Microbial CO₂ Capture and Bioconversion from Energy Systems as a Tool Supporting Decarbonization and Sustainable Development: A Comparative Review of Microalgae, Cyanobacteria, and Bacteria

Abstract

Microbial carbon dioxide (CO₂) utilization can supply biomass and chemicals while reducing fossil feedstock demand. This critical review compares microalgae, cyanobacteria, acetogens, hydrogen-oxidizing bacteria, and microbial electrosynthesis for integration with energy systems. Evidence is organized by measurement endpoint, reactor scale, and operating duration within a common process boundary. Photosynthetic cultivation favors sites with natural light, land, and local nutrients; annual productivity, harvesting, and contaminant transfer constrain implementation. Engineered cyanobacteria offer target metabolites when culture stability and product recovery are demonstrated. Hydrogen-oxidizing bacteria favor protein or polymer production with low-emission H₂, efficient gas delivery, and explosion protection. Acetogens suit anaerobic chemical production where reductant supply and product inhibition can be managed. Recent electrosynthesis results support reactor and recovery integration, while electricity demand, selectivity, and component durability determine whole-process feasibility. Laboratory experiments, pilot demonstrations, and sustained operation provide different levels of validation. Practical selection depends on complete energy inputs, product quality, and a defined market or treatment function. The climate benefit also depends on CO₂ origin, the displaced product, and subsequent carbon release; biomass formation alone does not establish durable removal.

1. Introduction

Rising temperatures and persistently high emissions justify the development of technologies that limit CO₂ release to the atmosphere and reduce the use of fossil carbon as a feedstock. According to the World Meteorological Organization (WMO), the global mean surface temperature in 2025 was approximately 1.43 °C above the 1850–1900 average [1]. The Global Carbon Budget 2025, published in 2026, provisionally estimated fossil CO₂ emissions in 2025 at 38.1 Gt, approximately 1.0% higher than in 2024 [2]. In the European Union, the low-carbon transition is guided by targets to reduce net emissions by at least 55% by 2030 relative to 1990, by 90% by 2040 under the 2026 amendment to the European Climate Law, and to achieve climate neutrality by 2050 [3]. The Fit for 55 package links delivery of the 2030 target to changes in energy systems, industry, transport, and fuel policy [4]. For bioconversion, the relevant requirement is to demonstrate a reduction in net emissions for a defined energy input and product use.

Physicochemical capture approaches include liquid absorption, adsorption on porous materials, membrane separation, and cryogenic processes [5]. Separated CO₂ can subsequently be injected into suitable geological formations, mineralized, or converted into chemical products [6]. Carbon capture and storage (CCS) and carbon capture and utilization (CCU) nevertheless serve different purposes: the former provides long-term isolation [5], whereas the latter supplies a feedstock for a subsequent process [6]. Biological CO₂ fixation can complement these approaches through biomass and metabolite synthesis at relatively mild temperatures and pressures [7]. When the carbon source is biogenic or atmospheric CO₂ and energy is supplied by sunlight or low-emission sources, the input of additional fossil carbon can be reduced [8]. Closing the carbon loop nevertheless requires accounting for subsequent carbon release, gas losses, and the energy consumed by all operations [6,9].

Microalgae, cyanobacteria, and non-photosynthetic bacteria differ primarily in how they obtain energy and electrons. Microbial systems exhibit different performance in biomass and metabolite production as well as CO₂ fixation. For Nannochloropsis, areal biomass productivity reached 19–24 g total solids (TS)·m⁻²·d⁻¹ in vertical photobioreactors (PBRs) compared to 12–15 g TS·m⁻²·d⁻¹ in the horizontal systems evaluated. These values do not quantify the percentage capture efficiency for the CO₂ supplied [10]. In hydrogen-oxidizing bacteria (HOB), gas feeding allows substantially higher cell concentrations to be maintained without illumination. In Cupriavidus necator cultures, medium optimization increased the final biomass concentration from 15 to 53 g TS·L⁻¹, while gas supply control reduced gas consumption by 67% relative to the reference configuration. In the reference configuration without gas recirculation, H₂ utilization was approximately 5% [11]. Experiments using real gases confirm the importance of operational constraints. Feeding Chlorella vulgaris with flue gas from biogas combustion increased the growth-phase biomass production rate to 0.0778 ± 0.0031 g volatile solids (VS)·L⁻¹·d⁻¹ but also caused acidification and earlier culture decline; the CO₂ fraction in the gas decreased from 13 ± 1 to 1 ± 0.5% [12]. In a study using flue gas from a municipal heating plant, sulfur oxides (SO_x) caused culture collapse on day eight. After desulfurization, the production rate reached 0.2810 ± 0.0162 g VS·L⁻¹·d⁻¹, and the reported CO₂ utilization efficiency (CO₂UE) increased from 24.1 to 41.4% [13]. A different constraint was observed with Synechocystis sp. PCC 6803: with the pH maintained at approximately 7.5, growth on gas containing 30% CO₂ was 2.7-fold slower than at 4% CO₂ [14]. Carbon availability, contaminant tolerance, and photosynthetic regulation therefore require separate assessments; increasing the CO₂ concentration alone does not guarantee process intensification.

Recent publications address complementary aspects of microbial CO₂ utilization. Sadvakasova et al. [15] connected microalgal metabolic engineering and carbon-concentrating mechanisms with bioprocess design, consortia, and hybrid cultivation. Zhang et al. [16] reviewed Wood–Ljungdahl energetics, genetic tools, and engineering for diversified acetogenic products. Rovira-Alsina et al. [17] examined electricity-driven protein production by HOB, including recycled nitrogen, gas utilization, downstream processing, economics, and commercialization. Jourdin et al. [18] modeled electrochemical and electricity cost effects on electrosynthesis feasibility. Ali et al. [19] integrated algal wastewater treatment, CO₂ capture, cultivation design, biomass valorization, and deployment barriers. Onyeaka and Ekwebelem [20] focused on bacterial fixation pathways as well as genetic and metabolic engineering. De Souza Sevalho et al. [21] combined systematic scoping and bibliometrics with fixation pathways, production platforms, products, and industrial constraints. The present review complements these contributions with a cross-platform comparison specifically for energy system CO₂ streams. Its synthesis links distinct measurement endpoints and whole-process boundaries to reactor scale, operating duration, and deployment constraints. Product carbon origin is separated from CO₂ feed recovery; compact reactor performance is considered together with reductant production, gas delivery, and recovery. Laboratory maxima, pilot demonstrations, and sustained operation support different conclusions rather than a universal efficiency ranking.

This review aims to determine the conditions under which microbial CO₂ capture and bioconversion from energy systems can provide process, energy, or climate advantages. It compares the strengths and limitations of microalgae, cyanobacteria, acetogenic bacteria, hydrogen-oxidizing bacteria, and electrobiological systems using data on gas composition, reactor scale, productivity, and product quality. Biomass utilization is considered only in relation to the captured carbon balance and process value. Figure 1 presents the comparison framework.

Figure 1. Microbial platforms for CO₂ utilization from energy systems and the scope of whole-process assessment. CBB, Calvin–Benson–Bassham cycle; WLP, Wood–Ljungdahl pathway; HOB, hydrogen-oxidizing bacteria; PHB, poly(3-hydroxybutyrate). Authors’ schematic based on [6,10,11].

2. Literature Selection and Analytical Criteria

Publications were identified through targeted searches of PubMed, Europe PMC, publisher resources, and institutional repositories, supplemented by citation tracking. Queries combined microorganism and platform names with CO₂ fixation, flue gas, gas fermentation, microbial electrosynthesis, pilot operation, carbon balance, and techno-economic assessment. The main publication period was 2010–2026, with a literature cut-off of 28 October 2026; earlier sources supported foundational mechanisms. The bibliography contains 135 sources, including contextual documents and a published correction. Reference identities were checked by DOI and title.

Comparative analysis included only publications addressing microbial inorganic carbon assimilation, cultivation or reactor performance, gas stream integration, or associated energy, environmental, and economic assessments. Experimental studies and process models supplied quantitative evidence; reviews and institutional documents supported mechanisms and context. Numerical comparisons required identifiable units and measurement bases. Evidence was appraised by feed identity, scale, duration, reported uncertainty, and balance coverage, distinguishing laboratory results, pilot demonstrations, and long-term operation. Missing operating details and inventories were not reconstructed. In the tables, n.r. means not reported and n.v. indicates not verified for the comparison; neither denotes zero performance.

Comparisons were grouped by gas flow CO₂ removal, biomass or product productivity, product carbon origin, electron recovery, and whole-system energy or climate performance. Concentrations and volumetric productivities use g·L⁻¹ and g·L⁻¹·d⁻¹ where justified; areal, electrode-normalized, illumination-hour, and biochemical bases remain explicit. Daily equivalents of hourly maxima are conversions, not measured 24 h averages. TS/VS and the stated ground, biofilm, electrode, and catholyte bases were retained. CO₂UE retains the definition used in each study.

3. CO₂ Stream Composition and Integration Conditions

A carbon-containing stream should be characterized by its origin, CO₂ fraction, flow rate, accompanying constituents, and temporal variability [22]. The CO₂ fraction affects its partial pressure and availability in the liquid phase but does not determine reactor loading alone: the same composition at a different flow rate implies a different carbon supply and contact time [23]. CO₂ dissolution changes inorganic carbon equilibria and pH; increasing the supply can therefore alleviate substrate limitation or impair growth conditions [13]. Fossil or biogenic origin also affects the climate assessment of the product [6].

The significance of individual constituents depends on the platform biology. O₂ serves as an electron acceptor for HOB and is produced during oxygenic photosynthesis, whereas it presents an obstacle during anaerobic acetogenesis that requires gas conditioning [24]. H₂ supplies reducing equivalents in non-photosynthetic systems, while N₂ may primarily dilute the feed or serve as a substrate for nitrogen-fixing cultures [25]. CO introduces a different carbon and electron balance from that of pure CO₂ [26]. Sulfur and nitrogen oxides, gas temperature, and substances introduced with the medium require assessment of tolerance and treatment costs [13]. Stream suitability depends on matching the complete feed to the organism and product.

The CO₂ source simultaneously affects carbon loading, gas conditioning, and the physiological requirements of the culture [27]. In a Tetraselmis subcordiformis study, biogas supplied for combustion contained 63.1 ± 1.2% CH₄, 32.9 ± 1.7% CO₂, 1.2 ± 0.3% O₂, and 3.4 ± 0.6% N₂, together with 70 ± 5 parts per million (ppm, by volume) H₂S. Combustion produced gas containing 13.4 ± 1.2% CO₂, 8.2 ± 0.9% O₂, and 75.8 ± 2.3% N₂; the corresponding fractions for coal–biomass co-combustion were 14.9 ± 1.7%, 6.9 ± 2.1%, and 73.7 ± 1.9% [23]. These compositions characterize specific streams and depend on fuel, excess air, and treatment. In PBRs with a working volume of 10 L, the gases were cooled to 20 °C, with a CO₂ supply of 8.0 mg·min⁻¹ and gas recirculation of 10.8 L·min⁻¹. The gas–liquid contact intensity was therefore an important contributor to the result, alongside the CO₂ concentration itself [23].

