1. J. Mater. Sci. Technol., Vol.23 No.2, 2007 223
Effect of Pulsed Current TIG Welding Parameters on Pitting
Corrosion Behaviour of AA6061 Aluminium Alloy
T.Senthil Kumar1) , V.Balasubramanian2)† , M.Y.Sanavullah3) and S.Babu2)
1) Department of Automobile Technology, Bharathidasan University, Tiruchirappalli 620 024, India
2) Department of Manufacturing Engineering, Annamalai University, Annamalainagar 608 002, India
3) Principal, V.M.K.V.Engineering College, Salem 636 308, India
[Manuscript received September 29, 2005, in revised form June 23, 2006]
Medium strength aluminium alloy (Al-Mg-Si alloy) has gathered wide acceptance in the fabrication of light
weight structures requiring a high strength-to weight ratio, such as transportable bridge girders, military
vehicles, road tankers and railway transport systems. The preferred welding process for aluminium alloy is
frequently TIG (tungsten inert gas) welding due to its comparatively easier applicability and better economy.
In the case of single pass TIG welding of thinner section of this alloy, the pulsed current has been found
beneficial due to its advantages over the conventional continuous current process. The use of pulsed current
parameters has been found to improve the mechanical properties of the welds compared to those of continuous
current welds of this alloy due to grain refinement occurring in the fusion zone. A mathematical model has
been developed to predict pitting corrosion potential of pulsed current TIG welded AA6061 aluminium alloy.
Factorial experimental design has been used to optimize the experimental conditions. Analysis of variance
technique has been used to find out the significant pulsed current parameters. Regression analysis has been
used to develop the model. Using the developed model pitting corrosion potential values have been estimated
for different combinations of pulsed current parameters and the results are analyzed in detail.
KEY WORDS: Pulsed current; Tungsten inert gas welding; Medium strength aluminium alloy;
Pitting corrosion; Design of experiments; Analysis of variance
1. Introduction Pulsed current tungsten inert gas (PCTIG) weld-
ing, developed in 1950s, is a variant of TIG welding
Weld fusion zones typically exhibit coarse colum- which involves cycling of the welding current from a
nar grains because of the prevailing thermal condi- high level to a low one at a selected regular frequency.
tions during weld metal solidification. This often re- The high level of the peak current is generally selected
sults in inferior weld mechanical properties and poor to give adequate penetration and bead contour, while
resistance to hot cracking. It is thus highly desirable the low one of the background current is set at a level
to control solidification structure in welds and such a sufficient to maintain a stable arc. This permits arc
control is often very difficult because of higher tem- energy to be used efficiently to fuse a spot of controlled
peratures and higher thermal gradients in welds in dimensions in a short time producing the weld as a se-
relation to castings and the epitaxial nature of the ries of overlapping nuggets and limits the wastage of
growth process. Nevertheless, several methods for re- heat by conducting into the adjacent parent material
fining weld fusion zones have been tried with some in a normal constant current welding. In contrast to
success in the past: inoculation with heterogeneous constant current welding, the fact that the heat en-
nucleants[1] , microcooler additions, surface nucleation ergy required to melt the base material is supplied
induced by gas impingement and introduction of phys- only during peak current pulses for brief intervals of
ical disturbance through torch vibration[2] . time allows the heat to dissipate into the base mate-
rial leading to a narrower heat affected zone (HAZ).
The use of inoculants for refining the weld fusion
The technique has secured a niche for itself in specific
zones is, as a matter of fact, not as successful as in
applications such as in welding of root passes of tubes,
castings because of the extremely high temperature
and in welding thin sheets, where precise control over
involved in welding and also due to the undesirable ef-
penetration and heat input are required to avoid burn
fect of inoculating elements on weld mechanical prop-
through.
erties at the level required for grain refinement. Other
Extensive researches have been performed in this
techniques like surface nucleation and microcooler ad-
process and reported advantages include improved
ditions were also turned down because of the compli-
bead contour, greater tolerance to heat sink varia-
cated welding set-ups and procedures associated with
tions, lower heat input requirements, reduced residual
their use. In this process, two relatively new tech-
stresses and distortion[4] . Metallurgical advantages of
niques, termed as magnetic arc oscillation and cur-
pulsed current welding frequently reported in litera-
rent pulsing, have gained wide popularity because of
ture include refinement of fusion zone grain size and
their striking promise and the relative ease with which
substructure, reduced width of HAZ, control of seg-
these techniques can be applied to actual industrial
regation, etc.[5] . All these factors will help in improv-
situations with only minor modifications of the exist-
ing mechanical properties. Current pulsing has been
ing welding equipment[3] .
