Technical Note: Calibration-Free pH Sensing of Ocean and Estuarine Waters

. An electrochemical based, all solid-state, calibration-free pH sensor is presented. The sensor is targeted for monitoring the pH of ocean and estuarine environments covering a salinity range from 10-35 ppt PSU [12] without the need for additional any [13] salinity measurements. The sensor performance is demonstrated in both laboratory and field conditions, 10 showing a pH of 8.33 in two sensors, against a measured pH of 8.34 from a freshly calibrated glass pH sensor. Excellent precision is shown, exhibiting a pH of 8.33 +/- 0.015 and 8.33 +/- 0.012 across multiple measurements. The field tests were conducted in an estuarine environment close to Oban in Scotland where the sensor was deployed for a period of three days. The sensor was validated against a sampled solution that measured a pH of 7.74, which was equivalent to the pH 7.74 obtained from ANB’s pH sensor at that exact time, and it was tested alongside a glass pH sensor during deployment, demonstrating its 15 stability over 3 days of testing. [11] The data highlighted the efficiency ability [14] of the sensor to monitor the tidal variations of pH in the estuarine environment.

3 suffer from reference drift and provide very accurate pH readings (Clayton and Byrne, 1993;Seidel et al., 2008;Rerolle et al., 65 2013;DeGrandpre et al., 2014). However, deployment issues around maintenance remain because of their need to mix the pH indicator dye with a seawater sample. This means they require deployment with optical dye bags which need to be replaced periodically and flowlines which can be blocked by biofouling species (Newhall et al., 2007;Rerolle et al, 2016;Chen et al., 2010).

70
In this paper, a new technology that exhibits several advantages from the technologies described above is presented. This relies on a redox active/pH active molecule that is combined with a redox active/pH inactive molecule to determine the pH of the seawater without the need of a calibration (Lu and Compton, 2014a). The main advantage of this novel technology is that any reference drift is tracked and accounted for in-situ, so it does not require any calibration. Moreover, the sensor does not require any special storage conditions and can be left dry, allowing for pH to be monitored in places where this has never been done 75 before, such as flood plain and tidal monitoring.

2.1 Laboratory Tests
Laboratory tests were conducted in temperature controlled stirred solutions. The solution was produced using synthetic sea salts (H2Ocean Natural Reef Salt, purchased from Maidenhead Aquatics, UK) diluted to the correct concentration using 80 deionised water. CO2 (BOC, 99%) was bubbled into the solution periodically to manipulate the pH. Temperature control was obtained through a temperature exchange loop placed inside a temperature-controlled chamber. The exchange loop temperature was manipulated through an external temperature-controlled bath (Fisher Scientific Isotemp 4100 R20 Refrigerated/Heated Bath Circulator) and held at 15°C for the time of the experiment. The solution was stirred using a Teflon overhead stirrer and the temperature measured using a k type thermocouple sheathed in polymer housing. The sensors were immersed in the tank 85 and connected to a control unit via a SubConn cable. The sensors were powered using a 12 V battery supply and each sensor used was switched on and off during the course of the experiment to ensure there was no bias in the experiment. The solution was sampled periodically throughout the course of the experiment and the pH measured using a calibrated glass ocean pH probe which was calibrated daily using pH 4, pH 7 and pH 10 standard buffers. [31] [37] The results of this are detailed in section 4 'Results and Discussion'. 90

2.2 Field Trials
The sensor was tested in Ardmucknish bay, located in Dunbeg, a village located outside Oban, Scotland ( To overcome the drawbacks of today's pH sensing devices an innovative approach is presented encompassing a solid-state, calibration-free sensor. The sensor has been developed to be plug and play to the end user, with little maintenance and no need for regular factory recalibrations. It has a range of 2-10 pH units and an accuracy of +/-0.05 pH units. Figure 1 2 shows the image of the sensor with the sensing element, or transducer, at the front end. This transducer is a replaceable part which can be easily replaced when required (vide infra), whilst the back end utilises a submersible connector through which power and 105 comms communications are passed to the control unit. On board the sensor is the electronic circuitry required to control the sensor and convert the raw measurement signal to the solution pH in a timestamped manner to the end user.

