Evaluation of welded titanium joints used with cantilevered

implant-supported prostheses
Christopher N. Hart, BDSc, Grad Dip Clin Dent,a and Peter R. Wilson, MDS, MS, PhDb
School of Dental Science, University of Melbourne, Melbourne, Victoria, Australia
Statement of problem. Early failure of laser-welded titanium implant frameworks in clinical practice has
prompted an investigation of the strength and durability of welded cantilevered titanium sections.
Purpose. The purpose of this study was to determine the effect that the use of filler wire in laser welding of
titanium cantilever frameworks had on the flexural strength and fatigue resistance of the welded joints.
Material and methods. Sixty titanium implant-supported frameworks with 12-mm cantilevers were fabri-
cated in 4 groups (n=15), using 3 different laser welding protocols with 0, 1, and 2 weld passes with filler
wire, and 1 conventional tungsten inert gas welding method. The volume of filler wire used (mean volumes
0, 1.7, 3.4, and 8.3 mm3) was determined by measurement of the length of wire before and after welding
each joint. Ten frameworks from each group were tested for ultimate flexural strength by loading the cantilevers
10 mm from the abutment. The remaining 5 frameworks from each group were similarly tested under a simu-
lated masticatory load of 200 N until failure, or to 1 million cycles. A 2-way analysis of variance was used to
examine the flexural strengths, and log-rank statistics were applied to cyclic test data (a=.05).
Results. There were significant differences between the 4 groups for ultimate flexural strength (P,.001) and
resistance to cyclic loading (P=.002). The volume of filler wire added was a significant predictor of ultimate flex-
ural strength (P=.03), and was a borderline determinant of the number of cycles to failure at 200 N (P=.05).
Each laser weld pass with filler wire roughly doubled the ultimate flexural strength and fatigue resistance of
the joint relative to the previous weld. Tungsten inert gas welding with efficient argon shielding deposited
the most filler wire and produced the strongest and most fatigue-resistant joints.
Conclusion. The ultimate flexural strength and fatigue resistance of cantilevered joints in laser-welded tita-
nium prostheses are improved by the use of filler wire. Tungsten inert gas welding with efficient argon shielding
can be used in situations when a high-strength joint is required. (J Prosthet Dent 2006;96:25-32.)

CLINICAL IMPLICATION
Cantilevered joints in prostheses are required to withstand occlusal loading. Increasing the
strength of welded joints in cantilever situations should ensure long clinical service of the
prosthesis.

T he recent use of titanium in implant-supported su-

perstructures for the treatment of edentulous patients
has been driven primarily by the desire to produce light,
strong frameworks rapidly, without the complications of
casting distortion and porosities.1 Fabrication of tita-
nium implant-supported complete-arch superstructures
incorporating cantilevers by welding is a routine proce-
dure. In particular, the use of the neodymium: yttrium-
aluminium-garnet (Nd:YAG) laser to weld titanium
permits controlled heat application, minimizing distor-
tion and allowing welding in the presence of veneering
materials.2-8 However, the failure of welded titanium
This article is based on a thesis submitted to the University of Mel-
bourne, in partial fulfillment of the requirements for the Degree
of Master of Dental Science.
a
Postgraduate student.
b
Associate Professor, Head of Prosthodontic Unit.
cantilever joints after short clinical exposure has promp-
ted an investigation into their mode of fabrication.
A study of welded joints of commercially pure tita-
nium (CPTi) formed using various welding techniques
showed that the equiaxed microstructure of the parent
metal was retained in the fusion zone with laser welding,
whereas electric-arc welding (also referred to as tungsten
inert gas (TIG) welding) created a heat-affected zone,
which resulted in joints that had predominantly acicular
microstructures.9 The study concluded that titanium
joints with acicular microstructures were highly resistant
to tensile stresses, but their fatigue strengths were low.
Authors of a similar study also reported that the hard-
ness and strength of welded titanium are affected by
the presence of oxygen, nitrogen, and hydrogen during
the welding and cooling processes.10 Both studies re-
ported that laser and TIG welding resulted in increased
hardness in the heat-affected zone, whereas electron
beam welding, which was performed under vacuum,
had no effect on the hardness in the heat-affected zone.

