International Journal of Heat and Mass Transfer 112 (2017) 991–1004

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

journal homepage: www.elsevier .com/locate / i jhmt

Condensation on hybrid-patterned copper tubes (I): Characterization ofcondensation heat transfer

http://dx.doi.org/10.1016/j.ijheatmasstransfer.2017.05.0390017-9310/� 2017 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.E-mail address: li01@cec.sc.edu (C. Li).

Mohammad Alwazzan, Karim Egab, Benli Peng, Jamil Khan, Chen Li ⇑Department of Mechanical Engineering, University of South Carolina, Columbia 29208, United States

a r t i c l e i n f o

Article history:Received 30 November 2016Received in revised form 5 May 2017Accepted 10 May 2017

Keywords:Dropwise condensationHydrophobicHybrid wettabilityEnhancementDroplet migration

a b s t r a c t

Condensation heat transfer performance can be improved by increasing the condensate removal rate.Commonly, this can be achieved by promoting dropwise condensation mode in which super/hydrophobiccoatings applied on the entire condenser surface. Herein, alternative mini-scale straight patterns con-sisted of hydrophobic (b) and less-hydrophobic (a) regions were formed on the condenser tubes. Theexistence of the two adjacent regions generates wettability gradient which can mitigate condensateand increase its removal rates. A parametric study was conducted to experimentally determine the influ-ence of (b/a) ratios on the heat transfer performance and droplet dynamic under saturation conditionnear the atmosphere pressure with the presence of non-condensable gases (air). The results reveal thatall patterned surfaces exhibited a drastic enhancement in terms of condensation heat transfer coefficientand heat flux compared to those of filmwise condensation. More interestingly, some (b/a) ratios signif-icantly outperformed a surface with a complete dropwise condensation. In addition, an optimum (b/a)ratio of (2/1) exists with b and a-regions widths of 0.6 mm and 0.3 mm, respectively. The heat transfercoefficient of the optimum ratio is peaked at a value of 85 kW/m2 K at a subcooling of 9 �C, which is4.8 and 1.8 times that of a complete filmwise and dropwise condensation, respectively. Our study alsoreveals that the b-regions served mainly as droplet nucleation sites with rapid droplets mobility; whereasthe a-regions promoted droplet removal from the neighboring b-regions, and served as drainage pathswhere condensate can be drained quickly under gravitational force. Furthermore, the existence of botha and b-regions on the condensing surface controls the droplets maximum diameters of the growing dro-plets on the b-regions. The maximum diameter is approximately 0.56 ± 3% mm, which is 26% the size ofthe droplets maximum diameter on a full b-region surface. In summary, this wettability-driven mecha-nism allows droplets to be removed from the condensing surface at higher rates, leading to a substantialenhancement in the condensation heat transfer coefficient.

� 2017 Elsevier Ltd. All rights reserved.

1. Introduction

Condensation is a required and yet widespread process that canbe found in many small-scale and industrial applications such as inpower plants, water desalination and harvesting, HVAC (Heating,Ventilation, and Air Conditioning), and dehumidification [1–4].Increasing the condensation performance efficiency can signifi-cantly reduce the energy and materials consumption as well ascapital and operational costs. Condensation heat transfer (HT) per-formance can be enhanced using diverse methods which are pre-sented throughout numerous research studies. One of the mostcommon methods is by promoting dropwise condensation (DWC)mode owing to its superior HT rate to that of filmwise condensa-

tion (FWC) mode [5–9]. In general, an effective DWC can beachieved when nonwetting superhydrophobic or incomplete wet-table surfaces are used as the condensing surface. These surfacesusually have higher drop nucleation site densities, coalescenceand droplet roll-off rates, resulting in a higher replenishment fre-quency at smaller droplet diameters [10–14].

Various methods have been developed to achieve DWC. Theseinclude using ion implantation to reduce the surface free energyby applying different alloy composites on condensers made fromvariety of materials, such as titanium, stainless steel, and alu-minum. The influence of different parameters, roughness, and oxi-dation effects on these surfaces was also considered [15–20]. Bothion-plating and ion-beam mixing were used to develop differentalloy layers on copper condenser surface to reduce its wettability[21]. Dynamic ion-beam mixed implantation technique also usedwith different surface processing conditions to form, for instance,

Nomenclature

A surface area of the tube (m2)C constant of Eq. (5)cp specific heat capacity (J/kg K)D diameter (m)F droplet departure frequency (1/s)g gravitational acceleration (m/s2)h heat transfer coefficient (W/m2 K)hfg latent heat of evaporation (J/kg)k thermal conductivity (W/m K)L tube length (m)_m mass flow rate (kg/s)Nu Nusselt numberP pressure (Pa)Pr Prandtl numberQ heat transfer rate (W)q00 condensation heat flux (W/m2)R resistance (K/W)Re Reynolds numberT temperature (K)U overall heat transfer coefficient (W/m2 K)

Subscriptsc coolantc:conv coolant convection inside the tubecond condensationCu copperc:w coolant at wall temperature

f filmi inneri:cs inner cross section of the tubein:c inlet of the coolantl liquidLMTD log mean temperature differenceo outerout:c outlet of the coolantsat saturation temperaturetot totalv vaporw wall

Greek symbolsD difference (–)q density (kg/m3)l dynamic viscosity (Pa s)

AbbreviationsHT heat transferHTC heat transfer coefficienta less-hydrophobic region(s)b hydrophobic region(s)R ratioDWC dropwise condensationFWC filmwise condensation

992 M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004

a polytetrafluoroethylene film on different substrates such asbrass, copper, stainless steel, and carbon steel [22]. Surfaces withlow wettability can also be formed by depositing layers of bariumstearate molecules using the Langmuir-Blodgett method [23].Another technique to reduce the surface wettability is to use chem-ical vapor deposition method, whereas an amorphous layer ofhydrogenated diamond-like carbon [8,24,25], or a polymer layer,was grafted on the condensing surface to achieve sustained DWC[26]. Wet chemical oxidation method was also used to form asuperhydrophobic nanowire layer of copper (II) hydroxide on acopper substrate to sustain high condensation performance undera relatively longer period of time [27].

In addition, a method combined chemical oxidation, etchingprocess, and chemical vapor deposition was used to deposit a sila-nized copper oxide layer featuring knife-like functionalized nanos-tructure on copper surfaces to achieve both Jumping-droplets andDWC modes [28,29]. Such a method is also used to deposit anultrathin layer of graphene coating on copper surface showingchemically stable and low thermal resistance hydrophobic surfaces[30,31]. Slippery surfaces exhibiting high droplets mobility andsweeping rates were prepared by lubricating the condensing sur-face consist of a layer of hierarchical micro-nanoscale texture[32,33]. These types of surfaces repulse micro-scale condenseddroplets due to the reduction of the liquid-solid pining effect.

Organic coatings formed by self-assembled monolayer (SAM)method were also widely used to reduce the free surface energyof the condensers and achieve DWC [34,35]. Despite the additionalthermal resistance induced by thick organic coatings, DWC for aperiod over 12,000 h (500 days) with a six times higher HT perfor-mance compared to FWC was realized [36]. By applying an ultra-thin film of (SAM)-n-octadecyl-mercaptan layer on a copper sur-face, an exultant DWC was achieved with an 300% and 180%enhancement over FWC under vacuum conditions for a period of100 h and 2600 h, respectively, or up to eight times under atmo-

spheric pressure [37,38]. Other studies using similar coatings ondifferent substrates such as gold-coated aluminum and copper-nickel tubes reported nine and fourteen times enhancements,respectively, over FWC under atmospheric conditions [39]. Dueto the simplicity of the procedure of SAM coatings method, it isadapted herein to reduce the wettability of the condensing surface.

Most of these methods are intended to reduce the surface wet-tability over the entire condensing surface to ensure a uniformDWC mode. However, promoting DWC can also be realized bymanipulating different wettability regions on the condenser sur-face at nano- or microscale levels. Such examples of this approachcan be inspired by nature, such as by mimicking the lotus leaves bymeans of short nanotubes deposited on micromachined posts [40],or by mimicking the Namib dessert beetles using hybrid surfaceswith high-contrast wettability patterns [41]. Other examples canfollow man-made designs such as using a hierarchical structurewhich combines both nanograss and micropyramidal architectures[42], introducing nanoscale hydrophilic lines on hydrophobic sur-faces [43], or by forming hydrophilic dots on hydrophobic surfaces[44], and many other studies [45–50]. However, these methods canbe too expensive and challenging to be implemented in industrialscale, especially when the condensing surface is in a tubing geom-etry. Moreover, these nano/micro-scale structures tend to beflooded at high condensation heat flux or at high degree of subcool-ing [28,49].

