Deposition of metallic colloids under sub-cooled nucleate boiling

Colloids and Surfaces A: Physicochem. Eng. Aspects 397 (2012) 85–91

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces A: Physicochemical and Engineering Aspects journal homepage: www.elsevier.com/locate/colsurfa

Deposition of metallic colloids under sub-cooled nucleate boiling Hitesh Bindra a,∗ , Barclay G. Jones b a b

Levich Institute, City college of New York, New York, NY 10031, United States Department of Nuclear Plasma and Radiological Engineering, University of Illinois at Urbana Champaign, Urbana 61801, United States

a r t i c l e

i n f o

Article history: Received 8 November 2011 Received in revised form 2 January 2012 Accepted 27 January 2012 Available online 7 February 2012 Keywords: Bubble dynamics Colloidal transport Evaporation Interaction potential Wall superheat

a b s t r a c t Heterogeneous nucleate boiling causes the deposition of colloidal particles and governs the growth rate of porous deposit layers on heating surfaces when dilute metallic colloidal suspensions are present in aqueous medium. Post-experimental analysis reveals that these deposits are found in the form of spots of variable thickness around the bubble nucleation site. It is shown here that thickness of these spots and deposit amount is directly related to evaporation rate which is dependent on heating wall temperature and fluid temperature. The experimental observations suggest that the deposition occurs around the contact line of bubble and extends underneath area of bubble micro-layer. Rate of transportation of particles to the wall is directly related to frequency of bubble nucleation and departure diameter of departing bubbles. These transported particles deposit onto the heating surface with an attachment probability, which is modeled here as a function of colloid–surface interaction potential. A model of deposition rate has been developed on the basis of evaporative flux Eq  , attachment probability patt and bulk concentration of colloidal particles C0 . This model provides qualitative explanation for experimental results and suggests that deposition flux has linear dependence on the quantity patt C0 Eq  . © 2012 Elsevier B.V. All rights reserved.

1. Introduction Heterogeneous nucleate boiling is the most common form of boiling observed in real life, with applications ranging from boiling water in a faucet to steam generators in power plants. It results in the deposition of impurities such as particulates, salts, and colloids, etc., which are present as a suspension or solute form in the heat transfer fluid, on the heating surface. Various boiling experiments [1,2] with suspended nano-particles showed that these particles form layered deposits on the heated surface which have an improvised impact on surface wettability. The reason attributing to this increase in surface wettability might be higher surface roughness for the deposit layers compared to the bare surface, leading to Wenzel state. Despite research efforts of a decade, it is still unknown how to control this surface roughness by deposition from these boiling experiments so that surface wettability can be tuned. In the real world, industrial scale boilers and nuclear power plants have experienced the deposition of corrosion products on the heat exchange surfaces and the adverse consequences associated with it such as CRUD induced power shift in pressurized water reactor (PWR) type of nuclear power plants [3]. The eroded corrosion products in the bulk stream comprise of metal oxides (size order ∼10 nm to ∼10 ␮m) in the form of dilute suspensions, which can be defined

∗ Corresponding author. Tel.: +1 212 650 6720; fax: +1 212 650 6660. E-mail address: [email protected] (H. Bindra). 0927-7757/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2012.01.038

as colloidal suspensions. The most common examples of such dilute colloidal suspensions are found in process industries with carbon steel piping, wherein the corrosion products mostly constitute of Fe2 O3 present in water in very dilute quantities (10 ppb to 1 ppm). The post examination of industrial scale heat exchange surfaces showed the formation of irregular, non-uniform and thick porous layers (100–300 ␮m). Whereas, laboratory experiments conducted for time scales such as minutes or hours show that these particles deposit selectively at nucleation sites on the heated surface in the form of circular discs and spots. It is difficult to correlate the deposition data from such laboratory experiments with the deposits on steam generators, evaporators, and nuclear fuel surfaces which operate for several months. Therefore, it is essential to understand the basic mechanism for this boiling assisted deposition and develop phenomenological models. Large scale laboratory experiments [4–6] were useful in formulating empirical relations between overall deposition rate of porous layers and operating flow, pressure, heat flux and concentration of the particles. Still, the basic mechanism of boiling-assisted deposition of colloids is unclear. The few consistent mechanistic observations of the prior experimental work are; (a) deposits are in the form of porous layers (b) experiments on small time scale such as minutes or hours show that deposits are in the form of isolated spots or rings representing active bubble nucleation sites. Pool boiling [7] and flow boiling [8] experiments with colloidal suspensions showed that spot-like circular deposit shapes obtained were similar to deposit shapes obtained with boiling of insoluble salt solutions

