# Fileset

[Manuscript_250616.pdf](https://mdr.nims.go.jp/filesets/41d1460b-1f67-4959-87b6-926464db6678/download)

## Creator

[Takuya Kawazu](https://orcid.org/0000-0001-8081-4167)

## Rights

This is the Accepted Manuscript version of an article accepted for publication in Japanese Journal of Applied Physics.  IOP Publishing Ltd is not responsible for any errors or omissions in this version of the manuscript or any version derived from it.  The Version of Record is available online at https://doi.org/10.35848/1347-4065/ade64e.[Creative Commons BY-NC-ND Attribution-NonCommercial-NoDerivs 4.0 International](https://creativecommons.org/licenses/by-nc-nd/4.0/)

## Other metadata

[Schottky barrier lowering in n-AlGaAs/GaAs/AlGaAs double heterojunctions by local light illumination](https://mdr.nims.go.jp/datasets/d36d551d-0485-4e63-843a-5878c6d40335)

## Fulltext

Title of Paper Goes Here:  Template for JJAP Regular Papers (Feb. 2017) 1 Schottky barrier lowering in n-AlGaAs/GaAs/AlGaAs double-1 heterojunction by local light illumination 2 Takuya Kawazu 3 National Institute for Materials Science, Tsukuba, Ibaraki 305-0047, Japan 4 *E-mail: KAWAZU.Takuya@nims.go.jp 5  6 Abstract: We investigated the wavelength dependence of the Schottky photocurrent (ISG) in 7 an n-AlGaAs/GaAs/AlGaAs double-heterojunction under local light illumination. ISG was 8 measured with the illumination from two light sources (Light A and B). Light A illuminated 9 the entire Schottky gate region at 60 μW/mm² with varying the wavelength (940, 975, and 10 1064 nm), while Light B (670 nm laser) locally illuminated the ungated region at power 11 levels (PB) ranging from 0 to 2.9 μW. At low PB (≤ 86 nA) and high PB (≥ 0.14 μW), ISG 12 increases and decreases respectively with the increasing photon energy of Light A. By 13 analyzing the experimental data using a hot carrier model, this behavior is attributed to the 14 Schottky barrier lowering induced by Light B. The reduction in the effective Schottky barrier 15 height increases with increasing PB, reaching about 0.5 – 0.6 eV. 16  17  18  19  20  21  22  23  24  25  26  27  28  29   Template for JJAP Regular Papers (Feb. 2017) 2 1. Introduction 1 Owing to the importance in photodetector applications, the infrared (IR) photoresponses of 2 Schottky barrier diodes have attracted strong interest for a few decades.1-23) A Schottky 3 potential barrier develops in the interface region of a metal film on a semiconductor.24) When 4 incident photons are absorbed through free-carrier transitions in the metal film,25) photo-5 carriers are generated and transfer to the semiconductor, resulting in the generation of a 6 photocurrent. The detectable photon energy is determined by the Schottky barrier height, not 7 by the semiconductor bandgap, in which the IR photo-sensing is potentially possible in the 8 energy rage of 0.2 – 0.6 eV by choosing an appropriate combination of a metal and a 9 semiconductor.5,6,9,25) Various efforts have been made to improve the photoresponses of 10 Schottky-barrier diodes, such as back illuminations,6,9,10) very thin metal gates,5),6),9) two 11 Schottky-barriers,17) gate edge illuminations,19,21) surface plasmon,14)-16) and optical 12 cavities.6,11,12) 13 In our previous studies,22,23) we investigated the Schottky photocurrent in n-AlGaAs/GaAs 14 and n-AlGaAs/GaAs/AlGaAs heterojunctions with metal gates. We observed that the 15 photocurrent induced by illuminating the Schottky gate region is significantly enhanced 16 when the ungated region is also illuminated. This enhancement is attributed to holes excited 17 by the illumination of the ungated region, which diffuse in the GaAs buffer layer and 18 subsequently reach the Schottky gate region. Although the model proposing that holes 19 contribute to the enhancement of the Schottky photocurrent has successfully explained the 20 experimental results, the precise mechanism through which these holes influence the 21 photocurrent remains unclear. 