A similar approach of combining two distinct wettabilityregions but at mini-scale patterns was developed earlier by Kuma-gai [51] to promote droplet shading using alternating hydrophobicand hydrophilic stripes in vertical and horizontal orientations onthe condensing surface. Higher HT performances were obtainedon these hybrid wettability condensing surfaces compared to a sur-face with complete FWC. However, their performances were stillbelow that a surface with complete DWC. Different configurationsand geometries of these hybrid-patterns were also adapted and

M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004 993

formed on a condensing surface by various studies. For example, acondensing surface inspired by the back of the Stenocara beetles inthe Namib Desert was developed by sintering hydrophobic coppermesh/gauze on a polystyrene hydrophilic flat sheet, and used forefficient fog harvesting [52]. Another condensing surface withhydrophobic and hydrophilic regions was created by etching inter-digitated patterned designs inspired by the vein network layout ofbanana leaves on the condensing surface. An improvement up to19%, compared to that of a bare surface without the patterns,was achieved in the overall condensation collection rate [53]. Sim-ilar patterns but with a staggered design were also developed,showing enhancements of about 34.4% and 30.5% in both the heattransfer coefficient (HTC) and water collection rate, respectively, atdry bulb temperature of 20 �C and a relative humidity of 80% [54].However, the condensation HTC was not computed under satu-rated conditions or a wide range of subcooling degrees (i.e. highercondensation rates), which may cause such a design to either holdits high HT performance or exhibit lower one, due to possibleflooding to the patterns or the high thermal resistance of the bridg-ing droplets that are found and explained in Part II of this study.Creating hydrophilic circular shaped islands-patterns on a back-ground of hydrophobic regions was also used and tested, showingabout 7.5% higher HT performance to that of a complete DWC sur-face [55,56]. In another study, straight stripes with different widthof hybrid patterns were formed on the condenser surface andexperimentally tested, showing an enhancement of about 23% incondensation HTC compared to that of complete DWC at a subcool-ing degree of 2.0 K [57]. The need to develop a condensing surfaceexhibiting higher rates of condensation under saturation condi-tions and a wide range of subcooling degree is essential for indus-trial applications and fundamental research of condensation.

In the first part of a two-part study, a mini-scale straight stripedpattern consisting of alternative hydrophobic (b) and less-hydrophobic (a) regions was formed around the surfaces of coppertubes. A parametric study was conducted to experimentally inves-tigate the influence of different (b/a) ratios on the condensation HTperformance. All the experiments were performed under saturatedconditions near the atmospheric pressure, the action of gravity,and the presence of non-condensable gases (air) with a concentra-tion ranged from 3.5% to 7%. The experimental results were thencompared to a complete DWC, FWC, and Nusselt model of FWC.A visualization study was also carried out during such experimentsto consider the droplet dynamics, which will be the main focus ofPart II of this study. The main objectives of this part of the studywere: first, to create wettability gradient on the condenser surface

Fig. 1. Concept of droplet migration mechanism between two different wettability regiothe neighboring a-regions. (b) The capillary-driven phenomena due to the existencewettability region to higher one, that is from b to a region, respectively. (c) Completgravitational force.

and enhance the condensation HT performance by mitigating con-densate via the capillary-driven phenomena; second, to find theoptimum (b/a) ratio that provides the maximum heat flux andHTC, third, to justify why some (b/a) ratios exhibit lower perfor-mance than the optimum, and finally, to better understand thedynamic of droplets undergoing water vapor condensation on acondenser surface with hybrid wettability.

2. Experiment

2.1. Concept

The main purpose for introducing a and b-regions on the con-densing surface is to control the maximum base diameter of thecondensing droplets by inducing capillary-driven phenomena.The existence of wettability gradient on the condensing patternedsurface allows condensates to be rapidly shed at a smaller diame-ter. As a result, more surface areas would be exposed to the steam,leading to a higher condensation rate. These a-regions can beformed in any configurations, scales, and ratios. In this study, weadopted a pattern with a parallel straight-stripes configurationowing to its manufacturing simplicity, which allows applying thispattern to large surface areas. Herein, the width of the b-stripeswas set at 0.6 mm throughout all experiments. Other b-stripeswidths (scale) such (1.0, 0.5, and 0.2 mm) were experimentallytested, but their performances were found to be lower than0.6 mm. Therefore, we choose 0.6 mm as the case in this study asfixed scales, whereas, the a-stripes width was varied from0.2 mm to 0.6 mm within an increment of 0.1 mm.

During a condensation on hybrid patterned surface, dropletsinitially prefer to nucleate on the a-region rather than on the b-region, however, with a continuous condensation, a liquid film cov-ers the a-region and reduces its surface temperature, eventuallythe droplets prefer the b–region due to its availability to the sur-rounding vapor [49,58]. Conceptually, in the straight patterndesign, each b-region is bordered by two a-regions, as schemati-cally shown in Fig. 1. When a growing condensing droplet comesin touch with the a-regions-boundary, an immediate migrationwould occur due to the capillary force established between thesetwo different wettability regions. For those droplets that are morelikely to originate near the center of the b-regions, their base-diameters can reach a maximum size that is equivalent to thewidth of the b-regions, Fig. 1a. After which the growing dropletboundary will touch the a-region and relocation would occur,Fig. 1b; eventually, the gravitational force will drain the relocated

ns. (a) A condensing droplet grows on the b-region until its boundary touches withof the two wettability neighboring regions, drags the droplet(s) from the lowere droplet removal and drainage occur within the higher wettability region under

994 M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004

droplets within the a-region (Fig. 1c), and completely departingthe condenser. Basically, the existence of wettability gradient onthe condensing surface mainly services first, in controlling themaximum diameter of the condensing droplets on the b-regions,second, in increasing the droplet removal rate from b to a-regions with the assist of the capillary driven force, and third, inincreasing the drainage rate due to the existence of the drainagepaths, i.e. the a-regions.

2.2. Test samples

Copper tubes (Cu-101 from McMaster-Carr, 99.99% purity) witha total length of 120 mm and effective condensing length of92 mm, a diameter of 6.35 ± 0.8 mm, and wall thickness of0.9 ± 0.15 mm were used as the testing sample for all experiments.Tube geometry instead of flat plate was considered due to its com-mon use in real applications as well as the scarcity of studies thatinvestigate condensation on tube configurations with hybrid wet-tability. It was a challenge to form a hybrid wettability patternon a surface of tube, fortunately with the method developedherein, we were able to successfully form a ‘‘straight stripes” pat-tern on copper tubes. Five samples with different (b/a) ratios werefabricated, and tested, then compared to a sample with a full b-region (F-b), a full a-region (F-a), and Nusselt model of FWC.Table 1 shows the samples categories that were used as the maincase in this study. The letter R in the Name/label column denotesto the ratio of the b to a-region, and the digit represents thevalue of that ratio. Notice that the b-regions width keptconstant throughout this case of the study, whereas, the a-regionwidth was varied to attain different (b/a) ratios as shown inthe table.

Fig. 2. A schematic drawing of the main procedures of preparing a hybridwettability pattern on the condensing surface of this study. (a) Fixing the printedtransparent sheet which carries the designed pattern on top of the photoresistsheet. (b) Exposing both sheets to a UV light source that is facing the transparencesheet side. Only the regions of the photoresist film that is exposed to the light willbe brittle which will be identical to the printed design of the transparent sheet. (c)Carefully fixing the photoresist sheet with the assist of the adhesive layer on top ofthe condensing surface. Applying a micro-abrasive blasting process on thephotoresist side; the fine stream can only target the condensing surface areasbeneath the brittle regions of the photoresist sheet. This allows the creation ofroughened regions matching the pattern design of the photoresist sheet (d)Removing the photoresist sheet residual from the condensing surface followed bycleaning procedure, then immersing the condensing surface that carried theroughened pattern in the SAM solution for an hour.

