Key words radical - diradical - diradicaloid - carbene - hydrocarbon - heterocycle - open-shell
- paramagnetic
1
Introduction
Radicals are fundamental intermediates[1 ] and they play crucial roles in numerous biological[2 ] and industrially important processes.[3 ] In general, radicals are short-lived species and, therefore, in a condensed phase,
spontaneously react with surrounding (solvent, precursor, or by-product) molecules,
undergo self-decomposition or dimerization to eventually form closed-shell products.[4 ] Consequently, synthetic access to thermally stable[5 ] radicals for detailed structural investigations has been a subject of intense fundamental
interest for augmenting our understanding about chemical bonding and electronic structure.[6 ] In addition, stable radicals are also highly promising class of compounds for their
potential applications in materials science and emerging areas.[7 ] This is because of their unique optical, electronic, and magnetic properties.[6b ]
[8 ] The potential of stable radicals as unique building blocks for the design of next-generation
advanced functional materials for applications in optoelectronics, spintronics, and
solar cells is substantial.[9 ] The incorporation of stable radicals into artificial molecular machines (AMMs) is
very appealing because an electron’s spin can be much easier and more precisely controlled,
even with a smaller energy input, than a charge.[7c ] In this context, carbon-centered radicals have special advantages because of their
planar trivalent nature, which promises structural diversity and elaborate designs.
Stable radicals are also of immense interest as active materials in organic radical
batteries (ORBs).[10 ] These so-called ‘eco batteries’ have the potential to be cheaper, safer, and longer-lasting
than current metal-based batteries and, thus, are of ever-growing importance in our
modern technology-based society.[11 ] This becomes evidently more appealing if such radicals are derived from 1,3-imidazoli(ni)um
precursors, which are promising electrolytes for battery applications. In recent years,
radical catalysis has also emerged as an important tool in organic synthesis.[12 ]
Stable Radicals and Radicaloids
In 1900, Gomberg isolated the first stable radical, the triphenylmethyl I (Scheme [1 ]).[13 ] This seminal report marked the dawn of stable radical chemistry. I , however, undergoes self-association in the condensed phase, yielding the σ-dimer
II . Although the delocalization of the unpaired electron over the phenyl rings contributes
to the stability of I , steric shielding of the central carbon atom that retains the most spin density is
considered the key factor. The use of bulky substituents at para - or meta -position(s) of the phenyl groups slows down the dimerization to some extent, but
no monomeric species could be isolated. Derivatives containing ortho -substituted or perchlorinated aryl groups were found to be remarkably stable. Because
of the synthetic modularity, triphenylmethyl derivatives remain the most well-known
class of stable carbon-centered radical hydrocarbons reported to date. Nonetheless,
only heavily substituted or chlorinated species are isolable as monomers.[14 ] The presence of two unpaired electrons in a molecule adds additional complexity
to the electronic structure. Molecules containing two unpaired electrons in two nearly
degenerate molecular orbitals are called diradicals, which may have parallel (triplet)
or antiparallel (singlet) spins.[8 ] Depending on the interaction between the unpaired electrons, singlet diradicals
are further classified as open-shell (OS) and close-shell (CS) singlets. It should
however be noted that the nature of the singlet wave function changes progressively
from an OS to a CS configuration. Thus, the region in which the triplet becomes strongly
destabilized and where the OS singlet becomes a CS singlet is not clearly defined.
The term ‘diradicaloids’ is used for molecules with partial singlet diradical nature
in their ground state.[15 ]
Scheme 1 The triphenylmethyl (trityl) radical I and its σ-dimer II
After the isolation of the trityl radical I , Thiele reported the phenylene bridged compound III (n = 1; Scheme [2 ]) in 1904.[16 ] In 1907, Tschitschibabin isolated a related derivative III (n = 2) featuring a biphenylene bridge.[17 ] In general, they can be described either as an OS diradical (IIIa ) or a CS quinoid (IIIb ) structure. This is, however, not very straightforward. The electronic structure
and properties of these so-called Thiele’s (TH ) and Chichibabin’s (CH ) hydrocarbons had been a subject of intense scientific discussion for more than 70
years, and this issue came to be known as ‘the Biradical Paradox’.[18 ] In 1986, Montgomery et al.[19 ] determined the solid-state molecular structures of these highly reactive species
by single-crystal X-ray diffraction. Currently, CH is generally described as a diradicaloid (a resonance hybrid of IIIa and IIIb ) with a significant diradical character, whereas TH is considered as a p -quinodimethane (p -QDM). This can be explained based on the recovery of one (TH ) and two (CH ) Clar’s aromatic sextet ring(s) in the OS diradical forms IIIa .[20 ] Obviously, the Müller’s hydrocarbon (MH )[21 ] with a terphenylene (I , n = 3) bridge reported in 1941 has higher diradical character compared with TH and CH . So, the MH is extremely reactive and it remains a challenging task to obtain pure samples for
reliable measurements.
