Method for eliminating runout of braze filler metal during active brazing

Abstract

Nanometers thick conformal coatings deposited by atomic-layer deposition (ALD) onto the metal surface of an active braze joint modifies the surface chemistry to eliminate excess braze filler metal flow. Unlike other means used to prevent braze filler metal runout, the thin ALD coating does not hinder next assembly processes, does not require post-braze cleaning, and does not alter the base material mechanical properties.

Claims

1. A method for eliminating excess runout of braze filler metal during active brazing, comprising: providing a metal object having a faying surface; applying a thin conformal coating of a metal oxide to the metal faying surface; providing a ceramic object having a faying surface; placing a braze filler metal having an active element between the metal-oxide-coated metal faying surface and the ceramic faying surface to form contacted surfaces; and heating the contacted surfaces to above the melt temperature of the braze filler metal to actively braze the faying surfaces and form a brazed joint, wherein the metal oxide coating prevents runout of the braze filler metal beyond the faying surface of the metal object.

2. The method of claim 1, wherein the thin conformal coating is applied by atomic-layer deposition.

3. The method of claim 1, wherein the metal object comprises iron, nickel, cobalt, or alloys thereof.

4. The method of claim 1, wherein the metal oxide comprises Al.sub.2O.sub.3.

5. The method of claim 1, wherein the braze filler metal comprises the active element and a metal alloy.

6. The method of claim 5, wherein the active element comprises zirconium, titanium, or hafnium.

7. The method of claim 5, wherein the metal alloy comprises copper and silver.

8. The method of claim 1, wherein the ceramic object comprises alumina or aluminum nitride.

9. The method of claim 1, wherein the thin conformal coating has a thickness of greater than 1 nm.

10. The method of claim 9, wherein the thin conformal coating has thickness less than about 20 nm.

11. A method for eliminating excess runout of braze filler metal during active brazing, comprising: providing an alumina object having a faying surface; applying a thin conformal coating of a metal oxide to the alumina faying surface; providing a metal object having a faying surface; placing a braze filler metal having an active element between the metal-oxide-coated alumina faying surface and the metal faying surface to form contacted surfaces; and heating the contacted surfaces to above the melt temperature of the braze filler metal to actively braze the faying surfaces and form a brazed joint, wherein the metal oxide coating prevents the active element in the braze filler metal from reacting with the alumina object, thereby preventing runout of the braze filler metal beyond the faying surface of the metal object.

12. The method of claim 11, wherein the thin conformal coating is applied by atomic-layer deposition.

13. The method of claim 11, wherein the metal object comprises iron, nickel, cobalt, or alloys thereof.

14. The method of claim 11, wherein the metal oxide comprises ZrO.sub.2, TiO.sub.2, or HfO.sub.2.

15. The method of claim 11, wherein the thin conformal coating has thickness of greater than 1 nm.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The detailed description will refer to the following drawings, wherein like elements are referred to by like numbers.

(2) FIG. 1 is a photograph of bare Kovar lids that are actively brazed to Diamonite cylinders.

(3) FIG. 2A shows the chemical reaction for depositing Al.sub.2O.sub.3 by a conventional chemical vapor deposition (CVD) process and the binary reaction sequence for depositing Al.sub.2O.sub.3 by an atomic-layer deposition (ALD) process. FIG. 2B shows the binary pulse-purge ALD process. FIG. 2C illustrates the surface reactions of the ALD binary sequence.

(4) FIG. 3A is a side-view photograph of a “tensile button” test sample comprising an oversized Kovar interlayer sandwiched between two Diamonite ceramic buttons before tensile testing. FIG. 3B is a photograph of the fracture surface of a bare Kovar test sample after testing. FIG. 3C is a photograph of the fracture surface of an ALD-coated test sample after testing.

(5) FIG. 4 is a graph of showing the percent of samples showing active braze filler metal runout as a function of Kovar interlayer surface treatment.

(6) FIG. 5 is a graph of average joint tensile strength for a variety of Kovar interlayer surface modifications tested.

(7) FIG. 6A is a cross-sectional scanning electron microscope (SEM) image of a bare Kovar braze joint. FIG. 6B is a cross-sectional SEM image of a ALD Al.sub.2O.sub.3-coated Kovar braze joint.

