Polybutylene terephthalate (PBT) environmental product declarations increasingly display a single global warming potential value in kg CO₂e/kg, but that numerical label is not a polymer property such as density or melting point. Published cradle-to-gate GWP100 results for unfilled PBT resins can span approximately 5.5–9.0 kg CO₂e/kg across different production routes, electricity grids, allocation methods, and LCA background databases. Some of the spread reflects real differences in p-xylene, dimethyl terephthalate or purified terephthalic acid, and 1,4-butanediol supply chains; some reflects inconsistent system-boundary modelling. A value extracted from a technical datasheet without the EPD registration, PCR version, modular boundary, and verification statement may compare figures that are not technically equivalent. The following sections identify the variables that must be fixed before a PBT carbon-footprint comparison is undertaken.

What Does the Published kg CO₂e/kg Actually Represent?

Under ISO 14025:2006, a PBT EPD is a Type III environmental declaration based on life-cycle assessment. The declared unit for resin is commonly 1 kg of pelletized unfilled PBT at the manufacturing gate. It is not a functional unit because it does not state part mass, stiffness, impact resistance, or service life. Modules A1–A3 include raw material extraction and processing, transport to the polymerization plant, and manufacturing. For PBT, A1 includes p-xylene oxidation to terephthalic acid or dimethyl terephthalate, synthesis of 1,4-butanediol, catalysts such as tetrabutyl titanate, heat stabilizers, and packaging. A3 includes melt-phase esterification or transesterification, polycondensation at approximately 240–260 °C, pelletizing, solid-state post-condensation where used, wastewater treatment, and internal logistics. A1–A3 cradle-to-gate data do not include converter drying, injection molding, transport to the final user, or end-of-life. If the PCR aggregates A1–A3 with C1–C4 and D, the headline value has a different meaning than a gate-to-gate resin declaration.

1,4-butanediol synthesis regularly drives PBT upstream carbon intensity because fossil BDO is produced through energy-intensive routes such as the Reppe process from acetylene and formaldehyde, butadiene-based processes, or propylene oxide routes. A coal-to-acetylene BDO chain in one region can exhibit a much higher A1 burden than a natural-gas-based butadiene route in another, with differences exceeding 2 kg CO₂e/kg in some published upstream datasets. The choice between DMT transesterification and PTA direct esterification alters methanol recovery, wastewater load, and reactor energy, but BDO contribution and grid electricity carbon intensity frequently dominate A1–A3. EPD datasets from ecoinvent or GaBi vary by version and by the treatment of co-products such as tetrahydrofuran, γ-butyrolactone, and exported steam. Supplier-specific gate-to-gate data are preferred over generic background data because melt polycondensation energy can differ widely between continuous and batch plants.

Cradle-to-Gate Versus Cradle-to-Grave Boundaries

Module separability is the dominant source of false comparison. A PBT resin EPD reporting A1–A3 at 7.0 kg CO₂e/kg and a PBT part EPD reporting A1–A5 plus C1–C4 at 11.0 kg CO₂e/kg cannot be ranked without recovering module-level values. Module A4 transport from resin producer to injection molder can range from 0.05 kg CO₂e/kg for regional trucking to 0.30 kg CO₂e/kg for longer intermodal or intercontinental shipments, depending on distance, load factor, and mode. Module A5 includes drying at 120–130 °C, injection molding electrical energy, mold temperature control, hydraulic oil, regrind, and scrap. C1–C4 includes decommissioning, transport to waste treatment, incineration or landfill, and credits from recovered heat or electricity. Module D reports loads and benefits beyond the system boundary. A value labelled “net” or “with credits” may be defensible only if the underlying A–C value and the D value are separately enumerated under the relevant PCR.

Comparing PBT grades by mass alone penalizes glass-fiber and mineral-filled compounds. Unfilled PBT typically exhibits a density of 1.31 g/cm³ and a tensile modulus of approximately 2,500 MPa measured according to ISO 527-2. A 30% short-glass-fiber PBT grade typically has a density of 1.53–1.62 g/cm³ and a tensile modulus of 9,500–11,000 MPa. If a component is stiffness-limited, the filled grade can be used at reduced wall thickness, so kg CO₂e/kg is not the correct performance denominator. The specific tensile stiffness of unfilled PBT is roughly 1,900 MPa·cm³/g, while that of a 30% glass-fiber PBT may be 6,200–6,800 MPa·cm³/g. At constant deflection, the mass of a flat plate scales with the inverse cube root of modulus, so the filled part’s mass multiplier is not the simple density ratio. Life-cycle analysts should request primary part mass, wall-thickness calculations, and mechanical datasheets alongside resin EPD values before choosing between neat and reinforced grades.

