# GaAs Optical Field Effect Transistor (OPFET): A high performance photodetector for automotive applications.

ABSTRACTPhotodetectors are important components in automotive industry. Sensitivity, speed, responsivity, quantum efficiency, photocurrent gain and power dissipation are the important characteristics of a photodetector. We report a high performance photodetector based on GaAs Metal- Semiconductor Field Effect Transistor (MESFET), with very high responsivity, excellent quantum efficiency, high sensitivity, moderate speed, tremendous gain and low power dissipation, surpassing their photodiode, phototransistor and other counterparts. A theoretical model of GaAs front illuminated Optical Field Effect transistor is presented. The photovoltaic and photoconductive effects have been taken into account. The gate of the OPFET device has been left open to make a reduction in the number of power supplies. The results are in line with the experiments. The device shows high potential in automotive applications.

CITATION: Gaitonde, J. and Lohani, R., "GaAs Optical Field Effect Transistor (OPFET): A High Performance Photodetector for Automotive Applications," SAE Int. J. Passeng. Cars--Electron. Electr. Syst. 9(1):2016.

INTRODUCTION

The automotive industry has been equipped with photodetectors for control of headlamps, back lights, dashboard lights, anti-glare mirrors, and windshield wipers. These photodetectors require high sensitivity, responsivity, speed, low power dissipation, etc and should operate at visible or infrared wavelengths depending upon the application. Many field effect transistor (FET)-based photodetectors functioning in the visible-infrared region have recently been reported. The following few paragraphs will give a brief review of some of these photodetectors.

CdS nanobelt (NB) Metal-Semiconductor Field Effect Transistor (MESFET)-based photodetectors showing a photoresponsivity of 2x[10.sup.2] A/W corresponding to an external quantum effciency (EQE) of 5.2x[10.sup.2] and rise and decay times of 137 [micro]s and 379 [micro]s respectively at 488 nm wavelength at an optical power density of 5.3 mW/c[m.sup.2] were fabricated [1]. An external photoresponsivity of 128 mA/W (internal responsivity of 642 mA/W) was achieved in silicon waveguide integrated germanium Junction Field Effect Transistor (JFET) photodetector under 1550 nm illumination at an optical power of 1.7 mW with a 3 V bias [2]. Black phosphorous FETs showed a responsivity of 4.8 mA/W and a rise time of 1 ms at 640 nm at an optical power of 10 nW at 200 mV drain bias and zero gate bias [3]. Field-effect transistor-based solution-processed colloidal quantum dot photodetector exhibiting broad spectral bandwidth in the Near-Infrared (NIR) region and a responsivity of 0.391 mA/W under 600 nm illumination at an intensity of 30 [micro]W/c[m.sup.2] was established [4].

CdTe nanowire FETs with a responsivity of 80.1 A/W, a photoconductive gain of 2.5x[10.sup.4]% and response and decay times of 0.7 s and 1 s respectively under 400 nm illumination at an optical power density of 78 [micro]W/c[m.sup.2] were demonstrated [5]. A photocurrent gain of 9.8 was attained in a photo JFET with a colloidal quantum dot absorber channel layer at 450 nm wavelength at 0.4 [micro]W optical power at a bias of 30 V [6]. A rise time of 1 0 [micro]s was obtained. Pentacene-based photodetector using vertical FET configuration showed a responsivity of 160 mA/W at 400 nm at a drain bias of -2 V [7]. Using a field-effect transistor configuration, a solution-processed NIR photodetector based on PbSe colloidal quantum dots was introduced. It showed a responsivity of 500 A/W at a drain bias and a gate bias of -40 V with 40 mW/c[m.sup.2] of 980 nm illumination [8]. Organic FETs utilizing neodymium phthalocyanine as light sensitive material exhibited a responsivity of 108 A/W (EQE of 20612%) under 650 nm illumination with an optical power density of 5.92 [micro]W/c[m.sup.2] at a drain bias of 50 V [9].

