简介概要

Correlation between process parameters,grain size and hardness of friction-stir-welded Cu-Zn alloys

来源期刊：Rare Metals2018年第5期

论文作者：Akbar Heidarzadeh Tohid Saeid

文章页码：388 - 398

摘要：In this study, the effects of tool rotational speed, tool traverse speed, and Zn content on the grain size and hardness of the friction-stir-welded（FSWed） Cu-Zn alloy joints were investigated. The microstructures of the joints were examined using optical microscope（OM） and scanning transmission electron microscope（STEM）.Vickers hardness test was conducted to evaluate the hardness of the joints. In addition, the relationships between the process parameters, grain size, and hardness of the joints were established. The results show that the developed relationships predict the grain size and hardness of the joints accurately. The Zn content of the alloys is the most effective parameter on the grain size and hardness, where the tool traverse speed has the minimum effect. The relationship between the hardness and grain size of the joints has a deviation from the Hall-Petch equation due to formation of high dislocation density inside the grains. At higher Zn amounts, the dislocation tangles with high density form instead of dislocation cells, and hence, lower conformity with the Hall-Petch relationship is observed.

详情信息展示

稀有金属(英文版) 2018,37(05),388-398

Correlation between process parameters,grain size and hardness of friction-stir-welded Cu-Zn alloys

Akbar Heidarzadeh Tohid Saeid

Faculty of Materials Engineering, Sahand University of Technology

收稿日期：13 September 2015

Correlation between process parameters,grain size and hardness of friction-stir-welded Cu-Zn alloys

Akbar Heidarzadeh Tohid Saeid

Faculty of Materials Engineering, Sahand University of Technology

Abstract：

In this study, the effects of tool rotational speed, tool traverse speed, and Zn content on the grain size and hardness of the friction-stir-welded(FSWed) Cu-Zn alloy joints were investigated. The microstructures of the joints were examined using optical microscope(OM) and scanning transmission electron microscope(STEM).Vickers hardness test was conducted to evaluate the hardness of the joints. In addition, the relationships between the process parameters, grain size, and hardness of the joints were established. The results show that the developed relationships predict the grain size and hardness of the joints accurately. The Zn content of the alloys is the most effective parameter on the grain size and hardness, where the tool traverse speed has the minimum effect. The relationship between the hardness and grain size of the joints has a deviation from the Hall-Petch equation due to formation of high dislocation density inside the grains. At higher Zn amounts, the dislocation tangles with high density form instead of dislocation cells, and hence, lower conformity with the Hall-Petch relationship is observed.

Keyword：

Friction stir welding; Grain size; Hardness; Cu-Zn alloy;

Author： Tohid Saeid,e-mail:saeid@sut.ac.ir;

Received： 13 September 2015

1 Introduction

Copper and brasses (Cu-Zn alloys) have vast industrial applications because of their special characteristics such as high electrical and thermal conductivities,good combinations of strength and ductility,and excellent resistance to corrosion.Therefore,there is a large demand for welding of these types of alloys.On the other hand,a high heat input condition is needed during conventional fusion welding of the copper alloys due to their high thermal conductivity.Consequently,the higher heat input conditions dispose the joints to distortion,solidification cracking,and high oxidation rate [ 1, 2, 3] .Also,low boiling temperature of zinc leads to its evaporation during fusion welding processes of brasses,which results in color change and formation of a porous and weak layer of copper or copper oxide.Moreover,it is notable that the zinc vapor is toxic and can be harmful to the health of welding operators [ 4] .It seems that friction stir welding (FSW),as a solid-state process,could be a good alternative to overcome the problems arising from conventional fusion welding of copper and Cu-Zn alloys [ 5] .

Despite variant investigations into the pure copper FSW [ 6, 7, 8, 9, 10, 11, 12, 13, 14] ,the studies in the case of brass alloys are rather limited [ 15, 16, 17, 18] .For example,Cam et al. [ 15] revealed that the best combination of the strength and ductility in the case of Cu-10 wt%Zn and Cu-30 wt%Zn brass alloys could be achieved at a traverse speed of 210 mm·min^-1and a rotational speed of 1600 r·min^-1.Emami and Saeid [ 16] showed that the higher traverse speeds or lower rotational speeds caused finer grain size of the Cu-30 wt%Zn joints and hence higher hardness values.Xie et al. [ 18] demonstrated that the higher rotational speeds caused coarser grain sizes in the joints of 5-mm-thick brass plates.Furthermore,they found that increasing the rotational speed had no noticeable effect on the tensile and yield strengths,but it increased the elongation.

