Hygroscopic growth of pure and mixed components
Hygroscopicity curves of AS and OA particles are shown in Fig. 6. The
optical images of the AS particle at the phase change points can be seen in
Fig. 7. The ERH of AS is determined to be 44.3 ± 2.5 % RH, which
generally falls into the range from 33 to 52 % RH reported in the
literature (Tang and Munkelwitz, 1994a; Cziczo et al., 1997; Dougle et
al., 1998; Laskina et al., 2015). The DRH of AS particles is observed to
occur at 80.1 ± 1.5 % RH, which agrees well with reported values of
80 % RH by EDB (Tang and Munkelwitz, 1994a) and 82.3 ± 2.5 % RH by micro-Raman spectroscopy
(Laskina et al., 2015). As shown in Figs. 6b and 8, the measured ERH of OA is 71 ± 2.5 % RH, which
deviates from the reported value of 51.8–56.7 % RH by
Peng et al. (2001) using the EDB technology. It is
worthwhile to point out that the conversion of OA droplets to OA
dihydrate at 71 % RH is inconsistent with the observation of
Peng et al. (2001). They observed that OA droplets
crystallized to form anhydrous OA rather than OA dihydrate
at 51.8–56.7 % RH. The discrepancy on the ERH of OA compared to that
reported by Peng et al. (2001) is likely due to the effects of substrate and
sample purity. The size of dry particles ranging from 10 to 20 µm in
our experiment is consistent with observation using EDB by Peng et al. (2001), which eliminates the influence of particle size. The
substrate supporting droplets may promote the heterogeneous nucleation of
OA while the levitated droplets in the EDB study can avoid induced
nucleation by the substrate. Ghorai et al. (2014) also reported the
potential effects of substrate on the efflorescence transition of
NaCl / dicarboxylic acid mixed particles. In addition, the OA purity in our
study is 99.0 % lower than that of 99.5 % in the study by Peng et al. (2001).
Thus, trace amounts of impurities in OA droplets acting as a
heterogeneous nucleus could contribute to crystallization and result in a
higher ERH of OA. Due to the effects of substrate and sample
purity, the heterogeneous nucleation should be responsible for the
discrepancy on the observed ERH of OA. The water content of the
supersaturated droplet at the onset of crystallization determines the form
of OA crystal generated, i.e., anhydrous OA or OA dihydrate. Due
to a higher ERH, OA droplets with more water content favor the
formation of a dihydrate after crystallization. It should be noted that our
experiment appears to be favored in the atmospheric environment, considering
that insoluble material such as mineral dust mixed with OA may play the role
of substrate thus facilitating the heterogeneous nucleation of OA aerosols.
The Raman growth factor of OA shows no obvious change between
∼71 and 6.6 % RH upon dehydration. At RHs lower than
5 %, the Raman growth factors drop abruptly due to the transformation of
crystalline H2C2O4⚫2H2O into anhydrous OA,
as also indicated by the Raman spectrum. It seems that the structure of
the anhydrous OA particle is not as compact as that of the dihydrate, seen in Fig. 8.
Thus, the loss of crystal water results in no obvious change in particle
size. During the hydration process, the Raman growth factor of OA shows a
slight increase at 19.6 % RH, which can be attributed to the conversion
of anhydrous OA to dihydrate. The transition point of anhydrous
OA to OA dihydrate agrees with previous studies (Braban
et al., 2003; Ma et al., 2013b; Miñambres et al., 2013). No
deliquescence behavior is observed for OA dihydrate even at 94 % RH,
consistent with earlier observations (Ma et al., 2013b; Miñambres
et al., 2013; Jing et al., 2016).
Optical micrographs of the oxalic acid (OA) particle at
(a) 71.5 % RH, (b) 71 % RH, (c) 6.6 % RH and (d) 5 % RH during the
dehydration process.
Hygroscopicity of OA / AS mixtures with OIRs of (a) 1:3, (b) 1:1 and
(c) 3:1 as a function of RH. The red and blue dashed lines indicate the ERH
and DRH, respectively. In panels (a) and (b), Raman growth factors of pure
AS and OA above 80 % RH in the dehydration process are also included for
comparisons. In the panel (c), Raman growth factors of pure AS and OA
dihydrate above 80 % RH during the dehydration process are also given for
comparisons.
Optical micrographs of the mixed oxalic acid / ammonium
sulfate (OA / AS) particle (OIR = 1:3) at phase change points. Dehydration:
(a) 36.1 % RH and (b) 34.4 % RH. Hydration: (c) 79.4 % RH and
(d) 81.1 % RH. In the panel (d), the visual solid in aqueous phase is marked with a red
dashed circle.
