What cooking does to a spice
Toasting, frying in fat, simmering, steeping, grinding and souring each change what is in a spice, and a good deal of that has been measured. These are the findings, one study at a time: what was done, under what conditions, what changed, and what the study does not let a cook conclude. Choose what you are doing and to which spice. The chemistry behind them is set out in why spices behave as they do.
Showing all 54 findings.
Toasting and dry roasting · Increases Capsaicin
Poblano, Jalapeño, Serrano, Habanero
Grilling fresh Mexican peppers at 210 °C under household conditions raised their measured capsaicinoid content, by anything from 2.6 to 924.9 per cent depending on the pepper type, where boiling the same peppers had lowered it.
Conditions: in air · 210 °C · Household grilling as done in Mexican cooking
What it means in practice:A charred fresh chilli is not made milder by the flame. Whether the rise reflects water driven out of the fruit, easier extraction from cooked tissue or both cannot be told from the published summary.
What this does not show: Read in the abstract only. The abstract does not say on what weight basis the contents were compared, and a grilled pepper has lost water, so the increase should not be read as capsaicinoid being made. The very wide range includes pepper types with almost none to begin with.
Measured on the spice
Toasting and dry roasting · Decreases Capsaicin
Seeds from Korean dried red pepper held 36.50 mg of capsaicin per kilogram raw. Two minutes of roasting at 100 °C left this statistically unchanged at 32.57 mg/kg; two minutes at 200 °C lowered it significantly, to 19.47 mg/kg. Dihydrocapsaicin fell from 16.27 mg/kg raw to 12.60 at 100 °C and 9.40 at 200 °C.
Conditions: in air · 100 and 200 °C in a convection oven · 2 min · Seeds separated from Korean dried red pepper
What this does not show: Seeds separated from the fruit, which carry very little capsaicinoid to begin with; the flesh and the ground spice were not tested. One lot and one roasting time.
Measured on the spice
Toasting and dry roasting · Forms 2,5-Dimethylpyrazine, 2,3,5,6-Tetramethylpyrazine, 3-Ethyl-2,5-dimethylpyrazine, 1-Methylpyrrole, 2-Acetylfuran
Raw red pepper seeds contained no detectable pyrazines. After two minutes at 100 °C, 3-ethyl-2,5-dimethylpyrazine and 2,3,5,6-tetramethylpyrazine had appeared; after two minutes at 200 °C eight pyrazines were present, 2,5-dimethylpyrazine the largest, and their summed relative peak area had risen from 4.985 to 27.700. Pyrroles, furans, carbonyls and alcohols rose in the same way.
Conditions: in air · 100 and 200 °C in a convection oven · 2 min
What this does not show: Relative peak areas against an internal standard, not concentrations, from one lot of seeds. The roasted and nutty odour notes are assigned from the literature on each compound, not from a tasting of these seeds.
Measured on the spice
Toasting and dry roasting · Forms Pyrazines, Maillard-type volatiles, volatile oil
Aniseed, Caraway, Coriander seed, Cumin, Fennel seed
In raw anise, caraway, coriander, cumin and fennel fruits, oxygenated monoterpenes and related compounds made up about 75 per cent of the volatiles, led by anethole, carvone, linalool, cuminaldehyde and estragole respectively. Roasting brought compounds the raw fruits did not contain: pyrazines and other volatiles of the Maillard browning reaction.
Conditions: in air · Raw and roasted fruits compared by headspace SPME GC-MS; 117 volatiles followed
What it means in practice:This is the measured basis for toasted seed spices smelling roasted and nutty as well as of themselves: the roasted notes are new compounds made in the pan, not the seed’s own aroma intensified.
What this does not show: Read in the abstract only, which does not give the roasting temperature or time, or the amounts of pyrazine formed. Headspace profiles show what is released above the seed, not what it contains.
Measured on the spice
Toasting and dry roasting · Increases Cuminaldehyde
Of five umbelliferous seed spices roasted and compared, cumin showed the greatest intensification of its own character: the level of cuminaldehyde detected in the headspace rose on roasting. Roasted cumin also had the highest peroxide value of the five, 14.2 milliequivalents of oxygen per kilogram, a measure of the oxidation of its fat.
Conditions: in air · Raw and roasted fruits compared by headspace SPME GC-MS
What this does not show: Read in the abstract only; roasting conditions are not stated. A rise in the headspace means more cuminaldehyde is being released from the seed, which may reflect broken oil ducts and not a larger amount in the seed. A study of distilled oil found conventional roasting retained cumin’s aldehydes less well than microwave heating.
Measured on the spice
Toasting and dry roasting · Decreases Estragole
Roasting lowered the level of estragole detected from fennel fruits, the opposite of what the same treatment did to cuminaldehyde in cumin. The estragole level of unroasted fennel also differed according to where the fruit had been grown. Roasted fennel had the lowest peroxide value of the five spices tested, 6.1 milliequivalents of oxygen per kilogram.
Conditions: in air · Raw and roasted fruits compared by headspace SPME GC-MS
What this does not show: Read in the abstract only; no roasting conditions or amounts. The fall was observed in the headspace, and whether the estragole evaporated or was chemically changed is not stated.
Measured on the spice
Toasting and dry roasting · Decreases p-Cymene, Thymoquinone
Thymoquinone and p-cymene were the major volatiles of nigella seed whether raw, conventionally roasted or roasted by microwave, and the levels of both fell with roasting.
Conditions: in air · Microwave roasting at 0.45 kW for 2, 4 and 8 min; conventional roasting · Headspace SPME GC-MS; 38 volatiles identified
What this does not show: Read in the abstract only, which gives no amounts and does not state the temperature of the conventional roasting.
