Inﬂuence of Copper on Trivalent Chromium Conversion Coating Formation on Aluminum

In the present work, modiﬁed copper-containing trivalent chromium conversion (TCC) coating processes for aluminum were investigated. The copper addition to the TCC bath was made for the purposes of reducing the generation of hydrogen peroxide and chromium (VI) species during the coating growth. The morphologies and compositions of the coatings were examined using high-resolutionelectronmicroscopy,energy-dispersiveX-rayspectroscopyandRamanspectroscopy.UVphotometricmeasurementswereemployedtodeterminetheamountofhydrogenperoxideintheTCCsolution.Theresultantcoatingscontainedzirconiumoxides,chromiumhydroxideandﬂuoridesandsulfateconstituents,aswellascopperoxidesandcopper-richdepositsthatwerepreferredcathodicsitesforoxygenreduction.Ofmostsigniﬁcance,nochromium(VI)speciesweredetectedinthecoatingsbyRamanspectra.Itissuggestedthatthisresultsfromreducedgenerationofhydrogenperoxide,asdisclosedbyphotometricmeasurements,atthecathodiccopper-richparticles,duetofavoringofthefourelectronoxygenreductionreaction.©TheAuthor(s)2017.PublishedbyECS.ThisisanopenaccessarticledistributedunderthetermsoftheCreativeCommons Trivalent chromium conversion (TCC) coatings are promising eco- friendly alternatives to chromate conversion coatings due to the lower toxicity of trivalent chromium species compared with hexavalent chromium. 1 The TCC coating solution can be regarded as a modiﬁed Zr-based conversion coating solution with addition of small amounts of trivalent chromium salts. 2,3 Our previous work used scanning electron microscopy and atomic force microscopy to reveal cracking and spalling of the coating formed on superpure aluminum, especially after a prolonged conversion treatment. This was associated with the fast kinetics of aluminum dissolution due to the attack by ﬂuorine ions in the reaction solution and the stress in the coating. In contrast, the TCC coating formed on AA2024 alloy suppressed the oxygen reduc- tion reaction, due to the physical barrier created by the Zr-/Cr-rich 1,6,7 and the coating displayed improved adhesion, the coating signiﬁcantly thinner than superpure

Trivalent chromium conversion (TCC) coatings are promising ecofriendly alternatives to chromate conversion coatings due to the lower toxicity of trivalent chromium species compared with hexavalent chromium. 1 The TCC coating solution can be regarded as a modified Zr-based conversion coating solution with addition of small amounts of trivalent chromium salts. 2,3 Our previous work used scanning electron microscopy and atomic force microscopy to reveal cracking and spalling of the coating formed on superpure aluminum, especially after a prolonged conversion treatment. 4 This was associated with the fast kinetics of aluminum dissolution due to the attack by fluorine ions in the reaction solution and the stress in the coating. 5 In contrast, the TCC coating formed on AA2024 alloy suppressed the oxygen reduction reaction, due to the physical barrier created by the Zr-/Cr-rich coatings, 1,6,7 and the coating displayed improved adhesion, although the coating was significantly thinner than that formed on superpure aluminum.
With respect to the effect of copper alloying, George et al. 8 investigated binary alloys containing 1, 5, and 25 at.% Cu prepared by magnetron sputtering and revealed a decrease in the coating growth rate of zirconium-based coatings with increase of the Cu/Al ratio. This was suggested to be due to the copper species present in the coating impeding the cation transport to the coating base. Furthermore, a layer of corrosion products formed at the coating base. Cerezo et al. [9][10][11] modified the Zr-based conversion coating solution by adding a small amount of copper salts (30-50 ppm). The copper components, which had a high deposition tendency, formed copper and/or copper oxide agglomerates on the substrate, which created local alkalinity that supported the coating formation. As a consequence, the coating thickness on the multi-metal substrate (AA6014, cold rolled steel and hot dip galvanized steel) was increased.
One concern about TCC coatings has been the possibility of oxidation of Cr(III) species to form Cr(VI) species either during or following coating growth on aluminum. Cr(VI) species were reported to be present in the freshly-formed TCC coatings, but were undetectable in the case of a fresh coating on AA2024 alloy. 4,12 A generally accepted hypothesis for the formation of Cr(VI) species is that hydrogen peroxide created by the oxygen reduction reaction can oxidize Cr(III) species in the coating. 12,13 Furthermore, the presence of hydrogen peroxide during the conversion treatment of aluminum * Electrochemical Society Fellow. z E-mail: alexander_qi87@sina.com specimens was supported by the results of a recent study that used UV spectrophotometry. 14 In the present work, modified TCC coating solutions were prepared by addition of either 0.05 g/L or 0.5 g/L CuSO 4 to a SurTec 650 solution (a commercial TCC coating solution). The influence of the additions on the kinetics of coating growth was investigated using high-resolution scanning electron microscopy and transmission electron microscopy. Raman spectroscopy and UV spectrophotometry were used to study the coating chemistry, particularly the chromium valence state in the coating and the presence of hydrogen peroxide in the reaction solution, respectively.