A lower CO₂ fraction does not necessarily make a stream less suitable. In a C. vulgaris experiment using flue gas from a municipal heating plant, untreated gas contained 5.92 ± 0.31% CO₂, 10.2 ± 2.1% O₂, 801 ± 91 mg SO_x·m⁻³, and 284 ± 21 mg nitrogen oxides (NO_x)·m⁻³. After additional treatment, the CO₂ fraction remained similar at 5.61 ± 0.22%, while SO_x was reported as undetectable. The growth-phase biomass production rate increased from 0.1472 ± 0.0121 to 0.2810 ± 0.0162 g VS·L⁻¹·d⁻¹ [13]. This comparison shows that the observed benefit can arise from inhibitor removal rather than an increased carbon supply. The study used PBRs with a working volume of 120 L and pH-triggered gas feeding, limiting direct comparison with continuously sparged cultures.

Oxygen levels compatible with photosynthetic processes or aerobic HOB cultivation present a substantial obstacle during acetogenesis. Mohr et al. investigated a sequential system in which Parageobacillus thermoglucosidasius first removed O₂ from a 50:50 CO/air mixture, after which Clostridium ljungdahlii utilized the H₂ and CO₂ produced. The first stage lasted 70 h; an initial 3.3 mmol CO yielded 0.52 mmol acetate, equivalent to 0.16 mol·mol⁻¹ CO and approximately 63% of the theoretical maximum [24]. This result demonstrates biological feed conditioning rather than acetogen tolerance to oxygen-containing gas. The additional stage increases the process duration and requires its own medium and energy balance.

Gas separated during biogas upgrading differs from the flue gas produced through biogas combustion, and methane retention matters to the overall energy balance. Photosynthetic O₂ production during contact with raw biogas can alter outlet gas quality even when CO₂ is effectively removed. Conversely, supplying separated CO₂ with external H₂ enables gas-fed systems to operate without light. A pilot demonstration of biological methanation using contaminated CO₂ from a biogas plant comprised two 1000 L trickle beds [28]. Such validation confirms the importance of real-stream operation, although archaeal activity cannot be directly extrapolated to bacteria or microalgae.

Carbon supplied as CO requires separate interpretation. CO oxidation can supply electrons while also producing CO₂; syngas fermentation is therefore not process-equivalent to the reduction of pure CO₂ with hydrogen [29]. An industrial pilot using engineered Clostridium autoethanogenum for acetone and isopropanol synthesis achieved continuous productivity of up to approximately 72 g·L⁻¹·d⁻¹ (daily equivalent of 3 g·L⁻¹·h⁻¹) and approximately 90% selectivity [26]. This is an advanced example of gas bioconversion, but its energetic conditions and electron balance differ from those of H₂–CO₂ operation. Any such comparison should account for the uptake of each gaseous constituent and CO₂ formation through CO oxidation.

Supplying a dissolved one-carbon (C₁) carrier reduces gas transfer dependence but adds a carrier production stage. Engineered Escherichia coli initially doubled in approximately 70 h on formate and CO₂, with a yield of 1.5 g TS·mol⁻¹ formate; adaptation reduced the doubling time below 8 h and increased the yield to 2.3 g TS·mol⁻¹ [30]. Bicarbonate can provide a dissolved carbon and alkalinity reservoir, as demonstrated by an alkalihalophilic cyanobacterium [31]. Carrier carbon and production energy therefore belong in the system boundary. Table 1 summarizes stream compositions and outcomes. For T. subcordiformis, the reported CO₂UE is estimated as 1.88 times the increase in dry biomass divided by supplied CO₂ mass [23]. This metric represents a biomass-derived CO₂ estimate rather than the biomass weight divided by the CO₂ mass alone or a measured inlet–outlet molar flow removal efficiency.

Table 1. Carbon stream compositions, operating parameters, and outcomes of biological utilization.

Microorganism or System	Gas Composition and Carbon Source	Operating Conditions and Parameters	Biomass or Product	CO2 Utilization and Fixation	Ref.
Biogas after treatment, before combustion	CH4 63.1 ± 1.2%; CO2 32.9 ± 1.7%; O2 1.2 ± 0.3%; N2 3.4 ± 0.6%; H2S 70 ± 5 ppm	Fuel supplied to the furnace in the T. subcordiformis experiment	Not applicable: fuel stream before combustion	Not applicable: characterization of the feed to the flue gas source	[23]
Tetraselmis subcordiformis, flue gas from biogas combustion	CO2 13.4 ± 1.2%; O2 8.2 ± 0.9%; N2 75.8 ± 2.3%; CO 141 ± 14 ppm; NOx 145 ± 33 ppm; SOx 83 ± 11 ppm	10 L; 20 ± 2 °C; CO2 8.0 mg·min−1; gas recirculation 10.8 L·min−1	0.264 ± 0.011 g VS·L−1·d−1; maximum 2.660 ± 0.123 g VS·L−1	CO2 fraction: 13.4 → 1.3%; reported reduction 90.3 ± 3.2%; CO2UE 43.1 ± 3.1%	[23]
Tetraselmis subcordiformis, coal–biomass flue gas	CO2 14.9 ± 1.7%; O2 6.9 ± 2.1%; N2 73.7 ± 1.9%; CO 91 ± 31 ppm; NOx 140 ± 19 ppm; SOx 139 ± 11 ppm	10 L; 20 ± 2 °C; CO2 8.0 mg·min−1; recirculation 10.8 L·min−1	0.286 ± 0.009 g VS·L−1·d−1; maximum 2.710 ± 0.091 g VS·L−1	CO2UE 46.7 ± 2.4%	[23]
Chlorella vulgaris, municipal heating plant flue gas	With SOx: CO2 5.92 ± 0.31%; O2 10.2 ± 2.1%; SOx 801 ± 91 and NOx 284 ± 21 mg·m−3. After desulfurization: CO2 5.61 ± 0.22%; SOx undetectable	120 L; 23 ± 2 °C; 270–300 µmol photons·m−2·s−1; light/dark 12:12 h; pH-controlled gas feeding	Growth-phase production: 0.1472 ± 0.0121 → 0.2810 ± 0.0162 g VS·L−1·d−1; 2.120 ± 0.123 g VS·L−1 after desulfurization	CO2UE: 24.1 → 41.4%	[13]
Nannochloropsis oceanica, model gas	CO2 6, 30, and 50%; N2 used as the balance gas	Irradiance of 20, 60, and 80 µmol photons·m−2·s−1, respectively	Biomass of 1.24, 2.12, and 2.45 g TS·L−1, respectively	Percentage utilization of the CO2 feed: n.v.	[32]
Synechococcus elongatus, capillary PBR	CO2 0.04, 10, and 15%; best growth at 10%	27 °C; initial pH 8.2; BG11; 3000 lx; medium 34.4–86.6 mL·min−1; batch and semi-continuous	Maximum 2.1 g TS·L−1; growth rate 1.76 d−1	Reported biofixation 2.08 g CO2·L−1·d−1	[33]
Acetogenic culture, trickle bed	H2/CO2 = 70:30; model gas	Gas 3.8 mL·min−1; medium 10 mL·min−1; pH adjusted to 7 ± 0.3 once daily; cell recycling	Acetate: approximately 1 g·L−1·d−1; maximum concentration 40.84 g·L−1	At least 85% of acetate carbon from CO2 without initial glucose addition	[34]
Cupriavidus necator H16, model gas	H2/CO2/O2 = 85:10:2, initial ratio	Dissolved O2 <1.6 mg·L−1; gradual increase in O2 supply	13.1 ± 0.4 g TS·L−1; PHB 79 ± 2%	Approximately 98% of the PHB produced originated from CO2	[35]
Ralstonia eutropha, diluted mixture	H2/O2/CO2/N2 = 3.6:7.6:12.3:76.5	144 h; nitrogen limitation during accumulation	PHB up to 70% of cell mass	CO2 utilization balance: n.v.	[36]
Parageobacillus thermoglucosidasius and Clostridium ljungdahlii, sequential system	CO/air = 50:50; subsequent utilization of the H2 and CO2 produced	50 mL medium; biological O2 removal over 70 h	0.52 mmol acetate from an initial 3.3 mmol CO; 0.16 mol acetate·mol−1 CO	63% of theoretical yield relative to CO; not a CO2 feed recovery measurement	[24]

TS, total solids; VS, volatile solids; PBR, photobioreactor; PHB, poly(3-hydroxybutyrate); BG11, cyanobacterial medium; CO₂UE, source-defined utilization metric; SO_x/NO_x, sulfur/nitrogen oxides; ppm, parts per million by volume. n.r., not reported; n.v., not verified for the comparison.

4. Assimilation Pathways and Energetic Constraints

4.1. The Calvin–Benson–Bassham Cycle and Carbon Concentration

The Calvin–Benson–Bassham (CBB) cycle comprises ribulose-1,5-bisphosphate (RuBP) carboxylation, 3-phosphoglycerate (3-PGA) reduction, and regeneration of the CO₂ acceptor. The carboxylating enzyme is ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO). Synthesis of one triose phosphate molecule from three CO₂ molecules requires nine adenosine triphosphate (ATP) molecules and six reduced nicotinamide adenine dinucleotide phosphate (NADPH) molecules, equivalent to three ATP and two NADPH per CO₂ [37]. Pathway comparisons must use the same product basis: the cost of producing a triose is not equivalent to that of synthesizing acetate, lipids, or protein. CBB stoichiometry excludes cellular maintenance, biosynthesis, and the energy required to supply substrates.

RuBisCO also reacts with O₂; local CO₂–O₂ availability affects carboxylation and competing reactions [38]. Carbon-concentrating mechanisms (CCMs) combine inorganic carbon transport and intracellular organization to increase CO₂ near the enzyme [39]. Cyanobacterial carboxysomes depend on microcompartment structure [40] and scaffold interactions [41]. In many microalgae, the chloroplast pyrenoid serves a corresponding role; Essential Pyrenoid Component 1 (EPYC1) mediates RuBisCO condensation in Chlamydomonas reinhardtii [42].

Carbon supply must match energy availability. In Nannochloropsis oceanica, the favorable CO₂ fraction rose from 6% to 50% as irradiance increased from 20 to 80 µmol photons·m⁻²·s⁻¹ [32]. Synechocystis instead grew 2.7-fold more slowly at 30% than at 4% CO₂ despite pH control at approximately 7.5 [14]. Process limitations therefore shift between carbon delivery, photon supply, and photosynthetic regulation according to the strain and conditions.

