Selective Ion Sensing Organic Electrochemical Transistors Suitable for Blood Analysis

Organic electrochemical transistors (OECTs) have gained considerable attention due to their ability to simultaneously transduce and amplify ion‐based biological signals into electronic signals in bioelectronics. In this study, functionalized OECTs capable of detecting specific ion concentrations in aqueous solution and blood serum are investigated and described. Sodium, the most concentrated cation in the human body, is chosen as the target analyte due to its critical role in maintaining normal bodily functions such as blood pressure, nerve and muscle function, and fluid balance. These sodium ion sensors work at low source‐drain voltage, Vds = 0.4 V and gate voltage, Vg = 0 V, and demonstrate a high sensitivity of 126 μA dec−1 and a high selectivity over different cations. Furthermore, the OECT biosensors are employed to determine sodium concentration in more complex environments and demonstrated a log‐linear response within the physiological range of sodium in blood serum, ranging from 100 to 160 mm. In the final part of the study, the transistor‐based sensor is fabricated in a small‐footprint neural probe configuration and its sensing capability is investigated. These characteristics open up new opportunities for applications in wearable and implantable electronics.


Introduction
The selective detection and quantification of ionic species in aqueous media is of great interest in numerous fields, including DOI: 10.1002/adsr.2023000973] For example, in agriculture, plant nutritional testing has been shown to be an effective method for increasing crop yields and providing resilience in fluctuating conditions. [4]In biomedical diagnostics, an electrolyte panel is commonly used as part of a routine blood screening to evaluate a patient's clinical state and is associated with morbidity and mortality monitoring. [5]At present, the detection and monitoring of electrolyte levels in the human body is mainly performed using laborious laboratory equipment, hindering the potential application of real-time monitoring, which is becoming increasingly necessary in new fields such as precision medicine or patient-specific therapeutics.[11][12] In OECTs, upon application of a gate voltage (V g ), ions from the electrolyte permeate into or out of the channel, which is typically composed of poly (3,4-ethylenedioxythiophene)  doped with poly-(styrenesulfonate) (PEDOT:PSS).This phenomenon leads to a volumetric alteration of the doping state, resulting in substantial changes in source-drain current (I ds ) at low gate voltages (below 1 V) and high signal amplification (transconductance values up to tens of mS).
While OECTs serve as excellent ionic activity sensors, they are unable to selectively detect or deviate specific ions in a given solution.To address this issue, a plasticized polyvinyl chloride (PVC) matrix-based membrane containing ionophores can be incorporated between the electrolyte and the PEDOT:PSS channel to provide ion selectivity. [13]This sensor operates on the principle of channel modulation of the OECT as ions diffuse into the membrane.The transportation of ions through the interface between the electrolyte and the membrane is governed by the Nernst equation.At equilibrium, diffusion of cations (for instance) across the membrane is counterbalanced by the electric field induced by the ions.As the cation concentration in the solution increases, the electric field required to counterbalance diffusion must increase, resulting in an additional potential being added to the PEDOT:PSS channel, thus leading to a more de-doped channel and a decrease in I ds .
Different groups have conducted extensive efforts to develop OECTs-based ion sensing devices. [14]Strand et al. [15] developed sorbitol doped PEDOT:PSS OECTs for K + detecting in whole plant sap.Keene et al. [16] produced an OECT-based wearable patch on a flexible SEBS substrate by printing and the device can monitor calcium and ammonium ions from sweat.Tseng et al. [17] integrated an OECT and a two-electrode potentiostat in a microelectrode array to determine ion concentration and ion species.To enhance the sensitivity, Han et al. [18] demonstrated ion-selective OECTs (IS-OECTs) with sensitivities beyond the Nernstian limit by inducing an electrochemical gate electrode in a new device architecture.Ghittorelli et al. [19] developed current-driven IS-OECTs with sensitivity exceeding about one order higher than the Nernst limit.The current-driven configuration resembles an inverter topology, where the pull-down transistor is replaced with a current generator.The potential of IS-OECTs to revolutionize biomedical research and applications is undeniable.Their unique characteristics, such as high sensitivity, low cost, and ease of fabrication, make them an ideal tool for a wide range of medical monitoring and analytical tasks.Furthermore, they are highly biocompatible and offer the possibility of long-term implantation within the human body.With their impressive properties, IS-OECTs offer great promise for revolutionizing biomedical research and applications, making them an exciting and promising technology.In this study, we designed and investigated an ion-selective membrane (ISM) integrated PEDOT:PSS OECT as a high sensitivity and selective sensor, particularly for sodium ion (Na+).The potential for clinical applications of these devices was then revealed by detecting Na + at different concentrations in human serum.The results obtained from IS-OECTs were found to be highly comparable to those obtained from standard ion selective electrodes (ISEs), further validating the effectiveness of these devices.Furthermore, the downsizing of these ion sensors into implantable neural probes has opened up new avenues for research in the field of neuroscience.By measuring concentrations of different ions in the cerebral fluid, these probes have the potential to greatly enhance our understanding of brain activity and neuron firing.

