3.1. Evaluation of As-Deposited Electrodes
In our previous study, Ag was used as the upper electrode material for the Ag2O/β-Ga2O3-based p-n junction photodetector. This study confirmed that the surface condition of the electrode film has a significant influence on the performance of the photodetector. Therefore, as a further study, we tried to fabricate an electrode with an improved surface and apply it to the device. Therefore, a thin film was made by co-sputtering Ag and AZO, and the structural, optical, and electrical properties of the as-deposited thin film on the glass substrate were evaluated.
The properties of the electrodes were evaluated prior to deposition on the Ag
2O/β-Ga
2O
3 photodetector. Surface morphology in thin films analyzed by SEM are described in
Figure 1a–e and
Figures S1 and S2. The SEM images shown in
Figure 1a–d are 20–50 nm thick Ag:AZO films deposited by co-sputtering with Ag and AZO. The films exhibited smooth and homogeneous surfaces, and grain boundaries increased with increasing thickness. The uniformity of the surface improves photosensitivity when used as a layer of a photodiode device by reducing the scattering and reflectivity of the irradiated light.
Figure 1e is an energy-dispersive X-ray spectrometer (EDX) image of a 20 nm sample deposited at 50 W, showing the distribution and atomic percentages of Ag, Al, and ZnO present in the thin film. Since AZO is an oxide, it has a lower deposition rate than Ag, and it is inferred that more Ag is contained in the component table of the as-fabricated sample.
Figures S1 and S2 show the surface SEM images according to the thickness of Ag and AZO thin films, respectively. It was observed to have a homogeneous surface like that of the Ag:AZO thin film.
The crystallographic characteristics of the as-fabricated Ag:AZO film for each thickness were evaluated by analyzing the XRD patterns in
Figure 2a. The Ag:AZO samples had polycrystalline structures, and both Ag and AZO peaks were observed. Ag peaks at 2θ = 38.12, 64.4, and 77.5° corresponded to the (111), (220), and (311) planes, respectively, and the AZO peak at 2θ = 44.8° corresponded to the (400) plane (Ag; ICDD card 01-087-0720, AZO; ICDD card 01-071-0968). There was no significant change based on thickness; however, compared with the peak intensity of the AZO electrode in
Figure S3, the (400) peak intensity was lower because the AZO content in the Ag:AZO sample was 5.6 times lower than that of Ag, as we inferred from the EDX results.
The crystallite size was calculated using the Scherrer equation based on the (111) plane, which is the preferential growth plane of the XRD pattern of Ag:AZO, and it was inferred that the crystallite size depends on the thickness. The Scherrer equation is as follows [
36,
37]:
In Equation (1), 𝜏 is the crystallite size,
θ is the Bragg angle,
K is the Scherrer constant, and 𝛽 is the full width at half maximum (FWHM) inversely proportional to the crystallite size 𝜏. Therefore, the larger the FWHM, that is, the wider the XRD pattern, the smaller the crystallite size. The crystallite size according to the thickness of the thin films is shown in
Figure 2b. As the thickness increased, so did the crystallite size, and it was inferred that Ag:AZO, which had the smallest FWHM, had the largest crystallite size. Crystallite size increases as the density of a material increases; hence, a flat surface would be made of dense crystallites.
The low roughness of the thin film is an important factor for electrodes used in photodetectors, as it reduces loss such as scattering and reflection of incident light energy [
38]. Therefore, the roughness of each thickness of the deposited samples was obtained using AFM in the range of 1 × 1 μm
2 images and is shown in
Figure 3,
Figures S4 and S5. The RMS values representing the surface roughness of the Ag:AZO thin film ranged from 0.6 to 1.0 nm (20–50 nm Ag:AZO thin film). However, the range of RMS values of the Ag and AZO thin films were 1.0 to 1.3 nm and 0.9 to 1.2 nm, respectively, indicating that the Ag:AZO thin film had the lowest roughness compared with the same thickness. The roughness of the Ag:AZO 20 nm thin film was reduced by 25% compared with the previously used Ag 20 nm thin film. The roughness of the thin film is a similar feature to the crystallite size calculated by the Scherrer equation based on the XRD pattern, and by this, it can be inferred that a smoother surface is formed in the thin film with a large crystallite size. In addition, because the roughness of the thin film has a great effect on the performance of the photodetector, it is expected that a photodetector with high responsivity can be manufactured using the Ag:AZO electrode.
