Etching Behavior and Dielectric Film Formation on Aluminum Foil Stocks for Electrolytic Capacitors: A Review

Abstract

This review delineates the mechanisms of pit nucleation and growth, establishes the foundational principles of etching technology, and presents the findings from investigations on the behavior of anodic dissolution and anodic film formation on high-purity aluminum foils for electrolytic capacitors based on electrochemical analyses and surface electron microscopic observations of etched surfaces. To elucidate pit nucleation and growth mechanisms, the effects of crystalline oxide and small amounts of lead on etching behavior were investigated. Pits initiate at cracks surrounding MgAl₂O₄ spinel or γ-Al₂O₃, resulting from the crystallization of the oxide film at metal ridges on the aluminum substrate. Using ultra-high-resolution field-emission scanning electron microscope (FE-SEM), high-angle backscattered electron (BSE) images revealed the presence of lead as the bright nanoparticles, approximately 10 nm in size, at the surface oxidation layer along rolling lines attributable to pick-up inclusions during hot rolling.

Pitting attacks predominantly occur in the oxidation layer owing to the less noble potential for tunnel dissolution in the initial DC etching phase. Increased titanium content within the aluminum foil accelerated hydrogen evolution in the pit and hydrous oxide formation during the cathodic half-cycle of alternating current etching. The crystallization of anodic oxide films around MgAl₂O₄ spinel crystals, formed in a boric acid solution, was observed using transmission electron microscopy (TEM). Round-shaped γ′-Al₂O₃ formed around the MgAl₂O₄ crystals and expanded across the surface as the formation voltage increased.

1. Introduction

Capacitance of aluminum foils in electrolytic capacitors is determined by their surface area post-etching. The etching technique is chosen based on the working voltage, with DC and AC etching being the popular methods for the anode foils of high- and low-voltage capacitors, respectively. High-purity aluminum foils, characterized by high cubicity, are preferred for DC etching due to the selective dissolution of the (100) planes. The initial pits evolve, subsequently forming tunnels [1]. The morphologies of these tunnel pits are influenced by the presence of oxide particles, dislocations, and lead segregation on the surface of the annealed foils [2, 3].

Conversely, a solid solution of trace elements proves to be more efficacious than surface segregation for the propagation of AC etch pits, attributable to the cyclic process of anodic dissolution and cathodic film formation [4].

The primary objective of this review is to elucidate the anodic behavior of aluminum foils during etching and to furnish essential information on technologies for manufacturing aluminum electrolytic capacitors. A mechanism for pit formation in aluminum foil during the initial stages of etching was proposed, and the impact of various factors, including impurities and electrochemical conditions, on the initial etching phenomena and anodic film formation was elucidated.

2. Behavior of DC Etching

2.1 Changes in electrode potential and pit nucleation during the early stage of DC etching [5]

Three types of high-purity (>99.99%) aluminum foil with varying lead contents (Specimen A: 0.1, B: 0.5, C: 0.8 mass-ppm) were utilized for the DC etching process. The foils had a thickness of 104 µm and a cubicity exceeding 95%. Each specimen underwent annealing at 560°C in an Ar gas atmosphere for 5 h. The specimens were then subjected to galvanostatic etching for up to 50 ms in HCl solutions ranging from 1.5 to 7.6 mol dm⁻³ at 70°C. A constant current density of 200 mA cm⁻² was applied using a potentiostat/galvanostat. Changes in potential were monitored via a digital oscilloscope. The morphology of the etched specimens was examined by resin replicates with a SEM. Additionally, the pit structures in films removed from the etched foil were observed using a TEM after being stripped in an iodine (I₂) methanol solution. The same specimens underwent anodic polarization at a sweep rate of 50 mV min⁻¹ in HCl solutions (0.1 to 7.6 mol dm⁻³) at 70°C. The specimens were either etched or polarized after a 30 s immersion in the electrolyte.

The electrode potential changes during the initial phase of DC etching are depicted in Fig. 1. Initially, the electrode potential was high but subsequently stabilized, with a transition period lasting a few milliseconds.

Fig. 1

Changes in electrode potentials for specimens A to C during the early stage of DC etching at a current density of 200 mA cm⁻² in 1.5 mol dm⁻³ HCl solution at 70°C [5].

SEM images of the resin replicas obtained from specimen C in Fig. 1 following initial DC etching for up to 10 ms are illustrated in Fig. 2. The pit structures exhibited a range from hemispherical to half-cubic, with clusters of hemispherical pits. Notably, pits showing the crystallographic dissolution of the (100) faces and facets were evident after 10 ms of DC etching. Other morphological variations are depicted in Fig. 3. Clusters of multiple pits were observed after 5 ms, as shown in Figs. 2(d) and 3(a). The remaining faces, presumably (111) faces showing crystallographic dissolution, were noted at the pit tips in Fig. 3(b). Therefore, the selective dissolution of the (100) faces commenced within the hemispherical pits, propagated by the nucleation process.

Fig. 2

SEM images of resin replicas obtained from specimen C etched at a current density of 200 mA cm⁻² in 1.5 mol dm⁻³ HCl solution at 70°C for (a) and (b) 2 ms, (c) and (d) 5 ms, (e) and (f) 7 ms, and (g) and (h) 10 ms. Stage tilts for all specimens are 45° except for (h), which is 0° [5].

