Corrosion Science 75 (2013) 106–113

Contents lists available at SciVerse ScienceDirect

Corrosion Science

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

Water and corrosion resistance of epoxy–acrylic–amine waterbornecoatings: Effects of resin molecular weight, polar group and hydrophobicsegment

0010-938X/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.corsci.2013.05.020

⇑ Corresponding author at: School of Resource and Environmental Sciences,Wuhan University, Wuhan 430072, PR China. Tel./fax: +86 27 68775799.

E-mail addresses: liumhb@126.com (M. Liu), clab@whu.edu.cn (X. Mao),zhuhua333@126.com (H. Zhu), linanwd@126.com (A. Lin), wangdh@whu.edu.cn(D. Wang).

Min Liu a, Xuhui Mao a, Hua Zhu a, An Lin a, Dihua Wang a,b,⇑a School of Resource and Environmental Science, Wuhan University, Wuhan 430072, PR Chinab State Key Laboratory for Corrosion and Protection, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, PR China

a r t i c l e i n f o a b s t r a c t

Article history:Received 7 January 2013Accepted 24 May 2013Available online 5 June 2013

Keywords:A. Organic coatingsB. EISC. Polymer coatings

A two-step esterification process is developed for the synthesis of epoxy–acrylic-grafted-copolymerwaterborne resins. The effect of synthesis parameters on water and corrosion resistance of the water-borne coatings is investigated. The results reveal that moderate increasing of the resin molecular weight(<8000 Da) and carboxyl content (<27 wt.%) increased the crosslinking property, thereby improved theanticorrosion performance of the coatings. Longer epoxy-octanoic hydrophobic chains can provide stron-ger shielding effect on the hydrophilic portion of the polymer matrixes. The polar group content in awaterborne resin can be optimized for better anticorrosion performance, whereas the optimal value iscoating-specific.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Increasing environmental pressures are forcing the coatingindustry to minimize the release of volatile organic compounds(VOCs) and result in a continuous shift from solvent-borne coatingto waterborne coating [1,2]. At present, waterborne organic coat-ings have been wildly employed as building coatings and woodpaints, while their application as anticorrosive coatings for metalis still limited [3–6]. As is well known, hydrophilic componentswith polar groups or ionic groups are necessary for water-solubleor hydrosol resins. However, these polar groups are believed toform water penetration channels in the polymer matrixes of water-borne coatings [7,8]. The formation of a waterborne film is a com-plicated process consisting of three stages [8–15]: particle packing;particle deformation and compression; and particle coalescence. Inthe particle coalescence stage, molecular rearrangement occurredand polymer particles with the groups of similar polarity tendedto aggregate [7,16]. As a result, the polar groups in the particlesform polar channels for water permeation, accelerating the wateruptake in waterborne coatings and deteriorating the corrosionresistance. Perez et al. [5] reported that the water apparent diffu-sion coefficient (Dapp) for waterborne acrylic paint was 10 timeshigher than that of organic-solvent paints. Mikols et al. [17–19]

found that water existed in the polymer matrix in two distinctforms: free water filling the microcavities and bound water com-bined with polar groups of the polymer network. Viktor et al.[20–22] investigated that absorbed water mostly located in thehydrophilic part of the polymer matrixes and caused irreversibleswelling of polymers. Vanderwel et al. [23–26] found that thewater absorption caused by the hydrogen bonding between waterand polar groups was common in hydrophilic coatings, and the ab-sorbed water induced the swelling of the crosslinked polymers andweakened the mechanical property of the coatings.

Many efforts have been made to improve the anticorrosiveproperty of waterborne coatings. An optional method is to improvethe crosslinking property [1,27–31]. Chemical reactions betweendifferent polar groups not only facilitate crosslinks and thusenhance the physical and chemical integrity of the coalesced film,but also reduce the number of polar groups and thereby lower thewater sensitivity of the coatings [15,27,31]. Reactive polar groupsfor coating systems usually include epoxy groups, hydroxyl groups,carboxyl groups, amino groups, N-methylolacrylamide groups,blocked isocyanates and acetoacetate groups, etc. [7,15,27–31].Crosslinking property of a coating is known to be impacted bythe number density of these reactive polar groups. However, it re-mains controversial regarding the exact relation between them.Many studies suggested that more crosslinked units were neededin each polymer chain to obtain better crosslinking property[7,15]. Others reported that an increase in density of reactive polargroups resulted in poor crosslinking property and reduced tensilestrength of the cured films [32–34]. High reactive group content

M. Liu et al. / Corrosion Science 75 (2013) 106–113 107

caused fast crosslinking reactions, thereby suppressed furtherpolymer interdiffusion and lowered the gel content of the film [35].

Besides, the molecular weight of the polymer particles has beenconsidered as another important parameter for crosslinked films[7,14,36–38]. It is well known that coatings with high resin molec-ular weight are superior in barrier performance, because moreinterdiffusion of polymer chains occurred between large resin mol-ecules and led to a better crosslinking property [7]. However, somestudies also reported that higher resin molecular weight may exertnegative influences on the properties of the coatings. With highmolecular weight resins, the waterborne coatings showed lowerdispersibility and compatibility, hence poor film smoothness andprotective property [14,36].

