Natural lithography with 100-nm-diameter SiO2 spheres followed by inductively coupled plasma etching was used to texture the surface of 4H-SiC for a wide-spectrum large-acceptance-angle anti-reflective layer. The surface showed low normal-incidence reflectance of < 5% over a wide spectrum from 250 nm to 550 nm. Photodiodes fabricated from the surface-textured SiC showed broader spectral and angular responsivity than SiC photodiodes with SiO2 antireflective coating. The textured SiC photodiodes showed peak responsivity of 116 mA/W, large angle of acceptance angle (< 2% decrease in responsivity at 50 o incident angle) and low dark current at 10V.
© 2011 OSA
Wide band-gap semiconductor photodiodes (PDs) have emerged as potential components for ultraviolet (UV) detection in numerous civil and military applications. They have the advantage of sensitivity to middle UV (200 nm to 300 nm) with reduced response to near UV (300 nm to 400 nm) or visible light. Visible-blind or solar-blind UV detection can be achieved without filters or with less critical requirements on filters than narrower band-gap materials such as Si. 4H-SiC is one of the most promising materials for middle UV detection. SiC avalanche photodiodes (APDs) have shown good responsivity from 250 nm to 310 nm with very low dark current [1, 2]. Due to the availability of high quality SiC epitaxial material, SiC APDs with diameter of up to 250 μm have been reported . In addition, SiC is well suited for harsh environments owing to its hardness and thermal stability.
Imaging applications frequently require wide field of view. For example, in missile plume detection  the light signal originates from unpredictable positions and in non-line-of-sight ultraviolet (NLOS UV) communication  a wide field of view is needed to detect scattered light signals. At present, dielectric anti-reflection coatings (ARC) are widely used as surface treatment techniques to increase the normal-incidence responsivity of PDs near the wavelength of interest. The angular dependence of the reflectance is given by the expression 
Surface texturing is a technique that has been employed to reduce reflectance loss in solar cells , light emitting diodes  and, recently, photodiodes [8, 9]. In this work, surface texturing is proposed as a surface treatment technique on SiC PDs for wide-spectrum responsivity enhancement, large acceptance angle and uniform angular responsivity. Nanosphere natural lithography  is used to create a random pattern with feature size ~100 nm. Since the material quality of the depletion region is critical for photodiodes to achieve low dark current, inductively coupled plasma (ICP) etching was used to transfer the nanoscale pattern onto the SiC surface as a less invasive process than laser texturing [8, 9]. The spectrum and angular responsivity of SiC PDs without surface antireflective treatment, with SiO2 antireflective coating and with surface texturing are compared in this paper.
2. Nanosphere surface texturing and device fabrication
The 4H-SiC wafer from which the PDs were fabricated consists of an n-doped 4H-SiC substrate and the following four epitaxial layers, from bottom to top: 2000-nm n+ buffer layer (ND = 3.0 × 1018 cm−3), 650-nm p- layer (NA = 2.8 × 1015 cm−3), 200-nm p layer (NA = 2.4 × 1018 cm−3) and 200-nm p+ cap layer (NA = 1019 cm−3). A 5% solution of 100 nm-diameter SiO2 spheres in water was purchased from Corpuscular Inc. A solution of surfactant Triton X-100 and Methanol (1:100 in volume) was added into the suspension . The volume ratio of the SiO2 suspension to Triton X-100 solution was 10:1. The SiO2 suspension was spun onto the wafer with spin rate of 2100 revolutions per minute. This created a uniform monolayer of SiO2 spheres that covered most of the surface area after the solvent evaporated. A 36 second Ar/Cl2 inductive coupled plasma (ICP) etch was then performed to transfer the nanosphere monolayer random pattern onto the SiC wafer surface. The etch depth was ~180 nm in the areas between the SiO2 spheres. A buffered oxide etch was used to remove any SiO2 residue. Figure 1 shows scanning electron microscopic (SEM) images of the textured surface.
The normal-incidence surface reflectance was measured with a Perkin Elmer Lambda 950 UV/Vis/NIR spectrometer. As shown in Fig. 2 , low reflectance from 250 nm to 600 nm was observed on the textured SiC surface (△ data points). In the wavelength range 250 nm - 350 nm, the spectral region where SiC exhibits photoresponse, the reflectance of the textured surface is in the range 2.0% to 3.6%, which is one order of magnitude lower than that of the SiC/air interface (■ data points). For wavelengths > 450 nm, the textured-surface reflectance gradually increases owing to the disparity between the textured feature size and the wavelength . The reflectance of SiC with a 220 nm-thick SiO2 antireflective coating (ARC) prepared by plasma-enhanced chemical vapor deposition (PECVD) is also shown in Fig. 2 (◊ data points). Low reflectances of 7.5% and 3.8% were achieved in narrow spectral regions centered at 270 nm and 450 nm, respectively.
