Simulation study of ZnO nanowire FET arrays for photosensitivity enhancement of UV photodetectors

Accepted Manuscript Simulation study of ZnO nanowire FET arrays for photosensitivity enhancement of UV photodetectors Z. Golshan Bafghi, N. Manavizadeh PII:

S0749-6036(18)31576-3

DOI:

10.1016/j.spmi.2018.08.028

Reference:

YSPMI 5873

To appear in:

Superlattices and Microstructures

Received Date: 1 August 2018 Revised Date:

23 August 2018

Accepted Date: 30 August 2018

Please cite this article as: Z.G. Bafghi, N. Manavizadeh, Simulation study of ZnO nanowire FET arrays for photosensitivity enhancement of UV photodetectors, Superlattices and Microstructures (2018), doi: 10.1016/j.spmi.2018.08.028. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Simulation Study of ZnO Nanowire FET Arrays for Photosensitivity Enhancement of UV Photodetectors

RI PT

Z. Golshan Bafghi, N. Manavizadeh* Faculty of Electrical Engineering, K. N. Toosi University of Technology, Tehran 1631714191, Iran. * [email protected]

SC

Highlights:

M AN U

• Three configurations of ZnO nanowire field effect transistor based ultraviolet photodetector is studied. • The impact of parallel and series structure is investigated on detector’s performance. • Absorption percentage and External Quantum Efficiency increase in the array configuration. • Photosensitivity increases in the parallel structure. • Response times of the detectors are in a range of µs. • Array structure stables response of photodetector in a variety of illumination angle.

AC C

EP

TE D

Abstract An ultraviolet photodetector based on zinc oxide nanowire field effect transistor is designed and simulated in three configurations including single nanowire, three parallel and array of six nanowires. Transient response of photodetectors is studied by mixed-mode simulation. Results demonstrate that the light/dark current ratio is increased to 10 for the parallel and array of six nanowires. Transmission percentage of undesired wavelengths increases to 80% in the array of six nanowires, additionally. Furthermore, parallel nanowires detectors enhance photosensitivity from 3.325 in single nanowire structure to 12.477 in parallel structure; however, it does not impact on quantum efficiency. On the other hand, the array of six nanowires structure raises external quantum efficiency substantially from 50% for single nanowire to 58%. Moreover, array structure stables the photodetector in a variety of illumination angles. Although, field effect transistor photodetector structure improves the performance of UV detector and boost photodetector response time, three parallel and array of six nanowires structures slightly increase response time to 8.4 and 9.9µs, respectively in comparison with single nanowire structure. The proposed UV detector receives an encouraging response to µW range of power light intensity. Keyword— UV photodetector, ZnO nanowire FET, Array structure, Quantum efficiency, Transient response, illuminating angle, Photosensitivity.

ACCEPTED MANUSCRIPT I. INTRODUCTION

AC C

EP

TE D

M AN U

SC

RI PT

Ultraviolet (UV) photodetectors are utilized in many commercial applications such as environmental monitoring (Ozone and pollution), water purification, high temperature flame detection, secure space communication and missile warning system [1]-[5]. Room temperature UV detectors have appealed to defense technologies, aerospace techniques, biological analysis and flame warning [6]. Considering unique physical, mechanical and electrical properties of Zinc Oxide; ZnO has become one of the proper materials for room temperature UV detector applications. ZnO with a wide band gap (3.3 eV) in room temperature, high exciton binding energy (60 meV) has led to a large number of researches working on fabrication of UV-range photonic and optoelectronic devices [7],[8]. A wide band gap semiconductor has many merits for high temperature and power operations, reducing electronic noise, raising breakdown voltages and sustainability in high electric fields [9]. ZnO which is a biocompatible and environment-friendly material plays a prominent role in state-of-the-art UV photodetectors for the two reasons: (1) There are different methods and technologies to synthesize and fabricate ZnO. (2) ZnO surface can adsorb oxygen molecules and a low-conductivity depletion layer is made up [10]. Moreover, the use of low dimensional (1D) wide band gap inorganic semiconductor nanostructures materials have attracted enormous attention in the fabrication of electronic devices. The electrical integration of synthesized 1D nanostructures has been achieved with lithography, which promises high speeds and greater device versatility [11], [12]. 1D nanostructures such as nanorod (NR), nanowire (NW), nanotubes (NT) and nanobelts enable their utilization as functional building blocks to build photodetectors with distinctive spectral selectivity and sensitivity. Therefore, in recent years, various types of 1D metal oxide nanostructures including binary oxides (such as ZnO, SnO2, Ga2O3, Nb2O5, TiO2 and WO3) and ternary oxides (such as Zn2SnO4, ZnGa2O4, Zn2GeO4, In2Ge2O7) have been successfully constructed into photodetectors [13]-[19]. Specially, nanowire devices among all ZnO nanostructures are attractive due to fabricating high-density devices. There are different methods to synthesize ZnO NWs such as chemical vapor deposition (CVD) [20], Molecular beam epitaxy [21], Electro-chemical deposition [22] and Hydrothermal. Not only is ZnO a low-cost material, but also ZnO NWs can be grown easily by any technique, especially hydrothermal techniques such as chemical bath deposition (CBD) with high density [23]. Hence, zinc oxide NWs have wide applications for lasers [24], light emitting devices (LEDs) [25], piezo-electronics [26], solar cells [27], spintronics [28], field effect transistors (FET), transparent and flexible electronics [29] and UV sensors [30] as well. In addition, researches have reported different structures for ZnO nanowire UV photodetectors including metal-semiconductor-metal [31]-[33], Schottky diode [34], p-n junction diodes [35], [36] and field effect transistors [37], [38]. Among all reported structures for ZnO-based UV detectors, FET-based technology is a promising variation to conquer with undesired high-power consumption for high performance submicron optical devices, owing to the superior electrical and optical properties offered by CMOS technology, both in analog and digital applications [39], [40]. In contrast, ZnO nanostructures demonstrate high photoresponse due to increase in surface-to-volume ratio [5], [41]. Various ZnO-based UV photodetectors have been announced with high photo-responsivity, however, the typical response and recovery times of these photodetectors are ranging from several tens of microseconds to few hours in most cases [36],[41]-[44]. In this paper, a UV photodetector based on ZnO NW field effect transistors arrays are designed and simulated in 3D environment. The UV photodetector based on back-gate FET is investigated in three configurations including single ZnO NW FET, three parallel NW FETs and array of six nanowires in the channel of field effect transistor. Important parameters such as dark and illumination currents, spectral response, source and available photocurrent, photosensitivity, angular dependence on responsivity and external quantum efficiency are studied. Moreover, UV sensor response/recovery time is evaluated in mixed-mode environment for all three structures.

