Test Cell for a Novel Planar MEMS Loop Heat Pipe Based on Coherent Porous Silicon

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

1Center for Microelectronic Sensors and MEMS, Department of Electrical Computer Engineering and Computer Science, University of Cincinnati,Cincinnati, Ohio, 45221

2Department of Mechanical, Industrial, and Nuclear Engineering, University of Cincinnati, Cincinnati, Ohio, 45221 - 0030

3Spacecraft Thermal Systems, NASA John H. Glenn Research Center at Lewis Field, 21000 Brookpark Road, Mailstop 301-2, Cleveland, Ohio 44135 - 3191, (513) 556 – 4774, dcytrynowicz@netscape.net

Abstract. Work towards the development of an innovative, potentially high power density, MEMS loop heat pipe is in progress at the Center for Microelectronic Sensors and M E M S at the University of Cincinnati. The design of the loop heat pipe is based upon the very unique coherent porous silicon technology, a technique in which vast arrays of micrometer - sized through - holes are photo - electrochemically etched into a silicon wafer perpendicular to the (100) surface. The initial mathematical model, the design, fabrication and characterization of the device in the open loop configuration were previously reported at this conference, STAIF 2002. This paper begins with a very brief explanation of the device and its theory of operation. The design of the device components and their production utilizing the various techniques of microelectronic and microelectromechanical fabrication are presented. The modifications made to the photon - induced, electrochemical etch process, which significantly increase the etch rate of the pores, are explained. Attention is given to the mathematical model of the planar, MEMS, loop heat pipe with respect to the generation of the dimensions of the components through a summary of the recent advances. The emphasis of this paper is upon the design, construction and the characterization of the evacuated closed loop test cell structure.

GENERAL DESCRIPTION OF THE MEMS LOOP HEAT PIPE

FIGURE 1. This is a schematic representation of the MEMS loop heat pipe and its components.

Heat In

� Heat Out

�

Liquid Transport Line

Vapor Transport Line Condenser

Liquid Reservoir Plate

Top “Hot” Plate

Coherent Porous Silicon Primary Wick

Evaporator

Stochastic Secondary Wick

Figure 1 is a schematic representation of the planar loop heat pipe being developed at the Center for Microelectronic Sensors and MEMS at the University of Cincinnati. It is a two-phase device designed for thermal control applications, such as the cooling of electronic devices and systems. The device consists of an evaporator, condenser, separate liquid and vapor transport lines, a primary and a secondary porous wick, and a working fluid. Conventional loop heat pipes are cylindrical in form and contain stochastic or amorphous porous wicks. The innovative device under development is planar and contains a primary wick fabricated with coherent porous silicon. The liquid reservoir or compensation chamber of the loop heat pipe contains the working fluid. Its function is to ensure that the primary wick remains wetted, and it controls the operating temperature of the loop. When a finite amount of heat energy is applied to the evaporator, the wick transfers the heat energy to the liquid-vapor interface inside the primary coherent porous silicon wick. The input heat energy is transformed into the latent heat of vaporization during the evaporation of the liquid, which creates a high vapor pressure. At the condenser the latent heat of vaporization is transformed back into heat energy by the condensation process. The condensation process maintains a lower vapor pressure on the condenser side where heat is extracted. The vapor phase fluid flow from the evaporator to the condenser occurs due to the difference between the pressures of the liquid and vapor phases. The large pores of the stochastic secondary wick, which is located in the compensation chamber, allow for a low pressure drop and sufficient capillary force to assist in keeping the primary wick wetted. The primary coherent porous silicon wick through its capillary action serves as the “engine” which pulls the liquid interface back up to the evaporation surface and the cycle repeats

FABRICATION OF THE LOOP HEAT PIPE EVAPORATOR COMPONENTS The evaporator section of the loop heat pipe consists of the top “hot plate, the liquid reservoir or compensation chamber plate, the coherent porous silicon primary wick and the quartz wool secondary wick.