Figure 9 presents hygroscopic growth of OA / AS mixtures with OIRs of 1:3,
1:1
and 3:1. As can be seen in Figs. 9a and 10b, mixed OA / AS droplets (OIR = 1:3)
exhibit efflorescence transition at lower (34.4 ± 2.0 %) RH
relative to ERH (44.3 ± 2.5 %) of pure AS. During the hydration
process, mixed particles start to slightly absorb water before deliquescence
at 81.1 ± 1.5 % RH (seen in Figs. 9 and 10). It can be seen in Fig. 10
that the size of the 1:3 mixed OA / AS particle at 79.4 % RH prior to
deliquescence appears to be larger than that after complete efflorescence.
The decrease in ERH and slight water uptake before deliquescence for 1:3
mixed particles is likely due to the effects of NH4HSO4 formed
upon dehydration. NH4HSO4 has a low ERH (22–0.05 %) and DRH
(40 %) (Tang and Munkelwitz, 1994a), which may affect the
nucleation and crystallization of AS upon dehydration and lead to slight
water uptake prior to the deliquescence of AS. The hygroscopic growth of
mixed particles upon dehydration is in fair agreement with that of pure AS
or OA. However, the Raman growth factors of mixed particles upon hydration
show a considerable decrease in comparison to that upon dehydration. The
discrepancies for Raman growth factor at high RH between the two processes
can be attributed to the formation of NH4HC2O4 and residual
solid OA, both of which have a high deliquescence point larger than 95 % RH (Schroeder and Beyer, 2016). During the hydration
process, NH4HC2O4 and OA in the mixed aerosols remains solid
even at high RH (also seen in Fig. 10d), resulting in less water uptake of
mixed particles. A similar phenomenon is also observed for NaCl / OA mixed
particles upon hydration due to the formation of less hygroscopic sodium
oxalate (Peng et al., 2016).
The mixed OA / AS droplets with an OIR = 1:1 first partially effloresce at
75.0 ± 1.6 % due to the crystallization of
NH4HC2O4, as indicated by Raman spectra. Then, the full
efflorescence occurs at 44.3 ± 2.5 % RH with the crystallization of
AS. The full ERH of 1:1 OA / AS mixed droplets is highly consistent with that
of pure AS. During the hydration process, the Raman growth factor of 1:1
mixed particles increases slightly at 35.5 % RH, and then remains almost
invariable until 77 % RH, which is likely due to the formation of hydrate.
The deliquescence transition occurs at 77 ± 1.0 % RH slightly lower
than DRH of AS, which agrees with literature results for AS particles
containing OA (Brooks et al., 2002; Jing et al., 2016). The water contents
of mixed droplets after deliquescence are significantly lower than those
upon dehydration. The Raman features at 494 and 874 cm-1
have confirmed the presence of solid NH4HC2O4 upon hydration
across all RHs studied (seen in Fig. 4), which should be responsible for the
decreasing water uptake of the mixed particles at high RH.
Optical micrographs of the mixed oxalic acid / ammonium sulfate
(OA / AS) particle (OIR = 3:1) at (a) 77.1 % RH, (b) 74.4 % RH, (c) 66.2 % RH
and (d) 64.4 % RH during the dehydration process.
The spatial distribution of chemicals
within mixed oxalic acid / ammonium sulfate (OIR = 3:1) particles
at 74.4 % RH upon dehydration. (a) Raman spectrum acquired on the
surface showing the shell mainly consisting of NH4HC2O4.
(b) Optical micrograph of a partially effloresced droplet composed of oxalic
acid / ammonium sulfate (OIR = 3:1) mixtures at 74.4 % RH upon
dehydration. (c) Raman spectrum obtained at the core of the droplet showing
the liquid phase dominated by OA and AS.
For mixed OA / AS droplets with an OIR = 3:1, the partial and full
efflorescence transition could be observed at 74.4 ± 1.0 and
64.4 ± 3.0 % RH, respectively (seen in Figs. 9 and 11). As seen in
Fig. 3c, the bands at 494, 1471 and 1654 cm-1 suggest the formation of
crystalline NH4HC2O4 at 74.4 ± 1.0 % RH. Figure 12
presents the spatial distribution of chemicals within mixed OA / AS
(OIR = 3:1) particles at 74.4 % RH. The characteristic peak of 980, 1050 and 1471 cm-1 is assigned to SO42-,
HSO4- and HC2O4-, respectively. The sharp
absorption at 874 cm-1 and obvious peak at 1471 cm-1 indicate the
abundant content of NH4HC2O4. The comparison of
characteristic peaks between inner and outer phase reveals that the major
component on the surface of a mixed OA / AS (OIR = 3:1) particle is
NH4HC2O4. In contrast to the surface, the obvious features of
980 and 1050 cm-1 at the core of the particle suggest that
(NH4)2SO4 and NH4HSO4 mainly exist in the inner
aqueous phase. During the dehydration process, crystalline
NH4HC2O4 in the outer phase acts as the heterogeneous
nucleus, leading to the crystallization of OA dihydrate and other
components in the inner phase. Thus, the full ERH of 3:1 OA / AS mixed
droplets is higher than that of pure AS (44.3 ± 2.5 % RH) and
NH4HSO4 (22–0.05 % RH). During the hydration process, Raman
growth factors of mixed particles slightly increase at 34.5 % RH. No
deliquescence transition or significant water uptake is observed over the RH
range studied. This phenomenon can be explained by the fact that most of
the AS in the mixtures has been converted into NH4HC2O4 and
NH4HSO4 or letovicite. Although NH4HSO4 with a low DRH
may contribute to water uptake of mixed particles, the minor
NH4HSO4 or letovicite formed in the mixtures is likely to be
coated by NH4HC2O4 and OA with a high DRH. Thus, the mixed
OA / AS particles with an OIR = 3:1 show no obvious hygroscopic growth upon
hydration due to the change in aerosol composition and morphology effects.