ACCEPTED MANUSCRIPT

II. DEVICE STRUCTURE, SIMULATION APPROACH AND THEORY

M AN U

SC

RI PT

As shown in Fig. 1, n-type silicon is considered as the transistor gate with thickness of 30 nm. The thickness of SiO2 layer as a gate dielectric is 20 nm in all three structures. ZnO nanowires can be synthesis in different shapes such as square, hexagonal, star-shaped, etc. [45]-[49]. In order to simulate devices in Cartesian coordinate, ZnO nanowires have a square cross-section with the dimensions 50nm × 50nm and length of 1µm. As shown in Fig. 1 (b) and (c), the gaps between each NW in array structures is 15 nm. As ZnO is inherently an n-type material, nanowires concentration is considered about 1E+18 cm-3, also source and drain electrode’s length is 1 18 of nanowires’ length due to its occupation of minimum exposed area of nanowires. However, the transistor is biased in subthreshold region, thus the channel of transistor depletes from carries and the dark current decreases. Under illumination of UV spectrum, excitons which are created in ZnO NW increase the current and the concentration of carriers saturates in the transistor channel, therefore the illumination current is made independently of varying voltage. Parallel UV photodetector structure is made up of three single nanowire FETs located in parallel (Fig. 1(b)). Gold is utilized as source and drain electrodes and all three electrodes are considered throughout nanowires. The array structure, shown in Fig. 1(c), is made up of two parallel structures (Fig. 1(b)) in series. In all simulations, the incident beam wavelength is considered about 380 nm which is in UV spectrum and its radiation angle is 90° and temperature is considered 300 K. Furthermore, to simulate devices, Shockley-Read-Hall (SRH) concentration-dependent lifetime and SRH recombination models are considered as (1); − = !1" − + exp + + exp where,

(b)

(c)

AC C

EP

TE D

(a)

Gap Figure 1: UV photodetectors based on ZnO nanowire FET structures. (a) Single nanowire FET, (b) Three parallel nanowires FET, (c) Array of six nanowires FET.

#$0

=

ACCEPTED MANUSCRIPT

$'(')* $'(')* $ + &$ $+ ,$ + -$ $+ ,$ # 0

345 6

= #7$8

−

9 + #7 8

−

RI PT

!3" $'(')* $'(')* .1 + & $+ , + - $+ , where n and p are carrier concentration in semiconductors, ETRAP is the difference between the trap energy level and the intrinsic Fermi level, TL is the lattice temperature on degrees Kelvin, nie is the effective intrinsic concentration and TAUN0 and TAUP0 are the electron and hole lifetime, respectively and Ntotal is the local impurity concentration [50]. In addition, AN, BN, AP, BP are ionization coefficients and CP, CN, EP and EN are lifetime coefficient while both coefficients depend on material. NSRHP and NSRHN are material dependent SRH lifetime Defaults, as well. Moreover, Auger generation and recombination models are regarded as (4) to consider three particles transition which is according to a mobile carrier is either captured or omitted [51]. 9 !4"

SC

=

./ !2"

Δ exp ;

=

5

TE D

M AN U

where AUGN and AUGP are Auger coefficients. Once the magnitude of electric field becomes significant, carriers are accelerated in the electric field and the velocity will begin to saturate. This effect leads to reduce the effective mobility since the magnitude of the drift velocity is the product of the mobility and the electric field component in the direction of the current flow. The Caughey and Thomas Expressions [52] are used to implement an electrical field-dependent and concentration dependent mobility physical models. Moreover, in presence of heavy doping, greater than 1E+18 cm-3, experimental work has shown that the pn product in silicon becomes doping dependent [53]. As the doping level increases, a decrease in band gap separation occurs, where the conduction band is lowered by approximately the same amount as the valence band is raised. The band gap narrowing model is simulated by a spatially varying intrinsic concentration nie defined according to (5), = !5"

