A general and versatile approach to thermally generated N-heterocyclic carbenes

Article abstract of DOI:10.1002/chem.200400196

The synthesis of N-heterocyclic carbene (NHC) adducts by condensation of diamines with appropriately substituted benzaldehydes is described. This simplified approach provides the NHC adduct without first having to generate the carbene followed by its protection. These adducts undergo thermal deprotection to generate N-heterocyclic carbene in situ. Adduct decomposition temperatures were investigated as a function of catalyst structure by using thermal analysis and spectroscopic techniques. Importantly, unlike adducts derived from chloroform, the new pentafluorobenzene-based adducts are more readily prepared and are stable at room temperature. The utility of these adducts as organic catalyst precursors for living ring-opening polymerization (ROP) of lactide, transesterification reactions, and the synthesis of N-heterocyclic carbene ligated organometallic complexes is also described.

Full text of DOI:10.1002/chem.200400196

FULL PAPER
A General and Versatile Approach to Thermally Generated
N-Heterocyclic Carbenes
Gregory W. Nyce,[a] Szilard Csihony,[b] Robert M. Waymouth,*[b] and James L. Hedrick*[a]
Abstract: The synthesis of N-heterocy-
clic carbene (NHC) adducts by conden-
sation of diamines with appropriately
substituted benzaldehydes is described.
This simplified approach provides the
NHC adduct without first having to
generate the carbene followed by its
protection. These adducts undergo
thermal deprotection to generate N-
heterocyclic carbene in situ. Adduct
decomposition temperatures were in-
vestigated as a function of catalyst
structure by using thermal analysis and
spectroscopic techniques. Importantly,
unlike adducts derived from chloro-
form, the new pentafluorobenzene-
based adducts are more readily pre-
pared and are stable at room tempera-
ture. The utility of these adducts as or-
ganic catalyst precursors for living ring-
opening polymerization (ROP) of lac-
tide, transesterification reactions, and
the synthesis of N-heterocyclic carbene
ligated organometallic complexes is
also described.
Keywords: carbenes
¥
nitrogen
heterocycles ¥ organic catalysis ¥
organometallic complexes
Introduction
Arduengo and others observed that the saturated (1,3-
bis(2,4,6-trimethylphenyl)imidazolin-2-ylidene) N-heterocy-
ꢀ
Although carbenes have historically played an important
function in organic chemistry as transient intermediates,[1]
significant advances in the isolation of heteroatom-substitut-
ed singlet carbenes and persistent triplet diarylcarbenes
have collectively renewed interest and the scope of possibili-
ties.[2] Wanzlick×s[3] pioneering studies of the chemistry of
bis-1,3-diphenyl imidazolin-2-ylidene carbenes[4] laid the
clic carbene (SIMes) cleanly undergoes a C H insertion re-
ꢀ
action with compounds containing acidic C H bonds to
form stable NHC adducts, whereas the corresponding unsa-
turated carbenes led to more complicated mixtures of prod-
ucts.[10] Arduengo implicated that this adduct yielded the
free carbene upon melting. Lappert[11] and Grubbs[12] have
utilized the chloroform adducts to generate transition-metal
carbenes and implicated that free carbenes were generated
in these reactions. Enders carried out analogous investiga-
tions on the methanol adduct of the 1,3,4-triphenyl-4,5-dihy-
dro-1H-1,2,4,-triazol-5-ylidene; thermolysis of this methanol
adduct cleanly liberated the free carbene and methanol.[13]
The elimination of alcohols from methanol or tert-butanol
[5]
groundworkfor Arduengo×s elegant studies on the synthe-
sis, isolation, and characterization of the first stable imidazo-
lin-2-ylidene and imidazol-2-ylidene carbenes.[6] The chemis-
try of N-heterocyclic carbenes (NHC) has become a major
area of research[7] as these stable carbenes have proven to
be outstanding ligands for transition metals[8] as well as
potent nucleophilic organic catalysts.[9]
adducts of carbenes has proven a useful strategy for generat-
14]
ing transition-metal carbene complexes,[12
but the role of
free carbenes in these processes has never been clearly dem-
onstrated.
