Grainy shea butter: the real mechanism, the fix and the prevention
Shea grain is beta-polymorph StOSt crystals that segregated from the olein and coarsened. The mechanism, a diagnosis table and a rescue procedure.
Grainy shea is the most common complaint in handmade balm, and the explanation repeated almost everywhere is chemically wrong. "The stearic acid separates out" describes something that does not happen. Shea's fatty acids are esterified into triglycerides, and nothing in a sealed jar of balm de-esterifies them. What separates is a class of whole triglyceride, and knowing which one tells you exactly what to do about it.
Grain is aggregates of beta-polymorph StOSt crystals: the high-melting stearin fraction of the butter crystallised largely on its own, segregated from the liquid olein, then coarsened by Ostwald ripening until a fingertip can feel it. Remelt well above 45 C, hold, then cool fast with stirring.
What grain actually is
Shea butter is not one fat. It is a mixture dominated by two triglyceride families: StOSt (stearic-oleic-stearic, often written SOS), which is high melting, and StOO, which is much lower melting. When a shea blend cools slowly, few nuclei form, and the StOSt-rich material crystallises first and largely alone. Dissimilar triglycerides do not co-crystallise happily. They form eutectics and phase separate, so the growing crystals are far richer in StOSt than the butter around them.
Those crystals are then unstable in the form they first take. Fats are polymorphic: the same molecules can pack in several different lattices, and they progress from loose, low-melting packings to tight, high-melting ones. Shea stearin held at 20 C is mostly beta prime after a single day and fully beta after a week (Ray and colleagues, 2013). Beta crystals are larger, denser and better formed. Finally Ostwald ripening takes over, with big crystals growing at the expense of small ones, helped along by every warm and cool cycle the product experiences. Hou and colleagues (2024) observed large feather-like beta crystals in shea stearin at the end of crystallisation, which is exactly the habit that reads as grit.
So the honest one-line mechanism is: segregation of StOSt, transformation to beta, coarsening by ripening. Retire "the stearic acid separates out". It is not merely loose language, it points makers at the wrong lever, because the fix is thermal and kinetic rather than compositional.
Why shea segregates when other butters do not
The reason is in the fractionation data. Crude shea splits industrially into a stearin and an olein, and the two fractions have genuinely different melting behaviour.
| Measure | Crude | Stearin | Olein |
|---|---|---|---|
| Palmitic C16:0 | 3.2% | 2.4% | 3.9% |
| Stearic C18:0 | 43.5% | 60.4% | 30.3% |
| Oleic C18:1 | 45.8% | 32.9% | 56.0% |
| Linoleic C18:2 | 5.9% | 2.9% | 8.1% |
| Slip melting point | 32 C | 36 C | 25 C |
| Unsaponifiables | 5.1% | 0.2% | 6.4% |
An eleven degree gap between the two fractions inside a single butter is the whole story. There is a wide temperature window in which the stearin is solid and the olein is still liquid, and any cooling that lingers in that window lets one crystallise out of the other. Crude shea is reported at roughly 30 to 47% StOSt and 15 to 42% StOO depending on provenance and study, so quote a range rather than a number: the tendency is universal, the severity is batch dependent. Note also that fractionation carries nearly all the unsaponifiable matter into the olein. Compare that with the more uniform hard butters in our comparison of butters, where kokum is the low-grain reference for exactly this reason.
Diagnose it before you remelt
| Observation | Likely cause | Confirm by | Fix |
|---|---|---|---|
| Fine sandy grit throughout, appeared days after pouring | StOSt segregation, now transformed to beta | Rub a smear hard: it smears out under pressure and warmth, leaving no discrete particle | Full remelt, hold, fast cool with stirring |
| Grit only in the top 2 to 3 mm | Slow surface cooling under a warm lid, or heat cycling in storage | Scrape the top layer and compare with material from the middle | Remelt the batch, cool uncovered until set, then lid |
| Hard lumps that survive rubbing | Undissolved wax poured too cool, not shea grain | Warm a sample to 45 C: shea grain clears, wax lumps persist | Remelt, raise pour temperature, strain |
| Grain returned within a day of a rescue melt | Melt reached about 40 C only, so seed crystals survived | Trust a thermometer, not the appearance of the liquid | Remelt to 70-80 C and hold 20-30 minutes |
| Worse than expected in a shea and coconut blend | Short-chain fats pack poorly against StOSt | Run a test batch with the coconut swapped for a longer chain oil | Cut the short-chain fat or the shea share |
| Smooth on despatch, gritty on arrival | Thermal cycling in transit driving ripening | Ask what the parcel's journey was | Reformulate for heat, ship cool, avoid summer stock sitting |
Fine on Monday, grainy by the weekend
The complaint has a characteristic timeline, and it matches the measured polymorph transformation. Shea stearin at 20 C is largely beta prime after a day and beta after a week, so a balm can leave the bench smooth, pass a next-day check, and be unmistakably gritty when a customer opens it on Saturday. A quality check made on the day of pouring is testing the wrong thing. Hold a retained sample ten days at room temperature before signing off a shea formula, and a second at 30 C if it will ship in summer.
