porous anodic alumina formed on aa6063 aluminum alloy in a ... text.pdf · under mild conditions...

NR 4/2016 INŻYNIERIA MATERIAŁOWA MATERIALS ENGINEERING 163

Porous anodic alumina formed on AA6063 aluminum alloy in a two-step process combining hard and mild anodization

Małgorzata NorekDepartment of Advanced Materials and Technologies, Faculty of Advanced Technologies and Chemistry, Military University of Technology,

Warszawa, Poland, *[email protected]

Two-step process combining hard (in 0.3 M oxalic acid solution) and mild (in 0.1 M phosphoric acid solution) anodization at voltages ranging between 120 and 180 V was applied to prepare porous anodic alumina (PAA) on AA6063 alloy foil. The influence of ethanol on geometrical parameters of the PAAs was also tested. The analysis of the geometrical parameters as well as the current/voltage vs time transients suggested that alloying elements present in the AA6063 alloys play an important role in the relaxation of mechanical stresses at the metal/oxide interface occurring during the PAA’s growth. The ef-fect contributed to comparable interpore distance Dc values (between 270÷370 nm) and similar regularity ratio parameter (RR) in the samples anodized in ethanol-free and ethanol-modified electrolyte, despite much larger average current densities registered for the former samples. The hexagonal pore ordering increased with the applied anodizing voltage.

Key words: hard anodization, guided anodization, porous anodic alumina (PAA).

Inżynieria Materiałowa 4 (212) (2016) 163÷171DOI 10.15199/28.2016.4.3© Copyright SIGMA-NOT MATERIALS ENGINEERING

1. INTRODUCTION

Highly-ordered porous anodic alumina (PAA) is one of the most attractive templates for nanofabrication [1, 2]. Its geometrical pa-rameters including pore diameter and pore arrangement, interpore distance, or PAA film thickness are easily controlled by anodizing conditions such as type of electrolyte, applied voltage, anodiza-tion time, or bath temperature [3, 4]. Usually PAA is fabricated on high-purity aluminium (99.999%) which by far hoists its produc-tion costs. Therefore, from economical point of view a replacement of the expensive high-purity aluminium for low-purity aluminium which is about two orders of magnitude cheaper is very desirable.

The structural and compositional features of aluminium sub-strate play an important role in formation of long-range ordered po-rous anodic alumina. The low-purity aluminium contains a variety of alloying elements, which can bring about new behaviors during anodization process, not observed during anodization of high-pu-rity aluminium. The PAA fabrication on AA1050 aluminium alloy was already a subject of quite a number of studies [5÷9]. In order to obtain high quality PAA on the low-purity aluminium new ap-proaches towards surface preparation were developed. For instance, the AA1050 surface roughness was improved by optimizing several electropolishing variables (electropolishing time, the volume per-centage of perchloric acid in ethanol-perchloric acid solution, bath temperature etc.) preceding the anodization process in a mixture of sulfuric and oxalic acids [10, 11]. Moreover, a special surface treatment of AA1050 alloy involving electropolishing, acid etching, stripping, or cold-rolling prior to anodization, were applied to im-prove geometrical and morphological parameters of PAA [12, 13]. The influence of bath aging on the PAA’s thickness, pore structure and the barrier layer thickness was also analyzed [14]. Even though all these treatments have resulted in PAA of better quality, the pore arrangement was always worse as compared to the arrangement in PAA produced on high-purity aluminium [15, 16].

The surface pretreatments of aluminium are often time-con-suming and offer only a limited upgrading of PAA parameters. So far, the anodization of technical purity aluminium was conducted under mild conditions (mild anodization, MA), involving low cur-rent densities. The MA process is performed within specific volt-ages, lower than a critical voltage above which it is not possible to

carry on a stable anodization. It was previously demonstrated that the current density is a key parameter governing the pores order-ing [17, 18]. An excellent long-range pore arrangement is achieved during high field, high current anodization (hard anodization, HA) on high-purity aluminium [19]. In this process, a stable electrolysis can be performed at voltages much higher than the critical ones, and the problems related with heat dissipation are overcome by forma-tion of a thick barrier layer at the beginning of the process during MA, before voltage is ramped to a target formation potential. Hard anodization was performed in oxalic [19÷21], sulfuric [22, 23], and phosphoric [24] acid solutions. Thanks to the application of high potentials new self-ordering regimes, with interpore intervals in the range of hundreds of nanometers, were found. Well ordered PAA templates with the pore intervals in this range are of great in-terest for fabrication of optical devices operating in UV or visible spectral region [25÷27]. The major drawback of PAA films pro-duced in one-step HA is the irregular resulting top surface, which is a consequence of voltage increase at the beginning of the HA process. A parallel pore alignment from top to bottom of PAA’s film thickness can be obtained during so-called guided anodization. In the guided anodization regular hexagonal Al nanoconcave arrays formed on aluminium after removing of the PAA produced in the first HA step are used as nucleation sites to maintain the ordered pore growth during second MA step, which is performed at constant voltage throughout the whole process [28, 29].

In this work, the effect of high-field anodization on geometrical parameters of porous anodic alumina formed on AA6063 alumin-ium foil is studied. The two-step process combining HA in 0.3 M oxalic acid solution and MA anodization in 0.1 M phosphoric acid solution is applied to fabricate PAAs in the voltage range between 120 and 180 V in water-based and in ethanol-modified electrolytes. It was previously observed that ethanol has an essential impact on anodization of high-purity aluminium and resulting PAA’s morpho-logical features such as hexagonal pores ordering and the interpore distance Dc [29]. The influence of ethanol on the PAA’s structure formed in the MA–HA anodization of the AA6063 alloy under the same experimental conditions as the ones applied during the anodi-zations of high-purity aluminium is examined.

164 INŻYNIERIA MATERIAŁOWA MATERIALS ENGINEERING ROK XXXVII

2. EXPERIMENTALCommercially available AA6063 aluminium alloy sheet with a thickness of about 0.5 mm was cut into rectangular specimens of size 2 cm × 1 cm. Before the anodization process the Al sheets were degreased in acetone and ethanol and subsequently electropolished in a 1:4 mixture of 60% HClO4 and ethanol at 0°C, and constant voltage of 20 V, for 2 min. Next, the samples were rinsed with etha-nol first, then with distilled water and finally dried. As prepared Al coupons were insulated at the back and the edges with acid resist-ant tape, and serve as the anode. A Pt grid was used as a cathode and the distance between both electrodes was kept constant (ca. 5 cm). A large, 1 L electrochemical cell, a powerful low-constant-temperature bath, and vigorous stirring (750 rpm) were employed in the anodizing process. An adjustable DC power supply with volt-age range of 0÷300 V and current range of 0÷5 A, purchased from NDN, model GEN750 TDK Lambda, was used to control the ap-plied voltage, and APPA 207 TRUE RMS multimeters were used to measure and transfer the registered current and potential data to a computer.

