The present invention relates to a composite solder paste that includes a mixture of a solder or solder-forming powder and a Ni-containing reinforcement powder to produce a high melting point solder joint using liquid phase diffusion bonding and having a Ni-stabilized hexagonal (Cu,Ni)6Sn5 phase as the matrix phase that bonds together the reinforcement powder.
The development of high-temperature lead-free solders is becoming increasingly crucial for use in power electronics in the automotive, aerospace, military, and energy production industries [references 2,5]. The currently used typical high-temperature solders are Pb-5Sn (wt. %) or Pb-10Sn (wt. %) that melt around 300° C. A typical solder reflow cycle will have a peak reflow temperature 20° C. to 40° C. above the melting point of the solder; therefore, the current typical high temperature solder has a reflow temperature around 330° C. [references 8,13]. The operating temperatures of these solders are usually less than 250° C., but it is critical that the solder be able to survive hierarchical solder reflow throughout all stages of lower temperature reflow processing [reference 2].
With today's increased power pumping through electronics, higher operating temperatures can contribute to brittleness in solder joints by several mechanisms upon thermal cycling [references 8,10,11]. In the past, this cracking did not occur because the microstructures formed by lead and tin did not consist of any brittle intermetallic compounds, but now Pb use is becoming much more restricted due to RoHS [references 2,18,19].
There is a need for a lead-free version of these high-Pb solder alloys that can be used at equal or lower processing temperatures, but can still withstand high operating temperatures. For example, a lead-free version would have a lower melting temperature, so that the reflow temperature could also be lower, but that with the ability to evolve during low temperature reflow to a microstructure with high temperature stability. The lead-free version also would have acceptable mechanical properties that avoid brittle phases, which is a problem for several of the alternative solders, including another type of composite paste approach that uses pure Cu powder as a filler powder [reference 8]. Alloy cost should also be well within reasonable bounds, unlike the Au—Sn solder system that seems to meet some desirable characteristics, but certainly has a cost problem.
The present invention provides a composite solder material for mixing with flux to provide a composite solder paste that meets the above need. The composite solder material includes a mixture of relatively low melting solder or solder-forming powder and a relatively high melting Ni-containing reinforcement powder that provides a source of Ni for incorporation into and stabilization of the high temperature (more ductile) hexagonal (Cu,Ni)6Sn5 phase upon cooling from reflow to room temperatures. The Ni-containing reinforcement powder may be present in the mixture in amounts to improve electrical conductivity across the joint.
The Ni-containing reinforcement powder can be present as a majority (50% by volume or more) of the metallic powders present to maximize conductivity improvement (and impact resistance) if a gas-based (e.g., formic-acid) fluxing approach is used on a blended powder (cold pressed) solder “wafer,” or more preferably as a minority (less than 50% by volume) to significantly reduce or eliminate porosity in the solder joint made with more conventional composite solder (liquid-based) paste material.
The present invention also provides a high melting point solder joint produced using the solder in a liquid phase diffusion bonding process and having the stabilized high temperature hexagonal (Cu,Ni)6Sn5 phase as a matrix phase with improved ductility that bonds together the Ni-containing reinforcement powder particles, which can provide improved electrical conductivity in the solder joint.
In practice of an illustrative embodiment of the present invention, liquid phase diffusion bonding transforms, with each reflow cycle, more of the low melting solder or solder-forming powder to the hexagonal (Cu,Ni)6Sn5 matrix phase, raising the final melting temperatures of the post-processed solder from 227 degrees C. to over 400 degrees C. and giving the solder the ability to withstand higher Joule-heating, all while improving resistance to solder joint cracking by eliminating the crack-promoting low temperature monoclinic Cu6Sn5 phase in the solder joint matrix.
Advantages and further details of the present invention will become more readily apparent from the following detailed description taken with the drawings which follow.
The present invention provides a composite solder material for mixing with flux to provide a composite solder paste. The composite solder material includes a mixture of relatively low melting solder or solder-forming powder and a relatively high melting Ni-containing reinforcement powder. The Ni-containing reinforcement powder provides a source of Ni for incorporation into and stabilization of the high temperature hexagonal (Cu,Ni)6Sn5 phase upon cooling to room temperature from reflow temperature(s).
