Supporting Information for: GaAs/AlGaAs Nanowire Photodetector
Supporting Information for: GaAs/AlGaAs Nanowire Photodetector
Xing Dai,1,† Sen Zhang,1,2,3,† Zilong Wang,1,4 Giorgio Adamo,4 Hai Liu,5 Yizhong Huang,5 Christophe Couteau,2,4,6 and Cesare Soci1,2,4,*
1
Division of Physics and Applied Physics, School of Physical and Mathematical Sciences, Nanyang Technological University, Singapore 637371 2
CINTRA CNRS-NTU-Thales, UMI 3288, Singapore 637553 3
4
Tang Optoelectronics Equipment Co., China 201203
Centre for Disruptive Photonic Technologies, Nanyang Technological University, Singapore 637371 5
6
School of Material Science and Engineering, Nanyang Technological University, Singapore 639798
Laboratory for Nanotechnology, Instrumentation and Optics, University of Technology of Troyes, 10000 Troyes, France
† These authors equally contributed to the work. * Corresponding author: [email protected]
Atomic composition. Energy dispersive X-ray (EDX) analysis was performed on an asgrown GaAs/AlGaAs nanowire with a thick shell to increase accuracy. The core GaAs nanowire was grown from a 40 nm diameter Au nanoparticle at 430 oC for 5 min. The AlGaAs shell was then grown at 630 oC for 10 min, leading to a core-shell nanowire with overall diameter of 680 nm, as shown in the inset of Fig. S1. All gas flow rates were the same as those reported in the manuscript. Compositional measurements were conducted point-by-point along the nanowire axis in a Jeol JSM-6700F SEM equipped with EDX detector. Fig. S1 shows Al content in the nanowire shell of ~14-18 %. The composition is rather uniform along the nanowire axis, showing some variation near the nanowire tip,
Atomic Composition (%)
most likely due to reactant diffusion on the sidewalls near the Au catalyst.
50 40 30 20 As Ga Al
10 0
0
5
10
Distance from base to top (m)
Figure S1. Atomic composition of a thick AlGaAs shell in a GaAs/AlGaAs nanowire as a function of position along the nanowire axis obtained by EDX spectroscopy. Inset: SEM image of the GaAs/AlGaAs nanowire with core diameter of ~40 nm and overall diameter of 680 nm. Scale bar is 1 m.
Photodetector fabrication. The single core-multishell GaAs/hight-T GaAs/AlGaAs nanowire was removed from the initial substrate by ultrasonication in ethanol solution and dispersed onto a SiO2/Si substrate patterned with alignment markers. The contact to the GaAs core was defined by electron beam lithography (EBL), followed by wet etch in a mixture of citric acid liquid (C6H8O7:H2O=1:1 by weight) and H2O2 (30%). Firstly, the sample was dipped into a solution of citric acid: H2O2= 20:1, which promotes a high AlGaAs etching rate ~3 nm/s, for AlGaAs shell removal. Consecutively, the sample was immersed into another solution of citric acid:H2O2=1:1 for 6 min, then rinsed in DI water, leaving only ~70-80 nm GaAs core. 100 nm thick Pt was then deposited on the GaAs core acting as one electrode, while the other electrode was deposited on the AlGaAs shell. Deposition was performed with a FIB system by direct exposure to a focused Ga+ ion beam and employing (CH3)Pt(CpCH3) as the Pt source.
Simulation details. To map the electron distribution in the cross-section of nanowires, self-consistent Schrödinger and Poisson equations were solved by using Silvaco-Atlas simulation. Fig. S1 shows the contour plots in linear scale of (a) GaAs nanowire (r=40 nm), (b) GaAs/high-T GaAs nanowire (40 nm/170 nm), and (c) core-multishell GaAs/high-T GaAs/AlGaAs nanowire (40 nm/170 nm/30 nm), corresponding to plots (log scale) in Fig. 1a-c. Electron confinement at the heterointerface can be clearly observed in Fig. S2c.
(a)
(b)
(c)
Electron Conc (cm-3)
1x1018
7.5x1017 5x1017 2.5x1017 1x1010
Figure S2. Simulated electron contour plot of the cross-sectional plane of (a) core GaAs nanowire (r=40 nm), (b) GaAs/high-T GaAs nanowire (40 nm/170 nm), and (c) coremultishell GaAs/high-T GaAs/AlGaAs nanowire (40 nm/170 nm/30 nm). The scale bar is in linear scale.