where,

= &7$. @ln

AC C

5

EP

$ $ E + [!ln " + &7$. -] F !6" &7$. $ &7$. $ that BGN.E, BGN.N and BGN.C are 9E-3, 1E+17 and 0.5 respectively in Slotboom set of defaults for use in (6). Finally, Fermi-Dirac carrier statistic model is regarded as well. Newton method is picked up to obtain desirable results in solving equations as well. The NewtonRichardson method is a variant of the Newton iteration that calculates a new version of the coefficient matrix. To Study the performance of photodetector, some significant parameters should be investigated. Quantum efficiency is one of the critical parameters. External quantum efficiency is obtained from (7), I3JK LKML . H. = !7" I N46O where Available photocurrent is the amount of photon absorbed by the device expressed by current density. Source photocurrent is the amount of current generated by the light source and obtained from (8); & R I N46O = Q U !8" ℎT V where Bn is the intensity of the beam, λ is the source wavelength, h is Planck's constant, c is the speed of light and Wt is the width of the beam [54]. Δ

ACCEPTED MANUSCRIPT

RI PT

Furthermore, Photosensitivity is the other key parameter that should be studied for photodetectors which is obtained from (9): I_LL4` KV N a IbK6c ℎ('(WX WY'YZY'[ = logE^! " !9" IbK6c Moreover, the electrical field vectors are conservative and determined by the gradient of the potential of each location (x,y,z) and magnetic field which are called Faraday’s and Poisson’s law (10), j k ef = −∇!Ψ ∓ * " − !10" Q k' where Ψ is potential, nie is intrinsic carrier concentration, TL is electron thermal equilibrium temperature and K is Boltzman’s constant. In addition, A is a magnetic vector which derives magnetic field B that defines as & = ∇ × f [51].

SC

III. RESULT AND DISCUSSION

AC C

EP

TE D

M AN U

As a general rule, UV photodetectors based on diode structures perform in reverse region where reversed biased saturation current is measured and affected by illumination [29], [36]. For UV detectors based on ZnO NW FET, the device should be biased in sub-threshold region to have fewer carriers in transistor’s channel. To find the appropriate operational point for each UV photodetector, Ib − mn diagrams in dark and illumination conditions have been obtained. As found out in Fig. 2, the back-gate voltages are swept from -9 to 5 V and drain-source voltages are considered constant at 5 V for all structures. According to Fig. 2(a), the proper gate voltage to operate in sub-threshold region is lower than -6.5 V where light and dark currents are considerably different. For three parallel and array of six NWs structures, the gate voltage should be lower than -5 V, as shown in Fig. 2(b) and (c). Hence, the greater the number of parallel nanowires in UV photodetector, the lower the biasing voltage is. In addition, as the insets of Fig. 2 show by illuminating ultraviolet, the absolute value of threshold voltage of transistors decreases.

RI PT

ACCEPTED MANUSCRIPT

Figure 3: Drain Current – Drain Voltage in different gate voltages. (a) Single NW, (b) Three parallel NWs, and (c) Array of six NWs. The inset shows light currents in Single nanowire structure. The light/dark current ratio is greater in three parallel and array of six nanowires structures due to the lower dark current. The channel saturation makes light current fairly independent of gate voltage.

AC C

EP

TE D

M AN U

SC

Moreover, the drain current in dark and illumination conditions is indicated in Fig. 3 in which transistors are biased at three voltages of Vg= -7.5, -7.75, -8V where the difference between current values is considerable. The order of light current due to incident light with 1 W/cm2 in all three structures is fairly independent of gate voltage as a result of saturation of the density of carriers that exist in transistor’s channel. The inset of Fig. 3(a) shows the difference in light currents order in various gate voltages in single nanowire structure. Fig. 3(b) and (c) demonstrate that drain current of parallel structure is 8 orders of magnitude less than that of single NW FET at Vg= -7.5V in dark condition. As a result of applying voltage to device, carrier concentrations which are gathered near silicon/SiO2 interface is not uniform in the detector width, so the induced carrier to the channel of transistors is not the same in all FETs. Fig. 4(a) indicates electron concentration and electrical field vectors in silicon back-gate of three parallel NWs structure in distance of 1nm under interface of Si/SiO2 while Vg= -8 V and Vds= 5V, the middle part of silicon has the least electron concentration and the concentration of electrons near the edge is higher. Electron concentration contours and electrical field vectors in ZnO NWs, 1nm under electrodes, are shown in Fig. 4(b) where the channel of transistor is placed. Electron concentration decreases totally in the channel of transistor; especially in the NW which is located at the middle. In addition, electric field is in the z-direction while the applied voltage to the drain-source terminal generates electric field in x-direction. Since there is an electric field along z-axis and the current is in x-direction, there should be another force imposed on electrons causing them to move in a different direction and away from the center. As the result of Faraday’s law of induction, the induced magnetic field from charged particles movement, there exists a magnetic field around the device which causes the discrepancy of the electrons concentration in nanowires. According to the Hall Effect coefficient, if there exists a magnetic field along y-axis and an electric field along z-axis, then there will be an electric current along x-axis (10). 10-5

20

10-10 10-15 10

10

-20

0 -9 -6 -3 0

10-25 10

-30

10

-35

(c)

-9 -7 -5 -3 -1 1 VG [V]

3 Dark Light

3

Figure 2: Drain Current – Gate Voltage characteristics of UV photodetectors in dark and illumination conditions. (a) Single nanowire, (b) Three parallel nanowires, and (c) Array of six nanowires. The proper biasing point is lower in parallel structure. Insets show threshold voltage in dark and illumination conditions.

