CC BY
34
y^K
At the electrochemical measurements of the hydrogen permeation through Al membranes, the unusual runs of the permeation transients have been recorded. X-ray analysis of Al membranes cathodically hydrogen charged in alkaline solution showed the metal straining due to the hydrogen ingress. The metal super saturation and the hydride formation have been discussed to be the possible sources of the strain. The experimental conditions to evaluate the hydrogen lattice diffusivity have been considered
HYDROGEN STRAINING EFFECTS IN AL
O.Chernyayeva, D. Lisovytskiy
Institute of Physical Chemistry of the Polish Academy of Sciences ul.Kasprzaka 44/52, 01-224 Warsaw, Poland olgacher@ichf.edu.pl
Introduction
Hydrogen solubility and transport are the key factors in consideration susceptibility to hydrogen embrittlement and to stress corrosion cracking of metals. In the case of Al, the data of hydrogen diffusivity at room temperature obtained by the electrochemical measurements of hydrogen permeation are very scattered. The hydrogen diffusion coefficient ranging is from 10-12 to 10-7 cm2/s [1-5]. Similarly as the data of hydrogen solubility ranges from 10-8 [6] to 10-6 [7] at/at (in equilibrium with gas hydrogen pressure of 105 Pa). The reason might be: (1) the state of Al surface, (2) the specific behavior of Al exposed to the water solution and (3) the complexity of hydrogen transport within Al. The above factors interfere the hydrogen ingress into the metal and its transport within Al and thus affect the calculated hydrogen diffusivity and solubility. Al surface has been usually coated with the oxide films, which cannot be destroyed by the cathodic polarization [8] if no special precautions are applied. On the other hand, cathodic polarization alkalizes the near surface electrolyte and modifies the corrosion processes, which in turn influence the hydrogen ingress in Al. Hydrogen permeation may be also disturbed by the complex hydrogen transport within the membrane due to the action of the elastic strain.
Straining the metal contained the foreign atoms results in their movement toward the expanded region (Gorsky effect [9]). In the case of hydrogen charging, the entering hydrogen atoms produce the self-straining of the metal, which causes the up-hill diffusion of hydrogen already present in the metal (Lewis effect [10]) and manifests itself in transition decrease in hydrogen permeation [10, 11]. Recently the similar effect has been observed at electrochemical measurements of hydrogen permeation through Al [12] and its alloys [13]. In [13] this observation has been accounted for the not specified surface effects. However, the similarity of the phenomena to that described for Pd and its alloys [10, 11] allows attribute it to the complex hydrogen transport within the membrane. Indeed, analysis of the appearance of permeation transients [12, 14] and the additional straining of Al due to cathodic polarization observed in tensile [12,
14, 15] and torsion [16] tests confirmed the hydrogen up-hill diffusion to be involved.
The aim of present work was to recognize the effects of phenomena altering the hydrogen permeation through Al in order to establish the hydrogen diffusivity and hydrogen uptake from the data of electrochemical measurements of hydrogen permeation. The special afford was directed on distinguishing the surface effects and the effects of the complex hydrogen transport and on the evaluation of origin of that complexity.
Experimental procedure
The samples were cut from the pure aluminum sheets, 2.5x10-3 to 2x10-1 cm thick. The preparation of the membranes (1 x 2 cm) for electrochemical measurements of hydrogen permeation in order to remove the air borne surface film and to standardize the surface state of specimens has been described elsewhere [15].
Hydrogen permeation was measured in specially modified double cell [17] enabling to coat the egress side of a membrane with Pd (Figure 1). The egress cell was filled with 0.01M NaOH electrolyte and the anodic potential +150mVHg/HgO was applied to the egress side of the membrane. The above procedure ensured that the anodic current (J), being the measure of hydrogen permeation rate was recorded into the egress cell until the steady state value (J~), under the various conditions in the ingress cell: open circuit potential, application of cathodic polarization (build up permeation transients) and at cessation of cathodic polarization (decay transients). The ingress cell was filled with the 0.01M NaOH test solution. The cathodic polarization of the ingress side of the membrane was done from 20 to 120 mA/cm2.