The CBB cycle also operates in non-photosynthetic bacteria. In C. necator, H₂ oxidation and aerobic respiration generate ATP and reductant. Cultivation with an initial H₂/CO₂/O₂ ratio of 85:10:2 and dissolved O₂ below 1.6 mg·L⁻¹ yielded 13.1 ± 0.4 g TS·L⁻¹ [35]. Engineering a bacterial CCM into E. coli provides another route to improve carbon delivery [43]; however, transport costs and system stability remain relevant to host selection. Figure 2 links the CBB cycle to carbon concentration.

Figure 2. Inorganic carbon concentration and assimilation through the Calvin–Benson–Bassham (CBB) cycle in microalgae and cyanobacteria. (A) Inorganic carbon transport and increased CO₂ availability near RuBisCO. (B) Carboxylation, reduction, RuBP regeneration, and diversion of triose phosphates into biosynthesis. CCM, carbon-concentrating mechanisms; RuBisCO, ribulose-1,5-bisphosphate carboxylase/oxygenase; RuBP, ribulose-1,5-bisphosphate; 3-PGA, 3-phosphoglycerate; ATP, adenosine triphosphate; NADPH, reduced nicotinamide adenine dinucleotide phosphate. Authors’ schematic based on [37,40,41,42,43].

4.2. The Wood–Ljungdahl Pathway and Energy Conservation

The Wood–Ljungdahl pathway (WLP) channels two carbon units into acetyl-CoA, a derivative of coenzyme A (CoA): the methyl branch reduces CO₂ through formate and tetrahydrofolate derivatives, whereas the carbonyl branch reduces it to CO [44]. The carbon monoxide dehydrogenase/acetyl-CoA synthase complex catalyzes the final combination of these units [45]. In homoacetogenesis, the overall product is acetic acid, with four H₂ molecules required stoichiometrically per two CO₂ molecules, as shown in Equation (1):

4 H₂ + 2 CO₂ → CH₃COOH + 2 H₂O

(1)

The ATP demand of the methyl branch is offset by substrate-level phosphorylation during acetate synthesis, whereas growth energy depends on additional coupling of redox reactions to an ion gradient [46]. CO₂ assimilation in Acetobacterium woodii has been linked to the generation and use of a sodium gradient [47]. Aceticella autotrophica, an obligate autotroph using an energy-converting hydrogenase complex, exhibits a different organization. This species had a doubling time of 4.5 h at 60 °C and pH 6.0 without added NaCl; the reported standard free energy change for acetate formation was approximately −104 kJ·mol⁻¹ [48]. This value does not represent the energy available in any reactor: at actual H₂ and CO₂ partial pressures and product concentrations, the thermodynamic driving force can be substantially smaller.

Engineering A. woodii yielded 438 mmol·L⁻¹ acetic acid, 7.6 mmol·L⁻¹ acetone, and 14.5 mmol·L⁻¹ isopropanol [49], retaining acetate as the main product. In C. autoethanogenum, ATP homeostasis affected carbon partitioning [50]. Product engineering must therefore preserve energy conservation while improving selectivity and formation rates. Figure 3 outlines WLP organization.

Figure 3. Organization of the Wood–Ljungdahl pathway (WLP) in acetogenic bacteria, including ATP synthesis mechanisms. The methyl branch supplies a methyl group, while the carbonyl branch supplies CO; ACS combines these with CoA to form acetyl-CoA. Subsequent conversion yields acetate, biomass, or reduced products. FDH, formate dehydrogenase; THF, tetrahydrofolate; CFeSP, corrinoid iron–sulfur protein; CODH, carbon monoxide dehydrogenase; ACS, acetyl-CoA synthase; CoA, coenzyme A; ATP, adenosine triphosphate. Authors’ schematic based on [45,47,48,50].

4.3. Alternative Natural Carbon Fixation Pathways

The reductive tricarboxylic acid (rTCA) cycle fixes CO₂ through ferredoxin-dependent carboxylation reactions [51]. Reductive conversion of succinyl-CoA and 2-oxoglutarate, followed by citrate cleavage, supplies acetyl-CoA and regenerates the acceptor [52]. Classical experiments established ferredoxin involvement [53]. On a common glyceraldehyde-3-phosphate (G3P) basis, biochemical analysis yields five ATP for rTCA versus nine for CBB, with different electron-donor requirements [37].

When Chlorobium limicola M1 was grown phototrophically at 30 °C in a 1 L bottle with a CO₂ gas phase, cell extract ATP-citrate lyase activity was 0.01–0.02 µmol acetyl-CoA·min⁻¹·mg⁻¹ protein; equimolar ADP and ATP caused approximately 50% inhibition [54]. These measurements describe enzyme regulation, which represents the relevant endpoint for this mechanistic example.

The 3-hydroxypropionate (3-HP) system of Chloroflexus aurantiacus combines acyl-CoA carboxylation and intermediate reduction [55]. Purified malonyl-CoA reductase had an activity of 10 µmol·min⁻¹·mg⁻¹ at 55 °C and an optimal pH of 7.8, whereas extracts of autotrophic and heterotrophic cells showed activities of 0.08 and 0.03 µmol·min⁻¹·mg⁻¹ protein, respectively [56]. While these pathways expand metabolic engineering options, such enzyme activities require cultivation-scale validation before flue gas applications. Figure 4 summarizes these mechanisms.

Figure 4. Simplified organization of selected natural carbon fixation pathways. (A) The reductive tricarboxylic acid (rTCA) cycle, shown in the ATP-citrate lyase (ACL) variant. Citrate cleavage regenerates oxaloacetate and supplies acetyl-CoA for biosynthesis. (B) The 3-hydroxypropionate (3-HP) bicycle of Chloroflexus aurantiacus. CoA, coenzyme A; ATP, adenosine triphosphate. Authors’ schematic based on [37,53,54,56].

4.4. Synthetic Pathways and Host Engineering

Synthetic pathways can operate in cell-free systems [57,58] or be introduced into hosts that previously depended on organic carbon. Host engineering must integrate assimilation with energy supply, growth, and whole-cell regulation, as demonstrated for E. coli [59] and Pichia pastoris [60].

The synthetic crotonyl-CoA/ethylmalonyl-CoA/hydroxybutyryl-CoA (CETCH) cycle combines 17 enzymes from nine organisms and fixes approximately 5 nmol CO₂·min⁻¹·mg⁻¹ protein in a cell-free system [58]. The protein-normalized rate characterizes the reaction network rather than a cultivation process.

Converting E. coli to growth with CO₂-derived biomass carbon required approximately 200 days of adaptive evolution and yielded a doubling time of 18 ± 4 h [59]. Formate oxidation still supplied energy and could release CO₂. A separate reductive-glycine strategy enabled formate-supported growth with a doubling time below 8 h after adaptation [30]. Both routes therefore require including the energy–substrate balance.

Engineering Pichia pastoris involved eight heterologous genes and three native gene deletions. With CO₂ as the carbon source and methanol supplying energy, maximum specific growth increased from 0.008 to 0.018 h⁻¹ after adaptation [60]. Deployment requires stable target product formation without continuous selection, acceptable cofactor and energy inputs, and recovery. Table 2 relates these mechanisms to their experimental endpoints.

Table 2. CO₂ assimilation pathways with examples of organisms, conditions, and outcomes.

Pathway and Organism	Feed and Study Conditions	Objective and Theoretical Outcomes	Experimental Outcome	Ref.
CBB; Nannochloropsis oceanica	Gas containing 6, 30, and 50% CO2; 20, 60, and 80 µmol photons·m−2·s−1	Triose synthesis: 3 CO2, 9 ATP, 6 NADPH; cycle requirement alone	Biomass of 1.24, 2.12, and 2.45 g TS·L−1, respectively	[32,37]
CBB; Synechococcus elongatus	0.04–15% CO2; 27 °C; initial pH 8.2; BG11; 3000 lx	Light-driven CO2 assimilation; CBB stoichiometry as above	2.1 g TS·L−1; biofixation 2.08 g CO2·L−1·d−1; growth rate 1.76 d−1	[33]
CBB; Cupriavidus necator H16	H2/CO2/O2 = 85:10:2; dissolved O2 <1.6 mg·L−1	CO2 assimilation; ATP and reducing power from H2 oxidation	13.1 ± 0.4 g TS·L−1; PHB 79 ± 2%; approximately 98% of PHB from CO2	[35,37]
WLP; Acetobacterium woodii	H2–CO2; continuous cultivation with complete cell retention and increased yeast extract supply	2 CO2 + 4 H2 → CH3COOH + 2 H2O	Almost 14 g TS·L−1; maximum 148 g acetate·L−1·d−1; inhibition at 8–12 g·L−1	[47,61]
WLP; Aceticella autotrophica 3443-3AcT	H2–CO2; 60 °C; pH 6.0; no added NaCl	Homoacetogenesis; standard ΔG approximately −104 kJ·mol−1 acetate	Doubling time 4.5 h	[48]
WLP; Acetobacterium woodii [pJIR750_ac1t1]	H2–CO2; continuous gas supply to the bioreactor; engineered solvent pathway	Redirection of acetyl-CoA flux toward acetone and isopropanol	438 mmol·L−1 acetic acid; 7.6 mmol·L−1 acetone; 14.5 mmol·L−1 isopropanol	[49]
rTCA; Chlorobium limicola M1	Phototrophy; 30 °C; 1 L bottle with a CO2 gas phase; enzyme studied separately	Model to G3P: 5 ATP versus 9 ATP for CBB; different electron donor set	Extract citrate lyase: 0.01–0.02 µmol acetyl-CoA·min−1·mg−1 protein; approximately 50% inhibition at ADP:ATP = 1	[37,53,54]
3-HP; Chloroflexus aurantiacus	Autotrophic and heterotrophic cells; malonyl-CoA reductase studied at 55 °C; optimal pH 7.8	Reduction of 1 malonyl-CoA consumes 2 NADPH; a 3-HP cycle step	Purified enzyme: 10 µmol·min−1·mg−1; extracts: 0.08 and 0.03 µmol·min−1·mg−1 protein	[56]
CETCH; cell-free system	17 enzymes from 9 organisms; CO2 and regenerated cofactors	Synthetic CO2 fixation network independent of natural cellular organization	Approximately 5 nmol CO2·min−1·mg−1 protein	[58]
Synthetic CBB; Escherichia coli	CO2 as carbon source; formate as energy substrate; approximately 200 days of adaptation	Autotrophic origin of new biomass carbon	Doubling time 18 ± 4 h; isotopic verification of carbon origin	[59]
Reductive glycine pathway; Escherichia coli	Formate and CO2; host adaptation	Growth on a dissolved C1 intermediate	Doubling time approximately 70 → <8 h; yield 1.5 → 2.3 g TS·mol−1 formate	[30]
Synthetic CBB; Pichia pastoris, CBS7435-derived constructs	CO2 as carbon source; methanol as energy substrate; 8 genes added and 3 deleted	Reconstruction of peroxisomal methanol assimilation for CO2 assimilation	Maximum specific growth rate 0.008 → 0.018 h−1 after adaptation	[60]

CBB, Calvin–Benson–Bassham cycle; WLP, Wood–Ljungdahl pathway; rTCA, reductive tricarboxylic acid cycle; 3-HP, 3-hydroxypropionate pathway; CETCH, synthetic crotonyl-CoA/ethylmalonyl-CoA/hydroxybutyryl-CoA cycle; G3P, glyceraldehyde-3-phosphate; ATP, adenosine triphosphate; ADP, adenosine diphosphate; NADPH, reduced nicotinamide adenine dinucleotide phosphate; CoA, coenzyme A; TS, total solids (dry biomass); PHB, poly(3-hydroxybutyrate); BG11, mineral culture medium for cyanobacteria.

