Effect of viscosity on photodegradation rates in complex secondary organic aerosol materials

Article abstract of DOI:10.1039/c5cp05226b

This work explores the effect of environmental conditions on the photodegradation rates of atmospherically relevant, photolabile, organic molecules embedded in a film of secondary organic material (SOM). Three types of SOM were studied: α-pinene/O₃

Full text of DOI:10.1039/c5cp05226b

Volume 18 Number 13 7 April 2016 Pages 8755–9306
PCCP
Physical Chemistry Chemical Physics
www.rsc.org/pccp
ISSN 1463-9076
PAPER
Sergey A. Nizkorodov et al.
Effect of viscosity on photodegradation rates in complex secondary
organic aerosol materials

PCCP
View Article Online
View Journal | View Issue
PAPER
Effect of viscosity on photodegradation rates in
complex secondary organic aerosol materials†
Cite this: Phys.Chem.Chem.Phys.,
2016, 18, 8785
a
a
a
b
Mallory L. Hinks, Monica V. Brady, Hanna Lignell, Mijung Song,
b
b
c
c
James W. Grayson, Allan K. Bertram, Peng Lin, Alexander Laskin,
Julia Laskin and Sergey A. Nizkorodov*
d
a
This work explores the effect of environmental conditions on the photodegradation rates of atmospherically
relevant, photolabile, organic molecules embedded in a film of secondary organic material (SOM). Three
types of SOM were studied: a-pinene/O
obtained by exposure of LSOM to ammonia (brown LSOM). PSOM and LSOM were impregnated with
,4-dinitrophenol (2,4-DNP), an atmospherically relevant molecule that photodegrades faster than either
3 3 3
SOM (PSOM), limonene/O SOM (LSOM), and aged limonene/O
2
PSOM or LSOM alone, to serve as a probe of SOM matrix effects on photochemistry. Brown LSOM contains
an unidentified chromophore that absorbs strongly at 510 nm and photobleaches upon irradiation. This
chromophore served as a probe molecule for the brown LSOM experiments. In all experiments, either the
temperature or relative humidity (RH) surrounding the SOM films was varied. The extent of photochemical
reaction in the samples was monitored using UV-vis absorption spectroscopy. For all three model systems
examined, the observed photodegradation rates were slower at lower temperatures and lower RH, conditions
that make SOM more viscous. Additionally, the activation energies for photodegradation of each system were
positively correlated with the viscosity of the SOM matrix as measured in poke-flow experiments. These
Received 1st September 2015,
Accepted 7th December 2015
À1
activation energies were calculated to be 50, 24, and 17 kJ mol for 2,4-DNP in PSOM, 2,4-DNP in LSOM,
and the chromophore in brown LSOM, respectively, and PSOM was found to be the most viscous of the
three. These results suggest that the increased viscosity is hindering the motion of the molecules in SOM and
is slowing down their respective photochemical reactions.
DOI: 10.1039/c5cp05226b
www.rsc.org/pccp
imposed by the increased viscosity of SOA have already been
shown to impact the particle growth rates and mechanisms as
Recent investigations into the phase state of secondary organic well as the oxidation rates of organics within the SOA.
Introduction
2
,6,9,12,13
aerosol (SOA) have suggested that under certain conditions,
both atmospheric and laboratory-generated SOA particles may humidity (RH) are expected to have a significant effect on the
Environmental conditions such as temperature and relative
1–6
1–3,11,13,16,21–24
behave as semi-solids or amorphous solids rather than liquids.
viscosity of SOA particles.
As either temperature
This realization has important implications for our understanding or RH of the system is increased, the secondary organic
of processes occurring within SOA material. The rheological material (SOM) from which the particle is made will decrease
properties of organic aerosols, such as viscosity, are expected to in viscosity. For example, Renbaum–Wolff et al. reported that
6
,7
play a role in the dynamics of particle growth, gas–particle the viscosity of a-pinene/O SOM increases by several orders of
3
8
–10
3
partitioning,
compounds,
reactive uptake on particle surfaces.
diffusion kinetics of water, oxidants, and other magnitude upon reducing the RH from 90% to 30%. The effect
the dynamics of particle aggregation, and of temperature on the viscosity of SOM has not yet been
1
1–15
16
1
7–20
Diffusion limitations studied, but we estimate that a decrease in temperature from
0 1C to 0 1C will decrease the viscosity of a-pinene/O SOM by
roughly two orders of magnitude, as discussed in our previous
2
3
a
Department of Chemistry, University of California, Irvine, CA 92697, USA.
E-mail: nizkorod@uci.edu
2
5
work.
b
c
Department of Chemistry, University of British Columbia, Vancouver, BC, Canada
Environmental Molecular Sciences Laboratory, Pacific Northwest National
Laboratory, Richland, WA 99352, USA
The goal of this work is to investigate the effect of SOM
matrix properties on the photochemical kinetics of photolabile,
organic molecules within the SOM by varying the temperature
and RH. Experiments were performed on three types of organic
d
Physical Sciences Division, Pacific Northwest National Laboratory, Richland,
WA 99352, USA
matrix including a-pinene/O
3
SOM (PSOM), limonene/O
3
SOM
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
c5cp05226b
(LSOM), and aged limonene/O
3
SOM obtained by exposure of
This journal is ©the Owner Societies 2016
Phys. Chem. Chem. Phys., 2016, 18, 8785--8793 | 8785

View Article Online
Paper
PCCP
LSOM to ammonia (brown LSOM). These particular systems be diffusion (and thus viscosity) limited. Diffusion limitations in a
were selected for two main reasons. First, PSOM and LSOM are more viscous matrix would allow more time for competing
common SOA materials in both outdoor and indoor environ- deactivation (reaction (3)) to occur, thus preventing formation of
ments: a-pinene is the most important precursor to global SOA 2,4-DNP photoproducts (reaction (4)). For example, 2,4-DNP*
2
6,27
and limonene is the most important indoor SOA precursor.
would need just a few nanoseconds to diffuse 0.6 nm (a reason-
Second, brown LSOM was selected to test both the effect of able estimate for the distance to the closest abstractable H-atom)
exposure to ammonia on the viscosity of LSOM and to examine in a moderately viscous solvent such as octanol (viscosity B7 Â
À3
a different type of photochemistry. (We did not do experiments 10 Pa s), but it would need several milliseconds to diffuse the
4
with PSOM exposed to ammonia because reaction of PSOM with same distance in PSOM (viscosity B10 Pa s). The latter time is
28
ammonia, unlike that of LSOM, does not lead to browning.)
more than sufficient for the dissipation of the electronic excitation
Although the elemental composition of LSOM and PSOM compo- energy into heat. The excited states lifetime of 2,4-DNP has not
nents is quite similar, with average O/C ratios of 0.39 and 0.37 been reported, but based on the triplet lifetimes of related
respectively, the viscosity of PSOM is likely to be greater than that nitrobenzene and nitrophenols (ortho, meta, and para) deter-
36
of LSOM because the products of a-pinene ozonolysis tend to be mined by Takezaki et al. we expect it to be o1 ns. Therefore,
29,30
more structurally rigid.