2.3. Surfaces preparation

2.3.1. Hybrid wettability patterned surfaceThe copper tube surface was mechanically polished (Grids #:

800, 1200, and 2500), rinsed with sulfuric acid, acetone, ethanol,and deionized water consecutively, then dried by a nitrogenstream. The hybrid pattern preparation was then applied followingthe procedure illustrated in Fig. 2. First, an ordinary laser printerwas used to print the selected art work (straight stripes) usingblack color ink on a regular transparent sheet. The transparentsheet was then fixed flatly on the top of the photoresist sheet (IKO-NICS Imaging - RapidMask High Tack – 2 mil thickness) in whichthe printed side was facing the photoresist sheet, Fig. 2a. Second,both sheets were exposed to an ultra violet (UV) light source forabout 2–3 min (IKONICS Imaging – Letralite – 15W), Fig. 2b.Because the transparent sheet side was facing the light source dur-ing the light exposure process, the black ink print of the transpar-ent sheet blocks the light source from reaching the photoresistsheet. As a result, the unexposed areas of the photoresist sheetremained at the same original rubbery condition, whereas the

Table 1Samples specifications and labeling.

b-stripes width [mm] a-stripes width [mm] Name/label (b/a)

1 0.6 0.2 R32 0.6 0.3 R23 0.6 0.4 R1.54 0.6 0.5 R1.25 0.6 0.6 R16 Full – F-b7 – Full F-a8 – – Filmwise

exposed areas of the photoresist sheet turned brittle. The photore-sist sheet was then carefully cut to a length matching the outer cir-cumference of the tube and wrapped around the copper tube withthe assist of its adhesive side facing the tube surface. Third, a sand-blasting process was applied using a micro abrasive jet machine(Airbrasive Jet Technologies, LLC, Micro Abrasive Blasting Unit-Model K) and aluminum oxide powder (S.S White TechnologiesInc., Accubrade-27). The sandblasting stream was applied using ahand held micro-jet nozzle (Airbrasive Jet Technologies, LLC, noz-zle Type C, 1.14 mm in diameter). The sandblasting stream wasaimed perpendicularly and about two to three inches from the sur-face of the tube which was held and rotated by hand inside anabrasive blast cabinet. The sand blasting stream only targeted theouter surface areas of the tube that were located beneath the brit-tle region of the photoresist sheet, Fig. 2c. Fourth, the mask wasremoved and the tube was cleaned and dried. Thus, the coppertube surface area with two-roughness regions was prepared.Finally, the sample was immersed in 0.0025 mol/L solution ofn-Octadecyl Mercaptan and ethanol for one hour at 70 �C followedby rinsing and drying procedures, Fig. 2d.

M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004 995

2.3.2. Hydrophobic surface (F-b)It followed the same initial polishing and cleaning procedure of

the hybrid surface, except no mask or sandblasting was applied.The sample was then immersed in the solution to create the SAMcoating.

2.3.3. Sandblasted surface (F-a)It followed same procedure of preparing F-b surface; however,

before the immersion in the n-Octadecyl Mercaptan and ethanolsolution, sandblasting process was applied uniformly on the entiresurface.

2.3.4. Hydrophilic surface (filmwise)The copper tube was polished, rinsed and dried as mentioned

above. Then it was immersed into the 30% H2O2 solution for about3.5 hrs at ambient temperature, then rinsed with ethanol anddeionized water and dried again. Thus, a copper tube with hydro-philic coating was prepared showing a static contact angle of about30� with water.

The SAM coating applied on the condenser herein has differenteffects on the final surface wettability which was related to thesurface roughness. Regardless of the reduction in the a-region wet-tability that could be due to a flooding for the surface texture ortransition to Wenzel state [58–62], our main aim is to create wet-tability gradient by the existence of neighboring b and a regionsduring the condensation process. The average advancing/recedingcontact angles were found to be 125 ± 10�/52 ± 4� for the SAMcoated b-regions and F-b, whereas 138 ± 4�/73 ± 10� for the SAMcoated a-regions and F-a. However, the surface static contactangles computed by imaging process during a condensation undersaturation conditions were found to be 90� ± 2� and 66� ± 4�,

5 μm

50 μm

(a)

(c)

m

β

α

β

α

Fig. 3. Scanning electronic microscopy images of the self-assembled monolayer coatingpatterned surface.

respectively. In addition, scanning electronic microscopy (SEM)images of the surface of b-region and a-region are shown inFig. 3a and b, respectively. SEM images of the interface linebetween the two regions were also considered, showing inFig. 3c and d.

Fig. 4a shows an image of R2 as an example of the finished pro-duct undergoing condensation test at a subcooling temperature ofabout 9 �C and saturation conditions near the atmosphere. In addi-tion, an enlarged image of a small section of the surface is showingin Fig. 4b. The enlarge image shows the b and a regions, which canbe identified by the light and dark stripes, respectively. A sche-matic of the enlarge image shown in Fig. 4c clarifies the tworegions.

2.4. Experimental setup

The experiment setup shown in Fig. 5 consists of four main sec-tions: (i) a condensation testing chamber, (ii) a water vapor gener-ating system, (iii) a water coolant system, and (iv) a dataacquisition/computer system. The 10 � 20 � 24 cm3 aluminumcondensation chamber is equipped with four cartridge heatersinstalled at the chamber’s corners and controlled by a proportional-integral-derivative (PID) temperature controller; in addi-tion, the chamber outside walls are wrapped with thermal insula-tion layers to prevent condensation on the chamber walls duringexperimentation. The chamber is also equipped with a 600 diameterwindow made of heat-resistant borosilicate glass and two flexible-polyimide-etched-foil-heaters (Briskheat-ENP-R-775-01) installedon the glass window. This is to avoid condensation on the innerwall of the window, allowing for clear visualization and observa-tion. A high-speed camera (Phantom v7.3) and a light-emitting-

5 μm

5 μm

(b)

(d)

β

α

α

α

on (a) b-region, (b) a-region, and (c, d) the interface between b and a-region of the

(a)

(b) 500 μμm (c)

100 mm

(c)

Fig. 4. Condensation image of deionized water vapor on patterned copper tube, R2, in the experimental chamber under saturation condition of the atmosphere at subcoolingtemperature of 9 �C is shown in (a). The b and a-regions are displayed in the enlarged image of a small section of the condenser surface area of (b), and illustrated by theschematic drawing in (c).

Fig. 5. A schematic drawing of the experimental setup. The blue dashed lines represent the coolant system. A close loop system was adapted to circulating the coolant from aheated water tank to the test section at a constant flow rate. The red solid lines show the path of the deionized water vapor. The condensate drainage is represented by adouble blue line. All the measurement instruments are connected to the data acquisition system as shown by the doted green lines. (For interpretation of the references tocolor in this figure legend, the reader is referred to the web version of this article.)

996 M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004

M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004 997

diode light source are used for visualization. Three calibrated T-type thermocouples and pressure transducer (Omega PX01C1-050A5T) were installed to monitor and record the saturated vaportemperature and pressure, respectively. Pre-calibrated fine tip K-type thermocouples were used for measuring the inlet and outletcoolant temperatures inside the tube.

A stainless steel water vapor reservoir (10 L capacities)equipped with two controlled heaters was used to generate steam.The main and auxiliary heaters have output powers of 3500W and550W, respectively, allowing for generating steam at a desiredrate. The generated steam was delivered to the condensationchamber through a 19.05 mm diameter and 1524 mm long vibrat-ing resistance stainless steel hose. A cord heater (BriskHeat-HTC451003-3657 mm long and diameter of 4.7 mm) was wrappedaround the hose and controlled by a temperature controller(BriskHeat-SDC120KC-A SDC) to assure a delivery of dry steam.Both the hose and the heater are wrapped with thermal insulationlayers.

A 0.056 m3 capacity galvanized-steel water tank, which isequipped with a stainless steel 2000 W-immersion-heater (withtemperature controller) and a T-type thermocouple, was used tosupply coolant at varied temperatures to a gear pump (SHURflo –GMBN4VA53 – 9.7 GPM – 125 Psi). Two valves installed at thepump discharge line and were used to control the water flow rate(as shown in Fig. 5). The coolant flow rate was monitored by a flowmeter (OMEGA-FMG-92, 0.26–6.6 GPM, with an accuracy of ±1%)that was integrated along the coolant line. Prior to the test section,a controlled 7315 mm long 1440W-cord-heater (BriskHeat-HWC1240) was wrapped around the delivered coolant line to pro-vide an additional temperature control of the inlet temperaturebefore the test section within a maximum range of approximate5 �C. After the test section, the water was pumped back to thewater tank for reuse. Two push-to-connect branch tee adapterswere used to connect the testing samples to the coolant line. Thetee adapters also allow the two fine-tips thermocouples to beinserted to the center of the coolant flow to measure the coolant’sinlet and outlet temperatures. A data acquisition system (Agilent34972A) was interfaced with a computer for data collection andrecording.

