Enforced Layer-by-Layer Stacking of Energetic Salts towards High-Performance Insensitive Energetic Materials

Article abstract of DOI:10.1021/jacs.5b07852

Development of modern high-performance insensitive energetic materials is significant because of the increasing demands for both military and civilian applications. Here we propose a rapid and facile strategy called the "layer hydrogen bonding pairing approach" to organize energetic molecules via layer-by-layer stacking, which grants access to tunable energetic materials with targeted properties. Using this strategy, an unusual energetic salt, hydroxylammonium 4-amino-furazan-3-yl-tetrazol-1-olate, with good detonation performances and excellent sensitivities, was designed, synthesized, and fully characterized. In addition, the expected unique layer-by-layer structure with a high crystal packing coefficient was confirmed by single-crystal X-ray crystallography. Calculations indicate that the layer-stacking structure of this material can absorb the mechanical stimuli-induced kinetic energy by converting it to layer sliding, which results in low sensitivity.

Full text of DOI:10.1021/jacs.5b07852

Communication
pubs.acs.org/JACS
Enforced Layer-by-Layer Stacking of Energetic Salts towards High-
Performance Insensitive Energetic Materials
Jiaheng Zhang,† Lauren A. Mitchell,‡ Damon A. Parrish,‡ and Jean’ne M. Shreeve*,†
†Department of Chemistry, University of Idaho, Moscow, Idaho 83844-2343, United States
‡Naval Research Laboratory, 4555 Overlook Avenue, Washington, D.C. 20375, United States
S
* Supporting Information
hydrogen-bonding-promoted assemblies have been reported.5
ABSTRACT: Development of modern high-performance
insensitive energetic materials is significant because of the
increasing demands for both military and civilian
applications. Here we propose a rapid and facile strategy
called the “layer hydrogen bonding pairing approach” to
organize energetic molecules via layer-by-layer stacking,
which grants access to tunable energetic materials with
targeted properties. Using this strategy, an unusual
energetic salt, hydroxylammonium 4-amino-furazan-3-yl-
tetrazol-1-olate, with good detonation performances and
excellent sensitivities, was designed, synthesized, and fully
characterized. In addition, the expected unique layer-by-
layer structure with a high crystal packing coefficient was
confirmed by single-crystal X-ray crystallography. Calcu-
lations indicate that the layer-stacking structure of this
material can absorb the mechanical stimuli-induced kinetic
energy by converting it to layer sliding, which results in
low sensitivity.
Basically, hydrogen bonding involves a donor and an acceptor
containing an acidic hydrogen atom and an electronegative atom
(N, O, F, etc.), respectively. In our design concept, a nearly flat
molecule or ion is required; therefore, aromatic rings with nitro,
amino, or azo functional groups are structurally preferred, while
imino and methylene linkages should be avoided in molecular
design. Based on acidic hydrogen atoms and electronegative
atoms, a target molecule/ion is divided into several H-bond
donor and acceptor units, each donor and acceptor unit including
one or more acidic hydrogen atom(s) and electronegative
atom(s), respectively. In general, for stable, highly dense
energetic materials, the constituent molecule/ion should possess
at least four units to cover four or more directions which form
dense infinite two-dimensional networks. The core of this
strategy is that HIEMs can be designed by stacking of infinite 2D
layers in which sequence targets are connected by pairing H-
bond donor−acceptor units. Therefore, an energetic compound
with suitable complementary donor−acceptor units is another
prerequisite of the approach. Using this approach, all three
representative energetic molecules are assembled as successive
infinite 2D layers, and each layer is built up through pairing units
(Figure 1). More importantly, the predicted layer combination
modes are consistent with their real crystal structures (Figures
S2−S4). As a matter of fact, both TATB and DAAzF have other
ayer-by-layer assembly of molecules is ubiquitous in the
Lnatural world. More importantly, the highly ordered
structures formed in this way usually show fascinating material
characteristics, which has inspired chemists to conduct intensive
research on such systems. In this context, leading progress has
been made in the construction of luminescent chromophores
and organic semiconductors by the assembly of building blocks.1
Meanwhile, such parallel face-to-face arrangements were also
discovered in the crystal structures of some high-performance
insensitive energetic materials (HIEMs) and found to greatly
affect their physicochemical and detonation properties.2 These
energetic materials, such as 2,4,6-triamino-1,3,5-trinitrobenzene
(TATB), 3,3′-diamino-4,4′-azofurazan (DAAzF), and 1,1-
diamino-2,2-dinitroethene (FOX-7), are confirmed to have
some key structural features, such as planar conjugated molecules
and layered crystal packing with reasonable interlayer spacing.3
However, unsolved challenges still remain in layer construction
and direction control to avoid a herringbone motif or other low-
order packing pattern. With this in mind, described herein is a
rapid and facile strategy called the “layer hydrogen bonding
pairing approach” for controlling the assembly of energetic
molecules and enforcing layer-by-layer stacking in the solid state.
