Effect of chromium doping on the structure and mechanical properties of anti-wear TiB2 coatings

TiB2-based coatings have been intensively developed due to their physical and mechanical properties, including excellent thermal stability and high hardness with good abrasion and corrosion resistance, which appear to be the most beneficial in industrial application. Previous investigations have shown that doping TiB2 with W, Ni and C can significantly reduce residual stresses and improve adhesion, making these coatings ideal on tools to machining aluminum alloys. The aim of this study was to analyze the effect of an Cr interlayer on the durability (adhesion) of the fabricated Ti1−xCrxB2 (x = 0; 0.03; 0.06; 0.10) films and determine the influence of Cr on their microstructure and mechanical properties. The structural characterization of Ti1−xCrxB2 coatings was carried out using X-ray diffraction, X-ray photoelectron spectroscopy, transmission electron microscopy and atomic force microscopy. To investigate the mechanical properties, nano-scratch and—hardness tests (NST, NHT) were performed, and fracture toughness of the substrate layer systems was determined. The use of an adhesive layer of pure Cr increased the adhesion of the coatings to the substrate. It is shown that the changes in Cr content not only affect the microstructure, mainly by decreasing the crystallite size (column width), but also the texture (preferred film orientation) and phase composition. The addition of chromium also has an effect on the mechanical properties of TiB2 films by reducing their hardness and Young’s modulus and increasing their fracture toughness (KIC).


Introduction
The dynamic development of materials engineering techniques in the field of surface coatings deposition processes for tool components makes it necessary to search for the simplest solutions (structures and compositions) ensuring stable and reliable operation for the longest possible time.
Such a trend is clearly observed in the production of antiwear coatings dedicated for common machining, but also for use in automotive industries [1], gas turbines [2] and metal-cutting tools [3]. Among commonly used hard antiwear coatings, TiN still holds a dominant industrial position, although its adhesion to metallic substrates is frequently shown to be weaker than for borides and carbides [4]. Nevertheless, nitrides have been regarded as more promising in this respect as they are softer and more ductile than borohydrides and carbides, with lower melting points, but higher chemical stability. Recently, new TiB 2 -based coatings have been intensively developed due to their potential physical and mechanical properties, such as high melting point, low compressive residual stress state [5], good abrasion [6], increased hardness [7], corrosion resistance, high friction resistance [8] and excellent thermal stability. However, surface defects (droplets and nodular defects) observed on hard coatings [9] deposited on machining tools often serve as crack initiation sites in the substrate during operation, leading to complete failure of the entire tool. This is unacceptable for coating materials, especially those used in antiwear systems. Therefore, ensuring a proper stress distribution, in addition to attaining high quality coatings, should be 1 3 80 Page 2 of 13 addressed. The best and simplest solution that significantly improves the adhesion of high-strength coatings and simultaneously relaxes the generated tensile stresses is the application of an interlayer as a transition zone. This solution is promising as it enables the consideration of coating materials that would otherwise be disqualified due to their stiffness and brittleness, while also significantly improving film-substrate adhesion [10]. Furthermore, it has been observed that the transition interlayer can divert or even stop discontinuities/cracks propagating through the coating. Although this may consequently lead to coating delamination, it protects the core/substrate under wear conditions. This mechanism is a principal approach used in the construction of layered composites [11]. High strength anti-wear coatings have been the subject of many studies over the past years [12]. They remain challenging and reveal interesting properties especially due to the use of different compositional configurations or changes in stoichiometry.
Ali et al. [13] deposited Ti as a sublayer for TiB 2 -Al 2 O 3 coatings to enhance the adherence between the coating and the substrate and also to increase the Young' modulus.
The effect of nitriding an H13 steel substrate on the hardness and adhesion properties of TiB 2 coatings was analyzed by Park et al. [14]. After the process, an Fe 4 N phase was identified in the steel substrate. Such surface treatment noticeably improved the adhesion of TiB 2 coatings (increase to ∼30 N). Ozkan [15] described the friction and wear enhancement of magnetron sputtered bilayer Cr-N/ TiB 2 coatings. He showed that combined bilayer ceramic and metallic coatings can be used in mechanical systems such as the moving parts of an internal combustion engine as lubricated conditions to improve the tribological behaviors of the steel surfaces.
