Influence of Fluorozirconic Acid on Sulfuric Acid Anodizing of Aluminum

aCorrosion and Protection Group, School of Materials, The University of Manchester, Manchester M13 9PL, United Kingdom bAlmetron Ltd, Wrexham LL13 9UZ, United Kingdom cGraduate School of Chemical Sciences and Engineering, Hokkaido University, Nishi-8, Kita-ku, Sapporo, Hokkaido 060-8628, Japan dDivision of Applied Chemistry, Faculty of Engineering, Hokkaido University, Nishi-8, Kita-ku, Sapporo, Hokkaido 060-8628, Japan

Chromium (VI)-free conversion treatments that include fluorotitanic and fluorozirconic acids in the bath formulations have been developed for aluminum alloys. [1][2][3][4][5][6] The resultant coatings comprise chromium (III) (if present in the bath), titanium and zirconium fluorides, oxides and hydroxides, which constitute the main part of the coating thickness, and a comparatively thin layer of aluminum oxyfluoride next to the substrate. In comparison with the findings from the relatively numerous studies of conversion coatings produced using fluoroacids, very little information is available on the use of fluoroacids in anodizing of aluminum and aluminum alloys. One study has been made of anodizing aluminum and aluminum alloys at a constant current density in fluoroboric acid solutions at temperatures ranging from 0 to 30 • C. 7 Porous films were formed, with a pore size and population density that depended on the concentration and temperature of the fluoroboric acid solution and the time of anodizing. However, the information provided on the morphology and composition of the films was very limited and the role of the fluoroacid in the growth of the film was not considered. More recent studies examined the influence of ammonium hexafluorosilicate 8 and ammonium fluoride 8,9 additions to oxalic acid on the formation of porous anodic films on aluminum. The effect of hexafluorosilicate ions was more pronounced than that of fluoride ions, suggesting that Si-F bonds may be broken under the high electric field to generate free fluoride ions. 8 Increasing the concentration of the fluorine species in the electrolyte progressively reduced the barrier layer thickness 9 and also the voltage during anodizing under a constant current. 8 It was suggested that fluoride additions led to a microporous outer region of the barrier layer. 9 Furthermore, fluorine was detected at the base of the inner, non-porous region. 9 It was proposed that the incorporated fluoride generated a negative space charge that promoted the injection of Al 3+ ions into the film thereby increasing the current density under a particular applied voltage. 9 The present study has examined the effects of additions of fluorozirconic acid on anodizing of aluminum in sulfuric acid. Sulfuric acid is commonly used in industry for anodizing of aluminum alloys. The aims of the study were to further understand the role of fluorine species in the formation of porous anodic films and to identify any potential benefits that might arise from additions of the fluoroacid to the anodizing bath. * Electrochemical Society Member. z E-mail: p.skeldon@manchester.ac.uk

Experimental
Specimens of size 3 cm x 2 cm were cut from of 99.99% aluminum foil of 0.3 mm thickness. They were then electropolished for 3 min in a mixture of 60% perchloric acid and ethanol, with a volume ratio of 20:80, at 278 K. After rinsing in ethanol and distilled water and application of a lacquer mask (Stopper 45 McDermid) to define a working area on one side of 2 cm 2 , anodizing was carried out at 22 V for 300 and 1200 s at 0 or 20 • C in 0.1 M sulfuric acid containing 0, 0.1, 0.5 or 1.0 wt% of 46% fluorozirconic acid (H 2 ZrF 6 ). The fluorozirconic acid contained free-fluoride, with a concentration of 10 g l −1 as measured by the supplier (Almetron Ltd). The molar concentrations of the fluoroacid and free-fluoride in the prepared electrolytes are given in Table I. Electrolytes were also prepared using 0.1 M sulfuric acid with additions of sodium fluoride (NaF -Fisher scientific 99.0+ %) at concentrations from 3.5 × 10 −3 M to 3.5 × 10 −1 M that were selected to cover the concentrations of total fluoride and free-fluoride in the fluoroacid-containing electrolytes. As well as anodizing in electrolytes prepared with 0.1 M sulfuric acid, films were also formed at 22 V in 1.0 M sulfuric acid at 0 or 20 • C with and without addition of 0.1 wt% fluorozirconic acid.
