ARTICLE pubs.acs.org/JPCA
Thermal Characterization of Aminium Nitrate Nanoparticles Kent Salo,† Jonathan Westerlund,† Patrik U. Andersson,† Claus Nielsen,‡ Barbara D’Anna,§ and Mattias Hallquist*,† †
Department of Chemistry, Atmospheric Science, University of Gothenburg, SE 412 96 G€oteborg Sweden Centre of Theoretical and Computational Chemistry, Department of Chemistry, University of Oslo, POB 1033 Blindern, N-0315 Oslo, Norway § Universite Lyon 1, Lyon, CNRS, UMR5256, IRCELYON, Institut de Recherches sur la Catalyse et L’Environnement de Lyon, Villeurbanne, F-69626, France ‡
ABSTRACT: Amines are widely used and originate from both anthropogenic and natural sources. Recently, there is, in addition, a raised concern about emissions of small amines formed as degradation products of the more complex amines used in CO2 capture and storage systems. Amines are bases and can readily contribute to aerosol mass and number concentration via acidbase reactions but are also subject to gas phase oxidation forming secondary organic aerosols. To provide more insight into the atmospheric fate of the amines, this paper addresses the volatility properties of aminium nitrates suggested to be produced in the atmosphere from acidbase reactions of amines with nitric acid. The enthalpy of vaporization has been determined for the aminium nitrates of mono-, di-, trimethylamine, ethylamine, and monoethanolamine. The enthalpy of vaporization was determined from volatility measurements of laboratory generated aerosol nanoparticles using a volatility tandem differential mobility analyzer set up. The determined enthalpy of vaporization for aminium nitrates range from 54 up to 74 kJ mol1, and the calculated vapor pressures at 298 K are around 104 Pa. These values indicate that aminium nitrates can take part in gas-to-particle partitioning at ambient conditions and have the potential to nucleate under high NOx conditions, e.g., in combustion plumes.
’ INTRODUCTION Aerosol particles are of general concern regarding their health effects, such as affecting the respiratory systems and inducing cardiovascular diseases.13 These effects are well established even though the detailed mechanism is not fully known. The chemical composition of aerosol particles is complex and their different properties will possibly influence the impact on human health. In addition to inorganic compounds thousands of different organic compounds contribute to aerosol mass.4 A complication of many organic compounds is that they may exist both in the gas and the aerosol phase, i.e. partitioning.5,6 This partitioning is important for two major reasons concerning health effects. First, aerosol particles and gases behave differently in the complex mass transport taking place in our respiratory-lung system. Second, the atmospheric lifetime is different if a compound is present in the gas or in the particulate phase. In the gas phase the removal process is chemical reactions, whereas particles are subject to both dry and wet deposition. A key physical property to describe the partitioning of any compound is the vapor pressure.5,6 Organic amines are low molecular weight bases with high vapor pressures (p0298K of 10 to 105 Pa). Still, it has recently been evident that these compounds can contribute to the atmospheric aerosol mass via a number of mechanisms. First, acidbase reactions can neutralize acidic aerosols by uptake in the condensed r 2011 American Chemical Society
phase and produce the corresponding salts, e.g., aminium chlorides, aminium sulphates, and aminium nitrates.79 Second, there are suggestions that aliphatic amines can form amides with organic acids in the condensed phase and thereby form stable macromolecules of low volatility.10,11 Third, the reactions of gaseous amines with atmospheric oxidants have been proved to lead to secondary organic aerosol (SOA) formation.1215 In the atmosphere, saturated aliphatic amines will react with Cl, NO3, and OH radicals by abstraction of a hydrogen either from the carbon or the nitrogen position, where theoretical studies suggest abstraction of the hydrogen in the carbon position to be favorable.16 The