(1.

6) ALUMINUM SOLDERING - A NEW LOOK

A. E. Gickler, F. H. LePrevost Jr.,* T. Pan, A. M. Jaoquin,# C. A. Blue, and M. L. Santella^
ABSTRACT
Soldering aluminum has never been a mainstream process for industries and manufacturers. The
main reasons are: tenacious aluminum oxide which foils most attempts to solder using
conventional means; its dissimilarity with many solders and base metals for potential galvanic
corrosion consequences; the varieties of aluminum alloys, gages and tempers with varying
soldering results; and the most often misunderstood and overlooked factor of how aluminum
accepts or rejects heat during soldering.
Soldering can be a very attractive joining method for aluminum with much less heat distortion
due to its lower process temperature than brazing and fusion welding. Soldering can be done
with either soft solders (Sn-based, lower temperature) or hard solders (Zn-based, higher
temperature), with appropriate fluxes to fit processing temperature ranges. Various aluminum
alloys have different solderability: 1xxx, 2xxx, 3xxx, 4xxx and 7xxx are easier to solder than
6xxx series alloys. 5xxx series alloys are the most difficult due to their magnesium content.
Soldering requires adequate heat on the component, not on the solder. Because of the high
thermal conductivity and reflectivity of aluminum, it has been found that neither oxy-acetylene
flame torch, plasma arc, laser, induction heater, nor thermal spray is capable of providing good
results. Through a collaborative project between Oak Ridge National Laboratory, Ford Motor
Company, and Johnson Manufacturing Company, a high density infrared source of 300kW
plasma lamp has been demonstrated to have successful results with 80 wt% Zn- 20 wt%Al and
flux at a soldering temperature of 490C. Mechanical tests showed that the joint area is stronger
than the parent material with minimum softening. Detailed process parameters have been studied
to provide a basis for future mass production.
INTRODUCTION
Soldering Aluminum is not a common process in industry today. The following study will
address why someone would want to solder aluminum and what methods or processes are
available. Also discussed will be choices of alloys and fluxes (if needed) and some interim
results from a study involving an automotive OEM (Ford Motor Company), a National
Laboratory (Oak Ridge National Laboratory) and a solder/flux supplier (Johnson Manufacturing).
Soldering is by definition a low temperature joining process. Therefore, less distortion of the
aluminum component is expected by soldering than by brazing, welding or other fusion joining
Johnson Manufacturing Company, Princeton, IA
Ford Motor Company, Ford Research Laboratory, Dearborn, MI
^
Oak Ridge National Laboratory, Metals and Ceramics Division, Oak Ridge, TN
*

