The direct anti-Markovnikov addition of mineral acids to styrenes

Article abstract of DOI:10.1038/nchem.2000

The direct anti-Markovnikov addition of strong Bronsted acids to alkenes remains an unsolved problem in synthetic chemistry. Here, we report an efficient organic photoredox catalyst system for the addition of HCl, HF and also phosphoric and sulfonic acids to alkenes, with complete regioselectivity. These transformations were developed using a photoredox catalyst in conjunction with a redox-active hydrogen atom donor. The nucleophile counterion plays a critical role by ensuring high reactivity, with 2,6-lutidinium salts typically furnishing the best results. The nature of the redox-active hydrogen atom donor is also consequential, with 4-methoxythiophenol providing the best reactivity when 2,6-lutidinium salts are used. A novel acridinium sensitizer provides enhanced reactivity within several of the more challenging reaction manifolds. This Article demonstrates how nucleophilic addition reactions mediated by photoredox catalysis can change the way electrophilic and homofugal precursors are constructed.

Full text of DOI:10.1038/nchem.2000

ARTICLES
PUBLISHED ONLINE: 13 JULY 2014 | DOI: 10.1038/NCHEM.2000
The direct anti-Markovnikov addition of mineral
acids to styrenes
Dale J. Wilger, Jean-Marc M. Grandjean, Taylor R. Lammert and David A. Nicewicz
*
The direct anti-Markovnikov addition of strong Brønsted acids to alkenes remains an unsolved problem in synthetic
chemistry. Here, we report an efficient organic photoredox catalyst system for the addition of HCl, HF and also phosphoric
and sulfonic acids to alkenes, with complete regioselectivity. These transformations were developed using a photoredox
catalyst in conjunction with a redox-active hydrogen atom donor. The nucleophile counterion plays a critical role by
ensuring high reactivity, with 2,6-lutidinium salts typically furnishing the best results. The nature of the redox-active
hydrogen atom donor is also consequential, with 4-methoxythiophenol providing the best reactivity when 2,6-lutidinium
salts are used. A novel acridinium sensitizer provides enhanced reactivity within several of the more challenging reaction
manifolds. This Article demonstrates how nucleophilic addition reactions mediated by photoredox catalysis can change the
way electrophilic and homofugal precursors are constructed.
lkyl halides are ubiquitous building blocks throughout
organic chemistry owing to their participation in radical
reactions1, nucleophilic substitution reactions2 and organo-
Previous anti-Markovnikov additions: heteronucleophiles
H
‘Triple relay
catalysis’
H2O
A
OH
metallic cross-coupling reactions3–5. Organofluorine compounds
are far more stable than the heavier halide congeners and are infre-
quently viewed as carbocation or radical synthons. Nevertheless,
organofluorine compounds have garnered special attention due to
their indispensable function in modern pharmaceutical chemistry6.
Despite their extensive utility, methods to synthesize alkyl halides
from alkenes via anti-Markovnikov-selective hydrohalogenation
reactions remain inadequate. Instead, alkyl halides are customarily
accessed by the free radical chlorination and bromination of acti-
vated C–H bonds7,8. Benzylic halides can also be accessed via the
Markovnikov addition of anhydrous hydrogen halides to styrenes9.
Alkenes are inexpensive and readily available starting materials, yet
there is currently no direct and general method for the selective anti-
Markovnikov addition of most hydrogen halides to alkenes.
Catalytic variants of this transformation represent an unmet need
in chemical synthesis10. For example, the peroxide-initiated addition
of hydrogen halides to alkenes proceeds with anti-Markovnikov
regioselectivity under fairly harsh conditions and is inherently
restricted to hydrogen bromide. Anti-Markovnikov addition reac-
tions of the other hydrogen halides, including HCl, are thermodyna-
Ref. 18
Hydration
H
‘Photoredox
catalysis’
Me
OAc
HOAc
Me
Hydroacetoxylation
Refs 22–24
Current report: anti-Markovnikov hydrohalogenation
R2
R2
H
Catalyst 1
R3
R3
Catalyst 2
R3NH X
R1
R1
X
450 nm LEDs
Me
X = Cl (18 examples)
F (18 examples)
1a
O
Me
Me
OR
P
O
OR
N
(3 examples)
Me
mically unfavourable and severely limited in scope and utility11,12
.
BF4
O
O
Modern procedures for the formal anti-Markovnikov addition of
hydrogen chloride necessitate the stoichiometric oxidation of orga-
nometallic reagents prepared by hydroboration13, hydrozircona-
tion14 or by the deoxychlorination of preformed alcohols (for
example, the Appel reaction)15,16. The products of formal anti-
Markovnikov hydrofluorination are similarly accessed by the
S
O
R
2
(2 examples)
S
S
H
R3
R3
or
2
activation and nucleophilic displacement of preformed alcohols Figure 1 | Previous and current reports on the anti-Markovnikov addition
(deoxyfluorination), although these procedures suffer from poor of nucleophiles to alkenes. Grubbs and co-workers have reported a formal
chemoselectivity17.
anti-Markovnikov hydration reaction for terminal styrenes that operates via a
The lack of catalytic methods for the anti-Markovnikov hydro- palladium–Brønsted acid–ruthenium triple relay catalysis mechanism18. Our
functionalization of alkenes has recently been addressed by several group has previously reported anti-Markovnikov hydroetherification,
groups, including our own. The Grubbs laboratory has reported a hydroamination and hydroacetoxylation reactions for alkenes that employ a
triple relay catalysis reaction that provides formal anti- photoredox catalyst in conjunction with a redox-active hydrogen atom donor.
Markovnikov hydration products from styrenes (Fig. 1)18. Hartwig Reported here is the first catalytic system for the anti-Markovnikov addition
has reported a number of Ru- and Rh-catalysed anti-Markovnikov of chloride, fluoride and several other strong Brønsted acids.
Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-3290, USA. e-mail: nicewicz@unc.edu
*
720
NATURE CHEMISTRY | VOL 6 | AUGUST 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2000
ARTICLES
Nucleophilic
addition
‘
’
X
Me
X
Alkene
radical cation
Me
H
Me
Me
X
S
PET
3a
HAT
Me
Me
Me
N
Me
Me
*
Ph
SH
2a
Me
S
S
Oxidation
Reduction
Me
BF4
Ph
N
Me
Me
hν
450 nm
Me
S
+H+
Me
N
Me
BF4
1a
Figure 2 | Proposed catalytic cycle for the anti-Markovnikov hydrohalogenation of styrenyl alkenes. The catalytic cycle begins with electronic excitation of
photoredox catalyst 1a by visible light (450 nm). The excited-state reduction potential (*E1/2RED = 2.0 V versus SCE) of the catalyst favours quenching by
photoinduced electron-transfer (PET) from the alkene. The resultant alkene radical cation is electrophilic at the homobenzylic position, and nucleophilic attack
by the halogen produces the more stable benzylic carbon-centred radical. The anti-Markovnikov hydrohalogenation product is produced via hydrogen atom
transfer (HAT) from the redox-active thiol 2a. The effective turnover of the photoredox catalyst by oxidation is believed to be coupled to the reduction of the
thiyl radical. A source of acidic protons is necessary to convert latent thiolate to the active hydrogen atom donor. In many cases a disulfide species may be
used as the redox-active hydrogen atom donor because homolysis of the weak S–S bond produces a competent catalytic component.
hydroamination reactions19,20. Recently, Buchwald has also reported incapable of initiating undesired Markovnikov reactivity27.
a Cu-catalysed method for the anti-Markovnikov hydroamination Reported herein is our approach to overcoming these challenges,
of aliphatic alkenes21. Our group has reported several methods for and the strategies we have found successful for anti-Markovnikov
the intra- and intermolecular anti-Markovnikov addition of hydrochlorination, hydrofluorination and addition of other highly
various oxygen and nitrogen nucleophiles using a photoredox cata- Brønsted acidic nucleophiles.
lyst in conjunction with various redox-active hydrogen atom
donors22–24
.
Although the expanding repertoire of anti- Results and discussion
Markovnikov heteronucleophile additions is impressive, arguably Initial focus was placed on the anti-Markovnikov hydrochlorination
the most generally useful anti-Markovnikov alkene hydrofunctiona- of β-methylstyrene derivatives (3), as the chloride anion has a
lization reactions would include the halogens and sulfonates, nucleophilicity very close to the acetate anion28, and styrene
because they are high-utility products and versatile intermediates cation radicals generated via laser flash photolysis are known to
for further chemical manipulation. Nonetheless, no direct catalytic react with the chloride anion29. Additionally, the homobenzylic
anti-Markovnikov hydrofunctionalization reactions have been chlorination products are challenging to access by other methods.
reported for the chloride, fluoride or sulfonate anions.
Using the 9-mesityl-10-methylacridinium (1a)30,31 catalyst first
Our previously reported anti-Markovnikov heteronucleophile reported by Fukuzumi, we were pleased to find that employing in
additions have relied on the intermediacy of alkene cation radicals situ generated anhydrous hydrogen chloride (from pivaloyl chloride
generated by single-electron oxidation by an acridinium photoredox and 2,2,2-trifluoroethanol, TFE) gave the anti-Markovnikov hydro-
catalyst (1a). Heteroatom nucleophiles react with alkene cation rad- halogenation adduct (Table 1, Method A). The desired photoredox
icals with anti-Markovnikov regioselectivity25,26. An appropriate pathway successfully outcompetes the undesired Markovnikov
redox-active hydrogen atom donor (2) has been critical to providing addition reaction with minimal optimization, as long as a thiophe-
synthetically useful yields. We hypothesized that a similar reaction nol co-catalyst is included as the redox-active hydrogen atom donor
manifold could access an anti-Markovnikov addition of strong (4a). Various chlorinated solvents could be used, as well as trifluor-
Brønsted acids to alkenes. A proposed catalytic cycle is presented otoluene. The reaction solvent, concentration and the addition of a
in (Fig. 2). Strong Brønsted acids, including the halogens and sulfo- catalytic base all effect the anti-Markovnikov regioselectivity
nates, present a unique set of challenges regarding unproductive side (Supplementary Table 1). Halogenation is observed exclusively at
reactions and the low nucleophilicity of their conjugate bases. In our the α position (Markovnikov addition) when the photoredox cata-
previous studies, protic nucleophiles have always been used because lyst and thiophenol are excluded (see Supplementary Section VIII
the active hydrogen atom donor is regenerated in situ by reduction or page 11). Yields for styrene substrates with electron-withdrawing
of the thiyl radical and subsequent protonation (Fig. 2)22–24. Even groups were moderate, with little to no Markovnikov addition
the most acidic nucleophiles reported to date—the carboxylic (Table 1, 4b–h). Substrates with electron-releasing substituents pro-
acids (CH3CO2H: pKa = 4.75 (H2O), 12.3 (DMSO))—are typically vided lower yields and at times favoured the undesired Markovnikov
NATURE CHEMISTRY | VOL 6 | AUGUST 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
721

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2000
ARTICLES
Table 1 | The anti-Markovnikov hydrochlorination of styrenes.
Method A with
anhydrous HCl
Method B with
2,6-lutidine HCl
O
Cl
H
H
Cl
Catalyst 1a
Me
N
Me
Cl
tBu
R2
R2
R2
(5 mol%)
PivCl (2 equiv.)
TFE (4 equiv.)
Thiophenol (2a, 20 mol%)
2,6-lutidine (10 mol%)
CHCl3 [0.4 M],
R1
R1
R1
Method A or
Method B
Cl
H
+
(2 equiv.)
