Role of Methanesulfonic Acid in Sulfuric Acid–Amine and Ammonia New Particle Formation

Aerosol nucleation accounts for over half of all seed particles for cloud droplet formation. In the atmosphere, sulfuric acid (SA) nucleates with ammonia, amines, oxidized organics, and many more compounds to form particles. Studies have also shown that methanesulfonic acid (MSA) nucleates independently with amines and ammonia. MSA and SA are produced simultaneously via dimethyl sulfide oxidation in the marine atmosphere. However, limited knowledge exists on how MSA and SA nucleate together in the presence of various atmospherically relevant base compounds, which is critical to predicting marine nucleation rates accurately. This work provides experimental evidence that SA and MSA react to form particles with amines and that the SA-MSA-base nucleation has different reaction pathways than SA-base nucleation. Specifically, the formation of the SA-MSA heterodimer creates more energetically favorable pathways for SA-MSA-methylamine nucleation and an enhancement of nucleation rates. However, SA-trimethylamine nucleation is suppressed by MSA, likely due to the steric hindrance of the MSA and trimethylamine. These results display the importance of including nucleation reactions between SA, MSA, and various amines to predict particle nucleation rates in the marine atmosphere.


■ INTRODUCTION
Particle nucleation in the atmosphere may impact cloud formation and the Earth's radiation balance. 1,2 Atmospheric nucleation occurs when gas-phase compounds react to form a stable cluster. Sulfuric acid (SA, H 2 SO 4 ), an oxidation product of sulfur dioxide (SO 2 ) and dimethyl sulfide (DMS, C 2 H 6 S), has been shown to nucleate in the atmosphere, and its concentration in the atmosphere typically correlates with particle nucleation rates. 3−9 While sulfuric acid is an important molecule for atmospheric nucleation, sulfuric acid concentration alone cannot explain observed particle nucleation rates. 3,10 Various compounds can react with sulfuric acid to form particles, including ammonia 11−13 and amines. 14−16 These basic compounds are found in the atmosphere and are emitted through biomass burning, animal husbandry, and chemical and industrial plants. 17 For sulfuric acid−amine systems, extensive work has been conducted to determine the acid−base reaction steps for forming these particles. 3,16,18−20 In addition, sulfuric acid−base nucleation has been observed in ambient air, and its rates are quantified in the atmosphere worldwide. 5,21−24 As sulfuric acid nucleation is an integral part of weather and climate; recent studies have incorporated sulfuric acid−ammonia and amine nucleation into global climate models to improve predictions of atmospheric aerosol number concentrations. 25 −27 In addition to sulfuric acid, recent studies have found that methanesulfonic acid (MSA, CH 4 O 3 S) also contributes to particle nucleation and growth in the atmosphere. 28,29 MSA is primarily found in coastal and oceanic regions 29 as it is an oxidation product of DMS, a marine emission. 30,31 Amines and ammonia are also emitted in a marine environment, mainly from phytoplankton. 32−34 Previous field measurements indicate that MSA exists at around 10−100% of SA concentration, 28,35 with laboratory measurements showing MSA can nucleate with amines like dimethylamine (DMA, (CH 3 ) 2 NH). 36 Despite MSA's importance in atmospheric nucleation and its prevalence in the marine environment, no global climate models currently account for MSA nucleation.
While SA, MSA, amines, and ammonia coexist in a marine atmosphere, limited information exists on how these compounds nucleate together to form particles. Chen et al. 37 have demonstrated that MSA-amine reactions can form particles at parts per billion level concentrations of reactants, higher than measured in the atmosphere. 37 In addition, Elm's computational chemistry results previously showed that the inclusion of MSA to the SA-base system creates a strong interaction between MSA, SA, and ammonia/amines that could potentially enhance the nucleation rates compared to the SA-base systems. 38,39 However, this study was limited to clusters containing up to two acid molecules, any number of base molecules, and no water. While these cluster binding energies suggest that MSA could influence sulfuric acid nucleation rates, experimental observations are required to uncover the dominant stepwise reactions between MSA, SA, and amines/ammonia.
The experimental study presented here examines the nucleation reactions in the SA-MSA-amine/ammonia systems. Mass spectrometer measurements of freshly nucleated clusters show that MSA is involved in the first steps of nucleation. Particle concentration measurements also show that MSA could enhance or suppress sulfuric acid−base particle formation rates depending on the base compound. Results demonstrate that the role of MSA in MSA-SA-base in new particle formation, i.e., nucleation and growth, is dependent on the ratio of SA and MSA concentrations and the interaction of MSA with the various basic compounds. Including MSA when modeling atmospheric new particle formation, especially in marine environments, is needed to accurately predict particle concentrations in the atmosphere.

