MEMS Loop Heat Pipe Based on Coherent Porous Silicon Technology

Debra Cytrynowicz1, Mohammed Hamdan2, Praveen Medis1, Ahmed Shuja1, H. Thurman Henderson1, Frank M. Gerner2, Eric Golliher3

1Center for Microelectronic Sensors and MEMS, Department of Electrical, Computer Engineering and Computer Science, (513) 556 – 4774, 2Department of Mechanical, Industrial, and Nuclear Engineering, (513) 556 – 2646, University of Cincinnati, Post Office Box 210030, Cincinnati, Ohio 45221-0030, 3Spacecraft Thermal Systems,

NASA John H. Glenn Research Center at Lewis Field, 21000 Brookpark Road, Mail-Stop 301-2, Cleveland, Ohio, 44135 – 3191, (213) 433 – 6575, dcytrynowicz@netscape.net, hamdanm@email.uc.edu, pmedis@ececs.uc.edu,,

ashuja@ececs.uc.edu, thenders@ececs.uc.edu, Frank.Gerner@uc.edu, eric.golliher@grc.nasa.gov Abstract. This paper discusses the theory, modeling, design, fabrication and preliminary test results of the MEMS loop heat pipe being developed at the Center for Microelectronic Sensors and MEMS at the University of Cincinnati. The emphasis is placed upon the silicon micro wick and its production through a novel technique known as Coherent Porous Silicon (CPS) Technology.

INTRODUCTION

A heat pipe is a two phase thermal device for cooling, which utilizes the significant heat of evaporation when the operating fluid (water in this case) changes to the vapor phase. The system is essentially adiabatic from the evaporator to the condenser, where the vapor again becomes liquid, thus dumping its latent heat. The resulting liquid then recirculates to the evaporator to repeat the cycle.

In contrast to this present work, single channel micro heat pipes were developed earlier in this lab by H. T. Henderson and F. M. Gerner (Badran et al, 1997, Gerner et al, 1994, Longtin et al, 1994), under NASA Glenn sponsorship, in micromachined (100) silicon. In this latter work, each glass enclosed silicon V-groove acts as a single wick due to the hydrophilic glass or silicon dioxide surface for the liquid meniscus at the groove apex. Such planar arrays can be effective heat spreaders but are inherently limited because of their planar two-dimensionality, i.e. the closed grooves lie in the silicon wafer or chip surface.

In the work herein, however, a silicon loop heat pipe (LHP) is being fabricated, where the condenser can be either on-chip or more typically removed some considerable distance away, connected by tubing from and back to the wick. In a LHP the wick serves as the “engine” which pulls the liquid interface back up to the evaporation surface by capillary action. A pump can be used to overcome the viscous losses of the recirculation system external to and within the wick; in our case a very novel silicon micro wick array serves the purpose of a pump in a completely passive system where waste heat serves as the energy source.

The unique Coherent Porous Silicon (CPS) wick is the breakthrough technology where vast arrays (at present 6 x 106 / cm2) of uniform micrometer-sized through-holes are concurrently etched perpendicular to the (100) silicon wafer or chip surface. Silicon dioxide capillary tubes are then grown inside the pipes or pores, which show enormous capillary action.

The assets of the novel CPS wick include (a) a flat or planar cooling surface which unlike cylindrical (incoherent sintered wick) configurations, lends itself to greater applications; (b) the surface “hot plate”, which may in many applications be integrated with the wick, is itself silicon and may contain ordinary electronic circuitry, and other optical and MEMS devices; (c) the pores can be sized, are uniform, and are coherently “stacked” to allow maximum

density and porosity; and (d) the manufacturing process is an inexpensive batch procedure (much like that of integrated circuits) which also allows shaping of the pores and package integration.

This paper gives preliminary results of the ongoing work, which projects theoretical cooling possibilities well in excess of 300 W/cm2.

