Recent Progress in Electrochemical pH-Sensing Materials and Configurations for Biomedical Applications

Review pubs.acs.org/CR Cite This: Chem. Rev. XXXX, XXX, XXX−XXX Recent Progress in Electrochemical pH-Sensing Materials and Configurations for Biomedical Applications M. T. Ghoneim,† A. Nguyen,‡ N. Dereje,§ J. Huang,# G. C. Moore,⊥ P. J. Murzynowski,∥ and C. Dagdeviren*,† Downloaded via MASSACHUSETTS INST OF TECHNOLOGY on March 26, 2019 at 13:14:40 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. † MIT Media Lab, §Department of Mechanical Engineering, and #Department of Aeronautics and Astronautics, Massachusetts Institute of Technology (MIT), Cambridge, Massachusetts 02139, United States ‡ Department of Biological Engineering, ⊥Department of Materials Science and Engineering, and ∥Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology (MIT), Cambridge, Massachusetts 02142, United States ABSTRACT: pH-sensing materials and configurations are rapidly evolving toward exciting new applications, especially those in biomedical applications. In this review, we highlight rapid progress in electrochemical pH sensors over the past decade (2008− 2018) with an emphasis on key considerations, such as materials selection, system configurations, and testing protocols. In addition to recent progress in optical pH sensors, our main focus in this review is on electromechanical pH sensors due to their significant advances, especially in biomedical applications. We summarize developments of electrochemical pH sensors that by virtue of their optimized material chemistries (from metal oxides to polymers) and geometrical features (from thin films to quantum dots) enable their adoption in biomedical applications. We further present an overview of necessary sensing standards and protocols. Standards ensure the establishment of consistent protocols, facilitating collective understanding of results and building on the current state. Furthermore, they enable objective benchmarking of various pH-sensing reports, materials, and systems, which is critical for the overall progression and development of the field. Additionally, we list critical issues in recent literary reporting and suggest various methods for objective benchmarking. pH regulation in the human body and state-of-the-art pH sensors (from ex vivo to in vivo) are compared for suitability in biomedical applications. We conclude our review by (i) identifying challenges that need to be overcome in electrochemical pH sensing and (ii) providing an outlook on future research along with insights, in which the integration of various pH sensors with advanced electronics can provide a new platform for the development of novel technologies for disease diagnostics and prevention. CONTENTS 1. Introduction 2. The Power of Hydrogen (pH) 2.1. Definition, Importance, and Analytical Formulation 2.2. Temperature Effect 3. Materials for Electrochemical pH Sensors 3.1. Overview 3.2. Thin Films and Nanostructures 3.2.1. Metal Oxides Thin Films 3.2.2. Polymers 3.2.3. Nanorods 3.2.4. Nanotubes 3.3. Summary and Conclusions 4. pH-Sensing Configurations 4.1. Ion Sensitive Field Effect Transistor (ISFET) 4.2. Extended Gate Field Effect Transistor (EGFET) 4.3. Interdigitated Electrodes (IDEs) 4.3.1. Hybrid IDEs 4.3.2. Capacitance IDEs 4.4. Resistance Variation 4.5. Summary and Conclusions 5. Sensing Standards and Protocols © XXXX American Chemical Society 5.1. 5.2. 5.3. 5.4. Inherent Properties of Components Input Resistance of Characterization Systems Surface Cleaning Surface Resetting (Intermittent Cleaning vs In Situ Discussion) 5.5. Time Plots and Analysis 5.6. Critical Point (Pc) for Response and Drift Determination 6. pH Regulation in the Human Body 6.1. Cells 6.2. Kidneys and Lungs 6.3. Blood 7. pH Sensing in Biomedical Applications 7.1. Ex Vivo 7.1.1. Urine Tests 7.1.2. Saliva Tests 7.1.3. Tooth Decay 7.2. In Vivo 7.2.1. Glioblastoma 7.2.2. Intracellular and Extracellular pH B C C C D D D D E H H I I I I K K K L M M M P P P R T V V V V W W W W W Y Y Z Received: October 30, 2018 A DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews 7.2.3. Oral Hygiene 7.2.4. Ischemia 7.2.5. Sweat Analysis 8. Status Quo 8.1. Wearable pH-Sensing Systems 8.2. Implantable pH Sensing Systems 9. Challenges 9.1. Stability of pH-Sensing Devices 9.2. Repeatability of pH-Sensing Devices 9.2.1. Mixed Versus Specific Reactions 9.3. Reproducibility of pH-Sensing Devices 9.4. Modeling of pH-Sensing Devices 10. Future Outlook on pH Sensing in Biomedical Applications Author Information Corresponding Author ORCID Notes Biographies Acknowledgments References Review Therefore, the sensing of this essential parameter is of prime interest in current biomedical research. However, biological systems are extremely complex and constitute a myriad of chemicals and interactions. It is the act of balancing these interactions between chemicals that sustains life. This balance is achieved through equilibrium states that mandate the rates of reactions and proper activity of various fluids, and consequently the proper pH value when H+ is concerned. To this end, pH inevitably plays a role in balancing and altering these equilibriums. At a macromolecular scale, nucleic acids and proteins contain proton dissociable groups, which interact with the pH of the direct environment. Specifically, enzymes essential to catalysisfunction within a specific pH range and can begin to denature at the extremes of this range. At the cellular level, the cell environment is buffered to maintain a consistent equilibrium within the cell; for example, the cytoplasm regulates under a phosphate buffer system.4 Entire systems are also affected by pH, such as the circulatory system with blood regulated by a bicarbonate buffer system.5 The excellent buffering ability of biological systems not only helps maintain proper equilibrium and pH ranges, but can also reliably indicate anomalies and diseases when deviations occur. Tumor cell detection is one such example. Tumors induce reduced vasculature and thus oxygen, which increases the rate of anaerobic energy production and promotes a significantly more acidic environment than neighboring tissue.6 This results from the H+ donating capacity of the byproducts of anaerobic energy production, such as lactic acid, which in turn increases local H+ activity. Lactic acid is an Arrhenius acid (i.e., dissociates in water/aqueous solutions to give H+), which consequently increases the acidity (i.e., activity of H+) and lowers the pH of body fluids. Therefore, tumor tissue can be differentiated, and its progression and growth can be monitored by monitoring the pH. Methods of monitoring pH within biological systems, however, can vary depending on the situational needs and restrictions. Given the importance and strict regulation of pH in biological systems, pH sensors research has attracted the interest of many researchers. Figure 1a depicts the trend in the number of Scopus database listed publications over the past decade with “pH sensor” in the title and biomedical applications mentioned in the manuscript text. This review provides an overview of pH sensors based on their material systems, sensing configuration, operating principles, and their suitability for biomedical applications. The regulation of pH in the human body and representative biomedical pH sensors are also discussed. Finally, state-of-the-art pH sensors are compared for suitability in biomedical applications, and insights, challenges, and future outlook are provided. The review is organized as depicted in Figure 1b. Sections 1 and 2 introduce the topic and basic definitions; Sections 3 and 4 focus on materials for pH sensors and discuss pH-sensing configurations and techniques; Section 5 discusses standards and protocols for pH-sensing systems; Sections 6 and 7 present a debrief on pH regulation in the human body, followed by highlights of specific examples on pH sensing in biomedical applications; Sections 8 and 9 discuss the status quo of wearable and implantable pH sensors, and the common challenges facing pH-sensing systems; finally, Section 10 provides a future outlook on pH-sensing systems in biomedical applications. Z AB AC AD AD AF AI AI AJ AK AL AM AO AP AP AP AP AP AQ AQ 1. INTRODUCTION In 1889, Herman Walther Nernst postulated that the ion concentration could be measured using electrode potential. This foundation, paired with Arrhenius’s definition of an acid as a proton donor, paved the way for a term specifically designated to describe hydrogen ion (H+)/hydronium ion (H3O+)/proton concentration. pH was first defined in 1909 by Soren Peder Lauritz Sorenson in conjunction with his novel acid colorimetric assay, which used a hydrogen electrode paired with a calomel reference electrode (RE).1 The RE maintains a constant potential, while the hydrogen electrode builds up a potential proportional to H+ concentration in a solution. The potential difference measured across the two-electrode system changes with the pH of the solution. While Sorenson’s assay failed to break into the field dominated by more inexpensive and less accurate pH paper sensors, his pH term has become an essential component of modern lexicons.1 Originally defined as the negative logarithm base 10 of the H+ concentration, pH has since been modified to be the negative logarithm base 10 of H+ activity.2 This amendment stems from interaction of ions within a solution, which can cause some ions to deviate from ideal behavior and effectively appear inactive. To account for this phenomenon, ion activity (also referred to as effective ion concentration) is used in the definition instead of concentration. Despite Sorenson’s definition and attempts to popularize electrodes in pH measurement, the glass electrode and acidimeter were the true developments that issued a new era of pH measurements. The glass electrode, invented by Duncan McInnes and Malcolm Dole in the 1920s, was capable of specific ion detection by means of a doped glass membrane.3 In addition, the acidimeter, developed by Arnold O. Beckham, enabled acid strength detection.3 These advances enabled accurate pH measurements and opened new routes for engineering even better sensors. Along with the well-defined term for H+ activity and progress in its measurement, the role of pH in biological systems has become more evident. The regulation of pH is essential to maintaining healthy equilibrium in biological environments to support life. Disturbances and variations in pH can be either the cause or effect of disease and dysfunction within a biological system. B DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review the Faraday constant. Hence, the difference between the solution under the test’s pH and the reference solution’s pH (i.e., pH(X) − pH(S)) can be linked to the measured EMF difference (i.e., EX − ES). pH as a function of H+ activity is now the most commonly accepted definition because of electrode dependence on ion activity. The Nernst equation links the H+ activity to pH.9 Another reason H+ activity presents a better mode of measurement is the dependency of pH on temperature. This dependency is best explained by the activity of H+ varying in a directly proportional manner with temperature. Though the history of the definition of pH has seen many changes and modifications, today’s measurements are based on the latest definition. Developing new pH-sensing techniques and applications based on previous developments must be unhindered by any confusion surrounding the measurements. 2.2. Temperature Effect The temperature dependency of pH measurements affects the consistency of results. Variations in temperature are known to cause changes in solution viscosity and ion mobility. Overall temperature can have two main effects on pH sensing: (i) reducing electrode accuracy and measurement speed and (ii) changing the results due to the coefficient of temperature variation of the material itself.10 A number of sources can cause electrode variations, including effects on the electrode sensitivity, isothermal point calibration, thermal and chemical equilibrium, and membrane resistance. H+ activity, as defined by the Nernst equation, varies with a temperature-dependent Nernstian slope constant. This effect is typically compensated for by the initial sample temperature. The intersection point of the calibration lines of differing temperatures, also known as the isothermal point, is ideally represented by the zero potential point. However, as real electrodes possess different coefficients of temperature variation and all contribute to the total potential, the isothermal point often deviates from the ideal situation. Imbalances in thermal equilibrium can also result in pH measurement drifts over time and can be corrected for by using temperature insensitive electrode materials or a carefully maintained thermal environment.11 Chemical equilibriums at the electrode/electrolyte interface are also affected by thermal variation, given that temperature can affect the solubility of the metal salt, leading to slow response and drift. The glass membrane resistance of pH-sensing electrodes increases with decreasing temperature, causing sluggish response, and at extremes, complete dysfunction. Taking into account these temperature effects will thus lead to greater accuracy of pH measurements and more repeatable results, and extend their applications in biomedical fields. Particularly for biomedical applications of pH sensors, temperature can greatly affect results. As demonstrated by the experiments performed by Rosenthaul et al., blood pH varies linearly with temperature.12 The reported dependence coefficients were −0.0147 pH/°C and −0.0118 pH/°C for human blood and plasma, respectively. In addition to solution pH variation with temperature, pH sensors also demonstrate temperature dependence for biomedical application. Huang et al.’s review of pH-sensing iridium oxide film demonstrates how temperature can affect measurements and how this effect can be predicted in practical applications.13 The investigation showed the intrinsic and predictable dependence of the Nernstian potential on temperature based on the Nernst equation by recording pH and temperature for four buffer solutions and temperatures. Figure 2 presents the actual measurements, with Figure 1. Trends in pH-sensing research and the organization. (a) Trend in the number of publications including “pH sensor” in the title (black squares) and “biomedical applications” in the abstract or text of the manuscript (red circles) between 2008 and 2018, collected from Scopus database. (b) The review content follows the depicted flowchart clockwise from (1) to (5). 2. THE POWER OF HYDROGEN (pH) 2.1. Definition, Importance, and Analytical Formulation When pH was first defined by Sorenson in 1909, he based his calculations on electromotive force measurements which could be used in conjunction with the Gibbs energy equation (eq 1): ΔG = ΔG° + RT ln Q (1) 0 ΔG is the Gibbs energy change, ΔG is Gibbs energy change under standard state, and R, T, and Q are the gas constant, absolute temperature, and the reaction quotient, respectively.7 Using the electromotive force, Sorenson defined pH as the negative logarithm base 10 of H+ concentration. With the advent of Lewis’s concept of ion activity, the Gibbs energy equation involved in Sorenson’s calculations was modified to substitute H+ activities for the previous H+ concentration values. In 1932, two main general definitions pervaded the scientific community: pH is equal to the negative logarithm base 10 of (i) concentration of H+ and (ii) activity of H+. In 1948 both definitions fell to criticism.8 Despite these issues, pH measurements were practiced under eq 2 (where pH is a function of the H+ activity): i RT yz zz pH(X) − pH(S) = (E X − ES)/jjj (2) k F ln 10 { where an unknown solution (X) with pH equal to pH(X) and an electromotive force (EMF) of EX is measured against a reference solution (S) of known pH value of pH(S) and an EMF of ES. F is C DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review H+ ions accumulate on the ISFET’s gate dielectric, the resulting electric field modulates the current in the channel. Although these measurements can drift over time and render ISFET devices progressively less sensitive, novel fabrication techniques allow a smaller size and glass-less structure.18 Interdigitated electrodes take advantage of new micro- and nanofabrication techniques to maximize the surface area to volume ratio, therefore maximizing the sensitivity of the biosensor.19 Additionally, the small size reduces material costs and power consumption. The configurations of pH-sensing systems are discussed in detail in Section 4. Nonetheless, it is the extensive materials library available with pH-sensing capabilities that made these configurations possible. Figure 2. Theoretical temperature dependence of a representative IrOx pH sensor showing −0.3, −0.8, −1.3, and −2 mV/°C at pH = 2, 4, 7, and 10, respectively.13 Reproduced with permission from ref 13. Copyright 2011 Elsevier. 3.2. Thin Films and Nanostructures 3.2.1. Metal Oxides Thin Films. pH sensors have also been developed with numerous thin film metal oxides, including ZnO, PtO2, PbO2, IrO2, Sb2O3, RuO2, TiO2, Ta2O5, WO3, RhO2, OsO2, PdO, CuO, and SnO2.20−27 These materials interface with the electrolyte results in the accumulation of H+ and hydroxide ions (OH−)20 and find applications in the electrochemical measurement of pH because of the charged nature of these ions. In such configurations, a support material is typically coated with the metal oxide to create a durable electrode.20 For example, a sensor developed with Ta2O5 has demonstrated Nernstian pH sensitivity (∼−56.19 mV/pH) in the pH range from 1 to 10, without suffering errors caused by acid corrosion.28 Moreover, atomic force microscopy (AFM) has shown that Ta2O5 has a smooth and uniform surface, without cracks or crystal grains after annealing, which is essential to a capacitancebased sensor, with an impedance characteristic that varies little in different pH solutions.28 However, out of all the possible pHsensing metal oxides mentioned, IrO2, ZnO, and doped ZnO thin films have manifested the most desirable qualities, such as high sensitivity and biocompatibility, and, consequently, most reported metal-based pH sensors utilize these materials. IrO2 pH sensors can be created with high sensitivity and even super-Nernstian response, a sensitivity higher than −59 mV/pH at room temperature.22 In addition, IrO2 sensors achieve stability over a large range of pH values, temperatures, pressures, and media, as well as minimal potential drift, excellent chemical selectivity, durability, no requirement for pretreatment, and are biocompatible.13,20,23,29 Furthermore, electroplated iridium is cost-effective, precise, and produces reproducible results.30−34 Another oxide, ZnO, is a transparent semiconductor with a direct band gap (Eg = 3.37 eV) and a large exciton binding energy (60 meV).21 Additional attractive qualities for pH sensing include its biosafety and biocompatibility, importance as a nanomaterial for integration with microsystems and biotechnology, polar and nonpolar surfaces, and the ability to signal each time a H+ binds to its surface because of its conductivity.35,36 Surface charge will develop when an electrolyte interacts with ZnO through physical adsorption of ions or charged species on the surface.36,37 ZnO also has a distinct amphoteric nature due to a high density of binding sites for H+ and OH−.37,38 In the presence of high concentration of H+, the diffusion of H+ leads to a higher surface potential, while a high concentration of OH− causes ZnO to give up a proton to OH− and create a lower surface potential.37 Moreover, the diverse and abundant ZnO nanostructures, such as nanowires (NWs),39−41 nanoflakes (NFs),42 nanobelts,43,44 nanobows,45 and nanohelices,46 open novel designs and applications for ZnO, which varying dependence of pH on temperature ranging from −0.3 to −2.0 mV/°C for pH 2−10, respectively; the clear relationship between the two highlights the temperature effect on pH sensors and films.13 To extend this example and apply it to pH sensing within an organism, potentially acidic tumor cell detection would necessarily account for the temperature dependence of the sensor to prevent confusing healthy cells with diseased ones. Indeed, these temperature effects in pH sensing necessitate temperature correction for accurate pH measurements. Due to the strong dependence of pH on temperature, recent studies aimed at measuring both temperature and pH at the same time.14,15 For instance, Zhang et al. reported on nanosensors for simultaneous monitoring of lysosomal pH and temperature, indicating the need for the temperature measurement to calibrate the pH measurement.14 3. MATERIALS FOR ELECTROCHEMICAL pH SENSORS 3.1. Overview Measurements within biological systems often demand special considerations, such as preserving the life of the organism or a necessary microscopic scalability, and their effect on the methodology of sensing. These considerations determine the suitability of various pH-sensing materials and their target application. Using indicator dye that is sensitive to pH and covalently attached to reagent paper presents an inexpensive and quick way of testing biological fluid.16 Electrochemical methodbased pH sensors span from the glass electrode of the early 1900s to the modern extended gate field effect transistor (EGFET) and ion sensitive field effect transistor (ISFET) configurations. The glass electrode pH sensor compares the potential of known to unknown H+ using a RE and a sensing half-cell. Though an accurate and reliable method, the glass electrode suffers in its need for repeated calibration and fragile construction, making it difficult to miniaturize and use effectively in vivo.16 EGFET pH sensors use the physical protonation and deprotonation reactions that cause a difference in the surface potentials at the interface between the electrolyte solution and the extended gate of a transistor. The resulting electric field modifies the conductance of the field effect transistor (FET), and the current flowing in the channel between the source and the drain terminals is used to measure pH.17 The ISFET is another technique where the whole transistor is immersed in the solution and its gate dielectric is exposed, i.e., replacing the transistor gate with the electrolyte. As D DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Figure 3. pH-sensing materials. (a) Potentiometric time-trace of a pH bandage sensor from pH 8.51 to 2.69, with sensitivity of −58.5 mV/pH and response time less than 20 s. Inset shows a digital image of the printed potentiometric sensor on an adhesive bandage. Electropolymerization of polyaniline (PANI) onto the printed carbon was achieved to act as the working electrode, and the deposited polyvinyl butyral (PVB)-based membrane acts as the reference electrode (RE).59 (b) Long-term stability of PANI sensors measured at pH 5 and 7 buffer solutions showing drift values of 0.64 mV/h and 0.49 mV/h, respectively.60 (c) Schematic of a flexible sensor array containing Ca2+, pH, and temperature sensors patterned on a flexible polyethylene terephthalate (PET) substrate. The inset shows a photograph of a flexible sensor array.68 (d) SEM image of dried inverse hydrogel opal.73 (e) Spherical hydrogel schematic, showing shrinking and swelling, r0 is the initial radius, r∞ is the maximum radius at swelling equilibrium, and Dcoop is the cooperative diffusion coefficient accounting for solvent diffusion and consequent polymer chains motion. (f) Response of poly(vinyl alcohol)poly(acrylic acid) (PVA−PAA) hydrogel quartz crystal microbalance sensor at 3−10 pH values, with sensitivity of 13.2 kHz/pH and swelling and shrinking times of 500 and 800 ms, respectively.75 (g) Normalized intensity of 1432 cm at 3−8 pH values. For each specific pH value, the surface enhanced Raman spectroscopy (SERS) measurements were performed 10 times, and the average results were adopted, with error bars representing the standard derivation.76 (h) Plots of the intensity ratio I1438/I1069 as a function of pH, in the 3−8 range, under 1% O2, 5% O2, 10% O2, and 15% O2 conditions.77 (i) Calibration curves of single-walled carbon nanotubes (SWCNTs) pH-sensing electrodes on glass in pH 3−11, showing sensitivity of −48.1 mV/pH, −36.2 mV/pH, −22.6 mV/pH, and −16.4 mV/pH for 200, 80, 20, and 5 passes, respectively.78 Reproduced with permission from refs 59, 60, 68, 73, 75, 76, 77, and 78. Copyright 2014 John Wiley and Sons, Inc. Copyright 2017 Elsevier. Copyright 2016 American Chemical Society. Copyright 2010 Elsevier. Copyright 2004 Elsevier. Copyright 2011 American Chemical Society. Copyright 2016 American Chemical Society. Copyright 2016 Elsevier. can easily be altered by slightly modifying the conditions for preparation.38,47−52 ZnO is often n-type in its natural state,53 but doping can be utilized to adjust ZnO conductivity for different purposes.54 Iron, for example, has shown multiple potential benefits as a dopant material through its use in controlling the electrical conductivity, energy band structure, and carrier concentration of ZnO. Furthermore, iron doping of ZnO results in the reduction of ZnO nanostructure dissolution rates.55,56 Aluminum is another common dopant for ZnO and results in increased pH sensitivity.57,58 One of the main problems with doping is that the dopant may disrupt material morphology, but arrays of wellaligned In doped ZnO nanorods (In:ZnO) have been reported.53 3.2.2. Polymers. Besides inorganic pH-sensing oxides, the ion-exchanging ability seen in conductive polymers serves well for potentiometric sensors and has earned them considerable attention when developing pH sensors.16 Yet, the type of polymer chosen depends on the application along with sensitivity and selectivity requirements.16 A popular polymer used is polyaniline (PANI) because of its high conductivity, ease of synthesis, and stability.16 Andrade et al. developed a potentiometric sensor embedded into an adhesive bandage using PANI as the working electrode and polyvinyl butyral polymer (PVB) as the reference electrode (Figure 3a, inset).59 The working electrode is where the reaction of interest takes place, resulting in a reduction potential that is pH dependent, and the reference electrode is an electrode of known stability and known potential in the pH range of interest. A potentiometric sensor uses the relation between test solution’s pH and the difference in reduction potentials between the working and reference electrodes. The results showed a Nernstian response of −58 mV/pH between pH 4.35 and 8 with a response time of less than 20 s (Figure 3a). After 1000 bending cycles, the bandage still performed well at −58.5 mV/ pH. Another flexible and thin pH sensor based on a PANI array was developed by Yoon et al., and its performance closely matched a commercial pH meter.60 Within a pH range of 2.38− 11.61, it demonstrated a linear Nernstian response of −60.3 mV/pH with a response time of less than one second. At a pH of E DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews 4−7 2−9.5 relative standard deviation of responses less than 4% after 5 h, and lost 5% of its sensitivity after a month drift of 0.7 mV/h, 1.1% error in pH value in 1 h 1−11 <10 s Ag/AgCl Nyein et al.68 OCP poly(vinyl chloride) (PVC) and n-cetylpyridinium hexafluorophosphate (CPFP) incorporated with quinhydrone (QH) PANI OCP Ping et al.71 graphene-PANI (Gr-PANI) amperometric Sha et al.70 PVB coated Ag/AgCl yes average slope of −62.5 mV/pH 4.35−8 <20 s electropolymerized PANI OCP Guinovart et al.59 polyvinyl butyral (PVB) polymer Ag/AgCl yes polyaniline (PANI) nanopillar array Yoon et al.72 120 mg of lithium perchlorate (LiClO4) and 10 μL of pyrrole (PPy) dissolved in 5 mL of acetonitrile Ag/AgCl pH range 2.38−11.61 drift of 0.64 mV/h at pH 5, 0.49 mV/h at pH 7 (from 5 to 12 h) yields a stable signal in less than 20 s <1 s linear Nernstian response of −60.3 mV/pH Nernstian response of −58 mV/pH and post 100 bending cylces: −58.5 mV/ph −50.14 μA/pH·cm2 in pH 1−5, and 139.2 μA/pH.cm2 in pH 7−11 −57.5 mV/pH drift of 0.25 mV/day <1 s followed a Nernstian response (∼ −60 mV/pH) yes stability F Ag response time sensitivity where the swelling characteristic time constant, τ, is found through eq 5, reference material biocompatibility (4) open circuit potential (OCP) OCP r(t ) = r0 + (r∞(max) − r0)e−t / τ sensing material and shrinking may be found through the eq 4, setup (3) ref Table 1. Summary of the Key Developments in Polymeric pH Sensors r(t ) = r0 + (r∞(max) − r0)(1 − e−t / τ ) Korostynska et al.16 5, it showed a potential drift of 0.64 mV/h and 0.49 mV/h at pH 7 from 5 to 12 h (Figure 3b).60 Flexible pH sensors are, especially, appealing for their integration ability with logic, memory, and other sensing devices from the growing field of flexible electronics61−67 for fully flexible wearable systems (Section 8.1). PANI has also been used in the development of a wearable electrochemical device with the purpose of continuous monitoring of Ca2+ and pH in body fluids, as illustrated in Figure 3c.68 This device had an average slope of −62.5 mV/pH when tested at pH 4−7 and a potential drift of 0.7 mV/h over 1 h, resulting in a 1.1% error in pH value.68 The biocompatibility of PANI was studied, and the results showed that PANI does not induce skin irritation or provoke any sensitization.69 Further studies have been performed on the combination of multiple polymers. For instance, PANI has proved useful in enhancing the performance of other materials, such as graphene (Gr).70 An amperometric sensor using the fabricated Gr-PANI composite demonstrated a shorter response time with an improved sensitivity at −50.14 μA/(pH·cm2) between pH 1−5 and 139.2 μA/(pH·cm2) between pH 7−11.70 A pH-sensing membrane was developed by incorporating the ionic ncetylpyridinium hexafluorophosphate (CPFP) and poly(vinyl chloride) with quinhydrone (QH).71 In the pH range of 2−9.5, the sensor showed a sensitivity of −57.5 mV/pH and a response time of less than 10 s.71 After a month, the sensor lost only 5% of its sensitivity.71 Another potentiometric pH sensor was developed by coating a platinum electrode with 0.5 μm thick mix of 120 mg of LiClO4 and 10 μL of pyrrole dissolved in 5 mL of acetonitrile.16 In the pH range 2−11, the electrode exhibited a response time of less than one second and a drift of 0.25 mV/ day. Conductive polymers offer a good option for pH-sensing materials, due to their ion-exchanging properties. Many polymers have been studied; the most popular being PANI. Table 1 summarizes key developments in polymeric pH sensors. Evidently, the table shows excellent sensitivity (−57.5 to −62.5 mV/pH) and stability (0.25−0.7 mV/h drift) for polymeric materials as pH sensors, with PANI in open circuit potential (OCP) as the most common system. 3.2.2.1. Hydrogels. There are many types of polymers, and stimuli-responsive hydrogels are a special class of them. In response to stimuli, they can characteristically alter their volume, absorbing, and releasing amounts of aqueous solution.79 As polymers with cross-linked molecule chains, they are very useful for detecting changes in temperatures, light, and even pH. Figure 3d shows a scanning electron microscopy (SEM) image of a dried inverse hydrogel opal. A key characteristic of hydrogels is their swelling and shrinking properties. On the basis of the Tanaka-Fillmore theory, when uninfluenced by the surrounding, the behavior of swelling for a spherical hydrogel (as represented in Figure 3e) can be found through eq 3, 2−11 Review DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews G 2−12 ∼0.66 nm/pH deviation of ±0.07 units pH over 50 min 1−11 4.43−8.07 0.07/pH white emission poly(vinyl alcohol)/poly(acrylic acid) (PVA/PAA) hydrogel 74 mV/pH nanofibers bovine serum albumin, coumarin 460, fluorescein, and 5(6)-carboxy- slope of 0.16/ x-rhodamine hydrogel pH unit <2 s Mishra et al.95 Shaibani et al.96 Benson et al.91l surface enhanced Raman spectroscopy (SERS) fiber optics potentiometric 1−12 3 sensors after 10 days display sensitivities of 4.282, 4.279, and 4.280 nm/pH in the lower pH region rise time: 24 s and fall time: 20 s 13 nm/pH acrylamide, N,N′-methylene diacrylamide, N,N,N,Ntetramethylethylenediamine and methacrylic acid hydrogel alginate solution, poly(diallyldimethylammonium chloride), poly(sodium 4-styrenesulfonate), and D-(+)-glucono-1,5-lactone acrylamide, bis(acrylamide) solutions, and methacrylic acid hydrogel fiber optics sensing material Table 2. Summary of the Key Variables of Hydrogel pH Sensors sensitivity response time stability and r is the final radius, r∞max is the maximal radius in the swelling equilibrium, r0 is the initial radius, t is the time, and Dcoop is the cooperative diffusion characteristic.74 However, a more complex model must be used when considering the surrounding environment. Swelling kinetics and properties greatly influence the response time of a hydrogel pH sensor. For instance, Richter et al. showed that swelling for a quartz crystal microbalance sensor, coated with poly(vinyl alcohol)-poly(acrylic acid) (PVA−PAA) hydrogel, had a short response time of 500 ms due to its high ionic strength, a shrinking time of 800 ms, and sensitivity of 13.2 kHz/ pH, in the 3−10 pH range.75 Figure 3f shows the response time from the experiment. In order to apply the hydrogel properties in pH sensors, sensor transducers (either optical, oscillating, or conductometric) are used to provide electrical signals from the hydrogel’s swelling properties.74 Although significant progress has been made in optical pH sensors over the past few decades,80−89 this review is dedicated to electrochemical pH sensorsdue to the extensiveness and importance of the topic. One example of a conductometric transducer is found in Sheppard et al.’s hydrogel-coated interdigitated electrode array, where resistance was shown to decrease with an increase in conductivity as the hydrogel swelled.90 Hydrogels are advantageously shown to be extremely sensitive (up to 10−5 pH units), inexpensive, efficient, and have diverse functions that are useful for a variety of applications.91 However, the disadvantage of hydrogels is that they have a small working range74 and require complex setups. Richter et al. found that initial readings from hydrogels may be inaccurate.92 Despite the disadvantages, hydrogels are still a promising material for pH sensing, as they have unique properties that allow the sensors to be ultrasensitive and utilized in many circumstances. Table 2 provides a summary of hydrogel pH sensors with important reported parameters such as sensitivity, response time, and stability. The table shows the excellent ability of hydrogels’ swelling and shrinking in the 500− 800 ms range, respectively. However, swelling of 120 s and shrinking of 130 s have also been reported, for a different type of hydrogels, highlighting the strong dependence of the response on the hydrogel’s composition. In addition, the versatility of hydrogels enables its usage with many configurations, including potentiometric, interdigitated electrodes (Section 4.3), and resistance variation (Section 4.4). Consequently, the sensitivity of hydrogel sensors can be quantified as a frequency/wavelength shift (kHz/pH or nm/pH), voltage difference (mV/pH), and change in resistance (Ω/pH), adding to the versatility of the material. For reference, the uncommon setups, such as quartz crystal microbalance, magnetoelastic, Raman spectroscopy, and White emission, mentioned in Table 2, are briefly explained in this paragraph. The other common setups, such as interdigitated electrodes, and resistance-based sensors are explained in Section 4. The process for quartz crystal microbalance uses microgravimetric transducer principles to measure changes in frequency as pH changes.75 Another sensing technique is through magnetoelastic sensing, which measures pH sensitivity through changes in resonance frequency. In this technique, a pickup coil detects a magnetic flux causes by mechanical deformations of the sensor, which occur as a result of magnetic field impulses.79 For surface enhanced Raman spectroscopy Zhao et al.93 You et al.94 pH range (5) setup r2 Dcoop ref τ= Review DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 3. Summary of the Representative Works on Nanorod Based pH Sensors ref Al-Hilli et al.101 Chen et al.98 Ma et al.77 Lee et al.99 Zong et al.76 sensing material reference material setup ZnO nanorod Ag/AgCl open circuit potential (OCP) iridium nitride nanorod Ag/AgCl extended gate field effect transistor (EGFET) surface enhance Raman spectroscopy (SERS) ion sensitive field effect transistor (ISFET) Raman scattering 4-nitrothiophenol on gold nanorods ZnO-based nanorod/gaterecessed AlGaN/GaN gold nanorods Ag/AgCl p-aminothiophenol biocompatibility yes sensitivity −59 mV/pH at room temp 1−14 current: 26 μA/pH voltage: −22.66 mV/pH 4−10 yes yes yes pH range 3−8 −57.66 mV/pH 4−12 3−8 different O2 concentrations (Figure 3h).77 The performance of an EGFET pH sensor fabricated with 1-D iridium nitrate nanorods was also investigated. However, over a pH range of 4− 10, the sensor exhibited a sensitivity level of only −22.66 mV/ pH, a value much lower than the theoretical limit.98 Table 3 summarizes key works on nanorod-based pH sensors. The table shows that even functionalized nanorods have relatively limited pH range, compared to polymeric thin films (Section 3.2.2) in general and hydrogels (Section 3.2.2.1) specifically, except for ZnO nanorods that showed a sensitivity of −59 mV/pH in the 1−14 pH range. 