**1. Introduction**

Semiconductor-based ultraviolet (UV) photodiodes have been continuously developed that can be widely used in various commercial, civilian areas, and military applications, such as optical communications, missile launching detection, flame detection, UV radiation calibra‐ tion and monitoring, chemical and biological analysis, optical communications, and astro‐ nomical studies, etc. [1-2]. All these applications require very sensitive devices with high responsivity, fast response time, and good signal-to-noise ratio is common desirable charac‐ teristics. Currently, light detection in the UV spectral range still uses Si-based optical photo‐ diodes. Due to the Si-based photodiodes are sensitive to visible and infrared radiation, the responsivity in the UV region is still low [3-5]. To avoid these disadvantages, wide-bandgap materials (such as diamond, SiC, III-nitrides and wide-bandgap II–VI materials) are under intensive studies to improve the responsivity and stability of UV photodiodes, because of their intrinsic visible-blindness [6].

Among them, zinc oxide (ZnO) is another wide direct bandgap material due to its sensitive and UV photoresponse in the UV region [7-9]. ZnO has attracted attention as a promising ma‐ terial for optical devices, owing to its large direct band gap energy of 3.37 eV and a large ex‐ citon binding energy of 60 meV at room temperature compared to other II-VI semiconductors [10-12]. Therefore, ZnO is promising for use in light-emitting diodes (LEDs), laser diodes (LDs), ultraviolet (UV) detection devices [12-15]. Several deposition methods have been em‐ ployed for the growth of ZnO layers, including metal-organic chemical vapor deposition (MOCVD), molecular beam epitaxy (MBE), pulsed laser deposition (PLD), sol-gel and spray pyrolysis [16-20]. The synthesis of *p*-type ZnO films with acceptable stability and reproduci‐

© 2012 Chen; licensee InTech. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Chen; licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

bility by means of indium and nitrogen codoping or other group-III elements and nitrogen codoping has recently been demonstrated [21-24].

Since the quality of ZnO materials plays a key role in determining the performance of UV photodiodes. This chapter reviews the recent progress in Si-based heterostructure (UV) pho‐ todiodes, including *p*-ZnO/*n*-Si UV photodiodes, and *p*-ZnO/SiO2 ultrathin interlayer/*n*-Si UV photodiodes. Furthermore, the optoelectronic and the magneto-enhanced characteristics (so called magneto-optical multiplication effects) of UV photodiode placed in a strong mag‐ netic field were elucidated.

#### **2. ZnO/Si UV photodiodes**

Fabrication of a *p*-ZnO/*n*-Si heterojunction photodiode was reported [25]. An N-In codoped p-type was deposited on a (111)-oriented silicon substrate by ultrasonic spraying pyrolysis method. Three aqueous solution, Zn(CH3COO)2 2H2O (0.5 mol/l), CH3COONH4 (2.5 mol/l), and In(NO3)3 (0.5 mol/l), were as the source of zinc, nitrogen, and indium, respectively. The atomic ratio of Zn/N is 1:2 for N-doped film, and Zn/N/In is 1:2:0.15 for N-In codoped film [21]. The n-type Si (111) wafers were used as the substrates, which were etched with HCl for 5 min before deposition. The aerosol of precursor solution was generated by the commer‐ cial ultrasonic nebulizer. *P*-type N-In codoped ZnO films were obtained by heating the sub‐ strate to 650 C, and were subsequently studied by Hall measurement. The hole concentration and mobility of *p*-ZnO were around 1×1017 cm-3 and approximately 46 cm2 /V-s, respectively. *P*-ZnO/*n*-Si structures were then fabricated. The Ni/Au ohmic contact layer was evaporated onto the p-type ZnO film as the anode electrode, and a Ti/Pt/Au electrode was formed on the backside of the n-type Si substrate as the cathode electrode. Then, the cross section of the com‐ pleted structure is shown in Figure 1. The ZnO film with thickness of about 1.3 μm was formed on silicon substrate.

