CA2413246A1 - Aluminum filler alloy containing sodium for fluxless brazing - Google Patents

Abstract

An aluminum filler alloy, particularly useful in fluxless controlled atmosphere brazing (CAB), contains from about 4 % to 20 % wt.% silicon (Si) and about 0.0008 % to 0.06 % sodium (Na). In addition to the sodium, the alloy may also contain bismuth (Bi) and/or potassium (K) in the ranges of about 0.0005 % to 0.03 % K, and from about 0.03 % to 0.133 % Bi. The filler metals can be clad to aluminum core alloys preferably from the 3XXX, 5XXX, or 6XXX
alloys series.

Description

ALUMINUM FILLER ALLOY CONTAINING SODIUM FOR
FLUXLESS BRAZING
PRIORITY CLAIM UNDER 35 U.S.C. ~119(,e1 The present application claims benefit under 35 U.S.C. ~119(e) of U.S.
Provisional Application 60/213,274 filed June 22, 2000.
BACKGROUND OF THE INVENTION
1. Field of the Invention The present invention relates, in general, to an ahuninum filler metal that contains silicon (Si) and intentional additions of sodium (Na) or Na + bismuth (Bi) or Na + potassium (K), or Na + Bi + K. The combination of a core alloy and the new filler metals is brazed in a furnace with an inert, non-oxidizing atmosphere without the need of a flux.

2. Description of the Background Art Fluxless brazing is not a new or revolutionary concept when brazing aluminum ~
.
materials. In the mid 1970's, fluxless brazing was first introduced to the world on a commercial scale as vacuum brazing. Vacuum brazing was the main joining process for aluminum for about 20 years. The principles behind vacuum brazing were relatively simple, i.e., Mg was needed in the filler alloy to sublime and breach the tenacious oxide film on the brazing sheet surface. In a commercial brazing cycle, the pressure level was reduced to about 5 X 10-5 Torr while simultaneously heating the furnace to the brazing temperature near 1112°F (600°C). Just before the onset of melting, the pressure-temperature combination allowed Mg in the filler metal to sublime and disrupt the Ah03 film on the surface of the sheet. During the heating cycle, Mg also reacts with the A1203 film to form MgAlzOø, a Mg-spinel. The Mg-spinel is more friable than A1203 and more easily disrupted when the Mg sublimed. Magnesium spinels are also more easily wetted by molten aluminum.
This combination of reaction, sublimation disruption, and wetting characteristics allowed vacuum brazing to become a commercial success.
Since Mg was needed in the filler alloy, Mg could also be used in the core alloy to boost core strength levels with out affecting brazeability. Consequently, vacuum brazing alloys typically consisted of a 3005-type alloy clad with a 4045-type filler metal containing Mg levels ranging from 0.25%-1.5% (4045+1.5% =4104). Typical core alloys had post-braze strength levels around 10 ksi (69 MPa).

