Dead Volume in Microplates: Geometry, Accuracy, and Why Your Wells Are Never Quite Empty
Plates are not troughs
When most liquid-handling scientists think about dead volume, they think about reagent troughs. Troughs are the obvious case: they hold bulk reagent, they have a fixed geometry, and the dead volume is a single number you add to your stock calculation.
Plates are different. In a microplate, dead volume isn't a bulk reagent planning problem. It's a per-well, per-transfer, per-tip problem that plays out at the 5 to 500 µL scale. The consequences of getting it wrong are also different: not a bulk reagent shortage, but pipetting errors that compromise assay results, a short transfer here, a missed well there, an air bubble that ruins a fluorescence read.
What creates dead volume in a microplate well
Dead volume in a microplate well has three main contributors.
1. Geometric dead volume. This is the liquid that sits below the minimum aspiratable Z-height, the same concept as in a trough but at a much smaller scale. In a standard 96-well flat-bottom plate, the well is approximately 6.5 to 6.7 mm in diameter and the working depth is typically 10 to 11 mm. If the minimum reliable aspirate height is 1 mm above the well bottom, the geometric dead volume per well is:
π x (3.3 mm)² x 1 mm = approximately 34 µL
For a 96-well plate, that's 34 µL per well multiplied by however many wells you're aspirating, which is significant when working with small volumes or precious samples.
2. Pipetting physics at low volumes. At very low remaining volumes, the liquid surface in a well is a small meniscus close to the well bottom. Surface tension, evaporation, and tip wetting effects all become significant. Even if the robot can physically reach the bottom, the accuracy of aspiration at the last 5 to 10% of volume decreases.
3. Carry-over geometry. When a tip aspirates from a well, a film of liquid remains on the exterior of the tip. When the tip exits the well, a droplet may also form at the tip orifice. These aren't dead volume in the traditional sense, but they represent liquid that doesn't reach the destination, and they're included in the practical "volume not transferred" accounting.
Dead volume by plate format
The geometric dead volume per well varies substantially by plate format. These figures assume a 1 mm minimum aspiratable height and approximate well cross-sections.
| Plate format | Well diameter (approx.) | Well depth (approx.) | Dead volume per well (geometric) |
|---|---|---|---|
| 96-well flat bottom | 6.5 to 6.7 mm | 10 to 11 mm | ~30 to 35 µL |
| 96-well round bottom | 6.5 mm | 10 to 11 mm | Tapers to near-zero (bowl shape) |
| 96-well V-bottom | 6.5 mm | 10 to 11 mm | Tapers to near-zero (V-shape) |
| 96-well deep well (2 mL) | 8.3 mm | ~42 mm | ~55 to 65 µL |
| 384-well flat bottom | 3.3 mm | 11 mm | ~9 µL |
| 384-well round bottom | 3.7 mm | ~11 mm | Tapers, ~2 to 5 µL |
| 1536-well | 1.5 mm | ~5 mm | Under 1 µL (tip access is the limiting factor) |
Round-bottom and V-bottom 96-well plates have an advantage here analogous to Tecan tapered troughs: the geometry concentrates remaining liquid at the lowest point, reducing dead volume relative to the flat-bottom equivalent. This is why low-volume assays often use round or V-bottom plates.
The 10% rule in the context of plates
The 10% dead volume rule of thumb has a different application in plates than in troughs.
For troughs, 10% is a planning buffer on total bulk volume. For plates, the rule is typically applied per-transfer, adding 10% to the per-well volume to account for pipetting imprecision at the low end of the tip's range.
This is a rougher approximation for plates than for troughs. Dead volume in a well is dominated by geometric factors (well cross-section and minimum aspiratable height), not by percentage of working volume. For a well holding 200 µL, the geometric dead volume of ~30 µL is 15%, already above 10%. For a well holding 1000 µL in a deep-well plate, the ~60 µL geometric dead volume is only 6%.
The 10% rule underestimates dead volume in small-volume applications and overestimates it in large-volume applications. For plate work, calculate from geometry or use empirical characterisation for your specific plate and protocol.
How tip accuracy affects your dead volume calculation
The geometric dead volume above assumes the robot can reliably aspirate down to 1 mm from the well bottom. Whether it actually can depends on the tip in use, the liquid class, and the calibration state of the instrument.
Tip volume range and accuracy
Each tip type has a working range within which pipetting accuracy is specified. Aspirating at or below the lower limit of that range introduces volume error; you may aspirate less than intended, or aspirate air along with liquid.
On the Hamilton STAR:
| Tip type | Nominal range | Reliable low-volume floor | Implication for dead volume |
|---|---|---|---|
| 1000 µL standard | 50 to 1000 µL | ~20 to 50 µL | Conservative minimum aspirate volume; larger well clearance needed |
| 300 µL standard | 1 to 300 µL | ~2 to 5 µL | Suitable for low-volume aspiration near well base |
| 50 µL low-volume | 0.5 to 50 µL | ~0.5 to 1 µL | Can aspirate very close to well bottom in accessible formats |
| 10 µL ultra-low volume | 0.1 to 10 µL | ~0.1 µL | Specialist use; significant positional precision required |
On the Tecan Fluent, DiTi tip ranges differ but the principle is the same: the larger the tip, the larger the minimum reliable aspiration volume, and the more well dead volume is effectively inaccessible.
If you're working with 50 µL working volumes in a 96-well flat-bottom plate using 1000 µL tips, your effective dead volume may be 30 to 40 µL, which is a significant proportion of your working volume. Switching to 300 µL tips and an appropriate liquid class could bring that floor down substantially.