Fig. 3

SEM images of resin replicas showing crystallographic dissolutions of (100) faces, etched at a current density of 200 mA cm⁻² in 1.5 mol dm⁻³ HCl solution at 70°C for (a) 5 ms for specimen B and (b) 15 ms for specimen A [5].

TEM images capturing the pit structure of specimen A after 50 ms of DC etching are presented in Fig. 4. Observations revealed films along the sidewalls and at portions of the tips within the half-cubic pits. The analysis of these films suggests that the sidewalls are passive, with pit tunnel growth driven by dissolution at the pit tips [1].

Fig. 4

TEM images of a film removed from specimen A etched at a current density of 200 mA cm⁻² in 1.5 mol dm⁻³ HCl solution at 70°C for 50 ms. Stage tilts for (a) and (b) are 0° and 45°, respectively [5].

The average pit dimensions, including widths and lengths of the half-cubic pits and pit densities across a 4000 µm² area, were quantified using SEM images for each specimen. The product NS, derived from pit densities N and average pit areas S, ranged between 2.0 × 10⁻² and 2.6 × 10⁻², as indicated in Table 1. The current densities per pit, calculated by I/NS, were between 7.7 and 10.0 A cm⁻², where I represents the applied current density. Conversely, pit growth rates dl/dt were determined using triangle waves of 2 Hz at 200 ± 60 mA cm⁻², then the current densities per one pit i_d calculated by (dl/dt)(zρF/M) [1] were 8.1∼9.8 A cm⁻². The values calculated by I/NS agreed with i_d, indicating that aluminum dissolves at the tips of the pits.

Table 1 Average pit sizes and densities measured using SEM images of 4000 µm² for specimen A–C etched at a current density of 200 mA cm⁻² in 1.5 mol dm⁻³ HCl solution at 70°C for 50 ms [5].

The growth rate of the hemispherical pit is given as [6]:

  

\begin{equation} \mathrm{dr}/\mathrm{dt} = (\mathrm{i}_{\text{L}}\mathrm{M})/(\mathrm{z}\rho\mathrm{F}) = (\mathrm{DC}_{\text{S}}\mathrm{M})/(\rho\mathrm{r}) \end{equation} (1)

Because the shapes of the initial tunnel pits are half-cubic, the radius of the pit r is given as w/2, and the integrated equation is

  

\begin{equation} \mathrm{w}/2 = \sqrt{(2\mathrm{DC}_{\text{S}}\mathrm{Mt})/\rho} \end{equation} (2)

where

• C_S: Concentration of saturated AlCl₃ in electrolyte (mol cm⁻³)
• D: Diffusivity of Al³⁺ in electrolyte (cm² s⁻¹)
• i_L: Limiting current density at the mouth of a hemispherical pit (A cm⁻²)
• r: Radius of hemispherical pit (cm)
• F: Faradaic charge
• z: Chemical electric charge of aluminum
• M: Atomic weight of aluminum
• ρ: Density of aluminum (g cm⁻³)
• t: Etch time (s)

Where C_S = 2.6 × 10⁻³ mol cm⁻³, D = 6.0 × 10⁻⁶ cm² s⁻¹ [7], and w indicates the average pit sizes given in Table 1. Thus, the formation of a pit requires a long period with the increasing size of the pit when the diffusion of Al³⁺ in the electrolyte is the rate-determining step. High potential periods almost agree with the calculated values.

Across all specimens (A to C), a strong correlation between log[Cl⁻] and the pitting potential is illustrated in Fig. 5. The pitting potentials (E_pit) are determined by the following equation:

  

\begin{equation*} \text{Specimen A:}\ \mathrm{E}_{\text{pit}} = - 0.901-0.162\log[\text{Cl$^{-}$}] \end{equation*}

  

\begin{equation*} \text{Specimen B:}\ \mathrm{E}_{\text{pit}} = - 0.932-0.181\log[\text{Cl$^{-}$}] \end{equation*}

  

\begin{equation*} \text{Specimen C:}\ \mathrm{E}_{\text{pit}} = - 0.942-0.190\log[\text{Cl$^{-}$}] \end{equation*}

The morphology of the aluminum surface following anodic polarization (current density; 0.1 A cm⁻²) is depicted in the SEM images presented in Fig. 6. At low Cl⁻ concentrations, crystallographic dissolutions occur randomly. With increasing Cl⁻ concentration, tunnel dissolutions that grow along the ⟨100⟩ direction, parallel to the original surface, were distinctly observed. The surface dissolution area of specimen C was larger than that of specimen A.

Fig. 5

Relationship between Cl⁻ concentration and pitting potential [5].

Fig. 6

SEM images of specimen surfaces after anodic polarization at a current density of 0.1 A cm⁻² in (a) and (b) 0.1 mol dm⁻³ and (c) and (d) 7.6 mol dm⁻³ HCl solutions at 70°C with a sweep rate of 50 mV min⁻¹. The images show changes in structures from clusters of crystallographic dissolutions to tunnel dissolutions with increasing Cl⁻ concentration in the bulk electrolytes. (a) and (c); and (b) and (d) are the images of specimens A and C, respectively [5].

