Technical Note: Real-Time Diagnosis of the Hygroscopic Growth Micro-Dynamics of Nanoparticles with Two-Dimensional Correlation Infrared Spectroscopy

Nanoparticles can absorb water to grow up and this will affect the light scattering behavior, cloud condensation nuclei properties, lifetime, and chemical reactivity of these particles. Current techniques usually assume the shapes of nanoparticles to be spherical in calculation of aerosol liquid water content (ALWC), which may result in large uncertainties when the shapes of nanoparticles show large deviations to the 20 spherical assumptions. Furthermore, current techniques are also difficult to identify the intermolecular chemical interactions of phase transition micro-dynamics during nanoparticle deliquescence process because their limited temporal resolutions are unable to capture the complex femtosecond-level intermediate states. In this study, the hygroscopic growth properties of nanoparticles with electrical mobility diameter of 25 approximately 100 nm and their phase transition interaction dynamics on molecular scale are characterized on real time by using the Fourier transform infrared (FTIR) and the two-dimensional correlation infrared (2D-IR) spectroscopic techniques. With the FTIR spectroscopy, we develop a novel real-time method for ALWC by constructing the absorption spectra of liquid water, and realized real-time measurements of water 30 content and dry nanoparticle mass to characterize the hygroscopic growth factors (GF) which show discrepancies to the extended aerosol inorganics model (E-AIM). We further explore the difference that the deliquescence points of sodium nitrate (SN) and oxalic acid (OA) compounds are lower than that of AS by using the 2D-IR spectroscopic analysis technique. We also identify the occurrence sequential order of 35 2 the hydration interactions and investigate the dynamic deliquescence process of the functional groups for AS and its mixture compounds. Both SN and OA compounds lower the deliquescence point of AS, but only AN can change the hydrolysis reaction mechanism for AS in AS/AN and AS/OA mixtures. This study can not only provide important information with respect to the difference in phase transition point under 40 different conditions, but also improve current understanding of the chemical interaction mechanism between nanoparticles (particularly for organic particles) and medium, which is of great significance for haze control across China. and the two-dimensional correlation infrared (2D-IR) spectroscopic techniques. We use a FTIR spectrometer and an extended aerosol inorganics model (E-https://doi.org/10.5194/acp-2021-763Preprint.Discussionstarted:15November and the measured hygroscopic growth properties from the FTIR spectra for the AS particles. The results show that the


Introduction
Nanoparticles have long atmospheric lifetimes of weeks to months. As the increase in relative humidity, the sizes of nanoparticles will grow up due to the absorption of water, which may have complex phases and mixing states (Riemer et al., 2019) that influence the light scattering behavior, cloud condensation nuclei properties, lifetime, 50 and chemical reactivity of the particles (Lee and Allen, 2012;Vogel et al. 2016;Abbott and Cronin, 2021). An improved knowledge of these complex phases and states are crucial for investigating the gas-particle interactions in the atmosphere. Since the particle size vs. water uptake relationship is influenced by mixing characteristics of various inorganic and organic compounds (Nguyen et al., 2016;Steinfeld and Pandis, 55 2016), characterizing the water-aerosol interactions is also critical for identifying the fate and transport of trace species in the Earth's system and their effects on air quality, radiative forcing, and regional hydrological cycling (Carlton et al., 2020;Fan et al., 2018).
Ammonium sulfate is an important constituent and a major source of atmospheric 60 nanoparticles originated from anthropogenic activities (Ruehl et al., 2016;Kirkby et al., 2011;Xu et al., 2020). Various techniques such as the hygroscopic tandem differential mobility analyzer (H-TDMA), the electrodynamic balance (EDB), and the environmental scanning electron microscope (ESEM) have been used to investigate the https://doi.org/10.5194/acp-2021-763 Preprint. Discussion started: 15 November 2021 c Author(s) 2021. CC BY 4.0 License. hygroscopicity of ammonium sulfate (Tang and Munkelwitz, 1977;Tang and 65 Munkelwitz, 1994;Gysel et al.,, 2002;Matsumura and Hayashi 2007). These methods can characterize the deliquescence or phase transition processes of particles down to nanoscale. However, they usually assume the shapes of nanoparticles to be spherical in calculation of aerosol liquid water content (ALWC), which may result in large uncertainties when the shapes of nanoparticles show large deviations to the spherical 70 assumptions. Furthermore, current techniques are also difficult to identify the intermolecular chemical interactions of phase transition micro-dynamics during nanoparticle deliquescence process because their limited temporal resolutions are unable to capture the complex femtosecond-level intermediate states.
Recent studies concluded that the phase transition processes of particles may 75 include multiple intermediate states and are more complex than those indicated in previous results. These intermediate states differ from one to the other and last less than 10 ms (Esat et al., 2018). A label-free photonic microscope which uses Bloch surface waves as its illumination source for imaging and sensing is capable to provide real-time measurements of the hygroscopic growth process of a single aerosol with particle 80 diameter of less than 100 nm (Kuai et al., 2020). This method can provide valuable insights into the deliquescence and phase transition mechanisms of particles but cannot determine chemical composition information of particle deliquescence or growth or phase transition processes. It is necessary to develop a method to characterize the intermolecular interaction mechanisms during hygroscopic growth of nanoparticles, 85 which is crucial to understand the physicochemical properties of atmospheric aerosol and the nanoparticle-water interactions in hygroscopic growth, and further for haze control purpose.
In this study, the hygroscopic growth properties of mixed nanoparticles containing (NH4)2SO4/NaNO3 (ammonium sulfate (AS)/sodium nitrate (AN)) and 90 (NH4)2SO4/oxalic acid (AS/OA) and the phase transition interactions of these particles on molecular scale are characterized on real time by using the Fourier transform infrared (FTIR) and the two-dimensional correlation infrared (2D-IR) spectroscopic techniques. We use a FTIR spectrometer and an extended aerosol inorganics model (E-https://doi.org/10.5194/acp-2021-763 Preprint. Discussion started: 15 November 2021 c Author(s) 2021. CC BY 4.0 License. AIM) to characterize and predict the hygroscopic growth of pure AS particle and the 95 AS/AN and AS/OA mixed particles, respectively. We further use the 2D-IR spectroscopic technique to analyze the intermolecular interactions of the phase transition with respect to different levels of relative humidity (RH). This study can not only provide important information with respect to the difference in phase transition point under different conditions, but also improve current understanding of the 100 chemical interaction mechanism between nanoparticles (particularly for organic particles) and medium, which is of great significance for haze control across China.