22 In this work, we investigate the Schottky photocurrent ISG in an n-AlGaAs/GaAs/AlGaAs 23 double-heterojunction under two light sources (Light A and B in Fig. 1). Light A was used 24 to illuminate the entire Schottky gate region with varying its wavelength, while Light B was 25 employed to selectively illuminate the ungated region with modulating the power levels PB. 26 By analyzing ISG using a hot carrier model, we show that the illumination of the Light B has 27 an effect equivalent to reducing the effective Schottky-barrier height. As PB increases, the 28 reduction δΦ in the effective Schottky barrier height increases, with δΦ reaching 29 approximately 0.5 – 0.6 eV. 30  31   Template for JJAP Regular Papers (Feb. 2017) 3 2. Experimental methods 1 By using molecular beam epitaxy, we prepared the heterojunction sample on a semi-2 insulating GaAs(100) substrate. We grew a 500-nm-thick GaAs buffer layer, a superlattice 3 layer (25 periods of 10 nm Al0.27Ga0.73As/2 nm GaAs), a 15-nm-thick GaAs quantum well 4 (QW) channel, a 10-nm-thick nondoped AlGaAs spacer, an 80-nm-thick Si-doped AlGaAs 5 layer, and a 10-nm-thick GaAs capping layer. Throughout the growth, the substrate 6 temperature was set at 580 ℃. Schematic drawings of the sample cross-section view and the 7 band diagram are illustrated in Figs. 2 and 3. 8 The heterojunction sample was processed into a Hall bar device with a Schottky gate. The 9 Hall bar structure (represented by the dot-patterned region in Fig. 4) was defined using 10 photolithography, followed by wet etching. To form the Schottky metal gate, aluminum was 11 deposited via vacuum evaporation to a thickness of approximately 100 nm (gray region in 12 Fig. 4). The Schottky gate region of the heterojunction, indicated by the broken-line 13 boundary, has a width (W) of 50 μm and a length (L) of 550 μm. The ohmic contacts between 14 the electrode pads and the QW channel (hatched circles in Fig. 4) were made by alloying 15 InSn at 400 ℃ for 2 minutes in an argon atmosphere. The sample was bonded using silver 16 paste in a ceramic package with a gold-plated surface.  17 Photocurrent measurements were performed at room temperature using the experimental 18 setup with the schematic of device connections depicted in Figs. 1 and 4. The Schottky 19 photocurrent (ISG in Figs. 1 and 4) from the source contact (S) to the Schottky metal gate (G) 20 was monitored using a source measure unit (Keithley 2635), capable of measuring DC 21 currents in the range of 0.1 fA to 1.53 A.  22 For the illuminate of the Schottky gate region, we used 940-, 975-, and 1064-nm laser 23 beams (Light A), and the entire heterojunction structure was irradiated by one of them with 24 an excited power density of about 60 µW/mm2. Since the photon energy of Light A is below 25 the bandgap energy of GaAs and AlGaAs, interband absorption does not occur, resulting in 26 GaAs and AlGaAs crystals being nearly transparent to the incident light. Consequently, Light 27 A has minimal influence in the ungated region, even though it illuminates the entire structure. 28 Light A reaches the Schottky metal-semiconductor interface either through transmission via 29 the gate metal or by reflection at the back surface, subsequently exciting electrons (Fig. 2). 