A two electrode cell was used for anodizing individual specimens, with a cylindrical aluminum sheet of area 200 cm 2 acting as the cathode. The electrolyte was stirred using a magnetic stirrer during the growth of the films. The applied voltage was provided by a GPR-100H05 (Good Will Instrument Co.) power supply. The current passed through the cell during anodizing was recorded on a PC with inhouse developed Labview software. When anodizing was ended, the specimens were immediately removed from the electrolyte and rinsed with deionized water.
Observations of the surfaces and cross-sections of films were made by scanning electron microscopy (SEM) using a Zeiss Ultra 55 microscope operated at 1.5 keV. Compositional analysis of films was carried out using energy-dispersive X-ray (EDX) analysis (Oxford Instrument XMAX 80) in a FEI Quanta 250 scanning electron microscope with a silicon drift detector. For EDX analysis the microscope was usually operated at an accelerating voltage in the range 3 to 6 kV. The accelerating voltage was selected to avoid significant contributions to the X-ray yield from aluminum in the substrate. Selected specimens were also analyzed using accelerating voltages of 10 and 15 kV. The detection limit of the EDX analysis was about 0.1 wt%. The accuracy of the resulting atomic ratios of F:Al and S:Al was estimated to be typically about ±15%. Transmission electron microscopy (TEM) of selected specimens was carried out using an FEI-Philips CM20 instrument, operated at 200 kV. Sections of nominal thickness 20 nm were prepared using a Leica Ultracut microtome, with initial trimming of specimens using a glass knife and final sectioning using a Micro-Star type SU diamond knife.
Elemental depth profiles of the anodized specimens were obtained using a Horiba Jobin-Yvon, RF-5000 glow discharge optical emission spectrometer (GDOES) at a frequency of 13.56 MHz and power of 50 W. Light emissions of characteristic wavelengths during sputtering using neon gas (1100 Pa) were monitored throughout the analysis with a sampling interval of 0.05 s. Use of neon gas was necessary to excite optical emission from fluorine. The wavelengths of the spectral lines used for the analyses were 396.152 nm (aluminum), 130.217 nm (oxygen), 685.602 nm (fluorine) and 339.197 nm (zirconium).  Figure 1 show the trends in the current density during the first 100 s of anodizing. A peak current density occurred for all specimens immediately on the application of the voltage due to rapid growth of a barrier layer under a very high initial electric field. The current density then rapidly decreased to a minimum value due to the reduction in the field as the film thickened. Around the minimum current density, embryo pores develop. For the control specimens, a steady current density associated with the growth of the major pores was subsequently achieved. In contrast, the presence of the fluoroacid led to a second peak in the current density, followed by a slow decline toward a steady value. The second peak occurred at an earlier time with increase of the concentration of the fluorozirconic acid, suggesting that the fluoroacid assisted the establishment of major pores. The effect of the concentration of sulfuric acid on the anodizing behavior was investigated using a 1.0 M sulfuric acid solution containing 0.1 wt% fluorozirconic acid. Films were formed for 1200 s at either 0 or 20 • C. The resultant current density-time curves are shown in Figure  1c, which also includes the curves for the fluoroacid-free solution. Similarly to the observations made with the 0.1 M sulfuric acid, the current density was enhanced by the fluoroacid, with the effect tending to decrease with increasing anodizing time. Figure 2 shows scanning electron micrographs of the resultant film surfaces. Porous surfaces were obtained at 0 • C for all fluorozirconic acid concentrations, and also at 20 • C with 0.1 wt% fluorozirconic acid. However, the addition of 0.5 and 1.0 wt% fluorozirconic acid resulted in a collapsed cell structure. Micrographs of cross-sections of the films, which were prepared by ultramicrotomy, are shown in Figure 3. The film thickness measured from the micrographs and the charge density passed in forming the films are listed in Table II  The thickness values of films that exhibited collapsed cell structures are shown in bold type in Table II Figure 1a. The latter growth rate compares satisfactorily with a rate of 0.48 nm s −1 calculated using the film thickness of Table II determined from the scanning electron micrograph of the film cross-section. The films reveal some damage due to the sectioning by ultramicrotomy using a diamond knife. However, a typical porous structure is evident with mainly unbranched pores. The barrier layers in the films were about 20 nm thick, indicating a formation ratio of about 0.9 nm V −1 , consistent with values reported in the literature for anodizing in sulfuric acid. 10 A similar barrier layer thickness was observed in a specimen anodized in the fluoroacid-free electrolyte (not shown).