The first anodization was carried out at 0°C in 0.3 M H2C2O4 water-based solution and in ethanol-water 1:4 volumetric mixture. Anodizing voltages were ranging from 120 to 180 V. Prior to appli-cation of a given voltage, the samples were pre-anodized at 40 V for 10 min. Then, the voltage was slowly increased to a target value and the samples were anodized for 2 h. As obtained alumina was chemi-cally removed in a mixture of 6 wt % phosphoric acid and 1.8 wt % chromic acid at 60°C for 3 h. Second anodization was performed in 0.1 M phosphoric acid solution with 1:4 v/v mixture of ethanol and water as a solvent, at 0°C under the same anodizing voltage as used in the first step.

Morphology of PAAs was studied using field-emission scanning electron microscope FE-SEM (FEI, Quanta) equipped with energy- dispersive X-ray spectrometer (EDS). The chemical composition analysis was performed at 20 kV, magnification of 2000, and with a constant distance of samples to the detector. In the case of PAA, the cross-sections were analysed. Each measurement was repeated three times and an average of the three measurements was taken to determine the chemical composition of a given sample.

To obtain geometrical parameters of the fabricated PAAs, Fast Fourier Transforms (FFTs) were generated based on three SEM im-ages taken at the same magnification for every anodizing voltage, and were further used in calculations with WSxM software [30]. To estimate regularity ratio, radial average was generated from each FFT image. The regularity ratio RR was estimated according to the following formula (1) [31]:

RR H

Wn

S= ⋅

1 23 2

//

(1)

where n is the number of pores, S is the analyzed surface area, H is the maximum intensity value of the FFT intensity profile, and W1/2 is the width of the intensity profile at half of its height. The average interpore distance Dc was estimated as an inverse of the FFT’s ra-dial average abscissa from three FE-SEM images for each sample.

3. RESULTS AND DISCUSSION

Directly after electropolishing the AA6063 alloy samples were sub-jected to hard anodization. In Figure 1 the current density/voltage vs time curves measured during hard anodization in water-based oxalic acid solution (Fig. 1a) and in electrolyte with 4:1 v/v water to ethanol mixture (Fig. 1b) are presented. The HA method con-sists of two stages: anodization under the increasing voltage and a stable-voltage anodization. Voltage increase up to a target value Ut is followed by the increase of the current density ia. The sharp cur-rent increase was previously associated with ionization of charged particles (acid anions, in this case C2O4

2– anions) in the high electric field, which are being incorporated into the PAA framework and re-lease primary electrons into the oxide conduction band [32]. These electrons are accelerated by the high electric field enhancing the av-erage current flow. In the current transients at voltages higher than 140 V, the curves demonstrate two maxima, the second maximum being smaller than the first one. After reaching the first maximum, the current density decreases.

The decrease may be associated with a shift of equilibrium be-tween the oxidation rate at the Al–Al2O3 interface and the dissolu-tion rate at the Al2O3–electrolyte interface towards a faster oxidation rate. As a result of the faster oxidation rate, the oxide barrier layer becomes thicker, which is translated into a drop of the current den-sity. As the voltage rises, the equilibrium is restored and the barrier layer gets thinner again recovering a smooth current flow, which is manifested in a second increase of the current density (the second maximum). After reaching the target voltage, the HA proceeds to the stable-voltage anodization, where the current density decreases exponentially owing to limited diffusion of the ionic species (e.g., C2O4

2–) from the electrolyte to the electrolyte-oxide interface [32]. As compared to the HA of high-purity aluminium under the same electrochemical conditions (0.3 M oxalic acid solution, 0°C etc.) [29] the current densities are around two times smaller. The same trend was observed during mild anodization of high-purity Al foil

Fig. 1. Current density/voltage vs time curves recorded during hard anodization of the AA6063 samples at 120, 140, 160, and 180 V in ethanol-free (a) and ethanol-modified (b) electrolyteRys. 1. Krzywe gęstości prądu/napięcia w funkcji czasu anodowania twardego stopu aluminium AA6063 dla napięć HA odpowiednio 120, 140, 160 i 180 V w czystym elektrolicie (a) oraz w elektrolicie modyfikowanym etanolem (b)

120 V 120 V

140 V 140 V

160 V 160 V

180 V 180 V


and AA1050 alloy in sulfuric and oxalic acid solutions: the current values were always lower for the AA1050 samples [15]. Moreover, the current densities at the target voltages change negligibly with the voltage. It appears that there is a boundary ia value which is not exceeded despite increasing of Ut. The basically different cur-rent curves acquired for AA6063 anodization as compared to the currents registered during anodization of high-purity Al indicate a strong influence of impurities on the anodization process.

The rough composition of AA6063 was estimated by EDS analy-sis and is gathered in Table 1. The alloy contains Mg (~0.6 at. %), Si (~0.3 at. %), and Fe (~0.2 at. %) as principal alloying elements. Apart from the principal elements other elements including Cu, Ti, Zn, Mn, and Cr were also detected (the remainder). A relatively high amount of oxygen (~3 at. %) suggests that the alloy was par-tially oxidized in air. The influence of various elements on the process as well as on the resulted anodic alumina was analysed previously [33÷36]. According to those analysis, the amount and distribution of insoluble impurities will strongly influence the qual-ity of anodic oxide layers formed on the aluminium substrate under specific conditions. Moreover, the presence of various additives can affect both average ions migration rates through the barrier oxides and field-assisted dissolution of oxide. In the studied cases it is very probable that the impurities in the AA6063 alloy hindered the ion migration through the barrier layer resulting in the lower recorded steady-state current densities as compared to those registered for high-purity aluminium [29]. Furthermore, in contrast to the HA of high-purity aluminium, there is a negligible difference between the ia values reached at the Ut between the samples anodized in water-based electrolyte (Fig. 1a) and in electrolyte with 4:1 v/v water to ethanol mixture (Fig. 1b). The influence of ethanol is manifested, however, in ia decaying process: in the case of anodization in water-based electrolyte the current decrease is more gentle than in the case of HA in electrolyte with 4:1 v/v water to ethanol mixture. The current density drops below 10 mA/cm2 after around 3500 s for the samples anodized in ethanol-free electrolyte (Fig. 1a), whereas in the ethanol-modified electrolyte the current exceeds the 10 mA/cm2 value well before 2000 s in the case of anodization at 120, 140, and 160 V, and after ~2250 s for the sample anodized at 180 V (Fig. 1b). A steady-state anodization in ethanol-free electrolyte is carried out at the ia of 6÷8 mA/cm2, whereas in ethanol-modified solu-tion the ia attained at the stable-voltage anodization stage is around 3÷5 mA/cm2. Accordingly, the average current density is higher in the case of samples anodized in water-based electrolyte as compared to the ones anodized in ethanol-modified acid solution (Fig. 2).