Experimental
Conversion treatments.-Specimens of high-purity aluminum, of dimensions of 30 × 20 × 0.3 mm, were electropolished in a mixture solution of 60 wt% perchloric acid and 80 v/v% ethanol (1:4 v/v, <10 • C) at 20 V for 4 min. After electropolishing, the specimens were rinsed in ethanol and deionized (DI) water, followed by drying in a cool air stream.
SurTec 650 chromitAL (SurTec Corp.), a commercial trivalent chromium solution (Zr:Cr atomic ratio = ∼0.7), was diluted with DI water in a ratio of 1:4 by volume. Subsequently, either 0.05 g/L or 0.5 g/L copper sulfate (CuSO 4 , Sigma-Aldrich, U.K.) was dissolved in the solution. The pH of the solutions was then adjusted to 3.9 by adding droplets of 1 wt% NaOH and/or 5 wt% H 2 SO 4 . In order to form the coatings the aluminum, specimens were immersed in the solutions for different times, followed by immersion in DI water at 40 • C for 120 s, rinsing in DI water for 5 s, and drying in a stream of cool-air. 4 In contrast, some aluminum specimens were immersed in 0.01 M sodium fluoride solution containing either 0.05 g/L or 5 g/L copper sulfate. The specimens were prepared to assist understanding of the coating growth in the copper-containing TCC bath. The pH value of the copper-containing NaF solutions was adjusted to 3.9 by adding 5 wt% H 2 SO 4 and 1 wt% NaOH.
Open-circuit potential (OCP) measurements during coating formation were made for selected specimens using a Solarton electrochemical workstation with a ModuLab software controller.
The exposed area of each specimen was ∼2.25 cm 2 . A saturated calomel electrode (SCE, E • = 0.241 V vs. NHE) was employed as the reference electrode.

Solution codes Compositions
F + Acid 0.01 M NaF, pH 3.9 Cu + F + Acid 0.05 g/L CuSO 4 and 0.01 M NaF, pH 3.9 Cu + Cr(III) + F + Acid 0.05 g/L CuSO 4 + 0.01 M Cr(NO 3 ) 3 and 0.01 M NaF, pH 3.9 Cu + Cr(III) + KZF + F + Acid 0.05 g/L CuSO 4 + 0.01 M Cr(NO 3 ) 3 + 0.015 M K 2 ZrF 6 and 0.01 M NaF, pH 3.9 Coating characterization.-The surfaces of coatings were revealed by the secondary electron images recorded by scanning electron microscopy (SEM), using a Zeiss Ultra 55 instrument with an energydispersive X-ray spectroscopy (EDS) facility. Accelerating voltages of 3 kV and 15 kV were used for SEM and EDS, respectively. The SEM/EDS data for each specimen were recorded at three different points and processed by the INCA software (version 4.09). A JEOL 2000 FX II instrument, operated at an accelerating voltage of 120 kV, was used to examine electron transparent thin cross sections of specimens that were prepared by ultramicrotomy as described previously. 15 In order to examine the coating chemistry, a Renishaw 2000 Raman instrument with a 514 nm argon laser excitation was used. The instrument was controlled using GRAMS/32 and WIRE Video View software. The integration time was 30 s, with 10 times accumulation to avoid the effect of stray light noise. 4 The area of the laser beam on the specimen surface was about 1 μm in diameter. Prior to the specimen examination, the argon laser was calibrated at 520 cm −1 using a standard silicon panel. Raman spectra were collected at the frequency range of interest between 1200 and 200 cm −1 . Triplicated measurements were made that showed good reproducibility.