In the field of solid-state engineering, efforts toward molecular
control have focused on the utilization of weak interactions
between functional groups.4 Of these interactions, hydrogen
bonding remains the most reliable; numerous examples of
Figure 1. Schematic diagram of hydrogen-bonding-driven layer
assembly and the face-to-face crystal packing for TATB, DAAzF, and
FOX-7.
Received: July 27, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b07852
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
assembly modes; however, these modes are energetically less
favorable and have less hydrogen bonding inside the pairing units
(Figures S5 and S6). After the layer assembly, further interlinking
of the 2D layers to 3D supramolecules is driven by π-stacked or
other weak interactions in these cases.
Scheme 1. Synthesis of 5-(4-Amino-furzan-3-yl)-1-
hydroxytetrazole-Based Energetic Salts
After successful analysis of existing examples using the
proposed approach, it was our intent to show that it can also
be used to modify and design novel HIEMs. In this current work,
5-(4-amino-furazan-3-yl)-1-hydroxytetrazole (1) was chosen as a
trial compound; based on its structure, this molecule contains
three H-bond acceptor units and one donor unit. The
composition of hydrogen bond donor−acceptor units in this
target molecule represents a major class of energetic materials. In
addition, the deprotonation of the N−OH moiety offers
possibilities for structural modification. The synthesis process
of 1 and detailed experimental steps are described in the
Supporting Information. Judging by our strategy, this neutral
compound is unable to evolve to a two-dimensional extension
layer by itself because of the shortage of H-bond donor units for
pairing. To verify this, a suitable crystal of 1 obtained by slow
evaporation of a water solution was characterized by using single-
crystal X-ray diffraction. Instead of obtaining a water-free
structure, 5-(4-amino-furazan-3-yl)-1-hydroxytetrazole monohy-
drate was obtained. It is seen from the crystallographic data that
compound 1 as well as one molecule of water forms in a
monoclinic system P21/n with four pairs of molecules per unit
cell (Figure S7). However, even with pairing involving the extra
two hydrogen bond donors from the water molecule, the packing
diagram clearly shows zigzag stacking rather than a layer (Figures
2 and S8).
These ionic derivatives are isolated in nearly quantitative yields
and are stable at room temperature for long periods. Character-
ization of these ionic energetic compounds was accomplished
through IR spectroscopy, multinuclear NMR spectroscopy, and
elemental analysis.
By applying the strategy to analysis of these newly prepared
energetic salts, we believe that, instead of complete pairing, some
cations provide excess bonding donor units, which might also
result in an imbalance and a low-order packing pattern. For
instance, hydrazinium salt (3) and diaminoguanidinium salt (8)
have two or more NH2 building blocks which can donate at least
four acid hydrogens to form hydrogen bonds. The single-crystal
X-ray structures of 3 and 8 confirm our assumption about their
packing pattern (Figures S9 and S10). In order to pair these
donor units, the 4-amino-furazan-3-yl-tetrazol-1-olate anion
packs into a wavelike type for 3 and a crossing type for 8. The
dihedral angle between neighboring faces is 14.8° for 3 and 69.5°
for 8 (Figure S11). After ruling out the possibility of these poly-
amino cations, our requirements became clearer, i.e., a conformer
with one H-bond acceptor unit and three donor units of its own.
