A

Ga

b

a

ARR1AA

KPCECP

1

rs[caasrpbppiecct[ee

o

0d

Synthetic Metals 161 (2011) 313–323

Contents lists available at ScienceDirect

Synthetic Metals

journa l homepage: www.e lsev ier .com/ locate /synmet

lkyd coatings containing polyanilines for corrosion protection of mild steel

.S. Goncalvesa, A.F. Baldisseraa, L.F. Rodrigues Jr. a, E.M.A. Martinib, C.A. Ferreiraa,∗

LAPOL/PPGEM, Universidade Federal do Rio Grande do Sul, BP 15010, 91501-970, Porto Alegre, RS, BrazilLaboratório de Eletroquímica, Instituto de Química, Universidade Federal do Rio Grande do Sul, PB 15003, CEP 91501-970, Porto Alegre, RS, Brazil

r t i c l e i n f o

rticle history:eceived 16 July 2010eceived in revised form5 November 2010ccepted 28 November 2010vailable online 7 January 2011

a b s t r a c t

A study was carried out to investigate the performance of anti-corrosion coatings obtained from alkydpaints containing polyaniline and polyaniline derivatives applied on carbon steel. The polyaniline, thepolyaniline derivates and the paints were characterized through FTIR and Raman spectroscopies andthermogravimetric analysis. Cyclic voltammetry studies showed that polyaniline and its derivates gaveelectroactive properties to the paints. Accelerated corrosion experiments (salt spray and humidity

eywords:olyanilineoatinglectrochemical impedance spectroscopy

chamber) revealed a significant improvement in the performance of the coatings which contained theelectroactive polymer compared to conventional coatings. These results were verified by evaluation ofthe electrical resistance and capacitance measurements of the films using electrochemical impedancespectroscopy. For some of the samples, the Raman spectra demonstrated the presence of an oxide layerat the coating/substrate interface composed basically of Fe2O3 and Fe3O4. The results of this study rein-

otectings

orrosionaint

force the possibility of prlayer of oxides using coat

. Introduction

Conducting polymers (CP) have been extensively studied inecent years due to a great variety of possible applications ineveral fields, such as in energy storage systems, electrocatalysis1–3], electrodialysis membranes [4–7], sensors [8–10] and anti-orrosion coatings [11–16]. They exhibit different oxidation statesnd behave as electronic or mixed conductors [17]. The role ofCP coating is to prevent contact of the substrate with a corro-

ive environment and reduce the corrosion rate. In some cases, theedox behavior of the coating provides the substrate with anodicrotection [18]. The level of protection against corrosion affordedy a CP coating is dependent on both its structural and electronicroperties [13,18,19]. Among the well known conducting polymersolyaniline (PAni) and polypyrrole (PPy) have gained particular

nterest from researchers over the past two decades owing to theirlectronic applications and their potential use as anti-corrosionoatings [13,20,21]. Polyaniline is one of the most readily preparedonducting polymers [22], it has controllable electrical conduc-ivity, excellent environmental stability and easy processability23,24]. It has several oxidation forms, namely emeraldine salt,

meraldine base, pernigraniline and leucoemeraldine, but only themeraldine salt form is conductive.

The mechanisms proposed to explain the anti-corrosion actionf polyaniline on ferrous surfaces are many, however, most authors

∗ Corresponding author. Tel.: +55 51 33089413; fax: +55 51 33089414.E-mail address: ferreira.carlos@ufrgs.br (C.A. Ferreira).

379-6779/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.synthmet.2010.11.043

ing carbon steel against corrosion through the formation of a protectivewhich contain electroactive polymers.

© 2010 Elsevier B.V. All rights reserved.

agree that the passivation of iron and its alloys is promoted by theformation of adjacent layers of iron oxides [18,25–28] or nitride[29]. Wessling and Lu et al. [25,26] observed that the emeraldinesalt is immediately reduced on contact with iron in the presenceof water, with consequent oxidation of the metal and formation ofa double layer of oxides. In the presence of oxygen, PAni oxidizeand enables the reconstitution of the oxide layer if the coating isdamaged.

Spinks et al. [30] demonstrated the ability of either PAni or PPycoatings to provide corrosion protection of steel. The experimen-tal evidence shows that the PC galvanically coupled to the metalcauses an anodic shift in the corrosion potential and some authorsargue that this shift causes passivation and anodic protection. Thismechanism is supported by studies showing stable oxide formationon the metal surface after contact with the conducting polymer.Rohwerder et al. [31] postulated that continuous coatings of a con-ducting polymer will fail to provide corrosion protection in thepresence of larger defects, which they cannot passivate, and willshow a fast break-down of the whole coating by fast reduction.This phenomenon is caused by high cation mobility in the reducedpolymer, which is due to the gradual transformation of the polymerinto a “freeway” for fast cation transport with increasing progressof the reduction front.

Andrew et al. [32] state in their patent that the emeraldine base

(EB) form of PAni can perform better than the emeraldine salt (ES).Wei et al. [33] reported a similar behavior for PAni-EB and acid-doped forms of PAni as protective coatings for cold rolled steel inan aqueous NaCl medium. The PAni-EB was found to offer goodcorrosion protection as evidenced by the increase in the corrosion

314 G.S. Goncalves et al. / Synthetic M

Table 1Standard composition of the paints.