3.2.4. Nanotubes. Besides nanorods, researchers have taken a step further to increase pH-sensing capabilities by using nanotubes.78,103−107 Nanotubes have a hollow center allowing solutions to come into contact with almost twice the amount of surface area it would on a nanorod. As a result, they are much more sensitive to their surrounding environment.103 Nanotubes also display high mechanical stability, mass production capability, ease of chemical functionalization, and adjustable electrical properties, making them a better candidate for pHsensing material.78 Carbon nanotubes, specifically, have been extensively studied for pH sensing. Li et al. proposed a microfluidic pH-sensing chip that was developed based on single-walled carbon nanotube thin films (CNTFs).104 The developed chip performed at a sensitivity of −59.71 mV/pH with a standard deviation of 1.5 mV/pH in a pH range of 3−11.104 The results of the experiments show the chip is suitable for practical uses and allow the detection of metabolic processes in cells.104 Qin et al. developed an inkjet printing process to deposit single-wall carbon nanotubes for pH sensing.78 Their results found that thicker films can be effective sensing materials in potentiometric electrodes.78 With 5 passes (<20 nm thick film), the sensitivity of the device was only −16.4 mV/pH compared to −48.1 mV/pH for one developed with 200 passes (∼700 nm thick film)78 (Figure 3i). Moreover, testing has been done on the modification of single-walled carbon nanotubes. Tsai et al. proposed oxygenplasma-functionalized CNTFs on polyimide substrates as the sensing material for an EGFET sensor. 107 The study demonstrated a sensitivity of −55.7 mV/pH in a pH range of 1−13 for plasma treated CNTF compared to an as-sprayed sensitivity of only −37.6 mV/pH.107 Gou et al. developed a sensor based on oxidized single-walled carbon nanotubes functionalized with the conductive polymer poly(1-aminoanthracene).105 Using a chemiresistor, they were able to produce a sensitive pH response that approached the Nernst limit, without the need for a RE.105 The development of multiwalled carbon nanotubes is further studied by several researchers. For instance, Jung et al. investigated the pH-sensing characteristics of a multiwalled (SERS), the surface plasmon resonance (SPR), where incident light causes conduction electrons at the interface to oscillate resonantly, near particular metal surfaces creates an enhanced electromagnetic field. This in turn increases the Raman scattering intensity significantly.94 Noteworthy, assessing Raman shifts via a single point can be misleading, due to nonuniformities and irregularities. Instead, a Raman map of a reasonable area would be more objective for Raman peaks and shifts observations.97 The pH-sensitive Raman molecule (MBA) is affected when a decreasing peak intensity ratio is created from a decreasing pH.94 The last method is white emission. With changes in pH, white-emitting hydrogels can change from white to a nonwhite color and allow for the detection of pH changes from the intensity ratios of principal color components.91 Because hydrogels can be used in a variety of setups, they have gained considerable attention like metal oxides and other nanostructures. 3.2.3. Nanorods. Compared to metal oxides and polymers, nanorods have gained more attention for pH sensing due to their higher surface-to-volume ratio, a characteristic imperative for improving the sensitivity of pH sensing.98 The nanorods can have a significant effect on sensing performance when applied to different testing setups. Different sensing materials have been studied; the most popular being ZnO due to its chemical stability, nontoxicity, electrochemical activity, fast response, and low costs.99,100 Other nanorod materials investigated were gold, iridium nitrate, and tungsten oxide. Al-Hilli et al. explored the electrochemical potential response of ZnO nanorods between a pH of 1 and 14. The results showed a sensitivity of −59 mV/pH at room temperature, performing better than a ZnO EGFET.101 A similar study explored the sensing characteristics of the ZnO-based nanorod using gaterecessed AlGaN/GaN ISFETs.99 The developed biosensor exhibited a sensitivity of −57.66 mV/pH in a pH range of 4− 12.99 The performance was attributed to the larger sensing area from combining ZnO nanorods and AlGaN/GaN.99 Another popular nanorod material studied is gold (Au) nanorods. One study demonstrated hydrochloric acid (HCl) treated gold nanorods (GNRs) as an intracellular pH (pHi) sensor based on the SERS method.76 The results, in the 3−8 pH range, are shown in Figure 3g.76 The study found that by reducing the cytotoxicity of GNRs with HCl treatment bioapplications become possible.76 Experimentation has also been done on GNRs coated with 4-nitrothiophenol as a SERS nanoprobe. The purpose of this experiment was to report a new nanoprobe for pH sensing under different levels of hypoxia by SERS.77 Hypoxia is a condition of low oxygen levels that can detrimentally affect cells and tissues.102 The nanoprobe proved to be effective in measuring between 4.5 and 7.5 pH values at H DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 4. Summary of the Representative Works on Nanotube-Based pH Sensors reference material ref sensing material Li et al.104 Gou et al.105 Tsai et al.107 single-walled carbon nanotubes (SWCNTs) SWCNTs functionalized with poly (acrylic acid) (PAA) oxygen-plasma- treated carbon nanotube thin films (CNTFs) Ag/AgCl Qin et al.78 Inkjet-printed SWCNTs Ag/AgCl Jung et al.106 multiwalled carbon nanotubes sheet decorated with nickel (MWCNT/Ni) Ag/AgCl Ag/AgCl setup open circuit potential (OCP) field effect transistor (FET) extended gate field effect transistor (EGFET) OCP resistance-based pH range biocompatibility response time (s) 3−11 yes ∼30 2−12 yes 3−7 1−13 yes 3−11 sensitivity −59.71 mV/pH standard deviation is 1.5 mV/pH as-sprayed CNTF: −37.6 mV/pH plasma treated CNTF: −55.7 mV/pH 7 200 passes: −48.1 mV/pH 80 passes: −36.2 mV/pH 20 passes: −22.6 mV/pH 5 passes: −16.4 mV/pH 2−10 carbon nanotube sheet decorated with nickel.106 They reported the pH-sensing properties to be highly dependent on the size of the nickel particles.106 Table 4 summarizes key nanotube developments for pH sensing. Carbon nanotubes have proven to be a superior material in terms of pH sensing due to their attractive characteristics, including their high surface-to-volume ratio, high mechanical stability, mass production capability, ease of chemical functionalization, and adjustable electrical properties.78 Additionally, nanotubes can be applied in open circuit potential setups, EGFET testing, electrical resistance test, and chemiresistance testing, making them a versatile sensing material. Their sensitivity has also shown to be improved by both oxidation during the fabrication process, and the incorporation of polymers or nickel particles in the development.105,106 These developments indicate the possibility of further enhancing the sensing capabilities of nanotubes. 4. pH-SENSING CONFIGURATIONS The sensing configurations for pH have a wide range from ISFET to resistance-based electrodes. All configurations have their unique advantages and are suited for in vitro and in vivo biomedical applications, given that the dimensions of insertable parts are scaled sufficiently. Although the traditional potentiometric configuration has been widely used in pH sensing for multiple decades, novel configurations have been recently introduced. This section discusses ISFET and EGFET configurations, which are closer in operation principles to traditional potentiometric configurations except for introducing a gating structure (transistor) to modulate current instead of voltage. We then discuss configurations involving interdigitated electrodes (IDEs) where two separate finger-shaped electrodes take on an interdigitated structure to utilize hybrid materials and variation in the capacitances. Furthermore, resistance-based configurations are discussed, where the resistance of the sensing material changes with pH. Finally, we conclude this section by summarizing the discussed configurations and providing remarks. 3.3. Summary and Conclusions Thin films and nanostructures provide a precise pH measurement option as the sensing material. Thin film metal oxides have proven to be useful in potentiometric sensors with electrode− electrolyte interfaces. Of the studied metal oxides, IrOx, ZnO, and doped ZnO films have displayed the most attractive characteristics for pH sensing due to their high sensitivity, facile fabrication methods, and biocompatibility. Conductive polymers have also served well for potentiometric sensors due to their ion-exchanging properties. The most commonly studied polymer is PANI because of its high conductivity, ease of synthesis, and stability. PANI has also shown to be effective in enhancing the performance of other materials, such as Gr. Nanorods have gained considerable attention due to their higher surface-to-volume ratio. The most popular nanorod material is ZnO considering its chemical stability, nontoxicity, electrochemical activity, fast response, and low cost. Gold nanorods have also been studied, but they require pretreatment, such as HCl, to increase their biocompatibility. However, nanotubes have proven to be a better alternative. The structure of nanotubes allows an extremely high surface-to-volume ratio, almost twice that of nanorods. Nanotubes also demonstrate high mechanical stability, mass production capability, ease of chemical functionalization, and adjustable electrical properties. An oxidation process or incorporating polymers or nickel particles in the development can further enhance the performance of nanotubes. Although various material choices are available, we ultimately select the sensing device material depending on the application and sensitivity requirements. 4.1. Ion Sensitive Field Effect Transistor (ISFET) ISFETs emerged in the 1970s with a milieu of advantages over the glass electrode, including significant durability compared to the more fragile glass electrode, easy storage without many necessary conditions, less measurement bias at extreme pH, and lower temperature dependence, making these sensors ideal for biomedical applications.108 ISFET devices typically require an RE. On the other hand, ISFETs respond quickly to pH changes, are highly sensitive, and are potentially miniaturizable. The ISFET sensor, consisting of source, gate, and drain terminals, monitors the current flow between the source and drain contacts as it responds to changes in the electric field between the gate and source terminals.18 The gate material defines the sensitivity and selectivity of the ISFET. For biocompatible sensors, an enzyme membrane can be used to coat the ion-selective gates, or biomolecules may be immobilized of the surface of the gates. Although ISFETs have many advantages, issues arise due to impurities in the semiconductor channel material and instability of the sensing membrane.109 Another source of variability in ISFET measurements originates from slow responding sites and a hydration effect resulting in a voltage drift. 4.2. Extended Gate Field Effect Transistor (EGFET) In 1983, J. van der Spiegel introduced the EGFET as an alternative to the ISFET for pH sensing.110 Later, Chi et al. modified Spiegel’s EGFET model structure comprising of a RE and a metal oxide semiconductor field effect transistor I DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Figure 4. Various pH-sensing configurations. (a) Typical extended gate field effect transistor (EGFET) setup. MOSFET stands for metal oxide semiconductor field effect transistor. (b) Plot of the relationship between drain voltage and drain current at a constant reference voltage (Vref) of 3 V from ZnO/Si nanowires-based EGFET with a sensitivity of −46.25 mV/pH in the 1−13 pH range.116 (c) Typical interdigitated electrodes (IDEs) field effect transistor setup. (d) Linear relationship between pH and capacitance for CuO nanoflower (NF) and nanorods (NR) IDEs in the 5−8.5 pH range, with sensitivity of 0.64 μF/pH for NR at 50 Hz.117 (e) Schematic of resistance-based pH configuration with single-walled carbon nanotubes (SWCNTs), the sensors had a sensitivity of 236.3 Ω/pH in the 5−9 pH range and response times of 2.26 s at pH 5 to 23.82 s at pH 9.118. Reproduced with permission from 116, 117, and 118. Copyright 2013 Elsevier. Copyright 2018 Elsevier. Copyright 2011 MDPI (Basel, Switzerland) under CC-BY3.0 https://creativecommons.org/licenses/by/3.0/. (MOSFET) connected to a sensing electrode.111 The RE has a stable electric potential, while the sensing electrode has an electric potential that is sensitive to changes in pH. The MOSFET is composed of a gate, source, drain, and body. A voltage, generated by the RE and the interaction of sensing electrode with H+ in solution, is applied to the gate to create a conducting channel. The conducting channel allows current to flow from the drain to the source (in the case of n-type MOSFET configurations).112 Figure 4a represents a typical EGFET setup, composed of the RE and the sensing electrode submerged in a solution. The RE is connected to a constant voltage source, while the sensing electrode is connected to a MOSFET configuration. The site-binding model, in which the surface potential (φ) at the sensing layer and electrolyte interface is determined by the number of binding sites on the sensing membrane, is used to derive the concentration of the H+ ions in the solution:113 2.303(pH pzc − pH) = i qφ 1 yz qφ + sinh−1jjjj zzz kT k kT β { sensitivity parameter. The surface sites per unit area, NS, is related to β, by the equation: β= 1/2 ( ) 2q2NS Kb Ka KTCDL (7) where CDL is the electrical double layer’s capacitance from the Gouy−Chapman-Stern model,113 Ka is the acid equilibrium constant, and Kb is the base equilibrium constant.114 The gate voltage of the transistor is related to the current (IDS) between the drain and source by the MOSFET expression. The drain-source voltage (VDS) relates to the current linearly before the current saturates. The current saturates when the drainsource voltage reaches the gate-source voltage (VGS) minus a threshold voltage (VT), necessary for establishing a conductive channel between source and drain. The complete current equation is given by ÑÉÑ ÅÄÅ 1 IDS = K nÅÅÅÅ(VGS − VT)VDS − VDS2 ÑÑÑÑ ÑÖ ÅÇ (8) 2 (6) where pHpzc is the pH value at the point of zero charge, k is the Boltzmann’s constant, T is the absolute temperature of the system in Kelvin, q is charge of the electron, and β is the where Kn is a technology constant,114 and for the saturation region, when VDS = VGS − VT, the relationship is defined by J DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review 1 K n[(VGS − VT)2 ] (9) 2 An important factor to consider in the EGFET is the material used in sensing electrode. The material must have a high sensitivity to the activity of H+ in the solution. An ideal sensing electrode would have a sensitivity of approximately −59.2 mV/ pH at room temperature. It is composed of an ion selective sensing membrane and conducting polymer, which converts charge carriers from ions to electrons. Several experiments have been conducted to find materials that are close to the ideal sensitivity for the sensing membrane. Common materials used include ZnO, SnO2, and IrO2. Table 5 shows representative EGFET sensors and their key properties, such as the sensing electrode and RE materials, the 4.3. Interdigitated Electrodes (IDEs) IDS = 4.3.1. Hybrid IDEs. IDEs are another popular configuration for pH sensing. They were popularized in the 1960s and have since been integrated into various biological sensing devices. IDEs are transducers that consist of two interdigitated electrode structures. IDEs have microgaps, which are gaps between the anode and cathode, allowing the electrode to exhibit sensitivity to changes in pH.120 In addition to the microgaps, the IDEs also have metal−semiconductor interdigitated extended gates that include “finger” electrodes with a sensitive membrane.121 Unlike the traditional EGFET, the IDEs configuration does not contain a separate reference electrode.121 IDEs are usually fabricated by photolithography.120 Figure 4c shows a typical interdigitated setup, i.e., the interdigitated extended gate field effect transistor (IEGFET).121 IDEs structures can also be hybrid structures, which contain a mix of organic and inorganic materials, such as a mix of polymers and glass.122 In the IDEs configuration, there is a constant voltage connected to a reference interdigitated electrode. The sensing interdigitated electrode is sensitive to the activity of H+, which generates a surface potential. The voltage applied to the gate of the MOSFET configuration is the superposition of the reference voltage and sensing electrode’s surface potential, allowing current to flow. IDEs are simple, easy to fabricate, stable, highly sensitive, and have a great potential to be miniaturized. Thus, IDEs offer many of the same benefits as the EGFET configuration, but have the added advantages of making experiments easier to carry out and inexpensive due to the lack of a separate bulky RE.121 Because of the IDEs’ unique features, the IDEs configuration presents itself as a promising alternative to the EGFET and ISFET. However, despite the advantages, IDEs also present some limitations. Like the EGFET configuration, the sensing material must be highly sensitive.117 Even though some IDEs’ materials, like CuO, are biocompatible, they may only have short-term stability and show a sub-Nernstian response.117 Hence, novel materials must be investigated for longevity and maximum sensitivity. 4.3.2. Capacitance IDEs. Capacitance-based interdigitated sensors depend on the changes that occur to the capacitance between the two interdigitated electrodes. Changing the charge distribution, the surface area of the electrodes, the dielectric properties, or conductivity can affect the capacitance between the electrodes.123 The surface area of the electrodes can be increased by adding interdigitated fingers. The basis of capacitance (C) can be found through the equation: Table 5. Summary of the Representative EGFET pH Sensors ref reference material maximum sensitivity (mV/ pH) pH range passivated ZnO Ag/AgCl −49.35 4−12 ZnO calcinated at 150 °C Al-doped ZnO Ag/AgCl −38 2−12 Ag/AgCl −57.95 1−13 ZnO/silicon nanowires (NWs) SnO2 Ag/AgCl −46.25 1−13 Ag/AgCl −56−58 2−12 sensing material Chiu et al.113 Batista et al.114 Yang et al.115 Li et al.116 Chi et al.111 maximum sensitivity, and the pH range. Figure 4b shows representative results from experiments utilizing ZnO/Si NWs in EGFET configuration at reference voltage (Vref) of 3 V, 1−13 pH range, exhibiting −46.25 mV/pH sensitivity. Compared to other configurations, the EGFET structure has several advantages, such as being easy to fabricate at a low cost, having a disposable gate, and having long-term stability.113 In addition, EGFET offers the advantage of isolating the electronics part (i.e., the transistor) from the chemical sensing part (i.e., the sensing electrode), in contrast to the ISFET where the transistor’s gate is exposed to the solution under test. However, the EGFET structure still poses some limitations. Common materials, such as zinc oxide, used for the sensing membrane of the working electrode, may have low sensitivity to pH due to impurities in the material used for the sensing membrane.114 Therefore, these materials must be modified through passivation113 or doping,115 which may be cost ineffective and time-consuming. In addition, the EGFET configuration also includes a RE, which may be expensive and bulky. C = εR ε0 A d (10) Table 6. Summary of the Representative IDEs pH Sensors ref Ali et al.121 Haarindraprasad et al.126 Lakard et al.122 Manjakkal et al.117 setup interdigitated extended gate field effect transistor (IEGFET) IEGFET interdigitated microarray potentiometric interdigitated impedance-metric sensing material variable measured maximum sensitivity pH range ZnO thin film current −22.4 mV/pH 4−11 ZnO nanostructured thin film current 3.72 μA/pH 2−10 polypyrrole thin film covered with a plastic polyvinyl chloride membrane CuO nanorods current −58 to −60 mV/pH 0.64 μF/pH at 50 Hz 2−11 K capacitance 5−8.5 DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 7. Summary of the Representative Resistance-Based pH Sensors ref sensing material Yang et al.119 Li et al.118 single-walled carbon nanotube (SWCNTs) on flexible parylene-C substrate dielectrophoresis aligned SWCNTs Copa et al.131 ZnO nanorods Chinnathambi et al.129 Nguyen et al.128 polyaniline functionalized electrochemically reduced graphene oxide (ERGO-PA) emeraldine salt polyaniline (ES- PANI) and poly(vinyl butyral) (PVB) blend film pH range sensitivity 236.3 Ω/ pH 1.71 Ω/ pH 0.28 MΩ/ pH lt d size of micro- electrodes size of gap (μm) 4−10 3 cycles height: 1 μm 4 5−9 10−15 cycles for 5 devices width: 6 μm 3 4−10 thickness: 100 nm length: 0.25−0.75 mm 4−9 1−8 response time (s) pH 5: 2.26 pH 6: 3.08 pH 7: 11.1 pH 8: 17.05 pH 9: 23.82 10 100 2 cycles 3850−3980 μm (length) × 20−50 μm (width) 20−120 oxides, such as zinc oxide and iridium oxide, and carbon materials, such as SWCNTs.127 In addition, some of these sensors include interdigitated electrodes, as mentioned in the previous subsection, to increase their surface area. The semiconductor device analyzer measures the resistance of the solution.118 The SWCNTs sensor reported in Figure 4e has a sensitivity of 236.3 Ω/pH in the 5−9 pH range with a response time that varies from 2.26 s at pH 5 to 23.82 s at pH 9. Unlike the ISFET and EGFET configuration, resistance-based pH sensors do not require an RE.105 It is important to note that resistance increases as pH increases due to the increase of OH− ions in the solution.119 The H+ and OH− interact with the material in the sensing layer, which contributes to a change in resistance that the semiconductor device analyzer measures.128 Yang el al. provided an example of a time vs resistance plot at 4−10 pH range, specifically using SWNTs on parylene as the sensing membrane material.119 Linear trend between pH and resistance has also been observed for electrochemically functionalized polyaniline reduced graphene oxide as the sensing membrane material.129 To evaluate sensor performance and repeatability, the normalized resistance can be found through the equation: where ε0 is the permittivity of free space, εR is the relative permittivity of dielectric material (i.e., the solution of interest), A is the surface area of finger electrodes, and d is the distance between the finger electrodes.124 For simplified IDEs, where edge effects can be neglected, the capacitance can be found through the equation: C = ηε repeatability (11) where η is the number of interdigitated fingers, ε is the permittivity (εRε0) of the sensitive coating film, l is the length of the interdigitated electrodes, t is the thickness of the electrodes, and d is the distance between the electrodes.124 A more complex model must be used to find the capacitance of IDEs larger than the nanoscale. For valid measurements, the capacitance behavior of the sensing probe must exhibit stability and specificity. Measuring the changes in capacitance is useful because it can detect minute changes in biological systems, which is critical for detecting changes in pH.123 Capacitance decreases with greater pH values, due to the increasing presence of OH− ions. The OH− ions increase the negative charge in the solution.125 Figure 4d illustrates this trend in an experiment that investigates the sensitivity of CuO nanoflowers (NFs) and nanorods (NRs) using interdigitated electrodes.117 The interdigitated NR CuO electrode showed 0.64 μF/pH at 50 Hz in the 5−8.5 pH range. Table 6 summarizes key IDE sensors, indicating the ability to integrate various structures with the interdigitated configuration, such as the metal oxide nanorods (discussed in Section 3.2.3), metal oxide thin films (Section 3.2.1), and polymeric thin films (Section 3.2.2). R − R min ΔR = Rr R max − R min where ΔR Rr (12) is the normalized sensor resistance, ΔR is the sensor resistance relative to the lowest sensor resistance, Rr is the sensor resistance range of the pH-sensing test, Rmax is the maximum sensor resistance, and Rmin is the minimum sensor resistance.118 This equation is used to evaluate resistive sensors’ sensitivity, repeatability, and reproducibility from device to device, which is important if the devices are produced on a large scale. Resistance-based pH sensors provide many advantages that make them ideal for pH sensing over other devices. The materials used, such as SWCNTs, are affordable.118 These devices also display high sensitivity and long-term stability.105 In one such experiment with carbon nanotubes, a device exhibited the same performance and calibration 120 days after the initial testing.105 One of the main advantages is that this sensor does not require a bulky and expensive RE. Furthermore, the small 4.4. Resistance Variation Resistance-based pH sensors are also promising for detecting the activity of H+. Resistance-based sensors have been traditionally used as gas and temperature sensors. However, they have also arisen as a promising alternative to the ISFET and EGFET for accurately measuring pH. Resistance-based sensors are composed of electrodes with a highly sensitive layer. These electrodes are connected, by external wiring, to a semiconductor device analyzer.118 Figure 4e shows the experimental setup. Common materials used for the sensing layer include metal L DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review ments. We then discuss standardization issues and recommendations for complete and accurate pH-sensing systems’ characterization, including the importance of time plots, their analysis, and identifying a convention for determining a critical point for sensitivity, response time, and drift objective assessment. Although this section is generally important for all pH-sensing applications, it has to be emphasized that for clinical biomedical applications, especially in vivo experiments, standards and protocols are indispensable. size of resistance-based pH sensors makes them ideal for miniaturization and optimal for in vivo applications.129 Nevertheless, there are still some challenges facing resistancebased pH sensors. Even though a similar fabrication process is followed, it is difficult to replicate the same behavior of sensing material on each electrode.118 Similar to the disadvantage of the hybrid interdigitated electrode configuration and EGFET setups, the sensing material of the resistance-based sensors must also be carefully selected because not all biocompatible materials have a high sensitivity and long-term stability. Hence, the remaining challenges for resistance-based pH sensors are electrode-to-electrode variability, the poor selectivity to H+’s (since other ions interacting with the surface can also affect the resistance), and limited material choices. These areas are subject to further research but can, generally, be mitigated through controlling processes variations, doping the material or functionalizing the surface with coatings that are selective to H+ (usually at the expense of sensitivity), and studying the resistance change behavior of more materials that are biocompatible such as IrOx and doped ZnO.130 Table 7 summarizes representative resistance-based pH sensors. It is worth mentioning that the sensitivity of resistance-based sensors varies significantly based on the choice of materials. For instance, Table 7 shows sensitivities of 1.71 Ω/pH, 236.3 Ω/ pH, and 0.28 MΩ/pH for polyaniline functionalized electrochemically reduced graphene oxide (ERGO-PA), dielectrophoresis aligned SWCNTs, and emeraldine salt PANI (ES- PANI) and PVB blend film, respectively. Also, the representative works show the limited pH range of operation, compared to the EGFET (Section 4.2) and IDEs (Section 4.3) configurations. 5.1. Inherent Properties of Components Despite the many advantages that the pH-sensing configurations offer, there are some inherent properties of device components that may affect biological systems and pH measurements, either enhancing or diminishing the accuracy for these devices. Although these intrinsic properties are often ignored, they play an important role in determining the performance of devices. For instance, the RE is a common component of pH-sensing configurations, yet its behavior is often unreported. Whether the RE is solution filled, usually with potassium chloride (KCl), or gel filled, usually with sodium chloride (NaCl), can have an effect on the stabilization time with a response time in the range of tens of seconds or tens of minutes, respectively.132 Common problems that REs may exhibit are that some electrodes may (i) suffer from leakage of inner electrolytes, (ii) get contaminated,105 and (iii) require frequent calibration, as some electrodes are less stable, especially ones that are microfabricated.132 In addition, the inherent properties and behaviors of sensing materials such as thickness, crystallinity, and composition must be considered. Specific sensing materials, such as zinc oxide and iridium oxide, must be nontoxic and biocompatible, especially for in vivo applications.116 These materials, unfortunately, may also dissolve or react with the solution and become modified on their surfaces with depositions from the test solutions. These depositions may cause sections of the sensing membrane to become insensitive to variations in H+ activity, due to the deposits blocking direct contact between the sensing material and the electrolyte.133 Furthermore, it has been determined that increased surface area has improved sensitivity of devices. For example, Chiu et al. determined that EGFETs with ZnO nanorod arrays had a greater sensitivity of 44.01 mV/pH than EGFETs with ZnO thin films (−38.46 mV/pH) in the 4−12 pH range.113 Moreover, Chen et al. showed that EGFETs with sensing membranes with larger contact areas had a greater sensitivity, using an EGFET device with a tin oxide/indium tin oxide (SnO2/ITO) sensing gate.134 Chen et al. also showed simulation results of the variation in sensitivity with electrode contact area from ∼−14 mV/pH at 0.1 mm2 to −55 mV/pH for areas greater than 0.8 mm2. Interestingly, in this experiment, beyond a certain contact area ∼0.8 mm2, sensitivity seems to saturate at the −55 mV/pH. It is worth mentioning that experimental observations do not back up the simulation results. In fact, higher sensitivity has been commonly attributed to a larger sensing area, even when the area is greater than 0.8 mm2.36,135,136 Additionally, crystallinity and composition of metal oxides also affect sensitivity. Batista et al. concluded that ZnO calcinated at 150 °C is amorphous and its composition also has zinc monoacetate, which yields a sub-Nernstian sensitivity of −38 mV/pH in an EGFET configuration.114 Other components that need to be assessed include the characterization instrument, glass electrode, and commercial transistor. The transistor often has drift that may cause the graph 4.5. Summary and Conclusions From the ISFET to resistance-based sensors, pH-sensing configurations have evolved to measure pH more accurately, affordably, and efficiently. The ISFET and EGFET configurations both measure the current between the drain and source and are highly sensitive, but use a bulky and expensive RE. Interdigitated electrodes configurations measure the current or capacitance and have interdigitated fingers that increase surface area. Resistance-based sensors measure the resistance of a solution affordably and quickly and have a great potential to be miniaturized with their lack of an RE. Ultimately, these devices offer promising solutions to measuring pH, but many steps need to be taken to improve their selectivity and assess their suitability for biomedical applications, such as surface treatment for selectivity against other common biological ions (i.e., Na+, K+, Ca2+), biocompatibility tests, and long-term stability assessment. 