Scheme 2 Diradical (IIIa ) and quinoid (IIIb ) resonance forms of Thiele’s (n = 1), Chichibabin’s (n = 2), and Müller’s HCs and selected examples of analogous dicationic bis(triarylamino)
(IV ) and benzannulated (V and VI ) derivatives
Stable diradicaloids delocalized over π-conjugated systems are appealing synthetic
targets because of their unique physical properties and potential applications in
nonlinear optics, molecular electronics, and organic spintronics.[22 ] In this context, TH , CH , MH , and related molecules serve as very promising model systems. Over the past few years,
different synthetic strategies have been developed to thermodynamically and/or kinetically
stabilize various analogues of TH , CH , and MH . Bis(triarylamine) dications IV (Scheme [2 ]) are isoelectronic to III , in which the Ph2 C• moiety of III is replaced by an Ar2 N+• unit.[23 ] The benzannulation approach has also been successfully employed to isolate stable
polycyclic hydrocarbons such as V .[9a ]
[24 ] In 2017, Wu et al. reported the oxindolyl-based carbon-centered diradicaloid VI , an analogue of the MH .[25 ] Synthetic strategies to various carbon-centered stable radicals have been summarized
in some recent review articles.[6b ]
[26 ]
N-Heterocyclic carbenes (NHCs VII ; Scheme [3 ]) are very versatile σ-donor neutral ligands and have been extensively explored in
transition-metal chemistry and catalysis.[27 ] Similarly, numerous compounds containing a low-valent main-group element,[28 ] including radicals (VIII ),[29 ] have been stabilized by the use of NHCs. In this context, the success of Bertrand’s
cyclic alkyl amino carbenes (CAACs, IX ) in stabilizing open-shell main-group or transition-metal species is remarkable.[30 ] This is because of the favorable electronic structure of CAACs. The smaller HOMO–LUMO
energy gap of CAACs compared with classical NHCs makes them not only stronger σ-donors
but also very efficient π-acceptors with respect to the latter.[31 ] The enhanced electrophilicity of CAACs originates from the low-lying π*(Ccarbene –N) orbital that allows captodative[32 ] interactions between the carbene (IX ) and a main-group moiety (E) in the corresponding derivatives (X ).[30` ]
[b ]
[c ]
[d ]
,
[33 ] Various carbon-centered stable radicals have been also isolated using CAAC-frameworks.[34 ] In contrast, stable organic radicals derived from NHCs (VII ) have so far remained limited to some specific compounds featuring a redox active
nitrogen or sulfur atom(s).[35 ]
Scheme 3 NHC (VII ) and NHC-stabilized main-group radicals (VIII ). CAAC (IX ) and CAAC-stabilized main-group radicals (X )
Our research foci of developing new catalytic structures for sustainable synthesis[36 ] and accessing open-shell main-group compounds[27b ]
[37 ] for advanced functional materials relies on the development as well as exploration
of unconventional carbon-donor ligands/scaffolds. For this purpose, we use classical
NHCs as fundamental structural motifs to design new ligand sets.[38 ] This is because of the easy synthetic accessibility and tunable steric as well as
electronic properties of NHCs. The available structural variables (Figure [1 ]) in NHCs offer a plethora of opportunities to fine-tune the properties of derived
molecules in many ways. These variables clearly appear to be more important and decisive
in addressing the stability issue of highly sought-after open-shell molecules. Therefore,
NHCs are an ideal choice for the design of stable radicals with adaptable properties.
1,3-Imidazolium salts XI (R' = H) are typical precursors to NHCs (VII ) (Scheme [4 ]). Their cyclic voltammograms (CVs) exhibit a very negative reduction potential (ca.
–2.3 V vs. Fc/Fc+ ) that is irreversible.[39 ] Not surprisingly, chemical reduction of XI (R' = H) with potassium in refluxing THF affords free NHCs (VII ). The formation of VII is believed to result from the decomposition of unstable radical XII -H.[39 ] Muon resonance spectroscopy (μRS) studies showed that VII reacts with Mu• (muonium, a hydrogen atom equivalent) to generate the radical XII -Mu, analogous to XII -H.[40 ] The XII -Mu species was calculated to feature the nonplanar radical center with 74.9% of the
spin density located at the carbene-carbon atom.