JULY 2006 THE JOURNAL OF PROSTHETIC DENTISTRY 25

THE JOURNAL OF PROSTHETIC DENTISTRY
Fig. 1. Laser welded with filler wire cantilevered frame on
brass welding apparatus.
Early protocols for welded titanium prostheses were
described by Jemt and Linden,11 and involved the use
of specially fabricated components. Reports of clinical
longevity are documented for titanium frameworks fab-
ricated following these guidelines,11-15 but not all laser-
welded frameworks have followed these protocols.16 As
customization of components is almost universal in fixed
prosthodontics, it is difficult to standardize the implant
components used in such situations, making adherence
to specific protocols problematic.
Mean maximum occlusal forces for implant-supported
prostheses have been estimated to be 144 N, and these
are almost 3 times greater than the forces generated in
masticating cycles with a mean of 50 N.17 For this study,
a load of 200 N was used for the cyclical fatigue testing
of the welded frameworks, the same as that used in a
study on the changes in component fit of implant-
supported prostheses subjected to simulated functional
loading.18 On the basis that an individual eats 4 meals
each day, it has been estimated that approximately
1800 masticating strokes are performed daily.19 Thus,
from these studies the life of a welded framework may
also be estimated.
The objective of this study was to examine, under
simulated functional loading, the ultimate flexural
strength and fatigue resistance of clinically sized tita-
nium cantilever frameworks fabricated by 3 laser-weld-
ing methods using various amounts of filler wire in
forming the joints, and to compare these with frame-
works fabricated using a standard TIG welding
procedure.
MATERIAL AND METHODS
Sixty titanium cantilever frameworks were fabricated
in groups of 15 using 4 different laser and TIG welding
procedures incorporating various amounts of filler wire.
The frameworks were then mounted on implant replicas
in a custom apparatus, and the cantilevers were subjected
26
HART AND WILSON
Fig. 2. End and side views of steel testing apparatus showing
stainless steel implants secured with 5-mm metric screws.
to ultimate flexural strength and fatigue resistance test-
ing procedures. A load of 200 N was used for the
cyclical fatigue testing of the welded frameworks.
A brass welding apparatus (50 3 20 3 20 mm) was
fabricated to mount 2 implant replicas with a space of
15.0 mm between the centers of the implants, and
each implant was secured from 1 side with a 5-mm met-
ric screw (Fig. 1). A matching apparatus was machined
from steel alloy in which to mount the fabricated frame-
works for testing. This apparatus had 5-mm metric
screws from both sides to secure each implant and to
maximize the holding force during testing (Fig. 2).
All titanium components and filler wires were made
from grade 2 commercially pure titanium (CPTi) (Tita-
nium International, Coolaroo, Australia). One hundred
and twenty nonhexed abutment cylinders (5.0 3 10.0
mm) with a 2.6-mm bore were machined from a
6-mm CPTi drawn rod, and 120 implant replicas (4.1 3
15 mm) were machined from 5-mm American Iron
and Steel Institute (AISI) 306 stainless steel. A 5-mm
flat surface was machined into the side of each implant,
2 mm from the bottom, into which a side-screw was fit-
ted for the purpose of firmly securing the implants in the
testing apparatus. One hundred twenty rectangular
blanks (5.0 3 2.0 3 12.0 mm) were cut from a
2.0-mm-thick CPTi plate, using a high-pressure abra-
sive water jet cutting device (Neptune; Farley LaserLab,
Rockford, Ill) to reduce the likelihood of creating heat-
affected areas in the material. Concave radii (2.5 mm)
were milled into both ends of 60 of the rectangular
blanks to form connectors that fit intimately between
the 2 abutment cylinders when assembled in the welding
apparatus. The remaining 60 blanks had the same radius,
machined at one end only, for joining to an abutment as
the cantilever section.
Stainless steel implant replicas were inserted and se-
cured in the brass welding apparatus. Titanium abut-
ments were mounted on the implant replicas and
secured using new stainless steel guide pins (Nobel
Biocare, Goteborg, Sweden) for each framework. The
abutments were examined for correct seating as the
guide pins were lightly tightened using a screwdriver.
Cantilever framework components were individually as-
sembled on the apparatus, fitting the connectors with 2
concave radii between the 2 abutments mounted on the
