Utilizing the structural memory effect of layered double hydroxides for sensing water uptake in organic coatings

Utilizing the structural memory effect of layered double hydroxides for sensing water uptake in organic coatings

Progress in Organic Coatings 51 (2004) 91–102 Utilizing the structural memory effect of layered double hydroxides for sensing water uptake in organic...

511KB Sizes 1 Downloads 23 Views

Progress in Organic Coatings 51 (2004) 91–102

Utilizing the structural memory effect of layered double hydroxides for sensing water uptake in organic coatings F. Wong, R.G. Buchheit∗ Fontana Corrosion Center, Ohio State University, Columbus, OH 43210, USA Received 29 March 2004; received in revised form 29 June 2004; accepted 6 July 2004

Abstract In this paper, we report results demonstrating the structural memory effect of synthetic calcined layered double hydroxide (LDH) Li2 [Al2 (OH)6 ]2 CO3 ·nH2 O powder immersed in a bulk electrolyte, exposed to humid air, and embedded in a water-permeable epoxy matrix. Reconstruction of calcined LDH by the structural memory effect can be detected by X-ray diffraction (XRD) leading to a novel approach for remotely and non-destructively detecting water uptake in optically opaque organic coatings. The LDH Li2 [Al2 (OH)6 ]2 CO3 ·nH2 O was synthesized by aqueous co-precipitation then calcined in air at temperatures in excess of 220 ◦ C to form a Li–Al mixed hydrated oxide powder. Reconstruction in a matter of days was observed when the calcined mixed oxide was immersed in 0.5M NaCl solution. During exposure to humid air, LDH reconstruction was slower occurring over a matter of weeks, perhaps in a deliquescent electrolyte. Paint-like coatings were made and applied to aluminum alloy 2024-T3 (Al–4.4Cu–1.5Mg–0.6Mn) substrates by adding the calcined LDH at a rate of 10 wt.% to a commercial epoxy. Coated substrates were then exposed to 0.5M NaCl solution and LDH reconstruction progressed over tens of days as the coating absorbed water. During these exposure experiments, XRD and electrochemical impedance spectroscopy measurements were made periodically to track LDH reconstruction and measure uptake of water in the coating via capacitance measurements. LDH reconstruction was tracked using the ratio of the {0 0 3} LDH diffraction peak to the {1 1 1} Al diffraction peak. Using the Brasher–Kingsbury equation, the volume fraction of water in the coating was estimated from capacitance data. Up to the point of apparent coating saturation (about 10 vol. %), the XRD peak height ratio varied linearly with the estimated coating water content. This result suggests that additions of calcined LDH to organic coating may lead to methods for sensing early-stage coating degradation due to water uptake and may give an advance warning of substrate corrosion. © 2004 Published by Elsevier B.V.

1. Introduction Synthetic layered double hydroxide (LDH) compounds are important industrial materials used as absorbents, polymer stabilizers, alkaline catalysts, acid neutralizing agents, and precursors in the formation of spinel ceramics [1,2]. LDHs are manufactured on an industrial scale with a manufacturing capacity that increased by at least 20,000 metric tons in 1999 alone [3,4]. LDH compounds consist of positively charged layers of mixed metal hydroxides separated by negatively charged layers of anions and water. The prototypical LDH compound ∗

Corresponding author. Tel.: +1 614 292 6085; fax: +1 614 292 9857. E-mail address: [email protected] (R.G. Buchheit).

0300-9440/$ – see front matter © 2004 Published by Elsevier B.V. doi:10.1016/j.porgcoat.2004.07.001

is the naturally occurring hydromagnesite, whose formula is Mg6 Al2 (OH)16 ·CO3 ·4H2 O [5,6]. Structurally, the compound consists of Mg(OH)2 layers, which carry a net positive charge due to the periodic substitution of Al3+ on Mg2+ sites. The positive charge on the Mg(OH)2 layers is offset by net negative charge on interleaving layers consisting of carbonate ions and water molecules. The artificial LDH used in this study, Li2 [Al2 (OH)6 ]2 CO3 ·nH2 O, is slightly different in that the positive charge on the metal hydroxide layer arises due to the presence of Li+ cations situated in normally unoccupied octahedral sites in an Al(OH)3 layer (Fig. 1). LDH compounds are usually synthesized as powders by titration of one metal salt solution with another to induce precipitation [7,8,9]. LDH coatings may be formed by sol–gel methods using metal alkoxides [10] and by reaction of a metal

92

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

Fig. 1. A schematic illustration of the structure of layered double hydroxides. Metal hydroxide layers are located on the top and bottom layers while anion layer is located in the middle.

