Table of Contents
- Water activity decides whether you need a desiccant packet at all
- Choosing the packet type by food group
- Choosing the right bead: silica gel, clay or molecular sieve
- Dose: the packaging decides, not the amount of cake
- Food-contact rules
- Desiccant packets for mooncakes
- How to place the packet and common mistakes
- Frequently asked questions
- At what water activity does mold stop growing?
- What is a food desiccant packet?
- Which desiccant packet type should I use for dry food?
- Can a desiccant packet be used directly in food?
- How much silica gel does a 20 x 25 cm bag need for 30 days?
- How many desiccant packets should a mooncake box use?
- Should mooncakes use a desiccant packet or an oxygen absorber?
- Can blue desiccant beads be used for food?
- Does Vietnam have a dedicated standard for food desiccant packets?
- Contact ICD Vietnam for advice on desiccant packets
- References
- Related articles
Reviewed and updated by ICD - Week 47/2026
A desiccant packet can only keep food dry (water activity below roughly 0.6 to 0.7) inside sealed packaging; the dose depends on how much water vapor the packaging lets through far more than on how much food is inside. Mooncakes with a moist filling have an aw of 0.77 to 0.82, so they need another approach; a packet placed in food packaging must come with food-contact documentation.
This article goes from water activity to choosing a packet type by food group, how to estimate the dose from the packaging, food-contact rules, and mooncakes specifically. Figures come from the sources listed at the end; where a calculation is my own estimate, I say so on the spot.
Water activity decides whether you need a desiccant packet at all
Food does not spoil because of “humidity” in general but because of the free water that microorganisms can use, called water activity (aw, on a scale of 0 to 1). Multiply aw by 100 and you get the equilibrium relative humidity a food creates when it sits sealed in air at a constant temperature: an aw of 0.7 corresponds to 70% RH (USDA Pathogen Modeling Program). So when a desiccant packet pulls the humidity inside a package down, it also pulls down the aw at the food surface, provided the food was already dry.
The table of minimum aw for microbial growth (Fontana) shows that below 0.60 no microorganism multiplies; dry-tolerant molds such as Eurotium amstelodami and E. chevalieri start at 0.70 to 0.71, many Aspergillus and Penicillium species at 0.75 to 0.83, and the sugar-loving yeast Saccharomyces rouxii goes down to 0.62. Literature on bakery shelf life gives 0.65 or below as the level that avoids most mold problems. A desiccant packet therefore suits dry food with an aw below roughly 0.6 to 0.7 to keep it crisp and stop mold; for food with a higher aw, the water sits inside the food itself and the packet cannot pull it out.
The figure 0.85 is often misquoted as an FDA limit for baked goods. In fact 21 CFR 113.3 uses 0.85 to define low-acid food (pH above 4.6 and aw above 0.85) subject to canning regulations; it is a legal classification threshold, not a level that is “safe for bakery products”.
The final check is to measure aw with a water activity meter, or to test in practice: pack samples with and without a desiccant packet, keep them in a humid chamber and compare after 2 to 4 weeks. 30 °C and 75% RH is the long-term stability test condition for pharmaceuticals in the ASEAN region (climatic zone IVb); as far as I could find, food has no equivalent common test condition, but you can borrow this level as a harsh condition for comparing two samples.
Choosing the packet type by food group
| Food group | Typical aw | Main risk | How to use a packet |
|---|---|---|---|
| Biscuits, snacks, rice crackers, nuts | Biscuits about 0.3 | Lose crispness when they take up moisture again | Silica gel in sealed packaging; it can over-dry types that need to stay soft |
| Tea, coffee, powdered spices | Instant coffee about 0.2 | Clumping, loss of aroma | Silica gel or clay, packet placed separately, not mixed into the powder |
| Dried fruit | About 0.6 | Close to the mold threshold if it takes up moisture again | Silica gel; avoid over-drying types that need to stay pliable |
| Mooncakes | 0.77 to 0.82 (Cantonese style) | Mold, yeast, rancid oil | See the dedicated section below |
| Noodles, rice vermicelli, dry farm produce exported by sea | Depends on the product | Moisture enters the carton throughout the voyage | Large packet in a sealed PE-lined carton, plus a container desiccant for the outer carton |
| Bread, fresh baked goods, meat | Bread 0.95; fresh meat 0.99 | Microorganisms, not moisture in the air | A desiccant packet is not suitable; use refrigeration or an oxygen absorber, MAP |
The aw values in the table come from the USDA Pathogen Modeling Program list, except mooncakes, which come from a 2025 study. Nuts, tea and spices have an aw that varies by type, so measure a real sample before settling on anything. Dry-filled cakes such as sweet lotus seed or mung bean, with little water, belong to the group near the top, while cakes filled with salted egg yolk or mixed filling, rich in oil, go rancid from oxygen more than from moisture; see the mooncake section below.
