Surface & Coatings Technology 204 (2009) 237–245

Contents lists available at ScienceDirect

Surface & Coatings Technology

j ourna l homepage: www.e lsev ie r.com/ locate /sur fcoat

Effect of nanoparticles on the anticorrosion and mechanical properties ofepoxy coating

Xianming Shi a,b,⁎, Tuan Anh Nguyen a, Zhiyong Suo c, Yajun Liu a, Recep Avci c

a Corrosion and Sustainable Infrastructure Laboratory, Western Transportation Institute, PO Box 174250, College of Engineering, Montana State University, Bozeman, MT 59717-4250, USAb Civil Engineering Department, 205 Cobleigh Hall, Montana State University, Bozeman, MT 59717-2220, USAc Imaging and Chemical Analysis Laboratory, Department of Physics, Montana State University, Bozeman, MT 59717, USA

⁎ Corresponding author. Western Transportation InstiP.O. Box 174250, Bozeman, MT 59717-4250, USA. Tel.: +994 1697.

E-mail address: Xianming_s@coe.montana.edu (X. S

0257-8972/$ – see front matter. Published by Elsevier Bdoi:10.1016/j.surfcoat.2009.06.048

a b s t r a c t

a r t i c l e i n f o

Article history:Received 10 January 2009Accepted in revised form 30 June 2009Available online 8 July 2009

Keywords:NanoparticleEpoxy coatingCorrosion resistanceNanoindentationSEMEISAFM

Homogeneous epoxy coatings containing nanoparticles of SiO2, Zn, Fe2O3 and halloysite clay weresuccessfully synthesized on steel substrates by room-temperature curing of a fully mixed epoxy slurrydiluted by acetone. The surface morphology and mechanical properties of these coatings were characterizedby scanning electron microscopy and atomic force microscopy, respectively. The effect of incorporatingvarious nanoparticles on the corrosion resistance of epoxy-coated steel was investigated by potentiodynamicpolarization and electrochemical impedance spectroscopy. The electrochemical monitoring of the coatedsteel over 28 days of immersion in both 0.3 wt.% and 3 wt.% NaCl solutions suggested the beneficial role ofnanoparticles in significantly improving the corrosion resistance of the coated steel, with the Fe2O3 andhalloysite clay nanoparticles being the best. The SiO2 nanoparticles were found to significantly improve themicrostructure of the coating matrix and thus enhanced both the anticorrosive performance and Young'smodulus of the epoxy coating. In addition to enhancing the coating barrier performance, at least anothermechanism was at work to account for the role of the nanoparticles in improving the anticorrosiveperformance of these epoxy coatings.

Published by Elsevier B.V.

1. Introduction

Epoxy has beenwidely used as a coatingmaterial to protect the steelreinforcement in concrete structures [1–3], because of its outstandingprocessability, excellent chemical resistance, good electrical insulatingproperties, and strong adhesion/affinity to heterogeneous materials.Epoxy coatings generally reduce the corrosion of a metallic substratesubject to an electrolyte in two ways. First, they act as a physical barrierlayer to control the ingress of deleterious species. Second, they can serveas a reservoir for corrosion inhibitors to aid the steel surface in resistingattack by aggressive species such as chloride anions.

Nonetheless, the successful application of epoxy coatings is oftenhampered by their susceptibility to damage by surface abrasion andwear [4,5]. They also show poor resistance to the initiation andpropagation of cracks [6]. Such processes introduce localized defectsin the coating and impair their appearance and mechanical strength.The defects can also act as pathways accelerating the ingress of water,oxygen and aggressive species onto the metallic substrate, resulting inits localized corrosion. Furthermore, being hydrophilic in nature,

tute, Montana State University,1 406 994 6486; fax: +1 406

hi).

.V.

epoxy coatings experience large volume shrinkage upon curing andcan absorb water from surroundings [7,8]. The pores in the curedepoxy coating can assist in the migration of absorbed water and otherspecies to the epoxy–metal interface, leading to the initiation ofcorrosion of the metallic substrate and to the delamination of thecoating.

The barrier performance of epoxy coatings can be enhanced by theincorporation of a second phase that is miscible with the epoxypolymer, by decreasing the porosity and zigzagging the diffusion pathfor deleterious species. For instance, inorganic filler particles atnanometer scale can be dispersed within the epoxy resin matrix toform an epoxy nanocomposite. The incorporation of nanoparticlesinto epoxy resins offers environmentally benign solutions to enhan-cing the integrity and durability of coatings, since the fine particlesdispersed in coatings can fill cavities [9–11] and cause crack bridging,crack deflection and crack bowing [12]. Nanoparticles can also preventepoxy disaggregation during curing, resulting in a more homogenouscoating. Nanoparticles tend to occupy small hole defects formed fromlocal shrinkage during curing of the epoxy resin and act as a bridgeinterconnecting more molecules. This results in a reduced total freevolume as well as an increase in the cross-linking density [13,14]. Inaddition, epoxy coatings containing nanoparticles offer significantbarrier properties for corrosion protection [15,16] and reduce thetrend for the coating to blister or delaminate.

238 X. Shi et al. / Surface & Coatings Technology 204 (2009) 237–245

This work examines the influence of nanoparticles, including SiO2,Zn, Fe2O3 and halloysite clay, on the surface morphology, antic-orrosion behavior and Young's modulus of epoxy coatings. It isexpected to shed more light on the fundamental mechanisms throughwhich nanoparticles interact with the epoxy matrix and thus provideguidance for the design of high-performance epoxy coatings used forcorrosion protection of steel.

