Optical and structural properties of nitrogen incorporated Ni doped diamond-like carbon thin films

Co-depositing of nickel (Ni) and nitrogen (N2) in the diamond-like carbon (DLC) matrix was synthesized by pulsed laser deposition (PLD) technique. In work presented here, N2 was injected as a reactive gas to optimize the optical, and structural properties of Ni:N-DLC thin films, although N2 to argon (Ar) gas ratio was kept constant. Statistical properties of thin films were calculated from the surface topographic images of 2Ni:N-DLC, 3Ni:N-DLC, and 4Ni-N-DLC samples. The characterization of samples has been carried out by by atomic force microscopy (AFM), Raman spectroscopy, X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), and ultraviolet–visible spectroscopy. XRD analysis shows the crystalline nature of deposited thin films. Band gap energies of 2Ni:N-DLC, 3Ni:N-DLC and 4Ni:N-DLC has illustrated that the background N2 gas can be used for the optimization of the electrical properties of Ni-DLC thin films. Optical properties showed that 2Ni:N-DLC and 3Ni:N-DLC can be used for electronic devices.


Introduction
Diamond-like carbon thin films were grown with different Techniques (Dai et al. 2018;Santiago et al. 2019;Zhou et al. 2016;Ghadai et al. 2021;Zarei and Momeni 2022). Pulsed laser deposition is more efficient due to producing high energy excited species in the plasma plume and making the change in the structural and chemical properties of the film that depends on the laser parameters and other variables (Qiu et al. 2017;Salah et al. 2016).
As known, DLC thin film was classified as an electrical insulation layer among the carbonous materials. However, scientists found that DLC has strong hardness, low fractional coefficient, chemical stability, biocompatibility, non-toxicity, heat conductivity, and superior anti-scratch properties, they hope to improve the electrical conductivity of DLC thin films. Diamond-like carbon thin films have been widely used in implantology, electronics, automotive industry, and sensors (Das et al. 2018;Písařík et al. 2018;Ferreira et al. 2018).
Metal doped diamond-like carbon have been studied in different papers. Cu-doped diamond-like carbon was widely used due to its anti-corrosion properties. The excellent corrosion resistance of Cu-DLC thin films to iron substrate made it super anti-corrosion thin films (Khamseh et al. 2018). The decrease of hardness/Young's modulus of the DLC film structure can be obtained by Si-doped DLC film which reduces the sp 2 C-C cluster and I D / I G ratio in Raman spectrum. It can be used in magnetic recording media and sliding bearings. The increase of the hardness/Young's modulus can be accessed by Fe-doped DLC films result in the increase of sp 3 content in DLC film structure. Fe-DLC thin film is useful for nonconductive tool coating due to its hardness and good adhesion (Ray et al. 2016).
Nitrogen-doped diamond-like carbon thin films have attracted great interest due to their remarkable electrical properties, and high optical transmittance (Son et al. 2017). Chen Zeng et al. used the direct-current magnetron sputtering technique to grow nitrogen incorporated DLC thin films with gradient decreased and increased N doping. They found that the gradient decreased N-doped diamond-like carbon posses the highest hardness (Zeng et al. 2017). The modification of optical and mechanical properties of N-doped DLC which was prepared by anode source was studied by Zarei Moghadam and the co-authors. They found the variation of bandgap value from 2.61 eV to 2.39 eV by increasing N 2 flow rate from 5 to 50 sccm (Moghadam et al. 2019). The mechanical properties of N-doped DLC were studied by Chen Zeng and the co-authors. They found that the friction coefficient of N-doped DLC varied from 0.04 to 0.01 by increasing N 2 flow rate from 5 to 50 sccm (Zeng et al. 2017). The influence of transition metal doping on the tribological properties of DLC films has been studied by Gayathri et al. in 2015. Ag, Ni, and Ti dopants had 56%, 61%, and 57% sp 3 content, respectively (Gayathri et al. 2015).
Previous reports have been shown various co-doping such as B:N-DLC, and Hf:N-DLC which lead to a significant increase in the electrical conductivity of diamond-like carbon films (Son et al. 2017;Qi et al. 2017). In this study, we focus on the co-doping of N 2 and Ni into the DLC matrix, and the influence of this co-deposition in the optical and structural properties of DLC thin films. Early works showed that Ni and N 2 co-doping were not discussed anymore. Additionally, some statistical properties have been derived from atomic force microscopy (AFM) images.