5

ACCEPTED MANUSCRIPT

Er !10" Jt Bv Consequently, applied voltage on nanowires will create an electric current along x-axis and total electric field along z-axis (Fig. 4). This will make a magnetic field in z-direction which causes reduction in electric current in parallel nanowires structures in comparison with single nanowire transistor due to the exerted force on electrons. Moreover, the magnitude of applied magnetic field is not the same for all nanowires due to different distance of NWs to the source of magnetic field, which is producing contours in Fig. 4. Thus, the light/dark current ratio rises in three parallel structure in comparison with single nanowire detectors. Moreover, as a consequence of magnetic field, the electron concentration of middle nanowire decreases and the probability of recombination of photogenerated electron-hole pairs declines, as well. Hence, more electrons could participate in light current. To compare the other characteristics of different arrangements, the gates voltages are considered as -8V.

Figure 6: Absorption diagram. The inset indicates Transmission diagram. The array of six nanowires transmits unsatisfied wavelengths.

TE D

Figure 5: Spectral Response of UV photodetectors. The behavior of spectral response is independent of different structures.

M AN U

SC

RI PT

Rp =

AC C

EP

Photo-illumination spectra of all three configurations of photodetector exposed to UV light with 1 W/cm2 intensity is shown in Fig. 5. Although the maximum photocurrent has been obtained between 200-390 nm wavelengths, photocurrent does not vary significantly in the rest of wavelengths. Accordingly, ZnO NW array structures follow the trend of single NW’s spectral response with increasing in photocurrent. Fig. 6 indicates absorption and transmission percentage of light with different wavelengths and with 1 W/cm2 intensity. A single NW UV photodetector approximately absorbs 65% of UV light as well as parallel structure. The absorption of 400-900 nm wavelength reduces significantly in array of

Figure 4: Electron concentration and electric field vector in (a) Silicon and (b) ZnO nanowires in three parallel nanowires structure. The electron concentration is at the least in the middle nanowire. As a result, the dark current declines in parallel structures.

Figure 9: Angle dependence of photodetectors. The array of six nanowire structure stables the device in all angles.

SC

Figure 8: Sensitivity of UV photodetectors. The parallel sand array of six nanowire structures raise photosensevity.

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

six NWs structure. Nevertheless, it reduces the percentage of absorption in UV wavelength (Fig. 6). Furthermore, parallel and array structures have a considerable effect on transmission spectrum. As indicated in the inset of Fig. 6, array of six nanowires structure transmits about 80% of spectra with 500-900 nm wavelength while a single NW detector just transmits 20% of that spectra. In order to compare photodetectors, some parameters have to be noticed. As mentioned in Section II, Quantum efficiency is one of the critical parameters to make a comparison between photodetectors. Fig. 7 indicates external quantum efficiency of a UV photodetector in three configurations. Available photocurrent, the numerator of external quantum efficiency, is the amount of photon absorbed by the device expressed by current density which is shown in Fig. 7(a). Maximum available current is between 350-400 nm wavelengths for 1 W/cm2 intensity in all three configurations. Array of six NWs structure have affected available photocurrent significantly and raised it 12 times in comparison with single NW FET UV photodetector due to increase in exposed area of detector. Source photocurrent is the amount of current generated by the light source. Fig. 7(b) shows source photocurrent is raised by increasing in the area of the device that was exposed to light while the other parameters were fixed. External quantum efficiency according to Fig. 7(c) is the same for single NW and parallel structure and it is about 50%. On the other hand, array of six NWs configuration has the maximum of 58% in 300-360 nm wavelength. Furthermore, in three parallel NWs available photocurrent and source photocurrent which depends on beam width increases in the same ratio, as a result, single and three parallel NW structures have the same external quantum efficiency. In fact, by connecting NWs FET

Figure 7: Quantum Efficiency of UV photodetectors. (a) Source photocurrent, (b) Available photocurrent and (c) External Quantum Efficiency. The array of six nanowires structure increases external quantum efficiency owing to impact of available photocurrent significantly.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

photodetectors in series, external quantum efficiency increases due to increase in length of the detector and increase in effective area. Photosensitivity is the other parameter that should be investigated for photodetectors. Fig. 8 indicates the sensitivity of a detector in different wavelengths. Array of six and three parallel NWs structures have better sensitivity due to having more parallel nanowires as transistor channel and fewer dark current. Hence, illumination current rises while dark current reduces by parallel structures and device photosensitivity increases. In contrast, series structure does not have a significant effect on sensitivity. Furthermore, photosensevity is affected by biasing point of transistors considerably. Although quantum efficiency and photosensitivity are critical parameters to compare photodetectors, illumination source angle and beam incidental angles affect all important parameters such as illuminated drain current and available photocurrent following by external quantum efficiency and photosensitivity. The impact of illumination angle on drain current under illumination is investigated in Fig. 9. As seen in the inset of Fig. 9, the angle varies from 0 to 90° and drain current in single NW structure is maximum in 30°. It can be attributed that a fraction of illumination light with 1 W/cm2 intensity meets ZnO/gold junction and turns it to excitons under drain electrode, as a result, the drain current increases in 30° [55]. Moreover, in single and parallel NW structures, drain current decreases considerably in about 80-85° where contacts shading near drain contact is maximum. Conversely, in the array of six NWs structure, drain current is approximately constant and there is not any reduction in drain current curve due to compensation by neighborhood nanowires. Nonetheless, drain current is the most at 10-35° due to having Schottky junction in illumination circumstances.

