Hindered dialkyl ether synthesis with electrogenerated carbocations

Article abstract of DOI:10.1038/s41586-019-1539-y

Hindered ethers are of high value for various applications; however, they remain an underexplored area of chemical space because they are difficult to synthesize via conventional reactions^1,2. Such motifs are highly coveted in medicinal chemistry, because extensive substitution about the ether bond prevents unwanted metabolic processes that can lead to rapid degradation in vivo. Here we report a simple route towards the synthesis of hindered ethers, in which electrochemical oxidation is used to liberate high-energy carbocations from simple carboxylic acids. These reactive carbocation intermediates, which are generated with low electrochemical potentials, capture an alcohol donor under non-acidic conditions; this enables the formation of a range of ethers (more than 80 have been prepared here) that would otherwise be difficult to access. The carbocations can also be intercepted by simple nucleophiles, leading to the formation of hindered alcohols and even alkyl fluorides. This method was evaluated for its ability to circumvent the synthetic bottlenecks encountered in the preparation of 12 chemical scaffolds, leading to higher yields of the required products, in addition to substantial reductions in the number of steps and the amount of labour required to prepare them. The use of molecular probes and the results of kinetic studies support the proposed mechanism and the role of additives under the conditions examined. The reaction manifold that we report here demonstrates the power of electrochemistry to access highly reactive intermediates under mild conditions and, in turn, the substantial improvements in efficiency that can be achieved with these otherwise-inaccessible intermediates.

Full text of DOI:10.1038/s41586-019-1539-y

Letter
https://doi.org/10.1038/s41586-019-1539-y
Hindered dialkyl ether synthesis with
electrogenerated carbocations
Jinbao Xiang1,2,9, Ming Shang1,9, Yu Kawamata1, Helena Lundberg1,3, Solomon H. reisberg1, Miao Chen1, Pavel Mykhailiuk1,4,5
Gregory Beutner6, Michael r. Collins7, Alyn Davies8, Matthew Del Bel7, Gary M. Gallego7, Jillian e. Spangler7, Jeremy Starr8,
Shouliang Yang7, Donna G. Blackmond1 & Phil S. Baran1*
,
Hindered ethers are of high value for various applications; however, presence of the requisite secondary alcohol—provides the ether in only
they remain an underexplored area of chemical space because 11% yield8.
they are difficult to synthesize via conventional reactions1,2
.
Distinct from sterically sensitive SN2 and strongly acidic carboca-
Such motifs are highly coveted in medicinal chemistry, because tion pathways, there is a third class of ether synthesis that has been
extensive substitution about the ether bond prevents unwanted known for many years but has remained largely underexplored. This
metabolic processes that can lead to rapid degradation in vivo. class (Fig. 1a, yellow inset) stems from the oldest synthetic organic
Here we report a simple route towards the synthesis of hindered electrochemical reaction, the Kolbe dimerization, which was discov-
ethers, in which electrochemical oxidation is used to liberate ered9 in 1847. In the so-called interrupted Kolbe variant, known as
high-energy carbocations from simple carboxylic acids. These the Hofer–Moest reaction10, electrolytic oxidation of a carboxylic acid
reactive carbocation intermediates, which are generated with low under mildly alkaline conditions generates a carbocation that can be
electrochemical potentials, capture an alcohol donor under non- captured by incipient nucleophiles10–18. A distinct advantage of this
acidic conditions; this enables the formation of a range of ethers reaction is the non-acidic generation of high-energy carbocations
(more than 80 have been prepared here) that would otherwise be directly from carboxylic acids.
difficult to access. The carbocations can also be intercepted by
We were therefore surprised to find a dearth of applications of the
simple nucleophiles, leading to the formation of hindered alcohols Hofer–Moest reaction in synthetic contexts (Fig. 1b). Indeed, much
and even alkyl fluorides. This method was evaluated for its ability to more complex catalytic systems that take advantage of photolytic con-
circumvent the synthetic bottlenecks encountered in the preparation ditions have been developed to make alkyl–aryl ethers19. There are
of 12 chemical scaffolds, leading to higher yields of the required probably two reasons for the limited exploration of electrolytic decar-
products, in addition to substantial reductions in the number of boxylative ether synthesis: first, the barrier to entry into electrosynthesis
steps and the amount of labour required to prepare them. The use has traditionally been high for a practicing synthetic organic chem-
of molecular probes and the results of kinetic studies support the ist20. Second, all Hofer–Moest systems known so far have required
proposed mechanism and the role of additives under the conditions solvent-quantities of the alcoholic nucleophile, which is untenable for
examined. The reaction manifold that we report here demonstrates complex ethereal substrates. The alcoholic solvent—in addition to func-
the power of electrochemistry to access highly reactive intermediates tioning as a reagent—permits current to pass and also acts as an electron
under mild conditions and, in turn, the substantial improvements sink to balance the electrochemical process, liberating hydrogen gas.
in efficiency that can be achieved with these otherwise-inaccessible
intermediates.