substrate with an alkaline metal salt solution containing suitable anions [11]. Differential scanning calorimetry and X-ray diffraction (XRD) measurements show that thermal treatment of LDHs up to about 200 ◦ C induces dehydration. Treatment up to 500 ◦ C induces dehydroxylation and loss of vaporizable anions such as carbonate, nitrate, oxalate, and acetate. Heating at up to 900 ◦ C leads to the formation of mixed metal oxides, and in some cases, spinels. These calcined LDHs possess a “structural memory effect” [9,12,13,14,15,16]. Remarkably, the partially or fully dehydrated mixed metal oxides produced from LDH precursors will rehydrate on contact with water, take up anions, and reform the original LDH structure with great fidelity. A schematic illustration of this transformation is depicted in Fig. 2. All of these structural transformations are readily detected by XRD. We are interested in exploiting this structural memory effect to sense early stage coating degradation remotely and non-destructive by XRD. In this scheme, LDH is added to an organic resin and applied to a metal substrate using conventional application methods like spraying or brushing. In the presence of condensed aggressive electrolytes, degradation of organic coatings that leads to corrosion dam-

Fig. 2. Illustration describing the structural state transformation upon calcination and rehydration of layered double hydroxide compound showing its structural memory effect.

age of an underlying metallic substrate occurs in several welldefined stages [17]. First stage involves uptake of water and ionic species dissolved in the electrolyte into the polymer coating. The next stage involves concurrent interfacial separation and formation of a bulk of electrolytes at the coating/metal interface. Once an aggressive electrolyte has been formed at the interface, corrosion of the substrate ensues. In terms of sensing coating degradation, detecting uptake of water into the coating is important since at this stage, the sequence of events leading to coating degradation has begun but significant interfacial separation and damage to the substrate have not yet occurred [18]. Given the foregoing, the objective of this study is to demonstrate the structural memory effect in the calcined LDH, Li2 [Al2 (OH)6 ]2 CO3 ·nH2 O, when exposed to aqueous solutions, humid air and when embedded in an epoxy matrix, and to develop the basis for remotely and non-destructively detecting the earliest stages of organic coating degradation due to water uptake.

2. Experimental procedures 2.1. LDH preparation Li2 [Al2 (OH)6 ]2 CO3 ·nH2 O LDH powder was synthesized by co-precipitation at room temperature in aqueous solutions [19]. Two solutions were prepared. In the first solution, 20.68g of 0.3 M LiNO3 and 37.51g of 0.1 M Al(NO3 )3 ·9H2 O were added to 200 ml of deionized water. The second solution consisted of 6.0 g of 0.15 M NaOH and 21.2 g of 0.2 M Na2 CO3 also added to 200 ml deionized water. These solutions were added at the same rate (50 ml/h) to a beaker containing 100 ml deionized water with continuous stirring resulting in the formation of a white precipitate. The resulting mixture was maintained at pH 10 by dropwise addition of 1M NaOH. After the addition was completed, the resulting precipitate was aged at 70 ◦ C for 24 h under continuous gentle stirring. The precipitate was then filtered and washed thoroughly with deionized water several times. The filtered

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

93

Fig. 3. Experimental setup for exposure of calcined LDH to a humid air chamber with humidity in the range of 85–90% RH.

precipitate was dried in an air oven at 70 ◦ C overnight. The LDH structure was verified by powder XRD. The LDH powder was then calcined at 220, 820 and 1020 ◦ C for 2 h. To demonstrate the powder rehydration process, the 220 ◦ C calcined LDHs were exposed to immersion in 0.5M NaCl and a humid chamber (Fig. 3). 2.2. Coating preparation and application For coating studies, LDH powder, calcined at 220 ◦ C, was dispersed into epoxy resin. In these experiments, 0.59 g of calcined LDH was thoroughly mixed with 3.75 g of bisphenol epoxy resin, which was the unpigmented version of a primer (code number 02Y040 by DEFT chemical coating, Irvine, CA). Immediately prior to coating application, 1.56 g of an amide-based hardener (catalyst for the unpigmented primer with code number 02Y040CAT, DEFT chemical coating, Irvine, CA) was added to the mixture. This resulted in a pigment loading level of about 10 wt.%. This mixture was applied on 100 mm × 200 mm × 2 mm aluminum alloy 2024-T3 sheets using a number 9 rolling bar. Prior to coating, the surfaces were degreased in an alkaline sodium carbonate/sodium silicate solution and deoxidized in a nitric acid-based solution. The coated substrate was allowed to dry overnight. This procedure was repeated five times resulting in a six-ply coating that was 126 ␮m in thickness. Coated samples were then cut into two pieces. One piece was used for coating capacitance measurements used to quantify water uptake, the other was used in XRD experiments to characterize structural transformations in the LDH additive. EIS cell with 32 cm2 working electrode area was immersed in 0.5 M NaCl while

