What do peptide peak tailing, resolution and theoretical plates measure?
Theoretical plates, resolution and peak capacity are linked quantities, not three unrelated scores. In countercurrent distribution, theoretical plates, resolution and peak capacity are treated as related separation quantities (DOI:10.1080/00372367108055559). In electrophoresis and sedimentation, variance generation is the common origin of a plate number, a resolution and a peak capacity (DOI:10.1080/01496396908052249).
For a research peptide HPLC peak, theoretical plates are calculated from retention time and a width. The half-height form, N = 5.54 (tR / W0.5)^2, is less sensitive to baseline noise than the base-width form, N = 16 (tR / W)^2. If the peak tails, half-height N can look healthier than base-width N, because the tail adds little height at 50 percent. The record should name the width convention, or two lots cannot be compared.
Resolution between the main peptide peak and a neighbour is commonly Rs = 2 (tR2 - tR1) / (W1 + W2), using baseline widths. The denominator of that expression is peak width. Tailing enters resolution mainly through the width term, and through any apex shift when the integrator sits the apex on a skewed profile. A resolution figure that does not name the neighbouring peak is not interpretable.
Tailing is reported as a tailing factor or as an asymmetry factor. The USP tailing factor at 5 percent of height is T = W0.05 / (2f), where f is the front width from the leading edge to the perpendicular dropped from the apex. Asymmetry is often the back-to-front width ratio at 10 percent of height. A field that says only tailing leaves the comparison ambiguous.
A tailed peptide peak can still integrate to a high area percent if a coeluting impurity sits inside the tail and is not split out. Plates, resolution and tailing are the checks on whether that integral was taken on a separated band.
How does a theoretical-plate simulation build a chromatographic peak shape?
The method of theoretical plates treats the column as equilibrated stages. Solute is transferred stage to stage, and the amount leaving the last stage versus transfer number is the simulated chromatogram. Deineka used that construction to simulate chromatographic peak shape, which makes the plate model a shape model rather than only an efficiency number (DOI:10.1134/s0036024408010160). With linear stages and a large stage count, the profile approaches a Gaussian whose variance grows with the number of transfers.
Peptide peaks often leave the ideal. The construction assumes a distribution coefficient that does not change with concentration, and a bed in which every streamline sees the same number of stages. Nonlinear adsorption produces a sharp front and an extended tail even though the nominal stage count has not changed. Empirical N from a measured width then no longer equals the stage count in the simulation.
Independent dispersion processes add their variances. Giddings framed plates through variance generation, so resolution and peak capacity follow once total variance is known (DOI:10.1080/01496396908052249). In a peptide chromatogram, observed variance sums packing, sample-introduction, tubing, detector-cell and slow adsorption-desorption contributions. Extra-column terms do not grow with column length as packing variance does. A short column can therefore show a modest plate count even when the packing is efficient, because extra-column dispersion is a large fraction of width.
When certificates are compared, a higher N may reflect column length, flow rate or introduced volume rather than peptide composition. A drop in N between a reference chromatogram and a sample chromatogram on the same column and day is more informative than the absolute count. Absolute plate counts copied from different column lengths are not a composition comparison. State column length, particle size and flow beside N before any lot-to-lot reading.
Why does peptide peak tailing reduce resolution when plate count still looks high?
Resolution uses the sum of peak widths. A tail increases baseline width more than it increases half-height width, so Rs can fall while half-height N still looks acceptable. The plate number of the main peak does not report that failure.
Optimisation for solutes that already tail is a different problem from maximising plates for a Gaussian solute. Sekulic and colleagues set out optimisation strategies specifically for solutes exhibiting peak tailing, which matches the problem when a peptide critical pair is skewed rather than merely broad (DOI:10.1016/s0021-9673(01)89158-5). Lengthening the column raises N for both peaks, but it does not remove a concentration-dependent or secondary-interaction tail. If the tail width grows with load, a higher plate count measured on a dilute system-suitability marker will not predict resolution of the sample peak at the analytical load.
Tailing can also appear when the isotherm is linear. Miyabe and colleagues examined peak tailing and column radial heterogeneity in linear chromatography, where a nonlinear isotherm is not the stated cause (DOI:10.1016/s0021-9673(98)00852-8). Radial differences in mobile-phase velocity or retention send some solute molecules ahead of others. The late fraction becomes a tail. Because the process is a property of the bed, it can affect a reference standard and a research peptide similarly.
Peak resolution and tailing are paired even outside liquid chromatography. Alpha-particle spectrometry of environmental samples treats resolution and tailing as joint descriptors of peak shape (DOI:10.1016/j.apradiso.2004.03.038). It is not a peptide method, but it shows that resolution without a tailing description is incomplete. On a research peptide chromatogram the same rule applies: Rs, N and a stated tailing factor are read as a set.
