Water vapor is the third abundant species in Earth’s atmosphere and plays important role in atmospheric chemistry. One famous example, involving water vapor, is the sequential photochemistry of ozone to produce OH radical. In addition, a few studies also showed the oxidizing ability of some radical, such as O3, HO2, and OH, increases when water is present.65,66 The impact of a slow reaction with water vapor is still huge because of the vast amount of water vapor in the atmosphere.
Before the new preparation method of Criegee intermediates was reported 16, studies of Criegee intermediate reaction with water were mostly based on monitoring the product ratio in an ozonolysis system at various humidity levels;67-75 thus, the ratio of relative reaction rate coefficient, e.g. reaction coefficient with water to that with SO2 or unimolecular decomposition, was determined.
In ethylene ozonolysis system, Atkinson et al.67 reviewed several ozonolysis studies and deduced a ratio of k(CH2OO+H2O)/k(CH2OO+SO2) = 5.7x105. Suto et al.73, who monitored the light scattering by particle, and Becker et al.72, who measured the
formation of H2O2, determined the relative reaction coefficients ratio with water and SO2
to be 2.3x104 and 8.3x104, respectively. Recently, Berndt et al.75, repeated the same measurement but detected H2SO4 with mass spectroscopy; they found that the yield of H2SO4 decreased when water was present and the extra loss rate, kloss, showed a quadratic dependence on [H2O].
, ≅ H O
A few group studied the reaction of CH2OO with water vapor either by direct or indirect measurement with the new photolysis method to prepare CH OO. Welz et al.16 directly
measured the decay of CH2OO by VUV-PIMS in the photolysis system of CH2I2/O2/H2O.
They failed to observe the reaction with water vapor and gave an upper limit of
k
H2O<4x1015 cm3 s1. Stone et al. 76and Ouyang et al.77 prepared CH2OO with the photolysis method and monitored the formation of product.Stone et al.76 used laser-induced fluorescence technique to monitor CH2O, which is thought to be the main product of most CH2OO reaction. They did experiments at 200 Torr and had [H2O] up to1.7x1017 cm3 (5 Torr at 298 K). Experimental results showed that formation rate of CH2O remained the same when water was present; they gave an upper limit of kH2O< 9x1017 cm3 s1.
Ouyang et al.77 investigated this reaction with an approach similar to ozonolysis study.
CH2OO was prepared by the photolysis method; they monitored the formation of NO3
when both NO2 and H2O were present. They varied [H2O] up to 22 Torr and deduced a rate coefficient of CH2OO with H2O, kH2O = 2.5x1017 cm3 s1, by taking kCH2OO+NO2 = 7x1012 cm3 s1.
C2 Criegee intermediates were also investigated in the diiodoalkan photolysis system
18,47 or the ozonolysis system.74 Berndt et al.74 studied the effect of water vapor on H2SO4
formation in the 2-butene ozonolyssis system, which produces CH3CHOO. Water vapor decreased the yield of H2SO4. Unfortunately, they could not give a solid conclusion because of the complexity of CH3CHOO, including syn- and anti-conformers. A five parameters analysis gave an effective rate ratio for reaction of CH3CHOO with water to SO2 =1.4x104, but the authors believed this data was over interpreted.
On the other hand, Taatjes et al 18 directly measured the decay of CH3CHOO by
VUV-PIMS, which can separate the contribution of two conformers due to the difference in ionization energy. This study showed that syn-CH3CHOO doesn’t react with water
vapor with kH2O< 2x1016 cm3 s1; in contrast, anti-CH3CHOO reacts rapidly with water vapor with kH2O = 1x1014 cm3 s1.