**Figure 1.** Schematic cross section of the completed structure. [25].

*<sup>R</sup>* <sup>=</sup> *<sup>I</sup> ph* / *Pinc* <sup>=</sup>*<sup>η</sup> <sup>q</sup>*

where *I ph* is the photocurrent, *P inc* is the incident power, and *η*, *q*, *ν* and *h* are the *QE*, the elec‐ tron charge, the frequency of incident light, and Planck's constant, respectively. Using Eq. (1), the values of responsivity and QE at 530 nm at biases of 1 V were 0.204 A/W and 47.73%, re‐ spectively. The values of responsivity and QE at 850 nm at biases of 1 V were 0.209 A/W and 30.49%, respectively. In contrast to conventional Si-based photodetectors, the ZnO film has been improved the responsivity in UV/blue region. However, the responsivity was degrad‐ ed in near infrared region (700-1100 nm). This result means that the portion of light with high‐ er energy, such as 400–500 nm, was absorbed by ZnO film and the portion of light with lower energy, such as 800-1000 nm, can completely incident into Si substrate and was absorbed. However, the responsivity owing to the ZnO film absorption occurring through the band-to-

*<sup>h</sup><sup>ν</sup>* (*<sup>A</sup>* / *<sup>W</sup>* ) (1)

Si-Based ZnO Ultraviolet Photodiodes http://dx.doi.org/10.5772/48825 197

Responsivity *R* is given by [27]

band did not observe in this work.

Figure 2(a) shows the plots of the *I-V* characteristics of the photodiodes measured in the dark (dark current) and under illumination (photocurrent, λ=530 nm) at reverse biases from 0 to 1 V. As shown in Fig. 2(a), it was found the photocurrent approximately 3.9×10-7 A and the dark current was approximately 8.87×10-9 A at a bias of 1 V. Therefore, it was found that a photocurrent to dark current contrast ratio is around two orders of magnitude. Figure 2(b) shows the plot of responsivity as a function of the wavelength for a *p*-ZnO/*n*-Si heterostruc‐ ture photodiode at a bias of 1 V. The photodiodes exhibited two higher responsive regions denoted as A and B, respectively. Region A at wavelength approximately from 400 nm to 700 nm was owing to ZnO film absorption occurring through the band-to deep level [26], and region B at wavelength approximately from 700 nm to 1000 nm was owing to Si sub‐ strate absorption occurring through the band edge.

**Figure 1.** Schematic cross section of the completed structure. [25].

Responsivity *R* is given by [27]

bility by means of indium and nitrogen codoping or other group-III elements and nitrogen

Since the quality of ZnO materials plays a key role in determining the performance of UV photodiodes. This chapter reviews the recent progress in Si-based heterostructure (UV) pho‐ todiodes, including *p*-ZnO/*n*-Si UV photodiodes, and *p*-ZnO/SiO2 ultrathin interlayer/*n*-Si UV photodiodes. Furthermore, the optoelectronic and the magneto-enhanced characteristics (so called magneto-optical multiplication effects) of UV photodiode placed in a strong mag‐

Fabrication of a *p*-ZnO/*n*-Si heterojunction photodiode was reported [25]. An N-In codoped p-type was deposited on a (111)-oriented silicon substrate by ultrasonic spraying pyrolysis method. Three aqueous solution, Zn(CH3COO)2 2H2O (0.5 mol/l), CH3COONH4 (2.5 mol/l), and In(NO3)3 (0.5 mol/l), were as the source of zinc, nitrogen, and indium, respectively. The atomic ratio of Zn/N is 1:2 for N-doped film, and Zn/N/In is 1:2:0.15 for N-In codoped film [21]. The n-type Si (111) wafers were used as the substrates, which were etched with HCl for 5 min before deposition. The aerosol of precursor solution was generated by the commer‐ cial ultrasonic nebulizer. *P*-type N-In codoped ZnO films were obtained by heating the sub‐ strate to 650 C, and were subsequently studied by Hall measurement. The hole concentration

*P*-ZnO/*n*-Si structures were then fabricated. The Ni/Au ohmic contact layer was evaporated onto the p-type ZnO film as the anode electrode, and a Ti/Pt/Au electrode was formed on the backside of the n-type Si substrate as the cathode electrode. Then, the cross section of the com‐ pleted structure is shown in Figure 1. The ZnO film with thickness of about 1.3 μm was formed

Figure 2(a) shows the plots of the *I-V* characteristics of the photodiodes measured in the dark (dark current) and under illumination (photocurrent, λ=530 nm) at reverse biases from 0 to 1 V. As shown in Fig. 2(a), it was found the photocurrent approximately 3.9×10-7 A and the dark current was approximately 8.87×10-9 A at a bias of 1 V. Therefore, it was found that a photocurrent to dark current contrast ratio is around two orders of magnitude. Figure 2(b) shows the plot of responsivity as a function of the wavelength for a *p*-ZnO/*n*-Si heterostruc‐ ture photodiode at a bias of 1 V. The photodiodes exhibited two higher responsive regions denoted as A and B, respectively. Region A at wavelength approximately from 400 nm to 700 nm was owing to ZnO film absorption occurring through the band-to deep level [26], and region B at wavelength approximately from 700 nm to 1000 nm was owing to Si sub‐

/V-s, respectively.

and mobility of *p*-ZnO were around 1×1017 cm-3 and approximately 46 cm2

strate absorption occurring through the band edge.

codoping has recently been demonstrated [21-24].

netic field were elucidated.

on silicon substrate.