Similarly, brown LSOM may be physi- diffusion limitations may arise in materials with viscosity compar-
cally and chemically different from LSOM because it contains able to that of PSOM.
30
3
products of reactions of NH with carbonyls. Understanding how
Excitation: 2,4-DNP + hn - 2,4-DNP*
(R1)
viscosity of SOA may play a role in the photochemical kinetics
of different types of atmospheric molecules trapped within them is
important in order to better predict the lifetimes of toxic
pollutants.
Ideally, these experiments should be done with submicron
aerosol particles, similar to the ones found in the atmosphere,
in order to maintain the correct time scale for the gas-to-
Reaction: 2,4-DNP* + R–H - 2,4-DNP–H + R (R2)
Deactivation: 2,4-DNP* - 2,4-DNP + heat
Further reactions: 2,4-DNP–H - products
(R3)
(R4)
Our previous study examined the photochemistry of 2,4-DNP
particle exchange of volatile products and reactants (e.g., mole- at a range of temperatures in three different environments:
2
5
cular oxygen dissolved in the particles) of the photochemical water, octanol, and PSOM. It was found that the rate of
processes occurring in particles. However, experiments performed photochemical degradation of 2,4-DNP in PSOM was more
with aerosol particles would be more difficult and suffer from strongly affected by temperature than the same reaction in
potential interference from competing gas-phase photochemical octanol. We proposed that viscosity of the matrix is a key factor
3
1
processes. We have chosen to simplify the experiments by affecting the rates of photoreactions occurring in PSOM.
working with the bulk phase of SOM collected on a substrate. However, other interpretations of the observations could not
For photochemical reactions that involve only condensed- be excluded, for example, a different reaction mechanism in
phase reactants and products, bulk SOM experiments should PSOM compared to octanol. To further investigate the possible
provide the same information as experiments with aerosols.
matrix effects on the photochemistry of 2,4-DNP, we extended
As PSOM and LSOM photodegrade relatively slowly, a strongly our measurements to various temperatures and RHs in two
absorbing, nitroaromatic pollutant 2,4-dinitrophenol (2,4-DNP) different types of matrices – PSOM and related material LSOM.
was dispersed in the SOM films to serve as a photochemical
The third system examined in this study is brown LSOM
probe molecule. This compound enters the environment pre- formed by exposing fresh LSOM to ammonia vapors and drying
dominantly from pesticide use, manufacturing plants, biomass the sample. It has a distinctive absorption band at 510 nm that,
32–35
burning, and motor vehicles.
It is known to be present in in combination with tails of UV absorption bands extending
surface and atmospheric waters and in organic matter. Although into the blue region of the spectrum, give the material a
3
0,37–40
mixing of 2,4-DNP and biogenic SOA is unlikely, we used 2,4-DNP characteristic brown color.
Lee et al. found that an
just as a probe of SOM matrix effects on photochemistry. The aqueous solution of brown LSOM readily photobleaches after
photodegradation of 2,4-DNP upon UV irradiation is accompanied several minutes of exposure to UV radiation; the peak at 510 nm
2
8
by a decay in its absorbance at 290 nm and an increase in decays and the brown color disappears. Because this system
absorbance in the visible region of the spectrum, where PSOM photodegrades quickly and possesses a distinct UV-absorbing
and LSOM do not absorb, due to the formation of photo- functionality, an additional probe molecule was not used.
2
5
products.
,4-DNP is expected to photodegrade by a mechanism not presently known, it is likely to be an oligomeric, nitrogen
shown in reactions (1)–(4). As 2,4-DNP is irradiated, it is excited containing compound formed by condensation reactions
Although the molecular nature of the actual chromophore is
2
3
0
to a triplet state (reaction (1)) and proceeds to abstract a hydrogen between SOA carbonyls and ammonia.
atom from a neighboring molecule possessing weakly-bound We find that all three model systems exhibit similar beha-
hydrogen atoms (reaction (2)). Saturated hydrocarbons, aromatic vior wherein the photodegradation rate is increased at higher
hydrocarbons, and water are not suitable hydrogen donors in this temperature or higher RH. While the temperature effects can
reaction, whereas alcohols and aldehydes readily participate in result from a number of factors, the fact that there is also an
the H-atom transfer. We assume that the rate of reaction (2) may RH dependence suggests that viscosity is an important factor in
8
786 | Phys. Chem. Chem. Phys., 2016, 18, 8785--8793
This journal is ©the Owner Societies 2016

View Article Online
PCCP
Paper
photochemical reactions taking place within matrices. Increased
viscosity may be hindering the motions of the molecules in the
SOM and therefore slowing down the photochemical reactions in
which they participate.
Experimental
SOM for each experiment was generated in a 20 L flow tube by
dark ozonolysis of an SOA precursor volatilized under a flow of
dry air. Typical mixing ratios in the flow tube were 70 ppm
for O3 and 10 ppm for the precursor, either a-pinene (98%,
Alfa Aesar) or limonene (97%, Sigma-Aldrich); the flow tube
residence time was about 4 minutes. After passing through a
charcoal denuder to eliminate excess ozone and organic vapors,
the particles were collected using a Sioutas slit impactor
Fig. 1 A schematic diagram of the relative humidity- and temperature-
controlled photoreaction box used to irradiate all samples at various
RH and temperatures.
(Cat. No. 225-370) equipped with a single stage D (0.25 mm
cut point at 9 SLM collection flow rate), which was customized
to accommodate 25 mm CaF2 windows as the impaction 5 1C, two Peltier coolers (TE Technology, Inc. TC-48-20) were
substrates. The average collection time of 45 minutes generally utilized in addition to the water cooler. The temperature was
produced about 2 mg of SOM deposited on the window.
monitored throughout the experiment using a type-K thermo-
For the first two sets of experiments (PSOM/2,4-DNP and couple mounted on the microscope slide.
