Slag–Refractory Reactions: Why Basicity Alone Cannot Predict Lining Life
The Core Decision: Compatibility Is a Reaction Path, Not a Family Label
A refractory can remain below its published temperature limit and still lose thickness rapidly. The decisive question is not simply whether the lining is called acidic, basic, or neutral. It is whether the process medium can wet the surface, enter connected pores or joints, dissolve a vulnerable phase, create a new liquid or solid phase, and then remove the altered material faster than a protective boundary can develop. That sequence is the practical meaning of slag refractory interaction.
Temperature matters because it controls reaction rate, viscosity, diffusion, phase stability, and thermal stress. But temperature alone does not identify the reaction partners or predict what they will form. Two liquids at the same temperature may impose radically different corrosion rates because their oxide activities, redox state, saturation, viscosity, wetting behavior, and flow conditions differ.
The broader service-envelope method explained in our guide to refractory material selection establishes what operating information must be collected. This article goes one level deeper. It shows how to convert that information into a reaction-path diagnosis, design a meaningful corrosion test, interpret the altered cross-section, and decide whether a proposed refractory is chemically defensible.
Compatibility rule: do not approve a refractory because its bulk chemistry appears compatible. Approve it only when the expected reaction path, reaction rate, altered-layer behavior, and field consequence are acceptable.
The Five-Stage Attack Sequence
Refractory corrosion is often described as if it were a single event. In practice, it is a chain. Breaking any critical link can extend lining life; overlooking one can make a laboratory ranking irrelevant.
Stage 1: Delivery and Wetting
The corrosive medium must first reach the surface. Delivery may occur through a static bath, turbulent slag, droplets, splash, vapor, condensate, entrained dust, metal flow, flame impingement, or a deposit that becomes liquid during an upset. Contact frequency and renewal rate matter because fresh liquid can maintain a chemical driving force that a stagnant test does not reproduce.
Wetting determines whether a liquid spreads, beads, or enters fine surface defects. It is influenced by interfacial energy, atmosphere, temperature, surface roughness, carbon condition, additives, and the chemistry of both phases. A low apparent porosity does not guarantee protection if cracks, joints, or strongly wetted pathways provide rapid access.
Stage 2: Infiltration and Penetration
Once contact is established, liquid can move through open pores, matrix channels, microcracks, brick joints, anchor interfaces, or poorly consolidated zones. Capillary pressure, pore radius, permeability, contact angle, liquid viscosity, pressure, and exposure time influence the depth and rate. A useful measure of refractory penetration resistance must therefore distinguish total open porosity from the connectivity and size distribution of the actual transport network.
Penetration can be damaging even when gross surface dissolution appears small. The infiltrated zone may densify, embrittle, expand, shrink, or develop a thermal expansion coefficient different from the unaltered core. During cooling, the penetrant may crystallize or form new phases. Repeated cycles then turn a chemical gradient into a mechanical spalling plane.
Stage 3: Dissolution at the Interface
At the contact surface, refractory components may dissolve into the liquid. The driving force depends on chemical activity and the distance from saturation, not simply the reported weight percentage of an oxide. The rate also depends on temperature, boundary-layer thickness, flow, diffusion, surface area, and whether reaction products slow or accelerate transport.
A defensible refractory dissolution mechanism identifies which phase is dissolving first. In a castable, the fine matrix can disappear while coarse aggregate remains temporarily intact. In a brick, the bonding phase or grain boundary can be more vulnerable than the principal aggregate. A high headline percentage of alumina or magnesia cannot reveal that sequence.
Stage 4: Reaction, Precipitation, and Phase Transformation
Dissolved species do not always remain in the liquid. They may precipitate at the interface, enter a solid solution, or react with refractory minerals to create new phases. Some products raise viscosity, block pores, or form a stable barrier. Others create low-melting liquids, volume change, weak zones, or brittle layers.
The resulting refractory reaction layer should be judged by continuity, adhesion, thickness, permeability, melting behavior, thermal expansion, and stability during cycling. Calling any reaction layer “protective” without testing these properties is unsafe. A dense layer that slows attack during steady operation may detach during shutdown and expose a fresh surface at the next start.
