DEHALOGENATION OF CHLOROARENES WITH SODIUM DIHYHDRIDOBIS(2-METHOXYETHOXO)ALUMINATE IN THE PRESENCE OF TRANSITION METAL COMPOUNDS

Article abstract of DOI:10.1135/cccc19941645

Dehalogenation of low-chlorinated arenes such as p-dichlorobenzene of chlorobenzene with the title hydride is accelerated in the presence of transition metal species formed in situ from the corresponding 2,4-pentanedionates.Their efficiency decreases in the order: Co = Ni = Pd> Cu >> Mn > Fe which results from the changes of their activity and stability.The dehalogenation is well described by a kinetic model consisting of the set of dehalogenation steps which are first order in the chloroarene combined with the catalyst deactivation which is second order in the transition metal compound.

Full text of DOI:10.1135/cccc19941645

Dehalogenation of Chloroarenes
1645
DEHALOGENATION OF CHLOROARENES WITH SODIUM
DIHYDRIDOBIS(2-METHOXYETHOXO)ALUMINATE
IN THE PRESENCE OF TRANSITION METAL COMPOUNDS
Jaroslav VCELAK and Jiri HETFLEJS
Institute of Chemical Process Fundamentals,
Academy of Sciences of the Czech Republic, 165 02 Prague 6-Suchdol, The Czech Republic
Received November 12, 1993
Accepted January 13, 1994
Dehalogenation of low-chlorinated arenes such as p-dichlorobenzene or chlorobenzene with the title
hydride is accelerated in the presence of transition metal species formed in situ from the correspond-
ing 2,4-pentanedionates. Their efficiency decreases in the order: Co ≈ Ni ≈ Pd > Cu >> Mn > Fe
which results from changes of their activity and stability. The dehalogenation is well described by a
kinetic model consisting of the set of dehalogenation steps which are first order in the chloroarene
combined with the catalyst deactivation which is second order in the transition metal compound.
Group I – III metal hydrides are in general poor agents for dehalogenation of nonacti-
vated low-chlorinated arenes such as chlorobenzene, chloronaphthalene etc. In several
cases, however, their efficiency substantially increased upon addition of transition met-
al salts. Such a promoting effect has been observed in the MgH2 dechlorination of
chlorobenzene with NiCl2, Ni(PPh3)4 (ref.1) and, to the lesser extent, also with other
metal chlorides2 (FeCl3 > CrCl3 > MnCl2 > CoCl2, TiCl3 . 3 THF >> VCl3) and in the
NaBH4 dehalogenation of bromo- and chlorobenzenes with NiCl2(PPh3)2 (ref.3), using
the above metal compounds in substoichiometric (catalytic) amounts with respect to
both the hydride and the substrate. Although NaH was found to be very poor reagent
for the more reactive bromobenzene4, the reaction has been readily effected in the
presence of Pd(II)-(long chain ether-phosphane) complexes as catalysts5.
On the contrary, in dehalogenations with LiAlH4, only stoichiometric amounts of
transition metal chlorides (NiCl2 > FeCl2 >> CoCl2) promoted chlorobenzene dechlori-
nation6. A similar situation has been recently found7 also in 4-chlorobiphenyl dechlori-
nation with NaBH2(OCH2OCH3)2 where approximately 1 equivalent of the substrate
could be dechlorinated per equivalent of NiCl2, demonstrating thus a noncatalytic
course of the reaction. When compared to the above hydrides, the behaviour of the
aluminum analogue, NaAlH2(OCH2OCH3)2 (SDMA), might be a promising exception,
based on the strong accelerating effect achieved8 in the dehalogenation of aryl halides
in the presence of catalytic amounts of Pd(II)-benzonitrile complexes.
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