VOLUME 96 NUMBER 1
HART AND WILSON
Fig. 3. Titanium framework and stainless steel implant components.
apparatus (Fig. 3). The cantilever section with 1 concave
radius was held in position using tweezers, with the ra-
dius firmly against an abutment, and, using the laser,
was tack-welded in the corners to hold the compo-
nents together for welding. The assembled frameworks
were welded in groups of 15 using 4 different welding
procedures: (1) LF0, a single circumferential laser weld
of each joint without filler wire; (2) LF1, an initial cir-
cumferential laser weld, followed by a second weld
pass with the addition of 0.6-mm-diameter CPTi filler
wire in each joint; (3) LF2, the components were laser
tack welded at the ends, and then a 0.6-mm groove
was cut into each joint to a depth of approximately 0.5
mm using a 0.6-mm noncontaminating separating disc
(Dedeco Intl Inc, Long Eddy, New York). Then the
joints were airborne-particle abraded with 50-mm alu-
minium oxide and cleaned. A laser weld was performed
in each groove, followed by 2 circumferential weld passes
incorporating 0.6-mm-diameter CPTi filler wire; and (4)
TIG, welds with the introduction of 0.9-mm-diameter
CPTi filler wire at intervals into the molten pool. The
lengths of the 0.6-mm- and 0.9-mm-diameter CPTi filler
wire (initially cut as 100-mm lengths) were measured
(mm) using 150-mm digital vernier callipers (Lion
[Aust] Consolidated, Croydon, Australia) prior to weld-
ing, and remeasured after welding each joint to allow
calculation of the volume (mm3) of filler wire used.
The length of filler wire was kept constant in each laser
weld pass; however, this was more difficult with TIG
welding, as the filler wire was fed into a molten pool.
All laser welds were made using an Nd:YAG laser
welder (LaserStar LYNX; BEGO, Bremen, Germany)
at 280 V with a 0.6-mm-diameter irradiating focus
and 10-ms pulses, as per the machine manufacturers in-
structions for welding titanium. High-purity argon gas
(.99.6%; BOC Gases, Melbourne, Australia) was used
to shield the weld at a flow rate of 15 L/min, with the
nozzle as close to the work-piece as practical (less than
10 mm).
Tungsten inert gas welds were made using a micro-
processor-controlled TIG welding machine (Invertec
JULY 2006
THE JOURNAL OF PROSTHETIC DENTISTRY
V160-T; Lincoln Electric Company, Cleveland, Ohio)
with a 1.6-mm electrode. A polarized direct-current
electrode negative (DCEN) arc was varied to a maxi-
mum of 35 amps, and a protective shield of high-purity
argon gas was provided through the torch nozzle at a
flow rate of 15 L/min. Additional gas shielding was pro-
vided continuously at a rate of 3 to 4 L/min to the rear
of the joint being welded, via a customized shield that
was connected directly to the gas regulator. Argon gas
shielding was provided to both sides of the joint during
the welding and cooling process to minimize oxygen
and nitrogen contamination of the metal in the heat-
affected zone.
The welded frameworks were then mounted on stain-
less steel implant replicas that were placed in the steel
testing apparatus, and new implant replicas were used
for each framework. New titanium alloy (Ti-Al6-V4)
abutment screws (CDI, Melbourne, Australia) were
used to fasten each framework to the implant replicas,
and were tightened to a torque of 35 N
cm using a man-
ual torque wrench (Manual Torque Wrench, Part No.
20-165; Nobel Biocare). The apparatus was mounted
in a vise at the base of a universal load testing device
(Instron Universal Testing Instrument 8501; Instron
Engineering, Canton, Mass). The assembly was posi-
tioned such that a V-shaped indenter attached to a
2000-N load cell would load the cantilever at a point
10.0 mm from the abutment cylinder (Fig. 4). The pa-
rameters of the loading device were set at a crosshead
speed of 2 mm/min, and the measurements were cali-
brated at zero deflection of the load cell. Stop limits
were set at 5.0-mm displacement and a maximum load
of 1900 N for safety reasons. Ten frameworks from each
group were subjected to flexural loading to failure using
the universal testing device, and the results were recorded.
The remaining 5 frameworks were placed in a closed-
loop servo-hydraulic testing apparatus (MTS 810
Material Testing System; MTS, Eden Prairie, Minn)
and set to oscillate at a frequency of 10 Hz in a half sinus-
oidal waveform with a cyclical load between 15 6 2 N
and 200 6 5 N. Automatic limiters stopped oscillation
27
THE JOURNAL OF PROSTHETIC DENTISTRY HART AND WILSON

