Technology
From a molecular fingerprint to meaningful insight.
Mid-infrared light identifies molecules. Upconversion makes the signal readable. Chemometrics turns it into live data.
Why mid-infrared?
In the mid-infrared, molecules absorb light at wavelengths that are characteristic of their chemical structure. That makes it possible to tell specific compounds apart — a molecular fingerprint — rather than inferring them from indirect parameters like gravity, temperature or pressure.
The challenge has been reading that signal practically: mid-infrared detectors are typically costly, noisy or require cooling.
Optical upconversion
Upconversion uses a nonlinear crystal to shift the mid-infrared signal into the near-infrared, where standard, mature detectors can capture it. The spectrum is then converted into a concentration using chemometrics.
The approach builds on more than 15 years of research in infrared upconversion spectroscopy at DTU.
- 01
Molecular absorption
Mid-infrared light passes through the sample. Target molecules absorb at unique wavelengths, creating a distinctive fingerprint.
- 02
Optical upconversion
A nonlinear crystal shifts the mid-infrared signal to near-infrared, where standard detectors can read it.
- 03
Data interpretation
A spectrometer captures the signal and chemometrics converts the spectrum into a concentration reading.
First application
Diacetyl in fermentation
Diacetyl is a natural by-product of yeast metabolism that creates a buttery off-flavour at very low concentrations — production thresholds are often around 50 parts per billion. Fermentation can’t be completed until it falls below threshold, yet it is still measured through manual, offline lab sampling.
Brewers therefore add conservative safety buffers. Direct, inline molecular measurement could replace guesswork with live insight. Beyond brewing, the same platform could reveal molecular changes in other industrial processes.
Development stage
Demonstrated
Initial spectral evidence
Lab measurements of diacetyl in water (10,000 ppb, liquid phase) show a distinct divergence from a pure water reference at around 1700 cm⁻¹ (5.88 µm), supporting the feasibility of detection against a high-water background.
2026 · Ongoing
Proof-of-concept validation
Validating the preferred technical route for real-time diacetyl monitoring: limit of detection, calibration, repeatability and choice of gas or liquid phase measurement.
2027 · Target
Pilot
A controlled pilot on one to three fermentation tanks with a lead brewing partner.
2028 · Target
Reference installation
Prototype on selected production tanks, compared against existing lab analysis.
2029+ · Target
Scale through partners
Deployment through brewery accounts and equipment and integration partners.