Quick answer: scale flow by the square of the diameter ratio — F₂ = F₁ × (d₂/d₁)². From 4.6 mm at 1.0 mL/min to 2.1 mm gives 0.21 mL/min. If the particle size also changed, multiply by dp₁/dp₂ as well to hold efficiency constant. Back pressure follows ΔP ≈ φηLu/dp², so halving particle size roughly quadruples pressure.
Why flow scales with the square of the diameter
The quantity that governs a separation is not how many millilitres per minute the pump delivers — it is how fast the mobile phase moves along the column. That speed, the linear velocity, is flow rate divided by cross-sectional area:
Cross-sectional area goes with the square of the diameter, so a column half as wide needs a quarter of the flow to run at the same speed. Get this wrong in the intuitive direction — halving the flow when you halve the diameter — and you are running at twice the linear velocity you intended, with peaks eluting early and efficiency down.
Here is the scaling worked out for the transfers people actually make:
| From ID | To ID | Factor (d₂/d₁)² | 1.0 mL/min becomes |
|---|---|---|---|
| 4.6 mm | 3.0 mm | 0.425 | 0.43 mL/min |
| 4.6 mm | 2.1 mm | 0.208 | 0.21 mL/min |
| 4.6 mm | 1.0 mm | 0.047 | 0.047 mL/min |
| 3.0 mm | 2.1 mm | 0.490 | 0.49 mL/min |
| 4.6 mm | 10 mm | 4.73 | 4.7 mL/min |
| 4.6 mm | 21.2 mm | 21.2 | 21.2 mL/min |
| 4.6 mm | 50 mm | 118 | 118 mL/min |
The preparative rows are the ones worth staring at. Scaling a 1 mL/min analytical method to a 50 mm prep column means 118 mL/min — which is why prep systems have entirely different pumps, and why solvent cost, not resolution, is usually what limits prep throughput. Injection volume scales the same way, by column volume ratio.
The particle-size correction most transfers forget
Diameter scaling alone is only half the job when the particle size changes too — which it almost always does when a method moves from HPLC to UHPLC. Efficiency is governed by reduced velocity, which normalises linear velocity against particle size:
Full scaling: F₂ = F₁ × (d₂/d₁)² × (dp₁/dp₂)
Move from 5 µm to 1.7 µm particles and the extra factor is 5 ÷ 1.7 = 2.94. So a full 4.6 mm / 5 µm → 2.1 mm / 1.7 µm transfer is:
- Diameter term: (2.1 ÷ 4.6)² = 0.208
- Particle term: 5 ÷ 1.7 = 2.94
- Combined: 1.0 × 0.208 × 2.94 = 0.61 mL/min
Not the 0.21 mL/min the diameter alone suggested. Run the smaller particles at 0.21 mL/min and you are well below the efficiency optimum, wasting both the column's capability and the analyst's time. The trade is pressure — see the next section.
Estimating column back pressure
Pressure drop across a packed bed follows Darcy's law, usually written for chromatography as:
φ ≈ 1000 (flow resistance, well-packed spherical particles) · η = viscosity · u = superficial linear velocity
Everything in that expression is linear except the particle diameter, which is squared and in the denominator. That single fact explains the shape of the whole industry: sub-2 µm particles give better efficiency, and cost roughly nine times the pressure of 5 µm at the same linear velocity.
| Column | Particle | Flow | Linear velocity | Est. ΔP |
|---|---|---|---|---|
| 4.6 × 150 mm | 5 µm | 1.0 mL/min | 1.00 mm/s | ~60 bar (873 psi) |
| 4.6 × 250 mm | 5 µm | 1.0 mL/min | 1.00 mm/s | ~100 bar (1,455 psi) |
| 4.6 × 150 mm | 3 µm | 1.0 mL/min | 1.00 mm/s | ~167 bar (2,425 psi) |
| 2.1 × 50 mm | 1.7 µm | 0.4 mL/min | 1.92 mm/s | ~333 bar (4,830 psi) |
| 2.1 × 100 mm | 1.7 µm | 0.4 mL/min | 1.92 mm/s | ~666 bar (9,660 psi) |
| 2.1 × 50 mm | 2.6 µm core–shell | 0.4 mL/min | 1.92 mm/s | ~142 bar (2,060 psi) |
Estimates assume η = 1.0 cP (about water at 20 °C) and φ = 1000. Real pressures vary with mobile phase, temperature, packing quality and the instrument's own tubing contribution. Treat these as "will this fit inside my pressure limit" figures, not specifications.
The last row is the commercially important one. A 2.6 µm core–shell particle delivers efficiency close to a sub-2 µm fully porous particle at less than half the pressure — which is what allows a conventional 400 bar HPLC to run near-UHPLC separations without being replaced.
Mobile phase viscosity — the term people forget
Pressure is directly proportional to viscosity, and mobile phase viscosity is not the average of its components. Both common reversed-phase mixtures peak somewhere in the middle:
| Mobile phase | Viscosity at 25 °C | Effect on pressure |
|---|---|---|
| Water | 0.89 cP | Baseline |
| Acetonitrile | 0.37 cP | Lowest of the common solvents |
| Methanol | 0.54 cP | Low on its own |
| ACN / water 20:80 | ~0.95 cP | The ACN–water maximum, only slightly above water |
| ACN / water 50:50 | ~0.85 cP | Below water — gradients often drop in pressure |
| MeOH / water 40:60 | ~1.7 cP | Nearly double water — the pressure peak of a MeOH gradient |
| MeOH / water 80:20 | ~1.1 cP | Falling back down |
This explains a phenomenon that confuses people daily: on a methanol gradient the pressure rises in the middle of the run and falls again at the end, while on an acetonitrile gradient it falls steadily. Nothing is wrong with the column in either case. If you are near your instrument's pressure ceiling with methanol, switching to acetonitrile or raising the column temperature to 40 °C buys you meaningful headroom — viscosity drops roughly 2% per degree.
