Winterization

The purpose industry practice, not published data

Cannabis and hemp trichomes and cuticle carry a substantial load of waxes and lipids: long-chain fatty acids and their esters, sterols, phospholipids and the cuticular wax layer that protects the plant surface. A non-selective extraction pulls them along with the cannabinoids, and they cause three specific downstream problems. In distillation they raise the residue fraction, foul the flask or the heated wall, hold product in a viscous matrix so it is lost to the residue, and generate colour by degrading at distillation temperature — a waxy feed is the most common reason a first-pass distillate is dark. In a finished product they cause visible cloudiness and a grainy texture, and in a vaporiser cartridge they separate, wick badly and taste of burnt fat. And in a full-spectrum or broad-spectrum oil they dilute the active fraction with mass that nobody is paying for. Winterization is the standard remedy: dissolve the crude in ethanol, chill it hard enough that the waxes are no longer soluble while the cannabinoids still are, and filter the precipitated wax out cold.

• Clarity and taste in the finished product.
• Distillation performance: less residue, less fouling, less product held in the residue, lighter colour.
• Removing mass that is neither active nor saleable.

Sources: Composite: extraction-equipment vendor documentation 2026* · Green DW 2019*

The mechanism: differential solubility

Solubility is temperature-dependent and the dependence is not the same for every solute in the same solvent. Cannabinoids, which are lipophilic but comparatively small and only moderately structured molecules, remain appreciably soluble in ethanol even at −40 °C and below. Plant waxes and long-chain lipids, which are large, saturated, highly ordered molecules that pack readily into a crystalline solid, lose solubility sharply as temperature falls and come out of solution as a flocculent white-to-yellow precipitate. That divergence is the entire separation. It also explains every parameter choice in the operation: the colder you go, the more completely the waxes precipitate and the finer the precipitate becomes; the longer you hold, the closer the system gets to equilibrium and the more completely nucleation and growth finish, giving a precipitate that filters instead of blinding the filter; and the more dilute the solution, the more cleanly the waxes separate from the viscous cannabinoid phase, because a concentrated solution is thick, traps precipitate, and filters badly. Note that this is a solubility separation and nothing else — no reaction occurs, no molecule is changed, and the recovered wax is the plant wax that was always there.

Sources: Composite: extraction-equipment vendor documentation 2026* · Mullin JW 2001*

Parameters: ratio, temperature and time contested industry practice, not published data

The working parameters in the trade are a crude-to-ethanol ratio of about 1:5 to 1:10 by volume, held at anywhere from −20 °C to −80 °C for 12 to 48 hours. Those three numbers trade against each other. A higher dilution (nearer 1:10) precipitates more completely and filters far more easily, at the cost of more solvent to buy, chill and recover, and a larger vessel; a tighter ratio nearer 1:5 saves solvent and time on the evaporator but gives a viscous solution, incomplete separation and a slow, blinding filtration. A colder hold is more complete and faster to reach an acceptable endpoint: at −20 °C, achievable in an ordinary laboratory freezer, an overnight to 48-hour hold is typical; at −40 °C to −80 °C, requiring a proper low-temperature chiller or freezer, 12 to 24 hours is usually sufficient and the result is more complete. Colder is better and there is a real difference between a −20 °C winterization and a −60 °C one in the clarity of the final product. Some operations run a two-stage cold hold, dropping the temperature in steps to get a coarser first precipitate that filters easily followed by a finer second one. What is not negotiable is that the hold must be long enough for the precipitate to form as a filterable solid rather than a colloidal haze; pulling a flask out of the freezer after two hours and filtering it produces a filtrate that looks clear and goes cloudy in the receiving vessel, because the precipitation was never finished.

Contested — caveat. Winterization parameters are trade practice with no controlled published optimum for cannabis crude. Published and vendor-recommended ratios range from about 1:3 to 1:20, temperatures from −10 °C to −80 °C, and hold times from a few hours to several days, and the right combination depends on the wax load of the specific input, which depends on the extraction platform, the plant part and the cultivar. Establish your own by running a ratio-and-temperature matrix and judging on filtrate clarity after a warm-back test plus final product appearance, not by adopting a number.