The lower currents registered during anodization in the presence of ethanol were ascribed to the ethanol cooling effect [24, 28, 32]. The cooling effect of ethanol was related with ethanol ability to lower the dissociation constant of the electrolyte and to decrease of temperature generated at the PAA barrier layer. These effects, in turn, lead to the reduction of the ion diffusion ability and smaller amount of acid anions in the barrier layer, and consequently to a rel-atively smaller contribution from the ionic current.

As formed anodic alumina was chemically removed and SEM images of the resulted nanoconcave arrays formed on aluminium are presented in Figure 3.

As compared to Al nanoconcaves formed on high-purity Al [29], the hexagonal pore arrangement is worse in all studied cases. There is also no clear ethanol influence on the pore ordering. The geomet-rical features of Al nanoconcaves, such as regularity of hexagonal pore arrangement and the pitch size, are directly transferred from Al concavities to the corresponding PAAs grown on their base during mild anodization in phosphoric acid solution at the same operating temperature (0°C) and at the same anodizing voltages. Therefore, the analysis of the SEM images of top part of the PAAs is equivalent with the SEM analysis of the Al nanoconcave arrays and will be presented below.

The current-time transients acquired during the second anodiza-tion step are given in Figure 4. The curves show typical behaviour

Table 1. Results of EDS chemical analysis of AA6063 foil and PAAs anodized at 120, 140, 160, and 180 V, at. % (Table 1 gives only ap-proximate values)Tabela 1. Wyniki analizy EDS folii AA6063 oraz dla anodowego tlenku wytworzonego w napięciach MA odpowiednio 120, 140, 160 180 V, % at. (tabela 1 podaje jedynie orientacyjne wyniki)

AA6063 120 V 140 V 160 V 180 V

Al 95.88 47.25 45.29 48.23 46.53

O 2.53 51.42 53.63 50.64 52.21

Mg 0.59 0.57 0.51 0.53 0.41

Si 0.26 0.24 0.12 0.12 0.14

Fe 0.22 0.06 0.05 0.04 0.06

Remainder 0.52 0.46 0.40 0.44 0.65

The decrease may be associated with a shift of equilibrium

between the oxidation rate at the Al-Al2O3 interface and the dissolution rate at the Al2O3-electrolyte interface towards a faster oxidation rate. As a result of the faster oxidation rate, the oxide barrier layer becomes thicker, which is translated into a drop of the current density. As the voltage rises, the equilibrium is restored and the barrier layer gets thinner again recovering a smooth current flow, which is manifested in a second increase of the current density (the second maximum). After reaching the target voltage, the HA proceeds to the stable-voltage anodization, where the current density decreases exponentially owing to limited diffusion of the ionic species (e.g., C2O4

2-) from the electrolyte to the electrolyte-oxide interface [32]. As compared to the HA of high-purity aluminum under the same electrochemical conditions (0.3 M oxalic acid solution, 0°C, etc.) [29] the current densities are around two times smaller. The same trend was observed during mild anodization of high-purity Al foil and AA1050 alloy in sulfuric and oxalic acid solutions: the current values were always lower for the AA1050 samples [15]. Moreover, the current densities at the target voltages change negligibly with the voltage. It appears that there is a boundary ia value which is not exceeded despite increasing of Ut. The basically different current curves acquired for AA6063 anodization as compared to the currents registered during anodization of high-purity Al indicate a strong influence of impurities on the anodization process.

The rough composition of AA6063 was estimated by EDS analysis and is gathered in Table 1. The alloy contains Mg (0.59 at%), (Si 0.25 at%), and Fe (0.22 at%) as principal alloying elements. Apart from the principal elements other elements including Cu, Ti, Zn, Mn, and Cr were also detected (the remainder). A relatively high amount o oxygen (2.53 at%) suggests that the alloy was partially oxidized in air. The influence of various elements on the process as well as on the resulted anodic alumina was analyzed previously [33-36]. According to those analysis, the amount and distribution of insoluble impurities will strongly influence the quality of anodic oxide layers formed on the aluminum substrate under specific conditions. Moreover, the presence of various additives can affect both average ions migration rates through the barrier oxides and field-assisted dissolution of oxide. In the studied cases it is very probable that the impurities in the AA6063 alloy hindered the ion migration through the barrier layer resulting in the lower recorded steady-state current densities as compared to those registered for high-purity aluminum [29]. Furthermore, in contrast to the HA of high-purity aluminum, there is a negligible difference between the ia values reached at the Ut between the samples anodized in water-based electrolyte (Fig. 1a) and in electrolyte with 4:1 v/v water to ethanol mixture (Fig. 1b). The influence of ethanol is manifested, however, in ia decaying process: in the case of anodization in water-based electrolyte the current decrease is more gentle than in the case of HA in electrolyte with 4:1 v/v water to ethanol mixture. The current density drops below 10 mA/cm2 after around 3500 s for the samples anodized in ethanol-free electrolyte (Fig. 1a), whereas in the ethanol-modified electrolyte the current exceeds the 10 mA/cm2 value well before 2000 s in the case of anodization at 120, 140, and 160 V, and after ~ 2250 s for the sample anodized at 180 V (Fig. 1b). A steady-state anodization in ethanol-free

AA6063 120 V 140 V 160 V 180 V Al [at%] 95.88 47.25 45.29 48.23 46.53 O [at%] 2.53 51.42 53.63 50.64 52.21

Mg [at%] 0.59 0.57 0.51 0.53 0.41 Si [at%] 0.26 0.24 0.12 0.12 0.14 Fe [at%] 0.22 0.06 0.05 0.04 0.06

remainder [at%] 0.52 0.46 0.40 0.44 0.65

electrolyte is carried out at the ia of 6-8 mA/cm2, whereas in ethanol-modified solution the ia attained at the stable-voltage anodization stage is around 3-5 mA/cm2. Accordingly, the average current density is higher in the case of samples anodized in water-based electrolyte as compared to the ones anodized in ethanol-modified acid solution (Fig. 2).

