1 and Figure 1. All calculations were carried out with the Gaussian 09 program suite26 and the Molpro 2012.1 program.27 The UQCISD geometries of this work, depicted in Figure 1, are generally intermediate to those reported by da Silva11 and Karton.12 The crucial relative energy of the TS or effective barrier, E0‡, is 7.34 or 7.15 kcal mol−1 at the 3-point or 2-point CBS-CC//QC levels, respectively (see Table 1), the former closest to the higher-level 7.3 kcal mol−1 result of Karton,12 and somewhat greater than the 6.6 kcal mol−1 of da Silva11 taking into account that here E0‡ or ΔH‡(0 K) = ΔH‡(298 K) + 1.0 kcal mol−1.12 The (less important) relative energies E of the reactant and product complexes and of the final products, all at the CC//QC level, are within 1.2 kcal mol−1 of the earlier ΔH(298 K) results; note that for these loose complexes and the products, the difference ΔH(0 K) − ΔH(298 K) is ≤0.3 kcal mol−1.12 The forward and reverse barriers from this work (CC//QC level) used for calculating the tunneling factor κ(T) in the asymmetric Eckart-barrier approximation are ≈14.4 and ≈25.0 kcal mol−1, respectively. The scaled M06-2X imaginary frequency of the TS is ν‡im = i × 1435.5 cm−1, not too different from ν‡im = i × 1354.3 cm−1 reported by da Silva.11 Such a fairly low value is expected for a concerted double H-shift with effective reduced mass μ ≈ 2 amu; it is close to the ν‡im values around i × 1400 cm−1 found at the M06-2X level for deuterium atom shifts facing similar barriers in peroxy radicals from deuterated isoprene.28 The resulting κ(298 K) of 17.7, evaluated using the Eckart tunneling code of Coote et al.,29 is somewhat higher than the κ(298 K) = 12.4 value found when adopting the Eckart-barrier parameters reported by da Silva.11 Table 2 lists the tunneling factor κ(T),

k1(T ) = 1.36 × 10−18 × exp( − 1270/T ) cm 3 molecule−1 s−1

The 2-point extrapolated E0‡ of 7.15 kcal mol−1 yields an only slightly different k1(298 K) of 2.56 × 10−20 cm3 molecule−1 s−1. When using the lower E0‡ = 6.6 kcal mol−1 and ν‡im = i × 1354.3 cm−1 from da Silva, but the vibration−rotation parameters (not available from da Silva) of this work, one obtains a 2 to 3 times higher k1(298 K) of 4.8 × 10−20 cm3 molecule−1 s−1. By adopting the free energy of activation ΔG‡(298 K) = 18.95 kcal mol−1 (standard state 1 atm) from Karton’s Supporting Information,12 in combination with the κ(298 K) = 17.7 of this work, k1 can also be evaluated using the thermodynamic TST formulation (4′), which can be similarly derived for this type of reaction as above: k1(T ) = κ(T ) × (kBT /h) × exp[−ΔG‡(T )/RT ] × T × 1.363 × 10−22 cm 3 molecule−1 K−1

280 285 298 315 320

QTS/(QVA QAF),a cm3 molecule‑1 4.27 4.22 4.09 3.95 3.91

× × × × ×

10−29 10−29 10−29 10−29 10−29

κ(T)b 32.24 26.84 17.74 11.61 10.44

k1(T), cm3 molecule−1 s−1 1.50 1.58 1.86 2.43 2.64

× × × × ×

(4′)

in which the reciprocal of the factor (T × 1.363 × 10−22 cm3 molecule−1 K−1) converts pressure in atm to concentration in molecules cm−3. The resulting k1(298 K) = 5.64 × 10−20 cm3 molecule−1 s−1 is somewhat higher than our k1 values above, likely due to Karton’s B3LYP-D3 calculated TS geometry12 being slightly less rigid than the M06-2X geometry of this work. In any case, all four k1 results above are more than 5 orders of magnitude lower than the k1(298 K) of 1.30 × 10−14 cm3 molecule−1 s−1 reported by da Silva.11 The reason for the latter high result is not clear. It must therefore be concluded on the basis of the present results, supported by data from Karton,12 that the vinyl alcohol → acetaldehyde tautomerization catalyzed by formic acid, even at the highest tropospheric HC(O)OH mixing ratios of ≈10 ppvb or [HC(O)OH] ≤ 2.5 × 1011 molecules cm−3, occurs at a rate of only 10−8 s−1, meaning that this process should be entirely negligible in the atmosphere, in all conditions. Given that the other carboxylic acids in the atmosphere, mainly CH3COOH, have similar concentrations as HCOOH and should react with VA by an identical mechanism as HCOOH, the methyl/alkyl group and the C−H hydrogen of HCOOH both being merely bystanders (see TS geometry, Figure S1), one expects very similar rate coefficients and therefore likewise negligible atmospheric contributions. Interestingly, the gas-phase VA → AA conversion reactions catalyzed by gas-phase inorganic oxoacids (HNO3, H2SO4, HClO4 and H3PO4) were also investigated by Karton,12 who found that the latter three acids with a second-row central atom lead to (nearly-)submerged barriers and should therefore be much more efficient than HC(O)OH. Based again on ΔG‡(298 K) data in Karton’s Supporting Information,12 and adopting