Hydroxylammonium energetic salts belong to an important
category of next-generation energetic materials. Among them,
dihydroxylammonium 5,5′-bistetrazole-1,1′-diolate (TKX-50)
and dihydroxylammonium 3,3′-dinitroamino-4,4′-azoxy-
furazanate show impressive performances as promising candi-
dates to replace currently used explosives; in addition, hydroxyl-
ammonium nitrate has already been used as a green propellant by
the U.S. Army.7 Different from other cations, the oxygen of the
hydroxylammonium cation can not only increase the oxygen
balance but also provide the acceptor character of the hydrogen
bond. Meanwhile, the adjacent hydrogen and -NH3 group of the
hydroxylammonium cation can also act as three H-bond donor
units for pairing with the 4-amino-furazan-3-yl-tetrazol-1-olate
anion. In Figure 2 is displayed the perfect match of a predicted
hydrogen-bonding-driven layer assembly by the hydroxyl-
ammonium cation and 4-amino-furazan-3-yl-tetrazol-1-olate
anion. A suitable colorless crystal of 4 was obtained by slow
recrystallization from an aqueous solution at room temperature,
crystallizing in the monoclinic space group P21/c with four
molecules per unit cell. The geometry of the anion is comparable
to that observed in its neutral precursor. The amino group and
N−O− moiety lie in the plane of the furazan and tetrazole rings,
as supported by the torsion angles (∼180°). As expected, the
arrangement of each layer is the same as the predicted mode in
Figure 2. The dihedral angle between neighboring layers is
exactly 0° for the crystal of 4, which further confirms the parallel
Figure 2. Layer assembly mode of hydroxylammonium 4-amino-
furazan-3-yl-tetrazol-1-olate (4) and its crystal packing diagram.
As deduced from the proposed strategy, the balance of H-bond
donor and acceptor units plays an important role in the layer
assembly. However, most of the energetic compounds are
structurally defined by nitrogen- and oxygen-containing hetero-
cycles, N-oxide, and nitro groups, which are usually classified as
H-bond acceptors.6 Therefore, conformers to act as H-bond
donor units are highly sought. We then carried out acid−base
reactions and metathesis reactions using equimolar amounts of
bases and hydrochloride salts of several guanidines, respectively,
to prepare a series of energetic salts 2−10 (Scheme 1; see
Supporting Information for detailed description). All the chosen
cations are nitrogen-rich synthons that contain amino group(s)
which are able to provide sufficient H-bond donor character.
B
DOI: 10.1021/jacs.5b07852
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
face-to-face arrangements (Figures S12 and S13). To identify the
tightness of the assembly layers, the distances between furazan
centroids and tetrazole centroids were measured to be 3.75 and
3.35 Å, respectively, both typical geometrical parameters of
aromatic face-to-face π-interactions (<4.00 Å).8 In addition to π-
interactions, the assembly of the layers into a 3D supramolecular
network also resulted from hydrogen bonding (1.72 Å) between
the N−O− moiety from the upper anion and amino group from
the bottom cation.
Both density and oxygen balance (oxygen content) play key
roles in dictating the performances of explosives.9 For common
HIEMs, the oxygen balance is usually not very high because of
the presence of amino group(s). Although they have a low ratio
of heavy atoms (O, Cl, F, etc.), their densities usually reach very
high values which are comparable to those of poly-nitro
explosives. As a matter of fact, the increase in density is
attributed not only to the substitution of heavy atoms/groups but
also to an enhanced molecular packing efficiency in the solid
state.5a,10 This is clear from a comparison of packing coefficients
(Table 1) for TATB, DAAzF, FOX-7, RDX, 1, 3, 4, and 8.
Table 2. Detonation Properties of Compounds 1−10
Compared with RDX and TATB
b
c
d
e
g
D
ΔfH
νD
P
FS
a
f
material
(g cm3)
(kJ g−1
)
(m s−1
)
(GPa)
IS (J)
(N)
1
1.790
1.688
1.695
1.803
1.748
1.736
1.642
1.689
1.710
1.770
1.820
1.930
2.61
2.19
2.79
2.25
3.01
2.47
2.01
2.30
2.61
0.87
0.36
−0.54
8601
8436
8759
9100
8600
8363
8173
8580
8859
8303
8748
8114
30.0
26.6
29.0
33.4
28.0
26.1
23.6
26.7
29.0
25.3
34.9
31.2
37
>50
46
>360
>360
>360
>360
>360
>360
>360
>360
>360
>360
120
2
3
4
>50
>50
>50
>50
>50
45
5
6
7
8
9
10
>50
7.4
RDX
TATB
50
>360
a
b
All the newly prepared materials are anhydrous for testing. Density
was measured by using a gas pycnometer at room temperature.
c
d
e
Calculated heat of formation. Detonation velocity. Detonation
f
g
pressure. Impact sensitivity. Friction sensitivity.