2. Materials and methods

2.1. Materials

The epoxy resin and its hardener used in this research wereobtained from Phoenix Resins Inc. (Cinnaminson, NJ, USA), commer-cially known as MAS epoxies-FLAGTM. The liquid epoxy resin was ablend of multifunctional low molecular weight diluents and the di-glycedal ether of bis-phenol-A, whereas the hardener was based onadduction reaction chemistry of aliphatic amines. The weight ratio ofthe epoxy resin to the hardener was 2:1. Zn nanoparticles with a meandiameter of 35 nm, a specific surface area of 30-50 m2/g and facetedmorphology were purchased from Accumet Materials Co. (Ossining,NY, USA). Spherical SiO2 and Fe2O3 nanoparticles, obtained from MTIcorp. (Richmond, CA, USA), have a mean diameter of 15 and 20 nmand a specific surface area of 440 and N30 m2/g, respectively.Halloysite nanoclay (Al2Si2O5(OH)4·2H2O+SiO2) featuring a hollowcylindrical structure was purchased from Reade Advanced Materials(Reno, NV, USA). The steel coupons purchased from Metal Samples(Munford, AL, USA) were of Cor-ten B type (UNS number K11430;density: 7.60 g/cm3; chemical composition: C 0.10–0.19%, Cr 0.40–0.65%, Cu 0.25–0.40%, Fe 97.0–98.2%, Mn 0.90–1.25%, P≤0.04%, Si0.15–0.30%, S≤0.05%, V 0.02–0.10%) with a surface area of 2 cm2.Sodium chloride (NaCl) and acetone were purchased from FisherScientific (Pittsburgh, PA, USA).

2.2. Nanocomposite preparation

Steel substrate preparation: A copper wire was electrically con-nected to one surface of each cylindrical steel coupon, and then thissurface and all the other surfaces except the one exposed to electrolytefor corrosion testing were sealed with a thick bulk epoxy resin. Afterepoxy curing, the unsealed coupon surface was polished on siliconcarbide (SiC) papers down to a grid size of 1000 with the aid of ametallographic grinding disc. After polishing, the sample surface wasrinsed with tap water, sonicated in de-ionized water and then rinsedwith acetone.

Coating preparation: Epoxy composites are usually prepared bydispersing nanoparticles into the epoxymatrix either with a solvent orthrough a heating process. However, the latter process is prone toclustering or agglomeration of nanoparticles, resulting in poordispersion. The use of solvent is beneficial for dispersal of nanopar-ticles in the resin, but the curing agents are usually added to themixture after the solvent is removed by vacuum evaporation, whichdeteriorates the homogeneity of the nanocomposites after curing,especially for a high nanoparticle loading. To solve this problem, thecuring agent can be added to themixture before removing the solvent,which is expected to improve the dispersion of nanoparticles in thecoating. In addition, the slurry can be directly applied on the surface ofmetallic substrates to form a uniform thin barrier coating. In this work,acetone was chosen as the solvent, since the analyses by Fouriertransform infrared spectroscopy (FTIR) and FT-Raman [17] indicatedthat the sonication processing in acetone did not induce chemicalchange in the epoxy network. Before mixing, both resin and itshardener were diluted separately by acetone with a 1:1 weight ratio.Nanoparticles, which account for 1 wt.% of the total weight of resinand hardener, were added to the resin-acetone solution, followed bystirring at speeds up to 1550 rpm (Model 14-503, Fisher Scientific,

Inc.) and sonication (Model 50 T, VWR, West Chester, PA) for 10 min.After that, the hardener-acetone solution was added to the mixture,followed again by stirring and sonication for 10 min. The steelsubstrate was dipped into the finally obtained mixture for one timeand then kept in a dry place at room temperature for 7 days to allowfull curing, which led to the formation of a uniform coating for theanticorrosion and surface indentation tests in this work.

2.3. Morphological study of coatings

The surface morphology and thickness of the obtained nanocom-posite coatings were studied using Field Emission Scanning ElectronMicroscopy (FESEM). The films were removed from the steel couponsurface, and then sputter-coated with a very thin Iridium layer(approximately 1–2 nm) to avoid the charging effect caused by thenonconductive nature of epoxy coatings and to get high resolutionwith this virtually grain-free coating material. The surface morphol-ogies and the cross-section were analyzed by a Zeiss Supra 55VP PGT/HKL system, which offers an ultra-high resolution at a relatively lowvoltage.

2.4. Electrochemical characterization of coatings

Electrochemical measurements were conducted using a three-electrode system. The epoxy-coated steel coupon served as theworking electrode, while the counter electrode and the referenceelectrode usedwere a platinum grid and a saturated calomel electrode(SCE) respectively. The coatings evaluated in the electrochemicalmeasurements had similar thickness as those used in the morpholo-gical study since they were prepared following the same procedures.The corrosive solutions tested included 0.3 wt.% and 3 wt.% aqueousNaCl solutions. Two methods were used to test the anticorrosiveperformance of these nanocomposite coatings: electrochemicalimpedance spectroscopy (EIS) and potentiodynamic weak polariza-tion. Over the 28-day immersion of the coated steel, the EISmeasurements were carried out periodically using a Gamry ECM8Multiplexer. The steel was polarized at ±10 mV around its opencircuit potential (OCP) by an alternating current (AC) signal with itsfrequency ranging from 10 kHz to 10 mHz (10 points per decade).

In the potentiodynamic weak polarization tests, the steel waspolarized around its corrosion potential (−30 mV to 30 mV/SCE vs.OCP by a direct current (DC) signal at a scan rate of 0.2 mV/s.Polarization resistance (Rp) is defined by the slope of the potential-current density plot at the corrosion potential. Corrosion current(Icorr) is calculated from Icorr=B/Rp, assuming B=26 mV for the steelcorrosion.