The carrier oil changes the risk, and almost nobody says so
The most useful published finding on this subject is one the craft internet has not absorbed. Pettersson, Alander and Norberg (2025) followed shea stearin crystallising in two different liquid carriers and found the carrier controlled both the speed and the coarseness of the result.
| Carrier | Time to beta form | Crystal habit | Relative firmness |
|---|---|---|---|
| Rapeseed oil (triglyceride) | About 2 days | Small, densely packed aggregates | Reference |
| Octyldodecanol (polar emollient) | 5 to 7 days | Fewer, much larger aggregates | Up to ten times lower |
Read that carefully, because the slower system is the worse one. Delaying transformation did not prevent it. It produced fewer nucleation sites, so the crystals that did form grew large, and the network they built was up to ten times weaker. Large crystals are precisely what a fingertip detects. The working rule that follows: polar emollients such as octyldodecanol, C12-15 alkyl benzoate and caprylic/capric triglyceride dissolve shea well and slow its crystallisation, and that raises grain risk rather than lowering it. If you are choosing between oils in our carrier oil comparison, a plain triglyceride oil is the safer partner for a high-shea formula, and a stable one such as jojoba behaves differently again because it is a liquid wax ester rather than a triglyceride and brings no StOSt of its own.
The second compositional rule is about chain length. Shea packs poorly with short-chain fats, which is why high coconut oil, palm kernel or babassu alongside a lot of shea is a reliably gritty combination. If a formula must carry both, keep the shea below roughly 20% and the short-chain fat below 10% of the total, weight percent of a formula totalling 100.
The "heat it to 175 F" rule, examined honestly
This is the most repeated number in balm making and it deserves a straight answer. The 175 F (79 C) figure has no traceable source for shea. It does coincide with a lipid laboratory convention: Marangoni's group melts cocoa butter at 80 C for 30 minutes to erase crystal memory, and other laboratories use 55 C for 20 minutes for the same purpose. Somewhere along the way a food science protocol for a different fat became a craft rule with a Fahrenheit number attached.
Theory says no universal threshold exists at all. Memory erasure is a time and temperature trade-off, hotter needing less time and cooler needing more (Pink and colleagues, 2020). What does have a number attached is the clear point of beta StOSt, at about 41 to 43 C. That is the figure that matters, and it explains why melting shea "until it looks liquid" at around 40 C is genuinely not enough: the butter flows, and the seeds survive.
Trebalm's position: hold well above 45 C, and treat 70 to 80 C for 20 to 30 minutes as a generous margin borrowed from lipid laboratories rather than a measured threshold for shea. Nobody has published what that heat does to shea's unsaponifiables and tocopherols, so it stays an open question rather than a settled one.
Fix this batch
Grain is fully reversible. Melting past the beta clear point destroys the crystals, and nothing is chemically degraded in the process, so a rescued batch is not a compromised batch.
- Scrape everything back into one vessel. Rescuing pot by pot guarantees an inconsistent result. Reprocess the whole lot, including any thin scrapings from the edges of tins.
- Melt in a water bath to 70-80 C and hold 20 to 30 minutes. Time at temperature is doing the work, not the moment the last lump disappears. Use a probe thermometer in the fat, not an infrared reading off the surface. See our notes on workshop equipment for what is worth owning.
- Cool with agitation through the crystallisation window. Stir steadily from about 60 C down to the point of visible thickening. Agitation multiplies nucleation sites, which is the mechanism that gives you many small crystals instead of few large ones.
- Drop the temperature quickly. A cold water bath, or 20 minutes in the fridge, not the freezer. Get through the window between roughly 36 C and 25 C without lingering.
- Add heat sensitive materials at the end. Essential oils, flavour and tocopherol below 50 to 60 C, after the hold and during the controlled cool.
- Pour at the right temperature. Tins 60 to 68 C, tubes 65 to 72 C, as set out under pour temperatures, then leave the poured containers uncovered until fully set.
- Retain a sample and wait ten days. A same-day check tells you nothing about the beta transformation.
Prevent the next batch
- Beta StOSt clear point: 41-43 C. Melting to 40 C is not enough.
- Melt and hold: 70-80 C for 20-30 minutes.
- Danger window on cooling: roughly 36 C down to 25 C. Move through it fast.
- Beta prime after 1 day at 20 C, beta after 1 week.
- Practical ceiling for shea in a grain-sensitive product: 20-25% of the formula.
- Assess grain at 10 days minimum, and after a 30 C heat challenge.
Formulation levers matter as much as process. Reducing shea and replacing part of it with a more uniform hard butter cuts risk directly, and a wax network of the sort described in balm texture science immobilises the liquid phase enough to slow ripening. For a rich body balm, splitting the hard fat between shea and kokum is usually a better move than trying to process pure shea perfectly.
