Introduction

The ability to distinguish between red and green is essential in daily life. This is not only true for the correct recognition of images and the surrounding in general1,2, but also for the recognition of signals, such as traffic lights, for the steering of autonomous vehicles, as prominent example3,4. All this becomes a real issue for people suffering red-green color blindness5. Compared to visual aids6, visible-light photodetectors (PDs) discriminating between red and green offer a clear solution. Undoubtedly, dual-wavelength PDs based on dual-polarity photoresponse are most elegant and reliable7,8,9,10,11,12,13,14,15,16compared to other PDs which often only exhibit dual-magnitude photoresponse17,18,19,20,21. For dual-polarity PDs, the wavelength distinction is achieved by the reversal of the sign of the photocurrent (PC) for different wavelengths bands. Various technologies have been exploited for the realization of dual-polarity PDs based on the combination of photoelectrochemical PDs, thermoelectric PDs and photovoltaic PDs, however, rarely for the red and green12.

Evidently, all-solid-state semiconductor photovoltaic PDs are most compact, robust and can be operated in the self-powered mode without external power supply. For such PDs, which are based on p-n heterojunctions and Schottky junctions, dual-polarity photoresponse requires two distinct depletion regions in series with opposite energy band bending. They are created in a stack of p-type and n-type semiconductors, preferably exhibiting a smaller energy bandgap at the bottom and a larger energy bandgap at the top. Selective light absorption and photocarrier generation in the depletion regions and the transfer of the photocarriers to the contacts leads to the PCs with opposite sign for different wavelengths. The PCs need to be balanced by the PD design, which has been discussed in detail22. This involves the choice of the semiconductor materials with different energy bandgaps and wavelength dependent absorption coefficients and adjusting the layer thicknesses and doping concentrations.

Here we demonstrate a dual-polarity, all-solid-state photovoltaic PD for red-green imaging, operated in the self-powered mode. It is based on a core-shell n-InGaN nanowire (NW)/p-Cu2O heterostructure functional absorber directly grown on a p-Si (111) substrate by plasma-assisted molecular beam epitaxy (PA MBE). For PDs, InGaN is the ideal material due to the large absorption coefficient, the tunability of the direct energy bandgap over the whole visible range from the UV to the near-infrared with increasing In content and the large carrier mobility. Therefore, for the red-green PD, InGaN is ideally suited for the bottom n-type, red-light absorber with bandgap energy below ~ 1.8 eV of red light23,24. For the required high In content InGaN is highly n-type conductive due to unavoidable defects, acting as donors. Moreover, such In-rich n-InGaN forms an ohmic contact with the p-Si substrate due to the alignment of the p-Si valence band with the n-InGaN conduction band, allowing for easy back-contacting schemes25,26. However, due to the high unintentional n-type conductivity, In-rich InGaN is problematic to be used for the top p-type, green-light absorber with an energy bandgap below ~ 2.2 eV and high In content above 25%. Here, p-type Cu2O offers the ideal solution. Cu2O is a p-type semiconductor due to Cu vacancies with an energy bandgap of 2.1 eV. Most important, the conduction band and valence band of Cu2O are well aligned with those of In-rich InGaN27,28,29,30. The successful use of p-Cu2O as hole collector for InGaN-based PDs has been demonstrated31,32. In this study, the fabrication, the structural and optical properties and the operation mechanism of the red-green core-shell n-InGaN NW/p-Cu2O dual-polarity PD are presented. Balanced positive and negative PCs are obtained for a several-µm-thick Cu2O layer. The role of the core-shell InGaN NW structure33,34is mainly to suppress surface recombination to enhance the overall PC. The basic dual-polarity PD operation parameters are given. Finally, the imaging capability of the PD with excellent red-green color discrimination is experimentally demonstrated for red and green test images displayed on a mobile phone screen35,36. This real-life red-green imaging together with the balanced red and green PCs are the key features of the presented PD.

