Engineering Guide · Vendor-Neutral

Wave-Mixed vs Stirred-Tank Bioreactors for Viral Vector: Which Mixing Platform Should You Use?

Wave-mixed vs stirred-tank bioreactor side-by-side comparison for viral vector manufacturing Wave-mixed Rocking bag · no impeller · low shear kLa 5-30 h⁻¹ · up to ~500 L VS Stirred-tank Impeller-driven · high mass transfer kLa 30-150+ h⁻¹ · up to 5,000 L
Figure 1: Wave-mixed bioreactors use a rocking platform to agitate a pillow bag with no internal mechanics; stirred-tank bioreactors use one or more impellers with baffles and a sparger. Same viral-vector cell lines, very different shear environments.
Quick Verdict

For viral vector manufacturing, wave-mixed dominates seed-train and clinical-scale batches under 200 L, especially for shear-sensitive VSV-G-pseudotyped lentivirus and insect-cell baculovirus. Stirred-tank single-use bioreactors take over above 500 L and are the default choice for commercial AAV production on stable HEK293 lines where high kLa and predictable scale-up matter more than gentle mixing.

Key differences at a glance

Side-by-side comparison

Factor Wave-mixed Stirred-tank
Mixing mechanism Rocking platform tilts sealed pillow bag; wave induces mixing at surface Rotating impeller(s) on a shaft; baffles disrupt tangential flow
Maximum shear rate ~10-100 s⁻¹ (gentle) ~500-5,000 s⁻¹ (impeller-dependent)
Typical kLa 5-30 h⁻¹ 30-150+ h⁻¹
Mixing time (θ95) 30-120 s 10-30 s
Vessel format Single-use only (gamma-sterilized bag) Single-use bag OR reusable stainless steel
Working volume range 0.3-300 L (WAVE 500 ceiling) 1-5,000 L
Instrument capital cost $30k-$150k (controller + platform) $250k-$600k+ (SUB + controller at 50-500 L)
Consumable per batch $200-$1,500 (bag + tubing set) $600-$5,000 (SUB bag + impeller assembly)
Integration with PAT Optical DO / pH via bag spots; limited port count Full port complement, Raman/NIR, capacitance

Values reflect typical published specifications for viral-vector applications. Your vendor's current datasheet takes precedence.

Wave-mixed bioreactors in detail

Wave-mixed bioreactors are single-use, sealed pillow bags that sit on a rocking platform. The rocker tilts back and forth at a defined angle (typically 6-12°) and rate (10-40 rocks/min), which sloshes the liquid inside the bag into a standing wave. The wave carries oxygen from the headspace into the liquid and homogenizes the bulk. There is no impeller, no seal, no baffle — nothing rotates inside the vessel and nothing enters the sterile boundary except the sparge/vent gas line, feed/harvest tubing, and any bag-integrated sensor spots.

The technology was commercialized by Cytiva's WAVE Bioreactor line in the late 1990s and is now offered by Sartorius Biostat RM, Thermo HyPerforma Rocker, Applikon AppliFlex ST, and several smaller vendors. Cytiva's WAVE 25 (10-25 L working volume) and Sartorius Biostat RM 20 are the most common bench-scale units in cell and gene therapy labs.

How it works

A pre-sterilized bag pre-fitted with optical DO and pH sensor spots is clamped onto the rocking platform, inflated with a controlled headspace of air/CO₂/O₂, and filled to 30-50% of nominal volume with medium and cells. The controller sets rocking angle, rocking rate, headspace gas composition and flow, and sparge gas if the bag has a microsparger. As the platform tilts, liquid rushes to the low end, breaks against the far bag wall, and returns — a standing wave that presents a large air/liquid interface for oxygen transfer with almost no shear on suspended cells. The maximum shear rate at the wall is typically 10-100 s⁻¹, one to two orders of magnitude below a Rushton or elephant-ear impeller running at 100-300 rpm.