According to the above literatures,the FSW parameters have a considerable effect on the microstructure and mechanical properties of the brass alloy joints.Thus,the correlation between FSW parameters,microstructure,and mechanical properties of the brass joints can be very useful for scientific and industrial applications.In this regard,one of the applicable methods is response surface methodology(RSM).Some workers have proved that the RSM can be applied successfully for FSW of different metals and alloys [ 19, 20, 21] .For instance,Rajakumar et al. [ 19] established mathematical models to evaluate the effect of different FSW parameters on the grain size,ultimate tensile strength(UTS),and hardness of the AA6061-T6 aluminum alloy joints using RSM.They showed that the rotational speed of1100 r·min^-1,welding speed of 80 mm·min^-1,axial force of 8 kN,shoulder diameter of 15 mm,pin diameter of5 mm,and tool hardness of HRC 45 resulted in the best mechanical properties.Palanivel et al. [ 20] used RSM to correlate the FSW parameters (tool rotational speed,traverse speed,and axial force) and UTS of AA5083-H1 11aluminum alloy joints.They disclosed that the joint welded at tool rotational speed of 1000 r·min^-1,traverse speed of69 mm·min^-1,and axial force of 13.3 kN had higher UTS.

In addition to developing empirical relationships between the FSW parameters and joint performances,correlating the microstructural features and mechanical properties is a key issue.One of the general methods to determine the relationship between microstructure and strength or hardness of the materials is Hall-Petch (H-P)equation.The H-P equation in terms of hardness can be expressed as follows [ 22] :

where H is the hardness,d is the average grain size,and H₀and k are suitable constants associated with hardness measurements,respectively.Moreover,k is the slope of H-P equation,which indicates the relative strengthening contribution of grain boundaries.It has been revealed that the H-P equation needs to be modified in the case of severe plastic deformation processes of the metals due to formation of the substructures [ 22] .Thus,in the case of friction-stir-welded alloys which undergo a severe plastic deformation,the H-P equation can be deviated from its linear relationship(Eq.(1)).For example,Park et al. [ 23] studied the effect of micro structure on Hall-Petch relationship in the case of the friction-stir-welded (FSWed) thixomolded AZ91D Mg alloy.They showed that the substructures affect the H-P relationship as well as the grain size of the joints.

Although some workers investigated the effect of FSW on the brass joint properties,an investigation into the correlation between FSW parameters,microstructure,and hardness of the brass joints with different amounts of Zn seems necessary.Furthermore,the H-P equation for the FSWed pure copper and brass joints has rarely been discussed according to both the grain size and substructure effects.Therefore,in present study,three kinds of alloys including pure copper,Cu-30 wt%Zn,and Cu-37 wt%Zn brasses were friction-stir-welded under different tool rotational and traverse speeds.The relationships between FSW parameters and joint features (grain size and hardness)were established using RSM.Moreover,the hardness and micro structure of the joints were correlated based on H-P relationship.

2 Experimental

The Cu-Zn plates with different contents of Zn (0 wt%,30 wt%,37 wt%) were used as base metals (BMs) with dimensions of 100 mm×100 mm×2 mm.The plates were annealed at 500℃for 1 h.In order to produce a doublephase structure,the Cu-37 wt%Zn BM was heated at 810℃for 70 min and then was quenched in water at room temperature.Then,the plates were stress-relieved at 250℃for 1 h.The microstructures of the different BMs are shown in Fig.1.

The Design Expert software was used to design the experiments and establish mathematical models.The analysis of variance (ANOVA) was performed to validate the developed models.The considered parameters with their levels and units and the experimental design matrix used in this study are summarized in Tables 1 and 2.The plates were friction-stir-welded at different rotational and traverse speeds according to Table 2.In all of the experiments,a tool with a cylindrical shoulder (12.0 mm in diameter) and a simple cylindrical pin (3.0 mm in diameter and 1.7 mm in length) made of H1 3 hot work tool steel were used.Also,the tilt angle of the tool relative to the normal direction of the plate surface was set constant at 2.5°.