An interesting discrepancy was shown; most studies on the ozonolysis system72-75 agreed that CH2OO reacts with water while researches using the photolysis method 16,76,77 failed to measure this reaction. Our group directly measured the time profile of CH2OO with UV absorption spectroscopyunder a wide range of pressure (100-760 Torr) and temperature (283-358 K). 44-46
Fig. 13 Representative time profile of CH2OO band under different humidity level at 298 K. Single exponential decay function is fitted to these traces and plotted in gray line. For this experiment set, Ptotal = 250 Torr, PO2 = 10 Torr, PCH2I2 = 6.7 mTorr and photolysis laser energy = 9.7 mJ cm2. Replotting from Ref.
44.
Fig. 13 shows typical decay traces of CH2OO at different humidity level within 335-345 nm. 44 The decay rate of CH2OO is faster at high humidity level. Within our detection window, not only the absorption of CH2OO but also the absorption of CH2I2 was observed.
-500 0 500 1000 1500 2000 2500 3000
-0.002 0.000 0.002 0.004 0.006 0.008 0.010 0.012
0.014 RH < 3%
RH = 21%
RH = 42%
RH = 63%
RH = 83%
Abs . (340 nm)
Delay Time / s
The negative baseline at long time (>3ms) was attributed to the depletion of CH2I2. Because the photolysis of CH2I2 is in picosecond time scale and the refreshed time of gas is about 1 second. We treated the kinetic behavior CH2I2 depletion as a step function; thus, a constant term was introduced with the single exponential decay function, A0e-kobs + B0. The highest [CH2OO] in our experiment was about 3x1012 cm3, which was far lower than [H2O] (>1x1016 cm3). The observed decay rate without water vapor is caused by reaction with byproducts (mostly react with iodine atom). The observed decay rate without water vapor, denoted as k0, was subtracted from each set of experiment in the
pseudofirstorder plot.
Fig. 14 Pseudofirstorder plot of CH2OO reaction with water vapor at 298K. Left panel shows results against water monomer concentration and right panel shows results against water dimer concentration.
There is no pressure dependence in the pressure range of 100-500 Torr. A linear function is fitted to the whole data for right panel. Replotting the graph from Ref. 44.
Fig. 14 shows the dependence of the pseudofirstorder decay rate on [H2O] under different pressure at 298K. The observed decay rate has a quadratic dependency on [H2O], indicating that two water molecules participate in this reaction. We also plotted the observed decay rate against water dimer concentration in the right panel; the effective
0 1 2 3 4 5 6 7 8
reaction rate coefficient with water dimer was determined to be k(H2O)2 = 6.5x1012 cm3 s1 at 298K.
The quadratic dependence explained why Welz et al.16 could not observe this reaction.
Although mass spectroscopy has high sensitivity, it is hard to build a high pressure reactor and couple it to high vacuum (<104 Torr). The small total pressure of their experiments (~4 Torr) limited [H2O] in the reactor. The highest [H2O] in the VUV-PIMS
measurement16 was [H2O] = 3x1016 cm3, resulting in a low water dimer concentration, [(H2O)2] = 2x1012 cm3, and a small effective decay rate, keff =13 s1, which is far below their detection limit.
To understand the reaction mechanism and implication of atmosphere (the average temperature on Earth is about 15°C), we studied the decay rate of CH2OO at different temperature.45,46 The pseudofirstorder plot in the temperature tange 283-353K is shown in Fig. 15.
Fig. 15 Pseudofirstorder plot on [H2O] at different temperature. Left panel shows the quadratic behavior becomes less pronounced at temperature. Right panel shows a second order polynomial fit for the data at 358 K. Adapted from Ref 45 and 46.
Two important features were revealed from the experiment. First, a negative temperature dependence was observed; this behavior was explained by the formation of water
complex, which was first postulated by Ryznkov et al. 78,79 They suggested that CH2OO and water dimer will first form a water complex and then becomes hydroxymethyl hydroperoxide (HMHP, HOCH2OOH) with one water molecule. The details of this reaction mechanism will be discussed below (Fig. 17).
