Corrosion Science 80 (2014) 517–522

Contents lists available at ScienceDirect

Corrosion Science

journal homepage: www.elsevier .com/ locate /corsc i

Short Communication

Hydrogen diffusion coefficients through Inconel 718 in differentmetallurgical conditions

0010-938X/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.corsci.2013.11.002

⇑ Corresponding author. Current address: Materials Technology Center, SouthernIllinois University, Carbondale, IL 62901, United States. Tel.: +1 618 453 7822.

E-mail address: isuni@siu.edu (I.I. Suni).

Josiah J.M. Jebaraj a, David J. Morrison b, Ian I. Suni a,c,⇑a Department of Chemical and Biomolecular Engineering, Clarkson University, Potsdam, NY 13699-5705, USAb Department of Mechanical and Aeronautical Engineering, Clarkson University, Potsdam, NY 13699-5705, USAc Materials Science and Engineering Program, Clarkson University, Potsdam, NY 13699-5705, USA

a r t i c l e i n f o

Article history:Received 18 April 2013Accepted 10 November 2013Available online 17 November 2013

Keywords:A. NickelB. Hydrogen permeationB. TEMC. Cathodic protectionC. Hydrogen embrittlement

a b s t r a c t

We report hydrogen permeation studies through cold rolled, solutionized, and precipitation hardenedInconel 718 foils. The effective hydrogen diffusion coefficient is considerably higher (5.3–6.8 � 10�11 cm2/s) for the solutionized Inconel 718 than for either the cold rolled (3.3–4.2 � 10�11 cm2/s) or precipitation hardened (2.1–2.9 � 10�11 cm2/s) specimens. Microstructural studiesindicate that the reduced hydrogen diffusion coefficients in the latter specimens arise from hydrogentrapping at dislocations and precipitates that are present at much lower concentrations in the solution-ized specimens. Also, repeated permeation transients provide evidence for irreversible hydrogen trappingin the cold rolled and precipitation hardened specimens, but such effects are insignificant in the solution-ized specimens.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The offshore oil and gas environment is highly corrosive due tothe presence of seawater, hydrogen sulfide gas (H2S), and wide ex-tremes of temperature and pressure. Corrosion is frequently ad-dressed through alloy selection and cathodic protection.However, cathodic protection causes the formation of atomichydrogen on the metal surface; and this may diffuse into the metal,leading to hydrogen embrittlement. Metallurgical factors that af-fect the susceptibility to hydrogen embrittlement include crystalstructure, vacancies, dislocations, precipitates, grain boundaries,and degree of cold work [1,2]. Furthermore, thermomechanicalprocessing that leads to increased strength and hardness almost al-ways increases the susceptibility to hydrogen embrittlement [3,4].Environmental factors include cathodic potential, temperature,pressure, stress, strain, and strain rate [4]. The material responseis difficult to predict due to the complex interplay between hydro-gen dissolution, hydrogen mobility through diffusion, formation ofcalcareous deposits, hydride formation in some alloys, and hydro-gen trapping at defects. Mechanisms that have been proposed toexplain hydrogen embrittlement include hydrogen-enhancedlocalized plasticity [5], hydrogen-induced decohesion [6,7], andformation of brittle hydrides [8].

Because hydrogen diffusion is the first step in the proposedmechanisms for hydrogen embrittlement, understanding hydrogendiffusion is critical to understanding how to prevent hydrogen-in-duced brittle failure. Hydrogen diffusion through materials suscep-tible to hydrogen embrittlement may be impacted by theapplication of metallurgical treatments that vary the concentrationof hydrogen trap sites [1,2], or by surface processes such as nitrid-ation [9,10] or electrodeposition [11,12] that hinder hydrogentransport. Hydrogen transport through Fe alloys that are suscepti-ble to hydrogen embrittlement has been widely studied in electro-chemical systems by the Devanathan–Stachurski method [13,14].