Table 1. Oxygen Content, Oxygen Balance (Ω), Calculated
Packing Coefficients (PC), and Crystallographic Density for
TATB, DAAzF, FOX-7, RDX, 1, 3, 4, and 8
employing EXPLO5 v6.01 (Table 2).12 The detonation
velocities and pressures lie in the ranges 8173−9100 m s−1 and
23.6−33.4 GPa, respectively. Compound 4 exhibits the highest
detonation performances (νD = 9100 m s−1; P = 33.4 GPa),
which are significantly superior to those of TATB (νD = 8114 m
s−1; P = 31.2 GPa) and comparable to those of the highly
explosive RDX (νD = 8748 m s−1; P = 34.9 GPa). In addition, all
salts exhibit excellent thermal stabilities (Supporting Informa-
tion), which can be attributed to extensive hydrogen bonding
interactions and strong cation−anion interactions. The decom-
position temperatures for 2−10, which fall in the range of 213−
277 °C, are all higher than that of RDX (210 °C).
a
material oxygen content (%)
Ω
(%) PC (%) crystal density (g cm−3
)
TATB
DAAzF
FOX-7
RDX
37.2
16.3
43.2
43.2
25.7
15.9
23.8
12.4
−55.8
78.1
78.0
78.0
73.6
74.0
74.0
77.8
69.4
1.938
1.728
1.907
1.806
1.732
1.706
1.822
1.608
−65.3
−21.6
−21.6
−47.1
−59.7
−47.5
−69.1
b
1
3
4
c
8
a
Oxygen balance (based on carbon dioxide) for CaHbOcNd, 1600(c −
Impact and friction sensitivities (IS and FS, respectively) are
high priorities for secondary explosives. Low sensitivities of
energetic materials can reduce the risk of serious and fatal
accidents during the manufacturing process and application.13 In
this work, the sensitivity values of 1−10 were determined by
b
a − b/2)/Mw, where Mw = molecular weight. The parameters of 1
were calculated with one lattice water. The crystal data of 8 were
c
collected at 296 K.
Because of the highly ordered layer-by-layer arrangements, the
packing coefficients of TATB, DAAzF, FOX-7, and 4 are about
4% larger than the others and result in positive crystal density
changes (∼0.1 g cm−3). The densities of 1−10 measured with a
gas pycnometer at 25 °C (Table 2) are in agreement with the
crystal data and follow a similar trend. Among them, 4 shows the
uncommon distinction of being an energetic salt which has a
slightly higher measured density than its nonionic precursor (1).
It is worth mentioning that 10 possesses the second-highest
measured density among the newly prepared energetic salts.
Since the guanylurea cation also has one H-bond acceptor unit
(carbonyl group) and three donor units (three amino groups), it
is most likely that it can also assemble a layer structure similar to
4, which leads to high packing efficiency and high density.
In this work, the heats of formation for 1 and the
corresponding anion were obtained by using the isodesmic
reaction approach. For energetic salts 2−10, the solid-state heats
of formation are calculated by employing Born−Haber energy
cycles (see Supporting Information for calculation details).11
Because of the presence of many N−N and N−O bonds, all
compounds have highly endothermic enthalpies of formation in
the range from 232.8 kJ mol−1 (10) to 760.9 kJ mol−1 (5). With
measured density values and the calculated heats of formation in
hand, detonation properties of 1−10 were obtained by
using standard Bundesanstalt fur Materialforschung (BAM)
̈
techniques. All compounds show very low sensitivities toward
mechanical stimulus (Table 2). Most of these salts have IS and FS
greater than 50 J and 360 N, respectively. In the case of 4, the
good detonation performances coupled with relatively high
thermal stability and excellent mechanical sensitivities suggest
that this material may be an attractive HIEM candidate as a
practical insensitive secondary explosive.
