Fabrication of a solution-processed low voltage TFT by using colloid 2D ZnO nanosheets and its application as a UV photodetector

Abstract

ZnO nanostructures have been extensively employed in optoelectronic devices because of their unique optoelectronic properties; however, these devices have been developed using physical vapor deposition techniques, which are costly and need a state-of-the-art fabrication facility. Hence, a solution-processed, cost-effective, low-temperature method is required for the large-scale fabrication of 2D material-based electronic devices. In this contribution, we report template, polymer, and surfactant-free wet chemical synthesis of 2D ZnO nanostructures having dimensions of ∼200 nm and thickness of ∼30 nm following the hydrothermal method. Detailed structural, morphological, and optical investigation revealed the formation of a pure hexagonal wurtzite phase of ZnO nanosheets. Utilizing the as-synthesized nanosheets, solution-processed thin film transistors (TFTs) are fabricated under low annealing temperatures that exhibit a high carrier mobility of 8.05 cm² V⁻¹ s⁻¹ and an on–off ratio of ∼10⁵. Also, these TFTs show high photosensitivity and can be used as UV detectors. Thus, our study highlights low-temperature facile fabrication of 2D ZnO TFTs, which may have promising applications in electronic displays, logic circuits, UV detectors, biosensors, and portable electronics.

---

1. Introduction

Numerous research works are ongoing worldwide into two-dimensional (2D) nanocrystals owing to their exceptional optoelectronic, mechanical, and electrochemical properties.^1–9 In 2D materials such as nanoflakes, nanosheets, nanoribbons, etc., the electrons are confined in one direction, and free movement of electrons occurs in the two-dimensional non-nanoscale plane. Because of its strong in-plane bonding and weak van der Waals forces, 2D materials are easily exfoliated to a few even single-layer sheets. These thin sheets are stable against agglomeration, buckling, and rolling and can eventually be processed like a standard thin film. As a result of their mechanical robustness, they can withstand mechanical stress and cycle strain, which is required for flexible electronics.^10,11 Additionally, excellent carrier mobility, high on/off ratio, low power dissipation, and low power consumption make 2D materials an ideal substance for flexible electronics.^12–14 Owing to the outstanding electrostatic integrity of 2D semiconductors caused by their sub-nanometer thickness and flexibility, the power consumption of the device in a standby state is drastically reduced by 100 000 times compared to that of traditional Si transistors.^15,16 Large-scale synthesis of 2D materials requires an easy, low-temperature, solution processed wet chemical method. For scalable, facile, non-expensive synthesis methods, 2D metal oxide semiconductors are very suitable and can even substitute silicon thin films in several applications.^17–20

Zinc oxide (ZnO) is one of the most widely used metal oxides for thin-film transistors (TFTs) because of its exceptional optoelectronic features, high electron mobility, abundant availability, and non-toxicity.^21–23 In 2012, ZnO-based TFTs were commercialized by the Sharp Corporation, and even Apple's iMAC employed them in the device.²⁴ However, all the commercially accessible TFTs are fabricated through the physical vapor deposition method, which is expensive and demands state-of-the-art fabrication facilities.^25,26 Solution-based fabrication methods such as spin coating, spray coating etc. are easily accessible, economical, and require fewer vacuum techniques. Meyers et al. fabricated a ZnO TFT through the spin coating method and obtained a high carrier mobility of 4.3 cm² V⁻¹ s⁻¹.²⁷ Though solution-based synthesis of various size and shaped ZnO nanostructures has been reported, to the best of our knowledge, there are very few reports of surfactant-free, facile solution-based synthesis of ZnO 2D nanosheets and no reports of solution-processed ZnO 2D TFTs. For device applications, surfactant-free nanocrystals are beneficial as surfactants have adverse effects on carrier mobility.^28–30

In this report, we have synthesized template, polymer and surfactant-free ZnO 2D nanosheets following a hydrothermal method and studied their structural, chemical, morphological, and optoelectronic properties thoroughly. The surfactant-free, template-free synthesis method offers a simple, cost-effective, and environmentally friendly approach to material fabrication. By eliminating the need for surfactants and templates, this method reduces impurities and enhances surface properties. The morphology obtained from this method is compared with that from surfactant and template-based ZnO nanoparticle synthesis and the results are tabulated in Table S1 (ESI†). The formation of a 2D nanosheet of ∼200 nm having a thickness of ∼30 nm can be seen from electron microscopy and atomic force microscopy studies, and its crystalline phase has been identified from an XRD study. These ZnO nanosheets are employed for the fabrication of a TFT under low annealing temperatures (≤350 °C). The device annealed at 110 and 350 °C (Fig. 1), exhibited a high carrier mobility of 6.59 and 8.05 cm² V⁻¹ s⁻¹, respectively, which is remarkable for a solution-processed undoped ZnO TFT and comparable to that of expensive low-temperature processed indium-doped ZnO TFTs. A solution-processed LiInSnO₄ thin film has been used as a gate dielectric of this TFT and is capable of operating these TFTs within 2 V, which is required for portable electronics. Also, the fabricated devices exhibit high UV light sensitivity with an external quantum efficiency (EQE) of ∼8 and 2% for 110 °C and 350 °C annealed devices, respectively. Besides, the detectivity of these TFTs is greater than 10¹⁰ Jones. Overall, this scalable synthesis of 2D ZnO not only enables us to fabricate high mobility, low operating voltage ZnO TFTs, but also shows high UV sensitivity, which has been discussed in the following sections.

Fig. 1 Schematic representation of solution-processed TFTs (p⁺⁺-Si/LiInSnO₄/ZnO NS + ZnO sol) with a semiconducting layer processed in (a) device A (110 °C), (b) device B (350 °C) and (c) device C (MIM structure of p⁺⁺-Si/LiInSnO₄/Al). A photograph of (d) device A and (e) device B.

2. Experimental

2.1. Materials and chemicals

Hexamethylenetetramine (C₆H₁₂N₄), zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O), and ethanol were obtained from SRL, India. LiInSnO₄, lithium acetate [C₂H₃LiO₂] (TCI >98.0%), anhydrous indium chloride [InCl₃, Alfa Aesar > 99.99%], and tin(II) chloride [SnCl₂, Sigma-Aldrich >99.99%] were purchased. Di-ethanolamine (DEA) and 2-methoxy ethanol (2-MEA) were purchased from Sigma-Aldrich. The chemicals used in this study were used as is without further purification.

