Development of A Novel Powder Cluster Wick Structure for LTCC Embedded Heat Pipes

Transcription

1 Development of A Novel Powder Cluster Wick Structure for LTCC Embedded Heat Pipes Guangnan Deng, W. Kinzy Jones Hybrid lab, Department of Mechanical Engineering Florida International University, University Park Campus Miami, Florida Abstract This paper presents a novel powder cluster wick structure used for capillary pumping embedded heat pipe in LTCC substrate. Unlike normal sintered powder wick, this wick structure was fabricated by silver powder cluster, which helps it to get high capillary pressure and maintain a high permeability at the same time. The powder cluster wick were fabricated and tested for its maximum mass transport rate at vertical orientation. Experimental data from normal sintered powder wick were also collected to compare with the new wick. The results show that the powder cluster wick has a much higher maximum mass transport rate, thus can provide higher capillary limitation for the heat pipe. Key words Wick, Capillary Limitation, LTCC, Heat Pipe 1. Introduction The continuous miniaturizations of electronic devices and the rapid increase in power density of advanced microprocessors and electronic components have placed an ever-increasing importance on the thermal management in electronic packaging. Complete integration of cooling system will be needed to meet the cooling requirement for present and next generation electronic packaging system. Due to its extremely high effective thermal conductivity and the passive cooling nature, integration of micro heat pipes/ spreaders into the electronic package of microelectronic devices has become an attractive technology. Works at FIU embedding micro heat pipes in LTCC substrate have demonstrated capacity of dissipating thermal density of 100W/cm 2 [1]. Since the thermal performance of heat pipe is determined by its wick structure, it is crucial to develop suitable wick structure for the integrated miniature heat pipe. The International Journal of Microcircuits and Electronic Packaging, Volume 25, Number 1, First Quarter, 2002 (ISSN ) 80

2 The two most important properties of a wick are the capillary pressure and the permeability. The capillary pressure gradient drives the liquid from condenser to the evaporator, against the liquid and vapor pressure losses, and adverse body forces such as gravity. Large capillary pressure is preferred because the stable working fluid circulation in a heat pipe can be achieved only under condition that the capillary pressure head is greater than or equal to the sum of pressure losses along the vapor-liquid path. The permeability determines the frictional losses of the fluid as it flows through the wick, To maximize the liquid flow though the wick for a given pressure drop P, high permeability is preferred. Capillary pressure can be described by the surface tension σ, the wetting angle θ, and the effective pore radius, r eff [2]: 2σ cosθ P = (1) r eff Based on this expression, small pore radius, a liquid with large surface tension and good wetting are required to get high capillary pressure. The analytical expression permeability of wick structure is varied with the type of the wick. For an array of straight pores in a wick of porosity ε, assuming fully developed laminar flow in the pores of the wick one can show that, the permeability of a wick can be described as [3]: 2 r K = ε (2) 8 According to this expression, large porosity and pore radius are needed to get high permeability. Note that the latter requirement contradicts that for a high capillary pressure. In this paper, we introduce a new kind of sintered powder cluster wick. This work is a portion of an ongoing effort to develop advanced cooling system in LTCC substrate. Unlike normal sintered powder wick, this wick is fabricated by using silver powder clusters, which is made from aggregated finer silver powders. The benefit of this powder cluster lies in that the small pores on the surface of the cluster can effectively provide high capillary pressure, while the large radius of the clusters and resulting higher porosity help the wick get high permeability. 2. Fabrication of Powder Cluster Wick The silver powders were initially aggregated into clusters by mixing with a thin layer of PVB solution. These mechanically bond clusters were then sintered in a box oven. The final powder clusters go through mesh to be finally categorized. The original silver powder size is 400 mesh, the powder cluster has a size of 100 mesh. Figure 1 shows a SEM picture of the powder clusters. The powder cluster wick has a larger porosity because its total pore volume is the combination of the pore volume among the small powders and the pore volume among the big clusters. Assuming η 1 is the porosity of the powder cluster. Since each of the powder clusters is made from small powders, given the porosity of the cluster as η 2, the total porosity of the wick is ( η1) 2 ηw = η1 + 1 η (3) While for the normal powders with the size of the cluster, the porosity of the wick is 81