86

H. Bindra, B.G. Jones / Colloids and Surfaces A: Physicochem. Eng. Aspects 397 (2012) 85–91

Nomenclature db f nc fc hlv Cp l v Vvdw Velec Vt ε ε0 rp e k  G n0 C0 T Tw Ts T∝  Ka patt Eq Eq  p D  ıf xcp

bubble departure diameter (mm) frequency of bubble nucleation (1/s) no. of nucleation sites frequency of bubble nucleation at each site (1/s) latent heat of vaporization (J/kg) specific heat (J/kg-K) liquid density (kg/m3 ) vapor density (kg/m3 ) Van der Waals potential energy (kT) electrostatic potential energy (kT) total interaction potential energy (kT) vacuum permittivity (8.85 × 10−12 F m−1 ) relative permittivity of water (55.8 at 100◦ C) radius of particle (nm) electron charge (1.6 × 10−19 C) Boltzmann constant (1.38 × 10−23 J/K) surface charge density (C/m2 ) free energy (J) zeta potential (mV) concentration of ions (M) bulk colloidal concentration (ppm(w/w)) temperature (◦ C) wall temperature (◦ C) saturation temperature (◦ C) bulk temperature (◦ C) Debye radius inverse (nm−1 ) attachment rate coefficient (m/s) attachment probability evaporation rate (mm3 /s) evaporation flux (mm3 /cm2 -s) density of particle (gm/cm3 ) diffusivity of colloid (cm2 /s) kinematic viscosity (cm2 /s) interaction force boundary layer thickness (nm) distance of closest approach (nm)

[9]. The fundamental studies of vapor bubble growth during boiling from hot surface suggest that there is a liquid layer present, having thickness of few microns, between vapor bubble and heated surfaces. So the actual contact area of the vapor bubble is very small for surfaces such as metals or metal oxides. It was proposed that drying out of this liquid layer beneath the bubble leaving suspended particles on the surface might be the mechanism of deposition [8]. However, this proposed model did not explain the observed effect of chemistry on attachment or re-entrainment of particles. The bubbles in these experiments had very small residence times (O–10 ms) and no visual data on micro-layer evolution or dry-out underneath the bubbles is available yet. In a similar but more controlled experimental environment known as ‘coffee-ring’ experiment, more consistent theory and easily reproducible data were provided. Quantitative modeling [10] proposed to predict rate of growth of deposit based on evaporation flux around drop contact line matched very well with deposition profile and data. Similar experiments [11] with evaporating droplet of colloidal suspension with DLVO (Derjaguin–Landau–Verwey–Overbeek) interactions showed significant impact on deposition pattern. The evaporation rate for such experiments is much slower and for droplets evaporating at faster rate the contact line does not remain pinned necessarily and makes jumps causing dispersed deposits [12]. This can be expected under pool boiling conditions as well where wall superheats make the evaporation process much faster than ‘coffee-ring’. Some recent reports with high fidelity experimental

techniques have been helpful in determining contact line evolution during vapor bubble growth on a heated surface [13]. The limited choices of materials, which are essential to observe attachment kinetics, during experimental simulation, make it impossible to have such detailed observations during colloidal deposition experiments. However, in sub-cooled boiling the bubble growth time is much faster than waiting for another bubble to appear. Therefore the frequency of bubble generation and departure at multiple sites determines the rate of evaporation on the heated surface rather than growth and evolution of a single bubble. In this work the deposition characteristics viz. rate, size and thickness of these deposits dependent upon boiling parameters i.e. bubble frequency and departure diameter have been examined experimentally and a model is formulated based on phenomenological understanding. The motivation for this work was to generate boiling assisted colloidal deposition data and model, excluding the effects from forced circulation and complicated water chemistry.

