Effect of barrier on the performance of sub ... - OSA Publishing

Effect of barrier on the performance of submonolayer quantum dot infrared photodetectors Jun Oh Kim,1,4 Zahyun Ku,2,4 Alireza Kazemi,1 Augustine Urbas,2 Sang-Woo Kang,3 Sam Kyu Noh,3 Sang Jun Lee,3,5 and Sanjay Krishna1,* 1

Center for High Technology Materials, University of New Mexico, NM 87106, USA Air Force Research Laboratory, Wright-Patterson Air Force Base, OH 45433, USA 3 Division of Industrial Metrology, Korea Research Institute of Standards and Science, Daejeon 305-340, South Korea 4 These authors contributed equally 5 [email protected] * [email protected] 2

Abstract: We report on the effect of confinement barriers on the performance of InAs/InGaAs sub-monolayer quantum dot infrared photodetectors. Two samples with different AlxGa1-xAs barrier compositions (x = 0.07 for sample A and x = 0.20 for sample B) were grown with fourstacks of sub-monolayer quantum dot. Sample A had a peak response at ~7.8 μm, whereas sample B demonstrated three peaks at ~3.5, ~5, and ~7.0 μm with the intensity of the peaks strongly dependent on the applied bias. At 77 K, sample A and B had a detectivity of 1.2 × 1011 cm.Hz1/2/W (Vb = −0.4 V bias) and 5.4 × 1011 cm.Hz1/2/W (Vb = −1.5 V bias), respectively. ©2014 Optical Society of America OCIS codes: (040.3060) Infrared; (230.5160) Photodetectors; (230.5590) Quantum-well, -wire and -dot devices; (020.6580) Stark effect.

References and links 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

E.-T. Kim, A. Madhukar, Z. Ye, and C. Campbell, “High detectivity InAs quantum dot infrared photodetectors,” Appl. Phys. Lett. 84(17), 3277–3279 (2004). V. Ryzhii, “The theory of quantum-dot infrared photodetectors,” Semicond. Sci. Technol. 11(5), 759–765 (1996). S. Krishna, “Quantum dots-in-a-well infrared photodetectors,” J. Phys. D Appl. Phys. 38(13), 2142–2150 (2005). J. Phillips, “Evaluation of the fundamental properties of quantum dot infrared detectors,” J. Appl. Phys. 91(7), 4590–4594 (2002). H. Lim, W. Zhang, S. Tsao, T. Sills, J. Szafraniec, K. Mi, B. Movaghar, and M. Razeghi, “Quantum Dot Infrared Photodetectors: Comparison Experiment and Theory,” Phys. Rev. B 72(8), 085332 (2005). H. S. Ling, S. Y. Wang, C. P. Lee, and M. C. Lo, “High quantum efficiency dots-in-a-well quantum dot infrared photodetectors with AlGaAs confinement enhancing layer,” Appl. Phys. Lett. 92(19), 193506 (2008). S. Chakrabarti, S. Adhikary, N. Halder, Y. Aytac, and A. G. U. Perera, “High-performance, long-wave (~10.2 μm) InGaAs/GaAs quantum dot infrared photodetector with quaternary In0.21Al0.21Ga0.58As capping,” Appl. Phys. Lett. 99(18), 181102 (2011). A. V. Barve, S. Sengupta, J. O. Kim, J. Montoya, B. Klein, M. A. Shirazi, M. Zamiri, Y. D. Sharma, S. Adhikary, S. E. Godoy, W.-Y. Jang, G. R. C. Fiorante, S. Chakrabarti, and S. Krishna, “Barrier Selection Rules for Quantum Dots-in-a-Well Infrared Photodetector,” IEEE J. Quantum Electron. 48(10), 1243–1251 (2012). G. Jolley, L. Fu, H. H. Tan, and C. Jagadish, “Effects of well thickness on the spectral properties of In0.5Ga0.5As/GaAs/Al0.2Ga0.8As quantum dots-in-a-well infrared photodetectors,” Appl. Phys. Lett. 92(19), 193507 (2008). G. Ariyawansa, V. Apalkov, A. G. U. Perera, S. G. Matsik, G. Huang, and P. Bhattacharya, “Bias-selectable tricolor quantum dot infrared photodetector for atmospheric windows,” Appl. Phys. Lett. 92(11), 111104 (2008). A. V. Barve, J. Montaya, Y. Sharma, T. Rotter, J. Shao, W.-Y. Jang, S. Meesala, S. J. Lee, and S. Krishna, “High temperature operation of quantum dots-in-a-well infrared photodetectors,” Infrared Phys. Technol. 54(3), 215– 219 (2011). G. Jolley, L. Fu, H. H. Tan, and C. Jagadish, “Influence of quantum well and barrier composition on the spectral behaviorof InGaAs quantum dots-in-a-well infrared photodetectors,” Appl. Phys. Lett. 91(17), 173508 (2007). J. C. Cambell and A. Madhukar, “Quantum-dot infrared photodetectors,” Proc. IEEE 95(9), 1815–1827 (2007). P. Aivaliotis, L. R. Wilson, E. A. Zibik, J. W. Cockburn, M. J. Steer, and H. Y. Liu, “Enhancing the dot density in quantum dot infrared photodetectors via the incorporation of antimony,” Appl. Phys. Lett. 91(1), 013503 (2007).

