A simple method for the measurement of metabolic heat production rates during solid‐state fermentations using β‐carotene production with Blakeslea trispora as a model system

Abstract

Solid‐state fermentation (SSF) technology has been rapidly developed for the past 10 years as a production platform for secondary metabolites, biofuels, food, and pharmaceuticals. Yet, the main drawback of SSF is the local temperature rise of up to 20 K, which potentially reduces the strain activity and inactivates heat sensible products. Due to the low heat capacity and thermal conductivity of mixtures of air with plant material, in comparison to aqueous suspensions in submerged fermentations, heat from metabolic processes is less efficiently dissipated. The exact knowledge of the metabolic heat generation during SSF processes is crucial to guide strategies against overheating. In this work, a simple method using a cost‐efficient multichannel instrument is proposed, which allows the determination of heat generation during SSF processes. This method was successfully tested and validated with Blakeslea trispora producing β‐carotene during growth on barley. Additionally, the consequences of the generated metabolic heat during SSF on temperature rise and water evaporation were discussed. Finally, changes in growth and product concentration could also be detected by the heat signal, implying the potential as a timesaving screening method.

Abbreviations

CDW

cell dry weight

DW

dry weight

MD

maximal deviation

OUR

oxygen uptake rate

RQ

respiration quotient

SFO

sun flower oil

SSF

solid‐state fermentation

1. Introduction

Solid‐state fermentation (SSF) attracts more and more attention as a potential production technology for biologically active secondary metabolites, fuels, food, industrial chemicals, and pharmaceutical products. The microorganisms grow on inert carriers with a nutrient film or moist solid plant materials with no or minimal‐free water. The interspace consists of air. SSF resembles the natural habitat of many microorganisms, especially fungi, and thus facilitates growth 1, 2, 3, 4.

Still, a major challenge is the local temperature rise of up to 20 K, which causes enzymes to lose activity and decomposes temperature sensible products, like β‐carotene 5, 6. The size of the temperature rise is a result of several effects. First, the low water content of the substrates in SSF causes a serious problem in heat removal, because the thermal conductivity of water is approximately 21 times higher than the thermal conductivity of air 7 and 4–5 times higher than the one of barley depending on the moisture content 8. Second, the heat capacity of media in submerged fermentation (mainly water with 4.19 kJ kg⁻¹ K⁻¹) is usually much higher than the heat capacity of typical SSF media (barley with 70% water: 3.38 kJ kg⁻¹ K⁻¹ and water saturated air: 1.05 kJ kg⁻¹ K⁻¹) 9, 10. A small heat capacity of the media results in a higher temperature rise for a given metabolic heat. Third, high growth rates especially in the early phase of exponential growth (when no or small limitations in nutrients and oxygen can be observed) result in a high metabolic heat production rate. Finally, the usual poor mixing in SSF hinders the transport of the evolved heat 11.

Applied measures to prevent overheating are, for example, heat exchange at the wall, at additional internal cooling installations (e.g. baffles or mixing devices), forced aeration, and moisture evaporation. For the best design of the chosen devices, the exact knowledge of the metabolic heat released during fermentation must be available for the model‐driven development of SSF processes 12. Up to now, the metabolic heats are typically roughly estimated from respirometric measurements using the oxycaloric equivalent of approx. –460 kJ mol‐O₂ ⁻¹ 12, 13, 14. It is known, that mixed fermentative‐respiratory metabolisms and biological combustion processes, which deviate from the oxycaloric equivalent, can lead to considerable miscalculations of the metabolic heat from respirometric measurements 15, 16, 17. Temperature increases in SSF, especially in large‐scale processes, could be more accurately predicted and prevented, if reliable, directly measured metabolic heat data were available. Isothermal calorimetry features performance characteristics that make it a very interesting candidate for measuring the demanded heat values. The method is a continuous real‐time analysis 18 and provides further advantages, for instance, small sample volumes, high reproducibility, and sensitivity 19, and applicability to any prokaryotic and eukaryotic strains 20, 21. However, the pioneering works with fungi did not consider typical SSF processes. Those were mainly focused on growth on agar, on food 22, and building materials 23, 24, 25. Therefore, a measurement method based on isothermal calorimetry for SSF processes has to be developed and to be tested.