5. Microalgae and Cyanobacteria

5.1. Biological Basis and Microorganism Selection

Microalgae and cyanobacteria both use light to generate the energy and reducing power required for CO₂ assimilation [62]. From a process perspective, they form a photosynthetic platform, although eukaryotic microalgae [42] and prokaryotic cyanobacteria [40] differ in cellular organization, carbon-concentrating mechanisms, and opportunities for metabolic modification. Their joint assessment should consider photon availability, inorganic carbon speciation, nutrients, and product recovery. Increasing the CO₂ supply does not resolve light limitation or nitrogen and phosphorus deficiencies [10].

Organism selection depends on medium properties and intended product use. Chlorella vulgaris is a representative host for freshwater cultivation [13], whereas Nannochloropsis [10] and Tetraselmis [23] require appropriate salinity. Among cyanobacteria, Arthrospira platensis is used for biomass production [63], while Synechococcus elongatus [64] and Synechocystis sp. PCC 6803 [65] also serve as hosts for secreted metabolite synthesis. Wastewater cultivation additionally requires assessment of mixotrophy and bacterial interactions. Biomass growth may incorporate organic carbon from the medium or CO₂ generated by its oxidation; consequently, gas removal alone does not establish the origin of the product carbon [66].

Photosynthetic biomass does not have a constant composition. In T. subcordiformis supplied with different gases, protein accounted for 28.9–30.0% TS, lipids for 7.2–8.1% TS, and carbohydrates for 36.5–37.2% TS; the total carbon content was 539–577 mg·g⁻¹ TS [23]. Product selection therefore changes the optimization criterion: maximizing biomass growth, sugar secretion, or polymer accumulation requires different conditions [67]. Cyanobacteria producing sucrose [64] or isoprene [65] allow metabolite recovery without harvesting all cells each time, but they require construct stability and effective product separation.

5.2. Operating Conditions and Process Parameters

The principal parameters are photon flux, optical path length, temperature, pH, salinity, nutrient supply, and CO₂ loading [68]. Gas delivery should ensure carbon availability while limiting acidification and outlet losses. In suspension cultures, increasing the cell concentration intensifies self-shading [10]; in attached systems, biofilm thickness and resistance within the support become important [69]. Reactor geometry therefore affects both gas transfer and light availability [70].

In the outdoor Nannochloropsis comparison introduced in Section 1, vertical configurations outperformed horizontal systems under the same climate. The maximum photosynthetic efficiencies were 4.2% for the vertical tubular reactor, 3.8% for the flat panel, 1.8% for the horizontal tubular system, and 1.5% for the raceway [10]. The flat panel yielded the highest average productivity and efficiency among vertical configurations. Annual designs must also allow for fouling, seasonal changes, and restart losses.

Rao et al. reported 1.24 g TS·L⁻¹ at 6% CO₂ and 20 µmol photons·m⁻²·s⁻¹, 2.12 g TS·L⁻¹ at 30% and 60 µmol·m⁻²·s⁻¹, and 2.45 g TS·L⁻¹ at 50% and 80 µmol·m⁻²·s⁻¹ with N. oceanica [32]. The contrasting Synechocystis response highlights strain-dependent regulation [14]. For S. elongatus in a capillary PBR, favorable conditions included 10% CO₂, an initial pH of 8.2, BG11 medium, 27 °C, and 3000 lx; medium recirculation ranged from 34.4 to 86.6 mL·min⁻¹. The culture reached 2.1 g TS·L⁻¹, a specific growth rate of 1.76 d⁻¹, and reported biofixation of 2.08 g CO₂·L⁻¹·d⁻¹ [33]. Lux cannot be converted to photon flux without the light spectrum.

Immobilization and biofilms increase cell retention but do not eliminate loading constraints [71]. In the system studied by Dębowski et al. using a model gas containing 25% CO₂, increasing the flow rate from 25 to 200 L·h⁻¹ reduced the percentage removal from approximately 40 to 5.9% [69]. In a year-round greenhouse pilot with 15 panels and 90 m² of biofilm area, the average productivity was 3.07 g TS·m⁻² system footprint·d⁻¹ [70]. The first result describes the effect of contact time; the second describes sustained operation with a specific geometry. The system-footprint basis and year-long averaging distinguish this result from seasonal PBR maxima. Figure 5 compares these cultivation configurations.

Figure 5. Photosynthetic cultivation in three configurations: (A) an open raceway with suspended cells; (B) a flat-panel photobioreactor (PBR); (C) a biofilm on a support surface. Authors’ schematic based on [10,69,70].

5.3. Research Findings and Deployment Constraints

Real-gas cultivation is constrained by inhibitors, acidification, and declining activity during the cultivation cycle [72]. Gas loading and conditioning should therefore be selected for sustained production rather than maximum short-term removal.

C. vulgaris on biogas-combustion flue gas reached 0.780 ± 0.058 g VS·L⁻¹, with a growth-phase production rate of 0.0778 ± 0.0031 g VS·L⁻¹·d⁻¹ [12]. A. platensis achieved 0.209 ± 0.017 g VS·L⁻¹·d⁻¹ and 2.040 ± 0.049 g VS·L⁻¹, with gas-phase CO₂ decreasing from approximately 13 to 1.3% [63]. Both studies also reported acidification and premature culture decline. The desulfurization experiment described in Section 1 and Section 3 identifies gas conditioning as an actionable intervention [13]; its value depends on the additional stable production relative to conditioning costs.

For T. subcordiformis on coal–biomass flue gas, the reported biomass production rate was 0.286 ± 0.009 g VS·L⁻¹·d⁻¹ with a maximum concentration of 2.710 ± 0.091 g VS·L⁻¹ and CO₂UE of 46.7 ± 2.4%. With biogas-combustion flue gas, CO₂UE was 43.1 ± 3.1%, whereas the reported reduction in the gas-phase CO₂ fraction was 90.3 ± 3.2% [23]. The biomass-based estimate and gas-fraction change quantify different endpoints. A separate 6000 L Chlorella LJ34 pilot reportedly removed 85.3% of the nitrogen and 76.7% of the phosphorus from an industrial concentrate supplied with flue gas containing 13% CO₂ [73]. This result supports an integrated treatment function, while biomass quality remains a separate requirement.

Engineered cyanobacteria enable targeted synthesis [74]. S. elongatus expressing CscB secreted up to 0.0361 g sucrose·L⁻¹·h⁻¹ of illumination and exceeded a cumulative sucrose production of 10 mmol·L⁻¹ [64]. An engineered PCC 7942 strain sustained 2,3-butanediol production for 21 d, with a cumulative production of 2.38 g·L⁻¹ and a maximum 3-day average rate equivalent to 0.236 g·L⁻¹·d⁻¹ [75]. Synechocystis PCC 6803 produced up to 0.148 g isoprene·L⁻¹·d⁻¹ without continuous antibiotic selection [65]. PHB accumulation during nitrogen starvation can instead use previously stored glycogen; a 21-day experiment recorded final optical densities of approximately 0.51 for an enhanced-lysis variant and 1.15 for the other variants and the wild type [76]. Product secretion, construct stability, and whole-cycle yield therefore guide process design.

Deployment prospects combine natural light, local nutrients, and a defined product market [77]. The main constraints are seasonality [10], dilute biomass recovery [78], and contaminants; wastewater-treatment value can materially affect economics [79]. Carbonate precipitation adds another carbon-containing product [80,81]. A newer cultivation model further identifies inoculum design and culture failure frequency as cost drivers [82]. Table 3 separates production, CO₂ metrics, and this economic benchmark.

Table 3. Microalgae and cyanobacteria: production, CO₂ metrics, and an updated cultivation-cost benchmark.

