Thermochimica Acta 458 (2007) 38–46 A comparison of various Gibbs energy dissipation correlations for predicting microbial growth yields J.-S. Liu a,1 , V. Vojinovi´ c a , R. Pati˜ no b , Th. Maskow c , U. von Stockar a,∗ a Laboratory of Chemical and Biochemical Engineering, Swiss Federal Institute of Technology, EPFL, CH-1015 Lausanne, Switzerland b Cinvestav-M´ erida, Departamento de F´ ısica Aplicada, Km. 6 carretera antigua a Progreso, AP 73 Cordemex, 97310 M´ erida, Yucat´ an, Mexico c UFZ Centre for Environmental Research, Department of Environmental Microbiology, Permoserstraße 15, D-04318 Leipzig, Germany Available online 19 January 2007 Abstract Thermodynamic analysis may be applied in order to predict microbial growth yields roughly, based on an empirical correlation of the Gibbs energy of the overall growth reaction or Gibbs energy dissipation. Due to the well-known trade-off between high biomass yield and high Gibbs energy dissipation necessary for fast growth, an optimal range of Gibbs energy dissipation exists and it can be correlated to physical characteristics of the growth substrates. A database previously available in the literature has been extended significantly in order to test such correlations. An analysis of the relationship between biomass yield and Gibbs energy dissipation reveals that one does not need a very precise estimation of the latter to predict the former roughly. Approximating the Gibbs energy dissipation with a constant universal value of -500 kJ C-mol -1 of dry biomass grown predicts many experimental growth yields nearly as well as a carefully designed, complex correlation available from the literature, even though a number of predictions are grossly out of range. A new correlation for Gibbs energy dissipation is proposed which is just as accurate as the complex literature correlation despite its dramatically simpler structure. © 2007 Elsevier B.V. All rights reserved. Keywords: Biomass yield; Metabolism; Thermodynamics of growth; Gibbs energy dissipation; Yield prediction; Growth yield database 1. Introduction Whenever cells are grown in any kind of culture, it is of utmost importance to obtain as high a biomass density as possible. The achievable biomass concentration affects the ease with which scientific research projects can be carried out in that it determines the amount of biological material that can be derived from a given cellular strain. In industrial biotechnology the biomass concentration determines the amount, the synthesis rate and the concentration of the target product that can be expected and thus represents a prime factor influencing the economic viability of the project. The biomass concentration that can be obtained is in turn determined primarily by the growth yield characterizing the respective strain. It is therefore of practical significance to develop methods for roughly predicting the achievable biomass yields even before launching a project and/or carrying out in- depth experimental work. ∗ Corresponding author. Tel.: +41 21 69 33191. E-mail address: urs.vonStockar@epfl.ch (U. von Stockar). 1 Present address: Abraxis BioScience Inc., 2045 North Cornell Avenue, Mel- rose Park, IL 60160, USA. Many different approaches for biomass yield prediction were formulated and reported in the literature. Early work was based on attempts to correlate measured biomass yields in terms of Y ATP [1–3], or in terms of energetic efficiencies [4–9], and many others. In his analysis of thermodynamics of metabolism of Saccharomyces cerevisiae with impaired growth and that of nor- mally growing cells, Battley correlated the biomass yield with the average free energy per C-mole of substrate using the effi- ciency of free energy conservation [5]. In 1972, Minkevic and Eroshin, used the enthalpic efficiency coefficient for biomass yield production. They stated that the energy stored per unit C atom is related to reducing power, i.e. the degree of reduc- tion [4]. Roels showed that biomass yields for aerobic growth appear to depend on the degree of reduction of the carbon and energy substrate [10]. He explained this by pointing out that the bioenergetic growth efficiency has to be inferior to 1 and by showing that this imposes an energy limitation on the biomass yield when microorganisms grow on energy poor substrates, whereas the biomass yields in growth on energy rich substrates can be assumed to be determined by a C-limitation. A more rigorous thermodynamic treatment was proposed by McCarty and later refined [11–14]. Although this method is 0040-6031/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.tca.2007.01.016