1. Introduction

Cuprous oxide (Cu₂O), which is a p-type semiconductor with a band gap of 2.0 eV, has been widely studied due to its merits such as nontoxic nature, abundant availability and low-cost production. Cu₂O thin films are generally applied to electrochromic devices [1], oxygen and humidity sensors [2,3] and particularly heterojunction solar cells [4,5]. Although many early works associated with this oxide were focused on the application based on the longitudinal photovoltaic effect across the heterojunction [6,7], few study has been reported for the lateral photovoltaic effect (LPE) measured on the surface of Cu₂O nano-films. Since the LPE in response to spot illumination was first discovered by Wallmark [8], similar studies have been conducted in many different systems and LPE has been used in a wide application for various of optical transducers and sensors, including position sensitive detectors, shaft encoders, optical alignment sensors and so on [9–13]. Compared with the previous researches, the structure of Cu₂O nano-films deposited on Si substrate reported in this work has presented a significantly giant lateral photovoltage (LPV) with the highest sensitivity reaching up to 113 mV/mm in PN mode. More interestingly, when an external voltage is applied, LPV in this system can be modulated and the position sensitivity is improved dramatically, almost 10 times as large as its initial value, which provides a huge potential prospect for manufacture of high-sensitivity sensors.

2. Experimental details

In this paper, the Cu₂O thin films were deposited on n-type Si (1 1 1) at room temperature by DC magnetron reactive sputtering. The thickness of the Si wafers is around 0.3 mm and the resistivity is in the range of 50-80 Ωcm. The base pressure of the vacuum system prior to deposition was 2.3 × 10⁻⁴ Pa. High pure Cu target (60 mm diameter) was used. An argon gas pressure of 2.3 Pa and an O₂ gas pressure of 0.7 Pa were maintained during deposition. The deposition rates, determined by stylus profile meter on thick calibration samples, were 3.7 Å/s. Serial Cu₂O films with a variety of thicknesses ranging from 10 nm to 100 nm were fabricated. These Cu₂O samples were later confirmed by a Bruker Discovery D8 X-ray diffractometer (XRD), as shown in Fig. 1. It is found that two XRD peaks at 2θ = 36.8^。and 61.9^。appeared, owning to the diffraction of the (111) and (220) planes of Cu₂O. The results were consistent with the previous work [14]. All the electrodes (about 1 mm in diameter) to the films were formed by alloying indium and showed perfect ohmic contact. Other experimental details are similar with our published papers on LPE [15].

 

Fig. 1 XRD pattern of a Cu₂O/Si structure by DC magnetron sputtering.

Download Full Size | PDF

3. Experimental results and discussion

The dependence of the induced photovoltage on the laser position with different typical Cu₂O film thickness is shown in Fig. 2(a). When each sample was scanned with a He-Ne laser (632 nm and 3 mW) focused on a roughly 50-µm diameter spot at the Si surface without any background light, a large LPV can be observed between electrodes on Cu₂O side. The LPV is the largest when the incident radiation spot is closest to the measurement electrodes and shows a monotonic linear decrease as the spot is scanned away from the electrodes, becoming zero at the midpoint of these two electrodes. The voltage is reversed when the light spot is moved across the center between the two contacts. Similar LPVs as a function of laser position have been observed in the Cu₂O film thickness of 58.0 nm, 63.2 nm, 66.9 nm and 74.3 nm, respectively. It shows the strongest LPE at the optimum value of 66.9 nm.

 

Fig. 2 (a) LPV measurement in the structures of Cu₂O/Si with different Cu₂O film thickness (58.0 nm, 63.2 nm, 66.9 nm and 74.3 nm). The bottom inset displays the schematic circuit of the V_AB measurement. LPVs in (b) PP and PN mode and (c) NN and NP mode as a function of laser position in Cu₂O(66.9 nm)/Si structure. The inset shows the layout of LPV measurement. A (2 mm), B (−2 mm), C (2 mm) and D (−2 mm) mark the electrodes.

Download Full Size | PDF

Figures 2(b) and 2(c) present the values of LPV in four different modes in the Cu₂O (66.9 nm)/Si structure, which are called as PN mode, PP mode, NN mode and NP mode. Here “P” and “N” represent “p-Cu₂O side” and “n-Si side”, respectively. The first capital letter represents on which side the LPV is measured, and the second capital letter represents on which side the laser illuminates. Compared with the LPV measured on the Si side (NP or NN mode), the LPV measured on the Cu₂O side (PP or PN) is lager due to the higher resistivity of the Cu₂O thin film than that of the Si wafer. In addition, the distances from electrodes to the junction are larger if measuring from Si side, which reduces the LPV. Interestingly, when LPV is measured on the Cu₂O side, PN mode produces a larger LPV than PP mode. Whereas, in the two modes that LPV is measured on the Si side, NP mode produces a larger LPV. Thus, if we change the measurement side, the laser illumination side which produces the larger LPV is altered. This result is intriguing but the mechanism is not well known and needs further investigation. As shown in Figs. 2(b) and 2(c), four modes all present a good linear LPV versus laser position, and the obtained sensitivities range from 25 mV/mm to a remarkable value of 113 mV/mm (achieved in PN mode).