Herein, we report how the generation of carbocations from unac-
tivated, aliphatic carboxylic acids—and their subsequent capture by
The Williamson ether synthesis3,4 is a long-established method by heteroatom nucleophiles—can be leveraged to provide a wide array of
which to synthesize primary alkyl ethers via SN2 substitution (Fig. 1a). hindered C–X bonds.
However, in contexts involving secondary or tertiary alkyl halides the
Several key literature precedents can be highlighted that aided this
reaction often derails, leading to elimination byproducts or to no development. It is known that carbon-based electrodes favour the
reaction at all. Hindered ether 1, which is a key intermediate in the desired carbocation generation, whereas platinum electrodes favour
synthesis of an aurora kinase modulator, exemplifies this commonly unproductive radical (Kolbe-type) dimerization21. It is also known that
faced challenge. Despite the documented utility of hindered ethers1,2, inert, non-oxidizable anions (for example, ClO4−) enhance cation-like
very little progress has been made in facilitating access to them. The reactivity in the Hofer–Moest reaction22. Mildly alkaline conditions
alternative workhorse method, the Mitsunobu reaction, also fails in are known to be beneficial for the desired carbocation formation22,23
.
such settings owing to the steric demands of the SN2 process and the Finally, simple undivided cells are generally used in this process, which
pKa requirements of the nucleophile5. To the best of our knowledge, suggests that cathodic reduction would not interfere substantially with
the most frequently used method for the synthesis of hindered dialkyl the reaction.
ether bonds still uses carbocation chemistry accessed from olefins
It became immediately apparent that limiting the amount of alco-
(hydroalkoxylation) under strongly acidic conditions6. Although this hol posed several considerable challenges: the decomposition of the
transformation has been known for nearly a century, its use is sub- carbocation due to the low nucleophilicity of alcohols; the competitive
stantially limited in scope owing to sluggish reactivity and a lack of trapping of the carbocation by water; the consumption of alcohols by
chemoselectivity7. For example, in order to prepare hindered ether 1, anodic oxidation; and the necessity of an external electron-acceptor in
a multi-step route via 4-hydroxyproline 2 (R = CO2Me) was used. The order to balance electrons. Figure 1c summarizes the results of around
synthesis required over 6 days of reaction time and gave the required 1,000 experiments (see Supplementary Information for an extensive
product in less than 4% overall yield; the key C–O bond-forming reac- sampling) that were undertaken in order to solve these problems.
tion—the treatment of methylenecyclobutane with BF3·Et2O in the Not surprisingly, initial exploratory experiments using the proposed
1Department of Chemistry, Scripps Research, La Jolla, CA, USA. 2The Center for Combinatorial Chemistry and Drug Discovery of Jilin University, The School of Pharmaceutical Sciences, Jilin
University, Jilin, People’s Republic of China. 3Department of Chemistry, KTH Royal Institute of Technology, Stockholm, Sweden. 4Enamine Ltd, Kiev, Ukraine. 5Chemistry Department, Taras
Shevchenko National University of Kyiv, Kiev, Ukraine. 6Chemical and Synthetic Development, Bristol-Myers Squibb, New Brunswick, NJ, USA. 7Department of Chemistry, La Jolla Laboratories,
Pfizer Inc, San Diego, CA, USA. 8Pfizer Medicinal Sciences, Groton, CT, USA. 9These authors contributed equally: Jinbao Xiang, Ming Shang. *e-mail: pbaran@scripps.edu
N A t U r e | www.nature.com/nature

reSeArCH Letter
a
Pathways to hindered dialkyl ethers
b Previous decarboxylative etheriꢀcations
Me
X
Photochemistry
R1
R1
[Ir] cat.
[Cu] cat.
R2
R3
X = LG or OH
R2
R3
+
O
CO2A*
HO
SN2; problematic on hindered tertiary systems
hQ
FG
Alkyl aryl ethers
FG
Phenols
Me
Alkyl NHPI esters
O
HO
BF3-cat. addition
(11% yield)
NCbz
OH
NCbz
Electrochemistry
Hydroalkoxylation; <4% overall yield (>6 d)
1, kinase
modulator intermediate
A
R
R1
R2
O
R1
R2
CO2H
2
Me
Alkyl
+
Alkyl–OH
CO2H
R3
R3
Anodic oxidation
This work
Direct decarboxylative etheriꢀcation
Typical acid scope
HO2C
Alcohol scope (solvent)
A
R2
R1
R2
OH
X
R1
MeOH
+
R1
R2
O
R6
R5
R1
R3
CO2H
R1
R–OH
CO2H
Me
OH
R2
CO2H
X = N, O, Si
R3
R4
R2
OH
Anodic
oxidation
R3
Alkyl acid
Carbocation
Hindered ethers
Me
Me
HO
Alkyl–CO2H (Me, nPr, etc.)
