An absorption based miniature heat pump system for electronics cooling

An absorption based miniature heat pump system for electronics cooling

international journal of refrigeration 31 (2008) 23–33 available at www.sciencedirect.com w w w . i i fi i r . o r g journal homepage: www.elsevier...

968KB Sizes 5 Downloads 216 Views

international journal of refrigeration 31 (2008) 23–33

available at www.sciencedirect.com

w w w . i i fi i r . o r g

journal homepage: www.elsevier.com/locate/ijrefrig

An absorption based miniature heat pump system for electronics cooling Yoon Jo Kim*, Yogendra K. Joshi, Andrei G. Fedorov The George W. Woodruff School of Mechanical Engineering, Georgia Institute of Technology, Atlanta, GA 30332, United States

article info

abstract

Article history:

The development of an absorption based miniature heat pump system is motivated by the

Received 16 April 2007

need for removal of increasing rates of heat from high performance electronic chips such

Received in revised form

as microprocessors. The goal of the present study is to keep the chip temperature near am-

23 May 2007

bient temperature, while removing 100 W of heat load. Water/LiBr pair is used as the work-

Accepted 10 July 2007

ing fluid. A novel dual micro-channel array evaporator is adopted, which reduces both the

Published online 17 July 2007

mass flux through each micro-channel, as well as the channel length, thus reducing the pressure drop. Micro-channel arrays for the desorber and condenser are placed in intimate

Keywords:

communication with each other using a hydrophobic membrane. This acts as a common

Cooling

interface between the desorber and the condenser to separate the water vapor from LiBr

Component

solution. The escaped water vapor is immediately cooled and condensed at the condenser

Electronic

side. For direct air cooling of condenser and absorber, offset strip fin arrays are used. The

Absorption system

performance of the components and the entire system is numerically evaluated and

Water-lithium bromide

discussed. ª 2007 Elsevier Ltd and IIR. All rights reserved.

Design Evaporator Microchannel Calculation Performance

Pompe a` chaleur miniaturise´e a` absorption pour les applications e´lectroniques Mots cle´s : Refroidissement ; Composant ; E´lectronique ; Syste`me a` absorption ; Eau-bromure de lithium ; Conception ; E´vaporateur ; Microcanal ; Calcul ; Performance

1.

Introduction

Recent advances in semiconductor technologies are accompanied by an accelerated increase in power density levels

from high performance chips such as microprocessors. According to the International Technology Roadmap for Semiconductors (ITRS), these chips are expected to have an average heat flux of 64 W cm2, with the maximum

* Corresponding author. E-mail address: [email protected] (Y.J. Kim). 0140-7007/$ – see front matter ª 2007 Elsevier Ltd and IIR. All rights reserved. doi:10.1016/j.ijrefrig.2007.07.003

24

international journal of refrigeration 31 (2008) 23–33

Nomenclature A Bo COP dh f G h i j k _ m N Nu P P~ Pe Pr q q00 R Re s St T t u v x xm xv z

area (m2) Boiling number, q00 =Gl coefficient of performance hydraulic diameter (m) friction factor mass flux of channel flow (kg m2 s1) convective heat transfer coefficient (W m2 K1) specific enthalpy (J kg1) Colburn j factor, St Pr2=3 thermal conductivity (W m1 K1) mass flow rate (kg s1) mass flux at the phase interface or membrane (kg m2 s1) Nusselt number, hdh =k pressure (Pa) perimeter (m) Peclet number, dh u=a Prandtl number, n=a heat transfer rate (W) heat flux (W m2) gas constant (J mol1 K1) Reynolds number, rudh =m channel or fin pitch (m) Stanton number, h=rucp temperature (K) fin thickness (m) velocity (m s1) specific volume (m3 kg1) mass fraction of LiBr mole fraction of LiBr vapor quality axial coordinate (m)

junction temperature requirement of nearly 90  C, by the year 2009 (ITRS, 2005). Conventional chip packaging solutions, which use convective air-cooling techniques, are facing difficulties in removing such a high heat flux under the limited space allocated to thermal management. This is particularly true for portable electronics and systems operating in harsh environments. A variety of novel alternative thermal solutions for electronics cooling have been reported and briefly reviewed in Table 1. The cooling systems can be categorized into passive and active. The passive cooling systems utilize capillary or gravitational force to circulate the working fluid, while the active cooling systems are driven by a pump or a compressor for augmented cooling capacity and improved performance. Also, the active systems driven by a compressor can be called refrigeration/heat pump systems, which may offer further increase in power removal by insertion of a negative thermal resistance into the heat flow path (Mongia et al., 2006). This study aims at developing the design of an absorption system, which can remove 100 W of heat from electronic components with the chip junction maintained at room temperature. Among different candidate technologies,