The effects of morphology on the hygroscopic growth of aerosols have been
reported for AS particles containing adipic acid
(Sjogren et al., 2007). The water uptake of AS
particles containing relatively high content of adipic acid could be
suppressed due to AS enclosed by the crust of solid adipic acid with a high
DRH.
The observed ERH for mixed droplets was
dependent on the molar ratio of OA to AS. The mixed
OA / AS droplets with an OIR of 1:3 are observed to effloresce completely at
34.4 ± 2.0 % RH relative to ERH of pure AS (44.3 ± 2.5 %) or
OA (71 ± 2.5 %). It can be seen that AS as a major fraction of the
particle does not promote the heterogeneous nucleation of OA. Meanwhile, the
crystallization of AS is also influenced due to the presence of OA. A similar
phenomenon was also observed for malonic acid / AS
mixtures with minor organic content (Braban and Abbatt, 2004; Parsons et
al., 2004). Braban and Abbatt (2004) found that the ERH of malonic
acid / AS mixed particles was considerably decreased compared to
that of pure AS for mass fractions of malonic acid less than
0.3. They concluded that the presence of AS in the
supersaturated droplet could exert the extra barrier to nucleation of
malonic acid crystals rather than play the role of a heterogeneous
nucleation site. As for 1:3 OA / AS mixed droplets, AS may also
inhibit the nucleation of OA at relatively high RH. With decreasing
RH, aqueous OA could enhance the viscosity of the droplet due to
hydrogen bond interactions (Mikhailov et al., 2009), thus limiting the
nucleation of AS and resulting in a lower ERH with respect to
the value of pure AS (Parsons et al., 2004). In the case of mixed OA / AS
droplets with an OIR of 1:1 and 3:1, the NH4HC2O4 formed at
∼75 % RH upon dehydration likely acts as a heterogeneous
nucleus for crystallization of other components, which increases the full
efflorescence point of mixed particles. One study indicated that
humic acid sodium salt (NaHA, Aldrich) could also promote the ERH of AS
(Badger et al., 2006). Similar to OA, succinic acid and adipic acid
have a high deliquescence point and low solubility. However, it has been
found that the efflorescence point of AS in mixed particles is
not elevated even when the content of succinic acid or adipic acid is more
than 50 % by mass or mole fraction (Ling and Chan, 2008; Yeung et al.,
2009; Laskina et al., 2015). The chemical nature of a solid determines its
ability to act as a heterogeneous nucleus (Braban and Abbatt, 2004). In
contrast to AS particles containing succinic acid or adipic
acid, our results suggest that the addition of OA into AS droplets may trigger partial and full crystallization of aerosols at
relatively higher RH upon dehydration due to the NH4HC2O4 product
acting as an effective nucleus.
During the deliquescence process, the OA / AS mixed particles with an OIR of
1:3 and 1:1 exhibit a slightly lower deliquescence point than that of pure
AS, consistent with previous observations of effects of
crystalline OA on deliquescence transition of AS
(Brooks et al., 2002; Wise et al., 2003; Jing et al., 2016). It should be
noted that prior literature results also showed that continuous or smooth
water uptake from low RH was observed for particles composed of AS and OA
with a mass ratio of 1.5 : 1 due to the fact that after drying
OA existed in an amorphous or liquid-like state that prevented
nucleation of AS even under dry conditions (Prenni et al.,
2003). In the present study, water uptake by the OA / AS mixed particles at
high RH upon hydration is dramatically lower than that upon dehydration and
significantly decreased with elevated OA content. This phenomenon
distinguishes itself from hygroscopic characteristics of typical water-soluble
mixtures in the literature. It has been found that hydration growth curves and
dehydration growth curves are typically merged above the deliquescence point for
mixed systems containing inorganic salts and water-soluble organic compounds
(Choi and Chan, 2002; Chan and Chan, 2003; Gysel et al., 2004; Clegg and
Seinfeld, 2006; Sjogren et al., 2007; Pope et al., 2010; Ghorai et al.,
2014; Estillore et al., 2016). In this study, Raman spectra and the
micrographs suggest the presence of solid NH4HC2O4 and
residual solid OA at high RH should be responsible for the decreased water
uptake during the hydration process. In contrast, Prenni et al. (2003)
reported that the hygroscopic growth of OA / AS mixed particles remained
unchanged at 90 % RH with OA mass fraction ranging from 0.01 to 0.4. In
addition, they also found that water uptake after deliquescence was well
described by the model method assuming complete dissolution of OA in aqueous
phase as well as no interactions between OA and AS, which was also observed
by Jing et al. (2016) using the HTDMA. The previous HTDMA studies for OA / AS
mixed particles indicate no composition change and no specific interactions
existing between OA and AS (Prenni et al., 2003; Jing et al., 2016).