• Ubiquitous
• Inexpensive
• Reactive
• Non-acidic generation
• Difficult to access
• Medicinally important
c
Reaction optimization
Byproducts (from 3)
Base
Electrolyte (0.1 M)
Me
Me Me
Me
Me Me
Me
Ph
Me Me
Ph
OH
+
Ph
Me
+
Ph
H
Ph
O
Ph
Ph
CO2H
HO
Ph
Additive, solvent, RT
+C/–C, 10 mA, 3 h
Me
7
Elimination
Me
Me
3
4
5
6
8
(0.2 mmol)
(0.6 mmol)
From radical
Hydration
Conditionsa
Yield (%)b
Entry
5
6
7
8
9
10
1
2
3
4
5
6c
7
8
Et3N, nBu4NClO4, DMF or acetone
K2CO3, nBu4NClO4, DMF
–
–
<1 <1
–
–
<1
Byproducts (from 4)
15 <1 40 <1
<1
2,4,6-collidine,nBu4NPF6, DMF
2,4,6-collidine,nBu4NPF6, CH2Cl2
2,4,6-collidine,nBu4NPF6, 3 Å MS, CH2Cl2
6
40
3
4
5
–
28 14 <1
–
Me
O
Me
via
Me
Ph
-
2
–
14
<1
–
6
3
7
6
–
Ph
+
Ph
O
O
Ph
–
43
79 (77)d
AgPF6, 2,4,6-collidine, nBu4NPF6, 3 Å MS, CH2Cl2
Me Me
–
9
10
As in entry 6, 1 equiv. alcohol
AgPF6, nBu4NPF6, 3 Å MS, CH2Cl2
<1 <1 <1
<1 <1
<1
Anodic oxidation
Ionization
34
–
22 54
–
Fig. 1 | Background and reaction development. a, The synthesis of
hindered ethers is a long-standing challenge in organic synthesis.
A; constant-current electrolysis; cat., catalyst; Cbz, carboxybenzyl;
LG, leaving group. b, Historical context and previous strategies for
decarboxylative etherification. FG, functional group; NHPI,
N-hydroxyphthalimide. c, Development and optimization of hindered
3 (0.2 mmol), 3.0 equiv. of alcohol 4 (except where designated). bYield
based on gas chromatography. All entries were performed in triplicate.
cConditions: acid 3 (0.2 mmol), alcohol 4 (0.6 mmol), AgPF6 (0.3 mmol),
2,4,6-collidine (0.6 mmol), nBu4NPF6 (0.1 M), 3 Å molecular sieves
(150 mg), dichloromethane (CH2Cl2; 3 ml), current (I) = 10 mA, 3 h.
dIsolated yield. DMF, N,N-dimethylformamide; RT, room temperature;
+C/−C represents the graphite electrodes.
ether synthesis depicted through electromechanistic analysis. aCompound
conversion of 3 and 4 to 5 as an example—based on the literature prece-
The patent literature is replete with examples of hindered ethers in
dent available11,22—led to only trace amounts of product (entry 1). The various pharmaceutical and materials applications. Although carbo-
conversion of the carboxylate was improved by choosing potassium cation-based routes from olefins predominate in the literature, elec-
carbonate or 2,4,6-collidine as a non-oxidizable base (entries 2, 3), trochemical access to such valuable entities has notable advantages in
although 6 and 7 were identified as major byproducts due to the pres- terms of the time required, the step count and the overall yield. Figure 2
ence of radical intermediate22 and elimination pathways, respectively. shows an abbreviated depiction of six such applications as well as the
These problems were effectively suppressed by changing the solvent more than 80 ethers prepared (see Supplementary Information for a full
to dichloromethane (entry 4), which resulted in a large increase of the listing of substrates as well as a comparison to previous routes). Primary
desired ether product. In dichloromethane, hydration of the carboca- carboxylic acids and certain secondary systems are not compatible with
tion (leading to 8) persisted (entry 4), which was suppressed by the electrochemical etherification, because the resultant carbocations are
addition of 3 Å molecular sieves (entry 5). It was also found that dichlo- not sufficiently stable. However, this limitation is inconsequential from
romethane was apparently reduced at the cathode (Supplementary a synthetic perspective, as those ethers can be easily prepared through
Fig. 2), acting as an electron sink. Past approaches using water or simple standard SN2-type approaches3–5. Acids bearing various functional
alcohols as the solvent did not need to address this issue, as the solvent groups are tolerated, such as Boc-protected amines (17), aryl and alkyl
can serve as both the reagent and an electron sink (via proton reduc- halides (25, 27, 30), olefins (26), esters (29, 39), enones (39), ethers (35,
tion). Accordingly, the addition of a better sacrificial oxidant (silver 36, 62) and even oxidation-prone boronic esters (28). Similarly, the
hexafluorophosphate, AgPF6) considerably improved the yield of 5, scope of alcohol coupling partners is vast and includes acid-labile and
leading to our optimized conditions (entry 6). The addition of AgPF6 oxidation-prone chiral secondary benzylic alcohols (17), deuterated
also completely suppressed the formation of 6 and 7, although this systems (49), azetidinyl alcohols (50), protected sugars (53) and olefin-
effect was found to vary by substrate and—in some cases—silver addi- containing alcohols (48, 51, 59), as well as alcohols containing acetals and
tives are not necessary at all. Negative controls confirmed the necessity esters (44), halides (15), nitriles (45), nitro groups (see Supplementary
of a slight stoichiometric excess of the alcoholic partner (entry 7). In Information) and even Lewis-basic heterocycles (42, 43, 45). The ability
the absence of base, no desired product was observed, with the major to tolerate chiral, ionizable secondary alcohols is worth emphasizing,
products being the ketone 9 and the ester 10 (entry 8).
