The attached graph gives a ...

The attached graph gives a reasonable indication of aluminium precipitation against pH. It's not my work but at the moment I can't remember where it came from. In any case its one I've used a lot in the past to explain why either high or low coagulation pH is bad in terms of aluminium solubilisation or resolubilisation.

I've also found that the exact pH limits depend entirely on water chemistry - so 6.4 might be optimum for one water but not for another. We dose acid at one of our plants to optimise the coagulation pH and maximise organics removal and have found that, in that water, any pH below 6.0 results in elevated soluble aluminium levels. Therefore we aim for a pH of about 6.2. Similarly a pH prior to filtration of 7.5 or higher can result in resolubilisation of aluminium from flocs.

As mentioned in previous answers undertaking a series of jar tests at different pHs can be used to work out where the limits are for your particular water chemistry.