#202226 - $15.00 USD (C) 2014 OSA

Received 3 Dec 2013; revised 17 Dec 2013; accepted 17 Dec 2013; published 7 Jan 2014

1 February 2014 | Vol. 4, No. 2 | DOI:10.1364/OME.4.000198 | OPTICAL MATERIALS EXPRESS 198

15. J. Shao, T. E. Vandervelde, A. Barve, A. Stintz, and S. Krishna, “Increased normal incidence photocurrent in quantum dot infrared Photodetectors,” Appl. Phys. Lett. 101(24), 241114 (2012). 16. F. Hopfer, A. Mutig, G. Fiol, M. Kuntz, V. A. Shchukin, V. A. Haisler, T. Warming, E. Stock, S. S. Mikhrin, I. L. Krestnikov, D. A. Livshits, A. R. Kovsh, C. Bornholdt, A. Lenz, H. Eisele, M. Dahne, N. N. Ledentsov, and D. Bimberg, “20 Gb/s 85°C error-free operation of VCSELs based on submonolayer deposition of quantum dots,” IEEE J. Sel. Top. Quantum Electron. 13(5), 1302–1308 (2007). 17. S. S. Mikhrin, A. E. Zhukov, A. R. Kovsh, N. A. Maleev, V. M. Ustinov, Y. M. Shernyakov, I. P. Soshnikov, D. A. Livshits, I. S. Tarasov, D. A. Bedarev, B. V. Volovik, M. V. Maximov, A. F. Tsatsul'nikov, N. N. Ledentsov, P. S. Kop'ev, D. Bimberg, and Z. I. Alferov, “0.94 μm diode laser based on Stranski-Krastanow and submonolayer quantum dots,” Semicond. Sci. Technol. 15(11), 1061–1064 (2000). 18. Z. C. Xu, D. Birkedal, J. M. Hvam, Z. Y. Zhao, Y. M. Liu, K. T. Yang, A. Kanjilal, and J. Sadowski, “Structure and optical anisotrophy of vertically correlated submonolayer InAs/GaAs quantum dots,” Appl. Phys. Lett. 82(22), 3859–3861 (2003). 19. T. Niermann, F. Kießling, M. Lehmann, J.-H. Schulze, T. D. Germann, K. Potschke, A. Strittmatter, and U. W. Pohl, “Atomic structure of closely stacked InAs submonolayer depositions in GaAs,” J. Appl. Phys. 112(8), 083505 (2012). 20. T. Kita, N. Tamura, O. Wada, M. Sugawara, Y. Nakata, H. Ebe, and Y. Arakawa, “Artificial control of optical gain polarization by stacking quantum dot layers,” Appl. Phys. Lett. 88(21), 211106 (2006). 21. T. Switaiski, U. Woggon, D. E. Alden Angeles, A. Hoffmann, J.-H. Schulze, T. D. Germann, A. Strittmatter, and U. W. Pohl, “Carrier dynamics in InAs/GaAs submonolayer stacks coupled to Stranski-Krastanov quantum dots,” Phys. Rev. B 88(3), 035314 (2013). 22. D. Z.-Y. Ting, S. V. Bandara, S. D. Gunapala, J. M. Mumolo, S. A. Keo, C. J. Hill, J. K. Liu, E. R. Blazejewski, B. Rafol, and Y.-C. Chang, “Submonolayer quantum dot infrared photodetector,” Appl. Phys. Lett. 94(11), 111107 (2009). 23. J. O. Kim, S. Sengupta, A. V. Barve, Y. D. Sharma, S. Adhikary, S. J. Lee, S. K. Noh, M. S. Allen, J. W. Allen, S. Chakrabarti, and S. Krishna, “Multi-stack InAs/InGaAs sub-monolayer quantum dots infrared photodetectors,” Appl. Phys. Lett. 102(1), 011131 (2013). 24. S. Sengupta, J. O. Kim, A. V. Barve, S. Adhikary, Y. D. Sharma, N. Gautam, S. J. Lee, S. K. Noh, S. Chakrabarti, and S. Krishna, “Sub-monolayer quantum dots in confinement enhanced dots-in-a-well heterostructure,” Appl. Phys. Lett. 100(19), 191111 (2012). 25. A. Polimeni, A. Patane, M. Henini, L. Eaves, and P. C. Main, “Temperature dependence of the optical properties of InAs/AlyGa1-yAs self-organized quantum dots,” Phys. Rev. B 59(7), 5064–5068 (1999). 26. Y. Huang and C. Lien, “Strong Stark effect of the intersubband transitions in the three coupled quantum well: Application to voltage-tunable midinfrared photodetectors,” J. Appl. Phys. 78(4), 2700–2706 (1995). 27. P. W. Fry, I. E. Itskevich, D. J. Mowbray, M. S. Skolnick, J. J. Finley, J. A. Barker, E. P. O’Reilly, L. R. Wilson, I. A. Larkin, P. A. Maksym, M. Hopkinson, M. Al-Khafaji, J. P. R. David, A. G. Cullis, G. Hill, and J. C. Clark, “Inverted Electron-Hole Alignment in InAs-GaAs Self-Assembled Quantum Dots,” Phys. Rev. Lett. 84(4), 733– 736 (2000).

