Desiccants and Secondary Containment: Building a Dry Box Where the Air Is 80% RH

In a storeroom that sits near 80% relative humidity you do not control the room. You control a box inside it. The workable answer to humid-climate storage is a sealed, low-permeability enclosure holding a specified quantity of desiccant and a way to read the interior humidity without opening it. Room dehumidification is a capital project that works only while the mains supply holds. A dry box costs a fraction of that, needs no power at all, and its failure mode is slow and visible.

What follows is the physical build: which barrier to buy, how much desiccant a sealed volume actually needs, how to confirm the box is doing what you think it is doing, and where heat and the grid break the plan. Every figure below comes from a published standard or a measurement paper. Where the basis is a manufacturer's data sheet rather than a standard, that is said plainly rather than dressed up.

What 80% RH does to a storage specification

Stability data is generated against a climate, never in the abstract. Much of West Africa, the Congo basin, the East African coast and the Indian Ocean islands fall into climatic Zone IVb - hot and very humid - for which the long-term storage condition is 30 °C ± 2 °C at 75% RH ± 5% RH ^[1]. That number has a history worth knowing. ICH withdrew its own Zone IV guideline, Q1F, in 2006 and left the hot-and-humid conditions to be defined regionally and by WHO, which is why the 30 °C/75% RH condition now lives in a WHO annex rather than an ICH one ^[1]. The practical reading is blunt: a shelf life established at 25 °C and 60% RH was not established for your storeroom, and nothing about repeating the manufacturer's number makes it apply.

Dryness is the variable worth defending because water is not an inert passenger in a lyophilized solid. In a published study of a lyophilized antibody formulation, the measured glass transition temperature of the cake fell from roughly 80 °C at 1% residual moisture to roughly 25 °C at 8% ^[2]. That is an in vitro measurement on one formulated material, not a claim about anything you happen to be holding, and it is the only piece of degradation science this article borrows - the chemical routes that residual water opens up are a subject in their own right. What it carries here is a single line: the more water a solid takes up, the lower the temperature at which it stops behaving like a solid. In a room that reaches 35 °C by mid-afternoon, that margin is the entire argument for a dry box.

The barrier does the work; the desiccant only buys time

The most common failure is a handful of sachets thrown into a container that does not seal. Desiccant capacity is finite; a leaking enclosure is an unlimited load, and the sachets are quietly exhausted within weeks while everyone assumes they are working. Electronics dry-packing makes the relationship explicit: the quantity of desiccant per bag is calculated from the bag's surface area and its water vapour transmission rate so that the interior stays below 10% RH at 25 °C for the intended storage period, and the barrier film itself is specified at no more than 0.002 g per 100 in² in 24 hours at 40 °C after flex testing ^[3]. It is the barrier specification, not the sachet, that makes a small quantity of desiccant sufficient.

Pharmacopoeial practice measures the same property from the other direction. USP <671> classifies containers by measured moisture permeation, holding sealed containers charged with anhydrous calcium chloride at 23 °C and 75% RH for 14 days and reporting the water gained per litre of container volume per day against defined limits for the 'tight' and 'well-closed' classes ^[4]. If a supplier cannot say which class a container meets, treat it as neither and place it inside something that can.

A dry box is secondary containment, not a substitute for the conditions lyophilised material is meant to be held at in the first place. In practice the build is layered, and each layer has exactly one job.

• Primary container: the vial or bottle as received, seal undisturbed, kept in its original secondary packaging only if that packaging is itself dry.
• Barrier layer: a heat-sealed moisture barrier bag, or a rigid case with a continuous gasket and positive latches. This is the layer that must be specified rather than improvised ^[3].
• Desiccant: bought by adsorption capacity, placed so that it cannot make contact with the primary container ^[5].
• Indicator: a humidity indicator card readable through the barrier, or a logger recording from inside it.
• Outer layer: an insulated crate or cabinet whose job is thermal mass and physical protection, never moisture control.

How much desiccant, and of which material

Desiccant has a defined unit, and that unit is the only sane basis for buying it. MIL-D-3464E defines one unit as the quantity that, at equilibrium with air at 25 °C, adsorbs at least 3.00 g of water at 20% RH and at least 6.00 g at 40% RH; the weight constituting a unit may not exceed 50.0 g; and the specification's performance requirement is that a single unit keeps a closed metal test box below 60% RH ^[5]. Sachets are usually sold by gram weight, which on its own tells you nothing about capacity. Ask the supplier for units, or for the adsorption isotherm.

Sizing is an accounting exercise rather than a guess. Add up the water the enclosure will admit over the interval you intend to go without servicing it: permeation through the barrier (area × transmission rate × days), the water already present in the air you sealed inside, and - the term almost everyone forgets - the water held in whatever else went into the box. Cardboard, foam inserts, paper labels and printed inserts are all reservoirs that will hand their load to the desiccant. Remove them or condition them first. Then specify units with margin, because the box will be opened more often than the plan assumes.

