Atmospheric Chemistry and Physics Discussions

General comments The paper addresses the inter-annual trends and variability of stratospheric mixing, covering the tropics through the mid-latitude to the high latitudes, emphasizing the differences between the northern and southern hemisphere where necessary. The technique used is the finite-time Lyapunov exponent. Linear trends of stratospheric mixing and correlation with the QBO signal are significant findings. The overall quality of the paper is high and the discussions of the results are pertinent and illuminating.


Introduction
The phase of atmospheric aerosol particles is important when trying to determine the radiative balance of the atmosphere (IPCC, 2001) as well as for chemical reactions occurring heterogeneously in the atmosphere (e.g.Ravishankara, 1997).For example, N 2 O 5 hydrolysis occurs heterogeneously on aerosol surfaces.This reaction has been shown to be a significant nighttime NO x sink and proceeds much more slowly on dry solid particles than on aqueous solution droplets, even if highly supersaturated (Mozurekewich and Calvert, 1988;Hu and Abbatt, 1997;Kane et al., 2001;Thornton et al., 2003).Hence, it is of great importance to understand phase transitions of atmospherically relevant aerosol under atmospheric conditions.Ammonium sulfate (AS) is frequently used as a model tropospheric aerosol due to the widespread prevalence of ammonium and sulfate species in the troposphere.However, it has become evident that atmospheric aerosol is chemically more complex, with both organic and mixed inorganic-organic aerosols widely present in the troposphere.Several initial studies have focused on the properties of water-soluble organic compounds (WSOC) that are found in the atmosphere.Dicarboxylic acids, which have been identified globally in atmospheric aerosol (Yao et al., 2002;Rohrl and Lammel, 2001;Narukawa et al., 1999;Narukawa et al., 2002), have been considered by several recent phase transition studies (Prenni et al., 2001;Parsons et al., 2004;Braban et al., 2003;Brooks et al., 2002).
Given the chemically mixed nature of aerosol, it is important to understand and quantify the effect of organic chemicals on the phase transition behavior of model inorganic tropospheric aerosols.To date several studies have considered mixtures of dicarboxylic acids and inorganic salts with respect to phase transitions or water uptake.Brooks et al. (2002) completed a study of the deliquescence relative humidities of ternary bulk solutions.Choi and Chan (2002) studied the water cycles of dicarboxylic acids with both sodium chloride and AS using an electrodynamic balance (EDB) technique.Three tandem differential mobility analyzer (DMA) studies have been performed.Cruz and Pandis (2000) studied mixtures of glutaric acid and AS.Prenni et al. (2003) studied the water uptake of binary AS-dicarboxylic acid mixtures at 3 compositions and Hameri et al. (2002) studied AS with several dicarboxylic acids.Using EDB, Lightstone et al. (2000) focused on the potential heterogeneous effect of succinic acid (SA) cores on the efflorescence RH (ERH) of mixed ammonium nitrate -SA particles.
Though several studies have been performed, this work constitutes one of the detailed case studies of a three-component aerosol.The first was of the AS-maleic acid aerosol system (Brooks et al., 2002(Brooks et al., , 2003)).Analysis of the results presented here and comparison with the literature noted above are used to more clearly understand the possible effects of WSOC on the phase transitions of atmospheric aerosol.In particular we focus on the effect of each species in solution on the phase transitions on the other.In our previous study (Braban et al., 2003) both MA and oxalic acid were observed to be stable as highly supersaturated solutions, therefore one immediate question was whether the presence of AS would inhibit or promote the crystallization 2951 of dicarboxylic acid particles.Conversely, would MA suppress the crystallization of AS?This is the first study to consider these processes in terms of the phase transitions of the individual components lending insight into the physical processes occurring and their relevance for atmospheric aerosol.

Experimental
The Aerosol Flow Tube -Fourier Transform-Infrared spectrometer system (AFT) used in this work has been previously described (Braban et al., 2003).Aerosol is produced using a commercial atomizer (TSI Model 3076).A fraction of the aerosol flow is discarded while the remainder either passes through a silica gel dryer (TSI 3062) or bypasses the dryer for the deliquescence and efflorescence experimental modes, respectively.The aerosol flow is mixed with a nitrogen flow, a variable fraction of which is saturated with water vapor.In most experiments the total N 2 flow was 2 l.min −1 though in some of the efflorescence mode experiments an additional 1 l.min −1 was added to study the low RH range in more detail.The total flow (2.3-3.3 l.min −1 ) then passes through a precooler-mixing volume before entering the main aerosol flow tube.