Fig. 4. TIG-welded specimen mounted on apparatus for Fig. 5. Laser-welded specimen with filler wire added, show-
cyclic load testing at 200 N. ing fracture through weld zone.

Table I. Descriptive statistics of ultimate flexural strength

(N) and weld filler volume (mm3)
Minimum Maximum Mean load Mean volume
Group load load (SD) (SD)

LF0 226 263 243.5 (10.5)

LF1 427 460 445.7 (10.0) 1.70 (0.13)
LF2 737 782 755.9 (15.2) 3.44 (0.15)
TIG 1180 1812 1277.2 (190.1) 8.39 (1.03)
n=10.

The Cox proportional hazards model was used to con-

sider the volume of filler wire as well as the censoring.

Fig. 6. Failure of TIG-welded specimen showing deformation

rather than fracture. Note large radius of weld.
when the amplitude exceeded 2.0 mm, as a safety mech-
anism, and at 106 cycles. If failures occurred during
fatigue loading, the number of cycles withstood was
recorded. Data were entered into a spreadsheet (Micro-
soft Excel; Microsoft Corp, Redmond, Wash) in prepa-
ration for statistical analysis. The microstructures and
hardness of the laser- and TIG-welded joints were not
examined in this study, as the effects of welding and
heat on titanium microstructure and hardness have
been previously reported.9-10
The results of the ultimate flexural strength tests were
analyzed using a 2-way analysis of variance (ANOVA),
accounting for the sample group and the volume of filler
wire used in each joint. Post hoc comparisons were
performed using the Tukey adjustment for multiple
comparisons. A standard nonparametric analysis, the
log-rank statistic, was performed on the results of the cy-
clic loading tests due to the presence of censored data,
and separate pair-wise comparisons were also made us-
ing the log-rank test between each of the pairs of groups.
RESULTS
All of the laser-welded CPTi cantilever joints failed
through brittle fracture within the weld in both tests
(Fig. 5), whereas the TIG welds failed only by plastic de-
formation and only in the flexural strength test (Fig. 6).
The first 3 cyclic loading tests of TIG-welded frame-
works were censored at 1,000,000 cycles. The testing
of the fourth TIG-welded framework stopped automat-
ically at 866,473 cycles when the abutment screws failed
in the apparatus, rendering the apparatus unusable.
Thus, the fifth TIG-welded framework could not be
tested.
The ultimate flexural strengths for all 4 groups
ranged from 226 N to 1812 N, with no overlap between
the results of any of the groups. Descriptive statistics of
the flexural strengths are summarized in Table I, and
represented by a box plot (Fig. 7). Each laser weld pass
with filler wire approximately doubled the ultimate flex-
ural strength of the previous weld. The volume of filler
wire used in each weld was another variable measured
and was consistent within each group. A plot of load
against the volume of filler wire added is presented
(Fig. 8). Across the range of volumes there appears to
be a roughly linear relationship with strength.