The Top “Hot” Plate Reservoir style top plates were fabricated on two-inch borosilicate glass wafers, which were 700 micrometers thick. A 0.5-micrometer layer of gold was deposited on a 0.05-micrometer adhesion layer of chromium by electron beam evaporation to serve as the masking layer during the etching of the glass. The masking layer was patterned by ultraviolet photolithography. The gold was etched with aqua regia (3 HCl : 1 HNO3 : 4 H2O) and the chromium was etched with CR-7. The glass was etched at room temperature in a forty-nine percent by weight aqueous solution of hydrofluoric acid. Figure 2(a) and 2(b) are schematic representations of the glass top plate, and Figure 2(c) is a photograph. (a) Cross sectional view. (b) Top view. (c) Photograph of the glass top “hot” plate. FIGURE 2. These are schematic representations and a photograph of the glass reservoir style top “hot” plate. Plate Area: 2.89 cm2, Reservoir Area: 1 cm2, Reservoir Depth: 300 µm, Plate Thickness: 700 µm.

The Coherent Porous Silicon Wick

The very novel coherent porous silicon technology is a technique in which vast arrays of micrometer sized through-holes are photo-electrochemically etched into a silicon wafer perpendicular to the bottom (100) surface. Lehmann introduced the procedure for the production of macroporous silicon (Lehmann, 1993). The production procedure used at the University of Cincinnati was discussed in previous papers (Cytrynowicz et al., 2002, Hamdan et al, 2002). Coherent porous silicon is produced on a two-inch n-type, phosphorus doped, (100) oriented, silicon wafer. A thermal oxidation produces the 0.5-micrometer layer of silicon dioxide, which serves as a masking layer during the phosphorus diffusion that follows. The phosphorus diffusion, which uses phosphorus oxychloride, is performed to produce an n+ region across the backside of the wafer for ohmic contact purposes. The masking oxide layer is

removed and then either a dry thermal oxidation is performed or silicon nitride is deposited by low pressure chemical vapor deposition. During the production of the inverted pyramidal etch pits in the silicon wafer either can serve as the masking layer. Both masking layers are patterned by ultraviolet photolithography. If silicon dioxide is used as the masking layer, the windows are opened in the oxide using a buffered oxide etchant (5 NH4F: 1 HF). If silicon nitride is used, the windows are opened in the nitride layer by a reactive ion etch using carbon tetrafluoride and oxygen. The inverted pyramidal etch pits are formed in the silicon wafer by a room temperature wet chemical etch using a forty-five percent by weight aqueous solution of potassium hydroxide. After the etch pits have been formed in the silicon, aluminum is deposited by electron beam evaporation onto the backside of the wafer over the n+ region. The aluminum ring electrode is patterned by ultraviolet photolithography and the excess aluminum is etched away using PAE, an aqueous solution of phosphoric acid, nitric acid, and de-ionized water (19 H3PO4: 1 HNO3: 4 H2O). Coherent porous silicon is produced by a photon assisted electrochemical etch in which an aqueous solution of hydrofluoric acid serves as the electrolyte. The wafer is contained in a Teflon etch rig and a platinum electrode provides the contact to the electrolyte. An external bias voltage sufficient to supply the critical current density is applied to the etch rig. The backside of the wafer is exposed to light from a two hundred and fifty watt quartz-tungsten-halogen lamp. This light produces the electron-hole pairs that drive the dissolution reaction. The pyramidal etch pits concentrate the applied electric field and the pores begin to etch anisotropically at the tips of the pits. Although pore growth can be in any of the six <100> directions, pore growth in this case is towards the light, which is the source of the electronic holes. The etching process can be performed with the silicon dioxide or silicon nitride masking layer or without any masking layer. The etching process begins with obtaining the current versus voltage characteristic of each individual wafer. Figure 3 is an example of such a current versus voltage characteristic curve. The bias voltage is stepped up from zero to twenty-five volts and the current is measured at each increase of the voltage. From this I - V curve the bias point, or critical current density, of each individual wafer is determined. During the actual etching process, the hydroxyl ions oxidize the silicon and the hydrofluoric acid etches away the oxide. There are two types of etching techniques utilized in the production of coherent porous silicon. The standard process involves the use of a weak electrolyte and proceeds at a very slow rate. The alternate process proceeds at an accelerated rate and it involves the addition of hydrogen peroxide and a surfactant to a more concentrated electrolyte solution, which respectively act as oxidizing and wetting agents. The etch rates of the two processes differ by a factor of eight. The through – holes are produced by three different methods: etching through the entire thickness of the wafer, mechanical lapping of the backside of the wafer, or wet chemical etching.

