As is well known in the early development for space applications, a capillary pumped loop utilizes the capillary forces developed in a fine-pore wick to circulate the working fluid, which can transfer a large amount of heat in a small temperature difference without a need of external pumping power. Stenger [1] at NASA/Lewis U.S.A. first proposed the initial CPL concept in the mid-1960s. But not until the late-1970s serious CPL development for spacecraft applications began. Then in the 1980s the concept of CPL was extensively tested and applied in flight and
3
Commercial experiments. Subsequently in 1993, Ku [2] conducted an overview of capillary pumped loop technology. Characterization of CPL and verification of CPL performance were addressed. He also presented some performance anomalies. These include 1) sudden deprime of evaporators: non-condensable gas accumulation inside the wick or vapor penetration into the wick during startup; 2) evaporator deprime during rapid power step down: very complex interaction among the reservoir, evaporator and condensers; 3) pressure oscillations during steady operation: when accompanied by changes in other conditions it was suspected to cause evaporator deprimes. An example of a deprime is that when the evaporator is starved of liquid due to vapor existing in the liquid side of the wick. The phenomenon of deprimes should be avoided in the CPL operation. If deprimes happen, the system will stop working and will be broken. Maldanik et al. [3] experimentally and analytically investigated physical mechanisms during a CPL startup. They found that the startup process could be characterized by three main stages of different durations. The first stage is a steady growth of temperature and pressure in the loop in the absence of vapor phase. The second stage occurs when boiling in the liquid is initiated, resulting in quick lowering of evaporator temperature and pressure. In the last stage both the temperature and pressure become constant. They also predicted the minimum heat load for a reliable startup. Ku [2] and Meyer et al. [4] also agreed with these findings.
Later Mo et al. [5, 6] experimentally examined how the CPL startup process was affected by applying an electric field to the evaporator wick. They noted that the induced Maxwell stress at the liquid-vapor interface tended to reduce the startup time and improve the system performance. Besides, the depriming phenomenon can also be prevented.
The pressure and temperature oscillations in steady operation of CPLs can significantly affect their performance. In an experimental study Kolos and Herold [7]
4
found that the oscillations of temperature and pressure were resulted from the presence of the bubbles in the evaporator core. Specifically, the bubble growth and collapse cause the oscillations. O’Connell and Ku [8] suggested the factors leading to the unsteadiness of CPL pressure, including wick permeability, size of transport lines, reservoir volume, power input, sink temperature, and reservoir set point. Besides, their experimental results showed that when the transport line diameter was reduced, the pressure oscillation became more severe. O’Connell and Hoang [9] examined effects of wick properties on pressure oscillation. They noted that a more revere pressure oscillation was caused by the reducing permeability of the wick with a smaller pore size.
The evaporator deprime has received some attention. Ku [2] indicated that a sudden change in the input power would lead to an evaporator deprime. A step wise input power was applied to an evaporator by Pouzet et al. [10] to study its fundamental response mechanisms. The oscillations of temperature and pressure following the power change were noted. Using acetone and ammonia as working fluids, Bazzo and Rienl [11] examined the CPL startup and its operating ability when the different heat loads were applied to the capillary evaporator. The experimental results show that heat transfer capacity is better for ammonia. The acetone in the evaporator was heated more quickly for most cases, which in turn resulted in the evaporator dry out. Regarding the CPL operating principles, Ku [12] used thermodynamic diagrams to identify the state of the working fluid in various components of the system for different operating conditions. He also discussed some CPL operating limits, such as the capillary limit, subcooling limit and vapor pressure limit. The capillary limit was considered as the key to the whole system. When the capillary limit was lower than the vapor pressure limit, the system would stop working. Thus the vapor pressure of the system needs to be controlled under the
5
capillary limit in order to prevent the system failure.
More attention has been paid to investigating the heat transfer characteristics for CPL systems. Dickey and Peterson [13] examined the effects of the input power and adverse gravitation with the evaporator located higher than the condenser on the CPL heat transfer characteristics. At the same time, they also developed a computer model for the loop and the results for the steady-state variations of the temperature were presented. Later Liao and Zhao [14-16] moved further to examine the CPL heat transfer capability affected by the capillary structure height and particle size and the entrance temperature. They found that at increasing imposed heat flux the heat transfer coefficient increased to a maximum value and then decreased afterwards.
