Atomic Layer Deposition of NiO to Produce Active Material for Thin-Film Lithium-Ion Batteries

Atomic layer deposition (ALD) provides a promising route for depositing uniform thin-film electrodes for Li-ion batteries. In this work, bis(methylcyclopentadienyl) nickel(II) (Ni(MeCp)2) and bis(cyclopentadienyl) nickel(II) (NiCp2) were used as precursors for NiO ALD. Oxygen plasma was used as a counter-reactant. The films were studied by spectroscopic ellipsometry, scanning electron microscopy, atomic force microscopy, X-ray diffraction, X-ray reflectometry, and X-ray photoelectron spectroscopy. The results show that the optimal temperature for the deposition for NiCp2 was 200–300 ◦C, but the optimal Ni(MeCp)2 growth per ALD cycle was 0.011–0.012 nm for both precursors at 250–300 ◦C. The films deposited using NiCp2 and oxygen plasma at 300 ◦C using optimal ALD condition consisted mainly of stoichiometric polycrystalline NiO with high density (6.6 g/cm3) and low roughness (0.34 nm). However, the films contain carbon impurities. The NiO films (thickness 28–30 nm) deposited on stainless steel showed a specific capacity above 1300 mAh/g, which is significantly more than the theoretical capacity of bulk NiO (718 mAh/g) because it includes the capacity of the NiO film and the pseudo-capacity of the gel-like solid electrolyte interface film. The presence of pseudo-capacity and its increase during cycling is discussed based on a detailed analysis of cyclic voltammograms and charge–discharge curves (U(C)).


Introduction
In recent years, a large number of studies have been actively conducted to study the possibility of using nickel oxide (NiO) nanofilms as electrocatalysts for water decomposition [1], chemical sensors [2], active components of solar cells [3], and antiferromagnetic layers [4].NiO is extremely promising as an anode material for thin-film lithium-ion batteries [5].The value of the theoretical electrochemical capacity for NiO (718 mAh/g [6]) is close to the values for several other promising anode materials, such as SnO 2 (790 mAh/g) [7], CoO (715 mAh/g) [8], and MnO (650 mAh/g) [8].The capacity of these oxide materials is more than twice the capacity of graphite (372 mAh/g) used in the industry of lithium-ion batteries [9].The stability of NiO nanofilms during charge/discharge cycling at high current densities [10] allows us to consider NiO as promising material for creating negative electrodes of portable energy sources of biosensors, microchips, pacemakers, and other devices.[11].
NiO nanofilms are produced using various methods [12] such as thermal spraying, pulsed laser deposition, sol-gel, spin-coating, dip-coating, chemical vapor deposition, and atomic layer deposition (ALD).ALD is the most promising technology because it provides control over the thickness and purity of coatings with high precision, and can deposit uniform surface coatings on of complex shape and even porous and high aspect ratio substrates [13][14][15].This method could be a crucial factor for transition from 2D to 3D solid-state batteries (SSB), which are structured on 3D substrates with high aspect ratio instead of planar substrates.It could increase the energy density of SSB with the same thickness of the electrode to maintain the required conductivity of the layers [16].ALD is based on a realization of the sequence of chemical reactions between gaseous reagents and the surface species of the substrate, separated in time by purges with an inert gas to prevent uncontrolled reactions between the reactants and the reaction products.Because of the self-limiting nature, this method allows the deposition of films in a layer-by-layer fashion and the control of the thickness with high precision [13].
When selecting the deposition conditions via ALD, it is necessary to consider the stability and reactivity of the precursors.Many precursors have been tested for ALD NiO so far, but the most frequently used are shown in Table 1: nickel(II) acetylacetonate (Ni(acac) 2 ), bis(2,2,6,6-tetramethylheptane-3,5dionate)nickel(II) (Ni(thd) 2 ), bis(cyclopentadienyl) nickel(II) (NiCp 2 ), and NiCp 2 -based compounds such as bis(methylcyclopentadienyl) nickel(II) (NiMeCp 2 ), bis(ethylcyclopentadienyl) nickel(II) (NiEtCp 2 ), etc.  NiCp 2 and substituted nickel cyclopentadienyls have relatively high vapor pressure and can be used for ALD.When using these reagents, the deposition is performed in a wide range of temperatures (150-300 • C), while the growth rate per cycle varies in a range from 0.004 nm [47] to 0.12 nm [32].When water is used as a counter-reagent, an extremely low growth rate is usually observed (≈0.004 nm/cycle [47]), which is apparently because of the low rate of the chemical reaction.In this regard, the use of more reactive reagents, such as O 2 (including the effect of plasma) and O 3 is the most promising [15].Indeed, when using NiCp 2 /O 3 [29,32] and NiMeCp 2 /O 2 plasma [49], the growth per cycle can reach 0.12 nm and 0.048-0.084nm, respectively.Despite the large number of papers on the ALD of nickel and NiO thin films using NiCp 2 , there are no studies on the use of oxygen remote plasma as a counter-reagent.
Consequently, the purpose of this work is to study the process of atomic layer deposition of NiO nanofilms using NiCp 2 , NiMeCp 2 , and oxygen remote plasma as a counter-reagent and to determine the optimal ALD conditions.The films obtained under optimal conditions were tested as an active material for thin-film lithium-ion cells.