The X-ray diffraction spectra were recorded before the cathodic polarization of specimens and after their polarization at 100 mA/cm2 in 0.01M NaOH for 30 min. The Siemens D5000 diffractometer and the Cu radiation have been used. To distinguish the effect of the size of the coherently scattered area (crystalline size) and that of the strain on the width
of the diffraction lines, the following equation has been used [18]:
B-cos6 1 4 -o . „ - = — +-sin 6
D X *E
(1)
where: P - line width; 8 - the scattering angle; X - wave length; D - the size of coherently scattered areas; o - internal stress; E - Young modulus.
ingress
egress
0.01N NaOH
+ 150mV
Pd
J
-
Figure 1. Scheme of electrochemical double cell for hydrogen permeation measurements.
The change of the slope of the (Pcos©)/X vs. sin© relationship and the change of the value of (Pcos©)/X corresponding to the sin© = 0 measured after cathodic polarization would reflect the effect of hydrogen charging on the strain and on the crystalline size of Al, respectively.
Results
Hydrogen permeation. After pouring the test electrolyte into the ingress cell and after the attaining the steady state values of the near electrode pH and of the open circuit potential in the ingress cell, the steady state hydrogen permeation (J~ocp) has been recorded in the egress cell, cf.Figure 2a.
on
I
V
> L___✓
/ 2
J
120 180 600 700 time, s
Figure 2. Change of the hydrogen permeation current as recorded in the egress cell after application (on) and after cessation (off) of the cathodic polarization in the ingress cell
(a)
b - inserted are the schemes of (1) the usual hydrogen permeation transient and (2) of the first part of the recorded build-up permeation transient with the marked parameters describing the permeation minimum.
As seen in Figure 2a, at application of cathodic polarization to the ingress side of Al membrane, despite the increase in the hydrogen fugacity, the hydrogen permeation rate did not increase, as has been usually observed (curve 1 in Fig-
ure 2b) but initially decreased and then increased up to the certain steady state value (Jc~). The permeation minimum formed on the build-up transient can be characterized by the time to attain it (tmin) and by its depth (AJ). Cessation of ca-thodic polarization caused the symmetrical effect: hydrogen permeation rate initially increased and then decreased after the attaining some maximum.
At application of higher polarization initially increased the depth of the permeation minimum, similarly as it has been observed for Pd [19]. However, the further increase in polarization (in this case at polarization higher than 40 mA/cm2) permeation minimum decreased and at certain polarization, no permeation minimum has been formed (Figure 3).
Figure 3. Effect of application of cathodic polarization on the build-up permeation transients (Al, 0.1 cm, 0.01M NaOH), were:
- arabic numbers at curves correspond to applied cathodic polarization, mA/cm2; - level "0" corresponds to the steady state hydrogen permeation under the preceding conditions.
For thin membranes, the small permeation maximum was followed by the double, oscillation type minimum. In the case of multiple minimum, the depth and the time to attain the first minimum were taken into account. Relationship between membrane thickness, applied polarization and the depth of the permeation minimum had the complex appearance (Figure 4). With increase in the membrane thickness, the depth of minimum first decreased, then increased and again decreased. The complexity of depth vs. thickness relationship, seen for thin membranes might be attributed to the complex appearance of minimum recorded for those samples. Increase in cathodic polarization decreased the permeation minimum at given polarization and smoothed the "-AJ vs. L" relationship.
<
-10
-20 -30 -40
0.00 0.05
0.10 L, cm
/ „ o
o-
^—■
■ /
40mA/cm2
-0- 60mA/cm2
. A . 80mA/cm2
100mA/cm2
0.15 0.20
Figure 4. Effect of the membrane thickness and the level of cathodic polarization (shown with the Arabic numbers) on the depth of the hydrogen permeation minimum
b
a
X-ray analysis. No hydride phase but the modification of the X-ray spectra have been observed in the hydrogen charged specimens. As follows from the effect of hydrogen charging on the width of consequent lines (Figure 5) the crystalline size decreased and the internal stress increased due to the cathodic hydrogen charging of Al.
0.150 0.125
c<
0.100
® m
O 0.075
o cm
^ 0.050
0.025
0.000 0.