SiC PDs with beveled mesa structure were fabricated from the textured wafer. The mesas were defined by Ar/Cl2 ICP etching. A 3.5-hour thermal oxidation at 1050 °C was performed for sidewall damage removal and surface passivation. Multilayer ohmic contacts on both p- and n-type SiC were formed by evaporation of Ni/Ti/Al/Ni (350 nm, 200 nm, 800 nm and 600 nm, respectively). Two control samples were fabricated from the same wafer, but without surface texturing. One control sample was prepared with the same processing described above and the other one had an additional 220-nm-thick PECVD SiO2 ARC deposited after the thermal oxidation.
3. Experimental results
3.1 Current-voltage characteristics and transient response
The current-voltage (IV) characteristics of 200 μm-diameter SiC PDs are shown in Fig. 3 . All three wafers exhibited dark currents < 0.1 pA (3 nA/cm2) for reverse biased up to 10 V. However, as a result of the roughness of the textured sidewall, the devices with surface texturing showed higher dark current at high reverse bias. The textured devices showed lower forward current, which is due to increased lateral resistance in the textured region. The higher resistance also resulted in lower RC-limited response speed. The temporal pulse response was measured using a frequency-quadrupled Nd:YAG laser (λ = 266nm) with 250 ps pulse width. The 3dB bandwidth was derived from the Fourier transform of the transient response. The textured devices showed a lower bandwidth (~55 MHz) than the nontextured PDs (~140 MHz).
3.2 Spectral responsivity
The normal-incidence spectral responsivity was measured by comparing the photoresponse of the device under test at reverse bias of 10 V with that of a calibrated UV enhanced Si photodiode. A Xeon lamp followed by a monochromator was used as the source of narrow spectrum light. The photoresponse was measured with a lock-in amplifier. As shown in Fig. 4(a) , all devices showed peak responsivity at ~280 nm. The maximum responsivity of the textured device is 116 mA/W (external quantum efficiency, EQE = 52%), which is higher than the 99 mA/W peak responsivity (43% EQE) of the nontextured device. As shown in Fig. 4(b), a broad spectral responsivity enhancement is observed on the textured device, especially at wavelengths longer than 300 nm where the enhancement is > 25%. However, this enhancement drops with decreasing wavelength and becomes insignificant for wavelengths shorter than 260 nm. This is counter to what would be expected based on the reflectance curves shown in Fig. 2. A possible explanation is that absorption of short wavelength light primarily occurs near the surface and the surface recombination rate after ICP etching is higher than that of the original epitaxial surface. It follows that more carriers generated near the textured surface recombine before they can contribute to the photocurrent. The devices with a SiO2 ARC showed peak responsivity of 110 mA/W (49% EQE); a maximum responsivity enhancement of ~20% relative to the bare surface was observed at 270 nm, the wavelength of one of the reflectance valleys in Fig. 2.
3.3 Angular responsivity
The angular dependence of the photoresponse was studied with the same light source used for the spectral responsivity measurements, which is similar to that employed in Ref . The light from the monochromator was coupled into a 600 nm-diameter UV optical fiber followed by a collimator lens. The collimator was held on a rotation mount with a 9 cm-long arm. The inset of Fig. 5 shows the profile of the light spot in the plane of the sample surface for normal incidence. It was measured by moving a detector with 100-μm-diameter active area across the light beam in two perpendicular directions (shown as x and y in the figure). The diameter of the uniform illumination region was ~1.5 mm. The angular responsivity R(θ) of the device can be calculated from the relationFig. 4, Ip(θ) and I p(0) are the photocurrents measured at incident angles of θ and 0°, respectively. The cos(θ) term accounts for the fact that the projected area of the detector on the normal plane of the light beam decreases with the increasing angle of incidence.
Figure 5 shows the angular responsivity of 350 μm-diameter PDs measured at 280 nm wavelength with incidence angles, θ, in the range 0 o to 84°. For small incidence angles both the detector with surface texturing and that with the SiO2 ARC demonstrated enhanced responsivity compared to that of the PD without antireflective surface treatment. With the increase of incidence angle, the responsivity of the SiC PDs with surface texturing or without antireflective surface treatment showed no significant responsivity variance until the angle of incidence reached 50°; a decrease of 5% was observed at 56°. The devices with SiO2 ARC exhibited more variation with the angle of incidence. Relative to normal incidence the responsivity decreased by 5% at 23° and was less than that of the PD without surface treatment at 34°. For θ > 60°, the responsivity of all three types of PDs decreased rapidly to 50% of the normal incidence values θ = 80°.
The angular responsivities of the SiC PDs with no surface antireflective treatment or with surface texturing are shown in Fig. 6 for wavelengths in the range 260 nm to 310 nm. For 0 ≤θ ≤46°, the SiC PDs without surface treatment showed uniform angular responsivity with a standard deviation of 0.7%. For each wavelength, the responsivity at 50° is only ~2% lower than that at normal incidence. The devices with textured surface also exhibited uniform photoresponse in this spectral range. The measured standard deviation for 0 ≤θ ≤46° was also 0.7% and the average responsivity at 50° was 1.6% lower than that at normal incidence.