1646
Vcelak, Hetflejs:
Our continuing interest in chemical destruction of noxious polychlorinated hydrocar-
bons9–11 led us recently to the study of kinetics of dehalogenation of a series of chlo-
robenzenes with SDMA (ref.12). The slow dehalogenation rate and incomplete
dechlorination of the low chlorinated arenes formed during the multistep dehalogena-
tion process resulted in a serious obstacle against the use of this dehalogenating agent
for the above mentioned purpose. However, in view of other convenient properties of
SDMA (high solubility in aromatic hydrocarbons, safe manipulation) and of its ability
to act as a very efficient activating agent in formation of transition metal hydrogenation
catalysts (cf. ref.13 and references therein), it seemed to us useful to explore this possi-
bility. The results of our attempt to accelerate dehalogenation of the least reactive chlo-
roarenes (chlorobenzene and 1,4-dichlorobenzene) by transition metal compounds as
promotors are reported briefly in the present work.
EXPERIMENTAL
Chemicals
Sodium dihydridobis(2-methoxyethoxo)aluminate, 70% solution in toluene (trade mark Synhydrid,
Synthesia Kolin) was used after the actual hydride concentration had been determined iodometri-
cally14. The solvents (toluene, dioxane, both analytical purity grade) were dried over sodium and re-
distilled. Chlorobenzene, 1,4-dichlorobenzene (chemical purity grade, Lachema Brno) and metal
2,4-pentanedionates (acac) (Ni(acac)2, Co(acac)2, Pd(acac)2, Co(acac)3, Cu(acac)2, Fe(acac)2, and
Mn(acac)2 – all Fluka chemical purity preparations) were used as obtained. Bis(benzonitrile)palla-
dium(II) dichloride was taken from laboratory stock.
General Dehalogenation Procedure
A thermostatted reaction vessel equipped with a thermometer and a reflux condenser was filled with
dry toluene or dioxane (10 ml). Then, 0.5 M toluene solution of 1,4-dichlorobenzene or chloroben-
zene (2 ml) were added, followed by the metal pentanedionate dissolved in toluene (3 ml). The mag-
netically stirred solution was warmed up to the reaction temperature at which SDMA (60% toluene
solution) was added in one portion. In all experiments SDMA was taken in threefold molar excess
with respect to the chlorine of the chlorobenzene to be dehalogenated. At regular intervals, 0.5 ml
samples of the stirred reaction mixture were withdrawn, the unreacted hydride immediately decom-
posed by aqueous 2 M HCl. After phase separation, the toluene layer was subjected to GLC analysis
(Hewlett–Packard 5890 instrument equipped with a capillary glass column (25 m × 0.25 mm)
covered by OV-101 silicone elastomer, programmed temperature from 35 °C to 200 °C, calibration
by authentic samples, except higher boiling by-products which were calibrated with the use of biphe-
nyl as a standard). The corresponding GC calibration parameters (A, B) are presented in Table I. The
structure of dehalogenation by-products was examined by GC-MS method, using a Varian 3500 ca-
pillary gas chromatograph coupled with a Finnigan MAT ITD 800 ion trap mass detector (EI, 70 eV),
equipped with a LEO PC 386 SX for data processing with a NIST library of 42 000 spectra (for
details of measurement conditions see ref.12).
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

Dehalogenation of Chloroarenes
Treatment of Experimental Data
1647
Time–concentration dependences for both the starting chlorobenzenes and their dehalogenated pro-
ducts obtained by GLC analysis were treated in terms of a simple model of stepwise dehalogenation,
written for 1,4-dichlorobenzene in the following way:
k1
k2
1,4-Cl2C6H4
ClC6H5
C6H6
III
I
II
The analysis showed that the catalyst activity is also time dependent, decreasing steadily during the
reaction. This fact was incorporated into the kinetic scheme of dehalogenation which had then the
following form:
d[cat]/dt = −kdeact [cat]y
(1)
(2)
(3)
(4)
d[I]/dt = −k1 [I]x [cat]
d[II]/dt = k1 [I]x [cat] − k2 [II]x [cat]
d[III]/dt = k2 [II]x [cat] ,
where kdeact, k1, and k2 are apparent rate constants of the catalyst deactivation and of respective deha-
logenation steps depicted in the above reaction sequence (I to II to III), x and y are reaction orders
in the catalyst and the substrate. The SDMA concentration, [SDMA], is not written in the rate equa-
tions, being considered as constant and included in the respective rate constants (dehalogenations
were carried out with the hydride in sufficient excess to justify this assumption). Because of variable
properties of the catalyst systems with the metal compound to SDMA molar ratio, the reaction order
in SDMA had not been determined and was thus expressed in the rate constant dimension as un-
TABLE I
Characteristics of data treatment
GLCa
[SDMA]d
Compound
Rb, %
kerrorc, rel.%
mol l-1
A
B
C6H6
1.01
1.19
1.82
0
–
–
6
5
–
ClC6H5
0.024
−0.027
57
61
0.81
0.81
1,4-Cl2C6H4
a
Parameters of GLC calibration curves, where [RX]actual = (PRX/(PST + PRX)) [A + B (PRX/(PST
+
PRX))] (weight concentration). b R Maximum conversion to which the given dehalogenation was fol-
c
lowed and the obtained results treated. The relative error of the rate constant determination.
d SDMA concentration used in the measurements (SDMA/2 : C6H6 − nCln /n molar ratio was always 3 : 1).
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

