The Impact of Fixed Oxide Charge Density on the Performance of InSb Infrared Focal Plane Arrays

M. Ghorbanzadeha, M. Asadb, V. Fathipourc, M. Fathipoura*

aMEMS and NEMS Lab, Department of Electrical and Computer Engineering, University of Tehran, Tehran, Iran;

bNanotechnology Research Center, Shiraz University cDepartment of Electrical Engineering and Computer Science, Northwestern University, Evanston,

USA

BIOGRAPHY Mostafa Ghorbanzadeh: started his study in electronics engineering at the Ferdowsi University of Mashhad, Iran, in 2006. He started his M.Sc. degree at the University of Tehran, Iran in 2011. In 2011 he began to work on his master thesis in the MEMS & NEMS Lab and UTCAD Lab under Prof. M. Fathipour supervision. His research interests are in optoelectronics, solid state physics, nanoelectronics and III/V semiconductors. Currently, he has focused his work on InSb IRFPA design, fabrication and characterization.

TECHNICAL ABSTRACT Recently, large-format and small-pitch InfraRed Focal Plane Arrays (IRFPAs) has been developed for high

resolution thermal images. By scaling, FPA’s Performance decreases due to increase in surface effects [1]. Surface effects can be controlled by controlling surface potential. Surface potential depends on fixed oxide charges in passivation layer [2]. Density of fixed oxide charge in passivation layer can be controlled by varying passivation conditions and methods [3]. These charges can be negative or positive for InSb substrate. For example [1] has reported that negative fixed oxide charge can be fabricated with combination of anodic oxide and Photo-CVD passivation layer.

In this work, we investigate the effect of fixed oxide charges density on dark current, crosstalk and quantum efficiency. Our carrier generation-recombination models includes Shockley-Read-Hall, optical generation/radiative recombination, auger recombination and Front-side surface recombination. We also use Band-to-Band Tunneling model. Furthermore we use suitable optical and band structure properties for InSb in our simulation studies. Three pixels are considered from which central pixel is illuminated and investigate performance of the FPA [4]. Our studies show that FPA’s performance may be optimized by controlling fixed oxide charge density in passivation layer. Specifically:

• Minimum dark current occurs for large enough positive oxide charge (point B in Fig.1). Larger charge densities initiate tunneling (point A). Thus this point is not a reliable working point. On the other hand, large negative oxide charge densities (point C in Fig.2) causes extremely low dark current and is thus more reliable

• Maximum quantum efficiency occurs for negative oxide charge (Fig.3)

• Minimum crosstalk occurs for large enough negative oxide charge density. However this is not the case when negative charge density is so large that surface space-charge regions from neighboring pixels interact.(Fig.4)

A tradeoff exists between dark current, quantum efficiency and crosstalk. In general the optimal solution lies in a fabrication method which introduces negative fixed oxide charge thus increases quantum efficiency and decreases dark current.

Keywords: infrared, focal plane array, InSb, fixed oxide charge, crosstalk, dark current, quantum efficiency *mfathi@ut.ac.ir; phone 98 218 208-4329; http://tcad.ut.ac.ir

References

[1] K. Liu, J. Luo, and L. Dai, “Evaluation of implanted InSb p+n diodes passivated with composite anodic oxide/SiOx stack,” physica status solidi (a), vol. 205, no. 10, pp. 2469–2475, 2008.

[2] T. Sun, S. Lee, and S. J. Yang, “The current leakage mechanism in InSb p+n diodes,” Journal of applied physics, vol. 67, no. 11, pp. 7092–7097, 1990.

[3] I. Bloom and Y. Nemirovsky, “Surface passivation of backside-illuminated indium antimonide focal plane array,” Electron Devices, IEEE Transactions on Electron Devices, vol. 40, no. 2, pp. 309–314, 1993.

[4] N. Guo, W. Hu, X. Chen, and C. Meng, “Numerical analysis of mid-wavelength InSb infrared focal plane arrays,” Numerical Simulation of Optoelectronic Devices (NUSOD), 2010.

[5] A. Grove, Physics and technology of semiconductor devices, vol. 143. Wiley New York, 1967.

Figure 1 Dependence of dark current density on positive fixed oxide charge density at 77K for Insb IRFPA with various bias voltages.

Figure 2 Dependence of dark current density on negative fixed oxide charge density at 77K for Insb IRFPA with various bias voltages.

Figure 3 Dependence of quantum efficiency on fixed oxide charge density at 77K for Insb IRFPA with various bias voltages.

Figure 4 Dependence of crosstalk on fixed oxide charge density at 77K for Insb IRFPA with various bias voltages.

Maximum quantum efficiency

Crosstalk increases due to surface space-charge regions from neighboring pixels interact.

B: Having minimum dark current but not reliable

A: Tunneling is initiated

Due to contribution of generation-recombination centers (at the InSb/oxide interface), dark current increses[5]

C: Dark current decreases, because surface under the passivation layer is inverted. Thus contribution of generation-recombination centers (at the interface) in dark current decreases [5].

- - - - -