- Email: [email protected]
Resonant tunneling diode photodetector with nonconstant responsivity
Optics Communications 355 (2015) 274–278
Contents lists available at ScienceDirect
Optics Communications journal homepage: www.elsevier.com/locate/optcom
Resonant tunneling diode photodetector with nonconstant responsivity Yu Dong a, Guanglong Wang a,n, Haiqiao Ni b,n, Jianhui Chen a, Fengqi Gao a, Baochen Li a, Kangming Pei b, Zhichuan Niu b a b
Laboratory of Nanotechnology and Microsystems, Mechanical Engineering College, Shijiazhuang 050000, China Institute of Semiconductors, Chinese Academy of Science, Beijing 100084, China
art ic l e i nf o
a b s t r a c t
Article history: Received 2 April 2015 Received in revised form 24 June 2015 Accepted 27 June 2015
Resonant tunneling diode with an In0.53Ga0.47As absorption layer is designed for light detection at 1550 nm. The responsivity of the detector is simulated by solving the Tsu–Esaki equation. The simulation results show that the responsivity of the detector is nonconstant. It decreases with the increment of the power density of the incident light. Samples of the detector are fabricated by molecular beam epitaxy. The experimental results show that the responsivity increases while the power density of the incident light decreases which agree with the simulation results. The responsivity reaches 4.8 108 A/(W/μm2) at room temperature and 5.0 109 A/(W/μm2) at 77 K when the power density of the incident light is 1 10 13 W/μm2. & 2015 Elsevier B.V. All rights reserved.
Keywords: Resonant tunneling diode Light detection Responsivity
1. Introduction Light detection at wavelength of 1550 nm with high sensitivity is of great importance for quantum communication [1–3]. At present, many types of detectors have been investigated such as avalanche photodiode (APD) and superconducting single-photon detector (SSPD) [4–7]. As APD rely on the avalanche multiplication effect, it has disadvantages of high dark count rate and afterpulse noise [8]. The SSPD is based on superconducting materials' sensitivity to the change in temperature. The SSPD is suitable for operation at 1550 nm with low dark count rate, high detection efficiency and speed. However, SSPD has to be operated at extremely low temperature which is a serious constraint for its applications[9,10]. In 2005, Blakesley [11] showed that resonant tunneling diode (RTD) which contains a quantum dot layer and absorption layer is capable of single photon detection around 850 nm. When the incident photons are absorbed by the absorption region, the photo-excited holes will move to the emitter side and be captured by the quantum dots layer, which causes holes accumulation and potential modulation near the double barrier structure (DBS). In 2007, the detection wavelength of this type of detector was extended to 1310 nm by utilizing In0.53Ga0.47As absorption layer [12]. This type of detector works at temperature lower than 77 K to
n
Corresponding authors. E-mail addresses: [email protected] (G. Wang), [email protected] (H. Ni).
http://dx.doi.org/10.1016/j.optcom.2015.06.064 0030-4018/& 2015 Elsevier B.V. All rights reserved.
achieve single photon sensitivity. In 2012, Hartmann [13] implemented light detection at 1.3 μm by adding a GaInNAs absorption layer into GaAs/AlGaAs RTD. Without a quantum dot layer, this type of detector is still capable of light detection with responsitivity of 1000 A/W at room temperature. The resonant tunneling diode photodetector (RTD-PD) has advantages of low working voltage, high quantum efficiency and low dark count rate, which is promising for light detection with high sensitivity at room temperature. In this paper, we add In0.53Ga0.47As absorption layer into InGaAs/AlAs RTD to implement light detection at 1550 nm with high sensitivity. We find out the responsivity of the RTD-PD is nonconstant, which has not been reported.
2. Structure of the RTD-PD The structure of the RTD-PD is shown in Fig. 1. The n-type In0.53Ga0.47As layers on the top and bottom act as collector and emitter. The DBS consists of 1.4 nm AlAs, 6 nm In0.53Ga0.47As and 1.4 nm AlAs. 500 nm intrinsic In0.53Ga0.47As absorption layer is placed between collector and the DBS for light absorption. 15 nm In0.53Ga0.47As spacer is placed between the DBS and emitter to prevent impurity scattering from emitter to the DBS. A ring contact is deposited on the top to form electrical contact and clear aperture. The diameter of the clear aperture and the ring contact is 25 μm and 29 μm respectively.