■ METHODS
Nucleation experiments were conducted using a clean and repeatable glass flow reactor as described in Fomete et al., 40 with pertinent details repeated here. The flow rate and temperature of the reactor are held constant at 4 LPM and 298−300 K (based on small fluctuations in room temperature), respectively, while relative humidity (RH) is 20%. There are three main injection flows into the flow reactor: nitrogen entrained with sulfuric acid, dry nitrogen, and humidified nitrogen. Sulfuric acid vapor is generated by passing nitrogen over a liquid sulfuric acid reservoir and injected at the top of the reactor. Sulfuric acid concentration is controlled by specifying the flow rate through the reservoir with concentrations ranging from 10 7 to 10 9 cm −3 . Humidified nitrogen and dry nitrogen streams are also injected into the top of the reactor to control the RH in the flow reactor and provide a dilution stream for sulfuric acid. The reactor has been continuously purged for ∼2 years with gaseous sulfuric acid, nitrogen, and water to remove potential contaminant compounds and ensure repeatable reaction conditions. 16,40,41 Baseline measurements of the sulfuric acid dimer (i.e., a cluster containing two sulfuric acid molecules and any water molecules that evaporate upon measurement) and particle concentrations as a function of sulfuric acid concentration are taken daily to ensure consistent measurements across all experiments. 40 MSA was injected into the flow reactor at 80− 120 sccm, and its concentration was varied from 10 7 to 10 10 cm −3 by adjusting the flow rate of N 2 over the liquid MSA reservoir. Gaseous ammonia (NH 3 , Wards Science 40 wt % in H 2 O), methylamine (MA, CH 3 NH 2 , Sigma-Aldrich 40 wt % in H 2 O), dimethylamine (DMA, (CH3) 2 NH, Acros Organics 40 wt % in H 2 O), and trimethylamine (TMA, C 3 H 9 N, Sigma-Aldrich 45 wt % in H 2 O) were generated and injected into the flow reactor using custom-built permeation tubes and a double dilution system. 40,42 No unexpected or unusually high safety hazards were encountered.
For the nucleation experiments, MSA was injected into the flow reactor, allowing SA and MSA to mix for ∼8 s. Amines/ ammonia were injected into the flow reactor and allowed to nucleate with MSA and SA for ∼2 s prior to measurement. This nucleation time is based on the centerline velocity from the flow parameters and distance to the measurement point.
Concentrations of MSA, SA, and the base compounds, as well as freshly formed molecular clusters, are measured using an atmospheric pressure chemical ionization quadrupole mass spectrometer known as the Minnesota Cluster CIMS (MCC). 16,43−45 Acetate and nitrate were used as reagent ions to ionize acidic molecular clusters. Nitrate was used for dimer cluster observations and most particle observations. Acetate ionization was used to measure SA and MSA concentrations during the particle observations for MA and TMA when varying [MSA]. The reaction rate constant was 2 × 10 −9 cm 3 s −1 for nitrate and 4.6 × 10 −9 cm 3 s −1 for acetate. 46−48 Additionally, it is assumed that MSA is ionized by nitrate and acetate at the same rate as SA, as no previous measurements have been conducted on its ionization rate constant. Hydronium ions (and its larger clusters) were used to ionize the basic gases. Ion signals are converted to concentrations using the method described in Fomete et al., 48 with a chemical ionization time of 0.02 s. Mass-dependent transmission efficiency values for the MCC were used to account for differences in the detection due to ion mass. 45 The systematic uncertainty of the MCC has been estimated to be a factor of two. 14 However, this uncertainty would affect all measurements equally and thus would have little impact on the overall trends in the data.