2. Experimental method Various tests are conducted at different wall temperatures and sub-cooling levels in quiescent pool conditions. A large cuboidal Plexi-glass tank is chosen to house the heater and test piece in such a way that the boiling can be observed on a small metal test-piece in a large aqueous medium. It is desired that heater test piece set up has reduced impact from flow disturbances and to provide a virtually infinite pool of dilute solution of metallic colloids (10 ppm). Due to such large pool of colloidal assembly, change in concentration levels is negligible during the experiments which are conducted for 2 h each. Experimental set-up is shown via a diagram in Fig. 1. Temperature of the bulk liquid in the tank can be independently controlled using immersion heaters at two ends of the tank to avoid thermal gradients in particular directions, thus maintaining uniform bulk temperature. Pressure of the system is atmospheric and corresponding saturation temperature is 100 ◦ C; bulk tank temperature T∝ is maintained at constant value with operation range from 85 ◦ C to 97 ◦ C. Arrangement is set to have an access of high-speed visualization of bubble dynamics with CCD camera. Heating wire made of nichrome is wrapped around grooves on the anodized aluminum hollow cylinder which is adhered to the solid copper block with ceramic paste. This provides a 100 W (30 W/cm2 on the boiling surface) maximum power rating of the heater with maximum current of 4.5 A. The whole heater assembly is insulated in radial direction and heat is conducted in axial direction to heat the liquid. Heat is conducted through the copper block to the test piece through liquid metal joint to provide uniform heat flux over 1 cm2 area of the test piece. Wall temperature and bulk temperature, important parameters to directly impact boiling process, are measured via thermocouples. K-type thermocouples are employed having accuracy of ±1.1 ◦ C to help in monitoring feedback for variable resistance power supplies. High thermal conductivity of test piece avoids wall response time lag during bubble growth process. Particles tested in the system are hematite due to their practical importance as corrosion product. Therefore, alumina having high thermal conductivity and with different iso-electric point (IEP) from hematite is selected as the test piece. IEPs for the surface (Al2 O3 ∼ 9.0) and particles (Fe2 O3 ∼ 6.6) allow low attachment resistance if the pH of bulk solution is in between, in these tests it is 7.5. Average roughness of test piece after surface treatment is less than 0.1 ␮m, measured with DEKTAK stylus profilometer. Nominal spherical particle size after sonication and dilute solution preparation is ∼34 nm for most of the tests. Particle size and zeta-potential values were recorded for each test performed with

H. Bindra, B.G. Jones / Colloids and Surfaces A: Physicochem. Eng. Aspects 397 (2012) 85–91

87

Fig. 1. (a) Diagram of experimental set-up for boiling in dilute colloidal assembly (20 cm × 30 cm × 15 cm); Thermocouples represented by 䊉 are used to measure wall temperature and bulk temperature (b) top view: test piece (Ø 2.5 cm) with heater contact at the base (Ø 1.0 cm) and (c) SEM scan of bare surface.

the help of Malvern Zetasizer. Test pieces are analyzed post experiments for the surface profile of deposit and amount of deposit. Deposit spots on the surface are analyzed with DEKTAK stylus profilometer. After dissolving the particles in nitric acid solvent, the solution is tested under plasma ion mass spectroscopy, PIMS.