1648
Vcelak, Hetflejs:
known (i.e. k1, k2 in [SDMA]−x min−1, kdeact in [cat]−1 [SDMA]−x min−1). The set of four differential
equations written above was solved numerically according to the Runge–Kutta method with simulta-
neous optimation of the equation parameters to experimental data with the use of the modified Mar-
quardt method15. The computations were performed with variable x and y quantities (0, 1, 2) and the
fit of the resulting three rate constants to experimental data for the whole course of the reaction was
evaluated by comparing the values of the mean standard deviations. Based on these data, the deha-
logenation under study is well described for the set x = 1, y = 2. Maximum conversions (R, %) to
which dehalogenation of a given chloro compound was followed and obtained data used in rate con-
stant computation are given in Table I, along with the average relative error in rate constant determi-
nation (kerror, in rel.%). The rate constants of dechlorination of both chlorobenzenes are listed in
Table II.
RESULTS AND DISCUSSION
Our earlier studies (cf. refs13,16–18) on formation of transition metal catalysts by the
reaction of metal precursors with SDMA showed that properties of such catalysts de-
pend strongly on a number of reaction parameters (solvent, reaction temperature, the
metal to SDMA molar ratio, the sequence of mixing catalyst components, the control-
led aging and stabilization, nature of metal precursor etc.), optimation of which is
largely empirical. As expected, a similar situation has been observed also in the deha-
logenations with analogous metal hydride–transition metal systems (cf. refs7,8).
For that reason we have examined first several different ways of preparation of these
dechlorinating agents, and evaluated their efficiency based on the l,4-dichlorobenzene
conversion achieved under identical reaction conditions. To avoid the ill reproducibility
of the two phase activations taking place on using insoluble transition metal precur-
sors18, we have chosen the soluble metal 2,4-pentanedionates. Preliminary experiments
with SDMA–Ni(acac)2 showed that the most convenient treatment is the one step re-
duction of Ni(acac)2 with SDMA in the presence of the dehalogenated compound at the
temperature of dechlorination reaction. This treatment was thus taken as a general pro-
cedure for preparing other SDMA–transition metal dechlorination systems studied (see
Experimental).
In order to characterize quantitatively the efficiency of the systems studied, some
preliminary experiments concerning kinetics of l,4-dichlorobenzene dehalogenation
were performed, using Ni(acac)2. Dependence of the initial rate of the dehalogenation
on the Ni chelate concentration is presented in Fig. 1. The curves indicate a complex
metal action. At first sight they resemble an analogous dependence reported by other
authors8 for dechlorination of 1-chloronaphthalene with SDMA–PdCl2(C6H5CN)2 (cf.
Fig. 2 in the quoted work). One of possible reasons of such behaviour was thought to
be agglomeration of metal species to bulkier metal particles with smaller specific sur-
face8 (and thus reduced dehalogenation activity). To obtain further information, we
have analyzed both the reported dependence8 and the above curves in log–log form.
While any statistically significant relation has not been found for the SDMA–Pd deha-
logenation system, the rate data for SDMA–Ni give linear correlation with different
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

