*Article* **Morphological, Optical, and Electrical Properties of p-Type Nickel Oxide Thin Films by Nonvacuum Deposition**

#### **Chien-Chen Diao 1, Chun-Yuan Huang 2, Cheng-Fu Yang 3,\* and Chia-Ching Wu 2,\***


Received: 9 March 2020; Accepted: 26 March 2020; Published: 29 March 2020

**Abstract:** In this study, a p-type 2 at% lithium-doped nickel oxide (abbreviation L2NiO) solution was prepared using Ni(NO3)2·6H2O, and LiNO3·L2NiO thin films were deposited using an atomizer by spraying the L2NiO solution onto a glass substrate. The sprayed specimen was heated at a low temperature (140 ◦C) and annealed at different high temperatures and times. This method can reduce the evaporation ratio of the L2NiO solution, affording high-order nucleating points on the substrate. The L2NiO thin films were characterized by X-ray diffraction, scanning electron microscopy, UV–visible spectroscopy, and electrical properties. The figure of merit (FOM) for L2NiO thin films was calculated by Haacke's formula, and the maximum value was found to be 5.3 <sup>×</sup> <sup>10</sup>−<sup>6</sup> <sup>Ω</sup>−1. FOM results revealed that the L2NiO thin films annealed at 600 ◦C for 3 h exhibited satisfactory optical and electrical characteristics for photoelectric device applications. Finally, a transparent heterojunction diode was successfully prepared using the L2NiO/indium tin oxide (ITO) structure. The current–voltage characteristics revealed that the transparent heterojunction diode exhibited rectifying properties, with a turn-on voltage of 1.04 V, a leakage current of 1.09 <sup>×</sup> <sup>10</sup>−<sup>4</sup> <sup>A</sup>/cm<sup>2</sup> (at 1.1 V), and an ideality factor of *n* = 0.46.

**Keywords:** lithium-doped nickel oxide; non-vacuum deposition; figure of merit; heterojunction diode

#### **1. Introduction**

At present, numerous applications, such as touch panels, light-emitting diodes, and solar cells, require transparent, conductive coatings [1–3]. Thus far, materials belonging to the transparent conducting oxide (TCO) family have been frequently used for this purpose. Most of the industry standard TCO are n-type wide bandgap oxides (Eg > 3.1 eV), such as In2O3, SnO2, and ZnO, whose conductivity can be further tuned by aliovalent doping or the formation of oxygen vacancies [4]. In contrast, the development of p-type TOS remains a challenge. Recently, semi-transparent p-type conducting films of the nickel oxide (NiO) have attracted considerable attention because of their importance in several scientific applications, including (i)material for electrochromic display devices [5,6], (ii) functional sensor layers in chemical sensors [7], (iii) transparent electronic devices [8] and (iv) the magnetic properties of nanoparticles [9–12]. A stoichiometric NiO thin film is an insulator at room temperature (resistivity is ~10<sup>13</sup> <sup>Ω</sup>·cm) [13]. Much effort has been made to explain the insulating behavior of NiO. NiO crystallizes in a rock-salt crystal structure, in which Ni cations have a nominal valence state of 2+ (3d8) in octahedral coordination (see Figure 1). Due to a strong electron correlation in 3d orbitals, it has an optical bandgap of 3.4–4.0 eV [14]. In addition, according to the literature, at temperatures above the Néel temperature (523 K), the crystal structure of NiO is cubic, whereas below

the Néel temperature, the crystals become slightly distorted and acquire a rhombohedral structure which accompanies the antiferromagnetic ordering [15].

NiO thin films can be grown by several chemical and physical methods, including magnetron sputtering [16,17], evaporation [18], the sol–gel method [19], laser ablation deposition [20], and spray pyrolysis (SP) [21,22]. NiO thin films with low resistivity (1.4 <sup>×</sup> 10−<sup>1</sup> <sup>Ω</sup>·cm) can be deposited by sputtering [23]. Compared with vacuum deposition, SP is a relatively simple, cost-effective nonvacuum deposition method for fabricating TCO thin films for large-area coating. However, the resistivity of the doped NiO thin films fabricated by SP is ~104 <sup>Ω</sup>·cm [24]; this resistivity is several orders of magnitude greater than that observed for sputter-deposited NiO thin films. Conventional SP involves spraying a nickel nitrate solution onto a preheated glass substrate at a temperature greater than 300 ◦C, followed by evaporation, solute precipitation, and pyrolytic decomposition. With the increase in the substrate temperature, the evaporation ratio of the solution on the substrate is extremely swift, leading to the formation of inferior NiO thin films. To solve this problem, a modified spray method was used in this study. First, the substrate temperature is slightly greater than the boiling point of the spray solution. The evaporation ratio of the spray solution decreases, affording high-order nucleation points by the spraying of the solution onto the substrate. The thin films were then formed and further annealed at high temperatures to afford a crystalline structure. Finally, high-quality thin films were obtained and subsequently applied as a photoelectric device.

To improve the conductivity of the NiO thin film, three improved mechanisms were used: (i) holes generated from nickel vacancies, (ii) oxygen interstitial atoms, and (iii) monovalent atoms used as a dopant. Monovalent atoms can be used as the dopant to increase the electrical conductivity of the NiO thin films [25,26]. In this study, a modified spray method was employed for the deposition of 2 at% Lithium (Li)-doped NiO (L2NiO) thin films with a high electrical conductivity. The monovalent atoms of Li can be substitution Ni atoms. In addition, the effects of annealing temperatures and times on the physical, optical, and electrical properties of the L2NiO thin films were investigated. X-ray photoelectron spectroscopy (XPS) was used to investigate the variations in the characteristics of the L2NiO thin films. Finally, a transparent heterojunction diode device comprising an L2NiO thin film and an indium–tin oxide (ITO) thin film was fabricated for future applications.