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 10: Transient response of UV detector to a periodic pulse light. (a) Single nanowire detector (b) Three parallel and (c) Array of six nanowires (d) recovery time for array of six NWs structure. The response time of the photodetector is in microsecond range; however, the parallel structure increases response time slightly.

AC C

EP

TE D

The most important parameter in UV photodetectors is response/recovery time. Response/recovery time of UV detector is affected by carrier concentration of ZnO nanowires, length and width of nanowires, source and drain contact materials and temperature. Fig. 10 indicates transient response of ZnO NW FET UV photodetector in all three configurations. The response time is studied in mixedmode environment and structures are exposed to a pulse light source with 16 µs width, 32 µs period and 1 W/cm2 intensity while gates and drain-source voltages are biased at -8 and 5 V, respectively. Response times of single, three parallel and array of six ZnO NWs FET-based UV detectors are about 4, 8.4 and 9.9µs, respectively, while the transistor’s channel is not saturated. Results demonstrate that response time of proposed UV detectors is less than those of structures reported in [31]-[36]. Recovery time of UV detector is considered when the current of detector reaches 10 percent of initial value and is about 100s (Fig. 10(d)). Although FET-based UV photodetectors have a long recovery time, it can be decreased by reducing concentration of ZnO and silicon (e.g. 1s for single NW structure), subsequently, photosensitivity decreases. It is due to substantial carrier concentration in transistor’s channel and it takes time to deplete considerably and drain current drops when the light

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

source is switched off; however, this ZnO NW FET-based UV detector is remarkably sensitive even in low light intensities. Fig. 11 indicates and summarizes drain photocurrent of three FET-based structures in different light intensities while gate voltages are biased at -8V and drain voltages are swept from 0 to 5 V. As observed, the drain current and photosensitivity of single nanowire structure are 2 µA and 7.85 in 1 µW/cm2 and 2.75µA and 7.88 in 1 W/cm2 intensity, respectively. Results revealed that the response and photosensitivity of ZnO NW FET-based UV photodetector in all three structures are the same. Consequently, our designed UV detectors based on ZnO NW FET arrays have a considerable sensing even in µWatt light intensity. Table I compares the current ratio, sensitivity, external quantum efficiency and response/recovery time of UV photodetectors presented in literature and this work. In proposed array FET detector, the dark current reduces dramatically and hence the light/dark current ratio rises greatly. The photosensitivity of array structure of ZnO nanowire FET obtained 12.47 which is greater than those of the other structures in Table I. FET-based detector boosts quantum efficiency to above 50 %, especially for array structure that reaches 58% greater than those of the other mentioned efficiencies. The response time of the proposed photodetector markedly decline to µs; however, its recovery time is greater than p-n hetero-structure detectors (Keramatnejad et al. 2014). The presented results corroborate the effectiveness of array structure to enhance the performance of FET-based UV photodetectors. Although the array structure raises response time infinitesimally comparing with the other FET structures, it improves sensitivity, quantum efficiency and stability of FET-based UV detector. In Fact, the array of six nanowires structure improves the detector performance in different aspects such a decrease proper biasing voltage, increase the light/dark current ratio due to decrease dark current, raise external quantum efficiency as a result of increase in effective exposed area and the number of absorbed photon and stables the detector in different illumination angles. Furthermore, the proposed FET based photodetector has small response time in µs range.

3.5 (c)

3 2.5 2 1.5

1 W/cm2 2 1 mW/cm 1 W/cm2

1 0.5 0

0

1

2 3 VD [V]

4

5

Figure 11: UV photodetector response to different light intensity. (a) Single nanowire, (b) Three parallel nanowires, and (c) Array of six nanowires structures. The proposed photodetector is sensitive to UV spectrum even in µw power in all three structures.

ACCEPTED MANUSCRIPT Table I: Comparison of critical parameters of our work with literature. Structure

Dark current [A]

Current ratio

Sensitivity

EQE

Response time

Recovery time

[23]

ZnO NW FET

1E-10

1.01E4

-

-

-

-

[35]

n-ZnO NW/P-Si

5.5E-8

5.2

3.72

-

7s

9s

[33]

Mn doped ZnO NR diode

2E-5

6

-

-

7s

29s

[31]

ZnO NR

7.28E-7

2.95

-

[32]

ZnO NW

1.05E-4

2.23

-

[13]

ZnO NW

1E-5

6.2

-

[36]

ZnO/Sb_ZnO NW

-

-

[12]

ZnO NW

-

-

This work

Single ZnO NW FET

1.3E-9

8.5E2

This work

Three parallel NW FET

1.1E-16

This work

Array of ZnO NW FET

1.1E-16

-

-

-

11s

43s

-

9.1s

56s

SC

49.5

1.43

-

30ms

30ms

2.85

25

23s

73s

3.325

51

4µs

105s

3E10

12.465

50

8.4µs

105s

3E10

12.47

58

9.9µs

105s

M AN U

TE D

IV. CONCLUSION

RI PT

Ref.

