Electrochemical Impedance Spectroscopy for Microbiological Pro-Cesses: on the Way to a Monitoring Tool for the Determination of Biomass

Novel approaches during industrial fermentation processes are not only restricted to classical pH, dO 2 and off gas analysis. There are many in-situ and online sensors based on different physical principles to determine biomass, product quality, metabolism and cell death. These process analytics are normally very cost intensive. One of the very important approaches relating to in-situ determination is the viable cell concentration (VCC). Based on this knowledge, an increased efficiency in monitoring and controlling strategies during cultivations can be provided. Electrochemical impedance spectroscopy (EIS) is a powerful tool in life science and has been heavily investigated for its potential to characterize the microbiological behavior during cultivation processes. For that purpose, EIS is used to monitor biomass in fermentation processes of Escherichia coli and Saccharomyces cerevisiae. A correlation of double-layer capacitance (CDL) and the cell density was found. The novelty of this approach has been proven with different state-of-the-art biomass measurements (dry cell weight (DCW) and optical density (OD)) in order to get an adequate verification of this method. Finally, it has been demonstrated that EIS measurement at low frequencies is a powerful monitoring tool in different modes (offline, online an inline) for microbiological cultivations. In addition to that, this approach is also perfectly suited to determining physio-logical states of the cells. for Microbiological Pro-Cesses: on the Way to a Monitoring Tool for the Determination of Biomass.


Introduction
For producing organic compounds such as food, drugs and bulk chemicals as well as in waste-to-value concepts, microbiological cultivations play a key role [1]. As probably supposed, microorganism, like bacteria and yeast are not only applied in pharmaceutical technology. For guaranteeing product quality and safety, process monitoring such as pH, dis-solved oxygen (dO 2 ), off gas analysis and biomass is state-of-the-art in today's industrial cultivation processes. However, in such applications mentioned above, process monitoring, i.e., process analytical technology (PAT) is of crucial importance to control the system and react timely upon metabolic changes. The biomass as an important measured value in bioprocesses can only be determined by using offline methods or high-end soft-sensor applications [2]. For that usage, these sensor systems are often performed as inline/online/at-line detection systems. The usual analyzing methods for these applications, such as for example high-performance liquid chromatography (HPLC) for metabolites, off-gas balance, and/or dielectric spectroscopy measurements, have proven to be helpful. For the determination of biomass, however, an accurate and reliable measurement systems [3,4] is still required. By determining viable cell concentrations (VCCs), the process parameters subsequently lead to more robust and reliable bioprocesses and contribute to a higher product quality. Moreover, this would enhance more flexibility in batchto-batch variations. The determination of VCC is performed as an offline measurement using flow cytometry or confocal microscopy [5,6]. Such methods are based on marker proteins and fluorescence probes, which allow screening between living/dead cells [5,6]. Besides the staining with propidium iodide, other stains can also mark the DNA of the cells, which also allows for these screenings [7]. Such advanced analytical tools are mainly used in the pharmaceutical industries due to their cost-intensiveness.
Therefore, ordinary bulk foods, which use yeast in their production process, such as the brewing and baking industry, manufacture their products in rather uncontrolled process conditions. Regarding this, it should be mentioned that the conditions for propagation and fermentation are crucial for the quality of the product itself. Indeed, an implementation of online measurements for determining the vitality in the brewing industry was historically founded by affordable, simple, robust and reproducible tests [8]. This is mainly due to the fact that the online and inline biomass measurement approaches are rather scarce, because the principles are based on physical measurements. Beside other physically based principles, such as optical measurements, many biosensors also use a change of an electrical signal for analysis. It is common to apply high frequency alternating current (AC) impedance spectroscopy with high field amplitudes, which is referred to as β-dispersion [9,10]. The interesting aspect here is, that cells with an integer cell membrane affect the relative permittivity between two electrodes in contrast to dead cells. Therefore, this signal is used for the estimation of VCCs. The details about the measurement principles can be found in [11][12][13][14]. The model organism for the application of AC measurements in the ß-dispersion range is yeast, it being a very important expression host for recombinant proteins [15][16][17] and of high interest because of its meaningfulness. In addition to that, there are approaches towards more complex expression systems, such as filamentous fungi and Chinese hamster ovary (CHO) cells, which have already been performed [18][19][20][21]. It is evident, that these measurements show a strong dependence upon the physical process parameters (such as aeration and stirring-causing gas bubbles, temperature shifts and pH gradients), and are furthermore, highly affected by changes in the media's composition dur-ing cultivation.
In addition, changes in the media during fermentation show effects on the signal amplitude and requires complex data analysis [22].
At this point it is necessary to note, that the ß-relaxation in the frequency range from 107 to 104 Hz is not the only one characterized by the relaxation phenomenon, which can be exploited for the determination of biomass. Especially, changes of the electrical double layer by the adsorption/desorption of cells at the electrode surface, are detectable at low frequencies in the millihertz range. The so-called α-dispersion, which is characterized from 104 to 10-2 Hz, can also provide valuable information. While during ß-dispersion effects, the en-tire cell is embedded within Wagner-Maxwell polarization, α-dispersion on the other side, shows preliminary ionic interactions and relaxation phenomena on the cell membrane itself [23]. Here one must distinguish between two effects, firstly, the cell itself, exhibiting differences in cell walls, the membrane compositions, size as well as shape and metabolism.
Secondly, physical parameters especially in the media (pH and ion concentrations) can play an influence on the potential distribution at the electrode double layer [24,25]. It has been proven, that EIS measurement in the α-dispersion regime provides valuable results for detection of bacteria in soil, food and feces-polluted water by using interdigitated electrodes [26][27][28][29][30][31][32][33][34]. First approaches towards process monitoring were shown by Kim et al. [35]. He Several studies have been carried out in our group on E. coli and S. cerevisiae and promising results for VCC determination throughout the whole cultivation range were achieved. We cultivated in batch phase as well as in the fed-batch, which led to high cell densities [36,37] and has been proven and tested in the brewery environment [38]. It is planned to carry out further investigations to correlate the total biomass to the extracted CDL and show the straight-forward application and the possibilities it holds. Especially a system for an ac-curate VCC measurement, such as vitality and cell number (living/dead) determination even in industrial-scale applications will be developed, which could bring immense facilitation and increase in the time needed to evaluate such processes.

