Because of physically separated sensing region from field-effect region, the extended-gate ion-sensitive field-effect transistor (EG-ISFET) has many advantages over the ISFET including light insensitivity, easy fabrication, and flexibility of sensor array arrangement [1-2]. However, in the EG-ISFET configuration, there is an additional capacitance derived from the passivation layer on the top of extended gate. As a consequence, part of sensing voltages drop across the derived capacitance and the transconductance of EG-ISFET is attenuated. It should be noted that a small transconductance would result in a less current variation and lead to a poor performance in sensitivity. Increasing the capacitance by increasing the two-dimensional area of the top metal layer of the extended gate can ameliorate this issue [3]. However, it will arise other issues such as increasing the fabrication cost at the expense of larger wafer area and decreasing the pixel-resolution of sensor array because of enlargement of each sensing pixel. To address this issue, in this study we propose a lifted three-dimensional (3D) metal architecture as the extended gate for an EG-ISFET. Compared with the conventional EG-ISFET with a planar electrode surface (Fig. 1A), the proposed lifted EG-ISFET, namely the LEG-ISFET, has two metal layers lifted beyond the planar surface (Fig. 1B). Obviously, the LEG-ISFET has larger liquid/solid interface surface area than the conventional EG-ISFET does. In other words, the LEG-ISFET should have a better transconductance than the conventional EG-ISFET does. At the same time, the LEG-ISFET keeps the same wafer area as the conventional EG-ISFET. To fabricate the proposed LEG-ISFET, conventional EG-ISFETs and LEG-ISFETs with different size of top metal layer are implemented on the same chip by a standard CMOS 0.35 μm process. Then the fabricated chip undergoes wet etching to remove passivation-free metal/via layers including the oxides surrounded by them to realize the lifted 3D metal structure as shown in Fig. 2. In experiments, the ID-VG curve of each FET device is measured using Ag/AgCl reference electrode in a pH7 buffer solution. Based on the slope of the measured curves, the transconductance can be extracted. To facilitate the comparison, Table 1 shows different symbols representing different device designs. Based on the experiment, it shows that the transconductance attenuates seriously in the conventional EG-ISFET with the smallest top metal area (Fig. 3A). Compared with the MOSFET with exactly the same structure, the transconductance measured from conventional EG-ISFET is about 3 times lower than that of the MOSFET with the same structure. As the top metal area increased, the transconductance of the conventional EG-ISFET is gradually increased. On the other hand, the transconductance of each proposed LEG-ISFET is at the same level as the intrinsic MOSFET (Fig. 3B). It should be noted that the smallest LEG-ISFET has a 3D electrode surface area estimated about 3 times to the 2D top-electrode area of the smallest EG-ISFET. However, the transconductance improvement of the smallest LEG-ISFET is roughly the same as the largest EG-ISFET, which has about 600 times the top-electrode area of the smallest EG-ISFET. This finding suggests that factors other than surface area should also account for the transconductance enhancement. It is interesting to investigate the underlying mechanism to further improve the performance of ISFET. In conclusion, a high-transconductance configuration of LEG-ISFET is proposed and verified. This development has potentials for realization of sensor array with high pixel-resolution and high sensitivity. [1] Kaisti, Matti. "Detection principles of biological and chemical FET sensors." Biosensors and Bioelectronics 98 (2017): 437-448. [2] Chi, Li-Lun, et al. "Study on extended gate field effect transistor with tin oxide sensing membrane." Materials Chemistry and Physics 63.1 (2000): 19-23. [3] Sohbati, Mohammadreza, and Christofer Toumazou. "Dimension and shape effects on the ISFET performance." IEEE Sensors Journal 15.3 (2015): 1670-1679. Figure 1
Read full abstract