**Electrochemical Evidence of non-Volatile Reduced Sulfur Species in Water-Soluble Fraction of Fine Marine Aerosols**

#### **Ana Cviteši´c Kušan <sup>1</sup> , Sanja Frka <sup>2</sup> and Irena Cigleneˇcki 1,\***


Received: 27 September 2019; Accepted: 30 October 2019; Published: 1 November 2019

**Abstract:** The traditional voltammetric method at the mercury electrode, and an acidification step developed for the determination of reduced sulfur species (RSS) in natural waters, was for the first time used for the quantification of RSS in the water-soluble fraction of fine marine aerosols collected at the Middle Adriatic location (Rogoznica Lake). The evidence of two types of non-volatile RSS that have different interaction with the Hg electrode was confirmed: mercapto-type which complexes Hg as RS–Hg and sulfide/S0-like compounds which deposits HgS. The analytical protocol that was used for RSS determination in aerosol samples is based on separate voltammetric studies of a methyl 3-mercaptopropionate (3-MPA) as a representative of mercapto-type compounds and sulfide as a representative of inorganic RSS. Our preliminary study indicates the presence of mainly RS–Hg compounds in spring samples, ranging from 2.60–15.40 ng m−<sup>3</sup> , while both, the mercapto-type (0.48–2.23 ng m−<sup>3</sup> ) and sulfide and/or S0-like compounds (0.02–0.26 ng m−<sup>3</sup> ) were detected in early autumn samples. More expressed and defined RS–Hg peaks recorded in the spring potentially indicate their association with biological activity in the area. Those samples were also characterized by a higher water-soluble organic carbon content and a more abundant surface-active fraction, pointing to enhanced solubility and stabilization of RSS in the aqueous atmospheric phase.

**Keywords:** reduced sulfur species; marine aerosols; water-soluble fraction; voltammetry; mercapto-type compounds; Rogoznica Lake

### **1. Introduction**

In different aquatic ecosystems, electrochemical methods have been widely used for the characterization and determination of different sulfur (S) species, comprising dissolved and/or particulate inorganic and organic S compounds, including thiols ([1–12] and references therein). Strong interaction between mercury electrode (Hg) and S species is a background for their electroanalytical determination and speciation at the Hg electrode. The methodology is based on the tendency of inorganic and organic reduced S species (RSS) to deposit a HgS layer [1–5,13–15] and/or RS–Hg complexes [9,10,15] (term "complex" refers to a different type of interaction between the analyte and the Hg electrode) during an accumulation step at the deposition potential (E*d*) around −0.2 V vs. the reference electrode (Ag/AgCl). In solutions containing sulfide anions, an insoluble HgS layer is formed during the deposition step at the Hg surface by the reversible process of a two-electron-transfer oxidation of Hg at potentials more positive than −0.5 V vs. the reference electrode (Ag/AgCl) (Equation (1)) [1–15]:

$$\rm{HS^{-}} + \rm{Hg^{0}} \leftrightarrow \rm{HgS(s)} + \rm{H^{+}} + \rm{2e^{-}}.\tag{1}$$

The same process occurs during the interaction of Hg with some other organic (thiourea, thiols, oxines, thioanions) and inorganic S species (polysulfides, dissolved and colloidal S, labile chalcogenides, i.e., CuS, PbS, HgS, FeS, Ag2S). During the potential scan toward negative values, layers of HgS and adsorbable RS–Hg complexes (in the case of DMS, DMDS, cysteine, glutathione) are reduced between −0.45 and −0.70 V (vs. Ag/AgCl) with facilitated reduction of the RS–Hg complexes at a more positive potential than −0.68 V (vs. Ag/AgCl) where HgS reduction occurs [2,9,10,15,16].