A bilayer coating of ~ 3.4 μm thickness, consisting of a nanocrystalline TiB 2 protective top layer and a TiN diffusion-barrier bottom layer deposited by chemical vapor deposition (CVD) method, was analyzed by Gruber et al. [16]. They used a TiN coating to primarily serve as a barrier against B diffusion into the substrate, which would otherwise compromise coating adhesion to cemented carbide cutting tools. The coating TiN succeeded in this role.
Besides the adhesion to the substrate, the next problem in TiB 2 coating is the low cracking resistance. These properties can be altered by the addition of certain elements to the TiB 2 structure. In the literature, the authors proposed a Ti-B-M system where M = Ni, Cr, Zr, C, W. A TiB 2 coating with 0-12.6 at% Ni contents was deposited using the magnetron sputtering (MS) method by Wang et al. [17]. They demonstrated that additions of Ni to form the cemented boride nanostructure could noticeably toughen the TiB 2 coatings. E. Contreras et al. [18] analyzed the TiB 2 coatings with doping of carbon. The doping of carbon in the TiB 2 led to a progressive reduction in the hardness and Young's modulus.
In our previous work [19], we presented that an addition of tungsten to a TiB 2 coating produced by MS had a considerable impact on the microstructure and increased the brittle cracking resistance.
Several classic chemical vapor deposition (CVD) [20] and vapor assisted techniques [21], including physical vapor deposition (PVD) [22], are used to deposit TiB 2 coatings. However, one of the most versatile deposition methods remains magnetron sputtering (MS) [23], which allows precise control of the ion bombardment conditions during growth, thus enabling extensive modification and improvement of coating properties [24]. Conducting the deposition process under low temperature conditions, as done in the physical vapor deposition (PVD) method, produces TiB 2 coatings that show great potential in tribological applications.
This paper discusses the structural changes, chemical and phase composition differences, and analyzes the mechanical properties of Cr-doped TiB 2 coatings. A magnetron sputtering method was proposed for the deposition of Ti 1−x Cr x B 2 coatings on steel substrates. The deposition parameters were optimized based on our previous studies [7]. To change the residual stress and increase the adhesion of the TiB 2 coating to the substrate, we used a Cr transition metal coating as a sublayer. To reduce or eliminate the problem of low fracture toughness, it was proposed to fabricate and analyze coatings with several variants of Cr doping concentration: Ti 1−x Cr x B 2 (x = 0; 0.03; 0.06; 0.1).

Coating fabrication
Ti 1−x Cr x B 2 coatings with different chromium content x = 0; 0.03; 0.06; 0.1 were deposited by means of direct current magnetron sputtering (DCMS) at the Łukasiewicz Research Network-Institute for Sustainable Technology in Radom (Poland). A Twin direct current (DC) magnetron setup was used, equipped with disc-shaped targets made of pure chromium (Cr, ∅140 mm) and titanium diboride (TiB 2 , ∅ 100 mm), both fixed on copper pads. The substrates were made from W320 hot worked steel, tempered to 50 ÷ 52 HRC. Before the deposition process, the substrates were thoroughly washed with pure alcohol and etched with argon ions (voltage of − 800 V) to remove surface impurities. During deposition, the argon pressure in the vacuum chamber was 4 × 10 -3 mbar, and the substrate bias − 100 V was applied. The power on the Cr target was varied in the range of 70 ÷ 165 W to attain various Cr concentrations, whereas the power on the TiB 2 target was held constant at 2000 W. The deposition time of each coating was 40 min. Their thicknesses were in the range of 1.2 ÷ 1.8 μm depending on the Cr content.

Coatings characterization
The microstructure of the Ti 1−x Cr x B 2 coatings was investigated by means of transmission electron microscopy (TEM, model JEOL JEM-2010ARP, 200 kV) by preparing crosssectional lamella manufactured via focused ion beam (FIB). Selected area electron diffraction (SAED) patterns, obtained using an ∅ 150 nm aperture, were used for phase analysis. They were interpreted using JEMS software (P. Stadelmann, Switzerland). Phase identification was facilitated using the identified inorganic crystal structures database ICSD (FIZ Karlsruhe, Germany) [25].
