Universal detection for peptides without a chromophore: charged aerosol detection and ELSD against UV

Reversed-phase chromatography with ultraviolet detection is the default purity method for synthetic peptides, and the default carries an assumption that is rarely stated explicitly: that every species eluting from the column absorbs ultraviolet light in proportion to how much of it is there. For the peptide backbone at 214 nm, that assumption holds reasonably well, because the amide bond itself is the chromophore and every residue contributes one. At 280 nm it holds only for sequences containing tryptophan, tyrosine, or cystine. And for a large class of process-related impurities — residual scavengers, protecting-group fragments, counterion-associated species, non-peptidic excipients carried through from synthesis — it does not hold at all. Those species can pass through a UV detector nearly invisibly while occupying real mass in the vial.

Aerosol-based universal detectors were developed to close that gap. Evaporative light scattering detection (ELSD) and charged aerosol detection (CAD) both respond to any analyte less volatile than the mobile phase, regardless of its optical properties. Neither replaces UV for peptide purity work, and neither is a drop-in substitute for the numbers on a certificate of analysis. But understanding what they measure clarifies something about UV that is easy to forget — that a chromatographic purity figure is a statement about detector response, not about mass.

Where the ultraviolet assumption fails

Peptide bonds absorb strongly in the far ultraviolet, with a maximum near 190 nm and useful analytical response in the 210–220 nm window where mobile-phase absorbance is still tolerable. Detection at 214 nm is therefore approximately proportional to the number of amide bonds in a molecule, which for closely related peptide impurities — deletion sequences, truncations, deamidated variants — makes area percent a serviceable proxy for molar composition. A des-Gly deletion impurity has one fewer amide bond than the target and responds within a few percent of it. This is the case that makes RP-HPLC-UV work as well as it does.

The proportionality degrades in two directions. Short fragments and single protected amino acids carry few amide bonds and are underrepresented relative to their mass. More importantly, wholly non-peptidic material may carry none. Trifluoroacetic acid, the most common counterion in preparative peptide purification, has essentially no absorbance above 220 nm at the concentrations involved. Triisopropylsilane, dithiothreitol, and other cleavage-cocktail scavengers absorb weakly or not at all in the analytical window. Sugars used as lyoprotectants in the final formulation — mannitol, trehalose, sucrose — are optically transparent across the entire ultraviolet range used for peptide work. A chromatogram can be clean at 214 nm while the solid in the vial contains substantial mass of material the detector never registered.

Amino acid analysis and quantitative NMR address part of this by measuring peptide content directly, and the gap between peptide content and chromatographic purity is precisely where this non-absorbing mass lives. What those techniques do not provide is a chromatographic picture — a separation in which the non-peptidic material appears as resolved peaks with retention times, so that its identity and number can be examined rather than inferred from a mass balance.

How aerosol detectors respond

Both ELSD and CAD begin the same way. Column effluent is nebulized into a fine aerosol, the mobile phase is evaporated in a heated drift tube, and what remains is a stream of dry analyte particles suspended in gas. Everything after that step is independent of chromophores, because the measurement is being made on solid particles rather than on molecules in solution.

ELSD passes that particle stream through a light beam and measures scattered light. Scattering intensity depends on particle size and number density, and the relationship to analyte mass is characteristically non-linear — response typically follows a power law with an exponent between roughly 0.7 and 1.8 depending on the analyte and the nebulization conditions. Calibration curves are therefore fitted in log–log space, and the dynamic range over which a single fit remains valid is modest, often around two orders of magnitude.

CAD replaces the optical step with a charging step. The dried particle stream is mixed with a corona-charged nitrogen flow, charge transfers to the particles in proportion to their surface area, and the resulting aggregate charge is measured by an electrometer. Because charge transfer scales with particle surface area rather than with scattering cross-section, CAD’s response is closer to linear than ELSD’s — still typically a power function, but with an exponent nearer unity and correctable in software on most modern instruments. Sensitivity is generally an order of magnitude better than ELSD, with detection limits in the low nanogram range on column for well-behaved analytes.

The property both share, and the reason they are called universal, is that response depends on the mass of non-volatile residue and only weakly on what that residue is. Reported response factors across chemically unrelated compounds typically fall within a factor of two to three, compared with the orders of magnitude that separate a tryptophan-containing peptide from mannitol at 280 nm. Uniform response is an approximation, not an identity — but it is a far better approximation of mass than ultraviolet absorbance is.

What this changes on a purity chromatogram

Running the same peptide sample with UV and aerosol detection in series produces two chromatograms that agree on peptide-related peaks and diverge everywhere else, and the divergence is the informative part.

Peaks present in CAD or ELSD but absent in UV are non-absorbing material: counterion-associated species, scavenger residues, formulation excipients, inorganic salts that survive nebulization. Their retention behavior is usually unhelpful — much of this material elutes at or near the void volume — but their presence and approximate magnitude are established rather than assumed.

The reverse case is equally instructive. Peaks large in UV and small in aerosol response are low-mass, strongly absorbing species: aromatic protecting-group fragments such as those released from Fmoc chemistry, or small aromatic impurities that inflate a 214 nm chromatogram out of proportion to the mass they represent. A purity figure computed by UV area percent is penalized by these species more than a mass-basis accounting would justify.

Neither observation says the UV number is wrong. Chromatographic purity by UV area percent is a defined, reproducible, and useful measurement, and comparability across laboratories depends on everyone computing it the same way. The point is that it answers a narrower question than its name suggests — what fraction of the ultraviolet-absorbing material eluting from this column is the target peak — and orthogonal detection makes the boundaries of that question visible.

Practical constraints

Aerosol detection imposes requirements that ordinary peptide methods often violate. Both detectors are destructive, so they sit at the end of a split or downstream of a UV cell. Both require fully volatile mobile phases: phosphate buffers, non-volatile ion-pairing agents, and involatile salts accumulate in the drift tube and produce baseline drift or persistent background. Peptide methods using trifluoroacetic acid are compatible in the sense that TFA is volatile, but TFA itself generates substantial background signal and suppresses response, which is one reason formic-acid or ammonium-formate methods are often preferred when aerosol detection is planned.

Gradient elution introduces a second complication. Nebulization efficiency depends on mobile-phase composition, so response for an identical mass of analyte changes across an acetonitrile gradient — typically increasing as organic content rises. Quantitative work across a wide gradient therefore requires either an inverse-gradient compensation flow or acknowledgment that comparisons are semi-quantitative. For the qualitative purpose of asking whether non-absorbing material is present and roughly how much, this is tolerable. For assigning a number to a certificate, it is a real limitation.

Volatile analytes are also invisible by construction. Anything lost with the mobile phase in the drift tube does not reach the detector, so residual solvents remain the province of headspace gas chromatography and water remains the province of Karl Fischer titration. Universal detection is universal only among non-volatile species.

Where it fits

For routine lot release of a well-characterized peptide with an established formulation, RP-HPLC with UV detection at 214 nm, supported by mass spectrometric identity confirmation and a peptide content assay, covers the analytical ground that matters. Aerosol detection adds most where the composition of the non-peptidic fraction is genuinely unknown: characterizing material from a new synthesis route, investigating a discrepancy between peptide content and chromatographic purity, evaluating a formulation whose excipient profile has not been independently confirmed, or examining a lot whose mass balance does not close.

Read that way, CAD and ELSD are less an alternative to ultraviolet detection than a way of asking what ultraviolet detection has been leaving out. The two chromatograms are answers to different questions, and the interval between them is a measurable quantity rather than a matter of inference.