5. SENSING STANDARDS AND PROTOCOLS Given the diverse materials (Section 3) and configurations (Section 4) available for pH sensing as well as the expanding trends in both, standards and protocols are essential. Without standards and common protocols, the impact of research and its usefulness would be limited, and it would not be possible to provide objective contexts where various materials and systems can be accurately benchmarked. Therefore, this section is dedicated to discussing issues pertaining to the inherent properties of the various pH-sensing components, such as REs, buffer solutions, and transistors (in the case EGFET configuration, Section 4.2), and the effect of measuring instrument input resistance (Rin) on results. Furthermore, surface conditioning is discussed in terms of pre-tests’ initial treatments, as well as intermittent cleaning between measureM DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Figure 5. Inherent properties of the device components and input resistance of measurement system. (a) The relationship between time, drain to source current (Ids)n and drift for a commercial CD4007 nMOS commonly used in EGFET, showing a drift of −0.31 μA/min at drain to source voltage (Vds) of 5 V and gate to source voltage (Vgs) of 3 V.132 (b) The relationship between time and potential with shunt resistance (R) of 1.6 GΩ.132 (c) The effect of ∼3 h cycling tests in pH 6, 7, and 8 on the ZnO surface through (bottom) digital image of pristine ZnO surface and (top) digital image of the ZnO surface after successive testing. (d) Scanning electron microscopy pictures of the Ta2O5 surface structure (bottom) after 5 and (top) 30 CIP cycles. Neither a visible degradation nor a destruction of the Ta2O5 films has been observed after CIP procedure.146 (e) √Ids vs pH of solution at different time instances at a reference voltage (Vref) of 3 V and Vds of 5 V, and corresponding sensitivity values (S = slope of the √Ids vs pH plot divided by the slope of the √Ids vs Vgs plot).132 (f) Drift characteristics of the undoped ZnO and Al doped ZnO (AZO) with different atomic percentages (atom %) nanostructured pH-EGFET sensors measured within pH = 7 for the duration of 12 h. AZO with 3 atom % Al is the most stable with 1.27 mV/ h drift.147 (g) Actual drain current plot with time for ZnO sensing film vs Ag/AgCl reference electrode (RE) (the dash lines indicate different time instances to highlight the change in relative and absolute current values with time).132 (h) The first derivative of part (g) with time, highlighting the suggested placement of the critical point (Pc) and the estimated drift rates for solutions of different pH values.132 Reproduced with permission from ref 132, 146, and 147, respectively. Copyright 2018 John Wiley and Sons, Inc. Copyright 2005 Elsevier. Copyright 2013 Hindawi under CC-BY-3.0 https://creativecommons.org/licenses/by/3.0/. to deviate from an ideal Nernstian response. Drift is caused by diffusion of H+ and OH− and variations in the sensing surface.137 Figure 5a shows the relationship between pH and drift for SnO2/ ITO EGFET and the relationship between time, current, and drift for a commercial CD4007 nMOSFET, commonly used as the transistor in EGFET configurations (inherent drift ∼−0.31 μA/min). Interestingly, drift seems to decrease slightly with time. In addition, Chen et al.’s work suggests an increase in drift value with the increase in pH. For instance, drift values of 0.884, 1.58, 1.71, 1.8, and 2.51 mV/h correspond to pH values of 2, 4, 6, 8 and 10, respectively, for the SnO2/ITO EGFET. On the other hand, drift values of 14.2, 8.9, and 2.5 μA/min for pH 6, 7, and 8, respectively, for ZnO/Au EGFET, i.e., lower gate voltage drift at higher pH.132 Hence, the observed trends here cannot be generalized as drift values, and trends would depend on the pH solution constituents as well as the sensing electrode materials. There also may be a leakage current into the gate of the transistor, leading to a loss of a couple of millivolts, but this does not affect sensitivity because it is common in all solutions and should practically cause a systemic error, mainly affecting the standard reduction potential calculations. A MOSFET device, which is part of the ISFET and EGFET setups, usually has high input resistance (Section 5.2 has further information) and input capacitance. For the EGFET and ISFET devices, the site binding model, discussed in Section 9.4, and electrochemical reactions at the surface of the sensing material affect the measurement of surface potential.138 There are many steps that are required in order to characterize components on pH-sensing configurations. The RE must be checked for stability and response time by calibrating it in different test solutions using a specialized instrument such as a digital pH meter with a high input resistance. For a transistor in the EGFET configuration, stability and drift can be determined by an instrument such as a semiconductor device analyzer. The test solutions must also be characterized to ensure that they are stable in their pH values after long periods of time and during exposure to ambient test environments. Exposure to the surrounding air may affect the pH of the solution. As mentioned previously, exposure to carbon dioxide (CO2) may lead to a decrease in pH. Test solutions are assessed using a traditional and reliable pH-sensing device. After these intrinsic properties N DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX ref Li et al.116 sensing material ZnO/silicon nanowires (SiNW) cleaning method (1) Cleaned with sulfuric acid (H2SO4:H2O2 = 3:1) setup stability pH range hysteresis: 9.74 mV EGFET 0.62 μA1/2/ pH hysteresis: 5.4 mV 2−12 high electron mobility transistor (HEMT) 37.17 μA/ pH hysteresis: resolution of 0.1 pH 7.0−8.0 electrolyte-insulatorsemiconducto r(EIS) −60.2 mV/pH hysteresis: 22.4 mV 4.0−10.0 −66 mV/ pH Rasheed et al.155 ZnO/Ag/ZnO Ding et al.156 AlGaN/GaN (2) Dried with nitrogen gas (1) Wafer was precleaned with acetone and isopropanol Oh et al.154 Al2O3/SiO2 (2) Dried with nitrogen (1) After etching, cleaned in a fuming nitric acid for an hour ZnO/SiNW (2) Piranha solution (1:1 solution of 97% H2SO4 and 30% H2O2), 6:1 buffered oxide etchant, and deionized water rinsing, each for 10 min (1) Ultrasonication inacetone, isopropyl alcohol and deionized water EGFET Coppa et al.157 ZnO (2) Immersed in a mixture of 5 M aqueous hydrofluoric acid and 0.02 M silver nitrate solution for 1 h at room temperature (3) After etching, immersed in 30 wt % nitric acid solution for 1 min (4) Rinsed with deionized water and air-dried (1) Rinsed with methanol for 5 s Schottky barrier diodes Kumar et al.158 ZnO (2) Dried in flowing nitrogen (1) Rinse in acetone ZnO (2) Clean in dimethyl sulfoxide (DSMO) (3) Final clean with toluene (99.9%) (after each step, dried in flowing nitrogen) (1) Silicon substrate cleaned with solution consisting of 40% H2SO4 and 60% H2O2 O Ali et al.159 repeatability −46.25 mV/pH (2) Rinsed with deionized water (3) Soaked in dilute hydrofluoric acid (HF:H2O = 1:100) (1) Ultrasonically cleaned in acetone deionized water Huang et al.136 sensitivity extended gate field effect transistor (EGFET) EGFET difference ratios: < 5% in pH = 5−13, 25% in pH = 1, 15% in pH = 3 1−13 Chemical Reviews Table 8. Summary of the Common Substrate Cleaning Techniques Used in pH Sensor’s Fabrication 2.0−12.0 −27.86 mV/pH Review DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review are characterized, pH testing and objective assessment of various configurations can take place. In most cases, the substrate surface is submerged in either an alcohol solution (acetone, methanol, etc.) and/or an acid (sulfuric acid, hydrosulfuric acid, piranha solution, nitric acid, etc.) and then rinsed by deionized water. The purpose of using alcohol and acid is to clean organic residues or any impurities off the substrate surfaces. The substrate is then dried either with paper towel, flowing nitrogen, or in air.153 Additionally, in the case of Ag or other ions assisted etching processes, the wafer is also cleaned in acid to remove silver dendrites and residual ions.154 Table 8 summarizes common substrate cleaning techniques (ex situ cleaning). The various different cleaning protocols, even for the same substrate material, is indicative of the need for a standard protocol to eliminate variations in sensing film quality due to the conditioning of the underlying substrate. Additionally, to ensure higher sensitivity and overall quality, it is important that the surface of the sensing material is kept clean of any residues during the manufacturing process of the sensing material and electrode structure. Different types of cleaning vary depending on the specific material and the setup used, but they all have commonalities, such as immersing the surface in acid or alcohol, and plasma treatment. Specifically, for sensors using ZnO, another common cleaning method that is often used is plasma treatment, often under ultrahigh vacuum (UHV) conditions.157 An important note on plasma treatment is that under short duration plasma treatment (∼2 min), using X-ray photoelectron spectroscopy (XPS) and AFM, the surface composition of the examined sensing material might change. For instance, it has been reported that plasma treatment of AlGaN barrier layer increased the Al−O bonds at the surface. The increase in Al−O bonds is important to an increase in pH sensitivity, while further exposure to plasma treating increases Ga−O over Al−O bonds, and significantly decreases the sensitivity. With this process, a smooth and clean surface with an ultrathin oxide membrane can be obtained, increasing the sensitivity and quality of the sensor.160 Hence, while short-term exposure to plasma could remove contaminants from the sensing surface, long-term exposure of the film to plasma could result in significantly lower sensitivity.160 In addition, the temperature, time, and pressure that the film is treated in should change based on many factors, most notably the crystallography of the material. For example, the settings are significantly different for ZnO (0001) and ZnO (0001).161 Table 9 summarizes the cleaning methods for a pH-sensing material surface (in situ cleaning). Similar to the observations made on Table 8, cleaning methods differ for the same sensing material. Furthermore, in this case it is the sensing surface itself that is being subjected to varying conditioning techniques. Given the dependence of pH measurements on sensing material’s surface quality, the conditioning protocol is critical to report. Similarly, investigating how the conditioning protocol has affected the surface is essential, for consistency in reporting and comprehensibility. 5.2. Input Resistance of Characterization Systems Along with assessing the inherent properties of components, it is important to assess and characterize the internal resistance (Rin) of the pH tools. pH full measurement cells (i.e., ones composed of at least a sensing electrode and a RE immersed in an electrolyte) have high impedance. Digital multimeters, semiconductor device analyzers, n-type and p-type commercial transistors, instrumentation amplifiers, electrometer, electrochemical analyzers, and pH meters each have their own input resistance. For instance, digital multimeters usually have a Rin in the range of 10 MΩ to 10 GΩ.139 Semiconductor device analyzers usually have an Rin greater than 1 TΩ.140,141 For OCP measurements, the input resistance of the potentiometer must be orders of magnitude larger than the resistance of the pH full cell. The pH full cell resistance includes the resistance of the glass sensing electrode or other sensing electrodes and the resistance of the RE.132 Figure 5b illustrates the relationship between time and OCP with a shunt resistance of 1.6 GΩ showing inaccurate results. The resistance of the pH full cell occurs as a result of the sensing material’s resistivity (ρ) and sensing geometry. For example, instruments measuring the potential of the glass electrode should have an Rin in the hundreds of GΩ to TΩ range, because the membrane resistance of the glass electrode is in MΩ.142 An instrument with high input resistance is correlated to smaller error, better stability, and higher and better pH response, because only a small portion of the current would travel through the instrument’s resistance.132,143,144 Overall, the ability of the pH device to perform successfully is dependent on the resistivity of the sensing material, the resistance of the measuring instrument, and the resistance of the pH cell. 5.3. Surface Cleaning The process of cleaning a sensor is extremely important in order to ensure reusability and confidence in measurement; however, the detailed process is rarely discussed.148 Often times, the importance of cleaning the sensor is underestimated. In fact, it has been reported that the sensor performance differs by as much as 13% with different cleaning procedures carried out.148 Since different cleaning procedures can cause such different results, not using a systematic cleaning process could lead to wrong conclusions. For example, a sensor’s high sensitivity might be mistakenly shadowed by undesired deposits if an inefficient cleaning procedure was used. While the cleaning process is rarely discussed, it is extremely important in ensuring sensor’s accuracy, quality, and objective reporting. In general, based on the sensor material and setup, three types of cleaning are performed: cleaning the substrate that the sensor is based on (ex situ), cleaning of the sensor material itself (in situ), and cleaning the electrode in between measurements. All three of these cleaning methods should be carried out to ensure maximum quality, reliability, and life span of the sensor. The initial step of cleaning happens on the substrate that the sensing film is deposited on. For example, silicon wafers are often used to grow silicon nanowire (SiNW) structures. To ensure that no impurities or any contaminant exists on the silicon substrates, steps should be taken to clean the surface of the substrate before growing microstructures. This step is crucial and previous studies have shown that cleaning the surface enhances the quality of the sensing device.149−152 5.4. Surface Resetting (Intermittent Cleaning vs In Situ Discussion) With time, pH electrodes naturally undergo aging effects, which slightly impacts their performance.145 This is exacerbated with coatings and contamination that occur with frequent use of the sensor (Figure 5c).132,145 Resetting the surface of sensors is imperative for obtaining reliable results when monitoring pH. However, its method depends on the sensing requirements and P DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX no degradation in 2 h testing periods difference between trails <0.01 mV extended gate field effect transistor (EGFET) ZnO Zhang et al.38 Kumar et al.158 ZnO i-ZnO (0001̅): 30 min, 525 °C, and 0.050 Torr ii-ZnO (0001): 60 min, 550 °C, 0.050 Torr (1) ZnO NWs and NW−NF hybrid structures treated with oxygen plasma for 30 s to remove impurities and organic contaminants (1) Argon sputter cleaning b-Au/ZnO Exposure to a 20 W remote 20% O2/80% He plasma as follows: b-Schottky barrier diodes −43.22m V/pH −58.7 mV/pH ion sensitive field effect transistor (ISFET) a-Surface studies only AlGaN/ GaN a-ZnO Wang et al.160 Coppa et al.157,161 (2) Immersed in ethanol solution with 20 μL of 5% 3-aminopropyl triethoxysilane (APTES), by volume, for 2 h with periodic supply of the vaporized solution (3) APTES surface rinsed with deionized water five times (1) Short-time O2 plasma treatment (1−5 min) AlGaN/ GaN Ding et al.156 (1) Treated in a UV/O3 chamber (400 W, 10 min) high electron mobility transistor (HEMT) 37.17 μA/ pH repeatability sensitivity setup cleaning method sensing material ref Table 9. Summary of the Intermittent Cleaning Methods for pH-Sensing Materials’ Surface given application. The two main methods are intermittent cleaning between each measurement, or in situ solutions. Intermittent cleaning is the most common form of resetting the surface of a sensor in research studies. Frequent cleaning of the surface minimizes the effects of contamination.145 This method allows for more consistent and accurate measurements of pH, by ensuring there is little to no residual substance from previously tested materials. A pH-sensing membrane based on an ionic liquid polymer composite developed by Ping et al. performed with a sensitivity of 57.5 mV/pH, close to the Nernstian expected value.71 In between each measurement, the membrane electrodes were washed with deionized water, and no hysteresis was observed.71 When testing p-aminothiophenol functionalized gold nanorods, they also showed well retained values.76 In between each of the measurements, the culture dishes were rinsed with phosphate buffer solution three times.76 Intermittent cleaning can also be applied to some implantable sensors. An implantable, battery-less, and wireless capsule with integrated pH sensors for gastroesophageal reflux monitoring was developed and implanted in the esophagus wall of a pig.162 Between each measurement, the esophagus was flushed with tap water to reset the sensor.162 The device consistently performed with sensitivities between −51.1 and −57.7 mV/pH.162 However, intermittent cleaning is not practical in real-world applications where measurements must be taken continuously and in real-time. Cleaning-in-place (CIP) is an in situ solution that exposes the senor to a chemical process before measurements in order to combat hysteresis. Many industries, including biotechnology, food, pharmaceutical, cosmetic, construction and building materials, and water purification, produce a large demand for in-line pH sensors.146,163,164 Since they cannot utilize glass electrodes due to strict regulations, which happen to be less prone to fouling, they rely on ion-sensitive field-effect transistors.132,146,163 Schöning et al. developed a “non-glass” pH sensor based on a Ta2O5-gate electrolyte−insulator− semiconductor structure.146 Before measurements, the sensor was subjected to cleaning in 4% NaOH solution at 80 °C for 15 min, then in 0.65% HNO3 solution at 80 °C for 5 min.146 The hysteresis seen during testing ranged between 1.5 and 9 mV, depending on the sensor type, number of CIP cycles, and the pH value of the buffer.146 The device performed with a Nernstian value of 57 mV/pH, which showed to be independent of number of CIP cycles, because after 30 cycles no degradation of the sensor surface was observed (Figure 5d). Linkohr et al. investigated the stability of AlGaN/GaN pH sensors that have undergone CIP treatments.163 They exposed their sensors to 1.5% NaOH at 80 °C for 30 min and then rinsed them with deionized water afterward.163 The sensor showed hysteresis below 3 mV in the pH range of 2−10, but then jumped to 25 mV with pH 12, indicating a negative effect from alkaline solutions.163 After 15 cycles, the sensitivity dropped from −57 mV/pH to −30.3 mV/pH, clearly showing how the process of CIP impacted the performance of the sensor.163 While CIP treatments do well in preventing hysteresis, they heavily impact the lifetime of these sensors, producing a significant challenge for researchers and increased costs.146,163 Intermittent cleaning enables researchers to obtain accurate results by diminishing the effect of hysteresis and improving the lifetime of sensors.145,146,163 CIP produces more realistic results for real-world applications. These treatments applied before measurements help prevent hysteresis, typically seen with cross contamination.146,163 However, they decrease the lifetime of a 2−9 4.0−9.0 7.0−8.0 Review stability pH range Chemical Reviews Q DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 10. Summary of the Surface Resetting Techniques for pH Sensors ref sensing material reference material Queeney et al.145 Ghoneim et al.132 Schöning et al.146 resetting method pH range sensitivity • Spray-jet ZnO Ta2O5 Ag/AgCl Ag/AgCl • Retractable housing • Applying a voltage in reverse direction • Briefly dissipating the charge from the sensing electrode Cleaning in 4% NaOH solution at 80 °C during 15 min and subsequently in 0.65HNO3 solution at 80 °C during 5 min 3 to 10 Linkohr et al.163 AlGaN/GaN Ag/AgCl Sensor is exposed to 1.5% NaOH at 80 °C for 30 min and rinsed in DI water afterward. 2 to 12 Ping et al.71 poly(vinyl chloride) and ncetylpyridiniu m hexafluoropho sphate incorporated with quinhydrone Ag/AgCl membrane electrodes washed with deionized water 2 to 9.5 Zong et al.76 Cao et al.162 gold nanorods p-aminothiophenol Ag/AgCl cell culture dishes were washed with phosphate buffered saline for three times washed with tap water 3 to 8 • −57 mV/pH, independent of number of CIP cycles • Hysteresis between 1.5−9 mV dependent on number of CIP cycles and pH value of buffer • After 15 cycles, sensitivity dropped from −57 mV/pH to −30.3 mV/pH • Hysteresis below 3 mV between pH 2 to 10 and 25 mV when pH 12 • −57.5 mV/pH • No hysteresis observed IrOx sensor due to the strong chemicals used.146,163 Other in situ methods that have been discussed include altering the electrical properties of the sensor, such as by temporarily applying a voltage in the reverse direction of the cell potential.132 Another method would be to model the equilibrium surface charge density as a function of pH, which could then be related to the measured potential.132 These approaches still require further investigation, but they open possibilities to improve in situ solutions for accurate pH sensing. Table 10 summarizes the surface resetting techniques discussed in this section. The sprayjet solution utilizes an intermittent jet of air or water directed at the tip of the pH electrode to clean the contamination coating. This is suitable for submersible systems. However, it is not suitable when the dilution of tested solution is an issue, which is definitely the case with biomedical applications. The same applies to all intermittent cleaning methods involving chemical or DI splashing treatment. In that case, the pH probe can be fixed in a retractable housing, where it can be retracted, washed away from the testing environment, and reinserted. This concept is suitable for biomedical applications as the sensing probe can be implanted and retracted at specific intervals for cleaning. However, when the system does not constitute a probe structure, as in wearable pH-sensing patches and systems (Section 8.1), even the retractable housing would be inapplicable. In that case, the most feasible option is either to apply an electrical signal to reverse the reaction that resulted in contamination coating, or periodic discharging of the previous measurements’ surface charge. 1.9 to 12 between −51.1 and −57.7 mV/pH depicted in Figure 5e, while in saturation mode, the sensitivity of a ZnO EGFET sensor jumps from −17.5 mV/pH to −58.1 mV/pH just by waiting until the critical point has been reached.132 In linear mode, the sensitivity jumps from −9.8 mV/ pH to −84.8 mV/pH, possibly indicating that an equilibrium was reached earlier.132 It is necessary to report the time plot in order to determine how long the electrodes must remain in a solution before an accurate measurement can be collected.132 Unfortunately, even with their vast importance, many papers utilizing EGFET and ISFET do not include time plots that show a response time in their discussion. Without referring to a time plot for corroboration, performance results are not as reliable as they should be. Nonetheless, several ISFET, EGFET, and potentiometric works include time plots to show stability and drift characteristics of a device; however, they rarely indicate when the critical point has been reached. Thus, the calculation of drift becomes arbitrary. For instance, Figure 5f shows drift values defined as the slope of the output voltage vs time plot between 5 and 12 h.107,147 The oxygen plasma treated CNTF in pH 7 had a drift value of 1.36 mV/pH,107 and the drift values for Al doped ZnO films with different Al atomic percentages (atom %) of 0, 1, 2, 3, 5, 7 atom % corresponded to 16.81, 13.59, 4.77, 1.27, 3.38, and 8.79 mV/h in pH 7.147 Although these observations and stating the exact range for calculating the drift are very useful, the lower limit of 5 h is not a common time frame for collecting a pH measurement. Therefore, an earlier time range that is feasible for collecting pH measurements (i.e., several minutes) would be most useful and expectedly would result in higher drift values. On the other hand, a wearable pH sensor using PANI in OCP configuration exhibited a drift of 0.7 mV/h.68 In the latter case, drift was calculated from an hour of continuous measurement in a 4 h time frame, most likely between the third and the fourth hour. Furthermore, a ZnO pH sensor in OCP configuration with ±3 mV/h drift was reported with no indication of either a time range or the window of calculation (likely within the first 15 min based on the reported stability plot).37 One solution to determining the critical point is to evaluate the first derivative of the time plot, demonstrated in Figure 5g,h.132 The point 5.5. Time Plots and Analysis When using a pH sensor, the device will require some time before it can obtain at least 90% of the full response.132 This time period is referred to as the response time, and it can range anywhere from a few seconds to several minutes depending on a number of factors. The final point after the complete response is defined as the critical point and is used in the calculation of the calibration plot for the device, making it extremely important.132 Taking the critical point too early or too late can have detrimental effects on the measured sensitivity value. As R DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 11. Summary of the Representative EGFET and ISFET Works, Highlighting the Limited Reporting of Time Plots ref setup reference material sensing material sensitivity Ghoneim et al.132 extended gate field effect transistor (EGFET) ZnO Ag/AgCl Wang et al.147 Li et al.168 EGFET aluminum-doped ZnO (AZO) Ag/AgCl • At 240 s: −58.1 mV/pH saturated, −84.8 mV/pH linear • At 0 s: −17.5 mV/pH saturated, −9.8 mV/pH linear Al dosage of 3%: −57.95 mV/pH indium tin oxide (ITO)/ polyethylene terephthalate (PET) iridium nitride nanorod Ag/AgCl −44.86 mV/pH Chen et al.98 ion sensitive field effect transistor (ISFET) EGFET Ag/AgCl Lee et al.99 ISFET Ag/AgCl Tsai et al.107 EGFET Ag/AgCl −55.7 mV/pH Batista et al.114 Ali et al.121 Chiu et al.169 Lin et al.170 EGFET ZnO-based nanorod/gate-recessed AlGaN/GaN oxygen-plasma-treated carbon nanotube thin films ZnO • Current: 26 μA/pH • Voltage: −22.66 mV/pH −57.66 mV/pH Rasheed et al.171 Yang et al.115 EGFET ZnO tantalum ZnO (ZnO:Ta) indium−gallium-zinc-oxide nanoparticles/silicon nanowire (IGZO/SiNWs) multilayer ZnO/Pd/ZnO structure EGFET Al-doped ZnO (AZO) Huang et al.136 EGFET Lee et al.172 time plot 240 s included 10 s included −38 mV/pH Ag/AgCl Ag/AgCl −22.4 mV/pH −41.56 mV/pH −50 mV/pH Ag/AgCl −52 mV/pH zinc oxide/silicon nanowire hybrid Ag/AgCl Al-dosage of a-0%: −35.23 mV/pH b-1.98%: −57.95 mV/pH c-3.35%: −55.61 mV/pH d-6.27%: −53.34 mV/pH • SiNWs: −52 mV/pH EGFET ZnO thin films and ZnO nanorods Ag/AgCl Chiu et al.113 EGFET ZnO thin films and nanorods Ag/AgCl Wang et al.57 Thanh et al.165 Maiolo et al.72 Chang et al.173 Fernandes et al.167 EGFET EGFET Al-doped ZnO nanostructures ZnO nanorods Ag/AgCl Ag/AgCl • ZnO/SiNW: −58 to −66 mV/pH • Unpassivated: 47.96 μA/pH • Passivated: 52.58 μA/pH • Unpassivated intrinsic-ZnO (i-ZnO) nanorod array: −44.01 mV/pH • Unpassivated i-ZnO thin film: −38.46 mV/pH • Passivated iZnO thin-film: −42.37 mV/pH • Passivated i ZnO nanorod array: −49.35 mV/pH −57.95 mV/pH −15.4 mV/pH EGTFT ZnO nanowalls Ag/AgCl −59 mV/pH EGFET ZnO thin films and nanowire array Ag/AgCl 48.6 μA/pH −36.9 mV/pH EGFET ZnO thin films saturated calomel Rosli et al.166 EGFET ZnO nanostructures/Au/ITO • Fluorine doped tin oxide substrate with Al % of 0%: −22.3 mV/pH 3%: −29 mV/pH 7%: −40.1 mV/pH 8%: −30.9 mV/pH • ITO substrate 3%: −23 mV/pH 7%: −26.6 mV/pH 8%: −33 mV/pH 10%: −30 mV/pH −38.2 mV/pH EGFET EGFET EGFET response time Low98,114,165−167 (or high57,99,107,115,136) reported sensitivities can be due to the sensing material used, the fabrication process followed, or an early (or late) extraction of values before (or after) reaching the critical point. This could result in incorrect where a constant value begins signifies that the critical point has been reached. Figure 5h indicates the critical point is at 240 s. It is strongly recommended to report the time plot, which clearly indicates the response time of their device. S DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 12. Summary of the Various Reports on pH Sensors, Indicating Response Times and Emphasizing the Absence of a Clearly Identified Critical Point ref Ghoneim et al.132 Goldstein et al.174 Grant et al.175 Li et al.104 Gou et al.105 Qin et al.78 Korostynska et al.16 Yoon et al.72 Guinovart et al.59 Ping et al.71 Rout et al.100 setup sensing material reference material critical point 240 at 240 s extended gate field effect transistor (EGFET) fiber optics ZnO • Optoelectronics • Electrochemical open circuit potential (OCP) FET OCP OCP • Silica optical fibers • IrOx single walled carbon nanotubes (SWCNTs) • N/A • Ag/AgCl Ag/AgCl 5400 SWCNTs functionalized with poly (acrylic acid) (PAA) Inkjet-printed SWCNTs 120 mg of lithium perchlorate (LiClO4) and 10 μL of pyrrole (PPy) dissolved in 5 mL of acetonitrile polyaniline (PANI) nanopillar array electropolymerized PANI Ag/AgCl Ag/AgCl Ag film 3−7 7 <1 Ag/AgCl polyvinyl butyral polymer (PVB) Ag/AgCl <1 <20 Au 150 OCP OCP OCP Ag/AgCl response time (s) pH sensitive dye contained within a H+ permeable envelope 42 poly(vinyl chloride) (PVC) and n-cetylpyridinium hexafluorophosphate (CPFP) incorporated with quinhydrone (QH) ZnO FET ∼30 <10 Table 13. Summary of the Representative pH Sensors and Their Reported Drift Values ref Chang et al.176 setup ion sensitive field effect transistor (ISFET) reference material sensing material ZrO2 Ag/AgCl critical point drift hours 1−7 Wang et al.147 Rigante et al.178 extended gate field effect transistor (EGFET) FinFET aluminum-doped zinc oxide (AZO) Ag/AgCl HfO2 Ag/AgCl Tang et al.168 ISFET Ag/AgCl Zhou et al.177 open circuit potential (OCP) indium tin oxide (ITO)/ polyethylene terephthalate IrO2 n-channel pH 3 −58.55 mV pH 5 −51.54 mV pH 7 −41.61 mV pH 9 −34.66 mV pH 11 −32.52 mV p-channel pH 3 13.33 mV pH 5 6.04 mV pH 7 −4.91 mV pH 9 −25.92 mV pH 11 −30.82 mV Al dosage of 3% best drift rate at 1.27 mV/h over 12 h • Single wire FinFET: drift time of 0.13 mV/h. • 3-wire FinFET: Drift time of 0.1 mV/h. • 5-wire FinFET:0.12 mV/h (all over 105 h) drift rate <1.7 mV/h over ∼8.5 h Ag/AgCl • Over 86 h: potential drift 0.3 mV/h • First 30 h: potential drift 0.6 mV/h conclusions. Including a time plot indicates the response behavior of the device, makes results and findings more reliable, and enables objective benchmarking of various sensing materials and systems. Table 11 summarizes various EGFET and ISFET works and time plots. An important observation for EGFET and ISFET reports is the necessity of identifying the critical time at which the output or transfer plots of the transistor are collected. This can be identified by collecting the time plots (Ids vs time) discussed in this subsection and identifying the critical point (discussed in detail in the next section, Section 5.6). In general, sensitivity values for ISFET and EGFET configurations vary widely from −22.4 to −59 mV/pH, and response times range from tens of seconds to several minutes. In addition, the EGFET configuration (Section 4.2) is gaining popularity because it offers the advantage of separating the electronics part from the sensing part, compared to the ISFET configuration (Section 4.1). 