2.1. Experimental results In these experiments wall temperature Tw and bulk temperature T∝ are controlled parameters. The boiling parameters, dependent on these two temperatures are bubble size, frequency and number of active nucleation sites. These boiling parameters are the measure of total evaporation rate and will be utilized in correlating deposition rate. These boiling parameters are hard to be predicted theoretically but previously developed models [14,15] help in understanding qualitative relation between controlled parameters and boiling parameters. Wall temperature is raised higher than saturation temperature of water at atmospheric pressure to nucleate bubbles. The magnitude difference between wall temperature Tw and saturation temperature Ts is called wall superheat and non-dimensionally represented by Jakob number (Ja = [Tw − Ts ]cpl l /v hlv ). Similar quantity, degree of bulk liquid sub-cooling is represented by sub-cooling number (Nsub = [Ts − T∞ ]cpl l /v hlv ). As described in Fig. 2, higher wall superheat (or higher Ja) experiments show deposit spots of larger diameter. The corresponding bubble dynamics high speed  visual ization reveals that average bubble departure diameter db and frequency f increases with increase in wall superheat.   Reports [14,16] in literature for bubble departure diameter db under pool

 

boiling have consistently shown that db ∝ Jan , where n is a positive real number. It is well established with experimental data in these reports that wall superheat (or Ja) directly impacts bubble departure diameter. The post experiment test surface pictures (Fig. 2a–d) show consistency with these reports and present that deposition spots have direct correspondence with bubble departure diameters. Test domain of Jakob number, Ja is such that number of active nucleation sites do not change and bubbles remain iso-

 3

lated. Quantification of evaporation rate i.e. Eq =  db nc f/6 is done by the high speed multiple bubble dynamics experimental data (Fig. 2d) obtained during ensemble of deposition experiments. Therefore, wall superheat directly impacts evaporation rate and also the amount of particles transported to the wall. Temperature distribution on the wall is solved numerically; the wall region above saturation temperature shows correspondence with the location of deposition spots (Fig. 2e). The corresponding quantification of deposit amount and thickness for each of those experiments will be detailed later.

Isolated deposit spots, such as one shown in Fig. 3a are analyzed using profilometer and scanning electron microscope (SEM) to understand the mechanism of deposition process. Most of the deposit spots were scanned by DEKTAK® stylus profilometer and showed valley (less thick) in the center (Fig. 3b). The internal region of the deposit spots which does not contain particles corresponds to dry hot radius during bubble nucleation [13]. This evidence as observed in various deposit spots and experimental observations of dry hot spot and cold spot radius with IR temperature data [13] suggest that deposit is formed outside the dry hot spot. This suggests deposit thickness increases outside the dry-spot and after attaining peak or plateau region it starts to decrease as evaporation rate decreases due to colder liquid outside the micro-layer region. This hypothesis is explained and presented in Fig. 3c. Evaporation in the liquid layer induces concentration, agglomeration and deposition of particles near triple phase contact line. SEM pictures show formation of clusters and agglomerates (Fig. 3d), although formation of agglomerates in the liquid phase during evaporation cannot be claimed with present data. In a simplified geometry, it is experimentally observed that evaporation effect dominates over diffusion process which leads to aggregation of particles near evaporating interface [17]. It can be postulated that liquid evaporation rate near wall is much higher which makes colloidal concentration and agglomeration dominate over diffusion and escape process. Variation in sub-cooling changes frequency f (product of number of active sites, nc and frequency at each site, fc ) of bubble nucleation at same or multiple sites on the surface exposed to same wall temperature. The bubble departure diameters and deposit spot diameters are not significantly affected by sub-cooling variation. Higher subcooling decreases bubble frequency (f) which means evaporation rate is reduced and vice versa. These experimental results are shown below and qualitative explanation of the frequency behavior is attributed to change in waiting time for the bubble nucleation [14,18,19]. Each bubble nucleation and departure is accompanied by particle transportation. Therefore, thickness of deposit spot is directly influenced by bubble frequency (Fig. 4d). Higher bubble frequency implies more evaporation at same spot and more particles transported to the wall, therefore thicker deposits are observed. At much higher subcooling (∼15 ◦ C), heated surface is observed to have uniform layer of light-colored deposit in background to deposits on bubble nucleation sites (Fig. 4c). This can be attributed to Soret effect, as hematite has negative Soret coefficient. The amount of deposit is very much less to have an accurate measurement and is negligible as compared to amount of deposit on the bubble nucleation sites. But this deposition rate related to Soret effect can become significant for larger particles (300–400 nm). Here, we present quantitative report of deposition rate variation with wall superheat and subcooling. Deposit spots on the test