Stage 5: Removal of the Altered Material
Corrosion becomes visible wear when dissolved material is transported away or when the altered layer is removed by flow, abrasion, impact, thermal stress, structural movement, or gas evolution. This is why a static cup test can understate damage in a high-turbulence zone. The test may allow a saturated boundary layer to remain in place, while the real process continuously strips it away.
Conversely, a test can overstate field attack if it forces unrealistically severe contact, uses excessive liquid volume, excludes a protective process coating, or exposes every sample face rather than the intended hot face. The test geometry must represent the hypothesized attack sequence rather than merely produce dramatic corrosion.
Why One Basicity Number Cannot Predict Lining Life
In steel and other oxide-slag systems, buyers often receive a single basicity value, commonly expressed as CaO/SiO₂. The slag basicity ratio is useful for initial classification, but it is not a complete compatibility model. It says nothing by itself about the concentration and oxidation state of iron, MgO saturation, alumina, manganese oxide, alkalis, fluorides, solids, viscosity, residence time, or liquid renewal.
| Variable beyond simple basicity | Why it changes the reaction path | Evidence to request |
|---|---|---|
| FeO, Fe₂O₃, and oxygen potential | They can change oxidation of carbon-containing linings, dissolution behavior, liquid-phase formation, and solid-solution chemistry. | Oxide analysis linked to process stage, oxygen practice, and sampling time—not one annual average. |
| MgO content and saturation | A liquid closer to saturation with MgO may have less driving force to dissolve a magnesia-based lining, although viscosity and process consequences must also be considered. | Temperature-specific saturation assessment, mineral phases, and actual slag samples. |
| Al₂O₃ and spinel-forming reactions | Alumina can change liquidus behavior and viscosity; spinel formation or solid-solution uptake may alter penetration and reaction-layer stability. | Complete chemistry, phase-equilibrium assessment, and cross-sectional mineralogy. |
| MnO and other low-viscosity contributors | These species can enter refractory phases, change slag viscosity, and increase transport through penetrated zones. | Traceable analysis by heat or campaign stage and a test slag that preserves the relevant concentration. |
| Fluorides, alkalis, borates, and other fluxes | Small additions may lower melting temperature, increase wetting, destabilize bonding phases, or create highly mobile liquids. | Detection limits low enough to capture minor but influential constituents and upset concentrations. |
| Temperature and solids fraction | A nominal composition can behave as a viscous multiphase mixture at one temperature and a mobile liquid at another. | Normal, peak, and local interface temperatures plus liquidus or phase-fraction calculations. |
| Viscosity, flow, and renewal rate | Flow controls boundary-layer thickness and removal of dissolved species or protective precipitates. | Velocity, stirring, gas injection, rotation, bath depth, residence time, and turbulence by zone. |
| Carryover and process sequence | Mixing two slags or introducing residual flux can create a short-lived composition more aggressive than either normal stream. | Startup, transition, carryover, cleaning, and upset samples in addition to steady-state data. |
Weight percentage also differs from chemical activity. Two slags with similar laboratory reports may contain different mineral or liquid phases and therefore different effective reaction potentials. Redox state can shift the valence of multivalent elements. Temperature can move the system across a liquidus boundary. Solid particles can either increase apparent viscosity or mechanically abrade a weakened surface.
The practical response is not to abandon basicity. It is to treat basicity as the first coordinate in a multidimensional exposure map. Compatibility should be assessed over the normal range, credible extremes, and transitional mixtures. The event that controls campaign life may occupy only a small part of operating time.
Three Forensic Files: Similar Temperature, Different Reaction Path
File A: A Steel Slag Line with a Decarburized Reaction Front
Suppose a magnesia-carbon lining performs well during stable production but loses rapidly after a change in oxygen practice and slag carryover. The peak temperature is unchanged. A simple explanation such as “the slag became more basic” may be wrong or incomplete.
The investigation should examine FeO and oxygen potential, carbon oxidation, MgO saturation, liquid volume, viscosity, stirring, arc or burner hot spots, and the thickness of the decarburized zone. Once carbon is removed from the hot face, slag can enter more readily, contact oxide grains, and change the local reaction path. A weakened matrix can then be removed by turbulence or thermal cycling.
Macroscopic wear depth alone cannot distinguish oxidation-first damage from direct dissolution. Sectioning should map residual carbon, infiltrated chemistry, cracks, and newly formed phases from the hot face into the unaltered core. If the decarburized depth correlates with penetration and wear, improving carbon protection or oxygen control may create more value than increasing nominal MgO content.