used by several investigators to obtain refined grains
in weld fusion zones and improvement in weld me-
† Prof., Ph.D.(IITM), to whom correspondence should be chanical properties[6,7] . However, reported research
addressed, E-mail: visvabalu@yahoo.com.

2. 224 J. Mater. Sci. Technol., Vol.23 No.2, 2007
Table 1 Important factors and their levels
Levels Peak current, P /A Base current, B/A Pulse frequency, F /Hz Pulse on time, T /%
Low (−1) 160 80 2 40
High (+1) 180 90 6 60
alloy to find out the feasible working limits of pulsed
Table 2 Experimental design matrix and test results current TIG welding parameters. AA4043 (Al-5%Si)
Expt No. P B F T ∗
PCP/mV aluminium alloy of 3 mm diameter was used as the
1 −1 −1 −1 −1 −500 filler metal and different combinations of pulsed cur-
2 +1 −1 −1 −1 −475 rent parameters were used to carry out the trial runs.
3 −1 +1 −1 −1 −590 The bead contour, bead appearance and weld quality
4 +1 +1 −1 −1 −540 were inspected to identify the working limits of the
5 −1 −1 +1 −1 −495 welding parameters. From the above analysis, the
6 +1 −1 +1 −1 −460 following observations were made.
7 −1 +1 +1 −1 −535 (1) When peak current was less than 160 A, incom-
8 +1 +1 +1 −1 −510 plete penetration and lack of fusion was observed. At
9 −1 −1 −1 +1 −550 the same time, when peak current was greater than
10 +1 −1 −1 +1 −505 180 A, undercut and spatter was observed on the weld
11 −1 +1 −1 +1 −600 bead surface.
12 +1 +1 −1 +1 −575
(2) When background current is lower than 80 A,
13 −1 −1 +1 +1 −525
the arc length was found to be very short and addi-
14 +1 −1 +1 +1 −490
15 −1 +1 +1 +1 −560 tion of filler metal became inconvenient. On the other
16 +1 +1 +1 +1 −530 hand, when the background current was greater than
Note: ∗ PCP—pitting corrosion potential
90 A, the arc became unstable and arc wandering was
observed due to its increased arc length.
(3) The bead appearance and contours appear to
work on the effect of pulsed current parameters on me- be similar to that of constant current weld beads when
chanical and metallurgical properties are very scant. pulse frequency was less than 2 Hz, while more arc
Moreover, no systematic study has been reported to glare and spatter was experienced if pulse frequency
analyze the influence of pulsed current parameters on was greater than 6 Hz.
mechanical and metallurgical properties. (4) When pulse on time was lower than 40%, the
Thus, in this investigation an attempt was made weld nugget formation was not so smooth due to in-
to develop a mathematical model to predict the effect complete melting of filler metal. On the contrary,
of pulsed current TIG welding parameters on pitting when the pulse on time was greater than 60%, the
corrosion behaviour of medium strength AA6061 alu- overmelting of filler metal and overheating of tung-
minium alloy using statistical tools such as design of sten electrode was noticed.
experiments, analysis of variance and regression anal-
ysis. 2.3 Developing the experimental design matrix
By considering all the above conditions, the fea-
2. Scheme of Investigation sible limits of the parameters have been chosen such
that the AA6061 aluminium alloy should be welded
In order to achieve the desired aim, the present in- without any weld defects. Due to narrow ranges of
vestigation was planned in the following sequence: (1) factors, we decided to use two level, full factorial de-
identifying the important pulsed current TIG welding sign matrix to optimize the experimental conditions.
parameters, which have influence on grain refinement Table 1 presents the ranges of factors considered and
in fusion zone and corrosion resistance; (2) finding the Table 2 shows the 16 sets of coded conditions used to
upper and lower limits of the identified parameters; form the design matrix, 24 (2 levels and 4 factors) fac-
(3) developing the experimental design matrix; (4) torial design. The 16 experimental conditions (rows)
conducting the experiments according to the design have been formed for main effects by using the for-
matrix; (5) recording the responses; (6) developing mula 2nc−1 for the low (−1) and high (+1) values;
mathematical models; (7) identifying the significant where ‘nc’ refers to the column number. For exam-
factors; (8) checking the adequacy of the developed ple, in Table 2, the first four rows are coded as −1 and
models. next four rows are coded as +1, alternatively, in the
third column (because nc=3 and therefore 23−1 =4).
