Influence of Pre-oxidation on Filamentary Carbon Deposition on 20Cr25Ni Stainless Steel
- 70 Downloads
The results of a pre-oxidation heat treatment at 930 °C in Ar/H2/H2O environments on a Si-bearing Nb-stabilised 20Cr25Ni austenitic steel are presented. The heat treatment was conducted under low pO2, achieved by the introduction of controlled amounts of moisture into the gas. The atmosphere promoted the formation of a continuous, dense, adherent, protective surface scale composed of Cr2O3 and MnCr2O4 with a thin Si-rich oxide at the oxide alloy interface. Samples with different oxide layer thicknesses were produced and further exposed at 700 °C, to a gas of nominal composition CO2/1%CO/1000 vpm C2H4 for 4 h. This gas mixture has a carbon activity greater than unity and readily forms filamentary carbon on the non-pre-oxidised alloy. This is catalysed by nickel particles formed intrinsically from the alloy during the early stages of oxidation of the unprotected surface. The oxide layers produced, as a result of the pre-oxidation process, could suppress carbon deposition onto the alloy; a significant reduction in carbon deposit was noted with an oxide of 125 nm thickness, and no deposit was found on the sample with an oxide thickness of 380 nm. The depth of depletion of chromium from the alloy correlated with the thickness of the oxide formed during the pre-oxidation heat treatment, but the chromium concentration at the oxide/metal interface remained at ~ 15–16 wt% and considered to be sufficient to reform a protective layer in the event of mechanical damage to the original. No additional chromium depletion of the alloy occurred during the 4-h deposition stage.
KeywordsPre-oxidation Carbon deposition Chromia Stainless steel
It has been recognised for many years that selective pre-oxidation of chromium-bearing alloys to form a chromia layer reduces or eliminates carbon deposition and carburisation in CO2-based environments [1, 2]. This benefit has also been observed in the steam cracking process where the onset of coking was delayed by the formation of a chromia layer at the surface of alloys . Studies to date have employed micrometre-thick chromia layers which can result in substantial depletion of chromium in the underlying alloy [1, 2, 4, 5]. Such depletion profiles are characteristic of the selective oxidation process  and can prejudice the ability of the alloy to reform a healing layer should the original layer crack or spall [7, 8].
The purpose of the present study was to examine the effect of thin sub-micrometre chromia layers on carbon deposition on a Nb-stabilised 20Cr25Ni austenitic stainless steel. It is known from previous studies on this alloy [2, 9] that such carbon deposition is filamentary in nature and is catalysed by nickel-rich metallic particles produced during the initial stages of oxidation in, typically, ethene-bearing CO2/1%CO gas. The catalytic nature of such particles has been well documented in the literature [10, 11, 12, 13, 14]. This process does not lead to carburisation of the underlying alloy as in “metal dusting” [15, 16, 17] but may be a precursor to this.
The aim of this paper is to present the results of a study designed to prevent the formation of the catalytic particles by the early development of a protective, continuous surface layer of chromia. The intention is to examine the efficacy of oxide layers appreciably thinner than those used previously [2, 4, 5]. The results presented here will find relevance in applications which may benefit from the formation of a highly protective oxide layer formed during a pre-oxidation heat treatment, e.g. solid state fuel cells and heat exchangers. The work here focuses on the effect of the pre-oxidising heat treatment on the composition of the initial oxide formed and also the effect this has on the early stages of exposure to an atmosphere with a high carbon activity. The presence or otherwise of filamentary carbon deposits is used as an indicator of the effectiveness of this approach in producing a highly protective continuous layer of chromia across the surface of the alloy.
A section of commercial alloy in the form of a can with a diameter of 15 mm and wall thickness of approximately 0.4 mm was supplied. The nominal composition of the alloy was 20.0Cr, 25.0Ni, 53.1Fe, 0.7Nb, 0.6Mn and 0.56Si wt%. The can was sectioned into approximately 15 mm long ring specimens using a slow speed saw and this was in turn sectioned into three samples of approximate final dimensions 15 mm by 10 mm. Great care was taken during the sectioning process to avoid stresses being induced from clamping of the can. Burrs produced during the sectioning process were removed using 1200-grit SiC paper, thus removing any possible extraneous sources of particles that might catalyse the deposition process. The samples were cleaned in ethanol in an ultrasonic bath prior to being subjected to the standard heat treatment for this alloy, i.e. 930 °C for 30 min.