Experimental
The ability for EIS measurement has already been successfully demonstrated in several microbiological processes. The monitoring of biomass for E. coli and S. cerevisiae requires a different procedure for cultivation and analytics, which shall be closer considered in the following [36][37][38]. After addition of DiBAC4 (bis-(1,3-dibutylbarbituricacid) trimethineoxonol) and Rh 414 dye diluted cultivation broth was measured using a Cy Flow Cube 8 flow cytometer (Sysmex-Partec, Bornbach, Germany). Rh 414 binds to the plasma membrane and visualizes all cells, while Di BAC is sensitive to plasma membrane potential and therefore distinction between viable and non-viable cells can be achieved. Detailed information on the viability assay can be found elsewhere [40]. Overall errors with this method were in the range of 0.5% to 1%. As less than 5% of dead cells were detected in all samples, DCW and VCC can be assumed to be equivalent.

Expression Host and Cultivation for S. cerevisiae
All cultivations were performed using the S. cerevisiae strain,  Detailed information on the viability assay can be found elsewhere [40]. The overall errors for this method were in the range of 0.5% to 1%. Sugar concentrations in the fermentation broth were determined using a Supelco C-610H HPLC column (Supelco, Bellefonte, PA, USA) on an Ultimate 300 HPLC system (Thermo

