The inhibition of N2O5 hydrolysis in sulfuric acid by 1-butanol and 1-hexanol surfactant coatings

Article abstract of DOI:10.1021/jp068228h

Gas - liquid scattering experiments are used to measure the fraction of N₂O₅ molecules that are converted to HNO₃ after colliding with 72 wt % H₂SO₄ containing 1-hexanol or 1-butanol at 216 K. These alcohols segregate to the surface of the acid, with saturation coverages estimated to be 60% of a close-packed monolayer for 1-hexanol and 44% of a close-packed monolayer for 1-butanol. We find that the alkyl films reduce the conversion of N₂O₅ to HNO ₃ from 0.15 on bare acid to 0.06 on the hexyl-coated acid and to 0.10 on the butyl-coated acid. The entry of HC1 and HBr, however, is enhanced by the hexanol and butanol films. The hydrolysis of N₂O₅ may be inhibited because the alkyl chains restrict the transport of this large molecule and because the alcohol OH groups dilute the surface region, suppressing reaction between N₂O₅ and near-interfacial H ₃O⁺ or H₂O. In contrast, the interfacial alcohol OH groups provide additional binding sites for HC1 and HBr and help initiate ionization. These and previous scattering experiments indicate that short-chain alcohol surfactants impede or enhance sulfuric acid-mediated reactions in ways that depend on the chain length, liquid phase acidity, and nature of the gas molecule.

Full text of DOI:10.1021/jp068228h

J. Phys. Chem. A 2007, 111, 2921-2929
2921
The Inhibition of N2O5 Hydrolysis in Sulfuric Acid by 1-Butanol and 1-Hexanol Surfactant
Coatings
Seong-Chan Park, Daniel K. Burden, and Gilbert M. Nathanson*
Department of Chemistry, UniVersity of Wisconsin-Madison, 1101 UniVersity AVenue,
Madison, Wisconsin 53706
ReceiVed: NoVember 30, 2006; In Final Form: February 16, 2007
Gas-liquid scattering experiments are used to measure the fraction of N2O5 molecules that are converted to
HNO3 after colliding with 72 wt % H2SO4 containing 1-hexanol or 1-butanol at 216 K. These alcohols segregate
to the surface of the acid, with saturation coverages estimated to be 60% of a close-packed monolayer for
1-hexanol and 44% of a close-packed monolayer for 1-butanol. We find that the alkyl films reduce the
conversion of N2O5 to HNO3 from 0.15 on bare acid to 0.06 on the hexyl-coated acid and to 0.10 on the
butyl-coated acid. The entry of HCl and HBr, however, is enhanced by the hexanol and butanol films. The
hydrolysis of N2O5 may be inhibited because the alkyl chains restrict the transport of this large molecule and
because the alcohol OH groups dilute the surface region, suppressing reaction between N2O5 and near-interfacial
H3O+ or H2O. In contrast, the interfacial alcohol OH groups provide additional binding sites for HCl and
HBr and help initiate ionization. These and previous scattering experiments indicate that short-chain alcohol
surfactants impede or enhance sulfuric acid-mediated reactions in ways that depend on the chain length,
liquid phase acidity, and nature of the gas molecule.
Introduction
the abundance nor the speciation of these organic molecules is
firmly established, but they are expected to range from small
molecules such as methanol and acetone to long-chain fatty
acids.21-24 These molecules may segregate to the surface of the
acid droplets and coat them, potentially impeding interfacial
transport and suppressing N2O5 hydrolysis. Organic coatings
have been recently suggested to be one explanation for the
variability in N2O5 hydrolysis rates on acidic particles over the
northeast United States.25 In this article, we focus on the effects
of two soluble alcohols, 1-butanol and 1-hexanol, on N2O5
hydrolysis in 72 wt % H2SO4 at 216 K and compare hydrolysis
rates with HCl and HBr uptake into the bare and film-coated
acids.
Insoluble and soluble organic species behave very differently
at the surfaces of aqueous solutions. Numerous experiments
show that insoluble, long-chain surfactants such as hexadecanol
(C16H33OH) form compact monolayers on water and sulfuric
acid that inhibit rates of water evaporation by 104 or more, with
resistances that grow exponentially with chain length.26 Kinks
in the chain moderate this resistance substantially, as demon-
strated recently by the 10-fold greater permeation of acetic acid
through films of cis-9-octadecen-1-ol (C18H36O) on water than
through the straight-chain 1-octadecanol.27 Permeation data are
scarce for short-chain surfactants: slightly soluble C8-C12
alcohols in water may impede H2O and CO2 transport,28 and
organic films can even enhance uptake when the gas is more
soluble in the surface film than in the subphase, as shown by
the adsorption of anthracene onto 1-octanol-coated water.29
Conversely, recent fluorescence experiments indicate that 1-oc-
tanol films on water suppress interfacial pH changes induced
by exposure to HNO3 or NH3.30
The aerosol-mediated conversion of N2O5 into HNO3 is a
key step in regulating ozone levels in the stratosphere and
troposphere and in denitrifying the troposphere through wet and
dry deposition of nitric acid.1-6 Because N2O5 is formed at night
from NO2 and NO3 and photolyzes back into these species
during the day, the reaction
N2O5 + H2O aerosol8 2HNO3
(1)
effectively converts NO2 and NO3 into the temporary reservoir
species HNO3. In the lower stratosphere, this reaction suppresses
the NO/NO2-catalyzed destruction of ozone while enhancing
both the HO/HO2 cycle (through photolysis of HNO3 into OH
and NO2) and the Cl/ClO cycle (by curtailing formation of
ClONO2) that reduce O3.7 Inclusion of this hydrolysis reaction
in models of the lower stratosphere and upper troposphere lead
to better agreement between predicted and observed O3 depletion
rates and ratios of nitrogen oxide species.8,9 In the global
troposphere, N2O5 hydrolysis occurs on a variety of aqueous
surfaces10 and is estimated to reduce NO and NO2 levels by
∼40% and O3 levels by ∼5% during winter months.11-13
The hydrolysis of N2O5 is typically mediated by submicron
sulfuric acid particles in the lower stratosphere that range from
40 to 80 wt % H2SO4 at temperatures from 200 to 240 K, with
wider variations in acidity in the upper troposphere due to
absorption of NH3.3,4 Laboratory studies demonstrate that N2O5
hydrolysis occurs with a probability near 0.1 in sulfuric acid
aerosols over a wide range of acid concentrations at low
temperatures.14-19 Field measurements indicate that, in the upper
troposphere and tropopause regions, these aerosol particles
contain significant concentrations of organic species.20 Neither
Most pertinent to the present study are experiments by
Thornton and Abbatt, which demonstrate that a monolayer of
hexanoic acid on synthetic seawater aerosol reduces N2O5
hydrolysis by 3- to 4-fold.31 Further experiments by McNeill
* To whom correspondence should be addressed. E-mail: Nathanson@
chem.wisc.edu.
10.1021/jp068228h CCC: $37.00 © 2007 American Chemical Society
Published on Web 03/24/2007

2922 J. Phys. Chem. A, Vol. 111, No. 15, 2007
Park et al.
Figure 1. Several pathways for an N2O5 molecule colliding into sulfuric
acid containing 1-hexanol. The τ values refer to the average liquid-
phase residence times of unreacted N2O5 and HNO3.
et al. show that sodium dodecyl sulfate coatings on natural
seawater and NaCl aerosols reduce hydrolysis of N2O5 10-fold.32
Folkers et al. and Anttila et al. also showed that aqueous sulfate
particles coated with thick, multilayer films formed from
monoterpene oxidation reduce N2O5 hydrolysis by factors of
10 or more.33,34
N2O5 molecules can follow a range of pathways upon
collision with a submonolayer organic film on sulfuric acid, as
illustrated in Figure 1 for acid coated with hexanol molecules
in different chain conformations.35,36 At thermal collision
energies, nearly all impinging N2O5 molecules will be momen-
tarily trapped at the surface, and only a small fraction will recoil
directly (inelastic scattering).37,38 Some of the thermalized N2O5
may then desorb from the surface before or after moving
between the hexyl chains, while others may permeate through
the porous film and undergo hydrolysis near the film-acid
interface or deeper into the acid.