Figure 1 Available structural variables in classical NHC-based radicals
Scheme 4 Deprotonation of XI (R = aryl groups; R' = H or an aryl group) with a base to NHCs (VII ) or MICs (XII ). Reduction of XI (R' = H) to an NHC via a transient radical XII -H. Reaction of VII with Mu to XII -Mu. One-electron reduction of XI (R' = Ar) to stable radicals XIV
These early studies clearly emphasize the requirement of a suitable substituent at
the C2-position of XI to stabilize putative NHC-radicals. We previously showed that various transition-metal[36 ] as well as main-group element[38 ]
[41 ] compounds featuring a mesoionic carbene (MIC), which may also be referred to as
abnormal NHCs (aNHCs),[42 ] XII can be readily prepared by the deprotonation of C2-arylated 1,3-imidazolium salts
XII (R' = Ar) and in no case was cleavage of the C2–CAryl bond observed (Scheme [4 ]). This encouraged us to reason that stable NHC-based radicals XIV should be synthetically accessible on one-electron reduction of the appropriate C2-arylated
1,3-imidazoli(ni)um salts. This account summarizes recent advances, particularly from
this laboratory, in NHC-based carbon-centered stable radicals chemistry.
2
NHC-Monoradicals
For the synthesis of desired C2-arylated precursors 3 , two protocols based on palladium[36e ] and nickel[36b ] catalysis have been developed in this laboratory. While the scope of palladium catalysis
is limited to aryl iodides, nickel catalysis protocol is equally compatible with aryl
bromides and iodides. Remarkably, bis-1,3-imidazoli(ni)um salts are also accessible
by two-fold C–C coupling of appropriate aryl dihalides with various NHCs (see below).
The yield of reactions in most cases is more than 80%. The product precipitates out
in the employed solvent (o -xylene, THF, toluene or 1,4-dioxane) and can be readily extracted by simple filtration.
No additional purification is required and the product can be used as such. Two typical
examples with unsaturated (1a ) and saturated (2 ) NHCs are shown in Scheme [5 ].
Scheme 5 Direct C2-arylation of NHCs 1a and 2 to 3a , 3b , 4a , and 4b
In contrast to the C2-protonated 1,3-imidazolium salts XI (R' = H), the CVs of C2-arylated 1,3-imidazoli(ni)um salts 3a , 3b (E
1/2 = –1.66 V, –1.91 V) and 4a , 4b (E
1/2 = –1.42 V, –1.49 V vs. Fc+ /Fc) exhibit one-electron reduction at a very low potential (Figure [2 ]).[43 ] Note, the reduction potential for 3 and 4 is not only significantly lower compared with XI -H (E
1/2 ≥ –2.3 V) but the process is also reversible. This indicates the accessibility and
stability of the corresponding radicals.
Figure 2 Cyclic voltammograms (CVs) of 3a ,b and 4a ,b (in CH3 CN/n Bu4 NPF6 , 100 mV s–1 vs. Fc+ /Fc)
Indeed, the radicals (IPrPh )• (5a ), (IPr4-DMP )• (5b ), (SIPrPh )• (6a ), and (SIPr4-Tol )• (6b ) are accessible as crystalline solids on chemical reduction of the corresponding
cations 3a ,b and 4a ,b with KC8 (Scheme [6 ]). Compounds 5 and 6 are deeply colored stable solids.[43 ]
Scheme 6 Reduction of 3a ,b and 4a ,b to stable radicals 5a ,b and 6a ,b
Figure 3 Solid-state molecular structure of the monoradical 6b ; H atoms are omitted for clarity
Freshly prepared samples of 5 and 6 were NMR silent but each of them exhibited a doublet EPR signal (see below), indicating
their paramagnetic character. Attempts to grow single crystals of 5a and 5b for X-ray diffraction failed because they slowly decompose in solution to give the
oxidized products. The unsaturated NHC (IPr) and its 1,3-imidazolium salts (XI , 3 ) are 6π-electron aromates. The lone-pair at the carbene carbon atom of IPr is the
s-type orbital and does not take part in the π-system. The IPr-derived radicals 5 are 7π-electron systems. Thus, the loss of the aromaticity in the IPr-derived radicals
5a ,b may account for their low stability. This is however not the case with the SIPr-derivatives.
Fortunately, the radicals 6 based on a saturated NHC (SIPr) are stable, and suitable single crystals can be grown
from a saturated n -hexane solution. The solid-state molecular structures of 6a and 6b determined by X-ray diffraction are distinct from the free NHC (SIPr) and the corresponding
1,3-imidazolinium salts 4a ,b . A solid-state molecular structure of a typical radical 6b is shown in Figure [3 ].