The sensitivities of energetic materials are related to the ease of
self-sustaining hot spots which are caused by mechanical
stimulus-induced lattice deformation (shear, slip, disorder,
etc.).14 It has been confirmed experimentally that explosives
with certain geometries are able to minimize the deformation
energy hot spots and act as desensitizers.2a,15 To gain a better
understanding of the desensitizing mechanism of layer-by-layer
assembly HIEMs, a force field was established for 4 and RDX
based on their unit cell geometries, and their abilities to handle
external forces were probed (Figure 3; see Supporting
Information for calculation details). For the purpose of
convenient comparison, the value of external force is converted
from mole units (kJ mol) into volume units (MJ m3) by dividing
by the simulated cell volume. According to our simulation, in the
case of 4, the internal stress along the a axis (0−76.0 MJ m3) and
the b axis (0−45.5 MJ m3) is much lower than that along the c axis
C
DOI: 10.1021/jacs.5b07852
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
ACKNOWLEDGMENTS
■
Financial support from the Office of Naval Research (NOOO14-
12-1-0536), the Defense Threat Reduction Agency (HDTRA 1-
11-1-0034), and CFD Research Corporation is gratefully
acknowledged.
REFERENCES
■
(1) Das, A.; Ghosh, S. Angew. Chem., Int. Ed. 2014, 53, 2038−2054.
(2) (a) Zhang, C.; Wang, X.; Huang, H. J. Am. Chem. Soc. 2008, 130,
8359−8365. (b) Zhang, C.; Xue, X.; Cao, Y.; Zhou, Y.; Li, H.; Zhou, J.;
Gao, T. CrystEngComm 2013, 15, 6837−6844. (c) Zhang, J.; Zhang, Q.;
Vo, T. T.; Parrish, D. A.; Shreeve, J. M. J. Am. Chem. Soc. 2015, 137,
1697−1704.
Figure 3. (a) Model for external forces acting on hydroxylammonium 4-
amino-furazan-3-yl-tetrazol-1-olate (4) crystal unit cell. Any mechanical
stimulus can be analyzed into three forces along three axes. (b) Internal
stress curve along the a and b axes for 4 and RDX. (c) Plot showing the
sliding constraint of the RDX crystal unit cell along the b axis.
(3) Ma, Y.; Zhang, A.; Zhang, C.; Jiang, D.; Zhu, Y.; Zhang, C. Cryst.
Growth Des. 2014, 14, 4703−4713.
(4) (a) Wood, P. A.; Feeder, N.; Furlow, M.; Galek, P. T. A.; Groom, C.
R.; Pidcock, E. CrystEngComm 2014, 16, 5839−5848. (b) Zhang, J.;
Parrish, D. A.; Shreeve, J. M. Chem. Commun. 2015, 51, 7337−7340.
(5) (a) Landenberger, K. B.; Bolton, O.; Matzger, A. J. J. Am. Chem. Soc.
2015, 137, 5074−5079. (b) Landenberger, K. B.; Bolton, O.; Matzger,
A. J. Angew. Chem., Int. Ed. 2013, 52, 6468−6471. (c) Bolton, O.;
Matzger, A. J. Angew. Chem., Int. Ed. 2011, 50, 8960−8963.
(0−219.8 MJ m3). It may be concluded that, in contrast to
compression deformation, the layer-by-layer HIEMs can readily
absorb mechanical stimuli by converting kinetic energy into layer
sliding to prevent the formation of hot spots. This also should be
the rationale for why high-performance insensitive energetic
materials can be used as desensitizers versus mechanical stimuli.
For comparison, the internal stress along the a axis and along the
b axis for RDX is calculated to be 0−365.0 and 0−115.0 MJ m3,
respectively, indicating that the slide is highly restricted because
of its packing mode.
