Acidity and the multiphase chemistry of atmospheric aqueous particles and clouds

The acidity of aqueous atmospheric solutions is a key parameter driving both the partitioning of semi-volatile acidic and basic trace gases and their aqueous-phase chemistry. In addition, the acidity of atmospheric aqueous phases, e.g., deliquesced aerosol particles, cloud, and fog droplets, is also dictated by aqueous-phase chemistry. These feedbacks between acidity and chemistry have crucial implications for the tropospheric lifetime of air pollutants, atmospheric composition, deposition to terrestrial and oceanic ecosystems, visibility, climate, and human health. Atmospheric research has made substantial progress in understanding feedbacks between acidity and multiphase chemistry during recent decades. This paper reviews the current state of knowledge on these feedbacks with a focus on aerosol and cloud systems, which involve both inorganic and organic aqueous-phase chemistry. Here, we describe the impacts of acidity on the phase partitioning of acidic and basic gases and buffering phenomena. Next, we review feedbacks of different acidity regimes on key chemical reaction mechanisms and kinetics, as well as uncertainties and chemical subsystems with incomplete information. Finally, we discuss atmospheric implications and highlight the need for future investigations, particularly with respect to reducing emissions of key acid precursors in a changing world, and the need for advancements in field and laboratory measurements and model tools.

However, it should be noted that literature values especially for the quenching with dissolved organic matter show a huge variation with typical values between 10 7 and up to few 10 9 M -1 s -1 (see above-mentioned references). Moreover, the simple mechanism scheme does not consider any triplet reactions with inorganics such as halides due to their very compound and triplet specific reactivity (see e.g. (Treinin and Hayon, 1976;Loeff et al., 1992;Loeff et al., 1993;Tinel et al., 2014)).
Consequently, the calculated concentrations represent upper limit values for the steady-state concentrations of PS*.
Finally, daytime mean concentrations were calculated for urban winter haze, rural aerosol and rural/urban cloud conditions based on the modelled PS* concentration profiles (see Table 1 in the main manuscript text) and applied for the comparison of the different S(IV) to S(VI) conversion pathways (see Fig 7). The obtained daytime mean concentrations, ranging from about few 10 -11 mol L -1 under deliquesced aerosol conditions to 10 -13 − 10 -12 mol L -1 under cloud conditions, are in a reasonable agreement with reported triplet concentrations in aerosols (2.3·10 −13 − 1.6·10 −10 mol L -1 estimated by Wang et al. (2020)), in fog samples (0.07 − 1.5·10 −13 (Kaur and Anastasio, 2018)) and in natural surface waters (10 −14 − 10 −13 (Zepp et al., 1985;Canonica et al., 1995)).  In the following paragraphs, the data compiled in Tables S3 and S4 are briefly discussed. It should be noted that individual data which have been already been compiled in former reviews are generally not repeated here.

Formaldehyde
The evaluation of the hydration constant of formaldehyde has been done several times (Bell, 1966;Ogata and Kawasaki, 1970;Doussin and Monod, 2013). Nevertheless, some studies were not included in these reviews. The values obtained Khyd. = 2190 at T = 293 K by Zavitsas et al. (1970), Khyd. = 2220 at T = 295 K by (Sutton and Downes, 1972), Khyd. = 2420 at T = 298 K by Lewis and Wolfenden (1973) and Khyd. = 2270 at T = 298 K by McDonald and Martin (1979) are slightly higher than the last recommended value Khyd. = 2000 at T = 298 K by Doussin and Monod (2013). Except the investigation of formaldehyde by Rivlin et al. (2015) with Khyd. = 2100 at T = 293 K no further recent measured value of the hydration equilibrium constant was added. Since there is no significant change in the recent reported values, the last recommendation suggested by Doussin and Monod (2013) should be applied.

Acetaldehyde
In contrast to formaldehyde, acetaldehyde has a significant smaller hydration equilibrium constant by a factor of 10 3 , due to the electronic influence of the CH3 group. The value recommended Khyd. = 1.43 at T = 298 K by Bell (1966) was suggested by a review of Tur'yan (2000) appears to be lower with Khyd. = 1.2 at T = 298 K. Further studies (Kurz, 1967;Lewis and Wolfenden, 1973;Sorensen and Jencks, 1987) which were not included in  or Doussin and Monod (2013) indicated the same Khyd. value. The last recent review by Doussin and Monod (2013) suggested the use of the recommended value from Tur'yan (2000), which as well be the recommendation from the present work. Table S3 summarizes the given Khyd. values from Bell (1966) and the references therein, Greenzaid et al. (1967b); Lewis and Wolfenden (1973); Buschmann et al. (1982) as well as Doussin and Monod (2013) and the references therein. The recommended value Khyd. = 0.85 from Doussin and Monod (2013) refers to a slightly higher value based on studies from the 1980's. Overall, it is recommended to use the suggested value from Doussin and Monod (2013).