In order to elucidate the effect of resin molecular weight andpolar group content on the barrier performance of waterbornecoatings, a two-step esterification process is developed in thisstudy to prepare various epoxy–acrylic-grafted-copolymer water-borne resins (EA resins), and the water and corrosion resistanceof their unpigmented waterborne coatings were evaluated. Unlikethe previously reported methods, the two-step esterification pro-cess was designed to accurately control the molecular weightand carboxyl content of the EA resins. Hence the prepared coatingsallow us to evaluate the effect of polar groups on the water andcorrosion resistance. The scientific hypothesis of this study is thatoptimum resin molecular weight and carboxyl content of the EAresins not only ensure good water dispersibility, but also promisefavorable crosslinking property and anticorrosion performancefor the coatings.

2. Materials and experimental

2.1. Chemical materials

Two types of bisphenol-A epoxy resins (E-12 and E-20), the typ-ical structure of which was depicted in Fig. 1, were purchased fromSinopec company (China) and were used as received. E-12 and E-20hold a number-average molecular weight (Mn) of �2000 and�1000, respectively. The curing agent was methylated amino resinat 80 wt.% solid content (Type SM5717, Sanmu Group Co., China).All the other chemical reagents were provided by SinopharmChemical Reagent Co. (China).

2.2. Preparation of epoxy–acrylic waterborne coatings

EA resins were synthesized by a two-step esterification process,as depicted in Fig. 2. The first esterification is to decrease the func-tionality of the epoxy resin. Epoxy monomers reacted with equalmole ratio n-octanoic acid at 105 �C for 100 min to produce theepoxy-octanoic ester (EP), and a certain amount of N,N-dimethy-lethanolamine (DMEA) was used as the catalyst. The mixed organicsolvents for the reaction include 50 wt.% propylene glycol methylether and 50 wt.% n-butyl alcohol. PA prepolymer was preparedvia a free radical polymerization of several acrylic monomers(see the recipe described in Table 1). At the beginning of reaction,one-third of portion I was added into a three-neck flask equippedwith a constant temperature magnetic blender. When the temper-ature rose to 75 �C, portion I, portion II and the remaining two-third of portion III were instilled into the flask by a constant flowpump. After 60 min of reaction, the temperature was elevated to

Fig. 1. General molecular structure of bisphenol-A epoxy re

85 �C and maintained till to the end of the dropwise addition(�90 min). Afterwards, the reaction system was kept at 85 �C foradditional 120 min. The second esterification step was the reactionbetween the prepared EP and PA prepolymers. Based on solid con-tents of the resultant solutions, the mass ratio of EP to PA was 3:7,and the reaction condition was the same as that of the first ester-ification. Finally, the obtained epoxy–acrylic graft copolymer (EAresin) was prepared.

For the preparation of an unpigmented epoxy–acrylic coating(EA coating), the EA resin was neutralized by DMEA and mixedwith amino resin (curing agent). Then the mixture of the two resinswas diluted to 45 wt.% solid content with distilled water prior touse. The addition amount of amino resin was determined by thereactive group contents of the EA resin.

2.3. Preparation of samples

Commercial aluminum foil and galvanized steel sheet, with thecompositions given in Table 2, were used as substrates for EA coat-ings. Samples on aluminum foils (150 mm � 150 mm � 0.01 mm)were used for water absorption test. Galvanized steel sheets(100 mm � 50 mm � 2 mm) were used as the substrates for elec-trochemical impedance spectroscopy (EIS) test and neutral saltspray (NSS) test. The metal substrates were degreased using ace-tone solvent and dried in air before the experiments. The coatingswere painted by roll coating method at ambient temperature(�25 �C), and were cured at 150 �C temperature for 30 min. Afterthe curing process, the samples were kept in desiccators for30 days before measured. The thicknesses of the obtained coatingsare 10 ± 2 lm, as measured by a coating thickness gauge (TT260,Beijing TIME Corp., China).

2.4. Measurements

Molecular weights of the PA and EA resins were determined bygel permeation chromatography (GPC) (2690D, Waters Corp., USA),using tetrahydrofuran (THF) as an eluent. Glass transition temper-ature (Tg) of the cured films was measured by differential scanningcalorimetry (DSC) (Q20, TA instruments, USA). The measurementswere carried out under the protection of nitrogen atmosphere witha scanning rate of 10 �C min�1 in the temperature range from�50 �C to 300 �C. The whole measuring process was operatedaccording to ASTM/D3418-82. Gel content, representing the insol-uble fraction of a cured film in a good solvent [7], was tested byextracting the cured film with refluxing 2-butanone in a Soxhletextractor for 10 h. Water absorption of the films was evaluatedfollowing a standard method (HG2-1612-1985). Viscosity of thecoatings was determined by a digital viscometer (DV-79, ShanghaiNirun Intelligent Technology Co., Ltd., China), at a constant temper-ature of 25 �C. Thermal decomposition behavior of the films wasexamined with a thermo-gravimetric analysis (TGA) (DiamondTG/DTA, PerkinElmer instruments, USA) under nitrogen flow. TheTGA spectra were acquired in the temperature range from 30 �Cto 650 �C at a heating rate of 5 �C min�1. NSS tests were conductedin a salt spray cabinet (Wuxi Tainuo Testing Equipments Co., Ltd.,China), with spraying NaCl solution (50 ± 5 g dm�3, pH = 6.5–7) at35 ± 0.5 �C, lasting for 600 h.

