Investigation of water adsorption and hygroscopicity of atmospheric 1 particles using a commercial vapor sorption analyzer 2 3

Water adsorption and hygroscopicity are among the most important physicochemical properties of aerosol particles, largely determining their impacts on atmospheric chemistry, radiative forcing, and climate. Measurements of water adsorption and hygroscopicity of nonspherical particles under subsaturation conditions are non-trivial because many widely used techniques require the assumption of particle sphericity. In this work we describe a method to directly quantify water adsorption and mass hygroscopic growth of atmospheric particles for temperature in the range of 5–30 °C, using a commercial vapor sorption analyzer. A detailed description of instrumental configuration and experimental procedures, including relative humidity (RH) calibration, are provided first. It is then demonstrated that for (NH 4 ) 2 SO 4 and NaCl, deliquescence relative humidities (DRHs) and mass hygroscopic growth factors measured using this method show good agreements with experimental and/or theoretical data from literature. To illustrate its ability to measure water uptake by particles with low hygroscopicity, we used this instrument to investigate water adsorption by CaSO 4  ∙ 2H 2 O as a function of RH at 25 °C. The mass hygroscopic growth factor of CaSO 4  ∙ 2H 2 O at 95 % RH, relative to that under dry conditions (RH < 1 %), was determined to be (0.450 ± 0.004) % (1 σ). In addition, it is shown that this instrument can reliably measure a relative mass change of 0.025 %. Overall, we have demonstrated that this commercial instrument provides a simple, sensitive, and robust method to investigate water adsorption and hygroscopicity of atmospheric particles.


Introduction
Atmospheric aerosol particles, directly emitted by natural and anthropogenic processes or secondarily formed in the atmosphere, have significant impacts on air quality, visibility, human health, and radiative and energy balance of the Earth system (Pöschl, 2005;Seinfeld and Pandis, 2006).The ability to uptake water is among the most important physicochemical properties of aerosol particles, and it largely determines their impacts on atmospheric chemistry and climate (Martin, 2000;Rubasinghege and Grassian, 2013;Farmer et al., 2015;Tang et al., 2016).The ability of aerosol particles to uptake water depends on particle composition, relative humidity (RH), and temperature (Martin, 2000;Tang et al., 2016).Under subsaturation conditions (RH <100%), the ability to uptake water is typically called water adsorption in surface science and hygroscopicity in aerosol science (Tang et al., 2016).Under supersaturation conditions, aerosol particles can be activated to cloud droplets (McFiggans et al., 2006;Petters and Kreidenweis, 2007) and ice particles if temperature is below 0 o C (Pruppacher and Klett, 1994;Vali et al., 2015).Hygroscopicity of atmospheric particles has been extensively investigated by a large number of studies, and many experimental techniques have been developed.These techniques have been summarized and discussed by a very recent review paper (Tang et al., 2016), and here we only mention widely used ones.For airborne monodisperse particles typically produced by a differential mobility analyzer (DMA), the hygroscopicity can be determined by measuring their diameters at dry (typically at RH <15% or lower) and humidified conditions (Swietlicki et al., 2008;Freedman et al., 2009;Robinson et al., 2013;Lei et al., 2014).Typically, the diameter change is determined by using a scanning particle mobility sizer (in which mobility diameters are measured) (Vlasenko et al., 2005;Swietlicki et al., 2008;Herich et al., 2009;Koehler et al., 2009;Wu et al., 2011) or aerosol extinction-cavity ring down spectrometry (in which optical diameters are measured) (Freedman et al., 2009;Attwood andAtmos. Meas. Tech. Discuss., doi:10.5194/amt-2017-56, 2017 Manuscript under review for journal Atmos.Meas.Tech.Discussion started: 6 April 2017 c Author(s) 2017.CC-BY 3.0 License., 2011).These techniques require an underlying assumption that particles are spherical.However, a few important types of particles in the troposphere, including mineral dust and soot, are known to be non-spherical (Veghte and Freedman, 2012;Ardon-Dryer et al., 2015).Therefore, although these techniques can provide useful information, it is difficult to quantitatively determine the amounts of water associated with non-spherical particles at a given RH (Tang et al., 2016).Single particle levitation techniques, which measure light scattering intensity to determine the size (and thus the hygroscopic growth) of levitated particles, also have similar drawbacks (Krieger et al., 2012).