Figure 7 illustrates the impact of varying HCl concentrations on the electrode potential of specimen A. As demonstrated in Fig. 8, the steady-state potential shifts to a less noble value with increasing Cl⁻ concentration, akin to the behavior observed for the pitting potential.

Fig. 7

Changes in electrode potentials of 3.0 to 7.6 mol dm⁻³ HCl solutions at 70°C for specimen A during the early stage of DC etching at a current density of 200 mA cm⁻² [5].

Fig. 8

Relationship between Cl⁻ concentration and electrode potential [5].

The limiting current density at the mouth of the hemispherical pit, i_L, is given as [7]:

  

\begin{equation} \mathrm{i}_{\text{L}} = 3\mathrm{FDC}_{\text{S}}/\mathrm{r} = 3\mathrm{F}\sqrt{(\mathrm{DC}_{\text{S}}\rho)/(2\mathrm{Mt})} \end{equation} (3)

The average current density for all the side walls of the half-cubic pit, i_w, is given by

  

\begin{equation} \mathrm{i}_{\text{w}} = \mathrm{i}_{\text{L}}/3 = \mathrm{F}\sqrt{(\mathrm{DC}_{\text{S}}\rho)/(2\mathrm{Mt})} \end{equation} (4)

A half-cubic pit comprises four half-square sidewalls and a square bottom face. Assuming the Cl⁻ concentration in the bulk solution and the half-cubic pit are $\text{C}^{0}{}_{\text{Cl}^{ - }}$ and $\text{C}_{\text{Cl}^{ - }}$, respectively, and considering the H⁺ concentration in the pit, C_H⁺ = 0, due to the uncertainty in the hydrolysis of AlCl₃, then:

  

\begin{equation} \mathrm{C}_{\text{Cl${^{-}}$}} = (\Delta\Sigma \mathrm{C}+2\mathrm{C}^{0}{}_{\text{Cl${^{-}}$}})/(1+1/\mathrm{z}) \end{equation} (5)

with the requirement to be electrically neutral, where

  

\begin{equation} \Delta\Sigma\mathrm{C} =(2\mathrm{i}_{\text{w}}\mathrm{r})/(\mathrm{zFD}) = (\mathrm{i}_{\text{w}}\mathrm{w})/(\mathrm{zFD}) \end{equation} (6)

ΔΣC represents the difference in total ion concentration between the pit and the bulk solution (ΣC_i-ΣC⁰_i).

Thus, the dashed line in Fig. 8 was derived using eqs. (4)–(6). Considering the impact of the I.R. drop on the potential, the dashed line more closely approximates the solid line.

The potential peak observed during the initial stages of DC etching was more noble than the steady-state potential. The formation of salt films is also examined. At high anodic potentials, a continuous aluminum chloride film was formed [8]. The formation of this salt film was facilitated by the concentration of Cl⁻ and the diffusion of Al³⁺. On the other hand, hemispherical and crystallographic dissolutions were observed at high potentials and pitting potentials, respectively. The current densities for the anodic polarization curves were limited above the pitting potentials. As Cl⁻ concentration in the bulk electrolytes increased, the pitting potentials shifted to less noble values, and the surface morphologies evolved from clusters of crystallographic dissolutions to tunnel dissolutions. Additionally, an increase in Cl⁻ concentration reduced the high-potential period and rendered the stability potential less noble. The solubility of AlCl₃, denoted as C_S, decreases with increasing HCl concentration, thereby enhancing the formation of a salt film [9] in the pit with rising Cl⁻ concentration.

A high-potential period is crucial for pit nucleation. Pits continue to nucleate within existing pits or other active sites until continuous aluminum chloride films form. During this period, the rate-determining step is the diffusion of Al³⁺ in the electrolyte. Following the growth of hemispherical pits, crystallographic dissolution of the (100) faces leads to the formation of half-cubic pits. These pits, with their facets, grow at a steady-state potential, considered the pitting potential.

2.2 Nucleation process of pits around crystalline oxide particle during DC etching [10]

Two types of high-purity aluminum foils with different magnesium contents (Specimens containing 41 mass-ppm and <1 mass-ppm Mg) were subjected to DC etching at a current density of 200 mA cm⁻² in a 1.5 mol dm⁻³ HCl solution at 70°C. TEM images of the oxide film from the specimen containing 41 mass-ppm Mg are presented in Fig. 9. Particles ranging in size from 0.2 to 0.3 µm were distributed across the surface. The particle indicated by the arrow in Fig. 9(b) was identified as a MgAl₂O₄ spinel crystal through selected-area diffraction and EDX analysis. Similar findings were obtained from the analyses of other particles.

Fig. 9

TEM images of oxide film on aluminum foil (specimen A) [10]. (a) and (b) Surface oxide film; (c) selected area diffraction pattern; and (d) EDX spectra.

TEM images of the surface oxide films from the specimen containing 41 mass-ppm Mg after DC etching for up to 15 s are shown in Fig. 10. Pits were observed around the MgAl₂O₄ spinel crystals after a 10 ms of DC etching. The structure of the pits evolved from hemispherical to half-cubic within 50 ms. The MgAl₂O₄ spinel crystals, some located along the rolling lines, remained at the center of the half-cubic pits. The pits clustered after 500 ms and 1 s, respectively. After 15 s, passive films with waves on the tunnel sidewalls were observed. The same morphologies were noted in the SEM images of the resin replicas, as depicted in Fig. 11. Initial pits are succeeded by dissolution beneath the particles. Films were observed on the sidewalls and parts of the tip of the half-cubic pits. Based on these observations, it was determined that the sidewalls are passive and that the tunnels expand due to facet dissolution at the pit tips.