Results and discussion

PD structural and optical properties

Figure 1a shows a cross-sectional overview scanning electron microscopy (SEM) image of the optimized PD structure. The core-shell InGaN NWs grown by plasma-assisted molecular beam epitaxy (PA MBE) on the p-Si (111) substrate are 300 nm long. The Cu2O layer electrodeposited on top is 6 μm thick and is covered with a 150-nm thick Indium Tin Oxide (ITO) layer. A cross-sectional SEM image of the as-grown InGaN NWs is shown in the inset. More detailed SEM images of the different PD parts are shown in Fig. S1 (Supporting information). The X-ray diffraction (XRD) omega-two-theta spectra of the InGaN NWs with the 6-µm thick Cu2O layer on top are shown in Fig. 1b. For the Cu2O layer, the sole appearance of the Cu2O (011) and (111) reflections at 29.58° and 36.44° evidences phase purity. The InGaN NW reflections cannot be seen because the Cu2O layer is too thick. For the bare InGaN NWs, next to the pronounced Si (111) reflection at 28.44°, a reflection at 33.56° is observed. This reflection is due to the InGaN NW shell with an In content of 30.1%, assuming full lattice relaxation. The corresponding bandgap energy is 2.06 eV taking a band bowing parameter b of 2.5 for slightly disordered InGaN. The room-temperature photoluminescence (PL) spectrum of the InGaN NWs in Fig. 1c exhibits the PL peak wavelength of 755 nm or PL peak energy of 1.64 eV which is assigned to the InGaN NW core. The corresponding In content is 42.5%. Lateral quantum confinement of electrons and holes in the NW core is neglected here, which underestimates the In content34. However, for the detector operation the optical bandgap energy of the NW core is relevant which is given by the PL peak energy. The dominance of the NW shell in the XRD spectra is due to the large shell-to-core volume ratio. The dominance of the core in the PL spectra is due to the relaxation of photogenerated carriers from the shell to the core or excitation below the shell bandgap energy. The room-temperature PL spectrum of the Cu2O layer, shown in Fig. S2 (Supporting information) exhibits the PL peak wavelength of 575 nm or PL peak energy of 2.15 eV. To recall, the core-shell NWs naturally form during growth at elevated temperature due to the interplay of In deposition, diffusion, desorption and incorporation on the m-plane NW sidewalls and c-plane NW top. The In desorption on the m-plane NW sidewalls sets in at a lower temperature than the In desorption on the c-plane NW top. This leads to a reduced supply of In atoms from the NW sidewalls to the NW top with respect to the direct supply of In atoms on the NW top. The consequence is an In-poor shell surrounding an In-rich core.

Direct evaluation of the present core-shell InGaN NW structure comes from the transmission electron microscopy (TEM) and energy dispersive X-ray (EDX) analysis. The TEM image in Fig. 1d shows a single NW with average diameter of 30 nm. In the more chemically sensitive high-angle annular dark-field (HAADF) image in the inset the In-rich InGaN NW core is visible as bright contrast surrounded by a darker contrast from the In-poor InGaN NW shell. The high-resolution TEM image of the InGaN NW core in Fig. 1e with the electron diffraction image in the inset reveal high crystal quality. Direct chemical resolution of the core-shell InGaN NW structure is obtained by the EDX elemental mappings of the In and Ga signals, overlaid in Fig. 1f. The corresponding intensity line scans of the In and Ga EDX signals across the NW along the yellow dotted line are shown in Fig. 1g. The EDX spectra shown in Fig. 1h are taken at the positions indicated in Fig. 1f. The In-rich InGaN NW core with about 20 nm diameter is clearly distinguished inside the In-poor InGaN NW shell with a quite well-defined interface. More EDX analysis of the complete PD structure, including the Si (111) substrate, the Cu2O layer and the ITO layer is presented in Fig. S1 (Supporting information).

Fig. 1
figure 1

Structure and composition of the solid-state PD. (a) Cross-sectional overview SEM image of the dual-polarity core-shell n-InGaN/p-Cu2O PD structure with the SEM image of the as-grown InGaN NWs in the inset. (b) ω−2θ XRD spectra of the bare core-shell InGaN NWs and the InGaN NW/Cu2O heterostructure on Si (111). (c) PL spectrum taken at room temperature of the core-shell InGaN NWs. (d) TEM image of a single core-shell InGaN NW with the corresponding HAADF image in the inset. (e) High-resolution TEM image of the InGaN NW core with the electron diffraction image in the inset. (f) Overlaid In and Ga EDX elemental mappings. (g) In and Ga EDX intensity line scans as indicated in (f). (h) EDX spectra taken at the core and shell positions indicated in (f).