Common mistakes & pro tips
- Scaling flow linearly with diameter. Halving the diameter needs a quarter of the flow, not half. This is the single most common flow-scaling error.
- Ignoring the particle-size term. Scaling only by diameter when moving to smaller particles leaves you running well below the efficiency optimum.
- Forgetting the instrument's own pressure contribution. Tubing, the injector and especially a narrow detector cell add pressure that the column calculation does not include. On UHPLC this can be tens of bar.
- Exceeding the column's rating, not just the pump's. Many columns are specified below their instrument's ceiling; the packing can be compressed permanently.
- Pro tip — change one thing at a time. When a transfer misbehaves, revert to the original flow on the new column first. If the chromatography is right and only the timing is off, the problem is scaling. If peak shape is wrong too, it is the column or the dwell volume.
- Pro tip — check compendial limits before you scale. USP <621> allows specified adjustments to flow and column dimensions without revalidation, but the allowances are bounded. Scaling outside them turns an adjustment into a change.
How to use this calculator
- Scale between columns — enter the original ID, particle size and flow, then the new column's ID and particle size. You get the scaled flow with and without the particle correction.
- Velocity & pressure — enter one column and one flow rate to see linear velocity, reduced velocity, and an estimated back pressure.
- Pick a mobile phase to set viscosity, or enter your own value in centipoise.
- Compare the reduced velocity to 3. That is roughly the efficiency optimum; well above it you are trading efficiency for speed, which is often the right trade.
Frequently asked questions
How do you scale HPLC flow rate between column sizes?
Scale by the square of the internal diameter ratio: F₂ = F₁ × (d₂/d₁)². Moving from a 4.6 mm column at 1.0 mL/min to a 2.1 mm column gives 1.0 × (2.1/4.6)² = 0.21 mL/min. Squaring is right because flow has to change with cross-sectional area, not with diameter, to keep the mobile phase moving along the column at the same speed.
What is linear velocity in HPLC?
Linear velocity is how fast the mobile phase actually travels down the column, in millimetres per second, as opposed to volumetric flow rate in millilitres per minute. It is flow rate divided by the column's cross-sectional area. It matters because efficiency depends on linear velocity: two columns of different diameters run at the same linear velocity give the same separation, while two run at the same volumetric flow rate do not.
How do you calculate HPLC back pressure?
Column pressure drop follows Darcy's law: ΔP = φ × η × L × u ÷ dp², where φ is a flow-resistance factor of roughly 1000 for a well-packed bed, η is mobile phase viscosity, L is column length, u is superficial linear velocity and dp is particle diameter. The dominant term is dp² in the denominator — halving particle size quadruples the pressure at the same velocity, which is the entire reason UHPLC systems are built for 1000 bar.
What is the optimal flow rate for an HPLC column?
Efficiency peaks at a reduced velocity of about 3, which corresponds to a linear velocity of roughly 3 × Dm ÷ dp. For 5 µm particles that is around 0.6 mL/min on a 4.6 mm column; for 1.7 µm particles on a 2.1 mm column it is around 0.4 mL/min. In practice most methods run faster than the theoretical optimum because the van Deemter curve is flat on its right-hand side for small particles — you lose very little efficiency and save a lot of time.
Does flow rate affect retention time?
Retention time is inversely proportional to flow rate — double the flow and every peak elutes in half the time. Retention factor k, however, does not change, because both the peak time and the void time halve together. That is precisely why k rather than retention time is the transferable quantity in a method, and why system suitability criteria are written in terms of k and resolution.
Why did my back pressure go up without changing the flow rate?
Most often a blocked inlet frit from particulates in the sample or mobile phase — the classic sign is pressure climbing over days while peak shape stays normal. Other causes are a mobile phase change to a more viscous mixture (methanol–water peaks near 1.7 cP at about 40% methanol, twice the viscosity of either pure solvent), a drop in column temperature, or a partially blocked in-line filter. Reverse-flushing the column off-detector often resolves a frit blockage.
References & further reading
- Van Deemter, J. J., Zuiderweg, F. J. & Klinkenberg, A. (1956), Chemical Engineering Science — the original plate-height/velocity relationship.
- Snyder, L. R., Kirkland, J. J. & Dolan, J. W., Introduction to Modern Liquid Chromatography, 3rd ed. — column permeability, reduced parameters and method scaling.
- Shimadzu, "Optimizing Your HPLC/UHPLC System — General Recommendations" — practical pressure and velocity guidance.
- USP General Chapter <621> Chromatography — permitted adjustments to flow rate, column length and particle size.
- Standard viscosity data for acetonitrile–water and methanol–water mixtures at 25 °C.
Back-pressure figures are estimates from the Darcy relationship with a generic flow-resistance factor; they exclude your instrument's tubing and detector contribution. Always confirm against the column's and pump's rated limits. See our accuracy policy.