Vacuum brazing was, however, not without its problems. Major problems included capital equipment costs that were high, fit-up tolerances that were carefully controlled, surface cleanliness had to be very good, vacuum furnace atmosphere had to be carefully controlled, and furnace maintenance was hazardous due to Mg-rich deposits that could ignite when cleaning. These factors in combination with the advent of the new NOCOLOKTM
brazing process Ied to the decline of vacuum brazing. Vacuum brazing is still being used to produce some heat exchangers that have yet to be converted to the NOCOLOKTM
process.
However, the vast majority of heat exchangers are currently brazed by the NOCOLOKTM
process.
The NOCOLOKTM process was quickly adopted by the commercial heat exchanger makers due to its robustness. Unlike the vacuum brazing process, NOCOLOKTM
brazed assemblies could be dirtier, fit-up tolerances were less critical, capital equipment costs were significantly less, and the leak rate of brazed cores was near zero. However, the NOCOLOKTM brazing process, has some limitations.
The NOCOLOKTM process was developed by Alcan in 1976, but it was not commercially adopted until about 1988. The NOCOLOKTM brazing describes a process in which a potassium cryolite (KAlF4-K3AlF6) powder is suspended in a water carrier and sprayed onto a heat exchanger surface. After spraying the potassium cryolite slurry onto the component, the water is evaporated in a drying oven before brazing. Next, the component to be brazed is placed into a control atmosphere brazing (CAB) furnace with a nitrogen atmosphere and brazed. During the heating cycle, the flux melts about 10°F (5°C) to 15°F
(8°C) below the melting point of the filler metal, i.e., 1065°F
(574°C). When the flux melts, it dissolves the oxide on the aluminum sheet surfaces to create a nascent surface that the f ller metal can now wet. After the filler melts and forms joints, the component is cooled to about 500°F (260°C) or lower before it exits the furnace.
Several researchers have recognized and acknowledged that the effectiveness of the NOCOLOKTM flux can be adversely affected by Mg in the filler metal or the core alloy.
There is considerable debate as to what the maximum level of Mg that can be tolerated in the core alloy. Although NOCOLOKTM brazing sheets can contain Mg in the core, the filler alloys have always been Mg-free. The industry consensus has been that core alloy Mg levels below about 0.3% can be brazed with a flux coverage of 3 to 5 g/m2. Yamaguchi found a linear relationship between increasing core Mg level and the flux loading required to counter the poisoning effect. A similar relationship was also found by Wong. Flux poisoning refers to a reaction that occurs between the KA1F4-K3A1F6 and Mg. The resulting product is MgF2, a very stable compound that prevents the NOCOLOKTM flux from dissolving the flux. On the surface, increasing the flux loading to counter increasing core alloy Mg levels appears to be an effortless way to solve the problem at hand. However, most of the commercial heat exchanger producers have found that increasing the flux loading in excess of Sg/m2 results in significantly higher production costs.
Although raising the Mg content in the core alloy raises the strength, when the NOCOLOKTM process was originally adopted, users of the process quickly learned that Mg containing alloys resulted in a flux penalty. As such, brazing sheet materials were typically comprised of Mg-free 4XX~~ filler alloys clad to 3003 (Mg-free) core alloys.
From Table 1 it is clear that raising the Mg level is desirable to increase the post-braze strength of the core alloy. Increasing the core alloy strength translates to thinner wall structures, lighter assemblies, and a potential costs savings since less material would be needed to make a heat exchanger. Since industry has not been using 0.5% Mg containing alloys, it is reasonable to assume that the flux cost and other processing penalties probably exceed the value of higher strength core alloys.
Table 1. Typical post-braze tensile strength levels in commonly available brazing sheet core alloys.
Alloy (Kaiser Version)AA3003 (K383)AA3003+0.3% Mg (K384)AA3005 (K366) UTS, ksi (MPA) 20 (138) 22 24 (166) YS, ksi (MPA) 8 (55) 8.4 10 (69) Elongation, % 18 21 17 Kaiser designation for the equivalent alloy processed by the long life practice.
As noted earlier, there are other issues aside from the cost penalty such as damage to the furnace by excess flux dripping off the pants and eroding the stainless steel retort, human exposure to the flux (considered to be a skin and eye irritant), and effluent disposal. These factors add to the cost of increased flux usage.
Over the past six or seven years, heat exchanger makers have been seeking ways to increase the core alloy strength without increasing the flux loading. In the mid-1990's, Kaiser Aluminum & Chemical Corporation developed a filler metal that was alloyed with a small, but controlled Li addition. This Li-containing filler metal allowed high Mg core alloys (i.e., 0.5% Mg or greater) to be NOCOLOKTM brazed without a flux penalty. This achievement resulted in an incremental gain for heat exchanger makers. However, the current thrust is to develop a fluxless brazing process (alloy, surface treatment, etc.).