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

7.5

8

0 5 10 15 20 25 30

Voltage, V

Cur

rent

, I, m

A

FIGURE 3. The Current Versus Voltage Characteristic Obtained Prior to the Coherent Porous Silicon Etch Process.

The pores were patterned in orthogonal and hexagonal arrays. Pore lengths ranged from three hundred micrometers to one millimeter. The diameter of the pores can range between one and thirty-five micrometers. The pore pitch, center-to-center spacing, can range between four and one hundred micrometers. For this application the pore diameters were five micrometers at a pitches of eight and ten micrometers. Pores were also fabricated with two-micrometer diameters at a pitch of four micrometers. The basic porosities of the arrays were twenty-five percent

Critical Current

(5µm@10µm, 2µm@4µm and thirty-nine percent (5µm@8µm). The theoretical upper limit to porosity is ninety-two percent, which is achievable through various post-processing techniques (Hölke, 1998; Mantravadi, 2000; Rajaraman, 2000; Ranganathan, 1999; Van Dyke, 2000). At present pores are patterned over the entire wafer surface. Either the full surface of the wafer is exposed to the electrolyte in the standard etch rig or two small circular areas approximately one centimeter in diameter are exposed to the electrolyte in the pieces rig. Figures 4(a) and 4(b) are photographs of coherent porous silicon prior to the production of the through – holes. Figure 5 shows the components of the pieces etch rig with an etched sample at the center of the photograph.

← 50µm→

FIGURE 4. The photographs display samples of coherent porous silicon before the production of the through-holes.

FIGURE 5. This photograph shows the components of the pieces rig and an etched sample.

The Liquid Reservoir or Compensation Chamber A lexan jig was constructed to serve as the liquid reservoir or compensation chamber bottom plate, to hold the quartz wool secondary wick and to house the evaporator package. The volume of the compensation chamber was one cubic centimeter. It was mechanically machined out of a lexan cube with an area of twenty-five square centimeters and a thickness of approximately 1.3 centimeters. A nine square centimeter stainless steel plate was attached by screws to the compensation chamber side of the lexan jig. Figures 6(a) and 6(b) are photographs of the lexan jig without and with the stainless steel back plate. This arrangement makes it possible to quickly change evaporator packages for testing purposes.

(a) (b)

FIGURE 6. Photographs of the Lexan Characterization Jig.

Etched Sample

Lexan Etch Mask With O-Rings

Copper Bottom Plate

(a) The pore diameter is two micrometers and the pore pitch is four micrometers.

(b) The pore diameter is five micrometers and the pore pitch is twenty micrometers.

The Stochastic Secondary Wick

The secondary wick was made of quartz wool fibers and was placed in the liquid reservoir or compensation chamber. The function of the secondary wick was to assist in keeping the primary coherent porous silicon wick supplied with the working fluid. The diameter of the fibers ranged from seven to nine micrometers. The porosity of the secondary wick was approximately ninety-two percent. The secondary wick filled about seventy percent of the compensation chamber volume.

