Effect of Forming Temperature and Sheet Temper on the Brazing Characteristics of Aluminum Alloy Sheet

Loading...
Thumbnail Image

Date

2018-09-20

Authors

Benoit, Michael

Advisor

Wells, Mary
Hansson, Carolyn

Journal Title

Journal ISSN

Volume Title

Publisher

University of Waterloo

Abstract

Automotive heat exchangers are fabricated by forming and brazing of multi-layered aluminum (Al) alloy sheets. The Al brazing sheets are comprised of two alloy layers: an AA3xxx core, which provides strength to the assembly, and an AA4xxx clad, which melts during brazing to provide filler metal for joints throughout the assembly. Warm forming has recently proven to be a promising technique to expand heat exchanger design possibilities, by increasing the material forming limits, and by enabling the use of higher strength materials, by reducing springback after forming. However, no consideration had been given to the effect of warm forming on downstream brazing and corrosion performance. The objective of the current research is to understand the effect of forming temperature and initial sheet condition on the brazing performance of Al brazing sheets. The Al brazing sheet used throughout the current work was industrially produced, with an overall thickness of 200 µm, and a single AA4045 clad layer comprising 10 % of the sheet thickness. The sheets were supplied in both the fully annealed (O) and the work hardened (H24) sheet tempers. Warm forming was initially simulated by performing interrupted tensile tests between room temperature (RT) and 250 °C, up to pre-determined levels of strain between 2 % and 12 %, at an average engineering strain rate of 6.6x10-4 s-1. The rate of liquid clad alloy depletion during simulated brazing was measured with differential scanning calorimetry, using a parameter referred to as the liquid duration time (LDT). A small LDT, caused by rapid depletion of the liquid clad due to penetration into the core alloy, was predicted to result in poor brazing performance. The LDT for the O-RT forming condition decreased from 44.2 min when no strain was applied to the sheet, to a minimum value of 29.7 min at 4 % strain, before increasing with the further application of strain. The LDT data were correlated with the post-braze sheet microstructures: when the LDT was decreasing, the core alloy was non-recrystallized and the phenomenon of liquid film migration (LFM) occurred during brazing, while for conditions where the LDT was increasing, the core alloy was characterized by coarse, recrystallized grains without LFM. The trend in the LDT data was in excellent agreement with prior studies where LFM had been observed, indicating the suitability of the LDT as a predictor of brazing performance. When the forming temperature was increased to 250 °C, the LDT decreased from 42 min at 0 % strain, to a minimum value of 26 min at 8 % strain, but did not increase at greater applied strains. The change in the LDT data after warm forming was attributed to an increased range of strains over which LFM occurred. Thus, brazing of O temper sheet formed at 250 °C was predicted to be impaired relative to RT formed sheet. Conversely, the LDT for H24 sheet was found to be independent of both applied strain and forming temperature, and a recrystallized core alloy was observed in all cases. While some dynamic recovery is believed to have occurred during warm forming, the H24 core alloy hardness was still in the same order of magnitude as RT formed sheet, so core alloy recrystallization could still occur. Consequently, brazing of H24 temper sheet was predicted to be insensitive to forming temperature. The difference in brazing characteristics of O-RT and O-250 °C conditions was confirmed from sagging distance experiments, again using warm formed tensile coupons. Maximum sagging of O-RT sheet occurred at 4 % strain, while complete rigidity was observed at higher strains. For O-250 °C samples, the sagging distance remained elevated between 2 % to 12 % strain. Similar to conditions with a low LDT, LFM was observed in the post-braze micrographs of forming conditions with large sagging distances (i.e. O-RT-4 % and O-250 °C-10 %), and transmission electron microscopy revealed a recovered sub-structure in front of the LFM grains. The sagging distance as a function of strain for O-150 °C samples was close to that of the RT formed sheet, which indicated that formability improvements could be achieved at this temperature, without altering the brazing characteristics of the sheet. The brazing predictions made using the LDT and sagging distance data were tested by brazing of scaled-down electric vehicle battery cooling plates, which were formed from both O and H24 sheet tempers, between RT and 250 °C. Simulated brazing of single formed plates revealed that the microstructures within the plate were in good agreement with the results from the simplified tensile test specimens, at comparable levels of strain. Formed plates of the same forming conditions were then brazed together, to create functional cooling plates. In all cases, plates were successfully brazed, and were capable of withstanding an applied internal pressure of 0.28 MPa. Furthermore, no obvious difference in the brazing performance was found between the various sheet temper-forming temperature combinations at the component-level, and warm forming was shown to not adversely impact the ability to form brazed joints. The LDT and sagging distance measurements taken from strained sheet specimens were shown to be inadequate to predict the brazing performance of warm formed O temper sheet in assemblies more representative of heat exchangers, since these metrics did not account for wetting and capillary flow of the liquid clad alloy. Microstructure analysis confirmed that the microstructures were similar to the warm formed tensile specimens, although certain microstructures not present in the tensile specimens were also observed, such as strain induced boundary migration in the O-250 °C condition. However, the strain rate in the plates was estimated to be in the order of 1.0x10-1 s-1, orders of magnitude higher than the tensile specimens, and the plates experienced a significantly higher local strain (25 %) at the location in question. Additional tensile tests performed up to 20 % strain at 150 °C and 250 °C, using strain rates between 6.6x10-4 s-1 and 6.6x10-2 s-1, revealed a dependence of the post-braze microstructure on the strain rate, due to increased strain rate sensitivity at elevated temperatures, and similar microstructures as observed in the plates were found for comparable strains and strain rates. It is concluded that warm forming, used to improve formability of Al brazing sheet, does not impair brazing performance. Brazing predictors, such as the LDT and sagging distance, are useful for studying interactions occurring within the sheet during brazing, but do not account for liquid clad flow, which is a major factor in brazed joint formation in real components. The microstructure evolution of O temper sheet during brazing depends on applied strain, strain rate, and forming temperature. The change in microstructure with changes in process variables also supports the deformation energy driving force for the LFM phenomenon. The H24 sheet was found to be insensitive to an increase in forming temperature, in terms of the post-braze microstructure, LDT, and ability to braze real components. It is recommended that the potential of the warm forming process be further investigated by forming full-scale components, and forming at higher temperatures to further improve forming limits and springback reduction. Finally, the relative corrosion performance of the different sheet tempers and forming temperatures must be more thoroughly investigated.

Description

Keywords

Brazing, Warm forming, Corrosion, Aluminum alloys, Automotive, Liquid film migration, Recrystallization, Strain induced boundary migration

LC Keywords

Citation