HILIC for polar peptides: retention logic where reversed-phase falls short

Reversed-phase HPLC is the default separation for synthetic peptide purity, and for most sequences it performs well. Retention is driven by hydrophobic partitioning onto an alkyl-bonded phase, and a gradient of increasing organic solvent elutes peptides roughly in order of hydrophobicity. The method has a structural weakness, however: peptides that are short, highly charged, or rich in polar residues interact weakly with the stationary phase and elute at or near the void volume, where they are poorly resolved from salts, scavengers, and other unretained material. Hydrophilic interaction liquid chromatography (HILIC) was developed for exactly this region of chemical space, and it has become a useful orthogonal tool for characterizing peptides that reversed-phase handles badly.

This post outlines the retention mechanism of HILIC, the method variables that govern peptide separations, the practical difficulties that limit its routine use, and what a HILIC result can and cannot add to the interpretation of a certificate of analysis.

The retention mechanism: a water layer, not a hydrophobic surface

In HILIC, a polar stationary phase (bare silica, amide, diol, zwitterionic sulfobetaine, or similar) is paired with a mobile phase that is predominantly organic, typically 70 to 95 percent acetonitrile, with a small aqueous fraction. The aqueous component preferentially adsorbs to the polar surface and forms a water-enriched layer. Analytes partition between the organic-rich bulk mobile phase and this immobilized aqueous layer, so the more polar the analyte, the greater its retention. Elution is achieved by increasing the aqueous fraction, which is the reverse of the reversed-phase gradient.

Partitioning is only part of the picture. Depending on the stationary phase and mobile phase pH, secondary interactions contribute substantially: electrostatic attraction or repulsion between charged peptide side chains and surface silanols or bonded charged groups, and hydrogen bonding with the amide backbone. For peptides this matters because the same molecule can carry multiple basic groups (Lys, Arg, His, the N-terminus) and acidic groups (Asp, Glu, the C-terminus). Retention on a charged phase therefore reflects a mixed mode in which both polarity and net charge play roles, and small changes in buffer strength can shift elution order in ways that surprise analysts accustomed to reversed-phase behavior.

One consequence is that selectivity is genuinely different from reversed-phase. Peptides that co-elute on a C18 column because they differ by a single polar residue or a deamidation event may separate on a HILIC phase, and the reverse is also true. This is the property that makes HILIC valuable as an orthogonal method: its failure modes do not overlap closely with those of reversed-phase, so an impurity masked in one system has a reasonable chance of being resolved in the other.

Method variables that matter for peptide separations

Several parameters dominate peptide HILIC performance, and each interacts with the others.

Buffer composition and ionic strength. Because electrostatic interactions contribute to retention, salt concentration is a primary control. Ammonium formate and ammonium acetate at roughly 5 to 20 mM are common choices, partly because they are volatile and compatible with mass spectrometry. Low ionic strength tends to produce strong, sometimes irreversible retention and broad peaks for multiply charged peptides, while higher ionic strength screens charge interactions and sharpens peaks but reduces MS signal through ion suppression. The choice is a compromise rather than an optimum.

Mobile-phase pH. pH determines the ionization state of both the peptide and the stationary phase. At low pH, silanols are largely protonated and acidic residues are neutral, so basic peptides carry a strong positive charge. At moderate pH, silanols ionize and can bind basic peptides strongly. Methods are often run at pH 3 to 5 for practical reasons, but selectivity can change materially across that range, and peptides with histidine are particularly sensitive near their side-chain pKa.

Organic solvent and gradient design. Acetonitrile is strongly preferred over methanol, which is protic and competes for the aqueous layer, weakening retention. Gradients usually begin at high organic content and decrease. Peptides are often retained through a relatively narrow window of aqueous fraction, so shallow gradients are common.

