Limiting Oxygen Index (LOI): Test Method, LOI Values of Plastics and How 11 Flame Retardant Systems Raise Them

The limiting oxygen index (LOI) is the lowest oxygen concentration, in volume percent of a flowing oxygen and nitrogen mixture, that still supports flaming combustion of a plastic specimen, and it is measured to ASTM D2863 and ISO 4589-2. Air holds about 21 vol % O2, so a polypropylene at 17.5 vol % O2 keeps burning in a room while a PVC cable compound at 26 vol % O2 does not: what does it take to move a polymer across that line?

This page sets out the test method, the sourced LOI values of plastics, 11 flame retardant systems and how much they move the number, which additives lower it instead, how the result reads against UL 94 and the cone calorimeter, and the specifications that name an LOI target. It starts where the family of plastic additives narrows to the flame retardants, because the limiting oxygen index is the first number a formulator checks when a compound has to stop burning.

LOI ranks materials under one laboratory condition that no real fire reproduces, so it screens formulations and it does not certify a finished product, a limit stated here before the method, the sourced data and the 11 flame retardant systems that move the number are worked through in turn.

In brief:

• Air: about 21 vol % O2
• Current standards: ASTM D2863-23e1 (D2863-19 superseded) and ISO 4589-2
• Lowest sourced neat polymer: polypropylene, 17.5 to 17.8 vol % O2
• Highest sourced compound on this page: intumescent polypropylene with 1.5 wt% ZnO, at 43.7 vol % O2
• Typical flame retardant content of a plastic product: 2 to 28 wt%

What Is the Limiting Oxygen Index (LOI)?#

The limiting oxygen index is the minimum concentration of oxygen, in volume percent of a flowing oxygen and nitrogen mixture, that just supports flaming combustion of a specimen burning candle-like from the top (ASTM D2863). The specimen stands vertically in the test column, is ignited at the top, and burns downward while the gas mixture flows upward past it; the value a laboratory reports is the oxygen fraction at which the specimen just keeps burning, not the fraction at which it stops. Reporting a concentration rather than a burn time or a temperature holds every other test variable fixed, which lets one number rank one material against another under one condition. The term also appears as "oxygen index (OI)" or "critical oxygen index" in older literature, and it is occasionally written as "limited oxygen index" on supplier and instrument pages, a spelling variant that does not change the test the acronym refers to.

The limiting oxygen index describes a material property measured under laboratory conditions; it is not a rating a finished product carries into service. A plastic's LOI depends on the polymer backbone, the specimen thickness and any flame retardant or filler in the formulation, so the same resin grade can report different LOI results from different sample preparations. Every value on this page therefore carries its host polymer and, where a value comes from a formulated compound rather than the neat resin, its loading.

How is the LOI value calculated?#

The LOI is calculated as 100 times the volumetric oxygen flow divided by the sum of the oxygen and nitrogen flows, so a mixture of 26 volume parts oxygen and 74 volume parts nitrogen gives an LOI of 26.0 vol % O2.

LOI = 100 × [O2] / ([O2] + [N2])

Worked example: 26 parts O2 + 74 parts N2 gives LOI = 100 × 26 / (26 + 74) = 26.0 vol % O2.

The operator does not compute this figure from a single burn or a stopwatch reading. The instrument steps the oxygen fraction of the gas mixture up and down around the point where the specimen just keeps burning, and the concentration recorded at that point, not a burn duration, is the reported LOI value.

What does an LOI above 21 vol % O2 mean?#

An LOI above about 21 vol % O2 means the material needs more oxygen than air contains to keep burning, so under the conditions of ASTM D2863 it stops burning once the igniting flame is removed. Values above about 26 to 28 vol % O2 are commonly described as self-extinguishing, but no standard defines that threshold, and the description comes from general reference sources rather than from ASTM D2863 or ISO 4589-2.

An LOI number by itself does not describe how a material behaves in an actual fire. Three factors change that outcome in ways the LOI test does not capture:

• Ignition source size and heat flux, which in a real fire are rarely as small or as controlled as the LOI test's igniter flame
• Flame spread over a real article, where geometry, orientation and surface area differ from a single upright specimen
• Smoke, acid gas and toxicity of the combustion products, none of which the LOI test measures

Is a high LOI the same as fire safety?#

No: the limiting oxygen index ranks materials under one laboratory condition and says nothing about how much heat a burning article releases, which is the quantity fire-hazard assessment uses. A compound can reach a high LOI and still release heat quickly once it does ignite, because the LOI test measures ignitability under a controlled gas stream rather than the heat release rate that a full-scale fire-hazard assessment is built around.