Component Quantity (wt%)

Alkyd resin 35.1Titanium dioxide 0.5Zinc oxide 3.0Green halogen 0.6

pons

osica(bEm

cadtpTb

2

2

moahbTmua

Hzwe

2

T(ltgpGp

a closed chamber with a relative humidity of 100% at 38 ◦C, in

Solvents 35.0Other 22.8Active component 3.0

otential and polarization resistance. This phenomenon may notriginate merely from the barrier effect of the coatings because theonconjugated polymers, such as polystyrene and epoxy, did nothow the same electrochemical behavior.

Chen et al. [34] reported that PAni-EB/epoxy resin coatingsffered efficient corrosion protection of mild steel in a 3.5% NaClolution, especially when the EB content was 5–10%. In a prelim-nary study, our group [35] observed that paints prepared withhlorinated rubber (CR) and PAni are able to protect mild steel inbetter way than paint prepared with only CR. Both forms of PAni

undoped, doped with HCl and DBSA) can offer better protection,ut the coating obtained from the paint prepared with 10% of PAni-B was found to have the best performance in the protection of theetal.In this study polymer coatings were obtained from alkyd paints

ontaining different types of CP (undoped and doped with differentcids). The aim of the work was to evaluate the influence of theopant of different PAni on the behavior of the coatings as well aso compare the performance of paints containing PC with a paintrepared using classical anticorrosion pigments (zinc chromate).he paints were prepared using equipment commonly employedy the paint industry.

. Experimental

.1. Synthesis of the conducting polymer

The conducting polymers were obtained by chemical poly-erization of the monomers o-ethoxyaniline (C2H5OC6H4NH2),

-methoxyaniline (CH3OC6H4NH2) and aniline (C6H5NH2), usingmmonium persulfate ((NH4)2S2O8) as the oxidant agent andydrochloric acid (HCl) for acidification of the medium. The ratioetween monomer and oxidant was the same for all syntheses.he (NH4)2S2O8 solution was added dropwise and the reaction wasaintained for 6 h at −5 (±3)◦C. The polymer obtained was filtered,

ndoped by stirring in an ammonium hydroxide (NH4OH) solutionnd then doped again with different acids.

Undoped polyaniline (PAni EB) and polyanilines doped withCl (PAni HCl), p-toluenesulfonic acid (PAni pTSA), dodecylben-enesulfonic acid (PAni DBSA), or camphorsulfonic acid (PAni CSA)ere obtained. Poly(o-methoxyaniline) (PM pTSA) and poly(o-

thoxyaniline) (PE pTSA) were doped with p-TSA.

.2. Paint preparation

The paints were prepared using pigment dispersion equipment.he components were weighed according to a defined formulationTable 1) and dispersed in a high rotation mechanical shaker andater mixed in a ball mill. Common alkyd resins are polyesters syn-hesized by reacting carboxylic acids such as phthalic anhydride,

lycerin and fatty acids from different vegetal oils [36]. The size ofarticles dispersion, evaluated with a Fineness of Grind Gage (BYKardner), was between 5 and 6 Hegman (40 and 30 �m). One of theaints contained zinc chromate (Zn2CrO4) as a pigment.

etals 161 (2011) 313–323

2.3. Preparation of the test specimens

The paints were applied on cold laminated SAE 1020 steel sheets(160 mm × 75 mm × 0.8 mm) using an airless gun. The test speci-mens were then left in a dust-free location until the total drying ofthe coatings.

2.4. Characterization of the samples

2.4.1. Infrared spectroscopyInfrared spectroscopy was carried out using a Perkin Elmer FTIR

spectrometer model Spectrum 1000. The samples of different poly-mers were mixed with KBr, pressed and then analyzed.

2.4.2. Raman spectroscopyRaman spectroscopy was performed with a Dilor-Jobin Yvon

Labram spectrometer equipped with a 1024-diode multichanneldetector using a He/Ne laser (17 mW) at an excitation wavelengthof 633 nm. The data were treated with the Labspec software andthe spectral interval ranged from 1200 to 2000 cm−1.

2.4.3. Thermogravimetric analysisThe thermogravimetric analysis was carried out with a TA

Instruments model TGA 2050 analyzer using Thermal Advantagesoftware for the data treatment. All samples were analyzed underN2 atmosphere using a heating rate of 20 ◦C/min.

2.4.4. Electrical conductivityThe conductivity measurements were obtained using the

four-point method. The polymeric material to be analyzed wascompressed into cylindrical forms with 25 mm diameter. The mea-surement system consists basically of a Signatone four-fixed tipholder (S-301-6 model) and a Keithley 2400 source.

2.4.5. Cyclic voltammetryThe cyclic voltammetry tests were conducted with an EG&G

potentiostat model PAR 373 controlled by the EG&G PAR M 270program. A three-electrode cell was used. An Ag/AgCl electrode wasused as the reference and a platinum grid as the counter electrode.The working electrode consisted of a platinum sheet onto which thepaints were applied, coating an area of 1 cm2. Platinum was usedbecause steel is much more active than the polymeric layers withinthe range of potential used in this test. The cyclic voltammetry wascarried out between −0.2 and 1.2 V (Ag/AgCl) with a scan rate of50 mV/s. The samples were submitted to 50 consecutive cycles intwo different media: 1.0 M HCl and 3.5 wt% NaCl.