Greeks a b d f ho k l m n r

thermal diffusivity (m2 s1) aspect ratio membrane thickness (m) two-phase frictional multiplier overall surface efficiency mass transfer coefficient (m s1) latent heat of vaporization (J kg1) dynamic viscosity (Pa s) kinematic viscosity (m2 s1) density (kg m3)

Subscripts a air d desorber e evaporator h heating fluid l liquid phase m membrane r refrigerant ref reference rm interface between refrigerant (water) and hydrophobic membrane s solution sat saturation sm interface between solution (LiBr þ water) and hydrophobic membrane triple triple point v vapor phase w wall wc cooling side wall wh heating side wall Superscript * modified

the absorption refrigeration offers the compactness, relative ease of scalability to varying cooling load, and relatively high COP, making it an attractive option for cooling of high performance electronics (Suman et al., 2004). Specifically, such a system can be considered as one of the candidates for cooling of the CPU in desktop PCs and also as a wearable cooling system, e.g., for workers exposed to hazardous materials, police wearing body armor, and military personnel exposed to nuclear, biological or chemical warfare agents (Drost and Friedrich, 1997). The working fluid is the water/ LiBr pair, where water and LiBr are used as refrigerant and absorbent, respectively.

2. Principal features of absorption based heat pump system Fig. 1 shows schematic diagram of an absorption based heat pump system, which mainly consists of an evaporator, an absorber, a desorber, a condenser, a liquid pump and expansion devices. One of the major advantages of such heat pump system is the utilization of waste heat around 90  C, which brings about significant reduction in operating costs

25

international journal of refrigeration 31 (2008) 23–33

Table 1 – Various thermal solutions for electronics cooling Type Pal et al. (2002)

Thermosyphon

Performance

Working fluid

Maximum cooling capacity: 80 W, average junction-to-ambient thermal resistance: 0.40 K W1 Maximum cooling capacity: 80 W, average junction-to-ambient thermal resistance: 0.95 K W1

Water

Remarks –

PF5060

Maydanik et al. (2005)

Loop heat pipe

Minimum junction-to-ambient thermal resistance: 0.58 K W1 at cooling capacity: 130 W

Ammonia and water

10–20 g

Jiang et al. (2002)

Electroosmotic pumping

Maximum cooling capacity: 38 W, junction temperature below: 120  C, junction-to-ambient thermal resistance: 2.5 K W1

Water



Wei and Joshi (2004)

Stacked micro-channel

Reduced thermal resistance and pressure drop

Water

Numerical analysis

Bintoro et al. (2005)

Impinging jet

Maximum cooling capacity: 200 W, maximum junction temperature: 90  C

Water

Substantial system volume

Fan et al. (2001)

Thermoelectric micro-cooler

Maximum cooling power density: 1000 W cm2



40  40 mm2

Mongia et al. (2006)

Vapor compression heat pump

Maximum cooling capacity: 50 W, junction-to-fluid thermal resistance: 0.25 K W1, minimum COP: 2.25

Isobutane

Notebook computer cooling

Drost and Friedrich (1997)

Absorption heat pump

Maximum cooling capacity: 350 W

Water/LiBr

Expected evaporator temperature over 40  C

and energy savings. The only component of this system with moving mechanical parts is the liquid pump, so that relatively quiet operation is possible and no lubricant is used. The coefficient of performance of absorption based heat pump system is defined as the heat removal capability of evaporator per heat/power supply to desorber, i.e., qe (1) COP ¼ qd

Fig. 1 – Schematic diagram of an absorption based heat pump system.

Power consumption of liquid pump is usually negligibly small. The cyclic process for the refrigerant loop is exactly the same as that of vapor compression refrigeration (heat pump) system, except the vapor compressor is replaced with a ‘chemical compressor’ which consists of absorber, liquid pump, solution heat exchanger and expansion device. The pressurization process in the chemical compressor starts in the absorber, where the refrigerant vapor from evaporator (state point 2) is exothermically condensed and absorbed into the strong LiBr solution (state point 10), resulting in weak LiBr solution at state point 5. Following the absorber, the LiBr solution is pressurized by the liquid pump. The solution heat exchanger preheats the weak LiBr solution of state point 6 to state point 7 using high temperature strong LiBr solution flow from the desorber. In the desorber, high pressure and high temperature superheated refrigerant water vapor is generated and desorbed from the weak LiBr

26

international journal of refrigeration 31 (2008) 23–33

solution and returns to the refrigerant loop. Meanwhile, the LiBr solution becomes the strong solution and returns to the absorber through solution heat exchanger and expansion device in sequence, which completes solution loop or chemical compression cycle. The condensation/absorption process at the absorber and the vaporization/desorption process at the desorber make use of liquid pump feasible to support pressure difference between the condenser and the evaporator. Although the presence of absorber and desorber increases the overall system volume, the displacement volume and power consumption for liquid compression are much smaller than those for vapor compression, so that chemical compression is more favorable to miniaturization than vapor compression where the volume size is limited by the electric motor.