For the heat treatment process, the samples were placed in alumina boats and positioned in a controlled gas, heat treatment furnace with extraction. The rig was sealed and checked for gas tightness. All sections of the circuit of the rig were purged for a total of 2 h using a dry gas mixture of 5% H2 in Ar at a flow rate of approximately 0.3 l min−1. Following this, the temperature of the furnace containing the samples was brought up to 930 °C, at a heating rate of 20 °C min−1. To introduce moisture into the gas mixture, the pipework had been split before entering the heat treatment furnace to provide two routes controlled by a valve, and this diverted the gas through a flask containing deionised water. It should be noted here that the bypass circuit and flask were included in the purging sequence. The valve was used to direct all or some of the gas through the flask, and in this way it was possible to control the moisture levels achieved. The moisture content of the gas was measured continuously during the heat treatment period using a Vaisala dew point and temperature transmitter DMT342 hygrometer sited at the entrance to the annealing furnace. The samples were held in this environment for 30 min, after which the alumina boat was moved, within the sealed furnace, to a forced cooled section and allowed to cool to room temperature before removal and storage.
Moisture levels in heat treatment gas with calculated pO2 from Eq. 2
27 ± 3
2 × 103
6.1 × 10−23
100 ± 15
5 × 102
8.4 × 10−22
2030 ± 75
2.5 × 101
3.4 × 10−19
The dissociation oxygen activities of NiO, Cr2O3 and Fe3O4 are 7.8 × 10−17, 2.7 × 10−32, and 2.8 × 10−22, respectively, calculated for the pure substances . The activities of the elements in the alloy will be less than unity so this approach provides a conservative estimate of the reactivity of the individual gas mixtures. From the calculated values given in Table 1, it can be seen that NiO is not expected to form in any of the three annealing gas mixtures used. However, Cr2O3 is expected to form in all cases and Fe3O4 is only expected to form in the gas mixture with the highest moisture content, i.e. the highest pO2.
Following the oxidising heat treatments, the samples were moved within the furnace out of the hot section of the furnace and allowed to cool to room temperature under the flowing gas before removal. Selected samples were further exposed to an atmosphere of nominal composition 1%CO, 1000 vpm C2H4, balance CO2 at 1 atmosphere total pressure at a temperature of 700 °C for 4 h. Details of the exposure protocol are as follows. The samples were placed in alumina boats and positioned within a purpose built rig at room temperature, and the rig is sealed and checked for gas tightness. The rig was purged with the same gas as before, i.e. 5% H2 in Ar, for 2 h. The rig consisted of two furnaces in series. After the initial purge, the first furnace was brought up to 700 °C; this was followed, after a further 1 h purge, by the second furnace, also set to 700 °C. The first furnace housed titanium foil used to remove residual oxygen from the test gases and ensure a low pO2 was achieved. After this, the gases were swapped to the carbon depositing gas and held at 700 °C for 4 h. At the end of this time period, the gas was swapped back to 5% H2 in Ar and the samples were cooled within the furnace. At all stages of the exposure, the flow rate of the gases was approximately 0.3 l min−1.
The samples were examined after the heat treatment stage and also after the final deposition exposure. Examination of the surfaces of the samples was performed using the JOEL 7000F Scanning Electron Microscope fitted with a Field Emission Gun (FEGSEM). This equipment has the capability of resolving the filamentary carbon fibres without the need for additional sample preparation, e.g. coating with a conductive layer. Cross sections were also prepared of each stage in the testing to provide information on the composition and thickness of the surface oxide and also record elemental depletion of the alloy. This analysis was achieved by taking cross sections using a dual-beam FEI Quanta 3D system consisting of a Focussed gallium Ion Beam (FIB) and a conventional field emission scanning electron microscope (SEM) column. Regions, characteristic of the surface of the samples, were selected, extracted and thinned by FIB. The initial stages of foil preparation used a 30-keV ion beam with the probe current reduced successively throughout the procedure, and a final polishing/cleaning was performed at 5 keV. The foils were examined using an FEI Tecnai F20 STEM equipped with an Oxford Instrument X-max SDD detector operating at 200 kV.
Oxidising Heat Treatment
Thickness of oxide layers and depth of chromium depletion related to pre-oxidation dew point, after the heat treatment and subsequent deposition stage
− 53 °C
− 41 °C
− 13 °C
Pre-oxidation heat treatment
Oxide thickness (nm)
125 ± 25
265 ± 25
380 ± 30
Cr depletion depth (μm)a
Oxide thickness (nm)
140 ± 20
270 ± 30
390 ± 30
Cr depletion depth (μm)a
The oxide type and distribution were consistent in all samples and thus found to be independent of the moisture level of the gases used. It should be noted that no Fe-rich oxides or (Ni, Fe)-rich particles were found in these samples. The oxides present and the absence of Fe and Ni oxides are consistent with the predictions of oxide formation under the gas compositions used. There was some evidence of localised enrichment of Fe at the alloy oxide surface which is consistent with the selective oxidation process, as shown in Fig. 6. The concentration of Cr within the alloy is sufficient, after the pre-oxidising heat treatment, to form a healing layer of Cr2O3 should the oxide be damaged. In this regard, the oxidation behaviour of the alloy would not be compromised by the pre-oxidation treatment.