5.6. Critical Point (Pc) for Response and Drift Determination In addition to time plots, Pc in pH sensing is an important characteristic of pH sensors that is often overlooked. The critical point is used to determine the response time, drift, and sensitivity of a device. Each of these ultimately determines the effectiveness of a device, so without knowing the critical point there is no reliable way to evaluate a device. T DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Figure 6. pH regulation in the human body. (a) Ion carriers (Na+-H+ exchanger (NHE)) regulation of cell pH.179 (b) Blood−brain barrier ion transporters and channels.191 (c) pH before birth and after birth, showing good correlation between fetal scalp pH value and outcome pH value and used to study the reliability of fetal scalp pH values,194 (d) Microfluidic device fabrication scheme for ZnO-based pH sensor with open circuit potential (OCP) configuration, with −43.71 mV/pH sensitivity, drift of 3 mV/h, and response times between 26 and 32 s in the 1.68−9.18 pH range.37 Reproduced with permission from refs 179, 191, 194, and 37, respectively. Copyright 2009 Springer Nature, Ltd. Copyright 2014 Elsevier. Copyright 2016 Springer Nature, Ltd. Copyright 2017 American Chemical Society. mV/h range for p-channel ISFET. For 12 h, an average drift of 1.27 mV/h was reported for AZO EGFET. For larger time ranges, drift values of 0.1 and 0.3 mV/h have been reported for 3-wire HfO2 FinFET and IrO2 in OCP configuration. The trend clearly shows that longer time ranges and later time frames exhibit lower drifts, and it would be inaccurate to compare drifts of configurations and materials if time range and time frames are inconsistent. Evidently, a standardized method for identifying the critical point is essential. The method discussed is to take the first derivative of the time plot and locate the point at which constant values begin, as illustrated in Figure 5g,h. The point at which this occurs is then to be defined as the critical point and used in the sensitivity calculation. Following standards and protocols for pH measurements is what ensures the validity of the measurement and extends its applicability to future and related works. Traditional pH measurements start with proper calibration of the tested device. This can be done using standardized buffer solutions such as the National Institute of Standards and Technology (NIST) recommended buffers, or in phosphate buffered solutions that mimic biological fluids. This is usually conducted by using a minimum of two calibration pH buffer solutions that cover the expected range of the desired pH measurements. A three-point calibration with three buffered solutions is recommended for a more precise calibration plot. Once calibration is complete, the device is washed in deionized water and used to measure the test solution. A traditional potentiometric measurement would output the voltage difference between the reference electrodes and sensing electrode versus time. The analysis of the plot is straightforward and response time, saturation value, and drift can be extracted. With the advancements in pH sensing and biomedical applications, the actual sensing environment is far more complex than the common standard buffer solutions. Hence, it is more feasible to calibrate in more customized environments that closely match the real one. To this end, arbitrary choices for calibration and testing might be not only acceptable but even more accurate. To cope with the emerging Response time is important for pH sensors because for reallife applications there may only be a small window of time to take a measurement, and thus a device must be able to respond within that time. Although the response time of a device is defined as the time it takes for a device to reach 90% of the full response, or the time it takes for a device to reach the critical point,132 many reports do not explain how this value was determined.16,59,71,72,78,100,104,105,174,175 Without reporting how the critical point was found or when it is achieved, there is no way to know if these are correct measurements or just arbitrary values. Table 12 summarizes various reports, indicating response times, and whether the critical point was identified. This table clearly shows the dispersion in response time reporting from <1 s to 90 min. Although that is possible based on the material system; there is 20 times variation for PANI in OCP (from <1 s to <20 s). With an arbitrary point for response calculation, it is highly subjective to compare materials or systems responses across different reports, unless the same work compares two materials or systems using the same subjective methodology. This highlights the need for the critical point convention and its usefulness in identifying proper sensitivity values and separating the full response from the onset of drift. Drift is another important characteristic of pH sensors; it determines the stability of a device. Over time, most electrodes suffer from potential drift, i.e., the slope of the output after the critical point (full response) has been achieved.176 Quantifying drift begins at the critical point, so without knowledge of when the critical point of a device occurs, it is difficult to assess drift and accurately determine suitability for long-term monitoring. Various reports147,168,177,178 show drift values; however, there is no explanation of how it was extracted. Thus, failing to report when or how the critical point was reached makes objective benchmarking of these reports infeasible. Table 13 summarizes representative pH sensors and their reported drift values. Notably, the lack of a common convention for calculating drift results in subjective time frames and ranges for calculating drift. When drift is collected in the first seven hours, the values are in the 4.6−8.4 mV/h range for ZrO2 n-channel ISFET and 0.7−4.4 U DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review acid buffering and excretion by the reclamation of filtered bicarbonate and its regeneration are essential to the renal acid− base balance.187 Lungs also function in the excretion of metabolically produced CO2. The lungs control the CO2 partial pressure (pCO2), or the pCO2 reflecting the amount of the gas dissolved in the blood, to counteract the influx of H+ and resulting acidosis or alkalosis due to the excessive retention or excretion of CO2.188 This pulmonary regulation of acid−base balance is based on ventilation adjustment. Ventilation, directed by chemoreceptors, afferent and efferent nerves, central nervous connections, and the skeletal muscles, occurs in response to both arterial pH and that of the cerebral spinal fluid (CSF).189 Worth mentioning, this ventilation must balance the need to control the pH and also the need for oxygenation. Arterial response is observed to be more rapid than CSF response. CO2 functions in these processes alongside its combination and dissociation products to eliminate acid waste and regulate extracellular pH (pHe).189 The lungs, by excreting carbonic acid, contribute to the buffer system, which controls blood pH.190 Blood pH varies inversely with pCO2 concentration. The pH of CSF differs from arterial pH in that carbonic acid among other molecules diffuse easily. This results in a direct effect of pCO2 through the blood− brain barrier and the necessity for constant action of the bicarbonate pump to maintain a normal pH.189 With a sudden increase or decrease in pCO2, respiratory acidosis or alkalosis occurs, severely changing CSF pH and possibly causing delirium or even a coma.188 trends in measurement technologies, providing as much information as possible about calibration, measurement, and analysis is essential, including specifications of characterization instruments. In this section, we highlighted the new aspects that are essential for establishing standards and protocols to cope with current progress. These include (i) accounting for extrinsic components effects, (ii) instrumentation, (iii) initial and intermittent surface conditioning, (iv) collecting and analyzing time plots in all cases, and (v) defining conventions that would enable consistent reporting. Nonetheless, at this stage there is no clear protocol that is widely approved and followed across the pH-sensing community. This highlights the need for careful reporting that includes all possible sources of error (intrinsic and extrinsic to the assessed system). Through proper reporting, various studies can be objectively benchmarked. Only then will patterns emerge for best practices, and new universal standards and protocols can be established. 6. pH REGULATION IN THE HUMAN BODY pH regulation in the human body is crucial for proper functionality and disease prevention. Strict regulation mechanisms exist at the cellular and organ levels. This section focuses on pH regulation in the extra- and intracellular environments as well as the organ level, specifically the kidney and lungs. We also discuss the essential role of blood in pH regulation throughout the body. 6.1. Cells pHi must be maintained within a strict range in order for cellular processes to proceed. This balance is regulated both within the membrane of the cell as well as certain organelles. These cellular compartments have inherent pH buffering capacities varied by intracellular weak acids and basis. An additional buffer in most mammalian cells is created by the hydration of CO2 and deprotonation of carbonic acid.179 These two buffering capacities compensate for changes in pHi. Acidification of the cell is prevented by membrane H+ pumps or proton coupling (Figure 6a). The energy to force these H+ against the electrochemical gradient can be provided by adenosine triphosphate (ATP).179 To recognize changes in pH and acknowledge the need for H+ transfer, membrane recognition proteins are utilized. Individual organelles, such as lysosomes, must also maintain specific pH values to perform. Organelles preserve their pH through similar methods like the cell, i.e., through membrane H+ transfer activity.179 Because of the excellent buffering abilities of the cells, a deviation in cellular pH usually indicates anomalies in functionality.179,180 For instance, cellular pH can become vitally important in recognizing and treating cancer cells because their pH is different from healthy cells.181,182 Hence, understanding and monitoring pH can enable recognition of cancerous growth. To this end, extensive recent studies have been carried out to sense both pHi and extracellular pH (pHe),180,183,184 including sensing in an in vivolike 3D environment185 and in vivo imaging.186 6.3. Blood Given the important role of blood pH and its effect on CSF, blood pH is tightly regulated in the human body, fluctuating between 7.35 and 7.45. The bicarbonate buffer system in the kidneys and the respiratory function of the lungs are the main regulatory functions for blood pH. Other methods for pH regulation include ion transporters and channels across different barriers, such as across the blood−brain barrier (Figure 6b).191 Specifically, the chloride−bicarbonate exchanger and the sodium−hydrogen ion channel found on the blood−brain barrier help regulate the pH level in both the blood and the brain. Changes in blood pH can be attributed to several things, including strenuous activity, environmental changes, and health complications. However, these mechanisms allow the blood pH to drop or rise outside that range for short periods of time without fear of complications. When participants were subjected to a short interval of high intensity exercise, their average pH dropped to 7.11.192 After five intervals of high intensity exercise, their average pH decreased to as low as 6.94.192 This continuous drop during exercise was due to the increase of CO2 in the blood. Once participants had longer time to rest, their respiratory function was allowed to increase their pH level back up to normal levels.192 Coso et al. investigated whether this result was affected by the physical condition of the person, i.e., trained versus untrained.195 This experiment showed that physical condition of individuals has little effect because the blood pH of both trained and untrained groups were remarkably similar.195 Although there are normal cases when blood pH drops out of the regulated range for short periods of time, concerns arise when the blood pH is not able to return to the safe range. This is a clear indication that there is some anomaly inside the body. Dangerous environmental changes significantly affect blood pH levels. Osborn investigated the effects of cardiac function on the 6.2. Kidneys and Lungs Regulation of pH levels in internal organs such as the kidneys and lungs are also vital to understand. For instance, kidneys contribute to organismal pH regulation by specific H+ buffering and secretory mechanisms. The bicarbonate buffer system entices the kidneys to reabsorb filtered HCO3− and convert it into excreted products.187 This reabsorption takes place in the proximal tubule. The distal nephron then excretes acid to be trapped in urine with either filtered anions or ammonia. Overall V DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 14. Summary of the Representative Studies Correlating pH to Physical Conditions ref testing condition Coso et al.195 bicycle exercise Hermansen et al.192 Momiyama et al.196 Kuehnle et al.194 treadmill or bicycle exercise out-of-hospital cardiac arrest pregnancy Osborn et al.193 Mani et al.37 hypothermia cancer blood sample capillary blood taken from finger capillary blood taken from fingertips fetal scalp blood and umbilical artery arterial blood lowest average pH 7.207 ± 0.051 for trained and 7.182 ± 0.080 for untrained individuals after one exercise interval (1 min workout and 4 min rest): 7.11, and after five exercise intervals: 6.94 unfavorable outcome: 6.93 ± 0.19 before birth: 6.98, and after birth: 6.90 light anesthesia and no artificial respiration: 7.16 pregnancy, hypothermia, and cancer. The next section discusses the expanding biomedical applications of pH sensing for both ex vivo (Section 7.1) and in vivo (Section 7.2) experiments. respiratory performance and blood pH by subjecting dogs to low temperatures and inducing hypothermia. Osborn showed hypothermia effect on pH under light anesthesia for an 11 kg dog with no artificial respiration.193 The results showed that when temperatures dropped, the respiratory function plummeted, and with that the arterial pCO2 increased.193 The rise in arterial pCO2 was assumed to be the cause of the drop in pH level, having an inverse relation with one another.193 Health complications are another important factor in blood pH levels. Momiyama et al. investigated the prognostic values of blood pH levels in patients resuscitated from out-of-hospital cardiac arrest and found that pH levels were much higher with an average of 7.26 in patients with favorable outcomes compared to those with unfavorable outcomes with an average of 6.93.196 A pH level of 7.05 was found to be the optimal cutoff level for favorable outcome in patients and the pH level of a patient with a favorable outcome never dropped below 6.95.196 Blood pH level is such a key factor in evaluating health that obstetricians use it to inform their decisions on their delivery method for pregnancies.194 When the fetal scalp blood pH was 7.20 or lower, a doctor would have to make the decision to perform a c-section or an instrumental vaginal delivery in order to avoid complications (Figure 6c shows good correlation between fetal scalp pH value and outcome umbilical pH value, used to study the reliability of fetal scalp pH values).194 However, results may not always be reliable and can be false at times. It was recommended to take at least two samples before making a decision as close to the time of delivery as possible.194 There is also direct correlation between pH and the presence of cancer cells in the blood (tumor cells). Mani et al. developed a working ZnO-based microfluidic pH sensor as a tool to examine circulating tumor cells (Figure 6d).37 The device achieved a Nernstian response of −43.71 mV/pH along with a high stability and drift of 3 mV/h, and a short response time between 26 and 32 s in the 1.68−9.18 pH range.37 When blood pH drops below 7.35, it is an important warning sign of functional anomalies inside the body. There are many causations for a drop in pH, such as strenuous activity that does not allow the blood to get enough oxygen, dangerous drops in body temperature that cause a rise in pCO2 levels, and health problems like cardiac arrest and pregnancy complications. The human body regulates this as much as possible with the help of the bicarbonate buffer system in the kidneys and the respiratory function of the lungs, so short-term drops in pH are usually normal. However, when a rise back to normal levels is not seen, concerns arise. Table 14 shows a summary of studies correlating pH to physical conditions. The importance of pH sensing in blood is evident through its myriad biomedical applications, which ranges from monitoring simple physiological changes such as exercising to more serious conditions as cardiac arrest, 7. pH SENSING IN BIOMEDICAL APPLICATIONS Among other pH applications, those of biomedical ones ranging from ex vivo to in vivo are numerous. This section discusses progress in pH sensors for both ex vivo and in vivo applications. Examples of ex vivo applications include the common urine and saliva tests, and the recent tooth decay assessment tests. On the other hand, the in vivo applications discussed include glioblastoma detection (the most frequent brain tumor), pHi and pHe sensing, oral hygiene assessment, monitoring of ischemic episodes, and sweat analysis. 7.1. Ex Vivo 7.1.1. Urine Tests. pH sensing of excreted bodily fluids is used to assess the condition of the patient. In particular, urine pH testing is a popular method, given the convenience of large sample collections and usefulness for assessing the treatment needs of a patient. The pH of the fluid acts as a biochemical marker, which is analyzed most commonly in two ways. Dipstick testing utilizes single-use test strips which report pH, presence of glucose, presence of certain proteins, and other important variables. 197 Though cheap and easy to use, dipsticks demonstrate significant pH measurement variability at more extreme values. Alternatively, urine pH can be found using a pH meter, which demonstrates more accuracy but requires personnel training and frequent calibration.197 Though the pH meter is ideal for guiding patient treatment decisions, the dipstick remains a valuable, if less precise, tool for patient use. Another factor to consider is that the pH of urine samples is unstable at higher temperatures, thus, affecting the results of analysis on tampered and old samples.198 This is relevant to the drug testing applications of urine pH testing. 7.1.2. Saliva Tests. Similar to urine tests, the pH of saliva is also often tested either as a medical guide or drug testing method. Difficulties arise in saliva pH testing because of the limited sample size and collection issues.199 Saliva is generally collected by spitting or stimulation by chewing or sucking. Once the sample is procured, a filtration device is often used to reduce the viscosity, making it easier to analyze later. Given that unstimulated saliva pH may range from 5 to 7, differences in these values can be quantified to indicate medical changes in the patient.200 Since these supplementary steps increase cost, while only providing a short window of sample viability, saliva testing is less common than urine testing. 7.1.3. Tooth Decay. Recently, applications of miniaturized pH sensors have gained increasing attention, especially in oral hygiene, due to its noninvasive nature, its small size, and its ability to quantify pH. For example, one of the common W DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Figure 7. pH sensing in biomedical applications (ex vivo tooth decay and in vivo glioblastoma and intra- and extracellular tests). (a) Experimental protocol for measuring pH of dental caries. The pH 7 buffer used after tooth measurement is to check the agreement of all measured points to ensure reproducible results. The final step is for sterilization before new measurements are taken.204 (b) Corresponding plot for carious (pH 6.16) vs healthy enamel (pH 6.99).201 (c) Data representation showing pH variation between healthy root (pH 6.85), arrested caries (pH 6.07), and active caries (pH 5.3). Statistically significant differences were accepted when the P-value is <0.05.204 (d) The reverse pH gradient evident within cancerous cells through multiple stages of acidosis (physiological extracellular pH (pHe) ≈ 7.4, acute acidosis pHe ≈ 6.8, and chronic acidosis pHe ≈ 6.7). pHi stands for intracellular pH.182 (e) The tumor itself can be identified by its acidic pH via nanosonophore assisted multispectral photoacoustic imaging.210 (f) Transport of protons from cytoplasm to extra-cellular fluid via a proton pump. ATP stands for adenosine triphosphate. (g) pHe map using chemical exchange saturation transfer (CEST) magnetic resonance imaging (MRI) in a subcutaneous Michigan Cancer Foundation-7 (MCF-7) mouse model with a Gaussian filter.211 Reproduced with permission from ref 204, 201, 204, 182, 210, and 211, respectively. Copyright 2018 American Chemical Society. Copyright 2016 Elsevier. Copyright 2018 American Chemical Society. Copyright 2013 Damaghi, Wojkowiak, and Gillies under CC-BY-3.0 https://creativecommons.org/licenses/by/3.0/. Copyright 2017 Springer Nature, Ltd. under CC BY 4.0 https://creativecommons.org/licenses/by/ 4.0/. Copyright 2015 John Wiley & Sons, Inc. ways to assist a dentist in evaluating dental caries, such as radiographic examination. However, early stages of teeth erosion cannot be detected using this method, due to its low sensitivity and high rate of false positives and negatives. Since dental caries are fairly common, a quantitative method to evaluating them would be highly beneficial; especially since early diagnosis would improve oral health, minimize tooth loss, and ultimately improve overall health and quality of life.204 Under normal conditions, the salivary pH is maintained around 6.7−7.3.205 When the pH of the saliva, specifically on applications of pH sensors is in detecting dental erosion. Dental erosion is defined as the loss of tooth structure by acid dissolution.201 The main cause of dental caries is due to the lack of oral hygiene, causing bacteria to grow on the teeth surface, and creating an acid as a byproduct.201 Figure 7a shows the sensing setup, where sensing and REs are both attached to the tooth.202 For the case of enamel cavities, the main clinical diagnosis is done by the visual cues and human judgment. However, this method is prone to human error and depends highly on the dentist’s experience and skill.203 There are other X DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Y 4.3−7.3 Ag/SiO2+ fluores cein isothio cyanate (FiTC) nanoparticles max standard error 0.08 pH 2−8 carbon microfiber standard error up to 0.6 pH Isfet silicon microelectrode −54 Mv/ Ph 3.77−7.27 each sample tested three times using different methods 5.3−6.8 Ta2O5 Chaisiwamongkhol et al.205 Gashti et al.213 Kaneto et al.212 Murakami et al.203 21 extracted human teeth (different types of cavities at different sites) 20 extracted carious tooth (divided into active/arrested) on the surface of teeth (6 subjects, age 25−28) commerical synthetic saliva and authentic human saliva attachment surface of a biofilm of the oral bacteria, Streptococcus salivarius Tabata et al.201 Isfet 4−8 −57.4 Mv/Ph 18 extracted human tooth samples Ratanaporncharoen et al.204 Isfet Ir/IroX 3−9 −56.96 Mv/Ph Ir/Irox 4.8−6.8 tantalum oxide (Ta2O5) ion sensitive field effect transistor (Isfet) Isfet sensing material setup testing surface/environment enamel teeth used For enamel samples Fujii et al. 7.2.1. Glioblastoma. Given that the pHe of solid tumors is known to be acidic, as shown in Figure 7d,e, the measurement of pH in this region is essential to monitoring and treating cancerous growth (physiological pHe ≈ 7.4, acute acidosis pHe ≈ 6.8, and chronic acidosis pHe ≈ 6.7).181,182,214 Current in vivo 202 7.2. In Vivo ref Table 15. Summary of the Key Works on Tooth Decay Ex Vivo Measurements sensitivity pH range standard error <0.3 pH repeatability dental plaque or in dental cavity, is below a critical value of 5.5, it is an indicator of potential dental decay. Therefore, pH measurement can be used to assist the diagnosis of dental caries. Indeed, multiple papers reported that the sensors were able to detect significant pH differences between sound enamel (pH 6.99) and carious enamel (6.16) as well as healthy root (pH 6.85), arrested caries (pH 6.07), and active caries (pH 5.30) (Figure 7b,c).201,204 Other studies have shown that irregular salivary pH could be a sign of diseases such as anxiety disorder and gingivitis, besides tooth decay.206−209 In general, there is great interest in developing a miniaturized pH sensor that works reliably on different areas of the mouth. However, there are numerous obstacles to this development. First, while the sensing surface of most sensors is extremely small (0.015 mm long and 0.75 mm wide204), the overall footprint of the pH sensor device is huge, and unsuitable for fine oral measurements. In addition, the rough surface of human teeth may also provide challenges to sensors with flat sensing surface.203 Researchers were able to overcome these issues by taking advantage of iridium oxide’s (Ir/IrOx) unique properties, of being mechanically strong and chemically inert. Ir/IrOx was utilized in a needle-like shape pH sensor. The probe diameter was 300 μm, and was capable of measuring the pH on all types of surfaceseven in deep cavities.201 Moreover, due to the sensor’s size, it is able to measure in between teeth, which has traditionally been the hardest. Using an Ir/IrOx needle as the sensing material effectively solved the issues of the sensor being too large and measuring on rough surfaces. In fact, the pH sensors that adopted IrOx as the sensing material have all shown relative success (near Nernstian sensitivity and very high repeatability) in ex vivo testing.201,204 As a result, iridium oxide is recommended for oral pH-sensing applications. Other sensing materials used, though less common, include tantalum oxide (Ta2O5) and carbon microfiber. Most reported pH sensors for monitoring tooth decay use the ISFET configuration (Section 4.1). Currently, most studies test teeth samples externally (ex vivo), before moving to in vivo testing, since the ex vivo testing environment is relatively more stable. While a lot of progress has been made on oral pH sensors, there is much work to be done to develop a model that can work in clinical situations, as the environment becomes more complicated as pH sensors are moved from an external tooth to within the mouth. Within the mouth, other factors may interfere with the pH sensor’s reading. For example, the effect of saliva on the pH sensor has not been fully investigated. Human saliva has numerous components, including many charged proteins. This may interfere with proper functionality of the pH sensor.204 Also, the effect of oxygenated saliva on the sensors has not been fully explored.205 Table 15 summarizes key works on tooth decay ex vivo measurements. For tooth decay applications, a sufficient pH range is 5−7 with calibration ranges usually extending farther on both sides, with ISFET configuration (Section 4.1) and metal oxides (Section 3.2.1), such as Ta2O5 and IrOx. relative standard deviation (RSD) among eight measurements were 0.09 and 1.67% in the first cycle, 0.06 and 1.09% in the second cycle, and 0.03 and 0.57% in the third cycle standard error for sound enamel area is 0.05 pH, for carious lesion area is 0.01 pH Review DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 16. Summary of the Representative pHe Sensing Works ref Chung et al.22 Munteanu et al.222 Marzouk et al.226 Das et al.227 sensing material experimental setup IrOx carbon fiber modified with reduced graphene oxide and syringaldazine IrOx palladium oxide thin film sensitivity pH range potentiometric voltammetric pH microsensor −69.9 ± 2.2 mV/pH −60 ± 2.5 mV/pH 4−10 potentiometric extended gate field effect transistor (EGFET) −63.5 ± 2.2 mV/pH −62.87 mV/pH 2−10 2−12 during the early stages of wound healing. In addition, cancer cells have been found to have a lower pHe than a normal cell (which in turn promotes drug resistance and increases invasiveness221). While these cancer cells maintain a close to normal pHi, their pHe is found to be more acidic, ranging from 6.2 to 6.9,219 than the surrounding blood and tissues, which have a pH of 7.4.222 This acidity is caused by the dependency on anaerobic metabolism where an excess amount of lactic acid is produced by glycolysis, due to insufficient removal from tumor vasculature.223 The acidity is also caused by a large amount of CO2.223 pHe can also have an effect on the uptake of anticancer drugs and how tumor cells respond to therapy.223 Because the maintenance of pHe plays a key role in physiological and cellular functions, it is important to have an accurate, precise, and reliable device that will determine the pHe value in vivo. There are many devices that are used for detecting pHe, including microelectrodes, radionuclide imaging, MRI relaxometry, and MRS.211 Microelectrodes are miniaturized electrodes with ion-selective membranes that are sensitive to changes in pHe . 224 One configuration that uses electrodes is the potentiometric pH sensor, which uses an ion-selective sensing electrode and RE to determine pHe. ISFET and EGFET configurations can also be used, but they require a transistor in their configurations.112 In addition, microelectrodes can be interdigitated to increase surface area.121 Radionuclide imaging uses PEmT to detect pH-sensitive radiolabeled probes.211 Fluorescence utilizes the properties of dyes to measure pHe optically.211 MRI relaxometry examines the pH-dependent relaxation rates.225 Figure 7g shows an example of a pHe map, using MRI and a Gaussian filter. MRS is a noninvasive technique that analyzes changes in pHe.211 These techniques and devices for measuring pHe must be highly sensitive, be highly selective, display long-term stability, and have the ability to be used in vivo. Table 16 shows different experimental set-ups to measure pHe and their sensitivities. In most cases, metal oxide (Section 3.2.1) sensing films are utilized in different configurations (Section 4) due to their biocompatibility and facile deposition methods. Although the reported pH ranges are relatively wide, pHe variation is usually within less than 1 pH unit around the neutral value (i.e., pH 7). 7.2.3. Oral Hygiene. Salivary pH marks an important biological marker for many bodily diseases, including periodontal disease such as gingivitis, and dental caries, which was discussed in Section 7.1.2. Due to saliva’s noninvasiveness and ease of collection, storage, and shipping, saliva has great potential for pH testing.206 While, in recent years, physical monitoring systemssuch as heart rate monitors and temperature sensorshave evolved, developing an accurate pH sensor is still a challenge in health monitoring. Overcoming this challenge would enable additional biomedical sensing applications and lead to a more personalized medical care.228,229 tumor pH measurement is largely performed by pH-sensitive positron emission tomography (PEmT) radiotracers, magnetic resonance spectroscopy (MRS), magnetic resonance imaging (MRI), and optical imaging.215 PEmT uses radiolabeled DMO to determine the pH gradient, but is largely inaccurate and imprecise. MRS and MRI monitor metabolic and physiologic processes. MRS uses chemical shifts between pH-dependent and -independent resonances to determine pH.216 Overall there is room for development in in vivo pH measurement of cancerous cells, and the methods available require consideration of their respective drawbacks and advantages. 7.2.2. Intracellular and Extracellular pH. pHi is also another useful biomedical indicator to suggest the conditions within a cell, and thus the health of that cell. In vivo measurement of pHi ideally should be sensitive and not affect the subject. There are a number of methods to obtain this pH, including nuclear magnetic resonance spectroscopy, pH microelectrodes, and pH-sensitive fluorescent reporters.217 One can determine pHi from the negative logarithm based 10 of the acid dissociation constant (pKa) and pHe of a weak acid or base exposed to the cell. Nuclear magnetic resonance uses the ratio between protonated and deprotonated phosphate groups to determine the pH with great accuracy. Once pH-sensitive microelectrodes are prepared and calibrated, they act as miniaturized pH meters and are best applied to larger cells to provide accurate results. Fluorescent indicator dyes can measure varying pH within a cell by close monitoring of the cell under a microscope. There are also pH-sensitive fluorescent proteins which act similar to the dye but can be coded within the cell itself.217 Consideration of each method’s sensitivity and accuracy and set up is necessary to determine the ideal intracellular sensor for a particular application. pHe is the pH of the extracellular fluid outside of the cell. Many mechanisms exist that export H+ ions, produced by oxidative metabolism and fermentation, into the extracellular fluid.211 Figure 7f is a simplified schematic of H+ ions being transported to the exterior of the cell from the interior by a proton pump, establishing a concentration gradient. Acid−base homeostasis plays a vital role in maintaining physiological and cellular responses.218 Healthy cells maintain a normal pHe of around 7.4 by biological buffers.219 In a highly acidic or basic pHe, cellular functions, such as enzyme activity and DNA synthesis, are greatly diminished or ceased completely.218 Thus, a deviation from the normal pHe may be an indication of a disease or a physiological abnormality. Another example is that deviation from the normal pHe can impair the immune response, especially in acidic environments, due to inhibition of lymphocyte activity.218 An instability of pHe can also be an indication of metabolic abnormalities.22 Furthermore, insulin resistance in skeletal muscle cells may be correlated to the lowered pHe, as the insulin receptor’s phosphorylation level, also known as activation, was diminished.220 Interestingly, pHe is found to range from 5.7 to 6.1 Z DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Figure 8. pH sensing in biomedical applications (in vivo oral hygiene and ischemia tests). (a) pH readings from commercial synthetic saliva with actual saliva.205 The green dotted line represents measurement from commercial synthetic saliva sample, and the blue dotted line represents the real saliva sample’s results, having pH values of 7.16 and 7.51 respectively. The cyclic voltammetry scan rate is 4 V/s to remove the influence of any oxygen reduction reaction, and the solid lines represent results from prepared synthetic saliva samples of various pH values. (b) A digital image and a (c) crosssection of the phone case setup that is used for colorimetric analysis of pH strips,231 and (d) test result from three male individuals (aged 25−37) during 16 h of regular day time, showing variation in pH.231 (e, f) An overview of the sodium and pH-sensing system, with Na+ sensitivity of 60 mV/ decade vs Ag/AgCl reference electrode (RE), long-term stability of 1 week, and low detectability limit of 10−4 mol/L.228 (g) Different components of the sensor inserted into the papillary muscle of the rabbit heart used to indicate pH drop with ischemic episodes.232 (h) 1,2-Naphthoquinone (1,2NQ) pH sensor insertion into the rat brain for in vivo ischemia testing. The results showed a normal pH of 7.21, 7.13, and 7.27 in the striatum, hippocampus, and cortex regions, respectively, and a decreased pH upon global cerebral ischemia of 6.75, 6.52 and negligible change in the striatum, hippocampus, and cortex regions, respectively. R.E., W.E., and C.E. stand for reference, working, and counter electrodes, respectively.233 (i) IrOx pH sensors array tested on the right ventricle of a human heart in OCP configuration, with a response time as low as 0.5 s and a sensitivity of 69.9 mV/pH. The results indicated a drop in pH from 7.4 to 6.55 during ischemic episodes. Location of pH sensors specified by colored circles (navy, pink, purple).234 (j) An overview and a zoom in on the same pH sensors array in part (i) on a flexible surface.234 (k) Time vs pH during ischemia periods, for a human heart showing a drop in pH from 7.4 to 6.6.234 Reproduced with permission from refs 205, 231, 228, 232, 233, and 234, respectively. Copyright 2017 Royal Society of Chemistry. Copyright 2013 Royal Society of Chemistry. Copyright 2018 National Academy of Sciences. Copyright 2002 Elsevier. Copyright 2016 American Chemical Society. Copyright 2013 John Wiley and Sons, Inc. Many different factors are in play for oral and salivary pH, studying the corrosiveness and effects of different sugary beverages on oral pH.230 pH measurement can either be done within the mouth, or the saliva can be collected in a test tube. In vivo testing is harder than ex vivo testing, since the uneven surface within the oral cavity including the human’s diet and time of the day during measurement. Having a pH sensor that can closely monitor pH variations may enable more targeted studies, for example, AA DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 17. Summary of the Representative Oral pH Sensing Works ref testing environment Watanabe et al.236 inside the test subject’s mouth Oncescu et al.231 human subjects Lee et al.228 in human subjects, and in solutions test subjects with different level of periodontal conditions commercial synthetic saliva and authentic human saliva collected human saliva Baliga et al.206 Chaisiwamongkho et al.205 Hans et al.230 setup an indicator strip, a reference strip and a flash diffuser sensor placed on a flexible film single electrode digital pH meter sensing material pH range sensor abilities IrOx 5.0−9.0 pH strip 5.0−9.0 operating time: 19 h, maximum error of 0.15 pH repeatability depends greatly on phone model (i.e., camera quality) single electrode digital pH meter in vitro maximum standard error is 0.11 pH 2−8 pH glass type electrode digital pH meter max standard error 0.08 pH standard error lower than 0.3 pH for coffee, Pepsi, fruit drink, but up to 1.25 pH for milk measure the pH of a testing strip in a self-designed case, by taking a picture and analyzing the color through a phone’s application.231 However, the accuracy of this system can be questionable, as many different factors (such as temperature and humidity) can alter the results. The case is shown in Figure 8b, and the exact design is shown in Figure 8c. The study showed great variation between different test subjects. Figure 8d shows test results from three male individuals (aged 25−37), during 16 h of regular day time, showing great variation in pH. Moreover, the smart-phone model and camera could greatly change the results. As a result, the system’s accuracy needs further improvements. Worth mentioning, sodium sensors also prove to be a good indicator of personal health. In addition, they can monitor sodium intake for people who have high blood pressure, hypertension, diabetes, and obesity. Thus, it could prove valuable to develop a system that integrates both a pH sensor and sodium sensor to monitor a wide range of different conditions. Lee et al. incorporated sodium sensors onto a stretchable plastic film, and then attached it onto a retainer.228 Figure 8e shows the how the sensor is attached, and Figure 8f shows the sensing platform, showing sensitivity for Na+ ions of −60 mV/decade concentration vs Ag/AgCl RE, long-term stability of 1 week, and low detectability limit of 10−4 mol/L. Overall, oral and salivary pH levels are accurate indicators of oral and overall body health. In most cases, the saliva is collected and tested externally, due to the ease of the procedure. While sensors that can be orally inserted can be a good alternative, there are only few research studies in this area. Finally, the use of sodium sensors should be considered and integrated with the pH sensor for both expanding applications and validation of sensor readings. Table 17 summarizes key works on oral pH sensors. Commercial and research quality pH sensors have been reported for oral pH studies, and the results indicate the validity of the investigated pH systems for oral pH assessment. Saliva pH lies in the 6−7 pH range with temporary perturbations taking place based on what is being eaten or drunk. An extended lower pH (∼5.3−6.16 pH) value is indicative of tooth decay (Section 7.1.3). 