2.5. Experimental procedure

For all experiments, the coolant flow rate inside the tubes wasmaintained at 3 ± 2% L/min. whereas, the inlet temperature of thecoolant was varied from 28 �C to 85 �C, corresponding to the Rey-nolds number ranging from 18,000 to 40,000, respectively. Thecontinuous increase of the inlet temperature during the experi-ment was caused by three heating sources: the immersion heaterof the water tank, the inlet heater, and the condensation. The rateof temperature increase and the data sampling rate were consid-ered prior to the data collection to assure collecting steady statedata. Before the coolant water was pumped in the coolant side,the dry steam was assured to be delivered to the condensationchamber until a steady state saturation condition was reachedfor at least 10 min. Monitoring the inside temperature and pres-sure of the chamber as well as the chamber wall temperatureallows for the confirming of a steady state saturation condition.The near atmosphere pressure inside the condensation chamberwas measured and found to be ranged between 0.106 to0.112 ± 0.002 MPa for all experiments. For each testing sample,three sets of experiments were conducted, two for HT characteriza-tions, and one for the visualization study. The two HT experimentswere performed before and after the visualization study, respec-tively, to confirm if there is any change in terms of HT performanceduring the experiments. In the visualization study, the window’sheaters and LED light source were applied. Whereas, during the

HT measurement experiments both the window’s heaters and lightsource were disabled to avoid errors resulting from radiation onthe temperature measurements.

2.6. Data reduction

Using the measured inlet ðTin:cÞ and outlet ðTout:cÞ temperaturesof the coolant side, the measured mass flow rate ð _mcÞ, and the cool-ant specific heat ðCp:cÞ, the total HT rate through only the tube sur-face ðQÞ can be determined by Eq. (1). The heat loss ðQlossÞ in Eq. (1)was computed by calibration experiment as detailed in Section 2.6.After ðQÞ was determined, the condensation heat flux ðq00 ¼ Q=AoÞcan be computed by considering the outer surface area of the tubeðAoÞ.Q ¼ _mcCp:cðTout:c � Tin:cÞ � Qloss ð1Þ

The overall HTC ðUÞ which is a function of only measuredparameters was determined by using ðQÞ, ðAoÞ, and the logarithmicmean temperature difference ðDTLMTDÞ as follows:

U ¼ QAODTLMTD

ð2Þ

DTLMTD ¼ ðTsat � Tin:cÞ � ðTsat � Tout:cÞln Tsat�Tin:c

Tsat�Tout:c

� � ð3Þ

where Tsat is the measured saturated temperature of the watervapor inside the condensation chamber.

To compute the condensation HTC ðhcond ¼ 1=AORcondÞ, the corre-sponding condensation resistance ðRcondÞ was calculated by consid-ering all resistances in the system as follow:

Rcond ¼ Rtot � Rw � Rc:conv ð4Þwhere, Rtotð¼ 1=AoUÞ is the total resistance, Rwð¼ lnðDo=DiÞ=2pkCuLÞis the wall radial conduction resistance, and Rc:convð¼ 1=AihcÞ is theconvection resistance inside the tube To determine the convectionHTC inside the tube ðhc ¼ NucKc=DiÞ, A Sieder-Tate correlation forthe coolant side was used to calculate the coolant Nusselt numberðNucÞ inside the tube [63].

Nuc ¼ CRe0:8c Pr13ðlc=lc:wÞ0:14 ð5Þ

where, C is a constant of 0.035 as determined experimentally in thisstudy [64], Re ¼ ðqc _mcDi=lcAi:csÞ Reynolds number, Pr Prandtl num-ber, lc viscosity of the coolant at average bulk temperature, and lc:w

the viscosity of the coolant at wall temperature. An iterative schemewas carried out to determine the reference temperatures at whichappropriate mean values of the fluid properties can be selected.After hcond was computed, the temperature difference between thetube’s outer surface and the saturated water vapor inside the cham-ber ðDTÞ was determined using the condensation HTC definition:

DT ¼ QAohcond

ð6Þ

2.7. Calibration of the system

To determine the heat transfer rate through only the condensersurface, a calibration experiment was conducted to account for theheat added to the system by any source but not the condensertube, such as the tube fittings and connections. In this experiment,a copper tube similar to the testing tubes and well insulated withmoisture resistance silicon foam on the outer surface was placed inthe testing section. A condensation experiment procedure identicalto all other experiments was then performed, and the HT rate wascomputed which in this case it presents the heat loss

15 20 25 30 35 40 45 500.015

0.020

0.025

0.030

0.035

0.040

Measured data Curve fitting

Qlo

ss [k

W]

Re×103

Fig. 6. An example of the measured HT rate based on the coolant side as a functionof Reynolds number inside the insulated tube during a calibration experimentshowing on the graph with green dot symbols. A curve fitting line presented by thedashed red line was used to consider the heat dissipated by any parts of theassembly but not the condensing tube, and used in Eq. (1) as Qloss . (Forinterpretation of the references to color in this figure legend, the reader is referredto the web version of this article.)

998 M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004

Qloss ¼ _mcCp:cðTout:c � Tin:cÞInsulated tube

� �. This accounted for all HT by

conduction or condensation through the tube fittings and connec-tions. Moreover, the calibration test also accounts for the thermo-couples temperature difference behavior associated with theincrease in the inlet and outlet temperatures within the testingrange. An example of the measured HT rate plotted versus Re ofthe coolant side is shown in Fig. 6. The curve fitting equation ofthe data in Fig. 6 was then used in the second term of the right sideof Eq. (1) to represent the amount of the heat loss during each con-

Fig. 7. Image of a smooth copper tube treated with H2O2 (Filmwise) undergoing filmwpressure.

0

50

100

150

200

250

300

350

0 5 10 15 20 25

Measured Calculated Nusselt's model

q" [kW

/m2 ]

ΔT [°C]

(a)

Fig. 8. Experimental results of condensation (a) heat flux and (b) HTC as a function of susaturation condition near the atmospheric pressure. The before refining two data sets idots), see Section 2.7, and by the computing method illustrated in the data reduction Sinterpretation of the references to color in this figure legend, the reader is referred to th

densation test. To compute accurately the condensation HTC, thesame calibration test was conducted every time when any of thethermocouples, flow meter, pressure transducer, working fluid,and the testing tube dimensions and material was changed.

2.8. Calibration of the experimental setup

The classic Nusselt model of FWC was used to validate and cal-ibrate the experimental setup as showing in Eq. (7).

hfilmwise ¼ 0:728qlðql � qvÞgh0

fgk3l

llDoDT

" #14

ð7Þ

where ql; qv ; g; h0fg ; kl; ll; and Do are the liquid density, vapor

density, gravity, corrected enthalpy of evaporationðh0

fg ¼ hfg ½1þ 0:68ðCp:lðTsat � TwÞ=hfgÞ�Þ [65], liquid thermal conduc-tivity, liquid dynamic viscosity, and the tube outer diameter,respectively. All these properties were computed at film tempera-ture ðTf ¼ Tw þ 0:25ðTsat � TwÞÞ [66], except for qv and hfg whichwere computed atTsat .

For experimental validation, a smooth copper tube treated withH2O2 (sample Filmwise) was used to conduct FWC condensationtest. The measured HT data was then compared to Nusselt modelof FWC. Multiple experiments were conducted to assure repeatableagreements of the results. Fig. 7 shows the FWC image of the testedsample during one of the condensation tests. For all experiments,FWC mode was maintained over the entire condensing surface.

For a further validation of the calculated method in Section 2.5,a direct measurement method of the surface temperature of thetube wall was used. A 0.508 mm diameter T-type calibrated ther-mocouple with the tip coated with thermal grease (ZM-STG2)was inserted in a 0.5 mm deep hole made on the tube wall (drillbit wire gauge-76). The inserted thermocouple was used to mea-sure the tube local wall temperature during the condensation

10 mm

ise condensation of water vapor under saturation condition near the atmosphere

0 5 10 15 20 250

10

20

30

40

50 Measured Calculated Nusselt's model

h cond

[kW

/m2 K

]

ΔT [°C]

(b)

rface subcooling temperature for ‘‘Filmwise” undergoing FWC of water vapor undern the graphs were determined by a direct surface measurement method (red solidection 2.5 (green hollow square), and compared with Nusselt model of FWC. (Fore web version of this article.)

M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004 999

experiment. Then a soldering procedure was followed to keep thethermocouple inside the hole and to prevent the influence of phasechange on the measurements. The soldering coating thickness wasabout 0.4 mm above the tube outer surface. The thermocouple holewas located at the center of the tube length and at the midpoint ofthe tube height. The conductivity between the thermocouple’s tipand both, the depth of the hole, and the soldering coating thicknesswere considered in the data reduction. The FWC results based onboth methods used in Section 2.5 and the direct measured methodwere compared and found to be in a good agreement with Nusseltmodel of FWC, as shown in Fig. 8. The red circular symbols repre-sent the data of the direct measurement using the thermocouples,whereas the green hollow square symbols represents the datacomputed from the calculated method showing in Section 2.5.

