Side-by-side guides for the decisions bioprocess engineers actually face. Every comparison includes a decision matrix, use-case fit, vendor landscape, and cost analysis. No paid placements — recommendations are based on published specifications and deployment patterns.
Choose the right in-line, at-line, or non-invasive sensor for your scale and modality. Covers dissolved oxygen, pH, biomass (capacitance vs optical), and process-analytical spectroscopy.
The measurement-principle decision. Optical has replaced polarographic as the cGMP default — no electrolyte, no warm-up, no O₂ consumption, better CIP/SIP durability. Polarographic still wins on response time and capital cost for high-OUR microbial.
Sensors · pHThe pH sensor measurement-principle decision. Glass electrodes win on accuracy (0.044 vs 0.072 pH discrepancy in the Fratz-Berilla 2024 head-to-head) and dynamic range; optical wins where the vessel is single-use because no glass-electrode form factor can be pre-integrated into a sealed gamma-sterilised bag.
Sensors · CO2The dissolved CO2 sensor measurement-principle decision. Severinghaus electrodes (Mettler InPro 5000i) own the cGMP installed base and the wider 0–1000 mbar range; solid-state mid-IR optical (Hamilton CO2NTROL) eliminates electrolyte refills for new stainless-steel builds; HPTS-dye optical (PreSens) is the only realistic option in single-use bags.
Sensors · Glucose · PATThe glucose monitoring placement decision. Inline Raman (Resolution Spectra ProCellics, Endress+Hauser Kaiser, Tornado HyperFlux) drives a 15-minute closed-loop feedback into the bioreactor controller; at-line YSI 2900D / Nova BioProfile Flex2 / Cedex Bio HT delivers ±2% reference accuracy in 60 seconds per aliquot. CHO commercial mAb runs both — Raman for control, YSI for the daily reference check.
Sensors · DO · VendorThe form-factor decision for optical DO. VisiFerm is an insertable probe with integrated Arc transmitter for reusable stainless steel vessels; PreSens is a gamma-sterilised spot glued inside single-use bags. Same physics, different vessels.
Sensors · Platform · VendorThe digital-platform decision one level up from the individual probe choice. Arc puts the transmitter inside the probe head and speaks Modbus RTU straight to the DCS; ISM keeps a separate M400 or M800 transmitter and exposes 4-20 mA / HART / PROFIBUS PA. ArcAir Bluetooth calibration vs iSense PC asset management. Coverage across DO, pH, biomass, and CO2.
Sensors · DO · VesselThe vessel-driven sensor decision. Single-use DO spots come pre-integrated in Sartorius, Thermo, Cytiva bags; reusable probes dominate stainless steel. Scope explicitly excludes pulse oximeters and automotive sensors.
Sensors · BiomassCapacitance (Aber FUTURA, Hamilton INCYTE) measures viable cell volume via RF impedance — the cGMP standard for mAb fed-batch feed control. Optical turbidity is cheaper but can't distinguish viable from dead cells.
Sensors · PATThe PAT analyzer decision. Raman wins on glucose, lactate, antibody in aqueous culture (the Paik 2018 parallel comparison); NIR wins on water content, glutamine, ammonium, and UF/DF monitoring.
Upstream operating mode and facility architecture decisions. Covers batch vs fed-batch vs perfusion, single-use vs stainless steel, and their economic crossover points.
The facility-architecture decision that determines your sensor choice, cleaning validation scope, and CapEx. Crossover point typically 2,000-5,000 L working volume depending on batch frequency and product lifetime.
SUB · VendorTriadic comparison of the three market-leading single-use stirred-tank bioreactor platforms. All three deliver equivalent mixing (<30 s) and kLa (>10 h−¹) at 50–2000 L, so the decision is CDMO installed base (Biostat STR at Samsung/WuXi/BI), scale ceiling (only HyPerforma DynaDrive reaches 5000 L with 12–20:1 turndown), or downstream integration (Xcellerex anchors Cytiva ReadyToProcess X-platform).