The catalytic nature of Ni-rich particles and their role in carbon deposition has been well documented [2, 11]. A mechanism for the production of nanometre-sized particles from the 20Cr:25Ni austenitic stainless steel has been proposed earlier . This involved the selective, internal oxidation of Fe and Cr from the alloy under low oxygen environments. Under these conditions, Ni would not oxidise but instead becomes concentrated as particles within the mixed oxides of Cr and Fe. On exposure to a gas with a high carbon activity, filamentary carbon deposits form on the surface of the alloy, preferentially at the centre of exposed grains. It has been shown more recently that these regions of the grains experience greater internal oxidation . The results presented here demonstrate a route that prevents this mechanism occurring by the early formation of a surface chromia layer.
The composition of the surface oxide formed during the oxidising heat treatment has been confirmed by TEM cross-sectional analysis to be composed of Cr2O3 with grains of a (Cr, Mn) spinel. Underlying these, a thin layer of Si was identified as was an enhancement of Nb. Significantly, no oxides of Ni or Fe nor metallic particles of these elements were found within the oxide. Calculation of the activities of the oxygen in the annealing gases amply explains the oxides observed, as oxides of Si, Cr, Mn and Nb are all thermodynamically stable under the conditions used here. It could be expected that the alloy heat-treated using the gas with the highest moisture content and thus highest oxygen activity would form Fe3O4. The lack of evidence of iron oxides in the surface oxide would suggest that the more protective oxides form rapidly and dominate the surface, thereby reducing the oxygen activity at the oxide/alloy interface still further to that in equilibrium with those more stable oxides.
The uniformity in the thickness and composition of the oxide across the surface and the lack of evidence of any enhanced grain boundary depletion shows that the conditions used for the oxidising heat treatment favour the formation of a uniform oxide across the whole surface of the alloy. This is in contrast to that observed for this alloy when exposed to lower temperatures, 500–700 °C, where relatively faster grain boundary diffusion compared with that in the lattice results in the formation of a protective surface oxide of chromium at the emerging grain boundaries which grow laterally [2, 6, 9, 25] until coverage of the surface is achieved . Once an oxide has been established on the surface of the alloy, the continued growth will be dominated by diffusion through the oxide and, predominantly, via grain boundaries [29, 30, 31, 32].
It has been known for many years that the formation of a chromia layer of a few micrometres in thickness can protect the underlying alloy from carbon deposition and carburisation in atmospheres of high carbon activity (> 1). This pre-oxidation procedure can lead to substantial and deleterious Cr depletion of the underlying alloy, however. The purpose of the present work was to examine whether much thinner oxide layers could also provide useful protection against carbon deposition.
A technique has been developed that produces adherent layers of chromia, together with crystals of MnCr2O4, on an austenitic steel pre-oxidised at 930 °C in Ar/H2/H2O mixtures. By control of the moisture content of this gas, oxide layer thicknesses in the range 125 nm to 350 nm were produced. This resulted in depletion of Cr from the alloy, but the remaining concentration was sufficient to ensure rehealing occurred if the surface oxide was damaged.
The ability of these layers to protect against carbon deposition was examined by exposure to a gas mixture of CO2/1%CO/1000 vppmC2H4 at 700 °C for 4 h. It has been shown that a significant reduction in filamentary carbon deposition was achieved with surface oxides as thin as 125 nm, although isolated carbon fibres were found associated with machining defects and pores in the oxide layer. Total suppression of carbon formation was achieved on the sample which possessed an oxide thickness of approximately 350 nm.
The authors are grateful to EDF Energy for their financial support for this work. The support provided by the Centre for Electron Microscopy at the University of Birmingham is also acknowledged.
- 4.J. E. Bainbridge, G. W. Horsley, J. A. Cairns, and P. Campion, UK patent: GB 2 097 821A. Carbon Deposition Inhibition (1982).Google Scholar
- 15.H. J. Grabke, R. Krajak, and J. C. Nava Paz, Corrosion Science 35, 1141 (1993).Google Scholar
- 18.I. Barin, O. Knacke, Thermochemical Properties of Inorganic Substances (Springer, Berlin, 1973).Google Scholar
- 19.E. K. Storm and N. H Krikirian, The Journal of Physical Chemistry; 63, 1747 (1959).Google Scholar
- 28.P. W. G. Simpson and H. E. Evans, British Nuclear Energy Society 265 (1985).Google Scholar
- 29.A. Atkinson and R. I. Taylor, in Transport in Nonstoichiometric Compounds, eds. G. Simkovich, V. S. Srubican (Plenum Publishing Corporation, New York, 1985) p. 285.Google Scholar
OpenAccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.