Oxygen transfer scales with rocking rate, angle, headspace overlay flow, and fill volume. A well-tuned WAVE 25 at 25 rocks/min, 8° angle, and 30% fill delivers kLa of 15-25 h⁻¹ — enough for HEK293 at densities up to ~4-6 million cells/mL, and enough for insect cells (Sf9, Hi5) up to ~8 million cells/mL. Above those densities, oxygen becomes the limit and process teams either switch to microsparging (which reintroduces bubble shear) or transfer to a stirred-tank platform.

When wave-mixed wins

Wave-mixed wins whenever the product or cell line is shear-sensitive. Lentivirus pseudotyped with the VSV-G glycoprotein is the canonical example: functional titer drops 50-80% when impeller tip speed exceeds ~1.0-1.5 m/s in a stirred tank because shear peels VSV-G off the envelope and destroys infectivity. Wave-mixed platforms preserve envelope integrity because there is no impeller and no local shear peak. Baculovirus in Sf9 or Hi5 insect cells is similarly favored — insect cells lack a cell wall and cleave at moderate shear, and Sf-cell-based AAV programs on the Sf9/rBV platform (Voyager, uniQure) commonly use wave-mixed platforms for the primary batch. T-cell expansion for CAR-T is the third pillar: Kymriah, Yescarta, and Breyanzi all specify wave-mixed expansion in published GMP protocols.

Wave-mixed also wins on capital cost, cleanroom footprint, and turnaround time. A WAVE 25 platform occupies ~1 m² of bench, needs no CIP/SIP infrastructure, and a bag changeover from batch N to batch N+1 takes 30-60 min instead of 8-24 h for a stainless-steel clean cycle. That matters when a clinical AAV or lentivirus program is running 15-30 batches/year in a shared cleanroom.

Stirred-tank bioreactors in detail

Stirred-tank bioreactors are the workhorse of biopharma. A cylindrical vessel with a height:diameter ratio of typically 2:1 or 3:1 holds the culture; one or two impellers on a central shaft agitate the broth; four vertical baffles disrupt tangential flow and force axial/radial mixing; a sparger at the bottom introduces oxygen as bubbles. The geometry and the impeller-baffle physics are the same whether the vessel is 5 L or 5,000 L, which is why stirred-tank scale-up is well characterized and stirred-tank platforms dominate commercial manufacturing.

For viral vector work the relevant options are single-use stirred tanks (SUBs). The three market-leading platforms — Sartorius Biostat STR, Thermo HyPerforma S.U.B., and Cytiva Xcellerex XDR — all deliver equivalent mixing (<30 s) and kLa (>10 h⁻¹) at 50-2,000 L and dominate the commercial AAV and mAb single-use install base. Traditional reusable stainless-steel STRs from Getinge (Applikon) and Eppendorf BioFlo still cover the bench range (0.5-40 L) for process development.

How it works

Impellers are chosen for the process. For low-shear mammalian and viral-vector work the two dominant options are pitched-blade (elephant-ear) and marine impellers — both deliver primarily axial flow with a modest radial component and keep tip-speed shear at 0.5-1.2 m/s at operational rpm. For higher-density fed-batch or microbial work, Rushton disc turbines drive stronger radial flow and higher kLa but with tip speeds of 1.5-2.5 m/s. Impeller diameter is typically 0.3-0.5 of vessel diameter, and impeller spacing on multi-impeller shafts is typically 1.0-1.5× impeller diameter.

Sparging is the second lever. A ring sparger under the lowest impeller introduces gas as bubbles; the impeller shears the bubbles into smaller ones, increasing interfacial area and driving kLa. For viral vector work the sparger is usually a microsparger (10-100 µm pore) to keep bubble size small and reduce local shear during bubble breakup. DO control is fed forward through impeller speed, gas flow, and O₂ fraction; pH is controlled by CO₂ addition (acid side) and NaHCO₃ or NaOH (base side).

When stirred-tank wins

Stirred-tank wins on scale, mass transfer, and PAT integration. Above 500 L there is no wave-mixed option — the WAVE 500 with 300 L working is the ceiling and every commercial-scale AAV program (>500 L, e.g. Novartis Zolgensma, Sarepta Elevidys, Roche Luxturna) runs stirred-tank. Stirred-tank kLa reaches 100+ h⁻¹ at production impeller speeds, which is what a stable HEK293 producer cell line at 8-12 million cells/mL needs for adequate O₂ supply. Full port complement lets you install Hamilton VisiFerm DO, Mettler InPro pH, Aber FUTURA capacitance, and Raman probes on the same vessel, which is difficult in a wave bag with its 4-6 port count.

