11638
J. Phys. Chem. A 2010, 114, 11638–11644
Size-Dependent Reactions of Ammonium Bisulfate Clusters with Dimethylamine Bryan R. Bzdek, Douglas P. Ridge, and Murray V. Johnston* Department of Chemistry and Biochemistry, UniVersity of Delaware, Newark, Delaware 19716 ReceiVed: July 9, 2010; ReVised Manuscript ReceiVed: September 8, 2010
The reaction kinetics of ammonium bisulfate clusters with dimethylamine (DMA) gas were investigated using Fourier transform ion cyclotron resonance mass spectrometry (FTICR-MS). Clusters ranged in size from 1 to 10 bisulfate ions. Although displacement of the first several ammonium ions by DMA occurred with near unit efficiency, displacement of the final ammonium ion was cluster size dependent. For small clusters, all ammonium ions are exposed to incoming DMA molecules, allowing for facile exchange (“surface” exchange). However, with increasing cluster size, an ammonium ion can be trapped in an inaccessible region of the cluster (“core” exchange), thereby rendering exchange difficult. DMA was also observed to add onto existing dimethylaminium bisulfate clusters above a critical size, whereas ammonia did not add onto ammonium bisulfate clusters. The results suggest that as the cluster size increases, di-dimethylaminium sulfate formation becomes more favorable. The results of this study give further evidence to suggest that ambient sub-3 nm diameter particles are likely to contain aminium salts rather than ammonium salts. Introduction The formation and growth of new atmospheric particles may play an important role in global climate. New particles in the low nanometer size range arise from the condensation of atmospheric gas-phase species. These particles can subsequently grow to ∼100 nm in diameter, where they may serve as cloud condensation nuclei (CCN).1-6 New particle formation (NPF) is a global phenomenon and frequently occurs over the continental boundary layer in regional events that can extend hundreds of kilometers.7,8 As a result, new particles that grow into the CCN size range may indirectly affect global climate by increasing cloud albedo and influencing precipitation patterns (aerosol indirect effect). These effects are poorly understood, especially compared to the understanding of the climatic effects of gas-phase species, which renders modeling aerosol climate effects difficult.9 Improved climate modeling necessitates a better understanding of the mechanisms of NPF and particle growth. Mounting evidence suggests that organic compounds, such as amines, drive the growth of newly formed particles. Observations of ambient particles during NPF events using a condensation particle counter battery indicate that 3 nm diameter particles contain a significant organic component.10 Volatility analysis of nucleated particles also suggests that these particles do not consist of just sulfuric acid, water, and ammonia, because the newly formed particles volatilize at temperatures higher than that for pure ammonium sulfate or sulfuric acid.11 A recent field study in Antarctica revealed that freshly nucleated particles had lower hygroscopicities than those of particles studied during nonnucleation events.12 Recent laboratory work indicates that organic aminium salts have the requisite physical properties: lower volatilities and hygroscopicities than those of ammonium salts.13 Amines have also been observed directly in particles during NPF events. Dimethylamine (DMA) has been observed in nucleated particles in Hyytia¨la¨, Finland.14 Other field studies have revealed the presence of amines in ultrafine particles during NPF events in urban, rural, and remote environments.13,15 * Corresponding author. Phone: (302) 831-8014. Fax: (302) 831-6335. Email:
[email protected].