Materials and Methods
Monocrystalline silicon wafers (surface orientation 111, the diameter 40 mm, Svetlana Co., LTD, Saint-Petersburg, Russia) and stainless-steel plates (316SS, Tob New Energy Technology Co., LTD, Xiamen, China, diameter of 16 mm) were used as substrates.Before deposition, silicon and stainless-steel substrates were cleaned in an ultrasonic bath in acetone and deionized water for 15 min.The substrates were then immersed for 5 min in 10% hydrofluoric acid to remove the native oxide.Then, the substrates were cleaned using Piranha etch (70% H 2 SO 4 and 30% H 2 O 2 ) for 20 min to remove organic residues and generate hydroxyl surface species.Finally, the samples were rinsed in double deionized water and dried in inert gas atmosphere [55].
The ALD of NiO was performed by the commercial Picosun R-150 setup in the Functional Materials Laboratory, SPbPU, at the temperature range from 150 to 400 • C and reactor base pressure 8-12 hPa.Bis(methylcyclopentadienyl) nickel (II) (Sigma-Aldrich, 97%, St. Louis, MO, USA) and bis(cyclopentadienyl) nickel (II) (Sigma-Aldrich, 99%) were used as the nickel-containing precursors.Precursors were kept in stainless-steel bottles (Picohot TM 200, Picosun Oy, Espoo, Finland) and sublimated at temperatures from 90 to 170 • C. Pulse times varied from 0.5 to 15 s, and remote oxygen plasma was used as a counter-reagent (oxidizer).The power of plasma was 3000 W, and the frequencies were 1.9-3.2MHz.The total oxygen plasma processing time was 19.5 s (Ar purge-0.5 s, 40 sccm, Ar + O 2 plasma-14 s, 90 sccm, Ar purge-5 s).Before performing each stage of the nickel-containing reagent pulse reactor was purged with nitrogen and pumped to a pressure of 8-12 hPa.
The spectroscopic ellipsometry parameters (Ψ and ∆) for NiO films were measured with Ellips-1891 SAG ellipsometer (CNT, Novosibirsk, Russia) in the wavelength range 350-1000 nm.The data were measured at the incidence angle of 70 • .«Spectr V.1.10»software was used for the construction of structural-optical model function and fitting.The model consisted of a silicon substrate, SiO 2 native layer (thickness 3nm) and NiO layer.«Spectr V.1.10»database of refraction and absorption coefficients was used for model construction.Then the experimental spectra and calculated model functions were fitted.All parameters except NiO film thickness were fixed.Finally, fixation of refraction and absorption coefficients was removed and fitting was repeated.The thickness value obtained at the final fitting was considered as the most accurate.The accuracy of the NiO film thickness was estimated at about 0.3 nm, and the gradient of the thickness was calculated using Equation (1): where GT is the thickness gradient and T max and T min are the maximum and minimum film thicknesses, respectively.X-ray reflectometry (XRR) and x-ray diffraction (XRD) studies in surface sensitive grazing incidence XRD (GIXRD) modes were performed using a Bruker D8 DISCOVER (Cu-Kα, Billerica, MA, USA) high-resolution diffractometer.XRD were conducted in the range of 35-80 • with a step of 0.1 • and an exposure of 1 s at each step.The angle of incidence of the primary beam was 0.7 • .XRR measurements were made in the range of angles from 0.3 • to 10 • with the increment 0.01 using symmetric scattering geometry.The obtained results were processed by the Rietveld method using TOPAS 5 software (XRD) and by the simplex method using LEPTOS 7.7 (XRR).
Scanning electron microphotographs (SEM) showing planar views and cross-sectional views were obtained using the SEM Zeiss Merlin in the Nanotechnology Interdisciplinary Resource Centre, SPbSU (CARL ZEISS, Oerzen, Germany).The microscope spatial resolution was around 1 nm, and magnification was up to 200,000.SE2 regime with an accelerating voltage of 30 kV was used.A total of 3-4 random positions on the sample surface were scanned.
X-ray photoelectron spectra (XPS) were registered with a Thermo Fisher Scientific Escalab 250Xi spectrometer in the Physical Methods of Surface Investigation Resource Centre, SPbSU (Thermo Fisher, Waltham, MA, USA).Before performing the measurement, the surface layer of samples was etched by Ar + ions with an energy of 500 eV for 30 s to remove adventitious carbon.The samples were excited by Al-Kα (1486.7 eV) X-rays in the vacuum of 7 × 10 −8 Pa.The sample charging was automatically compensated.
For characterization of the sample surface topography, adhesion, phase contrast, and conductivity, atomic force microscopy (AFM, Dimension Edge, Brucker, Billerica, MA, USA) with an Icon (Brucker) microscope was used.A silicon cantilever coated on the tip with Pd/Pt was used, and the semi-contact mode was used with a cantilever amplitude of 50 nm and an oscillation frequency of 50 kHz for phase contrast mapping.This semi-contact mode registers the peak force (peak force nanomechanical mapping) for topography and adhesion study.The amplitude of the non-resonant oscillation of the cantilever was 50 nm, the oscillation frequency was 2 kHz, and a voltage of 1.7 V was applied for the conductivity maps study.
The samples of NiO deposited on the 316SS surface were used for electrochemical studies.These thin-film electrodes were used to make CR2032 coin cells.Lithium foil, polyolefin porous film 2325 (Celgard, Charlotte, NC, USA), and TC-E918 (Tinci, Guangzhou, China) solution were used as the counter electrode, separator, and electrolyte, respectively.The composition of TC-E918 was the 1M solution of LiPF 6 in a mixture of organic carbonates (ethylene carbonate, propylene carbonate, diethyl carbonate, ethyl methyl carbonate, vinylene).Coin cells were assembled inside an argon glove box OMNI-LAB (VAC).Cyclic voltammetry (CV) was performed using a potentiostat PGSTAT302N+ (Autolab, Utrecht, the Netherlands) in the range of 0.01-3 V with a scan rate of 0.5 mV/s.Cyclic charge/discharge was performed using the battery testing system CT-3008W-5V10mA (Neware, Shenzhen, China) at room temperature in the potential range of 0.01-3 V, current 40 µA.