hydrogen abstraction will result in the formation of an alkyl or an alkylamino radical species that will promptly react further to form a wide range of closed shell products, e.g., aldehydes, amides, nitrosamides, and nitramines, that may contribute to the organic nitrogen fraction of the atmospheric aerosol. These three mechanisms may all contribute to the formation of particles that contain toxic and/or carcinogenic compounds. It is therefore of importance to establish the roles of these mechanisms and link them to the atmospheric conditions and sources of the amines. Received: May 27, 2011 Revised: September 1, 2011 Published: September 12, 2011 11671
dx.doi.org/10.1021/jp204957k | J. Phys. Chem. A 2011, 115, 11671–11677
The Journal of Physical Chemistry A
ARTICLE
Table 1. Physical Properties of the Aminium Nitrates Used in This Work compound AN
molecular weight (g mol1)
density (g cm3)
melting point (K)
80.04
1.741.76 e
443 f
γa (J m2) 0.1 m
fp
94.07
1.211.27 g
383 j
0.1 m
ep
108.09
1.201.23 g
350 k
0.1 m
ep
122.12
1.251.28 g
428 l
124.10
850.56
1.241.30 , 1.26 h
201n
4.93
735.36
256 n
5.34
672.00
0.1 m
311 n
5.70
821.76
0.0506 h
302 o,p 274 n
5.46
622.08
5.34
549.12
ep
MEA-nitrate
4.43
251 o,p
1.231.30 , TMA-nitrate
εii/kBd (K)
202 o,p
1.401.42 , DMA-nitrate
146 n
σiic (Å)
155 o,p
1.75 , MMA-nitrate
Vcb (cm3 mol1)
324 h
259 o,p EA-nitrate
108.09
1.21 i
286 i
0.0473 h
256 n 260 o,p
a Surface free energy. b Critical volume. c Interparticle distance where the potential is zero calculated according to Bird et al.39 d Depth of the potential energy well calculated according to Bird et al.39 e This work (BET-method). f Product data sheet. g This work (AMS-method), shape factor 1 assumed. h Greaves et al.46 i Atkien et al.45 j Mylrajan et al.58 k Walden et al.59 l Jain et al.60 m Estimated. n Calculated according to Lydersen et al.47 o Calculated according to Nanoolal et al.61 p Values used in this work.
A detailed list of sources of atmospheric amines was published by Ge et al.17 Major sources are of both natural and anthropogenic origin and examples are biological processes in the oceans, combustion, agriculture, and waste treatment. In addition to these sources, there is an increasing demand to use amines in carbon capture and storage techniques (CCS). The amines are used to trap CO2 in the flue gases in combustion plants and can be a potential additional emission source of the amines used and their degradation products.18 This source will be specific in its coemission with combustion related products, e.g., particles, VOC, NOx, and SO2. To establish the atmospheric fate of these amines and their products, there is a need to develop methods to determine their gas to particle partitioning.17,19 Recently, there have been large research efforts put into the measurements and modeling of the vapor pressures and enthalpies of vaporization/ sublimation of low volatile atmospheric relevant compounds, e. g., Salo et al.,20 Cappa and Jimenez,21 and references therein. In brief, there are methods that determine the thermal properties from evaporation of suspended particles e.g. using a volatility tandem differential mobility analyzer (VTDMA),2026 thermodenuder systems,27,28 electro dynamic balance (EDA), or optical tweezers.29,30 Alternatively, one may derive the vapor pressures from bulk samples, e.g., using Knudsen effusion cells.3136 Furthermore, there are several empirical computational methods to estimate vapor pressures as described in a recent assessment by Barley et al.37 The main caveat is how to measure the low vapor pressures needed to assess atmospheric transformation. The method applied and developed for the current work is a variant of the VTDMA technique earlier used and described by Jonsson et al.38 and Salo et al.20 This has now for the first time been applied to alkyl aminium nitrates that are supposed to be formed by the acid base reaction of alkyl amines with nitric acid in the atmosphere. Specifically, the objective of this study was to determine the volatility, i.e. the rate of evaporation of aminium nitrates and infer the enthalpies of vaporization and saturation vapor pressures.