2.4.6. Film thickness measurementThe magnetic induction method was applied to measure the

film thickness with a Fischer Instrumentation Ltd model DualscopeMP20 gage. The measurements were repeated 6 times and the aver-age and standard deviation were calculated.

The film thickness was measured and the average thicknessobtained was 50.6 (±1.1) �m, which was controlled to allow a com-parative analysis of the different coatings.

2.4.7. Humidity chamber experimentsThe humidity chamber experiments were carried out using

accordance with the ABNT standard, and also fulfilled the fol-lowing criteria: bubble size (NBR 5841/81), bubble density (NBR5841/81), rusting degree (NBR 5770/84) and underlayer migration(NBR 787/76).

G.S. Goncalves et al. / Synthetic M

Table 2Identification of the samples.

Sample Active component

T PAni HCl PAni HClT-PAni DBSA PAni DBSAT-PAni pTSA PAni TSAT-PAni CSA PAni CSAT-PAni EB Undoped PAniT-PM pTSA Poly(o-methoxyaniline) TSA

2

mq

eew[oEs

3

3

fTOTc

3

3

waotT

T-PE pTSA Poly(o-ethoxyaniline) TSAT-STD Zinc chromateT-BCO Without active component

.4.8. Electrochemical impedance spectroscopyThe impedance measurements were carried out with an EG&G

odel PAR 273 A/96 potentiostat, coupled to a Schlumberger fre-uency response analyzer model Solartron SI 1255 FRA.

For this experiment a three-electrode cell was employed, how-ver, the electrolyte was a 1.0 M HCl solution and the workinglectrode consisted of an SAE 1020 steel sheet on which the paintas applied. The exposed area of the coating was limited to 10 cm2

37]. The amplitude of sinusoidal perturbation was 10 mV at thepen circuit potential. The frequency range was 100 kHz to 10 mHz.xperiments were performed in duplicate and the results are pre-ented through Nyquist and Bode diagrams.

. Results and discussion

.1. Paint processing

Nine samples, with the same basic formula, were obtained, dif-ering only in the active component of protection against corrosion.he percentage of active component was 3 wt% related to the resin.ther means the sum of filler and active component, if present.able 2 provides a description of the samples and the active anti-orrosion component.

.2. Paint characterization

.2.1. Infrared spectroscopyInfrared analysis was carried out on all paint samples to verify

hether degradation of the electroactive polymer chain occurredfter the processing, and also to monitor for possible spectral effectsf interactions between the components of the basic formula andhe conductive polymer. Fig. 1 shows the spectra of samples T-BCO,-STD and T-PAni HCl.

Fig. 1. Infrared absorption spectra of

etals 161 (2011) 313–323 315

It was verified that the only changes in the spectra were thebands at 1494 cm−1 and 1465 cm−1 which appeared in the case ofthe paint containing PAni HCl. These bands are attributed to thepresence of the conductive polymer and are related to the stretch-ing of the C C link of the quinoid and benzenoid ring, present inthe spectrum of the pure polymer at 1558 cm−1 and 1474 cm−1. Thedisplacement of the bands to lower regions may be due to interac-tion between the conductive polymer and the resin of the paint.The other bands detected are attributable to the polymer (alkyd)used to produce the paints [38].

3.2.2. Raman spectroscopyThe Raman spectroscopy analysis highlighted the presence of

the active polymers in the paints. Fig. 2 shows the Raman spectrafor the PAni DBSA, T-PAni DBSA and T-BCO samples.

The spectrum for the paint sample T-PAni DBSA is, as a whole,the sum of the spectra of the polymer PAni DSBA and basic paintformula represented by the T-BCO sample, confirming that thepresence of the conductive polymer in the paint after the process-ing did not generated any significant modification in its chemicalstructure.

3.2.3. Thermal stabilityThe thermal stability characteristics of the coatings evaluated

through thermogravimetric analysis are very similar. As an exam-ple, Fig 3 shows the thermograms obtained for the samples T-PEp-TSA and T-BCO, respectively, which are almost identical, the onlydifference being the presence of around 3% conductive polymer inthe T-PE p-TSA sample.

According to Fig. 3, two small mass losses occur at betweenaround 130 ◦C, probably due to loss of humidity and residual sol-vents from the paint production process. From this temperature to600 ◦C it is possible to observe a considerable mass loss related todegradation of the base-resin of the paint. The peaks for the loss ofthe resin and the conductive polymer coincide (Fig. 4).

The results show that the coating can be used under conditionsthat do not surpass 170 ◦C, which is the onset temperature of theresin base degradation. The high residual percentage (49%) is dueto the amount of inorganic charges and additives used in the paintformula.

3.2.4. Cyclic voltammetry experimentsThe electrochemical stability of the paints was evaluated by

cyclic voltammetry experiments. The voltammograms display asimilar shape, as shown in Fig. 5. The paints containing PAni

T-BCO, T-STD and T-PAni HCl.

