- Email: [email protected]
Characterization of copper in phosphoric-acid-anodized 2024-T3 aluminum by Auger electron spectroscopy and Rutherford backscattering
Thin Solid Films, 84(1981) 155--160
155
PREPARATION AND CHARACTERIZATION
C H A R A C T E R I Z A T I O N O F C O P P E R IN P H O S P H O R I C - A C I D A N O D I Z E D 2024-T3 A L U M I N U M BY A U G E R E L E C T R O N SPECTROSCOPY AND R U T H E R F O R D BACKSCATTERING* J. S. SOLOMON
University o f Dayton Research Institute, Dayton, OH 45469 (U.S.A.) N. T. McDEVITT Materials Laboratory, Air Force Wright Aeronautical Laboratories, Wright-Patterson Air Force Base, OH 45433 (U.S.A.) (Received March 23, 1981 ; accepted April 8, 1981)
The effects of the electrochemical anodization of deoxidized 2024-T3 aluminum on copper were characterized by Auger electron spectroscopy and Rutherford backscattering. Anodization was performed in phosphoric acid at a constant potential. Data are presented which show that constant potential anodization of 2024-T3 is more efficient than that of aluminum in terms of oxide growth rates for short anodization times. However, the maximum anodic oxide thickness achievable on thc alloy is less than that achievable on the pure metal. Copper is shown to be enriched at the oxide-metal interface because of its diffusion from the bulk during anodization. Possible effects of copper on oxide growth mechanisms and oxide morphology are discussed.
1. INTRODUCTION
When aluminum is anodized in an aggressive electrolyte such as H3PO 4, the resulting anodic oxide has a porous structure with well-defined columns I-3. The oxide growth is limited by both the increasing current-retarding effects of the growing oxide film and the dissolving action of the acid electrolyte. The maximum oxide thickness is reached when the dissolution and growth rates become equal. The anodic oxide growth mechanism for 2024-T3 is essentially the same as that for aluminum. However, the morphology of the oxide layer may differ from that of the pure metal, as determined by scanning electron microscopy 4. This may be due in part to thc presence of intermetallic compounds, such as CuAI 2, and of microconstitui~nts, such as AI-Cu-Mg. Since these inclusions have different oxidation or dissolution rates from those of the AI-Cu solid solution, discontinuities within the anodic oxide layer can result 5. Furthermore, the presence of copper in aluminum has been shown to have considcrable effects on oxide growth rates 4-6. In this paper we discuss the effects of anodiza'don on the distribution of copper within the oxide and the oxide-metal interface of 2024-T3 aluminum and its possible influence on anodic oxide growth and morphology. The anodized pieces * Paper presented at the International Conference on Metallurgical Coatings, San Francisco, CA, U.S.A., April 6- I0, 1981. 0040-6090/81/0000-0000/$02.50
.4.3 Elsevier Sequoia/Printed in The Netherlands
156
J.S. SOI.OMON, N. T. Mcl)EVITT
were characterized by Auger electron spectroscopy (AES), Auger sputter profile analysis (ASPA) and Rutherford backscattering (RBS). 2. EXPERIMENTAI. DETAILS