the other was simply immersed in a beaker containing 0.5 M NaCl. 2.3. X-ray diffraction (XRD) XRD of powders and coatings was carried out using a Scintag XDS 2000 X-ray diffraction measurement system operated in theta-theta vertical goniometer mode. The instrument had 2 kW sealed tube CuK␣ X-ray source. To fully resolve diffraction patterns, scans were collected from 2 to 70◦ 2-theta at a scan rate of 2◦ /min. 2.4. Electrochemical impedance spectroscopy (EIS) Coating capacitances were measured to determine water uptake by the epoxy coatings. The details of this analysis are presented in the Section 3. Capacitances were extracted from EIS spectra collected from coatings immersed in an aerated 0.5M NaCl solution. Measurements were made using a three-electrode cell configuration in which the coated sample was the working electrode, platinum mesh was used as a counter electrode and a saturated calomel electrode was used as the reference electrode. Measurements were made using a Princeton Applied Research Model 273A potentiostat and a Solartron SI 1255 HF Frequency Response Analyzer. Spectra were collected over frequencies ranging from 5 mHz to 65 kHz using a 10 mV sinusodial voltage perturbation. Spectra were sampled at a rate of seven points per decade frequency. The experiment was carried out, and data were logged under computer software control using Scribner Associates ZplotTM EIS measurement software. EIS data

94

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

were analyzed and fitted using Scribner Associates’ ZviewTM software.

3. Results And discussion 3.1. Structural transformation in LDH powder The synthesis methods used in this study to produce the Li–Al–LDH yielded a compound whose XRD pattern was consistent with the one reported by Serna [20] and Poeppelmeier [8]. Serna and Poeppelmeier give the LDH composition as [LiAl2 (OH)6 ]+ ·X− ·nH2 O, where X− is one of several simple inorganic anions that can be accommodated in the anion interlayer without a change in the LDH crystal structure. Characterization of the compound by electron ˚ c= diffraction indicates a hexagonal unit cell (a = 5.32 A, ˚ and the presence of a superlattice in a gibbsite-like 22.7 A) metal hydroxide layer indicating cation ordering among octahedral sites. The stoichiometry of the compound reflects the fact that these sites are filled at a rate of 2:1 by Al3+ and Li+ , respectively. In gibbsite, the sites occupied by Li+ are normally vacant, hence the metal hydroxide layers carry a net positive charge that must be offset by the companion anion interlayer. When the interlayer anion is predominantly ˚ CO3 2− , Serna reports an {0 0 3} basal plane spacing of 7.6 A [20]. In general, calcination of LDH compounds at temperatures up to about 200 ◦ C results in the loss of structural water. Calcination at higher temperatures leads to volatilization of vaporizable interlayer anions such as carbonate, which forms CO2 gas, collapse of the layered structure, and the formation of a poorly crystalline mixed oxide. Calcination at temperatures above 800 ◦ C can lead to the formation of

spinels [19]. These transformations are readily detected by XRD. The Li–Al LDH examined in this study behaves in a manner consistent with the general behavior of LDH compounds during calcination. Fig. 4 shows XRD patterns of Li–Al LDH calcined at different temperatures ranging from 220 to 1020 ◦ C. Calcining the LDH powder at temperatures between 220 and 820 ◦ C, results in the formation of the mixed oxide phase indicating volatilization of carbonate and layer collapse. At 820 and 1020 ◦ C, strong sharp reflections emerge in the pattern consistent with the formation of ␥-LiAlO2 and LiAl5 O8 [8]. The structural memory effect is triggered by contacting the calcined LDH with water. Fig. 5 shows an example of the structural memory effect for Li–Al LDH calcined at 220 ◦ C as revealed by XRD. In this experiment, a calcined LDH powder sample was fully immersed in aerated 0.5M NaCl solution with stirring. Fig. 5a shows the diffraction pattern of the parent Li–Al LDH in the as-synthesized condition. Fig. 5b shows the diffraction pattern of the poorly crystalline mixed oxide formed by calcination. This pattern is characterized by at least three diffuse, but widely separated remnant diffraction peaks caused by dehydration and dehyroxylation [8]. After immersion in solution for 1 day, peaks characteristic of the calcined mixed oxide are still strong. Some weak parent LDH peaks are also present in the pattern (Fig. 5c). After 5 days immersion, the diffraction pattern of the parent LDH is almost fully recovered as illustrated in Fig. 5d. The reconstructed LDH patterns, both peak position and relative intensities, are well preserved compared to the parent LDH indicating good structural fidelity. There is a very long tail in the {0 0 3} reflection at about 12◦ 2␪ which is due to diffraction from the mixed metal hydroxide layers in the compound. Several factors may contribute to peak-broadening