Table 2. Partition Function Ratio QTS/(QVA QAF), Tunneling Factor κ(T), and Rate Coefficient k1(T) for Tautomerization of Vinyl Alcohol (VA) to Acetaldehyde (AA) Catalyzed by Formic Acid (FA) at Selected Temperatures T in the Range 280 to 320 K T/K

(5)

10−20 10−20 10−20 10−20 10−20

a

Partition function ratio includes reaction path degeneracy of 2 (two equivalent transition state enantiomers). bTunneling factor κ(T) for double-H shift from RC to PC, following the asymmetric Eckartbarrier approximation.

the total partition function ratio QTS/(QA × QB), and the rate coefficient k1(T) calculated using eq 4 for selected temperatures in the range 280−320 K. The QTS/(QA × QB) ratio based on our M06-2X ro-vibration parameters, and including the reaction path degeneracy of 2 (as there are two equivalent TS enantiomers), is found to be 4.09 × 10−29 molecule−1 cm3 at 298 K; the low ratio for a bimolecular reaction reflects the rigidity of the tight TS with high vibration frequencies for the transition modes, and is in quantitative agreement with the strongly negative ΔS‡(298 K) of −42.6 cal mol−1 K−1 (standard 4008

DOI: 10.1021/acs.jpclett.5b01787 J. Phys. Chem. Lett. 2015, 6, 4005−4011

Letter

The Journal of Physical Chemistry Letters

Figure 2. Annually and vertically averaged sink of vinyl alcohol (in hour−1) due to (a) reaction with OH, and (b) uptake by aqueous aerosols and clouds and subsequent tautomerization to acetaldehyde.

modeling study,9 the photoisomerization of AA is estimated to generate about 23 Tg VA yr−1, of which, neglecting dry and wet deposition, about one-third (8 Tg yr−1) is predicted to be converted back to AA due to heterogeneous uptake. Adopting the ca. 60% HC(O)OH yield predicted for the reaction of VA with OH,2 the global production of formic acid due to the reaction of VA with OH is estimated at ca. 10 Tg HC(O)OH yr−1. Note, however, that the newly proposed VA conversion would generally turn out slightly faster than the VA + OH reaction when the lower kOH of 2.6 × 10−11 cm3 molecule−1 s−1 estimated by the AOPWIN EPA program36 would be used in the model; also, this would result in substantially more VA converting to AA and less formic acid from the VA + OH reaction than above. In conclusion, in this work, the formic-acid catalyzed tautomerization of VA to AA is shown to be a very slow process with rate coefficient at ambient temperatures predicted to be as low as a few times 10−20 cm3 molecule−1 s−1orders of magnitude slower than an earlier theoretical estimate11and therefore entirely negligible in the atmosphere, while from the recent work of Karton12 it can be construed that the catalysis by inorganic second-row oxoacids in the gas phase is likewise unimportant because of the very low atmospheric mixing ratios of such compounds. On the other hand, it is argued that the heterogeneous uptake of VA by aqueous aerosols and cloud droplets, followed by fast liquid phase conversion to AA, should be an important VA sink of similar strength as its gas-phase reaction with OH that forms partly HC(O)OH, resulting in an overall atmospheric VA lifetime as short as 1 h on average, and leaving little room for any other chemical conversions of VA.