To evaluate the method, we chose a SSF process with Blakeslea trispora producing β‐carotene because carotenoids and, in particular, β‐carotene are reported to have positive effects on the health of different animals 26, 27, 28. There is a broad literature foundation that aims at increased productivity and efficiency of β‐carotene production 29, 30, 31, 32. Nevertheless, reports of SSF with B. trispora are not available yet. Some scientific reports on carotene production based on SSF with other, less productive organisms, like Mucor sp. 33, Penicillium sp. 34, and Rhodotorula glutinis 35, 36 do exist. However, those publications did not have the intent to tackle the problems of SSF due to limited mass‐ and energy transport 37, 38.

In summary and based on the current state of the art, the following tasks emerge. A calorimetric method to measure the metabolic heat of SSF processes has to be developed and validated. Potential correlations between the metabolic heat production of fungi and the rates of growth and product formation will be investigated. For that purpose, isothermal calorimetry will be tested on the example of β‐carotene production by Blakeslea trispora growing on barley using SSF. In addition, the ability of multichannel, cost‐effective isothermal calorimeters (cement calorimeter) will be tested and be proven for screening with the aim of optimizing SSF conditions.

2. Materials and methods

2.1. Strains

Blakeslea trispora plus mating type (ATCC 14059) and B. trispora minus mating type (ATCC 14060) were obtained from Leibnitz Institute DSMZ‐German Collection of Microorganisms and Cell Cultures and precultured as proposed by Roukas et al. 39.

2.2. Media and cultivation

Storage‐stable barley from a local supplier with a moisture content of 10% was mixed with deionized water and if necessary with additives. The mixtures of barley, water, and additives were autoclaved at 121°C for 20 min. Heat sensible additives like β‐ionone and thiamin were added as sterile filtered stock solutions prior to inoculation. Generally, additives were added in such a way that the moisture content of the samples was always identical at 70% ± 1% (maximal deviation—MD) after autoclaving and inoculation. Information on additives and their dosages was taken from different references and adjusted to SSF (Table 1).

Table 1.

List of applied additives

Additive	Concentration	Dimension	Reference
Sunflower oil (SFO)	10–50	g/kg	48, 55
Imidazol	0.5	g/kg	56
β‐Ionone	0.3	mL/kg	40
Thiamine	0.1	g/kg	57
Sodium malate	0.5	g/kg	58

All concentrations are related to dry weight (DW).

Ten milliliters of a sterile 2% span‐20 solution were used to rinse the spores from the petri dish. The spore count in the suspension was estimated with a cell counting chamber under microscope at a magnification of 400. The suspension was subsequently centrifuged and the pellet was suspended to obtain the according concentration for inoculation. The prepared substrate was inoculated with 0.5 mL spore stock solution to obtain a final concentration of 10⁵ spores g⁻¹ dry weight with a mating type ratio plus to minus of 1:10 as proposed earlier 40 and incubated at 28°C.

Three types of experiments were conducted. First, isothermal calorimetry was used to detect the microbial heat production. Secondly, since the sample amount in the ampoules, used for the calorimetric experiments, was too small to generate reliable biomass and product concentration data, growth experiments in falcon tubes were performed simultaneously. To ensure gas exchange, the screw caps on the falcon tubes were left partly open. Thirdly, the oxygen demand was analyzed in a respirometer for the basic substrate over a period of 10 days. In the calorimetric ampoules, 1 g of substrate was used. For the 50 mL falcon tubes as well as the respirometric analysis each, 20 g of substrate were used. These amounts guarantee a thin substrate layer of 0.5 cm to a maximum of 1.0 cm to minimize oxygen limitation during the cultivation time 19. Information about the oxygen limitation is given in the Supporting Information. Every day samples were taken in duplicate. The complete content of a falcon tube or respirometer bottle was thoroughly ground to ensure homogeneity for further analysis.

2.3. On‐line analysis

For off‐ and on‐line analysis the experimental error is given as SD or as MD. A cement calorimeter MC CAL^® (C3 Prozess‐ und Analysentechnik GmbH, Haar b. München, Germany) was used to monitor the metabolic activities of B. trispora on‐line. The instrument has 12 independent measurement channels that can be equipped with reaction ampoules up to 40 mL. The manufacture declares a thermal detection limit of 2·10⁻⁵ W. Figure 1 shows the instrument (A), the working principle (B), and a typical reaction vessel (C).

Figure 1.

Schematic diagram of the used calorimeter. (A) Picture of the instrument. (B) General, although simplified working principle. (C) Picture of ampoules.

