Solvent sensors based on amorphous ZnSnO thin-film transistors

Abstract

Amorphous zinc–tin oxide (a-ZTO) thin-film transistors (TFTs) were prepared using a combustion solution method. The properties of the a-ZTO films and a-ZTO TFTs were studied in detail. For applications, a-ZTO TFTs are used as solvent sensors. The a-ZTO TFTs exhibited strong sensitivity and selectivity in detecting solvents (e.g., cyclohexane, ethanol, and deionized water). The electron donors in the ZTO channels were determinated by the polarity of the solvent, which affects the location of the Femi level (E_f). Moreover, a feasible mechanism model related to the electron transfer channel (ETC) was proposed to explain the sensor behaviours. This model can be applied to most of the amorphous oxide TFT sensors. This work is expected to not only provide an insight into the fundamental understanding of the behaviours of amorphous oxide TFT biosensors, but also to offer a basic design guideline for device fabrication in this system.

---

Recently, amorphous oxide semiconductor (AOS) thin-film transistors (TFTs) have attracted great interest for flat panel and next-generation displays.^1–7 In particular, ZnInO (ZIO) based AOSs have been extensively investigated in the past few years.^8,9 The full-scale production of displays using InGaZnO TFTs has been carried out by manufacturers.^10–13 However, the toxicity, limited storage, and high cost of In prevent long-term application in all the fields. Therefore, indium-free AOSs are urgently pursued to meet the requirements. Amorphous ZnSnO (a-ZTO) emerges as one of the most promising indium-free alternatives, which has received more and more attention.^14,15

More recently, several reports have appeared for DNA detections using ZIO-based TFT biosensors.¹³ These reports promote the developments of oxide TFT sensors for detecting DNA, enzyme, and organic solvents. Unlike the traditional fluorescence techniques, TFT sensor is a “label-free” method, yielding many advantages such as high sensitivity and direct transduction.¹⁴ Furthermore, small-sized and high-density TFT arrays have an intuitive appeal as portable or packed-sized sensors at low cost, which will benefit the developments of a device platform for the direct and rapid medical diagnosis.¹⁵ Therefore, AOS TFT-based sensors exhibit a broad range of applications and potentially lower manufacturing costs.^16,17

In this communication, we develop new solvent sensors using a-ZTO TFTs. Based on a sole trigger, the electrostatic force between the targets and the detection area, a-ZTO based TFTs are effective to detect organic solvents including H₂O. Moreover, ZTO may be a favourable AOS for immobilization of biomolecules with low isoelectric points such as DNA through electrostatic interaction without any electron mediator. In our work, a combustion solution process is carried out to fabricate a-ZTO TFTs and the TFTs are used to detect different solvents (e.g., cyclohexane, ethanol, and deionized water) with high reliability and practicality. A feasible mechanism model related to electron transfer channel (ETC) is proposed to explain the sensor behaviours.

For a-ZTO precursors, solutions (Zn : Sn = 4 : 7 in molar ratio) were achieved by mixing the individual precursor and aging for 24 h. Precursors were synthesized by dissolving SnCl₂ with NH₄NO₃ or Zn(NO₃)₂ in 2-methoxyethanol, acetylacetone, and NH₃ (14.5 M) solutions, respectively. Then, a-ZTO channel films were prepared with a combustion process by spin-coating the precursors on 150 nm SiO₂/n⁺⁺ Si wafers (University Wafer) at 4300 r.p.m. for 35 s, followed by an annealing process at 300 °C for 30 min under air. To obtain the desired thickness, the above steps were repeated for several times. Al metal was employed for the source and drain electrodes by an electron beam evaporation technology. The resultant channel width and length were 1000 and 100 μm, respectively. The pure solvents (cyclohexane, ethanol, and deionized water) were dropped onto the exposed a-ZTO channel surface with a micropipette, and the devices were then set in ambient air to evaporate the solvents for detections.