Fig. 4. XRD patterns of a Li2 [Al2 (OH)6 ]2 CO3 ·3H2 O LDH powder at the as-synthesized (parent/original) form and at the calcined forms. Parent LDH transformed into its calcined forms after 2 h of calcination at a specific temperature indicated on the right side of the plot.

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

95

Fig. 5. XRD patterns showing structural transformation of the 220 ◦ C-calcined Li–Al–CO3 LDH during rehydration in 0.5M NaCl.

including layer disorder, incorporation of several different anions leading to different interlayer spacings or a reduction in LDH crystallite size [21]. In any case, the LDH pattern is sufficiently well reproduced that there is no difficulty in distinguishing the pattern of the reconstructed LDH from the pattern of the mixed oxide, and there is no doubt that the structural memory effect has been triggered. As has been reported in earlier studies [1], LDH reconstruction was observed to be slower for LDHs calcined at higher temperatures. LDH calcined at 820 ◦ C showed a mixed spinel and LDH pattern after as early as 7 up to 20 days of immersion in 0.5M NaCl (Fig. 6). In the context of sensing water uptake in coatings, the ability to mediate LDH reconstruction by calcination temperature may represent a simple and attractive approach to tailoring sensitivity to water uptake. Since main thrust of these experiments was to explore calcined LDH reconstruction. The majority of the experiments were carried out with LDH samples calcined at 220 ◦ C as these samples exhibited the most rapid reconstruction kinetics. LDH reconstruction is also triggered by exposure to humid air, although the reconstruction kinetics are considerably slower than that observed during exposure to aqueous solutions. Fig. 7 shows the reconstruction steps of the 220 ◦ C calcined LDH exposed to a humid chamber for 30 days. A comparative assessment of the reconstruction kinetics in air

and solution was made by plotting the peak height of the basal plane {0 0 3} LDH reflection as a function of exposure time in humid air or in solution. These data are shown in Fig. 8. The calcined LDH immersed in 0.5 M NaCl needs 5 days to fully reconstruct while more than 30 days are needed for an equivalent amount of reconstruction during exposure to humid air. 3.2. Reconstruction of calcined LDH in epoxy coatings due to water uptake Fig. 9 shows diffraction patterns from calcined LDHbearing epoxy coatings on Al alloy substrates after various lengths of exposure to 0.5M NaCl solutions. Initially, sharp diffraction peaks from the Al alloy substrate and a broad, diffuse peak from scattering by the epoxy are evident in the pattern. The three diffuse peaks characteristic of the calcined LDH are either masked by the broad epoxy scattering peak and the Al diffraction peaks or are of insufficient intensity on this scale to be distinguished from the background. After 1 day of exposure, a weak reflection at about 12◦ is evident in the pattern. This is the expected position for the most intense LDH reflection associated with diffraction from the {0 0 3} basal plane. With increasing exposure time, reflections from 20–24◦ and at 36◦ positions develop in the pattern as the peak

96

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

Fig. 6. XRD patterns showing rehydration of the 820 ◦ C-calcined Li–Al–CO3 LDH upon immersion in 0.5M NaCl. Mixed LDH and ␥-LiAlO2 and LiAl5 O8 pattern was observed up to 20 days of immersion.

height at 12◦ increases. These are all locations of intense reflections in the LDH pattern, and the development of the pattern with increased exposure time suggests reconstruction of the LDH additive due to water taken up by the epoxy coating. This result is the basis for sensing coating water uptake. To track the progress of water uptake, the ratio of the {0 0 3} LDH to the {1 1 1} Al peak height is used as a convenient internal standard [22]. This ratio is then measured as a function of exposure time. Implicit assumptions in this approach are that the {1 1 1} Al peak height does not vary significantly with exposure time, that the LDH formation kinetics are fast compared to coating water uptake kinetics and that the {0 0 3} reflection height scales with volume fraction of LDH [23]. Prior experiments show that Li–Al LDH begins to reconstruct in less than 1 day, which is rapid enough to track water uptake in epoxy, which occurs over the span of days and weeks. The variation in the {0 0 3} LDH/{1 1 1} Al peak height ratio (PHR) with exposure time is shown in Fig. 10. This PHR plot demonstrates three distinct stages. A steady increase in the PHR is observed in the first 8 days, followed by a plateau and a further increase after 53 days. Water uptake in organic coatings tracked by gravimetry follows a similar trend and

