Novel catalysts for dechlorination of polychlorinated biphenyls (PCBs) and other chlorinated aromatics

Article abstract of DOI:10.1039/b705756c

Diiron complexes of fluorene and fluorene* (1,2,3,4,5,6,7,8,9- nonamethylfluorene) have been found to be catalysts for the dechlorination of chlorinated aromatics, such as PCBs. The Royal Society of Chemistry.

Full text of DOI:10.1039/b705756c

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Novel catalysts for dechlorination of polychlorinated biphenyls (PCBs)
and other chlorinated aromatics{
Andrew E. D. Fletcher, James Moss, Andrew R. Cowley and Dermot O’Hare*
Received (in Cambridge, UK) 17th April 2007, Accepted 6th June 2007
First published as an Advance Article on the web 22nd June 2007
DOI: 10.1039/b705756c
into efficient systems for treating low level PCB contaminated
waste.
Diiron complexes of fluorene and fluorene* (1,2,3,4,5,6,7,8,9-
nonamethylfluorene) have been found to be catalysts for the
dechlorination of chlorinated aromatics, such as PCBs.
Given that the reaction proceeds via reduction of the metal we
have been screening a range of electron rich metal complexes in
order to try and discover new active catalytic systems. For
example, 19- and 20-electron iron g-arene species have been shown
previously to be powerful reducing agents but we found them to be
inactive catalysts for PCB dechlorination, since the ligand is
reduced by NaBH4.8 However, in the course of these investigations
we investigated the activity of two novel electron-rich
di-iron fluorene complexes, [(FeCp)2FluH]2+[PF6]2; (1)9 and
[(FeCp)2Flu*H]2+[PF6]2; (2)10 (Flu = g6-C13H10; Flu* = g6-
C13Me9H), Fig. 1. Electrochemical studies on both 1 and 2 show
they both exhibit two well resolved 1-electron reduction steps
(DE = 300 and 350 mV for 1 and 2 respectively) which
demonstrate substantial electronic communication between the
two iron centres mediated through the fluorene ligands. Here we
report the use of 1 and 2 as catalysts for the dechlorination of both
PCBs and other chlorinated aromatics. As far as we are aware this
is only the second example of a catalytic transition metal based
dechlorination system. The only previous report using iron
catalysts was the high temperature dechlorination using colloidal
iron particles.11
About a third of all persistent and toxic pollutants are chloro-
organic compounds,1 and collectively they represent a major
global environmental hazard. The worst examples are the toxic
polychlorinated biphenyls (PCBs). A total of 750,000 tonnes of
PCBs (as 60–70 congener mixtures) were produced between 1929
and 1976 when they were banned by international treaty. Since
then levels have been monitored in many locations around the
globe and PCBs have been shown to be still ubiquitous and
persistent. Their chemical inertness, which was once one of their
industrial advantages due to their resistance to attack by acids and
bases, now makes them such significant and persistent environ-
mental pollutants. Methods now exist for their destruction in bulk
quantities, usually by very high temperature incineration.
However, significant quantities of PCBs still reside in contami-
nated effluent, soils or sediments and to date there is no cost
effective chemical method for their destruction.
Although Kagan’s reagent (SmI2) is relatively efficient at
dechlorinating PCBs giving only mono or dichlorinated biphenyl,
it requires 1.7 equivalents of SmI2 per equivalent of PCB and is
only most effective when used with hexamethylphosphoramide
(HMPA), which is highly toxic.2 PdCl2(dppf) (dppf = 1,19-
bis(diphenylphosphino)ferrocene) uses very mild conditions, but
very little dechlorination of PCBs is observed even after 4 days
reflux.3
The crystal structure of 2[BF4]2 is shown in Fig. 2.{ Crystals
were grown by slow diffusion of diethyl ether into a saturated
solution of 2[BF4]2 in nitromethane. The crystal structure is similar
to the Flu0H analogue with the principal difference being that the
methyl substituent on the capping carbon (C1) causes the ligand to
twist at an angle of 15.5 degrees. When viewed as a space filling
model it becomes apparent that this change in geometry is
necessary to accommodate the methyl substituent.
In 1995 Liu and Schwartz reported the first catalytic molecular-
based dechlorination of chlorinated aromatic hydrocarbons. They
reported that titanocene dichloride (Cp2TiCl2; Cp = g-C5H5)
could be used as a catalyst for the dechlorination of PCBs
producing biphenyl as the only organic product and at the
relatively low temperature of 125 uC.4 They also showed that this
system would be catalytically active towards chlorinated benzenes.5
Using a reductive process rather than an oxidative process, such as
when using Fenton’s reagent,6 had the advantage that no harmful
side products could be produced by incomplete reaction. Knowles
and co-workers demonstrated that this approach could in principle
be applied to contaminated soils, although there were major
catalyst compatibility issues to be overcome.7 There is a clear need
to discover new catalytic systems so that they can be developed
Both 1 and 2 were tested as catalysts for the dechlorination of
Aroclor 1242, an industrial mixture of PCBs.§ The conditions
chosen were similar to those used by Schwartz and co-workers.
0.75 mmol of catalyst was combined with 19.2 mmol NaBH4,
20 mmol pyridine and 1.24 g of Aroclor 1242 (equivalent to
15 mmol of Cl). 27.5 cm3 of diglyme was used as solvent and the
reaction mixture heated to 125 uC overnight. Since many Fe arene
Chemistry Research Laboratory, Department of Chemistry, University
of Oxford, Mansfield Road, Oxford, UK OX1 3TA.
E-mail: dermot.ohare@chem.ox.ac.uk; Fax: 44 1865 285131;
Tel: 44 1865 285130
{ For crystallographic data in CIF format (CCDC 644165) and
experimental results see DOI: 10.1039/b705756c
Fig. 1 Structures of 1 and 2.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 2971–2973 | 2971