7.2.4. Ischemia. Out of the numerous biomedical applications pH sensors have, ischemia is certainly among the most important ones. Ischemia is defined as an inadequate blood supply to an organ or part of the body, especially the heart muscles. When oxygen supply is cut off to the cells, the cells are only able to create ATP through glycolysis, and produce H+ as a byproduct. As of now, ischemia is very hard to detect, and a sensor that can detect ischemic metabolism can result in proves to be a challenge, as current rigid sensors and plastic boards are not suited for oral insertion. As a result, there are not many reported cases where actual sensors are attached on the inside of the oral cavity, most likely due to the convenience of ex vivo testing. Developing a sensor that is compatible to work inside the human mouth requires extensive research, since the environment within the mouth (uneven surface, different temperatures, and disturbance from mouth-movement) can lead to inaccuracy and repeatability issues in the measurements. Meanwhile, carrying out tests on saliva outside of human body does not change its pH, and greatly increases the accuracy of measurement. Unlike tooth testing, where it is necessary to test directly on the tooth surface (since removing the tooth from the body is not plausible), saliva can be easily collected into a test tube. Thus, in most cases, commercial pH sensors are used to measure the pH externally for several reasons. For instance, although commercial pH sensors offer great consistency and accuracy, they are usually bulky and not easily implantable. Since a wide pH range is not needed (the normal pH of saliva is 6.7− 7.4), most sensors used in this application operate between a range of 5.0−9.0.230 As a result, during these ex vivo tests, it is simply easier and more efficient to use a commercial pH sensor instead of developing a new sensor, since a commercial pH sensor can accomplish the same purpose and is readily available for purchase. Developing pH sensors that can operate under the unique environment of the mouth still faces challenges. For example, the oxygenated environment of the mouth has an effect on pH measurement. Researchers were able to successfully combat this challenge, as shown in Figure 8a, as the carbon fiberbased pH sensor showed comparable results from commercial synthetic saliva and real saliva samples, where both are oxygenated biological samples with pH values of 7.16 and 7.51 respectively. This means that the sensor was able to overcome the interference caused by an oxygenated environment, using cyclic voltammetry with a scan rate of 4 V/s to remove the influence of any oxygen reduction reaction.205 Another example is studying the role that the physical state of food plays in its cariogenic potential. The longer the sugar is stuck to the teeth, the longer the bacteria act on sugars and produce acid, leading to development of dental caries. In this case, liquid sugars has lower cariogenic potential than solid and sticky sugars, as they tend to stick to the teeth surface due to their property of adherence.230 While many researchers focused on developing a miniaturized pH sensor that is flexible and insertable, Oncescu et al. overcame this challenge with a rather simple solution: the researchers developed a system that uses a smart-phone application to AB DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review stability, wide range of pH detection, rapid response times, minimal drift, outstanding chemical selectivity, and high durability.234 The sensing setup consists of a RE and a miniaturized sensing electrode or array. The size of the sensing surface on the sensing node is 1 mm, making it suitable for in vivo testing. Arrays of sub-millimeter-scale and precision pH sensors distributed on thin elastic membranes were also reported for in vivo testing.234 The structure and zoom-in view of the individual sensors is shown in Figure 8j. On the other hand, several studies used a nontraditional photometric sensor for ischemic pH measurements235 and achieved successful realtime monitoring. Using the photometric pH sensor, the severity of the episode is correlated with a decrease in pH and restricting the blood flow. In addition, it was found that during ischemic episodes, there is a noticeable decrease in the pH of the surrounding tissues. In Figure 8k, time (in minutes) is plotted against pH, and showed significant reduction in pH during ischemia for a human heart from a pH of 7.4−6.6. In conclusion, there are many obstacles facing the development of an accurate sensor for use in detecting ischemia, including the flexibility and ability of the sensor to work in the heart’s environment. The available studies suggest that traditional ways of pH analysis do not work well given the obstacles, and newer ways of pH sensing could result in better implantable sensors with higher sensitivity, repeatability, lower drift, and overall better quality. Table 18 summarizes works on pH sensors for monitoring ischemia. The physiological pH range for monitoring ischemia is pH 6−8, and, similar to pHe (Section 7.2.2), metal oxides (especially IrOx) are commonly used as the pH-sensing material. 7.2.5. Sweat Analysis. Sweat, a fluid produced by the body to reduce body temperature, is also a common external bodily fluid for measuring pH and is critical to assess the physiological health of an individual. Sweat is secreted by the sweat gland. Sweat has a pH range between 5 and 7, which is acidic in comparison to a neutral pH in blood. It is composed of different substances, both organic and inorganic.240 The substances include amino acids, minerals, lactic acid, urea, salts, fatty acids, and trace elements.241 Despite these many compounds in sweat, approximately 99% of sweat is composed of water.242 Since sweat includes cations such as sodium, potassium, and magnesium, it is important that the sensor to be used is selective for H+, so that the cations are not falsely accounted for while determining the measurement of pH. Sweat is a useful bodily fluid because it can be stored for the long-term, is easy to measure noninvasively, and is less prone to alterations compared to other body fluids.242 In fact, humans perspire at a rate between 300 and 700 mL/day, and during exercise, may perspire at a rate of 1.4 L/h. Thus, sweat is a plentiful source for pH measurements.243 Assessing the pH and analytes of sweat is important as it can be used for early detection of diseases in the human body, as it is correlated to the blood. One such disease is diabetes. Diabetes often causes more acidic pH levels in the body. With diabetes, an individual often develops diabetic ketoacidosis, a condition in which many ketones are produced, causing a very acidic blood pH.244 Another is cystic fibrosis, a condition that affects the digestive system and lungs. A basic pH (∼9) can be an indicator of cystic fibrosis, due to the lack of reabsorption of bicarbonate ions.245,245−247 In addition to being a useful way to detect diseases, sweat can also be used to detect drugs. According to the pH partition theory, bases are likely to accumulate in acidic individualized treatment and improve patient’s overall condition.237 In addition, pH variation is known to indicate metabolic function abnormality, and accurate monitoring of the pH can greatly assist clinicians by giving them valuable information about the condition of the organ, resulting in more accurate diagnosis and better treatment.234 Specifically, in the case of heart ischemia, researchers believed that both the magnitude of the pH shift and the duration of ischemia are important in the heart’s ability to resume normal function.237 In the brain, having an accurate pH sensor can allow scientists to gain a better understanding of the role pH plays in brain diseases.233 During ischemia periods before an Ischemic stroke,238 an inadequate blood supply prevents the accumulation of extracellular potassium ions, and cuts off the oxygen supply, therefore triggering the anaerobic metabolism that produces lactic acid. Most importantly, pH in the affected area usually experiences notable decrease. Since these events happen simultaneously, pH value and ion concentrations, such as potassium (K+), lactate (C3H5O3−) or sodium (Na+), are good indicators and early detectors of ischemia. In fact, during ischemic episodes, potassium ions’ concentration can increase up to four times its normal level.238 As a result, an integrated sensor, which has pH, K+, and other ion detectors integrated onto one sensor could be beneficial. Doing so would provide higher confidence in the data, due to the potential of validation of sensor readings. There is still much work to be done in order to develop a sensor that is suitable for biomedical use. As shown in previous works,238,239 the sensors developed all performed well in salt and buffer solutions, with high sensitivity and repeatability. However, when sensors are tested under in vivo conditions in living animal tissues, the measurements are less accurate and tend to fluctuate, with significantly lower sensitivity and repeatability. In addition, there are many other difficulties involving manufacturing miniaturized pH sensors used for in vivo testing. For example, glass electrode sensors are currently most commonly used; however, these electrodes are large and bulky, easy to break, and difficult to be miniaturized for implants.233 Moreover, the complex environment within the heart requires the sensor to have certain qualities, such as flexibility. The continuous movement, complex curved structure, low stiffness, and heterogeneous surfaces pose substantial engineering challenges for mapping the heart’s pH.234 In addition, due to the invasive nature of inserting a sensor into the body, the specific procedure is important. Also, the placement of the sensor may have an effect, due to the uneven surfaces of the human body. Figure 8g shows the configuration of the pH sensor tested on the rabbit’s heart papillary muscle. The device was used to observe drops in pH with ischemic episodes. Figure 8h shows the 1,2-naphthoquinone (1,2-NQ) pH sensor tested in a rat brain. The results showed a normal pH of 7.21, 7.13, and 7.27 in the striatum, hippocampus, and cortex regions, respectively, and a decreased pH upon global cerebral ischemia of 6.75, 6.52 and negligible change in striatum, hippocampus, and cortex regions, respectively. Figure 8i shows an array of IrOx pH sensors tested on the right ventricle of a human heart in OCP configuration, with a response time as low as 0.5 s and a sensitivity of −69.9 mV/pH. The results indicated a drop in pH from 7.4 to 6.55 during ischemic episodes. Researchers who used traditional methods mostly opted to use IrOx as the sensing material, citing its high sensitivity and AC DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX 4.0−10.0 −69.9 ± 2.2 mV/ pH 6.8−7.8 −24.2 mv/pH to −73.4 mv/pH max standard error 0.09 pH 5.8−8.0 Table 19. Summary of the Representative Works on pH Sweat Monitoring ref Caldara et al.243 Dang et al.247 Guinovart et al.252 Guinovart et al.251 Guinovart et al.253 Yin et al.68 Curto et al.254 develop a flexible sensor that can detect pH, K+, Na+ ions salt solutions to assess ischemia evaluate effectiveness of the miniature glass electrode and dog heart the photometric sensor for ischemia thin, stretchable sensor for monitoring ischemia explanted rabbit hearts and a donated human heart Anastasova et al.239 Steward et al.235 Chung et al.234 IrOx membrane dog heart Curto et al.249 IrOx develop a sensor which could detect ischemic metabolism Soller et al.237 1,2naphthoquin one rat brain (in vivo) in vivo monitoring meter for pH ratiometric microelectrochemical meter miniature glass electrode/ photometric sensor open circuit potential (OCP) 0.07 μA/pH tantalum oxide Ir/IrOx pH buffer solutions ischemic rabbit papillary muscle myocardial ischemia (heart) ischemic heart Rai et al. Marzouk et al.232 Zhou et al.233 ion sensitive (ISFET) sensing material 238 fluids such as sweat.248 Therefore, basic drugs usually accumulate in the individual’s sweat.242 Devices that measure the pH of sweat include fabric/flexible plastic-based sensors, and epidermal-based sensors. They can be worn during exercise and could be used for real-time monitoring of the sweat from an individual.249 The fabric/flexible plasticbased sensors have textiles and fabrics with special properties, and are in constant contact with the skin.250 Epidermal-based sensors (i.e., elastomeric stamps and tattoos) are usually printed directly on the skin.251 Table 19 summarizes works on pH sweat monitoring. Potentiometric (i.e., OCP) configuration is the most widely used in sweat analysis, due to the simplicity of the setup. mean pH at 2 h: standard error around 0.09 pH relative standard deviation of 0.9−3.4% ±0.05 pH standard error ±0.2 pH Review tested medium issue investigated ref Table 18. Summary of the Representative Works on pH Sensors for Monitoring Ischemia setup sensitivity pH range repeatability 6.5−8.0 6.4−7.4 Chemical Reviews setup pH range sensitivity response time (s) optical pH sensor 2−10 205 Hz/lux 110 potentiometric 5−9 −11.13 ± 5.8 mV/pH <8 potentiometric 3−11 potentiometric 3−7 ∼ −54 mV/pH <25 potentiometric 3−9 −90 mV/pH ∼ 50 potentiometric barcode pH sensor microfluidic platform optical pH sensor 4−7 1−12 −63.7 mV/pH <60 <20 4.5−8 8. STATUS QUO Complementing the discussed sections on various pH-sensing biomedical applications, from the ex vivo (Section 7.1) to the in vivo (Section 7.2), this section focuses on notable progress and the status quo for wearable and implantable pH-sensing systems. 8.1. Wearable pH-Sensing Systems Wearable electrochemical sensors are promising for a variety of biomedical applications, due to their noninvasiveness and ease of use. Examples include detecting hormone levels, oxygen concentration, ion concentration, and pH levels. Their noninvasiveness gives these devices a potential for large-scale use. Their ease of use makes them accessible to the general population. An ideal wearable electrochemical pH sensor would be miniaturized, while still preserving high performance, high reliability, and a Nernstian response.247 They should also have low manufacturing costs and great flexibility, to follow the contours of the human body.240 Examples of wearable sensors are illustrated in Figure 9. Figure 9a shows the conceptual design of a band-aid wearable pH sensor on a human model, utilizing cotton yarns dyed with carbon nanotube ink, with a response time <60 s, and sensitivity of −59 mV/pH in potentiometric configuration vs a Ag/AgCl RE in the 3−11 pH range.252 Figure 9b shows the structure of an ISFET pH sensor configuration for wound monitoring, using Al2O3 gate dielectric and sensitivity of −50 mV/pH in the 3.3−11.4 pH range.255 These pH sensors are useful for early detection of a disease, assessing human performance, and other useful applications.250 These pH sensors often measure pH from sweat, wounds, and saliva. Since there AD DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Figure 9. Wearable pH sensors. (a) Illustration of band-aid sensor on human model. The sensor utilized cotton yarns dyed with carbon nanotubes ink and showed a response time <60 s, and a sensitivity of 59 mV/pH in potentiometric configuration vs Ag/AgCl reference electrode (RE) in the 3−11 pH range.252 (b) Schematic of a wearable device integrating the ion sensitive field effect transistor (ISFET) configuration, with an Al2O3 gate dielectric and a sensitivity of −51.2 mV/pH in the 3.3−11.4 pH range.255 (c) Representative application of wearable pH electrochemical sensor.265 (d) Entire epidermal tattoo pH sensor experimental setup attached to the wrist to measure pH of human perspiration.251 (e) Wearable example of a tattoo epidermal-based sensor for biomarkers (such as lactate) assessment of sweat.250 (f) Schematic of a smart wound care pH sensor integrated with a uric acid sensor.263 (g) Schematic of a full-setup of a smart bandage, including an integrated pH sensor, temperature sensor, drug loaded hydrogel, and an electronic heater to release drugs on-demand.264 Reproduced with permission from refs 252, 255, 265 , 251, 250, 263, and 264, respectively. Copyright 2013 Royal Society of Chemistry. Copyright 2017 American Chemical Society. Copyright 2014 MDPI (Basel, Switzerland) under CC-BY-3.0 https:// creativecommons.org/licenses/by/3.0/. Copyright 2013 Royal Society of Chemistry. Copyright 2014 Elsevier. Copyright 2018 Electrochemical Society. Copyright 2018 John Wiley and Sons. are other substances in sweat and other outer bodily fluids, the device must be highly selective and only measure the activity of H+. In addition, some of these sensors may be wireless or stretchable to allow for ease of movement and portability,247 and may integrate other sensors as well, such as temperature and glucose sensors,256 as the integrated disposable sweat monitoring strip. The integrated sensors include pH, glucose, and temperature sensors. The pH sensor included a PANI sensing electrode in OCP configuration vs Ag/AgCl RE and showed nonlinear voltage dependence of pH in the 4−7 pH range.256 A common application of these wearable sensors is measuring the pH in sweat. As mentioned in Section 7.2.5, sweat is a very popular body fluid to measure due to its ease of measurement. However, it does contain other substances, both organic and inorganic, such as sodium ions and glucose.240 Sweat usually has a normal physiological pH on the more acidic side, ranging from 5 to 7 due to its composition of minerals, lactic acid, and urea.241 Measuring analytes in sweatsince it is correlated to blood can be an indicator for diseases, such as diabetes and hypochloremia.247 Wearable sweat sensors can be categorized into two types: fabric/flexible plastic-based sensors, and epidermal-based sensors. Fabric/flexible plastic-based sensors have constant contact with the skin.250 Textiles, which are tough, flexible, and react to external stimuli,257 are printed on the fabric AE DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 20. Summary of the Representative Wearable pH Sensor Developments ref setup sensing material pH range sensed medium <8 graphite-polyurethane composite 5−9 sweat potentiometric <60 cotton yarns dyed with carbon nanotubes ink 3−11 sweat potentiometric ∼ −54 mV/pH <25 3−7 sweat potentiometric −90 mV/pH ∼50 poly(aniline) (PANI) associated with the reversible emeraldine salt (ES)−emeraldine base (EB) transition PANI 3−9 sweat potentiometric optical pH sensor −63.7 mV/pH 4−7 4.5−8 sweat sweat potentiometric −58 ± 0.3 mV/ pH −51.2 mV/pH <20 PANI methyl red, bromocresol green, bromocresol purple and bromothymol blue dyes electropolymerized PANI 4.35−8 wounds Al2O3 3.3−11.4 sweat curcuma-dyed cotton/curcuma-dyed polyamide buffer solutions wounds sweat potentiometric Guinovart et al.252 Bandodkar et al.251 Karyakin et al.253 Nyein et al.68 Curto et al.249 Guinovart et al.262 Nakata et al.255 Rahimi et al.267 Bandodkar et al.251 response time (s) −11.13 ± 5.8 mV/pH −59 mV/pH Dang et al.247 Giachet et al.266 sensitivity ion sensitive field effect transistor (ISFET) textile-based optical sensor potentiometric −53 mV/pH 58 PANI curcuma dyed cotton: 6.5−8.5; curcuma dyed polyamide: 8.5−13.0 4−10 potentiometric −50 mV/pH 25 PANI 3−7 monitoring systems that are integrated with uric acid sensors,263 temperature sensors, drug loaded hydrogel, and an electronic heater to release drugs on-demand.264 Table 20 summarizes key wearable pH sensor developments. Despite the many advantages that these wearable, noninvasive pH sensors present, these wearable pH sensors still have serious limitations and challenges. With normal body movement, epidermal-based pH sensors may show mechanical malfunctions. Some pH sensors also may not have long-term stability and may not be sensitive to highly acidic or highly basic solutions.250 In addition, wearable sensors are often not worn over a long period of time, due to the physical discomfort it may cause for the user.260 Therefore, many improvements need to be made on these wearable devices in order to promote large-scale public use and accessibility. plastic-based sensors. These textiles must not affect the pH of the skin where the sensor is measuring.258 Fabrics that are often used include wool, cotton, and nylon, since they have the chemical and physical properties for an optimal electrochemical pH sensor. A representative wearable pH sensor is provided in Figure 9c. Epidermal-based PANI pH sensors in potentiometric configuration vs Ag/AgCl ink electrode, as shown in Figure 9d, measure pH by having conformal contact with the skin.251 These sensors exhibited a sensitivity of −50.1 mV/pH (increased with stretching up to −59.6 mV/pH), a response time between 10 and 25 s, and a batch- to-batch relative standard deviation of 4.63% (n = 4) in the 3−7 pH range. Similarly, elastomeric stamps and tattoo sensors (Figure 9e) printed directly onto the skin have been reported for detection of other biomarkers in sweat, such as lactae.250 These types of sensors use various pHsensing configurations, such as potentiometric, ISFET and EGFET configurations (discussed in Section 4). Epidermal pHsensing hydrogel fibers have also been reported for detecting wound healing and skin disorders, thus, overcoming the susceptibility to long-term degradation, as in the case of electrochemical electrodes exposed to sweat.259 Interestingly, wearable sensors are especially applicable for monitoring infant physiological health and, more importantly, are useful for early detection of potential life-threatening health conditions. Due to an infant’s inability to communicate verbally, sensors must provide clinicians and parents with critical health information at home or in a neonatal intensive care unit (NICU), while avoiding irritation, interruption of sleep, or causing stress to the infant.260 Another application of wearable pH sensors is measuring a wound’s pH. Wounds are particularly costly to a patient and may fail to heal properly.261 Whereas the pH of healthy skin tends to be more acidic, in the range 5−5.5, wound pH tends to be more basic, typically having a pH ranging from 7 to 8.5.262 There is a correlation between pH values and wound healing. The healing process may also be affected by the pH value beneficially, or detrimentally.262 Therefore, it is important to have tools that can accurately monitor a wound’s pH. Figure 9f,g shows schematics of different smart wound 8.2. Implantable pH Sensing Systems While wearable pH sensors may offer a noninvasive, easy-to-use option for users, they are limited to measuring only substances found outside of the body. This mainly includes fluids produced by the body, such as sweat, saliva, urine, and open wounds. Implantable sensors, on the other hand, allow for wider opportunities of application. Though significant work must still be done, development of a miniaturized, implantable sensor presents the opportunity to monitor real-time pH levels anywhere in the human body, including blood, the esophagus, and brain tissue. pH is an effective parameter in the blood for many circumstances, as discussed in Section 6.3. In these circumstances, such as for sickle cell disease, it is beneficial to determine pH measurements in vivo.174 A miniature fiber optic pH sensor was produced for physiological use and was tested in the jugular vein of a sheep.174 Utilizing fiber optics, a pHsensitive dye contained within a H+ permeable envelope is implanted in the area of interest, in this case the vein, and the optical density of the dye is then measured by illuminating the dye through a single strand fiber and sensing the back scattered light through another optic fiber strand that is connected to a remote light detector.174 The sensor functioned within the physiological pH range of 7.0−7.4, and when tested in vivo, the AF DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Figure 10. Implantable pH sensors. (a) Silica-based fiber optic (●) and IrOx electrochemical (■) measurements of tissue pH in response to injections of sodium bicarbonate into the peritoneal cavity at 3 and 18 min, showing a reasonable response time of ∼5 s, sensitivity of −57.9 mV/pH after 85 min, and a drift of 0.4 mV/h.175 (b) Analytical performance of intravascular sensors can be influenced by (i) thrombus formation on the sensor surface, (ii) “wall effect” caused by positioning the sensor near metabolically active endothelial cells, and (iii) vasoconstriction around the sensor.268 Reproduced with permission from refs 175 and 268. Copyright 2001 Elsevier. Copyright 2002 Elsevier. Table 21. Summary of the Representative Implantable pH Sensors ref Frost et al.268 sensing material reference material (1) Copolymer of n-butylmethacrylate and 2methacryloxyloxyethyl posphorylcholine. (2) pH indicator reagent dyes (e.g., phenol red). setup (1) Fiber optics (2) Microdialysis catheter, sensors fiber optics Goldstein et al.174 Cao et al.162 pH sensitive dye contained within a H+ permeable envelope IrOx Ag/AgCl frequency sensing Grant et al.175 • Silica optical fibers Ag/AgCl • Optoelectronics Ag/AgCl • Electrochemical potentiometric Hao et al.269 Wencel et al.270 • IrOx polyvinylcholride (PVC) matrices onto carbon fiber electrodes 8-hydroxypyrene-1,3,6-trisulfonic acid (HPTS) encapsulated in sol−gel matrix fiber optics location response time pH range 42 s 7.0−7.4 sensitivity blood vessels jugular vein of a sheep esophagus wall of a pig brain tissue of a rat brain tissue of a rat human tissue 5s 6.6−7 −51.1 to −57.7 mV/pH −57.9 mV/pH <1 s 6−8 −58.4 mV/pH <2 min 6−8 1.9−12 capsule with pH sensors in order to monitor gastroesophageal reflux disease (GERD).162 The pH sensor utilized iridium oxide as the sensing material vs Ag/AgCl RE, and demonstrated Nernstian values between −51.1 and −57.7 mV/pH in the pH range 1.9−12.162 When tested in the esophagus wall of a pig and compared to a commercial pH sensor, it showed comparable results and performed better when introduced to an alkaline solution of pH 11.162 Unlike the commercially available sensors, this device has no limit on monitoring duration and is a suitable option for monitoring real time pH.162 Table 21 summarizes key implantable pH sensors. With significant advantages seen in implantable pH sensors, the most important is the wide range of application possibilities. Several studies successfully monitored in real-time the pH level in blood, the esophagus, and the brain. However, the biological response of the body and the lack of reliable performance of the implanted devices pose substantial challenges.268 Hao et al. demonstrated a viable route to address the fouling of the sensing membrane due to biological deposits.269 This is done using an H+ selective membrane with polyvinyl chloride (PVC) matrices onto carbon fiber electrodes in a potentiometric setup. The sensor exhibited improved antifouling property and reversible and repeatable results, even after three hours in vivo (inside rat’s brain, with response time <1 s and −58.4 mV/pH sensitivity). Furthermore, Wencel et al. demonstrated a robust ratiometric fiber opticsbased pH sensor in human tissue.270 The sensor used a 8hydroxypyrene-1,3,6-trisulfonic acid (HPTS) in hydrogel matrix results were comparable, if not better than a commercial glass electrode, with a response time of 0.7 min.174 Implantable pH sensors can also help with the treatment of patients that had traumatic brain injury.175 Grant et al. developed silica-based fiber optic and IrOx potentiometric sensors to monitor the pH of brain tissue.175 When tested in vitro in blood, the fiber optic sensor demonstrated a Nernstian value of −57.9 mV/pH, and the electrochemical sensor showed similar results at −57.8 mV/ pH vs Ag/AgCl RE, with a response time of 5 s and a drift of 0.4 mV/h.175 Both were implanted in the brain of a Sprague− Dawley rat model, and sodium bicarbonate was injected into the peritoneal cavity of the rat to change the pH of the brain tissue.175 Both sensors reacted similarly; however, they displayed different results, due to a calibration issue (Figure 10a).175 The test ran for 50 and 165 minthrombus formation may occur for longer durations.175 Although miniature fiber optic pH sensors show promise in monitoring the pH in vivo, results are vastly affected by biological responses.268 Some of these responses include thrombus formation due to the absorption of proteins on the sensor’s surface, the “wall effect” caused by placing the sensor near metabolically active endothelial cells, and reduced blood flow from vasoconstriction around the sensor (Figure 10b).268 These responses are reduced when implanting in larger blood vessels.268 Implantable sensors can also be used in different organs. Cao et al. developed an implantable, battery-free, and wireless AG DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Figure 11. Stability, repeatability, and reproducibility assessment of pH sensors. (a) Drift at pH 6 with time for three different fin field effect transistor (FinFET) in liquid gate configuration devices normalized at the initial Vth at time (h) = 30: D1 (single wire with fin thickness (TFin) = 30 nm), D3 (three wire FinFET with TFin = 20 nm), and D4 (five wire FinFET with TFin = 30 nm). The device used HfO2 as gate dielectric and sensing material and silicon nanowire channels, showed a sensitivity of −57 mV/pH vs Ag/AgCl reference electrode (RE), and excellent stability when tested for 105 h. The single wire FinFET (D1) showed a drift of 0.13 mV/h, a 3-wire FinFET (D3) showed a drift time of 0.1 mV/h, and a 5-wire FinFET (D4) showed 0.12 mV/h.178 (b) Drift characteristics of the undoped ZnO and aluminum doped ZnO (AZO) nanostructured pH-EGFET sensors measured within pH = 7 for the duration of 12 h, with drift values ranging from 1 to 17 mV/h.58 (c) Fitting curves of 13 tests after 24 h of hydrothermal hydration of Ir/ IrO2 at 220 °C, recorded over 40 days. Measured cell potentials were carried out against a commercial Ag/AgCl (saturated KCl) RE; the device sensitivity range is −59 to −70.5 mV/pH.271 (d) Hysteresis of ∼5 μA measured in ZnO/SiNWs-based pH sensor in extended gate field effect transistor (EGFET) configuration, as a time vs drain source current, with sensitivity of −66 mV/pH.136 (e) pH plotted against drain source current for CuSbased pH sensor in EGFET configuration, with error bars typically representing standard deviation through pH 7−4−7−10−7 cycle and sensitivity of −23.3 mV/pH. The sensor exhibited repeatability with a relative standard deviation (RSD) of 0.04%, 0.02%, and 0.38% for glass, tungsten and Si substrates, respectively.274 (f) Test results in Coke, orange, coffee, and water from a commercial pH sensor and a flexible polyaniline (PANI) nanopillars-based pH sensor in OCP configuration vs Ag/AgCl RE, showing good repeatability. The columns represent average values from five readings, and the error bars typically represent the standard deviation of the measurements. The sensor has sensitivity of −60.3 mV/pH, drift of 0.64 mV/h in pH 5, drift of 0.49 mV/h in pH 7, response time of 1 s, and sustained performance over 1000 mechanical bending cycles.60 (g) Reproducibility test of five ZnO nanotube electrodes and five ZnO nanorod electrodes vs Ag/AgCl RE in OCP configuration, with a sensitivity of −45.9 and −28.4 mV/pH, respectively, showing excellent reproducibility, with error bars representing relative standard deviation of 5%.36 (h) Plot of normalized resistance and pH value across five single-walled carbon nanotube (SWNTs) sensors in resistance-based configuration with response time varying from 2.26 s in pH 5 to 23.82 s in pH 9 and sensitivity of 236.3 Ω/pH. A droplet has been placed and removed 10−15 times before the sensor showed the depicted stable response. The dots show the average values, and the error bars represent standard deviation from the five devices.118 (i) Plot showing pH value and output voltage of 14 indium tin oxide (ITO)/ polyethylene terephthalate (PET)-EGFET electrode samples and (j) Their sensitivity (S) (average of −50.1 mV/pH).143 Reproduced with permission from refs 178, 58, 271, 136, 274, 60, 36, 118, and 143, respectively. Copyright 2015 American Chemical Society. Copyright 2013 Hindawi under CC-BY-3.0 https://creativecommons.org/licenses/by/3.0/. Copyright 2017 Springer Nature, Ltd. Copyright 2013 Electrochemical Society. Copyright 2017 Elsevier. Copyright 2017 Elsevier. Copyright 2009 MDPI (Basel, Switzerland) under CC-BY-3.0 https://creativecommons.org/licenses/by/3.0/. Copyright 2011 2009 MDPI (Basel, Switzerland) under CC-BY-3.0 https:// creativecommons.org/licenses/by/3.0/. Copyright 2012 Elsevier. AH DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 22. Summary of the Stability Results of Various pH Sensors ref setup Wang et al.272 Wang et al.271 open circuit potential (OCP) OCP Wang et al.147 extended gate field effect transistor (EGFET) FinFET Rigante et al.178 Li et al.168 Kurzweil et al.20 ion sensitive Field effect transistor (ISFET) ISFET pH range sensing material reference material stability 1−13 IrO2 Ag/AgCl 1.68−12.47 Ir/IrO2 Ag/AgCl 1−13 aluminum-doped ZnO (AZO) Ag/AgCl 3−10 HfO2 Ag/AgCl Ag/AgCl 1.6−12.2 indium tin oxide (ITO)/ polyethylene terephthalate (PET) Ru/RuO2 • Single wire FinFET: drift time of 0.13 mV/h. • 3-wire FinFET: drift time of 0.1 mV/h • 5-wire FinFET: 0.12 mV/h (all over 105 h) drift rate <1.7 mV/h over ∼8.5 h Ag/AgCl 0.13 mV/pH (pH 4) 0.38 mV/pH (pH 7) 7.31 mV/pH (pH 10) (over 13 min) • Over 86 h: potential drift 0.3 mV/h Zhou et al.177 OCP 2.22−11.81 IrO2 Ag/AgCl Vanamo et al.273 OCP 1−9 poly(3,4-ethylenedioxyt hiophene) doped with poly(styrenesulfonate) anions (PEDOT/PSS) film Ag/AgCl average of 10 calibration curves over period of 2.5 years: k and slope = −58.4 mV/pH • E2 had a stable, gradual, negative potential drift • E3 remained relatively constant without much potential drift over 40 days Al dosage of 3% best drift rate at 1.27 mV/h over 12 h • First 30 h: potential drift 0.6 mV/h • After overnight short-circuiting, measured for 30 min: −3.24 mV/h and +4.08 mV/h time of 0.1 mV/h, and a 5-wire FinFET showed 0.12 mV/h (Figure 11a).178 Results of different testing set-ups show further improvements based on the type of material used for the sensing electrode. IrOx shows the promise of being a superior material for pH sensing in biological media. It demonstrated a fast and stable response in aqueous, nonaqueous, nonconductive, and corrosive media.20 When used as the pH-sensing material for a miniature multiparameter sensor chip, it demonstrated a drift value of 0.3 mV/h over 86 h, with a rate of 0.6 mV/h for the first 30 h.177 Research has also been done on the pH sensing and drift characteristics of hydrothermal AZO nanostructured sensors.147 With 20 measured samples, Al dosage of 3% has the best drift rate at 1.27 mV/h for 12 h (Figure 11b).147 Alternative RE materials have also been investigated. For instance, a solid state thin-film RE composed of titanium/gold/ silver/silver chloride (Ti/Au/Ag/AgCl) has been assessed for effectiveness as a RE.168 This fabricated RE performed comparably with the standard silver/silver chloride (Ag/AgCl) RE, at a drift rate of 1.7 mV/h over 8.5 h.168 Table 22 summarizes stability results of key pH-sensing reports. A popular method for fabricating oxide thin films for sensing electrodes is thermal oxidation. However, the dry films produced exhibit significant aging effects.271 Sufficient hydration during fabrication and preparation of electrodes has proven to be a key factor in reducing the amount of potential drift.271 One method studied was high-temperature hydrothermal hydration treatment where Ir/IrO2 electrodes were subjected to hydration at 220 °C for 24 h and then soaked in deionized water. Tested within a pH range of 1.68−12.47, electrodes that underwent this method exhibited good stability over the course of 40 days, attributed to the more orderly crystal arrangement and high content of OH− groups as shown in Figure 11c.271 Measured cell potentials were carried out against a commercial Ag/AgCl (saturated KCl) RE in OCP configuration, and the device on an optical fiber tip and exhibited reliable performance, indicated by the low drift (0.003 pH in 22 h in lab setup and 0.004/h in vivo). These studies highlight the rapid advancements in pH sensing, and collectively, show the importance of pH sensing in biomedical applications. 