LSOM/2,4-DNP), 2,4-DNP (99.4%, Sigma Aldrich) was added to
The RH within the box was controlled by mixing flows of dry
SOM by pipetting 100 mL of a 0.01 M 2,4-DNP/methanol (o2% RH) and wet (498% RH) air. Incoming dry air was first
solution onto the CaF window, partially dissolving the SOM. split into two separate flows controlled by two mass flow
2
The methanol was allowed to evaporate and 2,4-DNP to permeate controllers. The humidified flow was obtained by passing dry
the SOM for 30 minutes, after which the 2,4-DNP remained air through a Nafion drier/humidifier (PermaPure). The final
embedded in a thin film of SOM (as far as we could ascertain RH was set by adjusting the flow through each mass flow
based on the inspection of the film under optical and Raman controller. The RH inside the box was measured with a Vaisala
microscopes). It was necessary to keep the concentration of HMP237 probe.
embedded 2,4-DNP low enough to ensure that the viscosity of
Irradiation of each sample was performed with the output of
the SOM matrix was not significantly affected, yet also high a 150 W xenon arc lamp (Newport model 66902 lamp housing)
enough to absorb more radiation than the surrounding that was reflected by a 280–400 nm dichroic mirror and passed
SOM molecules. Therefore, we elected to use a mass ratio of through a 295 nm long-pass filter (Schott WG295) and UV band-
2
5
2,4-DNP : SOM around 1 : 50 to satisfy both requirements.
pass filter (Schott BG1) in order to isolate a band (290–400 nm)
For the third set of experiments, brown LSOM was produced of the actinic radiation that is most relevant for tropospheric
by exposing fresh LSOM to ammonia vapors as described in photochemistry. The light was then piped into the photoreaction
ref. 37. First, a thin film of LSOM was created by pipetting box through a liquid light guide (Edmunds #53-691) at a 151
1
00 mL of methanol onto the LSOM, allowing it to spread over angle with respect to the film normal (see Fig. 1).
the window, and evaporating the solvent for 30 minutes. The Absorption spectra of the SOM film were recorded using a
window was then inserted into a small glass petri dish, which custom spectrometer setup consisting of a D /W light source
2
was placed carefully into a larger petri dish containing a and an Ocean Optics USB4000 spectrometer. The radiation
solution of 0.1 M ammonium sulfate (499%, EMD) thus from the light source was transmitted between spectrometer
allowing it to absorb the vapors (estimated to contain 300 ppb components and across the sample using 600 mm optical fibers
4
1
NH using the AIM-II model), but not touch the actual solution. as is shown in Fig. 1.
3
The larger petri dish was covered with a lid and the entire unit
Viscosity measurements of SOM were performed using the
was stored in the dark. After two days of exposure, the window poke-flow technique described in ref. 3, 23 and 42. The method
was removed and dried under a flow of dry air for at least one builds on the qualitative approach introduced by Murray et al.
4
3
hour, after which the SOM film had browned.
to determine the flow characteristics of particles. Briefly, SOM
After each sample (2,4-DNP in SOM or brown LSOM) was was generated in the 20 L flow tube as described above and
prepared, it was irradiated inside the temperature and RH collected using the Sioutas impactor on substrates coated
controlled setup shown in Fig. 1. The CaF window loaded with with trichloro(1H,1H,2H,2H-perfluorooctyl)silane (Sigma-Aldrich).
2
sample was placed on top of a quartz microscope slide that was The coating procedure is described in detail in ref. 44. During
maintained at a desired temperature ranging from 0 1C to 35 1C collection, the aerosol particles from the flow tube coagulated
using a circulating water cooler connected to an aluminum on the hydrophobic substrates creating supermicron SOM parti-
heat sink. If needed, in order to achieve temperatures below cles. After collection, the substrates containing the supermicron
This journal is ©the Owner Societies 2016
Phys. Chem. Chem. Phys., 2016, 18, 8785--8793 | 8787

View Article Online
Paper
PCCP
particles were transported to the University of British Columbia of fluid flow is capable of providing both lower and upper limits
and stored at B280 K until the viscosity measurements were of viscosity that are consistent with literature or measured values
2
carried out.
when the viscosity of particles are in the range of 5 Â 10 to
6
42
The supermicron SOM particles on the hydrophobic sub- 3 Â 10 Pa s. The major source of uncertainty in the viscosity of
strates were placed inside a RH-controlled flow-cell to carry out the SOM arises from uncertainty in the physical properties of
4
4–46
viscosity measurements.
The flow-cell contained a small SOM that are used in simulations (i.e., values shown in Table 1).
hole in the top, which allowed a sterilized sharp needle (0.9 mm Particle to particle variability of t(exp,flow) is typically small.
 40 mm with a 10 mm tip, Becton-Dickson, USA) to pass
through and poke the particles. The needle was mounted on a
Results and discussion
micromanipulator (Narishige, model MO-202U, Japan) allowing
precise control of the movement of the needle. Prior to poking,
The absorption spectra of (a) 2,4-DNP in LSOM and (b) brown
the geometry of a particle was a spherical cap. Due to the force
LSOM taken during the course of photodegradation are pre-
by the needle while poking, the spherical cap geometry was
sented in Fig. 2. The inset of each graph corresponds to the
converted into a half-torus geometry. After the needle was
absorbance decay at the representative wavelength for each
removed, the material flowed to minimize the surface energy
system (290 nm for 2,4-DNP in LSOM and 510 nm for brown
of the system and eventually returned back to the spherical cap
LSOM). The combined absorbance from the SOM and the
geometry by filling the central hole. This process was moni-
reactant can be expected to follow eqn (1),
tored and recorded using a reflectance optical microscope
Àkt
(
Zeiss Axio Observer, 40Â objective) equipped with a CCD
A(t) = ASOM + A [b + (1 Àb) Â e
]
(1)
0
camera. The diameter of the equivalent hole area produced
by poking the particle was calculated based on the relationship
d = (4A/p) , where d is the equivalent area diameter of a hole
with area A. The experimental flow time, t(exp,flow), was deter-
mined as the time needed for d to reach half of the initial value
of the inner hole of the half torus.
where ASOM is the absorbance due to the SOM matrix, assumed
to be unchanged by photolysis, A is the starting absorbance
due to the reactant, k is the photodegradation rate constant,
and b is the ratio of the absorption coefficient of the photolysis
product(s) to that of the reactant at the wavelength of interest.
