Effect of physical ageing onto the water uptake in epoxy coatings

Effect of physical ageing onto the water uptake in epoxy coatings

Journal Pre-proof Effect of physical ageing onto the water uptake in epoxy coatings Yossra Elkebir, Stéphanie Mallarino, Dao Trinh, Sébastien Touzain...

882KB Sizes 1 Downloads 91 Views

Journal Pre-proof Effect of physical ageing onto the water uptake in epoxy coatings

Yossra Elkebir, Stéphanie Mallarino, Dao Trinh, Sébastien Touzain PII:

S0013-4686(20)30158-4

DOI:

https://doi.org/10.1016/j.electacta.2020.135766

Reference:

EA 135766

To appear in:

Electrochimica Acta

Received Date:

07 October 2019

Accepted Date:

23 January 2020

Please cite this article as: Yossra Elkebir, Stéphanie Mallarino, Dao Trinh, Sébastien Touzain, Effect of physical ageing onto the water uptake in epoxy coatings, Electrochimica Acta (2020), https://doi.org/10.1016/j.electacta.2020.135766

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier.

Journal Pre-proof Effect of physical ageing onto the water uptake in epoxy coatings Yossra ELKEBIR, Stéphanie MALLARINO, Dao TRINH, Sébastien TOUZAIN* Laboratoire des Sciences de l’Ingénieur pour l’Environnement, LaSIE UMR 7356 CNRS La Rochelle Université, Avenue Michel Crépeau, 17000 La Rochelle, France [email protected]

Abstract: This study aims to evaluate the effect of physical ageing onto the water uptake process in epoxy coatings during immersion in saline solution. DGEBA/Jeffamine230 stoichiometric polyepoxyde system was prepared as coatings onto aluminum Q-panels. Two different curing protocols were applied in order to obtain two fully cured polyepoxyde systems, with different amounts of physical ageing. DSC measurements allowed to measure the enthalpy associated to the physical ageing and the glass transition temperature. Coated samples were immersed in sodium chloride solution (30 g.L-1) at different temperatures (30°C, 40°C, 50°C and 60°C) and the water uptake was evaluated by electrochemical impedance spectroscopy (EIS), using the Brasher and Kingsbury equation. In parallel, the swelling of the coatings was in situ monitored using scanning electrochemical microscopy (SECM). At different solution temperatures, it was found that the water uptake values at equilibrium for coatings with physical ageing was lower than the values obtained with coatings without physical ageing. The swelling of coatings with physical ageing was also lower than the swelling of the coatings without physical ageing. A thermodynamic analysis of the diffusion parameters allowed to show that the densification of the polymer matrix induced by the physical ageing is the main factor that governs the water uptake process. This was explained by more interactions between polar groups of the polymer chains that decrease the number of polar groups for water diffusion.

Keywords Organic coatings; water uptake; physical ageing; EIS; swelling; SECM.

*

Corresponding Author: S. Touzain (E-mail: [email protected])

1

Journal Pre-proof I.

Introduction

The physical ageing of polymers may be encountered when the polymer network is below or close to the glass transition Tg, and it is a well-known phenomenon for bulk polymers [1, 2]. This affects the polymer microstructure and leads to a decrease of the free volume. Physical ageing can however be erased by heating the polymer above its glass transition temperature (Tg), so it is a reversible phenomenon. Physical ageing has then a significant impact onto the mechanical properties of the polymer [3-6] and this is clearly a concern for the industries that use epoxy resins, as physical ageing can compromise durability, reliability, and safety. Kong [6] studied water sorption in neat epoxy systems and composite and he showed that the diffusion rate and the water uptake at saturation was decreased in presence of physical ageing. On the other hand, Akele et al. [7] studied the effect of physical ageing and water sorption in polycarbonates and they concluded that physical ageing cannot modify significantly the rate of water diffusion in the polymer. Recently, Le Guen-Geffroy and al. [8] studied the coupling between physical ageing and plasticization for bulk epoxies hygrothermally aged in seawater and they showed that physical ageing was much faster in a wet environment than in dry conditions. The physical ageing of organic coatings has paid no such attention even since the important work of Perera [9]. In this study, the author claimed that physical ageing has an influence on the formulation, application and the service life behavior of organic coatings but the effect onto water uptake process was not evaluated. Shi and al. [10] studied the effect of physical ageing onto the degradation of organic coatings when submitted to accelerated weathering tests but without water. Olivier et al. [11] established a correlation between the loss of the barrier properties and the increase of the stress generated during hygrothermal ageing cycles of electrocoated paints. However, to date, it seems that no direct link was demonstrated between physical ageing and the kinetic of water uptake in organic coatings. In this work, the water uptake process of an epoxy resin system with different amounts of physical ageing is monitored. A model epoxy system DGEBA/Jeffamine 230 was used without pigment of fillers in order to obtain the response of the sole polymer. The different amounts of physical ageing were obtained thanks to controlled curing protocols. The water uptake was measured by electrochemical impedance spectroscopy (EIS) during immersion. However, our previous [12] work showed that the organic coatings swell during immersion so it was necessary to measure swelling by SECM in order to modify water sorption curves and finally, evaluate the kinetic of water diffusion. 2

Journal Pre-proof II.