Measurement Setup
Determination of VCC by using capacitance probes are widely spread for industrial applications. Regarding this, the physical principle of these systems based on β-dispersion (107-104 Hz).
Unfortunately, these probes show a high dependence on process parameters (e.g., stir-ring, temperature, pH, salt and substrate concentration, etc.) [32,42]. The goal was to devel-op and test the novel application for low frequency electrochemical impedance spectroscopy in online and inline measurements. In this regard we exploited a different physical phenomenon (α-dispersion) at frequencies below 10 kHz, which provides valuable information regarding the biomass concentration. By this physical principle deformation of ionic species around the cell membranes are detected. The dielectric response of the ap-plied EIS signal was proportional to the ionic charge gathered around the membrane of ad-sorbed cells on the electrode [43,44]. For the characterization of E. coli and S. cerevisiae [36,37] in lab-scale EIS measurements were applied in the range of 106 to 10-1 Hz with amplitudes of 100 to 250 mV using the Alpha-Ahigh-resolution dielectric analyzer (Novo control, Montabaur, Germany). The EIS measurement was also performed for an industrial propagation process took place in the Stiegl Brewery in Salzburg/Austria [38]. Therefore, EIS measurements were recorded in the range of 106 to 10-1 Hz with amplitudes up to 100 mV using the N4L PSM 1735 frequency analyzer (Newton 4 th Ltd., Leicester, UK). To be able to generate a viable and also straight-forward monitoring process tool that can be applied to indus-trial scale setups and could be also moved to different operating sites, a range of different properties and features that the setup and the probe required had to be carefully considered.
We were aiming for a new way of monitoring processes, developing a sophisticated probe, while at the same time keeping the ease of handling and the accuracy of the obtained data at the forefront of our mind. Through careful planning a stepwise approach was devel-oped to reach this goal. These steps involved the construction of an online and inline proto type. The testing of thereof at reallife conditions and the generation of an impedance signal and its link to a viable cell concentration, and finally the implementation of the probe in an online mode to prove the applicability of this novel biomass sensor [36][37][38]. Details on the fitting procedure and data evaluation are given in [37].   [38]; (b) impedance measurement principle including online measurement probed developed [36,37]. (c) Sketch of the inline probe prototype indicating the used material and wiring. Connection to the impedance analyzer was performed using four-point BNC (Bayonet Neill-Concelman) connector [37].

Online EIS Measurements
Various experiments and measurements were undertaken to evaluate the scope of the developed online probe. Before standard tests could be developed and trialed, the Nyquisit plot was fitted for various concentration of the bio media by complex nonlinear square fitting (CNLS). Furthermore, the effect of the media background was investigated and was dis-missed as not having an impact on the impedance signal due to the flat slope (k= 6.4x10-8) of the curve from the plotted clarified fermentation supernatant against the biomass concentration. It also proved beneficial to utilize higher amplitudes of 500 mV for the flow-through measurements in order to avoid unwanted changes of the double layer resistance resulting in negative resistances [37]. Due to the negligible effects of the media on the measurements also the double layer capacitance could be determined (Figure 2). The adaptability of this online probe can be further showcased by the promising results of both measurements of E. coli and S. cerevisiae [37,38]. Measurements with E. coli showed high reproducibility of the obtained results, as the correlation between DCW and CPEdl-Q (see Figure 3) of two separate cultivations were nearly ident (similar slopes and high R2 = 0.94 and 0.98).
Reproducibility remained high even when two cultivations with different specific growth rates were investigated, highlighting the inherent strength of ESI for process monitoring. Similar promising results were obtained from the S. cerevisiae cultivation.
As mentioned above they were again fitted with CNLS (see Figure   4). The next step was to test the probe further and move the online probe from the lab setting to large scale atline measurements in the Stiegl Brewery in Salzburg/Austria [38]. The industrial propagation was undertaken in a 1500 L reactor. During propagation the reactor tem-perature rose from 8-12 °C. Hence, the measurement system was cooled to 7.6-8 °C in order to maintain a steady temperature during EIS. The impedance measurement method for online measurements in the a-dispersion regime was successfully transferred to a large-scale industrial application. The results of both, lab scale and industrial scale results, can be seen below (see Figure 5). Furthermore, a correlation was obtained between the cells count-ed at the brewery and the impedance signal (see Figure   5b) was established and can be used for the direct measurement of the cells from now onwards.    Figure 6: (a) Impedance raw data in the Nyquist plot for an aerobic cultivation. Black squares represent the signal from the online probe-enlarged in the inlay-and red triangles, the inline probe at similar time stages. Capaci-tance of the inline probe is one order of magnitude lower (as a result of smaller electrode areas). (b) The de-pendence of the impedance signal (not smoothed), glucose consumption and ethanol production in an aerobic cultivation using the inline probe.

