One mismatched ion or stray microbe can turn an injector into a choke point. Data from tight sandstone to field case histories show why reservoir and production engineers are doubling down on compatibility tests—especially corefloods—before pushing a single barrel underground.
Waterflooding (secondary recovery) and chemical enhanced oil recovery (tertiary) keep barrels flowing—but the wrong water can suffocate a reservoir. In tight sandstone, mixing incompatible injection and formation waters precipitated calcium carbonate (CaCO₃) and aluminosilicate scales, halving core permeability (–50%) and shrinking median pore radius by ~21.6% (www.mdpi.com) (www.mdpi.com). In a documented well case, carbonate scaling drove up injection pressure to the point it was “making it impossible to inject water” (www.mdpi.com).
The takeaway, echoed across laboratory studies and field practice: rigorous compatibility testing is essential before field injection to avoid costly injectivity failures (www.mdpi.com).
Compatibility risks and damage mechanisms
Laboratory and field evidence converge on several failure modes. Mineral scale and precipitation—classically calcium carbonate/sulfate, iron sulfides, aluminosilicates—arise when mixed waters or temperature/pressure shifts push ions past solubility. Gong et al. describe “induction, damage, and stabilization” scaling stages in corefloods, ending in 50% permeability loss (www.mdpi.com).
Clay swelling and fines migration are just as unforgiving. Even a few percent expandable clays (e.g., smectite) can swell or disperse under dilute brines. In unconsolidated packs, 1.4–6.8% montmorillonite or kaolinite reduced permeability by 10–40% versus clay-free media; montmorillonite swelled ~39% and kaolinite ~15% by volume (high-porosity samples still saw ~2–5% perm loss), but low-perm rocks are far more vulnerable (www.researchgate.net) (www.researchgate.net).
Particles are another silent killer. Suspended solids—from silt to corrosion oxides or nickel fines—plug throats and form cakes. Bennion et al. list “suspended solids, corrosion products, emulsions, oil-wet hydrocarbon agglomerates” among damaging constituents (www.researchgate.net). A practical screen from field experience: keep total suspended solids (TSS) below about 5–10 mg/L and most particles under 1 μm to avoid formation plugging (benchmarks vary by formation). Usman’s pore-throat analysis drove an 11 μm absolute filter specification to prevent plugging (www.researchgate.net).
Corrosion and gas interactions compound the risk. Elevated dissolved O₂ or CO₂ accelerates tubing and equipment corrosion, seeding the water with oxides and sulfides; oxygen levels are often kept below 1 ppm for critical injectors. Targeted chemical control is routine, including oxygen scavengers and sour gas management. Chemical programs are typically delivered through controlled feeds; for example, facilities often rely on precise chemical metering via a dosing pump and on scavenging packages such as oxygen/H₂S scavengers.
Microbiological fouling—sulfate-reducing bacteria (SRB), iron bacteria—produces biofilms and H₂S souring. Standard microbiology (heterotrophic counts, Most-Probable-Number for SRB, PCR assays) flags loads; if counts exceed thresholds (often a few hundred CFU/mL), biocides or ozone treatment are recommended. A Chinese standard identifies SRB and iron bacteria as control indices for flood water and ties limits to core damage criteria (patents.google.com) (patents.google.com).
Organic/oil contamination and emulsification add another layer. Surface or recycled produced water can carry oil; rolling tests with reservoir oil reveal emulsion stability. Bennion et al. again flag “oil and grease, emulsions, oil-wet hydrocarbon agglomerates” as harmful (www.researchgate.net).
Laboratory screening sequence and models
Programs start with brine analysis: ion chromatography and ICP (inductively coupled plasma) define TDS (total dissolved solids), cations, anions, hardness, pH, alkalinity, H₂S, O₂, microbiology, oil content, and TSS. These feed screening models—Langelier index and PHREEQC/PhreeqC—to predict scale risks and supersaturation indices.
Static compatibility tests follow. Simple jar tests mix formation and candidate injection brines at ratios (1:1, 1:4, etc.) at ambient and reservoir temperature (e.g., 80–90 °C), then filter or centrifuge to weigh precipitates and ID minerals by XRD (X-ray diffraction) or XRF (X-ray fluorescence). These batch tests are qualitative—they “neglect factors such as flow rate”—so they reveal what can precipitate (e.g., “CaCO₃” vs “FeS”) but not the dynamic impact on flow (www.mdpi.com).