**2. ZnO/Si UV photodiodes**

196 Photodiodes - From Fundamentals to Applications

$$R = I\_{ph} \;/\; P\_{inc} = \eta \frac{q}{h\nu} (A / W) \tag{1}$$

where *I ph* is the photocurrent, *P inc* is the incident power, and *η*, *q*, *ν* and *h* are the *QE*, the elec‐ tron charge, the frequency of incident light, and Planck's constant, respectively. Using Eq. (1), the values of responsivity and QE at 530 nm at biases of 1 V were 0.204 A/W and 47.73%, re‐ spectively. The values of responsivity and QE at 850 nm at biases of 1 V were 0.209 A/W and 30.49%, respectively. In contrast to conventional Si-based photodetectors, the ZnO film has been improved the responsivity in UV/blue region. However, the responsivity was degrad‐ ed in near infrared region (700-1100 nm). This result means that the portion of light with high‐ er energy, such as 400–500 nm, was absorbed by ZnO film and the portion of light with lower energy, such as 800-1000 nm, can completely incident into Si substrate and was absorbed. However, the responsivity owing to the ZnO film absorption occurring through the band-toband did not observe in this work.

nm thick NW core. The aspect ratios of the RNDs, which were calculated using the averaged values of the lengths and diameters, were ~10 and 30 for RND2 and RND6, respectively. A magnified image showing a ZnO/Si NW, in which the ZnO shells were partly peeled off during the sample preparation. A uniform thickness of ZnO over the Si core is observed. The yellow dashed lines indicate the position of the interface between ZnO and Si, as shown

Si-Based ZnO Ultraviolet Photodiodes http://dx.doi.org/10.5772/48825 199

Figure 4(a) shows the photoresponsivity spectra under a forward bias of 0.5 V. It is clear the UV responsivities of RND2 and RND6 are higher than that of the planar thin film diode (PD) under a forward bias. Such as compared to a PD, a RND2 (6 μm) resulted in a ~2.7 times en‐ hancement of the UV responsivity at λ=365 nm in the forward bias. In addition, the en‐ hanced UV photoconductive response in ZnO NWs may be attributed to the presence of oxygenrelated hole-trap states at the NW surface [29]. As a result, RNDs can improve the UV photodetection sensitivity due to the high surface area to volume ratio. In this case, the UV responsivities at λ=365 nm were detected to be 0.23, 0.42, and 0.63 A/W for PD, RND2, and RND6, respectively. Owing to the short penetration depth, the carrier generation normal‐ ly occurs near the surface. It indicates surface scattering and recombination decrease the car‐ rier lifetime. Figure 4(b) shows the photoresponsivity spectra of RNDs compared to the PD under a reverse bias. The values of the visible/UV responsivity at λ=700 nm and 365 nm were 17.2 A/W for RND6 and 0.86 A/W for PD. It appears that the ZnO surface can be depleted by the surface oxygen absorption according to the hole-trapping mechanism [29]. There‐ fore, both the UV and visible photoresponsivities of the RNDs were better than that of a pla‐ nar PD, owing to the enlarged surface area to volume ratio, efficient carrier collection, and

**Figure 4.** Photoresponsivity spectra of the RNDs and PD measured under (a) forward and (b) reverse biases. Their en‐

ergy band diagrams and charge transport mechanisms are also depicted in the insets [28].

in Figs. 3(b) and 3(c).

improved light absorption.

**Figure 2.** a) The dark and illuminated (λ=530 nm) I-V characteristics of the *p*-ZnO/*n*-Si heterostructure photodiode. (b) The responsivity as a function of the wavelength for a *p*-ZnO/*n*-Si heterostructure photodiode at a bias of 1 V [25].

**Figure 3.** a) Schematic showing the configuration of the photoresponse measurement system used for the *n*-ZnO (shell)/*p*-Si (core) radial nanowire photodiodes. (b) A typical cross-sectional SEM image of the *n*-ZnO/*p*-Si NW arrays. (c) A magnified image showing the bottom region of a ZnO/Si NW [28].