Strong effects of the height between the liquid tube and vapor tube were experimentally investigated by Meyer et al. [4]. Similarly, Chen and Lin [17] studied the influences of the relative height between the evaporator and condenser, fluid inventory and power input. At increasing height the heat transfer capability of the CPL using FC-72 as working fluid is noted to increase. But, when it increases to certain height, thermal resistance does not reduce further. Instead the thermal resistance remains at a constant. This is attributed to the fact that for an increase in the height the potential energy increases. So, the mass flux in the CPL increases, which in turn results in a higher pressure drop in the liquid and vapor tubes. The experimental results also show that the fluid inventory has an optimal value. Similar conclusion was reached earlier by Miyasaka et al. [18].
Methods to improve CPL heat transfer capability were also examined by some research groups. Pohner and Antoniuk [19] showed that machining the vaporization enhancement grooves (VEGs) into the lands of the evaporator extrusion could raise the heat transfer coefficient. A fibrous slab wick running through the entire condenser tube to serve as a non-condensable gas trap was noted to increase the CPL reliability
6
and operational lifetime too. A similar measure was taken by Muraoka et al. [20] by using a porous structure inside the condenser to stabilize the interface between the liquid and vapor phases. Schweickart et al. [21] showed that adding a mechanical pump could effectively prevent the evaporator deprime.
Using the techniques of MEMS, Kirshberg and Yerkes [22]、Pettigrew and Kirshberg [23] and Meyer et al. [24] fabricated a miniature CPL system on a silicon chip. For traditional CPL an evaporator capillary structure is normally employed to produce capillary forces. However, for the tiny CPL system very small channels are etched on the chip to yield the required capillary forces. More uniform temperature distribution can be obtained and the resulting heat transfer capability increases. When a liquid is in contact with a solid surface in the tiny channel, the extended meniscus can be divided into three parts, the intrinsic meniscus region, the thin film region and the adsorbed region [25]. Park et al. [26, 27] utilized a mathematical model to investigate the transport phenomena and heat transfer characteristics in the thin film region in a micro-channel. They found the length and the thickness of the thin film region decrease exponentially at increasing heat flux, leading to higher capillary force and heat transfer coefficient.
Over the past numerical simulation was also conducted to elucidate the flow and heat transfer within the porous structure in the evaporator. Cao and Faghri [28, 29]
used a numerical analysis to simulate flat-plate type evaporator. The results show that some limitations exist between the wick and heater interface. Specifically, as the babble grows to a size to cover the heater surface, the capillary forces resulting from the porous structure disappear. The liquid-vapor interface is destroyed and this is similar to the boiling limit for traditional heat pipes. Zhao and Liao [30] moved further to consider the effects of the groove size and the thermal conductivity of the capillary structure. They noted that the low thermal conductivity of the capillary
7
structure could result in a very steep temperature gradient at the fin/porous structure interface. Zhao et al. [31] also investigated the flow in a vertical porous channel heated symmetrically along its vertical walls. Very different flow phenomena in the single-phase and two-phase flow motions in the porous structure were noted. In single- phase flow the mass flux increases with the heat flux, but in the two- phase flow the mass flux drops sharply for an increase in the heat flux. Hanlon and Ma [32]
developed a mathematical model to study the heat transfer capability for a sintered wick structure. The results show that the evaporation heat transfer coefficient can be enhanced by reducing the average particle size. They also proposed that thin film evaporation played an important role in the enhancement of evaporating heat transfer.
Some feature and major results obtained is the previous study of CPLs reported in the open literature are given in Table 1.1.