Dehalogenation of Chloroarenes
TABLE II
1649
Rate data and product composition for dehalogenation of l,4-dichlorobenzene (I) and chlorobenzene
(II) with SDMA–transition metal dehalogenation agents (for conditions see Experimental)
Products, %
c
Metal
compound
[cat]
YRX
%
b
b
b
Entry
kdeact
k1
k2
. 102,a
by-
product
II
III
1,4-Dichlorobenzene
1
2
3
4
5
6
7
8
9
–
–
–
0.0011
0.95
0.76
0.24
3.7
0.0014
0.12
0.17
0.089
1.4
15
51
67
58
34
65
70
39
49
64
62
60
78
50
45
14
13
41
44
35
27
47
54
–
2
3
7
6
2
3
8
–
–
–
1
1
4
0
0
1
0
–
7
PdCl2(C6H5CN)2
PdCl2 + 2 PPh3
PdCl2 + 4 PPh3
Ni(acac)2
3.3
3.4
3.2
4.5
1.7
16
16
6
0.25
0.61 32
Ni(acac)2
3.8
8.1
1.5
2.7 ~10−4
4.1 ~10−4
5.7
1.6
0.18
14
9
Ni(acac)2
4.8
23
1.4
0.26
~10−4
~10−4
~10−4
d
Ni(acac)2
4.7
–
d
Ni(acac)2
0.27
0.12
1.3
–
–
d
10 Ni(acac)2
–
–
11 Co(acac)2
12 Co(acac)2
13 Co(acac)2
14 Co(acac)3
15 Cu(acac)2
3.8
3.5
3.6
2.0
4.9
2.8
9.5
7.1
0.05
46
57
57
44
44
11
9
15
2
e
f
1.0
0.04
3.5
0.23
17
6
1.4
~10−4
~10−4
0.04
~10−4
3.4 ~10−4
0.15
0.093
0.11
1
16 Mn(acac)2
17 Fe(acac)3
3.6
2.1
1.3
3.0
2
8.9
1
Chlorobenzene
18
–
–
–
–
–
–
5.9 . 10−4
10
29
68
–
–
–
–
–
13
14
19 Co(acac)2
20 Co(acac)2
3.0
2.9
0.89
5.7
0.17
1.9
16
53
f
a Expressed as molar fraction. b Eqs (1) – (4), kdeact in [cat]−1 [SDMA]−x min−1, k1 and k2 in lx min−1
c
mol−x. YRX denotes conversions of I and II after 3 h reaction. d The incremental addition of Ni com-
pound; conversions of I after l h (entry 8), 2 h (entry 9), and 3 h (entry 10). e Dioxane as the solvent.
f At boiling point of the solvent (110 °C).
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

1650
Vcelak, Hetflejs:
slopes for both periods followed (0.41 (r = 0.975) for 10 min and 0.31 (r = 0.972) for
30 min). The variable fractional order to the metal demonstrates that the catalytic
properties of the formed metal species are subject to time changes, which should be
incorporated to a kinetic model of the dehalogenation. Treatment of experimental data
obtained with Ni(acac)2 and the other metal pentanedionates studied (for details see
Experimental) yielded the best fit for the model involving the set of dehalogenation
steps which were first order in the chloro compound, combined with the second order
catalyst deactivation. This model (Eqs (1) – (4)) described well the course of the deha-
logenation reaction, as demonstrated on dichlorobenzene dehalogenation in Fig. 2 as an
example. The corresponding rate constants, along with the product composition at the
end of the dehalogenation are collected in Table II.
Prior to discussion of the rate data for individual transition metal systems, the
composition of dehalogenation products deserves mentioning. As demonstrated in the
Table II, in all the experiments except the noncatalytic dehalogenation, the expected
products of hydrogenolytic cleavage (chlorobenzene, benzene) are accompanied by
several minor side products. Their total amount is expressed in Table II in mole %
determined with the use of biphenyl as the standard (see Experimental). Irrespective of
the metal chelate used as a metal precursor in 1,4-dichlorobenzene dechlorination, GC-
MS analysis confirmed the presence of 4 responses corresponding to the compounds of
the general formula C13H11Cl. Comparison of their mass spectra with those recorded in
the instrument library suggested the four possible types which could not be further
differentiated. These were: x-methyl-x′ -chlorobiphenyl (IV) and (x-chlorophenyl)-
phenylmethane (V) isomers, diphenylchloromethane (VI) or x-(chloromethyl)biphenyl
(VII) isomers.
x-ClC6H4CH2C6H5
x-CH3C6H4C6H4Cl-x′
V
IV
C6H5CH(Cl)C6H5
x-ClCH2C6H4C6H5
VI
VII
The products had relative elution volumes (with respect to toluene) within 3.10 – 3.38,
their relative distribution being 5, 30, 35, and 30 per cent. In spite of the already
stressed structural uncertainty, these products all would arise from homolytic cleavage
of the aryl−Cl bond of the chloroarene followed by interaction of the radicals formed
(ClC6H5 for IV and V, and Cl• for VI and VII) with the solvent. This cleavage is surely
•
mediated by the transition metal, as similar products have not been found in the SDMA
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

