Characterization of evaporation and boiling from sintered powder wicks fed by capillary action

TL;DR: In this article, the authors measured the dependence of thermal resistance on the thickness and particle size of sintered copper powder wick surfaces, both under evaporation and boiling conditions, and demonstrated that for a given wick thickness, an optimum particle size exists which maximizes the boiling heat transfer coefficient.

About: This article is published in International Journal of Heat and Mass Transfer.The article was published on 2010-09-01 and is currently open access. It has received 251 citations till now. The article focuses on the topics: Boiling & Nucleate boiling.

Summary

1. Introduction

• Heat pipes and vapor chambers are widely employed in electronics cooling and other applications which demand efficient transport or spreading of heat from a localized source of high heat flux.
• These studies also concluded that minimizing the particle layer thickness can reduce the conduction thermal resistance below the evaporating meniscus.
• For applications in which local heat fluxes demand in excess of 500 W cm 2 may be encountered, understanding the incipience of boiling, thermal resistance during boiling, local dryout limitations, and the dependence of these quantities on the wick properties, are important areas for investigation.
• This arrangement subjects the wick to a different fluid feeding mechanism than is encountered in a vapor chamber and the results from this study may thus not be applicable to vapor chambers.
• The effects of the different wick parameters such as pore size, particle diameter and layer thickness for such wicks under conditions of capillary-fed operation are not yet rigorously understood.

2. Experimental facility, sample preparation, and test procedures

• The rationale for the design of the test facility was to simulate the internal conditions of an operating heat pipe evaporator section and supply the working fluid (deionized water) to the porous wick surface in the same manner as occurring within an actual heat pipe.
• The sub-sections below describe the test chamber and liquid flow loop design which reproduces such conditions.

2.1. Test chamber, heater block assembly, and flow loop details

• Detailed diagrams of the test chamber and heater block assembly are shown in Fig. 2(a).
• The test chamber provides a saturated environment into which the sample dissipates heat and also allows for capillary-fed operation of the wick.
• Four T-type thermocouples are inserted into 1.19 mm diameter holes manufactured into the tip at a uniform separation distance of 7.62 mm.
• The outer flow loop which supplies the test chamber with saturated liquid during testing is shown in Fig. 2(c).

2.2. Sample fabrication

• All of the sintered copper powder wicks investigated in this study were manufactured at Thermacore, Inc.
• A layer of copper powder of the desired uniform thickness was sintered on to a square solid copper substrate using a high-temperature mold.
• The range of particle sizes chosen is representative of sintered wicks used in commercial applications.
• The sample is placed face-down into a recessed pocket in the graphite mold and is held off the surface of the mold by stainless steel pins which hold up the four corners of the exposed substrate.
• Before heating the solder joint, the sample surface is purged with argon supplied from a compressed gas line that is fed directly into the recessed pocket.

2.3. Test procedures

• Immediately prior to insertion into the test facility, all soldered samples are treated to remove any possible grease contamination Table 1 List of the sintered copper powder samples tested and their geometric properties.
• For this treatment, the samples are dipped in acetone and thoroughly rinsed with deionized (DI) water to remove any embedded grease.
• This method has been previously used for pre-test treatment of sintered copper powder wick materials [17].
• A steady state is defined to have been reached when the substrate temperature changes at a rate of less than 0.05 C/min time-averaged over a period of 10 min.
• The supplied heat flux is increased in increments to acquire different test data points for developing a boiling curve.

3. Data reduction and measurement uncertainties

• The recorded temperatures are averaged over the final 100 data points (5 min) of the acquired steady-state data.
• Spreading by lateral conduction in the copper substrate prevents further one-dimensional extrapolation of the temperature field to the front surface of the substrate.
• A comprehensive uncertainty analysis is performed and all the reported results include these uncertainty estimates.
• The uncertainty in the calculated heat flux is estimated by evaluating the uncertainty intervals for the slope and intercept of the linear fit to the temperature field in the heater block [18].
• All thermocouples utilize an external ice point reference junction resulting in an estimated temperature measurement uncertainty, UT, of ±0.3 K.

4. Experimental results and discussion

• Test results for all the sintered powder wick samples are presented in the form of boiling curves and the variation of thermal resistance with input heat flux; a representative set of such plots for the 900 lm thick samples is shown in Fig.
• None of the figures presented here indicate that a dryout or critical heat flux condition has been reached at the maximum heat flux shown.
• This is a considerably larger heat flux than has been observed in past studies of monoporous sintered powders and the difference may be attributed, in part, to the short fluid working distance which eliminates any capillary feeding limitations.
• In almost all cases, a slight increase in the resistance is observed at the highest heat fluxes tested.