EIS was performed with an electrochemical workstation (PAR-STAT 2273, Princeton Applied Research, USA) at room temperature

sin (n = �5.5–6 for the E-12, n = �2–2.5 for the E-20).

Fig. 2. The schematic diagram for the two-step esterification process.

Table 1Recipe for the synthesis of the acrylic polymer (total mass weight: 150 g).

Portion number Feed name Weight (g)

I Methacrylic acid (MAA) 13Methymethacrylate (MMA) 15n-Butyl acrylate (BA) 72

II Azoisobutyronitrile (AIBN) 1.25n-Dodecyl mercaptan (NDM) 1.25Ethyl acetate 2.5

III Butyl acetate 20Propylene glycol methyl ether (PM) 13n-Butyl alcohol 12

Table 2Chemical composition of metal substrates.

Element Material

Galvanized steel sheet (wt.%) Aluminum foil (wt.%)

Zn 99.7 0.003Al 0.058 99.5Mn 0.065 0.013Si 0.031 0.108Fe 0.015 0.350C 0.060 –P 0.004 –S 0.008 –Cu – 0.014Ti – 0.017

108 M. Liu et al. / Corrosion Science 75 (2013) 106–113

(�25 �C). A three-electrode cell arrangement was used in theexperiments [10,38]: the galvanized steel sheet coated with

modified coating was set to be the working electrode with a circu-lar tested area of �13 cm2; a saturated calomel electrode (SCE) wasthe reference; and a platinum plate electrode with a dimension of10 mm � 10 mm was the counter. The frequency range was 10�2 to105 Hz, and the amplitude of the sinusoidal voltage was 10 mV.

3. Results and discussion

3.1. Preparation of the EA resins

In the first esterification step, the epoxy equivalent and acidvalue of the reactants were monitored (see Fig. A1 in the Supple-mentary data): epoxy equivalent increased as a function of reac-tion time, while the acid value decreased gradually. Thisobservation suggested the consumption of octanoic acid, and theprogress of epoxide ring-opening reaction. At the end of the ester-ification, the epoxy equivalent increased by twice of the originalvalue, which meant that half of the epoxy rings were exactlyopened when the octanoic acid was exhausted. As described inthe experimental section, the remaining epoxy groups in the resul-tants are designed to react with acrylic polymers in the secondesterification.

The molecular weights of the prepared PA are listed in Table 3.It can be observed that the polydispersity index (PDI) was less thantwo for all PAs, while the number-average molecular weight (Mn)of the PA changed with the amount of the initiator (AIBN) andchain transfer agent (NDM). As the additive amount decreased,the Mn of the PA increased from 2844 to 7142 Da accordingly.The relative narrow molecular weight distributions of the PA

Table 3Molecular weights of the prepared acrylic polymers with 20 wt.% MAA content underdifferent additions of initiator (AIBN) and chain transfer agent (NDM).

Acrylic polymers AIBN (wt.%)a NDM (wt.%)a Mn (Da) PDIb

PA I 3 4 2844 1.89PAII 2 2 4326 1.99PAIII 1.5 1 5278 1.96PAIV 1 1 5793 1.96PAV 1 0.5 7142 1.83

a wt.% is based on the total weight of the ingradients in Table 1.b PDI represents the polydispersity index.

Fig. 3. GPC traces of the reactants (EP and PA) and the product (EA) in the secondesterification step for the synthesis of Coating EA-12-20-65.

M. Liu et al. / Corrosion Science 75 (2013) 106–113 109

(PDI < 2) also suggest that the acrylic polymerization proceeded ina controlled manner [39].

Using different EP and PA prepolymers, a series of EA resinswere synthesized in the second esterification step, as shown inTable 4. Measurements using GPC and DSC were conducted tosee if the two polymers were successfully grafted. As presentedin Fig. 3, Mn of the product (the EA resin) is approximately thesum of the Mn of the reactants (the EP and the PA), indicating thatthe EP was expectedly grafted with the PA (Fig. 2). The comparisonof the DSC curves of the films (Fig. 4) further convinced thesuccessful preparation of EA resins. Four kinds of resins, EP, PA,the prepared EA and the blend of the EP and PA, were all curedby AR, but the resultant films exhibited different DSC curves. Forthe case of individual resin, the relaxation transition occurred at�0 �C for PA, and �40 �C for EP. For the blended one (BEA), tworelaxation transitions, at around 0 and 40 �C, appeared in the curve,suggesting the separated status of the EP and PA in the blend. Incontrast, the curve of the prepared EA resin looks very differentfrom the other three curves, suggesting that a homogeneouscopolymer was successfully prepared via the chemical graftingprocess presented in Fig. 2.

The EA resins and the properties of their unpigmented coatingswere listed in Table 4. PA resins with different Mn, ranging from2800 to 7100 Da (see Table A1 in the Supplementary data), wereselected for the synthesis of EA resins with different molecularweights. Also, MAA addition in the preparation of PA was adjustedto achieve different carboxyl contents for the produced EA resins,since other ingredients are void of strong polar groups likecarboxyl. Thus, the MAA addition (the mass ratio of the MAA tothe total mass of the acrylic monomers, 8–27 wt.% in Table 4) isa simple indicator that reflects the carboxyl content, and thusthe polarity of the EA coatings. In the following sections, theMAA addition, instead of carboxyl content, was used fordiscussion.