because acidic methods for ether synthesis using such alcohols would
N A t U r e | www.nature.com/nature

Letter reSeArCH
2,4,6-collidine (3.0 equiv.)
nBu4NPF6 (0.1 M)
O
R1
R4
R4
R2
R3
R5
R6
R5
R6
R1
R2
+
OH
AgPF6 (1.5 equiv.)
3 Å MS, CH2Cl2 (3 ml), RT
+C/–C, 10 mA, 3 h
O
O
HO
R3
(0.2 mmol, 1.0 equiv.)
(3.0 equiv.)
a
Applications
Me
Me Me
CO2Me
NHCbz
CO2Me
O
Me
O
BocN
O
Br
NCbz
O
O
Me
O
NHCbz
O
Me
Me
Me
Me Me
Me
OH
12
1 step
1
11
1 step
13
1 step
14
1 step
15
1 step
Current route: 2 steps
(51%, 15 h)
Previous route: 3 steps
(<4% overall, >6 d)
(32%, 3 h)
5 steps
(31% overall, 2.5 d)
(40%, 3 h)
7 steps
(37% overall, >3 d)
(21%, 3 h)
6 steps
(24% overall, 2 d)
(42%, 3 h)
4 steps
(47% overall, >2 d)
(81%, 3 h)
2 steps
(<2% overall, >5 d)
Secondary alcoholsk
CD3
Me Me
D
Tertiary acids
Me
Me
Me
Me
Me
O
Me
O
Me
O
O
CD3
O
O
N
Me
j
49: 82%
Me
BocN
Me
Me
Me
Me
Me Me
O
Me
n
NCbz
n = 2, 18: 54%b,c / 40%c,d
17, 45%b,c
16, 65%
n = 1, 19: 40%b,c / 32%c,d
O
CF3
Me
48, 32%e
95% e.e.
Ts
R
f
50, 70%
46, 60%b
47, 39%e
Cl
O
O
Cl
O
R
O
Me
Cl
Me
Me
Me Me
R = H, 22: 61%e
R = Me, 23: 58% / 62%b,j / 50%d
R = PhCH2CH2, 24: 35%b,c / 28%c,d
Me
Me
Me Me
Me
25, 24%f,g / 43%f,g,h
Me
R = nPr, 20: 52%b,c
R = Bn, 21: 41%b
Me
Me
Me
H
O
O
Me
NBoc
Me
52, 66%e
Me
R
R
Me Me
O
O
Me
H
H
CO2Me
Me
O
O
Ph
O
Ph
O
51, 54%e
Ph
O
O
Me Me
O
f
X
53, 52%
Me Me
f
Me
5, 77%j / 78%f / 54%i
Me
26, 43%f,g / 31%d
R = R = –(CH2)5–, X = F, 27, 46%
R = R = Me, X = BPin, 28, 54%f
Tertiary alcoholsk
Me
Me
Me
O
Secondary acids
CO2Me
Me
Me
Me
BocN
Me
Cl
Me
Me
Me
Me
Me
Me
O
O
O
O
Me
O
Cl
54, 42%
55, 65%
56, 52%
Me
Me
Me
Me
f,g
Me
Me
29, 41%
30, 74%, d.r. = 1:1
f
31, 68%
Me
Me Me
O
Me
Me
O
Me Me
O
Me
16
O
O
O
Me
Me
CF3
f
59, 28%b,c
Me
57, 46%
58, 65%j/54%
i
f
32, 80%f,j / 48%i
34, 7%
33, 57%
Fluorinated ethersk
α-heteroatom acids
Me
OMe
O
CF3
Me
O
O
F
F
F
F
O
NCbz
BocN
Ph
O
N
O
O
O
O
O
f
36, 55%j
62, 88%f,g, d.r. = 1:1
35, 58%
37, 82%
f,g
f
60, 46%
61, 23%
Drugs/natural products
CF2H
Me
O
O
Me
Me
Me
OMe
O
CF3
CH2F
Me
Me
H
Me
O
Me
Me
O
Ph
f,g
Me
O
AcO
Me
Me
H
CH2F
H
Me
Me
Me
Me
f,g
f,g,j
63, 42%
64, 51%
65, 42%
38, 90%j, d.r. = 1:1
39, 35%b,c, d.r. = 1:1
Primary alcoholsk
PEG ethers
Ar
H
N
O
Me
N
Ph
Me Me
O
Ph
OMe
O
O
OH
Me Me
Me Me
41, 59%j
O
Ph
O
S
O
Cl
O
Me
O
f
H
Me
Ar = 2,4-diF
–Ph
66, 59%
Me Me
O
Cl
Me
Me
f
R
O
42, 68%e
40, 48%
Me Me
Ph
O
O
O
NC
Me
Me
O
Me
O
O
Me Me
O
O
O
Me
Et
BocN
CN
Me Me
N
N
Ph
O
N
N
Cl
Me Me
R
=
Ph
Me
O
O
Me
69, 25%b,c
f
67, 48%e,j
44, 52%
Me
f
Me
45, 42%e
Me
Et
68, 57%e
OMe
43, 50%