In an illustrative embodiment of the present invention, the low melting point solder or solder-forming powder can comprise a solder alloy, such as a low melting eutectic or near-eutectic Sn—Cu solder alloy, other solder alloys containing Sn, and/or metallic Sn so as to form the high temperature hexagonal (Cu,Ni)6Sn5 phase under reflow conditions. The selected composition of the solder or metallic Sn takes into account any expected alloy element uptake from the Ni-containing reinforcement powder and/or components being soldered, such as uptake of Cu and/or Ni from PCB pads, etc.
In another illustrative embodiment, the Ni-containing reinforcement powder can comprise alloy powder particles that contain a metal such as one or more of Cu, Co, and others alloyed with Ni, provided the reinforcement powder has a higher melting point than the solder or solder-forming powder. The reinforcement powder preferably comprises Cu—Ni alloy powder. A particular embodiment of the present invention employs Cu-5-15 weight % Ni alloy powder particles to this end. During reflow conditions, liquid phase diffusion bonding between the two powders results in the Ni stabilizing agent being incorporated in solid solution in the hexagonal (Cu,Ni)6Sn5 phase in an amount to stabilize the phase from reflow to room temperature. The selected composition of the reinforcement powder can take into account effects of any alloy element uptake from the components being soldered.
In still another illustrative embodiment of the present invention, the reinforcement powder is present as about 60 to about 90% by volume of the metallic powders present in the mixture sans flux. Preferably, in this embodiment, the reinforcement powders are present in an amount of about 60 to about 80% by volume of the mixture. An even more preferred amount of reinforcement powder is about 70% by volume of the mixture.
In a further embodiment of the present invention, the Ni-containing reinforcement powder can be present as a minority (less than 50% by volume) of the powder mixture sans flux to significantly reduce or eliminate porosity in the solder joint made with conventional composite solder (liquid-based) paste material. A preferred amount of the reinforcement powder is from about 10% by volume to less than 50% by volume, such as about 10 to about 40 volume % or about 10 to about 30 volume % of the mixture.
In still a further illustrative embodiment of the present invention, the reinforcement powder is employed in a particle size range that is similar to (substantially the same as) or larger than the particle size range of the solder or solder-forming powders in order to minimize particle surface area for wetting by the solder or solder-forming powder and to retain the solid reinforcement powder particles themselves in the final solder joint microstructure.
Practice of the present invention involves producing a high melting point solder joint produced by using the composite solder paste in a reflow liquid phase diffusion bonding process and having the stabilized high temperature hexagonal (Cu,Ni)6Sn5 phase as the matrix phase that bonds together the Ni-containing reinforcement powder particles, which may provide a high electrical conductivity network in the solder joint. The high melting point solder joint is free of the undesirable monoclinic Cu6Sn5 (η′) phase that can cause solder joint cracking upon cooling from the reflow temperature to room temperature as explained below.
In practice of the present invention, liquid phase diffusion bonding transforms, with each reflow cycle, more of the low melting solder or solder-forming powder to the hexagonal (Cu,Ni)6Sn5 matrix phase, raising the final melting temperatures of the post-processed solder from 227 degrees C. to about 400 degrees C. and giving the solder the ability to withstand higher Joule-heating, all while improving resistance to solder joint cracking by eliminating the crack-promoting, low temperature monoclinic Cu6Sn5 (η′) phase in the solder joint matrix at room temperature.
A series of high temperature (330° C.) reflow tests that were coupled to extended thermal aging at 250° C. were conducted with several types of Cu-alloy substrates in contact with a single type of Sn—Cu—Ni solder; namely, Nihon Superior's SN100C (Sn-0.65% Cu-0.05% Ni-60 ppmw Ge solder % in weight %) available from Nihon Superior (NS Building, 1-16-15 Esaka-Cho, Suita City, Osaka 564-0063, Japan) (see
The model joint sample microstructures shown in
To explain in more detail,
With respect to phase stabilization, on the Cu—Sn phase diagram, Cu6Sn5 is seen existing as two allotropic phases. One allotrope is the high temperature hexagonal Cu6Sn5 (q) and the other is the low temperature monoclinic Cu6Sn5 (η′). The fact that the interfacial (Cu)6Sn5 (η′) IMC forms during reflow without excessive porosity shows that there is adequate wetting and bonding of all of the substrate alloys by the solder alloy, but the tendency for cracking of the resulting IMC in the non-Ni solder joints of
The following Detailed Examples are offered to further illustrate the present invention without limiting the scope of the invention.
