L. D. Marciasini et al. / Tetrahedron Letters 55 (2014) 1702–1705
1703
Table 1
Catalytic system optimization
Entry
Catalyst
Solvent
Yielda (%)
1
—
THF
0
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
[TiCp2]
[TiCp2]
TiCp2(CO)2
(CptBu)2TiCl2
CpTiCl3
TiCp2Cl2
Cp⁄TiCl3
TiCp⁄2Cl2
(Indenyl)2ZrCl2
(tBuCp)2ZrCl2
Cp2ZrCl2
Cp2ZrHCl
[Cp2Co]PF6
Cp2Co
THF
30
56
60
61
62
72
67
61
75
68
68
74
61
69
76
74
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN
Cp2Ni
Cp2Ru
a
Isolated yield after pinacol treatment and purification.
transfer. A quick solvent screening showed that polar solvent fa-
vored greatly the reaction, most likely because of the poor diazo-
nium solubility in apolar media. Most titanocene derivatives were
found to efficiently catalyze the coupling reaction (Table 1, entries
3–9). Addition of alkyl lithium species to TiCp2Cl2 in various quan-
tities led to minor differences in catalytic activities. Effect of cyclo-
pentadienyl (Cp) ring replacement by the permethyl analog (Cp⁄)
was unclear (Table 1, entries 8 and 9). Overall the titanocene dichlo-
ride was found to be the most active and practical complex to use.
As other metals were used for dehydrogenation of amine
borane complexes, they were evaluated in the borylation process.
Indeed zirconium based catalysts31 also efficiently promote the
borylation reaction (Table 1 entries 10–13). Among all tested
complexes the Schwartz’s reagent turned out to be one of the most
efficient (Table 1 entry 13) despite few differences within the
results. Other metallocene derivatives were tried, including cobal-
tocene (Table 1 entry 15), cobaltocenium hexafluorophosphate
(Table 1 entry 14), nickelocene (Table 1 entry 16), and ruthenocene
(Table 1 entry 17) led to interesting results. As Schwartz’s reagent
is known to react with nitrile, we evaluated the imino zirconium
complex formed by reaction between Cp2ZrHCl and acetonitrile,
it gave similar yields than without preformation.
This reaction was then extended to the use of other aryldiazo-
nium salts. Overall the reaction proceeds smoothly with electron
donating or withdrawing group. In most cases yields remain in
the 60–80% range, which is quite decent considering they originate
from a multistep sequence: aminoborylation, alcohol solvolysis,
and transesterification by pinacol. In addition to methoxy groups
(Table 2, entries 2–5), electron poor arylboronates, bearing nitro
group (Table 2, entries 6 and 10), trifluoromethyl groups (Table 2,
entries 7 and 9), cyano groups (Table 2, entry 8) can be obtained
using this method. Not surprisingly, 4-benzoyl substituent led only
to reduction product (benzophenone, Table 2, entry 25).
other substituents. Even iodobenzeneboronates can be synthesized
using our method (Table 2, entries 23 and 24).
Overall, apart from arylketo and carboxaldehyde substituted
diazonium salts, this method was found to be quite general in
terms of chemical substitution, likely owing to the mild conditions
used during the process. Additionally, this method offers a prepa-
ration of boron derivatives which could appear more versatile than
the previous methodologies using bis(pinacolato)diboron. Indeed,
the resulting aryl amino borane can be transformed at will,
depending on the work up of the borylation reaction, into the boro-
nic acid, boronates, or trifluoborate salts. As an example, diisopro-
pylamino-(4-methyl)phenylborane 3a is transformed by treatment
with a MeOH/aq HCl mixture affording the corresponding boronic
acid 5a in 91% yield. Trifluoroborate salts are obtained from the
same intermediate by treatment with a saturated aqueous solution
of KHF2 leading to the borate 6a in 90% yield. Considering the
safety issues associated with diazonium salts manipulation, we
performed an experiment generating the salt in situ by reaction
of the 4-methylaniline with isoamylnitrite and BF3–Et2O. Yields
were comparable (70%) but the reaction was slower than with
the isolated diazonium salt (4 vs 2 h).
Even if our first rational for this reaction was based on the acti-
vation of the B–H bond by zirconocene and titanocene, we were dee-
ply intrigued by the fact that any zirconocene and titanocene seem
to promote this reaction. Moreover other metallocenes (Table 1 en-
tries 14–17) led to the same conversion. We then performed a series
of experiment to rule out the possibility of a classical ionic metal
centered mechanism. In the case of a titanium centered mechanism
the reaction is likely to start with the oxidative addition of the aryl
diazonium salt to the metallocene (Scheme 3, A1). The aryl transfer
from the metal center onto the boron could proceed through an
addition leading to an amidoborane (Scheme 3, A2)32 or via a metal
assisted
r
-bond metathesis (Scheme 3, A3).33
One advantage of this methodology is the complete compatibil-
ity with halide-substituted aryldiazonium salts. If fluorine arylbo-
ronates can be easily accessible using classical methods (Table 2,
entries 12–14), the presence of chlorine, bromine and even more
iodine, can lead to selectivity issues when performing borylation
with classical reagents (organolithium or palladium/nickel cataly-
sis). In our cases, chlorine (Table 2 entries 15–19) or bromine
(Table 2 entries 20–22) substituted diazonium salts led to the cor-
responding boronates in similar yields to those obtained with
However, despite our tries we never succeeded to obtain the
cationic metallocenium by reacting titanocene (generated in situ
from TiCp2Cl2) or zirconocene with a diazonium salt. In all cases,
degradation occurred, mostly via arylation of the Cp ring. Moreover
reaction of zirconocene dichloride 7 with one equivalent of phenyl
lithium led to the phenylzirconocenium chloride 8, which under-
went an ion metathesis by the addition of silver tetrafluoroborate
9. This sequence provided us a reference compound which was
never observed during the standard reaction. Moreover, we never
Marciasini, Ludovic D.