Experimental

1. Materials The epoxy resin was prepared from diglycidylether of bisphenol A (DGEBA, Sigma Aldrich, ref. 1675-54-3) cured with Jeffamine 230 (Sigma Aldrich, ref. 9046-10-0). All materials were used as received without further purification. A stoichiometric amount of DGEBA was added to the amine hardener and mixed at room temperature. The mixture was deposited onto aluminum Q-panels and coated panels were inserted in a mould constituted by two aluminum plates covered by a Teflon sheet and separated by thick spacers. The dry film thicknesses were about 120 ± 10 μm to 240 ± 15 μm for all coatings (measured by an Elcometer 311 Gauge Thickness), depending on the mixture quantity that was poured onto the Q-panels. 2. Curing/cooling protocol The curing protocol was constituted by different temperature plateaus, from ambient temperature to 120°C, as shown in Fig.1a. The final temperature, which is above the glass transition temperature Tg, was maintained for seven hours to allow a full cross-linking of the coatings. Then, the heating was switch off and two different cooling conditions were used. The first condition was to let the coated panels into the oven, waiting a natural decrease of the temperature. This allows the temperature to be slowly decreased and to maintain the coated panel at a temperature just below Tg for a long time, allowing the physical ageing. The second condition was to remove the mould containing the coated

Temperature (°C)

7h

a)

120

b)

Aluminum mould fan

100 1h

80 3h

60 1h

40 20 0

2

4

6 8 10 12 14 16 Time (h)

Figure 1: a) Curing protocol and b) cooling protocol to avoid physical ageing. panels from the oven after the 7h plateau and to insert the mould between two fans for a rapid

3

Journal Pre-proof cooling (Fig. 1b). The temperature is then quickly decreased well above Tg which allow to avoid physical ageing of the organic coating. The properties of coatings with or without physical ageing were then compared to those of free films, with or without physical ageing, obtained between two Teflon sheets and cured with the same procedure. FTIR, DSC and DMA experiments were realized (results not shown) and did not show any differences in initial properties. DMA experiments allowed to investigate the time/temperature relation thanks to the WLF phenomenology and the C1 parameter was used to calculate the free volume fraction [13]. 3. Experimental methods a. Differential Scanning Calorimetry (DSC) Glass transition temperature (Tg) measurements were performed by Differential Scanning Calorimetry (DSC) with a TA Instruments Q100. The epoxy resins were scanned from 20 to 130 °C at 10 °C.min-1 under nitrogen flow. The glass transition temperature Tg is taken at the half height of the change in heat capacity (middle of transition). Modulated Differential Scanning Calorimetry was also used to evaluate the physical ageing. A temperature modulation of ±0.32°C was superimposed onto the temperature scan (2°C/min) from 20°C to 140°C and from 140°C to 20°C. b. Electrochemical Impedance Spectroscopy (EIS) The immersion of coated aluminum Q-panels was realized with an O-ring seal impedance glass cell (Gamry) filled with the aqueous solution (NaCl 3 wt. %). The contact surface was about 16 cm2 and the immersion tests were performed at 30, 40, 50 and 60 °C. The water uptake was followed in situ with a two electrodes cell using a graphite counter electrode and the aluminum substrate as the working electrode. EIS measurements were performed with a Gamry REF 600 at the free corrosion potential using a 30 mV r.m.s perturbation (11 pts/decade). Each measurement is repeated at least twice to verify the repeatability and the accuracy of the method. From these measurements, the film capacitance (CHF) was determined using the real Re(Z) and the imaginary Im(Z) parts of the impedance at high frequency (f = 10 kHz) using [14]: ―𝐼𝑚(𝑍)

𝐶𝐻𝐹 = 2𝜋𝑓(𝑅𝑒(𝑍)2 + 𝐼𝑚(𝑍)2)

(1)

The high frequency has been chosen as 10 kHz since theoretically, in this high frequency domain, the total impedance of the system is governed by the coating capacitance [15, 16]. Then, the water uptake was calculated with the Brasher and Kingsbury [17] equation, since it is the most used model for water uptake estimation and has been proved to be more successful 4

Journal Pre-proof than more complex approaches [18-20]: 100.log

𝜒𝑉 =

(

𝐶𝐻𝐹(𝑡)

)

𝐶𝐻𝐹(𝑡 = 0)

(2)

𝑙𝑜𝑔𝜀𝑤

where W is the water permittivity; CHF(t) is the measured capacitance at any time t; CHF(t=0) is the measured capacitance at initial time. All the results are presented as function of the reduced time 𝜏 =

𝑡𝑖𝑚𝑒

1/2 -1 𝑡ℎ𝑖𝑐𝑘𝑛𝑒𝑠𝑠 (s .cm )

in order to remove the thickness effect. c. Scanning Electrochemical Microscopy (SECM) The Scanning Electrochemical Microscopy SECM (Biologic M470) was used to in-situ monitor the swelling of the coated substrates. The coated substrates were immersed in NaCl 3% w.t. solution with the addition of a 10 mM potassium ferrocyanide (K4[Fe(CN)]6) as a redox specie. SECM experiments were realized at the different immersion temperatures. A 50µm microelectrode was used as a tip. The tip was polarized at 0.6V/Ag/AgCl in order to get the oxidation of ferrocyanide ions. The current of the tip was measured during approaching the tip towards the substrate in the z-direction (approach curve) with a very slow velocity 0.5 µm/s. When the tip is positioned far away from the substrate (z = 0), the tip current is constant because the redox mediator concentration is constant. When the tip approaches closer to the substrate, the diffusion of the redox mediator is blocked by the insulating coating, its concentration decreases at the tip surface and therefore, the tip current decreases. The approach curve is then in the negative feedback mode. The distance between the tip and the coated substrate at the initial state is d0. During immersion in the saline solution, the coating thickness increases due to swelling and the distance d between the tip and the coated substrate becomes smaller than the initial distance d0. The coating thickness is therefore calculated as the evolution of the distance between the tip and the swollen sample. All details of the experiment can be found in previous studies [12, 21]. III.