Atomic Layer Deposition of NiO Thin Films
The necessary condition for correct ALD is the use of an excess of reagent vapors to saturate the surface during chemical reactions.This excess can be achieved by increasing the vapor pressure by heating the evaporator or/and by increasing the pulse time.For NiCp 2 , the vapor pressure at 100 • C is 320-370 Pa [56,57], while for Ni(MeCp) 2 , the vapor pressure at 73 • C is 133 Pa [58].These values are very small; therefore, to pulse the reagents in sufficient excess into the reactor, their heating is necessary.According to the literature, NiCp 2 is heated to temperatures from 40 to 150 • C [1,25,27,29,32,59,60].Such a large variation in temperatures is due to the technological differences in the device ALD setups.There are different types of reagent containers such as open source boats [29,36], stainless-steel bottles [61], and stainless-steel bubblers [47], which differ in their construction.Depending on the type of the container, the requirements for the reagent vapor pressure are different.If open source boats are used that are located inside the reactor, a lower temperature is required to evaporate reactants.For instance, NiCp 2 [29] and NiEtCp 2 [29] in open source boats are maintained at 40 • C [29,36] and 80 • C [29], respectively.
In this work, for the NiCp 2 and Ni(MeCp) 2 storage and pulse, we used stainless-steel bottles, which require the highest temperature of over 100 • C [60,61].We selected 90 • C as the first temperature of the container.At this temperature, the pulse of the reagent was not observed.A stable pulse was observed only at heating at 115 • C.
We studied films growths per cycle (GPC) at different container temperatures.In the case of NiCp 2 (Figure 1a), increasing the temperature of the evaporator from 115 to 145 • C leads to an increase in the GPC and an increase in the thickness gradient over the substrate area.According to the literature, the decomposition temperature of NiCp 2 is 186 • C [62] and is unstable with prolonged aging at high temperatures [63].Indeed, we found that with prolonged use of NiCp 2 in our containers at a temperature of 145 • C, its color changed from dark green to dark brown, thus forming a fine powder, indicating the decomposition of the reagent.For further work, we used the temperature of 115 • C, and to achieve an excess of reagent and saturation, we decided to increase the reagent pulse time.5 s, the GPC remained virtually unchanged.Thus, to saturate the surface of the substrate, applying a NiCp2 pulse for 5 s was sufficient.To determine the ALD window, the dependences of GPC on temperature were studied (Figure 2a).With an increase in temperature from 150 to 200 °С, an increase in GPC was observed, indicating an insufficient reactivity of the NiCp2 in this temperature range.In the range of 200-300 °C, GPC was remained practically unchanged and was approximately 0.011-0.012nm.This interval can be regarded as an ALD window.A further increase in temperature in the range of 300-400 °C led to an increase in the GPC of NiO, attributed to the partial decomposition of NiCp2.When using Ni(MeCp)2 as a reagent, an increase in the container temperature in the range of 100-130 °C did not affect the value of GPC (Figure 2b).With an increase in the container temperature in the range of 130-170 °C, a decrease in the GPC was observed.Furthermore, the color of the reagent changed from dark green to black, which was caused by the decomposition of Ni(MeCp)2 while operating in this temperature range.For further research, we used 130 °C.An increase in the pulse time from 1 to 3 s (container temperature (Tc)-130 °C, reactor temperature (Tr)-250 °C) did not affect the GPC; consequently, a pulse time of 1 s was used.When studying the effect of deposition temperature in the range of 200-350 °C, it was found that a noticeable film growth (GPC of about 0.01 nm) was observed in the range of 250-300 °C (Figure 2a).This GPC turned out to be lower than the observed GPC (0.048-0.084 nm) during the ALD of NiO films on W, Ru, Pt substrates, using direct oxygen plasma [49].Apparently, the low growth rate is associated with the lower reactivity of the removed plasma used in our process.To determine the necessary pulse time to achieve saturation, GPC dependence on the pulse time of NiCp 2 was studied.The experiments were performed at a reactor temperature of 300 • C. Increasing the pulse time in the range of 0.5-1 s sharply increased GPC.When the pulse time increased beyond 5 s, the GPC remained virtually unchanged.Thus, to saturate the surface of the substrate, applying a NiCp 2 pulse for 5 s was sufficient.
To determine the ALD window, the dependences of GPC on temperature were studied (Figure 2a).With an increase in temperature from 150 to 200 • C, an increase in GPC was observed, indicating an insufficient reactivity of the NiCp 2 in this temperature range.In the range of 200-300 • C, GPC was remained practically unchanged and was approximately 0.011-0.012nm.This interval can be regarded as an ALD window.A further increase in temperature in the range of 300-400 • C led to an increase in the GPC of NiO, attributed to the partial decomposition of NiCp 2 .
Coatings 2019, 9, x FOR PEER REVIEW 6 of 16 5 s, the GPC remained virtually unchanged.Thus, to saturate the surface of the substrate, applying a NiCp2 pulse for 5 s was sufficient.To determine the ALD window, the dependences of GPC on temperature were studied (Figure 2a).With an increase in temperature from 150 to 200 °С, an increase in GPC was observed, indicating an insufficient reactivity of the NiCp2 in this temperature range.In the range of 200-300 °C, GPC was remained practically unchanged and was approximately 0.011-0.012nm.This interval can be regarded as an ALD window.A further increase in temperature in the range of 300-400 °C led to an increase in the GPC of NiO, attributed to the partial decomposition of NiCp2.When using Ni(MeCp)2 as a reagent, an increase in the container temperature in the range of 100-130 °C did not affect the value of GPC (Figure 2b).With an increase in the container temperature in the range of 130-170 °C, a decrease in the GPC was observed.Furthermore, the color of the reagent changed from dark green to black, which was caused by the decomposition of Ni(MeCp)2 while operating in this temperature range.For further research, we used 130 °C.An increase in the pulse time from 1 to 3 s (container temperature (Tc)-130 °C, reactor temperature (Tr)-250 °C) did not affect the GPC; consequently, a pulse time of 1 s was used.When studying the effect of deposition temperature in the range of 200-350 °C, it was found that a noticeable film growth (GPC of about 0.01 nm) was observed in the range of 250-300 °C (Figure 2a).This GPC turned out to be lower than the observed GPC (0.048-0.084 nm) during the ALD of NiO films on W, Ru, Pt substrates, using direct oxygen plasma [49].Apparently, the low growth rate is associated with the lower reactivity of the removed plasma used in our process.When using Ni(MeCp) 2 as a reagent, an increase in the container temperature in the range of 100-130 • C did not affect the value of GPC (Figure 2b).With an increase in the container temperature in the range of 130-170 • C, a decrease in the GPC was observed.Furthermore, the color of the reagent changed from dark green to black, which was caused by the decomposition of Ni(MeCp) 2 while operating in this temperature range.For further research, we used 130 • C.An increase in the pulse time from 1 to 3 s (container temperature (Tc)-130 • C, reactor temperature (Tr)-250 • C) did not affect the GPC; consequently, a pulse time of 1 s was used.When studying the effect of deposition temperature in the range of 200-350 • C, it was found that a noticeable film growth (GPC of about 0.01 nm) was observed in the range of 250-300 • C (Figure 2a).This GPC turned out to be lower than the observed GPC (0.048-0.084 nm) during the ALD of NiO films on W, Ru, Pt substrates, using direct oxygen plasma [49].Apparently, the low growth rate is associated with the lower reactivity of the removed plasma used in our process.
Under the considered ALD conditions, the growth rates for both NiCp 2 and Ni(MeCp) 2 precursors were ~0.01 nm/cycle.The determined value is in the interval from 0.004 nm/cycle [47] (counter-reagent H 2 O) to 0.12 nm/cycle [29] (counter-reagent O 3 ) observed for NiO deposition when Ni(Cp) 2 , and its derivatives were used as the precursors.Since the size of cyclopentadienyl and methylcyclopentadienyl ligands in precursor molecules are relatively high, and thus the concentration of adsorbed molecules on the surface during chemisorption due to steric hindrance [64] is not sufficient to provide the subsequent formation of NiO monolayer during the full ALD cycle.
Considering the reported values of GPC for different counter reagents (Table 1) and results of the present study as a first approximation we assume that GPC increases among the series of counter-reagents: H 2 O, O 2 remote plasma, O 2 direct plasma, and O 3 .However, a more detailed research is needed a since GPC depends on many factors [65]: precursor size, chemical activity, surface species, deposition temperature, etc.
Based on the obtained experimental data, NiCp 2 and Ni(MeCp) 2 decompose during prolonged exposure at high temperatures (more than 115 and 130 • C, respectively).However, NiCp 2 is more thermostable, because for 6000-7000 cycles of ALD (about 50-60 h), a reagent pulse was observed, comparing to Ni(MeCp) 2 decomposed after 2500-3000 cycles of ALD (20-25 h).In addition, when using NiCp 2 , there is a wider ALD window (200-300 • C).Given the better thermal stability and a wider range of deposition temperatures, the use of NiCp 2 is more preferable in comparison with Ni(MeCp) 2 for deposition of NiO nanofilms by ALD.