Fig.1 OM images of BM in alloys with different Zn contents:a 0 wt%Zn (pure copper),b 30 wt%Zn (single-phase brass),and c 37 wt%Zn(double-phase brass)

下载原图

Table 1 Coded and actual values of parameters

After FSW,the microstructures of the joints were studied using optical microscope (OM,Olympus 100).The metallographic samples were cut from the joints transverse to the welding direction and then polished and etched with a solution of 20 ml nitric acid and 10 ml acetic acid.Clemex image analysis software was applied to measure the average grain size of the different joints.For deep physical study of the joints,scanning transmission electron microscopy(STEM,Hitachi S-4800) was used.The STEM samples were thin-polished and then double-jet electro-polished using a solution of HP0₄:CH₄O:H₂O=1:1:2 (volume ratio).The Vickers hardness test was performed for hardness measurement in the center of the joints using load of 0.5 N for 10 s.

3 Results and discussion

3.1 Empirical relationships

According to Eq.(2),the response parameter (Y),i.e.,mean grain size (D_av) of the stir zone (SZ) or SZ hardness is a function of input parameters,i.e.,tool traverse speed(A),tool rotational speed (B),and alloy type (here is Zn content)(C).Moreover,Table 3 shows that the Design Expert software suggests the quadratic model for both of the responses.Therefore,in this study,the empirical relationships were developed by means of a second-order polynomial regression model including the main and interaction effects of the input parameters as indicated in Eq.(3):

where X_i and X_j are independent variables,b₀ stands for the mean value of responses,and b_i,b_ii,and b_ij are linear,quadratic,and interaction constant coefficients,correspondingly.

Considering A,B,and C parameters,Eq.(3) can be stated as Eq.(4):

The coefficients of Eqs.(3) or (4) were calculated by Design Expert software using the following formulas [ 24, 25, 26, 27, 28] :

Finally,after coefficient calculations,the mathematical models for D_av and hardness (H) of SZs have been established as Eqs.(9) and (10),respectively,

The experimental and predicted values (by Eqs.(9),(10)) of D_av and hardness are summarized in Table 2.Also,the normal plots of residuals and the predicted response versus actual response plots are,respectively,illustrated in Fig.2a-d,for the response D_av and hardness.The normal probability plot indicates whether the residuals follow a normal distribution or not,in which case the points will follow a straight line.If the points do not follow a straight line,it means that a transformation of the response may provide a better analysis [ 29] .Figure 2a,b demonstrates that errors are extended normally because the residuals follow a straight line.Figure 2c,d reveals that the predicted response values are in good agreement with the actual ones within the ranges of the process parameters,because the data points are split evenly by the 45°line.In other words,there is a strong correlation between the model's predictions and its actual results.

The significance of the models and their coefficients can be determined according to the ANOVA results as shown in Tables 4 and 5.In ANOVA results,the F value,P value,R² and adjusted R² can be used to recognize the significance of the models and coefficients,where F value is a test for comparing curvature variance with residual (error)variance,P value is the probability of seeing the observed F value if the null hypothesis is true,and R² or R-squared is a measure of the amount of variation around the mean value explained by the model.In summary,larger F value,R² and adjusted R²,and lower P value disclose the more significant model and coefficients [ 30] .According to the ANOVA results for D_av (Table 4) and hardness (Table 5),the F value,P value,R² and adjusted R² for the predicted models of D_av,and hardness are 47.3300,<0.0001,0.9617,0.9414 and 91.8900,<0.0001,0.9799,0.9692,correspondingly.Thus,the results show that the developed models predict the responses adequately.

下载原图

Table 2 Design layout including experimental and predicted values

下载原图

Table 3 Results of different conducted models for responses of grain size and hardness,where 2FI being two-factor interaction model

Fig.2 Normal plots of residuals a,c and predicted response versus actual response plots b,d for responses:a,b D_av,and c,d hardness

下载原图

Table 4 ANOVA data for response D_av

Furthermore,in ANOVA tables,P values of smaller than 0.0500 confirm that a coefficient is significant,and P values of larger than 0.1000 validate that a coefficient is not significant [ 31] .Consequently,A,B,C,AB,BC,and C² are significant terms for both of the predicted models.Therefore,by considering thesignificant terms only,the final relationships can bestated as follows:

下载原图

Table 5 ANOVA data for response hardness

Fig.3 Counters at different conditions a-c and perturbation plots d for response D_av.Numbers in a-c being values of D_av

The F values demonstrate that the order of more significant terms in the relationships developed for D_av and hardness (Eqs.(11),(12)) are as follows,respectively:C>B>A>BC>C²>AB and C>B>A>C²>BC>AB.

