Kuo and Peles 2009 Flow Boiling of Coolant (HFE-7000) Inside Structured and Plain Wall Microchannels

7/25/2019 Kuo and Peles 2009 Flow Boiling of Coolant (HFE-7000) Inside Structured and Plain Wall Microchannels

1/9

C.-J. Kuo

Y. Peles1

e-mail: [email protected]

Department of Mechanical, Aerospace and

Nuclear Engineering,

Rensselaer Polytechnic Institute,

Troy, NY 12180

Flow Boiling of CoolantHFE-7000Inside Structured andPlain Wall MicrochannelsFlow boiling was experimentally studied using coolant HFE-7000 for two types of par-

allel microchannels: a plain-wall microchannel and a microchannel with structured re-entrant cavities on the side walls. Flow morphologies, boiling inceptions, heat transfercoefficients, and critical heat fluxes were obtained and studied for mass fluxes ranging

from G164 kg / m2 s to G3025 kg/ m2 s and mass qualities (energy definition) rang-

ing from x0.25 to x1. Comparisons of the performance of the enhanced andplain-wall microchannels were carried out. It was found that reentrant cavities wereeffective in reducing the superheat at the onset of nucleate boiling and increasing the heattransfer coefficient. However, they did not seem to increase the critical heat flux.DOI: 10.1115/1.3220674

Keywords: flow boiling, subcooled boiling, nucleate boiling heat transfer, critical heatflux, MEMS, microchannel

1 IntroductionFlow boiling is associated with a very high heat transfer coef-

ficient and has been extensively studied for numerous coolingapplications in conventional scale system 18. Because of thecontinuous advances in electronic technology and the correspond-ing power density increase, flow boiling in microdomains hasbeen a topic of great interest in the last decade 916. Key engi-neering parameters that are typically examined include onset ofnucleate boilingONB, heat transfer coefficient, critical heat fluxCHFconditions, flow patterns, and flow instabilities.

In conventional scale, structured reentrant cavities have beeneffective in reducing ONB, increasing heat transfer coefficient,and increasing CHF17,18. Several studies have been conductedto examine the performance of the reentrant cavity at the micro-

scale 14,15,1924. Kosar et al. 14 and Kuo and Peles 15studied the thermal performance and flow boiling patterns inreentrant-cavity microchannel using water. They argued that reen-trant cavities can promote bubble nucleation in microchannels,and thus, enhance heat transfer. Kuo and Peles 25,26conductedstudies of flow boiling instability of water in microchannels, andBhavnani and co-workers2224 conducted similar experimentswith dielectric fluid FC-72. Their results suggest that flow boilinginstability can be mitigated by forming structured reentrant cavi-ties in the channel wall and suppressing the frequently cited rapidbubble growth, which in turn increased CHF value. Certain flowconditions and fluid properties, such as surface tension and massflux have also been shown to affect bubble nucleation in bothmacro- and microscale channels 2731.

The current paper presents a study of flow boiling using coolantHFE-7000 through an array of five parallel microchannels in twotypes of devices: one with reentrant cavities on the side walls andanother with plain side walls. Flow boiling was recorded for bothsubcooled and saturated boiling for mass quality 0.25x1,mass flux ranging from G=164 kg / m2 s to G=3025 kg /m2 s,and heat flux up to 150.4 W / cm2. Flow patterns, boiling incep-

tion, heat transfer coefficient, and CHF were obtained and studied.Comparison between the two microchannel devices are presentedand discussed.

2 Device Overview

The microchannel device consists of five parallel microchan-nels, which are 10,000 m long, 200 m wide, and 250 mdeep, spaced 200 m apart. For the reentrant-cavity microchan-nel, each sidewall encompasses an array of reentrant cavitiesspaced 100 m apart 100 cavities on each side of the10,000 m long microchannel. An acute angle connects the7.5 m mouth to the 50 m inside diameter reentrant body. Ascanning electron microscopeSEMimage of the reentrant cavi-ties is shown in Fig. 1a. In order to minimize ambient heatlosses, air gaps were formed on the two ends of the side walls, andinlet and exit plenums were etched on the thin silicon substrate150 m. On the top, a Pyrex substrate sealed the device andallowed flow visualization. Figure 1b depicts a CAD model ofthe heater and thermistors on the backside of the device. For localtemperature measurement, three thermistors, which are 10 mwide and 300 m long Fig. 1c, were located 3400 m,6700 m, and 10,000 m downstream the channel inlet to-gether with electrical connecting vias. On top of the thermistorlayers, a 1 m silicon oxide layer was deposited for electricalinsulation. A heater was then formed on top of the oxide layer todeliver the heating power to the microchannels.