**Figure 1.** Crystal structure of the NiO thin film.

#### **2. Experimental Methods**

Lithium-doped nickel oxide (LNiO) thin films were deposited on a Corning glass substrate by the modified spray method. The spray solution was prepared by mixing nickel nitrate (Ni(NO3)2·6H2O, Alfa Aesar, MA, USA) and lithium nitrate (LiNO3, J.T. Baker, NJ, USA) in deionized (DI) water. A 1 M L2NiO spray solution was prepared by doping 2 at% Li in NiO. The modified spray method involved

spraying a L2NiO solution at 140 ◦C, which then evaporated, affording high-quality L2NiO thin films on the Corning glass substrate. The L2NiO thin films were deposited under the following conditions: solution volume = 40 mL, deposition rate = 10 mL/min. The distance between the Corning glass substrate and the nozzle was approximately 20 cm, and compressed air was used as the carrier gas. Annealing temperatures and times were 400–600 ◦C and 1–3 h, respectively, for the crystallization of the L2NiO thin films. Finally, to fabricate the transparent heterojunction diodes, L2NiO thin films were deposited on an ITO glass substrate, and the top and bottom aluminum (Al) electrodes were deposited by electron-beam evaporation. The surface morphology of the L2NiO thin films were examined by high-resolution scanning electron microscopy (HR-SEM, Hitachi, Japan). The resulting interface layer morphology between the L2NiO and ITO thin film was characterized by high-resolution transmission electron microscopy (HR-TEM, JOEL, Japan). The phase and crystallinity of the L2NiO thin films were measured by X-ray diffraction (XRD, Bruker, MA, USA) using CuKα radiation in the 2θ range of 20◦–80◦. The bonding state and element content of the L2NiO thin films were investigated using X-ray photoemission spectroscopy (XPS, ULVAC·PHI, Japan). The XPS using a monochromatic Al Kα X-ray (hν = 1486.6 eV) source was carried out at normal emission with an electron energy analyzer. The resistivity, carrier concentration, and mobility were measured by Hall effect measurements using the Van der Pauw method. The optical transmittance of the L2NiO thin films was measured using a UV–vis system (Agilent, CA, USA), and the transmittance spectrum was recorded as a function of the wavelength in the range of 200 to 1100 nm. The current–voltage (I–V) properties of the transparent heterojunction diode was measured using an HP4156 semiconductor parameter analyzer (Agilent, CA, USA).

#### **3. Results and Discussion**

Figure 2 shows the HR-SEM images of the L2NiO thin film with different annealing temperatures and times. The HR-SEM image of the L2NiO thin film which had annealed at 400 ◦C for 1 h revealed a smooth surface and no grain growth (Figure 2a). With the further increase in the annealing time to 3 h at 400 ◦C, the surface morphology revealed small grain sizes (Figure 2b). The average grain size of the film annealed at 400 ◦C for 3 h was 38 nm. Surface SEM morphologies shown in Figure 2c,d were compared by the increase in the annealing temperature of the L2NiO thin films from 500 ◦C to 600 ◦C for 3 h, and the grain sizes slightly increased. At annealing temperatures of 500 ◦C and 600 ◦C for 3 h, the average grain sizes of the L2NiO thin films were 45 nm and 58 nm, respectively. As a result of annealing at higher temperatures, the surface atoms on the substrate acquire more energy, and these atoms can move to suitable nucleation sites. In addition, the low activation energy ions doped in the thin film can easily escape from trap sites and transfer to nucleation sites. Crystalline thin films can be obtained when an increased number of better nucleation sites are formed on the substrate. In this study, the low activation energy of the Li ions doped in the NiO thin film leads to the increase in grain size with the increase in the annealing temperatures and times [27]. Compared with previous reports, the crystalline grain structure of the L2NiO thin films deposited by the modified spray method is better than that obtained by SP [28,29], because SP involves the deposition of the solution onto a preheated (>300 ◦C) substrate, but the evaporation ratio of the solution is extremely swift, affording poor nucleating points. Therefore, the surface morphology of the thin film is not good. The thickness of the L2NiO thin film with different annealing temperatures and times is shown in cross-section SEM images (Figure S1). The thickness of the L2NiO thin film annealed at 400 ◦C for 1 h was 202 nm. The thickness of the L2NiO thin films increased slightly as the annealing temperatures and times increased.

**Figure 2.** Surface SEM images of the L2NiO thin films as a function of annealing temperatures and times: (**a**) 400 ◦C for 1 h, (**b**) 400 ◦C for 3 h, (**c**) 500 ◦C for 3 h, and (**d**) 600 ◦C for 3 h.

The crystalline structure of the L2NiO thin films was examined by XRD using CuKα(λ = 0.1542 nm) radiation. Figure 3 shows the XRD patterns of the L2NiO thin films with different annealing temperatures and times. The observed XRD patterns of the L2NiO thin films were compared with the Joint Committee on Powder Diffraction Standards (JCPDS) data; they were in good agreement with the standard diffraction pattern of NiO (JCPDS card no. 47-1049). Diffraction peaks for the L2NiO thin films were observed at 2θ values of 37.3◦, 43.2◦, and 63.1◦, which correspond to the (111), (200), and (220) planes, respectively. The L2NiO thin films were polycrystalline without any other detectable secondary phase. Diffraction results revealed that the L2NiO thin film annealed at 400 ◦C for 1 h exhibited an approximate amorphous structure due to its weak-intensity diffraction peaks (Figure 3a). However, with the increase in the annealing temperatures and times from 400 ◦C to 600 ◦C and 1 to 3 h, respectively, diffraction intensities for the (111), (200), and (220) planes slightly increased (Figure 3b–d). The increase in the diffraction intensity was related to the grain sizes of the L2NiO thin films. Figure 3 (right side) also shows the grazing incidence angle X-ray diffraction patterns (GIAXRD) of the L2NiO films in the 2θ range of 42◦ to 45◦. The full-width half-maximum for the diffraction peak of the (200) plane of the L2NiO thin films decreased from 0.38 to 0.25. The crystallite size of the L2NiO thin films was then calculated using the Scherrer equation. With the increase in the annealing temperatures and times, the grain sizes increased from 39 nm to 60 nm. The results obtained for the various grain sizes were similar to those from SEM (Figure 2). In addition, with the increase in the annealing temperatures and times, the (200) plane was slightly shifted to high angles. According to Bragg's law (*n*λ = 2*dsin*θ) and *d* = *a*/*(h<sup>2</sup>* + *k2* + *l 2)* <sup>1</sup>/2, the lattice constant (a) slightly decreased from 4.178 Å to 4.169 Å with the increase in the annealing temperatures and times, indicating that the larger radius of Ni2<sup>+</sup> (0.69 Å) can be substituted by the smaller radius of Li<sup>+</sup> (0.68 Å); this subsequently leads to the decreased lattice constant of L2NiO thin films [30].

