Article | Open | Published:

# Modelling and Realization of a Water-Gated Field Effect Transistor (WG-FET) Using 16-nm-Thick Mono-Si Film

Scientific Reportsvolume 7, Article number: 12190 (2017) | Download Citation

## Abstract

We introduced a novel water-gated field effect transistor (WG-FET) which uses 16-nm-thick mono-Si film as active layer. WG-FET devices use electrical double layer (EDL) as gate insulator and operate under 1 V without causing any electrochemical reactions. Performance parameters based on voltage distribution on EDL are extracted and current-voltage relations are modelled. Both probe- and planar-gate WG-FETs with insulated and uninsulated source-drain electrodes are simulated, fabricated and tested. Best on/off ratios are measured for probe-gate devices as 23,000 A/A and 85,000 A/A with insulated and uninsulated source-drain electrodes, respectively. Planar-gate devices with source-drain insulation had inferior on/off ratio of 1,100 A/A with 600 μm gate distance and it decreased to 45 A/A when gate distance is increased to 3000 μm. Without source-drain electrode insulation, proper transistor operation is not obtained with planar-gate devices. All measurement results were in agreement with theoretical models. WG-FET is a promising device platform for microfluidic applications where sensors and read-out circuits can be integrated at transistor level.

## Introduction

By exploiting electrical double layer (EDL) concept and using electrical double layer capacitance (EDLC) as gate insulator, high channel control can be achieved at low gate voltages in field effect transistors (FETs)1,2. This provides a cheap solution to technologically and economically challenging problem of realizing an insulating layer as thin as possible without variations and pinholes. This phenomenon is explained by the combined Gouy-Chapman-Stern model3. Ionic liquids, ion-gels, aqueous and solid electrolytes are used to form EDL in these devices. They can use metal oxide semiconductors4,5,6, graphene7,8,9, carbon nanotubes10,11,12, organic semiconductors13,14,15 or Si16,17,18 as active channel layer.

Water as electrolyte material was used in organic field-effect transistors (OFETs) with probe-gate before19. Our group contributed to this concept by realizing a water-gated OFET with a planar-gate structure20. Planar-gate topology allows patterning of source, drain and gate electrodes on the same layer with a single photolithography step. It reduces fabrication process complexity and provides easy integration with fluidic channels. However, organic semiconductors are prone to ion diffusion which results in electrochemical doping21. Low charge carrier mobility and degradation due to environmental instability are other drawbacks of OFET devices.

In this paper, we present the realization and modelling of a water-gated field effect transistor (WG-FET) which uses 16-nm-thick mono-Si film as channel layer. A sketch of the WG-FET device is given in Fig. 1a. It combines the advantages of planar electrolyte-gated OFET design with the high performance of single crystalline Si layer.

Fluidic interfaces of these devices provide an integration platform for sensors and their read-out circuits at transistor level. Ultra-thin and high mobility characteristics of the channel layer provide better surface sensitivity in sensor applications and in-situ amplification in their read-out circuits. Two-dimensional electron gas (2DEG) feature of 16-nm-thick Si layer is exploited to obtain both in the same device22,23. Si nanowire FETs24,25 use similar concepts, however they use Si layers thicker than 25 nm for channel region and have more complicated fabrication steps. Label-free electronic sensors utilizing 30-nm-thick single crystalline Si layers have been shown as promising sensors26 but their usage as transistor or read-out circuit are not considered. Ion-sensitive FETs (ISFETs)27,28,29 resemble to these devices, however, they are bulk Si devices operating in depletion/inversion modes with large reference electrodes, and SiO2, Si3N4, Al2O3, Ta2O5 or other types of insulation layers. WG-FETs require simpler fabrication steps compared to these devices. Furthermore, neither planar and probe gate electrode topologies nor the effect of source and drain electrode insulation have been investigated and modelled before.

