SHF – Microfluidics 2006 - Toulouse, 12-14 December 2006 – G.P. Celata, M. Cumo, S.J. McPhail, G. Zummo – Single-Phase Laminar & Turbulent Heat Transfer in Smooth & Rough Microtubes SINGLE-PHASE LAMINAR & TURBULENT HEAT TRANSFER IN SMOOTH & ROUGH MICROTUBES G.P. Celata* + , M. Cumo°, S.J. McPhail*, G. Zummo* celata@casaccia.enea.it, maurizio.cumo@uniroma1.it, stephen.mcphail@casaccia.enea.it, zummo@casaccia.enea.it * ENEA, Institute for Thermal Fluid Dynamics Via Anguillarese 301, 00060 S.M. Galeria, Rome, Italy ° University of Rome La Sapienza Corso Vittorio Emanuele II, 244, Rome, Italy + (corresponding author) KEYWORDS Microtubes, local Nusselt number, wall axial conduction, peripheral heat loss ABSTRACT In the attempt to characterize the exact behaviour and the possible deviations from classical theory of microscale thermal fluid dynamics, an experimental campaign was carried out studying laminar and turbulent heat transfer in uniformly heated smooth glass and rough stainless steel microtubes from 0.5 mm down to 0.12 mm. Heat transfer in turbulent regime proved to be coherent – within the range of experimental accuracy – with the classic Gnielinski correlation for the Nusselt number. For the laminar case, an anomalous drop in Nusselt number for decreasing Reynolds number was observed in the smooth glass tubes. As the stainless steel tubes manifested relatively normal diabatic behaviour in this regime (apart from the evident influence of the thermal development region that increases heat transfer above the thermally fully developed value), the explanation of this unexpected diminution of the Nusselt number must be sought in the dispersion of heat, put in externally through the thin film deposited on the glass tube outer surface, to peripheral attachments to the test section. This distorts the measured energy balance of the experiment, especially as the convective force of the fluid diminishes, resulting in lower Nusselt numbers at lower Reynolds numbers. 1. INTRODUCTION The miniaturization of many appliances in biomedic, chemical and computer technology has brought with it increased demands for space-efficient high-performance heat dissipation and catalytic devices. Though much research on microscale level has already been done in recent years in the relevant fields, especially as regards hydrodynamic and heat transfer characterization, there is still much diversion of results to be discerned in the various experimental and numerical reports. The diversion occurs in different contexts: fluid drag of laminar, transient and turbulent single-phase flows, heat transfer of liquid and gas flows, and two-phase flow in adiabatic and heated microchannels; each with a vast class of problems relating to fluid type, phase or compressibility, channel shape and aspect ratio, surface properties and diabatic conditions. In the numerous experimental and theoretical investigations done over the past years, a number of contradictory conclusions have been drawn, but lately most inconsistencies seem proven to originate from the difficulty of experimental conditions and high sensitivity of models to measurement errors. One of the first studies on microscale heat transfer was done by [Wu & Little, 1984], where Nitrogen flowing through rectangular channels of 134 to 164 μm hydraulic diameter was investigated, both in laminar as in turbulent conditions. The authors found Nusselt numbers to be higher than predicted by conventional theory, and dependent on channel roughness. A new correlation was proposed with a stronger dependence on Reynolds number. [Peng & Peterson, 1996] also worked on steel rectangular channels, using water as a working fluid, and found a dependence of the Nusselt number on Reynolds and Prandtl numbers even in laminar flow. The overall magnitude was lower than theoretical predictions however.