28 VOLUME 96 NUMBER 1
HART AND WILSON
Fig. 7. Box plot of ultimate flexural strength (N) for 4 groups; outlier in TIG group represented by bar (1812 N).
Fig. 8. Scatter plot of ultimate flexural strength (N) compared to volume of filler wire added.
A 2-way ANOVA was fitted to the data, with both
volume and group as explanatory variables for load.
However, there was an outlier in the TIG group that vi-
olated the assumptions of the analysis. The results of the
ANOVA with the outlier removed are summarized in
Table II. There were highly significant differences
(P,.001) between the mean loads for the 4 groups,
even after adjusting for volume. This demonstrates
that not all groups had a common mean. The effect of
group was also highly significant (P,.001) for all 6
pair-wise differences between types of welds using the
Tukey adjustment for multiple comparisons. The
ANOVA also showed that there was a statistically signif-
icant linear relationship (P=.033) for load with volume
of filler wire after adjusting for group.
Calculation of least squares mean values for load,
adjusted for volume, showed that LF0 had a mean
JULY 2006
THE JOURNAL OF PROSTHETIC DENTISTRY
flexural strength of 291.5 N (SE mean 22.3 N), LF1
had a mean flexural strength of 468.4 N (SE mean
11.3 N), LF2 had a mean flexural strength of 752.4 N
(SE mean 5.6 N), and the TIG group was the strongest,
with a mean flexural strength of 1143.1 N (SE mean
34.1 N).
A statistical description of the results of the cyclic
loading fatigue testing at 200 N is presented in Table
III. Using a standard nonparametric analysis of these
data that allows for the censoring involved, the log-
rank statistic was 39.3 (df=3, P,.001), indicating that
there is significant evidence of a difference between
the groups. Separate pair-wise comparisons were also
made using the log-rank test between each of the pairs
of groups. All of these tests were significant, with
P=.002 in each case (or P=.012 after adjustment for
multiple comparisons using the Bonferroni method), a
29
THE JOURNAL OF PROSTHETIC DENTISTRY HART AND WILSON

Fig. 9. Scatter plot of cycles to failure at 200 N versus volume of filler wire added.

Table II. 2-way ANOVA for load, using adjusted sum of Table III. Descriptive statistics of number of cycles to
squares for tests without TIG outlier failure and weld filler volume (mm3)
Source df Seq SS Adj SS Adj MS F P Minimum Maximum Mean cycles Mean volume
Group cycles cycles (SD) (SD)
Volume 1 4937279 1431 1431 4.94 .033
Group 3 147356 147356 49119 169.68 ,.001 LF0 18955 20224 19794 (589)
Error 34 9842 9842 289 LF1 39125 75377 48601 (15059) 1.68 (0.14)
Total 38 5094478 LF2 125483 264863 208214 (58652) 3.29 (0.21)
TIG 886473 1000000 Censored 8.14 (0.83)
Seq SS, Sequential sum of squares; Adj SS, adjusted sum of squares; Adj MS,
adjusted mean square.