Results
The ionically selective OECT consists of a multifunctional layered structure that provides selective detection of target ions as described in Figure 1a.The device incorporates an active layer made of conducting polymer PEDOT:PSS, an ion reservoir layer made of Poly(sodium-4-styrene sulfonate) (PSSNa), and a PVCbased ISM.Analyte contacts with ISM directly.Ag/AgCl pellet electrode was chosen as the gate electrode since its electrode potential is constant in different aqueous solutions and its small size. [20]Details on the device fabrication are provided in the Experimental Section.
First, the pristine PEDOT:PSS OECT performance was characterized before the deposition of the ISM.The results demonstrate a typical transfer curve of PEDOT:PSS working in depletion mode such that a positive V g causes the dedoping of the channel and therefore, reduces I ds until it reaches the off state (Figure S1a, Supporting Information).Under the 1 mm NaCl solution, the PEDOT:PSS channel was completely dedoped at V g of 0.8 V and this resulted in an on/off ratio of ≈1.1 × 10 4 with a maximum transconductance G m of 2.3 mS (Figure S1b, Supporting Information).The on/off ratio and maximum G m of the device decreased to ≈2 and ≈1.2 mS, respectively, after the PEDOT:PSS channel was coated with an ISM.It is worth noticing that the shift of the peak G m toward lower bias (from V g = 0.1 V to V g = 0 V) after adding the ISM layer is attributed to the potential developed in the ISM.The deteriorated device performance clearly demonstrates the ISM acted as an ion barrier, hindering ions from the electrolyte solution from permeating into the PEDOT:PSS channel. [18]Other than the increases in the response time, the sluggish ion transportation in the membrane also leads to fewer ions able to permeate into the channel and electrochemically reduce the PEDOT:PSS. [21]Different methods have been suggested to tackle the decline in OECT performance caused by the ISM.One of these is to optimize the ISM composition, such as increasing the ratio of ionic sites to ionophores, thus enhancing the ionic conductivity of the matrix. [22]However, the effectiveness is limited by the availability of free vacant ionophores. [23]Another approach is by placing an internal electrolyte reservoir between the channel and ISM which can provide the required ions.Among different reported reservoir materials such as polyelectrolytes, [18] aqueous solution, [24] and hydrogels, [25] polyelectrolytes stand out among them due to the operability during device assembly and miniaturization.The performance of IS-OECTs can be improved by inserting a PSSNa reservoir layer between the ISM and channel.This modification allows the device to exhibit a maximum G m of ≈1.7 mS (Figure S1b, Supporting Information) and an on/off ratio of 200 (Figure 1b), demonstrating successful volumetric gating through the mobile Na + ions from the PSSNa layer.Figure 1b inset shows the output curve of the device.It is important to consider the effective applied gate potential when choosing operational voltages for an ion-sensing device.To ensure stable device operation and to avoid degradation due to hydrolysis, the V g applied should be well below 0.7 V.In our measurements, the V g applied was narrowed to a range of −0.2 to 0.6 V to maintain stable device performance.
The time constant () of ion injection into the channel was determined under a fixed Na + concentration (10 −3 m) of the analyte and a gate pulse.By applying an exponential decay fitting, the value of  was estimated to be ≈195 ms (Figure 1c).Although it is still slower than a normal OECT, it is already substantially faster than the previously reported for OECT with ion-selective membrane but no internal reservoir.Those devices have a relaxation time in the order of seconds. [26]Furthermore, the IS-OECT exhibited excellent electrochemical stability.1d shows that, after pulsing the V g with different biases over a period of ≈400 s and three cycles, no significant current shifts were observed with a fixed V ds of −0.2 V.
Before testing under different analytes, the baseline of the IS-OECT with PSSNa reservoir layer is set by DI water as shown in Figure 2a.This transfer curve will be used as the reference curve later.For the other analytes, when the sodium concentration increases, the potential between the analyte and the ISM will increase and thus more Na + will be injected from the PSSNa layer into the PEDOT:PSS and dedoping the channel.The electroanalytical performance of the IS-OECT device was investigated by measuring I ds under different Na + concentrations ranging from 10 −6 to 1 m.The measurements were performed at fixed V g = 0 and V ds = −0.4V as shown in Figure 2b.To achieve an adequate saturation of primary ion uptake and enhance the detection limit, the freshly prepared ion-selective membranes went through a conditioning procedure.1 mm NaCl diluted in DI water was used to precondition the functionalized transistors overnight before the test.The I ds shows step-wise decrease when the analyte is replaced by a higher Na + concentrations.The variation between DI water and 1 m is ≈500 A.The values of current changes, ΔI ds (I ds −I baseline , where I baseline is the channel current in DI water), were plotted against the correspondent Na + concentrations in Figure 2c.The error bars in the graph represented the maximum and minimum values measured from three devices.The device exhibited a log-linear response to Na + ions in the range of 10 −4 to 1 m, with a current sensitivity of 126 A dec −1 .These devices show two order of magnitude improvement compared with previously reported Na + ISM-based electrolyte gated-organic field effect transistor sensors. [24]he analyte/membrane potential (E m ), proportional to the analyte activity (i.e., concentration), is described by Nernstian Equation as follows: [20] E where k Boltzmann's constant, T temperature, n is the number of electrons transferred, e is the elementary charge, M is the concentration of analyte ion, and E 0 is a constant.The decrease of I ds reflects the change in effective gate potential V g,eff (V g,eff = V g +E m ).One can directly evaluate the V g,eff and the membrane potential, E m , by measuring the I ds drop in Figure 2a.The voltage-based sensitivity in the linear regime is 85 mV dec −1 (Figure 2d), higher than the theoretically predicted value of 59.2 mV dec −1 (Equation 1).Similar super-Nerstian response has been reported for OECTs operated by a polarizable gate electrode in electrolytes with different ion concentrations. [27]The uncompensated PSS− polyanions at the ISM/PSSNa interface can be regarded as an electrochemical gate with a capacitance. [18]They are formed by the transportation of Na + ions in the PSSNa layer from the ISM interface into the PEDOT:PSS channel.The limit of detection is 0.11 mm which confirms the ability of the IS-OECT to measure the variation of Na + ionic concentration in the body fluid.