The formation of free NHCs VII from their 1,3-imidazolium precursor XI -H is a formal two-electron reduction (Scheme [4 ]), whereas one-electron reduction of XI -H gives radicals XII -H. The value of C–N bond lengths and the N–C–N bond angle in the free NHC (2 ) is clearly larger compared with the corresponding protonated precursor (Table [1 ]).
Table 1 Selected Bond Lengths (Å) and Bond Angles (°) for the Free NHC (SIPr), Radical Precursors
(4a and 4b ), and Radicals (6a and 6b )
Compound
C1–C4
C4–C5
C1–N1b
N1–C1–N2
SIPr (2 )a
–
–
1.346(2)
105.0(1)
(SIPrPh )I (4a )
1.474(2)
1.395(2)
1.329(2)
111.8(2)
(SIPrPh )• (6a )
1.398(5)
1.430(5)
1.410(4)
107.7(3)
(SIPr4-Tol )I (4b )
1.477(4)
1.390(4)
1.333(4)
111.8(3)
(SIPr4-Tol )• (6b )
1.406(2)
1.429(2)
1.405(1)
108.5(1)
a Free NHC.
b The value is averaged.
An intermediate bond length and bond angle alteration trend can be expected for the
respective radicals. Indeed, the C1–N1/N2 bond lengths of 6a (av. 1.405(4) Å) and 6b (av. 1.404(2) Å) are considerably longer than those of 4a (av. 1.330(4) Å) and 4b (av. 1.329(4) Å) as well as of SIPr (2 ) (av. 1.346(2) Å).[44 ] The N1–C1–N2 bond angle of 6a (108.9(3)°) and 6b (108.5(9)°) is intermediate between the free carbene SIPr (105.0(1)°)[44 ] and the imidazolinium salts (2 -H)+ (114.6(4)°), 4a . (111.8(1)°) or 4b (111.0(1)°). Similarly, the C1–C4 bond length in 6a (1.398(5) Å) and 6b (1.406(2) Å) is shortened considerably in comparison with 4a (1.474(2) Å) and 4b (1.477(4) Å). Some bond lengths alteration across C
ipso
–C
ortho
bonds of the aryl (Ph or 4-Tol) substituent of 6a ,b is also observed in comparison with 4a ,b . The structural parameters are fully consistent with the radical character of 6a ,b and the partial delocalization of the unpaired electron over the aryl substituent
at the carbene carbon atom.
Figure 4 Calculated spin densities (isovalue 0.003 a.u.) and Mulliken spin density (ρspin ) numbers at the BP86+D3(BJ)/def2-TZVPP//BP86-D3(BJ)/def2-SVP level of theory for
6b ; H atoms are omitted for clarity
The localized Mülliken atomic spin densities (see Figure [4 ] for 6b ) clearly reveal that the unpaired electron in 5 and 6 is mainly located on the carbene carbon atom (37–45%). In addition, there are considerable
spin densities on the ortho - (19–23%) and para - (19–29%) carbon atoms of the aryl group, but only 8–11% on the ring nitrogen atoms.
This suggests that the radicals 5 and 6 are stabilized by delocalization of the unpaired electron on the C2-aryl substituent.
The EPR spectroscopic studies clearly support the paramagnetic nature of 5 and 6 . The room-temperature EPR spectra of 5 and 6 in THF show a doublet, which is a characteristic of monoradical species (Figure [5 ]). The doublet displays couplings with two nitrogen atoms and several hydrogen atoms
of the aryl substituent. The calculated hyperfine coupling (hfc ) constants at the PBE0/EPR-II//BP86+D3(BJ)/def2-SVP level of theory for 6a were found to be Aiso
(14 N) = 6.08, 5.71; A
iso
(1 H) = –16.12 (
ortho
-Ph), 7.27 (
meta
-Ph), –20.76 (
para
-Ph). The significant low Aiso
(14 N) value (4.10–6.09) for 5a ,b and 6a ,b is in good agreement with the computed spin density at the nitrogen atoms.
Figure 5 Experimental and simulated EPR spectra of 6a and 6b
After having isolated the monoradicals 5 and 6 , one obvious quest was to probe the feasibility of stable species featuring more
than one unpaired electron. The initial challenge was to access suitable starting
materials.
3
NHC-Diradical(oid)s
Compounds featuring two unpaired electrons are called biradicals, diradicals, or diradicaloids,
depending on the interaction between two spin dipoles. The presence of two unpaired
electrons adds additional complexity to the structure and properties of these entities.
