That's true. Rather have a player that can contribute 5v5 in multiple situations and provide versatility. However, perhaps Khaira can be utilized in his natural position at C? He has mentioned that he is more comfortable as a center as opposed to playing on the wing. I would want to look into the numbers regarding this.
As for 1st vs. 2nd pk unit deployment, your statements may be a general assumption unless there is data to back that up. I assume that the most important aspect of any PK (aside from the obvious system in place) is consistent line changes and fresh legs. As such, 1st and 2nd PK units will still face top opposition at a similar rate. I could be totally wrong about this too though and deployment rates variating between the two units is significantly difference.
My point wasn't a matter of deployment Vs top opposition (though there's an element of that); it was that the 1st unit is more likely to start PKs. This is for a number of reasons including fresher legs from fewer 5-on-5 mins. Also, this initial deployment is more often going to be against the top PP unit (to the point you raised) which means they're likely the best the opponent has at establishing and maintaining offensive pressure. So if you're more likely to:
a.) start the PK (for a variety of reasons), as such,
b.) have conceded the zone by default on a FO loss rather than the opponent having to work to gain it, and
c.) be starting the PK and therefore, be more likely to face the opponent's best offensive players (for maintaining off. zone pressure)
Then it stands to reason that you'll have a harder time and will concede more goals than the unit that more often comes on in relief during play/between icings. To illustrate the basis for my point, here are the # of def. zone faceoffs our main PKers were on-ice for on the PK:
Sheahan 179 (151mins)
Archibald 169 (139mins)
Nuge 56 (106mins)
Khaira 39 (98mins)
Draisaitl 67 (56mins)
Russell 27 (33mins)