3.1. Microstructure and crystal structure analyses
For the Cu-doped film series, deposited at pH = 6.85, a compact surface was achieved for the as-grown films with a small amount of Cu (x≤0.025). Figure 1a–c shows topographical AFM images of the surface of samples with x = 0.000, 0.015, and 0.025, respectively. The films reveal a uniform morphology without any cracks; no significant morphological differences were found among these films, except for an increase in the grain size, as listed in Table 1. However, the FE–SEM observations in Fig. 2a–c clearly display the degradation in the surface quality of the thin films with high Cu levels (x≥0.030). As x reaches 0.035, the film shows many pores on the surface as clearly seen in Fig. 2b. When x further increases to 0.040 (Fig. 2c), the porosity becomes much higher, resulting in a poorer microstructure and a decrease in thickness.
Table 1
Grain size, roughness, and thickness of films
x | Zn-free | Zn-added | |
| Mean grain size (nm) | Roughness (nm) | Thickness (nm) | Mean grain size (nm) | Roughness (nm) | Thickness (nm) |
0.000 | 50 | 17.7 | 420 | 4 | – | 300 |
0.015 | 60 | 22.5 | 450 | 7 | – | 280 |
0.025 | 70 | 23.1 | 400 | 10 | 4.1 | 280 |
0.030 | 70 | 31.8 | 400 | 13 | 6.7 | 270 |
Annealing temperature: 400 °C |
The average grain size was evaluated from surface images acquired by FE–SEM and AFM and using ImageJ software. |
The roughness was calculated using an AFM scanning probe image processor (SPIP–4.8.2) software. |
The thickness was evaluated from cross-sectional FE–SEM images. |
The (Zn,Cu) co-doped (Ni,Mn)3O4 thin films were prepared from nickel manganese chloride solutions by simultaneously adding 0.01 mol ZnCl2 and x (0.000–0.030) mol CuCl2 under the same conditions as for the Cu-doped-only films, except that the solution pH was adjusted to decrease from 6.85 to 6.65. The AFM photographs of as-grown Zn-added thin films with x = 0.000, 0.015, and 0.025 are also shown in Fig. 1. Comparing the AFM images of specimens with and without Zn indicates that both, the Zn-free and Zn-added films, have grains with a spherical morphology; however, adding Zn results in a smooth surface with much finer grains (Table 1). On the other hand, the thickness of each film series does not substantially change, varying within 400–450 nm for the Cu-doped and 270–300 nm for the (Zn,Cu) co-doped films (Fig. 2d and Table 1).
The crystal structure of the annealed Zn-free and Zn-added films was analyzed by XRD, and the results are shown in Fig. 3. As presented in Fig. 3a and b, all films are well-crystallized after heat treatment in air at 400 °C. The high crystallization of the LFD nickel manganite-based thin films after annealing at a moderate temperature has been discussed elsewhere [16, 17]. A tetragonal spinel structure is revealed for films without Cu (x = 0.000), regardless of the presence of Zn in their compositions. However, the (221) tetragonal spinel peak for Zn-added film shifts to a slightly higher angle, as observed in Fig. 3c, which can be attributed to the replacement of Mn2+ ions with Zn2+ ions at the tetrahedral sites of the spinel. As the ionic radius of the Zn2+ cation (0.58 Å) is smaller than that of the Mn2+ cation (0.66 Å) [13], this replacement decreases the lattice parameter in the spinel phase, thus shifting the corresponding peak. On the other hand, cubic symmetry (Fd–3 m space group) appears when x≥0.010 for both systems (referring to JCPDS card No. 84–0542), without any secondary phases. This finding demonstrates that Cu, rather than Zn, induces the transformation from the tetragonal to the cubic spinel structure in the present (Ni,Mn)3O4-based specimens. The (311) peak shows the highest intensity for all cubic spinel patterns, clearly indicating the preferential orientation of the films. The XRD pattern characteristics and changes in the intensity are very similar as x increases in both systems, even though the adding Zn has been reported to improve the crystallinity in the case of spin-sprayed nickel manganite thin films [19]. Moreover, the peak intensity uniformly increases with increasing x, demonstrating that Cu increases the crystallinity.
