Reservoir souring

Reservoir souring is the phenomenon when there is an increase of mass of hydrogen sulfide (H₂S) per unit mass of total produced fluids due to activities of sulphate reducing bateria (SRB) as a result of waterflood. Hydrogen sulfide is extremely toxic and corrosive and there is H₂S level requirement in the sales gas (<4ppm). The H₂S generation due to waterflood must be evaluated for any waterflooded project to ensure proper project design.

The phenomenon of unexpected increase in hydrogen sulphide concentrations in produced fluids from petroleum reservoirs has been observed in different parts of the globe. In recent years at least two major North Sea oil fields are reported to have recorded higher concentrations of hydrogen sulphide in produced fluids after seawater breakthrough occurred. Conversely there are still several fields with mature waterfloods, for example Forties, that have suffered few souring problems.

Mechanisms of souring

Sulphate Reducing Bacteria

Sulfate reducing bacteria (SRB) can be traced back to 3.5 billion years ago and are considered to be among the oldest forms of microorganisms, having contributed to the sulfur cycle soon after life emerged on Earth. These organisms "breathe" sulfate rather than oxygen, in a form of anaerobic respiration.

SRBs are widely distributed in oil production facilities and in seawater, which makes their introduction into water-flooded reservoirs. It has a wide range of metabolic mechanisms which allows sulphate reduction to proceed under many different environmental conditions at the expense of a range of electron donors and carbon sources. Under optimal conditions, it has been estimated that a sphere of porous rock approximately 7 metres in radius could support an SRB population capable of producing the 400 kg of H₂S per day observed for badly soured wells.

that under favourable conditions SRB's will convert SO4- sources to H₂S. In general, SRB, obtain energy for growth and reproduction from the oxidation of a range of organic materials which also serve as sources of carbon. Since SRB grow in the absence of oxygen, the oxidation of organics, such as acetic acid, is linked to the reduction of sulphate:

CH₃COO^- + SO₄^2- → 2HCO₃ + HS^-

SRB's utilise suitable carbon and energy sources in oxidising an organic substrate. It donates an electron, along an electron transport chain. Sulphate acts as the electron acceptor, being reduced to sulphide. Some SRB’s are able to use hydrogen (via hydrogenese enzymes) rather than organic compounds as electron donors.

4H₂ + SO₄^2- + H⁺ → 4H₂O + HS^-

In this case, the requirement for carbon is satisfied by organic compounds or from the fixation of carbon dioxide. This consumption of hydrogen is one way in which SRB are implicated in corrosion events in the oil industry.

Reduction of SRB Protection by Bisulphite Chemicals

To protect process and well completion of equipment against excessive corrosion, oxygen levels in sea water need to be reduced prior to injection. Mechanical means, gas counter current stripping or vacuum deaerators, are used. These achieve reductions to around 100 50 parts per billion(ppb). It is usual to add, continuously, small amounts of ammonium bisulphite or sodium sulphite to chemically scavenge further oxygen. Some operations, NH₄HSO₃ is often only used during maintenance shut downs of the mechanical units.

Such method can cause sulphite scavengers to react with the residual chlorine used as the primary biocide, thereby reducing protection against SRB's. Then a direct reaction of bisulphite/sulphite with the steel process equipment and tubing can also occur. The nature of the reduction reaction has been investigated, on both NH₄HSO₃ and Na2SO3.

Several qualitative tests have been performed, using mild steel coupons in deaerated 3.5% NaCl solutions, with pH adjusted to 5 or 6 with NaOH/HCl at ambient temperature. Bisulphite concentrations from 0.3 g/l to 12.5 g/l have been used, and the presence of sulphide is the confirmed.

