Duplex Aging and Gas Nitriding Process as a Method of Surface Modiﬁcation of Titanium Alloys for Aircraft Applications

: This study discusses the effect of a duplex aging + nitriding process on the wear resistance of an aged double-phase titanium alloy, BT22. Nitriding was applied simultaneously with the heat treatment of the alloy, which is advantageous over the conventional heat and surface treatment methods applied to titanium alloys. According to the results, the thickness of the case depth of the nitrided samples was 40–50 µ m. Moreover, nitrogen was uniformly dispersed in the substrate, which was indicated by the hardness tests. The average microhardness of the substrate material was 300 HV 0.01 , while the hardness of the top layer was 1190 HV 0.01 , which is an almost four-fold increase. The applied duplex treatment substantially affected the wear performance of the tested alloy. For the untreated alloy, the maximum coefﬁcient of friction was 0.8, while in the surface-modiﬁed sample, the maximum ﬂuctuations reached 0.6. The abrasive wear process was dominant in the nitrided samples, while delamination and adhesive wear were observed for the untreated specimens. The nitrided alloy exhibited double the wear resistance of the untreated samples. The proposed treatment does not require additional time or energy consumption, providing a substantial technological advantage over conventional methods. Though the alpha case reduces the mechanical performance of titanium, the nitriding of only the component sections intended to withstand friction will have a positive effect.

Although adhesion is vital for any coating, diffusion-saturation of metals is the gold standard. Titanium alloys easily absorb gases when heated above 250-300 • C [3]. This feature of titanium alloys also limits the type of working gases that can be used during heat treatment, as the introduction of nitrogen, oxygen or hydrogen into the surface should An alloy similar to BT22, Ti-55531, a modification of the alloy Ti-5553, was studied by [21][22][23]. Its mechanical properties can be altered by a variation in the nucleation process of an α-phase. Moreover, the niobium-and zirconium-modified alloy Ti-5321 of a Ti-Al-Mo-V system is stronger than most other β and near-β alloys and can also be used to replace both Ti-1023 and Ti-64 [24,25].
The nucleation of the ultrafine α phase in an experimental double-phase titanium alloy, Ti-64-57 (a Ti-64 alloy modified by 5Mo and 7Zr), from α", obtained as a result of quenching, produces more α/β interphases [26]. This forces dislocations to climb and block, therefore significantly increasing the alloy's strength [27].
The mechanical properties and microstructures of these alloys are well known already [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26], but their ability to be thermochemically processed has still not been discussed thoroughly enough. Some studies have been carried out on the BT22 alloy and the related family of alloys, mostly evaluating the impact of the process on the hardness and microstructure, not the wear resistance [27][28][29]. Therefore, our study proposed a method of improving the wear resistance of a heavyduty double-phase Ti-64, BT22 alloy by a duplex aging and thermochemical treatment. The tribological performance of the treated material was tested under dry sliding conditions in a pin on disc configuration. The study discusses both the coefficient of friction (COF) and the wear of the tribocouple. To support the results, SEM observations were used to determine the dominant wear modes of both treated and reference (non-heat-treated) samples. The results will allow academia and the industry to enhance the wear resistance of other β, near-β and α + β alloys.
Of the reviewed alloys (Table 1), the most frequently used alloy is Ti-64. For structural and biomedical applications, many methods of surface strengthening can be applied [30,31]. The wear resistance properties of this alloy were studied in depth. The application of nitriding in BT22 alloy was studied in [32,33], and the results were very promising.