Natural cycling of S compounds through the environment has taken on a new significance due to their involvement in the formation of atmospheric aerosol particles and their influence on global environmental processes and human health [17–19]. Atmospheric S contributes the most to atmospheric acid deposition being detrimental to ecosystems, harming aquatic biota, as well as to a wide range of terrestrial plant life [20]. Moreover, S-rich atmospheric particulate matter (PM) can serve as cloud condensation nuclei and participate in the cloud formation processes, which can ultimately affect the radiative balance of the atmosphere and the Earth's climate [21,22]. The main anthropogenic sources of atmospheric S are coal and oil combustion, oil refining, and smelting of copper ores ([23,24] and references therein). The widespread combustion of fossil fuels has greatly increased S emissions into the atmosphere, with the anthropogenic component now substantially greater than natural emissions on a global basis [25]. The main natural sources are oceanic phytoplankton [21], volcanoes [26,27], and geothermal fields [28]. Production of marine S compounds is mostly related to eutrophication phenomena and spreading of hypoxic–anoxic dead zones in the marine environment. A significant fraction of natural S emission occurs in the form of dimethylsulfide (DMS), produced by phytoplankton and zooplankton grazing, which is released from the ocean into the atmosphere, where it undergoes oxidation to form sulfur dioxide (SO2), dimethylsulfoxide, dimethylsulfone, methanesulfonic acid (MSA), and sulfate (SO<sup>4</sup> <sup>2</sup>−) [29–31].

The wide range of S oxidation states (from −2 to +6) as well as organic and inorganic forms present in ambient samples makes the characterization of aerosol S challenging [32]. Sulfur in aerosols is usually measured in the form of inorganic sulfate ion, and it is often assumed to be one of the most important forms of aerosol S. However, in addition to sulfate, other S compounds, even in smaller quantities, are present in ambient aerosols. Among these, the most abundant are sulfides, organosulfur/sulfate compounds, and polycyclic aromatic sulfur heterocycles (PAHSs) [33–35]. Even though different methodological approaches have been applied to resolve S content in aerosols, including ion chromatography, X-ray fluorescence, and inductively coupled plasma mass spectrometry, these cannot determine the oxidation state or directly identify the chemical form of aerosol S. Thus, there is still a need for accurate and direct method enabling speciation of different S species in small amounts such as those found within PM2.5 and/or PM<sup>10</sup> samples (usually no more than a few milligrams).

In this work, for the first time the water-soluble (WS) fraction of the fine marine aerosols was studied by voltammetric methods that are used thus far for RSS characterization in natural waters. An analytical protocol for the electrochemical RSS characterization in the WS fraction of aerosol samples is given based on separate voltammetric studies of a methyl 3-mercaptopropionate (3-MPA), and its mixture with sulfide. Here, 3-MPA is used as a representative for mercapto-type compounds (RS–Hg type) which gives a more positive reduction peak (around −0.60 V vs. Ag/AgCl at neutral pH) than standard HgS reduction (around −0.70 V vs. Ag/AgCl) usually considered for RSS quantification in different natural samples [1–8].

### **2. Experiments**

### *2.1. Materials and Methods*

The chemicals used were of reagent grade. The 3-MPA Me 3-MP stock solutions were prepared by dissolving 98% methyl 3-mercaptopropionate (Merck, Darmstadt, Germany) in N<sup>2</sup> degassed ultra-pure water (Milli-Q, 18.2 MΩ, total organic carbon (TOC) < 3 ppb). The inorganic sulfide stock solutions

were prepared by dissolving Na2S crystals (Sigma-Aldrich, Steinheim, Germany) in N<sup>2</sup> degassed solution of 2 <sup>×</sup> <sup>10</sup>−<sup>4</sup> M NaOH (Merck, Darmstadt, Germany). Working solutions were prepared by adding, small volumes of the 3-MPA and/or Na2S stock solutions to the previously N<sup>2</sup> degassed solutions of supporting electrolyte (0.55 M NaCl, pH ≈ 6). The above-mentioned conditions, including the supporting electrolyte selection, are chosen to be consistent and comparable with our previous measurements by the same approach [2,7,8,14,16,36]. Nevertheless, we tested the method using the electrolyte of lower ionic strength to mimic more atmospheric conditions (0.1 M NaCl) and did not detect any significant difference in the electrochemical response.