’ EXPERIMENTAL SECTION Volatility Measurements. A VTDMA system was used to
determine the volatility properties of alkyl aminium particles and
ammonium nitrate (AN). The aminium nitrates investigated were the nitrate salts of monomethylamine (MMA), dimethylamine (DMA), trimethylamine (TMA), ethylamine (EA), and monoethanol amine (MEA) with selected physical properties listed in Table 1. The system used in this study has been described in detail previously by Jonsson et al.38 and Salo et al.20 The thermal properties of aerosol particles were inferred from the rate of evaporation of individual suspended particles at selected temperatures. The sample aerosols were generated by nebulizing aqueous solutions of the pure compounds using a TSI 3076 constant output atomizer. The nebulized aerosols were dried with a silica diffusion dryer to relative humidities (RHs) < 5% but still the generated aerosol particles were assumed to be supersaturated droplets (see Salo et al.20). A narrow, close to monodisperse size distribution of the sample aerosol was achieved using a differential mobility analyzer (TSI 3071). The initial aerosol sample was already confined and by selection a fraction of the monodispers aerosol from the larger end of the initial size distribution eliminated any potential effect of multiple charges. The evaporation of the size selected particles was done using an oven consisting of a stainless steel tube (50 cm, 6 mm i.d.), embedded in a temperature controlled aluminum block. After the heating section a charcoal denuder (16 cm) was used to remove the evaporated vapors to eliminate any potential recondensation. Eight oven/denuder systems were mounted in parallel to enable swift changes in temperatures. Under standard operating conditions a flow rate of 0.3 L per minute (SLPM) were used. The change in the particle median diameter was monitored using a scanning mobility particle sizer (SMPS). The SMPS system was size calibrated using nebulized polystyrene latex spheres (PLS; Duke Standards, 3K/4K Series Particle Counter Standards). To illustrate and visually compare the volatility for different systems the change in diameters of these narrow size distributions were used to calculate the volume fraction remaining (VFR), eq 1. 3 Df ð1Þ VFR ¼ Di where Df is the final diameter after evaporation and Di is the initial diameter after the reference oven. It should be stressed that using Di measured with the SMPS system after the reference 11672
dx.doi.org/10.1021/jp204957k |J. Phys. Chem. A 2011, 115, 11671–11677
The Journal of Physical Chemistry A
ARTICLE
Figure 1. Temperature dependence of volume fraction remaining (VFR) for aerosol particles from AN-nitrate (pale-gray closed circles), DMA-nitrate, (open circles), MEA-nitrate, (closed circles), and suberic acid (gray closed circles). The two horizontal dashed lines mark the VFR interval used for calculations.
oven compensate for the effect of any potential evaporation in the selecting differential mobility analyzer and the tubing before the heating section. Thus, the difference between Di and Df is a result of the evaporation in the heated section. In Figure 1, the temperature dependence of VFR is shown for AN, DMA-, and MEA-nitrate together with data from suberic acid derived in a previous study.20 In order to extract and calculate vapor pressures from the experimental data a number of physical parameters were used (see Table 1 and eq 2). ! Z Df Dp Fi RT 4γi Mi 0 exp dDp ð2Þ pi ¼ f ðKni , αÞ Dp Fi RT 4Di, air ΔtMi
experimental conditions and the input parameters to eq 1, providing an upper limit on the investigated systematic errors of (18%, as described in the Results and Discussion section. Density Measurements. The density of three aminium salts has been measured using two methods. The first is based on direct measurements of the salt volume using a Brunauer EmmetTeller (BET) surface apparatus.41 The empty volume (V0) of a cell is determined by introducing N2, a known amount of the pure salt is placed in the cell, and then a primary vacuum is then applied at room temperature until the pressure is stable. Finally, N2 is introduced and a new volume V1 is determined. The difference, V0 V1, provides the volume of the salt. From the known mass (weighted before the experiment) the density is derived: F = m/V. Traces of water adsorbed to the salt were