2.2. Preparation of ZnO nanosheets

Surfactant-free ZnO nanosheets were synthesized via a hydrothermal method following our previous report.³¹ A 1 : 1 molar ratio of C₆H₁₂N₄ and Zn(NO₃)₂·6H₂O was mixed with 150 ml of deionized water and swirled for 5 min to form a suspended solution. Later, the suspension was transferred into a Teflon-lined autoclave and kept at 100 °C for 12 h. The white precipitate was cleaned several times with ethanol and deionized water. The resulting ZnO nanosheets (ZnO NSs) were dried overnight at 80 °C in a vacuum oven. The precursor solution of LiInSnO₄ was prepared by mixing solutions of three different salts, namely lithium acetate, anhydrous InCl₃, and SnCl₂, with equal atomic ratios. Initially, three individual solutions of these three different salts were prepared by dissolving them in a 2-MEA solvent with a molar concentration of 200 mM. After that, these three solutions were mixed in a 1 : 1 : 1 molar ratio in the ambient atmosphere to obtain a homogeneous precursor solution of LiInSnO₄.^32–34 Similarly, a precursor solution of ZnO of concentration 20 mM was prepared from zinc acetate (dehydrate) by dissolving it in 2-MEA. After stirring the solution under the ambient atmosphere for an hour, 20 μl DEA was added to stabilize the solution, and it was further stirred at 60 °C for 2 h to dissolve the solute precursor properly. For the pure ZnO NS dispersion, the ZnO NS powder was dispersed in ethanol maintaining a 25 mg ml⁻¹ concentration and was kept in an ultrasonic bath for one hour. Later, the supernatant of the dispersion was used for further studies. Finally, a mixed solution phase of precursor ZnO (Sol) and ZnO NSs was obtained by mixing both solutions in a 1 : 1 ratio in an ultrasonic bath for an hour.

2.3. Device fabrication

To investigate the transport properties of these ZnO NSs, two different thin film transistors (TFTs) are fabricated on a highly p-doped silicon (p⁺⁺-Si) substrate using LiInSnO₄ as a gate dielectric and ZnO NSs as a semiconductor channel. These TFTs are fabricated in a top contact bottom gated geometry. The channel of the TFTs is deposited by dispersing ZnO NSs in a precursor sol of ZnO (20 mM zinc acetate solution). The role of the low-concentration zinc acetate is to connect 2D ZnO for better charge transport. After deposition of the ZnO semiconducting channel, it was annealed at 110 °C and 350 °C to make two different devices, named device A (Fig. 1(a)) and device B (Fig. 1(b)), respectively. Two devices are also made without adding zinc acetate (Sol) and annealed at 110 °C and 350 °C to make two different reference devices named device A-R (Fig. S1a, ESI†) and device B-R (Fig. S1b, ESI†), respectively. In addition, a metal–insulator–metal (MIM) device (device C, Fig. 1(c)) has been fabricated for electrical characterization of the dielectric film, by depositing LiInSnO₄ thin film on p⁺⁺-Si, whereas Al has been used as a top electrode (Fig. 1(c)). For these device fabrications, initially p⁺⁺-Si with dimensions of 15 mm × 15 mm was cleaned with a dedicated soap solution, followed by water rinsing. The wafer was cleaned in an ultrasonic bath using water, acetone, and isopropanol solution for 15 minutes each to remove unwanted impurities, making the substrate smooth and compatible for thin film spin coating. Finally, to eliminate the surface oxidation, the substrate was placed in a plasma chamber for ten minutes.

The precursor sol of the LiInSnO₄ dielectric was spin-coated over a pre-cleaned p⁺⁺-Si substrate at 5000 rpm for 50 seconds, and then transferred to a furnace at 350 °C for 30 minutes. The process was repeated three times to achieve a sufficient thickness of the gate dielectric, which was finally annealed at 550 °C for an hour to get a compact film LiInSnO₄ with minimum defects. Then, the ZnO NS solution was spin-coated on top of this LiInSnO₄ thin film, which works as a semiconductor channel of the TFTs. In the case of device A-R and device B-R, a colloidal solution of ZnO NSs was spin-coated over the LiInSnO₄ dielectric film at a spinning speed of 2500 rpm for 60 s, whereas the mixed ZnO NS and ZnO precursor sol solution was spin-coated over the LiInSnO₄ dielectric film to fabricate device A and device B. After ZnO NS deposition (for device A-R) or after mixed ZnO NS and ZnO sol solution deposition (for device A), the device was annealed over a hotplate at 110 °C for 1 hour to remove the solvents. In contrast, the device was annealed in a muffle furnace at 350 °C for 1 hour (for device B-R and device B, respectively), which converted the precursor ZnO sol (zinc nitrate solution) to a ZnO film that can also work as an inter 2D ZnO NS connector, facilitating charge transport between the 2D ZnO NSs. Finally, aluminum (Al) electrodes of thickness ∼70 nm were deposited on top of the ZnO film through a shadow mask and work as the source and drain electrodes of the TFTs with a ‘width-to-length ratio’ (W/L) of 118 (23.6 mm/0.2 mm). Device C (MIM device) was fabricated in a similar way without using the ZnO layer and has been used for the characterization of the LiInSnO₄ dielectric thin film.

3. Results and discussion

3.1. Structural and morphological properties

Powder-XRD analysis was carried out to investigate the detailed crystal structure and phase composition of the as-synthesized ZnO NSs. The presence of characteristic diffraction peaks at 31.72°, 34.36°, and 36.18°, corresponding to (100), (002), and (101) crystal planes, respectively, confirmed the formation of the hexagonal wurtzite structure and the diffraction angles perfectly matched with the JCPDS (Joint Committee on Powder Diffraction Standards) 36-1451 data (Fig. S2, ESI†). It should be noted that the intensity ratio of the [(101)/(002)] diffraction peaks increases from 2.16 to 3.19 compared to the JCPDS value, suggesting the formation of a 2D sheet. Furthermore, pure hexagonal wurtzite phase was present because no impurity peaks or peak shifts were observed. The phase purity and the refinement of the unit cell parameters were performed by the Rietveld method using Profex, as shown in Fig. 2(a). The fitting parameters (χ² = 1.18 and goodness of fit (GOF) = 1.08) imply that the as-synthesized ZnO NSs are in a pure phase and the refined unit cell parameters (a = 3.2524 Å and c = 5.2102 Å) are in good agreement with the standard hexagonal wurtzite ZnO.³⁵

Fig. 2 (a) Rietveld refined XRD patterns and (b) W–H plot (β cos θ vs. 4 sin θ) of ZnO NS.