Experiments are carried out by incrementally increasing or decreasing the RH for deliquescence-mode or efflorescence-mode experiments, respectively.Both the flow tube and the pre-cooler/mixer are temperature jacketed and controlled using a circulator (Julabo FP88-MW).The temperature in the flow tube is measured using two T-type thermocouples.The thermal gradient over the observation region is less than 0.2 K in all the experiments performed and generally less than 0.1 K.
A length of 60 cm in the flow tube is monitored with the FTIR spectrometer (Bomem MB104).IR spectra are recorded over the wavenumber range 4000-500 cm −1 with a resolution of 4 cm −1 .The RH in the flow tube is measured using a gas-phase water line calibration, integrating the area under 1710 cm −1 -1690 cm −1 (Braban et al., 2003).
The calibration was carried out in two ways: (i) by mixing of known flows of dry and water-saturated N 2 at constant temperature; (ii) calibration using ice-saturated condi-tions between 213 K and 270 K.In general, the confidence level of reported relative humidities is ±1%.99.9% assay AS (Fisher) and 99% purity MA (Sigma-Aldrich Chemical Co.) were used to make the solutions.The percent dry mass composition of the solution was known.Compositions are described herein as the dry mass fraction of MA, f MA , i.e. f MA =0.1 represents 90% AS/10% MA.Infrared (IR) spectra were recorded as an average of 50 scans.From the IR spectrum the RH is measured using the gas phase water line calibration, as described above.The condensed-phase features of the aerosol IR spectrum are used to determine the phase state of the aerosol, and in particular the point at which phase transitions of the components occur.

Results
3.1 Ammonium sulfate + malonic acid infrared spectra In this study 12 solutions containing compositions between 100% AS and 100% MA were studied.In Figs. 1 and 2, IR spectra for 5 compositions (A-E) of AS and MA are presented at RH<1% and RH>80%, respectively.Gas phase water lines have been subtracted for clarity from the high RH spectra.Compositions A-E shown in each figure are f MA =0.10, 0.25, 0.49, 0.75 and 0.90.The AS component of the aerosol will be considered first.In Figs. 1 and 2, spectrum A, absorption features are seen between 2800-3300 cm −1 , 1420-1450 cm −1 , 1115 cm −1 and 620 cm −1 corresponding to N-H stretch modes, NH + 4 deformation mode and two sulfate bands, respectively (Cziczo and Abbatt, 1999).These features are present as the fraction of MA is increased until f MA =0.90, where the AS features are not discernable.The NH + 4 mode in Fig. 1A present at 1420 cm −1 is indicative of AS in the crystalline phase.The corresponding mode in Fig. 2A at 1455 cm −1 is indicative of the salt being in solution.
As the fraction of MA is increased changes are observed in the AS modes in the IR spectrum, specifically under the driest conditions (i.e. the spectra shown in Fig. 1, 2953 RH<1%) the AS modes are different from when f MA >0.25.In fact the changes occur when f MA >0.20 (spectrum not shown).When f MA >0.20, the AS modes are broader and the NH + 4 mode at 1420 cm −1 is shifted to higher wavenumbers (seen in Fig. 1C and D) and there is also a side peak evident (∼1050 cm −1 ) on the broadened mode at 1105 cm −1 .This is thought to be indicative of the AS component or a significant fraction of it, being aqueous rather than crystalline, i.e. at RH<1% the AS component does not fully crystallize out.This will be more fully discussed in Sect.3.2.In Fig. 2, the AS is clearly in the aqueous phase and the AS modes are broadened and decreased in intensity, with the AS mode shifted to ∼1450 cm −1 .Condensed phase H 2 O features can be seen between 2500-3500 cm −1 , ∼1600 cm −1 and <1000 cm −1 , being the OH stretch, the H-O-H bending mode and the H-bonding librations, respectively.In considering the MA component of the aerosol, in Fig. 1A, the carbonyl stretch at ∼1720 cm −1 due to the MA is just observable.As the fraction of MA is increased, this mode increases in intensity and other IR features due to MA become apparent in the IR spectrum between 2000 and 1000 cm −1 .In our previous aerosol studies, the carbonyl resolves into the two peaks at 1749 cm −1 and 1686 cm −1 when the MA is dry and crystalline (Braban et al., 2003).Also, features between 1500 and 1000 cm −1 are sharp and well resolved, and the modes at 912 and 630 cm −1 are present.The MA does not appear to be fully crystalline from the spectra at RH<1%, except at f MA =0.90 (Fig. 1E).
