Trends in Welding Research, Proceedings of the 8th International Conference Copyright © 2009 ASM International®

Stan A. David, Tarasankar DebRoy, John N. DuPont, Toshihiko Koseki, Herschel B. Smartt, editors, p 25-34 All rights reserved.
DOI: 10.1361/cp2008twr025 www.asminternational.org

Recent Developments of FSW Technologies:

Evaluation of Root Defects, Composite Refractory Tools for Steel Joining and
One-Pass Welding of Thick Sections Using Self-Reacting Bobbin Tools

W M Thomas and C S Wiesner

TWI Ltd, Cambridge, United Kingdom

Abstract
The toughness of welded components is of great
concern. Imperfections, especially surface breaking, can
exist in the root region which can affect structural
integrity. This paper describes results of impact testing
friction stir welded joints containing root imperfections.
The paper also looks at recent advances made in friction
stir welding of steel using refractory metal tools, and
describes a method of securing two different types of
refractory materials to provide a composite FSW tool.
The use of composite tools enables different ‘coupling'
relationships to be made between the shoulder and the
workpiece and the probe and the workpiece.
Furthermore, this paper describes some of the feasibility
work that has been carried out to develop self-reacting
(bobbin) friction stir welding of 25 mm thick aluminium
alloy material and 8mm thick 12% chromium steel.
Figure 1 The basic features of the ‘through-hole’ impact
Introduction technique. With the hole positioned in the middle of the
weld region.
The systematic development of Friction stir welding
(FSW) has led to a number of variants of the technology. The feature that characterises the new ‘through-hole’
The following describes preliminary studies being carried impact test is that the specimen is weakened by a hole in
out on Bobbin stir welding and the evaluation of weld the neutral axis perpendicular to the impact direction.
root defects. (1-3) The use of a precision reamed hole instead of a re-
entrant ‘V’, ‘U’ or ‘Keyhole’ notch means that the
Evaluation of FSW root defects toughness characteristics of the sample surface layers
can be evaluated. In addition, the hole provides a lower
and more diffuse stress concentration than the V notch.
A method of evaluating various (FSW) weld root defects
Moreover, comparisons can be made of the influence of
has been developed by the use of ‘through-hole’ impact
surface or weld root imperfections on the toughness of a
testing using the drop-weight method (Figure 1).
weld or a component. The toughness of the weld root
region is measured, in terms of the absorbed impact
energy needed to cause fracture of the specimen,
similarly to absorbed energy measured in Charpy and
conventional Drop-weight tests. (4-7)

The through-hole impact testing method is in some ways,

more representative of a real life performance as it
allows the fracture to propagate along the easiest route.
(Through-hole impact testing provides a more holistic
approach to an overall joint integrity).

25
 Retreating side Advancing side
The provision of a through-hole in the test specimen
sufficiently weakens the specimen, such that failure in
the sample occurs through, or close to, the hole.
Weakening of the specimen in this way is necessary to TMAZ
ensure that the specimen collapses and/or breaks during
testing and does not simply bend, and absorb very high
energies. The size and position of the hole can be
selected to ensure that total absorbed energy falls within
the range of capability of the test machine. The hole is
normally positioned in the middle of the specimen, but
can be varied above or below the centre line and the Parent HAZ Backing Plate
diameter can be changed to suit the material being Steel
tested. For example a 3mm diameter precision reamed
hole is used for aluminium and a 3.5 diameter precision a)
hole is used for steel. The main advantage of the
through-hole test is to provide freedom for fracture path
deviation. (7)

Fracture paths in root welds or clad layers that fail away

from the hole indicate the orientation of the weakest
route of fracture within the weldment. This ‘weakest’
fracture path provides some very useful information
about voids, incomplete penetration and the influence of
gross remnant oxide features that can be found in FSW
welds and about low toughness microstructures present.

Note: Another impact test that deserves mention is a

variant of the so called ‘Schnadt impact test’ whereby the b)
addition of a hardened steel pin is placed within a
comparatively large diameter hole (Typically the hole
diameter can be up to 50% of the test specimen
thickness (4)). In this method, the hole is, positioned off
centre, close to the striker side of the specimen and
opposite a ‘V’ notch on the tension side. The hardened
steel pin is believed to suppress plastic deformation
during impact. When used without a ‘V’ notch the
‘Schnadt impact test’ may be adapted to evaluate weld
root imperfections and when used in this way will provide
freedom for fracture path deviation, in a similar manner
in which the ‘through-hole impact test does. However,
the ‘Schnadt test’ has never received wide recognition
and has for practical purposes been forgotten mainly c)
because the Schnadt-specimens are more expensive to
produce.

