From a molecule with two nitrogen atoms to benzoyl chloride would seem to be quite a segué!
I believe that you were misled by the peak at 3490 cm-1, which you ascribed to the N-H stretch in a secondary amine, R2N-H.
Note that the IR spectrum was run with the sample present at a relatively high concentration, so much so that the very intense C=O absorption peak at 1757 cm-1 runs "off the chart". In this situation, one frequently observes "ovetones" in the spectrum that otherwise might be too weak to notice.
Note also that the errant (and much less intense) 3490 cm-1 absorption (overtone) is roughly twice the frequency of the fundamental absorption (1754 cm-1).
The situation is similar with musical overtones. As an example: With your right index finger hold down the piano key that corresponds to the C above middle C (i. e., simply lift the hammer off the piano string, thereby leaving that string free to vibrate). Then, with your left index finger strike the middle C piano key sharply. If the piano is in relative tune, along with the fundamental (261.6 Hz) you will hear the first overtone (frequency = 2 x 261.6 Hz). With both musical and spectroscopic examples the overtone is not as intense as the fundamental due to exponential decay.