processes. Soldering temperatures of 225C to 490C are well below the 661C aluminum
melting temperature although 490C is above most annealing temperatures. Stresses in the
aluminum from shearing, drawing, and heat-treating are changed by the localized heating
encountered during soldering and distortion may result. However, brazing and fusion welding
would have a much greater problem in this area. Techniques such as pre heating, non-continuous
joints and a careful selection of joint geometry can reduce the degree of distortion.
A lower capital investment in both equipment and facilities is required for soldering when
compared to brazing. A reduced capital investment in the tools used to produce the component
parts may also be possible since the fit tolerances required for soldering are generally greater
than for brazing. Disassembly, re-manufacturing and repair is much easier and more practical.
Also, soldering is nearly always done in air without the need for expensive reactive or inert
atmospheres.
Typical applications of aluminum soldering have been heat exchanger assembly,
electronic/electrical capacitor manufacturing, and light bulb manufacturing. None of these
applications requires high tensile strength, but, they do require the solder to not cause galvanic
corrosion over time.
PROCESSES
Fluxless processes for aluminum soldering include mechanical rubbing of aluminum with solder,
ultrasonic bath soldering, and thermal spray. Flux-based methods include heating the assembly
by induction, flame, infrared, hotplate, furnace, soldering iron, laser and arc lamp.
Soldering aluminum is not a common practice. The primary reason is that rapid formation of an
oxide layer, and the difficulty in removing that oxide layer, inhibits solder wetting of the
aluminum. When soldering copper, removal of the copper oxide is relatively easy with mild
organic and inorganic fluxes. Aluminum oxide is not so easily removed and requires strong
fluxes such as an organic amine based (up to 285C), inorganic fluxes (chloride or fluoride up to
400C) and complex fluoroaluminate salts (above 550C). The use of mechanical rubbing,
ultrasonics or thermal spray depends upon using the molten zinc to abrade away the aluminum
oxide layer to allow wetting of the aluminum. No flux is used. Tin/zinc soft solders (soft solders
usually have bulk shear strength around 35 MPa or 5,000 psi) are typically used with the first
two fluxes since their melting point is under 330C. The zinc component helps in preventing
galvanic corrosion. Zinc based hard solders (below 400 C melting point and bulk shear strength
approaching 175 MPa or 25,000 psi) use fluxes that require over-heating of the aluminum alloy.
If soft solder fluxes are used, the residue that remains after soldering must be removed. Solders
used for aluminum generally contain zinc with some lead, cadmium, tin, copper, and aluminum.
Tin/lead, for instance, can be used but a long-term electrochemical corrosion problem may
develop. With the anticipated worldwide ban on lead in solder, most industries have or are
switching to lead-free solders. This removes some of the more ductile and/or higher temperature
soft solders available for the materials engineer. Cadmium bearing solders have been effectively
banned due to worker health issues.

MATERIALS
Table 1 lists some of the lead-free and cadmium-free alloys that can be used to solder aluminum.
Some alloys are common, such as 91Sn9Zn, 70Sn30Zn, and 98Zn2Al. Others just appear in the
trade literature as being invented. Note that all of the solders in this list contain zinc.
92Sn/8Zn (eutectic)
91Sn/9Zn
80Sn/20Zn
75Sn/25Zn
70Sn/30Zn

98Zn/2Al
97Zn/3Al
95Zn/5Al
90Zn/10Al
85Zn/15Al

60Sn/40Zn
80Zn/20Al
Table 1 Soft and Hard Aluminum Solders

76Sn/21Zn/3Al
71.5Sn/25Zn/3.5Al
67Sn/27Zn/3Al
55Sn/33Zn/11Al/1Cu
67Sn/17Zn/15Al/1Cu
90.3Sn/9.1Zn/0.6Al (eutectic)

Aluminum alloys often have alloy additives to improve strength, rigidity, corrosion resistance,
machineability and formability. Some additives cause no problem for soldering, but magnesium
is the exception. Magnesium containing aluminum alloys (e.g. 5xxx and 6xxx series) are used
for extending the strength to weight ratio and better corrosion resistance in some applications.
However, the author(s) are not aware of any solder or flux that is very effective with magnesium
containing aluminum alloys. The magnesium oxide is too difficult to remove and does not allow
solder wetting to take place. Titanium and some exotic additives such as vanadium and
chromium may also cause solderability problems. The 1xxx (99% Al or higher), 2xxx (copper
added), 3xxx (Manganese added), 4xxx (silicon added) and 7xxx (zinc added) series are
generally solderable. The 5xxx (magnesium added) series is probably not solderable and the
6xxx (silicone and magnesium added) series may or may not be solderable depending upon the
individual alloy. The 6061 discussed later is definitely solderable, Also, the 2xxx series in sheet
form typically has a 6xxx clad which determines solderability.
In some cases the aluminum can be plated with nickel or coated with zinc by thermal spray or
other methods. These surfaces are readibly solderable. Soldering aluminum to other metals
(steel, galvanized steel, copper, brass, stainless, etc.) is also done, but with some difficulty since
the joint design must allow for differential thermal expansion and the same flux seldom works
for both metals. The simple job of heating the assembly at the joint area becomes difficult since
the aluminum conducts heat away from the joint very rapidly as compared to other metals. A
general rule of thumb in soldering is heat the component, not the solder. This allows the
substrate to transfer heat to the solder and melt the solder once it is up to the melting
temperature. Fluxes can insulate the solder from the substrate and cause the reactivity of flux to
expire before the solder has melted. Over-heating a flux can leave a hard residue that the solder
cannot penetrate through in order to wet the substrate. Cored soft solders normally eliminate this
problem since the flux is not released until the solder melts. Cored hard solder fluxes leave little
residue.