consequence of complete separation of the number of
cycles to failure between the groups. The specimens
endured between 18,995 cycles and censorship at
1,000,000 cycles. All 3 laser-welded groups failed below
300,000 cycles. The volume of filler wire introduced
into these welds was also consistent within each group.
A scatter plot of the relationship between the number
of cycles at 200 N to failure and the volume of filler
wire added for each of the groups is shown (Fig. 9).
When both filler wire volume and group were included,
the Cox proportional hazards model for censored data
showed that volume was a borderline predictor of the
number of cycles to failure (P=.05), whereas group
was not (P=.7).
DISCUSSION
A narrow range of results within each of the laser-
welded groups was observed for both ultimate flexural
strength and fatigue resistance tests. The welding condi-
tions, which include machine settings, shielding gas
flow, and volume of filler added, are determined by
the operator. If these parameters are matched, it should
be possible to reproduce the same results. The small var-
iation in the amounts of weld filler wire introduced into
any of the laser-welded joints is due to the fine control
the operator has over the volume of filler wire being
added, as it is introduced by overlapping small laser
pulses with the weld progressing by 0.1 mm to 0.2
mm per pulse. By maintaining consistent weld bead di-
mensions, close control over the amount and distribu-
tion of weld metal is possible. It is likely that this is one
of the reasons why little variation in strength within
experimental groups is found in most of the published
studies relating to laser-weld testing.3,10
Introducing filler wire increases the volume of weld
material, significantly improving the ultimate flexural
strength and also increasing the number of cycles en-
dured before fatigue failure. The significant differences
between the 3 laser-welded groups suggests that in-
creasing the volume of the weld is an important variable
that is likely related to the operator, as the operator has
direct control over the preparation of the joint and the
amount of filler wire introduced into the joint. By fol-
lowing standard welding procedures of preparing weld
joints, grinding a tapered end or cutting a groove into
the joint creates a deeper weld zone. The resultant use
of more filler wire to produce a larger weld increased
the strength of the joint.
The TIG-welded group was significantly different
from the laser-welded groups in flexural strength
tests, as the cantilevers all sustained loads in excess of