Results 1. Initial characterization

The ideal DGEBA/Jeffamine230 polymer network can be represented as shown in Fig. 2. The polymer chains are cross-linked through N atoms from the hardener, which also constitute polar groups as hydroxyl (–OH) or ether (-O-) groups. Between the macromolecular chains, voids exist and are related to the free volume. 5

Journal Pre-proof

: Polar site

Free volume HO

CH3

CH H2C

HO

CH

O

CH N

H2C

CH3

CH2

HC

CH n

CH2

Free volume

OH

CH2 N CH2 CH

OH

Figure 2: Schema of the ideal DGEBA/Jeffamine230 network. The surface of DGEBA/Jeffamine230 coatings obtained by both cooling protocols were scraped in order to get sufficient quantity to perform DSC experiments (Fig. 3) at the dry state (before immersion tests). With a rapid cooling rate of the coatings, the heat flow curve presents a smooth evolution with temperature that allows to evaluate the Tg value at about 91°C ± 1. With a slow cooling rate of the coatings, the heat flow presents an endothermic peak near the Tg value that makes impossible the Tg evaluation and this peak is related to physical ageing [1]. The interesting point is that both cooling protocols were then able to produce two kinds of coatings, with and without physical ageing (PA).

6

Journal Pre-proof

4

Without PA With PA

Heat Flow (W/g)

3 2 1

92°C

0 -1 -2 -3 -4 -5

40

60

80

100

120

Temperature (°C) Figure 3: DSC thermograms of the DGEBA/Jeffamin230 coatings with and without physical ageing (first scan) before immersion tests. Modulated DSC experiments were performed in order to measure the physical ageing (Fig. 4) and to evaluate properly Tg values before immersion tests. The non-reversing heat flow presents an endothermic peak when temperature is increased while an exothermic peak is obtained when the temperature is decreased. In fact, the curve presents an artefact that is due to the dynamic temperature solicitation when performing Modulated DSC, this artefact being present during heating and cooling ramps. That is why it is necessary to subtract the enthalpy of the exothermic peak from the total enthalpy of the endothermic peak, which contains the physical ageing quantity and the artefact response [22]. The amount of physical ageing fractions are presented in Table 1. It can be seen that the rapid cooling of the coatings allowed to get a low PA but did not succeed to avoid completely the PA. On the contrary, the slow cooling rate allowed to obtain a fraction of PA that is significant, along with no increase of Tg and a slight decrease of the free volume. It has to be understood that the PA induced a reorganization of the molecular network with an increase of the number of low-energy configurations between the polymer chains, causing a structure with a lower free volume.

7

Journal Pre-proof

0.15

0.10

0.96 J/g 0.09

0.06

1.37 J/g 0.03

0.02

0.05

-0.02

92°C

0.00 40

60

80

0.00

Rev Heat Flow (W/g)

Heat Flow (W/g)

0.12

Nonrev Heat Flow (W/g)

a)

100

120

0.00

Temperature (°C)

0.06

b)

Heat Flow (W/g)

0.09 0.06

0.98 J/g

0.03 0.00

2.93 J/g

-0.03 -0.06

Nonrev Heat Flow (W/g)

0.12

0.08 0.06

0.04

0.02

0.04 0.02 0.00

0.00

-0.02 -0.04

Rev Heat Flow (W/g)

0.15

-0.02

-0.09

-0.06

91°C

-0.12

-0.04 40

60

80

100

-0.08

120

Temperature (°C) Figure 4: MDSC thermograms of the DGEBA/Jeffamin230 coatings a) without physical ageing and b) with physical ageing (first scan) before immersion tests.

8

Journal Pre-proof

Without PA

With PA

Fraction of PA (J/g)

0.4±0.2

1.9±0.2

Tg (°C)

91±1

92±1

Free Volume (%)