GENERAL DESCRIPTION OF THE LOOP HEAT PIPE SYSTEM A Loop Heat Pipe, designed to transport heat over relatively long distances is being developed at the Center for Microelectronic Sensors and MEMS at the University of Cincinnati. The device consists of an evaporator, condenser, a liquid reservoir, and separate liquid and vapor lines. It has a liquid medium that is converted to a vapor at the evaporator and is then converted back to a liquid at the condenser. The pressure of vaporization is larger than the pressure of condensation and hence the vapor flows from the evaporator to condenser with no additional power input. Consequently, the LHP is passive or self-starting. This passive transport of heat is achieved by capillary pumping in the Coherent Porous Silicon wick structure, which is located in the evaporator. The entire cycle of evaporation and condensation occurs in a sealed, evacuated package. The liquid reservoir ensures that the wick remains wetted. Figure 1(a) and 1(b) (Ranganathan,1999) depict the LHP system. (a)

FIGURE 1. This is a representation of the loop heat pipe system. (a) Block Diagram, (b) Schematic Representation.

A finite heat flux is introduced into the device at the evaporator. The CPS wick located inside the evaporator functions as an engine to the liquid vapor interface or meniscus. Capillary forces maintain the separation between the liquid and vapor phases. The heat evaporates the liquid, creating a vapor pressure. The pressure drop results in a vapor phase fluid flow from the evaporator to the condenser where the vapor returns to the liquid state. The liquid must return to the evaporator once it is condensed. The evaporation process, the wetting properties and the surface tension of the liquid maintain a finite curvature of the liquid-vapor interface in a quasi steady state fashion. This translates into a pressure difference across the interface that is based on the capillary pressure of the microscopic pores of the CPS wick. The amount of liquid that evaporates in the wick is constantly replenished from the liquid phase reservoir by the capillary forces of the pores. This completes the cycle. The input energy is converted into the latent heat of vaporization at the evaporator. This energy is transported in the vapor phase to the condenser, where it is transformed back to heat energy by the condensation process. The processes of evaporation and condensation are nearly isothermal. The result is a very high effective thermal conductivity and the CPS source uses the waste heat to run the engine. Existing and proposed applications of the LHP system include electronic cooling, such as for laptop computers and high power electronics, waste heat rejecters in space vehicles, and solar concentrator cooling.

Input Heat Output Heat

Vapor Flow

Liquid Flow (b)

Condenser

Evaporator

MEMS BASED COHERENT POROUS SILICON MICRO WICK STRUCTURE

The evaporator section of the planar loop heat pipe consists of an evaporator “hot plate”, the coherent porous silicon wick, a liquid reservoir or compensation chamber, and separate liquid and vapor lines. The various components of the device are fabricated using standard microelectronic and MEMS techniques. Figure 2 gives a schematic representation of the device.

FIGURE 2. Components of the Loop Heat Pipe Evaporator.

The Coherent Porous Silicon Wick The very novel CPS technology was utilized in the fabrication of the porous wick in the loop heat pipe. The wick was made from a two-inch n-type (100) silicon wafer and the structure separates the top and bottom plates in a bonded sandwich structure. The wick consists of an array of coherent pores produced by the light-assisted electrochemical dissolution of silicon in an aqueous solution of hydrofluoric acid. Macroporous silicon technology, originally developed by Lehmann (Lehmann, 1997), allows the formation of coherent arrays of straight pores in n-type silicon. The technique was extended and applied by H. T Henderson’s group at the Center for Microelectronic Sensors and MEMS at the University of Cincinnati. Pore diameters achieved range between one and 35 micrometers and pore pitch ranges between four and 100 micrometers (Hölke et al, 1998, Rajaraman, 2000). Henderson’s group at the University of Cincinnati is also investigating coherent porous silicon for use as micro-needles, gas discharge lamps, and other microfluidic systems such as “chemistry laboratory on a chip” devices (Ranganathan, 1999, Rajaraman, 2000, Mantravadi, 2000, Van Dyke, 2000).