4.1. Effect of particle size on thermal performance

• The thermal performance of all three 900 lm thickness samples is compared in Fig. 5(b) to distinguish the impact of varying particle size.
• A minimum resistance is exhibited by the intermediate particle size, suggesting that an optimal pore size may exist for maximizing the boiling heat transfer coefficient.
• Each point on this plot corresponds to the averaged resistance over the nucleate boiling regime (200–500 W cm 2) for each sample.
• The trends are unchanged if averaged values over the entire heat flux range are included.
• The sample resistance is seen to decrease when the particle size is decreased from 355 to 106 lm (effective pore diameters from 150 to 45 lm).

4.2. Effect of wick thickness on thermal performance

• The effect of wick thickness on the thermal performance is illustrated by comparing the thermal resistances (averaged over the heat flux range of 200–500 W cm 2) for all three particle diameters in Fig.
• For all particle sizes, the thermal resistance is at a maximum for the median thickness sample and decreases slightly for the 600 and 1200 lm samples.
• It is noted, however, that the average spread in the results over the set of different samples is approximately 21%, which is less than the experimental uncertainty of 25.8% over this heat flux range.
• For only the 45–75 lm particle diameter samples, the spread in thermal resistance with respect to wick thickness is outside the predicted uncertainty, suggesting a possible dependence of the performance on thickness for this particle size.
• Visual observation of the sample surfaces during testing (as discussed in the next section) also confirmed the presence of boiling and bubble departure from the surface of all samples at heat fluxes above 100 W cm 2, independent of thickness.

4.3. Visualization of operating regimes

• It was observed from the thermal measurements in the preceding sections that the overall sample thermal resistance is highest at low heat fluxes (when evaporation is the main mode of two-phase transport) and sharply decreases and to a relatively constant value after the incipience of boiling.
• Detailed visualizations of the surface were recorded for each heat flux to aid in verification of the regime-based mechanisms proposed above.
• The thermal performance results for this test are presented in Fig. 10.
• To improve delineation between the vapor bubble, underlying particles, and fluid surface, a representative sketch of several of the frames is also shown in Fig. 12(b).
• This trend, together with visualization of the wick in these regimes, supports the explanation that this large resistance drop can be attributed to removal of the wick layer conductive resistance once boiling commences.

5. Conclusions

• A novel test facility was developed to characterize the thermal performance of wick structures used in heat pipes and vapor chambers, under operational conditions as seen in the actual devices.
• A parametric investigation of the dependence of the thermal performance on the wick thickness and the particle size was conducted.
• In addition, the surface of the wick was observed with high-speed visualization in situ during the testing.
• (3) A sharp reduction in the overall evaporator thermal resistance is observed corresponding to a transition from the evaporation to the boiling regime.
• This occurs as a result of a transition in the mode of heat transfer from thin-film evaporation at the free surface of a liquid-saturated wick to bubble nucleation at the substrate–wick interface, effectively eliminating the wick conduction resistance.

Figures

Fig. 3. Diagrams of (a) the sintered copper wick test sample where the dashed region indicates the location of the sintered wick sample and (b) the graphite soldering stand used for proper alignment of the solder joint. The 0.102 mm thick layer of solder is made of Pb–Sn–Ag–Sb (62–35.75–2–0.25%) and has a low melting temperature (Tmelt = 178 C) and a thermal conductivity of 50 Wm 1 K 1.

Fig. 11. Image and representative sketch of the fluid free surface and solid particles at a heat flux for which the evaporation regime is sustained for sample 250-355:600b. The top layer of particles can be seen protruding from the liquid free surface with the curved menisci outlined by reflection of the light source. The porous layer is approximately three particle diameters thick.

Fig. 4. Diagram of the copper heater block temperature measurement locations and nomenclature.

Fig. 10. The boiling curve and the variation of thermal resistance with heat flux for sample 250-355:600b. High-speed visualization of the sample surface during testing confirms that up to a heat flux of 55 W cm 2, only evaporation from the liquid–vapor interface at the top of the wick is present. This is followed by the incipience of boiling at the 75 W cm 2 flux. The (red) data point shown as a diamond symbol and the dashed lines represent the heat flux, superheat, and resistance directly prior to incipience of boiling as extracted from the transient test data. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)

Fig. 5. Steady-state thermal performance results for the 900 lm thickness test samples. The maximum heat flux test point does not correspond to a dryout condition for any of the curves. The results are presented in the form of (a) boiling curves, and (b) calculated sample thermal resistances as a function of heat flux. (Maximum uncertainties: q00 ±11.2 Wm 2, Rsample ±0.047 C/W, DT ±6.8 C.)