The ion selectivity of IS-OECTs is originated from ionophores embedded in the PVC membrane that selectively and reversibly bind with primary ions.To confirm the selectivity of our device to Na + , separate solution methods (SSM) based on common interfering ions found in biofluids such as potassium (K + ) and calcium (Ca 2+ ) were employed.In the SSM, the IS-OECT contacted with the same solution, and the currents I ds were measured concurrently in real-time while varying the ionic concentration (Figure 2e).Compared with K + and Ca 2+ , the devices showed reasonable selectivity toward Na + (Figure 2f).Relatively large ΔI ds response to K + and Ca 2+ ions were observed at high concentrations due to the fact that certain ions can penetrate through the ISM at high concentrations.The selectivity factors of the Naselective OECT to potassium and calcium ions were calculated to be−logk MPM Na, K = 1.9 and −logk MPM Na, Ca = 2.1 respectively, at their concentration of 0.1 m (Figure S2, Supporting Information).These values imply the selectivity of the IS-OFETs are able to clearly deviate the Ca 2+ or K + from Na + , and they are similar to the reported values using the same ionophores. [28]he reversibility and repeat testing capability of the IS-OECTs were confirmed by measuring the I ds response for Na + concentration from low to high (10 −6 → 1 m) and high to low (1 → 10 −6 m).As shown in Figure S3 (Supporting Information), I ds was slightly lower when the experiment was started from high concentrations, which is related to a small hysteresis induced in the ISM by the previously immersed analyte.Nevertheless, the difference was minor and can be avoided if sufficient time is allowed for the system to equilibrate.To demonstrate the stability of the device performance, the Na + sensors were tested over 2 weeks time.A ≈6% change in the baseline current was observed, and the devices exhibited stable sensitivity and sensing performance in the range of the linear response (10 −4 to 1 m) as shown in Figure S4 (Supporting Information).
Sodium ions in the blood are crucial for maintaining fluid balance, nerve function, and muscle function.They help to regulate water content in cells, transmit electrical impulses in the nervous system, and regulate muscle contraction.However, too much sodium in the blood can increase the risk of high blood pressure, leading to heart disease, stroke, and other health problems.The concentration of Na + in blood is considered as hyponatremia or hypernatremia when the concentration is below 135 mm or above 145 mm, respectively.A reliable tool to discriminate concentrations near this range is critical.Here, blood serum from human with a Na + concentration in the range of 110-160 mm were used to establish a calibration curve.The serum Na + concentrations were previously measured by ISE (Roche, 9180 Electrolyte Analyzer) to validate our I ds measurements.The testing protocol consists of first measuring the I baseline of the IS-OECT devices in DI water, removing the DI water, and then measuring the I ds in human serum.Figure 3a (red circle) shows the ΔI ds of the IS-OECT sensors and the sensitivity of the blood serum in between 100 and 160 mm is 494 ± 23 A dec −1 .Comparing with the calibration curve based on Na + dissolved in DI water in the same range (black square in Figure 3a), the decay in sensitivity is due to other components in the blood serum including antibodies, antigens, hormones, etc., which would interfere the voltage development in the ISM and introduce more confounders into the analysis.Based on the calibration curve in Figure 3a, we measured the Na + concentration of another ten serum samples and compared with the standard ISE results in Figure 3b.The results matched well and the linear fit curve had a correlation coefficient of 0.93.
To further unlock the potential of the device for real-time ion monitoring inside the human body, a flexible, miniaturized IS-OECT is developed onto implantable neural probe.Recently, our group has demonstrated flexible multichannel neural probes with spike recording and stimulation functions. [29]Since the current IS-OECTs capture specific ion fluxes, these neural probes can be potentially applied for direct brain activity sensing.In the brain, the Na + is involved in different electrophysiological activities.[32] In-vivo monitoring the dynamics of Na + has a major impact on our understanding of the detailed molecular mechanisms of the Na + in maintaining brain functions.Figure 4a shows the geometry of the tip of a probe containing OECTs with different channel width, W/length, L ratios.PEDOT:PSS based probe was fabricated as previously reported. [31]Briefly, the device fabrication involves 1) the deposition and patterning of Au film on parylene C substrate with a 5 nm thick Cr adhesion layer that defines the interconnects, 2) 100 nm thick film of PEDOT:PSS is patterned in islands to yield the channels, 3) a parylene-C film insulates the Au interconnects, allowing the electrolyte to come in contact only with the PEDOT:PSS channels.PSSNa and sodium ISM layers were then deposited on the channel in sequence.Since the probe has narrower Au interconnects width of 10 m, the larger interconnects resistance would lead to a decrease in the maximum I ds , and transconductance.Hence, a larger gate voltage is needed to attain the maximum transconductance (V g (max Gm) ).To partly overcome this issue, we increased the thickness of Au interconnects from 50 to 200 nm.The transfer curve of the device under DI water with W of 50 m and L of 5 m channel is shown in Figure 4b.It exhibited an on/off ratio of 10 3 , max G m of 2.3 mS, and V g (max Gm) of 0.17 V. Figure 4c,d shows the device responses to different Na + concentrations.The current sensitivity is 113 A dec −1 and voltage sensitivity is 54 mV dec −1 for the analyte ranging between 10 −4 and 1 M.These results confirm the potential of the probe for in vivo real-time ion monitoring.