### *2.2. Aerosol Sampling and Preparation*

Natural samples of atmospheric PM2.5 were collected by using the low-volume Sequential Sampler SEQ 47/50 (SEQ47/50) (Sven Leckel, Ingenieuburo GmbH, Berlin, Germany) on a pre-baked (450 ◦C for 4 h) glass fiber filters (GF/F) (Whatman, grade GF/F, d = 47 mm). The sample collection was conducted at a flow rate of 2.3 m<sup>3</sup> /h for 48 h.

Sampling was conducted throughout the spring and early autumn in 2016 at the Middle Adriatic location (next to marine Rogoznica Lake, 43.53◦ N, 15.95◦ E). According to the air-mass backward trajectory analysis by using the NOAA HYSPLIT model at 10 m above sea level continental air-mass inflows mostly affected sampling area in winter and autumn, while in spring and summer periods southern marine pathways had higher impact. From October to March the synoptic wind circulation was northerly, while from April to September the direction of wind changed to southerly with the domination of south western wind during the summer.

Collected aerosol samples were stored in Petri slides (Millipore Inc., Darmstadt, Germany) at −50 ◦C prior to analysis. Approximately 10% of the filter sample was cut in pieces and dissolved in 10–20 mL of ultra-pure water (Milli-Q, 18.2 MΩ, total organic carbon (TOC) < 3 ppb), ultrasonicated for 20 min, and stored at 4 ◦C overnight. Afterward, the solution was filtrated through GF/F filters of 0.7 µm pore sizes and further treated as WS aerosol fraction. For the electrochemical measurements the final ion strength of the WS solution was adjusted by adding saturated NaCl solution to correspond to 0.55 M NaCl electrolyte.

For sulfate (SO<sup>4</sup> <sup>2</sup>−) analysis approximately a quarter of each exposed filter was cut and dissolved in 10 mL of ultra-pure water (Milli-Q, 18.2 MΩ, total organic carbon (TOC) < 3 ppb), ultrasonicated for 20 min, and stored at 4 ◦C overnight. The solutions were filtrated through 0.22 µm PTFE disk filters and analyzed by ion chromatography (IC). SO<sup>4</sup> <sup>2</sup><sup>−</sup> anions were measured on a Dionex ICS 3000 ion chromatograph (Thermo Scientific, Sunnyvale, CA, USA) with a conductivity detector.

### *2.3. Electrochemical Instrumentation and Procedure*

The RSS measurements: The RSS measurements were performed with a µ-Autolab (Electrochemical Instrument Eco Chemie, Metrohm Autolab B.V., Utrecht, Netherlands) electrochemical analyzer connected to a 663 VA Stand Metrohm mercury electrode. The reference electrode was an Ag/AgCl/3 M KCl (Ag/AgCl electrode connected to the solution with an electrolyte bridge). A platinum electrode served as the auxiliary electrode. The cell volume was 25 mL.

The applied electrochemical technique was cathodic striping square wave voltammetry (CSSWV). Determination of the non-volatile RSS compounds in Hg was made by their deposition at positive electrode potentials (E*<sup>d</sup>* = −0.2 and −0.4 V vs. reference electrode Ag/AgCl) before and after the acidification step and purging by N<sup>2</sup> to remove volatile RSS as already described [2,7,16]. Namely, non-volatile RSS can be measured by voltammetry after acidification and purging of the sample with an inert gas to remove all volatile RSS. For the given experimental conditions, purging by N<sup>2</sup> was performed between 3 and 5 min. Cathodic scans with and without accumulation (acc. time, t*<sup>a</sup>* = 0–120 s) by stirring at the starting deposition potentials were run with an amplitude (A) of 25 mV and frequency (f) of 80 Hz. The sensitivity of the measurements with 120 s accumulation time is given by the limit of detection (LOD) and limit of quantification (LOQ), which for sulfide measurements

were 0.13 and 0.46 nM, respectively, and for 3-MPA the LOD was 0.66 nM and the LOQ 2.21 nM. Quantitation of RSS in the studied samples was based on external calibration, by using calibration curves for sulfide and 3-MPA in 0.55 M NaCl.

