The radioisotope 182 Hf (t 1/2 =8.9Ma) is of great interest for astrophysical applications as a chronometer for the early solar system or as possible live supernova remnant on earth. However, AMS measurements of 182 Hf are seriously influenced by the presence of the stable isobar 182 W, which cannot be separated at typical AMS energies. Previous studies revealed a possible suppression of 182 W against 182 Hf by extracting the negatively charged pentafluoride HfF5- from the ion source, leading to a detection limit for 182 Hf/ 180 Hf in the order of 10 −11 . However, this suppression behavior is in contrast to theoretical calculations of the electron affinity and recent measurements using SIMS instruments, where the achieved suppression cannot be reproduced. The aim of our study is to determine the effects of ion source background as well as further investigate the suppression of tungsten against hafnium by extracting negatively charged fluoride ions from different sample materials. The previously reported suppression factor of about 6000 could be increased to 36000 by careful tuning of the ion source using HfF 4 as sample material. The trend of the theoretical electron affinities could be reproduced using atomic tungsten and hafnium instead of HfF 4 as sample material. This supports the assumption that the major contribution of the tungsten background is not sputtered from the target matrix but comes from somewhere else in the ion source. Measurements from the second ion source show a higher background of tungsten and a lower suppression factor, i.e. careful design of the ion source is crucial. Moving the sputter beam over the target surface extending over the wheel holding the targets revealed the highest tungsten background was detected outside the sputter target position. Further investigations are necessary to locate the origin of the tungsten background in the ion source. Possible sources are the material used for the ion source construction or contaminations in the cesium used for sputtering.