Y. Dong et al. / Optics Communications 355 (2015) 274–278
275
3. Simulation of the responsivity of the detector
incident light ring contact
n-InGaAs 100 nm
where e is the electron charge, m* is the effective mass in emitter, kB is the Boltzmann constant, ℏ is the reduced Planck constant, T is the absolute temperature, Ef is the Fermi energy in emitter, E is the electron energy along the longitudinal growth direction of the detector, VD is the potential drop along the DBS, and D(E) is the transmission coefficient of the DBS which can be calculated by transfer matrix method [15], given by 1 DðEÞ ¼ M 11 M 11 ð2Þ
InGaAs 500 nm AlAs 1.4 nm InGaAs 6 nm AlAs 1.4 nm InGaAs 15 nm n-InGaAs 100 nm Fig. 1. Structure of the RTD-PD. The DBS is consisted of 1.4 nm AlAs, 6 nm In0.53Ga0.47As and 1.4 nm AlAs. A ring contact is deposited on the top to form electrical contact and clear aperture.
Incident light at wavelength of 1550 nm is absorbed by the In0.53Ga0.47As absorption layer, producing photogenerated electron–hole pairs. When the detector is under positive bias, the photogenerated holes will drift to the emitter side and accumulate at the interface between the DBS and the absorption layer. This causes band bending through the DBS, resulting to the variation of the tunneling current density of the RTD-PD. Meanwhile, if the incident light is not fully absorbed by the absorption layer, the light will be further absorbed by the spacer layer generating electron–hole pairs. The photogenerated electrons will accumulate near the interface between the DBS and the spacer layer, inducing holes on the other side of the DBS. This effect can promote the photo-response of the RTD-PD, but considering the thickness of the spacer layer is only 15 nm, this effect is negligible.
emitter collector + r
+
r
+
The responsivity of the detector comes from the variation of the tunneling current density (ΔJ). To quantify ΔJ, both the dark current density and the photocurrent density should be quantified. The tunneling current density of DBS can be calculated by the Tsu– Esaki equation [14] ! Z 1 þexp ðEf EÞ=kB T em kB T 1 dE J RT ¼ D ð E Þln ð1Þ 1 þ exp ðEf E eV D Þ=kB T 2π 2 ℏ3 0
incident light
VD+ΔVD Fig. 2. Schematic band diagram of the RTD-PD under light. Due to the lowering of the energy level of the double barrier structure and the repulsive force of the accumulated holes, each photogenerated hole keeps a spacing of r. The potential drop along the DBS (VD) will increase by ΔVD due to the holes accumulation.
where M11 is the element in the first row and first column of the transfer matrix. When the detector is under light, there will be photogenerated holes accumulating near the interface between the DBS and the absorption layer, which results to an increment of VD, as shown in Fig. 2. At first, the photogenerated holes will accumulate at the interface of the DBS and the absorption layer. However, due to the lowering of the energy level of the DBS and the repulsive force of the accumulated holes, subsequent photogenerated holes will not accumulate at the interface, but keep a distance away from it. Assuming that each photogenerated hole has a spacing of r, the increment of VD can be calculated by
ΔV D ¼
p 1 e X l þ ðn 1Þ r
4πε n ¼ 1
ð3Þ
where ε is the relative permittivity, p is the number of the accumulated holes, l is the thickness of the DBS. To calculate the total current density of the RTD-PD, the excess current component should be considered. The excess current is mainly composed of thermionic current, which can be calculated by 1=2 4kB TBJ RT J TH ¼ ð4Þ A UV D where B and A are the bandwidth and area of the RTD-PD respectively. According to equation (1)–(4), the current density of the RTDPD with varying numbers of accumulated holes is simulated as shown in Fig. 3. During the simulation, the temperature is set at 300 K, r is set as 5 nm, B is set as 100 kHz and A is set as 500 μm2. As there is a positive correlation between the power of the incident light and the numbers of the accumulated holes, the simulation results can represent the photo-response of the RTDPD. Negative differential resistance (NDR) is obtained in each curve in Fig. 3, resulting from the off-resonance between the electron energy and the resonance level of the DBS. When there are photogenerated holes accumulating near the interface of the DBS and the absorption layer, the peak voltages of the detector become lower. When p varies from 0 to 1, the peak voltage decreases by 0.03 V. However, when p varies from 200 to 300, the peak voltage only decreases by 0.01 V. Meanwhile, as RTD-PD possesses lower excess current density at lower bias voltage, the peak current density also decreases with the increment of p. Fig. 4 shows the calculated ΔJ with varying numbers of accumulated holes. When VD is lower than the peak voltage, ΔJ
276
Y. Dong et al. / Optics Communications 355 (2015) 274–278
Fig. 3. The current density variation with varying numbers of accumulated holes (p). The temperature is set at 300 K. The spacing of each accumulated holes is set as 5 nm. The bandwidth is set as 100 kHz and the area of the RTD-PD is set as 500 μm2.
