2.1. Current Status of Dissolved CO Measurement Methods in Syngas Fermentation

Measurement of dissolved CO (DCO) in syngas fermentation medium is a challenging task because of the low CO solubility, interference from other chemicals, and conditions of syngas fermentation process. CO is sparingly soluble in water with a Henry’s law constant about 121,561 kPaL/mol at 37 °C, which results in a saturated concentration approximate to 23.25 mg/L (ppm) for pure CO under one standard atmosphere pressure [1,25]. Due to other major components (H₂, N₂, and CO₂) in syngas, actual DCO concentration in fermentation medium is lower [6]. The fermentation medium contains many chemicals from syngas, nutrients for microorganisms, and fermentation products that may interfere with the DCO measurement [6]. H₂ is the most significant interfering component as many CO detection mechanisms, such as electrochemical [26] or conductivity sensors [27], also respond to H₂. Other chemicals, such as CH₄, NH₃, H₂S, and HCl from syngas and alcohols and organic acids, may also affect the accuracy of certain CO detection mechanism. Lastly, syngas fermentation operates at specific conditions, which limit applications of some CO detection mechanisms. The microorganisms exploited in the fermentation are strictly anaerobic, so methods that require O₂ during detection may be problematic [3]. Temperature and pH are also important for optimal cell growth of the microorganisms, which further limits the selection of CO detection mechanisms [1].

Currently, limited methods were introduced for DCO measurement in syngas fermentation. The CO-myoglobin assay method was reported to analyze gas–liquid mass transfer coefficients in syngas fermentation [28,29]. In our research, offline gas chromatography and fermenter gas mass transfer models were used to estimate the DCO concentration [30].

The CO-myoglobin assay method exploits metalloproteins, proteins with an iron ion cofactor such as hemoglobin and myoglobin, to detect dissolved phase CO by observing the changes in optical absorption spectra between CO-free and CO-bound metalloproteins [12]. DCO concentrations are obtained through predetermined fitting models between known DCO concentrations and optical absorption spectra of CO-bound metalloproteins [10,31]. However, the CO-myoglobin assay method requires complicated operation procedures and has a slow response (more than 30 min) [10]. In addition, the metalloproteins have limited lifespan, which implies that this method may not be appropriate for repeated, long-term dissolved CO measurements.

DCO concentration (mol/L) CCO,L in the bulk of liquid medium can be estimated through the liquid film mass transfer model with the help of gas chromatography [30]:

where CCO* is the DCO concentration (mol/L) in the interface surface in equilibrium, which can be calculated from Henry’s law based on the headspace partial pressure of CO. VL is the volume of fermentation medium and a is the area (m²) of the gas–liquid interface surface. kL,CO is the liquid film mass transfer coefficient (L/m^2·h), which is estimated beforehand. The molar rate of transfer (mol/h) −dnCOdt represents the consumption of CO during the fermentation, which can be obtained by measuring CO partial pressure in the inlet and outlet gas flow with gas chromatography. DCO concentration can be obtained by solving the Equation (4) with the partial pressure data from gas chromatography. However, the accuracy of this method is highly related to the correctness of the mass transfer model and the mass transfer coefficient kL,CO. Meanwhile the response of this method is slow due to the process to measure CO partial pressure from inlet, outlet, and headspace of the fermenter with gas chromatography.

Other DCO measurement methods were reported for medical or health applications, such as fluorescent optical sensors for CO imaging in tissues [32] and indirect methods for blood CO concentration measurement [33,34].

Fluorescent optical sensors fabricated with fluorescent proteins [13,32] and organic CO probes based on palladium catalyzed Tsuji–Trost reaction [35] were reported as a novel solution for in vivo CO imaging in animal tissues. The photoluminescence response triggered by reactions between CO and these fluorescent probes provides robust resistance to interference from other chemicals. These sensors were also reported with strong fluorescent response to dissolved phase CO in water-based solutions [32,36,37], which suggests that fluorescent optical sensors can have potential applications in DCO measurements for syngas fermentation. Simplicity and fast-response are the most appealing properties of fluorescent optical sensors [13,38]. However, fluorescent optical sensors may not be applicable for automatic, repeated DCO measurement, because their sensing reactions are conditional reversible with the aid of special reagents [13,38] or completely irreversible [36,37].

Determination of CO in blood can be indirectly measured with gas chromatography using a chemical CO extraction reagent. Reagents like formic acid [34] and ferricyanide [33] were reported to break down CO-bound metalloproteins to release gas phase CO from a fixed amount of blood samples. This approach circumvents the challenge of measuring DCO, but it is only applicable for blood CO measurement when metalloproteins exist. However, it is possible to use physical CO extraction methods for indirect DCO measurement in the syngas fermentation process.

Physical gas extraction methods are designed to extract dissolved gas by decreasing the partial pressure of the gas, such as the gas stripping technique [39], static headspace equilibration method [40], and vacuum extraction system [41,42]. Applications of semipermeable membranes in these methods were reported to achieve automatic, high-volume, rapid dissolved gas extractions [43,44,45]. In our opinion, membrane-aided vacuum extraction systems are the most practical DCO measurements methods with several automatic, in-house systems reported [43,44,46]. Dedicated dissolved gases measurement instruments, i.e., membrane introduction mass spectroscopy, were reported for applications in environmental science [47,48,49]. The successful applications of these gas extraction methods for gases with the same sparingly solubility, such as O₂ and N₂ [50], and noble gases, such as Ar, He, Ne, and Kr [43], suggest their potential applications in the DCO measurement. However, the sampling time of these gas extraction systems is around several minutes to hours, due to the time to establish new equilibrium in the gas extraction process, which are negatively related to the solubility of the target gas [45].

Despite the long sampling time and the complicated gas extraction system design, the introduction of physical gas extraction systems enables common gas phase CO sensor to indirectly measure DCO concentration. Current direct DCO measurement methods still have many obstacles to overcome for automatic, repeated DCO measurement, and the indirect DCO measurement methods helps to fulfill the current, urgent need from syngas fermentation while direct DCO measurement methods are still under improvement. A detailed review of gas phase CO detection mechanisms is performed to analyze their potential application in DCO measurement during syngas fermentation process, both directly and indirectly.