Nanofluid Properties and Their Effects on Convective Heat Transfer in an Electronics Cooling Application

Abstract

In the search for new, more effective coolant fluids, nanoparticle suspensions have shown promise due to their enhanced thermal conductivity. However, there is a concomitant increase in the viscosity, requiring an increase in pumping power to achieve the same flow rate. Studies of flow cooling in simple geometries indicate that there is a benefit to using nanofluids, but it is difficult to justify extending these results to the far more complicated geometries. Moreover, with the variability of property measurements found in literature, it is possible to show conflicting results from the same set of flow-cooling data. In this work we present a self-contained study of the properties and effectiveness of an alumina in water nanofluid. Flow-cooling is studied in an off-the-shelf fluid cooling package for electronics to examine the effects of the particulates in a practical scenario. We measure the thermal conductivity and viscosity of the same suspensions to assure consistent interpretation of our results. We find that, while there is no anomalous enhancement of the thermal properties or transport, there is a benefit to using a low volume fraction alumina nanoparticle suspension over using the base fluid alone. In fact, there is an optimal volume fraction (1%) for this nanofluid and electronics cooling system combination that maximizes the heat dissipated. However, we find that this benefit decreases as the volume fraction, and hence the viscosity, increases. Understanding where the trade-off between viscosity increase and thermal conductivity increase occurs is critical to designing an electronics cooling system using a nanofluid as a coolant.

Shelf stability of nanofluids and its effect on thermal conductivity and viscosity

“…Zeta potential is an electrophoretic property of heterogeneous fluids and is defined as the potential difference between the dispersed particles surface and the layered fluid attached to the particles. If the zeta potential value is below −30 mV or above +30 mV, the suspensions are considered more stable due to strong repulsive force between the charged nanoparticles [10][11][12][13]. This can be obtained with high surface charge density of nanoparticles.…”

Group analysis and numerical computation of magneto-convective non-Newtonian nanofluid slip flow from a permeable stretching sheet

“…), applications have significantly diversified in the past decade. New areas in which nanofluids have been deployed include bioconvection in microbial fuel cells , electronics cooling (Townsend and Christianson 2009), building physics and contamination control (Kulkarni et al 2009), pharmacological administration mechanisms (Mody et al 2013), peristaltic pumps for diabetic treatments and solar collectors (Rana et al 2013a). Experimental works have been complemented in recent years by numerous theoretical and computational simulations.…”

Experimental Investigation on the Thermal and Hydraulic Performance of Alumina–Water Nanofluids in Single-Phase Liquid-Cooled Cold Plates

“…In short, there is some difference between this model and the Maxwell model, albeit small. In Table 2, thermal conductivity values of select concentrations of alumina-water nanofluids calculated using the Maxwell model are reported and compared to published experimental Table 1 The density of alumina-water nanofluids (measured and calculated) measurements reported by Williams et al [17] and Townsend and Christianson [18]. The difference between the experimental measurements and Maxwell theory is also listed in Table 2 and considered acceptable agreement; note also that the INPBE study showed that the more detailed model proposed by Nan et al [14] only agreed with the measured thermal conductivity to an accuracy of 18% [10].…”

“…b Measured values for the baseline fluid (water) were compared to NIST data for temperatures between 20 and 80 C. c Values from Eq. (3) were compared to published experimental measurements reported by Williams et al[17] and Townsend and Christianson[18]. d Values from Eq.…”