2.1. Specimen preparation Circular pieces of 2024-T3 aluminum approximately 5 cm in diameter and 0.2 cm thick were first degreased with acetone. One set was then deoxidized in Oakite (obtained from Oakite Products of Canada Ltd., Bramlea, Ontario, Canada) for 15 min and rinsed in distilled H20. The deoxidized pieces were anodized at a constant potential of 10 V in a room temperature constant circulation bath of 1.0 M H3PO 4 for times of 1, 3, 5, 10, 20, 60 and 120 min. After anodization the specimens were rinsed with distilled H 2 0 and hot air dried. 2.2. Method ofanalysi~ 2.2.1. Auger electron spectroscopy The Auger instrumentation and data-handling scheme used in this work have been described previously v. A coaxial electron gun which was operated at 4 keV with a current density of 0.10 A c m - 2 was used to provide high energy excitation of the specimen. The ion beam used for sputter profiling in this work was generated with a Physical Electronics Industries model 94-191 sputter ion gun which was operated with a beam energy of 2 keV and an ion current density of approximately 3 ~A m m - 2 in a partial pressurc of argon of 6.6 x 10- ~ Pa. 2.2.2. Ruther[brd backscattering RBS was performed using a 2 MeV 4He" ion beam produced with a Van de Graaff generator. The backscattered ions were collected by a solid state detector at an angle 0 of 160 measured against the forward direction of the incident beam. The energy resolution of the detector was such that the value of dE/dx (in keV p.g cm - ~) for '*He ' ions was subject to an error of -+ 5'~';ior a depth resolution of 20 nm. Backscattering factors for copper and aluminum were calculated assuming the 2024T3 composition to be Alo ~sCuo 02 . 3. RESULTS Figure 1 contains anodization time versus oxide thickness curves determined from ASPA data for aluminum and 2024-T3 anodized in 1.0 M H 3 P O 4 at a constant potential of 10 V. The oxide on 2024-T3 has a limiting thickness of approximately 400 nm while growth continues on the pure metal (the maximum thickness on the metal was approximately 1000 nm). The copper in-depth profiles from Oakite-deoxidized alloy specimens which were anodized at 10 V for 1, 3, 5 and 10 min are shown in Fig. 2. These profiles all reflect an increase in copper at the oxide metal interface with the profiles from the specimens anodized for 1, 3 and 5 min showing copper enrichment at the interface. The profile from the specimen anodized for 10 min no longer shows a copper enrichment at the interface but a concentration level equal to that in the bulk. The depth distributions of the copper-enriched zones in the specimens anodized for 1.3
CHARACTERIZATION OF C H IN
H3PO 4- ANODIZED 2 0 2 4 - T 3 A1
157
and 5 min, based on the sputter rate for aluminum metal, are 34, 59 and 84 nm respectively. The data in Fig. 2 suggest that copper enrichment can be attributed solely to the chemical deoxidation treatment and that subsequent H3PO, t anodization depletes copper from the oxide-metal interface. However, since the copper profile from a non-etched and anodized specimen also showed copper enrichment at the oxide metal interface, enrichment can be attributed to the anodization process itself. Copper was not detccted within the anodic oxide layer, regardless of thickness or pre-anodization treatment. '~ 800
¢.
Ill
// < 0
o
5
~
~
~o
'
2~
6
ANOOIZAT1ON TIME(MIN)
,'o
~'o
3b
4"0
SPUTTER TIME (MiN)
Fig. 1. Oxide film thicknesses on aluminum ( - - ) and 2024-T3 (-- ) determined from ASPA and plotted as functions of the anodization time at 1O V in room temperature 1.0 M H3PO 4. Fig. 2. Auger sputter protiles of copper in anodic oxide films on Oakite-etched 2024-T3 anodized for 1.3. 5 and 10 min at 10 V in room temperature 1.0 M H3PO4.
