The pH Sensing Properties of RF Sputtered RuO2 Thin-Film Prepared Using Different Ar/O2 Flow Ratio

The influence of the Ar/O2 gas ratio during radio frequency (RF) sputtering of the RuO2 sensing electrode on the pH sensing performance is investigated. The developed pH sensor consists in an RF sputtered ruthenium oxide thin-film sensing electrode, in conjunction with an electroplated Ag/AgCl reference electrode. The performance and characterization of the developed pH sensors in terms of sensitivity, response time, stability, reversibility, and hysteresis are investigated. Experimental results show that the pH sensor exhibits super-Nernstian slopes in the range of 64.33–73.83 mV/pH for Ar/O2 gas ratio between 10/0–7/3. In particular, the best pH sensing performance, in terms of sensitivity, response time, reversibility and hysteresis, is achieved when the Ar/O2 gas ratio is 8/2, at which a high sensitivity, a low hysteresis and a short response time are attained simultaneously.

sensing diameters 2 mm and 1 mm, respectively, as shown in Figure 1. RuO2 thin films were deposited onto the platinum sensing electrode area of the electrode cell utilizing an RF magnetron sputtering system (using shadow masking). A 99.95% pure, 4-inch diameter RuO2 sputtering target was used to sputter 300 nm RuO2 thin-films in Ar/O2 gas ratio varying from 10/0 to 7/3. After pumping down the sputtering chamber to a base presure of 1 × 10 −6 Torr, Ar and O2 gases were introduced into the chamber through mass flow controllers. The deposition process was carried out in pure Ar and Ar/O2 gas mixture of ratios ranging from 9/1 to 7/3 at 1 mTorr process pressure and 100 W RF sputtering power. The substrate temperature was kept low, i.e., the substrate was not heated during the film deposition.

pH Measurement
The potential difference between the electrodes of the pH sensor in real-time was recorded using high performance digital multimeter (Agilent 34410A, http://www.agilent.com). A unity-gain buffer amplifier was developed and used for impedance matching and signal loss minimization. The electrochemical behavior of the sensor was investigated through potentiometric measurements. The potential difference between the RuO2 sensing electrode and the Ag/AgCl reference electrode was recorded with 20 s intervals for test solutions of pH 4, pH 7 and pH 10 (Rowe Scientific, Australia) at 22 °C (temperature of our cleanroom). The proof-of-concept experiments were designed to demonstrate the effects of the Ar/O2 gas ratio on the performance of the pH sensor. The calibration of commercial pH sensors is typically validated using standard pH values of pH 4, pH 7 and pH 10. Therefore, in this preliminary study, the same buffer solutions were chosen to investigate the performance of the developed pH sensor. A magnetic stirrer was used to continuously stir the test solutions for quicker equilibrium of ion concentrations around the sensing electrodes. As a comparator, test solutions were also measured by an Oakton pH meter (pH 700 benchtop meter), and the sensor was cleaned after each pH measurement using deionized (DI) water rinse followed by N2 drying.

pH Sensing Mechanism
Fog and Buck have proposed five possible interpretations for the pH response mechanism of metal oxides including RuO2, with the most accepted interpretation being oxygen intercalation, which is represented by the following equilibrium reaction [14,30]: where δ is the oxygen intercalation, MOx is a metal oxide with higher oxidation state and MOx-δ is a metal oxide with lower oxidation state. Note that in Equation (1), oxygen intercalation was assumed with a proton activity in the liquid phase and an oxygen activity in the solid phase. Pourbaix diagrams have also been widely adopted for explaining the mechanism governing the redox equilibrium between two insoluble ruthenium oxides, which is represented by the following reaction [31]: Up until now, RuO2, as an electronically conductive oxide of rutile structure, has been commonly used in electrocatalysis applications [16][17][18]. It is well-known that the high catalytic activity of RuO2 is due to several factors, including high surface area, orientation of RuO2 molecules, and/or the RuO2 nanocrystal structure [19]. Moreover, the effect of surface morphology of ruthenium oxide hydrates has been investigated by measuring the point of zero charge and examining, through cyclic voltammetry, the diffusion of tritium protons in pores, cracks and along grain boundaries [32]. Trasatti [33] has suggested the following general equilibrium for the proton exchange for RuO2: The Nernst's mathematical equation predicting the potential between the sensing and reference electrodes versus the pH value is given by [26][27][28]: where E 0 is the reference electrode potential, R is the gas constant, T is absolute temperature and F is Faraday's constant. The entire term . is called the Nernst slope. The Nernst slope is 59.16 mV/pH at 25 °C.