(8) FIG. 7A is an SEM image of the Kovar interface of a Diamonite/97Ag-1Cu-2Zr braze filler metal/10 nm ALD Al.sub.2O.sub.3-coated Kovar braze joint. FIG. 7B is a Ag electron dispersive spectroscopy (EDS) map of the interface. FIG. 7C is an Al EDS map of the interface. FIG. 7D is a Zr EDS map of the interface.

(9) FIG. 8A is an SEM image of the Diamonite interface of a Diamonite/97Ag-1Cu-2Zr braze filler metal/10 nm ALD Al.sub.2O.sub.3-coated Kovar braze joint. FIG. 8B is a Ag EDS map of the interface. FIG. 8C is an Al EDS map of the interface. FIG. 8D is a Zr EDS map of the interface.

(10) FIG. 9A is an SEM image of the Kovar interface of a Diamonite/97Ag-1Cu-2Zr braze filler metal/bare Kovar braze joint. FIG. 9B is a Ag EDS map of the interface. FIG. 9C is an Al EDS map of the interface. FIG. 9D is a Zr EDS map of the interface.

(11) FIG. 10A is an SEM image of the Diamonite interface of a Diamonite/97Ag-1Cu-2Zr braze filler metal/bare Kovar braze joint. FIG. 10B is a Ag EDS map of the interface. FIG. 10C is an Al EDS map of the interface. FIG. 10D is a Zr EDS map of the interface.

DETAILED DESCRIPTION OF THE INVENTION

(12) According to the present invention, a solution to the problem of runout is to mask the metal surface with a thin barrier layer coating. The barrier layer preferably has the properties of conforming to the features of the metal faying surface, limiting surface wetting or spread, not hindering the braze joint performance (i.e., the joint should provide a hermetic seal and have high mechanical strength), remaining intact after brazing, and not compromising subsequent assembly functions.

(13) More particularly, the invention is directed to a method to eliminate runout by applying a very thin conformal coating to the faying surface of the metal object using atomic-layer deposition (ALD). The ALD process inhibits deleterious reactions that occur when the active element (e.g., zirconium, titanium, or hafnium) in the liquid active brazing filler metal reacts with the ceramic and metal faying surfaces. Additionally, unlike other braze-stop materials applied to metal surfaces to prevent excess wetting and spread of the liquid filler metal, this conformal coating adheres completely to the metal surface and is thin enough that it can remain in and not interfere with subsequent joining processes, such as welding, and can be considered particulate free.

(14) ALD is a chemical vapor deposition (CVD) process that uses self-limiting surface reactions applied in a binary sequence, leading to atomic-layer controlled growth of a thin film. ALD is similar to conventional CVD, except that the ALD process breaks the CVD reaction into two half-reactions, keeping the precursor materials separate during the half-reactions. This is accomplished through sequential pulsing of special precursor vapors, each of which forms about one atomic layer on the surface during each reaction cycle, versus CVD that introduces multiple precursor materials simultaneously. Each half-reaction involves the reaction between a gas-phase precursor and a surface function group. The surface reaction continues until all functional groups are consumed and replaced with a new functional group which does not react with that gas-phase precursor. This self-limiting surface chemistry enables conformal coating of surfaces with monolayer precision. Application of the binary sequence ABAB . . . results in layer-by-layer highly conformal film growth.

(15) The ALD thin film can be grown using a chemical reactor. See J. W. Elam et al., Rev. Sci. Instrum. 73(8), 2981 (2002), which is incorporated herein by reference. The reactor uses a gas pulse switching method that allows introduction of gaseous and high vapor pressure liquid precursors into the reactor. Precursors can be introduced into the hot-wall reactor using N.sub.2 as a carrier gas. Exposure times for the A and B precursors and purge times between A and B pulses can be independently set. Deposition rate can be measured using an in-situ quartz crystal microbalance. Typically, the deposition rate per cycle is constant within a run and very reproducible from run to run.

(16) A binary reaction sequence for deposition of Al.sub.2O.sub.3 using trimethyl aluminum (TMA) and water as precursors is shown in FIG. 2A, where surface species are indicated by asterisks. As shown in FIGS. 2B and 2C, in the “A” flow, TMA is pulsed into the reaction chamber and reacts with surface hydroxyl groups, forming Al—OAl(CH.sub.3).sub.2 on the surface and evolving methane. TMA will continue reacting with the surface until all of the surface hydroxyl groups are consumed. The reaction chamber is then purged of the first reactants. Next, water vapor is pulsed into the chamber and reacts with the surface dimethyl aluminum groups, desorbing methane and once again forming a surface hydroxyl group. The chamber is then purged of the second reactants and the sequence is repeated. This reaction cycle is repeated until the desired film thickness is achieved. For the specific process used in the examples below to grow aluminum oxide films on Kovar by ALD, one complete cycle occurred every 32 seconds with a film thickness deposition rate of 1.1 Å (0.11 nm) per cycle.