30 The excited electrons then transfer to the QW, inducing the Schottky photocurrent ISG. 31   Template for JJAP Regular Papers (Feb. 2017) 4 For the illumination of the ungated region, we used a 670-nm-laser beam (Light B). Light 1 B was focused by a microscope objective lens, allowing the observation of the sample 2 surface and precise monitoring of the irradiation position. The laser spot, with a diameter of 3 approximately 50 μm, was positioned at the location indicated by 'L' in Fig. 4, about 1.4 mm 4 from the Schottky gate region. The photon energy of Light B, approximately 1.85 eV, 5 exceeds the bandgap energies of the n-Al₀.₂₇Ga₀.₇₃As layer (~1.78 eV), the GaAs buffer layer, 6 and the superlattice layer, resulting in absorption processes within these layers. However, 7 the absorption in the superlattice layer is expected to be the most significant process for 8 enhancing the Schottky photocurrent.23) When Light B is absorbed by the superlattice layer, 9 it generates electrons and holes (Fig. 3). The electrons migrate towards the QW and are 10 rapidly and uniformly redistributed within the QW due to the high conductivity of the two-11 dimensional electron gas (2DEG) channel. In contrast, the generated holes move toward the 12 GaAs buffer layer and accumulate there. The accumulated holes diffuse laterally and then 13 reach the Schottky gate region. The positive charge of the holes induces electric fields, 14 thereby enhancing the Schottky photocurrent ISG under the illumination of Light A. 15  16 3. Results and Discussion 17 3.1 Experimental results 18 Initially, the Schottky photocurrent ISG from the source contact (S) to the metal gate (G) was 19 measured under the conditions where the wavelength λA of Light A was 940 nm, and the 20 power PB of Light B was 0.57 μW. Figure 5 shows ISG for the on/off action of Light A and 21 B. 'A', 'B', and 'A+B' in the horizontal axis of Fig. 5 indicate the illuminations of Light A, 22 Light B, and both, respectively. 'off' represents that all the lights are off. When Light A 23 irradiates the sample ('A' in Fig. 5), the Schottky photocurrent flows from S to G, where ISG 24 is positive. Upon irradiation with Light A, photoelectrons are excited at the Schottky metal-25 semiconductor interface and subsequently transfer to the QW channel (Fig. 2), resulting in 26 the generation of the photocurrent from S to G. For the illumination of Light B only ('B' in 27 Fig. 5), the Schottky photocurrent is not generated (ISG ~ 0). Light B excites electrons and 28 holes (Fig. 3), but the electron-hole excitation itself does not induce the Schottky 29 photocurrent. When both Light A and B simultaneously illuminate the heterojunction ('A+B' 30 in Fig. 5), ISG increases by a factor of 4 compared to Light A only. Note that the IR 31   Template for JJAP Regular Papers (Feb. 2017) 5 photoresponse in the Schottky gate region is effectively enhanced by the illumination of 1 Light B, where the illumination position is far away (~ 1.4 mm) from the gate region. Due 2 to the separation of holes from the electrons that recombine with them by a thick superlattice, 3 the photo-generated and subsequently accumulated holes diffuse laterally in the GaAs buffer 4 layer over a significant distance, allowing them to reach the Schottky gate region (Fig. 3). 5 ISG under Light A illumination is enhanced by the electric fields induced by the positive hole 6 charge in the Schottky gate region.  7 A similar measurement was conducted with Laser B positioned at the GaAs substrate 8 location marked as L’ in Fig. 4. As indicated by the dotted line in Fig. 5, no increase in the 9 photocurrent for Light A was observed when exposed to Light B. This observation confirms 10 that the photocurrent increase is not due to substrate illumination or stray light, but requires 11 electron-hole excitation of the heterojunction by Light B. 12 Figure 6 shows the Schottky photocurrent ISG from S to G when the wavelength of Light 13 A is 940, 975, and 1064 nm. Here, the power PB of Light B is 20 nW. ISG are approximately 14 1.6, 3.1, and 3.6 nA for Light A illumination of 1064-, 975-, and 940-nm wavelengths, 15 respectively ('A' in Fig. 6). When both Light A and B simultaneously illuminate the 16 heterojunction ('A+B' in Fig. 6), ISG increase and are about 4.7, 5.8, and 6.2 nA for the 1064-, 17 975-, and 940-nm laser illuminations. 