Figure 4 shows the transmittance of the fabricated thin films in the wavelength range of 190 to 800 nm using a spectrophotometer. AZO, a transparent electrode, exhibited a high transmittance in visible light and a transmittance of 89% at a thickness of 20 nm. However, Ag exhibited a peak transmittance of 57% in the 320 nm wavelength band, and a low transmittance in the other wavelength bands except for the corresponding region. The Ag:AZO thin film fabricated via the co-sputtering of Ag and AZO exhibited optical properties of both Ag and AZO, and had a peak at 320 nm. It also exhibited a transmittance of 55 to 70% at a thickness of 20 nm.
The Hall meter was used to evaluate the electrical properties of the fabricated thin film to be used as an electrode. As shown in
Table 4, the sheet resistance, carrier concentration, and mobility were inferred, and after application to the device, changes in electrical properties were also observed after heat treatment of the electrode for post-heat treatment at 300 °C, to improve the crystallinity of Ag O [
39,
40,
41].
The sheet resistance of various thicknesses of deposited thin films decreased as the thickness increased, and in the case of the Ag:AZO thin film, the sheet resistance was at least six times lower than that of the AZO thin film, which is an existing electrode material, at all thicknesses. In addition, after annealing at 300 °C, the sheet resistance of the 20 nm Ag:AZO thin film was reduced by half from 72.8 to 38.1 Ω/Sq. Regarding carrier concentration with respect to temperature change, all three electrodes exhibited an increasing change, but there was a difference in the increase rate depending on the electrode. This increase in carrier concentration is because defects generated during thin film deposition using sputtering are reduced via heat treatment, thereby preventing loss due to trapping during the movement of carriers [
42]. However, owing to the increased electron concentration, the hole mobility decreased.
As shown in
Table 5, it was confirmed how the work function of Ag:AZO and co-sputtered Ag:AZO changed by measuring the work function of the deposited electrode. The work function of Ag:AZO is 4.66 eV, which is higher than Ag: 4.26 eV and AZO: 4.1 eV.
3.3. Characteristic Evaluation of Fabricated Device
Scheme 3 is a photodetector fabricated by sequentially depositing the p-type Ag
2O and electrodes evaluated above on a β-Ga
2O
3 substrate. As shown in the figure, all three electrodes with the same thickness were deposited on one device to evaluate the semiconductor characteristics and photoreaction.
Figure 6 is the I–V curve for each thickness of a Ag
2O/β-Ga
2O
3 photodetector based on the p-n junction deposited with a Ag:AZO electrode. Measurements were conducted in a dark room to eliminate the influence of light, and voltages from −2 to 4 V were applied to evaluate current changes according to voltage. The rectification characteristics of the p-n junction were observed at all thicknesses, and the off- and on-currents of the Ag:AZO 20 nm device were 2.56 × 10
−11A at 0 V and 5.16 × 10
−3 A at 1.73 V, respectively. The on/off ratio was calculated as 2.01 × 10
8.
Figure 7 shows the parameters calculated from the I–V curve of the photodetector shown in
Figure 5 based on the thermionic emission model [
44]. The parameters were on-resistance (R
on), barrier potential (
φB), and ideality factor (
n) based on thickness, and the thermionic emission model is as follows [
45].
In Equation (2), Js is the saturation current density, q is the charge, V is the voltage across the diode, Rs is the series resistance, and n is the ideality coefficient representing the deviation between the ideal diode and the actual non-uniform barrier and tunneling diode. k is the Boltzmann constant and T is the Kelvin temperature. In Equation (3), A is the contact area, A* is Richardson’s constant, and β-Ga2O3 has a value of 41 A/(cm2*K2).
As the electrode thickness increased, the on-resistance increased from 11.67 to 14.0 Ω·cm
2, the ideality coefficient increased from 1.42 to 2.42, and the barrier potential decreased from 1.58 eV to 1.29 eV. The change in the ideality factor is related to the interface state between the electrode and the p-type layer. As the thickness of the electrode increased, the stress and roughness of the material increased, causing defects at the interface between the p-type and the electrode, which hindered the capture or movement of carriers, leading to an increase in resistance [
46]. However, it was inferred that the barrier height increased as the thickness decreased because the roughness and interface defects decreased as the thickness of the electrode decreased. Defects present in the interfacial layer degrade the movement of carriers, so carriers move smoothly in samples with few defects, forming a wide depletion layer; hence, the barrier height increased [
47]. A photodetector with a high ideality coefficient is affected by charge trapping and carrier recombination, so it was judged that the device with the 20 nm Ag:AZO electrode had the best electrical characteristics [
48]. In addition, one of the causes of a high ideality coefficient is the hump phenomenon present in the I–V curve, which affects the current flow because of the high trap density due to the trapping of charged species in Ag and AZO combined crystals with different resistivity and conductivity. It was inferred that the hump phenomenon had occurred.
