Assessment of the Biogenic Souring in Oil Reservoirs under Secondary and Tertiary Oil Recovery

Abstract

The formation of hydrogen sulfide (H₂S) in petroleum reservoirs by anaerobic microbial activity (through sulfate-reducing microorganisms, SRMs) is called biogenic souring of reservoirs and poses a risk in the petroleum industry as the compound is extremely toxic, flammable, and corrosive, causing devastating damage to reservoirs and associated surface facilities. In this paper, we present a workflow and the tools to assess biogenic souring from a pragmatic engineering perspective. The retention of H₂S in the reservoir due to the reactions with iron-bearing rock minerals (e.g., siderite) is shown in a theoretical approach here and supported with literature data. Cases are provided for two fields under secondary (waterflooding) and tertiary flooding with microbial enhanced oil recovery (MEOR). The use of the Monte Carlo method as a numerical modeling tool to incorporate uncertainties in the measured physical/chemical/biochemical data is demonstrated as well. A list of studies conducted with different chemicals alone or in combination with various biocides to mitigate biogenic souring provides an overview of potential inhibitors as well as possible applications. Furthermore, the results of static and dynamic inhibition tests using molybdate are presented in more detail due to its promising mitigation ability. Finally, a three-step workflow for the risk assessment of biogenic souring and its possible mitigation is presented and discussed.

1. Introduction

The release of hydrogen sulfide (H₂S) due to anthropogenic activities may be a result of industrial processes, especially wastewater or waste treatment, but also gas and oil production and oil refining. One source for H₂S generation in hydrocarbon reservoirs is the interaction of hydrocarbons with sulfur-containing minerals like anhydrite (CaSO₄) to form carbonates in addition to H₂S. These reactions take place at high temperatures (>140 °C) in a process called thermogenic H₂S generation [1]. Another known type of H₂S generation in oil reservoirs is due to anaerobic microbial activity (through sulfate-reducing microorganisms, SRMs) and is referred to as biogenic reservoir souring.

Independent from its source, production of H₂S is always an unwelcome issue in the oil and gas industry as the compound is extremely toxic, flammable, and corrosive. Damage to reservoir and associated surface facilities caused by biogenic souring remains one of the largest techno-economic risks in all water/chemical injection oil recovery operations, estimated to cause over USD 100 billion in annual economic damage [2,3,4]. A number of studies from a wide range of disciplines have been published to investigate various aspects of its generation, transport in porous media, rock and fluid interaction, mitigation, and production–corrosion characteristics. Bastin et al. were among the first to provide evidence that H₂S can be generated by SRMs in oilfields [5]. Another pioneer, ZoBell, investigated the effect of microbial souring in relationship with water injection to increase oil production [6]. Several books and publications describe the microbial aspects of souring [7,8,9,10,11], and well-established expertise is existing on the issue, including methods of comprehensive analyses of key parameters. The efforts on the modeling and mitigation in the reservoir as well as in production systems are also subject of various publications [3,12,13,14,15,16]. Based on the theoretical, laboratory, and field results, for example, the use of nitrate to mitigate reservoir souring has been proposed [8,17,18,19,20]. Although nitrate has been used in many field applications [21,22,23], discussions are ongoing on its effectiveness, and various alternative mitigation chemicals and methods are proposed [24,25,26,27,28].

The prediction of biogenic souring as well as evaluating its impact on production performance is of crucial importance for the oil exploration and production industry. Despite recent studies that have made important contributions to the understanding of biogenic souring of reservoirs, pragmatic reservoir engineering perspectives providing guidance on how to deal with this difficult and costly challenge in field applications are still scarce. Accordingly, this paper focuses on the following five main objectives:

• Provide a survey of recent studies on the understanding of biogenic reservoir souring, the controlling parameters, and its mitigation;
• Emphasize the importance of mineral scavenging for the retention of H₂S in the reservoir based on a theoretical approach as well as laboratory measurements;
• Underline laboratory and numerical modeling efforts, approaches, and success rates, exemplified via two application cases on a waterflooding and a microbial enhanced oil recovery (MEOR) technology project;
• Use of the Monte Carlo method as a pragmatic reservoir modeling approach with the associated uncertainties;
• Present and discuss a workflow that can be used in similar studies to assess the risks of biogenic souring and potential mitigation measures.

2. Biogenic Souring: Mechanisms, and Impacts on Reservoir–Oil Production

The dissimilatory sulfate reduction, i.e., the conversion of sulfate to H₂S by SRMs is the major source of H₂S in biogenic souring and arguably one of the most relevant challenges for the petroleum industry. Dissimilatory sulfate reduction is exclusive to anaerobic conditions as typically encountered in oil reservoirs due to highly reducing environments. However, it should be noted that dissimilatory sulfate reduction is just one aspect of the biogeochemical sulfur cycle and that sulfur-containing compounds can be abiotically and biotically transformed in several ways. For instance, sulfur intermediates can serve as both an electron acceptor and donor simultaneously producing both sulfate and sulfide (so-called sulfur disproportionation, see Figure 1). In addition, sulfur-containing compounds can also be oxidized or assimilated into microbial cells for biosynthesis of amino acids and proteins [29] (Figure 1). But as no H₂S is excreted into the extracellular environment in those cases, the reader is referred to a number of excellent papers and reviews where the sulfur cycle is discussed in more detail [29,30,31].

Like all other biological cells, SRMs require chemical compounds to grow on, so-called substrates. These include molecules acting as an electron donor and acceptor to gain energy from biochemical redox reactions (dissimilatory metabolism). In addition, a carbon source which can be incorporated into the cell mass to allow cell growth is required (assimilatory metabolism). In the case of organic compounds, volatile fatty acids (VFA), the electron donor and carbon source can be the same molecule. In other cases, using inorganic substrates such as hydrogen (H₂) may only be utilized for harvesting energy via electron transfer to an electron acceptor. In that case, the carbon required for cell growth can be taken from organic molecules (heterotrophy) or inorganic carbon sources such as CO₂ (autotrophy). SRMs can use sulfate (SO₄²⁻) as well as less-oxidized sulfur forms like thiosulfate (S₂O₃²⁻) or sulfite (SO₃²⁻) as electron acceptors (Figure 1). For oil fields, free sulfate concentrations in injection waters with waterflooding vary from not present at all up to 30 mM (milliMol) when seawater is injected [32]. Another significant example of the injection of organic material into a reservoir is the application of specific techniques for enhanced oil recovery, namely the injection of nutrients in MEOR or the injection of (bio)polymers and (bio)surfactants. Hydrolysis of polymeric substances into monosaccharides and amino acids and subsequent in situ fermentation results in similar metabolites as described above.

SRMs are common in many environments and can be active as long as the temperature is within the biologically relevant range. Even if a reservoir has no documented history of either elevated sulfate levels or souring, souring can occur once sulfate is introduced, i.e., with waterflooding. Acetate (CH₃COO⁻) is a common substrate for SRMs as VFA and often presents in formation waters and is a potential metabolite produced during MEOR or in situ hydrocarbon degradation. H₂S generation in the presence of SO₄²⁻ and acetate can be described as follows:

$${{\mathrm{C}\mathrm{H}}_3\mathrm{C}\mathrm{O}\mathrm{O}}^{-}+{\mathrm{S}\mathrm{O}}_4^{2-}+{3\mathrm{H}}^{+}\to 2\mathrm{C}{\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{S}+2{\mathrm{H}}_2\mathrm{O}$$

(1)

The H₂S generated due to this microbial interplay should be assessed for the design of secondary and tertiary oil recovery applications such as MEOR or biopolymer flooding. In this context, it is not only important to consider the microbial cycling of sulfur compounds, but also abiotic reactions: this may include chemical sulfur and sulfide oxidation due to oxidants such as nitrite, but also burial as metal sulfides such as FeS (Figure 1).

In the reservoir, biogenic souring occurs only if specific initial and/or boundary conditions are met. Hence, the formation and transport of H₂S in the reservoir is controlled by the microbiological, geochemical/physicochemical, petrophysical, and thermodynamical parameters. The H₂S generated is transported in the different phases, namely oil, water, and gas, throughout the reservoir and may be retained in fluid phases and/or may react to solid metal sulfides in the reservoir. Once reaching a production well, the flow conditions and resulting thermodynamics in the wellbore cause the H₂S to arrive topside to partition mostly into the gas phase, with a potentially significant impact on the safety of the facilities.

3. Controlling Parameters of the Biogenic Souring in Oil Reservoirs

The assessment of biogenic reservoir souring needs a thorough analysis and evaluation of the parameters causing and controlling biogenic souring. To assess the magnitude of H₂S and its mitigation, the following parameters should be defined and evaluated.

Interventional studies involving animals or humans, and other studies that require ethical approval, must list the authority that provided approval and the corresponding ethical approval code.

3.1. Microbiology: SRM in the Reservoir and/or Injected Water

The subsurface harbors a vast and native microbial biosphere and due to the lack of oxygen in the deep subsurface anaerobic processes prevail: preferred metabolic pathways include fermentation, anaerobic hydrocarbon degradation, acetogenesis, methanogenesis, and other anaerobic respiration processes with alternative terminal electron acceptors including nitrate, iron (and other oxidized metals), and oxidized sulfur species (Figure 1). Since oxygen and nitrate are usually absent (unless anthropogenically injected) and metal oxides have usually been reduced over time, the most prevalent processes are fermentation, methanogenesis, and sulfate reduction, if there is a source of sulfate.