Figure 3 contains RBS spectra in the region 1.2-1.8 MeV for specimens etched and anodized for times up to 2 h. Contrary to the ASPA data, copper interracial enrichment is evident in all cases. Also evident in Fig. 3 is a copper-deficient zone (compared with the bulk level) extending approximately 95 nm below thc enrichment zone. Table I lists the width of the copper enrichment zone and the copper content in the bulk and at the interface for each of the anodized specimens. The values listed for the bulk composition are in good agreement for the nominal copper concentration of 4.5 wt.~; listed for the 2024-T3 alloy. Thc RBS data also showed the stoichiometry of the anodic oxide to be AI20 3. TABLE I RUTIIERFORD BACKSCAT'IERING DAFA OBTAINED WITH A 2 O A K I T E A N D A N O I ) I Z E D A T 10 V IN 1.0 M
Anodization time
Copper interJacial width
(min)
(nm)
0 1
3 5 10 20 60 120
38 38 48 57 76 91 85 91
MeV '*He ~
BEAM F R O M
2024-T3
E T C H E D IN
H3PO4
Copper content (wt.%) Bulk
lnterJace
4.8 4.1 4.1 4.1 4.1 3.9 4. I 4.1
7.0 6.3 5.0 4.6 4.5 4.1 4.2 4.2
158
J.S. SOLOMON, N. T. McDEVITT
The apparent disagreement in the copper distributions, as determined with the two techniques, can be explained in terms of the effect of the interface morphology on the ASPA depth resolution At/t 8. A serious problem with the use of ASPA to characterize interracial regions is the broadening effects on the sputter profile contours representing an interface, owing to poor depth resolution At/t. Mathieu and Landolt ~ have reported an interracial width in terms of the sputter time At for a non-porous-type anodic oxide on aluminum of approximately 30°,~i of the total thickness using a 3 keV ion beam.
C~
h
....i
.....
/~.
i t'!
.... - - - . . - .
,"iil
08
1--06
~._.5 __ IO.
~,__~_~ --r
,zoo
--
,4oo \~ ENERGY (keV) w --
.
_
,%'0o
L_
~
-
160 ZOO 360 ,(0o T H I C K N E S S [nm)
500
600
Fig. 3. RBS spectra (obtained using a 2 MeV 4He + beam) of the copper and phosphorus peaks from Oakite-etched 2024-T3 anodized in room temperature 1.0 M H~PO4 at 10 V; each spectrum is labeled with the anodization time (in minutes). Fig. 4. Depth resolution At/t (where t is the sputtering time in minutes) vs. the anodic oxide thickness sputtered from aluminum ( ~ ) and 2024-T3 (E3) anodized in room temperature 1.0 M H3PO a at 10 V.
Figure 4 contains plots of At/t versus oxide thickness obtained from oxygen profiles from aluminum and 2024-T3 anodized in 0.1 M H3PO 4. The difference in the respective levels (about 35% for aluminum and about 55~,,, for 2024-T3) of At/t for films thicker than 100 nm is most probably related to the difference in the respective oxide layer textures especially near the interface 9' 1.o.As a consequence of the relatively poor depth resolution of 55% for 2024-T3, interfacial features less than 55 nm wide in a specimen with an oxide film 100 nm thick cannot be characterized by ASPA with a high degree of confidence. 4. DISCUSSION
During the anodization of copper-containing aluminum alloys such as 2024-T3 there is a constant evolution of oxygen 6' ~~ which is probably related to a charge
CHARACTERIZATION OF C u IN H3PO 4- ANODIZED 2 0 2 4 - T 3 AI
159
transfer mechanism involving copper rather than aluminum, resulting in copper dissolution by the electrolyte. The net result is a lower concentration ofAl 3 + ions to combine with 0 2 - ions to form oxide. The consequence of this is reflected in Fig. 1 which shows a lower maximum oxide thickness on the alloy. However, the growth rate for the alloy is greater than that for the metal at the onset of anodization. This phenomenon is possibly related to the effect of the evolution of oxygen during the anodization of the alloy. According to Strehblow and Doherty 11 the emission of Oz during anodization prevents the formation of a continuous oxide film. In the case of the metal where little or no oxygen is evolved the initial oxide is a continuous dense layer. The thickness of this layer is voltage dependent with a thickness-to-voltage ratio of 1-2 nm V - 1 (ref. 12). This film forms in a very short period of time and causes a very rapid decrease in current density owing to its electrical