Ice Water Extraction and Bubble Hash: What Separates a Good Wash From a Bad One
Cold water does not extract anything — it makes resin glands brittle enough to fall off, and everything after that is a question of how gently you move the water.
1. Read This First: The Legal Position in Thailand
In Thailand, cannabis extracts containing more than 0.2 percent tetrahydrocannabinol by weight remain a Category 5 narcotic. That did not change when the plant material was rescheduled. As the law firm Nishimura & Asahi set out in its 2022 briefing on the Ministry of Public Health notification of 9 February 2022, the exemption covers only an “extract substance containing tetrahydrocannabinol (THC) not exceeding 0.2 percent per weight” from plants grown in Thailand.
Ice water hash is, by design, resin far above 0.2 percent, so a successful wash produces a controlled narcotic under Thai law. Regulation has tightened since: the Bangkok firm Tilleke & Gibbins describes the Notification of the Ministry of Public Health Re: Controlled Herbs (Cannabis), B.E. 2568, issued 26 June 2025, requiring a prescription on the official PT 33 form. What follows is for jurisdictions where processing is lawful.
2. The Principle: Cold, Water and Gentle Force
2.1 Why cold makes the glands let go
Resin sits outside the plant, in stalked glands weak where head meets stalk. Reference material on hashish states the mechanism directly: trichomes “may break away from supporting stalks and leaves when plant material becomes brittle at low temperatures”. Cold embrittles that join and stiffens the resin, so a detached gland behaves like a solid bead rather than a smearing droplet. Warm resin is glue; cold resin is gravel.
2.2 Mechanical, not chemical
Nothing dissolves. Cannabinoids are poorly soluble in water: it is a carrier and a cushion, not a solvent. Reference descriptions list the toolkit as ice, water, agitation, and filtration bags with various-sized screens. A wash therefore cannot improve on its input — it only sorts by particle size.
3. Trichome Anatomy That Matters to a Washer
3.1 The three gland types
Cannabis carries bulbous glands, capitate-sessile glands flush against the tissue, and capitate-stalked glands raised on a multicellular stalk. Only the last matters commercially: largest, most resin, most exposed to agitation, with a defined breaking point. Sessile glands release less readily; bulbous glands pass through nearly every mesh.
3.2 How large is a resin head, honestly
This figure is quoted constantly and sourced rarely. The primary morphology paper on cannabis trichome dimensions could not be retrieved for this article, so rather than invent a citation, here is what was verifiable.
Reference material on solventless extraction states that “the microns that are held in highest regards are the 73μ and 90μ, as this is where the resin heads reside”, and that full-spectrum material “will normally come from 45μ-159μ, as smaller and larger particles are likely to be too unrefined or broken stalks of the trichomes”. Working backwards from that, intact mature heads span roughly 45 to 160 micrometres, concentrated around 70 to 120.
Tighter figures repeated in grower communities — heads of 80 to 100 micrometres, or a canonical 90 — are practitioner consensus with no traceable published measurement behind them. Head diameter is a distribution shifting with cultivar and maturity, which is why a stack beats one screen: a lone 90 micrometre bag loses every head below 90 and holds every contaminant above it.
4. Bag Micron Ranges and What Each Fraction Holds
The 73 and 90 ratings and the 45-to-159 window come from the reference material above; the rest of this table is practitioner consensus.
| Rating in micrometres | Role | Typical contents | Destination |
|---|---|---|---|
| 220 to 250 | Work bag | Leaf, stem, plant material | Discarded |
| 190 | Coarse | Debris, broken stalks, clumps | Low grade |
| 160 | Top of spectrum | Oversized, clustered heads | Inspect closely |
| 120 | Upper prime | Large heads, some contamination | Often very good |
| 90 | Prime | Intact mature heads | Full melt candidate |
| 73 | Prime | Intact heads, smaller | Full melt candidate |
| 45 | Lower prime | Small heads, fragments | Melts, sometimes green |
| 25 to 35 | Fines | Debris, contaminants | Rarely full melt |
4.1 Reading “seventy-three to ninety” as a starting point, not a law
A useful default and a poor dogma. Large-glanded cultivars can put their best material in 120, small-glanded ones in 45. Judge fractions on look, smell and behaviour under heat, not on the printed number.
5. Starting Material: Fresh Frozen Versus Dried and Cured
5.1 Why fresh frozen dominates, and how to handle it
Fresh frozen material — cut and frozen within hours, never dried — dominates modern washing. Turgid tissue is less friable, so less leaf shreds into the water; aromatics that would evaporate during drying survive; and heads never dehydrated are fuller and rounder, improving size sorting and melt. The trade-off: it carries the plant’s full field microbiology into the bucket.