As shown in Fig. 7(a) , the angular responsivity of the PDs with SiO2 ARC fluctuated with wavelength, e.g., the ratio of measured responsivity at 50° to that at normal incidence varied from 81% (λ = 270 nm) to 101% (λ = 310 nm). A calculation of angular responsivity variance based on the change of surface reflection loss using Eq. (1) is shown in Fig. 7(b), which yields relatively good agreement between the simulation and the experiment data. It follows that the angular responsivity nonuniformity is primarily due to the angular dependence of the reflectance.
SiC photodiodes with nanosphere natural lithography textured surfaces were fabricated and their performance was compared to those of SiC photodiodes with SiO2 anti-reflective coating or without any anti-reflective surface treatment. At reverse bias of 10V, the surface texturing showed a wide-spectrum responsivity enhancement of up to 30%. The peak responsivity of the textured PDs is 116 mA/W, which is higher than that of the PDs with 220 nm SiO2 anti-reflection coating (110 mA/W) and PDs without anti-reflective surface treatment (99 mA/W). The acceptance angle and angular responsivity uniformity from 0° to 50° angle of incidence for the textured photodiodes is slightly better than that of those without anti-reflective surface treatment while those with SiO2 anti-reflection coating showed significant angular spectral nonuniformity. All the studied SiC photodiodes exhibited low dark current of < 0.1 pA at 10V reverse bias.
This work was supported by the U. S. Army Research Laboratory under cooperative agreement number W911NF-09-2-0019.
References and links
1. X. G. Bai, H. D. Liu, D. C. McIntosh, and J. C. Campbell, “High-detectivity and high-single-photon-detection-efficiency 4H-SiC avalanche photodiodes,” IEEE J. Quantum Electron. 45(3), 300–303 (2009). [CrossRef]
2. Q. G. Zhou, D. McIntosh, H. D. Liu, and J. C. Campbell, “Proton-implantation-isolated separate absorption charge and multiplication 4H-SiC avalanche photodiodes,” IEEE Photon. Technol. Lett. 23(5), 299–301 (2011). [CrossRef]
3. F. P. Neele and R. M. Schleijpen, “Electro-optical missile plume detection,” Proc. SPIE 5075, 270 (2003).
4. G. A. Shaw, M. L. Nischan, M. A. Iyengar, S Kaushik, and M. K Griffin, “NLOS UV communication for distributed sensor systems,” Proc. SPIE 4126, (2000).
5. H. A. Macleod, in Thin-Film Optical Filters, 4th ed. (Taylor & Francis Group, 2010).
6. J. H. Zhao, A. H. Wang, M. A. Green, and F. Ferrazza, “19.8% efficient “honeycomb” textured multicrystalline and 24.4% monocrystalline silicon solar cells,” Appl. Phys. Lett. 73(14), 1991–1993 (1998). [CrossRef]
7. R. Windisch, C. Rooman, S. Meinlschmidt, P. Kiesel, D. Zipperer, G. H. Dohler, B. Dutta, M. Kuijk, G. Borghs, and P. Heremans, “Impact of texture-enhanced transmission on high-efficiency surface-textured light-emitting diodes,” Appl. Phys. Lett. 79(15), 2315–2317 (2001). [CrossRef]
8. Z. Li, B. K. Nayak, V. V. Iyengar, D. McIntosh, Q. G. Zhou, M. C. Gupta, and J. C. Campbell, “Laser-textured silicon photodiode with broadband spectral response,” Appl. Opt. 50(17), 2508–2511 (2011). [CrossRef] [PubMed]
9. Z. H. Huang, J. E. Carey, M. G. Liu, X. Y. Guo, E. Mazur, and J. C. Campbell, “Microstructured silicon photodetector,” Appl. Phys. Lett. 89(3), 033506 (2006). [CrossRef]
10. H. W. Deckman and J. H. Dunsmuir, “Natural lithography,” Appl. Phys. Lett. 41(4), 377–379 (1982). [CrossRef]
11. J. C. Hulteen and R. P. Vanduyne, “Nanosphere lithography - a materials general fabrication process for periodic particle array surfaces,” J. Vac. Sci. Technol. A 13(3), 1553–1558 (1995). [CrossRef]
12. H. E. Bennett and J. O. Porteus, “Relation between surface roughness and specular reflectance at normal incidence,” J. Opt. Soc. Am. 51(2), 123 (1961). [CrossRef]
13. C. P. Chen, P. H. Lin, L. Y. Chen, M. Y. Ke, Y. W. Cheng, and J. J. Huang, “Nanoparticle-coated n-ZnO/p-Si photodiodes with improved photoresponsivities and acceptance angles for potential solar cell applications,” Nanotechnology 20(24), 245204 (2009). [CrossRef] [PubMed]