MIL-D-3464 classifies bagged desiccant by dusting behaviour rather than by chemistry - Type I general purpose, Type II non-dusting - and dry-packing practice calls for the non-dusting type inside a sealed barrier bag, which is the sensible default anywhere near containers that will be opened ^[3][5].

Verify the box instead of trusting it

A humidity indicator card is the cheapest instrument that exists for this job, and it should be readable without breaking the seal. Dry-packing practice specifies cards carrying at least three spots at 5%, 10% and 60% RH and defines how quickly they must respond: at 30 °C and 60% RH the 5% spot must begin changing within four minutes and complete the change within seven ^[3]. One procurement caution. The classic blue-to-pink cards use cobalt dichloride, which the European Chemicals Agency placed on the REACH Candidate List of substances of very high concern in 2008 on carcinogenicity grounds, with the entry later updated for reproductive toxicity ^[6]. Cobalt-free cards are the current default, older stock still circulates in many markets, and disposal of the old kind is a waste-stream question rather than a bin question.

A logger inside the box gives you the record a card cannot. Inexpensive capacitive sensors drift, and they tend to drift most in precisely the conditions you bought them for, so check them against a fixed point. ASTM E104 is the standard practice for maintaining a constant relative humidity over a saturated aqueous salt solution in a small closed container, applicable from dryness to near saturation between 0 °C and 50 °C ^[7]. The fixed points themselves come from the NBS compilation: at 25 °C, saturated lithium chloride holds 11.30 ± 0.27% RH, magnesium chloride 32.78 ± 0.16% RH, and sodium chloride 75.29 ± 0.12% RH ^[8]. Sodium chloride is ordinary table salt, and it lands almost exactly on the Zone IVb reference condition. Run the check in a separate sealed jar - the standard warns that saturated salt solutions are corrosive and that corrosive vapour sits above them ^[7].

Then write it down, because a box nobody logs is a box everyone is guessing about. USP <1079> frames storage and transport as a risk-management activity: identify the risks, define the mitigation, monitor the conditions, and decide in advance what happens when they go out of range ^[9]. WHO's model guidance for the storage and transport of temperature-sensitive products expects storage areas to be mapped and monitored rather than assumed adequate, and its technical supplements cover that qualification work in detail ^[10]. Neither document necessarily binds a private laboratory in most African jurisdictions, and it would be dishonest to imply otherwise - but both are free to read, and both are what an institutional buyer, a tender reviewer or an inspector reaches for when they want a yardstick.

Heat, power and the limits of regeneration

Regeneration is where a tidy plan meets the grid. Silica gel can be reactivated, and the specification is explicit about what counts as success: after the prescribed reactivation, bagged desiccant must retain at least 90% of its original unit adsorption capacity and at least 80% of its original adsorption rate ^[5]. That describes a controlled process - a fan oven held at the manufacturer's stated temperature for the stated time, then resealing while the material is still warm and dry. A frying pan and an estimate will not reach that figure, and half-regenerated desiccant is worse than none, because it still reads as protection.

Then the constraint nobody writing from a temperate climate accounts for. World Bank Enterprise Survey data records the share of firms reporting electrical outages and the average number in a typical month, and across much of sub-Saharan Africa outages are a routine operating condition rather than an incident ^[11]. Any humidity strategy resting on a compressor that must run continuously - a room dehumidifier, an air-conditioned store - fails on the day the grid does, usually without anyone noticing until the next stock check. A desiccant enclosure has no power requirement, degrades slowly, and announces its own degradation on the indicator card. Where power is unreliable and oven time is uncertain, the honest decision is often to stop regenerating altogether and treat desiccant as a consumable with a scheduled replacement date.

Heat is a separate variable from humidity, and the two combine into the failure most humid storerooms actually suffer: condensation. Air at 30 °C and 80% RH has a dew point close to 26 °C, so any surface colder than that wets within seconds of exposure. A container taken from cold storage into that room and opened immediately collects liquid water on and inside it, and no desiccant sachet undoes that. The retrieval is not free in itself either, since repeated warming and refreezing carries its own damage mechanisms - which is the argument for taking material out as few times as possible. The order of operations matters more than the equipment: seal the enclosure before cooling it, and let a sealed enclosure equilibrate to room temperature while still closed. This costs nothing and is skipped constantly.

A minimum viable dry box

1. Pick the enclosure on a specification, not on appearance: a gasketed rigid case with latches, or a heat-sealed barrier bag with a stated water vapour transmission rate ^[3].
2. Strip the contents of cardboard, paper and untreated foam, or condition them dry before they go in.
3. Charge it with desiccant bought by units of adsorption capacity, with margin for the openings you will actually make ^[5].
4. Fit a cobalt-free humidity indicator card where it can be read through or beside the seal ^[3][6].
5. Add a logger inside, and check that logger once against saturated sodium chloride at a known temperature before you trust its record ^[7][8].
6. Write the sealing date, the desiccant quantity and the replacement date on the outside, and log every opening.
7. Set a fixed review interval, treat any indicator change as the trigger to reseal with fresh desiccant, and keep the record - it is the only evidence the storage condition was ever met ^[9][10].