The pattern of the array of pores, pore diameter and center-to-center spacing or pitch, is predetermined by photolithography. The diameter of a pore is determined largely by the resistivity of the wafer. Inverted pyramidal etch pits are formed in the silicon by anisotropic wet chemical etching, establishing the etching sites. When properly biased, silicon etches in an aqueous solution of hydrofluoric acid if electronic holes are present in the material. Electron/hole pairs are generated across the bulk of an n-type silicon wafer under illumination. When an electric field is applied, the electronic holes are attracted to the tips of the etch pits and the anisotropic etching of coherent porous silicon results. Pore growth occurs at a specific critical current density, JPS, on the voltage – current curve. The holes produced are consumed at the pore tips in the formation of anodic sites, which results in an oxidation. The oxide formed is then etched by the hydrofluoric acid solution. If the generation area of the electronic holes is limited to the backside of the wafer, the pores tend to grow towards the light, the source of the electronic holes. On a (100) wafer the dominant direction of pore growth is in a vertical <100> direction. Branching can result and pores can grow along any of the six <100> directions when the critical current density is extended. This branching has applications in achieving cross-permeation.

Many of the wicks used in the initial loop heat pipe had pores with a diameter of five micrometers and a pitch of twenty micrometers. The porosity was very low, only 6.25 percent. The theoretical maximum porosity is ninety percent. The wicks are lapped to a thickness between 150 and 200 micrometers. Wicks of higher porosity are currently in the development stage. Pore diameters will be two micrometers or less and pore pitch will be four

Input Heat

Coherent Porous (100) Silicon Wick

Wick Bonds

Evaporator

Liquid Reservoir

Vapor Outputs

Liquid Input Secondary Wick

micrometers or less are feasible. The components will be packaged by means of wafer-to-wafer anodic bonding and the individual devices will be diced out. The manipulation of pore shape to increase wick porosity and the exploitation of the branching phenomenon to achieve cross-permeation are being investigated. Preliminary results have been demonstrated in square and hexagonal cross-sectional shapes. Figure 3 is a SEM micrograph of a sample of coherent porous silicon before through-holes had been produced.

FIGURE 3. This SEM micrograph shows a sample of coherent porous silicon. Pore Diameter: 5 µm, Pore Pitch: 20 µm, Wick Porosity: 6.25 %.

Evaporator Top Hot Plate The evaporator top "hot plate” is the contact to the heat source. It can be constructed with silicon or some other material. For these experiments, top plates were constructed in silicon as well as borosilicate, Pyrex 7740, glass. Although any size wafer can be used, two-inch (100) n-type silicon wafers were used. One top plate design contained two rectangular reservoirs connected by several micromachined rectangular V - grooves which collect the vapor. The second design contained a one square centimeter reservoir for vapor collection. The structures were etched to a depth of 212 micrometers into the (100) silicon wafer by traditional anisotropic etching in an aqueous solution of potassium hydroxide. A one micrometer layer of silicon dioxide served as the etch mask for both silicon designs. Figures 4(a) and 4(b) are photographs of the two silicon top plates fabricated for the experiments.

(a) (b) FIGURE 4. These are photographs of the two top plate designs. (a) Silicon Grooved Top “Hot” Plate, Individual Chip Size: 1.7 x 1.7 cm2, Reservoir Area: 1.04 x 1.1 cm2, Reservoir Depth: 212 µm, (b) Reservoir Style Silicon Top “Hot” Plate, Individual Chip Size: 1.2 x 1.2 cm2, Reservoir Area: 1 cm x 1 cm, Reservoir Depth������ �.

Bottom Liquid Reservoir Plate The bottom plates were fabricated on a two-inch borosilicate, Pyrex 7740, glass wafer, although the plate could also be made in silicon. The function of the experimental one square centimeter reservoir was to keep the wick wetted. The reservoir was etched into the glass wafer to a depth of 150 micrometers by means of traditional isotropic wet chemical etching in 49 weight percent hydrofluoric acid. A 5000 angstrom layer of gold on a 500 angstrom chromium adhesion layer served as the etch mask. Figure 5 is a photograph of a glass bottom plate. FIGURE 5. The glass liquid reservoir bottom plate is shown in this photograph. Individual Chip Size: 1.7 x 1.7 cm2, Reservoir Area: 1 cm2, Reservoir Depth: 150 µm.