These Examples illustrate blending together two types of Ni-containing powder, gas-atomized Cu-10Ni (wt. %) powder particles and atomized solder powder of Nihon Superior's SN100C (Sn-0.65% Cu-0.05% Ni-60 ppmw Ge) to form a composite solder blend.
Before going through the compounding of a full paste sample, the intended dual-powder composite model was tested as a porous compact that was infiltrated by a typical flux described below for the SN100C solder and reflowed to form simulated test joints. These test joints were useful for microstructure evaluation to study the effects of varying reflow time and temperature on the resulting joint microstructure. The test joints were to demonstrate that the composite solder joint microstructure could benefit from nickel additions through the phenomena of liquid-phase diffusion bonding (LPDB). The LPDB process takes advantage of the relatively low liquidus temperature of the SN100C solder powder combined with the relatively high diffusion kinetics of the Ni addition in both powders to rapidly form ductile intermetallic compound [(Cu,Ni)6Sn5] that would serve as a strong, crack-free matrix phase for the high electrical conductivity of the resulting Cu-10Ni particle network.
Model composite paste solder joints were produced with a variety of blends of gas-atomized Cu-10Ni (wt %) powder and Sn-0.7% Cu-0.05% Ni+Ge, Nihon Superior SN100C (wt. %) solder powder, according to the combinations listed in
The flux-soaked powder compacts were then tested using the setup illustrated in
At random, of the six (6) reflow combinations, four were repeated three times and two were repeated twice. At least sixteen micrographs of the same magnification were taken and analyzed for each tested sample. After analysis, the best ratio of powders was determined to be 70 volume percent Cu-10Ni with 30 volume percent SN100C. This combination resulted in the best continuous Cu—Ni conductive network while not sacrificing the SN100C needed to wet the Cu-10Ni particles together to form the desired Ni-stabilized hexagonal (Cu,Ni)6Sn5 IMC matrix phase. Therefore, the following results are for this particular model containing 70 volume percent Cu-10Ni.
The image analysis of sixteen micrographs per sample gave the averages shown in
Within this composite solder mixture of high-melting Cu—Ni powder and low melting SN100C powder, a process that can be termed “liquid phase diffusion bonding” (LPDB) occurs, where the SN100C matrix alloy interdiffuses with and consumes the surface of the solid Cu—Ni reinforcement powder above the matrix liquidus temperature. The resulting alloy that comprises the majority of the matrix is hexagonal Cu5.5Ni0.5Sn5, which has a higher melting temperature than the original matrix of SN100C.
For this composite solder technology, the term liquid-phase diffusion bonding or LPDB more accurately represents the phenomenon occurring within this composite microstructure between the melted solder alloy matrix and the slowly dissolving Cu-10Ni powders. Thus, a major portion of the bonding process actually involves rapid volume diffusion through the resulting intermetallic compound (IMC) phase that forms immediately upon melting of the solder alloy, rather than by the surface diffusion mechanism usually associated with solid-state sintering. It is quite interesting (see
An innovative advantage of the composite solder paste system is that with each reflow cycle, more and more of the SN100C solder alloy transforms into the hexagonal (Cu,Ni)6Sn5 IMC, raising the final melting temperature of the system post-processing from 227° C. to about 400° C. and giving the solder the ability to withstand higher Joule heating. Also, it is apparent from the microstructures of
Advantages, in addition to a lower reflow temperature, a shorter reflow time, a higher resulting melting temperature, and it being a lead-free/environmentally-friendly high temperature solder, are that the thermal expansion of the solder and most ceramic substrates are more equally matched and that the strength of the solder joint is theoretically increased. Estimates of the thermal expansion mismatch show a mismatch of 20-25 ppm/° C. between Pb-5Sn and alumina and a mismatch of only 5-10 ppm/° C. between Cu-10Ni and alumina [references 9,17,19]. Estimates of shear modulus for Pb-5Sn and Cu-10Ni show values on the order of 5 GPa and 50 GPa [reference 17], respectively [reference 15]. These estimated values are based on a “rule of mixtures” approach and remain to be experimentally tested.
