Patent Watch: Flavoring compositions having a roast aroma profile – US 2026/0090576 A1

Applicant: Firmenich SA Priority date: September 21, 2022 US filing date: September 14, 2023 Published: April 2, 2026 Status: Pending

Source: US 2026/0090576 A1 on Google Patents. All figures below are taken from the published application.

Anyone reading this application should keep two things apart: what the specification discloses, which is a great deal, and what the US claims currently cover, which is considerably less.

1. The core idea

The application describes a way to make a roast or smoke flavor from cereal bran rather than wood smoke.

Conventional smoke flavors are made by burning or pyrolyzing wood and condensing the smoke. That delivers the smoky, roasted phenols flavorists want, but high-temperature combustion also produces polycyclic aromatic hydrocarbons (PAHs) such as benzo[a]pyrene.

Firmenich's alternative is to dry-roast cereal bran under controlled conditions, collect the phenolic distillate that comes off, optionally add water afterward to recover more volatiles, and then blend the bran-derived phenols with other roast and meat aroma compounds.

The chemistry behind this is straightforward. Bran is rich in ferulic, coumaric, sinapic and related phenylpropanoic acids, and heating them produces phenols associated with smoke, char, grilled food and roasted meat. Wood fires run at around 600°C. The preferred treatment here is 200–250°C, ideally no higher than about 235–240°C.

The stated goal is a roast flavor for meat, bacon, pork and meat analogues that carries very little PAH or acrylamide.

2. The flavor chemistry

The composition is built in two layers.

The specification lists further optional materials: dimethyl-, trimethyl-, ethylmethyl-, ethyldimethyl- and acetylpyrazines, furfuryl mercaptan, furfuryl thioacetate, methylfurfural, acetic acid, 2-methyl-3-furanthiol, tetrahydromethylfuranthiol, trimethylamine and sotolone.

In other words, the product is not simply roasted bran. It is a bran-derived phenolic smoke fraction used as a base and then built up with familiar roast, Maillard and meaty aroma compounds.

One detail worth noticing: the main analytical tables quantify 13 phenols, but the first-group list in the claims has 16. The three extras, 4-vinylphenol, 4-propylguaiacol and 2,6-dimethoxyphenol, appear in the broader disclosure and claim language but were not measured in the principal experiments.

3. Keeping contaminants low

Avoiding the toxicant profile of conventional smoke condensates is a theme throughout the document.

The specification describes compositions with less than 300 ppb of PAH and/or acrylamide, with tighter preferred limits. For benzo[a]pyrene, it gives <10 ppb, then <5, <4, <3, <2 and <1 ppb. For acrylamide, the preferred levels step down from <300 to <200, <100, <50, <20 and <10 ppb.

Claim 1 is stricter than parts of that description. It requires less than 300 ppb PAHs and less than 300 ppb acrylamides, not one or the other.

4. Choice of bran

The disclosure covers rice, corn, wheat, oat, rye, barley, sorghum and spelt bran. Rice, corn and wheat are preferred, and corn emerges as the favorite. It gave the highest phenol yield and the best sensory results, it is widely available as a by-product of starch processing, and, unlike wheat, the inventors do not regard it as a major allergen.

Total yield of the 13 measured phenols after five hours at 235°C, per 100 g of bran:

Corn produced more than twice the phenols of rice in this experiment. Its profile was also different: corn was relatively high in 4-ethylguaiacol and 4-vinylguaiacol, while rice bran distillate was dominated by guaiacol.

5. The manufacturing process

The absence of water during roasting is deliberate. According to the inventors, water alters the reaction chemistry and can affect both aroma formation and contaminant levels.

Water is added on purpose after roasting, though, and this step turns out to matter a good deal in the experiments, as the next section shows.

6. Why the post-roast water step matters

In a representative run, 100 g of rice bran was roasted at 235°C for five hours, with condensate collected in five fractions over that period. Then 20 mL of water was added and distillation continued for another 30 minutes to give a sixth fraction.

The total was about 33 mL of distillate: 20.6 mL from the dry roast and 12.4 mL after the water was added.

That last fraction contained substantial additional 4-ethylphenol and 4-ethylguaiacol. The inventors put this down at least partly to the compounds' relatively high boiling points and conclude that hydrodistillation after roasting improves total phenol recovery.

For a flavor chemist, this is one of the more useful practical points in the document: use dry heat to generate the chemistry, then use water to strip out the higher-boiling phenols that would otherwise stay behind in the bran.

7. Temperature and time

Phenol yield generally rose with both time and temperature. For rice bran heated for five hours, total target phenols per 100 g of bran went from:

• 7.6 mg at 200°C
• 13.5 mg at 220°C
• 23.2 mg at 225°C
• 24.5 mg at 230°C
• 27.3 mg at 235°C
• 33.7 mg at 250°C

Guaiacol was the main product, at about 17 mg/100 g at 235°C compared with 8.1 mg at 220°C. Other significant contributors were 4-ethylphenol, 4-ethylguaiacol and 4-vinylguaiacol.

Although 250°C gave the highest yield, the application settles on about 235°C. Part of the stated reason is that the inventors wanted to stay below a temperature they link to "natural flavoring preparation" under the European regulatory framework they cite. That rationale is the applicant's own.

8. Atmosphere and pressure

The inventors also tested different atmospheres and pretreatments.

A gentle flow of either air or nitrogen increased total phenol formation compared with the reference run.

Vacuum at around 100 mbar generally lowered output, with one notable exception: 4-vinylguaiacol rose roughly eightfold. The inventors suggest that at reduced pressure this relatively early reaction product distilled off before it could break down further.

Pressure pretreatment with water or formic acid did not help enough to be worth pursuing.

The practical takeaway is that pressure and gas flow can be used to steer the phenolic profile, not just the yield.