Hydrogen peroxide measurement.-In order to investigate hydrogen peroxide generation, a photometric method using a titanyl reagent (3 M H 2 SO 4 , 0.05 M C 4 K 2 O 9 Ti and DI water of 1:1:4 by volume) was employed. 16 Aluminum specimens, of dimensions 30 × 20 × 0.3 mm, were first ground to a 1200 SiC grit finish, followed by 1 wt% HCl etching at room temperature for 5 s to activate the surface and rinsing in DI water for 15 min. Specimens were then immersed in 250 ml of each of the four solutions listed in Table I at a temperature of 40 • C. The solutions comprised (i) 0.01 M sodium fluoride, (ii) 0.01 M sodium fluoride + 0.05 g/L copper sulfate, (iii) 0.01 M sodium fluoride + 0.05 g/L copper sulfate + 0.01 M chromium nitrate + 0.015 M potassium fluorozirconate. The pH of each solution was adjusted to 3.9 by adding droplets of 5 wt% H 2 SO 4 . The solutions, which were contained in a glass beaker, were stirred to enhance mass transport to the substrate/solution interface. A cuvette was used to contain a mixture of 1 ml of the test solution and 2 ml of the titanyl reagent. The solution was collected after immersion of the aluminum for 0, 300, 600, 1200, 1800, 3600 and 5400 s. A UV-vis spectrophotometer (UV-3000, MAPADA) was employed to investigate the absorbance at 400 nm of the mixture solutions in the cuvette. The concentration of hydrogen peroxide in the test solutions was determined from the absorbance (A H 2 O 2 ) using the relation reported previously: 14 H 2 O 2 concentration (μM) = 2950 x A H 2 O 2 . Furthermore, a careful check was made using a 100 mM H 2 O 2 solution alone and mixed solutions containing 100 mM H 2 O 2 and either 0.05 or 0.5 g/L CuSO 4 . The photometric results showed that the presence of copper ions did not interfere with detection of hydrogen peroxide using the titanyl reagent. All measurements were repeated three times for a particular condition in order to ensure good reproducibility of the data with errors of less than 3%. All chemical reagents in the present work were of analytical grade. DI water of resistivity 18 M cm 2 was used to prepare solutions and to clean specimens.

Results and Discussion
Copper components in the reaction solution.- Figure 1a displays the OCP of the aluminum during immersion for 5400 s in 0.01 M sodium fluoride solution containing either 0.05 g/L or 0.5 g/L copper sulfate. At the OCP, the currents due to the anodic and cathodic reactions are of equal magnitude and vary as the coating develops on the substrate surface. 17 The OCPs initially decreased due to dissolution of the passive film on the aluminum in the presence of fluoride ions, which enhanced the oxidation of the substrate. 18 Figure 1b shows details of the early stages of immersion, showing the higher potential values for the solution of higher copper content, by about ∼99 mV at 300 s. Subsequently the OCPs increased toward a relatively constant value that was about 138 mV higher for the solution containing 0.5 g/L copper sulfate compared with the solution containing 0.05 g/L copper sulfate. These positive OCP shifts are associated with the presence of copper salts in the solution, which led to the deposition of copper at cathodic sites. The amount of deposited copper was greater in the solution containing 0.5 g/L copper sulfate, as shown by the SEM examination below. Figures 2a and 2b show scanning electron micrographs of the specimen surfaces after conversion treatments for 300 s in the solutions containing 0.05 and 0.5 g/L copper sulfate, respectively. EDS analyses (not shown here) revealed that the bright particles were copper-rich deposits that were formed at surface defects or residual flaws in the passive film. 19 The particle coverage was analyzed by the ImageJ software, employing a threshold to distinguish the matrix and particle regions. Using SEM images at the same magnification, areal coverages of 5.5 and 13.1% were obtained for the surfaces immersed in the solutions containing 0.05 g/L and 0.5 g/L copper sulfate, respectively. The greater coverage at the higher copper concentration led to the observed increase in the OCP.   (Table II). After conversion treatment for 300 s the sizes of the copper-rich particles increased and agglomerates of up to 10 μm in diameter were observed (Figs. 5a and 5b). Furthermore, extensive cracking and spalling of the coating occurred in the coating formed in the solution of higher copper content. EDS point analyses (Table  III) revealed higher amounts of chromium and zirconium at particle locations compared to the situations in the matrix. This is similar to the findings after immersion for 15 s (Table II). The analyses made at copper particle-free locations suggest that a small amount of copper may be incorporated into the coating. The results of EDS indicate the formation of a thicker deposited layer above the particles than elsewhere on the aluminum surface. This is due to the enhanced cathodic reaction at the particles, which generates a locally increased pH that promotes the deposition of the Zr-and Cr-rich coating material. 7,20 Figures 6 displays transmission electron micrographs of crosssections of the coatings at locations on the aluminum surface remote from copper-rich particles after conversion treatment for (a-c) 120 s and (d-f) 300 s, in the copper-free SurTec solution 3 (a, d) and solutions containing 0.05 g/L (b, e) and 0.5 g/L (c, f) copper sulfate. The crosssections reveal a chromium-and zirconium-rich outer layer above a thin aluminum-rich layer that contains oxygen and fluorine species. 3 The cracks through the coatings (Figs. 6c and 6d) may form due to dehydration of the coating that occurs in the laboratory air or under vacuum in the microscope. 5 The coating thicknesses were analyzed using the ImageJ software. The results are presented in Table IV. After a treatment of 120 s, the coatings were of similar thickness, about 44 nm. After a treatment of 300 s, the coating thickness had increased to the range 68 to 78 nm. The coating formed in the solution containing 0.5 g/L CuSO 4 revealed the lowest thickness. This is possibly associated with the locally enhanced coating deposition at the copper-rich particles, although variability in the thickness of the coating across the aluminum substrate cannot be ruled out.