Details of the instrument can be found at http://www.c3-analysentechnik.de/zementkalorimeter.php. The working temperature was 28.00 ± 0.01°C (MD). All measurements were done against 0.4 g water as reference, chosen to have the same order of magnitude of heat capacity as the sample. The heat capacity of the substrate was taken from Disney 9. The calorimeter was regularly calibrated using the internal electrical heater (Joule heat) located directly atop the Peltier elements. For further characterization of the measuring performance of the calorimeter, two types of nonbiological experiments were conducted. In the first set of experiments, the heat of a chemical reference reaction was measured (the base‐catalyzed hydrolysis of methyl paraben) and then compared to the heat flow calculated by using reference values for reaction enthalpy (–50.5 ± 4.3 kJ/mol—SD) and rate constant ((3.15 ± 0.11) ·10⁻⁴ s⁻¹—SD) from literature 41. In the second set of experiments, the Joule heat (P) of a 14.5 Ω resistor located inside of the water filled measuring ampoule was measured and compared with the calculated Joule heat. The heater was driven by a constant current source (10.6 mA). That current was monitored via a digital multimeter (Voltcraft VC270, Conrad Electronic AG, Wollerau, Switzerland) throughout the heater experiment. The maximal instrumental error was 0.05 mA and 0.4 Ω for the current (I) and for the resistance (R), respectively. The Joule heat was calculated using (Eq. (1)).

$$P=R{I}^2$$ (1)

The ampoules used as calorimetric vessel (Fig. 1C) were made of polypropylene. The total filling volume of the applied ampoules was determined to be 29.1 ± 0.2 mL (n = 12, SD).

The oxygen uptake rate (OUR) at 28°C was determined at the same time by a respirometer BSBdigi (SELUTEC GmbH, Hechingen, Germany) with 20 g inoculated substrate. CO₂ was absorbed by a mixture of solid NaOH and Ca(OH)₂. Consumed oxygen is immediately replaced by electrolytically produced oxygen to ensure a constant oxygen partial pressure. The electric signal is directly converted to oxygen demand with the help of the package software BSBdigi NT/LF (SELUTEC GmbH, Hechingen, Germany).

The respiration quotient was analyzed by the gas sensors for carbon dioxide BCP‐CO₂ and oxygen BCP‐O₂ with a range of 0 to 10 Vol% and 0.1 to 25 Vol%, respectively (BlueSens gas sensor GmbH, Herten, Germany). The software FERMVis (BlueSens gas sensor GmbH, Herten, Germany) was applied for data collection.

2.4. Off‐line analysis

The fungal biomass (c_CDW) was indirectly estimated by measuring the glucosamine content (c_GS) as proposed earlier 42, 43. The method was slightly modified and the modifications are detailed in the Supporting Information. The constant growth rate of the filamentous fungi B. trispora (+) and (–) mating type can be approximated by a strict linear correlation of biomass versus time (slope).

For product analysis, a sample of 4 g was lyophilized and afterwards extracted twice with 30 mL acetone with 1% butylhydroxytoluene as antioxidant. The extract was concentrated by a rotary evaporator, and suspended in acetone to achieve a volume of 1.5 mL solvent. The extract was filtered through a 0.22 μm PTFE membrane and a 10‐μL sample was injected into the RP‐HPLC column (Zorbax Eclipse XDB C18, 150 × 4.6 mm) by an autosampler (Hitachi AS‐4000, Merck KGaA). The composition of the mobile phase changed from 80:20 (acetone/water) to 100:0 within 7 min. In the subsequent half minute the original ratio was restored. The working temperature is 30°C. Peaks were detected by a VIS‐detector (Hitachi L‐4250, Merck KGaA) at a wavelength of 450 nm.

2.5. Validation of the biological calorimetric measurements

A linear relationship (oxycaloric equivalent) exists between the heat production rate ($\dot{Q}$) and the OUR 13, 14, 44 (Eq. (2)).

$$\dot{Q}=-460kJmo{l}^{-1}·OUR$$ (2)

Equation (2) is valid for growth on carbohydrates if partial fermentative metabolisms can be excluded and the combustion enthalpy of biomass and products fulfill the Thornton‐rule 14, 45. Barley grains contain mainly starch and just a little crude protein and lipids 46. Thus, the estimation of the heat production rate during growth on barley using the respirometrically determined OUR (Eq. (2)) was used to validate the biological calorimetric measurements. The oxycaloric equivalent can also be used to give a rough estimate of the time of aerobic growth (t) using the available oxygen in the calorimetric ampoule (n_O2) and the measured average heat production rate (${\dot{Q}}_{AV}$) (Eq. (3)).

