Marine sulfate-reducing bacteria cause serious corrosion of iron under electroconductive biogenic mineral crust

Iron (Fe0) corrosion in anoxic environments (e.g. inside pipelines), a process entailing considerable economic costs, is largely influenced by microorganisms, in particular sulfate-reducing bacteria (SRB). The process is characterized by formation of black crusts and metal pitting. The mechanism is usually explained by the corrosiveness of formed H2S, and scavenge of ‘cathodic’ H2 from chemical reaction of Fe0 with H2O. Here we studied peculiar marine SRB that grew lithotrophically with metallic iron as the only electron donor. They degraded up to 72% of iron coupons (10 mm × 10 mm × 1 mm) within five months, which is a technologically highly relevant corrosion rate (0.7 mm Fe0 year−1), while conventional H2-scavenging control strains were not corrosive. The black, hard mineral crust (FeS, FeCO3, Mg/CaCO3) deposited on the corroding metal exhibited electrical conductivity (50 S m−1). This was sufficient to explain the corrosion rate by electron flow from the metal (4Fe0 → 4Fe2+ + 8e−) through semiconductive sulfides to the crust-colonizing cells reducing sulfate (8e− + SO42− + 9H+ → HS− + 4H2O). Hence, anaerobic microbial iron corrosion obviously bypasses H2 rather than depends on it. SRB with such corrosive potential were revealed at naturally high numbers at a coastal marine sediment site. Iron coupons buried there were corroded and covered by the characteristic mineral crust. It is speculated that anaerobic biocorrosion is due to the promiscuous use of an ecophysiologically relevant catabolic trait for uptake of external electrons from abiotic or biotic sources in sediments.

The sum reaction is Note that a mere co-oxidation of HCOH and Fe 0 can result in the same equation.

Metal loss rates
The rate (velocity) of the loss of metal mass (m) by corrosion is = dm/dt or, if constant during an experimental time interval Δt, also = Δm/Δt. Division by density, ρ, and surface area, a, yields the thickness (θ) loss rate, = Δθ/Δt = Δm/(a ρ Δt). With SI units and ρ = 7.87 kg m If, for convenience, thickness is measured in mm, mass in mg, area in cm 2 , and time in yr, the thickness loss rate is (S11)

Corrosion current density
The rate (velocity) of the loss of metal amount (n, in mol) by corrosion is derived from mass (m) loss (see above) as = Δm/(M n v corr a Δt), with M a being the atomic mass. With n e as the number of electrons released per metal atom and the Faraday constant F, the current of electron loss is i corr = n e F = n n v corr e F Δm/(M a Δt). Division by surface area, a, yields the current density, (S12) If the corrosion rate is given as thickness (θ) loss instead of mass loss, the density, ρ, must be included: (S13) With SI units, n e = 2, M a = 55.85 · 10 −3 kg mol 1 (for iron) and ρ = 7.87 · 10 3 kg m −3 (mild steel EN 1.0330), and F = 96,485 C mol −1 , formulas (S12) and (S13) convert to i corr (A m 2 ) = 3.455 · 10 6 (A s kg ) (S17)

Crust conductivity and redox potential difference
Electrical current, I, through a cylindrical or cubical body of ohmic behavior is proportional to the applied voltage, V, and area perpendicular to current direction, a, and inversely proportional to length, d, i.e. I =  V a / d. The proportionality constant,  , is the specific conductance or conductivity (A V 1 m 1 ; S m 1 ,  1 m 1 ). The (argument of the vectorial) electrical field strength sustaining the current I is thus With For a corrosion current of i corr = 0.61 A m −2 (calculated from loss of Fe 0 ) and the determined conductivity of σ = 50 S m −1 , the field strength between metal and colonized crust would be only which is as low as 1.2 · 10 4 V across a crust with a realistic thickness of 1 cm. The voltage V is the difference between the operational redox potentials of the electronaccepting sulfate reduction (SR) and the electron-donating iron dissolution (FeDiss), V = Δ =  SR   FeDiss . Due to reaction overpotentials, Δ must be within the difference of the equilibrium redox potentials (E), i.e. V = Δ < ΔE = E SR  E FeDiss . E SR and E FeDiss are calculated from the half reactions and free energies of formation 1 (Thauer et al., 1977;Garrels and Christ, 1985), first for standard conditions at pH = 7 (E°', here used for E°p H7 ), and subsequently for near-real conditions assuming activities (in seawater) of {SO 4 2− } = 3 · 10 3 and {HCO 3 − } = 1.5 · 10 2 [from applied concentrations and activity coefficients (Stumm and Morgan, 1996)] and pH = 8.
The resulting ΔE = −0.35 V gives more than sufficient leeway for a 'self-adjusting' Δ during (the irreversible) electron withdrawal by corrosive SRB.