Why the freezer is the wrong tool
Shock cooling is popular because it visibly works on the day. The problem is what it produces. Cooling below about 18 C favours the alpha form, which is unstable and will still march toward beta over the following days. Freezer shock buys small initial crystals, not permanent stability, and it can also cause condensation on the surface of a cold container, which is a separate problem in an anhydrous product. Stirring through the cooling range is the better evidenced lever. Use a cold water bath or a short fridge hold to remove heat, and use agitation to control crystal number.
What the literature does not contain
There is no peer-reviewed literature on cosmetic shea graininess as such. Everything above is adjacent lipid science, mostly from food and fractionation research on shea stearin, applied to a cosmetic problem that nobody has formally studied. Saying so is part of the answer. Two corollaries follow. First, "refined shea does not go grainy" has no support, and refined stearin is more SOS-rich and higher melting than crude. Second, you cannot temper shea to form V. Forms I to VI are cocoa butter nomenclature, and cocoa butter tempering deliberately carries a few percent of surviving form V seed under controlled shear. Shea's stable end state is plain beta, and the goal is many small beta crystals produced by a nucleation burst. Same vocabulary, different physics, and the difference also shows up in the cracking and crumbling faults that brittle cocoa butter causes.
When it is not worth saving
Three cases. If the balm smells sharp, waxy or like old crayons, it is oxidising, and remelting a rancid batch gives you smooth rancid balm. If it carries botanical material or anything you cannot reheat safely, reprocessing changes more than the crystal structure. And if the batch is a dozen four gram tubes, the reprocessing time costs more than the ingredients are worth. Grain is a defect of feel, not of safety, so the honest commercial answer is usually to correct the process, pour again, and keep the gritty batch for your own use rather than sell a rescue you have not aged for ten days.
Frequently asked questions
Is grainy shea butter still safe to use?
Yes. Grain is a physical fault, not spoilage. The crystals are the same triglycerides that were always in the butter, rearranged into aggregates large enough to feel. Nothing has oxidised, nothing has degraded, and remelting restores the original material completely. Judge spoilage by smell instead: rancid shea smells sharp, waxy or like old crayons.
Why does shea go grainy but kokum does not?
Because shea contains two populations of triglyceride with very different melting behaviour, roughly a stearin fraction melting near 36 C and an olein fraction melting near 25 C. On slow cooling the stearin crystallises largely on its own and separates from the liquid. Kokum is far more uniform in composition, so there is much less to segregate.
Does melting to 175 F really erase crystal memory?
There is no traceable source for that specific figure applied to shea. What is defensible is that beta-form StOSt clears at about 41 to 43 C, so melting only until the butter looks liquid at around 40 C leaves seed crystals intact. Holding at 70 to 80 C for 20 to 30 minutes is a generous margin borrowed from lipid laboratories.
Should I put balm in the freezer to stop grain?
Not as a first choice. Cooling below about 18 C favours the unstable alpha form, which will still transform toward beta over the following days, so freezer shock buys small initial crystals rather than permanent stability. Stirring through the cooling range is the better evidenced lever, with a cold water bath or a short fridge hold as support.
Does refined shea butter avoid graininess?
No, and the claim has no support. Refining removes odour and colour, not the triglyceride mix that causes segregation. Refined shea stearin is more SOS-rich and higher melting than crude shea, so if anything the fractionated material is more prone to visible crystals. Process control matters far more than grade.
Can I fix grainy balm without remelting the whole batch?
Not reliably. Surface reflow with a heat gun smooths the top few millimetres and leaves the crystals underneath, so the grit returns the moment someone digs in. If the batch matters, remelt it entirely, hold well above the beta clear point, then cool fast with agitation.
Why did the grain come back a week after I fixed it?
Either the melt was not hot enough or long enough to destroy every seed, or the cooling was slow enough to let the stearin crystallise alone again. Storage cycling does the rest: repeated warming and cooling coarsens existing crystals by Ostwald ripening, which is why a smooth balm can turn gritty in a hot delivery van.
Sources and further reading
- Ray, J. and colleagues, Polymorphic behaviour of shea stearin, European Journal of Lipid Science and Technology, 115:1094, 2013.
- Pettersson, K., Alander, J. and Norberg, S., Crystallisation of shea stearin in liquid carriers, Journal of the American Oil Chemists' Society, 102(8):1237-1248, 2025.
- Abdel-Razek, A. G. and colleagues, Characterisation of crude shea butter and its stearin and olein fractions, Foods, 12(8):1626, 2023.
- Pink, D. A. and colleagues, Modelling crystal memory in triacylglycerol melts, Molecules, 25:5631, 2020.
- Hou, J. and colleagues, Crystallisation behaviour of shea stearin, Journal of the American Oil Chemists' Society, 101(11), 2024.
- Ghazani, S. M. and Marangoni, A. G., Molecular origins of polymorphism in cocoa butter, Annual Review of Food Science and Technology, 12:567, 2021.
Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.