Abnormal Photocurrent in Semiconductor p‐n Heterojunctions: Toward Multifunctional Photoelectrochemical‐Type Photonic Devices and Beyond

Semiconductor p‐n heterojunctions are important building blocks for modern electronic and photonic devices. Further combining semiconductor p‐n heterojunctions with light and electrolyte environment, interesting photoelectrochemical (PEC) phenomena can occur, which enriches the design principles of multifunctional devices. In fact, recent years have witnessed the emergence of PEC‐type photonic devices. For PEC‐type photonic devices, a key to realize multifunctionality is to control the photocurrent polarity of the photoelectrode. In this study, an abnormal photocurrent is reported from p‐InGaN/n‐GaN nanowire heterojunctions under a blue light illumination: although n‐GaN is transparent to the blue light (and thus optical absorption mainly occurs in p‐InGaN) and p‐InGaN in principle can only give negative PEC photocurrent, the detailed experiments show that positive PEC photocurrent can be generated from the p‐InGaN segment due to the existence of the built‐in electric field at the p‐n junction. This study shows a new route to control the photocurrent polarity in a semiconductor p‐n heterojunction photoelectrode. This unveiled role of the built‐in electric field is expected to impact the design of emerging PEC‐type photonic devices, as well as other novel photonic and electronic devices based on semiconductor nanowire p‐n heterojunctions.


Introduction
[3][4][5] Further combining semiconductor p-n heterojunctions with light and electrolyte environment, rich physical and chemical phenomena can happen, broadening the design DOI: 10.1002/aelm.202300274principles of multifunctional electronic and photonic devices.20][21][22][23][24][25] To realize multifunctionality in PECtype photonic devices, controlling the photocurrent polarity of the photoelectrode is a key.This further makes IIInitrides are particularly interesting.In fact, III-nitrides have a number of advantages for PEC-type photonic devices compared to other material systems.[28][29] Second, they have excellent chemical and mechanical stabilities, making it possible to have devices with long lifetime and applications in harsh environment for environmental monitoring. [30,31]Third, they have large and widely tunable bandgap energies, which provides unprecedent light detection capability as well as photocarrier transport engineering. [4,30,32]Moreover, their nanowire form shares the generic benefits of nanowires such as large surface to bulk volume ratio, high material quality, and electrical doping control that can modify the surface charge properties.36][37][38][39][40] This can be understood by the unique band bending for p-type and n-type III-nitride nanowires.Shown in Figure 1a, for p-type III-nitride nanowires, due to the downward band bending at the nanowire and electrolyte interface, the photogenerated electrons will drift to the nanowire surface and participate in hydrogen evolution reaction (HER), leading to a negative photocurrent. [19,26,37,38,41]In contrast, for n-type III-nitride nanowires, due to the upward band bending at the nanowire and electrolyte interface, the photogenerated holes will drift to the nanowire surface and consequently undergo oxygen evolution reaction (OER), leading to a positive photocurrent. [18,22,28,39]n this study, we report an abnormal positive PEC photocurrent from photoelectrodes made with p-InGaN/n-GaN nanowire heterojunctions under a blue light illumination.Although n-GaN is transparent to the blue light (as such, light absorption mainly occurs in the p-InGaN layer) and p-InGaN itself does not typically support positive PEC photocurrent, a positive PEC photocurrent can be obtained by exploiting the hole accumulation induced by the built-in electric field at the junction.This concept of hole accumulation is shown in Figure 1b.Under a 405 nm light illumination, the light absorption only occurs in the p-type InGaN segment.Due to the downward bending at the nanowire and electrolyte interface, the photogenerated electrons drift toward the surface; in the meantime, the photogenerated holes will drift away from the nanowire surface due to the downward band bending.However, due to the strong built-in electric field at the junction, the migration of the photogenerated holes toward the substrate will be hindered.As such, the photogenerated holes tend to accumulate in the p-InGaN segment.[42,43] These accumulated holes form a reservoir that can further lead to a positive PEC photocurrent as shown in detail in the present study, making it possible to have both negative and positive PEC photocurrent using p-type III-nitrides.

