Introduction

It is well known that the magnetization produced by ferromagnetic materials can be used to control the charge carrier flow in a nearby substrate, which serves as a fundamental process underlying many current technologies.1, 2 However, the literature on the integration of ferromagnetic materials into nanoscale semiconductor devices is rather limited.3 One-dimensional metal-oxide semiconductor nanostructures have been extensively investigated in the past because of their unique characteristics and broad potential for various applications.4, 5 In fact, a variety of optoelectronic devices based on one-dimensional nanowires (NWs) have been developed, such as emitters,6, 7 detectors8, 9 and transistors.10, 11 Recently, tin dioxide (SnO2) with a wide direct band gap of 3.6 eV and high quantum efficiency in the ultraviolet region has attracted a great deal of attention, and there are potential applications in many practical devices, such as visible-blind photodetectors and solar cells.12, 13, 14, 15 In addition, because NWs have a high surface-to-volume ratio, the surface of NWs can influence the conductivity markedly. Owing to charged defects located at SnO2 NWs surface, the upward band-bending forms a low-conductivity depletion layer. Once electron–hole pairs are photoexcited, holes will drift to the surface through the electric field, leaving unpaired electrons inside; thus, the recombination probability of electrons and holes reduces, and the lifetime of photo-generated carriers increases.16, 17 SnO2 NWs therefore have been developed to detect gas molecules, including CO, NO2, and so on18, 19 because the adsorbed gas molecules on the surface will extract charge carriers from SnO2 and react with adhered gas, resulting in altered conductivity. To improve the sensitivity of the devices, SnO2 NW has been functionalized with several different materials including catalytically active metals, p-type semiconductors, and semiconductor quantum dots.20, 21, 22, 23, 24 Through these previous efforts, the photocurrent (PC) gain of SnO2 NW devices can be not only greatly enhanced but also extended to a wider photoresponse spectrum range. In this paper, we provide an alternate approach to enhance the sensitivity of a single SnO2 NW photodetector that has two ferromagnetic nickel (Ni) electrodes. Our strategy is the first attempt at magnetic manipulation of photoresponse in a one-dimensional semiconductor device. Quite surprisingly, the enhancement factor ratio for magnetic to nonmagnetic electrode-containing devices exceeds 20 times, which is much larger than all of the previously reported values for semiconductor NW photodetectors. The underlying mechanism is interpreted in terms of the migration of electrons to the NW surface due to magnetized Ni electrodes, which will enhance both oxygen adsorbed and surface band-bending effects and lead to the amplification of PC gain. The results we describe here could therefore open up a new route for the creation of high-efficiency optoelectronic devices based on the coupling between ferromagnetic materials and nanostructured semiconductors.

Experimental procedures

The growth of SnO2 NWs is based on a vapor–liquid–solid process. In the synthesis process, an Au layer of 10 nm thickness was first deposited on M-plane (100) sapphire (Al2O3) to serve as the catalyst. Then, Sn powder was placed on a ceramic boat and put in a furnace with argon flow as a carrier gas at a flow rate of 200 sccm. The temperature was increased from room temperature to 1000 °C at a rate of 100 °C min−1. After a growth time of 5 min at 1000 °C, the SnO2 NWs were obtained. To reduce surface defects, the as-grown NWs were annealed at 1000 °C for 3 h. Figure 1a shows the field emission scanning electron micrograph image of the morphology of SnO2 NWs obtained by a JSM-6500F field emission scanning electron micrograph system (JEOL Ltd., Tokyo, Japan). The X-ray diffraction analysis of SnO2 NWs was carried out using a diffractometer (Panalytical X’pert PRO, PANalytical, Almelo, Netherlands) with Kα line as the incident radiation; X-ray diffraction is used to identify the structure of synthesized product. Most of the peaks can be perfectly indexed according to the tetragonal rutile structure of SnO2, as shown in the inset of Figure 1a. The transmission electron microscopy images of as-grown NWs and annealed NWs are shown in Figures 1b and c, respectively. We can clearly observe that the surface of the annealed NWs is smoother than that of the as-grown NWs.25

Figure 1
figure 1

(a) Field emission scanning electron microscope (FE-SEM) image of as-grown SnO2 nanowires (NWs). Inset: X-ray diffraction (XRD) pattern of as-grown SnO2 NWs. (b) Transmission electron microscopy (TEM) image of as-grown SnO2 NW. (c) TEM image of annealed SnO2 NW. (d) SEM image of single SnO2 NW device. The distance of the channel between two terminals is 8 μm, and the diameter of the NW is ∼50 nm.