Figure 8 shows I–V curves of three electrodes with thicknesses of 20 nm. Rectification characteristics of the p-n junction were observed in all devices using each electrode, and the off- and on-currents of the Ag device were 1.62 × 10
−11 A at 0 V and 6.23 × 10
−4 A at 1.71 V, respectively, with an on/off ratio of 3.85 × 10
7. In addition, the off- and on-currents of the AZO device were found to be 2.58 × 10
−11 A at 0 V and 7.03 × 10
−4 A at 1.71 V, respectively, resulting in an on/off ratio of 2.72 × 10
7. Therefore, the Ag:AZO electrode was found to have the largest on/off ratio.
Table 6 displays the photodetector’s electrical parameters for each electrode. AZO, as an oxide, has a higher resistance than other electrodes, resulting in a relatively high
Ron value compared with the other electrodes. However, it can be inferred that there are few interface defects with the p-type layer, as the ideality coefficient is closest to one and the potential barrier is high.
Figure 9 depicts the I–V curve as a function of the intensity of UV irradiation on the Ag:AZO 20 nm device. Reverse bias is generated by applying irradiating light with a forward bias to the device, and reverse current is generated because of the generated electron-hole pair. Therefore, as shown in the figure, it was inferred that the stronger the light intensity, the greater the reverse current.
Figure 10 shows the photocurrent densities of the photodetectors to which each electrode was applied over time, when irradiated with UV radiation with an intensity and wavelength of 1000 μW/cm
2 and 254 nm, respectively, in zero bias. Ag:AZO exhibited the highest photocurrent density of 31 μA/cm
2, compared with AZO and Ag, which had photocurrent densities of 27 μA/cm
2 and 18.6 μA/cm
2. There are two reasons for this phenomenon. First, because of the low surface roughness, loss due to scattering and the reflection of light energy incident on the surface of the thin film is relatively small. Second, because of the low roughness of the electrode, sufficient bonding between the interface layers is achieved, resulting in lower contact resistance. Therefore, it is expected that the largest photocurrent appears in the photodetector using the 20 nm Ag:AZO electrode with the lowest roughness.
In
Figure 11, the photocurrent density obtained by irradiating each element with UV from 100 to 1000 μW/cm
2 in zero bias is shown. As the light intensity increased, more electron-hole pairs were generated, and the photocurrent increased. The light intensity-photocurrent density of the devices exhibited a linear increase. Responsivity, which represents photosensitivity and is an important index for evaluating the performance of photodetectors, and detectivity, which is a performance index for the smallest detectable signal, were evaluated from the photocurrent density value based on the light intensity.
Figure 12 shows the responsivity and detectivity as a function of the light intensity of each device. The equations for calculating responsivity and detectivity are as follows [
49,
50,
51,
52].
In Equation (4),
Jphoto is the photocurrent density,
Jdark is the dark-current density, and
P is the irradiated-light intensity. In Equation (5),
e is the absolute value of charge, and
J is the dark-current density. The responsivity and detectivity of each device calculated with the above formulas are shown in
Figure 12. Responsivity and detectivity decreased as the intensity of light increased. At higher intensities, more electron-hole pairs were generated, which induced self-heating due to collisions and vibrations between electrons, increasing charge carriers, as well as increasing recombination rates. Therefore, the highest value was measured at 100 μW/cm
2, where the effect of self-heating was the least. The maximum values of responsivity and detectivity of each photodetector with 20 nm electrode applied were Ag:AZO: 56 mA/W, 6.99 × 10
11 Jones; Ag: 42.5 mA/W, 5.3 × 10
11 Jones; and AZO: 44.4 mA/W, 4.8 × 10
11 Jones. In order to check the response spectra of the photodetector, the same measurement was performed under 365 nm light, but no photo-response was observed
Rise and decay times, which evaluate the response speed of the device, are shown in
Figure 13. Response speed, one of the important factors in evaluating photodetectors, is determined by the size of the internal potential caused by the difference in band gap between p-type Ag
2O and n-type β-Ga
2O
3 and measured using photocurrent characteristics over time [
50,
53,
54,
55,
56,
57,
58,
59]. When a UV radiation of 1000 μW/cm
2 was used in zero bias, the rise and fall times of the Ag:AZO, Ag, and AZO devices were 31.4 and 22 ms, 26 and 43 ms, and 63 and 63.8 ms, respectively. Based on the characteristics of the fabricated photodetector, the photodetector using the 20 nm-thick Ag:AZO electrode with the lowest surface roughness exhibited the highest photo response and responsivity. Furthermore, based on the current flowing characteristics, it was inferred that the p-n junction structure was driven without an external power supply owing to the built-in potential [
60,
61]. As shown in
Figure S6, additional measurement with a light of 1000 μW/cm
2 was conducted to confirm the sustainability of the Ag:AZO photodetector after 6 months of fabrication.