Reactions Type that are suggested:

HSO₃^- + 7H⁺ + 3Fe²⁺ → 3Fe2+ + H₂S + 3H₂O

HSO₃^- + 5H⁺ + 3Fe²⁺ → 2Fe2+ + FeS + 3H₂O

The electrochemical nature of the reaction has been further investigated by potentiodynamic polarisation scans using a gold rotating disc electrode (RDE). Scan rates of 1 mv/sec and rotation rates of 10 Hz are used. The curves show a reduction reaction occurring at a potential of about 0.6 V (SCE) which is concentration dependent (in respect of reaction rate and the equilibrium potential) and substantially under diffusion control. The reaction has been identified as the formation of dithionite

S₂O₄^2- + 2H₂O → 2SO₃^2- + 4H⁺ 2e^-

where:

E_o = 0.416 0.1182 pH + 0.0295 log (SO₃^2-)² / (S₂O₄^2-)

This reaction was confirmed by polarisation of a platinum electrode to 0.70 V in a solution of NaCl/Na2SO3 followed by positive identification of dithionite by spectroscopic analysis.

Thermal Sulphate Reduction

Thermal Sulphate Reduction chemical is the direct reduction of sulphate by hydrocarbons in order to produce hydrogen sulphide. Whether the kinetics of these types of reactions are such that they may contribute to reservoir 'souring' is open to debate, although evidence for this has been growing in recent years.

At temperatures in the range 250 325°C, and modest pressures, many organic compounds are rapidly oxidised with high product yields. A requirement for this reaction is the presence of sulphur species in a lower valance state to initiate the reaction. Any S species of valance less than +6 will initiate, though H₂S appears most successful.

The mechanisms proposed by Toland initially involves the protonation of the sulphate ion:

Without initiation the reaction cannot proceed. However, in the presence of H₂S:

The thiosulphate formed is unstable and decomposes, in acidic conditions, to elemental sulphur and sulphate.

Reactive elemental sulphur is then available for the oxidation of various organic species. The sulphite formed can undergo further disproportionation reactions, oxidising more organic substrate and re generating H₂S to initiate further reaction.

Mineral Solubility

Of the minerals indigenous to petroleum reservoirs only pyrite of the sulphur species is likely to be present in any abundance. Some studies suggest that some minerals can have a scavenging effect upon H₂S.

FeS₂ + 4H⁺ + 2e Fe²⁺ + H₂S

Data on the pK of iron sulphide vary considerably, leading to unreliability of any prediction of the H₂S levels produced. To overcome this the solubility reaction for pyrite can be expressed as:

H+ + FeS₂ Fe²⁺ + HS^- + S_o

This avoids the problems associated with needing an accurate value for the second dissociation constant for hydrogen sulphide. Using this approach the solubility product can be calculated.

Ksp (pyrite) = a_Fe²⁺ . a_HS- / a_H⁺ = 3.98 x 10^-17

giving a pK (pyrite) of 16.4. Using a geochemical equilibrium model, pyrite solubilities can then be calculated for the systems of interest. Over the range of conditions prevalent in the reservoir during production of oil we conclude that it is unlikely that pyrite dissolution can give rise to significant quantities of sulphide species in solution levels are always much less than 1 ppm.

Partitioning of H₂S

Some of the mechanisms discussed have the potential for generating only relatively small amounts of H₂S and appear to be readily discounted. Understanding the partitioning and solubility behaviour of H₂S is important and can simply be exemplified by the following:

For H₂S generated from the produced oil

For H₂S generated from produced water

Where:

WOR = Water Oil ratio.

GOR = Gas Oil ratio

S2^- = Sulphide concentration in ppm (w/v)

4060 = Conversion constant for standard conditions of 15°C, 14.7 psi.

For 'sour' oil the H₂S concentration in gas rises in proportion to the amount present. For 'sour' water rapid increases of H₂S in the gas occur at high water cut for fixed concentrations in the water because of the limited availability of hydrocarbon gas. The distribution of H2S between oil and water, under reservoir conditions will be dependent upon many parameters; most importantly pH, salinity, temperature and pressure.

Mitigation

H₂S Scavengers

Zinc and iron based absorbers are cheap and react quickly, but can cause downstream oil/water separation problems. Aldehydes are cheap but slow reacting; a key point when this must take place between wellhead and separators. Strong oxidisers, like chlorine dioxide, have found favour since they are both quick in reaction and cause little production upsets. However, they are corrosive and require special metallurgies. New organic scavengers have been developed with some success but still requiring further improvements. Currently, triazine scavengers offer the highest efficiency and are building some track record of cost and performance.

