Oxidation-state-dependent photochemistry of sulfur-bridged anthracenes

Article abstract of DOI:10.1002/anie.201306236

What's up sulfur? The photochemical reactivity, including a mechanistic study of sulfur-bridged anthracenes is reported. The oxidation state of the bridging sulfur (SO_n) dictates the excited-state behavior of these molecules (see picture). Copyright

Full text of DOI:10.1002/anie.201306236

.
Angewandte
Communications
DOI: 10.1002/anie.201306236
Photochemistry
Oxidation-State-Dependent Photochemistry of Sulfur-Bridged
Anthracenes**
Peter R. Christensen, Brian O. Patrick, ꢀlise Caron, and Michael O. Wolf*
In chemical synthesis, functional group tolerance often
determines the utility of a reaction. Where one functional
group may inhibit reactivity, another may accelerate the rate
ten-fold.[1] In photochemical reactions, slight modifications in
functional groups can also have dramatic consequences, with
electronic and conformational configurations determining
how molecules behave in the presence of light.[2] In order to
better predict the outcomes and rationally design photo-
chemical syntheses, a broader understanding of functional
group behavior is needed. Organic sulfides, sulfoxides, and
sulfones are common functional groups relevant to pharma-
ceutical,[3] agrochemical,[4] and light-emitting compounds.[5]
The photochemistry of various alkyl and aryl sulfoxides has
been examined by Lee and Jenks with a focus on the
mechanism of pyramidal inversion at sulfur.[6] There are very
few reports, however, where the photochemistry of sulfur-
containing molecules is studied as a function of the oxidation
state at sulfur. We recently reported enhanced photolumi-
nescence (PL) of several soluble sulfur-bridged aromatic
molecules, whereby successive oxidation of the sulfur bridge
Scheme 1. Synthesis and photochemical products of sulfur-bridged
anthracenes (m-CPBA=meta-chloroperoxybenzoic acid, dba=dibenz-
ylideneacetone, and dppf=1,1’-bis(diphenylphosphanyl)ferrocene).
results in a systematic increase in the PL quantum yield
(FPL).[7] During the course of this study, we found that
analogous sulfur-bridged compounds, in which the aromatic
substituent is anthracene, exhibit comparatively low FPL in all
oxidation states. This unusual behavior stands in contrast to
the other aromatic groups that were examined and led us to
study the photophysics and photochemistry of sulfide, sulf-
oxide, and sulfone-bridged anthracenes in more detail.
Sulfide 1 was synthesized using palladium-catalyzed cross-
coupling[8] of 9-bromoanthracene and potassium thioacetate
(Scheme 1). Attempts to use CuI[9] or Pd0[10] catalysts to
couple 9-anthracenethiol and 9-bromoanthracene were
unsuccessful, possibly because of catalyst poisoning with the
electron-rich thiol. The corresponding sulfoxide 2 and sul-
fone 3 were obtained by oxidation of 1 with one or two
equivalents, respectively, of meta-chloroperoxybenzoic acid
(m-CPBA; details are found in the Supporting Information).
Sulfide 1 exhibits relatively weak photoluminescence and
dilute solutions are stable to irradiation at 365 nm for several
hours. Irradiation of sulfoxide 2 under the same conditions
results in rapid (seconds) loss of the bridging SO and
À
formation of a new carbon carbon bond yielding 9,9’-
bianthryl 4 in > 99% isolated yields. On the other hand,
irradiation of sulfone 3 using the same light source yields
a new anthracene dimer 5 containing a three-membered
episulfone ring. Extended (overnight) irradiation of 3 or 5 in
the presence of oxygen results in the formation of anthraqui-
none (6). Under the same extended irradiation conditions, no
anthraquinone is formed from 4. In the dark, the bridged
compound 5 reverts thermally back to 3.
Single crystals of compounds 1–3, and 5 were obtained
and the structures were determined using X-ray diffraction.[11]
Crystallographic data for 9,9’-bianthryl (4) were previously
reported.[12] The molecular structures of the series of sulfur-
bridged anthracenes 1–3 are qualitatively similar with only
minor differences observed in the solid-state geometries. The
intramolecular anthracene–anthracene centroid (An-An)
distance remains relatively constant throughout the series
(1 = 5.100 ꢀ, 2 = 4.984 ꢀ, 3 = 5.019 ꢀ), and the through space
distance between the bridge-head carbons are all about 2.9 ꢀ.