When Bio-Based BDO or Mass Balance Feedstock Accounting Is Used

Bio-based 1,4-butanediol derived from sugar fermentation or succinic acid hydrogenation introduces biogenic carbon flows that can make A1–A3 values appear low. Under EN 15804+A2:2019, biogenic CO₂ uptake in biomass is reported separately from fossil CO₂ emissions. Cradle-to-gate GWP may be depressed by a negative A1 uptake, while the same biogenic carbon may later be emitted in C1–C4 incineration. A purchaser who compares only A1–A3 between bio-BDO PBT and fossil-BDO PBT therefore sees an incomplete carbon balance. Mass-balance or book-and-claim feedstock attribution under ISCC PLUS or REDcert can further shift feedstock carbon burden without physical segregation. The EPD must state whether renewable electricity certificates, biomass allocation, or co-product credits are embedded in the calculation. Published supplier-specific LCIs for all bio-BDO PBT configurations are limited, so conversion-specific model reviews are required rather than reliance on generic market values.

Downstream processing energy is absent from A1–A3 resin EPDs but materially changes the final part footprint. PBT must be dried to below approximately 0.02% moisture before melt processing; typical drying conditions are 120 °C for 3–4 h in desiccant dryers. If a converter processes PBT at melt temperatures above 260 °C without adequate drying, hydrolysis reduces intrinsic viscosity and generates scrap. Injection molding grades with an intrinsic viscosity of 0.8–1.2 dL/g require tight temperature control; rejected parts raise the effective resin consumption per accepted functional part by the inverse of the yield. An EPD based on 1 kg of accepted part may therefore exceed the resin EPD value by more than A5 electricity alone. Scrap rates of 2–5% are common, but poor drying, hot-runner degradation, or thin-wall connector molds can push this above 10%. These operational losses are not visible in a resin-grade EPD.

How to Verify an EPD Without Mixing A1–A3 and C1–C4

Verification starts with the EPD registration number, programme operator, validity period, and underlying LCA report, not with the headline kg CO₂e/kg. The EPD must name its PCR and version. PBT-specific PCRs for plastic resins may prescribe mass or economic allocation for aromatics and BDO co-products, while construction-product PCRs under EN 15804+A2:2019 require fixed modular reporting and biogenic carbon separation. The GWP100 characterization factor version must also be checked: IPCC AR4, AR5, and AR6 use different radiative efficiency and decay assumptions for methane and nitrous oxide. A third-party verification statement under ISO 14025:2006 should be present for product-specific declarations. If the EPD does not separately report A1–A3, A4–A5, C1–C4, and D, the value should not be used in comparative material selection.

At end-of-life, PBT combustion stoichiometry provides a technical benchmark. The repeat unit C₁₂H₁₂O₄ contains approximately 65.4% carbon by mass; complete oxidation yields roughly 2.4 kg CO₂ per 1 kg of PBT, excluding facility burdens and energy recovery credits. Incineration with energy recovery may report C1–C4 below this value because exported electricity and heat are credited in module D, or because biogenic carbon is separated under EN 15804+A2:2019. Mechanical recycling is affected by collection, sorting yield, and quality loss caused by thermal history and hydrolytic degradation. An EPD that shows lower C1–C4 due to recycling must specify the substitution credit formula and the recycled content substitution ratio. Comparing C1–C4 values without matching waste treatment technology, transport distance, and energy recovery mix reproduces the same system-boundary error as comparing A1–A3 against A1–A5.

Standards and PCRs That Bound PBT EPD Comparability

The core life-cycle assessment framework is defined by ISO 14040:2006 and ISO 14044:2006. Environmental product declaration verification is governed by ISO 14025:2006. For PBT used in building and construction components, EN 15804+A2:2019 adds mandatory modularity and biogenic carbon reporting. ISO 14067:2018 defines product carbon footprint studies and requires the scope and system boundary to be explicitly stated. PBT compounds containing short glass fiber, mineral fillers, halogenated flame retardants, or antimony trioxide cannot be represented by neat resin EPDs because A1 includes additive production and A3 includes compounding energy. A compound EPD or transparent supplier LCA using the correct formulation is required. UL 94 flammability class, RoHS compliance, and REACH registration do not substitute for carbon footprint data.

Halogenated flame-retardant PBT grades used in electrical connectors and relays illustrate the limit of neat-resin data. A typical short-glass-fiber, brominated flame-retardant PBT contains multiple additives at loadings of 5–20% by mass. The antimony trioxide synergist and brominated epoxy or brominated polystyrene carry upstream burdens that can raise A1 GWP by 1–3 kg CO₂e/kg relative to unfilled resin, depending on additive chemistry and supplier location. At 0.8 mm wall thickness, UL 94 V-0 performance requires formulation-specific additive packages; the EPD must match the compound formulation, not a generic PBT family. Injection molders producing thin-wall connectors may also run higher melt temperatures and faster cycle times, increasing A5 electricity per accepted part. Published EPDs for such specific configurations are limited, so grade-specific LCA review is necessary before forming comparative claims.