Sources: Composite: extraction-equipment vendor documentation 2026* · Mullin JW 2001*

Filtration, and why it must stay cold industry practice, not published data

The filtration is where winterizations are won and lost. It is done under vacuum, through a Buchner funnel or a filter plate at laboratory scale and through a jacketed filter housing, plate-and-frame press or filter dryer at production scale. A filter aid — diatomaceous earth is the standard, sometimes with bentonite or activated carbon added when colour removal is also wanted — is used as a pre-coat bed on the medium and often also as a body feed mixed into the solution, because the precipitated wax is compressible and gelatinous and on its own it blinds a filter almost immediately; the rigid particles of the filter aid keep the cake permeable so flow continues. Media are normally stepped: a coarse pass in the region of 25 µm or above to take the bulk, then progressively finer passes at around 5 µm, 1 µm and sometimes 0.45 µm to catch the fines. The absolute requirement is that everything stays cold through the whole filtration. The waxes are only insoluble because they are cold; as soon as the solution warms toward ambient, the finer precipitate redissolves, passes straight through the filter, and reappears in the product when it is concentrated. In practice that means chilling the funnel, the receiving flask, the filter aid and the solvent used for rinsing, working quickly, keeping the vessel in a bath or a cold room rather than on a bench, and never leaving a part-finished filtration standing. A useful check on whether the operation worked is a warm-back test: take a sample of the filtrate, let it come to room temperature, and look for haze forming. If it hazes, the filtration was warm, incomplete or too coarse, and the material needs to go back.

• Vacuum filtration with a filter-aid pre-coat and often a body feed; the wax alone will blind any medium.
• Step the media coarse to fine rather than trying to do it in one pass.
• Pre-chill the funnel, the receiver, the filter aid and the rinse solvent.
• Warming during filtration redissolves the fines and silently undoes the work.
• Warm-back test on the filtrate is the honest check.

Sources: Composite: extraction-equipment vendor documentation 2026* · Green DW 2019* · Mullin JW 2001*

Yield loss, and how to read it contested industry practice, not published data

Winterization removes mass, and some of what it removes is product. The removed wax fraction from a CO2 or ethanol crude is often quoted in the region of 10 to 30 percent of crude mass, and a small further loss comes from cannabinoids occluded in the wax cake and from product held in the filter aid and the wetted equipment. That loss is not a failure, it is the operation working: the wax was never saleable and it would have been lost to the distillation residue anyway, usually taking more product with it. What matters is measuring it honestly. Weigh the crude in, weigh the recovered oil out after solvent removal, assay both, and keep a cannabinoid mass balance rather than a mass-only one — because a 25 percent mass loss with no cannabinoid loss is an excellent winterization, and a 12 percent mass loss with 8 percent of the cannabinoids gone into the cake is a bad one. Washing the cake with a small volume of cold ethanol recovers most of the occluded cannabinoid and is worth doing at any scale. The other quantity worth tracking is the apparent potency increase: removing non-active mass raises the assayed percentage of the remaining oil, and an operation that reports that increase as if it were a purification achievement rather than the arithmetic of removing wax is misreporting.

Contested — caveat. The 10-30 percent figure is trade experience across mixed inputs, not a measured constant. Wax load varies enormously with plant part (trim and leaf carry far more cuticular wax than flower), cultivar, cure and extraction platform, so a specific operation should determine its own figure and track it as a process-control metric.

Sources: Composite: extraction-equipment vendor documentation 2026*

Solvent removal, and what comes next industry practice, not published data

The filtrate is a dilute cannabinoid solution in ethanol, and the next operation is solvent recovery — a rotary evaporator at small scale, a falling-film evaporator at production scale, both covered on rotary-evaporation. The important point at the interface is that the winterized oil coming off the evaporator is not residual-solvent compliant and is not distillation-ready in that state. Ethanol retained in a viscous oil will flash in a distillation flask or, worse, in a wiped-film unit, so a deliberate devolatilisation step — a vacuum oven hold, or a low-temperature solvent-stripping pass on the still before the terpene cut — comes between winterization and the cannabinoid body cut. The standard sequence for a CO2 or ethanol crude going to distillate is therefore: extract, decarboxylate (order can vary), winterize, filter cold, recover solvent, devolatilise, then distil. Operations that skip the devolatilisation step diagnose it later as a vacuum problem they cannot hold, when it is actually solvent boiling out of the charge.