View Article Online
45.6% and 35.0% respectively. Schwartz has proposed that the
Cp2TiCl2 catalysed dechlorination of PCBs proceeds via a radical
process. We attribute the the non-quantitative conversion of all the
PCB congeners to biphenyl as due to coupling of the dechlorinated
hydrocarbons during the reaction. For comparison, tests using
Cp2TiCl2 under the same conditions gave an average biphenyl
yield of 55.7%. However, for both 1, 2 and Cp2TiCl2 no PCBs
could be detected in the reaction mixture by GC chromatography.
The lifetime of 2 as a PCB dechlorination catalyst was
investigated by doubling the quantities of pyridine, reducing agent
and Aroclor 1242 present in the reaction mixture. The same
quantity of catalyst was able to continue dechlorination for
48 hours. At this point the catalyst was no longer active and only
monochlorinated biphenyls and biphenyl could be detected. Under
the same reaction conditions Cp2TiCl2 remained active for much
longer. Cp2TiCl2 was able to dechlorinate 6 times the amount of
Aroclor 1242 used in a standard test (when NaBH4 and pyridine
were also increased proportionally) and was still functioning
after 336 hours, at which point the test was stopped. Calculating
a turnover number for these catalysts is difficult, since the number
of cycles completed until the catalyst ceases working depends on
the degree of chlorination of remaining molecules. However, it can
be estimated that 2 is able to complete approximately 100
dechlorination cycles per molecule whereas after 336 hours
Cp2TiCl2 had completed around 300 dechlorination cycles.
1,2,4,5-Tetrachlorobenzene is often used as simple mimic for the
complex PCB congener mixtures. Chlorinated benzenes have
similar reduction potentials to PCBs and so should be dechlori-
nated under similar conditions. We found that only 2 was effective
at dechlorinating 1,2,4,5-tetrachlorobenzene. The reason for 1
being unreactive is as yet unclear. The dechlorination pathway
could be followed qualitatively using GC chromatography. The
dechlorination was shown to proceed stepwise and via 1,4-
dichlorobenzene and 1,2-dichlorobenzene. 1,3-dichlorobenzene
was only detected in trace amounts. This can be rationalised by
considering the localisation of the negative charge during the
reduction.12 The colour changes for the reaction were the same as
on PCBs above. The identical behaviour of 2 towards these two
substrates would suggest potential applications in dechlorinating
other chlorinated aromatics. Similarly 1 could also be used,
however, it is less reducing than the methylated analogue, so one
would not expect it to have such wide applicability. It should be
possible, due to the stepwise nature of the dechlorination to
partially dechlorinate other polychlorinated substrates by quench-
ing the reaction after a predetermined time.
Fig. 2 Crystal structure of [(FeCp)2Flu*H](BF4)2; 2. Thermal ellipsoids
2
drawn at 50% probability, H atoms and BF4 counterions omitted for
clarity.
complexes are slightly light sensitive, tests were run both in light
and in darkness.
For both 1 and 2 after 24 hours no PCB could be detected in the
reaction mixture by GC chromatography. Biphenyl was the only
observed organic product. Initially the reaction mixture was
orange, and over 24 hours it alternated between orange and dark
brown, the final mixture being dark brown in colour. It is
proposed that this indicates the alternation between the Fe(II) and
Fe(I) states. A proposed reaction mechanism is shown in Fig. 3.
NaBH4 readily reduces one iron centre to Fe(I). This extra electron
can then be transferred to the PCB molecule in an analogous
fashion to the mechanism proposed by Schwartz for Ti(III). The
resulting PCB radical anion may then lose Cl2 before abstracting
2
2
?
H? from BH4 . The resulting BH3
is then scavenged by
complexation with the pyridine, as was demonstrated using 11B
NMR by Schwartz for the Cp2TiCl2 case.4
Quantitative analysis of PCB destruction indicated slight
differences in the catalytic performance between 1 and 2. The
methylated analogue, 2, was more active, giving an average yield
of 49.0% biphenyl compared to 39.1% for the non-methylated
analogue, 1. The yield dropped slightly in both cases when the
reactions were performed under normal laboratory lighting to
In conclusion, [(FeCp)2FluH][PF6]2; 1 and [(FeCp)2Flu*H]
[PF6]2; 2 are rare examples of molecular organotransition metal
compounds that are able to catalytically reductively dechlorinate
commercial PCB mixtures.
2 is also able to catalytically
dechlorinate tetrachlorobenzene. The redox activity and the mixed
valance nature of these complexes seem to be important
contributing factors to their activity.
We would like to thank the EPSRC for the award of a graduate
studentship to AEDF.
Notes and references
˚
{ Crystal data for 2[BF4]2: monoclinic, space group P21/n, a = 9.5417(3) A,
3
˚
˚
˚
b = 26.6042(10) A, c = 12.3888(4) A, b = u, V = 2994.81(18) A , 23181
measured reflections, 6378 independent (Rint 0.063). 477 parameters, R =
Fig. 3 Proposed catalytic cycle for dechlorination of PCBs.
2972 | Chem. Commun., 2007, 2971–2973
This journal is ß The Royal Society of Chemistry 2007