Acidity of Fe 2+
The pK a of an acid, viz. a reaction (acid ⇄ base + proton) with equilibrium constant K eq (also termed K a ) is ΔG°→ is the standard free energy of the forward reaction. Hydrated Fe 2+ ions tend to release a proton from the water shell according to which in thermodynamic data compilations and for calculations is usually simplified as Free energy with revised ΔG f°-value (Rickard and Luther, 2007) for Fe 2+ of −90.5 kJ mol 1 is ΔG°→ = +50.4 kJ mol 1 , yielding pK a = 8.8 (T = 298 K). Free energy with a formerly common ΔG f°-value for Fe 2+ of −78.87 kJ mol −1 yields ΔG°→ = +38.8 kJ and pK a = 6.6 (T = 298 K).
It is possible that also the ΔG f°-value for Fe(OH) + (−277.3 kJ mol 1 ) requires revision (origin of present value not investigated). Hence, the pK a -values calculated here may be regarded only as upper and lower limits of a range into which a fully revised pK a -value will fall.

Content of formed biomass in precipitated corrosion products
The fraction of biomass in the precipitated corrosion products is expressed as the quotient with m Bio indicating the biomass and m Min the mineral mass (superscript 'm' indicates mass ratio rather than molar ratio in other quotients). Because there is presently no convenient analytical method, the quotient is calculated from the amounts of iron oxidized and sulfate reduced. The mineral mass, m Min , is that of precipitated FeS and FeCO 3 , and possibly coprecipitated alkaline earth (Mg + Ca, here Ae) carbonates, AeCO 3 , i.e. m Min = m FeS + m FeCO3 + m AeCO3 . This is expressed via molecular masses (M) and amounts (n, mol) as m Min = M FeS n FeS + M FeCO 3 n FeCO3 + m AeCO3 . Because FeS precipitation scavenges all formed sulfide, n FeS = n SR, the amount of sulfate reduced. Assuming that all ferrous iron formed during EMIC and not precipitated as FeS is precipitated as FeCO 3 , the amount of the latter is total iron loss by EMIC minus sulfidic iron, i.e. n FeCO3 = n FeEMIC -n FeS or n FeCO3 = n FeEMIC -n SR. This yields for the mineral mass m Min = M FeS n SR + M FeCO 3 (n FeEMIC -n SR ) + m AeCO 3 (S28) The biomass (e.g., in g) formed per amount (e.g., in mol) of iron used for (attributed to) the anabolism (biosynthesis) is calculated from the predicted (Eq. 11, 12) yield coefficient Y FeAnb (biomass per iron oxidized by the anabolism) and the amount of iron needed for the anabolism, m Bio = Y Anab n FeAnab . Because n FeAnab = n ΔFe(0)  4 n SR (Eq. 8), the biomass is With equations (S28) and (S29), the quotient in equation (S27) (11) and (12) respectively.