Results and Discussion
Figure 2a,b show the titled-view and the top-view scanning electron microscopy (SEM) images of the p-InGaN/n-GaN nanowires.The significant increase of the nanowire diameter as the growth of InGaN segment is consistent with previous re-ports, e.g., Refs.[40,44] Mg and Si were the p-type and n-type dopants, respectively.The doping concentrations were similar to the previous reports. [19,37,45,46]Detailed discussion on the p-type doping can be found in the SupportingInformation.Figure 2c is a low-magnification transmission electron microscopy (TEM) image of a nanowire, with the inset showing a high-magnification TEM image.The crystalline planes are clearly seen from the inset image.The growth direction is also labeled.Detailed examination further suggests that such nanowires have excellent crystalline quality.Such nanowires were also examined under the scanning TEM (STEM) mode.Shown in Figure 2d-i is a lowmagnification high-angle annular dark-field (HAADF) image of the same nanowire as in Figure 2c, based on which elemental mapping for Ga and In was further carried out.The ADF signals for Ga and In are shown in Figure 2d-ii,iii, respectively.It is seen that while the Ga signal exits in the whole nanowire, the In signal only exists in the region roughly ≈150 nm from the nanowire root, confirming the formation of the p-InGaN/n-GaN junction.We have further estimated the In content in the InGaN segment through the room temperature photoluminescence experiments (Figure S1), and an In content of ≈26 mol.% was derived (Supporting Information).Therefore, the band edges of the present InGaN nanowire segment are able to straddle redox potentials of water. [27,29]he PEC experiments were subsequently performed.Figure 3a shows the experimental setup schematic.A photo of the photoelectrode is also shown in the inset of Figure 3a.The nonchopped light linear sweep voltammetry (LSV) is shown in Figure 3b.It is seen that under the dark condition, there is essentially no current, suggesting the absence of any electrochemical reactions at the nanowire surface.Under the 405 nm light illumination, a positive photocurrent appears as the potential changes.This is in contrast to the previous studies, wherein for p-type III-nitrides only negative photocurrent can be measured when the light absorption only occurs in the p-type III-nitrides, e.g., Refs.[22,27,37,40] Although in general III-nitride nanowires are   typically highly chemical resistant, to further confirm the photoresponse, the chopped-light LSV was also measured.Figure 3c shows a typical curve, confirming the positive photocurrent from the nanowire photoelectrodes.
To further understand the nature of the positive photocurrent, we have further measured the photocurrent from photoelectrodes made with n-GaN nanowires grown with the condition similar to that of the n-GaN nanowire segment in the p-InGaN/n-GaN nanowires, as well as the photocurrent from photoelectrodes made with p-InGaN nanowires grown with the condition similar to that of the p-InGaN nanowire segment in the p-InGaN/n-GaN nanowires under the 405 nm light illumination.The detailed SEM images of these nanowires, together with the discussion on the nanowire morphology, can be found in Supporting Information (Figure S2,S3).Figure 4 show the LSV results of such photoelectrodes.It is seen that for n-GaN nanowire photoelectrodes, only negligible photocurrent is measured, due to the transparency of n-GaN to the incident light, which rules out that the positive photocurrent is due to n-GaN.For p-InGaN nanowire photoelectrodes, it is seen that the photocurrent stays negative even under positive potentials, which indicates that the positive photocurrent measured from the p-InGaN/n-GaN nanowires is related to the built-in electric field at the junction.
We have further measured the photocurrent from p-InGaN nanowire photoelectrodes up to 2 V.The results are shown in the Supporting Information (Figure S4).Briefly, it is found that, although a positive electrochemical current is measured at larger potentials, the photoelectrochemical current remains negative, and no photoelectrochemical current is measured ultimately.
The above experiments therefore confirm the measurement of the abnormal positive PEC photocurrent, and such a positive photocurrent is related to the built-in electric field at the p-n junction.The appearance of the positive PEC photocurrent can be explained by the following: due to the presence of the built-in electric field at the p-n junction, the photogenerated holes tend to accumulate in the p-InGaN region, and further driven by the applied positive potential that favors the charge carrier transfer process of the photogenerated holes to the nanowire surface, [47] OER is made possible.Essentially, the built-in electric field leads to hole accumulation, and further driven by the applied positive potential, the holes migrate to the surface along the closed electric circuit direction, reducing downward band bending and consequently leading to OER.In the meantime, in this case, the photogenerated electrons will migrate toward the counter electrodes, leading to HER at the counter electrode.As such, a positive PEC photocurrent can be measured.
For the present nanowire p-n heterojunction photoelectrode, a positive potential of 0.5 V (versus Ag/AgCl) is required in order to obtain the positive PEC photocurrent, for potentials less than 0.5 V, the PEC photocurrent stays negative.While the presence of the negative PEC photocurrent under negative potentials can be explained by the accelerated charge carrier transfer process of the photogenerated electrons to the nanowire surface under negative potential, [8,31,36] the presence of the negative PEC photocurrent under positive potentials (but less than 0.5 V) could indicate the downward band bending nature of the p-type nanowires when in contact with the electrolyte, and thus HER is a more naturally favorable process at the nanowire surface (such that in order to obtain a positive PEC photocurrent an even larger applied positive potential is needed).
Further following the built-in electric field induced hole accumulation model described above, one would expect that using the 405 nm light under similar excitation conditions, the negative photoresponse from the p-InGaN/n-GaN nanowire photoelectrode would be weaker than that from the p-InGaN nanowire photoelectrode.This can be seen by comparing Figure 3b,4b.The time-dependent photocurrent at 0 V (versus Ag/AgCl) in Supporting Information (Figure S5) also confirmed this difference.
In the end, to provide more details of the abnormal positive photocurrent, the time-dependent photocurrent under different excitation powers at an applied potential of 1 V were further measured.From Figure 5a, it is seen that as the excitation power density increases, the photocurrent density increases.Figure 5b further plots the photocurrent density as a function of the excitation power density.A nearly linear response can be seen, suggesting that the photocurrent density might be limited by the photocarrier generation process. [48,49]Also shown in Figure 5b is the calculated responsivity as a function of the excitation power density.It is seen that the responsivity of the photoelectrode is nearly excitation independent, in the range of 0.3 to 0.4 mA W −1 .The time-dependent photocurrent density under different applied potentials were also measured.From Figure 5c, it is seen that stable photocurrent can be obtained, and the polarity of the photocurrent changes as the potential increases.Figure 5d,e further plot the photocurrent density and responsivity as a function of the applied potential for the negative photocurrent and the positive photocurrent, respectively.It is seen that in both cases, the photocurrent density increases nearly linearly with the applied potential.Moreover, responsivities up to ≈1 mA W −1 and 0.3 mA W −1 are measured for the negative and the positive photocurrent, respectively.