The value of b is 0 for non-absorbing products, smaller than
for weakly absorbing products, 1 at the isosbestic point, and
larger than 1 when products absorb stronger than the starting
compound.) Since we were interested only in the values of k, the
observed decays were fit using the simplified equation:
1
/2
0
4
7
(
1
Simulations of fluid flow were carried out using COMSOL
Multiphysics (version 4.3a) to convert t
values into
exp,flow)
(
2
3,42
viscosity.
In the simulations, a half-torus geometry with
an inner radius, R, and a tube radius, r, were required. These
values were set to match the experiments, and the viscosity was
varied until the modeled flow time, t(mod,flow), was equal to
Àkt
A(t) = const
1
+ const
2
+ e
(2)
t(exp,flow). The physical parameters needed for these simulations
include: slip length (the interaction between the fluid and solid to obtain rate constants for each set of experimental conditions.
surface), surface tension and density of the SOM, and the The use of eqn (2) automatically accounts for any uncompensated
contact angle at the particle–substrate interface. Upper and wavelength-independent offsets in the absorbance measurements.
lower limits of the parameters used in these simulations are
It should be noted that in the case of 2,4-DNP, the decay in
provided in Table 1. Previous validation experiments with absorbance at B250–320 nm is accompanied by an increase in
sucrose–water particles and high viscosity standards have the broad absorption band in the visible range (B400–450 nm)
shown that the poke-flow technique combined with simulations indicating formation of a brown carbon product (a result of
Table 1 Physical parameters used in the simulations to determine viscosities of a-pinene-derived SOM (PSOM) and limonene-derived SOM (LSOM). R
and r indicate the radius of the tube and the radius of the inner hole, respectively, for a half-torus geometry
a
b
c
d
Slip length
nm)
Surface tension
Density
(g cm
Contact angle
(1)
À1
À3
(
(mN m
)
)
PSOM
LSOM
Values for lower limit
Values for upper limit
5
40
75
1.30
1.30
70 (if r o 2R), 90 (if r 4 2R)
90 (if r o 2R), 70 (if r 4 2R)
10 000
Values for lower limit
Values for upper limit
5
23
72
1.47
1.67
60 (if r o 2R), 80 (if r 4 2R)
80 (if r o 2R), 60 (if r 4 2R)
10 000
Brown LSOM
Values for lower limit
Values for upper limit
5
23
72
1.47
1.67
65 (if r o 2R), 80 (if r 4 2R)
80 (if r o 2R), 65 (if r 4 2R)
10 000
a
b
62
Ref. 48–61. For PSOM, surface tension values were based on the viscosity of model compounds. For LSOM and brown LSOM, surface tension
values were based on the estimated surface tension of liquid limonene at 293 K (as the lower limit) and the surface tension of pure water at 293 K
63 c
(as the upper limit) values from ACD/Labs (chemspider.com) and Engelhart et al.
For PSOM, density values are based on Chen and Hopke,
64
65 d
2
009. For LSOM and brown LSOM, density values are based on values from Kostenidou et al.
Contact angles (70–901 for PSOM and 60–801 for
LSOM) were determined using 3-D fluorescence confocal images of the SOM particles on the substrates.
8
788 | Phys. Chem. Chem. Phys., 2016, 18, 8785--8793
This journal is ©the Owner Societies 2016

View Article Online
PCCP
Paper
Fig. 3 Products tentatively identified by LC-PDA-MS after 2,4-DNP
photolysis in isopropanol. Details leading to the assignments are provided
in the ESI.†
of many different molecules that would yield a complicated
2
9
background spectrum. The full results of the LC-PDA-MS
analysis, as well as experimental details, are provided in the
ESI.† We assigned the observed products of photodegradation
of 2,4-DNP by analogy with photochemistry of related com-
pounds, nitrobenzene and 2-nitrophenol. Based on the m/z
values and UV-vis absorption spectra of the eluted peaks, we
were able to assign products to the structures shown in Fig. 3. We
emphasize that we did not have standards for any of these
compounds for unambiguous identification, so the assignments
should be regarded as tentative. Nevertheless, our results strongly
2
suggest that photoreduction of –NO group(s) in 2,4-DNP is the
main mechanism of photodegradation. We observed peaks
Fig. 2 Representative absorption spectra recorded during the photo- corresponding to an –NH
(aniline) product as well as to an
2
degradation of (a) 2,4-DNP in LSOM and (b) brown LSOM. The insets show
–NO (nitroso) intermediate which is expected for this process.
the decay of (a) the 290 nm peak of 2,4-DNP and (b) the 510 nm peak
characteristic of the brown LSOM chromophore where the red trace is the
experimental decay and the black trace is the fit to eqn (2). The arrows
indicate the wavelength at which the spectra are changing significantly.
Based on the UV-vis absorption spectra (Fig. S9 of the ESI†) the
compounds containing an aniline group are likely the ones
responsible for the brown color of the 2,4-DNP photodegradation
products after the irradiation and cause the growth of the visible
absorption band at B400–450 nm. The formation of aniline
reduction of one of the –NO
below). The rate constants k determined from eqn (1) and (2) at environmental fate of 2,4-DNP because of the high reactivity of
90 nm (decrease in absorbance, b o 1, const 4 0) and 420 nm aniline compounds.
2 2
groups to –NH group as discussed products has significant implications for understanding the
2
2
(
increase in absorbance, b 4 1, const o 0) were the same within
Regarding the kinetics of 2,4-DNP photodegradation, as
discussed in Lignell et al., the effective activation energy of
2
2
5
the uncertainties of the fitting.
The increase in the visible absorbance of the irradiated reactions with low intrinsic barriers is determined by the
,4-DNP samples is an important observation for two reasons. surrounding matrix’s viscosity, which depends strongly on
2
First, it confirms recent findings that photochemical proces- temperature. The photodegradation of 2,4-DNP fits this
sing is capable of altering the light absorption properties profile – it has low activation energy in octanol and much
2
8,66,67
of brown carbon.