4-methoxythiophenol
(2b, 20 mol%)
3
Anti-Markovnikov
product (4)
Markovnikov
product (5)
hν 450 nm, 24 h
CHCl3/TFE (2.5:1)
[0.6 M], hν 450 nm, time h
H
H
H
H
Cl
H
H
Me
Me
Me
Cl
Me
Cl
Me
Br
Me
Cl
Cl
Cl
Cl
Cl
Cl
Br
4a
4b
4c
4d
4e
4f
Method A:
Method A:
Method A:
Method A:
Method A:
Method A:
*
*
*
*
*
*
70% , 58%†
59% , 58%†
65% , 61%†
72% , 71%†
64% , 63%†
62% , 63%†
4a:5a > 20:1‡
4b:5b > 20:1‡
4c:5c > 20:1‡
4d:5d > 20:1‡
4e:5e > 20:1‡
4f:5f > 20:1‡
Method B (5 h):
Method B (4 h):
Method B (16 h):
Method B (4 h):
Method B (24 h):
*
*
*
*
*
93% , 65%†
96% , 87%†
90% , 95%†
86% , 84%†
99% , 96%†
H
H
H
H
Me
H
H
Me
Me
Me
Me
Me
Me
Cl
Cl
Cl
Cl
Cl
Cl
F
Me
tBu
MeO2C
4g
4h
4i
4j
4k
4l
Method A:
Method A:
Method A:
28%
4i:5i = 0.7‡
Method A:
70%
4j:5j = 3.0‡
Method A:
45%
4k:5k = 1.1‡
Method A:
29%
*
*
*
*
*
*
74% , 64%†
51% , 53%†
4g:5g > 20:1‡
4h:5h > 20:1‡
Method B (48 h):
*
Method B (4 h):
Method B (6 h):
91% , 78%†
*
*
97% , 89%†, 93%§
85% , 85%†
Me
Me
H
H
H
Me
H
Cl
Cl
Cl
MeO
Me
Cl
Cl
H
H
Cl
Me
Br
Br
4m
4n
4o
4p
4q
4r
Method B (24 h):
Method B (24 h):
Method B (8 h):
Method B (24 h):
Method B (24 h):
Method B (48 h):
*
*
*
*
*
*
93% , 88%†
65% , 61%†
69% , 65%†
62% , 64%†
50% , 49%†
39% , 34%†
*Average yield of the desired anti-Markovnikov regioisomer 4 for two trials by 1H NMR against internal standard. †Isolated yield of the anti-Markovnikov regioisomer 4 for a single trial. ‡Average regioselectivity
for the anti-Markovnikov product (4:5) by 1H NMR for at least two trials (Method A only). §Isolated yield (2.98 g, 93%) for a single reaction prepared on the 18.9 mmol scale. At this scale, the reaction was
complete in 8 h using standard glassware (100 ml round-bottomed flask).
addition products (4i–l). Nonetheless, the inherent ability of the prohibited their further evaluation. We believe that
photoredox system to control regiochemistry in the presence of a 4-methoxythiophenol (2b) provides superior reactivity under
strong Brønsted acid was promising.
these conditions due to the increased basicity of its conjugate
thiolate32. The active hydrogen atom donor is presumably
regenerated in situ by protonation (Fig. 2), so the thiolate must be
capable of abstracting a proton from the weak ammonium acid.
The TFE co-solvent is included to solubilize the hydrochloride
salt. Using 2,6-lutidine hydrochloride (Table 1, Method B), the
anti-Markovnikov hydrohalogenation of β-methylstyrene is
complete in less than 5 h with a yield of 93%. Using the amine
hydrochloride salts, chloride addition is observed exclusively at
the β position and no competing Markovnikov addition was
observed for any substrate tested. This modification allowed for
the anti-Markovnikov hydrochlorination of electron-rich styrenes
with good chemical yields (4i,k,l). More interestingly, the use of
lutidinium hydrochloride allowed for the anti-Markovnikov
hydrochlorination of geminally disubstituted and monosubstituted
terminal styrenes (4n–r), a challenging subclass of alkenes within
photoredox catalysis22–24. Several α-methylstyrene derivatives gave
yields above 60%, while monosubstituted styrenyl alkenes
provided somewhat lower yields. Small amounts of thiol-ene
Anti-Markovnikov hydrochlorination using amine-HCl salts.
Despite the simplicity and complete atom economy inherent to
the direct addition of hydrogen chloride, we speculated that a
surrogate incapable of promoting competitive Brønsted acid-
catalysed
reactions
would
provide
anti-Markovnikov
hydrohalogenation with complete regioselectivity. Hydrochloride
salts derived from amines have the advantage of being easy to
handle, functional group-tolerant and, as a requirement for the
method, mildly acidic. A survey of various ammonium chloride
salts revealed that a combination of 2,6-lutidinium hydrochloride
with 4-methoxythiophenol (2b) as the redox-active hydrogen
atom donor was ultimately found to provide the best reaction
yields with acceptable reaction times (Fig. 3). Hydrochloride salts
derived from different amines provided vastly different rates and
chemical yields, with the 2,6-lutidine salt outperforming even the
closely related pyridine hydrochloride. Hydrochloride salts derived
from 2,6-di-tert-butylpyridine were similarly effective, but cost
722
NATURE CHEMISTRY | VOL 6 | AUGUST 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2000
ARTICLES
H
and danger associated with handling anhydrous hydrogen fluoride
has been mitigated by the development of several tempered poly-
hydrogen fluoride conjugates. Of particular note are Olah’s
reagent (pyridine-HF) and ether-hydrogen fluoride complexes,
which allow for the hydrofluorination of alkenes with
Me
Method B (5 h)
Me
Amine-HCl salt
(2 equiv.)
Cl
3a
4a
100
90
80
70
60
50
40
30
20
10
0
Markovnikov selectivity, albeit with limited substrate scope36,37
.
H
N
Yield
(4a)
More recently, the Boger38 and Hiroya39 groups have unveiled
radical-based approaches, which generate Markovnikov hydro-
fluorinated products under mild conditions. As previously men-
tioned, no direct catalytic anti-Markovnikov hydrofluorination has
been reported to date.