tunneling factors of order 10 as for reaction 1, the k(298 K) can be estimated to range from ≈10−15 (H3PO4) to ≈10−14 (H2SO4) cm3 molecule−1 s−1, much higher than k1 above. However, of the three oxoacids affording these high rate coefficients, only H2SO4 is known to be present in the atmosphere, and then only in very low measured gas-phase concentrations of order 105 to 107 molecules cm−3,30,31 offsetting the higher k and resulting in a first-order rate of only 10−9 to 10−7 s−1 for that case. Therefore, these reactions too are expected to have a negligible effect on VA in the atmosphere. On the other hand, we propose vinyl alcohol → acetaldehyde conversion on wet aerosols and in cloud droplets as an efficient process in the atmosphere. In aqueous solution, VA is wellknown to revert to its much more stable AA form nearinstantly, at a minimum rate of ≈1.5 × 10−2 s−1 near pH = 4 to 5, and much faster at both lower and higher pH, explained by acid- or base catalysis.32 Therefore, the limiting factors for conversion of atmospheric VA to AA in an aqueous phase, e.g., a cloud droplet or an aqueous aerosol, are gas-phase diffusion toward the aerosol/droplet surface and uptake upon collision. In absence of specific data for VA uptake on aqueous aerosols, the accommodation coefficient might be approximated by the measured temperature-dependent coefficient for ethanol uptake on water,33 about 0.01−0.02 at lower tropospheric temperatures. Using submicron aqueous (inorganic+organic) aerosol surface area densities calculated by a global chemistry transport model and validated against in situ measurements,34 the firstorder VA sink due to aerosol uptake, calculated following Schwartz,35 is estimated to range between 5 × 10−5 and 5 × 10−4 s−1 in the continental boundary layer, with an annual and global average of ca. 1.2 × 10−4 s−1, corresponding to a lifetime of about 2 h. As seen on Figure 2, the modeled heterogeneous VA sink over continents is of the same order as (although generally somewhat below) the gas-phase photochemical sink due to the reaction with OH calculated using the same model when adopting the reported theoretical rate constant estimate2 kOH ≈ 6.8 × 10−11 cm3 molecule−1 s−1. The modeled VA lifetime due to the two sinks considered is somewhat less than 1 h. With global acetaldehyde emissions (92 Tg yr−1) and photochemical production (102 Tg yr−1) similar to a previous



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpclett.5b01787. Geometries, absolute energies, zero-point vibration energies, harmonic vibrational frequencies, and rotational constants of all structures (PDF) 4009