Here we report hydrogen permeation studies through Inconel718 foils in three different metallurgical conditions: cold worked,solutionized, and precipitation hardened. Due to the much slowerrate of hydrogen diffusion through Ni-based alloys relative to Fealloys, few electrochemical hydrogen permeation studies havebeen reported in Ni-based alloys [12,15,16].

2. Materials and methods

2.1. Materials

Inconel 718 foil 50 lm thick was purchased from ESPI metalsand cut into 4.0 cm diameter disks. The as-received foil was re-ported by the manufacturer to be annealed and cold rolled, andoptical profilometry yielded a surface roughness (Ra) of 35 (3)nm on both sides of the specimen. We will refer to the as-receivedfoil as cold rolled. The composition of Inconel 718 foil is given in

Fig. 1. Schematic of the experimental cell for hydrogen permeation experiments.

518 J.J.M. Jebaraj et al. / Corrosion Science 80 (2014) 517–522

Table 1. Solutionized specimens were prepared according to AMS5663M [17] by heating the specimens in air at 976 �C for 0.5 hand then quenching to room temperature in oil. Precipitationhardened specimens were prepared by heating the solutionizedspecimens in air at 739 �C for 8 h, then at 635 �C for 10 h. The thinoxide layer that formed at elevated temperature was removed bymechanical abrasion, after which optical profilometry yields a sur-face roughness (Ra) of 65 (2) nm on both sides of the specimen.

2.2. Microstructural characterization

Specimens for microstructural analysis were prepared by grind-ing Inconel 718 foils with 600, 800, and 1200 grit SiC followed bypolishing with 0.05 lm alumina. The mirror finish surface wasthen etched with glyceregia (10 mL glycerol, 10 mL HCl, 5 mLHNO3) for 90 s. Transmission electron microscope specimens of3 mm diameter were punched from 50 lm thick foils. These spec-imens were electropolished using a double jet polisher at �10 �Cand 10 V in a solution of 450 mL methanol, 100 mL 2-butoxyetha-nol, and 50 mL perchloric acid. Microstructural characterizationwas accomplished using an Olympus PME optical microscope, aJEOL 7400 scanning electron microscope (SEM) with energy disper-sive X-ray spectrometer (EDX), and a JEOL JEM 1200 transmissionelectron microscope (TEM).

2.3. Hydrogen permeation measurements

A schematic of the experimental Devanathan–Stachurski cell forhydrogen permeation experiments is shown in Fig. 1. All electro-chemical experiments were performed at 50 �C using a circulatingwater bath. Prior to hydrogen permeation experiments, the speci-mens were coated with�20 nm Pd by vacuum evaporation on bothsides of the specimen. Following Pd coating, the specimens wereclamped between the charging and measurement cells, both ofwhich were made of Pyrex. The exposed membrane area withineach cell was about 5.7 cm2. In the charging cell, hydrogen wasgenerated at a constant current density of 1.6 mA/cm2 in 0.1 MNaOH. In the measurement cell, hydrogen transport through themembrane was quantified by measuring the anodic current be-tween the specimen and a Pt counter electrode at +300 mV vs. asaturated calomel electrode (SCE) in 0.1 M NaOH. The Inconel718 membrane was shared by both electrochemical cells, servingas the working electrode in the hydrogen measurement cell andthe cathode in the hydrogen generation cell. The constant currentdensity in the charging cell was supplied between the specimenand a Pt anode by a Mastech HY 3005D DC power supply, andthe anodic current in the measurement cell was monitored usinga Gamry model G-750 potentiostat, which was operated with afloating ground. During the absence of charging, decay transientswere obtained.

Prior to all experiments, the electrolytes were purged with highpurity Ar. Before initiating hydrogen generation in the cathodic(charging) cell, the anodic (measurement) cell was operated for

Table 1Chemical composition of Inconel 718.