How the LOI Test Works: ASTM D2863 and ISO 4589-2#

The LOI test runs in four steps: the specimen is clamped upright in the test column, an oxygen and nitrogen mixture flows upward past it, the top of the specimen is ignited, and the oxygen fraction is stepped up and down until the lowest concentration that still supports flaming combustion is found. ASTM D2863 currently stands at D2863-23e1, with D2863-19 superseded, and ISO 4589-2 covers the same ambient-temperature determination.

Table T1. The two current LOI standards.

ASTM D2863 and ISO 4589-2 are cited together in most specifications but are not declared equivalent by either body; check which standard a given value was measured to before comparing it with another. An elevated-temperature part of ISO 4589 exists but its designation is not verified in our source library and is therefore not stated here.

The method sits in the same family as every other bench method used for testing plastic additives, and it is one of the cheapest of them per specimen. A laboratory report has to state the standard and edition, the procedure used, and the specimen type and thickness together with the LOI value, because none of those figures is comparable to another laboratory's result without them.

Which specimens does ASTM D2863 accept?#

ASTM D2863 covers six specimen types, including self-supporting bars and sheet up to 10.5 mm thick with an apparent density above 15 kg/m³, and flexible film that has to be supported vertically. Peer-reviewed flame retardant studies commonly report bars measuring 100 by 6.5 by 3.2 mm or 125 by 6.5 by 3.2 mm, dimensions that reflect the research practice used to generate comparable formulation data rather than a single dimension the standard itself mandates.

Specimen thickness and self-support change the result a laboratory reports, which is why an LOI number published without a specimen description cannot be compared with another. A thicker bar and a thin film of the same compound do not report the same LOI, so a specification that names a target has to name the specimen geometry along with it.

ASTM D2863 procedures A, B and C#

ASTM D2863 defines three procedures: top-surface ignition (A), propagating ignition (B) and a short comparison against a specified minimum oxygen index (C).

• Top-surface ignition (Procedure A), which ignites the top face of the specimen and steps the oxygen concentration to find the index directly
• Propagating ignition (Procedure B), which ignites the specimen so the flame front propagates down its length
• A short comparison against a specified minimum oxygen index (Procedure C), used to confirm pass or fail at one stated concentration rather than to measure the full index

Procedure C is the one a specification uses when it states "LOI not less than 26 vol % O2": the laboratory does not measure the full index, it confirms that the specimen fails to sustain combustion at that stated concentration.

ASTM D2863 and ISO 4589-2 compared#

ASTM D2863 and ISO 4589-2 are both ambient-temperature oxygen index methods that report the result in the same unit, vol % O2. Specifications name one or the other, and datasheets often cite both. Neither body declares them equivalent, so a result should always name the method it was run to.

LOI Values of Plastics: What the Sourced Data Shows#

The sourced LOI of unmodified plastics in this reference runs from polypropylene at 17.5 to 17.8 vol % O2 to polycarbonate at 25 to 29 vol % O2, which places polypropylene below the oxygen content of air and polycarbonate above it. Table T2 lists the four neat polymers this reference can source a value for.

Table T2. Sourced LOI of unmodified plastics.

Only four unmodified polymers have a sourced LOI in our source library. A value without a specimen thickness and a method is not comparable, which is why this table is short rather than long.

Most commodity polyolefins and styrenics sit below 21 vol % O2, and some engineering polymers sit above it, a difference that comes from the polymer backbone itself rather than from an additive. Polypropylene at 17.5 to 17.8 vol % O2 and expanded polystyrene at about 18 vol % O2 both need less oxygen than air contains to keep burning, while polyamide 6 at 21.0 vol % O2 sits at the air line and polycarbonate at 25 to 29 vol % O2 sits above it without a flame retardant.