The double carbenylation of 1,4-dihaloarenes with NHCs under nickel catalysis affords
the desired dicationic species [(NHC)(C6 H4 )
n
(NHC)](X)2 (X = Br or I) (Scheme [7 ]). Salts 7 –9 can be isolated as white or pale-yellow crystalline solids in good to excellent (>80%)
yields. Both unsaturated (1a and 1b )[45 ] as well as saturated (2 )[46 ] NHCs can be used.
Scheme 7 Two-fold carbenylation of 1,4-dihaloarenes with NHCs to bis(1,3-imidazoli(ni)um)
salts (NHC)(C6 H4 )
n
(NHC)
The CVs of compounds [(IPr)2 (C6 H4 )](Br)2 7a and [(IPr)2 (C6 H4 )2 ](Br)2 7b (Figure [6 ]) exhibit two one-electron reductions in the –1.04 to –1.51 V region.[45 ] The first reversible one-electron reduction at E
1/2 = –1.04 V (7a ) and E
1/2 = –1.37 V (7b ) vs. Fc/Fc+ corresponds to a radical cation, which undergoes a second reduction at E
1/2 = –1.29 V (7a ) and E
1/2 = –1.51 V (7b ) to give the corresponding neutral species. A similar trend can also be seen for
the derivatives 9a and 9b based on the saturated NHC, SIPr (2 ).[46 ] Nonetheless, both redox events in 9a (E
1/2 = –0.80 and –0.95 V) and 9b (E
1/2 = –1.04 and –1.28 V) occur at a lower potential compared with 7a and 7b . Similarly, CVs of compounds featuring a terphenylene spacer, 7c , 8 , and 9c also exhibit two cycles consistent with two one-electron (quasi)reversible redox processes.[47 ] Masuda and co-workers reported two-electron reduction of a closely related system
[(SIPr)C≡C(C6 H4 )C≡C(SIPr)](Br)2 and isolated a stable cumulene [(SIPr)=C=C=(C6 H4 )=C=C=(SIPr)].[48 ] Moreover, Bertrand et al. also reported[30e ]
[30f ] stable cumulenes as well as mixed-valence compounds derived from CAACs that exhibit
considerable open-shell character. Very recently, some mixed-valence systems based
on carbene–pyrylium/pyridinium hybrids have been also reported.[49 ]
Figure 6 Cyclic voltammograms of [(IPr)2 (C6 H4 )](Br)2 7a (black line) and [(IPr)2 (C6 H4 )2 ](Br)2 7b (red line) in CH3 CN/Bu4 NPF6 (vs. Fc/ Fc+) in V
As indicated by the cyclic voltammetry studies, two-electron reduction of 7 –9 with a suitable reagent should lead to the formation of the corresponding neutral
compounds. Indeed, treatment of 7 –9 with KC8 afforded expected compounds 10 –12 in good to excellent yields (Scheme [8 ]). Ether open-shell (singlet or triplet) diradical (A ) or closed-shell quinoid (B ) resonance forms can be drawn to describe the structure of compounds 10 –12 . The weight of each form depends on the interaction between the unpaired electrons,
which, for a perfect quinoidal structure, results in a two-center-two-electron covalent
bond formation at the expense of an aromatic sextet(s). Compounds 10 –12 can be regarded as NHC-analogues of TH , CH , and MH (Scheme [2 ]). Therefore, it is logical to compare and analyze the structure and properties of
10 –12 with these classical HCs.
Scheme 8 KC8 reduction of 7 , 8 , and 9 to radical hydrocarbons 10 , 11 , and 12 , respectively. Diradical (A ) and quinoidal (B ) resonance forms of 10 –12
3.1
NHC Analogues of Thiele’s Hydrocarbon (TH)
The TH (1904)[16 ] and CH (1907)[17 ] were already reported in the beginning of 20th century; it was, however, only in
1986 when Montgomery et al. successfully solved the solid-state molecular structures
of these highly reactive species by single-crystal X-ray diffraction.[19 ] This provided a direct comparison of the C–C bond length of the central phenylene
ring and consequently settled the issues associated with ‘the Biradical Paradox’.