Phase analyses of the coatings were performed by means of grazing incidence X-ray diffraction (GI-XRD) using a PANanalytical EMPYREAN DY 1061 diffractometer equipped with a Cu Kα lamp (0.154 nm), at an angle of 3°. The effective penetration depth was determined for GRAZ-ING using Eq. (1) [26]: where G x is the fraction of the incident beam absorbed by the material before the beam reaches the maximum penetration depth Z (nm), μ is the linear attenuation coefficient of the material, α is the incidence angle (°) and 2θ is the angle of the diffracted beam (°).
X-ray diffraction patterns were fitted using Highscore Plus software and crystallographic information files (CIF). Crystallite sizes D (nm) were determined using Scherrer's equation Eq. (2) [27]: where K is a dimensionless shape factor, with a value close to unity, λ is the wavelength of X-ray source (nm), β is full width at half maximum (rad), θ is the diffracted angle (°).
X-ray photoelectron spectroscopy (XPS) was used to determine the chemical states of the elements in the coatings using a PHI5000 Versa Probe II instrument equipped with a surface charge analyzer (SCA). The monochromatic Al Kα source worked at 15 kV. Argon was used as the protective gas.
The surface topography of the coatings was analyzed via atomic force microscopy (AFM, model Veeco Dimension Icon SPM) with a NanoScopeV controller.
The hardness of Cr-doped TiB 2 coatings deposited on W320 steel substrates was investigated by nanoindentation using a CSM tester with a Berkovich diamond tip. A maximum force of 8 mN was applied to not exceed 10% of the coatings' thickness. For this load at least 8 measurements were made.
Young's modulus E was calculated based on the results obtained, as well as the corresponding standard deviations. The measurement results enabled determination of the H/E as the plasticity index or load factor, which is responsible for the maximum elastic deflection, when the coating is not destroyed and the H 3 /E 2 as a resistance to plastic deformation, which determines the load capacity of the coating.
To determine the K IC , a 350 mN normal force was used for each coating to initiate cracks at the indentation sites. The cracks were documented via scanning electron microscopy. The lengths of the cracks were measured and based on the Laugier model formula (Eq. 2) [28], the value of K IC was determined: where K IC -fracture toughness (MPa·m 1/2 ); x v -indenter geometry factor (for Laugier Equation x v = 0.016); E-Young modulus of the coating (GPa); H-hardness of the coating (GPa); P-the indentation load (mN); a-the length from the center to the corner of the indent (μm); l-the length of the cracks (μm), c = l + a, the total length from the center of the indent to the end of crack (μm).
Scratch tests were performed using a MicroCombiTester. They were conducted using a Rockwell indenter with a tip angle of 120°. The scratch length was set at 5 mm with the indenter moving at a constant speed of 5 mm/min, and a normal load linear increase from 1 to 30 N.

Influence of Cr on the phase structure
GI-XRD was performed on the fabricated coating at an angle of 3°, for which the penetration depth of the X-rays was calculated to be 1.2 µm. Consequently, peaks originating from the substrate were found in the diffraction pattern at 2θ = 64.7° and 82.2°.
The remaining peaks were assigned to the coating material, i.e., a hexagonal TiB 2 phase (pattern number 04-002-5184) and a TiB 1.94 phase (pattern number 04-001-9272). The presence of a non-stoichiometric phase is visible at 2θ = 28.16° as a peak broadening for the (100) TiB 2 phase. The boron deficiency may be due to its segregation into nanocolumnar grain boundaries during deposition and forming phases with Cr, but the quantity is insufficient to be identified by means of XRD. A second possibility is the simultaneous formation of a B-rich amorphous phase separating stoichiometric TiB 2 nanocolumn structures. Thörnberg et al. [29] identified tightly packed TiB 2 nanocolumnar structures with planar defects of Ti-enriched stacking faults, accommodating the B deficiency. These results clearly show that the addition of Cr to TiB 2 influences the basic interactions between dopants (defects) and surface/grain boundaries (GB). These relations may simply affect the energy balance leading to subsequent stabilization of nanostructures of coating [31]. Therefore, under the same deposition conditions, there are more surface/grain boundaries which have a clear effect on reducing the grain size in the coating material (Fig. 1).