SUMMARY OF THE INVENTION
The present invention is a product and process for fluxless brazing of aluminum materials in an inert, non-oxidizing atmosphere. The product is an aluminum filler metal that contains silicon and sodium. In addition, bismuth and/or potassium may be added with the sodium. The present brazing process includes brazing under controlled atomospheric conditions. Preferably, the brazing occurs in an inert atmosphere containing high levels of nitrogen or argon. The intentional addition of sodium or sodium in combination with potassium and/or bismuth results in a fluxless brazing with core alloys having a greater amount of magnesium. The increased levels of magnesium in the core alloy yield an increased post-brazing strength of the core alloy.
Several advantages are obtained by producing the present filler material. The present invention is a truly fluxless aluminum brazing process that does not require a vacuum furnace, NOCOLOKTM (fluoride flux) flux or other costly, unique capital equipment. By controlling the levels of sodium in the filler material, large fillet sizes rivaling the conventional NOCOLOKTM fluxed and brazed material are obtained. The filler material can be clad to core alloys preferably from the 3XX~~, 5, or 6XX~~ alloys described by the Aluminum Association, for example, a 3005-type alloy preferably containing about 0.5%
magnesium. Increasing the amount of magnesium in the core alloy results in an increased post-braze strength of the core alloy. The quality of the atmosphere in the furnace is important to obtaining the above identified advantages. Acceptable joints are formed when the inert gas contains preferably less than about 100 ppm of oxygen. Such preferred inert atmospheres include nitrogen or argon gases.
More specifically, the present invention comprises a filler alloy and a brazing process in a controlled, inert environment using the filler alloy, where the filler is comprised of about 4 to 20 wt% silicon and intentional additions of sodium, or sodium and bismuth, or sodium and potassium, or sodium, bismuth and potassium. More preferably, the levels of sodium range from about 0.0008 to 0.06 wt.%, potassium levels range from about 0.0005 to 0.03 wt. %, and bismuth levels range from about 0.03 to 0.133 wt.%. The balance of the material contains aluminum and incidental impurities, for example, iron, copper, and magnesium up to about 1 wt.% and manganese up to about 1.5 wt.%.
BRIEF DESCRIPTION OF THE DRAWINGS
The features and advantages of the present invention will become apparent from the following detailed description of a number of preferred embodiments thereof, taken in conjunction with the accompanying drawings, in which:
FIGS. 1A-1 C are sketches of a mini-radiator test sample used to evaluate the influence of the subject invention clad metal alloy compositions on the tube-to-header fillet size;
FIGS. 2A-2C are sketches of a "box-ring" sample used to evaluate the effect of sample geometry on the fillet size of the clad alloys constructed in accordance with the preferred embodiments of the present invention;
FIG. 3 is a graph of the average fillet areas of the fluxless brazing process of the present invention in a nitrogen atmosphere using a K366 (0.5% Mg) core with 4045 filler alloys containing sodium or sodium and bismuth, with the materials being cleaned by wiping with acetone;
FIG. 4 is a graph of the average f llet areas of the fluxless brazing process of the present invention in a nitrogen atmosphere using a K366 (0.5% Mg) core with 4045 filler alloys containing sodium or sodium and bismuth, with the materials being cleaned with a Trenton aqueous cleaner;
FIG. 5 is a graph of the average fillet areas of the fluxless brazing process of the present invention in a nitrogen atmosphere using a K366 (0.5% Mg) core with 4045 filler alloys containing sodium or sodium and bismuth, with the being cleaned by vapor degreasing in 1-1-1 trichloromethane; and FIG. 6 is a graph showing the comparison of the brazing filler alloys of the present invention with a number of control alloys.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
As has been discussed, the present invention relates generally to the brazing of aluminum articles in which sodium is intentionally added to an aluminum filler to facilitate fluxless brazing. The addition of sodium alone or in combination with potassium and/or bismuth to the filler material allows increased levels of magnesium to be introduced into the core alloy resulting in a core alloy of higher strength. The brazing is performed in a controlled environment having an inert atmosphere. Preferably, the controlled environment contains high levels of nitrogen or argon.
For the purposes of this invention, and as used hereinafter, the term "controlled atmosphere brazing" or "CAB" refers to a brazing process which utilizes an inert atmosphere, for example, nitrogen or argon in the brazing of aluminum alloy articles.