2.1 Identifying the important parameters The method of designing such a matrix can be found
From the literature[5–8] and our previous work[9] , in literature[10,11] .
the predominant factors which have greater influence For the convenience of recording and processing
on fusion zone grain refinement of pulsed current TIG the experimental data, upper and lower levels of the
welding process have been identified. They include: factors have been coded as +1 and −1, respectively,
(1) peak current; (2) background current; (3) pulse and the coded values of any intermediate levels can
frequency and (4) pulse on time. be calculated using the following expression[12] .
Xi = X − [(Xmax + Xmin )/2]/[(Xmax − Xmin )/2] (1)
2.2 Finding the working limits of the parameters
A large number of trial runs were carried out us- where Xi is the required coded value of a factor for
ing 5 mm thick rolled plates of AA6061 aluminium any value X from Xmin to Xmax ; Xmin is the lower

3. J. Mater. Sci. Technol., Vol.23 No.2, 2007 225
Table 3(a) Chemical composition (wt pct) of base metal and all weld metal
Type of material Mg Mn Fe Si Cu Al
Base metal (AA6061) 0.689 0.331 0.230 0.531 0.305 Bal
All weld metal (AA4043) 0.050 0.220 0.050 5.0 0.120 Bal
Table 3(b) Mechanical properties of base metal and all weld metal
Type of material Yield strength Ultimate tensile strength Elongation Vickers hardness, 0.05 kg
/MPa /MPa /(%)
Base metal (AA6061) 270 310 10 240
All weld metal (AA4043) 140 210 7 260
level of the factor and Xmax is the upper level of the where b0 is the average of responses (pitting corrosion
factor. potential); b1 , b2 , b3 ,......b15 are the coefficients that
depend on respective main and interaction factors and
2.4 Conducting the experiments and recording the re-
can be calculated by following expression[10] .
sponses
The base metal used in this investigation is a bi = Σ(Xi · Yi )/n (4)
medium strength aluminium alloy of AA6061 grade.
The chemical composition of the base metal was ob- where ‘i’ varies from 1 to n, in which Xi is the cor-
responding coded value of a factor and Yi is the cor-
tained using a vacuum spectrometer (ARL-Model:
responding response output value (pitting corrosion
3460). Sparks were ignited at various locations of the
potential) obtained from the experiment and ‘n’ is
base metal sample and their spectrum was analyzed
the total number of combinations considered (in this
for estimation of the alloying elements. The chemi-
case n=16).
cal composition of the base metal in weight percent is
Analysis of variance (ANOVA) method was ap-
given in Table 3. The polarization test was carried out
plied to find out the significance of main factors and
in non-deaerated 3.5% NaCl solution of pH 7. Analar
interaction factors. The higher order interactions (3
grade chemicals and double distilled water were used
and 4 factor interactions) are practically insignificant
for preparation of the electrolyte. The specimens were
prepared according to the metallographic standard. and hence not considered[13] . Yate s algorithm was
Specimens of 20 mm×40 mm (width and length) were used to calculate the sum of squares. Table 4 repre-
prepared to ensure the exposure of 10 mm diameter sents the Yate s algorithm and in the column marked
circular area in the weld region to the electrolyte. The (1), the upper half was obtained by adding successive
rest of the area was covered with an acid resistant lac- pairs of treatments and the lower half by subtracting
quer. A potentiostat (Gill AC) was used for this study successive pairs. Columns (2), (3) and (4) were ob-
in conjunction with an ASTM standard cell and per- tained in the same manner from the entries in columns
sonal computer. The corrosion rate was calculated by (1), (2) and (3), respectively. Each sum of square was
polarizing the specimen anodically and cathodically obtained by squaring the corresponding effect total
and by extrapolating the Tafel regions of anodic and and dividing the result by r. 2nf , where ‘r’ is number
cathodic curves to the corrosion potential. The inter- of replicates (trials) and ‘nf ’ is the number of chosen
section of these two lines at the corrosion potential factors. Further details regarding ANOVA method
yields the corrosion current density, icorr . The corro- and Yate s algorithm can be found in literature[10,11] .