316 G.S. Goncalves et al. / Synthetic Metals 161 (2011) 313–323

DBSA,

rar

rsPpcbtcN

3

ippftT4

Fig. 2. Raman spectra: (a) PAni

etained electrochemical activity and stability even after 50 cycles,condition necessary to protect against corrosion according to data

eported in the literature [25,26].For the paint samples, the voltammograms show current peaks

elated to redox pairs in the same potential region as those pre-ented by the conducting polymer samples (voltammograms of theC alone are not displayed), showing that the electroactive com-onent is the conducting polymer. The current densities of theoatings are lower than those observed for the polymers alone. Thisehavior is to be expected, since the matrix in which the conduc-ive polymers are dispersed is not electroactive, and the polyanilineoncentration added to the paint is low. The electrolyte (HCl oraCl) did not influence the cyclic voltammetry results.

.2.5. Humidity chamber experimentsHumidity chamber experiments were carried out for all coat-

ngs. The T-PAni HCl and T-PAni EB samples showed the besterformances, since there was no blister formation, even for time

eriods longer than 668 h. The T-PE pTSA sample showed blisterormation with low density after 72 h of the experiment; however,he result remained unaltered until the end of the experiment. The-PM pTSA sample showed blister formation with low density after8 h, while the T-BCO and T-STD samples showed blister formation

Fig. 3. TGA thermogram of T-BCO.

(b) T-PAni DBSA and (c) T-BCO.

with high density in less than 48 h. The T-PAni DBSA, T-PAni CSAand T-PAni pTSA samples showed bubble formation with high den-sity in less than 24 h. All samples showed minimal rusting at theend of the experiment with the exception of the T-PM pTSA samplewhich showed a small degree of rusting.

The aspect of the T-PAni EB sample at the beginning and at theend of the experiment (668 h), and of the T-STD samples at thebeginning and after 48 h of the experiment, can be seen in Fig. 6.

3.2.6. Raman spectroscopyRegardless of the mechanisms proposed to describe the pro-

tection against corrosion using conductive polymers, most authorsagree that it is promoted by the formation of a double layer of Fe2O3and Fe3O4 [26,39].

In order to identify the oxides present at the coating/substrateinterface, Raman analysis was carried out after the samples under-went the humidity chamber experiment. Figs. 7 and 8 show theRaman spectra obtained on analyzing different points of the coat-

ing/substrate interface for the T-PAni EB, T-STD and T-BCO samples,respectively.

The spectrum of Fig. 7 shows an oxide layer composed of Fe2O3and Fe3O4 that is typical of the samples T-PE pTSA, T-PM pTSA, T-PAni HCl and T-PAni EB. A spectrum of a Fe3O4 sample was added to

Fig. 4. TGA thermogram of T-PE pTSA.

G.S. Goncalves et al. / Synthetic Metals 161 (2011) 313–323 317

l 1 M (

toioat

aiw

Fig. 5. Cyclic voltammograms of polymer paints in HC

his figure to help to interpret the peak at 653 cm−1, characteristicf this oxide and absent in the spectrum of Fe2O3 as can be seenn Fig. 8. The formation of a layer mainly composed of Fe2O3 wasbserved for the samples T-PAni CSA, T-PAni DBSA, T-PAni pTSAnd T-STD. On the other hand, no oxide formation was observed in

he case of the T-BCO sample (Fig. 9).

The formation of the Fe2O3 oxide in the T-STD sample can bettributed to the zinc chromate present in this coating. The Cr+6

ons reduce to Cr+3, oxidizing the Fe0 to Fe+3, which reacts withater or oxygen to form ferric oxide.

a–c) and 3.5 wt% NaCl (d–f), v = 50 mV s−1, 50 circles.

Same samples containing conductive polymers induced the for-mation of a layer composed of Fe2O3 and Fe3O4. The formation ofdifferent types of oxide is a function of the redox potential and thespecies with lower potential will be preferentially formed.

It is proposed that the formation of Fe3O4 occurs when the oxy-

gen concentration at the oxide–metal interface is deficient (Eqs.(1)–(5)). Thus, the presence of Fe3O4 in only some of the samplescontaining conductive polymers may be related to the quantity anddistribution of the polymer in the matrix of the base paint. It is pos-sible that the larger the quantity and the better the spread of the

318 G.S. Goncalves et al. / Synthetic Metals 161 (2011) 313–323

(b) T-

cbt

F

(

[

2

[

p

Fig. 6. Humid chamber experiments: (a) T-STD 0 h,

onductive polymer the more compact the first layer of Fe2O3 wille, inhibiting the passage of oxygen through the oxide, leading tohe formation of the Fe3O4 layer.

e0 + H2O = (FeOH)ads + H+ + e− (1)

FeOH)ads + H2O = [Fe(OH)2]ads + H+ + e− (2)

Fe(OH)2]ads + H2O = Fe(OH)3 + H+ + e− (3)

Fe(OH)3 = Fe2O3 + 3H2O (4)

Fe(OH)2]ads + Fe2O3 = Fe3O4 + H2O (5)

The presence of Fe2O3 and Fe3O4 oxides in the samples whichrovided a greater degree of protection may indicate that the pro-

Fig. 7. Raman spectra obtained through the analysis in different

STD 48 h, (c) T-PAni EB 0 h and (d) T-PAni EB 668 h.

tection is, indeed, promoted by metal oxidation and active polymerreduction, creating a protective layer, in agreement with the mech-anisms proposed by other authors.