increases with the increment of VD and p. At 0.5 V, ΔJ has an increment of 232 A/cm2 when p varies from 1 to 100 while the increment is 49 A/cm2 when p varies from 100 to 200. Therefore, with the power of the incident light increasing, the increasing speed of ΔJ slows down, resulting to a decrement of the responsivity. The photosensitive area of the RTD-PD has great influence on the responsivity of the detector. When the power of the incident light is constant, the power density increases when the photosensitive area decreases, which leads to higher increment of the tunneling current density. To consider the effect of the photosensitive area on the simulation results, the tunneling current variation under incident light with power density of 1 W/μm2 is utilized to quantify the responsivity of the RTD-PD, the unit of which is A/(W/μm2). Considering that the absorption coefficient of In0.53Ga0.47As for incident light at 1550 nm and 300 K is 7800 cm-1 [16]. By calculation, there will be an optical loss of 32.3% with 500 nm In0.53Ga0.47As absorption layer, so the responsivity of the RTD-PD can be calculated by R ¼ 67:7% U
ΔJ U A pEp =A
ð5Þ
where Ep is the photon energy. However, the simulation is two-dimensional, if we treat the simulation as three-dimensional, its “diameter” should be equivalent to the diameter of a molecule, so its photosensitive area is 1 nm2. Fig. 5 shows the calculated responsivity of the RTD-PD at wavelength of 1550 nm, indicating that the responsivity decreases while the power density of the incident light increases. The slope of the linear fitting line (red line in Fig. 5) has a standard error of 0.01. According to the linear fitting results, the relationship between the responsitivity and the power density of the incident light (P) can be got, given by R ¼ 0:07 U P ð 0:85 7 0:01Þ
ð6Þ
4. Experimental results
Fig. 4. The variation of the current density (ΔJ) with varying numbers of accumulated holes.
Fig. 5. The calculated responsivity of the RTD-PD at wavelength of 1550 nm. The potential drop along the DBS is 0.5 V.
Samples of the RTD-PD are fabricated by molecular beam epitaxy on InP substrate. The mesas of the detector are etched by inductively coupled plasma and Ti/Pt/Au ring contacts are deposited on the top of the structure to form clear apertures. To test the responsivity of the RTD-PD, a laser working at 1550 nm with a power of 5 mW is used as the light source. The laser light is focused on the photosensitive area by an optical fiber. The diameter of the tip of the optical fiber is 20 μm. The effective ratio between the illuminated area and the photosensitive area is calculated to be 0.64, so that the responsivity of the detector will be underestimated, as not all the absorption region is responsible for light absorption. The power of the light can be changed by connecting optical fiber attenuators with different attenuation coefficient to the optical fiber. Fig. 6 shows the photoresponse of the RTD-PD with 500 μm2 photosensitive area at room temperature. The devices are oscillating in the NDR region due to the bistability of the NDR region [17]. The current density of the detector increases under light, and the current increment have a positive correlation with the power density of the incident light. In Fig. 6(a), the peak voltages are larger than the simulation results. The reason is that, apart from the DBS, there are extra potential drop in the epitaxial layer of the RTD-PD and the series resistance of the current meter. The peak current density is larger than the simulation results, as noise current also contribute to the total current of the RTD-PD. The responsivity of the RTD-PD is further tested at room temperature and 77 K. Fig.7 shows the tested responsivity of the
Y. Dong et al. / Optics Communications 355 (2015) 274–278
277
Fig. 6. The photo-response of the RTD-PD at bias voltage of (a) 0–1.2 V and (b) 0.87–0.875 V at room temperature. The photosensitive area of the tested detector is 500 μm2.
solving the Tsu–Esaki equation. The simulation results show that, while the power density of the incident light increases, the responsitiviy of the RTD-PD decreases. The experimental results also show that the responsivity of the RTD-PD is nonconstant, which agree with the simulation results. The research results is significant for understanding the detecting mechanism of the detector, and provide basis for optimizing the performance of the RTD-PD.
Acknowledgments
Fig. 7. The tested responsivity at room temperature (RT) and 77 K when the RTDPD is biased at 0.87 V.
This work is supported by the 973 Program of China (Grant nos. 2013CB932904 and 2012CB932701), National Natural Science Foundation of China (Grant nos. 61274125, 61274013, and 61435012), Hebei Provincial Natural Science Foundation, China (Grant no. A2015506019), the Special Foundation for National Key Scientific Instrument of China (Grant no. 2012YQ140005), the Open Fund of High Power Laser Laboratory of Chinese Academy of Engineering Physics (Grant no. 2013HEL03). References
RTD-PD at 0.87 V together with the simulated responsivity. Considering that the effective ratio between the illuminated area and the photosensitive area is 0.64, the experimental results are divided by 0.64. It is shown that, the responsivity increases while the power density of the incident light decreases which agree with the analysis above. The responsivity at 77 K is about 1 order of magnitude higher than that at room temperature, indicating that the RTD-PD has better performance at lower temperature. The responsivity reaches 4.8 108 A/(W/μm2) at room temperature and 5.0 109 A/(W/μm2) at 77 K when the power density of the incident light is 1 10 13 W/μm2. The simulation results are higher than the experimental results both at room temperature and 77 K, as the noise from the detector and the experimental devices reduce the performance of the RTDPD. By lowering the dark current of the RTD-PD with a pre-barrier structure [18] or growing a thicker absorption layer, the performance of the RTD-PD can be promoted significantly.