3. Design concepts and performance evaluation of components Analysis of the miniaturized micro-channel absorber is important due to its effect on the system performance. We modified the predictive model of Goel and Goswami (2005) for the liquid-cooled and air-cooled water/LiBr micro-channel absorber. It is assumed that the cooling plate is fully wet by the liquid film and thermodynamic equilibrium exists at the interface between aqueous LiBr solution and refrigerant (water) vapor. Also, mass transfer driven by thermal and pressure difference is regarded to be negligible. The effects of heat losses and non-condensable gases are not considered. Since the configuration of miniaturized micro-channel absorber is straightforward and any new physical concepts are not introduced, we only include the results of absorber analysis in the following calculations, and the detailed description of the absorber analysis (Kim et al., submitted for publication) is not addressed in this paper. Most of the attention is focused on the other critical components, to which novel concepts are adopted, for heat pump system relevant to electronics cooling. These are evaporator and condenser/desorber.

3.1.

Evaporator

One of the major drawbacks of two-phase flow inside microchannels is the large pressure drop through the evaporator. Narrow flow passages, large specific volume of the vapor phase and friction between phases offer more resistance to fluid transport. Vapor phase specific volumes and (dT/dP)sat, which is evaluated from Clausius–Clapeyron equation (Smith et al., 1996), of various refrigerants are plotted in Fig. 2. As shown, the vapor phase specific volume of water is drastically larger than that of other refrigerants, resulting in fast moving vapor phase with large velocity and thus significant pressure drop in evaporator. The pressure drop of two-phase water flow in micro-channel may cause drastic reduction in evaporator performance or blockage of microchannel with ice because of the high sensitivity of saturation temperature on pressure change and high triple point (Ttriple, water ¼ 273.16 K). To this end, configuration originally proposed by Kandlikar and Upadhye (2005) is adopted as the evaporator (Fig. 3) and

Fig. 2 – Vapor phase specific volumes and (dT/dP)sat of various refrigerants at 20  C evaluated using Clausius– Clapeyron equation (Smith et al., 1996) and thermal properties from REFPROP 6.0 (McLinden et al., 1998).

compared with single-pass micro-channel evaporator in this study. By providing dual-pass for the refrigerant flow inside micro-channels, both the flow length and the mass flux of each micro-channel are reduced by half so that the pressure drop can be significantly reduced. Dimensions and operating conditions of both single-pass and dual-pass micro-channel evaporators depicted in Fig. 3 are listed in Table 2. It is assumed that the refrigerant flows can be uniformly distributed without inlet headers for both configurations. Dual-pass evaporator has twice the number of channels and half the channel length compared to a single-pass evaporator. Considering the thermal and fluid flow in micro-channel evaporator, energy and momentum conservation equations for refrigerant flow, and an energy conservation equation for micro-channel wall are established based on the micro-channel heat sink model of Koo et al. (2001). di _  ho hr P~r ðTw  Tr Þ ¼ 0 m dz

(2)

" #     v 2 dP 2fl G2 ð1  xv Þ 2 ð1  xv Þ2 2 d ðx Þ ¼  þ fl þ G dz dh rl dz arv ð1  aÞrl

(3)

  d dTw kw Aw  ho hr P~r ðTw  Tr Þ þ q00 s ¼ 0 dz dz

(4)

In regard to the heat exchange between refrigerant and microchannel wall where a uniform heat flux is applied on one side, the overall surface efficiency ho is introduced (Koo et al., 2001). The vertical channel wall is considered as a fin and thus uniform wall temperature distribution is assumed. Several correlations for two-phase heat transfer coefficients have been proposed (Lazarek and Black, 1982; Tran et al., 1997; Lee and Lee, 2001; Warrier et al., 2002; Yu et al., 2002; Lee and Mudawar, 2005a). Considering the Boiling number ðBo ¼ q00 =GlÞ range (w1  102) in this study, the predictive