"Core" means an aluminum alloy which is the structural support for the aluminum alloy that is used as the filler. "Filler" means an aluminum alloy which is used to braze the core or other aluminum articles. "Cladding" is used to describe the use of the filler when it is overlaid on one or both surfaces of the core. Thereafter, the clad core is called a composite or a brazing sheet. "Fillet"
means a concave junction between two surfaces.
The presently preferred filler alloy contains aluminum as a major constituent, and also contains silicon and sodium. Sodium is intentionally added to achieve the desired beneficial properties, and may also be added to the alloy in combination with bismuth and/or potassium.
The addition of bismuth and/or potassium tightens the performance variation or improves overall performance of the alloy.
The preferred alloy has the above elements in the following ranges as measured by weight percent (wt. %). The amount of silicon in the filler alloy is from about 4% to 20%.
The preferred level of sodium is from about 0.0008% to 0.06%, with a more preferred range from about 0.005% to 0.05%, and a most preferred range from about 0.006% to 0.03%. If added to the alloy, the preferred level of bismuth is from about 0.03% to 0.133%, with a more preferred range from about 0.03% to 0.1%, and a most preferred range from about 0.03% to 0.08%. The preferred level of potassium, again if added to the alloy, is from 0.0005% to 0.03%, having a more preferred range from about 0.001% to 0.03%. The balance of the filler alloy contains aluminum and incidental impurities. The incidental impurities may comprise iron, copper, magnesium and manganese. The weight percentages for iron, copper and magnesium are typically below 1 %, and usually are below 0.3%. In the case of magnesium, the level must be kept below about 0.1%. The amount of manganese can be up to about 1.5%
without adversely affecting the brazeability of the alloy.
The filler alloy is generally employed in the form of a brazing sheet rolled from ingots having the desired alloy composition. The filler is applied to the surface of the aluminum core alloy through cladding regardless of the brazing process. Cladding of the aluminum core alloy with the filler is accomplished by methods well-known in the art, for example by pressure welding through a rolling process. Preferably, the material is partially annealed in dry nitrogen. After rolling and partial annealing, the sheets are cleaned and assembled into desired parts, for example, mini-radiators samples of the type illustrated in FIGS. lA-C.
FIG. 1A shows a top view of the pre-braze sample and FIG. 1B shows a side view of the pre-y braze sample. FIG. 1 C shows the post-braze sample. The parts are then brazed in a furnace containing an inert gas, preferably nitrogen or argon. The preferred incoming gas has about 100 ppm of oxygen or less.
As will be demonstrated by the following examples, sodium alone or sodium combined with potassium and/or bismuth, when added in specific amounts to the base aluminum alloy filler material, have been found to increase the strength of the alloys and increase the fillet sizes compared with materials not containing these elements. The intentional addition of these elements allows brazing to occur without the need of flux.
EXAMPLE PREPARATIONS
example Preparation for Examples 1-4.
To test the properties of alloys formed in accordance with the present invention, fifteen alloys were cast for brazing trials. The filler alloys, based on 4045, were alloyed with Na or Na combined with Bi and/or K. Different levels of these alloying elements were examined to better define the maximum operational window. A list of the alloys is shown in Table 2.
Table 2. Sample aluminum filler metals for fluxless brazing.
ID Si Fe Cu Mn M~ Na K Bi 025 9.53 0.02 0.004 0.003 0.003 0.036 054 9.14 0.02 0.004 0.003 0.003 0.011 055 10.56 0.03 0.005 0.024 0.001 0.050 056 11.32 0.03 0.003 0.010 0.000 0.040 057 9.71 0.02 0.001 0.003 0.014 0.026 058 9.14 0.02 0.001 0.002 0.001 0.050 0.065 059 9.62 0.02 0.001 0.001 0.000 0.019 0.109 060 9.51 0.02 0.001 0.002 0.000 0.029 0.127 061 8.82 0.03 0.004 0.020 0.035 0.040 0.080 062 9.10 0.02 0.001 0.002 0.000 0.010 0.127 063 9.55 0.02 0.001 0.002 0.000 0.026 0.274 064 9.31 0.02 0.003 0.002 0.000 0.015 0.161 065 9.17 0.02 0.002 0.002 0.000 0.035 0.182 066 9.30 0.02 0.001 0.002 0.001 0.040 0.002 070 9.42 0.02 0.001 0.002 0.000 0.050 0.026 0.160 AA4343 6.8-8.2 0.8 0.25 0.10 0.04 --- --- ---max max max max AA4045 9.0-11.00.8 0.30 0.05 0.05 --- --- ---max max max max ~ From the Aluminum Association: Each and Others are 0.05% max. and O.lSmax., respectively.
The filler alloys were clad, i.e. roll bonded, to a K366 core alloy. This alloy is a 3005-type alloy containing about 0.5% Mg and processed to develop long life corrosion performance. This core alloy was selected to show that the new filler metals could be fluxless brazed while clad to a high strength core alloy. The compositional range for the K366 core alloy is described in Table 3.
Table 3. Compositional range of the K366 and K:i83 core alloys.
Element Si Fe Cu Mn M Cr Ti Each Others K366 0.10 0.25 0.28-1.00- 0.39-0.05 0.05 0.05 0.15 max. max. 0.38 1.30 0.51 max. max. max. max.