Our previous studies of butanol and hexanol coatings on 56
to 68 wt % D2SO4 (0.20 to 0.30 mole fraction acid) at 213 K
led us to suspect that these films would not significantly inhibit
N2O5 passage into the acid because they do not form compact
monolayers.37-40 Each alcohol reacts with H2SO4 to form
mixtures of ROH, ROH2+, ROSO3-, and ROSO3H, which are
expected to be in roughly equal proportions in 70 wt % H2SO4
at 298 K.41 On the basis of surface tension measurements of 72
wt % H2SO4 mixed with butanol and hexanol at 295 and 250
K, the total surface concentration of all alkyl species at 216 K
is predicted to saturate at ∼3.0 × 1014 cm-2 (∼60% of a close-
packed monolayer) for 1-hexanol and ∼2.2 × 1014 cm-2 (∼44%
of a close-packed monolayer) for 1-butanol.38 Our first studies
indicated that butyl films on 56-68 wt % D2SO4 at 213 K do
not impede water evaporation and actually enhance HCl and
HBr uptake by providing extra OH surface groups that act as
protonation sites for HCl and HBr dissociation.37,39 When
hexanol is substituted for butanol, the entry of HCl is still
enhanced, except when the acid concentration is low enough
(56 wt % D2SO4) to reduce the fraction of HexOH2+ and allow
the hexyl film coverage to rise to ∼68% of a compact
monolayer.38 At this higher packing, HCl uptake is reduced from
0.68 to 0.59 and water evaporation is reduced by 20%. In
contrast to the enhanced uptake of HCl and HBr by the alcohol
films, the uptake of the basic molecule CF3CH2OH is nearly 1
at all acid concentrations and is impeded by less than 4% by
both butyl and hexyl films.
Figure 2. Molecular beam scattering apparatus and liquid reservoir.
The Teflon reservoir is sealed except for a 0.9 cm diameter hole through
which the N2O5 beam strikes the acid. Only one of the prechopper or
postchopper wheels is used at one time.
N2O5 hydrolysis differently than HCl and HBr uptake, and they
may act differently on sulfuric acid or other subphases. The
experiments below utilize 72 wt % H2SO4 or 70 wt % D2SO4
at 216 K, each 0.32 mole fraction acid, which are highly viscous
(∼1900 cP) and low vapor pressure (∼2 × 10-4 Torr)
liquids.42,43 They show that, although the hexyl and butyl films
significantly enhance HCl and HBr uptake, the conversion of
N2O5 to HNO3 drops by 60% and 33%, respectively, indicating
that submonolayer films can impede this near-surface reaction.
Experimental Procedure
N2O5 and Acid Preparation. N2O5 is synthesized by
oxidizing NO with O2 and O3 consecutively at room tempera-
ture, purifying the product by distillation under O3 and storing
it at 195 K.44 HNO3 formation is minimized by baking the
apparatus and by passing NO and dried O2 through a 195 K
cold trap to remove H2O. The melting point of the N2O5 sample
under argon was observed to be 39-42 °C, which encompasses
the literature value of 41 °C.45 During the scattering experiments,
the sample is held at 253 K, where the N2O5 vapor pressure is
estimated to be 8 Torr.46 Incident beams of 13 and 150 kJ mol-1
N2O5 are created by expanding pure N2O5 or N2O5 seeded in
750 Torr H2 through an unheated, 0.13 mm diameter glass
nozzle.
Surfactant/acid solutions are prepared by diluting 97 wt %
H2SO4 to 72 wt % with Millipore water, adding 1-hexanol (up
to 0.1 M) or 1-butanol (up to 1 M) and immediately cooling to
216 ( 0.5 K in the liquid reservoir shown in Figure 2. The
alcohols (Aldrich, >99% pure) are used without further
purification. Titrations indicate that the acid concentration
changes by no more than 0.5 wt % during the time it is in
vacuum (typically less than 3 days). The measured reductions
in surface tension upon adding alcohol are those expected when
adding soluble, small-molecule surfactants, indicating that the
alcohols are not contaminated with long-chain insoluble
species.40
A continuously renewed, vertical liquid film is formed by
rotation of a 5.0 cm diameter glass wheel partially submerged
in 60 mL of the acid solution.39 The acid-coated wheel is
skimmed by a cylindrical Teflon scraper, which removes the
outer 0.1 cm of acid and alcohol. The remaining ∼0.03 cm thick
acid film then passes in front of a 0.7 cm2 circular hole, where
it is intercepted by the N2O5 beam impinging at an incident
The observations above indicate that gas transport through
surface films is controlled by the chain length of the surfactant,
the identity of the gas molecule, and the underlying liquid.
Hexanol and butanol films on sulfuric acid may therefore alter

Inhibition of N2O5 Hydrolysis in Sulfuric Acid
J. Phys. Chem. A, Vol. 111, No. 15, 2007 2923
angle θinc of 45°. At a typical wheel speed of 0.17 Hz, the time
interval between scraping and exposure to the beam is 0.49 s.
Argon scattering measurements show that this time is sufficient
to allow the butyl and hexyl films to become reestablished at
the surface of the acid, as shown previously in refs 38 and 39.
The acid remains in front of the hole in the reservoir for
0.45 s and is exposed to the incident beam for a time texp equal
to 0.27 s.
Distinguishing N2O5 and HNO3 by Pre- and Postchopper
TOF Analysis. Time-of-flight (TOF) spectra of N2O5 and HNO3
exiting from the acid are recorded by a differentially pumped
quadrupole mass spectrometer oriented at an angle θfin ) 45°,
as shown in Figure 2. Electron-impact ionization is used to detect
the neutral molecules with an electron energy of 70 eV.
+
Unfortunately, the N2O5 parent ion is unstable and all stable
N2O5 ion fragments are also produced by HNO3 ionization. Here
we monitor the most abundant fragment, NO2+ (m/z ) 46), and
employ pre- and postchopper TOF analysis to distinguish N2O5
and HNO3 by their short (<10-6 s) and long (∼0.1 to 10 s)
residence times, respectively, in the acid. The pre- and
postchopper wheels are shown in Figure 2. With the postchopper
wheel in place and the prechopper wheel removed, the incident
beam strikes the liquid continuously and the exiting molecules
are chopped into 36 µs pulses upon exiting the acid. The arrival
times of these molecules at the mass spectrometer in the
postchopper spectrum therefore depend only on the velocities
of the exiting molecules. In contrast, the prechopper wheel slices
the incident beam into 50 µs pulses, and the total arrival time
at the mass spectrometer in the prechopper spectrum depends
on the residence times of the molecules in the acid solution as
well as their gas-phase flight times. As described below, the
HNO3 signal can be obtained by subtracting the prechopper
spectrum (containing only N2O5) from the postchopper spectrum
(containing both N2O5 and HNO3).
Figure 3. Pre- and postchopper TOF spectra of N2O5 and their
difference spectra following collisions of 150 kJ mol-1 N2O5 with
72 wt % H2SO4 containing (a) no hexanol and (b) 40 mM hexanol.
