Study on chemical corrosion properties of titanium alloy in 2A14 aluminum melt

Titanium alloy radiation rods have excellent physical and chemical properties compared to other materials, and are commonly used for ultrasonic casting of 2A14 aluminum alloy. However, titanium alloys are chemically corroded in high-temperature aluminum melts for a long time, making it difficult to precisely regulate the elemental composition during casting. In order to better understand the high-temperature chemical corrosion mechanism of titanium alloy radiation rods, this research looks into the corrosion morphology, weight loss, surface roughness, and reaction layer. The study’s findings suggest that the rate of chemical corrosion of titanium alloy in high-temperature aluminum melt is often inversely correlated with the degree of roughness, with the degree of roughness changing nonlinearly during the corrosion process. Titanium alloy weight loss rates with roughness Ra0.4 μm, Ra7.2 μm, Ra9.5 μm and Ra9.8 μm are 0.16 mg per min, 0.25 mg per min, 0.37 mg per min and 0.29 mg per min, respectively. The corrosion product of the chemical corrosion process is TiAl3, which is granular. Under varying roughness conditions, the solid-liquid interface of Al/Ti emerges reactants after 4 min, and the TiAl3 reaction layer arises after 12 min. Furthermore, the reaction layer with little roughness is flat and compact, whereas the reaction layer with great roughness is loose and contains many faults. At the same time, the growth rate of the reaction layer decreases slightly. And the greater the surface roughness, the greater the TiAl3 reaction layer grows at the titanium alloy matrix.


Introduction
During the ultrasonic casting process of alloys [1][2][3][4], the radiation rod is directly immersed in the melt.It is impacted by ultrasonic cavitation [5][6][7][8] as well as high-temperature aluminum melt, which causes corrosion and damage to the end face and sides of the radiation rod.In the high-temperature aluminum melt, the metal atoms in the radiation rod become active.A portion of metal atoms react with aluminum liquid to generate compounds that enter the aluminum melt, while the other portion of metal atoms directly enter the aluminum melt.This leads to impurities doping in the aluminum melt, which affects the performance of the aluminum melt.The study of high-temperature melt corrosion mechanisms has received increasing attention in order to avoid the short lifespan and high cost of metal caused by high-temperature corrosion damage.Corrosion in high-temperature melts is influenced by various factors such as melting, dissolution, and chemical reactions.The corrosion pattern will undergo modifications based on variations in the temperature of the metal melt, the types of solid/liquid metals, and the external environment [9][10][11].
Researchers have conducted research on the application of ultrasonic radiation rods [12][13][14][15][16] over the past few years.Shi et al [17] prepared an 2A14 aluminum alloy using a ceramic steel structure radiation rod.The grain refinement effect at the edge of the ingot is the best, with smaller secondary phases and more solute elements that can be dissolved.The ability to suppress segregation at the edge of the ingot is stronger.Liang et al [18] employed ultrasonic manufacturing for the casting of 35CrMo steel and examined the impact of radiation rod length on the performance of steel liquid.When the length was 135 mm, the transmission of ultrasound proved to be the most efficacious.They investigated the substance of radiation and found that silicon nitride ceramic radiators can effectively avoid corrosion in molten metal with high stability.Kong et al [19] prepared low-carbon steel by employing titanium alloy radiation rods and found that when ultrasonic treatment was employed, pearlite was broken, reducing its average length from 550 to 140 μm.The corresponding aspect ratio was decreased from 12 to 1. Zhang et al [20] had successfully produced an alloy of Al-Zr-Ti alloy by employing the ultrasonic casting technique.Grain refinement occurs when the undercooling is high enough to activate the primary intermetallic compounds or dispersed inoculants.Dai et al [21] analyzed the thermodynamics and kinetics of titanium alloys' oxidation at high temperatures, and they also conducted prospective research on the upcoming trend of titanium alloys' oxidation modification at high-temperatures.It was summarized in the following table 1 [17][18][19]22]: For the use of titanium alloy radiation rods in casting process, existing research mostly focuses on the corrosion of titanium alloy radiation rods caused by ultrasonic cavitation [23,24].Dong [25] studied the hightemperature oxidation performance of Ti-6Al-4V alloy in air.Very few researchers have studied how well ultrasound radiation rods can resist chemical corrosion at high-temperatures.As a results, this article conducts a thorough investigation into the corrosion phenomenon of titanium alloy radiation rods in high-temperature aluminum alloy melts.The titanium alloy (Ti-6Al-4V) used for ultrasonic casting is grade TC4 [23], which belongs to (α+β) titanium alloy family.Table 2 shows its composition.A number of sets of titanium alloy samples were taken from ultrasonic radiation rods and processed into different roughness, then soaked in hightemperature aluminum melt.The surface morphology and contour, weight loss and weight loss rate, roughness and the growth of the interface reaction layer were investigated.The solid-liquid interface reaction layer was identified using EDS and XRD.The generation time of corrosion products is additionally examined to discover the corrosion behavior of titanium alloy radiation rods with various roughness in aluminum melt.