4.8 ±0.4

4.1±0.4

Table 1: Initial properties of DGEBA/Jeffamine230 systems. 2. Sorption curves Typical EIS spectra obtained for a coating without physical ageing during immersion at 40°C in saline solution are presented in Fig. 5. A capacitive response is observed on the whole frequency range and all EIS experiments were similar to this behavior, for coating with and without PA, for all immersion temperatures. Thanks to the Brasher and Kingsbury equation (Eq. 2), sorption curves were calculated and are presented in Fig. 6 for coatings with and without physical ageing during immersion in NaCl 3wt.% solution at different temperatures. These curves present a rapid increase of the water uptake values at the beginning of sorption and tends to a saturation limit for longer immersion time. This is a typical pseudo-fickian behavior[23-26], with water uptake values at saturation that slightly increase with the immersion temperatures from 30°C to 50°C while a neat increase is obtained at 60°C. For coatings without PA, the water uptake values at saturation are close to 5% between 30°C and 50°C while for this temperature domain, the water uptake values at saturation are between 3.5% and 4.5% for coatings with PA. The lower water uptake values with PA can be related to a denser polymer network (smaller free volume fraction), with more polymer chains interactions that create part of the polymer network where water can not penetrate and/or decrease the free volume. The more important water uptake values at 60°C for coatings with and without PA can be related to a temperature that is close to the humid Tg (measured by DSC for water saturated coatings using high-pressure hermetic capsules [27]) which is 72°C. At 60°C, the rubber fraction of the polymer network is not negligible towards the vitreous fraction, which can explain a modification of the water uptake mechanisms and then, higher water uptake values at saturation for both kinds of coatings. From these sorption curves, the water uptake values at saturation can be estimated (Fig. 7). It is clear that the water uptake values are lower for coatings with PA but the difference with values obtained for coatings without PA tends to decrease when immersion temperature increases. It seems then that an increase of the immersion temperature erases the initial physical 9

Journal Pre-proof ageing, as already observed by Akele et al. [7]. This can be explained by an increase of the chain mobility with higher immersion temperatures which finally forces the physically aged polymer network to evaluate towards a structure close to that of the non-physically aged

-90

Phase (°)

-60

126s 7035s 22714s 39946s 57206s

-30

0 -2 10

10

-1

10

0

3419s 14091s 31339s 48577s 66723s 10

1

10

2

10

3

10

4

10

5

10

6

10

10

12

10

11

10

10

10

9

10

8

10

7

10

6

10

5

10

4

10

3

|Z| (.cm²)

coating.

7

Frequency (Hz) Figure 5: Typical EIS spectra obtained for a 240µm thick coating without physical ageing during immersion in NaCl solution at 40°C for 18 hours.

10

Journal Pre-proof

7 6

v (%)

5 4

Without PA 3

30°C 40°C 50°C 60°C

2 1 0 0

15000

30000

45000

60000

75000

1/2

 (s /cm)

7 6

v (%)

5 4 3

With PA 30°C 40°C 50°C 60°C

2 1 0 0

15000

30000

45000

60000

75000

1/2

 (s /cm)

Figure 6: Typical volume water uptake curves obtained from the Brasher and Kingsbury equation for coatings with and without PA aged in NaCl 3wt.% solution at different temperatures.

11

Journal Pre-proof

7.0

Without PA With PA

6.5 6.0

v (%)

5.5 5.0 4.5 4.0

Dry Tg : 92 °C Humid Tg : 72°C

3.5 3.0 20

30

40

50

60

70

Temperature (°C) Figure 7: Evolution of water uptake values V (obtained from the Brasher and Kingsbury equation) at saturation with the ageing temperature and with or without physical ageing. The volumic water uptake values v , from EIS experiments onto coatings, can be compared to the mass water uptake values m that were obtained from gravimetric measurements (with less than 0.1% error) during immersion of DGEBA/Jeffamine230 free films [28], thanks to a procedure described elsewhere [21]. It is then necessary to convert v values into mass values m_v (wt.%) thanks to [21], considering that the mass uptake is negligible compared to the initial coating mass: 𝑚𝐻2𝑂

𝜌𝐻2𝑂.𝑉𝐻2𝑂

𝜌𝐻2𝑂

𝜒𝑚_𝑣 = 𝑚𝑐𝑜𝑎𝑡𝑖𝑛𝑔 = 𝜌𝑐𝑜𝑎𝑡𝑖𝑛𝑔.𝑉𝑐𝑜𝑎𝑡𝑖𝑛𝑔 = 𝜒𝑣.𝜌𝑐𝑜𝑎𝑡𝑖𝑛𝑔

Eq. 3

where H2O is the water density and coating is the coating density (1.176g/cm3 ). The coating density is considered to be constant, since Kong [6] shown that the physical ageing only impacts the third digit of the density value. Table 2 shows that 𝜒𝑚_𝑣 values, obtained from EIS experiments and using Equation 3, are much higher than 𝜒𝑚 values, obtained from gravimetric measurements, which is often the case when comparing directly data obtained from EIS and data obtained with gravimetry [14, 29-32]. Table 2 shows that m_v values are higher than m values and that the analysis of EIS data leads to overestimated values of the water uptake. This will be discussed in following parts.

12

Journal Pre-proof

Without PA

With PA

30°C

40°C

50°C

60°C

m (wt.%)

3.0±0.1

2.9±0.1

2.9±0.1

2.9±0.1

v (vol.%)

4.8±0.2

4.9±0.2

5±0.2

6.5±0.2

m_v (wt.%)

4.1±0.2

4.1±0.2

4.2±0.2

5.4±0.2

m (wt.%)

2.4±0.1

2.5±0.1

2.7±0.1

2.7±0.1

v (vol.%)

3.1±0.2

3.3±0.2

4.4±0.2

6.2±0.2

m_v (wt.%)

2.6±0.2

2.8±0.2

3.7±0.2

5.2±0.2

Table 2: Mass and volume water uptake values from free films and coatings for the DGEBA/Jeffamine230 system. Thanks to EIS experiments, the capacitance at high frequency can be calculated using Eq. 1 and the coating capacitance at the beginning of immersion CHF(t=0) can be evaluated through a linear regression of the first points of the curve (Fig. 8). Then, the relative permittivity of the dry coatings can be calculated using: 𝜀𝑟(𝑡 = 0) =

𝐶𝐻𝐹(𝑡 = 0).𝑒

Eq. 4

𝜀0.𝑆

where e is the coating thickness (cm); S is the coating surface (cm2) under the electrochemical cell and  is the vacuum permittivity (8.85 10-14 F.cm-1). 𝜀𝑟(𝑡 = 0) was calculated with CHF(t=0) that was obtained at two different high frequencies (10kHz and 100kHz) in order to be sure that EIS response in the high frequency domain was only related to the organic coating.

