Digging deeper

Hiroshi Iwata, the researcher who identified the source of kabi-shu at the Matsuura brewery, published his method of detecting TCA in 2009 – and it is still in use today, including in the analysis of sake submitted to the Japan Sake Awards 2021 (also known as the National New Sake Appraisal, zenkoku shinshu kanpyoukai, 全国新酒鑑評会). 

Iwata-sensei commented during his interview for the Proof podcast episode that kabi-shu was fairly common when he was involved with the National New Sake Appraisal in the early 2000s, but in 2021 it was detected by judges in just 33/821 entries (0.04%), where two or more examiners indicated kabi-shu or four or more said there was a smell like paper or dust. (So his advice based on his experiences at Matsuura Shuzo clearly helped!)

The method used to analyse the competition sake was gas chromatography-mass spectrometry (GCMS), the same method Iwata-sensei used to pinpoint the cause of the problems at Matsuura Shuzo. The internet has many helpful explanations, but the short version is that chromatography splits a mixture (e.g. sake) into its distinct components, and mass spectrometry (plus a lot of analysis) identifies the components by the mass of the molecules.

GCMS was used to look for TCA (2,4,6-trichloroanisole) and TBA (2,4,6-tribromoanisole, which causes a similar smell) in these 33 sake. TCA was at undetectable levels in 6/33, while 9 sake each had ≤1.0 or 1.0-2.0 ng/L, 6 sake had 2.0-3.0 ng/L, 2 had 3.0-4.0 ng/L, and one had over 4.0 ng/L. The average value was 1.89 ng/L, and the maximum was 8.3 ng/L. There was less TBA, with it being undetectable in 27/33 sake, ≤1.0 in 5 sake, and 1.0-2.0 ng/L in just one sake. Average value was 0.81 ng/L, and maximum was 1.7 ng/L. The threshold to smell TCA  is just 1.0-2.0 ng/L.

Judges’ area at the 2021 Japan Sake Awards (zenkoku shinshu kanpyoukai/National New Sake Appraisal), from “Analysis of Sake Components Presented to the Japan Sake Awards 2021”

Breweries that submitted entries were asked about warehouse space, main and other varieties of rice used to make the sake, the percentage of the main rice variety, what prefecture produced the rice, polishing ratio, quantity of rice used in one batch, starter method, addition of alcohol, components of the sake submitted (alcohol, sake meter value/SMV, acidity, amino acidity), yeast variety, yeast varieties if starters were mixed, how many days the moromi fermented, maximum temperature of the moromi, maximum Baume of the moromi, kasu-buai (percentage of solids remaining after pressing), and pasteurisation.  

The researchers then analysed aroma components (ethyl acetate, isoamyl alcohol, isoamyl acetate, and ethyl caproate), glucose, TCA and TBA, and potential for generating DMTS as an indicator of hine-yasusa (老ねやすさ), susceptibility to the hine fault).

In the preliminary round, judges were asked to score:

• Aroma
  • Quality (superb – neutral – flawed)
  • Intensity (intense – neutral – lacking)
  • Ginjo aroma/rich aroma (fruity (banana)/isoamyl acetate – fruity (apple)/ethyl caproate – ethyl acetate – higher alcohols)
  • Wood/spice aromas (acetaldehyde – isovaleraldehyde – 4VG)
  • Koji, sweetness and burnt notes (koji – sweet aroma/caramel-like – burnt aroma)
  • Oxidation, deterioration and sulfur (hine-ka, namahine-ka, smell of yeast or kasu – sulfur)
  • Transferred smells (rubber – mold – earth – paper/dust)
  • Fatty and acidic smells (diacetyl – fatty acids – acidic smell)
  • Other (free entry)
• Flavour
  • Quality (superb – neutral – flawed)
  • Intensity (intense – neutral – lacking)
  • Depth (deep – neutral – light)
  • Aftertaste/pleasantness (short/clean – neutral – heavy)
  • Stimulation/texture (rounded/smooth – rough)
  • Characteristics (sweetness – acidity – umami – bitterness – astringency)
  • Other (free entry)
• Overall evaluation (superb – good – neutral – somewhat flawed – flawed)

The final round scoring simply asks for an overall evaluation of 1, 2 or 3, or else “not suitable for a prize”, in which case the judge has to say why.

The reports on what yeast was used was also interesting – 279 entries used Kyokai #1801, followed by a mixture of yeasts at 137, and Meiri yeast at 114. Nothing else was over 100, with 91 other/unknown.