Stirred-tank also wins when a program is moving from clinical to commercial and needs a validated scale-up path. The step from a 200 L Biostat STR to a 2,000 L Biostat STR is a well-characterized geometric scale-up: preserve P/V and impeller tip speed, adjust sparging to maintain kLa, and expect <10% shift in productivity. From a 20 L WAVE to a 200 L stirred-tank there is no geometric relationship — you re-optimize sparging, impeller speed, and DO setpoint, and CMC groups typically budget 3-6 months of tech transfer.

Pros and cons

Wave-mixed

Advantages

  • Lowest shear of any bioreactor format — preserves VSV-G envelope, insect-cell integrity, T-cell viability.
  • Fully closed, gamma-sterilized single-use bag. No CIP/SIP infrastructure, no sterile boundary breach on batch changeover.
  • Capital cost 3-5× lower than an equivalent-scale SUB. Fits the CGT clinical-scale budget.
  • Fast changeover: 30-60 min between batches vs 8-24 h for stainless CIP.

Disadvantages

  • Scale ceiling ~300 L working volume. Commercial programs above this must transfer platforms.
  • Limited kLa (5-30 h⁻¹) constrains cell density and productivity. Microsparging helps but reintroduces shear.
  • Fewer sensor ports; retrofitting PAT (Raman, capacitance) is difficult in a sealed bag.
  • Mixing time (30-120 s) is 3-5× slower than STR, making fast pH corrections and rapid feed additions harder.

Stirred-tank

Advantages

  • Scales from 5 L to 5,000 L with well-characterized geometric scale-up rules (P/V, tip speed, kLa).
  • High kLa (30-150+ h⁻¹) supports HEK293 densities up to 12-15 million cells/mL for productive stable-line AAV.
  • Full port complement for PAT, feed lines, sampling — the default platform for QbD control strategies.
  • Available in single-use (SUB) and reusable stainless — flexibility for clinical-through-commercial programs.

Disadvantages

  • Impeller shear can damage shear-sensitive viral vectors. VSV-G lentivirus titer drops 50-80% above tip speed ~1.0-1.5 m/s.
  • Capital cost 3-5× higher than equivalent-volume wave-mixed. 200 L Biostat STR ≈ $400k vs Biostat RM 200 ≈ $80k.
  • Larger footprint. A 200 L SUB with utilities needs ~10 m² of cleanroom vs ~3 m² for a Biostat RM 200.
  • Reusable stainless requires CIP/SIP and clean-hold validation — significant regulatory scope and turnaround time.

Which should you choose?

Pick based on the modality, the scale, and the shear tolerance of the vector. For viral vector programs the three deciding constraints are almost always: how shear-sensitive is the enveloped particle, what production scale does the clinical or commercial demand require, and where does the seed train fit in the overall train.

VSV-G lentivirus, any scale

Shear on the envelope destroys titer. Wave-mixed preserves 2-3× more functional particles than a stirred-tank at equivalent nominal production. Bluebird, Novartis, and academic vector cores default to wave-mixed for GMP clinical batches.

Choose Wave-mixed

Commercial AAV, ≥500 L

Wave doesn't scale here — WAVE 500 tops out at 300 L working. Stable HEK293 producer lines at 8-12M cells/mL need kLa >50 h⁻¹, which only a stirred-tank delivers. Every commercial in-vivo AAV program runs stirred-tank at production scale.

Choose Stirred-tank

Preclinical & early clinical, budget-constrained

Below 50 L, wave-mixed capital wins outright ($40k WAVE 25 vs $250k Biostat STR 50). Faster changeover shortens batch cycles. Move to stirred-tank when clinical demand pushes above 50-100 L or Phase 3 requires the commercial-transfer path.

Choose Wave-mixed

Seed train for large-scale AAV

A hybrid train is common: wave-mixed 10-50 L for seed expansion (gentle, single-use, fast turnaround), then transfer to a stirred-tank at production scale (200-2,000 L). Preserves gentle expansion where it matters and productive volume where it matters.