is interpreted as indicating diffusion-controlled water uptake [18]. 3.3. Equivalent circuit modeling and impedance spectra Electrochemical impedance spectroscopy (EIS) is sensitive to the earliest stages of coating degradation including water uptake. EIS was used here to gather coating capacitance data for estimating water uptake and to examine for the onset of coating delamination or substrate corrosion. EIS spectra collected from degrading coated metal substrates were fit to physically relevant equivalent circuit models by complex non-linear least squares (CNLS) fitting routine within ZviewTM [24]. The fitting routine returned numerical values for all the circuit elements present in the equivalent circuit model. Fig. 11a and b are examples of typical impedance spectra (Bode magnitude plots) with the corresponding physical and equivalent circuit models. The ECM on Fig. 11a, which is similar to a simplified Randle’s circuit, consists of a solution resistance (Rs ), a coating constant phase element (CPE), and a coating pore resistance (Rpore ). After prolonged exposure, the adhesion between coating and substrate weakens creating localized separations

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

97

Fig. 9. XRD patterns illustrating structural transformation of the calcined LDH-pigment in epoxy coating upon immersion in 0.5 M NaCl. The emergence of {0 0 3}, {0 0 6}, and {0 0 9} LDH reflections with time indicates rehydration of the calcined LDH-pigment. Numbers on the right side of the plot are time of immersion in days.

Fig. 7. XRD patterns showing slow rehydration of the 220 ◦ C-calcined Li–Al–CO3 LDH upon exposure in a humid chamber. The calcined LDH structure needed at least 30 days of exposure to get a distinguishable rehydrated LDH structure.

at the coating/substrate interface, which alters the coating and impedance response. Fig. 11b shows an ECM for the latter case that includes constant phase element for the double layer at the bottom of the defects (CPEdl ), and a polarization resistance of the substrate (Rp ), which is exposed at the base of the defect. These two ECMs were used to analyze the resulting EIS spectra in this study. Due to the non-ideal capacitive response, CPEs were used in ECM rather than perfect capacitors [24]. The following equation was used to extract the true capacitance, C from the

Fig. 8. A plot showing variation in rehydration kinetics as indicated by the difference in peak height slopes for two different types of exposures. The value for peak height is taken from the LDH reflection from the {0 0 3} plane at 12◦ .

Fig. 10. Rehydration of calcined LDH-pigment during immersion as shown by increasing peak height ratio. The peak height ratio is calculated by normalizing the intensity of {0 0 3} LDH reflection at 12◦ to intensity of {1 1 1} aluminum reflection at 38.5◦ .

98

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

Bode plots of EIS data collected from the coated samples as a function of exposure time are shown in Fig. 13. The portions of these curves with slopes near −1 in a Bode magnitude plot indicate capacitance-dominated responses while plateaus indicate resistance-dominated responses. After 53 days of exposure in NaCl solution, emergence of additional capacitive and resistive responses is observed at intermediate frequency region (200–1000 Hz). The appearance of the new responses after 53 days was taken as an indication of the onset of a new and enhanced episode of coating degradation perhaps involving coating/metal separation. The high frequency capacitance-dominated response in the Bode magnitude plot is due to the coating capacitance [24]. The resistance-dominated response observed at intermediate frequency region is still attributed to solution and pore resistances. These regions of the spectrum were fit to extract numerical value for the CPE and pore resistance for the coating. Impedance data at the low frequency region were not fully resolved and not included in the analysis. For short immersion times, the data were fitted to the simplified R–CPE circuit model (Fig. 11a) while for longer immersion times, the data were fitted to a more complex circuit model shown in Fig. 11b. 3.4. Estimation of coating water uptake

Fig. 11. (a) A physical model, a simplified Randles circuit model, and a typical Bode magnitude plot that represent an intact coating [24]. (b) Loss of adhesion between coating and metal substrate causes the emergence of the second RC constant at the complex equivalent circuit model and the Bode magnitude plot [24].