18 Figure 7 shows the Schottky photocurrent ISG when the power PB of Light B is about 0.86 19 μW. The other experimental conditions are the same as those in Fig. 6. When Light A and B 20 simultaneously illuminate the heterojunction ('A+B' in Fig. 7), ISG are about 19.2, 16.4, and 21 14.5 nA for the 1064-, 975-, and 940-nm laser illuminations. The values of ISG are 22 significantly higher than when PB = 20 nW ('A+B' in Fig. 6), which is due to the excitation 23 of a larger number of holes. Note that ISG increases with increasing λA at PB = 0.86 μW, while 24 ISG shows a downward trend with increasing λA at PB = 20 nW. 25 Next, we examined the Schottky photocurrent ISG by altering the wavelength λA of Light 26 A and the power PB of light B, with both Light A and B concurrently illuminating the 27 heterojunction. Figure 8 shows ISG as a function of the photon energy (hνA = hc/λA, where h 28 represents Planck's constant and c denotes the speed of light) of Light A. Here, the 29 experimental conditions are identical to those previously described, with the exception of PB. 30 Note that as hνA increases, ISG increases and decreases at PB ≤ 86 nA and ≥ 0.14 μW. When 31   Template for JJAP Regular Papers (Feb. 2017) 6 λA is 1064 nm (~ 1.17 eV), 975 nm (~ 1.27 eV), and 940 nm (~ 1.32 eV), ISG increase about 1 14.3, 6.5, and 5.1 times respectively with the increase in PB from 0 (red open circles) to 2.9 2 μW(gray inverted triangles); the smaller the photon energy of Light A is, the stronger the 3 enhancement effect of Light B becomes.  4  5 3.2 Schottky barrier lowering 6 According to a hot carrier model,17) the Schottky photocurrents ISG is given by 7 𝐼𝐼𝑆𝑆𝑆𝑆 = 𝑒𝑒𝑒𝑒𝑃𝑃𝐴𝐴2ℎ𝜈𝜈𝐴𝐴�1 −� 𝛷𝛷ℎ𝜈𝜈𝐴𝐴�2                                               (1) 8 where e is the elementary charge, A is the optical absorptance,Φ is the effective Schottky 9 barrier height, and PA is the incident optical power. In our case, PA corresponds to the power 10 of Light A and is about 60 μW/mm2×LW. The solid lines in Fig. 8 represent ISG derived from 11 Eq. (1), using Φ as a fitting parameter to match the experimental data corresponding to each 12 power level PB of Light B. The value of APA is set to 7×10-8 W (A ~ 0.042) and is determined 13 such that Φ (≡ Φ0) is 0.6 eV in the absence of Light B irradiation. The effective Schottky 14 barrier height Φ0 (= 0.6 eV) under dark conditions is an estimate derived from the JSG-VG 15 characteristics observed when a bias voltage VG is applied to the Schottky gate. The optical 16 absorptance A of 0.042 indicates that only 4.2% of the incident light contributes to 17 photocarrier generation. The low efficiency is primarily attributed to the high reflectivity of 18 the Schottky metal gate, where the 100-nm-thick aluminum layer is expected to reflect over 19 90% of the incident light. Improving efficiency can be achieved by reducing the thickness 20 of the metal gate and/or selecting an alternative gate material. 