File B: Glass Contact Where Product Quality Fails Before Thickness
In a glass-melting furnace, an AZS or zircon-based refractory can interact with alkali-bearing molten glass through diffusion, glassy-phase exudation, partial dissolution, pore formation, and transport of refractory-derived species into the melt. The operational consequence may appear as stones, cords, blisters, seeds, color variation, or other defects before the lining reaches a structural minimum thickness.
Here, compatibility cannot be reduced to corrosion depth. The chemistry and morphology of released particles matter, as do boundary-layer flow, temperature, glass composition, joint orientation, electrical boosting, and the location of the material relative to the glass-quality stream. A material with slower bulk loss can still be unacceptable if its corrosion products create more damaging defects.
Testing should reproduce the intended glass, temperature, contact geometry, and duration. The investigation should record recession, blistering, exudation, penetration, interface composition, and defect potential. An isothermal test may rank candidates under stable conditions, but field validation must include thermal gradients and flow.
File C: Aluminum Penetration and Internal Corundum Growth
In aluminum melting and holding equipment, molten metal and alloying elements can penetrate a refractory and react with silica-bearing or other susceptible phases. Fluxes, magnesium content, oxygen potential, bath turbulence, cleaning tools, and local temperature influence the result. The visible hot face may appear serviceable while reactions develop beneath it.
Internal oxidation-reduction reactions can produce alumina-rich growth, porosity, and stress. As the altered region expands, it can displace lining material, reduce vessel capacity, open cracks, or create areas that are difficult to clean. The controlling step may change with temperature: reaction chemistry can dominate in one range while metal penetration becomes rate-controlling in another.
Non-wetting additives and dense matrices may reduce initial entry, but they must be qualified with the actual alloy, flux, temperature, and atmosphere. A test using pure aluminum cannot automatically validate performance in a magnesium-bearing recycling operation with salt flux and frequent mechanical cleaning.
Microstructure Decides Which Route the Liquid Can Take
Bulk oxide chemistry identifies available reaction partners; microstructure determines how quickly they meet. This is why two products with similar data-sheet chemistry can show different corrosion behavior.
Porosity Is Not the Same as Permeability
Apparent porosity reports the volume of accessible pores under a defined test condition. Permeability describes how readily fluid can pass through a connected network. A material can have modest apparent porosity but contain connected cracks or coarse channels that permit rapid infiltration. Another can show a similar porosity value with fine, tortuous, poorly connected pores that slow transport.
Field processing changes this network. Excess castable water, inadequate vibration, rebound in a gunned layer, poor curing, explosive-spalling damage, brick joints, and thermal cracks can dominate the installed transport path. Laboratory specimens made under ideal conditions may therefore overstate refractory penetration resistance.
The Matrix Usually Reacts Before the Aggregate Skeleton
Fine particles, binders, impurities, and intergranular phases provide large surface area and continuous pathways. They may dissolve or react before dense coarse grains. As matrix support disappears, aggregates can be released mechanically even if the grains themselves remain chemically stable.
For this reason, a supplier should explain not only aggregate purity but also matrix composition, binder development, grain-size distribution, firing or sintering behavior, and expected phases at service temperature. A premium aggregate cannot compensate indefinitely for an unstable continuous matrix.
Protective Reactions Must Remain Protective Through Cycling
Some engineered systems use spinel formation, calcium hexaluminate, pore-blocking precipitates, viscosity increase, or controlled surface densification to slow attack. These mechanisms can be valuable, but they create their own design constraints. Excessive expansion may crack the lining. A reaction product may be stable at operating temperature but transform or detach during cooling.
The approval criterion should therefore include heat-up, steady exposure, cooling, and reheating. The best result is not automatically the thickest reaction zone. It is a thin or controlled altered zone that remains adherent, limits transport, avoids damaging stress, and does not contaminate the process.
