Corrosion fatigue behavior of 7050 aluminum alloys in different tempers

Corrosion fatigue behavior of 7050 aluminum alloys in different tempers

PII: Engineering Fracture Mechanics Vol. 59, No. 6, pp. 779±795, 1998 # 1998 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0013-...

2MB Sizes 2 Downloads 62 Views

PII:

Engineering Fracture Mechanics Vol. 59, No. 6, pp. 779±795, 1998 # 1998 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0013-7944/98 $19.00 + 0.00 S0013-7944(97)00173-2

CORROSION FATIGUE BEHAVIOR OF 7050 ALUMINUM ALLOYS IN DIFFERENT TEMPERS CHIH-KUANG LIN{ and SHENG-TSENG YANG Department of Mechanical Engineering, National Central University, Chung-Li 32054, Taiwan AbstractÐCorrosion fatigue (CF) experiments, including both high-cycle axial fatigue (S±N curve) and fatigue crack growth (FCG), have been performed on 7050 aluminum alloys in a 3.5 wt% NaCl solution as a function of aging treatment. The results of these environmental tests were compared with those obtained in laboratory air to characterize the e€ect of aging treatment on CF susceptibility. Fatigue resistance in both peak aged (T6) and overaged (T73) tempers was dramatically reduced by the aqueous chloride environment. The FCG rates for T73 condition were lower than the counterparts for T6 condition in both air and saline solution. 7050-T73 alloy exhibited longer fatigue lives in air but shorter ones in the corrosive environment as compared to the T6 temper. This may be attributed to the formation of more extensive and larger corrosion pits acting as crack nuclei to facilitate crack initiation, in the T73 tempered condition. Comparison of CF and stress corrosion cracking (SCC) results reveals that overaging treatments used to improve grain boundary characteristics and increase the intergranular SCC resistance might not guarantee an equivalent improvement in the resistance to transgranular CF cracking. # 1998 Elsevier Science Ltd. All rights reserved KeywordsÐ7050 aluminum alloy, corrosion fatigue, aging treatment, high-cycle axial fatigue, fatigue crack growth, stress corrosion cracking, pitting.

1. INTRODUCTION 7XXX SERIES aluminum alloys are widely used in the aircraft industry due to their high strength to density ratio. These materials (e.g. 7075-T6) can reach their peak strength through proper T6 aging treatment. However, they are very susceptible to intergranular stress corrosion cracking (SCC) in the T6 temper, especially in chloride-containing media in the short transverse direction of thick sections [1]. Overaging treatments (like T73 temper) were developed to improve the short transverse SCC resistance by increasing the grain-boundary precipitate size and spacing, but with some reduction of tensile strength from the T6 temper [1]. However, it cannot be assumed that alloys and tempers with good SCC resistance would show good resistance to corrosion fatigue (CF) as CF failures of aluminum alloys are characteristically transgranular [1]. For example, Jacko and Duquette [2] reported that no signi®cant di€erence in total fatigue lives could be detected between 7075-T6 and -T73 aluminum alloys when they were tested in the form of smooth axial fatigue specimens in aerated 0.5 N NaCl solution. In addition, Khobaib et al. [3] also reported that the fatigue crack growth (FCG) rates measured from pre-notched specimens in 0.1 M NaCl solution for the T6 and T73 conditions of 7075 alloy were almost the same at the stress intensity range (DK) between 8 to 15 MPa m1/2. Similar observation of no inherent advantage of FCG resistance for T73 over T6 temper in 0.2 M NaCl solution was also reported by Stoltz and Pelloux [4]. Other investigations on 7075 aluminum alloys [5, 6], however, found that overaging treatment T73 did exhibit superior FCG resistance to peak aged T6 temper in the environments of 3.5 wt% NaCl solution and 90% relative humidity air. The 7050 alloy was developed to obtain a combination of strength, fracture toughness, and SCC resistance superior to that provided by conventional high strength alloys like 7075-T6 [6]. Alloy 7050 was developed speci®cally for an optimum combination of the above properties in thick sections by reducing the quench sensitivity. Increased Cu content provided a good balance of strength and SCC resistance, while restriction of the impurity elements Fe and Si provided high toughness [6]. The enhancement of the SCC resistance and toughness of 7050 alloy, as compared to other conventional high strength 7XXX alloys, resulted from the e€ect of Cu on increasing the temperature range of GP zone stability [7]. Environmental enhancement of crack {Author to whom all correspondence should be addressed. 779