Working principle of WG-FET device is summarized for probe- and planar-gate topologies in Fig. 1b,c, respectively. Thin Si layer is moderately doped with Boron (~1015 cm−3). The WG-FET presented here is an accumulation type device. EDLs are formed on both gate electrode surface and Si active area. Positive charge carriers are attracted to the interface surface in Si layer with application of negative VGS to the gate electrode. This turns on the transistor. For a thin Si film based accumulation mode p-channel MOSFET, expression of channel current for linear regime (V DS > V GS − V FB) is given as

$${I}_{{\rm{acc}}}=\frac{W}{L}{\mu }_{s}{C}_{{\rm{ox}}}[({V}_{{\rm{GS}}}-{V}_{{\rm{FB}}}){V}_{{\rm{DS}}}-\frac{{V}_{{\rm{DS}}}^{2}}{2}]$$
(1)

where μ s is the hole surface mobility, C ox is the oxide capacitance, V GS is the applied gate voltage and V DS is the drain-source voltage difference. V FB is the flat-band voltage and it represents the threshold voltage for accumulation mode transistor since inversion does not occur30. The equation becomes

$${I}_{{\rm{acc}}}=\frac{W}{L}{\mu }_{s}\frac{{C}_{{\rm{ox}}}}{2}{({V}_{{\rm{GS}}}-{V}_{{\rm{FB}}})}^{2}$$
(2)

in saturation regime (V DS ≤ V GS − V FB).

In these equations, it is assumed that potential applied on the insulator layer throughout the channel is uniform and equal to V GS which is the case for traditional MOSFET architecture. In WG-FET topology, effective V GS is the voltage on the EDL which is formed on top of the active Si area. The value of this potential on any arbitrary point along the length of the device, V g_EDL(x), is a result of combined effects of source, drain and gate electrodes. Insulation on source and drain electrodes greatly reduces their effects on V g_EDL by introducing extra serial capacitances with low values. Whether being insulated or not, their effects should be taken into account in calculating the channel current to obtain a more realistic model. Therefore, if the channel length is L, voltage of a point on EDL insulation layer can be written in most general form as

$${V}_{g\_\mathrm{EDL}}(x)={k}_{1}{V}_{{\rm{GS}}}+{k}_{2}{V}_{{\rm{DS}}}+({k}_{3}{V}_{{\rm{DS}}}+{k}_{4}{V}_{{\rm{GS}}})\frac{x}{L}$$
(3)

where x is from 0 to L. In probe-gate setup, gate electrode is placed in the middle of the channel, on top of the active area. Therefore, its effect is symmetric throughout the channel length. In planar-gate setup, gate electrode is designed surrounding the active area. Symmetric design ensures equal gate-source and gate-drain distances as in Fig. 1a. These symmetric effects of the gate electrodes result in an approximately uniform potential distribution throughout the channel, so the contribution of V GS in V g_EDL should not depend on x which makes k 4 = 0.

To obtain an expression for the channel current, we can use the equation

$$I{\int }_{0}^{L}{\rm{d}}x=W{\mu }_{s}{C}_{{\rm{EDL}}}{\int }_{V\mathrm{(0)}}^{V(L)}[{V}_{g\_\mathrm{EDL}}(x)-{V}_{{\rm{thc}}}-V(x)]{\rm{d}}V$$
(4)

where C EDL is the capacitance of EDL. V thc stands for a threshold constant. It represents the cumulative effects of trapped charges at the Si/water interface with the flat-band voltage needed to form the channel in accumulation mode. If we use gradual channel approximation as $$V(x)=({V}_{{\rm{DS}}}/L)x$$ and substitute dV with $$({V}_{{\rm{DS}}}/L){\rm{d}}x$$, the channel current expression becomes

$$I{\int }_{0}^{L}{\rm{d}}x=W{\mu }_{s}{C}_{{\rm{EDL}}}{\int }_{0}^{L}[{V}_{g\_\mathrm{EDL}}(x)-{V}_{{\rm{thc}}}-\frac{{V}_{{\rm{DS}}}}{L}x]\frac{{V}_{{\rm{DS}}}}{L}{\rm{d}}x\mathrm{.}$$
(5)