The difference spectra correspond to thermal desorption of HNO3
molecules generated by N2O5 hydrolysis. N2O5 inelastic scattering (IS)
is fit by a dotted curve. The difference spectra and the thermal
desorption (TD) component of the N2O5 prechopper spectra are fit by
Maxwell-Boltzmann distributions shown with solid curves.
from a ratio of pre- and postchopper signals, as shown later in
eq 2. The reliability of this start-stop procedure has been tested
in three ways: (1) high-energy argon scattering from the stopped
The raw TOF signals are converted into relative N2O5 and
HNO3 signals by correcting for the different ionization cross
+
sections σNO for dissociative ionization of N2O5 and HNO3
2
into NO2+. These cross-sections were calculated by multiplying
wheel (texp ) 6 s) and continuously moving wheels (texp
)
0.27 s) are identical, implying that the packing of the butyl and
hexyl films remains the same on the stationary and moving
films,38,39 (2) measurements of the HBr f DBr and HCl f
DCl exchange fractions (ranging from 0.08 to 0.60) in D2SO4
using the stopped wheel differ by less than (0.02 from
measurements using a continuously moving wheel, and (3) the
N2O5 hydrolysis probabilities in bare 72 wt % H2SO4 when
measured from the stopped wheel and from the moving wheel
are each 0.15 ( 0.02 after correcting for the fraction of HNO3
remaining in the acid, as described in the Appendix.
Hexanol Solubility Measurements. The solubility of hexanol
in 72 wt % H2SO4 was measured by comparing hexanol mass
spectrometer signals from pure hexanol (of known vapor
pressure51) and from 0.02 to 0.1 M hexanol in the acid at
253, 243, and 233 K. The vapor pressures varied linearly with
hexanol concentration, and when extrapolated to 216 K, yield
a hexanol solubility of 3.2 ( 0.4 × 108 M atm-1 and an enthalpy
of vaporization of 78 ( 2 kJ mol-1. As discussed later, the
solubility of butanol is expected to be slightly lower.52
the total ionization cross section σtot of each parent molecule
+
by the branching ratio r for NO2 formation. We used the
following values to calculate σNO : σtot (N2O5) ) 7.0 Å2 from
+
2
ref 47, r (N2O5 f NO2+) ) 0.54 from ref 48, σtot (HNO3) )
6.3 Å2 from ref 49, and r (HNO3 f NO2+) ) 0.60 from ref 50.
Each value is measured at an electron energy of 70 eV, which
+
is equal to our experimental setting. Fortuitously, σNO is
2
3.8 Å2 for both N2O5 and for HNO3. The relative intensities of
HNO3 and N2O5 are therefore equal to the relative fluxes
obtained from the TOF spectra to within an estimated uncer-
tainty of (15% in the ionization parameters. This systematic
uncertainty is not included in the error bars reported for the
+
hydrolysis probabilities because σNO does not change the ratio
2
of probabilities for the bare and alcohol-coated acids.
Monitoring HNO3 Desorption. The mass spectrometer is
collimated to monitor a patch of the acid-coated wheel for a
time equal to the exposure time, texp, of 0.27 s. This observation
time is insufficient to observe the complete evaporation of the
HNO3 product, which desorbs over several seconds because of
its high solubility of 2 × 106 M atm-1 in 72 wt % H2SO4 at
216 K (as described in the Appendix). To monitor most of the
HNO3 desorption, the wheel is periodically stopped for 6 s and
the HNO3 desorption signal is recorded. The liquid is viscous
enough (1900 cP at 216 K) that it sags only slightly during this
stop period; a droop in the acid film does not distort the
measurements because the hydrolysis probability is determined
Results and Analysis
Hydrolysis of N2O5 by Bare Sulfuric Acid. Figure 3a
compares pre- and postchopper TOF spectra recorded at m/z )
46 (NO2+) following collisions of 150 kJ mol-1 N2O5 with
uncoated 72 wt % H2SO4 at 216 K. In the postchopper mode,
the arrival time distribution depends only on the velocities of

2924 J. Phys. Chem. A, Vol. 111, No. 15, 2007
Park et al.
TABLE 1: N2O5 Hydrolysis Probabilities and HCl and HBr HX f DX Exchange Fractions Measured for 72 wt % H2SO4 or 70
wt % D2SO4 at 216 K
N2O5 frxn
N2O5 hydrolysis
HCl f DCl
HBr f DBr
acid solution
(eq 2)a
probability γb
exchange fractionc
exchange fractionc
bare acid
180 mM BuOH
40 mM HexOH
0.12 ( 0.02
0.08 ( 0.02
0.05 ( 0.01
0.15 ( 0.02
0.10 ( 0.02
0.06 ( 0.01
0.08 ( 0.03
0.20 ( 0.02
0.16 ( 0.02
0.19 ( 0.02
0.63 ( 0.02
0.60 ( 0.02
a Measured in both 70 wt % D2SO4 and 72 wt % H2SO4, each 0.32 mole fraction acid. b The hydrolysis probability γ is obtained from frxn after
using eq A.1 to correct for HNO3 molecules accumulating in solution and assuming no production of alkyl nitrates. c Measured in D2SO4 by
H f D exchange. These values are equated with the entry probability of HX, either as molecular HX or as X- and H+/D+.38
the exiting N2O5 and HNO3 molecules. Therefore, both species
contribute to the TOF spectra measured at m/z ) 46, regardless
of their residence times in the acid. In the prechopper mode,
the TOF spectrum is a convolution of the gas-phase velocities
of the molecules and their residences times in the acid. The
average residence time of HNO3 is so long (∼0.2 s), and the
residence times of the individual molecules are so broadly
distributed, that the desorption times of the exiting HNO3
molecules are not correlated with the incident pulses. In this
case, the HNO3 desorption signal merges with the background.
This was shown previously in Figure 3 of ref 53 for collisions
of HNO3 with 70 wt % D2SO4 at 213 K. As described later, the
residence time of N2O5 lies in the opposite limit; it is shorter
than the minimum 10-6 s residence time that visibly alters the
prechopper TOF spectrum.38,53 Therefore, the prechopper spectra
measured at m/z ) 46 in Figure 3a is composed only of
unreacted N2O5 molecules and is equivalent to the postchopper
spectra of N2O5. By subtracting the prechopper spectrum
containing only N2O5 signal from the postchopper spectrum with
contributions from both N2O5 and HNO3 signals, we obtain the
postchopper TOF spectrum of HNO3.
The bimodal feature of the prechopper spectrum in
Figure 3a reveals that N2O5 molecules with high incidence
energy follow two distinct pathways upon colliding with the
surface, as depicted in Figure 1. The narrow, inelastic scattering
(IS) peak at early arrival times (high exit velocities) corresponds
to molecules that recoil directly from the surface, transferring
a portion of their kinetic energy through one or a few bounces.
The broader peak at later arrival times (lower exit velocities)
arises from N2O5 molecules that become thermalized through
dissipation of their excess energy and then thermally desorb
(TD). This prechopper TD component is fit well by a Maxwell-
Boltzmann (MB) distribution at the acid temperature of 216 K
with a mass of N2O5 (108 amu) and the difference spectrum is
fit well by an MB curve with a mass of HNO3 (63 amu),
supporting our experimental approach to identifying them.
The difference spectrum corresponds to thermalized N2O5
molecules that undergo hydrolysis and then desorb as HNO3,
while the prechopper TD signal represents the remaining
N2O5 molecules that thermally desorb before converting to
HNO3. Therefore, the fraction frxn of the thermalized N2O5
molecules that are converted into HNO3 is calculated from
component. Finally, frxn must be corrected for those HNO3
molecules that accumulate within the acid and do not desorb
over the 6 s observation time. This correction, described in the
Appendix, shows that 18% of the HNO3 remain behind and
diff
diff
therefore that I must be replaced by I /0.82. The corrected
TD
TD
hydrolysis probability γ is then found to be 0.15 ( 0.02, as
summarized in Table 1.
Hydrolysis of N2O5 by Hexyl-Coated Sulfuric Acid. The
effects of doping the 72 wt % H2SO4 solution with 40 mM
hexanol are displayed in Figure 3b. Surface tension measure-
ments indicate that, at this hexanol bulk concentration, the hexyl
surface concentration reaches its saturation value of 3.0 × 1014
cm-2, corresponding to 60% of a compact monolayer.38,54 Upon
addition of hexanol, the N2O5 direct scattering channel decreases
and the total (N2O5 + HNO3) desorption channel increases.
