Two labs heated doenjang and got opposite answers. Both are right.
Drafted with an AI writing tool and checked by a person. This ingredient has not been cooked and re-checked for this article yet. About
- Cooked by me
- not yet — this article is built from published sources
- Sources opened
- 9, each cited in the sentence that uses it
- Papers read in full
- 3, two of them table by table
- Heating studies that varied the time
- 0
The instruction is everywhere. Add the soybean paste at the end. Do not let it boil. Just One Cookbook puts it in capitals — “NEVER BOIL miso soup because it loses flavor and aroma” — and then extends the claim: “Miso loses its nutrients, flavor, and aroma at high temperatures.”
Three things are bundled into that sentence, and they are not equally supported. One has been measured twice, with opposite results. One has not been measured at all.
Measurement one: almost everything goes
In 2008 a group at Changwon National University cooked doenjang and measured what was left in the food afterwards (J. Life Science 18(11):1570-1577). Their method section gives the recipe:
된장스프(된장 5 g+물 50 ml) … 를 100ºC에서 10분간 조리하여
Five grams of doenjang in 50 ml of water, cooked at 100°C (212°F) for 10 minutes. They extracted the volatiles by solvent-assisted flavour evaporation, which recovers the compounds still present in the liquid — close to what reaches the person eating it.
The results from their Table 1, raw paste against cooked soup, in ng/g:
| compound | paste | after 10 min at 100°C |
|---|---|---|
| 3-Methylbutanol | 1,698.3 (125.8) | 69.0 (14.6) |
| 1-Octen-3-ol | 174.5 (18.4) | 53.7 (8.5) |
| Hexanal | 102.5 (7.3) | 50.4 (0.6) |
| Maltol | 3,631.6 (260.9) | 2,856.0 (153.4) |
| 2-Furanmethanol | 710.7 (118.9) | 711.4 (78.8) |
| Furfural | 126.4 (36.6) | 1,019.1 (38.1) |
| 3-(Methylthio)propanal | 131.1 (19.0) | 317.5 (35.6) |
3-Methylbutanol, one of the paste’s dominant alcohols, drops by 96%. Furfural, which is not in raw paste in any quantity, multiplies eightfold — it is made by the heat, not released by it. And methional, listed here as 3-(methylthio)propanal, more than doubles. Hold on to that last one.
The authors say plainly what they think happened:
신선한 과일향이나 꽃향과 같은 향들이 소실되었는데 이는 된장 가열 시에 저급에스테르화합물이 증기와 함께 휘발 소실된 것으로 생각되었다
The fresh fruit and floral notes were lost, they write, because the low-molecular esters volatilised away with the steam. Not broken by the heat. Carried off by it.
Measurement two: almost everything increases
The same year, researchers at the Korea Food Research Institute heated doenjang and measured the headspace — the air above the paste (Korean J. Food Sci. Technol. 40(3):271-276). The crucial detail is the container. Ten grams of paste went into a 30 mL headspace glass vial with 3 g of NaCl, and the vial was sealed with a silicon/teflon septum before being held at room temperature, 50°C (122°F) and 100°C (212°F).
Sealed. Nothing could leave.
Their totals, in ppm, for a commercial paste:
| group | room temp. | 50°C | 100°C |
|---|---|---|---|
| Total esters | 5.83 | 26.86 | 851.87 |
| Total acids | 4.50 | 13.92 | 32.99 |
| Total volatile phenols | 0.02 | 1.13 | 14.72 |
Total esters rise by a factor of 146. A single compound, ethyl hexadecanoate, goes from 0.02 to 717.97 ppm.
Read side by side with the first study, this looks like a contradiction. It is not.
What each one actually asked
The first group asked: after you cook it, what is still in the bowl? The second asked: while it is hot and sealed, what is in the air above it?
Both answers are correct, and together they describe one process. Heat does two things at once. It creates aroma compounds — Maillard products like furfural and maltol, long-chain esters forming from free fatty acids and ethanol — and it drives volatiles out of the liquid. In a sealed vial you see only the first, because the second has nowhere to go. In an open pan you see the net result, and the net result is loss.
