Introduction

Graphene, a two-dimensional (2D) carbon atomic crystal, has attracted substantial interest for electronic applications1 owing to its high intrinsic carrier mobility2, excellent mechanical flexibility3, optical transparency4, and unique ambipolar transport characteristics5. In particular, its ambipolarity suggests that carriers can be tuned continuously between electrons and holes by supplying the required gate biases, enabling a wide variety of applications, including memory6, frequency multipliers7, high-frequency oscillators up to the THz range8, and fast switches9. However, graphene-based transistors have not yet been implemented in real applications because of their low on/off ratio, which arises from graphene’s gapless band structure10. A high on/off ratio has been obtained by introducing a barristor, that is, a gated graphene/Si junction, because the high junction resistance can provide a sufficiently low off-state current11. However, these barristors usually do not have the typical advantages of graphene, such as mechanical flexibility, optical transparency, and ambipolar transport properties, because Si and a low-k dielectric are used. Recently reported barristor structures, such as graphene/MoS2 and graphene/WS2, have shown very low mobility (40–60 cm2/V s) compared to that of graphene12.

Graphene photodetectors are being developed13. The high mobility of graphene enables high-speed extraction of the photogenerated carriers. However, graphene photodetectors, which typically use the local potential gradient near the graphene–metal junctions, have shown low external quantum efficiency (EQE) and responsivity owing to poor absorption and low built-in potential14. To overcome this problem, graphene photodetectors require extensive junctions rather than local junctions, as well as high mobility and a high built-in potential. Unlike graphene-based lateral photodetectors, which have a rather small photosensing active area near the graphene–metal contact, the vertical barristor device generates a broad region of photocurrent throughout the vertical stack area15,16,17. In addition, barristor devices based on bonding with a high-absorption material may exhibit high absorption resulting from the high built-in potential. However, the 2D barristor devices have revealed limited optoelectronic performances due to their low carrier mobilities and poor gate tuning ratios12,14.

Here, we propose a field-effect device with a graphene/MoSe2 channel layer and a high-k ion-gel gate dielectric. The device shows a high on/off ratio (3.3 × 104) and ambipolar behavior that is controlled by an applied gate voltage. Modulation of the Fermi level (EF) of graphene by applying a gate voltage (VSG) is confirmed by the change in the Schottky barrier (SB) height (ΦB) at the graphene/MoSe2 junction. These field effects, including ambipolar behavior, are locally investigated using scanning photocurrent microscopy (SPCM). It is further demonstrated that an external electric field can be used to modulate the amplitude or even completely reverse the polarity of the photocurrent in the vertical junction of the graphene/MoSe2 barristor device. The strong gating effect of the device results in a higher EQE (66.3%), responsivity (285.0 mA/W), and gate tuning ratio (1.50 μA/V) compared to those of pristine devices. Therefore, our graphene/MoSe2 barristor with an ion-gel gate dielectric is a suitable candidate for use in ambipolar transistors with a high on/off ratio and gate-tunable broad-area photodetectors with a high EQE and responsivity.

Materials and methods

Single-layer graphene and few-layer MoSe2 were fabricated by mechanical exfoliation on a SiO2 (300 nm)/Si substrate. The SiO2/Si substrate was cleaned with hot piranha solution (H2SO4/H2O2 = 4:1) to remove organic matter from its surface18. The electrodes were fabricated using electron beam lithography (Tescan Mira 2 and Raith Elphy Quantum Plus) and electron beam evaporation (Daeki-Hi-tech DKEB-02-04). Poly(methyl methacrylate) (PMMA) C4 solution was spin-coated on the layers at 4500 rpm, followed by baking at 180 °C for 2 min. Electron beam lithography at a dose of ~280 μC/cm2 was used to define patterns on the spin-coated PMMA layer. Then, the Au (50 nm) source, drain, and gate electrodes were deposited on the graphene, MoSe2, and SiO2/Si substrate, respectively, by electron beam evaporation at a deposition rate of 0.4 Å/s and a pressure of 10−6 Torr. Graphene with two source electrodes was transferred to the sample on which MoSe2, two drain electrodes, and one gate electrode were deposited. The graphene/MoSe2 junction was formed between the source and drain electrodes. A polydimethylsiloxane well was located between the side gate and the graphene/MoSe2 channel and filled with an ion-gel dielectric for the SPCM measurements19,20.

Electrodes and markers were patterned on exfoliated graphene and MoSe2 layers over SiO2/Si substrates using electron beam lithography and electron beam evaporation systems. An ion gel was fabricated by gelation of a triblock copolymer in an ionic liquid21. An ultraviolet (UV) cross-linkable polyelectrolyte ion-gel dielectric layer was deposited on the graphene/MoSe2 heterostructure as a gate dielectric. The ionic liquid 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide, the monomer poly(ethyleneglycol) diacrylate (Mw = 575 g/mol), and the UV cross-linking initiator 2-hydroxy-2-methylpropiophenone were mixed at a weight-ratio of 88:8:4, and the mixed solution was dropped and spread on the graphene/MoSe2 structure. The dropped solution was solidified by UV exposure (365 nm, 100 mW/cm2) for 10 s. Finally, a side-gate graphene/MoSe2 device with channel dimensions of 10 μm × 30 μm was obtained.

Results and discussion

Figure 1a shows a schematic diagram of the graphene/MoSe2 barristor structure. High-quality single-layer graphene and few-layer MoSe2 samples were prepared by mechanical exfoliation on a SiO2 (300 nm thick)/Si substrate. The Au (50 nm) source, drain, and gate electrodes were deposited on the graphene, MoSe2, and SiO2/Si substrate, respectively, by electron beam evaporation. The device fabrication is described in detail in the “Materials and methods” section and Supplementary Information (Fig. S1).

Fig. 1: Graphene/MoSe2 barristor characterization.
figure 1

a Schematic diagram and b optical microscope image of the graphene/MoSe2 barristor device. c Schematic of the SPCM measurement setup equipped with a transport measurement system. d SPCM image of the graphene/MoSe2 barristor device obtained at VSG = −1.0 V.

Full size image

Figure 1b shows an optical microscope image of a graphene/MoSe2 barristor device. The blue and green areas surrounded by white dashed lines are graphene and MoSe2, respectively. Two source and drain electrodes were used to check the electrical characteristics of graphene and MoSe2, respectively. The electrical characteristics of the graphene/MoSe2 barristor device were checked between one source and one drain electrode. The SPCM setup, which includes a transport measurement system, is illustrated in Fig. 1c and the Supplementary Information (Fig. S2). The graphene/MoSe2 barristor device was illuminated by a diffraction-limited laser (spot diameter, ~500 nm; wavelength, 532 nm) while the device conductance was recorded as a function of the laser spot position. The small spot size (Supplementary Information, Fig. S3) enabled us to record the photoinduced electronic signal originating from light illumination on a specific part of the graphene/MoSe2 barristor. Because the photocurrent is influenced by the potential profile, an SPCM image can provide information on the local potential profile. We simultaneously obtained a reflected light image to determine the position of the laser spot. Figure 1d shows a photocurrent image of the graphene/MoSe2 barristor obtained at VSG = −1.0 V. The yellow areas denote the source and drain electrodes. The color scale refers to the photocurrent measured between the source and drain electrodes at zero bias. The red area near the graphene/MoSe2 junction corresponds to photocurrent flowing to the drain. The blue area near the MoSe2/drain electrode junction corresponds to photocurrent flowing to the source. The generated photocurrent signals are attributed to band bending and the resultant local electric field at the graphene/MoSe2 and MoSe2/drain electrode junctions. The opposite photocurrent directions indicate the opposite directions of the local electric fields at the graphene/MoSe2 and MoSe2/drain electrode junctions.

Figure 2a shows the source–drain current amplitude (|ISD|) vs. the source–drain voltage (Vbias) measured as VSG increases from −1.25 to 0.5 V (right panel) and from 0.5 to 1.25 V (left panel) in 0.25 V steps. The orange and blue arrows indicate directions of changes in |ISD| with increasing VSG in the p-type and n-type regions, respectively. The right panel shows typical p-type behavior, in which the current decreases as VSG increases from −1.25 to 0.5 V. By contrast, the left panel shows typical n-type behavior, in which the current increases as VSG increases from 0.5 to 1.25 V. We can confirm the ambipolar behavior (p-type and n-type characteristics) for −1.25 V < VSG < 1.25 V. Note that we can control the type of graphene/MoSe2 barristor device by changing the applied VSG, as in an ambipolar graphene channel device.

Fig. 2: Resistive switching behavior depending on the top electrode material.
figure 2

a |ISD|–Vbias curves measured as VSG increases from −1.25 to 0.5 V (right panel) and from 0.5 to −1.25 V (left panel) in 0.25 V steps. The orange and blue arrows indicate changes in the direction of |ISD| with increasing VSG in p-type and n-type regions, respectively. b Transfer characteristics of the graphene/MoSe2 barristor, which are measured as a function of VSG at a fixed Vbias of 10 mV. Schematic band diagrams of the graphene/MoSe2 barristor at c VSG < 0, d VSG > 0, and e VSG ≫ 0.

Full size image

Figure 2b shows the transfer characteristics of the graphene/MoSe2 barristor device, which were measured as a function of VSG at a fixed Vbias of 10 mV. These characteristics are different from those of graphene and MoSe2 devices (Supplementary Information, Figs. S4 and S5, respectively). Our device shows a high on/off ratio of 3.3 × 104, which substantially exceeds those reported for pure graphene transistors (2–20). The barristor exhibits distinct ambipolar behavior and remarkably high carrier mobility values, specifically, an electron mobility of 247 cm2/V s and a hole mobility of 182 cm2/V s at room temperature. These values are determined using the following expression for the field-effect mobility: μ = (1/CSG) × (dσ/dVSG), where CSG is the side-gate dielectric capacitance (Supplementary Information, Fig. S6). The conductivity is defined as σ = ISD/Vbias × L/W, where L is the length and W is the width of the barristor channel. Because this device does not have well-defined channel dimension, the mobility was calculated from the minimum value of L and the maximum value of W for identifying the lower bound of the mobility. The obtained mobility values have similar orders of magnitude to those reported for other barristor devices (40–60 cm2/V s), such as graphene/MoS2 and graphene/WS212, and for MoSe2 devices (50–160 cm2/V s)22,23,24,25.

The observed ambipolar behavior controlled by VSG can be explained by using the schematic band diagrams of the graphene/MoSe2 barristor at VSG < 0 V, VSG > 0 V, and VSG ≫ 0 V (Fig. 2c–e, respectively). ΦB1 and ΦB2 indicate the barrier heights for electrons and holes, respectively, at the graphene/MoSe2 interface. Because of the observed p-type behavior at VSG < 0 V, we can assume that ΦB1 > ΦB2 at VSG < 0 V (Fig. 2c). If the EF of graphene is near the valence band of MoSe2, holes become the majority carriers, so a p-type SB forms at the graphene/MoSe2 junction3. As the gate voltage increases, ΦB2 also increases, and hole transport becomes more difficult. Conversely, as the gate voltage decreases, ΦB2 also decreases, and hole transport becomes easier. The EF of graphene can be increased by the applied positive VSG so that ΦB1 = ΦB2 and even ΦB1 < ΦB2, as shown in Fig. 2d, e, respectively. When ΦB1 < ΦB2 at VSG ≫ 0 V, electrons become the majority carriers, and the graphene/MoSe2 barristor exhibits n-type behavior, as shown in Fig. 2e. As the gate voltage increases, ΦB1 decreases, and electron transport becomes easier. Conversely, as the gate voltage decreases, ΦB1 increases, and electron transport becomes more difficult. The substantial change in the EF of graphene is attributed to the small area of the Fermi surface and the introduction of the ion-gel gate dielectric. In addition, because MoSe2 has a small band gap and similarly high-mobility values (50–160 cm2/V s) for electrons and holes, the relative sizes of ΦB1 and ΦB2 can be controlled by modulating the EF of graphene.

To investigate the local field effect, we performed SPCM measurements of the graphene/MoSe2 barristor shown in Fig. 3a at various gate voltages. Figure 3b–d show photocurrent images of the barristor at various gate voltages between −1.0 and 1.0 V. Because of the side-gate structure and transparent ion-gel dielectric, the illumination light reaches the barristor device without substantial loss. The blue dashed line, red dashed line, and yellow areas denote graphene, MoSe2, and the electrodes, respectively. Unlike a graphene-based lateral photodetector, which has a rather small photosensing active area near the graphene–metal contact, our vertical barristor device clearly shows a broad area of photocurrent generation throughout the junction of the vertical graphene/MoSe2 stack. When a VSG of −1.0 V is applied from the side gate (Fig. 3b), the SPCM images show a red area corresponding to photocurrent flowing to the drain electrode at the graphene/MoSe2 junctions and a blue area corresponding to photocurrent flowing to the source electrode at the MoSe2/drain electrode junction. Under a VSG of 0.25 V (Fig. 3c), the photocurrent contrast signal at the graphene/MoSe2 and MoSe2/drain electrode junctions disappears. As VSG increases to 1.0 V, the photocurrent contrast at the MoSe2/drain electrode junction changes to red, whereas blue and red photocurrents are presented on the graphene/MoSe2 heterojunction (Fig. 3d). It seems that the blue and red photocurrents on the graphene/MoSe2 heterojunction may be caused by the spatial inhomogeneity of the graphene/MoSe2 device. Because we used the side gate configuration and a liquid ion gel, the field effect on the graphene/MoSe2 junction may depend on the location of the 2D planar device. The changes in photocurrent direction at both junctions demonstrate manipulation of the local potential profile and inversion of the band bending at each junction, which are obtained by varying VSG in this vertically stacked device. Importantly, because of the finite density of states in graphene and the weak electrostatic screening effect, the graphene/MoSe2 SB height can be effectively modulated by applying an external field through the side-gate electrode26,27,28.

Fig. 3: Field-effect change in the local photocurrent of the graphene/MoSe2 barristor device.
figure 3

a Reflection image and b–d SPCM images of the graphene/MoSe2 barristor device obtained at VSG values of −1.0 to 1.0 V. The blue dashed line, red dashed line, and yellow area denote graphene, MoSe2, and the electrodes, respectively. Schematic band diagrams of the graphene/MoSe2 device at e VSG < VF, f VSG = VF, and g VSG > VF, which correspond to ΦB1 > ΦB2, ΦB1 = ΦB2, and ΦB1 < ΦB2.

Full size image

Figure 3e–g show schematic band diagrams of the graphene/MoSe2 and MoSe2/drain electrode junctions at various VSG values. Modulation of the EF of graphene and MoSe2 can directly affect the band bending and photocurrent characteristics at the junctions. Because of the quantum capacitance and partial electrostatic transparency of graphene, the applied VSG not only modulates the EF of graphene but also penetrates graphene to accumulate/invert space charges within MoSe229. In graphene/MoSe2 barristor devices, the change in EF of graphene is very important. Therefore, we calculated ΔEF using \({\Delta}E_{\mathrm{F}} = \hbar \nu _{\mathrm{F}}\sqrt {\pi n}\). Here, \(\hbar\) is the Dirac constant, \(\nu _{\mathrm{F}}\) is the Fermi velocity of graphene (1.0 × 106 m/s), and n is the carrier density11,30. The hole carrier density at VSG = −1.50 V and electron carrier density at VSG = 1.45 V of our graphene/MoSe2 barristor device are determined to be 1.03 × 1013 and 4.88 × 1012 cm−2, respectively. The calculated ΔEF is −0.37 eV at VSG = −1.50 V and 0.26 eV at VSG = 1.45 V. The amount of change in EF of 0.63 eV at the graphene/MoSe2 barristor device obtained when VSG varies between −1.50 and 1.45 V is larger than that of other barristor devices11,30. We define VF as the applied VSG at which ΦB1 = ΦB2, which results in a flat band in MoSe2. When the EF value of graphene is reduced at VSG < VF, holes become the majority carriers, resulting in a p-type SB with downward band bending at the graphene/MoSe2 junction. Then, the photoexcited electrons near the graphene/MoSe2 junction drift toward the graphene to produce a positive photocurrent to the drain electrode (Fig. 3e). Similarly, p-type band bending causes photoexcited electrons near the MoSe2/drain electrode to drift toward the drain electrode, resulting in a negative photocurrent to the source electrode. When the EF values of graphene and MoSe2 are increased at VSG = VF, the band bending of MoSe2 at the interfaces becomes negligible, resulting in easy recombination of photoexcited electrons and holes before the photocurrent forms (Fig. 3f). When the EF value of graphene is higher at VSG > VF, electrons become the majority carriers, resulting in an n-type SB with upward band bending at the graphene/MoSe2 junction. Then, the photoexcited holes near the graphene/MoSe2 junction drift toward the graphene to produce a negative photocurrent to the source electrode (Fig. 3g). Similarly, n-type band bending causes the photoexcited holes near the MoSe2/drain electrode to drift toward the drain electrode, resulting in a positive photocurrent to the drain electrode. Various graphene/MoSe2 devices showed similar photocurrent images obtained at VSG = 0 V (Supplementary Information, Fig. S7).

Figure 4a shows the gate dependence of the local photocurrent measured along the black line in the inset SPCM image in Fig. 4a as VSG is swept from −1.0 to 1.0 V in 0.01 V steps. We measured the photocurrent by applying very low laser power (0.407 µW/cm2) without a bias voltage. At VSG < VF, positive (red contrast) and negative (blue contrast) photocurrents are generated at the graphene/MoSe2 (1) and MoSe2/drain electrode (2) junctions, respectively. This photocurrent polarity is a result of p-type band bending, in which the EF of graphene becomes closer to the valence band of MoSe2, as shown in the band diagram in Fig. 3e. This result implies that the photogenerated electrons at the MoSe2/drain electrode and graphene/MoSe2 junctions are collected by the drain and source electrodes, respectively, as reported in previous studies of semiconducting devices with p-type bending31,32,33. At each junction point, the photocurrent polarity is inverted at VSG ~ 0.25 V. In addition, the polarity changes continuously from positive to negative (negative to positive) near the graphene/MoSe2 junction (MoSe2/drain electrode) as VSG is swept from −1.0 to 1.0 V. At VSG > VF, negative and positive photocurrents are generated at the graphene/MoSe2 (1) and MoSe2/drain electrode (2) junctions, respectively. This photocurrent polarity can be interpreted as n-type band bending, in which the EF of graphene becomes closer to the conduction band of MoSe2, as shown in the band diagram in Fig. 3g. This result implies that the photogenerated holes at the MoSe2/drain electrode and graphene/MoSe2 junctions are collected by the drain and source electrodes, respectively, as reported in previous studies of semiconducting devices with n-type bending34. The graphene/MoSe2 barristor device shows ambipolar behavior with p-type and n-type characteristics at VSG < VF and VSG > VF, respectively. Because the intensity of the photocurrent was roughly proportional to the local electric field or the slope of the electrostatic potential, we qualitatively examined the electrostatic potential along the device by integrating the photocurrent line profile at VSG values ranging from −1.0 to 1.0 V in 0.25 V steps.

Fig. 4: Field-effect modulation of photocurrent at junctions in graphene/MoSe2.
figure 4

a Gate voltage dependence of photocurrent measured along the black line of the SPCM image (inset). b Electrostatic potential profile obtained by numerical integration of the experimental accumulated photocurrent amplitudes. The inset shows the photocurrent along the graphene/MoSe2 barristor device at VSG values ranging from −1.0 to 1.0 V in 0.25 V steps. Photocurrent vs. VSG data obtained at the c graphene/MoSe2 and d MoSe2/drain electrode junctions.

Full size image

Figure 4b presents the resulting electrostatic potential profile measured along the black line in the inset SPCM image in Fig. 4a. The measured photocurrent signal [Iph(x)] (Fig. 4b, inset) is proportional to the potential gradient as follows34:

$$I_{{\mathrm {ph}}}\left( x \right) \propto - \frac{{{\mathrm {d}}\varphi (x)}}{{{\mathrm {d}}x}}$$

where x denotes the position along the device channel, and φ(x) is the electrostatic potential. The potential changes dramatically near the graphene/MoSe2 (1) and MoSe2/drain electrode (2) contact regions, indicating SB formation at the graphene/MoSe2 junction (1) and MoSe2/drain electrode (2). Note that the change in potential at the graphene/MoSe2 junction with changing VSG is 13-fold larger than that at the MoSe2/drain electrode. This finding implies that the graphene/MoSe2 junction can become a potential element for tunable electronic or optoelectronic devices.

Figure 4c, d show photocurrent vs. VSG data obtained at the graphene/MoSe2 and MoSe2/drain electrode junctions, respectively. At the graphene/MoSe2 junction, the positive photocurrent decreases as VSG increases from −1.0 to 0.4 V, and the photocurrent becomes negative at VSG = 0.4 V. When VSG is −0.98 V, we obtain a very high gate tuning ratio of ~76.6 nA/V, which is defined as the ratio of the change in photocurrent amplitude to the change in VSG. At the MoSe2/drain electrode junction, negative and positive photocurrents are generated at −1.0 V < VSG < 0.15 V and 0.15 V < VSG < 1.0 V, respectively. When VSG is −0.71 V, we obtain a high gate tuning ratio of ~18.2 nA/V. We could obtain a higher gate tuning ratio of 1.50 μA/V at a graphene/MoSe2 junction in another barristor device (Supplementary Information, Fig. S8). These gate tuning ratios, which were measured at a very low laser power (0.407 µW/cm2), are much higher than those of other graphene-based devices (1–5 nA/V)35. In addition, the very high gating effect of our graphene/MoSe2 barristor structure with a high-k ion-gel gate dielectric dramatically changes the polarity of the photocurrent. The high gate-tunability of the photocurrent is significant for the facile modulation of sensitivity for optical sensors. In dark spaces (such as a tunnel and cinema) where the physical environment is limited, photosensitivity should be amplified so that sufficient photocurrent can be secured. In bright spaces (such as outdoors on a sunny day), photosensitivity can be reduced and the value of photocurrent flowing through the device can be constantly adjusted. Therefore, the development of high-gate tuning elements in optical devices is very necessary35,36.

Using the photocurrent response and input laser power, we can determine the EQE and photoresponsivity. The photoresponsivity (RI) is defined as the ratio of the electrical current response to the incident optical power (RI = Iph/Popt)16,37. The EQE (η) is defined as the number of carriers produced per photon and is expressed as

$$\eta = \frac{{I_{{\mathrm {ph}}}}}{{q{\it{{\Phi}}}}} = \frac{{I_{{\mathrm {ph}}}}}{q}\left( {\frac{{hv}}{{P_{{\mathrm {opt}}}}}} \right)$$

where Iph is the photocurrent, Φ is the photon flux (=Popt/hv), h is the Planck constant, v is the frequency of the light, q is the electron charge, and Popt is the optical power. At the graphene/MoSe2 junction, the EQE increases from 3.3% (corresponding to a photoresponsivity of 14.2 mA/W) at VSG = 0 V to 24.0% (corresponding to a photoresponsivity of 103.2 mA/W) at VSG = −0.98 V under excitation at a laser power of 0.407 µW/cm2. At the MoSe2/drain electrode junction, the EQE increased from 0.5% (corresponding to a photoresponsivity of 2.2 mA/W) at VSG = 0 V to 3.18% (corresponding to a photoresponsivity of 16.8 mA/W) at VSG = −0.71 V. The EQE at the graphene/MoSe2 junction is ~6.6-fold larger than that at the MoSe2/drain electrode junction at VSG = 0 V. The EQE (24%) observed at the vertical graphene/MoSe2 junction at VSG = −0.98 V is two orders of magnitude higher than those of lateral metal–graphene–metal photodetectors (EQE ≈ 0.1–0.2%)14,38,39. We could obtain a higher EQE of 66.3% and photoresponsivity of 285.0 mA/W at a graphene/MoSe2 junction in another barristor device (Supplementary Information, Fig. S8). This high EQE is due primarily to more efficient photon absorption in the broad area of the vertical barristor device and more efficient charge separation resulting from a much larger band offset and a much higher mobility16.

Conclusion

In summary, we fabricated a graphene/MoSe2 barristor with a high-k ion-gel gate dielectric. The graphene/MoSe2 device showed a high on/off ratio (3.3 × 104) and ambipolar behavior that was controlled by an applied gate voltage. SPCM measurements revealed that the graphene/MoSe2 barristor had very high EQE (66.3%) and photoresponsivity (285.0 mA/W) values, making it suitable for highly efficient photocurrent generation and photodetection. We further demonstrated that an electric field applied to the gate electrode could substantially modulate the amplitude and polarity of the photocurrent at the graphene/MoSe2 junction, resulting in a high gate tuning ratio (1.50 μA/V). Therefore, the graphene/MoSe2 barristor with a high-k ion-gel gate dielectric is a suitable candidate for use in ambipolar transistors (with a high on/off ratio) and gate-tunable photodetectors (with a high EQE and responsivity).