Measured on the spice
Toasting and dry roasting · Forms Methylpyrazine, 2,5-Dimethylpyrazine, Furans
Roasting nigella seed significantly raised its pyrazines and furans. Methylpyrazine and 2,5-dimethylpyrazine were the main pyrazines, formed at high concentration in seed roasted by microwave for 8 min and in conventionally roasted seed, less in seed given the shorter treatments.
Conditions: in air · Microwave roasting at 0.45 kW for 2, 4 and 8 min; conventional roasting
What this does not show: Read in the abstract only; no concentrations. The author suggests these compounds may affect the flavour of the roasted seed, but no sensory test is reported.
Measured on the spice
Toasting and dry roasting
Coriander seed, Fennel seed, Cumin
In Sri Lankan cooking the curry spice base, thuna paha, is either three seeds — coriander, fennel and cumin — or five aromatics, and dry-roasting is one of the pre-treatments that change its appearance and taste. Black curries are made with dark-roasted spices, coriander, fennel and cumin above all, and dried goraka deepens their dark colour.
Conditions: in air · Coriander, fennel and cumin seeds dry-roasted dark for a Sri Lankan black curry
What this does not show: A description of one national tradition, with no temperature, time or degree of roasting given. Nothing here says which compounds dark roasting forms in these seeds; the laboratory findings on roasted coriander and cumin elsewhere in this catalogue measured other conditions.
Documented practice
Toasting and dry roasting · Depends on conditions Cinnamaldehyde, benzaldehyde, γ-caprolactone, isomintlactone, γ-nonalactone
Roasting shifted the headspace of red Egyptian mahlab from aldehydes to lactones. Lactones rose from 6.8 to 24.2 per cent of the volatiles, with γ-caprolactone climbing from 2.14 to 6.58 per cent, isomintlactone from 1.19 to 8.31 and γ-nonalactone from 0.88 to 4.20. Benzaldehyde and (E)-cinnamaldehyde, the leading aldehydes of the raw seed, fell to 0.8 and 1.9 per cent. The authors link the lactone rise to the brief heating with which mahlab is traditionally used in confectionery.
Conditions: in air · 180 °C on a heating mantle · 15 to 20 min · Powdered seed, 5 g wrapped in aluminium foil; red mahlab from Egypt compared raw and roasted
What this does not show: Relative peak areas from headspace sampling, not concentrations, on a few seed lots; the odour notes attached to the lactones come from the literature and no sensory test was made of the roasted seed.
Measured on the spice
Frying and blooming in fat · Depends on conditions Capsaicin, pungency
Chilli powder fried in oil for 10 min gave up its capsaicinoids to the oil to an extent that depended on temperature. The oil held 14.8 mg/kg after frying at 130 °C, 20.9 at 150 °C, 19.4 at 170 °C and 7.4 at 190 °C, which the authors calculate as 73.5, 103.9, 96.4 and 36.8 per cent of what the powder contained. At 130 °C the release was incomplete; at 190 °C only about a third was recovered, which the authors attribute to serious loss of capsaicinoids at that temperature.
Conditions: in oil or fat · 130, 150, 170 and 190 °C · 10 min · Chilli powder fried in four times its weight of oil; proton NMR
What it means in practice:Hot oil does carry chilli heat efficiently, but not at any temperature: in this preparation oil that was too cool left heat behind in the powder, and oil at 190 °C returned little more than a third of it.
What this does not show: One chilli powder, one oil and a single frying time of 10 min; a tempering lasts seconds and a poured chilli oil cools as it meets the chilli. A recovery above 100 per cent shows the size of the measurement error.
Measured on the spice
Frying and blooming in fat · Depends on conditions Capsanthin, Capsanthin 5,6-epoxide, 13-Z carotenoid isomers, colour
Frying chilli powder in oil at 130 °C for 10 min transferred 97.0 per cent of its capsanthin and 93.2 per cent of its total carotenoids into the oil. At higher temperatures far less pigment survived: 37.1 per cent of the capsanthin at 150 °C, 23.1 at 170 °C and 14.9 at 190 °C. As the temperature rose, capsanthin 5,6-epoxide and 13-Z isomers increased, the red carotenoids were lost faster than the yellow, and browning products built up.
Conditions: in oil or fat · 130, 150, 170 and 190 °C · 10 min · Chilli powder fried in four times its weight of oil; proton NMR
What it means in practice:The red of chilli moves into oil almost completely at a moderate frying temperature and is destroyed by a hot one. Pigment and pungency did not share an optimum in this study: the colour was best extracted at 130 °C, the heat at 150 to 170 °C.
What this does not show: One chilli powder and ten minutes of frying, much longer than a spice spends in a tempering. The study did not test temperatures below 130 °C, so it cannot say where extraction begins to fall off at the cool end.
Measured on the spice
Frying and blooming in fat · Converts Capsaicin, Vanillin
The vanillin content of chilli oil rose steadily with frying temperature from 130 to 190 °C. Vanillin is a known thermal breakdown product of capsaicin, and the authors take its rise as confirmation that higher temperature drives the degradation of capsaicin in the oil.
Conditions: in oil or fat · 130 to 190 °C · 10 min
What this does not show: Vanillin can form from other constituents of chilli as well, and the study did not trace it to capsaicin directly. The amounts are far too small to be assumed perceptible.
Measured on the spice
Frying and blooming in fat · Retains Capsaicin, Capsanthin, colour, pungency
The carotenoids and capsaicinoids of chilli held up better under heat in an oil-based system, a hot pot base, than in a water-based one, chilli juice. Ultraviolet light significantly reduced both groups of compound in both systems, and storage at 4 °C significantly slowed their loss.
Conditions: in oil or fat · Hot pot base (oil system) compared with chilli juice (water system) under heat, light and storage
What this does not show: Read in the abstract only, which gives no figures or temperatures. The two systems differ in more than their medium, so the comparison is between two foods, not between oil and water alone.