[a] Dr. G. W. Nyce, Dr. J. L. Hedrick
Center for Polymeric Interfaces and Macromolecular Assemblies
IBM Almaden Research Center
We have recently shown that NHC are potent organic cat-
alysts for ring-opening polymerization of cyclic esters and
transesterification reactions.[9g,h] As part of our interests in
developing convenient methods for generating catalysts in
situ,[9h,j] our attention was drawn towards NHC adducts that
could be thermally activated[10,11] to form NHC for control-
led organocatalytic polymerization. For this strategy to be
successful, it was important to demonstrate that carbenes
can be liberated from these adducts efficiently in the ab-
sence of transition metals. In this report, we describe a gen-
eral strategy for generating a variety of carbene adducts
650 Harry Rd. San Jose, CA 95120 (USA)
Fax (+1)408-927-1632
E-mail: Hedrick@almaden.ibm.com
[b] Dr. S. Csihony, Prof. R. M. Waymouth
Stanford University
Stauffer 1, Room 205, Stanford CA 94305 (USA)
Fax (+1)650-736-2262
E-mail: waymouth@stanford.edu
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2004, 10, 4073 4079
DOI: 10.1002/chem.200400196
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4073

FULL PAPER
amount p-toluenesulfonic acid and Na2SO4 (50 mg) were then added to
the vial. The reaction mixture was stirred for 8 h; it was then filtered and
the solvent evaporated under reduced pressure to yield 6 as a light
brown powder (395 mg, 96%). 1H NMR (400 MHz, CDCl3, 258C): d=
3.7 3.9 (m, 2H), 3.9 4.1 (m, 2H), 6.5 (s, 1H), 6.7 6.8 (m, 2H), 6.8 6.9
(m, 1H), 7.2 7.5 ppm (m, 2H); 19F NMR (376 MHz): d=ꢀ143.2 (brs,
2F), ꢀ153.7 (m, 1F), ꢀ161.7 ppm (m, 2F); 13C NMR (100 MHz): d=
46.9, 68.1, 112.9, 118.7, 129.9, 144.4 ppm; m.p. 1408C; HRMS analysis
(ESI): m/z calcd [M+H]+: 391.1234; found: 391.1228.
from the corresponding diamines and aldehydes
(Scheme 1),[3] and show that thermolysis of these adducts
provide a convenient source of carbenes in a tunable
manner depending on the nature of the adduct.
1,3-Bis(2,4,6-trimethylphenyl)-2-(2,3,5,6-tetrafluorophenyl)imidazolidine
(7): By Using the synthetic procedure described for 4, the reaction of
2,3,5,6-tetrafluorobenzaldehyde (246 mg, 1.38 mmol) and N,N’-bis-(2,4,6-
trimethylphenylamino)ethane (408 mg, 1.38 mmol) afforded 7 as a white
crystalline solid (520 mg, 82%). 1H NMR (400 MHz, CDCl3, 258C): d=
2.2 (s, 6H), d 2.3 2.6 (br, 12H), 3.5 3.6 (m, 2H), 3.9 3.4 (m, 2H), 6.4 (s,
1H), 6.8 ppm (br, 4H); 19F NMR (376 MHz): d=ꢀ137.1 (m, 1F), ꢀ140.5
(m, 1F), ꢀ140.8 (m, 1F), ꢀ149.3 ppm (m, 1F); 13C NMR (100 MHz): d=
19.2 (br), 21.0, 51.4, 71.8, 104.1, 105.1 (t, J=22.4 Hz), 124.1 (m), 130.2
(br), 135.7, 139.5, 144.1 (m), 145.1 (m), 146.5 (m), 147.5 ppm (m); HRMS
analysis (ESI): m/z calcd [M+H]+: 457.2267; found: 457.2256.
1,3-Bisphenyl-2-(2,3,5,6-tetrafluorophenyl)imidazolidine: By using the
synthetic procedure described for 6, the reaction of N,N’-diphenylethyl-
enediamine (105 mg, 0.5 mmol) and 2,3,5,6-trifluorobenzaldehyde (89 mg,
0.5 mmol) afforded the product as an off-white powder (173 mg, 93%).
1H NMR (400 MHz, CDCl3, 258C): d=3.7 (m, 2H), 3.9 (m, 2H), 6.5 (s,
1H) 6.6 (d, 4H), 6.8 (t, 2H), 6.8 (m, 1H), 7.1 ppm (m, 4H); 19F NMR
(376 MHz): d=ꢀ139.0 (m, 2F), ꢀ143.7 ppm (br, 2F); 13C NMR
(100 MHz): d=46.9, 68.1, 106.4 (t, J=22.5 Hz), 112.7, 118.5, 120.6 (t, J=
14 Hz), 129.8, 143.9 (m), 144.6, 145.1 (m), 146.4 (m), 147.5 ppm (m); m.p.
1728C.
Scheme 1.