3 Device Fabrication, Experimental Apparatus, and

Procedures

3.1 Microchannel Fabrication Method. The microelectro-mechanical system MEMS devices were micromachined on apolished double-sided n-type 100 single crystal silicon waferemploying techniques adapted from integrated circuitICmanu-facturing. A 1 m thick high-quality oxide film was deposited onboth sides of the silicon wafer to shield the bare wafer surfaceduring processing and to serve as an electrical insulator. A layer of150 thick titanium was deposited by a CVC 601 sputter depo-sition system and patterned on the backside of the wafer to formthe thermistors. Electrical connectors of 0.2 m aluminum con-taining 1% silicon and 4% copper were subsequently formed inorder to create electrical connections to the thermistors. Follow-

1Corresponding author.Contributed by the Heat Transfer Division of ASME for publication in the JOUR-

NAL OF HEATTRANSFER. Manuscript received November 13, 2008; final manuscriptreceived May 19, 2009; published online October 15, 2009.

Journal of Heat Transfer DECEMBER 2009, Vol. 131 / 121011-1Copyright 2009 by ASME

wnloaded From: http://asmedigitalcollection.asme.org/ on 04/17/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use


2/9

ing, a 1 m thick plasma enhanced chemical vapor depositionPECVDoxide was deposited to insulate the thermistors and viasfrom the lower layer. The heater was then formed on top of theoxide layer by CVC sputtering deposition. A 70 thick layer oftitanium was initially deposited to enhance adhesion characteris-tics and was followed by sputtering a 1 m thick layer of Al-1%Si-4%Cu. Subsequent photolithography and concomitant wetbench processing created the heater on the backside of the wafer.Another 1 m thick PECVD oxide was deposited to protect theback side features during further processing.

Next, the microchannels were formed on the top side of thewafer. The reentrant cavities on the channel wall were also createdthrough the same step. The wafer was taken through a photoli-thography step and a reactive ion etching RIE oxide removalprocess to mask certain areas on the wafer, which were not to beetched during the deep reactive ion etching DRIEprocess. Thewafer was consequently etched in a DRIE process, and siliconwas removed from places not protected by the photoresist/oxidemask. The 2D structure of microchannels and reentrant cavities

was then formed. The DRIE process formed deep vertical trencheson the silicon wafer with a characteristic scalloped sidewall pos-sessing a peak-to-peak roughness of 0.3 m. A profilometerand SEM were employed to measure and record various dimen-sions of the device.

The wafer was flipped, and the backside was then processed tocreate an inlet, outlet, side air gap, and pressure port taps for thetransducers. A photolithography step, followed by a buffered ox-ide etch BOE 6:1 oxide removal process, was carried out tocreate a pattern mask. The wafer was then etched-through in aDRIE process to create the fluidic ports. Thereafter, electrical

contacts/pads were opened on the backside of the wafer by per-forming another round of photolithography and RIE processing.Finally, the processed wafer was stripped of any remaining resistor oxide layers and anodically bonded to a 1 mm thick polishedPyrex glass wafer to form a sealed device. After successfulcompletion of the bonding process, the processed stack was die-sawed to separate the devices from the parent wafer.

The MEMS device was packaged by sandwiching it betweentwo plates. The fluidic seals were forged using miniatureo-rings, while the external electrical connections to the ther-mistors and the heater were achieved from beneath throughspring-loaded pins, which connected the thermistors and theheater to electrical pads residing away from the main microchan-nel body.

3.2 Experimental Test Rig. The setup, as shown in Fig. 2,consists of three primary subsystems: the flow loop section, in-strumentation, and a data acquisition system. The test sectionhouses the MEMS microchannel devices and its fluidic and ther-mal packaging module. The microchannel device is mounted onthe fluidic packaging module through o-rings to ensure a completeleak-free system. The fluidic packaging delivers the working fluidand access to the pressure transducers. The heater, which is fab-ricated on the device backside, is wiredthrough electric padstothe power supply. The thermistors are also connected to a NationalInstruments SCXI-1000 series data acquisition system.

The main flow loop includes the microchannel device, a pulse-less gear pump, a reservoir, which consists of a deaerator unit anda heating element to control the inlet temperature, and a flowmeter. The test section heater is connected to a power supply withan adjustable dc current to provide power to the device. The ther-

mistors output signals are recorded by the data acquisition system.Simultaneously, the inlet pressure and test section pressure dropare collected, and the boiling process in the microchannels is re-corded by a Phantom V4.2 high-speed camera maximum framerate of 90,000 frames/s, and 2 s exposure timemounted over aLeica DMLM microscope. Calibration of the thermistors is per-formed prior to the experiment by placing the device in an ovenand establishing the resistance-temperature curve for each indi-vidual sensor.