780

CHIH-KUANG LIN and SHENG-TSENG YANG

growth under sustained load and cycle fatigue conditions has long been recognized as critical problem for high strength aluminum alloys. Most of the research work done with 7050 aluminum alloys has been con®ned to evaluating the SCC behavior in di€erent tempers (e.g. in two classic review articles [8, 9]). However, little work [6, 10, 11] has been done on the CF characteristics of 7050 alloy. More research is necessary to characterize the e€ect of aging treatment (T6 vs T73) on the CF resistance in chloride environment for 7050 alloys due to their extensive use in many components of primary load carrying structures of aircraft ¯ying in marine atmospheres. Most of the previous work (e.g. Refs [2±6, 10, 11]) discussing the CF behavior of 7XXX aluminum alloys was based on either the high-cycle fatigue (S±N curve) or FCG experiment. It was not possible to distinguish in these results the environmental e€ect on the fatigue crack initiation and propagation stages separately, for a given alloy composition and aging treatment. This research was therefore planned to characterize the in¯uence of environment on the fatigue crack initiation and propagation stages on a 7050 aluminum alloy in an aerated 3.5 wt% NaCl solution by systematic experiments, including both high-cycle axial fatigue (S±N curve) and FCG (da/dN ÿ DK). Two aging treatments, T6 (peak age) and T73 (overage), were applied to assess whether the overaging temper T73 would improve the CF resistance as e€ectively as it did for the SCC resistance. The results obtained in salt water were then compared to those obtained in laboratory air to characterize the e€ect of aqueous chloride environment on the reduction in fatigue resistance for the given alloy. A limited number of SCC tests under sustained loads by using the same types of smooth and pre-notched specimens used in the CF tests were also performed to make a comparison between the SCC and CF failure characteristics.

2. EXPERIMENTAL PROCEDURE The material used in this investigation was 7050 high strength aluminum alloy supplied in the form of 150-mm-thick rolled plate. The wt% composition of the major elements in the studied alloy was 5.56 Zn, 1.97 Mg, 1.80 Cu, 0.146 Zr, and Al (balance). Two di€erent tempers, T6 (peak aging) and T73 (overaging), were applied to the specimens. The T6 temper consists of solution treatment at 4708C  1 h, water quench, natural aging 7 days, and aging at 1208C  24 h. The T73 temper consists of solution treatment at 4708C  1 h, water quench, natural aging  7 days, and two-stage aging at 1208C  24 h + 1608C  30 h. Specimens used in both high-cycle axial fatigue and FCG tests were made in the short transverse ST orientation as speci®ed in ASTM Method E647-91 [12]. The monotonic tensile properties of both T6 and T73 conditions are listed in Table 1. These tensile properties were obtained by using tensile specimens having a cylindrical gage section of 6 mm in diameter and 25 mm in length. High-cycle axial fatigue tests were conducted as per ASTM Method E466-82 [13] in both laboratory air and 3.5% NaCl solution at room temperature to determine the stress±life (S±N) curves. FCG experiments were performed in accordance with ASTM Method E647-91 [12] on compact tension (CT) specimens under the same testing environments to determine the da/ dN ÿ DK relationship. The smooth-surface axial fatigue and pre-notched CT specimens were machined to the geometry and dimensions as shown in Fig. 1 according to the ASTM E46682 [13] and E647-91 [12], respectively. Both types of fatigue tests were conducted on a closedloop, servohydraulic machine under tension±tension load control with a load ratio R = 0.1. The cyclic load applied was a sinusoidal wave with a frequency of 20 Hz. The high-cycle axial fatigue tests were run to failure or to 2  106 cycles where specimen was considered to be a runout. The crack length in the FCG tests was determined by the compliance technique recommended by ASTM E647-91 [12] using a clip gage mounted on the front edge of the CT specimen to monitor Table 1. Monotonic tensile properties for 7050 aluminum alloy in di€erent tempers