3.2 Effect of parameters on Dav and hardness

The contour and perturbation plots for D_av are illustrated in Fig.3a-d.As well,the microstructures of the joints welded at different welding conditions are shown in Fig.4.According to Figs.3 and 4,larger traverse speeds and lower rotational speeds cause smaller D_av.Based on the fine and equiaxed grains in the SZ of the joints,it can be concluded that dynamic recrystallization (DRX) occurs during FSW.Since FSW is a hot deformation process due to the existence of heat and deformation,the grain size of the joints will be controlled by thermomechanical parameters such as strain rate and temperature [ 32] .The strain rate and temperature during thermomechanical processes can be related by Zener-Holloman as follows [ 33] :

Fig.4 OM images of joints welded under different welding conditions corresponding to experimental Nos.in Table 2:lower heat input,a No.1(0 wt%Zn),b No.10 (30 wt%Zn),and c No.19 (37 wt%Zn);higher heat input,d No.7 (0 wt%Zn),e No.16 (30 wt%Zn),and f No.25(37 wt%Zn)

where Z is Zener-Holloman parameter, is strain rate,T is temperature,Q is activation energy,and R is the gas constant. and T can be calculated using following equations,respectively [ 33] :

where R_m,r_e,and L_e,respectively,denote the half of tool rotational speed,the effective radius,and depth of the dynamically recrystallized zone;k andαare constants between 0.04-0.06 and 0.65-0.75,respectively;ωis tool rotational speed;v is tool traverse speed;and T_m is the melting point of the alloy [ 34] .In addition,it has been demonstrated that the grain size during thermomechanical processes has an inverse relationship with Z.Therefore,according to the Eqs.((13)-(15)),larger causes smaller D_av,where higher T leads to bigger D_av.It can be concluded that and T are challenging in determining the final D_av during the thermomechanical processes of the alloys.In this study,according to Figs.3 and 4,with the rotational speeds increasing (higher T and lower ) and the traverse speeds decreasing (higher T),D_av increases continuously.As a result,in present investigation,the dominant factor which controls the final D_av is T.

Furthermore,Figs.3 and 4 reveal that the higher Zn content of the alloys causes smaller D_av under the same process condition.This can be explained by the effect of Zn on the stacking fault energy (SFE) of the different alloys.The effect of SFE on the steady state minimum grain size (d_min) of the severely plastic deformed (SPDed) materials can be stated using the model developed by Mohamed [ 35] :

where b is Burgers vector,γ_SFE stands for amount of SFE,G is shear modulus,andαand q are constants.The d_minvalue of the materials processed by different SPD methods agrees with Mohamed model [ 36, 37, 38, 39] .Recently,Morishige et al. [ 40] have investigated the effect of Mg content on d_min value of the friction-stir-processed (FSPed) 5052 and5058 aluminum alloys.They showed that with the Mg content increasing,the d_min value of the FSPed alloys decreased due to lower SFE.According to the relationship reported by Gallagher [ 41] ,the effect of Zn content on the SFE of the Cu-Zn alloys (γ_Cu-Zn) could be expressed through the following equation:

Fig.5 Counters at different conditions a-c and perturbation plots d for response hardness.Numbers in a-c being values of hardness

where yo stands for SFE of the pure copper,x_Zn is Zn concentration, denotes the solubility limit of Zn at high temperature,and k_γis a dimensionless constant.Equations (16) and (17) reveal that the grain refinement can be achieved with Zn content increasing in Cu-Zn alloys during FSW.Similar results have been reported for the Cu-Zn alloys processed by high-pressure torsion (HPT) and by HPT followed by cold-rolling [ 42] .It seems that this effect of Zn in Cu-Zn alloys is very similar to that of Mg in AlMg alloys [ 40] ,which is attributed to the reduction in dislocation mobility and hence recovery rate in these types of solid solution alloys.

