Introduction

A finding that often surprises Life Cycle Assessment (LCA) practitioners is striking: metals systematically dominate 90 to 99% of toxicity scores calculated with USEtox, regardless of the product studied. Whether electronics, detergents, food, buildings, vehicles, textiles, or cosmetics, the result is almost always the same: metals overwhelm the scores, while organic substances often contribute less than 1%.

The European Joint Research Centre's analysis of 30 representative Environmental Footprint products confirmed this pattern: zinc dominates ecotoxicity in 16 out of 30 products, with copper and chromium as major contributors in most other cases (Saouter et al., 2018).

This observation is neither an error nor a modeling artifact. It results directly from three converging factors:

1. Intrinsic properties of metals (infinite persistence, bioaccumulation, chronic toxicity)
2. USEtox's intentional design (identifying the 10-20 dominant contributors)
3. Inventory data availability (metals exhaustively documented, organics largely absent)

This article explains why this dominance is scientifically expected and why the JRC now recommends separating metals, organics, and inorganics for meaningful interpretation.

What USEtox Is Designed to Do (and Not Do)

USEtox Identifies the 10-20 Dominant Contributors

Before understanding why metals dominate, it's essential to grasp what USEtox is meant to do. The official USEtox 2.0 documentation explicitly states (Fantke et al., 2018, DOI: 10.11581/DTU:00000011):

"In practice, this means that for LCA practitioners these toxicity potentials are very useful to identify the 10 or 20 most important chemicals pertinent for their comparative applications, while implying a motive to disregard hundreds of other substance emissions whose impacts are by far less significant."

The documentation further emphasizes:

"It is usually more meaningful and thus recommended to plot and compare toxicity impact scores on logarithmic scales, avoiding the over-interpretation of small differences of a factor <10 that may appear large on a linear scale."

The Logarithmic Scale Principle: Orders of Magnitude vs Precision

This logarithmic interpretation principle is fundamental and radically distinguishes USEtox from regulatory risk assessment:

USEtox operates on orders of magnitude:

• Differences of a factor 2 to 5 are considered negligible
• Only differences of factor 10, 100, 1000+ are meaningful
• Results span 15 orders of magnitude across all substances
• Individual uncertainty is ±3 orders of magnitude

Risk assessment requires precision:

• A 2-fold difference in toxicity can change a substance's classification
• A 20% change in exposure can shift a substance from "safe" to "of concern"
• DNELs and PNECs operate with safety factors of 10 to 1000
• Dose-response relationships require mg/kg precision

Practical consequence: In USEtox, two substances with characterization factors of 5×10³ and 2×10⁴ are essentially equivalent (both ~10⁴). In risk assessment, a substance with DNEL = 5 mg/kg and another with DNEL = 20 mg/kg represent significantly different safety profiles.

This design choice means that USEtox is intentionally built to compress toxicity into a handful of dominant contributors—and metals, due to their unique properties, naturally emerge as these dominants.

Metals Combine All Characteristics That Maximize USEtox Scores

USEtox characterization factors are calculated using the formula:

CF = FF × XF × EF

Where:

• FF (Fate Factor): Environmental persistence
• XF (Exposure Factor): Human or ecosystem exposure potential
• EF (Effect Factor): Intrinsic toxicity (inverse of HC20 or ED50)

Metals exhibit extreme values on all three factors simultaneously.

Infinite Environmental Persistence (FF → Maximum)

Metals are chemical elements—they cannot degrade.

Unlike organic molecules that undergo:

• Biodegradation (microbial decomposition)
• Photolysis (UV degradation)
• Hydrolysis (water-mediated degradation)
• Atmospheric oxidation (reaction with OH radicals)

Metals persist indefinitely in the environment.

Bioaccumulation and Food Chain Transfer (XF ↑↑)

Many metals bioaccumulate through food webs, creating multiple exposure pathways for humans:

Direct pathways:

• Drinking water (As, Pb, Cr, Cu)
• Inhalation of atmospheric particulates (Pb, Cd, As)

Indirect pathways via food chain:

Crops: Metals in soil → root uptake → edible plant parts

• Cd concentrates in leafy vegetables, cereals (rice)
• Pb in root vegetables
• Cu and Zn in all crops

Animal products: Soil/feed → livestock → meat, milk

• Cd accumulates in kidneys and liver
• As in poultry (historical use in feed additives)

Fish: Water → phytoplankton → zooplankton → fish → humans

• Methylmercury bioamplifies in predatory fish
• Cd in shellfish and crustaceans

USEtox models these pathways using empirical transfer factors (BCSF for soil-to-plant, BTF for animal products, BCF for fish) rather than detailed mechanistic modeling, but captures the cumulative effect of multi-pathway exposure.