One noteworthy point is that the closest intermolecular An-
An contacts decrease as the oxidation state of the bridging
sulfur increases (1 = 5.748 ꢀ, 2 = 4.478 ꢀ, 3 = 3.941 ꢀ; see
Figure S1 in the Supporting Information). This behavior is
consistent with that of anthracene derivatives possessing
electron-withdrawing groups (i.e. perfluoroanthracenes)
[*] P. R. Christensen, Dr. B. O. Patrick, ꢀ. Caron, Prof. M. O. Wolf
Department of Chemistry, University of British Columbia
2036 Main Mall, Vancouver, BC, V6T 1Z1 (Canada)
E-mail: mwolf@chem.ubc.ca
[**] This research was supported by the Natural Sciences and Engi-
neering Research Council of Canada. We thank Dr. Saeid Kamal
(LASIR) for helpful discussions and access to instrumentation.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201306236.
12946
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 12946 –12950

Angewandte
Chemie
which are known to exhibit contracted intermolecular pack-
ing compared to unsubstituted anthracene.[13]
The PL spectra for compounds 1–3 were obtained in
nonpolar cyclohexane (CHx), dichloromethane (Ch2Cl2), and
polar acetonitrile (MeCN). In Ch2Cl2, the photoluminescence
of 2 is broad, unstructured, and centered around 455 nm
(Figure 1B). Upon conversion to bianthryl (4) the overall
shape and maximum wavelength are the same as for 2,
however, there is a significant increase in the PL intensity.
Twisted intramolecular charge-transfer (TICT) states can
occur in tethered anthracenes when orbital overlap is minimal
(i.e. orthogonal anthracene units). For example, a TICT state
is observed in 9,9’-bianthryl, where the anthracene units adopt
a minimally overlapped, antiplanar orientation relative to one
another.[18] In more polar solvents the dipolar charge-sepa-
rated excited state in 9,9’-bianthryl is stabilized, and red-
shifted, structureless emission is observed. A pronounced red-
shift in the PL spectrum of 2 is observed as the solvent
polarity is increased from CHx to MeCN indicating that
emission from this SO-bridged anthracene is likely occurring
from a TICT state (Figure 2A).
The structure of the bridged dimer 5 contains a three-
membered episulfone ring with bond angles of about 608 at
each vertex. Bridged anthracene dimers are well-known; the
[4+4] cycloaddition of anthracene, first observed in 1867 by
Fritzsche,[14] has been reviewed several times.[15] Strained
À
carbon sulfur bonds have been synthetic targets for several
decades as precursors for introducing unsaturated bonds into
organic compounds,[16] yet this is, to our knowledge, the first
photochemical preparation of an episulfone.
1H NMR spectroscopy was employed to monitor the
conversion of 2 to 4, and of 3 to the bridged species 5
(Figure S2). Notably, the singlet at d = 8.68 ppm in the
spectrum of 3, which corresponds to the aromatic proton
directly opposite to the sulfur bridge, is shifted upfield at d =
4.64 ppm in the bridged species 5, typical of anthracene
photodimers.[15b] Under the same irradiation conditions no
change was observed in the 1H NMR spectrum of 1. The
formation of 5 was also monitored with {1H}-13C HSQC to
correlate the change in the aromatic singlet to the change in
the carbon atom at this position (Figure S3).
The solid-state photoreactivity of 2 and 3 was probed by
monitoring the powder X-ray diffraction (PXRD) patterns of
the two solids before and after irradiation with the same
365 nm lamp used in the solution experiments (Figure S4).
Even with overnight irradiation no change was observed in
the PXRD patterns of the two solids. Solution 1H NMR
spectra taken before and after irradiation of the solid samples
also show no evidence for the formation of photoproducts 4
and 5 in the solid state.