3.3 Experimental Procedures and Data Reduction. Thecoolant HFE-7000 was first degassed at atmospheric pressureTsat=34C for at least 24 h. Then, the system was pressurizedby helium to p=143 kPa Tsat=45C. The liquid flow rate wasfixed at desired values, and experiments were conducted after

steady thermal-hydraulic conditions were reached. The electricalresistances of the thermistors were also measured at room tem-perature. During the experiment, voltage was applied in 0.5 Vincrements to the test section heater, and the resistance data for theheater and the thermistors were recorded once steady thermal-hydraulic state was reached, at which the liquid flow rate, heatinput, and resistance data remained constant. Flow visualizationwas also performed through the experiment. Flow morphologies,boiling inception, flow instability, and critical heat flux were re-corded. The heat flux was then decrease in 0.5 V decrements, andthe heat input, thermistor resistance, and flow morphologies wereonce again recorded. The procedure was repeated for differentflow rates.

Fig. 1 a A SEM image of the reentrant cavities, b a CADmodel of the heater and the thermistors on the backside of themicrodevice, and ca CAD model of a single thermistor

121011-2 / Vol. 131, DECEMBER 2009 Transactions of the ASME



3/9

To estimate heat losses, electrical power was applied to the testsection after evacuating the coolant from the test loop. Once thetemperature of the test section became steady, the temperaturedifference between the ambient and test section was recorded withthe corresponding power. The plot of power versus temperature

difference was used to calculate the heat loss Qloss associatedwith each experimental data point. The average heat loss wasestimated to be about 8%.

Data obtained from the voltage, current, and pressure measure-ments were used to calculate the average single- and two-phasetemperatures and the heat transfer coefficients. The electrical inputpower P and heater resistanceRwere determined by the measuredvoltage Vand current I, respectively, with

P=VI 1

andR=V/I 2

The electrical resistance-temperature calibration curves of thethermistors were used for determining the thermistor temperaturefor each local position Tthermistor. The local surface temperatureT1 ,T2 , T3 at the base of the microchannels was then calculatedas

T1,T2,T3=Tthermistor PQlosst

ksAp3

where t,ks, and Apare the substrate thickness, thermal conductiv-ity of silicon, and the platform area, respectively.

Applying fin analysis, the overall fin efficiency is defined as

o=5fAf+At 5Af

At4

where f=tanhmH / mH, m=h2L0+ W / ksWL0, Af=2HL0, At= 5L0W+ 2H, andL0,W,H, andksare the channel length, width,height, and substrate conductivity, respectively.

The overall fin efficiency was iteratively estimated through Eq.4 to be equal or larger than 95%. Thus, the effective heat fluxqeff and the channel wall heat fluxq ch were defined as

qeff =P Qloss

Ap5

qch =P Qloss

At6

where A tis the total channel surface area. The local mass qualityat a distance L from the inlet is obtained by

x=PQlossL/L0GAscpTsat Tin

GAshfg7

where G, As, cp, and hfg are the mass flux, total channel crosssection area, specific heat, and latent heat of vaporization, respec-tively. The thermodynamic quality xis defined using energy bal-ance. The negative value ofxcorresponds to situations where theheat transfer into the flow is less than what it takes to warm up theentire liquid flow to saturation temperature. These situations still

allow inception of nucleate boiling flow regimes. The negativevalues of thermodynamic quality xshould not be confused withthe true vapor quality X define as ratio of cross-sectional vapormass flow rate to total mass flow rate that is often used in thetwo-phase literature.

Local heat transfer coefficient h can be obtained through thelocal surface temperature T, the mean liquid temperature Tl, andqch according to

h= qch

TTl8

whereTlis obtained by energy balance. To evaluate the heat trans-fer during the flow boiling process, the two-phase heat transfercoefficient h tp is also defined

htp= qch

TTsat9

Finally, the mean absolute error MAE is used to compare theexperimental results with correlations according to the followingexpression

MAE= 1

Mi=1

i=MUexp Utheo

Uexp 100% 10

where Uand Mare the parameter under investigation and numberof samples, respectively.

Fig. 2 Experiment setup

Journal of Heat Transfer DECEMBER 2009, Vol. 131 / 121011-3



4/9

Fig. 3 Flow morphologies: a bubbly flow for G=1615 kg/m2 s, qeff=65.8 W/cm2, and x=0.08; b oscillating single-phase liquid/singlebubble/slug tail for G=303 kg/m2 s, qeff =11.9 W/cm

2, and x=0.08; cchurn flow for G=303 kg/m2 s, qeff =34.4 W/ cm

2, andx=0.23;dwispy an-nular flow for G=303 kg/m2 s, qeff =38.9 W/ cm

2, and x=0.28; e invertedannular flow forG=303 kg/m2 s, qeff =53.0 W/cm

2, and x=0.75.