K383 0.10 0.3 0.20-1.00- 0.05 0.05 0.05 0.05 0.15 max. max. 0.40 1.30 max. max. max. max. max.

All core-clad composites (hereafter, referred to as materials) were rolled into sheets that are 0.013 inches (0.330 mm) thick. The clad layer (i.e., filler metal) was 10% of the total composite sheet thickness, i.e. 0.0013 inches (0.033 mm) tluclc. All materials were partial annealed to the H24 temper in dry nitrogen.
After rolling and partial annealing, the sheet samples were cleaned in different commercial cleaners and assembled into the mini-radiator samples depicted in FIG. 1. The mini-radiators approximate commercially produced radiator or heater core joints. All samples were brazed in a furnace with a nitrogen cover gas. The preferred incoming gas had about 100 ppm of Oz or less.
All of the sample materials were compared to conventionally NOCOLOKTM brazed materials that do not contain substantial Mg in the core or the filler metal.
This material was selected since it developed the largest possible fillets with flux coverages ranging from 3 g/mz to 5 g/m2. This range of flux coverage is used by most commercial heat exchanger manufacturers today.

A commercially made brazing sheet material was used as the control material.
The control material was a long life material with a core alloy designated K383 and filler alloy of AA4343. A detailed description of the core and filler alloy compositions is shown in Tables 2 and 3. The control material was 0.0125" (0.3175 mm) thick with the liner comprising 10%
of the total sheet thickness. Normally, the control material would be coated with 3 to 5 g/m2 of NOCOLOK flux, then brazed in a furnace with an inert atmosphere. However, the control material was compared to the new materials without the benefit of flux. The control material was given the same preparation and cleaning as the new materials.
Example Preraaration for Example 5.
A second type of test specimens, box rings, was used to determine the influence of inert gas furnace atmosphere on the quality of the braze joints. FIGS. 2A-C
show the box ring samples made with the brazing sheet. FIG. 2A shows the top view, FIG. 2B
shows the end view, and FIG. 2C shows the side view. The box ring samples permit evaluation of two joint geometries with one test specimen. In this test, the quality of the two joint geometries is visually rated and compared to a standard. A test specimen made with the sodium containing aluminum alloy filler material (material no. 025) was acetone cleaned and brazed without flux. Several test specimens were made and brazed in furnace atmospheres consisting of increasing oxygen levels in the nitrogen, as shown in Table 4.
Table 4. Calibrated nitrogen gas oxygen levels.
Run N en Level m Nitro en Level .

1 50 Balance 2 60 Balance 3 70 Balance 4 80 Balance 5 150 Balance After brazing, the samples were visually rated and compared to the test specimen brazed in nitrogen with 50 ppm of oxygen. The rating system was from 1 to 5. A
rating of 5 was assigned to the samples brazed in the nitrogen with a 50 ppm of oxygen.
The ratings were:
5 = Large fillets, no stitching (skips) on the upright and tangential rings.

