Two weeks apart, two manufacturers exported what their paperwork described as much the same product: a pultruded carbon fiber composite core, the load-bearing element inside overhead power conductors. The case records, reported on 14 October and again on 5 November, end in opposite places. One shipment was detained, classified as a controlled dual-use item, and penalized for moving without an export licence. The other was released on the same day its owner obtained a formal determination that the goods fell outside the control list.
The end use was not the difference. Both went into civil transmission lines. The difference was whether the exporter could put a parameter-level argument on the table, item by item, before the goods were detained.
What follows is our assessment of that gap, built from the two published records and from the way the control list is actually structured. It is analysis, not a legal opinion, and not a statement of fact about any individual shipment.
Two Shipments, Two Opposite Outcomes
The first case surfaced in mid-October. Officers held a carbon fiber composite core, examined it, and placed it inside a control entry covering carbon fiber composite cores for structural use. No licence had been applied for. The penalty that followed attached to the exporter, not to the buyer and not to the forwarder.
The second case was disclosed three weeks later, and it ran the other way. Goods of the same functional description, a stranded carbon fiber composite core for overhead conductors, were questioned at the border and held. The exporter’s representatives then assembled a file along five separate lines: the technical properties of the goods themselves, the applicability of the specific entry, a formal business consultation with the commerce ministry, the correspondence between goods held by different trading entities, and the customs identification procedure. On 23 October the determination came back clean. Not a dual-use item. The same day, the documents cleared and the goods left.
Set the two files side by side and the variable is not the product. It is the quality of the technical description the exporter could produce inside the first week.
End Use Cannot Settle The Question
Most export departments work from a simple assumption. If the buyer is a utility and the part ends up on a transmission tower, the goods are civil, and civil goods are not controlled. That reasoning collapses on its own terms, because the control list is not written in the language of end use.
A control entry describes a material state, and then a set of numbers. What the fiber is. How far it has been consolidated. What the modulus and tensile strength measure, what resin it carries, what physical form leaves the gate. End use enters the analysis late, and it enters as one input among several. It is never a standalone exclusion. That misunderstanding gets expensive, because it makes exporters skip the technical file at the exact moment the technical file is the only thing that decides the case.
Consolidation state is the variable people miss most. The same carbon fiber can leave a plant in four commercially distinct forms, and each one gets examined against a different part of the list.
| Form as it leaves the plant | What the entry examines | The question asked at the border |
|---|---|---|
| Dry fabric, woven or non-crimp | Areal weight, width, weave pattern and tow size | Which reinforcement description applies, and does the declared areal weight match the roll? |
| Tow, yarn or chopped strand | Filament count, sizing chemistry, cut length | Is this intermediate reinforcement, or already a semi-finished article? |
| Unidirectional tape or fabric prepreg | Resin system, resin content, measured modulus and strength | Which prepreg entry was checked, and are the measured values on file? |
| Consolidated laminate or finished part | Stack sequence, thickness, cure or consolidation state | Is this an article rather than a material, and who is the producer of record? |
None of those four rows is inherently safer than the others. A dry fabric is not “less controlled” than a prepreg. It is controlled, or not, against a different description. The error is assuming that moving one step up or down the value chain changes the answer, when in practice it changes the question.
The Parameters Behind The Decision
A second assumption does most of the remaining damage: “our fiber is industrial grade, not aerospace, so the threshold cannot apply.” The published entry for aircraft-grade carbon fiber prepreg sets its threshold at a tensile modulus of 50 GPa and a radial tensile strength of 500 MPa. In composite terms those are modest numbers. Most structural carbon fiber clears both comfortably, including plenty of material sold into industrial work that never goes near an aircraft program.
So the parameter test does not separate aerospace from industrial. It separates characterized material from uncharacterized material. An exporter holding a measured modulus, a measured tensile strength, a fiber areal weight and a resin content, all traceable to a batch number, has a technical position. An exporter answering “it is just standard carbon fiber” has an opinion. Those two replies are not the same reply, and the difference is not a matter of tone.
Form and width carry as much weight as grade. A fabric at 1270 mm and a slit tape at 25 mm can hold identical fiber and an identical resin system and still be declared differently, because the physical form changes how the goods are identified and which entry applies. That is not a loophole to work. It is a documentation burden to carry. Whatever the form, the description on the declaration has to match the goods on the pallet rather than the goods in the catalogue.
Resin system earns its own line. A thermoplastic system and a thermoset system with the same reinforcement can sit on different entries, and a declaration reading only “carbon fiber prepreg” leaves the officer to assume the more conservative case. Naming the polymer, the nominal resin content and the consolidation route costs nothing. It removes an assumption that will otherwise be made against you.
The Paperwork Gap That Decides Cases
The second case was won on procedure as much as on physics. Two failure points recur often enough to treat as standard risk.
The consultation applicant is not the exporter of record. When the entity requesting a ruling differs from the entity named on the export declaration, a reply covering a similar product does not automatically extend to the shipment in front of the officer. The gap has to be closed with a goods-consistency statement mapping producer, product name, model, technical parameters, manufacturing state and transaction documents onto each other, line by line. Similarity of product name is not the test. Correspondence of goods is.
The reply arrives after the detention. Where batches leave through different ports, a ruling obtained for one batch can land weeks after a second batch has already been held. At that point the shipment sits inside the identification procedure, and both the clock and the cost change. A file that would have taken days turns into a process measured in weeks.
There is a timing rule that gets overlooked. Once goods are questioned, the supporting material has to arrive inside a defined window. In the published account that window is seven working days. Miss it and the file is decided without your input. Seven working days is not enough time to commission a tensile test. It is enough time to retrieve a test report you already hold, which is the whole argument for keeping batch data on file instead of in a drawer at the plant.
What To Put In Your Sourcing File
Four documents do most of the work, and none of them is expensive to produce.
- A parameter-to-entry map. One row per model sold: fiber type and tow size, areal weight, width, resin system, measured modulus, measured tensile strength, consolidation state. Beside it, the specific entry that was checked and the conclusion reached. A product manager can build this in an afternoon, and it answers most of the questions that arrive later.
- A goods-consistency statement, kept current. Producer, product name, model, parameters, manufacturing state, plus the invoice and packing list references. Update it when a model changes, not when a container is held.
- Batch-level test data travelling with the goods. Areal weight, resin content, and where relevant porosity and interlaminar shear. A measured sheet turns a claim into evidence at the moment evidence is being asked for.
- A suspension clause in the contract. If a licence or a determination does not come through, delivery dates shift and neither side carries a penalty. Buyers accept this more readily than sellers expect, because the alternative is a held container and an argument about who pays for it.
One more item belongs in the cost model rather than in the file. Where a prepreg line carries a zero export rebate, that rebate cannot be treated as recoverable when a landed price is built. Quotes assuming a standard composite-goods rebate quietly overstate margin on exactly the grades most likely to be questioned. Aircraft-grade prepreg is the obvious case, since the published thresholds are low enough to catch most structural material.
For buyers, the question worth asking is not “is this controlled?” It is “which entry did you check, and what were the measured values you checked against?” A supplier who answers on one page is a lower-risk supplier than one who answers with an assurance. The difference rarely shows up in the price. It shows up in the first week of a dispute.
We supply carbon fiber fabrics, thermoplastic prepregs and laminates, and each grade carries its measured areal weight and resin content through to the shipping documents, which are the same numbers a customs officer or a customer’s quality engineer will eventually ask for. If you are assembling a specification file for a carbon fiber part and want the measured values written down before you commit, send us the model and the target parameters through our contact form. Our carbon fiber fabric selection guide is a reasonable place to start on the fabric side.