Solid deodorant stick: the hardness a twist-up needs on a hot commute

A deodorant stick at 18 percent beeswax and 4 percent candelilla with a 24 percent powder load, in a 70 ml twist-up poured at 76 to 82 C, tested to 40 C.

A deodorant stick asks more of an anhydrous formula than anything else a home maker attempts. It carries a quarter of its own weight in insoluble powder, stands 6 mm proud of the tube without support, survives a bag on a 30 C morning, and still lays down a film in a two-second stroke. This page gives the formula, the reason each number sits where it does, and the temperature at which it stops working.

• Total wax: 22%
• Powder load: 24%
• Pour: 76 to 82 C
• Fill: 62 g in a 70 ml tube
• Cosmetic, not an antiperspirant

What a twist-up has to survive, in numbers

A tin balm is held by its container. A stick is not. Once the elevator advances, the exposed 5 to 8 mm is a cantilever loaded sideways by the application stroke, and it fails either by shearing at the rim or by smearing product onto the tube edge. That sets a floor on hardness. The same stroke sets the ceiling: a stick hard enough never to break will not deposit anything, and the user reports it as doing nothing.

No cosmetic standard test covers this. Needle penetration under ASTM D1321 was written for petroleum wax and is only indirectly useful, since it measures a homogeneous wax rather than a filled oleogel. What you can do at the bench is fix three pass criteria and apply them to every batch:

• Advance and hold. Wind the stick out 8 mm at 22 C. It supports its own weight indefinitely and does not lean.
• Stroke test. Six firm passes across a forearm leave a continuous film, no crumbs at the tube rim and no visible flat spot worn into the dome.
• Drop test. Capped, from 1 m onto a hard floor. The stick may bruise. A stick that shatters inside the barrel has too much combined wax and powder.

Those are sensory criteria, and they repeat only as well as the person applying them. The instrumented alternatives, penetrometer and cone, are in measuring balm hardness.

The formula

All figures are weight percent of the total batch and add to exactly 100. The gram columns are for a 200 g trial batch, which fills three 70 ml tubes with enough left for the top-up pour, and for a 500 g batch, which fills seven.

A nominal 70 ml twist-up does not hold 70 g. The elevator and screw occupy the base, so the usable column is nearer 66 ml, and at a set density of about 0.94 g/ml for a powder-loaded stick that is 62 g. Weigh a filled tube before you commit to a fill weight on a label: tube geometry varies more between suppliers than the stated volume suggests. Closures and container conversions are in the packaging guide.

Why the wax is 22 percent, and split two ways

Twenty-two percent total wax is half again what a firm lip balm carries, and the extra is not there to make the stick harder to a fingernail. A filled system needs more continuous phase to bind. Powder particles do not join the wax network; they sit inside it and reduce the cross-section of wax that carries load. A 22 percent wax stick with 24 percent powder handles like a 14 to 15 percent wax stick with no powder at all.

The split matters as much as the total. Beeswax melts across roughly 61 to 65 C and yields before it breaks, which is what stops the cantilever snapping. Candelilla melts higher, near 68 to 73 C, and it is the component that decides whether the stick survives a warm bag. Four percent lifts the softening behaviour of the whole stick by a few degrees. Ten percent would too, and would also turn it glassy and prone to shattering in the drop test, because candelilla builds a stiff, brittle crystal network rather than a compliant one. That trade between wax types is the whole reason the structure is split rather than carried by one material.

The candelilla also sets the pour band, which is the point most recipes get wrong. You cannot pour this at 70 C. Below about 74 C some of the candelilla has already crystallised, and it seeds a coarse, gritty set that shows as a rough matte dome and a scratchy first stroke.

The powder load, and why it has to come down when the wax goes up

This is the distinguishing problem of the format. Odour control scales with powder, and structure scales with wax, but wax and powder both add to a single quantity that the stick cannot exceed: total non-deformable solids. Push either one up and the stick moves along the same path, from firm, to short, to crumbly, to a thing that breaks off at the rim on the first stroke. That is why a heat-resistant summer version cannot also be the high-powder version. Raising wax to hold 35 C uses part of the same budget the powder was spending.

Particle size sits underneath all of this. Coarse crystalline bicarbonate at 10 percent feels scratchy where a milled grade at the same percentage does not, most obviously on skin shaved that morning. Starch choice matters less: arrowroot and tapioca behave almost identically in an anhydrous base, and both are less abrasive than most clays, as compared in starches and silica.

Coconut oil at 25 percent, and the 24 C problem

Coconut oil earns its place as the only cheap material in the blend that is a hard fat in the tube and a thin liquid on skin. It melts over roughly 24 to 26 C, so it contributes real structure at 20 C and none at 30 C. That sharp transition is both its usefulness and its risk.

Run the arithmetic on solid content. At 20 C the stick's solids are 22 percent wax, most of the 25 percent coconut, and perhaps 40 to 50 percent of the shea, which is around 50 percent of the mass. At 30 C the coconut has gone entirely and the shea has given up most of its stearin, leaving little more than the 22 percent wax to hold the column up. A quarter of the formula changes state across a ten degree window that includes normal room temperature. That, not the wax, is why so many home-made deodorants that felt perfect in March collapse in July. The grades, the lauric fraction and the comedogenicity argument are covered in coconut oil in balms, and they are the reason a formula that behaved in March collapses in July.