In this paper, we present a theoretical model of GaAs Optical Field Effect Transistor or Optically Controlled MESFET (OPFET) by considering both photoconductive and photovoltaic effects into account. The first section presents the theory. The continuity equations in the channel and depletion regions have been solved analytically. The channel charge, and depletion charge due to photogeneration as well as charge due to doping have been calculated by numerical integration using Trapezoidal method. The total drain to source current has been computed using the model presented in [10]. The various detector parameters which reveal its performance are calculated. The second section shows the obtained results and their interpretation and analysis. The third section gives a summary or the conclusion of the work done.

The simulations show that the device has better performance in terms of detector responsivity, gain, and quantum efficiency over various detectors reported so far in the visible-NIR region. In addition, it has low power dissipation, low value of switching time and high sensitivity. This detector will be useful for the automotive industry employing visible-infrared photodetectors with benefits to the aerospace and commercial vehicle industries as well.

THEORY

The schematic structure of GaAs OPFET with floating gate is presented in Fig. 1. The substrate material is p-type semi-insulating (SI) GaAs whereas the active layer is of n-type GaAs. The active layer is uniformly doped. The radiation is allowed to fall on the transparent gate of the device. The gate is kept floating. The radiation is absorbed in the channel and substrate regions creating electron-hole pairs. The optically generated electrons move towards the channel while the holes move in the opposite direction. On the application of drain to source voltage, the electrons give rise to drain to source current which increase the conductivity of the channel. The holes crossing the junction develop a photovoltage which reduces the depletion width and in turn increases the drain to source current.

The extension of the Schottky junction depletion region in the channel measured from the surface under illumination is given by [11]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

where [[PHI].sub.B] is the Schottky barrier height, [DELTA] is the position of Fermi level below the conduction band, [epsilon] is the permittivity of GaAs, [V.sub.OP1] is the photovoltage generated across the Schottky junction and v(x) is the channel voltage which varies from 0 at the source end and [V.sub.DS] at the drain end. [V.sub.DS] is the drain to source voltage and [v.sub.gs] is the gate to source voltage.

Since the substrate is made semi-insulating and the channel is moderately doped, the p-n junction depletion width at the channel-substrate interface extends completely into the substrate whereas the depletion width is zero on the channel side. The photovoltage generated across the p-n junction depletion region has no effect on the detector characteristics.

Calculation of Photovoltage

The continuity equation for holes in the Schottky junction depletion region is given by [11]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

Here p is the hole density, [v.sub.y] is the carrier velocity in the y-direction, [[tau].sub.p] is the lifetime of holes, [alpha] is the absorption coefficient, [PHI] is the radiation flux density, y is the distance from surface towards the substrate.

The solution of Eq. (2) can be expressed analytically as

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)

using the boundary condition at y=[y.sub.dg,]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)

Then the photovoltage can be calculated as

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)

where p(0) is the hole density evaluated at y=0, [J.sub.s1] is the reverse saturation current density across the Schottky junction, q is the electronic charge, k is the Boltzmann constant and T is the temperature.

Calculation of Photogenerated Electron Density in the Channel

The continuity equation for electrons in the neutral channel region is represented by [11]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (8)

Here n is the electron density, [D.sub.n] is the diffusion constant of electrons, [[tau].sub.n] is the lifetime of electrons.

Equation (8) can be solved analytically and its solution is given as

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (9)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (10)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (11)

[L.sub.n] is the diffusion length of electrons, [L.sub.n] = [([D.sub.n] [[tau].sub.n]).sup.1/2].

The boundary condition used is at y=[y.sub.dg,]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (12)

Calculation of Photogenerated Electron Density in the Depletion Regions

The continuity equation in the depletion region is given by [11]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (13)

The analytical solution is given as

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (14)

Calculation of Drain to Source Current

The total conducting charge is given by,

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (15)

Here [Q.sub.d] is the charge due to doping; [Q.sub.ch], [Q.sub.dep1], and [Q.sub.dep2] are the charge due to photogeneration in the channel, Schottky junction depletion region, and substrate depletion region respectively.