molecular sieves (100 mg), CH2Cl2 (2 ml), I = 10 mA, 3 h. fAgClO4
(0.6 mmol) instead of AgPF6, nBu4NClO4 (0.1 M) instead of nBu4NPF6. g4.0
or 6.0 equiv. alcohol. h1.5 ml CH2Cl2, I = 7.5 mA, 4 h. i nBu4NClO4 (0.1 M)
instead of nBu4NPF6, no AgPF6. jReaction performed in triplicate; yield is
average of three runs. kSee Supplementary Information for more examples.
Boc, tert-butyloxycarbonyl; d.r., diastereomeric ratio; MS, molecular
sieves; Ts, tosyl.
Fig. 2 | Applications, and partial scope of hindered ether synthesis
via electrochemical decarboxylation. aSee Supplementary Information
for literature routes. bAgSbF6 (0.3 mmol) instead of AgPF6. cDBU
(1,8-diazabicyclo[5.4.0]undec-7-ene; 0.6 mmol) instead of 2,4,6-collidine.
dKSbF6 (0.3 mmol) instead of AgPF6. eAlcohol as limiting reagent,
conditions: alcohol (0.15 mmol), carboxylic acid (0.45 mmol), AgClO4
(0.6 mmol), 2,4,6-collidine (0.675 mmol), nBu4NClO4 (0.2 M), 3 Å
N A t U r e | www.nature.com/nature

reSeArCH Letter
CO2Me
CO2Me
2,4,6-collidine (1.5 equiv.)
nBu4NPF6 (0.1 M)
O
R1
R2
Nu
Commercially
available
R1
+
Nu
OH
Acetone (3 ml), RT
+C/–C (10 mA), 3 h
R2
R3
HO2C
70
71
HO2C
F
R3
Me OH
(0.2 mmol, 1.0 equiv)
F
Applicationsa
CO2H
H
O
CO2Me
Me
Me
Me
NH2
CO2Me
H
Me
H
MeO
HO
HO
MeO
HO
HO
72
73
74
75
76
77
Me Me
Current route: 2 steps
(56%, 9 h)
Previous route: 7 steps
(21% overall, >4 d)
1 step
(66%, 3 h)
9 steps
2 steps
(31%, 24 h)
7 steps
2 steps
(22%, 27 h)
14 steps
3 steps
(17%, 21 h)
7 steps
1 step
(61%, d.r. = 1.1:1, 3 h)
5 steps
(15% overall, >5 d)
(12% overall, 3 d)
(5% overall, >9 d)
(8% overall, 62 h)
(yield not reported)
Me
Tertiary and secondary acidsb
OH
Me
Me OH
Me
Me Me
OH
Me
O
N
18OH
OH
Me
Me
Me Me
OH
Me
R
H
Me
OH
N
R
R
O
MeO2C
Me
80, 95%c
Me OH
R = Ph, 78, 52%c
81,61% yield R = Boc, 82:32%c
R = Br, 84: 67%c
R = Bpin, 85:55%c
86,53%c from
(1S)-(–)-camphanic acid from dehydroabietic acid
87, 32%c (d.r. = 3:1)
88,40%c
from indoprofen
R = PhO, 79, 66%c
(67.1%
R = Ts, 83:69%c
18O labelled)c
Esteriꢀcation
O
Fluorination
Amidation
H
N
Me
F
Me
O
F
O
O
Me Me
F
F
F
F
N
F
F
O
O
CO2Me
O
N
Cl
Cl
Boc
89,31%d
90,36%d
91, 35%e
92, 58%e
93, 18%e
94, 62%e
95, 14%d
Scale-up
Me
Me
A
Me Me
Ph
CO2H
OH
Me Me
Ph CO2H
HO
O
H2O
+
+
MeO2C
MeO2C
81, 0.86 g (65%)g
3
71
(
)-4
(
)-5, 2.08 g (72%)f
1.97 g (12 mmol)
4.40 g (36 mmol)
1.53 g (7.2 mmol)
0.6 ml
Fig. 3 | Applications and partial scope of trapping electrogenerated
carbocations with other nucleophiles along with scalability
demonstration. aSee Supplementary Information for literature routes.