3.1 Transducer 120
This innovative calibration-free pH sensor is based on a voltammetric electrochemical technique, where a time-dependent potential is applied to each sensing electrode in turn and the resulting current is measured as a function of that potential (Bard, 2000). The pH is calculated using the resulting peak potential, as it is linearly associated with the pH [1]. The sensor utilizes a pH active molecule combined with a pH inactive molecule within a solid-state matrix (Batchelor-McAuley et al., 2011;Lu and Compton, 2014b). The combination of both The electrochemical response of both species [25] can determine the pH of 125 the seawater with no need for calibration, by tracking the performance of the reference electrode (pH inactive molecule) [26] through an additional in-situ electrochemical measurement (Dai et al., 2015a). As mentioned, the peak potential of the pH sensing electrodes moves with pH variation, allowing the determination of detection of the change in pH. At the same time, the pH inactive electrode tracks the reference drift, permitting an in-situ re-calibration of the electrodes, [27] and therefore, generating a calibration free pH measurement. Figure 2 3 shows how the peak potential moves with pH shifts to lower 130 potentials as the pH of the solution increases for the pH sensing electrodes ( Figure 23A) whilst the reference tracker potential stays stable ( Figure 3B), only changing in peak position when the reference electrode potential changes. Figure 3 was obtained by testing both pH sensing and reference trackers in synthetic seawater (H2Ocean Natural Reef Salt, Maidenhead Aquatics, UK). The pH of the solution was decreased by the addition of CO2, allowing the response of both the pH sensing and reference sensing elements to be monitored [3]. The pH was measured using a full ocean glass pH probe (Idronaut, Italy), which was 135 calibrated daily using standard buffers [6]. When the pH sensing electrodes and reference tracker are combined, subtracted values can be calculated, and based on the calibration algorithm shown in Fig. 3C, [28] the resulting pH can be determined.
The calibration of the sensor shows a sensitivity of 54.8 mV/pH unit in synthetic seawater, overlaying the calibration plot

155
The efficacy of this approach is shown in Figure 3 4. Figure 34A details the variation in peak potential of both the pH sensing and reference tracking systems as a function of time. After 10 h of deployment, a large change in peak potential was observed in the pH sensing system. Nevertheless, the reference tracker exhibited the same change in potential, meaning that the Ag/AgCl redox couple was impacted due to a change in the environment of the reference electrode.
[29] The tracking of the reference electrode allows the system to correct the change in the reference electrode potential in-situ. whilst Figure 34B shows the end 160 user result of pH as a function of time, displaying a pH 8.07 +/-0.04 over a 14 h experiment, [30] The data highlightsing the robustness of the sensing system where after 10 h of deployment there is a large change in the peak potential response of both the reference tracker and pH sensor, whilst no change is seen in the pH measurement owing to the reference electrode tracker correction [3]. Without the reference tracker this phenomenon may be wrongly attributed to a pH change by the end user. the reference electrode first, to track any changes in the reference electrode potential, and then the twelve pH sensitive electrodes in the transducer head, taking ca.15 seconds for each measurement.
[33] This data shows the reproducibility of the 180 two sensors taken in the same tank without any calibration prior to deployment. Across the entire set of measurements an average pH value of 8.33 +/-0.015 and 8.33 +/-0.012 for sensor 1 and sensor 2, respectively, were obtained. Without averaging the twelve pH measurements, the accuracy for each sensor was 8.33 +/-0. 04

3.2 Lifetime and Maintenance
The lifetime of the sensor is dependent on the number of measurements that each electrode records and not on the deployment time of the sensor. The electrodes in the transducer are repetitively cycled [35] and repetitive continual voltammetric measurements on each electrode ages the electrode causes a decay in the response of the electrode. This decay is seen through 200 a loss in the voltammetric peak size until the peak potential can no longer be deciphered. and maintenance is required. At this point the sensor notifies the end-user maintenance is required. It is for this reason that there are a number of electrodes on the front surface of the transducer of the sensor, (twelve electrodes are pH sensing, three are reference tracking and these are combined with two counter electrodes and a silver rod which acts as the conventional electrochemical reference electrode) in order to so as to maximise the lifetime of the sensor before maintenance is required. 205 The sensor has an onboard processor which assesses the voltammetric profile applies the one-time voltammetric sweep parameters to the electrochemical reaction occurring at each electrode. The resulting current received at the electrode due to the electrochemical reaction is then recorded, the data is analysed to determine the potential at which maximum current flows, which is then used to calculate the corresponding pH value.
[36] The response of each electrode is analysed after each measurement to ensure an accurate response is observed, when this is not the case the health number is increased, to alert the 210 end user that maintenance is required.
Maintenance of the sensor is through an abrasion of the transducer surface using an abrasion block (wet and dry paper, 320 grade). After the abrasion the sensor can be redeployed, and the health of the sensor returns to its original value. The transducer displays up to 15,000 measurements between abrasions and will last for approximately 25-30 abrasions before it will need 215 replacing. The sensor has been designed so that the replacement of the transducer is very simple for the end user.

3.3 Electronics and Communication
The sensor operates in either autonomous (stand-alone) or user-controlled mode through an RS232/485 communication protocol. In autonomous mode the sensor is connected to a power source and operates at a pre-set measurement rate. The maximum frequency for displaying a pH reading is approximately 15 sec, with the possibility of decreasing the measurement 220 rate depending on the end user requirements. The data is saved to an SD card and can be downloaded after deployment through a data retrieval unit via a USB connection to a PC.
In user-controlled mode the user defines the measurement rate and initiates the scan and shutdown commands. The data is returned to the control unit in real time. Table 1 summaries the supply voltage and current draw when the sensor is in various 225 modes. The pH output is calculated using an inbuilt algorithm based on the equation presented in Fig. 3C (y = -0.0548x + 0.6794 where x corresponds to pH and y to the peak potential (V)) [10], which combines a knowledge of the peak potential of both the pH sensing and reference tracking elements. As mentioned before, a potential is applied to each electrode and the resulting current is measured as a function of that potential, obtaining the resulting voltametric response, which will provide the peak 235 potential value needed for the calculation of the pH. and the temperature to produce the pH. Temperature is also needed for the algorithm to calculate the pH [9]. The temperature is measured using the onboard k type thermocouple. An example of the sensor output is shown in Fig. 5 6. After the system details (firmware version, serial number, date, time, scanning mode) the sensor outputs the date/time, pH, electrode number, temperature and health. As mentioned above, the health of the electrode corresponds to its condition and varies from 0 to 9. When the health of the majority of electrodes reaches 9, the transducer will 240 require an abrasion.