Coating and solution chemistry.-Raman spectroscopy, which had an analysis depth of several micrometers, was employed to examine the species present in the coatings by means of the vibrational shifts. 13,21,22 Figure 7 shows Raman spectra after conversion treatment of the aluminum for 300 s in the modified SurTec 650 solutions with 0.05 (Figs. 7a and 7b) and 0.5 g/L (Figs. 7c and 7d) CuSO 4 additions.
The spectrum of Figure 7a recorded at a region of coating on the aluminum substrate in the solution containing 0.05 g/L CuSO 4 reveals peaks for CuO at ∼280 and ∼620 cm −1 and for Cu 2 O at 213 and 690 cm −1 . 23,24 Copper metal is deposited initially at cathodic sites on the aluminum surface. Deposited copper may subsequently be oxidized if it becomes electrically isolated from the aluminum substrate. The isolation may be achieved by the formation of coating material, which EDS analysis indicated developed preferentially at the deposits. The presence of fluoride ions in the formed coatings promoted the oxidation reaction of isolated copper particles. [25][26][27] Stronger peaks due to CuO were revealed at the region of a particle (Fig. 7b) as evidenced by the peaks present at 280 cm −1 . Notably, a satellite peak at 328 cm −1 associated with CuO excitation was detected on the coated particles but not on the coated matrix (Fig. 7a). The presence of the copper-rich deposits (Figs. 4 and 5) can result in surface-enhanced Raman spectra (SERS), 28,29 which may explain the prominent Raman peaks that were associated with copper species, especially on the copper-rich particles. The Raman shifts at 450 and 540 cm −1 in Figure 7a are associated with the presence of zirconium oxide and chromium ((III) oxide/fluoride, respectively. 13 The Raman spectrum for the coating formed in the 0.5 g/L CuSO 4containing solution at a region free from deposited copper (Fig. 7c) was similar to that reported previously for a coating formed in a copper-free SurTec solution. 12 The Raman shifts at 470, 540 and 990 cm −1 are associated with the presence of zirconium oxide, chromium ((III) oxide/fluoride and sulfate, respectively. 13 In contrast, the coated Table II. SEM/EDS point analyses (wt%) of specimen shown in Figures 4a and 4b, N/D = not- ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 207.241.231.83 Downloaded on 2018-07-20 to IP   particles (Fig. 7d) disclosed the presence of CuO (262 and 660 cm −1 ), Cu 2 O (700 cm −1 ), chromium oxides/fluorides (545 cm −1 ) and zirconium oxide (448 cm −1 ). 4,13,30,31 The difference of the Raman peaks associated with the coating components in Figs. 7b and 7d, for example, CuO at 280 and 262 cm −1 , respectively, may result from the effect of particle size on scattering. 32 In addition, the Raman peak at 545 cm −1 represents the presence of a mixture of chromium components, such as chromium oxide (550 cm −1 ), chromium hydroxide (526 cm −1 ) and chromium fluoride (538 cm −1 ). 12,13 With respect to zirconium oxide, the Raman peaks at 448 and 450 cm −1 display negative shifts relative to the reported position at 476 cm −1 . 33 This may result from the effect of contributions of aluminum oxide/hydroxide (438 cm −1 ) 4,28 and zirconium fluorides (450 cm −1 ). 34 Notably, a Raman peak at 780-904 cm −1 , characteristic of Cr(VI), was not resolved in any Raman spectra in Figure 7. 12 The region of the Cr(VI) peak is highlighted in Figure 8 for a Raman spectrum recorded from the TCC coated aluminum matrix after conversion treatment for The absence of a detectable amount of Cr(VI) species in the present coatings is proposed to be related to reduced generation of hydrogen peroxide during conversion treatment of aluminum in the modified solutions, which is supported by the following measurements.
