A simple technique to incorporate microlenes with small photodiode arrays is demonstrated and analyzed. Using this method, the fill factor was increased from 2.6% to 22.4% for a two by two array. Simulation results are also shown. The photocurrent with microlens was approximately 8.6 times larger than without the microlens, which is consistent with simulation results.
©2010 Optical Society of America
Silicon carbide avalanche photodiodes (APDs) are becoming increasingly important in ultraviolet detection. They can potentially be used for applications, such as UV astronomy, chemical or biological reagent detection, and flame detection [1–3]. Currently, SiC devices are fabricated with diameters in the range of 30 μm to 250 μm. In some applications, larger active areas are needed. However, larger devices have more defects, which lead to higher dark current, instability, and lower yield. One approach to overcome this problem is to fabricate arrays of small-diameter APDs. Essentially all of the devices that have been reported to date require beveled sidewalls to avoid edge breakdown [3–6]. However, this requires additional space between devices and leads to low fill factors.
Combining microlenses with detectors is an effective way to increase the fill factor. Several approaches have been demonstrated. They can be divided into two categories: integrated microlenses and external microlenses. Glasses and polymers are candidate materials for integrated microlenses. They have been successfully fabricated on different detectors and emitters [7–10]. However, due to low optical transmittance in the UV range, it is impractical to fabricate integrated microlenses on SiC APDs using these materials. On the other hand, external microlenses can be utilized with UV detectors owing to a broad range of available materials. Typically, fused silica and CaF2 are used to fabricate UV microlenses because they have high optical transmittance in the UV range. Through microfabrication technologies, external microlens arrays can be fabricated [11,12]. However, the cost of custom designed external microlens arrays is usually expensive and their alignment to photodiode arrays can further increase the cost.
In order to improve the fill factor, a self-aligned microball lens integration  method to couple microlenses with a two by two SiC APD array for UV detection is presented in this paper. Using SU-8 photoresist as a “holder”, a 350 μm-diameter commercially available UV quartz microsphere lens (Humanity Co. (Japan)) can be combined with a 100 μm-diameter SiC APD. The transmission of these microsphere lenses in the wavelength range of 220nm to 440nm is approximately 90%. After mesa etching, the diameter of the APD active area is approximately 80 μm. The proposed structure is shown in Fig. 1 . With microlenses, the fill factor was increased from 2.6% to 22.4% in a two by two array structure. Numerical simulation is described in Section II. Section III reports the device fabrication and microlens incorporation. Section IV shows the device characteristics and the performance improvement afforded by the microlens.
2. Numerical simulation
In order to determine the optimum geometry parameters of the SU-8 photoresist and to estimate the tolerance to misalignment, numerical simulation using MATLAB has been employed. The microlens model, which is illustrated in Fig. 2 , assumes the following: The microspherical lens is vertically illuminated by parallel light. The light intensity is constant and 90% of the incident light is transmitted through the lens. The top contact will block light. The radii of lens and device are 175 μm and 40 μm, respectively. d is the distance between a refracted ray and the line of optical symmetry at the edge of the lens and x is the distance between the incident ray and the line of optical symmetry. A top view of the device is shown in Fig. 3 .
To design the size of SU-8 photoresist hole, the distance between refracted ray and the line of optical symmetry at the edge of the lens d was calculated. The simulation result shows that d is a function of x (the distance between incident ray and the central line), which is illustrated in Fig. 4 . In order for all the refracted light to be incident on the device, the radius of the SU-8 photoresist hole should at least be equal to the maximum value of d. Considering the fixed support from the photoresist and the maximum photoresist thickness, the radius of the hole was chosen as 160μm.
In order to design the thickness of the SU-8 photoresist, the variation of light coupled into the photodiode with the distance between the lens and the device was studied. At each distance, the light collected by the device was calculated. In Fig. 5 , the 175 μm-, 350 μm-, and 500 μm-diameter exhibit two peaks. The location of the second peak is determined by the contact geometry and the size of the lens. Since the distribution of the refracted light at a certain distance is not linear, it is possible to have most of the refracted light blocked by the contact. In theory, the optimum distance for 175 µm-radius microball lenses is approximately 25 µm. This is, however, difficult to achieve in practice and it was found empirically that the best results were achieved when the distance was zero. Therefore, the thickness of the photoresist could be in the range of 50 µm to 100 µm.
The alignment tolerance for this configuration has been calculated and is shown in Fig. 6 . If the thickness of the photoresist is less than ~100 µm, there will be a space up to 38 µm between the lens and the APD. Figure 6(a) shows the tolerance to lateral (Δx and Δy) and vertical (Δz) misalignment. Figure 6(b) illustrates the critical importance of the light source tilt degree on collection efficiency. If the tilt degree is less than 10°, acceptable coupling can be achieved.
3. Device fabrication and microlens integration
The SiC APD structure, shown in Fig. 7 , consists of a 200 nm p+ layer (Na = 1.1x1019cm−3), a 200 nm p layer (Na = 2x1018cm−3), a 480 nm p− layer (Na = 1.1x1016cm−3) and a 2 μm n+ layer (Nd = 4x1018cm−3) grown on an n-type substrate. The mesa was first etched by inductively coupled plasma (ICP). This was followed by passivation with 850 nm of silicon dioxide. Then, a 220 nm-thick silicon dioxide anti-reflection layer was deposited. Both p- and n-type metal contacts were deposited by e-beam evaporation in the following sequence: Ni(40 nm)/Au(100 nm) The contact pads were also deposited by e-beam evaporation. Based on the total active area of the four APDs in the array the fill factor is approximately 2.6%.