2.9. Data refining

The refining process includes averaging the mass flow rate val-ues that were kept constant throughout all the experiments toreduce the fluctuations in the readings, and presenting both theinlet and the outlet temperatures of the coolant by a curve fittingmethod. The refining procedure was applied for all experimentaldata reduction. Moreover, due to a continuous data collection dur-ing the condensation test data points were skipped to avoid pre-senting clattered data in the graphs. Fig. 9 shows the finalrepresentation of the data of Fig. 8 after applying the refiningprocedure.

0 5 10 15 20 250

50

100

150

200

250

300

350 Measured Calculated Nusselt's model

q" [kW

/m2 ]

ΔT [°C]

(a)

Fig. 9. The results of condensation (a) heat flux and (b) HTC as a function of surface subcoadding the error bars.

0 5 10 15 20 250

2

4

6

8

10

12 R3 R2 R1.5 R1.2 R1 F-β F-α

Perc

enta

ge o

f NCG

[%]

ΔT [°C]

(a)

Fig. 10. Percentage of NCG as a function of (a) surface subcooling degree, and (b) the

2.10. Non-Condensable gases measurement

The percentage of the non-condensable gases ðNCG %Þ wascomputed by comparing the chamber pressure that was measureddirectly via the pressure transducer ðP1Þ to that pressure computedby the measurement of the saturated temperature using the ther-mocouples ðP2Þ as follow:

NCG % ¼ P1 � P2

P1� 100 ð8Þ

For all experiments, the percentage of NCG was ranged between3.56% and 6.84%, as shown in Fig. 10.

2.11. Error propagation

The uncertainties in the measurements ðxRÞ associated withheat flux, the overall HTC, log mean temperature difference, andthe condensation HTC were considered in the data reduction pro-cess following Kline and McClintock method [67], showing in Eq.(9).

xR ¼ @R@t1

x1

� �2

þ @R@t2

x2

� �2

þ . . .þ @R@tn

xn

� �2" #1

2

ð9Þ

where, ðR ¼ Rðt1; t2; . . . tnÞÞ is a function of independent variablesðnÞ, xi is uncertainty in the variable ðtiÞ.

0 5 10 15 20 250

10

20

30

40

50

60(b) Measured

Calculated Nusselt's model

h cond

[kW

/m2 K

]

ΔT [°C]

oling degree for the same data sets presented in Fig. 8 after the refining process and

0.107 0.108 0.109 0.110 0.111 0.1120

2

4

6

8

10

12 R3 R2 R1.5 R1.2 R1 F-β F-α

Perc

enta

ge o

f NCG

[%]

Pressure [MPa]

(b)

measured saturated pressure inside the testing chamber for all tested samples.

800∆T=5°C

1000 M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004

2.12. Instruments uncertainties

Fivahede

∆T=6°C∆T=7°C

Experimental measurement

0 5 10 150

200

400

600

800

1000

ΔT [°C]

q" [kW

/m2 ]

g. 11. Experimental condensation HT performance as a functiopor condensation under saturation conditions near the atmosphat flux and HTC as functions of subcooling temperature for samtailed in Section 2.10.

Uncertainty

600 ∆T=8°C

∆T=9°C

Coolant inlet and outlet temperatures ðTin; ToutÞ 0.5 K

]

Flow rate of the coolant ð _mÞ 1%

m2

Saturated vapor temperature ðTsatÞ 0.2 K

400W/

Saturated vapor pressure ðPSatÞ 2%

" [k

Condensing surface area ðAÞ 2%

q

Sieder-Tate correlation HTC ðhiÞ 10% Error associated with heat flux ðq00Þ 4–9%

200

Error associated with HTC ðhcondÞ 13–57%

0 1 2 30

∞(β/α)

Fig. 12. Condensation heat flux as a function of the (b/a) ratio for sample R3, R2,R1.5, R1.2, R1, F-b, and F-a at 5, 6, 7, 8, and 9 �C of surface subcooling undersaturation pressure of (0.106–0.112 ± 0.002 MPa). The zero and infinity valuespresenting in the x-axis refers to sample Sandblasting and F-b, respectively.

3. Results and discussion

3.1. Heat transfer results

The condensation heat flux and HTC as a function of subcoolingtemperature of the patterned samples R2, R1.5, and R1.2 (seeTable 1 for details) are presented in Fig. 11a and b, respectively,and compared with Filmwise sample and Nusselt model of FWC.As expected, all the samples treated with SAM-coating exhibitedsignificantly higher HT performance compared to that of FWCmode. This is due to the influence of the surface energy reductioncaused by the SAM’s coatings. However, since the main focus ofthis study is to investigate the influence of the coexistence of aand b-regions on the condensing surface, sample F-b, and F-a werealso experimentally tested and included in the comparison ofFig. 11. The figure shows that among all samples, F-a has the low-est HT performance indicating that the roughened region did notcontribute in enhancing the HT performance due to the reductionof the surface wettability induced by the SAM coatings, instead itmainly acted as drainage path due its higher surface wettabilityduring condensation process.

More interestingly, Fig. 11 shows that R2 and R1.5 outperformnot only other samples, but also sample (F-b) which exhibits acomplete DWC. The results also show that R2 gained the upper-most improvement in terms of the heat flux and HTC based onthe wettability degree of a and b-regions. The heat flux and HTCof R2, at a subcooling of 9 �C, are 780 kW/m2 and 85 kW/m2 K,respectively. This demonstrates that a maximum enhancement inthe HT performance is approximately 1.8 and 4.8 times higher thanthose of a complete DWC and FWC, respectively.

20 25

R2 R1.5 R1.2 F-β F-α Filmwise Nusselt model

(a)

2

n of subcooling ðDT ¼ Teric pressure (0.106–0.e samples sets are sho

Moreover, the condensation HTC curves of all patterned sur-faces increased with the increase in the subcooling temperaturefollowed the same behavior of a complete DWC. This demonstratesthat the HT performances of the hybrid-patterned surfaces wereenhanced due to promoting the overall DWC mode [68]. If theDWC was dominant, the slope of HTC-DT curve would have thetendency to follow a positive steeper slope. That is the slope ofthe HTC-DT curves which start with the minimum value demon-strated in the case of R1.2, and gradually increase as the (b/a) ratioincreased up to the optimum ratio, R2, showing higher DWC HTCs.

3.2. Existence of an optimum (b/a) ratio

By examining the performances of all patterned samples pre-sented in Fig. 11, we confirmed that the condensation heat fluxand HTC was significantly influenced by the variation of the (b/a)ratio. The figure also displays that R2 has the upper most HT per-formance among all other samples, and hence the optimum (b/a)ratio. Any deviation from this ratio would result in a lower HT per-formance as illustrated by the graph of the heat flux versus (b/a)ratio in Fig. 12. In this figure, R1 and R3 were also included andcompared with all hybrid patterned samples at five different sub-

0 5 10 15 20 250

30

60

90

120

150(b) R2

R1.5 R1.2 F-β F-α Filmwise Nusselt model

ΔT [°C]

h cond

. [kW

/mK

]

sat � TwÞ for sample surfaces R2, R1.5, R1.2, F-b, F-a, and Filmwise undergoing water112 ± 0.002 MPa), and compared with Nusselt model of FWC. Both the condensationwing in (a) and (b), respectively. The error bars of all data computed by the method

0 5 10 15 20 250

200

400

600

800

1000

0 5 10 15 20 250

20

40

60

80

100(b) α=0.3 mm

α=0.4 mm α=0.5 mm α=0.6 mm α=0.7 mm F−β F−α Nusselt model

q" [kW

/m2 ]

ΔT [°C]

β = 1.0 mm

(a)

β = 1.0 mm

α=0.3 mm α=0.4 mm α=0.5 mm α=0.6 mm α=0.7 mm F−β F−α Nusselt model

h cond

. [kW

/m2 K

]

ΔT [°C]

Fig. 13. Condensation (a) heat flux and (b) HTC as a function of subcooling temperature for hybrid patterned samples with fixed b-regions width of 1.0 mm and varied a-regions width (from 0.3 to 0.7), compared with F-b, F-a, and Nusselt model of FWC undergoing water vapor condensation under saturation conditions near the atmosphericpressure (0.106–0.112 ± 0.002 MPa).

M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004 1001

cooling degrees (note: R1 and R3 were not included in Fig. 11 toavoid redundancy and clatter of the data).