Choose Wave-mixed → Stirred-tank

Real-world use cases

Typical setups where CGT teams have converged on one platform or the other.

Lentivirus, 200 L clinical
Wave-mixed WAVE 200 for VSV-G lentivirus

HEK293T triple-transfection in 50-100 L working volume, 4-6 million cells/mL at harvest. Wave preserves VSV-G functional titer at 10⁶-10⁷ TU/mL. Kymriah and Yescarta GMP master batch records specify wave-mixed at this scale.

AAV commercial, 2000 L
Stirred-tank Biostat STR 2000 for stable HEK293 AAV

Stable ELEVECTA producer line or transient triple transfection at 1,500-2,000 L working volume. Elephant-ear impeller, microsparger, capacitance-controlled feed. Titers 5×10¹³-2×10¹⁴ vg/L. Novartis, Sarepta, and CDMOs (Charles River, Lonza) at this scale.

Baculovirus/Sf9, 100 L
Wave-mixed Biostat RM 100 for AAV Sf9/rBV

Insect-cell-based AAV programs (Voyager, uniQure) use wave-mixed at 20-100 L to protect Sf9 viability during baculovirus infection. Higher-density Sf9 needs microsparging, which some teams achieve with a Sartorius Biostat RM microsparge bag.

CAR-T seed, 10 L
Cytiva Xuri W25 for T-cell expansion

Autologous CAR-T (Kymriah, Yescarta, Breyanzi) uses Cytiva Xuri Cell Expansion System (a wave-mixed W25 tuned for T cells) for 8-10-day expansion from 10⁸ to 10¹⁰-10¹¹ T cells. Closed, single-patient, gamma-sterilized bag.

Not sure which sensor set fits your wave-mixed or stirred-tank vessel?

Answer six questions about your modality, scale, and shear tolerance and get a ranked shortlist of DO, pH, biomass, and PAT sensors that match the vessel format. Covers optical spots for wave bags and inline probes for stirred tanks.

Open the Sensor Selection Tool

Cost and lifecycle considerations

Total cost of ownership includes three components

Capital cost (platform + controller) + recurring consumables (bag or SUB per batch, tubing sets, filters, sensor spots) + indirect costs (facility footprint, changeover labor, tech transfer between scales). Wave-mixed wins on capital and cleanroom footprint; stirred-tank wins on cost per gram of vector once scale exceeds ~200 L.

Capital economics favor wave-mixed at bench and small clinical scale. A Cytiva Xuri W25 or Sartorius Biostat RM 20 platform lands in the $40,000-$80,000 range with controller and DO/pH monitoring. An equivalent-volume stirred tank — a 20 L Sartorius Biostat B-DCU or a 15 L Applikon single-use — sits at $80,000-$150,000. At 200 L, the cost gap widens further: a WAVE 200 or Biostat RM 200 is ~$100,000-$150,000; a Biostat STR 200 or HyPerforma 200 is $350,000-$500,000.

Per-batch consumables reverse partially at scale. A 25 L wave bag with tubing set is $250-$500 (Cytiva ReadyToProcess or Sartorius Flexsafe RM). A 200 L Biostat STR SUB with impeller assembly is $2,000-$4,000. On a per-liter-of-vector basis, though, stirred-tank still wins at scale because a 2,000 L STR delivers 4-8× the batch yield of the 300 L wave ceiling per operator-shift. For clinical batches under 100 L, wave-mixed usually delivers a lower total-cost-per-batch; for commercial batches above 500 L, stirred-tank wins on cost per gram of vector.

Cost component Wave-mixed (200 L) Stirred-tank (200 L SUB)
Platform + controller$100k-$150k$350k-$500k
Consumables per batch$400-$1,500$2,000-$4,000
Changeover labor per batch0.5-1.0 h1.5-2.5 h
3-year TCO (20 batches/yr)~$300k~$650k

Vendor landscape

Major vendors in each camp, with one-line positioning notes.