Measured on the spice
Frying and blooming in fat · Depends on conditions 2-Methylpyrazine, Limonene, Camphene, (E)-2-Heptenal, volatile oil
The temperature at which hot oil is poured over ground chilli changed the aroma of the finished chilli oil significantly. Oil poured at 150 °C gave richer citrus, woody and grassy notes, with stronger signals for terpenes such as limonene and camphene. Oil at 180 °C gave more malty and fatty notes. Oil at 210 °C, and the traditional three pours at falling temperature, gave the strongest nutty aroma, and the signal for 2-methylpyrazine at 210 °C was about twice that at 150 °C.
Conditions: in oil or fat · Rapeseed oil poured at 210, 180 or 150 °C, or in three pours at falling temperature · Dried Erjingtiao chilli ground to pass a 50-mesh sieve · 30 mL of oil over 7 g of chilli; left covered for 24 h
What it means in practice:Cooler oil keeps more of the chilli’s own aroma and hotter oil trades it for roasted notes made in the pour. The traditional practice of pouring in stages at falling temperature sits between the two.
What this does not show: Signal intensities from ion-mobility spectrometry, not concentrations, and one chilli cultivar with one oil. The aroma descriptions come from the instrumental profile and a panel in one laboratory.
Measured on the spice
Frying and blooming in fat · Decreases Crocin, Tricrocin, Crocin-3, Crocetin, 13-cis-Crocetin, colour
Saffron heated in olive oil at 180 °C for 5 min lost about two thirds of its colour compounds: the total fell to 36 per cent of the uncooked level. Crocin dropped to 66 per cent and tricrocin to 26 per cent, while crocin-3 rose slightly and two compounds absent from raw saffron appeared, crocetin and its 13-cis isomer. The pattern is that of the sugars being split from the crocins, followed by partial isomerisation of the crocetin released.
Conditions: in oil or fat · 180 °C · 5 min · 0.08 g of Spanish saffron in 7.5 mL of olive oil
What it means in practice:Saffron’s colour compounds are sugar esters that dissolve in water, and hot oil neither extracts them well nor spares them. The finding supports the usual practice of steeping saffron in liquid, and gives no support to frying it.
What this does not show: One saffron and one frying condition, hotter and longer than a spice usually spends in oil. Peak areas at a single wavelength, so breakdown products that do not absorb there are not counted.
Measured on the spice
Frying and blooming in fat · Decreases Diallyl disulfide, Monoterpenes, Methylpyrrole, Aldehydes from the heated oil, volatile oil
Garlic, Ginger, Sichuan pepper
What is smelled when spices hit hot oil is aroma leaving the pan. Stir-frying garlic released abundant diallyl disulfide into the air; stir-frying ginger and Sichuan pepper released abundant monoterpenes and related terpenoids; all of the spices tested gave off methylpyrrole, and the aldehydes emitted by the heated oil itself rose roughly tenfold once spices were added. A separate study of a test kitchen, in which 16 herbs and spices were shallow-fried in rapeseed oil and a chicken curry was cooked, found the same pattern by stage: aldehydes while the oil heated, monoterpenes while the spices fried.
Conditions: in oil or fat · Spices stir-fried in corn oil; emissions measured in the air by on-line mass spectrometry
What it means in practice:Frying a spice in fat is not only an extraction into the oil. Part of the lightest aroma is driven off into the room in the same moment, which is one reason the time a spice spends in hot fat is kept short.
What this does not show: Both read in the abstract only, and both are air-quality studies: they measured what entered the air and not what stayed in the oil or the food, so the share lost cannot be stated. The first study names its Sichuan pepper as Zanthoxylum piperitum.
Measured on the spice
Frying and blooming in fat · Extracts Linalool, β-Myrcene, γ-Terpinene, Eucalyptol (1,8-cineole), Terpinen-4-ol, Linalyl acetate, Furfural, (E,E)-2,4-Heptadienal, volatile oil
In oil fried with Sichuan pepper, thirteen compounds were identified as the key contributors to aroma, and eleven of them came from the spice: terpenes and terpene alcohols including β-myrcene, γ-terpinene, eucalyptol, linalool, terpinen-4-ol and linalyl acetate, which carried the herbal, piney and citrus notes. Only two, (E,E)-2,4-heptadienal and furfural, were made in the frying, by oxidation of the oil and browning reactions; these carried the fatty and roasted notes.
Conditions: in oil or fat · Fried huajiao oil compared with raw and roasted Sichuan pepper and with heated rapeseed oil
What it means in practice:A fried Sichuan pepper oil is chiefly the spice’s own aroma moved into fat, with a small roasted contribution from the heated oil. This is a different balance from chilli oil, where the pour makes much of the aroma.
What this does not show: Read in the abstract only, which gives no frying temperature or time. One oil and one preparation, and the tingling compounds were not part of the study.
Measured on the spice
Frying and blooming in fat · Depends on conditions 2-Methylpyrazine, (E)-2-Heptenal, 2-Methoxy-3-(1-methylethyl)pyrazine, volatile oil
Sichuan pepper, Star anise, Fennel seed, Bay leaf, Tsaoko, Ginger
In a rapeseed oil fried with a mixture of Sichuan peppercorn, chilli, ginger, green onion, star anise, fennel seed, tsaoko and bay leaf, the frying temperature strongly influenced which volatiles formed. Frying at 160 °C gave the most diverse set, spanning aldehydes, ketones, terpenes, esters and alcohols. Seventy-three compounds distinguished the oil fried at 150 °C from that fried at 160 °C but only thirty-nine separated 160 from 170 °C, which the authors read as high heat reducing complexity.
Conditions: in oil or fat · 150, 160 and 170 °C · 15 min at the temperature the authors selected · Rapeseed oil fried with a mixture of spices, chilli and green onion
What this does not show: Read in the abstract only. One formulation of ten ingredients fried together, so no change can be assigned to any single spice, and three temperatures only ten degrees apart.