Connections to the System To make the connections to the liquid input and vapor output lines, one hole was drilled in the center of each of the reservoirs of the glass bottom and silicon top plates. A diamond-dusted drill bit, one millimeter in diameter, was used to drill the holes. Small sections of glass capillary tubing with an inner diameter of one millimeter were epoxied with JB KWIK Bond to the plates at the each of the holes. Silicone tubing was used for the liquid and gas transport lines. In actual operation the source of the input heat to the LHP could be something such as integrated electronics or a solar cell concentrator. To simulate this input heat during device characterization tests, a section of a Kapton heater strip was attached to the back of the top plate. A copper-constantan thermocouple was placed at the top and bottom of the wick, and at the heater to monitor the temperatures at these locations during the test sequence. The pressure at the vapor output lines was monitored by Honeywell model 24CDFA6D pressure sensor. The assembly was mounted on a copper-clad board. A needle valve was included in the circuit to manipulate the pressure. For these experiments the plates of the LHP were epoxied together. The plates will eventually be bonded together by anodic or fusion type boding techniques. Figures 6(a) and 6(b) are photographs of the packaged device. Figure 7 is a photograph of the device in the actual test setup.

(a) (b) FIGURE 6. Photographs depict the experimental one square centimeter loop heat pipe assembly. The lateral dimensions of the device are arbitrary. The device pictured is epoxied together at the edges, however, the norm is anodic or fusion type bonding. (a) Top View of LHP Evaporator, (b) Bottom View.

Thermocouple to the Heater

Thermocouple to the Bottom of the Wick

Kapton Heater Strip

Thermocouple to the Top of the Wick

L

FIGURE 7. The Experimental One Square Centimeter Loop Heat Pipe Assembly.

THE THERMAL ANALYSIS OF FLUID FLOW IN THE COHERENT POROUS SILICON CHIP

As previously stated, the LHP is a two-phase device that utilizes the capillary pressure generated in a porous wick to pump the working fluid through a sealed, closed loop. The vapor formed above the wetted wick in the evaporator is transported through the output vapor line to the condenser. The liquid that condenses is driven back to the evaporator by the capillary effect. This fluid circulation is passive in the sense that there are no external drivers, such as pumps, required for pumping to occur. They are driven alone by the waste heat that they transfer to the condenser.

FIGURE 8. The Liquid – Vapor Interface.

As shown in Figure 8, at a fully developed, laminar, steady state flow with constant properties, the pressure drop in the vapor region is given by Equation 1 and by Equation 2 in the liquid region. These pressure drops are due to the kinenamatic viscosity of the fluid. The pressure across the interface is given by the Young - Laplace relationship of Equation 3. The values of the pressure are given as in a condition of forward flow as Psat,v ≥ P2 > P1 > Psat,l . Some of the applied heat flux causes evaporation while the remaining heat flux leaks to the liquid lines. The heat flux that causes a phase change is known as the latent heat and it is related to the enthalpy difference between liquid and vapor as given in Equation 4.

Psat,l – P1 = - [8 l��� l� 2] (1)

Input Liquid Line

Output Vapor Lines

•

(1)

(2)

s

y

L mhfg •

R2 R1

T2

Tsat,v & Tsat,l

T1

Q

P2 – Psat,v = -�� vm(L –���� v� 2) (2)

rPP lsatvsat

σ2,, =− (3)

q = mhfg (4)

Combining Equations 1 through 4 produces Equation 5:

( )[ ]

+−−=− 212

182

ryyL

hA

q

rPP lv

fg

ννε

σ (5)