It was also verified that the coatings with better performance inthe humidity chamber were those containing greater amounts ofpure polymer.

3.2.7. Electrochemical impedance spectroscopy experimentsThe following coatings were selected as representatives of the

electrochemical impedance results: T-STD and T-BCO which areclassical formulations and were used to compare the performanceof coatings containing conductive polymers; T-PAni EB and T-PE

points of the coating/substrate interface for the T-PAni EB.

G.S. Goncalves et al. / Synthetic Metals 161 (2011) 313–323 319

e coa

prt

tH1iTatepiwm

b

Fig. 8. Raman spectrum obtained at th

TSA, which showed the best performance in the accelerated cor-osion experiments; and T-PAni pTSA, to observe the influence ofhe dopant acid. The results were compared to uncoated steel.

The Nyquist diagram for the uncoated sample showed capaci-ive arcs at the open circuit potential, soon after its immersion inCl (Fig. 10a). The polarization resistance (RP) fluctuated between0 and 20 � cm2. On the other hand, the solution resistance (Rs)

ncreased from 5 to 45 � cm2 after the first 24 h of the experiment.his Rs increase (shift of the arc on the real impedance axis) can bettributed to the intense formation of insoluble corrosion productshat diminish the conductivity of the medium next to the workinglectrode. After 144 h of the experiment a thick layer of corrosionroducts could be observed over the exposed surface and Warburg

mpedance appears on the Nyquist diagram represented by an axis

ith a 45◦ inclination from the horizontal, which characterizes aass transport process by diffusion.All samples coated with paints containing PAni showed similar

ehavior in the initial hours of the experiment (Figs. 11 and 12). A

Fig. 9. Raman spectrum obtained at the coat

ting/substrate interface for the T-STD.

line almost parallel to the imaginary impedance axis shows a capac-itive behavior in the first moments of immersion. At this point thecircuit of the electrochemical cell is open, with no contact betweenthe electrolyte and the working electrode due to the high electricalresistance of the film.

One hour after immersion all samples showed a capacitive arch,characteristic of an RC representation in parallel. The larger diame-ters of the arcs for the coated samples (Figs. 11a and 12a) comparedto the non-coated steel (Fig. 10a) show that R is related to the resis-tance of the film coating (Rf) and C to the film capacitance (Cf). Thediameter of these capacitive arcs fluctuates with immersion time,indicating a variation in the film resistance, Rf. In the first 3 h Rfdecreases and thereafter it shows an oscillatory behavior.

This decrease in the film resistance (Rf) during the initial

moments of immersion is attributed to the contact establishedbetween the working electrode and the electrolyte by the dif-fusion of the latter through the pores of the coating up to thesubstrate/coating interface. Some samples, such as T-PE pTSA and

ing/substrate interface for the T-BCO.

320 G.S. Goncalves et al. / Synthetic Metals 161 (2011) 313–323

150100500

0

50

100

150

200

1 H

z

100

mH

z

10 m

Hz

Z (ohm cm )

-Z (

ohm

cm

)

Uncoated steel

0 hour1 hour3 hours5 hours7 hours24 hours144 hours240 hours288 hours360 hours

-jZ''

(ohm

cm

2 )

Z' (ohm cm2)

10 20

0

10

30 35 40 45 50

0

5

10

15

20

10-3 10-1 101 103 105

5x100

101

1,5x101

2x101

2,5x101

3x101

Log (f,Hz)

Log

(|Z

|, oh

m.c

m)

Log

(|Z

|, oh

m.c

m2 )

Log (f,Hz)

Uncoated steel

0 hour 1 hour 3 hours 5 hours 7 hours 24 hours 144 hours 240 hours 288 hours360 hours

a

b

Fo

Tfid

raawp

afmcrtp

vmT

80006000400020000

0

2000

4000

6000

8000

1 H

z

100

mH

z

10 m

Hz

10 H

z

1 H

z

10 m

Hz

100

mH

z

T-Pani EB

0 hour1 hour3 hours5 hours7 hours24 hours144 hours240 hours288 hours360 hours

Z (ohm cm )

-Z (

ohm

cm

)

-Z (

ohm

cm

2 )

Z (ohm cm2)

10-3 10-1 101 103 105101

102

103

104

Log (f,Hz)

Log

(|Z

|, oh

m.c

m)Lo

g (|

Z|,

ohm

.cm

2 )

Log (f,Hz)

T-Pani EB

0 hour 1 hour 3 hours 5 hours 7 hours 24 hours 48 hours 216 hours 384 hours

b

a

ig. 10. Nyquist diagram (a) and Bode diagram (b) for uncoated steel, in HCl 1 M, atpen circuit potential.

-STD, unlike the others, still show high resistance values in therst hour of the experiment, probably due to the samples havingifferent porosity characteristics.