removed by a second series of experiments where the samples were dried at 85 °C under primary vacuum for 1.5 h . For some cases the procedure removed a part of the sample but sufficient material was always left for the subsequent density analysis. The second method provides the effective density from suspended particles generated in the same way as in the vapor pressure measurements which was measured by comparing the size distribution derived by a compact time-of-flight aerodyne aerosol mass spectrometer (AMS) and an SMPS system. The AMS measures the aerodynamic diameter (Dva) due to the acceleration of the particles in the vacuum, while the SMPS measures the mobility diameter (Dm) at 1 atm. The two diameters are related through the density of the particle (Fp) and the Jayne shape factor (Sp)42 as is seen in eq 4. (Sp = 1 for spherical particles and Sp < 1 for nonspherical particles, e.g., NH4NO3 has a shape factor of 0.8.42) The resulting densities are listed in Table 1. Dva ¼ Sp Fp ¼ Fef f Dm
Di
Equation 2 was used to derive the saturation vapor pressure (p0) of compound i at evaporative temperature T. Fi is the particle density, R is the gas constant, and Di,air is the diffusivity of molecule i in air which was calculated according to Bird et al.39 with parameters σii and εii, Mi is the molar mass of compound i. Δt is the median residence time of the aerosol assuming a fully developed laminar parabolic flow neglecting any radial transport due to diffusion and thermophoresis. A temperature correction was applied to the volumetric flow rate that was set and determined at room temperature. In the second term f(Kni, α) is the semiempirical Fuchs and Sutugin correction term for particle diameters in the transition regime40 (eq 3) where Kni is the Knudsen number and α is the accommodation coefficient and was set to unity in this work. f ðKni , αÞ ¼
1 þ Kni ð1 þ Kni Þ 1 þ 0:3773Kni þ 1:33Kni α
ð3Þ
In the third term of eq 2, γi is the surface free energy. Further details on how to derive the vapor pressures from these data and associated uncertainties are described in Salo et al.20 In summary, this type of instrument can measure vapor pressures of aerosol compounds in the range of 1011 to 104 Pa if the following assumptions are used: (1) the evaporating particles are spherical, (2) the evaporated molecules are not recondensing, (3) the surface free energy is constant and isotropic, and (4) any latent heat can be disregarded. In addition to reported statistical errors an uncertainty analysis was carried out on the basis of
ð4Þ
Synthesis. The MMA, DMA, TMA, and EA nitrate salts were prepared by passing a stream of the gaseous amine through dilute nitric acid solutions until the solutions were slightly alkaline. The MEA-nitrate salt was prepared by adding dilute nitric acid to an aqueous solution of the amine until the solution was slightly acidic. The excess amine/nitric acid was then driven off by heating the solutions to 50 °C under dry conditions until the salts eventually solidified. Quantum Chemistry Calculations. Quantum chemistry calculations were performed using Gaussian 03 programs.43 All structures were optimized in G3 calculations.44
’ RESULTS AND DISCUSSION Vapor Pressures and Enthalpies of Vaporization. A thorough analysis of the thermal properties of the nitrate salts of MMA, DMA, TMA, EA, and MEA have been done. Median diameter was determined as a function of temperature for all five compounds, and corresponding vapor pressures and VFR were calculated at respective temperature. From the calculated vapor pressure at each temperature the data was evaluated assuming a ClausiusClapeyron relationship, i.e., a linear dependence of log p0 versus 1/T. As an example, Figure 2 shows this type of data for the same compounds as in Figure 1. The resulting enthalpy of vaporization and estimated vapor pressure at 298 K for all the investigated compounds are listed in Table 2. In this work the reported enthalpies of vaporization (ΔHvap) was determined at temperatures somewhat higher 11673
dx.doi.org/10.1021/jp204957k |J. Phys. Chem. A 2011, 115, 11671–11677
The Journal of Physical Chemistry A
ARTICLE