The folk instruction is right. The usual explanation for it is not. Nothing is being destroyed by high temperature. It is leaving the pot.
The two studies also agree on more than the contradiction suggests. Both find the small fruity esters falling — ethyl butanoate declines in the sealed vial too, from 0.57 to 0.44 to zero. Both find Maillard products rising. They differ only on the total, and the total is the part that depends on whether there is a lid.
Japan has the time axis
Neither Korean study varied the cooking time. One used 10 minutes, the other 30. Nobody compared three minutes to twenty — which is precisely the comparison the advice depends on.
For that, the closest evidence is Japanese, and it is about miso rather than doenjang. Sugawara and Hosaka at Iwate University tracked miso suspension volatiles by SPME, heating to 80°C (176°F) (J. Home Economics of Japan 60(6):545-552):
The number of flavor components detected by gas chromatography-olfactometry increased when the miso suspension was heated to 80℃, more than 3 times the number at room temperature. However, about 40% of these volatiles could not be detected after heating at 80℃ for 3 min.
Three times as many compounds appear on heating — and then about 40% of them are gone within three minutes. Rise, then fall. Which is the sealed vial and the open pan, observed in sequence instead of in separate experiments.
Two of their specifics line up with the Korean paste across species, method and country. “3-Methyl-1-butanol, one of the main flavor components of miso soup, was greatly decreased at the initial stage of heating” — the compound that fell from 1,698.3 to 69.0 ng/g in Changwon. And “Methional contributed to the aroma degradation, being newly formed and increased by heating at 80℃” — the compound that rose from 131.1 to 317.5 ng/g in that same table. Two laboratories, two foods, two analytical methods, the same two molecules moving the same two ways.
Japan also has the instruction in writing
Korea’s cooking advice about doenjang circulates without a documented origin. Japan’s equivalent is published by the trade body. The Japan Federation of Miso Manufacturers Cooperatives states it directly:
「みそ」は調理の最後に入れましょう
Add miso at the end of cooking. And it gives the mechanism:
煮立ててしまうと、「みそ」のおいしい香りとうま味成分が蒸発したり、固まって溶けない成分に変わってしまいます。「みそ」が溶けて汁が熱せられたら火を止めてください。
Boil it hard, the booklet says, and the good aroma and umami compounds evaporate, or turn into components that clump and will not dissolve. Once the miso has dissolved and the soup is hot, turn off the heat.
The word the industry uses is 蒸発 — evaporation. The same mechanism the Changwon authors proposed from their data, written as consumer guidance by people selling the product.
For doenjang I could not find a counterpart. The Rural Development Administration’s technical material on traditional pastes covers manufacture and hygiene, not what to do with the finished paste in a pot. Korean food law has the same shape: the Food Code’s section on fermented pastes defines the products and sets four specifications, none of which concerns cooking. Korean standards govern how these pastes are made, not how they are used — the same gap that leaves gochujang’s heat scale voluntary.
The nutrient claim has nothing behind it
“Miso loses its nutrients, flavor, and aroma at high temperatures.” The aroma part is measured. The nutrient part, as far as I can establish, is not measured at all.
I looked for heating experiments on the things that would count — amino acid profile, isoflavones, enzyme activity — in doenjang and in miso. I found none. The sentence travels as a unit because it sounds like one claim, but one half of it rests on two laboratories and the other half rests on nothing I could locate.
The bacteria claim points the other way
The remaining piece of the folklore is that boiling kills the living culture. The one relevant measurement I found reports the opposite of what the claim implies (J. Korean Soc. Food Sci. Nutr. 43(10):1619-1626):
A 1.0~2.0 log reduction in total aerobes of Doenjang and Gochujang sauce was observed upon Joule heating at 80, 85, and 95°C.
A one-to-two log reduction, from a paste that carries eight to nine log while it matures, leaves a very large number behind.