international journal of refrigeration 31 (2008) 23–33

 f Re ¼ 24 1  1:3553b þ 1:9467b2  1:7012b3 þ 0:9564b4   0:2537b5

27

ð5Þ

The governing equations (2)–(4) are integrated and then discretized using the upwind scheme (Patankar, 1980). The discretized equations are simultaneously iteratively solved using successive under-relaxation (SUR). The properties of water are determined using REFPROP 6.0 (McLinden et al., 1998). Fig. 4 illustrates the temperature and pressure variation under normal heat load condition. For single-pass configuration, larger pressure drop (1.21 kPa) resulted, and the corresponding refrigerant saturation temperature drop was 7.8  C. Temperature difference between the refrigerant and the wall was about 10  C at the inlet and then, due to the decreased heat transfer coefficient, increased up to 13  C at the outlet of evaporator. A decrease in the heat transfer coefficient is attributed to the saturation temperature drop because the nucleate boiling, which is the dominant heat transfer mechanism in high Bo two-phase flow, is suppressed by the increase of density, velocity and viscosity differences between liquid and vapor phases. On the other hand, for dual-pass configuration, only 0.26 kPa pressure drop and 1.4  C temperature drop can be observed from Fig. 4b. Also, the wall temperature of 35  C was maintained along the length of the evaporator. Both the half channel length and half mass flux of dual-pass configuration contributed to reducing the pressure drop by more than one-fourth. The large pressure drop of singlepass configuration degrades the performance of the evaporator and the entire system. Due to the reduced inlet pressure into the chemical compressor, to supply the same refrigerant mass flow rate, more heat must be provided to the desorber and rejected from the absorber, resulting in an increase in the load assigned to the system. A 1000 W of heat load was assigned to both single-pass and dual-pass evaporators as an extreme operating condition. Again, the dual-pass configuration has more uniform temperature and pressure distribution along the evaporator length, and the pressure drop of 3.6 kPa is observed from Fig. 5. Since the inlet refrigerant temperature was 50  C, the pressure drop increase was less than the increases of mass flux and heat flux. Also, since the saturation temperature sensitivity to pressure variation, (dT/dP)sat, is only 0.0016 K Pa1 at 50  C, which is one-fourth of that at 25  C, the temperature drop observed was only 6.8  C. For both configurations, the surface temperatures were well below 80  C. Fig. 3 – Conceptual diagram of (a) single-pass and (b) dualpass micro-channel evaporator.

model of Yu et al. (2002) is adopted. The two-phase multiplier correlation of Lockhart and Martinelli (1949) is incorporated with the C value proposed by Lee and Mudawar (2005b) and used in Eq. (3). Also, the void fraction model of Rouhani and Axelsson (1970), one of the drift flux models particularly effective at low to medium reduced pressure, is adopted. The friction factor applicable to single phase laminar flow in a rectangular channel is given by:

3.2.

Condenser/desorber

Conventional configurations of condenser and desorber in absorption heat pump system are too bulky to be fitted into small scale envelope for electronics cooling. Fig. 6a shows a novel configuration of a combined micro-channel condenser and desorber. Micro-channel heat exchangers for the desorber and condenser are attached to each other on one side, and a hydrophobic membrane is placed between the desorber and the condenser. For air side heat transfer enhancement, offset strip fins are installed on the other side of the condenser. The desorber is heated by waste heat or other

28

international journal of refrigeration 31 (2008) 23–33

Table 2 – Dimensions and operating conditions of micro-channel evaporators Parameters

Single-pass evaporator Normal

Channel width (mm) Channel depth (mm) Channel length (mm) Channel wall thickness (mm) Number of channels Evaporator length (mm) Evaporator width (mm) Mass flow rate (g s1) Heat transfer rate (W) Inlet temperature ( C) Inlet vapor quality Material

0.5 1.0 30.0 0.3 38 30.0 30.7 0.042 100 25 0.02 Aluminum

Dual-pass evaporator

Extreme

Normal

Extreme

0.5 1.0 30.0 0.3 38 30.0 30.7 0.45 1000 50 0.02

0.5 1.0 15.0 0.3 76 30.0 30.7 0.042 100 25 0.02 Aluminum

0.5 1.0 15.0 0.3 76 30.0 30.7 0.45 1000 50 0.02

available heat source, and the refrigerant inside the LiBr solution is vaporized. The surface tension and capillary actions upon interaction with the hydrophobic membrane prevent the liquid solution from flowing through the membrane pores. Only the generated vapor can escape from the desorber to the

condenser through the hydrophobic membrane, which realizes the liquid–vapor separation. The escaped refrigerant vapor is immediately cooled in the micro-channel condenser, and the released heat is dissipated to the ambient by the aid of offset strip fin array.

Fig. 4 – Temperature and pressure variations of (a) singlepass and (b) dual-pass micro-channel evaporators under normal heat load conditions listed in Table 2.