bSee Supplementary Information for more examples. cH2O (0.1 ml) as
nucleophile. dCarboxylic acid or phenylacetonitrile as nucleophile
(0.6 mmol, 3 equiv.), conditions: AgClO4 (0.6 mmol, 3 equiv.),
2,4,6-collidine (0.6 mmol, 3 equiv.), nBu4NClO4 (0.1 M), 3 Å molecular
conditions: 18-crown-6 (0.72 mmol, 3.6 eq), AgClO4 (0.6 mmol, 3 equiv.),
2,4,6-collidine (0.6 mmol, 3 equiv.), nBu4NPF6 (0.1 M), 3 Å MS (150 mg),
CH2Cl2 (3 ml). fConditions for scale-up to ( )-5 (each reaction): 3
(2.4 mmol), ( )-4 (7.2 mmol), 2,4,6-collidine (3.6 mmol), nBu4NClO4
(0.1 M), 3 Å molecular sieves (450 mg), CH2Cl2 (9 ml) +C/−C (10 mA),
RT, 15 h. gConditions for scale-up to 81 (each reaction): 71 (1.2 mmol),
H2O (0.1 ml), 2,4,6-collidine (1.8 mmol), nBu4NPF6 (0.02 M), acetone
sieves (150 mg), CH2Cl2 (3 ml). eKF (0.72 mmol, 3.6 equiv.) as nucleophile, (9 ml), +C/−C (10 mA), RT, 12 h. Nu, nucleophile; pin, pinacolato.
lead to elimination or racemization (17 is formed in 95% enantiomeric synthetic pathways. Six such examples (72–77) are illustrated in Fig. 3
excess (e.e.)). Electrogenerated cations bearing fluorine atoms can also (see Supplementary Information).
be intercepted, opening up a range of organofluorine-containing ether
The addition of water to generate various tertiary and secondary
systems that were previously either difficult to access or unknown alcohols is a broadly applicable process, with selected examples depicted
(see Supplementary Information for full listing of fluorinated ethers). in Fig. 3 (for additional examples, see Supplementary Information).
Finally, the synthesis of polyethyleneglycol (PEG) ethers—which histori- Again, the chemoselectivity is on par with that observed in the syn-
cally has been laborious—can be achieved in a modular fashion using this thesis of hindered ethers: aryl bromides (84), boronic esters (85),
process (no PEG ethers analogous to 66–69 are known). To ensure the electron-rich aromatics (87), lactams (88), ethers (79) and esters (81)
robustness of the process, ten randomly selected examples in Fig. 2 were are tolerated. In preliminary studies, the possibility of adding other
run in triplicate, with yields varying by no more than 5% between runs. nucleophiles to the putative electrogenerated carbocations was also
As mentioned above, the use of simple alcohols such as methanol— studied. Carboxylates that are not capable of decarboxylation under
as well as water—is already known in electrochemical decarboxyla- these conditions could act as nucleophiles, thus providing hindered
tive processes10–18 (Fig. 1b). However, the scope of such processes is esters (89, 90)24. Useful organofluorine building blocks could also be
quite limited. The conditions developed here were therefore adapted accessed (91–94) in a process that might be of use in radiolabelling
for these related reactions (Supplementary Information). In order to studies, because the fluorine source used (the inexpensive salt potas-
render this reaction general, the choice of electrolyte (tetrabutylam- sium fluoride) is preferred in such situations25. Finally, using benzoni-
monium hexafluorophosphate, nBu4NPF6) and base (2,4,6-collidine) trile as a nucleophile led to the expected Ritter-type product (95), albeit
was crucial, whereas silver additives were found to be unnecessary. As in lower yield26. The robust and practical nature of this process was
with the synthesis of hindered ethers, the most convincing case for also demonstrated by the gram-scale preparation of 5 and 81 through
the use of this reaction stems from its ability to substantially truncate etherification and hydroxylation processes, respectively. In the case of
N A t U r e | www.nature.com/nature

Letter reSeArCH
11. Corey, E. J., Bauld, N. L., La Londe, R. T., Casanova, J., Jr & Kaiser, E. T. Generation
of cationic carbon by anodic oxidation of carboxylic acids. J. Am. Chem. Soc. 82,
2645–2646 (1960).
12. Luo, X., Ma, X., Lebreux, F., Markó, I. E. & Lam, K. Electrochemical
methoxymethylation of alcohols – a new, green and safe approach for the
preparation of MOM ethers and other acetals. Chem. Commun. 54, 9969–9972
(2018).
13. Bunyan, P. J. & Hey, D. H. The electrolysis of some aryl-substituted, aliphatic
acids. J. Chem. Soc. 1360–1365 (1962).
