3.1. Inhibiting Ability of Gallic Acid
The evaluation of the inhibiting ability of gallic acid was started from the determination of the aluminum corrosion potential. This was performed using the open circuit potential measurements. The OCP vs. time dependencies were determined in deaerated 0.5 M H
3PO
4 aqueous solutions containing gallic acid (
Figure 2). OCP decreases immediately when aluminum is exposed to an acidic solution because this solution rapidly dissolves a thin layer of Al
2O
3 formed on the aluminum electrode in air. Subsequent decrease in the OCP is much slower, and after circa 3 h the stationary value is reached. This corresponds to the corrosion potential E
corr and was calculated as the average of the OCP values obtained during the last 5 min of the corrosion experiment. It can be concluded that gallic acid does not influence E
corr. To confirm this, a statistical analysis of the corrosion potential was conducted using Equation (1). The following values were obtained: −1.151 ± 0.008 V and −1.154 ± 0.009 V vs. REF for
CGA = 0 and 38 mM, respectively.
The potentiodynamic polarization method is commonly applied in corrosion studies. It helps to understand the corrosion mechanism and offers the possibility to estimate the corrosion rate. In this work, the polarization curves were determined as a function of the initial concentration of gallic acid, in 0.5 M H
3PO
4 aqueous solutions (
Figure 3a).
The anodic polarization curves have a typical shape for aluminum with no linear part visible [
2]. Thus, only cathodic curves can be used for the determination of the corrosion current density. The kinetics of the anodic reaction, which is aluminum oxidation, does not depend on the concentration of gallic acid.
The cathodic polarization curves describe mainly a hydrogen evolution process. They exhibit a rather short linear part corresponding to about one decade of the current density. This occurs between the overpotential η = −0.1 and −0.2 V. When the overpotential is more negative than −0.2 V, the polarization curves deviate from linearity and reduction peaks appear. These peaks correspond to the reduction of gallic acid. When its concentration
CGA = 1 mM, the maximum of the reduction peak is located at η < −0.5 V and is not visible in
Figure 3a. The peak positions are shifted in an anodic direction as
CGA increases. Gallic acid is not reduced when −0.2 V < η < 0 V, because all the polarization curves in
Figure 3a overlap with the one obtained at
CGA = 0 mM. This also means that gallic acid does not influence the cathodic process in the corrosion cell when aluminum corrodes at the corrosion potential.
It remains unclear why the reduction peak also appears when
CGA = 0 mM. The following three explanations can be proposed. The first one is the reduction of impurities from the electrolyte. Orthophosphoric acid applied in all experiments was of analytical purity and contained 0.00005 wt. % of As, 0.00002 wt. % of Cu, 0.00005 wt. % of Pb, and 0.0005 wt. % of Fe. This corresponds to the following concentrations of impurities in the electrolyte: 29 μg∙dm
−3 of As, 11 μg∙dm
−3 of Cu, 29 μg∙dm
−3 of Pb, and 287 μg∙dm
−3 of Fe. These are rather low values and probably do not affect the shape of the cathodic polarization curve. The other explanation is the reduction of oxygen. The electrolyte was always thoroughly deoxygenated with Ar before the experiments. Argon was also applied to the electrochemical cell over the electrolyte’s surface during the experiments. Nevertheless, oxygen is produced on the counter electrode during cathodic polarization and may diffuse toward the working electrode and undergo reduction. The last explanation is the reduction of H
3PO
4, which is the most abundant in the studied system among the possible electroactive species. The possible reactions are listed in
Table 1. The driving force for these reactions is high because the reduction peak is centered at circa −1.45 V vs. REF (Ag|AgCl (3M KCl)).
The Tafel extrapolation method was applied to determine the corrosion current density i
corr and the Tafel slopes for the cathodic reactions (see example in
Figure 3b). The obtained Tafel slopes do not depend on
CGA and are within the range between −166 and −181 mV∙dec
−1. These values are more negative than −120 mV∙dec
−1 obtained for aluminum in acetic or sulfuric acid [
33]. This means that when η is between −0.2 V and 0 V there is an additional electrochemical process apart from hydrogen evolution. Corrosion current densities are between 36 and 49 μA∙cm
−2 and do not depend on the gallic acid initial concentration. These values correspond to linear corrosion rates between 0.39 and 0.53 mm∙year
−1. The corrosion rates are much lower when compared to those obtained for aluminum of technical purity in 0.5 M H
3PO
4: 1.96 mm∙year
−1 [
5] and 2.56 mm∙year
−1 [
34].