Dehalogenation of Chloroarenes
1651
dehalogenations of chlorobenzenes without addition of transition metal compounds in
both the previous12 and this work. Participation of the aromatic solvent in side products
formation seems to be confirmed by the strong suppression of this process on using
dioxane as the solvent (compare entries 13 and 14, Table II). The complete absence of
these products could not be expected, as also here the hydride was used as its toluene
solution.
Similarly, GLC analysis of the products of chlorobenzene dechlorination with some
SDMA–transition metals yielded two responses corresponding to the compounds of the
general formula C13H12, suggesting formation of diphenylmethane (VIII) or methylbi-
phenyl isomers (IX), based on their mass spectra. Their relative elution volumes were
2.85 and 2.94 with respect to toluene, with the relative distribution 65 and 35 per cent.
The rate data in Table II show that the overall effect of transition metal species
formed in SDMA–metal precursor systems on the course of the dehalogenation results
C6H5CH2C6H5
x-CH3C6H4C6H5
VIII
IX
FIG. 1
FIG. 2
Dependence of the rate of dechlorination, r
(mol l−1 min−1), of 1,4-dichlorobenzene with
SDMA–Ni(acac)2 on Ni(acac)2 concentration,
[cat] (mmol l−1). (For reaction conditions see
General Procedure in Experimental.) 1 For 30
min period, 2 for 10 min period
Comparison of experimental (points) and com-
puted data (curves) for SDMA dechlorination
of 1,4-dichlorobenzene (1) to chlorobenzene
(2) and benzene (3) (in dioxane, 70 °C,
Co(acac)2 – entry 12 in Table II, X stands for
molar fraction). Mean standard deviations of
experimental points are 0.033 X (1), 0.027 X
(2), and 0.008 X (3)
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