Which silanol and extra-column effects cause peptide peak tailing?
Residual silanols are a documented source of secondary interaction for proteins, peptides and their impurities on commercial size-exclusion columns. Javeri and colleagues added cyclodextrin to an SEC-HPLC mobile phase to mitigate those interactions with residual silanols, and associated that change with better separation and resolution (PMID:29946984). The result is method-specific. It does not authorise cyclodextrin in every reversed-phase peptide method, and it is not a lot-release limit. It does show that a tail, and the resolution lost with it, can come from the stationary-phase surface rather than from a low plate count in the packing.
Contemporary LC troubleshooting has set out strategies for peak tailing (DOI:10.56530/lcgc.na.br4774h5). A useful laboratory split is chemical tailing versus instrumental tailing. Chemical tailing includes silanol interaction, metal interaction, overload and a sample solvent that is stronger than the mobile phase. Instrumental tailing includes excessive tubing volume, a large detector cell relative to peak volume, a poor connection, and a void at the column inlet. Those causes change W, and therefore both N and Rs, without any change in the peptide sequence. A shared tail on a non-peptide marker and the peptide points to the column or the flow path, not to the peptide alone.
Instrumental tailing is not unique to peptide HPLC. In capillary gas chromatography, solvent-peak tailing has been reduced by pressure equalisation into a buffer volume on a removable splitless inlet (DOI:10.1002/jhrc.1240020907). That device is not an HPLC fitting. The transferable point is narrower: a tail can be an inlet or flow-path artefact, so a tailing factor on a certificate should be read beside introduced volume, extra-column hardware and column identity, not only beside the peptide name.
Record each hardware or mobile-phase change, because an undocumented column or additive change makes a later plate count incomparable.
How are peptide peak tailing resolution theoretical plates recorded on a research certificate?
A usable record states the formula, not only the number. For theoretical plates, state whether N used half-height width or baseline width, and state the retention time that entered the formula. For tailing, state USP tailing factor or asymmetry factor, and state the percent of height at which the width was taken. For resolution, name both peaks, not only the main peak, and state whether widths were baseline widths or half-height widths. Without those fields, peptide peak tailing resolution theoretical plates cannot be compared across laboratories or across software versions.
System suitability and sample results are different rows. A marker chromatogram can show acceptable N and tailing while the peptide sample, at a higher load or in a different diluent, tails and loses resolution from a related substance. The certificate should identify which chromatogram supplied each value, and it should carry the method identifier, column description, flow, temperature, gradient timetable and detection wavelength.
Integration events belong in the same block. Valley-to-valley, drop-perpendicular and skim events change apparent peak width, and therefore change both N and Rs, even when the raw detector trace is unchanged. If a tail is skimmed into the main peak, area percent rises and resolution against the shoulder disappears from the peak table. An Australian lot record should show lot number, sample name and chromatogram identifier together, so the tailing factor is tied to one chromatogram.
Numerical habits such as Rs at least 1.5, tailing factor not more than 2.0, or N at least 2000 are laboratory conventions for system suitability. They are not findings of the plate-theory papers cited here, and they are not purity or identity specifications. Passing them is not an identity conclusion and does not replace orthogonal mass confirmation of the main peak.
When should a Gaussian plate count be refused for a tailed peptide peak?
A Gaussian plate formula is a summary of variance under a shape assumption. When the chromatogram shows a visible tail, front or shoulder, that assumption has already failed. Forcing N from half-height width, then treating the number as a precise column efficiency, hides the part of the peak that most affects resolution. The appropriate record is the measured tailing factor, the resolution against the nearest named peak, and a statement that the plate count is indicative only.
A high-resolution alpha-particle study proposed a peak-shape model with an explicit high-energy tail (DOI:10.1140/epja/i2019-12827-x). That model is not transferable to peptide HPLC, and its parameters must not be copied onto a certificate. The relevant lesson is methodological: if the shape is skewed, the shape model should be stated, rather than implied by a single plate count. In chromatography, plate-model simulation is a way to study peak shape rather than to assume a Gaussian (DOI:10.1134/s0036024408010160).
Resolution and plates also do not identify the chemical species under the peak. A symmetrical main peak can still be the wrong peptide or an unresolved coelution. Orthogonal mass measurement, sequence-level fragment evidence where the method provides it, and a second chromatographic selectivity are the identity checks. Peak capacity, as used in the plate-and-resolution literature, only estimates how many bands could fit. It does not assign any band (DOI:10.1080/00372367108055559).