4 = Smaller fillets than a 5 rating, slightly irregular, but no stitching.
3 = Small fillets, no stitching.
2 = Small fillets, stitching apparent on either the upright or tangential rings.
1 = No visible fillet on the upright or tangential rings.
Control material coated with 1 g/m2 of flux was also tested, assembled into a box ring test specimen, and brazed in the atmospheres described in Table 4.
Duplicate test specimens were brazed for each nitrogen-oxygen mixture. The duplicate test specimens were not brazed at the same time, but in a separate trial.
All mini-radiator and box ring test specimens were brazed using the braze cycle described below.
a. Begin at 288°C (550°F), the sample enters the furnace at room temperature.
b. Once the sample has been sealed into the furnace, the chamher is evacuated to remove the lab atmosphere. Alternately, the furnace atmosphere can be purged with nitrogen before beginning the brazing thermal cycle. If the furnace is evacuated, the cycle takes two minutes. Purging requires that the equivalent of three to four furnace chamber volume exchanges occur before beginning the braze cycle.
c. Bleed in N2, with an OZ level below 100 ppm and a dew point below -45°C.
The gas flow rate was set at 7 liters/minute throughout the entire brazing cycle.
d. Once the sample has reached 163°C (325°F), the heating cycle was begun.
e. The test specimens typically required 22 minutes to reach 590°C
(1095°F).
~ Once the test specimen reached 590°C (1095°F), heating was terminated.
g. The test specimen temperature will coast up to 599°C (1110°F) in approximately 5 minutes.
h. When the test specimen temperature reached 599°C (1110°F), the gas flow was increased to its maximum to cool the test specimen.
i. Although the gas flow rate has been increased, the test specimen will continue to gain heat until it reaches approximately 604°C (1120°F), at which time the gas begins to cool the test specimen. The total soak time in the temperature range or 599°C (1110°F) to 604°C (1120°F) is about 4 minutes.

j. Once the test specimen has cooled to 571°C (1060°F) or lower, it is removed and air cooled.
After brazing, the test specimens were evaluated by microsectioning, mounting, and polishing through the centerline of the tube sections. The fillet areas of the tube-to-header samples were examined on an optical metallograph and measured with quantitative image analysis software.
EXAMPLES
Three of the subject Na-containing materials were examined in detail, i.e., materials identified as 025, 055, and 058 from Table 2. Also included in the test matrix was the control material. The control material was given cleaning treatments identical to those applied to the Na-containing materials. After cleaning, all materials were brazed under the same conditions and compared. Other materials were examined, but using only one cleaning technique, i.e., the Trenton cleaner. These materials were identified as 063, 066 and 070.
The testing revealed unexpected results, i.e., it is possible to braze aluminum alloys in a nitrogen atmosphere furnace without a flux. The performance of the new fluxless alloys are described in the examples shown below.
Test specimens made from the control material were also given identical cleaning treatments and brazed without flux. This was done to determine if airy of the cleaners had an unexpected effect on the brazeability of the new materials. If an effect was observed, it would not be possible to positively determine whether the new alloys are a success. Fillet areas of the new materials were compared to conventionally NOCOLOKTM fluxed and brazed test specimens. Two large data bases, developed in 1993 and 1998, containing information generated using the mini-radiator test specimen served as a baseline to judge the performance of the new materials. In order for the new materials to be considered a success, the brazing performance of the new materials must be similar to the incumbent NOCOLOKTM
flux brazing process.
After brazing the test specimens, the samples were metallographically examined to measure the fillet sizes. Typically, 16 tube-to-header joints were examined.
The data plotted in FIGS. 3-6 show the average fillet areas as a symbol. Error bars representing a 95%
confidence interval are shown either side of the average. This convention was used for all data developed on the sample materials.