1652
Vcelak, Hetflejs:
from both activity and stability of this species. Provided that the model proposed is not
formal, then deactivation processes take increasing relative importance due to the
higher power function with respect to metal precursor concentration (Eq. (1)). This can
be documented, for example, on the results obtained with PdCl2(C6H5CN)2 (entry 2)
and those aimed at stabilizing palladium species by addition of PPh3 during its forma-
tion (entries 3 and 4). They reveal that the increased conversion of l,4-dichlorobenzene
achieved with the PdCl2(C6H5CN)2 + 2 PPh3 system (67% vs 47%) results from the
significant suppression of catalyst deactivation (kdeact almost 3 times lower in the latter
case) while the overall dehalogenation activity remains practically the same (cf. com-
pensation of k1 and k2 changes). Further addition of PPh3 changes both parameters
analogously, resulting thus in the decrease of its efficiency (entry 4).
Several experiments aimed at evaluating the role of reaction conditions showed that
at least for the nonstabilized metal system (Ni(acac)2, entries 5 – 7), the higher activity
of the system prepared with the metal precursor in the lower concentration (entry 5) is
overweighed by its low stability (cf. kdeact 31.5 in entry 1 with e.g. 4.8 in entry 7). As
expected, the rate od deactivation increases with increasing temperature (cf. entries 11
and 13), and seems to depend on the solvent (cf. entries 11 and 12). In this connection
it is worth mentioning that in the recent study of dehalogenation of chlorobiphenyls
with NaBH2(OCH2CH2OCH3)2–NiCl2 (ref.7), the higher yields were obtained by por-
tionwise addition of NiCl2 during the reaction. However, the already discussed high
kdeact found for the metal species formed at lower metal precursor concentrations (com-
pare entries 5 and 7), which would be the case of incremental addition, speaks against
the applicability of this method in our case. The results obtained in such an experiment
using Ni(acac)2 (entries 8 – 10) did confirm this conclusion. Although the analysis of
the above data allowed us to specify the role of different reaction parameters, it has not
contributed much to the deeper understanding of this process. This question is under
further investigation.
In conclusion, the efficiency of the metal chelates studied deserves mentioning. The
l,4-dichlorobenzene conversions obtained under identical reaction conditions could rea-
sonably be taken for comparison. They demonstrate that except Mn and Fe chelates, the
other metal compounds give systems that speed up the dehalogenation reaction. Their
catalytic effect decreases in the order: Co ≈ Ni ≈ Pd > Cu >> Mn > Fe. The last entry
in Table II illustrates the efficiency of the systems examined in the chlorobenzene de-
halogenation. The use of SDMA–transition metal systems to effect dehalogenation of
PCB congeners will be reported in a subsequent work.
The authors thank Mr Z. Soukup (this Institute) for the excellent technical assistance with dehaloge-
nation experiments.
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

Dehalogenation of Chloroarenes
1653
REFERENCES
1. Carfagna C., Musco A., Pontellini R., Terzoni G.: J. Mol. Catal. 54, L23 (1989).
2. Carfagna C., Musco A., Pontellini R., Terzoni G.: J. Mol. Catal. 57, 23 (1989).
3. Lin S. T., Roth J. A.: J. Org. Chem. 44, 309 (1979).
4. Nelson R. B., Gribble G. W.: J. Org. Chem. 39, 1425 (1974).
5. Okano T., Yamamoto M., Noguchi T., Konishi H., Kiji J.: Chem. Lett. 1982, 977.
6. Ashby E. C., Lin J. J.: J. Am. Chem. Soc. 43, 1263 (1978).
7. Tabaei S. M. H., Pittman C. U., jr., Mead K. T.: J. Org. Chem. 57, 6669 (1992).
8. Simunek I., Kraus M.: Collect. Czech. Chem. Commun. 38, 1786 (1973).
9. Hetflejs J.: Chem. Listy 87, 407 (1993).
10. Friesova A., Sabata S., Rericha R., Hetflejs J.: Chem. Prum. 43/68, 104 (1993).
11. Sabata S., Friesova A., Rericha R., Hetflejs J.: Chemosphere 27, 1201 (1993).
12. Vcelak J., Friesova A., Rericha R., Hetflejs J.: Collect. Czech. Chem. Commun. 59, 1368
(1994).
13. Zavrel P., Svoboda P., Hetflejs J.: Collect. Czech. Chem. Commun. 57, 2135 (1992).
14. Strouf O., Casensky B., Kubanek V.: Sodium Dihydridobis(2-methoxyethoxo)aluminate (SDMA).
A Versatile Organometallic Hydride, p. 35. Academia, Prague 1985.
15. Marquardt D. W.: J. Soc. Industr. Appl. Math. 11, 431 (1963).
16. Svoboda P., Hetflejs J., Soucek J.: Chem. Prum. 30/55, 89 (1980).
17. Vankova Z., Svoboda P., Hetflejs J.: Chem. Prum. 31/56, 132 (1981).
18. Svoboda P., Dedek J., Hetflejs J.: Chem. Prum. 42/67, 110 (1992).
Translated by the author (J. H.).
Collect. Czech. Chem. Commun. (Vol. 59) (1994)