In summary, we have developed a rapid and facile approach to
design high-performance insensitive energetic materials, which is
based on the hydrogen-bonding-promoted layer assembly of
energetic molecules and therefore enforces layer-by-layer
stacking in the solid state. By applying this strategy, a novel
energetic material, hydroxylammonium 4-amino-furazan-3-yl-
tetrazol-1-olate (4) was designed, synthesized, and fully
characterized, which shows good detonation performances (νD
= 9100 m s−1; P = 33.4 GPa) and excellent sensitivities (IS > 50 J;
FS > 360 N). The single-crystal X-ray structure of 4 confirmed
expected layer-by-layer stacking, and the layer combination is
exactly the same as in the design mode. These positive results
suggest that this qualitative approach will enhance the future
prospects for not only HIEMs design but also other solid-state
materials. Computational analysis of internal stress indicates that
the layer-by-layer geometries of HIEMs can readily absorb
mechanical stimuli by convert kinetic energy into layer sliding
and result in lower sensitivities.
(6) (a) Klapotke, T. M.; Schmid, P. C.; Schnell, S.; Stierstorfer, J. Chem.
̈
- Eur. J. 2015, 21, 9219−9228. (b) Dippold, A. A.; Klapotke, T. M. J. Am.
̈
Chem. Soc. 2013, 135, 9931−9938. (c) Zhang, Q.; Zhang, J.; Qi, X.;
Shreeve, J. M. J. Phys. Chem. A 2014, 118, 10857−10865.
(7) (a) Zhang, J.; Shreeve, J. M. J. Am. Chem. Soc. 2014, 136, 4437−
4445. (b) Fischer, N.; Fischer, D.; Klapotke, T. M.; Piercey, D. G.;
̈
Stierstorfer, J. J. Mater. Chem. 2012, 22, 20418−20422. (c) Witze, A.
Nature 2013, 500, 509−510.
(8) Molca
4211−4217.
̌ ́ ́ ́
nov, K.; Sabljic, I.; Kojic-Prodic, B. CrystEngComm 2011, 13,
(9) (a) Fischer, D.; Klapotke, T. M.; Stierstorfer, J. Angew. Chem., Int.
̈
Ed. 2015, DOI: 10.1002/anie.201502919. (b) Wu, Q.; Zhu, W.; Xiao, H.
J. Mater. Chem. A 2014, 2, 13006−13015. (c) Zhang, J.; Shreeve, J. M. J.
Phys. Chem. C 2015, 119, 12887−12895. (d) Shen, C.; Wang, P.; Lu, M.
J. Phys. Chem. A 2015, 119, 8250−8255.
(10) Aakeroy, C. B.; Wijethunga, T. K.; Desper, J. Chem. - Eur. J. 2015,
̈
31, 11029−11037.
(11) Jenkins, H. D. B.; Tudela, D.; Glasser, L. Inorg. Chem. 2002, 41,
2364−2367.
(12) Suce
2013.
́
ska, M. EXPLO5, v6.01; Brodarski Institute: Zagreb, Croatia,
(13) (a) Politzer, P.; Murray, J. S. Adv. Quantum Chem. 2014, 69, 1−30.
(b) Politzer, P.; Murray, J. S. Cryst. Growth Des. 2015, 15, 3767−3774.
(c) Badgujar, D. M.; Talawar, M. B.; Asthana, S. N.; Mahulikar, P. P. J.
Hazard. Mater. 2008, 151, 289−305.
(14) Politzer, P.; Murray, J. S. J. Mol. Model. 2015, 21, 25−36.
(15) (a) An, Q.; Liu, Y.; Zybin, S. V.; Kim, H.; Goddard, W. A. J. Phys.
Chem. C 2012, 116, 10198−10206. (b) Sharia, O.; Tsyshevsky, R.;
Kuklja, M. M. J. Phys. Chem. Lett. 2013, 4, 730−734. (c) Aluker, E. D.;
Krechetov, A. G.; Mitrofanov, A. Y.; Zverev, A. S.; Kuklja, M. M. J. Phys.
Chem. C 2012, 116, 24482−24486.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b07852.
■
S
Synthesis, characterization data, calculation details, and
crystal refinements (PDF)
X-ray crystallographic files in CIF format for 1 CIF)
X-ray crystallographic files in CIF format for 3 (CIF)
X-ray crystallographic files in CIF format for 4 (CIF)
X-ray crystallographic files in CIF format for 8 (CIF)
AUTHOR INFORMATION
Corresponding Author
*jshreeve@uidaho.edu
■
Notes
The authors declare no competing financial interest.
D
DOI: 10.1021/jacs.5b07852
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX