Fabrication of Microchannel Devices Via Diffusion Bonding and

Fabrication of Microchannel Devices Via Diffusion Bonding and Transient Liquid Phase Bonding Purpose: To reduce the device size of microfluidic devices and heat exchanges to achieve higher efficiencies and better portability. Previous Work: Present Diffusion Bonding Future Transient Liquid Phase Bonding Transient liquid phase (TLP) bonding requires an interlayer between two parent metal shims. Using boron or phosphorus in a nickel interlayer will lower the melting point of the nickel hastening the bonding process, Figure 5. Goals of diffusion bonding: 1. Minimize channel deformation. Materials: Stainless Steel Shims 50 μm and 100 μm thick patterned with microchannels. • Pressure distribution plates • Shims coated with nickel boron or nickel phosphorus • Vacuum Hot Press • Optical Microscope 3. Minimize bonding temperature and pressure, and therefore costs. 4. Maximize bonding efficiency. Present Work: Figure 1: schematic of vacuum hot press illustrating basic configuration of shims and microchannels to bond the device. Q is the heat delivered to the system and F is the force applied by the press. Method: Diffusion bonding and transient liquid phase bonding require pressure and heat Figures 1 and 2. The vacuum hot press can apply several tons force and up to 1200 °C to a stack of shims. Vacuum Chamber and furnace Pressure Ram Figure 2: The vacuum hot press used to apply heat and pressure simultaneously to bond the shims. Purpose of applying pressure and heat: • Promote contact • Increase diffusion speed 1. Figure 3: Diffusion bonded stainless steel shims with channel deformation. Diffusion bonded shims at 1800 psi and 980 °C. The deformation is caused by too much pressure during bonding or cutting and polishing. The present work is on un-coated shims. Each stack was arranged to maximize the vacuum hot press working time. Table 1 illustrates the parameters used for this experiment. Table 1: Diffusion bonding design of experiment used to maximize the information from each run. The two parameters varied per run are the temperature ramp rate and the shim span, which is the distance between channels on the shim. Ramp up rate (°C/min) Shim span (μm) Dwell temperature (°C) Run 1 2 400 Run 2 8 Run 3 Run 4 Special thanks to Steve Leith. Todd Miller, Jack Rundel, Danielle Clair, and Phillip Harding for all of their help and support. CBEE 3. The concentration of suppressant decreases 4. Interlayer solidifies 5. Bond homogenizes Figure 5: The TLP bonding process. The interlayer melts, suppressant diffuses, and the bond solidifies In preparation for future work on this project, modeling has been done for the interlayer thickness vs dwell time, Figure 6. Dwell pressure (psi) Dwell duration (minutes) 120 980 1000 60 100 800 980 1000 60 8 400 980 1000 60 2 800 980 1000 60 The shims are stacked to maximize information from each run. The stack is composed of four distribution plates and two The temperature must remain below thicknesses of shim, 50 μm and 100 μm as the melting point of 316 stainless steel. shown in Figure 4, which will be cut using a wire EDM and inspected. 10 shims 50 μm thick Interlayer melts 2. Melting point suppressant diffuses into the parent metal 1065 °C Dwell time ( Minutes) • 2. Maximize the ratio of good channels to total channels. 1100 °C 80 1150 °C 60 40 20 0 5 shims 50 μm thick 5 shims 100 μm thick 10 shims 100 μm thick End blocks Figure 4: An example of the shim stacks used to maximize data collected per run. 0 20 40 60 80 Interlayer thickness (μm) Figure 6: Interlayer thickness vs dwell time in the vacuum hot press governed by the equations shown. D is the diffusion coefficient, tf is the dwell time, CαL is the critical concentration, Co is the initial concentration, and Cm is the concentration at the interface of melting point suppressant in the interlayer. Thinner interlayers correspond to lower dwell times. Higher temperatures reduce dwell time due to the exponential effect of the temperature on diffusion coefficient.