Influence of the Structural-Phase State of a Copper Substrate upon Modification with Titanium Ions on the Thermal Cyclic Resistance of a Coating Based on Zr-Y-O

Abstract

The results of investigation of the surface of a copper substrate modified with titanium ions are presented. The phase composition, the structure, and the morphology of the surface of the copper alloy modified by titanium ions have been investigated by X-ray, SEM, and TEM. It has been established that there are the intermetallic phases of the Cu-Ti equilibrium diagram in the surface layer during the treatment of copper by the titanium ions. A multilevel micro- and nanoporous structure is formed in the modified layer. It has been established that the structure-phase state and morphology of the surface layers of copper directly effects on the thermocycler resistance and adhesion of the Zr-Y-O coating. The thermocyclic resistance of the Zr-Y-O coating increases by an order of magnitude, the adhesion to the substrate is 2 times if the substrate surface is treated with titanium ions for 6 min.

1. Introduction

Methods of surface modification are widely used in the world. Today, there are many radiation-beam technologies (RBT), which are classified by energy carriers and take into account the main modifying factor. Among the RBT, the ion treatment of the materials of various instrument and the equipment is the most widespread method used in medicine, food, automotive, and agriculture industries [1,2,3].

Recently, the alloying of structural metal materials with various impurities has received increasing attention [4,5]. This is an effective and economical way to increase the durability of the parts operating under cyclic loads, contact fatigue, and abrasion. As a result, strong, durable, and corrosion-resistant layers on their surfaces are produced [6,7]. In work [8], a variant of the solution of the problem of modification of the transition region ‘coating substrate’ is proposed for a purposeful change of adhesion, strength, corrosion, and other properties of coatings. The solution is to increase the depth of the subsurface layer of the substrate being modified. The authors [8,9,10] believe that, in addition to the change in the content of impurity elements in the region of the interface, the reason for the improvement of the adhesion properties of the coating is the effect of ion implantation, such as the radiation damage in the surface layer of the substrate, which, in turn, causes a change in the conditions for the nucleation and a subsequent growth of the coating. Additionally, in addition to the phase change in the composition, irradiation of the surface of solids by ionic and plasma flows causes certain changes in the relief. Depending on the parameters of the irradiating flux and the conditions on the surface, these changes manifest both in the development and the smoothing of the relief. Surface modification can be a direct result of ionic or plasma exposure, for example, the result of inhomogeneous etching of a polycrystalline metal surface. In other cases, changes in the relief are the result of the phenomena initiated in the near-surface layers by ionic or plasma impact. These phenomena include ion-induced stresses, dislocation mobility, recrystallization, and changes in the composition of the near-surface layers. Under modification, various structural and phase changes occur which are determined by the parameters of the radiation exposure, which leads to a change in the optical, mechanical, and electrical properties of the surface layers of the material [11,12].

The authors in [13,14,15] discuss hardening of the surface layers of metals at using ion treatment which results in an increase in the dislocation density at a depth of up to 10 µm (the so-called long-range effect), as well as the formation of intermetallic phases. Therefore, it is advisable to choose those ‘metal-ion beam’ systems where intermetallic compounds are formed. In previous works, it was shown that the treatment of a titanium alloy with metal ions leads to a refinement of the surface structure, an increase in the adhesion to the substrate of the nanocomposite coating based on Si-Al-N [16].

These coatings are mainly used to protect aircraft parts from wear and corrosion. In this article, the problem is posed to improve the thermal cyclic resistance of the thermal barrier coatings based on Zr-Y-O deposited on a substrate of copper alloys for thermal protection of rocket engine nozzles, used for reusable activation in order to correct the orbit of a spacecraft. Copper and its alloys have a high coefficient of thermal expansion, while the ceramic coating material of yttrium zirconium oxide, on the contrary, has a much lower value. As a result, the thermal cyclic durability of such a thermal barrier coating does not exceed several cycles, and, accordingly, the service life of the rocket engine is insufficient for controlling a spacecraft during long-term space missions.

We suggest that one of the ways to solve this problem is the modification (structuring) of a substrate using copper alloys (water-cooled nozzle) by heavy metal ions (Zr, Ti) [17,18]. This intermediate operation allows one to form a microporous sublayer in the plane of the interface [19,20]. Then, ZrO₂-based ceramics stabilized by yttrium with the effect of transformation hardening is introduced into microspores, which can significantly decrease the value of the internal stresses of different levels: stresses of the first and second kind and the moment or local stresses. As a result, chains of closed pores are formed in the plane of the interface, which are optimal for obtaining high thermomechanical properties of thermal barrier coatings as was shown by the example of the Ti-Cu system when the ceramic coating Al-Si-N was deposited [21,22].

