Slight variation in NMR integrals of CH protons vs. CH2 or CH3

Question

While acquiring some NMRs I find that there is a consistent pattern in that protons attached to sp² carbons tend to have slightly smaller integrals than protons attached to sp³ carbons:

(peaks in aromatic region are from a para-disubstituted benzene ring; peak at 4.92 ppm is from a trisubstituted alkene; 4.50 is a NH; the remainder are aliphatic CH₂'s or CH₃'s.)

I calibrated the total sum of integrals to 21; one can see from the spectrum that the aromatic and olefinic peaks are slightly smaller than integral values, and the others are slightly larger. Obviously it doesn't hinder the analysis of the spectrum, but it was an interesting trend (the same was observed in multiple spectra).

Am I reading too much into this, or is there a reason behind it? I strongly suspect there is something - perhaps related to the marginally slower relaxation of these protons, since there are no geminal protons (only vicinal protons)?

If it is of relevance - a 60° pulse is used (pulse program zg60). The number of scans ns is 16, and the relaxation delay d1 is 1 second. I'm happy to provide any other necessary acquisition parameters, or to provide the full structure of the compound if it's necessary.

---

P/S I found something in Findeisen & Berger's 50 and More Essential NMR Experiments, in which they mention that the integral of an aromatic proton in strychnine is smaller than expected as it has the longest relaxation time. Even though the authors used ns = 1, too short a delay between the receiver gain adjust and the scan was used, which led to incomplete relaxation and hence a smaller integral.

Answer 1

That's to be expected: Less protons as spatially close neighbours, less dipolar coupling, slower relaxation.

For a relaxation/recycle delay of $\pu{1 s}$ and $T_1$ of perhaps $\pu{3 s}$, the Ernst angle is already only $44^\circ$, so your measurement is far from optimal anyway. Use a shorter pulse or longer recycle delay.

At one second, your are anyway dangerously close to generating echoes which overlay the signal from further scans, distorting the accumulated spectrum.