Full size image

PD operation mechanism

The photograph of the experiment and schemes of the PD structure and PC for red and green light illumination are shown in Fig. 2a and b with the core-shell n-InGaN NW/p-Cu2O heterostructure, the p-Si plus GaIn eutectic bottom contact and the ITO plus Ag mesh top contact. The polarity refers to the p-type Cu2O layer on top. To explain the dual-polarity PD operation mechanism, the PC under chopped illumination and the energy band structure of the core-shell n-InGaN NW/p-Cu2O heterostructure together with the photocarrier generation and transfer are schematically drawn for red light illumination in Fig. 2c,d and for green light illumination in Fig. 2e,f. The inner upward energy band bending is due to the Cu2O-InGaN p-n junction formation and the near-surface downward energy band bending is due to Cu2O surface states. Photocarrier separation in the associated depletion regions and internal electric fields drives the PC. The width of the depletion regions depends on the doping concentrations. Negative PC is obtained for depletion regions with upward energy band bending and positive PC for depletion regions with downward energy band bending. Due to the energy bandgaps of the InGaN NW core and Cu2O layer, red light is mainly absorbed by valence band-to-conduction band transitions in the InGaN NW core leading to negative PC. The Cu2O layer is transparent for the red light, apart from scattering and defect absorption, and absorbs only the green light by conduction band-to-valence band transitions, leading to positive PC. However, green light may also reach the inner InGaN NW core being absorbed there and leading, in addition, to negative PC. Balancing of the negative and positive PCs is easiest achieved by adjusting the thickness of the Cu2O layer for a given length of the InGaN NWs, which is pursued here.

For simplicity, the bottom and top contacts are omitted in Fig. 2d,f. The bottom p-Si substrate forms a near-ohmic contact with the core-shell n-InGaN NWs, see Fig. S3 (Supporting information), due to the band alignment of the Si valence band with the conduction band of In-rich InGaN, fully supporting the small PCs. The degenerate ITO top contact layer forms a Schottky-like junction with the p-Cu2O layer. Therefore, the Si substrate and the ITO top contact layer are not relevant for the PC generation, only for the photocarrier collection. There is no energy band bending and, therefore, no photocarrier separation takes place to drive a PC at an ohmic junction or within a degenerate semiconductor.

Fig. 2
figure 2

Experiment and operation mechanism of the dual-polarity solid-state PD. (a) Photograph of the experiment with red and green diode laser light illumination and top-view SEM image of the PD in the inset. (b) Scheme of the complete core-shell n-InGaN NW/p-Cu2O PD structure with ITO plus Ag mesh top contact and p-Si substrate plus GaIn eutectic bottom contact and experimental setup for red and green light illumination. (c–f) Scheme of the photocurrent under chopped illumination and the energy band structure of the core-shell n-InGaN NW/p-Cu2O heterostructure functional absorber together with the photocarrier generation and transfer, (c,d) for red light illumination and (e,f) for green light illumination.

Full size image

Considering in more detail the photocarrier separation and transfer, the negative PC generated by the red light absorbed in the core of the core-shell InGaN NWs near or in the inner depletion region with upward energy band bending is driven by the electrons moving downwards and the holes moving upwards. The electrons are freely transferred to the p-Si bottom contact. The holes are transferred to the ITO top contact, however, facing an energy barrier at the near-surface depletion region with downward energy band bending in the Cu2O layer. The positive PC generated by the green light absorbed in the Cu2O layer near or in the near-surface depletion region with downward energy band bending is driven again by the electrons moving downwards and the hole moving upwards. The electrons are freely transferred to the ITO top contact. The holes are transferred to the p-Si bottom contact, however, facing an energy barrier at the inner p-Cu2O/n-InGaN depletion region with upward energy band bending. At each energy barrier, the holes accumulate and create a photovoltage which reduces the energy barriers due to the band bendings. The residual energy barriers are overcome by thermal excitation, reducing the PC by the Boltzmann factor \(\:exp\left(-E/kT\right)\), where E is the residual energy barrier height, T is the absolute temperature and k is the Boltzmann constant.