Full size image

After the fabrication, the SnO2 NWs devices were made by the shadow-mask method,13 and Ni (200 nm)/Au (20 nm) electrodes were deposited on the SnO2 NWs to form ohmic contacts. A typical SEM image of a single SnO2 NW device is illustrated in Figure 1d. The distance of the channel between two terminals is 8 μm, and the diameter of the NW is ∼50 nm. A thermal annealing at 500 °C for 15 s after contact deposition was necessary to minimize the contact resistance. In our work, an external magnetic field provided by a pair of electromagnets is applied in the normal direction of the plane of the substrate. After the electrodes were magnetized, the NW device was transferred onto a stage under the microscope without external magnetic field. For the PC measurement, a He–Cd laser working at 325 nm was used as the excitation light source. A Keithley 236 measurement system (Keithley Instrument, Inc., Cleveland, OH, USA) was used to supply the dc voltage (0.1 V) and to record the PC.

Results and discussion

Figure 2 shows the dark currents of the device after the Ni electrodes were magnetized under different magnetic fields. We can clearly see that the dark current decreases with increasing magnetic field. This phenomenon of decreased dark current with increasing magnetic field can be understood as follows: when free electrons flow through the conducting channel of the NW, the electrons migrate to the surface of NW due to the Lorentz force caused by the magnetic field and trapped by the surface defects, the number of conducting electrons is therefore reduced and the resistance of NW increases. Moreover, the electron accumulation at the NW surface generates a depletion region near the surface of NW and a high-resistance zone is formed, which further increases the resistance of NW. When the magnetic field increases, the amount of the accumulated electrons is larger, and thus, the high-resistance zone near the surface of NW also becomes larger.22, 23 To be more quantitative, let us consider the Hall effect. The value of the Hall electric field EB can be expressed as , where is the drift velocity, and the magnetic field. As the number of accumulated charges is proportional to EB, the dark current therefore decreases with increasing magnetic field. In addition, we have performed a similar experiment for the magnetic field parallel to the axis of NW. In this case, the current does not change with increasing magnetic field. This lack of change in conductivity with a parallel magnetic field therefore provides additional evidence to support our interpretation. A more detailed examination of the decrease of dark current with increasing magnetic field is also discussed in a later section. Upon illumination with photon energy larger than the energy band gap of SnO2 (3.6 eV), the conductivity of SnO2 NWs increases greatly due to photogenerated electron–hole pairs. As shown in Figure 3a, the photoresponse of an NW device with a bias of 0.1 V under the illumination of He–Cd laser (325 nm) with the excitation intensity of 8 W m−2 can be clearly observed. Surprisingly, after the Ni electrodes were magnetized, the PC is greatly enhanced. For instance, the PC difference (Δi=ilight−idark) of the NW device can be enhanced by 590%, 1200%, 1750% and 2100%, as the electrodes were magnetized under the field of 350, 500, 750 and 1400 G, respectively.

Figure 2
figure 2

I–V characteristics of a single SnO2 nanowire (NW) device exposed to different magnetic field strengths.

Full size image
Figure 3
figure 3

(a) Photoresponse of a single SnO2 nanowire (NW) device with a bias of 0.1 V exposed to magnetic field strengths of 350, 500, 750 and 1400 G and the illumination of a He—Cd laser working at 325 nm with an excitation intensity of 8 W m−2. (b) The gain logarithmic plot versus illumination intensity of a single SnO2 NW device exposed to magnetic field strengths of 350, 500, 750 and 1400 G.

Full size image

Here, we investigate the PC gain of this NW photodetector with the magnetized electrodes. The current gain (Γ) of the photoresponse, a factor used to determine the efficiency of electrons induced by photon and collected during the PC measurement can be calculated using the expression13

where Δi is the current difference between the PC and dark current, q is the electron charge, hν is the photon energy of the incident light, P is the power of photon that the NW has absorbed and is expressed by P=I × l × d and I is the excitation intensity illuminating on the pattern, l and d are the length and the width of the NW, respectively, and η is the quantum efficiency, which is set to 1 for simplicity because the magnitude rather than absolute value is emphasized. Without a magnetic field, the calculated PC gain of the single SnO2 NW device is ∼2500 at an intensity of 1 W m−2 under He–Cd laser (325 nm) illumination. We then calculate the gain achieved by different magnetic fields caused by ferromagnetic electrodes. Quite interestingly, the gain of the photodetector can reach up to almost 20 000 when the electrodes are magnetized in a field of 1400 G. Compared with the PC gain for the unmagnetized electrodes, the enhancement factor is estimated to be 800%. According to the data shown in Figure 3b, the highest enhancement factor can even exceed 2000%. This value is much larger than all of the previously reported enhancement factors for functionalized NW photodetectors, as summarized in Table 1.26, 27, 28 Therefore, it is very desirable to understand the mechanism enabling the ultrahigh PC gain obtained here.