Plasma Nitriding
Plasma nitriding methods typically provide high hardness and case depth to treated surfaces. Moreover, the process is known to smoothen the surface profile of the sample [34]. However, various sharp edges may be burnt or overheated due to a local increase in the surface temperature; nevertheless, plasma nitriding is widely used to enhance the wear resistance of titanium alloys and steels [34].
Taktak and Akbulut [35] investigated the wear resistance of the titanium alloy Ti-6-4, which was nitrided in a plasma medium after detonation explosive shock treatment (DEST). After nitriding at temperatures up to 900 • C for 12 h, a maximum top surface hardness of 2500 HV 0.05. was reported. This value is much higher than found in other studies reviewed in this section. The authors did not explain it, but they probably obtained a dense compound layer consisting only of TiN due to the high strain induced by DEST. The wear tests showed [35] that nitriding significantly increases the wear resistance of titanium alloys. The primary wear mechanism of WC-Co vs. nitrided Ti-64 was fatigue-assisted abrasive wear.
Pure titanium was plasma-nitrided in [36]. The high hardness (1000-1300 HV 0.1 due to the formation of super-hard ε-Ti 2 N and δ-TiN phases) and relatively low thickness (5-15 microns) created an unwanted sevenfold difference in hardness between the diffusion layer and the substrate. The high hardness values in [35,36] may also be explained by the low load used for Vickers indentation. According to the authors of [36], plasma nitriding of pure Ti resulted in a significant reduction in COF.
Plasma nitriding of a Ti5Al4V2Mo alloy in a mixture of nitrogen and argon [37] for 3-4 h at temperatures 500-900 • C produced a rather hard surface (600-800 HV 0.05 ). Thus, we can conclude that using argon and reduced pressure does not promote significant layer thickness and hardness. In other work [38], a Grade 2 titanium alloy was plasma-nitrided at 650-950 • C (with 50 • C increments) for 8 h. The maximum surface hardness (1550 HV 0.05 ) and layer thickness (about 55 µm) were reported for plasma nitriding at 950 • C. However, increasing the nitriding temperature above 850 • C reduces the wear resistance of the alloy. Plasma nitriding using an N 2 -NH 3 gas mixture at temperatures of 700 and 750 • C was studied by Shen and Wang [39]. According to the results provided, a 2.3 µm compound layer for NH 3 and a 6.5 µm compound layer for the N 2 -NH 3 gaseous mixture was obtained, which is much less than that achieved at higher temperatures.
Plasma nitriding provides the greatest layer thickness and hardness when applied at temperatures above 850 • C. Low-temperature plasma nitriding does not lead to high thickness or hardness in the modified layers. Plasma nitriding is quite a complicated method. It requires sophisticated equipment [35][36][37] and cannot always be reproduced in common manufacturing conditions. It greatly improves corrosion resistance. The best wear resistance was reported in plasma-nitrided Ti alloys for specimens treated at a temperature range of 800-850 • C [38]. Increasing the temperature of plasma nitriding from 700 to 850 • C improves the wear resistance [38,39]. On the other hand, reducing the process temperature reduces the surface roughness [36] and the coefficient of friction, thus increasing the wear resistance [35,39]. Moreover, a low-temperature process can improve the corrosion resistance of an alloy by creating a protective barrier [40].

Laser-Assisted Nitriding
For titanium alloys, the novel plasma-enhanced pulsed-laser deposition system [41] allows one to obtain a relatively thin nitride layer (more than 100 nm over 1 h), which is not very effective. Compared with plasma nitriding, laser-assisted gas nitriding [42] requires much more power (1-5 kW vs. 400-650 W [36]) for processing. Moreover, the plasma nitriding method treats all the exposed surfaces, while laser-assisted methods affect only the scanned surfaces. However, they may be used for selectively treated areas and textured surfaces [43] or thermocycled coatings [44,45]. While laser nitriding involves surface remelting and requires precise control of the saturation parameters, it is also hard to process internal surfaces via this method (e.g., holes for the insertion of bronze landing gear bushings). Obtaining a thick coating is not always possible, but their effective production has been reported [46].

Gas-Blow Induction Heating Nitriding Method
Induction heating of a Ti64 alloy in a nitrogen atmosphere generates a nitrided layer in only a few minutes [47]. The best wear resistance of the alloy is observed when the treatment temperature is 900 • C, as both compound and diffusion layers are formed. Nevertheless, at this temperature, coarsening of the microstructure may be observed, which has an adverse effect on the alloy's mechanical properties. Moreover, the authors of [48] removed the compound layer and discovered that this improved the fatigue resistance while reducing the wear resistance.
Low temperatures (650 • C) during gas-blow induction heating nitriding after shot peening with fine particles produce both diffusion and compound layers in a relatively short exposure time. The hardness of treated Ti-64 specimens is 400-450 HV 0.025 , which is low. However, the highest substrate temperature can be observed not on the top of the sample but in the sublayer because the surface is cooled by gas flow during the induction heating process [49,50]. The gas-blow induction heating nitriding process has potential as a rapid nitriding technology, but it should be studied much more thoroughly than it has been to date.

Electron-Beam Vacuum Nitriding
This newly introduced method [51,52] uses an electron beam for nitriding under low (8-10 Pa) pressure. Pure titanium was used to manufacture the test specimens, and the results were quite positive. The strengthened layer thickness was 15-35 µm, and the surface micro hardness measured by the Vickers method at a load of 0.1 kg was up to 12 GPa (≈1200 HV 0.1 ). Though this technology may have a future as a type of plasma nitriding, it is necessary to develop it further for aerospace applications.