Reactive Sputtering Model
The growth of the metal-oxide films by sputtering in a reactive environment may be described as follows: When the sputtering target is a metal oxide the oxygen partial pressure is originated from the internal (target) and external sources (gas flow). At steady-state conditions, the flux of oxygen atoms, f, is equal to the flux of metal atoms, because of the target homogeneity. In this case, the target supplies molecules into the sputtering chamber with a rate equals to fAt, where At is the target surface. On the other hand, the external source of oxygen increases the oxygen partial pressure in the chamber by a gain factor or Naq0/V0, where Na is Avogadro's number, V0 is one mole volume under normal conditions, q0 is the oxygen flow. The oxygen concentration inside the chamber decreases because of the oxygen deposition onto both walls of the chamber and the substrate (receiving surface). Using the molecular kinetic theory, the flux of oxygen due to a partial pressure is expressed as: where is the mass of the oxygen molecule, is Boltzmann's constant and is the absolute temperature. Note that, as the oxygen gas passes through the system it is removed from the reactive chamber by the system pump at a rate equals to SNa/P0V0 where P0 is the atmosphere pressure and S is the pumping speed. Kissine et al. [34] have particularly developed a model for tin oxide (SnO2) prepared by varying Ar/O2 gas ratio during RF sputtering, showing that, at steady-state condition, the balance of oxygen partial pressure in the vacuum chamber will be governed by the following equation [34]: where α and β are the parts of the receiving surface covered by Sn and SnO phases respectively, η is a coefficient which shows the relative adsorption of oxygen atoms onto the SnO phase compared to their adsorption onto the Sn phase, q = 1+q0Na /V0fAt is the normalized oxygen flow with respect to the flow from the target, σ = SNa/(2P0/(2πmkT) 1/2 )V0As is the pumping efficiency, Ψ FAs/fAt, As is the surface of the film deposited on the substrate. Kissine et al. [34] have also made two assumptions, that (i) the surface occupied by the Sn phase (α) increases when Sn is deposited onto the SnO and SnO2 phases, however, it decreases because of the oxidation process; and (ii), the surface occupied by the SnO2 phase increases when SnO is oxidized, but, it decreases because of the deposition of Sn. Thus, at equilibrium: where γ is the part of the receiving surface covered by the SnO2 phase. Furthermore, Kissine et al. [34] have simulated the SnOx film content versus the oxygen flow and concluded that the O:Sn ratio is equal to about 1:1 when the oxygen flow from an external source is rather small, i.e., the q value is below 1/10, resulting in an amorphous structure. When the oxygen flow (i.e., q) is increased, the SnO2 phase dominates. Note that, however, the crystal perfection and the presence of oxygen vacancies depend on the q value. Due to the similarity of Sn and Ru atoms, the above analysis can be used to accurately calculate the stoichiometry of RuO2 film prepared under RF sputtering in Ar/O2 gas mixture.

Sensitivity
The sensitivity of pH sensor was validated by immersing the sensor in pH buffer solution of pH 4, pH 7 and pH 10 at 22 °C. Each sensor was tested three times in the same buffer solution in order to investigate the Nernstian response. Figure 2 shows the average of three tests of the potential versus pH values for different oxygen percentages over RuO2 sputtering film. The sensor showed super-Nernstian slopes in all cases, starting from 64.33, 72.50, 69.83, and 73.92 mV/pH for Ar/O2 gas ratio of 10/0, 9/1, 8/2, 7/3, respectively. The super-Nernstian response of the sensor could be attributed to the use of an RuO2 thin-film prepared by RF magnetron sputtering [29,35]. The results shown in Figure 2 demonstrate linear regressions with high correlation coefficient r 2 values between 0.9983 and 0.9964. The pH sensor resolution for different Ar/O2 gas ratio was reported as inverse of the sensitivity in the unit of pH/mV.
The increase in sensor potential with decreasing the argon/oxygen ratio could be attributed to a decrease in free carrier density as well as carrier mobility in the RuO2 thin film [36]. Cui et al. have particularly confirmed that the value for carrier mobility and carrier concentration is very sensitive to the structure of the Indium tin oxide (ITO) films prepared with various Ar/O2 gas ratio [37]. Decreasing the Ar/O2 ratio or increasing the percentage of oxygen enhances the pore formation in the RuO2 thin film thus reducing the carrier density. Furthermore, the enhanced pore formation (or defects) with decreasing the Ar/O2 ratio increases the scattering of the charge carrier, thus decreasing the carrier mobility [37]. Furthermore, the RuO2 thin-film porosity increases with decreasing the argon/oxygen ratio, thus providing larger sensing surface volume. Figure 3 shows scanning electron microscope (SEM) images of the RuO2 thin-films, for different Ar/O2 gas ratios. Table 1 shows the pH sensitivity, resolution and linearity of the RuO2 sensing film sputtered using different Ar/O2 gas ratios.

