CORROSION ENGINEERING

Field Measurement of Hydrogen

in Sour Gas Pipelines✫

M. Elboujdaini, V.S. Sastri, and R.W. Revie*

ABSTRACT segregation, and shape of nonmetallic inclusions.

Some of the environmental factors that can cause HIC
The hydrogen (H) concentration (COH) at the inside surface of are the partial pressures of hydrogen sulfide (H2S) and
linepipe steels containing different media was determined and carbon dioxide (CO2), temperature, pH of the medium,
compared with values obtained in field tests. Test media moisture, and the presence of aggressive ions.
included NACE TM0177, NACE TM0284, modified NACE Sour gas pipelines have failed as a result of HIC.4-5
TM0284, and field-produced water saturated with hydrogen Considerable effort has been expended by steel
sulfide (H2S). Results showed field conditions were more linepipe producers, users, and research organizations
severe than TM0284 but less severe than modified TM0284.
seeking to understand the mechanism of HIC so that a
Data suggested a medium buffered to a pH of 4.1 to 4.7
might reflect field conditions. Results also suggested diffusion laboratory test method can be developed to identify
coefficient values obtained from breakthrough times could be and quantify material susceptibility to HIC, which would
used in the calculation of COH in field tests without sacrificing allow for production of steels with greater resistance to
accuracy. It was shown that analysis of breakthrough time HIC.
data can provide some idea of the degree of the relative trap Previous work in collaboration with the Canadian
density in linepipe steels. Standards Association (CSA) Sour Service Task Force
KEY WORDS: carbon steel, cracking, field test, hydrogen, involved the evaluation and comparison of threshold
hydrogen-induced cracking, hydrogen permeation, hydrogen hydrogen concentrations (CthH) for HIC in 19
sulfide, pipelines, sour gas, trap density commercial linepipe steels.6 A correlation between the
diffusible hydrogen concentration (COH) in the coupon
INTRODUCTION and HIC was established, as was CthH, above which
cracking would occur.6
Linepipe steels used in sour service are prone to A major concern of linepipe producers and users
hydrogen-induced cracking (HIC), also known as is the extent to which laboratory test solutions reflect
stepwise cracking (SWC), depending on metallurgical field conditions.
and environmental factors.1-3 The metallurgical factors In response to this concern, experimental work
consist of alloying elements, microstructure, strength, was undertaken to determine the concentration of
hydrogen inside a pipe wall containing different test
solutions (such as NACE Standard TM01777 and
modified NACE Standard TM02848 [pH 3.5], NACE
✫
Submitted for publication October 1993; in revised form, March 1994.
* Energy, Mines, & Resources Canada, CANMET, 568 Booth St., Ottawa,
TM0284, and field-produced water) and to determine
K1A 0G1, Canada. COH in a loop in a field with and without added inhibitor.

0010-9312/94/000149/$5.00+$0.50/0
636 © 1994, NACE International CORROSION–AUGUST 1994
 CORROSION ENGINEERING

TABLE 1
Composition of AM-2 Steel(A)
C Si Mn P S Al Cr Ni Cu Nb Ti N Mo Ca

0.080 0.41 0.80 < 0.03 0.0016 0.13 0.027 0.015 0.011 0.023 0.019 0.0058 < 0.01 0.0040

(A)
C, carbon; Si, silicon; Mn, manganese; P, phosphorus; S, sulfur; Al, aluminum; Cr, chromium; Cu, copper; Ni, nickel; Nb, niobium; Ti, titanium;
N, nitrogen; Mo, molybdenum; and Ca, calcium.