2.5. Characterization of mechanical property

The mechanical properties of nanocomposite coatings weremeasured using the nanoindentation method based on Atomic ForceMicroscopy (AFM) [18]. For each sample, 256 pairs of force vs.displacement curves were obtained over the surface from a 5×5 μm2

area by subdividing the area into 16×16 equal-sized pixels andacquiring one pair of curves from the center of each pixel. Each pair ofcurves includes a loading and an unloading curve. The tip velocities forthese measurements varied between 0.3 and 0.5 μm/s. The stiffness ofeach sample was evaluated by fitting the corresponding loadingcurves to the Hertzian model which approximates the tip geometry asa hemisphere [16]. The curves are fitted in such a way that only theinitial part of the indention (up to 1000 nN loading force) was selectedfor fitting so that the indentation depth is limited to be less than100 nm, which is comparable to the tip radius. Such a smallindentation depth relative to the coating thickness, usually morethan 10 µm, is in consistence to the hemispherical tip geometryapproximation, and also minimizes the possible interference from the

239X. Shi et al. / Surface & Coatings Technology 204 (2009) 237–245

substrate contribution. Fitting the indentation data to the Hertzianmodel in some cases does not give very good results, but the obtainedYoung's modulus still remains close to the values obtained by othertechniques within 50% accuracy [19].

3. Results and discussions

3.1. Effect of nanoparticles on the morphology of epoxy coating

Typical top-view FESEM images of the control coating (plainepoxy, containing no nanoparticles) and the nanocomposite coatingsare shown in Fig. 1a–h. The average thickness of the epoxy coatingswas estimated from their cross-sectional view (as shown in Fig. 1f–h).We used freshly prepared epoxy-acetone solutions and followedidentical sample preparation procedures to prepare all the coatingsamples (with temperature, epoxy concentration, and pulling speed ofthe steel coupon out of coating solution stayed the same). It wasobserved that the plain epoxy coating and the epoxy coatingsmodified by nanoparticles of Zn, Fe2O3 and halloysite clay had asimilar thickness of approximately 40 μm, whereas the nano-SiO2

modified epoxy coating had a thickness of approximately 10 μm,according to the FESEM cross-sectional imaging of the coatings(Fig. 1f–h). This is mostly due to the significant reduction in theviscosity of the epoxy-acetone solution induced by the addition ofSiO2 nanoparticles.

Fig. 1a indicates that the cured plain epoxy coating has a relativelyhomogeneous morphology. The epoxy coating modified by SiO2

nanoparticles with a high specific surface area of 440 m2/g wasobserved to be much denser than the plain epoxy coating and showedno sign of nanoparticle agglomeration (Fig. 1c), partly attributable toits reduced internal stress inherent in the reduced coating thickness.The epoxy coating modified by Zn nanoparticles with a lower specificsurface area (30–50 m2/g) was also denser than the plain epoxycoating, but had some agglomeration of nanoparticles (Fig. 1b). In thecase of the epoxy coating containing Fe2O3 nanoparticles with a evenlower specific surface area (≥30 m2/g), even more agglomeration ofnanoparticles was observed along with aggravated microcracks (asshown in Fig. 1d). Based on the top-view and cross-sectional view ofthe coating, such microcracks were found to be localized near the topsurface of the coating and no crack was observed across the entirethickness of the coating layer or near the coating-steel interface. Wespeculate that the nanoparticles with higher specific surface area notonly served as better nano-fillers for the epoxy matrix, but also moreactively participated in the epoxy-curing process (possibly acting asnuclei for the growth of cross-linking epoxy-amine networks). Thesmall size of the nanoparticles is also advantageous since it enablestheir penetration into ultra-small holes, indentation and capillaryareas both in the coating matrix itself and at the metallic substrate. Inthe presence of halloysite clay, the nanocomposite epoxy coatingexhibited a textural structure with little agglomeration (see Fig. 1e),owing to the hollow cylindrical structure characteristic of thehoalloysite nanoparticles.

3.2. Effect of nanoparticles on the corrosion resistance of the coated steel

The corrosion potential, corrosion current, polarization resistanceand instantaneous corrosion rate were estimated from the measuredpotentiodynamic polarization curves of epoxy-coated steel.

Fig. 2a and b shows the temporal evolution of instantaneouscorrosion rate of the steel coated by various epoxy coatings, during the28-day immersion in 0.3 wt.% and 3 wt.% NaCl solutions, respectively.The incorporation of a small amount of nanoparticles (1% by totalweight of resin and hardener) into the epoxy coating significantlyreduced the corrosion rate of the epoxy-coated steel in bothelectrolytes, while the beneficial effect of nanoparticles was morepronounced in 0.3 wt.% NaCl solution than in 3 wt.% NaCl solution.

After 28 days of immersion, the nanoparticles reduced the corrosionrate of epoxy-coated steel by 638–2365 times in 0.3 wt.% NaCl solutionand by 11–910 times in 3 wt.% NaCl solution. For the steel protected bythe nanocomposite epoxy coatings, its corrosion rate increased by 20–1263 times when the chloride concentration increased from 0.051 M(0.3 wt.% NaCl) to 0.51 M (3 wt.% NaCl).

Fig. 3a and b shows the temporal evolution of polarizationresistance (Rp) of the steel coated by various epoxy coatings, in0.3 wt. % and 3 wt.% NaCl solutions, respectively. The incorporation ofnanoparticles into the epoxy coating significantly enhanced thepolarization resistance of the epoxy-coated steel in both electrolytes,while the beneficial effect of nanoparticles was more pronounced in0.3 wt.% NaCl solution than in 3 wt.% NaCl solution. It should be notedthat the measured Rp consisted of a component characteristic of thecoating-electrolyte interface inside the coating (indicating coatingporosity/compactness) and another component characteristic of thesteel-electrolyte interface (indicating charge transfer resistance). Forthe nonconductive coatings investigated, the Rp of the coated steel inthe relatively less corrosive 0.3 wt.% NaCl solution (Fig. 3a) can beused to estimate the relative void fractions within the coating matrix,assuming an inverse proportional relationship between the coatingcompactness and the measured Rp. On the other hand, the Rp of thecoated steel in the relatively more corrosive 3 wt.% NaCl solution(Fig. 3b) can be used to estimate the relative corrosion resistance ofthe steel at the steel-electrolyte interface.

