Why Is Sake Filtered, and What Does Muroka Change?
Most sake is filtered twice over. First the finished mash is pressed away from its solids; then the young sake is usually treated with powdered activated carbon to lift colour and strip off-notes. 無濾過 (muroka) means only that second step was skipped — the charcoal. It does not mean the sake arrived in the bottle untouched, and that distinction turns out to be the genuinely interesting part.
Three different things get called “filtration”
English collapses into one word what Japanese brewing keeps carefully apart.
上槽 (jōsō) is pressing. At the end of fermentation the mash is a thick porridge of sake, undissolved rice and koji. It gets squeezed — in a modern accordion press, or in the old boat-shaped fune — and what runs out is sake, while what stays behind is the white cake called sake kasu. Every sake goes through this. There is no such thing as unpressed sake.
澱引き (ori-biki) is settling. The freshly pressed liquid still carries a haze of fine rice and yeast. Left in a tank for a few days, most of it drops to the bottom and the clear sake is drawn off the top. This is gravity, not filtration.
濾過 (roka) is filtration proper, and this is the step the label is talking about. It covers plain mechanical filtration through a filter medium, and it covers treatment with powdered activated carbon. When a bottle says muroka, it is almost always the carbon that was left out. The National Research Institute of Brewing’s own outline of sake production lists pressing and filtration as two separate stages, which is a useful reminder that they are different operations doing different jobs (NRIB, The Story of Sake).
What activated carbon is actually for
Charcoal has three traditional jobs in a sake brewery: reducing unpleasant flavours, removing colour, and adsorbing impurities generally (Murayama et al., 2018).
The colour question is worth pausing on, because it is more chemistry than cosmetics. Fresh sake is not water-white; it carries a faint gold or green cast, and it picks up more colour during storage. Iron is the notorious accelerant. The NRIB specifies that water used for brewing must contain less than 0.02 ppm iron — considerably stricter than Japan’s drinking-water standard of 0.3 ppm — because excess iron turns sake reddish-brown and coarsens both taste and aroma. Carbon pulls that colour back out.
The off-note question is more specific still. Sake that has been stored badly develops ひね香 (hineka), an aged odour whose main culprit has been identified as 1,3-dimethyltrisulfane, or DMTS — sulphurous, faintly like cut onion. Its detection threshold in sake is around 0.18 μg per litre, which is to say a few parts per trillion of something you very much do not want. In one published experiment, a deliberately aged sake sitting at 0.25 μg/L was treated with activated carbon and came down to 0.09 μg/L, comfortably back under the threshold (Murayama et al., 2018). The fault disappeared. Charcoal works.
What carbon takes with it
Here is the catch, and it is the whole argument for muroka. Activated carbon is not selective. It works by drawing molecules into a vast network of micropores, and it has a particular appetite for hydrophobic ones — which is exactly what an aroma ester is.
In the same experiment, the same charcoal treatment that took DMTS down also took ethyl hexanoate from 1.4 mg/L to 0.6 mg/L. Ethyl hexanoate is one of the two esters that make up 吟醸香 (ginjōka), the apple-and-green-melon perfume that a ginjo brewer spends the whole winter coaxing out of a cold, slow fermentation. Carbon removed rather more than half of it, in the course of fixing an unrelated fault.
That trade-off is precisely why the researchers were testing gold nanoparticles on a silica support as an alternative adsorbent: it stripped DMTS harder than charcoal did while leaving the esters essentially intact. Worth saying plainly — that is a laboratory adsorbent, not something in your bottle. Nobody is filtering commercial sake through gold. But it maps the problem neatly: carbon is a blunt instrument that removes faults and virtues together.
So carbon is best understood as a dial rather than a switch. Turn it up and you get a paler, cleaner, more stable sake with less to say. Turn it down and you keep the aroma and you keep the risk.
Muroka is not “untouched” — the press cloth got there first
This is the part the romantic version of muroka tends to skip. If charcoal adsorbs esters, so does everything else porous that sake touches on its way out — starting with the cloth it is pressed through.
A team at the Industrial Technology Center of Okayama Prefecture collected eighteen used filter cloths from working breweries — cotton, polyethylene terephthalate and PVC — and analysed what had accumulated in them. The cloths were loaded with volatile compounds, esters prominent among them, and the profile differed by material: the synthetic cloths trapped both a greater quantity and a wider variety than cotton did. Then, following actual commercial pressing runs, they measured the sake before and after. Medium- and long-chain fatty acid ethyl esters dropped, while the corresponding fatty acids did not — the signature of selective adsorption onto the cloth rather than of simple dilution. In one daiginjo, ethyl decanoate emerged at roughly 15% of its pre-press level (Tanino et al., 2021).
There is no unedited sake. By the time a brewery decides whether to reach for the charcoal, the press cloth has already made its own choices about which esters survive. Muroka marks the point where deliberate editing stopped — not the point where editing never began.