3.2. Electrical properties
Figure 4a depicts the (ρ)–temperature (T) curve in the testing temperature range T = 295–368 K (22–95 °C) for the annealed Zn-free samples. This figure clearly shows that the resistivity decreases with increasing testing temperature, demonstrating a typical NTCR feature for the films. On the other hand, the lnR/T and reciprocal of the absolute temperature (1000/T) in Fig. 4b clearly present a nearly linear relationship over the measured temperature range. This linear dependence should be attributed to the electron hopping conduction mechanism, for which the R–T relationship is commonly described by the well-known expression: R = CTexp(Ea/kBT), where C is a constant, and T, Ea, and kB are the absolute temperature, activation energy for electrical conduction, and Boltzmann constant, respectively [3, 5]. Results similar to those in Fig. 4 were confirmed for all the annealed Zn-added films.
Figure 5a presents the RT resistivity (ρ) of films examined as a function of the Cu doping level, x. For the Zn-added system, the resistivity rapidly decreases to a minimum value of 880 Ω cm when x is increased from 0 to 0.02, and then increases slightly with further increases in x. In a similar manner, the Zn-free films also show a decrease in ρ with x ≤ 0.025 and a tendency to increase with x ≥ 0.03. The lowest ρ value of 200 Ω is achieved when x = 0.025. Compared with the Zn-free films at x = 0.000, an increase in resistivity from 6,300 Ω cm to 9,210 Ω cm is observed for the Zn-added films. Although the resistivity does not change much, this result clearly indicates that Zn doping deteriorates the conductivity. The Hall carrier concentration and mobility for the Zn-free films, as presented in Table 2, show that the carrier concentration systematically increases, whereas the carrier mobility decreases with Cu in the films. Therefore, the tendency of ρ to decrease in Fig. 5a should be attributed to the carrier concentration rather than the carrier mobility. On the other hand, the mobility of the charge carriers decreases in the Cu-rich compositions, possibly because the ionized Cu atoms act as scattering centers at a high carrier concentration, thereby decreasing the carrier mobility. Consequently, we suggest that the increased electrical conductivity of the Cu-doped films is mainly due to changes in the carrier concentration, caused by the combined effects of the composition and crystal structure.
Table 2
Carrier concentration and carrier mobility of Zn-free films
x | Carrier concentration (cm− 3) | Carrier mobility (cm2/V·s) |
0.000 | 4,44 × 1013 | 29.0 |
0.010 | 8.57 × 1013 | 25.4 |
0.015 | 1.37 × 1014 | 25.5 |
0.020 | 2.87 × 1014 | 24.7 |
0.025 | 7.35 × 1014 | 20.4 |
0.030 | 8.79 × 1014 | 16.0 |
In Cu-doped nickel manganite spinels, monovalent copper (Cu+) is known to reside at the tetrahedral sites, while bivalent copper (Cu2+) occupies both, the octahedral and tetrahedral sites of the spinel [20, 21]. The presence of Cu+ increases the amount of cations in the tetrahedral sublattices, which in turn displaces some Ni2+ cations from tetrahedral to octahedral sites. This process produces Mn4+ at octahedral sites to maintain the electrical neutrality, consequently increasing the Mn4+concentration. As more Mn4+ ions are generated, a larger number of sites are available for electron hopping, which could be the possible reason for the improved carrier concentration in Table 2, which contributes to the electrical conductivity of the present two types of films.
Many researchers have suggested that Zn2+ is located only at tetrahedral sites [10, 11, 22]. Thus, it is reasonable to infer that the Cu cations are expelled from tetrahedral sites and move into octahedral sites in the (Zn,Cu) co-doped compositions. This process changes Cu+ into Cu2+; consequently, the Cu+ cation concentration at the tetrahedral sites decreases. Moreover, as Cu+ changes into Cu2+, Mn4+ transforms into Mn3+ to maintain the electrical neutrality of the lattice [13], which eventually increases the electrical resistivity of the Zn-added films, as shown in Fig. 5a. The much finer grain size with added Zn, as shown in Fig. 1, should be also taken into account to better explain the differences between the conductivities of these two film systems.
Cu doping is found to decrease the absolute RT TCR from 3.21 to 2.38% K− 1 for the Zn-free system, as depicted in Fig. 5b. Clearly, Cu doping provides a significant decrease in the resistivity, but degrades the film sensitivity. A drop in TCR with Cu doping was also reported by Cho et al. [9] for spin-sprayed nickel manganite thin films, but the possible reasons for this behavior are still under investigation. On the other hand, the presence of Zn in the compositions is found to improve the TCR, which varies in the range of 2.82 to 3.54% K− 1 for the (Zn,Cu) co-doped system. The measured TCR values are even better than the results obtained for sputtered Ni–Co–Mn–O films, where the absolute TCR of ≥ 3% K− 1 was only achieved when the resistivity was higher than 1,000 Ω cm [23, 24]. Notably, in the present work, co-doping 0.01 mol Zn along with Cu remarkably enhances the TCR, whereas it only moderately deteriorates the conductivity. For instance, a high TCR of 3.02% K− 1 and a resistivity as low as 880 Ω cm are simultaneously achieved for the Zn-added film with x = 0.020. These excellent electrical features are interesting for infrared sensing applications such as microbolometers.