Laboratory studies have implicated amines and other organics as playing a crucial role in the initial stages of NPF. For example, recent kinetics experiments have indicated that heterogeneous reactions of alkylamines with sulfuric acid contribute to the growth of acidic particles.16 Researchers utilizing a tandem differential mobility analyzer observed that organic species, including trimethylamine, enhance nanoparticle growth.17 Recent work in our laboratory has examined the exchange of amines for ammonia in ammonium bisulfate and ammonium nitrate nuclei. Exchange was collision-limited, indicating that ambient salts in this size range would likely be aminium salts rather than ammonium salts, even if they were initially formed as ammonium salts.18 These experimental findings are in agreement with recent computational studies that predict a significant enhancement of amine levels in particles.19,20 Much theoretical and experimental effort has focused on the composition of the critical cluster, since it determines the nucleation rate.21-24 Ambient observations suggest that the nucleation rate is correlated with the sulfuric acid vapor concentration.7,8 Two main nucleation models have been advanced to explain this correlation. In the kinetic model,25-28 the critical cluster is assumed to form through bimolecular collisions of sulfuric acid-containing clusters, and the nucleation rate (J) is proportional to the square of the sulfuric acid concentration ([H2SO4]2). In the activation model,29,30 nucleation occurs through the activation of small clusters containing one sulfuric acid molecule, and J is directly proportional to the sulfuric acid concentration ([H2SO4]1). However, despite much emphasis on the composition of the critical cluster, comparatively little work has examined the mechanisms of growth and chemical tranformation from the critical cluster up to detectable size, despite evidence suggesting that the observation of NPF is governed by the competition between aerosol scavenging and growth, rather than just nucleation.22,30,31 Methods to probe and interpret aerosol kinetics are welldeveloped for large particles (>100 nm diameter). These methods are often used to study uptake mechanisms onto particles and to determine the rate-limiting step of a reaction.32-41 However, aerosol kinetics are not as well-understood for nanoparticles, in part due
10.1021/jp106363m 2010 American Chemical Society Published on Web 10/11/2010
Reactions of Ammonium Bisulfate Clusters, DMA
J. Phys. Chem. A, Vol. 114, No. 43, 2010 11639
SCHEME 1
to the limited instrumentation that can perform composition measurements in this size range.42 A proper understanding of the mechanisms of particle growth from the critical cluster size to the minimum observable size for most instruments (3 nm diameter) and beyond requires the systematic study of how the reaction kinetics for these particles change as a function of size. This work examines the kinetics of ammonium bisulfate cluster transformation when exposed to DMA gas as a function of cluster size with the intent to identify size-dependent reaction pathways. Experimental Section The experimental setup is similar to that in our previous work on amine exchange in smaller ammonium bisulfate clusters.18 Charged ammonium bisulfate clusters were introduced to a 7T Bruker Apex Qe Fourier transform ion cyclotron resonance mass spectrometer (FTICR-MS) operating in the positive mode by electrospray of a 0.5 mM solution of ammonium sulfate (Aldrich) in 70/30 acetonitrile/water. Electrospray produced an array of clusters that were predominantly singly charged. Ions for a specific cluster of interest were mass-selected and accumulated in a quadrupole. Ions were then transferred to the ICR cell where they were exposed to a constant pressure of DMA gas (Matheson Tri-Gas) at 2.4 ( 0.5 × 10-8 Torr that was introduced to the cell via a leak valve. Although multiply charged clusters were observed in the mass spectrum at the same exact mass as singly charged clusters, multiply charged clusters were easily and quantitatively differentiated from the singly charged clusters by the presence of isotopic peaks at fractional mass-to-charge (m/z) ratios, which cannot arise from a singly charged cluster. Analysis of the abundance of these fractional m/z peaks determined the intensity of the multiply charged clusters to be generally 97% singly charged clusters). Therefore, the effect of multiply charged clusters on kinetic analysis was minimal. In addition, singly charged dimethylaminium bisulfate clusters were introduced to the FTICR-MS by electrospray of a dimethylaminium sulfate solution (0.5 mM in 50/50 methanol/ water) made by combining equal proportions of solutions of 2.0 mM DMA (Fluka) in 50/50 methanol/water and 1.0 mM H2SO4 (Fisher) in 50/50 methanol/water. Isolated dimethylaminium bisulfate clusters were exposed to DMA gas to study the kinetics of DMA addition to the clusters. Finally, ammonium and dimethylaminium bisulfate clusters were exposed to ammonia gas (Matheson Tri-Gas) at 1.0 ( 0.2 × 10-7 Torr to determine the kinetics of ammonia addition to the clusters. A mass spectrum of ions in the ICR cell can be obtained at a specific trapping time. FTICR-MS provides high resolution and accuracy m/z measurements, which allows for the assignment of
