Hydrogen absorption in plasma-nitrided iron

Hydrogen absorption in plasma-nitrided iron

Acta Materialia 54 (2006) 1525–1532 www.actamat-journals.com Hydrogen absorption in plasma-nitrided iron Z. Wolarek, T. Zakroczymski * Department o...

455KB Sizes 0 Downloads 1 Views

Acta Materialia 54 (2006) 1525–1532 www.actamat-journals.com

Hydrogen absorption in plasma-nitrided iron Z. Wolarek, T. Zakroczymski

*

Department of Electrochemistry, Corrosion and Applied Surface Sciences, Institute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka 44-52, 01-224 Warsaw, Poland Received 30 May 2005; received in revised form 10 November 2005; accepted 10 November 2005 Available online 18 January 2006

Abstract Hydrogen absorption by membranes prepared from various parts of plasma-nitrided iron specimens was studied using electrochemical techniques. The membranes were charged with hydrogen by galvanostatic cathodic polarisation in 0.1 M NaOH. After cessation of hydrogen charging, the desorption rate of hydrogen was measured at both sides of the membrane. An analysis of the desorption rates enabled to one distinguish the amounts (concentrations) of the diffusible and the reversibly trapped hydrogen in the nitrided compound layer, the diffusion zone and the unnitrided iron substrate. Trapped hydrogen accounted for the majority of the total absorbed hydrogen. Hydrogen trapping was enhanced by nitride precipitate–iron matrix interfaces, and especially by the nitrided compound layer.  2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Iron; Plasma nitriding; Hydrogen; Diffusion; Solubility

1. Introduction Plasma nitriding involves chemical and metallurgical changes of surface/near surface regions. Nitrided layers on a substrate, besides their beneficial wear properties, may reduce hydrogen absorption by the substrate metal. This effect may be considered in terms of hindering the entry of hydrogen atoms into the outer nitrided layer (the surface effect), and hindering the transport of hydrogen through the whole nitrided layer (the barrier effect). The barrier effect was studied previously with reference to nitrided iron [1,2], carbon steels [3,4] and low-alloy highstrength steel [5,6] using the electrochemical permeation technique [7]. In our previous work [8] the diffusivity of hydrogen and the lengths of diffusion paths in each layer of nitrided iron were evaluated. It was found that the diffusivity of hydrogen in the outer compound layer of nitrides is relatively low (Dc = 1.9 · 108 cm2/s), being more than 4000 times lower than that in the unnitrided *

Corresponding author. E-mail address: [email protected] (T. Zakroczymski).

iron substrate (Ds = 8.0 · 105 cm2/s). The length of the diffusion paths in the compound layer and in the substrate may be identified with their thickness. It was also shown that the transport of hydrogen in the nitrided diffusion zone with dispersed Fe4N precipitates occurs mainly through the ferrite matrix and can be characterised by a diffusion coefficient close to that for the unnitrided iron. However, hydrogen has to bypass the nitride precipitates and, therefore, the real diffusion paths have different lengths and they may be longer than the thickness of a membrane. Aside from the transport of hydrogen, the absorption of hydrogen is the other important process characterising the behaviour of hydrogen in metals. With reference to nitrided iron under given charging conditions, this process may be roughly described by the total amounts (concentrations) of hydrogen absorbed in each layer. In turn, the total amount of hydrogen in a given layer includes the so-called ‘‘diffusible’’ hydrogen which occupies interstitial sites, being the solid solution component, as well as the ‘‘trapped’’ hydrogen associated with microstructural defects such as dislocations, grain boundaries, non-metallic inclusions and other internal interfaces.

1359-6454/$30.00  2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actamat.2005.11.018