Fig. 5 – Temperature and pressure variations of (a) singlepass and (b) dual-pass micro-channel evaporators under extreme heat load conditions listed in Table 2.

international journal of refrigeration 31 (2008) 23–33

29

corresponding to the solution side membrane temperature, Tsm. Since the pressure of the generated vapor is higher than any other bulk fluid pressure, which is locally non-equilibrium, the pressure difference between the solution side and the refrigerant side of the membrane acts as a driving force for vapor to escape through the hydrophobic membrane. However, due to the dissolved LiBr, the saturation vapor pressure must be reduced. Reverse osmosis caused by concentration difference exerts a negative effect on the driving force. Consequently, the final form of driving force equation is given by:      m DP ¼ Psat ðTsm Þ 1  xm sm  Prm 1  xrm  m m ð6Þ  Psat ðTsm Þxsm  Prm xrm The analytical model for membrane distillation of Schofield et al. (1987) is modified to predict the heat and mass transfer behaviors of the micro-channel condenser/desorber which is conceptually depicted in Fig. 6c. Energy and momentum conservation equations for air flow through offset strip fin array and an energy equation for heating fluid are additionally established. The equations as a set of governing equation for condenser/desorber are listed below. _r _s dm dm ¼ ¼ P~s N dz dz

(7)

 d _ s xs ¼ 0 m dz

(8)

_h m

dih  hh P~h ðTwh  Th Þ ¼ 0 dz

_ s is Þ _s dðm dm ¼ hs P~s ðTwh  Ts Þ þ hs P~s ðTsm  Ts Þ þ ism dz dz

(10)

_ r ir Þ _r dðm dm ¼ hr P~r ðTr  Twc Þ þ hr P~r ðTrm  Tr Þ þ irm dz dz

(11)

_a m

dia  ho ha P~a ðTwc  Ta Þ ¼ 0 dz

The details of mass transfer mechanism through the hydrophobic membrane pore are described in Fig. 6b. Continuous heating and cooling on each side support the temperature gradient inside the condenser/desorber micro-channel combination, as depicted in Fig. 6b. When the refrigerant vapor is generated at the surface of the hydrophobic membrane, its pressure, Psm, is the saturation vapor pressure of water

(12)

hs P~s ðTs  Tsm Þ ¼

_s dm km ðirm  ism Þ þ P~r ðTsm  Trm Þ dz d

(13)

hr P~r ðTrm  Tr Þ ¼

km ~ Pr ðTsm  Trm Þ d

(14)

  dP 2fa G2a ¼  dz dh;a ra Fig. 6 – (a) Conceptual diagrams, (b) mass transfer mechanism and (c) control volume of micro-channel condenser/desorber.

(9)

(15)

The pressure drop of refrigerant in the condenser is not considered because (i) the condensation process is less dissipating in nature than boiling, (ii) the saturation temperature in the condenser is relatively high, and (iii) the mass flux through each condenser micro-channel is significantly smaller than that in the evaporator due to the larger volume of condenser. As a result, the effect of the specific volume of vapor phase on the pressure drop is not significant. It is assumed that the vaporization takes place at the surface of the hydrophobic membrane, or vapor generated at any other location moves to the hydrophobic membrane surface without appreciable heat or mass exchange with bulk liquid. Also, the

30

international journal of refrigeration 31 (2008) 23–33

concentration polarization, which accounts for the concentration difference between the bulk fluid and the membrane interface, is neglected. Under the pressure difference given, mass transfer through the membrane is a combination of Knudsen diffusion and Poiseuille flow through the membrane. Schofield et al. (1990) proposed a membrane permeability model for Knudsen– Poiseuille transition region and thus the mass flux through the membrane is defined as: N ¼ azb DP

(16)

where z is the dimensionless membrane mean pressure (0.5(Psm þ Prm)/Pref), a is the membrane permeation constant, and b is 0 for Knudsen diffusion and 1 for Poiseuille flow. Heat transfer coefficients for laminar rectangular microchannel flow of heating fluid, liquid LiBr solution and single phase flows of subcooled and superheated refrigerant are determined using empirical correlation from Shah and London (1978). For the condensation process, the predictive model of Shah (1979) was used to evaluate the heat transfer coefficient during condensation inside small rectangular channel. Along the membrane, the heat transfer coefficients of LiBr solution and refrigerant are modified to account for mass transfer through the phase interface (Treybal, 1968). The air side heat transfer coefficient and friction factor for laminar flow through the offset strip fin array and the modified heat transfer coefficients involving mass transfer are determined by j and f correlation proposed by Manglik and Bergles (1994). The solution procedure for the governing differential equations (7)–(15) is basically the same as for the evaporator simulation, except that at the beginning of the desorption, a hybrid scheme is used to consider the diffusion caused by very small mass flow rate and thus Peclet number (Pe) of the vapor flow (Patankar, 1980). The properties for LiBr solution are evaluated by Yuan and Herold (2005). The dimensions and operating conditions used in the simulations are given in Table 3. Fig. 7 shows the overall performance of the micro-channel condenser/desorber under reference condition, for a refrigerant flow rate at the exit of the condenser of 0.042 g s1. From Fig. 7 it is observed that the heating fluid transfers its energy, around 133 W, to liquid