It is thought that this is not due to the low intensity of the MA features in the spectra because in other experiments, not presented here, the two peaks of the carbonyl and the lower wavenumber modes were clearly resolvable if the MA was effloresced even at very low extinctions.
In a solution containing AS and MA it is possible other salts including ammonium bisulfate, letovicite and ammonium hydrogen malonate could form if partial dissociation of aqueous MA occurs (Note that in a 6 M aqueous solution of MA, 1.5% of the MA is in the form of the singly de-protonated malonate ion MA − ; this fraction increases as the solution becomes more dilute.In 0.1 M, 10% of the MA is dissociated).The side-peak at 1050 cm −1 in Fig. 1C and D is at the same position as a bisulfate mode (Cziczo and Abbatt, 2000), however the most intense peak indicative of bisulfate at 1300 cm −1 is notably absent.It is thought that this shoulder is a broadening of the AS mode due to dissolution alone.There is also no evidence of a mode at 1550 cm −1 seen in ammonium malonate and bimalonate IR spectra (Charbonnier and Arnaud,  1973; Braban and Abbatt, in preparation, 2004 1 ).Therefore, IR observations imply that ammonium bisulfate, letovicite or ammoniated malonate species if present, are at very low levels below our detection limit of a few weight percent.

Analysis of phase transitions
For each composition, both an efflorescence mode and a deliquescence mode experiment were performed at 283 K and 293 K, respectively.In binary solutions, if a particle is dry, deliquescence is generally observed at the thermodynamically predicted deliquescence RH.Similarly for a ternary solution the lowest RH at which the solution is in equilibrium with solid components A and B (in this case AS and MA) can be calculated and measured.This invariant point of the ternary system is called the eutonic deliquescence RH (EDRH).Brooks et al. (2002) measured the EDRH for the AS-MA ternary system with bulk solutions as a function of temperature.Deliquescence mode experiments presented here were performed at 283 K to differentiate between the binary deliquescence processes, which occur at 283 K at RH∼81%, and the predicted EDRH at RH=73%.
To understand the phase of both the AS and MA components of the aerosol particles, the IR spectra were analyzed as a function of RH.In aerosol IR spectra there are 2 main types of features that may be analyzed to understand the phase of that component: the presence of condensed phase H 2 O and the IR features of the salt/acid.In previous studies the integrated area due to the OH stretch of condensed phase water has been used for this purpose.For example, in our previous study (Braban et al., 2003) the area under 3500-3000 cm −1 ), after the subtraction of the lowest RH spectrum, was used to analyze for the presence of condensed phase water.To account for differences in the aerosol populations used in different experiments the integrated area A H 2 O was normalized to the integrated area at RH=83%, at which point all aerosols were aqueous.This process did not alter any of the major features, but allowed greater clarity in observing the differences in the experimental results.Brooks et al. (2003) used a ratio method to study phase transitions in the AS-maleic acid system, using both the condensed phase water and acid modes.In the present work, in addition to the condensed phase water, we have analyzed the phase transitions of AS and MA IR features.The AS mode at 1420/1455 cm −1 shifts upon dissolution as discussed in Sect.Brooks et al., 2002).To quantify the phase transitions of MA, it can be seen that the mode at 665 cm −1 changes intensity and the broad mode at 918 cm −1 decreases greatly in intensity when MA becomes aqueous.Both of these features were used to analyze the phase of the MA.The results from the latter analysis for deliquescence mode experiments are shown in Fig. 3C.
The results presented in Figs.3-5 are discussed below using composition classification based on the observed behavior.In particular, results have been color coded with respect to the phase transition behavior of AS.In Fig. 3B, the blue double-ended arrow indicated the ±4 cm −1 resolution used in these experiments.The results and observations are summarized in Table 1.
