Freezing does not stop weight loss. In frozen cargo the temperature can stay on target while the product still loses weight. This is freezer burn. Its cause is not melting but sublimation: ice turns to vapour without passing through the liquid stage. According to USDA, freezer burn does not make food unsafe; it leaves grey-brown, leathery dry patches on the surface and lowers quality.
Sublimation: ice leaves without melting
The water in frozen product is in the form of ice crystals. Even at low temperature these crystals turn to vapour without passing through the liquid stage. The vapour leaves the surface and settles as frost on colder surfaces inside the packaging and on the container wall. Iowa State University's extension service explains that oxygen left in the package bleaches the surface, and that foods with a high water content are more prone to freezer burn.
Weight loss on a per-kilo sale
On a product sold by weight this is a direct loss, but the total weight of a tightly sealed pack does not change: the water inside the product turns to frost in the pack. The difference shows when the product itself is weighed with the frost removed, or when vapour leaves the pack. Example with invented figures: a 0.5% loss on 500 kg of net product is 2.5 kg; at 8 € per kilo that is a 20 € difference on one shipment.
Freezing and storage: where the loss happens
FAO's fish measurements give these losses by freezing method:
| Freezing method | Product | Weight loss |
|---|---|---|
| Air blast | IQF shrimp | 2 - 2.5% |
| Air blast | IQF haddock | 1.2% |
| Carbon dioxide tunnel | IQF haddock | 0.6% |
| Liquid nitrogen | IQF products | 0.3 - 0.8% |
| Contact (plate) freezer | Fish blocks | 0% |
FAO finds the claims that freezing itself produces freezer burn unfounded; freezer burn becomes visible after hours or days of storage.
Storage: same temperature, different loss
Unit: grams per square metre of exposed fish surface per day (g/m²).
| Store type | Average temperature (°C) | Daily loss (g/m²) |
|---|---|---|
| Pipe grid | -27.9 | 0.25 |
| Finned pipe grid | -25.4 | 2.30 |
| Unit cooler | -27.9 | 9.34 |
| Unit cooler | -29.3 | 4.96 |
| Jacketed | -15.0 | 4.06 |
| Unit cooler | -30.0 | 5.0 - 50.0 |
At the same -27.9 °C, a pipe grid store loses 0.25 g/m² a day and a unit cooler store loses 9.34 g/m²: a 37-fold difference (9.34 / 0.25). FAO lists temperature fluctuation, humidity, air velocity, heat transfer, packaging and product shape among the factors. Fish stored near a fan cooler dries quickly, and the edges and corners of fillets dry before the rest of the surface.
Swings speed the loss up
When the temperature rises, surface ice evaporates and the vapour moves into the empty space in the package. When the temperature falls, that vapour turns back to ice there. Each cycle grows the crystals and dries the surface further. The thermometer can show the target while the loss continues. In the logger history, look at peak values, time outside the limit and door openings, not at the average.
Glaze: the ice shell also goes on the scale
On fish and seafood the ice glaze protects the surface, but on the scale it is weight. Codex STAN 190-1995 excludes the glaze from the declared net contents and describes weighing the product after the ice glaze has been sprayed off with cold water. A 2010 study (Analytica Chimica Acta) measured 712 samples from 50 batches and found an average glaze of 8.7%. It set 6% as the lower limit for adequate protection and 12% as the upper level, above which glazing is excessive. None of the batches exceeded 12%.
| Glaze share | Net weight of a 10 kg gross load | Reading |
|---|---|---|
| 6% | 9.40 kg | Lower limit for adequate protection |
| 8.7% | 9.13 kg | Study average (712 samples) |
| 12% | 8.80 kg | Upper threshold; above it is excessive |
Calculation, with glaze as a share of gross weight: 10 kg × (1 - 0.06) = 9.40 kg; 10 kg × (1 - 0.12) = 8.80 kg. When glaze rises from 6% to 12% on the same load, 0.60 kg less is delivered, even if the product itself stays the same. The study also estimates that one percentage point of glaze is worth about 1 million euros a year in a market with low to moderate fish consumption, such as Belgium.
Codex defect and storage rules
Codex STAN 190-1995 defines excessive dehydration as a defect in frozen fillets: when more than 10% of the surface of the sample unit, or the defect area set for the pack size (150 cm² for units of 501-5000 g), shows excessive moisture loss, it counts as a defect. The same standard requires storage at -18 °C or colder on the label. If the target is -18 °C, the plate choice should also be made against that target.
How to measure the loss
- Pick samples of the same product code and packaging; record gross, tare and net weight at loading.
- At delivery use the same scale (or a traceable one) and the same tare method. For glazed product, spray off the ice glaze with cold water before weighing, as Codex describes.
- Measure the dried area and divide it by the total surface. Under Codex, a defect is more than 10% of the surface; compare the result with that limit.
- Write the scale ID, the lot and the surface finding on the form.
- If there is scale uncertainty or a packaging difference, do not report the difference as a definite product loss.
Limiting the loss: packaging and a steady temperature
The first step is tight, moisture-resistant packaging: the narrower the path for vapour, the slower the surface dries. The second is a transport environment that holds the temperature flat. While they melt, GEBHARDT eutectic plates hold the inside of the container close to their melting point. The plate table shows the options:
| Melting point | 390 × 295 × 34 mm | 480 × 280 × 34 mm | 615 × 378 × 34 mm |
|---|---|---|---|
| -3 °C | yes | yes | yes |
| -14 °C | - | yes | - |
| -18 °C | - | yes | - |
| -21 °C | yes | yes | yes |
With a -18 °C target, a plate melting at -18 °C works at the limit and leaves no margin; a -21 °C plate leaves 3 °C below the target, which is why it is the usual choice for frozen loads. Whether that margin holds for the product and route shows in the logger record. In a trial, product, route and packaging stay fixed and only the plate melting point changes.