Fig. 4a and b presents the Nyquist diagrams of the steel coated byvarious epoxy coatings, after 7-day immersion in 0.3 wt.% and 3 wt.%NaCl solutions, respectively. As shown in both figures, the Nyquistdiagrams derived from the EIS measurements featured two capacitiveloops, with the high-frequency loop (on the left) and the low-frequency loop (on the right) attributed to the resistance andcapacitance of the coating and of the steel-electrolyte interfacerespectively. These experimental EIS curves represent the electro-chemical process with two time constants, which are well separated.Equivalent electric circuit is generally used to interpret the EIS data.Equivalent electric circuit with two time constants had been used in[20] for epoxy coated steel with nano Ag pigment in 3.5 wt.% NaCl. Inthis research, the fitting of all EIS data was performed using a simpleequivalent electric circuit model (Fig. 5) with two time constants wellseparated. The obtained parameters are given in Tables 1 and 2, whereR1 and C1 are the resistance and capacitance of coating characteristicof its pore network structure (the coating-electrolyte interface insidethe coating), and R2 and C2 are the corrosion resistance of the steeland the double layer capacitance on the steel surface (the steel-electrolyte interface) respectively. If the coating was intact, anequivalent electric circuit with only one time constant (instead oftwo time constants as shown in Fig. 5) would be needed for analyzingthe EIS data. R0 is the solution resistance between the referenceelectrode and the working electrode (nanocomposite coated steel),which depends not only on the resistivity of electrolyte (ionicconcentration, type of ions, temperature and so on) but also on thegeometry of the area inwhich current is carried. R0 is not a property ofthe coating. Therefore, it is not technically or theoretically importantin the analysis of coating performance. The incorporation of a smallamount of nanoparticles (1% by total weight of resin and hardener)into the epoxy coating greatly increased the coating resistance R1 by6–1295 times and reduced the coating capacitance C1 by 1–112 times,indicating reduced coating porosity and improved barrier perfor-mance for corrosion protection of the steel substrate. It was alsoobserved that over the time of immersion in the corrosive electrolytesR1 decreased and C1 increased, indicating the entry of electrolyte intothe epoxy coatings, which is consistent with previous research [21,22].The incorporation of nanoparticles into the epoxy coating alsosignificantly increased the charge transfer resistance R2 by 3–186times and reduced the double layer capacitance C2 by 1–484 times,indicating enhanced corrosion resistance of the steel at the steel-

Fig. 1. SEM images of epoxy coatings. a): plain epoxy; b): with Zn nanoparticles; c): with SiO2 nanoparticles; d): with Fe2O3 nanoparticles; e): with halloysite clay; all of which weretop view at magnification level of approximately 100,000 times; f): cross-sectional view of the plain epoxy coating indicating a thickness of 37.5 μm; g) cross-sectional view of thenano-Zn modified epoxy coating indicating a thickness of 36.6 μm; and h) cross-sectional view of the nano-SiO2 modified epoxy coating indicating a thickness of 9.7 μm.

240 X. Shi et al. / Surface & Coatings Technology 204 (2009) 237–245

Fig. 2. Temporal evolution of corrosion rate of epoxy-coated steel in (a) 0.3 wt.% NaClsolution, and (b) 3 wt.% NaCl solution, as a function of nanoparticles.

Fig. 3. Temporal evolution of polarization resistance of epoxy-coated steel in (a) 0.3 wt.%NaCl solution, and (b) 3 wt.% NaCl solution, as a function of nanoparticles.

241X. Shi et al. / Surface & Coatings Technology 204 (2009) 237–245

electrolyte interface [23]. It should be noted that the charge transferresistances R2 obtained by fitting EIS were lower than the polarizationresistances Rp measured by potentiodynamic polarization curves. Apossible explanation is that surface area for steel-electrolyte interface,where steel was directly in contact with electrolyte, was lower than awhole surface area of steel coupon.

3.3. Effect of nanoparticles on the Young's modulus of epoxy coating

As the indentation depth in AFM experiments is less than 100 nmand the coating thickness is above 10 μm, the contribution of steelsubstrate to the final results is negligible. Fig. 6 presents the force-displacement curves for the control coating (plain epoxy, containingno nanoparticles) and the nanocomposite coatings, which wereobtained using the AFM-based nanoindentation technique. Theupper (black) curve and the lower (red) curve in each diagramcorrespond to the loading and the unloading processes of the inden-tation, respectively. A hysteresis between the loading and unloadingcurves was observed for all the tested samples, indicating a plasticdeformation of the epoxy coatings upon the indentation. The nearlylinear slope of the unloading curves in Fig. 6 indicates that the tipindentation resulted in very little elastic deformation and thusthe loading curve was fitted to obtain the mechanical properties ofthe coating. In this work we presume that all the coatings haveidentical mechanical properties over the whole thickness and theYoung's modulus obtained by fitting only the initial part of the loadingindentation process to the Hertzain model were presented in Fig. 7 asthe representative value of the whole coatings.