$$t=\frac{-460kJmo{l}^{-1}·n_{O2}}{{\dot{Q}}_{AV}}$$ (3)

n_O2 ((2.39 ± 0.02)·10⁻⁴ mol—SD) can be calculated from the oxygen content in air (ξ = 0.2095) 7, the air filled volume of the headspace (v = (28.2 ± 0.2)·10⁻⁶ m³—SD), the atmospheric pressure (p = 101,325 Pa), the universal gas constant (R = 8.31441 J mol⁻¹ K⁻¹), and the temperature (T = 301.15 K) via (Eq. (4)).

$$n_{O2}=\frac{\xi pv}{RT}$$ (4)

3. Results

3.1. Instrumental reproducibility

Most of the commercially available isothermal microcalorimeters have a working volume of a few milliliters, which is too small for a representative analysis of SSF processes. Special isothermal multichannel calorimeters originally designed to analyze the binding behavior of cement have the proper working volume, but the signal noise is often one magnitude worse than that of conventional microcalorimeters. However, yeast cultures, growing with the maximum specific growth rate, produce approximately 250 W kg_CDW ⁻¹ 47. Taking technically relevant substrate loads of SSF processes (>10 mg_CDW g_DW ⁻¹), heat production rates in the range of several W kg_DW ⁻¹ are expected, which are comfortable to measure even with this type of instrument.

For the further development, the instrumental reproducibility, accuracy, and stability are important. The baseline drift caused by diurnal temperature changes was determined to be <0.001 mW h⁻¹ (results not shown). The mass‐related signal drift was <0.003 (W kg_DW ⁻¹ h⁻¹) assuming a typical sample size of 1 g or 0.3 g_DW. A typical long‐term signal noise of <0.02 mW or < 0.07 W kg_DW ⁻¹ for inert samples was observed (see Supporting Information). The comparison of the measured heat trace of the methyl paraben hydrolysis (reference reaction) with the expected heat calculated from a thermokinetic model of that reaction demonstrates the accuracy of the calorimetric measurements (for details see Supporting Information). For the quantitative assessment of the accuracy, a constant heat signal is required. The constant heat signal was provided by an electrical heater located at the same position in the ampoule as the later samples. The measured current at the electrical heater (10.62 ± 0.08 mA, n = 8, SD) translates into an expected heat production rate of 1.63 mW (Eq. (1)). The calorimetric measurement delivers 1.55 ± 0.01 mW (n = 411, SD). Accordingly, the measurement error is approximately 0.08 mW (0.27 W kg_DW ⁻¹) or 4.9%. This deviation is within the error range from the combined error of the current and resistance measurement (0.05 mW, propagation of error in Eq. (1)).

3.2. Biological reproducibility

In order to evaluate the biological reproducibility, six identically treated samples for biomass determination and 12 samples for calorimetric monitoring were prepared. Inoculated material was incubated for 72 h prior to transfer into the calorimeter ampoules to monitor highly active biomass after the lag phase. Figure 2 compares metabolic heat (A) with fungal biomass growth in a time span from 72 to 114 h (B). Generally, attention has to be paid to the thermal equilibration for the first hour of incubation in the calorimeter. This period cannot be used for the interpretation of the biological data. From the second hour onwards, the calorimetrically measured heat evolutions were 358 ± 60 kJ kg_DW ⁻¹ (72–95.2 h, SD) and 638 ± 88 kJ kg_DW ⁻¹ (72–113.4 h, SD). The values correspond to a mean heat production rate of 4.4 ± 0.6 W kg_DW ⁻¹ (see the discussion section, SD). The amounts of formed biomass were 29 ± 3 mg_CDW g_DW ⁻¹ and 56 ± 13 mg_CDW g_DW ⁻¹ for 72 h and 114 h old cultures, respectively (Fig. 2B, SD). For the additional validation of the calorimetric measurements, equivalent respirometric measurements were performed with the result of OUR = (1.05 ± 0.02)·10⁻⁵ mol kg_DW ⁻¹ s⁻¹ (n = 6, SD). This corresponds to –4.8 ± 0.1 W kg_DW ⁻¹ (SD) using the oxycaloric equivalent (Eq. (2)).

Figure 2.