THE LOOP HEAT PIPE MODEL Coherent porous silicon (CPS) wicks can be fabricated with different pore diameters, lengths and center-to-center spacing or pore pitches. Optimizing these dimensions is important and requires an understanding of the application in which the wick is used. In this study the wick is used in the loop heat pipe evaporator, therefore, hydrodynamic and thermal models were developed to optimize the coherent porous silicon wick for the best performance.

Hydrodynamic Steady State Modeling of the Coherent Porous Silicon Wick As a Pump

Figure 7 describes the condition of the fluid flow inside the primary coherent porous silicon wick as being fully developed, laminar, and at steady state with constant properties. Equation (1) describes the hydrodynamic relation between pressure and mass flow rate under earlier assumptions. The minimum pressure will occur when the liquid-vapor interface is located at the top of the wick. This implies that the liquid-vapor interface will be stable and that the wick will remain fully saturated with working fluid. The pressure difference across the interface is predicted by the Young-Laplace equation, Equation (2). The wetting angle is given as zero, (θ = 0°). The pressure values are given in terms of a forward flow condition,

lsatvsat PPPP ,12, >>≥ . To determine the appropriate geometry for

the coherent porous silicon wick, the relation between the build up of pressure across the coherent porous silicon wick is developed and optimized to determine the best possible pore diameter.

−=−

21,

8

rA

LmPP

l

llsat ερ

µ & (1)

rPP lsatvsat

σ2,, =− (2)

From Equations (1) and (2):

−=−

21,

182

rL

A

m

rPP lvsat ν

εσ &

(3)

Optimizing Equation (3) for the optimum radius leads to the following:

+−==

32

, 1820

rLv

A

m

rdL

dPl

vsat

εσ &

σεA

Lvmr l

opt

&4= (4)

FIGURE 7. A Schematic Diagram of the Temperature Distribution Inside the Primary Wick with the Control Volume and the Boundary Conditions.

The result is shown in Figures 8 and 9. The optimum radius calculation is shown below. To develop a mathematical relation that describes the flow in the coherent porous silicon wick, the physics of the problem must be understood. A portion of the applied heat causes evaporation while some of it leaks to the liquid compensation chamber. The latent heat causes liquid to evaporate and increases the pressure. The pressure difference between the evaporator and condenser is developed due to the processes of evaporation and condensation. The circulation of the working fluid begins due to the pressure difference in the loop. By knowing the mass flow rate the pumping abilities of different wicks can be determined. The maximum pressure that the interface can hold occurs when the interface radius of curvature is the same as the pore diameter as shown below.

Heat Transfer and Temperature Distribution across the Coherent Porous Silicon Wick

A full understanding of the manner in which heat is transferred within the wick is required to understand the parameters that affect the design of the coherent porous silicon wick. The working fluid has to reach the compensation chamber as a sub-cooled liquid to prevent the formation of bubbles at the inlet to the compensation chamber or in the transport line tubing system. If a bubble forms inside the tubing system or just before the inlet to the compensation chamber, the fluid flow could be blocked and cause bulk flow movement. This type of movement reduces the performance of the loop heat pipe because the flow is interrupted and the performance of the loop heat pipe will exhibit some periodic behavior.

A special concern is given to ensure that the working fluid is in the sub-cooled state at the inlet to the compensation chamber. In the compensation chamber, the sub-cooled liquid begins to gain heat at a constant pressure. If the working fluid enters the compensation chamber as a sub-cooled liquid, the liquid will then gain the heat needed to reach the saturation conditions and saturated liquid will enter the primary wick. It is very important to understand how heat is transferred in the primary wick. Heat is conducted through the coherent porous silicon wick in the direction opposite to that in which heat is carried by the working fluid. Local temperature equilibrium is assumed, for example, the temperature is assumed to be constant in every cross section in the wick. Figure 7 and the previous discussion lead to the following energy equation.