Second, parts of a linear behavior exist above 349 K, indicative of the reaction with water monomer. The contribution from water dimer becomes comparable with the contribution from water monomer at high temperature. A second order polynomial is fitted to extract the rate information; the fitted result at 358 K data set is shown in Fig. 15 (right panel).
The quadratic part will become comparable to the linear part when [H2O]=1x1018 cm3 at 358 K. Reaction rate coefficient with water monomer of kH2O(358 K) = 7.3x1016 cm3 s1 is determined.
CH2OO reaction with water monomer was observed in some extreme conditions.46,80 Not only at high temperature, at which reaction with water dimer becomes slow, but also the reaction of CH2OO with water monomer can be observe under low humidity
environment. Recently, Berndt et al. 80 successfully measured the rate coefficient of CH2OO with H2O under ambient condition; they had highest [H2O] = 1x1015 cm3, with which [(H2O)2] is million times smaller than [H2O]. They determined the reaction rate coefficient with water monomer as kH2O(297 K) = 3.2x1016 cm3 s1; this value agrees with the direct measurement result from VUV-PIMS. 16
Fig. 16 Arrhenius plots of CH2OO reactions with water monomer and water dimer. Reaction with water dimer has a negative activation energy of 8.1 kcal mol1;45 in the contrast, reaction with water monomer has a positive Ea of ~3 kcal mol1.46 We added the data from Berndt et al80 for extending the temperature range of the monomer reaction. Replotting the figure from Ref. 45 and 46.
Arrhenius plot of rate coefficient of CH2OO reaction with water vapor is shown in Fig.
16. The reaction of CH2OO with (H2O)2 shows a strong negative temperature dependence with an activation energy Ea = 8.1±0.6) kcal mol1 by fitting all the data to the
Arrhenius form k(T) = Aexp(Ea/RT).45 With the calibration from multiple measurements at distinctive temperature, we suggested the reaction rate coefficient with water dimer to be k(H2O)2(298 K)= 7.4±0.6) cm3 s1 at 298 K. On the other hand, CH2OO reaction with H2O shows slightly positive temperature dependence by comparing our high temperature data with room temperature data from Berndt et al.; 80 the reaction rate coefficient of H2O at high and low temperature differ by a factor of ~2. The activation energy of reaction with H2O is estimated to be ~3 kcal mol1. 46
2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6
-15.6 -14.8 -12.4 -11.6 -10.8
-16.0 -15.2 -12.0 -11.2 -10.4
log10 [kmonomer / (cm3 s
1)]
Our works Berndt et al.
1000 / (T/K) log10 [kdimer / (cm3 s
1)]
Fig. 17 Proposed reaction mechanism of CH2OO reaction with water dimer.
The negative temperature dependence of reaction with water dimer could be explained by the formation of water complex 46,78 (see Fig. 17). Where kf and kr means the forward and reverse rate coefficient of complex formation and kp is the rate coefficient of double hydrogen transfer to the product.
CH2OO has strong zwitterionic character 29,38,81, resulting in a large dipole moment (4.1 D for CH2OO) 81; there are partial positive charge on the center carbon and partial negative charge on the terminal oxygen. This pre-reactive complex, which has a seven members ring, will go through double a hydrogen transfer process and end by release one water molecule.
Assuming a fast equilibrium between (H2O)2 and CH2OO, the effective rate coefficient with (H2O)2 becomes product of equilibrium constant of (H2O)2-CH2OO complex and kp. This assumption is reasonable because (i) the equilibrium constant of water complex is usually small. 65 The calculated equilibrium constant of (H2O)2-CH2OO is 7x1023 cm3 at 298 K (G°298 = 2.9 kcal mol1). 78 (ii) Strong dipole-dipole interaction between CH2OO and (H2O)2 makes a fast forward reaction.
The effective rate coefficient of CH2OO with water dimer is k(H2O)2(T) = Keq(T) kp(T) and the temperature dependence of k(H2O)2 is mainly controlled by Keq; higher complex concentration at low temperature speeds up this reaction.