1526

Z. Wolarek, T. Zakroczymski / Acta Materialia 54 (2006) 1525–1532

Nomenclature C0 Cd,0

Cd Cd,i

CHd,av CHt,av CH,av Dc Ds Dd,eff Dd,x

D(c + d)

ic i1 p i0p iH,L iHd,L iH,0

hydrogen concentration at the entry side of a membrane hydrogen concentration at the entry side (X = 0) of the membrane prepared from the nitrided diffusion zone hydrogen concentration in the membrane prepared from the nitrided diffusion zone section-dependent hydrogen concentration in the membrane prepared from the nitrided diffusion zone average concentration of the diffusible hydrogen average concentration of the trapped hydrogen average concentration of hydrogen (CHd,av + CHt,av) hydrogen diffusivity in the nitrided compound layer hydrogen diffusivity in the unnitrided iron substrate effective hydrogen diffusivity for the membrane prepared from the nitrided diffusion zone section-dependent hydrogen diffusivity for the membrane prepared from the nitrided diffusion zone effective hydrogen diffusivity for the membrane comprising the nitrided compound and diffusion layer cathodic (charging) current density steady-state permeation rate of hydrogen initial steady-state permeation rate of hydrogen desorption rate of hydrogen at the membrane exit side (X = L) desorption rate of the diffusible hydrogen at the membrane exit side desorption rate of hydrogen at the membrane entry side (X = 0)

The aim of the present work was to investigate the absorption of hydrogen in nitrided Armco iron using electrochemical techniques. The standard electrochemical permeation technique [7] was used to measure the permeation of hydrogen through a membrane and to identify the diffusible hydrogen. The electrochemical desorption technique [9,10], applied to both sides of a previously charged membrane, was used to determine the total amount of absorbed hydrogen, and hence the amount of reversibly trapped hydrogen. Efforts were made to ascribe the trapped hydrogen to particular kinds of traps. 2. Experimental The material used was a commercial Armco iron (C 0.031, Mn 0.12, Si 0.01, P 0.008, S 0.015, Cr 0.01, Ni 0.01, Cu 0.02, Al 0.03 wt.%) in the form of a hot-drawn,

iHd,0 k L Lc Ls Ld Ld,i L(c + d) La,d qH,L qHd,L qHt,L qH,0 qHd,0 qHt,0 qHd qHt qH t X

desorption rate of the diffusible hydrogen at the membrane entry side number of partial membranes thickness of a membrane thickness of the nitrided compound layer thickness of the membrane prepared from the unnitrided iron substrate thickness of the membrane prepared from the nitrided diffusion zone thickness of the partial membrane (slice) thickness of the membrane comprising the 2nitrided compound and diffusion zone length of diffusion paths in the ferrite matrix of the diffusion zone amount of hydrogen leaving a membrane at its exit side (X = L) amount of the diffusible leaving a membrane at its exit side amount of trapped hydrogen leaving a membrane at its exit side amount of hydrogen leaving a membrane at its entry side (X = 0) amount of diffusible hydrogen leaving a membrane at its entry side amount of trapped hydrogen leaving a membrane at its entry side total amount of diffusible hydrogen (qHd,L + qHd,0) total amount of trapped hydrogen (qHt,L + qHt,0) total amount of hydrogen (qHd + qHt) time distance

25 mm diameter rod. Plate specimens, 3 mm thick, were machined out from the rod perpendicular to its axis. The specimens were plasma nitrided on one side in a gas mixture of 80% N2 + 20% H2 under a pressure of 670 Pa at temperature of 540 C for 6 h, and cooled slowly in the nitriding chamber. This treatment produced a modified layer of about 1.3 mm thick. The modified layer consisted of a thin, about 10 lm, outer compound layer composed of nitrides Fe2–3N (e-phase) and Fe4N (c 0 -phase), and an inner, much thicker, diffusion zone with dispersed Fe4N precipitates and a solid solution of nitrogen in the iron matrix. Scanning electron microscopy images of the etched cross-sectional surface of the nitrided specimens and the method of preparing various membranes by abrading the nitrided specimen can be found elsewhere [8]. The following membranes were prepared:

Z. Wolarek, T. Zakroczymski / Acta Materialia 54 (2006) 1525–1532

1527

 One-layer membranes of thickness Ls = 0.4, 0.6, 0.8 and 1 mm, comprising only the unmodified iron substrate.  One-layer membranes of thickness Ld = 0.4, 0.6, 0.8 and 1 mm, comprising various sections of the diffusion layer. In this case, the compound layer was removed and the specimen was abraded from the other side to the corresponding depth.  Two-layer membranes of thickness L(c + d) = 1 mm, comprising the compound layer and a substantial part of the diffusion zone. The membranes were mounted between two electrochemical cells; the circular area exposed to solutions was 0.5 cm2. Initially, the membranes were charged at one (entry) side with hydrogen cathodically generated from 0.1 M NaOH at a current density ic = 20 mA/cm2. Depending on the kind of membrane, the entry side corresponded to the surface of the unnitrided iron substrate, to the outer surface of the compound layer or to the surface of the diffusion layer just beneath the compound layer. The membrane exit side, previously coated with a thin layer of Pd, was polarised at a constant anodic potential of 0.15 VHg|HgO|0.1 M NaOH (0.32 VNHE). The anodic current was a measure of the permeation rate of hydrogen. After 24 h of charging, the cathodic current was switched off and the entry side of the membrane was polarised immediately at the same potential of 0.150 VHg|HgO|0.1 M NaOH as the exit side. Subsequently, the anodic currents were recorded at both sides of the membrane. After subtraction of the background currents, the desorption rates of hydrogen were obtained. The permeation and desorption measurements were carried out at 30 C. Each measurement was repeated 2–4 times, and the evaluated values of hydrogen content are shown as the mean values with the standard error of the mean. 3. Results 3.1. Unnitrided iron substrate membranes 3.1.1. Desorption of hydrogen at the membrane exit side (X = L) The results for the desorption rate of hydrogen (iH,L), measured at the exit side of the 1 mm thick membrane are shown in Fig. 1(a). This desorption rate (continuous line) consists of the desorption rate of the diffusible hydrogen (iHd,L) and the desorption rate of hydrogen originating from traps (iHt,L). Since the diffusible hydrogen is able to move through the metal according to the laws of diffusion, the desorption rate iHd,L may be expressed as [11] ( !) 2 1 2Ls X ð2n þ 1Þ L2s 0 iHd;L ¼ ip 1  pffiffiffiffiffiffiffiffiffi exp  ; ð1Þ 4Ds t pDs t n¼0 where i0p is the steady-state permeation rate, just at the beginning of desorption (t = 0), Ls is the membrane thickness and Ds is the hydrogen diffusivity in the unnitrided

Fig. 1. An analysis of the desorption rate of hydrogen at the exit (a) and entry (b) side of the membrane comprising only the unmodified iron substrate.

iron. For a given i0p , knowing Ds [8] and Ls, one can reproduce the desorption rate of the diffusible hydrogen iHd,L (dashed line in Fig. 1(a)). The area under the iH,L curve corresponds to the total amount of hydrogen that left the membrane at its exit side, qH,L, and the area under the iHd,L curve determines the amount of the diffusible hydrogen, qHd,L. Consequently, the difference qH,L  qHd,L gives the amount of trapped hydrogen, qHt,L (Table 1). 3.1.2. Desorption of hydrogen at the membrane entry side (X = 0) A similar procedure may be applied to the desorption of hydrogen from the membrane at its entry side. In this case, the desorption rate of the diffusible hydrogen is given by [9] ( "  2 2 1 X Ls n Ls 0 iHd;0 ¼ ip 1 þ pffiffiffiffiffiffiffiffiffi exp  Ds t pDs t n¼0 !#) 2 1 X ðn þ 1Þ L2s þ exp  . ð2Þ Ds t n¼0 The measured desorption rate iH,0 and the reproduced rate iHd,0 are shown in Fig. 1(b). Integration of these curves leads to the amounts of hydrogen: qH,0 and qHd,0, and hence qHt,0 (Table 1). Finally, the total amounts of the diffusible hydrogen (qHd = qHd,L + qHd,0), the trapped hydrogen (qHt = qHt,L + qHt,0) and the hydrogen stored in the membrane before the interruption of charging (qH = qHd + qHt) can

1528

Z. Wolarek, T. Zakroczymski / Acta Materialia 54 (2006) 1525–1532

Table 1 Amounts of hydrogen desorbed from various iron membranes at their exit and entry side and divided into the amounts of the diffusible and trapped hydrogen Membrane thickness (mm) Unnitrided iron substrate membrane

Ls

0.4

0.6

0.8

1.0

Nitrided diffusion zone membrane

Ld

0.4

0.6

0.8

1.0

Nitrided compound + diffusion zone membrane

Lc + Ld

0.01 + 0.99

Hydrogen content, q · 108 (mol/cm2) qH,L qH,0 qH qH,L qH,0 qH qH,L qH,0 qH qH,L qH,0 qH