After 3 h of the experiment, all samples showed a minimum filmesistance value, except for the T-PAni EB sample which showedminimum value in the first hour. The Nyquist diagram shows

n inductive arch at low frequencies for some immersion times,hich may be related to oxidative dissolution of the substrate afterermeation of the coating by the electrolyte.

With the corrosive attack, the resistance values, both of the filmnd the electrolyte, increase again, the most likely cause being theormation of corrosion products which block the pores, making

ass transport more difficult. The increase in electrolyte resistancean be observed in the Nyquist diagram through the shift in theesistance rates at a frequency of 100 kHz to higher values andhis may be associated with the formation of insoluble corrosion

roducts.

For all samples, the film resistance increases until a maximumalue is reached at around 5 and 7 h of the experiment. The maxi-um resistance values for the film coatings vary with sample type.

he coatings containing electroactive polymers show values in the

Fig. 11. Nyquist diagram (a) and Bode diagram (b) for T-PAni EB in HCl 1 M at opencircuit potential after various times of exposure.

range of 4.5–6.0 k� cm2, while the standard sample showed a valueof 2.5 k� cm2. At this point of the experiment, the Nyquist diagramsno longer show inductive components at low frequencies, probablydue to interruption of the corrosion process with permanent oxideformation under the metallic surface.

This difference in the Rf values reveals the possibility that thecoating acts as a physical barrier, which separates away from themetallic substrate and moves to the medium and also forms a redoxpair, in which the conductive polymer reduces, oxidizing and pas-sivating the base metal.

After reaching the maximum resistance (Rf) values, the sam-ples showed different behaviors. In the case of the T-BCO sample,the resistance values continuously decreases until values close tothose of non-coated steel, which may be related to the subcuta-neous delamination of the coating, and subsequent detachment ofthe film.

The T-STD sample showed a significant decrease in the resis-tance values after they peaked. However, the values remainedstable at around 1.0 k� cm2 until the end of the experiment.

The T-PAni pTSA sample (Fig. 12) showed a similar behavior tothe T-BCO sample, with a continuous drop in the resistance until

G.S. Goncalves et al. / Synthetic M

500040003000200010000

0

1000

2000

3000

4000

5000

10 H

z

1 H

z

100

mH

z

10 m

Hz

Z (ohm cm )

-Z (

ohm

cm

)

Pani TSA

0 hour1 hour3 hours5 hours7 hours24 hours48 hours264 hours336 hours384 hours

-jZ''

(ohm

cm

2 )

Z' (ohm cm2)

0 200 400 600

0

200

400

600

10-3

10-2

10-1

100

101

102

103

104

105

106

101

102

103

Log (f,Hz)

Log

(|Z

|, oh

m.c

m)Lo

g (|

Z|,

ohm

.cm

2 )

Log (f,Hz)

T-Pani TSA

0 hour 1 hour 3 hours 5 hours 7 hours 24 hours 48 hours 264 hours 336 hours384 hours

b

a

Fo

ri

fce

2o

ta

pTmrc

under the film. For some samples, Rf reaches a constant value after50 h of immersion indicating that the development of paths which

ig. 12. Nyquist diagram (a) and Bode diagram (b) for T-PAni pTSA in HCl 1 M atpen circuit potential after various times of exposure.

eaching values similar to those of the non-coated steel. This behav-or is consistent with the humidity chamber experiments.

The T-PE pTSA sample showed stability in the maximum valuesor up to 24 h, and thereafter the resistance decreased to valueslose to 2.5 k� cm2 and then remained stable until the end of thexperiment.

The T-PAni EB sample (Fig. 11) showed a drop to values close to.5 k� cm2 after peaking, and then remained stable until the endf the experiment.

The formation of stable oxide layers, although not confirmed, ishe most likely cause of this behavior, since the resistance valuesre comparable to those of the oxide protectors.

The stability of the resistance at relatively high values for longeriods during the experiments in the case of the T-PAni EB and-PE pTSA samples indicates an improvement in the coating perfor-

ance compared to the reference samples T-BCO and T-STD. These

esults are in agreement with those obtained in the acceleratedorrosion experiments.

etals 161 (2011) 313–323 321

Electrolyte resistance (Re), film resistance (Rf) and film capaci-tance (Cf) values were obtained from the log |Z| versus log(f) Bodediagrams of the different systems (Figs. 11b and 12b). The Re val-ues were obtained through extrapolation from the high frequencyregion level and Rf from the low frequency region level (Re + Rf).

The slope of the curve in the intermediate frequency regionallows an evaluation regarding whether the system is predomi-nantly resistive, capacitive, or inductive, or even a combinationof these properties. All extrapolated lines displayed slope valuesclose to −1, showing that the capacitive behavior predominates inthe total impedance, in this frequency range. The linear correlationcoefficient was also calculated and all curves had values close to 1,showing a negligible dispersion of the points around the line.

The capacitance C value was determined from the Bode diagram,using the following equation:

C = 1

2�f∣∣Z

∣∣

(6)

where f is the frequency, in Hz.Fig. 13 shows the variation in the capacitance C with sample

immersion time.The T-PAni pTSA, T-PAni EB and T-BCO samples showed an

increase in C followed by a decrease in the first 5 h of the exper-iment. The T-PE pTSA and T-STD samples showed no capacitancepeaks in the first few hours. This may be related to the differences inthe porosity of the coatings. The non-coated steel sample showedhigh C values compared to the coated samples.

