Dear Christian, indeed it is related to the cut_occ_states option. If this option is false you also see the occupied states at negative energies. This is very useful in many case. If now you want to cut the occupied states, simply rerun xspectra using the cut_occ_states=.true. option and the spectrum will be smoothly cut out at the e_f you have specified. Note also that you don't need to perform again the full lanczos process but only to replot the curve, namely use the option xonly_plot=.true., in the namelist &input_xspectra provided you have kept the .sav file with all the informations on the previously performed lanczos process, namely the a and b coefficients. If you give a look at the diamond example in the examples/XSpectra_examples of the distribution then you will find all the informations you need.
Finally, if the cut_occ_states option takes too much time then there are other parameters to tune that can drastically reduce the time needed. To understand how cut_occ_states works you can look at Phys. Rev. B 54, 7334?7349 (1996) sections III A and III b. However I suggest you do not touch the parameters related to cut_occ_states unless you know what you do. All the best, M. -- * * * * Matteo Calandra, Charge de Recherche (CR1) Institut de Min?ralogie et de Physique des Milieux Condens?s de Paris Universit? Pierre et Marie Curie, tour 16, case 115 4 Place Jussieu, 75252 Paris Cedex 05 France Tel: +33-1-44 27 52 16 Fax: +33-1-44 27 37 85 http://www.impmc.jussieu.fr/~calandra