(17) To evaluate mechanical strength of the braze joint, “tensile button” test samples were prepared according to the ASTM-F19 standard test method. See ASTM International, ASTM F19-11(2016), “Standard Test Method for Tension and Vacuum Testing Metallized Ceramic Seals,” West Conshohocken, Pa. This test method establishes the standard procedures for conducting tension and vacuum tests on metal-ceramic seals to determine the bond strength of brazed, metallized ceramics.

(18) FIG. 3A is a photograph of a “tensile button” test sample comprising an oversized Kovar (Fe-29Ni-17Co) metal spacer sandwiched between two Diamonite alumina ceramic buttons by the active braze alloy 97Ag-1Cu-2Zr. The ceramic tensile buttons were cleaned prior to brazing using a 3-step solvent cleaning process followed by resintering at high temperature. The resintering process was accomplished in either wet hydrogen (28° C. dew point) at 1500° C. for 60 minutes, or air-fired at 1575° C. for 120 minutes. The resintering atmosphere appeared to have minimal impact on the excess flow/runout response, hermeticity, or joint strength. The 1.25 mm thick Kovar interlayers and 50 μm thick Ag-1Cu-2Zr braze metal preforms where also cleaned using the 3-step solvent cleaning process. Additionally, the Kovar substrates were subjected to a pickling acid etch process followed by a bright-dip, followed by rinsing in deionized water. The Kovar surfaces were either uncoated (“bare”) or ALD coated with 5-20 nm Al.sub.2O.sub.3. The samples were brazed in a high-vacuum furnace according to the following temperature cycle: 15° C./min from ambient to 925° C., soak 15 minutes; 10° C./min from 925° C./to 985° C., soak 5 minutes; 25° C./min from 985° C. to 925° C., soak 0 minutes; Furnace cool (uncontrolled) to ambient.

(19) This Kovar-silver-based braze metal-alumina combination frequently results in excess flow of the braze filler metal. The oversized Kovar interlayer allows the runout, excessive wetting, and unwanted spread of the braze filler metal to be easily evaluated. FIG. 3B shows the fracture surface of a test sample that did not have an ALD coating on the faying surfaces of the Kovar interlayer after tensile testing. As can be seen in this figure, excess braze flow was observed when sealing with the bare Kovar interlayer only. Conversely, as seen in FIG. 3C, no excess braze flow was observed when the Kovar surface was coated with a 10 nm ALD layer prior to brazing.

(20) FIG. 4 is a graph showing the percent of samples showing runout as a function of Kovar interlayer surface treatment. Some of the Kovar surfaces were dimpled with small mounds or bumps, approximately 25 μm tall, on the faying surface. Dimples can be effective in maintaining braze joint thickness uniformity, which can improve joint tensile strength. All 35 of the samples that were not treated with the ALD Al.sub.2O.sub.3 layer (“bare” Kovar) exhibited runout, both with and without dimples. Conversely, all samples in which the Kovar surfaces were ALD coated with 5-20 nm Al.sub.2O.sub.3 prior to brazing exhibited no runout, irrespective of the alumina ceramic or resintering atmosphere used. Thus, braze filler metal runout was completely eliminated with a 5 nm ALD alumina coating (some runout was observed with very thin, 1 nm thickness, ALD coatings). A preferred upper end of the thickness range is dependent on post-processing and joint strength requirements. All of the ALD-coated samples exhibited a hermetic seal (<5E-12 atm-cc/sec He leak rate). These results clearly demonstrate the effectiveness of using thin film coatings of ALD alumina to eliminate excessive braze filler metal flow for this system of materials whilst maintaining hermiticity.