Current density-time curves
Sodium fluoride had a generally similar effect on the anodizing curves as the fluoroacid: the charge density at 0 • C increased progressively with additions of 3.5 × 10 −3 , 3.5 × 10 −2 and 3.5 × 10 −1 M sodium fluoride by factors of about 4.2, 6.4 and 8.8 at 0 • C (Table III). At 20 • C, only the first two additions were made, which gave increases in the charge density by factors of about 1.3, 1.6 and 4.7. The highest addition was excluded from examination because of severe film dissolution. The films retained a porous surface structure for additions of 3.5 × 10 −3 and 3.5 × 10 −2 M at 0 • C, and to 3.5 × 10 −3 M at 20 • C. Disintegration of the porous structure due to chemical dissolution occurred for 3.5 × 10 −1 M at 0 • C, and collapse of the structure at 3.5 × 10 −2 M at 20 • C. The ratio of the film thickness to the charge density was in the range 0.44 to 0.54 nm cm 2 mC −1 at 0 • C, but was reduced by film dissolution to the range 0.38 to 0.26 nm cm 2 mC −1 at 20 • C. Increasing additions of sodium fluoride resulted in increments of the film thickness of 3.14, 6.49 and 8.05 at 0 • C for additions of 3.5 × 10 −3 , 3.5 × 10 −2 and 3.5 × 10 −1 M sodium fluoride, respectively, and of 0.36, 1.86 at 20 • C for additions of 3.5 × 10 −3 and 3.5 × 10 −2 M sodium fluoride.    at.%. Thus, the F:Al ratio was enhanced by a factor of 2.7 and 3.6 at 0 and 20 • C, respectively, as the fluoroacid concentration was increased from 0.1 to 1.0 wt%. The results suggest that the effect of the temperature on the F:Al ratio in the films formed was comparatively small. Figure 5 shows the relationship between the current density averaged over the 300 s time of anodizing and the measured F:Al ratio in films formed at 0 and 20 • C in the electrolytes containing fluorozirconic acid and sodium fluoride. The data were fitted by linear relationships, with the gradient at 0 • C being lower than those at 20 • C. Thus, a given fluoride addition to the electrolyte causes a greater increase in the current density at 20 • C than at 0 • C.
The S:Al ratio in the films formed in the 1 M sulfuric acid electrolytes for times of 300 s was increased to about 0.10, compared with values of about 0.06 in the electrolytes containing 0.1 M sulfuric acid (Table IV). The ratio is in reasonably good agreement with previous reports of S:Al ratios in films formed in sulfuric acid under similar conditions. 8,12 The fluorine contents of the films were about 1.0 and 0.7 at.% at 0 and 20 • C, respectively. The respective F:Al ratios were about 0.03 and 0.02, which are about 50 and 30% of those measured in the films formed in 0.1 M sulfuric acid. The film formed at 20 • C for 300 s was also analyzed using a range of accelerating voltages of the electron beam. The detected concentration of fluorine decreased from 0.7 at.% at 5 kV, to 0.5 at.% at 10 kV, and to 0.4 at.% at 15 kV. The O:Al was monitored to ensure that the X-rays were detected only from the film. Considering that a higher acceleration voltage results in increased beam penetration, the results indicate a reduction in the fluorine concentration with increasing depth in the film. In contrast, no significant change occurred in the sulfur concentration.  The F:Al ratio was also increased by increasing the concentration of sodium fluoride at 20 • C, with ratios of 0.03, 0.06 and 0.21 for additions of 3.5 × 10 −3 , 3.5 × 10 −2 and 3.5 × 10 −1 M sodium fluoride, respectively (Table IV) In order to determine the distribution of fluorine, EDX analyses were also made at three locations in a cross-section of a film formed for 1200 s at 20 • C in the electrolyte containing 1.0 wt% fluorozirconic acid. This specimen was chosen because the film was relatively thick, about 2.9 μm. The analyses were made in rectangular areas of size about 6 μm x 0.3 μm. The areas were located within the outer half of the film, in the middle of the film and within the inner half of the film. The O:Al atomic ratio was close to 1.5 in all of the areas, which confirmed that no drift occurred during the analyses. The analyses were carried out at least twice, using different regions of the crosssection. The results are shown in Table V. The fluorine concentration decreased by about 28% between the outer half of the film and the inner half of the film, whereas the sulfur concentration increased by about 36%. The decrease in the fluorine concentration with depth correlates with the decrease in the current density during film growth (Figure 1b). The increase in the sulfur concentration may be a consequence of a  Film compositions: GDOES.