1. Introduction In the past decade, quantum dot infrared photodetectors (QDIPs) [1–5] based on StranskiKrastanov (SK) quantum dots (QDs) have been widely explored for improved device performance (such as low dark current, higher operating temperature and multi-color detection) using various designs of heterostructures (e.g., quantum dot in-a-well: DWELL, resonant tunneling structure: RT-DWELL and confinement enhancing barrier: CE-DWELL) [6–8]. The DWELL design combines the advantages of a conventional quantum dot infrared photodetector (QDIP) such as low dark current and normal incidence operation with the design flexibility and reproducibility of a conventional quantum well infrared photodetector (QWIP) [6]. Various designs have been explored using SK dots [9–14]. However, one of the biggest limitations of this approach is the “pancake” shape of the dot, with a base of 20-30 nm and a height of 4-6 nm. This limits the 3D confinement in the quantum dot and reduces the ratio of normal incidence absorption to the off-axis absorption [15]. One of the alternative growth modes to the formation of SK QDs is a sub-monolayer (SML) deposition technique, which can achieve a much higher density, smaller size, better uniformity, and has no wetting layer as compared to the SK growth mode [16–21]. In the SML-QD design, less than 1 monolayer (ML) (hence the name, Sub-Mono-Layer) of InAs is epitaxially grown after which a thin (In)GaAs layer is grown. Following this, a second stack of InAs is grown which is vertically aligned to the first stack due to the strain effect. This process is repeated several times (depending on the number of stacks). Thus, the overall