If V g_EDL (x) is replaced with the corresponding voltage distribution function, channel current can be found as

$$I=\frac{W}{L}{\mu }_{s}{C}_{{\rm{EDL}}}[({k}_{1}{V}_{{\rm{GS}}}+{k}_{2}{V}_{{\rm{DS}}}-{V}_{{\rm{thc}}}){V}_{{\rm{DS}}}-\frac{1-{k}_{3}}{2}{V}_{{\rm{DS}}}^{2}]$$
(6)

and in saturation it becomes

$${I}_{{\rm{sat}}}=\frac{({k}_{3}+1)W}{2L}{\mu }_{s}{C}_{{\rm{EDL}}}{({k}_{1}{V}_{{\rm{GS}}}+{k}_{2}{V}_{{\rm{DS}}}-{V}_{{\rm{thc}}})}^{2}\mathrm{.}$$
(7)

In I sat expression of equation (7), k 1 represents the gating capability of the applied V GS. k 2 models the effect of drain electrode voltage on gating. When a negative V DS is applied, it affects the transistor as a competing gate electrode. Therefore, it is desirable to have higher k 1 and lower k 2 values. k 3 depends on both the effect of drain electrode and channel length. It acts as a common multiplier.

## Results

### Simulation Results

To verify the V g_EDL (x) function and extract parameters, electric field and voltage distribution simulations of WG-FET devices with probe- and planar-gate setups are performed with COMSOL Multiphysics 5.2. For planar-gate setup, two symmetric gate electrodes are placed with a specific distance away from the source and drain electrodes. Two different layouts with gate electrode distances of 600 μm and 3000 μm are used to see the effect of planar gate electrode position on V g_EDL (x). For each topology, versions with and without source-drain electrode insulation are simulated.

Device model used in simulations for WG-FETs with source-drain electrode insulation is given in Fig. 2a. Due to alignment errors, insulating layer, photoresist (PR) in this case, overlaps with channel area which creates covered regions. These regions cannot be controlled by gate, therefore they behave like series parasitic resistances, R px, to the transistor. Value of R px depends on fabrication process and can vary from sample to sample. In the case of devices with insulated source-drain electrodes, these parasitic effects are simulated using HSPICE as explained in the next section.

In simulations, voltage distribution on EDL is analyzed for different V GS and V DS values. Fig. 2b shows the simulation results for devices with source-drain electrode insulation. It can be seen that combined effects of drain and gate electrodes result in an uneven voltage distribution on EDL throughout the channel. In planar-gate setup with the closer gate electrode layout, effect of gate electrode on voltage distribution is lower compared to the probe-gate setup and it decreases further when the planar gate electrode is placed in 3000 μm distance.

In devices without source-drain electrode insulation, source and drain electrodes are directly in contact with the water droplet like the gate electrode. Due to absence of insulator overlap with channel area, series parasitic resistance R px does not exist, which is desirable for proper transistor operation. For these devices, difference between simulation results of probe- and planar-gate setups is more significant. In planar-gate setups, effect of gate electrode on voltage distribution is drastically lower and V g_EDL curves dissociate more for different V DS values, compared to probe-gate setup, as seen in Fig. 2c.

The effective voltage on the EDL insulation layer for the WG-FET device is not uniform throughout the channel when V DS is not equal to 0 V. Instead, it is a function of the channel position as V g_EDL (x) as indicated in equation (3). By making linear approximations in simulation results, parameters k 1, k 2, and k 3 are calculated for all WG-FET topologies as given in Table 1.