Sample preparation
The titanium alloy radiation rod is cut to a size of 25 mm ×25 mm×10 mm sample blocks.These samples were then processed to obtain different original roughnesses of the sample surfaces.The samples with different roughness are called A, B, C, and D. A sample was finely machined on a lathe, while B, C, and D were all rough machined.This created a clear metallic luster on the original surface of the sample.In experiments a sample with a certain degree of original roughness was used to compare the corrosion behavior of rough and smooth surfaces.To achieve the same roughness, the sample's surface was polished with 2000# sandpaper, then the samples were washed with alcohol and acetone, and finally dried with warm air.A graphite crucible containing 5.0 Kg of 2A14 aluminum alloy ingots was placed in a resistance furnace at 800 °C for melting.The aluminum alloy was completely melted, and then stirred to remove the oxide layer from the surface of the aluminum melt.The resistance furnace settings were then adjusted to lower the melt temperature to 700 °C for insulation [26,27].Figure 1 shows taht the titanium alloy sample block is placed at the bottom of the crucible and the sample is removed every 12 min, which is mainly used to study the differences in corrosion weight loss, roughness changes, and reaction layer growth of titanium alloys with different roughness levels in an aluminum melt environment at 700 °C.The sample was cleaned using physical and chemical methods after the corrosion experiment was complete, and then weighed.The titanium alloy was immune to chemical reactions with hydrochloric acid and sodium hydroxide, whereas Al could react with acids and bases.Hence, the samples were thoroughly cleansed with hydrochloric acid and sodium hydroxide respectively, followed by ultrasonic cleaning.The samples were then dried in warm air and weighed using an electronic scale with a precision of 0.1 micrograms.The weight of the samples was determined by taking an average of three measurements, and a weight loss rate graph was constructed based on this.
Another set of samples was carried out according to the above experimental process without cleaning the aluminum alloy attached to the surface of the titanium alloy in order to obtain the profile reaction layer of the titanium alloy samples at different stages of the reaction.A wire cutting machine was used to cut the sample profile, then it was ground and polished.Chemical corrosion experiments were conducted in aluminum melt for a duration of 1 min, 2 min, 3 minK and 11 min, respectively, with the aim of determining the precise generation time of Al/Ti interface reactants.The sample preparation and testing were identical to the above.

Microstructural characterization
A super depth of field microscope (VHX5000) was used to examine the sample's surface.The original surface roughness value of the sample was tested with an optical surface profiler (Wyko NT9100).High resolution scanning electron microscopy (SEM:MIRA3 TESCAN) revealed the morphology of the corroded sample surface and section reaction layer.Energy-dispersive x-ray spectroscopy (EDS:OxfordX-Max20) and x-ray diffraction (XRD) in a Rigaku 600 x-ray diffractometer were used to perform qualitative analysis and identify the reaction layer components.The XRD was operated at a scanning speed of 0.02°s −1 at 40 kV, using Cu Ka radiation (wavelength λ Ka = 1.54056Å).

Corrosion micromorphology
Figure 2 shows the microstructure of sample A before and after the corrosion happened.Figure 2(a) shows that the surface of sample A displays obvious machining marks, distinct contour lines, and a smooth surface without apparent pits.Figure 2(b) shows that the material's surface deformation is very small because the experiment only lasted a short time.But a certain amount of corrosion has happened and the roughness has also increased.There are pits of varying sizes distributed on the corroded surface, especially at the protrusions on the original surface of the material.This causes the originally sharp surface processing marks to become blurred due to corrosion.

