#hplc-massspec 2026-05-20

shallow gradients hide related substances behind the main peak. a steeper one separates and looks worse, honestly

i have a number at about a third of what i expected, is that a charge state

do you trust an area percent from a shallow gradient

need more sample

show me the spectrum

you need more sample for identity than for purity. plan for that when you post it

a certificate with a number and no chromatogram is asking you to trust the integration you cannot see

*Janoshik not the other one

is a tailing factor of anything specific a problem

a shoulder on a peak is information, and usually bad information, anyway

monoisotopic and average diverge more the bigger the molecule. say which one you are quoting

update on the earlier thing how would a truncation sequence show up

thats a charge state

why does integration change the number so much

i asked Janoshik for the raw trace and they sent it. that is the right answer to that question

retention time is a hint

214nm sees the amide backbone so it sees everything. 280 only sees aromatics

somebody explain how a 99.6% peak can be hiding a 2% impurity because i keep reading that and it sounds like nonsense

it is not nonsense but it is often stated badly. two things get mixed together

so salts and solvents just do not appear

essentially. acetate has almost no absorbance up at 214 to speak of. you will not see your counterion on a UV trace and that is exactly why the mg per vial number needs a separate test

the counterion thing is the bridge between this channel and #coa-reading, people bounce between the two arguments without noticing

ok co-elution i follow. how would you detect it

change the selectivity and see if the peak splits.
shallower gradient, or a different stationary phase, or swap TFA for formic acid in the mobile phase.
or run it into a mass spec and look at whether the mass across the peak is constant.
if the leading edge and trailing edge have different masses you had two things

that last one is the good test and it is why LC-MS beats LC alone for this

peak purity by diode array is the cheap version. same idea, spectra across the peak instead of masses

and baseline drift, is that the same problem

different problem, same consequence. a rising baseline during the organic ramp means the integrator has to decide where zero is, and small peaks near the end of the run vanish into it

same raw file, integrated twice

so comparing 99.6 from one lab against 99.2 from another is meaningless

at that resolution yes. 99.6 vs 94 is meaningful. 99.6 vs 99.2 is noise plus operator

the rule of thumb people here use is one to two points of area% is inter-lab noise and anything inside that is a tie

that rule of thumb has been argued about since 2024 and nobody has landed it properly. treat it as a convention not a fact

fair. it is a convention that keeps people from starting fights over rounding

the TFC lots i had done came back 98.7 and 99.4 from two services and i genuinely could not tell you which vial was better

were they the same batch

same batch, split sample, mailed on the same day

then you have a nice little inter-lab experiment and the answer is they agree

did not feel like agreement at the time

it never does

what about the Saber report someone posted with a completely flat baseline and four peaks total, is that suspicious

not inherently. a clean sample on a good system looks boring. but a trace with no noise at all, i mean literally none, is worth a second look because real detectors have noise

and check the y axis scale. you can flatten anything by zooming out

of course you can. brilliant

the axis trick is the most common way a real trace gets used dishonestly. nothing is faked, you just cannot see the small stuff

GL Biochem traces i have include the axis values which is basic and yet not universal