Selective ion sensing with high resolution large area graphene field effect transistor arrays

Accurate, rapid, and reliable ion concentration measurement at low concentrations is crucial across diverse applications, including medical diagnostics, environmental monitoring, and industrial process control. Potentiometric ion sensors, particularly solid-state sensors like ISFETs, offer advantages over chromatographic and spectrophotometric techniques due to their compactness, ease of integration, and real-time measurement capabilities. However, traditional ISFETs using ion-selective membranes suffer from limited selectivity, as ions other than the target ion can bind to the membrane, leading to unreliable measurements in multi-ionic analytes. This limitation can be addressed by using ISFET arrays targeting various ionic species, including known interferents, and employing Nikolskii-Eisenman analysis for concentration estimation. While previous ion-sensitive arrays have been demonstrated, they often fall short in achieving the resolution and detection limits needed for many real-time sensing applications. Silicon-based ISFETs, commonly fabricated using CMOS processes, are typically micro-scaled, limiting resolution due to low-frequency charge fluctuations. Efforts to improve resolution have involved exploring alternative materials and structures, but these have faced challenges in achieving the desired large-area arrays while maintaining high resolution. Graphene, with its large-area fabrication potential and high charge carrier mobility, is a promising material for overcoming these limitations. While early graphene ISFETs showed modest performance, recent advances have demonstrated high sensitivity and resolution for H+ and K+ measurement. This paper reports the development of a high-resolution, multi-ion sensing array based on large-area graphene ISFETs to address the challenges of ion interference and achieve real-time, simultaneous measurements of multiple ionic species important in agricultural runoff and water quality monitoring.

The paper reviews existing methods for ion concentration measurement, highlighting the limitations of chromatographic, spectrophotometric, and potentiometric techniques. It discusses the advantages and disadvantages of potentiometric ion sensors, particularly ISFETs, emphasizing the issue of cross-sensitivity and lack of selectivity in traditional designs. A table comparing various ISFET channel materials (Si, Si nanowires, Si FET + BJT, AlGaN/GaN HEMTs, MoS2, InN, carbon nanotubes, and graphene) and their respective sensitivities and resolutions is presented. This comparison highlights the superior performance of graphene ISFETs in terms of achievable resolution and detection limits. The literature also shows the use of Nikolskii-Eisenman analysis to overcome cross-sensitivity issues in potentiometric arrays, but prior work fell short of the desired resolution and detection limits for many applications. The authors also cite previous work on graphene ISFETs, showcasing their enhanced capabilities for improved resolution compared to existing materials. The selection of specific ions (K+, Na+, NH4+, NO3-, SO42-, HPO42-, and Cl-) is justified by their significance in agricultural runoff and the need for accurate water quality monitoring.

The researchers fabricated large-area graphene ISFETs (~cm²) using wafer-scale processing. A 100 mm diameter graphene monolayer, grown via chemical vapor deposition (CVD), was wet-transferred onto a fused silica wafer with a parylene C layer. Ti/Au contacts were evaporated to form source and drain contacts. The wafer was diced into individual devices, which were mounted onto printed circuit boards (PCBs). Ionophore membranes selective for K+, Na+, NH4+, NO3-, SO42-, HPO42-, and Cl- were then drop-cast onto the graphene surface and encapsulated with epoxy to create the ISFET array. Raman spectroscopy and Hall measurements were conducted to characterize the quality and properties of the graphene. The ISFETs were characterized individually by immersion in electrolytic solutions of controlled concentrations. The drain-source current (Ids) was measured versus electrolytic gate potential (Vref) at a constant drain-source bias voltage. The response of each ISFET to its target ion was analyzed to determine sensitivity using both Vnp (potential at minimum conductance) and Ids (at constant Vref) measurements. A parabolic fit was used to extract Vnp values from the transfer curves, and linear fits were used to determine sensitivity from Ids versus concentration plots. The impact of temperature and pH on ISFET sensitivity was investigated. Cross-sensitivity was assessed using the separate solution method, whereby individual ISFET responses to different ions were measured. The Nikolskii selectivity coefficients were extracted using linear fits of ISFET current versus ion concentration plots. The accuracy of these coefficients was confirmed via the mixed solution fixed interference method. The complete ISFET array was tested with multi-ion solutions, and ion concentrations were calculated from the measured currents using a modified Nikolskii-Eisenman equation, which accounts for current sensitivities in the presence of interfering ions. Finally, the array was used to monitor ion uptake by duckweed in an aquarium over three weeks.

The fabricated large-area graphene ISFETs achieved high resolution and accuracy in ion concentration measurements. The ISFETs exhibited sensitivities approaching the Nernstan limit for several ions, and linear response was observed down to concentrations below 10⁻⁵ M. The resolution for cation ISFETs was approximately 3 x 10⁻³ log concentration units, and slightly higher for anions (2 x 10⁻² log concentration units). The Nikolskii-Eisenman analysis successfully accounted for ion interference, with cation concentrations estimated to within ±0.01 log concentration units and anion concentrations within ±0.05 log concentration units. The separate solution method provided an efficient way to calibrate the ISFET array and determine selectivity coefficients, which were validated using the mixed solution method. The real-time, simultaneous measurement of seven ions (Na+, K+, NH4+, NO3-, SO42-, HPO42-, Cl-) in multi-ion solutions demonstrated the array's capacity for accurate and reliable analysis. The experiment with duckweed in the aquarium showcased the applicability of the array in monitoring dynamic ion concentrations in a complex environment, observing a significant reduction in ion levels over three weeks due to nutrient uptake by the plants.

The study successfully demonstrated a high-performance, multi-ion sensing array based on large-area graphene ISFETs. The use of graphene provided significant advantages in achieving high resolution and sensitivity, overcoming the limitations of traditional silicon-based ISFETs. The application of the Nikolskii-Eisenman analysis with a modified equation suitable for current measurements enabled accurate estimation of multiple ion concentrations despite cross-sensitivity. The demonstrated accuracy and resolution significantly exceed the capabilities of many existing techniques. The work highlights the potential of graphene ISFET arrays for real-time monitoring of ion concentrations in various applications. The aquarium experiment confirmed the utility of this technology for in situ monitoring of environmental systems.

This research successfully demonstrated a high-resolution, multi-ion sensing array using large-area graphene ISFETs. The array achieves high accuracy and selectivity through a modified Nikolskii-Eisenman analysis, enabling real-time, simultaneous measurement of seven ions. The simple calibration method and demonstrated performance in a complex environment make this approach suitable for various applications, particularly real-time water quality monitoring. Future research could focus on expanding the range of detectable ions, further optimizing the ionophore membranes for improved selectivity, and developing miniaturized, portable versions of this technology.

The study acknowledges limitations in the selectivity of certain ionophore membranes, leading to slightly lower accuracy in anion concentration estimation compared to cations. Improved ionophore membrane preparation could potentially reduce the variation in doping levels and further enhance performance. The study also notes that changes in ISFET performance require recalibration of selectivity coefficients. The current setup may also require further miniaturization for certain practical applications.