Biocide Treatments

Biocide treatments of injection wells have been tried to treat reservoir souring with little success. Failure has been attributed to the biocide not penetrating sufficiently deeply into the reservoir. Chlorine (at a residual level of > 10ppm) has been shown in sand packs to limit bacterial activity close to the inlet, but this could have severe consequences on corrosion in injecting wells.

Nitrate Treatments

Nitrate treatments change the bacterial population by encouraging nitrate reducing bacteria to use up all the available electron donors, preventing SRB’s from using them.

Sulphate Removal from Injection Water

Reverse Osmosis plants are used to reduce sulphate levels in seawater used for injection developments. These were built for scale control, but could have some impact on potential souring. No work has yet been undertaken as far as we know to establish whether such technology, which has enormous up front capital costs, could be effective in limiting reservoir souring by reducing the availability of the prime sources of SRB energy.

Reservoir Souring Prediction

To fully understand and be able to predict reservoir souring accurately, it is important that further progress is made in obtaining relevant data on plausible microbiological and chemical routes and the rate and extent of contributing reactions. It is equally important to account for the relative solubility/partitioning data of H2S between oil/water and gas phases in the reservoir and in process equipment and validate this by reliable analysis.

Modelling Capabilities

Reservoir models for forecasting H₂S production during the life of a field have been developed. Two of the souring forecasting models in common usage in the Oil industry are SourSimRL (Oil Plus Ltd) and Dynamic TVS (Rawwater Engineering Company Ltd). Both models have a track record in the industry; DynamicTVS was first published in 1993 (HMSO Publication OTH 92 385: Oilfield Reservoir Souring, ISBN 0-7176-0637-6) and SourSim (Nace 2006, Paper 06664). There is no public record of any benchmarking of the two models. However, both models are based upon similar input parameters and on separate occasions have been used to forecast the souring propensity of identical injector/producer pairs. The results of these comparisons are confidential to the Operators, however in broad terms the base case model forecasts were in accord. They have been used in several studies to enable metallurgical selection and cost-effective mitigation strategies to be considered.

The basis of the models depend on a number of factors, the most important of which is the rate at which water moves through the reservoir. Other uncertainties include the degree of scavenging by reservoir rock (particularly iron minerals such as siderite), watercut, the type of water produced (seawater versus PWRI/EWI), producing GOR and nutrient supply. Most H₂S models use ‘generic’ nutrient types to represent the various nutrients due to natural variation and uncertainties.

• SRBs + [C,N,P nutrients] + Sulphate ions → Sulphide → H₂S

Generally, the more nutrients and sulphate ions present in injected water, the greater the amount of H₂S generated by SRBs. As formation and injection waters can contain VFAs, ammonium ions, amine and phosphorus compounds (e.g. production chemicals), together with substances such as natural surfactants from produced hydrocarbons, produced water can be much richer in nutrients for bugs than injected seawater. This means that the souring potential associated with PWRI/EWI can be much higher than for seawater injection (by up to a factor of 2-3), provided salinity of the reinjected produced water is low (less than 100,000mg/l TDS). If the salinity is greater than 100,000mg/l TDS, the majority of the bacteria are likely to be killed and further souring will not occur. As H₂S moves through the reservoir, it can interact with the mineralogy, particularly the iron minerals. Sulphide ions react with iron ions dissolving from iron minerals to form iron sulphide solids. This type of reaction causes natural scavenging of H₂S generated by SRBs.

• Sulphide ions + Iron ions → Iron Sulphide precipitates

Reservoir Souring Schematic

The amount of H₂S produced by a well going sour is related to the amount of water produced multiplied by a sour water concentration (usually expressed as ppm w/w). Most of the H₂S is transported through the reservoir in injected water. This sour water concentration can be converted into a calculated gas phase H₂S concentration by partitioning the H2S between the gas, oil and water at surface (or a maximum value can be obtained by assuming all of this H₂S flashes into the gas phase).