3.3. Electrical stability
Figure 6a shows the resistance drift (ΔR/Ro) for films with and without added Zn after aging at 150 °C in air for 500 h. As expected, a high resistance drift of up to 10.5% is revealed for the Zn-free samples, which is much larger than that for the (Zn,Cu) co-doped specimens, which was below 6.1%. This result confirmed the strong effect of Zn on the electrical stability. The most conductive Zn-added film (x = 0.020) possesses ΔR/Ro = 5.4%, whereas the lowest ΔR/Ro = 5.2% is attained for one film where x = 0.025. Figure 6b shows the detailed aging history of the x = 0.020 Zn-containing film, which demonstrates that the film undergoes a sudden increase in resistance during the first 82 h of aging. Subsequently, the increasing tendency of the resistance tends to slow down and reaches a nearly saturated state after 215 h, with a resistance drift of 5.2%. Up to 500 h of aging, the overall resistance drift is 5.4%.
The mechanism underlying the drift in the resistance of Cu-doped nickel manganite spinels has been interpreted as the Cu+ ions at tetrahedral sites easily oxidizing to Cu2+ under aging conditions (at temperatures ≤300 °C in air), which causes Cu ions to migrate from the tetrahedral to the octahedral sites [25]. This finding implies that the resistance drift is associated with a decrease in Cu cations at tetrahedral sites. To confirm this finding, the oxidation states of Cu before and after aging were verified by XPS analysis in Cu2p3/2 region. The XPS experiments were performed on the most conductive specimens for each system. The order of binding energies in the Cu2p3/2 region is Cu+ (Te) < Cu2+ (Oc) < Cu2+ (Te) [26, 27], as presented in Fig. 7. The Cu2p3/2 XPS characteristics are listed in detail in Table 3. Figure 7a1–2 distinctly shows that for the Zn-free film with x = 0.025, aging decreases the Cu+ ion concentration at the tetrahedral sites and increases the Cu2+ ion concentration at the octahedral sites (see also Table 3). In contrast, in the case of the (Zn,Cu) co-doped film with x = 0.020, neither the Cu+ nor Cu2+ ions change much after 500 h of aging, as can be observed in Fig. 7b1–2 and Table 3. These smaller variations in the amounts of Cu+ and Cu2+ ions can be attributed to the fact that Zn2+ ions almost exclusively occupy the tetrahedral sites, and thus, the Cu ions with a valence of only 2 + predominantly reside at the octahedral sites. Therefore, the migration of Cu ions from the tetrahedral to the octahedral sites would be less pronounced, which possibly explains the higher stability of these films. In addition, the smooth surfaces in these film surfaces may reduce the adsorption of oxygen, which hinders the oxidation of Cu+ ions. Notably, the aging performance of the (Zn,Cu) co-doped films in this study, which attain a resistance drift of 5.2–6.1%, is better than that of the undoped nickel manganite films prepared by spin-spraying and annealed at 400 °C in Ar, which exhibited a resistance drift of 6.5% for the same aging time [28]. However, this drift is still higher than those of the Mn–Co–Ni–O thin films annealed at 650–800 °C in air (1.7–5.2%) [29].
Table 3
Characteristics of the XPS Cu2p3/2 spectra of Zn-free (x = 0.025) and Zn-added (x = 0.020) films
Samples | Binding energy (eV) | Peak intensity (area %) |
Cu+ (Te) | Cu2+ (Oc) | Cu2+ (Te) | Cu+ (Te) | Cu2+ (Oc) | Cu2+ (Te) |
Zn-free* | 931.1 | 933.0 | 943.4 | 37.1 | 34.0 | 28.9 |
Zn-free# | 931.1 | 934.0 | 934.3 | 20.8 | 42.5 | 36.7 |
Zn-added* | 930.9 | 933.9 | 934.5 | 24.4 | 32.2 | 43.2 |
Zn-added# | 931.0 | 933.8 | 934.6 | 20.2 | 34.1 | 45.7 |
* Before aging. |
# After aging at 150 °C in air for 500 h. |
Te and Oc refer to tetrahedral and octahedral sites, respectively. |