#202226 - $15.00 USD (C) 2014 OSA



confinement potential of the structure is governed by the number of stacks and the amount of indium rather than the stochastic strain driven process alone. Quantum dot detectors based on sub-monolayer designs have been reported by a few groups with promising results [22–24]. Due to the advantages of SML-QDs, the SML-DWELL design has attractive features such as increased normal incidence absorption, strong in-plane quantum confinement, and narrow spectral wavelength detection as compared with SK-DWELL. In this paper, we investigate the effect of confinement barriers on the performance of SML-DWELL detectors. The active regions of the detector consisted of 4 stacks 0.3 ML InAs with In0.15Ga0.85As spacers. It was found that sample A with a confinement-enhanced (CE) Al0.22Ga0.78As barrier had a single peak at 7.8 μm with a detectivity of 1.2 × 1011 cm.Hz1/2/W (Vb = −0.4 V bias) at 77 K. However, sample B with an Al0.20Ga0.80As barrier had three peaks at (~3.5 μm, ~5 μm, ~7 μm) due to various quantum confined transitions. Using a 1D Schrodinger solver, the peaks were assigned to various bound-to-bound and bound-tocontinuum transitions. Sample B had a reduced dark current which improved the 77 K peak detectivity to 5.4 × 1011 cm.Hz1/2/W although at a higher operating bias Vb = −1.5 V bias. 2. Growth of SML-DWELL and Photoluminescence measurement at room temperature The SML-DWELL device structure and active region for both samples (sample A: black and sample B: red) are illustrated in Figs. 1(a) and 1(b), respectively. First, a 200 nm thick buffer layer, a 600 nm thick bottom contact layer (n = 2 × 1018 cm−3) and an AlxGa1-xAs barrier were grown at 590°C. The active region consists of 10 periods of the 4 stacks SML-DWELL design. The AlxGa1-xAs barrier thickness and Al composition (x) for sample A and B are 48 nm and 0.07, and 50 nm and 0.20, respectively. The layers of sample A consists of 2 nm thick Al0.22Ga0.78As, 1 nm thick GaAs, 4 stacks of InAs SML-QDs embedded in a 5.3 nm thick In0.15Ga0.85As, 1 nm thick GaAs, and 2 nm thick Al0.22Ga0.78As. In sample B, the DWELL layer is the same as sample A except without the 2 nm thick Al0.22Ga0.78As. SML-QDs are formed by the multiple stack technique, which results in the vertical coupling between each stack of InAs SML-QDs. The active region of the 4 stacks SML-QDs were grown as follows: (i) 1.06 nm In0.15Ga0.85As were deposited on 1 nm thick GaAs (ii) 0.3 ML InAs were deposited on 1.06 nm thick In0.15Ga0.85As; this was repeated four times to make 4 stacks of 0.3 ML InAs. (iii) A 1.06 nm thick In0.15Ga0.85As was grown on the last 0.3 ML InAs layer (fourth), which was capped with 1 nm thick GaAs to prevent the out-diffusion of indium atoms. During the formation of SML-QDs, 10 seconds of a growth interrupt with arsenic was used. The growth temperature was 500°C. Photoluminescence (PL) measurements were performed at room temperature (300 K) using a 632.8 nm He-Ne laser with ~12 mW pump power. The PL emission was detected using a monochromator and an InGaAs photodiode with standard lock-in techniques. Figure 1(c) shows the normalized PL spectra of both samples as a function of wavelength at 300 K. Both samples show strong emission with one dominant peak at ~967 nm (sample A, ~1.281 eV) and ~978 nm (sample B, ~1.268 eV). It is interesting to note that no excited state is visible in the sample even at higher pump powers (not shown here). The full width at half maximum (FWHM) of the emission peak for sample A and B are 41 and 47 meV, respectively. The high emission peak energy (~1.25 eV) and narrow FWHM of the PL spectra indicate that SML-QDs in sample A and B are a smaller size and more uniform than conventional SK-QDs. The peak wavelength of sample A is slightly blue-shifted by 11 nm, as compared with sample B. This is probably attributed to a higher Al concentration in the AlGaAs barrier used to enhance the confinement. We believe that the shift in PL peaks is attributed to enhanced confinement due to the additional Al0.22Ga0.78As barriers and Al incorporation in the InGaAs QW during the AlGaAs growth, thereby leading to an InAlGaAs QW. Polimeni et al. reported that InAs/AlxGa1-xAs QDs resulted in blue-shifted emission energy as the Al concentration (x) increased from 0 to 0.8, which is due to the formation of smaller dots and Al incorporation in InAs dots during the (AlGa)As overgrowth [25]. Figures #202226 - $15.00 USD (C) 2014 OSA



1(d) and 1(e) show the schematic of conduction band diagram of sample A and B, respectively. The energy levels (dashed lines) are extracted with a semi-empirical estimate based on the known conduction band offsets between the materials, the photocurrent spectra and PL data.