Me
Me
Conversion (of 3a)
Me
NH3
Me
H
N
NH4
Merging photoredox catalysis with synthetic fluorination chem-
istry presents unique challenges not encountered within the
anti-Markovnikov hydrochlorination. Athough contemporary
radical-based alkene hydrofluorination methods employ electrophi-
lic fluorinating reagents such as N-fluorosulfonimide, N-fluoropyr-
idinium or N-fluoroammonium salts, a nucleophilic source of the
fluoride ion was required within the current photoredox system.
Although the nonpolarizable character (hardness) and basicity of
the fluoride ion render it a poor nucleophile, we were encouraged
by pioneering reports from the Arnold laboratory40–42. Arnold
and co-workers demonstrated that alkenes oxidized under photo-
chemical conditions could undergo aryl-fluorination by an analo-
gous mechanism41. Even though hydrofluorination products were
not observed in the report by Arnold, that key precedent demon-
strated that anionic fluoride will add into alkene radical cations
with the desired anti-Markovnikov regioselectivity.
H2N(iPr)2
HNEt3
Figure 3 | Effect of different amine hydrochloride salts on the anti-
Markovnikov hydrochlorination of β-methylstyrene. The average yield of
the anti-Markovnikov product (4a) and the conversion of alkene (3a) are
given for several amine hydrochloride salts after 5 h reaction time. Both
yields and conversions are the average for at least two trials by 1H NMR.
by-products were observed in the reactions of terminal styrenes. The
remaining mass balance was attributed to dimerization and
oligomerization. We are currently investigating modifications to
the photoredox catalyst and redox-active hydrogen atom donor,
which will provide higher chemical yields for terminal styrenes.
Other electron-rich non-styrenyl substrates such as trialkyl-sub-
stituted alkenes, enol ethers and enamides are oxidizable, but failed
to produce anti-Markovnikov hydrochlorination products using
Method B. The alkene radical cations emanating from these sub-
strates are highly stabilized and likely to be unreactive towards
nucleophilic attack by the chloride anion. Although substrates
with strongly electron-donating ortho and para methoxy substitu-
ents such as anethole (not shown) are easily oxidized by the photo-
redox catalyst and are reactive towards other nucleophiles, including
the acetate anion24, when using Method B these substrates proved
extremely sluggish. For example, anethole and the ortho-methoxy
analogue failed to proceed beyond 60% conversion (<48% yield,
1H NMR) even when high catalyst loadings (10 mol%) and long
reaction times (48 h) were used.
Our initial focus was on identifying a suitable fluoride source for
the transformation. Any number of common fluoride sources
including quaternary ammonium compounds, alkali–metal salts
paired with crown ethers or phase transfer reagents, triphenyldi-
fluorosilicate and Olah’s reagent caused catalyst decomposition or
otherwise failed to yield product. Analysis of catalyst (1a) solutions
1
by H NMR after treatment with caesium fluoride in acetonitrile
indicated demethylation to form 9-mesitylacridine. Treatment of
an N-phenyl analogue (1b) with caesium fluoride provided
adducts consistent with nucleophilic dearomatization of the acridi-
nium moiety43. These observations prompted us to investigate the
9-mesityl-2,7,10-trimethylacridinium catalyst (1c) reported by
Fukuzumi44 as well as a previously unreported 9-mesityl-2,7-
dimethyl-10-phenylacridinium catalyst (1d), in the hope of achiev-
ing increased catalyst turnover. Information regarding the full
optimization of this transformation with regards to catalyst identity,
solvent and the redox-active hydrogen atom donor are all provided
in the Supplementary Section X and XI or page 19.
Anti-Markovnikov hydrofluorination. The incorporation of
fluorine into pharmaceutical lead compounds as a bioisosteric
replacement for hydrogen atoms, hydroxyl groups and the other
halogens has become widespread in medicinal chemistry. This is
due to the element’s ability to modulate biological and physical
properties with minimal steric impact33,34. Fluorination is also
employed within materials and agrochemistry to modulate the
properties of polymers and pesticides, respectively. Another highly
specialized application of fluorinated compounds includes
positron emission tomography. Positron emission tomography
requires the use of synthetic methods to enrich 18F within
molecules used as radiotracers (for example, fluorodeoxyglucose,
FDG). Being the most electronegative element, fluorine exacts
special constraints during its incorporation within organic
molecules. Despite the importance of organofluorine chemistry,
there is still a widespread need for more practical and selective
fluorination methods35. We envisaged an immediate extension of
the methodology described above to provide the direct and
regioselective anti-Markovnikov hydrofluorination of alkenes
under mild conditions.
Ultimately, triethylamine trihydrofluoride (NEt3·3HF), a mild
and slightly acidic reagent, was found to be the only fluoride
source suitable for this transformation. This reagent is available
commercially and does not typically etch glassware45. The optimal
hydrogen atom donor for the anti-Markovnikov hydrofluorination
was 4-nitrophenyl disulfide (2c). All four of the catalysts (1a–d)
tested provided the desired anti-Markovnikov hydrofluorination
product under the optimized conditions. The more sterically hin-
dered catalyst (1d) generally provided higher chemical yields
when fluoride and other more basic nucleophiles (vide infra) were
employed, and was therefore used throughout the subsequent
hydrofluorination scope (Table 2). The spectroscopic and electro-
chemical properties of the different photoredox catalysts were not
significantly different (see Supplementary Section IV or pages 4–
8). The increased product formation is attributed to improved cata-
lyst stability and turnover number. Under the optimized conditions,
β-methylstyrene gave a 73% yield of the desired alkyl fluoride
(Table 2, 6a) with full conversion in less than 2 h. Similar to hydro-
chlorination using amine salts, exclusive anti-Markovnikov regios-
electivity was observed, with no evidence of fluorine incorporation
The synthetic attractiveness of directly fluorinating alkene C–C
π-bonds has been recognized previously. The considerable reactivity
at the
α
position for any styrenes tested in the alkene
NATURE CHEMISTRY | VOL 6 | AUGUST 2014 | www.nature.com/naturechemistry
723
© 2014 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2000
ARTICLES
Table 2 | The anti-Markovnikov hydrofluorination of styrenes.