Fig. 6. Steady-state thermal performance results for the 250–355 lm diameter particle test samples. The results are presented in the form of (a) boiling curves, and (b) calculated sample thermal resistances as a function of heat flux. (Maximum uncertainties: q00 ±11.3 Wm 2, Rsample ±0.047 C/W, DT ±6.9 C.)

Fig. 1. Cross-section of a vapor chamber showing the important thermal resistances.

Fig. 12. (a) Series of images documenting growth and departure of a single bubble from a nucleation site within the porous layer for sample 250-355:600b. In the initial frame, the liquid free surface is undisturbed. In the following frames, the arrows point to the bubble diameter as it grows in size before breaking through the free surface and being released. The images are extracted from video recorded at 10,000 frames per second (video is provided as Supplementary data). (b) Representative sketches of the nucleation site (corresponding to upper left image in part (a)) and bubble growth (corresponding to stages 2a and 2c in the photographs).

Fig. 8. Average boiling thermal resistance for all 9 test samples as a function of the sample wick thickness. (Maximum uncertainties: q00 ±11.6 Wm 2, Rsample ±0.047 C/ W, DT ±7.1 C.)

Fig. 7. Average boiling thermal resistance for all 9 test samples as a function of the average sample particle size. (Maximum uncertainties: q00 ±11.6 Wm 2, Rsample ±0.047 C/W, DT ±7.1 C.)

Fig. 9. Schematic illustration of the visualized two-phase regimes for wicked evaporation and boiling. Resistances to heat flow are also conceptually indicated to explain the heat transfer mechanisms.

Fig. 2. Diagrams of (a) the test chamber, (b) the heater block and sample assembly, and (c) the test facility flow loop.

Frequently asked questions

Q1. What are the contributions in "Characterization of evaporation and boiling from sintered powder wicks fed by capillary action" ?

The objective of the current study is to measure the dependence of thermal resistance on the thickness and particle size of such surfaces. 

Q2. What is the way to measure the temperature drop across the graphite wick?

The higher conductivity and the well-defined and minimized thickness of this interface are critical for minimizing the uncertainty in estimating the temperature drop across this interface, especially at the high heat fluxes investigated in this study. 

Q3. What is the main focus of past investigations?

Understanding the performance limitations imposed on the heat pipe operation by the available capillary pressure and the wick permeability has been the focus of past investigations. 

Q4. How many T-type thermocouples are inserted into the tip?

Four T-type thermocouples are inserted into 1.19 mm diameter holes manufactured into the tip at a uniform separation distance of 7.62 mm. 

Q5. How is the wick feeding and boiling mechanism maintained in a heat pipe?

To reproduce the wick feeding and evaporation/boiling mechanisms occurring in a heat pipe, the liquid level in the test chamber is maintained at a fixed height throughout the duration of each test by means of a carefully located drain on the side wall. 

Q6. What is the process of heating the solder joint?

Before heating the solder joint, the sample surface is purged with argon supplied from a compressed gas line that is fed directly into the recessed pocket. 

Q7. What is the thermal resistance of the sample at low heat fluxes?

It was observed from the thermal measurements in the preceding sections that the overall sample thermal resistance is highest at low heat fluxes (when evaporation is the main mode of two-phase transport) and sharply decreases and to a relatively constant valueafter the incipience of boiling. 

Q8. What is the thermal resistance for the smallest particles?

8. For all particle sizes, the thermal resistance is at a maximum for the median thickness sample and decreases slightly for the 600 and 1200 lm samples. 

Q9. How is the temperature at the back of the copper substrate calculated?

The temperature at the back of the copper substrate can then be calculated by extrapolation usingTsubstrate ¼ T4 q00x ðx5 x4Þ kCu þ ts ks ; ð2Þbecause the cross-sectional area remains uniform up to this location. 

Q10. How is the temperature of the sample compared to the boiling curve?

From the boiling curve, a nearly linear rise in the superheat temperature is observed as the heat flux is gradually incremented up to 55 W cm 2. 

Q11. What is the effect of minimizing the particle layer thickness?

These studies also concluded that minimizing the particle layer thickness can reduce the conduction thermal resistance below the evaporating meniscus. 

Q12. What are the uncertainties for the heat flux, superheat temperature, and thermal resistance?

The average (and maximum) uncertainties for the heat flux, superheat temperature, and thermal resistance are ±2.1%, ±25.5%, and ±25.8% (±3.9%, ±38.3%, and ±38.2%), respectively. 

Q13. What is the effect of the dryout conditions on the sample?

The increase is likely induced as the sample experiences dryout conditions over some regions, leading to localized vapor blanketing and a concomitant increase in the average surface temperature. 

Q14. What is the thermal resistance of the powder samples?

The calculated thermal resistances of these different test samples are presented in Figs. 5(b) and 6(b) as a function of heat flux.