-Elemental depth profiles in the anodic films were obtained by GDOES in order to demonstrate the presence of fluorine throughout the thicknesses of films formed for 300 s, thereby complementing the EDX analysis presented previously in Table IV. The profiling was restricted to films formed at 0 • C. Results are presented in Figure 6 for the films formed with different concentrations of fluorozirconic acid. The aluminum/film interface coincides with the decline of the oxygen intensity and the rise of the aluminum intensity as sputtering of the substrate commences. The times required to sputter through the films formed using 0.1, 0.5 and 1.0 wt% fluorozirconic acid were about 37, 68 and 98 s, respectively, correlating with an increasing thickness with increasing additions of fluorozirconic acid. Fluorine was present throughout the thickness of each films. The fluorine signal intensity increased with increase in the concentration of fluorozirconic acid. The signal included background noise, which was estimated from analysis of a control specimen formed in a fluoroacid-free electrolyte. The noise was estimated as about 50% of the total signal intensity of the specimen anodized with 0.1 wt% fluorozirconic acid. The fluorine analyses by GDOES are in qualitative agreement with the EDX analyses, which showed that the fluorine concentration doubled between additions of 0.1 and 1.0 wt% fluorozirconic acid. The fluorine appeared to be distributed uniformly through most of the film thickness, with apparently no depletion with depth, contrary to findings of EDX analysis of thicker films formed for 1200 s ( Table V). The absence of depletion is probably a consequence of faster diffusion of fluorine in the pores of the the thinner films analyzed by GDOES, and possibly also of changes in the sputtering rate of the film during profiling. Sulfur was present throughout most of the thickness of the films. The fluorine signal extended beyond the sulfur signal, which indicates the presence of a sulfur-free, fluorine-containing layer at the base of the films. Optical emission intensities at the wavelength for zirconium (not shown) were at the background level, indicating negligible amounts of zirconium in the films.
Film compositions: RBS.-The zirconium content in a film formed for 300 s in 0.1 M sulfuric acid containing 0.1 wt% fluorozirconic acid at 0 • C was measured by RBS, which is sensitive to the presence of heavy elements in a light matrix. The resultant spectrum revealed yields from oxygen, aluminum and sulfur, as shown in Figure 7. Comparison with a control specimen, anodized in the absence of fluoroacid, showed that only a pile-up background signal occurred in the energy range expected for zirconium in the film. The upper limit on the Zr:Al atomic ratio was 5 × 10 −5 . The S:Al atomic ratio was about 0.06 in agreement with the results of the EDX analyses. The upper limit on the Zr/S ratio was about 8 × 10 −4 . The Zr/S ratio was therefore lower by a factor ≥ 28 than the molar ratio of fluorozirconate ions to sulfate ions in the electrolyte (2.24 × 10 −2 ). The efficiency of anodizing was evaluated from the ratio of the charge due to the Al 3+ ions in the film (276 mC cm −2 ) and the charge passed in the anodizing cell (402 mC cm −2 ). The latter is consumed by oxidizing aluminum, with negligible contributions from side reactions, such as oxygen evolution. The charge ratio indicated that the film formed at an efficiency of about 0.69. Films on control specimens anodized in fluoroacid free electrolytes at 0 and 20 • C formed at efficiencies of 0.63 and 0.64, respectively. The results suggest that the fluoroacid had negligible influence on the efficiency of oxidation of the aluminum and that dissolution of the film was also negligible. In contrast, it was reported previously that the efficiency of oxidation of the aluminum was reduced progressively by increasing additions of ammonium hexafluorosilicate or ammonium fluoride to oxalic acid. 8 After anodizing for 80 min, at a temperature of 17 • C, and an ammonium hexafluorosilicate concentration of 2.5 mM, the efficiency was about 0.5 and reduced to about 0.1 at a concentration of 10 mM. The origin of the low efficiency was unexplained.  