Choosing the right bead: silica gel, clay or molecular sieve
The beads should only make indirect contact: they sit inside the packet, and the packet shell is the barrier against the food. For a detailed comparison of bead types see types of desiccant.
Silica gel works over a wide humidity range so it suits snacks, dried fruit and spices: the manufacturer states a capacity of more than 12% of its weight at 20% RH, more than 27% at 50% RH and up to about 35% at high humidity. The drawback is that it can over-dry soft food, and at low RH the beads hold less water. Montmorillonite clay takes up about 20 to 30% of its weight, works well at room temperature and loses effectiveness when hot, so it suits flour and grain in an air-conditioned warehouse. Molecular sieve (zeolite) takes up about 20 to 22% of its weight but holds humidity very low even when hot, which suits high-value goods; in exchange it over-dries easily and can only be regenerated at 200 to 315 °C.
Gel-type calcium chloride (the kind often used in containers) has a much higher capacity but can leak liquid and irritate the eyes, so it should not be placed next to food.
Dose: the packaging decides, not the amount of cake
Water vapor reaches a desiccant packet from three sources: the air trapped in the package at sealing, the water released by the food itself as the two reach equilibrium, and the vapor that permeates through the film and seals over the shelf life. The first source is tiny: 0.5 liter of air at 20 °C and 75% RH holds only about 0.006 g of water (17.3 g per m³ at saturation, CCI figure), while 1 g of silica gel holds 0.12 to 0.27 g. The third source is what decides how fast the packet fills.
Film permeability is measured as WVTR (grams of water through 1 m² of film per day). The table below uses values normalized to a 1 mil (25 micron) film at 38 °C and 90% RH. A film twice as thick lets through half as much, and a poor seal can wipe out the whole advantage of a good film.
| Film (25 micron) | WVTR, 38 °C/90% RH (g/m² per day) | Water entering a 0.1 m² bag over 30 days, estimate | Type A silica gel needed, estimate |
|---|---|---|---|
| Biaxially oriented PP (OPP) | 3.9 to 6.2 | 2 to 3 g | 8 to 12 g |
| HDPE | 4.7 to 7.8 | 3 to 4 g | 9 to 16 g |
| Cast PP | 9.3 to 11 | 5 to 6 g | 19 to 22 g |
| Oriented PET | 16 to 20 | 9 to 11 g | 32 to 40 g |
| LDPE | 16 to 23 | 9 to 12 g | 32 to 46 g |
Practical conclusion: a thin LDPE or PET bag of about 20 x 25 cm needs tens of grams of silica gel for 30 days, far more than the 1 to 5 g packets usually sold. To cut the dose, change to a better, thicker barrier film or press the seal firmly before adding more packets. That is why the line “one box uses 1 to 2 packets” cannot be trusted without the packaging type.
For export cartons lined with sealed PE, the dose calculation document is MIL-STD-2073-1E (US military packaging). For a sealed non-metal bag, the dose is U = 1.6 x bag area (ft²) + X x cushioning material weight; converted to metric that is about 17 units per m² of bag surface. One unit under MIL-D-3464E absorbs at least 3.00 g of water vapor at 20% RH and 6.00 g at 40% RH (25 °C). A 40 x 30 x 30 cm carton with a PE bag of about 0.66 m² needs around 11 units, that is, absorbs at least about 33 g of water at 20% RH. This formula was written for long-term storage of military goods, so I treat it as an upper bound for a food shipment delivered within a few weeks; general guidance is in how to use desiccant packets correctly.
Food-contact rules
The packet should only sit inside the packaging and not touch the food, and the packet shell must be strong enough not to tear so that beads do not fall into the food. Blue beads that turn pink contain cobalt chloride, which has been on ECHA’s Candidate List of Substances of Very High Concern (SVHC) since 28/10/2008 because of cancer and reproductive toxicity risk; Poison Control states clearly that indicating beads are not used to protect food or medicine. For food, choose packets without indicator and without cobalt.
In Europe, Regulation (EC) 450/2009 governs “active” materials and articles in contact with food, defined to include those that absorb substances from the environment around the food, and it gives as an example active substances held in small paper sachets. In my reading, a desiccant packet placed in food packaging falls within its scope. The regulation requires labeling in words (with a symbol if feasible) when the packet could be mistaken for something edible, and requires a written declaration of compliance to accompany the material.