In probe-gate setup with source-drain electrode insulation, approximately 51.5 % of the applied V GS and 1.7 % of the applied V DS are transferred as effective gate voltage. Planar-gate setup with 600 μm gate distance and source-drain electrode insulation, has inferior k 1 and k 2 values relative to probe-gate setup. When the gate electrode distance is increased to 3000 μm for the same setup, k 1 decreases to 0.315 while k 2 increases to 0.032, which indicates that the effect of gate electrode weakens with increasing planar gate electrode distance. For devices without source-drain electrode insulation, probe-gate setup has a comparable k 1 value with a slightly elevated k 2 value with respect to devices with source-drain electrode insulation. On the other hand, planar-gate setups have considerably poor k 1 and k 2 values. For planar-gate setup with 600 μm gate distance, k 1 value is halved whereas k 2 value is doubled compared to probe-gate setup. When the gate distance is increased to 3000 μm, they deteriorate further. For planar-gate setups without source-drain electrode insulation, k 1 and k 2 values are comparable as seen in Table 1. This implies that the effect of gate electrode voltage is no longer dominant on V g_EDL and the effect of drain electrode gets significant which is not desired for a proper transistor operation. This points to the significance of source-drain electrode insulation on planar gate devices.

### Experimental Results

All WG-FET topologies given in Table 1 are fabricated for electrical measurements. Each topology is also simulated with HSPICE by using equations (6) and (7) with corresponding k 1, k 2, and k 3 values for comparison. For devices with source-drain electrode insulation, series parasitic resistances due to insulator overlapping, R px, are added in HSPICE models. In Fig. 3a,b, I DSV DS and I DSV GS measurement results are given for probe-gate setup with source-drain electrode insulation. I ON and I OFF currents are measured as 184 μA and 8 nA, respectively, which gives an on/off ratio of 23,000 A/A. Transfer curve measurements for V DS = −1 V (Fig. 3b) point to an effective threshold voltage of −0.46 V. This corresponds to V thc value of −0.26 V. Both I DSV DS and I DSV GS measurement results are parallel with simulations. In experiments, no variation in transistor characteristics was noted due to vertical distance of probe gate electrode as long as it stayed in the water droplet. Also, lateral displacement of probe was tested up to 3 mm and no effect was noticed on measurement results. For planar-gate setup with 600 μm gate distance and source-drain electrode insulation, on/off ratio is found as 1,100 A/A. The effective threshold voltage is calculated as −0.56 V, which results in V thc value of −0.27 V. Decrease in on/off ratio is expected due to inferior k 1 and k 2 values. Simulations are in agreement with measurement results as seen in Fig. 3c. Similarly, for planar-gate setup with 3000 μm gate distance and source-drain electrode insulation, on/off ratio is found as 45 A/A and effective threshold voltage is calculated as −0.63 V. For this measurement, V thc is found as −0.13 V. Increasing gate electrode distance from 600 μm to 3000 μm, decreases on/off ratio further. Theoretical and experimental results are given in Fig. 3d.

In Fig. 4a, I DSV DS measurement results are given for probe-gate setup without source-drain electrode insulation. Inferior k 1 and k 2 values compared to its counterpart with source-drain electrode insulation suggest worse transistor performance. However, I ON and I OFF currents for this setup are measured as 596 μA and 7 nA, respectively, which gives an on/off ratio of 85,000 A/A. This can be explained due to the absence of R px. When there is no series parasitic resistance due to insulator overlapping, I ON is boosted which significantly increases the on/off ratio. The effective threshold voltage is calculated as −0.51 V, which results in V thc value of −0.26 V. Simulation results are also in agreement with measurements as in Fig. 4a. For planar-gate setup without source-drain electrode insulation, I DSV DS measurement results are given In Fig. 4b. Gate electrode distance is 3000 μm in this layout. Again, absence of R px manifests itself with high current levels like in the probe-gate layout. However, deteriorated transistor operation can be seen in this graph as expected from the comparable k 1 and k 2 values. It is hard to mention about a proper threshold voltage or on/off ratio about the planar-gate setup without source-drain electrode insulation. Theoretical simulations demonstrate similar behaviour and support the measurement results as seen in Fig. 4b.

