According to my experience, the behavior you are seeing could be a result of combined isotherm and transport behaviors. First, ACT isotherm params can have an impact on the boarding of the peak, please refer to Figure 10 (b) 3 in my paper for an example. One thing you could try is to fit isotherm params closer or within to your loading density range to fine-tune it, see if it improves. Second, transport part is usually more complex, there might be contributions from both pore and surface diffusion. The boarding of the peak could just be because the diffusivity is decreased when q is too high, especially when they are released at the same time during elution, see this topic for some discussions. Another possibility could be surface diffusion, I also recommend playing around with it and see if it helps.
If you want to isolate these two, I recommend doing static binding experiments at different pHs to fix ACT isotherm params. And then in the model you can play around with transport params.