1 Introduction

To achieve the climate goals, so-called green energy is becoming increasingly important. Changing weather conditions lead to fluctuating power outputs, which leads to the fact that the excess electricity needs to be stored. Transformation of electricity to hydrogen via electrolysis is an option, but the lack of adequate infrastructure for the storage and transport of the gas is a problem. This issue can be addressed by methanation of hydrogen and carbon dioxide to obtain natural gas:

$$4\,\mathrm{H}_{2}+\mathrm{CO}_{2}\rightarrow \mathrm{CH}_{4}+2 \mathrm{H}_{2}\mathrm{O}$$
(1)

A new approach is to use methanogenic archaea that perform the methanation process [1]. Such microorganisms can produce natural gas in an underground natural gas reservoir. With H2 involved in the process, hydrogen embrittlement has to be considered. This phenomenon, although known since 1874 [2], is not yet fully understood. In presence of H2 and CO2 as well as a chloride-containing, humid environment under anaerobic conditions, there are two main potential sources of absorbed hydrogen. The first one is the dissociation of gaseous H2:

$$\mathrm{H}_{2}\leftrightarrow 2 \mathrm{H}_{\mathrm{ad}}$$
(2)

The hydrogen molecule H2 dissociates into two adsorbed hydrogen atoms Had. According to Sieverts and Krumbhaar [3], the hydrogen solubility of metals increases with increasing temperatures. This is depicted in Sieverts’s law:

$$S=S_{0}\cdot \sqrt{p}\cdot e^{-\Updelta H/RT}$$
(3)

where S0 is the solubility constant, p the partial pressure, ∆H the heat of solution, R the universal gas constant, and T the absolute temperature [4]. When CO2 dissolves in water, carbonic acid H2CO3 is formed [5]:

$$\mathrm{CO}_{2}+\mathrm{H}_{2}\mathrm{O}\leftrightarrow \mathrm{H}_{2}\mathrm{CO}_{3}$$
(4)

The carbonic acid dissociates. The increasing concentration of H+ results in a lower pH, and the cathodic reaction

$$2\,\mathrm{H}^{+}+2 e^{-}\leftrightarrow \mathrm{H}_{2}$$
(5)

is promoted [6]. Simultaneously an anodic dissolution of iron takes place:

$$\mathrm{Fe}\leftrightarrow \mathrm{Fe}^{2+}+2 e^{-}$$
(6)

Not all of the reduced H+ ions recombine to H2 as shown in Eq. 5, some of the adsorbed hydrogen atoms (Had) can get absorbed (Hab):

$$\mathrm{H}^{+}+e^{-}\leftrightarrow \mathrm{H}_{\mathrm{ad}}\leftrightarrow \mathrm{H}_{\mathrm{ab}}$$
(7)

Thus, the second potential source of absorbed hydrogen is the cathodic corrosion reaction. In general, this source can be neglected when using corrosion resistant alloys (CRAs).

Whiteman and Troiano [7] stated that the amount of absorbed hydrogen necessary to produce hydrogen embrittlement is one or two orders of magnitude greater for austenitic stainless steels compared to steels with a bcc lattice. Thus, a duplex stainless steel, which is a mixture of both microstructures, is of interest for being tested to identify application limits in terms of hydrogen embrittlement.

The diffusion of hydrogen is much faster in ferrite compared to austenite, while the solubility of hydrogen atoms shows a reverse behaviour [8, 9]. Olden et al. [8]. stated that austenite acts as a hydrogen sink in a duplex stainless steel, attracting and absorbing the hydrogen atoms that diffuse in the nearby ferrite. Tohme et al. [10] recently confirmed this behaviour combining Scanning Kelvin Probe Force Microscopy (SKPFM) measurements with finite element (FE) modelling. The diffusion against the concentration gradient is a phenomenon known as “up-hill” diffusion, described by Darken [11] for carbon in steels as early as 1949.

Since it can be challenging to make suggestions for field service based on standard laboratory tests [12], simulating field conditions is very reasonable. To answer the question of whether steel grades are suitable for underground methanation, a routine for testing has been developed. An autoclave test bench was designed to simulate field conditions. In this paper, tests were conducted with the 2205 duplex stainless steel (UNS S31803).

2 Experimental Procedure

The tested steel grade presented in this paper was the duplex stainless steel UNS S31803, commonly known as 2205. Samples were taken from commercially available casing tube sections. The chemical composition of the investigated material is shown in Table 1.