Therefore, at this stage, an urgent task is to study the morphology, structure, and phase composition of the surface layer of a copper substrate modified with titanium ions, as well as to increase the thermal cycling resistance and adhesive strength of an alloy based on Zr-Y-O as a heat-shielding coating for use in aerospace engineering.

2. Materials and Methods

Copper alloy samples were modified under a continuous titanium ion beam with an energy ~0.5–2.5 keV and an ion current density of 2–20 mA/cm². For deposition the vacuum system KVANT-03MI (Tekhimplant Ltd., Tomsk, Russia) was used [23]. Samples were placed in the camera on the object table in front of the ion source for ion bombardment. The temperature of the samples during the ion bombardment was 900–1000 K. The technological parameters of the treatment are shown in

Table 1. Parameters of treatment of the copper alloy by titanium ion.

Samples	Bias, V	Treatment Time, min	Fluence, Ion/cm2
Initial copper	-	-	-
Treated copper	−900	2	0.6 × 1018
Treated copper	−900	3	0.9 × 1018
Treated copper	−900	6	1.8 × 1018
Treated copper	−900	12	3.6 × 1018

. After the ion beam treatment of the copper substrate by titanium ions, a Zr-Y-O coating was deposited using a magnetron sputtering technique. The magnetron was powered by a pulsed current source with pulse repetition frequency up to 50 kHz and a duty cycle of 80%. During coating sputtering, the substrates were heated to a temperature of 290 ± 10 °C. A molybdenum thermocouple was used to measure the temperature. The thickness of the sputtered Zr-Y-O protective coatings was 6 μm.

The structural phase state of the treated layers of the samples was studied by the TEM method using a JEOL-2100 device (Jeol Ltd., Akishima, Japan). Foils for the TEM studies were prepared by the cross-section method using an Ion Slicer EM-09100IS (Jeol Ltd., Akishima, Japan). The acceleration voltage was 2 kV, the polishing angle was 4.5°. The bright-field images together with the corresponding microdiffraction patterns were used to classify the structures, the grain size and the phase composition of the samples. The X-ray was also used to investigate the phase composition and the crystalline lattice parameters of the coatings. The JCPDS and PDF-2 (International Centre for Diraction Data, Campus Blvd, PA, USA) database was used to interpret diffractograms. The chemical composition of the coatings was determined by the energy dispersive X-ray analysis using a microanalyzer INCA-Energy (Oxford Instruments) with the built-in TEM JEOL-2100 and SEM LEO EVO-50XVP. The grain size, pore size, and dislocation density were determined by the secant method [24].

The resistance of the coatings to cracking and peeling when changing the temperature was determined from the results of thermal cycling of the samples according to the following regime: heating the sample to 1000 °C for 1 min, then forced cooling for 1 min to a temperature of 20 °C, photographing the surface of the sample on the side of the coating using a special Microscope DCM500 camera on an optical microscope BMG-160 (Carl Zeiss, Oberkochen, Germany), the data being transferred directly to the computer, then heating again. The total duration of each cycle, including all the stages of the process—heating, cooling, photographing—was 5 min. Photographing of the coating was also carried out before testing for thermal cycling resistance. For the criterion of thermal stability of the coatings, the number of the cycles until the detachment of 50% of the coating area from the sample surface was selected [22]. After that the tests were stopped. The adhesion of the coatings to the substrate was determined using a Revetest-RST instrument (CSM Instrument, Peseux, Switzerland). To obtain reliable results, three scratches were made on the surface of each coated sample.

3. Results

It is known that irradiation of a copper substrate with heavy ions can cause changes in the surface relief. The above phenomena include the occurrence of ion-induced stresses, the nucleation and the dislocations movement, recrystallization, and changes in the composition of the near-surface layer [13,15].

In Figure 1 one can see the surface morphology of the copper sample treated with titanium ions for 2, 6, and 12 min (which is the same as 0.6 × 10¹⁸, 1.8 × 10¹⁸, and 3.6 × 10¹⁸ fluence). At short modification times (2 min) ion etching begins on the surface of the copper substrate caused by different erosion rates of the adjacent surface areas with different crystallographic orientation. After the bombardment with a high-current flux of titanium ions, the surface becomes a cell-like structure. After two minutes of treatment, there appear etching pits, with some areas being etched faster or slower. When the treatment time increases up to 6 min, the etch pits, become deeper, with the other poorly etched elements protruding above the surface in the shape of whiskers. With the treatment time of 12 min, the latter start to be partially joined by horizontal bridges, which form cavities in the substrate surface layer (Figure 1i,j). Schematic illustration of the etched areas is shown in Figure 1d,h,m.

