Research
The brownie experiment
Upcycling Coffee Silverskin as a Functional Ingredient in Brownies
Miraya R. Dadabhoy · Food Marketing & Technology, September 2026
Read the full article(opens external site in a new tab)
Just like peanuts, coffee beans also have an outer skin like coating which flies away during the roasting process.
Where it began
A product-development internship at Beachville Coffee Roasters in Chennai, where she saw silverskin being discarded after roasting.
Apr 2025 – Jul 2026. The internship covered the science of roasting, including caramelisation and the Maillard reaction, and coffee product development: she developed a banana latte blend.
Step 1: The question
What if something a coffee roastery throws away could become a useful ingredient?
Step 2: The material
Coffee silverskin is the thin outer skin of the coffee bean. It comes away during roasting and is usually thrown out. Published research credits it with dietary fibre and phenolic antioxidants, and its coffee notes sit naturally with cocoa.
Step 3: The idea
Replace part of the flour in a cocoa brownie with silverskin, keeping the combined weight of flour and silverskin constant.
Step 4: The experimentation
- First round. Differences were too subtle to measure meaningfully.
- Second round. Bitterness, graininess and dryness rose sharply at each step, which set the upper boundary.
- Consumer trial. Five coded samples, tasted blind.
Three rounds of testing
Silverskin levels tested in each round. The consumer trial is highlighted.
- First round
- 0% to 5% in steps of 1
- Second round
- 0% to 25% in steps of 5
- Consumer trial
- 0%, 2%, 4%, 6% and 8%
She ran more than 15 trials under the mentorship of Harsha Pore, a food technologist and sensory research manager.
Step 5: The blind tasting
40
Tasters recruited
31
Valid responses
A blind tasting with members of the public outside a café. Samples were coded, served in random order, with water between each.
What they rated
- Overall liking
- Flavour
- Texture
- Appearance
- Aftertaste
- Bitterness
- Graininess
- Moistness
- Purchase likelihood
Step 6: What she found
Overall liking: lowest and highest
Mean overall liking on a 1–5 scale. 0%, 4% and 6% fell within this range; individual values were not reported. Differences were not statistically significant.
- 2% silverskin (lowest)
- 3.45
- 8% silverskin (highest)
- 3.79
- 0%, 4% and 6%
- Within range; not reported individually
Which sample did tasters prefer?
10 of 31 valid participants preferred the control (0%), more than any other sample.
- Preferred the control
- Preferred another sample
- Preferred the control (0%)
- 10 of 31
- Preferred another sample
- 21 of 31
Graininess: 0% and 8%
Mean graininess: 1.65 at 0%, 2.52 at 8%. Bitterness followed a similar pattern to graininess (values not reported).
- 0% silverskin
- 1.65
- 8% silverskin
- 2.52
Hearing the story behind the ingredient raised stated interest in buying again, but taste, texture and price remained the reasons cited most often.
“…the 8% formulation should be read as a promising upper working level, not a proven optimum.”
Step 7: What happens next
- Refine the recipe within the 6–8% range
- Control batch-to-batch variation
- Carry out nutritional and safety analyses before wider use
Step 8: The bigger idea
Food waste can sometimes be reconsidered as a resource.
“We live in a world which throws away a third of the food we grow…”