30 VOLUME 96 NUMBER 1
HART AND WILSON
1000 N. Ignoring volume, there are large average load
differences between the 4 groups; however, when vol-
ume is considered, the evidence for differences between
the groups is far less clear-cut. The ANOVA showed that
across the 4 groups there was a significant linear relation-
ship (P=.033) for load with volume. In fatigue testing,
3 of the TIG-welded frameworks endured 1,000,000
cycles of 200-N load, and the fourth framework stopped
at 860,473 cycles due to fracture of the abutment re-
taining screws. As there are such significant differences
between the groups, the effect of this failure on the re-
sults is minimal. The amount of filler wire used in the
TIG welding process was greater and more variable
than for the laser-welded specimens. A TIG welder in-
puts more heat, melting more metal and filler wire, to
make a larger, deeper, and stronger joint. However, var-
iations in the shape and depth of the TIG welds appear
to affect the strength of TIG-welded joints, as seen by
the variations for load with the volume of filler wire
used in the flexural strength tests.
The significant differences (P,.001) between the
groups suggest that the fatigue resistance may be more
likely related to the volume of filler wire added
(P=.05) than the group itself (P=.7). That is, after ac-
counting for a volume of weld effect, it was shown due to the more efficient protection of the heat-affected
that the strength of a TIG weld is not intrinsically differ-
ent from that of a laser weld.
The high outlier in the flexural strength test for the
TIG-welded group significantly affected the statistical
analysis. This may be indicative of the variation that
can be expected from this welding method; however,
the production of stronger welds is beneficial to the lon-
gevity of a prosthesis. The fact that the outlier was also
the specimen that used the most filler wire further rein-
forces the trend of greater filler volumes producing
stronger joints. When the technique for producing a
TIG weld is considered, it should be appreciated that
the device used was a commercial welding unit, designed
to be used on objects of a larger scale than the implant
components assembled. It is extremely difficult to have
as precise a control over a plasma arc as one can have
over a focused laser beam. The filler wire is melted
into the weld pool in essentially uncontrolled droplets.
The volume of these droplets is likely to be determined
by the diameter of the wire, the proximity to the arc, op-
erator dexterity, and possibly the angle at which the filler
wire is introduced to the weld pool. Hence, not only is
control over the amount of filler wire added to the mol-
ten weld pool difficult, but also the depth and distribu-
tion of the molten weld pool throughout the weld joint,
resulting in variations in flexural strength relative to the
volume of filler wire.
Conversely, for laser-formed welds, where the small
focal diameter and the localized nature of the weld
zone mean that control over the weld region is im-
proved, it is necessary to precisely align the beam under
JULY 2006
THE JOURNAL OF PROSTHETIC DENTISTRY
magnification, so that adequate heat is provided to form
a weld into both components. This is complicated by the
potential for reflection of the beam off the surface of the
weldment.
To melt filler wire, increased heat input is required,
which is normally associated with problems of distor-
tion, and consequent misfit of bounded sections will
occur. However, in the fabrication of cantilevers by
welding, this can be less of a problem, as adjustments
can be made after welding.2 By fabricating the cantilever
section first using either TIG welding or laser welding
with added filler wire, laser welds use less heat and may
then be used for creating the bounded sections between
abutments, where distortion is best minimized. Alterna-
tively, abutments with prefabricated cantilevers could
be commercially manufactured, and these joined using
laser welding with filler wire to form a cantilevered
superstructure.
It was noted that the TIG-welded joints in this study
failed only by plastic deformation, whereas others re-
ported that TIG welds failed by brittle fracture in the
heat-affected zone.9,10 Failure by plastic deformation
in this study indicates that the metal in the heat-affected
zone of the joint is not hard or brittle. This is most likely
zone during the welding and cooling processes by argon
flowing directly onto both sides of the weld area, rather
than welding beneath an argon blanket under a dome.
The effect of protecting the heat-affected zone from
oxygen and nitrogen during welding and, just as impor-
tantly, during the cooling process, was demonstrated by
studies that reported that electron beam welding of
CPTi carried out under a vacuum, and allowed to cool
under vacuum, resulted in no change in the hardness
of the metal in the heat-affected zone.9,10
It is not possible from the small sample size to draw
conclusions relating to minimum flexural strengths re-
quired to provide fatigue resistance under masticatory
loads. However, it is suggested that increasing the
amount of filler wire deposited in laser-welded joints
may improve the strength and fatigue resistance of the
framework.
The difficulty of applying the findings of these results
to clinical situations is related to the large clinical vari-