Structure and Composition of the Films Deposited on Silicon
Figure 3a shows a surface sensitive grazing incidence XRD (GIXRD) pattern of NiO film deposited on a silicon substrate using NiCp 2 and oxygen plasma at 300 • C as a result of 2300 ALD cycles.The XRD pattern clearly shows reflections (111), (002), (022), (311), corresponding to cubic Fm3m modifications of NiO, as well as the reflection (311) associated with the imperfection of a single crystal of silicon [29].The NiO cell parameter calculated by the Rietveld method was a = 0.416 nm, which is similar to the bulk value of NiO-0.418nm (ICSD file no.61324 (2006)).The high ratio of the intensity of reflections compared with the background indicates either the absence of the amorphous phase or its fraction is insignificant.The texture was refined using the Rietveld method and the March-Dollase parameter in the (111) direction was 0.79, which is close to 1 and it corresponds to low texturing.Lu et al. showed [29] that in the case of ALD using NiCp 2 , Ni(EtCp) 2 , and O 3 (Tr = 200-300 • C) on a silicon substrate (Si (100) with etched native SiO 2 layer), a cubic NiO phase is formed, but neither texture nor anisotropic strain distribution was observed.In addition, oriented growth of crystalline NiO on silicon with (in the present paper) or without a native layer of silicon oxide probably does not occur, unlike the cases when Pt [49], soda lime glass [20], MgO [51], and Al 2 O 3 [51] are used as substrates (Table 1).
According to ellipsometry and XRR (Figure 3b), the film is smooth (roughness less than 1 nm for a sample 28-30 nm in thickness) and consists of a homogeneous layer with a clear NiO/substrate interface and a slight thickness gradient (~2 nm) across the silicon substrate.The simulated curve (Figure 3b) was obtained by a simplex method using the software LEPTOS 7.7.The film density, determined on the basis of XRR curve treatment, is about 6.6 g/cm 3 (Table 2), which is close to the values for bulk NiO of 6.72 g/cm 3 [66].The linear appearance of the XRR curve is probably caused by the presence of a thin layer of SiO 2 on the surface of the silicon substrate.The χ 2 value for the modeled curve was 8.080 × 10 −3 .It should be noted that the density of nanofilms obtained by the ALD can differ significantly from the bulk phases because of steric factors (distribution of hydroxyl groups, size of the precursor) [67] as well as a larger volume fraction of defective areas (sub-surface layer).
XPS spectra of the deposited NiO films were measured to study the elemental composition, bonding states, and conformality.Prior to the XPS measurements, the as-deposited NiO films were etched for 30 s by argon ions to remove the outermost layers.Only nickel (~47%), oxygen (~41%), and carbon (~12%) were detected (Figure 4a).Ni/O ratio was 1.15.The excess of nickel and the presence of carbon may be due to the unreacted NiCp 2 molecules.No evidence of silicon was detected, so the resulting film can be considered continuous and conformal.crystal of silicon [29].The NiO cell parameter calculated by the Rietveld method was a = 0.416 nm, which is similar to the bulk value of NiO-0.418nm (ICSD file no.61324 ( 2006)).The high ratio of the intensity of reflections compared with the background indicates either the absence of the amorphous phase or its fraction is insignificant.The texture was refined using the Rietveld method and the March-Dollase parameter in the (111) direction was 0.79, which is close to 1 and it corresponds to low texturing.Lu et al. showed [29] that in the case of ALD using NiCp2, Ni(EtCp)2, and O3 (Tr = 200-300 °C) on a silicon substrate (Si (100) with etched native SiO2 layer), a cubic NiO phase is formed, but neither texture nor anisotropic strain distribution was observed.In addition, oriented growth of crystalline NiO on silicon with (in the present paper) or without a native layer of silicon oxide probably does not occur, unlike the cases when Pt [49], soda lime glass [20], MgO [51], and Al2O3 [51] are used as substrates (Table 1).According to ellipsometry and XRR (Figure 3b), the film is smooth (roughness less than 1 nm for a sample 28-30 nm in thickness) and consists of a homogeneous layer with a clear NiO/substrate interface and a slight thickness gradient (~2 nm) across the silicon substrate.The simulated curve (Figure 3b) was obtained by a simplex method using the software LEPTOS 7.7.The film density, determined on the basis of XRR curve treatment, is about 6.6 g/cm 3 (Table 2), which is close to the values for bulk NiO of 6.72 g/cm 3 [66].The linear appearance of the XRR curve is probably caused by the presence of a thin layer of SiO2 on the surface of the silicon substrate.The χ 2 value for the modeled curve was 8.080 × 10 −3 .It should be noted that the density of nanofilms obtained by the ALD can differ significantly from the bulk phases because of steric factors (distribution of hydroxyl groups, size of the precursor) [67] as well as a larger volume fraction of defective areas (sub-surface layer).XPS spectra of the deposited NiO films were measured to study the elemental composition, bonding states, and conformality.Prior to the XPS measurements, the as-deposited NiO films were etched for 30 s by argon ions to remove the outermost layers.Only nickel (~47%), oxygen (~41%), and carbon (~12%) were detected (Figure 4a).Ni/O ratio was 1.15.The excess of nickel and the presence of carbon may be due to the unreacted NiCp2 molecules.No evidence of silicon was detected, so the resulting film can be considered continuous and conformal.For a detailed study of the bonding states of the film Ni2p, O1s, and C1s spectra were analyzed (Figure 4b-d).Unfortunately, the quantitative chemical analysis of Ni-containing materials is challenging because of the complexity of the Ni2p peak shapes resulting from multiplet splitting, shake-up, and plasmon loss structures [68,69].The obtained Ni2p spectrum contains Ni2p3/2 and For a detailed study of the bonding states of the film Ni2p, O1s, and C1s spectra were analyzed (Figure 4b-d).Unfortunately, the quantitative chemical analysis of Ni-containing materials is challenging because of the complexity of the Ni2p peak shapes resulting from multiplet splitting, shake-up, and plasmon loss structures [68,69].The obtained Ni2p spectrum contains Ni2p 3/2 and Ni2p 1/2 peaks.Both of them can be well separated into five components using the Gaussian-Lorentzian function (Figure 4d).The most intensive peaks at 853.9 and 871.5 eV and corresponding satellite peaks (861.1 and 878.Thus, the film contains predominantly NiO.However, NiCp 2 residues (C-C, C-H, and Ni-C bonds) and Ni(OH) 2 /NiOOH oxide were also present in a small amount.
On the SEM images (Figure 5a,b) of the surface of the deposited NiO film, there were no visible structures or defects.The film was smooth, uniform, and conformal.The film thickness determined from cross-sectional SEM images varied in the range of 27-31 nm (Figure 5b), which was in good agreement with the thicknesses determined using spectroscopic ellipsometry and XRR.On the SEM images (Figure 5a,b) of the surface of the deposited NiO film, there were no visible structures or defects.The film was smooth, uniform, and conformal.The film thickness determined from cross-sectional SEM images varied in the range of 27-31 nm (Figure 5b), which was in good agreement with the thicknesses determined using spectroscopic ellipsometry and XRR.AFM topography (Figure 5c) and phase contrast (Figure 5d) images of the surfaces of NiO films, measures in different areas (0.5 µm × 0.5 µm, 1 µm × 1 µm, and 1.5 µm × 1.5 µm) were similar.The roughness values (Ra), determined for the 0.5 µm × 0.5 µm and 1.5 µm × 1.5 µm scan areas, were 0.32 and 0.34 nm, respectively.The calculated roughness is close to the values of the silicon substrate (0.30 nm) and the amorphous (<0.39 nm [41,48]) or crystalline (<1 nm [50]) films prepared by ALD.A larger roughness value (0.63 nm), determined by the XRR method, may be caused with a larger surface area of measurement.
AFM topography and phase contrast images contain two phases (light and dark areas with diameter 10-40 nm).The lateral size of these areas on the phase contrast images was smaller than on AFM topography (Figure 5c) and phase contrast (Figure 5d) images of the surfaces of NiO films, measures in different areas (0.5 µm × 0.5 µm, 1 µm × 1 µm, and 1.5 µm × 1.5 µm) were similar.The roughness values (R a ), determined for the 0.5 µm × 0.5 µm and 1.5 µm × 1.5 µm scan areas, were 0.32 and 0.34 nm, respectively.The calculated roughness is close to the values of the silicon substrate Coatings 2019, 9, 301 10 of 16 (0.30nm) and the amorphous (<0.39 nm [41,48]) or crystalline (<1 nm [50]) films prepared by ALD.A larger roughness value (0.63 nm), determined by the XRR method, may be caused with a larger surface area of measurement.
AFM topography and phase contrast images contain two phases (light and dark areas with diameter 10-40 nm).The lateral size of these areas on the phase contrast images was smaller than on the images of the topography.The observed phase contrast can be caused by the formation of the phases with different crystallographic orientations in substrate planes or by the presence carbon on the surface.

