Direct Arylation of Benzene
FULL PAPER
mocoupling. Most notably, significant amounts of by-prod-
ucts, derived from an apparent coupling reaction between
aryl bromide 1a with tert-butanolate, were additionally pres-
ent in this reaction mixture.[32] Shirakawa and co-workers re-
ported similar impurities when KOtBu was used in combina-
tion with bathophenanthroline as ligand, supporting the pos-
sibility of a competing reaction pathway involving an aryne
intermediate.[17] In our hands, the results with LiHMDS as
base were in any event significantly better compared to the
results obtained when using KOtBu (Table 4, entries 1 and
2).[33]
Experimental Section
General experimental details: All reaction mixtures were prepared under
argon atmosphere using an Aldrich AtmosBag. Solvents and chemicals
were obtained from standard commercial vendors and were used without
any further purification. Benzene (anhydrous, 99.8%), KOtBu (sublimed,
99.99%), DMEDA (99%), CoCl2 (97%), Co
(99.9%+), Fe(OAc)2 (95%), NiI2, NiCl2 (98%), Ni
(acac)3 (97%), Cu(acac)2 (97%), Cu(OAc)2 (98%), CuI (98%), Mn-
(acac)2, and Mn(acac)3 (tech.) were purchased from Sigma–Aldrich.
LiHMDS, CoBr2 (anhydrous, 97%), Co(acac)2, and Co(OAc)2 (anhy-
G
ACHTUNGTRENNUNG(acac)3
E
ACHTUNGTRENNUNG
R
G
ACHTUNGTRENNUNG
R
ACHTUNGTRENNUNG
E
ACHTUNGTRENNUNG
drous, 98%) were obtained from Alfa Aesar. 1,10-phenanthroline
(99+%) was purchased from Acros Organics.
Similar to the approach taken with the transition-metal-
catalyzed arylation protocols described above, we subse-
quently optimized both protocols using LiHMDS and
KOtBu by employing a high-T/p process window. As can be
seen in Table 4, for the additional optimization studies the
amount of LiHMDS was successfully reduced to two equiva-
lents. We also minimized the amount of the 1,10-phenan-
throline base to 15 mol%, still being able to achieve com-
plete conversion after 5 min at 2008C and a moderately high
product yield (Table 4, entry 7).
1H and 13C NMR spectra were recorded on a Bruker 300 MHz instru-
ment using CDCl3 as solvent. Chemical shifts (d) are reported in ppm
downfield from TMS as internal standard. The abbreviations s, d, t, q,
and m are used to indicate singlet, doublet, triplet, quadruplet, and mul-
tiplet signals, respectively. Melting points were determined on a Stuart
SMP3 melting point apparatus. GC-FID analysis was performed on
a Trace-GC (ThermoFisher) with a flame ionization detector using a HP5
column (30 mꢁ0.250 mmꢁ0.025 mm). After 1 min at 508C, the tempera-
ture was increased in 258C minÀ1 steps up to 3008C and kept at 3008C
for 4 min. The detector gas for the flame ionization was H2 and com-
pressed air (5.0 quality). GC-MS analysis was performed on a Trace-GC
Ultra–DSQ II-MS system (ThermoElectron, Waltham, USA). The GC
conditions were as follows: HP-5 MS column (30 m ꢁ 0.25 mm ID,
0.25 mm film, Agilent, Waldbronn, Germany), carrier gas helium 5.0, flow
1 mLminÀ1, temperature gradient identical to GC-FID. The MS condi-
tions were as follows: positive EI ionization, ionization energy 70 eV, ion-
ization source temperature 2808C, emission current 100 mA, full-scan-
mode. Silica gel flash chromatography separations were performed on
a Biotage SP1 instrument by using petroleum ether/ethyl acetate mix-
tures as eluent.