[Q.sub.n] can be calculated as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (16)

where [n.sub.ch], [n.sub.dep1], and [n.sub.dep2] are the photogenerated electron densities in the channel, Schottky junction depletion region, and substrate depletion region respectively; a is the active layer thickness; d is the surface to substrate thickness.

The total drain to source current is then calculated using the model described in [10]:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (17)

where the symbols have the same meaning as in [10]. The parasitic source and drain resistances involved in the calculation are obtained following a similar procedure as given in [12].

All the integrations concerned with the computation of charge are performed numerically using Trapezoidal method.

Since the gate is kept open, there is no leakage current flow at the gate junction. The operating ambient temperature may have impact on the device performance. However, it is subject to a separate and independent topic and beyond the scope of this paper.

Calculation of Detector Parameters Responsivity [13]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (18)

[I.sub.ph] is the photocurrent and P is the incident optical power.

(P = ZL[PHI]h?/[(1 - [r.sub.s])(1 - [r.sub.m])], h is the Planck's constant, v is the frequency of radiation, rm is the reflection coefficient at the metal entrance and [r.sub.s] is the reflection coefficient at the metal semiconductor interface.

Quantum Efficiency [13]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (19)

Photocurrent gain [14]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (20)

([I.sub.L] = qZL[PHI](1 - exp(-ad))is the primary photocurrent)

Power dissipation [15]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (21)

Switching time [15]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (22)

Sensitivity

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (23)

([[PHI].sub.high] is the higher flux density and [[PHI].sub.low] is the lower flux density)

All the above simulations have been carried out in MATLAB with MEX coding to accelerate the program.

RESULTS

The simulations have been performed to obtain the various detector parameters at wavelengths ranging from 350 nm-875 nm. Refer to Table 1 in Appendix for various parameters used in calculation. Fig.2 shows the responsivity at different wavelengths at a drain to source bias of 2.21 V and a flux density of [10.sup.12] /[m.sup.2]-s.

The graph in Fig. 2 shows that very high values of responsivity (~[10.sup.12]) are obtained over the range 450 nm-850 nm and it is constant. The band-gap energy of GaAs is 1.424 eV which corresponds to a wavelength of 872 nm. Thus any energy less than 1.424 eV (872 nm) will generate less response which has been proven in the graph where there is a greater than one order reduction in the responsivity at 875 nm. The graph clearly shows that the GaAs OPFET is operable in the visible-NIR region with no response in the UV range (300 nm-400 nm). The flux density of [10.sup.12] /[m.sup.2]-s corresponds to an optical power of 0.17 fW. The high response at such low level of illumination is attributed to the photovoltage generated at the Schottky junction. The photovoltage at a bias of 2.21 V and at a flux density of [10.sup.12] /[m.sup.2]-s is shown in Fig. 3 at different wavelengths.

The reponsivity as a function of optical power under an illumination of 600 nm at a bias of 2.21 V is depicted in Fig. 4.

As shown in Fig. 4, the responsivity decreases as the optical power increases. However, the responsivity at the highest optical power used in this simulation (0.17 mW) is quite a high value (22.8 A/W).

The quantum efficiency as a function of wavelength at a flux density of [10.sup.12] /[m.sup.2]-s and a bias of 2.21 V is presented in Fig. 5.

The quantum efficiency defines how efficiently the optical energy is converted into electrical energy. The quantum efficiency variation with wavength (Fig. 5) is similar to that of responsivity variation and can be interpreted in the same manner. Very high EQE values are obtained at [10.sup.12] /[m.sup.2]-s flux density due to the photovoltage produced across the Schottky junction. Since the responsivity and quantum efficiency are related quantities (R = q x EQE/h?), the EQE dependence on optical power will be similar to that of responsivity dependence as shown in Fig. 4.