Which is not an argument against muroka. It is an argument against reading it as purity. A muroka bottling is a brewery declining one specific intervention, for reasons it usually has thought about carefully. It is not a bottle that escaped the process.
Muroka is not a legally defined term
Japanese sake labelling is governed by the National Tax Agency’s Standard for Manufacturing Methods and Quality Labelling of Sake (清酒の製法品質表示基準), issued in 1989 and in force since April 1990. Alongside the eight special designation categories, it sets out exactly ten optional terms a producer may put on a label, each with a stated condition (National Tax Agency):
rice variety; region of production; years in storage; 原酒 (genshu) — no water added after brewing, beyond an adjustment of under 1% ABV; 生酒 (namazake) — no heat treatment at any point; 生貯蔵酒 (namachozōshu); 生一本 (kiippon); 樽酒 (taruzake); a narrow allowance for quality adjectives; and date of production.
Muroka is not on that list. It is a brewery convention rather than a regulated claim — which is why its precise meaning shifts a little between producers, though in practice it reliably means no activated carbon, with the sake still pressed and usually still passed through a plain mechanical filter. Compare that with nama, where the law tells you exactly what you are getting.
The same standard contains a detail worth borrowing. It forbids a label from claiming that a sake is 最高 (the highest), 第一 (the first) or 代表 (the representative) of its industry. A brewery is legally not permitted to tell you its muroka is the best sake. That is a reasonable instinct to adopt when reading one.
So what does muroka actually taste like?
Expect a deeper colour — more gold, sometimes a faint green — more weight and texture on the palate, and a broader flavour that includes things the charcoal would otherwise have taken out. Whether that reads as richer or as rougher depends on the sake and on you. The vocabulary that fits is the same one a wine drinker uses for an unfined, unfiltered red: more information, less polish.
One honest caveat. Muroka rarely travels alone. It is very often sold as muroka nama genshu, because a brewery inclined to skip the charcoal is usually also inclined to skip the pasteurisation and the dilution. Three decisions arrive in one bottle, and the difference you taste is not carbon alone. Our breakdown of muroka, genshu and nigori pulls those apart.
And, as with polishing ratios: this is a style, not a rank. A heavily charcoal-treated honjozo built to be served warm is doing its job well. So is an untreated junmai daiginjo that tastes of the whole rice grain. They are answers to different questions.
Frequently asked questions
Is muroka sake cloudy?
Usually not. Cloudiness is にごり (nigori), which comes from deliberately leaving rice solids in the sake at pressing — a different decision entirely. Muroka is typically clear, just less pale than a carbon-treated bottle. The two can be combined, which is why the terms are easy to confuse.
Does muroka mean unpasteurised?
No. Pasteurisation is 火入れ (hiire), a heat step at roughly 60–65°C, and it is a completely separate axis from filtration. A sake can be muroka and pasteurised, or filtered and unpasteurised. If a bottle is unpasteurised it will say nama, because that term is regulated.
Is there charcoal left in filtered sake?
No. The powdered carbon is stirred in, given time to adsorb, and then removed by filtration along with everything it has picked up. It is a processing aid, not an ingredient.
Does filtration change the alcohol content?
Not meaningfully. Alcohol content is set by fermentation and then, in most cases, by dilution with water before bottling — typically from around 20% down to about 15%. Sake that skips that dilution is labelled genshu. That is a separate word from muroka for a reason.
Should I serve muroka sake differently?
Treat it as you would the rest of the bottle’s description rather than the word muroka itself. If it is also nama, keep it cold and drink it reasonably soon after opening. If it is also genshu, it is probably 17–19% ABV and takes well to a little ice or a splash of water. The absent charcoal is not, on its own, a serving instruction.
Sources
- Murayama, H., Yamamoto, Y., Tone, M., Hasegawa, T., Kimura, M., Ishida, T., Isogai, A., Fujii, T., Okumura, M. & Tokunaga, M. (2018). Selective adsorption of 1,3-dimethyltrisulfane (DMTS) responsible for aged odour in Japanese sake using supported gold nanoparticles. Scientific Reports. DOI 10.1038/s41598-018-34217-w. Full text (PMC6207770)
- Tanino, Y., Ito, K., Hanafusa, H. & Miyake, T. (2021). Effect of filter cloths on the relationship between fatty acids and ethyl esters during automatic pressing of sake. Food Science and Technology Research, 27(3), 417–428. DOI 10.3136/fstr.27.417. Full text (J-STAGE, open access)
- National Tax Agency of Japan. Standard for Manufacturing Methods and Quality Labelling of Sake (清酒の製法品質表示基準), NTA Notification No. 8 of 1989, in force from April 1990. Official summary · Full text of the notification
- National Research Institute of Brewing. The Story of Sake (1st edition, January 2014). PDF
Written by Dr. Sake. We take sake apart — the chemistry, the history, the label — and put it back together in plain English. For weekly diagrams, label breakdowns and brewery notes, follow @thesakeanatomy on Instagram.
Keep exploring