The following examples were conducted to determine the effect of the volume fraction and particle size range of the Ni-containing powder (i.e. Cu-10 weight %) on solder joint porosity and solder joint composition. The performance of the solder pastes was determined by experimentation with reflow cycles on a hot plate using solder paste samples. The test setup used a boron-nitride coated stainless steel plate placed on a resistance-heated hot plate. A copper plate (substrate) was clamped by a stainless steel clamp fixture on the stainless steel plate to optimize the heating of the solder paste on the copper plate substrate. The setup was tested with various peak reflow temperatures along with peak reflow times described below.
Both single-sided and double-sided joints with Cu substrates were formed for analysis. In the case of the single-sided joints, the solder paste was spread on the copper plate substrate. In the case of double-sided joints, the experimental procedure was the same except that a second copper plate was placed on top of the solder paste residing on the lower copper plate.
In initial testing, a double-sided joint was made from composite solder paste that was formulated from the powder blend of 70 vol. % Cu-10Ni and 30 vol. % SN100C solder alloy mixed with flux of the type described above (J-STD 002C RMA type of flux). This was given a reflow profile of 10 seconds at the peak temperature of 250° C.
Next, a double-sided joint was made from a solder paste of 24 vol. % Cu-10Ni:76 vol. % SN100C powders and the flux with a reflow of 60 seconds at the peak temperature of 255° C. This second double-sided joint test using the solder paste of 24 vol. % Cu-10Ni:76 vol. % SN100C resulted in less solder joint porosity. The solder joint microstructure produced after reflow is seen in
A re-test of the solder paste of 70 vol. % Cu-10Ni and 30 vol. % SN100C mixed with the flux but using a reflow profile (250° C. for 60 seconds) resulted in a large amount of residual flux residue remaining trapped in the solder joint. An EDS image (elemental mapping) showed large amounts of residual carbon, apparently from flux residue that prevented the molten Sn alloy powders from coalescing after melting and, thus, prevented successful joint formation. From this result, it was determined that there wasn't a viable path for the gaseous decomposition byproducts of the flux to escape during reflow due to the effect of an elevated paste viscosity from the continuous network of Cu-10Ni powder (at 70 vol. %) that blocked the passageways. Other micrographs of the same joint (not shown) indicated that porosity (spherical, gas-type bubbles) was also significant in this joint, as another indication of trapped thermal decomposition byproducts of the residual flux residue left in the reflowed joint.
As a result, three additional solder paste samples were blended to contain 13 vol. % Cu-10Ni, 26 vol. % Cu-10Ni, and 36 vol. % Cu-10Ni (the balance being the SN100C solder alloy powder) and mixed with the flux. The three pastes were each run with a reflow profile of 15, 30, and 60 seconds at the peak reflow temperature of 255° C. to make single-sided solder joints with which to test the effect of the powder blend ratio and time on joint porosity; namely, a higher vol. % of SN100C powder compared to Cu-10Ni powder in the composite paste mixture was hoped to have an effect on reducing the paste viscosity during reflow and, thus, the amount of retained porosity due to trapped flux residue after reflow. This effect could be achieved by a lowered composite paste viscosity by increasing the Sn alloy content since it would be liquid above 227° C., allowing for easier pathways for the flux residue to escape. This lower viscosity was intended to prevent gas-bubble-related voiding.
As just mentioned, three different powder blend ratios were used:
These powder blends correspond to 41 wt. % Cu-10Ni:59 wt. % SN100C; 30 wt. % Cu-10Ni:70 wt. % SN100C; and 16 wt. % Cu-10Ni:84 wt. % SN100C, respectively.