Simplified calculation of the contribution of direct to total anaerobic corrosion
Whereas CMIC by sulfide leads to FeS as the only product (Eq. 4), EMIC leads in addition to non-sulfidic ferrous iron which tends to precipitate as carbonate (Eq. 5). Hence, the ratio of FeS to total Fe(II) in a crust should, in principle, allow to calculate the contribution of EMIC to MIC (EMIC and CMIC), the total corrosion due to the activity of SRB in the environment. Still, this is a rather formal treatment that does not consider higher levels of complexity, e.g. secondary (simultaneous or subsequent) reaction of sulfide from organotrophic SRB not only with the metal, but also with ferrous carbonate from EMIC (Fe 0  H 2 S, HS   FeCO 3 ), thus increasing the proportion of FeS (FeCO 3 + H 2 S → FeS + HCO 3  + H + ). Hence, EMIC could be the real (yet 'masked') cause of corrosion even if all ferrous iron in the crust is present as FeS.
The contribution (mol/mol) of EMIC to MIC is formally expressed as the quotient of the amount of iron corroded by EMIC, n FeEMIC , to the amount corroded by MIC, n FeMIC , the latter being identical with the measurable ferrous iron formed, n Fe(II) , so that An expression must be found which besides n Fe(II) includes the other measurable amount, n FeS , but no longer the not directly obvious n FeEMIC . This is achieved by including four other equations. Two of these are equations (7) and (9), respectively. Equation (8) cannot be applied, because MIC includes also organic electron donors in addition to Fe 0 for sulfate reduction, so that n FeCatab < 4 n SR . Furthermore, the characteristic product of EMIC is non-sulfidic iron, its amount being designated n FeNonS . This is ¾ of the amount oxidized by the catabolism (Eq. 5) plus the amount resulting from biosynthesis, i.e.
Finally, the amount n FeNonS is iron totally formed by MIC (EMIC + CMIC), n Fe(II) , less sulfidic iron, n FeS , so that For convenience, we arrange the coefficients and 'parameters' (q EMIC , q Anab ) of equations (S35)(S39) [order of rows below] by n Fe(II) , n FeS , n FeEMIC , n FeCatab , n FeAnab , n FeNonS as Because the two measurable 'variable' amounts, n Fe(II) and n FeS , are finally of interest, the matrix is converted .
(S41) . Kinetic aspects of the abiotic reaction of iron in circumneutral water, and direct (lithotrophic) iron corrosion by SRB. Availability of H + -ions at the metal surface and combination of adsorbed H-atoms to adsorbed H 2 are assumed to be rate-controlling steps ('bottle necks'), thus also controlling liberation of H 2 into water (Bockris and Reddy, 1970;Hamann et al., 2007)  A. Production of 'cathodic' hydrogen by reduction of H + ions (Fig. S1), and sulfide that could be formed by H 2 utilization by SRB (4 H 2 + SO 4 2 + 2 H +  H 2 S + 4 H 2 O). B. Original iron specimen (day 0), specimen with precipitate after 5 months (original) and after removal of precipitate (using HCl-hexamine). Fig. S4. Insensitivity of non-corrosive control strain HS3 towards Fe 2+ . Addition of H 2 to the culture including iron granules leads to rapid sulfide production (measured as sulfate loss).

Fig. S5.
Electro-technical scheme with approximate voltage drops of the split-coupon incubation device for conductance measurement of the biogenic crust formed on corroding iron. The device circumvents interference by the noticeable contact resistance between the iron wire and the iron coupon inside the incubated bottle (Fig. 3A). The plot in the lower part depicts the voltage drop along current flow. The outer voltage (V o ) is supplied and adjusted such that the voltage across the split (V s ) is kept at 0.20 V while the current (I) is being measured. The adjusted low voltage for measurement avoids electrolysis. Measurement of V s is carried out with a high-resistance voltmeter. V c1 and V c4 are the voltage drops due to contact resistance between the iron wire and the iron coupon (around 1 Ω), and V c2 and V c3 the arbitrarily assumed voltage drops due to the contact resistance between iron and the sulfidic crust. Voltage drop along the iron wire and the iron coupons is negligible (resistance by two and four orders or magnitude lower, respectively, than resistance of wire-coupon contact and the crust).      S10. Sulfide production (determined as sulfate consumption) and decrease of dissolved ferrous iron due to carbonate precipitation in long-term incubations of corrosive SRB. Strain IS4 (A) which was more alkali-tolerant than strain IS5 (B) grew up to higher pH [pH increase due to equation (5)] thus promoting precipitation according to Fe 2+ + HO  + HCO 3   FeCO 3 + H 2 O. This favored formation of micro-chimneys (Fig. 5C). Six cultures of each strain were incubated in parallel and sacrificed at different time points for SEM analysis (Fig. 4, Figs S6 to 8). Formation of crater-and chimney-like structures in cultures of strain IS4 coincided with the drop of [Fe 2+ (aq) ] below detection limit (0.2 mg/l). The initial pH was 7.3. Strain IS4 reached pH ≈ 9. Activity of strain IS5 ceased at pH ≈ 8.  Table S1. Compilation of corrosion rates recorded for (anoxic) natural and engineered environments, and for laboratory cultures of sulfate-reducing bacteria. Table S2. Vitamins in used media. Table S3. Conductivity values measured in an incubation device with split coupon with corrosive cultures of strains IS4 and IS5, and with sterile artificial seawater. Iron is provided as the sole source of electrons. Table S4. Electrical conductivity of selected substances.