Welds tested by ‘through-hole’ impact

Friction stir welds were produced in 12mm thick, 12%
chromium alloy steel using tools with a tapered probe
and three flats. Similar welding parameters of rotational
speed 584 rev, traverse rate 2.5mm/sec and axial force
of 32kN were used. A full penetration butt weld, an
incomplete penetration weld-on-plate and a butt weld
with an incomplete penetration defect were chosen for
evaluation using the ‘through-hole’ impact test.
d)

26
 fresh, untempered martensite. This was particularly in
the HAZs and to a lesser extent the central TMAZ of the
weld region. (9, 10 &11)

e)

Figure 3 Hardness survey mid-thickness in 12mm

thick 12% Cr steel weld

The results of the impact tests are shown in Table 1.

Table 1 ‘Through-hole impact’ test results - 12%

chromium alloy steel
f)
Specimen Tensio Absorbed Lateral Notes
No n Face Energy Expansi
Figure 2 Transverse macro & micro sections. (A Joules on mm
millimetre scale is shown for Figures a, c, & e.) Parent NA 129 1.60 -
material
Parent NA 117 1.72 -
a) Full penetration butt weld, showing attached material
stainless steel backing bar. Parent NA 119 1.44 -
b) Detail of the root region of the full penetration weld material
(Figure 2a), free from gross imperfections. Full Weld 100 1.54 -
penetration root
c) Incomplete penetration weld-on-plate. butt weld (1)
d) Detail of the imperfections at the bottom of the weld- Full Weld 96 1.49 -
on-plate region (Figure 2c). penetration root
e) Incomplete penetration butt weld, showing voids and butt weld (1)
Full Weld 98 1.59 -
surface breaking defect. penetration root
f) Detail of the surface breaking imperfection at the butt weld (1)
root of the incomplete penetration butt weld (Figure Weld-on- Weld 116 1.70 -
2e). Showing the original joint line with no plastic plate (2) root
Weld-on- Weld 111 1.64 -
deformation (at the weld root) which changes in plate (2) root
orientation by severe plastic deformation, as it gets Weld-on- Weld 112 1.57 -
closer to the central weld region. plate (2) root
Incomplete Weld 37 0.54 Fracture
Macrosections of the chosen friction stir welds are shown penetration root away
butt weld (3) from hole
in Figures 2a, c and e respectively. Microsections of the Incomplete Weld 36 0.53 Fracture
respective details of the root features of the welds are penetration root away
shown in Figures 2b, d and f. butt weld (3) from hole
Incomplete Weld 31 0.46 Fracture
penetration root away
A hardness traverse is shown in Figure 3 and indicates butt weld (3) from hole
increased weld region hardness owing to promotion of
27
 Weld-on-plate

Note: (1) No intentional imperfection (see Figure 2a&d) 25.0

Note: (2) Incomplete penetration (embedded
imperfection) weld-on-plate (see Figure 2b&e) 20.0

Note: (3) Incomplete penetration, butt weld with surface

breaking imperfection (see Figure 2c&f) 15.0

Load, kN
10.0
Figure 4 shows the results of dynamic testing as force
versus time graphs. Here, the force or “load” is recorded 5.0

using an instrumented striker. This provides information

about the deformation and fracture processes 0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
experienced by the specimen. It can be seen that the Time, mS

parent material test showed an initial broad load peak

over ~1.5ms and then a gradual drop off of load up to c)
~8ms. In contrast, the shape of the instrumentation Partial penetration butt weld

records taken from the incomplete penetration butt weld 25.0

had an initial peak that dropped to a very low load value

after ~0.6 ms. The differences in time to fracture are an 20.0

indication of the degree of ductility achieved and fracture 15.0

Load, kN
appearance. Although the peak force achieved in the
incomplete penetration butt weld was higher than the 10.0

parent material, because of its higher hardness (see

Figure 3) and therefore higher strength than the parent 5.0

material, the time to fracture is considerably shorter. 0.0

Furthermore, as will be shown, the tension side of the 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
Time, mS
specimens failed by brittle fracture, which is a rapid,
compared with slower, ductile fracture in the parent
d)
material.
Figure 4 Typical instrumentation records for:
a) Parent material
Parent material
b) Full penetration butt welds
25.0
c) Weld on plate
d) Incomplete penetration butt weld
20.0