14. Iwasaki, T., Hrorikawa, H., Matsumoto, K. & Miyoshi, M. Electrochemical
synthesis and reactivity of α-alkoxy α-amino acid derivatives. Bull. Chem. Soc.
Jpn. 52, 826–830 (1979).
15. Tajima, T., Kurihara, H. & Fuchigami, T. Development of an electrolytic system for
non-Kolbe electrolysis based on the acid−base reaction between carboxylic
acids as a substrate and solid-supported bases. J. Am. Chem. Soc. 129,
6680–6681 (2007).
16. Shtelman, A. V. & Becker, J. Y. Electrochemical synthesis of 1,2-disilylethanes
from α-silylacetic acids. J. Org. Chem. 76, 4710–4714 (2011).
17. Torii, S., Inokuchi, T., Mizuguchi, K. & Yamazaki, M. Electrolytic decarboxylation
reactions. 4. Electrosyntheses of 3-alkyl-2-cycloalken-1-ol acetates from
1-alkyl-2-cycloalkene-1-carboxylic acids. Preparation of dl-muscone from
cyclopentadecanone. J. Org. Chem. 44, 2303–2307 (1979).
18. Coleman, J. P., Lines, R., Utley, J. H. P. & Weedon, B. C. L. Electro-organic
reactions. Part II. Mechanism of the kolbe electrolysis of substituted
phenylacetate ions. J. Chem. Soc., Perkin Trans. 2 1064–1069
(1974).
large-scale etherification, the silver salt additive could be left out with
a only minimal effect on yield (78% in the presence of silver salt, com-
pared with 72% when this additive is omitted).
Extensive mechanistic studies were also undertaken in order to under-
stand the role of the additives and the nature of the reactive interme-
diate. In summary, the mechanism is likely to involve the rate-limiting
oxidation of a carboxylate on the anode to generate a carbocation,
followed by nucleophilic attack by an alcohol to afford the ether product
(see Supplementary Information for full details).
It is anticipated that the mild electrogeneration of carbocations
reported herein will find use in numerous settings in which standard
SN2 and carbocation-based approaches are unsuccessful in forming
hindered functionalized carbogenic frameworks.
Online content
Any methods, additional references, Nature Research reporting summaries,
source data, extended data, supplementary information, acknowledgements, peer
review information; details of author contributions and competing interests; and
statements of data and code availability are available at https://doi.org/10.1038/
s41586-019-1539-y.
19. Mao, R., Balon, J. & Hu, X. Decarboxylative C(sp3)–O cross-coupling. Angew.
Chem. Int. Ed. 57, 13624–13628 (2018).
20. Yan, M., Kawamata, Y. & Baran, P. S. Synthetic organic electrochemical methods
since 2000: on the verge of a renaissance. Chem. Rev. 117, 13230–13319
(2017).
21. Ross, S. D. & Finkelstein, M. Anodic oxidations. V. The Kolbe oxidation of
phenylacetic acid and 1-methylcyclohexaneacetic acid at platinum and at
carbon. J. Org. Chem. 34, 2923–2927 (1969).
22. Schäfer, H. J. Recent contributions of Kolbe electrolysis to organic synthesis.
Top. Curr. Chem. 152, 91–151 (1990).
23. Koehl, W. J. Anodic oxidation of aliphatic acids at carbon anodes. J. Am. Chem.
Soc. 86, 4686–4690 (1964).
24. Iwasaki, T., Horikawa, H., Matsumoto, K. & Miyoshi, M. An electrochemical
synthesis of 2-acetoxy-2-amino acid and 3-acetoxy-3-amino acid derivatives.
J. Org. Chem. 42, 2419–2423 (1977).
25. Huang, X., Liu, W., Hooker, J. M. & Groves, J. T. Targeted fuorination with the
fuoride ion by manganese-catalyzed decarboxylation. Angew. Chem. Int. Ed. 54,
5241–5245 (2015).
26. Eberson, L. & Nyberg, K. Studies on the Kolbe electrolytic synthesis. V. An
electrochemical analogue of the Ritter reaction. Acta Chem. Scand. 18,
1567–1568 (1964).
Received: 21 March 2019;Accepted: 12 July 2019;
Published online xx xx xxxx.
1. Roughley, S. D. & Jordan, A. M. The medicinal chemist’s toolbox: an analysis of
reactions used in the pursuit of drug candidates. J. Med. Chem. 54, 3451–3479
(2011).
2. Fischer, J. & Ganellin, C. R. (eds) Analogue-based Drug Discovery 206–217 (Wiley,
2006).
3. Williamson, W. Ueber die theorie der aetherbildung. Justus Liebigs Ann. Chem.
77, 37–49 (1851).
4. Kürti, L. & Czakó, B. Strategic Applications of Named Reactions in Organic
Synthesis 484–485 (Elsevier, 2005).