0=∂∂

−−∂∂

+ ∆+∆+ x

effxxpxx

effxp x

TAkTCm

x

TAkTCm && (5)

The same problem was tackled in a previous work (Hamdan et al, 2002). Following the same argument for the boundary conditions, the solution procedure will lead to results similar to those previously reported. The boundary conditions combined with Equation (5) produce the temperature distribution in the coherent porous silicon wick, Equation (8).

(a) Heat flux at the upper side of the coherent porous silicon wick is known.

FIGURE 8. Pressure Across the Wick is Plotted Versus the Pore Size for Different Heat Fluxes, (Hamdan, 2002).

FIGURE 9. Pressure Across the Wick is Plotted Versus the Pore Size for Different Wick Porosities, (Hamdan, 2002).

Q = 106 W/cm2

Q = 107 W/cm2

Q = 108 W/cm2

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

1E+2

1E+3

1E+4

1E+5

1E+6

1E+7

Pres

suer

Bui

ld U

p, d

P, (

Pa)

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

1E+4

1E+5

1E+6

1E+7

1E+8

Pre

ssue

r B

uild

Up,

dP,

(Pa

)

= 80 %

= 50 %

= 20 %

= 5 %

fgLx

eff hmQx

TAk &−=

∂∂

=

(6)

(b) Temperature at the bottom of the coherent porous silicon wick is known. In other words, the temperature of the compensation chamber is known, for example, the saturation temperature of the condenser is known.

( ) ccTxT == 0 (7)

( ) ( )1−

−+= − axaL

p

fg

pcc ee

C

h

Cm

QTxT

& (8)

The last equation represents the temperature distribution in the coherent porous silicon wick. It shows how heat is transferred through the wick. This equation can be used to calculate the amount of heat leaking to the compensation chamber; it could also be used to determine how much of heat is going into evaporation. Using the derived temperature distribution equation for the wick, one can calculate the heat leaking to the compensation chamber as well as the latent heat of vaporization. To find the heat leak and the latent heat of vaporization as a function of the evaporator temperature, the following limitations were applied.

(1) 0=∂

∂=

xeffleak x

TAkQ (9)

(2) ( ) evapTTxT max,2 == (10)

One can calculate the amount of heat needed for evaporation, the amount of heat leak to the compensation chamber and the amount of heat conducted through the coherent porous silicon matrix, see Equations (11) through (15).

paL

ccvfg

fg

fglatent

Ce

TTh

QhhmQ

−

−+

==

−1

& (11)

−

−=

1aL

ccvpleak

e

TTCmQ & (12)

( )ccvpsensible TTCmQ −= & (13)

sensiblelatenleak QQQQ ++= (14)

−

−+=

−aL

ccvpfg

e

TTCmhmQ

1&& (15)

Optimizing Equation (15) for the optimum wick thickness leads to the following:

( )( )

( )aL

aLccvp eae

TTCmdL

dQ −

−−

−−==

21

10 &

This indicates that to prevent heat leaking to the compensation chamber, one should have a very thick wick, which

satisfies the condition ∞→aL . Note that Ak

Cma

eff

p&= and ( ) fseff kkk εε −+= 1 . By combining Equations

(12) and (13), the following result is obtained.

1−= aL

leak

sensible eQ

Q

By rearranging the last equation and substituting for the value of a , one can represent the wick thickness as function of the heat leak, the sensible heat, the mass flow rate, the thermal capacity of the working fluid, the wick effective thermal conductivity, and the cross section area of the coherent porous silicon wick.