1.3 Objective
The above literature review clearly indicates that the basic flow and heat transfer mechanisms associated with the traditional CPLs have been extensively investigated. Besides, the CPL systems have been used successfully in large heat dissipation systems for spacecraft applications. But a flat shape evaporator has been proposed for CPU heat dissipation [33]. In this experimental study, we intend to develop and test an improved CPL system aiming at the high power density CPU cooling. A typical evaporator design is schematically shown in Fig. 1.2. More specifically, the evaporator consists of a grooved copper plate attached on the wick surface, acting as the heating zone, and the vapor can move in the grooves. Thus, the resistance to the vapor flow is reduced. In addition, it is inefficient to use only the upper surfaces of the fins for heating. The vertical and bottom surfaces of the grooved channels will also be covered by the thin wick. This will increase the heated surface
8
area for liquid vaporization, which is expected to enhance heat transfer in the evaporator. Effects of the input power, liquid inventory, cooling water temperature in the condenser, and relative height between the evaporator and condenser on the heat transfer performance of the CPL will be investigated.
References Working fluid Liquid inventories
Wick properties Applications
Ammonia 3.5 Maldanik et al. (1993)
0 32
775
pore radius = 1.3µm, polyethylene spacecraft
Ammonia 7.9 Meyer et al. (1993)
0 58
1129
pore radius = 1.5µm, sintered nickel
powder spacecraft R-134a 0.247
Mo et al. (1999)
0 26.5
50
pore radius = 1.5µm, polyethylene spacecraft
R-134a 3.21 Mo et al. (2000)
0 26
650 pore radius = 1.5µm, polyethylene spacecraft Ammonia 1.27
Ku (1996)
0 30
500
pore radius = 13, 16µm spacecraft
Ammonia 1.27 O’Connell and Hoang (1996)
0 30
500
pore radius = 8.3, 13.02, 16.12,
19.08µm spacecraft R-134a 5
Pouzet et al. (2004)
0 33
600
pore radius = 20µm spacecraft
Acetone, Ammonia 2.8 Bazzo and Riehl (2003)
0 30
100 pore radius = 20µm electronics cooling Ammonia 2.87
Dickey and Peterson (1994)
0 58
130
pore radius = 1~1.5µm, sintered
nickel powder electronics cooling ( )
Table 1.1 Summary of some features in previous CPL studies.
(℃)
9
References Working fluid Liquid inventories
Wick properties Applications
Water 32 Liao and Zhao (1999)
-0.5 168
887
particle diameters = 0.55, 1.09, 1.99,
2.56 mm electronics cooling Water 32
Liao and Zhao (2000)
1 144.6
256 particle diameters = 2.0 mm electronics cooling Water 25.9
Zhao and Liao (2000)
-3.5 135
740
particle diameters =1.09 mm electronics cooling
FC-72 2.5 Chen and Lin (2001)
50% 12 83
40 pore radius =10µm electronics cooling HCFCs-142b 7
Miyasaka et al. (1995)
40%
0 38
43.7
Capillary pipe diameters =2mm electronics cooling
Ammonia 3.16 Pohner and Antoniuk (1991)
0 32
500 pore radius = 20µm spacecraft Ethanol 3.81
Muraoka et al. (1998)
0 60
120
Sintered bronze electronics cooling
Water 375 Kirshberg and Yerkes (2000)
0 100
7.5
MEMS Groove height/width
=50/50µm electronics cooling Water 375
Pettigrew and Kirshberg (2001)
0 100
7.5
MEMS Groove height/width
=50/50µm electronics cooling ( )
References Working fluid Liquid inventories
Wick properties Applications
Water 224 Meyer and Dasgupta (2003)
0 100
4.8
MEMS Groove height/width
=30/22µm electronics cooling Water 4.12
Zhao and Liao (1994)
0 100
123.6 particle diameters = 0.55 mm humidistat Methanol 25.62
Tsai et al. (2005)
0 98
50.00
Copper screen electronics cooling
Water 35 Maydanik et al. (2005)
0 100
140
pore radius = 1~10µm, sintered
copper powder electronics cooling
,max
Tevap
∗ denotes maximum mean evaporator temperature allowed in the experiment.
Hg
∗ denotes condenser-evaporator relative height.
( )
Hg∗ cm Tevap,max∗
( )
( )
''
2 , m a x
, m a x e
e
q W
c m
Q W
Table 1.1 Continued (2)
(℃)
11
Fig.1.1 Schematic of a Capillary Pumped Loop [37].
EVAPORATOR
Qin
VAPOR
LIQUID