Volcanic cement blends face their toughest test in Algeria’s salt flats

Reinforced mortars made with volcanic supplementary cementitious materials have spent 360 days soaking in some of the harshest waters a civil engineer could choose, and the results overturn a comfortable assumption: the exposure chemistry, not the binder recipe, decides how fast the steel inside corrodes. In a new study, ten mortar formulations — an ordinary Portland cement reference plus nine blends containing raw perlite, natural pozzolan, or balanced combinations of the two — were monitored electrochemically across five media of escalating aggressiveness. Across every indicator measured, the same severity ranking emerged: freshwater was mildest, followed by seawater, then a sulfate solution, then a concentrated sodium chloride–magnesium chloride brine, with natural Sebkha water from the Grande Sebkha of Oran in Algeria proving the most aggressive of all.

The statistical verdict was unambiguous. A descriptive variance partition of the electrochemical records attributed between 61.9 and 73.1 percent of the variability in corrosion potential and corrosion current density to the exposure medium, while formulation accounted for only 5.4 to 5.9 percent. A formal two-way analysis of variance on the terminal apparent mass change of the complete specimens told the same story even more starkly: the medium explained 91.5 percent of the variance, the mortar formulation just 6.4 percent, with a small but significant interaction between the two. In other words, choosing the right volcanic blend matters — but where you build matters roughly ten times more.

The Sebkha result is the study’s most counterintuitive finding. The natural water, collected at Es Senia, contained 9.9 grams of dissolved solids per liter, was roughly 97 percent sodium chloride by ionic load, and carried only 10 milligrams per liter of magnesium and 97 milligrams per liter of sulfate — modest figures compared with seawater or the laboratory brine, whose chloride load reached about 52.7 grams per liter. Yet the Sebkha water drove the highest corrosion currents and the largest apparent mass losses. The authors propose a physically consistent explanation rooted in oxygen: in concentrated brines, dissolved oxygen solubility and cathodic oxygen transport fall, throttling the corrosion reaction itself, whereas moderately saline, sodium-chloride-dominated water sits near the classical salinity range where corrosion rates peak before oxygen limitation takes over. Because dissolved oxygen was not monitored, this mechanism remains a testable interpretation rather than a demonstrated cause.

The experimental campaign was unusually thorough for a mortar-scale screening study. Three replicate reinforced cylinders — each 50 millimeters in diameter and 100 millimeters long, with a centrally embedded ribbed FeE500 steel bar leaving 50 millimeters of exposed length — were tested for every formulation–medium combination, generating 1,800 electrochemical records over twelve 30-day cycles. Half-cell potential and corrosion current density were tracked throughout, and the whole specimens were repeatedly weighed without stripping corrosion products, yielding an apparent mass-change record that closely mirrored the Faradaic estimates computed from the currents: the parity check between the two independent scales showed a mean error of just 1.42 percent at the final cycle.

Those Faradaic numbers put the corrosion in context. Mean corrosion current densities at cycle 12 ranged from 8.16 to 14.11 microamperes per square centimeter across the five media, corresponding to penetration rates of roughly 0.095 to 0.164 millimeters per year. Even the freshwater exposure, nominally the gentlest, sustained mean currents near 6.6 microamperes per square centimeter over the monitoring window — evidence that the campaign captured active propagation rather than the passive initiation stage, since the specimens had already entered a propagation regime at the first measurement point under the accelerated immersion protocol.

On the materials side, both volcanic additions earned their pozzolanic credentials. X-ray diffraction showed perlite dominated by a broad glassy aluminosilicate hump between 20 and 35 degrees two-theta, while the natural pozzolan displayed sharper, more crystalline reflections from anorthoclase and quartz. The standardized Frattini test placed all twelve determinations for both materials firmly in the pozzolanic domain, with mean calcium depletion relative to the lime-saturation boundary of 51.9 percent for perlite and 37.1 percent for the pozzolan. Triplicate water-sorption tests on the raw powders added a distinctive descriptor: after 72 hours of immersion, perlite had absorbed 13.58 percent of its dry mass in water and the pozzolan 12.75 percent, and both retained more than 90 percent of that water at 97 percent relative humidity, releasing roughly half by 80 percent relative humidity — behavior the authors interpret as a possible moisture-buffering contribution rather than direct proof of transport benefits in hardened mortar.