All the coatings, except the one modified with SiO2 nanoparticles,showed Young's modulus ranging from ∼60 to ∼350 MPa, which are

comparable to the documented values of 20–560 MPa for organiccoatings obtained using similar methods [24]. The epoxy coatingmodified with SiO2 nanoparticles showed a significantly enhancedYoung's modulus of ∼2.5 GPa, which coincided with the much smallerdeformation hysteresis in Fig. 6c relative to other samples (Fig. 6a, b, d&e). A smaller increase, ∼30%, in Young's modulus was observed forthe nano-Zn modified epoxy coating. As shown in Fig. 7, themodificationwith nanoparticles does not always enhance the stiffnessof the epoxy coatings: the nanoclay and nano-Fe2O3 modified samplesshowed ∼30% and ∼25% decrease relative to the unmodified epoxycoating, respectively.

The mechanical properties of nanocomposites, represented byYoung's modulus in this paper, depend heavily on the integrity andinternal properties of the coating surface, since under mechanicalstress the micro-voids between the nanoparticles or between thepolymer matrix and the nanoparticles may become the origin ofcracks. For the epoxy coating modified by SiO2 nanoparticles with ahigh specific surface area of 440 m2/g, the distinct improvement in itsstiffness may be ascribed to the following mechanisms. First, thenanoparticles tend to occupy holidays such as pinholes and voids inthe thin-film coating and serve as the bridges in the interconnectedmatrix, causing a reduction of the total free volume and anenhancement of the cross-linking density of the cured epoxy [13,14].As such, the cured nanocomposite coating has reduced chainsegmental motions and improved stiffness. Second, the SiO2 nano-particles may act to prevent epoxy disaggregation during curing andresult in a more homogenous coating [13,14]. Finally, the SiO2

nanoparticles resulted in a reduced viscosity of the epoxy solutionin acetone and thus led to a thinner coating layer on the steel, whichdiminished the internal stress of the cured coating.

Fig. 4. EIS Nyquist diagrams for nanocomposite-epoxy-coated steel after 7 days (a) in 0.3 wt.% NaCl solution, and (b) in 3 wt.% NaCl solution. Their impedance components wereplotted at full scale (left side) and low scale (right side).

242 X. Shi et al. / Surface & Coatings Technology 204 (2009) 237–245

The addition of other nanoparticles did not affect the Young'smodulus of the epoxy coating as prominent as the SiO2 nano-particles. We tentatively correlate this result with the lower specificsurface areas of Zn and Fe2O3 nanoparticles (30–50 m2/g and≥30 m2/g, respectively) relative to SiO2. In the case of Fe2O3, theagglomeration of nanoparticles in the cured nanocompositecoating led to aggravated microcracks on the coating surface, asevidenced by FESEM imaging (Fig. 1d), which further weakened thecoating and resulted in the smallest Young's modulus among all thetested samples. This highlights the importance of good dispersionof nanoparticles in delivering desirable mechanical properties ofnanocomposite epoxy coatings.

Unlike the other nanoparticles used this work that possess anapproximately spherical geometry, the halloysite clay nanoparticlesfeature a nanotubular structurewith an averages diameter of 30 nmand

Fig. 5. Schematic drawing of the equivalent circuit. R0 is associated with the electrolyteresistance. R1 and C1 are the resistance and capacitance of coating, respectively. C2 is thecapacitance of the double layer. R2 is the charge transfer resistance at the steel-electrolyte interface.

lengths between 0.5 and 10 µm [25]. It has been reported that theaddition of montmorillonite nanoclay which has a layered structure((Na,Ca)0.33(Al,Mg)2(Si4O10)(OH)2·nH2O), enhanced the modulus ofepoxy coatings [26–28]. However, our results indicated that thenanoclay modified epoxy coating exhibited a loss in the Young'smodulus compared with the plain epoxy coating. Although currently itremains unclear whether the weakening of the nanoclay-epoxycomposite in our experiments is associated with the hollow cylindricalstructure of the halloysite clay nanoparticles, such a difference suggeststhat the nanostructures of clay materials may play a significant role inthe mechanical properties of epoxy composite coatings.

3.4. Role of nanoparticles in enhancing the anticorrosive performance ofepoxy coating

According to the EIS data after 7-day immersion in both 0.3 wt.%and 3 wt.% aqueous NaCl solutions (Tables 1 and 2), the incorporation

Table 1Parameters of the equivalent circuits after 7 days in 0.3 wt.% NaCl solutions.

Coating samples R0 (Ω∙cm2) R1 (Ω∙cm2) R2 (Ω∙cm2) C1 (F∙cm−2) C2 (F∙cm−2)

Plain epoxy 373.7 1.44E+05 7.29E+05 1.11E−09 3.58E−06Epoxy+nano-Zn 362.1 4.26E+07 8.68E+07 5.78E−10 1.08E−07Epoxy+nano-SiO2 128.8 1.24E+08 9.07E+07 4.85E−10 1.25E−08Epoxy+nanoclay 238.7 3.70E+06 4.47E+07 5.89E−10 1.50E−07Epoxy+nano-Fe2O3 498.1 4.96E+07 1.36E+08 6.64E−10 7.37E−09

Fig. 7. Young's modulus of epoxy coatings doped with different nanoparticles. For eachsample, the data were averaged from four nanoindentation curves randomly selectedfrom 256 curves obtained.

Table 2Parameters of the equivalent circuits after 7 days in 3 wt.% NaCl solution.