View Article Online
0.0785, wR = 0.0713. The data were collected at 150 K on a Enraf-Nonius
Kappa CCD diffractometer with graphite-monochromated MoKa radia-
tion. Intensity data were processed using the DENZO-SMN package.13a
Structure was solved using SIR92.13b Subsequent full-matrix least-squares
refinement was carried out using the CRYSTALS program suite.13c
§ In a typical dechlorination reaction a prepared mix of Aroclor 1242 (or
1,2,4,5-tetrachlorobenzene), pyridine and diglyme was added under N2 and
the reaction was heated to 125 uC and stirred. The chlorinated aromatic,
base and solvent were dried/degassed as appropriate using standard
techniques. 1 ml aliquots were taken at 4 h and 24 h for Aroclor 1242 and
1, 2, 4 and 24 h for 1,2,4,5-tetrachlorobenzene. These aliquots were
quenched immediately in 1.5 ml water and then stored in test tubes at
280 uC until analysis. Samples were thawed before diluting with 100 ml of
a 0.0025 mol dm23 of 1,2,4-trichlorobenzene in CHCl3. 1 mL of the organic
layer was then injected into a Finnigan TraceGC ultra with FID using a
25 m non-polar SGE BPX-05 column. The column was held at 60 uC for
4 min before ramping at 10 uC min21 up to 200 uC. This temperature was
then held for a further 4 min. Peak integrations were analysed by Thermo
Electron Corporations Chrom-Card data system Ver. 2.3. The results
quoted are averages of multiple runs.
2 S. A. Jackman, C. J. Knowles and G. K. Robinson, Chemosphere, 1999,
38, 1889–1900.
3 L. Lassova, H. K. Lee and T. S. A. Hor, J. Org. Chem., 1998, 63,
3538–3543.
4 Y. Liu, J. Schwartz and C. L. Cavallaro, Environ. Sci. Technol., 1995,
29, 836–840.
5 Y. Liu and J. Schwartz, Tetrahedron, 1995, 51, 4471.
6 D. L. Sedlak and A. W. Andren, Environ. Sci. Technol., 1991, 25,
1419–1427.
7 M. A. Wright, C. J. Knowles, J. Stratford, S. A. Jackman and
G. K. Robinson, Int. Biodeterior. Biodegrad., 1996, 38, 61–67.
8 P. Michaud, D. Astruc and J. H. Ammeter, J. Am. Chem. Soc., 1982,
104, 3755–3757.
9 Q. Dabirmanesh, S. I. S. Fernando and R. M. G. Roberts, J. Chem.
Soc., Perkin Trans. 1, 1995, 743–749.
10 J. Moss, J. Thomas, A. Ashley, A. R. Cowley and D. O’Hare,
Organometallics, 2006, 25, 4279–4285.
11 M. Takaoka, A. Fukatsu and N. Takeda, Toxicol. Environ. Chem.,
2000, 76, 95–109.
12 L. Lassova, H. K. Lee and T. S. A. Hor, J. Mol. Catal. A: Chem., 1999,
144, 397–403.
13 (a) Z. Otwinowski and W. Minor, Methods Enzymol., 1997,
276, 307–326; (b) A. Altomare, G. Cascarano, C. Giacovazzo and
A. Guagliardi, J. Appl. Crystallogr., 1994, 27, 1045–1050; (c)
P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout and
D. J. Watkin, J. Appl. Crystallogr., 2003, 36, 1487.
1 I. R. Danse, R. J. Jaeger, R. Kava, M. Kroger, W. M. London,
F. C. Lu, R. P. Maickel, J. J. McKetta, G. W. Newell, S. Shindell,
F. J. Stare and E. M. Whelan, Ecotoxicol. Environ. Saf., 1997, 38,
71–84.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 2971–2973 | 2973