**Figure 3.** X-ray diffraction patterns of the L2NiO thin films as a function of the annealing temperatures and times: (**a**) 400◦C for 1 h, (**b**) 400◦C for 3 h, (**c**) 500◦C for 3 h, and (**d**) 600◦C for 3 h.

The crystal structure parameter of the L2NiO thin films produced with a 600 ◦C annealing temperature for 3h was fitted using the cubic structural model, with the atomic positions being described in the space group Fm3m. The fitted profiles of the L2NiO thin films for XRD data at 600 ◦C annealing temperature for 3h is shown in Figure 4. The final refinement convergence of the L2NiO thin films produced with a 600 ◦C annealing temperature for 3h was achieved with χ<sup>2</sup> = 1.38, and the measured result agreed well with the simulation value. i.e., the lattice constant (a) of the L2NiO thin was 4.1686 Å, similar to the value calculated using Bragg's law. The refined values of all thin film are also tabulated in Table 1.

θ **Figure 4.** Rietveld refinement of the L2NiO thin films produced with an annealing temperature of 600◦C for 3 h.


**Table 1.** Refined values of the L2NiO thin films with different annealing temperatures and times.

Figure 5a shows the optical transmittance spectra of the L2NiO thin films in the 250–1100 nm range. For the L2NiO thin films annealed at 400 ◦C for 1 h and those annealed at 400 ◦C, 500 ◦C, and 600 ◦C for 3 h, average transmittance values in the visible region (400 to 700 nm) were 46.8%, 72.3%, 84.6%, and 87.9%, respectively. The increase in the average transmittance of the L2NiO thin films was related to the increase in the grain size and decrease in the grain boundary, leading to the low scattering effect in L2NiO thin films. Surface SEM images revealed that the grain size of the L2NiO thin films increased with different annealing temperatures and times; this result was in agreement with the optical transmittance results. In the ultraviolet range, with the increase in the annealing temperature from 400 ◦C to 600 ◦C at an annealing time of 1 h to 3 h, the absorption edge was shifted toward a short wavelength region. The blue-shift can be explained by the Burstein–Moss shift effect [32–34].

**Figure 5.** (**a**) Optical transmittance spectra and (**b**) optical bandgap of the L2NiO thin films as a function of annealing temperatures and times.

The optical energy band gap (Eg) is an important physical parameter that mainly determines the electrical and optical characteristics of materials. The energy band gap of the L2NiO thin films were determined by applying the Tauc and Davis-Mott models [35]:

$$(al\nu)^{1/2} = B(h\nu - E\_{\S}^{op})\tag{1}$$

where α is the absorption coefficient, h is Planck's constant, ν is the frequency of the incident photon, and *B* is the absorption edge width. The energy band gap was determined by the extrapolation of a straight linear region of the plots to *h*ν = 0 [36]. Figure 5b shows (α*h*ν) <sup>1</sup>/<sup>2</sup> vs. *h*ν for L2NiO thin films with different annealing temperatures and times. With the increase in the annealing temperatures from 400 ◦C to 600 ◦C and the annealing times from 1 h to 3 h, the energy band gap of the L2NiO thin films increased from 2.89 to 3.21 eV (Figure 5b), suggesting that the Burstein–Moss effect may affect the band gap shift [33,34]. In the Burstein–Moss effect, the band-gap shift is mainly related to the high carrier concentration and/or low effective mass. According to the Burstein–Moss effect, the divergence of the band gap is expressed as

$$
\Delta E\_{\mathcal{S}}^{BM} = \frac{h^2}{2m\_{\rm rc}^\*} \left( 3\pi^2 n \right)^{\frac{2}{3}} \tag{2}
$$

where Δ*EgBM* is the shift value between the doped semiconductor and undoped semiconductor, mcv\* is the reduced effective mass, and n is the carrier concentration. The absorption edge of the L2NiO thin films were observed in a shorter wavelength region because of the increase in the carrier concentration (*n*) (Figure 5a).

Figure 6 shows the optical energy band gap, carrier concentration (*n*), mobility (μ), and resistivity (ρ) of the L2NiO thin films with different annealing temperatures and times. All samples exhibited p-type properties. With the increase in the annealing temperatures and times, the mobility of the L2NiO thin films increased from 2.39 to 11.96 cm2/Vs because of the increase in the grain size and decrease in the grain boundary, causing the carrier to encounter less hindering materials, and subsequently resulting in increased carrier mobility. Meanwhile, the increase in the annealing temperatures and times caused an increase in the carrier concentration. The number of Li atoms substituting the sites of Ni atoms increased, leading to the substitution of a large number of Ni2<sup>+</sup> by Li ions in the normal crystal sites and creating holes at a high annealing temperature. Therefore, the carrier concentration of the L2NiO thin films increased; this result can be obtained by Equation (3).

$$\frac{1}{2}\text{O}\_2^{(\text{g})} + \text{Li}\_2\text{O} \Leftrightarrow 2\text{O}\_\text{x}^\text{O} + 2\text{Li}\_{\text{Ni}\text{i}} + 2\stackrel{\cdot}{\text{h}} \tag{3}$$

With the increase in the annealing temperatures and times, the Li concentration of the L2NiO thin films increased, as demonstrated by the XPS analysis shown in Table 2 and the XRD analysis shown in Figure 3 (right side). The resistivity of the film is known to be proportional to the reciprocal of the product of carrier concentration and mobility, as follows:

$$
\rho = 1 / \left( n \times \mu \right) \tag{4}
$$

**Table 2.** Elements of L2NiO thin films as a function of annealing temperature and time.


**Figure 6.** Optical band gap, carrier concentration, mobility, and resistivity of the L2NiO thin films as a function of annealing temperatures and times.

