Scripta Materialia 53 (2005) 14431448 www.actamat-journals.

com

Magnetization eect on hydrogen absorption in high-strength steels and its implications

nchez a, B. Mishra b, D.L. Olson F.J. Sa
a b

b,*

Department of Petroleum Engineering, Colorado School of Mines, Golden, CO 80401-1887, USA Department of Metallurgical and Materials Engineering, Colorado School of Mines, Center for Welding, Joining and Coating Research, Golden, CO 80401-1887, USA Received 2 April 2005; received in revised form 25 July 2005; accepted 10 August 2005 Available online 13 September 2005

Abstract Magnetization eect of hydrogen absorption in high-strength steels has been investigated and the implications of the results for magnetic ux leakage assessment used for determination of defects in line pipe steels are discussed. A strong and continuous remanent magnetic eld in steels at a level commonly residual from magnetic ux leakage assessment practices can result in an increase of up to 1.62 times in the absorbed hydrogen concentration. 2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Keywords: Magnetization; Remanent; Magnetic ux leakage; High-strength steels; Pipelines; Hydrogen solubility; Adsorbed hydrogen concentration

1. Introduction The most ecient method for transporting uids in the oil, gas, and petrochemical industries is through steel pipelines. Pipelines are designed conservatively to protect against high-pressure, thermal eects, ice mechanics, and environmental degradation. To satisfy these conditions, pipelines are constructed from thick wall pipes and highstrength steels. However, high-strength steels are susceptible to hydrogen induced cracking [1]. This susceptibility is aected primarily by the hydrogen concentration, steel microstructure, stress concentration, temperature, and environment. The solubility of hydrogen in steels is a function of eective hydrogen partial pressure resulting from the environment, steel composition and microstructure, and temperature. During oil and gas operations hydrogen may be introduced in steel from various sources, such as hydrogen-bearing environments during welding, heat-treatment, pickling

Corresponding author. Tel.: +1 303 273 3955; fax: +1 303 384 2189. E-mail address: dolson@mines.edu (D.L. Olson).

or service. The most common service source of hydrogen is the cathodic reduction of water, which can be present during cathodic protection for corrosion. Processes involving cathodic polarization, such as electrodeposition, and reduction of H2S-containing hydrocarbon uids are also sources of hydrogen. The integrity of pipelines, with high level of condence and reliance, is monitored using intelligent inspection tools (broadly known as intelligent pigs). One of the assessment methods is the in-line inspection technology of determination of localized magnetic ux leakage (MFL). This technique performs by magnetizing the pipe steel near the saturation ux density and then detecting a local ux leakage caused by anomalies in surface morphologies. The magnetic saturation ux density of high-strength pipeline steels is approximately 1.9 T, while the remaining density (remanence) is in the region of 1 T. Hydrogen is considered as an electron donor to the d-band in steel. Studies have shown that when a magnetic eld near a hydrogen-charged ferromagnetic material is abruptly changed a redistribution of hydrogen atoms occurs [2]. Hydrogen diuses to energetically more favorable sites inside the magnetic domain structure. Hydrogen also

1359-6462/$ - see front matter 2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.scriptamat.2005.08.021