These trends have been observed before in hyperthermal
collisions of Ar, HCl, and HBr with sulfuric acid containing
hexanol or butanol; they occur because the alkyl groups roughen
the surface and reduce its effective mass, increasing the
likelihood that N2O5 will thermalize upon collision.38,39
The most important change in Figure 3b is seen in the value
of frxn, which drops from 0.12 ( 0.02 to 0.05 ( 0.01 upon
addition of hexanol. This reduction may be caused by at least
three different factors: (1) a reduced hydrolysis efficiency
imposed by the surface hexyl film, (2) a reduction in the HNO3
desorption flux because the hexyl film increases the HNO3
residence time in the acid, and (3) reaction of N2O5 or HNO3
with the hexyl species to produce hexyl nitrate. We examined
the effect of added hexanol on the average HNO3 residence time
by measuring HNO3 desorption at two different observation
times of 0.27 and 6 s. These times were selected by continuously
spinning the acid-coated wheel at 0.17 Hz and by stopping the
wheel, respectively. As shown in Figure 4a and b, fewer HNO3
molecules desorb over the 0.27 s observation time than over
the 6 s period for both bare and hexyl-coated acids. The ratios
of 0.60 (bare acid) and 0.63 (hexyl-coated acid) are identical to
within our uncertainty, indicating that the hexyl film does not
measurably impede evaporation of HNO3.
Hexyl species at the surface or in the acid may also react
directly with N2O5 or with HNO3 generated by hydrolysis to
produce hexyl nitrate, CH3(CH2)5ONO2, according to the
reactions ROH + N2O5 f RONO2 + HNO3 and ROH +
HNO3 f RONO2 + H2O. Hexyl nitrate should, like hexanol
itself, evaporate slowly from solution, with dominant cracks in
the mass spectrometer at m/z ) 46, co-incident with NO2+, at
m/z ) 43, which does not overlap with N2O5 or HNO3 but does
overlap with hexanol, and at m/z ) 76, which is unique to hexyl
nitrate. Repeated searches at each m/z value showed no evidence
of a signal that could be assigned to the alkyl nitrate. We
conservatively estimate that the uncertainty in these searches
is ∼5% of the total HNO3 signal. The ratio of ionization cross
sections at NO2+ for hexyl nitrate and nitric acid is expected to
1
/2IHTDNO
1/2IdTiDff
3
frxn
)
≈
(2)
N2O5
TD
I
+ 1/2IHTDNO3 ITprDe + 1/2ITdiDff
where ITprDe(IN2O5) and ITdiDff(IHNO3) are the relative fluxes of
TD
TD
thermally desorbing molecules of the prechopper (N2O5) and
the difference (HNO3) TOF spectra, respectively. The factor of
1/2 in eq 2 arises from the stoichiometry of eq 1. The measured
exchange fraction from Figure 3a is 0.12 ( 0.02. The ( 0.02
error bar reflects one standard deviation in reproducibility of
(0.01 and an estimated (0.01 uncertainty in fitting the TD
1
be ∼ /3, based on the cracking patterns of each molecule50,55

Inhibition of N2O5 Hydrolysis in Sulfuric Acid
J. Phys. Chem. A, Vol. 111, No. 15, 2007 2925
Figure 4. HNO3 difference spectra obtained at two beam exposure
times, texp ) 0.27 s (]) and 6 s (0), following collisions of 150 kJ
mol-1 N2O5 with 72 wt % H2SO4 containing (a) no hexanol and (b) 40
mM hexanol. In each panel, the TOF signals are normalized to the
signal at texp ) 6 s.
Figure 5. Pre- and postchopper TOF spectra of N2O5 and the difference
spectra following collisions of N2O5 at a low incidence energy of 13
kJ mol-1 with 72 wt % H2SO4 containing (a) no hexanol and (b) 40
mM hexanol.
and assuming that the total ionization cross sections scale with
the number of electrons in each molecule.49 The relative TOF
signals N may then be converted to relative fluxes I ) N V by
multiplying by the ratio of their average velocities, which is
Vhexyl nitrate / Vnitric acid ) 0.7. These numbers imply that the
fraction of hexyl nitrate produced should not exceed 3 × 0.7 ×
0.05 ≈ 0.1 of the HNO3 flux. Thus, the formation of hexyl
nitrate may be as high as 10% of the formation of HNO3,
increasing the total loss of N2O5 into the hexyl-coated acid from
0.05 to 0.055, a difference that lies within the precision of the
measurement. This adjustment is not incorporated into the
analysis below, but it remains a source of uncertainty in our
experiments. In the absence of this nitration channel, we
conclude that the reduction in HNO3 desorption imposed by
the hexyl film is due to a reduction in the hydrolysis of N2O5.
the addition of 40 mM hexanol to 72 wt % H2SO4 decreases γ
by more than half.
Figure 6a displays the dependence of frxn on the bulk-phase
concentration of hexanol in 72 wt % H2SO4 at 216 K. The open
circles represent the surface concentration of hexyl species
calculated from surface tension measurements for 72 wt % H2-
SO4 at 295 K.54 This coverage follows a Langmuir-type
adsorption curve, rising rapidly at low-bulk hexanol concentra-
tions and then plateauing near 20 mM. The production of HNO3
follows an opposite trend: frxn decreases sharply and then also
plateaus. The correlation between frxn and hexyl surface coverage
implies that the reduction in hydrolysis with the addition of
hexanol is caused by the presence of hexyl species that segregate
at the surface rather than by hexyl species dissolved in the bulk.
Hydrolysis of N2O5 by Butyl-Coated Sulfuric Acid. We
also investigated changes in N2O5 hydrolysis upon adding
1-butanol, a soluble surfactant that is two CH2 groups shorter
than 1-hexanol. Figure 6b plots frxn against the bulk concentra-
tion of butanol following collisions of 160 kJ mol-1 N2O5 with
72 wt % H2SO4 at 216 K. The value of frxn drops from 0.12 (
0.02 for bare acid to 0.08 ( 0.02 for 180 mM butanol (44% of
a compact monolayer) and then does not change significantly
with further increasing butanol bulk concentration. As shown
in Table 1, the hydrolysis probabilities γ corrected for residual
HNO3 in the acid are 0.15 (bare acid) and 0.10 (180 mM
butanol). Just as with hexanol in panel a, the trend in frxn is
opposite to that of the surface segregation of butyl species (open
circles).
The TOF spectra in Figures 3 and 4 were generated by
collisions of N2O5 at an incident energy of 150 kJ mol-1 (84
RTliq), chosen because this high energy beam has a high flux
(estimated to be 6 × 1014 cm-2 s-1 or approximately one
monolayer s-1) and because the IS and TD components are well
separated at this energy. To determine if this high collision
energy skews the value of frxn, experiments were performed at
the much lower incident energy of 13 kJ mol-1 (7 RTliq), where
the N2O5 flux is reduced by half. As shown in Figure 5, nearly
all impinging N2O5 molecules thermalize on the surface, and
no IS channel can be discerned in the pre- or postchopper
spectra. The measured frxn value of 0.05 ( 0.02 matches the
value of 0.05 ( 0.01 found at high incident energy in Figure 3.
The invariance of frxn with incidence energy implies that frxn
may be considered to be a measurement of the hydrolysis
probability γ that is conventionally reported for thermal-energy
collisions at 2RTacid ) 3.6 kJ mol-1 (neglecting any production
of hexyl nitrate). As described in the Appendix, these values
must be corrected for HNO3 that remains behind in solution.