15.0
Although substantially broken, each half of the tested
Load, kN

specimen remained attached at the impact hinge region,

10.0 see Figure 5.
5.0

0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
Time, mS

a)
Full penetration butt weld

25.0

20.0

15.0
Load, kN

10.0

5.0

0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
Time, mS

b)

28

a) b)
c) d)
Figures 5a, b, c and d show typical broken impact
tested samples from:
a) Parent material.
b) Full penetration butt weld (free from gross
imperfections).
c) Weld-on-plate (weld with internal voids near the
root).
d) Incomplete penetration butt weld (weld with surface
breaking defects at the root of the weld)
Tests on the parent material gave the largest impact
values (121 Joules) with failure as shown in Figure 5a. A
similar, across the hole, fracture path occurred for all the
comparatively high impact energy test specimens,
namely the weld-on-plate, with internal voids (113
Joules, Figure 5c) and the butt weld free from gross
imperfection (98 Joules, Figure 5b). However, tests on
the incomplete penetration weld, with a surface breaking
imperfection at the weld root followed the orientation of
the original joint line and the severely plastically
deformed joint line (which was caused by welding) and
deviated around the hole into the coarse TMAZ (Figure
5d). The test specimens with surface breaking
imperfections gave comparatively low impact energy
(34.5 Joules). Severe plastic deformation, developed
during the welding operation caused a change in the
orientation of the original joint line, in the incomplete
penetration weld as shown in Figure 2f. This change in
orientation guided the fracture path into coarse
microstructure and around the ‘through-hole’ feature;
clearly, this type of surface breaking imperfection is a
worse stress concentrator than the through-hole. Two
main types of fracture path were observed, one across
the hole and one that deviated around the hole.
Direct comparison can only be made between welds that
fracture in the same manner. Welds that fracture across
the hole can only be compared directly with welds that
reveal a similar fracture path. Welds that deviate around
the hole can only be compared directly with welds that
also deviate around the hole.
29
In a similar way, but to a lesser extent, comparison
between weld and parent material cannot be made
directly. However, the results may well be judged in the
same fashion as parent material and weld metal Charpy
impact test results are often interpreted.
Consequently, while a deviated fracture path cannot be
compared directly with fracture across the hole it
indicates a more serious situation. Given that the hole
does intensify stress then it would be true to say that
when judged against the impact tests taken from the
parent material, the impact energy of the incomplete
penetration weld, with a surface breaking imperfection at
the weld root (Figure 7d) was reduced by at least 72%.
When judged against parent material the impact energy
needed to break the full penetration weld was reduced
by 19%, while the weld-on-plate energy despite the
presence of internal voids was reduced by 6.6%.
The effect of weld root surface breaking imperfections
can best be understood when the respective impact
energy is measured against that from the weld-on-plate
test specimen (with internal voids). Notwithstanding, the
different fracture paths present, differences in
microstructure are for the most part removed, which
means that the cause and effect of the geometry and
position of the imperfection (in this case incomplete
penetration) can be recognised. It is clear, therefore,
that surface breaking imperfections at the weld root are
serious. In this example the impact energy was reduced
by at least 69% owing to the presence of the surface
breaking imperfection at the weld root.
Examination of the fracture faces of the parent material
tests revealed that no coarse crystalline structure was
present, with the fracture surface being fibrous in
appearance. Conversely, extremely coarse crystalline
structure was present throughout the fracture face with
the incomplete penetration welds. Coarse crystalline
structure was also present on the ligament furthest away
from the root of all the butt welded specimens. These
differences in fracture appearance are reflected in the
absorbed energies and lateral expansions obtained from
these specimens, see Table 1.
FSW steel welding trials using composite refractory
tools
When welding steel, the tool shoulder reaches a high
temperature within a few seconds of making contact with
the plate surface and the colour indicates that it operates
at over 1200 degrees C. Many researchers have
reported work on the welding of ferrous materials, carbon
steels, stainless steels and nickel-based alloys using
both refractory metal and poly-crystalline cubic boron
nitride tools. Tool wear and catastrophic tool failure
remain issues and the friction stir welding over 20 mm in
thickness and over 20 metres in length remains a
technical challenge.
TWI has conducted work using composite tools
comprising of two different refractory metal alloys. The
material for the shoulder was selected to produce a
smooth surface finish and the material for the probe was
selected to have high strength and to achieve good
coupling to the steel at the welding temperature. The tool
was made using a ‘Morse taper’ principle which is an
ideal arrangement for securing the two materials, see
Figure 6.
Figure 6 Securing two different refractory materials by a
tapered friction coupling design
The concept of composite tools with the shoulder and
probe optimized for their particular function is being
pursued further with the aim of extending the length of
the weld that can be achieved and the range of materials
that can be welded.
The Morse type taper allows the combining of different
materials and provides a unique hot friction drive
coupling. The conical probe part compresses the tapered
bore of the shoulder to produce hoop stresses within the
shoulder. These hoop stresses can be accommodated
easily by increasing the thickness and/or diameter of the
shoulder.
The friction contact across the entire conical surface
area of the interface between inner and outer matching
taper provides a large amount of torque transmission, so
splines, bolts or keys and keyways and similar stress
raisers are not required.
30
The surface of the steel welds showed a uniform surface
ripple (caused by the final sweep of the trailing edge of
the rotating tool). The FSW weld was essentially smooth
and flush with the surface as shown in Figure 7.
Figure 7 Typical surface appearance of friction stir
welded 12% Cr alloy steel plate using a composite tool
The almost semicircular ripple in the weld tracks for steel
was essentially the same as those for aluminium FSW
welds.
Initial results in welding 12 mm thick 12% Cr steel have
proved encouraging with over 20m of weld being made
with a single tool in one metre stages (i.e. with 20 weld
starts and stops). The typical surface appearance of a
weld is shown in Figure 7 and a representative weld
cross section is shown in Figure 2a. Initial mechanical
tests showed tensile failure outside the weld region at
parent material strength levels and acceptable face, root
and side bends achieving 180degrees.
Bobbin stir welding
Self-reacting FSW has been shown to be effective for
joining hollow extrusions and lap joints. Essentially there
are two types of self-reacting techniques one known as
the ‘bobbin tool’ (1, 12 & 13) and one known as the
‘adaptive technique’ (AdAPT) (14, 15 & 16). The bobbin
technique provides a fixed gap between two shoulders,
while the adaptive technique enables adjustment of the
gap between the shoulders during the welding operation.
The self-reacting principle of the bobbin technique
means that the normal down force required by
conventional FSW is reduced. The reactive forces within
the weld are contained between the bobbin shoulders
(Figure 8).