-11

CHF

2

CHF (F/cm )

3.2x10

3.0x10

Linear Fit -11

-11

CHF(t=0)=2.85 10 2.8x10

-2

F.cm

-11

0

5000

10000

15000

20000

25000

30000

35000

1/2

 (s /cm)

Figure 8: Evaluation of the coating capacitance at the beginning of immersion (dry state).

13

Journal Pre-proof

4.4 4.2

r(t=0)

4.0 3.8

10kHz_without_PA 100kHz_without_PA 10kHz_with_PA 100kHz_with_PA

3.6 3.4

30

35

40

45

50

55

60

Temperature (°C) Figure 9: Evolution of the dry coating permittivity 𝜀𝑟(𝑡 = 0) with the immersion temperature and with or without physical ageing. Fig. 9 presents the evolution of the dry coating permittivity 𝜀𝑟(𝑡 = 0) obtained with the immersion temperature and with or without physical ageing. Both frequencies (10kHz and 100kHz) used to calculate the dry coating permittivity give the same trends. Without PA, the dry coating permittivity varies very little considering uncertainty which is in agreement with other works dealing with epoxy systems when the temperature range is far from the Tg value [33, 34]. On the contrary, the dry coating permittivity clearly increases with temperature for coatings with PA. These results show that the two different cooling rates were able to manage the physical ageing amount in DGEBA/Jeffamin230 coatings, before immersion. The lower dry coating permittivity for coatings with PA is well-known in epoxy systems [35]. It can be explained by the densification of the polymer network that decreases the ability of polar groups to move when the electrical field is applied. However, when the temperature is increased, the dry coating permittivity values for coatings with PA tends to equal the values obtained with coatings without PA. It seems that the increasing temperature in immersion tests allows to rejuvenate the coatings, which is a well-known process in physically aged polymers [1, 2, 7]. 3. Swelling measurements

14

Journal Pre-proof The swelling of organic coatings was measured by SECM at 30°C, 40°C and 50°C thanks to approach curves (Fig. 10). The immersion temperature of 60°C was not considered here since it was previously shown that a different sorption mechanism was involved. As it can be seen, the approach curves (normalized current in negative feedback) show a decreasing current value with increasing immersion time, meaning that the tip current due to the redox mediator decreases because the redox mediator diffusion is lowered sooner when the tip goes down to the swelling coating. 1.2 1.0

T = 30°C Without PA

Inorm

0.8 0.6

Immersion time increasing

0.4 0.2 0.0 -300

-250

-200

-150

-100

-50

0

z (m)

1.2 1.0

T = 30°C With PA

Inorm

0.8 0.6

Immersion time increasing 0.4 0.2 0.0

-140

-120

-100

-80

-60

-40

-20

0

z (m) Figure 10: Approach curves obtained by SECM during immersion at 30°C in saline solution of

15

Journal Pre-proof coatings without (110µm thick) and with physical ageing (240µm thick).

4.5 4.0

V/V (%)

3.5 3.0 2.5 2.0

Without PA

1.5

30°C 40°C 50°C

1.0 0.5 0.0

0

10000

20000

30000

40000

1/2

 (s /cm) 4.5 4.0 3.5

V/V (%)

3.0 2.5 2.0

With PA

1.5

30°C 40°C 50°C

1.0 0.5 0.0

0

10000

20000

30000

40000

1/2

 (s /cm) Figure 11: Swelling curves for coatings without and with physical ageing. From the approach curves, the swelling was evaluated (Fig. 11). As a general trend, it can be noticed that coatings without PA swell more than coatings with PA. In a previous work [12], we shown that for a same epoxy system, coatings swell more than free films because of internal 16

Journal Pre-proof stresses within the coating that develop during curing. As said before, the physical ageing allows a denser polymer network, where interactions between polymer chains are stronger and/or more numerous. It can then be reasonably proposed that the internal stresses are maintained tougher with strong polymer chain interactions in presence of PA while without PA, they can be relaxed more easily during plasticization by water molecules. Moreover, fewer polar groups are available to interact with water molecules in presence of PA which decreases the plasticization effect and then the water induced swelling. When the immersion temperature is increased, the swelling increases for coatings with and without PA. The measured swelling is the global swelling caused by temperature (thermal expansion and relaxation of internal stresses) and by water ingress (plasticization), so it is not surprising to observe final swelling values that increase with temperature, even if the relative effect of temperature and water ingress onto the swelling can not be distinguished at this stage. The sorption curves obtained with the Brasher and Kingsbury equation (Eq.2) can now be modified with swelling data [21] and new sorption curves are obtained and plotted vs. the reduces time 𝜏 =