AC C

EP

An ultraviolet photodetector based on ZnO nanowire field effect transistor have been designed and simulated in 3D environment and the device transient response is studied in mixed-mode environment of TCAD tools. Three structures based on FET photodetectors are studied including single, three parallel and array of six ZnO NW FET structures. Parallel structure affects the boundary of appropriate gate voltage; the greater the number of parallel nanowires in UV photodetector is, the less the biasing voltage in reversed bias is. Furthermore, parallel structure reduces the order of drain current in dark environment which is following an increase in Illumination / I Dark ratio. Absorption and transmission spectrum have been substantially impacted by parallel and series structure of nanowires in FET based UV photodetector and they rise the percentage of transmission to 80% in λ > 400nm. Moreover, series structure increases external quantum efficiency to 58% due to increase in available photocurrent considerably. On the other aspect, parallel structure raises photosensitivity of detector to 12.477 in comparison with a single nanowire; however, series structure does not have a tremendous effect on photosensitivity. The great merit of array structure is compensating the response of detector’s drain current in various angles and maintain the drain current constant at 3.3µA. Simulation results show that, the response times are 9.9, 8.4 and 4 µs for array of six NWs, three parallel NWs and single NW structures, respectively. Furthermore, recovery time of detector is about 100s. In conclusion, array structure improves the performance of UV detector on account of increasing external quantum efficiency and photosensitivity while it increases the response time slightly. This structure stables the detector’s response in a variety of illumination angles and the detector have

ACCEPTED MANUSCRIPT desired current response even in µWatt light intensity. Results indicate array structure in FET-based photodetectors can streamline the performance and efficiency of nanowire-based detectors; however, it raises the response time slightly.

V. REFERENCE

AC C

EP

TE D

M AN U

SC

RI PT

[1] C. S. Rout, M. Hedge, A. Govindaraj, Ammonia sensor based on metal oxide nanostructures, Nanotech. 18 (2007) 205504. [2] H. S. Bae, S. Im, Ultraviolet detecting properties of ZnO-based thin film transistors, Thin Solid Film 469 (2004) 75-79. [3] M. Liu, H. K. Kim, Ultraviolet Detection with Ultrathin ZnO Epitaxial Films Treated with Oxygen Plasma, Appl. Phys. Lett. 84 (2004) 173–175. [4] A. B. Yadav, A. Pandey, D. Somvanshi, S. Jit, Sol-Gel-Based Highly Sensitive Pd/N-ZnO Thin Film/N-Si Schottky Ultraviolet Photodiodes, IEEE Trans. on Electron Devices 62 (2015) 18791884. [5] M. R. Alenezi, S. J. Henley, S. R. P. Silva, On-Chip Fabrication of High Performance Nanostructured ZnO UV Detectors, Sci, Rep. 5 (2015) 8516. [6] E. Monroy, F. Omnès, F. Calle, Wide-Bandgap Semiconductor Ultraviolet Photodetectors, semiconductor Sci. and Tech 18 (2003) 4. [7] D. Zhang, X. Gu, F. Jing, F. Gao, J. Zhou, S. Ruan, High Performance Ultraviolet Detector Based on TiO 2 /ZnO heterojunction, J. Alloy and Comp. 618 (2015) 551-554. [8] L. Yang , W. Liu, H. Xu, J. Ma, C. Zhang, C. Liu, Z. Wang, Y. Liu, Enhanced near-UV electroluminescence from p-GaN/i-Al2O3/n-ZnO heterojunction LEDs by optimizing the insulator thickness and introducing surface plasmons of Ag nanowires, J. Mater. Chem. C 5 (2017) 3288-3295. [9] P. Pattanaik, S. K. Kamilla, D. P. Das, D. K. Mishra, Experimental and Simulated Study of Electrical Behavior of ZnO Film Deposited on Al Substrate for device applications, J. Mater. Sci. 25 (2014) 30623068. [10] Zh. Jin, L. Gao, Q. Zhou, J. Wang, High-Performance Flexible Ultraviolet Photodetectors Based on Solution-Processed Ultrathin ZnO/Au nanoparticle composite films, Sci. Rep. 4(2014) 4268. [11] L. Peng, L. Hu, X. Fang, Low‐dimensional nanostructure ultraviolet photodetectors. Adv. Mater. 25 (2013) 37. [12] L. Yin, H. Ding, Z. Yuan, W. Huang, C. Shuai, Z. Xiong, A simple and transparent well-aligned ZnO nanowire array ultraviolet photodetector with high responsivity, Optical Mater.80 (2018) 149-153. [13] J. Park, J. Lee, Y. Noh, K.H. Shin, D. Lee, Flexible ultraviolet photodetectors with ZnO nanowire networks fabricated by large area-controlled roll-to-roll processing. J. M. Chem. C 4 (2016) 7948-7958. [14] Y. Xie, L. Wei, G. Wei, Q. Li, D. Wang, Y. Chen, J. Jiao, A self-powered UV photodetector based on TiO 2 nanorod arrays, Nanosc. Res. Let 8 (2013) 188. [15] X. Zhou, Q. Zhang, L. Gan, X. Li, H. Zhang, High—Performance Solar‐Blind Deep Ultraviolet Photodetector Based on Individual Single‐Crystalline Zn2GeO4 Nanowire, Adv. Func. Mater. 26 (2016) 704-712. [16] H. Chen, K. Liu, L. Hu, A. A. Al-Ghamdi, X. Fang, New concept ultraviolet photodetectors, Mater. Today 18 (2015) 493-502. [17] R. Sun, G.G. Wang, Z.C. Peng, Fabrication and UV photoresponse of GaN nanowire-film hybrid films on sapphire substrates by chemical vapor deposition method. Mat. Lett. 217 (2018) 288-291. [18] Q. An, X. Meng, K. Xiong, Y. Qiu, Self-powered ZnS nanotubes/Ag nanowires MSM UV photodetector with high on/off ratio and fast response speed. Sci. Rep. 7 (2017) 4885. [19] W. Tian, H. Lu, L. Li, Nanoscale ultraviolet photodetectors based on onedimensional metal oxide nanostructures. Nano Res. 8 (2015) 382-405. [20] Zh. Zhan, L. Xu, J. An, H. Du, Z. Weng, W. Lu, Direct Catalyst-Free Chemical Vapor Deposition of ZnO Nanowire Array UV Photodetectors with enhanced photoresponse speed, Adv. Eng. Mater. 19 (2017) 8. [21] A. Zhang, G. Zheng, C. M. Lieber, General Synthetic Methods, In: Nanowires. Nanoscience and Tech., Springer, pp. 15-37, 2016. [22] M. Ghoul, Z. Braiek, A. Brayek, I. B. Assaker, N. Khalifa, J. B. Naceur, A. Souissi, A. Lamouchi, S. Ammar, R. Chtourou, Synthesis of Core/Chell ZnO/ZnSe Nanowires Using Novel Low-Cost Two-Steps Electrochemical Deposition Technique, J. Alloy and Comp. 647 (2016) 660-664.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