Three caveats belong with that figure, and they are substantial. The samples were sauce products rather than home-made stew. Joule heating passes current through the food and has a different thermal history from a pot on a burner. And the counts were of total aerobes and Bacillus cereus spores — lactic acid bacteria were not enumerated separately. What survives in an actual pot of doenjang jjigae remains unmeasured.
What I could not verify
- How much is lost at three minutes versus twenty. No Korean study varied the heating time; the two above used a single duration each. The 40% figure at three minutes is from miso, not doenjang.
- Any measurement of nutrient loss on heating. Searched for and not found, in either food.
- What survives in a real pot. The microbial figure comes from Joule-heated sauce, counts total aerobes rather than lactic acid bacteria, and cannot be transferred to a home stew.
- Whether the first study’s pan was covered. The paper states the temperature and the time but not the vessel. The authors’ own explanation — that esters left with the steam — implies an open one, but implication is not a method section.
- The cooking conditions in a third Korean study. Joo and Shin (2004) analysed doenjang stew aroma, but compared two stews rather than heated against unheated, and the Korean body text of the PDF is encoded as glyph codes I could not extract, so I could not read their temperature or time.
- Where the Korean version of the instruction comes from. I checked a Rural Development Administration researcher’s own published explainer on doenjang; it contains no sentence about heating, boiling, timing or nutrient loss. The advice is repeated widely in Korean without a traceable source document.
- Whether pyrazines rise or fall. The two studies disagree on this group, and the Changwon table’s pyrazine rows sit in a column layout I did not transcribe directly from the page image, so I have left their figures out rather than risk misreading them.
FAQ
Does boiling doenjang destroy its flavour?
Heating does not destroy the aroma compounds so much as drive them off. One Korean study measured 3-methylbutanol falling from 1,698.3 to 69.0 ng/g after 10 minutes at 100°C, while another, heating the paste inside a sealed vial where nothing could escape, measured total esters rising from 5.83 to 851.87 ppm. The compounds leave with the steam.
Should I add doenjang at the end of cooking?
Japan's miso trade association says so in writing, and explains why — the aroma evaporates. I could not find the equivalent instruction in any Korean government or research-institute document.
Does heating destroy the nutrients in doenjang?
English pages say so. I found no measurement of it — not for amino acids, isoflavones or enzyme activity, in doenjang or in miso.
Does boiling kill the beneficial bacteria?
The one measurement I found points the other way. Joule heating at 80–95°C cut total aerobes by 1 to 2 log, from a paste that carries 8 to 9 log while it ferments. That still leaves a great deal behind, though the sample was a sauce product and lactic acid bacteria were not counted separately.
Sources
- Jeong, Cho & Cha, Volatile flavour compounds in cooked Omandungi-Doenjang stew, J. Life Science 18(11):1570-1577, 2008
- Lee & Ahn (Korea Food Research Institute), Changes in volatile compounds of doenjang on heating, Korean J. Food Sci. Technol. 40(3):271-276, 2008
- Sugawara & Hosaka, Headspace analysis of miso soup volatiles on heating by SPME, J. Home Economics of Japan 60(6):545-552, 2009
- Sugawara et al., Aroma of heated miso suspension, J. Home Economics of Japan 60(5):453-459, 2009
- Japan Federation of Miso Manufacturers Cooperatives, Miso o Shiru (educational booklet, PDF)
- Cho et al., Effect of ohmic heating on microbial reduction in doenjang and gochujang sauce, J. Korean Soc. Food Sci. Nutr. 43(10):1619-1626, 2014
- Joo & Shin, Flavour compounds and sensory characteristics of doenjang stew, Korean J. Food Sci. Technol. 36(2):202-210, 2004
- Ministry of Food and Drug Safety, Korean Food Code, Chapter 5 §12 Jang (fermented pastes and sauces)
- Just One Cookbook, Homemade Miso Soup
Anything not cited above is general cooking knowledge. How these articles are written is explained on the About page.