Considering in more detail the balance of the negative and positive PCs, the red light creates a negative PC because it is selectively absorbed in the n-InGaN NW core. The green light, however, is absorbed not only in the p-Cu2O layer but can be also absorbed in the n-InGaN NW core to create positive and negative PCs. The total PC is the sum of both the negative and positive PCs. Therefore, to balance the negative and positive, red and green PCs, the Cu2O layer has to be thick enough to almost fully absorb the green light and to avoid that too many photocarriers reach the inner n-InGaN NW/p-Cu2O heterojunction by diffusion. For optimization of the Cu2O layer thickness, electrochemical photodetection is adopted because of the ease of PD fabrication. The PD as anode and a Pt cathode are immersed in 1.5 M NaI solution. For the balancing of the red and green PCs this has no influence because the near-surface downward energy band bending of Cu2O is governed by Cu2O surface states and largely independent of the top contact being an electrolyte or ITO. In Fig. 3 the dependence of the red and green electrochemical PCs on the Cu2O layer thickness is plotted.

Fig. 3
figure 3

Balancing the red and green electrochemical PCs by the Cu2O layer thickness. PC densities for 650-nm red light illumination and for 520-nm green light illumination as a function of the Cu2O layer thickness. The illumination power density is 5 mW/cm2.

Full size image

For too thin Cu2O layers, the negative PC generated in the inner depletion region of the core-shell n-InGaN NWs by the green light overcompensates the positive PC generated in the near-surface depletion region of the Cu2O layer. In this case, only negative PCs are observed for both the red and green light, which decrease with increasing Cu2O thickness. Eventually, the green PC becomes positive and saturates when only the green light absorption near or in the near-surface depletion region of the Cu2O layer is effective and negligible green light reaches the inner depletion region of the core-shell n-InGaN NWs. Therefore, from the dependence of the green PC on the Cu2O layer thickness an absorption length of 1.5–2 μm follows from the exponential law of absorption, when the green PC is decreased by the factor e−1, where e is Euler’s number. This corresponds to an absorption coefficient of Cu2O for the green light of 5–7 × 103 cm−1, which is significantly lower than that of InGaN of 2–3 × 105 cm−1, in agreement with the literature37,38. Therefore, a several µm thick optimum Cu2O layer is to be expected. However, as mentioned, also the red PC decreases in magnitude with increasing Cu2O layer thickness. This is attributed to red light scattering and defect absorption in the Cu2O layer. Therefore, the Cu2O layer shall not be too thick to diminish the red light too much to reach the core-shell InGaN NWs and also not to add a too large series resistance. The 6-µm thick Cu2O layer results in the right balance which, however, is at the cost of the total PC and, hence, responsivity and detectivity. This compromise cannot be avoided for obtaining dual-polarity PD functionality.

PD operation parameters

The basic operation parameters of the PD for dual-polarity/wavelength photodetection are given in Fig. 4 for the 6-µm thick Cu2O layer. Figure 4a shows a plot of the red and green PC densities as a function of the illumination power density between 5 and 50 mW/cm2. The original measurements are shown in Fig. S4(Supporting information). In dual-wavelength photodetection, these are the calibration curves to determine the red or green illumination power density after the wavelength has been identified by the sign of the PC. Important is the good linearity, quantified by the exponents of the power laws fitting the PCs. The exponent is 0.97 for the 650-nm red PC and 0.95 for the 520-nm green PC. Both exponents are close to one, indicating negligible trapping and non-radiative losses due to defects acting as non-radiative recombination centers. The responsivities and external quantum efficiencies (EQEs) calculated from the measured PCs, the illumination power densities and the dark current30 for the red and green PCs are plotted in Fig. 4b, which are rather constant for both wavelengths. To complete, the deduced specific detectivities are plotted in Fig. 4c for the measured dark current density of 20 pA/cm2.

To provide more confidence to the sign change of the PC between the red and green, illumination by 450, 635 and 808 nm diode lasers is added in Fig. 4d to the 520 and 650-nm green and red diode lasers, from the blue to the near-infrared. The illumination power density is 10 mW/cm2. For 450-nm illumination in the blue, the PC is positive, as expected. For 635-nm illumination in the light-red, the PC is already negative. For 808-nm illumination in the near-infrared, within the InGaN NW core bandgap, no PC is detected, confirming that the Si substrate has no contribution. This is due to the near-ohmic p-Si/n-InGaN NW junction even if the 808-nm illumination, for which the InGaN NWs and the Cu2O layer are transparent, reaches the Si substrate to generate photocarriers therein. The time resolved PC traces for 650-nm red illumination and 520-nm green illumination are shown in Fig. 4e,f. The rise and decay times, defined as the times for the PC to reach 90% or 10% of the stationary PC, are 11 and 7 ms for the red PC and 21 and 18 ms for the green PC. The rise and decay times of the red PC are somewhat shorter than those of the green PC because of the excellent optoelectronic properties of InGaN for PC generation.