Inline EIS Measurements
Besides versatile online impedance measurements, also an inline probe was developed [36]. This probe could be, among other things, used to estimate the viable cell concentrations in aerobic and anaerobic cultivations. In order to be able to compare inline and online measurements, data was recorded alternating both probes in aerobic and anaerobic cultivations. The data of the aerobic cultivation with the inline probe shows a reduction of the capacity by an order of a magnitude compared to the online measurements.
Further relevant details for the measurements with the inline probe can be seen in Figure 6. These results were then used to fit the biomass data. Both, the aerobic and anaerobic cultivations, show a linear fit between impedance signals and offline DCW beyond a threshold of 1g/ 1 DCW, as shown in Figure 7c and 7d. Furthermore, the correlation of the calculated biomass compared to the offline biomass was established. The expected exponential growth of the cultures were accurately monitored. In summary, for a defined minimal media the inline probe showed reproducibility and stable results (max. volume = 10-20 L) [36]. As defined media are however somewhat considered as a model media, a complex malt extract medium was subsequently investigated. Online measurements in such media can be cumbersome or simply not possible, as in the case of OD measurements. Hence, EIS could represent a valuable alternative. In this case again both inline (see Figure 7b) as well as online measurements (see Figure 7a) were performed to compare the obtained data. For the inline signal a distinctive decrease for the capacitance with subsequent increase of ethanol growth can be seen, while the drop in the online measurements is less pronounced. The interpolated data of both measurements was plotted against DCW. While both signals appear to be simi-lar in appearance compared to the defined media beforehand, a shift in the signal intensity can be observed.
In summary it can be said that the inline measurements setup as well as the online probe definitely have their merits, especially when it comes to the monitoring of complex media such as malt extract or molasses where other optical online methods reach their limitations. Figure 7: (a) Impedance signal over cultivation time for the online probe using only malt extract as growth media. The line (orange) shows the interpolation procedure. As for related aerobic cultivation, a drop in the impedance is observed after consumption of sugars. (b) Impedance signal raw data, interpolated and smoothed for the inline probe. (c) Normalized impedance signal vs. DCW for the online probe using malt extract compared to defined media. (d) Normalized impedance signal vs. DCW for the inline probe using malt extract and defined media with glucose.

Conclusion
Within this studies new online and inline probes based on EIS at low frequencies for the de-termination of VCCs for S. cerevisiae were tested. At the beginning, cultivations were monitored by using a prototypical developed online probe for pharmaceutical E. coli fed-batch cultivations. Fortunately, this batch cultivations on defined media for aerobic and anaerobic growth showed stable results, regardless of the carbon source or concentrations. A noval inline probe was designed and tested in aerobic and anaerobic cultivations. Therefore, the inline probe has been applied in a defined media and was compared to the online probe. During the process a good description of the biomass growth was achieved.
Apart from the determination of the biomass during the cultivation, physiological states could be deter-mined, depending on the respiratory condition of the cells. This measurement setup which has been developed for biomass is highly beneficial, especially in complex media such as malt extract or molasses. For instance, optical online methods would not be suitable in such optically dense media. Therefore, the developed system shows high potential for monitoring cell growth. Using such online or inline probes allow real-time determination of the biomass. In further studies a novel DOI: 10.26717/BJSTR.2021. 35.005682 27556 EIS analyzer at low frequency for low-cost application, such as food sector, will be developed. In addition to that new applications for VCC measurement for monitoring algae are being evaluated.