Results

Equation 5 was plotted as shown in Figures 9 through 12. Figure 9 shows the characteristic pumping curves of three silicon wicks with different pore radii and the same porosity. The build-up of pressure will decrease as the mass flow rate increases due to increased friction losses. For a mass flow rate of approximately zero, no heat flux, the pressure across the wick is determined by the capillary pressure. Therefore, at small heat flux values, the build up of pressure increases inversely with the pore radius. At high mass flow rates when the pressure across the wick is approximately zero, the pressure across the wick is dominated by friction losses. The reason for such a non-monotonic relation is that at high mass flow rates the build up of pressure across the wick will increase inversely with the squared value of the pore radius. The pressure drop through the system will increase as the mass flow rate increases. For a fully developed, laminar flow in a smooth pipe, a linear relation between the mass flow rate and the pressure is expected as shown in Figure 9. Figure 10 shows the optimum value of pressure build-up across the wick as a function of different heat flux values. Capillary pressure dominates at low flow rates close to zero while friction losses dominate at high flow rates. Stated another way, pressure across the wick at a low flow rate will increase as a function of (1/r) while at a high flow rate it will decrease as a function of (1/r2). Therefore, a non-monotonic relation is expected between the pressure across the wick and the pore radius as shown in Figure 10. It also shows that a small reduction in the pore size from the optimum value will tremendously reduce the pressure built up. By increasing the mass flow rate, such as by increasing the amount of heat flux removed, the maximum pressure build-up that can be developed across the wick or through the LHP will be reduced.

Figures 10 and 11 show the same relation but for different positions of the interface. The interface position is believed to have a large effect on the pressure built up across the wick due to the friction losses in the vapor and liquid regions. For water, the friction loss in the vapor region is 100 times larger than that in the liquid region for the same mass flow rate. The kinematic viscosity of the vapor is 100 times larger than that of the liquid; therefore, keeping the vapor region as small as possible is recommended. The pressure drop due to friction loss is minimal when the wick is fully saturated. Figure 12 shows that porosity will enhance the performance of the wick tremendously since a higher pressure build up can be created due to less of a pressure loss.

•

•

FIGURE 9. Characteristic Curves for Different CPS Wicks. FIGURE 10. The Build Up of Pressure Versus

Pore Size As a Function of the Input Heat.

FIGURE 11. The Build Up of Pressure Versus Pore Size FIGURE 12. The Build Up of Pressure Versus As a Function of the Interface Position. Pore Size As a Function of Wick Porosity.

Preliminary Experimental Results Preliminary experiments were conducted using the different types of LHP packages fabricated at University of Cincinnati. A specially designed polycarbonate jig was used to perform some of the permeability and bubble burst through tests. The initial tests performed were permeability, porosity and capillary tests, tests for package leakage and dry loop tests. The last tests conducted in the characterization sequence were open loop and closed loop tests. All tests were performed in a cleanroom, and utmost care was exercised to prevent contaminants, such as dust, et cetera, from entering the system, and clogging the exceptionally small pores. Purified de-ionized water was used as the working liquid in all of these experiments.

40 8 0 120 160 200 Heat Flux, Q, (W/m 2 )

0

2

4

6

8 P

ress

ure,

P, (

104 P

a)

For Y = 0.5

System r = 5 µ m

r = 2.5 µ m r = 2 µ m

1E-8 1E-7 1E-6 1E-5 1E-4 Radius of the Pores, r (m)

1E+3

1E+4

1E+5

1E+6

1E+7

Bui

ld U

p of

Pre

ssur

e, d

P, (

Pa)

1E+2

Y = Wick Q = 100 Wm2 Porosity = 5 %

Y = 1.0 Y = 0.75 Y = 0.5 Y = 0

Bui

ld U

p of

Pre

ssur

e, d

P. (

Pa)

Radius of the Pores, r, (m) 1E-9 1E-8 1E-7 1E-6

1E+4

1E+5

1E+6

1E+7

1E+8 For Different Porosity

Q = 100 Wm2 Fully Saturated Wick

ε = 5 % ε = 20 % ε = 50 % ε = 80 %

Bui

ld U

p of

Pre

ssur

e, d

P, (

Pa)

Radius of the Pore, r, (m)

1E-6 1E-5

1Ε+2

1Ε+3

1E+4

1E+5

1E+6

Q – Heat Transfer (W/m2) Fully Saturated Wick

5 % Porosity Q = 1MW/m2

Q = 10 W/m2

Q = 100

1E+7

1E-8 1E-7 1E-4

The Permeability Test

As shown in Figure 13, the liquid is pumped through the silicon wick with a specific pressure head. The mass flow rate was calculated by measuring the liquid collected per unit time. Since the experimental results agreed with the theoretical model (Bejan, 1995) the porous model may be applied to the CPS wick. Figure 14 shows the theoretical and experimental test results.