specimens were then rinsed in deionized water and re-anodized at 5 mA cm −2 in 0.1 M ammonium pentaborate solution at 20 • C and the voltage surge at the start of anodizing was measured. The thickness of the barrier layer was then calculated from the voltage surge using a ratio of 0.9 nm V −1 , as suggested by the previous TEM examination of films. Figure 8 shows the dependence of the barrier layer thickness on the immersion time. The results indicate a linear relationship between the barrier layer thickness and the immersion time. The dissolution rates determined from the best-fit lines to the data are given in Table VI The voltage surges at the start of re-anodizing of specimens in ammonium pentaborate solution were about 17.8 and 18.4 V for specimens anodized in the presence of 0.1 wt% fluorozirconic acid at 0 and 20 • C, respectively, and 17.5 and 15.5 V for specimens anodized in the electrolyte containing 1.0 wt% fluorozirconic acid. The difference in the surge voltages may be due to the differences in the film compositions, which affect the magnitude of the electric field required for film growth. Furthermore, chemical dissolution of the barrier layer at the termination of anodizing and prior to rinsing of the specimen in de-ionized water may have a significant influence on the voltage surge at 20 • C for the electrolyte containing 1.0 wt% fluorozirconic acid. For instance, from the measured chemical dissolution rate, a loss of barrier layer thickness of about 1.4 nm is predicted to be possible in Table VI 15 Outward migrating Al 3+ ions are ejected at the pore bases to the electrolyte. 16 Only anion species are normally incorporated into the films from the electrolyte, e.g. sulfate, 11,12,17,18 phosphate 18,19 or oxalate 20 ions from sulfuric, phosphoric and oxalic acids. In the absence of film growth at the film/solution interface, located at the pore base, by outward migration of Al 3+ ions, cation species, such as Zr 4+ ions, originating from the electrolyte are not expected to be present in porous films.
The presence of fluorine and the absence of zirconium in the present films may be due to the following reactions: The oxide species produced in (1) may dissolve in the electrolyte or Zr 4+ may be ejected from the film surface under the influence of the electric field. HF may be released to the electrolyte and could attack the film or may provide a source of fluoride for incorporation into the film. Fluoride ions produced by progressive dissociation of fluorozirconic anions, according to (2), may be directly incorporated into the film or combine with H + ions to form HF. The subsequently generated ZrF 3 + ions would migrate away from the film surface. Alternatively, the following reactions may occur: The charge on the ZrOF + should also cause it to migrate from the oxide surface. An analogous species, TiOF + , is considered to be stable in solutions of titanium and fluoride ions. 21 The absence of zirconium species in the anodic films contrasts with the deposition of zirconium oxide during formation of conversion coatings on aluminum alloys in baths containing fluorozirconate ions. The deposition is favored at cathodic sites on the alloy surface where the pH is locally increased by reduction of oxygen, resulting in the reaction: 22 On the contrary, during anodizing, the anodic oxidation of the aluminum generates increased acidity according to the reaction: Thus, conditions for the deposition of zirconium oxide do not occur.
The similarity of the porosities of the films formed at 0 and 20 • C in electrolytes containing a particular concentration of fluorozirconic acid is unexpected from the results of dissolution of the barrier layer post anodizing, which show an increase by about a factor of ten in the rate of dissolution at 20 • C compared with 0 • C. This may be due to differences in the composition and temperature of the pore electrolyte during film growth. The generation of HF from fluorozirconate ions may also play a role in enhancing dissolution at 0 • C.