CPE parameters returned from the fitting routine [25]: C = Y 1/n R(1−n)/n pore

(1)

In this expression, Y represents the CPE constant, and n is the CPE exponent and Rpore is the pore resistance that is in parallel with the CPE as in Fig. 11a. Fig. 12 shows complex plane plots of impedance spectra collected as a function of exposure time in 0.5M NaCl for an Al alloy substrate coated with calcined LDH-bearing epoxy. As exposure time increases, the higher frequency semi-circle becomes smaller. Since solution resistance is relatively constant, the progressive decrease in the size of the semi-circles indicates a decrease in the pore resistance. Decreasing pore resistance signifies degradation of the coating because less protection is provided. At low frequencies, incompletely resolved semi-circles are present. These are not completely resolved due to the lower frequency limit imposed during the measurements. Nonetheless, the lower frequency response is characteristic of a diffusional impedance response or an additional RC response due to substrate corrosion [26].

True coating capacitance values were calculated using Eq. (1), and were plotted versus time in Fig. 14. This figure shows three distinct stages of behavior, much like that demonstrated in the PHR versus time plot of Fig. 10. To estimate the amount of water content in the coating at each measurement point, the Brasher–Kingsbury equation was applied to the capacitance data. This equation provides a simple mathematical relationship between coating capacitance, C, and volume fraction of water in an organic coating, Xv [27]: Xv =

100 × log(C/C0 ) log(80)

(2)

In this expression, the constant 80 is the relative permittivity of the absorbed water as assumed by Brasher and Kingsbury in their original analysis. Parameter C0 is the dry coating capacitance, which is the capacitance value at immersion time t = 0. The C0 was obtained by a linear extrapolation of C versus t data to zero immersion time [28]. Fig. 15 shows the linear extrapolation for this study, which leads to a C0 value of 2.41 × 10−9 F/cm2 . Coating water content versus exposure time is shown in Fig. 16. First-stage behavior occurs over the initial 8 days of exposure. This is characterized by rapid, homogeneous, diffusional water uptake by the coating. Rapid water uptake at the beginning of the exposure period is expected considering the rapid increase in the PHR at short exposure times shown in Fig. 10. Second stage behavior is characterized by a period of slow or no water uptake. This type of response has been interpreted as indicating coating saturation [29]. The coating appears to saturate at about 15 vol.% water, and there is little

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

99

Fig. 12. Complex plane plots from EIS measurements of the AA2024-T3 substrate coated with calcined LDH-pigment epoxy coating during immersion in 0.5 M NaCl. The exposed working electrode area was 32 cm2 (inset: complex plane plots at larger scale).

change in the capacitance (or the PHR, Fig. 10) from 8 to 53 days exposure. In third-stage behavior, bulk water is believed to condense locally, perhaps at the coating metal interface. Strictly speaking, the Brasher–Kingsbury equation can no longer be used to accurately estimate water content under these conditions [29], however, data points were plotted in Fig. 16 to distinguish third-stage behavior on the plot. The end of saturation period may also indicate the start of coating breakdown due to condensation of bulk water in the coating [28,29], or the onset of coating/metal separation. In Fig. 10, the PHR data suggest that during the third stage of behavior, the amount of rehydrated LDH increases significantly. This response mimics closely the water uptake behavior shown in Fig. 16. The notion that the formation of bulk water in the coating or accumulation of water at the coating/metal interface may lead to increased rehydration of the pigment seems reasonable, but how this might occur is presently unknown.

In terms of corrosion sensing, it is noted that the emergence of a new R–C time constant in the EIS data and the onset of water condensation are coincident at about 53 days exposure. No direct visual evidence of substrate corrosion was noted on these surfaces, but it is reasonable to conclude that the coating/metal interface had proceeded to an advanced state of degradation and corrosion had initiated or was imminent. 3.5. A model for coating water uptake based on LDH pigment rehydration Coating water uptake and calcined LDH rehydration demonstrate similar three-stage time dependencies during exposure to aqueous solutions. This suggests that a linear relationship exists between the coating water content, Xv and the fraction of LDH pigment rehydrated as expressed by the PHR. A plot of Xv versus PHR is shown in Fig. 17. Linear

100

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

Fig. 15. Extrapolation of coating capacitance to coating capacitance, C0 , of 2.41 × 10−9 F/cm2 .

√ t → 0 yields to a dry

regression of the data set yields the following model: Xv = −4.96 + 213 PHR with a linear correlation coefficient R of 0.904. In the data set used for this regression analysis, two extremely outlying data points and data points beyond the saturation plateau were excluded (not shown on the plot). The high correlation coefficient value indicates a linear dependence of Xv on the PHR and suggests that tracking rehydration of the calcined LDH pigment by XRD can use used to estimate coating water uptake.

