Searching geometries' optimization for hydrogen storage systems

I have a problem on geometries optimization in Orca software. Here http://www.scientific.net/JNanoR.5.113 the author reports the binding energies for everyone of the compounds C4H4 and C5H5 with metals Sc, Ti and V and hydrogens from 1h2 to 5h2. The results that i obtained are not as good as i would like. I don't really know which is the problem. My input files look like this one...

! PAL8 ! PBE SlowConv opt 6-31g

% scf Damp fac 0.8 ErrOff 0.02 Min 0.1 Max 0.99 end MaxIter=5 Convergence tight end

%geom MaxIter 150 MaxStep 0.7 TolRMSG 1e-5 TolMaxG 3e-4 TolRMSD 2e-4 TolMaxD 4e-4 end

C -0.863882000 1.159379000 -0.787529000 C 0.602862000 1.173569000 -0.914692000 C 0.728530000 1.264876000 0.549616000 C -0.737951000 1.249837000 0.677176000 H -1.690132000 1.249997000 -1.488771000 H 1.294372000 1.285972000 -1.746060000 H 1.550317000 1.459624000 1.234634000 H -1.434551000 1.440432000 1.489825000 V -0.042977000 -0.667265000 -0.009272000 H -0.265419000 -1.153706000 1.733377000 H 0.563867000 -1.201048000 1.628150000 H 1.581940000 -1.234345000 -0.607194000 H 1.660882000 -1.281386000 0.226254000 H 0.192778000 -1.316728000 -1.692013000 H -0.633826000 -1.396229000 -1.572196000 H -1.726923000 -1.344892000 -0.169167000 H -1.654126000 -1.188132000 0.652049000
 * 1) ---Este es un comentario---
 * xyz 0 2

Energies are obtained like the difference on energies between no hydrogens binded and 1 h2 binded plus h2's energy. And that way for everyone of the hydrogen binding energies..

The problem is that there are some values in which my calculations converge on energies but that is not true for all energies. I don't know how to search ground states in optimizations... May be, the problem could be solved that way...

Can anyone help me please??