1,3-Bis(2,4,6-trimethylphenyl)-2-(trifluoromethyl)imidazolidine:,
N,N’-
Experimental Section
Bis-(2,4,6-trimethylphenylamino)ethane (205 mg, 0.69 mmol) with tri-
fluoroacetaldehyde ethyl hemiacetal (120 mg, 0.82 mmol) were placed in
a 20 mL scintillation vial equipped with a stir bar. Several drops of tolu-
ene were then added to the reaction vial with a catalytic amount of p-tol-
uenesulfonic acid and anhydrous magnesium sulfate. The reaction mix-
ture was then placed in a 408C oil bath and stirred for 24 h. The reaction
mixture was diluted with a small amount of CH2Cl2 and passed through a
small column of basic alumina to afford a white powder. The product
was then passed through a small column of silica gel (CH2Cl2) to afford
pure white powder (175 mg, 68%). 1H NMR (400 MHz, CDCl3, 258C,):
d=2.3 (s, 6H), 2.4 (brs, 12H), 3.4 (m, 2H) 3.8 (m, 2H), 5.0 (q, 1H, J=
4.1 Hz), 6.9 ppm (brs, 4H); 19F NMR (376 MHz): d=ꢀ78.1 ppm (d, J=
4.1 Hz, 3F); 13C (100 MHz): d=19.1 (br), 20.3 (br), 21.1, 51.1, 74.9 (q, J=
33.0 Hz), 130.1, 135.9, 139.4 ppm; m.p. 1778C.
General methods: Commerical reagents and solvents were purchased
from Aldrich and used without further purification. Deuterated NMR
solvents were purchased from Cambridge Isotope Laboratories. 1H and
13C NMR spectra were recorded on a Bruker Avance 400 spectrometer.
1H and 13C NMR chemical shifts were referenced to the residual solvent
peak. Gel permeation chromatography was performed in THF on a
Waters chromatograph equipped with four 5 mm Waters columns (300î
7.7 mm) connected in series with increasing pore size (10, 100, 1000, 105,
106 ä). A Waters 410 differential refractometer and 996 photodiode array
detector were employed. Modulated differential scanning calorimetry
(MDSC) measurements were recorded on a TA Instruments DSC 2920
with a ramp rate of 108minꢀ1 under a nitrogen atmosphere. Thermal
gravimetric analysis (TGA) measurements were recorded on a TA In-
struments Hi-Res 2950 with a ramp rate of 108minꢀ1 under a nitrogen
1,3-Bisphenyl-2-(trifluoromethyl)imidazolidine: N,N’-Diphenylethylene-
diamine (105 mg, 0.5 mmol) and trifluoroacetaldehyde ethyl hemiacetal
(80 mg, 0.55 mmol) were combined in a 20 mL scintillation vial equipped
with a stir bar. Several drops of toluene were then added to the reaction
vial. A catalytic amount of p-toluenesulfonic acid and anhydrous magne-
sium sulfate were then added to the reaction vial. The vial was placed in
a 708C oil bath and stirred for 24 h. The reaction mixture was diluted
with a small amount of CH2Cl2 and passed through a small column of
basic alumina to afford a white powder. The product was then passed
through a small column of silica gel (CH2Cl2) to afford the desired prod-
uct as a pure white powder (140 mg, 89%). 1H NMR (400 MHz, CDCl3,
258C): d=3.7 (m, 2H), 3.8 (m, 2H), 5.7 (q, 1H, J=4.7 Hz), 6.9 (m, 6H),
7.3 ppm (m, 4H); 19F NMR (376 MHz): d=ꢀ75.3 ppm (d, J=4.6 Hz,
3F); 13C (100 MHz): d=47.1, 73.9 (q, J=32.7 Hz), 114.1, 119.7, 129.7,
146.0 ppm; m.p. 788C.
purge.
1,3-Bis(2,4,6-trimethylphenyl)-2-(trichloromethyl)imidazolidine
was prepared by literature methods.[15]
1,3-Bis(2,4,6-trimethylphenyl)-2-(pentafluorophenyl)imidazolidine (4): In
a
5 mL scintillation vial, 2,3,4,5,6-pentafluorobenzaldehyde (201 mg,
1.7 mmol) was dissolved in a minimal amount of glacial acetic acid. N,N’-
Bis-(2,4,6-trimethylphenylamino)ethane (305 mg, 1.03 mmol) was placed
into the vial. The reaction mixture was stirred, after which the reaction
mixture was completely homogeneous. After 1 min the reaction mixture
felt warm to the touch and within 30 min a large amount of precipitate
had formed. The precipitate was washed with cold methanol. The precipi-
tate was then dissolved in CH2Cl2 and forced through a short silica plug
that had been previously treated with triethylamine. Evaporation of the
CH2Cl2 solution afforded 4 as a white crystalline solid (346 mg, 71%).