Particle concentrations were measured with a 1 nm versatile water condensation particle counter (vwCPC, TSI 3789). 49 The conditioner on the vwCPC was set to 1°C, and the initiator was set to 99°C. The nucleation time of 2 s was chosen to ensure that freshly formed particles did not grow beyond ∼1 nm in diameter during the short nucleation time. Longer nucleation times would result in higher coagulation losses, which would obscure the reaction formation pathways. In comparison, shorter nucleation times would mean particles are smaller than 1 nm and would not get measured by the vwCPC. Figure S1 shows the difference in particle counts of the vwCPC at a 1 nm setting vs the 2 nm setting (conditioner set to 2°C and the initiator set to 90°C) while injecting 7 pptv of TMA and [MSA] = 4 × 10 9 cm −3 into the sulfuric acid flow reactor. Particle concentrations decreased by over 97% when changing the vwCPC temperatures from the 1 nm to the 2 nm setting. The decrease in particle concentration indicates that almost all of the particles formed are 1−2 nm. ). Dimer clusters likely had water and base molecules attached that evaporated upon measurement. 50 Throughout the discussion of monomers and dimers, the convention used is a monomer or dimer refers to the number of acid molecules in the cluster rather than the total number of molecules. An increase in acid dimer concentrations is a useful indicator of particle formation, and likely, the acid dimers contained a base molecule that evaporated when the cluster was ionized. Figure  indicates that there are little to no molecular interactions occurring between acid molecules when no base is present in the reactor. The SA·SA dimer without a base is likely forming via ion-induced clustering (IIC) as the binding free energy of the uncharged cluster is weak at −5.5 kcal/mol. 45 IIC occurs within the inlet of the MCC and is where ions that have been electrically charged by the reagent ion (e.g., nitrate) continue to react and form clusters with other neutral molecules within the flow reactor. 40 Similarly, MSA·SA and MSA·MSA may also form via IIC in the SA and MSA injection conditions. IIC likely influences all of the dimers as SA·SA, MSA·SA, and MSA·MSA have similar computed binding free energies of −5.5, −5.1, and −5.4 kcal/mol, respectively. Note that all referenced binding free energies are provided by Elm 38,39 and summarized in Table S1.
[MSA·MSA] remains low and unchanged for all of the base molecules, implying that either the MSA·MSA·base or MSA·base clusters are unstable at 298− 300 K. In addition, Elm 39 has shown that the computed MSA· base free energies range from −3.4 to −8.7 kcal/mol. 39 In contrast, SA·base free energies are generally stronger, ranging from −5.6 to −12.6 kcal/mol. 39 38 The binding free energies for SA·SA·NH 3 ·NH 3 and MSA·SA·NH 3 · NH 3 are −27.0 and −23.6 kcal/mol, respectively. 38 The more strongly bonded SA·SA·NH 3 ·NH 3 cluster indicates that ammonia shows some preference to form SA dimer and  38,39 In contrast, the free energies of the monomers are not similar as SA·MA is more strongly bonded at −7.2 kcal/mol when compared to MSA·MA at −3.9 kcal/mol. 38 also increases. This trend is expected since a cluster is more likely to collide with SA in this high SA regime, and MA has previously been shown to help form relatively stable sulfuric acid clusters. 14,16,53 For DMA = 16 pptv in Figure 1, the [SA·SA] and [MSA·SA] increase by over 50% compared to the no DMA system. The higher [SA·SA] is not surprising, given DMA'ss wellestablished high stability effects on sulfuric acid clusters. 16  This result agrees with binding free energies where MSA·SA· DMA is −28.2 kcal/mol and MSA·SA·TMA is −24.9 kcal/mol. SA·TMA is also more strongly bonded (−12.6 kcal/mol) than MSA·TMA (−8.7 kcal/mol). 38 Additionally, there is a significant decrease in the binding free energy of MSA·SA· TMA·TMA (−31.9 kcal/mol) when compared to SA·SA· TMA·TMA (−41.5 kcal/mol). This decrease in the free energy for MSA·SA·TMA·TMA indicates that at larger cluster sizes, TMA is much more likely to react with SA, which matches the high fraction of [SA·SA] for TMA in Figure 2. Therefore, the main pathway for dimer formation for the TMA-MSA-SA system is TMA reacting with SA.
Comparison of the observed dimer concentrations of the TMA-SA-MSA and MA-SA-MSA systems leads to disagreement with computed free energies. Specifically, MSA·SA·TMA has similar binding free energies as MSA·SA·MA (−24.9 and −24.2 kcal/mol, respectively). Furthermore, TMA has stronger binding free energies with MSA than MA (−8.7 and −3.  Scheme 1 summarizes the reaction pathways for the SA-MSA-base system. The likely formation pathway for these amines/ammonia starts with the reaction of SA and base. For DMA and MA, the monomer adds either a SA molecule or an MSA molecule at similar rates. From Scheme 1B, the first step is still the addition of a TMA molecule to SA; however, for TMA, the more energetically favorable next step is the addition of SA instead of MSA. This reaction pathway is consistent with the cluster observations from Figure 1 that show a significant increase in the [SA·SA] when TMA is present in the flow reactor. Scheme 1C shows how ammonia has relatively weak interactions with both acid molecules and likely only slightly prefers reacting with SA.