Corrosion weight loss rate
The weightlessness method was used to characterize the corrosion degree of titanium alloy materials in hightemperature aluminum melts, considering the convenience of experimental equipment and the accuracy of experimental results.The initial weight, post experimental weight, and total weight loss of each group of samples with different roughness are presented in table 3. The weight loss rate of titanium alloy samples in 700 °C aluminum melt over time was summarized and plotted in figure 3, in order to visually see the magnitude of the weight loss rate of the material at different experimental periods.The formula for determining the rate of weight loss: Where v t , m t , and t respectively represent the weight loss rate, weight, and experimental duration at that time.
Figure 3 shows that if the surface is smooth, the less weight the block sample loses, and vice versa.In four groups of samples, the weight loss rate remained relatively stable under high-temperature aluminum melt corrosion.The average weight loss rates for samples A, B, C, and D are 0.16 mg per min, 0.25 mg per min, 0.37 mg per min, 0.29 mg per min, respectively.But the Ra9.8 sample loses weight a little less than the Ra9.5 sample.This means that the rougher the surface of the sample, the more weight it loses.
Figures 4(a) and (b) show the surface morphology of Ra9.5 and Ra9.8 samples, respectively.It can be seen that the Ra9.5 sample is a little rougher, but its surface is smoother and its maximum surface protrusion is over 40 μm.But the height of the convex body on the surface of Ra9.8 is not more than 25 μm.Figures 4(c) and (d) depict the morphology curves of two samples at their respective sampling lengths.The maximum size pit found on the Ra9.5 sample has a height of 44.7 μm and a width of 272.8 μm, which is approximately twice that of the Ra9.8 sample.The greater width of the aluminum melt results in a greater contact area between the pit and the   aluminum melt.The aluminum melt rapidly spreads on the surface of the titanium alloy, which leads to a higher rate of diffusion, chemical reaction, and recombination of Ti atoms in the aluminum melts.On the contrary, the surface roughness of a smooth sample is smaller, and the fewer pits there are on the surface, the smaller the contact area with the aluminum melt.The diffusion rate of exposed Ti atoms in aluminum melt is hindered, resulting in a decrease in the chemical reaction rate and a decrease in their weight loss rate.It is important to note that the weight loss rate is influenced by Rz (micro roughness cross height) and Ry (maximum contour height), but these two parameters are not the main research focus of this article.

Surface roughness
In order to further investigate the mechanism of the influence of original surface roughness on hightemperature corrosion of titanium alloys, curves of the variation of different original surface roughness over time were plotted based on the measured data during the experimental process, as shown in figure 5.The samples with an initial surface roughness of Ra0.4 μm, Ra7.2 μm, Ra9.5 μm and Ra9.8 μm are subsequently increased to Ra0.7 μm, Ra7.7 μm, Ra10.1 μm and Ra10.2 μm, respectively, following a 48 min soak in high temperature aluminum melt.From the results, it can be seen that the roughness of each sample has changed little and improved just somewhat.However, the roughness increase of the sample with high roughness is greater.Its variation law is usually the same as the above weightlessness.
The height difference between a material's surface and its roughness is related to how rough it is.The rougher the surface, the higher the height difference will be.The diffusion rate of Ti atoms from the matrix to the Al melt is more different on the surface with a bigger original roughness because the difference in vertex height between convex and concave is bigger.With the increase in reaction time, the contour of the edge of the bulge and the slope of the surface will become more obvious, which will further increase the height difference of the original surface contour.
Figures 6(a) and (b) show the surface contour curves of sample A (Ra = 0.4) during the experiment time of 0 min and 48 min, respectively.As shown in the figure, the original surface of the material had a small contour fluctuation, and only the local concave-convex body was sharp.However, after 48 min, the fluctuation of the material surface was intensified, especially the original concave-convex body was more clearly displayed on the matrix and appeared more sharp.The maximum contour height difference of a concave-convex body increased from 4 μm to 6 μm.