21 As shown by the solid lines in Fig. 8, the dependence of the measured Schottky 22 photocurrent ISG on the photon energy hνA is well described by Eq. (1), with Φ being adjusted 23 for each power level PB of Light B. Notably, the transition of ISG from an increasing to a 24 decreasing trend as a function of hνA can be attributed to variations in Φ. For a fixed 25 excitation power PA of the Light A, the number of incident photons decreases inversely with 26 increasing hνA, as described by the term 𝐴𝐴𝑃𝑃𝐴𝐴 ℎ𝜈𝜈𝐴𝐴⁄  in Eq. (1). Conversely, the fraction of 27 photo-excited electrons that surmount the Schottky barrier increases with increasing hνA, 28 where the fraction is given by the term �1 −�𝛷𝛷 ℎ𝜈𝜈𝐴𝐴⁄ �22�  in Eq. (1). The overall trend 29 whether increasing or decreasing of ISG as a function of hνA is determined by the interplay of 30   Template for JJAP Regular Papers (Feb. 2017) 7 these two factors. Differentiating Eq. (1) with respect to hνA reveals that ISG increases with 1 hνA for hνA < 4Φ, whereas it exhibits a decreasing trend for hνA > 4Φ. 2 Although the hot electron model (Eq. (1)) qualitatively aligns with the experimental data 3 presented in Fig. 8, its quantitative accuracy remains unsatisfactory. Specifically, the 4 measured Schottky photocurrent ISG increases with increasing the photon energy hνA under 5 illumination with Light B at PB = 86 and 57 nW (black and pink open triangles in Fig. 8). In 6 contrast, the hot electron model predicts an inverse trend (black and pink solid lines in Fig. 7 8). This discrepancy primarily arises from an error in estimating the effective Schottky 8 barrier height Φ0 (= 0.6 eV) under dark conditions, which will be discussed in the following 9 section. Furthermore, the deviation between the model predictions and experimental results 10 can be attributed to factors such as measurement uncertainties (estimated error in ISG of 11 approximately 5 – 6%) and simplifications inherent in the hot electron model, as well as the 12 omission of electron photo-excitation in the QW23) and the wavelength dependence of the 13 optical absorptance A. The solid circles in Fig. 9 represent the effective Schottky barrier 14 height Φ, determined via the fitting procedure described above, plotted as a function of the 15 excitation power PB of Light B. As PB increases from 0 to 2.9 μW, Φ decreases from 16 approximately 0.6 to 0.03 eV. It should be noted that the reliability of these values requires 17 careful consideration, as the obtained value of Φ is highly sensitive to the estimation error 18 of Φ0 under dark conditions. A discussion of this issue will be provided in the following 19 section. To mitigate this uncertainty, we conduct our analysis using the effective Schottky 20 barrier reduction δΦ ≡ Φ0−Φ, which is less influenced by the magnitude of Φ0 (see the 21 following section). The solid circles in Fig. 10 represent the effective Schottky barrier 22 reduction δΦ as a function of the excitation power PB of Light B. As PB increases, δΦ 23 increases, and the maximum reduction in the effective Schottky barrier height is about 0.5 – 24 0.6 eV. In a Schottky barrier diode, the barrier energy is governed by the work function of 25 the gate metal, the electron affinity of the semiconductor, and the surface states at the metal-26 semiconductor interface. For many III-V compounds, the Schottky barrier energy is 27 predominantly determined by the surface states. The intrinsic energy Φint of such a Schottky 28 barrier is reduced by the image charge of the electrons emitted in an electric field. The 29 lowering ΔΦ0 in Schottky barrier energy is proportional to the square root of the electric 30 field E0 present at the semiconductor side of the Schottky metal-semiconductor interface, as 31   Template for JJAP Regular Papers (Feb. 2017) 8 expressed by 1 ∆Φ0 = � 𝑒𝑒𝐸𝐸04𝜋𝜋𝜀𝜀𝑠𝑠,                                                        (2) 2 where εs is the permittivity of the semiconductor. For a Schottky diode with a barrier height 3 of 0.6 eV, ΔΦ0 is estimated to be about 0.057 eV. Note that Φ0 defined above corresponds to 4 Φint-ΔΦ0. When