Second, it shows that, depending higher activation energy in PSOM, likely due to the high
2
5
on the system, photochemical processes in brown carbon are viscosity of the PSOM film. The activation energy of each
capable of creating light-absorbing compounds, not just system studied in the current work was determined using the
destroying (photobleaching) them. Our work provides evidence Arrhenius plot shown in Fig. 4. This plot illustrates the relation-
that brown carbon has a dynamic absorption spectrum that ship between temperature and rate constant for each system
can be altered on atmospherically relevant time scales via (including 2,4-DNP in PSOM shown previously in ref. 25) under
photochemistry.
dry conditions. The results indicate that the activation energy
À1
To investigate the mechanism of 2,4-DNP photodegradation, of the photodegradation of 2,4-DNP in LSOM (24 Æ 1 kJ mol
)
À1
we performed LC-PDA-MS (liquid chromatography coupled to is lower than that in PSOM (48 Æ 6 kJ mol ). If these activation
a photodiode array detector and an electrospray ionization energies are actually viscosity-dependant, then this suggests
high-resolution mass spectrometer) measurements on irra- that PSOM has a higher viscosity than LSOM. Another observa-
diated samples of 2,4-DNP in isopropanol (0, 30, and 60 min tion of the current work is that both of the reactions that
of irradiation in a quartz cuvette by the same light source as occurred in LSOM (2,4-DNP and brown LSOM) had similar
À1
À1
used for the film experiments). Isopropanol was used as the activation energies (24 Æ 1 kJ mol and 16 Æ 5 kJ mol
,
solvent in this analysis, as SOM is very complex and is made up respectively) implying that the viscosities of LSOM and brown
This journal is ©the Owner Societies 2016
Phys. Chem. Chem. Phys., 2016, 18, 8785--8793 | 8789

View Article Online
Paper
PCCP
Fig. 4 Arrhenius plots of the photodegradation rate of 2,4-DNP in PSOM
(
red), 2,4-DNP in LSOM (blue), and brown LSOM (green) under dry
conditions. The slopes correspond to activation energies of 48, 24, and
À1
1
6 kJ mol , respectively. Markers with no error bars represent experi-
ments that were performed once. All other markers are the average of 2–6
data points obtained at each temperature.
LSOM are similar. This would be consistent with minor com-
6
8
positional differences between the two materials.
In order to test this hypothesis, the viscosities of LSOM
and PSOM generated in our flow tube were experimentally
determined from poke flow experiments at different RHs,
specifically, by measuring how quickly a half-sphere of SOM
Fig. 5 (a) Average experimental flow time from the poke-flow experi-
ments as a function of relative humidity. The markers correspond to the
returns to its original shape after being distorted by a needle. average of 4–11 individual poke flow measurements. The error bars
The results of these experiments are shown in Fig. 5. Fig. 5a represent one standard deviation for the repeated measurements.
(
b) Simulated ranges of viscosities as determined by upper and lower
depicts the average experimental flow time, which is taken for
the equivalent area diameter to decrease to 50% of its initial
diameter. The PSOM had a longer experimental flow time than
either LSOM or brown LSOM. For example, under dry condi-
tions the experimental flow time of PSOM is longer by approxi- RH 4 95%,
limits. The error bars are dominated by the uncertainties in the fitting
parameters listed in Table 1.
6
9,70
but the increase in the photodegradation rate
mately an order of magnitude than the experimental flow times is also observed at lower RH. Therefore we do not think that
of LSOM and brown LSOM, strongly suggesting a higher LLPS is responsible for the RH dependence of the photo-
viscosity. Unlike the clearly different poke flow times in PSOM degradation rate. The change of RH seemed to have a larger
and LSOM, the error bars on the absolute viscosity values effect on the brown LSOM system than either of the 2,4-DNP/
shown in Fig. 5b overlap because of the sensitivity to the model SOM systems. This is most likely due to the fact that 2,4-DNP
2
5
parameters needed to convert the poke flow times into the photodegrades extremely slowly in water. As water molecules
viscosity values. However, the simulated upper and lower limits are introduced to the film, the rate of photodegradation may be
of viscosity also suggest that PSOM is more viscous than LSOM increased due to decreased viscosity, but it is partly offset by
or brown LSOM. These results are qualitatively consistent with the dilution of the molecules that 2,4-DNP can react with. The
larger apparent activation energies for the photodegradation of photodegradation of brown LSOM, on the other hand, is not
2,4-DNP in PSOM vs. LSOM.
affected by the presence of water molecules. Therefore, the
It is also clear from Fig. 5 that the viscosity of SOM is increase in the rate of photodegradation due to decreased
strongly related to RH, which is in agreement with previous viscosity will be greater for brown LSOM than for 2,4-DNP in
3
measurements. Irradiation of all systems under a range of RHs LSOM or PSOM which may explain the larger RH dependence
revealed that as RH was increased, the photoreaction rate seen in brown LSOM.
constant increased (Fig. 6). The number of data points in
We should point out that viscosity is not the only property of
Fig. 6 is limited because each point requires several days of SOM that can change with RH. It is conceivable that the
experiments, however, the increase in the rate with RH is clear. presence of water in the SOM film changes the molecular
We interpret this to be the result of water molecules softening composition by means of hydration of aldehydes, hydrolysis
the SOM matrix and allowing for the photoexcited molecules to of anhydrites, and other reactions involving water. For example,
diffuse to their reaction partners faster. We note that liquid– hydrolysis reactions of organonitrates in aerosols have been
7
1
liquid phase-separation (LLPS) is possible in PSOM at very high observed at the RH levels used in this work. While these
8790 | Phys. Chem. Chem. Phys., 2016, 18, 8785--8793
This journal is ©the Owner Societies 2016

View Article Online
PCCP
Paper
demonstrated by Fig. 6b (note the small vertical error bars
calculated from the repeated experiments). The red dots corre-
spond to samples of 2,4-DNP in PSOM that were dried/
humidified under a 0.3 SLM flow rate, and the light blue dot
represents the samples that were dried under a 6.5 SLM flow
rate. The fast dried samples with a low viscosity core had a
photodegradation rate that was a factor of B3 higher than the
slow dried samples with a high viscosity core. This provides
additional indirect evidence that the viscosity of SOM plays an
important role in the photodegradation of these systems.