Wave-mixed vendors

Stirred-tank single-use vendors

Frequently asked questions

What is the difference between wave-mixed and stirred-tank bioreactors?
Wave-mixed bioreactors use a rocking platform to agitate a single-use pillow bag with no impeller inside, generating gentle wave-induced mixing with kLa typically 5-30 h⁻¹. Stirred-tank bioreactors use one or more mechanical impellers in a cylindrical vessel to drive mixing and mass transfer, achieving kLa 30-150+ h⁻¹. Wave-mixed excels for shear-sensitive cells and volumes up to ~500 L; stirred-tank dominates production scale from 200 L to 5,000 L.
Are wave-mixed bioreactors better for viral vector production?
For lentivirus production, wave-mixed is often preferred because VSV-G pseudotyped particles are shear-sensitive and titer drops sharply above impeller tip speeds of ~1.0-1.5 m/s. For AAV, both platforms work: wave-mixed dominates clinical batches under 200 L and lentiviral packaging, while stirred-tank scales up to 2,000 L for commercial AAV programs on stable HEK293 lines. The rule of thumb is enveloped viruses favor wave; capsid-based vectors like AAV work in both.
What is the maximum scale of a wave bioreactor?
Cytiva WAVE 500 tops out at 300 L working volume in a 500 L pillow bag, and Sartorius Biostat RM 200 reaches 100 L working volume in a 200 L bag. Above 300 L, wave-mixed platforms lose kLa headroom and vendors themselves recommend switching to stirred-tank single-use bioreactors like Sartorius Biostat STR, Thermo HyPerforma S.U.B., or Cytiva Xcellerex XDR.
Why do lentivirus producers prefer wave-mixed bioreactors?
The VSV-G glycoprotein used to pseudotype most clinical lentivirus vectors is fragile; shear breaks the envelope and destroys infectivity. Published data show lentivirus functional titer drops 50-80% at impeller tip speeds above 1.0 m/s. Wave-mixed platforms have no impeller and generate maximum shear rates of ~10-100 s⁻¹ compared to 500-5,000 s⁻¹ in a stirred tank, so wave-mixed preserves ~2-3× more functional titer for the same nominal particle production. That is why Kymriah, Yescarta, and Breyanzi wave-mixed lentivirus batches are the CGT baseline.
Can you scale up from wave-mixed to stirred-tank?
Yes, but not by geometric scale-up. Wave-mixed has no impeller so P/V and tip-speed criteria don't translate. The practical scale-up path is to match kLa and pCO₂ target, then re-establish DO/pH/feed control in the stirred tank. Expect a re-optimization run: transient AAV titer often shifts ±20-30% between platforms until sparging, impeller speed, and DO setpoint are tuned. Most CGT programs run the seed train in wave-mixed and the production reactor in stirred-tank.
What is the cost difference between wave-mixed and stirred-tank bioreactors?
Wave-mixed capital is 3-5× lower ($30,000-$120,000 for a WAVE 25 or Biostat RM 20 system vs $250,000-$600,000 for a 200-500 L single-use stirred tank plus controller). Consumable per batch is comparable at small scale ($200-500 for a wave bag, $600-1,500 for a 50-200 L SUB bag). At above 500 L, stirred-tank wins on cost per gram of vector because wave doesn't scale that high.
Is the WAVE bioreactor still used in commercial manufacturing?
Yes. Wave-mixed platforms sit in the seed train and clinical-stage production of most approved CGT products. Kymriah, Yescarta, and Breyanzi all use wave-mixed platforms in T-cell expansion. AAV and lentivirus programs from Bluebird, Novartis, and Spark have used WAVE bioreactors for GMP batches. For pivotal-scale commercial AAV (above 500 L), programs typically transfer to stirred-tank single-use platforms because wave-mixed scale is limited.
How does mixing time compare between wave-mixed and stirred-tank?
Stirred-tank mixing time (θ95) is typically 10-30 seconds across scales when P/V is preserved. Wave-mixed mixing time is 30-120 seconds and depends on rocking rate, angle, and fill volume — bags below 30% fill mix in less than 30 s, while bags at 60-70% fill can take 60-120 s. For fast pH or feed titration, stirred-tank is better; for gentle bulk homogenization of shear-sensitive cells, wave-mixed is adequate.

Resources and references