The surface-active substance (SAS) measurements: SASs were determined by the electrochemical method of alternating current (a.c.) voltammetry with out-of-phase mode using a µ-Autolab (Electrochemical Instrument Eco Chemie, Metrohm Autolab B.V., Utrecht, Netherlands) electrochemical analyzer connected with 663 VA Stand Metrohm mercury electrode equipped as described above. A.c. voltammetry is based on the SAS adsorption effects at the hydrophobic surface of Hg (at E*<sup>d</sup>* = −0.6 V) [2,36 and references therein]. The measurements were performed under following conditions: accumulation potential E*<sup>d</sup>* = −0.6 V, t*<sup>a</sup>* = 30 s, f = 75 Hz, amplitude *A* = 10 mV, and phase angle = 90◦ . The concentration of SAS in each sample was expressed in equivalents (mg L−<sup>1</sup> ) of a model surfactant Triton-X-100 (T-X-100). The detection limit of SAS determination is 0.02 mg L−<sup>1</sup> equivalent of T-X-100, with LOQ of 0.05 mg L−<sup>1</sup> . All electrochemical measurements were done at room temperature (21 ± 1 ◦C).

In addition, water-soluble organic carbon (WSOC) content was determined by the hightemperature catalytic oxidation (HTCO) method at a TOC-VCPH instrument (Shimadzu, Japan) as described previously [7,16,36]. The WSOC sample aliquot (15 mL) was acidified with 2 M HCl to pH ~ 3 in order to eliminate the inorganic carbonates. The concentration of each sample was calculated as an average of three to five replicates. The quantification limit was 0.228 M for dissolved organic C with reproducibility of 5%.

### **3. Results and Discussion**

### *3.1. Electrochemical Behavior of 3-MPA*

During voltammetric experiments, accumulation of 3-MPA on Hg at E*<sup>d</sup>* = −0.2 V resulted in the formation of a 3-MPA–Hg complex (Equation (2)), which was then reduced by scanning the potential toward more negative values. The reduction of the formed complex occurred at around −0.6 V by giving a reversible and reproducible reduction peak according to process explained by the reverse of Equation (2) [9,37]:

$$2\text{RSH} + \text{Hg} \leftrightarrow \text{(RS)}\\2\text{Hg} \text{(ads)} + 2\text{H}^+ + 2\text{e}^-. \tag{2}$$

The RSH stands here for thio-type RSS (DMS, DMDS, cysteine, methanethiol, gluthatione). In general, the position and the height of the 3-MPA–Hg reduction peak is sensitive to the pH of the bulk solution as already noted for other RSS studied at the Hg surface [2,3,12,15]. Therefore, in this study, measurements were done at the same conditions, pH ~ 6. By prolonging the accumulation time with stirring (t*<sup>a</sup>* = 0–120 s) at the starting E*<sup>d</sup>* = −0.2 V the height of the 3-MPA–Hg reduction peak increases, I*<sup>p</sup>* as shown for the 9 nM 3-MPA solution (Figure 1a). The same effect is obtained by increasing the 3-MPA bulk concentrations (Figure 1b) used for the preparation of a calibration curve (Figure 1c).

**Figure 1.** Cathodic striping square wave voltammetry (CSSWV) curves showing (**a**) accumulation effect in 9 nM methyl 3-mercaptopropionate (3-MPA) electrolyte solution; (**b**) effect of increasing concentration of the 3-MPA, and (**c**) calibration curves for the 3-MPA and Na2S, with insertion of CSSWV curves for increasing sulfide concentration in electrolyte solution. Experimental conditions: E*<sup>d</sup>* = −0.2 V, A = 25 mV, f = 80 Hz, t*<sup>a</sup>* = 120 s.

