Producing Safe Drinking Water Using an Atmospheric Water Generator (AWG) in an Urban Environment

: Alternative new technologies are urgently needed to overcome the rapidly increasing global water scarcity. Atmospheric dew water is a potential source of potable water, as the earth’s atmosphere contains billions of tons of fresh water (98% in a vapor state). The atmospheric water generator (AWG) converts water vapor into liquid water and is a promising solution for water scarcity. We provide the ﬁrst comprehensive analysis of the chemical proﬁles of water produced for several months by an AWG in the city of Tel Aviv, Israel. Metals, inorganic ions, volatile organic compounds (VOCs), and semi-VOCs were analyzed in the dew water. The main elements found were ammonium, calcium, sulfate, and nitrate. Location of the sampling site in an urban residential area, between major tra ﬃ c routes, likely a ﬀ ected the chemical composition of the produced dew water. Nevertheless, the produced water nearly always (day and night in di ﬀ erent seasons) met the WHO and Israeli drinking water standards. Thus, even in a highly developed urban environment, the AWG o ﬀ ers an excellent alternative source of safe drinking water throughout the year.


Global Water Scarcity and Alternative Drinking Water Sources
Freshwater scarcity is an ever-growing concern for human society [1]. In its last report, the World Economic Forum listed water crises as one of the top five global risks to Man worldwide, in terms of impact and likelihood [2]. Currently, over 2 billion people across the globe are experience high water scarcity, a number that is expected to rise with population growth and the intensity of climate change [3].
Solutions for water scarcity are based on various water-saving strategies, reclaiming of used water, and water production. Seawater desalination by reverse osmosis is the most common and efficient water-production method [4]. However, it requires a large saline or brackish water source and is therefore not applicable in regions with no access to brackish or sea water. Moreover, desalination plants and long pipelines require large capital investments. Distillation is therefore not relevant to poor and noncoastal regions, many of which suffer from chronic severe water scarcity. Production of atmospheric water is another potential source of potable water. The earth's atmosphere is a huge and renewable water resource, containing approximately 12,900 billion tons of fresh water [5]-98% vapor and the rest in a condensed state (clouds and fog).

Study Area
The sampling station is located in northern Tel Aviv (Figure 1), near the campus of Tel Aviv University (32°07′09″ N; 34°48′17″ E), 39 m above sea level, 16 m above the ground. The station is located within the largest urban and industrial metropolitan area in Israel, 1 km from major roads and about 2 km from the Mediterranean coast. This site is adjacent to an air monitoring station operated by the Israeli Ministry of Environmental Protection that continuously monitors a wide range of meteorological and air parameters (data can be retrieved from: http://www.sviva.gov.il/English/env_topics/AirQuality/Pages/NationalAirMonitoringNetwork.asp x). Water samples were collected from June 2019 to March 2020 under different climatic conditions, at different hours and over several seasons.

AWG Apparatus
The sampling station is located in northern Tel Aviv (Figure 1), near the campus of Tel Aviv University (32°07′09″ N; 34°48′17″ E), 39 m above sea level, 16 m above the ground. The station is located 1 km from major roads and about 2 km from the Mediterranean coast.
The unique AWG was designed and built by WaterGen Ltd., an Israeli-based company that is a leader in the manufacture of commercial systems to produce water from the atmosphere. The apparatus was designed to meet the specific research requirements (e.g., does not contain any posttreatment technology) and is not used for commercial purposes.
The air-into-water ( Figure 2) process takes place as the air enters the AWG at ambient temperature. When it meets the evaporator, condensation occurs, and water is collected into the water tank. More specifically, from the evaporator, the vapor passes through a compressor where it reaches high pressure and temperature and condenses into a liquid state in the condenser by emitting the