Conclusion
In summary, we have shown that high-T/p process windows
can be used to dramatically increase the efficiency of the
transition-metal-catalyzed direct arylation of benzene with
aryl bromides. A high-throughput screening under these re-
action conditions revealed that several metals (in particular
Co complexes) are able to catalyze this arylation in the ab-
sence of a ligand, while for other metals (Cu, Ni, Cr) the
Screening experiments were carried out in 5 mL Pyrex reaction vials with
PTFE seals and screw caps in a SiC heating block with a 6ꢁ4 deep well
matrix (24MG5 rotor, Anton Paar GmbH).[26] The high thermal conduc-
tivity and effusivity of silicon carbide allowed rapid heating (within
5 min) of the reaction vials in the heating block, which was preheated on
a conventional hot plate with minimal deviations in temperature across
the block. The actual temperature inside the vials depended on the se-
lected temperature and the boiling point of the used solvent, but roughly
corresponded to the hotplate temperature.[26] The suggested temperature/
pressure limit of the vials was 2008C/20 bar.
presence of
a ligand, such as 1,10-phenanthroline or
DMEDA, is required to allow direct arylation. The high-T/p
reaction screening additionally revealed a novel Ni- or Cu-
catalyzed amination reaction that converts aryl bromides
into the corresponding anilines when using LiHMDS as am-
monia equivalent. In the presence of 1,10-phenanthroline as
ligand, the selectivity in this process switches toward direct
arylation and the formation of biaryls. This impressive selec-
tivity control by a simple additive is of significant interest
for further mechanistic studies. As an additional result from
a comparative study, it was discovered that LiHMDS ap-
pears to be a superior base compared to KOtBu in the 1,10-
phenanthroline-mediated, transition-metal-free organocata-
lytic direct arylation of benzene with aryl bromides.
The initial results obtained from high-throughput parallel
reactions screening at 1608C were further optimized by em-
ploying a high-T/p autoclave made out of sintered SiC. At
a reaction temperature of 2008C (about 15 bar) the efficien-
cy of the direct arylation protocols was dramatically in-
creased, allowing, in most cases, full conversion to be ach-
ieved within 5 min, while retaining high selectivity at often
much reduced catalyst loadings. This constitutes a significant
improvement to the published protocols in this area and will
undoubtedly contribute to the further development of this
rapidly growing field of modern synthetic organic chemistry.
For experiments using 10 mL SiC vessels, a Monowave 300 single-mode
microwave reactor from Anton Paar GmbH (Graz, Austria) was used
(Figure S1).[31b] The reaction temperature was monitored by an external
infrared sensor (IR) that was housed in the side-walls of the microwave
cavity, measuring the surface temperature of the reaction vessel, as well
as an internal fiber-optic (FO) temperature probe (ruby thermometer).
In all instances, the temperatures recorded by the IR and FO probes
were different (FO 5–258C higher than IR), mainly depending on IR
sensor calibration. Temperature control through the internal FO probe
(IR as slave) was generally a beneficial choice for Pyrex vessel experi-
ments, in which the selected target temperature was attained by dielectric
heating of the solvent. The IR sensor turned out to be superior to the FO
probe when SiC vessels were used. Therefore, temperature controlling
was carried out by using IR (FO as slave) after previous calibration using
FO (IR as slave).
All biaryls, synthesized in this work, are known in the literature. Their
1
structure and purity was confirmed by H and 13C NMR spectroscopy and
GC-MS analysis. The NMR data are given in the Supporting Informa-
tion.
General procedure for parallel-screening reactions in SiC plates
(Figure 2, Table S1 in the Supporting Information): LiHMDS (65 mg,
388 mmol) and the respective catalyst (19 mmol) were introduced into the
5 mL Pyrex vials under an argon atmosphere. Then, 4-bromoanisole (1a,
16 mL, 128 mmol), dodecane (29 mL, 128 mmol), and benzene (2 mL) were
Chem. Eur. J. 2012, 18, 5047 – 5055
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5053