Method C
Mes
Catalyst 1d
1d
H
Et3N•3HF (1–2 equiv.)
Me
R2
(5 mol%)
Me
Me
4-nitrophenyl
disulfide (2c, 25 mol%)
CHCl3 [0.25 M],
R1
R1
N
Method C
F
BF4
Ph
hν 450 nm, time h
(3)
Anti-Markovnikov
product (6)
H
H
H
H
Cl
H
H
Me
Me
Me
Cl
Me
Me
Br
Me
F
F
F
F
F
F
Cl
Br
6a
6b
6c
6d
6e
6f
Method C (2 h):
Method C (12 h):
Method C (12 h):
Method C (12 h):
Method C (24 h):
Method C (24 h):
*
*
*
*
*
*
73% , 44%†
74% , 58%†
76% , 63%†
75% , 59%†
75% , 55%†
65% , 47%†
H
H
H
H
Me
H
H
Me
Me
Me
Me
Me
MeO
Me
F
F
F
F
F
F
F
Me
tBu
MeO2C
6g
6h
6i
6j
6k
6m
Method C (12 h):
Method C (12 h):
Method C (6 h):
Method C (6 h):
Method C (6 h):
Method C (12 h):
*
*
*
*
*
*
85% , 30%†
43% , 36%†
66% , 43%†
71% , 43%†
63% , 57%†, 69%‡
52% , 41%†
OMe
H
H
Me
H
H
H
Me
Me
Me
F
Me
Me
S
F
H
S
F
F
F
F
6s
6t
6n
6u
6v
6w
Method C (12 h):
Method C (6 h):
Method C (12 h):
Method C (12 h):
Method C (12 h):
Method C (12 h):
*
*
*
*
*
*
75% , 46%†
54% , 29%†
43% , 36%†
65% , 48%†
34% , 17%†
54% , 46%†
*Average yield of the desired anti-Markovnikov hydrofluorination product 6 for two trials by 1H NMR against internal standard. †Isolated yield of the anti-Markovnikov hydrofluorination product 6 for a single trial.
‡Isolated yield (1.34 g, 69%) for a single reaction prepared on the 10.0 mmol scale. At this scale, the reaction was complete in 12 h using standard glassware (100 ml round-bottomed flask).
hydrofluorination. A variety of substituted β-methylstyrene deriva- atom donor provided no observable yield within control exper-
tives were tested. Derivatives possessing both electron-withdrawing iments for both the anti-Markovnikov hydrochlorination and
(6b–h) and electron-releasing (6i–k,t,u) substituents are well toler- hydrofluorination. We are currently investigating new reaction con-
ated. A para-fluorinated substrate undergoes fluorination in high ditions and attempting to expand the scope of these transformations
yield (6g, 85%), whereas a substrate with a more electron-withdraw- to include more alkyl-substituted alkenes and terminal styrenes.
ing carbonyl substituent was slower to react (6h, 43%). Although
monosubstituted styrenes were not accessible substrates using Anti-Markovnikov addition of mineral acid equivalents. To
Method C, α-methylstyrene provided a moderate yield for the expand the generality of the photoredox catalysis system, the
desired primary fluoride (6n, 43%). Knowing the potential influence addition of organic oxyanions to alkenes was tested. The analysis
of this method within medicinal chemistry, two heteroaromatic sub- began with phosphates, which are similar in acidity to HF, and
strates were tested and found to be competent reaction partners (6v, then sulfonates with increasing acidity were examined. Organic
w). The anti-Markovnikov hydrofluorinated adducts displayed phosphates are common synthetic intermediates in organic
increased volatility and were difficult to visualize during chromato- chemistry and are ubiquitous biologically relevant compounds48.
graphy. This resulted in lower isolated yields than were observed by Similar to HF, the organic phosphates tested were incapable of
1H NMR.
promoting Markovnikov addition reactions in the halogenated
Our previously reported anti-Markovnikov functionalization solvents tested. The formation of an amine salt was not necessary;
methods have all successfully included tri-alkyl substituted instead, the phosphoric acids were initially used with only a
alkenes. These are more demanding substrates under standard catalytic amount of base (Table 3, Method D). Similar to the anti-
photoredox conditions due to their higher oxidation potentials Markovnikov hydrofluorination, the photoredox catalyst 1d
OX
1/2
(E ≈ 2.0 V versus SCE) and steric impediment during reaction. typically gave higher yields when phosphoric acids were used as
We were pleased to observe the formation of the anti- nucleophiles (Supplementary Table 4). Diphenyl disulfide (2d)
Markovnikov hydrofluorination product when methylcyclopentene was found to be the optimal redox-active hydrogen atom donor
(6x, see Supplementary Section, page 25: 42% yield, 70% conver- under these conditions, rendering the anti-Markovnikov
sion) was examined. Unfortunately, the product could not be iso- hydrooxyphosphorylation products in good yields when dibutyl
lated due to challenges described in the previous paragraph. The and dibenzyl phosphate were used (7a and 8a, 86% and 89%,
identity of the desired product was confirmed by 1H and 19F respectively). Diphenyl phosphate gave a much lower yield for the
NMR and by comparison with literature values46. It is worth β-oxyphosphorylated product (9a, 12%), indicating a strong
noting that although our laboratory has previously reported TFE electron-withdrawing effect of the phenyl substituents. Similar to
as a potential hydrogen atom donor within a mechanistically dis- the anti-Markovnikov hydrochlorination, we found that using the
similar transformation47, omission of a redox-active hydrogen 2,6-lutidine salt of diphenyl phosphate (Method E) was necessary
724
NATURE CHEMISTRY | VOL 6 | AUGUST 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2000
ARTICLES
these transformations and depends strongly on the identity of the
nucleophile used. We expect this methodology to fundamentally
impact the manner in which chemists construct electrophilic
reagents. Our group is currently working on adaptations of this
methodology, which will allow for the control of absolute stereo-
chemistry during anti-Markovnikov addition reactions.