Microstructure
Figure 7(a) shows the profile microstructure of the corrosion layer of sample A after 48 min of soaking in 700 °C aluminum melt.The figure shows that the titanium alloy is corroded by the aluminum melt to form a layer material.EDS was used to analyze the components of layered materials, and the results of energy spectrum analysis showed that the atomic ratio of Al and Ti at point 1 was approximately 3:1.EDS functions as an elemental analysis method and does not have the capability to identify precise compounds.But it could be  confirmed that the layered material was TiAl 3 comparing the Al-Ti binary phase diagram and combining with reference [25].
Figure 8 shows SEM images of reaction layers from four groups of samples with different roughness, soaked in melt for different times.Titanium alloy samples with different roughness will react with Al, and the thickness of the reaction layer will increase with the increase in experiment time.At 4 min, the reactants began to form at the Al/Ti solid-liquid interface at 4 min.When soaked for 12 min, a reaction layer about 1 μm thick appeared at the solid-liquid interface.When soaked for 48 min, the thickness of the reaction layer increased to 5 μm.However, the reaction layer is a little different in density.The more roughness of the sample, the more loose the reaction layer.
The morphology of the reaction layer at the solid-liquid interface exhibited slight variations among all groups.Figure 8(c) shows that the thickness of the reaction layer formed at the Al/Ti interface is uniform and without obvious defects, and the interface between the matrix and the reaction layer is smooth at 48 min.Figures 8(f), (i), (l) show that the thickness of the reaction layer at the interface is inconsistent and there are many cracks.TiAl 3 with granular is distributed near the reaction layer and floats away from the interface.This happens because sample A is smooth, and Ti atoms dissolve on the surface at the same rate.This makes TiAl 3 reaction layers grow at similar rates in different places.Therefore, a dense reaction layer is formed.But samples B, C and D have a relatively large roughness and the chemical reaction rate is different, resulting in a loose texture of the TiAl 3 reaction layer.On the less smooth surface, the Ti atoms can diffuse further away from the matrix into the aluminum melt.So that the TiAl 3 can be separated from the reaction layer.At the same time, it also explains the change in the weight loss rate of the titanium alloy sample in section 3.2, which is caused by the change in the diffusion rate of the reaction layer particles to the aluminum melt.
Figure 9 shows the growth curves of the interfacial reaction layer over time for samples with different roughness soaked in aluminum melt.The complete reaction layer did not appear until 4 min for each sample, so the thickness of the interface reaction layer was set to 0 at this time.The thickness of the reaction layer at the solid-liquid interface of samples A, B, C and D increased from 0 to 3.2, 4.0, 4.7 and 4.7 μm during the experiment period of 4 to 48 min.It is also found that the growth rate of the reaction layer decreases with time.After the titanium alloy is immersed in the aluminum melt, the titanium element will dissolve and diffuse into the aluminum liquid for chemical reaction, forming the TiAl 3 compound.A dense reaction layer will soon form at the Al/Ti solid-liquid interface, as shown in figure 8(b).If Ti in the matrix wants to enter the aluminum liquid and react with Al, it needs to cross the TiAl 3 layer after a certain time.With the thickness of TiAl 3 layer increasing, it is harder for Ti to escape from the matrix, so the growth rate of the reaction layer also slows down.However, due to the poor density of the reaction layer formed by samples B, C and D, it is less difficult for Ti atoms to escape, which results in a higher formation rate.The formation rate is consistent with its roughness.

XRD analysis of reaction layer
In order to further confirm the phase of the reaction layer, the components of reaction layer that were generated at 12 min were analyzed by XRD.XRD examination of sample implies that there were almost only α-Al and TiAl 3 present in the sample A, as shown in figure 10.Some unknown phases occurred, but with very minor fraction.So this substance can be confirmed as TiAl 3 combined with the above EDS spectrum.