Element Composition (wt%)

C 0.03Cr 18.36Fe 17.98Mn 0.12Mo 2.91Ni 53.43Cb (Nb) 5.02Ti 1.06Other 1.09

5–6 days until the anodic current dropped to �30 nA/cm2, and re-mained approximately constant. This ensures that all oxidizableimpurities are removed from the electrolyte in the measurementcell before starting the hydrogen permeation transient. This mayalso allow hydrogen initially dissolved in the Inconel 718 foil todiffuse into the measurement cell and be oxidized.

3. Results and discussion

3.1. Microstructural analysis of Inconel 718 membranes

Optical microscopy of the three types of specimens indicatedthat they all had fairly equiaxed grains. Using the average linearintercept method, the three conditions exhibited similar grain sizesof about 21 lm. EDX analyses of the three conditions indicate thatthe chemical compositions are identical to within measurementaccuracy. Fig. 2 shows TEM micrographs of the three material con-ditions. The cold rolled microstructure shown in Fig. 2a consists ofdense bundles of dislocations, reflecting the severe deformationprocessing required to form such thin sheets. Fig. 2b shows thatthe solutionized microstructure appears to be single phase withisolated dislocations. Fig. 2c clearly exhibits the contrast that isfound in c0/c00 precipitation hardened Inconel 718 [18].

3.2. Hydrogen diffusion coefficients within Inconel 718 membranes

Figs. 3–5 illustrate repeated hydrogen permeation transientsthrough cold rolled, solutionized, and precipitation hardened Inco-nel 718 membranes, respectively. The hydrogen permeation tran-sients were analyzed, assuming constant hydrogen concentrationat the charging surface of the membrane, by the time-to-break-through, time lag, Fourier, and Laplace methods [1]. Eqs. (1)–(4),were used to calculate effective diffusion coefficients for the risetransients, while Eqs. (5) and (6) were used for decay transients.The results of these analyses are given in Tables 2–4. Table 5 sum-marizes the average hydrogen diffusion coefficient obtained foreach permeation transient for each membrane condition.

Time lag method D ¼ L2

6tlagð1Þ

Breakthrough time method D ¼ L2

15:3tbð2Þ

Fig. 2. Tem images of the cold rolled (a), solutionized (b), and precipitationhardened (c) Inconel 718 membrane.

Fig. 3. Repeated hydrogen permeation transients through a cold rolled Inconel 718membrane.

Fig. 4. Repeated hydrogen permeation transients through a solutionized Inconel718 membrane.

Fig. 5. Repeated hydrogen permeation transients through a precipitation hardenedan Inconel 718 membrane.

J.J.M. Jebaraj et al. / Corrosion Science 80 (2014) 517–522 519

Table 2DH in cold rolled Inconel 718 at 50 �C.

Method Diffusion coefficient (�10�11 cm2 s�1)

1st transient (rise) 1st transient (decay) 2nd transient (rise) 3rd transient (rise) 3rd transient (decay)

Time lag 1.5 3.5 3.3Breakthrough time 2.9 3.9 3.9Fourier analysis 2.2 3.6 3.7 3.5 3.5Laplace analysis (slope) 2.7 3.4 3.6 3.9 4.2Laplace analysis (intercept) 2.1 3.9 3.9 4.0 3.9Range 1.5–2.9 3.4–3.9 3.5–3.9 3.3–4.0 3.5–4.2

Table 3DH in solutionized Inconel 718 at 50 �C.

Method Diffusion coefficient (�10�11 cm2 s�1)

1st transient (rise) 1st transient (decay) 2nd transient (rise) 2nd transient (decay) 3rd transient (rise)

Time lag 6.5 6.0 5.3Breakthrough time 6.8 6.5 6.1Fourier analysis 6.3 5.3 6.3 5.3 6.3Laplace analysis (slope) 6.2 6.2 6.3 6.2 5.8Laplace analysis (intercept) 6.5 5.9 6.1 6.0 5.9Range 6.2–6.8 5.3–6.2 6.0–6.5 5.3–6.2 5.3–6.3

Table 4DH in precipitation hardened Inconel 718 at 50 �C.