High Chronic Toxicity (EF ↑↑)

Metals exhibit diverse mechanisms of chronic toxicity generating high effect factors:

USEtox effect factors (EF) are inversely proportional to toxicity thresholds: lower thresholds → higher EF → higher characterization factors.

The Multiplicative Effect

Because CF = FF × XF × EF, metals achieve CFs several orders of magnitude higher than most organics.

USEtox documentation confirms that characterization factors vary over 15 orders of magnitude across substances, with metals systematically occupying the highest positions.

Where Do Metal Emissions Come From in LCA?

Most metal emissions contributing to toxicity scores in LCA do not originate from the studied product itself. They come from well-documented upstream industrial processes embedded in virtually every product system.

Mining and Mineral Processing

Primary source of metals to environment in LCAs:

Mine tailings:

• Sulfide ore processing releases: Cd, Pb, As, Zn, Cu, Hg
• Tailings dams leach metals for decades after mine closure
• Wind erosion disperses metal-contaminated dust

Ore beneficiation:

• Flotation circuits discharge wastewater with dissolved metals
• Slag disposal sites leach metals to groundwater

Metal Refining and Smelting

Pyrometallurgical processes:

• Roasting of sulfide ores → SO₂ + volatilized As, Cd, Hg, Pb to air
• Smelting → dust emissions containing Zn, Pb, Cu, Cd
• Electrolytic refining → wastewater with Cu, Ni, Zn

These emissions are systematically included in every LCA using:

• Metals in product (electronics, vehicles, buildings)
• Steel (Cr, Ni, Mn emissions during production)
• Grid electricity (coal power emits metals)

Fossil Fuel Combustion

Coal combustion is the largest anthropogenic source of several metals. Every kWh of coal electricity in ecoinvent includes emissions of Hg, As, Cr, Ni, Pb, Zn to air.

Oil and natural gas combustion emit smaller but significant quantities of Ni, V (heavy crude oil) and Hg ("sour" gas).

Waste Incineration

Municipal waste contains metals from:

• Pigments in plastics (Cd, Cr, Pb)
• Printing inks (Cu, Zn, Ni)
• Electronic components (Ag, Au, Cu, Sn, Pb)
• Batteries (if not sorted): Cd, Hg, Pb, Ni

Emission pathways:

• Volatilized metals to air: Hg, Cd, Pb, Zn
• Bottom ash leachate to water: Cu, Zn, Pb
• Fly ash disposal: concentrated metals

Transportation Infrastructure

Tire and brake wear:

• Tires: Zinc (activators and accelerators) in formulation
• Brake pads: Cu, Sb, Ba (friction materials)
• Worn particles deposit on road surfaces → wash into stormwater

Road surface wear:

• Asphalt contains trace metals
• Guardrails (galvanized steel) corrode → Zn runoff

These emissions are included in LCAs of vehicles and transportation services.

Agriculture

Fertilizer application:

• Phosphate fertilizers contain Cd (naturally occurring in phosphate rock)
• Cumulative effect over decades → soil Cd accumulation

Pesticides:

• Copper fungicides (vineyards, orchards): repeated application
• Old vineyard soils can contain high Cu concentrations

Animal manure:

• Copper and zinc from feed additives (growth promoters in pig farming)
• Repeated spreading → soil accumulation

Why These Emissions Are Visible in LCA

Life cycle inventory databases thoroughly document metals because:

1. Regulatory monitoring: Large industrial facilities must report metal emissions (PRTR registers)
2. Long monitoring history: Metal emissions have been regulated since 1970s-1980s
3. Analytical simplicity: ICP-MS measures all metals simultaneously in one sample
4. Mass balance approaches: Input/output metal accounting in industrial processes

In contrast, organic emissions are often:

• Not monitored individually (aggregated as "VOCs," "TOC," "COD")
• From diffuse sources (consumer products) without emission factors
• Not subject to mandatory reporting for many substances
• Analytically complex (require GC-MS or LC-MS with substance-specific methods)

Result: LCA databases "see" metal emissions clearly across all life cycle stages, while organic emissions are largely invisible.