A large group of known SRMs belong to the bacterial classes and domains of Deltaproteobacteria and Firmicutes, respectively. Sulfate-reducing bacteria (SRB) were the first microorganisms identified in oil fields [5], and several lower temperature oil reservoirs contain SRB. Some sulfate reducers are phylogenetically affiliated with the archaea and are designated as sulfate-reducing archaea, or SRA. In this paper, the term SRB is preferred as they are considered to be the main cause of H₂S generation in the reservoir. Two well-known genera of SRB include the non-sporulating Desulfovibrio and the spore-forming Desulfotomaculum. Other genera of SRB, including Desulfomonas, Desulfobacter, Desulfobacterium, Desulfobulbus, Desulfococcus, Desulfonema, Desulfosarcina, Desulfomicrobium, and Thermodesulfobacterium have also been identified and classified as H₂S generating microbes. Besides, some anaerobic extremophilic archaea as well as mesophilic bacteria such as Bacillus, Pseudomonas, Proteus Clostridium, and Thermoproteus are reducing elemental sulfur to sulfide [7]. In general, SRB are a mixed group of morphologically and nutritionally diverse, strictly anaerobic bacteria which utilize sulfate as a terminal electron acceptor, reducing it to H₂S. SRB are present in almost all types of habitats, and although being obligate anaerobes, they can also be found dormant in a wide variety of oxygenated environments and are now well recognized for their involvement in microbial induced corrosion (MIC) [7]. Larsen et al. provide a list of the SRB detected in various fields in the North Sea [33], whereas Gieg et al. [34] and Immanuel et al. [26] provide a list of various SRMs reported in various produced waters with the corresponding salinity and temperature conditions. Nowadays various methods are used for the identification and enumeration of bacteria. These include 16S rDNA sequencing for identification and fluorescence microscopy, quantitative fluorescent in situ hybridization (qFISH), or quantitative polymerase chain reaction (qPCR) and reverse transcription qPCR (RT-qPCR) for enumeration of the total (i.e., dead and alive) as well as biologically active cells [7].

The kinetics of H₂S generation via microbial sulfate reduction has been investigated by many authors. Tang et al. provide insights on sulfate reduction biokinetics and the governing parameters using various dynamic setups such as bioreactors and sand packs [29]. These studies suggest that the volumetric sulfate reduction rate in the presence and use of various energy and carbon sources by SRB varies depending on operating conditions such as pH, sulfate concentration, and, finally, temperature. In a reservoir, the biokinetics of microbial sulfate reduction mainly depend on the discrepancy between the prevailing reservoir temperature and the optimum growth temperature of the corresponding SRB, but in terms of reservoir engineering, the reaction times are generally quite fast, i.e., being in the range of several days.

The growth rate of SRB in the presence of substrates can be predicted by the well-known half-empirical Monod equation [35]. With variations of the basic equation, Monod kinetics were assessed to account for different products, microbial species, or substrate inhibition [36,37,38]. The basic form of Monod equation with the growth rate r_g is provided as follows:

$$r_g=r_{max}\frac{S}{K_S+S}$$

(2)

in which S is substrate concentration, r_max is the maximum growth rate, and K_S is the half rate constant at which the rate is r_max/2.

3.2. Pressure and Temperature

As discussed above, temperature plays a crucial role on the activity of SRB. Most SRB are thought to be active from low up to moderate reservoir temperatures (10–70 °C), whilst thermophilic SRB (thermodesulfobacterium) are reported to be active up to 85 °C [29,34]. According to a limited number of studies, the pressure ranges in oil reservoirs do not seem to have an inhibiting effect on the activity of SRB and H₂S production [39]. Figure 2 depicts the results of a study on water samples from several North Sea fields showing the activity regions of SRB as a function of pressure and temperature. As can be concluded, the activity of SRB is limited by temperature rather than pressure with tolerable pressure limits reaching 500 bar, which is higher than the prevailing pressure in most oil reservoirs. SRB can also be present in injected seawater and are capable of growing in the reservoir in the so-called thermal viability shell (TVS), i.e., the lower-temperature zone which results from the injection of colder water below the reservoir temperature ([39]; see next chapters).

3.3. Chemical–Physicochemical Conditions

3.3.1. Source of Carbon, Energy, and Nutrients

Within the microbial catabolism (i.e., harvesting energy) of SRB, various chemical compounds may serve as electron donor as described in Equation (1) and Section 2. These include hydrogen and various organic carbon compounds, for example VFAs and their salts, such as formate, acetate, propionate, butyrate, and lactate. Amino acids, furfural, sulfur compounds, (polar) aromatic hydrocarbons, and saturated hydrocarbons are among the other compounds which may be utilized by SRB [29]. However, VFAs are among the organic carbon compounds easiest to metabolize and typically present in any formation water to a certain degree. In the case of produced water re-injection, VFAs are usually also present in the injection medium. In the case of injection of biosurfactants and biopolymers for oil recovery or in the application of MEOR, VFAs are generated during various biotic and abiotic degradation processes. According to Vale et al. [40], the mixture of biosurfactant and oil increased or decreased SRB activity depending on the surfactant concentration (high/low, respectively). Given that such organic carbon compounds are also incorporated into the cells as biomass, the dissolved organic carbon (DOC) content in the water is another useful parameter to assess the concentration of potentially bioavailable carbon [41].

3.3.2. Salinity

High total dissolved solids (TDS) brines will exert a high osmotic pressure on bacterial cells. Laboratory data and unpublished field data indicate that the levels of H₂S that can be generated in high-salinity brines (TDS of higher than 150 g/L) are lower compared to equivalent lower-salinity brines. On the other hand, presence and activity of SRB has been observed in higher-salinity reservoirs of up to 180 g/L [27]. With seawater as the injection medium, a lower salinity environment is established which provides favorable conditions for the growth of SRB.

3.3.3. Partitioning

In common terminology and when discussing H₂S, often only the gas phase is considered. However, H₂S may also partition to quite some degree into the liquid phases, especially under elevated pressure. In water, it may then further dissociate into aqueous sulfide ions (HS⁻ and S²⁻), depending on the prevailing pH. In oil reservoir engineering, it is much more valuable to assess the partitioning of all sulfidic species in gas, oil, and water. In the reservoir, where oil and water co-exist, H₂S is a mobile chemical compound, transitioning between the oil and the water phase. If the oil phase is in a saturated phase, i.e., if the pressure is lower than the saturation pressure (bubble point), then a gas phase is formed in the reservoir in addition to liquid phases. A comparable situation occurs during the downhole flow to the wellhead into the separators; due to the pressure drop, H₂S degasses along with other light components from the liquid phases. The fraction in the gas phase is defined by the H₂S partial pressure and is a function of a number of thermodynamic parameters. Prediction of three-phase equilibrium, or partitioning between the phases present, is important for both subsurface reservoir conditions and wellhead/separator conditions.

The solubility of H₂S in liquids under the normally prevailing temperature and pressure conditions in reservoirs is increasing with increasing pressure and decreasing with increasing temperature. This can be predicted by using the partial-pressure concept via Henry’s Law. At low pressure, Henry’s constant depends on temperature and, to a lesser degree, on pressure. At high pressures, it is necessary to consider the pressure dependency of Henry’s constant as proposed by [42]. For water, salinity needs to be considered as well. The solubility of H₂S in water decreases as the salinity of water increases [43]. Figure 3 shows the solubility of H₂S in pure water and in brine with 6 molar (M) NaCl salinity as a function of pressure at 80 °C. The pH value also plays a role which is discussed in the following chapter.

The diversity of organic compounds in various crude oils may markedly affect the solubility of H₂S in oil and its distribution between crude oil and water. Heavy, more aromatic oils provide a higher H₂S solubility than lighter, paraffinic oils [44].

Figure 4 shows the oil–water partitioning coefficient (K_ow) of H₂S in hexadecane and in a medium-heavy oil (29°API) calculated with PVTSim [45] as a function of pressure. Calculations are performed for 1% and 10% NaCl brines. As expected, the H₂S preferentially partitions in crude oil, which is heavier and more complex than n-hexadecane. On the other hand, it can also be deduced from the figure that, for the high-salinity brine case, the K_ow is higher for the same oil type, as the solubility of H₂S in the brine decreases.

Figure 3. Solubility of H₂S in pure and saline water (6 M NaCl). Data from [46].

Figure 3. Solubility of H₂S in pure and saline water (6 M NaCl). Data from [46].

3.3.4. pH

Most SRB have an optimum growth at around pH 7 but are also active in the pH range of 5–8 commonly encountered in oil reservoirs [18]. Hence, pH as a parameter in souring relates to the partitioning of H₂S in existing phases rather than the inhibition of SRB activity at very-low and high pH values. As the pH of water increases above 5, the concentration of the undissociated H₂S decreases in a predictable manner by dissociating to the bisulfide ion (HS⁻). This bisulfide ion and the further dissociated sulfide ion (S²⁻), which exists under alkaline conditions, are unable to partition from the water to either the liquid hydrocarbon or the gas phases. Therefore, upon contact with alkaline water, any H₂S in the hydrocarbon phase will not only partition into the water, but also dissociate until the partitioning and dissociation reactions reach equilibrium. This equilibrium state may or may not support the existence of significant H₂S in the gas phase. Figure 5 shows the K_ow for two different pH values (pH = 8 and pH = 6) as a function of pressure where the effect of pH on the partitioning of H₂S into oil and water phases can be clearly observed.