Bench · VendorTriadic comparison of bench and PD bioreactor families. Applikon (Getinge) is single-vessel glass/SU on one BioXpert stack (150 mL – 20 L); Eppendorf covers 250 mL – 40 L with 1–8 vessels on BioFlo 320 and rigid-wall BioBLU single-use vessels; Sartorius Ambr is the high-throughput arm (24–48 parallel vessels, robotic liquid handling) at 5–10× the capital per unit. Peer-reviewed Ambr scale-down literature (Nienow, Xu, Manahan) confirms equivalence to 5 L bench and 2,000 L Biostat STR when kLa is matched.
Process ModeThe upstream operating-mode decision. Volumetric productivity data, COGS analysis, hybrid intensified approaches for mAb. Triadic comparison with real-world case study data.
EconomicsThe manufacturing-architecture cost comparison. Capital intensity, operational complexity, and regulatory risk traded off across batch and continuous biomanufacturing approaches.
Perfusion · Cell RetentionThe cell retention device decision for perfusion bioreactors. ATF (Repligen XCell ATF) holds mAb sieving at 88-95% across full runs thanks to alternating diaphragm flow; peristaltic-pump TFF decays 95% to 52% from cell-lysis fouling, but a centrifugal-pump TFF (Levitronix) closes the gap and integrates directly into continuous downstream.
Process IntensificationThe process intensification decision for CHO mAb fed-batch. Intensified fed-batch (high inoculation density 5-20 million cells/mL via N-1 perfusion seed) roughly doubles titer in the same N-stage vessel — 8-12 g/L vs 4-6 g/L, in a 10-12 day run vs 14-15 day standard. Capex of $300-800k per seed train ATF/TFF retrofit. Decision matrix for capacity-constrained mAb, greenfield ≥500 kg/yr, and clinical-to-commercial scale-up.
Viral vector and gene therapy manufacturing platform decisions. Producer cell line versus transient transfection choices for AAV and lentivirus programmes.
The next CAR-T fork — autologous and allogeneic are both ex vivo. In vivo dosing (targeted LNP with CAR mRNA, or T-cell-pseudotyped lentivirus) replaces 3–6 week manufacturing with a single IV. Covers Capstan CPTX2309 (Lilly, CD8-tLNP + CD19 CAR), Umoja VivoVec, Interius INT2104, and Orna. Projected COGS $2–10k per dose vs $100–150k autologous; tLNP-mRNA CAR is transient (days–weeks) and repeat-dosable; lentiviral in vivo mirrors ex vivo persistence. Best fit: autoimmune (lupus, MG, SSc, RA) for tLNP, oncology at scale for lentiviral. Zero approvals as of Aug 2026.
CAR-T · ManufacturingThe cell therapy manufacturing architecture decision. Autologous is the only approved architecture as of July 2026 (7 products) and delivers months-to-years persistence, but costs $100–150k COGS per dose and takes 3–6 weeks vein-to-vein. Allogeneic amortises a single donor batch across 100–1,000 doses for ~$5–20k projected COGS and 5–7 day time-to-infusion, but requires TRAC/B2M/CD52 gene edits and shorter 1–3 month persistence. Covers ALPHA3 pivotal readout, gene-edit toolbox, iPSC platforms, and the four-scenario decision matrix.
Viral Vector · PlatformThe vector-selection decision for gene therapy. AAV wins for in vivo delivery to non-dividing tissue with transgenes under 4.7 kb (Zolgensma, Luxturna, Elevidys, Hemgenix); lentivirus wins for ex vivo modification of dividing cells (Kymriah, Yescarta, Zynteglo, Skysona) and larger 8 kb payloads. Covers particle size, integration, yield (1E11-1E14 vg/mL vs 1E7-1E10 TU/mL), bioreactor scale (200-2000 L vs 50-500 L capped by VSV-G toxicity), cold chain, and per-batch cost (~$2M vs ~$1.5M). Vendor landscape covers Charles River, Cytiva, Lonza, AGC, Oxford Biomedica, WuXi, MilliporeSigma, Yposkesi.