Table 4EA coatings prepared by EA resins with different molecular weights, carboxyl contents an

Coating type(EA-XXa-XXb-XXc)

Epoxy resin MAAaddition (wt.%)

Mn of PA in the resin

Mn (Da) PDI

EA-12-20-50 E-12 20 2844 1.89EA-12-20-65 E-12 20 4326 1.99EA-12-20-73 E-12 20 5278 1.96EA-12-20-80 E-12 20 5793 1.96EA-12-20-92 E-12 20 7142 1.83EA-12-8-73 E-12 8 5138 1.99EA-12-13-75 E-12 13 5533 1.91EA-12-27-74 E-12 27 5362 1.90EA-20-8-83 E-20 8 5963 1.95EA-20-13-80 E-20 13 5490 1.93EA-20-20-82 E-20 20 5793 1.96EA-20-27-81 E-20 27 5682 1.83

a Type number of epoxy resin (12 and 20 represent E-12 and E-20, respectively).b MAA addition (wt.%).c Abbreviation for the resin molecular weight.

3.2. Effect of molecular weight on water and corrosion resistance

‘‘EA-12-20-XX’’ in Table 4 represents a group of unpigmentedEA coatings prepared by the resins with different molecularweights. The water absorption of their cured films, as shown inTable 4, constantly decreases from 0.779% to 0.597% when the re-sin molecular weight increases. However, the decrease of waterabsorption seems to be weak when the resin molecular weightreached 8000 Da (from 0.601% to 0.597%). Fig. 5a shows the bodeplots of galvanized steel samples coated with different EA coatings.The initial log|Z| value increased as the resin molecular weight in-creased, although the increasing trend became not evident after8000 Da. This phenomenon is in accordance with the wateradsorption trend, suggesting that a higher molecular weight waspreferable for water and corrosion resistance.

In Table 4, it is notable that the increase in the molecular weightof the EA resins basically results in higher gel content (from 89.55%to 98.62%), even though the increasing trend was not obviouswhen the molecular weight increased to 8000 Da. This observationmeant that the crosslinking property, which is reflected by the gelcontent [7], was improved with higher resin molecular weight. Dueto the improved crosslinking property, better water and corrosionresistance of the samples was observed for the samples with higherresin molecular weights.

In order to quantitatively evaluate the anticorrosion perfor-mance of the EA coatings, the evolution of the low-frequency

d epoxy monomers.

Resin molecularweight

Waterabsorption (%)

Gel content (%) Viscosity(mPa s)

5000 0.779 89.55 364.16500 0.685 91.40 841.27300 0.620 98.33 2636.98000 0.601 98.31 5172.19200 0.597 98.62 6255.27300 1.144 89.53 –7500 0.772 92.357400 1.135 99.398300 1.002 87.958000 0.809 92.038200 1.075 97.478100 2.540 99.67

Fig. 4. DSC curves of the films prepared by different polymers cured with aminoresins. EP represents the epoxy-octanoic ester (EP-12); PA represents the acrylicpolymer (Mn = 4326; MMA addition is 20 wt.%); EA represents the copolymerproduct of EP and PA (the resin for Coating EA-12-20-65). BEA represents thephysical blend of the EP and the PA. All films were cured with amino resins.

Fig. 5. (a) Bode plots of the galvanized steel samples with EA coatings. Sampleswere immersed for 2 h prior to measurement. (b) Time dependence of the low-frequency impedance modules (|Z|f=10mHz) of the galvanized steel samples with EAcoatings. Testing was conducted in 3.5 wt.% NaCl solution. The coatings includedEA-12-20-50 ( ), EA-12-20-65 ( ), EA-12-20-73 (�), EA-12-20-80 ( ) and EA-12-20-92( ).

110 M. Liu et al. / Corrosion Science 75 (2013) 106–113

impedance module (|Z|f=10mHz) of the coated samples during a long-term immersion was depicted in Fig. 5b [40]. As can be seen,because of the penetration of water and electrolyte [41], theimpedance of the samples all presented a decreasing trend. Inthe case of the samples with resin molecular weights more than8000 Da (EA-12-20-80 and EA-12-20-92), the |Z|f=10mHz experi-enced a fast decrease in the time duration of 0–250 h. After500 h of immersion, the one that exhibited the best corrosionresistance was EA-12-20-73 (|Z|f=10mHz > 107 X cm2), followed byEA-12-20-80 and EA-12-20-92. Namely, there is an optimal resinmolecular weight for the long-term corrosion resistance of thecoating, and too high molecular weight may deteriorate the anti-corrosion performance. In Table 4, the viscosities of the coatingsprepared by ‘‘EA-12-20-XX’’ resins are listed. A dramatical increasein viscosity from around 300 to 6000 mPa s can be observed as theresin molecular weight increased. High viscosity resulted in poorpaintability of the coatings, causing defect points and inferiorsmoothness of the cured film [14,36]. For the EA-12-20-80 andEA-12-20-92 samples, we thought the excessive viscosityaccounted for their lower corrosion resistance in comparison withthe optimal one (EA-12-20-73) in EA-12-20-XX coating series. Asdiscussed above, increasing the resin molecular weight within acertain range seems to be a good way to improve the anticorrosionperformance of the coatings. Since the coatings with resin molecu-lar weights higher than 8000 Da was too viscous to provide goodpaintability, and hence did not show the best anticorrosion perfor-mance, in the following sections, the EA resins with molecularweights of 7000–8000 Da were selected for further study.