3 Figure1. SEM picture of a powder cluster η w = η 1 (4) In this investigation, wicks were successfully made using this kind of clusters. Since the capillary limit is one of the most commonly met limitation for miniature heat pipe, experimental study on the capillary limit of this cluster wick was performed. 3. Experimental Setup The capillary limit of the wick structure was characterized by testing its maximum mass transport rate of the wick, which is determined by the maximum capillary pressure, permeability, and thickness of the wick. To test the maximum mass transport rate of the porous wick structure, an experimental test structure similar to the working operation of the heat pipe was set up. The experimental setup is shown schematically in Figure 2 The experimental set-up use a testing plate with the wick sintered onside. The testing plate is made of 4 layers of LTCC tape and has a dimension of WLT 1x100 x. 5 mm. Porous silver powder wick was sintered on one side of the testing plate. On the other side, a serpentine platinum thick film resistor was sintered at one end to act as the heat source. The heater was insulated by using fire glass blanket. The power used by the platinum heater was supplied by a regulated DC power supply with 3A and 60 Volt capacities. The heat load was adjusted by varying the input voltage and current used by the heater, which were measured by two digital multi-meters. Power (P) was then calculated using the measured voltage (V) and current (I) by the relation P=V*I. To measure the temperature of the evaporate section; a type K thermocouple was bonded on the wick, right above the heater. Experiment data of the thermocouple was collected by a Scanning Thermocouple thermometer. 82

4 Power Line Aluminum foil Test Plane Heater Distilled Water Thermocouple Wick DC Power Supply Thermocouple Wire Scale Computer Figure 2. Experimental setup for testing mass transport rate The testing plate was secured vertically on a beaker. Distilled water was added to the beaker so that certain length of the wick is dipped into the water. The height of the water level can be varied so that the maximum mass transport rate of the wick can be tested at different wick length. The testing cell, including the beaker, testing plate and the power line, thermocouple onside, was put a Denver Instrument XP-600 digital scale. To eliminate the influence of the weight of the power line and the thermocouple wire, all the lines and wires are hung on a laboratory stand. In addition, these lines were tightly secured on the stand so that the weight under test would not change when the lines and wires are incidentally touched or moved during test. Since the water in the beaker can lose its weight because of its diffusion into the ambient condition, the opening of the beaker was covered by a layer of aluminum foil to eliminate the weight loss due to the diffusion from the water surface. The weight loss due to the water diffusion from the wick surface was included into the mass transport rate because it was also provided by the wick. However it was found to be as small as g/min, and thus was negligible anyway. 4. Experimental Procedure When power is input into the heater, the test setup would loss weight due to the evaporation of the water in the wick. Unless its maximum capacity was meet, the capillary force of the wick can continuously supply water to the evaporate area. The mass transport rate at given input power level can be tested by measuring the loss of the water over a certain period of time of operation. At any one of the power levels, the power was set and the wick allowed to reach steady state. Then, it was allowed to operate over a certain period of time, 10 minutes in our case. Since the temperature oscillated in a range of C during the test, so the temperature was 83