88

H. Bindra, B.G. Jones / Colloids and Surfaces A: Physicochem. Eng. Aspects 397 (2012) 85–91

Fig. 2. Higher wall-superheat results in larger deposit size and higher deposit amount. Sub-cooling levels of 6 ◦ C (Nsub = 16.85) at atmospheric pressure conditions are maintained for all three experiments conducted for 2 h. (Section of test piece post-experiment as visible from naked eye), (a) Ja = 5.6; (b) Ja = 11.2; (c) Ja = 22.4; (d) bubble dynamics data obtained from high speed CCD visualization and (e) temperature (◦ C) profile on test piece corresponding to Ja = 11.2, Nsub = 16.85.

Fig. 3. (a) Surface analysis of one of the deposit spots as seen by naked eye; (b) Stylus profilometer scan along deposit diameter (V–V) showing cavity in the center region; (c) hypothesis to explain deposition profile with valley and peaks in (b), and d) SEM image showing agglomerates of 34 nm particles (scan of rectangular box in (a)).

H. Bindra, B.G. Jones / Colloids and Surfaces A: Physicochem. Eng. Aspects 397 (2012) 85–91

89

Fig. 4. Deposit as visible from naked eye with change in subcooling for same wall superheat of 4 ◦ C, (a) Nsub = 16.85; b) Nsub = 33.7; c) Nsub = 42.1; (d) thickness profile from spot with Nsub = 33.7 and (e) decreasing frequency (nc f) on increase in subcooling.

pieces were re-suspended in acid solution for PIMS to quantify the amount of mass deposited for different set of experiments. It reveals direct dependence of deposition rate on overall evaporation

 3

rate Eq =  db nc f/6. Data from our experiments and previous work by Lister and Cussac [6] for different range of evaporative flux is shown in Fig. 5. Evaporative flux is equal to overall evaporation rate per unit heated area. In congruence with previous studies [4], it is found that deposition rate is linear function of colloidal

0.005

0.004

2

Deposition rate ( g/cm -sec)

Pool boiling(dp~34 nm)

0.003

Hematite/Alumina Flow boiling Lister[6](dp~600 nm)

Ja=22.4/Nsub=33.7

Magnetite/Incoloy

Ja=11.2/Nsub=16.85

0.002

Ja=11.2/Nsub=33.7 Ja=5.6/Nsub=16.85

0.001

1

10

Ja=11.2/Nsub=42.1

100 3

2

Evaporative flux (mm /cm of heated area)/sec Fig. 5. Deposition rate as a function of boiling dynamics parameters. Particle concentration is 5 ppm (w/w) for both experimental set-ups. Pool boiling deposition rate is obtained by post-experimental dissolution of hematite particulate and analyzed using PIMS.

concentration for similar evaporative flux. With the concentration up to 20 ppm and experimental time of 2 h, no significant effect on boiling dynamics was observed. This information is used to obtain a linear model of evaporation induced particle transport and attachment kinetics. However this analysis and further model development is limited to very dilute colloidal concentrations.