3.3. Effect of carboxyl content on water and corrosion resistance

The EA coatings with similar resin molecular weights (7300–7500 Da) but different carboxyl contents were denoted by‘‘EA-12-XX-7X’’ in Table 4. The MAA contents include four levels,8, 13, 20 and 27 wt.%. It is observed that the water absorption de-creases from 1.144% to 0.620% with the increased carboxyl content(MAA addition from 8 to 20 wt.%). However, for the coatings withthe highest MAA addition of 27 wt.%, the water absorption returnsto 1.135%, being equal to that with 8 wt.% MAA addition. A similartrend is observed in the bode plots of their coated samples (Fig. 6).After 2 h of immersion, the samples with higher MAA addition(20 wt.% and 27 wt.%) basically exhibited higher impedance, and

EA-12-20-73 sample held the highest |Z|f=10mHz of 109.4 X cm2

(Fig. 6a). At longer times of immersion, EA-12-20-73 sample al-ways kept the highest impedance, showing the best barrier perfor-mance for metal substrates (Fig. 6b). However, the EA-12-27-74sample demonstrated a fast decrease in the impedance values soonafter the start of the immersion. The |Z|f=10mHz value became evenlower than that of EA-12-13-75 sample after nearly 70 h of immer-sion (Fig. 6b), indicating that the 27 wt.% MAA addition greatlyweakened the anticorrosion performance of coating. The observa-tion of the 600 h NSS test (see Fig. A2 in the Supplementary data)further confirmed the results obtained in the impedance measure-ments. Rust spots appeared on the samples coated by CoatingEA-12-8-73 (at 124 h), EA-12-27-74 (at 264 h), and EA-12-13-75(at 360 h), while EA-12-20-73 sample did not exhibit any visiblerust spots at the end of the exposure (at 600 h).

The gel contents listed in Table 4 demonstrate that the cross-linking property of the films was improved by increasing the car-boxyl content, and the highest gel content of 99.39% wasachieved when the MAA addition was 27 wt.% (Coating EA-12-27-74). Considering the one with 20 wt.% MAA addition (CoatingEA-12-20-73) demonstrates the best water and corrosion resis-tance, it can be concluded that the improvement on crosslinkingproperty does not always result in better anticorrosive propertyof the coating. The variation of the relation was obviously associ-ated with the excess carboxyl content at 27 wt.% MAA addition.

Fig. 6. (a) Bode plots of the galvanized steel samples with EA coatings. Sampleswere immersed for 2 h prior to measurement. (b) Time dependence of the low-frequency impedance modules (|Z|f=10mHz) of the galvanized steel samples with EAcoatings. Testing was conducted in 3.5 wt.% NaCl solution. The coatings includedEA-12-8-73 ( ), EA-12-13-75 ( ), EA-12-20-73 (�) and EA-12-27-74 ( ).

Fig. 7. TG curves of the films prepared by different EA coatings (different MAAadditions), after 1000 h of immersion in distilled water.

Table 5The weight loss of the EA films revealed by TG curves.

Coating type Dm (%) Weight loss (%)

100 �C 200 �C Part I Part II

EA-12-13-75 1.15502 1.51030 1.15502 0.35528EA-12-20-73 0.55031 0.96395 0.55031 0.41364EA-12-27-74 0.92078 1.98352 0.92078 1.06274

Table 6Tg values of the EA coatings with different carboxyl contents before and afterimmersion.

Coating type Tg (�C) DTg

Before immersion After immersion for 250 h

EA-12-13-75 50.26 47.82 2.44EA-12-20-73 51.29 49.91 1.38EA-12-27-74 51.88 47.52 4.36

M. Liu et al. / Corrosion Science 75 (2013) 106–113 111

We thought the effect of carboxyl content on the anticorrosion per-formance of the EA waterborne coatings seemed like a double-edged sword: increase in the carboxyl content can improve thecrosslinking property, thus enhancing the barrier performance;

on the other hand, higher carboxyl content also increases the affin-ity of the polymer matrix to water, thereby weakening the corro-sion resistance of coatings.

In order to verify the role of the carboxyl content on the corro-sion property of an EA coating, TGA and DSC were employed toinvestigate the absorbed water in the unpigmented films. Fig. 7shows the TGA curves of the EA films after 1000 h of immersionin distilled water. According to the TGA curves, the weight losseswithin the temperature regions of 0–100 �C (Part I in Table 5),and 100–200 �C (Part II in Table 5) can be calculated, respectively.As generally recognized [21], the weight loss in the temperaturerange of 0–100 �C is associated with free water in the capillariesand microvoids within the polymer matrix, while the weight lossin the 100–200 �C range is related to the bound water combinedwith the hydrophilic groups of the polymer matrix [26]. It is ob-served that Coating EA-12-13-75, which has the lowest gel content,showed the largest weight loss between 0 and 100 �C (Part I), sug-gesting that inferior crosslinking property did lead to largerabsorption of free water; Coating EA-12-27-74, which has thehighest carboxyl content, showed the largest weight loss between100 and 200 �C (Part II), indicating that excess carboxyl groupsaccelerated the absorption of bound water. For Coating EA-12-20-73, the Part II weight loss is slightly larger than that of CoatingEA-12-13-75, whereas its total absorbed water (sum of Part I andPart II weight loss) is evidently the lowest. Water uptake wasthought to result in the degradation of the polymer matrix, andthus the reduction of Tg and mechanical properties [3,4]. As shownin Table 6, three films all revealed evident reductions of Tg after250 h of immersion in distilled water. The change extent in Tg

reduction (DTg) basically reflected the sequence of their adsorbedwater amounts.