Another important parameter is the gain achieved in the photodetection process. The photocurrent gain at different wavelengths is shown in Fig. 6 at a fux density of [10.sup.12] /[m.sup.2]-s and a bias of 2.21 V.

Very high values of optical gain are achieved as shown in Fig. 6 over the visible-NIR region. The graph of photocurrent gain versus optical power is shown in Fig. 7 at a wavelength of 600 nm and a bias of 2.21 V. As can be seen, the gain decreases as optical power increases.

The photovoltaic effect is dominant at very low intensities. As the optical power increases, the photoconductivity increases. However, the photovoltage generated also increases with optical power as shown in Fig. 8. Thus, it is the combination of photovoltaic and photoconductive effects at higher light intensities.

Figure 9 shows the photoresponses of GaAs OPFET at different optical power levels under 600 nm illumination. The figure depicts a high sensitivity at all intensities.

The detector switching time is presented in Fig. 10 as a function of optical power. Very low values of switching time are obtained. The response times decrease with incident optical power.

Power dissipation is another important figure of merit of photodetectors. Fig. 11 shows the power dissipation at different optical power levels under an illumination of 600 nm at a bias of 2.21 V. The power dissipation is in the milliwatts or one-tenths of milliwatts range thus showing that the device can be regarded as a low power device.

CONCLUSION

A theoretical model of GaAs OPFET has been presented which takes into account both the photovoltaic and photoconductive effects. The continuity equations in the channel and depletion regions have been solved analytically for electron and hole density. The charge components have been obtained numerically using Trapezoidal rule and the drain-to-source current has been calculated following the model presented in [10]. The various detector parameters have been evaluated. The device exhibits high values of responsivity, quantum efficiency, and photocurrent gain which are higher than the values reported in the review. The values are very high at extreme low intensities and decrease as optical power increases. The device shows superior response in the visible-NIR region thus showing its operation in this spectral range and also high sensitivity. Furthermore, low values of switching time and power dissipation are obtained. Hence the device will also work as a high speed low power device. The device is operating with only one power supply voltage since the gate has been left floating which is an additional advantage. Also below 3 V power supply has been used. Thus, the device will serve as a high performance photodetector for automotive applications.

FUTURE RESEARCH CONSIDERATIONS

Possible future research may include device reliability and latch-up conditions caused by boundary effects.

REFERENCES

[1.] Yu, Y., Dai, L., Wen, X., Wu, P., et al. "High-performance single CdS nanobelt metal-semiconductor field-effect transistor-based photodetectors." ACS Applied Materials & Interfaces 2(10): 2724-2727, 2010

[2.] Jian, W., Yu, M., Lo, G., Kwong, D., et al. "Silicon waveguide integrated germanium JFET photodetector with improved speed performance." IEEE Photonics Technology Letters 12(23): 765-767, 2011

[3.] Michele, B., Groenendijk, D., Blanter, S., Steele, G., et al. "Fast and broadband photoresponse of few-layer black phosphorus field-effect transistors." Nano letters 14(6): 3347-3352, 2014

[4.] Yang, S., Zhao, N., Zhang, L., Zhong, H., et al. "Field-effect transistor-based solution-processed colloidal quantum dot photodetector with broad bandwidth into near-infrared region." Nanotechnology 23(25), 2012

[5.] Mehrdad, S., Davami, K., Kheirabi, N., Baek, C., et al. "Single-crystalline CdTe nanowire field effect transistors as nanowirebased photodetector." Physical Chemistry Chemical Physics 16(41): 22687-22693, 2014

[6.] Valerio, A., Kramer, I., Labelle, A., Sutherland, B., et al. "Photojunction Field-Effect Transistor Based on a Colloidal Quantum Dot Absorber Channel Layer." ACS nano 9(1): 356-362, 2015