These results suggest that certain types of photochemical
reactions may be suppressed in condensed-phase environment
of the highly-viscous aerosols relative to the same reactions in
common organic solvents. In particular, photochemical processes
involving secondary reactions of electronically-excited organic
molecules with the matrix constituents are likely to be affected
by the diffusion limitations, and therefore, by the viscosity. This
scenario clearly applies to 2,4-DNP, which photodegrades by a
reaction of its triplet excited state with suitable hydrogen atom
donors (reactions (1)–(4)). The photodegradation rate of the
unidentified chromophore in the brown LSOM is similarly
affected by viscosity. However, we want to emphasize that not
all organic photochemical reactions will be suppressed in viscous
aerosols by the same mechanism. Direct photolysis processes
occur on faster time scales and do not require the excited
molecules to diffuse through the matrix before the reaction. Such
Fig. 6 (a) The rate constants as a function of relative humidity for the
photodegradation of 2,4-DNP in PSOM 2,4-DNP in LSOM, and brown
LSOM. Each marker with error bars represents the average of all the data reactions are less likely to be affected by the material viscosity; for
points at each RH. These experiments were repeated 2–6 times. The example, Norrish-I splitting of carbonyls can be efficient even in
À6
À6
slopes of the fit were (4.7 Æ 0.1) Â 10 , (2.9 Æ 0.9) Â 10 , and (8 Æ 2) Â
75
crystals. On the other hand, caging effects could be more
À6 À1
1
0
s /% RH for 2,4-DNP in PSOM 2,4-DNP in LSOM, and brown LSOM
significant in highly-viscous solvents, and result in suppression
of direct photolysis quantum yields. Therefore, the effects of
viscosity on photochemical reactions need to be investigated on
respectively. Plot (b) demonstrates the effect of drying speed.
reactions may also contribute to the observed changes in the a case by case basis.
photodegradation rate, the physical effect of RH on the SOM
These results could help in the interpretation of atmospheric
viscosity likely dominates over the chemical effect of RH on the lifetimes of particulate nitro-aromatic compounds, some of
molecular composition of SOM. Additional experiments with which are known toxins. In cold and dry segments of the atmo-
materials that are not reactive towards water are needed to fully sphere, such as over the polar regions and near the tropopause,
separate the chemical and physical effects of RH.
these compounds could survive longer if they are trapped inside
Koop et al. and other groups have previously found that the highly-viscous particles. In contrast, the same compounds could
viscosity of an aerosol particle at a particular RH strongly depends degrade faster under warm, humid conditions or when they are
2,72–74
on the rate at which the system is dried/humidified.
It has unprotected on the particle surface. It is possible that other types
been suggested that drying quickly will cause water molecules of photolabile organic compounds can similarly be protected
nearest the surface of the particle to evaporate quickly, creating a from photodegradation under cold, dry conditions. It is also
highly viscous ‘‘crust’’ that prevents water molecules in the interior possible that photosensitized reactions can similarly be sup-
of the particle from escaping. This results in an aerosol particle pressed by the viscosity of the SOM matrix. These effects will
2,72
that contains a lower viscosity core.
Drying more slowly allows need to be investigated in future studies. It also remains to be
more time for water molecules within the particle to diffuse to the seen if these conclusions can be applied to other types of SOA,
outer layers and evaporate, causing the particle to be more uni- including those formed from anthropogenic sources.
form in terms of viscosity. Thus the interior viscosity of the slowly
dried particle is actually expected to be higher than that of the
interior of the quickly dried particle.
Conclusions
In our studies, the lowest RH experiments were performed
2
5
under two different flow rates of dry air (6.5 SLM and 0.3 SLM). In our previous work, we found that the photodegradation
We found that the samples that were dried under a high flow rate of 2,4-DNP in viscous PSOM had a significantly stronger
rate of dry air photodegraded more quickly than samples dried temperature dependence than the same process in much less
under the slower flow rate. This was a reproducible effect, viscous octanol. We suggested that this effect was due to the
This journal is ©the Owner Societies 2016
Phys. Chem. Chem. Phys., 2016, 18, 8785--8793 | 8791

View Article Online
Paper
PCCP
fact that temperature has a stronger effect on the viscosity of
PSOM than on the viscosity of octanol. The increased viscosity
may be hindering the motion of electronically excited 2,4-DNP
molecules within SOM, slowing its photodegradation. In order
to examine this effect, we expanded the scope of the previous
measurements to investigate the effect of relative humidity on
the photodegradation kinetics of 2,4-DNP in PSOM. We also
7 R. A. Zaveri, R. C. Easter, J. E. Shilling and J. H. Seinfeld,
Atmos. Chem. Phys. Discuss., 2013, 13, 28631–28694.
8 T. D. Vaden, D. Imre, J. Ber ´a nek, M. Shrivastava and
A. Zelenyuk, Proc. Natl. Acad. Sci. U. S. A., 2011, 108,
2190–2195.
9 M. Shiraiwa, A. Zuend, A. K. Bertram and J. H. Seinfeld,
Phys. Chem. Chem. Phys., 2013, 15, 11441–11453.
investigated LSOM as an alternative organic matrix for 2,4-DNP 10 I. Riipinen, J. R. Pierce, T. Yli-Juuti, T. Nieminen,
photodegradation, and carried out similar experiments with a
completely different photochemical system, specifically brown
LSOM obtained by exposure of LSOM to ammonia. In all cases,
as the viscosity of the SOM was increased by cooling the
material or exposing it to dry air, the reaction rate decreased.
S. H ¨a kkinen, M. Ehn, H. Junninen, K. Lehtipalo, T. Pet ¨a j ¨a ,
J. Slowik, R. Chang, N. C. Shantz, J. Abbatt, W. R. Leaitch,
V. M. Kerminen, D. R. Worsnop, S. N. Pandis, N. M.
Donahue and M. Kulmala, Atmos. Chem. Phys., 2011, 11,
3865–3878.
The activation energy for the 2,4-DNP photodegradation in 11 D. L. Bones, J. P. Reid, D. M. Lienhard and U. K. Krieger,
PSOM and LSOM were correlated with the explicitly measured
viscosity of the materials.
Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 11613–11618.
12 M. Shiraiwa, M. Ammann, T. Koop and U. P o¨ schl, Proc. Natl.
Acad. Sci. U. S. A., 2011, 108, 11003–11008.
13 B. Wang, R. E. O’Brien, S. T. Kelly, J. E. Shilling, R. C. Moffet,
Acknowledgements
M. K. Gilles and A. Laskin, J. Phys. Chem. A, 2015, 119,
4498–4508.
P. L., A. L., J. L., and S. N. acknowledge support by the U.S.
Department of Commerce, National Oceanic and Atmospheric
Administration through Climate Program Office’s AC4 pro-
gram, awards NA13OAR4310066 (PNNL) and NA13OAR4310062
1
4 H. C. Price, J. Mattsson, Y. Zhang, A. K. Bertram, J. F. Davies,
J. W. Grayson, S. T. Martin, D. O’Sullivan, J. P. Reid,
A. M. J. Rickards and B. J. Murray, Chem. Sci., 2015, 6,
4876–4883.