Fig. 12 shows that when (b/a) ratio was reduced lower valuesthan R2’s ratio, the performance was reduced as the case in R1.5and R1.2. Whereas, when (b/a) ratio was increased to higher valuesthan the one of R2, the HT performance was reduced as the case inR3. The heat flux and HTC of R1.5 were higher than F-b, howeverlower than the optimum ratio. The maximum heat flux and HTCof R1.5 were about 757 kW/m2 and 71 kW/m2 K at a subcoolingdegree of �10.7 �C, respectively. On the other hand, the maximumheat flux and HTC for R1.2 were about 633 kW/m2 and 42 kW/m2 Kat a subcooling of 15.0 �C, respectively, showing a lower heat fluxthan not only R2 and R1.5, but also lower than that of the completeDWC. The irregular drop in the heat flux of R1.2 in Fig. 12 may beexplained by reaching a point at which the DWC on the b-regionscan be dominated by the droplet bridging phenomenon (explainedin detail in part II). The 0.5 mm width of the a-region of sampleR1.2 allowed for high droplet bridging coverage rate, whichinduces a high thermal resistance and reduces the condensationrate (See Fig. 12 of part II).

1.8 ms 2.1 ms

0.0 ms 0.3 ms

Fig. 14. Time lapse snapshots via the recorded visualization during a condensation tesdroplets during condensation on a hybrid patterned surface for a sample with (b/a)0.112 ± 0.002 MPa. The Red lines highlight a droplet undergoing direct migration from bmigration from b to a-region, followed by bouncing and coalescing actions with neighborcolor in this figure legend, the reader is referred to the web version of this article.)

In addition, Fig. 12 suggests that (b/a) ratio of the hybrid pat-terned surface plays a major role in enhancing the condensationHT performance and by choosing a proper scale and ratio (i.e. band a widths) the performance can be maximized. In this study,based on the wettability degree of b and a-regions, the optimum(b/a) ratio was found to be (2/1). Whereas, based on the averagemaximum droplet base diameter of the hydrophobic surface (F-b), which was 2.2 ± 0.02 mm, the scale (i.e. the width of b-regions) was found to be 0.6 mm, which is about 27% of the basediameter of the average maximum droplets as observed on a F-bsurface. Likewise, we contemplate that the optimum ratio wouldchange if the wettability of b and a-region varied.

In addition, other pattern scales (i.e. other b-regions widths)were also studied, such as 0.5, 0.7, and 1.0 mm; however the opti-mum scale found and considered herein was the 0.6 mm, and cho-sen to be the main scale in this study investigation. Fig. 13 showsan example of the HT performance results of a group of patternedsamples having a scale of 1.0 mm. Whereas, the a-regions width ofthis group was varied from 0.3 mm to 0.7 mm within an incrementof 0.1 mm. The figure shows that all the samples but one exhibited

5 mm

2.4 ms

500 μm

0.6 ms

t on a hybrid/patterned surface. It illustrates the dynamic of the capillary drivenof (0.7/0.3) at a subcooling degree of 16.2 �C, and saturated pressure of 0.106–to a-region, whereas green lines highlight a larger droplet undergoing first a directing droplet(s) as highlighted by the blue lines. (For interpretation of the references to

1002 M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004

lower HT performance than that of F-b. In addition, the figureshows that there also exists an optimum ratio based on this b-regions width of 1.0 mm which was for a-regions width of0.4 mm. However, the HT performance of this optimum ratio wasnot significantly higher than the sample with a complete DWC(F-b).

3.3. Droplet dynamics

Careful attentions were paid to investigate why some patternedsamples with certain (b/a) ratios exhibited higher HT performancethan others. In other words, why there exists an optimum ratio?Using the integrated visualization system, a close observationwas given to the droplet motion mechanism during condensation.We observed that the b-regions with lower wettability actedmainly as nucleation sites where DWC was dominant, whereasthe a-regions with higher wettability acted mainly as drainagesites, promoting droplets migrations from the adjacent b-regions.In addition, besides the coalescence of condensates, and the directvapor condensation on both the condensates and condenser sur-faces [11,14], two unique capillary driven droplet mitigation mech-anisms were observed herein owning to existence of neighboring aand b regions.

The first reported mechanism is due to the relatively smallerdroplet on the b-region and near the b-a boundary. When a grow-ing droplet boundary touched the a-region, a direct droplet migra-tion and confinement within the a-region occurred and followedby a drainage action under the gravitational force (see the redhighlight in Fig. 14). The second witnessed mechanism isaccounted to the migration process of a relatively larger droplet(droplet that was nucleated near the center of the b-region or awayfrom the a-regions boundaries). In this case, migrations to the a-region occurred with larger momentum compared to the formercase. As highlighted in green in Fig. 14, the large droplet entersthe a-region from one side to bounce partially out the other sideboundary while its base still pinned to the a-region to reach andcoalesce with other droplets in the neighboring b-regions. Thiscontributes to coalescence and drainage actions of more droplets;

g 0 ms 5 ms 1

Fig. 15. Bridging phenomena under formation action. Time lapse snapshots via the recorof 13.5 �C, and saturated pressure of (0.106–0.112 ± 0.002 MPa).

(a)

g

(b) 0 μs

Fig. 16. Mechanism of droplet under migration action from b to a-region due to the coalereach a diameter equivalent to the width of b-region is illustrated by the (a) schematisubcooling temperature of 16.2 �C, and saturated pressure of (0.106–0.112 ± 0.002 MPa)

hence, higher rates of surface renewal can be achieved. To summa-rize, combining regions with two different wetting properties (i.e.b and a) on a condensing surface may allows for a better dropletdrainage, and a higher droplet departure frequency; and by intro-ducing a-region on b surface within a suitable scale and rationratio, the condensation rate can be increased to outperform a sur-face with a complete DWC.

Moreover, the experimental results and the integrated visual-ization show that decreasing (b/a) ratio leads to a dominant FWCmode due to an increase of a-regions, yet resulting in a lower HTperformance. On the other hand, increasing (b/a) ratio which canbe realized by two different approaches: either by increasing thewidth (scale) of the b-regions or by decreasing the width of thea-regions. The formal approach was not adapted herein, and thewidth kept constant throughout this study due to the scale effecton the dynamic of the growing droplets on the b-regions. Basically,by increasing the scale (i.e. the width of the b-stripes), the dropletson the b-regions can grow to larger droplet base diameters; hence,the thermal resistance can be increased. With further increase inthe b-stripes, the HT performance will be similar to F-b exhibitingalmost a complete DWC mode with maximum droplets diameterclose to 2.2 ± 0.02 mm as witnessed on F-b. The latter approachwhich was maintained by keeping the width of b-stripesunchanged while varying the a–stripes width was adapted herein.We noticed that by decreasing the a-regions to less than the opti-mum ratio, the droplets on the two neighboring b-stripes can bejoined (bridged) to each other across the middle a–stripe to forma larger droplet that covered three stripes all together (i.e. twoa–stripes and one b-stripe). We observed such interesting phe-nomena (Bridging) increased the thermal resistance due to thelow conductivity of the larger bridging water droplets. With a fur-ther decrease of the a–stripes, more bridging droplets occurred,resulting in different bridging types, such as bridging-dropletsbetween two, three, four, or even five neighboring a-stripes. Thetime lapse images of Fig. 15 demonstrate an example of bridgingdroplet forming on R1.2 surface during condensation. The figureshows an already existing bridging droplet covering two a-stripes was coalescing with surrounding droplets to form a larger

0 ms 15 m 40 ms

4 mm

ded visualization during vapor condensation on R1.2 surface at a subcooling degree

0.46 μs 0.622 μs

5 mm

0.92 μs

scence of small droplets near the a-boundaries which prohibits a growing droplet toc and (b) time lapse images for a patterned sample with (b/a) ratio of (0.7/0.3) at.

M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004 1003

bridging droplet covering three a-stripes. This can induce a higherthermal resistance and yet deteriorates the condensation HT per-formance. More details about the Bridging phenomena will be pre-sented in Part II of this study.

Furthermore, we witnessed that the measured average maxi-mum base diameter of droplets on the b-regions during the con-densation was not identical to the b-regions width asconceptually predicted, but it was 94 ± 3% of width. This is a resultof the higher migrating rate of smaller droplets near the a-boundaries coalescing with the large droplet and joining it withthe a-region before the large droplet boundary reaches the b-region. An illustration of such action is presented in the schematicof Fig. 16a, and by the time lapse images of Fig. 16b.

It was also noticed that the migration rate is at its peak valuenear the a-boundaries; while as it is going farther, bigger dropletsand lower population rates were observed, indicating that the pos-sible maximum droplet diameter is probably originated near thecenter of the b-region or further as possible from the surroundinga boundaries similar to the observation found by Ghosh et al. [53].

