Freshwater quality detection is important for pollution control. Three important components of water quality are pH, ammonia and dissolved H2S and there is an urgent need for a high-precision sensor for simultaneous and continuous measurement. In this study, all-solid-state electrodes of Eh, pH, NH4+ and S2− were manufactured and mounted to a wireless chemical sensor with multiple parameters. Calibration indicated that the pH electrode had a Nernst response with slope of 53.174 mV; the NH4+ electrode had a detection limit of 10−5 mol/L (Nernst response slope of 53.56 mV between 10−1 to 10−4 mol/L). Ag/Ag2S has a detection limit of 10−7 mol/L (Nernst response slope of 28.439 mV). The sensor was cylindrical and small with low power consumption and low storage demand to achieve continuous in-situ monitoring for long periods. The sensor was tested for 10 days in streams at Trawsgoed Dairy farm in Aberystwyth, UK. At the intensively farmed Trawsgoed, the concentration of NH4+ in the stream rose sharply after the application of slurry to adjacent fields. Further, the stream was overhung with extensive vegetation and exhibited changes in pH, which correlated with photosynthetic activity. Measurements of S2− were stable throughout the week. Our data demonstrate the applicability of our multiple electrode sensor.
Globally, agriculture is the largest user of fresh water but is also a source of contamination through chemical runoff. Within freshwater, pH, ammonia and dissolved H2S are three important indicators of quality. pH is an important measurement of water chemistry and reflects both characteristics of the underlying bedrock and the amount of plant growth and organic material. The latter reflects the release of CO2 as a result of the decomposition of organic matter and also active respiration, both of which lead to the formation of carbonic acid (H2CO3). Plant and algal growth also reflect run-off of nitrogen (N)-fertilizers, whether synthetic or as biological wastes, leading to the eutrophication of water courses (Ip et al. 2004). These N inputs are reduced to free ammonia (Xu et al. 2008). Dissolved H2S can originate from various sources (coal-based power plants, natural-gas processing, refineries, smelter operation) and can be regarded as a major air pollutant entering the atmosphere and causing acid rain (Primavera et al. 1998). Continuous measurement of dissolved H2S in water is in need to monitor any accidental releases, particularly from industrial sources (Miloshova et al. 2003). Therefore, monitoring pH, dissolved H2S and ammonia, ideally continuously, is very important for the agricultural water environment, drinking water and fresh water. In situ is a particularly attractive target as some ions in water solution are unstable and easily change to other forms.
The traditional methods for ion measurement include ion chromatography (Rey et al. 1998), fluorometry, spectrophotometry (Goyal et al. 1988) and titration (He et al. 2000). However these methods are time-consuming and cannot be used for continuous in-situ monitoring. Ion selective electrodes (ISEs), which give a potentiometric signal that obeys the Nernst equation, are currently attracting considerable attention (Mikhelson 2013). ISEs have a high selectivity to certain ions, and are able to detect low concentrations. Further, ISEs are applicable for in-situ monitoring due to their quick response time (Michalska 2012). The most widely used ISEs are glass electrodes (Gonçalves et al. 2011; Mikhelson 2013), which contain an internal reference solution. However, these are difficult to miniaturize and are easy to damage (Chen et al. 2007; Cheng et al. 2011). The development of all-solid-state electrodes represents a considerable recent improvement (Cattrall & Freiser 1971; Mikhelson 2013). These have a lower resistivity and are smaller than glass electrodes, and have eliminated the need for an internal reference solution. The high performance of all-solid-state electrodes makes it possible to integrate these other electrodes for long-term continuous in-situ monitoring in the water environment.
In this study, we describe the construction of all-solid-state electrodes for Eh, pH, NH4+ and S2−, mounted to a wireless chemical sensor with multiple parameters. We demonstrate the sensor's efficacy in the continuous in-situ measuring of Eh, pH, NH4+ and S2−in freshwater streams in an active agricultural farm (Trawsgoed, UK; 52°20′38″N, 3°57′4″W). Thus, extensive use of this sensor could allow the monitoring of agricultural runoff in varying environments.