Experimental
The films were grown on Si (100) substrates which were rinsed for 10 min by methanol, acetone, and distilled water. Then, substrates (1 cm × 1 cm) have been submerged into the Hydrofluoric acid (HF, 2%) solution three times for 1 min for removing the probable oxide layers prior to placement in the deposition chamber. All substrates were mounted onto a static plate. each Pure Ni strip (99%) with the size of 0.5 cm × 5 cm placed tangentially ( Fig. 1) on the highly pure graphite target (99.9%, 2 inches). Nd:YAG laser (λ = 1064 nm$, E max = 250 mj, PulseWidth = 12 ns) with 10 Hz repetition rate and 200 mj energy was used to deposit Ni-DLC thin films. The rotation of the target holder with the speed of 1.5 rev/ min was due to the homogeneous thin film deposition. 6300 laser shots stroke the designed target in 14 min. The target to substrate angle and distance was about 43 0 and 4 cm, respectively. The perched argon (Ar) gas was used for the pre-sputtering process.
In the experimental procedure, the presence of Ar gas was essential due to not producing nitride compounds on the surface of the graphite target. The N 2 and Ar plasma produce energetic nitrogen species, which can react easily to the surface of DLC films.
Various collision gases, Ar, and N 2 (Ar/N 2 ≈6) were bled into the deposition chamber through the gas valve, simultaneously. The pressure of backfill gas in the deposition chamber was 4 × 10 -5 mbar during film growth. The Initial pressure was about 4 × 10 -6 mbar. Different dopants of Ni strips were 2 (2Ni),3 (3Ni), and 4 (4Ni) strips as seen in Fig. 1. After film growth, the chamber was evacuated, and a strip was removed from the graphite target. Then, the deposition process has been continued for other samples.
Energy-dispersive X-ray spectroscopy (EDS) analysis has been done using MIRA III, TESCAN, Czech Republic. Raman spectroscopy was carried out by µRaman, Avantes (532 nm, 100% laser power), Netherlands. AFM micrographs were obtained using FemtoScan (Non-contact mode), Russia. UV-Vis spectra were recorded by Shimadzu, Japan. Grazing XRD analysis done using Philips PW1730, Netherlands. XPS device for analyzing was Bes-Tech, Germany.

Results and discussion
The topographic images of thin films (Fig. 2) were obtained by AFM analysis. The aggregation has been seen by increasing Ni dopant from 2 to 4Ni. Nanoparticles are spherically distributed inhomogeneously on the Si substrate as seen in Fig. 2. One of the statistical parameters, RMS roughness, describes the variety of height fluctuations on the surface. RMS roughness, average roughness, and surface skewness were computed using the two-dimensional discrete surface profile (AFM images) consisting of N × N (N = 512) points. Scientists showed that RMS roughness is not enough for describing the surface roughness (Petrik et al. 1998), but it is valuable to compare the RMS roughness of Ni:N-DLC thin films. Calculated results by Matlab software can be seen in Table 1.
The highest surface roughness value is for the 4Ni:N-DLC which is most likely caused by additional Ni strips. Scientists experience prove that the graphite target should be changed as often as possible due to the relation between the smoothness of the Ni:N-DLC thin films and the quality of the graphite target that limited our deposition procedure. Our experience proved that the lack of the substrate temperature in the experimental section caused the higher amounts of the roughness.
The positive sign of the skewness (R sk ) shows that the data points (the height of nanoparticles) are right-skewed toward the data average (Ramezani et al. 2020). The skewness values were reported in Table 1.
Film thickness was measured from the cross section of all three samples. These amounts for 2Ni, 3Ni, and 4Ni were 148.87 ± 5 nm, 167.21 ± 5 nm, and 189.13 ± 7 nm, respectively.
The structural properties of DLC thin films have been characterized by the Ferrari-Robertson equation. The size of graphite cluster (L a ) can be written as (Ferrari and Robertson 2004):  where C is ≈ 0.0055. I D /I G is the ratio of the D-line to G-line in the Raman spectroscopy.
As known: where the optical band gap (E g ) has been measured by Tauc equation (Khan et al. 2019): where h is Planck's constant and refers to absorption coefficient. The transformation of measured reflectance spectra to the corresponding absorption spectra can be done by applying the Kubelka − Munk function (Makuła et al. 2018).