1444

nchez et al. / Scripta Materialia 53 (2005) 14431448 F.J. Sa

occupies the interstices or holes in the iron alloy lattice introducing a lattice strain. Since straining the iron alloy lattice results in less d-orbital overlap, then the less d-band lling occurs due to less action of the Pauli exclusion principle resulting in a change in both the shape and the lling of the d-band. This structural change results in a signicant alteration in electronic entropy contributions. Thus, there is a relationship between the hydrogen concentration and the magnetic properties of the steel: magnetic properties change as hydrogen concentration changes, and vice versa. The aim of the present study is to demonstrate experimentally that the magnetic induction generated by MFL inspection tools aects considerably the hydrogen content in pipeline steels. It is shown that the concentration of hydrogen in steels is a function of the surrounding magnetic eld strength as well as the hydrogen-bearing environment pressure, steel composition and microstructure, and temperature. The experimental investigation was carried out on three pipeline steel grades under strong and uniform magnetic elds. 2. Theoretical approach Since hydrogen atoms donate their electrons to the dband of transition metal, these bestowed electrons modify electron spin conguration and, in consequence, the material behavior under a magnetic eld. This phenomenon varies the entropy of the system resulting in a variation of the external work. A thermodynamic expression of equilibrium is utilized to derive the relationship between the magnetic eld strength and the activity of hydrogen. At constant pressure and temperature, the auxiliary function dG = external work. Modifying the derivation from the Nernst equation by adding the second external work term (DM B) to the electrochemical potential term (nFE), the dierential Gibbs free energy, dG, is given by dG DM B ne F E; 1 where DM is the change in magnetization, B is the magnetic ux density, ne is the number the electrons exchanged in the electrochemical reaction, F is Faradays constant, and E is the electrochemical potential. The region of hydrogen solubility is considered where the appropriate reaction, in terms of electrochemistry, is given as M H e ! M H s ; 2

Ln [H]
Ln [H] B=Bmax

M Slope = RT

Ln [H] B=0

Bremaining

Bmax

B (gauss)

Fig. 1. Schematic representation of Eq. (5) with illustration of ln[H] as a function of magnetic ux density (B).

shown as M H(s) in reaction (2). For a given alloy composition, the activity of the metal, M, may be taken as unity. At equilibrium, dG = 0, and for a given metal composition, the hydrogen concentration as a function of the change in magnetization is given by   ne FE0 DM B. 4 lnH 2:303 pH RT RT Rearranging Eq. (4), the hydrogen content under a magnetic ux density is determined by   DM HBB HB0 exp B; 5 RT where the equilibrium hydrogen content [H]B=B is determined under a magnetic ux density (B), and [H]B=0 is the hydrogen solubility under no applied magnetic eld and is given by   ne FE0 2:303 pH 6 H B0 exp RT at constant pH, temperature and for xed alloy composition, [H]B=0, is a constant. Eq. (6) shows how the absorbed hydrogen concentration increases with magnetic ux experience. This expression indicates that the hydrogen solubility will increase exponentially with increasing magnetic ux [3,4]. Rearranging Eq. (5), the graphical expression is developed, as shown in Fig. 1, where the slope of the straight line, (DM/RT), is a material property. 3. Experimental approach Several specimens were obtained from sectioned pipes to analyze the microstructure and to determine the chemical composition and mechanical properties. The results from the chemical analysis and the mechanical tests satised the requirements of the API specication 5L for the grades X52, X70, and X80 steels (Tables 1 and 2). Using standard optical microscopy examination, the microstructure revealed typical characteristics for these low carbon highstrength pipeline steels. Steel grades X52 and X70 present a mixed ferrite structure with small concentration of

where H(s) is a hydrogen atom in solid solution in the solid metal matrix, M. The half-cell electrochemical potential, E, in Eq. (1) belongs to reaction (2). Therefore, Eq. (1) becomes    RT H 0 dG DM B ne F E ln ; 3 ne F MH where E0 is the standard half-cell electrode potential for reaction (2) and is dependent on the composition of the metal, M, used. Note that [H] is used for simplicity to represent the concentration of hydrogen in the metal, which is

 nchez et al. / Scripta Materialia 53 (2005) 14431448 F.J. Sa Table 1 Chemical composition of tested steels (average) Material X52 X70 X80 API API API steel steel steel X52weld PSL2 X70weld PSL2 X80weld PSL2 Carbon (%) 0.0608 0.0704 0.07091 0.22 max. 0.22 max. 0.22 max. Manganese (%) 1.31 1.41 1.35 1.40 max. 1.65 max. 1.85 max. Phosphorus (%) 0.005 0.010 0.004 0.025 max. 0.025 max. 0.025 max. Sulfur (%) 0.0070 0.0031 0.0040 0.015 max. 0.015 max. 0.015 max. Titanium (%) 0.010 0.04 max. 0.06 max. 0.06 max.