the heterojunction is irradiated with Light B, photo-excited holes move 5 toward the substrate and diffuse to the Schottky gate region (Figs. 1 and 3), and their positive 6 charge induces an electric field Eind in addition to E0 in Eq. (2). Such an electric field causes 7 an additional lowering δΦ of the Schottky barrier energy. Eind is anticipated to be 8 proportional to the hole density, which is expected to be proportional to the excitation power 9 PB of Light B, resulting in 10 𝑒𝑒𝐸𝐸𝑖𝑖𝑖𝑖𝑖𝑖4𝜋𝜋𝜀𝜀𝑠𝑠= 𝛼𝛼𝑃𝑃𝐵𝐵,                                                         (3) 11 where α is a constant with a dimension of J·s. From Eqs. (2) and (3), we obtain 12 ∆Φ0 + 𝛿𝛿Φ = �𝑒𝑒(𝐸𝐸0+𝐸𝐸𝑖𝑖𝑖𝑖𝑖𝑖)4𝜋𝜋𝜀𝜀𝑠𝑠= �ΔΦ02 + 𝛼𝛼𝑃𝑃𝐵𝐵.                                 (4) 13 The solid line in Fig. 10 represents Eq. 4 with ΔΦ0 set to 0.057 eV and α as the fitting 14 parameter. At low power PB (<~ 0.06 μW) of Light B, Eq. 4 agrees with the experimental 15 data for α = 5.9×10-13 eV·s, indicating the validity of our explanation. In contrast, at high PB 16 (>~ 0.1 μW), the experimental values of δΦ deviate from those predicted by Eq. 4. This 17 discrepancy may be caused by the increase of electron-hole recombination in the superlattice 18 layer at high PB. When the heterojunction is irradiated with Light B, photo-excited electrons 19 (holes) are driven towards the QW (GaAs buffer) by the built-in electric field (Fig. 3(b)). As 20 PB increases, the accumulation of holes in the GaAs buffer becomes more pronounced. The 21 charge of accumulated holes weakens the built-in electric field. As a result, the 22 recombination rate of holes increases before reaching the GaAs buffer, where Eq. 3 is no 23 longer valid. This issue is important for improving the device performance and is left for 24 future work. 25  26 3.3 Reliability of extracted Schottky barrier height 27 Before presenting the concluding remarks, we discuss the reliability of the extracted 28 effective Schottky barrier height Φ and its reduction δΦ (≡ Φ0−Φ). The absorbivity of 29   Template for JJAP Regular Papers (Feb. 2017) 9 aluminum has been reported to be around 8% for incident light with the wavelength range 1 of 940 – 1064 nm.26,27) Based on the Fresnel formula, the reflectance of a 100-nm-thick 2 aluminum film is estimated to be around 94%. However, the actual optical absorptance is 3 strongly affected by several factors, including the presence of an oxide layer, surface 4 roughness, and contaminants,28) making precise estimation difficult. Furthermore, in our 5 experimental setup, Light A illuminates the entire heterojunction structure, including both 6 the Schottky gate and ungated regions. Due to the high optical transparency of GaAs and 7 AlGaAs crystals to Light A, a portion of the light incident on the ungated region may also 8 reach the Schottky gate region by reflection at the back surface. This additional contribution 9 further complicates the precise estimation of the optical absorptance A. 10 Rather than assuming the optical absorptance A a priori, we first estimated the effective 11 Schottky barrier height Φ0 (= 0.6 eV) under dark conditions based on the JSG-VG 12 characteristics. We then compared Eq. (1) with the experimentally measured dependence of 13 the photocurrent ISG on the photon energy hνA in the absence of Light B (open circles in Fig. 14 8) to determine A. This analysis yielded A ~ 0.042 (4.2%), which was adopted in the previous 15 section. However, in this procedure, imprecise evaluation of Φ₀ introduces errors in the 16 extraction of A, potentially leading to considerable uncertainty in the determination of the 17 Schottky barrier height Φ under illumination. 