Build the Corrosion Test Around a Specific Hypothesis
Useful refractory corrosion testing begins with a written hypothesis. Examples include: “high-FeO slag oxidizes the carbon-containing matrix before dissolution,” “alkali vapor condenses in the cooler superstructure and attacks the bond,” or “metal penetration creates the layer that later spalls during cycling.” The method must be capable of confirming or rejecting the proposed sequence.
| Method or evidence tool | Question it can answer | Important blind spot |
|---|---|---|
| Static cup or crucible test | Relative reaction, penetration, phase formation, and containment of a selected liquid under controlled temperature and time. | Limited liquid renewal and flow can allow saturation or protective layers not present in service. |
| Rotary slag test | Comparative corrosion under renewed dynamic contact; ASTM C874-20(2026) provides a current standardized method. | Rotation, sample geometry, slag volume, atmosphere, and thermal profile may still differ from the plant. |
| Rotary finger or immersion test | Flux-line attack, dynamic penetration, and comparison of several specimens in a common corrosive medium. | Results depend strongly on specimen placement, rotation, bath control, and interpretation of maximum wear depth. |
| Isothermal molten-glass test | Corrosion resistance and interface behavior against a defined glass; ASTM C621-09(2022) is relevant. | It may not represent furnace flow, joints, gradients, volatilization, or long campaign duration. |
| Vapor-attack test | Response of superstructure materials to volatile species; ASTM C987-10(2023) addresses furnace-superstructure vapor attack. | Plant condensation, mixed vapor chemistry, deposit cycling, and leakage can be difficult to reproduce. |
| Oxidation-resistance test | Stability of susceptible phases in a specified atmosphere; ASTM C863-00(2022) addresses silicon carbide refractories. | Oxidation alone may not reproduce simultaneous slag, metal, or mechanical exposure. |
| Thermodynamic and phase-equilibrium calculation | Possible stable phases, saturation, liquid fraction, and chemical driving forces over composition and temperature. | Equilibrium does not predict reaction rate, transport, wetting, crack formation, or installation defects. |
| Controlled field panel | Installed performance under real process variability, gradients, maintenance, and flow. | Poorly matched zones, changing operations, or missing baseline measurements can invalidate the comparison. |
Avoid naming a test without defining its conditions. Report sample orientation, dimensions, preparation, pre-firing, exposed faces, liquid chemistry, liquid-to-refractory ratio, atmosphere, temperature calibration, heating rate, dwell time, cycling, agitation, refresh schedule, cooling method, repeats, and measurement procedure.
Specifications should also be checked for obsolete references. ASTM currently lists C768-99, the drip-slag practice, as withdrawn in 2004. A company may still use a useful internal method derived from it, but the procurement document should call it an internal procedure and provide the complete method rather than present it as a current ASTM requirement.
Severity does not equal relevance. A test that destroys every candidate may show only that the condition was extreme. A test that causes no measurable change may lack resolution. The objective is to reproduce the controlling reaction path with enough differentiation to support a decision.
Read the Cross-Section, Not Just the Final Shape
A photograph of a worn sample is not a complete corrosion result. Maximum recession can rank gross loss, but it cannot identify whether the cause was dissolution, penetration, oxidation, phase transformation, cracking, or mechanical removal. Reliable refractory corrosion analysis combines dimensional measurement with a mapped cross-section.
Separate the Zones Before Taking an Average
A tested or used refractory may contain a surface deposit, reaction or precipitation zone, infiltrated zone, decarburized or oxidized zone, cracked transition, and unaltered core. Averaging their chemistry can erase the gradient that explains the failure.
Mark distance from the original hot face and preserve orientation. Take samples from the surface inward at defined intervals. Compare the attacked area with an unexposed reference from the same lot where possible. Record whether cracks existed before sectioning and whether loose material was lost during removal.
Use Each Analytical Technique for the Question It Can Answer
- Optical microscopy: maps cracks, penetration fronts, aggregate release, reaction zones, and macroscopic texture.
- SEM with EDS: examines local morphology and elemental distribution across grains, matrix, interfaces, and penetrated areas.
- X-ray diffraction: identifies crystalline phases, while recognizing that amorphous or minor phases may require additional methods.
- Electron-probe or elemental mapping: provides higher-resolution concentration profiles where diffusion and reaction fronts matter.
- Carbon analysis: helps define oxidation or decarburization in carbon-containing products.
- Porosity and permeability measurements: compare transport structure before and after exposure, provided specimen condition is controlled.
- Thermal and mechanical testing: evaluates whether the altered zone has become weak, brittle, expanded, or unstable through cycling.