FIGURE 13. This schematic shows the permeability test setup.

FIGURE 14. The graph displays the results of the preliminary permeability tests.

The Capillary Test

A U-tube manometer as shown in Figure 15 was used to evaluate the capillary performance of the silicon wick. The silicon wick was initially placed in a position lower than that of the liquid vessel to ensure the wetting of the CPS wick. While the wick was held at that same position, the liquid vessel was gradually lowered until the column began to recede. The difference in fluid levels, h, was measured.

FIGURE 15. The capillary pressure test setup is schematically represented.

Pressure Head

Pressure Gauge

Collecting Reservoir

CPS Wick

0

20

40

60

80

0 0.5 1 1.5 2 2.5 3 3.5

Voltage

LHP - Dry LHP - Wet Pressure

Hea

ter

Tem

pera

ture

h CPS Wick

Liquid Reservoir

Package Leakage Test

For a predetermined amount of time, the package was immersed in a bath of water and the loop was pressurized with nitrogen. A steady pressure was maintained by means of a pressure gauge. Readings of the pressure gauge as well as the visual observations were made. A similar test was conducted with a vacuum applied. According to the gauge pressure, both tests showed that the package could easily handle pressures ranging from negative seven to positive seven pounds per square inch.

Dry Loop Tests and Wet Loop Tests

These tests were conducted to determine whether there was evaporation occurring inside the package and whether there was fluid flow in the loop. A dry loop test was performed first. Temperature readings of the wick were taken at each increase in voltage to the heater. A similar test was then conducted using water in the loop as the working fluid. The results of the second test, when compared with those of the first, showed a decrease in temperature, which might be proof of a mass flow rate due to evaporation in the CPS wick. In one such instance the package had a wick with five-micrometer diameter pores spaced twenty micrometers apart. This wick demonstrated a temperature drop of 120 C from the dry test to loop with water as the working fluid. The temperature of the chip was 65º C. Figure 16 graphically displays a comparison between the results obtained in the dry and wet loop tests

FIGURE 16. The Volumetric Flow Rate Versus the Pressure Drop.

The Secondary Wick A common problem in microfluidic devices is the ability to handle two phases. In a microchannel when a gas is allowed to form in the channel with liquid surrounding it, high pressures are needed to remove gas bubbles. In earlier experiments with the CPS wick structure, observations determined the presence of vapor on the liquid side of the wick. This has been experimentally shown to occur in the start-up stage of operation, when the water in the reservoir is heated due high conduction of heat through the wick and water channels. This causes vapor to form in the reservoir, which can impede the capillary action of the wick. In the earlier design there was no integral method to combat this transient effect. In this case once the vapor bubble existed there was no way to eliminate it during operation. The volume occupied by vapor as compared to that of the same mass of water was also shown to lead to a drying-out effect in the package. At high heat fluxes, a substantial amount of vapor produced can result. Due to the difference between the densities of water and vapor (~1000 times), the vapor can fill the top chamber and begin to

0.E+00

4.E-08

8.E-08

1.E-07

2.E-07

2.E-07

0 2000 4000 6000 8000 10000

Pressure Drop, (Pa)

Vol

umet

ric

Flo

w R

ate,

(m

3 /s)