X-ray photoelectron spectroscopy (XPS) coupled with argon ion sputtering was applied to obtain an elemental distribution and to identify the chemical states across the coatings' thickness. The spectral lines B 1s, Ti 2p, Cr 2p 3/2 (coating material), and Fe 2p 3/2 (steel substrate) were analyzed. Additionally, C 1s and O 1s lines were identified on the coatings' surface as organic contaminations. Firstly, the surface of each sample was analyzed, followed by measurements performed after every 20 min of argon ion sputtering until reaching the steel substrate. The change in chemical composition for each sample as a function of sputtering time is shown in Fig. 2. The thickness scale is not presented as the Ar ion sputtering rate was not determined. It can be seen that the elements were uniformly distributed across the coatings' thickness, with an increased Cr content near the steel substrate resulting from the Cr interlayer.
The interlayer Cr was observed after different sputtering durations (Fig. 2 a-200 min, b-120 min, c-140 min, 2d-120 min) due to the different sputtering rate. Figure 3 presents the O 1s and Ti 2p XPS spectral lines gathered from the coating's surface. The presence of oxygen O1s is a consequence of the adsorption of contaminants from the environment, forming the oxides layer on the surface. The O 1s oxygen line consists of two peaks. The first one at 531 eV is typical for the C=O type bond (contaminations). The second one observed at a higher binding energy (BD) value of 533 eV indicates the presence of organic bonding H-O-C on the coating's surface. The position of the Ti 2p 3/2 peak at 454.5 eV indicates that this is metallic Ti, but we see also satellite or TiO (Ti(2)) as a passivation layer [32]. Peaks resulting from the contaminations layer disappear after the first cycle of sputtering. The XPS spectra of the Ti 1−x Cr x B 2 coatings (x = 0.03, 0.06, 0.1) after 100 min of Ar ion sputtering are presented in Figs. 4, 5 and 6.
The Ti 2p 3/2 and Ti2 p 1/2 peaks located at binding energies 454.5 eV and 460.5 eV, respectively, are due to the presence of Ti bonded to B in TiB 2 [32]. A small content of TiO was detected at BE = 456.4 eV (Fig. 4). The binding energy of B 1s at 187.5 eV results from the metallic state of boron in TiB 2 [33]. Two peaks were identified within the Cr 2p spectral line. The binding energy of 574.3 eV is typical for the metallic state of chromium. The second peak at BE = 576 eV (Figs. 4, 5, 6) is typical for Cr on + 3 oxidation state.
Microstructural analysis of pure and Cr-doped TiB 2 coatings was performed via TEM. The thickness of the observed coatings (Fig. 7a-d) varied depending on the Cr interlayer and was estimated to be 1.8 µm, 1.2 µm, 1.4 µm and 1.5 µm for Ti 1−x Cr x B 2 (x = 0; 0.03; 0.06; 0.1), respectively. A columnar crystalline structure was detected in all the coatings. Additionally, it was found that the column width was influenced by the Cr content in the coatings and was in the range of 79 ÷ 30 nm for pure TiB 2 , 30 ÷ 23 nm, 50 ÷ 20 nm and 23 ÷ 17 nm for coatings containing Ti 1−x Cr x B 2 (x = 0.03, 0.06, 0.1), respectively. Hexagonal close-packed (hcp) TiB 2 phase was identified in all coatings. SAED patterns of the Ti 1−x Cr x B 2 coatings (x = 0; 0.03; 0.06) determined the coatings to be monocrystalline (SAED patterns marked as 1 ÷ 3 in Fig. 7); however, the diffraction spots were stretched. In contrast, the diffraction pattern of the Ti 0.9 Cr 0.1 B 2 coating (SAED pattern 4 in Fig. 7) was in the form of continuous rings, which was linked to the doped coating's narrow column widths. Similar results have been previously obtained for Ni-doped [17] and W-doped [19] TiB 2 coatings. Reducing the column widths in TiB 2 coatings with higher chromium content can significantly affect the properties of these materials. An amorphous transition layer of Cr about 400 nm in the pure TiB 2 coating, as well as about 275 nm in the Ti 1−x Cr x B 2 (x = 0.03; 0.06) and 260 nm in the Ti 1−x Cr x B 2 (x = 0.1) are present between the coating and the substrate material. The purpose in depositing this layer was to change the residual stress between the TiB 2 coating and the steel substrate which influence the adhesion of the TiB 2 coating to the substrate. This kind of intermediate layer was previously applied by Berger et al. [5] between a TiB 2 coating and a cemented carbide substrate. He explained that this kind of amorphous or nanocomposite interlayer can be formed when the energy and mobility of the adsorbed atoms is insufficient, as it is in the case of low bias and low substrate temperature. Also, Lofaj et al. [34] in a pure TiB 2 coating magnetron sputtered onto austenitic stainless steel applied amorphous intermediate layer and explained its formation by the gradual evolution of the sputtering parameters, such as low substrate temperature during the initial deposition stage.