than 298 K and the saturation vapor pressures (p0) were extrapolated to 298 K using the ClausiusClayperon relationship. It should be noted that provided data is assumed to be derived from evaporation of a liquid and not from a crystalline phase, i.e., enthalpy of vaporization and not enthalpy of sublimation, see Salo et al. for a detailed discussion on the phase of investigated nanoparticles.20 Furthermore, ΔHvap was obtained by assuming that it is constant over the studied temperature interval, as evident by a linear dependence of log p0 versus 1/T, Figure 2. Density of Aminium Nitrates. Table 1 lists the results from the AMS and the BET measurements after removal of any adsorbed traces of water at 85 °C for 1.5 h. The BET method was considered more accurate in determining the absolute density of the salt than the AMS method. In general the two methods are in quite good agreement for DMA- and TMA- nitrate salts with a maximum variability of (10%. For MMA-nitrate salt the SMPSAMS analysis gave a density of 1.23 ( 0.03 g cm3 while the BET method gave 1.40 ( 0.02 g cm3. Such discrepancy may suggest that suspended MMA-nitrate salt particles are possibly nonspherical and taking as reference the density calculated with the BET method a shape factor of 0.88 is derived. The calculations in this work were performed using the densities measured with the BET method. Sensitivity Analysis. Statistical errors at the 95% confidence level are reported for each amine nitrate and given in Table 2. Selected systematic errors were considered in an uncertainty
analysis. Here estimated uncertainties in the parameters used in eq 2 were evaluated including experimental uncertainties e.g. related to flow rates and evaporative temperatures. This detailed analysis was applied on the evaluation procedure and experimental data for TMA-nitrate and a summary is presented in Table 3. The main conclusion is that the uncertainty of these parameters did not impact ΔHvap by more than 1% (i.e., the slope of the ClausiusClayperon plot). The uncertainties were larger for the vapor pressures extrapolated to 298 K. The sample flow and the evaporative temperature of the oven are believed to be controlled within (0.05 LPM and (1 K resulting in a change in vapor pressure of less than 18% and 11%, respectively. Surface free energy for the aminium nitrates is scarcely reported in the literature. Only values for MEA- and EA-nitrate were found.45,46 For the other compounds it was estimated to be 0.1 J m2. A fairly large perturbation (0.010.2 J m2) resulted in less than 17% change in vapor pressure. The interparticle distance where the potential is zero was calculated from the critical volume which is estimated using the Lydersen group contribution method.47 For the amine salts this estimation was not trivial and in the estimates the group contribution of a nitro group was used instead of the nitrate group. The uncertainty of using a nitro group instead of a nitrate group was assumed to be no larger than the resulting critical volume when adding and withdrawing an extra nitro group. This changed the vapor pressure by about 10%. Depth of the potential energy well (ε/kB) is calculated using melting temperatures of the aminium nitrate with an uncertainty of (3 K which resulted in less than 1% change in vapor pressure. The calibration of the SMPS system using PLS generates a correction factor with a statistical error of approximately (1.1%. Applying that error in the analysis changed the obtained p0 no more than 2%. The densities used in this work were measured as described earlier and also reported in the ADA 2010 final report (Annex).48 An estimated error of (0.1 g cm3 changed the vapor pressure by less than 11%. Ammonium Nitrate. The obtained enthalpy of vaporization for ammonium nitrate is of the same magnitude as in previous studies by Brandener et al. and Hildenbrandt et al using two alternative techniques.35,49 The saturation vapor pressure for AN nanoparticles of 6.79 104 Pa in this work is lower than the values reported by the earlier studies, see Table 2. The reason for this is unclear but we note the large discrepancy between the two previous studies. Recently, Hildebrandt et al. and Chien et al. provided new insights into the evaporative mechanisms.34,35,50 In
Figure 2. Temperature dependence of saturation vapor pressures of ammonium nitrate (pale-gray closed circles), DMA-nitrate (open circles), MEA-nitrate (closed circles), and suberic acid (dark-gray closed circles). The error bars are the 95% confidence interval of the linear regression.