Formulation effects did exist, but they were exposure-dependent rather than universal. At equal total replacement levels, the balanced binary perlite–pozzolan blends consistently showed lower terminal apparent mass change than the corresponding single-SCM mortars, and Tukey comparisons confirmed these differences statistically. The pozzolan-rich formulations fared worst in the harshest media: M20PZ and M30PZ spent 50 to 75 percent of the twelve monitoring cycles in the active corrosion zone — half-cell potentials below minus 350 millivolts, indicating more than 90 percent corrosion probability under the ASTM C876 classification — when exposed to the two saline end-members. Perlite-rich mortars, meanwhile, showed the highest accessible porosity, yet porosity alone did not rank the corrosion response, reinforcing the authors’ argument that no single material indicator can stand in for reinforcement durability.

The mechanistic picture the authors assemble is one of coupled, multi-ion attack. Chloride acts mainly at the steel–mortar interface, locally destabilizing the passive film that the alkaline pore solution normally maintains. Magnesium ions work on the matrix itself, exchanging with calcium in calcium-silicate-hydrate to form non-cohesive magnesium–silicate–hydrate and brucite while depressing pore-solution pH and eroding the alkalinity that underpins passivity. The sulfate medium combined sodium and magnesium sulfate salts deliberately, exercising both the classical ettringite-and-gypsum pathway and the magnesium-driven decalcification pathway in one solution. Semi-quantitative energy-dispersive X-ray analysis of exposed specimens traced the chemical signature of this evolution, with local calcium-to-silicon ratios falling from 1.91 in the reference mortar to 1.55 in a balanced blend and 1.33 in the pozzolan-rich mix — illustrative local indicators, the authors caution, not bulk phase quantification.

To turn the monitoring data into something a maintenance engineer could use, the team fitted a power law of the form Δm = At^α to all 50 cumulative Faradaic mass-loss series. The exponent exceeded one in every series, indicating supra-linear acceleration of cumulative loss, with mean values ranging from 1.108 in Sebkha water to 1.224 in freshwater. Validated by a time-ordered holdout — calibrated on the first eight cycles, tested on the last four — the model predicted the withheld data with a mean error of 2.13 percent. Inverting the fitted law against a 2,000-milligram maintenance-alert threshold, equivalent to about 0.16 millimeters of average radial penetration or a 6 percent local cross-section reduction of the test bar, produced threshold times spanning 359 days for the pozzolan-rich M30PZ in Sebkha water to 756 days for the reference mortar in freshwater. The authors are explicit that this is a comparative early-warning ranking within a deliberately bounded two-year screening horizon, not a structural service-life prediction.

Indeed, the study’s own diagnostics argue against over-extending any single kinetic curve. When a two-segment power law with a fixed midpoint breakpoint was compared against the single global exponent, 45 of the 50 series favored the broken model, and the late-window exponent exceeded the early-window exponent in every medium — direct evidence that the apparent corrosion rate drifts even within a controlled, continuous-immersion campaign. Combined with the acknowledged limitations — no electrochemical impedance spectroscopy to separate solution resistance, apparent rather than chemically stripped mass measurements, a single Sebkha grab sample, and continuous immersion that cannot reproduce the evaporative concentration and wetting–drying cycling of real sabkha environments — the authors frame their formulation trends as testable hypotheses for prioritizing candidates, not universal rankings. The practical message for engineers designing repair mortars and cover materials for coastal, sulfate-bearing, or salt-flat-adjacent infrastructure is clear: volcanic blends should be screened in representative coupled media, because performance in freshwater or single-salt tests simply does not transfer to the chemically tangled waters where these materials will actually serve.

Subject of Research: Corrosion behavior of reinforced mortars containing volcanic supplementary cementitious materials under chloride, sulfate and natural Sebkha exposure

Article Title: Corrosion of reinforced mortars incorporating volcanic supplementary cementitious materials under chloride–sulfate and natural Sebkha exposure

Article References: Brahimi, S., Hamadache, M., Hifi, M., & Mouli, M. (2026). Corrosion of reinforced mortars incorporating volcanic supplementary cementitious materials under chloride–sulfate and natural Sebkha exposure. Case Studies in Construction Materials, 25, Article e06622. https://doi.org/10.1016/j.cscm.2026.e06622

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06622

Keywords: reinforced mortar, supplementary cementitious materials, natural pozzolan, perlite, steel corrosion, chloride ingress, sulfate attack, Sebkha water, electrochemical monitoring, power-law kinetics, concrete durability, Algeria