Anot
plane and the H-C-H plane is zero for the ground state CH2OO. Energy difference between reactant and transition state of water monomer reaction is about 3 kcal mol1, consisted with the estimated activation energy of kH2O. 46
On the other hand, the structure of (H2O)2-CH2OO complex is very close to the transition state structure. The energy increase due to deformation of transition state structure is not comparable with the stabilization energy of complex formation; thus, this reaction pathway shows a negative temperature dependence, which agrees with the experimental activation energy (-8.1 kcal mol1 for experiment, 45-6 kcal mol1 for theory) 46.Water complex plays an important role in this case. According to activated complex theory (ACT) and thermodynamics, the activation energy of a reaction relates to the enthalpy change (H) but not the energy change (E). The discrepancy between experiment and theory may be explained by the thermal energy difference.
This is the second example of reaction involving water dimer. Hydration of SO3 to form sulfuric acid also exhibits a quadratic dependence on [H2O] and a large negative
activation energy of Ea = 9 kcal mol1,83 which is attributed to the formation of stable pre-reactive complexes with (H2O)2. Theoretical calculation shows that the barrier decreases from 28 kcal mol1 to 11 kcal mol1 for reaction with water monomer and water dimer, respectively. 84 This kind of water-catalyzed reaction involving hydrogen atom transfer was reviewed by Kumar et al. 85, which received great attention in atmospheric chemistry due to the abundance of water concentration.
Reactions of small Criegee intermediate with water trimer and tetramer was also investigated theoretically. 79 The reaction rate increases with the number of water molecule but the effect of water cluster is negligible in the atmosphere because the concentration of large water cluster drops dramatically while the increase of rate
coefficient is limited by collision frequency. Moreover, Zhu et al. 86 calculated the effect of water surface reaction with CH2OO; they found that a new reaction channel opens at the water surface with a picoseconds lifetime of CH2OO. This theoretical prediction should be tested.
Theory predicted HMHP as the product of CH2OO water reaction. 78,79 This agrees with the product analysis of ethylene ozonolysis.68,69 Early oznolysis study showed that the yield of HMHP reaches a limit,
HMHP = 0.4, when [H2O]>1.5x1017 cm3 at 298K, while
HMHP is almost zeros at dry condition. 69 Recently, Nakajima et al. 87 observed the formation of HMHP in a jet beam. By changing the co-reactant from H2O to D2O, they confirmed that the extra H and OH group in HMHP were attributed to the co-reactant.Due to the large exothermicity, HMHP may decompose into hydroxymethoxy
(H2C(O)OH) and OH radical.88 Then, H2C(O)OH radical will react with oxygen to form HO2 and HCOOH (kH2C(O)OH+O2 = 3.5x1014 cm3 s1 at 298 K).58 The formation of HCOOH was observed both in ethyleneozonolysis system 68and photolysis of CH2I2/O2
by microwave spectroscopy 87 when water was present.
Interestingly, a study of ethylene ozonolysis 70 showed that the yield of H2CO is independent of humidity. These observations may explain why Stone et al. 76 didn’t observe the change of H2CO formation rate. In the ozonolysis system, H2CO is mainly produced from the decomposition of POZ; the product of CH2OO reaction with water vapor, HMHP, will finally become HCOOH but not H2CO. In CH2I2 photolysis system, it is possible taht most CH2O is due to reactions involving ICH2OO; forexample, 2
ICH2OO → 2 CH2IO + O2, CH2IO → CH2O + I.
