Communication
Table 2. (Continued)
Entry Aniline
17
Product (8)
Flow[b] 8/9/7 Yield [%][c]
3
1:0:0 83[d]
[a] Syringe 1 was loaded with a methanol solution containing 1 and
methyl acrylate while syringe 2 contained MeSO3H, tert-butylnitrite, and
[Pd(OAc)2] in DMF. The two syringes were flowed together into one
DuPont FEP tube 1152 mL (1/8 ODꢁ0.06 ID) at RT at the specified flow
rate. After traversing the reaction tube, the reaction mixture emptied into
a quenching flask containing triethylamine in ethyl acetate. [b] The flow
rate reported is in mLminÀ1. [c] After collecting the effluent from the reac-
tor in the Et3N quenching solution, product was extracted into ether, con-
centrated, and purified by column chromatography. Yield is based on the
volume of the starting material solutions infused after reaching steady
state. [d] 5 mol% of [Pd(OAc)2] used.
Scheme 3. Diazotization/Mizoroki–Heck reaction quenching study. [a] CD3OD
was used in place of CH3OH/DMF.
ing. While the conditions that are used with tBuONO are anhy-
drous, one must also remember that diazonium salts are also
made using NaNO2 in water,[1] thus they are quite stable in
water.
To examine the above concerns, a series of control experi-
ments were conducted (see Scheme 3). First, a diazotization so-
lution of 2 in MeOH/DMF (1:1) was added to a flask containing
the remaining reactants necessary for the Mizoroki–Heck reac-
tion in an ether/water mixture at RT (Scheme 3a). To mimic the
reactions in this report, the mixture was stirred for 4 h and the
layers were separated and worked up as usual (e.g., see
Table 2). The cross-coupling proceeded to give an 8% yield of
6 and 54% when the quenching solution was stirred for 24 h.
When the quenching solution contained only water
(Scheme 3b), 5% conversion was attained after 4 h, and 18%
after 24 h. It would appear that reports in the literature using
any of these methods to quench the reactor effluent need to
be revisited.
that one gets from reading the literature on this protocol is
that diazotization is generally favoured by electron-withdraw-
ing groups on the ring and an ortho substituent appears to en-
hance salt lifetime, presumably due to steric effects.[8,9] Moving
the nitro group to the 4-position and leaving the ortho site
empty (8b) did not adversely impact the process (Table 2,
entry 2), although completely removing it did reduce conver-
sion through to coupling (entries 14 and 15). Further, the pres-
ence of electron-donating groups were very nicely tolerated
(entries 3, 4, 7, 9, 11, and 13). The process also works on styryl-
based coupling partners (entry 17).
Careful study of the literature involving not only diazonium
salt use in flow, but flow chemistry in general indicates that
proper control and quenching experiments are not always con-
ducted to support the claims made in the manuscripts. Many
authors report flow rates, residence times, and yields that
relate to a specific reaction’s performance in the reactor, yet
they merely flow the effluent from the reactor into a flask to
collect it over the course of hours, or in some cases even days.
Unless the reaction requires something affixed to the reactor
lumen, for example, a catalyst flow bed, most reactions could
simply continue to proceed in the collection vessel. Ironically
the controlled contact time of reactants is one of the key fea-
tures that enables flow chemistry to produce such clean, high-
yielding and reproducible transformations. Not properly
quenching reactor effluent may undo much of what is gained
by working in flow in the first place. Consequently, the claims
made in a number of flow chemistry reports have yet to be va-
lidated. Even flowing effluent into a cooled flask is no guaran-
tee that the reaction has terminated; it must be chemically
quenched and validated to be so. In the flow diazonium salt
literature, reported ‘quenches’ include flowing reactor effluent
into empty chilled flasks, into water, and into a water/ether
mixture. As many diazonium salt reactions are actually con-
ducted at 08C,[1,2] it is safe to reject the first quench method
straight off. The quenches involving water are equally perplex-
To ensure that the results derived from the new protocols
developed in this manuscript were meaningful, a quenching
strategy was developed.[10] It is known that Et3N effectively
causes dediazotization,[13,14] so a quenching solution of Et3N
(one equiv relative to 1) in ethyl acetate was created. Diazoni-
um salt 2 was created in batch and added in one shot to the
1
quenching solution and after 60 s a H NMR spectrum of the
resultant solution was recorded (Scheme 3c).[10] Dediazotization
proceeded to approximately 67%, while addition to a solution
containing 2 equiv of Et3N (Scheme 3d) led to complete forma-
tion of 3. To confirm that dediazotization in the quenching
flask is fast enough to preclude any Mizoroki–Heck coupling
from occurring, that is, all diazonium salt is destroyed rapidly,
the cross-coupling components were added to the quenching
solution (Scheme 3e). This time when 2 was added, zero cou-
pled product was observed confirming that this new quench-
ing strategy is efficient. When the reactions are performed in
flow, even with only 1.0 equiv of Et3N present, the flow of ef-
fluent is slow enough to ensure that the base is always in vast
excess of the diazonium salt to ensure thorough quenching.
Thus, the results presented in this manuscript originate solely
from what happens in the reactor and the conclusions can be
viewed with confidence from that perspective.
&
&
Chem. Eur. J. 2014, 20, 1 – 6
4
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!