Coating samples R0 (Ω∙cm2) R1 (Ω∙cm2) R2 (Ω∙cm2) C1 (F∙cm−2) C2 (F∙cm−2)

Plain epoxy 200.1 352.5 2.42E+04 6.64E−09 1.64E−05Epoxy+nano-Zn 366.4 4.24E+04 9.63E+04 3.35E−09 7.06E−06Epoxy+nano-SiO2 765.9 2.51E+03 7.28E+05 5.85E−11 7.39E−07Epoxy+nanoclay 254.2 3.60E+04 3.11E+06 1.78E−09 6.86E−06Epoxy+nano-Fe2O3 200.3 4.57E+05 9.03E+05 7.61E−10 4.31E−07

243X. Shi et al. / Surface & Coatings Technology 204 (2009) 237–245

of nanoparticles increased the coating resistance R1 and the chargetransfer resistance R2 while reducing the coating capacitance C1 andthe double layer capacitance C2. This suggests that at least twopossible mechanisms contributed to the enhanced corrosion protec-tion of nanocomposite epoxy coating. First, nanoparticles improvedthe quality of the cured epoxy coating, reduced the porosity of thecoating matrix, and zigzagged the diffusion path available bydeleterious species, leading to improved barrier performance of theepoxy coating. Second, nanoparticles improved the adherence of the

Fig. 6. Curves of force vs. displacement of a) plain epoxy; epoxy doped b) with Zn nanoparticles; c) with SiO2 nanoparticles; d) with Fe2O3 nanoparticles; e) with halloysite claynanoparticles.

244 X. Shi et al. / Surface & Coatings Technology 204 (2009) 237–245

cured epoxy coating to the underlying substrate and altered thephysiochemical properties of the coating-steel interface, the specificpathway of which is dependent on the type of nanoparticles asdescribed below.

Table 3 provides the open circuit potential (OCP) data of epoxy-coated steel in salt solutions, as a function of nanoparticle type, NaClconcentration and exposure duration. It should be cautioned that theOCP reading of the coated steel was contributed both by the corrosionpotential of the steel itself and by the electrical resistance of thecoating layer.

Incorporating 1 wt.% of nano-Zn particles with a mean diameter of35 nm into the epoxy coating reduced the corrosion rate of the epoxy-coated steel by 739 and 11 times, respectively, after 28-day immersionin 0.3 wt.% and 3 wt.% aqueous NaCl solutions. In the less corrosiveelectrolyte (0.3 wt.% NaCl), the OCP of the steel protected by the nano-Zn modified epoxy coating was 0.052 VSCE and −0.229 VSCE at 24 hand 672-hours respectively, which were significantly higher than thatof the plain-epoxy-coated steel (−0.457 VSCE and −0.584 respec-tively, as shown in Table 3).We speculate that due to the low dosage ofZn in epoxy coating, nano-Zn nanoparticles were quickly consumed toform ZnO, which worked as both an anodic-type inhibitor and a goodnano-filler to significantly inhibited corrosion of bare steel. This nobleshift in the OCP induced by the Zn nanoparticles decreased over timein 0.3 wt.% NaCl and diminished in the more corrosive electrolyte(3 wt.% NaCl), likely due to the cathodic dissolution of ZnO. It meritsfurther investigation to see whether a higher dosage of ZnOnanoparticles can provide better long-term anti-corrosive perfor-mance for the epoxy coating.

Zinc-rich primers have been extensively and successfully used forcorrosion protection in heavy-duty environments, often involvingmuch higher loading of zinc powder with mm and μm grain sizes[29,30]. Therefore, we also compared the performance of an epoxycoating with ordinary Zn particles (b150 μm, 1% by total weight ofresin and hardener) and that of the nano-Zn modified epoxy coating.In the first 24 h of exposure to 3 wt.% NaCl solution, the steel protectedby the nano-Zn-modified epoxy coating showed a significantly lowercorrosion rate (and similar OCP) than the steel protected by theordinary-Zn-modified epoxy coating. Yet over time such nano-effectdiminished and by the 9th day of immersion the two Zn-modifiedcoatings offered comparable corrosion protection for the underlyingsteel. It merits further investigation whether increasing the loading ofnano-Zn particles in the epoxy coating would significantly improve itslong-term anticorrosive performance.

Incorporating 1 wt.% of nano-Fe2O3 particles with amean diameterof 20 nm into the epoxy coating reduced the corrosion rate of theepoxy-coated steel by 2365 and 910 times, respectively, after 28-dayimmersion in 0.3 wt.% and 3 wt.% aqueous NaCl solutions. In additionto enhancing the coating barrier performance, Fe2O3 nanoparticlesserved as anodic-type corrosion inhibitor to significantly reduce thecorrosion of the epoxy-coated steel in both electrolytes. In the lesscorrosive electrolyte (0.3 wt.% NaCl), the OCP of the steel protected by

Table 3Open circuit potential of epoxy-coated steel in salt solutions, as a function ofnanoparticle type, NaCl concentration and exposure duration.

Coating on the steel OCP of coated steel in0.3 wt.% NaCl (V, vs. SCE)

OCP of coated steel in3 wt.% NaCl (V, vs. SCE)

24-hour 672-hour 24-hour 672-hour

Plain epoxy −0.457 −0.584 −0.603 −0.396a

Epoxy+nano-Zn 0.052 −0.229 −0.598 −0.613Epoxy+nano-SiO2 −0.597 −0.548 −0.551 −0.597Epoxy+nanoclay −0.170 −0.237 −0.583 −0.507Epoxy+nano-Fe2O3 0.060 −0.049 −0.437 −0.336

a The OCP of this plain-epoxy-coated steel generally increased over time withsignificant fluctuations and this high OCP readingmay be resulted from the formation ofcorrosion product on steel surface.

the nano-Fe2O3 modified epoxy coating was 0.060 VSCE and −0.049VSCE at 24h and 672-hours respectively, which were significantlyhigher than that of the plain-epoxy-coated steel (−0.457 VSCE and−0.584 respectively, as shown in Table 3). This noble shift in the OCPinduced by the Fe2O3 nanoparticles decreased over time in 0.3 wt.%NaCl and was much less apparent in the more corrosive electrolyte(3 wt.% NaCl). It should be noted that the nano-Fe2O3 particles did notdisperse very well in the epoxy coating (as shown in Fig. 1d) andbetter anti-corrosive performance of the nanocomposite coating canbe expected once better dispersion of the nanoparticles is achieved.