3.3.5. Rock Mineralogy/Formation Water Composition

Due to precipitation reactions, H₂S retention in the subsurface is assumed to be significant if the reservoir fluids contain dissolved ferrous iron Fe(II) or if the rock includes iron-rich minerals. This is referred to as mineral scavenging and, for exemplary purposes, we focus here on siderite (FeCO₃) only. If present in the reservoir rock’s flow paths, siderite may react with H₂S to form pyrrhotite (FeS) or other (partially amorphous) sulfidic minerals:

$$\mathrm{F}\mathrm{e}\mathrm{C}{\mathrm{O}}_3+{\mathrm{H}}_2\mathrm{S}\to {\mathrm{H}}_2\mathrm{O}+\mathrm{C}{\mathrm{O}}_2+\mathrm{F}\mathrm{e}\mathrm{S}\downarrow$$

(3)

A theoretical approach based on the precipitation reaction detailed above can provide an idea on the retention potential of H₂S: Consider 1 m³ of oil reservoir with a porosity of 25% that is filled with formation water containing 0.5 mM SO₄²⁻. According to Equation (1), the total consumption of the present sulfate would yield a stoichiometric concentration of 0.5 mM H₂S (or 0.125 mol per 1 m³ reservoir volume based on the aqueous phase in the porous media). If the formation consists of 1% wt. siderite, this corresponds to more than 20 kg (172.6 mol) siderite in a 1 m³ reservoir, given a rock density of 2650 kg/m³. With a conservative assumption that only 1% (1.72 mol) of this siderite amount is in contact with the water phase and hence available to react with H₂S according to Equation (3), there is a significant excess of reactive siderite compared to the available H₂S by a factor of approx. 15-fold. This example using conservative input data, however, shows the relevance of mineral scavenging in oil reservoirs if the formation mineralogy contains iron-bearing minerals in contact with H₂S.

In addition to the above detailed retention mechanism of the reservoir rock, the divalent cation content of the formation water may well cause the sulfate to precipitate as Ca²⁺, Ba²⁺, or Sr²⁺ sulfates (e.g., CaSO₄, BaSO₄, and SrSO₄), depending on oversaturation and temperature; these precipitation reactions usually have faster kinetics than the microbial consumption of the sulfate:

$${\mathrm{B}\mathrm{a}}^{2+}\left(\mathrm{o}\mathrm{r} {\mathrm{S}\mathrm{r}}^{2+} \mathrm{o}\mathrm{r} {\mathrm{C}\mathrm{a}}^{2+}\right)+{\mathrm{S}\mathrm{O}}_4^{2-}\to {\mathrm{B}\mathrm{a}\mathrm{S}\mathrm{O}}_4\left(\mathrm{o}\mathrm{r} {\mathrm{S}\mathrm{r}\mathrm{S}\mathrm{O}}_4 \mathrm{o}\mathrm{r} {\mathrm{C}\mathrm{a}\mathrm{S}\mathrm{O}}_4\right)\downarrow$$

(4)

Beyond the effect on sulfate, sulfides may be precipitated also by reaction with dissolved ferrous iron (II) in the formation water in a manner very much like the siderite retention mechanism outlined above:

$${\mathrm{F}\mathrm{e}}^{2+}+{\mathrm{S}}^{2-}\to \mathrm{F}\mathrm{e}\mathrm{S}\downarrow$$

(5)

Low concentrations of ferrous iron cations are often found in reservoir waters, especially if iron-bearing minerals are present within the reservoir.

The mitigation capability of iron-bearing minerals has been investigated and discussed in the literature based on experimental evidence. In a recent study, Graham et al. presented the results of an experimental study consisting of batch and dynamic experiments (sand packs and core floods) and they concluded that the magnitude of the two proposed mechanisms for H₂S reservoir scavenging (dissolution/precipitation) depends on the stability of the Fe-bearing minerals and provided a scavenging capacity of 0.5 to 13 mg H₂S/g rock over a range of initial pH values and temperatures [47]. Kögler et al. reported the H₂S retention capacity of crushed formation material containing siderite (10% wt.) to be approximately 0.14 mg/g [27]. The data available in the literature are listed in

Table 1. Data on mineral scavenging of H₂S from the available references (static: batch experiments carried out with flasks filled with rocks and H₂S containing fluids; dynamic: experiments displacing H₂S containing fluids through cores of reservoir rocks).

Type of Evaluation	Rock Type	H2S Retention [mg/g]	Reference
Dynamic	Cores from various reservoirs, k = 5200 mD	0.014	[48]
Dynamic	Cores from various reservoirs, k = 0.4 mD	0.0196	[48]
Dynamic	Cores from various reservoirs, k = 8200 mD	0.55	[48]
Dynamic	Cores from various reservoirs, k = 740 mD	1.95	[48]
Static	Crushed cores with siderite content (<814 μm)	0.24	[49]
Dynamic	Cores with siderite content; from the available data	0.003	[49]
Static	Crushed rock (100–315 mm) with 20% wt. siderite	0.5–3.7	[47]
Static	Crushed rock (<100 mm) with 20% wt. siderite	2.08–9.70	[47]
Dynamic	Sand packs filled with sand with 4% wt. siderite	1.03–1.74	[47]
Dynamic	Sand packs filled with formation rock with 10% wt. siderite	0.14	[27]
Theoretical	Based on stoichiometry for 1% wt. siderite content	0.34	This study

. Although the experimental setups and procedures are subject to uncertainty, the values emphasize the validity of the above theoretical approach and demonstrate the importance of mineral scavenging in mitigating biogenic souring.

Temperature has the largest effect on mineral scavenging. Higher temperatures yield larger capacities, which can be attributed to accelerated kinetics. The composition of the reservoir rocks can be estimated from petrographic evaluations of recovered core materials, but the degree to which they contribute to the H₂S scavenging in a reservoir is difficult to predict. The mineralogy is generally not homogeneously distributed, and petrophysical parameters like permeability also play a role. In addition to iron carbonates, other minerals can contribute to the H₂S retention, both positively and negatively. Especially, the presence of clays is mentioned to adsorb H₂S, thus increasing the mineral-scavenging capacity. High porosity and permeability may translate to a lower specific surface area of the rock and hence a lower contact between scavenging minerals and H₂S-bearing phases. Wettability can affect the access of H₂S present in both phases to the minerals as well. In addition, hydrodynamic heterogeneity is also a factor that can affect mineral scavenging of H₂S in the reservoir; reservoirs with high heterogeneity allow flow paths with higher permeability and the generated H₂S can flow in the corresponding phases to the production wells without contacting the carbonates.

3.4. Petrophysical Conditions

3.4.1. Permeability

It is known that SRB can differ in shape and size with an average length of one µm, like most other bacteria. Experimental studies reveal a linear relationship between straining rate of spherical bacteria and the ratio of bacterial diameter to average pore size. Gruesbeck and Collins suggested that the ratio between bacteria diameter to the pore size must be in the range of 0.17 to 0.25 for straining-induced plugging to occur [50]. This corresponds to the claim made by Davis and Updegraff that the pore throat sizes of the rock should be at least twice the cell sizes to allow passage of microbes through the rock without substantial plugging [51]. Generally, rod-shaped bacteria tend to adopt a preferred orientation in direction of the flow along their major axis [52]. For free movement and transport of the SRB throughout the reservoir, the permeability or, more precisely, the average pore size of the reservoir studied should be large enough (k_avr > 100 mD, pore diameter > 10 µm). If SRB are introduced into the reservoir via injection water and if the permeability is too low to transport the SRB deeper into the reservoir, H₂S can only be generated in a limited volume of the reservoir. Potentially, biofilms may form in the near-wellbore region which may accelerate the souring. On the other hand, lower permeability means a higher contact area, thus increasing H₂S retention via mineral scavenging in the reservoir, if the rock mineralogy allows.

3.4.2. Heterogeneity

Reservoir heterogeneity can increase the effects of biogenic souring because of at least two aspects: Via higher permeability channels, the soured water and/or oil will reach the production wells faster, therefore allowing less time for retention and scavenging reactions which would remove H₂S from the fluids; the resulting rapid breakthrough of H₂S to producers is another drawback. The other negative aspect is the lower contact area between soured fluids and reservoir rock, which again implies a lower amplitude of retention and scavenging of sulfate and H₂S in the reservoir. The heterogeneity can be quantified with the help of various methods like the Dykstra–Parsons (DP) coefficient [53].

An overview of the controlling parameters of biogenic souring and their influences in terms of reservoir engineering is provided in

Table 2. Overview of controlling parameters of biogenic souring and their influences.