Conclusion
In this work we report an abnormal photocurrent from p-InGaN/n-GaN nanowire heterojunctions.It is found that under the 405 nm light illumination, although the n-GaN nanowire segment is transparent and the light absorption only occurs in the p-InGaN layer, a positive photocurrent can be obtained from the p-InGaN nanowire segment, due to the built-in electric field at the p-n junction induced extra degree of freedom in controlling the photocarrier dynamics.This directly pinpoints the role of the built-in electric field in controlling photocarrier dynamics.Such a built-in electric field induced photocarrier dynamics control not only applies to the blue band, but also to other bands and other semiconductor heterojunctions.This finding, therefore, could redefine the design principle of multifunctional PEC-type photonic devices for a wide range of applications in sensing, selective solar fuel production generation, switchable logic for information processing, and so on; and depending on the applications one may want to utilize or minimize the effect of the built-in electric field, e.g., in the previously reported photoelectrodes for dual-wavelength photodetection, one would like to minimize the effect of the builtin electric field. [6,7,16,17]More importantly, this finding may also impact other novel semiconductor nanowire p-n heterojunction based photonic and electronic devices, beyond PEC-type photonic devices.

Figure 1 .
Figure 1.a) Downward band bending in p-type III-nitride nanowires that derives HER, and upward band bending in n-type III-nitride nanowires that derives OER, correlating to the negative and positive photocurrent, respectively.b) Schematic of the p-InGaN/n-GaN heterojunction nanowires under a 405 nm light illumination.The built-in electric field at the junction leads to hole accumulation in the p-type segment.These accumulated holes form a reservoir for a positive PEC photocurrent (details are described in the main text, together with the experimental evidence), making it possible to have both negative and positive PEC photocurrent with p-type III-nitrides.CB and VB denote conduction band minimum and valance band maximum, respectively.

Figure 2 .
Figure 2. a) and b) Titled-and top-view SEM images of the p-InGaN/n-GaN nanowires.c) A low-magnification TEM image of a nanowire, with the inset showing the high-magnification TEM image.d) HAADF image of the same nanowire and the corresponding Ga and In ADF signal mapping.

Figure 3 .
Figure 3. a) Schematic of the PEC experiments setup, as well as the dominant PEC process for the p-InGaN/n-GaN nanowire photoelectrodes at V > 0.5 V and V < 0.5 V (versus Ag/AgCl).b) and c) Non-chopped light LSV and chopped light LSV for the p-InGaN/n-GaN nanowire photoelectrodes, respectively.The light excitation density was 7.5 Mw cm −2 .

Figure 5 .
Figure 5. a) The time-dependent photocurrent density at different excitation power densities at 1 V. b) Photocurrent density and responsivity versus the excitation power density.c) The time-dependent photocurrent density at different applied potentials at an excitation power density of 4.2 mW cm −2 .d) and e) photocurrent density and responsivity versus the applied potential for the negative photocurrent and positive photocurrent, respectively.