𝑡𝑖𝑚𝑒

𝑡ℎ𝑖𝑐𝑘𝑛𝑒𝑠𝑠 (Fig. 12). Rigorously, the reduced time should be also

modified by the swelling data. However, we checked that the thickness evolution does not significantly change the reduced time since the swelling is only a few percent so the reduced time was only related to the initial coating thickness. The modified sorption curves for coatings without PA ageing give now the same volume water uptake value at saturation (about 3.4%) which is close to the mass water uptake value at saturation obtained with free films (about 3% for all immersion temperatures, see Table 2). The modified sorption curves do not depend on the temperature anymore, because the effect of temperature was included in the swelling curve obtained from SECM and is then removed in the modified sorption curves. The modified sorption curves for coatings with PA present volume water uptake values at saturation between 2.3 and 2.4% and seem to be still dependent on the immersion temperature, even if this effect is quite tiny. This tendency is the same as that observed with free films where the mass water uptake values at saturation were temperature-dependent (see Table 2) and comprised between 2.4 and 2.7%. These results are a new example showing that EIS and gravimetry can give very similar water uptake values when the electrochemical data are corrected by swelling data.

17

Journal Pre-proof 4.0 3.5 3.0

V_corr (%)

2.5 2.0 1.5

Without PA

1.0

Xv_corr_30°C Xv_corr_40°C Xv_corr_50°C

0.5 0.0

0

20000

40000

60000

80000

1/2

 (s /cm) 3.0 2.5

V_corr (%)

2.0 1.5 1.0

With PA Xv_corr_30°C Xv_corr_40°C Xv_corr_50°C

0.5 0.0

0

20000

40000

60000

80000

1/2

 (s /cm) Figure 12: Sorption curves (obtained from the Brasher and Kingsbury equation) modified by swelling data for coatings without PA and with PA. 4. Diffusion kinetics The sorption curves can be fitted using the Fick law (Eq. 3) in order to calculate the diffusion coefficient D of water using [36-38]: 18

Journal Pre-proof

(

8

1



𝜒𝑣(𝜏) = 𝜒∞ 𝑣 . 1 ― 𝜋2∑𝑛 = 0(2𝑛 + 1)2.𝑒𝑥𝑝

(

))

― (2𝑛 + 1)2.𝜋2.𝐷.𝜏2 4

Eq. 3

Where 𝜒𝑣(𝜏) is the volume water uptake at the reduced time 𝜏 =

𝑡𝑖𝑚𝑒

∞ 𝑡ℎ𝑖𝑐𝑘𝑛𝑒𝑠𝑠; 𝜒𝑣 is the

water content at saturation and D is the diffusion coefficient. An example of the fitting results are presented in Fig. 13 for a coating without physical ageing. Initial sorption curves and curves modified by the swelling were considered in order to evaluate if the swelling correction had an influence onto the determination of diffusion coefficients (Table 3).

5

v (%)

4 3 2

v

1

Fick fit v_corr

0

Fick fit 0

5000

10000

15000

20000

25000

1/2

 (s /cm) Figure 13: Typical sorption and modified sorption curves (obtained from the Brasher and Kingsbury equation) fitted using Eq. 3 for a coating without physical ageing during immersion at 50°C.

30°C

40°C

50°C

D (cm²/s) calculated from

Without PA

2.0 ±0.5 10-9

5.6 ±0.5 10-9

13.2 ±0.5 10-9

initial sorption curves

With PA

2.1 ±0.5 10-9

4.4 ±0.5 10-9

10.7 ±0.5 10-9

D (cm²/s) calculated from

Without PA

2.1 ±0.5 10-9

5.0 ±0.510-9

13.3 ±0.5 10-9

modified sorption curves

With PA

1.9 ±0.5 10-9

4.4 ±0.5 10-9

12.4 ±0.5 10-9

19

Journal Pre-proof Table 3: Values of the diffusion coefficient D of water calculated thanks to Eq. 3 with initial and modified sorption curves. From Table 3, it appears that initial and modified sorption curves give the same values of the diffusion coefficient. It is then important to note that the correction of initial sorption curves by the swelling data allows to obtain the same water uptake values at saturation than free films and the same features of kinetic diffusion. It can be also remarked that the diffusion coefficients are slightly lower for coatings with PA than for coatings without PA and that they increase with higher immersion temperatures. Thus, the diffusion process can be characterized by Arrhenius’ law (Eq .4) [38]: 𝐷 = 𝐷0.𝑒𝑥𝑝

( ) ―Δ𝐻 𝑅𝑇

Eq. 4

Where D0 is the pre-exponential factor for the diffusion coefficient, H is the enthalpy of diffusion, R is the perfect gas constant and T is the temperature. Plotting ln(D)=f(1/T) allows to obtain thermodynamic parameters D0 and H (Table 4), as previously detailed [39].