Silvaco-Atlas simulation was also used to simulate carrier transport in the photodetector both in the dark and under illumination. Two carrier mobility models are considered, including concentration-dependent low field mobility model and fielddependent mobility model. The first one allows the system to look up for doping concentration related mobility in a default table, while the second one includes velocity saturation effect on mobility in high electric field. In addition, several carrier generation and recombination models are taken into account when solving the Boltzmann transport equation: concentration-lifetime dependent Shockley-Read-Hall (SRH) recombination, auger recombination, and optical recombination.1 Optical spectrum is calculated using full wave simulation software COMSOL Multiphysics® assuming a TM polarized wave normally incident on the nanowire. Optical absorption is obtained by integrating the product of Electric Field and Current Density over the hexagonal cross-section of the nanowire.2 The wavelength range is span using 5 nm scanning step.
Photocurrent dependence on illumination intensity. The light intensity dependence of the photocurrent of a single core-multishell nanowire is shown in Fig. S3. Data were collected at room temperature under green laser illumination (λ=532 nm), with intensities of 2, 10 and 20 mW/cm2. The photocurrent at small bias is linearly proportional to the applied voltage between core and shell electrodes.
Figure S3. Current-voltage characteristics of a core-mutlishell GaAs/high-T GaAs/AlGaAs nanowire photodetecor measured at room temperature as function of laser intensity, with excitation wavelength of 532 nm.
Photo-responsivity and detectivity determination. Responsivity, the ratio of electrical
output to the optical input, is calculated following the equation Ri
i ph PD in
P
i ph I in APD
, where
i ph , PinPD , I in , APD are photocurrent, light power incident onto the surface of the nanowire detector, light intensity, and the exposed surface area of the nanowire photodetector,
respectively. The exposed surface area used for calculation is estimated to be half of the nanowire’s cylindrical surface area whose dimensions are taken from SEM image of the device, APD
1 2 r PDl , where r PD is the nanowire radius, l is the length of the core2
shell portion of the nanowire between the electrodes as shown in Fig 3c.3 Responsivity measurement was done with conventional amplitude modulation technique: a Xe lamp was used as white light source and dispersed by a monochromator within the range of 300 to 900 nm and modulated by a mechanical chopper at frequency of 138 Hz. Monochromatic light intensity was determined by a calibrated reference photodiode. Time constant of the lock-in amplifier was set to 300 ms, which corresponds to 0.42 Hz equivalent noise bandwidth. The resulting responsivity spectrum of the nanowire device at room temperature, shown in Fig. 4(a), corresponds well with typical GaAs photoresponse. This indicates a limited contribution from the AlGaAs shell. The nanowire has high photo-responsivity for wavelengths between 400 and 890 nm. At λ =855 nm, the peak responsivity of the nanowire device is 0.57 A/W, which is higher than commercial planar GaAs photodetectors.4 Specific detectivity (D*) was calculated to assess the sensitivity of the photodetector D*
against
noise
A(f ) Ri 2 NEP inoise
in
terms
of
noise
equivalent
power
(NEP):
A(f ) , where A is the detector area, f is the noise bandwidth,
inoise is the noise current and Ri is the above mentioned photo-responsivity. Measured responsivity and dark current were used to estimate the value of specific detectivity. Among all possible noise sources, 1/f noise current is removed effectively by the lock-in
amplifier, while thermal noise current i jo
4kT (f ) , where R is shunt resistance in the R
detector, is estimated to be two orders of magnitude lower than shot noise from dark current.3 Therefore specific detectivity was calculated taking into account only shot noise from dark current, given that the photocurrent is smaller than the dark current. At 2 V bias D*
and Ri 2 noise
i
=855
A(f )
i ph I in 2qid A
nm,
the
specific
7.20 1010 cm Hz / W
(at
detectivity room
is
temperature,
modulation frequency of 138 Hz, and with noise bandwidth of lock-in amplifier of 0.42 Hz). The specific detectivity of the nanowire photodetector is limited by its dark current.
References 1. https://dynamic.silvaco.com/dynamicweb/jsp/downloads/DownloadManualsAction.do?
req=silen-manuals&nm=atlas, Page 1456. 2.
Kempa, T. J.; Cahoon, J. F.; Kim, S.-K.; Day, R. W.; Bell, D. C.; Park, H.-G.;
Lieber, C. M., Proc. Natl. Acad. Sci. U. S. A 2012, 109, 1409-1412. 3.
Soci, C.; Zhang, A.; Bao, X. Y.; Kim, H.; Lo, Y.; Wang, D. L., J. Nanosci.
Nanotechnol. 2010, 10, 1430-1449. 4.
http://search.newport.com/?q=*&x2=sku&q2=818-BB-45, Biased Photodetector,
400-900nm, GaAs, 12.5 GHz.