Experiment description
The experimental system includes a nanoparticle generation system, a 105 humidification system, and a FTIR analysis system. Nanoparticles with diameters of ∼100 nm (volume equivalent diameter (Dve)) are aerosolized by an atomizer (model 255, MetOne), dried by a diffusion dryer (model 3062, TSI), sorted into specific diameters (Dve) by a differential mobility analyzer (DMA; model 3082, TSI), and finally deposited onto a 3 cm × 3 cm zinc selenide (ZnSe) substrate ( Figure 1) inside a 110 sample cell through a cone-shaped hole. The sheath-to-sample flow ratio of the DMA is set to be 10:1(the sheath flow is 10 L/min and the sample flow is 1 L/min), which can produce an effective mobility for the measured aerosol with size ranging from 14.9 to 673.2 nm. The nanoparticles with a Dve of ∼100 nm are selected for deposition. After about a deposition time of 12 h, the substrate is sealed inside the sample cell to obtain This time interval is used to stabilize the atmospheric condition inside the sample cell.

120
The FTIR spectrometer is equipped with a KBr beam splitter and a liquid nitrogencooled mercury cadmium telluride (MCT) detector for measuring the absorption spectra of the samples. A He-Ne laser metrology keeps the FTIR instrument in a good optical alignment. The FTIR spectrometer saves middle infrared (MIR) spectra with a spectral range of 800 to 4000 cm −1 , spectral resolution of 4 cm -1 , and repeat times of 64. The then are used to derive the liquid water content using the optical constants of water in the infrared provided by Downing and Williams(Downing and Williams, 1975). We quantify the nanoparticles masses using a simple procedure described in our previous studies without any treatment to the substrates (Wei et al., 2019).

130
After baseline correction, the infrared spectra are normalized into 2D-IR spectra with the 2D Shige software (Kwansei-Gakuin University, Japan) (Noda and Ozaki, 2014). The wavenumber regions ranging from 2800-3800cm −1 and from 800-1400cm −1 which cover the absorption features of almost all identifiable functional groups of interest are selected for analysis. In present work, the red and blue colors in 135 the 2D-IR spectra represent positive and negative correlations, respectively.

Sample description
In this study, all chemical reagents are produced by Aladdin Reagent Inc. (reagent grade, 99.8% purity), and the water is obtained from an ultrapure water system (Direct-Q3, Millipore). Table 1