sion potential and corrosion current density were ob-
tained for each Tafel plot directly from the personal 3.1 Final mathematical model
computer attached to the polarization set-up. ANOVA test results are presented in Table 5.
From the ANOVA test results, it is evident that all
3. Mathematical Model the main factors (P, B, F, T ) and few interaction fac-
tors (BF and BT ) were considered to be significant.
In order to represent the PCP of the joint, the Hence the final model was developed including only
response function can be expressed as follows[10–12] . these significant factors and given below.
PCP = f (peak current (P ), base current (B ), (PCP) = {(−528) + 16.88(P ) − 27.5(B ) +
pulse frequency (F ), pulse on time (T ))
PCP = f (P , B , F , T ) (2)
14.38(F ) − 4.38(T ) + 6.88(BF ) − 3.13(BT )}mV
The model selected includes the effects of main (5)
factors and first order interaction of all factors. It
is a portion of power series polynomial expressed as 3.2 Checking adequacy of the developed model
follows. Coefficient of correlation ‘r’ is used to find how
close the predicted and experimental values lie and it
PCP = b0 + b1 (P ) + b2 (B) + b3 (F ) + b4 (T ) + b5 (P B)+ is calculated using the following expression.
b6 (P F ) + b7 (P T ) + b8 (BF ) + b9 (BT )+ r 2 = Explained variation/Total variation =
b10 (F T ) + b11 (P BF ) + b12 (P BT ) +
¯ ¯
Σ(PCPp − PCP)2 /Σ(PCPe − PCP)2 (6)
b13 (P F T ) + b14 (BF T ) + b15 (P BF T ) (3)

4. 226 J. Mater. Sci. Technol., Vol.23 No.2, 2007
Table 4 Yate s algorithm to calculate sum of squares for pitting corrosion potential (PCP)
Y [1] [2] [3] [4] SS
−500 −975 −2105 −4105 −8440 4E+06 1
−475 −1130 −2000 −4335 270 4556.3 P
−590 −955 −2230 135 −440 12100 B
[+] −540 −1045 −2105 135 −10 6.25 PB
−495 −1055 75 −245 230 3306.3 F
−460 −1175 60 −195 −20 25 PF
−535 −1015 70 15 110 756.25 BF
−510 −1090 65 −25 −20 25 P BF
−550 25 −155 105 −230 3306.3 T
−505 50 −90 125 0 0 PT
−600 35 −120 −15 50 156.25 BT
[−] −575 25 −75 −5 −40 100 P BT
−525 45 25 65 20 25 FT
−490 25 −10 45 10 6.25 PFT
−560 35 −20 −35 −20 25 BF T
−530 30 −5 15 50 156.25 P BF T
Table 5 ANOVA (analysis of variance) test results for pitting corrosion potential
Factors
Sum of squares (SS) Degrees of freedom (d.o.f.) Mean squares (SS/d.o.f.) Fratio (MS/error)
Main factors
P 4556.25 1 4556.25 72.9
B 12100 1 12100 193.6
F 3306.25 1 3306.25 52.9
T 3306.25 1 3306.25 52.9
Two factors
∗
PB 6.25 1 6.25 0.1
∗
PF 25 1 25 0.4
∗
PT 0 1 0 0
BF 756.25 1 756.25 12.1
∗
BT 156.25 1 156.25 2.5
∗
FT 25 1 25 0.4
Error 312.5 5 62.5 –
Total 24550 15 – –
Note: ∗ F(1,5,0.95) =6.41. Therefore, P B, P F , P T , BT & F T are not significant at 95% confidence level
where PCPp is predicted (using the above model) pit- stresses built up during freezing exceed the strength of
ting corrosion potential value for the given factors; the solidifying weld metal. The commonly used meth-
PCPe is the experimental value for the correspond- ods to reduce the tendency for solidification cracking
¯
ing factors; PCP is the average of experimental pit- include: altering weld metal composition, through the
ting corrosion potential values. The value of ‘r’ for addition of a filler wire, close process control, and con-
the above developed model is found to be 0.92, which trolling the grain structure within the fusion zone. It
indicates high correlation between experimental and is widely accepted that by changing the weld s grain
predicted values. structure, from coarse columar to fine equiaxed, bet-
ter cohesion strength can be obtained, and the re-
4. Discussion maining eutectic liquid present during the final stage
of solidification can be fed more easily and the pre-
The mathematical model developed in the above formed cracks may be healed[14,15] .