5. Conclusion In summary, the nonconstant responsivity of RTD-PD is analyzed in this paper. By building the physical model of the responsivity of the RTD-PD, the current density variation with varying numbers of accumulated holes is simulated through
[1] R.H. Hadfield, Single-photon detectors for optical quantum information applications, Nat. Photonics 3 (2009) 696. [2] L. Zhou, Y.B. Sheng, Efficient single-photon entanglement concentration for quantum communications, Opt. Commun. 313 (2014) 217. [3] J.S. Pelc, Q. Zhang, C.R. Phillips, L. Yu, Y. Yamamoto, M.M. Fejer, Cascaded frequency upconversion for high-speed single-photon detection at 1550 nm, Opt. Lett. 37 (2012) 476. [4] A. Restelli, J.C. Bienfang, A.L. Migdall, Single-photon detection efficiency up to 50% at 1310 nm with an InGaAs/InP avalanche diode gated at 1.25 GHz, Appl. Phys. Lett. 102 (2013) 141104. [5] W.J. He, B. Sima, Y.F. Chen, H.D. Dai, Q. Chen, G.H. Gu, A correction method for range walk error in photon counting 3D imaging LIDAR, Opt. Commun. 308 (2013) 211. [6] R.H. Hadfield, J.L. Habif, J. Schlafer, R.E. Schwall, S.W. Nam, Quantum key distribution at 1550 nm with twin superconducting single-photon detectors, Appl. Phys. Lett. 89 (2006) 241129. [7] C. Schuck, W.H.P. Pernice, H.X. Tang, NbTiN superconducting nanowire detectors for visible and telecom wavelengths single photon counting on Si3N4 photonic circuits, Appl. Phys. Lett. 102 (2013) 051101. [8] Y. Kang, H.X. Lu, Y.H. Lo, D.S. Bethune, W.P. Risk, Dark count probability and quantum efficiency of avalanche photodiodes for single-photon detection, Appl. Phys. Lett. 83 (2003) 2955. [9] T. Yamashita, S. Miki, W. Qiu, M. Fujiwara, M. Sasaki, Z. Wang, Temperature dependent performances of superconducting nanowire single-photon detectors in an ultralow-temperature region, Appl. Phys. Express 3 (2010) 102502. [10] S. Kohjiro, F. Hirayama, H. Yamamori, S. Nagasawa, D. Fukuda, M. Hidaka, White noise of Nb-based microwave superconducting quantum interference device multiplexers with NbN coplanar resonators for readout of transition edge sensors, J. Appl. Phys. 115 (2014) 223902. [11] J.C. Blakesley, P. See, A.J. Shields, B.E. Kardynał, P. Atkinson, I. Farrer, D. A. Ritchie, Efficient single photon detection by quantum dot resonant tunneling diodes, Phys. Rev. Lett. 94 (2005) 067401.
278
Y. Dong et al. / Optics Communications 355 (2015) 274–278
[12] H.W. Li, B.E. Kardyna, P. See, A.J. Shields, P. Simmonds, H.E. Beere, D.A. Ritchie, Quantum dot resonant tunneling diode for telecommunication wavelength single photon detection, Appl. Phys. Lett. 91 (2007) 073516. [13] F. Hartmann, F. Langer, D. Bisping, A. Musterer, S. Höfling, M. Kamp, A. Forchel, L. Worschech, GaAs/AlGaAs resonant tunneling diodes with a GaInNAs absorption layer for telecommunication light sensing, Appl. Phys. Lett. 100 (2012) 172113. [14] R. Tsu, L. Esaki, Tunneling in a finite superlattice, Appl. Phys. Lett. 22 (1973) 562. [15] S. Vatannia, G. Gildenblat, Airy’s functions implementation of the transfermatrix method for resonant tunneling in variably spaced finite superlattices, IEEE J. Quantum Electron. 32 (1996) 1093.
[16] E. Zielinski, H. Schweizer, K. Streubel, H. Eisele, G. Weimann, Excitonic transitions and exciton damping processes in InGaAs/InP, J. Appl. Phys. 59 (1986) 2196. [17] K.L. Jensen, F.A. Buot, Numerical simulation of intrinsic bistability and highfrequency current oscillations in resonant tunneling structures, Phys. Rev. Lett. 66 (1991) 1078. [18] K. Maezawa, H. Matsuzaki, M. Yamamoto, T. Otsuji, High-speed and low-power operation of a resonant tunneling logic gate, IEEE Electron Device Lett. 19 (1998) 80.