(UCI). MH and HL acknowledge support by the National
1
1
1
1
1
2
2
5 A. Zelenyuk, D. Imre, J. Ber ´a nek, E. Abramson, J. Wilson and
M. Shrivastava, Environ. Sci. Technol., 2012, 46, 12459–12466.
6 R. M. Power, S. H. Simpson, J. P. Reid and A. J. Hudson,
Chem. Sci., 2013, 4, 2597–2604.
Science Foundation (NSF) grant CHE-0909227. MB was sup-
ported by the NSF summer research undergraduate experience
(RUE) program. MS, JG, and AB were supported by the Natural
Sciences and Engineering Research Council of Canada. The
LC-PDA-MS measurements were performed at the W. R. Wiley
Environmental Molecular Sciences Laboratory (EMSL),
national scientific user facility located at PNNL, and sponsored
by the Office of Biological and Environmental Research of the
US. DOE. PNNL is operated for US DOE by Battelle Memorial
Institute under Contract No. DEAC06-76RL0 1830.
7 J. D. Hearn and G. D. Smith, Phys. Chem. Chem. Phys., 2005,
7, 2549–2551.
a
8 D. A. Knopf, L. M. Anthony and A. K. Bertram, J. Phys. Chem.
A, 2005, 109, 5579–5589.
9 M. Kuwata and S. T. Martin, Proc. Natl. Acad. Sci. U. S. A.,
2012, 109, 17354–17359.
0 S. Zhou, M. Shiraiwa, R. D. McWhinney, U. Poschl and
J. P. D. Abbatt, Faraday Discuss., 2013, 165, 391–406.
1 V. Perraud, E. A. Bruns, M. J. Ezell, S. N. Johnson, Y. Yu,
M. L. Alexander, A. Zelenyuk, D. Imre, W. L. Chang,
D. Dabdub, J. F. Pankow and B. J. Finlayson-Pitts, Proc.
Natl. Acad. Sci. U. S. A., 2012, 109, 2836–2841.
References
1
2
3
C. Kidd, V. Perraud, L. M. Wingen and B. J. Finlayson-Pitts,
Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 7552–7557.
T. Koop, J. Bookhold, M. Shiraiwa and U. Poeschl, Phys. 22 H. C. Price, B. J. Murray, J. Mattsson, D. O’Sullivan, T. W.
Chem. Chem. Phys., 2011, 13, 19238–19255.
Wilson, K. J. Baustian and L. G. Benning, Atmos. Chem.
Phys., 2014, 14, 3817–3830.
L. Renbaum-Wolff, J. W. Grayson, A. P. Bateman,
M. Kuwata, M. Sellier, B. J. Murray, J. E. Shilling, S. T. 23 M. Song, P. Liu, S. Hanna, S. Martin and A. Bertram, Atmos.
Martin and A. K. Bertram, Proc. Natl. Acad. Sci. U. S. A., 2013,
10, 8014–8019.
A. Virtanen, J. Joutsensaari, T. Koop, J. Kannosto, P. Yli-
Chem. Phys. Discuss., 2015, 15, 1131–1169.
24 A. P. Bateman, H. Belassein and S. T. Martin, Aerosol Sci.
Technol., 2013, 48, 42–52.
1
4
Pirila, J. Leskinen, J. M. Makela, J. K. Holopainen, U. Poschl, 25 H. Lignell, M. L. Hinks and S. A. Nizkorodov, Proc. Natl.
M. Kulmala, D. R. Worsnop and A. Laaksonen, Nature, 2010,
67, 824–827.
A. P. Bateman, A. K. Bertram and S. T. Martin, J. Phys. Chem.
A, 2015, 119, 4386–4395.
Acad. Sci. U. S. A., 2014, 111, 13780–13785.
26 R. K. Pathak, A. A. Presto, T. E. Lane, C. O. Stanier, N. M.
Donahue and S. N. Pandis, Atmos. Chem. Phys., 2007, 7,
3811–3821.
4
5
6
V. Perraud, E. A. Bruns, M. J. Ezell, S. N. Johnson, Y. Yu, 27 T. Wainman, J. Zhang, C. J. Weschler and P. J. Lioy, Environ.
M. L. Alexander, A. Zelenyuk, D. Imre, W. L. Chang, Health Perspect., 2000, 108, 1139–1145.
D. Dabdub, J. F. Pankow and B. J. Finlayson-Pitts, Proc. 28 H. J. Lee, P. K. Aiona, A. Laskin, J. Laskin and S. A.
Natl. Acad. Sci. U. S. A., 2012, 109, 2836–2841.
Nizkorodov, Environ. Sci. Technol., 2014, 48, 10217–10226.
8792 | Phys. Chem. Chem. Phys., 2016, 18, 8785--8793
This journal is ©the Owner Societies 2016

View Article Online
PCCP
Paper
2
3
3
3
9 D. E. Romonosky, A. Laskin, J. Laskin and S. A. Nizkorodov, 52 V. S. J. Craig, C. Neto and D. R. M. Williams, Phys. Rev. Lett.,
J. Phys. Chem. A, 2015, 119, 2594–2606. 2001, 87, 054504.
0 A. Laskin, J. Laskin and S. A. Nizkorodov, Chem. Rev., 2015, 53 D. C. Tretheway and C. D. Meinhart, Phys. Fluids, 2002, 14,
15, 4335–4382. L9–L12.
1 C. George, M. Ammann, B. D’Anna, D. J. Donaldson and 54 J. T. Cheng and N. Giordano, Phys. Rev. E: Stat., Nonlinear,
S. A. Nizkorodov, Chem. Rev., 2015, 115, 4218–4258. Soft Matter Phys., 2002, 65, 031206.
2 P. J. Shea, J. B. Weber and M. R. Overcash, in Residue 55 S. Jin, P. Huang, J. Park, J. Y. Yoo and K. S. Breuer, Exp.
1
Reviews, ed. F. Gunther and J. Gunther, Springer, New York,
983, vol. 87, ch. 1, pp. 1–41.
3 M. M. Sanagi, M. Miskam, W. A. Wan Ibrahim, D. Hermawan
and H. Y. Aboul-Enein, J. Sep. Sci., 2010, 33, 2131–2139.