Figure 5. Dependence of (Pcos0)/X vs. sin© for diffraction lines measured before and after cathodic polarization (iC = 100 mA/cm2) of Al.
Discussion
Cathodic polarization of Al may decrease the hydrogen ingress into the metal due to suppression the corrosion processes or due to increase in the near electrode pH. However, promoting of corrosion and increase in the near electrode pH by the addition of NaOH has been found to increase the hydrogen permeation, similar as in the case of cessation of cathodic polarization. Therefore, near electrode processes in electrolytes are not the reason for unusual hydrogen permeation [15].
The reason for the observed hydrogen permeation transients might be the complex hydrogen transport within the membrane affected by the up-hill diffusion. Hydrogen is present in metal before the application of cathodic polarization as shows hydrogen permeation recorded at open circuit potential, Jocp~ in Figure 2a. Cathodic polarization produces the straining of metal, as been undoubtedly confirmed in tensile tests [14]. The strain propagation has been always faster than diffusion. Therefore the conditions for the hydrogen up-hill diffusion [10, 15] are fulfilled and it can be taken into account. The summary hydrogen flux J, consisting of the initial flux Jocp~, the flux Jin originated from the gradient of hydrogen concentration (inward one) and the flux Jb originated from the gradient of strain (backward one) may be written as:
J = Jocp~ + Jin - Jb (2)
According to analysis done in [11], the short time permeation is governed by the backward flux, whereas the inward flux dominates the long time permeation.
The unusual hydrogen permeation has been characterized, beside the symmetry in the appearance of build-up and decay transients (Figure 2a) and reversibility of the effect by the following experimentally observed features:
• formation of the transition maximum and the multiple minimum;
• influence of membrane thickness on the effect;
• diminishing and vanishing of the effect with the increase in polarization.
In Lewis effect [10], the driving force for the up-hill diffusion is the expansion of metal lattice by the entering hydrogen atoms, which mostly depends on the hydrogen concentration. Since the solubility of hydrogen in Al is much lower than in Pd (CH = 10-2 at/at [20]) the elastic strain introduced by the dissolved hydrogen in Al is not enough to produce the Lewis effect, as discussed in [12]. However, the metal straining might be also produced by the solid phase layer, adherent to the metal surface. The surface Al2O3 layer producing the metal straining high enough to affect the hydrogen transport cannot be the source of the observed effects since it forms at anodic polarization [21] and has been not influenced by the cathodic polarization [8]. However, at cathodic polarization of Al in alkaline solution, the phase thermodynamically stable is AlH3, which transforms to Al(OH)3 at cessation of cathodic polarization [22, 23]. As has been analyzed in [14], the hydride can produce the mismatch strain, similar to that imposed by the oxides. On the other hand, hydroxide has the gel nature [23] and thus, cannot strain the metal lattice. Therefore, the formation and decomposition of hydride at application and at cessation of cathodic polarization of Al, producing straining and relaxation of the ingress surface of metal have been proposed [14] to be responsible for the symmetry of the build-up and decay permeation transients and for the reversibility of the hydrogen permeation minimum.
The up-hill diffusion causes the relaxation of strain [24], which in turn affects the redistribution of the mobile species within the metal [9]. This coupling of the strain induced hydrogen transport and the hydrogen induced strain relaxation produces the oscillation of minimum, which has been theoretically predicted in [11] and observed in work [14]. The more thin membranes, the more obvious oscillation and the shorter period of oscillation have been recorded. The observations agree with the data [25] showing the increase in the oscillation frequency and the decrease in the damping with decrease in the thickness of vibrating membrane. The formation of small initial permeation maximum in the case of thin membranes (Figure 4) has been explained [19] by mechanically induced bending of membrane and contraction of internal volume causing the additional egress of hydrogen.
The shallow permeation minima formed for thin membranes may be attributed to the formation of complex minimum in the case of such membranes. With the increase in the membrane thickness, the effect of disturbance of the hydrogen transport at the ingress surface of the membrane on the permeation transients became less pronounced (Figure 4).