The UV/Vis absorption spectra of compounds 1–3 show
the structured “fingerprint” profiles common of anthracenes
(Figure S5). While the spectra for 1–3 are all structured, they
are red-shifted, and have less well-defined vibronic structure
than other tethered anthracenes. The presence of pronounced
vibronic structure in electronic absorption profiles indicates
minimal p–p overlap between anthracenes in the ground
state.[17] Upon irradiation, the absorption spectrum of 2
exhibits loss of the band centered at 410 nm and an increase in
the vibronic structure between 315 and 385 nm as bianthryl
(4) is formed (Figure 1), indicating that significantly more p–
p orbital overlap is present in 2 than in 4. Photoconversion of
3 to 5 results in a decrease in the structured absorbance
between 325–425 nm (Figure S6) attributed to the introduc-
tion of new sp3 carbons.
Figure 2. A) PL spectra of 2 taken in CHx (g), CH2Cl2 (a), MeCN
(c), and B) solvent-dependent kinetics Àln([2]/[2]o) for the conver-
sion of 2 to 4. [2]o =1 ꢁ10À5
m
The PL spectra of compounds 1 and 3 display complex,
solvent-dependent profiles. In CHx a combination of struc-
tured PL between 390–450 nm is present along with broad,
unstructured emission at wavelengths larger than 450 nm
(Figure 3). Moreover, the broad long-wavelength emission is
most prominent in nonpolar solvents. Previous studies have
shown that anthracenes tethered together with alkyl linkers
can exhibit broad emission resulting from a combination of
excimer (Ex), and charge-transfer (CT) excited states,
Figure 1. Change in the A) electronic absorption spectra, and B) emis-
sion spectra with photoconversion of 2 [b] to 4 [c] in CH2Cl2.
Figure 3. PL spectra of compounds A) 1, and B) 3 in CHx (g),
CH2Cl2 (b), and MeCN (c).
Angew. Chem. Int. Ed. 2013, 52, 12946 –12950
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12947

.
Angewandte
Communications
whereas structured, shorter-wavelength emission results from
delocalized inter-anthracene excited states.[17] For com-
pounds 1 and 3, the observation that broad emission is most
prominent in nonpolar media indicates that this emission
results from Ex states, rather than CT states.
The kinetics of the formation of 4 from 2 were followed by
monitoring the changes in absorbance as a function of
irradiation time in different solvents (Figure 2B). The loss
of 2 was fit to a single exponential decay function yielding
pseudo first-order rates of reaction in dilute solution. It was
found that a slight increase in the rate of formation of 4 occurs
in nonpolar solvent (CHx) whereas in dichloromethane and
MeCN the rates are identical. This indicates that formation of
4 likely involves both Ex and CT states.
The rate of formation of 4 also depends significantly on
the presence of oxygen (Figure 4C). Argon-purged samples
of 2 in CH2Cl2 exhibit rates of conversion nearly five times
faster than in the presence of oxygen (Tables S2 and S3)
suggesting that the transformation of 2 to 4 involves a triplet
state.[19] In contrast, the formation of 5 shows no dependence
on the presence of oxygen (Figure 4D), suggesting that the
rate-limiting step in this reaction does not involve a triplet
state. It has previously been shown that the photodimeriza-
tion of anthracene proceeds through an excited singlet
state.[15]
The room-temperature photoluminescence lifetimes (tPL
)
of compounds 1–3 were found to be dependent on the
oxidation state of the bridging sulfur. All of the compounds
exhibit prompt (ꢀ 1 ns) tPL while compounds 2 and 3 have
multiexponential lifetimes with a substantially longer (about
30 ns) component (Table 1). The longest lived component was
The increase in emission intensity If = F/Fmax (F = area
under photoluminescence spectrum) also exhibits single
exponential kinetics, directly proportional to the loss of 2,
suggesting the formation of 4 occurs in an intramolecular
fashion (Figure 4A). To provide further evidence for intra-
Table 1: Room-temperature photoluminescence lifetimes for com-
pounds 1–4 collected in argon atmosphere.