Figure 8 Bobbin tool showing self-contained reactive
forces
Trials in 25mm thick 6082-T6 aluminium using the above
arrangement produced good quality welds. Metallurgical
sections showing the width of the larger diameter (drive
side) shoulder and the smaller opposed shoulder are
shown in Figure 9. Unlike single sided stir welds, the
weld profile revealed is narrower in the mid-thickness
than at the shoulder regions. Several flow features within
the thermo-mechanically affected zone (TMAZ) are
shown in Figure 9.
Advancing side Retreating side
Figure 9 Macrosection of Bobbin weld in 25 mm thick
6082-T6 aluminium
The hardness distribution across the transverse direction
in the 25mm thick 6082-T6 aluminium weld is shown in
Figure 10. The minimum hardness is located in the HAZ
near the interface between the TMAZ and the HAZ.
31
Figure 10 Hardness survey mid-thickness in 25mm thick
6082-T6 aluminium weld
Three point bend testing confirmed that the weld
provided good mechanical integrity (Fig 11).
Figure 11 Three point bend test on 25 mm thick
6082 T6 aluminium alloy bobbin weld
Bobbin type tools, are similar to other standard FSW
tools that are driven from one side, in that the tool
behaves as a rotating cantilever. The use of a tapered
probe provides for a more uniformly stressed tool which
displaces substantially less material during welding than
a cylindrical pin type probe. The use of a tapered probe
for the bobbin tool enables a proportional reduction in the
diameter of the lower shoulder of the bobbin tool. A
reduction in the lower shoulder diameter results in lower
frictional contact and resistance, therefore less torque
and bending moment on the tool. The additional frictional
contact provided by the lower shoulder and the absence
of a backing anvil, which acts as a heat sink, means that
the operating temperature will be higher than that of
similar conventional welds. Tool design and process
conditions will need to be adjusted to allow for the
welding travel speed to be increased benefiting from this
additional heat generation.
Bobbin welds essentially eliminate partial penetration,
lack of penetration or root defects. Preliminary trials have
shown that lap welds produced by the bobbin technique
have less problems with the adverse orientation of the
notch at the edge of the weld.
Certain bobbin welds can reveal a mid-thickness ‘blip’
that appears on the advancing side, see figure (13). Non-
optimised welds can also be characterised by
imperfections that appear in the mid-thickness near the
‘blip region’ of the weld on the advancing side see Figure
12. These imperfections are usually caused by
insufficient static to dynamic volume ratio of the probe to
provide an adequate flow path.
Figure 12 Non-optimised bobbin welds showing a
mid-thickness ‘blip’ and imperfections on the
advancing side. 25mm thick 6082-T6 aluminium alloy.
Double driven bobbin techniques (DDB)
For certain applications bobbin tools that are driven from
both ends are envisaged (Figure 13a and b).
a)
32
b)
Figure 13 Bobbin tool
a) Driven from both ends
b) Driven from both ends and reactive force
applied from both ends
With both sides of a fixed-gap bobbin tool driven, the
probe part of the tool no longer behaves as a rotating
cantilever. A bobbin tool that is driven from both ends
and designed for uniform stress, means that the aspect
ratio of the probe can be altered (decrease in cross-
section area and/or increase in length) to provide an
improved flow path. However, while the torque and
bending forces can be shared between both ends, the
cross-section of the probe must be able to accommodate
the reactive forces that tend to push the shoulders apart.
Double-driven and double-adaptive bobbin
techniques (DDDAB)
The concept of a double-driven bobbin also includes the