LiBr solution and thus the temperature decreases along its flow direction. A small amount (around 11 W) of this energy was used to increase the temperature of the LiBr solution to near the inlet temperature of the heating fluid, while most of the energy was used to vaporize the refrigerant. It is interesting that around 19 W of heat was dissipated by conduction through the membrane, which represents internal energy leakage caused by indirect sensible heat exchange between the condenser and the desorber. The remaining 122 W of heat was transferred to the ambient, condensing the refrigerant to subcooled water and increasing the air temperature to 39  C. By combining the condenser and desorber into one component, significant size reduction is realized and a high temperature difference is sustained by the heating and cooling on each side so that the mass transfer through the membrane is significantly enhanced. However, as mentioned, considerable amount of internal energy loss takes place via conductive coupling through a membrane, which increases the heat load of desorber and condenser. The air side pressure drop through offset strip fin array was negligibly small (34 Pa).

4.

System performance

To determine the COP, temperature/pressure at each state point, mass flow rate, and capacity of each component, a simple analysis is carried out by applying lumped mass and energy conservation equations to each state point using inlet and outlet information of each component. The performance of micro-channel structure absorber with integrated array of offset strip fin (Fig. 8) was numerically evaluated using a model similar to the one used for the condenser/desorber. For the complete absorption of refrigerant water into LiBr solution, 106 W of heat was rejected to ambient and the air side pressure drop was 85.4 Pa. The isentropic efficiency of the pump is assumed to be 100%, and the expansion process is considered isenthalpic. The operating conditions at each point are listed in Table 4, and the Du¨hring plot for absorption based miniature heat pump system is provided in Fig. 9. State point 3 is a fictitious point representing condenser inlet. The system operating

Table 3 – Dimensions and operating conditions of micro-channel condenser/desorber Dimensions Parameters Channel width (mm) Channel depth (mm) Channel length (mm) Channel wall thickness (mm) Number of channels Fin width (mm) Fin height (mm) Fin strip length (mm) Fin thickness (mm) Number of fins Overall length (mm) Overall width (mm) Material

Operating conditions Values 0.5 0.3 100 0.3 125 2.1 22 32.5 0.7 35 100 100.3 Aluminum

Parameters

Values 1

Heating fluid mass flow rate (g s ) Heating fluid inlet temperature ( C) Solution mass flow rate (g s1) Solution inlet temperature ( C) Solution pressure (Pa) Solution inlet concentration Refrigerant mass flow rate (g s1) Refrigerant pressure (Pa) Air velocity (m s1) Air inlet temperature ( C)

1.0 90 0.296 55.5 12,344 0.403 0.042 12,344 3.5 30

international journal of refrigeration 31 (2008) 23–33

31

Fig. 8 – Conceptual diagram of micro-channel absorber.

as a result of evaluation of the evaporator, absorber and condenser/desorber performances. Also, the heating/cooling capacities of evaporator, absorber, desorber and condenser are found to be 100 W, 110 W, 135 W, and 125 W, respectively, resulting in the COP of 0.74. The parasitic internal heat leakage inside the condenser/desorber was estimated as 20 W. Since the operating pressure of the evaporator is higher than that of a conventional absorption based heat pump system, the predicted COP of the absorption refrigeration system is higher than that of conventional absorption system. In addition, the higher evaporator operating pressure allows the system cycle

Fig. 7 – (a) Temperature, (b) mass flow rate and (c) LiBr mass fraction variations of micro-channel condenser/desorber under reference operating conditions listed in Table 3.

curve (solid line) deviates from the constant mass fraction line of LiBr, which means the condenser outlet and desorber inlet are in a subcooled state. The operating conditions listed in Table 4 are approximately identical to those obtained earlier

Fig. 9 – Du ¨ hring plot for absorption based miniature heat pump system.