AWG Apparatus
The sampling station is located in northern Tel Aviv (Figure 1), near the campus of Tel Aviv University (32 • 07 09" N; 34 • 48 17" E), 39 m above sea level, 16 m above the ground. The station is located 1 km from major roads and about 2 km from the Mediterranean coast.
The unique AWG was designed and built by WaterGen Ltd., an Israeli-based company that is a leader in the manufacture of commercial systems to produce water from the atmosphere. The apparatus was designed to meet the specific research requirements (e.g., does not contain any post-treatment technology) and is not used for commercial purposes. The air-into-water ( Figure 2) process takes place as the air enters the AWG at ambient temperature. When it meets the evaporator, condensation occurs, and water is collected into the water tank. More specifically, from the evaporator, the vapor passes through a compressor where it reaches high pressure and temperature and condenses into a liquid state in the condenser by emitting the heat back into the environment. The expansion process occurs at the expansion valve where the temperature and pressure of the vapor decrease. The vapor enters the evaporator where evaporation occurs naturally due to the reduction in pressure, and when ambient air contacts the vapor, the cycle starts again.
Water 2020, 12, x FOR PEER REVIEW  4 of 19 heat back into the environment. The expansion process occurs at the expansion valve where the temperature and pressure of the vapor decrease. The vapor enters the evaporator where evaporation occurs naturally due to the reduction in pressure, and when ambient air contacts the vapor, the cycle starts again. The unique AWG ( Figure 3) used in this research produces water at a minimum temperature of 10 °C and cools the ambient air to 8 °C below the dew point, with an average airflow of 400 m 3 /h and a compressor power of 1000 W. It has two 2 L water tanks that are completely separate from each other and the apparatus can be programmed to produce water at a specific time for each container. The device is 35 cm high, 25 cm long, and 14 cm wide.
It should be reiterated that the unique AWG does not have any air/water-filtration systems or water-treatment technologies.

Sampling Methods
Water samples were taken directly from the AWG, at the highest possible level of sterility and caution, to avoid external contamination. The water from the AWG container was collected into clean 2 L glass bottles. The bottles were closed and shaken. Water was then immediately portioned into the various test tubes as needed. The test tubes were stored in a 4 °C refrigerator until analyses.
Each water sampling was accompanied by two types of blanks that were subjected to all of the same processes. The first blank contained 900 mL distilled water in a 1 L glass bottle that was closed with a stopper. Its purpose was to check that there is no contamination in the process (test tubes, vials, or refrigerator) and that there is no contamination in the instrument used to perform the analysis. The second (environmental) blank contained 900 mL distilled water in a 3 L open glass beaker with a diameter of 20 cm. Its purpose was to investigate whether and how the environment affects the composition of the water, and to identify specific contaminants entering the water. The bottles were placed next to the apparatus throughout the experiment (Figure 3). The unique AWG (Figure 3) used in this research produces water at a minimum temperature of 10 • C and cools the ambient air to 8 • C below the dew point, with an average airflow of 400 m 3 /h and a compressor power of 1000 W. It has two 2 L water tanks that are completely separate from each other and the apparatus can be programmed to produce water at a specific time for each container. The device is 35 cm high, 25 cm long, and 14 cm wide.
It should be reiterated that the unique AWG does not have any air/water-filtration systems or water-treatment technologies.

Sampling Methods
Water samples were taken directly from the AWG, at the highest possible level of sterility and caution, to avoid external contamination. The water from the AWG container was collected into clean 2 L glass bottles. The bottles were closed and shaken. Water was then immediately portioned into the various test tubes as needed. The test tubes were stored in a 4 • C refrigerator until analyses.
Each water sampling was accompanied by two types of blanks that were subjected to all of the same processes. The first blank contained 900 mL distilled water in a 1 L glass bottle that was closed with a stopper. Its purpose was to check that there is no contamination in the process (test tubes, vials, or refrigerator) and that there is no contamination in the instrument used to perform the analysis. The second (environmental) blank contained 900 mL distilled water in a 3 L open glass beaker with a diameter of 20 cm. Its purpose was to investigate whether and how the environment affects the composition of the water, and to identify specific contaminants entering the water. The bottles were placed next to the apparatus throughout the experiment (Figure 3).