AAV · PlatformThe AAV manufacturing platform decision. Transient wins on speed (weeks to first batch, any serotype). Stable wins on cost and consistency once annual demand exceeds ~1E17 vg — eliminates the ~$100k/g GMP plasmid spend that drives ~40% of transient AAV cost. Covers Cytiva ELEVECTA, Asimov AAV Edge (6E15 vg/L), Lonza and Charles River nAAVigation timelines (5-9 months), batch CV, full:empty capsid ratio, and 3-year TCO at 20 batches/year.
Production format decisions upstream of platform choice. Titer, downstream complexity, cost per gram, and approved-product landscape for antibody-based biologics.
Capture and polishing resin choices that drive per-gram cost of goods. Dynamic binding capacity, alkaline stability, ligand chemistry, and base matrix tradeoffs for mAb purification.
The 4-way Protein A capture resin decision. MabSelect PrismA leads on raw DBC (77-80 mg/mL at 6 min) and 0.5 M NaOH lifetime (~90% retained at 150 cycles); Toyopearl AF-rProtein A HC-650F is the strongest methacrylate alternative (>65 g/L, 200+ cycles at 0.1 M NaOH); Amsphere A3 is the credible second-source to Cytiva; MabSelect SuRe LX remains the legacy commercial baseline.
Downstream · ClarificationThe primary clarification decision for CHO mAb, E. coli, and AAV harvest. Depth filtration alone wins below 2,000 L on capex and gentleness (95-99% recovery, no CIP); a disc-stack centrifuge (BTPX, Culturefuge, GEA) plus depth filter polish wins above 10,000 L on per-batch consumables. Includes 3-year TCO, scale break-points (60-150 L/m2 depth filter ceiling), and modality fit for shear-sensitive viral vectors.
Downstream · ChromatographyThe capture-step efficiency decision. Continuous (PCC, SMB, MCSGP) delivers 2-4x productivity and 30-60% less resin at 2-3x the capex; batch single-column wins on flexibility, validation maturity, and clinical-scale economics. Covers ICH Q13 regulatory landscape, perfusion pairing, and a 3-year TCO at 1000 L, 4 g/L mAb. Vendors: Cytiva AKTA pcc, Sartorius BioSMB, YMC Contichrom CUBE 30.
Downstream · TFF · FormulationThe TFF mode-selection decision. UF and DF are two operational modes of the same cassette: UF concentrates by pulling permeate without replacement; DF holds volume constant by adding fresh buffer at the permeate rate. Covers the UF→DF→UF sequence, diavolume math (C/C0 = exp(-N)), Donnan effects at high concentration, and a 3-year cost model for 1,000 L mAb final formulation. Vendors: Pellicon, Sartocon, KrosFlo, Cadence.
Regulatory and compendial framework choices that decide how single-use components, cleaning, and container-closure systems appear in your filing.
Promoter, vector, and host-strain decisions upstream of every recombinant protein run. Strength, leakiness, copy-number, and host-strain compatibility tradeoffs.
Carbon footprint, life cycle assessment and process mass intensity for biomanufacturing. Which tooling you actually need, and what it costs.
Commercial suites run €6,000–12,000 per seat per year, and the licensed background database rather than the software is what you are paying for. Screen first to find the hotspot; buy a licence when the number has to be externally verified. None of them ships a bioprocess model.
Sustainability · LCA methodOne asks what share of emissions is yours, the other what emissions change because you exist. On one 2,000 L batch the method is worth 5%, but it decides whether a grid-region choice is worth 5.1× or nothing at all. When to use each, and how to build a defensible marginal factor.
Sustainability · FacilityModelled at matched scale and titer, the two land within 0.2% of each other: 3,980 vs 3,971 kg CO2e/kg. The plastic burden and the cleaning burden nearly cancel, the answer flips at a grid intensity of 479 g CO2e/kWh, and 89% of both footprints is the cleanroom they share.
The Bioreactor Sensor Selection Tool takes 6 questions about scale, modality, vessel, parameters, environment, and budget — and returns ranked sensor recommendations with vendor links. Covers every comparison above, plus CO₂, temperature, and glucose.