3.2.1 Results: 0≤f MA <0.10 (pink and red symbols) Pure AS aerosol (Fig. 3, pink ), take up very little H 2 O, A H 2 O ∼0, as RH increases until the DRH is reached, measured here as RH=81±2%.This value is in agreement with previously measured values (Braban et al., 2001;Cziczo and Abbatt, 1999;Xu et al., 1998;Onasch et al., 1999).Similarly, in the efflorescence experiment (Fig. 4, pink ), A H 2 O decreases as RH is lowered.At RH=33±2%, A H 2 O decreases to a small and approximately constant value when efflorescence occurs.This RH is also consistent with other studies (Cziczo and Abbatt, 1999).Deliquescence and efflorescence are also seen clearly in the shift of the NH + 4 mode in Figs.3B and 4B.In deliquescence mode when there is a very minor fraction of MA present in the aerosol, f MA =0.01 and 0.06, there is an increase in A H 2 O with RH even when the RH<DRH AS , though the amount of condensed phase water is small.From the analysis of the ammonium mode, the AS component is crystalline at all RH<DRH AS and at DRH AS the ammonium mode shifts to ∼1450 cm −1 , indication deliquescence of the ammonium sulfate.It is not possible to analyze the MA mode at 912 cm −1 due to the minor fractions of MA.Qualitatively, from the observation of the carbonyl mode discussed in Sect.3.1 and the observation of a small increase of condensed phase water as a function of RH we believe that the MA does not appear to be crystalline at RH<DRH AS .However we also note that this increase may be due to the presence of adsorbed water at higher RHs, given that a small amount of water uptake is observed before deliquescence in the single component AS aerosol.These results will be further discussed in Sect. 4. In efflorescence experiments aerosol crystallized at the ERH of pure AS, to within the error of the experiments.
3.2.2Results: 0.10≤f MA ≤ 0.25 (orange symbols) The deliquescence-mode experiments for the composition range 0.10≤f MA ≤0.25 are presented as orange symbols in Fig. 3. Water uptake occurs over the complete RH range.However the major increase in A H 2 O occurs close to the EDRH for the AS-MA 2957 system at this temperature (Brooks et al., 2002).This observation is consistent with the shift in the NH + 4 mode which also occurs at approximately 73% RH for both the f MA =0.10 and 0.25.There is insufficient MA to analyze the phase of the MA quantitatively, however there is water uptake at all relative humidities and, qualitatively, from the IR spectra the MA does not appear to be fully dry.Thus, it is thought that the "dry" aerosol (RH<1%) used in the deliquescence mode experiments consists of solid AS with some supersaturated solution of MA.
In efflorescence mode experiments, it can be seen from both the A H 2 O and the NH + 4 plot (Fig. 4A and B) that the crystallization of AS in these aerosols is significantly lowered compared to that of pure AS.This process is not full efflorescence as there is still condensed phase water associated with the aerosol, therefore the term crystallization is used to describe the phase transition of the AS component of the aerosol.
3.2.3Results: 0.25<f MA ≤0.90 (green symbols) For compositions 0.25<f MA ≤0.90 water uptake occurs at all relative humidities in deliquescence-mode experiments and there is no hysteresis in the A H 2 O plots between the efflorescence and deliquescence mode experiments (Figs.3A and 4A).This is consistent with the IR spectra in Fig. 1B, C and D being similar to the equivalent spectra in Fig. 2 and also with neither component of the aerosol having crystallized.This observation is confirmed by virtually no abrupt shift in the NH + 4 mode (Fig. 3B) for compositions f MA =0.40 and 0.50.It is thought that this is not due to any masking of the shift by the presence of the MA mode at 1444 cm −1 , as one would still see a significant shift at the DRH or EDRH particularly when AS is the major fraction.For these compositions however it is noted that there is a gradual shift in this mode to slightly higher wavenumbers as RH increases.Possible reasons why this might occur will be discussed in Sect. 4. It could be argued that f MA =0.75 has a considerable shift in wavenumber at the EDRH, however this shift initiates at about 50% RH, quite unlike the behavior of particles with a lower f MA (e.g.f MA =0.25).
3.2.4Compositions: 0.90<f MA ≤1.00 (blue and purple symbols) Deliquescence mode experiments for both f MA =0.92 and 0.98 show a small amount of water uptake prior to deliquescence at the EDRH.Although there is too little AS present to study the position of the NH + 4 mode, in Fig. 3C the MA mode at 923 cm −1 is plotted as a function of RH.In deliquescence-mode experiments, after the aerosol is exposed to the silica dryer and mixed with 2 l.min −1 N 2 , the initial RH is less than 1%.
The MA appears crystalline until the major water uptake and deliquescence occurs at the EDRH, however there still seems to be some water uptake at RH<EDRH.Also, there is some loss of peak area before the water is seen in the condensed phase in the pure MA aerosol.