4.0 ± 1.5 30.7 ± 4.0 34.7 ± 5.5 7.4 ± 2.6 39.9 ± 4.7 47.3 ± 7.3 7.4 ± 4.1 44.6 ± 5.1 52.0 ± 9.2 14.2 ± 4.8 51.3 ± 7.7 65.5 ± 12.6

qH,L qH,0 qH qH,L qH,0 qH qH,L qH,0 qH qH,L qH,0 qH

23.4 ± 4.2 62.2 ± 9.2 85.6 ± 13.4 29.9 ± 3.7 75.2 ± 10.1 105.1 ± 13.8 35.9 ± 4.6 87.0 ± 11.1 122.9 ± 15.7 48.8 ± 4.9 100.4 ± 9.8 149.2 ± 14.7

qH,L qH,0 qH

0.8 ± 0.2 64.6 ± 12.1 65.4 ± 12.3

qHd,L qHd,0 qHd qHd,L qHd,0 qHd qHd,L qHd,0 qHd qHd,L qHd,0 qHd

0.23 ± 0.09 0.47 ± 0.18 0.70 ± 0.27 0.37 ± 0.11 0.74 ± 0.23 1.10 ± 0.34 0.47 ± 0.13 0.93 ± 0.26 1.40 ± 0.39 0.65 ± 0.17 1.30 ± 0.33 1.95 ± 0.50

qHt,L qHt,0 qHt qHt,L qHt,0 qHt qHt,L qHt,0 qHt qHt,L qHt,0 qHt

3.8 ± 1.4 30.2 ± 3.8 34.0 ± 5.2 7.0 ± 2.5 39.2 ± 4.5 46.2 ± 7.0 6.9 ± 4.0 43.7 ± 4.8 50.6 ± 8.8 13.5 ± 4.6 50.0 ± 7.4 63.5 ± 12.0

qHd

8.3 ± 2.2

qHt

77.4 ± 11.2

qHd

12.1 ± 3.7

qHt

93.0 ± 10.1

qHd

14.4 ± 3.5

qHt

108.5 ± 12.2

qHd

15.2 ± 4.1

qHt

134.0 ± 10.6

qHd,L qHd,0 qHd

0.10 ± 0.02 0.20 ± 0.04 0.29 ± 0.06

qHt,L qHt,0 qHt

0.7 ± 0.2 64.4 ± 12.1 65.1 ± 12.2

be evaluated. The above procedure was applied to the other membranes with different thickness and the corresponding amounts of hydrogen are given in Table 1. 3.2. Nitrided diffusion zone membranes As an example, the measured desorption rates of hydrogen from the membrane of thickness Ld = 1 mm prepared from the diffusion zone of the nitrided specimen is shown in Fig. 2(a) and (b), for the membrane exit and entry side, respectively. The areas under the iH,L and iH,0 curves correspond to the amounts of hydrogen qH,L and qH,0 which left the membrane at its exit and entry side, respectively. Their sum gives the total amount of hydrogen stored in the membrane before desorption, qH (Table 1). The mathematical approach based on Eqs. (1) and (2), successfully used to distinguish the amount of diffusible hydrogen from the amount of trapped hydrogen in the unnitrided iron substrate, cannot be applied for the membrane containing the nitride precipitates, because the real diffusion paths through the ferrite matrix, La,d, are not equal to the membrane thickness L. It was revealed earlier [8] that the diffusion paths have various lengths, always longer than the membrane thickness L. As a result of different La,d, the effective (apparent) diffusion coefficient of hydrogen, Dd,eff, evaluated from the permeation transients

Fig. 2. Desorption of hydrogen at the exit (a) and entry (b) side of the membrane comprising a section of the nitrided diffusion zone.

