Are negative peaks in HPLC always a problem?

No. With a refractive-index detector (RID) negative peaks are often normal: the detector measures the RI difference versus the mobile phase, so an analyte with a lower RI than the eluent gives a peak that points down. On a UV detector a negative peak usually does signal something to fix, a sample-diluent mismatch, the wrong wavelength, or a vacancy peak.

Why do I get negative peaks with a refractive index detector?

An RID compares the refractive index of the column effluent against a reference of pure mobile phase. When an eluting band has a lower refractive index than the mobile phase, the signal drops below baseline, a negative peak. It is the expected physics, not a fault: in the same run some analytes can be positive and others negative.

How do I fix negative peaks caused by the sample solvent?

Dissolve the sample in the mobile phase, or a weaker solvent than the mobile phase, and reduce the injection volume. A diluent that absorbs differently from the eluent disturbs the baseline near the solvent front and can push early peaks negative; matching the diluter removes the disturbance.

Can a UV detector show negative peaks?

Yes. If the analyte absorbs less than the mobile-phase background at the detection wavelength (for example with a UV-absorbing additive in the eluent), the band depletes the background and reads negative. A DAD reference wavelength that overlaps the analyte band can also subtract into a negative excursion.

← Back to blog

Troubleshooting 7 min read Jun 30, 2026

Negative peaks in HPLC: causes and fixes

TL;DR: A negative peak means the detector signal dropped below the baseline as the band eluted, the band gave less response than the mobile phase. On a refractive-index detector that is often normal physics; on UV it usually points to the sample diluent, the detection wavelength, or a vacancy peak. Read which one by its fingerprint, then apply the matching fix.

What a negative peak actually is

Every detector reads a baseline set by the mobile phase and reports the difference a band makes as it passes. A normal (positive) peak is "more signal than the mobile phase". A negative peak is the opposite: for a moment the eluting zone produces less signal than the eluent it displaced. So the question is never "why is it upside down", it is "why does this band read lower than the background". Four answers cover almost every case.

Cause 1, Refractive index (RID): usually expected

A refractive-index detector compares the refractive index of the column effluent against a reference cell of pure mobile phase. If an eluting analyte has a lower refractive index than the mobile phase, the signal dips below baseline, a negative peak. This is the detector working correctly, not a fault. In one isocratic run you can legitimately see some analytes positive and others negative depending on each one's RI relative to the eluent.

Fingerprint: only happens on RID (or other bulk-property detectors); direction depends on the analyte, and reverses if you change the mobile-phase composition.

Fix: nothing to "fix" if the peak integrates, just integrate it as a negative peak. If you need everything pointing up, change the mobile phase so its RI sits below all analytes, or switch to a detector that responds to the analyte directly (UV, ELSD, MS).

Cause 2, Sample diluent stronger or different from the mobile phase

If you dissolve the sample in a solvent that absorbs differently from the mobile phase (or is simply a stronger eluent), the injected plug disturbs the baseline. You get a negative dip near the solvent front and distorted, sometimes negative, early peaks.

Fingerprint: the disturbance sits at or just after t₀; worst for early-eluting peaks; changes if you change the injection solvent or volume.

Fix: dissolve the sample in the mobile phase (or a weaker solvent than the mobile phase) and reduce the injection volume so the diluent does not overload the head of the column.

Cause 3, The analyte absorbs less than the mobile phase (UV)

If the mobile phase carries a UV-absorbing additive and your analyte absorbs less at the detection wavelength, the band depletes the background absorbance as it passes, a real negative peak at the analyte's retention time. The same logic powers indirect UV detection on purpose.

Fingerprint: negative peaks at true analyte t_R on a UV/DAD; the eluent has an absorbing component; flips to positive if you move to a wavelength where the analyte absorbs more than the background.

Fix: choose a detection wavelength where the analyte absorbs more than the eluent, remove the absorbing additive, or, on a DAD, check that the reference wavelength does not overlap the analyte's absorbance.

Cause 4, Vacancy (system) peaks

When the sample is missing a component that is present in the mobile phase (a buffer salt, an additive), the injected plug locally dilutes that component. As that "hole" travels and elutes, the detector sees less of the additive, a negative vacancy peak at a fixed time.

Fingerprint: appears at fixed times tied to mobile-phase components, not to your analytes, and shows up even in a blank injection.

Fix: match the sample matrix to the mobile phase (add the same additive to the diluent), or simply recognise and ignore the system peak once identified with a blank.

See it without an instrument. In the free PureAnalyt HPLC simulator the refractive-index detector can show inverted peaks depending on the analyte and mobile phase, so you can watch when a negative peak is expected chemistry rather than a fault, and compare it against the same sample on UV or MS.

A quick decision checklist

Only on RID, direction depends on analyte? → refractive index. Expected; integrate it.
Dip at the solvent front, early peaks worst? → sample diluent. Dissolve in mobile phase, inject less.
Negative at the real analyte t_R on UV, absorbing eluent? → wavelength/background. Change λ or remove the additive.
Fixed times, present in a blank? → vacancy peak. Match the matrix or ignore it.
Whole trace mirrored? → detector zero/polarity. Re-autozero and check settings.

A negative peak is just information pointing down. Match the pattern, decide whether it is real chemistry (RID) or an artefact (diluent, wavelength, vacancy), and you will know in seconds whether to fix it or simply integrate it. For the broader picture, see how to read an HPLC chromatogram and why HPLC peaks tail.

Open the HPLC simulator → Free · no signup · runs in your browser