Fig. 1. (a) Schematic view of the SML-DWELL device structure. Sample A and B were fabricated with a 410 × 410 µm2 mesa with the circular aperture of 300 µm diameter for normal incidence. The active region consists of 10 periods of 4 stacks of 0.3 ML InAs SMLQDs layer. (b) Diagrams of the active region for sample A and B are shown in the upper (black) and lower (red) parts. InAs/InGaAs SML-QDs are placed between the GaAs QW layer and the AlxGa1-xAs barrier, which is composed of a 2 nm thick Al0.22Ga0.78As layer and a 48 nm thick Al0.07Ga0.93As layer for sample A and a 50 nm thick Al0.20Ga0.80As layer for sample B. (c) Room temperature photoluminescence (PL) data obtained with He-Ne laser excitation are plotted for sample A and B. The PL peak wavelength of sample A is blue-shifted by about 11 nm as compared with sample B, which results from the presence of Al0.22Ga0.78As confinement enhancing barrier. (d) and (e) Schematic of conduction band diagram of sample A and B, respectively. The energy levels in the DWELL structure (dashed lines) can be estimated with the PL and spectral response data.

#202226 - $15.00 USD (C) 2014 OSA



Fig. 2. (a) Spectral response of sample A as a function of applied bias at 77 K. The peak wavelengths of sample A is ~7.8 μm with a narrow bandwidth (Δλ/λp), which is related the bound-to-bound transition. (b) Spectral response of sample B under −0.5 V. Two peaks are observed at ~5 and 7 μm. The response at ~3.5 μm is also visible at low bias (c) and high bias (d).

3. Fabrication of SML-DWELL device and Barrier dependent spectral response with various bias values Following the growth, 410 × 410 µm2 mesa n-i-n devices were fabricated using standard optical lithography and inductively coupled plasma etching. Ohmic contacts were created by sequentially depositing Ge/Au/Ni/Au using e-beam evaporation and then performing rapid thermal annealing. A schematic of the fabricated device with a circular aperture of 300 µm in diameter is shown in Fig. 1(a). The devices were mounted on a leadless chip carrier (LCC) with silver epoxy, wire-bonded and loaded in a cryostat with a KBr window. The spectral responses were recorded using a Thermo-Nicolet Fourier transform infrared spectrometer. Figure 2(a) shows the spectral response (SR) of sample A measured for normal incidence with various bias values at 77 K. The dominant peak of sample A is ~7.8 μm with a narrow spectral bandwidth (Δλ/λp) of ~14%. This clearly indicates that the peak in sample A is possibly due to a bound-to-bound transition, specifically a transition between the SML-QD ground state and the excited state of the QW. Figure 2(b) shows the SR of sample B at Vb = – 0.5 V and 77 K. Two dominant response peaks are observed around 5 and 7 μm. Moreover, a weak response at ~3.5 μm is also visible at low and high bias as shown in Figs. 2(c) and 2(d). The spectral bandwidth of sample A (~14%) is broader than sample B (~9% at ~7 μm), which may be a result from the excited state of the QW because it is close to the Al0.07Ga0.93As conduction band edge.

#202226 - $15.00 USD (C) 2014 OSA



4. Bias dependent multicolor response and Improved device performance Sample B exhibited bias dependent multicolor response. Figure 3 shows a 2D contour map of the normalized SR intensity of sample B as a function of the wavelength and applied bias at 77 K. Under low applied bias (Vb = –0.5 ~–1 V), there is a peak at ~5 μm (ΔE ~248 meV) and a second peak at ~7 μm (ΔE ~177 meV) with similar peak intensity, which have ~12% and ~9% bandwidth (Δλ/λp), respectively. This narrow bandwidth suggests that both peaks result from bound-to-bound transitions from a state in the dot to a state in the well. The ~5 μm peak is due to the transition from the ground state of SML-QD to a second excited state of the QW and the ~7 μm peak is due to the transition from the ground state of the SML-QD to the first excited state of the QW as shown in the inset of Fig. 3. The response at ~5 μm has the highest peak intensity at –0.5 V, and then it gradually decreased with increasing bias voltage. Additionally, another peak at ~3.5 μm (ΔE ~354 meV) can be observed at a lower applied bias than Vb = −1.4 V bias as shown in Fig. 2(c). This peak is attributed to the transition between the ground state of the SML-QD and the continuum state of the well (bound-tocontinuum transition). The ~3.5 μm peak can also be observed at an applied bias higher than Vb = −1.4 V as shown in Fig. 2(d) and inset of Fig. 3. However, the SR peak intensity is much weaker than the other two peaks for all biases, due to much smaller absorption coefficients. Only ~7 μm peak is observed for high applied bias. For the bound-to-bound transition, the oscillation strength is large and the escape probability is small (as compared with low bias), hence the peak resulting from the bound-to-bound transition is dominant at high bias. These three peaks can also be observed at positive bias. Moreover, these three peaks are shifted to longer wavelengths as the applied bias increases, which can be explained by the quantum confined stark effect [26,27].