Therefore, with the increase in the carrier concentration and mobility, the resistivity of the L2NiO thin films decreased from 4.73 to 1.08 Ω·cm. Compared with previous reports, the resistivity of L2NiO thin films was slightly less than those of undoped NiO thin films [36,37].

Figure 7 shows the Ni 2p3/<sup>2</sup> XPS spectra of the L2NiO thin films with different annealing temperatures and times. During the Gaussian fitting process, binding energies for the NiO and Ni2O3 peaks were observed at 854.0 eV and 855.8 eV, respectively, in the Ni 2p3/<sup>2</sup> XPS spectra of the L2NiO thin films annealed at 400 ◦C for 1 h (right side of Figure 7) [38,39]. The Ni2O3 and NiO peaks corresponded to Ni3<sup>+</sup> and Ni2+, respectively. With the increase in annealing temperatures and times for the L2NiO thin films, the intensity of the Ni3<sup>+</sup> bonding state slightly increased over that of the Ni2<sup>+</sup> bonding state. This result is related to the insertion of an excess amount of oxygen ions in the interstitial sites of the L2NiO thin films due to annealing conducted in the atmosphere; this leads to the formation of Ni3<sup>+</sup> ions and holes, which can be represented by Equation (5).

$$\frac{1}{2}\text{O}\_2^{(\text{g})} \Leftrightarrow \text{O}\_i'' + 2\stackrel{\cdot}{\text{h}} \tag{5}$$

Meanwhile, the Ni 2p3/<sup>2</sup> results were also confirmed from the O1s XPS spectra of the L2NiO thin films with annealing temperatures and times, as shown in Figure 8. With the increase in the annealing temperature and times, the intensity of the O1s peak increased appreciably. In the O1s XPS spectra, the deconvolution of the electron binding energy of NiO (529.3 eV) and Ni2O3 (531.7 eV) was observed for the L2NiO thin films [38,40,41]. The increase in the magnitude of the Ni3<sup>+</sup> bonding state was slightly greater than that of the Ni2<sup>+</sup> bonding state, demonstrating that the hole carrier concentration of the L2NiO thin films increased with the annealing temperatures and times. The increased hole carrier concentration of the L2NiO thin films was in agreement with the Hall measurement (Figure 6).

**Figure 7.** Ni 2p3/<sup>2</sup> XPS spectra of the L2NiO thin films as a function of annealing temperatures and times: (**a**) 400◦C for 1 h, (**b**) 400◦C for 3 h, (**c**) 500◦C for 3 h, and (**d**) 600◦C for 3 h.

**Figure 8.** O 1s XPS spectra of the L2NiO thin films as a function of annealing temperatures and times. (**a**) 400◦C for 1 h, (**b**) 400◦C for 3 h, (**c**) 500◦C for 3 h, and (**d**) 600◦C for 3 h.

Figure 9 shows the figure-of-merit (FOM) of the L2NiO thin films with different annealing temperatures and times. To deposit L2NiO thin films with high transmission and low resistivity, the FOM values for the L2NiO thin films with different annealing temperatures and times were calculated by using Haacke's equation [42]:

$$\text{FOM} = \frac{T^{10}}{R\_s} \tag{6}$$

where *T* is the average optical transmittance at 400–700 nm and *Rs* is the sheet resistance of the L2NiO thin films. With the increase in annealing temperatures and times, the FOM of the L2NiO thin films increased (Figure 9). The maximum FOM (5.3 <sup>×</sup> 10−<sup>6</sup> <sup>Ω</sup>−1) was obtained for the L2NiO thin films annealed at 600 ◦C for 3 h. FOM results revealed that L2NiO thin films exhibit satisfactory optical and electrical characteristics for photoelectric device applications.

**Figure 9.** FOM values for L2NiO thin films as a function of annealing temperatures and time.

From the above results, it can be seen that the carrier concentration, mobility, and conductivity characteristics of the Li ions doped in NiO thin film improved compared to non-doped NiO thin film. For the possible fabrication of transparent heterojunction diodes, the L2NiO thin films were then deposited at an annealing temperature of 600 ◦C and an annealing time of 3 h onto an ITO glass substrate to form a p–n junction structure. The ITO thin film was deposited by the sputtering method. The thickness of the ITO thin film was 1000 Å with a 92% visible-light transmittance and a resistance of 12 Ω·cm, as shown in Figure S2a,b. Figure 10 shows the current–voltage (I–V) curve of the L2NiO/ITO transparent heterojunction diode. The I–V results confirmed that the fabricated L2NiO/ITO transparent heterojunction diode exhibited rectifying behavior with the use of aluminum (Al) as the electrodes. Before the measurement of the I–V properties of the L2NiO/ITO transparent heterojunction diode, ohmic contacts were confirmed to be present between the Al electrodes and the L2NiO and ITO thin films. Under forward bias, the turn-on voltage for the L2NiO/ITO transparent heterojunction diode was ~1.04 V (Figure 10a); this value is less than that (2.57 V) for a p-NiO/n-TZO diode [43], 2.5 V for a p-CuO/n-ZnO diode [44], and similar (1 V) to that for a p-NiO/n-ZnO diode [45]. Under a reverse voltage, the leakage current was 1.09 <sup>×</sup> 10−<sup>4</sup> A/cm<sup>2</sup> at 1.1 V for the L2NiO/ITO transparent heterojunction diode (Figure 10b). The rectification ratio (R) for the L2NiO/ITO transparent heterojunction diodes was calculated using Equation (7) as follows to obtain a value of 17.3 (at 1.1 V):

$$R = \frac{\text{Forward current}}{\text{Revenue current}}\tag{7}$$

The ideality factor (*n*) can be calculated from the slope of the linear region of the forward-bias log(I)–V curve, which can be derived from Equation (8) and Figure 11:

$$n = \frac{q}{kT} \times \left[\frac{dV}{d\ln(I)}\right] \tag{8}$$

where *k* is the Boltzmann constant, *T* is the temperature in kelvin, and *q* is the electron charge. The ideality factor for the L2NiO/ITO transparent heterojunction diode was *n* = 0.46, which was less than the ideal value of *n* = 1 (Equation (8)). The high leakage current and low ideality factor result from imperfections between the heterojunction interfaces of the L2NiO and ITO thin films. Ajimsha et al. reported that in an oxide layer, these imperfections were caused by the presence of different crystal-type structures [46].