Kim et al. [28] were demonstrated utilizing radial heterojunction nanowire diodes (RNDs) array consisting of *p*-Si/*n*-ZnO NW core/shell structures which were fabricated using confor‐ mal coating by atomic layer deposition (ALD). Vertically dense Si NW arrays were prepared by Ag-induced electroless etching of *p*-type Si wafers. After formation of the Si NW arrays, the ALD technique was used to conformably coat a *n*-type ZnO thin film on the high aspect ratio Si NWs, as shown in figure 3(a). The properties of long (6 μm) and short (2 μm) nano‐ wire photodiodes, denoted as RND2 and RND6, respectively. The typical diameter of the *n*-ZnO/*p*-Si NW arrays was 350-400 nm, which consisted of a 100 nm thick shell and a 150-200 nm thick NW core. The aspect ratios of the RNDs, which were calculated using the averaged values of the lengths and diameters, were ~10 and 30 for RND2 and RND6, respectively. A magnified image showing a ZnO/Si NW, in which the ZnO shells were partly peeled off during the sample preparation. A uniform thickness of ZnO over the Si core is observed. The yellow dashed lines indicate the position of the interface between ZnO and Si, as shown in Figs. 3(b) and 3(c).

Figure 4(a) shows the photoresponsivity spectra under a forward bias of 0.5 V. It is clear the UV responsivities of RND2 and RND6 are higher than that of the planar thin film diode (PD) under a forward bias. Such as compared to a PD, a RND2 (6 μm) resulted in a ~2.7 times en‐ hancement of the UV responsivity at λ=365 nm in the forward bias. In addition, the en‐ hanced UV photoconductive response in ZnO NWs may be attributed to the presence of oxygenrelated hole-trap states at the NW surface [29]. As a result, RNDs can improve the UV photodetection sensitivity due to the high surface area to volume ratio. In this case, the UV responsivities at λ=365 nm were detected to be 0.23, 0.42, and 0.63 A/W for PD, RND2, and RND6, respectively. Owing to the short penetration depth, the carrier generation normal‐ ly occurs near the surface. It indicates surface scattering and recombination decrease the car‐ rier lifetime. Figure 4(b) shows the photoresponsivity spectra of RNDs compared to the PD under a reverse bias. The values of the visible/UV responsivity at λ=700 nm and 365 nm were 17.2 A/W for RND6 and 0.86 A/W for PD. It appears that the ZnO surface can be depleted by the surface oxygen absorption according to the hole-trapping mechanism [29]. There‐ fore, both the UV and visible photoresponsivities of the RNDs were better than that of a pla‐ nar PD, owing to the enlarged surface area to volume ratio, efficient carrier collection, and improved light absorption.

**Figure 2.** a) The dark and illuminated (λ=530 nm) I-V characteristics of the *p*-ZnO/*n*-Si heterostructure photodiode. (b) The responsivity as a function of the wavelength for a *p*-ZnO/*n*-Si heterostructure photodiode at a bias of 1 V [25].

**Figure 3.** a) Schematic showing the configuration of the photoresponse measurement system used for the *n*-ZnO (shell)/*p*-Si (core) radial nanowire photodiodes. (b) A typical cross-sectional SEM image of the *n*-ZnO/*p*-Si NW arrays.

Kim et al. [28] were demonstrated utilizing radial heterojunction nanowire diodes (RNDs) array consisting of *p*-Si/*n*-ZnO NW core/shell structures which were fabricated using confor‐ mal coating by atomic layer deposition (ALD). Vertically dense Si NW arrays were prepared by Ag-induced electroless etching of *p*-type Si wafers. After formation of the Si NW arrays, the ALD technique was used to conformably coat a *n*-type ZnO thin film on the high aspect ratio Si NWs, as shown in figure 3(a). The properties of long (6 μm) and short (2 μm) nano‐ wire photodiodes, denoted as RND2 and RND6, respectively. The typical diameter of the *n*-ZnO/*p*-Si NW arrays was 350-400 nm, which consisted of a 100 nm thick shell and a 150-200

(c) A magnified image showing the bottom region of a ZnO/Si NW [28].

198 Photodiodes - From Fundamentals to Applications

**Figure 4.** Photoresponsivity spectra of the RNDs and PD measured under (a) forward and (b) reverse biases. Their en‐ ergy band diagrams and charge transport mechanisms are also depicted in the insets [28].