Consensus handling practice, without a published parameter set: remove large stems, avoid handling flower by hand, freeze fast, hold frozen until it enters water. Repeated thawing and refreezing is blamed for degraded washes, on the reasoning that ice crystals rupture tissue and release pigment.
5.2 Mould and moisture in a humid tropical climate
In a 2023 review in Frontiers in Microbiology, Gwinn, Leung, Stephens and Punja catalogue the toxigenic genera of concern — Aspergillus, Penicillium, Fusarium, Mucor, Alternaria, Cladosporium and Trichoderma — and identify water activity as the controlling variable. Most fungi grow best at a water activity of 0.83 to 0.99, below 0.7 prevents growth of most, and some Penicillium and Aspergillus species survive down to 0.62 to 0.7.
Two findings bear on hash directly. Concentrating resin concentrates what came with it: they record ochratoxin A in 39 percent of resin samples against 27 percent of herbal samples. And a 2020 review in Frontiers in Pharmacology by Park and colleagues flags “sweat curing” as a contamination risk, because of the high water activity inside stacked material.
Where humidity stays high most of the year, material must never sit warm and wet. The window between the end of the wash and the start of drying is the most dangerous period: wet hash is a dense, high-water-activity mass of exactly what fungi prefer.
5.3 Washing dried and cured material
Dried and cured input still works and was the historic norm. It needs no freezer chain, but sheds more plant matter into the water, yields lower, and gives a duller aroma. Where freezing is unreliable — a real constraint in much of the region — well-dried material washed carefully beats badly handled fresh frozen.
6. Water, Ice and Agitation
The consensus target is water as close to freezing as the ice will keep it, generally described as under about 4 degrees Celsius, with ice present throughout so the temperature never drifts up. That figure is craft practice, not a published parameter. Washers prefer low-mineral water, reverse osmosis or distilled, reasoning that minerals and chlorination products show in the patty as haze or off-taste — again consensus, not measurement. Water is a microbiological input too, so clean water matters in the tropics. Ice is used generously, at least as much by weight as plant material.
Agitation intensity is the main quality variable, and most operators err in the same direction. Gentle agitation removes heads ready to come off; aggressive agitation removes those plus stalk fragments, leaf tissue and chlorophyll, so yield rises while quality collapses. The same material, bags and water give full melt or green paste depending on nothing but how hard the slurry was moved.
7. The Wash
7.1 Method, duration and successive cycles
By hand with a paddle, by gentle machine, or by stirring in a lined vessel, the target motion is a slow even circulation that lifts and turns the material rather than beating it. Material soaks cold first; convention runs to short opening cycles of a few minutes with longer, firmer cycles later. The first wash pulls the ripest, most exposed heads and gives the cleanest, palest result; each later wash reaches deeper, yield falling, colour darkening, melt declining. Most run three to five cycles, keeping fractions separate.
7.2 When to stop
Stop when the water comes off green rather than amber, when returns no longer justify the contamination they bring, or when the material breaks down into pulp. Chasing one more cycle is the commonest way to spoil a good batch: low-grade output contaminates everything it joins.
8. Collection and Drying
8.1 Sieving and particle size
Patties are rinsed cold, pressed gently to remove free water, then broken down. Reducing the wet mass to small uniform particles, commonly over a fine sieve or micro-grater onto a cold surface, increases surface area so water leaves fast and evenly. A thick patty dries outside-in, and the inside is where things go wrong.
8.2 Freeze drying compared with air drying
Drying is where most quality is lost: the resin survived the wash intact, and it is drying that costs aroma and invites microbial growth.
| Factor | Freeze drying | Cold air drying |
|---|---|---|
| Mechanism | Frozen water sublimed under vacuum | Evaporation into cool moving air |
| Typical duration | Roughly one to two days | Several days to two weeks |
| Aroma retention | Highest | Lower; long exposure costs aroma |
| Colour | Usually lightest | Often darkens |
| Microbial risk | Low; frozen, dries fast | Higher; long window at usable water activity |
| Main barrier | Equipment cost | Needs a cold, dry, clean space |
| Best suited to | Full melt fractions | Pressing grades |
Durations and outcomes above are practitioner consensus; no published drying-parameter study was accessible for this article. Air drying wants a cold, dark, low-humidity space with light air movement and no direct warmth, spread thinly and turned — in the tropics, a refrigerated space with dehumidification, not a room.
8.3 Microbial risk when wet hash dries badly
Wet hash sits at the top of the water activity range where, per the review above, most fungi grow best. A dense patty left at room temperature in humid air spends hours or days where those organisms thrive, and cannot be salvaged: later heating destroys neither mycotoxins already formed nor ochratoxin A. Dry it cold, fast, thin.