The corrected values are listed in the third column of Table 1,
which are just slightly higher than frxn. The observed reduction
in the hydrolysis probability γ from 0.15 to 0.06 reveals that
The data in panels a and b may be used to plot γfilm/γbare
directly against the surface coverage θfilm in panel c, where θfilm
) nsurf/nmax is the fraction of a compact monolayer covered by
the film and nmax is ∼5 × 1014 cm-2 in the close-packed
configuration.56 The surface concentrations nsurf were obtained
from the surface tension measurements at 295 and 250 K and
linearly extrapolated to 216 K.39,54 Panel c indicates that N2O5

2926 J. Phys. Chem. A, Vol. 111, No. 15, 2007
Park et al.
N2O5 hydrolysis, HCl and HBr entry is significantly enhanced
by the surface films.
Discussion
It is striking that the butyl and hexyl films on 72 wt %
H2SO4 at 216 K impede N2O5 hydrolysis because these same
films enhance HCl and HBr uptake into 70.2 wt % D2SO4, as
shown in Table 1. H/D isotope effects are not responsible for
the different outcomes; the same reduction in N2O5 hydrolysis
by the hexyl film is observed using 70 wt % D2SO4 and 72 wt
% H2SO4, which are each 0.32 mole fraction acid.
Previous studies in our laboratory suggest that the entry of
HCl and HBr into D2SO4/D2O solutions containing hexanol at
213 K is controlled by two competing effects: hexyl chains at
the surface impede gas transport through the film, while the
basic hexanol OD groups assist HX entry by providing extra
surface sites for HCl and HBr dissociation.38 As the underlying
solution is made more acidic, the hexyl and butyl films become
more porous and HCl and HBr molecules can more easily reach
the alcohol OD groups and protonate them. In particular, a
saturated hexyl film on 68 wt % D2SO4, which corresponds to
62% coverage of a compact monolayer, enhances the entry of
HCl from 0.14 (bare acid) to 0.23 and of HBr from 0.29 to
0.63, whereas a saturated hexyl film on 56 wt % acid,
corresponding to 68% coverage, reduces HCl entry from 0.69
to 0.58. N2O5 lacks the acidic proton of HCl and HBr and cannot
follow this pathway. In 70 wt % H2SO4 at cold temperatures,
N2O5 hydrolysis is postulated to proceed primarily through an
acid-catalyzed channel consisting of N2O5 + H+ f HNO3 +
+
+
NO2 and NO2 + H2O f HNO3 + H+.17,57 The surface
+
alcohol OH and OH2 groups could potentially substitute for
H2O and H+ as reactants and convert N2O5 into HNO3 and alkyl
nitrate. The reduction in N2O5 hydrolysis we observe and the
absence of a signal attributable to the alkyl nitrate indicate that
these reactions do not readily occur and that the surface alcohol
molecules instead impede the conversion of N2O5 to HNO3.
This reduction in N2O5 hydrolysis may be framed within two
scenarios motivated by the steady decrease in the fractional
hydrolysis rate γhexyl/γbare with hexanol surface coverage θhexyl
shown in Figure 6c: (1) the hexyl chains block entry of N2O5
into the acid, and (2) the hexyl head groups reduce the hydrolysis
rate in the near-interfacial region. The steady decrease in Figure
6c is described by the relation γhexyl/γbare ≈ 1 - (1 ( 0.2)θhexyl
over the range of θhexyl ) 0 to 0.6, where θhexyl is the fraction
of the surface area covered by hexyl chains in the all-trans
configuration (∼20 Å2 in area and ∼10 Å long).56,58 Clifford
et al. have recently observed, for 1-octanol films on water, that
the interfacial hydrolysis of HNO3 and NH3 also decreases
linearly with increasing film coverage.30
Figure 6. Fraction frxn of N2O5 converted to HNO3 vs (a) hexanol and
(b) butanol bulk concentration in 72 wt % H2SO4 at 216 K. The hexyl
and butyl surface concentrations at 295 K are plotted against the right-
hand axis as open circles. The error bars represent (1σ for 5
measurements (bare acid), 3 (40 mM hexanol), and 3 (180 mM butanol).
The other points represent single measurements. (c) The ratio of
hydrolysis probabilities γfilm/γbare vs fractional surface coverage θfilm
from the data in panels a and b and the Appendix. The θfilm values
were obtained from surface tension data at 295 and 250 K and
extrapolated to 216 K.54 The dashed line is a fit to the hexanol data.
hydrolysis decreases approximately linearly with hexanol surface
coverage, such that γhexyl/γbare ) 1 - mθhexyl with m ) 1.0 (
0.2 (dashed line). The single cluster of butanol values, combined
with the uncertainty in determining θfilm, makes it difficult to
determine if the shorter-chain alcohol impedes hydrolysis less
effectively at constant surface coverage.
Collisions of HCl and HBr with Hexyl-and Butyl-Coated
Sulfuric Acid. To gauge the effects of hexyl and butyl films
on HCl and HBr uptake into the acid, we also measured the
extent of HX f DX exchange following collisions of 100 kJ
mol-1 HCl and 150 kJ mol-1 HBr from 70 wt % D2SO4 at 216
K containing no surfactant, 40 mM hexanol, and 180 mM
butanol. As discussed in refs 37, 38, and 53, the fraction fexch
of thermalized HX molecules that undergo D f H exchange
can be interpreted as the entry probability of HCl and HBr into
the acid, either as molecular HX or as X- and H+ following
dissociation at the surface. Table 1 shows that, in contrast to
Within the first scenario, the passage of N2O5 into the acid
is limited to motions though fluctuating gaps between the hexyl
chains, which become less prevalent as more hexyl species
segregate to the surface at higher bulk-phase concentration and
become more tightly packed. These gaps arise because, at
intermediate coverages, the hexyl chains should adopt a range
of configurations with varying spacings and relative positions,
as pictured in simulations of butanol and heptanol on water35,36,39
and inferred from neutron reflection studies of hexanol on
water58 and sum frequency generation studies of hexanol on
59.5 wt % H2SO4.59 The approximate scaling between γhexyl
/
γbare and 1 - θhexyl implies a constant “blocking power” for
each added hexyl chain. This simple correlation may be
accidental, however, because the various hexyl chain conforma-
tions project different surface areas and span different lengths,

Inhibition of N2O5 Hydrolysis in Sulfuric Acid
J. Phys. Chem. A, Vol. 111, No. 15, 2007 2927
and the range of these structures becomes more restricted as
the chain density increases. In particular, this scaling will likely
fail as θ approaches 1 and the chains become highly aligned
and compact; for insoluble long-chain surfactants, gas perme-
ation is observed to decrease exponentially with surface
concentration at high coverage.26
which a reaction occurs is equal to (D/k)1/2, where D is the N2O5
diffusion coefficient and k is the hydrolysis rate constant.
Reference 17 predicts that this depth is approximately 1 Å for
72 wt % H2SO4 at 216 K. In accord with this estimate, the
thermal-desorption component of the prechopper TOF spectrum
in Figures 4, 5, and 6 can be fit well with a Maxwell-Boltzmann
distribution unconvoluted with a residence time distribution,
indicating that the average bulk-phase residence time τ of intact
N2O5 is less than 2 × 10-6 s.53 In this case, the average diffusion
The model above may be analyzed quantitatively in the limit
that hydrolysis occurs far from the surfactant layer, which acts
only to impede the entry of N2O5 into the acid. The hydrolysis
probability may then be expressed as independent resistances
using 1/γhexyl ) 1/Rhexyl + 1/Γreact, where Rhexyl is the probability
that N2O5 passes through the hexyl film and enters the acid
depth of 0.6(Dτ)1/2 is less than 10 Å for D e 10-8 cm2 s-1 42,64
.
