While it is true a canister is more efficient at Co2 usage and a tank from what I hear can last for years, maintaining Co2 @ correct levels is possible without much Bps in a sump environment if
proper diffusion method is used, I know this because I am doing it. I am maintaining a limegreen DC with only 5Bps and my setup consists of a 90gal, a CPR CS150 overflow in which I have placed a Durso standpipe, not to retain Co2 but too reduce noise and a 30gal sump. The water from my tank is not directed underwater, it is filtered through a micron sock which gets dispersed over another layer of coarse filter media, then finally over a Biowheel and I have no special covers to retain Co2.
IMO I think the sump / Co2 loss explained above is exaggerated and you do not lose much Co2 as stated. If I can maintain a limegreeen DC with 5Bps on a 90gal how much less is needed in a canister setup?
While I will suggest a canister filter for ease of use over a sump don't be discouraged in using your sump on your tank as needed Co2 levels can be maintained. Take a look at this
Reactor, this is what I am using currently with excellent results and prior to this I was misting with a Rhinox 5000.
Now the important thing with a sump setup is the final flowrate that is used 'tank turnover rate' this should be around 600gph with
less being better, anything over and you are asking for trouble, remember this is not 'saltwater' :-D.
Here are some plants in my 90gal sump setup....see my sig also........
When I was misting with Rhinox as you can see
healthy unstunted growth
Using a sump on a planted tank is possible as I am using one now and I have no fancy setup to prevent loss of Co2.