The limiting oxygen index (LOI) is one of the most widely used small-scale indicators of flammability for thermal insulation materials such as EPS, XPS, polyurethane foam, and phenolic foam. This article explains how LOI tests are conducted under ISO 4589-2, ASTM D2863, and GB/T 2406.2, how to interpret the resulting oxygen concentration values, and which material, specimen, and conditioning factors most strongly influence the outcome. The analysis below helps engineers, quality managers, and specifiers evaluate LOI data correctly and avoid common misinterpretations.
Technical Overview
The limiting oxygen index (LOI) is defined as the minimum concentration of oxygen, expressed as a volume percentage in a flowing mixture of oxygen and nitrogen, that will just sustain flaming combustion of a vertically mounted specimen after ignition. Because ordinary air contains approximately 20.9% oxygen, an LOI of 21% serves as the practical dividing line: materials below this value will continue to burn in air once ignited, while materials above it tend to self-extinguish under ambient conditions.
For thermal insulation products, LOI functions primarily as a rapid, reproducible screening test that reflects the relative ignitability and burning persistence of polymers and foams. It is not, by itself, a complete assessment of fire hazard, because it does not capture heat release, flame spread over assemblies, or smoke production. Accurate LOI result analysis therefore requires a clear understanding of the test conditions, specimen geometry, material behavior such as melting or dripping, and the statistical scatter inherent in the method.
- LOI is measured on a volumetric oxygen concentration basis and reported to the nearest 0.1%.
- The test is a small-scale laboratory method performed on individual specimens, not on installed systems.
- Results are highly sensitive to specimen thickness, conditioning, and flame retardant distribution.
- LOI data support, but do not replace, classification tests under schemes such as GB 8624 or EN 13501-1.
Testing Samples
Combustible Foamed Plastics
Expanded polystyrene (EPS) and extruded polystyrene (XPS) are the most frequently tested insulation foams. Unmodified polystyrene typically yields LOI values in the range of 17 to 19 vol%, meaning it burns readily in air, while flame-retardant grades reach 24 to 30 vol% depending on additive loading. Rigid polyurethane (PU) foam generally falls between 18 and 22 vol%, whereas polyisocyanurate (PIR) foam, which benefits from isocyanurate crosslinking and strong char formation, commonly achieves 28 to 35 vol%.
Phenolic and Specialty Foams
Phenolic foam is among the best-performing organic insulation materials, with typical LOI values above 35 vol% and frequently exceeding 40 vol%. Its aromatic structure promotes rapid char formation that insulates the underlying material and suppresses flame propagation, producing distinctive post-test appearance characterized by thick intumescent char rather than melt flow.
Mineral Fiber Insulation
Rock wool and glass wool products are non-combustible mineral-based materials. In LOI testing they often cannot be ignited at any achievable oxygen concentration, and accredited third-party laboratories typically report such results as "does not ignite" or as an LOI exceeding the measurement range. For these materials, non-combustibility testing under ISO 1182 is generally the more appropriate method, and LOI data mainly serve as confirmatory evidence.
Specimen Preparation and Conditioning
Under ISO 4589-2 and the equivalent GB/T 2406.2, rigid self-supporting specimens are typically cut to 80 to 150 mm in length, 10 mm in width, and 10 mm in thickness, with the specific geometry category recorded because LOI values cannot be compared between different thicknesses. Specimens are conditioned at 23 ± 2 °C and 50 ± 5% relative humidity for a minimum of 88 hours in accordance with ISO 291, unless the product specification mandates a dry or heat-aged condition. Surfaces should be tested as supplied, since removing skins or fused facings can change the measured index.
Testing Items
A complete LOI determination for insulation materials includes both the numerical index and a structured set of qualitative observations that are essential for sound result analysis.
- Determination of the limiting oxygen index value in vol%, reported to 0.1%.
- Ignition behavior, including time to ignition and afterflame time after removal of the igniter.
- Extent of burning relative to the standard 50 mm flame spread criterion or the 3-minute observation period.
- Flaming drips, molten collapse, and flaming particles falling from the specimen.
- Afterglow, charring, intumescence, and shrinkage behavior.
- Scatter among replicate specimens, which reveals material inhomogeneity.
Testing Methods
Standards and Apparatus
The principal standards are ISO 4589-2 (also published as EN ISO 4589-2), ASTM D2863, GB/T 2406.2, and JIS K 7201-2. The procedures are broadly harmonized but differ in details such as permitted specimen dimensions, ignition technique, and calculation approach, so the applied standard must always accompany the reported value. The apparatus consists of a heat-resistant glass chimney, a vertical specimen holder, calibrated oxygen and nitrogen flow controls or mass flow controllers, and a small propane-fueled diffusion-flame igniter.
Test Procedure
The test is performed inside the chimney with a controlled upward flow of premixed oxygen and nitrogen at a linear velocity of approximately 40 ± 10 mm/s through the column.
- Mount the specimen vertically with its upper end at least 100 mm below the chimney top.
- Set an estimated starting oxygen concentration based on material type and prior data.
- Apply the igniter flame to the specimen top for up to 5 seconds, repeating as permitted by the standard until sustained flaming is established.
- Observe whether burning persists for 3 minutes or reaches the 50 mm extent criterion.
- If the criterion is exceeded, reduce the oxygen concentration; if the specimen extinguishes early, increase it, using progressively finer adjustment steps to bracket the transition point.
- Repeat with sufficient specimens to complete a valid bracketing sequence, typically requiring around 15 specimens for a full determination.