Fig. 10

TEM images of nucleation of pits around MgAl₂O₄ crystals during the early stage of DC etching [10]. (a) 5 ms, (b) 10 ms, (c) 50 ms, (d) 500 ms, (e) 1 s and (f) 15 s.

Fig. 11

SEM images of resin replicas during the early stage of DC etching. Stage tilts for all specimens are 45°. (a) 5 ms, (b) 10 ms, (c) and (d) 50 ms, (e) 500 ms, and (f) 1 s.

Figure 12 presents TEM images of the pit tips initiated after 50 ms of DC etching. MgAl₂O₄ spinel crystals and facet dissolution were distinctly observed using a preparation method involving jet electropolishing on one side.

Fig. 12

TEM images of pit structures after 50 ms DC etching by jet electropolishing from one side [10]. Tip of the pit with (a) MgAl₂O₄ and (b) facet.

TEM images of the surface oxide films from the specimen containing <1 mass-ppm Mg after 50 ms of DC etching are displayed in Figs. 13 and 14. Particles were distributed over the aluminum foil surface, measuring approximately 0.2 µm in size and identified as γ-Al₂O₃ crystals through selected-area diffraction and EDX. γ-Al₂O₃ crystals remained at the center of the half-cubic pit, with a portion distributed along the rolling lines, exhibiting behavior similar to that of MgAl₂O₄ spinel crystals.

Fig. 13

TEM images of pit around γ-Al₂O₃ crystal after 50 ms DC etching (<1 mass-ppm Mg) [10]. (a) Bright field, (b) dark field, (c) selected area diffraction pattern, and (d) EDX spectra.

Fig. 14

TEM images of pit distribution after 50 ms etching [10]. (a) Pits around γ-Al₂O₃ crystals and (b) cluster of pits along the rolling direction.

TEM images of specimens containing 41 mass-ppm and <1 mass-ppm Mg, prepared using an ultramicrotome, are shown in Fig. 15. Both MgAl₂O₄ spinel and γ-Al₂O₃ crystals were observed at the metal ridges of the aluminum substrates.

Fig. 15

TEM images of ultramicrotomed cross-section of the surface of aluminum foils [10]. (a) MgAl₂O₄ (41 mass-ppm Mg) and (b) γ-Al₂O₃ (<1 mass-ppm Mg).

2.3 Effect of electrolyte immersion time on tunnel pit formation [11]

TEM images of the pits for specimen containing <1 mass-ppm Mg, etched at a current density of 200 mA cm⁻² in 1.5 mol dm⁻³ HCl solution at 70°C for 5 s after being immersed in electrolyte for 0–180 s, are shown in Fig. 16. Clusters of a few micrometers, consisting of multiple cubic pits, were distributed in the specimens after immersion for 0–30 s (a–c). Tunnel pits smaller than 1 µm were observed in specimens immersed for 60 s (d). Step dissolution occurred over all surfaces of specimens immersed for 120 s and 180 s (e, f).

Fig. 16

TEM images of pits after 5 s of DC etching using immersed specimens in electrolyte for (a) 0 s, (b) 10 s, (c) 30 s, (d) 60 s, (e) 120 s, and (f) 180 s [11].

SEM was used to observe anodic oxide film replicas to investigate the distribution of tunnel pits at low magnification. Figure 17 displays SEM images of pits under the same etching conditions as Fig. 16, showing a gradual decrease in pit clusters and an increase in tunnel pits with longer immersion times from 0 s to 60 s. The tunnel pits were most uniformly distributed in specimens immersed for 30 s and 60 s but significantly decreased with surface dissolution in specimens immersed for 120 s and 180 s.

Fig. 17

SEM images of tunnel pits after 5 s of DC etching immersed in electrolyte for (a) 0 s, (b) 10 s, (c) 30 s, (d) 60 s, (e) 120 s and (f) 180 s [11].

Figure 18 illustrates changes in electrode potential during initial DC etching. The high-potential period was extended, and the potentials at peak and steady states shifted to less noble values with increasing immersion time. For example, steady-state potentials for specimens immersed for 0 s and 180 s were −350 mV/SCE and −880 mV/SCE, respectively.

Fig. 18

Changes in electrode potential during the early stage of DC etching [11].

TEM was used to observe surface oxide films stripped from specimens immersed for 0–180 s, as shown in Fig. 19. Dissolutions around γ-Al₂O₃ crystals and fine pits with a diameter of 20 nm were noted after a 30 s immersion, as indicated by arrows. The γ-Al₂O₃ crystals were partially dissolved on some surfaces after a 60 s immersion, as indicated by arrows. Then, they were completely removed after more than a 120 s immersion. Surface dissolutions caused by γ-Al₂O₃ crystals enlarged along the rolling lines.

Fig. 19

TEM images of surface dissolutions of aluminum foils immersed in electrolyte for (a) 0 s, (b) 10 s, (c) 30 s, (d) 60 s, (e) 120 s, and (f) 180 s [11].