From analysis of the collection of SEM (montage) micrographs (not shown) of the solder joint microstructures produced in this way using a reflow profile of 255° C. for 60 seconds, it was concluded that increasing the amount of Sn-containing powders (SN100C) in the blend ratio helps to decrease overall porosity, a more important characteristic of joint quality that results from the reflowed paste, similar to
Solder joint microstructures made using the three solder pastes described above but using 15 seconds of reflow were evaluated. SEM images indicated the average area percent of each phase within the microstructure for all times measured with each vol. % of Cu-10Ni. For example, for 36 volume % Cu-10Ni, the IMC area (IMC matrix phase) of the joint was 45.7% of the total joint area after only 15 seconds of reflow. For 26 volume % Cu-10Ni, the IMC area of the joint was 46.2% of the total joint area after only 15 seconds of reflow. For 13 volume % Cu-10Ni, the IMC area of the joint was 31.7% of the total joint area after only 15 seconds of reflow. The total fraction of IMC phase after only 15 seconds of reflow was quite remarkable and was dependent on the amount of available Cu-10Ni interfacial area.
Moreover, as the Sn content of the paste blends was increased (due to less Ni—Cu powder), more IMC was formed with longer holds (e.g. 15 seconds and 60 seconds) at the peak reflow temperature. An increase in IMC was observed clearly, particularly for the samples with 36 vol. % and 13 vol. % of Cu-10Ni (pre-reflow volume % of Cu-10Ni powder).
Another observation found from these tests was that increasing the time of peak reflow does not reduce porosity. With longer reflow times, the solder paste only allows for more diffusion of the IMC, which largely increases the viscosity of the paste as it solidifies. In order for the flux to escape, the solder paste should have a lower viscosity upon initial reflow, which calls for the Cu-10Ni filler phase to be less than 50 vol. %, such from about 10 volume % to less than 50 volume % of the blend sans flux, in spite of the effect on solder joint electrical conductivity. The conductivity difference between the IMC and Cu-10Ni in the literature is actually relatively minor and that of the IMC is still within acceptable values. This lowered viscosity allows the flux to remain on the powders just until the tin melts and wets the solid powders. As soon as the wetting occurs, the residue is free to escape through the liquid tin. Therefore, increasing the amount of liquid phase upon reflow reduces joint porosity.
The composition of each solder joint phase produced using these paste blends stayed consistent with the intended joint composition, no matter what ratio of SN100C to Cu-10Ni was used, implying that the powder ratio does not affect the composition of the resulting phases.
The particle size range of the Cu-10Ni powder was next investigated as to its effect, if any, on the porosity seen in the initial testing described above to produce a double-sided joint from composite solder paste that was formulated from the powder blend of 70 vol. % Cu-10Ni and 30 vol. % SN100C solder alloy and flux given a reflow profile of 10 seconds at the peak temperature of 250° C. This initial blend contained the same size of both powders. The Cu-10Ni powder was of Type 4 (25-32 μm), and the SN100C powder was also of Type 4 (25-32 μm).
In these additional tests, the size of each type of powder was tested for its effect on solder joint porosity of the composite paste. For this tests, the Cu-10Ni as-atomized powder was size classified to 20-38 μm and to 5-20 μm, which essentially matches the Type 4 and Type 6 size classifications for solder powder. Since SN100C powder was provided in both Type 4 (25-38 μm) and Type 6 (5-15 μm), this made it possible to produce paste samples called B, C, and D. Table 1 shows a summary of the different composite paste powder blends tested.
Each powder blend from Table 1 mixed with the above flux as a paste was spread onto a copper plate to form a single-sided solder paste joint wherein the copper plate was held in a stainless steel clamp fixture that was placed on the boron nitride-coated stainless steel plate residing on the hot plate surface to act as a heat sink in order to maintain a consistent hot plate temperature for reflow. A thermocouple was also clamped to the setup to monitor the temperature of the solder during the experimental runs.