1
2[b]
3[c]
4
t1 [ns][a]
t2 [ns]
ꢀ1
ꢀ1
ꢀ1
ꢀ1
–
27.0(1)
29.2(1)
27.6(1)
[a] In all cases t1 was shorter than the measurable minimum. [b] Errors
shown are standard deviations. [c] t3 not shown, see Tables S2, and S3
for details.
found to increase as the oxidation state of the bridging sulfur
increases, tPL: 1 < 2 < 3 both in the presence of air, and when
the solution is purged with argon. Previous studies[17] have
correlated longer-lived bi-exponential tPL in tethered anthra-
cenes to the rate of deactivation of eximer states (longer tPL
indicates longer lived excimer states). The longer tPL of
compounds 2 and 3 suggests that long-lived excited-state
species may be responsible for the observed reactivity. While
the excited state of 1 is deactivated rapidly through either
nonradiative intersystem crossing or through photolumines-
cence, 2 and 3 can undergo larger structural rearrangements
in the excited state.
The relatively short photoluminescence lifetimes
observed for all compounds (nanoseconds) are consistent
with emission from singlet states, however, the slower photo-
chemical conversion of 2 to 4 in the presence of oxygen
suggests that a nonemissive triplet state may be playing a role
in this reaction. To probe the possible involvement of a triplet
state in the formation of 4 from 2, the effect of adding a triplet
sensitizer, [Ru(bpy)3]2+ (bpy = 2,2’-bipyridine) on the rate of
the photochemical reaction was tested. Ruthenium poly-
pyridyl complexes are excellent triplet sensitizers and have
found several applications as photocatalysts in organic
chemistry.[20]
Figure 4. Conversion of A) 2 to 4, and B) 3 to 5 in CH2Cl2. The plots
show the change in concentration of 2 ([2]/[2]o), and 3 ([3]/[3]o) in the
~
^
presence of oxygen ( ) and purged with argon ( ) as well as emission
&
intensity (If) in the presence of oxygen ( ) and purged with argon (*)
as a function of the irradiation time (t) in seconds. C) Plot of Àln([2]/
[2]o), and D) Àln([3]/[3]o) for the conversion of 2 to 4, and 3 to 5 in the
&
presence of air ( ), and purged with argon (*).
molecular conversion of 2 to 4, a dimethyl-substituted
analogue of 2 (2Me2) was irradiated in the presence of 2,
and the reaction products were characterized (Scheme S1). If
the cross-product 10-methyl-9,9’-bianthryl were to form, this
would indicate that the conversion of 2 to 4 occurs through
a diffusional, intermolecular process. Analysis of the reaction
products shows only the two homo-coupled products 4, and
10,10’-dimethyl-9,9’-bianthryl (4Me2) which strongly suggests
intramolecular reactivity.
Kinetic data for the conversion of 3 to 5 was collected by
monitoring the loss in long-wavelength absorbance at 415 nm
as a function of time (Figure 4B). The rate of formation for 5
was found to exhibit single exponential kinetics slower than
that of the conversion of 2 to 4.
Ruthenium polypyridyl dyes are particularly useful for
sensitizing organic reactions involving triplet states since
intersystem crossing is very fast in these compounds resulting
in high quantum yields of long-lived (microseconds) triplet
states.[21] In addition, the metal-to-ligand charge-transfer
(MLCT) states most commonly involved in rapid triplet
12948 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 12946 –12950

Angewandte
Chemie
formation occur at low energies, typically in the visible
portion of the spectrum.[21] Following work by Islangulov and
Castellano,[22] wherein the photodimerization of anthracene
was sensitized by a ruthenium dye, the conversion of 2 to 4
was measured in the presence of [Ru(bpy)3]2+. This species is
suitable as a triplet photosensitizer for monitoring the
conversion of 2 to 4 because the MLCT band has a maximum
around 450 nm and absorption is present to about 500 nm,
allowing for selective excitation of the [Ru(bpy)3]2+ in the
presence of 2. Dilute solutions of 2 were irradiated at 475 nm
in the presence and absence of [Ru(bpy)3]Cl2 (Figure S9)
using a pulsed nanosecond laser as the light source (details in
the Supporting Information). There is a substantial enhance-
ment in the rate at which 2 is lost in the presence of the dye,
providing strong evidence for the involvement of a triplet
state in the formation of 4.
states as evidenced by PL and measured rates of conversion to
4 in various solvents. Second, the conversion of 2 to 4 likely
involves intersystem crossing to a reactive triplet state. SO2-
bridged anthracene 3 appears to possess the largest amount of
Ex character in the excited state as evidenced by its long-lived
PL, and the geometry of the resulting photochemical product
(5).