Parameter	Range, Influence
Solubility H2S	Higher in oil than in water; higher in complex oil compositions containing asphaltenes and aromatics; lower in brines with increasing salinities
Partitioning, Kow, Kog, Kwg	Depending on the H2S solubility in oil and water phases, salinity, temperature, pressure, pH
pH	Over reservoir typical ranges, not relevant for the activation of SRB in oil reservoirs. The higher the pH, the higher the solubility in water, thus the lower the Kow
Water chemistry	Divalent cations in the formation water may precipitate with both sulfate (Ba2+, Sr2+, and Ca2+) and sulfide (Fe2+, Zn2+), acting as scavengers
Rock mineralogy	The presence of iron-bearing minerals is crucial for the in situ retention of H2S via precipitation
Temperature	Most SRB are not active at temperatures higher than 80 °C
Pressure	Not relevant effect on SRB activity for most common pressures in oilfields, therefore no relevant effect on H2S generation
Sulfate	One of the main components of biogenic souring, concentration has a strong impact on the possible H2S concentration generated
Salinity	Certain SRB species live in higher-salinity brines up to ca. 180 g/L; H2S solubility in water decreases as the salinity increases (salting-out effect)
Presence of SRB	If no indigenous SRB exist in the reservoir no biogenic souring is expected; however, SRB can be introduced with injection water
Permeability	Low permeability, small pore radius limits the transport of microbes therefore their planktonic movement; no relevant effect on H2S transport once generated
Heterogeneity	Important for the transport of SRB and H2S; if high, less retention in the reservoir and rapid breakthrough of H2S to the producers can be expected
Carbon sources	Completes the souring reaction; VFA can be an important source; oil components may also be a nutrient source for SRB

.

4. Laboratory and Numerical Modeling Efforts—Types and Applications

Although biogenic souring has been observed in numerous oil fields where seawater is injected, predictive capabilities are still limited. In a general sense, if souring has occurred in a certain part of the field, the generated information can, if well documented, also be parametrized and utilized for other parts of the reservoir. However, field simulations of biogenic souring are not straight forward as spatial and timely uncertainties in the parameters play a role on the development of the conceptual models, and thus on the numerical implementations. In the related literature, three conceptual models are frequently discussed and applied.

4.1. Conceptual Reservoir Models of Biogenic Souring

4.1.1. Mixing Type Souring Pattern

In this concept, it is assumed that microorganisms thrive in the mixing zone of nutrient-rich formation water and sulfate-rich injection brine [54]. Given this model, produced H₂S should peak with the injection water breakthrough, before it eventually flattens out, due to the mixing front moving beyond the ideal temperature window and due to the loss of ideal mixing ratios (i.e., depletion of available sulfate due to consumption and dilution of injection brine).

4.1.2. Biofilm Formation

A biofilm buildup in the proximity of the water injectors is assumed to take place. Hence, the model attributes the generation of H₂S during seawater injection to the sessile (e.g., attached cells) SRB in the biofilm, and all reactions responsible for reservoir souring take place in the vicinity of the injection well. Here, the temperature reduction due to continuous injection enables optimized conditions for microbial activity, sustained by crude oil itself as a carbon source (biodegradation of oil components or water-soluble BTEX) [48]. With this model, the H₂S production begins with injection water breakthrough as well, but slowly increases over time with a growing population of SRB in the biofilm. The nutrient concentration of the injection water is one of the major controlling factors.

4.1.3. Thermal Viability Shell (TVS)

This concept is based on the correlation of experimental data that relates the consumption of sulfate to the temperature and pressure in the presence of SRB [39]. In its original form, the model assumes a constant concentration of sulfate in seawater and does not include the chemical constraints (limited nutrient availability) effects on the generation of H₂S. Thus, with the profile of temperature and pressure in a reservoir, the generation rate of H₂S is obtained with a simple stoichiometry function of the reduced sulfate. In the original model, the retention and partitioning of H₂S are ignored. The model attributes the delay in the produced H₂S to the lag of the temperature front with respect to the injection front, i.e., thermal vs. hydraulic radius, by calculating them via analytical methods.

4.2. Model Implementations

Maxwell and Spark described the possibility for hybrid models for field simulation, and in a recent paper on the Gullfaks (Norway) field, this could be shown for the H₂S formation in an oil field flooded with seawater [55,56]. Indeed, the numerical reservoir models could, and in most of the cases should, contain all the relevant aspects of the three concepts, in addition to the general rules of transport in porous media including effect of controlling parameters presented in Section 3. Figure 6 depicts the formation and transport of biogenic H₂S in the reservoir schematically. Assuming that SRB are present in the reservoir in sessile or planktonic form, injection of sulfate would cause the formation of H₂S due to the reaction provided in Equation (1). If the temperature in the reservoir is too high for the activity of the SRB (>80 °C), the injection of cooler water may also transport SRB into the reservoir and the decreasing reservoir temperature in the reservoir can form favorable conditions for SRB. The SRB may adhere/adsorb onto the rock surface forming biofilm layers. The biofilm formation kinetics are known to occur within a relatively short time and the biofilm density is expected to decrease with increasing distance from the injector. As these biofilms potentially contain high numbers of SRB, these are the main sources of H₂S generation, especially around the injectors as theorized by the biofilm concept.

The above concepts for injection, transport, and retention in porous media and production are numerically modeled by various researchers and engineers, at laboratory and/or field scale. A summary of these attempts is provided in

Table 3. List of the relevant studies on the numerical modeling of biogenic souring.

Year	Model Title	Overview	Ref.
2004	Modelling using commericial reservoir simulators, CMG-STARS	Kinetic models for microbial growth and metabolite production to simulate up-flow packed bioreactor tests	[65]
2009	A souring model in a 3D finite difference compositional non-isothermal reservoir simulator	The partition coefficient depending mainly on temperature and only weakly on pressure, to determine the H2S concentration in the two phases	[72]
2009	Uncertainty analysis using CMG-CMOST	H2S concentrations presented as a probabilistic range	[68]
2010	Modelling biosouring mitigation using commercial reservoir simulator CMG-STARS	Designing and optimizing nitrate injection programs as part of an overall management strategy of microbial souring in produced water recycle injection operations	[66]
2010	3D modelling applied in numerical simulator MUFTE; not compatible with reservoir simulation	A model for the transport of H2S in an oil- and water-saturated porous medium with different retention mechanisms	[73]
2012	Using UTCHEM, a reactive flow simulator	The simulator is used to predict the onset of reservoir souring and the effectiveness of nitrate injection	[72,74]
2013	SourSim®RL, a reservoir souring code to be coupled with commercial reservoir simulators	Considers temperature distribution, SRB growth, H2S partitioning and transport mitigation with nitrate module and oil biodegradation also implemented;	[75,76]
2014	Souring simulator coupled with geochemical code PHREEQC is used	The coupling scheme to the geochemical solver is explicit, the partitioning of components between the phases including evaporation of H2O is handled fully implicitly	[62]
2014	A numerical simulator including the growth rate equation for the SRB by modifying a simulator of MEOR	The simulation suggested that SRB generates H2S only around the injection well because of a temperature drop there; souring can be prevented more surely by heating up to 50 °C, or reducing ethanol in the injection water	[77]
2016	TOUGHREACT for isotopic and biogeochemical systems; kinetics with Monod expression	With data from column experiments as constraints, a first reactive transport model of a new candidate inhibitor, perchlorate; compared with nitrate	[78]
2017	To simulate multiphase flow and bio-chemical reactions combining TMVOC and TOUGHREACT	An approach to utilize genomic information to constrain the biological parameters needed for modeling souring using data derived from oil reservoir studies	[59]
2017	Attempt to model reservoir souring by using the multi purpose 3D reservoir simulator UTCHEM	The work differs from earlier works by explicitly determining parametric values required for Monod kinetic model as function of salinity, temperature and pH	[58]
2019	Field simulation based on the use of the reservoir model in combination with SourSim®RL based on measured data	The initial phase consisted of the history match to define the most likely souring mechanism in the field. The forecast considered various scenarios with a range of sensitivities on carbon nutrient and sulphate levels	[60]
2020	A souring multi-domain workflow considering dynamic reservoir characteristics	An integrated approach from reservoir to facilities, and presents a case study by using commercial reservoir and souring simulators	[79]
2020	A coupled reactive transport model using PHREEQC to study the effect of mineralogical composition	The results show that chalk containing anhydrite produces more H2S in comparison to limestone; rock containing iron minerals can inhibit the H2S generation in porous media	[63]
2021	Use of fuzzy logic to combine basic principles with sparse data for practical uses	A fuzzy logic model built around the reaction kinetics and then conditioned with field data. It is reported that the model matches the published field data fairly well	[64]
2022	Use of commercial simulator CMG-STARS in combination with uncertainty analysis (CMOST)	Experimental data were used for the calibration of the model; the model was applied to predict potential souring using Monte-Carlo analysis	[67], this study
2023	A Thermo-Hydro-bio-Chemical (THbC) model coupled with industry-standard simulations	The model has been tested and successfully tunned using the history of the field data for predicting microbial sulfate reduction in an oil field in the Danish North Sea	[80]

. Numerical simulators UTCHEM and TOUGH2 were adapted successfully for modeling the various aspects of H₂S formation and transport at the laboratory scale [57,58,59]. Field modeling efforts are provided by various authors starting with [48]. More recent numerical simulators such as REVEAL by PETEX and the “Oil Plus 3D” can better represent H₂S transport within the reservoir, but uncertainties about the controlling factors and limits for SRB activity have to be considered [60,61]. At the moment, “SourSim” can be combined with any commercial reservoir simulator. Several sophisticated souring models use geochemical algorithms to predict the biogenic souring and its mitigation using various chemicals [62,63,64]. State-of-the-art commercial reservoir simulators are (and should be) able to model all reservoir features of the biogenic souring [65,66,67]. From a pragmatic point of view, the main issue is to calibrate and validate the models. A validated reservoir model is mandatory as an initial step of a reliable souring modeling work. For the second step, the souring prediction, the model should be adapted to reflect reservoir souring processes, which in turn requires chemical and physical input from laboratory analyses of the field samples and measurements. As a rule of thumb, the higher the reliability and accuracy of the input data, the higher the validity of the reservoir souring model. Therefore, physical/chemical modeling and analysis at laboratory and field scales are essential to support the numerical studies. Recent studies provide good examples of such calibration and validation efforts to predict and mitigate reservoir souring based on produced water sulfate concentration monitoring and lab-scale dynamic testing [68,69,70,71]. A compilation of the laboratory experiments and measurements that can support numerical souring studies are provided in

Table 4. List of the experiments and analyses that can support reservoir modeling studies.