Comparison of calorimetric growth signal with increase in fungal biomass. (A) Calorimetry: the measurements occurred with 72 h old cultures. Six of 12 samples are shown exemplarily (the samples with the lowest and highest evolved heat are included). (B) Parallel growth experiment in falcon tubes: the increase in fungal biomass (CDW—cell dry weight) is shown after 72 h (white) and 120 h (gray) (n = 4; error bars are standard deviation).

3.3. Potential of calorimetry for SSF process optimization by single additives

It has been shown that the presented calorimetric method provides results with good reproducibility and accuracy and that the measured heat production is related to fungal growth. For that reason, it can be hypothesized, that the calorimetric method can be used to screen the influence of different growth conditions (e.g. additives, nutrients, water activity, gas composition, pressure etc.) on fungal growth. Because product formations are often related to the growth rate or to biomass concentration, it can be further hypothesized, that the method is even suited to optimize SSF product formations. To test these theses, the effect of different concentrations of sun flower oil (SFO), ranging from 0 to 5 %, on heat evolution as well as on biomass growth and product formation were monitored over a time period of 10 days (Fig. 3) starting with freshly inoculated barley. SFO is suited as test additive, because it has already been shown to improve the fungal growth and β‐carotene production in liquid media 48.

Figure 3.

Pattern of growth and product formation of B. trispora on barley in presence of different amounts of sun flower oil (SFO). (A) Shows the typical heat evolution and (B) illustrates the growth rate and β‐carotene production rate in dependency on the SFO concentration (n = 5; error bars are SD). The experiment started with freshly inoculated barley.

Through the addition of 5% SFO, the biomass growth rate increases from 7.6 ± 0.5 mg_CDW g_DW ⁻¹ d⁻¹ (without SFO, SD) to 11.5 ± 0.6 mg_CDW g_DW ⁻¹ d⁻¹ (5% SFO, SD). The formation rates of β‐carotene increased from 1.9 ± 0.3 μg g_DW ⁻¹ d⁻¹ (without SFO, SD) to 4.3±0.4 μg g_dm ⁻¹d⁻¹ (5 % SFO, SD). The calorimetric data does not only support the application of the linear growth rate model, but also indicate the positive effect of SFO.

Further, the oxygen decrease and the carbon dioxide increase were analyzed for cultivation with 5% SFO addition. A respiration quotient (RQ) of 0.88 ± 0.05 (SD) was calculated over a period of 10 days (data not shown). The RQ accounts for the ratio of carbon dioxide evolution divided by oxygen consumption. RQ values below 1.0 are typical for the utilization of fat (from SFO).

3.4. Further additives

Concluding from the experiments with SFO, it became obvious that multichannel calorimetry is principally suited to analyze the influence of growth promoting additives on fungal growth. To generalize this screening capability of the calorimetric method, it would be necessary to consider a range of different additives. Several substances with a proclaimed positive effect on product formation were tested (Fig. 4). Based on the calorimetric data, a positive effect of SFO and a negative effect of β‐ionone were observed. The effect of β‐ionone results from a prolonged lag phase. All other additives did not significantly influence the metabolic heat evolution. A similar picture emerges from the much elaborated biomass growth analysis with glucosamine as cell indicator. The growth rate was significantly increased by addition of 5% SFO (11.5 mg_CDW g_DW ⁻¹ d⁻¹) and a little with thiamin (9.0 mg_CDW g_DW ⁻¹ d⁻¹). Likewise, the production rates of β‐carotene were strongly improved by SFO addition (4.3 μg g_DW ⁻¹ d⁻¹) and a little by thiamin addition (2.4 μg g_DW ⁻¹ d⁻¹). The reference values were 7.6 ± 0.5 mg_CDW g_DW ⁻¹ d⁻¹ and 1.9 ± 0.3 μg g_DW ⁻¹ d⁻¹ (SD) for biomass and β‐carotene, respectively. Imidazole, β‐ionone, and malate had no positive effect on growth and product formation, in contradiction to the literature. These additives even reduced both biomass and product formation. Imidazole is known to quench the last metabolic steps of the β‐carotene biosynthesis, so that lycopene is accumulated. Indeed, peaks for lycopene were seen in the HPLC chromatogram but they were below the detection limit of 0.15 μg g_DW ⁻¹.

Figure 4.

Influence of different additives on growth and product formation rate of B. trispora. (A) Shows the typical heat evolution and (B) illustrates the biomass formation and β‐carotene production rate (n = 5; error bars are SD).

