Buying Corrugated Cartons in China Without Getting Burnt on Strength

The first carton order I placed from China was for a client selling cast aluminium garden fittings. I specified "double wall, strong" and got a container of boxes that felt solid on the dock in Foshan in March and folded down the middle by the time they reached a warehouse in Rotterdam in July. The boxes were not defective. My specification was.

Corrugated board is the one category where the buyer almost always owns the failure, because the factory builds exactly what the drawing says and the drawing says nothing. These are the things I now put in writing before a single sheet is cut.

ECT or burst, not "five layer, strong"

China exports board graded two ways. The US-influenced route is ECT, edge crush test in kN/m or lb/in, which predicts stacking strength reasonably well. The older route is Mullen burst in kgf/cm² or psi, which measures puncture resistance and correlates poorly with compression. A factory will quote whichever you ask for, so pick one and stay consistent.

Practical mapping I have found reliable for goods under 20 kg per carton:

Single wall, B flute, ECT 32 lb/in (about 5.5 kN/m): light retail items, individual units inside an outer.
Single wall, C flute, ECT 44 (about 7 kN/m): the workhorse for small parcel shipping.
Double wall, BC flute, ECT 51 to 61 (about 8.7 to 10.4 kN/m): anything over 15 kg or stacked three high in a container.
Triple wall, ECT 71 plus: machinery and export crates where the box is part of the structure.

Note that the Chinese domestic grades are often written as five layer or seven layer, and a five-layer box can be anything from ECT 40 to ECT 60 depending on liner grammage. Layer count is a construction description, not a strength grade. Get the ECT number.

The liner grades hide the real cost

Board is liner plus medium. A five-layer box built with 120 g/m² test liner on the outside, 90 g medium, and 110 g on the inside looks identical to one built with 175 g kraft on the outside and 140 g medium. Both may pass the printed ECT on a dry day, and they behave very differently after a month in a humid container.

What I ask for now: liner grade named (kraft, test liner 1, test liner 2, or white top), grammage for each ply, and the flute profile. For export cartons going into a sea container I default to at least 175 g kraft outer and 140 g medium on the load-bearing plies, and I accept that it costs 15 to 25 percent more than the cheapest quote. That premium is smaller than one damaged pallet.

Humidity is the single biggest derating factor

Corrugated loses roughly 40 to 50 percent of its compression strength between 50 percent and 90 percent relative humidity. A container crossing the equator routinely hits 85 percent RH inside for days, and a warehouse in a monsoon climate is no better. If your box needs ECT 44 in Rotterdam in January, it needs to be specified as if it were working at ECT 55 in dry conditions, or treated.

Two mitigation routes. Heavier board, which is what most people do. Or moisture-resistant treatment, which in China usually means a water-resistant adhesive (often called "waterproof glue", starch with a resin additive) and sometimes a wax or acrylic coating on the liners. Ask which one the factory means when they say waterproof, then put a sample in a sealed bag with a wet sponge for 72 hours and test compression again. I do this with every new factory and it has never once matched what was claimed.

Also worth knowing: the pallet matters as much as the box. A box with a stated 400 kg stacking limit on a pallet that flexes will fail at 200 kg. If you are shipping over 8 cartons high, specify the pallet type and check the deck board spacing.

Print, die-cut, and the tolerances nobody documents

Flexo print on corrugated is coarse by packaging standards. Registration is typically plus or minus 1.5 to 2 mm on a standard two-colour job, and fine text under 6 pt will fill in. If the artwork has a thin keyline or a small barcode, expect problems. Barcodes need a quiet zone and should be printed on the flattest panel, with at least a 1.5 mm bar width at 100 percent magnification.

Die-cut dimensional tolerance is usually plus or minus 2 mm on a box of 400 mm, which is fine until your product fits with 1 mm clearance. Internal dimensions should be specified as minimums, not nominals, and checked on the first article with a tape measure at three points per dimension.

Two more small things that have cost me real money. One, the flap gap. If the top flaps meet with a 3 mm gap on a badly set die, fine powder or small parts escape and the box looks cheap. Specify flap overlap of 2 to 3 mm and check it on the sample. Two, the score line. A score pressed too deep cracks the liner on a fold and the box fails at the corner within a week. Fold a sample flat and back three times; if you can see white liner fibre at the score, the die is too sharp.

What the paperwork should actually contain

For export to the EU or US, the two documents that matter are not certificates in the usual sense. First, a declaration that the board and any adhesive comply with the applicable food contact rules if the box touches food directly or indirectly, which for corrugated means the German BfR recommendation XXXVI or the US FDA 21 CFR 176.170 route, backed by a migration test if the product is fatty. Second, a statement on recycled content and heavy metals for packaging waste compliance; most factories can provide one because their liner suppliers already do.

Also confirm the factory's own ISPM 15 position if they supply pallets, and get the treatment mark on the pallet photographed at loading, not promised. A container held at destination for a missing IPPC stamp costs more than the entire carton order.

The sample protocol that replaced my intuition

Now I order three sets of samples from two factories: the cheapest quoted spec, the spec I want, and the spec one grade above. I load them with the actual product, stack them the way the container will, and leave them in a room at 30 degrees C and 80 percent RH for a week. Then I measure compression and open them. It costs around 400 dollars including freight and takes ten days, and it is the only reliable way I have found to see which quote is actually cheaper.