Table 3 | The anti-Markovnikov addition of phosphoric and
sulfonic acids.
H
Me
Conditions:
Method D* or
Method E†
X
Me
H
H
Methods
OBu
A
O
O
The general procedure for the anti-Markovnikov hydrochlorination of styrenes
using 2,6-lutidine hydrochloride (Method B) was as follows. In a typical experiment,
a flame-dried 2 dram vial was equipped with a magnetic stir bar and 9-mesityl-10-
methylacridinium tetrafluoroborate (1a, 30.8 mg, 0.0771 mmol, 5.0 mol%). The vial
was placed under an inert atmosphere and 2,6-lutidine hydrochloride (443 mg, 3.08
mmol, 2.0 equiv.) was added. The solvent CHCl3/TFE (2.5:1 (vol:vol), 2.8 ml) was
added to a concentration of ∼0.55 M. Liquid substrate (β-methylstyrene 3a, 200 µl,
1.54 mmol, 1 equiv.) was added via microsyringe after the solvent. The hydrogen
atom donor, 4-methoxythiophenol (2b, 37.9 µl, 0.308 mmol, 20 mol%), was added
via microsyringe. The vial was sealed with a Teflon-coated septum cap, and the
reaction mixture was irradiated (450 nm) for the time indicated. Upon completion,
saturated aqueous sodium bicarbonate was added carefully and the two phases were
allowed to separate. The organic phase was collected, and the aqueous phase was
extracted with two portions of dichloromethane. The combined organic portions
were passed through a short plug of SiO2. The solvent was removed under reduced
pressure. NMR yields were taken by integration against a known quantity of
hexamethyldisiloxane (HMDS) in CDCl3. The final products were isolated by silica
gel chromatography using the conditions listed. The procedures for the
anti-Markovnikov hydrofluorination (Method C), anti-Markovnikov
hydrooxyphosphorylation (Methods D and E) and anti-Markovnikov
hydrooxysulfonylation (Method E) are all presented in the Supplementary
Sections VII and XV.
P
P
OBu
OBn
O
O
Me
Me
7a
8a
Method D (48 h):
86%‡, 79%§
Method D (8 h):
89%‡, 71%§
H
H
OPh
O
O
O
S
P
OPh
Me
O
O
Me
Me
10a
9a
Method D (48 h):
12%§
Method E (24 h):
72%‡, 67%§
Method E (16 h):
94%‡,86%§
H
H
O
O
O
O
S
S
Ph
CF3
O
O
Me
11a
Me
Compounds 1a, 1b and 1d are available commercially from Aldrich with the
following catalogue numbers, respectively: 794171, 793221 and 793876.
Not observed
(Method E)
Method E (24 h):
37%‡, 35%§
Received 10 January 2014; accepted 16 May 2014;
published online 13 July 2014
*Method D: The alkene (β-methylstyrene) was combined with catalyst (1d, 5 mol%), diphenyl
disulfide (2d, 10 mol%), catalytic base (2,6-lutidine, 10 mol%) and either dibutyl, dibenzyl or
diphenyl phosphate (2 equiv., free acid) in CH2Cl2 (0.7 M) and then irradiated at 450 nm for
the time indicated. †Method E: The alkene (β-methylstyrene) was combined with catalyst (1d,
5 mol%), 4-methoxythiophenol (2b, 20 mol%), and the appropriate 2,6-lutidine salt of the
desired phosphate or sulfonate nucleophile (2 equiv.) in CH2Cl2 (0.5 M) and then irradiated at
450 nm for the time indicated. ‡Average yield of the desired anti-Markovnikov addition product
for two trials by 1H NMR against internal standard. §Isolated yield of the anti-Markovnikov
addition product for a single trial.
References
1. Togo, H. Advanced Free Radical Reactions for Organic Synthesis (Elsevier, 2004).
2. Carey, F. A. & Sundberg, R. J. Advanced Organic Chemistry 5th edn
(Springer, 2007).
3. Rudolph, A. & Lautens, M. Secondary alkyl halides in transition-metal-catalyzed
cross-coupling reactions. Angew. Chem. Int. Ed. 48, 2656–2670 (2009).
4. Terao, J. & Kambe, N. Cross-coupling reaction of alkyl halides with Grignard
reagents catalyzed by Ni, Pd, or Cu complexes with π-carbon ligand(s). Acc.
Chem. Res. 41, 1545–1554 (2008).
5. Binder, J. T., Cordier, C. J. & Fu, G. C. Catalytic enantioselective cross-couplings
of secondary alkyl electrophiles with secondary alkylmetal nucleophiles: Negishi
reactions of racemic benzylic bromides with achiral alkylzinc reagents. J. Am.
Chem. Soc. 134, 17003–17006 (2012).
6. Purser, S., Moore, P. R., Swallow, S. & Gouverneur, V. Fluorine in medicinal
chemistry. Chem. Soc. Rev. 37, 320–330 (2008).
7. Smith, M. B. & March, J. in Substitution Reactions: Free Radicals in Advanced
Organic Chemistry 6th edn, Ch. 14 (Wiley, 2007).
8. Fuller, S. E., Smith, J. R. L., Norman, R. O. C. & Higgins, R. Selectivity in the
halogenation of hexane by tertiary aminium radicals from the
photodecomposition of N-halogenoammonium perchlorates. J. Chem. Soc.
Perkin Trans. 2, 545–552 (1981).
9. Kharasch, M. S. & Kleiman, M. Synthesis of polyenes. III. A new synthesis of
diethylstilbestrol. J. Am. Chem. Soc. 65, 11–15 (1943).
10. Beller, M., Seayad, J., Tillack, A. & Jiao, H. Catalytic Markovnikov and anti-
Markovnikov functionalization of alkenes and alkynes: recent developments and
trends. Angew. Chem. Int. Ed. 43, 3368–3398 (2004).