Discussion
At the beginning of contact between Al/Ti liquid and solid, Ti atoms dissolve and diffuse in the aluminum melt.At the same time, some Al atoms move around in the aluminum melt and into the Ti matrix.The diffusion mode is mainly composed of Ti atoms diffusing into liquid aluminum and auxiliary Al atoms diffusing into the titanium matrix.Based on the diffusion characteristics, the concentration of Ti atoms decreases from the surface of the titanium alloy to the aluminum melt.Therefore, the rapidly saturated Ti atoms near the matrix can immediately react with Al atoms to form the compound TiAl 3 , which is only distributed on the matrix surface as scattered particles at the beginning.The diffusion rate of Ti atoms in titanium alloys with small surface roughness is consistent with the increase of infiltration time due to the uniform height of the surface profile.As a result, a uniform and dense reaction layer gradually forms at the liquid-solid interface of Al/Ti.The appearance of the reaction layer makes it harder for Ti atoms to escape to the melt, so the concentration near the wall decreases and the growth rate of the TiAl 3 reaction layer from the Ti matrix to the Al melt also decreases.The reaction layer is uniform in all places, and the growth rate is roughly the same, but it decreases slightly with time.It is closely bound to the Ti matrix.
The matrix of titanium alloy with a rough surface has different surface profiles and more cracks than the smooth one.Therefore, the actual contact area of Al/Ti is greater than that of a surface with less roughness, and its diffusion rate is also greater.The uneven distribution of Ti atoms concentration diffused from titanium alloy are attributed to the varying height of contour on the surface, resulting in the formation of a TiAl 3 reaction layer with varying thickness and discontinuity.The interior of the reaction layer is also relatively loose, which is susceptible to the flow of aluminum liquid and consequently separates from the interface into the aluminum melt.The loose reaction layer on the rough surface makes TiAl 3 particles break away more easily.The growth rate of reaction layer on rough surfaces is greater than that on smooth surfaces.With the increase of reaction time, the reaction layer of TiAl 3 grows from the Ti base to the Al liquid side, and the growth rate also decreases, but the growth rate is still higher than that of the smooth surface.

Conclusions
In this paper, the corrosion properties of titanium alloys with different surface roughness were studied, and the surface morphology and profile, weight loss and weight loss rate, roughness, and growth of the interfacial reaction layer were analyzed.The reaction layer at the solid-liquid interface was identified by EDS and XRD  phase analysis, and the formation time of corrosion products was observed.The corrosion behavior of a titanium alloy radiation rod with different roughness in aluminum melt was investigated.The conclusions are as follows: (1) The chemical corrosion of titanium alloys occurs in aluminum melts.The rougher the surface of titanium alloy, the higher the weight loss rate, and the weight loss rate is nonlinear during the corrosion process.The weight loss rates for titanium alloys with roughness of Ra0.4 μm, Ra7.2 μm, Ra9.5 μm and Ra9.8 μm are 0.16 mg per min, 0.25 mg per min, 0.37 mg per min and 0.29 mg per min, respectively.The weight loss rate is influenced by surface smoothness.The weight loss rate of a surface with poor flatness and roughness is greater than that of a surface with good flatness and roughness.
(2) The Al/Ti interfacial material formed by chemical corrosion between titanium alloy and aluminum melt is TiAl 3 compound.Under different roughness conditions, reactants appear at the Al/Ti solid-liquid interface in about 4 min, and TiAl 3 reaction layer appears in 12 min.
(3) The reaction layer at the interface with low roughness is flat and dense, while the reaction layer at the interface with high roughness is loose and has many defects.Furthermore, the growth rate of the reaction layer decreased slightly with the reaction time.The higher the surface roughness, the higher the growth rate of the TiAl 3 reaction layer on the matrix of titanium alloy material.

Figure 3 .
Figure 3.The curve of weight loss rate curve.

Figure 4 .
Figure 4.The surface morphology of samples, (a) Ra9.5, and (b) Ra9.8.(c) and (d) shows the morphology curves of two samples at their respective sampling lengths.

Figure 7 .
Figure 7. (a) Profile topography of the reaction layer, (b) EDS element analysis.

Figure 8 .
Figure 8. SEM images of reaction layers with different experimental duration.

Figure 9 .
Figure 9. Reaction layer growth curves of samples with different roughness.

Figure 10 .
Figure 10.XRD pattern of sample A at 12 min.

Table 1 .
Summary of different types of radiation rods and alloys.

Table 2 .
Contents of major solute elements in the TC4 (wt%).