1445

Table 2 Mechanical properties of tested steels Sample API specication 5L Minimum yield strength psi (MPa) X52 X70 X80 52,000 (358.5) 70,000 (482.6) 80,000 (551.6) Minimum UTS psi (MPa) 66,000 (455.0) 82,000 (565.4) 90,000 (620.5) Tensile test Yield strength psi (MPa) 75,612 (521.3) 82,619 (569.6) UTS psi (MPa) 81,771 (563.8) 94,161 (649.2) 100,000 (689.5)

carbides at grain boundary besides the coarse grains, few grains of pearlite, and greater ne ferrite grain concentration in grade X52. In contrast, steel grade X80 presents an acicular ferrite structure with sub-grain formation (Fig. 2).

Fig. 2. Microphotography of X52 (a), X70 (b), and X80 (c) steel samples showing: X52X70, mixed ferrite structure with low concentration of carbides at grain boundary and few evidences of pearlite. X80, acicular ferrite structure with heavy sub-grain formation. Etching: Nital 2%.

To establish the magnetic ux density required to saturate the steels, magnetization measurements were performed. Magnetization hysteresis loops were measured with a transverse-eld vibrating-sample magnetometer at room temperature under maximum applied elds of 0.6 T (600 G). Magnetization was computed as magnetic moment per unit mass of metal and expressed in units of kilogauss (kG). Fig. 3a shows the magnetization curves obtained from the test. The saturation magnetic ux density (Bs) for the longitudinal (L) samples of two steels (X52 and X70) is approximately 1.8 T (18,000 G), while for the radial (R) or transverse samples the saturation value is 1.6 T (16,000 G). Fig. 3b and c show the remanence points of the pipeline steels. Remaining magnetic ux density in both lateral and radial samples is approximately 1 T (10,000 G) for both grades. This value represents the magnetic ux density remaining in the pipeline steel after pigging operation is completed using MFL tools. These saturation and remaining magnetic ux density values are consistent with previous results obtained from 36 gas transmission pipe steel samples by Nestleroth and Crouch [5]. The hydrogen concentration was determined using the LECO-Hydrogen Determinator RH-404. Because this elemental analyzer requires 1 g samples, several 4 5 10 mm (0.157 0.197 0.394 in.) parallelepiped-shape specimens are sectioned from the selected pipeline steels to run hydrogen-charging experiments. One group of samples for each steel grade was used for the hydrogen-charging experiment without an applied magnetic eld and the second group for the experiment run under a specic strong, uniform magnetic eld. A magnetic induction of approximately 2 T (20,000 G) was selected. This experimental saturation magnetic ux density is to ensure a maximum magnetic saturation density for all the steel grades. The rst group of steel samples was mounted with epoxy resin. The epoxy resin reduces the galvanic cell eects in the hydrogen-charging system and ensures complete

1446
2

nchez et al. / Scripta Materialia 53 (2005) 14431448 F.J. Sa

(a)
1.5

Magnetic Flux Density (TESLA)

1 0.5 0 -0.5
X70R

-1 -1.5 -2

X70L X52R X52L

-8

-6

-4

-2

Magnetic Field Strength (KOe)

20

(b)
15 10 5 0 -5 -10 -15 -20 -1

Magnetic Flux Density (kG)

Fig. 4. Schematic diagram of hydrogen cathodic charging apparatus for the experiment without a magnetic eld applied to the sample (a). Steel sample-magnet setup is submerged in charging solution (1 N H2SO4) (b).

Magnetic Field Strength (kOe)

20

(c)
15 10 5 0 -5 -10 -15

The test was performed and the experimental para-meters were controlled electronically using a galvanostatic unit and commercial measurement software. To determine the behavior of hydrogen concentration as a function of charging time, the charging experiments were run varying the exposure time in intervals of 1, 2, 3, 6, and 8 h. For the hydrogen-charging experiments under an applied magnetic eld, the samples were placed between two sets of permanent magnets inducing a magnetic ux density of 2 T (20,000 G). The steel sample-magnet assembly was later submerged in the charging solution, as shown in Fig. 4b. Electrochemical implications related to the magnet surface and the steel sample exposed area were carefully considered. Magnets and steel samples were electrically isolated using coating and isolating spacers. 4. Results, interpretation, and signicance