**Figure 10.** Current–voltage curve of the L2NiO/ITO transparent heterojunction diode: (**a**) forward current and (**b**) reverse current.

**Figure 11.** Log current–voltage curve of the L2NiO/ITO transparent heterojunction diode.

To further investigate the interface between the L2NiO and ITO thin films, TEM images were recorded. Figure 12a shows a magnified TEM image of the interface between the L2NiO thin film annealed at 600 ◦C for 3 h and the ITO thin film; a clear layer was present at the interface between the two materials. The interfacial layer thickness was 55 Å; this layer was thought to be NiO2, Ni2O3, and Ni3O4, because during spraying, Ni and O can be easily combined with In or Sn in ITO. This result can be attributed to the relatively high bond energy of Ni–O (1029 kJ/mol) compared with those of Sn–O (531.8 kJ/mol) and In–O (320 kJ/mol) [47–49]. From the XPS result, it was hypothesized that the interfacial layer between the NiO and ITO was Ni2O3. Figure 12b shows the selected-area electron diffraction (SAED) pattern of the boundary between the L2NiO and ITO thin films. The corresponding SAED pattern exhibited (200) NiO, (211) ITO, and (002) Ni2O3 diffraction rings, confirming that the L2NiO (including NiO and Ni2O3) and ITO thin films are polycrystalline; this result also confirmed that a thin Ni2O3 interfacial layer was present between the L2NiO and ITO thin films. The thin Ni2O3 interfacial oxide layer rendered a high leakage current and low ideality factor for the L2NiO/ITO transparent heterojunction diode.

**Figure 12.** (a) TEM image and (b) SAED pattern of the L2NiO/ITO transparent heterojunction structure.

#### **4. Conclusions**

In this study, a modified spray method was used for the deposition of high-quality 2 % Li-doped NiO (L2NiO) thin films. The L2NiO thin films exhibited a cubic (NaCl-type) structure, and the lattice constant of the L2NiO thin films slightly decreased from 0.4178 Å to 0.4169 Å with the increase in annealing temperatures and times. As the smaller radius of Li<sup>+</sup> (0.6 Å) was substituted by the larger Ni2<sup>+</sup> (0.69 Å), the number of substituted Li<sup>+</sup> increased, leading to a decrease in the lattice constant of the L2NiO thin films. According to the Burstein–Moss shift theory, the optical energy band gap (Eg) of the L2NiO thin films increased from 2.89 eV to 3.21 eV with the increase in the annealing temperatures and times because of the increase in the carrier concentration. In Hall measurements, the carrier concentration and mobility of the L2NiO thin films increased, leading to a decrease in the resistivity from 4.73 Ω·cm to 1.08 Ω·cm with the increase in the annealing temperatures and times. The optimum FOM (5.3 <sup>×</sup> <sup>10</sup>−<sup>6</sup> <sup>Ω</sup>−1) was obtained for the L2NiO thin films annealed at 600 ◦C for 3 h, for which the resistivity and average transmittance were 1.08 Ω·cm and 87.9%, respectively. Finally, the transparent heterojunction diode comprising a p-type L2NiO thin film and an n-type ITO thin film was successfully fabricated. Its properties included (1) a turn-on voltage of 1.04 V, (2) a leakage current of 1.09 <sup>×</sup> <sup>10</sup>−<sup>4</sup> <sup>A</sup>/cm<sup>2</sup> (at 1.1 V), (3) a rectification ratio of 17.3, and (4) an ideality factor of 0.46. The high leakage current resulted from the Ni2O3 thin layer between the heterojunction interfaces of the different crystal-type structure with the L2NiO and ITO thin films. Therefore, the L2NiO film was

shown to possess satisfactory properties for applications including transparent diode, electrochromic display, and solar cell devices.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2079-4991/10/4/636/s1, Figure S1: Cross-section SEM images of the L2NiO thin films as a function of annealing temperatures and times: (a) 400 ◦C for 1 h, (b) 400 ◦C for 3 h, (c) 500 ◦C for 3 h, and (d) 600 ◦C for 3 h. Figure S2: (a) Cross-section SEM image and (b) optical transmittance spectra of the ITO thin film.

**Author Contributions:** C.-C.D. participated in I-V measurement and design the transparent heterojunction diode. C.-Y.H. participated in XRD analysis and Rietveld refinement of the XRD data. C.-F.Y. and C.-C.W. participated in the design experimental of Li doped NiO films. C.-C.W. participated in the Hall, SEM, and XPS analysis of Li doped NiO films and fabrication of Li doped NiO films. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors acknowledge the financial support of the Ministry of Science and Technology (MOST 108-2221-E-143-001, MOST 108-2622-E-143-001-CC3). The authors gratefully acknowledge the use of high-resolution scanning electron microscope and multipurpose X-Ray thin-film micro area diffractometer equipment belonging to the Instrument Center of National Cheng Kung University.

**Conflicts of Interest:** The authors declare that they have no conflict of interests.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Electrochemical Deposition of Silicon-Carbon Films: A Study on the Nucleation and Growth Mechanism**

**Nina K. Plugotarenko 1, Tatiana N. Myasoedova 1,\*, Mikhail N. Grigoryev <sup>2</sup> and Tatiana S. Mikhailova <sup>1</sup>**


Received: 13 November 2019; Accepted: 6 December 2019; Published: 10 December 2019

**Abstract:** Silicon-carbon films have been deposited on silicon and Al2O3/Cr-Cu substrates, making use of the electrolysis of methanol/dimethylformamide-hexamethyldisilazane (HMDS) solutions. The electrodeposited films were characterized by Raman spectroscopy and scanning electron microscopy, respectively. Moreover, the nucleation and growth mechanism of the films were studied from the experimental current transients.