11. Kharasch, M. S. & Mayo, F. R. The peroxide effect in the addition of reagents to
unsaturated compounds. I. The addition of hydrogen bromide to allyl bromide.
J. Am. Chem. Soc. 55, 2468–2496 (1933).
to obtain a high chemical yield (94%). The substituent effects
observed within the phosphoric acid scope prompted us to
investigate sulfonic acids, the acidity of which can be fine-tuned
by changing the nature of the alkyl or aryl substituent attached to
the central sulfur atom. The pKa values for sulfonic acids can vary
by more than ten orders of magnitude within a common solvent,
from approximately –2 for methane sulfonic acid to –13 for
trifluoromethane sulfonic acid (both in H2O)49,50. Sulfonates are
primarily employed as leaving groups, so their use within
nucleophilic addition reactions may provide
a variety of
interesting new methodologies. In accordance with our previous
hypothesis regarding the importance of thiolate basicity,
4-methoxythiophenol (2b) functions as the most efficient
hydrogen atom donor whenever a 2,6-lutidine salt is used as the
nucleophile. The anti-Markovnikov hydrooxymesylation product
is obtained with
a good yield (10a, 72%), while the
benzenesulfonate salt provides a much lower yield (11a, 37%) for
the reaction. The effects of electron-withdrawing groups on the
nucleophilicity of the sulfonate anions are further reinforced in
that no addition products are formed when the triflate salt is used.
12. Mayo, F. R. Free radical addition and transfer reactions of hydrogen chloride
with unsaturated compounds. J. Am. Chem. Soc. 76, 5392–5396 (1954).
13. Brown, H. C. & De Lue, N. R. Organoboranes for synthesis. 12. The reaction of
organoboranes with nitrogen trichloride. A convenient procedure for the
conversion of alkenes into alkyl chlorides via hydroboration. Tetrahedron 44,
2785–2792 (1988).
14. Zheng, J. et al. Synthesis of diverse well-defined functional polymers based on
hydrozirconation and subsequent anti-Markovnikov halogenation of 1,2-
polybutadiene. Macromolecules 45, 1190–1197 (2012).
Conclusions
We have demonstrated that a photoredox catalysis system in con-
junction with a redox-active hydrogen atom donor can easily be
adapted for the anti-Markovnikov addition of strong Brønsted
acids, including several halogens, phosphates and sulfonates. The
choice of redox-active hydrogen atom donor is critical within
15. Appel, R. Tertiary phosphane/tetrachloromethane, a versatile reagent for the
chlorination, dehydration, and P–N linkage. Angew. Chem. Int. Ed. 14,
801–811 (1975).
NATURE CHEMISTRY | VOL 6 | AUGUST 2014 | www.nature.com/naturechemistry
725
© 2014 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2000
ARTICLES
16. Squires, T. G., Schmidt, W. W. & McCandlish, C. S. Jr. Zinc chloride catalysis in
the reaction of thionyl halides with aliphatic alcohols. J. Org. Chem. 40,
134–136 (1975).
39. Shigehisa, H., Nishi, E., Fujisawa, M. & Hiroya, K. Cobalt-catalyzed
hydrofluorination of unactivated olefins: a radical approach of fluorine transfer.
Org. Lett. 15, 5158–5161 (2013).
17. Furuya, T., Kuttruff, C. A. & Ritter, T. Carbon–fluorine bond formation. Curr.
Opin. Drug Discov. 11, 803–819 (2008).
18. Dong, G., Teo, P., Wickens, Z. K. & Grubbs, R. H. Primary alcohols from
terminal olefins: formal anti-Markovnikov hydration via triple relay catalysis.
Science 333, 1609–1612 (2011).
19. Takaya, J. & Hartwig, J. F. Mechanistic studies of ruthenium-catalyzed anti-
Markovnikov hydroamination of vinylarenes: intermediates and evidence for
catalysis through π-arene complexes. J. Am. Chem. Soc. 127, 5756–5757 (2005).
20. Takemiya, A. & Hartwig, J. F. Rhodium-catalyzed intramolecular, anti-
Markovnikov hydroamination. Synthesis of 3-arylpiperidines. J. Am. Chem. Soc.
128, 6042–6043 (2006).
21. Zhu, S., Niljianskul, N. & Buchwald, S. L. Enantio- and regioselective
CuH-catalyzed hydroamination of alkenes. J. Am. Chem. Soc. 135,
15746–15749 (2013).
40. Mangion, D. & Arnold, D. R. Photochemical nucleophile–olefin combination,
aromatic substitution reaction. Its synthetic development and mechanistic
exploration. Acc. Chem. Res. 35, 297–304 (2002).
41. Chan, M. S. W. & Arnold, D. R. Photochemical nucleophile–olefin combination,
aromatic substitution (photo-NOCAS) reaction, part 15. Investigations
involving fluoride anion as the nucleophile and the effect of fluorine substitution
on the relative stability of the reaction intermediates. Can. J. Chem. 75,
1810–1819 (1997).
42. Arnold, D. R., McManus, K. A. & Chan, M. S. W. Photochemical nuclophile–
olefin combination, aromatic substitution (photo-NOCAS) reaction, part 13.
The scope and limitations of the reaction with cyanide anion as the nucleophile.
Can. J. Chem. 75, 1055–1075 (1997).
43. Lin, Y-C. & Chen, C-T. Acridinium salt-based fluoride and acetate
chromofluorescent probes: molecular insights into anion selectivity switching.
Org. Lett. 11, 4858–4861 (2009).
22. Hamilton, D. S. & Nicewicz, D. A. Direct catalytic anti-Markovnikov
hydroetherification of alkenols. J. Am. Chem. Soc. 134, 18577–18580 (2012).
23. Nguyen, T. M. & Nicewicz, D. A. Anti-Markovnikov hydroamination of
alkenes catalyzed by an organic photoredox system. J. Am. Chem. Soc. 135,
9588–9591 (2013).