At increased cathodic polarization the hydrogen uptake and thus, the imposed strain should increase and the more pronounced permeation minimum should be formed. This indeed has been observed at the increase in cathodic polarization from 20 to 40 mA/cm2 (Figure 3). However, the further increase in polarization decreases the depth of the permeation minimum for all the studied membranes (Figure 4). It should be taken into account that the increase in the elastic stress is limited by the plastic deformation of the metal whereas the amount of entering hydrogen increases at increased cathodic polarization. Therefore, at high cat-hodic polarization the inward hydrogen flux prevails over the backward one and suppresses it, which results in the less pronounced permeation minimum and even in its vanishing (Figure 3). The possibility of the plastic deformation of Al
sin ©
due to the cathodic polarization has been proved by the increase in the dislocation density within the surface layer of Al, polarized in H2SO4 solution [26]. The change of crystalline size and of the internal stress observed in present work ( Figure 5) also confirms the plastic deformation of the surface layer at application of high cathodic polarization.
The complex hydrogen transport disturbs the hydrogen permeation through the membrane and the hydrogen diffusion coefficient cannot be calculated by simplified method. In such a case, the equations taking into account the summary hydrogen flux [11, 14, 15] should be used. However, the complexity of those equations and the simplifications done at their derivation do not allow their use for evaluation of hydrogen diffusivity from experimental data.
Since the Fick's flux dominates the hydrogen permeation for a longer time [11], the part of the buildup permeation transigen, close to the may be used to calculate the hydrogen diffusivity. In [27], the application of graphical methods taking into account the Fourier and the Laplace approaches, which provide the reasonable accuracy for 0.98J» >J > 0.3J» has been thoroughly discussed. As analyzed in [15], the Fourier method allows the more precise graphic determination of diffusion coefficient, which has been derived from the slope of ln(1-J/J~) against the time (t) relationship:
ln(1 - J) -
nD,
L2 1
(3)
The calculated hydrogen diffusion coefficients and those taken from the literature reveal the similar dependence: they decrease with the decrease in the membrane thickness (Figure 6). This suggests that the material near the ingress surface produces the hampering effect on the hydrogen transport.
^ 0.100 ®
M
O 0.075 O
oa
0.25 0.50
sin ©
Figure 6. Effect of the membrane thickness on the calculated hydrogen diffusion coefficient.
The retardation of hydrogen transport and the increase in the content of permeable hydrogen within the near surface metal reveal the interaction of permeated hydrogen with the traps. The traps are created at cathodic polarization, as confirmed the X-ray analysis (Figure 5) revealing the metal distortion and the formation of internal stresses. Therefore, the formation of the hydrides or at least the high supersaturation of the surface layers of Al with hydrogen may be expected at application of the high cathodic polarization. Since the difference in the lattice parameters of AlH3 and Al, the high stresses and the plastic deformation of Al can be expected due to the hydride formation. For the same reason, the process of hydride formation cannot occur over the all surface or propagate to the deep parts of the metal. Therefore, not the hydride phase, but the formation of the patched hydride-like
surface layer with the highly stressed metal islands can be postulated. The similar straining effect and the metal plastic deformation may also produce by high supersaturation of Al with hydrogen if hydrogen concentration stresses [28] overcome the elastic limit. However, although the hydride phase has not been yet confirmed, the assumption of its formation seems to be reasonable. According to [22, 23] the hydride phase transforms to the gel hydroxide one providing reversibility of the. No such an effect may occur in the case of the supersaturated material.
Conclusions
1. The unusual runs of hydrogen permeation rate noticed upon the application and the cessation of cathodic polarization of Al in alkaline solution are accounted for by the complex transport of hydrogen within the membrane.
2. The hydrogen up-hill diffusion affecting the hydrogen flux is caused by the strain introduced into the membrane at cathodic polarization, confirmed in tensile tests.
3. The source of the strain has been suggested to be the formation of the patchy, hydride like layer on the Al surface at the application of cathodic polarization in alkaline solution, which decomposes at the cessation of cathodic polarization.
4. To calculate the hydrogen diffusivity from the electrochemical measurements of hydrogen permeation, the phenomena affecting the experimental results should be determined in order to apply the experimental conditions providing undisturbed diffusion transport of hydrogen. The metal straining should be also considered at the discussion of the hydrogen embrittlement and the stress corrosion cracking of Al.