Methodology
The hygroscopic growth factor (GF), indicating the water uptake ability of aerosol particles, is defined as GF = Dwet/D0, where Dwet (cm) is the mean diameter of the particles at the designated RH and D0 (cm) is the mean initial diameter of the dry particles at room temperature. In present work, room temperature is assumed to be 25°C 155 and the RH varies from 50% to 95%. The GF used for investigation of hygroscopic growth properties of nanoparticles can be calculated via equations (1) to (4), Where V0 (cm 3 ) is the initial volume of the dry nanoparticle at approximately 25°C, and Vwater (cm 3 ) is the water volume contained in the nanoparticle at the designated RH; Mwater (g) and Mi (g) are the calculated water mass and the mass of the i th pure component at the designated RH, respectively; ρwater (g/cm 3 ) (approximately 1 g/cm 3 ) 165 and ρi (g/cm 3 ) are the densities of water and the i the pure component, respectively, and i is the number of pure component.
We use the E-AIM (UNIFAC) following the Zdanovski-Stokes-Robinson (ZSR) method to predict GF (http://www.aim.env.uea.ac.uk/aim/aim.php). The basis of this method is described as, Where i  is the volume fraction of the i th pure component in the dry mixture, GFi is the GF of the i th pure component.  predicted and measured Mwater/M0 results are generally consistent throughout the humidification process. The strong peaks observed at 3250 cm −1 and 1112 cm −1 at the 185 initial RH of 45% are the stretching vibration peak of OH and the symmetrical stretching vibration (νs) peak of the sulfate, respectively (Wang et al., 2017;Ná jera and Horn, 2009;Gopalakrishnan et al., 2005). With the increase in RH between 45% and 80%, the peak position of the symmetrical stretching vibration of the sulfate (1112 cm −1 ) starts to redshift slowly (Figure 3), which indicates that water molecules have been 190 attached to the surface of the solid AS, and the sulfate is then bonded with these water molecules to form a hydrogen bond during this hydration process (Yeşilbaş and Boily, 2016). The area of OH reflects the liquid water content in the ZnSe substrate. We find that, in the meantime, the area and position of the OH stretching peak did not change significantly, which indicates that no hygroscopic growth of the AS nanoparticles 195 occurs. All these behaviors are captured on real time by the FTIR spectra. (Wang et al., 2019;Tang et al., 2016;Ná jera and Horn 2009;Martin 2000) When the RH reaches 80%, the peak position of the sulfate shifts to 1099 cm −1 , the O-H stretching peak is still the same as that in initial humidity condition (45%) but the area of the peak increase abruptly from 0.25 to 5.47 ( Figure 3). This indicates that the 200 nanoparticles have absorbed water rapidly and transformed from the crystalline phase to the aqueous phase. According to the E-AIM predictions and the results from previous studies (Estillore et al. , 2016;Cruz and Pandis, 2000;Tang 1982), this process is called the deliquescence, and the RH at this stage is referred to as the deliquescence RH (DRH).

Results and discussion
When deliquescence occurs, NH4 + molecules hydrated with SO4 2− are replaced with 205 H2O molecules, which leads to the redshift in the symmetrical stretching vibration peak for the sulfate (Dong et al., 2007). Tang et al. (1982), Cruz and Pandis (2000), and Estillore et al. (2016) have used the photonic microscope to observe hygroscopic growth properties of big-size particles. Our method and the size of particle are different from previous studies, but we obtained a consistent DRH to those in Tang et al. (1982), 210 Cruz and Pandis (2000), and Estillore et al. (2016). Since the particle size (~100 nm) in this study is much smaller than those in previous studies and is not influenced by Kelvin effect, we can capture the hygroscopic growth properties of nanoparticles on real time with the FTIR spectroscopy.
The AS nanoparticle continues to be humidified after deliquescence, resulting in a 215 further increase in OH area due to continuous water uptake. However, the peak position of the sulfate keeps constant regardless of RH, indicating that the AS is still in the aqueous phase after deliquescence. As the increase in RH, the nanoparticle volume increases but its mass keeps constant, resulting in a decrease in concentration and peak area. Figure S1

Hygroscopic growth of pure and mixed-component nanoparticles
With the results derived from the FTIR measurements, we calculated the GFs for both pure and mixed nanoparticles via equation (4) and investigated their variabilities with respect to the changes in RH. Figure 6 compares the measured and predicted GF for both pure and mixed nanoparticles under the humidity conditions from 50% to 95%.

245
The measured and predicted GF are in good agreement. The GF can be obtained precisely using the H-TDMA technique via a direct measurement to the aerosol diameter. In this study, the GFs for both pure and mixed compounds are calculated with liquid water content and the relative masses of dry compound obtained from FTIR The GF curves can be used to investigate the sensitivity of particle volume to RH.
If the RH is less than 80%, the AS nanoparticles are in a stable crystal state that is immune to water and the size of these particles keeps constant. At the RH of 79.9 ± 255 0.10%, deliquescence occurs, the nanoparticle volume grows up sharply by up to approximately 3.1 times and transforms from the crystalline to aqueous phase. As RH keeps increasing after deliquescence, the AS nanoparticles become fully liquid droplets, and their volumes keep increase due to further water uptake.
Since both the AS/AN and AS/OA mixed nanoparticles absorb liquid water below 260 their DRH, their GF curves differ from the pure AS particle, and their DRH values were lower than that of the AS. The results are in good agreement with previous studies(Seinfeld and Pandis, 2016) which measure GF by H-TDMA. As a result, FTIR measurement technique in this study provide a real-time method to characterize the hygroscopic growth of aerosols.

