1. Introduction

In the past few decades, organic semiconductors (OSCs) have attracted a lot of interest for the application of optoelectronics devices. Organic light-emitting diodes (OLEDs) have become a promising candidate for the next-generation flat panel displays, and organic solar cells are also challenging the commercial inorganic solar cells due to the intriguing combination of low weight and cost. In comparison with the transverse photovoltaic effect used in the solar cells, lateral photovoltaic effect (LPE) also has a wide range of applications, especially position-sensitive detectors (PSDs). Under a nonuniform irradiation on a semiconductor device (such as p-n junction), an additional photovoltage parallel to the plane of the film can be observed [1–3]. Besides in p-n junctions, the effect has been observed in various structures, such as metal/semiconductor superlattices, metal-oxide-semiconductor (MOS), metal-semiconductor (MS) structures and perovskite heterostructures [4–9]. Recently, large LPE was also found in nanometer composite films on Si substrates [10–12]. Till now, only few studies have been reported for LPE in organic semiconductor related devices [13]. Thus, it is interesting to study what will happen if the upper layer is replaced by metal doped organic semiconductor films in the MOS/MS structures. Recently, a crossover of magnetoresistance was found in Co-Alq₃ granular films on Si substrates, which originates from the channel switching of carrier transport between upper Co-Alq₃ granular film and inversion layer underneath in the MOS-like structure [14]. In this paper, we investigated the lateral photovoltaic effect in Co-Alq₃ granular films on Si substrate, and discussed its possible mechanism.

2. Experimental details

Co-Alq₃ granular films were deposited on silicon substrates using co-evaporation technique. To study the influence of substrate on LPE, three kinds of n-type Si(111) wafers with different resistivities (2-4Ω-cm, 70-80Ω-cm and >1000Ω-cm) were used. Before deposition, in addition to rinsing with distilled deionized water and cleaning with acetone, HF pretreatment (8mol/L, 10min) was carried out to remove the native oxide layer of Si wafers. Finally, in order to remove HF residuals, the etched substrates were rinsed with distilled deionized water and dried with gaseous nitrogen . The base pressure of the evaporation chamber is about 1 × 10⁻⁴Pa. Volume ratios of Alq₃ and Co are obtained by adjusting the evaporation rate of the two materials. The thicknesses of the films are about 80nm. Both the evaporation rate and the film thickness are determined by a quartz crystal thickness monitor (Sigma SQM-160).

The as-deposited samples were cut into 5×12 mm², and two indium electrodes were placed on the Co-Alq₃ film surface or the Si substrate. The schematic drawing of the LPE measurement is shown in the inset of Fig. 1 . The samples were scanned spatially with a semiconductor laser (10mW, 532nm) focusing on a roughly 200μm-diameter spot at the surface in the dark. The open-circuit photovoltage was measured using a Keithley 2001 multimeter. I-V curves were measured using a Keithley 2400 sourcemeter. All the measurements in this paper were performed at room temperature.

 

Fig. 1 LPV as a function of the laser position (x) observed in Co_0.4(Alq₃)_0.6/Si (2-4 Ω-cm) sample. The upper inset shows the partial enlarged drawing of the LPV curve in reverse mode and the bottom inset shows the schematic drawing of LPE measurement.

Download Full Size | PDF

3. Results and discussions

Figure 1 shows the dependence of the induced photovoltage on the position of light for Co_0.4(Alq₃)_0.6/Si (2-4 Ω-cm) sample at room temperature, where V_AB denotes the LPV between the indium electrodes on the granular film surface and V_CD denotes the LPV on the Si substrate respectively. For V_AB, the LPV shows a monotonic linear positive dependence on the position within the range of two contacts, which is similar to that of inorganic LPV devices [5,6]. The optimal LPV was obtained near the electrodes, reaching about 60mV. When the laser spot passed the center, the sign of LPV was reversed. At the center of the two contacts, no obvious LPV was measured. The open-circuit position sensitivity, which means the change of the output voltage for displacement of the laser spot, is 34.7mV/mm, close to the LPV value in inorganic LPV devices [5,6]. While for V_CD, it only shows very small negative dependence of LPV on the position, with the open-circuit position sensitivity being about 1mV/mm, as seen in the upper inset of Fig. 1. The results are different from those of MS junctions in the early studies, where LPV was only observed on the side of semiconductors, while no obvious LPV was observed on the metal side. Recently, large LPV was also observed at upper layer side in MS structures due to the appropriate resistance of thin films [6,11].

In order to understand the mechanism of LPE, we investigated the transverse I-V curve of the above sample, as shown in Fig. 2 . The cathode contact is located on the Si substrate, and the anode contact is located on the upper Co-Alq₃ film. The I-V curve shows typical nonlinearities and rectifying current-voltage behavior, indicating the transport characteristic of Schottky barrier (SB) in the MS-like structures.