Arrhenius parameters DGEBA/Jeffamine230 ΔH (kJ/mol)

D0 (cm²/s)

Coating without PA

78

5.1 104

Coating without PA with swelling correction

75

1.5 104

Coating with PA

69

1.5 103

Coating with PA with swelling correction

72

5.1 103

Table 4: Thermodynamic parameters for water diffusion in DGEBA/Jeffamine230 coatings without and with physical ageing (PA). From table 4, it can be seen that both D0 and H seem to be lower in presence of physical ageing. However, this tendency should be validated with more experiments at different immersion temperatures. If this tendency would be confirmed, it could be explained that, as said before, physical ageing allows a denser polymer network where polar groups from the 20

Journal Pre-proof macromolecular chains can interact. These polar groups are then less available to interact with diffusing water molecules. A diffusing water molecule can then access more easily another polar group by a less energetic jump, in other words a lower H value. Moreover, if less polar groups are involved in the diffusion process, it also means that less diffusing pathways are possible. There are less choices for the diffusing water molecules to jump from a polar group to another one which means smoother pathways for diffusion and a lower pre-exponential factor D0.

IV.

Conclusions

Two model DGEBA/Jeffamine230 epoxy networks were applied onto aluminum Q-Panel and different cooling procedures allowed to obtain different amount of physical ageing (PA). The epoxy system with PA was denser with a lower free volume. The sorption curves obtained during immersion experiments between 30°C and 60°C showed that water uptake at saturation was lower in presence of PA. It was proposed that polar groups are less available for water diffusion so less water molecules can be absorbed by the physically aged network. Moreover, the swelling of coatings with PA was lower than that of coatings without PA and this was related to stronger or more numerous polymer-polymer interactions. It can then be proposed that physical ageing has a strong effect onto the plasticization phenomenon. The kinetic of water diffusion was evaluated using sorption curves, corrected or not by the swelling. It was shown that the swelling correction does not affect the diffusion coefficients of water: they are lower in the case of coatings with physical ageing. It appears that the control of physical ageing in organic coatings is a key parameter for the service life and the maintain of anticorrosive properties.

Acknowledgements: Authors would like to thank the Région Poitou-Charentes (France) for PhD financing.

21

Journal Pre-proof V.

Bibliography

[1] L.G.E. Struik, Physical Aging in Amorphous Polymers and Other Materials, Elsevier Sci. Publ. Comp., Amsterdam‐Oxford‐New York, (1978). [2] J.M. Hutchinson, Physical aging of polymers, Progress in Polymer Science, 20 (1995) 703-760. [3] Y. Zheng, R.D. Priestley, G.B. McKenna, Physical aging of an epoxy subsequent to relative humidity jumps through the glass concentration, Journal of Polymer Science Part B: Polymer Physics, 42 (2004) 2107-2121. [4] G.M. Odegard, A. Bandyopadhyay, Physical aging of epoxy polymers and their composites, Journal of Polymer Science Part B: Polymer Physics, 49 (2011) 1695-1716. [5] A. Lee, G.B. McKenna, Effect of crosslink density on physical ageing of epoxy networks, Polymer, 29 (1988) 1812-1817. [6] E. Kong, Physical aging in epoxy matrices and composites, in: K. Dušek (Ed.) Epoxy Resins and Composites IV, Springer Berlin / Heidelberg1986, pp. 125-171. [7] N. Akele, F. Thominette, D. Paris, M.F. Pays, J. Verdu, Physical ageing and water sorption in polycarbonate, Journal of Materials Science Letters, 15 (1996) 1001-1002. [8] A. Le Guen-Geffroy, P.Y. Le Gac, B. Habert, P. Davies, Physical ageing of epoxy in a wet environment: Coupling between plasticization and physical ageing, Polymer Degradation and Stability, 168 (2019). [9] D.Y. Perera, Physical ageing of organic coatings, Progress in Organic Coatings, 47 (2003) 61-76. [10] X. Shi, B.M.D. Fernando, S.G. Croll, Concurrent physical aging and degradation of crosslinked coating systems in accelerated weathering, Journal of Coatings Technology and Research, 5 (2008) 299-309. [11] M.G. Olivier, A.P. Romano, C. Vandermiers, X. Mathieu, M. Poelman, Influence of the stress generated during an ageing cycle on the barrier properties of cataphoretic coatings, Progress in Organic Coatings, 63 (2008) 323-329. [12] G. Bouvet, D. Trinh, S. Mallarino, X. Feaugas, S. Touzain, In situ monitoring of organic coating swelling by dynamic mechanical analysis and scanning electrochemical microscopy, Progress in Organic Coatings, 96 (2016) 13-18. [13] D. Nguyen Dang, S. Cohendoz, S. Mallarino, X. Feaugas, S. Touzain, Effects of curing program on mechanical behavior and water absorption of DGEBA/TETa epoxy network, Journal of Applied Polymer Science, 129 (2013) 2451-2463. [14] A.S. Castela, A.M. Simões, Water sorption in freestanding PVC films by capacitance measurements, Progress in Organic Coatings, 46 (2003) 130-134. [15] G.W. Walter, The application of impedance methods to study the effects of water uptake and chloride ion concentration on the degradation of paint films--II. Free films and attached/free film comparisons, Corrosion Science, 32 (1991) 1085-1103. [16] S. Duval, M. Keddam, M. Sfaira, A. Srhiri, H. Takenouti, Electrochemical impedance spectroscopy of epoxy-vinyl coating in aqueous medium analyzed by dipolar relaxation of polymer, Journal of the Electrochemical Society, 149 (2002) B520-B529. [17] D.M. Brasher, A.H. Kingsbury, Electrical measurements in the study of immersed paint coatings on metal. I. Comparison between capacitance and gravimetric methods of estimating water-uptake, Journal of Applied Chemistry, 4 (1954) 62-72. [18] R.G. Duarte, A.S. Castela, M.G.S. Ferreira, A new model for estimation of water uptake of an organic coating by EIS: The tortuosity pore model, Progress in Organic Coatings, 65 (2009) 197-205. [19] J.M. Sykes, A variant of the Brasher–Kingsbury equation, Corrosion Science, 46 (2004) 515-517. [20] S.A. Lindqvist, THEORY OF DIELECTRIC PROPERTIES OF HETEROGENEOUS SUBSTANCES APPLIED TO WATER IN A PAINT FILM, Corrosion, 41 (1985) 69-75. [21] C. Vosgien Lacombre, G. Bouvet, D. Trinh, S. Mallarino, S. Touzain, Water uptake in free films and coatings using the Brasher and Kingsbury equation: a possible explanation of the different values obtained by electrochemical Impedance spectroscopy and gravimetry, Electrochimica Acta, 231 (2017) 162-170. [22] T.A. Instruments, Modulated Differential Scanning Calorimetry Traing Course, La Rochelle University, 2018. [23] A.R. Berens, H.B. Hopfenberg, Diffusion and relaxation in glassy polymer powders: 2. Separation of diffusion and relaxation parameters, Polymer, 19 (1978) 489-496. [24] A.R. Berens, Diffusion and relaxation in glassy polymer powders: 1. Fickian diffusion of vinyl chloride in poly(vinyl choride), Polymer, 18 (1977) 697-704. [25] N. Coniglio, K. Nguyen, R. Kurji, E. Gamboa, Characterizing water sorption in 100% solids epoxy coatings, Progress in Organic Coatings, 76 (2013) 1168-1177. [26] D. Nguyen Dang, B. Peraudeau, S. Cohendoz, S. Mallarino, X. Feaugas, S. Touzain, Effect of mechanical stresses on epoxy coating ageing approached by Electrochemical Impedance Spectroscopy measurements, Electrochimica Acta, 124 (2014) 80-89. [27] I. Merdas, F. Thominette, J. Verdu, Humid aging of polyetherimide. II. Consequences of water absorption 22