Temper T6 T73

Ultimate strength (MPa)

Yield strength (MPa)

Elongation in 25 mm (%)

544 492

482 448

3.2 3.6

Corrosion fatigue behavior of 7050 aluminum alloys

781

Fig. 1. Geometries of specimens: (a) axial fatigue test specimen; and (b) compact tension (CT) specimen.

the crack-mouth-opening displacement during testing. All CT specimens in FCG tests were ®rst precracked in laboratory air to obtain fatigue precrack that satis®es the requirements of ASTM E647-91 [12]. Most experiments were duplicated to check for repeatability. Special acrylic cells were fabricated for fatigue testing in the corrosive environment. For the high-cycle axial fatigue test, the cell enclosed the gage section of the specimen in a fully submerged condition during fatigue testing as shown in Fig. 2(a). For the FCG tests, the cells on each side of the CT specimen surfaces enclosed the areas of the crack growth path as shown in Fig. 2(b). In both set-ups, the 3.5 wt% NaCl solution circulated through the cells from an external reservoir and returned to the reservoir by means of a small lift pump. The salt water solution in the reservoir had a pH value of 7.16 both before and after the tests. All environmental tests were conducted under free corrosion condition, i.e. no external electric potential was applied to the specimens. A special vacuum bag sealant having very low sti€ness and high adhesion capability was used to seal the cells on the specimens. A check-up FCG test using a CT specimen with acrylic cells attached without NaCl solution circulated was conducted in laboratory air to compare with that in laboratory air without acrylic cells attached. No visible di€erence in the FCG rates from these two comparative tests could be discovered indicating the use of this sealant did not interfere with the compliance of the loading system and the corrosion FCG data obtained in this manner are reliable. In order to characterize the di€erence between CF and SCC behavior for this particular aluminum alloy in the T6 and T73 tempers, a limited number of smooth axial and CT specimens were subjected to selected constant loads in the same corrosive environment to obtain the data

782

CHIH-KUANG LIN and SHENG-TSENG YANG

Fig. 2. Schematic view of the experimental set-up for corrosion fatigue tests: (a) axial fatigue; and (b) fatigue crack growth.

of failure time and crack velocity. Characterizations of the fracture surface morphology were made by scanning electron microscopy (SEM). Selected unbroken CT specimens were polished and etched on the side surfaces to observe the crack propagation path by optical microscopy. 3. RESULTS AND DISCUSSION 3.1. Environmental e€ect on the fatigue behavior Figure 3 is a plot of the fatigue life as a function of maximum cyclic stress for 7050-T6 and -T73 smooth specimens tested in both air and 3.5 wt% NaCl solution. The straight solid and dashed lines in Fig. 3 represent the best-®t S±N curves by a simple power law. The results of the FCG tests using CT specimens were plotted as (da/ dN) vs (DK) in Fig. 4. The experimental results presented in Figs 3 and 4 indicated that there were two signi®cant environmental e€ects of the sodium chloride solution on the fatigue behavior of the 7050 alloys tested in each temper:

Corrosion fatigue behavior of 7050 aluminum alloys

783

Fig. 3. Comparison of S±N curves in di€erent environments for 7050 aluminum alloy in di€erent tempers (arrows designate runout tests; number label designates repeated data points).

(i) a reduction in fatigue life; and (ii) an enhancement in FCG rate. The fatigue limits (de®ned as the maximum cyclic stresses corresponding to the fatigue life of 2  106 cycles) for both 7050T6 and -T73 alloys in air and aqueous chloride environments are given in Table 2. The percent reduction in fatigue limit by the corrosive environment is 37.5% for T6 temper and 67.4% for T73 temper. These comparisons show that T73 temper is more susceptible to the NaCl solution than T6 temper in terms of CF performance of smooth specimens. Corrosion fatigue is commonly de®ned as the damage and failure of a material resulting from the combined action of cyclic stresses and an aggressive environment. The signi®cant reduction in the fatigue limit by chloride environment for the 7050 alloys in each aged condition might be partially attributed to the formation of corrosion pits. The corrosion pits on the

Fig. 4. Comparison of FCG curves in di€erent environments for 7050 aluminum alloy in di€erent tempers.

784

CHIH-KUANG LIN and SHENG-TSENG YANG Table 2. Comparison of cyclic fatigue strengths in air and 3.5 wt% NaCl Endurance limit (at 2  106 cycles), Smax (MPa)

Temper

Air

3.5 wt% NaCl

Percentage change

T6 T73

200 215

125 70

ÿ37.5% ÿ67.4%

smooth surfaces of axial fatigue specimens may become the sites of stress concentration and thus enhance crack initiation leading to a shorter fatigue life. This can be supported by the SEM fractography observations. Fatigue fracture surface morphologies near the crack initiation sites for 7050-T6 and -T73 smooth specimens tested in salt water in high-cycle fatigue regime are shown in Figs 5 and 6, while those counterparts in air are shown in Fig. 7. It can be seen that fracture of CF smooth specimens originated from corrosion pits which were absent in the specimens tested in air. However, at high cyclic stress regions, the detrimental e€ect of the aggressive solution was reduced as shown by the convergence of the S±N curves in di€erent environments for each temper condition. This implies that as the applied cyclic stress increased, the dominant CF crack initiation mechanism might change from pitting to another mechanism, namely anodic slip dissolution. It has been reported for other aluminum alloys that CF crack nucleation process was believed to be controlled by a slip-induced anodic dissolution mechanism at high cyclic stresses (or strain rates) and dominated by pitting mechanism at lower cyclic stresses (or strain rates) [14±16]. This is due to the fact that high stresses provide sucient mechanical driving force to produce slip steps and that, in the case of a short fatigue life, the corrosion pit may not become large enough to act as a crack nucleus. Hydrogen embrittlement has been proposed to account for the enhanced FCG rate in corrosive environments for aluminum alloys [10, 14, 17±20]. Corrosion fatigue and surface chemistry studies in high strength aluminum alloys suggest that hydrogen produced by the reaction of water vapor with the freshly formed fracture surface is responsible for the embrittlement process and an enhancement in crack velocity, by an active path dissolution mechanism [20]. The fractography results support such a hydrogen embrittlement mechanism for CF crack propagation since the featureless and cleavage-like zone (labels C in Figs 5 and 6) could result from localized embrittlement of the alloy. Figure 7 shows a typical river pattern emanating from the initiation site (label i) and clearly shows the ``feathery'' nature of the river lines on a CF specimen tested in air. Figures 5 and 6, on the other hand, both show two distinct zones emanating from the initiation pit in saline solution, the ®rst zone being essentially featureless, cleavage-like (labels C), while the second zone exhibits a modi®ed river pattern. This type of fracture surface characteristic, which was also observed in other 7XXX aluminum alloys, has been attributed to the hydrogen embrittlement mechanism[2, 17]. Therefore, it is suggested that the hydrogen embrittlement mechanism plays a very important role in increasing the FCG rate in the aqueous environment, particularly in the intermediate and high DK ranges, for both 7050-T6 and -T73 alloys tested. Such enhancement of FCG rate would therefore contribute in part to the reduction of fatigue lives for smooth specimens tested in the aqueous chloride environment. Figure 8 shows the typical striation morphologies observed on the fracture surfaces of fatigue specimens tested in air and 3.5 wt% NaCl solution. The fatigue striations shown in Fig. 8(a) are very smooth and characterized by the featureless rumplings as a result of extensive plastic blunting by shear at the crack tip. This striation feature is the so-called ``ductile striations,'' commonly observed in dry or ambient air for 7XXX aluminum alloys [2±4, 21, 22]. On the other hand, the so-called ``brittle striations'' as shown in Fig. 8(b) are characterized by their roughened appearance accompanied by arrest markings caused by stepwise cleaving of the surface. The major environmental attack and brittle fracture are typical of such micrograph which was generally observed in aqueous chloride solutions for Al±Zn±Mg aluminum alloys [3, 4, 21, 22]. These striation morphology observations provide further evidence of the environment-assisted embrittlement processes that occur at the crack tip and enhance the FCG rate for the 7050-T6 and -T73 alloys tested.

Corrosion fatigue behavior of 7050 aluminum alloys

Fig. 5. SEM fractography of corrosion fatigue for 7050 aluminum alloy in T6 temper near the fracture origin: (a) low magni®cation; and (b) high magni®cation (i: crack initiation site; c: cleavage area).

Fig. 6. SEM fractography of corrosion fatigue for 7050 aluminum alloy in T73 temper near the fracture origin: (a) low magni®cation; and (b) high magni®cation (i: crack initiation site; c: cleavage area).

785

786

CHIH-KUANG LIN and SHENG-TSENG YANG

Fig. 7. SEM fractography of fatigue in air for 7050 aluminum alloy in T6 temper near the fracture origin (i: crack initiation site).

Fig. 8. SEM micrographs of striations of axial fatigue specimens in various environments for 7050 aluminum alloy in T6 tempers: (a) in air; and (b) in 3.5 wt% NaCl.

Corrosion fatigue behavior of 7050 aluminum alloys

Fig. 9. Comparison of pitting on the surfaces of corrosion fatigue specimens for 7050 aluminum alloy in di€erent tempers (p: pitting area).

Fig. 10. SEM micrograph of pitting for 7050 aluminum alloy in T73 temper (p: pitting area).

787

788

CHIH-KUANG LIN and SHENG-TSENG YANG

Corrosion fatigue behavior of 7050 aluminum alloys

789

Table 3. Life of 7050 aluminum alloy under static loading in 3.5 wt% NaCl Temper

Applied stress = 350 MPa Life (h)

T6 T73 a

19.1 19.3 >168.7a >143.7b

Specimen did not break. Loading history: 350 MPa (143.7 h) 4 400 MPa (10.6 h) 4 450 MPa (45.3 h) 4 475 MPa (5.4 h) 4 broken.

b

3.2. Comparison of CF resistance in di€erent tempers It is seen in Figs 3 and 4 that 7050-T73 alloys are superior to 7050-T6 ones in terms of fatigue resistance in air due to the longer fatigue lives for smooth specimens and lower FCG rates for CT specimens. However, the counterpart results in a 3.5 wt% NaCl solution show a di€erent picture. The T73 temper still provides more resistance to FCG in saline water but becomes less e€ective in inhibiting the evolution of fatigue damage as the CF lives of smooth specimens are remarkedly shorter than those in T6 condition. Apparently, there is no inherent advantage to the T73 temper for CF resistance in smooth surfaces as for SCC resistance, when compared to T6 temper. The S±N curves obtained in the chloride environment suggest that the environment-assisted crack nucleation mechanism is more e€ective in smooth specimens of T73 condition than in T6 ones and signi®cantly reduces the fatigue life even though the FCG rates are lower in T73 condition. This also implies that crack initiation is the predominant stage in determining the CF life for the 7050-T6 and -T73 alloys tested and can be attributed to the early nucleation at stress concentrators like corrosion pits. The inferior performance in the S±N response, particularly at low cyclic stresses, for smooth specimens of T73 condition can be attributed to the formation of profuse and larger pits as shown in Fig. 9. The fatigue lives for the T6 and T73 smooth speci-

Fig. 11. Stress corrosion crack growth rate curves of 7050 aluminum alloy in di€erent tempers (data for T73 temper were taken from the study of Speidel[8]).

790

CHIH-KUANG LIN and SHENG-TSENG YANG

mens shown in Fig. 9 are of the same order and around 106 cycles. It is clearly observed in Fig. 9 that pitting phenomenon is much more severe on the circumferential smooth surfaces for the T73 specimen as compared to the T6 ones. Figure 10 is a higher-magni®cation view of the pitting phenomenon observed on the smooth surfaces of T73 specimens. Some of the pits on the T73 specimen shown in Fig. 9 are large enough to be visible by naked eyes while no such scale of corrosion pits were detected on the smooth surfaces of T6 specimens. These large pits can e€ectively become crack nuclei due to stress concentration e€ect and enhance the fatigue crack initiation resulting in a signi®cant reduction in fatigue life. Pitting has been found to occur preferentially around constituent particles in 7XXX aluminum alloys [17] and to show a quicker growth kinetics for aluminum alloys with longer aging time [14]. As the precipitates in the overaged T73 temper are widely spaced and larger than those in the peak aged T6 temper, pits formed around the precipitates which are heavily attacked by the chloride environment are thus larger and become e€ective stress concentrators in the T73 specimens. Figure 6(b) provides evidence of pits as crack nuclei by showing the microcracks emanating from the heavily pitted areas in a 7050-T73 smooth specimen. In this regard, pitting-induced fatigue crack nucleation is enhanced in the overaged 7050-T73 alloy rather than in the peak aged 7050-T6 alloy, resulting in a greater percentage decrease in fatigue limit by the saline solution for the 7050-T73 alloy despite the lower FCG rates observed in T73 condition. 3.3. Comparison of CF and SCC behavior In addition to CF experiments, a limited number of stress corrosion (SC) experiments under sustained constant loads in a similar aqueous environment were conducted to qualitatively characterize the di€erence between CF and SCC mechanisms. For smooth axial specimens, an applied static stress of 350 MPa was selected to exercise the SC tests in duplication for both aging conditions. Table 3 lists the SC results in time to failure for both tempers. The overaged T73 temper apparently exhibits a superior SCC resistance to the peak aged T6 temper in NaCl solution. This is inferred from the fact that both T6 smooth specimens last about 19 h before failure, while the two T73 smooth specimens did not fail at the time of 143.7 and 168.7 h, respectively, when the tests were terminated. The SC experiment on one of the T73 specimen (>143.7 h) was continued until failure by increasing the static stress levels stepwise as described in the footnotes of Table 3. This is done so to obtain a failed T73 smooth specimen by SCC mechanism for fractography analysis. Crack growth rate (CGR) data from SC experiments on the CT specimens are shown in Fig. 11. As no detectable crack extension (or extremely low CGR, <<10ÿ6 mm sÿ1) was observed in the T73 specimen for a certain period of time in the present work, the data for T73 temper presented in Fig. 11 was taken from the study of Speidel [8]. The data from Speidel's study [8] were obtained for the same alloy (7050-T73) under the same environment (3.5% NaCl solution) and specimen orientation (short transverse) but in a di€erent specimen geometry (double cantilever beam, DCB). The CGRs of SC in region II (plateau region) for T73 condition are signi®cantly lower than those in T6 condition by more than two orders of magnitude indicating, again, the T73 temper is superior to T6 temper in terms of SCC resistance. The current SC results obtained from smooth and pre-notched specimens are consistent with previous conclusions drawn on conventional high strength aluminum alloys that there is an inherent advantage of SCC resistance for overaging treatment rather than for peak aging treatment [8]. The superiority of T73 to T6 temper on SCC resistance of smooth specimen may be attributed to the remarkedly lower CGR by the enhanced resistance to intergranular cracking as pitting phenomenon is still more severe in the T73 specimens than in the T6 ones. Figure 12 shows the circumferential surfaces for the smooth axial specimens in T6 and T73 tempers after SC test. It is clearly seen in Fig. 12 the pitting phenomenon on the smooth surface is much more severe for T73 condition than for T6 condition. Note that the counterpart photograph of the CF specimens presented in Fig. 9 also shows the similar phenomenon. These results indicate that T73 temper is more vulnerable to pitting in the aqueous chloride environment regardless of the loading mode (i.e. cyclic or sustained load). However, the resistance to crack propagation may play the most important role in determining the time to failure under sustained loads as the cracking

Corrosion fatigue behavior of 7050 aluminum alloys

Fig. 12. Comparison of pitting on the surfaces of stress corrosion cracking specimens for 7050 aluminum alloy in di€erent tempers (p: pitting area).

Fig. 13. SEM fractography of stress corrosion cracking for 7050 aluminum alloy in T73 temper (i: crack initiation site).

Fig. 14. Corrosion pit growth path for stress corrosion cracking specimens in T73 temper.

791

792

CHIH-KUANG LIN and SHENG-TSENG YANG

Fig. 15. Comparison of crack growth path for 7050 aluminum alloy in T6 tempers: (a) fatigue in air; (b) fatigue in 3.5 wt% NaCl; and (c) SCC in 3.5 wt% NaCl (cracks are indicated by the arrows).

Corrosion fatigue behavior of 7050 aluminum alloys

793

Fig. 16. Comparison of crack growth rates in di€erent loading conditions for 7050 aluminum alloy in T6 temper.

pattern shifts from transgranular fracture for CF to intergranular fracture for SCC resulting in the signi®cant di€erence in CGR between CF and SCC. Figure 13 shows the fracture surface of a T73 smooth axial specimen failed by SCC mechanism. It can be seen that the main crack still originated at a pitting area and propagated in an intergranular mode in contrast to the transgranular propagation mode observed in the counterpart CF specimen shown in Fig. 6. Figure 14 shows the growth path of a corrosion pit for an SC smooth specimen in T73 temper. This micrograph was taken from the cross section of the gage section in an axial specimen after being axially cut, polished, and etched. The corrosion pit (label p) shown in Fig. 14 seems to nucleate within a grain and develops to become a microcrack by growing in a direction perpendicular to the loading direction. It is worth noting that once this microcrack reaches the ®rst grain boundary it begins to grow along the grain boundary. This provides further evidence of intergranular cracking mechanism for 7050 aluminum alloys subjected to sustained loads in NaCl solution. Further comparison of cracking pattern in CF and SCC is given in Fig. 15. These micrographs were taken from the side surfaces of the unbroken CT specimens. The cracks essentially grow by a transgranular type of fracture under cyclic loading regardless of the environment while the cracking pattern shifts to an intergranular type of fracture under sustained constant loads, as evidenced by Fig. 15. Corrosion pits were primarily responsible for the crack nucleation in salt water under both cyclic and sustained loads, while their roles in determination of failure time were rivalled by the crack growth characteristics in di€erent loading modes. A comparison of the CGRs obtained for 7050-T6 alloys under the given testing conditions is given in Fig. 16 where crack velocity (da/dt) is plotted vs the maximum applied stress intensity factor (Kmax). At the intermediate stress intensity region, the FCG rates in 3.5 wt% NaCl solution are remarkably greater than those counterpart CGRs of SCC by two orders of magnitude. Hence, given a similar pit or initial crack size, it would take much longer for the crack to propagate from its initial size to ®nal critical size under static loading simply because of the much lower CGR compared with cyclic loading. The extent of the di€erence in CGR between T6 and T73 tempers under static and cyclic loads might explain why smooth specimens in T73 temper exhibit shorter CF lives but much longer time to failure by SCC mechanism compared with T6 ones. The CGR in T73 condition is superior to that in T6 condition by less than ®ve fold in CF (Fig. 4), but by two orders of

794

CHIH-KUANG LIN and SHENG-TSENG YANG

magnitude in SCC (Fig. 11). The much lower CGRs in SC for T73 smooth specimens apparently make up for the disadvantage of profuse pitting to crack nucleation and extend the time to failure to a much greater extent in comparison with T6 ones. The overaging T73 temper was developed to improve the intergranular SCC resistance by formation of the larger and widely spaced precipitates at the grain boundary [1]. These coarser grain-boundary precipitates can become e€ective hydrogen trapping sites by trapping localized hydrogen present at the grain boundary and reduce the detrimental hydrogen e€ects [23±25]. Apparently, this theory is also con®rmed in the current study for 7050 alloy. However, overaging the 7050-T6 alloy to the T73 temper might not guarantee an equivalent improvement in the CF resistance simply because the cracking pattern shifts from intergranular to transgranular fracture for CF loading. The present results indicate that the peak aged condition has a far superior CF resistance for smooth surfaces in spite of its greater FCG rates. This favorable CF property of the peak aged material seems to arise from its less severe pitting phenomenon on smooth surfaces. It is hoped that the present results and discussions provide helpful information for the fundamental understanding of the CF and SCC behavior in high strength aluminum alloys. 4. CONCLUSIONS 1. Comparison of S±N curves revealed that T73 temper was more susceptible to the NaCl solution than T6 temper as 7050-T73 alloy exhibited longer fatigue lives in air but shorter ones in the corrosive environment than did 7050-T6 alloy. This higher degree of susceptibility to corrosion fatigue for T73 temper in tests of smooth specimens is attributed to the formation of extensive and larger corrosion pits acting as crack nuclei to facilitate crack initiation. 2. Crack initiation at corrosion pits played a more important role in determining the corrosion fatigue lives of smooth specimens than did the crack propagation stage in the saline solution. This is inferred from the fact that T73 condition exhibited lower fatigue crack growth rates but shorter fatigue lives than did T6 condition in salt water. 3. The stress±corrosion±cracking resistance of 7050-T6 alloy was remarkedly improved by overaging to the T73 temper with an enormous reduction in crack growth rate and increase in time to failure, even though pitting phenomenon was more severe in T73 temper than in T6 temper. Accordingly, the lifetimes for smooth specimens subjected to sustained constant loads in the corrosive environment were predominantly controlled by the crack propagation stage over the crack initiation stage. 4. Overaging the 7050-T6 alloy to the T73 temper improved the corrosion±fatigue±cracking resistance to a less extent than it did the stress±corrosion±cracking resistance, simply because the cracking pattern changed from intergranular to transgranular fracture for corrosion fatigue loading.

AcknowledgementsÐThis work was funded by the National Science Council of the Republic of China under Contract no. NSC-86-2216-E-008-009.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Hatch, J. E., Aluminum: Properties and Physical Metallurgy. American Society for Metals, Metals Park, OH, 1984. Jacko, R. J. and Duquette, D. J., Metallurgical Transactions A, 1977, 8A, 1821±1827. Khobaib, M., Lynch, C. T. and Vahldiek, F. W., Corrosion, 1981, 36, 285±292. Stoltz, R. E. and Pelloux, R. M., Corrosion, 1973, 29, 13±17. Selines, R. J. and Pelloux, R. M., Metallurgical Transactions, 1972, 3, 2525±2531. Bucci, R. J., Engineering Fracture Mechanics, 1979, 12, 407±441. Staley, J. T., Metallurgical Transactions, 1974, 5, 929±932. Speidel, M. O., Metallurgical Transactions A, 1975, 6A, 631±651. Burleigh, T. D., Corrosion, 1991, 47, 89±98. GuÈrbuÈz, R., Doruk, M. and SchuÈtz, W., Journal of Materials Science, 1991, 26, 1032±1038. Pao, P. S., Gao, M. and Wei, R. P., Scripta Metallurgica, 1985, 19, 265±270.

Corrosion fatigue behavior of 7050 aluminum alloys

795

12. ASTM E647-91 Standard, Standard test method for measurement of fatigue crack growth rates. American Society for Testing and Materials, Philadelphia, PA. 13. ASTM E466-82 Standard, Standard practice for conducting constant amplitude axial fatigue tests of metallic materials. American Society for Testing and Materials, Philadelphia, PA. 14. Chen, G. S. and Duquette, D. J., Metallurgical Transactions A, 1992, 23A, 1563±1572. 15. Rebiere, M. and Magnin, T., Materials Science and Engineering A, 1990, 128, 99±106. 16. Bayoumi, M. R. ., Engineering Fracture Mechanics, 1993, 45, 297±307. 17. Corsetti, L. V. and Duquette, D. J., Metallurgical Transactions, 1974, 5, 1087±1093. 18. Dervenis, C. P., Meletis, E. I. and Hochman, R. F., Materials Science Engineering A, 1988, 102, 151±160. 19. Holroyd, N. J. H. and Hardie, D., Corrosion Science, 1983, 23, 527±546. 20. Wei, R. P., Pao, P. S., Hart, R. G., Weir, T. W. and Simmons, G. W., Metallurgical Transactions A, 1980, 11A, 151±158. 21. Forshyth, P. J. E., Stubbington, C. A. and Clark, D., Journal of Institute of Metals, 1961±1962, 90, 238±239. 22. Stoltz, R. E. and Pelloux, R. M., Metallurgical Transactions, 1972, 3, 2433±2441. 23. Scamans, G. M., Journal of Materials Science, 1978, 13, 27±36. 24. Mueller, M., Bernstein, I. M. and Thompson, A. W., Scripta Metallurgicaa, 1983, 17, 1039±1042. 25. Rajan, K., Wallace, W. and Beddoes, J. C., Journal of Materials Science, 1982, 17, 2817±2824. (Received 21 August 1997, in final form 4 November 1997, accepted 7 November 1997)