In the United States, the FDA assesses the legal status of food-contact materials substance by substance, for each substance that could migrate into food: it must be listed in 21 CFR parts 174 to 179, be GRAS, have a prior-sanction letter from before 1958, be exempt under the regulatory threshold, or have an effective food contact notification. So the phrase “FDA compliant” on a packet says nothing without documentation for each component.
In Vietnam, QCVN 12-1 to 12-4 apply to plastic, rubber, metal and glass packaging and utensils in contact with food. I did not find a dedicated standard for desiccant packets. So when buying, ask the supplier for:
- A declaration or certificate of compliance tied to the exact product code and wrapping material
- Toxicity and composition test results
- The production process and quality control procedure
- For exports: paperwork for the destination market (EU, United States and so on)
The “do not eat” text on the packet should still be kept so users do not confuse it with a seasoning packet, and the packet is always kept away from children; see are desiccant packets poisonous.
Desiccant packets for mooncakes
Mooncakes come in two groups: dry fillings with little water such as lotus seed and sweet mung bean, and moist, oily fillings such as mixed filling and salted egg yolk. A 2025 study of Cantonese mooncakes measured an aw of 0.77 to 0.82 and classed them as intermediate-moisture foods. Against the minimum-aw table above, this range sits in the zone where many molds and yeasts can grow. Here the water sits in the filling, not in the air of the box, so by my reasoning no desiccant packet can pull the filling below the mold threshold. I have not found an aw figure for each type of Vietnamese mooncake filling.
A study by Kanpairo (2024) followed mooncakes for 42 days at 29 ± 2 °C in three kinds of packaging: a PP bag, a PP bag with an oxygen absorber, and a vacuum nylon-LLDPE bag. The two causes of spoilage named were water migrating from the filling to the crust and changing the texture, and oil going rancid. The vacuum nylon-LLDPE bag slowed microorganisms and fat oxidation best, giving a 35-day shelf life at room temperature while still meeting the Thai standard TIS 1454/2012. The difference from the other two was statistically significant (p below 0.05). The lesson is that good barrier packaging is worth more than one extra packet in an ordinary bag.
Oxygen absorbers also have limits. Toyama and colleagues (Japan, 1980) found that an oxygen absorber stopped mold completely for 5 weeks in semi-moist confectionery, but in bean-filled cakes with a relatively high aw, CO2 and ethanol from yeast still appeared, and high-aw cakes still spoiled from yeast and bacteria despite the oxygen absorber; the authors concluded that hygiene and microbial control are needed right in production. The difference between the two packet types is covered in oxygen absorber vs desiccant packet.
For the packer, the order of priority is: a recipe that keeps the filling aw low, production hygiene, good barrier packaging, and only then the packet. A desiccant packet works best when each cake is already sealed in its own bag before going into the box, in which case the packet deals with the humidity in the box, not the water in the filling. I have no reliable source for the number of packets per box, so this article gives no figure.
How to place the packet and common mistakes
Choose packaging with low water vapor permeability: a bag with an aluminum layer, a good barrier film or a plastic box with a tight gasket, because permeable packaging fills the packet quickly (WVTR table above). Place the packet above or beside the product, not pressing on the cake and not inside a bag of powder. A silica gel packet left out in humid air takes up moisture immediately before use, so take it out of its sealed bag only just before packing. A packet used in food should not use color-changing indicator beads; replace it by the calculated shelf life and check with an aw meter or hygrometer.

Common mistakes:
- Using blue cobalt beads for food.
- Using a desiccant packet for oily cakes while ignoring oxygen, barrier packaging and hygiene, so the cakes still go rancid.
- Leaving the packet in humid air for a long time before packing.
- Choosing permeable packaging and then adding more packets instead of changing the film.
- Overdosing and drying out food that needs to stay soft.
- Not keeping certification records, which causes trouble with exports or complaints.
ICD industrial desiccant packets (calcium chloride, clay) are for cartons and are not meant to go into food. For long export shipments, see the humidity threshold in safe humidity for storing goods and ICD desiccant packets for the outer carton.
Frequently asked questions
At what water activity does mold stop growing?
Below 0.60 no microorganism grows per Fontana’s table. The most dry-tolerant molds such as Eurotium start at about 0.70, and bakery literature gives 0.65 or below as the level that avoids most mold problems.
What is a food desiccant packet?
It is an activated silica gel or clay packet wrapped in paper or nonwoven fabric, placed in sealed packaging to keep food dry with indirect contact. The packet must have food-contact documentation from the supplier and must not use cobalt indicator beads.
Which desiccant packet type should I use for dry food?
Cobalt-free silica gel suits a wide humidity range; clay suits flour and grain in an air-conditioned warehouse; molecular sieve is for high-value goods needing very low humidity. A moisture barrier in the packaging matters just as much as the bead type.