ExperimentaTheoretical

cause a large build up of pressure on the top side of the wick. If the pressure build up is greater than the capillary force of the pores in the CPS, the water will burst out of the wick. This phenomenon was observed in the start-up of the wick during previous experimental testing of the packaged device. As in the previous situation, this presents a problem during device operation. In a microgravity environment no means exists to bring the water in the reservoir back to the primary wick. The experimental observation of these two effects led to the decision that a secondary wick could be utilized to handle any vapor on the reservoir side of the package. Classically the secondary wick in a LHP consists of a reticulated structure similar to the primary wick. The main difference will be the larger size pores to allow for a low pressure drop and sufficient capillary force to bring the liquid back to wet the primary wick. Instead of the traditional secondary wick it was decided to use a network of glass fiber to act as a secondary wick. The general idea was to render the effort of fabricating a secondary to a minimum as to conserve time. Initial tests were run on the use of quartz fiber as a secondary wick. There were many reasons for the choice of this material as the secondary wick. Unlike other fibers, such as cotton, utilized as a wicking material, quartz fibers do not absorb water. The water simply wets the surface of the quartz because it is strongly hydrophilic. Initial tests were run to determine whether the ten-micrometer fibers could be packed and the capillary forces could achieve a four-millimeter water level rise. Upon initial investigation, up to a five centimeter rise in water level was achieved. Due to the success of this test, absorption tests were run to determine the porosity of the secondary quartz wool wick. In Figure 17 the absorption indicates the maximum porosity at approximately 92 percent. The absorption begins to decrease after this point with increased packing density. The sketches below the graph depict the phenomena that occur as the packing density increases. At a porosity of 95 percent there will be areas where the fibers are not close enough to accommodate the flow of the water. At even higher packing densities, the fibers begin to break and glass particles fill some of the voids. This graph was fit to a second-degree polynomial as shown. The equation for this polynomial will have its peak value at 0.85 grams of glass for a volume of five cubic centimeters. This corresponds to a porosity of 92 percent. Permeability tests of the secondary wick were performed as shown in Figure 18. The graph demonstrates the linear relationship between the flow rate and the pressure drop. The experiments indicate that the quartz should provide the necessary capillary pressure to deliver water to the primary wick while maintaining a low pressure drop.

FIGURE 17. The effect of different packing densities on absorption is graphically displayed. The bold line represents the capillary force. The data for the absorption test was fit to a second degree polynomial and demonstrated a very good match.

1.0834 0.7547

0.5053

2

3

4

5

0 0.4 0.8 1.2 Grams of Quartz Glass

Mill

ilite

rs o

f W

ater

Abs

orbe

d

Porosity: 95.41 % Porosity: 92.00 % Porosity: 90.51 %

Capillary Force

y = -6.498x2 + 11.095x – 0.1401 R2 = 1

y = 4.384E-11x + 9.564E-09R2 = 9.892E-01

0.E+004.E-088.E-081.E-072.E-072.E-072.E-073.E-073.E-074.E-074.E-074.E-075.E-07

0 2000 4000 6000 8000 10000 12000

Pressure Drop ( Pa)

Vol

umet

ric

Flo

w R

ate

( m

3 /s)

FIGURE 18. The Permeability Test of the Secondary Wick: The graph demonstrates a linear relationship between the pressure drop and the flow rate.

FUTURE RESEARCH NASA seeks to use this technology to develop advanced space power systems that carry all of the functionality of current power converters, for example, but in much smaller packages. To accomplish this, the micro LHP will remove a large amount of heat from a very small space. The power density of spacecraft electronics must increase to enable lighter, smaller and less costly satellite subsystems. NASA is bringing this micro LHP technology from Technology Readiness Level (TRL) 1 to 2. Prior to this effort, basic principles had been observed and reported. The UC CPS evaporator demonstrated that evaporation could occur and surface tension would maintain liquid in the wick. During this current effort, the concept for an actual micro loop heat pipe to include a secondary wick and compensation chamber was formulated. Further planned work will bring the technology to TRL 3. The analytical and experimental critical function of the micro LHP will be demonstrated. Variables, such as the choice of the optimum working fluid, such as methanol, toluene, ammonia, water, or some other fluid, will be explored. Modifications of the micro LHP geometry to increase performance will be the subject of follow-on work. Advances will be made in the CPS production process, such as control of pore formation through the increased control of the various process parameters. A greater understanding of the fundamental process of CPS formation will be achieved. Eventually, it is hoped that a high heat flux-capable micro LHP will be combined with an actual silicon power converter to demonstrate performance as a subsystem. Flight-testing aboard the International Space Station or Shuttle is also desirable for future validation of the silicon CPS micro LHP technology.