9. CHALLENGES Evidently, throughout Sections 4−8, there are common challenges that persist, mainly those pertaining to the reliability of pH-sensing measurements. This section discusses reliability issues, such as stability, repeatability, and reproducibility. All these issues are critical milestones for a pH-sensing system to be suitable for biomedical applications and mass production. Furthermore, the modeling and theoretical aspects of pHsensing mechanisms are discussed, and the challenges facing accurate predictions that comply with experimental results are highlighted. 9.1. Stability of pH-Sensing Devices Although significant progress has been made on the improvement of pH-sensing systems, stability is still a challenge. Electrodes suffer from potential drift over time which makes it practically difficult to obtain consistent values. This phenomenon is known to be influenced by a number of factors including the testing setup, sensing electrode material, RE material, and fabrication method. Most glass electrodes and ISFET configurations show poor long-term stability.20,147 The EGFET, on the other hand, has shown better results.147 A less commonly used setup is the FinFET in liquid gate configuration. The device had critical features of 20 nm, used HfO2 as gate dielectric and sensing material, and used a silicon nanowire channel. This device showed a sensitivity of −57 mV/pH vs Ag/AgCl RE and had excellent stability when tested for 105 h. The single wire FinFET showed a drift of 0.13 mV/h, a 3-wire FinFET showed a drift AI DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 23. Summary of the Repeatability Assessments of Various pH Sensors ref Huang et al.136 Sabah et al.274 Huang et al.283 Wu et al.284 Yoon et al.60 Prats-Alfonso et al.282 Nguyen et al.275 Salvo et al.285 sensing material ZnO/Si nanowires (NWs) CuS/glass CuS/W CuS/ Si Si NWs reference material Ag/AgCl repeatability setup extended gate field effect transistor (EGFET) EGFET EGFET touch panel film (TPF) polyaniline (PANI) nanopillar IrOx Ag/AgCl Ag/AgCl The sensor is slightly more sensitive to acidic than basic solutions hysteresis: CuS/glass: 0.48 mV CuS/W: 11.72 mV CuS/Si thin film: 11.05 mV hysteresis: about 0.8 mV hysteresis: relative standard deviation (RSD) of less than 2% in the 3−13 pH range hysteresis: 1.9 mV compared to initial voltage of 473.1 mV AgCl hysteresis: 1.5−0.5 mV OCP IrOx AgCl OCP graphene oxide (GO) AgCl Standard deviation is 0.2 pH units in 4−7 pH range and 0.4 pH units in pH 2 Repeatability from three trails showed a standard deviation of ±0.2 pH units Ag/AgCl Ag/AgCl sensitivity range is −70.5 mV/pH with a constant intercept at 0 pH (standard reduction potential difference between sensing and RE) of 800.5 mV over the 40 days. When the same device was hydrated in DI water at room temperature for 24 h, a significant drift in the intercept at 0 pH from 901 to 563 mV was observed. Another method studied is referred to as the carbonate melt oxidation. A uniform iridium oxide film is coated on the surface of an iridium metal wire through oxidation of the wire in a carbonate melt.272 In another study, using an Ir/ IrOx electrode, 10 calibration curves were obtained in pH 1−13 over the span of 2.5 years for a long-term stability test. The results showed a stable linear response with the calibration curves overlapping, eliminating the need for frequent calibration.272 Short circuiting has also been proven to reduce the need for frequent calibration, increasing the stability of electrodes.273 After short circuiting two identical solid-contact ion-selective electrodes overnight, both drifted slightly toward their original potentials, one at −3.24 mV/h and the other at 4.08 mV/h.273 ISFET open circuit potential (OCP) OCP standard deviation (RSD= × 100) for the first and second pH mean measurements is then calculated. The repeatability is obtained by taking the difference between the RSD values at the same pH values. Figure 11d shows a representative hysteresis plot for ZnO coated SiNWs sensing electrode in EGFET configuration with a sensitivity of −66 mV/pH and a hysteresis value of ∼5 μA. Another example for assessing pH sensors’ repeatability is by calculating standard deviations for repeated measurements. Figure 11e,f shows an example of a repeated plot from pH 7−4− 7−10−7 cycle, and multiple measurements in various environments, respectively, with error bars typically representing standard deviation from multiple runs (not explicitly defined in reports).60,274 Figure 11e shows pH plotted against drain source current for CuS-based pH sensor in the EGFET configuration, with standard error bars showing repeatability through pH 7−4−7−10−7 cycle and sensitivity of −23.3 mV/ pH. The sensor exhibited repeatability with RSD of 0.04%, 0.02%, and 0.38% for glass, tungsten and Si substrates, respectively, and Figure 11f shows test results in Coke, orange, coffee and water from a commercial pH sensor and a flexible PANI nanopillars-based pH sensor in OCP configuration vs Ag/ AgCl RE, showing good repeatability. The columns represent average values from five readings, and the error bars typically represent the standard deviation. The sensor has sensitivity of −60.3 mV/pH, drift of 0.64 mV/h in pH 5, drift of 0.49 mV/h in pH 7, and a response time of 1 s. The flexible PANI pH sensor sustained performance over 1000 mechanical bending cycles, making it suitable for integration with other high performance flexible electronic components that have been demonstrated to survive such bending behavior for mechanically dynamic applications.276−281 Nguyen et al. showed three measurement runs overlaid as another route for displaying pH sensor’s repeatability, indicating acceptable repeatability from an IrOxbased pH sensor in OCP configuration vs Ag/AgCl RE with sensitivity of ∼−60 mV/pH and response time of 30 s.275 Materials used for the sensor play a significant role in the sensor’s repeatability. While most materials show relatively good repeatability, as shown in Table 23 (RSD of 2% for hysteresis, and pH standard deviation of less than 0.2 pH units across multiple trials), those that have a sensing electrode made of IrOx showed slightly lower repeatability. In an IrOx sensor, the residual standard deviation of the slopes (sensitivity) from 10 independent measurements is 3.4%.282 Similarly, Nguyen et al.275 utilized IrOx sensors and reported maximum pH standard 9.2. Repeatability of pH-Sensing Devices Like stability, repeatability is one of the biggest challenges facing pH sensors. Repeatability of a pH sensor means that it behaves the same way and outputs similar results every time it is in a similar solution. Although perfect repeatability is almost impossible to achieve, minimizing deviation is necessary in order to have confidence in the pH sensor’s measurement. This issue becomes especially important as the precision of the data increases, especially in critical applications, such as biomedical sensors. Factors that contribute to the repeatability of a sensor include the sensor setup, sensing and RE’s material, and fabrication process. Normally, the repeatability is measured by carrying out multiple trials of pH measurement in the same buffer solution, and calculating the standard deviation between trials. In this case, the quantitative measure of a sensor’s repeatability is its deviation (in pH units), i.e., the lower the deviation, the better the repeatability. Another route to evaluate repeatability is hysteresis, defined as a measure of how much the pH sensor is impacted by its previous reading (i.e., history). In EGFET, hysteresis is, typically, measured via a series of (Ids vs time) curves, where the drain-source current (Ids) is plotted against time. Cycles of pH measurements were carried out between pH values 2 and 12, for 1 min each. The exact pH for the buffer solutions and time submerged in each buffer could change based on the particular experiment. The relative standard deviation AJ DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review electrode, and Eref the standard reduction potential of the RE. The chemical reaction that takes place at the RE is shown below in eq 21, and has a reduction potential of around 220 mV deviation across three trails of 0.4 pH units, significantly higher than the average of 0.2 pH units for graphene oxide. The reasoning given for this behavior is the subtle differences of nanoscale pore sizes in the iridium oxide film, quality, and uniformity of Ag/AgCl REs.275 Hence, repeatability might especially be an issue when using IrOx-based pH sensors. On the other hand, pH sensors using Si NWs have consistently shown relatively better repeatability. A possible reason for Si NW’s good repeatability is that the single crystal Si NWs structure is more reliable and stable.283 Furthermore, a hybrid of Si NW and ZnO sensor exhibited a larger surface area and more binding sites than the pristine Si NWs for adsorbing additional ions, thus improving overall quality of the sensor.136 Sabah et al. compared different kinds of copper sulfide hybrid sensors.274 The hybrid material also played an essential role in the sensor’s stability. While the method of fabrication, loop cycle, and loop length are all kept the same, changing the sensor’s material from CuS/W to CuS/glass decreased the sensor’s hysteresis from 11.78 mV to 0.48 mV.274 Finally, the setup of a sensor does not seem to have a significant correlation with its repeatability. Similar setups had sensors that exhibited both good and poor repeatability. Overall, while many details factor into the repeatability of the pH sensor, the most important one is the material used in the sensor. Another insight regarding repeatability is that sensors with good overall quality, such as high sensitivity, rapid response time, and low drift, have higher repeatability. 9.2.1. Mixed Versus Specific Reactions. A set of chemical reactions with associated standard reduction potential needs to be accounted for when extracting values for surface potential and pH. In the case of ZnO, for example, eqs 13−19 below are all possible chemical reactions that can take place at the sensing surface.132 These reactions are included to demonstrate the fact that all possible reactions involving Zn ion redox reactions have negative potential, contrary to the typically reported positive values. This indicates competing mixed reactions where Zn ion redox is not a dominant reaction. However, during pH measurement, many different reactions (some expected, some unexpected) can happen between the sensing surface and the solution. Solution tested is not always pure and may contain ions that can also react with the sensing surface and cause unwanted exchanges. Under ideal conditions, only a few specific reactions happen during pH measurement and the sensor follows the Nernstian behavior (−59.18 mV/pH). Zn 2 + + 2e F Zn − 0.7618 V (13) ZnO2 2 − + 2H 2O + 2e F Zn + 4OH− − 1.215 V (14) ZnOH+ + H+ + 2e F Zn + H 2O − 0.497 V (15) Zn(OH)4 2 − + 2e F Zn + 4OH− − 1.199 V (16) ZN(OH)2 + 2e F Zn + 2OH − 1.249 V (17) ZnO + H 2O + 2e F Zn + 2OH − 1.260 V (19) − − AgCl + e− F Ag + Cl− The reduction potential for REs are known, Ag/AgCl electrodes (which are used in almost all pH sensor setups) is known to be around 220 mV. Following the slope calculated from tests, the standard reduction potential (in mV) can be calculated at pH 0. At pH 0, the pH term in eq 20 equates to zero, the potential of RE is around 220 mV, and the voltage difference is the measurement of the pH sensor. Worth mentioning, this is a simplified example, whereas in practical cases there are other elements in the pH cell that consume portions of the observed voltage measurements. For instance, reference electrodes structure includes a porous or semipermeable membrane that allows ionic exchanges between the reference solution and the solution under test. The difference in ions’ mobilities and concentrations within the reference electrode (such as 3 M KCl) and the outside solution (i.e., calibration buffers or test solutions) creates a concentration gradient and potential barrier. This is referred to as the liquid junction potential (LJP). The LJP cannot be directly measured and typically consumes a fraction millivolt to few millivolts across, resulting in ±0.01 to 0.05 pH error in the pH measurement. If sensing electrode’s reduction potential is close to one of the known equations, then it can be reasonably concluded that the specific reaction is dominant. And expect a stable and repeatable performance. Sometimes, the effect of mixed potential can be obvious. In their paper, Ghoneim et al.132 calculated the standard reduction potential at the ZnO electrode to be +869 mV, while eqs 13−19 all have negative values, indicating clear mixed potential in this particular case. The involvement of mixed potential could result in many different unwanted consequences. For example, the sensitivity of the sensor will deviate from the ideal Nernstian response (higher or lower depending on the exact reactions that are happening on the surface). In addition, because the mixed potential reactions can lead to ions accumulating on the surface and ions dissolving in the solution, mixed potential could lead to corrosion and passivation of the sensing surface.132 In their paper, Meruva et al.,286 it was stated that the mixed potential was correlated to the partial pressure level of oxygen (pO2). The reason behind it is that under higher pO2, more oxygen resides on the sensing surface and interacts with the sensing surface more readily than in normal pO2 conditions.286 In another case, Macdonald et al.287 developed a W/WO3 pH sensor to test the pH under high temperature and pressure. The sensor showed lower values compared to the Nernstian sensitivity. The reduced sensitivity is attributed to a mixed potential. The provided explanation was that this behavior is expected because the W/WO3 sensor is not an equilibrium system, but rather displays a mixed potential resulting from a balance between a partial anodic process (eq 22) and a partial cathodic reaction (e.g., eq 2).287 132 In their paper, Ghoneim et al. explained ways to test if multiple reactions are taking place. The actual measured full cell potential is the difference between the two half-cell reactions, varying depending on the pH of tested solution given by eq 20: Ecell = Esense − Eree f − 59.18 mV*[pH] (21) W + 3H 2O → WO3 + 6H+ + 6e− H + + e− → (20) 1 H2 2 (22) (23) As shown, it is usually hard to identify what mixed potential reactions are taking place. Most of the time, it is a speculation. Hence, it is essential to understand that sometimes the In the equation, Ecell represents the measurement voltage difference, Esense the standard reduction potential of the sensing AK DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 24. Summary of the Representative pH-Sensing Works and the Likely Associated Reactions ref sensing material standard reduction potential (mV) reported potential W+3H2O → WO3 + 6H+ + 6e− W/WO3 Macdonald et al.287 H + + e− → Meruva et al.286 ionophore tridodecyla mine (TDDA) IrOx Chung et al.234 Yang et al.288 Lonsdale et al.289 Batista et al.114 Ghoneim et al.132 1 H2 2 +465 mixed potential +922 Ir(OH)3 ⇌ IrO(OH)2 + H+ + e− IrOx reduced graphene (rGO) RuO2 +676 Ir(OH)3 ⇌ Ir(OH)2O− + H+ IrO(OH)2 ⇌ IrO(OH)O− + H+ 2[IrO2(OH)2·2H2O)]2− + 3H+ + 2e ⇌ [Ir2O3(OH)3·3H2O]3 Ru(IV)O2 + e− + H+ ⇌ Ru(III)O(OH) ZnO +636 ZnO +869 +611 likely reaction mixed potential IrOx dominant reactions mixed potential involving multiple oxidation states of Ir and/or reactions with negative standard reduction potential RuO2 dominant reactions mixed potential involving oxygen evolution oxygen evolution reactions mixed potential involving oxygen evolution Table 25. Summary of the Representative Reports on pH Sensors and Their Reproducibility Results ref setup Fulati et al.36 Voigt et al.290 Bartic et al.291 Li et al.118 potentiometric Lue et al.143 extended gate field effect transistor (EGFET) ion sensitive field effect transistor (ISFET) interdigitated electrodes (IDEs) resistance-based reproducibility sensitivity 5 ZnO nanotube electrodes/5 ZnO ZnO nanotube: −45.9 mV/pH ZnO nanorod electrodes nanorod: −28.4 mV/pH pH cycles −54−59 mV/pH sampling experiments of the drain current in time 5 sensors −62 mV/pH 30 samples in 3 different runs −47.2, −48.6, and −51.0 mV/pH 236.3 Ω/pH speculation can result in wrong conclusions about what is actually happening at the sensing surface. One potential way to is to scan the surface to see if there are structural changes to deduce the possible mixed potential reactions. It is reasonable to deduce that corrosion and passivation of the sensing surface may indicate the occurrence of mixed potential. Another challenge is that it is not only hard to understand the exact reactions, but also no empirical solution exists. There are major difficulties to control specific reactions that occur on the sensing surface, and such methods are rarely discussed in recent reports. When mixed potential is interfering with the pH sensor’s proper functionality, it may be helpful to change the sensing material. In conclusion, mixed potential happens when certain ions in the solution/environment react with the sensing surface in an undesired way. A mixed potential artifact might show up as an unexpected standard reduction potential value or serious repeatability and stability issues. Further research is needed in order to accurately determine mixed potentials and suggest solutions to circumvent its effects. Table 24 summarizes representative works and the likely associated reactions. Although there are always competing mixed reaction, a relatively higher stability is correlated with a dominant reaction involving the ions at the sensing surface, pertaining to their reversible oxidation−reduction reactions. sensing material ZnO nanotubes/ZnO nanorods diamond-like carbon (DLC) and Ta2O5 poly (3-hexylthiophene) semiconducting polymer single-walled carbon nanotubes (SWNTs) indium tin oxide (ITO) thin films pH range 4−12 1−12 2−10 5−9 2−12 9.3. Reproducibility of pH-Sensing Devices Reproducibility, in which the same response to the same stimuli is given from device to device, is not to be confused with repeatability, in which the same response is given by the same device multiple times to the same stimuli, as shown in the previous subsection. Reproducibility of the response across multiple devices is important because devices often need to be replaced due to deterioration of materials and due to hygienic reasons.243 In pH reproducibility tests, multiple devices are usually evaluated in specific buffer solutions. Ideally, devices should be reproducible in terms of sensitivity, stability, drift, and response time. Fulati et al. reported a great example of reproducibility measurements across 10 pH potentiometric sensor electrodes (5 ZnO nanotube sensing electrodes and 5 ZnO nanorod sensing electrodes) at pH 6. In their paper, tests of 5 ZnO nanotube electrodes and 5 ZnO nanorod electrodes vs Ag/AgCl RE in OCP configuration, with sensitivity of 45.9 and 28.4 mV/pH, respectively, showed excellent reproducibility, relative standard deviation of 5%.36 Figure 11g shows the results from the reproducibility test by Fulati et al. Diamond-like carbon thin films (DLC)- and Ta2O5-ISFET reproducibility and pH response for low pH values are also shown.290 Another example of good reproducibility is in Bartic et al.’s experiment, using interdigitated electrodes with organic semiconductors.291 In Li et al.’s experiment of SWNTs sensors, in resistance-based configuration with response time varying from 2.26 s in pH 5 to 23.82 s in pH 9 and sensitivity of 236.3 Ω/pH, reproducibility AL DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Figure 12. Modeling of pH sensors. (a) A simplified schematic of the space charge region (SCR), Helmholtz region (HR), and Gouy region (GR). (b) Dependence of potential on capacitance components at the semiconductor/electrolyte interface.300 (c) Schematic of examples of point defects and extended defects. Reproduced with permission from ref 300. Copyright 2009 MDPI (Basel, Switzerland) under CC-BY-3.0 https://creativecommons. org/licenses/by/3.0/. constant, and T is the temperature in Kelvin.292 Another important model to consider is the Nernst equation: was reported across five sensors, finding the normalized resistances, as provided in Figure 11h, with a standard deviation of less than 5%.118 Noteworthy, a droplet has been placed and removed 10−15 times before the sensor showed the depicted stable response. To test the reproducibility of an ITO/PET electrode on EGFETS in the wide pH range between 2 and 12, Lue et al. reported on testing 14 ITO/PET electrodes. These 14 electrodes showed an average sensitivity of 50.1 mV/pH with a standard deviation of ±4.1 mV/pH. Figure 11i shows the output voltage response of the samples at different pH, and Figure 11j shows the variation in sensitivity across the 14 samples.143 In addition, Lue et al. prepared 30 samples in 3 different runs, reporting sensitivities of −47.2, −48.6, and −51.0 mV/pH across the runs.143 These samples were also evaluated in solutions of different temperatures of 25, 40, and 50 °C. At high temperatures, sensitivity dramatically decreases to −21.8 mV/ pH. Lastly, these samples were evaluated over a long period of time (55 days) and still showed a sensitivity of higher than −45 mV/pH and had linearity above 98.5%, which shows that these electrodes have both great stability and reproducibility.143 Table 25 summarizes representative reports on pH sensors and their reproducibility results. Although few examples of good reproducibility have been reported, reproducibility discussions and results are rarely reported in most pH-sensing works. E= (25) where E is the reduction potential, R is the universal gas constant, T is the temperature in Kelvin, F is Faraday’s constant, z is the number of electrons, [X]out is the concentration of ions outside the cell, [X]in is the concentration of ions inside the cell.293 Particularly for the glass electrode, many models for the potential response have been developed. One model is the Donnan boundary potential model, stating that H+ and sodium ion diffusion through the glass membrane and the potential generated was caused by the difference of the diffusion rates of different ions.292 However, later, it was discovered that H+ ions do not diffuse into the glass membrane.292 Nikolsky, in 1937, theorized the ion exchange equilibrium theory, whereby establishing that the exchange of the protons by a sodium ion on a glass site leads to a potential difference.294 Later, in 1967, Durst presented the idea that adsorption of H+ on the glass surface leads to a potential response.295 In 1994, Bauke’s ideas were published, which stated that the glass surface groups and the ions in aqueous solution are in dynamic equilibrium. The potential on the glass membrane and the potential difference between the glass and solution is caused by a dissociation mechanism.292,296 A negative potential on the glass membrane is produced by a net- charge density by negatively charged groups.292,296 Cheng also proposed his hypothesis for the potential mechanism, whereby an electrode is a double-layer capacitorbased on the Guoy-Chapman model and Poisson−Boltzmann equation.292,297 Interestingly, this viewpoint does not consider the potential difference to be caused by redox reactions. Instead, a potential is generated through the equation: 9.4. Modeling of pH-Sensing Devices Many models have been developed to understand the mechanism of how pH is measured from the pH-sensing configurations. Devices, which measure the electrical potential, have sensing materials that are sensitive to changes in the activity of H+.292 An ideal device would be highly selective to H+ and has an ideal Nernstian sensitivity of −59 mV/pH. To understand how pH is found through electrical potentials, the following equation is used: i F yz zz pH = pHS + (E − ES)jjj k 2.30RT { [X]out RT ln ZF [X]in Ecapacitance = Eindicator + Ereference (24) where pHS is the pH of the standard reference solution, E is the cell potential of the solution, ES is the potential of the standard reference solution, F is Faraday’s constant, R is the universal gas (26) The sensing electrode and RE derive their potentials form the capacitance law, E = q/C, where q is the charge density and C is the capacitance.297 This model considers how H+ and OH− AM DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review electrons from electrode to solution as the redox potential of the electrolyte is lower than the Fermi level.300 The Helmholtz region is adjacent to the semiconductor surface and is a result of the adsorption of ions or of surface bonds between the solution species and surface.300 Whereas the capacitance of the space charge region is considerably dependent on potential, the capacitance in the Helmholtz region has a little dependence. It is important to note that the potential in this region is entirely dependent on the interactions between semiconductor and electrolyte solution.300 The concentration of H+ near the solid surface [H+S ] is found through the following equation: absorb to the electrode surface in acidic and basic solutions, respectively.297 Since Cheng’s proposal was drawn from the Gouy−Chapman model, it is important to consider how this model explains a double layer of a surface. At the electrical interface, Helmholtz proposed that charges at the interface formed two layers of opposite charges.298 However, this model did not account for adsorption of ions on the surface and diffusion of ions, and Guoy and Chapman determined there is a Gouy−Chapman diffuse double layer composed of an uneven distribution of anions and cations.299 Because this model did not account for highly charged double layers, the Gouy−Chapman model was later modified by Stern to create the Gouy−Chapman-Stern model of what occurs at the metal/electrolyte interface to consider the finite size of an ion. It establishes that there is a Stern layer of ions adsorbing to the electrode surface along with the Gouy− Chapman diffuse layer.299 These concepts are important because the electrode that measures pH interacts with H+ and other ions at the metal/electrolyte interface, thus creating a capacitance, which is used to find the potential. Overall, there are three regions in the model on the metal oxide/electrolyte solution interface where capacitance is measured: the space charge region in the semiconductor, the Helmholtz region, and Gouy diffuse layer, as mentioned earlier. A simplified schematic is provided in Figure 12a of the regions for reference. The total capacitance is found through the equation: 1 1 1 1 = = = c cSC cH cG i eψ y HS+ = [Hb+] expjjj− 0 zzz k kT { where is the Boltzmann distribution bulk concentration. Counter ions could also neutralize the surface charge by adsorbing to the surface. In this region, the capacitance, which is assumed to be constant, is found through the equation: σ0 CH = ψ0 − ψβ (32) where σ0 is the surface charge and ψ0 and ψβ is the surface potential of the solid and mean potential of absorbed counterions at the plane, respectively.300 For pH devices in general, one model used to explain the mechanism is the site-binding model, which examines the chemical and electrical interactions on the surface of an oxide and solution. Based on the Gouy−Chapman−Stern model, an oxide surface, which contains a surface charge from H+ and OH−, has two layers of constant capacitance and also includes an outer diffuse layer, as mentioned previously.138 The model states that the changes in the surface potential voltage (i.e., at the sensing layer and electrolyte interface) is a function of the number of binding sites on the sensing membrane.113 The SiteBinding Model is given by eq 33. (27) where CSC is the capacitance of the space charge region, CH is the capacitance of the Helmholtz region, and CG is the capacitance of the Gouy region.300 Correspondingly, the potential is found through: V = VSC + VH + VG (28) Figure 12b shows the relationship between potential and capacitance components at the electrolyte/semiconductor interface. The space charge region in the semiconductor, between the semiconductor surface and the electrode bulk, is where the majority carriers are depleted and an associated electric field arises.300 The equation for capacitance (CSC) in this region can be found through the equation: CSC ÅÄÅ ÑÉ−1/2 ÅÅ 2 ji kT zyzÑÑÑÑ j Å = = ÅÅ jV − Vfb − zÑ ÅÅ qε0εSND jj ∂U q zz{ÑÑÑÑÖ k ÅÇ 2.303(pH pzc − pH) = i qψ 1 yz qψ + sinh−1jjjj zzz kT k kT β { (33) where pHpzc is the pH value at the point of zero charge, k is the Boltzmann’s constant, T is the absolute temperature of the system in Kelvin, and q is charge of the electron, and β is the sensitivity parameter (defined in eq 34).301 ∂Q S ÄÅ ÉÑ1/2 ÅÅ ij kT yzzÑÑÑÑ Å j Å Q S = qNDW = ÅÅ2qε0NDjjV − Vfb − zÑ j ÅÅ q zz{ÑÑÑÑÖ k ÅÇ (31) [H+b ] β= (29) 2q2NS(K aKb)1/2 KTCDL (34) where NS is surface sites per unit area, where CDL is the electrical double layer’s capacitance, Ka is the acid equilibrium constant, and Kb is the base equilibrium constant.113 When qψ β≫ (35) kT where QS, the space charge, is given by (30) then the surface potential can be simplified into q is the elementary charge, ε0 is the permittivity of a vacuum, εS is the semiconductor dielectric constant, ND is the density of an electron donor, V is the electrode potential, Vfb is the flat band potential, W is the width of the space charge layer, k is the Boltzmann’s constant, and T is the temperature in Kelvin.300 Specifically, for a n-type semiconductor electrode at open circuit, there is usually a positive charge and the band edges are upwardly bent by an energy of VSC due to the transfer of ψ = 2.303 kT β (pHpzc − pH) q β+1 (36) On the other hand, the exact eq 33 has to be used when: qψ β≪ (37) kT AN DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Table 26. Summary of the Extracted Electrode Potentials for Representative ZnO and IrOx pH Sensors ref material/structure standard reduction potential (mV) reported reaction Chung et al.22 IrOx +922 IrOx redox reactions (mixed oxidation states) Yang et al.288 IrOx-rGO +676 2[IrO2(OH)2·2H2O]2− + 3H+ + 2e ⇔ [Ir2O3(OH)3·3H2O]3− Batista et al.306 Mani et al.37 ZnO ZnO +636 +746 Fulati et al.36 ZnO nanorods/ nanotubes ZnO +428.6/+627.2 For acid: ZnO + 2OH− + H2O ⇔ Zn(OH)2− 4 For base: ZnO + 2H3O+ + 3H2O ⇔ Zn(H2O)2+ 6 ZnO + H+ ⇔ ZnOH+ +869 mixed potential Ghoneim et al.132 mixed potential 10. FUTURE OUTLOOK ON PH SENSING IN BIOMEDICAL APPLICATIONS Throughout the review, we have discussed available materials for pH and structures for pH sensing (Section 3), common configurations for pH-sensing systems (Section 4), sensing standards and protocols (Section 5), pH regulation in the human body (Section 6) and relevant biomedical applications (Section 7), progress in wearable and implantable pH sensors for biomedical applications (Section 8), and the challenges facing pH-sensing systems and measurements (Section 9). In this section we summarize the learnings from the extensive knowledge base on pH sensing for biomedical applications and provide a future outlook on the field. The strict pH balance maintained by efficient regulation in biological systems makes pH sensing for biomedical applications highly advantageous. Minor deviations in pH can be used as an early indication of malfunction or disease as well as an alarming signal before the rise or propagation of other diseases. As a result, research on pH sensing for biomedical applications has gained sustained interest for decades. For instance, implantable pH sensors (in vivo) have been reported since the 1980s (details can be found in Section 8.2). With advancement in microfabrication technology and nanotechnology manipulation capabilities, new pH-sensing materials, structures, and techniques have emerged. Among the investigated polymeric