HPLC vs UPLC: What Changes When a Peptide Is Analyzed at Higher Pressure

HPLC vs UPLC: one separation, two operating ranges

The formal definitions settle most of the argument. IUPAC’s 2017 recommendations on the terminology of separation methods say that liquid column chromatography “utilizing small particles (2–20 μm) and a relatively high inlet pressure”, up to 50 MPa, is often called high-performance liquid chromatography. When particles smaller than 2 µm and inlet pressures up to 150 MPa (1,500 bar) are used, the same entry continues, the method is characterized as ultra-high-performance liquid chromatography (UHPLC) or ultra-performance liquid chromatography (UPLC). One technique, two operating ranges.

The second name has a commercial history. In March 2004 Waters Corporation introduced what its annual report described as “a novel HPLC-type technology”: columns of very small particles plus an instrument built “to accommodate the increased pressure and narrow chromatographic bands that are generated by these small particles”. UPLC is a Waters trademark, and Waters describes the technology as columns of 1.7–1.8 µm particles with mobile phase delivered at up to 1,030 bar (15,000 psi). Because the trademark belongs to one company, UHPLC is the neutral generic term, and the one used from here on.

The physics is older than the hardware. MacNair, Lewis and Jorgenson published ultrahigh-pressure reversed-phase separations in packed capillary columns in 1997, and a 2006 paper by Nguyen and colleagues noted that the approach had been described about 30 years earlier from the fundamental chromatographic equations, while compatible systems and columns “have been introduced in the market in 2004 only”.

Why smaller particles sharpen peaks

A peak is a band of molecules that has spread out on its way through the column. That spreading is measured as plate height, which IUPAC defines as the column length divided by the plate number: the smaller the plate height, the more plates fit into a given length and the narrower the peaks.

In 1956 van Deemter, Zuiderweg and Klinkenberg identified “longitudinal diffusion and resistance to mass transfer” as causes of non-ideal band broadening. Column performance is still “usually evaluated” with the “well-known” van Deemter equation, as Felletti and colleagues put it in a 2020 study of sub-2 µm columns. Waters’ primer on small particles describes three contributions:

• Uneven paths. Molecules take slightly different routes around the particles; the effect depends mainly on particle size and how well the bed is packed.
• Diffusion along the column. Molecules drift forwards and backwards as they travel; this matters most at low flow rates.
• Slow mass transfer. Molecules diffuse in and out of each particle’s pores, and some lag behind. This term grows with flow rate and with the square of particle size, so small particles keep their efficiency at high flow.

IUPAC builds the particle-size dependence into its vocabulary: the reduced plate height is “plate height divided by average particle diameter”. Well-packed columns of fully porous particles typically reach a minimum reduced plate height of about 2 (McCalley gives 1.9–2.1). Assuming that value for both, a 5 µm packing bottoms out near a 10 µm plate height and a 1.7 µm packing near 3.4 µm, and the same 100 mm column holds roughly 10,000 plates in the first case and 29,000 in the second.

Resolution improves less than the plate count does. IUPAC defines peak resolution as the difference between two retention times divided by the peaks’ average width at the base, and peak width shrinks only with the square root of the plate number. Tripling the plates therefore improves resolution by a factor of about 1.7. The extra efficiency can be spent on speed instead: keep the ratio of column length to particle diameter constant and the plate count stays about the same, so 50 mm of 1.7 µm particles holds roughly as many plates as 150 mm of 5 µm particles — and small-particle columns also have higher optimum flow rates.

What extra resolution reveals in a peptide sample

Synthetic peptides carry impurities that closely resemble the product. US Food and Drug Administration guidance on certain generic synthetic peptide drugs describes peptide-related impurities as “amino acid sequences related to, but different from, that of the active ingredient”, arising from insertion, deletion or other modifications such as oxidation. Some can elute close to the main peak, which is where reading the chromatogram itself tells you more than the headline figure.

That is where resolution matters. When two peaks overlap, the integration software has to divide the area between them; when they overlap completely there is nothing to divide, and the impurity is reported as product. Resolving a hidden impurity moves area in one direction only: out of the main peak and into the impurity list. The same material can therefore post a lower purity on a higher-resolution method without having changed at all, as the schematic below shows.

Guidance for peptide medicines points the same way. For these generic synthetic peptides, the FDA recommends “sensitive and high resolution analytical procedures (e.g., UHPLC-HRMS)” and asks applicants to identify each peptide-related impurity at 0.10 percent of the drug substance or more. That is a drug framework, and this article makes no claim about which rules apply to any laboratory reagent; it is cited only as a measure of the resolving power considered necessary for peptide impurity work.

Even UHPLC has limits. When Bosc-Bierne and colleagues separated a common pool of 32 synthetic T-cell epitope peptides on a 2 µm capillary column in 2024, six “highly overlapped”, and the authors noted that separating such mixtures “may be limited, even when UHPLC is applied”. Resolution narrows the blind spot; it does not remove it.