section has been written in C program and the devel- Another way of reducing the susceptibility to so-
oped C program has been used to estimate the pitting lidification cracking is through fusion zone grain re-
corrosion potential of the pulsed current TIG welded finement, which confers the further benefit that the
AA6061 aluminium alloy welds for different combi- weld metal mechanical properties are improved. Var-
nations of pulsed current parameters. Predicted val- ious grain refinement techniques have been discussed
ues were plotted and displayed in Fig.1. The plotted in the literature for aluminium alloy welds, e.g. elec-
graphs can be effectively used to understand the ef- tromagnetic stirring, current pulsing, torch vibration
fect of pulsed current parameter, such as peak current, and inoculation. Of these, pulsed current welding
base current, pulse frequency and pulse on time, on technique has gained wide popularity because of their
pitting corrosion resistance of TIG welded AA6061 striking promise and the relative ease with which
aluminium alloy joints. Figure 2 reveals the fusion these techniques can be applied to actual industrial
zone microstructure of the welded joints. situations with only minor modifications to the exist-
ing welding equipment[17] .
4.1 Effect of pulsed current parameters on fusion zone In general, the formation of equiaxed grain struc-
grain size ture in CCTIG (continous current tungsten inert gas)
Solidification cracking occurs when the thermal weld is known to be difficult because of the remelting

5. J. Mater. Sci. Technol., Vol.23 No.2, 2007 227
Fig.1 Effect of pulsed current parameters on pitting corrosion potential
of heterogeneous nuclei or growth centers ahead of the up may not be effective in welding because of the
solid-liquid interface. This is due to the high tem- small size of the fusion welds and the fine interden-
perature in the liquid, thus making survival nuclei drite spacing in the weld microstructure. Thus grain
difficult. The microstructural evolution in weld fu- refinement observed in the PCTIG welds is therefore
sion zone is also influenced in many ways by current believed to be due to other effects of pulsing on the
pulsing, principally, the cyclic variations of energy in- weld pool shape, fluid flow and temperature. The con-
put into the weld pool cause thermal fluctuations, one tinual change in the weld pool shape is particularly
consequence of which is the periodic interruption in important. As the direction of maximum thermal
the solidification process. As the pulse peak current gradient at the solid-liquid interface changes contin-
decays the solid-liquid interface advances towards the uously, newer grains successively become favourably
arc and increasingly becomes vulnerable to any distur- oriented. Thus, while each grain grows only a small
bances in the arc form. As current increases again in distance, more grains grow resulting in a fine-grained
the subsequent pulse, growth is arrested and remelt- structure[14] .