The linear corrosion rate describes well the studied system because of the uniform character of the corrosion. This was confirmed microscopically (
Figure 4). SEM analysis of the corroded aluminum specimens did not reveal any significant influence of gallic acid on the corrosion process. The uniform character of corrosion is related to the high purity of aluminum. Usually the corrosion of aluminum alloys, even in acidic solutions, has a localized character, due to the presence of numerous intermetallic particles in their microstructure [
35].
3.2. Corrosion Mechanism
Electrochemical impedance spectroscopy was used to provide a deeper understanding of the system studied. First, the impedance spectra were recorded at the open circuit potential, after 3 h of exposure of aluminum in the corrosive environment, as a function of
CGA (
Figure 5 and
Table 2). The impedance spectra for
CGA = 1 and 10 mM are not shown, because they are essentially the same as those obtained for
CGA = 0 and 38 mM.
The impedance spectra were quantitatively analyzed using the electrical equivalent circuits. The spectra recorded at OCP were approximated using the equivalent circuit presented in
Figure 6a. This circuit consists of resistors
R1–
R4, constant phase elements
CPE1,
CPE2, and the capacitor
C1. The physical meaning of
R1,
R2, and
CPE1 is straightforward, while the interpretation of
R3,
R4,
CPE2, and
C1 is complicated and requires detailed knowledge of the system studied.
Resistors
R1 and
R2 correspond to the solution resistance and the charge transfer resistance, respectively. The latter takes into account both anodic and cathodic processes because the spectra were recorded at the OCP.
CPE1 corresponds to the double layer capacitance. The electrical capacity of the double layer is usually not modeled using the capacitor because it describes the idealized electrode/electrolyte interface with only one time constant of the charge transfer process. In reality, there is always a certain distribution of these time constants. Thus, the constant phase element is typically used to obtain the capacitance of the electrical double layer. The impedance of
CPE is given by Equation (2) [
36]:
where
T is a parameter corresponding to the capacitance,
α is the parameter describing the deviation from the purely capacitive behavior,
ω is the angular frequency, and
j is the imaginary number.
The capacitance of the electrode’s double layer,
Cdl, can be calculated using Equation (3), assuming the surface distribution of time constants of the charge transfer process [
36]:
Elements
R3,
R4,
CPE2, and
C1 correspond to the adsorption of e.g., reaction intermediates and/or corrosion inhibitor as well as the formation of the protective layer onto aluminum. Impedances of these elements are related to the rate constants of the aforementioned processes and changes in the surface coverage in the case of both adsorption and the formation of the protective layer. These dependencies are complex and interrelated [
36,
37].
The fitting results together with their statistical analysis are provided in
Table 2. This analysis confirms that gallic acid does not influence the corrosion process of aluminum in the orthophosphoric acid solution. Polarization resistance
Rp, which is the sum of the four resistances, equals 1013 ± 97 Ω∙cm
2 and 936 ± 45 Ω∙cm
2 for
CGA = 0 and 38 mM, respectively. This means virtually the same value of corrosion rate in both cases. The same values of
T2, α
2,
R3,
C1, and
R4 obtained for
CGA = 0 and 38 mM indicate that gallic acid molecules do not adsorb onto aluminum. Therefore, these elements of the equivalent circuit correspond to adsorption of reaction intermediates and/or the formation of the protective layer. Adsorption of reaction intermediates seems to be plausible only for the cathodic process, where adsorbed hydrogen atoms are formed [
36]. The charge transfer related to aluminum oxidation in acidic solutions is believed to be too rapid to observe the formation of any intermediates [
38]. The protective layer that could be formed and dissolved simultaneously is Al
2O
3. This explains the lack of the linear part on the anodic polarization curves in
Figure 3a. Double layer capacitances also support the hypothesis of partial coverage of the electrode with Al
2O
3, because they are smaller when compared to the typical value for metallic electrodes, 20 μF∙cm
−2 [
39]:
Cdl = 9.1 ± 0.6 μF∙cm
−2, and 9.6 ± 0.3 μF∙cm
−2 for
CGA = 0 and 38 mM, respectively.