Previous studies involving the extrusion of SO from
sulfoxides have proposed both triplet radical,[23] and singlet
homolytic[24] cleavage mechanisms. While no cross-products
are formed when 2Me2 is irradiated in the presence of 2 this
does not allow for the unambiguous distinction between
À
radical and homolytic cleavage of the carbon sulfur bond.
In summary, novel photochemistry of sulfur-bridged
anthracenes is presented. The excited-state reactivity was
found to depend on the oxidation state of the bridging sulfur
atom. Sulfide (S) bridged anthracene does not react in the
excited state, and the sulfoxide (SO) and sulfone (SO2)
analogs react rapidly to form different products. Sulfoxide-
bridged anthracene was found to liberate SO and form
bianthryl in an intramolecular fashion. The formation of an
episulfone-bridged compound is observed when SO2-bridged
anthracene is subjected to photoirradiation. The photoreac-
tivity of SO and SO2-bridged anthracenes is believed to result
from longer-lived excited states, notably through the forma-
tion of reactive triplet states, as confirmed for compound 2 by
triplet sensitization. The role of excimer formation in the
excited-state is also believed to play a large role in the
formation of both 4 and 5. This chemistry, with optimization
of the synthetic conditions, is expected to be useful for the
synthesis of longer aromatic oligomers and polymers.
Different excited-state deactivation pathways can explain
the disparity in photochemical reactivity observed for 1–3
(Scheme 2). We propose that initially, direct excitation (DE)
Received: July 17, 2013
Revised: September 19, 2013
Published online: November 4, 2013
À
Keywords: C C coupling · mechanisms · photochemistry ·
photosensitization · sulfur
.
[1] Y. S. Chung, B. F. Duerr, P. Nanjappan, A. W. Czarnik, J. Org.
Chem. 1988, 53, 1336 – 1338.
[2] H. Kotani, K. Ohkubo, S. Fukuzumi, J. Am. Chem. Soc. 2004,
126, 15999 – 16006.
Scheme 2. Proposed mechanism for the observed oxidation-state-
dependent photochemistry of sulfur-bridged anthracenes.
[3] a) G. A. Piazza, A. L. K. Rahm, M. Krutzsch, G. Sperl, N. S.
Paranka, P. H. Gross, K. Brendel, R. W. Burt, D. S. Alberts, R.
Pamukcu, D. J. Ahnen, Cancer Res. 1995, 55, 3110 – 3116; b) L.
Liu, J. E. Stelmach, S. R. Natarajan, M. H. Chen, S. B. Singh,
C. D. Schwatrz, C. E. Fitzgerald, S. J. OꢁKeefe, D. M. Zaller,
D. B. Schmatz, J. B. Doherty, Bioorg. Med. Chem. Lett. 2003, 13,
3979 – 3982.
[4] C. Lamberth, J. Sulfur Chem. 2004, 25, 39 – 62.
[5] a) E. Hubijar, A. Papadimitratos, D. Lee, A. Zakhidov, J. P.
Ferraris, J. Phys. Chem. B 2013, 117, 4442 – 4448; b) Q. Zhang, J.
Li, K. Shizu, S. Huang, S. Hirata, H. Miyazaki, C. Adachi, J. Am.
Chem. Soc. 2012, 134, 14706 – 14709; c) M. M. Oliva, J. Casado,
J. T. L. Navarrete, S. Patchkovskii, T. Goodson III, M. R. Harp-
ham, J. S. S. de Melo, E. Amir, S. Rozen, J. Am. Chem. Soc. 2010,
132, 6231 – 6242.
of anthracene occurs, as evidenced by the structured UV/Vis
absorption profiles. After this DE, one or more excited states
(CT and Ex), involving both of the bridged anthracenes,
result, evidenced by the solvent dependence of the PL
spectra.
