Photoconductivities in monocrystalline layered V2O5 nanowires grown by physical vapor deposition
© Chen et al.; licensee Springer. 2013
Received: 7 September 2013
Accepted: 12 October 2013
Published: 25 October 2013
Skip to main content
© Chen et al.; licensee Springer. 2013
Received: 7 September 2013
Accepted: 12 October 2013
Published: 25 October 2013
Photoconductivities of monocrystalline vanadium pentoxide (V2O5) nanowires (NWs) with layered orthorhombic structure grown by physical vapor deposition (PVD) have been investigated from the points of view of device and material. Optimal responsivity and gain for single-NW photodetector are at 7,900 A W-1 and 30,000, respectively. Intrinsic photoconduction (PC) efficiency (i.e., normalized gain) of the PVD-grown V2O5 NWs is two orders of magnitude higher than that of the V2O5 counterpart prepared by hydrothermal approach. In addition, bulk and surface-controlled PC mechanisms have been observed respectively by above- and below-bandgap excitations. The coexistence of hole trapping and oxygen sensitization effects in this layered V2O5 nanostructure is proposed, which is different from conventional metal oxide systems, such as ZnO, SnO2, TiO2, and WO3.
Vanadium pentoxide (V2O5) is the most stable crystallization form and is also the most applicable in the industry among vanadium oxide systems such as VO, VO2, and V2O3. The orthorhombic layered structure of V2O5 promises a high ionic storage capacity for energy storage applications. Recently, its quasi-one-dimensional nanostructures such as nanowires (NWs), nanobelts (NBs), and nanotubes have gained substantial attention. Due to high surface-to-volume ratio and high surface activity, V2O5 1D structures for various applications, such as field emitters[2–5], transistors[6, 7], chemical sensors[8–10], and lithium batteries[11–14], have been developed.
In addition, V2O5 with a direct optical bandgap at visible-light region (Eg = 2.2 to 2.7 eV)[2, 15–18] also inspires the studies of optoelectronic applications such as photodetection[2, 19], optical waveguide, and high-speed photoelectric switch. Although device performance of the individual NW has been demonstrated in several studies, fundamental photoconduction (PC) properties and their corresponding surface effects were less studied than the known hopping transport[6, 21–24]. The potential difference of the transport properties of nanomaterials grown by different approaches was also less known. In this paper, we report the study of photoconductivities of V2O5 NWs grown by physical vapor deposition (PVD). The performance of the single-NW device and intrinsic PC efficiency of the material have been defined and discussed. The results are also compared with the reported data of the V2O5 counterpart synthesized by hydrothermal approach. The probable PC mechanisms that originated from the bulk and surface under above- and below-bandgap excitations are also proposed.
V2O5 NWs were grown by PVD using high-purity V2O5 powder as the source material and mixed O2/Ar as the carrier gas. The growth temperature was 550°C, and the pressure was 0.3 Torr. The details of material growth can be found in our earlier publications[25, 26]. The morphology, structure, and crystalline quality of the as-grown V2O5 NWs were characterized by field-emission scanning electron microscopy (FESEM), X-ray diffraction (XRD), Raman spectroscopy, high-resolution transmission electron microscopy (HRTEM), and selected-area electron diffraction (SAD). Electrical contacts of the two-terminal single-NW devices were fabricated by focused ion beam (FIB; FEI Quanta 3D FEG, FEI Company, Hillsboro, OR, USA) deposition using platinum (Pt) as the metal electrode. Individual NWs were dispersed on the insulating Si3N4/n-Si or SiO2/n-Si template with pre-patterned Ti/Au microelectrodes prior to FIB deposition. Electrical measurements were carried out on an ultralow-current leakage cryogenic probe station (TTP4, LakeShore Cryotronics, Inc., Westerville, OH, USA). A semiconductor characterization system (4200-SCS, Keithley Instruments Inc., Cleveland, OH, USA) was utilized to source dc bias and measure current. He-Cd gas laser and diode laser were used to source excitation lights with wavelengths (λ) at 325 and 808 nm for the PC measurements, respectively. The incident power of laser was measured by a calibrated power meter (Ophir Nova II, Ophir Optronics, Jerusalem, Israel) with a silicon photodiode head (Ophir PD300-UV). A UV holographic diffuser was used to broaden laser beam size (approximately 20 mm2) to minimize error in power density calculation.