Table 1 Comparison of the enhancement factor for 1D inorganic semiconductor NW-photodetectors obtained in the present work and previous reports
Full size table

According to the reports by Binet et al. and Muñoz et al.,30 the relationship between the PC gain and the illumination intensity is very informative when considering the underlying mechanism. As shown in Figure 3b, the gain plot (logarithmic) versus intensity shows that there is a turning point in each curve and that the value of the light intensity for the turning point to occur increases with increasing magnetization of the ferromagnetic electrodes. First, the existence of a critical value of turning point is considered. For the excitation intensity below the critical value, Γ versus intensity follows an inverse power law (that is, Γ ∝ I−κ), and the value of the exponent κ is ∼0.1–0.2. According to the theoretical work of the κ value reported previously, the behavior in this regime can be understood in terms of oxygen-related hole-trap filling mechanism.31 More explicitly stated, when the NW is exposed in air and adsorbs oxygen molecules at its surface, the adsorbed oxygen molecules will capture electrons from the NW and become negatively charged. This relationship can be described by the expression:

Under the light illumination, the photoinduced holes will migrate to the surface owing to the surface electric field and discharge negative-charged oxygen molecules, as described by the expression:

The neutralized oxygen molecules are photodesorbed from the surface. Therefore, the presence of hole-trap states will prolong the photoinduced electron lifetime, and thus, the gain can be amplified.

Once the excitation intensity exceeds the critical value, oxygen-related hole-trap states will be filled, which will change the electron-hole recombination behavior.13 At intensities exceeding the critical value of the turning point, the value of the exponent κ increases to a value of 0.6–0.8 as calculated from Figure 3b. Based on theoretical investigations,30, 31, 32, 33 the PC gain is now dominated by the modulation of surface space charge region (SCR).30 The existence of oxygen vacancies on the surface of an SnO2 NW causes the accumulation of free electrons on the NW surface.32 Consequently, the low-conductivity depletion region near the surface, referred to as SCR, is formed because of the existence of the upward band bending. Once electron–hole pairs are photogenerated, the photoinduced holes drift to the surface readily due to the built-in electric field, leaving unpaired electrons inside; thus, photogenerated electrons and holes are spatially separated. The spatial separation of electrons and holes reduces the electron-hole recombination rate; therefore, the electron lifetime is increased, and the photoresponse is enhanced. According to the simulation by Garrido et al.,33 the exponent κ of the PC induced by the modulation of surface SCRs is between 0.5 and 0.9, which is in good agreement with our experimental measurements.

Next, we discuss the gain behavior under the magnetic field caused by the magnetized Ni electrodes with the magnetization in the direction normal to the substrate plane. As the current flows through the conducting channel of NW, the electrons will migrate to the surface of NW due to the Lorentz force, and then, the amount of oxygen adsorbed by NW will be increased. Under this circumstance, not only is the oxygen-related hole-trapping effect is enhanced but also the surface band bending becomes more pronounced, as illustrated in Figure 4. Therefore, the pumping intensity that is needed to transfer the PC gain from the hole-trap filling to SCR mechanism will increase, and the critical value of the turning point therefore becomes larger when the strength of the magnetic field increases.

Figure 4
figure 4

Photocurrent gain mechanism of both oxygen-related hole trapping and surface band-bending effects with the assistance of magnetic field.

Full size image

To support the above interpretation, a more detailed examination of the behavior caused by Ni electrodes has been performed. Figure 5a shows the hysteresis curves of an Ni thin film exposed to an external magnetic field with the direction normal to the plane of the sample.34, 35 The hysteresis loops show that a larger remanence magnetization (mr) can be obtained with a higher maximum of external field (Bmax). As shown in Figure 5b, the relationship between mr and Bmax can be fit with a logarithmic function, that is, mr ∝ log (Bmax). We now examine the correlation of the decrease of dark current with increasing mr. In the model of oxygen molecules detection, the dependence of the resistance of NW (RNW) on the oxygen partial pressure (p(O2)) can be expressed as follows:36, 37

Figure 5
figure 5

(a) Hysteresis curves of nickel (Ni) thin film with the direction of external magnetic field perpendicular to the plane of the sample. (b) Remanence magnetization (mr) plot versus the maximum of external field (Bmax).

