On a GTO point of view you should only play "fold or shove" with this stack depth. But GTO is only optimal against players who either play GTO themselves or who "circle around GTO".
On a practical, exploitative point of view, against a weak player who limps hands like 68o 9bb deep, you most likely played the hand perfectly. He happened to outdraw you, but you need to consider his whole range, the profit you will make when he hits top pair or can't fold second pair, etc
The flop check is good too usually on this rather dry board. Let him get draws on the turn with his air and also let him try to steal turn. Both will happen way more often than him outdrawing you. Ev check usually > ev flop bet here (would be different with more depth though).
The analysis posted above on this matter is very shallow.
You have to realize in particular that against the specific 68o hand that the SB has, the ev of a flop bet and of a flop check is (relatively) close. The case were you lose ev when delay cbeting against this specific 68 hand is when he doesnt hit his str8 neither on the turn nor on the river, and doesn t bluff turn by check raising your delay cbet nor donk shove river his missed draw. Then you lose the second bet (if you bet flop he would have called and called the turn barrel no matter what the turn is). He will miss only 66% of the time or so, unless he s very fit or fold he will crai turn sometimes, or do stuff like donk shove bluff river, etc..
Also, he rarely has a good draw on the flop on this dryish board. Only 2oesd, no flush draws.
Don't over estimate protection value. Here your protection value also comes from geting gutshots to fold, either possibly on the flop cbet or on the turn barrel when unimproved, and also from not letting second pair and other bluff catcher see a river or third pair 5out you turn (I assume second pair will call the flop cbet but rarely the turn shove - unless "sucked out" like a flopped Q on turn A/K). The ev obtained from inducing turn bluffs from your opponent (which can be high unless he's a 100% passive and/or fit or fold fish) and the ev obtained from letting him improve to a draw or a hand weaker than aces is much higher than the protection value.
Important to note, if the SB has a Q, betting flop achieves nothing special, there is no protection value in it, the hand will go all in no matter what the turn and river are(I doubt he fold a Q even if the turn is a K or A) and the action choice between cbet flop or delay cbet has no impact on ev at all.
By checking you may also lose a bit of value against hands like second pair who could call a bet on flop but might fold vs a delay cbet when the turn is another overcard. But in compensation you also sometimes get looked up by weaker made hands when delay cbetting on a brick (like third pair peeling one bet, or 2nd pair deciding to call both turn+river bricks just because he expected you to cbet strong hands).
Basically the more passive and less likely to bluff your opponent is, the better it gets to bet flop rather than check, but on average, against a random fish limper in a donkament, I would expect the check to have more value (with only 2bets depth).