Fig. 2. Average current density for the analysed samples as a function of a target voltageRys. 2. Średnia gęstość prądu dla analizowanych próbek w funkcji przy-łożonego napięcia

for guided anodization where three stages can be distinguished: very high current density at the beginning of the process, its sub-sequent drop within 1–2 seconds, and finally its slower exponential decrease to a steady-state value within about 1 second. The three regimes can be attributed to the following phenomena: an electro-lytic reaction of water at the beginning, very rapid barrier layer for-mation and, at the end, a stable pores growth. In the guided MA the stage related with the competition among pores, which in the current-time curve is manifested in a moderate current increase pre-ceding the steady-state anodization [37, 38], is missing. This is due to the existing periodic pattern of nanoconcaves after the first ano-dization step at the aluminium surface that defines the pore nuclea-tion sites and eliminates the pore competition. The growth position of the pores is limited by the pre-patterned design and is restricted to the inner corner/electrolyte interface, where a maximum dissolu-tion rate of Al3+ is localized [39]. The time needed to proceed to the steady-state anodization is very short in the MA of AA6063 al-loy. A similar process took about 2 minutes during the MA of high-purity aluminium [29]. This may indicate that the dissolution rate of Al3+ in the inner corner/electrolyte interface is much faster in the low-purity aluminium, which again underlines a different mecha-nism operating in anodization of the two types of aluminium.

The SEM images of resulting PAAs are presented in Figure 5. The pores in the PAAs are characterized by irregular shape with


obtained after a two-step HA-MA process performed on high-purity aluminium in ethanol-free electrolyte (Dc in the range of 250 and 300 nm) [29]. In the anodization of high-purity aluminium it was suggested that ethanol, in addition to a local cooling effect, most probably has some impact on stress relaxation which arises from the volume expansion associated with the oxidation reaction at the metal-oxide interface during the initial stages of PAA film growth [40, 41]. The volume expansion during PAA film formation at the aluminium/oxide interface is associated with a repulsive interaction force (mechanical stress) and increases with the current density. This mechanical stress has to remain within a certain range of ac-ceptable values, in order not to deteriorate the pore arrangement [42]. At the same time the interpore distance increases. Although the effect of ethanol is clearly visible in the lower average current densities (Fig. 2), the comparable interpore distances in the sam-ples anodizes with and without ethanol suggest that the alloying elements in AA6063 foil have a much stronger influence on the re-laxation process than ethanol.

The ζ parameter, defined as ζ = Dc/U, is in the range between 2.1÷2.3 characteristic for HA process (Fig. 7a). The ζ values of around 2.0 nm/V were also determined previously in various HA processes [19, 22, 23, 43]. The reduced ζ for HA process was ex-plained by a reduced voltage dependence of pore diameter and the barrier layer thickness of PAA under high electric filed at the pore bottom. In HA processes a non-linear dependence of interpore dis-tance on applied voltage was also observed [44]. It was found that at high anodizing potentials the Dc is not solely determined by poten-tial Ut, but is also influenced by the current density ia (high electric field). The corresponding equation can be written as follows [ 23, 24, 44]:

DU

A B iC

c

t

a= + ⎛⎝⎜

⎞⎠⎟

exp

(2)

(A, B and C are constants which are dependent on the anodization conditions). Upon increasing of Ut, Dc increases with the Ut, as demonstrated in Figure 7a. On the other hand, the ζ parameter will decrease with the increase of the ia. Accordingly, it can be seen in Figure 7a that while Dc increases with Ut, ζ decreases with in-creasing Ut (and with increasing average current density, Fig. 2), in agreement with Eq. (2). However, according to Eq. (2) when Ut

The lower currents registered during anodization in the presence of ethanol were ascribed to the ethanol cooling effect [24, 28, 32]. The cooling effect of ethanol was related with ethanol ability to lower the dissociation constant of the electrolyte and to decrease of temperature generated at the PAA barrier layer. These effects, in turn, lead to the reduction of the ion diffusion ability and smaller amount of acid anions in the barrier layer, and consequently to a relatively smaller contribution from the ionic current.

As formed anodic alumina was chemically removed and SEM images of the resulted nano-concave arrays formed on aluminum are presented in Fig. 3.

Fig. 3. SEM micrographs of Al nanoconcave arrays obtained after a removal of the formed oxides in HA carried out in ethanol-free (a÷d) and ethanol-modified (e÷h) electrolyte at 120 (a, e), 140 (b, f), 160 (c, g), and 180 V (d, h); the 1 μm bar is for all imagesRys. 3. Mikrografie SEM nanowgłębień Al otrzymanych w wyniku anodowania twardego HA w czystym elektrolicie (a÷d) i w elektrolicie modyfikowa-nym etanolem (e÷h) odpowiednio dla napięć HA 120 (a, e), 140 (b, f), 160 (c, g), and 180 V (d, h); wskaźnik 1 μm dotyczy wszystkich mikrografii SEM

Fig. 4. Current density vs time transients recorded during the second mild anodization for the samples anodized on base of the Al nanocon-caves prepared in HA process conducted in ethanol-free electrolyteRys. 4. Krzywe gęstości prądu w zależności od czasu anodowania mięk-kiego MA przeprowadzonego w czystym elektrolicie

rather low circularity. Apart from a population of relatively larger pores, tiny pores are occasionally observed (the arrows in Fig. 5). With the voltage increase both pore size distribution and the pores regularity are improving. Moreover, a number of the small pores seem to decrease.