Method Diffusion coefficient (�10�11 cm2 s�1)

1st transient (rise) 1st transient (decay) 2nd transient (rise) 2nd transient (decay) 3rd transient (rise)

Time lag 1.5 2.7 2.6Breakthrough time 1.4 2.3 2.3Fourier analysis 1.5 2.1 2.2 2 2.5Laplace analysis (slope) 1.8 2.6 2.6 2.9 2.3Laplace analysis (intercept) 1.5 2.2 2.4 2.7 2.3Range 1.4–1.8 2.1 – 2.6 2.2–2.7 2–2.9 2.3–2.6

Table 5Average DH for each hydrogen permeation transient in each membrane.

Specimen Diffusion coefficient (�10�11 cm2 s�1)

1st transient (rise) 1st transient (decay) 2nd transient (rise) 2nd transient (decay) 3rd transient (rise)

Cold rolled 2.3 3.6 3.7 3.9 (3rd decay) 3.7Solutionized 6.5 5.8 6.2 5.8 5.9Precipitation hardened 1.5 2.3 2.4 2.5 2.4

520 J.J.M. Jebaraj et al. / Corrosion Science 80 (2014) 517–522

Fourier method rise transientð Þ jj1¼ 1� 2 exp �p2Dt

L2

� �ð3Þ

LaPlace method ðrise transientÞ jj1¼ 2

p1=2

L

ðDtÞ1=2 exp � L2

4Dt

!

ð4Þ

Fourier method ðdecay transientÞ jj1¼ 2 exp �p2Dt

L2

� �ð5Þ

LaPlace method ðdecay transientÞ jj1¼1� 2

p1=2

L

ðDtÞ1=2 exp � L2

4Dt

!

ð6Þ

In the above equations: L is the sample thickness; j is the perme-ation current density at time t; j1 is the steady state permeationcurrent density; tlag is the lag time, the point on the permeationcurve at which j = 0.63⁄j1; and tb is the breakthrough time, foundby extrapolating the linear portion of the initial rise transient toi = 0.

The results of Figs. 3–5 and Tables 2–5 yield evidence regardingthe concentrations of both reversible and irreversible hydrogentrap sites for membranes in the three different metallurgical con-ditions. For the solutionized Inconel 718 membrane, the effectivehydrogen diffusion coefficients extracted from all rise and all decaytransients are indistinguishable. This indicates that irreversiblehydrogen trap sites are insignificant for the solutionized Inconel718 membrane [1,19,20]. This suggests that lattice diffusion maydominate the behavior of the solutionized Inconel 718 specimens,since these have the lowest defect density compared to the coldrolled and precipitation hardened specimens.

On the other hand, the effective hydrogen diffusion coefficientfor the first rise transient is substantially lower than those ob-tained from subsequent transients for the cold rolled and precipi-tation hardened specimens. This phenomenon is consistent withthe presence of irreversible hydrogen trapping sites, which slowhydrogen transport during the first transient, but are already filledand have no effect on subsequent permeation transients [1,19,20].This behavior likely arises from the high concentration of defectsthat act as irreversible hydrogen trap sites in these specimens.For example, the cold rolled Inconel 718 contains a high dislocation

Table 6Apparent hydrogen solubility (Capp) in each membrane.

Specimen Hydrogen solubility (mmol cm�3)

1st transient (rise) 2nd transient (rise) 3rd transient (rise)

Cold rolled 2.34 1.52 1.90Solutionized 0.71 0.81 1.06Precipitation hardened 1.35 1.23 1.51

J.J.M. Jebaraj et al. / Corrosion Science 80 (2014) 517–522 521

density, while the precipitation hardened Inconel 718 has a highconcentration of c0 and c00 precipitates.