Measured on the spice
Frying and blooming in fat
In the traditional vegetarian cooking of South India, seasoning with spices fried in fat is recorded as the last step of a dish, done after the cooking and not at its start. Pongal, rice cooked with green lentils, is finished with generous amounts of cumin seeds and peppercorns sautéed in ghee; curd rice, vegetable dishes, lentil preparations and rasam are each described as being seasoned at the end, rasam with a final sprinkle of asafoetida.
Conditions: in oil or fat · Whole spices fried in ghee and added to the dish once it is cooked
What this does not show: A description of practice in one regional tradition, written to document dishes. It gives no temperatures or times and does not use the words tempering or tadka. No measurement was found of what a tempering lasting seconds extracts from a whole seed, so the chemistry recorded for longer frying should not be assumed to describe it.
Documented practice
Simmering and boiling · Decreases Curcumin
Turmeric boiled for 10 or 20 min, or pressure cooked for 10 min, lost between 27 and 53 per cent of its curcumin, the largest loss coming from pressure cooking. Cooking it together with red gram (pigeon pea) made no difference to the loss.
Conditions: in water · Boiling for 10 or 20 min; pressure cooking for 10 min · Turmeric alone and with red gram; curcumin measured by HPLC after freeze-drying
What it means in practice:Curcumin is not indestructible in a wet dish: a simmered or pressure-cooked dal keeps part of its curcumin, not all of it. The study measured the compound, and says nothing about how the colour of the dish looked.
What this does not show: Read in the abstract only. One laboratory, and the abstract does not give the variety of turmeric, the amount of water or the loss for each treatment separately. Curcumin was measured in a solvent extract of the freeze-dried cooked spice.
Measured on the spice
Simmering and boiling · Converts Curcumin, Ferulic acid, Vanillin, Vanillic acid
When curcumin is degraded by the heat of domestic cooking, the three major products are ferulic acid, vanillin and vanillic acid. The molecule breaks at the diketone bridge that joins its two halves, and at the first carbon of the chain linking the two aromatic rings.
Conditions: Heat treatment of the kind met in domestic cooking; products isolated by LC
What this does not show: Read in the abstract only, which does not state the conditions or the yield of each product. Identifying the products says nothing about whether they can be tasted or smelled in a dish.
Measured on the pure compound
Simmering and boiling · Decreases Piperine
Black pepper boiled for 10 or 20 min, or pressure cooked for 10 min, lost between 16 and 34 per cent of its piperine, with the largest loss in pressure cooking. The loss was somewhat lower when red gram was present.
Conditions: in water · Boiling for 10 or 20 min; pressure cooking for 10 min · Black pepper alone and with red gram or tamarind
What this does not show: Read in the abstract only. The abstract does not say whether the pepper was whole or ground, or what the piperine became. Pungency itself was not tasted.
Measured on the spice
Simmering and boiling · Decreases Piperine, Total extractable oleoresin, pungency
Both piperine and the total extractable oleoresin of black pepper declined on heating between 50 and 120 °C, and the decline followed first-order kinetics: a constant fraction of what remains is lost in each interval of time, faster at higher temperature. The same loss was followed in open-pan cooking, in pressure cooking and in a slow cooker.
Conditions: 50 to 120 °C held steady, and under changing temperature · Also followed in open-pan cooking, pressure cooking and a slow cooker
What this does not show: Read in the abstract only, which reports the form of the loss but no rate constants or percentages, so its size at any given temperature cannot be stated here.
Measured on the spice
Simmering and boiling · Decreases Capsaicin
Dried red pepper boiled for 10 or 20 min, or pressure cooked for 10 min, lost between 18 and 36 per cent of its capsaicin, the largest loss coming from pressure cooking. Neither red gram nor tamarind, nor the two together, changed the loss.
Conditions: in water · Boiling for 10 or 20 min; pressure cooking for 10 min · Dried red pepper (Capsicum annuum), alone and with red gram or tamarind
What it means in practice:Capsaicin is more durable than most aroma compounds but it is not untouched by wet cooking: this study found a measurable part of it gone after minutes of boiling. Most of the heat remained.
What this does not show: Read in the abstract only. The cultivar and the form of the pepper are not stated there, and the fate of the lost capsaicin was not determined. Other studies of fresh peppers report smaller losses on boiling.
Measured on the spice
Simmering and boiling · Decreases Capsaicin
Poblano, Jalapeño, Serrano, Habanero
Across fresh Poblano, Chilaca, Caribe, Jalapeño, Serrano, Habanero and Manzano peppers, boiling at 96 °C under household conditions caused moderate losses of capsaicinoids, between 1.1 and 28.1 per cent. The proportions of capsaicin, dihydrocapsaicin and nordihydrocapsaicin to one another were similar in raw and cooked peppers.
Conditions: in water · 96 °C · Household boiling as done in Mexican cooking; eight pepper types
What this does not show: Read in the abstract only, which does not give the boiling time or say whether the missing capsaicinoids were destroyed or had moved into the water. The range across pepper types is wide.
Measured on the spice
Simmering and boiling · Decreases Capsanthin
In diluted paprika juice held at 80, 90 and 100 °C, capsanthin declined with heating time, and the kinetic model fitted to the data predicted half-lives of 27.47, 21.23 and 15.23 min at the three temperatures for juice containing 2850 mg of capsanthin per litre.
Conditions: in water · 80, 90 and 100 °C · Held for 0 to 16 min · Diluted paprika juice; capsanthin by HPLC
What this does not show: Read in the abstract only. A juice of the fresh fruit, a water system, not paprika powder dispersed in the fat of a dish. The half-lives are model predictions from heating runs of at most 16 min.