It is interesting to note that Fe2O3 nanoparticles were previouslyreported to alter the magnetic properties of epoxy resin as well [31].Traditionally, Fe2O3 pigment with mm and μm grain sizes was used inprotective paint as corrosion inhibitor, and its protective mechanismwas considered physical rather than chemical. By mechanicallystrengthening the paint film, reducing moisture permeation throughthe film, and screening out destructive UV radiation, Fe2O3 pigmentswas excellent auxiliary pigments in metal primers and top coats [30].With improved dispersion in the epoxy matrix, the nano-Fe2O3

particles are expected to provide long-term anticorrosive perfor-mance of the epoxy primers.

Incorporating 1 wt.% of nano-SiO2 particles with a mean diameterof 15 nm into the epoxy coating reduced the corrosion rate of theepoxy-coated steel by 983 and 32 times, respectively, after 28-dayimmersion in 0.3 wt.% and 3 wt.% aqueous NaCl solutions. This isconsistent with a previous study [32], where the incorporation of1 wt.% nano-SiO2 particles improved the anticorrosive performance ofthe epoxy coating on 2024-T3 aluminum alloy. The OCP data in Table 3suggest that in both electrolytes the corrosion protection offered bythe SiO2 nanoparticles had more to do with the improvement in thecoating pore network than any modification of the coating-steelinterface. The SiO2 nanoparticles tend to occupy holidays in the thin-film coating and serve to bridge more molecules in the interconnectedmatrix, leading to increased cross-linking density of the cured epoxyas well as improved corrosion protection for the steel substrate.

Incorporating 1 wt.% of halloysite clay nanoparticles into the epoxycoating reduced the corrosion rate of the epoxy-coated steel by 638 and614 times, respectively, after 28-day immersion in 0.3 wt.% and 3 wt.%aqueous NaCl solutions. In the less corrosive electrolyte (0.3 wt.% NaCl),the OCP of the steel protected by the halloysite clay nanoparticlesmodified epoxy coating was −0.170 VSCE and −0.237 VSCE at 24 h and672-hours respectively, which were significantly higher than that ofthe plain-epoxy-coated steel (−0.457 VSCE and−0.584 respectively, asshown in Table 3). The experimental results suggest that halloysitenanoclay caused the noble shift in the OCP of the epoxy-coated steel andsignificantly inhibited its corrosion, in the case of 0.3 wt.% NaCl. In themore aggressive electrolyte (3% NaCl), even though the halloysite claystill provided strong corrosion protection for the steel substrate, thenoble shift in the OCP was no longer evident, likely due to the cathodicdissolution of aluminum oxide or hydroxide.

Our findings are consistent with previous research in terms ofbeneficial effect of nanoclay on the anticorrosive performance ofcoatings. By incorporating 2–6 wt.% montmorillonite nanoclay intothe polyurethane coating, Chen-Yang et al. [33] demonstrated that thecorrosion rate of coated stainless steel in 5 wt.% NaCl solution wasreduced by about 10–30 times. Yeh et al. [34] reported that thecorrosion rate of coated steel in 5 wt.% NaCl solution could be reducedby about 10 times when incorporating 1 wt.% montmorillonitenanoclay into the siloxane-modified epoxy coating. In our study, thehalloysite nanoclays were found to drastically enhance the polariza-tion resistance (Rp) of the coated steel, which was 15MΩ∙cm2 after 1 hin 3 wt.% NaCl. This value is much higher than Rp of the steel coated bypolyurethane/montmorillonite nanocomposite coatings (300 KΩ∙cm2

after 30 min in 5 wt.% NaCl, [33], and Rp of the stainless steel coated bysiloxane-modified epoxy/montmorillonite nanocomposite coatings(400–600 KΩ∙cm2 after 5 h in 5 wt.% NaCl, [34]).

245X. Shi et al. / Surface & Coatings Technology 204 (2009) 237–245

4. Conclusions

Nanoparticles of Zn, SiO2, Fe2O3 and halloysite clay were success-fully dispersed into epoxy resin matrix at a concentration of 1% by thetotal weight of epoxy resin and its hardener. The electrochemicalmonitoring of the coated steel over 28 days of immersion in both0.3 wt.% and 3 wt.% NaCl solutions suggested the beneficial role ofnanoparticles in significantly improving the corrosion resistance ofthe coated steel, with the Fe2O3 and halloysite clay nanoparticlesbeing the best. The potentiodynamic weak polarization test revealedthat after 28 days of immersion the nanoparticles reduced thecorrosion rate of epoxy-coated steel by 638–2365 times in 0.3 wt.%NaCl solution and by 11–910 times in 3 wt.% NaCl solution. The EISmeasurements indicated that the incorporation of nanoparticlesincreased the coating resistance R1 and the charge transfer resistanceR2 while reducing the coating capacitance C1 and the double layercapacitance C2. In addition to enhancing the coating barrierperformance, at least another mechanism was at work to accountfor the role of the nanoparticles in improving the anticorrosiveperformance of these epoxy coatings.

The epoxy coating modified with SiO2 nanoparticles showed asignificantly enhanced Young's modulus of ∼2.5 GPa. Nonetheless, themodificationwith nanoparticles does not always enhance the stiffnessof the epoxy coatings. A ∼30% increase in Young's modulus wasobserved for the nano-Zn modified epoxy coating, whereas thenanoclay and nano-Fe2O3 modified samples showed ∼30% and ∼25%decrease relative to the unmodified epoxy coating, respectively.