Electrical contacts of single V2O5 NW devices were examined by dark current versus applied bias (id-V) measurements. Figure 1f depicts typical id-V curves measured at room temperature of 300 K for the V2O5 NW with d at 400 ± 50 nm and the inter-distance between two contact electrodes (l) at 7.3 μm. A representative FESEM image of the individual V2O5 NW device is also shown in the inset of Figure 1f. The id-V curve reveals a linear relationship, indicating the ohmic contact condition of the NW device. Room temperature conductivity (σ) was estimated at 13 ± 3 Ω-1 cm-1. A similar σ can be reproduced from the other samples with a d range of 200 to 800 nm. The σ level is more than one order of magnitude higher than that (σ = 0.15 to 0.5 Ω-1 cm-1) of individual V2O5 NWs in previous reports in which small polaron hopping is attributed to the transport mechanism[23, 24].
where PNW is the incident optical power on the projected area (A) of the measured NW and can be calculated as PNW= IA = Idl]. The calculated R versus I result according to the measured ip values in Figure 2b is depicted in Figure 2c. The result shows that R increases from 360 to 7,900 A W-1 gradually and saturates at a near-constant level while intensity decreases from 510 to 1 W m-2. While comparing the optimal R with that of earlier reports, the value at 7,900 A W-1 is over one order of magnitude higher than that (R ~ 482 A W-1) of V2O5 NWs synthesized by hydrothermal approach. Even if the comparison is made at similar power densities in the range 20 to 30 W m-2, the PVD-grown V2O5 NW still exhibits higher R at approximately 2,600 than the reference data by a factor of 5. In addition, compared to other nanostructured semiconductor photodetectors, the R of the V2O5 NW device is higher than those of ZnS NBs (R ~ 0.12 A W-1), ZnSe NBs (R ~ 0.12 A W-1), ZnO nanospheres (R ~ 14 A W-1), and Nb2O5 NBs (R ~ 15 AW-1) and is lower than those of GaN NWs (R ~ 106 A W-1) and ZnS/ZnO biaxial NBs (R = 5 × 105 A W-1).
where E is the photon energy, e is the elementary electron charge, and η is the quantum efficiency. To simplify the calculation, the η is assumed to be unity. The calculated Γ versus I result is also plotted in Figure 2c. The maximal Γ of this work at approximately 3 × 104 is also over one order of magnitude higher than that (Γ = 1328) of the hydrothermal-synthesized V2O5 NWs. Compared with other nanostructured semiconductor devices, the Γ of the V2O5 NW is higher than those of ZnS NBs (Γ ~ 0.5 A W-1), ZnSe NBs (Γ ~ 0.4 A W-1), ZnO nanospheres (Γ ~ 5 A W-1), Nb2O5 NBs (Γ ~ 6 A W-1), and WO3 NWs (Γ ~ 5×103 A W-1) and is lower than those of ZnO NWs (Γ ~ 2 × 108 A W-1), SnO2 NWs (Γ ~ 9 × 107 A W-1), GaN NWs (Γ ~ 106 A W-1), and ZnS/ZnO biaxial NBs (Γ = 2 × 106 A W-1).
In addition, the power-dependent behavior of R (or Γ) could imply the potential hole trapping PC mechanism. The unintentionally doped V2O5 semiconductor has been confirmed to exhibit n-type conducting[6, 22, 39]. Under low power density, the photoexcited holes are totally captured by certain defects which function as a hole trap. The hole trapping effect leaves unpaired electrons which exhibit a long lifetime (τ). As photocurrent is linearly dependent on carrier lifetime, i.e., ip ∝ τ, the long-lifetime electron will substantially enhance and dominate the photocurrent generation. As the τ of electron which is decided by the hole trapping time is now a constant, R (or Γ) will be independent of the excitation power, i.e., R (or Γ) = const. Once the power exceeds a critical value (trap filling intensity), the photogenerated hole density is much higher than the trap density and the traps will be fully occupied. Under this condition, the trapping effect can be ignored and photocarriers will follow the bimolecular recombination mechanism[40–42]. The recombination after trap filling results in the decrease of τ with the increase of I, making an intensity-dependent R (or Γ) following an inverse power law, i.e., R (or Γ) ∝ I-k, where the theoretical k = 1/2. The aforementioned model agrees with the two-stage power-dependent R (or Γ) result in Figure 2c and ip in Figure 2b. The trap filling intensity is roughly at 5 W m-2, and the fitted k value is 0.62 ± 0.04 for the V2O5 NWs.
According to literature reports, the photoconductivity of metal oxide semiconductor NWs, such as ZnO, SnO2, TiO2, and WO3, mostly follow a common oxygen-sensitized (OS) PC mechanism[36, 37, 43–45]. The mechanism is controlled by the interaction of foreign oxygen molecule and semiconductor in the near surface area. According to the OS model, the PC process includes four steps: (1) In the dark and in the atmospheric ambience, as oxygen plays a role of electron trap state in the metal oxide semiconductor surface, through oxygen adsorption, the electron is captured on the surface and creates negatively charged surface states (or oxygen ions) [O2(g) + e- → O2-(ad)]. The effect induces an enhanced upward bending of the energy band at the surface. (2) Under light illumination, electron–hole pairs are generated [hυ → e- + h+] and (3) subsequently separated by the surface electric field or band bending. (4) The excess holes are attracted by the surface and recombine with negative-charged oxygen on the surface [h+ + O2-(ad) → O2(g)]. The result leaves unpaired electrons with prolonged lifetimes, which is similar to the hole trapping effect in the bulk. Recombination can only take place when oxygen molecules re-adsorb on the surface as that in step 1.