Analytical Methods
Metal composition was analyzed by inductively coupled plasma-optical emission spectrometry (ICP-OES; SPECTRO ARCOS series, Kleve, Germany), which was calibrated to detect and quantify a variety of metal elements in the range of several micrograms per liter to hundreds of milligrams per liter. The list of analyzed metals is given in Section 3.1.1. Sample preparation consisted of acidifying a 40-mL sample of water with 400 µL nitric acid (65%).
Major inorganic Ions were analyzed by ion chromatograph (IC; Integrion HPIC, Dionex Integrion, ThermoFisher Scientific, Waltham, MA, USA). The IC was calibrated to detect and quantify a variety of ions in the range of several micrograms per liter to several milligrams per liter. The IC column was a Dionex IonPac AERS 500, 4 × 250 mm, with a flow of 1 mL/min and mobile phase composition of: 0-11 min 23 mM KOH, 11-16 min 50 mM KOH, and 16-22 min 23 mM (KOH). Ammonium was analyzed using the ion-selective electrode (ISE) method with a Two-Channel Benchtop pH/mV/ISE meter with a Hanna HI4101 electrode. The limit of quantification (LOQ) of ammonium was 100 µg/L. Sample preparation consisted of placing 20 mL water produced by the machine in a tube and adding 1 drop of 32% HCl.
The mass spectrometer was set to selected ion monitoring (SIM) mode for both PAHs and PCBs, along with an internal standard (IS) that was used to improve the accuracy and precision of the quantitative analysis. The list of analyzed SVOCs is given in Appendix A. The calculations were performed in accordance to a calibration curve with a concentration range of 0.2-500 (µg/L) for each analyte (PCB and PAH). The correlation (R2) for all compounds was not less than 0.995. The sample was prepared by placing 100 mL water produced by the machine into a test tube and adding 5 g NaCl, 25 µL IS with a concentration of 50 mg/L, and 5 mL hexane; the tube was vigorously shaken manually. After phase separation, 1 mL from the top fraction (the organic phase) was placed into a GC vial.
Headspace (MPS 2XL autosampler, Gerstel, Linthicum, MD, USA) GC/MS technique was used to analyze VOCs. The GC/MS parameters were: 624Sil MS column (60 m, 250 µm, 1.4 µm, Restek, Bellefonte, PA, USA); oven temperature-initial 40 • C, hold for 2 min then raise to 240 • C at 8 • C/min; injection volume of 1000 µL; injector temperature of 220 • C; injection mode-splitless with purge time of 0.05 min; carrier gas-helium; transfer line of 280 • C. Headspace parameters were: incubation temperature of 80 • C, incubation time of 20 min, syringe temperature of 90 • C, agitator speed of 500 rpm and fill speed of 200 µL/s; 10 mL of sample was taken for the analysis. The mass spectrometer was set to SIM mode, and the IS used in the VOC analysis was fluorobenzene. The list of analyzed VOCs is given in Appendix B. The calculations were performed in accordance to the calibration curve with a concentration range of 0.1-500 (µg/L) for each analyte. R2 for all compounds was not less than 0.995.
pH was determined using a calibrated pH meter (MA235, Mettler Toledo, Columbus, OH, USA). NH 4 + was detected by a Hanna HI5222 selective electrode.

Overall Dew Water Quality Meets Drinking Water Standards
Dew water is the result of water vapor condensation. The dew collector (or substrate) and the atmosphere give the dew water unique chemical characteristics. We analyzed various chemicals in the dew water-metals, inorganic ions, NH 4 + , VOCs, and SVOCs. The results were compared to the drinking water guidelines of the [24,25] standards. A total of 64 water samples were collected from June 2019 to March 2020 under different climatic conditions, at different hours of the day and over several seasons. None of the measured chemicals (except nickel and benzo[a]pyrene in rare occasional samples), exceeded drinking water standards, although the standard deviation (SD) between samples was significant due to the varied climatic conditions. The median value for both blanks was less than the LOQ for the metals, ions, and NH 4 + . We had some concerns about the active dew collector inside the AWG apparatus becoming contaminated. However, examination showed this not to be the case, as the values on subsequent days were not affected by each other (Appendix C). Moreover, some chemical elements that constituted a major part of the device, such as iron, chromium, molybdenum, and aluminum, were not found at all or only in very small quantities (less than 5 µg/L) in the dew water.

