The short answer
Three published formulas answer this, not one. A sealed flexible bag works from interior surface area, a sealed rigid metal container works from volume, and moisture-sensitive electronic parts follow a third method again.
The familiar rule of 1.2 units per cubic foot is the metal-container formula. Applied to a barrier bag it answers a question nobody asked.
Every desiccant calculation starts with the same question and ends with a number, and the number depends on a choice the calculators make without saying so. Surface area or volume changes the answer entirely.
A unit is not a weight either. One unit is a performance threshold, so it works out to a different number of grams for silica gel than for clay.
This reference covers what a unit means, the three formulas, and the constants that go with each. It also covers the dunnage term nobody quotes and the bag spec the whole sum rests on.
What a Unit Actually Is
A unit of desiccant means a level of performance, not a mass, and the military spec sets two thresholds a material has to clear.
One unit adsorbs at least 3.0 grams of water vapor at 20 percent relative humidity, and at least 6.0 grams at 40 percent, both measured at 25°C.
Because materials take up water at different rates, a unit weighs whatever it takes to hit those numbers. One supplier reference puts a unit near 30 grams of silica gel, around 35 grams of clay and about 18 grams of fiber desiccant.
So ordering by weight and ordering by units are different orders. Swapping clay for silica gel at the same gram weight changes the protection.
Ordering gets muddled because packets sell both ways. A pack labeled in grams and a pack labeled in units are not the same thing until someone checks the material.
Three Formulas, Not One
Two of the three come from the same military standard, which publishes them side by side for different container types. The third comes from the electronics industry.
| Case | Source | Formula | Constant |
|---|---|---|---|
| Sealed flexible bag or wrap | MIL-STD-2073-1, Formula I | U = C × A | C = 0.011 per sq in, or 1.6 per sq ft |
| Sealed rigid all-metal container | MIL-STD-2073-1, Formula II | U = K × V | K = 0.0007 per cu in, or 1.2 per cu ft |
| Moisture-sensitive electronic parts | IPC/JEDEC J-STD-033 | U = 0.005 × A | simplified form, A in sq in |
A is interior surface area and V is enclosed volume, and both military formulas then add a term for dunnage, covered further down.
Carry the units through every time. The same letter C stands for 0.011 or 1.6 depending on whether the area arrives in square inches or square feet, and mixing them is a factor-of-144 mistake.
The container picks the formula. Two of the three come from one standard, which is how the wrong one travels.
The Cubic-Foot Rule Is for Metal Boxes
Search for this and the answer comes back as roughly 1.2 units per cubic foot of enclosed space, and that figure is real: it is Formula II.
Formula II applies to sealed rigid all-metal containers. Formula I, in the same standard, covers sealed flexible containers and runs on interior surface area instead.
The logic behind the split is simple. Moisture crosses a flexible barrier through its film, so the area of that film governs. A sealed metal box leaks mainly at its seals, and carries whatever air it shut in.
A flat bag with a large footprint and almost no depth exposes a lot of film and shuts in very little air. Volume-based math misses that case badly.
Put numbers on it. Take a bag 12 by 18 inches holding a part an inch thick, which gives roughly 490 square inches of interior surface and about 0.13 cubic feet of space.
Formula I turns that area into about 5.4 units. Formula II turns the volume into 0.15 units, and the two constants agree with each other whichever unit system you pick.
One answer is more than thirty times the other, and the bag has not changed — only the formula has.
Dunnage Carries Its Own Water
Both military formulas add a term for every pound of dunnage sealed inside the barrier, and the coefficients are not close together.
Cellulosic material takes 8.0 units per pound, while bound fibers take 3.6, glass fibers 2.0, and synthetic foams 0.5.
That is a sixteenfold spread between paper and foam. Wrap a part in kraft or sit it in plastic foam, and the same bag needs a very different count.
The upshot is that cushioning choices are moisture choices too. Picking a foam by packaging foam density sets the drop performance, and it sets part of the desiccant bill as well.
The Bag Sets the Answer
Desiccant holds only so much water, so the sum works only if the barrier lets water in at the rate assumed. That rate is a spec, not a trait of the plastic.