Gas Nitriding
Gas nitriding is probably the simplest of all nitriding methods. It requires only a furnace under good pressure control at the specified temperature ranges. The dynamic gas flow allows internal surfaces to be treated with the same efficiency as flat specimens. The nonoxidizing atmosphere during gas nitriding allows the heat treatment of the alloy in one step with thermochemical processing [31,32].
Gas nitriding was used to study the strengthening of the newly developed TZ20 (Ti-19.2Zr-6.48Al-3.86V-Hf-Na) alloy [53]. Increasing the temperature from 500 to 650 • C increased the weight gain from 0.62 to 3.67 mg/cm 2 (592%). The duration of the process was 260 min, and the nitrogen pressure in the reaction chamber was 0.05 MPa. The nitrided layer's thickness was 1-7 µm after 1 h, and the maximum value was 12-14 µm. The surface hardness increased from 470 HV 0.01 to 870 HV 0.01 . In another study [54], intermittent vacuum nitriding of a TB8 (Ti-15.32Mo-3.23Al-2.86Nb) alloy at 780 • C gave much better results: a hardness of 850-900 HV 0.1 and a nitride layer thickness of up to 120 µm. We may conclude that an increase in temperature and nitrogen gas pressure results in a thicker compound layer and higher hardness, but this also strongly depends on the alloy's composition. The influence of higher temperatures under atmospheric pressure is presented in [55]: increasing the temperature and exposure time resulted in thicker and harder surfaces. A method of gas nitriding in an ammonia medium also produced good results [39,56]: at 1100 • C and atmospheric pressure, the layer's thickness was 40-50 µm after 5 h.
Gas nitriding is not effective for β-titanium alloys but is good for double-phase alloys in the region of the β-transus temperature [57]. Vacuum preheating has a beneficial effect on the gas nitriding process [58]. This helps to remove oxygen and other gaseous impurities from the surface and facilitates the thermal diffusion process. Regarding the BT22 titanium alloy, the standard heat treatment combines diffusion annealing [11] from the betta region (750-850 • C) and slow cooling. This results in a uniform microstructure with a good combination of mechanical properties. High strength (up to UTS 1400 MPa) may be obtained after aging. This alloy may be used both in annealed and aged conditions.
The discs (samples) had a diameter of 55 mm and a height of 5 mm. The pin counterparts had a diameter of 8 mm and a length of 10 mm (the BT22 material in as-supplied condition). The friction surfaces of the discs (before thermochemical processing) and pins were ground to a roughness of R a 0.53 µm. Discs were degreased in an ultrasonic alcohol bath before the nitriding and wear tests.
The discs' and pins' dimensions and the "all-over" treatment of discs made it possible to perform four separate tests on each sample (See Figure 1b,c).
nonoxidizing atmosphere during gas nitriding allows the heat treatment of the alloy in one step with thermochemical processing [31,32].
Gas nitriding was used to study the strengthening of the newly developed TZ20 (Ti-19.2Zr-6.48Al-3.86V-Hf-Na) alloy [53]. Increasing the temperature from 500 to 650 °C increased the weight gain from 0.62 to 3.67 mg/cm 2 (592%). The duration of the process was 260 min, and the nitrogen pressure in the reaction chamber was 0.05 MPa. The nitrided layer's thickness was 1-7 μm after 1 h, and the maximum value was 12-14 μm. The surface hardness increased from 470 HV0.01 to 870 HV0.01. In another study [54], intermittent vacuum nitriding of a TB8 (Ti-15.32Mo-3.23Al-2.86Nb) alloy at 780 °C gave much better results: a hardness of 850-900 HV0.1 and a nitride layer thickness of up to 120 μm. We may conclude that an increase in temperature and nitrogen gas pressure results in a thicker compound layer and higher hardness, but this also strongly depends on the alloy's composition. The influence of higher temperatures under atmospheric pressure is presented in [55]: increasing the temperature and exposure time resulted in thicker and harder surfaces. A method of gas nitriding in an ammonia medium also produced good results [39,56]: at 1100 °C and atmospheric pressure, the layer's thickness was 40-50 μm after 5 h.
Gas nitriding is not effective for β-titanium alloys but is good for double-phase alloys in the region of the β-transus temperature [57]. Vacuum preheating has a beneficial effect on the gas nitriding process [58]. This helps to remove oxygen and other gaseous impurities from the surface and facilitates the thermal diffusion process. Regarding the BT22 titanium alloy, the standard heat treatment combines diffusion annealing [11] from the betta region (750-850 °C) and slow cooling. This results in a uniform microstructure with a good combination of mechanical properties. High strength (up to UTS 1400 MPa) may be obtained after aging. This alloy may be used both in annealed and aged conditions.
The discs (samples) had a diameter of 55 mm and a height of 5 mm. The pin counterparts had a diameter of 8 mm and a length of 10 mm (the BT22 material in as-supplied condition). The friction surfaces of the discs (before thermochemical processing) and pins were ground to a roughness of Ra 0.53 μm. Discs were degreased in an ultrasonic alcohol bath before the nitriding and wear tests.