Z. Wolarek, T. Zakroczymski / Acta Materialia 54 (2006) 1525–1532

1529

and taking the membrane thickness as the diffusion distance, depends on the membrane thickness. This coefficient characterises a given membrane as a whole. In turn, the dependence Dd,eff on L enabled the determination of the so-called section-dependent diffusivity, Dd,x, which characterises the transport of hydrogen at a given section of the diffusion zone. Taking into account the Dd,eff  L and Dd,x  X dependences given in Ref. [8], the following approach to evaluate the amounts of the diffusible hydrogen is proposed. The steady-state permeation rate i1 p through the nitrided diffusion zone membrane of thickness L = Ld is given by Fick’s first law i1 p ¼ F

Dd;eff C d;0 ; Ld

ð3Þ

where Cd,0 is the hydrogen concentration in the membrane at its entry side (X = 0) and F is the Faraday constant. Knowing experimental values of i1 p and Dd,eff, the hydrogen concentration Cd,0 can be calculated for each membrane. In order to evaluate the real concentration of hydrogen inside the membrane at distance X, the membrane may be divided into k partial membranes (slices) of thickness Ld,i (where i = 1, 2, 3, . . . , k), and a pertinent value of Dd,x can be ascribed to each partial membrane. According to the flux continuity, the steady-state permeation rate of hydrogen can be written as i1 p ¼

Dd;x ðC d;i1  C d;i ÞF ; Ld;i

ð4Þ

where Cd,i  1 and Cd,i are the hydrogen concentrations at the entry and exit side (section) of membrane i, respectively. Using an appropriate combination of Eqs. (3) and (4), the concentration of hydrogen inside the membrane, Cd, can be evaluated. The hydrogen concentration gradient obtained for the membrane of thickness Ld = 1 mm is shown in Fig. 3. In

Fig. 4. Steady-state concentration gradient of the diffusible hydrogen in the membranes with various thickness and comprising different parts of the nitrided diffusion zone.

contrast to the line gradient for the unnitrided iron membrane, this gradient is concave and it reflects the dependence of Dd,x on X [8]. The area under the Cd–X curve determines the amount of the diffusible hydrogen qHd. Finally, subtracting qHd from qH, the amount of the trapped hydrogen qHt was obtained (Table 1). The above approach was used to determine the hydrogen amounts and concentrations in the others membranes with various thicknesses, i.e., with different structure [8]. The evaluated concentration gradients are shown in Fig. 4. It should be noted that the thinner the membrane and, consequently, the less variable Dd,x, the less the concavity of the concentration gradient. Values of qH, qHd and qHt for all studied membranes are given in Table 1. 3.3. Composite membrane of thickness L(c + d), comprising the compound layer and the diffusion zone Since it was impossible to prepare a membrane consisting only of the nitride compound layer, information on the absorption of hydrogen in this layer was obtained indirectly by an analysis of the desorption of hydrogen from the membrane consisting of the whole compound layer and a significant part of diffusion zone (Fig. 5). Knowing the effective diffusion coefficient of hydrogen for this membrane, D(c + d) = 1.5 · 106 cm2/s [8], the desorption rates of the diffusible (mobile) hydrogen iHd,L and iHd,0 were obtained using Eqs. (1) and (2), respectively. The evaluated values of hydrogen content are given in Table 1. 4. Discussion 4.1. Unmodified iron substrate

Fig. 3. Steady-state concentration gradient of the diffusible hydrogen in the nitrided diffusion zone membrane.

In spite of relatively mild charging conditions (an alkaline solution, a moderate cathodic current density), the

1530

Z. Wolarek, T. Zakroczymski / Acta Materialia 54 (2006) 1525–1532

Fig. 6. Average concentration of the diffusible hydrogen (CHd,av), the trapped hydrogen (CHt,av) and their sum (CH,av) as a function of the thickness of the unnitrided iron substrate membranes.

Fig. 5. An analysis of the desorption rate of hydrogen at the exit (a) and entry (b) side of the membrane comprising the compound layer and the diffusion zone of the nitrided iron.

age concentration of the diffusible hydrogen is practically independent of the membrane thickness, whereas that of the trapped hydrogen is much higher and it decreases with membrane thickness. This also suggests that hydrogen occupies traps in the superficial layers initially. 4.2. Nitrided diffusion zone