The FFT analysis along with the radial average of the respective SEM images are presented in Figure 6. As can be seen there is no difference in the FFT rings appearance between the samples ano-dized on base of Al nanoconcavities prepared during HA in aque-ous (Fig. 6a÷d), and ethanol-modified electrolytes (Fig. 6e÷h). The points in the corners of hexagons become, however, more distinct as the voltage increases. The interpore distances determined from the radial average of the respective FFT images for both sets of samples are gathered in Figure 7a. As expected the Dc increases with the ap-plied voltage. On the other hand, there is a negligible influence of ethanol on Dc. This situation is basically different from the situation encountered during anodization processes of high-purity alumini-um, where above 140 V the Dc has become considerably larger for the samples anodized in ethanol-modified electrolyte [28]. Moreo-ver, the Dc values between 270 and 370 nm are larger than those


Fig. 5. SEM micrographs of PAAs (top view) fabricated on base of the Al nanoconcave obtained in HA process conducted in ethanol-free (a÷d) and ethanol-modulated (e÷h) electrolyte at 120 (a, e), 140 (b, f), 160 (c, g), and 180 V (d, h); the bar 1 μm is for all imagesRys. 5. Mikrografie SEM anodowego tlenku aluminium (widok z góry) wytworzonego na bazie nanowgłębień Al otrzymanych w wyniku anodowania twardego HA w czystym elektrolicie (a÷d) i elektrolicie modyfikowanym etanolem (e÷h) odpowiednio dla napięć MA 120 (a, e), 140 (b, f), 160 (c, g), and 180 V (d, h); wskaźnik 1 μm dotyczy wszystkich mikrografii SEM

Fig. 6. Results of Fast Fourier Transforms, FFTs along with FFTs intensity profiles (below a given FFT image) of the respective SEM images of PAAs presented in Figure 5Rys. 6. Wyniki szybkiej transformaty Fouriera FFTs wraz z profilami intensywności (poniżej odpowiednich obrazów FFT) wygenerowane na podsta-wie odpowiednich mikrografii SEM anodowego tlenku aluminium przedstawionych na rysunku 5


is kept constant, ia is a key parameter governing Dc. This means that under constant Ut, Dc should decrease with increasing current density. Despite higher average current densities in acid-only elec-trolyte as compared to ethanol-modified one (Fig. 2), the Dc values are comparable. This behaviour may be linked with similar ia values reached at Ut for both sets of samples (regardless of the voltage value) and again may indicate that the stress relaxation process and the factors influencing the relaxation can be essential in elucida-tion of this phenomena. Therefore, analysis of Eq. (2) supports the assumption that the insoluble impurities in AA6063 contribute to mechanical stress relaxation at the metal/oxide interface during PAA growth, and partly replace ethanol in stress relaxation in the samples anodized in acid-only electrolyte.

The degree of pore ordering was estimated based on Eq. (1). The analysis is summarized in Figure 7b. As in the case of Dc, the RR are comparable for both sets of samples, although it seems that the parameter becomes a little bit larger at higher voltages for samples anodized in ethanol-modified electrolyte, in contrast to what was found in HA-MA anodization of high-purity aluminium [29]. More-over, the RR parameter increases around two times with the applied voltage in both cases. In general, the RR values are lower than the corresponding values determined for PAAs formed on high-purity aluminium [29]. This signifies that the impurities (the alloying el-ements) in the AA6063 alloy influence not only to the interpore distance but play also a significant role in degrading of pore order-ing. Recently, a new method for characterizing the quality of less ordered PAA patterns in terms of uniformity, was developed [45].

In Figure 8 the PAAs growth rate as a function of the applied voltage along with the respective SEM images of PAA cross-section are presented. The PAA growth rate increases with voltage as the ion migration rate through the barrier layer becomes faster under the higher electric field. However, in reference to the PAA growth rate on pure aluminium [29], the rate of the anodic oxide growth on AA6063 aluminium is lower. The lower rate is presumably caused by the presence of various amount of insoluble alloying elements in aluminium, which can block the entrance of electrolyte ions to the barrier layer and thus slow down the oxidation process. The EDS analysis of the PAA cross-sections suggests that quite a portion of alloying elements transfer to the oxides (Tab. 1). Particularly Mg is the element which seems to be easily incorporated in the PAA framework. Also Si traps quite easily in the PAA walls. On the other hand, the amount of Fe in the PAAs is much lower than in AA6063 sheet implying that this element generally does not incorporate to the PAA matrix. The inclusion of Mg and Si into the PAA matrix

Fig. 7. The evolution of interpore distance Dc and ζ parameter as a function of MA voltage for PAAs fabricated on base of Al nanoconcavities prepared in HA process conducted in ethanol-free and ethanol-modified electrolyte (a), and the relation of RR parameter with the applied voltage for both groups of PAAs (b)Rys. 7. Zależność odległości między nanoporami Dc i parametru ζ (a)oraz parametru regularności RR (b) od przyłożonego napięcia MA dla anodowe-go tlenku aluminium wytworzonego na bazie nanowgłębień Al otrzymanych w wyniku anodowania twardego HA w czystym elektrolicie i elektrolicie modyfikowanym etanolem

were previously observed and were associated with a formation a non-uniform layer thickness and a scalloped substrate/oxide inter-face [34÷36]. The features are also observed in the studied samples. Moreover, it was shown that the elements, such as Si, are also an-odized during MA, forming Si particles surrounded by Si oxide. Owing to significantly reduced anodization of Si alumina invaded beneath the Si particle and eventually occluded the partially anod-ized Si particle in the film [36]. This gave rise to a non-uniform growth of the aluminium oxide pores.

The effect can be also seen in Figure 9 which presents the SEM images of the top and bottom parts of the resulted PAAs. Although generally the pores are well ordered throughout the whole PAAs thickness, occasionally the parallel arrangement of pores is dis-turbed by the presence of entities originating from the alloying elements. Furthermore, the long-range pore ordering is improving with the voltage increase. The voltage of 180 V is close to the self-ordering phenomenon occurring in 0.1 M phosphoric acid solution

Fig. 8. PAAs growth rate vs voltage determined for the PAAs fabricat-ed on base of Al nanoconcavities prepared in HA process carried out in ethanol-free electrolyte along with the respective PAA cross-sectionsRys. 8. Szybkość wzrostu matryc z anodowego tlenku aluminium w funk-cji przyłożonego napięcia wytworzonych na bazie nanowgłębień Al otrzy-manych w wyniku anodowania twardego w czystym elektrolicie


and the proximity to this condition in cooperation with the guidance of periodic concavities in the aluminium surface aided the paral-lel pores growth throughout the entire layer. Therefore, it can be concluded that a formation of a long-range ordered porous alumina on low purity AA6063 aluminium alloy is possible at high elec-tric fields.