The discs' and pins' dimensions and the "all-over" treatment of discs made it possible to perform four separate tests on each sample (See Figure 1b  The details of thermochemical processing are shown in Figure 2. Before the heating started, the air was evacuated from the furnace, reaching a vacuum of 10 µPa to remove the oxide film and reduce the presence of oxygen and hydrogen. The details of thermochemical processing are shown in Figure 2. Before the heating started, the air was evacuated from the furnace, reaching a vacuum of 10 μPa to remove the oxide film and reduce the presence of oxygen and hydrogen. Before the first aging stage, the nitrogen was let into the furnace until the pressure reached PN2 = 10 5 Pа. To utilize pure nitrogen, we passed it through a box with silica gel (to remove water; ambient temperature) and a gas trapper (a box with titanium chips heated to t 1 = 800 °C and t 2 = 900 °C). After each treatment procedure, the titanium chips were replaced, and the silica gel was vacuum-baked at 180 °C for 2 h. Next, aging at 750 °C was applied for 4 h, followed by aging at 850 °C for 4 h with the simultaneous diffusion saturation of the surface in the nitrogen atmosphere.
The microstructure was studied before and after nitriding in normal sections. The hardness and microstructure of the nitride layer were studied on tapered sections ( Figure  3). The sample was cut apart, then one of the parts was polished and etched for microstructural examinations; the other one was only polished and used for microhardness tests. Etching created a somewhat rough surface, and this reduced the repeatability and accuracy of measurements. Moreover, placing the indentations in multiple lines ( Figure  1b) allowed to keep an appropriate distance between them and to get any desired vertical pitch. The tapered section made the nitrided layer thicker and facilitated the study.
To treat the surface, Kroll's reagent with the following composition was used for etching: 5% of HNO3, 3% of HF and 92% of H2O. For the phase analyses of the nitrided surface, a DRON-3.0 X-ray diffractometer was utilized (Bregg-Brentano 2θ configuration, monochromatic Cu Kα radiation). Before the first aging stage, the nitrogen was let into the furnace until the pressure reached P N2 = 10 5 Pa. To utilize pure nitrogen, we passed it through a box with silica gel (to remove water; ambient temperature) and a gas trapper (a box with titanium chips heated to t 1 = 800 • C and t 2 = 900 • C). After each treatment procedure, the titanium chips were replaced, and the silica gel was vacuum-baked at 180 • C for 2 h. Next, aging at 750 • C was applied for 4 h, followed by aging at 850 • C for 4 h with the simultaneous diffusion saturation of the surface in the nitrogen atmosphere.
The microstructure was studied before and after nitriding in normal sections. The hardness and microstructure of the nitride layer were studied on tapered sections ( Figure 3). The sample was cut apart, then one of the parts was polished and etched for microstructural examinations; the other one was only polished and used for microhardness tests. Etching created a somewhat rough surface, and this reduced the repeatability and accuracy of measurements. Moreover, placing the indentations in multiple lines (Figure 1b) allowed to keep an appropriate distance between them and to get any desired vertical pitch. The tapered section made the nitrided layer thicker and facilitated the study. The microstructure of the modified layers of titanium alloys were examined using an optical microscope. For investigating the detailed microstructural and wear tracks, an SEM-FIB DualBeam Scios 2 (Thermo Scientific, USA) was used. Vickers microhardness was measured with a PMT-3M instrument under an indenter load of 10 g. The nitrided layer's thickness was measured by the metallographic and hardness profile methods [59].
We used a T11 pin-on-disc tribometer for the wear tests. Samples were weighed before and after the experiment using Radwag XA/210Y analytical balances with an accuracy of 0.01 mg. As the pins' installation diameter R values (see Figure 1c) were dissimilar surface, a DRON-3.0 X-ray diffractometer was utilized (Bregg-Brentano 2θ configuration, monochromatic Cu Kα radiation).
The microstructure of the modified layers of titanium alloys were examined using an optical microscope. For investigating the detailed microstructural and wear tracks, an SEM-FIB DualBeam Scios 2 (Thermo Scientific, USA) was used. Vickers microhardness was measured with a PMT-3M instrument under an indenter load of 10 g. The nitrided layer's thickness was measured by the metallographic and hardness profile methods [59].
We used a T11 pin-on-disc tribometer for the wear tests. Samples were weighed before and after the experiment using Radwag XA/210Y analytical balances with an accuracy of 0.01 mg. As the pins' installation diameter R values (see Figure 1c) were dissimilar (26 and 18 mm), the rpm was adjusted (266 and 184 respectively) to provide a constant sliding speed of 0.5 m/s and a sliding distance of 300 m and contact pressure of 0.2 MPa.