Microorganism or System	Gas Composition and Carbon Source	Operating Conditions and Parameters	Biomass or Product	CO2 Utilization and Fixation	Comparison Limitation	Ref.
Chlorella vulgaris	Flue gas before and after desulfurization; CO2 5.92 ± 0.31 and 5.61 ± 0.22%	120 L; 23 ± 2 °C; 270–300 µmol photons·m−2·s−1; light/dark 12:12 h; pH-controlled gas feeding	0.1472 ± 0.0121 → 0.2810 ± 0.0162 g VS·L−1·d−1; 2.120 ± 0.123 g VS·L−1	CO2UE 24.1 → 41.4%; air control 53.8%	Gas conditioning and higher CO2 supply; different costs among configurations	[13]
Tetraselmis subcordiformis	Coal–biomass flue gas; CO2 14.9 ± 1.7%; O2 6.9 ± 2.1%; N2 73.7 ± 1.9%	10 L; 20 ± 2 °C; CO2 8 mg·min−1; 270–300 µmol photons·m−2·s−1; light/dark 14:10 h	0.286 ± 0.009 g VS·L−1·d−1; maximum 2.710 ± 0.091 g VS·L−1	CO2UE 46.7 ± 2.4%; biomass-to-CO2 factor 1.88	Biomass-derived estimate; not a complete carbon balance	[23]
Tetraselmis subcordiformis	Flue gas from biogas combustion; CO2 13.4 ± 1.2%; O2 8.2 ± 0.9%; N2 75.8 ± 2.3%	10 L; parameters as above; real gas	0.264 ± 0.011 g VS·L−1·d−1; maximum 2.660 ± 0.123 g VS·L−1	Reported CO2 fraction reduction 90.3 ± 3.2%; CO2UE 43.1 ± 3.1%	Growth and gas-fraction reduction have different optima	[23]
Arthrospira platensis	Flue gas from biogas combustion; approximately 13% CO2	Batch cultivation; comparison with air; pH decline during the cycle	0.209 ± 0.017 g VS·L−1·d−1; 2.040 ± 0.049 g VS·L−1	Gas-phase CO2 fraction approximately 13 → 1.3%	Early culture decline; active-phase rate is not a whole-cycle average	[63]
Nannochloropsis, vertical systems	CO2 supplied to outdoor cultures	Same climate; vertical and horizontal geometries; experimental season	Vertical: 19–24 g TS·m−2 ground area·d−1; horizontal: 12–15	Common CO2 balance: n.r.; maximum photosynthetic efficiency of the vertical tubular system 4.2%	Seasonal productivity and light-use efficiency; no CO2 recovery percentage	[10]
Nannochloropsis oceanica	Model gas; CO2 6, 30, and 50%	Irradiance of 20, 60, and 80 µmol photons·m−2·s−1, respectively	Biomass 1.24, 2.12, and 2.45 g TS·L−1	CO2 feed utilization balance: n.v.	Final concentrations; CO2 fraction matched to light supply	[32]
Microalgae, biofilm PBR	Engine exhaust gas; greenhouse cultivation	15 panels; 90 m2 biofilm; one year of operation; periodic harvesting	Mean 3.07 g TS·m−2 system footprint·d−1	Common CO2 utilization metric: n.v.	Annual footprint-based mean; averaging period and configuration differ	[70]
Immobilized microalgae	Model gas containing 25% CO2	Flow rates of 25 and 200 L·h−1; cells within a support	Cell retention; comparable biomass productivity: n.r.	CO2 removal approximately 40 and 5.9%, respectively	Removal percentages at different flow rates require conversion to mass flow	[69]
Chlorella LJ34	Laboratory: 20% CO2. Pilot: flue gas containing 13% CO2 and industrial concentrate	Laboratory selection followed by a 6000 L tubular PBR	Laboratory: 1.26 g TS·L−1. Pilot: 85.3% nitrogen and 76.7% phosphorus removal	Authors’ laboratory carbon fixation metric: 0.19 g·L−1·d−1	Laboratory and pilot results are distinct; nutrient removal is not a CO2 balance	[73]
Synechococcus elongatus	Gas containing 0.04, 10, and 15% CO2; best growth at 10%	Non-wetted and hydrophilized hollow fibers; 27 °C; initial pH 8.2; BG11; 3000 lx	Maximum 2.1 g TS·L−1; growth rate 1.76 d−1	Reported biofixation 2.08 g CO2·L−1·d−1	Batch and semi-continuous modes and wetting state; no equivalence between lx and µmol photons	[33]
Synechococcus elongatus with CscB transporter	CO2 assimilation; photosynthesis; sucrose-secreting construct	Heterologous proton/sucrose transporter; rate normalized to illumination time	Sucrose: up to 0.0361 g·L−1·h−1 of illumination; over 10 mmol·L−1	Percentage recovery of the CO2 feed: n.r.	Product secretion differs from 24 h productivity and gas balance	[64]
Synechococcus elongatus PCC 7942	50 mmol·L−1 NaHCO3 in BG11; inorganic carbon supplementation	25 mL culture; 30 °C; continuous light; daily 10% medium replacement; 21 d production	Cumulative 2,3-butanediol production 2.38 g·L−1; maximum 3-day average rate 0.236 g·L−1·d−1	CO2 feed recovery not applicable to this bicarbonate-fed comparison	Cumulative product differs from peak rate; include bicarbonate carbon	[75]
Synechocystis PCC 6803	CO2 as carbon source; photosynthesis	Engineered isoprene synthesis; phenotype maintained without continuous antibiotic selection	Isoprene up to 0.148 g·L−1·d−1	Comparable percentage CO2 utilization: n.r.	Construct stability and metabolite amount do not independently establish the gas balance	[65]
Nannochloropsis oceanica IMET1 and biomineralization	1 L: 10% CO2. 500 L: 5%; NaHCO3 0.02 mol·L−1	Alkalinity control; CaCO3 precipitation; model gas	1 L: 1.52 g TS·L−1 on day 12; second product: carbonate precipitate	Additional cell retention in precipitate: 9.3% at 1 L; 60% on day 40 at 500 L; maximum 63.2 g CO2·m−2·d−1 on day 35	Gas balance and sources of Ca2+, NaHCO3, and alkalinity; precipitate stability	[81]
Chlorella HS2 and biomineralization	CO2 and CaCl2 addition; medium providing alkalinity	Semi-continuous cultivation; comparison with no-addition control	Cumulative biomass production increased 4.67-fold	CO2 fixation metric increased 10.9-fold	Fold changes relative to the control do not establish absolute flux or net removal	[83]
Raceway cultivation model	Raceway scenarios; gas n.v.	Seed train, cultivation, dewatering; failure sensitivity (2025 model)	Minimum selling price (USD·t−1 biomass): 1130–1200 semi-continuous; 1380–2040 batch	CO2 feed recovery: n.v.	Model; fuel conversion excluded	[82]

TS, total solids; VS, volatile solids; CO₂UE, source-defined metric; BG11, cyanobacterial medium. n.r., not reported; n.v., not verified for the comparison.

6. Acetogenic Bacteria

6.1. Biological Basis and Microorganism Selection

Acetogens use the Wood–Ljungdahl pathway to assimilate C₁ units and conserve energy, enabling anaerobic production of acetate and other metabolites without illumination [45]. In H₂–CO₂ operation, carbon and electrons are supplied separately [47]; in CO-containing gas, one molecule can serve both functions [50]. Distinguishing these feeds is fundamental to assessing the amount of CO₂ actually converted into product and the energy input associated with the reductant [84].

Strain selection depends on energy conservation, product tolerance, growth, and engineering potential. A. woodii uses a sodium gradient [47], whereas C. ljungdahlii uses proton-coupled electron transport involving the Rnf complex [85]. A comparison of eight less-characterized acetogens with A. woodii identified S. ovata and A. wieringae as promising hosts [86]. Their process performance still depends on cultivation and recovery conditions.

Redirecting synthesis from acetate to a solvent must preserve the cell’s capacity to obtain energy [87]. In C. autoethanogenum, changes in ATP homeostasis affected carbon partitioning [50]. Metabolic engineering can increase product value but requires parallel comparison of selectivity, formation rate, and gas consumption. Acetogenesis is particularly attractive when low-emission H₂ and CO₂ can be supplied continuously and the product can be recovered before inhibitory concentrations are reached.

6.2. Operating Conditions and Process Parameters

The parameters governing the process rate include the anaerobic conditions, H₂ and CO₂ partial pressures, gas–liquid transfer, temperature, pH, as well as cell and product residence times [88]. An insufficient H₂ supply limits both carbon reduction and the energetic driving force. Acidification changes the fraction of undissociated acid, while increasing cell retention without adequate product recovery can sustain biomass with declining activity [61].

The operating range is organism-dependent. Aceticella autotrophica 3443-3Ac^T had a doubling time of 4.5 h at 60 °C and pH 6.0 without added NaCl [48]. A thermophilic mixed-culture biofilm reactor was operated at 55 °C and a hydraulic retention time (HRT) of 1 d [89]. The trickle-bed reactor studied by Steger et al. was supplied with H₂/CO₂ = 70:30 at a gas flow rate of 3.8 mL·min⁻¹ and a medium recirculation rate of 10 mL·min⁻¹; the pH was adjusted to 7 ± 0.3 once daily [34]. These represent different cultivation strategies rather than the limits of a shared operating tolerance.

Product recovery and biocatalyst maintenance can be more important than simply increasing the cell concentration. In A. woodii cultures with microfiltration, complete cell retention decoupled the biomass residence time from the medium exchange rate, but inhibition was still observed above 8–12 g acetate·L⁻¹ [61]. Medium composition also requires control: yeast extract supplies growth factors and organic carbon, while glucose used during trickle-bed start-up changes product partitioning [34]. Their contributions must be included in the assessment of autotrophy.

6.3. Research Findings and Deployment Constraints

For a fixed target product, volumetric productivity determines vessel capacity, product concentration affects separation duty, and cell retention controls washout. Joint optimization is more informative than maximizing any one of these variables.

S. ovata cultures achieved 17.96 g acetate·L⁻¹·d⁻¹, a specific rate of 21.03 g acetate·g⁻¹ TS·d⁻¹ and a final concentration of 32.2 g·L⁻¹ [86]. Continuous A. woodii cultivation with full cell retention reached almost 14 g TS·L⁻¹ and a maximum of 148 g acetate·L⁻¹·d⁻¹ under the highest dilution rate and increased yeast extract supply [61]. These results underscore that performance depends on process selection rather than species-specific differences alone.

A thermophilic biofilm at 55 °C achieved 42.4 g acetate·L⁻¹ in batch and up to 10.5 g·L⁻¹·d⁻¹ continuously; acetate represented more than 98.5 and 99.1% of the total volatile acids, respectively [89]. A trickle-bed reactor without initial glucose addition produced approximately 1 g acetate·L⁻¹·d⁻¹, with at least 85% of the acetate carbon derived from CO₂ [34]. This last percentage is a product carbon origin metric.

The engineered A. woodii product distribution described in Section 4.2 retained acetate as the main stream [49]. The industrial C. autoethanogenum solvent pilot reached approximately 90% selectivity with a daily-equivalent productivity of up to 72 g·L⁻¹·d⁻¹ on CO-containing gas [26], but switching to H₂–CO₂ changes the electron and energy supply. Reductant availability, product recovery, and real-gas stability determine deployment. Table 4 provides daily unit comparisons where the source basis permits them.

Table 4. Acetogens, HOB, and electrosynthesis: production and CO₂ metrics with selected daily unit conversions.