Based on all the evidences above, the role of carboxyl contenton water resistance can be illustrated in Fig. 8. Too low carboxylcontent results in lower crosslinking level, and bigger water pene-tration channels including capillaries and microvoids are left be-tween polymer particles (see Fig. 8a). Hence, the largest amountof the absorbed free water (Part I in Table 5) was observed for Coat-ing EA-12-13-75. Excess carboxyl content, even though it enablesbetter crosslinking property, induces the aggregation of waterdue to the strong hydrogen bonding between water moleculesand hydrophilic groups, resulting in swelling of polymer and waterpermeation (see Fig. 8c). An optimal carboxyl content is the key tobalance the two sides of a coin; namely, favorable matrixes withtighter crosslinks and relative lower polarity can decrease the per-meation and absorption of both free and bound water (Fig. 8b). In

Fig. 8. The schematic diagram of water permeation in the EA coatings: (a) low MAA addition; (b) optimal MAA addition; and (c) high MAA addition.

112 M. Liu et al. / Corrosion Science 75 (2013) 106–113

this study, 20 wt.% MAA addition is demonstrated to provide theoptimum carboxyl content for enhancing water and corrosionresistance.

3.4. Effect of hydrophobic segments on water and corrosion resistance

Except for the molecular weight and carboxyl content, epoxymonomer was also a concerned factor for the anticorrosion perfor-mance of coatings. Two bisphenol-A epoxy resins, E-12 and E-20,which had different polymer chain lengths (Fig. 1), were used toprepare two series of EA coatings (denoted by ‘‘EA-12-XX-XX’’and ‘‘EA-20-XX-XX’’ in Table 4).

Coating EA-12-20-73 demonstrated the best anticorrosion per-formance in the coating series of EA-12-XX-XX since it showed thelowest water absorption (Table 4), highest bode impedance duringlong-term immersion (Figs. 5 and 6), and the best corrosion protec-tion on metal substrate in the salt spay test (Fig. A2 in Supplemen-tary data). In the case of EA-20-XX-XX series, the minimum waterabsorption of 0.809% (Table 4), as well as the highest impedance(initial and long times immersion, see Fig. 9), occurred at 13 wt.%MMA addition (EA-20-13-80), instead of at 20 wt.% MAA addition(EA-20-20-80). This observation implied that the optimal MMAaddition, or optimal polar group content, is actually dependent

Fig. 9. Time dependence of the low-frequency impedance modules (|Z|f=10mHz) ofthe galvanized steel samples with EA coatings during the immersion in 3.5 wt.%NaCl solution. The coatings included EA-20-8-83 ( ), EA-20-13-80 ( ), EA-20-20-82 (�) and EA-20-27-81 ( ).

on the type of the epoxy resin. Among the two series of coatings,the initial |Z|f=10mHz of the EA-12-20-73 sample is 109.4 X cm2,being higher than that of all other samples. Its |Z|f=10mHz value stillremained above 107 X cm2 after 500 h of immersion, demonstrat-ing the most admirable performance for corrosion protection.Furthermore, comparisons can be carried out between the coatingsderived from different epoxy resins: EA-12-13-75 vs EA-20-13-80,EA-12-20-73 vs EA-20-20-82, and EA-12-27-74 vs EA-20-27-81.The coatings in each group have similar resin molecular weights,gel contents, and equivalent carboxyl content due to the identicalMAA addition. As can be seen from Table 4, the EA coatings pre-pared with E-12 resin always showed lower water absorption.For example, for the group with 20 wt.% MAA addition, the wateradsorption of Coating EA-20-20-82 increased by 1.7 times as thatof Coating EA-12-20-73. The evolution of impedance (Figs. 6band 9) also confirmed that the coatings prepared with E-12(EA-12 series) were superior to the EA-20 series on anticorrosionperformance.

Because epoxy monomers with different chain lengths (E-12and E-20) were used for the preparation of EP resin, the resultantEA resins finally possess different lengths of hydrophobic epoxy-al-kyl chains. Since other factors like molecular weight and carboxylcontent were controlled at the same level, we thought the differ-ences of water and corrosion resistance between EA-12 and EA-20 coatings were associated with the lengths of the hydrophobicepoxy-alkyl chains. Longer hydrophobic chains in the moleculesof the EA-12 series resins may provide stronger shielding effecton the hydrophilic parts and lower the water sensitivity of thepolymer particles: the aggregation of absorbed water around thepolar groups can be hindered to some extent, and the water per-meation takes longer pathway due to the presence of hydrophobicalkyl chains. The shielding effect of alkyl chains also explains thedifferent optimal carboxyl contents for EA-12 and EA-20 coatings:E-12 resin, which holds the alkyl chain nearly two times longerthan that of E-20, can shield more hydrophilic groups in the EAmolecule. Thus, the optimal MAA addition for EA-12 series coatingsis 20 wt.%, being higher than the 13 wt.% for the EA-20 seriescoatings.