The similar analyses of the MA mode described above and shown in Fig. 3C were performed for the efflorescence mode experiments presented in Fig. 4.Only in the case of pure MA was efflorescence observed at 6% RH (Braban et al., 2003).For clarity the data are not shown in a third panel in Fig. 4. However to emphasize, no changes in the IR were observed in any of the MA modes during efflorescence mode experiments.The minimum RH attainable in efflorescence-mode is approximately 3%, thus this implies that MA efflorescence has been lowered to below this value.
In summary, for compositions f MA =0.92 and 0.98, aerosols were observed to deliquesce at the EDRH, with some H 2 O uptake prior to it.Efflorescence was not quantitatively observed for either component in these experiments as it was not possible to observe the AS modes.In efflorescence experiments MA clearly did not crystallize implying the addition of even low levels of AS suppresses the efflorescence of MA to <3% RH. 2959 4 Discussion and atmospheric implications

Summary of results
Experiments have been performed with the AFT to identify phase transitions of aerosol composed of ammonium sulfate (AS) and malonic acid (MA).When f MA <0.1, phase transitions occurred at the DRH AS and ERH AS .However between 0.1<f MA <0.25 deliquescence occurred at the eutonic DRH for the AS-MA system and the crystallization RH of AS was lowered.Between 0.25<f MA <0.90, no distinct deliquescence or efflorescence was observed, with particles taking up water at all relative humidities.Thus the crystallization of both MA and AS is suppressed by the presence of the other component.The crystallization RH of MA was <3% when any AS was present.When f MA >0.25, AS did not crystallize even when RH <1%.Thus, the presence of either second component increased the range of conditions under which the particles were in the solution phase.In particular it only requires a very low level of AS or MA to shift deliquescence to the EDRH, and if there is a significant AS and MA fraction, water uptake occurs at all relative humidities.This behavior, should it be widely applicable, will have a large impact on the phase of chemically mixed aerosol in the atmosphere.
The results for the AS phase transitions (as distinct from total aerosol particle phase transitions) are summarized in Fig. 5.To follow the processes occurring in the aerosol, in Fig. 5 the RH at which the NH + 4 mode is observed to start to shift as a function of RH has been plotted for the different aerosol compositions.That MA suppresses the crystallization of the AS is clearly seen, implying that in the atmosphere this would allow the particles to remain as a metastable supersaturated solution to low relative humidities.The lowering of the deliquescence RH of AS, initially to the eutonic RH, is also clear.We have included in Fig. 5 the points at which the NH + 4 mode begins to shift in deliquescence mode experiments where it is thought that full efflorescence of the AS has not occurred.Two hypotheses are suggested as to why the NH + 4 mode should shift in such a way.First, it is possible a fraction of the AS has crystallized (as the aerosol is taken to a lower RH in deliquescence mode experiments compared to the efflorescence mode experiments).If so, the shift may be due to a solubility of AS in supersaturated MA solution.The second hypothesis is that the AS/MA supersaturated solution chemical environment changes as a function of RH, such that the NH + 4 mode shifts.At this stage it is not possible to verify either hypothesis.

Comparison with other literature and atmospheric implications
AS-MA mixtures have been studied by three other groups (Hameri et al., 2002;Choi and Chan, 2002;Prenni et al., 2003).In the first two of the studies a 1:1 mixture (by mass and by mole, respectively) was studied and water uptake at all relative humidities was observed, which is supported by the results presented herein.Of these, Hameri et al. (2002) noted that the deliquescence of AS was not visible indicating that the deliquescence properties of the AS-MA particles are completely different to that of pure AS.Choi and Chan (2002), whose experiments were performed with an EDB, reported a crystallization RH of 45.9-46.9% and a deliquescence RH range of 57.8-73.7%(at 298 K), noting that the particles absorb a significant amount of water before the DRH.However, it is not clear that the entire particle crystallizes at the reported ERH, as water is still present below RH=45% even to very low relative humidities.If the observations of this study are taken into account, it seems likely that the particle may have partially crystallized.Differences in efflorescence relative humidities between EDB and AFT experiments have been noted, as the particles in EBD studies are significantly larger than those in tandem DMA and AFT experiments.EDB results can give a higher ERH than AFT experiments, which can possibly be attributed to the larger size of the particles, the volume based probability of a crystal embryo and longer residence time allowing mass transfer limited processes to occur.However, in order to systematically compare the results from the different experimental techniques it has been suggested that it would be relevant to calculate the relative nucleation rates and the supersaturation of the solutions at crystallization (A. Bertram, personnal communication).Prenni et al. (2003) studied the deliquescence and hygroscopic growth of ASdicarboxylic acid ternary mixtures.For the AS-MA system at 100:1 and 10:1 ratios 2961 they saw no effect on the AS phase transition.This is to be expected as at the temperature of their experiment, 303 K, the EDRH≈DRH AS .At 2.3:1 AS:MA continuous water absorption at all relative humidities was observed, which is in agreement with this study.As mentioned previously, Brooks et al. (2003) investigated the AS-maleic acid ternary aerosol system in a study analogous to this.Similar to MA, maleic acid is highly soluble.They report that maleic acid aerosols start taking up water at RH=20% and continue taking up water until the DRH at 89% at 273 K, at which water uptake accelerates.The internally mixed AS-maleic acid particles were observed to deliquesce at a lower RH than either of the single component systems, which is consistent with the results of this study.Brooks et al. (2003) attribute the uptake of water below the deliquescence point to water associated with the surface.However, this may also be due to partial crystallization of the ammonium sulfate leaving the maleic acid in a supersaturated solution, which would be more consistent with the results we report here.