Greenslade
There are several techniques which can be applied to quantify the amount of water associated with non-spherical particles at given temperature and RH.For example, adsorbed water can be measured by FTIR by its IR absorption at around 3400 and 1645 cm -1 (Goodman et al., 2001;Frinak et al., 2005;Ma et al., 2010a).However, it is non-trivial to convert IR absorption intensity to the amount of adsorbed water (Schuttlefield et al., 2007a;Tang et al., 2016).Several previous studies have used quartz crystal microbalance (QCM) to measure DRH and mass hygroscopic growth of particles (Schuttlefield et al., 2007b;Arenas et al., 2012;Liu et al., 2016;Yeşilbaş and Boily, 2016).The frequency change of the quartz crystal in a QCM, according to the Sauerbrey equation, is proportional to the change in mass of particles loaded on the crystal (Schuttlefield et al., 2007b).In addition, the amount of water associated with particles at a given RH can also be determined by measuring the change in water vapor pressure before and after exposure of particles to water vapor (Ma et al., 2010b;Ma et al., 2012), in a manner similar to determination of Brunauer-Emmett-Teller surface area.In theory the electrodynamic balance can be used to investigate the mass hygroscopic growth of nonspherical particles (Chan et al., 2008;Lee et al., 2008;Pope, 2010;Griffiths et al., 2012).To our knowledge, however, this technique has not been applied to mineral dust and soot particles yet.
In this work we have developed an experimental method to investigate water adsorption and hygroscopicity of atmospheric particles, using a vapor sorption analyzer which is commercially available.We note that two groups have used similar techniques to measure water adsorption by CaCO3 and Arizona Test dust (Gustafsson et al., 2005) and DRH of malonic acid, sodium oxalate, and sodium malonate (Beyer et al., 2014;Schroeder and Beyer, 2016).Nevertheless, the performance of this technique has never been systematically evaluated.
To validate this experimental method, we have determined DRHs of six compounds as a function of temperature from 5 to 30 o C, and the measured DRHs, varying from ~20% to ~90% RH, show excellent agreement with literature values.In addition, mass hygroscopic growth factors (MGF) of (NH4)2SO4 and NaCl have been measured as a function of RH at 25 and 5 o C, and the measured MGF agree very well with those predicted by the E-AIM model (http://www.aim.env.uea.ac.uk/aim/aim.php;last accessed: 11 January 2016).We show that this instrument can measure a relative mass change (due to water uptake) of <0.025% within 6 hours and <0.05% within 24 hours, and the accuracy of mass change measurement is mainly limited by baseline drifts.These features make this instrument particularly useful for laboratory studies of water adsorption by nonspherical particles and/or particles with low hygroscopicity.

Experimental section
The instrument used in this work is a vapour sorption analyser (Q5000SA) manufactured by TA Instruments (New Castle, DE, USA).The first part of this section provides a general description of this instrument, and the second part describes experimental methods used in this work.

Instrument description
Figure 1 shows the schematic diagram of the vapor sorption analyzer used in this work to measure hygroscopicity and water adsorption of particles of atmospheric relevance.This instrument consists of two main parts: 1) a high-precision balance used to measure the mass of

High-precision balance
The balance simultaneously measures the mass of an empty pan (serving as a reference) and a sample pan which contains particles under investigation.Each pan is connected to the balance by a hang-down wire which has a hook at the lower end to hold the pan.The balance is housed in a chamber which is temperature regulated.To avoid moisture condensation, the balance chamber is purged with a 10 sccm (standard cubic centimetre per min) N2 flow regulated by a mass flow controller (MFC3).
The balance has a dynamic range of 0-100 mg.Typical dry mass of particles used in our experiments are around 10 mg or less so that the total mass of particles due to adsorption of Atmos.Meas. Tech. Discuss., doi:10.5194/amt-2017-56, 2017 Manuscript under review for journal Atmos.Meas.Tech.Discussion started: 6 April 2017 c Author(s) 2017.CC-BY 3.0 License.
water at high RH does not exceed the upper limit of the balance.The stated sensitivity of the balance is <0.1 μg with a weighing accuracy of ±0.1%, a weighing precision of ±0.01%, and a signal resolution of <0.01 μg.The 24-h baseline drift is stated to be <5 μg for an empty metalized quartz pan at 25 o C and 20% RH.This is equivalent to a relative mass change of <0.05% if the sample mass is 10 mg.As shown Section 3.3, experimental tests do suggest that such performance can be reached.