4. Conclusion

In this work, a two-step esterification process was designed tosynthesize EA grafted copolymers for waterborne coatings. It wasfound that, in a certain range, moderate increase in the resinmolecular weights and the carboxyl contents can improve thecrosslinking property of the coatings, thus enhancing the waterand corrosion resistance. Increasing viscosity at high molecular

M. Liu et al. / Corrosion Science 75 (2013) 106–113 113

weight (>8000 Da) leads to poor paintability of coatings and dete-riorates the long-term anticorrosion performance. Excess carboxylgroups cause significant increase in water permeation and absorp-tion, and thus result in faster decrease of mechanical and anticor-rosive properties. The hydrophobic chains in the molecules of EAresins favor the anticorrosive property of coatings due to theshielding effect on the hydrophilic portion. Moreover, the shieldingeffect is more evident as longer alkyl chains exhibit. Coating EA-12-20-73, which was prepared with E-12 epoxy resin and20 wt.% MAA addition, demonstrates the best water and corrosionresistance.

The two-step esterification process developed in this studyallows the precise control on the carboxyl contents in the copoly-mers, therefore, enables an in-depth study on the effect of polargroups on the water and corrosion resistance of waterborne coat-ings. It was verified that, although the polar groups are inevitablefor waterborne resin, their content can be optimized to achievethe better crosslinking property and lower water absorption, thusimproved anticorrosion performance can be obtained for water-borne coatings.

Acknowledgements

This research was financed by National Natural Science Founda-tion of China (Nos. 50771111 and 51071112), the Program for NewCentury Excellent Talents in Universities (NCET-08-0416), and theFundamental Research Funds for the Central Universities (GrantNo. 20112050202007).

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.corsci.2013.05.020.

References

[1] C. Andersson, New ways to enhance the functionality of paperboard by surfacetreatment – a review, Packag. Technol. Sci. 21 (2008) 339–373.

[2] V.D. Athawale, R.V. Nimbalkar, Waterborne coatings based on renewable oilresources: an overview, J. Am. Oil Chem. Soc. 88 (2011) 159–185.

[3] J. Jang, E.K. Kim, Corrosion protection of epoxy-coated steel using differentsilane coupling agents, J. Appl. Polym. Sci. 71 (1999) 585–593.

[4] W.G. Ji, J.M. Hu, J.Q. Zhang, C.N. Cao, Reducing the water absorption in epoxycoatings by silane monomer incorporation, Corros. Sci. 48 (2006) 3731–3739.

[5] C. Perez, A. Collazo, M. Izquierdo, P. Merino, X.R. Novoa, Characterisation of thebarrier properties of different paint systems Part II. Non-ideal diffusion andwater uptake kinetics, Prog. Org. Coat. 37 (1999) 169–177.

[6] P.A. Sorensen, S. Kiil, K. Dam-Johansen, C.E. Weinell, Anticorrosive coatings: areview, J. Coat. Technol. Res. 6 (2009) 135–176.

[7] J.W. Taylor, M.A. Winnik, Functional latex and thermoset latex films, JCT Res. 1(2004) 163–190.

[8] A. Wegmann, Chemical resistance of waterborne epoxy/amine coatings, Prog.Org. Coat. 32 (1997) 231–239.

[9] G.L. Brown, Formation of films from polymer dispersions, J. Polym. Sci. 22(1956) 423–434.

[10] W.A. Henson, D.A. Taber, E.B. Bradford, Mechanism of film formation of latexpaint, Ind. Eng. Chem. 45 (1953) 735–739.

[11] V.B. Miškovic-stankovic, D.M. Drazic, M.J. Teodorovic, Electrolyte penetrationthrough epoxy coatings electrodeposited on steel, Corros. Sci. 37 (1995) 241–252.

[12] Y.J. Park, J.H. Kim, Film formation from reactive latex particles: influence ofintraparticle crosslinking on mechanical properties, Colloid Surf. A –Physicochem. Eng. Asp. 153 (1999) 583–590.

[13] A.F. Routh, W.B. Russel, A process model for latex film formation: limitingregimes for individual driving forces, Langmuir 15 (1999) 7762–7773.

[14] L.-S. Tang, M. Zhang, S.-F. Zhang, J.-Z. Yang, High performance waterborneaminoacrylic coatings from the blends of hydrosols and latexes, Prog. Org.Coat. 49 (2004) 54–61.

[15] M.J. Yang, J.D. Zhang, Design and cross-linking reaction of blend of reactivepolymer latex particles, J. Mater. Sci. 40 (2005) 4403–4405.

[16] M.C. Goh, D. Juhue, O.M. Leung, Y. Wang, M.A. Winnik, Annealing effects on thesurface structure of latex films studied by atomic force microscopy, Langmuir9 (1993) 1319–1322.

[17] W.J. Mikols, J.C. Seferis, A. Apicella, L. Nicolais, Evaluation of structural changesin epoxy systems by moisture sorption–desorption and dynamic mechanicalstudies, Polym. Compos. 3 (1982) 118–124.

[18] J.M. Zhou, J.P. Lucas, Hygrothermal effects of epoxy resin. Part I: the nature ofwater in epoxy, Polymer 40 (1999) 5505–5512.