To further study the LPE in this structure, an external negative bias voltage V_BD ( = V_B-V_D) is applied on electrodes B and D. Figure 3 illustrates the LPV (V_AB) as a function of laser position when the bias is −3 V. Without laser irradiation, the value of V_AB in this biased mode is about 2.7 V. When the laser spot is scanned from electrode D to C on the Si surface, however, a bias-assisted large LPV is observed. The range of V_AB reaches up to 1500 mV, whereas with no bias the initial range is 400 mV as illustrated in Fig. 2(b). Generally, the process of LPV enhancement with laser position shows two phases: (a) in the part where the laser spot is close to the D electrode (we call it region 1), V_AB rises rapidly from 2.2 V to 3.0 V in the range of 0.8 mm; (b) the changing rate of V_AB drops when the laser spot is close to the C electrode (region 2). Compared with no biased mode, the linearity in region1 deteriorates as the correlation coefficient [11] drops from 0.98 to 0.92. However, the dramatically enhanced sensitivity (974 mV/mm) indicates that a very small displacement of laser spot can lead to the giant change of V_AB.

 

Fig. 3 The LPV (V_AB) as a function of laser position in Cu₂O (66.9 nm)/Si structure when a bias voltage of −3 V is applied on electrodes B and D. The inset displays the schematic setup for LPV measurement, where A (2 mm), B (−2 mm), C (2 mm), and D (−2 mm) denote the electrodes.

Download Full Size | PDF

In addition, we also investigate the characteristic of V_AB in response to different bias voltages (V_BD from −1 V to −6 V). Table 1 shows the sensitivity measurements when laser spot is scanned through two regions between electrodes C and D. The sensitivities increase with rising of V_BD which implies they can be modulated by the bias voltage. When bias V_BD is −6 V, in region 1, the maximal LPV sensitivity 1114mV/mm is 10 times larger than its initial value of 113 mV/mm. To our knowledge such high sensitivity has not reached ever before. In region 2 the rising magnitude is much smaller as the bias increases.

Table 1. (Sensitivities measured on AB side in different bias voltage)

View Table

The dependence of LPV on the distance between two electrodes with a bias voltage of −3V is shown in Fig. 4. The range of LPV amplifies with the decrease of L and short L can boost the position sensitivity in region 1 greatly. Moreover, the turning point between two regions varies when the electrodes’ distance changes. The width of region 1 with large LPV changing rate will occupy more proportion of the whole distance between two electrodes as the L decreases.

 

Fig. 4 The experimental results of LPV (V_AB) measured in structure of Cu₂O(66.9 nm)/Si with a negative bias voltage of −3 V, in which the distances between AB ( = 2L) are 3.0 mm, 4.0mm, 5.0 mm and 6.0 mm, respectively. The inset shows the schematic setup for LPV measurement with a bias voltage of −3 V.

Download Full Size | PDF

The fully understanding of why LPV changing rate presents the two-region pattern and the sensitivity in region 1 increases significantly is not well available. After a careful inspection, we have tentatively generalized a qualitatively physical explanation. When there is no bias voltage on the electrodes B and D, the LPV sensitivity between two lateral electrodes can be presented as [16,17]

In the case of applying a −3V bias on the B and D electrodes, V_AB and V_CD are measured as 2.7V and −62 mV. Thus, the voltage V_AC on the A and C electrodes is −23.8 mV, much smaller than BD bias. In region 1, due to the larger bias V_BD, the recombination rate of photo-induced electron-hole pairs in Si wafer decreases (P becomes smaller) so that more holes are injected into the Cu₂O film, which means$N_0$increases and thus sensitivity $\kappa$improved greatly. However, the density of transition holes into Cu₂O film changes a little in region 2 because of small bias V_AC. We consider the different bias of V_BD and V_AC has a main and significant impact on this sensitivity variation in region 1 and 2.

Another factor which may influence this phenomenon is the variation of holes transport length in Cu₂O film. With the bias voltage, an extra electric field from electrode A to B will be formed in the lateral direction. The external electric field decays from electrode B to A because of the resistance effect. For the sake of simplicity, we suppose the electric fields in region 1 and 2 are approximately constant and the field in region 1 is much larger than that in region 2. Therefore the diffusion length${\lambda }_f$is different in the same direction and in the reversed direction of electric field. We note ${\lambda }_1$ and ${\lambda }_2$ as the holes transport length in the direction of BA and AB. So, in region 1 close to electrode B, the electric field is larger so that ${\lambda }_1$ decreases remarkably and ${\lambda }_2$ increases. In region 2, however, both ${\lambda }_1$ and ${\lambda }_2$ have a small change since the electric field is comparatively low. The effect on the sensitivity$\kappa$in virtue of the competition between the reduction of ${\lambda }_1$ and the increase of ${\lambda }_2$is difficult to evaluate theoretically. Nevertheless, it is reasonable to think this change of holes transport lengths plays a minor role in LPV sensitivity compared with the above-mentioned increasing transition holes$N_0$, since the variations of ${\lambda }_1$and ${\lambda }_2$ have a reversed effect on the sensitivity according to Eq. (1).

In addition, as two electrodes are close to each other, which means L is shorter, the LPV sensitivity increases. Moreover, the region affected by the bias voltage is comparatively enlarged within a short L, thus implying region 1 with large LPV changing rate will occupy more proportion of the distance between two electrodes as shown in Fig. 4. The turning point of LPV curve is probably determined by the instinctive conductive character of material and the thickness of Cu₂O [18]. The systematic search of this property is expected to be further investigated.

4. Conclusion

In this study, we report a giant lateral photovoltaic effect modulated by external bias in a Cu₂O/Si structure as its surface is illuminated by a laser. The experimental results have shown that LPE sensitivity greatly increases with the applied biased voltage. Although the nonlinearity increases as the distance L reduces, its high position sensitivity implies the potential utilizing in high accuracy measurement and position detection by signal processing. A mechanism about the effect of increased photo-generated holes has been presented to explain this result. The field-induced variation of diffusion length of these carriers is also discussed. We believe this remarkable LPV and high sensitivity are useful for the development of novel and multifunctional photo sensors.