Calculation and Reporting
The LOI is calculated as 100 × VO2 / (VO2 + VN2), where the flows are the calibrated volumetric flow rates of oxygen and nitrogen at the transition condition. The test report must state the standard used, specimen dimensions and category, conditioning history, ignition method (for example, top-surface ignition or propagating ignition under ASTM D2863), the calculated index, and all behavioral observations. Flow meters must be corrected for ambient temperature and pressure to keep measurement uncertainty within the roughly ± 0.5 to 1.5 percentage point repeatability normally achievable by competent operators.
LOI Result Analysis and Interpretation
Classification Benchmarks
Although LOI thresholds are indicative rather than universal pass/fail limits, the following bands are widely used for interpreting insulation material data.
- Below 21 vol%: combustible in normal air; the material sustains flaming after ignition (typical of unmodified EPS, PU, and many organic foams).
- 21 to 24 vol%: marginal performance; slow, intermittent burning in air.
- 25 to 28 vol%: flame-retardant territory, typical of treated EPS and XPS products.
- 28 to 32 vol%: strong self-extinguishing tendency, characteristic of PIR foams.
- Above 32 vol%: highly flame-resistant organic materials, such as phenolic foam.
- No ignition or values above 60 vol%: effectively incombustible by this method, typical of mineral wools.
In the Chinese regulatory framework, flame-retardant EPS supplied to B1 grade under GB 8624 is commonly specified with an LOI of at least 30 vol%, and EN 45545-2 for railway applications lists minimum LOI requirements for certain component categories. Always verify the threshold against the governing product specification, because individual specifications differ.
Factors That Influence Results
- Specimen thickness: LOI generally decreases with increasing thickness for many thermoplastics and foams, so only results from identical geometries should be compared.
- Density and cell structure: finer, more uniform cells and higher density often shift LOI upward in foamed plastics.
- Flame retardant content and aging: additive migration and thermal aging can reduce LOI over time; retesting after simulated service aging is advisable for critical applications.
- Moisture content: absorbed water artificially raises the measured index, which is why standardized conditioning is mandatory.
- Temperature: LOI decreases at elevated temperatures, so room-temperature results cannot be extrapolated to fire-exposure conditions.
- Operator and apparatus effects: igniter dwell time, flow calibration, and judgment of the burning criterion are the main sources of inter-laboratory variation.
Common Data Anomalies
Bimodal results across nominally identical specimens usually indicate non-uniform flame retardant distribution or density variation within the board or bun, and should prompt review of production records. Flaming drips can cause apparent propagation that does not reflect the bulk polymer behavior, while severe melting may require supplementary specimen support. Excessive shrinkage of fiber products around the holder and edge ignition effects on poorly cut specimens are further frequent sources of misleading data. A professional result analysis reports the mean index together with the observed scatter and any anomalous burning behavior.
Correlation with Other Fire Tests
LOI shows only an approximate correlation with UL 94 ratings and does not predict performance in intermediate-scale scenarios measured by the single burning item test under EN 13823 or cone calorimetry under ISO 5660. A material with a high LOI can still exhibit significant smoke production or heat release. For a defensible fire safety dossier, LOI should be combined with complementary reaction-to-fire data and, where required, with classification testing under GB 8624 or EN 13501-1 performed by accredited third-party laboratories.
Applications
LOI result analysis supports a wide range of engineering and regulatory activities across the insulation industry.
- Formulation development and optimization of flame retardant loading in EPS, XPS, PU, and PIR foams.
- Incoming inspection and production quality control with rapid lot-release verification.
- Supporting documentation for fire classification submissions under GB 8624 and related frameworks.
- Material qualification for railway, marine, and aerospace programs that cite ISO 4589-2 requirements.
- Comparative product selection during specification and value engineering.
- Failure analysis and post-incident investigation of insulation materials.
FAQ
What LOI value is considered flame retardant for insulation foam?
As a practical guideline, treated EPS and XPS products are expected to reach at least 24 to 26 vol%, while PIR and phenolic foams typically exceed 28 vol%. Specification sheets for flame-retardant polystyrene under the GB framework commonly require 30 vol% or higher, so the governing specification must always be checked.
Why is 21% the critical oxygen index threshold?
Ambient air contains about 20.9% oxygen by volume. A material whose LOI is below 21% can sustain flaming combustion in ordinary air, whereas a material above 21% requires an oxygen-enriched atmosphere to keep burning and will generally self-extinguish in normal conditions.
Why do LOI results differ between laboratories?
Differences typically arise from specimen geometry, conditioning, flow calibration, igniter technique, and interpretation of the burning criterion. Variations of up to about 1.5 percentage points between competent laboratories are considered normal for this method.
Is a high LOI sufficient proof of fire safety?
No. LOI is a small-scale ignitability and burning-persistence indicator only. It does not measure heat release, flame spread on assemblies, or smoke toxicity, so it must be supplemented by intermediate and large-scale reaction-to-fire tests for regulatory classification.
Can LOI be measured on mineral wool?
Testing is possible, but rock wool and glass wool usually cannot be ignited at achievable oxygen concentrations, and the result is reported qualitatively. Non-combustibility testing under ISO 1182 is the appropriate method for these products.
How should borderline or inconsistent results be handled?
Borderline values within roughly one percentage point of a specification limit should trigger repeat testing with fresh specimens, verification of conditioning and flow calibration, and review of specimen homogeneity. Persistent scatter indicates material variability that should be addressed at the production stage rather than hidden by averaging.