Microstructure and Phase Analyses
After nitriding, the surface color changed from silvery to yellow-gold, indicating the presence of titanium nitrides [53]. The results of the XRD studies are shown in Figure 4 and Table 2.  The microstructure of the modified layers of titanium alloys were examined using an optical microscope. For investigating the detailed microstructural and wear tracks, an SEM-FIB DualBeam Scios 2 (Thermo Scientific, USA) was used. Vickers microhardness was measured with a PMT-3M instrument under an indenter load of 10 g. The nitrided layer's thickness was measured by the metallographic and hardness profile methods [59].
We used a T11 pin-on-disc tribometer for the wear tests. Samples were weighed before and after the experiment using Radwag XA/210Y analytical balances with an accuracy of 0.01 mg. As the pins' installation diameter R values (see Figure 1c) were dissimilar (26 and 18 mm), the rpm was adjusted (266 and 184 respectively) to provide a constant sliding speed of 0.5 m/s and a sliding distance of 300 m and contact pressure of 0.2 MPa.

Microstructure and Phase Analyses
After nitriding, the surface color changed from silvery to yellow-gold, indicating the presence of titanium nitrides [53]. The results of the XRD studies are shown in Figure 4 and Table 2.   Therefore, the nitride layer predominantly consisted of unsaturated Ti 2 N, with a minor presence of TiN and Ti 2 O. The formation of a single-phase TiN compound layer was avoided. Figure 5a,b show two micrographs of the alloy obtained before heat treatment (light field: α phase; dark field: β phase. There are two types of α-particles present in Figure 5b: coarse and fine. The coarse grain boundary phase (red arrows) had a length of 10-60 µm and a thickness of 1-3 µm. Fine randomly oriented α colonies inside the β grains (yellow arrows) had a length of 3-20 µm and a thickness typically less than 1 µm. This microstructure is common for BT22 alloys in annealed equilibrium (as supplied) conditions [11]. The grain boundary α phase helps in measuring the average β grain size: 30-50 µm. The content of the α phase nearly equaled the quantity of the β phase (the β-phase content was 45-52%). The measured hardness of the material was 300 HV 0.01 . Figures 5a and b show two micrographs of the alloy obtained before heat treatment (light field: α phase; dark field: β phase. There are two types of α-particles present in Figure 5b: coarse and fine. The coarse grain boundary phase (red arrows) had a length of 10-60 μm and a thickness of 1-3 μm. Fine randomly oriented α colonies inside the β grains (yellow arrows) had a length of 3-20 μm and a thickness typically less than 1 μm. This microstructure is common for BT22 alloys in annealed equilibrium (as supplied) conditions [11]. The grain boundary α phase helps in measuring the average β grain size: 30-50 μm. The content of the α phase nearly equaled the quantity of the β phase (the β-phase content was 45-52%). The measured hardness of the material was 300 HV0.01.
The microstructure of the alloy after double-step nitriding is presented in Figure 5c. The thickness of the nitrided layer was 50 μm (the scale of Figure 5c is recalculated, taking into account the oblique section geometry, see Figure 1a). The α phase particles became finer (as seen in Figure 5a,c) and led to an increase in the alloy's surface hardness up to 1190 HV0.01 ( Figure 6). As the grain boundary phase changed its morphology, it was hard to measure the size of the β grains.

Hardness
The microhardness measurements indicated that the thickness of the thermodiffusion layer was 40-50 μm, which is typical for this surface strengthening technology. The surface hardness was uniform, which is evidence of a stable saturation process with a modified layer of the same thickness over the specimen. The hardness results of the three specimens (mean values + SD) are presented in Figure 6. Additionally, three indentations to measure the surface hardness were made on the top of each sample. The subsurface microhardness was 715 HV0.01. Similarly, metallographic studies indicated the same thickness of the nitrided layer: 40-50 μm. The hardness of the untreated material was 300 HV0.01. The microstructure of the alloy after double-step nitriding is presented in Figure 5c. The thickness of the nitrided layer was 50 µm (the scale of Figure 5c is recalculated, taking into account the oblique section geometry, see Figure 1a). The α phase particles became finer (as seen in Figure 5a,c) and led to an increase in the alloy's surface hardness up to 1190 HV 0.01 ( Figure 6). As the grain boundary phase changed its morphology, it was hard to measure the size of the β grains. The maximum surface hardness achieved by the method was 1190 HV0.01, which is more than that found for the method described in [37] for the titanium alloy Ti5Al4V2Mo (600-800 HV0.05), roughly the same as that in [33] (1550 HV0.05) and [29] (1000-1300 HV0.1), but less than the 2500 HV0.05 found in [28]. It should also be mentioned that lower loads during the microhardness test produced greater hardness as the indenter pierces the thinner top layer of the alloy. Naturally, this top layer has higher hardness. To compare the hardness profile, the strengthened layer's thickness were nearly the same as in [38,41,55] (in the range of 15-50 μm) but less than that found in [53], namely 120 μm.

Hardness
The microhardness measurements indicated that the thickness of the thermodiffusion layer was 40-50 µm, which is typical for this surface strengthening technology. The surface hardness was uniform, which is evidence of a stable saturation process with a modified layer of the same thickness over the specimen. The hardness results of the three specimens (mean values + SD) are presented in Figure 6. Additionally, three indentations to measure The maximum surface hardness achieved by the method was 1190 HV 0.01 , which is more than that found for the method described in [37] for the titanium alloy Ti5Al4V2Mo (600-800 HV 0.05 ), roughly the same as that in [33] (1550 HV 0.05 ) and [29] (1000-1300 HV 0.1 ), but less than the 2500 HV 0.05 found in [28]. It should also be mentioned that lower loads during the microhardness test produced greater hardness as the indenter pierces the thinner top layer of the alloy. Naturally, this top layer has higher hardness. To compare the hardness profile, the strengthened layer's thickness were nearly the same as in [38,41,55] (in the range of 15-50 µm) but less than that found in [53], namely 120 µm.

