Reading a chromatogram

This lesson is a fixed routine for reading the trace, detector signal against time, and the peak table printed under it. It uses the terms from Purity: what HPLC measures without explaining them again. With practice, the seven steps take a couple of minutes and tell you whether a chromatogram supports the purity printed above it.

What are you looking at on the axes?

Step 1: read the axes. The horizontal axis is time in minutes from the start of the run. The vertical axis is detector response, often in milli-absorbance units (mAU) for a UV detector. The vertical scale is usually set automatically so the main peak fills the frame, which squashes small peaks near the baseline. A vertically zoomed view, sometimes shown as a second panel, is more useful for seeing impurities.

Note the run time. A trace that stops shortly after the main peak cannot show anything that elutes later.

Which peaks are not from the sample?

Step 2: find the baseline and the non-sample peaks. Near the very start there is usually a disturbance as the sample solvent reaches the detector, often called the solvent front or void. Gradient methods can also show small system peaks that come from the mobile phase, not from the sample. Laboratories identify these by running a blank, the same method with no sample, and comparing. ICH Q2(R2) describes using blank samples or "signals in an appropriate baseline region" when estimating detection and quantitation limits (ICH Q2(R2)).

A good report either excludes blank peaks from the table and says so, or shows the blank trace alongside. If a peak at 1.5 minutes appears in both the blank and the sample, it should not be counted as an impurity, and its absence from the table is correct rather than suspicious.

Where is the main peak, and does it look right?

Step 3: locate the main peak and its retention time. It is normally the tallest peak. If the report includes a reference standard run, the main peak's retention time should match the standard's under the same method, a real but limited identity check.

Step 4: look at its shape. A well-behaved peak is narrow and close to symmetrical. Watch for three shapes:

• A shoulder. A bump on the side of the main peak, where it fails to return smoothly to the baseline. This often means a closely eluting impurity that was not fully separated.
• Tailing. A long, slow return to the baseline after the peak. It makes the end point of integration harder to place and can hide a small impurity in the tail.
• A flat top. The peak is cut off at the top because the detector was overloaded. Areas from a saturated peak are not reliable.

The EMA guideline on synthetic peptides warns about "the risk of co-eluting impurities" and says that when they are observed as one peak, the 1.0% qualification threshold applies to them unless justified (EMA, synthetic peptides guideline, 2025). A visible shoulder on a trace reported at 99.5% purity deserves a question.

What do the small peaks tell you?

Step 5: list the small peaks and where they sit. Note each one's retention time relative to the main peak. Peaks just before or just after the main peak are often near copies of the peptide, such as deletion sequences. Peaks far away are more likely to be unrelated material or degradation products.

Ask whether the small peaks are large enough to measure reliably. ICH Q2(R2) states that a signal-to-noise ratio of 3:1 is generally acceptable for the detection limit, and at least 10:1 for the quantitation limit. A bump only slightly taller than the baseline noise can be seen but not reliably measured, which is why methods report only peaks above a set threshold, 0.1% for synthetic peptides under the Ph. Eur. as cited in the EMA guideline.

Does the peak table agree with the picture?

Step 6: read the peak table. Typical columns are peak number, retention time, area, height and area percent. Check that every visible peak above the threshold has a row, that the rows are in time order, and that the main peak's row is the one with the reported purity.

Step 7: check the arithmetic. Add the area percent column. It should total 100, within rounding. Divide the main peak's area by the sum of areas and confirm it matches the reported purity. The chromatogram tool does both checks from the numbers you type in.

Can you spot the problems? A worked example

This peak table, invented for teaching, is attached to a report that claims 99.4% purity. Work through steps 6 and 7 before reading the answers below it.

• The arithmetic. 492,300 divided by 499,300 is 98.60%, which matches the table, not the 99.4% on the front page. The report's headline does not match its own data.
• Peak 4. At 12.5 minutes it sits just after the main peak (relative retention time about 1.03) and is 0.92%, close to the 1.0% qualification threshold. On the trace, check whether it appears as a separate peak or a shoulder.
• Peak 1. At 1.4 minutes, very early, with no area percent. If a blank run shows the same peak, excluding it is correct. If no blank is shown, ask.

Nothing here proves the sample is poor; 98.6% may be a perfectly good result. The problem is that the headline cannot be traced to the data, which is a fault in the report, not necessarily in the vial.

What other warning signs appear on chromatograms?

Published methods show the detail a full method description carries: column, dimensions, flow rate, temperature and detection wavelength, such as the C18 column, 1.5 mL/min, 25°C and 220 nm described for a leuprolide reference standard (McCarthy et al., Pharmaceutical Research, 2023). A buyer's report will not carry all of that, but it should carry enough to show the trace belongs to the sample. The HPLC-UV method page lists what a complete method description includes, and the retatrutide page shows what a standard test reports for that compound. Next, the same habit applied to the other raw output: Reading a mass spectrum.

Check your understanding

1. A peak at 1.3 minutes appears in both the blank and the sample run. Should it count as an impurity? No. It comes from the system or solvent, not from the sample.
2. The peak table's area percent column adds up to 101.6. What does that suggest? A transcription or integration error. The table should total 100 within rounding.
3. A main peak has a visible bump on its trailing side. What might that be? A closely eluting impurity that was not fully separated, which may be counted as product and raise the purity figure.