Abstract

The electrical characterization of p-Silicon (Si) and n-Zinc oxide (ZnO) nanorod heterojunction diode has been performed. ZnO nanorods were grown on p-Silicon substrate by the aqueous chemical growth (ACG) method. The SEM image revealed high density, vertically aligned hexagonal ZnO nanorods with an average height of about 1.2 μm. Electrical characterization of n-ZnO nanorods/p-Si heterojunction diode was done by current-voltage (I-V), capacitance-voltage (C-V), and conductance-voltage (G-V) measurements at room temperature. The heterojunction exhibited good electrical characteristics with diode-like rectifying behaviour with an ideality factor of 2.7, rectification factor of 52, and barrier height of 0.7 V. Energy band (EB) structure has been studied to investigate the factors responsible for small rectification factor. In order to investigate nonidealities, series resistance and distribution of interface state density (NSS) below the conduction band (CB) were extracted with the help of I-V and C-V and G-V measurements. The series resistances were found to be 0.70, 0.73, and 0.75 KΩ, and density distribution interface states from 8.38 × 1012 to 5.83 × 1011 eV−1 cm−2 were obtained from 0.01 eV to 0.55 eV below the conduction band.

1. Introduction

Zinc oxide (ZnO) is a promising and environment friendly semiconductor with a large exciton binding energy (60 meV) and a direct and wideband gap (3.37 eV). It has superior physical properties such as high breakdown electric field, high electron saturation velocity; radiation tolerance, and thermal conductivity [1], which enable it for making high-power and high-temperature devices. Being a transparent conducting oxide, it has a potential to replace indium tin oxide (ITO) in photovoltaics. Short-wavelength light emitters and detectors working at near-ultraviolet (UV) wavelengths can be made from ZnO.

Electronic-grade bulk, thin film, and nanostructured ZnO can be grown easily and cheaply on a variety of substrates by different growth or deposition methods along with the tailoring of band gap [25]. Homojunctions of ZnO are difficult to realize because reproducible p-type with desired hole concentration without diminution is still under research [6, 7]. However, heterojunctions of ZnO are realized with various semiconductors most commonly Si, GaN, and SiC for the fabrication of LEDs, Photodetectors, solar cells, biosensors, etc. [811].

For the realization of ZnO heterojunction, silicon is the most attractive choice because of its low cost and large area wafers. Therefore, fabrication of Si/ZnO heterojunction by different techniques has been a topic of interest, and several studies have been reported on their characterizations and device fabrication for potential applications [1215].

In the present work, we are reporting the growth, fabrication, and characterization results of p-Si/n-ZnO nanorod heterojunction diodes. The structural, optical, and electrical characteristics are studied by scanning electron microscopy (SEM), photoluminescence (PL), current-voltage (I-V), and capacitance-voltage (C-V) measurements. Electrical parameters including barrier height (φB), ideality factor (n), series resistance (Rs), doping density (Nd), and interface states density (NSS) are also reported.

2. Materials and Methods

ZnO nanorods have been grown on p-Si substrates of (100) orientation with doping of 1016 cm−3 and 1.38 Ω cm conductivity. Prior to growth, the substrates were ultrasonically cleaned for 15 minutes with each acetone and methanol and then rinsed with deionized water sequentially to remove contamination and dust from the surface.

ZnO nanorods were grown by the renowned low-temperature aqueous chemical growth (ACG) method [1619]. Initially, a seed layer solution was prepared by diluting zinc acetate dehydrate in methanol. Few droplets of seed solution were spin-coated on the substrates. The coating step was repeated three times, and then samples were heated at 250°C in air for 20 min to yield layers of ZnO on the substrate. For the preparation of solution, analytic reagent grade hexamethylenetetramine (HMT) (C6H12N4) and zinc nitrate (Zn(NO3)26H2O) 0.022–0.075 mM were used without further purification. The solution was prepared with an equimolar concentration of zinc nitrate and hexamethylenetetramine by dissolving in 200 ml of deionized water at room temperature. In order to get a homogeneous solution it was stirred at 60°C for 30 minutes by magnetic stirring for complete mixing. The samples were then placed in solution with seed layer faced down with a certain angle. The container of solution was sealed with an aluminium foil and heated at 96°C for 4 hours. After growth process, the samples were washed with deionized water for removing organic salts and blow-dried with nitrogen.