materials, PANI repeatedly exhibited desired properties with reliable and stable results, especially for wearable pH-sensing applications. IrOx exhibited extraordinary low drift values, Nernstian sensitivity, fast response time, biocompatibility, and repeatable and reproducible results. It is one of the most investigated materials for pH sensors and the fact that multiple reports from various research groups across the globe repeatedly reported its excellent pHsensing properties is a testament to its potential. The reliable performance of IrOx pH sensors can be attributed to the stable reaction at its surface as compared to other materials such as ZnO. Table 26 shows extracted electrode potentials for representative ZnO and IrOx pH sensors. The calculated standard electrode reduction potentials for IrOx are relatively closer to the reduction potential of redox reactions involving iridium ions than the case of calculated standard reduction potentials for ZnO electrodes versus the reduction potential of redox reactions involving zinc ions. This indicates that the reactions in the case of IrOx electrodes are likely dominated by the iridium ion redox reactions. However, in the case of ZnO and for the saturation region, when VDS = VGS − VT, the relationship is defined by 1 K n[(VGS − VT)2 ] 2 mixed potential involving oxygen evolution binding sites and can either overestimate or underestimate the potential, depending on the defect polarity.305 Common defects are shown in Figure 12c. For this site binding model, it is important to note that it falls short when considering the crowding effect, low selectivity, and defects, which is mentioned later in this subsection. or the EGFET configuration, there is a MOSFET, which allows a current to flow. A voltage is generated by the activity of H+ in the solution and the reference voltage. The voltage of the transistor is related to the current (IDS) between the drain and source by the MOSFET expression. The drain-source voltage (VDS) relates to the current linearly before the current saturates. The current saturates when the drain-source voltage reaches the difference between the gate-source voltage (VGS), which is related to the voltage of the RE, and modified threshold voltage (VT). For the linear region, the relationship is defined by the equation: ÉÑ ÄÅ Ñ Å 1 IDS = K nÅÅÅÅ(VGS − VT)VDS − VDS2 ÑÑÑÑ ÑÖ ÅÇ (38) 2 IDS = likely reaction dominant IrOx redox reactions (multiple oxidation states) dominant IrOx redox reactions and other mixed potentials mixed potential involving oxygen evolution mixed potential involving oxygen evolution (39) 114 where Kn is the conduction parameter. Despite the usefulness of these models, there are some limitations and defects that occur at the interface that interfere with the accuracy of the models. One such effect is the crowding effect. Above a critical electrical potential, ion concentration saturates and causes the potential profile for the electrolyte diffusion layer to change, repelling counterions, which in turn reduces the ion concentration at the surface and causes a lower capacitance.302 In this case, the pH is underestimated. However, the opposite can also occur, and there can be a greater pH sensitivity due to the crowding effect of counterions in buffer solutions, causing larger H+ concentrations at the sensor surface. As a result, the models mentioned earlier do not accurately predict pH as they fail to take the crowding effect into account. In addition, the sensor must be selective to H+. Possible cations, such as sodium, could affect the measurement by being accounted for in the model instead of H+.303 Due to a lack of selectivity, the pH predicted by these models can be greatly overestimated because the models account for other cations in the electrolyte solution. Defects are also a factor to consider in these models. There may be point defects, defective with either a vacancy, interstitial impurity, or substitutional impurity at a single atom, or extended defects, defective at multiple atoms or lattice sites,304 that may affect the accuracy of the models. As mentioned previously, the site binding model predicts that the number of binding sites affects the potential. Thus, any defect can affect the number of AO DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review electrodes, the discrepancy indicates a mixed potential likely dominated by other reactions instead of the zinc ion’s redox. This might explain the relative higher stability of IrOx and its wider use as a pH-sensing material. With the expanding investigated materials library for pH sensing and their various nanostructures, the different sensing configurations and testing instruments, and the myriad biomedical applications demonstrated in implantable and wearable pH sensors, standards and protocols become crucial. Furthermore, the complexity of the sensing mechanism modeling and the stability and repeatability challenges require further investigations. Table 27 shows a representative list of ment, in a timely fashion (∼ few minutes). If a similar application is targeting the immediate response to therapeutic drugs, higher sensitivity and resolution would be necessary with faster response times (∼ few seconds). On the other hand, monitoring mouth pH can accommodate fairly larger systems that can detect variations of ∼0.2 pH units in the 5.5−7 pH range, and a response time of a few minutes can be accommodated too. Therefore, as pH-sensing systems evolve toward higher sensitivities and detectability, smaller dimensions, faster responses, and long-term reliability and biocompatibility, new biomedical applications would become possible, helping to diagnose and prevent more diseases. Table 27. Summary of the Recommendations for Standardizing and Benchmarking pH Sensing Materials and Systems using EGFET Configuration.132 AUTHOR INFORMATION category intrinsic characterization system properties intrinsic sensing film properties buffers and solutions properties experimental techniques Corresponding Author *E-mail: canand@media.mit.edu. ORCID description input resistance M. T. Ghoneim: 0000-0002-5568-5284 C. Dagdeviren: 0000-0002-2032-792X intrinsic time constant stability Of all commercial components: (a) characterization instrument (b) commercial transistor (c) glass electrode (d) reference electrode (RE) • composition • crystallinity • thickness • resistivity • compositions • concentrations • conditioning surfaces before measurements • intermittent cleaning between measurements Notes The authors declare no competing financial interest. Biographies Mohamed T. Ghoneim is currently pursuing his postdoctoral training at the Conformable Decoders group at Massachusetts Institute of Technology (MIT), Media Lab. He earned his Ph.D. degree in Electrical Engineering from King Abdullah University of Science and Technology. His research interests include electrochemistry, reliability, microfabrication, flexible electronics, and biomedical applications. Canan Dagdeviren is Assistant Professor of Media Arts and Sciences and LG Career Development Professor of Media Arts and Sciences at MIT, where she leads a research group called Conformable Decoders. Prof. Dagdeviren earned her Ph.D. in Materials Science and Engineering from the University of Illinois at Urbana−Champaign. Her collective research aims to design and fabricate mechanically adaptive electromechanical and electrochemical systems to convert the patterns of nature and the human body into beneficial signals and energy. recommended reporting parameters for the EGFET pH-sensing configuration. For biomedical applications, another important criterion is biocompatibility (acute and long-term tests) of the materials, especially in the case of implantable pH sensors. A complete systematic reporting would help highlight common issues, benchmark materials and reports, and provide an extensive knowledge base for more informed studies. This is critical in the pursuit for empirical solutions to challenges such as modeling, hysteresis, stability, and repeatability. Similar protocol approaches for piezoelectric energy harvester characterization were reported.307,308 Evidently, the integration of various pH sensors with advanced electronics has provided a new platform for the development of novel pH-sensing technologies for disease diagnostics and prevention. Moving forward requires not only expanding available materials and techniques but also addressing the key challenges affecting the reliability and repeatability of pH measurements, reproducibility of sensors, and establishing a convention for standards and protocols for systematic and comprehensive reporting. Finally, sensitivity, accuracy, and precision requirements of pH sensors for biomedical applications vary widely based on the targeted application. For instance, differentiating tumors from healthy cells requires detecting a difference of less than 0.7 pH units in the range of 6.7−7.4, with probes in the micro-/ nanometer diameter range to access the external cell environ- Athena Nguyen is currently an undergraduate at MIT, Department of Biological Engineering. In Summer of 2018, she became an undergraduate researcher under the supervision of Prof. Canan Dagdeviren of the MIT Media Lab. She is interested in nanosensors and other devices for biomedical applications. Naomi Dereje is an undergraduate student at MIT pursuing a bachelor’s degree in mechanical engineering with a minor in statistics and data science. In Summer, 2018, she worked with and Conformable Decoders group at the Media Lab under the supervision of Prof. Canan Dagdeviren. She has also worked in the Biomedical Engineering Department at the Catholic University of America in Washington D.C. Her research interests include product design and development. Jiayao (Kyle) Huang is a current undergraduate student at MIT majoring in aerospace engineering. He previously worked in the Space Systems Lab under Professor David Miller, and Conformable Decoders in MIT’s Media Lab under the supervision of Prof. Canan Dagdeviren. His research interests include materials and structures, and propulsion. Grace C. Moore is currently pursuing an undergraduate degree in materials science and engineering at MIT. In Spring 2018, she worked with Prof. Canan Dagdeviren at the MIT Media Lab as an undergraduate researcher. She has also performed undergraduate research in convection batteries with Prof. Fikile Brushett. Her research AP DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review (20) Kurzweil, P. Metal Oxides and Ion-Exchanging Surfaces as pH Sensors in Liquids: State- of-the-Art and Outlook. Sensors 2009, 9, 4955−4985. (21) Wei, A.; Pan, L.; Huang, W. Recent Progress in the ZnO Nanostructure-Based Sensors. Mater. Sci. Eng., B 2011, 176, 1409− 1421. (22) Chung, H. J.; Sulkin, M. S.; Kim, J. S.; Goudeseune, C.; Chao, H. Y.; Song, J. W.; Yang, S. Y.; Hsu, Y. Y.; Ghaffari, R.; Efimov, I. R.; et al. Stretchable, Multiplexed pH Sensors with Demonstrations on Rabbit and Human Hearts Undergoing Ischemia. Adv. Healthcare Mater. 2014, 3, 59−68. (23) Yuqing, M.; Jianrong, C.; Keming, F. New Technology for the Detection of pH. J. Biochem. Biophys. Methods 2005, 63, 1−9. (24) Głab, S.; Hulanicki, A.; Edwall, G.; Ingman, F. Metal-Metal Oxide and Metal Oxide Electrodes as pH Sensors. Crit. Rev. Anal. Chem. 1989, 21, 29−47. (25) Santos, L.; Neto, J. P.; Crespo, A.; Nunes, D.; Costa, N.; Fonseca, I. M.; Barquinha, P.; Pereira, L.; Silva, J.; Martins, R.; et al. WO3 Nanoparticle-Based Conformable pH Sensor. ACS Appl. Mater. Interfaces 2014, 6, 12226−12234. (26) Zaman, S.; Asif, M. H.; Zainelabdin, A.; Amin, G.; Nur, O.; Willander, M. CuO Nanoflowers as an Electrochemical pH Sensor and the Effect of pH on the Growth. J. Electroanal. Chem. 2011, 662, 421− 425. (27) Manjakkal, L.; Cvejin, K.; Kulawik, J.; Zaraska, K.; Szwagierczak, D.; Socha, R. P. Fabrication of Thick Film Sensitive RuO2-TiO2 and Ag/AgCl/KCl Reference Electrodes and Their Application for pH Measurements. Sens. Actuators, B 2014, 204, 57−67. (28) Chen, M.; Jin, Y.; Qu, X.; Jin, Q.; Zhao, J. Electrochemical Impedance Spectroscopy Study of Ta2O5 Based EIOS pH Sensors in Acid Environment. Sens. Actuators, B 2014, 192, 399−405. (29) Lee, I.-S.; Whang, C.-N.; Lee, Y.-H.; Hwan Lee, G.; Park, B.-J.; Park, J.-C.; Seo, W.-S.; Cui, F.-Z. Formation of Nano Iridium Oxide: Material Properties and Neural Cell Culture. Thin Solid Films 2005, 475, 332−336. (30) Ges, I. A.; Ivanov, B. L.; Schaffer, D. K.; Lima, E. A.; Werdich, A. A.; Baudenbacher, F. J. Thin-Film IrOx pH Microelectrode for Microfluidic-Based Microsystems. Biosens. Biosens. Bioelectron. 2005, 21, 248−256. (31) Yamanaka, K. Anodically Electrodeposited Iridium Oxide Films (AEIROF’s) from Alkaline Solutions for Electrochromic Display Devices. Jpn. J. Appl. Phys. 1989, 28, 632−637. (32) Bezbaruah, A. N.; Zhang, T. C. Fabrication of Anodically Electrodeposited Iridium Oxide Film pH Microelectrodes for Microenvironmental Studies. Anal. Chem. 2002, 74, 5726−5733. (33) Spǎtaru, T.; Roman, E.; Spǎtaru, N. Electrodeposition of Cobalt Oxide on Conductive Diamond Electrodes for Catalytic Sensor Applications. Rev. Roum. Chim. 2004, 49, 525−530. (34) Lee, W. H.; Lee, J.-H.; Choi, W.-H.; Hosni, A. A.; Papautsky, I.; Bishop, P. L. Needle- Type Environmental Microsensors: Design, Construction and Uses of Microelectrodes and Multi-Analyte MEMS Sensor Arrays. Meas. Sci. Technol. 2011, 22, 042001. (35) Al-Hilli, S. M.; Al-Mofarji, R. T.; Klason, P.; Willander, M.; Gutman, N.; Sa’Ar, A. Zinc Oxide Nanorods Grown on TwoDimensional Macroporous Periodic Structures and Plane Si as a pH Sensor. J. Appl. Phys. 2008, 103, 014302. (36) Fulati, A.; Usman Ali, S. M.; Riaz, M.; Amin, G.; Nur, O.; Willander, M. Miniaturized pH Sensors Based on Zinc Oxide Nanotubes/Nanorods. Sensors 2009, 9, 8911−8923. (37) Mani, G. K.; Morohoshi, M.; Yasoda, Y.; Yokoyama, S.; Kimura, H.; Tsuchiya, K. ZnO- Based Microfluidic pH Sensor: A Versatile Approach for Quick Recognition of Circulating Tumor Cells in Blood. ACS Appl. Mater. Interfaces 2017, 9, 5193−5203. (38) Zhang, Q.; Liu, W.; Sun, C.; Zhang, H.; Pang, W.; Zhang, D.; Duan, X. On-Chip Surface Modified Nanostructured ZnO as Functional pH Sensors. Nanotechnology 2015, 26, 355202. (39) Pachauri, V.; Vlandas, A.; Kern, K.; Balasubramanian, K. SiteSpecific Self-Assembled Liquid-Gated ZnO Nanowire Transistors for Sensing Applications. Small 2010, 6, 589−594. interests include electrochemistry, convection chemistry, electrochemical energy storage, and biomedical applications. Philip J. Murzynowski is a current undergraduate student at MIT majoring in computer science and electrical engineering. He previously worked in the Space Systems Lab under Prof. David Miller, and Conformable Decoders in MIT’s Media Lab under the supervision of Prof. Canan Dagdeviren. His research interests include cyber security, signal processing, and computer architecture. ACKNOWLEDGMENTS The authors thank David Sadat for assisting with the manuscript preparation and the MIT Media Lab for financial support. REFERENCES (1) Lesney, M. S. A Basic History of Acid- From Aristotle to Arnold. Am. Chem. Soc. 2003, 12, 47−50. (2) Khan, M. I.; Mukherjee, K.; Shoukat, R.; Dong, H. A Review on pH Sensitive Materials for Sensors and Detection Methods. Microsyst. Technol. 2017, 23, 4391−4404. (3) Myers, R. J. One-Hundred Years of pH. J. Chem. Educ. 2010, 87, 30−32. (4) Salis, A.; Monduzzi, M. Not Only pH. Specific Buffer Effects in Biological Systems. Curr. Opin. Colloid Interface Sci. 2016, 23, 1−9. (5) Salek, S. S.; Van Turnhout, A. G.; Kleerebezem, R.; Van Loosdrecht, M. C. M. pH Control in Biological Systems Using Calcium Carbonate. Biotechnol. Bioeng. 2015, 112, 905−913. (6) Tannock, I. F.; Rotin, D. Acid pH in Tumors and Its Potential for Therapeutic Exploitation. Cancer Res. 1989, 49, 4373−4384. (7) Kilpatrick, M.; Kilpatrick, M. The Teaching of the Theory of the Dissociation of Electrolytes. II. The Definition of pH. J. Chem. Educ. 1932, 9, 1010−1016. (8) MacInnes, D. A. Criticism of a Definition of pH. Science 1948, 108, 693−693. (9) Covington, A. K.; Bates, R. G.; Durst, R. A. Definition of pH Scales, Standard Reference Values, Measurement of pH and Related Terminology. Pure Appl. Chem. 1985, 57, 531−542. (10) Barron, J. J.; Ashton, C.; Geary, L. The Effects of Temperature on pH Measurement. Tsp 2011, 1, 1−7. (11) Chiang, J. L.; Jan, S. S.; Chou, J. C.; Chen, Y. C. Study on the Temperature Effect, Hysteresis and Drift of PH-ISFET Devices On the basis of Amorphous Tungsten Oxide. Sens. Actuators, B 2001, 76, 624− 628. (12) Rosenthal, T. B. The Effect of Temperature on the pH of Blood and Plasma in Vitro. J. Biol. Chem. 1948, 173, 25−30. (13) Huang, W.-D.; Cao, H.; Deb, S.; Chiao, M.; Chiao, J. C. A Flexible pH Sensor Based on the Iridium Oxide Sensing Film. Sens. Actuators, A 2011, 169, 1−11. (14) Zhang, W.; Abou El-Reash, Y. G.; Ding, L.; Lin, Z.; Lian, Y.; Song, B.; Yuan, J.; Wang, X. A Lysosome-Targeting Nanosensor for Simultaneous Fluorometric Imaging of Intracellular pH Values and Temperature. Microchim. Acta 2018, 185, 533. (15) Pietsch, C.; Hoogenboom, R.; Schubert, U. S. Soluble Polymeric Dual Sensor for Temperature and pH Value. Angew. Chem., Int. Ed. 2009, 48, 5653−5656. (16) Korostynska, O.; Arshak, K.; Gill, E.; Arshak, A. Review on Stateof-the-Art in Polymer Based pH Sensors. Sensors 2007, 7, 3027−3042. (17) Ahmed, N. M.; Kabaa, E. A.; Jaafar, M. S.; Omar, A. F. Characteristics of Extended-Gate Field-Effect Transistor (EGFET) Based on Porous n-Type (111) Silicon for Use in pH Sensors. J. Electron. Mater. 2017, 46, 5804−5813. (18) Lee, C. S.; Kyu Kim, S.; Kim, M. Ion-Sensitive Field-Effect Transistor for Biological Sensing. Sensors 2009, 9, 7111−7131. (19) Rajapaksha, R. D. A. A.; Hashim, U.; Fernando, C. A. N. Design, Fabrication and Characterization of 1.0 Μm Gap Al Based Interdigitated Electrode for Biosensors. Microsyst. Technol. 2017, 23, 4501−4507. AQ DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Polyaniline Nanopillar Array Electrode. J. Colloid Interface Sci. 2017, 490, 53−58. (61) Rojas, J. P.; Torres Sevilla, G. A.; Ghoneim, M. T.; Inayat, S. B.; Ahmed, S. M.; Hussain, A. M.; Hussain, M. M. Transformational Silicon Electronics. ACS Nano 2014, 8, 1468−1474. (62) Ghoneim, M. T.; Hussain, M. M. Study of Harsh Environment Operation of Flexible Ferroelectric Memory Integrated with PZT and Silicon Fabric. Appl. Phys. Lett. 2015, 107, 052904. (63) Ghoneim, M. T.; Rojas, J. P.; Young, C. D.; Bersuker, G.; Hussain, M. M. Electrical Analysis of High Dielectric Constant Insulator and Metal Gate Metal Oxide Semiconductor Capacitors on Flexible Bulk Mono-Crystalline Silicon. IEEE Trans. Reliab. 2015, 64, 579−585. (64) Ghoneim, M. T.; Kutbee, A.; Ghodsi Nasseri, F.; Bersuker, G.; Hussain, M. M. Mechanical Anomaly Impact on Metal-OxideSemiconductor Capacitors on Flexible Silicon Fabric. Appl. Phys. Lett. 2014, 104, 234104. (65) Torres Sevilla, G. A.; Ghoneim, M. T.; Fahad, H.; Rojas, J. P.; Hussain, A. M.; Hussain, M. M. Flexible Nanoscale High-Performance FinFETs. ACS Nano 2014, 8, 9850−9856. (66) Ghoneim, M. T.; Zidan, M. A.; Salama, K. N.; Hussain, M. M. Towards Neuromorphic Electronics: Memristors on Foldable Silicon Fabric. Microelectron. J. 2014, 45, 1392−1395. (67) Ghoneim, M. T.; Rojas, J. P.; Hussain, A. M.; Hussain, M. M. Additive Advantage in Characteristics of MIMCAPs on Flexible Silicon (100) Fabric with Release-First Process. Phys. Status Solidi RRL 2014, 8, 163−166. (68) Nyein, H. Y. Y.; Gao, W.; Shahpar, Z.; Emaminejad, S.; Challa, S.; Chen, K.; Fahad, H. M.; Tai, L. C.; Ota, H.; Davis, R. W.; et al. A Wearable Electrochemical Platform for Noninvasive Simultaneous Monitoring of Ca2+ and pH. ACS Nano 2016, 10, 7216−7224. (69) Humpolicek, P.; Kasparkova, V.; Saha, P.; Stejskal, J. Biocompatibility of Polyaniline. Synth. Met. 2012, 162, 722−727. (70) Sha, R.; Komori, K.; Badhulika, S. Amperometric pH Sensor Based on Graphene- Polyaniline Composite. IEEE Sens. J. 2017, 17, 5038−5043. (71) Ping, J.; Wang, Y.; Wu, J.; Ying, Y.; Ji, F. A Novel pH Sensing Membrane Based on an Ionic Liquid-Polymer Composite. Microchim. Acta 2012, 176, 229−234. (72) Maiolo, L.; Mirabella, S.; Maita, F.; Alberti, A.; Minotti, A.; Strano, V.; Pecora, A.; Shacham-Diamand, Y.; Fortunato, G. Flexible pH Sensors Based on Polysilicon Thin Film Transistors and ZnO Nanowalls. Appl. Phys. Lett. 2014, 105, 093501. (73) Shin, J.; Braun, P. V.; Lee, W. Fast Response Photonic Crystal pH Sensor Based on Templated Photo-Polymerized Hydrogel Inverse Opal. Sens. Actuators, B 2010, 150, 183−190. (74) Richter, A.; Paschew, G.; Klatt, S.; Lienig, J.; Arndt, K. F.; Adler, H. J. P. Review on Hydrogel-Based pH Sensors and Microsensors. Sensors 2008, 8, 561−581. (75) Richter, A.; Bund, A.; Keller, M.; Arndt, K. F. Characterization of a Microgravimetric Sensor Based on pH Sensitive Hydrogels. Sens. Actuators, B 2004, 99, 579−585. (76) Zong, S.; Wang, Z.; Yang, J.; Cui, Y. Intracellular pH Sensing Using P-Aminothiophenol Functionalized Gold Nanorods with Low Cytotoxicity. Anal. Chem. 2011, 83, 4178−4183. (77) Ma, D.; Zheng, J.; Tang, P.; Xu, W.; Qing, Z.; Yang, S.; Li, J.; Yang, R. Quantitative Monitoring of Hypoxia-Induced Intracellular Acidification in Lung Tumor Cells and Tissues Using Activatable Surface-Enhanced Raman Scattering Nanoprobes. Anal. Chem. 2016, 88, 11852−11859. (78) Qin, Y.; Kwon, H. J.; Subrahmanyam, A.; Howlader, M. M. R.; Selvaganapathy, P. R.; Adronov, A.; Deen, M. J. Inkjet-Printed Bifunctional Carbon Nanotubes for pH Sensing. Mater. Lett. 2016, 176, 68−70. (79) Ruan, C.; Zeng, K.; Grimes, C. A. A Mass-Sensitive pH Sensor Based on a Stimuli- Responsive Polymer. Anal. Chim. Acta 2003, 497, 123−131. (80) Lin, J. Recent Development and Applications of Optical and Fiber-Optic pH Sensors. TrAC, Trends Anal. Chem. 2000, 19, 541−552. (40) Park, J.; Lee, S.; Lee, J.; Yong, K. A Light Incident Angle Switchable ZnO Nanorod Memristor: Reversible Switching Behavior Between Two Non-Volatile Memory Devices. Adv. Mater. 2013, 25, 6423−6429. (41) Khaderbad, M. A.; Choi, Y.; Hiralal, P.; Aziz, A.; Wang, N.; Durkan, C.; Thiruvenkatanathan, P.; Amaratunga, G. A. J.; Rao, V. R.; Seshia, A. A. Electrical Actuation and Readout in a Nanoelectromechanical Resonator Based on a Laterally Suspended Zinc Oxide Nanowire. Nanotechnology 2012, 23, 025501. (42) Wen, W.; Wu, J.-M.; Wang, Y.-D. Large-Size Porous ZnO Flakes with Superior Gas- Sensing Performance. Appl. Phys. Lett. 2012, 100, 262111. (43) Gao, P. X.; Ding, Y.; Wang, Z. L. Crystallographic OrientationAligned ZnO Nanorods Grown by a Tin Catalyst. Nano Lett. 2003, 3, 1315−1320. (44) Pan, Z. W.; Dai, Z. R.; Wang, Z. L.; Hughes, W. L.; Lao, C.; Wang, Z. L. Nanobelts of Semiconducting Oxides. Science 2001, 291, 1947− 1949. (45) Hughes, W. L.; Wang, Z. L. Formation of Piezoelectric SingleCrystal Nanorings and Nanobows. J. Am. Chem. Soc. 2004, 126, 6703− 6709. (46) Gao, P. X.; Ding, Y.; Mai, W.; Hughes, W. L.; Lao, C.; Wang, Z. L. Conversion of Zinc Oxide Nanobelts into Superlattice-Structured Nanohelices. Science 2005, 309, 1700−1704. (47) Chang, Y.-C.; Yang, W.-C.; Chang, C.-M.; Hsu, P.-C.; Chen, L.-J. Controlled Growth of ZnO Nanopagoda Arrays with Varied Lamination and Apex Angles. Cryst. Growth Des. 2009, 9, 3161−3167. (48) Cho, S.; Jung, S.-H.; Jang, J.-W.; Oh, E.; Lee, K.-H. Simultaneous Synthesis of Al-Doped ZnO Nanoneedles and Zinc Aluminum Hydroxides through Use of a Seed Layer. Cryst. Growth Des. 2008, 8, 4553−4558. (49) Kong, X. Y.; Wang, Z. L. Spontaneous Polarization-Induced Nanohelixes, Nanosprings, and Nanorings of Piezoelectric Nanobelts. Nano Lett. 2003, 3, 1625−1631. (50) Mazeina, L.; Picard, Y. N.; Prokes, S. M. Controlled Growth of Parallel Oriented ZnO Nanostructural Arrays on Ga2O3 Nanowires. Cryst. Growth Des. 2009, 9, 1164−1169. (51) Wei, A.; Sun, X. W.; Xu, C. X.; Dong, Z. L.; Yang, Y.; Tan, S. T.; Huang, W. Growth Mechanism of Tubular ZnO Formed in Aqueous Solution. Nanotechnology 2006, 17, 1740−1744. (52) Zhou, H. L.; Shao, P. G.; Chua, S. J.; van Kan, J. A.; Bettiol, A. A.; Osipowicz, T.; Ooi, K. F.; Goh, G. K. L.; Watt, F. Selective Growth of ZnO Nanorod Arrays on a GaN/Sapphire Substrate Using a Proton Beam Written Mask. Cryst. Growth Des. 2008, 8, 4445−4448. (53) Ng, A. M. C.; Chen, X. Y.; Djurisic, A. B. ZnO Nanostructures for Optoelectronics : Material Properties and Device Applications. Prog. Quantum Electron. 2010, 34, 191−259. (54) Avrutin, V.; Silversmith, D. J.; Morkoç, H. Doping Asymmetry Problem in ZnO: Current Status and Outlook. Proc. IEEE 2010, 98, 1269−1280. (55) Aydin, C.; Abd El-Sadek, M. S.; Zheng, K.; Yahia, I. S.; Yakuphanoglu, F. Synthesis, Diffused Reflectance and Electrical Properties of Nanocrystalline Fe-Doped ZnO via Sol- Gel Calcination Technique. Opt. Laser Technol. 2013, 48, 447−452. (56) George, S.; Pokhrel, S.; Xia, T.; Gilbert, B.; Ji, Z.; Schowalter, M.; Rosenauer, A.; Damoiseaux, R.; Bradley, K. A.; Mädler, L.; et al. Use of a Rapid Cytotoxicity Screening Approach To Engineer a Safer Zinc Oxide Nanoparticle through Iron Doping. ACS Nano 2010, 4, 15−29. (57) Wang, J. L.; Yang, P. Y.; Hsieh, T. Y.; Juan, P. C. Ionic pH and Glucose Sensors Fabricated Using Hydrothermal ZnO Nanostructures. Jpn. J. Appl. Phys. 2016, 55, 01AE16. (58) Wang, J.-L.; Yang, P.-Y.; Hsieh, T.-Y.; Hwang, C.-C.; Juang, M.H. PH-Sensing Characteristics of Hydrothermal Al-Doped ZnO Nanostructures. J. Nanomater. 2013, 2013, 1−7. (59) Guinovart, T.; Valdés-Ramírez, G.; Windmiller, J. R.; Andrade, F. J.; Wang, J. Bandage- Based Wearable Potentiometric Sensor for Monitoring Wound pH. Electroanalysis 2014, 26, 1345−1353. (60) Yoon, J. H.; Hong, S. B.; Yun, S. O.; Lee, S. J.; Lee, T. J.; Lee, K. G.; Choi, B. G. High Performance Flexible pH Sensor Based on AR DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review (81) Li, L.-L.; Sun, H.; Fang, C.-J.; Xu, J.; Jin, J.-Y.; Yan, C.-H. Optical Sensors Based on Functionalized Mesoporous Silica SBA-15 for the Detection of Multianalytes (H+ and Cu2+) in Water. J. Mater. Chem. 2007, 17, 4492. (82) Islam, S.; Bakhtiar, H.; Bidin, N.; Riaz, S.; Naseem, S. Sol−gel Based Thermally Stable Mesoporous TiO 2 Nanomatrix for Fiber Optic pH Sensing. J. Sol-Gel Sci. Technol. 2018, 86, 42−50. (83) Islam, S.; Bakhtiar, H.; Naseem, S.; Aziz, M. S. B. A.; Bidin, N.; Riaz, S.; Ali, J. Surface Functionality and Optical Properties Impact of Phenol Red Dye on Mesoporous Silica Matrix for Fiber Optic pH Sensing. Sens. Actuators, A 2018, 276, 267−277. (84) Tsou, C.-J.; Chu, C.; Hung, Y.; Mou, C.-Y. A Broad Range Fluorescent pH Sensor Based on Hollow Mesoporous Silica Nanoparticles, Utilising the Surface Curvature Effect. J. Mater. Chem. B 2013, 1, 5557−5563. (85) Peng, H.-S.; Chiu, D. T. Soft Fluorescent Nanomaterials for Biological and Biomedical Imaging. Chem. Soc. Rev. 2015, 44, 4699− 4722. (86) Pospíšilová, M.; Kuncová, G.; Trögl, J. Fiber-Optic Chemical Sensors and Fiber-Optic Bio-Sensors. Sensors 2015, 15, 25208−25259. (87) Wolfbeis, O. S. Fiber-Optic Chemical Sensors and Biosensors. Anal. Chem. 2008, 80, 4269−4283. (88) Capel-Cuevas, S.; Cuéllar, M. P.; de Orbe-Payá, I.; Pegalajar, M. C.; Capitán-Vallvey, L. F. Full-Range Optical pH Sensor Based on Imaging Techniques. Anal. Chim. Acta 2010, 681, 71−81. (89) Fritzsche, M.; Barreiro, C. G.; Hitzmann, B.; Scheper, T. Optical pH Sensing Using Spectral Analysis. Sens. Actuators, B 2007, 128, 133− 137. (90) Sheppard, N. F.; Lesho, M. J.; McNally, P.; Shaun Francomacaro, A. Microfabricated Conductimetric pH Sensor. Sens. Actuators, B 1995, 28, 95−102. (91) Benson, K.; Ghimire, A.; Pattammattel, A.; Kumar, C. V. Protein Biophosphors: Biodegradable, Multifunctional, Protein-Based Hydrogel for White Emission, Sensing, and pH Detection. Adv. Funct. Mater. 2017, 27, 1702955. (92) Richter, A.; Howitz, S.; Kuckling, D.; Arndt, K. F. Influence of Volume Phase Transition Phenomena on the Behavior of HydrogelBased Valves. Sens. Actuators, B 2004, 99, 451−458. (93) Zhao, Y.; Lei, M.; Liu, S. X.; Zhao, Q. Smart Hydrogel-Based Optical Fiber SPR Sensor for pH Measurements. Sens. Actuators, B 2018, 261, 226−232. (94) You, Y. H.; Nagaraja, A. T.; Biswas, A.; Hwang, J. H.; Cote, G. L.; McShane, M. J. SERS-Active Smart Hydrogels with Modular Microdomains: From pH to Glucose Sensing. IEEE Sens. J. 2017, 17, 941−950. (95) Mishra, S. K.; Zou, B.; Chiang, K. S. Wide-Range pH Sensor Based on a Smart-Hydrogel- Coated Long-Period Fiber Grating. IEEE J. Sel. Top. Quantum Electron. 2017, 23, 284. (96) Shaibani, P. M.; Etayash, H.; Naicker, S.; Kaur, K.; Thundat, T. Metabolic Study of Cancer Cells Using a pH Sensitive Hydrogel Nanofiber Light Addressable Potentiometric Sensor. ACS Sensors 2017, 2, 151−156. (97) Ghoneim, M. T.; Smith, C. E.; Hussain, M. M. Simplistic Graphene Transfer Process and Its Impact on Contact Resistance. Appl. Phys. Lett. 2013, 102, 183115. (98) Chen, S. X.; Chang, S. P.; Chang, S. J. Investigation of InN Nanorod-Based EGFET pH Sensors Fabricated on Quartz Substrate. Dig. J. Nanomater. Biostructures 2014, 9, 1505−1511. (99) Lee, C. T.; Chiu, Y. S. Photoelectrochemical Passivated ZnOBased Nanorod Structured Glucose Biosensors Using Gate-Recessed AlGaN/GaN Ion-Sensitive Field-Effect- Transistors. Sens. Actuators, B 2015, 210, 756−761. (100) Rout, C. S.; Kulkarni, G. U.; Rao, C. N. R. Electrical and Hydrogen-Sensing Characteristics of Field Effect Transistors Based on Nanorods of ZnO and WO2.27. J. Nanosci. Nanotechnol. 2009, 9, 5652. (101) Al-Hilli, S. M.; Al-Mofarji, R. T.; Willander, M. Zinc Oxide Nanorod for Intracellular pH Sensing. Appl. Phys. Lett. 2006, 89, 173119. (102) Ahmad, Y.; Sharma, N. K.; Ahmad, M. F.; Sharma, M.; Garg, I.; Srivastava, M.; Bhargava, K. The Proteome of Hypobaric Induced Hypoxic Lung: Insights from Temporal Proteomic Profiling for Biomarker Discovery. Sci. Rep. 2015, 5, 1−20. (103) Tran, T.; Kwon, J.; Lee, K.; Lee, J.; Ju, B. pH Sensor Using Carbon Nanotubes as Sensing Material. In IEEE ICCE’06. First International Conference on Communications and Electronics; IEEE: Hanoi, Vietnam, 2007; pp 1−4. (104) Li, C. A.; Han, K. N.; Pham, X. H.; Seong, G. H. A Single-Walled Carbon Nanotube Thin Film-Based PH-Sensing Microfluidic Chip. Analyst 2014, 139, 2011−2015. (105) Gou, P.; Kraut, N. D.; Feigel, I. M.; Bai, H.; Morgan, G. J.; Chen, Y.; Tang, Y.; Bocan, K.; Stachel, J.; Berger, L.; et al. Carbon Nanotube Chemiresistor for Wireless pH Sensing. Sci. Rep. 2015, 4, 1−6. (106) Jung, D.; Han, M. E.; Lee, G. S. PH-Sensing Characteristics of Multi-Walled Carbon Nanotube Sheet. Mater. Lett. 2014, 116, 57−60. (107) Tsai, W.; Huang, B.; Wang, K.; Huang, Y.; Yang, P.; Cheng, H. Functionalized Carbon Nanotube Thin Films as the pH Sensing Membranes of Extended-Gate Field-Effect Transistors on the Flexible Substrates. IEEE Trans. Nanotechnol. 2014, 13, 760−766. (108) Jiao, L. H.; Barakat, N. Ion-Sensitive Field Effect Transistor as a pH Sensor Ion-Sensitive Field Effect Transistor as a pH Sensor. J. Nanosci. Nanotechnol. 2013, 13, 1194−1198. (109) Yuqing, M.; Jianguo, G.; Jianrong, C. Ion Sensitive Field Effect Transducer-Based Biosensors. Biotechnol. Adv. 2003, 21, 527−534. (110) van der Spiegel, J.; Lauks, I.; Chan, P.; Babic, D. The Extended Gate Chemically Sensitive Field Effect Transistor as Multi-Species Microprobe. Sens. Actuators 1983, 4, 291−298. (111) Chi, L. L.; Chou, J. C.; Chung, W. Y.; Sun, T. P.; Hsiung, S. K. Study on Extended Gate Field Effect Transistor with Tin Oxide Sensing Membrane. Mater. Chem. Phys. 2000, 63, 19−23. (112) Hussin, M. R. M.; Ismail, R.; Syono, I. Simulation and Fabrication of Extended Gate Ion Sensitive Field Effect Transistor for Biosensor Application. In Computer Applications for Security, Control and System Engineering; Springer: Heidelberg, 2012; pp 396−403. (113) Chiu, Y. S.; Tseng, C. Y.; Lee, C. T. Nanostructured EGFET pH Sensors with Surface- Passivated ZnO Thin-Film and Nanorod Array. IEEE Sens. J. 2012, 12, 930−934. (114) Batista, P. D.; Mulato, M. ZnO Extended-Gate Field-Effect Transistors as pH Sensors. Appl. Phys. Lett. 2005, 87, 143508. (115) Yang, P.-Y.; Wang, J.-L.; Chiu, P.-C.; Chou, J.-C.; Chen, C.-W.; Li, H.-H.; Cheng, H.-C. pH Sensing Characteristics of Extended-Gate Field-Effect Transistor Based on Al-Doped ZnO Nanostructures Hydrothermally Synthesized at Low Temperatures. IEEE Electron Device Lett. 2011, 32, 1603−1605. (116) Li, H. H.; Yang, C. E.; Kei, C. C.; Su, C. Y.; Dai, W. S.; Tseng, J. K.; Yang, P. Y.; Chou, J. C.; Cheng, H. C. Coaxial-Structured ZnO/ Silicon Nanowires Extended-Gate Field- Effect Transistor as pH Sensor. Thin Solid Films 2013, 529, 173−176. (117) Manjakkal, L.; Sakthivel, B.; Gopalakrishnan, N.; Dahiya, R. Printed Flexible Electrochemical pH Sensors Based on CuO Nanorods. Sens. Actuators, B 2018, 263, 50−58. (118) Li, P.; Martin, C. M.; Yeung, K. K.; Xue, W. Dielectrophoresis Aligned Single-Walled Carbon Nanotubes as Ph Ensors. Biosensors 2011, 1, 23−35. (119) Yang, C. F.; Chen, C. L.; Busnaina, A.; Dokmeci, M. R. SingleWalled Carbon Nanotube Based pH Sensors on a Flexible Parylene-C Substrate. In Proceedings of the 31st Annual International Conference of the IEEE Engineering in Medicine and Biology Society: Engineering the Future of Biomedicine, EMBC 2009; IEEE: Minneapolis, MN, USA, 2009; pp 4102−4105. (120) Nordin, N.; Hashim, U.; Azizah, N. Interdigitated Electrode (IDE) Based on Titanium Dioxide (TiO2) Thin Films for Biosensor Application. In AIP Conference Proceedings; AIP Publishing: Selangor, Malaysia, 2016; Vol. 1733, p 020084. (121) Ali, G. M. Interdigitated Extended Gate Field Effect Transistor Without Reference Electrode. J. Electron. Mater. 2017, 46, 713−717. AS DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review (122) Lakard, B.; Segut, O.; Lakard, S.; Herlem, G.; Gharbi, T. Potentiometric Miniaturized pH Sensors Based on Polypyrrole Films. Sens. Actuators, B 2007, 122, 101−108. (123) Mazlan, N. S.; Ramli, M. M.; Abdullah, M. M. A. B.; Halin, D. S. C.; Isa, S. S. M.; Talip, L. F. A.; Danial, N. S.; Murad, S. A. Z. Interdigitated Electrodes as Impedance and Capacitance Biosensors: A Review. In AIP Conference Proceedings; AIP Publishing: Krabi, Thailand, 2017; Vol. 1885, p 020276. (124) Tsouti, V.; Boutopoulos, C.; Zergioti, I.; Chatzandroulis, S. Capacitive Microsystems for Biological Sensing. Biosens. Bioelectron. 2011, 27, 1−11. (125) Farehanim, M. A.; Hashim, U.; Azizah, N.; Fatin, M. F.; Azman, A. H. Fabrication of Interdigitated Electrodes (IDEs) Using Basic Conventional Lithography for pH Measurement. In AIP Conference Proceedings; AIP Publishing: Penang, Malaysia, 2017; Vol. 1808. (126) Haarindraprasad, R.; Hashim, U.; Gopinath, S. C. B.; Kashif, M.; Veeradasan, P.; Balakrishnan, S. R.; Foo, K. L.; Poopalan, P. Low Temperature Annealed Zinc Oxide Nanostructured Thin Film-Based Transducers: Characterization for Sensing Applications. PLoS One 2015, 10, e0132755. (127) Chih, A.; Yang, F.; Chen, C. L.; Dokmeci, M. R. Single-Walled Carbon Nanotube Based pH Sensor. Carbon N. Y. 2009, 124, 12419. (128) Nguyen, H. D.; Nguyen, T. H.; Hoang, N. V.; Le, N. N.; Nguyen, T. N. N.; Doan, D. C. T.; Dang, M. C. pH Sensitivity of Emeraldine Salt Polyaniline and Poly(Vinyl Butyral) Blend. Adv. Nat. Sci.: Nanosci. Nanotechnol. 