Fig. 4 Flow map forG=1543025 kg/m2 s




5/9

3.4 Uncertainty Analysis.The uncertainties of the measuredvalues are obtained from the manufacturers specification sheets,while the uncertainties of the derived parameters are calculatedusing the method developed by Kline and McClintock 32. Un-certainty in the mass fluxG, total heat fluxq, temperatureT,mass qualityx, and heat transfer coefficienthare estimated tobe 3%, 1%, 1C, 3%, and 9%, respectively. Consider-ing that the heat losses in this study are estimated see Sec. 3.3tobe about 8%, the uncertainty in qeff or q ch is estimated as 4%.

4 Results and Discussion

4.1 Flow Morphologies. Flow patterns similar to conven-tional scale channelexcept for the oscillating single-phase liquid/single bubble/slug tail discussed below were observed for theplain microchannels and the microchannels with reentrant cavi-ties: single-phase liquid flow, bubbly flow Fig. 3a, oscillatingsingle-phase liquid/single bubble/slug tailFig. 3b, churn flowFig. 3c, wispy annular flow Fig. 3d, and inverted annularflowFig. 3e. A flow map Fig. 4is also presented to identifythe relationship between flow patterns and mass flux/mass qualityfor both microchannels. At low mass fluxes, oscillating single-phase liquid/single bubble/slug tail was noticeable immediatelyfollowing boiling inception, where a single bubble grew rapidly,formed a vapor slug, which occupied the entire microchannelcross section, and expanded both downstream and upstream. The

vapor slug then traveled downstream and the microchannel wastemporally occupied by liquid. This type of intermitted flow boil-ing characteristic is often termed as rapid bubble growth in somestudies, and is an important flow boiling instability mode in mi-crochannels 25. The rapid bubble growth flow pattern appearedin both plain and enhanced microchannels for low mass fluxes.However, it extended to higher mass flux in the plain microchan-nel G514 kg / m2 s for structured surface microchannels andG779 kg /m2 s for plain microchannels. For higher mass flux,during subcooled flow boiling, a much less violent bubble forma-

tion was observed with small bubbles departing from the wall,forming a bubbly flow. In this regime, a more uniform bubblenucleation process was generally observed for the microchannels

with reentrant cavities. For both types of microchannels, duringthe transition to saturated flow boiling, a churn flow pattern wasalso observed. For low mass fluxes, flow oscillation ceased toexist with the transition to churn flow at higher mass qualities.This suggested that flow oscillation in microchannels for low sur-face tension fluid is not as violent as for water and is not neces-sarily associated with premature CHF conditions see Sec. 4.3.Wispy annular flow patterns prevailed as the mass quality furtherincreased. When the thermal-hydraulic condition approaches CHF,inverted annular flow was observed for low mass fluxes G

Fig. 5 Effective heat flux at ONB for different mass fluxes

Fig. 6 Heat transfer coefficient as a function of channel heat flux for different mass fluxes




6/9

389 kg /m2 sfor both microchannels, which was characterizedby a liquid core surrounded by vapor sublayers. However, whilethe heat transfer coefficient decreased, the surface temperaturewas steady and did not indicate CHF. It appears that very thinliquid layers were presented adjacent to channel walls. The in-crease in surface temperature indicates that some dry spots startedto form on the channel walls. It also appears that the liquid corewas adjacent to the insolated top Pyrex surface, which suggeststhat the void fraction was significantly higher than it appears onFig. 3e as indicated by high vapor mass quality Fig. 4. CHFwas observed in close proximity to the inverted annular flow.

Figure 5 shows the heat flux at ONB as a function of mass fluxfor both types of microchannels. Significant lower heat flux atboiling inception was observed for the microchannels with reen-trant cavities. These results are consistent with existing studies forflow boiling in macroscale channels 17,18. Likewise, a studyperformed by the current authors on similar device using water15showed the same trendi.e., reentrant cavities provided muchlower heat flux at ONB. It was suggested that the reentrant cavi-ties are very effective in triggering boiling at much lower super-heated surface temperatures than plain channels. As a result, themicrochannels with reentrant cavities had a longer and morestable subcooled boiling region.

4.2 Heat Transfer Coefficient.The local heat transfer coef-ficient as a function of channel heat flux for both microchannels isshown in Fig. 6. Note that the heat transfer coefficient h is defined

based on the wall to mean liquid temperature difference

h=q/TTl= q/Tsat + Tsub for Tl Tsatx 0q/Tsat for Tl=Tsatx 0

11

Low heat transfer coefficients h10,000 W / m2 k were ob-served for single-phase flow. Suppression of boiling inceptiontemperature overshoot is apparent for the plain microchannels,as discussed in Sec. 4.1, which is shown by an extended single-phase region. For mass fluxes below 1120 kg /m2 s, as the chan-nel heat flux increased, a sudden increase in the heat transfercoefficient for the plain microchannels was detected, accompaniedwith boiling inception at local mass quality of x0. For thisdevice, subcooled flow boiling merely existed for G

