What is causing the harmonics at non-whole numbers? How does bad impedance matching etc cause this?
I was wondering if you have any examples with a a spectral density chart showing these other harmonics, I was having some trouble recalling the name of what I was looking for 
The details would need you to re-take RF Engineering course 
But short answer is that when there is impedance mismatch, there is RF leakage, this RF leakage is what generates the harmonics. Think of it as a leak in a water pipe.
I do not have spectral analyzer data that shows this for Zigbee vs WiFi. but I have a lot of data to show this for cell phone towers, which unfortunately I cannot share. But the same RF principles apply everywhere.
And for the less technical folks trying to follow along:
I had a professor who liked visual metaphors.
Standard washer and dryer right next to each other.
Put running shoes in the dryer so that it’s going thunk, thunk, thunk.
Now watch the bottle of fabric softener on the washer next to it start walking across the top of the washer and then fall off.
That’s sideband lobe interference. 
(It’s obviously not a perfect metaphor, but I like it because it’s easy to remember and helps explain why sometimes the problem is the device next to the one that has the symptom.)
And for those who want a technical but not too technical explanation (again with graphs), the following paper is also good.
http://www.metrictest.com/catalog/white_papers/Fundamentals%20of%20Interference%20App%20Note.pdf