4. CONCLUSIONS

Two-step process combining HA and MA anodization of AA6063 aluminium alloy at voltages 120÷180 V in water-based oxalic acid

solution and in electrolyte with 4:1 v/v water to ethanol mixture resulted in fabrication of porous anodic alumina with interpore distance between 270 and 370 nm, respectively. The presence of impurities influenced the current density vs time curves as well as the hexagonal pore ordering. The analysis demonstrated that the al-loying elements in the AA6063 play an important role in the relaxa-tion of mechanical stresses at the metal/oxide interface occurring during the PAA’s growth. The effect contributed to comparable Dc values in the samples anodized in ethanol-free and ethanol-mod-ified electrolyte despite much larger average current densities in the case of the former samples. Moreover, the impurities lowered

Fig. 9. The upper and bottom parts of the respective PAA cross-sections fabricated at 120, 140, 160, and 180 VRys. 9. Mikrografie SEM przekrojów górnej i dolnej części matryc z anodowego tlenku aluminium wytworzonych w napięciach MA odpowiednio 120, 140, 160 i 180 V


the hexagonal pore arrangement as compared to the arrangement in the PAAs prepared on high-purity aluminium. The pore order-ing was improved under application of higher anodization voltages. The porous anodic alumina obtained under anodization of low-cost AA6063 alloy can be applied for instance in various filtration sys-tems or in a large scale nanofibers production applicable particu-larly in the fields where the interpore distance corresponding to the spectral UV/blue range is important.

ACKNOWLEDGMENTS

The research was financed by National Science Centre (Deci-sion number: DEC-2012/07/D/ST8/02718). The work has been fi-nancially supported by the Polish Ministry of Science and Higher Education, Project: LAPROMAW (POIG.02.01.00-14-071/08/00).

REFERENCES

[1] Jani A. M. Md, Losic D., Voelcker N. H.: Nanoporous anodic aluminium oxide: Advances in surface engineering and emerging applications. Prog-ress in Materials Science 58 (2013) 636÷704.

[2] Lee W., Park S.-J.: Porous anodic aluminium oxide: anodization and templated synthesis of functional nanostructures. Chem. Rev. 114 (2014) 7487÷7556.

[3] Chen X., Yu D., Cao L., Zhu X., Song Y., Huang H., Lu L., Chen X.: Fabrication of ordered porous anodic alumina with ultra-large interpore distances using ultrahigh voltages. Mater. Res. Bull. 54 (2014) 116÷120.

[4] Chi C.-S., Lee J.-H., Kim I., Oh H.-J.: Effects of microstructure of alu-minium substrate on ordered nanopore arrays in anodic alumina. J. Mater. Sci. & Technol. 31 (2015) 751÷758.

[5] Aerts T., Dimogerontakis Th., Graeve I. De, Fransaer J., Terryn H.: In-fluence of the anodizing temperature on the porosity and the mechanical properties of the porous anodic oxide film. Surface & Coatings Technol-ogy 201 (2007) 7310÷7317.

[6] Fernández-Romero L., Montero-Moreno J. M., Pellicer E., Peiró F., Cor-net A., Morante J. R., Sarret M., Müller C.: Assessment of the thermal stability of anodic alumina membranes at high temperatures. Materials Chemistry and Physics 111 (2008) 542÷547.

[7] Montero-Moreno J. M., Belenguer M., Sarret M., Müller C. M.: Produc-tion of alumina templates suitable for electrodeposition of nanostructures using stepped techniques. Electrochimica Acta 54 (2009) 2529÷2535.

[8] Shih T.-S., Wei P.-S., Huang Y.-S.: Optical properties of anodic aluminium oxide films on Al1050 alloys. Surface & Coatings Technology 202 (2008) 3298÷3305.

[9] Runge M. R., Hossain T.: Interfacial phenomena in 7000 series alloys and their impact on the anodic oxide. Materials Today: Proceedings 2 (2015) 5055÷5062.

[10] Yu C.-U., Hu C.-C., Bai A., Yang Y.-F.: Pore-size dependence of AAO films on surface roughness of Al-1050 sheets controlled by electropolish-ing coupled with fractional factorial design. Surface & Coatings Technol-ogy 201 (2007) 7259÷7265.

[11] Bai A., Hub C.-C., Yang Y.-F., Lin C.-C.: Pore diameter control of anodic aluminium oxide with ordered array of nanopores. Electrochimica Acta 53 (2008) 2258÷2264.

[12] Stępniowski W. J., Michalska-Domańska M., Norek M., Twardosz E., Florkiewicz W., Polkowski W., Zasada D., Bojar Z.: Anodization of cold deformed technical purity aluminium (AA1050) in oxalic acid. Surface & Coatings Technology 258 (2014) 268÷274.

[13] Montero-Moreno J. M., Sarret M., Müller C.: Influence of the aluminium surface on the final results of a two-step anodizing. Surface & Coatings Technology 201 (2007) 6352÷6357.

[14] Schneider M., Kremmer K.: The effect of bath aging on the microstructure of anodic oxide layers on AA1050. Surface & Coatings Technology 246 (2014) 64÷70.

[15] Zaraska L., Sulka G. D., Szeremeta J., Jaskuła M.: Porous anodic alumina formed by anodization of aluminium alloy (AA1050) and high purity alu-minium. Electrochimica Acta 55 (2010) 4377÷4386.

[16] Michalska-Domańska M., Norek M., Stępniowski W. J., Budner B.: Fab-rication of high quality anodic aluminium oxide (AAO) on low purity alu-minium — A comparative study with the AAO produced on high purity aluminium. Electrochimica Acta 105 (2013) 424÷432.

[17] Ono S., Saito M., Asoh H.: Self-ordering of anodic porous alumina indu-ced by local current concentration: burning, Electrochem. Solid-State Lett. 7 (2004) B21÷B24.

[18] Ono S., Saito M., Ishiguro M., Asoh H.: Controlling factor of self-ordering of anodic porous alumina. J. Electrochem. Soc. 151 (2004) B473÷B478.

[19] Lee W., Ji R., Gösele U., Nielsch K.: Fast fabrication of long-range or-dered porous alumina membranes by hard anodization. Nature 5 (2006) 741÷747.

[20] Li Y. B., Zheng M. J., Ma L.: High-speed growth and photoluminescence of porous anodic alumina films with controllable interpore distances over a large range. Appl. Phys. Lett. 91 (2007) 073109.

[21] Cheng C., Ngan A. H. W.: Fast fabrication of self-ordered anodic porous alumina on oriented aluminium grains by high acid concentration and high temperature anodization. Nanotechnology 24 (2013) 215602.

[22] Schwirn K., Lee W., Hillebrand R., Steinhart M., Nielsch K., Gösele U.: Self-ordered anodic aluminium oxide formed by H2SO4 hard anodization. ACS Nano 2 (2008) 302÷310.

[23] Li Y., Ling Z. Y., Chen S. S., Wang J. C.: Fabrication of novel porous an-odic alumina membranes by two-step hard anodization. Nanotechnology 19 (2008) 225604.

[24] Li Y., Zheng M., Ma L., Shen W.: Fabrication of highly ordered nanopo-rous alumina films by stable high-field anodization. Nanotechnology 17 (2006) 5101÷5105.

[25] Wang B., Fei G. T., Wang M., Kong M. G., Zhang L. D.: Preparation of photonic crystals made of air pores in anodic alumina. Nanotechnology 18 (2007) 365601.

[26] Shang G. L., Fei G. T., Zhang Y., Yan P., Xu S. H., Ouyang H. M., Zhang L. D.: Fano resonance in anodic aluminium oxide based photonic crystals. Scientific Reports 4 (2014) 3601.

[27] Tsao Y.-C., Fisker C., Pedersen T. G.: Nanoimprinted backside reflectors for a-Si:H thin-film solar cells: Critical role of absorber front textures. Optics Express 22 (2014) A651÷A662.

[28] Han X. Y., Shen W. Z.: Improved two-step anodization technique for or-dered porous anodic aluminium membranes. J. Electroanalytical Chem. 655 (2011) 56÷64.

[29] Norek M., Dopierała M., Stępniowski W. J.: Ethanol influence on arrange-ment and geometrical parameters of aluminium concaves prepared in a modified hard anodization for fabrication of highly ordered nanoporous alumina. J. Electroanal. Chem. 750 (2015) 79÷88.

[30] Horcas I., Fernandez R., Gomez-Rodriguez J. M., Colchero J., Gomez-Herrero J., Baro A. M.: WSXM: a software for scanning probe microscopy and a tool for nanotechnology. Rev. Sci. Instrum. 78 (2007) 013705.

[31] Stępniowski W. J., Nowak-Stępniowska A., Presz A., Czujko T., Varin R. A.: The effects of time and temperature on the arrangement of anodic alu-minium oxide nanopores. Mater. Charact. 91 (2014) 1÷9.

[32] Yi L., Zhiyuan L., Xing H., Yisen L., Yi C.: Investigation of intrinsic mechanisms of aluminium anodization processes by analyzing the current density. RSC Advances 2 (2012) 5164÷5171.

[33] Habazaki H., Shimitzu K., Skeldon P., Thompson G. E., Wood G. C.: Formation of amorphous anodic oxide films of controlled composition on aluminium alloys. Thin Solid Films 300 (1997) 131÷137.

[34] Fratila-Apachitei L. E., Tichelaar F. D., Thompson G. E., Terryn H., Skel-don P., Duszczyk J., Katgerman L.: A transmission electron microscopy study of hard anodic oxide layers on AlSi(Cu) alloys. Electrochim. Acta 49 (2004) 3169÷3177.

[35] Fratila-Apachitei L. E., Terryn H., Skeldon P., Thompson G. E., Duszczyk J., Katgerman L.: Influence of substrate microstructure on the growth of anodic oxide layers. Electrochim. Acta 49 (2004) 1127÷1140.

[36] Fratila-Apachitei L. E., Duszczyk J., Katgerman L.: Voltage transients and morphology of AlSi(Cu) anodic oxide layers formed in H2SO4 at low tem-perature. Surf. Coat. Technol. 157 (2002) 80÷94.

[37] Parkhutik W. P., Shershulsky V. I.: Theoretical modelling of porous oxide growth on aluminium. J. Phys. D: Appl. Phys. 25 (1992) 1258÷1263.

[38] Li F. Y., Zhang L., Metzger R. M.: On the growth of highly ordered pores in anodized aluminium oxide. Chem. Mater. 10 (1998) 2470÷2480.

[39] Chen S. H., Chan D.-S., Chen C.-K., Chang T.-H., Lai Y.-H., Lee C.-C.: Nanoimprinting pre-patterned effects on anodic aluminium oxide. Jpn. J. Appl. Phys. 49 (2010) 015201.

[40] Singh G. K., Golovin A. A., Aranson I. S.: Formation of self-organized nanoscale porous structures in anodic aluminium oxide. Phys. Rev. B 73 (2006) 205422.

[41] Chu S. Z., Wada K., Inoue S., Isogai M., Yasumori A.: Fabrication of ide-ally ordered nanoporous alumina films and integrated alumina nanotubule arrays by high-field anodization. Adv. Mater. 17 (2005) 2115÷2119.

[42] Nielsch K., Choi J., Schwirn K., Wehrspohn R. B., Gosele U.: Self-ordering regimes of porous alumina: the 10 porosity rule., Nano Lett. 2 (2002) 677÷680.

[43] Lee W., Nielsch W., K., Gösele U.: Self-ordering behavior of nanoporous anodic aluminium oxide (AAO) in malonic acid anodization. Nanotech-nology 18 (2007) 475713.

[44] Yi L., Zhiyuan L., Xing H., Yisen L., Yi C.: Formation and microstructures of unique nanoporous AAO films fabricated by high voltage anodization. J. Mater. Chem. 21 (2011) 9661÷9666.

[45] Toccafondi C., Stępniowski W. J., Leoncini M., Salerno M.: Advanced arrangement analysis of anodic alumina formed by short anodization in oxalic acid. Mater. Charact. 78 (2013) 79÷86.


Porowaty anodowy tlenek aluminium wytworzony na stopie aluminium AA6063 w dwuetapowym

procesie łączącym twarde i miękkie anodowanieMałgorzata Norek

Katedra Zaawansowanych Materiałów i Technologii, Wydział Nowych Technologii i Chemii, Wojskowa Akademia Techniczna, ul. Kaliskiego 2, 00-908 Warszawa, Polska, [email protected]

Inżynieria Materiałowa 4 (212) (2016) 163÷171DOI 10.15199/28.2016.4.3© Copyright SIGMA-NOT MATERIALS ENGINEERING

Słowa kluczowe: anodowanie twarde, aluminium techniczne, anodowanie sterowane, etanol, porowaty anodowy tlenek aluminium (ATA).

1. CEL PRACYPorowaty anodowy tlenek aluminium (ATA) jest jedną z najczęściej używanych matryc do wytwarzania szerokiej gamy nanostruktur z różnych materiałów znajdujących zastosowania w optoelektro-nice, plazmonice czy medycynie. Do produkcji wysokiej jakości nanomateriałów jest wymagana matryca ATA charakteryzująca się dużym stopniem heksagonalnego uporządkowania nanoporów oraz ich równoległym ułożeniem wzdłuż całego przekroju. Ano-dowanie twarde (wysokoprądowe) jest procesem, który skutkuje wytworzeniem matryc ATA o takich właśnie cechach. Ponadto pro-ces ten znacznie skraca czas wytwarzania matrycy oraz powoduje rozszerzenie zakresu możliwych do osiągnięcia parametrów geo-metrycznych ATA, tj. odległości między porami. Zwykle najlepszej jakości matryce uzyskuje się przez anodowanie aluminium o wy-sokiej czystości (5N). Wysokie koszty ultraczystego Al powodu-ją, że poszukuje się materiałów alternatywnych do produkcji ATA, takich jak stopy aluminium. W pracy przeprowadzono dwuetapowy proces anodowania łączący anodowanie twarde (HA) w pierwszym etapie i anodowanie miękkie (MA) w drugim na stopie aluminium AA6063. Łączone procesy HA i MA umożliwiają uzyskanie hek-sagonalnie ułożonych, równoległych porów w całym przekroju matrycy ATA. Przeprowadzono kompleksową analizę matryc ATA wytworzonych w różnych napięciach z zakresu 120÷180 V. Zbada-no wpływ etanolu na proces anodowania i budowę matrycy przez porównanie krzywych prądowych zarejestrowanych w trakcie ano-dowania z parametrami geometrycznymi oraz stopniem uporządko-wania nanoporów obliczonych metodą FFT.

2. MATERIAŁ I METODYKA BADAŃ

Blacha z aluminium AA6063 (próbki o wymiarze 2 cm × 1 cm) zosta-ła wyelektropolerowana w mieszaninie 1:4 60% HClO4 oraz etanolu w 0°C, przy stałym napięciu równym 20 V, przez 2 min. Następnie tak przygotowane próbki poddano anodowaniu w 0,3 M roztworze kwasu szczawiowego, w temperaturze 0°C, przy napięciu z zakresu 120÷180 V. Anodowane były dwie serie próbek: I seria była anodo-wana w czystym roztworze wodnym, II seria w roztworze zawiera-jącym 1:4 v/v wody do etanolu. Uzyskany tlenek wytworzony w I etapie (HA) usunięto w mieszaninie 6% mas. kwasu fosforowego i 1,8% mas. kwasu chromowego w 60°C przez 3 h. Następnie próbki zostały poddane II etapowi anodowania (MA) w 0,1 M roztworze kwasu fosforowego, w temperaturze 0°C i przy tych samych napię-ciach, co zastosowane w I etapie. Na podstawie mikrografii SEM warstwy powierzchniowej analizowanych próbek wygenerowano obrazy FFT. Z wygenerowanych pierścieni FTT obliczono odległo-ści między porami Dc oraz parametr uporządkowania nanoporów RR.

3. WYNIKI I ICH DYSKUSJAUzyskane wyniki pokazały, że matryce ATA wytworzone na alu-minium AA6063 charakteryzują się gorszym stopniem uporząd-kowania nanoporów w porównaniu z ATA wytworzonym na czy-stym Al. Pierwiastki stopowe obecne w AA6063 powodują per-turbacje w uporządkowanym wzroście matrycy ATA. Dodatkowo zaobserwowano, że obecność pierwiastków stopowych przyczy-niła się do znacznego obniżenia gęstości prądowej w odniesie-niu do gęstości prądowej zarejestrowanej podczas anodowania czystego Al, co prawdopodobnie było spowodowane spowolnie-niem dyfuzji jonów w warstwie barierowej przez obecne tam za-nieczyszczenia. Nie zauważono znaczącego wpływu etanolu na proces anodowania, jak i na parametry geometryczne matrycy ATA w odróżnieniu do wyraźnego wpływu tego modyfikatora na proces anodowania czystego aluminium. Ponieważ mechanizm porządkowania porów matrycy jest ściśle związany z pojawie-niem się naprężeń ściskających na granicy metal/tlenek podczas wzrostu tlenku (duża różnica stałych sieciowych pomiędzy Al i Al2O3), efekt ten wytłumaczono silnym wpływem pierwiastków stopowych obecnych w AA60633 na proces relaksacji tych naprę-żeń. W efekcie nie wystąpiły odpowiednie warunki mechaniczne do formowania się zwartej, heksagonalnie uporządkowanej ma-trycy ATA. Stopień heksagonalnego i wzdłużnego uporządkowa-nia porów polepszył się wraz ze wzrostem przyłożonego napięcia anodowania.

4. PODSUMOWANIE

Stop aluminium AA6063 poddano dwuetapowemu procesowi anodowania łączącemu HA w 0,3 M roztworze kwasu szczawio-wego oraz MA w 0,1 M roztworze kwasu fosforowego w napię-ciach z zakresu 120÷180 V. Badano wpływ etanolu na proces oraz parametry geometryczne uzyskanych matryc ATA. Analiza zarówno krzywych prądowych w funkcji czasu, jak i parametrów geometrycznych matryc ATA pozwoliła na sformułowanie wnio-sku, że pierwiastki stopowe obecne w AA6063 w znaczący sposób przyczyniają się do relaksacji naprężeń mechanicznych genero-wanych podczas wzrostu ATA na granicy metal/tlenek. W efekcie ATA charakteryzowały się podobnymi wartościami parametru Dc (270÷370 nm) oraz podobnym stopniem uporządkowania nanopo-rów zarówno dla serii próbek poddanych HA w czystym elektro-licie, jak i modyfikowanym etanolem, pomimo większej średniej gęstości prądu zarejestrowanej dla pierwszej serii. Stopień upo-rządkowania porów wzrastał wraz z przyłożonym napięciem ano-dowania.

porous anodic alumina formed on aa6063 aluminum alloy in a ... text.pdf · under mild conditions...

Documents