Wear Resistance Test
Specimens were tested separately, with four tests per specimen condition. The test results are shown in Figure 7. The mean wear loss of the untreated titanium alloy was 16.55 µg, and the mean wear loss of the nitrided alloy was 8.24 µg (mean value + SD). Therefore, the nitrided alloy is superior when compared with untreated material by a magnitude of 2. The nitrided layer separates the friction surfaces, created by a natural or friction-induced oxide film [60], thus preventing contact between the two metals. The maximum surface hardness achieved by the method was 1190 HV0.01, which is more than that found for the method described in [37] for the titanium alloy Ti5Al4V2Mo (600-800 HV0.05), roughly the same as that in [33] (1550 HV0.05) and [29] (1000-1300 HV0.1), but less than the 2500 HV0.05 found in [28]. It should also be mentioned that lower loads during the microhardness test produced greater hardness as the indenter pierces the thinner top layer of the alloy. Naturally, this top layer has higher hardness. To compare the hardness profile, the strengthened layer's thickness were nearly the same as in [38,41,55] (in the range of 15-50 μm) but less than that found in [53], namely 120 μm.

Wear Resistance Test
Specimens were tested separately, with four tests per specimen condition. The test results are shown in Figure 7. The mean wear loss of the untreated titanium alloy was 16.55 μg, and the mean wear loss of the nitrided alloy was 8.24 μg (mean value + SD). Therefore, the nitrided alloy is superior when compared with untreated material by a magnitude of 2. The nitrided layer separates the friction surfaces, created by a natural or friction-induced oxide film [60], thus preventing contact between the two metals. The friction coefficient (COF) values are given in Figure 8. As one can see, regardless of the nitriding treatment applied, after saturating the BT22 alloy's surface with nitrogen, the variation reduced, and the average COF value dropped from 0.6 to 0.5. This is also reflected in the weight loss of the samples (Figure 7), where a clear advantage of nitriding over nontreated samples can be seen. The friction coefficient (COF) values are given in Figure 8. As one can see, regardless of the nitriding treatment applied, after saturating the BT22 alloy's surface with nitrogen, the variation reduced, and the average COF value dropped from 0.6 to 0.5. This is also reflected in the weight loss of the samples (Figure 7), where a clear advantage of nitriding over nontreated samples can be seen.

The Wear Tracks' Topography
The wear track analyses indicated a substantial difference in the dominant wear modes between the nontreated and the duplex-treated samples (Figure 9). The untreated alloy's wear track topography is given in Figure 9a. The general appearance of the friction surface indicates nonuniform wear with signs of deformation, metal flow, adhesive wear

The Wear Tracks' Topography
The wear track analyses indicated a substantial difference in the dominant wear modes between the nontreated and the duplex-treated samples (Figure 9). The untreated alloy's wear track topography is given in Figure 9a. The general appearance of the friction surface indicates nonuniform wear with signs of deformation, metal flow, adhesive wear and scuffing. On the other hand, in the duplex-treated sample (Figure 9b), the dominant wear mode is abrasive wear, was accompanied by plowing. In some spots, the diffusion layer may be seen, which proves that some of the layers survived the dry friction test.