Fig. 3. Contour plot of the normalized SR of sample B as a function of the wavelength with the applied bias at 77 K. At low bias, two peaks are observed at ~5 μm and ~7 μm with narrow bandwidth. Both peaks are probably due to the transition between the ground state of SML-QD and the first (~7 μm) / second (~5 μm) excited state of the QW. These peaks are shifted to longer wavelengths as the applied bias increases, which results from the quantum confined stark effect. Moreover, another peak is visible at ~3.5 μm as shown in the inset to the figure.

Zero field-of-view dark current measurements were undertaken as a function of applied bias and temperature. Figure 4(a) shows the dark currents of the two samples at 77 K. Both

#202226 - $15.00 USD (C) 2014 OSA



samples have symmetric dark current values in the positive and negative bias, which is due to the symmetry of the design. As can be seen, sample B has a dark current that is three orders of magnitude lower than sample A. The reduction in dark current of sample B is associated with the increased quantum confinement and reduced thermionic emission current because of the higher Al composition of the barrier (Al0.20Ga0.80As) as compared with sample A (Al0.07Ga0.93As). In sample A, even though the CE barriers are higher (Al0.22Ga0.78As) than in sample B and were placed on both sides of the QW (InGaAs/GaAs), dark current tunneling through the barriers is possible since they are only 2 nm thick. Radiometric measurements were undertaken, using a 900 K calibrated blackbody source and a network analyzer. The measured peak responsivities of sample A and B are ~0.45 A/W (at 7.8 μm, Vb = −0.4 V bias) and ~1.3 A/W (at 7 μm, Vb = −1.5 V bias). The values for the responsivity are reported at the optimal bias with maximum signal to noise ratio. The detectivity (D*) is calculated using the following equation: D* = R(A∆f)1/2/in, where R is the responsivity, A is the area of the detector, ∆f is the bandwidth, and in is the noise current, respectively. D* as a function of bias voltage for both samples are also compared at 77 K using f/2 optics. As shown in Fig. 4(b), the peak detectivity (highest D*) of 1.2 × 1011 (at 7.8 μm) and 5.4 × 1011 cm.Hz1/2/W (at 7 μm) were achieved at −0.4 V and −1.5 V for sample A and sample B, respectively. It is obvious that the higher D* of sample B (than sample A) is mainly due to the low dark current and high responsivity. This measurement did not account for substrate scattering, which can increase the responsivity and detectivity.

Fig. 4. (a) Dark current of sample A and B at 77 K. Dark current of sample B is lower than sample A by over 3 orders of magnitude because of the high Al composition in AlGaAs barrier as current blocking layer. (b) Detectivity of both samples at 77 K.

5. Conclusion In summary, we have investigated reduced dark current performance of SML-DWELL detectors with two different AlGaAs barrier compositions. The SML-DWELL with the Al0.20Ga0.80As barrier (sample B) had dark current more than three orders of magnitude lower than the SML-DWELL with barrier compositions of Al0.22Ga0.78As and Al0.07Ga0.93As (sample A) as well as a 4.5 times higher detectivity (D*). Moreover, three peaks have been observed at ~3.5, ~5, and ~7.0 μm due to the bound-to-continuum and the bound-to-bound transition. The optimum bias for sample B (−1.5 V) is higher than sample A (−0.4 V), which is owing to the higher energy of the AlGaAs barrier. In addition, D* of sample B under low bias (−0.6 V) is ~1 × 1011 cm.Hz1/2/W at both ~5 μm and ~7 μm, which is suitable for a Focal Plane Array. Acknowledgments This work was supported by AFRL contracts FA4600-06-0003 and FA9453-13-1-0284. We also acknowledge the Korea Research Institute of Standards and Science grant, JP2012-0001.

#202226 - $15.00 USD (C) 2014 OSA