1H NMR (400 MHz, CDCl3, 258C): d=2.3 (s, 6H), d 2.3 2.6 (br, 12H),
3.5 3.6 (m, 2H), 3.9 3.4 (m, 2H), 6.4 (s, 1H), 6.9 ppm (br, 4H); 19F NMR
(376 MHz): d=ꢀ136.3 (m, 1F), ꢀ148.6 (m, 1F), ꢀ155.8 (m, 1F),
ꢀ163.1 ppm (m, 2F); 13C NMR (100 MHz): d=19.2 (br), 20.9, 51.4, 71.8,
129.9, 130.1, 130.6 (br), 130.5, 130.6, 131.8, 135.3, 135.9, 139.3, 143.8 ppm;
HRMS analysis (ESI): m/z calcd [M+H]+: 475.2173; found: 475.2188.
1,3-Dimethyl-2-(pentafluorophenyl)imidazolidine: A stirring solution of
N,N’-dimethylethylenediamine (113 mg, 1.2 mmol) in diethyl ether
(5 mL) was cooled to ꢀ788C. Pentafluorobenzaldehyde (248 mg,
1.2 mmol) was added to the stirring solution and was allowed to warm to
room temperature. The reaction mixture was stirred at room temperature
for 30 min, after which the solvent was removed under reduced pressure
to yield the desired product as colorless oil (284 mg, 89%). The product
decomposed rapidly in pure isolated form, but was stable over 24 h in
dilute solution. 1H NMR (400 MHz, CDCl3, 258C): d=2.2 (s, 6H), 2.6
1,3-Bisphenyl-2-(pentafluorophenyl)imidazolidine (6): N,N’-diphenyl-
ethylenediamine (200 mg, 0.94 mmol) and 2,3,4,5,6-pentafluorobenzalde-
hyde (230 mg, 0.94 mmol) were placed in a 20 mL scintillation vial,
equipped with a stir bar, and dissolved with CH2Cl2 (5 mL). A catalytic
4074
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4073 4079

N-Heterocyclic Carbenes
4073 4079
(m, 2H), 3.3 (m, 2H), 4.1 ppm (s, 1H); 19F NMR (376 MHz): d=ꢀ142.9
(m, 1F), ꢀ155.2 (t, J=21.1 Hz, 2F), ꢀ163.0 ppm (m, 2F).
tion mixture. The vial was then capped and sealed with electrical tape.
The vial was taken out of the glovebox and place in a 658C oil bath.
After 10 min, the solution color changed from colorless to light yellow.
After 3 h of stirring, the reaction was removed from the oil bath and
quenched with several drops of water. The polymer was precipitated
from cold methanol and isolated by filtration. The polymer was dried
1,3-Dimethyl-2-(2,3,5,6-tetrafluorophenyl)imidazolidine: A stirring solu-
tion of N,N’-dimethylethylenediamine (136 mg, 1.5 mmol) in diethyl
ether (5 mL) was cooled to 08C. 2,3,5,6-Tetrafluorobenzaldehyde
(267 mg, 1.5 mmol) was added and the solution was allowed to warm to
room temperature. The reaction mixture was stirred at room temperature
for 30 min, after which the solvent was removed under reduced pressure
to yield the desired product as light yellow oil (346 mg, 93%). 1H NMR
(400 MHz, CDCl3, 258C): d=2.3 (s, 6H), 2.7 (m, 2H), 3.4 (m, 2H), 4.2 (s,
1H), 7.0 ppm (m, 1H); 19F NMR (376 MHz): d=ꢀ139.9 (m, 2F),
ꢀ143.6 ppm (br, 2F); 13C (100 MHz): d=40.8, 54.2, 83.4, 105.9 (t, J=
84.9 Hz); 120.6 (t, J=44.3); 144.8 (m), 147.3 (m), 163.4 ppm (m).
under reduced pressure to yield poly-l-lactide as
a white powder
(192 mg, 80%).
Kinetic measurements with SImMesHCCl3: A solution of SImMesHCCl3
(0.056m, 150 mL, 8.4 mmol) [D6]benzene was diluted with [D6]benzene
(150 mL 0.028m) and charged into teflon screw-capped NMR tube along
with carbon disulphide (5 mL, 84 mmol) and SImMesCS2 (1 mg, 2.6 mmol).
The NMR tube was placed into a Varian 300 MHz NMR instrument cali-
brated to constant temperature with ethylene glycol and the disappear-
ance of the 2-H proton of the imidazolidine ring was monitored versus
the residual proton signal of the [D6]benzene solvent. Standard error
analysis was used assuming a 10% error in proton integrals and a negligi-
ble error in the temperature (Table S2 in the Supporting Information).