18 For intimate contacts with a clean interface, the Schottky barrier height Φ0 between GaAs 19 and Al has been reported to be 0.93±0.02 eV.29) This value is frequently reduced due to 20 interfacial contamination, the presence of an intermidiate insulating layer, or deep-level 21 impurities. In studies of n-AlGaAs/GaAs field-effect transistors, Φ₀ is sometimes 22 approximated as 0.7 eV.30) Accordingly, we evaluate the optical absorptance A using Φ₀ = 23 0.7 eV instead of 0.6 eV, yielding A ~ 0.061 (6.1%). The Schottky photocurrent ISG was then 24 derived from Eq. (1), treating Φ as a fitting parameter following the approach in the previous 25 section, with A set to ~ 0.061 instead of ~ 0.042 [Fig. 1(S) in Supplementary Data]. The 26 dependence of ISG on hνA remains well described by Eq. (1) even with A ~ 0.061, suggesting 27 that the error in the value of Φ0 has a minor effect on the overall agreement between Eq. (1) 28 and the experimental data. The discrepancy observed in the previous section between the 29 increasing and decreasing trends of ISG with respect to hνA (PB = 86 and 57 nW in Fig. 8) is 30 slightly improved when the analysis is conducted using A ~ 0.061. 31   Template for JJAP Regular Papers (Feb. 2017) 10 By contrast, the value of the Schottky barrier height Φ obtained through the fitting 1 procedure is significantly affected by the estimation error of Φ0 under dark conditions. We 2 evaluated Φ under the assumption that Φ0 is 0.5, 0.6, and 0.7 eV and found that the obtained 3 Φ strongly depends on the assumed value of Φ0 [Fig. 2(S) in Supplementary Data], implying 4 that the extracted Φ values is not as reliable. This uncertainty can be effectively reduced by 5 considering the Schottky barrier reduction δΦ (≡ Φ0−Φ) [Fig. 3(S) in Supplementary Data], 6 where the values of δΦ exhibit a dependence on PB that is nearly independent of the assumed 7 value of Φ0. The result confirms that δΦ is a robust quantity, even when errors in Φ may be 8 present. 9  10 4. Conclusions 11 The Schottky photocurrent ISG in an n-AlGaAs/GaAs/AlGaAs double-heterojunction was 12 studied for the irradiation of two light sources (Light A and B). Light A, which is a 940 nm, 13 975 nm, or 1064 nm laser, illuminates the entire Schottky gate region at an intensity of 60 14 μW/mm². Light B, a 670 nm laser, is used to locally illuminate the ungated region with a 15 power PB ranging from 0 to 2.9 μW. In the low PB range (≤ 86 nA) and high PB range (≥ 16 0.14 μW), ISG exhibits an increase and decrease respectively with the increasing photon 17 energy of Light A. By analyzing the experimental data using a hot carrier model, this 18 behavior is well explained by the Schottky barrier lowering induced by Light B. As PB 19 increases, the reduction δΦ of the effective Schottky barrier height increases, where the 20 maximum δΦ is about 0.5 – 0.6 eV.  21  22  23  24  25  26  27  28  29  30  31   Template for JJAP Regular Papers (Feb. 2017) 11 References 1 1) R. Williams and R. H. Bube, J. Appl. Phys. 31, 968 (1960). 2 2) W. G. Spitzer, C. R. Crowell, and M. M. Atalla, Phys. Rev. Lett. 8, 57 (1962). 3 3) D. W. Peters, Proc. IEEE 55, 204 (1967). 4 4) F. D. Shepherd, A. C. Yang, and R. W. Taylor, Proc. IEEE 58, 1160 (1970). 5 5) E. Elabd, T. Villani, and W. Kosonocky, IEEE Electron Devices Lett., 3, 89 (1982). 6 6) W. F. Kosonocky, F. W. Shallcross, T. S. Villani, and J. V. Groppe, IEEE Trans. Electron Devices, 32, 7 1564 (1985). 8 7) J. M. Mooney, J. Appl. Phys. 65, 2869 (1989). 9 8) B. Y. Tsaur, C. K. Chen, and J. P. Mattia, IEEE Trans. Electron Devices 11, 162 (1990). 10 9) C. K. Chen, B. Nechay, and B. Y. Tsaur, IEEE Trans. Electron Devices 38, 1094 (1991). 11 10) T. A. Yost, A. Madjar, and P. R. Herezfeld, IEEE Transactions on Microwave Theory and Techniques 12 49, 1900 (2001). 