For the fabrication of heterojunction, an insulating layer of polymethyl methacrylate (PMMA) was spin-coated on the samples to fill the gaps between the nanorods. PMMA layer prevents the carrier cross-talk among the nanorods. It also isolates the electrical contacts, deposited at the top of nanorods, from reaching the Si substrate. Then, oxygen plasma cleaning was performed to remove excess PPMA from the top of the nanorods. About 150 nm thick ohmic contact of aluminium (Al) has been evaporated in the vacuum chamber on p-Si substrate. Al/Pt nonalloyed circular contacts were then evaporated to form ohmic contacts to n-ZnO nanorods. The thickness of Pt/Al contacts was 50/60 nm with a diameter of 0.58 mm and specific contact resistance of 1.2 × 10−5 Ω-cm−2. ZnO nanorods’ growth and device fabrication steps are shown in Figure 1.

The structural characteristics of ZnO nanorods were studied by using the JEOLJSM-6301F scanning electron microscope (SEM). Current-voltage (I-V) and capacitance-voltage (C-V) measurements have been performed by using Keithley SCS-4200 by placing the sample on a probe station.

3. Results and Discussion

The morphology and size distribution of ZnO nanorods were studied by SEM. Vertically aligned, dense array of hexagon-shaped nanorods with a mean diameter of 180–300 nm and approximate height of 1.4 μm are revealed as shown in Figure 2. Although the nanorods were not perfectly aligned on the substrate, they had a tendency to grow vertical on the substrate with an almost uniform distribution. The standard aspect ratio (SAR) of 4.66–7.77 has been obtained by dividing the length of rods with its diameter which is in agreement with reported values for hydrothermally grown ZnO nanorods [20, 21].

Current-voltage (I-V) characteristics of the Si/ZnO pn heterojunction have been studied at room temperature for a voltage range from–10 to +10 V as shown in Figure 3; inset shows the schematic of Si/ZnO heterojunction.

These characteristics show a nonlinear rectifying behaviour with turn-on voltage of 0.6 V. The current rectification factor (IForward/IReverse) was found to be 52 measured at ± 5 V. The values of ideality factor (n) and barrier height (ϕΒ) are extracted by using the Shockley equation based on the thermionic emission model given by equations (1) and (2).where A is the area (10.86 × 10−3 cm2), Io is the saturation current, kB is Boltzmann’s constant, and A is the Richardson constant having a value of 32 A/cm2 K2 for n-ZnO. The Richardson constant is extracted from A = 120 ×  m/mo A/cm2K2, where m is the effective mass for ZnO taken as 0.275mo for electrons [17].

The values of ideality factor and barrier height obtained are 2.7 and 0.70 V, respectively. A high value of ideality factor is attributed to series resistance and presence of interface sates.

To study the charge transport mechanism of Si/ZnO heterojunction, the logarithmic plot of current-voltage (Log I-Log V) under forward bias has been studied.

For our heterojunction, the LogI–LogV plot is having three regions as shown in Figure 4. Region-I is low forward voltage (V < 0.6 V) with a linear (IV) current-voltage relationship representing an ohmic current transport mechanism. This implies that tunnelling may be responsible for charge transport. The small value of current is due to limited carrier injection from the electrodes into the semiconductor under low biases.

Region-II is for moderate forward voltages (0.8 V < V < 3.4 V). For this region, an exponential increment in current is observed with an increase in the voltage following the relation IeKV, where K is a constant known as injection efficiency constant. The higher values of K indicate higher carrier injection [15]. For our heterojunction, the value of K is 1.02 V−1 obtained by fitting the I-V curve in region-II. The small value of K indicates that the thermal emission of carriers induced by surface states and traps is very low [17]. This implies that charge transport involves recombination tunnelling mechanism which is usually observed in wide band gap heterojunction.

Region-III is the high junction voltage region (3.4 V < V < 10 V). In this region, current-voltage characteristics was found to follow power law (IVL), where L is a constant, and for our diodes, the value of L is L∼1.5. The value of L less than 2 ascribes that the current is attributed to the trap charge limiting current (TCLC) which is associated with the distribution of trapped charges in a junction. This behaviour is usually observed in materials with low free charge carrier concentration [2225].