Phase transition dynamics of pure AS nanoparticles
Although FTIR measurements can be used to characterize the liquid water content and functional groups contained in the nanoparticles during the humidification process, it is difficult to separate the absorption peaks of the nanoparticles (especially for organic compounds) since these absorption peaks are overlapped. In contrast, 2D-IR 270 spectroscopic technique can resolve the overlapping peaks (McKelvy et al., 1998;Du et al., 2021) and, more importantly, can provide detailed information about the dynamic deliquescence processes of the functional groups (Noda and Ozaki, 2014). Synchronous correlation maps reflect the simultaneous changes that occur in the two separate spectral intensity variations. Asynchronous spectra can be used to identify the occurrence  Furthermore, two main blue/negative (1112, 1097) and (3250, 1112) auto-peaks were also observed for the AS nanoparticles, which indicates that the O-H stretching peak intensity is increasing, while the symmetrical stretching vibration for the sulfate when hydrated with NH4 + in the solid is decreasing. The blue/negative (1112, 1097) auto-290 peak indicates that the intensity of the symmetrical stretching vibration for the sulfate in the aqueous AS increased, while that in the solid AS decreased, and these two chemical bonds can be converted into each other. This behavior can thus be explained by the fact that NH4 + particles hydrated with SO4 2− are being replaced by H2O molecules with increase in RH.

295
In addition to the synchronous map, the asynchronous map indicates the sequential changes in the spectral intensities in response to the hygroscopic activities. Two main red/positive (3250, 1097) and (1112, 1097) and two main blue/negative (3250,1112) and (1097, 1112) auto-peaks are observed for the AS nanoparticles, which indicated that the peaks changed in the order from (1112) cm −1 > (3250) cm −1 > (1097) cm −1 . The 300 intensity of the symmetrical stretching vibration for the sulfate when hydrated with NH4 + in the solid would decrease, and then water molecules would attach to the surface of the solid (NH4)2SO4 (Yeşilbaş and Boily, 2016); finally, NH4 + particles hydrated with SO4 2− are replaced by the H2O molecules and the AS nanoparticles then become fully liquid droplets. This indicates that the surface-limited processes may control the water 305 transport to the AS. (3250, 1320) and (1112, 1097) auto-peaks are observed, which indicates that the peaks 315 change in the order from (1320) cm −1 > (3250) cm −1 > (1097) cm −1 > (1112) cm −1 . This indicates that the nitrate would be in the aqueous solution at a lower RH than the sulfate because the AN has a lower DRH (RH=74.3±0.4%, (Seinfeld and Pandis, 2016;Tang and Munkelwitz, 1993)), and the characteristics of the nitrate when hydrated with water differ from those of the sulfate. Furthermore, the hydrolysis reaction mechanism for the 320 sulfate in AS/AN may differ from that for the pure sulfate. One possible explanation for this phenomenon is that the nitrate would begin to absorb water at low RH, which enhances the dissolution of the AS. Therefore, the NH4 + particles that would be hydrated with the sulfate are replaced by the H2O molecules, and the intensity of the symmetrical stretching vibration for the sulfate in the solid AS nanoparticles would then  and one main blue/negative (3250, 1112) auto-peaks are observed, which indicated that the peaks change in the order from (1112) cm −1 > (3250) cm −1 > (1080) cm −1 . Therefore, 335 the hydrolysis reaction mechanism for the sulfate in AS/OA may be similar to that for the pure sulfate. The 2D-IR measurements could thus provide a real time method to characterize the dynamic variability of the nanoparticles during the hygroscopic growth process.

340
In this work, we demonstrate use of FTIR spectroscopy to measure the hygroscopic growth properties of assembled nanoparticles and combine this technique with 2D-IR spectroscopy to identify the occurrence sequential order of the hydration interactions and provide detailed information about the dynamic deliquescence processes of the functional groups. This approach enabled measurement of the water content and the dry 345 nanoparticle mass to characterize the hygroscopic GF and also further investigation of the deliquescence process, with results that were matched well with those obtained from

Author contribution
XW designed the experiment and wrote the paper with contributions from all coauthors; HG contribute to science discussions and suggested analyses; HD and JZ 365 prepared for the humidification system; YC, JW, YY and JL contributed to this work by providing constructive comments; YS contributed to this work by providing constructive comments, review, and editing.