Primary objectives were to compare COH values Finally, the HMD was attached to the pipe with a
obtained in laboratory tests to those measured in the clamp.
field and to identify the laboratory test medium that
most closely reflected field conditions. Field tests were Diffusion Coefficients
made on an aerial test loop. Under normal conditions, It was necessary to determine the diffusion
pipes would be buried and usually cathodically coefficient (D) for H in steel so that COH at the inside
protected. wall of the pipe could be calculated. The experimental
setup and the procedure for determining D was the
EXPERIMENTAL same as used by Devanathan (i.e., electrochemical
charging).10 The solution consisted of dilute sulfuric
The electric resistance-welded (ERW) linepipe acid (H2SO4) of pH 2.0 containing arsenic trioxide
steel (152.4-mm diam, 7.4-mm wall thickness), (As2O3), a poison for H atom recombination, at a
designated AM-2, was supplied by the CSA Sour concentration of 10 mg/L. Buffered saline solution
Service Task Force. Chemical composition of the steel saturated with H2S was not used for determining D
is given in Table 1. because of the formation of iron sulfide (FeS) film on
H flux was measured using a liquid H monitoring the internal pipe surface and its effect on the diffusion
device (HMD) in a one-sided charging experiment of H. D values were determined at 25°C, 40°C, and
using pipe sections 15 cm long and filled with solution. 60°C, and the activation energy was calculated so that
The pipe section was fitted with polycarbonate covers the data could be used to calculate D at other desired
on top and bottom. The top cover had openings for gas temperatures.
inlet and outlet, as well as for the temperature probe The schematic of the pipeline showing the flow of
and heater. The solutions were deaerated with liquid is given in Figure 1(a). The linepipe steel was the
nitrogen (N2) overnight, saturated with H2S, heated to same steel used in the laboratory experiments. The
60°C, and transferred into the pipe. H flux through the external surface of the pipe was prepared in the same
steel was measured in the following media at 60°C: way as in the laboratory experiments described earlier.
— NACE TM0284, synthetic seawater (ASTM D The HMD was installed at the 3 o’clock position. Solid-
1141) without heavy metal ions, pH range 4.8 to 5.4 state HMDs developed by Morris, et al., and identified
after saturating with H2S; as Ch 4, Pt 181, and Pt 180 were attached to the outer
— NACE TM0177, 5 wt% sodium chloride (NaCl), surface of the pipe at the 3 o’clock, 6 o’clock, and 12
0.5 wt% acetic acid, initial pH 2.8 saturated with H2S; o’clock positions, respectively.11 Experimental
— Modified NACE TM0284, pH 3.5 saturated with arrangement of the probes in the field tests is shown in
H2S; and Figure 1(b).
— Field process water, pH 6.9 (before passing
H2S) saturated with H2S.
The design, installation, and operation of the HMD
RESULTS AND DISCUSSION
used have been described earlier.3,9 All paint and
corrosion scale were removed from an area up to COH at the inside surface of the pipe was
80 mm from the center of the monitoring location. The determined using the HMD attached to the outside
area to be monitored (32-mm diam) was polished with surface of the pipe and by recording the H permeation
600-grade paper to remove pits and scratches. An curve as a function of time. A schematic curve is
organic solvent was used to clean the surface. shown in Figure 2. Peak COH values were calculated
Masking tape was placed on the area to be monitored, from the observed maximum of the permeation current
and epoxy solution was applied with a fine brush. After using the relationship:
the film was dry, 600-grade paper was used to smooth
the ridges. The masking tape was removed, and the imaxL
monitoring area was degreased using a solvent. peak COH = (1)
DF

CORROSION–Vol. 50, No. 8 637

CORROSION ENGINEERING

(a) (b)

FIGURE 1. (a) Schematic of H probe arrangement in field test and (b) photo of the experimental arrangement.

