Rapid Dihydrogen Cleavage by Persistent Nitroxide Radicals under Frustrated Lewis Pair Conditions

Article abstract of DOI:10.1002/chem.201602058

Persistent radicals undergo hydrogen atom abstraction reactions with a great variety of substrates, but not with dihydrogen. It has now been found that the TEMPO radical splits dihydrogen under mild conditions in the presence of the strong bulky B(C₆F₅)₃boron Lewis acid. The reaction is thought to proceed by a typical frustrated Lewis pair mechanism with the TEMPO radical acting as the active Lewis base. The reaction was analyzed by DFT, which indicates that no significant spin density on the hydrogen atoms is accumulated along the H₂splitting reaction path.

Full text of DOI:10.1002/chem.201602058

DOI: 10.1002/chem.201602058
Communication
&
Radical Reactions
Rapid Dihydrogen Cleavage by Persistent Nitroxide Radicals
under Frustrated Lewis Pair Conditions
Xin Tao,[a] Gerald Kehr,[a] Xiaowu Wang,[a] Constantin G. Daniliuc,[a] Stefan Grimme,[b] and
Gerhard Erker*[a]
cussed with suitable examples and a rationale will be dis-
Abstract: Persistent radicals undergo hydrogen atom ab-
straction reactions with a great variety of substrates, but
cussed.
The reaction pattern of the ubiquitous TEMPO nitroxide radi-
not with dihydrogen. It has now been found that the
cal can be changed from oxygen- to nitrogen-centered reactiv-
TEMPO radical splits dihydrogen under mild conditions in
ity by Lewis acid coordination (see Scheme 2).[7] We, therefore,
the presence of the strong bulky B(C6F5)3 boron Lewis
treated the TEMPO radical (3) with the strong boron Lewis acid
B(C6F5)3 (4)[8] and isolated the respective adduct 5 as orange
acid. The reaction is thought to proceed by a typical frus-
trated Lewis pair mechanism with the TEMPO radical
crystals in around 60% yield. The compound was characterized
acting as the active Lewis base. The reaction was analyzed
by spectroscopy and by X-ray diffraction (see Figure 1). It
by DFT, which indicates that no significant spin density on
shows a bent structure at oxygen and a planar-tricoordinate
the hydrogen atoms is accumulated along the H2 splitting
geometry at nitrogen.
reaction path.
Hydrogen atom abstraction is one of the most important and
most frequently encountered free radical reactions.[1] Reactive
transient radicals as well as persistent, isolable radicals under-
go this reaction type alike.[2] Most frequently, H-atom abstrac-
C
tion from carbon–hydrogen bonds is taking place, but H re-
moval from a few other element–hydrogen bonds takes place
as well.[3] Dihydrogen, HꢀH, is not amenable to the hydrogen
atom abstraction reaction scheme, mostly because of the ther-
modynamic restrictions imposed by the remaining hydrogen
atom (H) reaction product.[4] We have now found a remarkable
exception: we have observed that persistent nitroxide radicals
1[1–3,5] can also react with H2 to give the respective diamagnetic
products 2 in high yield under mild conditions, provided the
reaction is performed under frustrated Lewis pair (FLP) condi-
tions (see Scheme 1).[6] This remarkable reaction will be dis-
Scheme 2. Reaction of the TEMPO/B(C6F5)3 system with hydrogen sources.
We reacted the adduct 5 with a slight excess of 1,4-cyclo-
hexadiene. The reaction went to completion within about 1 h
at room temperature. An in situ experiment showed that ben-
zene was formed as expected, and we found that two TEMPO
derived products were formed, namely the ammonium-type
system 7 and a rearrangement product 6.
Scheme 1. Reaction of nitroxide radicals with H2 under FLP conditions.
We then exposed the TEMPO/B(C6F5)3 adduct 5 to dihydro-
gen. In this case the reaction was even faster. It went to com-
pletion within 10 min at room temperature and it principally
gave the same products 6 and 7. The ratio of these two prod-
ucts very much depended on the actual reaction conditions.
Reaction of concentrated TEMPO/B(C6F5)3 (1:2) mixtures in ben-
zene in the presence of excess borane gave predominantly the
salt 6, whereas the reactions in a 1:1 stoichiometry in dilute
solutions in pentane gave predominantly 7. Typically, the salt 6
was isolated pure from a reaction mixture in benzene in 66%
yield and compound 7 was obtained from a pentane mixture
[a] Dr. X. Tao, Dr. G. Kehr, Dr. X. Wang, Dr. C. G. Daniliuc, Prof. G. Erker
Organisch-Chemisches Institut, Westfꢀlische Wilhelm-Universitꢀt Mꢁnster
Corrensstraße 40, 48149 Mꢁnster (Germany)
E-mail: erker@uni-muenster.de
[b] Prof. S. Grimme
Mulliken Center for Theoretical Chemistry
Institut fꢁr Physikalische und Theoretische Chemie
Universitꢀt Bonn, Beringstraße 4, 53115 Bonn (Germany)
Supporting information for this article can be found under http://
dx.doi.org/10.1002/chem.201602058.
Chem. Eur. J. 2016, 22, 1 – 5
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Communication
of 7. The reaction of the TEMPO/B(C6F5)3 adduct was also car-
ried out with D2 under analogous conditions. The [O]ꢀD and
2
[N]ꢀD groups were positively identified by their H NMR reso-
nances (see the Supporting Information for details).
Because the iminium ion product 6 appeared to form
through a rearrangement process from the original tetrame-
thylpiperidine framework, we wanted to perform the H2/
B(C6F5)3 reaction with a persistent N-oxyl radical that was inert
C
to such rearrangement. For that purpose we chose the FLPNO
radical 9, which was readily prepared by treatment of the frus-
trated Lewis pair 8[10] with nitrogen monoxide (NO) as we had
previously described.[11] We exposed the pale blue solution of
C
the FLPNO radical 9 to H2 (1.5 bar) in CD2Cl2 solution at room
temperature in the presence of B(C6F5)3. After three days, the
blue color of the radical had disappeared and we monitored
the NMR spectra of the FLPNOH product 10.[11] The boron
Lewis acid had not been consumed during this process, in
Figure 1. Molecular structure of the TEMPO/B(C6F5)3 adduct 5 (thermal ellip-
soids are shown with 15% probability level). Selected bond lengths (ꢁ) and
angles (8): B1ꢀO1 1.594(3), O1ꢀN1 1.311(2); B1-O1-N1 134.7(2), SB1CCC 334.1,
SN1CCO 355.8.
C
principle the reduction reaction of FLPNO radical 9 with dihy-
drogen is catalytic in B(C6F5)3. This was confirmed by a series of
experiments (for details see the Supporting Information) and
we also performed the analogous reaction of 9 with D2 to give
the respective FLPNOD product 10-D (see Scheme 3).
under the latter condition as a white solid analytically pure in
82%. Both these products were characterized by spectroscopy
and by X-ray diffraction. Compound 7 has a structure that is
similar to that of its precursor 5, but it features a typical am-
monium-type nitrogen coordination (SN1CCO 336.68; the struc-
ture is depicted in the Supporting Information). Compound 6
(see Figure 2) shows the rearranged iminium cation structure
[N1ꢀC5 1.292(3), SN1CCC 359.98][9] and its independent complex
[(C6F5)3B-O(H)-B(C6F5)3]ꢀ counteranion [B1ꢀO1 1.561(3) ꢁ, B2ꢀ
O1 1.571(3) ꢁ, B1-O1-B2 141.7(2)8].
C
Scheme 3. Reaction of the P/B FLPNO radical with H2 in the presence of
B(C6F5)3.
C
This result strongly indicated that the reaction of TEMPO
with H2 in the presence of B(C6F5)3 initially led to the formation
of the respective TEMPOH product (11) and that this then un-
derwent the secondary reactions to the products 6 and 7 sub-
sequently in the presence of the strong boron Lewis acid
B(C6F5)3 (4). This was confirmed by the reaction of the inde-
pendently synthesized reduction product TEMPOH with
B(C6F5)3. We carried out this reaction under the typical reaction
conditions that we had applied for the TEMPO/H2/B(C6F5)3 reac-
tion. Indeed, treatment of the TEMPOH with B(C6F5)3 in ben-
zene at room temperature predominantly gave the iminium
salt 6, whereas in pentane it furnished the alternative product
7 predominantly. Consequently, we assume a nitrenium ion re-
arrangement pathway[9] to give 6 in competition to a hydro-
gen-shift reaction to give 7, both starting from TEMPOH, in
situ generated in the TEMPO/H2/B(C6F5)3 reaction (see
Scheme 4).
Figure 2. Molecular structure of the salt 6 (thermal ellipsoids are shown with
50% probability level).
It remains to be clarified how TEMPOH is generated from
the TEMPO radical and dihydrogen in the presence of the
B(C6F5)3 boron Lewis acid. Because a direct H-atom abstraction,
that is, a typical radical reaction,[1] can probably be ruled out,
we assume a frustrated Lewis pair (FLP) pathway.[6] This is
strongly supported by the results of a DFT analysis of the
system. The calculations were conducted with our well-estab-
1
We located in the H NMR spectrum the [O]ꢀH signal of the
anion of compound 6 at d: 6.63 ppm (CD2Cl2) and the [N]ꢀH
resonance of compound 7 at d: 5.90 ppm (CD2Cl2). The [O]ꢀH
1
signal of the anion of compound 6 was located in the H NMR
spectrum (d: 6.63 ppm in CD2Cl2) as was the [N]ꢀH resonance
&
&
Chem. Eur. J. 2016, 22, 1 – 5
www.chemeurj.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communication
Scheme 4. Reaction of TEMPOH with B(C6F5)3.
lished[12a,13] standard high accuracy DFT protocol,[14a] which is
based on PBEh-3c structure optimizations[14f] with subsequent
PW6B95-D3/def2-TZVP[14c,d] single-point energy computations
(see the Supporting Information for details). The final thermo-
chemical values include free solvation energies based on the
COSMO-RS DFT-based continuum solvation model[14b] and ther-
mostatistical corrections[14e] from energy to free energy at
298 K (standard conditions are 1 atm pressure and c=1 molLꢀ1
concentration in solution). The calculations show that the
B(C6F5)3/TEMPO adduct 5 is in a slightly endergonic equilibrium
with its components. According to our calculation, the
B(C6F5)3/TEMPO pair (3/4) is able to undergo a FLP reaction
with dihydrogen. The H2 cleavage reaction has a calculated
Scheme 5. Mechanistic Scheme of the B(C6F5)3-induced 2TEMPO+H2!
¼
2TEMPOH reaction with DFT-calculated Gibbs energies [DG (298 K) in kcal
molꢀ1 in benzene solution given relative to the value of compound 5
(DG=0)] and the proposed stabilizing follow-up reactions.
In the transition state, the HꢀH molecule (with a moderately
elongated bond length of 0.816 ꢁ) is located between the
TEMPO oxygen and the boron atom. The TS structure resem-
bles those of the corresponding B(C6F5)3/H2/phosphane sys-
tems.[15] According to the DFT calculations, the H2ꢀB distance
amounts to 1.528 ꢁ and the H1-H2-B angle is 132.68, which
leads to a B···H1 separation of 2.174 ꢁ. The oxygen–H1 distance
is 1.542 ꢁ and the O-H1-H2 angle amounts to 178.188. The spin
¼
Gibbs free activation energy DG (298 K) of 17.7 kcalmolꢀ1. It
uses the TEMPO oxygen Lewis base and the B(C6F5)3 Lewis acid
in a cooperative heterolytic dihydrogen cleavage[12] to give the
zwitterionic radical product 14 (DG= +10.8 kcalmolꢀ1). The re-
action scheme is shown in Figure 3 and Scheme 5.
¼
density in the transition state 13 lies on the TEMPO nitrogen
(0.52) and oxygen (0.39), but is quite different from the spin
distribution in the free TEMPO radical (N: 0.44, O: 0.51). The
spin density on the split hydrogen molecule in the TS is practi-
cally zero, indicating that TEMPO merely acts as a typical base
component in this FLP system. The resulting zwitterionic inter-
mediate radical (14) has the unpaired electron predominantly
on nitrogen (N: 0.64, O: 0.28). It shows a very close H1···H2 con-
tact (1.337 ꢁ) and a nearly linear O-H1···H2 arrangement (162.28;
H1···H2-B: 145.18, H2ꢀB: 1.238 ꢁ, OꢀH1: 1.017 ꢁ). Because this re-
action is, so far, an endergonic process, it needs a suitable
follow-up reaction sequence to make the overall TEMPO/H2/
B(C6F5)3 process energetically feasible.
It is well known that the protonated TEMPO, that is, the
TEMPOH radical cation 14+ is readily reduced by TEMPO to
+
C
give TEMPOH (11) plus the oxoammonium cation 15.[16,17] We
assume that 15 is not stable under our applied conditions but
is readily reduced by the [HB(C6F5)3]ꢀ hydridoborate of the in-
termediate 14 to give a second molar equivalent of TEMPOH
Figure 3. DFT-calculated reaction scheme of the H2 splitting reaction by the
¼
B(C6F5)3/TEMPO pair [DG (298 K) in kcalmolꢀ1 in benzene solution relative
to compound 5 (DG=0); values in parentheses are computed enthalpies in
C
(11). The overall B(C6F5)3-induced reaction of two TEMPO
solution].
Chem. Eur. J. 2016, 22, 1 – 5
www.chemeurj.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Communication
d) G. I. Likhtenshtein, J. Yamauchi, S. Nakatsuji, A. I. Smirnov, R. Tamura,
Nitroxides Applications in Chemistry, Biomedicine, and Materials Science
Wiley-VCH, Weinheim, 2008; e) C. Galli, P. Gentili, O. Lanzalunga, Angew.
Chem. Int. Ed. 2008, 47, 4790–4796; Angew. Chem. 2008, 120, 4868–
4874.
equivalents with dihydrogen to give two TEMPOH product
molecules is overall calculated to be exergonic by about
ꢀ11 kcalmolꢀ1 [18]
.
C
The TEMPO radical (3) is inert toward dihydrogen; it does
[4] a) G. Dixon-Lewis, M. M. Sutton, A. Williams, Discuss. Faraday Soc. 1962,
33, 205–212; b) R. Knutti, R. E. Bꢂhler, Chem. Phys. 1975, 7, 229–243;
c) J. Lede, J. Villermaux, React. Kinet. Catal. Lett. 1977, 6, 225–230;
d) R. J. Morton, R. X. Kaiser, Planet. Space Sci. 2003, 51, 365–373.
[5] H.-G. Korth, K. U. Ingold, R. Sustmann, H. de Groot, H. Sies, Angew.
Chem. Int. Ed. Engl. 1992, 31, 891–893; Angew. Chem. 1992, 104, 915–
917.
[6] a) D. W. Stephan, G. Erker, Angew. Chem. Int. Ed. 2010, 49, 46–76;
Angew. Chem. 2010, 122, 50–81; b) D. W. Stephan, G. Erker, Top. Curr.
Chem. 2013, 332, 1–350; c) D. W. Stephan, G. Erker, Top. Curr. Chem.
2013, 334, 1–317; d) D. W. Stephan, G. Erker, Angew. Chem. Int. Ed.
2015, 54, 6400–6441; Angew. Chem. 2015, 127, 6498–6541.
[7] a) J. J. Scepaniak, A. M. Wright, R. A. Lewis, G. Wu, T. W. Hayton, J. Am.
Chem. Soc. 2012, 134, 19350–19353; b) T. D. Nguyen, A. M. Wright, J. S.
Page, G. Wu, T. W. Hayton, Inorg. Chem. 2014, 53, 11377–11387.
[8] a) A. G. Massey, A. J. Park, J. Organomet. Chem. 1964, 2, 245–250;
b) A. G. Massey, A. J. Park, J. Organomet. Chem. 1966, 5, 218–225.
[9] a) O. D. Gupta, R. L. Kirchmeier, J. M. Shreeve, J. Am. Chem. Soc. 1990,
112, 2383–2386; b) H. Henry-Riyad, S. Kobayashi, T. T. Tidwell, ARKIVOC
2005, vi, 266–276.
not undergo the hydrogen abstraction reaction from H2, but
C
we have found that TEMPO (3) readily reacts with dihydrogen
in the presence of the strong bulky boron Lewis acid B(C6F5)3
to give a mixture of the diamagnetic products 6 and 7. We
could show that the systems 6 and 7 are most likely follow-up
products of the initially formed TEMPOH product, which subse-
quently reacts in a pair of competing stoichiometric reactions
with B(C6F5)3. The initial TEMPO+H2 reaction is probably cata-
lyzed by B(C6F5)3, similar to the analogous FLPNO+H2 reaction.
According to the theoretical study, a typical frustrated Lewis
pair pathway is likely for the B(C6F5)3-induced reaction of
C
TEMPO radical with H2. The presence of the strong B(C6F5)3
Lewis acid has opened a frustrated Lewis pair pathway by
which in the overall reaction two molar equivalents of TEMPO
are reduced by molecular hydrogen (HꢀH) to give two
TEMPOH equivalents; however, this reaction cannot take place
directly but needs a mechanistic “detour,” such as that depict-
ed, for example, in Scheme 5. Our observation further indicates
that frustrated Lewis pair chemistry can provide powerful tools
for finding new reactions,[11,13a,19] here even extending its scope
to the chemistry of persistent free radicals.
[10] P. Spies, G. Erker, G. Kehr, K. Bergander, R. Frçhlich, S. Grimme, D. W. Ste-
phan, Chem. Commun. 2007, 5072–5074.
[11] A. P. Cardenas, B. J. Culotta, T. H. Warren, S. Grimme, A. Stute, R. Frçhlich,
G. Kehr, G. Erker, Angew. Chem. Int. Ed. 2011, 50, 7567–7571; Angew.
Chem. 2011, 123, 7709–7713.
[12] a) L. J. Hounjet, C. Bannwarth, C. N. Garon, C. B. Caputo, S. Grimme,
D. W. Stephan, Angew. Chem. Int. Ed. 2013, 52, 7492–7495; Angew.
Chem. 2013, 125, 7640–7643; b) M. Sajid, G. Kehr, C. G. Daniliuc, G.
Erker, Angew. Chem. Int. Ed. 2014, 53, 1118–1121; Angew. Chem. 2014,
126, 1136–1139; c) T. Mahdi, D. W. Stephan, J. Am. Chem. Soc. 2014,
136, 15809–15812; d) D. J. Scott, M. J. Fuchter, A. E. Ashley, J. Am.
Chem. Soc. 2014, 136, 15813–15816; e) D. J. Scott, M. J. Fuchter, A. E.
Ashley, Angew. Chem. Int. Ed. 2014, 53, 10218–10222; Angew. Chem.
2014, 126, 10382–10386.
[13] a) M. Sajid, L.-M. Elmer, C. Rosorius, C. G. Daniliuc, S. Grimme, G. Kehr, G.
Erker, Angew. Chem. Int. Ed. 2013, 52, 2243–2246; Angew. Chem. 2013,
125, 2299–2302; b) L. E. Longobardi, L. Liu, S. Grimme, D. W. Stephan, J.
Am. Chem. Soc. 2016, 138, 2500–2503.
[14] a) R. Ahlrichs, M. Bꢃr, M. Hꢃser, H. Horn, C. Kçlmel, Chem. Phys. Lett.
1989, 162, 165–169; TURBOMOLE V7.0 2015, Universitꢃt Karlsruhe und
Forschungszentrum Karlsruhe GmbH, 1989–2007; b) A. Klamt, J. Phys.
Chem. 1995, 99, 2224–2235; F. Eckert and A. Klamt, COSMOtherm, COS-
MOlogic GmbH & Co. KG, Leverkusen, Germany, 2012; c) Y. Zhao, D. G.
Truhlar, J. Phys. Chem. A 2005, 109, 5656–5667; d) F. Weigend, R. Ahl-
richs, Phys. Chem. Chem. Phys. 2005, 7, 3297–3305; e) S. Grimme, Chem.
Eur. J. 2012, 18, 9955–9964; f) S. Grimme, J. G. Brandenburg, C. Bann-
warth, A. Hansen, J. Chem. Phys. 2015, 143, 054107.