This experimental limit implies that hydrolysis often occurs in
a region containing alcohol head groups and perhaps CH2 groups
as well.
and Γreact is the bulk-phase reactivity.34,60,61
Γreact does not vary
The data in Figure 6 therefore suggest two possible causes
for the suppression of N2O5 hydrolysis: the surface alkyl chains
may pack tightly enough to block the entry of some N2O5
molecules into the acid, while the alcohol head groups may
reduce the reactivity of N2O5 that do enter the acid, perhaps by
reducing the concentration of both H3O+ and H2O near the
film-acid interfacial region.
upon addition of hexanol because hydrolysis occurs only within
the bulk acid, where little hexanol is present. The range of values
of Γreact are obtained from γbare ) 0.15 in Table 1 using the
maximum value of Rbare ) 1 and its minimum value of 0.15,
yielding Γreact(Rbare ) 1) ) 0.18 and Γreact(Rbare ) 0.15) ) ∞
(every entering molecule reacts). In the latter case when Γreact
) ∞, Rhexyl is equal to γhexyl and decreases linearly with θhexyl
.
When Γreact is 0.18, Rhexyl starts at 1 for the bare surface and
drops more sharply than γhexyl. At the highest surface coverage
of θhexyl ) 0.60, the resistance equation predicts that Rhexyl varies
between 0.09 (Γreact ) 0.18) and 0.06 (Γreact ) ∞). This analysis
suggests that, if hydrolysis occurs far from the film region, then
the saturated hexyl film must stop 90% or more of the
thermalized N2O5 molecules from entering the acid.
Conclusions and Atmospheric Implications
Hexanol and butanol dissolved in 72 wt % H2SO4 at 216 K
form loosely packed surface films corresponding to ∼60% and
∼44% of a compact monolayer at their asymptotic coverages,
respectively. The conversion of N2O5 to HNO3 is noticeably
impeded by these surface films, with reaction probabilities that
drop from 0.15 to 0.06 for the hexyl film and to 0.10 for the
butyl film. These changes are smaller than those observed on
seawater using hexanoic acid (3- to 4-fold reduction),31 most
likely because the alcohol films do not pack as tightly on acidic
The contrasting behaviors of N2O5 and HCl or HBr may be
ascribed in part to their different sizes; the ∼70 Å3 molecular
volume of N2O5 is approximately twice that of HCl and HBr,
making it more difficult for N2O5 to pass through gaps between
the alkyl chains. Thus, HCl and HBr may permeate more easily
through the film and reach the underlying acid, where they can
bond to and protonate the alcohol OH groups. Size arguments
alone cannot justify a lower mobility of N2O5 through the film,
however, because the molecular volumes of CF3CH2OH and
N2O5 are very similar, but CF3CH2OH passes through the hexyl
and butyl films on nearly every collision.38 This difference points
to a potentially critical role for solute OH groups in aiding
transport within the porous film, perhaps via hydrogen bonds
with H2O, H3O+, or alcohol OH groups that straddle the film.62
+
subphases due to charge repulsion among the ROH2 head
- 40,41
groups and formation of ROSO3
.
The experiments demonstrate that short-chain alcohols can
significantly impede N2O5 hydrolysis by sulfuric acid droplets
if the alcohols are plentiful enough to form saturated surface
films. However, it is unlikely that butanol and hexanol are
present in the background upper troposphere or lower strato-
sphere to reach significant surface coverages. The two shortest
alcohols, methanol and ethanol, have been identified at con-
centrations of ∼1000 ppt and ∼50 ppt (parts per trillion by
volume), respectively, in polluted regions of the upper pacific
troposphere at 10-12 km.23,65 Our measured solubility of
hexanol in 72 wt % H2SO4 at 216 K of 3 × 108 M atm-1 permits
an estimate of the gas-phase concentration required to create a
saturated hexyl film. For a typical pressure of 160 Torr at an
altitude of 12 km, we estimate that the ∼40 mM hexanol bulk
concentration for a saturated hexyl film requires ∼600 ppt of
gas-phase hexanol. The solubility of butanol should lie between
We may also consider the opposite scenario in which N2O5,
like CF3CH2OH, passes easily between the alkyl chains, but
hydrolysis itself is suppressed by the interfacial alcohol species.
Within this second picture, the 1 - θhexyl behavior in Figure 6c
reflects the role of the hexyl head groups, rather than the hexyl
chains, in reducing the rate of reaction. Previous studies indicate
that N2O5 hydrolysis in 70 wt % sulfuric acid is expected to
occur very close to the surface.16,17,63 In this region, the hexyl
-
the hexanol value and that of ethanol of 2 × 108 M atm-1 52
; in
-OH and -OSO3 end groups may locally interfere with
protonation of N2O5 by interfacial H2SO4 or H3O+, which is
potentially the key step that initiates hydrolysis in acidic
solutions.17,57 In the limit in which hydrolysis occurs solely in
the surface region and the hexyl film does not impede transport,
the resistance equation reduces to 1/γhexyl ≈ 1/S + 1/Γsurf (where
S, the trapping probability, approaches 1 at thermal collision
energies, as shown in Figure 5).60 In this case, Γsurf scales
approximately with γhexyl (for γhexyl e 0.15) and thus would be
roughly proportional to 1 - θhexyl, which represents the fraction
of surface area not covered by head groups.
this case, the ∼200 mM butanol concentration for a saturated
butyl film requires ∼4000 ppt of gas phase butanol.
These minimum gas-phase concentrations rise rapidly in more
dilute sulfuric acid because of lower alcohol solubilities. For
60 wt % H2SO4 at 216 K, which is closer to aerosol acidities
near the tropopause region,9 the solubility of the alcohols is
expected to be near ∼4 × 106 M atm-1 52
. This lower solubility
requires the gas-phase hexanol concentration to increase to an
even more improbable value of 50 000 ppt to create a saturated
surface film. Butanol and hexanol alone are therefore not
sufficiently surface active in sulfuric acid in the upper tropo-
sphere or lower stratosphere to impose barriers to N2O5
hydrolysis. These short-chain alcohols are just one of many
different types of organic molecules, soluble and insoluble,21-24
Our experiments directly support the prediction that N2O5
hydrolysis takes place in a shallow layer near the surface,
potentially in the vicinity of the hexyl or butyl species.
According to continuum models,17,60 the average depth over

2928 J. Phys. Chem. A, Vol. 111, No. 15, 2007
Park et al.
that may coalesce within or on aerosol droplets and segregate
to the surface to form mixed monolayers of widely varying
porosity and reactivity.27 We hope to extend the alcohol
measurements here to soluble organic molecules with different
functional groups to determine their surface activity in sulfuric
acid and their ability to impede or enhance gas-liquid transport
and interfacial reactions.
HNO3 desorption was also measured for a much shorter
observation time of 0.27 s using a continuously rotating wheel.
In this case, fremain rises from 0.18 to 0.52. Despite this large
correction, the final value for γ(N2O5) was determined to be
the same in the bare acid, 0.15 ( 0.02, suggesting that the stop-
start technique may not be necessary to determine hydrolysis
probabilities even when the HNO3 product dissolves for long
times.
No corrections are necessary for HCl dissolution in 72 wt %
H2SO4 at 216 K because τ is so short (τ ) 2 × 10-4 and fremain
) 0.03). HBr dissolves for longer times in the bare acid (τ )
0.002 s and fremain ) 0.09), but again drops to 2 × 10-4 s in the
alkyl-coated acids because R ≈ fexch is larger. The value for
HBr entering the bare acid in Table 1 has been corrected for a
10% increase due to HBr molecules accumulating in solution
over the 0.27 s exposure and observation time.
Acknowledgment. We are grateful to the Air Force Office
of Scientific Research for supporting this work, to the Camille
and Henry Dreyfus Foundation for providing a postdoctoral
fellowship in environmental chemistry to S.-C.P., and to the
Vilas Foundation of the University of Wisconsin. We also thank
John Morris for performing early studies of N2O5 on bare
sulfuric acid, James Krier and Casey Harris for surface tension
measurements, and Steven Brown for advice on N2O5 synthesis.