Full size image

In our case, the oxygen partial pressure surrounding the NW surface is proportional to the number of electrons accumulated on the NW surface, and the number of accumulated electrons is also proportional to the magnetic field generated by mr as described above. Thus, a simple relationship can be drawn as follows:

According to the previous report, the exponent n has a value of 1/4 at room temperature.37 As shown in Figure 6, the exponent n≈1/4 can be used to fit our experiment data quite well in the low mr region. However, in the high mr region, the exponent deviates from 1/4 very significantly, which can be explained as follows. The conductivity of NW is now dominated by the SCRs mechanism.33 The current difference (δI) with and without magnetic field can be expressed as follows:

Figure 6
figure 6

Resistance of nanowire (NW) (RNW) plot versus remanence magnetization (mr). RNW is proportional to at low mr region, and RNW is proportional tos at high mr region. Inset: the current difference with and without magnetic field (δI) is proportional to .

Full size image

where

ΔΨ0 is the intrinsic barrier height, ɛ is the permittivity, Nd is the doping level, VT=kT/q, A is Richardson constant=1.2 × 106 A m−2 K−2 and ΨB is the barrier height induced by the magnetic field with a linear relationship. Thus, the correlation between δI and mr can be expressed as . As shown in the inset of Figure 6, our measurements can be described well by the above expression. Furthermore, a more careful examination shows δI is equal to iwithout−imeasured=V/(Rwithout−RNW), where Vis the applied bias. Thus, the relationship between RNW and mr can be expressed as follows:

where C is a constant. The experimental data shown in Figure 6 indicates that the fitting curve is also in good agreement with the above theoretical prediction. Therefore, all of the novel properties observed in our measurements induced by the magnetized Ni electrodes can be well understood, including the decrease of dark current, the occurrence of the critical point as well as ultrahigh gain phenomenon.

Finally, an additional experiment of PC gain versus light intensity has also been performed for the as-grown NW device, as shown in Figure 7. The PC gain of the device without the assistance of magnetic field can reach up to 8800 due to more surface defects than that of the annealed NW device.38 With the effect of the magnetized Ni electrodes, the PC gain can be as high as 4.7 × 104, and the highest enhancement factor can also have a large value of 3000%. This result demonstrates that the ultrahigh gain phenomenon can also be obtained even if the pristine NWs have a higher state of surface defects. It is worth noting that a very high-gain value up to 107 on SnO2 NW photodetector has been reported previously.39 However, in that case, the ultrahigh gain phenomenon was observed with an electrode separation of ∼2 μm under an external bias of 1 V. Those experimental conditions are quite different from ours. In our study, the electrode separation is 8 μm, and the external bias is 0.1 V. For a fair comparison, we should examine the definition of PC gain in detail. The current gain can also be expressed as follows:13

Figure 7
figure 7

The gain logarithmic plot versus illumination intensity of as-grown single SnO2 nanowire exposed to magnetic field strengths of 350, 500, 750 and 1400 G.

Full size image

where τlife is excess carrier lifetime, t is the transit time of the carrier between electrodes, μ is carrier mobility, V is applied bias and l is the distance between electrodes. Based on the same experimental condition, the calculated gain value for our as-grown SnO2 NW can be up to 3.2 × 106. With the Ni electrodes, a gain value of up to 1.5 × 107 can be reached, which is comparable to the ultrahigh gain reported previously.39 Therefore, there is no doubt that an ultrahigh enhancement factor can be achieved in our devices under conditions similar to previous experiments.

Conclusion

In summary, ultrahigh gain in photodetectors based on single SnO2 NWs with ferromagnetic Ni electrodes has been demonstrated. The underlying mechanisms are attributed to both oxygen-related hole-trapping and surface band-bending effects induced by the magnetized ferromagnetic electrodes. The PC gain value can be enhanced by >2000% compared with that of the devices without magnetized electrodes. This large enhancement factor is much greater than all previously reported values for functionalized NW photodetectors. Our results demonstrate a very simple and feasible way to achieve ultrahigh gain photodetectors; these devices can be included in many other systems and represent a new avenue for the creation of high-efficiency optoelectronic devices based on the coupling between ferromagnetic materials and nanostructured semiconductors.