32

international journal of refrigeration 31 (2008) 23–33

Table 4 – Operating conditions of each state point State point 1 2 3 4 5 6 7 8 9 10

_ (g s1) m

x (kg kg1)

i (J kg1)

T ( C)

P (Pa)

xv (kg kg1)

0.042 0.042 0.042 0.042 0.296 0.296 0.296 0.254 0.254 0.254

– – – – 0.403 0.403 0.403 0.470 0.470 0.470

162,740 2,543,693 2,662,740 162,740 71,302 71,302 121,941 154,537 95,532 95,532

25.0 23.6 87.5 38.9 35.0 35.0 55.5 71.6 45.8 42.5

3168 2913 12,344 12,344 2913 12,344 12,344 12,344 12,344 2913

0.032163 1 Superheated Subcooled 0 Subcooled Subcooled 0 Subcooled 1.04E05

to be located far from the crystallization line. The solution heat exchanger may contribute to reduce the load assigned to the desorber and absorber, so that the COP can be enhanced and the size of the system can be further reduced. For instance, if the solution heat exchanger has 20 W of heat exchange capacity, then COP can be increased to 0.87 and the heating/cooling capacity of the desorber and absorber can be decreased to 115 W and 90 W, respectively.

5.

Conclusions

The feasibility of absorption based heat pump for cooling of electronics is theoretically evaluated. To miniaturize the system, dual-channel evaporator, micro-channel heat/mass exchanger, and hydrophobic membrane condenser/absorber were incorporated. Relatively low temperature, around 30  C was achievable with heat removal capability of 100 W from electronic components. The representative evaporator has dimensions of 30 mm  30 mm  3 mm, and the suitable condenser/desorber and absorber have identical size of 100 mm  100 mm  45 mm, and are the largest components of the system. Considering the size of the other components including a micro-pump and connecting tubes, the entire heat pump system can be fitted into a 150 mm  150 mm  100 mm size envelope. It should be emphasized that the analytical models and results obtained in this study convincingly demonstrate the feasibility of using miniature absorption refrigeration system for electronics cooling applications, and also describe an expected envelope of design and operating characteristics. However, an experimental validation is needed for quantitative support of our theoretical conclusions.

Acknowledgment The authors are acknowledged the support of the Korea Research Foundation Grant funded by the Korean Government (MOEHRD) (KRF-2005-214-D00232) and Interconnect Focus Center one of the research centers funded under the Focus Center Research Program, a Semiconductor Research Corporation Program.

references

Bintoro, J.S., Akbarzadeh, A., Mochizuki, M., 2005. A closed-loop electronics cooling by implementing single phase impinging jet and mini channels heat exchanger. Appl. Therm. Eng. 25, 2740–2753. Drost, M.K., Friedrich, M., 1997. Miniature heat pump for portable and distributed space conditioning applications. In: IECEC-97, Proceedings of the Thirty-Second Intersociety Energy Conversion Engineering Conference, Honolulu, Hawaii, 27 July–1 August, pt. 2, pp. 1271–1274. Fan, X., Zeng, G., LaBounty, C., Bowers, J.E., Croke, E., Ahn, C.C., Huxtable, S., Majumdar, A., Shakouri, A., 2001. SiGeC/Si superlattice microcoolers. Appl. Phys. Lett. 78 (11), 1580–1582. Goel, N., Goswami, D.Y., 2005. Analysis of a counter-flow vapor flow absorber. Int. J. Heat Mass Transf. 48, 1283–1292. International Technology Roadmap for Semiconductors, Assembly and Packaging, 2005 Ed. Jiang, L., Mikkelsen, J., Koo, J.M., Huber, D., Yao, S., Zhang, L., Zhou, P., Maveety, J.G., Prasher, R., Santiago, J.G., Kenny, T.W., Goodson, K.E., 2002. Closed-loop electroosmotic microchannel cooling system for VLSI circuits. IEEE Trans. Comp. Packag. Technol. 25 (3), 347–355. Kim, Y.J., Joshi, Y.K., Fedorov, A.G. Analysis and simulation of air-cooled microchannel absorber for absorption based miniature electronics cooling system, submitted for publication. Kandlikar, S.G., Upadhye, H.R., 2005. Extending the heat flux limit with enhanced microchannels in direct single phase cooling of computer chips. In: Proceedings of 21st IEEE Semi-Therm Symposium, San Hose, CA, 15–17 March, pp. 8–15. Koo, J.M., Jiang, L., Zhang, L., Kenny, T.W., Santiago, J.G., Goodson, K.E., 2001. Modeling of two-phase microchannel heat sinks for VLSI chips. In: Proceedings of MEMS’01 Conference, Interlaken, Switzerland, January. Lazarek, G.M., Black, S.H., 1982. Evaporative heat transfer, pressure drop and critical heat flux in a small vertical tube with R-113. Int. J. Heat Mass Transf. 25 (7), 945–959. Lee, H.J., Lee, Y.L., 2001. Heat transfer correlation for boiling flows in small rectangular horizontal channels with low aspect ratios. Int. J. Multiphas. Flow 27, 2043–2062. Lee, J., Mudawar, I., 2005a. Two-phase flow in high-heat-flux micro-channel heat sink for refrigeration cooling applications: part II – heat transfer characteristics. Int. J. Heat Mass Transf. 48, 941–955. Lee, J., Mudawar, I., 2005b. Two-phase flow in high-heat-flux micro-channel heat sink for refrigeration cooling applications: part I – pressure drop characteristics. Int. J. Heat Mass Transf. 48, 928–940.