Measured on the spice
Simmering and boiling · Retains Crocin, Tricrocin, Crocin-3, Crocetindial, colour
Saffron boiled in water for 20 min kept most of its colour compounds. Crocin stood at 81 per cent of its uncooked level, tricrocin at 74 per cent and crocin-3 at 93 per cent, and the total of the measured crocetin compounds at 88 per cent. The minor compound crocetindial disappeared.
Conditions: in water · 20 min at the boil · 0.08 g of Spanish saffron in 50 mL of water; HPLC peak areas at 440 nm
What it means in practice:Twenty minutes of boiling costs saffron little of its colour. This says nothing about its aroma, which the study did not follow through cooking.
What this does not show: One purchase of Spanish saffron, in plain water with nothing else in the pot, measured as peak areas, not weights. Picrocrocin and safranal were not measured after cooking.
Measured on the spice
Simmering and boiling · Decreases Crocin, Crocetin, colour
Cooked into rice dishes, saffron’s colour compounds were recovered at far lower levels than from saffron cooked alone. A paella-type dish, saffron boiled with rice for 20 min, returned about 20 per cent of the original total; a pilaf-type dish, in which saffron was first fried with the rice in oil and then boiled, returned about 8 per cent and no detectable crocin, most of what remained being crocetin. The authors state that what becomes of the missing compounds is not yet known.
Conditions: Boiled with rice for 20 min; or fried with rice in oil for 10 min, then boiled for 15 min · 0.08 g of saffron with 500 g of rice, 300 mL of water and 30 mL of olive oil
What this does not show: A recovery measurement: compounds bound to rice, or changed into forms the method does not detect, count as lost, and the authors do not claim they were destroyed. The dishes were visibly coloured all the same, and one trial of each was made.
Measured on the spice
Simmering and boiling · Retains Picrocrocin, Safranal
Picrocrocin proved highly stable in water. Its half-life ranged from more than 3400 h in saffron extract at 5 °C to 9 h for the purified compound in water at 100 °C. In the purified compound at 100 °C, safranal was confirmed to form during the first 5 h, with a yield of up to 7.4 per cent.
Conditions: in water · 5 to 70 °C for saffron extracts; 100 °C for purified picrocrocin · Aqueous extracts and purified compound; second-order kinetics
What it means in practice:Over the length of time a dish is cooked, very little of saffron’s bitter compound breaks down, and only a small part of what does becomes aroma. Cooking is not a significant way of making safranal.
What this does not show: Read in the abstract only. Extracts and pure compound in the laboratory, in the dark; the abstract notes that the rate in whole extracts behaved irregularly with temperature, peaking at 35 °C.
Measured on the pure compound
Simmering and boiling · Decreases enzyme activity
The three mustards do not share one enzyme. Myrosinase extracted from brown mustard (Brassica juncea) was the most active, at 2.75 units per mL, against 1.50 for black (Brassica nigra) and 0.63 for yellow (Sinapis alba). Their resistance to heat differed as well: after 10 min at 70 °C the yellow mustard enzyme retained no activity, the brown 35 per cent and the black 59 per cent.
Conditions: in water · 70 °C · 10 min · Myrosinase extracted from black, brown and yellow mustard seed
What it means in practice:Warmth that would silence yellow mustard entirely leaves brown and black mustard still able to make some of their heat. Advice about hot liquid killing the pungency of mustard holds most strictly for the yellow seed.
What this does not show: Read in the abstract only. Enzyme extracted into solution; inside an intact dry seed it is better protected. One source of each seed, and the amount of pungent product formed was not measured.
Measured on the pure compound
Simmering and boiling · Decreases enzyme activity
Partially purified myrosinase from yellow mustard seed (Sinapis alba) began to be inactivated by heat at 60 °C, and the loss of activity between 65 and 75 °C was studied in detail; it did not follow a single simple rate.
Conditions: in water · Inactivation from 60 °C; kinetics studied at 65 to 75 °C · Partially purified Sinapis alba myrosinase in a model system
What this does not show: Read in the abstract only. A purified enzyme in buffer; the figure applies to yellow mustard and should not be carried over to brown or black mustard, whose enzymes another study found more heat-stable.
Measured on the pure compound
Simmering and boiling · Converts Allyl isothiocyanate, Diallyl sulfide, Diallyl disulfide, Diallyl trisulfide, Allyl thiocyanate, N,N'-Diallylthiourea
Mustard seed, Horseradish, Wasabi
Allyl isothiocyanate boiled in water for an hour breaks down. The products that could be extracted from the water included diallyl sulfide, disulfide, trisulfide and tetrasulfide, the same family of compounds found in cooked garlic, together with allyl thiocyanate; the major product left in the water was N,N′-diallylthiourea.
Conditions: in water · 100 °C under reflux · 1 h · Pure allyl isothiocyanate in water
What it means in practice:The sharpness of brown mustard, horseradish and wasabi is lost to long wet cooking by chemical breakdown as well as by evaporation, and the breakdown products are sulfur compounds of the kind that smell of garlic.
What this does not show: Read in the abstract only, which gives no yields. The pure compound in water for a full hour under reflux, where vapour is returned to the flask; in an open pan much of it would leave as vapour first.
Measured on the pure compound
Simmering and boiling · Decreases Allicin, enzyme activity
The activity of garlic alliinase followed a bell-shaped curve with temperature, highest between 35 and 40 °C and falling rapidly above 40 °C. In solution the enzyme was also short-lived: kept at 37 °C it was almost completely deactivated over the storage period tested, and it kept its activity only when frozen or dried.
Conditions: in water · Assayed across a range; highest activity at 35 to 40 °C · Enzyme and substrate held 5 min at each temperature before mixing · pH 8 buffer
What it means in practice:The enzyme that gives raw garlic its bite is undone by quite gentle warmth, which is consistent with garlic heated whole, before it is cut, staying mild.
What this does not show: Extracted enzyme in buffer. The study gives the shape of the curve and not a temperature and time at which a whole clove’s enzyme is destroyed; heat reaches the centre of a clove slowly.