Can a desiccant packet be used directly in food?
Only indirectly: the packet sits inside the packaging but is not mixed into the food, and beads must never be poured onto the food.
How much silica gel does a 20 x 25 cm bag need for 30 days?
It depends on the film. By the estimate in this article, a 25 micron OPP bag needs about 8 to 12 g, while a thin LDPE or PET bag of the same size needs about 32 to 46 g. This is the author’s estimate; measure in practice by weighing the packet before and after.
How many desiccant packets should a mooncake box use?
This article has no sourced figure. Cantonese mooncakes have an aw of 0.77 to 0.82, so a desiccant packet is not the main measure; prioritize the recipe, hygiene and good barrier packaging such as a vacuum nylon-LLDPE bag.
Should mooncakes use a desiccant packet or an oxygen absorber?
An oxygen absorber stops mold well in semi-moist confectionery but cannot stop yeast in high-aw fillings (Toyama 1980), and in a 42-day test a PP bag with an oxygen absorber did no better than a vacuum nylon-LLDPE bag (Kanpairo 2024). A desiccant packet only handles the humidity inside the box.
Can blue desiccant beads be used for food?
No. Blue beads that turn pink contain cobalt chloride, which is on ECHA’s SVHC list; Poison Control states that indicating beads are not used to protect food.
Does Vietnam have a dedicated standard for food desiccant packets?
I did not find a dedicated standard. QCVN 12-1 to 12-4 apply to plastic, rubber, metal and glass packaging. You need a written food-contact declaration from the supplier, and for exports follow the rules of the destination market.
Contact ICD Vietnam for advice on desiccant packets
ICD Vietnam Industrial Manufacturing Co., Ltd.
Northern Vietnam: 3rd Floor, Thang Long A1 Building, Thon Bau, Thien Loc Commune, Hanoi, 0983 797 186 / 090 345 9186 / 090 5859 186
Southern Vietnam: 551/212 Le Van Khuong Street, Tan Thoi Hiep Ward, District 12, Ho Chi Minh City, 098 6784 186
Email: sales@icdvietnam.com.vn
Zalo: Chat on Zalo now
References
- Aqualab, Water activity guide for food manufacturers - aqualab.com
- Neutec, The shelf life of bakery products simplified with water activity - neutecgroup.com
- USDA, Pathogen Modeling Program Online: Water Activity in Food - usda.gov
- Fontana A.J., Minimum Water Activity Limits for Growth of Microorganisms, trong Water Activity in Foods (Wiley, 2020) - doi.org
- 21 CFR 113.3, Definitions (low-acid foods: pH above 4.6 and aw above 0.85) - ecfr.gov
- ASEAN Guideline on Stability Study of Drug Product (long-term zone IVb condition: 30 °C/75% RH) - asean.org
- Poly Print, Water Vapor Transmission Rate (WVTR table normalized to 1 mil, 38 °C/90% RH) - polyprint.com
- Canadian Conservation Institute, Technical Bulletin 33: Silica Gel, Passive Control of Relative Humidity (2018) - publications.gc.ca
- Non-Indicating Silica Gel datasheet (capacity by humidity, regeneration at 120-150 °C) - farnell.com
- MIL-STD-2073-1E, Standard Practice for Military Packaging (Method 50, desiccant dose formula) - s3vi.ndc.nasa.gov
- MIL-D-3464E, Desiccants, activated, bagged, packaging use and static dehumidification (1987) - protectivepackaging.net
- Poison Control, Silica Gel (safety if swallowed, cobalt indicator, methyl violet) - poison.org
- ECHA, Candidate List of substances of very high concern (cobalt dichloride, 28/10/2008) - echa.europa.eu
- Commission Regulation (EC) No 450/2009 on active and intelligent materials and articles intended to come into contact with food - eur-lex.europa.eu
- FDA, Determining the Regulatory Status of Components of a Food Contact Material - fda.gov
- QCVN 12-1:2011/BYT, Hygiene and safety for synthetic plastic packaging and utensils in contact with food - nafiqad5.gov.vn
- Tseng L.-Y. et al., Enhancing the quality indicators and shelf-life of Cantonese mooncakes…, Applied Food Research (2025) - sciencedirect.com
- Kanpairo K., Effect of Package Type on Quality Change of Mooncake during Storage, Recent Science and Technology (2024) - tci-thaijo.org
- Toyama R. et al., Storage of High and Intermediate Moisture Confectionaries by Using Free Oxygen Absorber, Nippon Shokuhin Kogyo Gakkaishi 27(5), 1980 - jstage.jst.go.jp