779 kg/

m2

s. As indicated in Sec. 4.1, rapid bubble growth wasobserved for low mass fluxes especially in the plain channel as aresult of high superheat temperature. It was argued by Kuo andPeles 33 that in microchannels, the heat transfer enhancementduring nucleate boiling was associated with vigorous flow agita-tion of the bulk laminar liquid flow caused by the bubble forma-tion and motion. For plain microchannels, at the stage wherebubble grew hastily and form slug or annular flow downstream,the sudden liquid mixing caused by the rapid bubble growth en-hanced the heat transfer coefficient significantly. However, thisenhancement is rarely manageable as it often associated with flowinstability. For mass fluxes higher than 1606 kg/ m2 s, significantenhancements of heat transfer coefficient were observed 30%for the reentrant-cavity microchannel compared with the plain mi-crochannel. This result is in agreement with the previous study ofKuo and Peles on water boiling flow in similar microchannels15, in which the heat transfer coefficient was enhanced by up to30% for reentrant-cavity microchannel compared with plain mi-crochannel for high mass fluxes G303 kg /m2 s for water.The enhanced heat transfer coefficient was a result of a moreconsistent and uniform distribution of bubbles for high mass flux.

As the channel heat flux further increased x0, a transition tochurn flow, and later to wispy annular flow patterns, was ob-served, as indicated in Sec. 4.1, and convective boiling began toprevail. With the transition to saturated flow boiling and the cor-responding flow pattern transition, bubble nucleation graduallydiminished and the reentrant cavities ceased to be active; underthese conditions, the reentrant-cavity microchannel did not seem

to perform better than the plain microchannel. This was especiallysignificant for high mass qualities, as shown in Figs. 6a6c.The result concurs with flow visualization, where the flow patternsfor both channels showed good agreement.

Fig. 7 aqCHF , bBoCHF, and cxe,CHFas a function of massflux




7/9

4.3 Critical Heat Flux.In conventional scale, the CHF con-dition is primarily a function of mass flux G, exit quality xe,system pressure p, channel hydraulic diameter dh, channellength L, and fluid properties 34. The dependency of qCHF onmass flux and channel exit quality are examined here. Figure 7ashows qCHF as a function of mass flux for both types of micro-channels. Very similar values ofqCHF were observed for the plainmicrochannels and the channels with reentrant cavities. As indi-cated in Sec. 4.1, the annular and inverted annular flow patterns atthe channel exit and the relatively high exit quality at CHF con-ditions, suggests that CHF is triggered by liquid dryout. Withoutthe premature CHF caused by boiling instability, the heat transfermechanisms are similar for both microchannels at conditions im-mediately prior to dryout, and, thus, the CHF values are similar.Note that for the low mass fluxes, the CHF was not affected by therapid bubble growth flow pattern at low mass qualities. Thissomewhat contradicts early studies 25,35, which indicated thatflow oscillation can trigger premature CHF. It appeared that therapid bubble growth observed at low qualities and low mass fluxesis not sufficiently violent to trigger premature CHF. Once the massquality increases beyond a certain threshold x0, flow oscilla-tions are suppressed and are no longer detrimental factors control-ling CHF.

It is also shown in Fig. 7athat CHF increases with mass fluxfor both microchannels. The functional dependency of CHF andmass flux can be reduced to

qCHF

= 3.03G0.5

12with a MAE of 8.5%. Boiling number at CHF condition BoCHFwere also obtained. Kuo and Peles35suggested that the BoCHFis constant for water in plain microchannels, while in the currentinvestigation, BoCHFdecreased with mass flux Fig. 7b.

It has been argued that for dryout mechanism, the exit quality atCHF conditions xe,CHF decreases with increasing mass flux36,37. The reduction in xe,CHF at high mass fluxes for dryoutmechanism is suggested to be a result of increased droplet entrain-ment in the vapor core depleting liquid from the wall 36 orinterfacial wave instabilities induced by shear or surface tensionforces38. Figure 7cshowsxe,CHFas a function of mass flux for

both microchannel devices. Significant reduction in exit qualitieswith increasing mass flux can be observed for flow boiling ofHFE-7000. The results concur with the data of Kosar and Peles37 for R123 in 223 m hydraulic diameter microchannels.However, they conflict with the data of Kuo and Peles 35 forwater flow boiling in plain microchannels, which suggested CHFto be independent of mass quality. The contradicting conclusionmay be a result of considerably smaller surface tension of coolantsuch as HFE-7000 and R123 than water. Considering the surfacetension to be an important variable determining the liquid entrain-ment and the interfacial waves, the droplet and interface waveformations for HFE-7000 are very different from water.