For procurement of research material in Australia, these metrics support a documentation comparison. They help a laboratory decide whether a chromatogram was suitable for an area-percent calculation and whether two lots were run under comparable conditions. They are not evidence of a biological outcome and not use instructions. Research-use analytical framing stops at identity, purity method, system suitability and lot traceability.
Frequently asked questions
What is the difference between theoretical plates and resolution?
Theoretical plates describe band variance relative to retention time for one peak. Resolution compares the gap between two retention times with the sum of the two peak widths. A peptide peak can show a high plate count and still fail resolution if a neighbour is close, or if tailing widens the base. Plates are a property of one band. Resolution is a property of a pair. A certificate should name the pair used for Rs rather than print a bare number.
Does a high plate count rule out peptide peak tailing?
No. Half-height plate formulas barely see a low, long tail, so N can remain high while the baseline width used for resolution is already compromised. Radial bed heterogeneity and slow secondary adsorption can add a tail without lowering the nominal stage count in a simple Gaussian calculation. Read N beside a tailing or asymmetry factor measured at a stated height, and beside the chromatogram, not instead of them.
How can residual silanols affect peptide peak shape?
Residual silanols on commercial size-exclusion columns can interact with proteins, peptides or their impurities and distort separation. A published SEC-HPLC method added cyclodextrin to the mobile phase to mitigate those interactions and reported better separation and resolution. That finding is method-specific. It does not mean every reversed-phase peptide method should add cyclodextrin, and it is not a purity limit for a research lot.
Which tailing value should a certificate state?
State the formula, the peak height at which width was measured, and whether the result is a USP tailing factor or an asymmetry factor. A bare label such as tailing 1.2 cannot be compared with another laboratory's asymmetry 1.2. Also state whether the value came from a system-suitability chromatogram or from the sample chromatogram, because those are different chromatographic events and can disagree.
Can tailing hide a related-substance peak?
Yes. If a minor component elutes inside the tail and the integrator assigns one baseline, area-percent purity can look high while resolution against that component is not demonstrated. Skim or drop-perpendicular events change the apparent width and therefore both N and Rs. Orthogonal mass measurement, not a larger plate count alone, is what assigns an unresolved shoulder to a chemical species.
Should alpha-spectrometry tail models be used on a peptide lot record?
No. Papers on alpha-particle spectra discuss peak resolution, tailing and high-energy tail models for a different detector and a different physical process. They are a reminder that skewed peaks need an explicit shape model. Peptide lot records should use the HPLC method, the stated plate and tailing formulas, and orthogonal identity data. They should not import a spectrometry tail parameter as a release limit.
References
- DOI:10.1080/00372367108055559 — Theoretical Plates, Resolution, and Peak Capacity in Countercurrent Distribution — Separation Science — 1971
- DOI:10.1080/01496396908052249 — Generation of Variance, “Theoretical Plates,” Resolution, and Peak Capacity in Electrophoresis and Sedimentation — Separation Science — 1969
- DOI:10.1134/s0036024408010160 — Simulation of the shape of a chromatographic peak by the method of theoretical plates — Russian Journal of Physical Chemistry A — 2008
- DOI:10.1016/s0021-9673(01)89158-5 — Optimization strategies for solutes exhibiting peak tailing — Journal of Chromatography A — 1989
- DOI:10.1016/s0021-9673(98)00852-8 — Peak tailing and column radial heterogeneity in linear chromatography — Journal of Chromatography A — 1999
- DOI:10.1016/j.apradiso.2004.03.038 — Peak resolution and tailing in alpha-particle spectrometry for environmental samples — Applied Radiation and Isotopes — 2004
- PMID:29946984 — Use of Cyclodextrin as a Novel Agent in the SEC-HPLC Mobile Phase to Mitigate the Interactions of Proteins or Peptide or their Impurities with the Residual Silanols of Commercial SEC-HPLC Columns with Improved Separation and Resolution — Pharm Res — 2018
- DOI:10.56530/lcgc.na.br4774h5 — The Evolution of LC Troubleshooting: Strategies for Improving Peak Tailing — LCGC North America — 2023
- DOI:10.1002/jhrc.1240020907 — Minimizing solvent peak tailing by pressure equalization into a buffer volume. A simple device applied to an easily removable splitless injector for capillary GC — Journal of High Resolution Chromatography — 1979
- DOI:10.1140/epja/i2019-12827-x — A peak shape model with high-energy tailing for high-resolution alpha-particle spectra — The European Physical Journal A — 2019
Research use only
This article is provided for laboratory research and educational purposes only. Products referenced are not for human or veterinary use. ClaraScience makes no therapeutic, medical, or efficacy claims, and nothing here constitutes medical advice.