From the XRD pattern, the crystallite size of the NS was calculated by the Debye–Scherrer equation for the highest intensity peak (101):

where D = crystallite size, λ = wavelength of X-rays, k = shape factor (0.89), β = FWHM (full width at half maximum) (in radians) and θ = angle of diffraction (in radians), and was found to be ∼56 nm. It was also observed that with an increase in diffraction angle, the crystallite size rises consistently. Moreover, from macrostrain analysis, it was learned that macrostrain enlarges gradually with an increase in θ, implying a strong dependency on crystalline size. Hence, crystalline size is also calculated following the Williamson Hall plot (W–H plot) (Fig. 2(b)). In principle, the W–H plot largely depends on the approximate formulae for size broadening (β_L), and strain broadening (β_e), which varies differently to Bragg angle θ.

The average crystalline size was calculated as ∼48 nm using eqn (2), where β_e is the strain broadening, β_L is the size broadening, ε is the lattice strain, θ is the Bragg's angle, and L is the crystallite size; the K value is 0.9 and is in good agreement with the previously estimated crystallite size (∼56 nm) from the Debye–Scherrer equation.

β_tot = β_e + β_L = Cε tan θ + (Kλ/L cos θ) (2)

The dislocation density, calculated from the above equation (eqn (2)), was found to be 4.45 × 10¹⁴ m². A higher value of dislocation density defines greater mechanical strength of the synthesized nanosheets, and the negative slope of the W–H plot indicates that macrostrain cannot be a dominant source of broadening.³⁶ From these results it can be inferred that the influence of domain/crystallite size and macrostrain is negligible, indicating the formation of a less defective compound.

The morphological evolution of the synthesized nanosheets was performed through AFM and electron microscopies. The formation of a 2D nanosheet having a length of ∼200 nm was established by AFM analysis, as displayed in Fig. 3(a) and (b). As revealed by the height image (Fig. 3(b)), the thickness of the nanosheet was ∼30 nm. Beyond that, FESEM images also showed the formation of a 2D sheet-like structure having dimensions of ∼200 nm (Fig. 3(c)). From the EDS analysis, the atomic ratio of Zn and O was found to be 1 : 1 (Fig. 3(d)), confirming the formation of pure ZnO NSs as revealed earlier from the p-XRD study. To further analyze the individual NS morphology and crystal structure, high-resolution transmission electron microscopy (TEM) was performed. The bright-field TEM image confirms the formation of thin nanosheets having a length of ∼400 nm and a width of ∼300 nm, respectively. From Fig. 3(e), the 2D structure of the as-synthesized ZnO NSs is visible. Lattice fringes with an interplanar distance d = 0.282 nm corresponding to the (100) plane of the hexagonal wurtzite structure of the ZnO nanocrystal are distinctly identified (Fig. 3(f)). Furthermore, the synthesis of phase pure ZnO NSs with a wurtzite structure is confirmed by the fast Fourier transform (FFT) pattern displayed in the inset of Fig. 3(f). The lattice parameters that were determined from the FFT pattern closely match with the reported values.

Fig. 3 (a) and (b) AFM images of ZnO NSs; inset shows the height profile; (c) FESEM image of ZnO NSs (scale bar 200 nm); (d) EDS profile of ZnO NSs; (e) bright-field transmission electron microscopic image of ZnO NSs and (f) high-resolution TEM image of ZnO NSs with a d-spacing of 0.282 nm corresponding to the (100) reflection (enlarged d-spacing is shown in the inset) along with Fourier transform of a thin and isolated area of the sheet marked with a red square region. The lattice fringe region of the crystalline profile with a computed length scale is shown in the right inset.

3.2. Optical properties

After confirming the formation of pure phase 2D ZnO NSs, we investigated the optoelectronic properties and band structure of the as-synthesized compound as the band gap of the material plays an important role in fabricating efficient TFT devices.³⁷ The optical properties of the as-synthesized NSs were studied through UV-Vis diffuse reflectance spectroscopy (UV-Vis DRS). Fig. S3 (ESI†) shows the optical absorption and reflectance spectra of the ZnO NSs. ZnO shows an excitonic peak in the UV region. The absorption edge occurs at 430 nm and is attributable to electron transition from the valence band to the conduction band (O2p-Zn3d).³⁸ The broad absorption band that ranges towards longer wavelengths might be owing to the movement of the electronic cloud on the overall skeleton of the ZnO NS.³⁹ The band gap of the ZnO was calculated theoretically from the reflectance spectrum using the Kubelka–Munk function (F(R)), according to P. Kubelka and F. Munk's theory (1931).⁴⁰ The as-recorded reflectance spectrum can be converted to the matching absorption spectra by applying the Kubelka–Munk function (F(R_∞)) and using eqn (3)

(F(R_∞)hν)^γ = A(hν − E_g) (3)

where h = Plank's constant, ν = frequency of the photon, A = proportionality constant, E_g = band gap of the material and γ = nature of electron transition (= 2 for direct allowed transitions). Using the above equation, the band gap of the ZnO NSs was calculated to be 3.2 eV (Fig. 4(a)). On further theoretical investigation (inset of Fig. 4(a)) along with valence band XPS (Fig. S4a, ESI†), the concluded value leads to an understanding of the conduction band edge being at E_CB −0.44 eV and the valence band maximum being at 2.76 eV.

Fig. 4 (a) Kubelka–Munk plot of the ZnO NSs; (b) Raman spectra of the as-synthesized ZnO NSs; high-resolution core level (c) Zn 2p and (d) O 1s XPS spectra of the ZnO NSs.