Magnetic Flux Density (kG)

-20 -1

Magnetic Field Strength (kOe)

Fig. 3. Magnetization test results: (a) maximum saturation value is approximately 1.8 T (18,000 G), (b) and (c) remanence points in steel grades X52 and X70, respectively, are approximately 1 T (10,000 G).

submerging of the charged area in the solution (Fig. 4a). The cathodic charging was carried out electrolytically in a fresh 500 ml solution of 1 N H2SO4, using a platinum wire as anode and the steel sample as the cathode. The temperature (Troom) and pH (0.4) of the solution were continuously monitored, and the current rate was xed at 0.05 A/cm2.

With the purpose of assessing the eect of a strong, continuous magnetic eld on absorbed hydrogen content in pipeline steels, several steel grades X52, X70, and X80 samples were cathodically tested. A set of specimens was run without a magnetic eld and another set under a 2 T (20,000 G) magnetic induction to allow comparison. The absorbed hydrogen concentration at dierent cathodic charging times has been plotted in Fig. 5. The plotted results indicate that a strong and continuous magnetic eld causes a tremendous eect on absorbed hydrogen concentration in high-strength pipeline steels.

 nchez et al. / Scripta Materialia 53 (2005) 14431448 F.J. Sa

Hydrogen Concentration after Charging Pipeline Steel Grades X52, X70, and X80

1447

9 8

7 6 5 4 3 2 1 0 0 1 2 3 4 5 6 7 8 9
X52 X52-Magnet X70 X70-Magnet X80 X80-Magnet

Charging Time (hrs)

Fig. 5. Absorbed hydrogen concentration in pipeline steel grades X52, X70, and X80 at dierent cathodic charging times. It shows for comparison the experimental results without magnetization and under a 2 T (20,000 G) magnetic induction (indicated as Grade-Magnet).

After reaching the saturation level (plateau), the hydrogen content in steel samples under a 2 T magnetic induction is up to 6.3 times the concentration under cathodic charging conditions without an applied magnetic eld. According to Fig. 5, the saturation magnetic ux density points of pipeline steel grade X52, X70, and X80 are reached approximately after 4, 2, and 1 h of cathodic charging in 1 N sulfuric acid, respectively. Correlating Eq. (5) with the results in Fig. 5, the equilibrium point of absorbed hydrogen concentration under initial experimental conditions (without magnetization, B = 0) is represented by [H]B=0, i.e. the natural logarithm at magnetic ux density of zero (ln[H]B=0). This point is reached at plateau level; that is, [H]B=0 = 1.29 ppm where ln[H]B=0 = 0.2231. On the other hand, the plateau level of the magnetic ux density curve represents the absorbed hydrogen concentration at the specic magnetic induction of 2 T (20,000 G) and charging time and current density. For example, at [H]B=2T = 7.9 ppm, Eq. (5) for steel grade X52 becomes   DM ln7:9 ppm ln1:29 ppm 20; 000 G: 7 RT It is well known that the slope of this straight line, DM/RT, is a property of the assessed pipeline steel under the experimental conditions and is related to the magnetic
Table 3 Experimental data from Fig. 5 at magnetic ux density of 20,000 G Parameter Steel X52 Without magnetization Equilibrium hydrogen concentration [H] (ppm) ln[H] [H]/[H]B=0 ratio DM/RT 1.250 0.223 6.3 0.0001 Magnetization 7.900 2.067