Table 2. Enthalpy of Vaporization and Calculated Vapor Pressures this study compound AN
MMA-nitrate DMA-nitrate TMA-nitrate MEA-nitrate EA-nitrate
p0298
4
(10
Pa)
6.79+2.2 1.7
2.88+0.6 0.5 3.92+1.2 0.9 5.37+1.1 0.9 0.89+0.4 0.3 2.68+14 2.2
literature
1
ΔrH298a
ΔHvap (kJ mol ) 72 ( 5.3 (309 330 K)c
1
(kJ mol )
50.6
p0298
4
(10
ΔHvap (kJ mol1)
Pa)
14.3 b, 30.1*,b
89.9 b, 83.4*,b
22.4 d
97.2 ( 12.6 e (332.8 360.4 K)c
f,+
107 ( 20.9 g,+ (321.0 360.4 K)c n.a.
c
59.3
3.45 n.a.
63 ( 3.9 (315 339 K)c
62.7
n.a.
n.a.
66 ( 3.2 (308 333 K)c
68.6
n.a.
n.a.
74 ( 3.9 (330 348 K)c
55.5
n.a.
n.a.
54 ( 18 (323 343 K)c
61.5
n.a.
n.a.
65 ( 2.3 (318 345 K)
Enthalpies of reaction for R1R2R3N(g) + HNO3(g) f R1R2R3NH+NO3 from G3 calculations. b Brandner et al.49 c Valid for the given temperature interval. d Extrapolated to 298 K using data by Chien et al.34 e Hildenbrand et al.35 f Extrapolated to 298 K using data from Hildenbrand et al.35 g Hildenbrand et al.50 * Over liquid. + Sublimation, NH4NO3 (c) f NH4NO3 (g). a
11674
dx.doi.org/10.1021/jp204957k |J. Phys. Chem. A 2011, 115, 11671–11677
The Journal of Physical Chemistry A
ARTICLE
Table 3. Sensitivity Analysis for Vapor Pressures Derived from TMA-Nitrate Aerosol Particles
Table 4. Physical Properties at 298 K of the Pure Amines Used in This Work
P0 (298 K) factor
base value
4
Pa)
p0e (kPa)
ammonia
9.25
72.5
23.26 (195406 K a)
1020 351 205
MMA DMA
10.66 10.73
154 180
26.11 (180430 Ka) 26.64 (181438 K a)
+11
TMA
9.81
254
23.44 (156433 K a)
222
7
MEA
9.50
225
51.53 (181438 Ka)
0.041
10.81c
207
27.35 (192456 K a)
145
0.05 LPM +0.05 LPM
Tvapa
303 333 K
1 K
5.37+0.57 0.36
+1 K 0.09 J m2
5.37+0.88 0.68
2
+0.44 Å
% change
+17 13
+0.1 J m 5.7 Å
ΔHvapd (kJ mol1)
+18 16
0.3 LPM
σc
Vcd (cm3 mol1)
(10
sample flow
0.1 J m2
pKab
perturbation
5.37+0.93 0.84
γb
compound
5.37+0.66 0.53
0.52 Å
+12 10
ε/kBd
821.76 K
+5.76 K 5.76 K
5.37+0.01 0.01
1
correction factor e
1.152
+0.011
5.37+0.13 0.04
+2
Ff
1.27 g cm3
0.011 +0.1 g cm3 0.1 g cm3
>1 5.37+0.56 0.39
+11 7
a
Evaporative temperature. b Surface free energy. c Interparticle distance where the potential is zero estimated using the Lydersen method.47 d Depth of potential energy well. e Correction factor from SMPS calibration. f Density according to Table 1.