References
1. Ishikawa T., McLellan R.B. Acta metal, 1986, 34, 1091.
2. Hashimoto H., Kino T. J. Phys. F: Metal Phys, 1983, 13, 1157.
3. Choo W.Y., Bernstein I.M., Met. Trans, 1984, 15A, 1953.
4. Chu Youngchang, Yin Heyun, Latanisition R.M. Hydrogen diffusivity in pure aluminum by using electrochemical hydrogen permeation, Proc. 4th International Conference Hydrogen and Materials, China, 1988, p.114.
5. Braun R., et al. Hydrogen Transport and Cracking in Meta-ls.In: Turnbull A. ed. Inst. of Mater., Teddington, 1994.
6. Bond G.M., Robertson I.M., Birnbaum H.K. Acta Met, 1988, 36, 2191.
7. Eichnauer W., Hattenbach K., Pebler A. Z. Metallkunde, 1961, 52, 682.
8. Moon S-M., Pyun S-I. Corrosion Sci, 1997, 39, 399.
9. Gorsky W.S. Phys Z Sowjetunion, 1935, 8, 451.
10. Lewis F.A., Magennis J.P., McKee S.G., Ssebuwufu P.J.M. Nature, 1983, 306, 673.
11. Baranowski B. Diffusion in Elastic Media with Stress Fields. In: Sieniutycz S., Salomon P., editors. Adv. In Thermodyn., Flow, Diffusion and Rate Processes. Taylor and Francis: NY, 1992, p.168.
12. Lunarska E., Chernyayeva O. Proc. International Conference Environmental Degradation of Engineering Materiale „EDEM 2003", Bordeaux, France 2003. PH10.
0.75
13. Danielson M.J. Corros Sci, 2002, 44, 829.
14. Lunarska E., Chernyayeva O., Int. J. Hydrogen Energy 2006, 31, 237.
15. Lunarska E. Chernyayeva O. Hydrogen diffusivity and straining effect at cathodic polarization of Al in NaOH solution. In: Shipilov S.Jones R.H., Olive J.-M., Rebak R.B. eds. Environment-Induced Cracking of Materials. Volume 1: Chemistry, Mechanics and Mechanisms. Elsevier: Amsterdam-London-N.Y., Oxford, 2008, 1, p. 227.
16. Lunarska E., Chernyayeva O., Spivak L., J. Alloys and Compounds, 2005, 404-406, 269.
17. Devanathan V.M.A., Stachurski Z., J Electrochem Soc, 1964, 111, 619.
18. Taylor A. X-ray metallography. John Wiley & Sions Inc., New York-London, 1961.
19. Dudek D., Collect. Chech. Chem. Commun, 2005, 70, 113.
20. Peisl H. Hydrogen Induced Lattice Deformation. In: Alefeld G., Volkl J., editors. Hydrogen in Metals 1. Basic Properties. SpringerVerlag Berlin Heidelberg New Jork, 1978, p.69.
21. Bradhurst D.H., Llewelyn Leach J.S., J Electrochem Soc, 1966, 113, 1245.
22. Perrault G.G., J Electrochem Soc, 1979, 126, 199.
23. Antropov L.I., Vrjosek G.G., Fateev Ju.F., Zashchita metallov, 1975, XI, 300.
24. Konobejevskij S., Zhurnal eksperimentalnoj i teoreticheskoj fiziki, 1943, 6, 200.
25. Su C.M., Wen Y., Wutting M., Journal de Physique IV, 1996, 6, C8-757.
26. Watson J.W., Shen Y.Z., Meshii M., Metallurgical Transaction A, 1988, 19A, 2299.
27. Turnbull A., Saenz de Santa Maria M., Thomas N.D., Corrosion Sci, 1989, 29, 89.
28. Goltsov V., Ryumshina T., Smirnov L., Glukhova Zh., Kotelva R., Theory of hydrogen elasticity phenomenon, In: Goltsov V. editor. Progress in Hydrogen Treatment of Materials. Intern. Association for Hydrogren Energy, Donetsk - Coral Gables, 2001, p. 95.