Xing Dai,1,† Sen Zhang,1,2,3,† Zilong Wang,1,4 Giorgio Adamo,4 Hai Liu,5 Yizhong Huang,5 Christophe Couteau,2,4,6 and Cesare Soci1,2,4,*
1
Division of Physics and Applied Physics, School of Physical and Mathematical Sciences, Nanyang Technological University, Singapore 637371 2
CINTRA CNRS-NTU-Thales, UMI 3288, Singapore 637553 3
4
Tang Optoelectronics Equipment Co., China 201203
Centre for Disruptive Photonic Technologies, Nanyang Technological University, Singapore 637371 5
6
School of Material Science and Engineering, Nanyang Technological University, Singapore 639798
Laboratory for Nanotechnology, Instrumentation and Optics, University of Technology of Troyes, 10000 Troyes, France
† These authors equally contributed to the work. * Corresponding author: [email protected]
Atomic composition. Energy dispersive X-ray (EDX) analysis was performed on an asgrown GaAs/AlGaAs nanowire with a thick shell to increase accuracy. The core GaAs nanowire was grown from a 40 nm diameter Au nanoparticle at 430 oC for 5 min. The AlGaAs shell was then grown at 630 oC for 10 min, leading to a core-shell nanowire with overall diameter of 680 nm, as shown in the inset of Fig. S1. All gas flow rates were the same as those reported in the manuscript. Compositional measurements were conducted point-by-point along the nanowire axis in a Jeol JSM-6700F SEM equipped with EDX detector. Fig. S1 shows Al content in the nanowire shell of ~14-18 %. The composition is rather uniform along the nanowire axis, showing some variation near the nanowire tip,
Atomic Composition (%)
most likely due to reactant diffusion on the sidewalls near the Au catalyst.
50 40 30 20 As Ga Al
10 0
0
5
10
Distance from base to top (m)
Figure S1. Atomic composition of a thick AlGaAs shell in a GaAs/AlGaAs nanowire as a function of position along the nanowire axis obtained by EDX spectroscopy. Inset: SEM image of the GaAs/AlGaAs nanowire with core diameter of ~40 nm and overall diameter of 680 nm. Scale bar is 1 m.
Photodetector fabrication. The single core-multishell GaAs/hight-T GaAs/AlGaAs nanowire was removed from the initial substrate by ultrasonication in ethanol solution and dispersed onto a SiO2/Si substrate patterned with alignment markers. The contact to the GaAs core was defined by electron beam lithography (EBL), followed by wet etch in a mixture of citric acid liquid (C6H8O7:H2O=1:1 by weight) and H2O2 (30%). Firstly, the sample was dipped into a solution of citric acid: H2O2= 20:1, which promotes a high AlGaAs etching rate ~3 nm/s, for AlGaAs shell removal. Consecutively, the sample was immersed into another solution of citric acid:H2O2=1:1 for 6 min, then rinsed in DI water, leaving only ~70-80 nm GaAs core. 100 nm thick Pt was then deposited on the GaAs core acting as one electrode, while the other electrode was deposited on the AlGaAs shell. Deposition was performed with a FIB system by direct exposure to a focused Ga+ ion beam and employing (CH3)Pt(CpCH3) as the Pt source.
Simulation details. To map the electron distribution in the cross-section of nanowires, self-consistent Schrödinger and Poisson equations were solved by using Silvaco-Atlas simulation. Fig. S1 shows the contour plots in linear scale of (a) GaAs nanowire (r=40 nm), (b) GaAs/high-T GaAs nanowire (40 nm/170 nm), and (c) core-multishell GaAs/high-T GaAs/AlGaAs nanowire (40 nm/170 nm/30 nm), corresponding to plots (log scale) in Fig. 1a-c. Electron confinement at the heterointerface can be clearly observed in Fig. S2c.
(a)
(b)
(c)
Electron Conc (cm-3)
1x1018
7.5x1017 5x1017 2.5x1017 1x1010
Figure S2. Simulated electron contour plot of the cross-sectional plane of (a) core GaAs nanowire (r=40 nm), (b) GaAs/high-T GaAs nanowire (40 nm/170 nm), and (c) coremultishell GaAs/high-T GaAs/AlGaAs nanowire (40 nm/170 nm/30 nm). The scale bar is in linear scale.