We also thank the reviewers for their valuable comments.
References and Notes
(1) Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the Upper and Lower
Atmosphere; Academic Press: New York, 2000; Chapters 9.C and 12.C.
(2) Solomon, S. ReV. Geophys. 1999, 37, 275.
Appendix: Determining the Fraction of HNO3 Remaining
in Solution
(3) Hanson, D. R.; Lovejoy, E. R. J. Phys. Chem. 1996, 100, 6397.
(4) Jacob, D. Atmos. EnViron. 2000, 34, 2131.
We determine the fraction of N2O5 molecules that are
converted into HNO3 by measuring the relative flux of HNO3
molecules that desorb from solution. Because some of the HNO3
molecules created by hydrolysis remain behind in solution over
(5) Rodriguez, J. M.; Ko, M. K. W.; Sze, N. D. Nature 1991, 352,
134.
(6) Hanson, D. R.; Ravishankara, A. R.; Solomon, S. J. Geophys. Res.
Atmos. 1994, 99, 3615.
the measurement time texp, the measured flux IHNO in eq 2 is
3
(7) Wennberg, P. O.; Cohen, R. C.; Stimpfle, R. M.; Koplow, J. P.;
Anderson, J. G.; Salawitch, R. J.; Fahey, D. W.; Woodbridge, E. L.; Keim,
E. R.; Gao, R. S.; Webster, C. R.; May, R. D.; Toohey, D. W.; Avallone,
L. M.; Proffitt, M. H.; Loewenstein, M.; Podolske, J. R.; Chan, K. R.; Wofsy,
S. C. Science 1994, 266, 398.
(8) Fahey, D. W.; Kawa, S. R.; Woodbridge, E. L.; Tin, P.; Wilson, J.
C.; Jonsson, H. H.; Dye, J. E.; Baumgardner, D.; Borrmann, S.; Toohey,
D. W.; Avallone, L. M.; Proffitt, M. H.; Margitan, J.; Loewenstein, M.;
Podolske, J. R.; Salawitch, R. J.; Wofsy, S. C.; Ko, M. K. W.; Anderson,
D. E.; Schoeberl, M. R.; Chan, K. R. Nature 1993, 363, 509.
(9) Hendricks, J.; Lippert, E.; Petry, H.; Ebel, A. J. Geophys. Res.
Atmos. 1999, 104, 5531.
(10) See, for example, Badger, C. L.; Griffiths, P. T.; George, I.; Abbatt,
J. P. D.; Cox, R. A. J. Phys. Chem. A 2006, 110, 6986 and ref 19.
(11) Dentener, F. J.; Crutzen, P. J. J. Geophys. Res. Atmos. 1993, 98,
7149.
(12) Tie, X.; Emmons, L.; Horowitz, L.; Brasseur, G.; Ridley, B.; Atlas,
E.; Stround, C.; Hess, P.; Klonecki, A.; Madronich, S.; Talbot, R.; Dibb, J.
J. Geophys. Res. 2003, 108, 8364.
TD
too small. This flux can be corrected by estimating the fraction
of HNO3 molecules that remain dissolved in the acid. The
predicted thermal-desorption flux from a fresh liquid continu-
ously exposed for a time texp is Itrue ) Ienter(1 - (erfc[(texp/τ)1/2]‚
TD
etexp/τ),66 where Ienter is the equivalent flux of HNO3 entering
the solution generated by the impinging N2O5 molecules. This
flux depends on the characteristic residence time τ of the HNO3
molecules in solution. It is given by τ ) D(4H*RT/Rth V )2,
where H* is the overall solubility of HNO3 in M atm-1, D is
the HNO3 diffusion constant, Rth is the HNO3 entry probability
averaged over a Boltzmann distribution of collision energies at
temperature T, and V is the thermal velocity of HNO3. Over
the time t ) 0 to τ, the desorption flux rises from 0 to 57% of
the flux entering solution. We use the values H* ) 2 × 106 M
atm-1 for HNO3 in 72 wt % H2SO4 at 216 K from ref 43, D )
7 × 10-9 cm2 s-1 from ref 42, and Rth ) 1 from ref 53 to obtain
τ ) 0.2 s.
(13) Evans, M. J.; Jacob, D. J. Geophys. Res. Lett. 2005, 32, L09813.
(14) Hanson, D. R.; Ravishankara, A. R. J. Geophys. Res. Atmos. 1991,
96, 17307.
(15) Zhang, R. Y.; Leu, M. T.; Keyser, L. F. Geophys. Res. Lett. 1995,
22, 1493.
The fraction fremain of HNO3 molecules that accumulate within
the acid over the time texp is given by eq 8c in ref 67, where it
is also graphed:
(16) Fried, A.; Henry, B. E.; Calvert, J. G.; Mozurkewich, M. J. Geophys.
Res. Atmos. 1994, 99, 3517.
(17) Robinson, G. N.; Worsnop, D. R.; Jayne, J. T.; Kolb, C. E.;
Davidovits, P. J. Geophys. Res. Atmos. 1997, 102, 3583.
(18) Williams, L. R.; Manion, J. A.; Golden, D. M.; Tolbert, M. A. J.
Appl. Meteorol. 1994, 33, 785.
(19) Kane, S. M.; Caloz, F.; Leu, M.-T. J. Phys. Chem. A 2001, 105,
6465.
(20) Murphy, D. M.; Thomson, D. S.; Mahoney, T. M. J. Science 1998,
282, 1664.
(21) Singh, H.; Chen, Y.; Tabazadeh, A.; Fukui, Y.; Bey, I.; Yantosca,
R.; Jacob, D.; Arnold, F.; Wohlfrom, K.; Atlas, E.; Flocke, F.; Blake, D.;
Blake, N.; Heikes, B.; Snow, J.; Talbot, R.; Gregory, G.; Sachse, G.; Vay,
S.; Kondo, Y. J. Geophys. Res. Atmos. 2000, 105, 3795.
(22) Singh, H.; Chen, Y.; Staudt, A.; Jacob, D.; Blake, D.; Heikes, B.;
Snow, J. Nature 2001, 410, 1078.
(23) Apel, E. C.; Hills, A. J.; Lueb, R.; Zindel, S.; Eisele, S. J. Geophys.
Res. 2003, 108, 8794.
(24) Tervahattu, H.; Juhanoja, J.; Vaida, V.; Tuck, A. F.; Niemi, J. V.;
Kupiainen, K.; Kulmala, M.; Vehkamaki, H. J. Geophys. Res. Atmos. 2005,
110, D06207.
texp
true
TD
Ientertexp
-
I
dt
∫
0
fremain(texp) )
Ienter exp
t
) (τ/texp){erfc[(texp/τ)1/2] etexp/τ + 2(texp/πτ)1/2 - 1)}
(A.1)
For texp ) 6 s and τ ) 0.2 s, fremain is 0.18, indicating that 18%
of the HNO3 molecules remain in the acid even over times much
longer than τ. In this case, IdTiDff should be divided by 0.82 in eq
2 to obtain the true relative flux. This correction increases
frxn(bare) from 0.12 to 0.15, frxn(hexyl) from 0.05 to 0.06, and
frxn(butyl) from 0.08 to 0.10 in Table 1, where the corrected
(25) Brown, S. S.; Ryerson, T. B.; Wollny, A. G.; Brock, C. A.; Peltier,
R.; Sullivan, A. P.; Weber, R. J.; Dube, W. P.; Trainer, M.; Meagher, J. F.;
Fehsenfeld, F. C.; Ravishankara, A. R. Science 2006, 311, 67.
(26) Barnes, G. T. Colloids Surf. A. 1997, 126, 149.
(27) Gilman, J. B.; Vaida, V. J. Chem. Phys. A 2006, 110, 7581.
quantities are labeled γ and represent our best estimates of the
hydrolysis probabilities. The ratios of frxn before and after
correction remain essentially unchanged, such that γfilm/γbare
rxn(film)/frxn(bare).