Microorganism or System	Gas Composition and Carbon Source	Operating Conditions and Parameters	Biomass or Product	CO2 Utilization and Fixation	Ref.
Sporomusa ovata	Model H2–CO2; continuous supply	Controlled stirred reactor; standardized acetogen comparison	Acetate: 17.96 g·L−1·d−1; 21.03 g·g−1 TS·d−1; final 32.2 g·L−1	Comparable CO2 feed recovery: n.v.	[86]
Acetobacterium woodii	Model H2–CO2; medium containing yeast extract	Continuous cultivation with microfiltration and complete retention	Almost 14 g TS·L−1; maximum 148 g acetate·L−1·d−1; inhibition at 8–12 g·L−1	Common CO2 utilization metric: n.v.	[61]
Mixed culture, thermophilic biofilm	H2–CO2; model gas	Hollow fibers; 55 °C; batch and continuous operation at HRT 1 d	42.4 g acetate·L−1 in batch; maximum 10.5 g·L−1·d−1 continuously	CO2 balance: n.v.; acetate >98.5 and >99.1% of total acids	[89]
Mixed culture, trickle bed	H2/CO2 = 70:30; no initial glucose in configuration TB3	Gas 3.8 mL·min−1; medium 10 mL·min−1; daily pH adjustment to 7 ± 0.3; cell recycling	Acetate approximately 1 g·L−1·d−1; maximum 40.84 g·L−1	At least 85% of acetate carbon from CO2 according to the TB3 balance	[34]
Acetobacterium woodii [pJIR750_ac1t1]	H2–CO2; continuous feeding	Solvent pathway engineering	Acetic acid 438; acetone 7.6; isopropanol 14.5 mmol·L−1	Percentage CO2 utilization balance: n.v.	[49]
Clostridium autoethanogenum	CO-containing gas; carbon and electron sources differ from H2–CO2	Continuous industrial pilot; acetone- and isopropanol-producing constructs	Up to approximately 72 g product·L−1·d−1 (daily equivalent of 3 g·L−1·h−1); approximately 90% selectivity	This result does not describe reduction of pure CO2 with hydrogen	[26]
Cupriavidus necator H16	H2/CO2/O2 = 85:10:2, initial ratio	Dissolved O2 controlled below 1.6 mg·L−1	13.1 ± 0.4 g TS·L−1; PHB 79 ± 2%	Approximately 98% of the PHB obtained originated from CO2	[35]
Ralstonia eutropha, diluted mixture	H2/O2/CO2/N2 = 3.6:7.6:12.3:76.5	144 h; nitrogen limitation	PHB up to 70% of cell mass	Percentage CO2 balance: n.v.	[36]
Cupriavidus necator, model-based control	H2–CO2–O2; coordinated gas and medium supply	Autotrophic high-cell-density cultivation	53 g TS·L−1; approximately 75% PHB; 67% lower gas consumption	Saving concerns total gas consumption relative to the reference configuration, distinct from CO2 recovery	[11]
Cupriavidus necator, pressurized fermentation	H2–CO2–O2	1.5 and 3 bar; 4 bar also tested	PHB 10.8 and 29.6 g·L−1, respectively	Common CO2 recovery balance: n.v.	[90]
HOB, continuous SCP production	CO2 and electrolytic H2–O2	6 A; 133 A·m−2; HRT 1.55 d; pneumatically mixed reactor	0.89 g TS·L−1·d−1; 72% protein; productivity 7.4-fold higher than in batch	Approximately 94% CO2 utilization; nearly 100% H2 and O2	[91]
HOB, 100 L pilot	CO2 and electrolytic gases	Coupled electrolysis and fermentation; atmospheric pressure	10.8 ± 0.6 g TS·L−1·d−1; maximum approximately 12; approximately 60% crude protein	Comparable CO2 feed recovery: n.v.	[92]
MES, porous biocathode	CO2; electrons from the circuit	Approximately 102 A·m−2 electrode area; mature biocatalyst	685 ± 30 g acetic acid·m−2 electrode area·d−1; up to 11 g·L−1	CO2 recovery 94 ± 2%; electron recovery 100 ± 4%	[93]
MES, carbon foam	CO2; electrons from the circuit	Continuous operation; 8 A·L−1 catholyte	Maximum 18.72 g acetate·L−1 catholyte·d−1; equivalent to 0.78 g·L−1·h−1; acetate coulombic efficiency up to 35%	Common CO2 recovery: n.r.; acetate energy efficiency up to 12%	[94]
MES; Thermoanaerobacter kivui	CO2; cathodically supplied reductant	Compact plate reactor; continuous thermophilic operation	Maximum 84 g acetic acid·L−1 catholyte·d−1; equivalent to 3.5 g·L−1·h−1; titer 14 g·L−1	Complete CO2 and energy inventory: n.v.	[95]
MES coupled to electrodialysis	CO2; electrical electron supply	Bubble-column MES with extraction/concentration; full compartment basis n.v.	Mean acetate production 4.32 g·L−1·d−1; reported maximum concentration 98.1 g·L−1	Whole-process coulombic efficiency 84%; NaOH use 98.7% lower; complete energy inventory n.v.	[96]

HOB, hydrogen-oxidizing bacteria; HRT, hydraulic retention time; PHB, poly(3-hydroxybutyrate); SCP, single-cell protein; MES, microbial electrosynthesis; ED, electrodialysis. n.r., not reported; n.v., not verified for the comparison.

7. Hydrogen-Oxidizing Bacteria

7.1. Biological Basis and Microorganism Selection

Hydrogen-oxidizing bacteria use H₂ as an electron and energy source, with O₂ as the terminal respiratory electron acceptor [35,97]. CO₂ assimilation is therefore coupled to the supply of three gaseous substrates, the demand for which changes with growth and biomass composition. Production without illumination is an advantage; however, dependence on H₂ generation, gas transfer, and the control of their relative supplies are constraints [11].

C. necator H16 is a well-characterized host for biomass and PHB production [35,98]. Choosing between protein or polymer changes the cultivation requirements: growth and protein synthesis require available nitrogen [25], whereas nitrogen limitation can redirect carbon into storage material [36]. A final cell concentration that includes polymers is therefore not equivalent to the active biomass concentration or protein yield.

The nitrogen source also affects the culture structure and metabolic cost. Studies have compared ammonium assimilation with N₂ fixation [25] and examined co-cultures of nitrogen-fixing bacteria [99]. Single-cell protein (SCP) production may increase product value [100], but the crude protein content calculated from nitrogen must be complemented by amino acid composition, digestibility, and contaminant measurements [101]. For waste gas cultivation, product specifications must be assessed together with feed conditioning [102].

7.2. Operating Conditions and Process Parameters

Control should match H₂, CO₂, and O₂ supplies to consumption rates while accounting for gas lost at the outlet. Increasing pressure improves substrate availability but changes the compression energy demand [90]; increasing the O₂ supply must be consistent with the cellular activity and equipment conditions [103]. Feed gas composition and dissolved O₂ concentration describe different levels of control and should be reported separately [35].

The C. necator H16 gas cultivation example used explosion-protected equipment [35]. A diluted H₂/O₂/CO₂/N₂ mixture at a ratio of 3.6:7.6:12.3:76.5 was used over 144 h in another configuration [36]. Continuous electrolysis-coupled cultivation operated at 6 A, equivalent to 133 A·m⁻², and an HRT of 1.55 d [91]. These configurations require separate consideration of headspace composition, dissolved O₂, and medium exchange.

Hydrogen–oxygen cultivation also requires an operating safety envelope for feed lines, headspace, and off-gas. Dissolved O₂ control alone does not demonstrate a nonflammable gas phase. The H16 study used explosion-protected equipment, separate gas supplies, gas monitoring, non-return protection, inert purging, and strong ventilation [35]; dilution was explored in another approach [36]. Scale-up must assess gas accumulation during normal operation, start-up, shutdown, and feed or mixing failure. Monitoring, interlocked isolation, ventilation, and appropriately rated equipment add capital, maintenance, and utility costs.

A two-stage process comprising growth on glycerol followed by PHB accumulation from CO₂ increases the initial biocatalyst mass available for the gas-fed phase [104]. Product formation must nevertheless be accounted for separately from biomass originating from organic carbon. Models separating residual biomass and polymers show that their formation rates and limitations differ [105]. Similarly, medium and gas supply optimization must be considered jointly, rather than attributing the entire improvement to pressure or a single constituent.

7.3. Research Findings and Deployment Constraints

HOB assessment should prioritize biomass productivity and composition, substrate consumption, and energy per unit of the intended protein or polymer. Nitrogen-based crude protein estimates require complementary quality measurements, while polymer accumulation must be distinguished from residual cell growth.

The laboratory H16 culture described in Section 4.1 contained 79 ± 2% PHB, with approximately 98% of the polymer derived from CO₂ [35]. PHB reached 70% of the cell mass under nitrogen limitation in a diluted gas mixture [36]. Model-based gas control and medium optimization enabled approximately 75% PHB at a biomass concentration of 53 g TS·L⁻¹ [11]. These findings support polymer production, while reductant losses and recovery determine whole-process performance.

In pressurized fermentation, the PHB concentration increased from 10.8 g·L⁻¹ at 1.5 bar to 29.6 g·L⁻¹ at 3 bar, but increasing the pressure to 4 bar did not maintain a simple trend [90]. Studies of automatic O₂ control likewise demonstrated the importance of matching supply to demand [103]. Intensification therefore requires confirmation of productivity and energy consumption across the full cycle, alongside final concentration; increased gas solubility alone does not guarantee a proportional benefit.

Continuous SCP cultivation reportedly achieved 0.89 g TS·L⁻¹·d⁻¹, 72% protein, approximately 94% CO₂ utilization, and nearly complete H₂ and O₂ utilization; productivity was 7.4-fold higher than the batch comparison [91]. A separate atmospheric pressure 100 L pilot reported 10.8 ± 0.6 g TS·L⁻¹·d⁻¹, with a maximum near 12 g TS·L⁻¹·d⁻¹, and approximately 60% crude protein [92]. Testing an energy advantage requires matched electricity, gas delivery, and recovery inventories per unit of SCP.

HOB deployment requires low-emission energy and H₂ [106], stable cultivation, and verified product quality. Photovoltaic–protein models include energy-generation land and equipment in addition to the fermenter [107]. Gas supply interruptions, restart losses, and safe feeding protocols also affect plant utilization. Table 4 records production and gas utilization endpoints separately.

8. Microbial Electrosynthesis and Hybrid Systems

8.1. Biological Basis and Microorganism Selection

Microbial electrosynthesis (MES) supplies electrons for CO₂ reduction from an external electrical circuit [108]. The biocatalyst may be a pure acetogenic culture or an enriched microbial community [109]. Electron transfer occurs through interaction with the electrode or via H₂ generated at the cathode [93]. Biofilm presence alone does not establish which mechanism dominates [110]; culture selection should be linked to the electron supply mechanism and product selectivity.

This process configuration partly uses the same metabolic functions as gas-fed acetogenesis [111] or HOB cultivation. Hybrid systems couple water electrolysis with microbial CO₂ assimilation [112], while photosensitization can supply reducing equivalents through a semiconductor [113]. A comparison with a photosynthetic reactor therefore requires accounting for electricity generation and the regeneration of non-biological components.

8.2. Operating Conditions and Process Parameters

Key parameters include electrode potential, current, whole-cell voltage, cathode area, catholyte volume, pH [114], and the residence times of the medium and biocatalyst. Increasing the current can increase electron and H₂ supplies without necessarily increasing the fraction of electrons directed to the target metabolite. Comparisons should distinguish electrode area from reactor volume and coulombic efficiency from energy efficiency [94].