Butanal
The hydration of butanal (Khyd. = 0.43) was first revised by Bell (1966). Similar values have been determined by Greenzaid et al. (1967b); Lewis and Wolfenden (1973) compiled in Table S3. The recent evaluation by Doussin and Monod (2013) considering more recent investigations, which indicated higher values. The recommendation from Doussin and Monod (2013) as well as our recommendation can be given with Khyd. = 0.60.

Pivaldehyde
The investigation compiled by Greenzaid et al. (1967b); Lewis and Wolfenden (1973) as well as references included in Doussin and Monod (2013) supports the recommended value. A smaller value was found by Lienhard and Jencks (1966), but it has to be mentioned that the derived values tends to be too low in this study.

Methylglyoxal
A further hydration constant of the dicarbonyl compound was found in the literature, which was not considered by the evaluation in Doussin and Monod (2013). The Khyd. of methylglyoxal was reported with a value of 2700 (Wasa and Musha, 1970), 1279 (Montoya and Mellado, 1994) and 565 (Creighton et al., 1988). The new recommended value is suggested to be the average value of all three studies, with a derived Khyd. = 1512. Nevertheless, even with this value the majority (>99.9%) of the methylglyoxal is present in its gem-diol.

Acetone
In case of acetone, the reported values agree very well. Bell (1966) recommended a very low hydration equilibrium constant for acetone, which was also obtained by an investigation of Greenzaid et al. (1967b).

Diacetyl
The recommendation from Doussin and Monod (2013) for Diacetyl was based on more recent results from the references therein. A higher value Khyd. = 3.3 was earlier suggested by Bell (1966). Later on a smaller value was obtained by Lewis and Wolfenden (1973) (Khyd. = 0.244) and by Buschmann et al. (1982) (Khyd. = 2.1). Therefore, we also recommend the given value from Doussin and Monod (2013).
Further hydration constants for simple aldehydes and ketones have been evaluated in Doussin and Monod (2013). Since there are no further recent studies found in the literature, in many cases these available recommended values can be followed.
>99 -1700, results in an uncertainty of less than 1% for the appearance of the keto form of the protonated glyoxylic acid.
Since the revision compiled by Tur'yan (1998) no recent determination of hydration constants of glyoxylic acid was found, with exception for the deprotonated glyoxylate from Leitzke et al. (2001). Nevertheless, Doussin and Monod (2013)and the present work follow the recommendation given by Tur'yan (1998).

Pyruvate
The variability of the hydration equilibrium constant Khyd.2 of the pyruvate is reported in the literature in a range of Khyd.2 < 0.1 at T = 298 K. The present study recommends a Khyd. = 0.08 of pyruvate, as an average value from the data at T = 298 K in Table S4. This value is slightly higher than recommendation by Doussin and Monod (2013).

Mesoxalic acid
In case of the mesoxalic acid only the fully protonated dicarboxylic acid equilibrium constant was reported by Strehlow (1962) (Khyd. = 100) and by Le Henaff (1968) (Khyd. = 99). The derived recommended value KHyd. = 100 suggested from this present review is very close to the value of Khyd. = 99 suggested by Doussin and Monod (2013). Unfortunately, there are no further data reported concerning the pH dependency of mesoxalic acid.
In general, the pH dependent behaviour of the apparent hydration constant Kapp., similar to pyruvic acid, is expected because of the connected equilibria. The α-oxocarboxylic acids were found to exist in equilibrium with the hydrate (gem-diol). The  Table S4).

2-imidazol-carboxaldehyde
In addition to the compounds discussed up to here, a similar pH dependency of the hydration constant Khyd. was reported by Ackendorf et al. (2017) for 2-imidazol-carboxaldehyde or 2-IC (cf . Table S3). Under acidic conditions, the photochemical active side-chain aldehyde group undergoes the equilibrium reaction with water and forms the gem-diol while the imidazole ring will be fully protonated. In general, 2-IC behaves as a double-basic acid.
Regarding further studies of multifunctional carbonyl compounds, based on the findings discussed here, future studies should consider the complex equilibria which can result in complex dependencies on hydration of carbonyl compounds on acidity.

Computational chemistry
For simple aldehydes and ketones the comparison in Doussin and Monod (2013) of calculated towards measured hydration constants, leads to reasonable derived values for T = 298 K (Guthrie, 2000;Hilal et al., 2005;Gomez-Bombarelli et al., 2009;Raventos-Duran et al., 2010). Nevertheless, for molecules that are more complex the values obtained by calculation methods scatter in a broader range around the determined equilibrium constant, partly with a poor level of agreement. Apparently, there is quite some potential for improvement for numerical predictions of hydration constants, especially for more complex species.