Why Organic Chemicals Appear Negligible in USEtox

Several structural factors cause organic chemicals to disappear from USEtox results:

LCI Database Underreporting

Organic emissions are systematically underreported:

• Formulation ingredients (surfactants, fragrances, preservatives) rarely have emission factors
• Consumer product use emissions are absent in most LCI databases
• Industrial organic emissions aggregated as "Organic substances, unspecified" (no CF available)

Rapid Degradation → Low Fate Factors

Many environmentally relevant organics degrade rapidly. Even with high intrinsic toxicity, short environmental residence → low FF → low overall CF.

Sorption Reduces Bioavailability

Highly lipophilic organics (log Kow > 5-6) sorb strongly to sediments and soils:

• Reduced aqueous bioavailability
• Lower crop uptake
• Lower XF despite high Kow

Paradox: Very hydrophobic substances – usually having a high toxicity - may have lower exposure factors than moderately hydrophobic ones

System Boundary: Far-Field vs Near-Field

USEtox far-field models continental to global scale distribution. Local emissions from product use:

• Are diluted to continental-scale concentrations
• May be excluded if classified as "indoor" emissions (separate near-field module)
• Lack inventory data in most LCA databases

Result: Organic chemicals relevant to consumer exposure (cleaning products, cosmetics, air fresheners) may be entirely absent from the LCA inventory, while metal emissions from electricity production hundreds of kilometers away are fully quantified.

Why the JRC Recommends Separation by Chemical Class

Following the 2013-2018 Environmental Footprint Pilot phase, the JRC identified critical problems (Saouter et al., 2018):

Problems identified:

• Metals dominate 95 to 99% of toxicity scores in most product categories
• Organic substances contribute <1% and become invisible
• Stakeholders conclude "only metals matter" → ignore organic hazards in formulations
• Substitution decisions are distorted: companies focus on reducing metal-linked processes (switching from coal to renewable energy) while ignoring potentially hazardous organic ingredients

JRC Solution: Separate reporting by chemical class

Metals (and metalloids)

• Persistent elements
• Reported separately to reflect their dominance
• Acknowledges that USEtox metal modeling has limitations (no speciation, no background concentration consideration)

Organic substances

• Carbon-based molecules
• Reported separately to restore visibility
• Allows comparison within organic class (e.g., solvent A vs solvent B)

Inorganic substances (non-metallic)

• Acids, salts, minerals (e.g., ammonia, sulfates, phosphates)
• Often have different fate/exposure mechanisms

Benefits of the Split:

1. Prevents misleading "zero impact" impression for organics
2. Enables meaningful comparisons within each class
3. Better alignment with chemical policy (REACH focuses heavily on organics)
4. Acknowledges model limitations for metals (USEtox was developed primarily for organics)
5. Supports informed decision-making: 
6. Reduce metal emissions through energy/material choices
7. Reduce organic hazards through formulation optimization
8. Both matter, but in different contexts

Conclusion

Metals dominate USEtox toxicity scores due to the convergence of three factors:

1. Intrinsic properties: Infinite persistence + bioaccumulation + high chronic toxicity
2. USEtox design: Intentionally focuses on top 10-20 contributors + uses logarithmic scale
3. LCA data availability: Metals thoroughly documented, organics largely absent

This dominance is not an error or modeling artifact—it is the expected output when evaluating long-term, global-scale toxic pressure from life cycle emissions.

Key implications:

• Metal dominance reflects life cycle-wide environmental burden, not use-phase safety
• Organic chemicals matter in other assessment contexts (risk assessment, near-field exposure)
• The JRC split (metals / organics / inorganics) is scientifically necessary
• USEtox results should always be interpreted on logarithmic scale (factor 10 threshold)
• Small differences (2×, 5×) are within uncertainty and should not drive decisions
• USEtox cannot assess whether consumer products are "eco-friendly" or "safe for consumers"—product ingredient contributions will be masked by life cycle metals

For practitioners:

• Expect metals to dominate in LCA toxicity results
• Use separate reporting to restore visibility of organic contributions
• Recognize that USEtox far-field answers a specific question: "Which life cycle emissions cause the greatest cumulative toxic pressure?"
• This question is different from "Which product ingredients pose the greatest consumer risk?"—both questions are valid and necessary, but require different tools

USEtox was designed to identify substances that dominate long-term global toxic pressure. For sound physical and chemical reasons, these substances are predominantly metals.