Contents lists available at ScienceDirect
Optics Communications journal homepage: www.elsevier.com/locate/optcom
Resonant tunneling diode photodetector with nonconstant responsivity Yu Dong a, Guanglong Wang a,n, Haiqiao Ni b,n, Jianhui Chen a, Fengqi Gao a, Baochen Li a, Kangming Pei b, Zhichuan Niu b a b
Laboratory of Nanotechnology and Microsystems, Mechanical Engineering College, Shijiazhuang 050000, China Institute of Semiconductors, Chinese Academy of Science, Beijing 100084, China
art ic l e i nf o
a b s t r a c t
Article history: Received 2 April 2015 Received in revised form 24 June 2015 Accepted 27 June 2015
Resonant tunneling diode with an In0.53Ga0.47As absorption layer is designed for light detection at 1550 nm. The responsivity of the detector is simulated by solving the Tsu–Esaki equation. The simulation results show that the responsivity of the detector is nonconstant. It decreases with the increment of the power density of the incident light. Samples of the detector are fabricated by molecular beam epitaxy. The experimental results show that the responsivity increases while the power density of the incident light decreases which agree with the simulation results. The responsivity reaches 4.8 108 A/(W/μm2) at room temperature and 5.0 109 A/(W/μm2) at 77 K when the power density of the incident light is 1 10 13 W/μm2. & 2015 Elsevier B.V. All rights reserved.
Keywords: Resonant tunneling diode Light detection Responsivity
1. Introduction Light detection at wavelength of 1550 nm with high sensitivity is of great importance for quantum communication [1–3]. At present, many types of detectors have been investigated such as avalanche photodiode (APD) and superconducting single-photon detector (SSPD) [4–7]. As APD rely on the avalanche multiplication effect, it has disadvantages of high dark count rate and afterpulse noise [8]. The SSPD is based on superconducting materials' sensitivity to the change in temperature. The SSPD is suitable for operation at 1550 nm with low dark count rate, high detection efficiency and speed. However, SSPD has to be operated at extremely low temperature which is a serious constraint for its applications[9,10]. In 2005, Blakesley [11] showed that resonant tunneling diode (RTD) which contains a quantum dot layer and absorption layer is capable of single photon detection around 850 nm. When the incident photons are absorbed by the absorption region, the photo-excited holes will move to the emitter side and be captured by the quantum dots layer, which causes holes accumulation and potential modulation near the double barrier structure (DBS). In 2007, the detection wavelength of this type of detector was extended to 1310 nm by utilizing In0.53Ga0.47As absorption layer [12]. This type of detector works at temperature lower than 77 K to
n
Corresponding authors. E-mail addresses: [email protected] (G. Wang), [email protected] (H. Ni).
http://dx.doi.org/10.1016/j.optcom.2015.06.064 0030-4018/& 2015 Elsevier B.V. All rights reserved.
achieve single photon sensitivity. In 2012, Hartmann [13] implemented light detection at 1.3 μm by adding a GaInNAs absorption layer into GaAs/AlGaAs RTD. Without a quantum dot layer, this type of detector is still capable of light detection with responsitivity of 1000 A/W at room temperature. The resonant tunneling diode photodetector (RTD-PD) has advantages of low working voltage, high quantum efficiency and low dark count rate, which is promising for light detection with high sensitivity at room temperature. In this paper, we add In0.53Ga0.47As absorption layer into InGaAs/AlAs RTD to implement light detection at 1550 nm with high sensitivity. We find out the responsivity of the RTD-PD is nonconstant, which has not been reported.
2. Structure of the RTD-PD The structure of the RTD-PD is shown in Fig. 1. The n-type In0.53Ga0.47As layers on the top and bottom act as collector and emitter. The DBS consists of 1.4 nm AlAs, 6 nm In0.53Ga0.47As and 1.4 nm AlAs. 500 nm intrinsic In0.53Ga0.47As absorption layer is placed between collector and the DBS for light absorption. 15 nm In0.53Ga0.47As spacer is placed between the DBS and emitter to prevent impurity scattering from emitter to the DBS. A ring contact is deposited on the top to form electrical contact and clear aperture. The diameter of the clear aperture and the ring contact is 25 μm and 29 μm respectively.
Y. Dong et al. / Optics Communications 355 (2015) 274–278
275
3. Simulation of the responsivity of the detector
incident light ring contact
n-InGaAs 100 nm
where e is the electron charge, m* is the effective mass in emitter, kB is the Boltzmann constant, ℏ is the reduced Planck constant, T is the absolute temperature, Ef is the Fermi energy in emitter, E is the electron energy along the longitudinal growth direction of the detector, VD is the potential drop along the DBS, and D(E) is the transmission coefficient of the DBS which can be calculated by transfer matrix method [15], given by 1 DðEÞ ¼ M 11 M 11 ð2Þ
InGaAs 500 nm AlAs 1.4 nm InGaAs 6 nm AlAs 1.4 nm InGaAs 15 nm n-InGaAs 100 nm Fig. 1. Structure of the RTD-PD. The DBS is consisted of 1.4 nm AlAs, 6 nm In0.53Ga0.47As and 1.4 nm AlAs. A ring contact is deposited on the top to form electrical contact and clear aperture.