The capacitance-voltage (C-V) characteristics measured at the frequency of 1 MHz at room temperature are shown in Figure 5. By plotting the inverse squared of area per junction capacitance (A2/C2) against the applied reverse voltage (V), the built-in potential (Vbi), barrier height (), and doping concentration (Nd) can be extracted. The built-in potential is obtained by linear extrapolation of A2/C2 to the voltage axis (Figure 5). The intercept with the applied voltage axis yields a built-in potential of 0.8 V. The doping concentration of 1.3 × 1013cm3 has been extracted from the slope of the linear segment of the curve by using equation (3) where εS is the relative permittivity of ZnO having a value of 8.2. The barrier height of 1.12 V is obtained by using the following equation:where NC is the effective density of states in the conduction band (CB). Its value is 3.5 × 1018/cm3 for ZnO, obtained from the following equation:where mnI the effective mass of electrons given by . The barrier height obtained from the C-V measurement is comparatively larger than the barrier heights obtained from the I-V measurements. It is due to barrier inhomogenities, image force, and surface defects [26]. Secondly, interface traps do not respond to applied AC signals sometimes and do not contribute to capacitance at higher frequencies [27, 28].

At high forward voltages, the I-V characteristics of devices are deviated from linearity due to the presence of series resistance and interface states. The current-voltage (I-V) characteristics based on thermionic emission (TE) are given by the following equation:where RS is the series resistance which plays a crucial role and affects the device behaviour. We extracted series resistance by the method proposed by chueng [24, 29]given by the following equation:

The plot of dV/d (lnI) vs. current is shown in Figure 6. On linearly fitting the curve, we obtained series resistance from the slope of the curve and ideality factor from its y-intercept. The obtained values of series resistance and ideality factor are 0.75 KΩ and 5.7, respectively.

Furthermore, the barrier height and series resistance are also extracted by plotting function H (I) vs. current given by the following equations:

The plot of H (I) vs. current is shown in Figure 7. It is observed that H (I) increases linearly with increasing current. The value of series resistance obtained from the slope of the curve is 0.73 KΩ, and barrier height is calculated from y-intercept using the ideality factor obtained by equation (6) and is found to be 0.70. The values of series resistance calculated from the plot of H (I) vs. current is close to the value obtained from the plot of dV/d (lnI) versus current, implying the consistency of cheung’s method. Series resistance extracted by the direct method (ΔV/ΔI) from current-voltage characteristics is 0.77 KΩ.

The electrical parameters extracted for p-Si/n-ZnO heterojunction have been compared with the literature as shown in Table 1. The values of series resistances are 0.77 KΩ, 0.73 KΩ, and 0.75 KΩ obtained by different methods for our heterojunction. These values are on higher side however less than the values of 3.2 KΩ and 2.69/2.70 KΩ reported by [30, 33], respectively. Barrier height and ideality factor are in agreement with reported values. High values of ideality factor obtained for Si/ZnO heterojunction diode are attributed to barrier inhomogeneity, series resistance, and presence of defects at the junction. The doping concentration for our junction is lower than the reported values in the literature. For ZnO, this unintentional n-type doping is attributed to oxygen vacancies which act as donor-like native point defects. A small value of Nd may be due to less number of oxygen vacancies. High series resistance and small doping concentration are the limiting factors for majority carrier’s injection under forward bias and reducing the rectification factor.

The smaller values of rectification factor and turn-on voltage are sometimes attributed to the small value of conduction band (CB) offset of Si and ZnO [28, 34, 35]. Therefore, energy band (EB) diagrams have been studied. The energy band diagrams of for Si/ZnO heterojunction are made according to the Anderson rule. EB structure for isolated regions before making the contact is shown in Figure 8(a) which depicts that p-Si and n-ZnO form type-II staggered gap heterojunction. Both conduction and valence bands have band offset originated due to difference in electron affinities and band gaps.