ing of the growing dendrites can also occur. Current The weld pool solidification during fusion welding
pulsing also results in periodic variations in the arc begins with the epitaxial growth of grains from par-
forces and hence an additional fluid flows that low- tially melted zone grains along the fusion boundary,
ers temperatures in front of the solidifying interface. at the interface between the base metal and fusion
Furthermore, the temperature fluctuations inherent zone. The partially melted grains provide excellent
in pulsed welding lead to a continual change in the sites for growth with the growth rate exceeding the
weld pool size and shape favoring the growth of new nucleation rate in this region. Epitaxial growth across
grains. It is also noted that effective heat input for the weld pool results in long and oriented columnar
unit volume of the weld pool would be considerably grains. The epitaxial grains are the final result of
less in pulse current welds, so the average weld pool continuing growth of the partially melted grains from
temperatures are expected to be low[16,17] . the fusion boundary. Epitaxial growth requires that a
It is important to note that while dendrite frag- minimal degree of undercooling prevail. In contrast,
mentation has frequently been cited as a possible the nucleation of new grains both at and near the fu-
mechanism, evidence for this has not been found. It sion boundary necessitates a free energy barrier to be
was suggested that the mechanism of dendrite break- overcome. Consequently, no undercooling is necessary

6. 228 J. Mater. Sci. Technol., Vol.23 No.2, 2007
Fig.2 Micrographs of fusion zone region: (a) Joint 1 (D=40 µm), (b) Joint 2 (D=30 µm), (c) Joint 3 (D=65 µm),
(d) Joint 4 (D=55 µm), (e) Joint 5 (D=35 µm), (f) Joint 6 (D=20 µm), (g) Joint 7 (D=50 µm), (h) Joint
8 (D=40 µm), (i) Joint 9 (D=55 µm), (j) Joint 10 (D=35 µm), (k) Joint 11 (D=75 µm), (l) Joint 12
(D=60 µm), (m) Joint 13 (D=50 µm), (n) Joint 14 (D=30 µm), (o) Joint 15 (D=60 µm ), (p) Joint 16
(D=45 µm)
for nucleation. To initiate nucleation in the weld de- metallics are the initiation sites for pitting in Al-Zn-
posit and concurrently promote epitaxial grain refine- Mg-Cu alloys. The pitting is due to local dissolution
ment, it is essential to either increase the driving force, of the matrix or to dissolution of the intermetallics be-
i.e. degree of undercooling, or reduce the free energy cause there is galvanic coupling between intermetallics
barrier by introducing trace amounts of zirconium or and matrix. The intermetallics containing Cu and Fe
titanium to the aluminium weld pool[18] . are cathodic with respect to matrix and promote dis-
solution of the matrix, while Mg-rich intermetallics
4.2 Effect of pulsed current parameters on pitting are anodic with respect to the matrix and dissolve
corrosion preferentially[21,22] . In general, the pitting corrosion
The microstructure of AA6061 exhibits inter- resistance of AA6061 aluminium welds was found to
metallics and strengthening particles. The inter- be lower than that of the base metal. This can be at-
metallics are formed during casting and ingot homoge- tributed to the presence of segregation products in as
nization due to interaction between alloying elements solidified welds. The poorest corrosion resistance ex-
and impurities present in the alloy. In AA6061 the hibited by continuous current welds can be attributed
Mg2 Si intermetallics undergo phase transformation to the presence of continuous network of grain bound-
and change their morphology during ingot homoge- ary precipitates mainly containing magnesium rich η
nization, but they are not affected by solution heat phase. Region adjacent to the grain boundary is ex-
treatment and aging of the alloy. The strengthen- pected to be depleted in magnesium due to the pres-
ing particles have composition Mg2 Si and size in the ence of magnesium rich η at grain boundaries. These
nanometer range. Their precipitation in the matrix areas containing lower amounts of magnesium are the
during aging provides strength to the alloy. In ad- preferred locations for corrosion.
dition, the strengthening particles precipitate at the A relatively more uniform distribution of pits was
grain boundaries strongly affecting the resistance of observed in pulsed current welds and this is due to
intergranular corrosion of the alloy[19,20] . the absence of a continuous grain boundary precip-
However, Al7 Cu2 Fe and (Al,Cu)6 (Fe,Cu) inter- itates and to lower microsegregation of silicon and

7. J. Mater. Sci. Technol., Vol.23 No.2, 2007 229
magnesium in these welds. This could be attributed REFERENCES
to convection in weld pool due to current pulsing. The
grain boundary corrosion reported here is also simi- [1 ] R.P.Simpson: Weld. J., 1977, 56(3), 67s.
lar to weld decay generally observed in unstabilized [2 ] J.G.Garland: Metal Constru., 1974, 6(4), 121.
stainless steels due to depletion of chromium from lo- [3 ] K.Prasad Rao: In Proc. National Conf. on Recent
cations near the grain boundary as a consequence of Advances in Materials Processing, Annamalai Nagar,
chromium carbide precipitation at the grain bound- India, 2001, 176.
ary. [4 ] P.Ravi Vishnu: Weld. World, 1995, 35(4), 214.
