Specifying Color Consistency for Artificial Olive Trees in Airport Terminal Fit-Outs: Batch Codes, Delta E Tolerance and Sample Retention

Executive Summary

Airport terminal fit-outs that specify artificial olive trees at 3–6 m should write color as a number rather than a photograph: ΔE ≤2.0 against a signed golden sample, measured with a spectrophotometer in three zones per leaf under D65, with a batch code printed on every carton. Trees sit on Ø48–76 mm galvanized cores carrying 1,800–2,700 leaves/m³, and interior foliage carries NFPA 701 performance. The color clause is what keeps a second phase from arriving a visibly different green.

Grey fiberglass trunk shells racked by height in the workshop, crown assemblies waiting behind them.

1. Failure Modes & Commercial Pitfalls

Color is the one variable an airport buyer cannot inspect on arrival without a reference. A canopy that is 15% too light does not fail a snag list, cannot be rejected on a certificate, and cannot be repaired on site. It simply reads wrong for the next ten years, in a building where every passenger walks past it twice.

Five failures account for most of the disputes.

Showroom match, terminal mismatch. Two greens that agree under a 6500 K viewing booth can separate under the 3000 K downlights of a departures hall, because the pigments metamerise — their reflectance curves differ even where the L\a\b\* values agree. A sample approved on a bench, not under the actual lighting, is an approval of the wrong object.

Drift inside a single order. A 20-tree package rarely leaves the moulding floor in one run. Three setups over five weeks, a new masterbatch lot in week four, and the last six trees carry a green that is 2–3 ΔE away from the first six. Nobody notices in the crates; everybody notices when the row is assembled.

Phase-two matching against a drifting reference. Eighteen months later the second phase is matched to the original standard while the installed trees have moved 2–4 ΔE under UV and cleaning. The new units are technically correct and visually the odd ones out, because the reference on site is no longer the reference in the file.

The fire-retardant package moving the hue. Halogen-free intumescent systems are typically loaded at 4–8% and push a green toward yellow while dropping gloss. A supplier who changes FR route between batches changes the color without changing the "color code" on the order.

Gloss read as color. The same pigment at 8 GU and at 25 GU reads as two different greens at 20 m under a linear light run. In a glazed terminal the specular highlight does the rest.

Each of these ends in the same place: re-leafing canopy panels, airside access permits, escort hours, and a lead time measured in weeks — for a defect that a two-line clause on any artificial tree order would have prevented.

2. Material Science & Structural Engineering Standards

Color in an injection-molded leaf is not a coating. It is pigment dispersed in the polymer, and every other additive in that polymer can move it.

Masterbatch and let-down. Pigment is supplied as a masterbatch and dosed into the base resin at roughly 2–5%. The carrier resin has to be compatible with the base — a masterbatch built on the wrong carrier disperses unevenly, and uneven dispersion shows up as streaking on the leaf face rather than as a measured shift. Pigment dispersion is checked by filter pressure value on incoming lots, and it is the reason two "identical" greens can differ before anyone touches the mould. The same discipline applies to a date palm crown, where one frond presents a large uninterrupted surface and a dispersion fault reads at ten metres.

Opacifiers and stabilisers. Titanium dioxide at 2–6% gives hiding power; raising it lightens the mix and flattens the shadow zone. The UV package — HALS at 0.3–0.8% plus a UV absorber — carries a slight warm cast of its own. When pigment and stabiliser are dosed from separate masterbatches, they can drift out of step across a long run: the color holds while the fade path changes.

Fire-retardant chemistry. An intumescent additive at 4–8% loading lowers gloss and shifts hue. A topical back-coating leaves a film that changes sheen and wears unevenly at handling points, so the same tree can show two gloss levels after a year of cleaning.

Measuring a leaf, not a panel. A molded olive leaf is 25–45 mm long with a raised midrib and a gradient from base to tip. A single spot reading on it is noise. A workable protocol: three zones per leaf (light face, mid tone, shadow edge), five leaves per tree, ten trees per batch, 8 mm aperture, 45/0 geometry, D65/10° observer, values recorded as L\a\b\* and reported as ΔE per ASTM D2244.

Two curved trunks in brown finish outside the factory, photographed in daylight — the finish that has to be matched, not just the leaf green.

Accept and reject numbers. Batch against batch, side by side at 1.5 m under D65: ΔE ≤2.0. Batch against the signed golden sample: ΔE ≤2.5 on delivery, with ≤3.0 allowed after accelerated exposure. Visual backup on the ISO 105-A02 grey scale at grade 4–5, and a crocking check to ISO 105-X12 at grade 4 dry and 3–4 wet, which catches pigment that transfers to a cleaning cloth.

Retention, not memory. Three sets of leaf swatches at 300 × 300 mm plus one complete branch, sealed, signed and dated: one held by the buyer, one by the factory, one with a third party. Re-approve on a 12-month cycle. A swatch ages differently from an installed canopy, so the file needs both the original standard and the last measured site reading.

3. Life Safety, Fire Codes & Environmental Compliance

A color change is a compliance event, not only an aesthetic one. NFPA 701 reports are issued for a specific construction — fiber, weight, FR chemistry and finish — and pigment loading is part of that construction. Moving from a light olive to a dark olive at the same FR loading usually stays inside the tested family; changing the FR route or the coating to reach a color does not, and the report no longer describes the product being shipped. The practical request is the description page of the report, with the colour range and FR chemistry written on it, not just the pass line.