5. Swamy, K. C. K., Kumar, N. N. B., Balaraman, E. & Kumar, K. V. P. P. Mitsunobu and
related reactions: advances and applications. Chem. Rev. 109, 2551–2651 (2009).
6. Beyerman, H. C. & Heiszwolf, G. J. Reaction of steroidal alcohols with isobutene.
Usefulness of t-butyl as a hydroxyl-protecting group in a synthesis of
testosterone. Recl. Trav. Chim. Pays-Bas 84, 203–212 (1965).
7. Smith, M. B. & March, J. March’s Advanced Organic Chemistry 1037–1041
(Wiley, 2007).
8. Abraham S. et al. Aurora kinase compounds and methods of their use.
International patent no. WO2011088045A1 (2011).
9. Kolbe, H. Beobachtungen über die oxydirende wirkung des sauerstofs, wenn
derselbe mit hülfe einer elektrischen säule entwickelt wird. J. Prakt. Chem. 41,
137–139 (1847).
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
10. Hofer, H. & Moest, M. Mittheilung aus dem elektrochemischen Laboratorium
der Königl, Technischen Hochschule zu München. Justus Liebigs Ann. Chem.
323, 284–323 (1902).
© The Author(s), under exclusive licence to Springer Nature Limited 2019
N A t U r e | www.nature.com/nature

reSeArCH Letter
Acknowledgements Financial support for this work was provided by Pfizer,
Inc., the National Science Foundation (CCI Phase 1 grant 1740656), and the
National Institutes of Health (grant number GM-118176). China Scholarship
Council and Jilin University supported a fellowship to J.X., Zhejiang Yuanhong
Medicine Technology Co. Ltd supported a fellowship to M.S., The Hewitt
Foundation supported a fellowship to Y.K., The Swedish Research Council
supported a fellowship to H.L., Fulbright Fellowship supported a fellowship to
P.M., and S.H.R. acknowledges an NSF Graduate Research Fellowship Program
(no. 2017237151) and a Donald and Delia Baxter Fellowship. We thank
D.-H. Huang and L. Pasternack for assistance with NMR spectroscopy; J. Chen
for measuring the high-resolution mass spectroscopy data, and A. Rheingold,
C. E. Moore and M. A. Galella for X-ray crystallographic analysis.
Methods
Here we describe a typical procedure for the decarboxylative etherification. Further
experimental details are provided in the Supplementary Information.
General procedure for decarboxylative etherification. With no precautions to
exclude air or moisture, the ElectraSyn vial (5 ml) with a stir bar was charged
with carboxylic acid (0.2 mmol, 1.0 equiv.), alcohol (0.6 mmol, 3.0 equiv.), 2,4,6-
collidine (0.6 mmol, 3.0 equiv.), nBu4NPF6 (0.3 mmol, 1.5 equiv.), 3 Å molecular
sieves (150 mg), AgPF6 (0.3 mmol, 1.5 equiv.) and CH2Cl2 (3.0 ml). The ElectraSyn
vial cap equipped with anode (graphite) and cathode (graphite) were inserted into
the mixture. After pre-stirring for 15 min, the reaction mixture was electrolysed
at a constant current of 10 mA for 3 h. The ElectraSyn vial cap was removed, and
electrodes were rinsed with Et2O (2 ml), which was combined with the crude
mixture. Then, the crude mixture was further diluted with Et2O (30 ml). The
resulting mixture was washed with 2 M HCl (20 ml) and NaHCO3 (aq.) (20 ml),
dried over Na2SO4 and concentrated in vacuo. The crude material was purified
Author contributions J.X., M.S. and P.S.B. conceived the project. J.X., M.S., Y.K.,
H.L., D.G.B. and P.S.B. designed the experiments and analysed the data. J.X. and
M.S. developed the electrochemical decarboxylative methods and performed
their applications. H.L. and Y.K. carried out the mechanistic study. J.X., M.S.,
S.H.R., M.C., P.M., G.B., M.R.C., A.D., M.D.B., G.M.G., J.E.S., J.S. and S.Y. conducted
by preparative thin-layer chromatography to furnish the desired product. Full experiments to demonstrate the substrate scope. P.S.B. wrote the manuscript.
J.X., M.S., Y.K., S.H.R., P.M., H.L. and D.G.B. assisted in writing and editing the
experimental details and characterization of new compounds can be found in
the Supplementary Information.
manuscript.
Competing interests The authors declare no competing interests.
Data availability
Additional information
supplementary information is available for this paper at https://doi.org/
10.1038/s41586-019-1539-y.
Correspondence and requests for materials should be addressed to P.S.B.
Reprints and permissions information is available at http://www.nature.com/
reprints.
The data that support the findings of this study are available within the paper and
its Supplementary Information. Metrical parameters for the structures of (2R)-77
and (11R)-138 are available free of charge from the Cambridge Crystallographic
Data Centre (https://www.ccdc.cam.ac.uk/) under reference numbers 1918528
and 1903823, respectively.