Written and maintained by CASRAI Editorial Board
Last updated
Column volume, void volume and dwell volume are three different numbers that get casually lumped together as “the volume of the system,” but they answer three different questions: how much liquid the column tube physically holds, how much of that liquid actually moves through the packed bed with the mobile phase, and how much volume sits between the pump’s mixer and the column inlet before a gradient change even reaches the packing. Getting them right matters for method transfer, gradient reproducibility and knowing how long to equilibrate before the next injection. This guide works through the arithmetic for each, with numbers you can substitute your own column and flow rate into, plus the step-gradient procedure for measuring dwell volume on your own instrument rather than trusting a spec sheet.
Column volume (Vc): the geometric volume of the tube
Column volume is the total internal volume of the empty column tube — packing material, mobile phase and all — calculated purely from its dimensions:
Vc = π × r² × L
where r is the internal radius (half the internal diameter) and L is the packed bed length, both converted to the same length unit before multiplying, with the result in the corresponding volume unit (cm³ = mL when both dimensions are in cm).
| Column | ID | Length | Calculation | Vc |
|---|---|---|---|---|
| Conventional analytical HPLC | 4.6 mm | 150 mm | π × (0.23 cm)² × 15 cm | 2.49 mL |
| Narrow-bore UHPLC | 2.1 mm | 50 mm | π × (0.105 cm)² × 5 cm | 0.17 mL |
Vc is a purely geometric number — it says nothing about how much of that space the packing material itself occupies, which is what void volume corrects for.
Void volume (V0): the volume mobile phase actually occupies
The packing material displaces part of the column’s geometric volume, so the mobile phase only occupies a fraction of Vc — the interparticle (interstitial) space between particles plus the intraparticle space inside their pores that the mobile phase can access. That fraction is the column’s total porosity, and it depends on particle morphology:
| Particle type | Typical total porosity (εT) | Why |
|---|---|---|
| Fully porous silica (3–5 µm) | ≈0.65–0.70 | Standard reference range; most rule-of-thumb calculations use 0.65. |
| Superficially porous / core-shell (≈2.6–2.7 µm) | ≈0.50–0.60 | The solid, non-porous core displaces mobile-phase volume the fully porous particle wouldn’t. |
| Monolithic silica | ≈0.80–0.85 | Combined macropore (flow-through) and mesopore structure gives a higher total porosity than any particulate bed. |
The working approximation used across the industry is:
V0 ≈ 0.65 × Vc (for a conventional fully porous particle bed)
Applied to the two columns above at εT = 0.65:
- 4.6 × 150 mm: V0 ≈ 0.65 × 2.49 mL = 1.62 mL
- 2.1 × 50 mm: V0 ≈ 0.65 × 0.17 mL = 0.11 mL
This geometric estimate is a starting point, not a substitute for measuring it directly. The direct method: inject an unretained marker (uracil is standard for reversed-phase; sodium nitrate or thiourea are also used) and record its retention time, t0. Void volume is then V0 = t0 × F, where F is the flow rate. At 1.0 mL/min on the 4.6 × 150 mm column above, a t0 of 1.62 minutes corresponds to the 1.62 mL calculated geometrically — the two methods should agree within the uncertainty of the 0.65 approximation. A measured V0 that’s meaningfully off the geometric estimate is worth investigating (packing settling, void formation at the frit, or a superficially-porous particle you’d assumed was fully porous).
Void volume is what the retention factor (k’) calculation actually depends on: k’ = (tR − t0) / t0, where tR is an analyte’s retention time. An error in V0 propagates directly into every k’ value calculated from it, which is one reason method-transfer protocols specify how V0 was determined rather than leaving it to a generic assumption.
Dwell volume (gradient delay volume): the volume the column never sees
Dwell volume is a different quantity entirely — it isn’t a property of the column at all. It’s the volume between the point where the mobile-phase composition is actually set (the proportioning valve or mixer) and the column inlet: mixer volume, connecting tubing, the autosampler flow path if it’s inline, and any inline filter or guard column. When a gradient method changes %B, that new composition has to physically travel through the dwell volume before it reaches the top of the packed bed — so the column experiences every gradient step later than the pump’s own timeline says it started.
Dwell volume varies by pump architecture, not just by manufacturer: low-pressure quaternary-gradient (LPGE) systems mix all four channels before the pump head, so the mixer plus pump volume is part of the dwell path and typical dwell volumes run roughly 1,000–1,500 µL. High-pressure binary-gradient (HPGE) systems mix after two separate pump heads, closer to the column, and typically run lower — roughly 100–400 µL on conventional HPLC, often under 100 µL on UHPLC systems built specifically to minimize it. These are broad ranges, not a substitute for measuring your own instrument — the same nominal pump model can differ meaningfully by tubing length, mixer size and whether an inline filter is installed.