Raman spectra of 2Ni:N-DLC, 3Ni:N-DLC, and 4Ni:N-DLC are shown in Fig. 3a-c. The vibration frequencies of N-DLCs have lied between 1300-1600 cm −1 for ring-like (1) Fig. 3 Raman spectra of a 2Ni:N-DLC b 3Ni:N-DLC, and c 4Ni:N-DLC molecules. Two common bands D-line and G-line are due to the breathing and stretching mode of C-C sp 2 bonds. The third peak which is assigned as 1 is attributed to the distorted C-C sp 3 stretching vibration (Chicot et al. 2010) appeared in Fig. 3a-c. The companion mode of 1 mode always appeared as 2 . It can be seen in Fig. 3b. It is obscured by the tail of the G-line in Fig. 3a-c. The G-line position in Fig. 3a (attributed to the Raman spectroscopy of 2Ni:N-DLC) is at 1504.8 cm −1 . It shows that this thin film is in the second stage of the Ferrari et al. report namely the N-C graphite phase. The G-line position is moved to higher wavenumbers in Fig. 3b, c. It indicates a transition from N-C graphite to a-C as Ferrari-Robertson trajectory (Ferrari and Robertson 2004). Although, Raman comparison (G-line) with one of the papers showed that it may possess crystalline structure (Pang et al. 2010). The shift toward lower wavenumbers decreases the tensile stress (Gayathri et al. 2015; Zarei and Momeni 2022). Full-Width Half-Maximum (FWHM) of G-line shows the distortion of the bond angle and bond length in C-C sp 2 bonds. In other words, FWHM illustrates the structural disorders and it shows the cluster size of C-C sp 2 sites (Gayathri et al. 2012). 2Ni:N-DLC sample with the highest FWHM G (Table2) indicates the small clusters or the formation of defects. the second sample (3Ni:N-DLC) is the most unstrained sample with larger sp 2 clusters or the lowest defects. I D /I G ratio can be a good estimation of sp 3 /sp 2 ratio. Additionally, it can determine the sp 2 cluster size. The lowest I D /I G ratio for 4Ni:N-DLC attributes to the increase in sp 3 amount in this thin film. All information of Raman spectra (Fig. 3a-c) is tabulated in Table 2.
As illustrated in Fig. 4, The RMS Roughness values of 2Ni:N-DLC, 3Ni:N-DLC, and 4Ni:N-DLC thin films were drawn in terms of I D /I G ratio. It is depicted that the statistical properties of diamond-like carbon thin films are related to the structural properties of these thin films.
The XRD pattern of Ni:N-DLC films were illustrated in Fig. 5. Three different structures have been seen and match with all diffraction peaks. The peak position for the diamond in 2Ni:N-DLC appears at 2ϴ = 75.34° which can be indexed as (220) plane of the diamond. XRD pattern of 3Ni:N-DLC contains two peaks located at 2ϴ = 42.44°, and 2ϴ = 44.10°. As seen in Fig. 6, the XRD traces relating to the 4Ni:N-DLC film show two distinct peaks for diamond structure at 2ϴ = 42.46°, 2ϴ = 44.10° (Xiao et al. 2019). NiO diffraction peak can be seen at 2ϴ = 78.64° in all three spectra ( 10-0280). Graphite and C-N peaks could not consider due to the signal-to-noise ratio. The average crystallite size is calculated by the Scherrer formula d = Kλ/Bcosϴ; K ≈ 0.89 which is the Scherrer constant; λ is the X-ray wavelength and B is related to the full width at half-height of the peaks. The crystallite size for Ni(111) for 2Ni:N-DLC, 3Ni:N-DLC, and 4Ni:N-DLC are 3.34 nm, 5.18 nm, and 5.70 nm, respectively. The dislocation density values were estimated using δ = 1/d 2 , where δ is dislocation density and d is the crystallite size (Naveena et al. 2022). The calculated values are 0.09, 0.04, and 0.03 lines/(nm) 2 for 2Ni:N-DLC, 3Ni:N-DLC, and 4Ni:N-DLC samples (Ni (111)). XRD shows the crystallite nature of all thin films. It is obvious that the band gap energy should be measured to found that what is the role of Ni 3 N and diamond structure in the electronic properties of the films. Before that, we must obtain the percent of nitrogen in the thin films which we could not calculate in EDS.  Figure 5a shows X-ray photoelectron spectroscopy analysis (XPS) for 2Ni:N-DLC. Inelastic background has been subtracted and Voigt function fitting was conducted in the C1s peak of the XPS spectra (Fig. 5b-d). As seen, the C1s spectra in 2Ni:N-DLC thin film were deconvoluted to five peaks at 284.63 eV (sp 2 C-C), 285.24 eV (sp 3 C-C), 286.31 eV (sp 2 C=N), 287.34 eV (sp 3 C-N), and 288.68 eV (C=O). The five deconvoluted C1s Increasing binding energy from 2Ni:N-DLC to 3Ni:N-DLC sample, maybe is due to bonding to more electronegative atoms. Five deconvoluted sub-peaks of the C1s XPS spectrum of the 4Ni:N-DLC were at 284.50, 285.19, 286.25, 287.41, and 288.78 eV, which corresponded to the sp 2 C-C, sp 3 C-C, sp 2 C=N, sp 3 C-N, and C=O, respectively. It is expected to have the more formation of sp 2 C-C bonds in N Incorporated samples in comparison with the nitrogen-free doped diamond-like carbon thin films (Zarei and Momeni 2022;Son et al. 2017). The N content was 8 at. % and it was approximately constant in the Ni:N-DLC films ( Table 3).