The purpose of refractory microstructure analysis is not to create attractive microscope images. It is to connect material structure to a causal sequence: which phase changed first, how the liquid traveled, what new phases formed, where cracks initiated, and why the altered material was removed.
Convert Observations into a Cause-and-Evidence Statement
A useful conclusion reads like this: “The liquid entered connected matrix pores; Fe- and Mn-bearing species formed a low-viscosity infiltrated zone; the bond dissolved preferentially; differential expansion generated a crack parallel to the hot face; and flow removed the detached layer.” Each clause should be supported by a measurement or mapped observation.
An unhelpful conclusion reads: “The refractory was incompatible.” It identifies neither the cause nor the corrective action. Good post-mortem refractory analysis should separate the initiating mechanism from the final visible damage and state what evidence would falsify the proposed explanation.
Create a Compatibility Evidence Card for Every Candidate
A one-page evidence card helps engineering, procurement, operations, and suppliers compare the same variables. It prevents a product data sheet from being mistaken for an application approval.
| Evidence-card field | Required content |
|---|---|
| Candidate identity | Grade, formulation family, manufacturing location, lot, production date, storage condition, and approved revision. |
| Service zone | Exact equipment location, lining construction, exposure faces, thickness, joints, anchors, and adjacent materials. |
| Contacting media | Typical and extreme liquid, gas, vapor, metal, slag, ash, flux, deposit, and cleaning chemistry with sampling context. |
| Thermal and redox envelope | Normal, peak, upset, and cycling temperatures; atmosphere; oxygen potential; pressure; and transitions. |
| Proposed reaction path | Delivery, wetting, penetration, first phase attacked, reaction products, altered-layer behavior, and removal mechanism. |
| Test method | Complete procedure, reference standard where applicable, deviations, specimen preparation, repeats, and controls. |
| Measured outcomes | Recession, penetration depth, altered-layer thickness, cracks, mass change, phase formation, contamination, and uncertainty. |
| Field correlation | Comparable references, operating differences, trial-zone design, baseline, inspection method, and acceptance criteria. |
| Limitations | Conditions not represented, unresolved mechanisms, material sensitivities, and assumptions requiring monitoring. |
| Change-control triggers | Formulation, raw-material source, manufacturing site, process feed, fuel, flux, operating practice, or lining-design changes that require review. |
The card should be version-controlled. If a supplier changes an aggregate source, binder, antioxidant, fine-matrix component, firing schedule, or manufacturing site, the reaction path may change even when the commercial grade name remains unchanged. Likewise, a plant change in scrap mix, raw material, oxygen practice, fuel, flux, or recycling rate can invalidate a historical approval.
Turn Evidence into a Purchase or Trial Decision
| Decision status | Technical condition | Required action |
|---|---|---|
| Reject | Unacceptable low-melting reaction, rapid penetration, destructive oxidation or reduction, contamination, unstable altered layer, or no credible evidence for the actual exposure. | Remove from the shortlist or redesign the process and lining zone before reconsideration. |
| Conditionally acceptable | Reaction path appears manageable only inside a defined temperature, chemistry, atmosphere, or operating window. | Write those limits into operating control, monitoring, alarms, and supplier documentation. |
| Controlled field trial | Laboratory evidence is favorable but scale, cycling, installation, flow, or process variability remains uncertain. | Use matched zones, baseline dimensions, traceable lots, defined stop rules, and scheduled inspections. |
| Approved | Laboratory mechanism, installed quality, field performance, and post-service evidence agree within the defined envelope. | Maintain change control and continue monitoring variables that govern the reaction path. |
Purchase price should be compared only after incompatible candidates are removed. A lower-cost product should not compensate for a destructive reaction path. Conversely, the most expensive or highest-purity material is not automatically the best if it forms an unstable interface, contaminates the product, or cannot tolerate the installed thermal cycle.
Field-trial acceptance criteria should include more than total campaign life. Measure local wear rate, penetration or reaction depth where samples are available, repair consumption, shell temperature, product quality, operating variability, and the condition of adjacent joints or anchors. A longer campaign caused by mild production cannot validate a material.
Process Control Can Change Compatibility Without Changing the Refractory
Material selection and process operation are coupled. Changing slag saturation, oxygen potential, flux addition, bath mixing, carryover, temperature, residence time, or cleaning practice can alter molten slag attack even when the installed lining remains the same.