L2
Dt = (3)
b 15.3 tb

where L is the thickness, and tb is the breakthrough

time obtained from H permeation curves in both the
laboratory and field tests.
COH at the inside surface of the pipe was
calculated using the same imax for any given test
solution but for different values of D (Table 2):
— DE obtained from the electrochemical charging
method,
— Dtb1 obtained by averaging all the values
obtained from laboratory and field tests using
FIGURE 2. H permeation curve. breakthrough time, and
— Dtb2, averages of all D values obtained for five
different linepipe steels using tb.
COH values increased with increasing severity
where imax is the permeation current density, L is the of the medium (i.e., decreasing pH). DE of 4.8 x
wall thickness of the steel pipe, D is the diffusion 10–6 cm2/s was slightly higher than the Dtb1 value of
coefficient of H in the steel at the operating 4.6 x 10–6 cm2/s. This was reflected in very nearly
temperature, and F is the Faraday constant. As the same values for COH. The Dtb2 value of 7.47 x
expected, the inhibitor addition resulted in reduced imax. 10–6 cm2/s was higher than the DE value of 4.8 x
D was calculated using the relationship: 10–6 cm2/s. Using the value of 7.47 x 10–6 cm2/s
resulted in decreased values of COH, as shown in
0.5 Table 2. COH values were in the range of 3.2 µmol/cm3
i(t1) t2 –L2 1 1 to 8.4 µmol/cm3, which was in the expected range.
= xe – (2)
i(t2) t1 4D t1 t2 COH obtained in the field using the three D values
are given in Table 3. Comparison of the COH values
where i(t1) and i(t2) are the current densities at times t1 obtained in the field tests with the laboratory test
and t2, respectively. Four pairs of values of i(tn) and tn results showed that the most aggressive field
were taken from the permeation curves [i(t1) = 5% imax, environment resulted in a COH value between those in
i(t2) = 10% imax, i(t3) = 15% imax, and i(t4) = 20% imax] and NACE TM0284 and in modified NACE TM0284 test
were used to calculate six values of D, from which an environments. In other words, the field environment
average value was calculated. D values also were was more severe than NACE TM0284 and less severe
calculated using the relationship: than modified NACE TM0284.

638 CORROSION–AUGUST 1994

 CORROSION ENGINEERING

TABLE 2
Laboratory Test Data on COH in AM-2 Steel at 60°C(A)
imax L/F DE x 10–6 Peak COH Dtb1 x 10–6 Peak COH Dtb2 x 10–6 COH
Medium pH (x 10–5) (cm2/s) (µmol/cm3) (cm2/s) (µmol/cm3) (cm2/s) (µmol/cm3)

Field-produced water 6.9 2.4 4.8 5.0 4.6 5.2 7.47 3.2
NACE TM0284 5.4 4.03 4.8 8.4 4.6 8.8 7.47 5.4
Modified NACE TM0284 3.5 4.8 4.8 10.0 4.6 10.4 7.47 6.4
NACE TM0177 2.8 6.24 4.8 13.0 4.6 13.6 7.47 8.4

(A)
Solution presaturated with H2S and no inhibitor added.

TABLE 3
Field Test Data for H in AM-2 Steel
imax L/F DE x 10–6 Peak COH Dtb1 x 10–6 Peak COH Dtb2 x 10–6 COH
Medium (x 10–5) (cm2/s) (µmol/cm3) (cm2/s) (µmol/cm3) (cm2/s) (µmol/cm3)

Well A, 9.5% H2S 4.32 4.8 9.0 4.6 9.4 7.47 5.8
Well B, 15.8% H2S 2.70 4.8 5.6 4.6 5.9 7.47 3.6
Well C, 17.6% H2S(A) 1.44 4.8 3.0 4.6 3.1 7.47 1.9

(A)
With added inhibitor.

COH values obtained by using the three D values

are plotted against the pH of the medium in Figure 3.
The data show that the two lines using D values of
4.8 x 10–6 and 4.6 x 10–6 cm2/s were nearly the same.
The data corresponding to the Dtb2 value of 7.47 x
10–6 cm2/s gave a line with COH values lower than the
values calculated using the DE value of 4.8 x 10-6 cm2/s,
but within the range expected from this type of study.
COH values obtained in the field test at well A
corresponded to COH at a pH value of 4.4 to 4.7. This
suggested a laboratory test medium buffered to a pH of
4.4 to 4.7 would reflect the field conditions more
closely than NACE TM0284 or modified NACE
TM0284. Similar analysis of data available in the
literature showed that a test medium of pH of 4.1 was
likely to reflect some site-specific field conditions better
than NACE TM0284 or modified NACE TM0284 of pH
3.5.12
FIGURE 3. COH vs pH of test medium: ⌬ = D value 4.6 x
It was obvious that the Dtb1value obtained from the 10–6 cm2/s, ● = D value 4.8 x 10–6 cm2, and 䡬 = D value 7.47 x
permeation curve for calculating the COH value could be 10–6 cm2/s.
used in field tests instead of obtaining D by a
complicated and tedious electrochemical charging
method, without sacrificing accuracy. It also was clear
that a laboratory test medium buffered to a pH of 4.1 to values. Estimated values of DE then can be used to
4.7 would reflect field conditions more closely. calculate COH values without sacrificing accuracy.
Figure 4 shows the ratio DE/Dtb2 plotted against the
thickness of three linepipe steels of different chemical CONCLUSIONS
compositions at 60°C. At 60°C, the ratio DE/Dtb2
changed from 0.90 to 0.55 in going from 1.7 mm to ❖ Comparison of COH measured at the inside surface
4.3 mm in thickness. These data can be used to of sour service linepipe steel in the laboratory and in
estimate the value of DE from a knowledge of Dtb2 the field showed a laboratory test solution buffered to