≈
f

Inhibition of N2O5 Hydrolysis in Sulfuric Acid
J. Phys. Chem. A, Vol. 111, No. 15, 2007 2929
(28) For a review of gas transport through surfactant films, see the
introduction of ref 39.
(29) Mmereki, B. T.; Chaudhuri, S. R.; Donaldson, D. J. J. Phys. Chem.
A 2003, 107, 2264.
(52) Michelson, R. R.; Staton, S. J. R.; Iraci, L. T. J. Chem. Phys. A
2006, 110, 6711.
(53) Morris, J. M.; Behr, P.; Antman, M. D.; Ringeisen, B. R.; Splan,
J.; Nathanson, G. M. J. Phys. Chem. A 2000, 104, 6738.
(30) Clifford, D.; Bartels-Rausch, T.; Donaldson, D. J. Phys. Chem.
Chem. Phys. 2007, 9, 1362.
(31) Thornton, J. A.; Abbatt, J. P. D. J. Phys. Chem. A 2005, 109, 10004.
(32) McNeill, V. F.; Patterson, J.; Wolfe, G. M.; Thornton, J. A. Atmos.
Chem. Phys. 2006, 6, 1635.
(33) Folkers, M.; Mentel, T. F.; Wahner, A. Geophys. Res. Lett. 2003,
30, 1644.
(34) Anttila, T.; Kiendler-Scharr, A.; Tillman, R.; Mentel, T. F. J. Phys.
Chem. A 2006, 110, 10435.
(54) Krier, J. M.; Nathanson, G. M., to be submitted for publication.
(55) Fraser, R. T. M.; Paul, N. C. J. Chem. Soc. B 1968, 659.
(56) Tikhonov, A. M.; Pingali, S. V.; Schlossman, M. L. J. Chem. Phys.
2004, 120, 11822.
(57) Hallquist, M.; Stewart, D. J.; Baker, J.; Cox, R. A. J. Phys. Chem.
A 2000, 104, 3984.
(58) Li, Z. X.; Lu, J. R.; Thomas, R. K.; Rennie, A. R.; Penfold, J. J.
Chem. Soc., Faraday Trans. 1996, 92, 565.
(59) Van Loon, L. L; Minor, R. N; Allen, H. C. J. Phys. Chem. A 2007,
submitted for publication.
(35) Chen, B.; Siepmann, J. I.; Klein, M. L. J. Am. Chem. Soc. 2002,
124, 12232.
(60) Hanson, D. R. J. Phys. Chem. B 1997, 101, 4998.
(36) Daiguji, H. J. Chem. Phys. 2001, 115, 1538.
(37) Lawrence, J. R.; Glass, S. V.; Park, S.-C.; Nathanson, G. M. J.
Phys. Chem. A 2005, 109, 7458.
(38) Glass, S. V.; Park, S.-C.; Nathanson, G. M. J. Phys. Chem. A 2006,
110, 7593.
(39) Lawrence, J. R.; Glass, S. V.; Nathanson, G. M. J. Phys. Chem. A
2005, 109, 7449.
(40) Torn, R. D.; Nathanson, G. M. J. Phys. Chem. B 2002, 106, 8064.
(41) Williams, G.; Clark, D. J. J. Chem. Soc. 1956, 1304.
(42) Williams, L. R.; Long, F. S. J. Phys. Chem. 1995, 99, 3748.
(43) Carslaw, K. S.; Clegg, S. L.; Brimblecombe, P. J. Phys. Chem.
1995, 99, 11557. Carslaw, K. S.; Peter, T.; Clegg, S. L. ReV. Geophys.
1997, 35, 125. Calculations were performed at mae.ucdavis.edu/wexler/
aim.htm.
(44) Davidson, J. A.; Viggiano, A. A.; Howard, C. J.; Dotan, I.;
Fehsenfeld, F. C.; Albritton, D. L.; Ferguson, E. E. J. Chem. Phys. 1978,
68, 2085.
(45) Lowry, T. M.; Lemon, J. T. J. Chem. Soc. 1935, 692.
(46) Handbook of Chemistry and Physics; Linde, D. R., Ed.; CRC
Press: New York, 2001; Section 6, p 69.
(47) Antony, B. K.; Joshipura, K. N.; Mason, N. J. Int. J. Mass Spectrom.
2004, 233, 207.
(61) See ref 31 for a similar resistor-model analysis of N2O5 hydrolysis
in sea salt aerosol coated with hexanoic acid.
(62) A more polarizable solute might be expected to be more soluble in
the alkyl film and therefore move through it more easily. The polarizabilities
of N2O5 and CF3CH2OH, however, are estimated to be 8 and 5 Å3,
respectively, in the opposite order of their apparent mobility. This
observation led us to focus on the OH group rather than the CF3CH2 group
as a facilitator of the permeation of CF3CH2OH. Our measurements also
indicate that H2O moves without resistance through the hexyl and butyl
films. For polarizability estimates, see Miller, K. A.; Savchik, J. A. J. Am.
Chem. Soc. 1979, 101, 7206, and Wincel, H; Mereand, E; Castleman, A.
W. J. Phys. Chem. 1995, 99, 1792.
(63) Hanson, D. R.; Lovejoy, E. R. Geophys. Res. Lett. 1994, 21, 2401.
(64) The average diffusion depth over a time τ/2 is estimated by
calculating z(t) ) [∫c(z,t) dz]/[∫c(z,t) dz] ) 0.80(Dt)1/2, where c(z,t) is
the concentration profile of the gas in the liquid for initial exposure at t )
0. This profile is c(z,t)/c(t f ∞) ) erfc(z/(4Dt)1/2) - exp(z/(Dτ))1/2 exp(t/
τ)erfc((t/τ)1/2 + z/(4Dt)1/2). For a total residence time of τ and diffusion in
one direction of time t ) τ/2, z(τ/2) ≈ 0.6(Dτ)1/2. This profile is obtained
from the diffusion equation and the boundary condition -D∂c(z,t)/∂z|z)0
) (D/τ)1/2(c(t f ∞) - c(z, t)). See also the Appendix of ref 66 and page 72
of Carslaw, H. S; Yaeger, J. C. Conduction of Heat in Solids; Oxford
University Press; Oxford, 1959.
(48) O’Connor, C. S. S.; Jones, N. C.; O’Neale, K.; Price, S. D. Int. J.
Mass Spectrom. 1996, 154, 203.
(49) Jimenez, J. L.; Jayne, J. T.; Shi, Q.; Kolb, C. E.; Worsnop, D. R.;
Yourshaw, I.; Seinfeld, J. H.; Flagan, R. C.; Zhang, X.; Smith, K. A.; Morris,
J. W.; Davidovits, P. J. Geophys. Res. 2003, 108, 13.
(50) O’Connor, C. S. S.; Jones, N. C.; Price, S. D. Int. J. Mass Spectrom.
1997, 163, 131.
(65) Singh, H. B.; Salas, L. J.; Chatfield, R. B.; Czech, E.; Fried, A.;
Walega, J.; Evans, M. J.; Field, B. D.; Jacob, D. J.; Blake, D.; Heikes, B.;
Talbot, R.; Sachse, G.; Crawford, J. H.; Avery, M. A.; Sandholm, S.;
Fuelberg, H. J. Geophys. Res. 2004, 109, D15S07.
(66) Hanson, D. R.; Ravishankara, A. R. J. Phys. Chem. 1993, 97, 12309.
(67) Robinson, G. N.; Worsnop, D. R.; Jayne, J. T.; Kolb, C. E.; Swartz,
E.; Davidovits, P. J. Geophys. Res. Atmos. 1998, 103, 25371.
(51) Schmeling, T.; Strey, R. Phys. Chem. Chem. Phys. 1983, 87, 871.