To utilize calcined LDH as a true corrosion sensing pigment, a relation between water content in the coating and initiation of substrate corrosion is needed. In this study, substrate corrosion is assumed to commence when the coating water content has exceeded saturation. A specific peak height ratio that corresponds to that particular value of volume fraction of water can be linked to the beginning of substrate corrosion. By constructing regression models, water uptake through to water saturation can be characterized. However, further study is needed to confirm that the onset of corrosion is strongly linked to water supersaturation. To check for the reproducibility of the relationship between Xv and PHR, the experiments and analyses described here were carried out on two additional replicate coatings subject to identical exposures in 0.5 M NaCl solution. Three stage time-dependent behavior of Xv and PHR, as well as the linear

Fig. 14. An increasing coating capacitance of the coated substrate during immersion in 0.5 M NaCl follows the 3-stage pattern. Values were obtained from curve fitting of the Bode magnitude plots to the corresponding ECM.

Fig. 16. A plot showing the 3-stage variation in coating water content during immersion of the coated sample in 0.5 M NaCl. Coating water content was computed using Brasher–Kingsbury equation.

Fig. 13. Bode magnitude (above) and phase angle (below) plots from EIS measurements of the AA2024-T3 coated substrate upon immersion in 0.5M NaCl. The exposed working electrode area was 32 cm2 . Numbers indicate time of immersion in days.

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

101

if swelling occurred, it did not result in significant coating degradation. 4. Conclusions

Fig. 17. Linear correlation between coating water content variation from water uptake and {0 0 3}LDH/{1 1 1}Al peak height ratio variation from calcined LDH-pigment rehydration. Data points beyond saturation period and other outliers were excluded from the fit.

relation between Xv and PHR were demonstrated [30]. Additional experiments illustrating the utility of the technique for sensing water uptake under salt spray exposure conditions and conditions of alternate immersion are also reported in reference [30]. 3.6. Practical considerations The utility of LDH pigments for sensing water uptake in coating systems is dependent on several factors. These pigments would be most useful in coating systems designed to resist uptake of an electrolyte. The pigments must anticipate substrate corrosion. Mechanical damage to the coating excepted; coating electrolyte uptake must occur before substrate corrosion, thus reconstruction of the LDH pigment is well tuned as an early indicator of possible substrate damage. Immersion–emersion experiments have shown that LDH reconstruction is not reversible [30]. This has consequences for intermittent wet-dry exposures. Under these conditions, the LDH pigment would reconstruct during wet periods, but not during dry periods. When the pigment was interrogated, the extent of the reconstruction process would reflect cumulative water exposure. On the basis of experiments conducted so far, the impact of the LDH pigment on coating corrosion protection and adhesion appears minimal. The coatings used in this study suffered only minor amounts of corrosion damage during exposure and coating remained adherent to the substrate. The amount of damage observed indicates that the presence of the pigment does not lead to enhanced substrate corrosion or dramatic loss in adhesion. It is easy to imagine that the LDH might swell upon water uptake. However, the results or observations collected in this study indicated that

Li2 [Al2 (OH)6 ]2 CO3 ·nH2 O LDH demonstrates a structural memory effect through calcination and rehydration by taking up water from environments. The calcined LDH powder with a poorly crystalline mixed oxide solid solution exhibits faster rehydration kinetics than more crystalline forms such as ␥-LiAlO2 and LiAl5 O8 . Studies show that exposure of the calcined LDH to solution induces rehydration faster than exposure to humid air. Reconstruction during humid air exposure is interpreted as solution phase reconstruction in electrolytes formed by capillary condensation or deliquescence. Overall, the calcined LDH exhibits better structural fidelity in reconstruction during solution exposure. When calcined LDH powder is dispersed in an organic resin and applied to an aluminum alloy substrate as a coating, the rehydration transformation is retained and can be exploited as a form of sensing during exposure to aqueous environments. The rehydration transformation of calcined LDH can be tracked remotely by XRD and is indicated by an increase in the intensity of the diffraction pattern of the hydrated form of the LDH compound. The extent of the rehydration transformation can be tracked using the peak height ratio (PHR) of the {0 0 3} LDH diffraction peak to a reference diffraction peak from the substrate. For the purposes of relating the PHR to coating water content, the volume fraction of water, Xv , in the coating was estimated from capacitance data using Brasher–Kingsbury equation. Both PHR and Xv increases exhibit the three-stage “sigmoidal” curve shape, and regression analysis shows that Xv is linearly related to PHR. With this model, XRD can be used to assess water uptake by the coating. Evidence suggests that substrate damage or coating/metal separation occurs at the end of the coating saturation period, which is detectable using XRD methods described here. Acknowledgments This study is funded by Air Force Office of Scientific Research under MURI contract no. F49602-01-1-0352. Appendix A The following is an example of the data-fitting procedure to extract the numerical values of CPE (Y), CPE (n), and Rpore . Coating capacitance C was calculated using Eq. (1). These figures are Bode magnitude and phase angle plots of AA2024-T3 with calcined Li–Al–LDH pigmented epoxy coating fitted to a simplified Randles circuit (Fig. A.1).