Although the photoconductivity of the V2O5 NWs has been confirmed to be dominated by the bulk under band-to-band (λ = 325 nm) excitation, the sub-bandgap excitation using the 808-nm wavelength (E = 1.53 eV) was also carried out to further characterize the layered 1D nanostructure. Figure 4b depicts the photoresponses under the sub-bandgap light illumination at different I and at V = 0.1 V in air and vacuum ambiences for the V2O5 NW with d = 800 nm and l = 2.5 μm. As the values of photoresponse at sub-bandgap excitation are much less than the inter-bandgap excitation, the I of the 808-nm wavelength was operated at a relatively high range of 408 to 4,080 W m-2. Under high-power condition, the sub-bandgap excitation generates an observable photoresponse and the ip is linearly dependent on I. The ip versus I curves in air and vacuum ambiences are also plotted in the inset of Figure 4b. The monotonic linear dependence of ip and I is different from the two-stage power dependence for the band-to-band excitation in Figure 2b, implying the different PC mechanisms. The response time at a few seconds by 808-nm excitation is also much faster than that (τ > 100 s) by 325-nm excitation.
where v is the carrier drift velocity which is equal to the product of mobility (μ) and applied electric field (F), i.e., v = μF, where, Γ can be rewritten as. Accordingly, Γ depends on l and V. In terms of engineering application, photodetectors can be operated with high Γ by shortening l and increasing V. However, to objectively compare the intrinsic PC efficiency of the materials, the artificial factors have to be excluded.
The calculated Γn versus I using the data of Γ (Figure 2c) or ip (Figure 2b) for the V2O5 NW measured at V = 0.1 V under 325-nm (E = 3.82 eV) and 808-nm (E = 1.53 eV) illuminations are illustrated in Figure 4c. One data point of hydrothermal-synthesized V2O5 NWs calculated according to the data in (E = 2.76 eV) is also plotted for comparison.
After excluding the artificial contributions of l and V, the Γn of our PVD-grown V2O5 NWs at approximately 6 × 10-3 cm2 V-1 is two orders of magnitude higher than that (Γn ~ 5 × 10-5 cm2V-1) of the hydrothermal-synthesized ones for the similar I = 25 ± 5 W m-2. This result indicates the PVD-grown NWs exhibit a higher efficiency for photocarrier transport and photocurrent generation than the hydrothermal ones. The PVD (or thermal evaporation) approach usually provides better control for crystal growth, and the growth temperature at 550°C is also relatively high in comparison with that in the hydrothermal method (synthesis at 205°C). Accordingly, it is inferred that the higher PC efficiency (or Γn) originated from a higher crystalline quality in this PVD-grown V2O5 nanostructure.
In addition, Figure 4c also shows that the Γn at 325-nm excitation is also much higher than that at 808-nm excitation. The optimal (saturation) Γn at λ = 325 nm is 1.7 ± 0.2 × 10-2 cm2 V-1 which is over three orders of magnitude higher than that (Γn = 4.7 ± 0.6 × 10-6 cm2 V-1) at λ = 808 nm in air ambience. The Γn enhanced in the vacuum can also be observed therein. The analysis quantitatively demonstrates the difference of PC efficiency induced by above- and below-bandgap excitations. As Γn linearly depends on η and τ and the volume for optical absorption (or η) of the bulk by inter-bandgap excitation is much higher than that of the surface under sub-bandgap excitation, it is proposed that η plays an important role on the Γn difference for the wavelength-dependent PC. The relatively long photoresponse time (or τ) could also contribute to the higher Γn under inter-bandgap (325 nm) excitation.
Finally, it is noted that the PC mechanism based on the small polaron hopping transport has been proposed by Lu et al.. The very short lifetimes in the range of 1 to 1,000 μs are usually one of the criteria to manifest the polaron hopping mechanism. However, the typical lifetimes in this study either under 325 or under 808 excitation are at the orders of magnitude from seconds to hundred seconds, which is at least three orders of magnitude higher than the relaxation time of small polaron. The substantial difference could allow us to explain the PC mechanism on the basis of the conventional band conduction model (as shown in Figure 5) for monocrystalline semiconductors. The free electron-dominant conduction mechanism could also offer a probable explanation for the relatively higher σ in the PVD-grown V2O5 NWs in comparison with the literature data of which hopping is the dominant factor for charge conduction[23, 24].
Photoconductivities of the PVD-grown V2O5 NWs with monocrystalline orthorhombic structure have been investigated. In addition to the device performance, the PVD-grown V2O5 NWs exhibit two orders of magnitude higher PC efficiency (or Γn) than their hydrothermal-synthesized counterparts. In addition, the PC mechanism has also been studied by the power, environment, and wavelength-dependent measurements. Both the bulk-controlled (hole trapping effect) and surface-controlled (oxygen-sensitization effect) PC mechanisms have been observed under above- and below-bandgap excitations, respectively. Understanding of the transport properties in this layered V2O5 1D nanostructure could enable us to design the electronic, optoelectronic, and electrochemical devices by a more efficient way.
Ruei-San Chen would like to thank the financial support of the Taiwan National Science Council (grant nos. NSC 99-2112-M-011-001-MY3 and NSC 99-2738-M-011-001) and the National Taiwan University of Science and Technology (NTUST).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.