Clay dispersion and swelling assays are next in clay‑rich reservoirs. Shaker tests agitate crushed shale/clay in injection brine at downhole temperature and measure turbidity or sedimentation volume; linear swelling tests track thickness changes in clay pellets; volumetric swelling uses bentonite columns. Such results gate whether to adjust brine composition or add inhibitors.
Microbial screening incubates anaerobes (for SRB) and aerobes, quantifying CFU/mL and running “biocide efficacy” trials. Where continuous disinfection is needed, facilities often deploy biocides or non‑chemical systems—ultraviolet units with a 99.99% pathogen kill rate and low operating cost—via equipment like an ultraviolet reactor.
Particle and filtration tests close the loop. TSS is quantified by filtering (e.g., 0.45 μm), then gravimetry; particle size distributions come from laser diffraction or coulter counters; filtration trials push water through 5–100 μm meshes or ceramic membranes to profile resistance. These data set the filtration train—often including automated screens—to match pore-throat realities; cascaded mesh steps are frequently implemented using an automatic screen. Fine polishing is typically assigned to replaceable media like a cartridge filter, especially when studies (e.g., Usman’s) indicate an 11 μm absolute cutoff to prevent plugging (www.researchgate.net).
Dynamic coreflood protocol and benchmarks
The cornerstone test is the compatibility coreflood. Intact core plugs (1–1.5″ diameter) are cleaned/dried and baseline permeability measured. Plugs are then saturated with formation water and aged with reservoir crude (to restore wettability). After baseline injectivity, the candidate fluid (surface water, treated produced water, or specialized brine) is injected at reservoir temperature and pressure. Flow rates target field-representative shear; volumes span tens of pore volumes (PV), typically 10–50 PV, to capture progressive damage.
Interpretation centers on (a) permeability change versus injected PV, (b) any relative permeability/residual oil shifts, and (c) effluent chemistry and solids. Gradual pressure rise indicates plugging; sharp jumps signal severe blockage. Restoration slugs of clean water or chemical washes often test reversibility. Benchmarks are clear: many view ≥90–95% return permeability (≤5–10% damage) as “compatible,” with a practical target of <20% permeability loss. A Sinopec method set quality limits so that after 50 PV injection, the core’s permeability loss was ≤20% (patents.google.com). Results showing >20–30% permanent loss typically trigger a re‑evaluation (more treatment, tighter filtration, or a different source).
Case outcomes underline the stakes. Usman (2015) found a 25% produced‑water/75% freshwater blend crushed core permeability by ~80% relative to pure freshwater; adding 2000 ppm biocide plus 11 μm filtration cut the damage, with ~47% loss instead of 80% (www.researchgate.net) (www.researchgate.net). A coreflood comparing 100% freshwater versus 50% produced water showed the 50% produced‑water flood yielded 16% less oil (OOIP) (www.researchgate.net). Gong et al. likewise reported ~20% pore radius reduction and 50% perm loss from scaling in tight rock; in an actual well, carbonate scaling raised pressure “making it impossible to inject water” (www.mdpi.com) (www.mdpi.com).
When chemical disinfection is part of the mitigation plan confirmed by corefloods, facilities often standardize delivery hardware; for instance, placing the microbiology program on a dedicated dosing pump can hold the 2000 ppm biocide slug precisely where tests show it works.
Advanced characterization toolset
NMR (nuclear magnetic resonance) T₂ distributions and X‑ray CT imaging localize and visualize damage in cores before and after floods. Gong et al. used NMR to show losses in medium‑scale pore volumes consistent with measured scale deposition (www.mdpi.com). Microfluidic micromodels—transparent chips patterned with reservoir‑like pore networks—provide real‑time views of fines and precipitate clogging.
Effluent analytics complete the picture: solids by XRD for mineralogy; ion profiles (ICP/IC) PV‑by‑PV to see which salts drop out and when; gel permeation testing in polymer floods to check degradation.