+= 1ln

, leak

sensible

lp

eff

Q

Q

Cm

AkL

& (16)

Thermodynamic Temperature - Entropy Cycle for the Loop Heat Pipe

The loop heat pipe process can be described on a temperature – entropy diagram as shown in Figure 10. At Point (1) vapor begins to develop, the interface moves into the vapor space where it becomes superheated. The superheated vapor pressure increases due to the expansion of the vapor, while the interface prevents the interface from bursting through the wick. The low temperature in the condenser will create a low pressure, which results in a pressure gradient in the vapor line. This pressure gradient forces the vapor to move to the condenser, where it begins to condense. From the evaporator outlet to the condenser inlet the process will be isothermal, since the pipe is assumed to be adiabatic and the pressure drops are due to frictional loss in the vapor line. At the condenser, the superheated vapor will lose the sensible heat until it reaches the saturation condition, when a phase change begins to occur. The vapor starts to condense at the interface. The condensed liquid at Point (4) starts to lose more sensible heat in an isobaric process, Point (5). The sub-cooled liquid will move through the liquid line from the condenser to the lower pressure at the compensation chamber, Point (6). The capillary effect will drive the condensed liquid to the evaporator in an attempt to keep the wick wetted. The flow reaches the inlet of the compensation chamber, where the minimum pressure exists in the loop. The process from Point (5) to (6) will be isothermal since the transport lines are adiabatic and the pressure is decreasing due to frictional head loss. This region has the lowest pressure in the entire loop; therefore, it has a higher chance of initiating vapor bubble formation. This proposes and validates the need for the compensation chamber to be located at that point. The capillary pumped loop has its compensation chamber connected to this point through the liquid transport line, while in the loop heat pipe the compensation chamber is located at that point. From Point (6) to (7) the flow is moving due to the capillary effect of the secondary wick, while there is a pressure drop occurring simultaneously in the secondary wick. It is assumed that the pressure gain due to secondary wick is equal to the pressure drop due to friction in the secondary wick, therefore the process from Point (6) to (7) is assumed to be isobaric. For the loop heat pipe, Point (7) represents the saturation condition in the compensation chamber. A two-phase fluid exists in the compensation chamber, which means the primary wick will not be totally wetted. A secondary wick is introduced to ensure continuous wetting of the primary wick. The wetted primary wick will form and hold a stable interface. This interface will prevent the high-pressure vapor that forms above the interface from bursting the interface through the wick. The maximum pressure that the interface can hold depends upon the radius of curvature of the pores as well as the solid and liquid properties. It is obvious that Point (1) will have the maximum pressure in the loop while Point (7) will have the minimum. The quality of Point (7) depends on the size of the compensation chamber. It is necessary to have the pressure difference between Points (1) and (7) less than or equal to the capillary pressure developed in the wick. It is also important to maintain Point (6) in liquid saturation conditions or in the sub- cooled region to prevent any bubble formation in the liquid line.

The most important process in the loop occurs from Point (6) to Point (1), which takes place in the primary wick. It is very difficult to plot in the temperature-entropy diagram since it represents a two-phase fluid with a pressure increase across the interface from the liquid to the vapor. The process from Point (6) to (8) occurs in the superheated liquid region. Point (8) falls between the saturated liquid condition at the high temperature and the spinodal curve.

FIGURE 10. Schematic Diagram of the Loop Heat Pipe Cycle Presented in a Temperature Versus Entropy Thermodynamic Diagram.

THE TEST CELL Figure 11 is a photograph of the test cell assembled for the MEMS based planar loop heat pipe. A 40-mil diamond plated solid drill bit was used to drill two holes with diameters of approximately one-millimeter into the glass top plate. One hole served as the connection to the output vapor line and the other served as an input for the type T thermocouple to monitor the temperature at the top of the primary wick and the connection to the pressure sensor. A 0.7-micrometer layer of gold over a 0.06-micrometer adhesion layer of chromium was deposited onto the backside of the top plate. The top plate was bonded to the coherent porous silicon wick using a commercial product. Sections of stainless steel tubing were connected to the backside of the evaporator top “hot” plate using silver solder. Self-adhesive Kapton strip heaters were attached to the backside of the top plate to simulate the input heat.

FIGURE 11. The Test Cell Structure for the Planar, Microelectromechanical Loop Heat Pipe Under Development.