Example 1 In Example 1, the sample materials were cleaned by wiping their surfaces with acetone only. The purpose of the acetone wipe was to establish baseline brazing performance of the sample materials. Acetone, as a cleaner was selected since it easily removes most common oils and greases from aluminum sheet surfaces without chemically reacting or interacting with the aluminum substrate. After acetone cleaning, the mini-radiator specimens were placed into a furnace normally used for control atmosphere brazing (CAB) of specimens coated with NOCOLOKTM flux. The test specimens were brazed by using the fiunace thermal cycle described earlier. None of the acetone cleaned test specimens were NOCOLOKTM fluxed. As we can see in FIG. 3, the sample materials developed fillet areas that were equivalent to the control materials treated with NOCOLOKTM flux.
However, the control material that was not fluxed and acetone wiped did not form any fillets.
xam 1e 2 Another, commonly used, commercially available aqueous cleaner is made by Trenton. Trenton cleaner is a commercial caustic cleaner which contains NaOH.
Sample materials were cleaned in the Trenton cleaner and were fluxless brazed as was done in Examples 1. The data plotted in FIG. 4 indicates that the Trenton cleaner had a beneficial effect on the sample materials. However, the control material cleaned in the Trenton cleaner did not produce any fillets. Cleaning the sample materials in the Trenton cleaner produced fillet areas that were about the same size as the control materials that were NOCOLOKTM
fluxed and brazed.
The results of this cleaner suggest that the cleaner can have an influence on the performance of the subject fluxless brazing materials.
Example 3 In the eaxly days of fluxless brazing in a vacuum furnace, vapor degreasing was used to clean the aluminum brazing materials used then. Today, vapor degreasing is no longer in use due to its adverse effects on the ozone layer. The new fluxless brazing materials were cleaned by vapor degreasing and fluxless brazed in a nitrogen atmosphere, as done in the previous examples. The results shown in FIG. 5 indicate that if all the oils are removed from the surfaces of the subject sample materials, that the samples develop fillet areas that are equal to those formed when NOCOLOKTM fluxing and brazing.

Example 4 Based on the work described in Examples 1-3, the Trenton cleaner in combination with the sodium containing aluminum fluxless brazing materials developed the largest fillets.
A larger number of the material compositions described in Table 2 were examined. All materials were cleaned in the Trenton cleaner and fluxless brazed. The data plotted in FIG. 6 shows the fillet areas developed in the mini-radiator test specimens after fluxless brazing. It was found that all the new materials do not develop acceptable fillet sizes relative to the NOCOLOKTM fluxed control materials. However, there were some materials that developed acceptable fillet sizes relative to the control material treated with NOCOLOKTM flux.
The data shown in FIG. 6 clearly shows that increasing the Bi to fairly high levels results in increasingly small fillets. FIG. 6 also shows that the combination of Na and K
result in the largest mini-radiator fillet areas. However, the effect of high levels of K and Bi combined with Na results in small fillet areas. The data plotted in FIG. 6 indicates that each element must be maintained in a critical range, individually or in combination, to develop fillet sizes rival NOCOLOKTM fluxed and brazed materials.
xam 1e 5 Table 5 ranks the sensitivity of the Na-containing materials to the quality of the atmosphere.
Table 5. Effect of increasing oxygen levels in nitrogen on the brazeability of the new Na-containing materials Run No. Oxygen Nitrogen Na-Only Control Level, ppm Level Rating with 1 g/mz flux Ratin 1 50 Balance 5 5 2 60 Balance 4 5 3 70 Balance 3 5 4 80 Balance 3 3 5 150 Balance 1 2 Rating system:
5 = Large fillets, no stitching (skips) on the upright and tangential rings.

4 = Smaller fillets than a 5 rating, slightly irregular, but no stitching.
3 = Small fillets, no stitching.
2 = Small fillets, stitching apparent on either the upright or tangential rings.
1 = No visible fillet on the upright or tangential rings.
It was determined that adding any of the highly reactive elements like Na to a filler metal dramatically increases need for tight control of the furnace atmosphere.
When the furnace atmosphere oxygen level exceeds about 80 ppm, the brazeability of the subject Na-containing material is severely affected. By comparison, the control material coated with as little as 1 g/m2 of NOCOLOI~TM flux produced acceptable fillets in nitrogen with 80 ppm of oxygen present. When the oxygen levels rose to 150 ppm, the Na-containing materials failed to produce adequate fillets. The examples and results show that the Na-containing materials can be successfully brazed without flux, but only when the furnace atmosphere is of high quality. Acceptable joints can be formed when the inert gas contains less than about 100 ppm of oxygen.
Although the present invention has been disclosed in terms of a number of preferred embodiments, it will be understood that numerous modifications and variations could be made thereto without departing from the scope of the invention as defined by the following claims:

Claims (28)

Claims What is claimed is:

1. A process for controlled atmosphere brazing comprising, brazing an aluminum alloy without flux in a controlled atmosphere, using a filler alloy comprising, a) 4-20 wt.% silicon;
b) 0.0008-0.06 wt.% sodium; and c) the balance aluminum and incidental impurities.