use of a double-adaptive technique whereby both
shoulders can be adjusted and a load applied from both
ends see Figure 13. The latter arrangement will reduce
the reactive forces transmitted through the probe and
enable FSW to tackle thicker plate material than
currently possible. The (DDDAB) concept is expected to
increase the welding speed significantly above that
which is possible using conventional bobbin techniques
and may even provide welding speeds faster than
conventional FSW for thick plate welding.
The use of bobbin type techniques requires run-on and
run-off regions for the tool. Bobbin techniques are best
suited to flat two-dimensional applications but could be
developed for more complex shapes.
Bobbin stir welding of steel
There are a number of features that make Bobbin stir
welding of steel attractive. Two shoulders provide
sufficient heat generation from both sides of the
workpiece without any heat lost through the anvil support
plate. The containment of reactive forces within the tool
itself means that compressive deformation (Squashing)
of the probe does not occur. The probe part of a
conventional FSW tool is subjected to multi-axial forces
comprising: torsion, bending and compression. The
probe part of a single-sided bobbin tool is also subjected
to multi-axial forces comprising comparatively higher
levels of torsion and bending with only tensile forces
being applied through the probe.
Figure 14 shows a good surface appearance without
shielding gas. Use of an air blast aimed at the lower
shoulder provides a modest thermal management effect.
This means that compressive deformation of the probe is
eliminated.
Figure 14 The weld surface finish of a bobbin weld using
a refractory metal alloy, probe and shoulder combination
tool to weld 8mm thick 12% chromium steel (Rotational
Speed 584 rev, traverse rate 1.25 mm/sec. no axial
force)
Figure 15 Macrosection of a bobbin weld
33
Bend test carried out on both sides of a bobbin weld are
shown in Figure 16.
Figure 16 Bend tests carried out on both sides of a
Friction stir bobbin weld in 8 mm thick 12% chromium
steel. This weld was made with a refractory metal alloy
combination tool.
Concluding remarks
The ‘through-hole’ impact test is an effective comparative
tool for evaluating microstructures and root defects in
welded joints. The specimen is relatively easy to prepare
and only requires small amounts of material. The
‘through-hole’ impact test can be used to provide
guidance in forecasting the likely ‘fitness for service’ of
welds for particular applications.
The fracture path and surface of the ‘through-hole’
impact specimen reveals useful information about the
fracture process from two distinct regions of the
specimen; the tensile ligament that fails first in the test
and the second ligament that gives an indication of bulk
weld metal toughness. Moreover, the presence of a root
defect and/or zone of weakness is likely to be exposed
by the use of the ‘through-hole’ impact test.
Further work, will be undertaken at TWI to evaluate and
develop the ‘through-hole’ impact technique over a range
of materials especially weld root and clad layer integrity.
The study will include a change and reduction of the size
and position of the hole and for brittle weldments the
possibly of eliminating the need of any stress
concentrator.
The use of the Morse taper friction coupling principle for
composite tools made from refractory materials has been
demonstrated at high loads and at operating temperature
of over 1300 degrees C. Moreover, the use of composite
tools has enabled the frictional coupling between the
weld region around the probe and the weld surface
region adjacent to the shoulder to modify appropriately.
Further developments in this area are expected to
extend the use of FSW in joining steels, stainless steels