Pipetting accuracy at the low end of the range
Even within the specified working range, accuracy isn't constant. Coefficient of variation (CV) increases as you approach the minimum volume for any tip type. For a 1000 µL tip aspirating 50 µL, CV may be 3 to 5%. For a 300 µL tip aspirating 50 µL, CV may be 1 to 2%. For a 50 µL tip aspirating 50 µL, CV may be below 1%.
For dead volume calculations in plate work, this accuracy gradient matters. The last aspiration from a well, when remaining volume is close to the dead volume threshold, will have the highest CV. Design your protocol so that the final aspiration from any well isn't the most accuracy-critical transfer in your assay.
The critical role of labware definitions in plate work
In a trough, a labware definition error of 0.5 mm means wasted bulk reagent. In a plate, the same error means systematic per-well pipetting errors across your entire assay. See the Venus worktable for how labware definitions are built and placed in the first place.
What the labware definition controls
The labware definition for a microplate in Venus or FluentControl specifies:
Well X/Y positions: where each well is located relative to the plate origin. A definition built for an SBS-standard plate that doesn't match your actual plate's well layout will cause tip offset errors; the tip enters the well off-centre, which alters the effective Z-clearance, can cause tip-wall contact, and introduces carry-over risk.
Z-bottom (well depth): the Z-position of the well base. If this is wrong, the robot aspirates at the wrong height. Too high and dead volume increases. Too low and the tip contacts the well base, causing pressure errors (on Hamilton TADM / cLLD) or physical tip damage.
Well shape: flat, round, or V-bottom. The shape affects how the robot calculates LLD tracking in the well and whether it should use fixed-height aspiration or LLD-based aspiration.
Common labware definition failure modes in plate work
Using an SBS-standard definition for a non-standard plate. Not all 96-well plates are identical. Variations in well depth, wall thickness, and skirt height exist between manufacturers. A definition from Manufacturer A won't correctly describe a plate from Manufacturer B, even if both are nominally "96-well flat-bottom."
Ignoring the plate bottom profile. Many labs use a generic flat-bottom definition for plates that are actually slight round-bottom or have a different base radius. This makes no difference to high-volume transfers but matters substantially when aspirating close to the well base.
Not updating definitions after a consumable change. Switching plate suppliers mid-project is common. The definition doesn't update automatically. A definition built for your previous plate supplier's geometry may have a Z-bottom error of 0.5 to 1 mm relative to the new supplier's product, invisible in high-volume transfers but significant when aspirating near the dead volume threshold.
Labware teaching for plates: what it changes
Teaching a plate position on a Hamilton STAR or Tecan Fluent verifies that the software model of the deck matches physical reality. For plates, teaching matters at two levels. For the general principles behind this process, see teaching your deck.
Carrier-level teaching
The plate carrier must be correctly positioned on the deck. On the Hamilton STAR, carriers are loaded via the Autoload system or placed manually. If the carrier isn't seated flush, whether tilted, slightly offset, or on a warped deck track, the entire plate position is shifted. A 0.5 mm Y-offset on a carrier translates to a 0.5 mm Y-offset on every well in the plate.
In a 384-well plate with 4.5 mm well pitch, a 0.5 mm carrier offset means the tip is positioned 11% of the way toward the adjacent well. This is within the well at high volumes but becomes a source of cross-contamination risk and dead volume variability at low volumes.
Well-level teaching for high-density plates
For 384-well and 1536-well plates, individual well Z-bottom verification is standard practice. The well depths are smaller, the dead volumes are smaller, and the positional tolerances are tighter. A 0.3 mm error in Z-bottom that's negligible in a 96-well plate may represent 30 to 50% of the dead volume in a 384-well plate.
For 96-well plates in routine applications, carrier-level teaching and a correct labware definition are typically sufficient. For 384-well and below, or for low-volume applications in any format, well-level verification adds meaningful accuracy.
LLD in plates: when to use it and when not to
Capacitive LLD (cLLD) in plates works well in some formats and poorly in others.
Good LLD targets: tubes, deep-well plates. The liquid surface is well above the well base and has a large cross-section, so cLLD tracks reliably.
Challenging LLD targets: standard 96-well plates, where shallow, narrow wells mean LLD sensitivity must be set correctly for the well diameter. See the Hamilton cLLD sensitivity table:
| cLLD sensitivity setting | Recommended vessel |
|---|---|
| 1 (Very High) | 384-well plates |
| 2 (High) | 96-well round-bottom |
| 3 (Medium) | 96-well flat-bottom |
| 4 (Low) | Tubes |
LLD isn't appropriate for aspiration at fixed low volumes close to the well bottom. Once you know the remaining volume is close to dead volume, fixed-height aspiration (with a defined height from well bottom) is more reliable than LLD tracking. LLD at very low volumes in small wells risks detecting the well wall rather than the liquid surface.
The general strategy: use LLD for the main aspiration sequence, then switch to fixed-height aspiration for the final low-volume draw if your protocol requires aspirating close to the dead volume threshold.
Dead volume checklist for microplates
- Calculate dead volume from well geometry (cross-section x minimum aspiratable height), not from the 10% rule
- Choose plate format (flat, round, V-bottom) to match your working volume range
- Select the smallest appropriate tip type for your working volume; this directly reduces the effective dead volume floor
- Verify your labware definition against the physical plate, especially when changing plate suppliers
- Teach carrier positions correctly; for 384-well and below, verify individual well positions
- Use cLLD at the correct sensitivity setting for your plate format
- For final low-volume aspirations near the dead volume threshold, switch to fixed-height aspiration rather than LLD tracking
- Don't plan a protocol with working volumes that bring you routinely to within 20% of the geometric dead volume limit
ProtocolPilot generates optimised, ready-to-run scripts for Tecan and Hamilton from your assay parameters — flagging dead volume, labware, and liquid class gaps before you run a single tip.
Try ProtocolPilot →