The important question arises as to why gallic acid does not adsorb onto aluminum. Gallic acid is a weak acid with pK = −4.152 [
40]. This means that at pH = 6.0 it is almost completely (98%) dissociated. When the solution becomes acidic, gallic acid is protonated. Thus, it was important to check whether at pH = 1.1 gallic acid exists as the neutral molecule or if further protonation is possible, e.g., of the oxygen from the carbonyl group. The protonation of gallic acid can be studied using UV-Vis spectrophotometry because it produces a bathochromic shift of the absorption bands. Gallic acid has two absorption bands in the ultraviolet part of the spectrum. These bands are related to π → π* transitions [
41]. At pH = 6.0, when gallic acid is almost completely dissociated, these bands are centered at 212 and 259 nm. They are only slightly shifted at pH = 4.4, where circa 64% of gallic acid molecules are dissociated, and much stronger at pH = 1.1 (
Figure 7a). The absorption spectra obtained for gallic acid dissolved in H
3PO
4 and HCl at this pH are the same. This means that there are no specific interactions between gallic and orthophosphoric acids.
The absorption spectrum obtained in 0.5 M H
3PO
4 can be compared with the one calculated for an undissociated gallic acid molecule using the TD-DFT method (
Figure 7b). Positions of absorption bands in both spectra are comparable. This, and agreement with the already published data [
41,
42], strongly suggest that during the corrosion tests, gallic acid was in the form of an undissociated molecules. For additional confirmation, the UV-Vis absorption spectrum of the gallic acid molecule with protonated oxygen atom from the carbonyl group was calculated. This protonation process produces a strong bathochromic shift of the absorption band centered at λ = 278 nm to λ = 341 nm, and an additional absorption band also appears at 388 nm. There are no absorption bands at λ > 320 nm in the experimentally obtained spectra. Therefore, the oxygen atom from the carbonyl group is not protonated at pH = 1.1.
Adsorption onto a metallic electrode is related to the electrical charge of both counterparts. The sign of the electrical charge on aluminum depends on its potential in the electrolyte. In this work, the corrosion process was studied mainly at the corrosion potential. Its average value, circa −1.15 V vs. REF, is more negative when compared to the potential of zero charge of aluminum, −0.73 V vs. REF [
43]. The value of corrosion potential is between the potentials of anodic and cathodic areas in the corrosion cell. This means that, under studied conditions, at least the anodic areas are negatively charged. The electrical charge at the cathodic areas cannot be deduced in this manner; it can be positive of negative. The interaction of gallic acid with the metallic surface is possible via hydroxyl or carbonyl groups. All of them are negatively charged as is shown in
Figure 8. Thus, the probability of adsorption of gallic acid on the anodic areas of the corrosion cell is low and the anodic process in the corrosion cell is not inhibited. In fact, there is an overall positive charge at the benzene ring, but it apparently does not facilitate adsorption on anodic areas of corroding aluminum. Further studies, e.g., with local electrochemical methods, are necessary to explain why gallic acid does not adsorb at the cathodic areas in the corrosion cell. Adsorption of the neutral molecule should be possible at a potential close to the potential of zero charge [
44]. Adsorption of gallic acid onto aluminum oxide is possible at pH = 1.1, because the oxide surface is positively charged then [
45].
The negatively charged surface of aluminum attracts cations from the electrolyte. These are, in addition to hydronium, aluminum complexes such as [AlPO
4H
2]
2+ and [AlPO
4H]
+. Their presence in the solution at pH = 1.1 can be suggested based on thermodynamic calculations (
Figure 9). Complexes of gallic acid with aluminum are not formed because they are less stable than complexes with phosphate anions [
40,
46]. This confirms the results of corrosion experiments. The formation of stable complexes of aluminum with gallic acid would result in an increased corrosion rate.
Impedance spectra recorded at the OCP characterize both anodic and cathodic processes in the corrosion cell. These processes were studied separately using the following methodology. First, aluminum corroded freely at the OCP for 3 h, and the impedance spectrum was recorded. Subsequently, the working electrode was polarized either −0.1 V or +0.1 V vs. OCP. The constant potential was then held for 1 h, and the impedance spectrum was recorded. This procedure was carried out for
CGA = 0 and 38 mM, and again, there was no influence of gallic acid on the studied processes (
Figure 5a,b,
Table 3 and
Table 4).
The impedance spectrum characterizing the anodic processes, i.e., obtained at η = +0.1 V, consists of two time constants (
Figure 5a) and was approximated with the electrical equivalent circuit presented in
Figure 6b. This circuit can describe the passivation process [
37], where
R3 and
C1 correspond to the formation of the protective layer. The impedances of these two elements depend on the charge transfer resistance and describe how the electric current changes when the surface is gradually covered with the protective layer or when the resistivity of this layer increases. When the electrode potential is below the active-passive transition potential
R3 has a positive value and dissolution of the protective layer prevails. Nevertheless, a certain deviation from the Tafelian behavior is visible in the polarization curve. The higher the value of
R3, the more significant this deviation is. The active-passive transition occurs when
R3 goes to infinity. When the metal is in the passive state,
R3 is negative [
37].