4. Discussion

SSF is a fast‐developing technology to produce a wide range of products. Nevertheless, heat accumulation poses one of the major challenges. Therefore, the presented work summarizes the results of a simple method for the direct measurement of heat produced during SSF at the particular example of B. trispora growing on barley. Additionally, the publication evaluates effects of potent supplementations to a basic substrate.

First of all, a simple calorimetric technique for measurement of heat production rates during SFF processes was developed. The parameters of our simple, robust and cost‐effective multichannel example device were related to the sample size (signal drift: 0.003 W kg_DW ⁻¹ h⁻¹, signal noise (peak to peak): <0.07 W kg_DW ⁻¹). With the help of chemical reference experiments and via electrical heater experiments, we were able to demonstrate that the method's accuracy is in the relevant measurement range. The lower measuring limit (0.07 W kg_DW ⁻¹) of the method results from an assumed signal‐to‐noise ratio of 5 and the SD of the observed signal noise (10 h measurement). Taking the heat production rate of B. trispora growing on barley (4.47 ± 0.73 W kg_DW ⁻¹, see below) as a typical value, the accuracy as well as the signal to noise ratio is in an acceptable range.

Two potential reasons for the observed variance of the calorimetric measurement of SSF samples can be assumed. First, the biological heterogeneity can be one reason. This thesis is supported by a similar relative standard deviation (RSD) for the calorimetric measurement (16.7% (measuring period 72–95.2 h); 13.8% (measuring period 72–113.4 h)) and for the determination of the increase of fungal biomass (10.3% (measuring period 73–95.2 h); 23.2% (measuring period 73–113.4 h)). Second, a water loss of 0.023 kg kg_DW ⁻¹ d⁻¹ is expected if the variance of the calorimetric measurements is attributed to evaporation (see Supporting Information for details). This value corresponds to the observed water loss. The evaporation is facilitated by the more as one magnitude higher air volume in the ampoule in comparison to the sample volume. The large air/sample volume ratio was required to ensure sufficient oxygen supply.

The biomass related heat yield of 12.2 kJ g_CDW ⁻¹ (measuring period 72–95.2 h) or 11.3 kJ g_CDW ⁻¹ (measuring period 72–113.4 h)) can be estimated by dividing the produced heat by the increase in fungal biomass. These values are in the range of reported heat yields of different yeast strains growing aerobically on different mono‐ or disaccharides (11.7–13.6) kJ g_CDW ⁻¹ 49.

The observed approximate linear increases of heats versus time indicate a constant heat production rate. Due to the linearity, the mean heat production rate of 4.47 ± 0.73 W kg_DW ⁻¹ (measuring period 72–95.2 h, SD) or 4.40 ± 0.60 W kg_DW ⁻¹ (measuring period 72–113.4 h, SD) can be calculated from the data (see Supporting Information: Eq. (2)). Data from Oriol et al. gave about 10 W kg_DW ⁻¹ for the initial growth phase of Aspergillus niger on cut bagasse with glucose 50. In comparison, the heat evolved by B. trispora is in the same order of magnitude. The 50% difference might be due to the easily available substrate source for the A. niger, whereas in this study starch from barley has been the main nutrient source of the complex medium. In general, differences in culture conditions may also fully account for this deviation.

The respirometrically determined OUR ((1.05 ± 0.02)·10⁻⁵ mol kg_DW ⁻¹ s⁻¹ (n = 6, SD)) can be utilized for a rough estimation of a corresponding heat production rate of 4.8 ± 0.1 W kg_DW ⁻¹ (Eq. (2), SD). This value is in the range of the calorimetrically measured heat production rate of 4.4 ± 0.6 W kg_DW ⁻¹ (SD). The apparently sixfold higher fluctuations of the measurement values in calorimetry are probably caused by the 20‐fold lower sample size, in comparison to respirometry.

From the directly measured heat production rate, a temperature rise of 1.4 K h⁻¹ for barley with 70% water can be estimated under adiabatic conditions (see Supporting Information for details). Under isothermal, adiabatic conditions the metabolic heat will be used to evaporate water with a rate of 0.045 kg kg_DW ⁻¹ d⁻¹ (see Supporting Information for details). Admittedly, the estimates reflect physical limiting cases. In reality, evaporation, heat accumulation and heat dissipation to the environment will occur simultaneously and the real temperature rise will be less dramatic. However, the estimations illustrate the importance of exact data to estimate the self‐heating in SSF processes and may be incorporated in existing modeling.