Conclusion
In conclusion, our study demonstrates the successful functionalization of ISM in OECTs for the detection of sodium ions.Our sensor has shown excellent sensitivity (126 A dec −1 , 85 mV dec −1 ), dynamic range, stability, and reversibility compared to the state-of-the-art OECTs.Moreover, the ability of our IS-OECTs to detect Na + in blood serum within the physiological range suggests their potential for point of care applications.Furthermore, our platform can be adapted to sense a range of ionic species by simply modifying the ISM composition with the appropriate ionophore molecule.By advancing device architecture, these sensors can be further miniaturized and integrated onto neural probes, while maintaining their electrical performance and ion sensing capabilities.We anticipate that the integration of IS-OECTs onto neural probes will pave the way for the development of a new generation of multifunctional bioelectronic devices for monitoring brain activities.
Device Fabrication: The OECT devices were fabricated with standard microfabrication techniques.The Au/Cr (50 nm/5 nm) source/drain electrodes were deposited on clean glass substrates with a thermal evaporator (VNANO, VZZ-300S) and patterned by photolithography and lift-off process.Two 1.5 m-thick parylene layers were deposited successively, where the first used as electrical isolation layer and the second used as sacrificial layer.An adhesion promoter, Silane 174 and an antiadhesive, diluted (2% v/v in DI water) Micro 90 were used prior to each layer deposition, respectively.The OECT channels and contact pads were opened via successive photolithography process and reactive ion etching (RIE) steps (CIONE, Femto Science).PEDOT:PSS was then spin-coated on UV-ozone treated substrate through a 0.45 m polytetrafluoroethylene filter.After baking for 1 min at 100 °C, the sacrificial parylene layer was peeled off to define the OECT channels.The devices were, then, hard baked for 30 min at 140°C and soaked in DI water for 2 h to remove the excess of low molecular weight molecules from PEDOT:PSS.The electrolyte reservoir, PSSNa mixture was spin-coated on the PEDOT:PSS channels at 1000 rpm followed by a crosslinking process (1 h 130 °C baking).The ISM solution was finally drop-cast on top of the PSSNa film and dried overnight at room temperature to form the ISM film.Before ISM deposition, devices were soaked in 0.