Published LOI tables for the same nominal polymer often disagree, because specimen thickness, self-support, grade and the test procedure all change the result. This page publishes only values that carry a source, which is why Table T2 lists four polymers rather than the ten or more that some general reference tables show: our source library holds no sourced LOI for PVC, PE, PS, ABS, PMMA, POM, PTFE, PEEK or PSU, and this page does not estimate one.

How Flame Retardants Change the LOI#

Flame retardants raise the limiting oxygen index in five ways: by releasing water and diluting the flame (mineral), by building a char barrier (intumescent and phosphorus), by poisoning the flame chemistry (halogen with antimony trioxide), by boosting another flame retardant at a fraction of a percent (synergists), and, in the opposite direction, by being an additive that feeds the flame instead. Flame retardants make up 2 to 28 wt% of a plastic product depending on the polymer, the target rating and the processing route, a range Chea and colleagues report in a 2025 review, adapting data from Hahladakis et al. (2018).

Which chemistry does the lifting depends on the polymer, and the full classification sits under flame retardants for plastics. Aluminum trihydrate releases water from about 200 °C, an intumescent ammonium polyphosphate system needs 22 to 30 wt% in polypropylene to reach UL 94 V-0, and a halogen and antimony trioxide pairing acts in the gas phase rather than by diluting or charring the melt. The five sections below take each mechanism in turn, in the loading order Table T3 uses.

Table T3. 11 flame retardant systems and their sourced LOI.

Every row is a published formulation, not a recommended recipe. Loadings are given in the unit the source uses (phr for the PVC compound, wt% elsewhere). A loading that works in one grade of one polymer does not transfer. The row 3 zinc borate, row 5-6 zinc oxide, row 8 aluminum diethylphosphinate, row 9 expandable graphite, row 10 organophosphate flame retardants and row 11 flame retardants for polyurethane foam mentions each carry a substance or family page in this site's build, activated as plain-text mentions become links on that page's publication day.

Sources: Huber Advanced Materials cable brochure (rows 1-2); Materials, 2024, doi.org/10.3390/ma17184553 (row 3); Polymers, 2026, doi.org/10.3390/polym17202734 (rows 4-6); PMC12113799, ISO 4589-2 (row 7); PMC12899044, doi.org/10.3390/polym18030351 (row 8); Polymers, 2025, doi.org/10.3390/polym15204100 (row 9); PMC11642039 (row 10); Polymers, 2025, doi.org/10.3390/polym17182459 (row 11).

Mineral flame retardants: ATH and MDH#

Mineral flame retardants raise the LOI by releasing water into the flame: aluminum trihydrate starts to release its water at about 200 °C and absorbs 1051 J per gram doing it, which cools the surface and dilutes the fuel. Magnesium hydroxide works the same way at a higher decomposition temperature, which is why it replaces aluminum trihydrate in polymers processed too hot for ATH to survive intact.

A published lead-free PVC cable compound reaches 26 to 27 vol % O2 with 45 to 100 phr of aluminum trihydrate and 5 phr of zinc borate, and it passes UL 94 V-0 at 3 mm. A polypropylene halogen-free cable compound with 185.7 phr of coated magnesium hydroxide reaches 30.2 vol % O2. Grades, coatings and the loading penalty each mineral system carries are compared under mineral flame retardants (ATH and MDH).

Loading is the price of the mineral route: a low-smoke, zero-halogen compound runs about 60 % mineral filler by weight, mechanical properties fall as that share climbs, and a typical halogen-free flame-retardant cable compound carries 160 to 180 phr of ATH or MDH.

Intumescent and phosphorus flame retardant systems#

Intumescent and phosphorus flame retardants raise the LOI by building a swollen char layer that keeps oxygen away from the melt, and in polypropylene an ammonium polyphosphate system needs 22 to 30 wt% to reach UL 94 V-0. The char forms as the phosphorus component reacts with a carbon-forming agent and a blowing agent in the formulation, and the intumescent route processes up to about 220 °C, which sets its upper limit in polypropylene compounding.