unique elemental formulas to reactant and product ions. A plot of ion abundance as a function of trapping time (reaction profile) reveals the progress of the sequential reactions. Since the substitution reaction is exothermic,18 each incoming DMA molecule induces decomposition of the cluster as well as the displacement of a single ligand. These minor reaction-induced decomposition channels are not enhanced by adding unreactive argon collision gas. As will be discussed later, the reaction-induced decomposition was too small to affect quantitative kinetic analysis of the predominant simple displacements. All data were fit to the kinetic models using the simplex method of nonlinear fitting embodied in the Solver function of the Microsoft Excel program. Determination of second order rate constants required knowledge of the absolute pressure of gas in the ICR cell. However, due to effects associated with the magnet and the polarizability of the gas being measured, the ICR cell pressure reading given by the ionization gauge did not correspond to the true ICR cell pressure. The absolute gas pressure in the ICR cell was determined by the equation
Ptrue ) Pgauge × Kmagnet ×
( ) RN2
Rgas
(1)
where Ptrue is the true ICR cell pressure, Pgauge is the pressure reading on the ionization gauge, Kmagnet is an empirical correction factor for the effect of the magnet on the pressure reading (ratio of the pressure reading when the gauge is out of the magnetic field to when it is in the magnetic field), and R is the polarizability of the gas being measured. Because the ionization gauge is calibrated for N2, the ratio of the polarizability of N2 to the polarizability of the gas being measured is required to determine the absolute pressure of the gas being measured.43 Results and Discussion Amine Substitution into Ammonium Bisulfate Clusters. The sequential displacement of ammonia by DMA in ammonium bisulfate clusters is described by Scheme 1. Figure 1 presents representative mass spectra for the sequential displacement of ammonia by DMA in [(NH4)8(HSO4)7]+ (the “8-7” ammonium bisulfate cluster) at (a) 0 and (b) 10 s after exposure to DMA. Cluster assignments are given in the form a(b)-c-d, where a represents the number of ammonium ions initially in the unreacted cluster, b represents the progress of the displacement reaction (i.e., the number of ammonium ions that have been displaced by DMA), c represents the number of bisulfate ions in the cluster, and d (indicated when nonzero) represents the number of neutral DMA molecules that have
11640
J. Phys. Chem. A, Vol. 114, No. 43, 2010
Figure 1. Mass spectra for the reaction of 8-7 ammonium bisulfate with DMA at t ) (a) 0 and (b) 10 s reaction time. Cluster ions are identified in the form a(b)-c-d, where a represents the number of ammonium ions initially in the unreacted cluster, b represents the progress of the reaction (i.e., the number of ammonium ions that have been displaced by DMA), c represents the number of bisulfate ions in the cluster, and d (indicated when nonzero) represents the number of neutral DMA molecules that have added onto the cluster (not displacement). Ions containing sodium are indicated by the “Na-” prefix.
added onto the cluster (not displacement). The addition of DMA to a cluster is described by Scheme 2. In addition, clusters substituted with sodium are indicated with the “Na-” prefix. At 0 s reaction time (Figure 1a), the predominant peak is that of the pure 8-7 ammonium bisulfate cluster (8(0)-7) at m/z 822.991, for which mass selection was performed in the quadrupole. Mass selection was accomplished by setting the m/z of interest to that of the 9-8 cluster (m/z 937.985). Although a relatively small amount of the 9-8 cluster (9(0)8) is isolated, the predominant ion isolated corresponds to the 8-7 cluster, likely due to decomposition of metastable 9-8 clusters during isolation. For all examined clusters, isolation was accomplished by selecting for a cluster larger by one ammonium bisulfate neutral, [(NH4)(HSO4)]. Some of the 7-6 cluster (not present in the displayed m/z range) is SCHEME 2
Bzdek et al. also isolated in small intensity. In addition, clusters in which a sodium ion has replaced an ammonium ion (Na-7(0)-7 and Na-8(0)-8) were observed. Finally, a very small amount of the first DMA substitution for each cluster (8(1)-7, Na-7(1)7, 9(1)-8, and Na-8(1)-8) was detected owing to residual DMA in the system. As the reaction progresses (Figure 1b), the ion intensity for each sequential substitution builds, reaches a maximum value, and subsequently decreases as the different substitution products are formed and react away (e.g., 8(5)-7, 8(6)-7, 8(7)-7), eventually resulting in the fully substituted dimethylaminium bisulfate cluster (8(8)-7). In this case, as was the case for most clusters studied, the addition of a single DMA molecule to the fully substituted cluster (8(8)-7-1) was also observed at the end of the reaction. In addition to a signal from m/z values corresponding to the cluster of interest, also present are signals from the partially and fully substituted 7-6 clusters. The signals for this cluster in part come from the small amount of it that was isolated initially, but also from decomposition of the 8-7 cluster as a result of the exothermic displacements. Previous work has indicated that initial substitution of DMA for ammonia is governed by the difference in proton affinity values between the incoming and departing ligands, whereas subsequent substitutions may be governed by the difference in solvation enthalpies between the two.18 For the reaction of DMA with ammonium bisulfate, initial displacement is highly favorable (∆G0 < -16 kJ/mol).18 Since these displacements occur in near-vacuum conditions, there is no sink for the energy that is released from the substitution. As a result, this released energy induces decomposition of the cluster through the loss of an ammonium bisulfate neutral. Throughout the course of the reaction, the summed intensity of all 8-7 clusters decreased by about 30% relative to the summed intensity of all 7-6 clusters. As discussed later, this decrease in intensity of the 8-7 clusters is too small to significantly affect the kinetic analysis. Unidentified peaks in both spectra are relatively small in intensity and, in fact, comprise