Figure 1 .
Figure 1.Device structure and electrical characteristics.a) Schematic structure of the IS -OECT.b) Transfer characteristic of IS-OECT measured with 10 −3 M NaCl at V ds = −0.2V. Inset shows the output characteristic performed by sweeping V g from 0 to 0.8 V with a step of 0.16 V. c) Temporal response of I ds in response to pulsed V g of 0.4 V with NaCl concentration of 10 −3 M. Inset shows an exponential decay function fitted line of the curve.d) Transient response of I ds with V g = −0.2,−0.4,−0.6, 0.2, 0.4, 0.6 V in sequence for pulse width of 10 s, and constant V ds = −0.2V for three cyclical pulses.

Figure 2 .
Figure 2. Testing of Na + -selective OECT devices.a) Transfer and transconductance curves measured with DI water at V ds = −0.4V. b) Real-time current response to Na + ion concentrations at fixed V g = 0 V and V ds = −0.4V. c) Calibration curves of the average of three different devices measured current and d) the calculated membrane potential to NaCl solutions with different concentrations (error bars indicate maximum and minimum Selectivity test for the sensor with different interfering ions: e) Real-time current response to K + and Ca 2+ and f) Comparison of current response to Na + , K + and Ca 2+ at different concentrations.

Figure 3 .
Figure 3. a) Calibrate IS-OECTs with Na + aqueous solution (n = 5) and blood serum within 100-160 mm.b) Real sample analysis and comparison with standard ISE.

Figure 4 .
Figure 4. a) Photograph of the tip of the probe.b) Transfer (V ds = 0.4 V) and transconductance of the Na + selective-OECTs probe operated with DI water.c) Calibration curves of the measured current and d) the calculated membrane potential to NaCl solutions with different concentrations.

1 m
NaCl solution overnight to keep Na ions in the film while removing any excess compounds.Characterization: All output, transfer, and ion concentration response measurements were performed with Keithley 2636 SourceMeter using custom LabVIEW software.Temporal responses for OECTs and ISOECTs were recorded with a Keysight B1500 semiconductor parameter analyzer equipped with a fast measurement unit (WGFMU).A small PDMS well was used to confine the analyte solution.The electrical characteristics of OECTs w/o ISM were measured with 10 −3 m NaCl as an electrolyte and Ag/AgCl as gate electrodes.Ion concentration increase/decrease was controlled by successively adding the corresponding solutions /DI water to a fixed volume of DI water in the well.Statistics: Figure 2c,d values are presented with average value with n = 3. Error bars indicate maximum and minimum values.Figure 3a (black line) values are presented with mean ± SD with n = 5.Statistical analysis was done with Origin Pro software (version 8.5).