The bridges are clearly seen in the cross-section of the copper samples bombarded with titanium ions during different treatment times. If the treatment time is short, there are no defects in the cross-section, with only insignificant changes remaining on the surface. With the treatment time of 6 min the pits become deeper. Over the cross-section, one can see the mesh structure and pores in some places. With maximal treatment time, the pores and their size increase. In [20,21] it is shown that the best coating is produced at treatment time of 6 min, under which a continuous mesh pattern without extended pores occurs. In this case, the mesh is completely filled with the coating material. As a result, the maximum adhesion of the coating to the substrate is achieved. At the maximum value of the treatment time, the hollow channels—i.e., the big pores along the coating are peeled off or destroyed.

In Figure 2, one can see the TEM images of a cross-section of a copper treated with titanium ions at the processing times of 6 and 12 min and distribution of the elements by depth for these samples. One can see that the structure of the surface layer of both samples is strongly inhomogeneous. There are very small formations (10–20 nm) and rather large grains (up to 500 nm). The elemental composition is also inhomogeneous over the cross section of the samples. With a processing time of 6 min (Figure 2c) the titanium content is about 40 at % at a depth of 1 µm. After 1 µm in depth, the titanium content decreases by half (20 at %) and after 2 µm, the titanium content gradually decreases to zero at a depth of 4 µm. The titanium content changes somewhat differently in the sample with a treatment time of 12 min (Figure 2d). At shallow depths (up to 500 nm), the titanium content is more than 40 at %, then it decreases down to 20 at % and then gradually titanium content increases up to 50 at % at a depth of 3 µm in depth. Only after 3 µm in depth does the titanium content decrease, reaching zero at a depth of 6 µm.

The pores appear in the modified layer of the sample treated for 6 min, and then after treatment for 12 min their number increases substantially and they are located everywhere. They are clearly visible in the material of both samples (Figure 1 and Figure 2). Figure 3 shows TEM images of part of the surface with pores (a, b) and the pore size distribution for the samples with processing times of 6 min and 12 min. It was found that the average pore size for both samples was approximately the same and its amounts (10 nm). For a sample with a processing time of 6 min, the pores are evenly distributed over the entire surface; their maximum size is no more than 30 nm. For a sample with a processing time of 12 min, the maximum pore size reaches 50 nm. It is interesting to note that in the second case, the pores concentrate between the treated layer and the copper substrate. As a result, the pore size increases, and the modified layer is practically peeled off.

Figure 2 show the electron microscopic images of the modified surface of the samples treated for 6 min (Figure 2a) and 12 min (Figure 2b). It can be seen that, in the sample with 6 min of processing, the grains have a size of about 300–500 nm, while in the sample with 12 min of processing, they are significantly smaller—i.e., about 100–150 nm. In the annealed coarse-crystalline state, the average grain size of copper was about 100 microns [25]. That is, the treatment of the copper substrate with titanium ions leads to a significant grain fining. A dislocation structure is observed in samples treated for 6 min at a distance of 3–4 µm from the treated surface. The dislocation density is 3 × 10¹² m². This value of dislocation density is typical for a weakly deformed material [26]. For a sample processed within 12 min, there are only single dislocations in the field of view. Apparently, in this case, the temperature effect of the ion beam takes place simultaneously with the deformation one. The flexible extinction contours in the grains indicate a high level of elastic internal stresses, whereas the absence of dislocations in the grains and the grain size up to 150 nm indicates a partial course of dynamic recrystallization in the ion treatment process.

The X-ray investigation studies have shown that depending on the time of treatment, the phase composition of the copper substrate can vary greatly. According to the Cu-Ti phase diagram, there are the following phases: TiCu, Ti₂Cu₃, Cu₄Ti, Cu₂Ti, and Cu₃Ti in this system [24]. The types of the crystal lattices, phases, space groups, and the theoretical parameters of the crystal lattices were taken from the PDF-2 databank.

X-ray studies have shown that the surface layer mainly consists of the Cu₃Ti phase having an orthorhombic crystalline lattice. When the processing time increases to 12 min, the predominant phase becomes CuTi with a tetragonal lattice P4/nm, and there is a small quantity of the Cu₃Ti and Cu₄Ti₃ phases in the surface layer [27].

In

Table 2. Phase composition and parameter of the crystalline lattice of the copper treated under titanium ions.

Samples	Phases	The Crystalline Lattice of Copper, a, (Å)
Initial copper M1	Cu	3.614± 0.001
Treatment for 2 min	Cu3Ti, CuTi	3.615 ± 0.001
Treatment for 3 min	Cu3Ti, CuTi	3.614 ± 0.001
Treatment for 6 min	Cu3Ti, CuTi	3.615 ± 0.001
Treatment for 12 min	Cu3Ti, Cu4Ti3, CuTi	3.620 ± 0.001

the phase composition, the parameter of the crystalline lattice of the copper treated by titanium ions are given. One can see that the crystalline lattice parameters of the copper practically do not change when the time of the ion treatment is 2, 3, and 6 min it is close to the tabular value of copper. The change is significant only when the time of the ion treatment reaches 12 min. In this case, the surface layer of the copper is saturated with titanium ions (the atomic radii of copper and titanium are 1.28 and 1.46 Å, respectively).