Measuring dwell volume on your own instrument
The standard test isolates the pump/mixer/tubing path from the column entirely, so it measures the system’s dwell volume and nothing else:
- Replace the column with a zero-dead-volume union (or a very short piece of tubing) so the column’s own void volume doesn’t contribute to the result.
- Prepare mobile phase A as water (or your normal weak solvent) and mobile phase B as the same solvent with a small amount of a UV-absorbing tracer added — 0.1% acetone or 0.1% dimethyl sulfoxide are both standard choices, monitored around 254–265 nm.
- Program a step gradient, not a ramp: hold 0% B briefly, then step immediately to 100% B and hold, at a fixed flow rate.
- Record the detector trace. The absorbance stays near baseline, then rises through an S-shaped breakthrough curve as the 100% B front arrives, then plateaus.
- Read the dwell time (tD) off that curve. Two conventions are both used in practice and give slightly different numbers: reading the time at 50% of the plateau absorbance (the simpler, more common approach), or extrapolating the steepest linear portion of the rising edge back to the pre-step baseline (closer to some pharmacopeial method-transfer protocols). Pick one and note which you used — the two can disagree by several tenths of a minute at typical flow rates, enough to matter for a tight method transfer.
- Convert to volume: VD = tD × F.
Worked example: at 1.0 mL/min, a measured tD of 1.0 minute gives VD = 1.0 mL — consistent with a quaternary LPGE system in the typical range above. At 0.4 mL/min on a UHPLC binary system, a measured tD of 0.25 minutes (15 seconds) gives VD = 0.1 mL. This is exactly why a gradient method developed on a high-dwell-volume quaternary system doesn’t reproduce on a low-dwell-volume UHPLC system without adjustment — the effective gradient the column actually experiences starts earlier, and a delay/isocratic hold segment sized to the dwell-volume difference is the usual fix during transfer.
Equilibration volume: how much to flush before the next injection
Before the next injection, the system has to (1) flush the dwell volume back down to the initial %B, and (2) push enough mobile phase through the packed bed for the stationary phase itself to re-equilibrate at that composition — not just for the composition reaching the column inlet, but for the phase-mobile phase partition to restabilize. As a starting point:
Equilibration volume ≈ (N × V0) + VD
where N is commonly 5–10 column volumes for a reversed-phase C18 method returning to a stable initial composition; HILIC methods typically need substantially more (often 15–20+ column volumes) because the immobilized water layer that drives HILIC retention re-equilibrates more slowly than a reversed-phase surface does. Applied to the two example columns at N = 10:
- 4.6 × 150 mm at 1.0 mL/min: (10 × 1.62 mL) + 1.0 mL = 17.2 mL → ≈17.2 minutes before the next injection.
- 2.1 × 50 mm at 0.4 mL/min: (10 × 0.11 mL) + 0.1 mL = 1.2 mL → ≈3.1 minutes before the next injection.
The gap between those two numbers is the practical payoff of the whole calculation: for a short, narrow-bore UHPLC column on a low-dwell-volume system, equilibration is a small fraction of the run; for a conventional-format column on a high-dwell-volume quaternary system, it can rival the gradient run time itself, which is a real throughput cost worth knowing before you build a sequence around it.
Frequently asked questions
Does void volume change over a column’s lifetime?
Yes — packing settling, frit fouling or partial channel formation can shift measured V0 over time, which is exactly why re-measuring it with an unretained marker periodically (rather than trusting the original geometric estimate indefinitely) is a useful column-health check, alongside monitoring backpressure and plate count.
Why do vendors quote different dwell volumes for the “same” pump?
Because dwell volume includes the connecting tubing and any inline filter or guard column, not just the pump and mixer — two identically-specified systems set up with different tubing lengths or an added inline filter will have measurably different dwell volumes. That’s the core reason to measure it on your own configuration rather than citing a datasheet figure.
Do I need to know column volume to calculate void volume?
Only for the geometric estimate. If you measure V0 directly with an unretained marker (t0 × F), you never need Vc at all — the geometric route is useful mainly as a sanity check or when you don’t have an unretained-marker run available yet.
For the column-hardware side of these calculations — particle type, porosity and stationary-phase selection — see HPLC Column Selection: Stationary Phases & Selectivity. For how dwell volume differs between HPLC and UHPLC platforms and what that means for method transfer, see HPLC vs UPLC vs UHPLC. For the broader troubleshooting picture this arithmetic feeds into, see HPLC: Columns, Mobile Phases, and a Peak-Problem Troubleshooting Table.