The crystal phase identification was investigated by an X-ray diffraction (XRD, Bede D1) system with Cu Kα radiation (λ = 0.15406 nm) over the 2θ range of 10–90°. Cross-section microstructure and element line sweeping were carried out using high-resolution transmission electron microscopy (HRTEM, TecnaiG2F20 S-Twin). Thermogravimetric analysis (TGA) and differential thermal analysis (DTA) were carried out by TA Q600 under a N₂ atmosphere at 5 °C min⁻¹ in the temperature range of 25 °C to 400 °C. X-ray photoelectron spectroscopy (XPS, Thermo ESCALAB 250) measurements were performed to investigate the chemical bonding states with a monochromatic Al-Kα (1486.6 eV) X-ray source. Moreover, the signal responses of these biosensors were measured at RT in the dark using an Agilent E5270B.

Fig. 1(a) shows the XRD pattern of the ZTO film. The prepared film is amorphous because there is no obvious diffraction peak of the crystalline phase. The halo peak around 23° is attributed to the glass substrates.^2,6 Fig. 1(b) depicts the TGA and DTA of ZTO precursors. The ZTO DTA reveals a sharp, intense exotherm at about 80 °C, corresponding to the onset of combustion and corresponding abrupt and gradual mass losses. The self-energy generating combustion chemistry makes a crucial role in driving the reaction, thus completing the whole film fabrication.¹⁸ The low decomposition temperature enables flexible sensors possible. Fig. 1(c) and (d) display the HR-TEM image, selected area electron diffraction (SAED) pattern, and cross-section HR-TEM image with energy dispersive X-ray (EDX) line profiles of the ZTO film. As shown in the images, the prepared film is indeed amorphous phase without Zn or Sn related aggregation or lattice-ordered groups.^6,19 From the cross-sectional HR-TEM, the film is dense and with a thickness of about 35 nm. The EDX profile also indicates that the Zn, Sn and O elements distribute uniformly along the cross-sectional film. The relative peak intensity of Zn-K and Sn-L reveals that the molar ratio of Zn to Sn is approximately 4 : (5–8) in the matrix, which fits well with that in the solution.

Fig. 1 (a) XRD patterns of a-ZTO films, (b) TGA and DTA of ZTO precursors, (c) HR-TEM and SAED images of a-ZTO films, and (d) cross-section HR-TEM and EDX line profiles of a-ZTO films.

XPS measurements have been carried out to investigate the Zn, Sn and O chemical bonding states. During the analysis, all the binding energy data are calibrated by taking C 1s reference at 284.8 eV to compensate for any charge-induced shifts. Fig. 2(a) and (b) show the Zn 2p_3/2 and Sn 3d_5/2 XPS spectra of the a-ZTO film, which are located at about 1021.9 and 486.7 eV, respectively, indicating the characteristic values of fully oxides.²⁰ Fig. 2(c) depicts the O 1s XPS spectrum of the a-ZTO film. By using the Gaussian curve approximation, the typical O 1s peak is deconvoluted into three binding peaks: (1) O_I, M–O–M lattice species at 530.4 eV, (2) O_II, oxygen-deficient regions at 531.6 eV, and (3) O_III, weakly bound (M–OR) species such as H₂O, O₂, –OH and –CO₃ at 532.3 eV.²¹ In general, the O_II peak area is proportional to the oxygen vacancy (V_o) concentration,²⁰ so the ratio of S_OII/S_all is used to obtain a relative V_o concentration in the a-ZTO film, where S_OII represents the integral area of O_II peak and S_all is the total area of the three peaks. Here, a ratio of 16.37% is obtained, indicating an appropriate value.²²

Fig. 2 (a) Zn 2p_3/2, (b) Sn 3d_5/2, (c) O 1s XPS spectra of a-ZTO film.

Fig. 3 depicts the typical output and transfer characteristics of a-ZTO TFTs. The field effect mobility (μ_FE) and the threshold voltage (V_th) are deduced from the following equation,

where I_DS, V_DS, V_GS, C_i, W, and L are the drain current, drain–source voltage, gate bias, capacitance per unit area of the gate insulator (0.023 μF cm⁻²), channel width, and channel length, respectively. The subthreshold swing (SS), a significant parameter to indicate the interface trap density, is calculated by

Fig. 3 (a) Output and (b) transfer characteristics of a typical a-ZTO TFT.