Morphology of the Films Deposited on Stainless Steel
To further study the electrochemical characteristics of ALD NiO films, we deposited films on stainless-steel substrates using NiCp 2 and O 2 plasma.The temperature of 300 • C and 400 or 2300 cycles were used.Figure 6 shows the AFM surface topography, adhesion force maps (characterizing the interaction between the AFM probe and the sample surface) and the conductivity images for stainless-steel substrates and substrates with NiO film deposited using 400 ALD cycles.The deposition of the film did not change the surface morphology (Figure 6a,b).There was only a slight smoothing of the topography.Thus, NiO film probably had a slight tendency to grow more intensively in the grooves of the substrate.A slight contrast on the adhesion map of the stainless steel can be caused by a roughness; in deeper areas, the interaction force between the AFM probe and the surface was higher (Figure 6c).After deposition of the NiO film, the lateral difference between the adhesive forces (i.e., the forces required to detach the probe from the surface) became more significant (Figure 6d).On the map of the adhesion forces of the sample with a NiO film, a structure with pronounced contrast and clearly defined boundaries was traced (lateral size about 50 nm), which apparently indicates the presence of two phases.In addition, the adhesion forces images of steel with a NiO film (Figure 6d) showed structures with clear contrast and borders (lateral size about 50 nm).The presence of two phases can be explained by the semi-crystal structure of the film, in which ordered crystalline phases are present along with amorphous regions [74].phases with different crystallographic orientations in substrate planes or by the presence of carbon on the surface.