Metals
Although a few of the examined metals are essential to human health, excess amounts of others can have severe negative health effects, and their quantity in the dew water must be monitored. A total of 24 metals were studied in this research ( Table 1). The concentrations of the common metals zinc, aluminum, and copper were an order of magnitude lower than the drinking water standards (Figure 4). Zinc exceeded the LOQ 58 times (n = 64), with a median value of 11 µg/L. The main source of zinc pollution is the abrasion of vehicle tires. However, previous studies have also shown the impact of several long-range transport processes on the characteristics of aerosol over Israel [26]. Thus, during the winter, zinc originated mostly from local (Israeli) sources (vehicles), whereas during the summer it was part of the long-range aerosol transport from southwestern Europe [27].  Copper (Cu) was detected in all dew water samples (median value of 140.5 µg/L). Its main sources in the atmosphere are metal production, industrial combustion of lubricants, coal combustion, and vehicular road dust originating from abrasion of vehicle parts [28,29]. Thus, copper concentration in the atmosphere is expected to be higher in urban areas than in rural ones. Other studies have also found that copper concentrations in dew water are lower than the drinking water standards by at least an order of magnitude. For example, dew water studies in an urban area near Paris, France [30], and an urban, semiarid area in Chile [31] found 15 µg/L and 38.5 µg/L copper (median values), respectively.
Nickel was the only analyzed metal that exceeded the drinking water standard in some analyses ( Figure 5). It exceeded the IL drinking water standards seven times (n = 64) with a maximum value of 677 µg/L, median value below the LOQ (5 µg/L), and a significant standard deviation of 101.05 µg/L. Note that the nickel concentration exceeded the LOQ value 23 times, while 6 out of 7 results that exceeded the drinking water threshold were sampled in July 2019. Nickel has an important function in enzyme activity. However, in large quantities-more than about 0.8 mg/day for adults-it can critically damage human health. It may cause lung cancer, erode the intestines, and more [32]. It is therefore important to monitor the quantities of nickel in dew water. Nickel is mostly derived from fuel oil combustion [26] and, therefore, apparently originated from vehicle emissions from the nearby traffic. Another potential source of nickel is sea water, despite a relatively low concentration (4.8 × 10 −4 mg/L) [33]. To the best of our knowledge, this is the first report on nickel concentrations in dew water. Because it exceeded the drinking water standards in a few instances, further studies should monitor its presence.    We also detected low concentrations of calcium (Ca 2+ ), magnesium (Mg 2+ ), sodium (Na + ), and potassium (K + ) ions, which are considered to be major ions in water geochemistry ( Figure 6). These metals ions do not have either WHO or IL drinking water thresholds. It should be noted that Mg 2+ deficiency (median value 35.5 µg/L, SD = 50.6) and Ca 2+ deficiency (median value 176 µg/L, SD = 446) in drinking water can cause various health problems, such as tooth loss, rickets, and cardiac infarction [32]. Therefore, the dew water should be supplemented with these essential minerals to a concentration of 25-50 mg/L [34], as recommended for desalinated water.   Other metals were found in negligible quantities (a few micrograms per liter, below their LOQ) and did not exceed the drinking water standards. There is no minimum required concentration for these substances in drinking water (Table 1). However, it is important to keep track of them as they can warn of potential contamination, possibly from the AWG (e.g., iron).
The metals in the water originate from both local emissions (transportation, industry, and marine aerosols) and long-range atmospheric transport. Nevertheless, the overall levels of the metals in the produced dew water were much lower than the WHO and IL drinking water standards.

Inorganic Ions
The concentrations of the seven ions that were analyzed in the dew water were far below the WHO or IL drinking water standards (Table 1). NO 3 − , SO 4 2− , and chloride (Cl − ) values were lower by two orders of magnitude than the drinking water standards, and nitrite (NO 2 − ) values by one order of magnitude.

Ammonium
Since NH 4 + in the drinking water does not have any immediate relevance to health, there is no standard value in the WHO or IL guidelines. Nevertheless, a threshold odor concentration for NH 4 + of about 1500 µg/L has been reported, and a value of 35,000 µg/L has been set for the taste threshold [24]. NH 4 + in dew water exceeded the LOQ value (100 µg/L) 57 out of 64 times (89%), with a median value of 757.53 µg/L (SD = 537.85). Therefore, NH 4 + concentration does not affect the quality of the water produced by the AWG.