The linear relationship between the 3-MPA concentration and I*<sup>p</sup>* which is a direct measure of the reduction process at the Hg, is obtained for the wide concentration range, between 3.0 and 500 nM for an accumulation step with stirring (t*<sup>a</sup>* = 120 s). For the given experimental conditions at concentration of 500 nM, saturation of the electrode was achieved. Considering the concentrations of RSS expected in marine aerosol samples, we focused on the calibration in the lower concentration range, between 2 and 50 nM, as presented in Figure 1c. For comparison, the calibration curve of the Na2S is presented in the same graph. Typical voltammetric curves for increasing sulfide concentration are given as an insert in the same Figure. The height of the standard HgS reduction peak, I*p,* increases sharply with increasing the concentration of Na2S in the bulk solutions, being about four times more sensitive than the I*<sup>p</sup>* of the same 3-MPA concentrations (Figure 1c), as already documented for thiols, Na2S, and S<sup>0</sup> [38,39]. The peak potential of the HgS reduction, if compared with the 3-MPA-Hg reduction, is moving as well more negatively with the increase of the bulk sulfide concentration, i.e., the amount of the formed HgS layer [13,14]. Such a difference in sensitivity is used in this study as an advantage in characterization of mixtures containing organic and inorganic RSS, i.e., the 3-MPA and sulfide.

Variation of the deposition potentials (at the given experimental conditions) from E*<sup>d</sup>* = 0.00 to −0.50 V has been already shown to influence electrochemistry of the organic RSS at the Hg surface [9,10,13,40], i.e., the appearance and height of the revealed reduction peaks. Such a property of setting the electrode at the unique deposition potential of maximum response for the selected model compounds was successfully used in the analytical protocol for thiol characterization in seawater and freshwater samples [6,9–11,40,41]. Depending on the deposition potential, dimethylarsinyl-ethanol sulfide [13] and similar labile compounds would be oxidized and deposit the HgS layer at the Hg surface. The same behavior is also characteristic for thiourea [42] and some other low-molecular-weight thiols (LMWTs) [9,10] which deposit HgS around 0 V. The observed behavior is caused by the different stability of the RS–Hg and HgS-type compounds as well as different lability of organic RSS at the studied Hg electrode potentials. Here, by changing the deposition potential from −0.2 toward −0.4 V, for the solution containing 3-MPA, the height of the 3-MPA–Hg reduction peak is significantly decreasing by depositing at more negative potentials (Figure 2). Sulfide and elemental sulfur (S<sup>0</sup> ) in the same range of deposition potentials would not change their electrochemical behavior [13].

**Figure 2.** Effect of different deposition potentials, E*<sup>d</sup>* = −0.2 V (red lines), and E*<sup>d</sup>* = −0.4 V (blue lines) and different accumulation times for electrolyte solution containing 9 nM 3-MPA. Experimental conditions: A = 25 mV; f = 80 Hz.

Usually in the adopted electrochemical measurements of the RSS in natural samples, the acidification and purging step by N<sup>2</sup> is used to make a distinction between the volatile and non-volatile fraction of the RSS [2,3,7,13,16]. Sulfide-type RSS are considered volatile, while elemental sulfur (S<sup>0</sup> ) and S<sup>0</sup> -like compounds (including some LMWTs that deposit HgS and are not sensitive to purging in acidic conditions) are considered non-volatile RSS that mainly contribute to the HgS peak in natural samples. The 3-MPA is shown to be quite stable in water solution. Experiments with 3-MPA approved its stability in water solution within several hours. As distinguished from sulfide which disappears completely from acidic solution (pH = 3–5) in 3–5 min depending on the concentration, the 3-MPA is insensitive to acidification and purging with nitrogen (see Figure 3). This different behavior of 3-MPA and sulfide in acidic conditions enables their direct electrochemical determination in the solution.