102

F. Wong, R.G. Buchheit / Progress in Organic Coatings 51 (2004) 91–102

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] Fig. A.1. Fitting of the Bode magnitude and phase angle plots of AA2024T3 coated with calcined Li–Al–LDH bearing epoxy using the simplified Randles circuit.

[21] [22]

Element

Freedom

Value

Error

Error (%)

CPE (Y) CPE (n)

Free (+) Free (+)

8.65E-8 0.781

3.34-9 0.00452

3.86 0.578

Rpore

Free (+)

7.81E5

11400

1.46

C = Y 1/n R(1−n)/n pore 1/0.781

= (8.65 × 10−8 ) = 4.06 × 10

−8

F.

(A.1) (1−0.781)/0.781

× (7.81 × 105 )

[23] [24] [25] [26]

[27] [28] [29] [30]

A.J. Marchi, C.R. Apesteguia, Appl. Clay Sci. 13 (1998) 35. V. Rives, M. Ulibarri, Coord. Chem. Rev. 181 (1999) 61. Anon, Chem. Eng. News 77 (1999) 12. Anon, Chem. Week 161 (1999) 27. S. Miyata, Clays Clay Miner. 23 (1975) 369. H.F.W. Taylor, Min. Mag. 39 (1973) 377. K. Itaya, H.-C. Chang, I. Uchida, Inorg. Chem. 26 (1987) 624. K.R. Poeppelmeier, S.J. Hwu, Inorg. Chem. 26 (1987) 3297. W.T. Reichle, J. Catal. 94 (1985) 547. T. Lopez, E. Ramos, P. Bosch, M. Asomoza, R. Gomez, Mater. Lett. 30 (1997) 279. R.G. Buchheit, M.D. Bode, G.E. Stoner, Corrosion 50 (1994) 205. E.T. Devyatkina, N.P. Otsupalo, N.P. Tomilov, A.S. Berger, Russ. J. Inorg. Chem. 28 (1983) 801. M.A. Drezdzon, Inorg. Chem. 27 (1988) 4628. S. Miyata, T. Hirose, Clays Clay Miner. 26 (1978) 441. A. Nemudryi, V.P. Isupov, N.P. Kotsupalo, Russ. J. Inorg. Chem. 28 (1983) 1575. M.A. Ulibarri, J.M. Fernandez, F.M. Lbajos, V. Rives, Chem. Mater. 3 (1991) 626. A.S. Castela, A.M. Simoes, Prog. Org. Coat. 46 (2003) 55–61. E.P.M.v. Westing, G.M. Ferrari, J.H.W. deWit, Corros. Sci. 36 (1994) 960. F. Millange, R.I. Walton, D. O’Hare, J. Mater. Chem. 10 (2000) 1713. C.J. Serna, J.L. Rendon, J.E. Iglesias, Clays Clay Miner. 30 (1982) 180. S. Velu, V. Ramkumar, A. Narayan, C.S. Swamy, J. Mater. Sci. 37 (1997) 957. R. Jenkins, R.L. Snyder, Introduction to X-ray Diffractometry, Wiley, New York, 1996, p. 371. B.D. Cullity, Elements of X-Ray Diffraction, 2nd ed., AddisonWesley, Boston, MA, 1978, pp. 406–407. J. Titz, G.H. Wagner, H. Spahn, M. Ebert, K. Juttner, W.J. Lorenz, Corrosion 46 (3) (1990) 221–229. G.J. Brug, A.L.G.v.D. Eeden, M. Sluyters-Rehbach, J.H. Sluyters, J. Electroanal. Chem. 176 (1984) 275. F. Deflorian, V.B. Miskovic-Stankovic, P.L. Bonora, L. Fedrizzi, Corrosion Science, NACE International, June 1994, pp. 441– 442. D.M. Brasher, A.H. Kingsbury, J. Appl. Chem. 4 (1954) 62. F. Bellucci, L. Nicodemo, Corrosion 49 (1993) 235–236. F. Deflorian, L. Fedrizzi, S. Rossi, P.L. Bonora, Electrochim. Acta 44 (1999) 4243–4244. F. Wong, Master’s Thesis, The Ohio State University, Columbus, OH, 2004.