For future research, it would be important to investigate thecombined use of Fe2O3, SiO2 and halloysite clay nanoparticles andexamine the potential synergy between them to allow simultaneousimprovements in both anticorrosion and mechanical properties ofepoxy coating. A nano-Fe2O3 and/or nanoclay modified epoxy primerwith a nano-SiO2 pigmented topcoat would be an interesting systemto explore. It is also necessary to investigate ways to improvedispersion of the nanoparticles in the coating matrix and make thetransition from solvent-based to waterborne epoxy. More in-depthstudy of the effect of nanoparticles on epoxy-curing dynamics andkinetics would further advance the knowledge base of suchnanocomopsite coating systems. To enable long-term anticorrosiveperformance of the nanocomposite coatings, one should alsoinvestigate the potential application of the nanoparticles as reservoirsfor the storage and prolonged lease of corrosion inhibitors. Forinstance, corrosion inhibitors can be absorbed by the nanoparticlesand released only during contact with moisture, thus endowing thecoating with self-healing properties [16].

Acknowledgements

This work was supported by the Research and InnovativeTechnology Administration under U.S. Department of Transportationthrough the University Transportation Center research grant. Wewould like to specially thank the ICAL facility at Montana StateUniversity for their help with the use of AFM and FESEM.

References

[1] F. Galliano, D. Landolt, Prog. Org. Coat. 44 (2002) 217.[2] A. Talo, O. Forsén, S. Yläsaari, Syn. Met. 102 (1999) 1394.[3] V.B. Miskovic-Stankovic, M.R. Stanic, D.M. Drazic, Prog. Org. Coat. 36 (1999) 53.[4] B. Wetzel, F. Haupert, M.Q. Zhang, Compo. Sci. Technol. 63 (2003) 2055.[5] M.Q. Zhang, M.Z. Rong, S.L. Yu, B. Wetzel, K. Friedrich, Macromol. Mater. Eng. 287

(2002) 111.[6] S. Yamini, R.J. Young, Polymer 18 (1977) 1075.[7] D. Perreux, C. Suri, Compo. Sci. Technol. 57 (1997) 1403.[8] C. Loos, G.S. Springer, J. Comp. Mater. 13 (1979) 131.[9] K. Lam, K.T. Lau, Compo. Struc. 75 (2006) 553.[10] G. Shi, M.Q. Zhang, M.Z. Rong, B. Wetzel, K. Friedrich, Wear 254 (2003) 784.[11] Hartwig, M. Sebald, D. Putz, L. Aberle, Macromol. Symp. 221 (2005) 127.[12] F. Dietsche, Y. Thomann, R. Thomann, R. Mulhaupt, J. Appl. Polym. Sci. 75 (2000) 396.[13] N.Huong, Improvementofbearing strengthof laminatedcompositesbynanoclayand

Z-pin reinforcement, PhD. Dissertation, University of New South Wales, Australia(2006).

[14] O. Becker, R. Varley, G. Simon, Polymer 43 (16) (2002) 4365.[15] L.H. Yang, F.C. Liu, E.H. Han, Prog. Org. Coat. 53 (2005) 91.[16] S.V. Lamaka, M.L. Zheludkevich, K.A. Yasakau, R. Serra, S.K. Poznyak, M.G.S.

Ferreira, Prog. Org. Coat. 58 (2007) 127.[17] H. Miyagawa, M.J. Rich, L.T. Drzal, J. Polym. Sci. B 42 (2004) 4384.[18] Z. Suo, F.T. Arce, R. Avci, K. Thieltges, B. Spangler, Langmuir 22 (2006) 3844.[19] F.T. Arce, R. Avci, I.B. Beech, K.E. Cooksey, B. Wigglesworth-Cooksey, J. Chem. Phys.

119 (2003) 1671.[20] L. Xue, L. Xu, Q. Li, J. Mater. Sci. Technol. 23 (4) (2007) 563.[21] V.B. Miskovic-Stankovic, D.M. Drazic, M.J. Teodorovic, Corros. Sci. 37 (1995) 241.[22] V.B. Miskovic-Stankovic, D.M. Drazic, Z. Kacarevic-Popovic, Corros. Sci. 38 (1996)

1513.[23] Z.Z. Lazarevic, V.B. Miskovic-Stankovic, Z. Kacarevic-Popovic, D.M. Drazic, Corros.

Sci. 47 (3) (2005) 823.[24] M.R. Vanlandingham, S.H. McKnight, G.R. Palmese, J.R. Elings, X. Huang, T.A.

Bogett, R. Eduljee, J.W. Gillespie, J. Adhesion 64 (1997) 31.[25] G.W. Brindley, Clays and clay minerals 1 (1952) 33.[26] B. Qi, Q.X. Zhang, M. Bannister, Y.W. Mai, Compo. Struc. 75 (2006) 514.[27] L. Wang, K Wang, L. Chen, Y. Zhang, C. He, Composite A 37 (2006) (1890).[28] G. Khanbabaei, J. Aalaie, A. Rahmatpour, A. Khoshniyat, M.A. Gharabadian, J. Macromol.

Sci. B 46 (2007) 975.[29] C.M. Abreu, M. Izquierdo, M. Keddam, X.R. Novoa, H. Takenouti, Electrochim. Acta

41 (15) (1996) 2405.[30] T.C. Patton, Pigment Hand Book, John Wiley & Sons, New York, NY, 1973.[31] J. Mira, J.A. Lopez-Perez, J. Rivas, M.A. Lopez-Quintela, R. Caciuffo, D. Rinaldi, D.

Fiorani, IEEE Trans. Magn. 33 (5) (1997) 3724.[32] H. Shi, F. Liu, L. Yang, E. Han, Prog. Org. Coat. 62 (2008) 359.[33] Y.W. Chen-Yang, H.C. Yang, G.J. Li, Y.K. Li, J. Polym. Res. 11 (2004) 275.[34] J. Yeh, H. Huang, C. Chen, W. Su, Y. Yu, Surf. Coat. Technol. 200 (2006) 2753.