The phosphinate route in polyamides reaches the same LOI band at a fraction of the loading: 15 wt% of an aluminum diethylphosphinate system gives PA6 an LOI of 30.7 vol % O2 and UL 94 V-0 at 3.2 mm, and a 20 wt% package of expandable graphite, aluminum diethylphosphinate, melamine polyphosphate and organoclay gives a glass-fibre-reinforced PA6 (25 % glass fibre) an LOI of 32.0 vol % O2 with a peak heat release rate of 103 kW/m². The full intumescent and phosphorus route for polypropylene, including the pentaerythritol and metal-oxide variants covered above, is set out under flame retardants for polypropylene.

Processing temperature decides which route applies: intumescent ammonium polyphosphate systems are limited to about 220 °C, while the engineering polymers that use the phosphinate route process at 240 to 320 °C.

Halogen flame retardants and antimony trioxide#

Halogenated flame retardants raise the LOI by acting in the gas phase, and antimony trioxide multiplies that effect by carrying the halogen into the flame as antimony halide, interrupting the radical chain reaction that sustains combustion. Our source library holds no sourced LOI value for a halogen and antimony trioxide system in a specific polymer, so this section describes the mechanism and leaves the number to the flame retardant pages.

Antimony trioxide carries a harmonised Carc. 2 classification under Regulation (EC) No 1272/2008, which is why halogen-free routes are specified in many cable and rail programmes even where a halogen and antimony pairing would reach a comparable LOI. The classification, supply risk and permitted uses of antimony trioxide sit on its substance page.

Flame retardant synergists that raise the LOI at 1 wt% or less#

Synergists can move the LOI further per unit of loading than the flame retardant they support. In a 2026 peer-reviewed study, adding 0.25 wt% of zinc oxide or manganese oxide to an intumescent polypropylene raised its LOI to 30 vol % O2 and reached UL 94 V-0, and raising the zinc oxide addition to 1.5 wt% lifted the same system to 43.7 vol % O2, the highest sourced value on this page.

The ceria substitution is the clearest case, because the total additive load did not change: a 2026 peer-reviewed study found that swapping 1 wt% of the intumescent package for 1 wt% of layered cerium oxide raised the LOI from 29.4 to 32.6 vol % O2, while cutting total heat release by 38.9 % and total smoke release by 74.3 %. A separate formulation, 10 wt% zinc borate combined with 10 wt% magnesium hydroxide in a calcium-carbonate-filled polypropylene, reaches 29.4 vol % O2 through mineral synergy rather than a metal-oxide addition. Every number in this section comes from a single published study, so it is stated as something a formulation can reach, not as a target every grade will hit.

Plasticizers, fillers and other additives that lower the LOI#

Three additive groups pull the limiting oxygen index down: plasticizers, which add combustible organic mass, inert fillers, which replace flame retardant rather than polymer, and blowing agents, which open the surface to oxygen. Our source library records that plasticizer type changes the LOI of PVC but holds no sourced value for the size of that change per phr, so the direction is stated here and the number is not.

• Plasticizers in flexible PVC, which add combustible organic mass without contributing char or flame inhibition
• Inert mineral fillers, which dilute the flame retardant package rather than the polymer itself
• Foamed and cellular structures, which open more surface to the oxygen stream
• Glass fibre, which can wick the flame along the specimen and worsen a UL 94 result at the same LOI

A flexible PVC compound carries tens of phr of combustible ester, and the classes are compared under plasticizers for plastics. The published lead-free PVC cable reference formulation carries 55 phr of a plasticizer alongside 45 to 100 phr of aluminum trihydrate, and it still needs that mineral loading to reach 26 to 27 vol % O2. Inert mineral fillers work in the opposite direction from the flame retardant beside them, and the fillers for plastics family has its own page rather than a link from this table.

LOI, UL 94, the Cone Calorimeter and the Glow Wire Test#

The limiting oxygen index answers how much oxygen a material needs to keep burning, UL 94 answers whether it self-extinguishes and drips, the cone calorimeter answers how much heat it releases, and the glow wire answers whether a hot electrical contact can ignite it. Four bench tests answer four different questions, and a single compound can pass one and fail another because each test isolates a different variable: oxygen concentration, dripping behaviour, heat release rate under radiant flux, or ignition by a hot wire.

Table T4. Which fire test answers which question.