Aluminum can become much less difficult to solder providing certain conditions are met and
most importantly, proper heat can be applied. Understanding what methods of heat work best for
soft soldering or hard soldering has been the subject of several years of investigation by Johnson
Mfg. Co., Ford Motor Company and Oak Ridge National Laboratory. The desire to take a fresh
look at all available heat sources is based on a shared opinion that aluminum soldering has
numerous potential applications for the automotive industry.
Due to its low melting temperature, aluminum and aluminum alloys may be annealed or
tempered at temperatures as low as 325-350C in a relatively short time. This suggests that any
joining process approaching these temperatures for even a brief interval will begin to alter the
properties of the parent metals being joined. Overheating may result in stress relieving, sagging
or warpage, altering hardness, temper, surface condition, re-alloying of the base metal in the
immediate joint area, or even melting of the base metals.
Generally speaking, soft solders do not pose a significant risk to the parent material properties
from heating, provided the parts are not held at soldering temperatures for an extended period of
time. However, in some cases, exposure of aluminum to the molten alloy for even a short period
of time may result in re-alloying of the parent metal within the heat-affected zone (HAZ). This
will change its properties and may cause what appear to be heat cracks that emanate from the realloyed area since soft solders contain tin with lead, zinc, silver, or copper.
Hard-solders fall into the zinc-based family of alloys. The most widely used of these is
98Zn/2Al, having a melting range of 376-385 C. These alloys have been used at temperatures up
to 550C using all available heat sources. Lead-free composite body solders, tin-copper-zinc
alloys, also fall under the hard solder category due to their high liquidus temperature of 500C
and their semi-solid nature. Soldering aluminum with soft or hard-solders may be accomplished
by either chemical or mechanical means, but generally not both. Capillary action by molten
solder between parts generally requires the use of flux.
Under laboratory conditions certain heating methods have been found to produce excellent
soldering results. However it is not always possible, to duplicate these methods for in plant
production. Using a ceramic hot plate and either torch or soldering iron enables one to, in effect
slow down the process so it can be better understood and repeated. More often, soldering
aluminum requires a continuous application of modest heat, rather than highly intense heat like
an oxy-acetylene flame. In addition, localized heating may also be required to enable solder to
flow into a joint or capillary along a seam.
HEATING METHODS
The following heating methods were explored to determine the most likely candidates for
soldering automotive panels on an assembly line. For this application open flame cannot be
considered. Selected solders were tested using Plasma Arc wire feed equipment on aluminum
panels that were spot-welded to provide an overlap joint with a ditch for filling. This high-energy
heat source produced mixed results, but in the final analysis, it was observed that the heat is so
localized that neither soft, nor hard solders could flow consistently well between both panels.