Electrode construction and sensor design
Ir/Ir(OH)x-pH electrode preparation and laboratory test
The Ir wires were ultrasonically cleaned in an HCl bath and then rinsed in deionized water. In the three-electrode system of the CHI760D electrochemical workstation, the Ir electrode was used as the working electrode, the Ag/AgCl electrode as the reference electrode and the Pt electrode as the auxiliary electrode. Cyclic voltammetry (CV), with scanning in 5% LiOH solution (0.1 M), was used to form the Ir electrode with a layer of Ir(OH)x film (Zhang et al. 2017).
pH buffer solutions of 4.00, 6.86, 9.18 were used to calibrate the Ir/Ir(OH)x-pH electrode. The pH electrode was calibrated using a CHI760D electrochemical workstation.
NH4+ electrode preparation and laboratory test
An Ag wire was polished with alumina powder and ultrasonically cleaned in an HCl bath. It was then rinsed in deionized water and dried in the open air. A nano-Ag layer was electroplated to the Ag wire for better conductivity. The anode and the cathode of the electroplating system were all Ag wires. A solution of 0.1 M AgNO3 was used as the electrolyte.
The conductive polyaniline (PANI) layer was formed by the three-electrode system. The CV method was used in 0.1 M aniline hydrochloride. Then the wire was dipped in a saturated solution of copolymer aniline (CPANI) and 2,5-dimethoxyanilineas as the second layer. Finally, ammonium ionophore I (6.9%), potassium tetrakis (4-chlorophenyl)borate (0.7%) and 2-nitrophenyloctyl ether (92.4%) was dissolved in N,N-dimethyl formamide (DMF) and the wire dipped into this solution to form the third layer (Huang et al. 2015).
A series of NH4Cl solutions from 1 × 10−5 to 1 × 10−1 M were prepared. The NH4Cl solution consisted of 0.5349 g NH4Cl and was made up to 100 mL in a volumetric flask with deionized water to obtain a 1 × 10−1 M NH4Cl solution. 1 × 10−2 M to 1 × 10−5 M NH4Cl solution were stepwise diluted by the 1 × 10−1 M standard NH4Cl solution with deionized water. The NH4+ electrode was calibrated using a CHI760D electrochemical workstation.
Ag/Ag2S electrode preparation and laboratory testing
A 3 cm Ag wire was polished and electroplated with a nano-Ag layer using the same method as described in 2.2 and then immersed to the 0.1 M Na2S solution for 10 min. Ag2S formed the outermost layer of the electrode (Ding et al. 2015).
A series of Na2S solutions ranging from 1 × 10−7 to 1 × 10−1 M were prepared. To avoid S2− oxidation, a sulfur antioxidant buffer (SAOB) was prepared in place of deionized water when preparing the Na2S solution. Stocks of 1 L of SAOB contained 80 g of NaOH, 35 g of ascorbic acid (VC), 67 g of EDTA-2Na, and 35 g of NaCl. The Na2S solution consisted of 0.7804 g Na2S and was made up to 100 mL in a volumetric flask with SAOB to obtain a 1 × 10−1 M Na2S solution. 1 × 10−2 M to 1 × 10−7 M Na2S solutions were stepwise diluted by the 1 × 10−1 M standard Na2S solution with SAOB. The Ag/Ag2S electrode was calibrated using a CHI760D electrochemical workstation.
Multiple-parameter sensor design
Figure 1 shows the structural diagram of the chemical sensor. The outer shape is a cylinder with a length of 230 mm and 75 mm in diameter. The outer covering is made up of 30CrMnSi, which is a compound metal that confers added strength to protect the sensor against environmental impacts. The sensor contains a communication port, a detector part outside and a circuit board and a battery. The pH electrode, NH4+ electrode, Ag/Ag2S electrode, Eh electrode and the Ag/AgCl reference electrode were integrated into the detector part (Figure 2(a)). The Eh electrode was made of a platinum wire and the Ag/AgCl reference electrode was prepared by the melting method (Zhong & Yexiang 1998). To better connect the electrode to the detector part, the platinum wire and Ir wire were connected to the Ag wire by gas welding. The exposed metal parts, which the connect electrode and electric wire, were wrapped individually by shrinkable tube and an epoxy resin layer to avoid corrosion and short circuiting.