A porous biocathode experiment used approximately 102 A·m⁻² electrode area [93], while a continuous carbon-foam system operated at 8 A·L⁻¹ catholyte [94]. Geometry is therefore needed to compare current intensity. Membrane acid extraction limits product accumulation [115] but introduces electrical resistance and durability requirements [116].

Reactor development should couple biological productivity to gas recovery and separation energetics [18]. In particular, H₂ leaving a cathode-fed culture must be included in the electron and energy accounting [94].

8.3. Research Findings and Deployment Constraints

For acetate production, the volumetric rate and product concentration provide useful engineering comparisons when the volume basis and averaging interval are explicit.

A porous biocathode reportedly achieved 685 ± 30 g acetic acid·m⁻² electrode area·d⁻¹, up to 11 g·L⁻¹, 94 ± 2% CO₂ recovery in the product, and 100 ± 4% electron recovery [93]. The carbon-foam system achieved a daily-equivalent maximum of 18.72 g acetate·L⁻¹·d⁻¹ (0.78 g·L⁻¹·h⁻¹). The combined H₂ and acetate coulombic and energy efficiencies were approximately 80–100% and 35–42%; acetate alone reached maxima of 35% and 12%, respectively [94]. This identifies product selectivity, rather than total electron recovery, as the relevant improvement target.

Recent MES benchmarks change the rate assessment. A compact plate reactor with Thermoanaerobacter kivui reportedly reached a daily-equivalent maximum of 84 g acetic acid·L⁻¹ catholyte·d⁻¹ (3.5 g·L⁻¹·h⁻¹), at 14 g·L⁻¹ [95]. An ED–MES system reported a mean acetate rate of 4.32 g·L⁻¹·d⁻¹, 84% whole-process coulombic efficiency, and 98.7% lower NaOH consumption [96]. While these results shift attention toward sustained selectivity and integrated recovery, they do not encompass a complete energy inventory. Furthermore, the NaOH saving is not an equivalent reduction in total cost.

Coupling water splitting with C. necator yielded a reported CO₂ reduction energy efficiency of nearly 50% relative to the electricity input and approximately 180 g of CO₂ removed·kWh⁻¹ [117]. CdS-based photosensitization enabled acetic acid synthesis by a non-photosynthetic bacterium [113]. Solar and electrical conversion efficiencies retain different denominators [118]; catalyst regeneration and constituent release also require assessment.

Further intensification should combine active biofilm retention [119] with target product recovery and acceptable whole-cell energy demand [18,94]. Table 4 preserves reactor volume and measurement period qualifications alongside daily unit rates.

9. Reactors, Gas Transfer, and Operational Stability

9.1. Substrate Supply and Membrane Functions

In gas-fed bioprocesses, the production rate depends on substrate delivery to the active biomass [120]. For suspensions, this is described in part by the volumetric mass transfer coefficient (k_La); biofilms introduce additional resistance within the biofilm and support. Increasing the contact area, recirculation, or pressure can improve gas supply while also increasing energy demand. Configurations should therefore be assessed in terms of gas transfer, biological performance, and operational stability [121].

A comparison of hollow-fiber, trickle-bed, and stirred reactors found maximum mass transfer coefficients of 1062, 421, and 114 h⁻¹, respectively. However, these measurements used O₂ as the test gas, and the best-performing hollow-fiber module was made from non-porous polydimethylsiloxane [121]. These results describe specific configurations rather than H₂ or CO₂ transfer measurements during cultivation.

For CO in a composite hollow-fiber module, a maximum of 946.6 h⁻¹ was measured at a liquid recirculation rate of 1500 mL·min⁻¹ and a gas overpressure of 206.8 kPa; the tested ranges were 300–1500 mL·min⁻¹ and 34.5–206.8 kPa [122]. Intensification required changes to the hydraulic conditions and pressure, which are both associated with energy consumption. A comparison with a bubble-based culture should consider production and gas utilization per unit of energy input rather than the highest transfer coefficient alone.

Membranes can separate gases, deliver substrates, or retain cells. Direct CO₂ delivery to an H₂-fed biofilm has been demonstrated [123]. In S. elongatus, hollow-fiber wetting and gas composition affected both transfer and cultivation [33]. Cell-retaining filtration instead decouples the biomass residence time from the medium exchange rate, as demonstrated for A. woodii [61]. Figure 6 distinguishes gas delivery from cell retention; active membrane area, culture volume, and pumping or filtration energy are separate design quantities.

Figure 6. Membrane functions in bioreactors. (A) H₂ and CO₂ delivery. (B) Cell retention by a submerged filtration membrane, with independent culture sparging and permeate recovery. SRT, solids retention time; HRT, hydraulic retention time. Authors’ schematic based on [61,123,124].

9.2. Biofilms, Retention, and Variable Supply

In biofilms supplied with gas from the fiber side and nutrients from the liquid side, increasing thickness does not proportionally increase active biomass. A mixed H₂–CO₂ culture on hollow fibers reached 7.4 g·L⁻¹ acetate, 1.8 g·L⁻¹ butyrate, 0.98 g·L⁻¹ caproate, and 0.42 g·L⁻¹ caprylate, with reported complete H₂ utilization [124]. A broader product range was an advantage, whereas separation of the mixture remained a constraint. H₂ uptake efficiency in such a system does not establish target compound selectivity or isolation cost.

Retention limits washout of slow-growing cells but can also sustain low-activity biomass and inhibitor accumulation. Assessing functional recovery after cleaning requires parallel measurements of active biomass, production rate, and substrate consumption [61]. In photosynthetic cultures, a support or contactor additionally changes the light distribution [33]. Increasing the CO₂ transfer area can shift the limitation toward photon availability, which should be assessed under comparable illuminations and geometries [10].

Variable renewable energy supply makes restart performance important. Two 1000 L biological methanation beds tolerated H₂ interruptions of 12–72 h and recovered previous performance within 6–12 h [28]. Another 800 L bed was validated over approximately 450 days with two shutdowns [125]. These archaeal examples provide benchmarks for test duration and interruption protocols.

Year-round biofilm operation [70] complements seasonal PBR measurements [10] by revealing annual performance. For hydrogen-based systems, gas control improvements [11] should likewise be tested across start-up and interruptions. Table 5 summarizes the equipment evidence.

Table 5. Reactor configurations and quantitative examples of intensification.

Configuration	Experimental Parameters	Quantitative Data	Comparison Caveat	Ref.
Hollow-fiber module, trickle-bed, stirred reactor	O2 transfer comparison; polydimethylsiloxane hollow fibers, 6 mm packed-bed spheres, stirred tank	Maximum coefficients: 1062, 421, and 114 h−1, respectively	O2 test gas; not a measurement of CO2 utilization by biomass	[121]
Composite hollow-fiber module	CO; recirculation 300–1500 mL·min−1; overpressure 34.5–206.8 kPa	Maximum 946.6 h−1 at 1500 mL·min−1 and 206.8 kPa	CO and equipment-specific result; comparison requires pumping and compression inputs	[122]
S. elongatus capillary PBR	0.04–15% CO2; medium 34.4–86.6 mL·min−1; 27 °C; 3000 lx	Best growth at 10% CO2; biomass 2.1 g TS·L−1; reported fixation 2.08 g CO2·L−1·d−1	Effects of wetting and operating mode; lux not equated with photon flux	[33]
Immobilized PBR	Model gas containing 25% CO2; flow rates of 25 and 200 L·h−1	Removal approximately 40% and 5.9%, respectively	Removal percentage decreases at higher loading; mass balance per unit time required	[69]
Biofilm PBR	15 panels; 90 m2 biofilm area; engine exhaust; one year of operation	Mean 3.07 g TS·m−2 system footprint·d−1	Annual mean per unit system footprint, not per unit biofilm area	[70]
Gas-fed mixed-culture biofilm	H2–CO2 supplied through hollow fibers	Acetate 7.4; butyrate 1.8; caproate 0.98; caprylate 0.42 g·L−1; complete H2 utilization reported by the authors	Concentrations in a product mixture; H2 uptake does not establish individual acid selectivity	[124]
Pilot methanation beds, reference case	Two 1000 L reactors; H2 supply interruptions	Interruptions of 12–72 h; previous performance restored in 6–12 h	Archaea; test scope is methodologically useful without extrapolating physiology to bacteria	[28]
Long-term methanation, reference case	800 L trickle-bed reactor	Approximately 450 days of operation; two shutdowns	Evidence of operating duration for a specific system; does not establish maturity of the whole group	[125]

10. Capture Balance, Energy, and Costs

10.1. Consistent Metrics and System Boundary

CO₂ removal efficiencies should be calculated using inlet and outlet molar flow rates rather than the difference in volume fractions alone. O₂ production, H₂ uptake, humidity, and auxiliary gas supply change the total flow rate. In the T. subcordiformis experiment, comparing 13.4% inlet CO₂ with 1.3% at the outlet, alongside an increase in O₂ from 8.2 to 21.3% [23], does not automatically yield a reduction in CO₂ molar flow identical to the reduction in its fraction. Equation (2) provides the basis for calculating gas removal efficiency:

$${\eta }_{{\mathrm{CO}}_2}={\displaystyle \frac{q_{{\mathrm{CO}}_2,\mathrm{in}}-q_{{\mathrm{CO}}_2,\mathrm{out}}}{q_{{\mathrm{CO}}_2,\mathrm{in}}}}\times 100\%$$

(2)

where q_CO2,in and q_CO2,out denote the inlet and outlet CO₂ molar flow rates, respectively, and η_CO2 is the percentage removal efficiency.

Biologically retained carbon includes increases in cellular and metabolite carbon, accounting for initial biomass, organic carbon inputs, and changes in dissolved inorganic carbon. The relationship between biomass productivity and the equivalent CO₂ flux uses the measured carbon fraction and a molar mass ratio of 44:12, as shown in Equation (3). A carbon fraction of 50% corresponds to 1.83 kg CO₂ per kilogram of newly formed TS biomass. This is not a universal constant: T. subcordiformis contained 539–577 mg total carbon·g⁻¹ TS, whereas productivity was reported on a VS basis [23]. Conversion therefore requires the TS/VS relationship and identification of carbon origin.

$$r_{{\mathrm{CO}}_2,\mathrm{biomass}}=P_{\mathrm{X}}\times f_{\mathrm{C}}\times {\displaystyle \frac{44}{12}}$$

(3)

where P_X is biomass productivity on a TS basis, f_C is the mass fraction of carbon in TS, and r_CO2,biomass is the equivalent CO₂ retention rate. Units of mass and time, as well as the areal or volumetric reference basis, must be consistent.

A common cradle-to-grave boundary covers CO₂ origin and conditioning; electricity, heat, and H₂ generation; gas delivery; nutrients and cultivation; product recovery; equipment manufacture and replacement; and product use and end of life. Energy and emissions should be reported per functionally equivalent product or treatment service, with the same counterfactual and allocation rules. Avoided fossil production and captured carbon must not receive duplicate credits. Table 6 applies this boundary to platform selection; missing inventories are retained as gaps rather than assigned zero burdens.