4. Conclusions

The concept of introducing alternative parallel straight stripesconsisted of hydrophobic (b) and less-hydrophobic (a) regions atdifferent (b/a) ratios was applied on the surface of copper tubes.A parametric study was conducted to experimentally test and visu-ally observe surfaces with different (b/a) ratio underwent satu-rated vapor condensation near the atmospheric conditions.Accordingly, the following conclusions can be drawn:

(a) The condensation HT performance can be significantlyenhanced by the coexistence of b and a-regions. Comparedto the condensation HT performance of a sample with acomplete FWC and DWC, the hybrid patterned surface R2showed an enhancement of 480% and 180% in the HTC,respectively.

(b) There exists an optimum (b/a) ratio corresponding to wetta-bility degree of the b and a-regions adapted in this study. Itwas found that increasing the a-region width can lead to adominant FWC mode; however, decreasing it can increasethe bridging phenomena and hence reduce the condensationrate. Careful attention should be given when selecting thescale and ratio of b to a-regions. We expected that the scaleand ratio would vary depending on the wettability degree ofboth regions.

(c) The visualization analysis of the condensates dynamic dur-ing condensation on the hybrid patterned surfaces revealedtwo main capillary-driven phenomena responsible for theenhanced condensation HTC. The first is given for the dro-plets that were migrating directly from b to a-regions. Thesecond is for those relatively larger droplets that boundwhile their base pinned to a-region during the migration,and collecting and draining droplets from the neighboringb-regions.

(d) The coexistence of b and a-regions on the condenser surfacenot only provide better droplet drainage but also allow forthe maximum diameter of the growing condensates on thehydrophobic (b) regions to be controlled during thecondensation.

Acknowledgements

We gratefully acknowledge funding received by the NationalScience Founding (NSF) under award 1357920, and the Office

Naval Research (ONR) under grant N000140810080 for supportingthis work in part.

References

[1] D. Seo, C. Lee, Y. Nam, Influence of geometric patterns of microstructuredsuperhydrophobic surfaces on water-harvesting performance via dewing,Langmuir 30 (51) (2014) 15468–15476.

[2] H.G. Andrews, E.A. Eccles, W.C.E. Schofield, J.P.S. Badyal, Three-dimensionalhierarchical structures for fog harvesting, Langmuir 27 (7) (2011) 3798–3802.

[3] K. Hong, R.L. Webb, Performance of dehumidifying heat exchangers with andwithout wetting coatings, J. Heat Transfer-Trans. ASME 121 (4) (1999) 1018–1026.

[4] J.R. Lara, M.T. Holtzapple, Experimental investigation of dropwisecondensation on hydrophobic heat exchangers part I: dimpled-sheets,Desalination 278 (1–3) (2011) 165–172.

[5] F. Eslami, J.A.W. Elliott, Thermodynamic investigation of the barrier forheterogeneous nucleation on a fluid surface in comparison with a rigidsurface, J. Phys. Chem. B 115 (36) (2011) 10646–10653.

[6] J.W. Rose, Dropwise condensation theory and experiment: a review, Proc. Inst.Mech. Eng. Part A-J. Power Energy 216 (A2) (2002) 115–128.

[7] K. Rykaczewski, J.H.J. Scott, Methodology for imaging nano-to-microscalewater condensation dynamics on complex nanostructures, Acs Nano 5 (7)(2011) 5962–5968.

[8] G. Koch, D.C. Zhang, A. Leipertz, Condensation of steam on the surface of hardcoated copper discs, Heat Mass Transfer 32 (3) (1997) 149–156.

[9] E. Schmidt, W. Schurig, W. Sellschopp, Versuche über die Kondensation vonWasserdampf in Film- und Tropfenform, Technische Mechanik undThermodynamik. 1(2) 53–63..

[10] D. Quere, Rough ideas on wetting, Physica A 313 (1–2) (2002) 32–46.[11] P. Meakin, Droplet deposition growth and coalescence, Rep. Prog. Phys. 55 (2)

(1992) 157–240.[12] J. Feng, Y.C. Pang, Z.Q. Qin, R.Y. Ma, S.H. Yao, Why condensate drops can

spontaneously move away on some superhydrophobic surfaces but not onothers, Acs Appl. Mater. Interfaces 4 (12) (2012) 6618–6625.

[13] X. Ma, J.W. Rose, D. Xu, J. Lin, B. Wang, Advances in dropwise condensationheat transfer: Chinese research, Chem. Eng. J. 78 (2–3) (2000) 87–93.

[14] J.L. Viovy, D. Beysens, C.M. Knobler, Scaling description for the growth ofcondensation patterns on surfaces, Phys. Rev. A 37 (12) (1988) 4965–4970.

[15] M.H. Rausch, A.P. Fröba, A. Leipertz, Dropwise condensation heat transfer onion implanted aluminum surfaces, Int. J. Heat Mass Transfer 51 (5–6) (2008)1061–1070.

[16] M.H. Rausch, A. Leipertz, A.P. Froba, Dropwise condensation of steam on ionimplanted titanium surfaces, Int. J. Heat Mass Transfer 53 (1–3) (2010) 423–430.

[17] A. Bani Kananeh, M.H. Rausch, A.P. Fröba, A. Leipertz, Experimental study ofdropwise condensation on plasma-ion implanted stainless steel tubes, Int. J.Heat Mass Transfer 49 (25–26) (2006) 5018–5026.

[18] Q. Zhao, B.M. Burnside, Dropwise condensation of steam on ion-implantedcondenser surfaces, Heat Recovery Syst. Chp 14 (5) (1994) 525–534.

[19] Z. Qi, D.C. Zhang, J.F. Lin, Surface materials with dropwise condensation madeby ion-implantation technology, Int. J. Heat Mass Transfer 34 (11) (1991)2833–2835.

[20] M.H. Rausch, A. Leipertz, A.P. Froba, On the characteristics of ion implantedmetallic surfaces inducing dropwise condensation of steam, Langmuir 26 (8)(2010) 5971–5975.

[21] L.Z. Zhang Dongchang, Lin Jifang, A new method for achieving dropwisecondensation (II) new surface materials for dropwise condensation, J. Chem.Ind. Eng. (China) 3 (2) (1988) 263–271.

[22] X.H. Ma, J.B. Chen, D.Q. Xu, J.F. Lin, C.S. Ren, Z.H. Long, Influence of processingconditions of polymer film on dropwise condensation heat transfer, Int. J. HeatMass Transfer 45 (16) (2002) 3405–3411.

[23] Q. Zhao, D.C. Zhang, J.F. Lin, G.M. Wang, Dropwise condensation on L-B filmsurface, Chem. Eng. Process. 35 (6) (1996) 473–477.

[24] G. Koch, D.C. Zhang, A. Leipertz, M. Grischke, K. Trojan, H. Dimigen, Study onplasma enhanced CVD coated material to promote dropwise condensation ofsteam, Int. J. Heat Mass Transfer 41 (13) (1998) 1899–1906.

[25] J. Robertson, Diamond-like amorphous carbon, Mater. Sci. Eng. R-Rep. 37 (4–6)(2002) 129–281.

[26] A.T. Paxson, J.L. Yaguee, K.K. Gleason, K.K. Varanasi, Stable dropwisecondensation for enhancing heat transfer via the initiated chemical vapordeposition (iCVD) of grafted polymer films, Adv. Mater. 26 (3) (2014) 418–423.

[27] D. Torresin, M.K. Tiwari, D. Del Col, D. Poulikakos, Flow condensation oncopper-based nanotextured superhydrophobic surfaces, Langmuir 29 (2)(2013) 840–848.

[28] N. Miljkovic, R. Enright, Y. Nam, K. Lopez, N. Dou, J. Sack, E.N. Wang, Jumping-droplet-enhanced condensation on scalable superhydrophobic nanostructuredsurfaces, Nano Lett. 13 (1) (2012) 179–187.

[29] M.-K. Kim, H. Cha, P. Birbarah, S. Chavan, C. Zhong, Y. Xu, N. Miljkovic,Enhanced jumping-droplet departure, Langmuir 31 (49) (2015) 13452–13466.

[30] D.J. Preston, D.L. Mafra, N. Miljkovic, J. Kong, E.N. Wang, Scalable graphenecoatings for enhanced condensation heat transfer, Nano Lett. 15 (5) (2015)2902–2909.

1004 M. Alwazzan et al. / International Journal of Heat and Mass Transfer 112 (2017) 991–1004

[31] J. Rafiee, X. Mi, H. Gullapalli, A.V. Thomas, F. Yavari, Y. Shi, P.M. Ajayan, N.A.Koratkar, Wetting transparency of graphene, Nat. Mater. 11 (3) (2012) 217–222.