Since the first rise transient will fill irreversible hydrogen trapsites, all subsequent rise and decay transients only sample revers-ible hydrogen trap sites [1,19,20]. For those transients, the solu-tionized Inconel 718 specimen exhibits a consistently highereffective diffusion (6.1 � 10�11 cm2/s) coefficient than the coldrolled (3.8 � 10�11 cm2/s) and precipitation hardened(2.4 � 10�11 cm2/s) specimens. This suggests that the higher defectdensities in cold rolled and precipitation hardened specimens alsocontribute to higher densities of reversible hydrogen trap sites.

The hydrogen effective diffusion coefficients reported in Table 3for all permeation transients in solutionized Inconel 718 showexcellent agreement. Similarly, the diffusion coefficients reportedin Tables 2 and 4 for the second and subsequent permeation tran-sients in cold rolled and precipitation hardened Inconel 718 alsoshow excellent agreement. These results can be compared to gasphase measurements of hydrogen transport through Inconel 718which exhibit much greater scatter in the reported diffusion coef-ficients [21]. Turnbull et al. have observed an increase in effectivediffusivity of hydrogen by electrochemical hydrogen permeationexperiments in solutionized specimens compared to direct agedand precipitation hardened Inconel 718 at 80 �C, which is generallyconsistent with our results [16]. However, their results in solution-ized and precipitation hardened specimens lacked informationregarding the effective diffusivity of hydrogen in multiple perme-ation transients. The excellent reproducibility in the current resultscan be attributed to both the measurement of repeat hydrogenpermeation transients in the same specimen, and conductingmeasurements at low temperature. To the best of the author’sknowledge, the current results appear to be the first literaturereport of repeated rise and decay hydrogen permeation transientsin Ni alloys.

3.3. Hydrogen solubility within Inconel 718 membranes

In addition to measuring the effective hydrogen diffusion coef-ficient (Deff), the results of Figs. 3–5 can also be used to estimatethe apparent hydrogen solubility (Capp) [22]:

Capp ¼i1L

nFDeffð7Þ

where i1 is the steady state permeation current density, L is themembrane thickness, n is the number of electrons transferredduring hydrogen oxidation/reduction, and F is Faraday’s constant.Table 6 summarizes the hydrogen solubilities obtained from thepermeation transients in Figs. 3–5. These results are consistent withthe arguments above about the relative defect density in the differ-ent specimens. The solubility is lowest for the solutionized Inconel718, since this has a lower density of dislocations, precipitates, andother defects than the cold rolled and precipitation hardenedspecimens.

4. Conclusions

The results of this study indicate that the effective hydrogen dif-fusion coefficient is considerably higher (5.3–6.8 � 10�11 cm2/s)

for solutionized Inconel 718 than for either cold rolled (3.3–4.2x10�11 cm2/s) or precipitation hardened (2.1–2.9 � 10�11 cm2/s) specimens. The reduced hydrogen diffusion coefficient in coldrolled and precipitation hardened specimens arises from hydrogentrapping at dislocations and precipitates, respectively, which arepresent at much higher concentrations than in the solutionizedspecimens. In addition, differences between hydrogen transportduring the first and subsequent permeation transients shows evi-dence for irreversible hydrogen trapping in cold rolled and precip-itation hardened Inconel 718, but such effects are insignificant insolutionized specimens.

Acknowledgement

The authors gratefully acknowledge the support of this researchby General Electric Oil and Gas.

References

[1] A. Turnbull, M.S. de Santa Maria, N.D. Thomas, The effect of H2S concentrationand pH on hydrogen permeation in AISI 410 stainless steel in 5% NaCl, Corros.Sci. 29 (1989) 89–104.

[2] S.L. Chou, W.T. Tsai, Effect of grain size on the hydrogen-assisted cracking induplex stainless steels, Mater. Sci. Engin. A 270 (1999) 219–224.

[3] K.A. Esaklul, T.M. Ahmed, Prevention of failures of high strength fasteners inuse in offshore and subsea applications, Engin. Failure Anal. 16 (2009) 1195–1202.

[4] M.R. Louthan, Hydrogen embrittlement of metals: a primer for the failureanalyst, DOE Document WSRC-STI-2008-0062.