The surface topography of the Ti 1−x Cr x B 2 coatings was analyzed using atomic force microscopy (AFM) in tapping mode by performing 500 × 500 nm 2 area scans (Fig. 8), and the results are summarized in Table 1. The topography of all samples is typical for coatings with a columnar structure deposited by physical vapor deposition techniques [31]. In general, they have a very homogeneous and well-developed structure with no surface defects. The roughness varies slightly with the concentration of Cr up to 6 at%. For the Ti 0.9 Cr 0.1 B 2 coating, the R a parameter was found to be slightly higher. The increase of this value can be explained by the previously discussed change in preferred growth orientation and crystallite size.

Influence of Cr on mechanical properties
Based on Olivier and Pharr's method and obtained load-displacement curves, the hardness (H), Young's modulus (E) and their squares were calculated. The results are presented   Fig. 9 and summarized in Table 2. The hardness and Young's modulus measurements were carried out without exceeding a penetration depth of approx. 9% of the coatings' thickness (max. 90 nm).
The hardness of the pure TiB 2 coating was 37.7 GPa, which is higher than what has been previously reported in the literature [11,18]. Cr doping affected the material's hardness, firstly by increasing it to 38.5 GPa (maximum value) with 3% at. Cr, and later decreasing it to 35.9 and 36.3 GPa for 6 and 10% at. Cr, respectively. A slight decrease in Young's Modulus with increasing Cr content was registered. Changes in hardness and Young's Modulus were previously observed in coatings doped by Si or W [11,19]. Several factors can contribute to this. One of them is the preferred  crystallographic orientation of the fabricated coatings. TiB 2 coatings with a (001) orientation are known to yield the highest hardness compared to other TiB 2 coatings with a different orientation [29]. The second cause is due to residual stresses generated in the coating. As the residual stresses increase, the hardness of TiB 2 coatings also increases [34].
Another cause of increased hardness can result from the appearance of nanometric grains or an amorphous phase rich with B. Nanometric grains can impede dislocation generation and motion, and the amorphous phase can inhibit slip at grain boundaries [11]. The last factor may be the presence of a non-stoichiometric boron poor TiB 1.94 phase, which has a higher hardness than stoichiometric TiB 2 [29]. In all the mentioned potential causes of higher hardness for the Ti 0.97 Cr 0.03 B 2 coating is the growing volume fraction of the B tissue phase around the TiB 2 grains in overstoichiometric TiB 1.94 . For coating with higher contents of Cr (x = 0.06; 0.1) the hardness decreases. In these coatings, we observed an increase in the non-stoichiometric phase but also a change in the preferred orientation and a decrease the width of the columns. So, in these coatings grain size and texture are another mechanism that play a more major role, than those seen in the Ti 0.97 Cr 0.03 B 2 coating.