It should however be noted that the purification of TH , CH , and related compounds is extremely difficult and the presence of even 0.01% monoradical
impurity results in a doublet EPR signal. This may mask the weak intensity triplet
expected for a triplet diradical, if it exists in a thermal equilibrium with an open-shell
singlet diradical.[50 ] Structural parameters of TH and CH along with the NHC-derivatives 10a , 10b and 12a , 12b , are given in Table [2 ]. The calculated bond lengths and bond angles for the respective closed-shell (CS)
singlet and open-shell triplet (T) are also summarized in Table [2 ]. As can be seen from these data, the bond length alteration (BLA) for the phenylene
bridge in NHC-derivatives 10a and 12a is comparable to that of TH (0.10 Å). Therefore, the formation of 10a and 12a is accompanied by the loss of benzenoid structure of 7a and 9a . Experimental structures of 10a and 12a determined by X-ray diffraction show close resemblance with those calculated for
closed-shell systems. Indeed, a pure sample of 12a is EPR silent and exhibits well-resolved 1 H and 13 C NMR signals for the NHC and (C6 H4 ) units. As described by Nakano,[51 ] the open-shell nature of a system can be expressed in terms of its diradical character
(y = 0 for the close-shell and y = 1 for the pure singlet diradical). The diradical character for compound 10a and 12a was calculated to be zero (y = 0) and thus they are closed-shell species. The estimated value for the TH (y = 0.31) is, however, considerably higher.[50 ] While the ground state for both 10a and 12a is a closed-shell singlet, the calculated singlet-triplet energy gap (ca. 27 kcal/mol)
is however rather small.
3.2
NHC Analogues of Chichibabin’s Hydrocarbon (CH)
The BLA for compounds 10b (Figure [7 ]) and 12b (0.08 Å), NHC-analogues of the CH , is smaller compared with 10a and 12a (0.10 Å), but larger with respect to that of CH (0.05 Å). The calculated diradical character for 10b (y = 0.41 at B3LYP/def2-SVP level of theory) and 12b (y = 0.43 at B3LYP/def2-TZVPP level of theory) is higher than that of 10a and 10b but it is considerably lower when compared with that of CH (y = 0.72). The singlet-triplet energy gap of 10.7 kcal calculated at B3LYP/def2-SVP
for 10b and 12b is smaller than their monophenylene counterparts 10a and 12a (ca. 27 kcal/mol).
Figure 7 Molecular structure of 10b ; H atoms are omitted for clarity
Table 2 Selected Bond Lengths (Å) and Angles (°) Determined by X-ray Diffraction for the
Precursors (7a , 7b , 9a , 9b ), NHC-Hydrocarbons (10a , 10b , 12a , 12b ), Thiele’s (TH ) and Chichibabin’s (CH ) HCs;[19 ] and Calculated (B3LYP/def2-SVP) for 10a,b and 12a,b in Close-Shell (CS ) and Triplet (T ) Ground States.
Compound
Cα –N
Cα –C
i
C
i
–C
o
C
o
–C
o′
N–C–N
C
i′
–C
i′
ΣN
BLAb
TH
Ph2 C(C6 H4 )CPh2
–
1.381
1.449
1.346
–
–
–
0.10
7a
[(IPr)(C6 H4 )(IPr)](Br)2
1.474
1.391
1.390
107.8
–
360.0
–
10a
[(IPr)(C6 H4 )(IPr)]
1.375
1.453
1.354
104.6
–
350.7
0.10
CS
[(IPr)(C6 H4 )(IPr)]a
1.393
1.458
1.364
103.9
–
358.4
0.09
T
[(IPr)(C6 H4 )(IPr)]a
1.435
1.427
1.386
105.9/104.8
–
359.7
0.04
9a
[(SIPr)(C6 H4 )(SIPr)](I)2
1.326
1.478
1.397
1.390
113.2
–
356.4
–
12a
[(SIPr)(C6 H4 )(SIPr)]
1.404
1.371
1.451
1.349
108.2
–
349.4
0.10
CS
[(SIPr)(C6 H4 )(SIPr)]a
1.413
1.390
1.461
1.362
105.8
–
355.1
0.10
T
[(SIPr)(C6 H4 )(SIPr)]a
1.407
1.445
1.421
1.391
109.0
353.6
0.03
CH
Ph2 C(C6 H4 )2 CPh2
–
1.415
1.424
1.372
–
1.448
–
0.05
7b
[(IPr)(C6 H4 )2 (IPr)](Br)2
1.464
1.394
1.388
107.1
1.492
359.2
–
10b
[(IPr)(C6 H4 )2 (IPr)]
1.388
1.441
1.359
104.6
1.413
354.0
0.08
CS
[(IPr)(C6 H4 )2 (IPr)]a
1.404
1.449
1.369
104.2
1.416
359.0
0.08
T
[(IPr)(C6 H4 )2 (IPr)]a
1.436
1.428
1.391
105.1
1.470
358.0
0.04
9b
[(SIPr)(C6 H4 )2 (SIPr)](Br)2
1.331
1.466
1.395
1.383
111.5
1.494
355.5
–
12b
[(SIPr)(C6 H4 )2 (SIPr)]
1.390
1.386
1.443
1.359
107.9
1.408
347.3
0.08
CS
[(SIPr)(C6 H4 )2 (SIPr)]a
1.403
1.398
1.450
1.367
107.0
1.415
352.9
0.08
T
[(SIPr)(C6 H4 )2 (SIPr)]a
1.409
1.424
1.433
1.387
107.6
1.481
351.4
0.04
a Calculated for closed-shell (CS ) and triplet (T ) states.