Figures 12(a) and 12(b) are photographs of the lexan jig containing the evaporator section and the compensation chamber of the loop heat pipe. The two input lines were soldered to the stainless steel plate using silver solder. One line served as the input to the compensation chamber for the working fluid. The other served as an input fill line. Gauge pressure sensors were attached to the input liquid and the output vapor transport lines.

(a) Bottom view. (b) Top view.

FIGURE 12. This photograph displays the lexan jig containing the evaporator section of the loop heat pipe. Figure 13 is a photograph of the condenser. It was constructed from a section of copper tubing seven centimeters in length with a nominal outer diameter of 0.125 inches and a nominal inner diameter of 0.07 inches. The copper condenser tube was silver soldered to the stainless steel fluid transport lines. The cooling water reservoir was constructed from lexan. It had an inner volume of seventy-five cubic centimeters. Type T thermocouples were externally attached to the input and output of the condenser tube.

Coherent Porous Silicon Wick

Top “Hot” Plate

Compensation Chamber

Evaporator

Condenser

FIGURE 13. This photograph displays the copper condenser tubing encased in the lexan cooling water reservoir.

The liquid and vapor transport lines are pictured in Figure 14. The transport lines were constructed from stainless steel tubing with a nominal inner diameter of 1.29 millimeters and a nominal outer diameter of 1.56 millimeters. The length of the liquid transport line was twelve centimeters and the vapor line was seven centimeters in length. An additional stainless steel tube was connected to the liquid transport line for filling purposes.

FIGURE 14. This photograph displays the stainless steel liquid and vapor transport lines.

Anodic Bonding of the Components Surface roughness is a major issue for the bonding of coherent porous silicon samples. Half wafer samples etched in the coherent porous silicon pieces rig using a silicon nitride etch mask proved to be most successful in the anodic bonding experiments. After the pores had been opened into through-holes the backside of the samples were polished using colloidal silica. The gold-coated Pyrex top plates were also polished on the non-coated side. The samples were immersed in Opticlear™ for several hours to remove the lapping wax. The samples were then immersed for five minutes each in acetone, methanol and running de-ionized water. The samples were etched in a buffered oxide etchant (6 NH4F:1 HF) for two minutes and were then immersed in a base clean solution (1 NH4OH: 5 H2O:1H2O2) at eighty degrees Celsius for eight minutes. The gold-coated top plates were bonded to the half-wafer coherent porous silicon samples in a nitrogen atmosphere at four hundred degrees Celsius at a negative 1200 volts. Successful anodic bonding resulted after 1.5 hours. Proprietary processes have also been developed locally for the anodic bonding of silicon to silicon.

Copper Condenser Tube Cooling Water Reservoir

Liquid Transport Line

Vapor Transport Line Fill Line

EXPERIMENTAL ARRANGEMENTS AND OBSERVATIONS Preliminary tests have been conducted to evaluate the performance of the MEMS, planar, closed loop heat pipe system. Comparisons are being made between the theoretical and experimental results. The results of these efforts will be reported at this conference.

Characterization Of the Coherent Porous Silicon Wicks The coherent porous silicon wicks were tested for permeability and capillary pressure prior to packaging. Figure 15 shows a sample wick, which was etched in the pieces rig, encased in its lexan characterization package. Figure 16 is a photograph of the setup for the characterization of the primary wicks.

FIGURE 15. Photograph Showing a Sample Coherent Porous Silicon Wick in the Lexan Characterization Package. (The Wick was Produced Using the Two Piece Etch Rig.)

FIGURE 16. Photograph of the Coherent Porous Silicon Wick Characterization Arrangement. Permeability was evaluated by measuring the change in pressure across the wick at various constant mass flow rates. The working fluid, de-ionized water, was injected into the wicks using a syringe pump. Capillary pressure was evaluated by measuring the change in pressure with respect to a change in the height of the liquid reservoir. From these measurements the effective porosity and pore diameter values were calculated.