Turning lab results into operating specs
The endgame is operational limits and a treatment train that holds them. If TSS risk is high, designs add finer sand filtration and multistage clarification to meet targets such as TSS below about 5–10 mg/L and particle size mostly under 1 μm (rule of thumb from field experience). Plants commonly specify equipment such as a high‑rate clarifier—see a clarifier—followed by granular media like a sand/silica filter for load reduction.
If bacterial counts are high, systems add continuous biocide or UV; ultraviolet reactors provide a 99.99% pathogen kill rate without chemicals at low cost and are applied via an ultraviolet unit. When oxygen and sour gas are the problem, oxygen scavengers and H₂S management are deployed with packages such as oxygen/H₂S scavengers. Where clay swelling tests fail, water is reformulated (e.g., electrolyte adjustments or clay inhibitors).
The Sinopec protocol shows how lab thresholds become standards: by testing TSS, oil content, bacteria, etc., allowable maximum values are derived so that a ~50 PV coreflood causes only 20% permeability loss; examples include TSS ≤ 3–5 mg/L, median particle diameter ≤ 0.5 μm, and free oil ≤ 50 ppm (patents.google.com). Field sensors—turbidity meters, pH, ionic probes—are then used for real‑time compliance.
Ultimately, thorough lab evaluation turns uncertain risk into measurable controls. As Indonesian literature emphasizes, “tinjauan perlu dilakukan … agar tidak terjadi kerusakan formasi” (study interactions so formation damage does not occur) (journal.lemigas.esdm.go.id).
Method references and case data
Predictive models and lab methods cited include Langelier and PHREEQC/PhreeqC for water chemistry screening; static jar testing at ambient and reservoir temperature; clay swelling and dispersion assays (linear swelling, shaker tests, volumetric measures); microbiology (heterotrophic counts, MPN for sulfate‑reducers, PCR); particle tests (TSS by 0.45 μm filtration and gravimetry, laser diffraction, filtration through 5–100 μm meshes); and compatibility corefloods at reservoir temperature/pressure over 10–50 PV with effluent ICP/IC and XRD/XRF. Key thresholds cited include oxygen often held below 1 ppm for critical injectors; microbial interventions when counts exceed a few hundred CFU/mL; and field rules of thumb such as TSS < 5–10 mg/L with particle size mostly < 1 μm (benchmarks vary by formation). Compatibility benchmarks commonly target ≥90–95% return permeability (≤5–10% damage), with a working goal of <20% loss; Sinopec’s criterion of ≤20% loss after 50 PV is one example (patents.google.com).
Coreflood case data include: Gong et al. (tight sandstone) showing CaCO₃‑driven scaling that reduced permeability by 50% and shrank median pore radius by ~21.6%, with “induction, damage, and stabilization” phases and a field case “making it impossible to inject water” (www.mdpi.com) (www.mdpi.com); and Usman (2015) reporting ~80% permeability loss for a 25% produced‑water/75% freshwater blend versus pure freshwater, improved to ~47% loss with 2000 ppm biocide plus 11 μm filtration, and 16% lower oil (OOIP) recovery for a 50% produced‑water flood versus 100% freshwater (www.researchgate.net) (www.researchgate.net).
References: Sugihardjo (2004) (journal.lemigas.esdm.go.id); Usman (2015) (www.researchgate.net) (www.researchgate.net); Bennion et al. (2001) (www.researchgate.net); Gong et al. (2024) (www.mdpi.com) (www.mdpi.com); Aksu et al. (2015) (www.researchgate.net); China Petrochemical Corp (2022), patent CN115544754A (patents.google.com). Sources: Peer‑reviewed industry/academic studies and standards have been used, including Chinese and Indonesian technical publications.jpñ (All citations above correspond to these sources, e.g.: Gong et al., 2024 (www.mdpi.com) (www.mdpi.com); Usman 2015 (www.researchgate.net) (www.researchgate.net); Bennion et al. 2001 (www.researchgate.net); Aksu et al. 2015 (www.researchgate.net); Sugihardjo 2004 (journal.lemigas.esdm.go.id); Sinopec patent CN115544754A 2022 (patents.google.com)).