Type	Why, How	Methods, Analytics
Microbiological analyses	Analyses for: Microbial activity and composition; presence of SRMs, incubation;to growth; microbial community composition	Enumeration by MPN, ATP, OD600, DAPI, PCR, qFISH, DNA extraction and qPCR, batch incubations amplification of 16S rRNA	[7,9,33,55,81,82]
Bioreactors Sandpacks, Corefloods Micromodels	Souring and its mitigation under dynamic conditions, biofilm formation; transport of H2S under simulated reservoir conditions	Sandpack experiments, columns filled with formation materials; flow in micromodels, corefloods with analogue outcrops and reservoir cores	[17,27,29,83,84,85]
Chemical analyses	To determine the chemical composition of injected and produced water and oil; sulfide, sulfate, nitrate, VFA, cations, etc. Gas compositions; H2S, other gases	Spectrophotometric determination HPLC, GC, isotopic analysis	[9,30,86]
Petrophysical measurements	The permeability and pore radius/ throat as well as the porosity for the transport; model; mineralogy of the reservoir rock for reactivity; heterogeneity for transport in the reservoir; oil-water saturations for H2S partitioning	Routine and special core analysis, X-Ray measurements, well tests for heterogeneity, DP coeff. solubility of H2S in water and oil	[87,88,89]
Field-well measurements	Water cut for H2S partitioning, gas compositions for detecting H2S; corrosion monitoring; geology and production history for reservoir modelling production history for reservoir modelling	Routine field measurements and analyses, Draeger tubes corrosion tests; well tests	[12,56,90,91]

.

5. Case Analyses

Two cases of using commercial reservoir simulators for assessing and predicting the biogenic souring based on measured data and observations are presented in the following sections.

5.1. Case 1: A Field in North Sea under Seawater Flood

In this example, seawater injection into a green North Sea oilfield is considered. Due to absence of historic data from the field for the provided case, souring input parameters were taken from the literature and adjusted using in-house expertise as well as field analogues in the area [56,68]. Due to topside limitation, sulfate removal from seawater was not considered; the sulfate concentration in the injection water was measured to be about 2780 mg/L, whereas sulfate present in the formation water was in the range of 4–10 mg/L. The formation water shows a salinity similar to seawater, but with a significant concentration of acetate (VFA). Hence, sufficient carbon sources are available for SRB. With regards to phosphorous and nitrogen, both were assumed to be present in concentrations not limiting the growth of SRB. The reservoir temperature is just below 100 °C and thus outside the microbially tolerable zone, but would be affected by water injection, leading to lower temperature around the injector wells. The spacing between injectors and producers is between 1000 and 2500 m; the average permeability is around 200 mD. An assumed 100% conversion of all VFA would lead to about 240 ppm_w H₂S, which was used as a first conservative estimate in the study. Laboratory evaluations utilizing relevant seawater samples from the area indicated that the generation of H₂S is a realistic scenario, but the magnitude of the conversion could not be determined reliably.

To obtain a more realistic estimate for the resulting H₂S concentrations in the reservoir, a simulation study with CMG STARS (Release 2020) [92] was set up using the 3D reservoir model previously developed (Figure 7) [93]. Seawater is injected in the wells WI1, WI2, and WI3, and OP1, OP2, and OP3 are production wells. OP1 and OP3 are completed in the oil leg, while OP2 is completed in the water leg. The temperature variations due to water injection were accounted for by using the thermal option. No iron-bearing minerals were identified in subsurface samples, and, therefore, H₂S retention was not considered to ensure conservative souring estimates. The model that includes chemical reactions for the souring process was implemented to predict the reservoir, production, and injection performance over the field’s lifetime. This approach not only allows for the assessment of H₂S production over the life of the field, but also serves as an input to studies for topside mitigation needs and inorganic scale assessments. Especially on the former, high uncertainties regarding H₂S production rates may lead to design changes for the H₂S removal facilities.

The three production wells show distinct features in terms of H₂S production. As OP1 and OP3 are supported by injection into the oil leg, the effective mixing zone between injection and formation water is limited for these wells. This leads to short peaks of H₂S production in the simulation model, with an abrupt increase and a declining trend thereafter, when the mixing zone moves out of the favorable temperature window. In the case of the OP2 well, however, with the corresponding injection well being completed into the water zone, large efficient mixing between the nutrient rich formation and sulfate/SRB-rich injection water would lead to a steadily increasing H₂S production over time.

For the provided case, however, the history matching of production data in the future may help to reduce the impact of the input uncertainty on the model. As can be seen in Figure 8, the breakthrough of injection water (and production of H₂S) for OP1 is predicted to happen significantly earlier than for OP2 and OP3. This feature may allow for history matching with a subsequent tuning and re-running of the souring model to predict H₂S production at OP2 and OP3 with reduced uncertainties.

As a result of the simulation study, a basic mitigation strategy to reduce the H₂S production was defined to allow for H₂S scavenging at OP1. It is important to highlight that the mitigation strategy is only a part of the total approach for the asset. Intense treatment of the injection water aims to reduce the overall occurrence of souring as well.

5.2. Case 2: An Onshore Field on a MEOR Application

A second example is provided on the potential biogenic souring during a tertiary recovery application, namely MEOR. The study is described and discussed in previous papers in detail [27,93,94]. Biogenic souring was determined as one of the major risks at the beginning of the project and efforts to determine its magnitude and to develop potential mitigation started early. In the framework of the studies, numerical modeling using the experimental data delivered was a major element accompanying this project. After the laboratory analysis of both the injection and formation water, microbial growth was studied under various feeding conditions and important aspects of the assessment could be specified as provided in

Table 5. Properties of the field studies (Case 2) for biogenic souring.

Property	Data
Formation salinity	Injection water: highly saline, up to 160 g/L, sulfate content around 70 mg/L. Produced water: highly saline up to 180 g/L, no detectable sulfate
Formation rock mineralogy	Sandstone with siderite content of 2–10% wt.
SRB	Microbiological analysis of the samples from the field confirmed the presence of Desulfovibrionales (Desulfohalobiaceae) in the produced water as well as in the injection water; the field is under PWRI
Organic acids, VFA	Organic acids (e.g., acetate, lactate, formate) are products of the MEOR process. Their amount is high enough for complete reduction of the available sulfate by SRB
Pressure	30–70 bar under reservoir conditions; no detrimental effect on the SRB
Temperature	Injection water calculated 20–25 °C at the bottom hole during injection, reservoir temperature between 37° and 41 °C. Since SRB are present in the reservoir already, they can be active in a range valid for the reservoir studied
pH	Formation water pH~6.5, VFA due to MEOR can reduce the pH down to 5 (not considering pH-buffering effects of the formation rock); pH is in the range still allowing the activity of SRB
Permeability	Average permeability and the pore diameter are kavr 300 mD and 20 μm, respectively. This is high enough for the free movement of the SRB throughout the reservoir. On the other hand, the reservoir is highly heterogeneous
Injection rate	150 m3/day, including a defined concentration of nutrients for in situ MEOR
Partitioning coefficient	Based on high salinity formation water and with the assumption of a pH of 6, Kow is taken as 30 in the base case

. To summarize: SRB are present in the reservoir and the thermodynamic, petrophysical, and feeding conditions (conversion of injected nutrients to VFA during MEOR) are optimal for biogenic souring. This was confirmed with batch as well as dynamic experiments, as shown in Figure 9a,b. In Figure 9a, the result of a batch experiment is shown where MEOR nutrient was added to formation water alone as well as formation water containing formation sand and oil. It was concluded that after the initial incubation time (3 days), corresponding to the consumption of MEOR nutrient and subsequent formation of organic acids as potential SRB substrate, H₂S starts to form and is concurrent with a decrease in sulfate concentration. In the presence of sand and oil, the formation of H₂S is delayed and remains at a lower range until the 8th day. Figure 9b shows the result of a sand pack experiment in which sand is used as filling material as shown in Figure 9c. Here, it can be noted that H₂S is detected at the outflow only after a 30-pore volume (PV) injection, much later in comparison to the MEOR products acetate and ethanol, which are shown in the figure as total VFA. Supplementary experiments showed that only approximately 3% of the VFA generated by MEOR application are needed to transform all the existing sulfate to H₂S. This indicates that the generated H₂S in this case is limited by the amount of available sulfate. Furthermore, the H₂S concentrations measured at the outflow (sand packs) and the headspace (batch experiment) are significantly lower than the theoretical values calculated based on the initial sulfate values. This implies that other mechanisms like retention of sulfate and H₂S in the system play a role in the total amount of H₂S measured. To be able to predict the extent of the souring in the field application, numerical modeling was applied. The commercial reservoir simulator CMG STARS [92] was used to simulate laboratory continuous up-flow sand pack experiments. Observed compositional changes along the length of the sand pack are matched for various injection scenarios. A similar approach was used previously to model MEOR in a field application in the project. The MEOR modeling concept is based on modeling the metabolism kinetics by replacing the Monod equation with the available Arrhenius equation in STARS. The kinetic parameters were calibrated by using the laboratory data, including dynamic experiments [95].