2014, 5, 045001. (129) Chinnathambi, S.; Euverink, G. J. W. Polyaniline Functionalized Electrochemically Reduced Graphene Oxide Chemiresistive Sensor to Monitor the pH in Real Time during Microbial Fermentations. Sens. Actuators, B 2018, 264, 38−44. (130) Zhang, Y.; Nayak, T. R.; Hong, H.; Cai, W. Biomedical Applications of Zinc Oxide Nanomaterials. Curr. Mol. Med. 2013, 13, 1633−1645. (131) Copa, V.; Tuico, A.; Mendoza, J.; Ferrolino, J.; Vergara, C.; Salvador, A.; Estacio, E.; Somintac, A. Development of ResistanceBased pH Sensor Using Zinc Oxide Nanorods. J. Nanosci. Nanotechnol. 2016, 16, 6102−6106. (132) Ghoneim, M. T.; Sadraei, A.; de Souza, P.; Moore, G. C.; Bazant, M. Z.; Dagdeviren, C. A Protocol to Characterize pH Sensing Materials and Systems. Small Methods 2019, 3, 1800265. (133) Kakooei, S.; Ismail, C.; Ari-Wahjoedi, B. An Overview of pH Sensors Based on Iridium Oxide: Fabrication and Application. Int. J. Mater. Sci. Innov. 2013, 1, 62−72. (134) Chen, P. Y.; Yin, L. Te; Cho, T. H. Optical and Impedance Characteristics of EGFET Based on SnO2/ITO Sensing Gate. Life Sci. J. 2014, 11, 871−875. (135) Oh, H.; Lee, K. J.; Baek, J.; Yang, S. S.; Lee, K. Development of a High Sensitive pH Sensor Based on Shear Horizontal Surface Acoustic Wave with ZnO Nanoparticles. Microelectron. Eng. 2013, 111, 154− 159. (136) Huang, B.-R.; Lin, J.-C.; Yang, Y.-K. ZnO/Silicon Nanowire Hybrids Extended-Gate Field-Effect Transistors as pH Sensors. J. Electrochem. Soc. 2013, 160, B78−B82. (137) Chen, P.-Y.; Yin, L.-T.; Shi, M.-D.; Lee, Y.-C. Drift and Light Characteristics of EGFET Based on SnO2/ITO Sensing Gate. Life Sci. 2013, 10, 3132−3136. (138) Catts, J. G.; Langmuir, D. Adsorption of Cu, Pb and Zn by ΔMnO2: Applicability of the Site Binding-Surface Complexation Model. Appl. Geochem. 1986, 1, 255−264. (139) Keysight Technology. Keysight Technologies, Digital Multimeters. Datasheet 2015, 1−22. (140) Shee, D. B1500A Semiconductor Device Analyzer. (141) Keysight Technologies. Keysight B1500A Semiconductor Device Analyzer - Data Sheet, www.keysight.com/find/b1500a. (142) Mettler Toledo. InLab Sensors, 2017. (143) Lue, C. E.; Wang, I. S.; Huang, C. H.; Shiao, Y. T.; Wang, H. C.; Yang, C. M.; Hsu, S. H.; Chang, C. Y.; Wang, W.; Lai, C. S. pH Sensing Reliability of Flexible ITO/PET Electrodes on EGFETs Prepared by a Roll-to-Roll Process. Microelectron. Reliab. 2012, 52, 1651−1654. (144) Lin, J. L.; Chu, Y. M.; Hsaio, S. H.; Chin, Y. L.; Sun, T. P. Structures of Anodized Aluminum Oxide Extended-Gate Field-Effect Transistors on pH Sensors. Japanese J. Appl. Physics, Part 1 Regul. Pap. Short Notes Rev. Pap. 2006, 45, 7999−8004. (145) Queeney, K. M. Automated pH Cleaning and Calibration Systems - A Modular Approach. Technical Papers of ISA 2004, 749− 758. (146) Schöning, M. J.; Brinkmann, D.; Rolka, D.; Demuth, C.; Poghossian, A. CIP (Cleaning- in-Place) Suitable “Non-Glass” pH Sensor Based on a Ta2O5-Gate EIS Structure. Sens. Actuators, B 2005, 111−112, 423−429. (147) Wang, J. L.; Yang, P. Y.; Hsieh, T. Y.; Hwang, C. C.; Juang, M. H. PH-Sensing Characteristics of Hydrothermal Al-Doped ZnO Nanostructures. J. Nanomater. 2013, 2013, 1−7. (148) Nascimento, R. A. S.; Mulato, M. Mechanisms of Ion Detection for FET-Sensors Using FTO : Role of Cleaning Process, pH Sequence and Electrical Resistivity. Mater. Res. 2017, 20, 1369−1379. (149) Vasquez, R. P.; Lewis, B. F.; Grunthaner, F. J. X-Ray Photoelectron Spectroscopic Study of the Oxide Removal Mechanism of GaAs (100) Molecular Beam Epitaxial Substrates in in Situ Heating. Appl. Phys. Lett. 1983, 42, 293−295. (150) Ada, M. Y.; Ide, Y. Direct Observation of Species Liberated from GaAs Native Oxides during Atomic Hydrogen Cleaning. Jpn. J. Appl. Phys. 1994, 33, L683−L685. (151) Yamaguchi, K.; Qin, Z.; Nagano, H.; Kobayashi, M.; Yoshikawa, A.; Takahashi, K. Atomically Flat GaAs(001) Surfaces Obtained by High-Temperature Treatment with Atomic Hydrogen Irradiation. Japanese J. Appl. Physics, Part 2 Lett. 1997, 36, L1367. (152) King, S. W.; Barnak, J. P.; Bremser, M. D.; Tracy, K. M.; Ronning, C.; Davis, R. F.; Nemanich, R. J. Cleaning of AlN and GaN Surfaces. J. Appl. Phys. 1998, 84, 5248−5260. (153) Hach Company; pH Electrode Cleaning & Maintenance Guide; 2014; Vol. 1. (154) Oh, J. Y.; Jang, H. J.; Cho, W. J.; Islam, M. S. Highly Sensitive Electrolyte-Insulator- Semiconductor pH Sensors Enabled by Silicon Nanowires with Al2O3/SiO2 Sensing Membrane. Sens. Actuators, B 2012, 171−172, 238−243. (155) Rasheed, H. S.; Ahmed, N. M.; Matjafri, M. Z. Ag Metal Mid Layer Based on New Sensing Multilayers Structure Extended Gate Field Effect Transistor (EG-FET) for pH Sensor. Mater. Sci. Semicond. Process. 2018, 74, 51−56. (156) Ding, X.; Yang, S.; Miao, B.; Gu, L.; Gu, Z.; Zhang, J.; Wu, B.; Wang, H.; Wu, D.; Li, J. Molecular Gated-AlGaN/GaN High Electron Mobility Transistor for pH Detection. Analyst 2018, 143, 2784−2789. (157) Coppa, B. J.; Fulton, C. C.; Kiesel, S. M.; Davis, R. F.; Pandarinath, C.; Burnette, J. E.; Nemanich, R. J.; Smith, D. J. Structural, Microstructural, and Electrical Properties of Gold Films and Schottky Contacts on Remote Plasma-Cleaned, n -Type ZnO{0001} Surfaces. J. Appl. Phys. 2005, 97, 103517. (158) Kumar, K.; Hughes, G.; Mcglynn, E.; Biswas, M. XPS Study of Cleaning Procedures of ZnO (0001), (000−1), (10−10) Surfaces. In International Conference on Solid Films and Surfaces; Dublin, Ireland, 2008; p 1. (159) Ali, G. M.; Ra’ad, H. D.; Abdullateef, A. A. pH Sensing Characteristics of EGFET Based on Pd-Doped ZnO Thin Films Synthesized by Sol-Gel Method In Technological Advances in Electrical, Electronics and Computer Engineering (TAEECE), 2015 Third International Conference on; IEEE: Beirut, Lebanon, 2015; pp 234−238. (160) Wang, L.; Bu, Y.; Ao, J. P. Effect of Oxygen Plasma Treatment on the Performance of AlGaN/GaN Ion-Sensitive Field-Effect Transistors. Diamond Relat. Mater. 2017, 73, 1−6. (161) Coppa, B. J.; Fulton, C. C.; Hartlieb, P. J.; Davis, R. F.; Rodriguez, B. J.; Shields, B. J.; Nemanich, R. J. In Situ Cleaning and Characterization of Oxygen- and Zinc-Terminated, n-Type, ZnO{0001} Surfaces. J. Appl. Phys. 2004, 95, 5856−5864. (162) Cao, H.; Landge, V.; Tata, U.; Seo, Y. S.; Rao, S.; Tang, S. J.; Tibbals, H. F.; Spechler, S.; Chiao, J. C. An Implantable, Batteryless, and Wireless Capsule with Integrated Impedance and pH Sensors for AT DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review Gastroesophageal Reflux Monitoring. IEEE Trans. Biomed. Eng. 2012, 59, 3131−3139. (163) Linkohr, St.; Pletschen, W.; Schwarz, S. U.; Anzt, J.; Cimalla, V.; Ambacher, O. CIP (Cleaning-in-Place) Stability of AlGaN/GaN pH Sensors. J. Biotechnol. 2013, 163, 354−361. (164) Behnood, A.; Van Tittelboom, K.; De Belie, N. Methods for Measuring pH in Concrete: A Review. Constr. Build. Mater. 2016, 105, 176−188. (165) Van Thanh, P.; Nhu, L. T. Q.; Mai, H. H.; Tuyen, N. V.; Doanh, S. C.; Viet, N. C.; Kien, D. T. Zinc Oxide Nanorods Grown on Printed Circuit Board for Extended-Gate Field- Effect Transistor pH Sensor. J. Electron. Mater. 2017, 46, 3732−3737. (166) Rosli, A. B.; Abd Patah, N. D. H.; Rosdan, M. A.; Marbie, M. M.; Juhari, M. H.; Shariffudin, S. S.; Herman, S. H.; Rusop, M. Influence of Metal Catalyst for Zinc Oxide Nanostructures Grown by TCVD Method for Extended-Gate FET Sensor Application. In Proceedings RSM 2013:2013 IEEE Regional Symposium on Micro and Nano Electronics; IEEE: Langkawi, Malaysia, 2013; pp 171−174. (167) Fernandes, J. C.; Mulato, M. ZnO Thin Films Applied as pH Sensor in EGFET Devices. In MRS Proceedings; Cambridge University Press, 2015; Vol. 1805, pp mrss15−2136685. (168) Li, Q.; Tang, W.; Su, Y.; Huang, Y.; Peng, S.; Zhuo, B.; Qiu, S.; Ding, L.; Li, Y.; Guo, X. Stable Thin-Film Reference Electrode on Plastic Substrate for All-Solid-State Ion- Sensitive Field-Effect Transistor Sensing System. IEEE Electron Device Lett. 2017, 38, 1469−1472. (169) Chiu, Y. S.; Lee, C. T.; Lou, L. R.; Ho, S. C.; Chuang, C.-T. Wide Linear Sensing Sensors Using ZnO:Ta Extended-Gate FieldEffect- Transistors. Sens. Actuators, B 2013, 188, 944−948. (170) Lin, J. C.; Huang, B. R.; Yang, Y. K. IGZO NanoparticleModified Silicon Nanowires as Extended-Gate Field-Effect Transistor pH Sensors. Sens. Actuators, B 2013, 184, 27−32. (171) Rasheed, H. S.; Ahmed, N. M.; Matjafri, M. Z.; Al-Hardan, N. H.; Almessiere, M. A.; Sabah, F. A.; Al-Hazeem, N. Z. Multilayer ZnO/ Pd/ZnO Structure as Sensing Membrane for Extended-Gate FieldEffect Transistor (EGFET) with High pH Sensitivity. J. Electron. Mater. 2017, 46, 5901−5908. (172) Lee, C. T.; Chiu, Y. S.; Ho, S. C.; Lee, Y. J. Investigation of a Photoelectrochemical Passivated ZnO-Based Glucose Biosensor. Sensors 2011, 11, 4648−4655. (173) Chang, S.-P.; Li, C.-W.; Chen, K.-J.; Chang, S.-J.; Hsu, C.-L.; Hsueh, T.-J.; Hsueh, H.-T. ZnO-Nanowire-Based Extended-Gate Field-Effect-Transistor pH Sensors Prepared on Glass Substrate. Sci. Adv. Mater. 2012, 4, 1174−1178. (174) Goldstein, S. R.; Peterson, J. I.; Fitzgerald, R. V. A Miniature Fiber Optic pH Sensor for Physiological Use. J. Biomech. Eng. 1980, 102, 141−146. (175) Grant, S. A.; Bettencourt, K.; Krulevitch, P.; Hamilton, J.; Glass, R. In Vitro and in Vivo Measurements of Fiber Optic and Electrochemical Sensors to Monitor Brain Tissue pH. Sens. Actuators, B 2001, 72, 174−179. (176) Chang, K. M.; Chang, C. T.; Chao, K. Y.; Lin, C. H. A Novel PH-Dependent Drift Improvement Method for Zirconium Dioxide Gated PH-Ion Sensitive Field Effect Transistors. Sensors 2010, 10, 4643−4654. (177) Zhou, B.; Bian, C.; Tong, J.; Xia, S. Fabrication of a Miniature Multi-Parameter Sensor Chip for Water Quality Assessment. Sensors (Switzerland) 2017, 17, 1−14. (178) Rigante, S.; Scarbolo, P.; Wipf, M.; Stoop, R. L.; Bedner, K.; Buitrago, E.; Bazigos, A.; Bouvet, D.; Calame, M.; Schönenberger, C.; et al. Sensing with Advanced Computing Technology: Fin Field-Effect Transistors with High-k Gate Stack on Bulk Silicon. ACS Nano 2015, 9, 4872−4881. (179) Casey, J. R.; Grinstein, S.; Orlowski, J. Sensors and Regulators of Intracellular pH. Nat. Rev. Mol. Cell Biol. 2010, 11, 50−61. (180) Shen, Y.; Liang, L.; Zhang, S.; Huang, D.; Zhang, J.; Xu, S.; Liang, C.; Xu, W. Organelle- Targeting Surface-Enhanced Raman Scattering (SERS) Nanosensors for Subcellular pH Sensing. Nanoscale 2018, 10, 1622−1630. (181) Gerweck, L. E.; Seetharaman, K. Cellular pH Gradient in Tumor versus Normal Tissue: Potential Exploitation for the Treatment of Cancer. Cancer Res. 1996, 56, 1194−1198. (182) Damaghi, M.; Wojtkowiak, J. W.; Gillies, R. J. pH Sensing and Regulation in Cancer. Front. Physiol. 2013, 4, 370. (183) Susumu, K.; Field, L. D.; Oh, E.; Hunt, M.; Delehanty, J. B.; Palomo, V.; Dawson, P. E.; Huston, A. L.; Medintz, I. L. Purple-, Blue-, and Green-Emitting Multishell Alloyed Quantum Dots: Synthesis, Characterization, and Application for Ratiometric Extracellular pH Sensing. Chem. Mater. 2017, 29, 7330−7344. (184) Chen, S.; Hong, Y.; Liu, Y.; Liu, J.; Leung, C. W. T.; Li, M.; Kwok, R. T. K.; Zhao, E.; Lam, J. W. Y.; Yu, Y.; et al. Full-Range Intracellular pH Sensing by an Aggregation- Induced Emission-Active Two-Channel Ratiometric Fluorogen. J. Am. Chem. Soc. 2013, 135, 4926−4929. (185) Chandra, A.; Singh, N. Cell Microenvironment pH Sensing in 3D Microgels Using Fluorescent Carbon Dots. ACS Biomater. Sci. Eng. 2017, 3, 3620−3627. (186) Düwel, S.; Hundshammer, C.; Gersch, M.; Feuerecker, B.; Steiger, K.; Buck, A.; Walch, A.; Haase, A.; Glaser, S. J.; Schwaiger, A.; Schilling, F. Imaging of pH in Vivo Using Hyperpolarized 13C-Labelled Zymonic Acid. Nat. Commun. 2017, 8, 15126. (187) Koeppen, B. M. Renal Regulation of Acid-Bace Balance. Adv. Physiol Educ. 1998, 20, 275−S132. (188) Cohen, R. S. Acidosis and Alkalosis. In Fetal and Neonatal Brain Injury, 4th ed.; 2009; Vol. 9780521888, pp 402−408. (189) Mitchell, J. H.; Wildenthal, K.; Johnson, R. L. The Effects of Acid-Base Disturbances on Cardiovascular and Pulmonary Function. Kidney Int. 1972, 1, 375−389. (190) Mark, P. R. G. Renal Physiology; Departments of Electrical Engineering, Mechanical Engineering, Massachusetts, and the HarvardMIT Division of Health Sciences and Technology, 2004; pp 2−32. (191) O’Donnell, M. E. Blood-Brain Barrier Na Transporters in Ischemic Stroke. Adv. Pharmacol. 2014, 71, 113−146. (192) Hermansen, L.; Osnes, J. B. Blood and Muscle pH after Maximal Exercise in Man. J. Appl. Physiol. 1972, 32, 304−308. (193) Osborn, J. J. Experimental Hypothermia: Respiratory and Blood pH Changes in Relation to Cardiac Function. Am. J. Physiol. 1953, 175, 389−398. (194) Kuehnle, E.; Herms, S.; Kohls, F.; Kundu, S.; Hillemanns, P.; Staboulidou, I. Correlation of Fetal Scalp Blood Sampling pH with Neonatal Outcome Umbilical Artery pH Value. Arch. Gynecol. Obstet. 2016, 294, 763−770. (195) Del Coso, J.; Hamouti, N.; Aguado-Jimenez, R.; MoraRodriguez, R. Respiratory Compensation and Blood pH Regulation during Variable Intensity Exercise in Trained versus Untrained Subjects. Eur. J. Appl. Physiol. 2009, 107, 83−93. (196) Momiyama, Y.; Yamada, W.; Miyata, K.; Miura, K.; Fukuda, T.; Fuse, J.; Kikuno, T. Prognostic Values of Blood pH and Lactate Levels in Patients Resuscitated from Out-of- Hospital Cardiac Arrest. Acute Med. Surg. 2017, 4, 25−30. (197) Kwong, T.; Robinson, C.; Spencer, D.; Wiseman, O. J.; Karet Frankl, F. E. Accuracy of Urine pH Testing in a Regional Metabolic Renal Clinic: Is the Dipstick Accurate Enough? Urolithiasis 2013, 41, 129−132. (198) Cook, J. D.; Strauss, K. A.; Caplan, Y. H.; LoDico, C. P.; Bush, D. M. Urine PH: The Effects of Time and Temperature after Collection. J. Anal. Toxicol. 2007, 31, 486−496. (199) Crouch, D. J. Oral Fluid Collection: The Neglected Variable in Oral Fluid Testing. Forensic Sci. Int. 2005, 150, 165−173. (200) Kidwell, D. A.; Holland, J. C.; Athanaselis, S. Testing for Drugs of Abuse in Saliva and Sweat. J. Chromatogr., Biomed. Appl. 1998, 713, 111−135. (201) Tabata, M.; Ratanaporncharoen, C.; Asano, A.; Kitasako, Y.; Ikeda, M.; Goda, T.; Matsumoto, A.; Tagami, J.; Miyahara, Y. Miniaturized Ir/IrOx pH Sensor for Quantitative Diagnosis of Dental Caries. Procedia Eng. 2016, 168, 598−601. (202) Fujii, M.; Kitasako, Y.; Sadr, A.; Tagami, J. Roughness and pH Changes of Enamel Surface Induced by Soft Drinks in VitroAU DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review (223) Š vastová, E.; Hulíková, A.; Rafajová, M.; Zat’Ovičová, M.; Gibadulinová, A.; Casini, A.; Cecchi, A.; Scozzafava, A.; Supuran, C. T.; Pastorek, J.; et al. Hypoxia Activates the Capacity of Tumor-Associated Carbonic Anhydrase IX to Acidify Extracellular pH. FEBS Lett. 2004, 577, 439−445. (224) Hu, J.; Stein, A.; Bühlmann, P. Rational Design of All-SolidState Ion-Selective Electrodes and Reference Electrodes. TrAC, Trends Anal. Chem. 2016, 76, 102−114. (225) Carneiro, a. a. O.; Vilela, G. R.; De Araujo, D. B.; Baffa, O. MRI Relaxometry: Methods and Applications. Braz. J. Phys. 2006, 36, 1−7. (226) Marzouk, S. A. M.; Ufer, S.; Buck, R. P.; Johnson, T. A.; Dunlap, L. A.; Cascio, W. E. Electrodeposited Iridium Oxide pH Electrode for Measurement of Extracellular Myocardial Acidosis during Acute Ischemia. Anal. Chem. 1998, 70, 5054−5061. (227) Das, A.; Ko, D. H.; Chen, C. H.; Chang, L. B.; Lai, C. S.; Chu, F. C.; Chow, L.; Lin, R. M. Highly Sensitive Palladium Oxide Thin Film Extended Gate FETs as pH Sensor. Sens. Actuators, B 2014, 205, 199− 205. (228) Lee, Y.; Howe, C.; Mishra, S.; Lee, D. S.; Mahmood, M.; Piper, M.; Kim, Y.; Tieu, K.; Byun, H.-S.; Coffey, J. P.; et al. Wireless, Intraoral Hybrid Electronics for Real-Time Quantification of Sodium Intake toward Hypertension Management. Proc. Natl. Acad. Sci. U. S. A. 2018, 115, 5377−5382. (229) Wang, X.; Liu, Z.; Zhang, T. Flexible Sensing Electronics for Wearable/Attachable Health Monitoring. Small 2017, 13, 1−19. (230) Hans, R.; Thomas, S.; Garla, B.; Dagli, R. J.; Hans, M. K. Effect of Various Sugary Beverages on Salivary PH, Flow Rate, and Oral Clearance Rate amongst Adults. Scientifica 2016, 2016, 3−8. (231) Oncescu, V.; O’Dell, D.; Erickson, D. Smartphone Based Health Accessory for Colorimetric Detection of Biomarkers in Sweat and Saliva. Lab Chip 2013, 13, 3232−3238. (232) Marzouk, S. A. M.; Buck, R. P.; Dunlap, L. A.; Johnson, T. A.; Cascio, W. E. Measurement of Extracellular PH, K+, and Lactate in Ischemic Heart. Anal. Biochem. 2002, 308, 52−60. (233) Zhou, J.; Zhang, L.; Tian, Y. Micro Electrochemical pH Sensor Applicable for Real-Time Ratiometric Monitoring of pH Values in Rat Brains. Anal. Chem. 2016, 88, 2113−2118. (234) Chung, H.-J.; Sulkin, M. S. Ultrathin, Stretchable, Multiplexing pH Sensor Arrays on Biomedical Devices With Demonstrations on Rabbit and Human Hearts Undergoing Ischemia. Adv. Healthc. Mater. 2015, 3, 59−68. (235) Tait, G. A.; Young, R. B.; Wilson, G. J.; Steward, D. J.; MacGregor, D. C. Myocardial pH during Regional Ischemia: Evaluation of a Fiber-Optic Photometric Probe. Am. J. Physiol. 1982, 243, H1027−31. (236) Watanabe, T.; Kobayashi, K.; Suzuki, T.; Oizumi, M.; Clark, G. T. A Preliminary Report on Continuous Recording of Salivary pH Using Telemetry in an Edentulous Patient. Int. J. Prosthodont. 1999, 12. (237) Soller, B. R.; Stahl, R. F.; Knox, M. Microsensors for Monitoring Myocardial Ischemia: The Concept of pH Time Integral. In Engineering in Medicine and Biology Society, 1993. Proceedings of the 15th Annual International Conference of the IEEE; IEEE: San Diego, CA, USA, 1993; pp 1582−1583. (238) Rai, P.; Jung, S.; Ji, T.; Varadan, V. K. Ion-Sensitive Field Effect Transistors for pH and Potassium Ion Concentration Sensing: Towards Detection of Myocardial Ischemia. In Nanosensors and Microsensors for Bio-Systems 2008; International Society for Optics and Photonics: San Diego, CA, USA, 2008; p 69310I. (239) Anastasova, S.; Kassanos, P.; Yang, G. Z. Multi-Parametric Rigid and Flexible, Low- Cost, Disposable Sensing Platforms for Biomedical Applications. Biosens. Bioelectron. 2018, 102, 668−675. (240) Curto, V. F.; Coyle, S.; Byrne, R.; Angelov, N.; Diamond, D.; Benito-Lopez, F. Concept and Development of an Autonomous Wearable Micro-Fluidic Platform for Real Time pH Sweat Analysis. Sens. Actuators, B 2012, 175, 263−270. (241) Callewaert, C.; Buysschaert, B.; Vossen, E.; Fievez, V.; Van de Wiele, T.; Boon, N. Artificial Sweat Composition to Grow and Sustain a Mixed Human Axillary Microbiome. J. Microbiol. Methods 2014, 103, 6−8. Applications of Stylus Profilometry, Focus Variation 3D Scanning Microscopy and Micro pH Sensor. Dent. Mater. J. 2011, 30, 404−410. (203) Murakami, K.; Kitasako, Y.; Burrow, M. F.; Tagami, J. In Vitro pH Analysis of Active and Arrested Dentinal Caries in Extracted Human Teeth Using a Micro pH Sensor. Dent. Mater. J. 2006, 25, 423− 429. (204) Ratanaporncharoen, C.; Tabata, M.; Kitasako, Y.; Ikeda, M.; Goda, T.; Matsumoto, A.; Tagami, J.; Miyahara, Y. pH Mapping on Tooth Surfaces for Quantitative Caries Diagnosis Using Micro Ir/IrOx pH Sensor. Anal. Chem. 2018, 90, 4925−4931. (205) Chaisiwamongkhol, K.; Batchelor-Mcauley, C.; Compton, R. G. Amperometric Micro pH Measurements in Oxygenated Saliva. Analyst 2017, 142, 2828−2835. (206) Baliga, S.; Muglikar, S.; Kale, R. Salivary PH: A Diagnostic Biomarker. J. Indian Soc. Periodontol. 2013, 17, 461. (207) Seethalakshmi, C.; Jagat Reddy, R. C.; Asifa, N.; Prabhu, S. Correlation of Salivary PH, Incidence of Dental Caries and Periodontal Status in Diabetes Mellitus Patients: A Cross- Sectional Study. J. Clin. Diagnostic Res. 2016, 10, ZC12−ZC14. (208) Rathnayake, N.; Åkerman, S.; Klinge, B.; Lundegren, N.; Jansson, H.; Tryselius, Y.; Sorsa, T.; Gustafsson, A. Salivary Biomarkers for Detection of Systemic Diseases. PLoS One 2013, 8, No. e61356. (209) Yoshizawa, J. M.; Schafer, C. A.; Schafer, J. J.; Farrell, J. J.; Paster, B. J.; Wong, D. T. W. Salivary Biomarkers: Toward Future Clinical and Diagnostic Utilities. Clin. Microbiol. Rev. 2013, 26, 781− 791. (210) Jo, J.; Lee, C. H.; Kopelman, R.; Wang, X. In Vivo Quantitative Imaging of Tumor pH by Nanosonophore Assisted Multispectral Photoacoustic Imaging. Nat. Commun. 2017, 8, 471. (211) Enriquez-Navas, P. M.; Gillies, R. J. Measuring PHi and PHe by MRS. eMagRes. 2015, 4, 643−650. (212) Kaneniwa, N.; Otsuka, M. Evaluation of Acidogenicity of Commercial Isomaltooligosaccharides Mixture and Its Hydrogenated Derivative by Measurement of pH Response under Human Dental Plaque. Chem. Pharm. Bull. 1985, 33, 1660−1668. (213) Parvinzadeh Gashti, M.; Asselin, J.; Barbeau, J.; Boudreau, D.; Greener, J. A Microfluidic Platform with pH Imaging for Chemical and Hydrodynamic Stimulation of Intact Oral Biofilms. Lab Chip 2016, 16, 1412−1419. (214) Hashim, A. I.; Zhang, X.; Wojtkowiak, J. W.; Martinez, G. V.; Gillies, R. J. Imaging pH and Metastasis. NMR Biomed. 2011, 24, 582− 591. (215) Zhang, X.; Lin, Y.; Gillies, R. J. Tumor pH and Its Measurement. J. Nucl. Med. 2010, 51, 1167−1170. (216) Gallagher, F. A.; Kettunen, M. I.; Day, S. E.; Hu, D.-E.; Ardenkjær-Larsen, J. H.; Zandt, R.; Jensen, P. R.; Karlsson, M.; Golman, K.; Lerche, M. H.; Brindle, K. M. Magnetic Resonance Imaging of pH in Vivo Using Hyperpolarized 13C-Labelled Bicarbonate. Nature 2008, 453, 940−943. (217) Malhotra, D.; Casey, J. R. Intracellular pH Measurement. In eLS; 2015; pp 1−7. (218) Lardner, A. The Effects of Extracellular pH on Immune Function. J. Leukocyte Biol. 2001, 69, 522−530. (219) Paradise, R. K.; Lauffenburger, D. A.; van Vliet, K. J. Acidic Extracellular pH Promotes Activation of Integrin Αvβ3. PLoS One 2011, 6, No. e15746. (220) Hayata, H.; Miyazaki, H.; Niisato, N.; Yokoyama, N.; Marunaka, Y. Lowered Extracellular pH Is Involved in the Pathogenesis of Skeletal Muscle Insulin Resistance. Biochem. Biophys. Res. Commun. 2014, 445, 170−174. (221) Kenney, R. M.; Boyce, M. W.; Whitman, N. A.; Kromhout, B. P.; Lockett, M. R. A PH- Sensing Optode for Mapping Spatiotemporal Gradients in 3D Paper-Based Cell Cultures. Anal. Chem. 2018, 90, 2376−2383. (222) Munteanu, R. E.; Stǎnicǎ, L.; Gheorghiu, M.; Gáspár, S. Measurement of the Extracellular pH of Adherently Growing Mammalian Cells with High Spatial Resolution Using a Voltammetric pH Microsensor. Anal. Chem. 2018, 90, 6899−6905. AV DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review (242) Jadoon, S.; Karim, S.; Akram, M. R.; Kalsoom Khan, A.; Zia, M. A.; Siddiqi, A. R.; Murtaza, G. Recent Developments in Sweat Analysis and Its Applications. Int. J. Anal. Chem. 2015, 2015, 164974. (243) Caldara, M.; Colleoni, C.; Guido, E.; Re, V.; Rosace, G. Optical Monitoring of Sweat pH by a Textile Fabric Wearable Sensor Based on Covalently Bonded Litmus-3- Glycidoxypropyltrimethoxysilane Coating. Sens. Actuators, B 2016, 222, 213−220. (244) Levitt, D. G.; Schoemaker, R. C. Human Physiologically Based Pharmacokinetic Model for ACE Inhibitors: Ramipril and Ramiprilat. BMC Clin. Pharmacol. 2006, 6, 159−169. (245) Nikolajek, W. P.; Emrich, H. M. pH of Sweat of Patients with Cystic Fibrosis. Klin. Wochenschr. 1976, 54, 287−288. (246) Luckie, D. B.; Krouse, M. E. Cystic Fibrosis: Does CFTR Malfunction Alter pH Regulation? In Genetic Disorders; 2013; pp 319− 344. (247) Dang, W.; Manjakkal, L.; Navaraj, W. T.; Lorenzelli, L.; Vinciguerra, V.; Dahiya, R. Stretchable Wireless System for Sweat pH Monitoring. Biosens. Bioelectron. 2018, 107, 192−202. (248) Caplan, Y. H.; Goldberger, B. A. Alternative Specimens for Workplace Drug Testing. J. Anal. Toxicol. 2001, 25, 396−399. (249) Curto, V. F.; Fay, C.; Coyle, S.; Byrne, R.; O’Toole, C.; Barry, C.; Hughes, S.; Moyna, N.; Diamond, D.; Benito-Lopez, F. Real-Time Sweat pH Monitoring Based on a Wearable Chemical Barcode MicroFluidic Platform Incorporating Ionic Liquids. Sens. Actuators, B 2012, 171−172, 1327−1334. (250) Bandodkar, A. J.; Wang, J. Non-Invasive Wearable Electrochemical Sensors: A Review. Trends Biotechnol. 2014, 32, 363−371. (251) Bandodkar, A. J.; Hung, V. W. S.; Jia, W.; Valdés-Ramírez, G.; Windmiller, J. R.; Martinez, A. G.; Ramírez, J.; Chan, G.; Kerman, K.; Wang, J. Tattoo-Based Potentiometric Ion-Selective Sensors for Epidermal pH Monitoring. Analyst 2013, 138, 123−128. (252) Guinovart, T.; Parrilla, M.; Crespo, G. A.; Rius, F. X.; Andrade, F. J. Potentiometric Sensors Using Cotton Yarns, Carbon Nanotubes and Polymeric Membranes. Analyst 2013, 138, 5208−5215. (253) Karyakin, A. A.; Vuki, M.; Lukachova, L. V.; Karyakina, E. E.; Orlov, A. V.; Karpachova, G. P.; Wang, J. Processible Polyaniline as an Advanced Potentiometric pH Transducer. Application to Biosensors. Anal. Chem. 1999, 71, 2534−2540. (254) Curto, V. F.; Coyle, S.; Byrne, R.; Angelov, N.; Diamond, D.; Benito-Lopez, F. Concept and Development of an Autonomous Wearable Micro-Fluidic Platform for Real Time pH Sweat Analysis. Sens. Actuators, B 2012, 175, 263−270. (255) Nakata, S.; Arie, T.; Akita, S.; Takei, K. Wearable, Flexible, and Multifunctional Healthcare Device with an ISFET Chemical Sensor for Simultaneous Sweat pH and Skin Temperature Monitoring. ACS Sensors 2017, 2, 443−448. (256) Lee, H.; Song, C.; Hong, Y. S.; Kim, M. S.; Cho, H. R.; Kang, T.; Shin, K.; Choi, S. H.; Hyeon, T.; Kim, D. Wearable/Disposable SweatBased Glucose Monitoring Device with Multistage Transdermal Drug Delivery Module. Sci. Adv. 2017, 3, e1601314. (257) Trovato, V.; Rosace, G.; Colleoni, C.; Plutino, M. R. Synthesis and Characterization of Halochromic Hybrid Sol-Gel for the Development of a pH Sensor Fabric. In IOP Conference Series: Materials Science and Engineering; IOP Publishing: Corfu, Greece, 2017; Vol. 254. (258) Caldara, M.; Colleoni, C.; Guido, E.; Re, V.; Rosace, G.; Vitali, A. Textile Based Colorimetric pH Sensing - A Platform for Future Wearable pH Monitoring. In Proceedings - BSN 2012:9th International Workshop on Wearable and Implantable Body Sensor Networks; IEEE: London, UK, 2012; pp 11−16. (259) Tamayol, A.; Akbari, M.; Zilberman, Y.; Comotto, M.; Lesha, E.; Serex, L.; Bagherifard, S.; Chen, Y.; Fu, G.; Ameri, S. K.; et al. Flexible PH-Sensing Hydrogel Fibers for Epidermal Applications. Adv. Healthcare Mater. 2016, 5, 711−719. (260) Zhu, Z.; Liu, T.; Li, G.; Li, T.; Inoue, Y. Wearable Sensor Systems for Infants. Sensors 2015, 15, 3721−3749. (261) Sen, C. K.; Gordillo, G. M.; Roy, S.; Kirsner, R.; Lambert, L.; Hunt, T. K.; Gottrup, F.; Gurtner, G. C.; Longaker, M. T. Human Skin Wounds: A Major Snoballing Threat to Public Health and Economy. Wound Repair Regen. 2009, 17, 763−771. (262) Guinovart, T.; Valdés-Ramírez, G.; Windmiller, J. R.; Andrade, F. J.; Wang, J. Bandage- Based Wearable Potentiometric Sensor for Monitoring Wound pH. Electroanalysis 2014, 26, 1345−1353. (263) Roychoudhury, S.; Umasankar, Y.; Jaller, J.; Herskovitz, I.; Mervis, J.; Darwin, E.; Hirt, P. A.; Borda, L. J.; Lev-tov, H. A.; Kirsner, R.; Bhansali, S. Continuous Monitoring of Wound Healing Using a Wearable Enzymatic Uric Acid Biosensor. J. Electrochem. Soc. 2018, 165, B3168−B3175. (264) Mostafalu, P.; Tamayol, A.; Rahimi, R.; Ochoa, M.; Khalilpour, A.; Kiaee, G.; Yazdi, I. K.; Bagherifard, S.; Dokmeci, M. R.; Ziaie, B.; et al. Smart Bandage for Monitoring and Treatment of Chronic Wounds. Small 2018, 14, 1703509. (265) Stoppa, M.; Chiolerio, A. Wearable Electronics and Smart Textiles: A Critical Review. Sensors 2014, 14, 11957−11992. (266) Giachet, F. T.; Vineis, C.; Sanchez Ramirez, D. O.; Carletto, R. A.; Varesano, A.; Mazzuchetti, G. Reversible and Washing Resistant Textile-Based Optical pH Sensors by Dyeing Fabrics with Curcuma. Fibers Polym. 2017, 18, 720−730. (267) Rahimi, R.; Ochoa, M.; Tamayol, A.; Khalili, S.; Khademhosseini, A.; Ziaie, B. Highly Stretchable Potentiometric pH Sensor Fabricated via Laser Carbonization and Machining of Carbon− Polyaniline Composite. ACS Appl. Mater. Interfaces 2017, 9, 9015− 9023. (268) Frost, M. C.; Meyerhoff, M. E. Implantable Chemical Sensors for Real-Time Clinical Monitoring : Progress and Challenges. Curr. Opin. Chem. Biol. 2002, 6, 633−641. (269) Hao, J.; Xiao, T.; Wu, F.; Yu, P.; Mao, L. High Antifouling Property of Ion-Selective Membrane: Toward in Vivo Monitoring of pH Change in Live Brain of Rats with Membrane-Coated Carbon Fiber Electrodes. Anal. Chem. 2016, 88, 11238−11243. (270) Wencel, D.; Kaworek, A.; Abel, T.; Efremov, V.; Bradford, A.; Carthy, D.; Coady, G.; McMorrow, R. C. N.; McDonagh, C. Optical Sensor for Real-Time pH Monitoring in Human Tissue. Small 2018, 14, 1803627. (271) Wang, D.; Yang, C.; Xia, J.; Xue, Z.; Jiang, D.; Zhou, G.; Zheng, X.; Zheng, H.; Du, Y.; Li, Q. Improvement of the Ir/IrO2 pH Electrode via Hydrothermal Treatment. Ionics 2017, 23, 2167−2174. (272) Wang, M.; Yao, S.; Madou, M. A Long-Term Stable Iridium Oxide pH Electrode. Sens. Actuators, B 2002, 81, 313−315. (273) Vanamo, U.; Bobacka, J. Instrument-Free Control of the Standard Potential of Potentiometric Solid-Contact Ion-Selective Electrodes by Short-Circuiting with a Conventional Reference Electrode. Anal. Chem. 2014, 86, 10540−10545. (274) Sabah, F. A.; Ahmed, N. M.; Hassan, Z.; Abdullah Almessiere, M. Influences of Substrate Type on the pH Sensitivity of CuS Thin Films EGFET Prepared by Spray Pyrolysis Deposition. Mater. Sci. Semicond. Process. 2017, 63, 269−278. (275) Nguyen, C. M.; Huang, W.; Rao, S.; Cao, H.; Tata, U.; Chiao, M.; Chiao, J. Sol-Gel Iridium Oxide-Based pH Sensor Array on Flexible Polyimide Substrate. IEEE Sens. J. 2013, 13, 3857−3864. (276) Ghoneim, M. T.; Zidan, M. A.; Alnassar, M. Y.; Hanna, A. N.; Kosel, J.; Salama, K. N.; Hussain, M. M. Thin PZT-Based Ferroelectric Capacitors on Flexible Silicon for Nonvolatile Memory Applications. Adv. Electron. Mater. 2015, 1, 1500045. (277) Ghoneim, M.; Hussain, M. Review on Physically Flexible Nonvolatile Memory for Internet of Everything Electronics. Electronics 2015, 4, 424−479. (278) Kutbee, A. T.; Bahabry, R. R.; Alamoudi, K. O.; Ghoneim, M. T.; Cordero, M. D.; Almuslem, A. S.; Gumus, A.; Diallo, E. M.; Nassar, J. M.; Hussain, A. M.; Khashab, N. M.; Hussain, M. M. Flexible and Biocompatible High-Performance Solid-State Micro-Battery for Implantable Orthodontic System. npj Flex. Electron. 2017, 1, 7. (279) Alfaraj, N.; Hussain, A. M.; Torres Sevilla, G. A.; Ghoneim, M. T.; Rojas, J. P.; Aljedaani, A. B.; Hussain, M. M. Functional Integrity of Flexible N-Channel Metal− oxide−semiconductor Field-Effect Transistors on a Reversibly Bistable Platform. Appl. Phys. Lett. 2015, 107, 174101. AW DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review (280) Shaikh, S. F.; Ghoneim, M. T.; Bahabry, R. R.; Khan, S. M.; Hussain, M. M. Modular Lego-Electronics. Adv. Mater. Technol. 2018, 3, 1700147. (281) Ghoneim, M. T.; Alfaraj, N.; Torres-Sevilla, G. A.; Fahad, H. M.; Hussain, M. M. Out-of- Plane Strain Effects on Physically Flexible FinFET CMOS. IEEE Trans. Electron Devices 2016, 63, 2657−2664. (282) Prats-Alfonso, E.; Abad, L.; Casañ-Pastor, N.; Gonzalo-Ruiz, J.; Baldrich, E. Iridium Oxide pH Sensor for Biomedical Applications. Case Urea-Urease in Real Urine Samples. Biosens. Bioelectron. 2013, 39, 163−169. (283) Huang, B. R.; Hsu, C. L.; Wang, Y. K.; Yang, W. L. Core-Shell PN Junction Si Nanowires as Rapid Response and High-Sensitivity pH Sensor. IEEE Sens. J. 2017, 17, 3967−3974. (284) Wu, S.; Wu, Y.; Lin, C. High Performance EGFET-Based Ph Sensor Utilizing Low- Cost Industrial-Grade Touch Panel Film As the Gate Structure. uTAS Proc. 2014, 15, 2131−2133. (285) Salvo, P.; Calisi, N.; Melai, B.; Dini, V.; Paoletti, C.; Lomonaco, T.; Pucci, A.; Di Francesco, F.; Piaggesi, A.; Romanelli, M. Temperature-and PH-Sensitive Wearable Materials for Monitoring Foot Ulcers. Int. J. Nanomed. 2017, 12, 949−954. (286) Meruva, R. K.; Meyerhoff, M. E. Catheter-Type Sensor for Potentiometric Monitoring of Oxygen, pH and Carbon Dioxide. Biosens. Bioelectron. 1998, 13, 201−212. (287) Macdonald, D. D.; Liu, J.; Lee, D. Development of W/WO3 Sensors for the Measurement of pH in an Emulsion Polymerization System. J. Appl. Electrochem. 2004, 34, 577−582. (288) Yang, J.; Kwak, T. J.; Zhang, X.; McClain, R.; Chang, W. J.; Gunasekaran, S. Digital pH Test Strips for In-Field pH Monitoring Using Iridium Oxide-Reduced Graphene Oxide Hybrid Thin Films. ACS Sensors 2016, 1, 1235−1243. (289) Lonsdale, W.; Wajrak, M.; Alameh, K. Effect of Conditioning Protocol, Redox Species and Material Thickness on the pH Sensitivity and Hysteresis of Sputtered RuO2electrodes. Sens. Actuators, B 2017, 252, 251−256. (290) Voigt, H.; Schitthelm, F.; Lange, T.; Kullick, T.; Ferretti, R. Diamond-like Carbon-Gate PH-ISFET. Sens. Actuators, B 1997, 44, 441−445. (291) Bartic, C.; Palan, B.; Campitelli, A.; Borghs, G. Monitoring pH with Organic-Based Field- Effect Transistors. Sens. Actuators, B 2002, 83, 115−122. (292) Graham, D. J.; Jaselskis, B.; Moore, C. E. Development of the Glass Electrode and the pH Response. J. Chem. Educ. 2013, 90, 345− 351. (293) Pantelis, D. M.; Experton, J.; Martin, C. R. Rearranging the Nernst Equation to Make a Dosage-Controllable Membrane Delivery System. J. Electroanal. Chem. 2018, 819, 73−77. (294) Scholz, F. Nikolsky’s Ion Exchange Theory versus Baucke’s Dissociation Mechanism of the Glass Electrode. J. Solid State Electrochem. 2011, 15, 67−68. (295) Durst, R. A. Mechanism of the Glass Electrode Response. J. Chem. Educ. 1967, 44, 175. (296) Baucke, F. G. K. The Modern Understanding of the Glass Electrode Response. Fresenius' J. Anal. Chem. 1994, 349, 582−596. (297) Cheng, K. L. A New Concept for PH-Potential Calculations. J. Chem. Educ. 1999, 76, 1029. (298) Mauro, A. Space Charge Regions in Fixed Charge Membranes and the Associated Property of Capacitance. Biophys. J. 1962, 2, 179− 198. (299) Oldham, K. B. A Gouy-Chapman-Stern Model of the Double Layer at a (Metal)/(Ionic Liquid) Interface. J. Electroanal. Chem. 2008, 613, 131−138. (300) Al-Hilli, S.; Willander, M. The pH Response and Sensing Mechanism of N-Type ZnO/Electrolyte Interfaces. Sensors 2009, 9, 7445−7480. (301) Yates, D. E.; Levine, S.; Healy, T. W. Site-Binding Model of the Electrical Double Layer at the Oxide/Water Interface. J. Chem. Soc., Faraday Trans. 1 1974, 70, 1807−1818. (302) Parizi, K. B.; Xu, X.; Pal, A.; Hu, X.; Wong, H. S. P. ISFET pH Sensitivity: Counter- Ions Play a Key Role. Sci. Rep. 2017, 7, 1−10. (303) Huang, W. D.; Cao, H.; Deb, S.; Chiao, M.; Chiao, J. C. A Flexible pH Sensor Based on the Iridium Oxide Sensing Film. Sens. Actuators, A 2011, 169, 1−11. (304) Kofstad, P. Defects and Transport Properties of Metal Oxides. Oxid. Met. 1995, 44, 3−27. (305) Pilon, L.; Wang, H.; d’Entremont, A. Recent Advances in Continuum Modeling of Interfacial and Transport Phenomena in Electric Double Layer Capacitors. J. Electrochem. Soc. 2015, 162, A5158−A5178. (306) Batista, P. D.; Mulato, M. ZnO Extended-Gate Field-Effect Transistors as pH Sensors. Appl. Phys. Lett. 2005, 87, 143508. (307) Su, Y.; Dagdeviren, C.; Li, R. Measured Output Voltages of Piezoelectric Devices Depend on the Resistance of Voltmeter. Adv. Funct. Mater. 2015, 25, 5320−5325. (308) Su, Y.; Li, S.; Huan, Y.; Li, R.; Zhang, Z.; Joe, P.; Dagdeviren, C. The Universal and Easy-to-Use Standard of Voltage Measurement for Quantifying the Performance of Piezoelectric Devices. Extrem. Mech. Lett. 2017, 15, 10−16. AX DOI: 10.1021/acs.chemrev.8b00655 Chem. Rev. XXXX, XXX, XXX−XXX