3.3 Hardness-grain size correlation

The effects of FSW parameters and Zn content on the hardness of the joints are illustrated in Fig.5 using counter and perturbation plots.According to Fig.5,lower rotational speeds and higher traverse speeds cause higher hardness values,which is due to smaller grain size as explained in Sect.3.2 (Figs.3,4).Furthermore,at constant FSW parameters,the hardness values increase with Zn content of the alloys increasing.This can be due to the smaller grain size (Figs.3,4) and solid solution strengthening effect of Zn in alloys with higher amounts of Zn.

For correlation between hardness and grain size of the joints,the H-P relationships were estimated according to the data in Table 2.As shown in Fig.6,the H-P relationships for the joints of different alloys can be stated as follows:

where d refers to mean grain size of the joints.According to the R² values (0.94,0.75,and 0.74,respectively,for alloys with 0 wt%,30 wt%,37 wt%Zn),the H-P equations have deviation from their linear relationships.The reason of this deviation is the fact that in the H-P relationship,only the high angle grain boundaries are considered as obstacles to the dislocation movement [ 43] .Thus,the existence of substructures such as different dislocation structures,precipitates,and second phase particles can affect the H-P relationship [ 44, 45, 46, 47] .These microstructural features prevent the dislocation movement and pin them at an interval smaller than the grain size and hence reduce the effect of grain size on the hardness.

Fig.6 Plots for H-P relationships of alloys with different Zn contents:a 0 wt%Zn (pure copper),b 30 wt%Zn (single-phase brass),and c 37 wt%Zn (double-phase brass)

Fig.7 STEM images of joints for alloys with different Zn contents:a 0 wt%Zn (pure copper),b 30 wt%Zn (single-phase brass),and c 37 wt%Zn (double-phase brass)

Some investigations have shown that the FSW causes formation of fine and equiaxed grains with high density of dislocations [ 2, 5] .The STEM images of the joints for different alloys are shown in Fig.7,revealing that the grain interiors contain a high density of dislocations.According to Fig.7,the high density of dislocations,especially in the case of Cu-30 wt%Zn and Cu-37 wt%Zn alloys,causes deviation from linear H-P relationship.Furthermore,from R² values and slope of the H-P relationships shown in Fig.6 (146,124,and 105 are slop values,respectively,for alloys with 0 wt%,30 wt%,and 37 wt%Zn),with the Zn content increasing (lower slope and R²),the effect of grain size decreases in the H-P equation.The lower SFE of the alloys containing higher Zn amounts prevents the dislocation mobility and annihilation and hence causes formation of high density of dislocation with tangle structures(Fig.7b,c).In contrast,higher SFE of the pure copper results in more mobility and annihilation of dislocations and therefore lower dislocation density with cell structures(Fig.7a).It seems that the tangle structure of the dislocations (Fig.7b,c) pins the mobile dislocations more than the annihilated and cell structures (Fig.7a).

4 Conclusion

In this study,the effects of FSW parameters and Zn content on the D_av and hardness of the Cu-Zn alloy joints were investigated,and the relationships between parameters and responses were correlated.RSM was used to correlate the process parameters (tool traverse speed,tool rotational speed,and Zn content of the alloys) and the responses (D_avand hardness).The ANOVA data show that the developed relationships can predict the responses accurately.The order of the more significant and effective parameters on the D_avand hardness of the joints are as follows:Zn content>tool rotational speed>tool traverse speed.The effect of Zn is due to its influence on the SFE of the Cu-Zn alloys,where the effects of tool traverse and rotational speeds result from their influence on the Zener-Holloman parameter.In addition,the relationships between hardness and D_av of the joints show a deviation from H-P equation.The origin of this deviation is the formation of dislocations with high density in grain interiors.According to the slope and R² of H-P equation,the pure copper (0 wt%Zn) joints have more conformity with the H-P linear relationship compared to those of the brass(30 wt%and 37 wt%Zn) joints.This behavior is due to the effect of Zn content on the SFE and hence its influence on the formation of different dislocation structures,i.e.,dislocation cells and tangles in the case of pure copper and brass joints,respectively.