Preheating was also tried to reduce heat dissipation. While this did promote better capillary flow
between the panels, hot cracking became evident in some samples due to the rapid build up of
heat from the plasma arc. Heat sinks might be used to prevent the build up of heat in unwanted
areas and promote rapid cooling of solder joints.
Using powerful lasers (diode and CO2), solders were tested on aluminum panels of the same
alloy, having an overlap joint with ditches of various widths. Unlike steel, the highly reflective
surfaces of both aluminum and the solder made it difficult to bring the panels and the filler alloy
to uniform temperature to effect wetting. Best results were achieved by defocusing the laser's
beam and moving it slightly off center to spread the heat where needed. One problem still
remained, i.e., positioning solder on the work piece or feeding it through the beam must be done
in such a way as not to shade the joint, or melt the filler metal prematurely. Dual-beam lasers
may be required to accomplish this task.
Induction provides a rapid, delocalized heating technique for even large parts. Unfortunately,
induction heating did not provide sufficient heat to melt the solder when either laid upon the
panel, or being wire fed into the joint; this was due to its small contact area between the solder
and the heated surface. Secondary heat is required to bring the work piece and solder to the
proper temperature simultaneously, enabling solder to flow into the joint, fill ditches, etc.
Overall, induction heat used for soldering aluminum may be a good subject for continued
investigation.
Both tin-based and zinc-based alloys have also been tried using the thermal spray to deposit on
aluminum body panels as well as galvanized steel. This process does not enable solders to
capillary or flow between panels, plus the bond was not as good, and porosity was higher than
expected in early trials. More work is presently being done in this area to develop the means to
apply as body fillers to automotive seams requiring a Class-A finish.
INTRODUCTION TO VEHICLE STUDY
Most automotive companies use light gage steel for the vehicle body panels. Ford Motor
Company, along with other automotive companies, has been considering the use of 6xxx
aluminum alloys to save weight for increased performance and fuel economy. Aluminum is also
less expensive compared to alternatives such as fiberglass and exotic plastics. Aluminum also
has strength and durability comparable to steel body structure.
Due to aluminums high affinity for oxygen, it develops a tenacious oxide layer. This oxide
layer makes it extremely difficult to fuse aluminum-based materials through conventional
resistance spot welding techniques. A prospective method for joining such panels is to use spot
welding to temporarily hold panels in place. This would be followed by hard soldering the
aluminum body panels together, incorporating the use of a strong flux. Flux would, of course,
serve to eliminate the oxide layer making it possible to join the parts with a metallurgical bond.
Hard soldering could be accomplished through the use of a newly-developed high-density
infrared (HDI) source. In other words, a radiation source with a wavelength below the visible

spectrum of light, would be used to heat both the solder and joint (Fenton, 1963). An HDI
source is preferable because of aluminums extremely high reflectivity and thermal conductivity.
This HDI source allows the entire solder area to be heated extremely quickly, before a large
amount of heat can be transferred away from the joint, negating the need to heat the entire panel.
Ford is currently cooperating with Oak Ridge National Laboratory to investigate the use of
ORNLs unique 300 kW plasma lamp as an HDI source for soldering aluminum. This HDI
technology is combined with a robot to develop a computer controlled process.
The goal of this project was to create a strong metallurgical joint between 6061 aluminum
panels, supplied by Ford Motor Company, using the plasma lamp as the HDI source to solder the
joint. The Johnson Manufacturing Company supplied the solder and flux used.
MATERIALS AND METHODS
All experiments used the same solder designated as a zinc-aluminum alloy composed of 80
weight percent zinc and 20 weight percent aluminum. Externally applied flux paste was used in
conjunction with the solid solder wire. The melting range of the solder is 370 to 430C. 6061
aluminum has a melting point of 652C (Metals Handbook, Vol. 2).
Laboratory experiments were performed in a highly controlled 33 kW infrared (IR) furnace to
observe interfacial phenomena as a function of time, temperature, and fluxing agent. Specimens
consisted of reasonably uniform weight samples of flux core solder or solid solder and externally
applied flux on small pieces of 6061 aluminum. Different temperatures as well as different
furnace ramp times were evaluated. Experimentation was performed with a target furnace
temperatures of 450C, 475C, 480C, 490C, and 500C. It was established that 490C was the
best target temperature for melting and wetting and was used for the majority of the experiments.
IR furnace ramp times were varied from less than one minute to thirty minutes. Holding times
were thirty seconds for all runs.
After sufficient experiments were performed in the 33 kW IR furnace, similar samples were used
for testing with the plasma lamp facility. In the lamp, an electrical arc is struck between two
tungsten electrodes at a very high current. A quartz tube surrounds the electrodes. The tube is
kept cool by a continuously circulating jet of water. The arc is struck inside the tube between the
two electrodes in a non-reactive argon atmosphere, which serves to increase the life of the
electrodes during lamp use. The electric arc itself is what creates the infrared radiation.
Peak power density of the radiation is determined by the amperage at which the lamp is
programmed to run and the distance of the lamp from the surface being heated. Different
infrared focal lengths were evaluated. The lamp head used for the above brazing experiments
had a 4.0 cm focal length. Prior to experimentation, a model was used with the Telluride
program developed at Los Alamos National Laboratory specifically for the Oak Ridge National
Laboratory plasma lamp to determine the amperage needed to achieve a temperature of at least
490C. This current was predicted to be 410 amps by the model (Rivard). Each experiment was
performed with the long focal length lamp to provide a calculated peak power density of 330
W/cm2.

