Márcia M. C. Ferreira (PQ)
Instituto de Química - UNICAMP, 13083-970, Campinas - SP
marcia@iqm.unicamp.br
 

          The PAH (Polycyclic Aromatic Hydrocarbon) compounds have been for a long time  a
focus of great  attention  by  the  scientific  community  due  to  their   impact  on  health  and
environment.    Some   of   these    compounds    such    as    benzo[a]anthracene,   chrysene,
dibenzo[a,h]anthracene  and  benzo[a]pyrene  are  mutagens   and  carcinogens.  Usually,  the
PAHs are introduced into the environment as  a result of anthropogenic  activities which  have
increased dramatically  in the  last  20  years,  the major source being associated with coal  tar.
PAHs  have  been  detected  in  the  atmosphere,  waters,  soils,  sediments  and  food.  Their
physicochemical properties such  as  aqueous  solubility, volatility  (e.g. Henry's  law constant,
air/wate partition coefficient),  hydrophobicity  or  lipophilicity (e.g. n-octanol/water  partition
coefficient)  vary  widely.  These compounds  are  also  quite  involatile,  show relatively  low
vapor pressure and resistance to chemical reactions.  As  a  result,  their distribution in aquatic
systems,  the  atmosphere  and  soil  differ  significantly.  Also,  PAHs  show  a   tendency  to
accumulate in biota, soils, sediments,  and to be transported through the atmosphere far  away
from  their  source  discharge.
          So, meaningfull  health  and  environmental assessment requires reliable data  on  their
physicochemical properties.  Unfortunately,  very often these information are not available  in
the literature,  and  frequently  the  reported data have poor precision  and accuracy  specially
when   dealing   with   hydophobic  and   involatile   chemicals.   These   are   mainly  due  to
experimental  diffculties  from  procedures  such  as  preparing,  handling  and  analyzing  the
solutions  when  low  concentration  is  involved  (on trace levels - ppb).
          However,  with  the  advent  of  inexpensive  and  rapid computation there has been a
remarkable   growth   interest   in   the   area   of   QSPR   (Quantitative   Structure-Property
Relationships), which uses multivariate methods to model relevant properties as a function of
molecular structure parameters (called descriptors). A large number of descriptors have  been
proposed in the literature.  The  nature  of  the  most commonly used  descriptors  (structural,
topological,  electronic  and  geometric)   and  the  extent   to  which  they  encode  structural
molecular features  that are related to some specific physical property,  are at the heart of any
QSPR  study.
          This work focuses on a  subgroup  of  48 PAH  compounds composed exclusively  by
nonsubstituted  six-membered  aromatic  rings.  In  a  first  step,  physicochemical  properties
relevant for environmental studies such as  the boiling temperature (Tb),  the  retention index
(RI),  n-octanol/water  partition  coefficient  (Kow)  and solubility  (S)  are  related  with  the
descriptors  (X  matrix).    In  a second  step,   these  properties  were  incorporated  into  the
descriptor matrix to build several  quantitative  structure-property  relationships  and to obtain
prediction rules for the soil sorption coefficient (log Koc),  bioconcentration factor (BCF) and
Henry  law  constant  or  air-water  partition  coefficient.

((CNPq)