To determine more about the 2D-microstructure, crystallization, phase purity, structural disorder, and vibrational characteristics; a room-temperature Raman spectroscopic study was conducted. The hexagonal wurtzite structure of the ZnO nanoparticles has a P63mc space group.^41,42 Theoretically, ideal ZnO crystals show first-order Raman scattering owing to the vibration of the optical phonons near the Brillouin zone at the Γ point.⁴¹ From group theory, the optical modes corresponding to the hexagonal wurtzite structure are specified in eqn (4)

Γ_opt = A₁ + 2B₂ + E₁ + 2E₂ (4)

Here, the first-order Raman-active modes are A₁, E₁, and E₂. The B₁ modes, often known as silent modes, are typically inactive considering the Raman selection rule.^41,42Fig. 4(b) shows the Raman spectra corresponding to different atomic vibrations of the as-synthesized ZnO NSs, where the red arrows indicate the main direction of translation. The incident light from the 532 nm laser is directed along the [0001] orientation axis. Two distinct peaks at 101 and 437 cm⁻¹, respectively, corresponding to E₂(low)-Zn sublattice vibration and E₂(high)-oxygen sublattice vibration of the ZnO wurtzite structure, were observed. The 2E₂(low) and E₂(high)–E₂(low) modes emerged at 202 cm⁻¹ and 333 cm⁻¹, respectively.⁴³ Though the peak at 576 cm⁻¹ represents the 1st order A₁ Raman active mode, indicating the presence of a zinc blende phase, the minimal intensity of the mode implies the formation of phase pure wurtzite ZnO NSs, supporting the XRD study.

To identify the chemical states of the surface elements and intricate electrical structures, X-ray photoelectron spectroscopy (XPS) analysis was performed. From the wide survey spectra, the presence of Zn, O, and C elements was observed, indicating that Zn and O are the primary substances and that the produced nanosheets are completely free of irrelevant impurities (Fig. S4b, ESI†). The binding energies are calibrated using the reference C 1s emission at 284.5 eV. The core level Zn 2p spectrum shows two peaks at 1044.5 and 1021.5 eV, respectively, corresponding to Zn 2p_1/2 and Zn 2p_3/2 spin–orbit splitting, confirming the presence of a pure Zn²⁺ oxidation state at the nanosheet surface (Fig. 4(c)). The binding energy difference of 23 eV for spin–orbit splitting pairs well with the earlier report. The high-resolution O1s core level spectrum can be deconvoluted into two peaks at 532.6 and 529.5 eV, respectively (Fig. 4(d)). The peak at lower binding energy is due to the presence of lattice oxygen in the ZnO wurtzite structure. The broad shoulder at 532.6 eV is owing to the presence of adsorbed hydroxyl groups on the ZnO NS surface and O–C bond.^44,45

3.3. Characterization of a dielectric thin film

Electrical characterization of the LiInSnO₄ dielectric thin film has been performed with the MIM device (device C) structure. The capacitance vs. frequency (C–f) study of device C indicates that the areal capacitance of the film decreases with frequency; particularly above 10³ Hz, it reduces very rapidly (Fig. S5a, ESI†). The areal capacitance of the film at low frequency (∼20 Hz) is high (>200 nF cm⁻²). Moreover, the areal capacitance of the dielectric is ∼180 nF cm⁻² at 50 Hz frequency. The current vs. voltage (I–V) characteristics of device C indicate that the DC conductivity of the LiInSnO₄ dielectric thin film is very poor (Fig. S5b, ESI†), which is required for use as the gate dielectric of a TFT. The details of these C–f and I–V studies are discussed in the ESI† section (Fig. S5). No visible peak has been observed in the GIXRD pattern of the LITO dielectric thin film, confirming the amorphic nature (Fig. S6a, ESI†). The AFM study of this LiInSnO₄ (Fig. S6b and c, ESI†) thin film indicates that its roughness is ∼0.406 nm, which is sufficiently low for use as a gate dielectric of a TFT.⁴⁶ Moreover, the cross-sectional SEM images of the LiInSnO₄ dielectric layer and the ZnO nanosheet film annealed at 110 °C and 350 °C, are presented in Fig. S6(d)–(f) (ESI†), respectively. The estimated thickness of the LiInSnO₄ layer is 59 nm and those of the ZnO mixed film annealed at 110 °C and 350 °C are 19 nm and 20 nm, respectively. The optical transparency of these films is very high (>85%) because of their high optical band gap (∼5.16 eV), which is shown in the ESI† (Fig. S7).

3.4. Electrical characterization

The electrical characterization of device A and device B was performed under ambient atmospheric conditions, as shown in Fig. 5. During the output characteristics study, drain current (I_D) is measured by varying the drain voltage (V_D) from 0 V to 2 V for a fixed gate bias (V_G), whereas the gate voltage has been tuned from −0.5 to 2 V for different sets of characteristics. The I_Dvs. V_D plots of device A and device B are shown in Fig. 5(a) and (c), respectively. These data clearly show that device B has a relatively lower drain current than device A, whereas device B has better current saturation than device A. However, in both cases, the required external voltage can be within the 2.0 V limit, which is important for portable electronics applications. Similarly, during transfer characteristics (I_Dvs. V_G), V_G varies from −2 to 2 V, whereas V_D was 1 V. The transfer characteristics of device A and device B are shown in Fig. 5(b) and (d), respectively. These characteristics clearly show that device B has a high positive threshold voltage (V_th), whereas device A has a negative threshold voltage. Besides, device B has ∼10⁴ times higher on/off ratio (2.05 × 10⁵) than device A. The effective carrier mobility and subthreshold swing (SS) of the TFTs are measured by using the following equations;⁴⁷where C is the areal capacitance of the dielectric thin film and μ is the effective carrier mobility (eqn (5)) of the TFT. Besides, W and L are the width and length of the channel in the TFTs. The extracted carrier mobility of device A and device B, which are obtained from Fig. 5(b) and (d) are 6.60 (cm² V⁻¹ s⁻¹) and 8.05 (cm² V⁻¹ s⁻¹), respectively. The subthreshold swings (SSs), which are extracted by using eqn (2) are 202 mV dec⁻¹ and 1049 mV dec⁻¹ for device B and device A, respectively. The high SS value of device A originated due to its low on/off ratio. The histogram of the statistical analysis of the mobility, threshold voltage (V_th), and subthreshold swing (SS) of the 12 devices of device A are shown in Fig. S8(a), (b) and (c) (ESI†), while Fig. S8(d), (e) and (f) (ESI†) for device B, respectively. The comparative TFT parameters of device A and device B are shown in Table 1 and a comparison with earlier reported work has been presented in Table 2.