The military barrier specification, MIL-PRF-131, caps water vapor transmission at 0.02 grams per 100 square inches per 24 hours as received, and 0.03 after aging. Both limits apply after the material has been flexed.
That detail matters more than it looks. Flat film numbers flatter a bag that someone will crease, so the spec tests the creased state instead.
Flex conditioning has its own method, and ASTM F392 exists to evaluate flex-formed pinhole failures and the effect of flexing on transmission rates.
The same spec defines three classes, and Class 2 carries a limit worth checking before anyone orders it: combined contents inside the barrier must not exceed 10 pounds.
Seals matter as much as film. A bag with a perfect barrier and a weak seal is a bag with a hole in it, and the seal is where hand-made packs tend to fail.
Electronics ask for far tighter film. A dry pack bag under J-STD-033 must hold 0.002 grams per 100 square inches per 24 hours at 40°C, measured after flex conditioning. That is roughly ten times tighter, and its desiccant form asks about half the units per square inch.
Electronics Run on a Clock
For moisture-sensitive components the desiccant is one control among several, and the governing number is time out of the bag rather than units inside it.
Floor life runs by moisture sensitivity level, measured at factory ambient conditions at or below 30°C and 60 percent relative humidity.
- Level 1 — unlimited, at or below 30°C and 85 percent humidity.
- Level 2 — one year, dropping to four weeks at level 2a.
- Level 3 — 168 hours, then 72 hours at level 4.
- Level 5 — 48 hours, then 24 hours at level 5a.
- Level 6 — bake before use, every time.
Drying by an allowed method resets that clock, while running the parts through a reflow oven does not, which is the trap because the heat feels like it should count.
The same standard sends the desiccant back to the military definition, calling for material meeting MIL-D-3464 Type II. So the unit stays the unit even where the formula changes.
Indicator Cards and When to Bin One
A card inside the bag reports whether the system held, and the electronics standard asks for three spots, reading 5, 10 and 60 percent relative humidity.
Those cards were traditionally made with cobalt chloride, which a 1998 European directive classified as toxic with carcinogenic potential at stated concentrations. Cobalt-free versions now exist, several built on copper chloride.
Reuse is the part people get wrong, because ordinary cards are reversible, so a dried card reads low again and looks fine.
A card that has already reached 60 percent should come out of service, because that exposure spoils its accuracy at the low end — exactly the end that matters.
Cards cost very little, and replacing one costs less than guessing whether the last trip ruined it.
Common Questions
How many desiccant units does a barrier bag need?
Multiply the bag’s interior surface area in square inches by 0.011, then add the dunnage term for anything sealed inside. Working in square feet, the constant becomes 1.6 instead. Round up, since part units are not sold.
Is a unit of desiccant the same as a gram?
No. A unit is an adsorption threshold, so it lands near 30 grams for silica gel and around 35 for clay. Buying by weight across two materials changes the protection.
Can desiccant be dried out and used again?
Some materials dry out again with heat and some degrade, so the supplier’s data governs rather than a general rule. Inside a sealed pack the more useful question is whether the bag held at all, and that is what the indicator card answers.
Does desiccant replace baking for moisture-sensitive parts?
No. Desiccant keeps a sealed bag dry, while floor life governs the time those parts spend outside it. Once floor life runs out, drying is the only reset.
What these formulas are. They are published methods for specifying a quantity, not a prediction for one package in one warehouse. Storage time, humidity, seal quality and handling all move the real outcome. The standards behind these figures sit behind paywalls, so the controlling document governs any contract.
Key takeaways
- Three published formulas answer this question, and the container type decides which one applies.
- The familiar 1.2 units per cubic foot is the sealed rigid metal container case; flexible bags run on interior surface area at 0.011 per square inch.
- A unit is an adsorption threshold of 3.0 grams at 20 percent humidity and 6.0 grams at 40 percent, so its weight shifts with the material.
- Dunnage adds its own demand, from 0.5 units per pound for synthetic foam up to 8.0 for cellulosic material.
- For moisture-sensitive electronics the binding limit is floor life, and only drying resets it, never reflow.