TABLE 1 Chemical composition of the investigated material

The steel had a pitting resistance equivalent number (PREN) of 35.7. This value was calculated according to the following equation [13]:

$$\text{PREN}=\mathrm{Cr}+3.3 \mathrm{Mo}+16\,\mathrm{N}[\mathrm{wt}\% ]$$
(8)

The mechanical properties of the investigated steel grade are listed in Table 2. Tensile tests were performed on small, non-standard tensile specimens with an initial gauge length of 25 mm and a diameter of 3 mm. The specimens were drawn at room temperature with a crosshead speed of 0.1 mm/min.

TABLE 2 Specified Minimum Yield Strength (SMYS), Yield Strength (YS), Ultimate Tensile Stress (UTS) and fracture elongation (A) of the investigated material

The corrosion tests were conducted within autoclaves made of UNS N06625 (Alloy 625). Fig. 1 shows one of the used autoclaves.

Fig. 1
figure 1

Autoclave made of Alloy 625

Each autoclave contained three different specimens: an immersion specimen (Fig. 2a) i) for measuring the hydrogen uptake, a coupon (Fig. 2a) ii) for determining the presence of pitting or other corrosion phenomena, and a small tensile specimen (Fig. 2a) iii) for a constant load test (CLT). The load was applied to the CLT specimen with a spring made of a cobalt-base alloy and ceramic nuts (Fig. 2b), the latter ensuring electronic decoupling of the specimen from the more noble spring. The specimens were connected with PTFE parts.

Fig. 2
figure 2

Specimen assembly to be mounted in the autoclave. a Three specimens connected with PTFE parts: Immersion specimen (i), coupon (ii), and constant load specimen (iii). b Small tensile specimen with spring and ceramic nuts for constant load test

The specimen assemblies were mounted in the autoclaves. The vessels were evacuated and purged with argon several times to get very low partial pressures of oxygen and other atmospheric gases. Further the autoclaves were filled with an aqueous test solution and test gases (Fig. 3a, b). The last step was to mount the autoclaves on rotating shafts within a heated chamber (Fig. 3c).

Fig. 3
figure 3

Filling and mounting of the autoclaves: a Filling of autoclaves with aqueous test solution. b Pressing the test gas into the autoclave. c Several autoclaves mounted on rotating shafts within a heated chamber

Tests were performed with two different partial pressures of H2 gas: 20 bar and 100 bar. Also, the influence of 5 bar of CO2 gas was investigated. In more than half of the tests, an aqueous NaCl solution (brine) with a chloride concentration of 15,000 mg/l was used. The tests were conducted at 25 °C as well as at 80 °C (near field conditions) and lasted for 30 days. Thus, the material was tested under 10 different conditions.

The load for the CLTs was 90% of the specified minimum yield strength (SMYS). The load was applied by compressing a spring with a defined load and connecting it with the respective specimen and two ceramic nuts. To simulate the periodically changing conditions in the gas well, the autoclaves were rotated with a speed of 1 RPM. Consequently, the specimens were periodically wetted with the aqueous electrolyte.

Directly after the test, the immersion specimens were removed from the autoclaves and immediately cooled in liquid nitrogen. The cooled specimens were ground with silicon carbide paper (grit 120), rinsed with acetone and blow-dried quickly prior to hydrogen analysis. The hydrogen content was measured with a thermal conductivity cell after hot extraction at 950 °C. At the end of the tests, the constant load specimens were examined for possible fractures that occurred. Unbroken specimens were examined with a stereo microscope to see if any cracks or pits were visible. The coupons were examined for pits or other visible corrosion phenomena.

3 Results

None of the specimens loaded at 90% of the SMYS broke under the tested conditions. None of the unbroken specimens showed visible cracks or pits under the stereo microscope. Fig. 4 gives an overview for the results of the constant load tests.

Fig. 4
figure 4

Results of the constant load tests

Also, on the corrosion coupons, no pits or other visible corrosion phenomena were detected. The results of the hydrogen analyses of the 2205 tested in autoclave wheel tests at 25 °C are illustrated in Fig. 5. Where two or more tests under same conditions were performed, the mean value and standard deviation are shown in the diagrams. For individual tests, a deviation of ±15% is shown.

Fig. 5
figure 5

Hydrogen content for 2205 duplex stainless steel (UNS S31803) tested in different media in autoclave wheel tests for 30 days at 25 °C and 1 RPM. The chloride concentration of the brine was 15,000 mg/l

In comparison to the uncharged condition, which is shown in each diagram as the leftmost bar, a hydrogen uptake by dry H2 gas at 20 bar or 100 bar and 25 °C was not clearly detectable. A similar result was found for 20 bar of H2 and 5 bar of CO2 gas under wet conditions (with brine). At 100 bar of H2 in the presence of brine, a hydrogen content of 4.01 ppm was detected, which corresponds to an uptake of 1.76 ppm. Hydrogen contents for tests in the same media, but at 80 °C are shown in Fig. 6.