Silvaco-Atlas simulation was also used to simulate carrier transport in the photodetector both in the dark and under illumination. Two carrier mobility models are considered, including concentration-dependent low field mobility model and fielddependent mobility model. The first one allows the system to look up for doping concentration related mobility in a default table, while the second one includes velocity saturation effect on mobility in high electric field. In addition, several carrier generation and recombination models are taken into account when solving the Boltzmann transport equation: concentration-lifetime dependent Shockley-Read-Hall (SRH) recombination, auger recombination, and optical recombination.1 Optical spectrum is calculated using full wave simulation software COMSOL Multiphysics® assuming a TM polarized wave normally incident on the nanowire. Optical absorption is obtained by integrating the product of Electric Field and Current Density over the hexagonal cross-section of the nanowire.2 The wavelength range is span using 5 nm scanning step.
Photocurrent dependence on illumination intensity. The light intensity dependence of the photocurrent of a single core-multishell nanowire is shown in Fig. S3. Data were collected at room temperature under green laser illumination (λ=532 nm), with intensities of 2, 10 and 20 mW/cm2. The photocurrent at small bias is linearly proportional to the applied voltage between core and shell electrodes.
Figure S3. Current-voltage characteristics of a core-mutlishell GaAs/high-T GaAs/AlGaAs nanowire photodetecor measured at room temperature as function of laser intensity, with excitation wavelength of 532 nm.
Photo-responsivity and detectivity determination. Responsivity, the ratio of electrical
output to the optical input, is calculated following the equation Ri
i ph PD in
P
i ph I in APD
, where
i ph , PinPD , I in , APD are photocurrent, light power incident onto the surface of the nanowire detector, light intensity, and the exposed surface area of the nanowire photodetector,
respectively. The exposed surface area used for calculation is estimated to be half of the nanowire’s cylindrical surface area whose dimensions are taken from SEM image of the device, APD
1 2 r PDl , where r PD is the nanowire radius, l is the length of the core2
shell portion of the nanowire between the electrodes as shown in Fig 3c.3 Responsivity measurement was done with conventional amplitude modulation technique: a Xe lamp was used as white light source and dispersed by a monochromator within the range of 300 to 900 nm and modulated by a mechanical chopper at frequency of 138 Hz. Monochromatic light intensity was determined by a calibrated reference photodiode. Time constant of the lock-in amplifier was set to 300 ms, which corresponds to 0.42 Hz equivalent noise bandwidth. The resulting responsivity spectrum of the nanowire device at room temperature, shown in Fig. 4(a), corresponds well with typical GaAs photoresponse. This indicates a limited contribution from the AlGaAs shell. The nanowire has high photo-responsivity for wavelengths between 400 and 890 nm. At λ =855 nm, the peak responsivity of the nanowire device is 0.57 A/W, which is higher than commercial planar GaAs photodetectors.4 Specific detectivity (D*) was calculated to assess the sensitivity of the photodetector D*
against
noise
A(f ) Ri 2 NEP inoise
in
terms
of
noise
equivalent
power
(NEP):
A(f ) , where A is the detector area, f is the noise bandwidth,
inoise is the noise current and Ri is the above mentioned photo-responsivity. Measured responsivity and dark current were used to estimate the value of specific detectivity. Among all possible noise sources, 1/f noise current is removed effectively by the lock-in
amplifier, while thermal noise current i jo
4kT (f ) , where R is shunt resistance in the R
detector, is estimated to be two orders of magnitude lower than shot noise from dark current.3 Therefore specific detectivity was calculated taking into account only shot noise from dark current, given that the photocurrent is smaller than the dark current. At 2 V bias D*
and Ri 2 noise
i
=855
A(f )
i ph I in 2qid A
nm,
the
specific
7.20 1010 cm Hz / W
(at
detectivity room
is
temperature,
modulation frequency of 138 Hz, and with noise bandwidth of lock-in amplifier of 0.42 Hz). The specific detectivity of the nanowire photodetector is limited by its dark current.
References 1. https://dynamic.silvaco.com/dynamicweb/jsp/downloads/DownloadManualsAction.do?
req=silen-manuals&nm=atlas, Page 1456. 2.
Kempa, T. J.; Cahoon, J. F.; Kim, S.-K.; Day, R. W.; Bell, D. C.; Park, H.-G.;
Lieber, C. M., Proc. Natl. Acad. Sci. U. S. A 2012, 109, 1409-1412. 3.
Soci, C.; Zhang, A.; Bao, X. Y.; Kim, H.; Lo, Y.; Wang, D. L., J. Nanosci.
Nanotechnol. 2010, 10, 1430-1449. 4.
http://search.newport.com/?q=*&x2=sku&q2=818-BB-45, Biased Photodetector,
400-900nm, GaAs, 12.5 GHz.