VOCs and SVOCs
To the best of our knowledge, this is the first study of VOCs and SVOCs in dew water. The SVOC analysis was performed on 25 of the 64 water samples. Of all 32 analyzed compounds (Appendix A), only fluoranthene (C 16 H 10 ) and benzo[a]pyrene (C 20 H 12 ), which belong to the PAH group, exceeded the LOQ value in two dew samplings. Fluoranthene was found twice in the dew at 0.03 µg/L, but it does not have a standard value in the WHO or IL drinking water guidelines. The values of dew samples that exceeded the LOQ for benzo[a]pyrene were 0.67 and 0.19 µg/L, and the IL drinking water standard is 0.5 µg/L, thus, only one sample exceeded the IL (and WHO) drinking water standards.
Benzo[a]pyrene is formed when a material such as gasoline or wood burns incompletely and the residue sticks to the air particles. Its origin can be vehicle fuel emissions, especially from diesel engines [35], or the industry around the Tel Aviv metropolitan area [36]. As a byproduct of incomplete combustion and pyrolysis of carbon-containing fuels, benzo[a]pyrene has been identified in ambient air, surface water, and drinking water. Exposure to benzo[a]pyrene is associated with adverse birth outcomes and lung cancer [37], and, therefore, its presence in dew water should be further monitored to examine whether its rise above the drinking water standards was indeed a one-time event. Both fluoranthene and benzo[a]pyrene exceeded the LOQ values in the same dew samples, from 8 March 2020. However, climate conditions and air data did not show any extreme values on that day. In conclusion, except for this exceptional case, the dew water complied with the drinking water guidelines for the analyzed SVOC substances.
VOC analysis was performed on 25 of the 64 water samples. Of all 27 analyzed compounds (Appendix B), only chloroform (CHCl 3 ) exceeded the LOQ, once, with a value of 0.16 µg/L on March 2020. Nevertheless, neither IL (80 µg/L) nor WHO (300 µg/L) drinking water standards were exceeded. Anthropogenic sources of chloroform in the atmosphere could derive from plastics production, water-treatment plants, pesticide production, or hazardous substance-treatment facilities. Natural sources of chloroform are volcanic eruptions and emissions from marine algae and fungi [38]. Because the sampling site was in an urban area, the chloroform is likely to have originated from anthropogenic sources.

pH and Neutralization Factor
The pH values of the dew samples (n = 47) ranged from 6.5 to 7.9, with a median value of 7.3. While the WHO does not have a pH standard, the IL standard ranges between 6.5 and 9.5. pH values of dew water vary significantly across sites, ranging between 4.0 and 7.9, with wide daily variations due to the variable sources of ions. However, in most studies, the mean pH values usually remain fairly close to neutral [16]. The average pH of dew in the Israeli Negev desert is 7.4 [39], which is almost the same as the dew water pH found in this study.
The basic components (Mg 2+ , Ca 2+ , NH 4 + , and K + ) react with the acidic elements (NO 3 − and SO 4 2− ) in dew water. During these reactions, acidity is lowered, and water pH can become neutral. To determine the role of base components in the neutralization of nitric and sulfuric acids, a neutralization factor (NF) can be calculated according to the formula [16,40]: where [X] is the concentration of the elements responsible for neutralization (all concentrations are in units of µEq/L). The strength of neutralization decreases along the order NH 4 + (3.95) > Ca 2+ (1.35) > Mg 2+ (0.33) > K + (0.0). The high concentration of NH4+ compared to the other elements in dew water implies that it is the strongest neutralization factor, followed by Ca 2+ , Mg 2+ , and K + . Usually, NH 4 + or Ca 2+ are considered the most dominant neutralizing elements in dew water. Similar to our results, the neutralization effect of K + is nearly negligible in all studies [19,30,41].

Chemical Composition of Dew Water Produced during the Day and Night
Atmospheric characteristics can vary between day and night. During the day, the ambient temperature is higher, and the relative humidity is usually lower than at night because warmer air has the potential to hold more water vapor. During the daytime, more anthropogenic pollutants associated with industrial activity and traffic are emitted than at night. In addition, the daytime is characterized by photochemical reactions [42,43]. This is the first examination of the effect of time (day vs night) on the chemical composition of water produced by the AWG. The samples were divided into two groups: daytime-dew samples produced between 06:00 and 17:59 (n = 37) and night-time-dew samples produced between 18:00 and 05:59 (n = 27) ( Table 2). The hours are in local time (IST). The statistical analysis was performed with a two-tailed Mann-Whitney U test [44] (the nonparametric version of Student's t-test), because the data were comprised of independent samples. Despite the known different atmospheric characteristics between day and night, we did not find any significant difference between the time (day/night) of dew water production and its chemical characteristics, except for Ca 2+ . These results indicate that the AWG can produce potable water 24 h a day.

Conclusions
Our comprehensive analyses suggest that the production of dew water by an AWG in a large urban area can provide safe drinking water, during the day and at night, throughout the year. There was high variability in the concentrations of most substances between the water samples. Special attention and monitoring efforts should be payed to nickel and benzo[a]pyrene that were the only compounds that exceeded the drinking water standards. We demonstrate that the atmospheric water produced by AWGs can be a potential source of potable water, which may assist in dealing with the severe water scarcity existing across the globe, and specifically in remote and inland regions. The chemistry of the