9. Grading: The Melt Scale and Its Limits
The one-to-six star melt scale below is practitioner consensus, with no standards body behind it.
| Rating | Label | Behaviour under heat | Residue |
|---|---|---|---|
| 1 star | Food grade | Does not melt; chars | Heavy black |
| 2 star | Food grade | Softens, may bubble slightly | Heavy |
| 3 star | Half melt | Partial, uneven bubbling | Noticeable |
| 4 star | Half to full melt | Bubbles, incomplete | Light to moderate |
| 5 star | Full melt | Melts cleanly, sustained bubble | Minimal |
| 6 star | Full melt, top tier | Melts fully and evenly | Close to none |
What the scale measures is purity of the intact-head population: contaminants such as plant fragments, ruptured heads and cell debris do not melt, so they char and leave residue. Reference material on hashish agrees: where a high level of pure tetrahydrocannabinol is present, the product is almost transparent and starts to melt at the point of human contact. The same source gives the aroma check: good hash smells fragrant and aromatic, poor material distinctly mouldy or musty. Other bench indicators are colour from near-white through cream to amber, a dry sandy texture, and clumping under gentle warmth.
9.1 How much of the scale is subjective
A great deal. Colour correlates with quality but is cultivar-dependent, and a darker fraction can outperform a paler one. Melt is judged by eye, on non-standard equipment, often by the person who made it. The scale is shared vocabulary, not measurement: identical material can be rated a grade apart by two competent people.
10. Yield Expectations
Yield should be quoted as a range and against a stated basis, because fresh frozen material is mostly water: the same percentage against fresh frozen weight and against dry weight mean very different things.
For a related solventless step, reference material on hash rosin gives a published comparison of 0.3 to 8 percent for rosin against 10 to 20 percent for hydrocarbon extraction — solventless ranges are legitimately wide. For ice water hash itself no accessible published dataset was found, and circulating figures are practitioner consensus: poor material in low single digits against fresh frozen weight, competent material higher, purpose-grown higher still, with only a portion landing in full-melt grades. Total yield and full-melt yield are different quantities. The same source notes that growing for hash differs from growing for flower, and that some cultivars are deceptive about yield.
11. Faults and Their Causes
| Fault | Likely cause | Correction |
|---|---|---|
| Green colour | Chlorophyll from ruptured leaf; over-agitation | Gentler agitation, colder slurry |
| Visible plant matter | Work bag too coarse; too many cycles | Better pre-filtration; stop earlier |
| Poor melt, heavy residue | Non-melting debris; immature heads | Judge fractions separately |
| Sour or fermented smell | Wet material held warm; slow drying | Dry cold, thin, fast |
| Hay or musty smell | Contaminated input; mould | Reject the input |
| Dark colour, good aroma | Oxidation, late cycles, or cultivar | Judge by melt and smell |
| Very low yield, clean | Under-agitation or low-resin cultivar | Check a known cultivar |
| Gritty, will not press | Incomplete drying; hard water | Finish drying; softer water |
12. Storage, Ageing and Pressing
Fully dried hash is stored cold, dark and airtight, brought to room temperature before opening so condensation does not form. The enemies are heat, light, oxygen and residual moisture — and moisture turns a storage question into a microbiological one, because hash that was not fully dried can support growth, not merely degrade.
Pressing loose resin into a temple ball is a traditional finishing step: the material is worked with gentle warmth and pressure until individual heads rupture and coalesce into one homogeneous mass — the transformation reference material describes in pure sieved resin becoming gooey and pliable under heat. Pressing excludes air, slowing oxidation; it homogenises the batch; and it changes character, since rupturing heads releases volatiles and pressed material is often called rounder and less bright than the same resin loose — a trade-off, not an upgrade.
13. What Actually Separates a Good Wash
Cold that never lapses. Clean input. Agitation gentle enough to feel unproductive. Fractions judged individually rather than dumped together. Drying treated as the most important step, not the last. Almost every failed wash traces to chasing yield with force or rushing the dry — patience problems, not technique.
What does your stack look like, and which bag consistently gives your best fraction? We are especially interested in how members in high-humidity parts of the region manage the wet window between collection and dry. If you have washed the same cultivar both fresh frozen and dried-and-cured, tell us what changed. And if you have data rather than impressions, post it: this topic has far more folklore than measurement.
This article is part of the Asiannabis Community educational series on extraction and post-harvest processing. Content is for educational purposes within jurisdictions where cannabis processing is legally permitted.