Morphology of the Films Deposited on Stainless Steel.
To further study the electrochemical characteristics of ALD NiO films, we deposited films on stainless-steel substrates using NiCp2 and O2 plasma.The temperature of 300 °С and 400 or 2300 cycles were used.Figure 6 shows the AFM surface topography, adhesion force maps (characterizing the interaction between the AFM probe and the sample surface) and the conductivity images for stainless-steel substrates and substrates with NiO film deposited using 400 ALD cycles.The deposition of the film did not change the surface morphology (Figure 6a,b).There was only a slight smoothing of the topography.Thus, NiO film probably had a slight tendency to grow more intensively in the grooves of the substrate.A slight contrast on the adhesion map of the stainless steel can be caused by a roughness; in deeper areas, the interaction force between the AFM probe and the surface was higher (Figure 6c).After deposition of the NiO film, the lateral difference between the adhesive forces (i.e., the forces required to detach the probe from the surface) became more significant (Figure 6d).On the map of the adhesion forces of the sample with a NiO film, a structure with pronounced contrast and clearly defined boundaries was traced (lateral size about 50 nm), which apparently indicates the presence of two phases.In addition, the adhesion forces images of steel with a NiO film (Figure 6d) showed structures with clear contrast and borders (lateral size about 50 nm).The presence of two phases can be explained by the semi-crystal structure of the film, in which ordered crystalline phases are present along with amorphous regions [74].Because NiO is a semiconductor, its deposition leads to the expected reduction in current amplitude (surface electrical conductivity, Figure 6c,f).At the same time, the concentration of the regions with the maximum conductivity over the sample surface increases (Figure 6f).This change can positively affect the electrochemical activity of the coating.The presence of regions with increased conductivity (lateral size of about 50 nm) can be caused by the presence of carbon in the structure or on the film surface.Because NiO is a semiconductor, its deposition leads to the expected reduction in current amplitude (surface electrical conductivity, Figure 6c,f).At the same time, the concentration of the regions with the maximum conductivity over the sample surface increases (Figure 6f).This change can positively affect the electrochemical activity of the coating.The presence of regions with increased conductivity (lateral size of about 50 nm) can be caused by the presence of carbon in the structure or on the film surface.