Journal Pre-proof on thermomechanical properties, Journal of Applied Polymer Science, 77 (2000) 1445-1451. [28] Y. Elkebir, Influence de la microstructure et du vieillissement par relaxation structural sur les mécanismes d'absorption d'eau dans les résines polyépoxydiques, Ph. D. thesis, La Rochelle University, 2019. [29] F. Bellucci, L. Nicodemo, Water transport in organic coatings, Corrosion, 49 (1993) 235-247. [30] A.S.L. Castela, A.M. Simões, M.G.S. Ferreira, E.I.S. evaluation of attached and free polymer films, Progress in Organic Coatings, 38 (2000) 1-7. [31] A.S. Castela, A.M. Simões, Assessment of water uptake in coil coatings by capacitance measurements, Progress in Organic Coatings, 46 (2003) 55-61. [32] V.N. Nguyen, F.X. Perrin, J.L. Vernet, Water permeability of organic/inorganic hybrid coatings prepared by sol-gel method: a comparison between gravimetric and capacitance measurements and evaluation of non-Fickian sorption models, Corrosion Science, 47 (2005) 397-412. [33] Y. Sun, Z. Zhang, K.S. Moon, C.P. Wong, Glass transition and relaxation behavior of epoxy nanocomposites, Journal of Polymer Science, Part B: Polymer Physics, 42 (2004) 3849-3858. [34] L. Vouyovitch, L. Flandin, J.-L. Bessede, N.D. Alberola, Evolutions of microstructure and dielectric behavior of epoxy based insulator-insulator composites over long periods of time, Journal of Applied Polymer Science, 100 (2006) 3454-3464. [35] M.B.M. Mangion, G.P. Johari, Relaxations in thermosets. VI. Effects of crosslinking on sub‐Tg relaxations during the curing and aging of epoxide‐based thermosets, Journal of Polymer Science Part B: Polymer Physics, 29 (1991) 437-449. [36] J. Crank, G.S. Park, Diffusion in polymers, Academic Press1968. [37] C. Carfagna, A. Apicella, L. Nicolais, The effect of the prepolymer composition of amino-hardened epoxy resins on the water sorption behavior and plasticization, Journal of Applied Polymer Science, 27 (1982) 105-112. [38] M. Huldén, C.M. Hansen, Water permeation in coatings, Progress in Organic Coatings, 13 (1985) 171-194. [39] G. Bouvet, N. Dang, S. Cohendoz, X. Feaugas, S. Mallarino, S. Touzain, Impact of polar groups concentration and free volume on water sorption in model epoxy free films and coatings, Progress in Organic Coatings, 96 (2016) 32-41.

23

Journal Pre-proof CRediT author statement Y. Elkebir : Investigation; S. Mallarino: Investigation; Visualization; Writing - Review & Editing; Methodology, Conceptualization, supervision. D. Trinh: Formal analysis; Investigation; Visualization; Writing - Review & Editing; Methodology. S. Touzain: Investigation; Visualization; Writing - Review & Editing; Methodology, Conceptualization, project administration, supervision, Writing - Original Draft.

Journal Pre-proof

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: