Nitronyl nitroxide radicals as organic memory elements with both n- and p-type properties

Article abstract of DOI:10.1002/anie.201004899

Can't fight the SEEPR: Simultaneous electrochemical electron paramagnetic resonance reveals that a molecule containing the nitronyl nitroxide (NN) radical (structure and red layer) is redox-active, with switchability between oxidized and reduced states. An organic NN radical device utilizes the dual p- and n-type properties in a memory device. Copyright

Full text of DOI:10.1002/anie.201004899

Communications
DOI: 10.1002/anie.201004899
Molecular Electronics
Nitronyl Nitroxide Radicals as Organic Memory Elements with Both
n- and p-Type Properties**
Junghyun Lee, Eunkyo Lee, Sangkwan Kim, Gyeong Sook Bang, David A. Shultz,*
Robert D. Schmidt, Malcolm D. E. Forbes, and Hyoyoung Lee*
Organic molecules are being actively explored for use in
logical devices, either as individual memory elements or as
components embedded in small organic and polymeric
materials.[1] Conventional inorganic semiconductor devices
are limited in terms of performance improvement owing to
increased costs for device fabrication as well as physical
limitations on minimum feature dimensions. Organic
memory, however, is a possible substitute for both volatile
and non-volatile memory devices. It has the advantages of
facile tailoring through organic synthesis, simple device
fabrication (even upon flexible substrates), and very low
power consumption. Volatile organic memory is expected to
be applied towards dynamic random access memory
(DRAM), which typically requires a data refresh every few
milliseconds, while non-volatile organic memory can be
applied to read-only memory (ROM) and flash-type
memory. Several types of organic and polymeric materials
have been reported for this purpose, such as organic semi-
conductors,[2] charge-transfer complexes[3] (including redox-
active compounds),[4] and metal-nanoparticle-dispersed thin
films.[5] Recently, a new type of organic memory has been
added to this list, namely organic radical molecules (nitroxide
radicals, NOC) that contain an unpaired electron that is
capable of undergoing oxidation or reduction by applied bias
voltages.[6]
In 1901, Piloty and Schwerin succeeded in the synthesis
and isolation of porphyrexide, the first organic nitroxide.[7]
The most prominent member of this class of compounds is the
2,2,6,6-tetramethylpiperidine-N-oxyl radical (TEMPO).[8]
TEMPO and many other NO radicals belong to the category
of persistent radicals.[9] Since this pioneering work, the
Nakahara group has reported the synthesis of a polymeric
TEMPO radical derivative, poly(2,2,6,6-tetramethylpiperidi-
nyloxy methacrylate) (PTMA), for an organic radical bat-
tery.[10] The research group of Nishide extended this work
toward applications, such as radical batteries as cathode
active materials,[11] organic light-emitting diodes as hole-
injection layers,[12] and memory as p-type redox active
materials.[6a] The TEMPO radical is easily oxidized to yield
the corresponding oxoammonium salt, returning to the
TEMPO radical by a p-type one-electron reduction.[13]
However, for the complete circuit of the organic semi-
conducting device using PTMA, an n-type redox active
material as a partner to the p-type material is required.
Some previously reported polymer-based organic radical
memory devices required additional organic layers, such as
an electron-accepting layer for n-type[6a] and even a metal-
particle-dispersed dielectric layer for actuation of the organic
memory device.[6c] While previous research into polymer-
based organic radical memory has led to significant advances,
for a complete organic radical memory circuit it is crucial to
find new organic radical molecules that demonstrate switch-
ability and present both p- and n-type properties within the
molecule. Also, new molecules can facilitate understanding of
the origin of memory effects and whether that effect is
induced by the organic radical alone or whether other
environmental or chemical factors must be considered.
[*] Dr. J. Lee, E. Lee, S. Kim, Dr. G. S. Bang, Prof. H. Lee
NCRI, Center for Smart Molecular Memory
Department of Chemistry, Sungkyunkwan University
Suwon 440-746 (Republic of Korea)
Fax: (+82)31-299-5934
Herein we report novel molecular radical memory
behavior using a stable organic radical molecule. We have
synthesized and characterized the nitronyl nitroxide (NN)
radical molecule 2-(3’-tert-butyl-4’,5’-dimethoxymethoxybi-
phenyl-4-yl)-4,4,5,5-tetramethylimidazolidine-1-oxyl-3-oxide
(NN-Ph-CatMOM2)[14] (see also the Supporting Information).
The NN radical possesses one unpaired electron that is
E-mail: hyoyoung@skku.edu
Prof. D. A. Shultz, Dr. R. D. Schmidt[+]
Department of Chemistry, North Carolina State University
Raleigh, NC 27695-8204 (USA)
Fax: (+1)919-515-8920
E-mail: david_shultz@ncsu.edu
Prof. M. D. E. Forbes
Department of Chemistry
The University of North Carolina at Chapel Hill
Chapel Hill, NC 27599-3290 (USA)
ꢀ
delocalized across the two equivalent N O groups
(Scheme 1).[15]
Owing to delocalization, the oxidized and reduced states
of the NN radicals were expected to be stabilized over a wide
window of applied voltages, leading to a high switchability for
the NN radical memory. The ability of the NN-Ph-CatMOM2
to act as both electron donor and acceptor was investigated by
cyclic voltammetry (CV) and simultaneous electrochemical
electron paramagnetic resonance (SEEPR) spectroscopy
under an applied voltage.[16] Cyclic voltammograms were
[+] Present address: Department of Chemistry
The University of North Carolina at Chapel Hill
Chapel Hill, NC 27599-3290 (USA)
[**] This work was supported by the Creative Research Initiatives
research fund (project title: Smart Molecular Memory) of MEST/
NSF and the National Science Foundation (CHE-0943975).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201004899.
4414
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4414 –4418

suggests that NN-Ph-CatMOM2 exhibits rapid heterogeneous
electron-transfer kinetics.[18]
The SEEPR of the NN radical with zero applied potential
(neutral state) consists of five lines (g ꢁ 2.00) in the ratio of
1:2:3:2:1 arising from the hyperfine coupling of two equiv-
alent nitrogen atoms (nuclear spin I = 1) to the electron spin
(Figure 2).[19] The SEEPR revealed a near complete disap-
Scheme 1. Reversible redox mechanism of a NN-Ph-CatMOM2 radical
molecule.
obtained using an electrochemical analyzer and simultane-
ously recording the EPR spectra during the CV scan.[16] In
general, the CV of the related NN radical compounds with an
electron-donating group present a reversible redox couple at
positive voltages, while NN radical compounds with an
electron-withdrawing group showed
a reversible redox
couple at
a
negative voltage.[17] The CV of NN-Ph-
CatMOM2 is depicted in Figure 1.
Figure 2. SEEPR of 98.1 mm NN-Ph-CatMOM2 in N,N-dimethylforma-
mide containing 100 mm TBAPF6 using a gold helical electrode,
measured versus a silver wire reference electrode.
pearance of the strong EPR signal of NN-Ph-CatMOM2 upon
application of a + 1.2 V potential (oxidized state). The EPR
signal was regenerated upon returning the potential to
+ 0.1 V (neutral state; Figure 2). This reversible redox
couple was attributed to the oxidation of the NN radical to
the oxoammonium cation of NN-Ph-CatMOM2. Interestingly,
the SEEPR signal for the NN-Ph-CatMOM2 disappeared
under an applied potential of ꢀ0.4 V (reduced state), which
corresponded to the aminoxy anion, and was subsequently
regenerated at ꢀ0.05 V (Figure 2). These results demonstrate
the stabilization of the oxidized or reduced form of the NN
radical. Thus, the redox-active organic NN radical possesses
the necessary features for molecular memory controlled by an
applied voltage. There was an appreciable formation of an
impurity, iminonitroxide (IN) observed during the SEEPR
experiments, as evidenced by the simulation in Figure 2 (light
green hashed line, ꢀ0.05 V), which shows a mixture of about
58% NN and 42% IN.
A vertically structured metal–insulator–metal (MIM)
diode for the fabricated organic radical memory is shown in
Figure 3. The device is composed of a single layer of NN-Ph-
CatMOM2 molecules as the redox-active layer between two
gold electrodes. Gold was selected for both electrodes, as it is
known to have high oxidation resistance. This setup allows a
clear demonstration of whether or not the origin of the
memory effect comes from an intrinsic property of the NN
radical molecule. The organic layer was spin-coated onto the
bottom electrode from NN-Ph-CatMOM2 dissolved in aceto-
Figure 1. Cyclic voltammogram of a solution of 2.0 mm NN-Ph-
CatMOM2 in acetonitrile containing 0.1m TBAPF6 at a sweep rate of
0.1 Vsꢀ1 and using a glassy carbon electrode.
NN-Ph-CatMOM2 displayed a reversible redox wave at
0.82 V versus Ag/AgCl, which was assigned to the corre-
sponding NN/oxoammonium cation couple. This value was
consistent with those observed in other nitroxides and NN
radical compounds bearing electron-donating groups.[17] The
redox reversibility could be ascribed to resonance delocaliza-
tion of the oxoammonium cation. Importantly, NN-Ph-
CatMOM2 also exhibited a clear reversible redox wave at
ꢀ0.74 V vs Ag/AgCl, which corresponds to the reduction of
the NN radical to the corresponding aminoxy anion. This
value was similar to nitroxide radical compounds having
electron-withdrawing groups.[17] These redox waves were
observed over more than 50 cycles without decomposition.
The peak separation (EpaꢀEpc, with anodic peak potential Epa
and cathodic peak potential Epc) for the anodic redox wave
(positive region) of NN-Ph-CatMOM2 was about 70 mV at
scan rates of 100 mVsꢀ1. However, the peak separation in the
cathode redox wave (negative region) was twice as wide (ca.
140 mV), while this value was narrower than that of the
nitroxide compound reported in the literature.[17] This result
Angew. Chem. Int. Ed. 2011, 50, 4414 –4418
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4415

Communications
ꢀ1.8 V, that is negatively charged aminoxy anion. Below
ꢀ1.8 V, the device sharply switched to the OFF state.
Surprisingly, despite the fact that some of NN-Ph-CatMOM2
molecules were converted into IN-Ph-CatMOM2 when
returning back from the reduced to neutral state in the
solution (Figure 2), the hysteresis of the solid state of the
metal/NN-Ph-CatMOM2/metal junction showed high repro-
ducibility.
Stable conductivity switching behavior makes it possible
for nonvolatile molecular memory phenomena to be tested
under a voltage pulse sequence in a write–read–erase–read
(WRER) cycle. In such a cycle, the high (write) and low
(erase) conducting states were repeatedly induced and the
read states monitored in between the high and low conducting
states. A section of the voltage sequence and corresponding
current from the device is shown in Figure 5. The device can
be programmed to a high conductivity state using a + 3.0 V
Figure 3. Cross-sectional view of the device layout for the NN-Ph-
CatMOM2 organic memory device.
nitrile. The thicknesses of the top and bottom electrodes were
80 and 30 nm, respectively. The thickness of the spin-coated
radical film could be controlled from 17 to 100 nm, as
measured by ellipsometry.
As we expected, the ON/OFF ratios sharply increased
with an increase of radical film thickness. The ON/OFF ratios
of the radical film were 2.5 (17 nm), 10 (35 nm), and 1000
(70 nm). As the film thickness increased, the OFF current
sharply decreased, leading to an increased ON/OFF ratio.
However, when the thickness of the radical film was over
100 nm, the I–V curve showed an insulating behavior.
Furthermore, as the film thickness increased, threshold bias
voltage for the ON current increased (Supporting Informa-
tion, Figure S2). Figure 4 shows the I–V characteristics of the
Figure 4. Hysteretic I–V characteristics of the molecular device (Au/
NN-Ph-CatMOM2/Au). The I–V characteristics of the devices were
recorded by scanning the applied voltage from 0 to +3 V and then to
ꢀ3 V, followed by a reverse scan from ꢀ3 to 0 V.
NN-Ph-CatMOM2 with a thickness of about 70 nm. These
curves were recorded by scanning the applied voltage initially
from 0 to + 3 V and then to ꢀ3 V, followed by a reverse scan
from ꢀ3 to 0 V. The positive bias corresponds to a positive
voltage applied to the top metal pad, whereas a negative bias
corresponds to a negative voltage applied to the top pad. The
I–V curves of NN-Ph-CatMOM2 were symmetric, dipolar,
stable, and reproducible for thousands of scans. As the
positive bias voltage increased, the current starts to flow
(OFF state) and suddenly increased at a threshold voltage of
1.9 V to a conductance state that is a positively charged
oxoammonium cation (ON state). The conductive ON state
was maintained even in the reverse sweep, ranging from 1.9 to
Figure 5. a) Input and b) output of write–read–erase–read (WRER)
cycles of a molecular device containing NN-Ph-CatMOM2 for rewrit-
able data storage applications. Voltages: W +3.0, R 0.5, E ꢀ3.0, and
R 0.5 V. c) Current in the ON and OFF states as a function of the
number of WRER cycles. 1000 cycles were tested under vacuum.
4416
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4414 –4418

spectra were obtained on a Bruker Avance 400 spectrometer. NN-Ph-
CatMOM2 was prepared as described in the literature.[14] Cyclic
voltammograms were obtained using a CHI 660 A electrochemical
analyzer. An Ag(s) j AgCl(s) j KCl (satd) reference electrode and a
platinum wire auxiliary electrode were used. All electrochemical
measurements were carried out in a completely degassed solution.
The thickness of the thin film was measured using a spectroscopic
ellipsometer (Model J. A. Woollam VASE) at an incident angle of 60–
708 and in a wavelength range between 200 and 1000 nm.
pulse and to a low conductivity state using a ꢀ3.0 V pulse,
with multiple current measurements for reading at + 0.5 V.
The WRER cycles can be repeatedly performed in an excess
of 1000 cycles (Figure 5c). This device showed no significant
degradation after several hundred WRER cycles. Figure 6
SEEPR experiments were conducted on a JEOL FA100 EPR
spectrometer at the X-band (9.434 GHz) equipped with a JEOL
helical electrolytic cell (JEOL ES-EL30) comprising a 5 mm diameter
quartz tube, gold helical working electrode extending the length of
the resonant cavity, gold auxiliary electrode insulated with Teflon and
concentric to the helix, and a silver wire reference electrode. The
applied bias was controlled with a Cypress Systems, Inc. Omni-101
microprocessor controlled potentiostat. Instrument parameters were
kept constant and set to the following values: Center field 3360 G,
sweep width ꢂ 40 G, 100 kHz modulation width 1 G, amplitude 1.6 ꢀ
100, time constant 0.3 s, scan time 2 min, one scan. A 98.1 mm solution
of NN-Ph-CatMOM2 was prepared in anhydrous N,N-dimethylfor-
mamide containing 100 mm TBAPF6 (TBA = tetra-n-butylammo-
nium). The sample solution (0.8 mL) was added to the electrolytic
cell, the cell inserted into the resonant cavity, and the electrode leads
attached to the potentiostat. The NN EPR signal was observed to
decrease in intensity immediately under a positive (+ 1.2 V) applied
bias. Returning to near neutral (+ 0.1 V) bias restored the NN signal,
with the appearance of a small impurity. The signal was again
decreased in intensity immediately under negative (ꢀ0.4 V) bias, and
again returned to a near neutral (ꢀ0.05 V) bias, with significant
growth of the impurity signal. The final spectrum (ꢀ0.05 V) was
simulated using WinSim to obtain accurate hyperfine coupling
constants (aN).[20] This simulation shows the presence of two radical
species: 58% nitronylnitroxide radical (ꢀaN = 7.58 G) and 42%
iminonitroxide radical (ꢀaNa = 9.29 G and aNb = 4.20 G; see the
Supporting Information, Figure S1).
Preparation of the Au/NN-Ph-CatMOM2/Au device and charac-
terization: Ti/Au metal with a thickness of 50 ꢁ/300 ꢁ was deposited
onto a SiO2/Si substrate by electron-beam evaporation. The surface
roughness was approximately 5 ꢁ (root-mean-square), as measured
by atomic force microscopy (AFM) applied after a rapid thermal
annealing process at 5008C for 30 s. Subsequently, the bottom
electrodes were patterned with a line width of 60 mm using a stepper
and a lift-off process. The bottom electrode was cleaned with piranha
solution (98% H2SO4/30% H2O2 = 3:1 v/v) for 3 min. This was
followed by washing several times with deionized water and ethanol
and a final drying under a stream of nitrogen gas. For fabrication of
the bulk organic memory device, NN-Ph-CatMOM2 in acetonitrile
solution was spun onto an gold substrate at 200 rpm for 10 s, followed
by 1000 rpm for 30 s. The final speed was 2000 rpm for 10 s. The thin
film was dried in a vacuum at room temperature overnight. A gold
electrode (80 nm thick) was evaporated through a shadow mask onto
a molecular layer using the electron beam evaporator under 5.0 ꢀ
10ꢀ7 Torr. SEM-EDS and DSC measurements were used to probe film
formability and crystallinity (Supporting Information, Figures S3 and
S4). The sample was tested in a variable-temperature probe station
(Keithley 4200 semiconductor characterization system unit) under a
vacuum of less than 10ꢀ3 Torr.
Figure 6. Retention times of the ON and OFF states of the NN-Ph-
CatMOM2 device, probed with voltages at 0.5 V. The ON and OFF
states were induced by pulses of +3.0 and ꢀ3.0 V, respectively.
shows the typical retention time measured under inert
conditions. Once the device is switched to the ON state by
applying a positive voltage at + 3.0 V, this state was retained
after 700 s with no degradation. When the ON state was
switched back to the OFF state by a negative voltage pulse
applied at ꢀ3.0 V, the OFF state also was sustained.
In summary, cyclic voltammetry and SEEPR revealed that
the NN radical molecule was a redox-active molecule with
demonstrated switchability in both oxidized and reduced
states. The organic NN radical device bearing a crossbar
structure was easily fabricated by a wet process and showed
both p- and n-type properties that functioned as a memory
device. Based on the simple gold–molecule–gold device
structure, we conclude that the origin of the switchability
and memory phenomena of the NN radical memory device
originated from the organic NN radical molecule, NN-Ph-
CatMOM2. The WRER cycles of the robust radical memory
device showed no significant degradation after several
hundred WRER cycles, even though only one radical thin
layer was used. The transformation of NN to IN observed in
solution cannot be directly related to degradation in the solid-
state device. Such a transformation in the device would be
manifest in an attenuation of current. However, as seen in
Figure 5, the current is stable over at least 1000 cycles. This is
the first known development of voltage-driven organic radical
memory employing only one active layer of organic small
molecules having both p- and n-type properties. Although the
low ON–OFF ratio and switching cycles remained far from
satisfactory in comparison with current silicon technology, our
research on molecularly inherent radical memory provides
new insight into the design of voltage-driven functional
radical molecules.
Received: August 6, 2010
Revised: February 5, 2011
Published online: April 7, 2011
Experimental Section
Organic solvents were freshly distilled from appropriate drying
reagents prior to use. All starting materials were purchased from
Aldrich and used without further purification. 1H and 13C NMR
Keywords: electron paramagnetic resonance · memory devices ·
molecular electronics · nitrogen radicals · organic radicals
.
Angew. Chem. Int. Ed. 2011, 50, 4414 –4418
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4417

Communications
[8] O. L. Lebedev, S. N. Kazarnovskii, Tr. Khim. Khim. Tekhnol.
1959, 2, 649.
[9] D. Griller, K. U. Ingold, Acc. Chem. Res. 1976, 9, 13.
[10] a) K. Nakahara, S. Iwasa, M. Satoh, Y. Morioka, J. Iriyama, M.
Suguro, E. Hasegawa, Chem. Phys. Lett. 2002, 359, 351; b) H.
Nishide, K. Oyaizu, Science 2008, 319, 737.
[11] H. Nishide, S. Iwasa, Y. J. Pu, T. Suga, K. Nakahara, M. Satoh,
Electrochim. Acta 2004, 50, 827.
[12] T. Kurata, K. Koshika, F. Kato, J. Kido, H. Nishide, Chem.
Commun. 2007, 2986.
[13] P. Krzyczmonik, H. Scholl, J. Electroanal. Chem. 1992, 335, 233.
[14] D. A. Shultz, K. E. Vostrikova, S. H. Bodnar, H. J. Koo, M. H.
Whangbo, M. L. Kirk, E. C. Depperman, J. W. Kampf, J. Am.
Chem. Soc. 2003, 125, 1607.
[15] A. Caneschi, D. Gatteschi, R. Sessoli, Acc. Chem. Res. 1989, 22,
392.
[16] a) A. H. Maki, D. H. Geske, J. Chem. Phys. 1959, 30, 1356;
b) D. H. Geske, A. H. Maki, J. Am. Chem. Soc. 1960, 82, 2671;
c) I. B. Goldberg, A. J. Bard, J. Phys. Chem. 1971, 75, 3281;
d) R. D. Allendoerfer, G. A. Martinchek, S. Bruckenstein, Anal.
Chem. 1975, 47, 890.
[1] J. C. Scott, L. D. Bozano, Adv. Mater. 2007, 19, 1452.
[2] a) S. Mꢂller, C. Perlov, W. Jackson, C. Taussig, S. R. Forrest,
Nature 2003, 426, 166; b) D. Tondelier, K. Lmimouni, D.
Vuilaume, Appl. Phys. Lett. 2004, 85, 5763; c) L. D. Bozano,
B. W. Kean, V. R. Deline, J. R. Salem, J. C. Scott, Appl. Phys.
Lett. 2004, 84, 607.
[3] a) R. S. Potember, T. O. Poehler, D. O. Cowan, Appl. Phys. Lett.
1979, 34, 405; b) T. Oyamada, H. Tanaka, K. Matsushige, H.
Sasabe, C. Adachi, Appl. Phys. Lett. 2003, 83, 1252; c) Q. D.
Ling, Y. Song, S. J. Ding, C. X. Zhu, D. S. H. Chan, D. L. Kwong,
E. T. Kang, K. G. Neoh, Adv. Mater. 2005, 17, 455; d) Y. Song,
Q. D. Ling, C. X. Zhu, E. T. Kang, D. S. H. Chan, Y. H. Sang,
D. L. Kwong, IEEE Electron Device Lett. 2006, 27, 154; e) Q. D.
Ling, Y. Song, S. L. Lim, E. Y. H. Teo, Y. P. Tan, C. X. Zhu,
D. S. H. Chan, D. L. Kwong, E. T. Kang, K. G. Neoh, Angew.
Chem. 2006, 118, 3013; Angew. Chem. Int. Ed. 2006, 45, 2947.
[4] a) Y. Gofer, H. Sarker, J. G. Killian, T. O Poehler, P. C. Searson,
Appl. Phys. Lett. 1997, 71, 1582; b) K. Seo, A. V. Konchenko, J.
Lee, G. S. Bang, H. Lee, J. Am. Chem. Soc. 2008, 130, 2553; c) J.
Lee, H. Chang, S. Kim, G. S. Bang, H. Lee, Angew. Chem. 2009,
121, 8653; Angew. Chem. Int. Ed. 2009, 48, 8501.
[17] T. Suga, Y. J. Pu, S. Kasatori, H. Nishide, Macromolecules 2007,
40, 3167.
[18] T. Suga, Y. J. Pu, K. Oyaizu, H. Nishide, Bull. Chem. Soc. Jpn.
2004, 77, 2203.
[19] G. M. Rosen, S. Porasuphatana, P. Tsai, N. P. Ambulos, V. E.
Galtsev, K. Ichikawa, H. J. Halpern, Macromolecules 2003, 36,
1021.
[5] R. J. Tseng, J. Huang, J. Ouyang, R. B. Kaner, Y. Yang, Nano
Lett. 2005, 5, 1077.
[6] a) Y. Yonekuta, T. Kurata, H. Nishide, J. Photopolym. Sci.
Technol. 2005, 18, 39; b) Y. Yonekuta, K. Susuki, K. Oyaizu, K.
Honda, H. Nishide, J. Am. Chem. Soc. 2007, 129, 14128; c) Y.
Yonekuta, K. Honda, H. Nishide, Polym. Adv. Technol. 2008, 19,
281; d) K. Oyaizu, H. Nishide, Adv. Mater. 2009, 21, 2339.
[7] O. Piloty, B. Schwerin, Chem. Ber. 1901, 34, 2354.
[20] D. R. Duling, J. Magn. Reson. Ser. B 1994, 104, 105.
4418
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4414 –4418