44. Ohkubo, K. et al. Simultaneous production of p-tolualdehyde and hydrogen
peroxide in photocatalytic oxygenation of p-xylene and reduction of oxygen
with 9-mesityl-10-methylacridinium ion derivatives. Chem. Commun. 46,
601–603 (2010).
24. Perkowski, A. J. & Nicewicz, D. A. Direct catalytic anti-Markovnikov addition of
carboxylic acids to alkenes. J. Am. Chem. Soc. 135, 10334–10337 (2013).
25. Neunteufel, R. A. & Arnold, D. R. Radical ions in photochemistry. I. The 1,1-
diphenylethylene cation radical. J. Am. Chem. Soc. 95, 4080–4081 (1973).
26. Gassman, P. G. & Bottorff, K. J. Photoinduced lactonization. A useful but
mechanistically complex single electron transfer process. J. Am. Chem. Soc. 109,
7547–7548 (1987).
27. Pasto, D. J. & Gadberry, J. F Stereoselectivity and mechanism of acid-catalyzed
additions of acetic acid to (E)- and (Z)-2-butene in acetic acid. J. Am. Chem. Soc.
100, 1469–1473 (1978).
28. Pearson, R. G., Sobel, H. R. & Songstad, J. Nucleophilic reactivity constants
toward methyl iodide and trans-[Pt(py)2Cl2]. J. Am. Chem. Soc. 90,
319–326 (1968).
29. Johnston, L. J. & Schepp, N. P. Reactivity of radical cations: characterization of
styrene radical cations and measurements of their reactivity toward nucleophiles.
J. Am. Chem. Soc. 115, 6564–6571 (1993).
45. Franz, R. Über Trishydrofluoride tertiärer Amine und ihren Einsatz als
Fluorierungsmittel. J. Fluorine Chem. 15, 423–434 (1980).
46. Shellhamer, D. F. et al. Reaction of aminosulfur trifluorides with alcohols:
inversion vs. retention. J. Chem. Soc. Perkins Trans. 2, 973–977 (1996).
47. Wilger, D. J., Gesmundo, N. J. & Nicewicz, D. A. Catalytic
hydrotrifluoromethylation of styrenes and unactivated aliphatic alkenes via an
organic photoredox system. Chem. Sci. 4, 3160–3165 (2013).
48. Corbridge, D. E. C. Phosphorus: Chemistry, Biochemistry and Technology 6th edn
(CRC Press, 2013).
49. Marziano, N. C., Sampoli, M. & Gonizzi, M. On the Mc activity coefficient
function describing solute and solvent equilibria: methane sulfonic acid aqueous
solutions. J. Phys. Chem. 90, 4347–4353 (1986).
50. Raamat, E. et al. Acidities of strong neutral Brønsted acids in different media.
J. Phys. Org. Chem. 26, 162–170 (2013).
Acknowledgements
30. Fukuzumi, S. & Ohkubo, K. Selective photocatalytic reactions with organic
photocatalysts. Chem. Sci. 4, 561–574 (2013).
The authors acknowledge support from The David and Lucile Packard Foundation.
31. Benniston, A. C. et al. Charge shift and triplet state formation in the 9-mesityl-
10-methylacridinium cation. J. Am. Chem. Soc. 127, 16054–16064 (2005).
32. Bordwell, F. G., Zhang, X-M., Satish, A. V. & Cheng, J-P. Assessment of the
importance of changes in ground-state energies on the bond dissociation
enthalpies of the O–H bonds in phenols and the S–H bonds in thiophenols.
J. Am. Chem. Soc. 116, 6605–6610 (1994).
33. O’Hagan, D. Understanding organofluorine chemistry. An introduction to the
C–F bond. Chem. Soc. Rev. 37, 308–319 (2008).
34. Wang, J. et al. Fluorine in pharmaceutical industry: fluorine-containing drugs
introduced to the market in the last decade (2001–2011). Chem. Rev. 114,
2432–2506 (2014).
Author contributions
D.J.W., J-M.M.G. and T.R.L. performed the experiments. D.A.N. guided the research.
D.J.W. performed the experiments on the optimization and substrate scope for the anti-
Markovnikov hydrochlorination. D.J.W. and J-M.M.G. performed the experiments on
the optimization and substrate scope for the anti-Markovnikov hydrofluorination. D.J.W.,
J-M.M.G. and T.R.L. performed the experiments on the optimization and substrate scope
for the anti-Markovnikov phosphate additions. J-M.M.G. performed the experiments on
the optimization and substrate scope for the anti-Markovnikov sulfonate additions. D.J.W.,
J-M.M.G. and D.A.N. co-wrote the manuscript.
35. Liang, T., Neumann, C. N. & Ritter, T. Introduction of fluorine and fluorine-
containing functional groups. Angew. Chem. Int. Ed. 52, 8214–8264 (2013).
36. Olah, G. A., Masatomo, N. & Kerekes, I. Synthetic methods II1.
Hydrofluorination of alkenes, cyclopropane and alkynes with poly-hydrogen
fluoride/pyridine (trialkylamine) reagents. Synthesis 12, 779–780 (1973).
37. Bucsi, T. et al. Stable dialkyl ether/poly(hydrogen fluoride) complexes: dimethyl
ether/poly(hydrogen fluoride), a new, convenient, and effective fluorinating
agent. J. Am. Chem. Soc. 124, 7728–7736 (2002).
Additional information
Supplementary information and chemical compound information are available in the
online version of the paper. Reprints and permissions information is available online at
www.nature.com/reprints. Correspondence and requests for materials should be
addressed to D.A.N.
38. Barker, T. J & Boger, D. L. Fe(III)/NaBH4-mediated free radical hydrofluorination Competing financial interests
of unactivated alkenes. J. Am. Chem. Soc. 134, 13588–13591 (2012).
The authors declare no competing financial interests.
726
NATURE CHEMISTRY | VOL 6 | AUGUST 2014 | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.