Carbon footprint of additive manufacturing: why every database disagrees

You want to report the carbon footprint of a titanium LPBF part. You open four LCA databases and get four different emission factors for Ti-6Al-4V powder. The spread isn't rounding error — it's 3× to 5×. Which one do you use, and why do they differ so much?

This article explains the system-boundary choices that cause the disagreement, shows you how to read each database correctly, and tells you which to cite for different purposes.

The four major databases

All four are legitimate. They disagree because they make different choices about what counts.

Why they disagree: system boundaries

1. Cradle-to-gate vs. cradle-to-grave

Most emission factors for materials report cradle-to-gate: mining ore + refining + manufacturing the feedstock, up to the factory gate. They don't include:

• Transportation to your machine
• Energy consumed during printing
• Post-processing (HIP, machining, heat treat)
• End-of-life (recycling, landfill)

Some databases (ecoinvent in particular) offer cradle-to-grave datasets for some materials, which are much higher. Never mix a cradle-to-gate factor with a cradle-to-grave comparison.

2. Powder vs. wrought material

Ti-6Al-4V powder for LPBF is not the same as a wrought titanium billet. Powder atomisation (gas atomisation or plasma rotating electrode process) consumes significant additional energy:

• Wrought Ti-6Al-4V: ~50 kgCO₂e/kg (ecoinvent)
• Gas-atomised Ti-6Al-4V powder: ~85–130 kgCO₂e/kg depending on process

Many databases only have wrought datasets. When you use the wrought factor for powder, you undercount by 30–60 %.

3. Background energy system

The electricity mix assumed in the dataset matters enormously for energy-intensive metals.

• ecoinvent uses a regional market electricity mix for the country of production (e.g., Chinese titanium sponge uses Chinese grid).
• GREET uses US average grid intensity.
• ICE uses the UK grid for most datasets.

Chinese titanium production runs on a coal-heavy grid; Norwegian aluminium smelting uses hydropower. Same metal, very different carbon intensity.

4. Allocation method

When a smelter produces multiple outputs (e.g., aluminium ingot + slag), the emissions must be allocated across outputs. ecoinvent uses economic allocation by default; some datasets use mass or energy allocation. The choice can shift reported carbon by 20–40 %.

5. Data vintage

The electricity grid decarbonizes over time. An emission factor from 2015 is meaningfully different from one from 2023. ICE v3.0 has not been updated since 2019; its electricity-intensive materials are likely conservative (overestimates today's grid).

What each database does well

ecoinvent v3.10

Strengths: Most comprehensive, regularly updated, country-specific electricity mixes, transparent documentation, used for formal Environmental Product Declarations (EPDs).

Weaknesses: Commercial license ($); few powder-specific datasets (most are wrought or ingot).

When to cite: Formal LCA, ISO 14044 compliant studies, EPDs, supplier assessments.

Ti-6Al-4V coverage: Wrought Ti-6Al-4V alloy available; no dedicated LPBF powder dataset. Typical: 65–80 kgCO₂e/kg (wrought, cradle-to-gate, global average).

IDEMAT 2024

Strengths: Free, maintained by Delft for engineering design education, covers AM-specific materials (some powder datasets).

Weaknesses: Smaller scope than ecoinvent, less granular regional coverage.

When to cite: Student projects, early-stage design comparison, teaching.

Ti-6Al-4V: Approximately 70–90 kgCO₂e/kg.

ICE v3.0 (University of Bath)

Strengths: Free, UK-focused, widely accepted in UK construction and procurement contexts.

Weaknesses: Last updated 2019, limited to ~130 materials, no powder datasets.

When to cite: UK projects; quick comparison to structural materials (concrete, steel, timber).

Ti-6Al-4V: ~81 kgCO₂e/kg (wrought, cradle-to-gate).

GREET 2023 (Argonne)

Strengths: Free, rigorous US-centric energy analysis, excellent for comparing AM vs. CNC in North American context, covers AM energy consumption in lifecycle.

Weaknesses: US-specific; thinner on non-US materials; designed for vehicle applications.

When to cite: North American manufacturing; comparing AM to machining from billet; fuel/energy comparisons.

Ti-6Al-4V: ~60–75 kgCO₂e/kg.

The electricity factor problem

Material production carbon is only half the picture. The electricity consumed during printing must also be attributed. A typical LPBF machine draws 8–12 kW and prints for 20–200 hours per part.

Grid emission factors vary by factor of 10+ across regions:

A 10 kW LPBF machine running for 50 hours produces:

• Norway: 6 kgCO₂e (printing energy only)
• China: 290 kgCO₂e

For an aluminium part where the material factor might be 10–20 kgCO₂e/kg, the print location dominates the total footprint.

How to report correctly

Follow ISO 14067 (Carbon footprint of products):

1. State your system boundary explicitly: cradle-to-gate, or cradle-to-gate plus printing energy, or full cradle-to-grave.
2. Cite the database and version: "Ti-6Al-4V cradle-to-gate, ecoinvent v3.10, global average market dataset."
3. State the grid source: "Printing energy attributed using IEA 2024 EU average grid, 0.233 kgCO₂e/kWh."
4. Report the range: If you're comparing AM to CNC, report both the material factor and the energy factor separately, so the reader can substitute their own grid intensity.
5. Don't use powder factors from wrought datasets unless you explicitly add an atomisation penalty (typically +30 to +60 %).

Practical example — Ti-6Al-4V LPBF bracket (100 g part)

The 1.3× spread here is modest because titanium production dominates. For aluminium (more electricity-intensive at smelting), the spread between databases is wider, and the grid factor is relatively more important.

Use the Carbon Footprint Calculator to compute all combinations for your part in one table.

The honest answer

There is no single "correct" emission factor. The right approach is:

1. Report the full range from all credible databases.
2. State your grid assumption.
3. Use the same methodology consistently when comparing AM vs. conventional manufacturing — the relative comparison is more defensible than any absolute number.

The Carbon Footprint Calculator on this site computes all 16 combinations (4 material databases × 4 grid databases) and returns the min/median/max range, so you can report with appropriate uncertainty.

Related tools

• Carbon Footprint Calculator — kgCO₂e per part across all databases and grid regions
• DfAM Design Rules Checklist — reduce waste (and carbon) before you build
• Cost-Per-Part Estimator — cost and carbon are the two sides of AM justification