I’ve been tracking a batch for 6 weeks. The vendor’s public COA lists 97.8% by HPLC-UV at 220 nm, one main peak at 11.2 min. I split the same vial and sent two aliquots to independent labs. Lab A: 93.2%, plus an extra peak at 8.4 min, 2.1% area.
Lab B: 96.9%, no extra peak, but their method runs 15 min with 0.1% TFA. Same sample, different columns and gradients. Is that 8.4 min peak real or a method artifact? What raw data would you ask both labs for—chromatograms, blanks, spiking?
That 8.4 min peak smells like method. Ask for a blank injection and DAD scan before blaming the sample.
I ran into almost the same split: COA 98.2%, lab A 94.1%, lab B 97.6%. Turns out lab A used 214 nm and counted a shoulder, while lab B used 220 nm with a 0.5% inetgration threshold.
Once I got the raw chromatograms, the missing peak was there but below their cutoff. Ask both labs for integration events, baseline zoom, and whether they report area normalization or external standard. The purity number alone won’t settle it. Also check if their system suitability passed on the day.
A drifting baseline can create a small ghost peak that looks like 2% area.
Lower purity isn’t automatically correct. Need the blank.
Column dimensions matter. 150 x 4.6 vs 100 x 2.1 shifts dead volume. Compare k' values.
Six weeks unfrozen? That 4.6% drop could be real.
If Lab B saw no extra peak, their LOQ might be above 2.1%. Ask for LOQ and a spike at that level.
I’d ask for forced degradation. If the 8.4 min peak grows with heat or light, it’s real. If it shifts with mobile phase pH, method artifact.
I did a quick heat stress on my retained sample: the 8.4 min peak grew from 0.6% to 2.3% after 18 h at 60 °C. That felt less like an artifact and more like something I should chase. Now I’m re-running with a photo diode array to check peak purity.
Yep, I plotted k' against lab and my 100 x 2.1 column gave different elution windows than their 150 x 4.6. The extra bump at 8.4 min matched the void marker on theirs? Mine had it at 7.9. Dead volume chat first, purity chat later.
Asked Lab B for LOQ and a spike at 2.1%. They sent method summary but no raw data yet. I’m not accepting 'no peak' until I see the baseline zoom and their integration threshold. Diary note: follow up Monday.
Blank first. Always. My blank had a ghost at 8.4 min; changed everything.
214 vs 220 nm changed my shoulder area by 0.9%. Integration threshold 0.5% hid it completely. I’ve started reporting both area normalization and external standard so I can compare apples to apples.
System suitability passed, but the tailing factor was 1.8 on the day. I keep wondering if that shoulder was partially co-eluting. I’m asking for the suitability chromatogram, not just the pass/fail line.
Six weeks unfrozen would worry me too. I’d run a fresh aliquot and a frozen control side by side before blaming the method. If the drop tracks storage, that’s real sample behavior, not an HPLC artifact.
Tried mobile phase pH 3.0 vs 4.5. The 8.4 min peak shifted to 7.6 min at higher pH, so I’m leaning artifact for that one. The main peak stayed put. Now I need to check ionic strength too.
Login to reply to this thread.
Login / Sign up