The density of the interfacial trap states (N_t) at the ZTO channel/SiO₂ interface can be determined using the following equations,²²

where e, K, T, and q are the Euler's constant, Boltzmann constant, temperature and quantity of one electron, respectively. For our a-ZTO TFTs, the device behaviours are I_on/I_off = 10⁵–10⁶ with a low off-current in an order of about 10⁻¹¹ A, μ_FE = 1–3 cm² V⁻¹ s⁻¹, and V_th = −1 to 2 V. In what follows the TFTs are used to detect the solvents. In every detection case, three TFT devices were applied and the changes of the transfer characteristics for every solvent are almost the same.

The solvent sensors based on a-ZTO TFTs are demonstrated using nonpolar cyclohexane and polar C₂H₅OH and H₂O as the solvents, and the corresponding I_DS–V_GS characteristics are shown in Fig. 4. For nonpolar cyclohexane (Fig. 4(a)), there is little change of I_on/I_off (in the 3.11 × 10⁵ to 3.21 × 10⁵ range) and V_th (in the 0.35–0.53 range) after immobilizing cyclohexane for aging time of 0 and 360 min. This phenomenon indicates the TFT detector is not sensitive to this nonpolar solvent. While for polar C₂H₅OH (Fig. 4(b)), the I_on/I_off value decreases from 6.62 × 10⁵ to 2.42 × 10⁵ after immobilization (0 min), accompanied with the negative V_th shift from 0.98 V to −5.49 V. After aging 360 min under air, the electrical responses of the solvent sensor are I_on/I_off = 5.05 × 10⁵ and V_th = 1.22 V, recovering to the original state. However, for strong polarity H₂O (18.26 MΩ) (Fig. 4(c)), the TFT loses the transfer characteristic with the I_DS value of 10⁻⁴ to 10⁻⁵ A after immobilization (0 min). After 360 min, the device displays an I_on/I_off of 3.46 × 10⁵ and V_th of −0.70 V, also recovering to the original state. Detailed value changes can be seen in Table 1. It is obvious to see that the a-ZTO TFT sensors are very sensitive to different solvents and exhibit a good selectivity. Notably, since the a-ZTO TFT sensors can recover to the original state after durations, they can be used repeatedly in practical applications. In general, electron injections from the channel surface or the channel matrix to the electron transfer channel (ETC) at the interface contribute to the changes of electrical properties.^23,24 It is expected that this effect can also be applied in our case. The detailed sensing mechanism schematics of a-ZTO TFTs to the various solvents (cyclohexane, C₂H₅OH and H₂O) are shown in Fig. 5. In AOS TFTs, the AOS film contains some shallow levels above the Femi level (E_f) but below the bottom of conduction band (E_c), which act as electron trapping centres.^25,26 At V_DS ≥ V_GS − V_th, the energy band of the AOS channel bends to the channel/insulator interface due to the induced field (E). The electrons tend to accumulate at the channel/insulator interface, thus forming the ETC, and the E_f near the interface moves above onto the shallow trapping state level.²⁵ Meanwhile, the shallow trapping levels are all filled with electrons, which are captured from the channel during the applied bias. After immobilizing with cyclohexane on the a-ZTO back-channel, no extra electrons are injected into the channel due to the nonpolarity nature determined by the small dielectric constant (2.10). Generally, the larger dielectric constant corresponds to the stronger polarity. Therefore, amounts of accumulated electrons near the channel/insulator interface are not changed. The shallow trapping levels and E_f keep original, as Fig. 5(a) displays. After aging for 0 and 360 min, the transfer characteristics do not change. The small SS and N_t changes (from original to 360 min) also indicate this phenomenon, as shown in Table 1. However, using strong polarity liquids like absolute alcohol (dielectric constant of 24.5), the C₂H₅OH molecules cause “donor effects” on the channel through the electron-donating hydroxyl group,²⁷

C₂H₅OH = C₂H₅OH⁺ + e (4)

extra electrons are injected into the channel and accumulate at the interface, as shown in Fig. 5(b). The E_f further moves above onto the shallow trapping state level and close to the E_c through the doping. While the shallow trapping levels are still filled with electrons. By the solvent donor effects, the V_th exhibits an evident negative shift with increased I_DS (Fig. 5(b)), which is consistent well with this proposed mechanism. The SS (from 0.35 to 1.09 V per decade) and N_t (from 7.20 × 10¹¹ to 2.52 × 10¹²) changes also indicate that the interface trap states are indeed increased by the injected electrons. After aging 360 min, the TFT sensor recovers to the original state (shown in Table 1). Naturally, using the strong polar solvent H₂O (dielectric constant of 78.5), it is quite reasonable to suppose that more extra electrons are generated into the ETC,^28–30

H₂O = H₂O⁺ + e (5)

Fig. 4 Transfer characteristics of a-ZTO TFTs as solvent sensors for detecting solvents: (a) cyclohexane, (b) C₂H₅OH, and (c) H₂O.

Table 1 Electrical parameters (I_on/I_off, V_th, SS, and N_t) of a-ZTO TFT sensors for detecting different solvents

Sample	Cyclohexane			C2H5OH			H2O
	Original	0 min	360 min	Original	0 min	360 min	Original	0 min	360 min
Ion/Ioff (× 10^5)	3.21	3.21	3.11	6.62	2.42	5.05	5.45	—	3.46
Vth (V)	0.35	0.53	0.37	0.98	−5.49	1.22	−0.27	—	−0.70
SS (V per decade)	0.89	0.87	0.93	0.35	1.09	0.46	0.25	—	0.70
Nt (× 10^12 cm^−2)	2.01	1.98	2.11	0.72	2.52	0.98	0.47	—	1.57

---

Fig. 5 Schematic diagram of mechanism model to explain behaviours of a-ZTO TFT sensors in detecting different solvents: (a) cyclohexane, (b) C₂H₅OH, and (c) H₂O.

The massive electron injection induces the fact that the E_f tends much closer to the E_c, accompanied with fully filled electrons at the shallow defect levels, as Fig. 5(c) displays. Thus, the large I_DS is obtained and so as to make the TFT lose the transfer characteristic at 0 min (Fig. 4(c)). After aging 360 min, the sensor performance also recovers to the original state in the main. The analysis can be confirmed by the SS and N_t values, as listed in Table 1. Note that our recent study reveals that the a-ZTO TFT can be used to detect such biomaterials as DNA and enzyme, which will be reported elsewhere. The model presented here is also applied to these biosensor behaviours.

In summary, solvent sensors based on a-ZTO TFTs have been successfully demonstrated. The a-ZTO TFTs were fabricated using a combustion process and displayed a strong sensitivity in detecting solvents (e.g., cyclohexane, ethanol, and deionized water) with a good selectivity. A feasible mechanism model was developed to explain the sensor behaviours. The donor effects by the solvents determinate the location of Femi level (E_f), thus making the TFT different response. It is a reasonable and feasible model not only for our TFT detections, but also can be applied to other amorphous TFT detections.

Acknowledgements

This work was supported by National Natural Science Foundation of China under Grant no. 51372002 and Municipal Science & Technology Project of Wuhu City under Grant no. 2013zd18.

Notes and references

1. T. H. Chao, M. J. Chang, M. Watanabe, M. H. Luo, Y. J. Chang, T. C. Fang, K. Y. Chen and T. J. Chow, Chem. Commun., 2012, 48, 6148.
2. Q. J. Jiang, L. S. Feng, C. J. Wu, R. J. Sun, X. F. Li, B. Lu, Z. Z. Ye and J. G. Lu, Appl. Phys. Lett., 2015, 106, 053503.
3. M. G. Kim, M. G. Kanatzidis, A. Facchetti and T. J. Marks, Nat. Mater., 2011, 10, 382.
4. K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano and H. Hosono, Nature, 2004, 432, 488.
5. C. J. Wu, X. F. Li, J. G. Lu, Z. Z. Ye, J. Zhang, T. T. Zhou, R. J. Sun, L. X. Chen, B. Lu and X. H. Pan, Appl. Phys. Lett., 2013, 103, 082109.
6. Q. J. Jiang, J. G. Lu, J. P. Cheng, X. F. Li, R. J. Sun, L. S. Feng, W. Dai, W. C. Yan and Z. Z. Ye, Appl. Phys. Lett., 2014, 105, 132105.
7. Y. P. Wang, J. G. Lu, X. Bie, Z. Z. Ye, X. Li, D. Song, X. Y. Zhao and W. Y. Ye, Appl. Surf. Sci., 2011, 257, 5966.
8. J. W. Hennek, J. Smith, A. Yan, M. G. Kim, W. Zhao, V. P. Dravid, A. Facchetti and T. J. Marks, J. Am. Chem. Soc., 2013, 135, 10729.
9. H. W. Zan, C. C. Yeh, H. F. Meng, C. C. Tsai and L. H. Chen, Adv. Mater., 2012, 24, 3509.
10. W. B. Jackson, R. L. Hoffman and G. S. Herman, Appl. Phys. Lett., 2005, 87, 193503.
11. H. Q. Chiang, J. F. Wager, R. L. Hoffman, J. Jeong and D. A. Keszler, Appl. Phys. Lett., 2005, 86, 013503.
12. K. H. Lim, K. Kim, S. Kim, S. Y. Park, H. Kim and Y. S. Kim, Adv. Mater., 2013, 25, 2994.
13. Y. C. Shen, C. H. Yang, S. W. Chen, S. H. Wu, T. L. Yang and J. J. Huang, Biosens. Bioelectron., 2014, 54, 306.
14. F. Yan, S. M. Mok, J. Yu, H. L. Chan and M. Yang, Biosens. Bioelectron., 2009, 24, 1241.
15. S. J. Kim, J. Jung, K. W. Lee, D. H. Yoon, T. S. Jung, S. R. Dugasani, S. H. Park and H. J. Kim, ACS Appl. Mater. Interfaces, 2013, 5, 10715.
16. S. J. Kim, B. Kim, J. Jung, D. H. Yoon, J. Lee, S. H. Park and H. J. Kim, Appl. Phys. Lett., 2012, 100, 103702.
17. J. Jung, S. J. Kim, K. W. Lee, D. H. Yoon, Y. G. Kim, H. Y. Kwak, S. R. Dugasani, S. H. Park and H. J. Kim, Biosens. Bioelectron., 2014, 55, 99.
18. M. G. Kim, M. G. Kanatzidis, A. Facchetti and T. J. Marks, Nat. Mater., 2011, 10, 382.
19. M. Kim, J. H. Jeong, H. J. Lee, T. K. Ahn, H. S. Shin, J. S. Park, J. K Jeong, Y. G. Mo and H. D. Kim, Appl. Phys. Lett., 2007, 90, 212114.
20. M. G. Kim, H. S. Kim, Y. G. Ha, J. Q. He, M. G. Kanatzidis, A. Facchetti and T. J. Marks, J. Am. Chem. Soc., 2010, 132, 10352.
21. M. G. Kim, J. W. Hennek, H. S. Kim, M. G. Kanatzidis, A. Facchetti and T. J. Marks, J. Am. Chem. Soc., 2012, 134, 11583.
22. Y. S. Rim, D. L. Kim, W. H. Jeong and H. J. Kim, Appl. Phys. Lett., 2010, 97, 233502.
23. E. Fortunato, P. Barquinha and R. Martins, Adv. Mater., 2012, 24, 2945.
24. P. Wu, J. Zhang, J. G. Lu, X. F. Li, C. J. Wu, R. J. Sun, L. S. Feng, Q. J. Jiang, B. Lu, X. H. Pan and Z. Z. Ye, IEEE Trans. Electron Devices, 2014, 61, 1431.
25. M. Kimura, T. Nakanishi, K. Nomura, T. Kamiya and H. Hosono, Appl. Phys. Lett., 2008, 92, 133512.
26. J. Zhang, X. F. Li, J. G. Lu, Z. Z. Ye, L. Gong, P. Wu, J. Huang, Y. Z. Zhang, L. X. Chen and B. H. Zhao, J. Appl. Phys., 2011, 110, 084509.
27. X. Y. Xue, Y. J. Chen, Y. G. Wang and T. H. Wang, Appl. Phys. Lett., 2005, 86, 233101.
28. J. K. Jeong, H. W. Yang, J. H. Jeong, Y. G. Mo and H. D. Kim, Appl. Phys. Lett., 2008, 93, 123508.
29. J. S. Park, J. K. Jeong, H. J. Chung, Y. G. Mo and H. D. Kim, Appl. Phys. Lett., 2008, 92, 072104.
30. J. Zhang, X. F. Li, J. G. Lu, N. J. Zhou, P. J. Guo, B. Lu, X. H. Pan, L. X. Chen and Z. Z. Ye, RSC Adv., 2014, 4, 3145.

---