The electrode potential transitions from high to steady-state potential during the early stages of DC etching. Pits evolve from hemispherical to half-cubic structures. Pitting attacks occur around the MgAl₂O₄ spinel or γ-Al₂O₃ crystals, both at the aluminum substrates’ metal ridges. Since the initial cracks develop in the amorphous oxide layer at the metal ridges due to tensile stress, γ-Al₂O₃ crystals are found by the direct reaction between molecular oxygen and the bare aluminum surface through the cracks, as indicated by Shimizu et al. [12]. Pits appear initiated at cracks associated with the crystallization of the oxide films.

Pretreatment by immersion for 10 min at room temperature in 1 mol dm⁻³ HCl has been adopted for DC etching in order to obtain a uniform dispersion of tunnels by Alwitt et al. [1]. In this experiment, the areas of dissolution around γ-Al₂O₃ crystals were extended with increasing immersion time in the electrolyte prior to DC etching. When the immersion time in the electrolyte is prolonged from 0 s to 60 s, pits are initiated uniformly because of the increase in active sites. However, excess active sites on the surfaces formed above 60 s delay the tunnel pit growth.

2.4 Segregation of lead on high-purity aluminum foil [13]

The lead distribution can be observed by BSE images at a low accelerating voltage (1.7 kV) using FE-SEM [14]. An example is shown in Fig. 20. The samples are high-purity aluminum foils containing 0.6 mass-ppm and 5 mass-ppm Pb based on 17 mass-ppm Si, 10 mass-ppm Fe, and 58 mass-ppm Cu, with a thickness of 110 µm for high-voltage electrolytic capacitors. By comparing the in-lens image (a) of the foil surface with the BSE image (b) of the same field of view, high-contrast fine particles (arrows in (b)) were observed along the ridges of the raised area in the rolling direction or cracks perpendicular to the rolling direction. High-angle BSE images observed in the sample containing 5 mass-ppm Pb revealed the presence of lead detected by EDS analysis as brighter nanoparticles with size on the order of 10 nm (arrow in (c)).

Fig. 20

SEM images of Pb particles on the surface of aluminum foils containing (a) and (b) 0.6 mass-ppm and (c) and (d) 5 mass-ppm Pb [13]. (a) In-lens SE (1.7 kV), (b) and (c) BSE (1.7 kV) and (d) EDS spectra.

Figure 21(a) presents a SEM image at a low accelerating voltage (0.66 kV) of Al foil containing 0.6 mass-ppm Pb. The surface features both raised and flat regions, with noticeable cracks perpendicular to the rolling direction on some ridges, as highlighted by their oval shape. The in-lens image at 1.7 kV of the same area reveals grainy or linear low-contrast sections in the elevated areas aligned with the rolling direction, while the flat regions exhibit fewer low-contrast areas (b). Fine Pb particles exhibit periodic increases and decreases, with a cycle of approximately 30 µm, mirroring the morphology of the rolled surface, which alternates between raised and flat areas (c).

Fig. 21

SEM images of the surface of aluminum foil containing 0.6 mass-ppm of Pb [13]. (a) SE (0.66 kV), (b) in-lens SE (1.7 kV) and (c) changes in the number of Pb particles perpendicular to the rolling line.

A TEM image of the cross-sectional sample of Al foil containing 0.6 mass-ppm Pb, prepared via focused ion beam (FIB) processing, disclosed non-uniform surface layers with a thickness around 0.5 µm, identified as the surface oxidation layer. Oxygen presence within this layer was confirmed through EDS mapping analysis, as depicted in Fig. 22(a).

Fig. 22

TEM images of the FIB cross-sections of the aluminum foil containing 0.6 mass-ppm Pb [13]. (a) Cross-sectional image and EDS O map, (b) peeled layer and EDS O map, (c) divided surface oxide layer, and (d) metallographic image beneath flat area of surface morphology.

The surface oxidation layer also showed peeled (b) and divided sections (c), with oxygen detected across the entire layer by EDS analysis. Cracks resulting from the division directly exposed the aluminum substrate. Furthermore, the TEM image of the cross-section just beneath the flat areas did not reveal any high-density dislocation bands or cell structures of dislocations (d) [2].

When the cross-sectional sample of the ridge of the aluminum foil containing 8.2 mass-ppm Pb was processed with a cross-section polisher perpendicular to the rolling direction, as illustrated in Fig. 23, fine Pb particles were detected near cracks (c), within voids in the surface oxidation layer, and at the interface between the aluminum substrate and the surface oxidation layer (d) in BSE images using FE-SEM. No fine Pb particles were observed in the flat areas, as indicated in (b).

Fig. 23

SEM images of the cross-sectional BSE image of the aluminum foil containing 8.2 mass-ppm Pb [13]. (a) Surface perpendicular to the rolling line, (b): basis metal under flat area, and (c) and (d) surface oxidation layer.