A model of the reservoir sector selected for the MEOR field pilot is used to investigate the potential souring without mitigation. The model is further improved to simulate the H₂S dissipation in the reservoir by adding sulfate as another component to the system (formation and injection waters) and formed metabolites are applied in chemical reaction options (similar to Equation (1)) to generate H₂S with an assumed complete sulfate consumption. An injector with an injection rate of 150 m³/day and two producers with a total production of 150 m³/day are the well components of the model. The general characteristics of the model are provided in

Table 6. General characteristics of the sector model used in the study.

Grid	Corner Point
Number of cells	36(x) × 31(y) × 16(z)
Cell size Refinement	50 × 50 × 1.3 m 10 × 10 × 1.3
Porosity	0.018–0.291
Permeability	1–6130 mD (5 rock types)

. The model consists of 16 horizontal layers with varying permeabilities (Table 6; five rock types with permeabilities between 1 and 6130 mD) reflecting the vertical heterogeneity of the reservoir. Figure 10 shows the production predictions from one of the producer wells, situated 150 m away from the injector. As can be seen, the water cut is above 96% resulting in low oil production and, therefore, low H₂S content due to the preferred partitioning into the oil phase. H₂S breakthrough to the nearest production well (150 m to injection well) is calculated to be around one year after the injection start and reaches a maximum after 5 years. Note that in these runs, the retention of H₂S is neglected. Although the amount of H₂S dissolved in oil is larger than in water, the total H₂S production from the main producer’s water is more relevant during the years 1–2.5 due to the water cut of around 96%. With a reduction in the water cut with time, as demonstrated in related MEOR simulations, the absolute amount of H₂S being produced with the oil phase shows a steeper increase compared to the fraction from the water phase.

The siderite content of the reservoir rock reaches up to 10% wt. and the formation water contains ~1.2 mM ferrous iron. The corresponding retention reactions with generated H₂S may be modeled using the chemical options of the simulator. As experimental input was missing in this case, the geochemical reactions of H₂S are modeled by using Langmuir type adsorption curves for low and high cases. Figure 11 compares the H₂S concentration in oil for the 4th layer of the model after 8 years of MEOR application for (a) no retention and for the cases in which retention is introduced as (b) very-low with 0.001 mg iron sulfide per g of rock mg/g, (c) low with 0.01 mg/g, and the most realistic high case with (d) 0.1 mg/g rock.

As can be concluded for this case, a H₂S breakthrough to the full extent (i.e., similar concentrations near the injector and producer) is only observed for the no-retention case while the low- as well as the high-retention cases show only little or no breakthrough at all (realistic scenario: d). Note that in the retention runs, the potential changes in the porosity and permeability due to mineral precipitation were not considered.

In this case, the retention of H₂S and/or sulfate in the reservoir is the main controlling parameter. At the same time, the retained amount is one of the most uncertain parameters. To highlight this point, a sensitivity analysis was performed, and the results were evaluated with Monte Carlo simulations by using the CMOST option of CMG [92]. Mineral scavenging, sulfate content of injection water, and injection rates were selected as sensitive parameters and entered into a Monte Carlo uncertainty analysis as statistical distributions between the minimum and maximum values estimated from lab and field data; the sulfate content and injection rates were taken from field observations, while the values for mineral scavenging were estimated from the measured mineralogical compositions of the reservoir rock [27].

Figure 12 shows the statistical distribution of the H₂S amount produced from the P-1 with low (P10), high (P90) and mean (P50) probability cases. As shown in the figure, the total H₂S production rate in the most probable case (P50) is 0.349 kg/day, while the low-(P10) and high (P90)-probability cases are 0.158 kg/day and 0.541 kg/day, respectively. Since conclusive experimental data under reservoir conditions are not always available, mineral scavenging as well as the following parameters can be used for the prediction of H₂S formation and production:

-

Sulfate content of the injection/formation water: The sulfate content of both the injection and formation water can be determined from the water analyses of the injection and production waters. However, if sulfate is introduced via the injection water, and if the formation water does not contain sulfate in the first place, it might be challenging to estimate the timely and spatial distribution of sulfate in the reservoir;

-

Amount of substrate available for SRB in the reservoir is a critical parameter and independent of whether SRB are found in injection water and/or found in the reservoir. In the case, of a typical MEOR application, VFAs are generated as the product of in situ fermentation reactions and these are available in excess for the complete reduction of sulfate. In this case, they cannot be used in an uncertainty analysis;

-

Partitioning coefficients: The partitioning of the H₂S between the phases in the reservoir as well as in the wellbore can be determined well if the compositions and thermodynamic conditions are known. Yet, a careful examination of the existing literature implies that there is still a margin of uncertainties, especially for the oil/water partitioning coefficients.

6. Mitigation of Biogenic Souring

The mitigation of biogenic souring is subject to numerous scientific and engineering studies and field applications. Biogenic souring is commonly inhibited by adding biocides to surface water loops or injection fluids. However, these are often environmentally questionable and associated with excessive costs and a limited efficiency [3,34]. Furthermore, their usage is heavily regulated. In addition, most biocides are non-specific and cannot be used to target SRB exclusively if the growth of other microbial species is desirable. This issue is relevant in MEOR as well as in oil-to-methane conversion projects. Once the risk of biogenic souring is defined, some proactive measures such as removal of the components responsible for biogenic souring (sulfate and VFAs) may be helpful in limiting the extent of souring, but possibly with little impact on long-term trends. Co-injection of a chemical that serves as a souring inhibitor should be part of the mitigation measures in most cases. A list of the recent studies on potential inhibitors for biogenic souring is provided in

Table 7. Summary of potential inhibitors for biogenic souring.

Chemical/Application Mode	Conclusion	Ref.
Nitrate
Draugen North Sea oilfield (~70 °C): The seawater used for injection is treated by filtration, de-oxygenated using the scavenger sodium bisulfite, and sterilized using sodium hypochlorite plus biocide	70–80 mg/L in injection water; the dosage could be reduced to 50 mg/L, was successful during the first 3 months. Stopping the treatment for 6 weeks caused quick rebound of H2S, resuming nitrate addition brought the souring under control again	[25]
Injection of nitrate in Nigeria Bonga field; tested first in laboratory experiments e.g., sandpacks	One year after the start of the project, there was only limited amount of water production and not yet H2S	[22]
Up-flow packed bioreactors inoculated with water produced from an oil field and injected with lactate	Nitrate gives less SRB inhibition, under experimental conditions is concluded that use of nitrite is more favorable than use of nitrate	[82]
Amendment of nitrate (5 mM NaNO3) in a water-oil separation tank in an offshore platform in Brazil	The conclusion was that nitrate treatment was effective in this surface facility, but continuous application would be necessary to suppress SRB	[13]
The injection water with low sulfate concentration (~1 mM). Nitrate (2.4 mM, 150 ppm) was added at the water plants continuously or with week-long pulses of 14 mM	For low temperature fields, with low sulfate concentration injected, the increase in souring can be stopped by nitrate injection. A pulsing strategy appears to give better results than continuous injection	[96]
Ca(NO3)2 continuously applied at 150 ppm field wide. A biocide, acrolein, was also applied in weekly batches to treat the injection lines and prevent nitrate-fed biofilm	After one month of treatment, H2S in the produced gas from one gathering system decreased from an average of 170 ppmv to 110 ppmv. A corresponding decrease in H2S in the produced water was observed	[14]
Column experiments to investigate the effects of nitrate injection on SRB inhibition and MEOR	Resulted in significant sulfide control (40–100% reduction) within few days	[84]
Nitrite
Using up-flow packed bioreactors both nitrate and nitrite are effective sulfide removers	Nitrite is a stronger SRB inhibitor. Nitrate gives less SRB inhibition, because it is only partially converted to nitrite	[82]
Testing various inhibitors among them nitrite in full field numerical simulations	1 mM of nitrite can reduce souring by 92%, while perchlorate (1 mM) or nitrate (1 mM) reduce sulfide accumulation by 57% and 80% respectively	[97]
Molybdate
Batch/dynamic experiments and field test as a part of a MEOR project; a Na2MoO4 concentration of 0.5 mM was found as optimum	Molybdate inhibited the activity of SRB efficiently, while still enabling MEOR due to its specific inhibition mechanism. Adsorption is an issue!	[27]
Perchlorate
Columns with marine sediment flushed with coastal water amended with yeast extract and nitrate, chlorate, or perchlorate	The study indicates that (per)chlorate show great promise as inhibitors of sulfidogenesis in natural communities and provides insight into which organisms and respiratory processes are involved	[98]
With data from sandpack experiments, a reactive transport model compared perchlorate with nitrate.	Model runs previously matched with experimental data suggest that perchlorate is more effective than nitrate on a per mole of inhibitor basis.	[78]
Experiments under different matrix and inoculation conditions to evaluate treatment efficiency	Advantages of souring control with (per)chlorate treatments, and the application of galvanic signal as an economic, in-situ monitoring approach	[99]
A combination of laboratory experiments and numerical models to compare sulfide inhibitors	Perchlorate inhibits sulfate reduction by promoting sulfide oxidation	[97]
Others
A high-throughput screening strategy to identify inhibitors of SRMs, quantitatively ranked the selectivity and potency of compounds and synergistic interactions	Zinc pyrithione is the most potent inhibitor of H2S generation and is several orders of magnitude more potent than commonly used industrial biocides.	[4]
Some novel mitigation agents are cited	Adjuvants, anthraquinone, bacteriophage, diphenyliodonium salts	[18]
Combined
Microbiocide and nitrate are applied, either alone or in combination, in many seawater injection systems as controls to mitigate MIC or reservoir souring or both	Souring experiments, does not reflect the field data in that the dramatic decreases in H2S production seen in the model are not typical of the much less dramatic changes observed in the field	[100]
MICs of six broad-spectrum biocides and two specific metabolic inhibitors and fractional inhibitory concentration indexes (FICIs) for controlling a SRB	Nitrite was synergistic (FICI < 1) with all but one biocide due to its specific inhibition of dissimilatory sulfite reductase. Hence, combining nitrite with biocides allows more efficient and cost-effective control of SRB	[101]
Four new biocides and one commonly used biocide were tested in bioreactors operated for 591 days	SRB activity could recover within a period varied from 15 to 60 days. Neem-oil (NO) (1.5% vol/vol) and Dazomet (0.5% vol/vol) were the most efficient in controlling SRB	[85]
Sandpacked bioreactors simulating moderately high temperature oil reservoir for souring control	Co-injection of 2 mM nitrate and 0.75 mM BAC completely ceased H2S production and controlled souring.	[19]

. The most relevant applications include the following.