These results show that the best on/off ratio is obtained with probe-gate setup for devices with both insulated and uninsulated source-drain electrodes as expected from V g_EDL simulations. Higher k 1 value indicates that higher effective gate voltage can be obtained on the device. In combination with low k 2, a high on/off ratio is expected. For planar-gate topology, the setup with smaller gate distance gives a higher on/off ratio relative to the one with further gate electrode. When the gate distance was increased from 600 μm to 3000 μm, on/off ratio decreased drastically. These findings are in agreement with the device parameters given in Table 1 and supported by the HSPICE simulations. Effects of source-drain electrode insulation are also tested. For probe-gate setup, transistor operation is affected slightly due to uninsulated source-drain electrodes, since it results in a little lower k 1 and a little higher k 2 values as shown in Table 1. However, they result in higher on/off ratio compared to the devices with insulated source-drain electrodes, because these devices do not have series parasitic resistances, R px. On the other hand, for planar-gate setup without source-drain electrode insulation, I DSV DS measurement results show considerable deviations from proper transistor operation due to increasing effect of drain voltage on V g_EDL. Therefore, insulation of source-drain regions is crucial for WG-FET devices with planar-gate topology.

## Discussion

A novel WG-FET device using 16-nm-thick mono-Si film is presented. Devices with probe- and planar-gate topologies, and with and without source-drain electrode insulation are modelled and simulated. Device parameters, k 1 and k 2, model the gating effect of the gate and drain electrodes, respectively. Parameter k 3 has both drain voltage and channel length contributions. As these parameters are extracted from electric field and voltage distribution simulations, current-voltage relations of the transistors are obtained. All modelled and simulated devices are fabricated and tested. Current levels, on/off ratios, and effective threshold voltage values are extracted from measurements and compared with theoretical calculations, verifying them. The threshold constant parameter, V thc, is found approximately as −0.26 V from the measurements.

Best on/off ratios are obtained with probe-gate setups. 23,000 A/A and 85,000 A/A are measured for devices with insulated and uninsulated source-drain electrodes, respectively. Removing insulation layers on source-drain electrodes in probe-gate devices improves overall device performance even though it results in inferior k 1 and k 2 values. This is because it eliminates a more important factor, the serial resistances under the insulated regions as a result of inevitable alignment errors. In probe-gate devices, lateral or vertical displacement of gate electrode does not result in a significant variation in transistor operation.

Gate electrode distance and source-drain insulation are very critical in planar-gate devices. Their already low on/off ratio decreases from 1,100 A/A to 45 A/A when gate distance is increased from 600 µm to 3000 μm in devices with source-drain electrode insulation. Planar-gate devices with uninsulated source-drain electrodes do not have any proper transistor characteristics. Therefore, insulation of source-drain electrodes is essential for planar-gate setups.

WG-FET is a promising device especially for microfluidic applications. Si/water interface, transistor behaviour and easy fabrication make it suitable to implement and integrate sensors and read-out circuits together.

## Methods

### WG-FET Fabrication

WG-FET devices with probe- and planar-gate topologies are fabricated with a three-mask photolithographic process. First, 16-nm-thick single crystalline Si layer is patterned as active layer on top of 145-nm-thick buried oxide layer. Trilogy etch (126 HNO3:60 H2O:5 NH4F) is used to pattern Si after a lithographic step. Then, Al layer is thermally evaporated on top with approximately 200 nm thickness. After deposition of Al, a second photolithographic step and wet etch are used to pattern source, drain and planar gate electrodes. To obtain ohmic contacts between Al electrodes and Si layer, device samples are thermally annealed at 475 °C for 5 minutes. Annealing process is done under continuous nitrogen flow to prevent any unwanted surface deposition.

A PR layer with 4 μm thickness is used for both source-drain and field oxide insulation. Thermal SiO2 thin films are known for having charge traps on interface surface when in contact with water molecules31, therefore it is important to insulate field oxide regions to avoid undesired effects of trapped surface charges around the Si active area, and obtain more repeatable results (see Supplementary Material).

Electrical contacts are established with source, drain and planar gate electrodes using silver epoxy. Then, a de-ionized water droplet is placed on top covering both planar gate electrode and the Si active region to complete the WG-FET device. Fabrication steps are summarized in Fig. 5a. A micrograph of a fabricated sample device and a picture of the experimental setup are given in Fig. 5b,c, respectively.