• Filtrate goes to rotary or falling-film evaporation for bulk solvent recovery.
• Winterized oil is not residual-solvent compliant and is not ready for a wiped-film feed.
• Devolatilise deliberately before distillation, or spend the run fighting your own vacuum.

Sources: Composite: extraction-equipment vendor documentation 2026* · International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use 2021*

Which platforms make it mandatory industry practice, not published data

Winterization is a consequence of the extraction platform. CO2-extracted material, particularly from a supercritical run at high density, carries a heavy wax and lipid load and essentially always requires winterization before distillation; the selectivity that makes high-density supercritical CO2 productive is the same non-selectivity that brings the cuticular wax with it. Ethanol-extracted material also generally requires it, warm-ethanol crude emphatically so, and cold-ethanol crude to a lesser degree because the cold extraction already left a large part of the wax in the biomass — which is exactly the point of running ethanol cold, and is a reason to think of cold extraction as a partial winterization performed in advance. Hydrocarbon-extracted material often needs much less and sometimes none, because butane and propane are poor solvents for the polar and the very-long-chain material, and a cold hydrocarbon run on fresh-frozen biomass can produce a crude clean enough to distil directly. Solventless material is not winterized at all in the normal sense; ice-water hash and rosin are mechanically separated resin and there is no solvent solution to chill, though rosin is sometimes cold-filtered or given a cold-ethanol treatment when it is destined for a distillate or an isolate stream rather than sold as a solventless product.

Sources: Composite: extraction-equipment vendor documentation 2026* · Rovetto LJ 2017* · Qamar S 2021*

Safety, as part of the operation

Winterization puts a large volume of cold flammable solvent through a vacuum filtration, and each of those three words carries a hazard. Cold: at −40 °C and below, ethanol and dry-ice or cascade-chilled surfaces cause contact cold burns quickly, and cold does not feel as urgent as heat so people tolerate contact for longer — insulated cryogenic-rated gloves, not nitrile alone, and no bare-skin contact with cold metal. A dry-ice and ethanol bath at about −78 °C sublimes CO2 continuously, which displaces air in a low or enclosed space, so it needs ventilation and ideally low-level CO2 monitoring. Flammable: a volume of ethanol several times the volume of crude is being handled, transferred, filtered and evaporated, so the room needs ventilation, ignition-source control, appropriate electrical fittings, bonding and grounding on metal transfer vessels, and correct flammable-liquid storage under the flammable-liquids code. Vacuum: filtration flasks and funnels are evacuated glassware, so the same inspection discipline applies as anywhere else — check for star cracks and chips in good light, use only vacuum-rated vessels, never an Erlenmeyer under vacuum unless it is explicitly rated, and shield or cage large filtration flasks. Add one hazard specific to this step: diatomaceous earth is a respirable dust and crystalline-silica-containing grades are a recognised inhalation hazard, so filter aid is handled with respiratory protection and wetted rather than poured dry into a breathing zone. And handle the spent cake as what it is — a solvent-wet, flammable waste stream that must not be left in an open container in the workspace.

• Cryogenic-rated gloves; cold burns accumulate without the warning heat gives you.
• Dry-ice baths displace air — ventilate, and consider low-level CO2 monitoring.
• Large ethanol volumes: ventilation, ignition control, bonding and grounding, code-compliant storage.
• Vacuum filtration glassware gets the same star-crack inspection as a distillation flask.
• Diatomaceous earth is a respirable dust hazard; wet it, and wear respiratory protection.
• The spent filter cake is flammable solvent-wet waste; contain and dispose of it as such.

Sources: National Research Council (US) 2011* · National Fire Protection Association 2024* · Occupational Safety 2024* · Composite: extraction-equipment vendor documentation 2026*