Table 6. Common resource requirements, validation evidence, and deployment limits of microbial platforms.

Platform and Target	Resource and Feed Requirements	Scale and Operating Evidence	Main Implementation Limits	Ref.
Microalgae: biomass or treatment	Natural light and illuminated area; local nutrients; cooling, pH control, and contaminant removal	Laboratory/outdoor comparisons; 6000 L pilot; one-year biofilm operation	Annual productivity, land, and harvesting/dewatering; contaminant carryover can restrict product use	[10,13,70,73]
Engineered cyanobacteria: metabolites	Light, inorganic carbon, and compatible medium; no external H2	Laboratory secretion and isoprene production; 21 d butanediol experiment	Construct stability, product concentration, and recovery over repeated cultivation cycles	[64,65,75]
Acetogens: acetate or solvents	No illumination; low-emission H2–CO2 and anaerobic gas handling; CO feeds require a separate balance	Controlled and continuous retention studies; an industrial solvent pilot used CO-containing gas	Gas transfer, reductant demand, inhibition, and acid/solvent recovery	[26,34,61]
HOB: protein or PHB	No illumination; low-emission H2 and controlled CO2–O2; include electrolyzer and energy-generation land	Laboratory PHB cultivation; atmospheric pressure 100 L protein pilot	Explosion protection, energy per saleable product, biomass separation, and product quality	[35,92,107]
MES: acetate or other chemicals	No illumination; electricity, CO2 delivery, and management of cathodic H2	Continuous reactors and integrated electrodialysis; these cases do not establish full-scale durability	Cell voltage, target product selectivity, extraction power, and electrode/membrane life	[94,96]

Authors’ synthesis under the common boundary in Section 10.1. Scale and operating duration describe the cited configurations, not an entire platform. Land includes energy generation where relevant. Comparison requires the same product or treatment function and counterfactual; incomplete inventories preclude a common net-emission ranking.

Electron accounting in H₂-fed and electrochemical processes should include target products, biomass, byproducts, and unconsumed H₂. The selectivity contrast in Section 8.3 [94] illustrates why aggregate electron recovery is insufficient for product valuation. Energy inventories should include gas treatment and delivery, mixing, temperature control, illumination, electrolysis, and recovery.

10.2. Energy Demand and Illustrative Cost Cases

Equation (1) yields a stoichiometric minimum of approximately 0.092 kg H₂ per kilogram of CO₂ converted to acetic acid. At an explicitly assumed 50 kWh·kg⁻¹ H₂, the electrolysis-only minimum is 4.6 kWh·kg⁻¹ CO₂, costing EUR 0.23–0.46 at illustrative electricity prices of EUR 0.05–0.10·kWh⁻¹. Alternatively, purchasing H₂ at an assumed EUR 2–6·kg⁻¹ yields a reductant cost of EUR 0.18–0.55·kg⁻¹ CO₂; this replaces the electrolysis electricity term. These are sensitivity calculations, not market quotations or plant costs. Actual gas losses, biomass formation, compression, and recovery increase the requirements.

The 2020 MES assessment [18] provides a model-based case rather than a plant energy measurement. It includes reactor assets, electricity, medium, and CO₂, initially without product separation; recovery is assessed in a separate purification scenario. No external H₂ production unit is required in this electrochemical configuration. Electricity costs were varied from EUR 0.03 to 0.09·kWh⁻¹; the base scenario became profitable near EUR 0.045·kWh⁻¹ or below. Product use lay outside the model boundary. This threshold depends on the modeled performance and cannot be generalized to present-day MES installations.

Historical costs provide context rather than a common price scale [126]. A 2016 model estimated EUR 3.4·kg⁻¹ TS for cultivation and projected EUR 0.5·kg⁻¹ after ten years; the latter was not a measured outcome [127]. A 2017 floc-based assessment reported EUR 0.25–0.50·m⁻³ wastewater treated and EUR 5.29–8.07·kg⁻¹ TSS [79]. Aquafeed assessment also depends on lighting and product value [128]. Different cost years, functions, and boundaries prevent direct price ranking [129].

A newer algal model [82] covers inoculum production, raceway cultivation, and dewatering. The minimum biomass selling prices were USD 1130–1200·t⁻¹ for semi-continuous operation and USD 1380–2040·t⁻¹ for batch operation. The semi-continuous advantage disappeared at a modeled mean time to culture failure of 26 d; batch cultivation reached comparable costs at productivities above 28 g·m⁻²·d⁻¹. This strengthens the emphasis on culture reliability and seed-train design; the prices exclude subsequent fuel conversion and are not costs per tonne of CO₂ removed.

A pilot-based algal fuel assessment [130] followed direct flue gas feeding through PBR cultivation, flotation and centrifugation, hydrothermal liquefaction, H₂-assisted upgrading, and fuel use. Reactor, compressor, and LED electricity consumption was 0.93 kWh·kg⁻¹ dry algae, excluding building climate control; upgrading H₂ was derived from natural gas reforming with CO₂ capture. At the model electricity price of USD 69·MWh⁻¹, the fuel costs were USD 12–16 per gasoline gallon equivalent. The net emissions were approximately 68 g CO₂-equivalent·MJ⁻¹ for fossil CO₂ and −4 for biogenic CO₂. These scenarios included source and displacement credits but excluded construction burdens. The negative result is conditional on those boundaries [131]. Protein assessments likewise require the appropriate product function and the published correction to earlier calculations [132]. LCA and TEA should therefore be interpreted for a defined configuration [133].

10.3. Carbon Fate and Remaining Inventory Gaps

Biomineralization adds a mineral carbon output to biomass production. N. oceanica IMET1 was tested at 1 and 500 L with 10 and 5% CO₂, respectively, and 0.02 mol·L⁻¹ NaHCO₃. At 1 L, biomass reached 1.52 g TS·L⁻¹ on day 12, with an additional carbonate-associated capture of 9.3%. At 500 L, additional cell retention reached 60% on day 40, while the maximum reported capture rate was 63.2 g CO₂·m⁻²·d⁻¹ on day 35 [81]. CaCl₂-assisted Chlorella HS2 cultivation increased the cumulative biomass production 4.67-fold and the source fixation metric 10.9-fold [83]. Net retention additionally requires reagent origin and alkalinity accounting, phase stability, and end-of-life evidence; dissolved carbon inputs cannot be credited again as newly captured gaseous CO₂.

11. Deployment Barriers and Development Prospects

Deployment requires matching the platform to the feedstock, energy, land, and product demand [134]. Waste CO₂ integration may combine treatment and production value [79], but the relevant comparator must provide the same function [6]. Table 6 translates the evidence into engineering selection criteria under the system boundary defined in Section 10.1.

The scale constraint on photosynthesis can be estimated from published productivities. For a hypothetical stream of 1 t CO₂·h⁻¹, a carbon fraction of 50%, and a productivity of 19–24 g TS·m⁻²·d⁻¹ from the outdoor vertical configuration experiment [10], Equation (3) yields approximately 55–69 ha of productive area to fix the entire stream. This calculation excludes service spacing and seasonal deterioration. The productivity range reflects only the experimental period; it is not a forecast of annual plant performance. Deployment may be justified for part of the CO₂ stream if cultivation also treats a liquid stream or produces a product of sufficient value to cover the inputs.

Integration with waste streams requires contaminant control and product specifications [135]. The industrial concentrate pilot [73] supports a combined treatment function; desulfurization [13] supports stable cultivation. Neither observation alone demonstrates that biomass meets food or feed specifications because removed contaminants may enter the product.

For HOB, energy-generation land and electrolysis capacity must be included alongside the compact fermenter [107]. Safe gas delivery, biomass recovery, and quality assurance also contribute to the equipment and operating requirements.

For acetogens, the central decisions concern H₂ delivery, retention, inhibition, and acid or solvent recovery [34,61]. New MES rate benchmarks strengthen the case for process intensification, but electricity consumption and extraction remain decisive. Selecting MES over separate electrolysis and fermentation requires a demonstrated whole-process advantage [18,94].

Technology selection should first test light, land, and low-emission reductant availability, then gas compatibility and the intended product. Productivity, recovery, stability, and cost should subsequently be compared within that application. Table 6 provides this sequence without assigning a universal platform ranking.

Laboratory results establish biological feasibility or short-term rate potential; pilot demonstrations establish operation at a stated scale and feed. The 6000 L algal [73] and 100 L HOB [92] pilots demonstrate scale, while one-year biofilm monitoring [70] addresses seasonal operation. Approximately 450 d for archaeal methanation [125] remains a reactor stability comparator. Neither scale nor continuous operation alone establishes industrial durability. Further validation should test feed variability, shutdown/restart, material life, and product quality. Figure 7 links these requirements to specific installations.

Figure 7. Evidence required to assess the deployment potential of biological CO₂ conversion.

12. Conclusions

Microbial CO₂ utilization should be selected for a defined site, feedstock, and product. Practical advantage depends on matching biological conversion to energy supply, recovery, and market requirements.

Microalgae and cyanobacteria suit sites with natural light, land, and local nutrients. Annual net productivity per occupied area, seasonality, harvesting, and contaminant transfer govern scale-up. CO₂ tolerance does not establish tolerance to the complete flue gas; desulfurization, pH control, and biomass quality remain important for integration with waste streams.

Acetogens enable anaerobic chemical production with reliable reductant supply and effective product recovery. Cell retention does not eliminate product inhibition or gas transfer limitations. Results from CO-containing feeds require separate validation for H₂–CO₂ operation because their carbon and electron balances differ.

HOB support protein and polymer production with low-emission H₂, safe gas feeding, and verified product quality. Protein-rich biomass requires nutritional and purity assessment. Compact fermenters must be assessed together with electrolysis, gas recirculation, reductant losses, and biomass separation.

Engineered cyanobacteria require stable product formation without continuous selection and effective recovery. While recent MES benchmarks improve rate feasibility, selectivity, electricity demand, electrode/membrane life, and extraction determine whole-process advantage.

Climate assessment must include CO₂ origin, energy and reductant production, conditioning, cultivation, recovery, equipment, and product fate against a functionally equivalent reference. Biomass and biodegradable products generally retain carbon temporarily. Durable removal requires verified storage. Biomineralization also requires accounting for bicarbonate and alkalinity origin, reagent consumption, and mineral stability.

The priority evidence for technological readiness comprises closed mass and energy balances, sustained operation on real gases, recovery after interruptions, material durability, and product quality. Economic assessments should test energy prices, emission intensity, plant availability, and separation duties, with missing inventories stated explicitly.