A specification almost always names a rating and a thickness rather than an oxygen index, and the criteria are set out under UL 94 flammability ratings. UL 94 V-0 requires each specimen's afterflame time at 10 seconds or less, a total afterflame of 50 seconds or less across five specimens, no afterflame or afterglow beyond 30 seconds after the second flame application, no burning to the clamp and no cotton ignition from flaming drips, while V-2 allows flaming drips that ignite the cotton indicator below the specimen. Every UL 94 rating on this page is written with the thickness it was measured at, because the same compound reports a different rating at a different wall thickness.

Why LOI and UL 94 rank the same compound differently#

A compound can gain LOI and lose its UL 94 rating, because the two tests treat dripping in opposite ways: a specimen that drips away from the flame in the LOI column reads as harder to burn, while in UL 94 the same drips can ignite the cotton indicator and push the rating from V-0 to V-2.

LOI is a screening and ranking number for formulation work; UL 94 at a stated thickness is what a specification and a UL Yellow Card actually list. Cone calorimeter peak heat release correlates poorly with UL 94 for intumescent systems, a point Schartel B. (BAM Berlin) and Hull T.R. (UCLan) set out in their 2007 paper "Development of fire-retarded materials: interpretation of cone calorimeter data" in Fire and Materials, so three bench tests run on one compound can point in three different directions.

Which Specifications Ask for an LOI Value?#

The limiting oxygen index appears as a stated requirement mainly in cable and insulation specifications: PVC cable compounds are commonly specified at not less than 26 vol % O2, and ASTM C578 asks expanded polystyrene insulation for more than 24 vol % O2 against the roughly 18 vol % O2 of unmodified EPS.

Table T5. Where an LOI number is specified.

PVC cable compounds: the LOI 26 target#

PVC cable compounds are commonly specified at a limiting oxygen index of not less than 26 vol % O2, and a published lead-free reference formulation reaches 26 to 27 vol % O2 with 45 to 100 phr of aluminum trihydrate and 5 phr of zinc borate, alongside 100 phr of PVC K70, 55 phr of a DIDP plasticizer, 2.7 phr of a lead-free stabilizer and 10 phr of chalk, and passes UL 94 V-0 at 3 mm. Unmodified PVC is often quoted with a high LOI in general reference tables, but our source library holds no sourced value for it, so this page gives the compound values it can source and not a figure for the neat polymer.

Zinc borate at 3 to 6 phr alongside aluminum trihydrate reduces dripping and smoke in the same compound class. The stabilizer, plasticizer and filler package that sits around the flame retardant is on additives for PVC.

EPS and XPS insulation: the ASTM C578 requirement#

Expanded polystyrene without a flame retardant sits at about 18 vol % O2, while ASTM C578 asks foam insulation for more than 24 vol % O2, which is the gap the flame retardant package has to close. Hexabromocyclododecane (HBCD) filled that gap until it was listed under Stockholm Annex A in May 2013 and its REACH Annex XIV sunset date passed on 21 August 2015; polymeric brominated flame retardants have replaced it in current EPS and XPS formulations.

The polymeric flame retardants that replaced HBCD in foam are listed under additives for polystyrene.

Wire, cable and rail specifications that use other fire tests instead#

European cable and rail programmes do not classify on the limiting oxygen index: cables are classed by EN 50399 under the Construction Products Regulation, and rail interiors by EN 45545-2 hazard levels. The B1ca class under EN 50399 limits flame spread to 1.75 m or less, total heat release (THR1200) to 10 MJ or less, peak heat release to 20 kW or less and FIGRA to 120 W/s or less, with B2ca, Cca and Dca set at progressively less demanding levels as recorded in supplier and association summaries. Regulation (EU) 2024/3110 applies from 8 January 2026 and places cables in product family 31 under the Construction Products Regulation.

In rail interiors, a glass-fibre-reinforced PA66 with an Exolit OP phosphinate system reaches hazard level HL3 under requirement set R22 of EN 45545-2, and a glass-fibre PBT with the same phosphinate class reaches HL2. The three cable systems and their full additive packages are compared under additives for wire and cable compounds.

How to Select Flame Retardant Additives with LOI: 7 Steps#

Use the limiting oxygen index in seven steps: record the target, check the processing temperature, screen by LOI, fix the specimen geometry, add synergists last, confirm by UL 94 at the specified thickness, then re-check smoke, mechanics and regulatory status.