Electrochemical Properties of the Films Deposited on Stainless Steel
The initial stainless steel (316SS) and stainless steel with NiO coating (Tr = 300 • C, 2300 ALD cycles) were studied with the use of cyclic voltammetry.During the investigation of stainless steel, a moderate current increase was observed (Figure 7a curves with circle and plus markers).The shape of cathode and anode curves of stainless steel with deposited NiO film (solid and dashed lines) differed from the initial substrate.The stronger increase of current in areas characterizing the performance of NiO (1.5-0.05V (anode area) and in 1.3-2.7 V (cathode area) was observed.During the first cathode scan of NiO film, a pronounced increment of current appeared in the 0.8-0.2V range that was caused [75] by two processes.The first was a partial reduction in organic solvents (ethylene carbonate, propylene carbonate, diethyl carbonate, ethyl methyl carbonate) and functional additives (vinylene carbonate) of electrolytes.The reduction process led to the formation of a gel-like solid electrolyte interface (SEI) film comprising LiF, Li 2 CO 3 , and other compounds on the active layer (NiO) [76].The second process was the NiO conversion reaction with lithium as shown in Equation ( 2) [77]: Coatings 2019, 9, x FOR PEER REVIEW 11 of 16 The initial stainless steel (316SS) and stainless steel with NiO coating (Tr = 300 °С, 2300 ALD cycles) were studied with the use of cyclic voltammetry.During the investigation of stainless steel, a moderate current increase was observed (Figure 7a curves with circle and plus markers).The shape of cathode and anode curves of stainless steel with deposited NiO film (solid and dashed lines) differed from the initial substrate.The stronger increase of current in areas characterizing the performance of NiO (1.5-0.05V (anode area) and in 1.3-2.7 V (cathode area) was observed.During the first cathode scan of NiO film, a pronounced increment of current appeared in the 0.8-0.2V range that was caused [75] by two processes.The first was a partial reduction in organic solvents (ethylene carbonate, propylene carbonate, diethyl carbonate, ethyl methyl carbonate) and functional additives (vinylene carbonate) of electrolytes.The reduction process led to the formation of a gel-like solid electrolyte interface (SEI) film comprising LiF, Li2CO3, and other compounds on the active layer (NiO) [76].The second process was the NiO conversion reaction with lithium as shown in Equation ( 2) [77]: Two peaks were observed on the anode curve.The most intense (maximum position 2.2 V) characterizes nickel oxidation (Equation ( 3)), which provided the main capacity of the coating under investigation.An appearance of the second maximum of current amplification at 1.6 V is related to additional pseudo-capacity which is caused by the partial decomposition of the gel-like lithiumcontaining film [78].The initial compounds (Ni, Li2O, SEI film) and reactions occurred for the first five anode scans are the same consequently the position of peaks and their intensity should remain the same.The almost complete absence of changes in anode curve form indicates stable performance and retention of electrical contact between NiO and the substrate during cycling.

Ni + Li2O
During the second scan the maximum cathode current increment shifts to higher potentials (≈1V), its intensity diminishes.Observed peak characterizes the interaction between lithium and NiO coated by SEI film, therefore, the reaction during the second scan differs from the first one.The calculated value of oxidation-reduction potential for NiO conversion reaction is equal to 1.79 V [79].Thus, a shift of cathode current maximum from 0.41V to 1V indicates a decrease of polarization.During subsequent scans (3-5), the shape of cathode curves was similar to the shape of the cathode curve observed for the second scan.That indicates the occurrence of the same processes during 2-5 scans (cycles of charge/discharge).
The shape of cathode and anode curves and their evolution during the first five cycles characterizing the behavior of the deposited NiO film in conversion reaction with lithium is similar to results reported earlier for bulk NiO [75,80].Two peaks were observed on the anode curve.The most intense (maximum position 2.2 V) characterizes nickel oxidation (Equation ( 3)), which provided the main capacity of the coating under investigation.An appearance of the second maximum of current amplification at 1.6 V is related to additional pseudo-capacity which is caused by the partial decomposition of the gel-like lithium-containing film [78] During the second scan the maximum cathode current increment shifts to higher potentials (≈1V), its intensity diminishes.Observed peak characterizes the interaction between lithium and NiO coated by SEI film, therefore, the reaction during the second scan differs from the first one.The calculated value of oxidation-reduction potential for NiO conversion reaction is equal to 1.79 V [79].Thus, a shift of cathode current maximum from 0.41V to 1V indicates a decrease of polarization.During subsequent scans (3)(4)(5), the shape of cathode curves was similar to the shape of the cathode curve observed for the second scan.That indicates the occurrence of the same processes during 2-5 scans (cycles of charge/discharge).
The shape of cathode and anode curves and their evolution during the first five cycles characterizing the behavior of the deposited NiO film in conversion reaction with lithium is similar to results reported earlier for bulk NiO [75,80].
The shape of the charge/discharge curves of the tested thin film is similar to the shape of curves characterizing the interaction of bulk NiO with Li + [75,80].Because the current increase in the anode area for initial stainless steel was negligible (Figure 7a), the discharge capacity of modified steel is related to processes occurring with NiO coating.The initial measured discharge capacity of NiO film was 50 µAh.Based on film thickness (28-30 nm), density (6.6 g/cm 3 ), and tested sample diameter (1.58 cm), the specific capacity was calculated and amounted to 1336-1379 mAh/g (≈20.4 µA/cm 2 ).The obtained value exceeded the theoretical value stated for NiO (718 mAh/g) [6,81].Similar discharge capacity values were reported earlier (1240 mAh/g [81] and 1256 mAh/g [82]), and excess can be caused by contributions of pseudo-capacity to the measured value.The estimated value of NiO film weight was 38 µg, and the specific discharge current was 1050 mA/g (0.8-0.9 C).With an increase of charge/discharge cycles from 6 to 223, the shape of observed curves U(C) was similar to each other but the discharge capacity augmented.
To understand the cause of the capacity increase, CV measurements were repeated after 223 cycles.The intensity of cathode maximum diminished (less lithium and NiO participate in conversion Equation ( 2)), and moved forward low voltages (0.8V)., thereby leading to an increase in polarization.When comparing the 5th and 224th cathode scan curves, one can note the increase of current in 0.01-0.6V area that might be related to the insertion of lithium to SEI film.Although two anode peaks were observed at 224 discharge, almost in the same positions (1.7 V and 2.2 V), the ratio of maximum intensities changed.The amplitude of current at 1.7 V became more pronounced and surpassed the amplitude of the current at 2.2 V. Therefore, augmentation of discharge capacity during cycling was caused by the increase of pseudo-capacity provided by gel-like lithium-containing SEI film.