**Keywords:** nucleation; growth; electrochemical deposition; silicon-carbon films

#### **1. Introduction**

The diamond-like carbon (DLC) films are extremely alluring for their high mechanical hardness, high electric resistivity, biocompatibility, chemical inertness, low coefficient of friction, and optical transparency in the infrared range [1–3]. The issue of stress and poor adhesion to the substrate in DLC films is a persistent problem that could be solved by incorporation of other elements (W, Ti, Al, Si, etc.) [4–6]. Therefore, the incorporation of silicon is rather promising in order to obtain amorphous silicon-carbon films.

Silicon-carbon films are very promising materials for microelectronic devices operating in aggressive environments [7]. These films are used for gas sensors, ultracapacitors, field emission devices, and other applications in aggressive environments. There are many techniques for producing these films, such as magnetron sputtering [8], ion sputtering, chemical vapor deposition, pulsed laser deposition, electrochemical deposition from molten salt, and the sol-gel method [9–11]. However, the applications of these techniques have been limited, owing to the sophisticated equipment and precise experimental conditions, including high vacuum and high temperature. It was experimentally shown that most materials that can be deposited from the vapor phase can also be deposited in a liquid phase using electrochemical techniques and inversely [10]. The application of the liquid deposition techniques is a good prospect due to such advantages as low consumption of energy, low deposition temperature, availability for large area deposition on complicated surfaces, and the simplicity of the setup. There are some reports that have demonstrated the possibility of the electrochemical deposition of DLC films from the organic liquids such as methanol [12], acetonitrile [13], dimethylsulfoxide [14], and lithium acetylide in dimethylsulfoxide [15], in ambient conditions. However, earlier, we reported the electrochemical deposition of silicon-carbon films from methanol/ethanol and hexamethyldisilazane (HMDS) solution [16,17]. However, in the development of the synthesis of a new material, the deposition kinetics is one of the first components to be studied in detail to ensure reproducibility. Currently, there is no information about the deposition mechanisms of silicon-carbon films from organic liquids onto different substrates.

Electrochemical methods allow setting and controlling the overpotential, control charge, current, the volume of the deposited solution, and a number of nuclei comparatively easily in the system, so they are suitable for the study of the nucleation and growth of a new phase. The analysis of potentiostatic current transients allows getting more information on the mechanism and kinetics of the electrodeposition [18].

The aim of the present study is to investigate the mechanisms of the nucleation and growth of silicon-carbon films onto silicon and Al2O3/Cr-Cu substrates through experimental potentiostatic current transients. The surface morphology, as well as structural and phase composition of the films were determined from scanning electron microscopy and Raman spectra investigations, respectively.

#### **2. Materials and Methods**

#### *2.1. Synthesis of Silicon-Carbon Films*

In this communication, the silicon-carbon films were deposited on silicon (100) (the resistivity was 4.5 Om·cm) and Al2O3 substrates with a size of 12 <sup>×</sup> 17 mm2. In the first step, the silicon substrate was dipped in the HF solution (≈ 15%) for a few minutes, and the conducting layer (Cr-Cu) was sputtered on the surface of the Al2O3 substrate by the magnetron technique. The substrate was mounted on the negative electrode, and graphite was mounted on the positive electrode. The distance between the substrate and the positive electrode was set to 10 mm. The deposition was done from two types of solution: (1) a methanol and HMDS solution; (2) a dimethylformamide (DMF) and HMDS solution. HMDS was dissolved in analytically pure methanol/DMF, with the volume ratio of HMDS to methanol (DMF) of 1:9. The films were deposited for 30 min. The applied potential was 180 and 500 V, for methanol-HMDS and DMF-HMDS solutions, respectively.

A schematic diagram of the experimental setup is shown in Figure 1:

**Figure 1.** Schematic structure of electrolytic deposition system (1, glass cell; 2, dielectric cover; 3, graphite anode; 4, cathode substrate; 5, solution; 6, thermocouple; 7, clamps; 8, thermal table; 9, voltmeter of the thermocouple; 10, ammeter; 11, high-voltage voltmeter; 12, power supply).

#### *2.2. Characterization*

The film morphologies were investigated using scanning electron microscopy (SEM; SEM Zeiss Merlin compact VP-60-13, Stavropol, Russia). Raman spectra were recorded at ambient temperature using a Raman Microscope, Renishaw plc (Stavropol, Russia, resolution 2 cm<sup>−</sup>1, 514 nm laser).

#### **3. Results and Discussion**

#### *3.1. Characterization*

During the deposition for a composite film from the DMF-HMDS solution, we found that the current density increased from 35 mA/cm2 to 54–57 mA/cm2 with deposition time. In the case of the methanol-HMDS solution, the current density decreased slightly from 50 mA/cm2 to 44 mA/cm<sup>2</sup> and increased from 50 mA/cm2 to 55 mA/cm2 during the film deposition onto the silicon and Al2O3 substrate, respectively (Figure 2).

**Figure 2.** Experimental potentiostatic current transients for the deposition of silicon-carbon films on silicon (**a**) and Al2O3 (**b**) substrates.

The surface morphology of the films changes under varying technological conditions. The production of silicon-carbon materials is associated with thermodynamically nonequilibrium processes, which cause the formation of inhomogeneities as the films grow due to the self-organization of the structure. Figures 3 and 4 shows the SEM micrographs of the deposited films. From the figures, it can be seen that films deposited from the methanol-HMDS solution and DMF-HMDS solution on the silicon substrate are composed of compact grains. The average grain size was about 90, 60, and 170 nm for the films, deposited from the methanol-HMDS on the silicon substrate, from the DMF-HMDS solution on the silicon substrate, and from the methanol-HMDS on the Al2O3 substrate, respectively. The silicon-carbon films deposited from the DMF-HMDS solution on the Al2O3 substrate characterized by a powdery structure without large grains. Therefore, the histograms of the grain size distributions were built (Figure 5).