The positive effect of SFO addition to the substrate on growth and product formation has also been detected by the heat production, as both parameters directly correlate. Our results further support the thesis that SFO can be used by B. trispora as an easily digestible C‐source. This assumption is strongly supported by the measured RQ in a medium with 5% SFO. RQ values below 1.0 are typical for the utilization of fat (mainly from SFO) or protein as carbon source 51.

Different additives retrieved from literature were analyzed. Next to SFO, the supplementation with the vitamin thiamine could enhance both, biomass and product formation. Although stated otherwise, malate and β‐ionone had no positive effect and imidazole even reduced the vitality of the fungi. Imidazol quenches the last two metabolic steps from lycopene to β‐carotene. As there was still a production rate of 0.8 μg g_DW ⁻¹ d⁻¹ β‐carotene with addition of imidazole, it is suspected that the homogenous distribution of additives is a limiting factor.

The heat evolution data for the additives show a correlation toward growth, implying that calorimetry might be a fast and accurate method to distinguish positive and negative influences toward the growth and production behavior of microorganisms on solid substrate.

5. Concluding remarks

A method for quantifying the metabolic heat during SSF processes was developed and successfully tested. A simple (and cost‐efficient) multichannel calorimeter was found to be sensitive enough to quantify such processes. Further advantages of the method are the real‐time monitoring of metabolic activity, the noninvasive measurement and the much lower workload in comparison to conventional methods (e.g. glucosamine biomass estimate). In addition, calorimetry is developing into a high‐throughput method due to the invention of multichannel instruments. Today, up to 48 channels can be measured simultaneously in available instruments (for instance see: http://www.tainstruments.com; http://www.symcel.se). The development of inexpensive high‐throughput devices with a sample volume of >20 mL is conceivable and technically no longer a problem. The unique weakness of the method is the need to ventilate ampoules daily. Otherwise, oxygen depletion and metabolic waste accumulation in the form of carbon dioxide are the consequences 17, 52. This final problem can be overcome in future research through the development of an ampoule sealing that is permeable to oxygen and carbon dioxide and impermeable to water 53. In principle, the most competing technology (i.e. respirometry) also allows high throughput measurements, but the calculated heat production rate is potentially biased by deviations in the oxycaloric equivalent due to sample composition and potential anaerobic zones. However, the technically possible combination of respirometry and calorimetry has potential to reveal information about the peculiarities of the substrate and anaerobic zones that are important for designing SSF processes. Methods for the simultaneous measuring of CO₂, O₂, and heat on mL‐scale 53, at larger scales 16, 54, and theories for data interpretation 45 are already examined.

The proposed simple calorimetric method is potentially suited for the screening of optimal SSF conditions (additives, nutrients, bed size, pH, water activity, gas composition, and pressure) in a simple manner. This was demonstrated at the example of the influence of additives on the fungal growth and the production of β‐carotene with the filamentous growing fungus B. trispora. Still, the findings indicate a problem with homogeneous distribution, which is especially true for small concentrations. Future approaches should also focus on the mixing of solid substrates and their influence toward biomass due to elevated shear stress. Yet, the presented findings are of further importance for the design of next‐generation SSF processes.

Practical application

Metabolic heat evolved during solid‐state fermentation (SSF) may lead to a temperature rise and desiccation of the substrate, which is especially true for the technical scale due to an unfavorable ratio of the heat exchange area to reactor volume and poor mixing. The consequence is inactivation of the production strain and degradation of heat‐sensitive products. In case the exact metabolic heat production is known, appropriate measures can be adopted concerning the reactor design, e.g. additional heat exchange areas, and the operational parameters, e.g. water supplementation. Therefore, this study presents a simple and cost‐effective method to measure the metabolic heat under the particular conditions of SSF processes. Furthermore, the method is suitable for a time‐saving substrate and supplementation screening.

The authors have declared no conflicts of interest.

Nomenclature

I	[A]	current
nO2	[mol]	available oxygen in the calorimetric ampoule
P	[W]	joule heat
p	[Pa]	pressure
$\dot{Q}$	[W]	Heat production rate
R	[Ω]	resistance
T	[K]	temperature
t	[s]	time
v	[m3]	volume of the ampule head space

Greek symbols

ξ

[%]

oxygen content in air

Supporting information

Supporting Information