3. Model description In this paper, we propose a model of particle transportation based on the experimental information on evaporative flux. If there are suspended or colloidal particulate, or even non-volatile dissolved particulates, during the process of bubble growth, these particulates do not evaporate as their solvent does. Single or multiple bubbles contribute to the transport of the colloids to the surface. The evaporative flux Eq  results in particle flux, C0 Eq  on the heated surface if the bulk concentration is C0 . These transported particles are subjected to attachment forces from the surface. The attachment kinetics and rate of attachment for particle surface interaction has been derived in the literature [20] with the help of interaction force boundary layer thickness and surface potential on particles and substrate. It has been postulated previously using first order kinetics for attachment rate, with the relation between wall flux and rate of attachment as Jsurf = D ∇ C + CD ∇ Vt /kT = Ka C. Multiplying both sides of this equation with exp(Vt /kT) for 1-d case and integrating from distance of closest approach (xcp = 0.3 nm) [21] to interaction force boundary layer thickness (ıf ), which is approximated to be equal to Debye length −1 =

εε0 kT/e2 z 2 n, yields relation for

 ıf

attachment coefficient as Ka = D[

xcp

[exp((Vt − Vt (ıf ))/kT )]dx]

−1

.

The attachment probability (patt ) of the particles can be

90

H. Bindra, B.G. Jones / Colloids and Surfaces A: Physicochem. Eng. Aspects 397 (2012) 85–91

0.005

Interaction Potential Energy(kT)

b Pool boiling data Hematite/ Alumina Model Flow boiling data[6] Magnetite/Incoloy Model

0.004

2

Deposition flux ( g/cm -sec)

a

0.003

0.002

0.001

5

0 0 -5

50

100

150

200

250

300

Distance(nm)

-10

-15

Fe2O3/Al2O3 Fe2O3/Fe2O3

-20

-25

1

10

100 3

2

Evaporative flux (mm /cm of heated area)/sec Fig. 6. (a) Comparison of experimental and model results for two different set-ups and (b) total interaction potential energy of 34 nm hematite particles in interaction with alumina (attractive) and hematite surface (repulsive).

1

estimated as Ka ıf /D = [

Xcp

[exp((Vt − Vt (ıf ))/kT )]dX]

−1

. Total

interaction potential energy Vt = Vel + VvdW has two components in accordance with DLVO theory which can be evaluated as; electrostatic potential (Hogg–Healey–Fuerstenau model) Vel = εrp [( p + s )2 ln(1 + exp(− x) + ( p − s )2 ln(1 − exp(− x)]. and Van der Waals potential Vvdw = − A/6[ln(x/(x + 2rp )) + 2rp (x + rp )/ x(x + 2rp )] for interaction between spherical particle and flat surface. Zeta-potentials ( ) are obtained experimentally and substituted to calculate electrostatic potential for particle (p) and surface (s). Hamaker  constant for  two substances  interacting in a medium is A =



Ap −



Aw



As −



Aw . The hamaker

constants for materials, labeled with subscript p – particle, w – water and s – surface, are obtained from literature (Table 1). Using the interaction potential, we can obtain patt and the particle deposition flux which is proportional to the quantity patt C0 Eq  . The purpose of this simple model is to qualitatively and quantitatively understand the deposition under sub-cooled boiling. However, due to complex nature of this problem further detailed modeling and high fidelity experimentation is required for validation.

(a) multi-layer deposition and (b) bubble departure driven microconvection currents in fluid. After few cycles of deposition, new depositing particles come in interaction with deposited layers of similar particles rather than alumina base substrate which changes the interaction potential to repulsive from attractive range (Fig. 6b). In order to capture this phenomenon, attachment coefficient should be continuously modified based upon variation in area fraction of substrate coverage. Current model considers only bare surface for calculating interaction potential. Micro-convection currents due to bubble departure may result in erosion and re-entrainment of particles back into the bulk fluid. Departure of multiple bubbles from same site or various sites may result into fluctuations and stresses which may result in erosion of already deposited particles. The prediction of such transported particles requires detailed study of interaction of fluid stresses and material stress at microscopic scale. Similar model relating attachment probability with interaction potential and fluid stresses was developed for single phase particle transport and deposition [23,24]. Other reports in literature on similar deposition experiments do not enlist data for bubble dynamics and zeta-potential. Without this data, no prediction can be made for those systems, in other words they cannot be used for model validation purposes.