**Figure 3.** SEM micrographs of the silicon-carbon films deposited onto the silicon substrate from the methanol-hexamethyldisilazane (HMDS) (**a**) and DMF-HMDS (**b**) solutions.

**Figure 4.** SEM micrographs of the silicon-carbon films deposited on the Al2O3 substrate from the methanol-HMDS (**a**) and DMF-HMDS (**b**) solutions.

The scatter of grain size values for the films on silicon substrates lied in the range from 20 nm to 200 nm. Grains with sizes of 50 and 80 nm predominated for the films deposited from the methanol-HMDS and DMF-HMDS solutions, respectively. For the films deposited onto the Al2O3 substrate, the histogram of the grain size values distribution was characterized by the absence of pronounced maxima. It was evident that the range of grain sizes for the films deposited from the methanol-HMDS solution was much narrower than for those deposited from the DMF-HMDS solution and was in the range of 60–150 nm.

**Figure 5.** Histograms of the grain size distributions of the silicon-carbon films deposited on silicon (**a**) and Al2O3 (**b**) substrates.

The Raman spectra of the films with the deconvolution of the D and G peaks, deposited on silicon and Al2O3 substrates, are shown in Figure 6a,b, respectively.

The silicon-carbon films deposited from the methanol-HMDS solution investigated in this work were complex heterogeneous objects (Figure 6a). The Raman spectra contained the lines in the range that was characteristic of the SiC polytypes. The samples were characterized by the presence of the hexagonal 6H SiC polytype with the impurities of the rhombohedral 15R SiC phase. Furthermore, the bands attributed to the Si–C bond and nanocrystalline diamond (ND) were observed. The spectrum of the silicon-carbon film deposited on the Al2O3 substrate shifted to a lower wavenumber. The deconvolution of the Raman spectra allowed us to find out "hidden" peaks. Deconvolution was carried out on a minimum number of Gauss peak components for which their resulting curve described the experimental curve with confidence >0.99%. Therefore, in the resulting Gauss deconvolution, three peaks were observed at 1361, 1524, and 1627 cm<sup>−</sup>1. The peaks centered at 1361 and 1524 cm−<sup>1</sup> corresponded to the conventional D and G bands. The broadening in the G band at the higher wavenumber side was due to the presence of the D' band at 1627 cm<sup>−</sup>1. The appearance of the D' peak proved that silicon-carbon films were highly defective structures [19]. The relative intensity ratio of the D peak to G peak (ID/IG) of the silicon-carbon films deposited from the methanol-HMDS solution was 1.05 for the films on both types of substrates.

**Figure 6.** Raman spectra with the deconvolution of the D and G peaks (under the Raman spectra) of the silicon-carbon films deposited onto the silicon (1) and Al2O3 (2) substrates from the methanol-HMDS (**a**) and DMF-HMDS (**b**) solutions (D\* and D' peaks characterize disorder carbon).

Raman spectra of silicon-carbon films deposited from the DMF-HMDS solution could be characterized by the presence of the D peak and the G peak (Figure 6b). The spectrum of the silicon-carbon film on the silicon substrate was also characterized by the D + G scattering peak.

In the spectrum of the silicon-carbon film deposited on the silicon substrate, the position of the D and G peaks was 1386 and 1587 cm<sup>−</sup>1, respectively (Figure 6a), while the position of the D and G peaks was 1438 and 1597 cm<sup>−</sup>1, respectively, in the spectrum of silicon-carbon film, deposited on the Al2O3 substrate [20]. Furthermore, the G peak of the silicon-carbon film deposited on the silicon substrate shifted to a lower wavenumber, and the full width at half maximum of the G peak was also larger than that of the silicon-carbon film, deposited on the Al2O3 substrate. The high intensity of the D peak confirmed the existence of unsaturated hydrocarbons on the surface of SiC nanoparticles [21]. The bands attributed to the hexagonal 6H SiC polytype were observed.

The deconvolution of the D and G bands of the films deposited from the DMF-HMDS solution was also carried out as shown in Figure 6b. The D\*, D, and G peaks were found. It should be noted that the D\* peak has been found in disordered carbons. Some reports have attributed the D\* peak to the sp3 rich phase of disordered amorphous carbons [22]. The D and G peaks were centered at 1405 (1400) cm−<sup>1</sup> and 1600 (1584) cm<sup>−</sup>1.

Furthermore, it was seen that the relative intensity ratio of the D peak to G peak (ID/IG) of the silicon-carbon film deposited from the DMF-HMDS solution was higher than for the films deposited from the methanol-HMDS solution and reached ~1.29. The smaller ratio corresponded to smaller free carbon clusters [23].

#### *3.2. Mechanism Study*

The structure and morphology of silicon-carbon films depends on the nucleation and growth mechanism.

Potentiostatic transient measurement is an important method for studying the initial kinetics of electrocrystallization reactions [24–26].

The existing models of electrochemical deposition were based on two main ideal mechanisms for new phase nucleation on the electrode surface: instantaneous nucleation and progressive nucleation. In the case of instantaneous nucleation, all active centers are filled almost simultaneously, and further, slow growth of nuclei occurs due to the introduction of new atoms. In the presence of inhomogeneities on the surface of the substrate, germ growth first occurs at the most active centers, so with progressive nucleation, the nuclei simultaneously emerge and continue to grow. It is assumed that there is a constant supersaturation of the precursor concentration under potentiostatic conditions. Besides, both kinetic controlled and diffusion controlled growth mechanisms of a new phase on the surface are possible.

The model of 3D multiple nucleations with kinetic controlled growth was described by Isaev [18]. Instantaneous nucleation is described by:

$$\frac{\dot{j}}{\dot{y}\_{\text{max}}} = 2.34 \frac{t}{t\_{\text{max}}} \omega \left( 1.50 \frac{t}{t\_{\text{max}}} \right) \tag{1}$$

where *j* is the current density, *t* is time, and *t*max is the time at the maximum current.