13 11) M. Casalino, L. Sirleto, L. Moretti, and I. Rendina, Semicond. Sci. Technol. 23, 075001 (2008).  14 12) M. Casalino, L. Sirleto, L. Moretti, M. Gioffrè, G. Coppola, and I. I. Rendina, Appl. Phys. lett. 92, 15 251104 (2008). 16 13) S. Zhu, M. B. Yu, G. Q. Lo, and D. L. Kwong, Appl. Phys. Lett. 92, 081103 (2008). 17 14) A. Akbari and P. Berini, Appl. Phys. Lett. 95, 021104 (2009). 18 15) P. Berini, Adv. Opt. Photonics 1, 484 (2009). 19 16) C. Scales, I. Breukelaar, and P. Berini, Opt. Lett. 35, 529 (2010). 20 17) C. Scales and P. Berini, IEEE J. Quantum Electron. 46, 633 (2010). 21 18) T. Kawazu, T. Noda, Y. Sakuma, and H. Sakaki, Appl. Phys. Lett., 106, 022103, (2015). 22 19) T. Kawazu, T. Noda, and Y. Sakuma, Jpn. J. Appl. Phys. 56, 04CG04, (2017). 23 20) T. Kawazu, T. Noda, and Y. Sakuma, Appl. Phys. Lett., 112, 072101, (2018). 24 21) T. Kawazu, T. Noda, and Y. Sakuma, Jpn. J. Appl. Phys. 58, SIIB05, (2019). 25 22) T. Kawazu, Jpn. J. Appl. Phys. 59, 124003, (2020). 26 23) T. Kawazu, AIP Advances 14, 125007, (2024). 27 24) E. Spenke, Electronic Semiconductors (McGraw-Hill, New York, 1958). 28 25) W. P. Dumke, Phys. Rev. 124, 1813 (1961). 29 26) Bass M. Handbook of optics, Vol. 1, 2nd ed. (McGraw-Hill, New York, 1995).  30 27) L. Tunna, W. O’Neill, A. Khan, and C. Sutcliffe, Opt. Lasers Eng. 43, 937 (2005).  31 28) K. Howard, S. Lawson and Y. Zhou, Welding Journal 85 (5), 101 (2006). 32 29) S. M. Sze, Physics of Semiconductor Devices (Wiley, New York, 1981). 33 30) J. H. Davies, The Physics of Low-dimensional Semiconductors (Cambridge University Press, Cambridge, 34 1998) 35  36   Template for JJAP Regular Papers (Feb. 2017) 12 Figure Captions 1 Fig. 1. Schematic drawing of sample cross-sectional view. 2  3 Fig. 2. Schematic drawings of the cross-section view (a) and the band diagram of n-4 AlGaAs/GaAs/AlGaAs double-heterojunction (b) when Light A is irradiated.  5  6 Fig. 3. Schematic drawings of the cross-section view (a) and the band diagram of n-7 AlGaAs/GaAs/AlGaAs double-heterojunction (b) when Light B is irradiated. 8  9 Fig. 4. Schematic drawing of Hall bar structure. 10  11 Fig. 5. Photocurrent ISG for on/off action of Light A and B. 12  13 Fig. 6. Photocurrent ISG for on/off action of Light A and B. 14  15 Fig. 7. Photocurrent ISG for on/off action of Light A and B. 16  17 Fig. 8. Photocurrent ISG as a function of photon energy hνA of Light A. 18  19 Fig. 9. Effective Schottky barrier height as a function of power PB of Light B. 20  21 Fig. 10. Schottky barrier reduction δФ as a function of power PB of Light B. 22  23  24  25  26  27  28  29  30  31   Template for JJAP Regular Papers (Feb. 2017) 13  1 Fig.1. 2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26   Template for JJAP Regular Papers (Feb. 2017) 14  1 Fig.2. 2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25   Template for JJAP Regular Papers (Feb. 2017) 15  1 Fig.3. 2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24   Template for JJAP Regular Papers (Feb. 2017) 16  1 Fig.4. 2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26   Template for JJAP Regular Papers (Feb. 2017) 17  1 Fig.5. 2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24   Template for JJAP Regular Papers (Feb. 2017) 18  1 Fig.6. 2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24   Template for JJAP Regular Papers (Feb. 2017) 19  1 Fig.7. 2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24   Template for JJAP Regular Papers (Feb. 2017) 20  1 Fig.8. 2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22   Template for JJAP Regular Papers (Feb. 2017) 21  1  2  3  4  5  6  7  8 Fig.9. 9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31   Template for JJAP Regular Papers (Feb. 2017) 22  1 Fig.10. 2