 

Fig. 2 The transverse I-V curve for Co_0.4(Alq₃)_0.6/Si (2-4 Ω-cm) sample

Download Full Size | PDF

The LPE mechanism of such metal/semiconductor or Schottky-Barrier structures has been well established, which is similar to p-n junctions [1,2,6]. In the Co-Alq₃/Si structure, there exists a Schottky field near the interface between Co-Alq₃ granular film and Si substrate. When laser is incident onto the Co-Alq₃/Si structures, electron–hole pairs are generated inside Si substrate at light position. Due to the Schottky field, the minority holes transit into the upper Co-Alq₃ film, and electrons remain in the n-type silicon substrate, which leads to a transverse photovoltage. In our LPE experiment, the light is focused on a roughly 200μm-diameter spot on the surface of the samples. On account of nonuniform illumination, the presence of excess remaining electrons and injected holes gives rise to a non-equilibrium distribution, resulting in a gradient of lateral voltage. In light to the charge of carriers, the position dependence of LPV is reversible for that measured on the p side and n side, as shown in Fig. 1. Moreover, the good linearity of LPV indicates the uniformity of the upper Co-Alq₃ film.

In our previous works, the oxide layer of Si substrate plays important roles in the transport properties of the samples [14]. To study the influence of oxide layer on LPE, we fabricated a corresponding sample without HF treatment for substrate. In the sample, there exists a thin SiO₂ layer (about 2.1nm) between Co-Alq₃ film and Si substrate. The thickness of SiO₂ layer was determined by a M2000 Spectroscopic Ellipsometer (J. A. Wollam). Figure 3 shows the position dependence of LPV for Co_0.4(Alq₃)_0.6/SiO₂/Si (2-4 Ω-cm) sample at room temperature. As we can see, the LPV is obviously smaller than that of the corresponding sample without oxide layer, with a sensitivity of about 5-6mV/mm for both sides. The decrease of LPV may be related to the blocking effect of SiO₂ layer, which decreases the possibility of the tunneling of carriers from Si substrate to upper granular film, thus causing decreased LPV in MOS structure. Similar phenomenon was also observed in Ti/Si structures [15].

 

Fig. 3 the position dependence of LPV for Co_0.4(Alq₃)_0.6/SiO₂/Si (2-4 Ω-cm) sample

Download Full Size | PDF

As we can see in Fig. 3, the LPV shows similar position dependence for both sides of the sample, which is substantially different from the conventional p-n junction theory, where LPV should be reversible on both sides, as shown in the above sample without SiO₂ layer. The unusual phenomenon was also observed in LSMO/SNTO p-n junction, Cu₂O/Si and Al doped ZnO/Si structures. It may be attributed to the Dember effect or thermal effect [7,8,11,16]. However, the exact mechanism is still unclear. As a result of the blocking effect of insulate layer, fewer holes are transmitted into upper granular film layer. Therefore, more unseparated carriers (electrons and holes) directly diffuse along the lateral direction on Si side, leading to Dember process. The much larger mobility of electrons than that of holes induces the separation of electron-hole pairs in Si substrate, resulting in different transient distribution of electrons and holes. The electrons diffuse farther away from the light spot than the holes, thus causing a higher electric potential near the light spot [7,8]. However, owing to thin thickness of the SiO₂ layer, a part of excess holes still transit into upper layer and diffuse along the lateral film, which contributes to the presence of conventional LPE. In our samples, the competition between the conventional LPE and Dember effect determines the transient distribution of electric potential in the structure, bringing about the above unusual LPV dependence in Fig. 3.

As is known to us, the Si substrate also plays an important role in the physical process of LPE. To investigate the influence of substrate on lateral photovoltaic effect in the structure, we deposited the Co_0.4-(Alq₃)_0.6 films on three kinds of substrates with different resistivities. The samples are labeled with A, B and C respectively (resistivities of Si wafer: 2-4Ω-cm, 70-80Ω-cm and >1000Ω-cm, with HF pretreatment). Figure 4 shows the LPV for the three samples. The LPE behavior of Samples B and C are similar to that of Sample A. The phenomenon of reversible position dependence of LPV is also observed in Samples B and C, showing the conventional LPV mechanism in MS-like structures. It's worth noting that the LPV values of the three samples are quite different. The open-circuit position sensitivities for granular film sides are 34.7mV/mm (Sample A), 10.3mV/mm (Sample B) and 4.1mV/mm (Sample C) respectively, as shown in Fig. 4(a). The largest LPV_AB was observed in Sample A, where dopant density is greater than that of Samples B and C. While for Si substrate sides, the optimal LPV_CD was found in Sample C, reaching a surprisingly high value of 78mV/mm, as shown in Fig. 4(b).

 

Fig. 4 the position dependence of LPV for Co_0.4(Alq₃)_0.6/Si samples with different resistivities of Si wafer: 2-4Ω-cm, 70-80Ω-cm and >1000Ω-cm

Download Full Size | PDF

The dependence of LPE on Si substrates can be explained as following. For the sample with low resistivity Si substrate (Sample A), there exists a larger built-in field at interface induced by the higher doping level than that of Samples B and C. When light is incident onto Sample A, more holes transit into the upper layer, thus leading to larger LPV_AB in Sample A. Moreover, for the sample with high resistivity Si substrate (Sample C), fewer impurities of Si substrate cause less combination of non-equilibrium carriers, which brings about the largest LPV_CD in Sample C.

4. Conclusion

In summary, large lateral photovoltaic effect (LPE) was observed in Co-Alq₃ granular films on silicon substrate at room temperature. The insertion of oxide layer shows decreased irreversible LPV, which possibly results from the blocking effect of the insulator layer. The dependence of lateral photovoltaic effect on substrates was also briefly investigated.