Progressive nucleation can be expressed as:

$$\frac{j}{j\_{\text{max}}} = 2.25 \omega\_2 \left( 1.34 \frac{t}{t\_{\text{max}}} \right) \tag{2}$$

where ω(x) = exp- <sup>−</sup>*x*<sup>2</sup> *<sup>x</sup>* <sup>0</sup> exp(ξ2)d<sup>ξ</sup> is Dawson's integral:

$$\alpha \mathbf{z}(\mathbf{y}) = \exp(-y^3) \int\_0^y \left(y^2 - \xi^2\right) \exp\left(3y\xi^2 - 2\xi^3\right) \mathrm{d}\xi \tag{3}$$

The model of controlled nucleation was offered by Scharifker and Hills [27]. They considered the 3D nucleation model given that over time, the diffusion zones of individual nuclei overlap, which leads to a slowdown in germ growth. Instantaneous nucleation and growth are described by:

$$\left(\frac{j}{j\_{\text{max}}}\right)^2 = \frac{1.9542}{\left(t/t\_{\text{max}}\right)} \left\{1 - \exp\left[-1.2564 \left(t/t\_{\text{max}}\right)\right]\right\}^2 \tag{4}$$

Progressive nucleation can be expressed as:

$$\left(\frac{j}{j\_{\text{max}}}\right)^2 = \frac{1.2254}{\left(t/t\_{\text{max}}\right)} \left\{1 - \exp\left[-2.3367(t/t\_{\text{max}})^2\right]\right\}^2 \tag{5}$$

The experimental current–time transients shown in Figure 2 were analyzed using these expressions and experimentally obtained values for *<sup>j</sup>*max and *<sup>t</sup>*max. First, the dependences of ln <sup>1</sup> <sup>−</sup> *<sup>j</sup>* √ *t* (*j* √ *t*)max from *t* and *t* <sup>2</sup> were built in order to determine instantaneous or progressive nucleation.

Figure 7 shows graphs of electrodeposition transients characteristic of instantaneous nucleation.

**Figure 7.** Semilogarithmic dependencies calculated from the current transients for the film deposition solution on the silicon substrate from the methanol-HMDS solution (1); on the Al2O3 substrate from the methanol-HMDS solution (2) and DMF-HMDS solution (3).

Progressive nucleation is described by Figure 8.

**Figure 8.** Semilogarithmic dependencies calculated from the current transients for the film deposition from the DMF-HMDS solution onto the silicon substrate.

As shown in the figures, for all straight lines, a high approximation confidence value was set. The comparison was based on the standard error value. Deviations from linearity were caused by concurrent processes in the solution and on the substrate: molecules' dissociation, heating of the solution, and the formation of silicon-carbon and carbon bonds, characterizing the different growth rates.

In Figure 9, the model and experimental dependencies are presented. The analysis of the semilogarithmic and (*j*/*jm*) vs. (*t*/*tm*) dependencies showed that the mechanism of nucleation and growth of silicon-carbon films from the methanol-HMDS and DMF-HMDS solutions on the Al2O3 substrate was well described by Equation 1 for instantaneous nucleation (Figure 9b,d). The experimental current transients represented in the coordinates (*j*/*jm*) vs. (*t*/*tm*) for the deposition of silicon-carbon film from the DMF-HMDS solution on the silicon substrate demonstrated the characteristic features of the diffusion controlled growth model (Figure 9c).

The deposition of the silicon-carbon films from the methanol-HMDS solution onto the silicon substrate was characterized by the instantaneous nucleation with kinetically controlled growth (Figure 9a), while the model for instantaneous nucleation with diffusion controlled growth fit the growth mechanisms of the new phase from the methanol-HMDS and DMF-HMDS solutions on the Al2O3 substrate.

All the experimental dependences of *(j*/*jm*) vs. (*t*/*tm*) except the deposition from DMF-HMDS solution on the silicon substrate demonstrated the higher current density compared to the model for first two minutes due to the dissociation of molecules in precursors.

**Figure 9.** *Cont.*

**Figure 9.** Experimental and model dependences of the current density on the deposition time of silicon-carbon films from: (**a**) methanol-HMDS solution on a silicon substrate; (**b**) methanol-HMDS solution on a Al2O3 substrate; (**c**) DMF-HMDS solution on a silicon substrate; (**d**) DMF-HMDS solution on a Al2O3 substrate.

#### **4. Conclusions**

The silicon-carbon films were successfully deposited on silicon and Al2O3/Cu-Cr substrates from organic solutions. The films deposited from the methanol-HMDS solution were mostly characterized by the presence of the hexagonal 6H SiC polytype with the impurities of the rhombohedral 15R SiC phase. Raman spectra of silicon-carbon films, deposited from the DMF-HMDS, solution can be characterized by the presence of the D peak, G peak, and D + G scattering peaks of carbon and 6H SiC polytype peaks. It was shown that the nucleation and growth mechanisms depend on the nature of the solution and substrate.

**Author Contributions:** Conceptualization, N.K.P. and T.N.M.; data curation, N.K.P. and T.N.M.; formal analysis, N.K.P. and T.N.M.; funding acquisition, T.N.M. and N.K.P.; investigation, M.N.G. and T.S.M.; methodology, N.K.P., T.N.M., and M.N.G.; project administration, T.N.M. and N.K.P.; resources, T.N.M., M.N.G. and N.K.P.; software, N.K.P.; supervision, T.N.M.; validation, T.N.M. and N.K.P.; visualization, T.N.M., M.N.G., N.K.P., and T.S.M.; writing, original draft preparation, N.K.P. and T.N.M.; writing, article and editing, T.N.M. and T.S.M.

**Funding:** This work was financially supported by the Ministry of Education of Russia, under Contract No. 14.575.21.0126 (the unique identifier for the contract is RFMEFI57517X0126).

**Acknowledgments:** The authors acknowledge "The Center for the Collective Use of Scientific Equipment" of the North-Caucasus Federal University (Stavropol, Russia) for Raman spectroscopy and SEM investigations.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study in the collection, analyses, or interpretation of data; in the writing of the manuscript; nor in the decision to publish the results.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