hydrogen located in traps composed most (97–98%) of the total amount of absorbed hydrogen (Table 1). Since this hydrogen was reversibly trapped at room temperature, one can suppose that structural defects such as dislocations, grain boundaries, and non-metallic inclusions were mainly involved in hydrogen trapping. It can be concluded from Table 1 that the fraction of the diffusible hydrogen desorbed at the membrane exit side (qHd,L) is 1/3 while that at the entry side (qHd,0) is 2/3 of the total amount of the diffusible hydrogen (qHd). This is obvious because the model Eqs. (1) and (2), assuming a linear concentration gradient of hydrogen across the membrane at the beginning of desorption (or constant hydrogen diffusivity Ds), could be applied to evaluate qHd,L and qHd,0, respectively. In the case of the trapped hydrogen, its amount leaving the membrane at its entry side is more than 2/3 of the total amount, and the thinner the membrane the greater the ratio qHt,0/qHt. This means that the trapping of hydrogen occurred preferably at the membrane entry side, where the concentration of hydrogen dissolved in the lattice was relatively high, probably above some critical value [12]. The amounts of the various forms of hydrogen (qHd, qHt and qH), expressed in moles of H per unit area (Table 1), characterise a given membrane rather than its material. These values divided by the membrane thickness give the average concentrations CHd,av, CHt,av and CH,av, which are more reliable for comparison and more characteristic for the material. As seen in Fig. 6, the relatively low aver-

The recalculated values of the amounts of hydrogen (qHd, qHt and qH, Table 1) into corresponding average concentrations CHd,av, CHt,av and CH,av, are shown in Fig. 7. Generally, these concentrations are markedly higher than that for the unnitrided iron substrate (Fig. 6), and the thinner the membrane the distinctly higher the concentration of the trapped hydrogen. Taking

Fig. 7. Average concentration of the diffusible hydrogen (CHd,av), the trapped hydrogen (CHt,av) and their sum (CH,av) as a function of the thickness of the nitrided diffusion zone membranes.

Z. Wolarek, T. Zakroczymski / Acta Materialia 54 (2006) 1525–1532

into consideration the specific microstructure of the nitrided diffusion zone, in which the concentration of the Fe4N nitride precipitates decreased with depth, one can deduce that the nitride precipitates were additional, effective traps for hydrogen. Moreover, since the solubility of hydrogen in the iron nitrides is very low [1,13], one can suppose that the enhanced trapping of hydrogen occurs on the interfaces of the nitride precipitates with the ferrite matrix. The enhanced trapping of hydrogen by the nitride precipitates in the diffusion zone can be quantitatively estimated by comparison of hydrogen trapping in the unnitrided iron substrate membranes (Fig. 8). The level represented by the dotted line one may ascribe to the reversible trapping of hydrogen by common traps such as dislocations, grain boundaries and non-metallic inclusions. Consequently, the additional trapping in the nitrided diffusion zone, i.e., the difference between the dashed line and the dotted one, should be ascribed to the nitride precipitates. 4.3. Nitrided compound layer The amounts of hydrogen obtained for the membrane comprising the nitrided compound and the diffusion zone (Table 1) relate to the whole membrane. However, proportions of individual quantities indirectly give information on the absorption of hydrogen by the compound layer. The ratio qHt,0/qH,0 equals more than 0.99. In other words, hydrogen detected as reversibly trapped hydrogen had practically to originate from the compound layer. Taking the thickness of this layer Lc  10 lm, the average concentration of the trapped hydrogen CHt,av equals 6.4 · 104 mol/cm3, i.e., it is about 33 times higher than the average concentration of trapped hydrogen in the nitrided diffusion zone of the 0.4 mm thick membrane.