5 recorded every minute and the even temperature during the testing period was used. The weight loss during this period of time was recorded, and the mass transport rate was calculated by the relationship: M m = (5) t The primary objective of the test is to find the maximum mass transport rate of the wick at different wick length. To do this, the input power level was slowly increased and the mass transport rate at each of the power level was tested. The maximum mass transport rate is the one when the dry out occur, which can be monitored from the temperature at the evaporate section. 5. Results and Discussion Figure 3 shows the temperature at various input power level for wick height of 7.5cm. From the experiment data, it was found that, with the increase of the input power, the temperature rise steadily at the beginning. However when temperature approach to 100 C, it reach a plateau. During this period the temperature rise very slowly with the increase of the input power. This is because the evaporation of the water removed enormous heat from the wick, thus keep the temperature almost constant. The plateau can continue as long as the wick can provide enough water for evaporating. When the maximum mass transport rate was met, the dry-out occurred and the temperature was found to jump suddenly. The relation between the maximum mass transport rate and the wick length was shown in Figure 4. From the figure it is seen that the mass transport rate decrease lineally with the increasing wick height. This result is logical in that less capillary force was spent to overcome the weight of the water at the decreased wick height. The maximum mass transport rate tested is (g/min), at height of 4.5cm. In this investigation, a wick made by normal sintered silver powder (mesh size=100) was also fabricated to compare with the porous powder cluster wick. The experimental data for wick height of 4.5 cm were plotted in Figure 5. To provide comparisons, the data of the powder cluster wick at height of 4.5cm and 7.5cm were also plotted in the same figure. The experimental data were also listed in Table 1. From Figure 5 it could be seen that the wick made by normal silver powder has a much shorter plateau. The dry out was found to occur at input power level of as low as 22W. The according maximum mass transport rate was 0.44(g/min). While for the wick made by silver powder cluster, the plateau is very long, indicating large mass transfer ability (g/min) for h=4.5cm. Also it could be noticed that the plateau is shorter for wick height of 7.5cm. This is because the mass transport rate increased with the decreased wick height. However its maximum mass transport rate (0.54g/min) is even higher than the 4.5 cm normal powder wick. Note the maximum transport rate is also a function of the cross section area of the wick structure. To get the maximum transport rate per unit area, the cross section areas of the wicks were measured. For the wick made by silver powder cluster, the cross section of the wick is approximate rectangle with a dimension of WL 0.8 x 11 mm 2. So the area can be calculated by relation S = W x L = 8.8 mm 2. Using the mass transport data at wick height of 4.5 cm, the maximum mass transport rate per unit cross-area is (g/min x mm 2 ). For the normal powder wick, the cross section area is about 9.1 mm 2. At the wick height of 4.5 cm, 84

6 Temperature ( C) Power (W) Figure 3. Temperature plateau of powder cluster wick at height of 7.5 cm. Maximum Mass Transfer Rate (g/min) Wick Length (cm) Figure 4. The relation between the maximum mass transport rate and the wick length 85

7 Temperature ( o C) Power (W) powder cluster h=4.5 normal powder h=4.5 powder cluster h=7.5 Power (W) Figure 5. Temperature plateau of powder cluster wick and normal powder wick Table 1. Comparison of powder cluster wick with normal powder wick Wick type Wick length (cm) Maximum mass transport rate (g/min) Power level at dry out (W) Normal silver powder Silver powder cluster

8 the maximum transport rate per unit cross-area is (g/min x mm 2 ). Obviously normal powder wick has much lower mass transport ability than the powder cluster wick, which may due to the higher permeability of the latter. 6. Conclusions A novel silver powder cluster wick was developed for capillary pumping in embedded heat pipe in LTCC substrate. The use of powder cluster helps wick to get high capillary pressure and maintain a high permeability at the same time. Test results shows that the powder cluster wick has a much higher maximum mass transport rate than normal powder wick, thus can provide higher capillary limitation for the heat pipe. References [1] Embedded Micro Heat Pipes/Spreader in LTCC Ceramics, Proceeding of the Principle Investigators Conference, DAPAR HERETIC Program, Atlanta, 2001 [2] Faghri, A., 1995, Heat Pipe Science and Technology, Taylor and Francis, Washington, D.C. [3] N.Gernert,G.Baldassarre,J.Gottschlich, Fine Pore Loop Heat Pipe Wick Structure Development, Proceedings of Aerospace Atlanta Conference, SAE paper #961319, pp.1-10,