This creates an important management question: is the proposed protective mechanism under material control, process control, or both? If resistance depends on maintaining MgO saturation, a stable coating, limited FeO, or a specific atmosphere, the responsible operating variables must be measured and governed. A process-dependent protection mechanism cannot remain an unwritten assumption in the refractory specification.
Optimization also has boundaries. Adjusting slag chemistry to protect the lining can affect metal refining, viscosity, energy, yield, emissions, downstream processing, and waste. The correct decision minimizes total process risk rather than maximizing refractory life in isolation.
Focused FAQ
Does a chemically neutral refractory resist every slag?
No. “Neutral” is a broad classification, not a claim of inertness. Alumina, carbon, silicon carbide, chromia, and related systems can still undergo dissolution, oxidation, reduction, penetration, phase transformation, or product-contamination reactions under specific conditions.
Is a higher CaO/SiO₂ ratio always safer for a basic refractory?
No. The ratio does not capture FeO, MgO saturation, alumina, manganese oxide, fluorides, alkalis, liquid fraction, temperature, viscosity, redox state, or flow. Those variables can change attack even when the ratio remains constant.
What is the difference between corrosion and penetration?
Corrosion involves chemical alteration or dissolution of the refractory. Penetration is movement of liquid or vapor-derived species into pores, cracks, or joints. Penetration may precede corrosion and can create a mechanically unstable zone even when surface recession is limited.
Can lower apparent porosity guarantee better slag resistance?
No. It may help, but connected pore structure, permeability, cracks, joints, wetting, matrix chemistry, installation water, and thermal damage also control liquid entry. The installed lining can behave differently from an ideally prepared laboratory specimen.
Which corrosion test is best?
There is no universal best method. A static test may suit initial chemical screening; a rotary or immersion test may better represent renewed dynamic contact; a vapor test addresses superstructure exposure; and a field panel captures installation and operating variability. The correct method tests the proposed failure mechanism.
Why should a tested sample be sectioned?
The exterior shows final wear but not the sequence beneath it. A cross-section can reveal penetration depth, decarburization, reaction products, dissolved matrix, cracks, pore blocking, and the transition to unaltered material. These observations identify whether a candidate failed for the reason originally predicted.
Can thermodynamic software replace a corrosion test?
No. It can identify possible phases, saturation, liquid fraction, and chemical driving force under defined assumptions. It does not directly predict kinetics, wetting, transport, flow, crack formation, installation defects, or removal of an altered layer. It is a screening and interpretation tool.
When should a refractory compatibility approval be repeated?
Repeat or review the approval after meaningful changes in formulation, raw-material source, manufacturing site, installation method, feed, fuel, alloy, flux, slag practice, redox condition, operating temperature, cycling, or failure consequence. A stable grade name does not prove an unchanged reaction path.
What should a buyer ask a refractory supplier to provide?
Ask for the proposed reaction mechanism, complete test conditions, raw measurements, sample sections, relevant phase or microstructure evidence, repeatability, comparable field references, limitations, installation controls, and formal notification of material changes. A single corrosion photograph or percentage improvement is not enough.
Industry-Level Conclusion
Refractory compatibility is not a temperature rating or a three-family classification. It is the outcome of delivery, wetting, penetration, dissolution, reaction, phase formation, and removal under a specific thermal, chemical, mechanical, and operational envelope. The controlling link can change by zone and over the course of a campaign.
High-quality decisions begin with complete process samples and a written reaction hypothesis. They use tests that reproduce the relevant transport and atmosphere, then examine the full altered cross-section rather than only the final worn shape. Laboratory evidence, field trials, operating records, and removed-lining analysis must tell the same causal story.
When buyers adopt this standard of evidence, they stop asking whether a refractory is broadly “acidic,” “basic,” or “neutral” and start asking a more useful question: what will this material become after months of contact with our real process medium? That question leads to a measurable compatibility decision, clearer supplier accountability, and a more defensible estimate of refractory service life.
#SlagRefractoryInteraction #RefractoryCorrosion #SlagChemistry #RefractoryTesting #RefractoryMicrostructure #MoltenSlag #LiningLife #PostMortemAnalysis #SteelmakingRefractories #GlassFurnaceRefractories #AluminumFurnace #IndustrialMaterials #FailureAnalysis #MaterialsEngineering