resistivity. Almost simultaneously with the formation of the oxide the electrolyte begins to dissolve the oxide non-uniformly, creating thinner areas (pores) on which the oxide grows in a columnar fashion at a faster time-dependent rate. Since the alloy is uninhibited at the onset of anodization by the formation of a continuous oxide film, current not consumed by copper initially forms oxide on the alloy at a faster overall growth rate than on the pure metal. Another possible consequence o f O 2 evolution due to the presence of copper is the effect on oxide morphology. Oxygen evolution most probably occurs at surface sites containing higher copper concentrations. Only when the copper concentration at one of these sites is decreased by the formation of a soluble compound can oxide growth begin. However, if the local copper concentration increases with anodization time oxide growth will be retarded or possibly stopped. A'ccording to the data presented above copper is being replenished at the o x i d e metal interface while being consumed in the anodic process. The f'eplenishment of copper has been shown to be a diffusion process relative to the creation of vacancies at the oxide- metal interface TM 14. Because of the competition between the time-dependent processes of diffusion and dissolution, local copper concentrations are constantly changing. Consequently, oxide growth does not occur at fixed rates at particular sites during the course of anodization and the oxide morphology on the alloy may not be one of well-defined columns. 5. CONCLUSIONS
The presence of copper can significantly affect anodic oxide growth behavior during the electrochemical anodization of 2024-T3 aluminum in H3PO 4. At the onset of anodization the presence of copper can inhibit the formation of a uniform dense oxide film normally observed in the anodization of the pure metal. The result is a slightly greater growth rate compared with that for the pure metal. The anodization process results in the bulk diffusion of copper to the oxidemetal alloy interface where, after a period of time, a steady state enriched concentration zone of copper is established. This zone broadens with longer anodization time. The steady state enrichment of copper is the result of the cofnpetition between the two time-dependent mechanisms of diffusion and electrolyte dissolution. Because of this cyclic process growth rates may vary at local
160
J.S. SOLOMON, N. T. MC'I)EVITT
sites which in turn can result in a somewhat different oxide morphology from that produced on the pure metal. ACKNOWLEDGMENT
The work performed by J. S, Solomon was a partial requirement for his Master of Engineering dcgree and was sponsored by the Air Forcc Wright Aeronautical Laboratories, Air Force Systems Command, Contract F33615-78-C-5102. REFERENCES
G . E . T h o m p s o n , R. C. Furneau, G. (7. Wood, J. A. Richardson and J. S. Goode. Nature (l,on&m). 272 (1978) 433. 2 J.P. O'Sullivan and G. C. Wood. Proc. R. Soc. London, Ser. A. 317 (1970) 511. 3 J . U . Diggle, T. C. Downie and C. W. Goulding, Chem. Rev., 69 (1969) 365. 4 J.S. Solomon and D. E. Hanlin, Appl. SurJ~ Sci., 4 (1980) 307. 5 F. Keller, G. W. Wilcox, M. Tosterud and C. J. Slunder, Met. Alloy.~, 10 (1939) 219. 6 G . H . Kissin, B. E. Deal and R. V. Paulson, in G. H. Kissin (cd.), The Finishing ~! Aluminum, Reinhold. New York, 1963, p. 13. 7 J.S. Solomon and W. I.,. Baun. J. Vac. Sci. Technol., 12(1975) 375. 8 H.J. Mathieu and D. Landolt, in R. Debrozemsky, F. Rudenauer, 1. Viehbock and A. Breth (eds.). Proc. 7th Int. Vacuum Congr. and 3rd Int. (*ot![i on Solid SurJaces, Vienna. 1977, Berger. Vienna, 1977, p. 2213. 9 S. H o f m a n n and J. Erlewein, 7tlin Solid Films, 43 (1977) 275. 10 H.J. Mathieu, D. E. McClure and D. Landolt, Thin Solid k)'hns, 38 (1976) 281. I 1 H . H . Strehblow and C. J. Doherty, J. Electrochem. Sot., 125 (1978) 30. 12 R.C. Furneau, G. E. T h o m p s o n and G. C. Wood, Corros. Sci., 18 (1978) 853. 13 11. H. Strehblow, C. M. Mellian-Smith and W. M. Augustyniak, J. Eh'~'trochem. Sot., 125 (1978) 915. 14 W . D . Mackintosh, F. Brown and H. H. Plattner, d. Ele~'trochem. Sot., 121 (1974) 1281. 1
155
PREPARATION AND CHARACTERIZATION
C H A R A C T E R I Z A T I O N O F C O P P E R IN P H O S P H O R I C - A C I D A N O D I Z E D 2024-T3 A L U M I N U M BY A U G E R E L E C T R O N SPECTROSCOPY AND R U T H E R F O R D BACKSCATTERING* J. S. SOLOMON
University o f Dayton Research Institute, Dayton, OH 45469 (U.S.A.) N. T. McDEVITT Materials Laboratory, Air Force Wright Aeronautical Laboratories, Wright-Patterson Air Force Base, OH 45433 (U.S.A.) (Received March 23, 1981 ; accepted April 8, 1981)
The effects of the electrochemical anodization of deoxidized 2024-T3 aluminum on copper were characterized by Auger electron spectroscopy and Rutherford backscattering. Anodization was performed in phosphoric acid at a constant potential. Data are presented which show that constant potential anodization of 2024-T3 is more efficient than that of aluminum in terms of oxide growth rates for short anodization times. However, the maximum anodic oxide thickness achievable on thc alloy is less than that achievable on the pure metal. Copper is shown to be enriched at the oxide-metal interface because of its diffusion from the bulk during anodization. Possible effects of copper on oxide growth mechanisms and oxide morphology are discussed.
1. INTRODUCTION
When aluminum is anodized in an aggressive electrolyte such as H3PO 4, the resulting anodic oxide has a porous structure with well-defined columns I-3. The oxide growth is limited by both the increasing current-retarding effects of the growing oxide film and the dissolving action of the acid electrolyte. The maximum oxide thickness is reached when the dissolution and growth rates become equal. The anodic oxide growth mechanism for 2024-T3 is essentially the same as that for aluminum. However, the morphology of the oxide layer may differ from that of the pure metal, as determined by scanning electron microscopy 4. This may be due in part to thc presence of intermetallic compounds, such as CuAI 2, and of microconstitui~nts, such as AI-Cu-Mg. Since these inclusions have different oxidation or dissolution rates from those of the AI-Cu solid solution, discontinuities within the anodic oxide layer can result 5. Furthermore, the presence of copper in aluminum has been shown to have considcrable effects on oxide growth rates 4-6. In this paper we discuss the effects of anodiza'don on the distribution of copper within the oxide and the oxide-metal interface of 2024-T3 aluminum and its possible influence on anodic oxide growth and morphology. The anodized pieces * Paper presented at the International Conference on Metallurgical Coatings, San Francisco, CA, U.S.A., April 6- I0, 1981. 0040-6090/81/0000-0000/$02.50
.4.3 Elsevier Sequoia/Printed in The Netherlands
156
J.S. SOI.OMON, N. T. Mcl)EVITT
were characterized by Auger electron spectroscopy (AES), Auger sputter profile analysis (ASPA) and Rutherford backscattering (RBS). 2. EXPERIMENTAI. DETAILS
2.1. Specimen preparation Circular pieces of 2024-T3 aluminum approximately 5 cm in diameter and 0.2 cm thick were first degreased with acetone. One set was then deoxidized in Oakite (obtained from Oakite Products of Canada Ltd., Bramlea, Ontario, Canada) for 15 min and rinsed in distilled H20. The deoxidized pieces were anodized at a constant potential of 10 V in a room temperature constant circulation bath of 1.0 M H3PO 4 for times of 1, 3, 5, 10, 20, 60 and 120 min. After anodization the specimens were rinsed with distilled H 2 0 and hot air dried. 2.2. Method ofanalysi~ 2.2.1. Auger electron spectroscopy The Auger instrumentation and data-handling scheme used in this work have been described previously v. A coaxial electron gun which was operated at 4 keV with a current density of 0.10 A c m - 2 was used to provide high energy excitation of the specimen. The ion beam used for sputter profiling in this work was generated with a Physical Electronics Industries model 94-191 sputter ion gun which was operated with a beam energy of 2 keV and an ion current density of approximately 3 ~A m m - 2 in a partial pressurc of argon of 6.6 x 10- ~ Pa. 2.2.2. Ruther[brd backscattering RBS was performed using a 2 MeV 4He" ion beam produced with a Van de Graaff generator. The backscattered ions were collected by a solid state detector at an angle 0 of 160 measured against the forward direction of the incident beam. The energy resolution of the detector was such that the value of dE/dx (in keV p.g cm - ~) for '*He ' ions was subject to an error of -+ 5'~';ior a depth resolution of 20 nm. Backscattering factors for copper and aluminum were calculated assuming the 2024T3 composition to be Alo ~sCuo 02 . 3. RESULTS Figure 1 contains anodization time versus oxide thickness curves determined from ASPA data for aluminum and 2024-T3 anodized in 1.0 M H 3 P O 4 at a constant potential of 10 V. The oxide on 2024-T3 has a limiting thickness of approximately 400 nm while growth continues on the pure metal (the maximum thickness on the metal was approximately 1000 nm). The copper in-depth profiles from Oakite-deoxidized alloy specimens which were anodized at 10 V for 1, 3, 5 and 10 min are shown in Fig. 2. These profiles all reflect an increase in copper at the oxide metal interface with the profiles from the specimens anodized for 1, 3 and 5 min showing copper enrichment at the interface. The profile from the specimen anodized for 10 min no longer shows a copper enrichment at the interface but a concentration level equal to that in the bulk. The depth distributions of the copper-enriched zones in the specimens anodized for 1.3
CHARACTERIZATION OF C H IN
H3PO 4- ANODIZED 2 0 2 4 - T 3 A1
157
and 5 min, based on the sputter rate for aluminum metal, are 34, 59 and 84 nm respectively. The data in Fig. 2 suggest that copper enrichment can be attributed solely to the chemical deoxidation treatment and that subsequent H3PO, t anodization depletes copper from the oxide-metal interface. However, since the copper profile from a non-etched and anodized specimen also showed copper enrichment at the oxide metal interface, enrichment can be attributed to the anodization process itself. Copper was not detccted within the anodic oxide layer, regardless of thickness or pre-anodization treatment. '~ 800
¢.
Ill
// < 0
o
5
~
~
~o
'
2~
6
ANOOIZAT1ON TIME(MIN)
,'o
~'o
3b
4"0
SPUTTER TIME (MiN)
Fig. 1. Oxide film thicknesses on aluminum ( - - ) and 2024-T3 (-- ) determined from ASPA and plotted as functions of the anodization time at 1O V in room temperature 1.0 M H3PO 4. Fig. 2. Auger sputter protiles of copper in anodic oxide films on Oakite-etched 2024-T3 anodized for 1.3. 5 and 10 min at 10 V in room temperature 1.0 M H3PO4.
Figure 3 contains RBS spectra in the region 1.2-1.8 MeV for specimens etched and anodized for times up to 2 h. Contrary to the ASPA data, copper interracial enrichment is evident in all cases. Also evident in Fig. 3 is a copper-deficient zone (compared with the bulk level) extending approximately 95 nm below thc enrichment zone. Table I lists the width of the copper enrichment zone and the copper content in the bulk and at the interface for each of the anodized specimens. The values listed for the bulk composition are in good agreement for the nominal copper concentration of 4.5 wt.~; listed for the 2024-T3 alloy. Thc RBS data also showed the stoichiometry of the anodic oxide to be AI20 3. TABLE I RUTIIERFORD BACKSCAT'IERING DAFA OBTAINED WITH A 2 O A K I T E A N D A N O I ) I Z E D A T 10 V IN 1.0 M
Anodization time
Copper interJacial width
(min)
(nm)
0 1
3 5 10 20 60 120
38 38 48 57 76 91 85 91
MeV '*He ~
BEAM F R O M
2024-T3
E T C H E D IN
H3PO4
Copper content (wt.%) Bulk
lnterJace
4.8 4.1 4.1 4.1 4.1 3.9 4. I 4.1
7.0 6.3 5.0 4.6 4.5 4.1 4.2 4.2
158
J.S. SOLOMON, N. T. McDEVITT
The apparent disagreement in the copper distributions, as determined with the two techniques, can be explained in terms of the effect of the interface morphology on the ASPA depth resolution At/t 8. A serious problem with the use of ASPA to characterize interracial regions is the broadening effects on the sputter profile contours representing an interface, owing to poor depth resolution At/t. Mathieu and Landolt ~ have reported an interracial width in terms of the sputter time At for a non-porous-type anodic oxide on aluminum of approximately 30°,~i of the total thickness using a 3 keV ion beam.
C~
h
....i
.....
/~.
i t'!
.... - - - . . - .
,"iil
08
1--06
~._.5 __ IO.
~,__~_~ --r
,zoo
--
,4oo \~ ENERGY (keV) w --
.
_
,%'0o
L_
~
-
160 ZOO 360 ,(0o T H I C K N E S S [nm)
500
600
Fig. 3. RBS spectra (obtained using a 2 MeV 4He + beam) of the copper and phosphorus peaks from Oakite-etched 2024-T3 anodized in room temperature 1.0 M H~PO4 at 10 V; each spectrum is labeled with the anodization time (in minutes). Fig. 4. Depth resolution At/t (where t is the sputtering time in minutes) vs. the anodic oxide thickness sputtered from aluminum ( ~ ) and 2024-T3 (E3) anodized in room temperature 1.0 M H3PO a at 10 V.
Figure 4 contains plots of At/t versus oxide thickness obtained from oxygen profiles from aluminum and 2024-T3 anodized in 0.1 M H3PO 4. The difference in the respective levels (about 35% for aluminum and about 55~,,, for 2024-T3) of At/t for films thicker than 100 nm is most probably related to the difference in the respective oxide layer textures especially near the interface 9' 1.o.As a consequence of the relatively poor depth resolution of 55% for 2024-T3, interfacial features less than 55 nm wide in a specimen with an oxide film 100 nm thick cannot be characterized by ASPA with a high degree of confidence. 4. DISCUSSION
During the anodization of copper-containing aluminum alloys such as 2024-T3 there is a constant evolution of oxygen 6' ~~ which is probably related to a charge
CHARACTERIZATION OF C u IN H3PO 4- ANODIZED 2 0 2 4 - T 3 AI
159
transfer mechanism involving copper rather than aluminum, resulting in copper dissolution by the electrolyte. The net result is a lower concentration ofAl 3 + ions to combine with 0 2 - ions to form oxide. The consequence of this is reflected in Fig. 1 which shows a lower maximum oxide thickness on the alloy. However, the growth rate for the alloy is greater than that for the metal at the onset of anodization. This phenomenon is possibly related to the effect of the evolution of oxygen during the anodization of the alloy. According to Strehblow and Doherty 11 the emission of Oz during anodization prevents the formation of a continuous oxide film. In the case of the metal where little or no oxygen is evolved the initial oxide is a continuous dense layer. The thickness of this layer is voltage dependent with a thickness-to-voltage ratio of 1-2 nm V - 1 (ref. 12). This film forms in a very short period of time and causes a very rapid decrease in current density owing to its electrical resistivity. Almost simultaneously with the formation of the oxide the electrolyte begins to dissolve the oxide non-uniformly, creating thinner areas (pores) on which the oxide grows in a columnar fashion at a faster time-dependent rate. Since the alloy is uninhibited at the onset of anodization by the formation of a continuous oxide film, current not consumed by copper initially forms oxide on the alloy at a faster overall growth rate than on the pure metal. Another possible consequence o f O 2 evolution due to the presence of copper is the effect on oxide morphology. Oxygen evolution most probably occurs at surface sites containing higher copper concentrations. Only when the copper concentration at one of these sites is decreased by the formation of a soluble compound can oxide growth begin. However, if the local copper concentration increases with anodization time oxide growth will be retarded or possibly stopped. A'ccording to the data presented above copper is being replenished at the o x i d e metal interface while being consumed in the anodic process. The f'eplenishment of copper has been shown to be a diffusion process relative to the creation of vacancies at the oxide- metal interface TM 14. Because of the competition between the time-dependent processes of diffusion and dissolution, local copper concentrations are constantly changing. Consequently, oxide growth does not occur at fixed rates at particular sites during the course of anodization and the oxide morphology on the alloy may not be one of well-defined columns. 5. CONCLUSIONS
The presence of copper can significantly affect anodic oxide growth behavior during the electrochemical anodization of 2024-T3 aluminum in H3PO 4. At the onset of anodization the presence of copper can inhibit the formation of a uniform dense oxide film normally observed in the anodization of the pure metal. The result is a slightly greater growth rate compared with that for the pure metal. The anodization process results in the bulk diffusion of copper to the oxidemetal alloy interface where, after a period of time, a steady state enriched concentration zone of copper is established. This zone broadens with longer anodization time. The steady state enrichment of copper is the result of the cofnpetition between the two time-dependent mechanisms of diffusion and electrolyte dissolution. Because of this cyclic process growth rates may vary at local
160
J.S. SOLOMON, N. T. MC'I)EVITT
sites which in turn can result in a somewhat different oxide morphology from that produced on the pure metal. ACKNOWLEDGMENT
The work performed by J. S, Solomon was a partial requirement for his Master of Engineering dcgree and was sponsored by the Air Forcc Wright Aeronautical Laboratories, Air Force Systems Command, Contract F33615-78-C-5102. REFERENCES
G . E . T h o m p s o n , R. C. Furneau, G. (7. Wood, J. A. Richardson and J. S. Goode. Nature (l,on&m). 272 (1978) 433. 2 J.P. O'Sullivan and G. C. Wood. Proc. R. Soc. London, Ser. A. 317 (1970) 511. 3 J . U . Diggle, T. C. Downie and C. W. Goulding, Chem. Rev., 69 (1969) 365. 4 J.S. Solomon and D. E. Hanlin, Appl. SurJ~ Sci., 4 (1980) 307. 5 F. Keller, G. W. Wilcox, M. Tosterud and C. J. Slunder, Met. Alloy.~, 10 (1939) 219. 6 G . H . Kissin, B. E. Deal and R. V. Paulson, in G. H. Kissin (cd.), The Finishing ~! Aluminum, Reinhold. New York, 1963, p. 13. 7 J.S. Solomon and W. I.,. Baun. J. Vac. Sci. Technol., 12(1975) 375. 8 H.J. Mathieu and D. Landolt, in R. Debrozemsky, F. Rudenauer, 1. Viehbock and A. Breth (eds.). Proc. 7th Int. Vacuum Congr. and 3rd Int. (*ot![i on Solid SurJaces, Vienna. 1977, Berger. Vienna, 1977, p. 2213. 9 S. H o f m a n n and J. Erlewein, 7tlin Solid Films, 43 (1977) 275. 10 H.J. Mathieu, D. E. McClure and D. Landolt, Thin Solid k)'hns, 38 (1976) 281. I 1 H . H . Strehblow and C. J. Doherty, J. Electrochem. Sot., 125 (1978) 30. 12 R.C. Furneau, G. E. T h o m p s o n and G. C. Wood, Corros. Sci., 18 (1978) 853. 13 11. H. Strehblow, C. M. Mellian-Smith and W. M. Augustyniak, J. Eh'~'trochem. Sot., 125 (1978) 915. 14 W . D . Mackintosh, F. Brown and H. H. Plattner, d. Ele~'trochem. Sot., 121 (1974) 1281. 1