The aluminum foil with 0.6 mass-ppm Pb underwent etching in 1 mol dm⁻³ HCl and 3 mol dm⁻³ H₂SO₄ at 80°C for durations ranging from 100 ms to 30 s under a direct current (DC) of 250 mA cm⁻². Pitting attacks commenced around the ridges of the substrate in the initial stages of DC etching. As the etching time increased, the dissolution area extended across the entire surface. However, surface dissolution was hindered in parts of the flat area (a), as demonstrated in Fig. 24. Observations via SEM of the oxide film replicas, marked by the dashed line in (a), revealed that tunnel pits in the flat area were notably shorter. The onset of pit growth was delayed (b, c).

Fig. 24

SEM images of surface dissolutions of aluminum foils containing 0.6 mass-ppm Pb after DC etching for (a) 100 ms to 30 s and 40 V anodic oxide film replicas (tilt 45°) formed at 50 mA cm⁻² in 0.8 mol dm⁻³ of ammonium adipate electrolyte at 85°C after DC etching for (b) 5 s and (c) 15 s (arrows indicate etching areas) [13].

Aluminum cross-sections from both hot-rolled (thickness: 4 mm) and cold-rolled (thickness: 0.4 mm) sheets containing 0.6 mass-ppm Pb were examined using TEM and SEM. Surface oxidation layers ranging from 0.2 to 0.5 µm were identified along the rolling direction on the surfaces of both sheets, as depicted in Fig. 25. During the cold rolling process, the surface oxidation layers consisted of multiple strata (b) and (c), with the layer observed to be a grain-refined surface layer (b).

Fig. 25

TEM images of the FIB cross-sections of the aluminum sheet containing 0.6 mass-ppm Pb after hot rolling (a) 4 mm^t and cold rolling (b) and (c) 0.4 mm^t [13]. (a) and (b) TEM images and (c) SEM images.

Figure 26 presents SEM images and EDS oxygen distribution maps (3 kV) for the aluminum sheets containing 0.6 mass-ppm Pb after various stages: hot rolling ((a): 4 mm^t), cold rolling ((b): 1 mm^t, (c): 0.4 mm^t) and final rolling for foil after final annealing ((d): 0.13 mm^t), respectively.

Fig. 26

SEM images and EDS oxygen distribution maps (3 kV) for the aluminum sheets containing 0.6 mass-ppm Pb after various stages: hot rolling ((a): 4 mmt), cold rolling ((b): 1 mmt, (c): 0.4 mmt) and final rolling for foil after final annealing ((d): 0.13 mmt) [13].

Oxygen streaks were detected along the rolling direction in all samples, suggesting the persistence of the surface oxidation layer in contact with the rolls during both hot and cold rolling processes.

TEM and EDS mapping analyses (focusing on O and Al surface analyses) of the surface oxidation layers on the product foil (0.13 mm^t) before and after final annealing, in the cross-sectional direction, are displayed in Fig. 27(a) and (b). Before and after final annealing, the surface oxidation layers exhibited grain-refined structures, with oxygen prominently detected within and between these layers.

Fig. 27

TEM images of the FIB cross-sections and EDS O and Al maps of the aluminum sheet containing 0.6 mass-ppm Pb. (a) Before final annealing and (b) after final annealing.

From the results discussed, it is concluded that the surface oxidation layer results from a phenomenon known as pick-up inclusion, whereby thin layers of aluminum oxide are pressed into the rolled surface during hot rolling [15]. It was determined that these surface oxidation layers act as segregation sites for lead particles during the final annealing process.

3. Behavior of AC Etching (Effects of Titanium on Capacitors during Early AC Etching) [16]

The initiation and propagation of pits in high-purity aluminum foils during etching with a 5 Hz rectangular AC in a 1 mol dm⁻³ HCl solution at 30°C under a current density of ±400 mA cm⁻² were investigated. This study focused on potential changes and pit structure observations conducted using TEM and SEM. High-purity aluminum foils, 100 µm thick and containing 1 and 6 mass-ppm Ti, were examined to discuss the effects of titanium on pit propagation. Films stripped from the etched foils by immersion in I₂ methanol solution and samples prepared by jet electropolishing were used for the TEM observations.

A TEM image of pits produced in one AC cycle etching are shown in Fig. 28. Pits formed at the subgrain boundaries were observed using a preparation with jet electropolishing from one side. TEM images showcasing the surface oxide films from specimens containing 1 mass-ppm Ti and 6 mass-ppm Ti after 5–50 AC cycle etchings are illustrated in Fig. 29. After five AC etching cycles, clusters of multiple pits were observed at the edges of cubic pits. Particles of the etch film, measuring approximately 0.03 µm were only formed in the specimen containing 6 mass-ppm Ti after 10 AC cycles, as highlighted by the arrow in Fig. 29(c). In the specimen with 6 mass-ppm Ti after 50 AC etching cycles, the pit walls were enveloped by thick films, as indicated by the arrow in Fig. 29(b). The anodic oxide film replicas were observed under the SEM to investigate the propagation of pits at low magnifications. Figure 30 shows SEM images of pits for the specimens containing 1 mass-ppm Ti and 6 mass-ppm Ti after 10–250 cycle AC etchings. Pits occurred over all the surfaces and then uniformly propagated in the specimen containing 1 mass-ppm Ti. In the specimen containing 6 mass-ppm Ti, clusters of cubic pits with a size of about 10–20 µm were observed in addition to the non-etched areas.

Fig. 28

TEM image of pits formed at subgrain boundaries [16].