Measured on the pure compound
Simmering and boiling · Decreases Allicin
Allicin in water broke down faster the warmer the solution, across 45 to 95 °C, and faster at higher starting concentration. The loss was best described by two first-order processes running together, which the authors take to mean that allicin degrades by more than one route.
Conditions: in water · 45, 55, 65, 75, 85 and 95 °C · 30 to 1200 min · Synthesised allicin at 20 mg/L in water
What this does not show: Synthesised allicin in plain water. The same study shows that substances present in garlic, certain polyphenols with the enzyme peroxidase and several amino acids, change the rate, so the kinetics of the pure compound are not those of crushed garlic in a pan.
Measured on the pure compound
Simmering and boiling · Depends on conditions Cinnamaldehyde, Coumarin, Cinnamic acid, Cinnamyl alcohol, 2-Hydroxycinnamaldehyde
What hot water draws out of cassia bark shifts with the temperature and the time. Between 70 and 95 °C and between 2 and 12 min, hotter and longer extraction raised the coumarin and cinnamyl alcohol content of the extract and lowered its content of cinnamaldehyde, cinnamic acid and 2-hydroxycinnamaldehyde.
Conditions: in water · 70, 75, 80, 85, 90 and 95 °C · 2 to 12 min · Microwave-assisted extraction of Cinnamomum cassia bark with water
What it means in practice:A long, hot infusion of cassia is not simply a stronger version of a short one. In this study it held less of the compound that smells of cinnamon and more coumarin.
What this does not show: Read in the abstract only, which gives directions and significance but no concentrations. Microwave-assisted laboratory extraction, and cassia only; it does not speak for Ceylon cinnamon, and it does not say whether the cinnamaldehyde was lost as vapour or chemically changed.
Measured on the spice
Microwave heating · Retains Cuminaldehyde, volatile oil
When the volatile oil was distilled from cumin seed heated by microwaves and from seed roasted conventionally, the microwave-heated samples had retained the characteristic flavour compounds, the aldehydes, better than the roasted ones. The authors note that in Indian practice most spices are roasted before they are added to food.
Conditions: Microwave heating at several power levels against conventional roasting at several temperatures
What this does not show: Read in the abstract only; the yields and compositions are in the paper’s tables, which were not read, so no figure can be given. Measured on distilled oil, which counts what remains in the seed, whereas a headspace study that found cuminaldehyde rising on roasting counts what the seed gives off; the two do not contradict each other but cannot be compared directly.
Measured on the spice
Microwave heating · Depends on conditions Pyrazines, Aldehydes, Alcohols, Esters, volatile oil
How fenugreek is heated decides which kind of volatile it gives off. Sprouting, boiling and roasting all increased its aldehydes, alcohols, esters and alkanes. Microwave heating did not; there the dominant volatiles were pyrazines, which the authors describe as carrying a chocolate-like aroma.
Conditions: Raw, sprouted, boiled, roasted and microwaved seed and flour; headspace SPME GC-MS
What this does not show: Read in the abstract only, which gives no conditions or amounts. Sotolon, the compound usually held responsible for the characteristic smell of fenugreek, is not mentioned in it.
Measured on the spice
Steeping and soaking in water · Extracts Crocin, colour
The crocins of ground saffron moved into water and hardly at all into oil. The highest yield in the main series, 85.4 µg per mg of saffron, came from water at 60 °C after 20 min; hotter water and a longer infusion of 30 min both gave less, while pure olive oil never took up more than 3.1 µg/mg. In a further series at lower temperature, water at 20 and 40 °C extracted crocins better than water at 60 °C, and the authors conclude that heating is not needed to extract the colour.
Conditions: in water · 10 to 100 °C · 10, 20 and 30 min · 0.25 g of ground La Mancha saffron in 15 mL of water, olive oil or mixtures of the two
What it means in practice:Saffron gives its colour to water, and warm or even cool water does it as well as hot. The study infused ground saffron with constant stirring, so whole threads left to stand will be slower.
What this does not show: Ground saffron from one protected-origin pool, stirred mechanically in a laboratory incubator. The decline at higher temperature and longer time is attributed by the authors to breakdown of crocins, which they did not measure directly.
Measured on the spice
Steeping and soaking in water · Extracts Picrocrocin
Picrocrocin, the bitter compound of saffron, passed into water easily and was the most stable of the compounds followed: its extraction showed no significant difference across temperatures from 10 to 60 °C. The highest value, 47.6 µg per mg of saffron, was reached at 100 °C in 10 min. Into pure oil it moved only slightly, reaching at most 5.13 µg/mg after 30 min at 100 °C.
Conditions: in water · 10 to 100 °C · 10 to 30 min · Ground saffron in water, olive oil or mixtures of the two
What this does not show: One saffron, ground and stirred. The bitterness of the infusions was not tasted, so the figures describe the compound and not the perceived bitterness.
Measured on the spice
Infusing in oil · Extracts Safranal
Safranal, the principal aroma compound of saffron, behaved in the opposite way to the colour. In pure water it was found only in traces. It was recovered best where oil was present, reaching 12.3 µg per mg of saffron at 80 °C in a mixture of one part water to two of oil. At 100 °C with water present its level fell significantly, while in pure oil it went on rising with temperature and time.
Conditions: in oil or fat · 60, 80 and 100 °C · 10 to 30 min · Ground saffron in water, olive oil or mixtures of the two
What it means in practice:The colour and the aroma of saffron want different media: water takes the one, fat holds the other. A dish that contains both, as most saffron dishes do, has somewhere for each to go.
What this does not show: The fall at 100 °C in the presence of water was measured in the liquid, so it does not distinguish safranal driven off as vapour from safranal broken down. Safranal is also formed from picrocrocin on heating, which the design cannot separate from extraction.