DOI: 10.1021/acs.jpclett.5b01787 J. Phys. Chem. Lett. 2015, 6, 4005−4011

Letter

The Journal of Physical Chemistry Letters



Reactions: Hydrogen Abstraction from Aldehydes by OH. J. Am. Chem. Soc. 2001, 123, 2018−2024. (16) Hermans, I.; Jacobs, P.; Peeters, J. Pronounced Non-Arrhenius Behaviour of Hydrogen-Abstractions from Toluene and Derivatives by Phthalimide-N-oxyl Radicals: a Theoretical Study. Phys. Chem. Chem. Phys. 2008, 10, 1125−1132. (17) Mansergas, A.; Anglada, J. M. The Gas-Phase Reaction between O3 and HO Radical: a Theoretical Study. ChemPhysChem 2007, 8, 1534−1539. (18) Smith, M. C.; Chang, C.-H.; Chao, W.; Lin, L.-C.; Takahashi, K.; Boering, K. A.; Lin, J. J.-M. Strong Negative Temperature Dependence of the Simplest Criegee Intermediate CH2OO Reaction with Water Dimer. J. Phys. Chem. Lett. 2015, 6, 2708−2713. (19) Zhao, Y.; Truhlar, D. G. The M06 Suite of Density Functionals for Main Group Thermochemistry, Thermochemical Kinetics, Noncovalent Interactions, Excited States, and Transition Elements: Two New Functionals and Systematic Testing of Four M06-Class Functionals and 12 Other Functionals. Theor. Chem. Acc. 2008, 120, 215−241. (20) Laury, M. L.; Boesch, S. E.; Haken, I.; Sinha, P.; Wheeler, R. A.; Wilson, A. K. Harmonic Vibrational Frequencies: Scale Factors for Pure, Hybrid, Hybrid Meta, and Double-Hybrid Functionals in Conjunction with Correlation Consistent Basis Sets. J. Comput. Chem. 2011, 32, 2339−2347. (21) Pople, J. A.; Head-Gordon, M.; Raghavachari, K. Quadratic Configuration Interaction. A General Technique for Determining Electron Correlation Energies. J. Chem. Phys. 1987, 87, 5968. (22) Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Head-Gordon, M. A Fifth Order Perturbation Comparison of Electron Correlation Theories. Chem. Phys. Lett. 1989, 157, 479−483. (23) Kendall, R. A.; Dunning, T. H., Jr.; Harrison, R. J. Electron Affinities of the First-Row Atoms Revisited. Systematic Basis Sets and Wave Functions. J. Chem. Phys. 1992, 96, 6796−806. (24) Woon, D. E.; Dunning, T. H., Jr. Gaussian-Basis Sets for Use in Correlated Molecular Calculations. 3. The Atoms Aluminum Through Argon. J. Chem. Phys. 1993, 98, 1358−1371. (25) Peterson, K. A.; Wilson, A. K.; Woon, D. E.; Dunning, T. H., Jr. Benchmark Calculations with Correlated Molecular Wave Functions XII. Core Correlation Effects on the Homonuclear Diatomic Molecules B2 - F2. Theor. Chem. Acc. 1997, 97, 251−259. (26) 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.; et al. Gaussian 09, revision D.01; Gaussian, Inc.: Wallingford, CT, 2009. (27) Werner, H.-J.; Knowles, P. L.; Knizia, G.; Manby, F. R.; Schütz, M. Molpro: a General-Purpose Quantum Chemistry Program Package. WIREs Comput. Mol. Sci. 2012, 2, 242−253. (28) Peeters, J.; Müller, J.-F.; Stavrakou, T.; Nguyen, V. S. Hydroxyl Radical Recycling in Isoprene Oxidation Driven by Hydrogen Bonding and Hydrogen Tunneling: the Upgraded LIM1Mechanism. J. Phys. Chem. A 2014, 118, 8625−8643. (29) Coote, M. L.; Collins, M. A.; Radom, L. Calculation of accurate imaginary frequencies and tunneling coefficients for hydrogen abstraction reactions using IRCmax. Mol. Phys. 2003, 101, 1329−1338. (30) Berresheim, H.; Elste, T.; Tremmel, H. G.; Allen, A. G.; Hansson, H.-C.; Rosman, K.; DalMaso, M.; Mäkelä, J. M.; Kulmala, M.; O’Dowd, C. D. Gas-Aerosol Relationships of H2SO4, MSA, and OH: Observations in the Coastal Marine Boundary Layer at Mace Head, Ireland. J. Geophys. Res. 2002, 107 (D19), 8100. (31) Bardouki, H.; Berresheim, H.; Vrekoussis, M.; Sciare, J.; Kouvarakis, G.; Oikonomou, K.; Schneider, J.; Mihalopoulos, N. Gaseous (DMS, MSA, SO2, H2SO4 and DMSO) and particulate (sulfate and methanesulfonate) sulfur species over the northeastern coast of Crete. Atmos. Chem. Phys. 2003, 3, 1871−1886. (32) Capon, B.; Zucco, C. Simple enols. 2. Kinetics and mechanism of the ketonization of vinyl alcohol. J. Am. Chem. Soc. 1982, 104, 7567−7572.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research was sponsored by the Belgian Science Policy Office (BELSPO) under the contract SD/CS/05A (Project BIOSOA) in the frame of the Science for Sustainable Development program.