Laboratory Testing
Testing of the sensor in the laboratory has focused on lifetime stability and the ability of the sensor to respond in real time to 250 changes of pH in seawater solutions. Three different sensors were used to monitor the pH of a seawater solution for a week, and during this time CO2 was added sporadically bubbled in [39] to the solution sporadically to vary the pH, to pH 6.50, after which the seawater was left to degas between CO2 injections. Figure 7 details the plot of pH as a function of time for all three sensors. Each individual electrode's pH value has been presented instead of the average of the twelve, in order to improve the tracking of any small pH change. Initially only sensor 2 was tested, and after a day, sensor 1 was incorporated to the same 255 seawater/CO2 tank. Only two of the three sensors were tested in parallel at any one time: sensors 1 and 2 at the beginning, and sensor 3 was swapped with sensor 1 during the course of the experiment.
[8] The data of the three sensors showed favourable agreement between them, confirming the reproducibility between the sensors. They reacted immediately to the pH changes, allowing the end user to detect any spontaneous pH change in the environment. The sensor's measured pH values were validated using a calibrated full ocean pH sensor glass probe by measuring the pH of a small sample at certain times, assuming 260 the efficacy of the glass pH probe after being freshly calibrated. The pH of the sensor and the glass pH probe can only be compared when the seawater was in a stable condition, i.e. before adding CO2 or after letting it degas completely. In the first test, before CO2 was injected, sensor 2 presented a pH of 8.39 +/-0.31 for all individual twelve electrodes over a day in seawater, while the glass pH probe measured a pH of 8.42 when freshly calibrated.

Field Trials
The Field Trials were conducted over a period of three days (8 th -10 th September 2021), with a pre-set delay time of 15 min.
between measurements. In this case the sensor was placed in autonomous mode, powered directly from a 12 V power supply with no interface control. The data was saved to the internal memory of the sensor and downloaded after deployment. Figure   8 details the data obtained from the sensor through a plot of pH variation (purple) and temperature (black) as a function of time 280 over the 3 days. During the course of the trials, a water sample (ca. 250 mL) was taken in a plastic bottle and a pH of 7.73 was instantly measured using a calibrated glass pH electrode, which was equivalent to the pH 7.74 +/-0.047 obtained from ANB's pH sensor at that exact time. The pH of the sample was instantly measured in order to have the same temperature conditions as in the sea.
[7] The sensor presented a deviation of +/-0.048 pH units over the entire deployment, which is in accordance with the accuracy stated in the 'technical details' section. [4] [23] The data output from the sensor after 36 h of measuring on 285 a 15 min measurement interval showed the bi-daily semi-diurnal variations in pH ranging from 7.7 up to 7.9, consistent with the natural ebb and flow of the tidal waterways at Oban. At high tide, the river water mixes with the incoming seawater, producing a higher pH value. The pH variation with tides was quite noticeable, a result of the high tidal coefficient in those 3 days. Tidal coefficients are referred to as the difference in water height between high and low tides, on a scale of 20 to 120.
The sensor was deployed from the 8 th -10 th September when high tidal coefficients -101, 99 and 91 occurred (Oban, 2021). During the final day of the trial a glass electrode was deployed alongside the solid-state calibration free sensor. The 315 corresponding data comparing the solid-state calibration free sensor's response to the glass electrode is shown in Figure 9. It can be clearly seen the glass electrode displayed a very unstable pH reading was unstable during the course of its deployment, with values ranging between 7.50 and 7.90 in pH, endorsing the advantage of using solid-state the calibration-free sensor against the glass electrode for deployment in sea water.

Conclusions
The results demonstrate the use of a new solid-state, calibration-free pH sensor in the monitoring of estuarine and ocean waters.
The sensor was shown to respond effectively to pH in laboratory environments whereby the pH of the seawater solution was manipulated by the sporadic injection of CO2 into the mixed solution. Sensors were run in conjunction to each other showing the intra inter [43] reproducibility of the sensors and compared to a sample of solution whose pH was measured using a freshly 330 calibrated full ocean glass electrode. Excellent Favourable agreement was observed throughout the five seven-day experiment.
The field trials were conducted in an estuarine environment close to Oban in Scotland where the sensor was deployed for a period of three days. The sensor was validated against a sampled solution and tested alongside a glass pH sensor system. The data highlighted the efficiency ability [14] of the sensor to monitor the tidal variations of pH in the estuarine environment.