INTRODUCTION
Friction Stir Welding (FSW) is a relatively new joining process that is presently
attracting considerable interest among the researchers. The FSW is a solid state
welding/joining process where a machine rotates, plunges and then traverses
a specially shaped FSW tool along a joint/abutting edge to form a weld (Thomas,
1991). The rotation action and the specific geometry of the FSW tool generates
friction and mechanical working of the material which in turn generates the
heat and the mixing necessary to transport the material from one side of the
joint line to the other for welding. As the friction stir welding tool plays
a critical role in the formation of the weld, proper selection of the tool profile
yields a good weldability. As pointed out by Mishra and
Ma (2005), most of the tool designs are based on intuitive concepts. For
example, the length of the shank is chosen in such a way that it must be held
firmly by the machine during the welding process. Therefore, in this work, a
systematic study of the effects of the tool pin profile on the strength and
the hardness of the friction stir welded section of AA6063 alloy plates was
investigated.
MATERIALS AND METHODS
AA6063 plates were prepared using standard procedures as per the dimensions
shown in Fig. 1 for making butt joints. The factors/parameters
chosen for the study were axial load, rotational Speed and welding speed at
three levels along with the variation in the tool profile (square and pentagonal
cross section).

Fig. 1: 
Test plate after FSW, all dimensions are in mm 
The details of the square and the pentagonal pin profile tools used for the
welding are shown in Fig. 2a, b, respectively.
The tool shoulder diameter was fixed as 18 mm based on the earlier studies (Elangovan
and Balasubramanian, 2008a; Elangovan and Balasubramanian,
2007; Elangovan and Balasubramanian, 2008b). The
length and the diameter of the tools were arrived by FEM analysis as per the
details given in Karthikeyan and Mahadeven (2010). The
friction stir welding was performed using a Czechoslovakian made vertical milling
machine fitted with a specially designed retrofit as shown in the Fig.
3 (Minton and Mynors, 2006).

Fig. 2(ab): 
Friction stir welding tools, (a) Square pin profile and (b)
Pentagonal pin profile, All dimensions are in mm 
The experiments were performed based on the BoxBehnken experimental design
of response surface methodology. For three factors, the BoxBehnken design offers
the advantages of requiring a fewer number of runs (Kim
et al., 2006; Balasubramanian, 2008). The
process parameters and the design matrix are given in Table 1
and 2, respectively.
RESULTS AND DISCUSSION
In the present study the hardness and the tensile strength at the welded section
were taken as the responses parameters.

Fig. 4: 
Tensile test specimen, All dimensions are in mm 
Table 1: 
The process parameters 

Table 2: 
Design matrix 

The Hardness is was measured using Rockwell Hardness ‘H’ scale with
the 3.125 mm ball indenter and under a load of 60 kgf. The tensile tests were
carried out with the gradual application of the load in a Universal Testing
Machine. Figure 4 shows the dimensions of the tensile test
specimens cut out longitudinally from the welded plates. Fifteen experiments
were carried out randomly for each tool profile as per the recommendations of
the design matrix table in order to avoid any bias. The results of the hardness
and tensile test are presented in Table 3.
Regression modelling: The response function representing the properties can be expressed as the function of the axial load, rotational speed and traverse speed for FSW operation as shown in Eq. 1:
For the second order relations, with ‘k’ number of factors, the model will be of the regression type given by the expression 2:
Table 3: 
Response tabulation 

Table 4: 
Coefficient values 

With three factors and three levels, the polynomial is expressed as Eq. 3:
where, b_{0} is constant, b_{1}, b_{2}, b_{3} are coefficients of linear terms, b_{11}, b_{22}, b_{33} are coefficients of second order terms, b_{12}, b_{13}, b_{23} are the coefficients of the interaction terms.
The linear and the second order terms provide the effect of the individual factors and the interaction terms provide the combined effect of the parameters.
Determination of the values of coefficients: The calculated values of the coefficients for different responses are presented in Table 4. After determining the coefficients, the regression models are developed.
Regression models: The second order regression equations developed to
predict the hardness and the tensile strength at the weld section obtained by
the two different tool profiles are given by the Eq. 47.
For square pin profile:
For pentagon pin profile:
Adequacy check: The adequacy of the models so developed was tested using
the analysis of variance technique (ANOVA).
Table 5: 
Adequacy check for the square pin profile tool 

Table 6: 
Adequacy check for the pentagon pin profile tool 

Table 7: 
Results of confirmation tests 


Fig. 5(ae): 
Weld bead structures for square tool profile, (a) L: 1.5 KN,
S: 1000 rpm, F: 0.34 mm sec^{1}, (b) L: 1.5 KN, S: 710 rpm, F:
0.24 mm sec^{1}, (c) L: 1.1 KN, S: 1000 rpm, F: 0.5 mm sec^{1},
(d) L: 1.5 KN, S: 1400 rpm, F: 0.24 mm sec^{1} and (e) L: 1.1 KN,
S: 710 rpm, F: 0.34 mm sec^{1} 
The results of the ANOVA are given in Table 5 and 6.
From the tables, it was found that the calculated fratio (6.96) was lower than
that from the statistical table f (3, 3, .05) = 9.27, indicating the adequacy
for the models at a 95% confidence level. Hence, the developed models were considered
to be adequate to predict the hardness and the tensile strength at the weld
section for the square and the pentagon pin profiles.
Confirmation tests: Tests were conducted to verify the regression Eq.
36. The weld runs were made using the same values of
the rotational speed, the welding speed and the axial force for each model;
but in randomized combinations of the process parameters for square and pentagon
pin profiles. The results obtained are presented in Table 7.
Comparison of weld bead, hardness and tensile strength for square profile
and pentagonal tool profile: Figure 5 and 6
shows that the weld bead structures of the AA6063 alloy plates welded using
the square profile and the pentagon profile tool respectively for the same process
parameters.

Fig. 6(ae): 
Weld bead structures for pentagon tool profile, (a) L: 1.5
KN, S: 1000 rpm, F: 0.34 mm sec^{1}, (b) L: 1.5 KN, S: 710 rpm,
F: 0.24 mm sec^{1}, (c) L: 1.1 KN, S: 1000 rpm, F: 0.5 mm sec^{1},
(d) L: 1.5 KN, S: 1400 rpm, F: 0.24 mm sec^{1}, (e) L: 1.1 KN,
S: 710 rpm, F: 0.34 mm sec^{1} 
From the figures it is observed that 6063 alloy plates welded using square
profile pin has a good weld bead structure when compared to pentagon profile
tool. From the experimental results shown in Fig. 7a the plates
welded with the square profiled tool had a higher hardness value at the weld
bead than the pentagonal profiled tool. Similarly, the tensile strength pertaining
to the square profiled pin tool showed a constant improvement, while the plates
welded with the pentagonal profiled tool fluctuated slightly depending upon
the process parameters as shown in Fig. 7b.
CONCLUSIONS
Based on the experimental results, it is found that the friction stir welding
on the Aluminium alloy 6063 using two different friction stir welding tools
(i.e.) the square profiled tool and the pentagonal profiled tool, exhibits their
individual effects on the mechanical properties. The following conclusions are
arrived from the above study:
• 
The plates welded with the pentagonal profiled tool exhibits
mechanical properties at certain process parameters and the results are
satisfactory. Yet further study has to be done to on the pentagon profile
tool to evaluate the bead structure and the mechanical properties 
• 
The comparative evaluation of the results produced by the friction stir
welding of AA6063 by the square and the pentagonal profiled tools reveal
that the square profiled pin produces comparatively better results 