The inability of standard density functionals to correctly predict charge-transfer excitation bands is a well-known problem. This shortcoming has been attributed to the incorrect treatment by the functionals of the long-range exchange potential. Accordingly, a number of long-range corrected schemes have been proposed to deal with this situation. We have applied the CAM-B3LYP variant to the zincbacteriochlorin-bacteriochlorin complex, a system which highlights this problem, and have found a significant improvement in the calculation of its excitation bands over other functionals. In particular, the spuriously low charge-transfer states found previously do not appear for CAM-B3LYP. (c) 2005 Elsevier B.V. All rights reserved.