and nickel based alloys.
The fixed gap bobbin technique shows promise for
welding 25 mm thick aluminium plate and 12 mm thick
steel plate with out the need for an axial down force.
The reactive forces contained within the tool helps to
prevent axial compressive deformation of the probe. The
bobbin technique provides full penetration welds free
from lack of penetration and associated root defects. The
developments in bobbin tool welding of enclosed seams
such as extrusions will with certain applications eliminate
the need for internal backing bars to support the weld
region.
Acknowledgements are made for the support and
contributions provided by Dr P Woollin, Dr H G Pisarski,
Mr D G Staines, Dr I M Norris, Dr R L Jones, Dr D J
Marks, Mr J C Godden, Mr M F Gittos and Prof. W
Lucas.
References
1. Thomas, M W, Nicholas, E D, Needham, J C,
Murch, M G, Temple-Smith, P and Dawes, C
J. ‘Improvements relating to friction welding’.
European Patent Specifications 0615 480 B1.
2. Thomas, M W, Norris, I M, Staines, G D, and
Lucas, W: ‘The evaluation of root defects in
FSW by ‘through-hole’ impact testing –
preliminary studies’ Journal of Materials 2005.
3. Thomas, M W, Pisarski, H G, Norris, I M ,
Marks, D J, and Godden, J C: An investigation
of through-hole impact testing of weld root
imperfections in friction stir welding 12 per cent
chromium alloy steel. Proc. IMechE Vol 222,
Part B: J Engineering Manufacture, August
2008.
4. Boyd, G M: ‘The assessment of notch ductility
by a variety of notch tests’, Institute of Welding
Symposium, on Notch bar testing and its
relation to welded construction. London, 5
December 1951, P11-18.
5. Matthews, W T: ‘The Role of Impact Testing in
Characterising the Toughness of Materials’,
Impact Testing of metals, ASTM STP 466,
American Society for Testing and Materials,
1970, p.3-20.
34
6. Hughes, R K, and Ritter, J C: ‘Blunt Notch
Toughness Testing’, Mechanical Testing of
Materials, Conference Proceedings, Melbourne,
Victoria, Australia, 1994.
7. Hadley, I: ‘Charpy testing laser welds – the
significance of fracture path deviation’ TWI,
Bulletin, Vol. 42 No. 5, Sept/Oct 2001.
8. O’Brien R C, ‘Impact and fatigue
characterisation of selected Ferrous P/M
materials, presented at the Annual Metallurgy
Conference, Dallas, Texas, May 17-20, 1987.
9. Gooch T G and Ginn B J, ‘Heat-affected zone
toughness of SMA welded 12% CR martensitic-
ferritic steels’, Welding Journal, Nov 1990, 69
(11) 431s – 440s.
10. Gooch T G, Woollin P and Haynes A G,
‘Welding metallurgy of low carbon 13%
chromium martensitic steels’, Paper presented at
the Conference on Supermartensitic Stainless
Steels ’99 in Brussels Belgium 27-28 May 1999.
11. Onoro J, ‘Martensite microstructure of 9-12%
steels weld metals’, Journal of Materials
Processing Technology, Volume 180, Issues 1-
3, 1 December 2006, pp 137-142.
12. Colligan, K J, and Pickens J R ‘Friction Stir
Welding of Aluminium Using a Tapered Shoulder
Tool,’. ‘Friction Stir Welding and Processing III,
eds K V Jata, Mohoney, R S Mishra, and T J
Lienert, TMS Annual Meeting, San Francisco,
2005, pp 161-170.
13. Graham L D ‘Low Cost Portable Fixed-Gap
Bobbin Tools FSW Machine’, poster
presentation at the 86th Annual AWS
Convention/2005 Welding Show.
14. Thomas M W and Sylva G: ‘Developments of
Friction Stir Welding’, ASM Materials Solutions
2003, Conference & Exposition, 13-15 October
2003 Pittsburgh, Pennsylvania, USA.
15. Marie F, Allehaux D, and Esmiller B,
‘Development of the Bobbin Tool Technique on
various aluminium alloys’ TWI’s Fifth
International Symposium on Friction Stir
Welding, Metz, France, 14-16 September 2004.
16. Sylva G, and Edwards R, ‘A Feasibility study
for self Reacting Pine Tool Welding of Thin
Section Aluminum’, TWI’s Fifth International
Symposium on Friction Stir Welding , Metz,
France, 14-16 September 2004.