It can be observed that
R3 = 374 Ω∙cm
2, obtained for the anodically polarized electrode (
Table 4), and this is significantly higher than both
R3 = 150 Ω∙cm
2 and
R4 = 136 Ω∙cm
2 obtained at the OCP (
Table 3). Thus, anodic polarization enhances the formation of the protective layer. This should decrease the double layer capacitance of the electrode because the dielectric permittivity of Al
2O
3 is lower when compared to that of water. This seems to occur in the studied system.
Cdl decreases from 8.6 to 7.8 μF∙cm
−2 when
CGA = 0 mM and from 9.3 to 8.2 μF∙cm
−2 when
CGA = 38 mM. These, however, are small differences, and further statistical analysis is necessary to prove it. The protective layer is simultaneously formed and dissolved in an orthophosphoric acid solution, and either the surface coverage or the film thickness is insufficient to induce the active–passive transition. Gallic acid does not influence the passivation process; therefore, virtually the same values of
C1 and
R3 were obtained for
CGA = 0 and 38 mM.
The impedance spectrum characterizing the cathodic processes consists of the three time constants, with the inductive loop at the low-frequency range (
Figure 5b). The impedance spectrum was approximated using the electrical equivalent circuit presented in
Figure 6c. This circuit consists of resistors
R1–
R4, constant phase elements
CPE1 and
CPE2 as well as the inductor L
1. Resistors
R1 and
R2 correspond to the solution and the charge transfer resistance, respectively. The constant phase element
CPE1 represents the double layer capacitance. Resistors
R3 and
R4,
CPE2, and L
1 can be related to the adsorption of two different species [
47].
The results of approximation of the impedance spectra obtained at OCP and subsequent cathodic polarization are presented in
Table 5 and
Table 6 respectively. Cathodic polarization caused a very small change in the double layer capacitance.
Cdl decreased with polarization from 8.5 to 8.2 μF∙cm
−2 when
CGA = 0 mM and from 9.9 to 9.2 μF∙cm
−2 when
CGA = 38 mM. On the contrary to the anodic polarization, the cathodic one caused a significant decrease in charge transfer resistance (
R2 in all equivalent circuits). This is understandable because cathodic polarization should facilitate charge transfer from the electrode to hydrogen ions in the electrolyte. More interesting is the nature of the inductive loop appearing upon cathodic polarization. The most important cathodic process in the studied system is hydrogen evolution. Typically, this process is modeled with a simple equivalent circuit, such as presented in
Figure 6b [
36]. This might suggest that the inductive loop corresponds to the adsorption process that is not related to hydrogen evolution. It has already been mentioned that there is more than one cathodic process in the system studied. The following possibilities were proposed: reduction of impurities from the electrolyte, oxygen generated at the counter electrode, and orthophosphoric acid. The latter seems to be the most plausible. That is because the first two of them should be diffusion-controlled due to the low concentration of either impurities or oxygen in the electrolyte. Diffusion processes, contrary to adsorption, do not manifest in the impedance spectra in the form of the inductive loop [
36].
An additional corrosion experiment was performed to verify whether the inductive loop is related to the presence of orthophosphoric acid in the studied system. During this experiment, aluminum corroded freely in a 0.1 M HCl aqueous solution for 5 h. Then, the impedance spectrum was recorded. Subsequently, the electrode was polarized cathodically vs. OCP, η = −0.1 V. The potential was held for 1 h and the impedance spectrum was recorded (
Figure 10).
The impedance spectrum obtained at the cathodic polarization can be approximated using the equivalent circuit from
Figure 6b, which is typical for hydrogen evolution [
36] (results in
Table 7). This confirms the assumption that the inductive loop obtained in 0.5 M H
3PO
4 at η = −0.1 V (
Figure 5b) is related to the adsorption of a reduced form of orthophosphoric acid.
It is also interesting to compare the impedance spectra obtained at the OCP in 0.1 M HCl and 0.5 M H
3PO
4. The former exhibit the inductive loop at the low frequencies, thus the equivalent circuit from
Figure 6c was applied (
Table 8). Similar spectra were recorded for iron in an acidic solution, at the potential below the active-passive transition potential [
37]. Because the data points at the low frequencies are scattered and their number is too small a reliable determination of
R4 was not possible. Nevertheless, this value is very high. This fact, together with the high value of the charge transfer resistance,
R2, indicates that at the same pH, orthophosphoric acid is much more corrosive than hydrochloric acid. This, in turn, can be related to the formation of stable aluminum complexes in the H
3PO
4 aqueous solution.