Fig. 4
figure 4

Basic operation parameters of the solid-state PD for dual-wavelength photodetection. (a) PC density, (b) responsivity and EQE, and (c) specific detectivity as a function of the 650-nm red and 520-nm green illumination power density. (d) PC density under 450, 520, 635 and 808-nm blue to near-infrared illumination with 10 mW/cm2 illumination power density. (e,f) PC time traces for 650-nm red illumination and 520-nm green illumination. The rise and decay times are indicated.

Full size image

PD imaging performance

For imaging under realistic conditions, the signal-to-noise ratio (SNR) and the resulting minimum detectable illumination power density (MDP) are important, setting the limit to the ability to distinguish between red and green. The SNR is the ratio of the mean PC and the PC noise current, which is the standard deviation of the PC. Figure 5a shows a plot of the red and green SNRs as a function of the illumination power density from 0.5 to 2.5 mW/cm2 in the low-power density regime, where the illumination power densities of the red and green diode lasers can still be controlled reliably. The noise measurements are not limited by the measurement setup but by the 3 dB bandwidth of the solid-state PD of 17 Hz, given by \(\:{f}_{3dB}=0.35/{t}_{rise}\), where the longer green PC rise time is taken. Notably, the detector bandwidth is large enough for video imaging. The minimum detectable illumination power density is defined for a 3 dB SNR, that is, when the mean PC is 1.4 times larger than the noise current, indicated in Fig. 5a. By extrapolation, shown in the inset of Fig. 5a, the minimum detectable illumination power density is 0.24 mW/cm2 for the red light and 0.22 mW/cm2 for the green light. This suffices to recognize the environment in the day light and images on the display of a mobile phone, TV, or smart watch.

Good mobile phone displays have a maximum brightness of 1700–1800 cd/m2. Candela is the SI unit for the brightness as perceived by the human eye, taking into account the sensitivity of the human eye, which is largest in the green. The conversion to mW/cm2 is 1 cd steradian/m2 = 1.464 × 10−4 mW/cm2 at 555 nm, where the sensitivity of the human eye is maximum 1. Therefore, 1700 cd/m2 corresponds to a power density of 0.25 mW/cm2 for the 555-nm green light within a solid angle of 1 sr. About the same power density is received by the detector for a color field larger than the area of the detector at 1-cm distance, as in the experimental setting, taking into account that the pixels of the display radiate into a rather small solid angle. For perceived RGB white light the brightness of the red light in cd is slightly smaller than the brightness of the green light by less than a factor of two. Therefore, for both the 650-nm red light and the 520-nm green light, the power densities in mW/cm2 for an image on the mobile phone screen are slightly above the minimum detectable illumination power densities measured for the n-InGaN NW/p-Cu2O PD in Fig. 5a. This imaging capability is demonstrated for the PD scanning the red and green test images on the mobile phone screen, shown in the photograph in Fig. 5b together with the experimental setup. Being above the 3 dB SNR, the red and green colors are unambiguously distinguished by the sign and sign reversal of the PC with the dark current as offset, shown in Fig. 5c,d, as anticipated.

Fig. 5
figure 5

Dual-wavelength imaging and red-green color discrimination of a mobile phone display with the solid-state PD. (a) Signal-to-noise ratio as a function of the 650-nm red and 520-nm green illumination power density. In the inset, the 3 dB signal-to-noise ratio is indicated giving the red and green minimum detectable illumination power densities by extrapolation. (b) Photograph of the mobile phone with the red and green test images on the display and photodetector measurement setup. (c,d) 650-nm red and 520-nm green photocurrent traces when scanning the red and green color fields with the dark current as offset. The insets show the photographs of the experiment.

Full size image

Conclusions

A red-green, dual-polarity, photovoltaic photodetector was demonstrated for imaging, operated in the self-powered mode without external power supply. The functional absorber comprised a core-shell n-InGaN nanowire/p-Cu2O heterostructure with an inner depletion zone with upward energy band bending and a near-surface depletion zone with downward energy band bending producing the negative and positive photocurrents for red and green light illumination. The structural and optical properties of the dual-polarity photodetector were assessed in detail and the operation mechanism was discussed, based on the selective absorption of red and green light in the smaller bandgap energy InGaN NW core and the larger bandgap energy Cu2O layer. For balanced positive and negative photocurrents, a several-µm-thick Cu2O layer was required. The basic operation parameters of the dual-wavelength photodetector were given, including the dependency of the photocurrent, responsivity and specific detectivity on the red and green illumination power density, as well as the response rise and fall times. The imaging capability of the dual-polarity photodetector with superior red-green discrimination was experimentally demonstrated under realistic conditions for red and green test images displayed on a mobile phone screen. This builds the bridge to further applications such as the steering of autonomous vehicles by traffic lights as one prominent example where red and green light signals and indicators are encountered and need to be recognized and processed in daily life.

Experimental section

InGaN growth

The growth of the core-shell InGaN NWs on p-Si (111) substrates was performed by PA-MBE. Before growth, the native oxide on the Si surface was removed by etching in 10% HF aqueous solution for 2 min. The cleaned Si substrate was then loaded into the MBE buffer chamber and degassed at 300 °C for 1 h. After transferring into the MBE growth chamber, InGaN growth was performed with the reactive nitrogen source set to 390 W radio-frequency (RF) power and 2.5 standard cubic centimeters per minute (sccm) molecular N2 flow rate. The In and Ga source temperatures were 733 and 800 °C, leading to slightly N-rich growth conditions with about 0.2 μm/h growth rate for compact InGaN layers. The InGaN growth temperature was 530℃ which is above the onset of In desorption at around 500 °C. The growth time was 30 min for the 300 nm long InGaN NWs.

Cu2O deposition and PD fabrication

The Cu2O layer was prepared by electrodeposition on the core-shell InGaN NWs in a three-electrode electrochemical cell. For electrode fabrication, the InGaN NW sample was contacted with GaIn eutectic on the Si substrate back side, mounted on a Cu stripe, fixed to a glass plate and encapsulated with insulating epoxy resin, leaving a 0.05 cm2 opening for contact with the electrolyte. The InGaN NW electrode, a platinum mesh electrode and a saturated calomel electrode (SCE) were connected as working electrode, counter electrode and reference electrode to an electrochemical workstation (CHI600E). The aqueous electrolyte contained 0.4 M CuSO4 and 3 M lactic acid, adjusted to a pH of 12 by adding 5 M NaOH. The deposition voltage was − 0.4 V versus SCE. During the deposition under magnetic stirring with 180 rpm, the temperature was set to 35 °C by a hot plate. The deposition was followed in real-time by recording the deposition current density and charge density, shown in Fig. S5 (Supporting information). 1 C/cm2 corresponds to 1 μm compact Cu2O layer thickness. A large number of Cu2O depositions with Cu2O thicknesses between 0.7 and 7 μm were carried out until the best 6-µm Cu2O layer thickness was identified, which led to well-defined and well-balanced dual-polarity and dual-wavelength photodetection. Electrochemical photodetection was adopted for this optimization because of the ease of PD fabrication. The samples were connected to the electrochemical workstation as working electrode together with a Pt counter electrode immersed in 1.5 M sodium iodide solution. For the final PD, revealing similarly well-balanced PCs, 150 nm ITO was deposited on the top of the 6-µm thick Cu2O layer by magnetron sputtering (AJA International, ATC Orion 8) through a mask opening, defining the detector area of 0.05 cm2. Metallization was by a 50 nm thick silver mesh contact deposited by thermal evaporation for maximized carrier collection (INFICON SQC-310, Shenyang Kecheng Vacuum Tech Co, Ltd). The Si substrate back side was contacted by GaIn eutectic.

Characterization

The structural properties were assessed by SEM (ZEISS Gemini 500), XRD (BRUKER D8 ADVANCE) and TEM (JEM-2100 h 200 KV) with an attached spectrometer (Bruker Dual QUANTAX 200) for EDX analysis. The optical properties were assessed by PL spectroscopy at room temperature using the green 532-nm line of a Nd-YAG solid-state laser as excitation source and a single monochromator with attached Si charge-coupled device (CCD) for detection. The PC measurements were performed using a source meter (Keithley 2601B) and diode lasers with 650 and 520-nm wavelengths as illumination sources. The power density is uniform over the detector area. The PC response times were measured using a chopper and current amplifier (SR570).