Humidity chamber
The humidity chamber is used to regulate the temperature and RH under which hygroscopicity and/or water adsorption of particles are investigated.Inside the humidity chamber are housed a reference chamber (in which an empty pan is connected to the balance) and a sample chamber (in which a sample pan is connected to the balance).A dry N2 flow (regulated by MFC2) is delivered through a humidifier and then mixed the second dry N2 flow (regulated by MFC1).The total flow is set to 200 sccm, and the ratio of these two flows can be adjusted in order to regulate the final RH.After mixing, the flow is then split into two flows, one delivered into the reference chamber and the other into the sample chamber.Therefore, both chambers should have the same temperature and RH.This can be verified by measurements using two RH sensors, as shown in Figure 1.The main advantage of using a reference chamber and a sample chamber is that the amount of water adsorbed by the empty pan and the attached wire can be simultaneously determined (and automatically subtracted using the provided software) under the same condition when water uptake by particles under investigation is being measured.In addition, the effect of buoyancy, which varies with RH (Beyer et al., 2014;Schroeder and Beyer, 2016), is also automatically taken into account by using an empty pan as the reference.
Temperature inside the humidity chamber can be adjusted from 5 to 85 o C with a stated stability of ±0.1 o C, and RH can be varied between 0 and 98%, with an absolute accuracy of

Other features
Q5000SA is equipped with a programmable autosampler designed to deliver sample pans into the humidity chamber.The autosampler can host up to 10 sample pans; however, in order to minimize contamination by lab air, only one sample pan is uploaded into the autosampler immediately prior to the measurement.The instrument status is displayed on a touch screen for local operation.Q5000SA can also communicate with a computer via Ethernet.Two software packages are provided by the manufacturer: 1) TA Instrument Explorer Q Series is used to control the instrument, program measurement procedures, and log experimental data; 2) TA Universal Analysis can be used for graphing experimental data in real time, data analysis, and exporting data.Experimental data can be sampled with frequencies up to 1 Hz.

Experimental procedures
In our work two major experimental protocols are developed to 1) determine the DRH and 2) quantify water adsorption and/or mass hygroscopic growth.Corresponding experimental procedures are detailed below.

DRH determination
Based on the standard recommended by American Society for Testing and Materials International (ASTM, 2007) and TA Instruments (Waguespack and Hesse, 2007), an experimental method has been developed in this work to determine the DRH of a given sample.
It consists of the following three major steps: 1) after the sample pan is properly located in the humidity chamber, temperature is set to the given value; 2) after temperature is stabilized, RH is set to a value which is ~5% (when change/difference in RH is mentioned in this work, it always means the absolute value) higher than the anticipated DRH and the system is   the second period (120-160 min, as shown in Figure 2a) is plotted as a function of RH, and the RH (57.6% in this case) at which the sample mass reached the maximum value is equal to the measured DRH (ASTM, 2007).
Measurement of a DRH usually takes ~3 h in total, and experimental data such as RH and sample mass are recorded with a time resolution of 10 s.If the DRH is unknown, we can increase the upper RH and decrease the lower RH used in the measurement so that the DRH falls into the RH range.Increasing RH range used in the measurement will of course lead to the increase of experimental time required.

Quantification of water adsorption and/or mass hygroscopic growth
The following experimental procedures are used to determine the amount of water adsorbed by a material (i.e.mass hygroscopic growth factors): 1) a sample pan is delivered into the humidity chamber and temperature in the humidity chamber is set to a given value; 2) after temperature becomes stable, RH in the humidity chamber is set to 0% and the sample is equilibrated with the environment until its mass change is <0.05% within 30 min; 3) RH is increased to a given value and the sample is equilibrated with the environment again until its mass change is smaller than a certain value (typically 0.05% for less hygroscopic materials such as CaCO3 and fresh soot, and 0.1% for more hygroscopic materials such as (NH4)2SO4 and NaCl) within 30 min; 4) RH is further increased to another given value and the sample is equilibrated with the environment.The following assumptions are made to convert the mass of adsorbed water to its surface coverage (Tang et al., 2016): 1) particles are spherical, having a uniform diameter of 1 μm and a density of 2.5 g cm -3 , and 2) the average surface area that an adsorbed water molecule occupies is 1×10 -15 cm 2 .Under these assumptions, a mass change of 0.05% (relative to the dry mass) due to adsorption of water is equal to a surface coverage of 0.7 monolayers for adsorbed water.