Ouyang et al. 77 did not measure the absolute rate coefficient but the branching ratio of products. In their study, they believed that NO is formed from reaction of CH OO with
NO2; the formation of NO3 decreases because of competition between NO2 and H2O for CH2OO. However, Caravan et al.89 failed to observe NO3 but found adducts signal, CH2OONO2, by VUV-PIMS. NO3 in the photolysis system is more feasible to form from iodine chemistry with NO2, INO2 + IONO2 → NO3 + NO2 + I2,and not Criegee
chemistry.89
The reactivity of syn-conformer Criegee intermediates with water vapor was investigated in the TME ozonolysis system.Berndt et al. 74 found that water vapor has no effect on H2SO4 formation. This structure-dependent reactivity was also predicted by theoretical work in 2004.78 After a hard work on synthesizing 2,2-diiodopropane, our group directly measured the reaction of (CH3)2COO with water vapor.48
Fig. 19 Observed decay rate of (CH3)2COO under different water concentration. The observed decay rate is different because the [(CH3)2COO]0 is changed. Linear fitting gave us either positive slope or negative slope. This reaction is slower than our detection limit. Adapted from Ref. 48.
Fig. 19 shows the observed decay rate of (CH3)2COO under different experimental conditions.48 The rate coefficient for reaction with H2O was too slow to measure. This
0 1 2 3 4 5 6 7
0 250 500 750 1000 1250 1500
Exp #W1, 200 Torr Exp #W5, 200 Torr Exp #W2, 200 Torr Exp #W6, 200 Torr Exp #W3, 200 Torr Exp #W7, 400 Torr Exp #W4, 200 Torr Exp #W8, 500 Torr
k obs / s1
[H2O] / 1017cm-3
experimental result agrees with theoretical calculations 78 and the previous ozonolysis study 74. Base on the scattering of data, we determined an upper bound of (CH3)2COO bimolecular rate coefficient with H2O or (H2O)2 to be 1.5x1016 cm3 s1 for H2O and 1.3x1013 cm3 s1 for (H2O)2. 48
Lin et al.also calculated the reaction of CH3CHOO 46 and (CH3)2COO 90 with water vapor and measured the reaction rate coefficient of anti-CH3CHOOwith water vapor by UV absorption spectroscopy 47. The structures of Criegee intermediates strongly affect reactivity with (H2O)2. A simple explanation of this observation is that methyl group will block water molecule from attacking the central carbon. For reactions with H2O, this structure dependence is not as significant as (H2O)2 because the reaction bottleneck is deformation of Criegee intermediate at the transition state.
In summary, people relied on indirect measurement to understand the reaction of small Criegee intermediates with water vapor 67-75 before the photolysis method was
reported.16,18 No solid conclusion was made until recently. First, the accuracy of theory is not high enough. Although theory 78 successfully predicted the structure dependence, the predicted rate coefficients coverd a broad range. 9 The reaction rate coefficient of small Criegee intermediates with water vapor ranged from 1016–1013 cm3 s1 for
anti-conformers and 10
21–1017 cm3 s1for syn-conformers.9 Second, the potential complexity of ozonolysis system makes this situation messy; over-simplified reaction schemes were usually assumed in data analysis, leading to the wrong interpretation. This situation changed when directly measurements of small Criegee intermediates were employed in a wide pressure and humidity range, 44-49 such that these experimental data can be directly compared with ozonolysis system. 67-75 Methodology of ozonolysisstudies and calculation accuracy can be calibrated by the results from direct measurement.
Both experimental 18,45,47and theoretical 9,46,78works have proved that anti-conformer Criegee intermediates in the typical tropospheric condition react with water vapor and form hydroxyalkyl hydroperoxide, which releases OH radical and reacts with O2 to from organic acid and HO2.88 This reaction may increase the oxidizing capacity of the
atmosphere and organic acids lead to the formation of secondary organic aerosol.3 For
syn-conformers Criegee intermediates, they do not react with water vapor and have
potential to oxidize other trace gases.Due to the difficulty of precursor synthesis, direct measurement of larger Criegee intermediate is limited. Since Criegee intermediates with complex structures are formed in the atmosphere,60,61 study of the ozonolysis system and theoretical calculations are still the key to understand the role of Criegee intermediates in atmospheric chemistry.