Pressure, heat and speed: the practical trade-offs

Small particles are not free. IUPAC’s flow resistance parameter ties a column’s permeability to the square of its particle diameter, and McCalley spells out the consequence: because pressure is inversely proportional to the particle diameter squared, a 1.7 µm column needs about nine times the pressure of a 5 µm column at the same flow velocity. And since their optimum flow rate is higher, small-particle columns are meant to be run faster still — hence instruments rated well above the 50 MPa ceiling IUPAC associates with HPLC.

Pressure also brings frictional heat, which McCalley notes “is caused by percolation of the mobile phase through the packed bed” and which matters most when the flow is pushed through very small particles. Nguyen and colleagues discussed frictional heating and solvent compressibility as the potential problems of high pressure, while reporting that systems working at up to 1,000 bar were not influenced by them.

Narrow peaks are also easily spoiled by the instrument itself. Waters attributes the slow uptake of very small particles to “the negative influence of instrument band spreading and limited pressure range”: tubing, fittings and detector volume broaden a peak again after the column has sharpened it, so a sub-2 µm column on a system not built for it can give back part of its advantage. Whether the extra efficiency went into resolution or speed shows only in the method details. The table sets out UHPLC vs HPLC on the points that bear on a purity figure.

Why purity figures from different methods don’t compare

When a certificate reports purity as an area percentage — the main peak’s area divided by the total integrated area — every term in that ratio depends on the method. Four things move it:

• Resolution and selectivity. Whatever co-elutes with the main peak is counted as product. A different column, particle size or gradient can separate peaks that another method merges; Bosc-Bierne and colleagues note that a shallower gradient is expected to increase both resolution and peak capacity.
• Detection wavelength. UV area reflects absorbance, not amount. Kuipers and Gruppen measured the peptide bond’s molar extinction coefficient at 214 nm as 923 M⁻¹ cm⁻¹, with tryptophan about 30 times higher and phenylalanine, tyrosine and histidine about six times higher, so an impurity that has lost a tryptophan is under-counted.
• Integration. Where the software starts and ends each peak, how it splits overlaps and which small peaks it ignores all change the totals; in the 32-peptide study, two peptides were unresolved in some replicates with the perpendicular-drop method.
• The window. Anything that elutes after the gradient ends, or never leaves the column, is not in the total at all.

For comparing gradient methods, the preferred single figure is peak capacity: in Bosc-Bierne and colleagues’ words, “the number of peaks that can be separated within a retention window”, or in practice “the gradient run time divided by the average peak width”. Narrower peaks from sub-2 micron particles raise it in the same run time, though the authors caution that the highest peak capacity does not necessarily separate a specific sample most usefully.

Whether a purity method can see the impurities that matter is the business of analytical method validation. For two certificates, the working rule is simpler: compare figures from the same method, ideally the same laboratory, or regard the difference as unexplained.

What to check in a certificate’s method details

A line that says only “UPLC peptide analysis” or “HPLC purity” tells you the pressure range of the instrument, not the quality of the separation. These details do:

• The column: chemistry, particle size, length and internal diameter, which largely set how many plates the separation had.
• The gradient: starting and ending solvent composition and how long the change took. Gradient slope changes resolution and peak capacity, so methods on similar columns can still differ.
• The detection wavelength, such as the 214 nm used in the peptide study above. Figures recorded at different wavelengths are not the same measurement.
• The chromatogram and an integration table listing every counted peak with its retention time and area, plus any peaks excluded as solvent or blank signals.
• A separate identity result. A UV trace says nothing about mass; our guide to how LC-MS confirms a peptide’s identity explains what that line can and cannot show.
• A batch number that matches the vial, because a report is evidence about the sample that was tested and nothing else.

What’s established and what isn’t

Reading certificates when buying in Canada

Whether a certificate comes from a laboratory in Canada or abroad, the same habits apply when you compare research reagents:

• Read the method before the percentage. The acronym gives the pressure range; the column, gradient, wavelength and trace tell you what the number means.
• Compare like with like. Lots tested by the same method in the same laboratory can be compared directly; figures from different methods or laboratories can differ for reasons unrelated to the material.
• Read a lower figure from a sharper method with care, not alarm. It may simply be counting impurities that a broader method hid.
• Keep purity and identity apart. Look for the identity result as a line of its own.

We are a supplier ourselves, so apply the same checks to the certificates on our Lab Testing page as you would to anyone else’s. If a method detail you need is not on a document, ask us through the Contact page rather than assuming.

HPLC-tested research peptides

Janoshik test reports for the research peptides below are listed on our Lab Testing page. Look for the method behind each purity figure, as this guide suggests.

BPC-15715-amino-acid research peptideView product GHK-CuCopper tripeptideView product TB-500Thymosin beta-4 research peptideView product KPVLys-Pro-Val tripeptideView product

Frequently asked questions