The resistance of the coating to elastic deformation can be determined by the plasticity index H/E. The H 3 /E 2 ratio permits estimation of the material's ability to dissipate energy as plastic deformation during loading. When the H/E ratio is greater than 0.1, the coatings exhibit high     [36] In this coating, the hardness is highest. The H 3 /E 2 value of also higher in the Ti 0.97 Cr 0.03 B 2 coating. Chen X. [36] show that a coating with a larger H 3 /E 2 is less likely to be plastically deformed and should therefore have higher toughness. Next, the materials' fracture toughness (K IC ) was determined by measuring the length of the indentation (Berkovich tip) cracks (Fig. 10) made on each coating at a load of 350 mN. Analysis, based on the Laugier model, showed the presence of radial and circular cracks around the indentation marks in all studied coatings. The circular cracks seen related to the increased hardness of the coatings. The presence of radial cracks suggests a tendency towards brittle fractures. The calculated fracture toughness values (K IC ) are presented in Table 2. As the Cr content increases, the fracture toughness (K IC ) increases from 1.88 to 8.5 MPa m 1/2 . This indicates an apparent increase in the ductility of the coatings.
This effect can be explained by changes in the microstructure of the coatings with increased chromium content. The presence of a non-stoichiometric TiB 1.94 phase or an amorphous phase can greatly increase the K IC value [29].
The generation of a non-stoichiometric TiB 1.94 phase (with a lack of B atoms) is a consequence of boron migration and the formation of a boron-enriched amorphous phase separating the stoichiometric TiB 2 nanocolumn structures. Another explanation may be the nanocolumnar structure of the coating. When the width of columns decreases the direction of crack propagation is changed. The energy of a crack is reduced, and cracks may disappear [23].
Scratch testing was performed to evaluate the adhesion of the hard coatings to the substrate. Figure 11 shows the scratch tracks for critical and maximum loads performed on the Ti 1−x Cr x B 2 (x = 0; 0.03; 0.06; 0.1) coatings, respectively. The surface of each mark after the scratch test was observed using an optical microscope (OM) (Fig. 11). For all investigated samples, three scratches were made up to a maximum load of 30 N. For the TiB 2 coating, the first cohesive cracks appeared at a load of 2.4 N (Fig. 11a) and an increase in the load (to 5.6 N) caused the intensification of the cracking process along with the generation of scratch track fragments (Fig. 11a). For the coating with an addition of 3% at. Cr (characterized by the highest hardness), the first cohesive cracks appeared at a lower load 1.8 N (Fig. 11c). The destruction process was similar to that of the TiB 2 coating. For the other coatings, the load at which cohesive cracks appeared increased to 3 N, as seen for Ti 0.9 Cr 0.1 B 2 (Fig. 11g). For all coatings, increasing the load to a maximum value of 30 N did not result in complete coating detachment, which clearly indicates very good adhesion to the substrate. (Fig. 11b, d, f, h). The estimated critical load value L c1 in this case is much higher than that observed for the Ti-B-C coating without adhesive layers in reference [31]. It is also higher than that of the TiB 2 coating with a Ti adhesion layer [9], where the L y force value for the best coating variant was only 0.86 N. Therefore, it can be clearly stated that the application of a pure Cr adhesion layer and

Conclusions
In this work, the mechanical properties and adhesion of magnetron sputtered Cr-doped TiB 2 coatings with an Cr buffer layer were studied. All Ti 1−x Cr x B 2 (x = 0; 0.03; 0.06; 0.1) coatings were characterized by homogeneous nanocolumnar structures, and the thickness of the coatings were in the range of 1.2 ÷ 1.8 µm. The width of the columns (crystallites) decreased with increasing chromium content (from 27 ÷ 10 nm) and XRD phase analysis for all coatings identified two dominant phases: TiB 2 and TiB 1.94 . With the Cr doping content, the texture of the films also changed. These differences in structure strongly influenced the properties of the coatings. The addition of chromium in the TiB 2 coating resulted in a decreased hardness and Young's modulus and an increased fracture toughness (K IC ), except for Ti 0.97 Cr 0.03 B 2 . In this coating, the high hardness and reduced adhesion can be explained by the contribution of a non-stoichiometric TiB 1.94 phase with a hardness increase compared to TiB 2 . As the Cr content increases, the volume of TiB 1.94 increases; however, the column width (crystallite size) decreases simultaneously and the texture changes. These two effects have a significant impact on improving fracture toughness.