1531

4.4. Hydrogen entry vs. hydrogen absorption Finally, it is worth comparing the intensity of hydrogen entry and the resulting lattice solubility of hydrogen with the ability of each part of the nitrided iron to absorb hydrogen in any form. Under given charging conditions, the entry and lattice solubility are assessed by the concentration of the diffusible hydrogen just beneath the entry side of a given membrane, C0. In turn, the absorption ability can be assessed by the average total concentration of hydrogen, CH,av. Since the amounts of the trapped hydrogen were much higher than those of the diffusible hydrogen (Figs. 6 and 7), the average concentration of the trapped hydrogen, CHt,av, can be considered as a sufficient measure of hydrogen absorption. The pertinent values of C0 and CHt,av are compared in Table 2. As seen, the concentration C0 for the nitrided diffusion zone is about one order of magnitude higher than that for the unnitrided iron. This means that nitrogen atoms are not in the way of hydrogen atoms occupying interstitial sites in the ferrite matrix. This is not surprising because the solubility of nitrogen in the ferrite at room temperature is extremely low (about 0.001 wt.% [14,15]) and the majority of interstitial sites are still available for hydrogen. However, this aspect does not explain the observed higher solubility of hydrogen. Therefore, one may think that the higher solubility of hydrogen in the nitrided diffusion zone is a result of the lower effective diffusivity of hydrogen atoms (or longer diffusion paths). Probably, the slower effective transport of hydrogen through a membrane has an impact on the dynamic equilibrium between hydrogen atoms adsorbed on and absorbed beneath the iron surface. In turn, the solubility of the diffusible hydrogen in the nitrided compound layer is much lower than that in iron. This is understandable since the compound layer differs from iron (the surface is no longer metallic) and the concentration of nitrogen in this layer is higher [14,15]. The absorption or trapping ability (CHt,av), slightly increased by the additional trapping by the nitride precipitates, is even higher in the nitrided compound layer (Table 2). One may suppose that this intense trapping is related to pores within the compound layer. Accumulation of hydrogen within the compound layer of the nitrided iron was reported elsewhere [16]. Table 2 Concentration of the diffusible hydrogen beneath the entry side (C0) and the average concentration of the trapped hydrogen (CHt,av) for different parts of nitrided iron, cathodically charged under the same conditions: 0.1 M NaOH, ic = 20 mA/cm2 Unnitrided iron substrate Nitrided diffusion zone Nitrided compound layer a

Fig. 8. Trapping of hydrogen in the unnitrided iron substrate and in the nitrided diffusion zone membranes.

b c

C0 (mol/cm3)

CHt,av (mol/cm3)

3.6 · 107a 4.2 · 106b 6.0 · 108

0.9 · 105c 1.9 · 105c 6.4 · 104

Average value for all membrane thickness. Average value of Cd,0 (Fig. 4). Values for 0.4 mm thick membranes (Figs. 6 and 7).

1532

Z. Wolarek, T. Zakroczymski / Acta Materialia 54 (2006) 1525–1532

5. Conclusions 1. Modification of an iron surface by plasma nitriding strongly influenced the entry, lattice dissolution and trapping of hydrogen. However, the effect of individual nitrided layers (zones) on these processes was different. 2. The entry of hydrogen into the nitrided layer was much less than that into the ferrite. 3. Hydrogen absorbed in each layer occurred mainly as trapped hydrogen. 4. The nitride precipitate–ferrite matrix interfaces were additional traps aside from common structural defects. 5. The trapping of hydrogen by the nitride compound layer was much greater than that by the diffusion zone. References [1] Zakroczymski T, Lukomski N, Flis J. J Electrochem Soc 1993;140: 3578.

[2] Zakroczymski T, Lukomski N, Flis J. Corros Sci 1995;37:811. [3] Brass AM, Chene J, Pivin JC. J Mater Sci 1989;24:1693. [4] Bruzzoni P, Bru¨hl SP, Gomez JAB, Nosei L, Ortiz M, Feugeas JN. Surf Coat Technol 1998;110:13. [5] Fassini FD, Zampronio MA, de Miranda PEV. Corros Sci 1993;35:549. [6] Lesage J, Chicot D, Bartier O, Zampronio MA, de Miranda PEV. Mater Sci Eng A 2000;282:203. [7] Devanathan MAV, Stachurski Z. Proc R Soc A 1962;270:90. [8] Wolarek Z, Zakroczymski T. Acta Mater 2004;52:2637. [9] Nanis L, Nambodhiri TKG. J Electrochem Soc 1972;119:691. [10] Zakroczymski T. J Electroanal Chem 1999;475:82. [11] McBreen J, Beck W, Nanis L. J Electrochem Soc 1966;113: 1218. [12] Bockris JO’M, Subramanyan PK. J Electrochem Soc 1971;118: 1114. [13] Jack DH. Acta Metal 1976;24:137. [14] Hansen M, Anderko K. Constitution of binary alloys. New York (NY)/Toronto: McGraw Hill; 1958. [15] Chatterjee-Fischer R. Wa¨rmebehandlung von Eisenwerkstoffen, Nitrieren und Nitrocarburieren. Sindelfingen: Expert Verlag; 1986. [16] Lunarska E. Mater Corros 2000;51:1.