b Bond length alteration (BLA) = (C
i
–C
o
-C
o
–C
o′
).
This suggests that compounds 10b and 12b should be considered as singlet diradicaloids. Multireference CASSCF(2,2)+NEVPT2/def2-SVP
calculations for model systems 10b-M and 12b-M (where Dipp groups of 10b and 12b are replaced by methyl groups) also support their diradicaloid character. The extent
of pyramidalization of the nitrogen atom (Table [2 ]) of the 1,3-imidazol(in)e rings can also account for the quinoid character of 10a ,b and 12a ,b . As indicated by the sum of the angles at the nitrogen atom (ΣN), the nitrogen atoms
in the precursors 7a (ΣN = 360.0°) and 9a (ΣN = 359.2°) containing the unsaturated NHC (IPr) are planar. One of the nitrogen
atoms in 10a (ΣN = 350.7°) and 12a (ΣN = 354.0°) is, however, pyramidalized. A rather large pyramidalization in the
former suggests its higher quinoid character compared with the latter. A similar trend
can also be seen in compounds 10b and 12b derived from the saturated NHC (SIPr).
3.3
NHC Analogues of Müller’s Hydrocarbon (MH)
Figure 8 Solid-state molecular structure of 12c ; H atoms are omitted for clarity
The MH
[21 ] with a terphenylene (III , n = 3, Scheme [3 ]) bridge has higher diradical character compared to TH and CH . This can be rationalized according to Clar’s aromatic sextet rule[20 ] as the OS diradical form IIIa preserves three aromatic rings. Consequently, the MH is extremely reactive and it remains a challenging task to obtain pure samples for
reliable measurements. Nonetheless, stable analogues of MH have been reported by using appropriate synthetic approaches (see above). The exceptional
stability of 10a,b and 12a,b prompted us to prepare stable NHC-analogues of MH . Fortunately, the nickel catalysis protocol with a slightly higher loading (10 mol%)
gave access to the desired dications 7c , 8 , and 9c (n = 3; Scheme [7 ]) using three different NHCs. Consistent with the CVs of 7c , 8 , and 9c , their two-electron reduction with KC8 afforded compounds 10c , 11 , and 12c (n = 3; Scheme [8 ]) as stable crystalline solids after workup. Structural parameters (Table [3 ]) of 10c , 11 , and 12c (Figure [8 ]) clearly reveal the enhanced weight of the open-shell diradical form compared with
related CH -analogues (10b and 12b ) (Table [2 ]). The BLA is in the 0.05–0.07 Å range for the central and terminal phenylene rings
of the terphenylene unit. This suggests that compounds 10c , 11 , and 12c still have a considerable contribution of the quinoid resonance form, which is nonetheless
smaller compared to that of 10a,b and 12a,b .
The BLA calculated for the RB3LYP-optimized structures for singlet diradicaloids was
found to be similar to those of the solid-state data of 10c , 11 , and 12c obtained by X-ray diffraction (Table [3 ]). As expected, elongation of the spacer-size leads to enhancement of the diradical
character of 10c (0.65), 11 (0.64), and 12c (0.65). The diradical character calculated for the TH (0.28), CH (0.59), and MH (0.82) at the same level of theory (PUHF/6-311G*) indicate that the replacement of
the Ph2 C moiety with an NHC leads to the reduction of diradical character. This may be due
to the difference in the delocalization of frontier molecular orbitals concerning
the diradical character.