The Wear Tracks' Topography
The wear track analyses indicated a substantial difference in the dominant wear modes between the nontreated and the duplex-treated samples (Figure 9). The untreated alloy's wear track topography is given in Figure 9a. The general appearance of the friction surface indicates nonuniform wear with signs of deformation, metal flow, adhesive wear and scuffing. On the other hand, in the duplex-treated sample (Figure 9b), the dominant wear mode is abrasive wear, was accompanied by plowing. In some spots, the diffusion layer may be seen, which proves that some of the layers survived the dry friction test. More detailed studies of the surfaces were carried out using SEM. Figures 10 and 11 compare the wear scars of the materials. On the surface of the untreated alloy (Figure 10a), signs of plowing and metal outflows are intensive, although the area around this portion seems to be undamaged. The high-magnification image indicates the formation of secondary frictional layers on the surface, which may have originated from both the parent or the counterpart material (Figure 10b). However, it is challenging to claim whether the counterpart material was adhesively transferred because the substrate was self-mated (untreated BT22). Figure 10c indicates the intensive cold plastic deformation of the surface layer and the metal portions beginning to delaminate (arrowed). We may suppose that in this case, most of the wear particles formed as a result of deformed metal tears flowing out of the actual friction contact sites (see Figure 10d). Moreover, some wear debris was shaped by repeated rolling between the discs' and counterparts' surfaces ( Figure 10e). Such spheres are supposed to be harder than the parent material and are capable of causing damage, as seen in Figure 8a.
On the contrary, the surface of the nitrided specimen's wear track was flat ( Figure  11a). The actual contact occurs more often along lighter strips of the surface, containing less nitrogen (due to the worn-off top layer). The black spots present in Figure 11a indicate the wear products. Details are shown in Figure 11d More detailed studies of the surfaces were carried out using SEM. Figures 10 and 11 compare the wear scars of the materials. On the surface of the untreated alloy (Figure 10a), signs of plowing and metal outflows are intensive, although the area around this portion seems to be undamaged. The high-magnification image indicates the formation of secondary frictional layers on the surface, which may have originated from both the parent or the counterpart material (Figure 10b). However, it is challenging to claim whether the counterpart material was adhesively transferred because the substrate was self-mated (untreated BT22). Figure 10c indicates the intensive cold plastic deformation of the surface layer and the metal portions beginning to delaminate (arrowed). We may suppose that in this case, most of the wear particles formed as a result of deformed metal tears flowing out of the actual friction contact sites (see Figure 10d). Moreover, some wear debris was shaped by repeated rolling between the discs' and counterparts' surfaces ( Figure 10e). Such spheres are supposed to be harder than the parent material and are capable of causing damage, as seen in Figure 8a.
On the contrary, the surface of the nitrided specimen's wear track was flat (Figure 11a). The actual contact occurs more often along lighter strips of the surface, containing less nitrogen (due to the worn-off top layer). The black spots present in Figure 11a indicate the wear products. Details are shown in Figure 11d,e. Figure 9b indicates a possible brittle fracture in the surface. As the images were taken in BSE mode, the brighter areas correspond to the metallic substrate, while the areas saturated with nitrogen atoms are represented by the darker areas. The nature of other light spots around this area originated from their counterparts but not the substrate material. This place may be one of the few sites of adhesion wear typical of untreated titanium alloys [61].
Moreover, some plastic deformation and smearing of the Ti alloy can be seen, along with places of interest appearing as light spots on the surface. They contain a minimum of light elements (N and O) and may be formed due to adhesion and material transfer from their counterparts (untreated BT22). The size of powdery wear debris is 0.2-1 µm, with some coarser 3 µm particles covering the recesses and other "friction-free" parts of the surface (Figure 11d,e). Wear debris on the surface is darker than the sublayer (Figure 11e). The debris mainly resulted from the deterioration of the nitrided layer. The arrows show areas of further surface destruction and the formation of new wear particles. Their sharp edges contribute to increased wear loss. Their dimensions are close to the width of the grooves. They may also be impregnated into the soft counterpart surface, creating even more severe friction conditions. with places of interest appearing as light spots on the surface. They contain a minimum of light elements (N and O) and may be formed due to adhesion and material transfer from their counterparts (untreated BT22). The size of powdery wear debris is 0.2-1 μm, with some coarser 3 μm particles covering the recesses and other "friction-free" parts of the surface (Figure 11d,e). Wear debris on the surface is darker than the sublayer ( Figure  11e). The debris mainly resulted from the deterioration of the nitrided layer. The arrows show areas of further surface destruction and the formation of new wear particles. Their sharp edges contribute to increased wear loss. Their dimensions are close to the width of the grooves. They may also be impregnated into the soft counterpart surface, creating even more severe friction conditions.

Discussion
The composition, microstructure and microhardness of the surface laye a high amount of Ti2O, posed a question regarding the oxygen source. We the source was the natural oxide film on the surface of the alloy and tha amount was added from the nitrogen or during the cutting of the alloy into theless, its presence is considered to be beneficial, as it might decrease the brittleness of the surface. It was reported in [62] that relatively soft titanium have as solid lubricants. The hardness of Ti2N is also much lower than that o has a positive effect on the wear resistance [63].
The coefficient of the friction of the untreated BT22 alloy in the current s and that of the nitrided alloy was 0.5. This is much higher than the value o and nearly the same as the result in [35] (0.45). In other studies, the reported 0.7 [42], which is within the confidence interval (0.25-0.7). Compared with alloy, nitriding during aging provided stable and lower values for the coef tion and doubled the wear loss reduction (Figures 7 and 8). The maximum