9. Contaminant results

The key safety data come from rice bran distillates made at 235°C for five hours.

In the five dry-roast fractions, neither benzo[a]pyrene (limit of quantification 0.5 µg/kg) nor acrylamide (limit of quantification 10 µg/kg) was detected.

In the post-water hydrodistillate, benzo[a]pyrene was again not detected, but acrylamide came in at 109 µg/kg (109 ppb).

The application compares that figure with the 300 ppb EU benchmark it cites for certain bran-containing breakfast cereals. It also points out that because the flavor is used at low levels, acrylamide in the finished food would be diluted further.

These results are the main experimental support for the low-contaminant claim.

10. Sensory evaluation

Flavorists evaluated rice, wheat and corn prototypes made by roasting at about 235°C for three hours, then adding water and distilling. All three gave a recognizable roast character at a tasting level of around 500 ppm. Wheat and corn scored better than rice, and corn was preferred overall.

The specification says the flavors can add roast, charred or smoky notes to savory products, particularly pork and bacon, and also mentions uses such as whisky.

11. Building a finished flavor

Example 2 shows that the bran distillate is meant to be a base, not necessarily a finished flavor in its own right.

The preferred corn bran extract is topped up with methylcyclopentenolone, "furfuryl," furfural, furaneol, thiazoline, ethyl lactate, thymol and trimethylamine. It can then be formulated further with yeast extract; organic acids such as tartaric, malic, citric or lactic acid, or sodium diacetate or disuccinate; herbs and spices; garlic and onion powder; and sugar beet powder.

For specific end uses, the roast base can be combined with an existing ham, beef, chicken, turkey, hot dog, burger or sausage flavor.

The result is effectively modular: a bran-derived smoke and char base, plus Maillard and roast compounds, plus savory ingredients, plus a target meat profile.

12. Applications

The list of possible end products is long: seasonings and sauces, meat and seafood, soups, noodles, rice, pasta and potatoes, cheeses and fats, snacks, meat and dairy analogues, pet food and beverages. Both conventional meat and plant-based products are covered.

For general food use, the specification gives a very wide range of 0.001–1000 ppm, with 1–100 ppm described as most preferred in one passage. The meat-substitute formulations elsewhere in the document use much higher percentages of a finished flavor composition.

One embodiment prefers adding the flavor before the food is cooked, roasted or grilled, so that it becomes part of the cooked flavor profile.

13. Meat alternatives

A large share of the specification deals with meat analogues. It describes combining the roast composition with plant proteins (pea, soy or wheat), plant fibers, non-animal fats, binders, umami materials, salts, maskers and bitterness blockers, colorants, and optional iron or heme systems.

The worked examples include a vegan hot dog, a vegan burger-style product and chicken nuggets. Versions containing the roast composition were reported to have a more pleasant roast aroma than controls without it.

14. What the current US claims cover

For anyone assessing scope, this is the section that matters most.

The US claims on Google Patents run to 19, but many have been canceled. The live claims are 1, 4, 7, 8, 9, 16, 18 and 19.

Claims 2, 3, 5, 6, 10–15 and 17 are shown as canceled.

Reading claim 1

Claim 1 does not require all 16 first-layer phenols; one is enough, although a real composition would likely contain many. Likewise, only one of the eight second-layer compounds is required.

What it does require is the combination. A bran phenolic extract on its own does not fall within the literal wording of claim 1 unless it also contains one of the second-layer compounds.

The claim also has a product-by-process element, since the first-layer compound has to be obtained by roasting cereal bran.

15. Specification versus claims

The international PCT application, WO 2024/061741 A1, originally included claims covering:

• use of the composition to provide roast aroma
• heating cereal bran at 200–250°C for 1–5 hours without added water
• adding the second-layer flavor compounds
• heating under air or nitrogen
• adding water afterward and hydrodistilling
• recovery by condensation, hydrodistillation or steam distillation

In the US listing, claims 10–15 are canceled. The roasting process is still described in detail in the specification, but no independent US manufacturing process claim survives among the claims currently shown. US protection now rests on the low-contaminant composition, consumer products that contain it, and its use to impart roast aroma.

That difference is significant for anyone reading the application for competitive intelligence or freedom-to-operate purposes.

16. The inventive concept in summary

Taken as a whole, the technical teaching comes down to this. Cereal bran serves as a renewable, food-grade source of phenolic precursors. Roasting it at moderate temperatures without added water converts its phenylpropanoic acids into smoke and roast phenols, which are collected as a distillate instead of being generated from burning wood. Adding water after roasting recovers more of the higher-boiling phenols. The phenolic fraction is then built up with complementary roast, Maillard and meaty compounds to give a roast flavor low in PAHs and acrylamide, suitable for meat, plant-based meat and other foods.

Several supporting findings sit around that core:

• Corn is the preferred bran.
• Gas flow raises phenolic yield.
• Reduced pressure can shift the profile toward specific compounds, notably 4-vinylguaiacol.
• Post-roast hydrodistillation improves recovery.
• The resulting base is modular and can be developed into bacon, pork, ham, beef, chicken, grilled, charred or smoke flavors.

Takeaway

This is not a patent on a single bacon flavor. It is a platform application for making a smoke and roast phenolic flavor system from roasted cereal bran as an alternative to wood-derived liquid smoke, pairing that roast fraction with other aroma compounds and keeping PAH and acrylamide levels low.

Commercially, claim 1 is the one to watch: a specified bran-derived phenol, plus one specified secondary aroma compound, with PAHs and acrylamide each below 300 ppb. Claims 7–9 cover the corn, wheat and rice versions. Claim 16 extends to foods containing the composition, claim 18 covers use at 10–2000 mg/kg, and claim 19 covers using the composition to give a product roast aroma.

###