4. Discussion 5. Conclusions Comparing experimental data with the simplified model quantity patt C0 Eq  (Fig. 6a) for two different sets of experimental conditions, the data correspond closely with the predicted deposition flux when multiplication factor is 0.2. The consistency of the linear variation with evaporation flux and constant multiplication factor for two different experimental set-ups confirms that model captures the significant physical processes. There can be several reasons for the multiplication factor less than 1; some of them are;

Table 1 Material properties to evaluate interaction potential.

Particle/surface pH Debye radius, −1 (nm) Zeta potential (mV) Hamaker constant A (kT)

This work

Turner et al. [22]

Hematite/alumina 7.5 540 Hematite −20; alumina +20 1.5

Magnetite/incoloy 7.5 540 Magnetite [22] −17; incoloy +15 1.5

The experimental data generated in this work helps to relate conventional controlled parameters like Jakob and subcooling numbers, for evaluating boiling dynamics and evaporation rate, with the corresponding deposition shape, structure and amount of metallic colloids. The experimental evidence suggests that there is a very clear distinction between the shape, structure and amount of the deposits with changing boiling parameters i.e. bubble frequency and bubble departure diameter. Pool boiling in a virtually infinite dilute colloidal medium gives the facility to perform and show these effects qualitatively and quantitatively, whereas flow boiling experiments conducted before did not report the independent effect of boiling on deposition. High speed observations reveal that bubble departure diameters increase with increasing wall temperature. Bubble frequency is reduced with increase in the sub-cooling although no impact was observed on bubble diameter. Corresponding deposit pictures and thickness profiles post-experiment show that deposit size i.e. diameter directly corresponds to average bubble departure diameter and thickness

H. Bindra, B.G. Jones / Colloids and Surfaces A: Physicochem. Eng. Aspects 397 (2012) 85–91

corresponds to frequency. Spectroscopy weight measurements show that deposition flux is directly dependent upon evaporation flux Eq  . The classical models such as DLVO theory can be conveniently used to formulate attachment probability patt . Deposition flux is linearly proportional to patt C0 Eq  and for two different experimental test conditions the proportionality constant is 0.2. More experimental data for large scale systems with information on zetapotential and boiling parameters will be able to consolidate the model further. However, with currently limited or no information on bubble contact line evolution, agglomeration kinetics in evaporating micro-layer and multi-layer surface attachment kinetics, it is difficult to provide robust detailed model, invent mitigation techniques and control the porosity and wettability of porous layers on surfaces. In situ experimental observations such as continuous thickness, agglomeration kinetics and weight monitoring might help in resolving the impact of multi-layers on attachment kinetics. Acknowledgments We thank Dr. Charles Marsh, CERL and Dr. T Nam-dinh, Idaho National Lab for the support and guidance in understanding of the problem. Research for this publication was partly carried out in the Center for Microanalysis of Materials, University of Illinois at Urbana-Champaign, which is partially supported by the U.S. Department of Energy under grant DEFG02-91-ER45439. References [1] S.J. Kim, I.C. Bang, J. Buongiorno, L.W. Hu, Effects of nanoparticle deposition on surface wettability influencing boiling heat transfer in nanofluids, Appl. Phys. Lett. 15 (2006) 89. [2] H.D. Kim, M.H. Kim, Effect of nanoparticle deposition on capillary wicking that influences the critical heat flux in nanofluids, Appl. Phys. Lett. 1 (2007) 91. [3] J. Deshon, D. Hussey, B. Kendrick, J. McGurk, J. Secker, M. Short, Pressurized water reactor fuel crud and corrosion modeling, JOM J. Miner. Met. Mater. Soc. 63 (2011) 8. [4] M. Basset, J. McInerney, N. Arbeau, D. Lister, The Fouling of alloy-800 heat exchange surfaces by magnetite particles, Can. J. Chem. Eng. (2000) 78. [5] N. Arbeau, W. Cook, D. Lister, The early stages of deposition of magnetite particles onto alloy-800 heat exchange surfaces under sub-cooled boiling conditions, in: ECI Conference on Heat Exchanger Fouling and Cleaning: Fundamentals and Applications, Senta Fe, New Mexico, USA, 2006.