Measured on the spice
Acid and alkali · Retains Curcumin
Cooked in the presence of tamarind, turmeric lost 12 to 30 per cent of its curcumin, against 27 to 53 per cent when it was boiled or pressure cooked without it.
Conditions: in acid · Boiling for 10 or 20 min; pressure cooking for 10 min · Turmeric cooked in the presence of tamarind, a sour ingredient
What it means in practice:The smaller loss with a sour ingredient is consistent with curcumin being more stable in acid than in neutral or alkaline water, which solution studies show independently.
What this does not show: Read in the abstract only. The pH of the cooking liquid is not given, so the protection cannot be attributed to acidity on this evidence alone; tamarind brings other substances besides acid.
Measured on the spice
Acid and alkali · Decreases Curcumin, Vanillin, Ferulic acid, Feruloyl methane
Pure curcumin dissolved in phosphate buffer at pH 7.2 and 37 °C was about 90 per cent decomposed within 30 min. Across pH 3 to 10 the breakdown was pH-dependent and faster under neutral and basic conditions than in acid. Vanillin, ferulic acid and feruloyl methane were identified as minor products, and vanillin increased with time.
Conditions: in water · 37 °C · pH 3 to 10; the figure quoted is for pH 7.2 · Pure curcumin in 0.1 M phosphate buffer
What this does not show: Read in the abstract only. Curcumin dissolved in a buffer is far more exposed than curcumin held inside turmeric particles, starch and fat in a dish; the same abstract reports much slower breakdown in media containing serum. A pharmaceutical study at body temperature.
Measured on the pure compound
Acid and alkali · Decreases Curcumin, Demethoxycurcumin, Bisdemethoxycurcumin
In aqueous solution the three turmeric pigments broke down faster as the solution became more alkaline, the rate rising steeply from pH 7.45 to a maximum at about pH 10.2. The half-lives of curcumin, demethoxycurcumin and bisdemethoxycurcumin were 900, 1700 and 2200 h at pH 7.45, and 0.4, 1.0 and 5.0 h at pH 10.2. Bisdemethoxycurcumin was the most resistant of the three.
Conditions: in alkali · pH 7.45 to above 10.2 · Purified pigments and commercial oleoresin in aqueous solution
What it means in practice:An alkaline ingredient such as bicarbonate of soda is the condition under which turmeric colour is least stable. The figures are for pigment in solution and should not be read as the life of the colour in a batter or a dough.
What this does not show: Read in the abstract only, which does not state the temperature. Its half-life at pH 7.45 is very much longer than the breakdown another solution study reports near the same pH at 37 °C; the two used different media and cannot be reconciled from their abstracts.
Measured on the pure compound
Acid and alkali · Depends on conditions Capsaicin
In a heated model system, capsaicinoid content fell faster at higher temperature whatever the starting concentration. Acidity mattered: at a fixed starting concentration the loss was slowest at neutral pH, faster under alkaline conditions and fastest under acidic ones. Higher starting concentrations slowed the rate of loss as heating went on.
Conditions: pH Acidic, neutral and alkaline compared · Model system; initial capsaicinoid concentrations of 200 to 1000 ppm
What this does not show: Read in the abstract only, which gives directions and no rates, temperatures or pH values. A model system, so it cannot be read as the behaviour of chilli in a vinegar sauce; a cooking study found tamarind made no difference to capsaicin loss.
Measured on the pure compound
Acid and alkali · Depends on conditions 4-Hydroxybenzyl isothiocyanate, Sinalbin (4-hydroxybenzyl glucosinolate), Thiocyanate ion, pungency
Yellow mustard’s pungent product does not last in water. The enzyme releases 4-hydroxybenzyl isothiocyanate from sinalbin, and that compound then breaks down, giving thiocyanate ion. Its half-life was 321 min at pH 3.0 and only 6 min at pH 6.5, and more alkaline conditions shortened it further.
Conditions: in water · pH 3.0 to 6.5 · Sinapis alba seed meal and partly purified extract in buffered water
What it means in practice:For yellow mustard, acid does two different things: it slows the enzyme that makes the heat, and it preserves the heat once made. Mixed with plain water the pungency forms and then fades within minutes to hours; in vinegar what has formed keeps far longer.
What this does not show: Read in the abstract only. A study of seed meal as a soil amendment, not of table mustard, and the pungency was not tasted. It concerns Sinapis alba only; brown and black mustard make a different, volatile isothiocyanate.
Measured on the spice
Acid and alkali · Converts Allyl isothiocyanate, Allylamine, Carbon disulfide, Allyl dithiocarbamate, Diallylthiourea, Allyl thiocyanate
Mustard seed, Horseradish, Wasabi
Held at 80 °C in buffered water, allyl isothiocyanate decomposed at every acidity tested, to different products. In alkaline buffer the major products were allylamine, allyl dithiocarbamate, diallylthiourea and carbon disulfide; at pH 6 and pH 4 carbon disulfide and allylamine were the major ones. It also isomerised in part to allyl thiocyanate.
Conditions: in water · 80 °C · pH Buffers of pH 4, 6 and 8 · Allyl isothiocyanate at 1000 mg/kg
What this does not show: Read in the abstract only, which names the products and not the rates, so it does not show whether acid slows the loss. A buffer model at a single temperature.
Measured on the pure compound
Acid and alkali · Depends on conditions Allicin, Alliin, enzyme activity
Alliinase, the enzyme that makes allicin when garlic is crushed, worked within a narrow band of acidity. Its activity peaked at pH 7.0 and fell away sharply below pH 5.0 and above pH 8.0.
Conditions: in acid · 25 °C · pH Buffers from pH 1 to 8 · Alliinase extracted from garlic, assayed with alliin
What it means in practice:Garlic crushed straight into vinegar or lemon juice meets conditions in which its enzyme works poorly, so less of the sharp compound forms than when it is crushed first and the acid added afterwards.