However, it should be noted that the HTDMA studies did not perform
measurements for the dehydration process such that aerosols underwent rapid
drying on the time scale of seconds, i.e., the total residence time for
transformation of droplets into dry particles in the drying section of the HTDMA
is typically tens of seconds (Prenni et al., 2003; Jing et al., 2016), much
shorter than (10–12 h) in our study. In the HTDMA
experiments, the combination of faster drying and smaller particles with
submicron size implies that the aqueous phase obtained higher
supersaturations than in our present study (Rosenoern et al., 2008), leading
to less dissociation of OA and thus less HC2O4-
formed in the droplets as well as the inhibited formation of
NH4HC2O4. The fast evaporation of water from the surface of
an aqueous droplet upon rapid drying could result in a higher surface
concentration of solutes than the slow drying process (Treuel et al., 2011).
The higher surface concentration of OA corresponds to less
formation and hence decreased supersaturation of HC2O4-. Due
to the dependence of the nucleation rate on the extent of supersaturation, it
can be expected that the nucleation of NH4HC2O4 is suppressed
within OA / AS mixed droplets undergoing rapid drying.
(a) Raman spectra of equal molar mixed OA / AS particles after
the rapid drying process at various RH values upon hydration. The Raman spectrum
(black short dash) at 2.5 % RH obtained from the slow drying process is
also given for comparison. (b) Deliquescence curve of OA / AS mixtures with
an OIR of 1:1. The hygroscopic curve (olive symbols) of particles after the slow
drying process is also included for comparison. The blue dashed lines
indicate the DRH.
Considering the potential effects of drying time on the reactions between OA
and AS, we explored the hygroscopicity of OA / AS particles with an OIR of
1:1 after the rapid drying process. The mixed OA / AS droplets undergo dehydration to
form dry particles in 3–5 min. We observed one-step
efflorescence of rapidly dried particles (1:1, molar ratio) occurring at
47 ± 2.5 % RH, compared to the two-step efflorescence of
slowly dried particles occurring at 75 and 44.3 % RH.
The Raman spectra and hygroscopic curve upon hydration for OA / AS particles
with an OIR of 1:1 are presented in Fig. 13. The obvious discrepancies can
be observed for spectra at ∼2 % RH between the two drying
processes. After the rapid drying process, the spectra at ∼2 % RH
show the feature of crystalline AS (974 cm-1: νs(SO42-)) and anhydrous OA (1710 cm-1: ν(C=O), 1479 cm-1: νs(COO)). Meanwhile, no characteristic peaks for
NH4HC2O4 (494 cm-1: δ(COO), 874 cm-1: ν(C-C), 1729 cm-1: ν(C=O), 1469 cm-1: νs(COO))
and NH4HSO4 (874 cm-1: δ(S-OH)) can be identified in
the spectra. It is clear that the drying time for transformation of droplets
into dry particles has an impact on the reactions of OA with AS in the
aerosols due to particle-phase processes under kinetic control. Previous
studies found the longer drying time could lead to greater nitrate depletion
between nitrates and organic acids, which results from slow reaction and
diffusion in the viscous aerosols (Wang and Laskin, 2014).
The Raman growth factors of mixed particles with an OIR of 1:1 also increase
slightly at 36.5 % RH due to the formation of OA dihydrate, as indicated by
the Raman feature. The deliquescence transition of mixed particles occurs at
79.3 % RH. After deliquescence, Raman growth factors of mixed particles
after the rapid drying process are lower than that after the slow drying process,
which may be caused by the fact that at high RH the hygroscopic growth of AS
is slightly lower than that of NH4HSO4 formed in the particles
after the slow drying process (Tang and Munkelwitz, 1977). In
addition, it is found that after deliquescence OA dihydrate remains solid in
the mixed particles undergoing rapid drying.