Fig. 5 (a) Output and (b) transfer characteristics of device A (ZnO NS + ZnO sol gel)(110 °C); (c) output and (d) transfer characterstics of the device B (ZnO NS + ZnO sol gel)(350 °C).

Table 1 Summary of the device parameters of devices A and B

Device configuration	On/off ratio	Subthreshold swing (mV decade^−1)	Mobility (cm^2 V^−1 s^−1)	Threshold voltage Vth (V)	Number of trap states N. (cm^−1)
Device A	0.76 × 10^2	1050 ± 8.23	6.5 ± 0.18	−1.5 ± 0.0038	1.76 × 10^13
Device B	2.05 × 10^5	205.5 ± 3.10	7.95 ± 0.0008	0.71 ± 0.0017	2.83 × 10^12

---

Table 2 Comparison with earlier reported ZnO TFTs

Device	Processing technique	Operation voltage (volt)	Threshold voltage (Vth) volt	On–off ratio	Carrier mobility (cm^2 V^−1 s^−1)	Subthreshold swing (mV decade^−1)	Interface trap state density (cm^−2)	Ref.
ZnO TFT	Photolithography & wet etching	10	−2.0	1.3 × 10^5	0.15	108	—	48
ZnO TFT	ALD	4	−0.82	980	5.17	824	—	49
ZnO/SiN TFT	RF magnetron sputtering	10	—	2.7 × 10^7	5.2	—	—	50
ATO/ZnO TFT	Solution deposition	40	—	10^5	5–6	—	—	51
SiO2/ZnO TFT	RF sputtering	30	20	2.7 × 10^7	5.1	—	—	52
ZnO TFT	RF magnetron sputtering	20	∼0	1.6 × 10^6	1.2	3000	1.9 × 10^12	53
ZnO nanoparticle-based TFT	Solution process	40	−9	10^2	0.0004	—	—	54
ZnO TFT	ALD	30	−12.5	1.28 × 10^6	8.82 × 10^−3	1210	—	55
Device B	Solution process	2	0.71 ± 0.0017	2.05 × 10^5	7.95 ± 0.0008	205.5 ± 3.10	2.83 × 10^12	This work

---

Besides, the electrical characterization of the reference TFTs (ZnO film without ZnO sol) was also investigated, as shown in Fig. S9 (ESI†). It has been observed that the channel current for both low and high-temperature annealed devices is very low, and the carrier mobility for both devices is 0.5 cm² V⁻¹ s⁻¹ as listed in Table S2 (ESI†). Both these devices show reasonably good UV sensitivity, as shown in Fig. S10 (ESI†). However, their photoresponse speed is slower than those of devices A and B (Fig. S11, ESI†). The EQE spectra of these reference devices are very similar to those of devices A and B, but limited to 3% (Fig. S12, ESI†). Both these devices also show a shifting of the threshold voltage with light illumination (Fig. S13, ESI†).

3.5. Optical response of the devices

The optical response under different intensity UV light (with λ_max = 365 nm) illumination of the devices is shown in Fig. 6. Fig. 6(a) and (b) show the variation in transfer characteristics under UV light illumination of device A and device B, respectively. It can be noted that the variation of the depletion mode current is significantly higher than the accumulation mode current. Also, the variation of the depletion mode current of device B is significantly higher than that of device A. The variation in output characteristics of device A and device B of the drain current is shown in Fig. 6(c) and (d), respectively, which clearly shows that the current variation of device B is also significantly larger than that of device A. Besides, under the illumination of UV light, there is a remarkable shift in threshold voltage towards the negative gate voltage, valued from −1.54 V to −2.34 V and 0.72 V to 0.16 V for device A and device B, respectively.

Fig. 6 Transfer characteristics under UV light illumination of (a) device A and (b) device B, and output characteristics under UV light illumination of the TFTs in (c) device A and (d) device B, under UV illumination.

The optical parameters, like photosensitivity (S), photo responsivity (R), and detectivity (D*) of the devices have been calculated by using the following equations,⁵⁶

where I_Ph, P_opt, S, and I_dark refer, respectively, to photocurrent, power of incident light, active area, and dark current of the device. q, I_light, and h are the electronic charge (1.6 × 10⁻¹⁹ Coulomb), device current under light conditions, and Planck's constant (6.62 × 10⁻³⁴ m² kg s⁻¹).

The gate bias-dependent photosensitivity (S) of device A and device B under different intensities of UV-light illumination has been illustrated in Fig. 7(a) and (b), respectively, where V_G is varied within −2 V to 2 V with V_D = 1.0 V. From the sensitivity curve (Fig. 7(b)), it can be noticed that device B has a UV light sensitivity value of 225 at V_G ∼ 0.2 V, which is around two orders higher than that of device A (Fig. 7(a)). However, the detectivity (D*) of device A in depletion mode is higher than that of device B, and this is reversed in accumulation mode. Overall, both devices have detectivity (D*) >10¹⁰ Jones. This comparative investigation indicates that the photosensitivity and detectivity of device B (at 0 V gate bias) are significantly higher than those of device A, although their detectivity (D*) is in the same order. A summary of these device parameters is given in Table 3.

Fig. 7 Photosensitivity of the devices under UV illumination: (a) device A and (b) device B. Comparative plot of responsivities and detectivities of (c) device A and (d) device B.

Table 3 Summary of the optical parameters of device A and device B

Device	Responsivity (A W^−1) UV-illumination	Detectivity (Jones) (× 10^9) at VG = 0 V UV-illumination	Sensitivity (VG = 0 V) UV-illumination	Response time
				τ Rise (s)	τ Fall (s)
Device A	17.4	12.4	0.79	3.41	37.6
Device B	0.48	5.58	224.85	6.22	40

---

The transient photo responses of device A and device B are presented graphically in Fig. 8, showcasing their distinctive trait of swift response onset yet prolonged relaxation times, indicating a long transit time of photogenerated carriers. These experiments have been performed under zero gate bias with a 1 V drain bias with a UV illumination of 0.3 W m⁻². Moreover, the transient response of device A and device B was recorded under prolonged (>2 min) UV illumination of intensity 0.3 W m⁻², as shown in Fig. S14 (ESI†).

Fig. 8 Transient photo response of the TFTs in (a) device A and (b) device B under UV illumination, and corresponding single cycles of the TFTs in (c) device A and (d) device B.