[7.] Dan, Y., Zhang, L., Wang, H., Wang, Y., et al. "Pentacene-Based Photodetector in Visible Region With Vertical Field-Effect Transistor Configuration." Photonics Technology Letters, IEEE 27(3): 233-236, 2015

[8.] Wang, H., Li, Z., Fu, C., Yang, D., et al. "Solution-Processed PbSe Colloidal Quantum Dot-Based Near-Infrared Photodetector." Photonics Technology Letters, IEEE 27(6): 612-615, 2015

[9.] Lei, S., Li, Y., Ren, Q., Lv, W., et al. "Toward Ultrahigh Red Light Responsive Organic FETs Utilizing Neodymium Phthalocyanine as Light Sensitive Material." IEEE Transactions on Electron Devices, 63(1): 452-458, 2016

[10.] Chen, T., and Shur, M. "Analytical models of ion-implanted GaAs FET's." IEEE Transactions on Electron Devices, 32(1): 18-28, 1985

[11.] Roy, N., and Pal, B.,"Frequency-Dependent OPFET Characteristics with Improved Absorption under Back Illumination", Journal of Lightwave Technology, 18(4): 604-613, 2000

[12.] Byun, Y., Shur, M., Peczalski, A., and Schuermeyer, F. "Gate-voltage dependence of source and drain series resistances and effective gate length in GaAs MESFETs." IEEE Transactions on Electron Devices, 35(8): 1241-1246, 1988

[13.] Sze, S., and Kwok, K., "Physics of semiconductor devices" 3rd Edition, pp. 27, 2007

[14.] Madheswaran, M., Rajamani, V. and Chakrabarti, P. "Quasi-two-dimensional simulation of an ion-implanted GaAs MESFET photodetector." Microwave and Optical Technology Letters, 26(4): 247-254, 2000

[15.] Shur, M., and Eastman, L. "Current-voltage characteristics, small-signal parameters, and switching times of GaAs FET's." Electron Devices, IEEE Transactions on, 25(6): 606-611, 1978

[16.] Hudait, M., and Krupanidhi, S.,"Doping dependence of the barrier height and ideality factor of Au/n-GaAs Schottky diodes at low temperatures", Physica B: Condensed Matter, 307(1-4), pp. 134, 2001

CONTACT INFORMATION

Jaya V. Gaitonde, Electronics and Telecommunication Dept.

Farmagudi, Ponda,Goa, 403401.

jayagaitonde@rediffmail.com

Mobile: (+91) 9420687029.

Jaya Gaitonde and R B Lohani

GOA College of Engineering

APPENDIX

APPENDIX

Table 1. Various parameters used in calculation Parameter Name Value Ref Unit [mu] Low field 0.5 [13] ([m.sup.2]/V. electron s) mobility Z Charnel 100x[10.sup.-6] [11] (m) width L Channel 3x[10.sup.-6] [11] (m) Length a Active layer 0.15x[10.sup.-6] [11] (m) thickness d Surface to 1x[10.sup.-6] [11] (m) substrate thickness [[PHI].sub.B] Schottky 0.86 [16] (eV) Barrier Height [N.sub.dr] Equivalent 4.95x[10.sup.22] ([m.sup.-3]) constant doping concentration [DELTA] Position of 0.09 (eV) Fermi level below the conduction band [v.sub.y] Carrier 1.2x[10.sup.5] [11] (m/s) velocity in the y direction [[tau].sub.p] Lifetime of [10.sup.-8] [11] (s) holes [[tau].sub.n] Lifetime of [10.sup.-6] [11] (s) electrons

Printer friendly Cite/link Email Feedback | |

Author: | Gaitonde, Jaya; Lohani, R.B. |
---|---|

Publication: | SAE International Journal of Passenger Cars - Electronic and Electrical Systems |

Article Type: | Report |

Date: | May 1, 2016 |

Words: | 3330 |

Previous Article: | Reconstruction of wheel forces using an intelligent bearing. |

Next Article: | MEMS oscillators with improved resilience for harsh automotive environments. |

Topics: |