Metallographic analysis was performed on various small samples from each the IR furnace and
plasma lamp to examine the interface and void content of the specimens.
Lap joint specimens were made with the plasma lamp. They consisted of two plates spot-welded
together, with a one inch by two overlap. The individual parts were the same thickness as the
original small samples, as that they were made from the same material. The specimens were
three inches long due to the one-inch overlap for the lap joint. Spot welds were created using a
new spot welding machine that combined the application of a punching force and electric current
to fuse the surfaces together. It was predicted that 600 amps at a 1.8 second pulse time would be
needed with the HDI lamp to reach a desirable temperature for hard soldering (Rivard & Sabau).
The model predicted this as the current and pulse time for the proper heat transfer for melting
and wetting through capillary action through the joint. However, these parameters did not
provide the desired results, therefore, experimentation was performed with various pulse times,
currents, and specimen positions under the lamp. A short focal length lamp (1.0 cm) was used.
Experiments with this lamp were run at 280 amps and 300 amps for various time increments,
since prior experiments for Ford with this material produced reasonable results with the same
amperage on this lamp. The two samples from Ford were tested at 300 amps for a pulse time of
3.8 seconds. Using the trend line equation Power Density = 0.2611 x Amperage1.3529 developed
on Excel, the power density of 280 amps is calculated to be 534 W/cm2. The power density of
300 amps was measured to be 562 W/cm2 through prior experimentation with the short focal
length lamp (Rivard).
Mechanical testing was performed to complete the experiments with the one inch by three-inch
specimens. Main concentration was given to yield strength (with 0.2 percent offset), ultimate
tensile strength, and maximum strain.
RESULTS
Solder weight differences had negligible effect on experiments due to the small size of the
samples relative to the capabilities of the heat sources. Experiments with the IR furnace showed
long ramp times of fifteen to thirty minutes, and lower temperatures of 450C and 475C,
yielded a large amount of porosity throughout the solder and also at the interface. This is
illustrated by Figure 1, showing the result of a 15 minute run to a high temperature of 476C
using flux cored solder. Good wetting occurred at shorter times and temperatures of 490C or
higher. This is shown in Figure 2, where a flux core solder sample is heated to a high
temperature of 492C in 5 minutes. Ramp times of less than one minute with target temperatures
of 490C yielded excellent results for each solder. This is shown in Figure 3 for flux core solder,
and Figure 4 for solid solder with flux.
The small specimens used in the plasma lamp showed exceptional wetting. Figure 5 illustrates
the excellent result of solid solder with flux run at 410 amps for a pulse time of 1.8 seconds.
Figure 6 shows the wetting was acceptable with the flux core sample ran at 410 amps for 1.7
seconds; however flux core specimens show more porosity with the lamp than solid samples.