It was observed that the photocurrent in device A reached saturation after approximately 40 seconds, while device B reached saturation after around 35 seconds. Both devices maintained stable photocurrent for about 80 seconds (device A) and 171 seconds (device B). After the removal of UV illumination, both devices recovered within approximately 160 seconds. Consequently, the rise and fall times under prolonged UV illumination were calculated as 40.2 seconds and 165 seconds for device A, and 31.5 seconds and 165 seconds for device B, respectively. The devices exhibit relatively longer response times, which could be attributed to the persistent characteristics of the metal oxide semiconductors (ZnO NS + ZnO sol).^57–59 Interestingly, the devices show a good photoresponse w.r.t other reported metal oxide-based UV photodetectors, as tabulated in Table 4.

Table 4 Comparison of the optical parameters of the device with reported UV photodetectors

Device	Operating voltage (V)	Photosensitivity	Detectivity (Jones)	Responsivity (A W^−1)	Response time		Ref.
					Rise time (s)	Fall time (s)
Si/SiO2/AGZO/Al	15	—	NA	0.26	35.5	55.5	60
Si/SiO2/ZnO/PPR	60	—	NA	0.001034	—		61
SiO2/ZnO/Al	20	—	NA	0.00556	—		62
Si/SiO2/ZnO	50	—	3.35 × 10^10	1.99			63
Li2ZnO/SnO2	2	—	1.1 × 10^10	0.12	7	9	64
Ta2O5/a-IGZO	2	—	NA	4.75	—		65
a-IGZO/(aligned-SnO2-NW)	50	—	NA	8	10		66
Ga2O3/SiO2/a-IGZO	5	—	NA	3.2	—		67
Device II-B	2	224.85	1.24× 10^10	17.4	3.41	37.6	This work

---

3.6. External quantum efficiency

External quantum efficiency (EQE) of both the devices has been measured under 2 V drain bias with zero gate bias, which has been presented in Fig. 9, which clearly shows that both devices do not show any photosensitivity in the visible region. However, both devices have significant EQE values in the UV region. Under such measurement conditions, device A has a higher % EQE (i.e. 8%) as compared to device B, which has an approximate value of 2%.

Fig. 9 (a) External quantum efficiency of (a) device A and (b) device B.

3.7. The operating mechanism of phototransistors

When light illuminates the TFT in accumulation mode, photocurrent can be generated due to the illumination of UV light, which is capable of exciting a valence band electron to the conduction band of the ZnO NSs (Fig. 10(a)). However, visible light photons do not have sufficient energy to excite the valence band electron of ZnO NSs; rather, it transmits through the ZnO NS layer. Besides, it was also observed that the V_th of the TFTs is shifted due to the UV light illumination, which can be explained by the fixed charge generation in the channel. In this specific ZnO TFT, under dark and accumulation mode, there are initially no accumulated holes in the channel. However, under UV illumination, photo-generated holes can become trapped at the dielectric/semiconductor interface due to their immobile nature. This positive fixed charge in the interfacial gate dielectric causes a negative shift in the n-channel TFT's threshold voltage. As light intensity increases, the hole density in the channel also rises, leading to more hole trapping at the interface, which further shifts the V_th. On the other hand, in depletion mode, the energy band of the ZnO semiconductor bends in the opposite direction to the accumulation mode (Fig. 10(b)). Under the dark condition of this operation mode, channel current originates from the DC conductivity of the channel. As soon as UV light is illuminated, additional carriers are generated, which increases this DC conductivity significantly. However, such a larger variation is not observed in the accumulation mode due to the pre-existing high-density accumulated electrons of the channel.

Fig. 10 Energy band diagram of the proposed phototransistor in (a) accumulation mode and (b) depletion mode.

4. Conclusion

In summary, in this contribution, we have synthesized surfactant-free 2D ZnO nanosheets following the wet chemical solution method in an autoclave. The facile and low-cost synthesis method can be scaled up easily. The well-characterized ZnO NSs were employed to fabricate TFTs, exhibiting excellent charge transport properties with a high on/off ratio. The ZnO channel of the TFT annealed at 350 °C showed a carrier mobility of 8.05 cm² V⁻¹ s⁻¹, and the on/off ratio is ∼2 × 10⁵, which are significantly higher compared to the TFT annealed at 110 °C. Moreover, the UV photosensitivity of the higher temperature-annealed TFT is significantly higher than that of the lower temperature TFT, although the EQE value of device A (ZnO NSs annealed at 110 °C) is higher than that of device B (ZnO NSs annealed at 350 °C). The optical response of the TFT under UV illumination exhibits a high shift of threshold voltage towards negative gate voltage, making the device an efficient UV photosensor. The EQE measurement also proves the inertness of the TFT in the visible region. Overall, our study not only shows the facile synthesis of 2D ZnO NSs for the fabrication of efficient TFTs, but also paves the way to explore future research on low-temperature solution-based large-area fabrication of portable, flexible, and transparent devices.

Data availability

The data that support the results of this study are available via the following link: https://drive.google.com/drive/u/4/folders/1kl6uL-e2EQOFDAiyD0XZOYGQQcTgcjQ6.

Conflicts of interest

The authors declare no competing financial or non-financial interests.

Acknowledgements

This work is supported by ANRF, India. SS acknowledges the Start-up Research Grant, Science and Engineering Research Board (SERB) for financial support (file no. SRG/2022/001389), and S. Sadhu acknowledges the support from the DST Inspire Faculty Award (DST/INSPIRE/04/2021/000742). The authors extend their gratitude to the Central Instrument Facility Centre at IIT (BHU), Varanasi, India, for providing access to atomic force microscopy (AFM) measurement facilities and to CIF at the School of Materials Science and Technology IIT (BHU) for facilitating GIXRD characterization of the films. Akhilesh Kumar Yadav and Utkarsh Pandey acknowledge IIT (BHU) for the provision of the PhD Fellowship.