CONCLUSIONS With the utilization of the novel Coherent Porous Silicon structure, a MEMS based planer Loop Heat Pipe capable of handling high heat fluxes is being developed. The theoretical heat flux handling capability is equal to and possibly greater than 300W/cm2. By using the key coherent porous silicon based structure the freedom exists to modify such parameters as pore diameter, pitch, length, and shape. This allows the thermal behavior of the loop heat to be modeled more specifically. The optimization of the LHP for different space applications is simplified because of the ability to better control the characteristics of the final fabricated wick. No auxiliary driving input power is needed because the high capillary force produced by the wick’s micron-sized pores will drive the system.

ACKNOWLEDGMENTS

Major support for the project was received from the NASA John H. Glenn Research Center at Lewis Field (GRC) under grants NAG3 – 2281 and NAS3-01120 and NASA Langley Research Center (LaRC) under NASA grants NAS1 – 00123 and POL - 14449, and TEEC 50482SP SUBNASA. We would like to thank the Advanced Power and On-Board Propulsion Division of NASA’s Cross Enterprise Technology Development Program, especially Eric Golliher, Kenneth Mellott, and Richard Shaltens of GRC and Christopher Moore of LaRC for their continued support of and input to the project, and also for their efforts on our behalf. The research team would also like to thank the following people at the University of Cincinnati for their assistance with the various activities of the project. Srinivas Parimi and Srikoundinya Punnamaraju have been responsible for the bonding aspects of the project. Sowmya Suryamoorthy has been of assistance with the fabrication of the components.

REFERENCES Badran, B., Gerner, F. M., Ramadas, P., Henderson, H. T., Baker, K. W., “Experimental Results of Low – Temperature Silicon Micromachined Micro Heat Pipe Analysis Using Water and Methanol as Working Fluids,” Experimental Heat Transfer, Volume 10, Number 4, 1997, pp.253 – 272. Bejan, A., Convection Heat Transfer, John Wiley & Sons, Incorporated, New York, 1995. Gerner, F. M., Badran, B., Henderson, H. T., Ramadas, P., “Silicon - Water Micro Heat Pipes,” Micro and Molecular Heat Transfer, Editors: Tien and Kotake, Begel House, 1994. Gerner, F. M., Badran, B., Henderson, H. T., Ramadas, P., “Silicon – Water Micro Heat Pipes,” Thermal Science and Engineering, Volume 2, Number 1, 1994, pp. 90 – 97. Hölke, A., Pilchowski, J., Henderson, H. T., Saleh, A., Kazmierczak, M., Gerner, F. M., Baker, K., “Coherent Macro Porous Silicon as a Wick Structure in an Integrated Microfluidic Two – Phase Cooling System,” Proceedings of the SPIE Conference on Microfluidic Devices and Systems, Santa Clara, California, September 21 – 22, 1998. Lehmann, V., “The Physics of Macropore Formation in Low Doped N – Type Silicon,” Journal of the Electrochemical Society, Volume 140, October 1993, pp. 2836 – 2843. Longtin, J. P., Badran, B., Gerner, F. M., “A One – Dimensional Model of a Micro Heat Pipe During Steady State Operation,” Journal of Heat Transfer Volume 116, 1994 pp. 709 – 715. Mantravadi, N., “MEMS – Based Development of a Silicon CPS Wick for Loop Heat Pipe Applications,” Master’s Thesis, University of Cincinnati, Cincinnati, Ohio, 2000. Rajaraman, S., “Silicon MEMS – Based Development and Characterization of Batch Fabricated Microneedles for Biomedical Application,” Master’s Thesis, University of Cincinnati, Cincinnati, Ohio, 2000. Ranganathan, S., Hölke, A., Pilchowski, J., Henderson, H. T., Kazmierczak, M., Gerner, F. M., Baker, K., “An Integrated Passive Cooling System for Space Applications,” Proceedings of the Second International Conference on Integrated Micro/Nanotechnology for Space Applications, Pasadena, California, April 11 – 15, 1999. Van Dyke, B., “Development of Coherent Porous Silicon for Use in Biological and Optical Applications,” Master’s Thesis, University of Cincinnati, Cincinnati, Ohio, 2000.