differential inductor

Got some problems with the measurement of differential inductor ? Recently there are more and more papers describing the advantages of using differential inductor. Like Q enhancement, smaller area etc..
And papers always show the Q difference in differential and sigle-ended mode operation for an inductor, is that from measurement or just simulation? If it's from measurement, then how could we send the differential signal into the "DUT" if it's two port only ? I mean, of course you can try to send a differential signal by using balun ..., but how could you measure the two port S parameters at the same time? Or it must be a
three port system (with center tap to third GSG) so that you can do measuremnt on differential driven ? Any suggestion ?

I think that you can simply see the differential inductor as a single ended one with a center tap.
To have a model use the simmetry respect to center tap, so you can do your measurements with normal two port parameters and center connected to gnd. As you have the 2-port S-paramters you can "measure" the differential characteristics of the inductor on a simulator using ideal balun.
Hope it can help.
Mazz