### Experimental Setup

Current-voltage measurements are performed with probe- and planar-gate setups. Probe-gate setup is formed by immersing an external probe into the water droplet just on top of the active Si area. For planar-gate experiments, layouts with electrode distances of 600 μm and 3000 μm are used to examine the effect of gate distance on transistor performance. All devices are tested with and without source-drain electrode insulation for comparison. All measurements are carried out with Keithley 4200SCS semiconductor characterization system. Applied voltage is limited to −1.0 V to prevent electrochemical reaction at electrode/water interfaces.

### Data Availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

## Additional information

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## References

1. 1.

Fujimoto, T. & Awaga, K. Electric-double-layer field-effect transistors with ionic liquids. Phys. Chem. Chem. Phys. 15, 8983–9006 (2013).

2. 2.

Du, H., Lin, X., Xu, Z. & Chu, D. Electric double-layer transistors: a review of recent progress. J. Mater. Sci. 50, 5641–5673 (2015).

3. 3.

Schoch, R. B., Han, J. & Renaud, P. Transport phenomena in nanofluidics. Rev. Mod. Phys. 80, 839–883 (2008).

4. 4.

Misra, R., McCarthy, M. & Hebard, F. Electric field gating with ionic liquids. Appl. Phys. Lett. 90(052905), 1–3 (2007).

5. 5.

Yuan, H. et al. High-density carrier accumulation in ZnO field-effect transistors gated by electric double layers of ionic liquids. Adv. Funct. Mater. 19, 1046–1053 (2009).

6. 6.

Althagafi, T. M., Algarni, S. A., Naim, A. A., Mazher, J. & Grell, M. Precursor-route ZnO films from a mixed casting solvent for high performance aqueous electrolyte-gated transistors. Phys. Chem. Chem. Phys. 17, 31247–31252 (2015).

7. 7.

Inaba, A., Yoo, G., Takei, Y., Matsumoto, K. & Shimoyama, I. A graphene FET gas sensor gated by ionic liquid. Proc. IEEE 26th Int. Conf. MEMS 969–972 (2013).

8. 8.

Inaba, A., Takei, Y., Matsumoto, K. & Shimoyama, I. Ionic liquid-gated graphene FET array with enhanced selectivity for electronic nose. Proc. IEEE 27th Int. Conf. MEMS 326–329 (2014).

9. 9.

Brown, M. A., Crosser, M. S., Leyden, M. R., Qi, Y. & Minot, E. D. Measurement of high carrier mobility in graphene in an aqueous electrolyte environment. Appl. Phys. Lett. 109(093104), 1–4 (2016).

10. 10.

Ozel, T., Gaur, A., Rogers, J. A. & Shim, M. Polymer electrolyte gating of carbon nanotube network transistors. Nano Lett. 5, 905–911 (2005).

11. 11.

Kiga, N. et al. CNT-FET gas sensor using a functionalized ionic liquid as gate. Proc. IEEE 25th Int. Conf. MEMS 796–799 (2012).

12. 12.

Melzer, K. et al. Flexible electrolyte-gated ion-selective sensors based on carbon nanotube networks. IEEE Sens. J. 15, 3127–3134 (2015).

13. 13.

Panzer, M. J. & Frisbie, C. D. Exploiting ionic coupling in electronic devices: Electrolyte-gated organic field-effect transistors. Adv. Mater. 20, 3177–3180 (2008).

14. 14.

Popescu, D., Popescu, B., Brandlein, M., Melzer, K. & Lugli, P. Modeling of electrolyte-gated organic thin-film transistors for sensing applications. IEEE Trans. Electron Devices 62, 4206–4212 (2015).

15. 15.

White, S. P., Dorfman, K. D. & Frisbie, C. D. Operating and sensing mechanism of electrolyte-gated transistors with floating gates: Building a platform for amplified biodetection. J. Phys. Chem. C 120, 108–117 (2016).

16. 16.

Nikolaides, M. G. et al. Silicon-on-insulator based thin-film resistor for chemical and biological sensor applications. Chem Phys Chem 4, 1104–1106 (2003).

17. 17.