• Sour Water Concentration = Mass of H₂S Produced / Mass of Water Produced

Uncertainties

In most cases, it is impossible to quantify the effect of the uncertainties on the overall H₂S profiles due to the number of factors involved.

• Reservoir Heterogeneity and Breakthrough Times

The presence of high permeability streaks or fractures may increase the rate of souring by channelling water through a limited volume of rock to the producers. As Clair consists of a fractured reservoir, the uncertainty in breakthrough time and fractured nature of the reservoir is large.

• Interactions with and Distribution of Mineralogy

The variability of iron mineral distribution and its interaction with injection water will determine how much natural scavenging actually occurs in practice. Other clays and minerals may also have some impact on local pH conditions and interaction with iron species. It is very difficult to quantify precisely how much scavenging is likely to occur in practice, as it is also influenced by kinetics of the dissolution/precipitation/ion exchange reactions taking place. Representation of all these possible interactions with a few adsorption parameters in the Statoil model means that comparison would preferably be made with other fields of similar mineralogies to Clair. As no direct analogue has yet been evaluated from a souring point-of-view, low scavenging potential must be assumed, particularly if fractures are present. For low scavenging potential with no siderite, adsorption parameters suggested in the literature were used for the Statoil model runs. Provided the value selected is in the right ballpark, the overall H2S profile should not be too sensitive to this parameter. Orders of magnitude difference can have a very large impact.

• Surface Partitioning

The influence of kinetics, fluid composition, temperature and pressure on H₂S partitioning between gas, oil and water are not very well known. Estimates from thermodynamic calculations and field experience elsewhere tend to suggest the range of values used in the summary tables will apply in practice. Sensitivity to the actual value of the partition coefficients may be up to a factor of 3 in terms of gas phase H₂S concentrations.

• Effect of Reservoir Temperature

The reservoir temperature may create a smaller region around the injection wells in which the SRB can survive, although most H₂S generation is thought to occur close to the injection wellbore. There is no allowance for temperature or pressure in the Statoil model. The actual temperature of the injection water may have an impact on the rate of conversion of sulphate to sulphide. The variety of bacterial populations (mesophiles, thermophiles, hyperthermophiles) means that these species can survive even at very high temperatures, and if the rate of conversion of sulphate to sulphide is determined mainly by a limiting nutrient, the overall effect of a change in injection water temperature will be small.

• Nutrient Input

The precise concentration of nutrients being injected into the reservoir and their effects on population growth of SRB (and other bacteria) are difficult to assess. Much of the modelling has been derived from history-matching of field situations based on known or estimated information from the field or laboratory.

• Future Reservoir Management

Any differences in production/injection rates could have an impact on the overall H₂S profile (up or down). For example, prolonged production at higher watercuts than those assumed in the profiles could increase H₂S concentrations. Water shut-off could immediately reduce H₂S concentrations in a well by over 50% by reducing the partitioning effect at lower watercut and by shutting off sour water. Delay in water breakthrough times by reducing rates or increasing injector-producer distances could also be potential souring mitigation techniques.