Experimental Section
Full experimental details can be found in the Supporting Informa-
tion. CCDC 1471837 (5), 1471838 (6) and 1471839 (7) contain the
supplementary crystallographic data for this paper. These data can
be obtained free of charge from The Cambridge Crystallographic
Data Centre.
Acknowledgements
Financial support from the European Research Council (G.E.)
and the DFG in the framework of the Leibniz program (S.G.) is
gratefully acknowledged. We cordially thank Professor Armido
Studer for helpful discussions. C.G.D. is responsible for X-Ray
crystal structure analysis, S.G. is responsible for computational
chemistry.
[15] S. Grimme, H. Kruse, L. Goerigk, G. Erker, Angew. Chem. Int. Ed. 2010, 49,
1402–1405; Angew. Chem. 2010, 122, 1444–1447.
[16] a) B. M. Hoffman, T. B. Eames, J. Am. Chem. Soc. 1969, 91, 2169–2170;
b) V. A. Golubev, V. D. Sen, I. V. Kulyk, A. L. Aleksandrov, B. Acad. Sci.
USSR Ch. 1975, 24, 2119–2126; c) V. D. Sen, V. A. Golubev, J. Phys. Org.
Chem. 2009, 22, 138–143.
Keywords: boron · dihydrogen splitting · frustrated Lewis
pair · persistent radical · radical reactions
[1] a) L. Tebben, A. Studer, Angew. Chem. Int. Ed. 2011, 50, 5034–5068;
Angew. Chem. 2011, 123, 5138–5174; b) A. Studer, D. P. Curran, Angew.
Chem. Int. Ed. 2016, 55, 58–102; Angew. Chem. 2016, 128, 58–106.
[2] a) P. Gray, A. A. Herod, A. Jones, Chem. Rev. 1971, 71, 247–294; b) F. Re-
cupero, C. Punta, Chem. Rev. 2007, 107, 3800–3842; c) H. Karoui, F. Le
Moigne, O. Quari, P. Tordo in Stable Radicals: Fundamentals and Applied
Aspects of Odd-Electron Compounds (Ed.: R. G. Hicks), Wiley, Hoboken,
2010, pp. 173–220.
[17] A. M. Wright, J. S. Page, J. J. Scepaniak, G. Wu, T. W. Hayton, Eur. J. Inorg.
Chem. 2013, 3817–3820.
[18] For a comparison, see: E. A. Haidasz, D. Meng, R. Amorati, A. Baschieri,
K. U. Ingold, L. Valgimigli, D. A. Pratt, J. Am. Chem. Soc. 2016, 138,
5290–5298.
[19] Y. Hasegawa, C. G. Daniliuc, G. Kehr, G. Erker, Angew. Chem. Int. Ed.
2014, 53, 12168–12171; Angew. Chem. 2014, 126, 12364–12367.
[3] a) M. Lucarini, E. Marchesi, G. F. Pedulli, J. Org. Chem. 1998, 63, 1687–
1693; b) J. L. Pitters, P. G. Piva, X. Tong, R. A. Wolkow, Nano Lett. 2003, 3,
1431–1435; c) C. Jones, R. P. Rose, New J. Chem. 2007, 31, 1484–1487;
Received: May 3, 2016
Published online on && &&, 0000
&
&
Chem. Eur. J. 2016, 22, 1 – 5
www.chemeurj.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communication
COMMUNICATION
Persistent radicals undergo hydrogen
atom abstraction reactions with a great
variety of substrates, but not with dihy-
drogen. It has now been found that the
TEMPO radical splits dihydrogen under
mild conditions in the presence of the
strong bulky B(C6F5)3 boron Lewis acid.
The reaction is thought to proceed by
a typical frustrated Lewis pair mecha-
nism with the TEMPO radical acting as
the active Lewis base.
&
Radical Reactions
X. Tao, G. Kehr, X. Wang, C. G. Daniliuc,
S. Grimme, G. Erker*
&& – &&
Rapid Dihydrogen Cleavage by
Persistent Nitroxide Radicals under
Frustrated Lewis Pair Conditions
Chem. Eur. J. 2016, 22, 1 – 5
www.chemeurj.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