The excited-state deactivation depends heavily on the
oxidation state of the bridging sulfur. For S-bridged anthra-
cene 1, the excited state may be deactivated either through
short-lived excimer fluorescence, or through other nonradia-
tive pathways, most likely including a large degree of
intersystem crossing. For SO-bridged anthracene 2, the
excited state likely involves a combination of CT and Ex
[6] W. Lee, W. Jenks, J. Org. Chem. 2001, 66, 474 – 480.
[7] P. R. Christensen, J. Nagle, A. Bhatti, M. O. Wolf, J. Am. Chem.
Soc. 2013, 135, 8109 – 8112.
Angew. Chem. Int. Ed. 2013, 52, 12946 –12950
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12949

.
Angewandte
Communications
[8] N. Park, K. Park, M. Jang, S. Lee, J. Org. Chem. 2011, 76, 4371 –
4378.
[9] a) C. Chen, Z. Weng, J. F. Hartwig, Organometallics 2012, 31,
8031 – 8037; b) C. Uyeda, Y. Tan, G. C. Fu, J. C. Peters, J. Am.
Chem. Soc. 2013, 135, 9548 – 9552.
[10] J. F. Hartwig, Acc. Chem. Res. 2008, 41, 1534 – 1544.
[11] CCDC reference numbers 950417, 950418, 950419, 950420
contain the supplementary crystallographic data for this paper.
These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
[16] a) W. L. Mock, J. Am. Chem. Soc. 1975, 97, 3673 – 3680; b) D.
Suꢂrez, J. A. Sordo, T. L. Sordo, J. Phys. Chem. 1996, 100, 13462 –
13465; c) T. Harel, E. Amir, S. Rozen, Org. Lett. 2006, 8, 1213 –
1216.
[17] W. Rettig, B. Paeplow, H. Herbst, K. Mꢃllen, J. P. Desvergne, H.
Bouas-Laurent, New J. Chem. 1999, 23, 453 – 460.
[18] a) N. Mataga, H. Yao, T. Okada, W. Rettig, J. Phys. Chem. 1989,
93, 3383 – 3386; b) J. D. Simon, S. Su, J. Chem. Phys. 1987, 87,
7016 – 7023.
[19] N. J. Turro, Modern Molecular Photochemistry, Benjamin/Cum-
mings, Menlo Park, 1978.
[12] J. B. Kyzioł, J. Zaleski, Acta. Cryst. 2007, 63, 1235 – 1237.
[13] a) W.-Q. Deng, W. A. Goddard, J. Phys. Chem. B 2004, 108,
8614 – 8621; b) J. F. Tannaci, M. Noji, J. L. McBee, T. D. Tilley, J.
Org. Chem. 2008, 73, 7895 – 7900.
[20] H. Ikezawa, C. Kutal, K. Yasufuku, H. Yamazaki, J. Am. Chem.
Soc. 1986, 108, 1589 – 1594.
[21] J. M. McCusker, Acc. Chem. Res. 2003, 36, 876 – 887.
[22] R. R. Islangulov, F. N. Castellano, Angew. Chem. 2006, 118,
6103 – 6105; Angew. Chem. Int. Ed. 2006, 45, 5957 – 5959.
[23] a) G. W. Griffin, A. Manmade, J. Org. Chem. 1972, 37, 2589 –
2600; b) R. C. White, B. E. Arney, Jr., H. Ihmels, Beilstein J. Org.
Chem. 2012, 8, 1208 – 1212.
[14] J. Fritzsche, Pract. Chem. 1867, 101, 7363.
[15] a) H. Bouas-Laurent, A. Castellan, J.-P. Desvergne, R.
Lapouyade, Chem. Soc. Rev. 2001, 30, 248 – 263; b) H. Bouas-
Laurent, A. Castellan, J. P. Desvergne, R. Lapouyade, Chem.
Soc. Rev. 2000, 29, 43 – 55; c) H. D. Becker, Chem. Rev. 1993, 93,
145 – 172.
[24] R. S. Grainger, B. Patel, B. M. Kariuki, L. Male, N. Spencer, J.
Am. Chem. Soc. 2011, 133, 5843 – 5852.
12950 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 12946 –12950