Measurement and Modeling of Crack Conditions during the Environment-assisted Cracking of an Al-zn-mg-cu Alloy

Although it is well known that Al-Zn-Mg-Cu alloys are susceptible to intergranular environment-assisted cracking (EAC) in some temper conditions, the relative contribution of hydrogen embrittlement (HE) and anodic dissolution (AD) to the crack advance process remains uncertain. This work reports the results of combining experimental and modeling data of crack potentials to help elucidate the role of AD to the EAC process. Micro-reference electrodes were used to make in situ measurements of the crack potential during Stage II cracking of AA 7050. The crack tip potential was –0.85 ± 0.05 VSCE; near-tip potential gradients were large, ∼ 1 V/cm. Modeled crack potential distributions were strongly dependent on the value of the crack tip opening and the presumed presence or absence of a resistive saltfilm at the crack tip. Although quantitative assessment of the contribution of AD to crack propagation is confounded by a lack of knowledge of the tip dissolution area, AD cannot be excluded as playing a significant role in the crack advance process of AA 7050. Introduction High-strength Al-Zn-Mg-Cu alloys are susceptible to intergranular (IG) environment-assisted cracking (EAC) in aqueous chloride [1]. However, the relative contribution of hydrogen embrittlement (HE) and anodic dissolution (AD) to the crack tip damage process during aqueous cracking are not known [1-3]. Mechanistic understanding is lacking, in part because the crack’s tip geometry, chemical and electrochemical conditions are not well characterized. The goal of this research was to characterize the crack potential for an EAC-susceptible 7000series aluminum alloy (AA) and the associated crack tip geometry and environment to help elucidate the contribution of AD to the crack propagation process. This objective was addressed by measurement of the crack potential via in situ reference electrodes during crack growth tests acquired under mechanical and electrochemical control combined with modeling of the steady-state crack potential distribution. A one-dimensional (1-D) model was developed to assess the influence of crack geometry and tip and wall dissolution rate. Comparison of model results to measured profiles allows a determination of the applicability of various modeling assumptions. In this work, the effects of assumed tip geometry and salt film properties are presented. Metal salts have been implicated as playing a significant role in the localized corrosion and cracking in some alloy-environment systems. Modeling work of Simonen et. al. [4] predicted that a resistive NiCl2 film forms at the tip of long cracks of Ni-base alloys at the EAC threshold stress intensity. Artificial pits were used to determine the conductivity (κ) of NiCl2 films, κ = 1.8x10 1/Ω-cm [5] and in a following study, the measured Ni-salt conductivity gave the best prediction of the Stage I crack growth rate-stress intensity dependence when compared to experimental observations [4]. Pitting of high-purity Al at an applied electrode potential (EApp) between –0.70 and –0.52 VSCE was characterized by the formation of an aluminum oxychloride film (Al(OH)2Cl and Al(OH)Cl2) on the electrode surface [6, 7] These transitory compounds were also reported by Foley [8] to exist in pits of Al. Beck [9, 10] noted that a high resistance, bi-layer salt film covered the electrode surface at high anodic potentials (> +4 VSCE) in concentrated AlCl3 solutions. The salt film was proposed to consist of anhydrous AlCl3 next to the metal surface and an outer hydrated layer (AlCl3⋅6H2O), the total salt film thickness is estimated to be ∼ 100 nm for an overpotential of 1 V and metal dissolution rate of 0.1 A/cm. Note that this dissolution rate would correspond to a crack growth rate of 3.3x10 mm/s. Beck [9] reported that although Al salt films may form at EApp < +4 VSCE and low salt concentrations, they are not stable due to hydrogen bubble-induced mixing within the cavity. However, cracks have more restricted mass transport than pits and thus a tip film may be more stable at the lower current densities, EApp, or in less concentrated bulk solutions than observed for pits. In the present work, the effects of tip film conductivity and thickness on the ECrack profile was examined. Experimental Material Thick (152 mm) plate of Aluminum Association (AA) 7050 was provided by the Aluminum Company of America (Alcoa) and heat treatments were performed at the Alcoa Technical Overpotential is defined as the difference between the actual potential measured (or applied) and the reversible potential. Center. The chemical composition was Al-6.09Zn-2.14Mg-2.19Cu-0.11Zr-0.05Si-0.09Fe (wt. %). Blanks were resolutionized for 2 hr at 466°C, cold water quenched, stretched to relieve residual stresses, heated for 4 hr at 121°C for microstructure stabilization, and subsequently aged 20 hr at 118°C plus 12 hr at 154°C. This heat treatment simulates the peak aged (T651) condition. Some material was aged using single-step, non-commercial heat treatments consisting of aging for 6 or 12 hr at 163°C after the stabilization treatment. Material aged 6 hr at 163 °C was similar to T651 in its mechanical and EAC behavior where as the 12 hr aging treatment at 163°C produced a material with mechanical and EAC properties between that of T651 and overaged, EAC-resistant T7451 [11]. The resulting microstructure was partially recrystallized, with an elongated grain structure of variable size but typically 100 μm (S) x 200 μm (T) x 600 μm (L) [11]. Mechanical properties for T651 material were: σYS = 530 MPa (S), Elongation = 8 % (S), E = 66 GPa and KIC = 21.5 MPa√m (S-L) [11]. Aqueous EAC Testing and Crack Potential Measurements Wedge-open loaded (WOL) specimens (12.7 mm thickness x 52.3 mm width x 50.8 mm height), machined in the S-L orientation with the crack plane at the T/4 position in the plate, were anodized 48 hr at +0.30 VSCE in 0.5 M Na2CrO4, natural pH. Fracture specimens were clamped in a plastic chamber to expose the crack and notch to the 0.5 Na2CrO4 + 0.05 M NaCl, pH 9.2 test solution. Figure 1 illustrates the test apparatus. The fracture mechanics procedures are described in detail elsewhere [11, 12]. The low [Cl]/high [CrO4] electrolyte permitted application of a range of EApp without initiation of stable pitting (pitting potential, EPit ≈ -0.230 VSCE for AA 7050-T651) and limited crack wake corrosion which facilitated post-test fractography. The alkaline bulk solution was also desirable for studying changes in crack solution pH [13]. A 3-electrode configuration and potentiostat were used to control the potential of the grounded WOL specimen. WOL specimens were corrosion fatigue pre-cracked (10 Hz, Kmax = 10 MPa √m, R = 0.3) in the test solution at –0.675 VSCE, then rapidly loaded to a fixed crack mouth opening displacement (CMOD) in a computer-controlled servohydraulic test frame. During EAC growth, initial and final stress intensities (K) were ≈ 1415 and 9-12 MPa√m, respectively. Crack length was calculated via an elastic compliance solution [14, 15] and corrected to the optically measured initial and final crack length [11]. Crack growth rate was determined by linear regression of crack length vs. time data. To reservoir From pump Apply CMOD Controlling reference electrode Plexiglass chamber Micro-reference electrode Figure 1: Schematic of aqueous EAC test-cell illustrating location of micro-reference electrodes for in situ crack potential measurements. Just prior to fatigue pre-cracking, 1.0 mm (Microelectrodes Inc., NH, USA) or 1.6 mm (Diamond General Development Corp., MI, USA) diameter reference electrodes were inserted into holes in the top face of the fracture specimen such that the electrode’s sensing interface coincided with the crack plane. Figure 1 illustrates the position of the reference electrode used to measure the crack electrode potential (ECrack). Reference electrode signals, shielded from ambient electromagnetic interference, were conditioned through a high input impedance (100 GΩ) operational amplifier and recorded with a computer controlled data acquisition system at a rate of up to 1 point/s. Crack Potential Modeling Measured crack potential profiles were modeled in 1-D in order to assess the parameters that most influence the crack potential distribution. Parameters examined included crack tip geometry, crack tip and flank current density, and solution conductivity, the latter being a function of the presumed presence or absence of a precipitated film at the tip. Model Outline. The quasi steady-state potential distribution was estimated using Ohm’s Law which states that a potential change (∆E) results from the passage of current (I) through a medium of finite resistance (R),