[5] T. Matsumoto, J. Eastman, H.K. Birnbaum, Direct observations of enhanceddislocation mobility due to hydrogen, Scr. Metall. 15 (1981) 1033–1037.

[6] J.G. Morlet, H.H. Johnson, A.R. Troiano, A new concept of hydrogenembrittlement in steel, J. Iron Steel Inst. 189 (1958) 37–44.

[7] H.A. Wriedt, R.A. Oriani, Effect of tensile and compressive elastic stress onequilibrium hydrogen solubility in a solid, Acta Metall. 18 (1970) 753–760.

[8] B.J. Makenas, H.K. Birnbaum, Phase changes in the niobium–hydrogen SystemI: accommodation effects during hydride precipitation, Acta Metall. 28 (1980)979–988.

[9] J. Cwiek, M. Baczyñska, Behaviour of nitrided layers subjected to influence ofhydrogen, Arch. Mater. Sci. Engin. 43 (2010) 30–41.

[10] A. Gajek, Z. Wolarek, T. Zakroczymski, Behaviour of hydrogen in gas nitridediron studied by electrochemical permeation and desorption techniques,Corros. Sci. 58 (2012) 260–266.

[11] G. Zheng, B.N. Popov, R.E. White, Surface treatment for mitigation of hydrogenabsorption and penetration into AISI 4340 steel, J. Electrochem. Soc. 140(1993) 3153–3158.

[12] B.N. Popov, G. Zheng, R.E. White, Electroplating of thin films of bismuth ontotype 4340 steel and Alloy 718 to prevent hydrogen embrittlement, Corros. Sci.51 (1995) 429–435.

[13] M.A.V. Devanathan, Z. Stachurski, The adsorption and diffusion of electrolytichydrogen in palladium, Proc. Royal Soc. A 270 (1962) 90–102.

[14] M.A.V. Devanathan, Z. Stachurski, W. Beck, A technique for the evaluation ofhydrogen embrittlement characteristics of electroplating baths, J. Electrochem.Soc. 110 (1963) 886–890.

[15] J.M. Zagal, H.F. López, O. Flores, J.L. Albarran, L. Martínez, Microstructuraleffects on the hydrogen permeation of an Inconel alloy 690, Corros. Sci. 50(2008) 3371–3377.

[16] A. Turnbull, R.G. Ballinger, I.S. Hwang, M.M. Morra, M. Psaila-Dombrowski,R.M. Gates, Hydrogen transport in nickel-base alloys, Metall. Trans. A 23(1992) 3231–3244.

[17] AMS 5663, Nickel alloy, Corrosion and heat resistant, bars, forgings, and rings52.5Ni 19Cr 3.0Mo 5.1Cb 0.90Ti 0.50Al 19Fe consumable electrode or vacuuminduction melted 1775�F (968 �C) solution and precipitation heat treated,(2009).

[18] D.F. Paulonis, J.M. Oblak, D.S. Duvall, Precipitation in nickel-base alloy 718,Trans. ASM 62 (1969) 611–622.

[19] E. Owczarek, T. Zakroczymski, Hydrogen transport in a duplex stainless steel,Acta Mater. 48 (2000) 3059–3070.

522 J.J.M. Jebaraj et al. / Corrosion Science 80 (2014) 517–522

[20] C. Gesnouin, A. Hazarabedian, P. Bruzzoni, J. Ovejero-Garcıa, P. Bilmes, C.Llorente, Effect of post-weld heat treatment on the microstructure andhydrogen permeation of 13CrNiMo steels, Corros. Sci. 46 (2004) 1633–1647.

[21] J. Xu, X.K. Sun, Q.Q. Liu, W.X. Chen, Hydrogen permeation behavior in IN 718and GH 761 superalloys, Metall. Mater. Trans. A 25 (1994) 539–544.

[22] H. Addach, P. Berçot, M. Rezrazi, J. Takadoum, Study of the electrochemicalpermeation of hydrogen in iron, Corros. Sci. 51 (2009) 263–267.