Discussion
The composition, microstructure and microhardness of the surface layer, which had a high amount of Ti 2 O, posed a question regarding the oxygen source. We suppose that the source was the natural oxide film on the surface of the alloy and that some other amount was added from the nitrogen or during the cutting of the alloy into discs. Nevertheless, its presence is considered to be beneficial, as it might decrease the hardness and brittleness of the surface. It was reported in [62] that relatively soft titanium oxidides behave as solid lubricants. The hardness of Ti 2 N is also much lower than that of TN, and this has a positive effect on the wear resistance [63].
The coefficient of the friction of the untreated BT22 alloy in the current study was 0.6, and that of the nitrided alloy was 0.5. This is much higher than the value of 0.25 in [36] and nearly the same as the result in [35] (0.45). In other studies, the reported COF was 0.6-0.7 [42], which is within the confidence interval (0.25-0.7). Compared with the untreated alloy, nitriding during aging provided stable and lower values for the coefficient of friction and doubled the wear loss reduction (Figures 7 and 8). The maximum value of the friction coefficient for the untreated alloy was 0.8. This may be evidence of adhesion wear. In Figure 8a, one may see an initial rapid increase in the COF of the untreated alloy. The natural oxide film wears off from both the specimen and the counterpart at the initial moment of friction. Oxide-free metallic surfaces firmly adhere to each other. When the friction process becomes stable, the friction-induced oxide film separates the surfaces, and the coefficient of friction drops to lower values. The other reason for the high friction coefficient is the deformation of initial surface asperities and the increased number of contact points on the matting surfaces [64]. For nitrided specimens, the diffusion layer was strong enough and protected the surfaces from free metal contact, and the friction coefficient was stable from the beginning. Fluctuation in the coefficient of friction was much less and reached a value of 0.6 in a very short time. This is more evidence of the wear mechanism being other than adhesive [35].
In Figures 10 and 11, it is clear that the two groups of specimens presented different wear mechanisms. This is additionally seen in Figure 8. Thus, for the untreated alloy, the friction process typically begins from the reconfiguration of the as-ground surface. Due to plastic deformation, the asperities change their shape and the number of contact points increases [64]. The COF oscillations and morphology peculiarities, seen especially in Figure 10b, indicate the sticking of contacting surfaces. Furthermore, these deformed asperities and the sides of the wear grooves are deformed into thin foils (Figure 10c), which tear off and become wear products. After being processed multiple times by the counterpart, these secondary wear products may adhere back to the surface and subsequently fracture into much smaller particles (Figure 10d) or may roll into large wear particles (Figure 10e) and cause severe damage to the surface (Figure 10a). A detailed analysis of the wear particles (Figure 10e) indicates the possibility of agglomeration when several torn-off particles roll into one. Therefore, for this tribocouple, we may conclude that the wear mechanism was predominantly adhesion-delamination.
For nitrided specimens, the wear mechanism was different. Nitrogen in the surface reduced the capacity for adhesive wear, despite the counterpart being fabricated of untreated BT22 titanium alloy, as in the previous case [63]. Therefore, less evidence of adhesive wear was seen. The harder and more brittle modified layer produced wear particles of submicrometer size. As may be seen in Figure 11b,d, they can be assumed to have formed as a result of the destruction of the "horns" formed during friction. They continued to damage the surface (Figure 11e), producing fine wear grooves (Figure 11e). The nature of the "horns" is questionable. They may result from the adhesion [63] or destruction of the leading edge of the metal outflows (Figure 11c), through oxidation or by fatigue wear mechanisms. Considering the nature of the friction coefficient diagrams (Figure 8), the adhesion is a supplementary wear mechanism, while fatigue and abrasion are dominant. The possibility of adhesion and abrasion wear in titanium alloys has been proven in [65]. Importantly, due to small amounts of TiN being formed, three-body abrasive phenomena were not detected.
The wear loss of the nitrided alloy was half that of the untreated alloy. Nitriding in specially designed conditions decreases the wear loss by a magnitude of 3-5 [35], or even 7 [34]. In our case, an increase in wear resistance and surface hardness were benefits obtained by combining heat treatment and thermochemical processing. No additional time, energy or equipment was used. Taking this into account, the result begs to be noticed. As mentioned, we used the untreated BT22 titanium alloy as a counterpart, which may have contributed to the increased wear by adhering to and deteriorating the nitrided layer.
This surface strengthening method was studied only for the duplex aging process at 750-850 • C. However, BT22 alloys may be used in various conditions (quenching, annealing, annealing/quenching and aging) [11,12], and the highest strength (σ u up to 1400 MPa, [11]) can be achieved after quenching and aging at 480-650 • C, which is not enough for nitriding. It has to be stressed that while the wear resistance of the alloy can be improved due to nitriding, some of the mechanical properties of the alloy will have to be sacrificed. The alpha case formed during the proposed thermochemical treatment is suitable for the limited industrial application of titanium alloys. Therefore, it is necessary to find the combination of process parameters to minimize the impact on the performance of the parts, to reduce the thickness of the compound layer and to expand the nitrogen diffusion zone [66].

Conclusions
The studied duplex method of gas nitriding during the regular heat treatment of a BT22 titanium alloy increased the surface microhardness from 300 to 715 HV 0.01 . The maximum microhardness value of 1190 HV 0.01 was achieved on the top surface of the sample. According to XRD studies, a multiphase layer consisting of Ti 2 N, TiN, Ti 2 O and the metal phase was formed. The proposed thermochemical process improved the wear resistance of the BT22 alloy two-fold. Nitriding reduced the adhesion interactions between the sample and the counterparts. The wear mechanism of the untreated alloy in the pin-on-disc configuration was adhesion-delamination wear. For the nitrided alloy, the dominant wear process was fatigue wear, accompanied by adhesion. The tree-body abrasion phenomenon on the friction surface was not observed. This method (duplex treatment: nitriding during aging) reduced the steps in the process and the manufacturing costs of wear-tolerant titanium alloys. The method also achieved twice the wear resistance with no additional time and energy consumption.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.