The lap joint specimens did not produce consistent results. Some specimens experienced
excellent wetting of the joint during one run, and then a complete absence of wetting at the joint
on an identical run intended for confirmation of the first results. Solder wetting away from the
joint, either on the top section or the bottom section was observed with little to no solder contact
with the joint. This was likely due to poor heat transfer and will be discussed later in this paper.
Samples fabricated at Ford of similar size were already spot welded together, and worked
extremely well under the short focal length lamp. Figures 7 and 8 show what the samples looked
like before and after soldering, respectively.
Mechanical testing of the two samples from Ford after successful soldering showed the average
yield strength (at 0.2 percent offset) to be 69.6 MPa (10,099.5 psi) and the average tensile
strength to be 199.2 MPa (28,462.3 psi). The maximum strain average was 12.7 percent. As
shown in Figure 9, necking occurred on each specimen both above the spot weld and below the
soldered joint. Each specimen fractured on the top section near the spot weld.
DISCUSSION AND CONCLUSION
The quality of the interface increased with increasing temperatures and decreasing ramp times in
the 33 kW IR furnace for each solder type. This strongly suggested that pulse soldering with an
HDI source would be very effective. ASM states that, in general, for production work, both
temperature and time are kept at a minimum (which is) consistent with good quality, (Schwartz,
1987). After testing each solder type on small samples under the lamp, metallographic analysis
revealed that the solid solder presented the highest quality interface; therefore, the lamp would
likely favor use of the solid solder. Flux cored solder in larger specimen sizes was considered
inappropriate for pulse soldering due to its constant moving around prior to melting and wetting.
This is likely because of vaporization of the flux core during HDI heating and volatile outgassing at points on the surface prior to the solder fully melting. The solid solder, as suspected,
was more suited for HDI pulse brazing.
It was suspected that the contact interface between the two pieces composing the specimens was
poor due to bending from force applied by the spot welding machine. It is also very likely that
the weld contact itself was flawed by a surface oxide layer not removed prior to electrical spot
welding, suggested by the frailty of the specimens after fabrication. Such oxide layers are
actually destroyed in ultrasonic spot welding. These factors combined to cause inadequate heat
transfer, which yield poor results when soldering the parts. The perfectly repeatable success with
the Ford samples and solid solder further supports this theory.
After mechanical testing, it was noted that the Ford-produced solder joints were perfectly intact,
thus mechanical characteristics obtained were actually those of the material itself. Fracture
above the spot weld during mechanical testing reflects that the spot weld functioned as a stress
concentrator for crack initiation. The properties obtained through mechanical testing fell
between the wrought and T4 temper values for 6061 aluminum (Metals Handbook, Vol. 2),
suggesting that there was some tempering induced by the HDI heating and air cooling. Even a
simple joint, when properly designed and made, will have strength equal to or greater than that of
the as-soldered parent metal, (Brazing, 2001). In general, this is an excellent hard solder joint.

After mechanical testing, it appears that use of an HDI source to pulse solder the aluminum parts
with solid solder and externally applied flux would be feasible for automobile body
manufacturing. However, further conformational experimentation should be performed.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.

ASM International (1990). Properties of Wrought Aluminum and Aluminum Alloys.

Metals Handbook, 10th Edition. Vol. 2, pp. 102-103.
ASM Desk Editions Online Metals Handbook Desk Edition. Brazing and Soldering.
ASM International (2001). Retrieved 1 August 2002 from
http://www.asminternational.org/asm/servlet/Navigate
Fenton, E. A. (1963). Brazing Manual. New York, NY: American Welding Society,
Inc. pp. 9
Rivard, John D.
Sabau, Adrian S.
Schwartz, M. M. (1987). Brazing. Metals Park, OH: ASM International

Figures

Figure 1. High porosity with flux cored solder sample created by ramping to 476C in 15
minutes.

Figure 2. Good wetting flux core example created by ramping to 492C in 5 minutes.

Figure 3. Excellent flux cored solder interface created by ramping to 510C in 41 seconds.

Figure 4. Good solid solder interface created by ramping to 513C in 30 seconds.

Figure 5. Excellent interface provided by solid solder with flux ran at 410 amps for 1.8 seconds.

Figure 6. Good wetting, but more porosity with flux core solder ran at 410 amps for 1.7
seconds.

Figure 7. Ford specimensreverse side shows what pre-soldered joint looks like.

Figure 8. Ford specimenssoldered joint with solid solder.

Figure 9. Necking occurred above and below the soldered joints during tensile testing.