2.
The optimal temperature range for ALD using NiCp 2 was 200-300 • C, while using Ni(MeCp) 2 , the range is much narrower and is 250-300 • C. Growth per cycle for both precursors was 0.011-0.012nm.Thus, NiCp 2 is more thermally stable than Ni(MeCp) 2 , has a wider ALD window, and as a result is more suitable for the deposition of NiO films.
The film contains predominantly NiO.However, carbon as cyclopentadienyl residues and Ni(OH) 2 /NiOOH oxide are also present in small amounts.

5.
The specific discharge capacity (1336-1379 mAh/g) of nickel oxide films deposited on steel substrates significantly exceeded the theoretical capacity of bulk NiO (718 mAh/g).Moreover, with repeated cycling, capacity increased.Based on the experimental data, we assume that such a high capacity is formed by the conversion capacity (reduction and oxidation of NiO) and pseudo-capacity due to the formation of a gel-like SEI film.

Figure 1 .
Figure 1.The change of growth per cycle (GPC) with: (a) an increase in the evaporation temperature of NiCp2; and (b) pulse time.Substrate temperature was 300 °С.Error bars indicate the thickness gradient.

Figure 2 .
Figure 2. GPC as a function of the deposition temperature for NiCp2, Ni(MeCp)2, and O2 plasma (a) and GPC as a function of the Ni(MeCp)2 container temperature (b).Error bars indicate the thickness gradient.

Figure 1 .
Figure 1.The change of growth per cycle (GPC) with: (a) an increase in the evaporation temperature of NiCp 2 ; and (b) pulse time.Substrate temperature was 300 • C. Error bars indicate the thickness gradient.

Figure 1 .
Figure 1.The change of growth per cycle (GPC) with: (a) an increase in the evaporation temperature of NiCp2; and (b) pulse time.Substrate temperature was 300 °С.Error bars indicate the thickness gradient.

Figure 2 .
Figure 2. GPC as a function of the deposition temperature for NiCp2, Ni(MeCp)2, and O2 plasma (a) and GPC as a function of the Ni(MeCp)2 container temperature (b).Error bars indicate the thickness gradient.

Figure 2 .
Figure 2. GPC as a function of the deposition temperature for NiCp 2 , Ni(MeCp) 2 , and O 2 plasma (a) and GPC as a function of the Ni(MeCp) 2 container temperature (b).Error bars indicate the thickness gradient.

Figure 3 .
Figure 3. GIXRD (a) and XRR (b) patterns for NiO thin film deposited on silicon substrates using NiCp2 and O2 plasma at 300 °C during 2300 ALD cycles.

Figure 3 .
Figure 3. GIXRD (a) and XRR (b) patterns for NiO thin film deposited on silicon substrates using NiCp 2 and O 2 plasma at 300 • C during 2300 ALD cycles.

Figure 5 .
Figure 5. Plan-view and cross-sectional SEM images (a,b), AFM topography and phase contrast images (c,d) of NiO nanofilms deposited by ALD on silicon substrates using NiCp2 and O2 plasma at 300 °C during 2300 ALD cycles.

Figure 5 .
Figure 5. Plan-view and cross-sectional SEM images (a,b), AFM topography and phase contrast images (c,d) of NiO nanofilms deposited by ALD on silicon substrates using NiCp 2 and O 2 plasma at 300 • C during 2300 ALD cycles.

Figure 6 .
Figure 6.AFM 3D surface topographies adhesion maps (a,b); conductivity maps (c,d); and surface electrical conductivity (e,f) of stainless-steel substrate and NiOx nanofilm deposited on a stainless-steel substrate.

Figure 7 .
Figure 7. Cyclic voltammograms of stainless steel 316SS and stainless steel with NiO film (a), and charge/discharge curves of stainless steel with NiO film (b).

Figure 7 .
Figure 7. Cyclic voltammograms of stainless steel 316SS and stainless steel with NiO film (a), and charge/discharge curves of stainless steel with NiO film (b).
. The initial compounds (Ni, Li 2 O, SEI film) and reactions occurred for the first five anode scans are the same consequently the position of peaks and their intensity should remain the same.The almost complete absence of changes in anode curve form indicates stable performance and retention of electrical contact between NiO and the substrate during cycling.Ni + Li 2 O → NiO + 2Li + + 2e − ,

Table 1 .
Atomic layer deposition of NiO using different reagents.
Note: 1 mesoporous chromium terephthalate; 2 nanotubes; 3 carbon nanotubes; A: the reflexes related to crystalline NiO are absent on XRD diagram; C: presence of at least one reflex related to cubic phase of NiO; and Cryst: crystal phase not disclosed.

Table 2 .
Results of XRR data treatment for sample deposited using NiCp 2 at 300 • C and 2300 ALD cycles.

Table 2 .
Results of XRR data treatment for sample deposited using NiCp2 at 300 °С and 2300 ALD cycles.