permeability of the material. Dierent pipeline steels generate dierent straight-line slopes because they have diverse strength levels and microstructures. It is commonly expected that the magnetic properties of pipeline steels will vary with pipe grade. Fig. 5 illustrates the graphical expression of the experiment using the experimental data from Table 3. The [H]/[H]B=0 ratios equal 6.3, 3.1, and 1.6, for steel grades X52, X70, and X80, respectively, leading to the following statement: a magnetic induction of 20,000 G and 0.05 A/cm2 current density may increase more than six times the absorbed hydrogen concentration of pipeline steels. A question to be addressed is: how much is the absorbed hydrogen content increase under the remanent magnetization left by the MFL tool after inspection operations? This fundamental question can be addressed by the analysis of the in-between points from Fig. 6. As the remanent magnetic induction (Br) in the steel grades is approximately 10,000 G, the potential absorbed hydrogen content increases approximately up to 1.62 times (Table 4). However, this situation is possible if the magnetization is constant and does not experience alterations with aging. Surface anomalies could also increase the magnetic induction over these measured values from test coupons. Surface anomalies such as cracks (at crack tip) and dents (coldworked areas) increase considerably the magnetic ux density within the MFL process. They can reach a magnetization level greater than the remanent, thus the potential

H-Concentration (ppm)

Hydrogen Concentration vs. Magnetic Flux Density Steel Grades X52, X70, and X80
1.E+01

Ln [H]

1.E+00

X52 X70 X80

1.E-01 0 5,000 10,000 15,000 20,000

B (Gauss)

Fig. 6. Graphical representation of Eq. (5).

Steel X70 Without magnetization 2.100 0.742 3.1 0.00005 Magnetization 6.500 1.872

Steel X80 Without magnetization 2.800 1.030 1.6 0.00002 Magnetization 4.500 1.504

1448

nchez et al. / Scripta Materialia 53 (2005) 14431448 F.J. Sa

Table 4 Determination of intermediate points at magnetic ux density of 10,000 G Parameter at 10,000 G Steel grade X52 ln[H] [H]B=10kG (ppm) [H]B=0 (ppm) [H]B=10kG/[H]B=0 ratio 0.606 1.834 1.250 1.47 X70 1.223 3.398 2.100 1.62 X80 1.258 3.517 2.800 1.26

Acknowledgments The authors acknowledge and appreciate the research support of the US Department of the Interior (Mineral Management Services), US Department of Transportation (Oce of Pipeline Safety1), the assistance of the Magnetics Group of the Electromagnetics Division in the Electronics and Electrical Engineering Laboratory at NIST (Boulder), especially the scientic discussions with Dr. Ron Goldfarb, and the experimental suggestions from Dr. Yeong-Do Park (Hyundai Motor, South Korea). References
[1] Corrosion: fundamentals, testing, and protection, vol. 13A. Materials Park (OH): American Society of MetalsASM; 2003. [2] Pepperho W, Acet M. Constitution and magnetism of iron and its alloys. Berlin: Springer-Verlag; 2001. [3] Niyomsoan S, Termsuksawad P, Olson DL, Mishra B, Kaydanov VI, Gaura Z. The relationship between the thermoelectric power and phase structure in AB2 hydrogen storage materials. J Mater Sci Eng A 2004;291:26471. [4] Termsuksawad P. A fundamental investigation of hydrogen absorption in AB5 type hydrogen storage materials and its characterization by magnetic and electronic techniques. Doctoral Thesis, Colorado School of Mines, 2003. [5] Nestleroth JB, Crouch AE. Variations of magnetic properties in pipeline steels. Prepared for US Department of Transportation. No DTRS56-96-C-0010. Washington, DC, 1997.

absorbed hydrogen concentration under these conditions would increase more than three times. 5. Conclusions 1. A strong, continuous remanent magnetic eld causes a signicant eect on absorbed hydrogen concentration in high-strength pipeline steels. 2. At a remanent magnetic induction of 1 T, the adsorbed hydrogen concentration in high-strength pipeline steels increases by approximately 60%. At saturation magnetization (during pigging), the concentration may be as much as six times higher. 3. The strength level and the microstructure of pipeline steels inuence the eect of magnetization on hydrogen absorption in the tested steel. 4. Surface anomalies such as cracks and dents increase signicantly the magnetic ux density within the MFL process. Such localized magnetic induction would suggest a further increase of the potential absorbed hydrogen concentration.

1 For additional information: http://primis.rspa.dot.gov/matrix/ PrjCoHome.rdm?&prj=125.