The cross-sections of the resulting single-sided solder paste joints for Blends A and B, viewed under a scanning electron microscope (SEM) in backscatter mode indicated that Paste Blend B clearly seemed to produce lower porosity than Paste Blend A. Using quantitative metallography of at least fifteen (15) SEM images of the resulting joint from each reflow run, it was found that of at least fifteen (15) SEM images of the resulting joint from each reflow run, it was found that the images of Paste Blend A contain an average of 3.73 area % porosity and the images of Paste Blend B contain an average of 0.87 area % porosity. The SEM images also revealed a much smoother substrate bond when using the smaller particle size range Type 6 powder than when using the Type 4 powder. In addition, the intermetallic compound or IMC (Cu, Ni)6Sn5 seemed much more dispersed in the Sn matrix phase for the Type 6 powder.
When a double-sided solder paste joint was created using Paste Blend B, SEM images revealed that the interface remained connected and smooth. The porosity with the double-sided joint also remained low, averaging a porosity value of less than 1% of the image area.
These results indicate that the smaller the SN100C powder, the less porous the joint.
Energy dispersive spectroscopy (EDS) was performed on the solder joints in
Thus, in addition to the decrease in joint porosity, the smaller Type 6 powders sizes provided extra benefits in terms of spreading of the intermetallic compound that seems more dispersed in the joint made with the smaller powders. This quicker spread of the IMC should aid in the eventual goal of expanding the IMC across the entire joint interface. When the entire joint consists of a network of the nickel-modified IMC with possibly some residual pockets of Cu—Ni, the strength and ductility of the joint could improve. The composite solder paste joint will then take on the properties of a (Cu,Ni)6Sn5 matrix (Tm˜525° C.)2 containing pockets of a stronger and more conductive Cu—Ni phase.
After these findings, the optimal ratio of Cu-10Ni to SN100C was calculated by determining the matching pre-reflow powder combination composition with that of an entire joint of Cu5.5Ni0.5Sn5, this composition containing 50.12 at. % Cu, 4.58 at. % Ni, and 45.30 at. % Sn. By blending the paste to contain 66 vol. % Sn-alloy powder (SN100C) and 34 vol. % Cu-10Ni powder, the composition of the fully transformed solder joint should be 48.9 at. % Cu, 5.9 at. % Ni, and 45.2 at. % Sn.
From the above Examples, it is apparent that the dual-phase low melting solder blend (SN100C/Cu-10Ni) transforms to a composite solder joint for high temperature use. A preferred embodiment of the lead-free solder contains 70 volume percent Cu-10Ni powder blended with 30 volume percent SN100C powder, if a gas-based fluxing method is used on blended solder wafers to minimize porosity and trapped flux residue. Another preferred embodiment of the lead-free solder contains 13 volume percent Cu-10Ni powder blended with 87 volume percent SN100C powder to reduce or eliminate solder joint porosity. The resulting (Cu,Ni)6Sn5 IMC compositions of this solder system have an approximate 5 at. % content of Ni and showed no cracking due to the added solid solution nickel. There were no significant trends in area fraction or composition of the phases based on time or temperature. Due to this, the present invention provides for a suitable alternative lead-free high temperature solder with improved processing parameters.
Although the above Examples used a particular flux to fabricate the solder joints tested, the present invention is not so limited and envisions use of fluxes other than the flux described above, wherein the alternate flux is more volatile (i.e., is a gas-based flux like formic acid) and/or is more quickly removed during initial reflow such that flux selection alone, volume fraction selection alone, and/or particle size range selection alone, or in combination, provide option(s) that can be used to reduce and/or eliminate solder joint porosity.
The composite paste pursuant to the present invention is designed for rapid insertion into normal PCB assemblies wherein accommodating the switch to the solder paste pursuant to the invention would be nearly negligible in a commercial production setting. In addition, very similar processing parameters of those previously used in industry for Sn—Cu eutectic based paste could still be implemented successfully.
References, which are incorporated herein by reference:
Although certain illustrative embodiments of the present invention are described in detail above, those skilled in the art will recognize that various changes and modifications can be made therein without departing from the scope of the invention as set forth in the appended claims.
This application claims benefits and priority of U.S. provisional application Ser. No. 62/284,487 filed Oct. 1, 2015, the entire disclosure of which is incorporated herein by reference.
This invention was made with government support under Grant No. DE-AC02-07CH11358 awarded by the U.S. Department of Energy. The government has certain rights in the invention.
Number | Date | Country | |
---|---|---|---|
62284487 | Oct 2015 | US |