The T-BCO and T-PAni pTSA samples showed a continuousincrease in C for the duration of the experiment, which may indi-cate that the electrolyte is continuously entering the film pores, orthat the metal surface under the film is increasing in area becauseof a corrosion process.

The T-STD, T-PE pTSA and T-PAni EB samples showed stable Cvalues after the first 50 h of the experiment, evidencing that all filmpores were filled with electrolyte. The maximum C value for the T-STD sample was approximately 2.5 × 10−5 F/cm2, while for the T-PEpTSA and T-PAni EB samples the maximum capacitance rates werearound 8.0 × 10−6 F/cm2 after almost 400 h of the experiment.

The capacitance values observed at the end of the experimentfor the T-PE pTSA and T-PAni EB samples are of the same order ofmagnitude as the C values of compact oxides, which upholds thepossibility that a protective oxide is formed due to the presence ofthe conductive polymers, in agreement with the results obtainedin the previous experiments described.

Regardless of the protection mechanism, the low capacitancevalues of the T-PE pTSA and T-PAni EB samples show a significantimprovement in the performance of these coatings compared to thevalues for the T-BCO sample, which is the reference and containsthe classical anticorrosion pigment zinc chromate.

The Rf values of the film coatings were plotted as a function oftime after the first hour of immersion and are presented in Fig. 14.

As previously observed, the resistance of the coated samples ini-tially increases and then decreases, reaching stable values withinthe experimental period for some samples. The high Rf at the begin-ning of the test shows the dielectric character of the film thatseparates the metallic substrate from the solution. The Rf valuedrops over time, due to the absorption of electrolyte by the filmthrough the pores until it reaches the metal and leads to corrosion

the electrolyte moves through in the film pores [37], or the trans-port of corrosion products from the metallic substrate out of thefilm, reaches a stationary state.

322 G.S. Goncalves et al. / Synthetic Metals 161 (2011) 313–323

Fig. 13. Capacitance versus time for different coatings.

4003002001000

0

1000

2000

3000

4000

5000

6000

Time (hours)

Res

iste

nce

(ohm

.cm

2 )

Uncoated steel T-BCO T-STD T-Pani pTSA T-Pani EB T-PE pTSA

Rf (

ohm

.cm

2 )

Tim

0 10 20 30

0

2000

4000

6000

rsus t

4

np

s

co

wop

Fig. 14. Film resistance ve

. Conclusions

Infrared and Raman spectroscopy demonstrated that there waso degradation of the polymers during incorporation into the baseaint.

All paint samples presented electroactivity in 1 M HCl and someamples presented electroactivity in 3.5% NaCl.

Incorporation of undoped PAni and PAni doped with HCl intoonventional coatings leads to an improvement in the performance

f the coatings in the humidity chamber.

Raman spectroscopy analysis of the coating samples that under-ent the humidity chamber experiment revealed the presence

f an oxide layer at the coating/substrate interface, mainly com-osed of Fe2O3 and Fe3O4 in the samples containing PAni EB,

e (hours)

ime for different coatings.

PAni HCl and PE pTSA. The fact that the oxide layer is presentin the samples that showed the best results constitutes strongevidence that the protection in this experiment was promotedby the presence of the oxide. Samples for which only Fe2O3 wasdetected displayed a lower performance in the corrosion experi-ments.

Electrochemical impedance spectroscopy experiments carriedout in 1.0 M HCl aqueous solutions showed that, after the first 24 hof experiment, some coatings had capacitance values of the same

order of magnitude as those of protective oxides. In these samesamples, both the resistance and capacitance remained stable after15 days of immersion.

Overall, the results presented herein reinforce proposals thatprotection of carbon steel against corrosion occurs through the for-

etic M

mf

A

f

R

[

[

[[[

[

[

[

[

[

[

[

[

[[

[[[[[

[

[[[

[

[[

G.S. Goncalves et al. / Synth

ation of a protective layer of oxides when using coatings obtainedrom paints containing electroactive polymers.

cknowledgements

The authors would like to thank to CNPq, FAPERGS and CAPESor the financial support of this research.

eferences

[1] N. Gospodinova, L. Terlemezyan, Progr. Polym. Sci. 23 (1998) 1443–1484.[2] A.M.P. Hussain, D. Saikia, F. Singh, D.K. Avasthi, A. Kumar, Nucl. Instrum. Meth-

ods Phys. Res. B: Beam Interact. Mater. Atoms 240 (2005) 834–841.[3] Y.-Z. Su, W. Dong, J.-H. Zhang, J.-H. Song, Y.-H. Zhang, K.-C. Gong, Polymer 48

(2007) 165–173.[4] K.M. Manesh, P. Santhosh, A.I. Gopalan, K.-P. Lee, Electroanalysis 18 (2006)

1564–1571.[5] G. Chen, Z.Y. Wang, D.G. Xia, L. Zhang, R. Hui, J.J. Zhang, Adv. Funct. Mater. 17

(2007) 1844–1848.[6] S. Lupu, A. Mucci, L. Pigani, R. Seeber, C. Zanardi, Electroanalysis 14 (2002)

519–525.[7] F.D.R. Amado, E. Gondran, M.A.S. Rodrigues, J.Z. Ferreira, C.A. Ferreira, J. Membr.