Knopfmacher, O. et al. Nernst limit in dual-gated Si-nanowire FET sensors. Nano Lett. 10, 2268–2274 (2010).

18. 18.

Kaisti, M. et al. An ion-sensitive floating gate FET model: Operating principles and electrofluidic gating. IEEE Trans. Electron Devices 62, 2628–2635 (2015).

19. 19.

Kergoat, L. et al. A water-gate organic field-effect transistor. Adv. Mater. 22, 2565–2569 (2010).

20. 20.

Yaman, B., Terkesli, I., Turksoy, K. M., Sanyal, A. & Mutlu, S. Fabrication of a planar water gated organic field effect transistor using a hydrophilic polythiophene for improved digital inverter performance. Org. Electron. 15, 646–653 (2014).

21. 21.

Kergoat, L., Piro, B., Berggren, M., Horowitz, G. & Pham, M.-C. Advantages in organic transistor-based biosensors: From organic electrochemical transistors to electrolyte-gated organic-field-effect transistors. Anal. Bioanal. Chem. 402, 1813–1826 (2012).

22. 22.

Ertop, O., Sonmez, B. G. & Mutlu, S. Realization of a planar water-gated field effect transistor (WG-FET) using 16-nm-thick single crystalline Si film. Procedia Eng. 87, 76–79 (2014).

23. 23.

Sonmez, B. G., Ertop, O. & Mutlu, S. Improved repeatability in planar water-gated field effect transistor (WG-FET) with 16-nm-thick single crystalline Si film. Procedia Eng. 168, 1739–1742 (2016).

24. 24.

Chen, Y., Wang, X., Erramilli, S., Mohanty, P. & Kalinowski, A. Silicon-based nanoelectronic field-effect pH sensor with local gate control. Appl. Phys. Lett. 89(223512), 1–3 (2006).

25. 25.

Stern, E., Vacic, A. & Reed, M. A. Semiconducting nanowire field-effect transistor biomolecular sensors. IEEE Trans. Electron Devices 55, 3119–3130 (2008).

26. 26.

Stern, E. et al. Label-free immunodetection with CMOS-compatible semiconducting nanowires. Nature 445, 519–522 (2007).

27. 27.

Bergveld, P. Thirty years of ISFETOLOGY: What happened in the past 30 years and what may happen in the next 30 years. Sensor Actuat. B-Chem. 88, 1–20 (2003).

28. 28.

Chakrabarti, S. et al. Negative voltage modulated multi-level resistive switching by using a Cr/BaTiO x /TiN structure and quantum conductance through evidence of H2O2 sensing mechanism. Sci. Rep. 7(4735), 1–13 (2017).

29. 29.

Kumar, P. et al. Cross-point resistive switching memory and urea sensing by using annealed GdO x film in IrO x /GdO x /W structure for biomedical applications. J. Electrochem. Soc. 164, B127–B135 (2017).

30. 30.

Colinge, J.-P. Conduction mechanisms in thin-film accumulation-mode SOI p-channel MOSFET’s. IEEE Trans. Electron Devices 37, 718–723 (1990).

31. 31.

Nicollian, E. H., Berglung, C. N., Schmidt, P. F. & Andrews, J. M. Electrochemical charging of thermal SiO2 films by injected electron currents. J. Appl. Phys. 42, 5654–5664 (1971).

Download references

## Acknowledgements

This work is supported by Scientific and Technological Research Council of Turkey (TUBITAK) under project EEEAG 114R080.

## Author information

### Affiliations

1. #### Department of Electrical and Electronics Engineering, Bogazici University, Istanbul, 34342, Turkey

• Bedri Gurkan Sonmez
• , Ozan Ertop
•  & Senol Mutlu

### Contributions

B.G.S. wrote the main manuscript, B.G.S. and O.E. conceived and conducted the experiments, S.M. planned and supervised the project. All authors reviewed the manuscript.

### Competing Interests

The authors declare that they have no competing interests.

### Corresponding author

Correspondence to Senol Mutlu.

## About this article

### DOI

https://doi.org/10.1038/s41598-017-12439-8

## Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.