Fig. 29

(a) and (b) TEM images of pits produced by 5 to 50 AC cycle etching and (c) etch films developed by 10 AC cycle etching [16].

Fig. 30

SEM images of pit distributions and structures [16].

The polarization behaviors of specimens containing 1 mass-ppm Ti were examined through cyclic voltammetry [17]. Cyclic voltammograms for these specimens, obtained in a 1 mol dm⁻³ hydrochloric acid solution at 30°C are depicted in Fig. 31. The peaks in the anodic direction diminished with an increasing number of cycles. The breakdown (E_b) and protection (E_pp) potentials were identified at 0.25 V/SCE and −0.95 V/SCE, respectively, for both specimens. The addition of 6 mass-ppm Ti slightly shifted the onset potentials for hydrogen evolution (E_H) to more noble values, and the rate of hydrogen evolution was observed to increase. These findings suggest that titanium present in high-purity aluminum foils promotes hydrogen evolution within the pits. It is proposed that the hydrogen reduction, leading to the formation of etch films on the pit walls, results from a local pH increase during the cathodic half-cycle in AC etching [4]. Therefore, it can be inferred that the excessive formation of etch films, induced by traces of titanium in the aluminum foil, inhibits pit propagation. Further investigation is warranted to discuss in detail the impact of titanium on hydrogen evolution and storage reactions.

Fig. 31

Cyclic voltammograms of specimens [16]. Scan rate: 9.9 V s⁻¹.

4. Behavior of Dielectric Film Formation (Growth of Anodic Oxide Films on Aluminum with MgAl₂O₄ Spinel Crystals) [18]

The behavior of magnesium in the anodic films formed on the surfaces of aluminum foils containing 78 mass-ppm Mg was investigated through TEM observations and EDX analyses. The specimens were annealed at 575°C for 5 h in an Ar gas atmosphere with an oxygen concentration of less than 10 ppm. MgAl₂O₄ spinel particles, measuring 0.1 to 0.2 µm were dispersed across the specimen surfaces. Anodic films were generated at a current density 50 mA cm⁻² in a boric acid solution at 90°C for 2 min. Carbon replicas and cross-sections of these anodic films were prepared for TEM examination.

Figure 32(a) presents TEM images of the extractive carbon replicas of the crystalline anodic films formed in boric acid solutions. Crystalline oxide γ′-Al₂O₃ with round shapes [19–21] was identified around MgAl₂O₄ particles, expanding across the surface as the forming voltage increased up to 60 V. Magnesium was not observed in γ′-Al₂O₃ when the voltage exceeded 100 V, and cracks appeared in the crystalline oxide particles within the anodic films formed at 150 V (b). At 300 V, the crystalline oxide was directly divided by a crack (c). The formation of voids and cracks is attributed to the higher migration rate of Mg²⁺ compared to Al³⁺ under a high electric field.

Fig. 32

TEM images of extractive carbon replicas of crystalline anodic films formed in boric acid solutions at 50 mA cm⁻² up to 40 V, 60 V, 100 V, and 150 V for 2 min [18]. (a) Extractive carbon replicas, (b) EDX spectra and (c) TEM image of ultramicrotomed cross-section of anodic film formed in boric acid solution at 50 mA cm⁻² up to 300 V for 2 min.

5. Conclusions

This review clarified the pit nucleation and growth mechanisms and established the fundamentals of etching technology by describing the results of an investigation on the behavior of the anodic dissolution and anodic film formation of high-purity aluminum foils for electrolytic capacitors based on electrochemical analyses and surface electron microscopic observations of the etched surfaces. The following conclusions can be drawn from the findings.

1. (1)    During the initial stages of DC etching, pits transition from hemispherical to half-cubic shapes. Hemispherical pits initially form and expand under the diffusion control of aluminum ions within the bulk electrolyte during periods of high potential. Pits exhibiting facets grow under a steady-state potential.
2. (2)    Pitting attacks are observed around MgAl₂O₄ spinel or γ-Al₂O₃ crystals. Both MgAl₂O₄ spinel and γ-Al₂O₃ crystals are found at the metal ridges on the aluminum substrate. Pit initiation occurs at cracks, which are linked to the crystallization of the oxide film.
3. (3)    As the electrolyte immersion time before DC etching increases, the areas of surface dissolution around the γ-Al₂O₃ crystals on the aluminum foils become more extensive. With prolonged immersion, pits initiate uniformly across the surface due to a rise in the number of active sites, leading to a delay in tunnel pit growth due to excess active sites.
4. (4)    High-angle BSE imaging reveals the presence of lead in the form of bright nanoparticles, approximately 10 nm in size, located at the surface oxidation layer along the rolling lines. This is attributed to pick-up inclusions during hot rolling.
5. (5)    Increased titanium content within aluminum foils enhances hydrogen evolution within pits and the formation of hydrous oxides, or etching films, during the cathodic half-cycle of AC etching.
6. (6)    When aluminum foil containing MgAl₂O₄ spinel particles is anodized, round-shaped γ′-Al₂O₃ crystals form around the MgAl₂O₄ crystals, growing in size with an increasing formation voltage.

Acknowledgments

The author is thankful for the financial support from The Light Metal Educational Foundation, Inc. of Japan.

