# Monitoring of the Biotechnological Production of Dihydroxyacetone Using a Low‐Field 1H NMR Spectrometer

## TL;DR

This paper introduces a real-time monitoring system using low-field NMR to track the biotechnological production of dihydroxyacetone, improving resource efficiency in cosmetic industry processes.

## Contribution

The novel use of low-field 1H NMR spectrometers for real-time bioprocess monitoring of DHA production is presented.

## Key findings

- Low-field 1H NMR spectrometers enabled rapid, real-time monitoring of DHA production.
- A univariate model tracked substrate and product concentrations with independent calibration.
- Principal component analysis allowed semi-quantitative process monitoring without calibration.

## Abstract

The concept of sustainable production necessitates the utilization of waste and by‐products as raw materials, the implementation of biotechnological processes, and the introduction of automated real‐time monitoring for efficient use of resources. One example is the biocatalyzed conversion of the reusable by‐product glycerin by acetic acid bacteria to dihydroxyacetone (DHA), which is of great importance to the cosmetic industry. The application of compact spectrometers enables the rapid measurement of samples while simultaneously reducing the consumption of resources and energy. Yet, this approach requires comprehensive data preprocessing and, on occasion, multivariate data analysis. For the process monitoring of the production of DHA, a low‐field 1H nuclear magnetic resonance (NMR) spectrometer was implemented in on‐line mode. Small‐volume samples were taken from a bypass and transferred to the spectrometer by an autosampler. Complete analysis within minutes allowed real‐time process control. To this purpose, reliable automated spectral preprocessing preceded the creation of a univariate model. The model enabled the acquisition of process knowledge from chemical kinetics and facilitated the tracking of both substrate and product concentrations, requiring independent calibration. As a second multivariate approach, principal component analysis was utilized to monitor the process in a semi‐quantitative manner without the necessity for calibration. The results of this study are beneficial for real‐time monitoring applications with the objective of exerting control over the process in question while minimizing expenditure.