The sp 2 to sp 3 ratio can be obtained from the area ratio of the two peaks (sp 2 C-C, sp 3 C-C) which was decreased from 1.077 in 2Ni:N-DLC to 0.52 in 4Ni:N-DLC. Figure 5e demonstrates N1s peak of 2Ni:N-DLC sample has an approximately good coincidence with other papers (Zoua et al. 2005). Ni XPS peak can not be deconvoluted to other distinct peaks due to low intensity. There are some contradictions in the Raman spectroscopy and the XPS analysis of N corporated doped DLC thin films.
The relation of optical band gap E g with the absorption coefficient is expressed by the Tauc relation in Eq. (3). Indirect electronic transition in metal:N-DLC films was reported before in the different papers. Corresponding bandgap can be obtained by plotting (F(R).h ) 1∕2 vs. h for allowed indirect transition and prolong the linear portion of it to F(R) =0 value. It is known as the Kubelka − Munk function plot which can be seen in Fig. 7. It is evident that the least bandgap is about E g = 1.64(eV) for 2Ni:N-DLC sample. The Bandgap for 3Ni:N-DLC and 4Ni:N-DLC are E g = 1.81(eV) and E g = 2.30(eV), respectively. The plot covers the change in bandgap (Fig. 7) for different atomic percent of Ni in the N-DLC matrix. From Fig. 8, the bandgap increased gradually with the increase of Ni strips which greatly matches with Raman spectra (Fig. 3). It is shown that the variations in the number of sp 2 C-C clusters have a good relationship with the bandgap. Hence, a plot of I D /I G versus 1/E g 2 can be useful. In Fig. 8, the increased Ni dopant result in decreasing 1/ E g 2 and I D /I G values. As illustrated in Fig. 8, we can fit a linear line for five or more points. The size of the graphite cluster is derived from Eq. 1. The L a values for 2Ni:N-DLC, 3Ni:N-DLC, and 4Ni:N-DLC are about 0.762 nm, 0.709 nm, and 0.586 nm by sequential.

Conclusion
In summary, we changed the target arrangement in PLD deposition technique to optimize the optical, and the structural properties of crystallite Ni:N-DLC thin films. As seen, our calculation indicated that I D /I G was low enough, but this parameter was not a good estimate for describing sp 2 /sp 3 bonds in nitrogen incorporated doped diamond-like carbon films. The ratio of the sp 2 /sp 3 amount from XPS analysis indicated high quality Ni:N-DLC samples with about 37%-51% sp 3 bonds. Ni doped diamond-like carbon thin films were insulator (Zarei and Momeni 2022). The injection of the constant amount of N 2 gas in the chamber was due to having an adamant semiconductor which can be used in electronic devices. As a result, the optical band gap increased by elevating Ni dopants. 2Ni:N-DLC has a large intrinsic optical band gap in an order of the 1.64 eV that is about the band gap of other semiconductors like MoS 2 thin film. The size of the graphite cluster was the other measurement which was the least for 4Ni:N-DLC sample with the highest dopant. By the comparison of graphite cluster size, XPS spectra, and the band gap energy, E g experienced a decrease by increasing the size and amount of sp 2 clusters. As XPS results indicated, by increasing the incorporation of oxygen in the structure of thin films, XPS peak shift can be observed for 3Ni:N-DLC sample. By comparison of XRD analysis and Band gap value, it is evident that additional Ni dopant in 4Ni:N-DLC enhanced the sp 3 C fraction of DLC. In addition, 2Ni:N-DLC and 3Ni:N-DLC were good semiconductors in nature due to existing dominant Ni 3 N bonds. The statistical analysis indicated a rough surface with right-skewed height data. It was evident that most data fell to the right of average value. This study is continuing.