Stationary-phase chemistry. Bare silica, amide, and zwitterionic phases show distinct selectivity for peptides. Zwitterionic phases carry both positive and negative charges and tend to give more balanced behavior toward acidic and basic analytes, while bare silica is more strongly cation-exchange in character. Column equilibration is slower than in reversed-phase, often requiring a generous number of column volumes between runs to re-establish the water layer.

Practical difficulties that limit routine use

HILIC has a reputation for being less forgiving than reversed-phase, and for peptides several issues recur.

The first is sample solvent. A peptide dissolved in water or a high-aqueous diluent and injected onto a HILIC column behaves as a strong solvent plug, because water is the strong eluting component. The result is distorted, split, or broadened peaks. Samples generally need to be dissolved in a high-organic diluent, yet many peptides, particularly polar ones, are poorly soluble at 70 percent acetonitrile or above. Analysts commonly resort to a compromise composition and small injection volumes, which can limit sensitivity and complicate the comparison of results between laboratories.

The second is solubility and on-column behavior more broadly. A peptide that dissolves readily in water may aggregate or precipitate as the organic fraction rises, and this can occur silently, with loss of material at the column head reflected as low recovery rather than an obvious artifact. A purity figure from a method with unverified recovery should be read with caution.

The third is reproducibility. Retention times in HILIC are more sensitive to small variations in water content, buffer preparation, and temperature than reversed-phase retention times. Mobile phases that are pre-mixed and left standing can change composition through evaporation of acetonitrile, and buffer salts can precipitate if the organic fraction is too high for the chosen concentration. System suitability criteria, including a defined critical pair, become correspondingly more important.

The fourth is detection. HILIC mobile phases are generally compatible with low-wavelength UV, but ammonium formate and acetate absorb below about 220 nm, raising baseline noise during a gradient. Volatile buffers favor mass spectrometric detection, and the high organic content actually improves electrospray desolvation efficiency, which is one reason HILIC-MS has been explored for polar peptide characterization.

What HILIC adds to the interpretation of a COA

For a routine research-grade peptide with ordinary hydrophobic character, a well-run reversed-phase method with orthogonal mass confirmation answers most questions, and HILIC is rarely necessary. Its value is concentrated in specific situations: short or highly hydrophilic sequences, peptides rich in Lys, Arg, or Ser/Thr/Gly that elute near the void in reversed-phase, and cases where a known polar impurity such as a deletion sequence lacking a hydrophobic residue may be poorly resolved by the primary method.

Equally important is what a HILIC chromatogram does not establish. A single HILIC purity percentage, like an area-percent figure from any UV method, depends on equal detector response across species, and it measures chromatographic purity rather than peptide content. Counterions, residual water, and residual solvents remain invisible to the detector and unresolved from the void in most conditions, so they are better addressed by amino acid analysis, quantitative NMR, or Karl Fischer titration as discussed elsewhere in this journal. A HILIC result also inherits the recovery question described above: if some fraction of the sample is lost on-column, the peaks that remain may overstate purity.

Where two orthogonal methods are reported, the useful information lies in agreement and disagreement between them. Concordant purity values from reversed-phase and HILIC increase confidence that no major co-eluting species is being counted toward the main peak, whereas a discrepancy is a signal worth investigating rather than averaging. Neither method is a validated procedure for most research peptides in the formal sense of ICH Q2, because no analyte-specific validation exists, so the comparison functions as corroborating evidence rather than as proof.

Closing perspective

HILIC and reversed-phase are best understood as complementary views of the same sample, each with characteristic blind spots. Reversed-phase emphasizes hydrophobic differences and under-resolves the polar extreme; HILIC emphasizes polarity and charge and brings its own constraints around solubility, equilibration, and sample diluent. For analysts reading a COA, the practical question is less which technique was used than whether the method chosen was appropriate to the peptide in hand, whether recovery and system suitability were addressed, and whether the reported number is described as what it is: a comparative chromatographic measure under defined conditions. All materials discussed here are research-use-only compounds, and the analytical principles apply to characterization in laboratory settings.