The combined observations demonstrate that MA and DMA can form stable clusters with MSA. Ammonia is likely too weak of a stabilizing base to cluster with MSA. TMA showed an appreciable increase in [MSA·SA] compared to no TMA, but [SA·SA] still dominates the total dimer concentration. Thus, TMA likely reacts primarily with SA. This result is possibly due to TMA's greater steric hindrance that does not allow TMA to react as readily with MSA beyond the monomer.
Clusters larger than the dimer were observed in the MSA-SA-base systems. Small amounts of SA timer and tetramer were observed for the amines/ammonia. However, there were likely other trimers that included amines and MSA that were not being measured due to cluster fragmentation inside the MCC or during the atmospheric ionization with nitrate. Future work should explore cluster ionization to better understand how to measure larger molecular clusters.
Particle Observations. Since clusters are larger than the dimer fragment in the MCC, particle measurements were taken with the vwCPC for the SA-MSA-base systems to better understand nucleated particles' formation rates. Figure 2 Figure S2 shows that this decrease in particle counts was due to MSA suppressing SA-TMA nucleation and not coagulation. The observed decrease in particle concentration is likely due to the steric hindrance in the SA-MSA-TMA system, where MSA binds up available TMA (as MSA·TMA), preventing further reaction with MSA or SA. This result also agrees with the dimer cluster    Figure 3B shows the particle concentration results for the SA-MSA-MA system. MSA and MA concentrations were varied with [SA] = 5 × 10 7 cm −3 . Each curve represents a different ratio of [SA]/[MSA] from 0.55 to 0.07 for MA. Figure 3B shows the opposite phenomena to Figure 3A, where increasing [MSA] by an order of magnitude leads to an increase in the particle concentrations up to a factor of three. The increase in particle concentrations signifies that MA can nucleate with MSA and SA. MA likely nucleates with both acids because MA is a significantly smaller molecule than TMA and less likely to be sterically hindered when colliding and reacting with MSA. MSA reacting with MA to form particles is also in agreement with computational chemistry results that showed equal stabilities of the SA·SA·MA cluster and the MSA· SA·MA cluster.

ACS Earth and Space Chemistry
Particle concentrations for the SA-MSA-MA system are still lower than SA-MSA-TMA until [SA]/[MSA] < 0.55. This difference between base systems is due to the first step in monomer cluster formation, which is likely the new particle formation rate-limiting step. Though MA can add to MSA and SA, these clusters (SA·MA and MSA·MA) are still more weakly bonded than those with TMA (specifically SA·TMA). Thus, while MA enhances particle formation rates for the SA-MSA system, TMA remains a potent nucleating compound with SA and still plays an important role in atmospheric marine nucleation.

■ CONCLUSIONS
Methanesulfonic acid (MSA) impacts the initial steps of cluster formation in the SA-MSA-base system. Specifically, adding MSA to the SA-base system allows more pathways for dimer cluster formation, as indicated by the formation of the MSA dimer and the SA-MSA heterodimer. The fraction of [SA·SA], [MSA·MSA], and [SA·MSA] formed out of the total dimer concentration also indicates which base compounds are reacting more readily with MSA than others. Methylamine and dimethylamine react with SA and MSA to form relatively equal ratios of [MSA·SA] and [SA·SA], indicating that MSA is readily reacting with MA and DMA. However, [MSA·SA] in the case of TMA is significantly lower than the [SA·SA], indicating that TMA is preferentially nucleating with SA. There are only minor increases to dimer concentrations for ammonia, indicating that any interactions between ammonia and MSA are likely negligible compared to SA and ammonia.
Particle measurements showed that while MSA enhances sulfuric acid nucleation in the case of MA, it may suppress nucleation in the case of SA-TMA. Introducing MA into the SA-MSA system saw an increase in particle concentrations at higher concentrations of MSA, whereas TMA saw a decrease in particle concentrations at high MSA concentrations. However, in the atmosphere, MSA concentration typically never surpasses that of SA, so while the laboratory results show suppression in the case of TMA, the ratios of acids may not be atmospherically relevant.
Overall, results indicate that MSA is an important contributor to atmospheric nucleation and can increase particle formation rates for SA-base nucleation. Including MSA in models will be especially important as anthropogenic emissions of SO 2 decrease, thus decreasing the [SA]/[MSA] ratio. These conclusions indicate that nucleation models that account for MSA-SA-base nucleation are required to better predict particle number concentrations in atmosphere, especially in the marine environment.