Thermal test chip

There are three indirect, nondestructive measurement techniques to determine thermal resistance acoustic microimaging, x-ray, and thermal test chips. The acoustic microimaging and x-ray methods can be used in both development and production, but the thermal test chip is restricted to development. The first two indirect thermal techniques find the amount of voiding in the thermal path and, through the use of thermal modeling, calculate the thermal resistance. The thermal test chip can only find the thermal resistance capability of the physical design. 

The thermal resistance 0 of the microelectronic assembly is obtained in an indirect manner. The test assembly is the same as that of the actual assembly witii one exception — the active device is replaced with the thermal test chip. Measuring the jimction temperature of the test chip, the case temperature of the test DUT, and the power dissipation of the resistive heater, tiie thermal resistance of the test assembly is calculated with the formula ... 

If the thermal test chip were the same size as the actual active device and mounted in the test assembly with the same die-attach material, then the thermal resistance of the test assembly would be the same as the actual assembly. However, the chances of the thermal test die being the same size as the actual die are remote. Therefore, to find the thermal resistance for the actual assembly, the thermal resistance of the die (0oie) arid the die attach ( Die Att) need to be scaled. The equation for the die thermal resistance is... 

EIA/JESD51-4, E.J.S., Thermal Test Chip Guideline (Wire Bond-Type Chip), Eebruary, 1997. 

More sophisticated, multisensor test chips contain test structures for the evaluation of thermal resistance, electrical performance, corrosion, and thermomechanical stress effects of packaging technologies (Mathuna, 1992). However, multisensor chips have to be driven by computerized test equipment(Alderman, Tustaniwskyi, and Usell, 1986 Boudreaux et al., 1991) in order to power the chip in a manner that... 

Oettinger, EE 1984. Thermal evaluation of VLSI packages using test chips—A critical review. Solid State Tech. (Feb) 169-179. 

Curves of Figure 19 compare the data published for (a) boron nitride [37,40] (b) aluminium (c) diamond-[37-39] (d) aluminium nitride [37-42] (e) crystalline silica. It can be seen that, at 45 vol.%, the maximum thermal conductivity achieved with diamond powder is 1.5 W m K, while crystalline boron nitride at 35 vol.% affords 2.0Wm K. The thermal conductivity of silver-filled adhesives was studied by using silicon test chips attached to copper and molybdenum substrates [43]. The authors outline the importance of the shape factor A, related to the aspect ratio of the particles, to achieve the highest level of thermal conductivity. Another study reports the variation of the effective thermal resistance, between a test chip and the chip carrier, in relation to the volume fraction of silver and the thickness of the bond layer [44]. The ultimate value of bulk thermal conductivity is 2 W m at 25 vol.% silver. However, the effective thermal conductivity, calculated from the thermal resistance measurements, is only one-fifth of the bulk value when the silicon chip is bonded to a copper substrate. 

FIG. 9 Solder joint resistance over time for several thermal gradients created by varying test chip power dissipation and substrate temperature. (Courtesy of IBM Corp.)... 

Tests at the boiling point should be conducted with minimum possible heat input, ana boiling chips should be used to avoid excessive turbulence and bubble impingement. In tests conducted below the boiling point, thermal convection generally is the only source of liquid velocity. In test solutions of high viscosities, supplemental controlled stirring with a maraetic stirrer is recommended. 

---