References

1. D.E. Broussard, M.D. Lux, K.P. Havik; European Petroleum Conference, October 1982, London; EU 378, p885.
2. R. Cord Ruwisch, W. Kleinitz, F. Widdel; J. Petroleum Technology, Jan 1987, p97.
3. T.N. Nazina, E.P. Rozanova, S.I. Kuznetsov; J. Geomicrobiology, V(4), N(2), 1985, p103.
4. B.N. Herbert, P.D. Gilbert, H. Stockdale, R.J. Watkinson; SPE of AIME, Offshore Europe Conference, 1985, SPE 13978.
5. Asfandiyarov, F.A., Kildibekov, I.G., K.R. Nizamov; Jan 1984, Neft Khoz No.1, pp38 41.
6. A.M. Ametov, V.M. Entov, Izv Akad Nauk SSSR, Mekh Zhidkosti Nov Dec 1981, Gaza No.6, pp84 91.
7. G.L. Olson, W.P. Iverson, F.E. Brinckman; Amer Chem Div Preprints March 1983, V28, No.3, p769.
8. J.R. Postgate, Phil Trans Royal Soc London, SER B.V. 298, No.1093, p583, 82.09.13.
9. B.N. Herbert, Microbiol Group Symp (Birmingham, London, Proc p47, 76.10.06).
10. A.D. Forbes, Book IBN 0 85501 325 7 Heyden 7 Son Inc., Philadelphia; p169, 1980.
11. T.N. Nazina, E.P. Rozanova, Mikrobiologiya, Jan Feb 1980, V49, No.1, p123.
12. E.P. Rozanova, V.N. Bykov, A.L. Baldina, T.A. Kosogorova, Mikrobiologiya, March April 1983, V4, No.2, p347.
13. A.L. Baldina, V.N. Bydov, T.A. Kosogorova, V.A. Pokrovskii, E.P. Rozanova, Mikrobiologiya March April 1972, V41, No.2, p362.
14. G.S. Rogers, Ground Water, Jan Feb 1986, V24, No.1, p93, US Geological Survey.
15. K.C. Lunden, T.M. Stastny, NACE Corrosion 85 (Boston, 85.03.25 29) Preprint No.296, 7 PP, 1985.
16. D.W.S. Westlake, Int Doe Bartlesville Energy Technol Center Et Al Microbial Enhancement of Oil Recovery Conf (Afton, Okla, 82.05.16 21) Proc p102. Feb 1983.
17. M.V. Ivanov, S.S. Belyaev, Int DOE Bartlesville Energy Technol Center Et Al Microbial Enhancement of Oil Recovery Conf (Afton, Okla, 82.05.16 21) Proc p48, Feb 1983.
18. K.S. Laurinavichus, A. Ya Obraztsova, S.S. Belyaev, M.V. Ivnov, Amer Chem Soc Petrol Chem Div Preprints V28, No.3, p805, March 1983.
19. R.M. Siebert, 6th Annu Sept Gulf Coast Sect Res Conf (Austin, Texas, 85.12.01 04).
20. R KH Khazipoz, V.N. Kravchuk, T.V. Naumova, Neft Khoz, May 1982, No.5, pp61 63.
21. S.S. Belyayev, A.U. Lein, M.V. Ivanov, Geochem Int, 1981, V18, No.2, p59, 1981.
22. F.D. Cook, A. Jobson, D.W.S. Westlake, Chem Geol, Feb 1979, V24, Nos. 3 4, p355.
23. P. Jones; SPE Seminar on Produced Water Problems, Aberdeen, 14th May 1987.
24. A. Shere, C. Bond; SPE Seminar on Produced Water Problems, Aberdeen, 14th May 1987.
25. R.W. Mitchell; Journal of Petroleum Technology, June 1978, p877.
26. C. Oloman; J. Electrochem. Soc., Electrochemical Technology, 1070, V117, N(12), p1604.
27. M.S. Chao; J. Electrochem. Soc., 1986, V133, N(5), p954.
28. M. Pourbaix; Atlas of Electrochemical Equilibria in Aqueous Solutions, p535.
29. F. Cotton, G. Wilkinson; Advances in Inorganic Chemistry, Pull. Wiles, p535.
30. K.P. Whittingham, J.A. Hardy; UK Corrosion, Harrogate, 4th 6th Nov. 1985, p271.
31. E. Buck, S. Jones; Offshore Technology Conference, Houston, May 1986; OTC5269, p399.
32. H.W. Feely, J.L. Kulp; Bulletin of American Assoc. of Petroleum Geologists, 1957, 41, (8), p1802.
33. W.G. Toland, J. American Chemical Society, 1960, 82, p1911.
34. W.L. Orr; 7th International Meet on Organic Chemistry.
35. J.W. Moore, F.J. Millero, J.C. Cornwell and D. Rickard; Earth Sciences Review, No.1985.
36. D Thrasher: H2S Forecasting Model. BPX report POB/61/93
37. E Sunde, T Thorstenson, J E Vag and M S Espandal. SPE 25197
38. D Thrasher. 1993 Prudhoe Bay H2S Update. BPX report POB/60/93
39. D Thrasher. Scheihallion Reservoir Souring Potential BPX report POB/026/96