The summer stick: the same product without coconut oil

Take the coconut out and you must replace both its solid contribution below 24 C and its liquid contribution above. Shea and a little extra wax do the first, jojoba and sunflower the second, and the powder drops to 20 percent to keep total solids inside the budget above.

The summer stick is firmer at 20 C, applies with less payoff, holds its shape roughly five degrees higher and costs more, because jojoba is dear per gram. To change one variable rather than swap formulas, put the percentages into the formula converter and hold the wax to oil ratio fixed while you move the powder.

Method

1. Weigh phase A. Both waxes, the coconut oil, shea, sunflower and castor into one stainless or heatproof glass jug, tared between additions. Sift phase B into a separate bowl to break any lumps. Weigh phase C into a small dish.
2. Melt at 78 to 82 C. Stand the jug in simmering water. Hold until the liquid is clear with no candelilla haze, which takes longer than beeswax alone. Do not push past 90 C.
3. Add the powders hot. Sift phase B into the melt in three additions, stirring each in fully before the next. A lump of bicarbonate dropped in whole never disperses, and ends up as a gritty spot in one tube.
4. Stir until it thickens. Off the heat, stir continuously. At around 84 to 86 C the suspension is thin and the powders sink in seconds. Keep stirring as it cools and you will feel the viscosity climb sharply near 82 C, which is the point the powders stop settling.
5. Add phase C at about 78 C. Stir in the tocopherol and fragrance immediately before pouring and accept a small loss of the most volatile notes. There is no window in this formula where the melt is both below 60 C and still pourable.
6. Pour at 76 to 82 C. Wind the elevator down, hold the tube upright and fill to 2 mm below the rim in one movement. Stop and restart and you leave a seam that later becomes a break line.
7. Top up after 20 minutes. A powder-loaded stick shrinks more than a plain balm. Reheat the residue to about 80 C and pour a thin second layer to give a level, slightly domed top.
8. Cool upright and cure 48 hours. Room temperature, uncapped until firm, then capped. Judge hardness at 48 hours, not two, because both waxes are still crystallising. Why the window is so narrow is explained at pour temperatures.

Bicarbonate: dose, latency and the swap

Sodium bicarbonate works, and it is also the commonest reason a home-made deodorant gets abandoned. Craft formulas run it from 5 to 25 percent, and reports of burning, redness and a raised, itchy patch in the armpit fold cluster at the upper end. Ten percent is a compromise: enough to do something, low enough that most skin tolerates it.

The mechanism is dose of dissolved alkali, not pH on a label. Bicarbonate is freely soluble, around 9 to 10 g per 100 ml of water at 20 C, so it dissolves completely in sweat and presents all of its buffering capacity at once against skin that normally sits between about pH 4 and 6. A fat base holds it there, the site is occluded, and it is often shaved. The pattern is delayed: comfortable for one to three weeks, then a rash over a couple of days that clears within days of stopping. It is a cumulative irritant reaction, which is why the first fortnight tells you nothing.

Two responses are sensible. Patch test before committing a whole formula, using the method in how to patch test a balm, while accepting that a 48 hour patch cannot reproduce a three week cumulative exposure. Or swap the material: magnesium hydroxide substitutes close to one for one by weight and is around four orders of magnitude less soluble, so it cannot deliver a large dissolved dose however much is present. The full replacement, with zinc ricinoleate, is at bicarb free deodorant balm. If the reaction persists without bicarbonate, suspect the fragrance first and the shaving second.

Heat testing at 30, 35 and 40 C

Test the stick, not the bulk. A jug of the same formula tells you little, because the failure you care about is geometric: a column losing the ability to stand up. Use finished, cured tubes, capped and upright, in a still oven with a probe thermometer beside them. Let each equilibrate for the stated hold, then advance it 8 mm and apply it warm.

Record what happens after the excursion, not only during it. A stick that slumped at 40 C and reset overnight is not restored: the wax recrystallised without shear on a different cooling curve, and it comes back harder, duller and slightly grainy. That is the mechanism behind balm changed after shipping, and it means a complaint can describe a batch that left you correct.

For a small seller the honest position is that neither formula posts safely through a European or North American summer. Insulated mailers and cold packs buy hours, not days, so the useful lever is dispatch timing rather than formulation, as set out at shipping in hot weather. Past 40 C the fault stops being cosmetic and becomes balm melting in the heat.

Failure modes

What this formula cannot do

It cannot stop you sweating, and it will not perform identically on two people, because axillary odour depends on an individual's skin microbiota as much as on the product. Anyone who has reacted to bicarbonate before should make the magnesium hydroxide version. It does not go on broken or freshly shaved skin, and it treats nothing.

The decision rule is short. If the product will live indoors below 25 C, make the base formula and take the payoff coconut oil gives you. If it will be posted or carried in a bag in a hot climate, make the summer variant and accept a drier application. If you want the heat resistance and the 24 percent powder load together, the format is wrong: use a 60 ml jar, where the container carries the load and powders can run to 30 percent without the stick geometry ever being tested.

Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.