RH calibration
RH in the humidity chamber is measured by two RH sensors, as shown in Figure 1.RH measurements of vapor sorption analyzers and/or thermogravimetric analyzers can be calibrated/verified by determining the DRH of a reference material with a well-defined DRH (ASTM, 2007).In this work, NaBr provided by TA Instruments is used as the reference material (Waguespack and Hesse, 2007).We compare our measured DRHs of NaBr at six different temperatures with those reported by a previous study (Greenspan, 1977).The results are summarized in Table 1, suggesting that the differences between our measured and previous reported DRHs is <1% RH for temperatures ranging from 5 to 30 o C. The agreement is excellent, especially considering that 1) the stated RH measurement of our instrument has an accuracy of ±1% and 2) DRH values reported by Greenspan (1977) typically have errors of Atmos.Meas. Tech. Discuss., doi:10.5194/amt-2017-56, 2017 Manuscript under review for journal Atmos.Meas.Tech.Discussion started: 6 April 2017 c Author(s) 2017.CC-BY 3.0 License.±0.5%.Further inspection of results compiled in Table 1 reveals that the difference is larger at lower temperature and becomes smaller at higher temperature.(Greenspan, 1977).DRH values reported by Greenspan (1977), widely accepted as standard values, are recommended by the instrument manufacturer (Waguespack and Hesse, 2007) and also used in this study to calibrate our measure RH by taking into account the difference between our measured DRHs and those reported by Greenspan (1977) for NaBr at different temperatures.
All the RHs reported in this work (except measured DRHs of NaBr listed in Table 1) have been calibrated.In our work we have not verified RH measurements for temperature higher than 30 o C because the atmospheric relevance is limited.It should be pointed out if necessary, RH calibration can also be carried out at higher temperature (up to 85 o C) using the same procedure.

DRH measurements
Using the experimental method detailed in Section 2. with those reported by a previous study (Greenspan, 1977).Figure 3b shows the comparison of our measured DRHs with those reported in literature for Mg(NO3)2•6H2O as a function of  2 suggests that the absolute difference between our measured and previously reported DRHs is typically <1%.

Mass hygroscopic growth measurements
(NH4)2SO4 and NaCl are important components found in tropospheric aerosol particles, and their hygroscopicity has been well understood.They have also been widely used as standard materials for validation of hygroscopicity and cloud condensation nucleation activity measurements (Good et al., 2010;Ma et al., 2010b;Tang et al., 2015).In our work we have measured mass hygroscopic growth factors of (NH4)2SO4 and NaCl as a function of RH (0%, 30%, 60%, and 90%) at two different temperatures, with the purpose to further assess the performance of our new method.gives us further confidence that the method developed in this work is reliable for hygroscopicity measurements of atmospheric particles.A1).As shown in Figure 5, the ability of CaSO4•2H2O to uptake water is very limited, with the mass ratio of adsorbed water to dry particles is (0.450±0.004)% (1 σ) at 95% RH.This is qualitatively consistent with two previous studies which suggested that the cloud condensation nucleation activity of calcium sulfate aerosol particles is very low (Sullivan et al., 2009b;Tang et al., 2015).More detailed comparison and discussion are beyond the scope of this paper and will be addressed in a following publication.

Conclusion and outlook
The ability to uptake water vapor under sub-saturation conditions is one of the most important physicochemical properties of atmospheric particles, largely determining their  (Kelly and Wexler, 2005;Laskin et al., 2005;Zhang et al., 2008;Sullivan et al., 2009a;Han et al., 2013;Denjean et al., 2015;Tang et al., 2016).In the future, this instrument will be used to investigate water adsorption and hygroscopicity of mineral dust and soot particles before and after chemical processing.

Figure 1 .
Figure 1.Schematic diagram of Q5000SA used in this work.MFC: mass flow controller.High purity N2 is used for all the three gas flows regulated by MFC1, MFC2, and MFC3, respectively.

Figure 2 .
Figure 2. Typical experimental data in determination of DRH at a given temperature (NaBr at 25 o C as an example) using Q5000SA by linearly decreasing RH. (a): Change of normalized sample mass (solid curve, left y-axis) and RH (dashed curve, right y-axis) as a function of time; (b): Change of normalized sample mass as a function of RH when RH decreased linearly from 62% to 54%.

