Negdo's answer is 100% correct. This is meant as a "but if you really want to anyway..." follow-up.
The main challenge with designing an encounter for this party is almost certainly action economy. One monster gets one turn's worth of actions per round; the party gets 9! Any way you slice it, that's a lot of opportunity for the players to work together (which they should be doing!) to put the enemy in the worst position possible, and that enemy has precious little ability to respond. Second to that is HP economy: the players can typically force the single boss to spend a lot more of its HP per round than that boss can force the players to spend (at higher level, HP can become less important as spells can disable enemies without directly affecting their HP).
For a boss encounter (especially a solo one), this GM would lean heavily on Legendary Actions and Lair Actions, as well as interesting terrain features.
Both Legendary and Lair actions help with the action economy, letting the foe get a few more actions in per round. They allow the boss to both act and react (narratively, not just by using their reaction). With a good selection of Legendary/Lair actions, the single boss has at least some ability to act even if they lose initiative and the rest of the party looses everything they've got!
Legendary/Lair actions also help with the HP economy if they have the chance to damage PCs and/or heal the boss (though, be a little careful of the latter: I've seen plenty of players get annoyed when foes had the gall to actually use their healing abilities on themselves).
Interesting terrain features - ideally - encourage the party to change up their tactics a bit, round-by-round. I've played (and played with) rogue archers whose bread-and-butter is "move, cunning action hide, sneak attack" or "steady aim, sneak attack" every round: it's effective, but it's also repetitive. Heck: I've run/played combat encounters in 3.5, PF1, 4e, and 5e (and that's just the direct D&D lineage games) in which nobody moved after the first round! So: make the terrain dynamic - maybe there are stalactites that fall, changing sight lines and creating difficult terrain (and making people make DEX saving throws to get out of the way without getting bonked on the head!); maybe the floor is fragile, and there's a chance that it gives out when a creature stands on it for a full round.
So, why not throw out a CR number as a baseline? The dirty little secret of CR is that it is - at the absolute best - a very rough guideline, suggestive that an enemy of CR X is probably easier for a "standard" party to deal with than an enemy of CR X+1. Even with that limitation, there are still a million and one discussions about whether creature X's CR is right, especially when creature Y's CR is higher. It's a starting place, but leagues away from the destination of "good encounter" or "challenging encounter".