Under EN 13501-1, classification is granted to a product family whose composition and FR system are defined in the test report. Class B-s1,d0 is achievable for treated PE foliage, but the family definition decides whether a new color sits inside it. Ask for that definition in writing before approving a color that was not in the original submission.

Indoor air quality follows the same logic. Solvent-borne topcoats used to adjust sheen are the usual source of odor complaints in enclosed terminals with recirculating HVAC. Water-based or in-mold color avoids the issue; where a topcoat is used, the VOC and SVHC statement should reference the coating batch, not the finished tree alone.

One more point that belongs to the maintenance file rather than the submittal: cleaning agents attack gloss before they attack color. A caustic or solvent cleaner used on a matte leaf leaves a patch that reads lighter even though the pigment is intact. The cleaning instruction and the color specification have to be issued together.

4. Technical Comparison Table

The height column drives the core and the ballast; the color column does not move with height at all. It moves with exposure, with the number of phases, and with whether the tree will be read against a linear light run or against daylight. A five-star hotel atrium that pairs two trees side by side is the hardest case in the table, because any difference between them is exactly what the eye is looking for.

5. Structural Anchoring & MEP Coordination

The base plate is where color stops being a sample decision and becomes a fixed one.

An olive canopy is asymmetric. One face is fuller, one side shows more trunk, and the leaf angle sets how much specular light comes back to the viewer. When a tree is fixed on a chemical anchor through a base plate, its rotation is set by the bolt-circle position and the anti-rotation pin at the moment the resin cures. A 15–20° error in that template turns the fuller face away from the sightline, and the same approved color then reads flat and slightly grey from the passenger route. The template has to be set from the approved sightline, and the engineering calculations for the base should carry that rotation in degrees, not only the bolt circle, so the installer has a number to work to. That also fixes the sequence: the color approval is signed before the anchor holes are drilled, never after.

Ballast collar planters keep that freedom. A tree standing in a weighted collar can be rotated after installation and re-aimed once the space is lit, at the cost of a floor load of 90–560 kg depending on height and exposure. In a commercial atrium where tenant lighting is commissioned after handover, that rotation is often the difference between one row and two greens. Where the slab cannot take an anchor and the tree is above roughly 6 m, the collar is usually the only compliant route; where an anchor is feasible, movement after curing is not.

The lighting interface matters as much as the fixing. Terminal downlights at 3000 K and CRI 90 render an olive green warmer than the 4000 K line lighting used in the same building. Two identical trees can read differently under the two systems, which is a lighting problem that gets reported as a color problem. Ask for the CCT and CRI of the final installation, and approve the swatch under it.

Air movement completes the picture. Sprinkler throw and high-velocity air curtains polish the leaf face and dry it faster after cleaning, so a dust film builds and clears unevenly along a row. On a glazed roof, laminated glass passes most of the 380–400 nm band, and that is the band that drives fading — an interior tree under a glass roof needs the same UV-stabilised grade as an outdoor one.

Moulded leaves laid out in bulk for sorting on the workshop floor, with rolled trunk pieces alongside.

6. Commercial Contractor FAQs

Q: How should color tolerance be written into an artificial tree purchase order?

State the measurement, not the adjective. Name the reference (a signed golden sample held by both parties), the instrument geometry (45/0, 8 mm aperture, D65/10°), the sampling grid (three zones per leaf, five leaves per tree, ten trees per batch), the limit (ΔE ≤2.0 batch to batch, ≤2.5 against the sample on delivery), and the consequence (re-leafing at the supplier's cost). A clause that says "color to match approved sample" has no rejection mechanism.

Q: Can a second phase be matched to trees that have been installed for two years?

Yes, but the reference has to change. Measure the installed canopy on site with a portable spectrophotometer, average ten trees, and issue that reading as the phase-two standard. Matching the new batch to the original file produces trees that are technically correct and visibly newer. Where the measured drift exceeds about 4 ΔE, the honest options are re-leafing the older trees or accepting a deliberate contrast.

Q: Does a color change require a new NFPA 701 test?

It depends on what else changed. Within a defined product family, at the same FR chemistry and the same fiber and weight, a pigment change is usually covered by the existing report. Change the FR route, add a topical coating to reach a deeper tone, or move to a different base polymer, and the tested construction no longer matches what ships. The report's description page is the document that settles it.

Q: Why do two trees that matched in the sample room look different in the terminal?

Three common causes: metamerism between a 6500 K booth and 3000 K downlights, a gloss difference of 15 GU or more on the same pigment, or two masterbatch lots separated by a production gap. All three are measurable before shipment, and none of them are visible in a photograph.

Project Engineering Support & CAD Requests

We issue the color package with the shop drawings rather than after them: signed leaf swatches, the batch code list per tree, the spectrophotometer readings for each production run, and the FR report description page that names the color range. Our engineering team will check a proposed ΔE limit against the resin and FR system before it goes into a contract, and we hold retention samples for five years so a later phase has something real to match against. Send the lighting schedule and the installation sequence through WhatsApp and the color clause can be settled before the anchor template is set.

👉 Chat with Structural Engineering on WhatsApp (Instant Project Review)