[32] S. Anand, A.T. Paxson, R. Dhiman, J.D. Smith, K.K. Varanasi, Enhancedcondensation on lubricant-impregnated nanotextured surfaces, Acs Nano 6(11) (2012) 10122–10129.

[33] T.S. Wong, S.H. Kang, S.K.Y. Tang, E.J. Smythe, B.D. Hatton, A. Grinthal, J.Aizenberg, Bioinspired self-repairing slippery surfaces with pressure-stableomniphobicity, Nature 477 (7365) (2011) 443–447.

[34] Z. Lan, X.H. Ma, S.F. Wang, M.Z. Wang, X.N. Li, Effects of surface free energy andnanostructures on dropwise condensation, Chem. Eng. J. 156 (3) (2010) 546–552.

[35] X.H. Ma, S.F. Wang, Z. Lan, B.L. Peng, H.B. Ma, P. Cheng, Wetting modeevolution of steam dropwise condensation on superhydrophobic surface in thepresence of noncondensable gas, J. Heat Transfer-Trans. ASME 134 (2) (2012).

[36] P.J. Marto, D.J. Looney, J.W. Rose, A.S. Wanniarachchi, Evaluation of organiccoatings for the promotion of dropwise condensation of steam, Int. J. HeatMass Transfer 29 (8) (1986) 1109–1117.

[37] S. Vemuri, K.J. Kim, B.D. Wood, S. Govindaraju, T.W. Bell, Long term testing fordropwise condensation using self-assembled monolayer coatings of n-octadecyl mercaptan, Appl. Thermal Eng. 26 (4) (2006) 421–429.

[38] S. Vemuri, K.J. Kim, An experimental and theoretical study on the concept ofdropwise condensation, Int. J. Heat Mass Transfer 49 (3–4) (2006) 649–657.

[39] A.K. Das, H.P. Kilty, P.J. Marto, G.B. Andeen, A. Kumar, The use of an organicself-assembled monolayer coating to promote dropwise condensation ofsteam on horizontal tubes, J. Heat Transfer-Trans. ASME 122 (2) (2000) 278–286.

[40] C.H. Chen, Q.J. Cai, C.L. Tsai, C.L. Chen, G.Y. Xiong, Y. Yu, Z.F. Ren, Dropwisecondensation on superhydrophobic surfaces with two-tier roughness, Appl.Phys. Lett. 90 (17) (2007).

[41] Y.M. Hou, M. Yu, X.M. Chen, Z.K. Wang, S.H. Yao, Recurrent filmwise anddropwise condensation on a beetle mimetic surface, Acs Nano 9 (1) (2015) 71–81.

[42] X.M. Chen, J. Wu, R.Y. Ma, M. Hua, N. Koratkar, S.H. Yao, Z.K. Wang,Nanograssed micropyramidal architectures for continuous dropwisecondensation, Adv. Funct. Mater. 21 (24) (2011) 4617–4623.

[43] Y. Yamada, T. Ikuta, T. Nishiyama, K. Takahashi, Y. Takata, Droplet nucleationon a well-defined hydrophilic-hydrophobic surface of 10 nm order resolution,Langmuir 30 (48) (2014) 14532–14537.

[44] Y. Yamada, K. Takahashi, T. Ikuta, T. Nishiyama, Y. Takata, W. Ma, A. Takahara,Tuning surface wettability at the submicron-scale: effect of focused ion beamirradiation on a self-assembled monolayer, J. Phys. Chem. C 120 (1) (2016)274–280.

[45] M.H.M. Grooten, C.W.M. van der Geld, Surface property effects on dropwisecondensation heat transfer from flowing air-steam mixtures to promotedrainage, Int. J. Thermal Sci. 54 (2012) 220–229.

[46] D.-J. Huang, T.-S. Leu, Fabrication of high wettability gradient on coppersubstrate, Appl. Surf. Sci. 280 (2013) 25–32.

[47] M.M. Derby, A. Chatterjee, Y. Peles, M.K. Jensen, Flow condensation heattransfer enhancement in a mini-channel with hydrophobic and hydrophilicpatterns, Int. J. Heat Mass Transfer 68 (2014) 151–160.

[48] H. Bai, L. Wang, J. Ju, R. Sun, Y. Zheng, L. Jiang, Efficient water collection onintegrative bioinspired surfaces with star-shaped wettability patterns, Adv.Mater. 26 (29) (2014) 5025–5030.

[49] A. Kumar, G.M. Whitesides, Patterned condensation figures as opticaldiffraction gratings, Science 263 (5143) (1994) 60–62.

[50] W. Lei, Z.-H. Jia, J.-C. He, T.-M. Cai, Dynamic properties of vibrated drops on asuperhydrophobic patterned surface, Appl. Thermal Eng. 62 (2) (2014) 507–512.

[51] S. Kumagai, S. Tanaka, H. Katsuda, R. Shimada, On the enhancement offilmwise condensation heat transfer by means of the coexistence withdropwise condensation sections, Exp. Heat Transfer 4 (1) (1991) 71–82.

[52] Y. Wang, L. Zhang, J. Wu, M.N. Hedhili, P. Wang, A facile strategy for thefabrication of a bioinspired hydrophilic-superhydrophobic patterned surfacefor highly efficient fog-harvesting, J. Mater. Chem. A 3 (37) (2015) 18963–18969.

[53] A. Ghosh, S. Beaini, B.J. Zhang, R. Ganguly, C.M. Megaridis, Enhancing dropwisecondensation through bioinspired wettability patterning, Langmuir 30 (43)(2014) 13103–13115.

[54] P.S. Mahapatra, A. Ghosh, R. Ganguly, C.M. Megaridis, Key design and operatingparameters for enhancing dropwise condensation through wettabilitypatterning, Int. J. Heat Mass Transfer 92 (2016) 877–883.

[55] A. Chatterjee, M.M. Derby, Y. Peles, M.K. Jensen, Condensation heat transfer onpatterned surfaces, Int. J. Heat Mass Transfer 66 (2013) 889–897.

[56] A. Chatterjee, M.M. Derby, Y. Peles, M.K. Jensen, Enhancement of condensationheat transfer with patterned surfaces, Int. J. Heat Mass Transfer 71 (2014) 675–681.

[57] B. Peng, X. Ma, Z. Lan, W. Xu, R. Wen, Experimental investigation on steamcondensation heat transfer enhancement with vertically patternedhydrophobic-hydrophilic hybrid surfaces, Int. J. Heat Mass Transfer 83(2015) 27–38.

[58] J. Bico, U. Thiele, D. Quere, Wetting of textured surfaces, Colloid Surf. A-Physicochem. Eng. Asp. 206 (1–3) (2002) 41–46.

[59] C.W. Yao, T.P. Garvin, J.L. Alvarado, A.M. Jacobi, B.G. Jones, C.P. Marsh, Dropletcontact angle behavior on a hybrid surface with hydrophobic and hydrophilicproperties, Appl. Phys. Lett. 101 (11) (2012) 5.

[60] C.W. Yao, J.L. Alvarado, C.P. Marsh, B.G. Jones, M.K. Collins, Wetting behavioron hybrid surfaces with hydrophobic and hydrophilic properties, Appl. Surf.Sci. 290 (2014) 59–65.

[61] A. Lafuma, D. Quere, Superhydrophobic states, Nat. Mater. 2 (7) (2003) 457–460.

[62] P.T. Wang, J.W. Su, M.Y. Shen, M. Ruths, H.W. Sun, Detection of liquidpenetration of a micropillar surface using the Quartz crystal microbalance,Langmuir 33 (2) (2017) 638–644.

[63] E.N. Sieder, G.E. Tate, Heat transfer and pressure drop of liquids in tubes, Ind.Eng. Chem. 28 (1936) 1429–1435.

[64] J.W. Rose, Heat-transfer coefficients, Wilson plots and accuracy of thermalmeasurements, Exp. Thermal Fluid Sci. 28 (2–3) (2004) 77–86.

[65] W.M. Rohsenow, Heat transfer and temperature distribution in laminar-filmcondensation, Trans. ASME 78 (1956) 1645–1648.

[66] W.M. Nagle, G.S. Bays, L.M. Blenderman, T.B. Drew, Heat transfer coefficientsduring drop-wise condensation of steam, Trans. Am. Inst. Chem. Eng. 31(1935) 593–604.

[67] J.K.A.F.A. McClintock, Describing uncertainties in single-sample experiments,Mech. Eng. (1953) 3–8 (January, 1953).

[68] J.W. Rose, Condensation heat transfer, Heat Mass Transfer 35 (6) (1999) 479–485.