Sci. 234 (2004) 139–145.[8] F.D.R. Amado Jr., L.F. Rodrigues, M.A.S. Rodrigues, A.M. Bernardes, J.Z. Ferreira,

C.A. Ferreira, Desalination 186 (2005) 199–206.[9] F.D.R. Amado Jr., L.F. Rodrigues, M.M.C. Forte, C.A. Ferreira, Polym. Eng. Sci. 46

(2006) 1485–1489.10] M.A.S. Rodrigues, E. Gondran, C. Korzenowski, A.M. Bernardes, J.Z. Ferreira, J.

Membr. Sci. 279 (2006) 140–147.11] P.M. Carrasco, M. Cortazar, E. Ochoteco, E. Calahorra, J.A. Pomposo, Surf. Interf.

Anal. 39 (2007) 26–32.

12] H. Yoon, M. Chang, J. Jang, Adv. Funct. Mater. 17 (2007) 431–436.13] K. Maksymiuk, Electroanalysis 18 (2006) 1537–1551.14] C.A. Ferreira, P.C. Lacaze, S.J. Aeiyach, J. Aaron, Electrochim. Acta 41 (1996)

1801–1809.15] A. Meneguzzi, C.A. Ferreira, M.C. Pham, M. Delamar, P.C. Lacaze, Electrochim.

Acta 44 (1999) 2149–2156.

[

[

[

etals 161 (2011) 313–323 323

16] X.-H. Wang, J. Li, J.-Y. Zhang, Z.-C. Sun, L. Yu, X.-B. Jing, F.-S. Wang, Z.-J. Ye, Synth.Met. 102 (1999) 1377–1380.

17] A.M. Meneguzzi, C. Pham, J.C. Lacroix, C.A. Ferreira, J. Electrochem. Soc. 148(2001) B121–B126.

18] D.M. Lenz, M. Delamar, C.A. Ferreira, J. Electroanal. Chem. 540 (2003)35–44.

19] K. Fraoua, S. Aeiyach, J. Aubard, C.A. Ferreira, J. Adhes. Sci. Technol. 13 (4) (1999)517–522.

20] E. Armelin, R. Oliver, F. Liesa, J.I. Iribarren, F. Estrany, C. Alemán, Prog. Org. Coat.59 (2007) 46–52.

21] B.N. Grgur, M.M. Gvozdenovic, V.B. Miskovic-Stankovic, Z. Kacarevic-Popovic,Prog. Org. Coat. 56 (2006) 214–219.

22] J.I. Iribarren, E. Armelin, F. Liesa, J. Casanovas, C. Alemán, Mater. Corros. 57(2006) 683–688.

23] W.-C Chen, T.-C. Wen, A. Gopalan, Synth. Met. 128 (2002) 179–189.24] C.A. Ferreira, S. Aeiyach, A. Coulaud, P.C. Lacaze, J. Appl. Electrochem. 29 (1999)

259–263.25] B. Wessling, Adv. Mater. 6 (1994) 226–228.26] W.K. Lu, R.L. Elsenbaumer, B. Wessling, Synth. Met. 71 (1995) 2163–2166.27] T. Schauer, A. Joos, L. Dulog, C.D. Eisenbach, Prog. Org. Coat. 33 (1998) 20–27.28] M. Fahlman, S. Jasty, A.J. Epstein, Synth. Met. 85 (1997) 1323–1326.29] K. Fraoua, S. Aeyiach, J. Aubard, C.A. Ferreira, P.C. Lacaze, J. Adhes. Sci. Technol.

13 (4) (1999) 517.30] G.M. Spinks, A.J. Dominis, G.G. Wallace, D.E. Tallman, J. Solid State Electrochem.

6 (2002) 85–100.31] M. Rohwerder, A. Michalik, Electrochim. Acta 53 (2007) 1300–1313.32] T.P. McAndrew, A.G. Gilicinski, L.M. Robeson, US 5,441,772 (1995).33] Y. Wei, J. Wang, X. Jia, J.M. Yeh, P. Spellane, Polymer 36 (1995)

4535–4537.34] Y. Chen, X.H. Wang, J. Li, J.L. Lu, F.S. Wang, Corros. Sci. 49 (2007)

3052–3063.35] A.F. Baldissera, D.B. Freitas, C.A. Ferreira, Mater. Corros. 60 (2010) 790–801.36] B. Müller, U. Poth, Coatings Formulation: An International Textbook, Vincentz,

Hannover, 2006, p. 290.

37] W.S. Araujo, I.C.P. Margarit, M. Ferreira, O.R. Mattos, Electrochem. Acta 46

(2001) 1307–1312.38] Infrared Spectroscopy Atlas Working Committee, An Infrared Spectroscopy

Atlas for the Coatings Industry, vol. 1, 4th ed., Federation of Societies for Coat-ings Technology, Pennsylvania, 1991, p. 510.

39] J. Wessling, Posdorfer, Electrochem. Acta 44 (1999) 2139–2147.