2. The process for controlled atmosphere brazing of claim 1, wherein said filler alloy comprises 0.005-0.05 wt. % sodium.

3. The process for controlled atmosphere brazing of claim 2, wherein said filler alloy comprises 0.006-0.03 wt. % sodium.

4. The process for controlled atmosphere brazing of claim 1, wherein said filler alloy further comprises at least one element selected from the group comprising 0.0005-0.03 wt.
potassium and 0.03-0.133 wt. % bismuth.

5. The process for controlled atmosphere brazing of claim 4, wherein said filler alloy comprises both 0.0005-0.03 wt. % potassium and 0.03-0.133 wt. % bismuth.

6. The process for controlled atmosphere brazing of claim 4, wherein said filler alloy comprises 0.005-0.05 wt. % sodium.

7. The process for controlled atmosphere brazing of claim 6, wherein said filler alloy comprises 0.006-0.03 wt. % sodium.

8. The process for controlled atmosphere brazing of claim 4, wherein said filler alloy comprises 0.001-0.03 wt. % potassium.

9. The process for controlled atmosphere brazing of claim 4, wherein said filler alloy comprises 0.03-0.1 wt. % bismuth.

10. The process for controlled atmosphere brazing of claim 9, wherein said filler alloy comprises 0.03-0.08 wt. % bismuth.

11. The process for controlled atmosphere brazing of claim 1, further comprising, providing a core alloy upon which said filler alloy is clad.

12. The process for controlled atmosphere brazing of claim 11, wherein said core alloy is selected from the group consisting of 3XXX, 5XXX and 6XXX series alloys.

13. The process for controlled atmosphere brazing of claim 1, wherein said controlled atmosphere is a non-oxidizing gas.

14. The process for controlled atmosphere brazing of claim 13, wherein said non-oxidizing gas is nitrogen or argon.

15. The process for controlled atmosphere brazing of claim 13, wherein said non-oxidizing gas contains less than 100 ppm of oxygen.

16. The process for controlled atmosphere brazing of claim 1, wherein said incidental impurities include no more than 0.1 wt% magnesium and not more than 1.5 wt%
manganese.

17. The process of claim 1, further including the steps of providing an aluminum article to be brazed and cleaning said article with a caustic NaOh containing cleaner prior to brazing said aluminum article with said filler alloy.

18. An aluminum filler alloy comprising, a) 4-20 wt.% silicon;
b) 0.0008-0.06 wt.% sodium, and c) the balance aluminum and incidental impurities.

19. The aluminum filler alloy of claim 18, wherein said alloy comprises 0.005-0.05 wt. % sodium.

20. The aluminum filler alloy of claim 19, wherein said alloy comprises 0.006-0.03 wt. % sodium.

21. The aluminum filler alloy of claim 18, wherein said alloy further comprises at least one element selected from the group comprising 0.0005-0.03 wt. %
potassium and 0.03-0.133 wt. % bismuth.

22. The aluminum filler alloy of claim 21, wherein said alloy comprises 0.005-0.05 wt. % sodium.

23. The aluminum filler alloy of claim 22, wherein said alloy comprises 0.006-0.03 wt. % sodium.

24. The aluminum filler alloy of claim 21, wherein said alloy comprises both 0.0005-0.03 wt. % potassium and 0.03-0.133 wt. % bismuth.

25. The aluminum filler alloy of claim 21, wherein said alloy comprises 0.001-0.03 wt. % potassium.

26. The aluminum filler alloy of claim 21, wherein said alloy comprises 0.03-0.1 wt.
bismuth.

27. The aluminum filler alloy of claim 26, wherein said alloy comprises 0.03-0.08 wt.
bismuth.

28. The aluminum filler alloy of claim 18, wherein said incidental impurities include not more than 0.1 wt% magnesium and not more than 1.5 wt% manganese.