The results of the EDX analyses in Table IV reveal that the F:Al ratios in films formed in electrolytes containing 3.5 × 10 −2 and 3.5 × 10 −1 M sodium fluoride are of a similar order as in films formed in the electrolytes containing 0.1 and 1.0 wt% fluorozirconic acid, respectively. However, Table I shows that the concentrations of the fluoroacid and free-fluoride are roughly one and two orders of magnitude lower than the concentrations of the sodium fluoride. This suggests that fluoride ions and/or hydrofluoric acid generated from fluorozirconate ions in Reactions 1 to 4 are the main source of fluoride ions in the film. Such a suggestion is consistent with the observation made in earlier published work that the addition of hexasilicate ions to oxalic acid has a greater influence on film growth than additions of fluoride ions. 8 A previous study of fluoride incorporation into barrier-type films showed that fluoride ions migrate inwards in anodic alumina at about twice the rate of O 2− ions. 23 Consequently, a fluoride-rich layer should be present next to the substrate and also along the cell boundaries in the present films, as illustrated in the schematic diagram of Figure 9. The presence of a fluoride-rich layer at the film base was found by XPS analysis of a film formed in oxalic acid with added fluoride, 9 and is suggested to be present in the films formed in sulfuric acid according to the GDOES profiles of Figure 6. The cells in the present films probably comprise three regions of different composition: a fluoride-rich region at the cell boundary, an intermediate region consisting of oxide and fluoride, formed since sulfate ions migrate inward more slowly than O 2− and fluoride ions, and a region next to the pore/electrolyte interface containing oxide, fluoride and sulfate. A similar cell structure occurs in porous films formed on titanium 24,25 and zirconium, 26 where fluoride ions also migrate faster than O 2− ions.
The fluoride ions cause an increase in the growth rate of the film. However, the effect decreases with time due to the reduction in the rate of diffusion of fluoroacid anions and fluoride ions to the pore bases, and hence in the concentration of fluoride ions incorporated into the films. According to the EDX analyses of Table IV, the concentrations of fluoride ions in the films formed for 300 s in the presence of a fluoroacid at both 0 and 20 • C were similar to, or significantly greater than, C839 the concentration of sulfate ions in the films. Furthermore, comparison with the control specimens showed that the incorporation of fluoride did not significantly influence the incorporation of sulfate. The same applied also to films formed in the presence of sodium fluoride at concentrations of 3.5 × 10 −2 M and 3.5 × 10 −1 M. This behavior contrasts with observations made for barrier type films formed in a range of mixed electrolytes. 27 For example, the incorporation of molybdenum species from a molybdate-tungstate electrolyte reduces the incorporation of tungsten species.
From the present results, it is not possible to identify the role of the fluorine species in enhancing the rate of film growth, and more than one mechanism may be operative. The authors have previously proposed that the pores in films formed in sulfuric acid at 20 • C at current densities above about 2 mA cm −2 develop by flow of oxide from the barrier layer into the cell walls. 12 Thus, the increased film growth rate is possibly due to an enhancement in the flow of the film material of the barrier layer due to the incorporation of units of AlF 3 into the film. These species may alter the bonding with in the anodic alumina and decrease the viscosity of the oxide. The flow may also be assisted by higher compressive stresses within the barrier layer due to the volume per mole of aluminum of AlF 3 being greater than that of Al 2 O 3 by a factor of about 1.6. The continuous etching of the barrier layer during film growth by HF generated at the base of the pores may be a further factor in the enhancement of the current density. From the average rates of film growth, determined from film thicknesses in Table II, and the chemical dissolution rates of the films in Table VI, the ratios of the film dissolution rate to growth rates are calculated to be 0.5 × 10 −3 and 1.2 × 10 −3 for anodizing in the electrolyte containing 0.1 wt% fluoroziconic acid at 0 and 20 C, respectively, and 1.5 × 10 −3 and 9.5 × 10 −3 for anodizing in the electrolyte containing 1 wt% fluorozirconic acid at 0 and 20 • C, respectively. However, the dissolution rates may be accelerated during film growth by the electric field and Joule heating and by an enhanced concentration of fluoride ions and hydrofluoric acid generated from the fluorozirconic acid in Reactions 1 to 4. Thus, the significance of dissolution to the film growth is uncertain from the present results. increases the growth rate. However, film dissolution is also enhanced. This can lead to collapse of the porous structure at the film surface, which is promoted by increased fluoroacid/fluoride concentration and electrolyte temperature. 4. The films contain oxide, sulfate and fluoride species. Zirconium species are not incorporated into the films formed in electrolytes containing fluorozirconic acid. The incorporation of fluoride ions has negligible influence on the concentration of incorporated sulfate ions. 5. The migration rate of fluoride ions is faster than that of sulfate and oxide ions. Hence, it is expected that a fluoride-rich layer is present at the base of the barrier layer and at the cell boundaries.