The pedestal was formed using a layer of SU-8 3000, a high contrast, epoxy based photoresist. Different thickness could be obtained by changing the spin speed or choosing different members of SU-8 3000 family. If the area of wafer is on the order of 1 cm by 1 cm, thickness up to 120 µm can be achieved. After the pedestal was formed, the lens was manually placed on the pedestals. Better mechanical stability could be realized with UV-cured epoxy.
4. Device characteristics and photocurrent enhancement with microlens
Figure 8 shows the current-voltage characterisics of a 80 μm-diameter SiC APD with and without the lens. A broadband UV light source was used for the photocurrent measurement. It can be seen that the photo current with lens is approximately 8.6 times larger than the photo current without the lens. It follows that the fill factor of a two by two array with microlenses is ~22.4%.
The quantum efficiency was measured with a xenon lamp source, a calibrated UV- enhanced silicon photodetector, a monochromator and a lock-in amplifier. Figure 9 shows the quantum efficiency of a typical 80μm-diameter SiC APD. At 266 nm, the external quantum efficiency reached a peak value of 42%. At unity gain, the photocurrent with and without the lens was also measured versus wavelength. Figure 10 shows that the improved photoresponse is independent of wavelength.
The photocurrent improvement can be written asFig. 5. For alignment tolerance, the coupling efficiency is more sensitive to lateral displacement (Δx and Δy) than vertical. Based on the measurement, we estimate that δ is approximately 0.8. Therefore, the calculated improvement due to the lens is 9.2, which is consistent with the measured result.
In conclusion, a new method to couple microlenses to a two by two SiC APD array was demonstrated. A model was also developed based on this structure. Light transmission and alignment tolerance were calculated depending on different factors. The improvement of the fill factor is demonstrated by the experiment results. With microball lens the photocurrent was approximately 8.6 times larger than without the lens, which is consistent with the calculated result.
References and links
1. G. A. Shaw, A. M. Siegel, J. Model, and D. Greisokh, “Recent progress in short-range ultraviolet communication,” Proc. SPIE 5796, 214–225 (2005). [CrossRef]
2. H. Liu, D. McIntosh, X. Bai, H. Pan, M. Liu, and J. C. Campbell, “4H-SiC PIN recessed window avalanche photodiode,” LEOS 2007 20th Annu. Meet. IEEE Lasers Electro-Opt. Soc. (2007), pp. 517–518.
3. H. Liu, D. McIntosh, X. Bai, H. Pan, M. Liu, J. C. Campbell, and H. Y. Cha, “4H-SiC PIN recessed window avalanche photodiode with high quantum efficiency,” IEEE Photon. Technol. Lett. 20(18), 1551–1553 (2008). [CrossRef]
4. A. Vert, S. Soloviev, J. Fronheiser, and P. Sandvik, “Solar-blind 4H-SiC single-photon avalanche diode operating in geiger mode,” IEEE Photon. Technol. Lett. 20(18), 1587–1589 (2008). [CrossRef]
5. H.-Y. Cha, S. Soloviev, S. Zelakiewiez, P. Waldrab, and P. M. Sandvik, “Temperature dependent characteristics of nonreach-through 4H-SiC separate absorption and multiplication APDs for UV detection,” IEEE Sens. J. 8(3), 233–237 (2008). [CrossRef]
6. J. Hu, X. Xin, X. Li, J. H. Zhao, B. L. VanMil, K. Lew, R. L. Myers-Ward, C. R. Eddy Jr, and D. K. Gaskill, “4H-SiC visible-blind single-photon avalanche diode for ultraviolet detection at 280 and 350 nm,” IEEE Trans. Electron. Dev. 55(8), 1977–1983 (2008). [CrossRef]
7. H. Liu, S. Chen, S. Dong, and X. Yi, “Bionic ommatidia based on microlens array,” Opt. Eng. 48(6), 063401 (2009). [CrossRef]
8. Y. Ee, P. Kumnorkaew, R. A. Arif, H. Tong, H. Zhao, J. F. Gilchrist, and N. Tansu, “Optimization of light extraction efficiency of III-nitride LEDs with self-assembled colloidal-based microlenses,” IEEE J. Sel. Top. Quantum Electron. 15(4), 1218–1225 (2009). [CrossRef]
9. H. Yang, C. Chao, M. Wei, and C. Lin, “High fill-factor microlens array mold insert fabrication using a thermal reflow process,” J. Micromech. Microeng. 14(8), 1197–1204 (2004). [CrossRef]
10. B. Lee, K. J. Cha, and T. H. Kwon, “Fabrication of polymer micro nano-hybrid lens array by microstructured anodic aluminum oxide (AAO) mold,” Microelectron. Eng. 86(4-6), 857–860 (2009). [CrossRef]
11. P. Nussbaum, R. Völkel, H. Herzig, M. Eisner, and S. Haselbeck, “Design, fabrication and testing of microlens arrays for sensors and Microsystems,” Pure Appl. Opt. 6(6), 617–636 (1997). [CrossRef]
12. P. Ruffieux, T. Scharf, I. Philipoussis, H. P. Herzig, R. Voelkel, and K. J. Weible, “Two step process for the fabrication of diffraction limited concave microlens arrays,” Opt. Express 16(24), 19541–19549 (2008). [CrossRef] [PubMed]
13. Y.-H. Huang, C.-C. Yang, T.-C. Peng, M.-C. Wu, C.-L. Ho, and W.-J. Ho, “Alignment tolerance Enlargement of a high-speed photodiode by a self-positioned microball lens,” IEEE Photon. Technol. Lett. 18(1), 112–114 (2006). [CrossRef]