The EB diagram at equilibrium is shown in Figure 8(b). The CB offset (ΔEC) of 0.15 eV is obtained by applying the Anderson rule given by ΔEC = qχZnO-qχSi, where the electron affinities of ZnO (qχZnO) and Si (qχSi) are 4.2 eV and 4.05 eV, respectively [36, 37]. Due to the small value (0.15 eV), there might be a flow of electrons from the conduction band of Si to the conduction band of ZnO. But in the valence band, the larger value of valence band offset (ΔEV) prevents the movement of holes from ZnO to Si, where ΔEV = (qχZnO + EgZnO)–(qχSi + EgSi) = ΔEC + ΔEg = 2.4 eV. Here, the band gaps of ZnO (EgZnO) and Si (EgSi) are 3.37 eV and 1.12 eV, respectively [38]. The position of Fermi level for ZnO is obtained by using equation (10), where Nd is the donor concentration obtained from A2/C2 vs. V plot:

As the carrier concentration in p-Si is higher than ZnO, the depletion region will be primarily extended in ZnO. Thus, the current transport in the heterojunction is considered predominantly by the flow of electrons from n-ZnO to p-Si [39]. Therefore, when the heterojunction is forward biased, there will be flow of electrons from ZnO into Si due to the lowering of junction barrier as shown in energy band diagram in Figure 8(c). The current increases with the increase in forward bias due to injection and recombination of electrons in p-Si. However, diode current tends to become constant at higher forward bias due to series resistance. A nominal increase in the reverse current will take place with increase in reverse voltage due to increase junction barrier as shown in Figure 8(d) of reverse bias.

The quality of the interface and density of interface states play a critical role in semiconductor devices. Several device parameters such as barrier height, ideality factor, and rectification factor get affected due to the presence of interface states [40]. The recombination in these thin insulating interfacial layers is sometimes the main loss mechanism in heterojunction especially in heterojunction solar cells.

The origin of these interface states are exactly not known; however, they are introduced intentionally at the interface due to extended air exposure, chemical reaction during fabrication processes, and incomplete covalent bonds at the interface [41, 42]. Such defects produce a large density of interface states distributed in the energy and also cause leakage currents.

The density of interface states (NSS) can be extracted using I-V and C-V measured values using the following equation [43]:where εi and εS are the permittivity of insulator (oxide) and semiconductor having values of 3.9 and 8.2 for SiO2 and ZnO, respectively. Here, the values of εi/δ and n (V) are extracted using the following equations:where n () is the voltage-dependent ideality factor, δ is the thickness of oxide layer, and is its capacitance, obtained from the following equation:

The energy distribution of interface states below the conduction band is obtained by using the following equation:where is the effective barrier height which can be obtained from equation (16).

The distribution of interface states below conduction band is shown in Figure 9. The interface state density is found high near the conduction band edge (EC) and decreased monotonically with energy approaching the centre of the energy gap. The interface states density decreased from 8.38 × 1012 to 5.83 × 1011 eV−1 cm−2 and from Ec −0.01 to EC—0.55 eV below EC.

The distribution of densities of interface states (Nss) below the conduction band (Ec) is compared with the literature as shown in Table 2. The extracted density for our heterojunction is having a moderate range of values as compared with the reported values. As the interface states play an imperative role in electronic devices, they must be low in order to reduce the recombination and improve the junction performance.

4. Conclusion

Electrical characterization of p-Si/n-ZnO nanorod heterojunction diode has been performed. The values of barrier height, ideality factor, rectification factor, and built-in potential have been extracted. For heterojunction, low values of rectification factor are attributed to small CB offset; therefore, the energy band diagram of the heterojunction has been studied and a small value of (0.15 eV) CB offset was observed. Series resistance and interface states are attributed to the high values of ideality factor; therefore, series resistance and densities of interface states have been extracted. The values of series resistance were found to be 0.75, 0.73, and 0.77 KΩ extracted by different methods. By I-V and C-V, the density of interface states from 8.38 × 1012–5.83 × 1011 eV−1 cm−2 and from Ec − 0.01 to EC − 0.55 eV below EC was obtained. The value of interface states must be kept low in order to lessen the surface recombination and tunnelling.

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The authors declare that there are no conflicts of interest.

Acknowledgments

The authors appreciate the support and facilities by the Sultan Qaboos Oman IT chair office and Electronic Design Center at the NED University of Engineering and Technology.