In efflorescence studies they observed maleic acid ERH of 17% at 273 K and mixed aerosols effloresced at 27±3% RH, which implied the maleic acid was heterogeneously crystallized by the AS, in contrast to the suppression of the ERH that we observe.Both of our studies conclude that the presence of water-soluble organics internally mixed with AS aerosol could increase the range of conditions under which aerosol is in a solution phase.
The behaviour observed in this study may have significant atmospheric relevance.It is clear from this work and other studies that MA can inhibit the crystallization of AS.Other chemical systems have been observed to behave similarly (Prenni et al., 2003).It can be seen that the AS also does not promote the heterogeneous nucleation of MA, even when present as a major fraction of the aerosol.This behavior is in contrast to several studies that have investigated heterogeneous nucleation in aerosol particles, including Lightstone et al. (2000) and Martin et al. (2001).If sufficiently large nuclei of a solid are present, that solid has the potential to act as a heterogeneous nucleus.However, whether the process will occur frequently is dependant on the chemical nature of the system and the size of the inclusion.Lightstone et al. (2000) studied deliquescence, efflorescence and water activity in mixed ammonium nitrate -succinic acid particles in which both the presence of solid inclusions of succinic acid and fully dissolved succinic acid were investigated.Ammonium nitrate particles do not effloresce down to ∼0% RH whereas Lightstone et al. found that the presence of an SA core increased the ERH of AN.Similarly, it has been shown in studies of particles with inorganic "refractory" inclusions that efflorescence may be catalyzed, i.e. heterogeneous nucleation occurs (Han and Martin, 1999;Martin et al., 2001).In the experiments presented here AS does not act as a heterogeneous crystallization site for the MA, when the AS crystallizes.This is possibly due to the barrier to nucleation of MA crystals or the presence of residual AS in the supersaturated solution fraction of the droplet adds an extra barrier to the formation of critical embryonic crystals of MA.
In summary, this study has shown that in order to gain understanding of phase transitions in mixed phase systems it is necessary to identify the individual species that crystallize and those that remain in solution, as it is somewhat ambiguous to refer to deliquescence and efflorescence.The AS-MA mixed system shows a range of behaviors, as summarized in Table 1, most of which include significant water uptake at all atmospheric relative humidities and lack of complete crystallization.Therefore, in the absence of heterogeneous nuclei, it is likely the particles would exist as either saturated or supersaturated droplets under most conditions.+ 4 mode: Ammonium sulfate observed to go into solution, 283 K (♦); Crystallization of ammonium sulfate, 293 K (black filled ).Eutonic composition and EDRH for ammonium sulfate -malonic acid -H 2 O ternary system at 283 K is shown with the dashed line (Brooks et al., 2002).See text for details.
3.1.The results from the condensed phase water and AS mode analyses of deliquescence and efflorescence mode experiments are presented in Figs. 3 and 4, respectively.For clarity not all compositions are shown in every figure.In Figs.3A and 4A, the normalized A H 2 O is plotted as a function of RH.In Figs.3B and 4B, the position of the NH + 4 mode is plotted as a function of RH.The vertical arrows in Fig. 3 represent the DRH of MA (DRH MA , Arrow 1), AS (DRH AS , Arrow 2) and the EDRH for an AS-MA solution (Arrow 3;

Fig. 5 .
Fig. 5. Phase transition points of ammonium sulfate in ammonium sulfate-malonic acid mixtures as determined by shift in NH