Incident light at wavelength of 1550 nm is absorbed by the In0.53Ga0.47As absorption layer, producing photogenerated electron–hole pairs. When the detector is under positive bias, the photogenerated holes will drift to the emitter side and accumulate at the interface between the DBS and the absorption layer. This causes band bending through the DBS, resulting to the variation of the tunneling current density of the RTD-PD. Meanwhile, if the incident light is not fully absorbed by the absorption layer, the light will be further absorbed by the spacer layer generating electron–hole pairs. The photogenerated electrons will accumulate near the interface between the DBS and the spacer layer, inducing holes on the other side of the DBS. This effect can promote the photo-response of the RTD-PD, but considering the thickness of the spacer layer is only 15 nm, this effect is negligible.
emitter collector + r
+
r
+
The responsivity of the detector comes from the variation of the tunneling current density (ΔJ). To quantify ΔJ, both the dark current density and the photocurrent density should be quantified. The tunneling current density of DBS can be calculated by the Tsu– Esaki equation [14] ! Z 1 þexp ðEf EÞ=kB T em kB T 1 dE J RT ¼ D ð E Þln ð1Þ 1 þ exp ðEf E eV D Þ=kB T 2π 2 ℏ3 0
incident light
VD+ΔVD Fig. 2. Schematic band diagram of the RTD-PD under light. Due to the lowering of the energy level of the double barrier structure and the repulsive force of the accumulated holes, each photogenerated hole keeps a spacing of r. The potential drop along the DBS (VD) will increase by ΔVD due to the holes accumulation.
where M11 is the element in the first row and first column of the transfer matrix. When the detector is under light, there will be photogenerated holes accumulating near the interface between the DBS and the absorption layer, which results to an increment of VD, as shown in Fig. 2. At first, the photogenerated holes will accumulate at the interface of the DBS and the absorption layer. However, due to the lowering of the energy level of the DBS and the repulsive force of the accumulated holes, subsequent photogenerated holes will not accumulate at the interface, but keep a distance away from it. Assuming that each photogenerated hole has a spacing of r, the increment of VD can be calculated by
ΔV D ¼
p 1 e X l þ ðn 1Þ r
4πε n ¼ 1
ð3Þ
where ε is the relative permittivity, p is the number of the accumulated holes, l is the thickness of the DBS. To calculate the total current density of the RTD-PD, the excess current component should be considered. The excess current is mainly composed of thermionic current, which can be calculated by 1=2 4kB TBJ RT J TH ¼ ð4Þ A UV D where B and A are the bandwidth and area of the RTD-PD respectively. According to equation (1)–(4), the current density of the RTDPD with varying numbers of accumulated holes is simulated as shown in Fig. 3. During the simulation, the temperature is set at 300 K, r is set as 5 nm, B is set as 100 kHz and A is set as 500 μm2. As there is a positive correlation between the power of the incident light and the numbers of the accumulated holes, the simulation results can represent the photo-response of the RTDPD. Negative differential resistance (NDR) is obtained in each curve in Fig. 3, resulting from the off-resonance between the electron energy and the resonance level of the DBS. When there are photogenerated holes accumulating near the interface of the DBS and the absorption layer, the peak voltages of the detector become lower. When p varies from 0 to 1, the peak voltage decreases by 0.03 V. However, when p varies from 200 to 300, the peak voltage only decreases by 0.01 V. Meanwhile, as RTD-PD possesses lower excess current density at lower bias voltage, the peak current density also decreases with the increment of p. Fig. 4 shows the calculated ΔJ with varying numbers of accumulated holes. When VD is lower than the peak voltage, ΔJ
276
Y. Dong et al. / Optics Communications 355 (2015) 274–278
Fig. 3. The current density variation with varying numbers of accumulated holes (p). The temperature is set at 300 K. The spacing of each accumulated holes is set as 5 nm. The bandwidth is set as 100 kHz and the area of the RTD-PD is set as 500 μm2.