ation. The magnitude of the applied masticatory and
parafunctional forces is not easily quantifiable, and the
dimensions of a superstructure will be dictated by
the space available for the definitive restoration and
the veneering material selected. A conservative estimate
of 50 maximal occlusal forces per day equates to 18,000
cycles per year. Hence, it could be inferred that the ob-
servations of these laser-welded groups represent clinical
failures of up to 10 years function. This would suggest
that the use of laser welding without filler wire to pro-
duce titanium frameworks for definitive prostheses
should be cautioned against, unless the strength of the
31
THE JOURNAL OF PROSTHETIC DENTISTRY
weld is improved by increasing the volume of filler wire
added. Its application to provisional restorations may be
appropriate.
Laser welding presents a method of joining titanium
that can produce consistent results with low distortion,
and may be useful in creating implant-bounded sections.
TIG welding can be used to produce consistently strong
welded titanium joints, but the technique is more
demanding.
CONCLUSION
Within the limitations of this study, it was concluded
that the ultimate flexural strength (P=.03) and fatigue
resistance (P=.05) of cantilevered joints in laser-welded
titanium prostheses are influenced by the volume of filler
wire used in producing the welds. Early failure of pros-
theses can be expected where laser welding is performed
without the addition of filler wire. Increasing the size
and depth of a laser weld by joint preparation allows
for increased filler wire deposition, resulting in increased
strength and durability of the joint. Tungsten inert gas
welding, which deposited the largest volume of filler
wire, produced the strongest joints. The strength of
TIG-welded joints, however, is dependent on the effec-
tive argon gas shielding of the heat-affected zone during
the welding and cooling processes.
The authors thank Mr Joseph Palamara, Mr Peter Vryionis, and Mr
George Thalassinos for their technical assistance, and Professor Ian
Gordon for his invaluable assistance with the statistical design and
interpretation.
REFERENCES
1. Sjogren G, Andersson M, Bergman M. Laser welding of titanium in
dentistry. Acta Odontol Scand 1988;46:247-53.
2. Wee AG, Aquilino SA, Schneider RL. Strategies to achieve fit in implant
prosthodontics: a review of the literature. Int J Prosthodont 1999;12:
167-78.
3. Liu J, Watanabe I, Yoshida K, Atsuta M. Joint strength of laser-welded tita-
nium. Dent Mater 2002;18:143-8.
4. Neo TK, Chai J, Gilbert JL, Wozniak WT, Engelman MJ. Mechanical prop-
erties of titanium connectors. Int J Prosthodont 1996;9:379-93.
5. Berg E, Wagnere WC, Davik G, Dootz ER. Mechanical properties of
laser-welded cast and wrought titanium. J Prosthet Dent 1995;74:250-7.
32
HART AND WILSON
6. Roggensack M, Walter MH, Boning KW. Studies on laser- and plasma-
welded titanium. Dent Mater 1993;9:104-7.
7. Yamagishi T, Ito M, Fujimura Y. Mechanical properties of laser welds of
titanium in dentistry by pulsed Nd:YAG laser apparatus. J Prosthet Dent
1993;70:264-73.
8. Gordon TE, Smith DL. Laser welding of prosthesesan initial report.
J Prosthet Dent 1970;24:472-6.
9. Wiskott HW, Doumas MT, Scherrer SS, Susz C, Belser UC. Microstruc-
tures of brazings and welds using grade 2 commercially pure titanium.
Int J Prosthodont 2001;14:40-7.
10. Wang RR, Welsch GE. Joining titanium materials with tungsten inert gas
welding, laser welding, and infrared brazing. J Prosthet Dent 1995;74:
521-30.
11. Jemt T, Linden B. Fixed implant-supported prostheses with welded tita-
nium frameworks. Int J Periodontics Restorative Dent 1992;12:177-84.
12. Jemt T, Henry P, Linden B, Naert I, Weber H, Bergstrom C. A comparison
of laser-welded titanium and conventional cast frameworks supported by
implants in the partially edentulous jaw: a 3-year prospective multicenter
study. Int J Prosthodont 2000;13:282-8.
13. Ortorp A, Jemt T. Clinical experiences of implant-supported prostheses
with laser-welded titanium frameworks in the partially edentulous jaw:
a 5-year follow-up study. Clin Implant Dent Relat Res 1999;1:84-91.
14. Bergendal B, Palmqvist S. Laser-welded titanium frameworks for implant-
supported fixed prostheses: a 5-year report. Int J Oral Maxillofac Implants
1999;14:69-71.
15. Bergendal B, Palmqvist S. Laser-welded titanium frameworks for fixed
prostheses supported by osseointegrated implants: a 2-year multicenter
study report. Int J Oral Maxillofac Implants 1995;10:199-206.
16. Johansson LA, Ekfeldt A. Implant supported fixed partial prostheses: a ret-
rospective study. Int J Prosthodont 2003;16:172-6.
17. Haraldson H, Carlson GE. Bite force and oral function in patients with
osseointegrated oral implants. Scand J Dent Res 1977;85:200-8.
18. Hecker DM, Eckert SE. Cyclic loading of implant-supported prostheses:
changes in component fit over time. J Prosthet Dent 2003;89:346-51.
19. Graf H. Bruxism. Dent Clin North Am 1969;13:659-65.
Reprint requests to:
DR PETER R. WILSON
SCHOOL OF DENTAL SCIENCE
FACULTY OF MEDICINE, DENTISTRY AND HEALTH SCIENCES
THE UNIVERSITY OF MELBOURNE
711 ELIZABETH STREET
MELBOURNE, VICTORIA, 3000
AUSTRALIA
FAX: 61 3 9341 0339
E-MAIL: prwilson@unimelb.edu.au or cnhart@mac.com
0022-3913/$32.00
Copyright 2006 by The Editorial Council of The Journal of Prosthetic
Dentistry.
doi:10.1016/j.prosdent.2006.05.003
VOLUME 96 NUMBER 1