6.1. Nitrate/Nitrite

The relevance of nitrate in preventing H₂S formation in sewage systems in the framework of biocompetitive exclusion has been known for almost a century and was demonstrated in the laboratory with oilfield brine microcosms and bioreactors decades ago [89]. However, using nitrate to treat souring in oil production operations began only in the 1990s [102,103,104]. Because of its low cost, easy availability, and high solubility in water phases, it had been claimed as a preferred solution over the use of biocides. A high number of studies report its field application, e.g., [14,22,56,105]. The proposed mechanisms for reducing H₂S formation through nitrate or nitrite modification include the following: (1) the preferential use of nitrate as an electron acceptor instead of sulfate by some species of SRB, thus suppressing microbial sulfate reduction; (2) suppression of SRB activity as a result of competition between heterotrophic NRB and SRB for the same electron donors with SRB being out-competed; and (3) inhibition of SRB activity via intermediate compounds generated during nitrate reduction by heterotrophic NRB or NR-SOB (nitrate-reducing, sulfide-oxidizing bacteria), e.g., NO, N₂O, and, especially, nitrite, NO₂⁻. Nitrite is claimed to be an effective souring inhibitor agent while also mediating the oxidation of sulfide [24,64]. In general, reported nitrate and nitrite concentrations applied in laboratory and field applications range from 10 mM up to 100 mM. However, some species of SRB possess high nitrite reductase activity which allows them to overcome this inhibition by reducing nitrite to ammonia. Veshareh and Nick investigated a novel approach of evaluating different reaction pathways to design reaction models that reflect governing microbial processes during the souring processes in a set of batch and flow experiments, suggesting dissimilatory nitrate reduction to ammonium as the main reaction pathway [20]. Further, they illustrate the inhibition by nitrite as the major mechanism of the nitrate treatment process and introduce an inhibitory nitrate injection concentration: a Required Inhibitory Nitrate Concentration (RINC) of 15 mM is supposed to inhibit souring, regardless of the nitrite inhibition effect and the distance between injection and production wells. Furthermore, they demonstrate that the ratio of the nitrite–nitrate reduction rate can be used to estimate nitrate treatment effectiveness.

However, using nitrate or nitrite as inhibitor agents bears the potential for significant drawbacks which were studied in the recent literature:

-

The fundamental inhibition mechanisms are well understood, but the complex biological, biochemical, and abiotic interactions in a reservoir cannot be predicted fully. In recent studies [56], the inhibitory effect previously claimed to be due to nitrate injection on the Gullfaks field could be assigned to mixing and biofilm models instead and was potentially not related to the nitrate application;

-

Continuous and well-defined dosages of high nitrate concentrations are needed for direct inhibition (>10 mM). A sub-optimal or insufficient nitrate treatment strategy may result in higher H₂S production from production wells despite a lower total amount of generated H₂S inside the reservoir; but unconsumed nitrate in the case of excessive nitrate injection could increase the need for post-production water treatment [106];

-

The ability of some SRB, e.g., desulfovibrio desulfuricans, to use both nitrate and sulfate as electron acceptors represents an additional risk, because souring-associated microorganisms can be unintentionally enriched [4,107];

-

Also, depending on the composition of the reservoir brine (e.g., presence of dissolved calcite and/or iron) various types of mineral precipitation can be observed. Nitrite injected or formed as a fist intermediate during nitrate injection is very reactive towards Fe(II) and can lead to Fe(II) oxidation and the subsequent formation of low-soluble Fe(III) compounds. Concurrent to iron reduction, calcite minerals can be precipitated; the precipitation of calcite during bacterial activity is known as microbially induced calcite precipitation (MICP). Precipitation of minerals in porous media decreases the porosity and permeability, thus decreasing the injection-production capacity of the reservoir. The mechanism of calcite precipitation during anaerobic denitrification is not completely understood, but it is supposed to be a combination of several effects including a local pH increase, the production of CO₂ (see Equations (2)–(4)), the degradation of Ca²⁺-complexing organic acids, and enhanced nucleation via exopolymeric substances (EPS) [108];

-

There are concerns that nitrate or nitrite can exacerbate corrosion due to their oxidizing potentials; it was reported that during nitrate addition, corrosion rates increased due to MIC caused by nitrate-reducing activity forming nitrite or elemental sulfur [108,109].

6.2. Molybdate

Originally proposed by Peck [110], molybdate has been studied as an SRB inhibitor in recent years for relevant technical applications such as in biogas generation [111] or within the oil and gas industry [4,27]. As an analogue to the sulfate ion, molybdate targets the first step of microbial dissimilatory sulfate reduction; very specifically, it targets the sulfurization of ATP to adenosine-5-phosphosulfate (APS) under the release of pyrophosphate. The involved enzyme, ATP sulfurylase or sulfate adenylyl transferase, is unable to discriminate between sulfate and molybdate. The resulting molybdate intermediate is not stable, leading to gradual depletion of intracellular ATP and eventually to cell lysis, after which the molybdate is still available for further reactions [110,111,112]. Despite this catalytic mode of action, its low toxicity, and its low environmental hazard class, molybdate has not been employed in the petroleum industry widely. In addition to higher costs compared to other inhibitors, retention of molybdate in the reservoir by adsorption or precipitation has been considered an obstacle for its broad implementation.

Kögler et al. mentioned that molybdate inhibited the activity of SRB very efficiently, while still enabling a planned in situ MEOR due to its specific inhibition mechanism [27]. As shown in Figure 13, in batch experiments, 0.5 mM molybdate inhibited SRB growth more efficiently than 100 mM nitrate. The experiments also showed that a preflushing of molybdate delayed microbial sulfate reduction under dynamic (sand pack) conditions. However, significant molybdate adsorption onto reservoir rock occurred, and the maximum Langmuir saturation was estimated to be 34 μM_Molybdate per gram of rock. Residual oil allowed for further propagation of molybdate in sand pack reservoir models, but a pH < 6 limited the efficiency of molybdate due to its rapid adsorption. Molybdate adsorption should be minor on predominantly negatively charged quartz sand due to electrostatic repulsion underlying the significance of accessory mineral components such as gypsum, pyrite, or oxidized pyrite equivalents present in reservoir brine which can make up most of the fluid–rock contact, thus increasing the adsorption [27].

Under favorable souring conditions, the experiments also demonstrated the localized formation of macroscopic iron sulfide precipitations in sand packs with a corresponding four-fold permeability decrease after 40 days of continuous operation. The calculated mean sulfate reduction rate was 52 μM_SO4/h with molybdate preflushed sand packs showing a delay in sulfate reduction. This suggests an inhibitory effect, even if molybdate is partially adsorbed. In the corresponding pilot field of the same project, MEOR performed as a “huff and puff” application, sodium molybdate is reported to be injected alongside a nutrient solution with a concentration of 0.1 g/L, and no H₂S production was reported during the back production [67]. No other field applications were reported until now. More field trials are required to decide on the potential of molybdate as an inhibitory agent.

6.3. Perchlorate

The inhibition of microbial sulfate reduction by using perchlorate was investigated by many authors [109,113]. No effluent sulfide was observed when packed columns inoculated with marine sediment were flushed with coastal water amended with yeast extract and perchlorate (10 mM). Veshareh et al. suggest that, on a field scale, perchlorate inhibits H₂S accumulation by inhibiting sulfate reduction and promoting sulfide oxidation. Their field-scale numerical simulations predict that injection of 1 mM of nitrite can reduce souring by 92%, while perchlorate (1 mM) or nitrate (1 mM) have the potential to reduce H₂S accumulation by 57% and 80%, respectively. To date, no field application using perchlorate is reported [97].