REFERENCES

• 1)   R.S.  Alwitt,  H.  Uchi,  T.R.  Beck and  R.C.  Alkire: Electrochemical tunnel etching, J. Electrochem. Soc. 131 (1984) 13–17. doi:10.1149/1.2115495
• 2)   M.  Kawai,  K.  Yamada,  T.  Umezaki,  A.  Konishi,  H.  Watanabe and  Y.  Miura: Microstructure of surface area and etching behavior of high-purity aluminum foils for electrolytic capacitors, J. JILM 54 (2004) 211–217. doi:10.2464/jilm.54.211
• 3)   H.  Tsubakino,  A.  Nogami,  M.  Terasawa,  T.  Mitamura,  T.  Yamanoi,  A.  Kinomura and  Y.  Horino: Surface segregation of lead in high-purity aluminum during heat treatment, J. JILM 52 (2002) 410–416. doi:10.2464/jilm.52.410
• 4)   C.K.  Dyer and  R.S.  Alwitt: Surface changes during A.C. etching of aluminum, J. Electrochem. Soc. 128 (1981) 300–305. doi:10.1149/1.2127408
• 5)   N.  Osawa and  K.  Fikuoka: Pit nucleation behavior of aluminium foil for electrolytic capacitors during early stage of DC etching, Corros. Sci. 42 (2000) 585–597. doi:10.1016/S0010-938X(99)00117-1
• 6)   T.R.  Beck and  R.C.  Alkire: Occurrence of salt films during initiation and growth of corrosion pits, J. Electrochem. Soc. 126 (1979) 1662–1666. doi:10.1149/1.2128772
• 7)   T.R.  Beck: Size distribution of etch pits in Aluminum, Electrochim. Acta 33 (1988) 1321–1327. doi:10.1016/0013-4686(88)80121-X
• 8)   T.R.  Beck: Salt film formation during corrosion of aluminum, Electrochim. Acta 29 (1984) 485–491. doi:10.1016/0013-4686(84)87098-X
• 9)   T.R.  Beck: Electrical properties of aluminium chloride film on corroding aluminium, Electrochim. Acta 30 (1985) 725–730. doi:10.1016/0013-4686(85)80119-5
• 10)   N.  Osawa and  K.  Fukuoka: Effects of crystalline oxide on pit nucleation of aluminum foil for electrolytic capacitors during D.C. etching, J. Surf. Finish. Soc. Jpn. 50 (1999) 643–647. doi:10.4139/sfj.50.643
• 11)   N.  Osawa and  K.  Fukuoka: Effects of electrolyte immersion time on pit formation of high-purity aluminum foil for electrolytic capacitors during DC etching, J. Surf. Finish. Soc. Jpn. 50 (1999) 824–828. doi:10.4139/sfj.50.824
• 12)   K.  Shimizu,  R.C.  Furneaux,  G.E.  Thompson,  G.C.  Wood,  A.  Gotoh and  K.  Kobayashi: On the nature of “Easy Paths” for the diffusion of oxgen in thermal oxide films on aluminum, Oxid. Met. 35 (1991) 427–439. doi:10.1007/BF00664713
• 13)   N.  Osawa,  M.  Tomino,  T.  Hayashi,  K.  Ueda and  T.  Motoi: Segregation of lead on high purity aluminum foil for electrolytic capacitors, J. Surf. Finish. Soc. Jpn. 73 (2022) 504–511. doi:10.4139/sfj.73.504
• 14)  K. Shimizu and T. Mitani: A New World of Scanning Electron Microscopy, (The Nikkan Kogyo Publishing Prodyction Co. Ltd., 2008) pp. 10–12.
• 15)  M. Muramatsu: Study on rolling tribology of aluminum plate, Research group report No. 59, (J. Jpn. Inst. Light Metals, 2013) pp. 140–160.
• 16)   N.  Osawa and  K.  Fukuoka: Effects of titanium on pit formation in aluminum foil for electrolytic capacitors during early AC etching, J. Surf. Finish. Soc. Jpn. 50 (1999) 819–823. doi:10.4139/sfj.50.819
• 17)   W.M.  Moore,  C.T.  Chen and  G.A.  Shirn: A voltammetric investigation of AC corrosion phenomena at an aluminum electrode in hydrochloric acid, Corrosion-NACE 40 (1984) 644–649. doi:10.5006/1.3593902
• 18)   N.  Osawa and  K.  Fukuoka: Growth of anodic oxide films on aluminum with MgAl₂O₄ spinel crystals, J. Surf. Finish. Soc. Jpn. 51 (2000) 710–713. doi:10.4139/sfj.51.710
• 19)   C.  Crevecoeur and  H.J.  Wit: The anodization of heated aluminum, J. Electrochem. Soc. 134 (1987) 808–816. doi:10.1149/1.2100577
• 20)   K.  Kobayashi,  K.  Shimizu and  H.  Nishibe: The structure of barrier anodic films formed on aluminum covered with a layer of thermal oxide, J. Electrochem. Soc. 133 (1986) 140–141. doi:10.1149/1.2108509
• 21)   K.  Shimizu,  K.  Kobayashi,  G.E.  Thompson and  G.C.  Wood: The apparent induction period for γ-Al₂O₃ development in thermal oxide films on aluminium, Oxid. Met. 36 (1991) 1–13. doi:10.1007/BF00938453