When the sensor is in use and submerged in water, a semi-hollow cap covers the detector part. The semi-open design of the detector part not only guarantees full contact between the electrodes and water but also prevents the chemical membrane being damaged by rapid rushing water and greatly avoids biofouling.
Figure 3 illustrates the working principle of the data acquisition system. First, the chemical sensors detect the chemical signals and convert chemical signals into electrical signals, which then go through the filter, are amplified into the main processor, and stored in a Flash memory function. The watertight connector may be linked to the host computer for real-time communication. The whole data collection can be controlled by the host computer's software. The built-in clock and switch chip of the processor were calibrated by the host computer software, which also initiates the sensor's work mode to collect signals. The multi-parameter sensor can then be disconnected from the host computer at the start of the work mode of the processor and used for in-situ observations. At the end of the observation period, the multi-parameter sensor and the host computer can be reconnected to read the stored data through the watertight connector.
Figure 2(b) shows the structure of the circuit board. Each electrode was connected to a certain port from ‘CON 1’ to ‘CON 4’ with a reference electrode at the center, which was connected to ‘CON 8’. ‘RS232’ is the communication interface, connecting to signal lines. According to the circuit board design, the real-time data can be stored in a Flash chip when carrying out fieldwork and retrieved later after connecting to software. The frequency of data collecting can be set at a minimum of 1 s or greater intervals. Therefore, the sensor can be used in multiple environments without constraints of distance as the sensor works and stores data independently without connecting to a host computer. The two system mode, sleep mode and work mode, consume 1.0 mA and 3.5 mA of electric current respectively. The input impedance of the system is 1013 Ω (25°C), which guarantees long-term use with two AA Li batteries. The data are stored in a Flash chip with 128 M bit storage capacity as internal storage. If measuring frequency is set as once per 1 minute, the storage can guarantee 2 years' monitoring. When the storage is full, new data will cover old data automatically. Due to its lower power consumption and low storage demand, the sensor can achieve continuous in-situ monitoring for at least 6 months. The circuit board and the batteries are sealed in the cylinder of 30CrMnSi, and are not harmful to the surrounding environment.
The life cycle of the whole system is about 2 years; it needs service to change electrodes and batteries every 6 months.
Polyaniline (PANI) and silver nano particles, which are bacteriostatic, are used for preparation of NH4+ and Ag2S electrodes respectively (Tian 2010; Xiu et al. 2012). The 0.25 mm diameter Ir wire used for the pH electrode is thinner and hard to adhere. This avoids the problem of biofouling over its lifetime.
Field trial locations
A 10-day fieldwork was conducted between September 17–26th September 2016 at Trawsgoed Farm, Aberystwyth, UK. Trawsgoed Farm is the location of a 350-cow commercial dairy herd. As such, the farm produces a large volume of slurry from its livestock. Several ditches stretch across the farm and feed into the River Ystwyth some 28 km from its mouth. Given this potential for wide-ranging impact on local water quality, it is appropriate to engage in monitoring streams around this farm.
The multiple-parameter sensor, which can detect Eh, pH, NH4+ and S2−, was positioned in the ditch of the farm (52°20′38″N, 3°57′4″W). The position has been indicated as a red spot in Figure 4. The ditch was about 0.5 m deep and 1.0 m wide, surrounded with green plants. During the field trial, slurry was applied by shallow injection into the farmland adjacent to this stream on 18 September and 19 September. As an independent measure of pH, this was measured at a position close to the sensor using a Thermo Scientific Orion, ROSS pH Electrode. pH was measured at around midday on 19, 20, 21, 22, 23, 24, 25 and 26 September.
To understand the daily variation of chemical parameters, the sensor was set to collect data every 1 minute during the field trial. The sensor was taken back and connected to the host computer to extract the stored data at the end of each period of fieldwork. The obtained potential value was converted to a concentration value based on the calibration result. To reduce the effect of any abnormal data, the average of every half an hour was calculated based on the obtained 1 minute measurements.
RESULTS AND DISCUSSION
Figure 5(a) shows the potential-pH response of the Ir/Ir(OH)x-pH electrode. The slope of the fitting line is 53.174, close to the Nernst theoretical value. The correlation coefficient at R2 > 0.996 indicated good linearity when fitting the calibration curve. Figure 5(b) shows the calibration results of the NH4+ electrode in NH4Cl solution from 10−5 to 10−1 M. The potential gradient from 10−5 to 10−4 was a little smaller compared with others, the slope of the fitting line from 10−4 to 10−1 M was 53.56 with a coefficient R2 > 0.998, which is also close to the Nernst theoretical value. The Ag/Ag2S electrode has a high detection limit. Figure 5(c) shows the calibration results in Na2S solution from 10−7 to 10−1 M. The fitting line was 28.439 with a coefficient R2 > 0.997.
Measuring environmental pH value change
As an initial indication of sensor performance, the pH detected by the sensor at Trawsgoed farm was compared to that detected by a glass pH electrode with an internal reference solution. The pH values for the measuring period ranged between 6.5 and 7.4 (Figure 6(a)). Table 1 compared the result of pH value tested by pH probe and sensor, the latter is shown as red spots in Figure 4. The average difference of pH value between the pH probe and sensor was 0.18, which reveals the accuracy of the pH electrode of the sensor. In Table 1, the pH given by the probe is in the range between 6.6 to 7.36, while that given by the sensor is between 6.89 to 7.20. It seems that the latter is more stable.
|Time||Temperature (°C)||pH||Variation (%)|
|Time||Temperature (°C)||pH||Variation (%)|
apH value given by Thermo Scientific Orion, ROSS pH electrode.
bpH value given by the sensor.
Examining the sensor derived pH values over a diurnal cycle indicated a regular daytime rhythm (Figure 7). It decreased to the lowest value at around pH 6.5 at ∼6:40 AM in the morning and increased to a peak at around 7.2 at 14:00 PM in the afternoon. Given that the sensor was located in an area of dense vegetation, this feature almost certainly reflected the shifting balance between respiration and photosynthesis. At night, plants absorb oxygen and produce carbon dioxide because of respiration to increase of H2CO3 in water resulting in a decrease of pH. During daylight, plants absorb carbon dioxide and produce oxygen because of photosynthesis, leading to a relative decline in H2CO3 and increase in pH. The fact that this pattern was easily detected by the sensor is an indicator of its sensitivity in the detection of pH.
NH4+ concentration change
The concentration of NH4+ at Trawsgoed farm (Figure 6(c)) proved to be stable at ∼1.0 × 10−5 M for the first two days of the measuring period. However, immediately after the injection of slurry into neighboring fields, the NH4+ concentration rose sharply to 1.6 × 10−2 M (19 September, 2016). During the following days, this concentration appeared to be being diluted due to the continual flow of water in the area, so that the NH4+ concentration fell progressively to 3.9 × 10−4 M by 21 September. There was a rapid decrease to 1.6 × 10−5 M that correlated with a rain shower on the evening of 21 September that would have appeared to accelerate the dilution process. These results clearly indicate that the application of slurry increases the concentration of NH4+ in the stream and most likely in the Ystwyth River to which it is a tributary.
The concentration of S2− fluctuated slightly but averaged to ∼1.0 × 10−6 M (Figure 6(d)). This was quite a low concentration, which the application of slurry did not significantly change in the water.
All solid-state electrodes of Eh, pH, NH4+ and S2− were manufactured and mounted to a wireless chemical sensor with multiple parameters. Under laboratory conditions, we demonstrated continuous and simultaneous measurement of Eh, pH, NH4+ and S2−.
The sensor was field tested in streams at Trawsgoed farm in Aberystwyth, UK. The result in Trawsgoed indicates that pH value exhibits regular daytime fluctuations due to plant metabolism. The average difference in pH value between a commercial pH probe and the constructed sensor was only 0.18. The trends of pH and Eh stay fitted with the known negative correlation functional relationship between Eh and pH. The concentration of NH4+ rose sharply when slurry was applied to neighboring fields.
From the above, the electrodes of the sensor show high accuracy and sensitivity to certain ions. Regarding the features of small volume, less power consumption and low storage demand, the sensor will have a broad application in water environment detection.
This work was supported by the Open Research Fund of Laboratory of Marine Ecosystem and Biochemistry, SOA (LMEB201701) and UK-China Joint Research and Innovation Partnership Fund (PhD Placement Programme). The authors would like to thank the staff at Trawsgoed Farm sites (Aberystwyth University) for their help during the measuring periods.