Table 3 Selected Bond Lengths (Å) and Angles (°) Determined for 7c , 9c , 10c , and 12c by X-ray Diffraction
C
i
–C
o
C
o
–C
m
C
p
–C
p′
C
p′
–C
o′
C
o′
–C
o′
BLAa
7c
1.392
1.390
1.482
1.396
1.390
0.00
10c
1.436
1.364
1.431
1.425
1.367
0.07 [0.06]c
9c
1.398
1.387
1.481
1.396
1.389
0.01
12c
b
1.436
(1.435)
1.364
(1.369)
1.433
(1.434)
1.425
(1.429)
1.372
(1.370)
0.07 [0.05]c
0.07 [0.06]c
a Bond length alteration (BLA) = (C
i
–C
o
-C
o
–C
m
). The values for C
i
–C
o
, C
o
–C
m
, and C
p′
–C
o′
are averaged.
b 12c contains two molecules in the asymmetric unit (values of the second molecule).
c BLA for the middle phenylene ring [C
p′
–C
o′
-C
o′
–C
o′
].
The singlet-triplet energy gap for 10c (7.24), 11 (7.60), and 12c (7.40 kcal/mol) is significantly smaller compared to the related monophenylene 10a (38.28) and biphenylene 12a (18.56 kcal/mol) derivatives at the same level of theory. While this implies the
singlet ground state for compounds 10c , 11 , and 12c , the triplet state may be populated to some extent at moderate temperature given
a relatively small ΔE
S–T . Indeed, the EPR spectra of 10c , 11 , and 12c (Figure [9a ]) each exhibit a half-field signal at 298 K for the forbidden ΔMs = 2 transition, a characteristic of triplet diradicals. The calculated spin-density
distribution maps for the singlet state of 10c , 11 , and 12c (Figure [9b ]) suggest that the unpaired electron density is mainly located at the carbene-carbon
atoms and the terminal phenylene rings.
Figure 9 (a) Solid-state X-band EPR spectrum of 12c at 298 K with inset showing the half-field signal. (b) Odd-electron density map (yellow
mesh represents the isosurface with the contour value of 0.003 a.u) of 12c
4
Conclusion and Outlook
Synthetic access to fundamentally unique stable radicals and diradicaloids based on
classical NHCs has been accomplished. The results reveal the impact of an aryl substituent
at the C2-position of NHCs on the stability of derived radicals. Most strikingly,
the desired starting materials can be readily prepared in almost quantitative yield
by the direct C2-arylation of commercially available NHCs either under nickel or palladium
catalysis. The methodology is straightforward and both the pre-catalysts Ni(COD)2 and Pd2 (dba)3 are commercially available. As the nature of aryl substituents has a direct influence
on the stability and property of the derived radicals, further efforts to expand the
scope of C2-arylation strategy would be valuable. Direct C2-arylation of air-stable
1,3-imidazoli(ni)um salts, instead of free NHCs, to the desired C2-arylated products
is more appealing as it excludes the synthesis and handling of air-sensitive carbenes.
This obviously relies on the development of new catalytic protocols, which is also
important to broaden the scope of this transformation to different NHCs with variable
steric and electronic features.
Systematic investigations are required to identify the key parameters that direct
the physical (optical, electronic, magnetic) properties of NHC-based radicals and
diradical(oid)s. Instead of simple aryl groups, the use of other π-conjugated substituents
at the C2-position of NHCs is an appropriate means to further enhance the stability
as well as to control the properties of derived radicals. In this context, furan,
thiophene, pyridine, pyrazine, cyclopentadiene, and other heterocycles containing
a heavier main-group element (P, As, Se or Te) are interesting candidates.
Diradicaloids are considered as very promising singlet-fission (SF) materials for
their application in organic solar cells.[52 ] Molecular systems with too small ΔE
S–T gaps and those with triplet ground states are not suitable for efficient SF because
they have large deviation from 2E(T1) = E(S1), leading to high energy loss and slow
SF rate.[52a ]
[53 ] Because of the open-shell singlet ground state and a small singlet-triplet gap,
NHC-based diradicaloids hold promise for the design of SF materials. A facile access
to stable NHC-diradicaloids is expected to prompt further progress in synthesis, physicochemical
measurement, and theoretical analysis on the open-shell compounds with variable diradical
characters. Innovative NHC-radical(oid)s based on polycyclic aromatic hydrocarbon
(PAH) spacers (e.g., naphthalene, anthracene, pyrene, and other acenes) are important
targets we are working on.
NHC-diradicaloids are also interesting as candidates for open-shell (non)linear optical
materials.[22 ]
[54 ] Future targets include mixed-valence (MV) compounds, high-spin (tri- and polyradicals)
molecules, and non-Kekulé diradicals based on NHC and other singlet carbene frameworks.
The incorporation of NHC-based radicals in mechanically interlocked molecules (MIMs)
has enormous potential in supramolecular chemistry and materials science and thus
is worth pursuing. Among MIMs, the synthesis of catenanes and rotaxanes containing
NHC-based pro-radical frameworks is under way in this laboratory.