Gambill and Lienhard39proposed that a practical limitationof the maximum heat flux exists for boiling. They suggested thatthe maximum achievable CHF for pe /pc0.01 is 10% of themaximum heat flux calculated from the kinetic theory. Figure 8shows the ratio of the critical heat flux to the maximum heat fluxfrom the kinetic theory qCHF / qmkv as a function of dimensionlessexit pressurepe /pc. Results from several previous studies for CHFin microchannels using water and different coolants are also pre-sented15,16,35,37,40,41. The CHF data of the present study aremuch lower than Gambill and Lienhards limitation, which mightsuggest that the CHF in microchannels are fundamentally lowerthan in conventional scale channel. However, in the current study,a maximum qCHF / qmkv value of 0.02 was obtained for pe /pc 0.058 and G= 3025 kg/ m2 s, which is considerably higher thanthe values obtained by Kosar and Peles 37 for a similar pe /pc

value. Higher q CHF / qmkv value was obtained as the mass flux in-creased for the same system pressure. The results show that higherheat flux is achievable in microchannel with higher mass flux ordifferent working fluid. Further studies are needed to be con-ducted at even higher mass fluxes to explore the heat flux limita-tion in microchannels.

5 Conclusion

Flow boiling in parallel microchannels with low surface tensionfluid HFE-7000 was experimentally studied. Two types of micro-channels were examined: plain-wall microchannel and microchan-nel with structured reentrant cavities on the side walls. Flow mor-

Fig. 8 The ratio of the highest measured critical heat flux to the maximum heat flux from the kinetic theoryqCHF /qmkv as a

function of dimensionless exit pressure pe/pc




8/9

phologies, boiling inceptions, heat transfer coefficients, andcritical heat fluxes were obtained and studied for mass fluxesranging from G=164 kg / m2 s to G=3025 kg / m2 s. Compari-sons between the enhanced and plain-wall microchannels wereperformed. The key findings of this study are as follows:

1. Similar flow patterns were observed for both microchannels:bubbly flow, oscillating single-phase liquid/single bubble/slug tail, churn flow, wispy annular flow, and inverted annu-lar flow. However, transition lines between flow patternsshowed some discrepancy between the two devices. It wasobserved that oscillating single-phase liquid/single bubble/slug tail extend to higher mass fluxes in the plain microchan-nel.

2. Delay of boiling was observed for flow boiling in the plainmicrochannels, while the wall superheat was found to besignificantly reduced at ONB for the reentrant-cavity micro-channels.

3. Rapid bubble growth caused a step wise increase in the heattransfer coefficient for low mass flux in the plain microchan-nel.

4. Heat transfer coefficient was found to be enhanced by up to30% for the reentrant-cavity microchannel compared withthe plain microchannel. However, this enhancement dimin-ished at high mass qualities where convective boiling pre-vailed.

5. Dryout was established to be the CHF mechanism under the

current thermal-hydraulic condition. This concurs with theannular flow pattern and the heat transfer performance priorto CHF conditions for both microchannels. Without the pre-mature CHF caused by boiling instability, reentrant cavitiesdid not enhance the CHF for HFE-7000.

Acknowledgment

This work was supported by the Office of Naval Researchpro-gram officer: Dr. Marl Spector. The microdevice was fabricatedin Cornell Nanofabrication FacilityCNF, a member of the Na-tional Nanotechnology Infrastructure Network, which is supportedby the National Science FoundationNSFunder Grant No. ECS-0335765, Cornell University, its users, and the industrial affiliates.The authors would like to extend their gratitude to the staff andstudents of the CNF.

NomenclatureAp platform area heating surface area above the

heater m2As total channel cross section area m2At total channel surface area m2Bo Boiling number, q / Ghfg

G mass flux kg / m2 sH channel height m

h heat transfer coefficientW / m2 Chfg latent heat of vaporization J/kghsp single-phase heat transfer coefficient

W /m2 CI electrical currentA

ks thermal conductivity of the substratesiliconW /m C

L distance from the inlet of the microchannel mL0 channel length m

p pressurekPaP electrical powerW

q heat flux W / cm2qch channel wall heat flux W /cm

2qeff effective heat flux W / cm

2

Qloss heat loss WR electrical resistance

Re Reynolds number, vD/t thickness of the silicon substratem

T local surface temperatureCTin inlet temperature CTl mean liquid temperatureC

Tthermistor thermistor temperature CTw wall superheat, TTCTsat saturation superheat, TTsatCTsub subcooled temperature, TsatTC

V electrical voltageV

W

channel width mv velocitym/sX true vapor mass qualityx thermodynamiclocal mass quality

Greek

densitym3 /s viscosityN s / m2 surface tensionN/m

Subscripts

ch channeleff effective

l liquidsat saturation

sub subcooled

tp two-phasev vapor

References1 Thome, J. R., 1990, Enhanced Boiling Heat Transfer, Hemisphere, Washing-

ton, DC.2 Zeng, L. Z., and Klausner, J. F., 1993, Nucleation Site Density in Forced

Convection Boiling, ASME J. Heat Transfer, 115, pp. 215221.3 Del Valle, V. H. M., and Kenning, D. B. R., 1985, Subcooled Flow Boiling at

High Heat Flux, Int. J. Heat Mass Transfer, 28, pp. 19071920.4 Kandlikar, S. G., 1990, A General Correlation for Two-Phase Flow Boiling

Heat Transfer Inside Horizontal and Vertical Tubes, ASME J. Heat Transfer,112, pp. 219228.