References

1. Y. Lei, T. Zhang, Y. C. Lin, T. Granzier-Nakajima, G. Bepete, D. A. Kowalczyk, Z. Lin, D. Zhou, T. F. Schranghamer, A. Dodda, A. Sebastian, Y. Chen, Y. Liu, G. Pourtois, T. J. Kempa, B. Schuler, M. T. Edmonds, S. Y. Quek, U. Wurstbauer, S. M. Wu, N. R. Glavin, S. Das, S. P. Dash, J. M. Redwing, J. A. Robinson and M. Terrones, ACS Nanosci. Au, 2022, 2, 450–485.
2. L. Cheng, C. Zhang and Y. Liu, J. Am. Chem. Soc., 2019, 141, 16296–16302.
3. F. Schwierz, J. Pezoldt and R. Granzner, Nanoscale, 2015, 7, 8261–8283.
4. D. S. Schulman, A. J. Arnold and S. Das, Chem. Soc. Rev., 2018, 47, 3037–3058.
5. J. Meng, T. Wang, H. Zhu, L. Ji, W. Bao, P. Zhou, L. Chen, Q. Q. Sun and D. W. Zhang, Nano Lett., 2022, 22, 81–89.
6. E. Kahn, M. Liu, T. Zhang, H. Liu, K. Fujisawa, G. Bepete, P. M. Ajayan and M. Terrones, Mater. Today, 2020, 37, 74–92.
7. Q. Zhang, C. Liu and P. Zhou, iScience, 2023, 26, 106673.
8. F. Xue, X. He, J. R. D. Retamal, A. Han, J. Zhang, Z. Liu, J. K. Huang, W. Hu, V. Tung, J. H. He, L. J. Li and X. Zhang, Adv. Mater., 2019, 31, e1901300.
9. Z. Zhang, S. Wang, C. Liu, R. Xie, W. Hu and P. Zhou, Nat. Nanotechnol., 2022, 17, 27–32.
10. A. K. Katiyar, A. T. Hoang, D. Xu, J. Hong, B. J. Kim, S. Ji and J. H. Ahn, Chem. Rev., 2024, 124, 318–419.
11. J. H. Shin, J. Kwon, J. U. Kim, H. Ryu, J. Ok, S. Joon Kwon, H. Park and T.-I. Kim, npj Flexible Electron., 2022, 6, 32.
12. N. Li, Q. Wang, C. Shen, Z. Wei, H. Yu, J. Zhao, X. Lu, G. Wang, C. He, L. Xie, J. Zhu, L. Du, R. Yang, D. Shi and G. Zhang, Nat. Electron., 2020, 3, 711–717.
13. Y. Xiong, D. Xu, Y. Feng, G. Zhang, P. Lin and X. Chen, Adv. Mater., 2023, 35, 2206939.
14. R. Frisenda, A. J. Molina-Mendoza, T. Mueller, A. Castellanos-Gomez and H. S. J. van der Zant, Chem. Soc. Rev., 2018, 47, 3339–3358.
15. J. Puebla, J. Kim, K. Kondou and Y. Otani, Commun. Mater., 2020, 1, 24.
16. X. Hu, K. Liu, Y. Cai, S.-Q. Zang and T. Zhai, Small Sci., 2022, 2, 2200008.
17. F. Wang, J. H. Seo, G. Luo, M. B. Starr, Z. Li, D. Geng, X. Yin, S. Wang, D. G. Fraser, D. Morgan, Z. Ma and X. Wang, Nat. Commun., 2016, 7, 10444.
18. V. Krishnamurthi, T. Ahmed, M. Mohiuddin, A. Zavabeti, N. Pillai, C. F. McConville, N. Mahmood and S. Walia, Adv. Opt. Mater., 2021, 9, 2100449.
19. B. Y. Zhang, K. Xu, Q. Yao, A. Jannat, G. Ren, M. R. Field, X. Wen, C. Zhou, A. Zavabeti and J. Z. Ou, Nat. Mater., 2021, 20, 1073–1078.
20. S. Y. Kim, S.-i Kim, M. K. Kim, J. Kim, S. Mizusaki, D.-S. Ko, C. Jung, D.-J. Yun, J. W. Roh, H.-S. Kim, H. Sohn, J.-H. Lim, J.-M. Oh, H. M. Jeong and W. H. Shin, Inorg. Chem. Front., 2021, 8, 4482–4487.
21. S. A. Patil, P. B. Jagdale, A. Singh, R. V. Singh, Z. Khan, A. K. Samal and M. Saxena, Small, 2023, 19, e2206063.
22. F. H. Alshammari, P. K. Nayak, Z. Wang and H. N. Alshareef, ACS Appl. Mater. Interfaces, 2016, 8, 22751–22755.
23. L. Jiang, J. Li, K. Huang, S. Li, Q. Wang, Z. Sun, T. Mei, J. Wang, L. Zhang, N. Wang and X. Wang, ACS Omega, 2017, 2, 8990–8996.
24. J. F. Wager, Inf. Disp., 2016, 32, 16–21.
25. X. Yu, T. J. Marks and A. Facchetti, Nat. Mater., 2016, 15, 383–396.
26. Z. Chai, S. A. Abbasi and A. A. Busnaina, ACS Appl. Electron. Mater., 2023, 5, 2328–2337.
27. S. T. Meyers, J. T. Anderson, C. M. Hung, J. Thompson, J. F. Wager and D. A. Keszler, J. Am. Chem. Soc., 2008, 130, 17603–17609.
28. D. Darwis, N. Holmes, D. Elkington, A. L. David Kilcoyne, G. Bryant, X. Zhou, P. Dastoor and W. Belcher, Sol. Energy Mater. Sol. Cells, 2014, 121, 99–107.
29. S. Iravani, RSC Sustainability, 2023, 1, 72–82.
30. M. Ameri, M. F. Al-Mudhaffer, F. Almyahi, G. C. Fardell, M. Marks, A. Al-Ahmad, A. Fahy, T. Andersen, D. C. Elkington, K. Feron, M. Dickinson, F. Samavat, P. C. Dastoor and M. J. Griffith, ACS Appl. Mater. Interfaces, 2019, 11, 10074–10088.
31. A. Bhuin, S. Udayakumar, J. Gopalarethinam, D. Mukherjee, K. Girigoswami, C. Ponraj and S. Sarkar, Sci. Rep., 2024, 14, 10406.
32. U. Pandey, N. K. Chourasia, N. Pal, S. Biring and B. N. Pal, IEEE Trans. Electron Devices, 2022, 69, 1077–1082.
33. N. Pal, B. Thakurta, R. Chakraborty, U. Pandey, V. Acharya, S. Biring, M. Pal and B. N. Pal, J. Mater. Chem. C, 2022, 10, 14905–14914.
34. U. Pandey, A. K. Yadav, N. Pal, P. K. Aich and B. N. Pal, J. Mater. Chem. C, 2023, 11, 15276–15287.