increases with the increment of VD and p. At 0.5 V, ΔJ has an increment of 232 A/cm2 when p varies from 1 to 100 while the increment is 49 A/cm2 when p varies from 100 to 200. Therefore, with the power of the incident light increasing, the increasing speed of ΔJ slows down, resulting to a decrement of the responsivity. The photosensitive area of the RTD-PD has great influence on the responsivity of the detector. When the power of the incident light is constant, the power density increases when the photosensitive area decreases, which leads to higher increment of the tunneling current density. To consider the effect of the photosensitive area on the simulation results, the tunneling current variation under incident light with power density of 1 W/μm2 is utilized to quantify the responsivity of the RTD-PD, the unit of which is A/(W/μm2). Considering that the absorption coefficient of In0.53Ga0.47As for incident light at 1550 nm and 300 K is 7800 cm-1 [16]. By calculation, there will be an optical loss of 32.3% with 500 nm In0.53Ga0.47As absorption layer, so the responsivity of the RTD-PD can be calculated by R ¼ 67:7% U
ΔJ U A pEp =A
ð5Þ
where Ep is the photon energy. However, the simulation is two-dimensional, if we treat the simulation as three-dimensional, its “diameter” should be equivalent to the diameter of a molecule, so its photosensitive area is 1 nm2. Fig. 5 shows the calculated responsivity of the RTD-PD at wavelength of 1550 nm, indicating that the responsivity decreases while the power density of the incident light increases. The slope of the linear fitting line (red line in Fig. 5) has a standard error of 0.01. According to the linear fitting results, the relationship between the responsitivity and the power density of the incident light (P) can be got, given by R ¼ 0:07 U P ð 0:85 7 0:01Þ
ð6Þ
4. Experimental results
Fig. 4. The variation of the current density (ΔJ) with varying numbers of accumulated holes.
Fig. 5. The calculated responsivity of the RTD-PD at wavelength of 1550 nm. The potential drop along the DBS is 0.5 V.
Samples of the RTD-PD are fabricated by molecular beam epitaxy on InP substrate. The mesas of the detector are etched by inductively coupled plasma and Ti/Pt/Au ring contacts are deposited on the top of the structure to form clear apertures. To test the responsivity of the RTD-PD, a laser working at 1550 nm with a power of 5 mW is used as the light source. The laser light is focused on the photosensitive area by an optical fiber. The diameter of the tip of the optical fiber is 20 μm. The effective ratio between the illuminated area and the photosensitive area is calculated to be 0.64, so that the responsivity of the detector will be underestimated, as not all the absorption region is responsible for light absorption. The power of the light can be changed by connecting optical fiber attenuators with different attenuation coefficient to the optical fiber. Fig. 6 shows the photoresponse of the RTD-PD with 500 μm2 photosensitive area at room temperature. The devices are oscillating in the NDR region due to the bistability of the NDR region [17]. The current density of the detector increases under light, and the current increment have a positive correlation with the power density of the incident light. In Fig. 6(a), the peak voltages are larger than the simulation results. The reason is that, apart from the DBS, there are extra potential drop in the epitaxial layer of the RTD-PD and the series resistance of the current meter. The peak current density is larger than the simulation results, as noise current also contribute to the total current of the RTD-PD. The responsivity of the RTD-PD is further tested at room temperature and 77 K. Fig.7 shows the tested responsivity of the
Y. Dong et al. / Optics Communications 355 (2015) 274–278
277
Fig. 6. The photo-response of the RTD-PD at bias voltage of (a) 0–1.2 V and (b) 0.87–0.875 V at room temperature. The photosensitive area of the tested detector is 500 μm2.
solving the Tsu–Esaki equation. The simulation results show that, while the power density of the incident light increases, the responsitiviy of the RTD-PD decreases. The experimental results also show that the responsivity of the RTD-PD is nonconstant, which agree with the simulation results. The research results is significant for understanding the detecting mechanism of the detector, and provide basis for optimizing the performance of the RTD-PD.
Acknowledgments
Fig. 7. The tested responsivity at room temperature (RT) and 77 K when the RTDPD is biased at 0.87 V.
This work is supported by the 973 Program of China (Grant nos. 2013CB932904 and 2012CB932701), National Natural Science Foundation of China (Grant nos. 61274125, 61274013, and 61435012), Hebei Provincial Natural Science Foundation, China (Grant no. A2015506019), the Special Foundation for National Key Scientific Instrument of China (Grant no. 2012YQ140005), the Open Fund of High Power Laser Laboratory of Chinese Academy of Engineering Physics (Grant no. 2013HEL03). References
RTD-PD at 0.87 V together with the simulated responsivity. Considering that the effective ratio between the illuminated area and the photosensitive area is 0.64, the experimental results are divided by 0.64. It is shown that, the responsivity increases while the power density of the incident light decreases which agree with the analysis above. The responsivity at 77 K is about 1 order of magnitude higher than that at room temperature, indicating that the RTD-PD has better performance at lower temperature. The responsivity reaches 4.8 108 A/(W/μm2) at room temperature and 5.0 109 A/(W/μm2) at 77 K when the power density of the incident light is 1 10 13 W/μm2. The simulation results are higher than the experimental results both at room temperature and 77 K, as the noise from the detector and the experimental devices reduce the performance of the RTDPD. By lowering the dark current of the RTD-PD with a pre-barrier structure [18] or growing a thicker absorption layer, the performance of the RTD-PD can be promoted significantly.