REFERENCES

(1) Archibald, A. T.; McGillen, M. R.; Taatjes, C. A.; Percival, C. J.; Shallcross, D. E. Atmospheric Transformation of Enols: A Potential Secondary Source of Carboxylic Acids in the Urban Troposphere. Geophys. Res. Lett. 2007, 34, L21801. (2) So, S.; Wille, U.; da Silva, G. The Atmospheric Chemistry of Enols: A Theoretical Study of the Vinyl Alcohol + OH + O2 Reaction Mechanism. Environ. Sci. Technol. 2014, 48, 6694−6701. (3) Paulot, F.; Wunch, D.; Crounse, J. D.; Toon, G. C.; Millet, D. B.; DeCarlo, P. F.; Vigouroux, C.; Deutscher, N. M.; González Abad, G.; Notholt, J.; et al. Importance of Secondary Sources in the Atmospheric Budgets of Formic and Acetic Acids. Atmos. Chem. Phys. 2011, 11, 1989−2013. (4) Stavrakou, T.; Müller, J.-F.; Peeters, J.; Razavi, A.; Clarisse, L.; Clerbaux, C.; Coheur, P.-F.; Hurtmans, D.; De Mazière, M.; Vigouroux, C.; et al. Satellite Evidence for a Large Source of Formic Acid from Boreal and Tropical Forests. Nat. Geosci. 2012, 5, 26−30. (5) Millet, D. B.; Baasandorj, M.; Farmer, D. K.; Thornton, J. A.; Baumann, K.; Brophy, P.; Chaliyakunnel, S.; de Gouw, J. A.; Graus, M.; Hu, L.; et al. A Large and Ubiquitous Source of Atmospheric Formic Acid. Atmos. Chem. Phys. 2015, 15, 6283−6304. (6) Taatjes, C. A.; Hansen, N.; McIlroy, A.; Miller, J. A.; Senosiain, J. P.; Klippenstein, S. J.; Qi, F.; Sheng, L.; Zhang, Y.; Cool, T. A.; et al. Enols are Common Intermediates in Hydrocarbon Oxidation. Science 2005, 308, 1887−1889. (7) Andrews, D. U.; Heazlewood, B. R.; Maccarone, A. T.; Conroy, T.; Payne, R. J.; Jordan, M. J. T.; Kable, S. H. Photo-tautomerization of Acetaldehyde to Vinyl Alcohol: A Potential Route to Tropospheric Acids. Science 2012, 337, 1203−1206. (8) Clubb, A. E.; Jordan, M. J. T.; Kable, S. H.; Osborn, D. L. Phototautomerization of Acetaldehyde to Vinyl Alcohol: A Primary Process in UV-Irradiated Acetaldehyde from 295 to 335 nm. J. Phys. Chem. Lett. 2012, 3, 3522−3526. (9) Millet, D.; Guenther, A.; Siegel, D. A.; Nelson, N. B.; Singh, H. B.; de Gouw, J. A.; Warneke, C.; Williams, J.; Eerdekens, G.; Sinha, V.; et al. Global Atmospheric Budget of Acetaldehyde: 3D Model Analysis and Constraints from In-Situ and Satellite Observations. Atmos. Chem. Phys. 2010, 10, 3405−3425. (10) Yamada, T.; Bozzelli, J. W.; Lay, T. H. Kinetic and Thermodynamic Analysis on OH Addition to Ethylene: Adduct Stabilization, Isomerization, and Isomer Dissociations. J. Phys. Chem. A 1999, 103, 7646−7655. (11) da Silva, G. Carboxylic Acid Catalyzed Keto-Enol Tautomerizations in the Gas Phase. Angew. Chem., Int. Ed. 2010, 49, 7523−7525. (12) Karton, A. Inorganic Acid-Catalyzed Tautomerization of Vinyl Alcohol to Acetaldehyde. Chem. Phys. Lett. 2014, 592, 330−333. (13) Singleton, D. L.; Cvetanovic, R. J. Temperature-Dependence of Reactions of Oxygen-Atoms with Olefins. J. Am. Chem. Soc. 1976, 98, 6812−6819. (14) Alvarez-Idaboy, J. R.; Mora-Diez, N.; Vivier-Bunge, A. A Quantum Chemical and Classical Transition State Theory Explanation of Negative Activation Energies in OH Addition to Substituted Ethenes. J. Am. Chem. Soc. 2000, 122, 3715−3720. (15) Alvarez-Idaboy, J. R.; Mora-Diez, N.; Boyd, R. J.; Vivier-Bunge, A. On the Importance of Prereactive Complexes in Molecule-Radical 4010

DOI: 10.1021/acs.jpclett.5b01787 J. Phys. Chem. Lett. 2015, 6, 4005−4011

Letter

The Journal of Physical Chemistry Letters (33) Jayne, J. T.; Duan, S. X.; Davidovits, P.; Worsnop, D. R.; Zahniser, M. S.; Kolb, C. A. Uptake of gas-phase alcohol and organic acid molecules by water surfaces. J. Phys. Chem. 1991, 95, 6329−6336. (34) Stavrakou, T.; Müller, J.-F.; Boersma, K. F.; van der A, R. J.; Kurokawa, J.; Ohara, T.; Zhang, Q. Key chemical NOx sink uncertainties and how they influence top-down emissions of nitrogen oxides. Atmos. Chem. Phys. 2013, 13, 9057−9082. (35) Schwarz, S. C. In Chemistry of Multiphase Atmospheric Systems; Jaeschke, W., Ed.; Springer Verlag: Heidelberg, Germany, 1986; pp 415−471. (36) AOPWIN, US EPA Estimation Programs Interface Suite for Microsoft.

4011

DOI: 10.1021/acs.jpclett.5b01787 J. Phys. Chem. Lett. 2015, 6, 4005−4011