5 Kandlikar, S. G., 1991, Development of a Flow Boiling Map for Subcooledand Saturated Flow Boiling of Different Fluids Inside Circular Tubes, ASMEJ. Heat Transfer, 113, pp. 190200.

6 Liu, Z., and Winterton, R. H. S., 1991, A General Correlation for Saturated

and Subcooled Flow Boiling in Tubes and Annuli, Based on a Nucleate PoolBoiling Equation, Int. J. Heat Mass Transfer, 34, pp. 27592766.7 Warrier, G. R., and Dhir, V. K., 2006, Heat Transfer and Wall Heat Flux

Partitioning During Subcooled Flow Nucleate BoilingA Review, ASME J.Heat Transfer, 128, pp. 12431256.

8 Kandlikar, S. G., 2007, Discussion: Heat Transfer and Wall Heat Flux Parti-tioning During Subcooled Flow Nucleate BoilingA Review, ASME J. HeatTransfer, 129, pp. 13001301.

9 Jiang, L., Wong, M., and Zohar, Y., 2001, Forced Convection Boiling inMicrochannel Heat Sink, J. Microelectromech. Syst., 101, pp. 8087.

10 Zhang, L., Koo, J., Jiang, L., A sheghi, M., Goodson, K. E., and Santiago, J. G.,2002, Measurements and Modeling of Two-Phase Flow in MicrochannelsWith Nearly Constant Heat Flux Boundary Conditions, J. Microelectromech.Syst., 111, pp. 1219.

11 Kandlikar, S. G., 2002, Fundamental Issues Related to Flow Boiling in Min-ichannels and Microchannels, Exp. Therm. Fluid Sci., 26, pp. 389407.

12 Qu, W., and Mudawar, I., 2003, Flow Boiling Heat Transfer in Two-PhaseMicro-Channel Heat SinksI. Experimental Investigation and Assessment ofCorrelation Methods, Int. J. Heat Mass Transfer, 46, pp. 27552771.

13

Qu, W., and Mudawar, I., 2003, Flow Boiling Heat Transfer in Two-PhaseMicro-Channel Heat SinksII. Annular Two-Phase Flow Model, Int. J. HeatMass Transfer, 46, pp. 27732784.

14 Kosar, A., Kuo, C.-J., and Peles, Y., 2005, Boiling Heat Transfer in Rectan-gular Microchannels With Reentrant Cavities, Int. J. Heat Mass Transfer,482324, pp. 48674886.

15 Kuo, C.-J., and Peles, Y., 2007, Local Measurement of Flow Boiling in Struc-tured Surface Microchannels, Int. J. Heat Mass Transfer, 502324, pp.45134526.

16 Roday, A. P., Borca-Tasciuc, T., and Jensen, M. K., 2008, The Critical HeatFlux Condition With Water in a Uniformly Heated Microtube, ASME J. HeatTransfer, 130, pp. 012901.

17 Roser, R., Thonon, B., and Mercier, P., 1999, Experimental Investigations onBoiling of N-Pentane Across a Horizontal Tube Bundle: Two-Phase Flow andHeat Transfer Characteristics, Int. J. Refrig., 22, pp. 536547.

18 Chang, J. Y., and You, S. M., 1997, Boiling Heat Transfer Phenomena From




9/9

Micro-Porous Surfaces in Saturated FC-72, Int. J. Heat Mass Transfer,4018, pp. 44374447.

19 Zhang, L., Wang, E. N., Goodson, K. E., and Kenny, T. W., 2005, PhaseChange Phenomena in Silicon Microchannels, Int. J. Heat Mass Transfer,488, pp. 15721582.

20 Kosar, A., Kuo, C.-J., and Peles, Y., 2005, Reduced Pressure Boiling HeatTransfer in Rectangular Microchannels With Interconnected Reentrant Cavi-ties, ASME J. Heat Transfer, 12710, pp. 11061114.

21 Kuo, C.-J., Kosar, A., Peles, Y., Virost, S., Mishra, C., and Jensen, M. K.,2005, Bubble Dynamics During Boiling in Enhanced Surface Microchan-nels, J. Microelectromech. Syst., 156, pp. 15141527.