(21) The test buttons were also subjected to tensile testing to determine the strength of the joints with and without the ALD coatings. FIG. 5 is a graph of tensile strength at failure for a variety of Kovar modifications tested. The Diamonite tensile button assemblies incorporating undimpled and uncoated Kovar interlayers had tensile strengths slightly higher, on average, than undimpled 10 nm thick ALD Al.sub.2O.sub.3, 94 Mpa (13.5 ksi) versus 83 MPa (12.0 ksi) with standard deviations of 17 MPa and 19 MPa respectively. For the brazed Diamonite tensile buttons incorporating dimples, however, the results were nearly equal for the uncoated interlayers and 10 nm thick ALD Al.sub.2O.sub.3 coated interlayers: 100 MPa (14.5 ksi) average tensile strength with a standard deviation of 14 MPa for the uncoated Kovar versus 102 MPa (14.8 ksi) with a standard deviation of 25 MPa for the 10 nm ALD coated samples. Thus, there was minimal loss in joint strength with the ALD Al.sub.2O.sub.3 coating as compared to a bare Kovar joint.

(22) FIG. 6A shows a cross-sectional scanning electron microscope (SEM) image of a bare Kovar braze joint, showing excessive reaction length and filler metal runout on the unbrazed Kovar surface. FIG. 6B is a cross-sectional SEM image of a ALD Al.sub.2O.sub.3-coated Kovar braze joint, showing much reduced reaction length and runout on the Al.sub.2O.sub.3-coated Kovar surface. It is apparent from this figure that the coverage of the Al.sub.2O.sub.3 coating should be sufficient such that the liquid braze metal does not make direct contact to or wet any bare Kovar metal.

(23) SEM images and electron dispersive spectroscopy (EDS) maps were obtained to show the reactions at the braze joint interfaces for ALD Al.sub.2O.sub.3 coated Kovar and bare Kovar braze joints. FIGS. 7A-D show SEM and EDS maps of a 10 nm ALD Al.sub.2O.sub.3-coated Kovar braze joint at the Kovar—braze filler metal interface. FIGS. 8A-D show SEM and EDS maps of the 10 nm ALD Al.sub.2O.sub.3-coated Kovar braze joint at the Diamonite—braze filler metal interface. FIGS. 9A-D show SEM and EDS maps of a bare Kovar braze joint at the Kovar—braze filler metal interface. FIGS. 10A-D show SEM and EDS maps of the bare Kovar braze joint at the Diamonite—braze filler metal interface. All of the Zr EDS maps (FIGS. 7D, 8D, 9D, and 10D) show a continuous band of zirconium at both the Kovar and Diamonite interfaces. These continuous layers confirm that there is adequate active element material, Zr, to bond to two individual oxide faying surfaces. Recall the redox reaction mentioned previously, 3Zr+2Al.sub.2O.sub.3.Math.4Al+3ZrO.sub.2. In this reaction, the alumina ceramic, Diamonite, is reduced by the atomic zirconium in the liquid braze filler metal, and the “free” atomic aluminum is brought into the liquid braze filler metal as a result. This highly reactive aluminum in the liquid filler metal is then free to react with exposed surfaces on the Kovar interlayer, forming aluminide compounds (FIG. 9C). These aluminide reactions drive the excessive wetting and spreading behavior on the bare Kovar interlayer surface surrounding the braze joint. In samples having thin ALD Al.sub.2O.sub.3 layers deposited on the Kovar interlayers, the same interfacial reactions occur at the two faying surfaces with the zirconium in the liquid braze filler metal. In these samples, however, the alumina is reduced at both interfaces and the atomic aluminum also mixes with the liquid braze filler metal. However, at this point the aluminum is unable to react further as no “bare” Kovar surfaces are exposed, thus no runout occurs.

(24) Alternatively, rather than coating the metal faying surface with Al.sub.2O.sub.3, the nonmetal (i.e., alumina) faying surface can be coated with ZrO.sub.2, TiO.sub.2, or HfO.sub.2 to act as a barrier layer to prevent the Zr in the liquid braze filler metal from reacting with the alumina ceramic. ZrO.sub.2, TiO.sub.2, and HfO.sub.2 can be deposited by ALD. By doing this, the metal substrate can remain coating-free, if the designer chooses to do so. This enables the designer to choose the substrate—metal or nonmetal—that is most prudent to receive the ALD coating. Because metallurgical analysis performed to date has not revealed evidence of excessive wetting/flow due to “free” Zr in the molten filler metal (as described above has been observed with “free” Al), the filler metal will remain aluminum free.

(25) The present invention has been described as a method for eliminating runout of braze filler metal during active brazing. It will be understood that the above description is merely illustrative of the applications of the principles of the present invention, the scope of which is to be determined by the claims viewed in light of the specification. Other variants and modifications of the invention will be apparent to those of skill in the art.