The results of photometric measurements of hydrogen peroxide formed in the solutions listed in Table I are presented in Figure 9. The curves show that the rate of generation of hydrogen peroxide is highest at the start of the period of immersion and subsequently slows to a relatively constant value. In the 0.1 M NaF solution (pH = 3.9, adjusted by 5 wt% H 2 SO 4 droplets), the concentration of hydrogen peroxide rapidly increased up to ∼90 μM at 300 s and then rose more slowly to ∼170 μM at 5400 s. The greatest amount of hydrogen peroxide was detected in the acidified solution containing 0.01 M sodium fluoride. The amount was significantly reduced by the addition of 0.05 g/L copper sulfate, for example, to ∼47 and 100 μM after immersion for 1200 and 5400 s, respectively, compared with ∼145 and 180 μM H 2 O 2 respectively. However, with further modifications of the solution composition by additions of either 0.01 M Cr(NO 3 ) 3 or 0.01 M Cr(NO 3 ) 3 + 0.015 M K 2 ZrF 6 , the concentration of hydrogen peroxide was decreased to still lower levels of 18 and 10 μM, respectively after immersion for 5400 s. Notably, the copper-free SurTec 650 solution contained zirconium and chromium salts, sodium fluoride and sulfates and the atomic ratio of zirconium relative to chromium is 0.7. 35 Therefore the solution containing 0.01 M Cr(NO 3 ) 3 , 0.015 M K 2 ZrF 6 and 0.1 M NaF simulates the commercial SurTec solution used in this work. In comparison with our previous work, 14 a significant reduction of hydrogen peroxide generation was revealed during aluminum immersion in the copper-free solutions containing 0.01 M Cr(NO 3 ) 3 or 0.01 M Cr(NO 3 ) 3 + 0.015 M K 2 ZrF 6 . The results indicate that the addition of copper sulfate to the reaction solutions can inhibit the hydrogen peroxide generation during conversion treatment of aluminum even after prolonged immersion. One reason is associated with the enhanced cathodic reaction around the copper-rich particles in the oxygen reduction reaction. Jakab et al. used the L x LT copper surface to simulate the copper coverage on AA2024 alloy, and the results revealed that the four-electron charge-transfer oxygen reduction reaction rate increased linearly with copper coverage. 36 Furthermore, Colley et al. applied copper microelectrodes (25 μm diameter) to disclose the number electrons involved in the reduction reaction increased toward 4 as the mass transport rate decreased. 37 Thus, it is proposed that the reduction in hydrogen peroxide generation observed with the addition of copper sulfate to the conversion coating solution is associated with promotion of a four electron oxygen reduction reaction (ORR) rather than a two-electron reaction. Notably, our previous paper revealed a copper-enriched layer, a few nanometers thick, at the base of conversion coatings. 4 Its formation is associated with the nobility of copper relative to aluminum. The enrichment develops during pre-treatments of the substrate and is maintained during coating growth. No evidence has been found of an influence of the enrichment on Cr(VI) formation. In contrast, the dissolved oxygen in the coating bath was involved in the generation of Cr(VI), which we attributed to the production of hydrogen peroxide in the two-electron reduction of oxygen. 4 In addition, the reduction mechanism will be investigated by the authors in future studies using a rotating disk electrode. 38,39 In addition, the effect of copper additions to the reaction solution on chromate formation during corrosion and air exposure of the coated aluminum also requires examination.

Conclusions
1. Copper-rich deposits were formed on aluminum after conversion treatment in either 0.1 M sodium fluoride solutions or Sur Tec 650 conversion coating solutions, containing additions of copper sulfate. The copper-rich deposits led to increased an open-circuit potential due to an increased rate of the oxygen reduction reaction. 2. The coatings formed in the copper-containing solution (0.5 g/L) revealed numerous surface cracks and regions of spallation, while such coating defects were relatively negligible in the coatings formed in the modified solution containing 0.05 g/L CuSO 4 . 3. Raman spectra of the coated matrix and particles revealed the presence of copper oxides, zirconium oxides, chromium hydroxide and fluorides and sulfate constituents, while hexavalent chromium species were undetectable. 4. UV photometric measurements were successfully applied to demonstrate the generation of hydrogen peroxide during the conversion treatment of aluminum in the trivalent chromium solutions. 5. The addition of copper sulfate to a TCC solution can reduce the amount of hydrogen peroxide generated by the oxygen reduction reaction by promoting a four-electron reaction at the copper-rich regions.