35. S. Venkatesan, S. Suresh, P. Ramu, J. Arumugam, S. Thambidurai and N. Pugazhenthiran, Results Chem., 2022, 4, 100637.
36. J. I. Langford, R. J. Cernik and D. Louër, J. Appl. Crystallogr., 1991, 24, 913–919.
37. E. Fortunato, P. Barquinha and R. Martins, Adv. Mater., 2012, 24, 2945–2986.
38. M. Khan, P. Ware and N. Shimpi, SN Appl. Sci., 2021, 3, 528.
39. W. Muhammad, N. Ullah, M. Haroon and B. H. Abbasi, RSC Adv., 2019, 9, 29541–29548.
40. P. Makuła, M. Pacia and W. Macyk, J. Phys. Chem. Lett., 2018, 9, 6814–6817.
41. M. Silambarasan, S. Shanmugam and T. Soga, Int. J. ChemTech Res., 2015, 7, 1644–1650.
42. S. Guo, Z. Du and S. Dai, Phys. Status Solidi B, 2009, 246, 2329–2332.
43. Y. Song, S. Zhang, C. Zhang, Y. Yang and K. Lv, Crystals, 2019, 9, 395.
44. S. Gandla, S. R. Gollu, R. Sharma, V. Sarangi and D. Gupta, Appl. Phys. Lett., 2015, 107, 152102.
45. X. Li, X. Li, B. Zhu, J. Wang, H. Lan and X. Chen, RSC Adv., 2017, 7, 30956–30962.
46. V. Acharya, N. Pal, A. Sharma, U. Pandey, M. Suthar, P. Kumar Roy, S. Biring and B. N. Pal, Mater. Sci. Eng., B, 2023, 289, 116270.
47. U. Pandey, N. Pal, V. Acharya, P. K. Aich, A. K. Yadav and B. N. Pal, IEEE Sens. J., 2023, 23, 11482–11489.
48. I. R. Kaufmann, O. Zerey, T. Meyers, J. Reker, F. Vidor and U. Hilleringmann, Nanomaterials, 2021, 11, 1188.
49. A. M. Ma, M. Shen, A. Afshar, Y. Y. Tsui, K. C. Cadien and D. W. Barlage, IEEE Trans. Electron Devices, 2016, 63, 3540–3546.
50. T. Hirao, M. Furuta, T. Hiramatsu, T. Matsuda, C. Li, H. Furuta, H. Hokari, M. Yoshida, H. Ishii and M. Kakegawa, IEEE Trans. Electron Devices, 2008, 55, 3136–3142.
51. B. S. Ong, C. Li, Y. Li, Y. Wu and R. Loutfy, J. Am. Chem. Soc., 2007, 129, 2750–2751.
52. L. Zhang, H. Zhang, Y. Bai, J. W. Ma, J. Cao, X. Jiang and Z. L. Zhang, Solid State Commun., 2008, 146, 387–390.
53. P. Carcia, R. McLean, M. Reilly and G. Nunes Jr, Appl. Phys. Lett., 2003, 82, 1117–1119.
54. S. Lee, Y. Jeong, S. Jeong, J. Lee, M. Jeon and J. Moon, Superlattices Microstruct., 2008, 44, 761–769.
55. S. Kwon, S. Bang, S. Lee, S. Jeon, W. Jeong, H. Kim, S. C. Gong, H. J. Chang, H.-H. Park and H. Jeon, Semicond. Sci. Technol., 2009, 24, 035015.
56. E. Fortunato, P. Barquinha and R. Martins, Adv. Mater., 2012, 24, 2945–2986.
57. T. A. Heuser, C. A. Chapin, M. A. Holliday, Y. Wang and D. G. Senesky, J. Appl. Phys., 2022, 131, 155701.
58. H. Yoo, I. S. Lee, S. Jung, S. M. Rho, B. H. Kang and H. J. Kim, Adv. Mater., 2021, 33, 2006091.
59. U. Pandey, N. Pal, A. Ghosh, S. Suman, S. Biring and B. N. Pal, Nanoscale, 2024, 16, 8504–8513.
60. W.-L. Huang, S.-P. Chang, C.-H. Li and S.-J. Chang, Electronics, 2021, 10, 631.
61. B. S. Sannakashappanavar, A. B. Yadav, K. Singh and T. A. Sai, Micro Nanostruct., 2023, 179, 207581.
62. S. Park, B. J. Kim, S. J. Kang and N.-K. Cho, J. Korean Phys. Soc., 2018, 73, 1351–1355.
63. B. S. Sannakashappanavar, A. B. Yadav, C. Byrareddy and N. Narasimha Murty, J. Electron. Mater., 2020, 49, 5272–5280.
64. A. K. Singh, N. K. Chourasia, B. N. Pal, A. Pandey and P. Chakrabarti, IEEE Trans. Electron Devices, 2020, 67, 2028–2034.
65. T. Chang, C. Chiu, W. Weng, S. Chang, T. Tsai and Z. Huang, Appl. Phys. Lett., 2012, 101, 261112.
66. X. Liu, X. Liu, J. Wang, C. Liao, X. Xiao, S. Guo, C. Jiang, Z. Fan, T. Wang and X. Chen, Adv. Mater., 2014, 26, 7399–7404.
67. T. Chang, C. Chiu, S. Chang, T. Tsai, T. Yang, Z. Huang and W. Weng, Appl. Phys. Lett., 2013, 102, 221104.

---

Footnotes

† Electronic supplementary information (ESI) available: Schematic representation of solution-processed TFTs, p-XRD pattern of the as-synthesized ZnO NS, UV-vis absorbance and reflectance spectra, XPS survey spectrum, capacitance vs. frequency curve of the LiInSnO₄ dielectric, XRD and AFM images of the (a) 2D and (b) 3D LiInSnO₄ dielectric, transmittance vs. wavelength (c), Tauc plot of the LiInSnO₄ dielectric, output characteristics of reference devices at different temperatures, transfer characteristics under UV light illumination, transient photo response, external quantum efficiency, voltage vs. intensity variation under UV illumination, and TFT parameters of the devices. See DOI: https://doi.org/10.1039/d4tc05180g

‡ These authors contributed equally.

---