What this does not show: Enzyme extracted from garlic and assayed in buffer. In a crushed clove the reaction is very fast and the tissue buffers itself, so how much an acid dressing suppresses in practice was not measured.
Measured on the pure compound
Acid and alkali · Retains Allicin, Diallyl sulfide, Diallyl disulfide, Diallyl trisulfide
Allicin was more stable in acid than in alkali. In aqueous solution more of it remained at pH 3.0 to 6.0 than at pH 8.0 to 9.0, and under neutral to alkaline conditions the amounts of its breakdown products were significantly higher.
Conditions: in acid · pH 2 to 9 · Synthesised allicin in aqueous solution
What it means in practice:Acid has opposite effects on the two stages of garlic’s sharpness: it hinders the enzyme that makes allicin and protects the allicin already made. The order in which garlic is crushed and soured therefore matters to the result.
What this does not show: Synthesised allicin in buffer. The abstract and text give the direction of the effect; the size of the difference in a sauce, where other ingredients also react with allicin, was not measured.
Measured on the pure compound
Crushing and cutting · Depends on conditions Capsaicin, pungency
Habanero, Carolina Reaper, Ghost pepper, Jalapeño, Cayenne pepper, Serrano
When the fruits of 21 chilli cultivars from four Capsicum species were dissected, the placenta, the pale tissue that carries the seeds, held the highest concentration of capsaicinoids in every species, followed by the flesh, with the seeds lowest in every cultivar. In Capsicum chinense the averages were 2600 mg per 100 g of dry weight in the placenta, 1738 in the flesh and 748.7 in the seeds. The seeds nevertheless made up about two thirds of the dry weight of the fruit. A separate study of Tabasco pepper found the placenta the tissue richest in capsaicinoid-related compounds, by a factor of about 35.
Conditions: Fruits of 21 cultivars cut by hand into pericarp, placenta and seeds; no heat applied
What it means in practice:Cutting out the pale ribs and core removes the most concentrated source of heat in a chilli; the seeds come away with them, which is probably why the seeds are credited with it. In the hottest cultivars the flesh carries a great deal of capsaicinoid as well, so trimming reduces the heat without removing it.
What this does not show: Seeds dissected by hand had been lying against the placenta, so part of what was measured on them may have come from that contact. Concentrations are per dry weight; because seeds are dry and heavy, their share of a whole fruit’s total is larger than their concentration suggests.
Measured on the spice
Crushing and cutting · Forms Pyrrole-based purple-blue pigments, Pyrrole-based yellow pigments, colour
The blue-green colour that processed garlic sometimes takes on is a mixture of many pyrrole-based compounds, not one pigment. Some are purple to blue, absorbing at 565 to 600 nm, and some yellow, absorbing at 420 to 450 nm; green is what the eye makes of the two together. Each colour compound incorporates two molecules of an amino acid.
Conditions: Model systems reproducing the discoloration of processed garlic; isotope-labelled precursors
What this does not show: Read in the abstract only. Model systems, with the structures described as tentative. The abstract does not address the conditions, such as acidity or the age of the garlic, under which the colour forms.
Measured on the pure compound
Crushing and cutting · Converts 1-Propenyl cysteine sulfoxide, Propyl cysteine sulfoxide, Methyl cysteine sulfoxide, Pyruvate, enzyme activity
When an onion is crushed, the precursor responsible for its tear-producing effect, 1-propenyl cysteine sulfoxide, is consumed almost at once: its hydrolysis was close to complete between 5 and 20 s after maceration. The other two flavour precursors, the propyl and methyl compounds, were only about half consumed after 5 s and reacted no further. How much of each was present, and how far each reacted, depended on the cultivar and on the sulfur the plant had grown with.
Conditions: Measured from 5 s to 2 h after maceration · Rapidly macerated bulbs; precursors by HPLC
What it means in practice:The sharpest part of an onion’s chemistry is over within seconds of cutting, so nothing done to the cut onion afterwards can prevent it; only what is done before cutting can.
What this does not show: Read in the abstract only. Bulbs macerated completely in the laboratory, which breaks far more cells than a knife; the volatile products themselves were not measured.
Measured on the spice
Grinding
Fennel seed, Cinnamon (Ceylon), Cloves, Ginger, Korarima
Spice users in the markets of Wolaita, southern Ethiopia, describe mekelesha as fennel seed, cinnamon, clove, ginger, cardamom and pepper ground together in a mortar and pestle into a powder, which is added to doro wot. The same informants grind ginger, rue seed and garlic in a mortar for the fresh sauce datta.
Conditions: Whole dried spices ground together by hand in a mortar and pestle
What this does not show: What the informants of one district say, with no proportions, particle size or timing. It does not show that pounding gives a different result from milling.
Documented practice
Blanching · Retains Gingerol, pungency
Blanching and freeze-drying had no measurable effect on the [6]-, [8]- and [10]-gingerol content of ginger, whereas cooking and processing it into paste and into senbei crackers significantly lowered all three. Gingerols in every product went on declining gradually in storage.
Conditions: in water · Fresh ginger and ginger products; gingerols extracted in methanol and measured by HPLC
What this does not show: Read in the abstract only, which gives directions and significance but no amounts, temperatures or times. It does not say what the lost gingerol became.
Measured on the spice
How to read a finding
A finding is as general as the study that made it.
Each card is one result from one study. Its conditions are given in the study’s own terms, because a laboratory oven held at a fixed temperature is not a pan, and ten minutes in oil is not a tempering of a few seconds. A card marked as measured on the pure compound concerns the molecule in solution and not the spice. Where only the summary of a paper could be read, the card says so and states only what the summary states. Where a card says what a finding means in practice, that reading is ours and is labelled as judgement. One kind of card is different: a documented practice records what cooks do and measures nothing.
Where no finding appears for a spice or a treatment, none has been read. It does not mean that nothing happens.