Similar observations have also been made by other [5 ] A.A.Gokhale, Tzavaras, H.D.Brody and G.M.Ecer: In
investigators[8,23] in Al-Li alloy welded using AA5356 Proc. Conf. on Grain Refinement in Casting and
filler metal. Further, they opined that the aging up Welds, St. Louis, MO, TMS-AIME, 1982, 223.
to peak strength results in increased precipitation of [6 ] G.Madhusudhan Reddy, A.A.Gokhale and K.Prasad
equilibrium η at the grain boundaries, thus providing Rao: J. Mater. Sci., 1997, 32, 4117.
numerous anode-cathode cells. Overaging coarsens [7 ] H.Yamamoto: Weld. Int., 1993, 7(6), 456.
the precipitates and also results in precipitate agglom- [8 ] G.Madhusudhan Reddy, A.A.Gokhale and K.Prasad
eration, leading to a reduction in the density of precip- Rao: J. Mater. Sci. Technol., 1998, 14, 61.
itates as well as minimizing chemical inhomogeneity [9 ] V.Ravisankar and V.Balasubramanian: In Proc. Int.
around the precipitates due to diffusion effects. This, Conf. on IMPLAST, New Delhi, India, 2003b, 224-
232.
therefore, results in a relatively decreased tendency
[10] G.E.P.Box, W.H.Hunter and J.S.Hunter: Statistics for
for corrosion in the overaged condition, as compared
Experiments, John Wiley & Sons, New York, 1978.
to that in the underaged and peak aged conditions.
[11] D.C.Montgomery: Design and Analysis of Experi-
ments, John Wiley & Sons, New York, 1991.
5. Conclusions [12] J.Ravindra and R.S.Parmar: Metal Constru., 1987,
19, 45.
(1) Generally, peak current and pulse frequency [13] I.Miller, J.E.Freund and Johnson: Probability and
have direct proportional relationship with the pitting Statistics for Engineers, New Delhi: Prentice of Hall
corrosion resistance of the welded joints, i.e. if the of India Pvt. Ltd., 1999.
peak current is increased, the pitting corrosion resis- [14] S.Kou and Y.Le: Weld. J., 1986, 65.
tance will be increased. The similar effect is observed, [15] A.F.Norman, K.Hyde, F.Costello, S.Thompson,
when frequency is increased. S.Birley and P.B.Pragnell: Mater. Sci. Eng., 2003,
(2) Base current and pulse on time have inverse A335, 188.
proportional relationship with the pitting corrosion [16] T.Shinoda, Y.Ueno and I.Matsumoto: Trans. Jpn.
resistance, i.e. if the base current is raised, the pitting Weld. Soc., 1990, 21, 18.
corrosion resistance will be decreased. The similar in- [17] G.Madhusudhan Reddy: Proceedings of ISTE Sum-
fluence is noticed when pulse on time is increased. mer School on Recent Developments in Materials Join-
(3) The developed mathematical model can be ef- ing, Annamalai University, India, 2001.
fectively used to predict the pitting corrosion poten- [18] D.C.Lin, T.S.Wang and T.S.Srivatsan: Mater. Sci.
tial of PC TIG welded AA6061 aluminium alloy joints. Eng., 2003, A335, 304.
[19] R.P.Wei, C.M.Liao and M.Gao: Metall. Mater.
Trans. A, 1998, 29, 1153.
[20] P.S.Pao, S.J.Gill and C.R.Feng: Scripta Mater., 2000,
Acknowledgements
43, 391.
The authors would like to thank Defence Research &
Development Organization (DRDO), New Delhi for the [21] J.K.Park and A.J.Ardell: Metall. Trans. A, 1983, 14,
financial support rendered to carryout this investigation. 1957.
The authors also would like to thank the Department of [22] J.K.Park and A.J.Ardell: Scripta Metall., 1988, 22,
Manufacturing Engineering, Annamalai University for ex- 1115.
tending the facilities of Metal Joining Laboratory and Ma- [23] D.Hu, Y.Zhang, Y.L.Liu and Z.Y.Zhu: Corrosion,
terial Testing Laboratory to carryout this investigation. 1993, 49, 491.