Figure
Figure2ashows changes of RH and sample mass (normalized to that at 0 min) as a function of time in an experiment to measure the DRH of NaBr at 25 o C. RH was kept at 62% in the first 120 min during which the sample mass increased with time.After that, RH was linearly decreased to 54% with a rate of 0.2% RH per min, and during this period the sample mass first continued to increase to a maximum value and then decreased with time.The sample mass in Atmos.Meas.Tech.Discuss., doi:10.5194/amt-2017-56,2017   Manuscript under review for journal Atmos.Meas.Tech.Discussion started: 6 April 2017 c Author(s) 2017.CC-BY 3.0 License.All the processes are programmed, with the flexibility to choose the number of RH steps and the corresponding RH values.Experimental data such as RH and sample mass are recorded with a time resolution of 30 s. Relevant experimental results will be presented in detail in Sections 3.3 and 3.4.
Figure 4 show the comparison of mass hygroscopic growth factors measured by our work with those predicted by the E-AIM model (Wexler and Clegg, 2002).It can be concluded from the comparison that our measured mass hygroscopic growth factors agree well with theoretical values for (NH4)2SO4 and NaCl at both 5 and 25 o C.This Atmos.Meas.Tech.Discuss., doi:10.5194/amt-2017-56,2017 Manuscript under review for journal Atmos.Meas.Tech.Discussion started: 6 April 2017 c Author(s) 2017.CC-BY 3.0 License.

Figure 4 .
Figure 4. Comparison of mass hygroscopic growth factors measured in this work with these predicted by the E-AIM model.(a) (NH4)2SO4 at 5 o C; (b) (NH4)2SO4 at 25 o C; (c) NaCl at 25 o C; (d) NaCl at 5 o C.

Figure 5 .
Figure 5. (a) Measured mass hygroscopic growth factors (normalized to the mass at 0% RH) of CaSO4•2H2O as a function of RH up to 95% RH.(b) Time series of RH and normalized mass of CaSO4•2H2O particles with a dry mass of ~9.05 mg during a hygroscopic growth experiment.This experiment was carried out at 25 o C.
impacts on atmospheric chemistry and climate.In this work, we have developed a new experimental method to investigate interactions of particles with water vapor under subsaturation conditions at different temperatures from 5 to 30 o C, using a commercial vapor sorption analyzer.Operation temperature can be increased up to 85 o C, though the atmospheric relevance is limited.We have provided a detailed description of instrument configuration as well as experimental procedures to determine DRHs and mass hygroscopic growth factors.For the temperature range we have covered in this work (5-30 o C), our measured DRHs of six different compounds with DRHs ranging from ~20% to ~90%, show excellent agreement with those reported in literature.In addition, mass hygroscopic growth factors measured in our work at different RH values agree well with those predicted by the E-AIM model for (NH4)2SO4 and NaCl at 5 and 25 o C. Therefore, we have demonstrated that experimental methods developed in our work can reliably measure DRHs and mass hygroscopic growth factors from 5 to 30 o C. To test the ability of this instrument to measure hygroscopic growth of compounds with low hygroscopicity, we have determined mass hygroscopic growth factors of CaSO4•2H2O at 25 o C. It has been found that the ability of CaSO4•2H2O to uptake water is very limited.The mass of water adsorbed by CaSO4•2H2O at 95% RH is only (0.450±0.004)% of its dry mass.It has also been observed that this instrument can measure a mass change of <0.025% within 6 hours and <0.05% within 24 h, and accuracy of mass change determination is mainly limited by baseline drifts.With such an accuracy, this instrument is particularly useful for quantitative determination of water adsorption and/or hygroscopicity of non-spherical particles such as mineral dust and soot.Atmospheric aging processes are known to alter water adsorption, hygroscopicity, and cloud condensation nucleation activity of mineral dust and soot particles Atmos.Meas.Tech.Discuss., doi:10.5194/amt-2017-56,2017 Manuscript under review for journal Atmos.Meas.Tech.Discussion started: 6 April 2017 c Author(s) 2017.CC-BY 3.0 License.

Table 1 .
Comparison of DRHs (in %) of NaBr at different temperatures measured in our study with those reported in literature

Table 2 .
Comparison of DRHs measured by our study with those reported in literature