[23] G. H. Shen, A. R. Tandio, F. C. Hong, Hydrothermally Synthesized Ultrathin Zinc Oxide Nanowires Based Field-Effect Transistors, J. Thin Solid Film 618 (2016) 100-106. [24] TPH. Sidiropoulos, R. Röder, S. Geburt, O. Hess, S. A. Maier, C. Ronning, R. F. Oulton, Ultrafast ZnO Nanowire Lasers: Nano-Plasmonic Acceleration of Gain Dynamics at The Surface Plasmon Polariton Frequency, IEEE Laser and Electro-Optics Conference,CA,USA (2016). [25] L. Yang, W. Liu, H. Xu, J. Ma, C. Zhang, C. Liu, Enhanced near-UV electroluminescence from pGaN/i-Al 2 O 3/n-ZnO heterojunction LEDs by optimizing the insulator thickness and introducing surface plasmons of Ag nanowires. J. Mater. Chem. 5 (2017) 3288-3295. [26] P. Wang, Y. Fu, B. Yu, Y. Zhao, L. Xing, X. Xue, realizing room-temperature self-powered ethanol sensing of ZnO nanowire arrays by combining their piezoelectric, photoelectric and gas sensing characteristics, J. Mater. Chem. A 3 (2015) 3529-3535. [27] H. Yaghoubi, M. Schaefer, S. Yaghoubi, D. Jun, R. Schlaf, J. T. Beatty, A. Takshi, A ZnO Nanowire Bio-Hybrid Solar Cell, Nanotech. 28 (2016) 5. [28] M. Aras, S. Güler-Kılıç, Ç. Kılıç, Doping-induced spin-orbit splitting in Bi-doped ZnO nanowires, Phys. Rev. B 95 (2017) 155-404. [29] Zh. Zheng, L. Gan, H. Li, Y. Ma, Y. Bando, D. Golberg, T. A. Zhai, A Fully Transparent and Flexible Ultraviolet–Visible Photodetector Based on Controlled Electrospun ZnO‐CdO Heterojunction Nanofiber Arrays, Adv. Func. Mater. 25 (2015) 5885-5894. [30] A. Bonnano, M. Morello, M. Crepaldi, A. Sanginario, S. Benetto, V. Cauda, P. Civera, D. Demarchi, A Low-Power 0.13-µm CMOS IC for ZnO-Nanowire Assembly and Nanowire-Based UV Sensor Interface, IEEE Sensors J. 15 (2015) 4203-4212. [31] M. Y. Chaung, H. C. Yu, Y. K. Su, C. H. Hsiao, T. H. Kao, C. S. Huang, Y. C. Huang, J. J. Tsai, S. L. Wu, Density-Controlled and Seedless Growth of Laterally Bridged ZnO Nanorod For UV Photodetector Application, J. Sensors and Actuators B 202 (2014) 810-819. [32] T. Saha, N. Q. Guo, N. Ramakrishnan, Zinc Oxide Nanostructure-Based Langasite Crystal Microbalance Ultraviolet Sensor, IEEE Sensors J. 16 (2016) 2964-2970. [33] Ch. O. Chey, X. Liu, H. Alnoor, O. Nur, M. Willander, Fast piezo-resistive sensor and UV photodetector based on Mn-doped ZnO nanorods, J. Phys. Status Solidi 9 (2015) 87-91. [34] Y. N. Lee, J. K. Shin, Y.T. Lee, M. Ishida, W. Lee, SOI-Based Schottky Barrier Diode Array for Ultraviolet Line-Scanner, IEEE Sensors J. 15 (2015) 1727-1731. [35] K. Keramatnejad, F. Khorramshahi, S. Khatami, E. Asl-Soleimani, Optimizing UV detection Properties of n-ZnO nanowire/p-Si heterojunction Photodetectors by Using a Porous Substrate, Opt. Quant. Elec. 47 (2015) 1739-1749. [36] Zh. Wang, R. Yu, X. Wang, W. Wu, ZL. Wang, Ultrafast Response p-Si/n-ZnO Heterojunction Ultraviolet Detector Based on Pyro-Phototronic Effect, Adv. Mater. 28 (2017) 6880-6886. [37] X. Zheng, Y. Sun, X. Yan, X. Chen, Z. Bai, P. Lin, Y. Shen, Y. Zhao, Y. Zhang, Tunable Channel Width of a UV-Gate Field Effect Transistor Based on ZnO Micro-Nano Wire , J. RSC adv. 4 (2014) 18378-1838. [38] L. W. Ji, C. Z. Wu, T. H. Fang, Y. J. Hsiao, T. H. Meen, W. Water, Z. W. Chiu, K. T. Lam, Characteristics of Flexible Thin-Film Transistors with ZnO Channels, IEEE Sensors J. 13 (2013) 49404943. [39] V. Sorianello, L. Colace, S. Rajamani, G. Assanto, Design and Simulation of Optically Controlled Field Effect Transistors, J. Phys. Status Solidi C 11 (2014) 81–84. [40] J. Michel, J. Liu, L. C. Kimerling, High-performance Ge-on-Si Photodetector, J. Nat. Photon 4 (2010) 527-534. [41] Z. Pan, X. Zhao, W. Peng, X. Qi, Y. He, A ZnO-Based Programmable UV Detection Integrated Circuit Unit, IEEE Sensors J. 16 (2016) 7919–7923. [42] S. P. Ghosh, K. C. Das, N. Tripathy, G. Bose, D. H. Kim, T. I. Lee, J. M. Myoung, J. P. Kar, Ultraviolet photodetection characteristics of zinc oxide thin film and nanostructures, IOP Conference Series: Mater. Sci. Eng.115 (2016) 012035. [43] M. W. Chen, J. R. Retamal, C. Y. Chen, J. H. He, Photocarrier Relaxation Behavior of a Single ZnO Nanowire UV Photodetector: Effect of Surface Band Bending, IEEE Elect. Dev. Lett. 33 (2012) 411413. [44] Z. X. Li, Z. Rong, Effect of Surface Adsorbed Oxygen, Applied Voltage, and Temperature on UV Photoresponse of ZnO Nanorods, Chin. Phys, B 24 (2015) 107703. [45] Zh. Wen, L. Zhu, Z. Zhang, Z. Ye, Fabrication of gas sensor based on mesoporous rhombus-shaped ZnO rod arrays, Sensors and Actuators B 208 (2015)112-121.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