4 A. Albinet, C. Minero and D. Vione, Chemosphere, 2010, 80,
Fluids, 2004, 37, 825–833.
56 P. Joseph and P. Tabeling, Phys. Rev. E: Stat., Nonlinear, Soft
Matter Phys., 2005, 71, 035303.
57 N. Chiara, R. E. Drew, B. Elmar, B. Hans-J u¨ rgen and
S. J. C. Vincent, Rep. Prog. Phys., 2005, 68, 2859.
58 C.-H. Choi and C.-J. Kim, Phys. Rev. Lett., 2006, 96, 066001.
1
3
3
3
3
3
3
753–758.
5 M. A. J. Harrison, S. Barra, D. Borghesi, D. Vione, C. Arsene 59 L. Joly, C. Ybert and L. Bocquet, Phys. Rev. Lett., 2006,
and R. Iulian Olariu, Atmos. Environ., 2005, 39, 231–248. 96, 046101.
6 M. Takezaki, N. Hirota and M. Terazima, J. Phys. Chem. A, 60 L. Zhu, C. Neto and P. Attard, Langmuir, 2012, 28,
997, 101, 3443–3448. 3465–3473.
7 K. M. Updyke, T. B. Nguyen and S. A. Nizkorodov, Atmos. 61 L. Li, J. Mo and Z. Li, Phys. Rev. E: Stat., Nonlinear, Soft
Environ., 2012, 63, 22–31. Matter Phys., 2014, 90, 033003.
8 T. B. Nguyen, P. B. Lee, K. M. Updyke, D. L. Bones, J. Laskin, 62 R. Tuckermann and H. K. Cammenga, Atmos. Environ.,
1
A. Laskin and S. A. Nizkorodov, J. Geophys. Res.: Atmos.,
012, 117, D01207.
9 D. L. Bones, D. K. Henricksen, S. A. Mang, M. Gonsior,
2004, 38, 6135–6138.
63 G. J. Engelhart, A. Asa-Awuku, A. Nenes and S. N. Pandis,
Atmos. Chem. Phys., 2008, 8, 3937–3949.
2
3
A. P. Bateman, T. B. Nguyen, W. J. Cooper and S. A. 64 X. Chen and P. K. Hopke, Indoor Air, 2009, 19, 335–345.
Nizkorodov, J. Geophys. Res.: Atmos., 2010, 115, D05203.
0 T. B. Nguyen, A. Laskin, J. Laskin and S. A. Nizkorodov,
Faraday Discuss., 2013, 165, 473–494.
1 S. L. Clegg, P. Brimblecombe and A. S. Wexler, J. Phys. Chem.
A, 1998, 102, 2137–2154.
2 J. W. Grayson, M. Song, M. Sellier and A. K. Bertram, Atmos.
Meas. Tech. Discuss., 2015, 8, 877–903.
3 B. J. Murray, A. E. Haddrell, S. Peppe, J. F. Davies, J. P. Reid,
D. O’Sullivan, H. C. Price, R. Kumar, R. W. Saunders,
65 E. Kostenidou, R. K. Pathak and S. N. Pandis, Aerosol Sci.
Technol., 2007, 41, 1002–1010.
4
4
4
4
66 R. Zhao, A. K. Y. Lee, L. Huang, X. Li, F. Yang and J. P. D.
Abbatt, Atmos. Chem. Phys. Discuss., 2015, 15, 2957–2996.
67 H. Forrister, J. Liu, E. Scheuer, J. Dibb, L. Ziemba, K. L.
Thornhill, B. Anderson, G. Diskin, A. E. Perring, J. P.
Schwarz, P. Campuzano-Jost, D. A. Day, B. B. Palm, J. L.
Jimenez, A. Nenes and R. J. Weber, Geophys. Res. Lett., 2015,
42, 2015GL063897.
J. M. C. Plane, N. S. Umo and T. W. Wilson, Atmos. Chem. 68 A. Laskin, J. Laskin and S. A. Nizkorodov, Chem. Rev., 2015,
Phys., 2012, 12, 8575–8587. 115, 4335–4382.
4 M. Song, C. Marcolli, U. K. Krieger, A. Zuend and T. Peter, 69 J. F. Pankow, Atmos. Environ., 2003, 37, 3323–3333.
4
4
4
Atmos. Chem. Phys., 2012, 12, 2691–2712.
5 T. Koop, A. Kapilashrami, L. T. Molina and M. J. Molina,
J. Geophys. Res.: Atmos., 2000, 105, 26393–26402.
70 M. D. Petters, S. M. Kreidenweis, J. R. Snider, K. A. Koehler,
Q. Wang, A. J. Prenni and P. J. Demott, Tellus, Ser. B, 2006,
58, 196–205.
6 A. K. Bertram, S. T. Martin, S. J. Hanna, M. L. Smith, 71 S. Liu, J. E. Shilling, C. Song, N. Hiranuma, R. A. Zaveri and
A. Bodsworth, Q. Chen, M. Kuwata, A. Liu, Y. You and
S. R. Zorn, Atmos. Chem. Phys., 2011, 11, 10995–11006.
7 P. C. Reist, Aerosol science and technology, McGraw-Hill,
L. M. Russell, Aerosol Sci. Technol., 2012, 46, 1359–1369.
72 H. J. Tong, J. P. Reid, D. L. Bones, B. P. Luo and U. K.
Krieger, Atmos. Chem. Phys., 2011, 11, 4739–4754.
73 D. J. Burnett, F. Thielmann and J. Booth, Int. J. Pharm., 2004,
287, 123–133.
4
1993.
4
4
8 E. Schnell, J. Appl. Phys., 1956, 27, 1149–1152.
9 N. V. Churaev, V. D. Sobolev and A. N. Somov, J. Colloid 74 B. Zobrist, V. Soonsin, B. P. Luo, U. K. Krieger, C. Marcolli,
Interface Sci., 1984, 97, 574–581.
T. Peter and T. Koop, Phys. Chem. Chem. Phys., 2011, 13,
5
5
0 K. Watanabe and H. Udagawa, AIChE J., 2001, 47, 256–262.
3514–3526.
1 J. Baudry, E. Charlaix, A. Tonck and D. Mazuyer, Langmuir, 75 M. A. Garcia-Garibay and L. M. Campos, ChemInform, 2004,
001, 17, 5232–5236. 35, DOI: 10.1002/chin.200419246.
2
This journal is ©the Owner Societies 2016
Phys. Chem. Chem. Phys., 2016, 18, 8785--8793 | 8793