CORROSION–Vol. 50, No. 8 639

CORROSION ENGINEERING

pH 4.1 to pH 4.7 saturated with H2S would reflect field

conditions closely.
❖ D values obtained from tb could be used to calculate
COH in field tests without sacrificing accuracy.
❖ Analysis of tb data can be informative in giving some
idea on the relative trap density in linepipe steels.

ACKNOWLEDGMENTS

The authors acknowledge partial funding of this

project by CSA Sour Service Task Force and field work
by Y. Lafrenière.

REFERENCES

1. G.J. Biefer, MP 21, 6 (1982): p. 19.

2. R.W. Revie , V.S. Sastri, G.R. Hoey, R.R. Ramsingh, D.K. Mak, M.T.
Shehata, Corrosion 49, 7 (1993): p. 531-532.
3. R.W. Revie, V.S. Sastri, M. Elboujdaini, R.R. Ramsingh, Y.A. Lafrenière, FIGURE 4. Ratio of DE/Dtb2 vs thickness (t): 䡬 = D value
Corrosion 49, 7 (1993): p. 531-535.
4. A. Ikeda, T. Kaneke, I. Hashimoto, M. Takeyama, Y. Sumitomo, et al., 7.47 x 10–6 cm2/s.
“Development of Hydrogen-Induced Cracking (HIC) Resistant Steels and
HIC Test Methods for Hydrogen Sulfide Service,” T. Proc. Symp. Effect
of Hydrogen Sulfide on Steel, 22nd Ann. Conf. Metallurgists, Canadian
Institute of Mining and Metallurgy, Edmonton, Alta., Canada , Aug. 22-
24, 1983 (Montreal, Canada: CIM, 1983). 8. NACE Standard TM0284-87, “Evaluation of PIpeline Steels for
5. W. Bruckoff, “Rupture of a Sour Gas Line Due to Stress-Oriented Resistance to Stepwise Cracking” (Houston, TX: NACE, 1987).
Hydrogen-Induced Cracking — Failure Analyses, Experimental Results, 9. M.G. Hay, “An Electrochemical Device for Monitoring Hydrogen Diffusing
and Corrosion Prevention,” CORROSION/85, paper no. 389 (Houston, Through Steel,” CIM Conference of Metallurgists, Hydrogen Sulfide
TX: NACE, 1985). Symposium, Edmonton, Alta., Canada, 1983 (Montreal, Canada: CIM,
6. G.R. Hoey, R.W. Revie, R.R. Ramsingh, D.K. Mak, M.T. Shehata, 1983).
“Determination of Minimum Diffusible Hydrogen Concentration Causing 10. M.A. Devanathan, Z. Stachurski, Proc. Roy. Soc. A270 (1962): p. 90.
Hydrogen-Induced Cracking of Linepipe Steels,” Metals Technology 11. D.R. Morris, et al., Ironmaking and Steelmaking 16, 6 (1989): p. 429.
Laboratories Division Report MTL 88-41(TR), CANMET, Energy, Mines, 12. M.G. Hay, “Correlation of Laboratory Hydrogen-Induced Cracking Test
and Resources Canada, Ottawa, Ont., Canada (1988). Environments with Field Sour Gas Environments Using Hydrogen
7. NACE Standard TM0177-90, “Testing of Metals for Resistance to Sulfide Permeation Measurements,” CORROSION/91 paper no. 13 (Houston,
Stress Cracking at Ambient Temperatures” (Houston, TX: NACE: 1990). TX: NACE, 1991).

640 CORROSION–AUGUST 1994