[46] J. Cho, Q. Lin, S. Yang, J. G. Simmons, Y. Cheng, E. Lin, J. Yang, J. V. Foreman, H. O. Everitt, W. Yang, J. Kim, Sulfur-Doped Zinc Oxide (ZnO) Nanostars: Synthesis and Simulation of Growth Mechanism, Springer Nano Research 5 (2012) 20-26. [47] Q. B. Lin, S. Q. Wu, Z. Z. Zhu, Formation of star nanowires of sulfur-doped zinc oxide: Ab initio calculations, AIP Adv. 6 (2016) 095219. [48] M. Taghavi, V. Mattoli, B. Mazzolai, C. Filippeschi, L. Beccai, Synthesizing tubular and trapezoidal shaped ZnO nanowires by an aqueous solution method, RCS NanoScale 5 (2013) 3505-3513. [49] J. H. Lee, C. Allen, Synthesis of tetrapod-shaped ZnO nanowires by directive oxidation of Al-Zn mixture, J. Ceramic Society of Japan 117 (2009) 111-114., vol. 117. [50] M.E. Law, E. Solley, M. Liang, D.E. Burk, Self-consistent model of minority-carrier lifetime, diffusion length, and mobility, IEEE Elec. Dev. Lett. 12 (1991) 401-403. [51] S. Selberherr, Analysis and simulation of semiconductor devices. Springer Science & Business Media, 1984. [52] D.M. Caughey, R.E. Thomas, Carrier mobilities in silicon empirically related to doping and field, Proc. IEEE, 55 (1967) 2192-2193. [53] J.W. Slotboom, The pn-product in silicon, Sol. St. Elec. 20 (1977) 279-283. [54] ATLAS User’s Manual, Silvaco Int., Santa Clara, CA, USA, 2016. [55] Sh. Lu, J. Qi, S. Liu, Z. Zhang, Z. Wang, P. Lin, Q. Liao, Q. Liang, Y. Zhang, Piezotronic Interface Engineering on ZnO/Au-Based Schottky Junction for Enhanced Photoresponse of a Flexible SelfPowered UV Detector, Appl. Mater. Inter. ACS 16 (2014) 14116-14122.