91

[6] D. Lister, F. Cussac, Modeling of particulate fouling on heat exchanger surfaces: influence of bubbles on iron oxide deposition, Heat Transfer Eng. 10–11 (2009) 30. [7] H. Bindra, Effect of boiling in deposition of metallic colloids, PhD Thesis, University of Illinois, 2010. [8] S. Uchida, Y. Asakura, S. Heroaki, Deposition of boron on fuel rod surface under sub-cooled boiling conditions—an approach towards understanding AOA occurrence, Nucl. Eng. Des. (2011) 241. [9] N.B. Hospeti, R.B. Mesler, Deposits formed beneath bubbles during nucleate boiling of radioactive calcium sulfate solutions, AIChE J. 4 (1965) 11. [10] R.D. Deegan, O. Bakajin, T.F. Dupont, G. Huber, S.R. Nagel, T.A. Witten, Capillary flow as the cause of ring stains from dried liquid drops, Nature 6653 (1997) 389. [11] R. Bhardwaj, X. Fang, P. Somasundaran, D. Attinger, Self-assembly of colloidal particles from evaporating droplets: role of the pH and proposition of a phase diagram, Langmuir 11 (2010) 26. [12] K. Lee, M. Duchamp, G. Kulik, A. Magrez, J.W. Seo, S. Jeney, A.J. Kulik, L. Forro, R.S. Sundaram, J. Brugger, Uniformly dispersed deposition of colloidal nanoparticles and nanowires by boiling, Appl. Phys. Lett. 17 (2007) 91. [13] C. Gerardi, J. Buongiorno, Lin-wen Hu, T. McKrell, Study of bubble growth in water pool boiling through synchronized, infra red thermometry and high speed vide, Int. J. Heat Mass Transfer (2010) 53. [14] B.B. Mikic, W.M. Rohsenow, P. Griffith, On bubble growth rates, Int. J. Heat Mass Transfer 4 (1970) 13. [15] G. Son, V.K. Dhir, N. Ramanujapu, Dynamics and heat transfer associated with a single bubble during nucleate boiling on a horizontal surface, J. Heat Transfer (1999) 121. [16] L.Z. Zeng, J.F. Klausner, D.M. Bernhard, R. Mei, A unified model for the prediction of bubble detachment diameters of boiling systems I. Pool boiling, Int. J. Heat Mass Transfer 9 (1993) 36. [17] R.G. Shimmin, A.J. DiMauro, P.V. Braun, Slow vertical deposition of colloidal crystals: a Langmuir–Blodgett process? Langmuir 15 (2006) 22. [18] R.L. Judd, The influence of subcooling on the frequency of bubble emission in nucleate boiling, J. Heat Transfer (1989) 111. [19] N. Zuber, Nucleate boiling. The region of isolated bubbles and the similarity with natural convection, Int. J. Heat Mass Transfer 1 (1963) 6. [20] L.A. Spielman, S.K. Friedlander, Role of the electrical double layer in particle deposition by convective diffusion, J. Colloid Interface Sci. 1 (1974) 46. [21] M. Elimelech, R. Williams, J. Gregory, X. Jia, Particle Deposition and Aggregation, Butterworth-Heinemann, 1998. [22] C.W. Turner, D. Lister, D.W. Smith, Deposition and removal of sub micron particles of magnetite at the surface of alloy 800, System Corrosion and Engineering Branch, AECL Research, Chalk River Laboratories, 11015, 1994, pp. 1–14. [23] N. Epstein, Elements of particle deposition onto nonporous surfaces parallel to suspension flows, Exp. Therm. Fluid Sci. 4 (1997) 14. [24] N. Kallay, E. Matijevic, Particle adhesion and removal in model systems IV. Kinetics of detachment of hematite particles from steel, J. Colloid Interface Sci. 1 (1981) 83.