1. Record the target: the LOI value, the UL 94 rating and the thickness the specification actually names.
2. Check the processing temperature of the host polymer before choosing a chemistry.
3. Screen candidates by LOI first, because the test uses small specimens and little material.
4. Fix the specimen thickness and method for the whole screening series, so the values stay comparable.
5. Add the synergist only after the base system is at its loading limit.
6. Confirm the shortlist by UL 94 at the specification thickness, not by LOI.
7. Re-check smoke, acid gas, mechanical properties and regulatory status before the formulation is frozen.

Engineering polymers that process at 240 to 320 °C should rule out aluminum trihydrate and most ammonium polyphosphate systems at step 2, which is why the phosphinate and expandable graphite routes dominate polyamide and polyester formulations. These seven steps sit inside the general framework for how to select plastic additives, which covers dosage, compatibility and processing constraints beyond flame retardancy alone.

Narrow the candidate list first with the flame retardant selector by polymer and UL 94 rating before requesting samples, since a shortlist built on LOI and processing temperature alone still needs a UL 94 confirmation at the specification thickness.

Get quotes for the flame retardant system you shortlisted. Tell us the polymer, the target LOI and the UL 94 thickness through the plastic additive supplier finder, with flame retardant manufacturers and suppliers named in the same request.

What Else Do You Need Besides an LOI Value?#

An LOI value alone does not release a compound for production: smoke density, acid gas, dripping, mechanical retention and the regulatory status of the flame retardant all have to pass as well. LOI is a small-scale ranking tool, not a fire-hazard metric, and every one of those additional checks uses a different bench test and a different pass criterion.

Smoke, acid gas and flaming drips#

A compound that reaches a high LOI can still fail on smoke: aircraft interior materials are limited to a specific optical density of 200 at four minutes, measured to ASTM E662, and cable acidity is classed separately under IEC 60754. Zinc borate at 3 to 6 phr alongside aluminum trihydrate reduces dripping and smoke in PVC, and the 2026 ceria-substitution study cut total smoke release by 74.3 % at the same time it raised LOI. Zinc borate and the molybdates that cut optical density further are compared under smoke suppressants. Cable acidity under IEC 60754 sets an a1 class at a conductivity below 2.5 µS/mm and a pH above 4.3.

The regulatory status of the flame retardant you choose#

An LOI result proves nothing about whether the flame retardant that produced it may still be used: hexabromocyclododecane was listed under Stockholm Annex A in May 2013 and its REACH Annex XIV sunset date passed on 21 August 2015, long after it had reached the LOI targets of EPS insulation. Antimony trioxide carries a harmonised Carc. 2 classification under Regulation (EC) No 1272/2008, and Canada's Prohibition of Certain Toxic Substances Regulations, 2025 (SOR/2025-270, registered 12 December 2025) sets incidental limits of 100 mg/kg for HBCD and 1,000, 500 and 10 mg/kg for three PBDE groups, with end dates to 2044 for Dechlorane Plus. Current bans, sunset dates and product standards for every flame retardant class are tracked under flame retardant regulations. The health effects behind these listings sit on a separate flame retardant health effects page and are mentioned, not detailed, here.

What does "limiting oxygen concentration" mean?#

The limiting oxygen concentration (LOC) is a process-safety property of flammable gas, vapour and dust mixtures, and it is not the limiting oxygen index: LOI describes a solid specimen burning in a controlled oxygen and nitrogen stream. The two share an acronym family and nothing else, and this reference does not publish LOC values.

Is the LOI test the same for fabrics and foams?#

The apparatus and the underlying principle are the same, but the specimen holder and the specimen type are not, and a value measured on one specimen type cannot be compared with a value from another. Fabric LOI values are covered by textile standards outside the scope of this reference and are not published here; ASTM D2863's own specimen scope covers self-supporting bars and sheet up to 10.5 mm thick and flexible film supported vertically, which is the boundary this page works within.

Where can you get LOI testing done?#

Any fire-testing laboratory accredited for ASTM D2863 or ISO 4589-2 runs the test, and the report has to name the method, the procedure, the specimen thickness and the specimen type for the value to be usable. This reference names no laboratory and quotes no price, since comparing accredited labs and their turnaround times sits outside what a plastics-additive reference can verify.