6.4. Other Inhibitors

The use of other chemical inhibitors, among them tungstate and mono-fluorophosphate, were investigated at the laboratory scale in the literature. Klueglein et al. indicate that, in batch experiments, mono-fluorophosphate only inhibited H₂S formation at very high concentrations (1 g/L), where strong precipitation was observed with the high-salinity brine specific to the case studied [114]. Carlson et al. (2017) developed a high-throughput screening strategy to identify potent and selective inhibitors of SRM, quantitatively ranked the selectivity and potency of hundreds of compounds, and identified previously unrecognized SRM-selective inhibitors and synergistic interactions between inhibitors [4]. They identified zinc pyrithione as the most potent inhibitor of SRB indicating that it is several orders of magnitude more potent than commonly used industrial biocides. Greene et al. investigated the combined use of nitrate, nitrite, and molybdate with biocides based on the fractional inhibitory concentration index (FICI) [101]. They affirm the usefulness of such combinations and report a strong synergy between nitrite and the biocides glutaraldehyde, benzalkonium chloride, and bronopol. Although molybdate is a specific SRB inhibitor, they found that molybdate was less suitable than nitrite. Prajapat et al. studied various souring control treatments by evaluating them in sand packs [19]. Results revealed that continuous injection of 4 mM nitrate and 1.5 mM benzalkonium chloride (BAC) alone was not able to control souring. However, co-injection of 2 mM nitrate and 0.75 mM BAC completely ceased H₂S production and controlled souring, highlighting the efficiency of a combination of biocides due to synergistic effects.

6.5. Limits and Perspectives

Eventually, one of the biggest restraints is the limited capability of current models and numerical simulations to capture the complex biogeochemistry of biogenic reservoir souring: a comprehensive model needs to consider not only microbial sulfate reduction, but several competitive, complementary, and partially interfering microbial metabolic pathways such as fermentation, acetogenesis, and nitrate reduction (potentially coupled to sulfide oxidation and iron reduction). The kinetics of these processes are determined not only by various carbon sources and electron acceptors, but also on the characteristics of microbial growth, i.e., adsorption and biofilm formation versus planktonic growth. Further, growth kinetics might differ among oilfields according to their physicochemical parameters (such as salinity and temperature) which shape, in turn, the microbial community. Modeling efforts such as reported by Veshareh and Nick are a good start into this direction [20]. These microbiological considerations are further complicated by geochemical parameters which need incorporation into existing models, namely the reservoir mineralogy, to account for processes such as sulfide scavenging or pH buffering capacities which in turn can impact both sulfide speciation and microbial activity [115]. However, such modeling efforts will require the transdisciplinary collaboration of microbiologists, geochemists, and reservoir engineers.

Additionally, it will be necessary to think beyond universal biocides: depending on the reservoir mineralogy [108], the use of nitrate might be an option, but innovative, selective and environmental-friendly alternatives to inhibit reservoir souring, such as molybdate, should also be considered. Also, more research needs to be conducted on more exotic mitigation techniques such as sulphate removal, anthraquinone, adjuvants, and bacteriophages [18].

Finally, all potential mitigation techniques need to be tested in the field and not only experimentally in the lab. Here, it is important that companies take both the potential risk to employ new souring strategies as well as report those to the scientific public.

7. Workflow for the Assessment of Biogenic Souring

Risk assessment of waterflooding, chemical (biopolymer and biosurfactant) injection, or MEOR projects should consider biogenic souring as one of the relevant risks that may affect project scope, cost, and schedule. From a project management point of view, the risk should be defined and quantified as soon as possible, and mitigation measures should be planned and taken as per plan. Obviously, this work should be conducted based on an assessment plan considering state-of-the-art related technology and experiences.

Kalpakci et al. described a decision strategy which is one of the earliest examples of its kind [116]. Numerous studies apply assessment procedures, all of which have in common that they include experimental and numerical parts for the definition and mitigation of souring. In a list of strategies to reduce the risk of reservoir souring provided by Gieg et al., the removal of the sulfate (reverse osmosis, nanofiltration, dilution of sulfate containing injection water, etc.) or microbial substrates leading to H₂S generation (VFAs and nutrients) with various methods are cited as proactive measures [34]. Besides, biocides and/or nitrate/nitrite/molybdate are proposed as mitigation agents, as well as chemical scavengers as remediation measures on the producing side.

According to our laboratory and field experience, also taking into account previous similar attempts and the latest technological developments, the evaluation of biogenic souring and its mitigation can be carried out in three successive but interdependent steps:

7.1. Step 1: Preliminary Assessment

The first step is the definition of the risk in terms of its impact on various stages of oil production. In this step, the probability of the risk, its potential level in various stages of operations, its impacts on schedule, and project economics and scope are defined based on preliminary data and methods. For example, applying the collected data to the stoichiometry provided in Equation (1), can provide a good initial estimate of the probability and extent of reservoir souring. HSE issues should be evaluated at this first step, and if these issues prove difficult to resolve, the future of the project can be reevaluated.

7.2. Step 2: Modeling and Quantification

In the second step, the quantification and spatial and temporal distribution of souring can be assessed using reservoir models supported by experimental and field data. As discussed in previous chapters, reservoir models calibrated with field data are used to describe the geology and hydrodynamics first. Modeling H₂S generation, its partitioning and transport in the subsurface complete the souring reservoir model, which then helps to predict the H₂S concentration in the production as a function of time. To improve the reliability of the calculations, this part of the study should be supported by laboratory experiments and field measurements (Table 4) where applicable. The calibrated reservoir and experimentally validated H₂S models coupled to each other are the prerequisites for realistic predictions. However, the remaining uncertainties caused by geological and petrophysical as well as other parameters controlling biogenic souring can impact the prediction quality. In this case, it is proposed to perform the predictions using statistical approaches. As discussed in previous chapters, retention of H₂S, concentration, and reactions of sulfate with components of formation water as well as partitioning of H₂S into in situ phases may be appropriate parameters whose values should be considered in the context of probability distributions when supported by theoretical and experimental analyses. Modern reservoir simulators include such options like the Monte Carlo method to introduce the uncertainties and estimate the target parameters; at the end of this assessment, H₂S production can be estimated as function of time and location within provided probability ranges from low and high to mean (P10, P90, and P50) values. The identification of biogenic souring mitigation measures, if needed and doable with available tools and data, should also be included in this step. The impact of using various chemicals and biocides on the reservoir performance can also be assessed with the validated reservoir souring model based on experimental data. The spatial distribution of the chemicals to be injected can be predicted as a function of time, therefore providing a possibility to define and fine tune the injection strategies considering the full field scale. On the other hand, the use of chemical inhibitors such as nitrate can be harmful to the environment if, for example, they contaminate groundwater. An environmental impact assessment (EIA) should be planned to assess whether and under what circumstances the chemicals may be harmful to the environment and to consider remedial measures if necessary.

7.3. Step 3: Monitoring

Monitoring of the biogenic souring and the response of its mitigation should be part of the workflow. Field measurements of various parameters, such as concentrations of H₂S, sulfate, and mitigation chemicals, can be used to assess the success of mitigation measures and to re-evaluate the models used. The models can then be recalibrated based on the data and re-used for updating the mitigation operations and their success. In other words, the measurements are used to reduce the uncertainties in each step of modeling thereby increasing the quality of the predictions and mitigation measures.

The workflow is shown in Figure 14 The experimental work provides data for the numerical modeling, and each improvement of the experimental data leads to an improvement of the numerical predictions by reducing the uncertainties of the input parameters. The workflow thus results from the interaction of all steps in a kind of continuous improvement cycle.

8. Conclusions

Biogenic souring is a common risk in secondary (waterflooding) and tertiary oil production (EOR) operations. Both the reservoir and the production and injection history vary from case to case, so there are no one-size-fits-all solutions. However, many years of expertise and the experience of related industries can help increase the reliability of biogenic souring assessment and management. With this in mind, we evaluated recent and relevant studies on biogenic souring and combined the results with our practical expertise from a pragmatic reservoir-engineering perspective to support evaluation and decision quality on the topic. The conclusions can be summarized as follows:

-

Technological developments in microbiology and petroleum industry simulation tools and techniques provide a sufficiently good analytical background for the prediction of H₂S formation and its spatial and temporal distribution in the reservoir. The predictive capability depends heavily on the reliability and representativeness of the model created. The process of calibrating the simulation models requires high-quality, reliable data from theoretical work, chemical–physical modeling, and field measurements;

-

H₂S retention in the reservoir can be significant if reservoir rock contains iron-bearing minerals like siderite; as shown in a MEOR case analysis, even minor siderite content can inhibit the H₂S generation totally;

-

Considering the uncertainties in the associated inputs as well as in the geological unknowns, the use of the Monte Carlo method is suggested for numerical evaluation, where the uncertainties of each associated parameter can be considered to define the H₂S production in terms of a probability distribution. State-of-the-art commercial reservoir simulators provide user friendly interactive codes for such analysis;

-

Nitrate and nitrite are the best-known and best-studied chemicals for biogenic souring inhibition, although their reported effectiveness in field applications is controversial. Molybdate and perchlorate are the emerging candidates for SRB inhibition, but further investigation is needed for environmentally friendly, safe, and economical applications;

-

A workflow for assessing and managing biogenic souring in oil reservoirs during secondary and tertiary oil recovery is presented. The workflow provides a continuous cyclic process to improve the quality of the assessment and decisions based on new laboratory and field data in each cycle;

-

Further research is required for environmentally friendly, safe, and economical applications. This includes advanced testing of existing inhibitors and the development of new chemicals that can effectively combat reservoir souring and its severe impact on production. Easy-to-use analytical methods to detect the chemicals produced and using them to calibrate numerical simulation codes are essential. Reports on field applications to monitor H₂S production in operations would certainly support the research work.