Response Time
The response time of a potentiometric pH sensor is defined as the transit time required for its potential to reach 90% of an equilibrium value after immersing the sensor in a test solution [5,14]. To measure the response time, the potentials between the electrodes of the developed pH sensors were recorded for 10 min with 20 s intervals for buffer solutions of pH 4, pH 7 and pH 10 at 22 °C for different sensors employing RuO2 electrodes developed with various Ar/O2 gas ratios in the range 10/0 to 7/3. The response time for each sensor was calculated from the average of three test runs. For an Ar/O2 gas ratio of 10/0, the measured response times for pH 4, pH 7 and pH 10 were 3 s, 3 s and 160 s, respectively. For an Ar/O2 gas ratio of 9/1, the measured response times for pH 4, pH 7 and pH 10 were 3 s, 3 s and 60 s, respectively. However, when the Ar/O2 gas ratio varied from 8/2 to 7/3, the developed sensor displayed a much faster response time of 3 s for all the tested pH values of pH 4, pH 7 and pH 10. Note that, for the Ar/O2 ratios 8/2 and 7/3, the measured potential value reached 90% of the equilibrium potential value before the first 3 s of the 20 s interval. For the Ar/O2 ratios 10/0 and 9/1, the response time was not 3 s for all tested pH values. Table 2 shows the response time versus the Ar/O2 ratios for different pH values. Moreover, no change in the response time was displayed when the Ar/O2 gas ratio of the RuO2 film was further reduced below 7/3. These experiments demonstrated the importance of optimizing the Ar/O2 gas ratio for minimizing the response time of the pH sensor. Table 2 shows the response time versus the Ar/O2 ratios for different pH values.  0  3  3  160  9/1  3  3  60  8/2  3  3  3  7/3  3  3  3 Furthermore, the pH sensor exhibited a fast response time for acidic solutions, however, for an alkaline solution (pH 10) the response time of the pH sensor was dependent on the porous properties of the RuO2 sensing film, hence, on the Ar/O2 gas ratio. For an Ar/O2 gas ratio of 8/2, the average pore size is relatively large, compared to 10/0 Ar/O2 gas ratio, making it easier to trap the large-size OH − ions and this decreases the time needed to equilibrate the liquid in the pores of RuO2 sensing film. For an Ar/O2 gas ratio of 7/3, the response time was similar, however, the hysteresis was higher [38].

Stability
For stability testing, the pH sensor was immersed for 10 min three times in three different test solutions. For each 10 min interval the pH was periodically measured at a sampling time of 20 s. After each 10 min test, the sensor was cleaned with DI water followed by N2 drying. Figure 4 shows the measured potential (the average value from the three tests) versus time for different pH values. The sensor showed stable output potential for all pH values. The experimental results shown in Figure 4 demonstrate a stable pH sensor's response for an Ar/O2 gas ratio of 8/2.

Reversibility and Hysteresis
To investigate the reversibility of the pH sensor the pH was sequentially switched between pH 4, pH 7 and pH 10 in forward and backward orders at 300 s intervals without cleaning or drying the pH sensor, and the measured potential versus time was monitored. The switching cycle was repeated three times. Figure 5 shows the average recorded potentials (averaged over three tests) for different Ar/O2 gas ratio in the pH 4, pH 7 and pH 10 loops. The experimental results demonstrate an excellent reversibility and stable pH sensor's response for an Ar/O2 gas ratio of 8/2.
Typically, when the electrochemical potential of a pH solution is measured several times, different potentials are generated between the electrodes. This phenomenon is called hysteresis and is elaborated elsewhere [27,38]. Figure 6 shows the hysteresis voltage for different Ar/O2 gas ratios, for pH 4, pH 7 and pH 10 loops. It is noteworthy to mention that the acid loop hysteresis was smaller for the Ar/O2 gas ratio 8/2 compared to that for the Ar/O2 gas ratio 10/0. The increase in oxygen percentage increases the porosity of the RuO2 film as well as the overall sensing volume, thus resulting in faster ion diffusion through the sensing film. Tsai et al. have thoroughly investigated the effect of sensing area on the hysteresis for both acidic and alkaline loops [39].  According to Liao et al. [36], the potential different between the pH sensitive film and the electrolyte depends on the pH at the point with zero charge (pH ). This phenomenon is governed by the site-binding model described by the following equation [39]: where pH log p p is the pH value at the point of zero charge, ϕ is the surface potential of the electrolyte/insulator interface with respect to the electrolyte bulk, a and b are equilibrium constants, p a= −log ( a), p b= −log ( b), is the Boltzmann's constant, T is the temperature of the system in Kelvin, β is a dimensionless pH sensitivity parameter, given by [39]: is the surface site density, and is the double-layer capacitance and , are dissociation constants for potential determining ion and counter-ion surface reactions.
It is evident that the pH sensor structure prepared with an Ar/O2 gas ratio of 8/2 results in minimum hysteresis (<4 mV) for all the tested pH range (4-10).

Conclusions
In this paper, the effect of Ar/O2 gas ratio on the performance of RF sputtered RuO2 thin-film pH sensors has been experimentally investigated. Several 300 nm thin film RuO2 sensing electrodes, prepared by varying Ar/O2 gas ratio from 10/0 to 7/3 during RF sputtering, have been developed and their sensitivity, response time, stability, reversibility and hysteresis properties for pH sensing have been investigated. Experimental investigations have shown that an Ar/O2 gas ratio of 8/2 results in a RuO2 thin-film of excellent pH sensing properties, namely high sensitivity, low hysteresis and faster response, using a conventional RuO2 sputtering target. The optimized pH sensor structure has demonstrated a super-Nernstian response of 69.83 mV/pH, good stability and reversibility. The developed pH sensor can further be miniaturized as a lab-on-a-chip device and has application in biological analyses, water quality monitoring, chemical and environmental monitoring and in vivo clinical tests.