The cross-section of the copper treated with titanium ions was investigated by TEM. Figure 4 and Figure 5 shows the cross-section of the copper treated by titanium ions at different treatment times. It is shown that the cross section of the treated sample was highly non uniform. There are several areas differing in structure and phase composition. The first area is the copper almost in the original state. Further one can see a transition area, where there are Cu₃Ti, CuTi phases for a sample modified for 6 min (Figure 4c). As for the sample with the treatment time of 12 min, phases with a higher copper content appear such as Cu₃Ti, Cu₄Ti (Figure 5c,e).

Thermocyclic resistance of Zr-Y-O coatings was investigated on samples with preliminary ionic treatment of the coating with titanium ions for 6 min and on samples without treatment. When studying the structure of the copper modified with titanium ions, we have found that the surface treatment for 6 min gives a structure with a continuous mesh pattern and small pores. In this case, the mesh structure is completely filled with the coating material. As a result, the maximum adhesion of the coating to the substrate is achieved, which leads to a higher thermocyclic resistance of the Zr-Y-O coating. It can be seen that the coatings on the samples without preliminary ionic treatment begin to peel off already after two cycles and after the third cycle, 50% of the coating peels off (Figure 6a–c). As for the sample with the preliminary treatment of the substrate with titanium ions for 6 min, the delamination of the coating begins after 33 cycles. Half of the coating will peel off after 40 cycles (Figure 6d–f). That is, the treatment of the substrate surface with titanium ions leads to an increase in the thermal cyclic resistance of the coating by almost an order of magnitude.

The adhesive strength of the coating to the protected surface is an important parameter affecting the thermal cycling resistance of the coating [28]. The adhesion of the coatings to the substrate was determined for samples with pre-processing 6 min and without treatment of the copper substrate with titanium ions. To obtain reliable results, three scratches were made on the surface of each coated sample. It was found that adhesion was 2 times higher for the samples with ionic treatment compared those without ionic treatment and are, respectively 4.1 ± 0.3 and 8.0 ± 0.4 N.

Thus, to increase the thermal cyclic resistance of heat-shielding coatings, we used the method of producing a transition layer by modifying the surface layer of the substrate with a high-current flux of titanium metal ions. The top layer of the substrate turns into a highly effective relaxing material that prevents the destruction of the fragile Zr-Y-O ceramic coating for a sufficiently long time. In this case, doping, nanostructuring of the substrate surface as well as the appearance of nano- and microporosity improve the relaxation characteristics of the three-layer system of “coating—the modified substrate layer—the main volume of the substrate”. It is this complex multilevel structure that makes it possible to increase the thermal cyclic resistance and adhesion of the coating to the substrate.

4. Conclusions

The condition and properties of the surface layer of the substrate are of great importance for increasing the thermocyclic resistance of the ‘heat–protective coating–substrate’ system. It is along the surface of the substrate on the coating (the interface area) that a sharp jump occurs in the change in the structural-phase state and the physical-mechanical properties of the ‘heat–protective coating–substrate’ system, the maximum localization of elastic stress fields occurs. The condition of the substrate surface largely affects the formation of the structure and the properties of the coating itself. Under processing the surface layer of a copper substrate with a flow of ions, the surface layer is modified to a depth of up to 6 microns, its chemical composition and the structural-phase state change. The morphology of the surface acquires a fine-cellular structure with an average cell size ranging from 100 to 600 nm depending on the processing mode. The density of dislocations increases and the level of local internal stresses increases too. The distribution of dislocations over the volume of the surface layer is non uniform. There is a strong fragmentation of the grains of the surface layer to a depth of 1 to 6 μm, the transition of the structure from micrograin to nanograined occurs.

With an increase in the ion energy and ion current density, the content of embedded titanium ions and the depth of their penetration into the copper substrate significantly increase. The content of titanium atoms for the sample with a processing time of 6 min gradually decreases from 40 at % a distance of 1 μm from the surface and decreases to zero at a depth of 4 µm. For a sample, after 12 min of treatment, the alloying element penetrates to a depth of 6 μm.

When the treatment times increase, such phases as Cu₄Ti₃ and Cu₄Ti are observed. With maximum treatment time of 12 min, the CuTi phase becomes dominant.

Thus, modification the copper by titanium ions results in complex, multilevel structures which leads to an increase in the thermal cycling resistance of the Zr-Y-O coating by an order of magnitude and increase in the adhesion strength by 2 times.