[19] J.M. Zhou, J.P. Lucas, Hygrothermal effects of epoxy resin. Part II: variations ofglass transition temperature, Polymer 40 (1999) 5513–5522.

[20] V. Baukh, H.P. Huinink, O.C.G. Adan, S.J.F. Erich, L.G.J. van der Ven, Water–polymer interaction during water uptake, Macromolecules 44 (2011) 4863–4871.

[21] G. Lendvay-Gyorik, T. Pajkossy, B. Lengyel, Water uptake of water-borne paintresin films as studied by impedance spectroscopy and gravimetry, Prog. Org.Coat. 59 (2007) 95–99.

[22] G.Z. Xiao, M.E.R. Shanahan, Swelling of DGEBA/DDA epoxy resin duringhygrothermal ageing, Polymer 39 (1998) 3253–3260.

[23] S. Cotugno, D. Larobina, G. Mensitieri, P. Musto, G. Ragosta, A novelspectroscopic approach to investigate transport processes in polymers: thecase of water–epoxy system, Polymer 42 (2001) 6431–6438.

[24] S.O. Han, L.T. Drzal, Water absorption effects on hydrophilic polymer matrix ofcarboxyl functionalized glucose resin and epoxy resin, Eur. Polym. J. 39 (2003)1791–1799.

[25] P.A. Tamirisa, D.W. Hess, Water and moisture uptake by plasma polymerizedthermoresponsive hydrogel films, Macromolecules 39 (2006) 7092–7097.

[26] G.K. van der Wel, O.C.G. Adan, Moisture in organic coatings – a review, Prog.Org. Coat. 37 (1999) 1–14.

[27] A. Blencowe, J.F. Tan, T.K. Goh, G.G. Qiao, Core cross-linked star polymers viacontrolled radical polymerisation, Polymer 50 (2009) 5–32.

[28] R.G. Joshi, T. Provder, P. Ziemer, W.J. Mao, W.D. Shen, F.N. Jones, Investigationof the effect of precoalescence or postcoalescence crosslinking on filmformation, properties, and latex morphology, J. Coat. Technol. Res. 6 (2009)47–65.

[29] H. Mouaziz, A. Larsson, D.C. Sherrington, One-step batch synthesis of highsolids monodisperse styrene/glycidyl methacrylate and styrene/methacrylicacid emulsion copolymers, Macromolecules 37 (2004) 1319–1323.

[30] G. Tillet, B. Boutevin, B. Ameduri, Chemical reactions of polymer crosslinkingand post-crosslinking at room and medium temperature, Prog. Polym. Sci. 36(2011) 191–217.

[31] J.D. Zhang, M.J. Yang, Y.R. Zhu, H. Yang, Synthesis and characterization ofcrosslinkable latex with interpenetrating network structure based onpolystyrene and polyacrylate, Polym. Int. 55 (2006) 951–960.

[32] F. Mazuel, C. Bui, B. Charleux, E. Cabet-Deliry, M.A. Winnik, Interdiffusion andself-cross-linking in acetal-functionalized latex films, Macromolecules 37(2004) 6141–6152.

[33] J. Yoon, P.A. Lovell, Model reaction studies and crosslinking of dimethyl-meta-isopropenylbenzyl isocyanate containing ambient crosslinkable latex,Macromol. Chem. Phys. 209 (2008) 279–289.

[34] S.W. Zhang, R. Liu, J.Q. Jiang, H.Y. Bai, Film formation and mechanicalproperties of the alkoxysilane-functionalized poly (styrene-co-butyl acrylate)latex prepared by miniemulsion copolymerization, Prog. Org. Coat. 65 (2009)56–61.

[35] H.H. Pham, M.A. Winnik, Polymer interdiffusion vs cross-linking in carboxylicacid-carbodiimide latex films. Effect of annealing temperature, reactive groupconcentration, and carbodiimide substituent, Macromolecules 39 (2006)1425–1435.

[36] D.R. Bauer, G.F. Budde, Crosslinking chemistry and network structure in highsolids acrylic–melamine coatings, Ind. Eng. Chem. Prod. Res. Dev. 20 (1981)674–679.

[37] D.R. Miller, C.W. Macosko, Network parameters for crosslinking of chains withlength and site distribution, J. Polym. Sci., B: Polym. Phys. 26 (1988) 1–54.

[38] J.T. Zhang, J.M. Hu, J.Q. Zhang, C.N. Cao, Studies of water transport behaviorand impedance models of epoxy-coated metals in NaCl solution by EIS, Prog.Org. Coat. 51 (2004) 145–151.

[39] G. Pan, L. Wu, Z. Zhang, D. Li, Synthesis and characterization of epoxy–acrylatecomposite latex, J. Appl. Polym. Sci. 83 (2002) 1736–1743.

[40] L.-K. Wu, J.-T. Zhang, J.-M. Hu, J.-Q. Zhang, Improved corrosion performance ofelectrophoretic coatings by silane addition, Corros. Sci. 56 (2012) 58–66.

[41] B.V. Jegdic, J.B. Bajat, J.P. Popic, S.I. Stevanovic, V.B. Miškovic-Stankovic, The EISinvestigation of powder polyester coatings on phosphated low carbon steel:the effect of NaNO2 in the phosphating bath, Corros. Sci. 53 (2011) 2872–2880.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具