Fig. 6
figure 6

Hydrogen content for 2205 duplex stainless steel (UNS S31803) tested in different media in autoclave wheel tests for 30 days at 80 °C and 1 RPM. The chloride concentration of the brine was 15,000 mg/l

At 80 °C there was no detectable hydrogen uptake in 20 bar dry H2 gas or in the wet, CO2 containing atmosphere, while 100 bar dry H2 led to a measurable hydrogen uptake. In the presence of brine, H2 gas led to a hydrogen content of 6.00 ppm (3.75 ppm uptake) at 20 bar and 14.21 ppm (11.96 ppm uptake) at 100 bar.

Both at 25 °C and 80 °C, the presence of NaCl brine with a chloride concentration of 15,000 mg/l was found to promote the hydrogen uptake due to H2 gas. The higher temperature (80 °C compared to 25 °C) led to significantly higher hydrogen contents in the specimens exposed to 20 bar or 100 bar H2 gas and brine. Dry H2 gas only led to hydrogen uptake at 100 bar and 80 °C, although the measured amount was not significant.

Using stereomicroscopy, no signs of corrosion were found on any coupon. The presence of 6.00 and 14.21 ppm after testing with brine at 80 °C compared to 2.77 and 3.08 ppm without brine is a strong argument for hydrogen uptake during a corrosive process. While the α‑phase is less noble, the γ‑phase is more noble [14]. In the brine, the α‑phase serves as the anode, while the γ‑phase represents the cathode at which adsorbed atomic hydrogen is formed. Only when there is a high hydrogen partial pressure, the adsorbed hydrogen cannot recombine against this “back pressure” and is absorbed resulting in the high described concentration (6.00 and 14.21 ppm). Lack of brine and/or hydrogen pressure in the autoclave does not result in a substantial hydrogen absorption. Without brine, corrosion may not occur. Corrosion may occur without the presence of hydrogen gas, but adsorbed atomic hydrogen then tends to recombine to its molecular form which is released to the gas.

Although hydrogen contents up to 14.21 ppm were measured, no substantial embrittlement at 90% of the SMYS was observed in the constant load tests. In previous tests on carbon steels, it was found that the API 5CT [15] P110, which has the same SMYS as the duplex, had a hydrogen content of 0.38 ppm after being tested under same conditions without a failure. The significant difference in hydrogen content can be explained with the different microstructure (tempered martensite vs. ferrite-austenite). Austenitic structures provide more hydrogen traps, where the small atoms can get stuck. Since the amount of absorbed hydrogen required to produce hydrogen embrittlement of these structures is one or two orders of magnitude greater than that of steels with bcc lattice alone, the mixture of both—present in the investigated material—survives the CLT despite a hydrogen content of 14.21 ppm.

In tests on the austenitic stainless steel Alloy 28 (UNS N08028), a maximum hydrogen uptake of 7 ppm was found, while in this work on 2205 duplex stainless steel a hydrogen uptake of 12 ppm was measured after charging under the same conditions. Despite the much lower austenite content with its high solubility for hydrogen atoms, the 2205 absorbs more hydrogen than the fully austenitic material. The unique behaviour of duplex stainless steel can be described with the up-hill diffusion phenomenon mentioned in the introduction: due to the rapid diffusion, the ferrite transports the hydrogen atoms very well to the austenite, where the solubility and thus the driving force for the diffusion into it is very high. The ferrite seems to permanently “pump” hydrogen atoms into the austenite, resulting in higher hydrogen uptake compared to a martensitic or fully austenitic material. This effect seems to be further enhanced by the galvanic corrosion of the α‑phase of the duplex steel being coupled to the more noble γ‑phase as described above.

4 Conclusions

  • No cracks occurred under the constant load of 90% SMYS within 30 days of testing in rotated autoclaves containing up to 100 bar hydrogen gas with or without brine (15,000 mg/l chloride).

  • The 2205 duplex stainless steel (UNS S31803) did not show a corrosive attack after the immersion in brine with CO2 at 25 °C and 80 °C (evaluation with stereo microscope).

  • The presence of brine was found to promote hydrogen uptake.

  • A temperature of 80 °C led to significantly higher hydrogen contents in the specimens exposed to 20 or 100 bar H2 gas and brine compared to 25 °C.

  • A maximum hydrogen content of 14.21 ppm (uptake of 11.96 ppm) was found in the test with 100 bar H2 and brine at 80 °C after 30 days.

  • In the hydrogen uptake, significant differences to other steels were measured. The two-phase α‑γ-microstructure and thus the individual solubility and diffusion behaviour as well as the different corrosion potential have a decisive influence on the hydrogen uptake during the 30 days of testing.

  • The 2205 duplex stainless steel (UNS S31803) seems to be suitable for application in an underground microbiological methanation facility.