addition, AN evaporation measured in the present study is on the upper limit of the evaporation values that could be determined with the VTDMA setup. The limiting factor is the evaporation of a too volatile compound in the differential mobility analyzer and in the reference oven (298 K) providing difficulties in obtaining a stable reference diameter. This is reflected in the data shown in Figure 1 where the AN thermal profile has significantly larger scattering than the other thermal profiles. The enthalpy of vaporization measured in this work is somewhat lower than the recent values of 97.2 ( 12.6 kJ mol1 reported by Hildebrandt et al. based on measurements of gas phase composition above a bulk phase sample.35 Alkyl and Alcohol Aminium Nitrates. For the first time measurements of the vapor pressures and enthalpies of vaporization of alkyl aminium nitrates are reported. The corresponding amines, EA, MMA, DMA, and TMA are rather volatile compounds with similar physical properties, see Table 4. The measured vapor pressures and enthalpies of vaporization for the nitrate salts of these amines are within experimental errors the same: p0 ranging from 2.68 to 5.37 104 Pa, ΔHvap ranging from 54 to 66 kJ mol1. For the alcohol aminium nitrate (MEA-nitrate), the derived vapor pressure (6.2 105 Pa) is significantly lower than for the aliphatic aminium nitrates. The corresponding enthalpy of vaporization is also significantly different, 74 kJ mol1 compared to 5466 kJ mol1 for the aliphatic aminium nitrates. This is attributed to the additional polarity due to the hydroxyl group; well-known to reduce vapor pressure and increase enthalpy of vaporization. This is clearly seen when comparing the literature vapor pressures (0.041 and 145 kPa) see Table 4 and enthalpies of vaporization (51.53 and 27.35 kJ mol1) for MEA and EA, respectively. A question is by which mechanism the evaporation occurs. There are in principle three distinct routes of evaporation. (1) The salt transfers to the gas phase as a molecular entity. In this process the measured enthalpy is the enthalpy of vaporization. Recently, Hildebrant et al. demonstrated that 20% of AN evaporate following this pathway.50 (2) The precursor amine
EA a
Valid for the given temperature interval. b Handbook of chemistry and physics.62 c At 303 K. d Yaws. 63 e Calculated at 298 K using Antoine parameters provided by Yaws et al .63
and nitric acid evaporates as neutral molecules. Here the evaporation occurs as a result of proton transfer at the surface of the particle. The energy barrier for this proton transfer is not well-defined but it must be less than the enthalpy of reaction for the gas phase reaction R1R2R3N + HNO3 f R1R2R3NH+NO3, which has been estimated in quantum chemistry calculations as displayed in Table 2. There are, however, additional steps in this evaporation process. The limiting step is either the dissociation of the aminium nitrate forming HNO3 and the corresponding amine on the surface that subsequently evaporates, or the limiting step is the evaporation of HNO3 or the amines from the surface. The range of enthalpies measured was thus compared to HNO3 surface adsorption data to elucidate the possibility for a limitation by HNO3 leaving the surface. Here information is scarce and the quantity strongly depends on surface properties, e.g. the number of potential hydrogen bonds. For instance HNO3 on ice is found to be bonded by in between 30.3 and 54.0 kJ mol1 5153 and on hexane soot by 4867 kJ mol1.54,55 These values are all somewhat lower than the measured enthalpy change in the present study; however, the surface properties of the amine nitrate are different compared to soot and ice surfaces, and it is impossible without further work to establish if this might be a plausible route. Support for this mechanism may be the relatively small differences in enthalpy between the alkyl aminium nitrates. In the third possible route (3) the evaporation is via gaseous ions and this pathway requires too much energy to occur under tropospheric conditions. Regarding route (1) and (2), they are both possible and since no measurements of the gas phase composition were done it is difficult to distinguish between these. In another attempt to shed light on the mechanism Figure 3 shows derived vapor pressure at 298 K and enthalpy of vaporization versus corresponding proton affinity of the base forming the nitrate salt. For the alkyl aminum nitrates the values of vapor pressures and enthalpies are all similar within the statistical errors and any dependence on proton affinity becomes inconclusive. However, a significant feature was the lower vapor pressure and higher enthalpy of vaporization for MEA nitrate and the inorganic ammonium nitrate. MEA has a significantly lower pKa and proton affinity than the other amines but the influence on the vapor pressure of the nitrate salt was rather assigned to interaction of the hydroxyl group with surrounding molecules, i.e., intermolecular interactions. There are several models available to calculate vapor pressures and enthalpies of vaporization, even though limitations exist when coming to high molecular weight compounds with vapor pressures