FUTURE RESEARCH

The top “hot” plates will be fabricated silicon wafers with a thickness of one millimeter. The bottom liquid reservoir plates will be fabricated from borosilicate glass. The mathematical model has predicted that the optimal wick thickness is 1.9 millimeters. Work is in progress towards the production of thicker coherent porous silicon wicks. The selective patterning and etching of only specific areas will be attempted. The Kapton strip heaters will be replaced by resistive heaters, which will be integrated onto the backside of the top plates to simulate the input heat

Coherent Porous Silicon Wick

Harvard Apparatus 22 Syringe Pump

Characterization Jig Containing the Coherent

Porous Silicon Sample

Microswitch PK8869 4

Pressure Sensor

HP 6177C DC Current Source

HP E2373A Multimeter

source. Effort will be devoted to improving the system and the technique for the production of coherent porous silicon. A greater understanding of the physics of the dissolution reaction will be obtained. Future research also includes the design, construction, and characterization of a permanent filling station, which will inject the working fluid into the evacuated closed system. More tests will be designed and conducted to evaluate the performance of the loop heat pipe system with respect to the mathematical model. The transfer of heat through the various system components, as well as the various temperature and pressure relationships will be analyzed. Data collection will be automated.

CONCLUSIONS

A suitable test cell has been constructed to evaluate the performance of the non-cylindrical closed loop heat pipe system and the validity of the mathematical model. Initial tests produced results similar to those predicted by the theory. Tests, which are in progress, are being conducted in which the actual system performance is being compared to that predicted by the theory. Data collected during these tests will be presented at this conference.

ACKNOWLEDGMENTS Major support for the project was received from the NASA John H. Glenn Research Center at Lewis Field (GRC) under grant NAS3 – 01120 and the Graduate Student Research Program. We would like to thank the Advanced Power and On-Board Propulsion Division of NASA’s Cross Enterprise Technology Development Program, especially Richard Shaltens, Eric Golliher, and Ken Mellott of GRC, for their continued support of and input to the project, and for their many efforts on our behalf. The authors would like to thank the following people for their contributions to this project: Srinivas Parimi and Srikoundinya Punnamaraju (anodic bonding), Sowmya Suryamoorthy (one-millimeter wick fabrication), Ahmed Shuja (etch rate acceleration). Mr. Parimi was also instrumental in the automation of the coherent porous silicon etching system. The authors would also like to thank Vik Kapoor of the University of Toledo for his cooperation and assistance in the component fabrication aspects of the work.

REFERENCES

Cytrynowicz, D., Hamdan, M., Medis, P., Shuja, A., Henderson, H. T., Gerner, F. M., Golliher, E., “MEMS Loop Heat Pipe Based on Coherent Porous Silicon Technology,” in proceedings of Space Technology and Applications International Forum (STAIF-2002), edited by M. El-Genk, AIP Confrence Proceedings 608, Melville, NY, 2002, pp 220 –232.

Hamdan,M., Cytrynowicz, D., Medis, P., Shuja, A., Henderson, H. T., Gerner, F. M., Golliher, E., “Loop Heat Pipe Development by Utilizing Coherent Porous Silicon (CPS) Wicks,” Proceedings of the Eighth Intersociety Conference on Thermal and ThermomechanicalPhenomena in Electronic Systems, IEEE, 0-7803-7152-6/02, San Diego, California, 2002, pp. 457 - 465.

Hölke, A., Pilchowski, J., Henderson, H. T., Saleh, A., Kazmierczak, M., Gerner, F. M., Baker, K., “Coherent Macroporous Silicon as a Wick Structure in an Integrated Micro-Fluidic 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.

Mantravadi, N., “MEMS-Based Development of a Silicon Coherent Porous 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 Applications,” Master’s Thesis, University of Cincinnati, Cincinnati, Ohio, 2000.

Ranganathan, S., Hölke, A., Pilchowski, J., Henderson, H. T., Saleh, A., 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, Pasedena, 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.