5. Conclusion In summary, the nonconstant responsivity of RTD-PD is analyzed in this paper. By building the physical model of the responsivity of the RTD-PD, the current density variation with varying numbers of accumulated holes is simulated through
[1] R.H. Hadfield, Single-photon detectors for optical quantum information applications, Nat. Photonics 3 (2009) 696. [2] L. Zhou, Y.B. Sheng, Efficient single-photon entanglement concentration for quantum communications, Opt. Commun. 313 (2014) 217. [3] J.S. Pelc, Q. Zhang, C.R. Phillips, L. Yu, Y. Yamamoto, M.M. Fejer, Cascaded frequency upconversion for high-speed single-photon detection at 1550 nm, Opt. Lett. 37 (2012) 476. [4] A. Restelli, J.C. Bienfang, A.L. Migdall, Single-photon detection efficiency up to 50% at 1310 nm with an InGaAs/InP avalanche diode gated at 1.25 GHz, Appl. Phys. Lett. 102 (2013) 141104. [5] W.J. He, B. Sima, Y.F. Chen, H.D. Dai, Q. Chen, G.H. Gu, A correction method for range walk error in photon counting 3D imaging LIDAR, Opt. Commun. 308 (2013) 211. [6] R.H. Hadfield, J.L. Habif, J. Schlafer, R.E. Schwall, S.W. Nam, Quantum key distribution at 1550 nm with twin superconducting single-photon detectors, Appl. Phys. Lett. 89 (2006) 241129. [7] C. Schuck, W.H.P. Pernice, H.X. Tang, NbTiN superconducting nanowire detectors for visible and telecom wavelengths single photon counting on Si3N4 photonic circuits, Appl. Phys. Lett. 102 (2013) 051101. [8] Y. Kang, H.X. Lu, Y.H. Lo, D.S. Bethune, W.P. Risk, Dark count probability and quantum efficiency of avalanche photodiodes for single-photon detection, Appl. Phys. Lett. 83 (2003) 2955. [9] T. Yamashita, S. Miki, W. Qiu, M. Fujiwara, M. Sasaki, Z. Wang, Temperature dependent performances of superconducting nanowire single-photon detectors in an ultralow-temperature region, Appl. Phys. Express 3 (2010) 102502. [10] S. Kohjiro, F. Hirayama, H. Yamamori, S. Nagasawa, D. Fukuda, M. Hidaka, White noise of Nb-based microwave superconducting quantum interference device multiplexers with NbN coplanar resonators for readout of transition edge sensors, J. Appl. Phys. 115 (2014) 223902. [11] J.C. Blakesley, P. See, A.J. Shields, B.E. Kardynał, P. Atkinson, I. Farrer, D. A. Ritchie, Efficient single photon detection by quantum dot resonant tunneling diodes, Phys. Rev. Lett. 94 (2005) 067401.
278
Y. Dong et al. / Optics Communications 355 (2015) 274–278
[12] H.W. Li, B.E. Kardyna, P. See, A.J. Shields, P. Simmonds, H.E. Beere, D.A. Ritchie, Quantum dot resonant tunneling diode for telecommunication wavelength single photon detection, Appl. Phys. Lett. 91 (2007) 073516. [13] F. Hartmann, F. Langer, D. Bisping, A. Musterer, S. Höfling, M. Kamp, A. Forchel, L. Worschech, GaAs/AlGaAs resonant tunneling diodes with a GaInNAs absorption layer for telecommunication light sensing, Appl. Phys. Lett. 100 (2012) 172113. [14] R. Tsu, L. Esaki, Tunneling in a finite superlattice, Appl. Phys. Lett. 22 (1973) 562. [15] S. Vatannia, G. Gildenblat, Airy’s functions implementation of the transfermatrix method for resonant tunneling in variably spaced finite superlattices, IEEE J. Quantum Electron. 32 (1996) 1093.
[16] E. Zielinski, H. Schweizer, K. Streubel, H. Eisele, G. Weimann, Excitonic transitions and exciton damping processes in InGaAs/InP, J. Appl. Phys. 59 (1986) 2196. [17] K.L. Jensen, F.A. Buot, Numerical simulation of intrinsic bistability and highfrequency current oscillations in resonant tunneling structures, Phys. Rev. Lett. 66 (1991) 1078. [18] K. Maezawa, H. Matsuzaki, M. Yamamoto, T. Otsuji, High-speed and low-power operation of a resonant tunneling logic gate, IEEE Electron Device Lett. 19 (1998) 80.