international journal of refrigeration 31 (2008) 23–33

Lockhart, R.W., Martinelli, R.C., 1949. Proposed correlation of data for isothermal two-phase two-component flow in pipes. Chem. Eng. Prog. 45, 39–48. Maydanik, Y.F., Vershinin, S.V., Korukov, M.A., Ochterbeck, J.M., 2005. Miniature loop heat pipes – a promising means for electronics cooling. IEEE Trans. Comp. Packag. Technol. 28 (2), 290–296. Mongia, R., Masahiro, K., DiStefano, E., Barry, J., Chen, W., Izenson, M., Possamai, F., Zimmermann, A., Mochizuki, M., 2006. Small scale refrigeration system for electronics cooling within a notebook computer. In: ITHERM’06, Proceedings of the Tenth Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronics Systems, San Diego, USA, 30 May–2 June, pp. 751–758. McLinden, M.O., Klein, S., Lemmon, E., Peskin, A., 1998. NIST thermodynamic and transport properties of refrigerants and refrigerant mixtures database (REFPROP), Version 6.0. National Institute of Standards and Technology, Gaithersburg, Maryland, USA. Manglik, R.M., Bergles, A.E., 1994. The thermal–hydraulic design of the rectangular offset-strip-fin compact heat exchangers. In: Shah, R.K., Kraus, A.D., Metzger, D. (Eds.), Compact Heat Exchangers. Hemisphere, Washington, pp. 123–150. Pal, A., Joshi, Y.K., Beitelmal, M.H., Patel, C.D., Wenger, T.M., 2002. Design and performance evaluation of a compact thermosyphon. IEEE Trans. Comp. Packag. Technol. 25 (4), 601–607. Patankar, S.V., 1980. Numerical Heat Transfer and Fluid Flow. Hemisphere, Washington, DC. Rouhani, Z., Axelsson, E., 1970. Calculation of void volume fraction in the subcooled and quality boiling region. Int. J. Heat Mass Transf. 13, 383–393. Suman, S., Joshi, Y.K., Fedorov, A.G., 2004. Cryogenic/sub-ambient cooling of electronics: revisited. In: ITherm 2004, Las Vegas, Nevada, 1–14 June.

33

Smith, J.M., Van Ness, H.C., Abbott, M.M., 1996. Introduction to Chemical Engineering Thermodynamics, fifth ed. McGrawHill, New York, NY. Schofield, R.W., Fane, A.G., Fell, C.J.D., 1987. Heat and mass transfer in membrane distillation. J. Membrane Sci. 33, 299–313. Schofield, R.W., Fane, A.G., Fell, C.J.D., 1990. Gas and vapour transport through microporous membranes. I. Knudsen– Poiseuille transition. J. Membrane Sci. 53, 159–171. Shah, R.K., London, A.L., 1978. Laminar flow forced convection in ducts. In: A Source Book for Compact Heat Exchanger Analytical Data. Advances in Heat Transfer, Suppl. 1. Academic Press, NewYork. Shah, M.M., 1979. General correlation for heat transfer during film condensation inside. Int. J. Heat Mass Transf. 22, 547–556. Tran, T.N., Wambsganss, M.W., Chyu, M.C., France, D.M., 1997. A correlation for nucleate flow boiling in small channels. In: Shah, R.K. (Ed.), Compact Heat Exchangers for the Process Industries. Begel House, New York, pp. 353–363. Treybal, R.E., 1968. Mass Transfer Operation. McGraw-Hill, New York. Wei, Y., Joshi, Y., 2004. Stacked microchannel heat sinks for liquid cooling of microelectronic components. J. Electron. Packag. 126, 60–66. Warrier, G.R., Dhir, V.K., Momoda, L.A., 2002. Heat transfer and pressure drop in narrow rectangular channels. Exp. Therm. Fluid Sci. 25, 53–64. Yu, W., France, D.M., Wambsganss, M.W., Hull, J.R., 2002. Twophase pressure drop, boiling heat transfer, and critical heat flux to water in a small-diameter horizontal tube. Int. J. Multiphas. Flow 28, 927–941. Yuan, Z., Herold, K.E., 2005. Thermodynamic properties of aqueous lithium bromide using a multiproperty free energy correlation. Int. J. Heat. Ventilat. Air-Condition. Refrig. 11 (3).