22 Pate, D. T., Jones, R. J., and Bhavnani, S. H., 2006, Cavity-Induced Two-Phase Heat Transfer in Silicon Microchannels, Thermomechanical Phenom-ena in Electronic Systems,Proceedings of the 10th Intersociety Conference on

Thermal and Thermomechanical Phenomena and Emerging Technologies inElectronic Systems (ITherm 2006), pp. 7178.23 Jones, R. J., Pate, D. T., and Bhavnani, S. H., 2007, Control of Instabilities in

Two-Phase Microchannel Flow Using Artificial Nucleation Sites, Proceed-ings of the ASME InterPack Conference (IPack 2007), pp. 247358.

24 Naveenan, T., Jones, R. J., Pate, D. T., and Bhavnani, S. H., 2008, ThermalCharacteristics of Two-Phase Flow of a Dielectric Fluid of Surface-AugmentedMicrochannels, 11th IEEE Intersociety Conference on Thermal and Thermo-mechanical Phenomena in Electronic System ITherm 2008, pp. 189196.

25 Kuo, C.-J., and Peles, Y., 2008, Flow Boiling Instabilities in Microchannelsand Means for Mitigation by Reentrant Cavities, ASME J. Heat Transfer,1307, p. 072402.

26 Kuo, C.-J., and Peles, Y., 2009, Pressure Effect on Flow Boiling Instabilitiesin Parallel Microchannels, Int. J. Heat Mass Transfer, 521-2, pp.271280.

27 Yin, C.-P., Yan, Y.-Y., Lin, T.-F., and Yang, B.-C., 2000, Subcooled FlowBoiling Heat Transfer of R-134a and Bubble Characteristics in a HorizontalAnnular Duct, Int. J. Heat Mass Transfer, 4311, pp. 18851896.

28 Lie, Y. M., and Lin, T. F., 2006, Subcooled Flow Boiling Heat Transfer andAssociated Bubble Characteristics of R-134a in a Narrow Annular Duct, Int.

J. Heat Mass Transfer, 491314, pp. 20772089.29 Martin-Callizo, C., Palm, B., and Owhaib, W., 2007, Subcooled Flow Boiling

of R-134-a in Vertical Channels of Small Diameter, Int. J. Multiphase Flow,33, pp. 822832.

30 Bertsch, S. S., Groll, E. A., and Garimella, S. V., 2008, Refrigerant FlowBoiling Heat Transfer in Parallel Microchannels as a Function of Local VaporQuality, Int. J. Heat Mass Transfer, 51, pp. 47754787.

31 Revellin, R., Agostini, B., Ursenbacher, T., and Thome, J. R., 2008, Experi-mental Investigation of Velocity and Length of Elongated Bubbles for Flow ofR-134a in a 0.5 MM Microchannel, Exp. Therm. Fluid Sci., 32, pp. 870881.

32 Kline, S., and McClintock, F. A., 1953, Describing Uncertainties in Single-Sample Experiments, Mech. Eng.Am. Soc. Mech. Eng., 751, pp. 38.

33 Kuo, C.-J., 2009, Flow Boiling in Structured Surface Microchannels, Ph.D.dissertation, Rensselaer Polytechnic Institute, Troy, NY.

34 Collier, J. G., and Thome, J. R., 1994, Convective Boiling and Condensation,3rd ed., Oxford University Press, New York.

35 Kuo, C.-J., and Peles, Y., 2008, Critical Heat Flux of Water at Sub-Atmospheric Pressures in Microchannels, ASME J. Heat Transfer, 1307, p.072403.

36 Carey, V. P., 1992, Liquid-Vapor Phase-Change Phenomena, Taylor and Fran-cis, London.

37 Kosar, A., and Peles, Y., 2007, Critical Heat Flux of R-123 in Silicon-BasedMicrochannels, ASME J. Heat Transfer, 1297, pp. 844851.

38 Revellin, R., and Thome, J. R., 2008, A Theoretical Model for the Predictionof the Critical Heat Flux in Heated Microchannels, Int. J. Heat Mass Transfer,515-6, pp. 12161225.

39 Gambill, W. R., and Lienhard, J. H., 1987, An Upper Bound for the CriticalBoiling Heat Flux, Proceedings of ASME-JSME Thermal Engineering JointConference, Vol. 3, pp. 621626.

40 Kuan, W. K., and Kandlikar, S. G., 2006, Experimental Study on SaturatedFlow Boiling Critical Heat Flux in Microchannels, Proceedings of the 4thInternational Conference on Nanochannels, Microchannels and Minichannels,pp. 4552.

41 Wojtan, L., Revellin, R., and Thome, J. R., 2006, Investigation of SaturatedCritical Heat Flux in a Single, Uniformly Heated Microchannel, Exp. Therm.Fluid Sci., 308, pp. 765774.


Documents

Kuo and Peles 2009 Flow Boiling of Coolant (HFE-7000) Inside Structured and Plain Wall Microchannels