1,3-Dimethyl-2-(2,3,6 trifluorophenyl)imidazolidine: A stirring solution
of N,N’-dimethylethylenediamine (144 mg, 1.6 mmol) in diethyl ether
(5 mL) was cooled to 08C. 2, 3,6-Trifluorobenzaldehyde (256 mg,
1.6 mmol) was added and the solution was allowed to warm to room tem-
perature. The reaction mixture was stirred at room temperature for
30 min, after which the solvent was removed under reduced pressure to
yield the desired product as light yellow oil (346 mg, 93%). 1H NMR
(400 MHz, CDCl3, 258C): d=2.3 (s, 6H), 2.6 (m, 2H), 3.3 (m, 2H), 4.1 (s,
1H), 6.8 (m, 1H), 7.1ppm (m, 1H); 19F NMR (376 MHz): d=ꢀ119.4 (br,
1F), ꢀ137.3 (br, 1F), 142.8ppm (m, 1F).
1,3-Bisphenyl-2-(3,5-bis(trifluoromethyl)phenyl)imidazolidine: By using
the synthetic procedure described for 6, the reaction of N,N’-diphenyl-
ethylenediamine (105 mg, 0.5 mmol) and 3,5-bis(trifluoromethyl)benzal-
dehyde (122 mg, 0.5 mmol) afforded the product as an off-white powder
(206 mg, 95%). 1H NMR (400 MHz, CDCl3, 258C): d=3.8 (m, 2H), 3.9
(m, 2H), 6.7 (s, 2H) 6.8 (s, 2H), 6.8 (m, 2H), 7.2 (m, 4H), 7.7 (m, 1H),
7.9 ppm (s, 2H); 19F NMR (376 MHz): d=ꢀ63.1 ppm (s, 6F); 13C NMR
(100 MHz): d=46.9, 114.4, 119.3, 122.2, 122.6 (m), 124.9, 127.9 (br),
129.7, 132.1 (q, J=34 Hz), 144.6, 145.6 ppm (m); m.p. 1388C.
Kinetic measurements with SImMesHC6F5: Carbon disulphide (6.5 mL,
108 mmol) was injected into a solution of SImMesHC6F5 (0.036m, 300 mL,
10.8 mmol) in [D6]benzene in a teflon screw-capped NMR tube under ni-
trogen. SImMesCS2 (1 mg, 2.6 mmol) was added. The NMR tube was
placed into a calibrated, constant temperature Varian 300 MHz NMR in-
strument and the disappearance of the protons on the imidazolidine
backbone was followed. The integral values were calibrated with the ben-
zene peak.
Dependence of kobs on the concentration of carbon disulphide: The reac-
tion with a solution of SImMesHC6F5 (0.036m, 0.3 mL, 10.8 mmol) in
[D6]benzene was repeated with 20 times (216 mmol) and 30 times
(324 mmol) excess of carbon disulfide at 68.58C. The slight decrease in
kobs observed for [CS2]=0.072m and 1.080m is just outside our experi-
mental error and may be indicative of an association of CS2 with the
adduct.
1,3-Bis(2,4,6-trimethylphenyl)-2-(3,5-bis(trifluoromethyl)phenyl)imidazo-
lidine: By using the synthetic procedure described for 4,the reaction of
3,5-bis(trifluoromethyl)benzaldehyde (220 mg, 0.90 mmol) and N,N’-bis-
(2,4,6-trimethylphenylamino)ethane (270 mg, 0.90 mmol) afforded the
desired product as a white crystalline solid (407 mg, 87%). 1H NMR
(400 MHz, C6D6, 258C): d=2.1 (s, 6H), 2.3 (br, 12H), 3.5 3.6 (m, 2H),
3.9 3.4 (m, 2H), 5.9 (s, 1H), 6.7 (br, 4H), 7.5 (s, 1H), 7.9 ppm (s, 2H);
19F NMR (376 MHz, C6D6): d=ꢀ62.9 ppm (s, 6F); 13C NMR (100 MHz,
C6D6): d=20.1 (br), 20.9, 51.1, 80.1, 122.2 (m), 122.9, 125.7, 129.4 (br),
130.7 (br), 136.1, 136.9 (br), 139.5, 146.8 ppm; m.p. 1398C.
Dependence of kobs on [SImMesHC6F5]0: A solution of SImMesHC6F5
(0.036m, 150 mL, 5.4 mmol) in [D6]benzene was diluted to 0.016m with
[D6]benzene (190 mL) and the kinetics were measured with 20-fold excess
(108.0 mmol) of carbon disulfide at 68.58C. The observed kobs was identi-
cal within experimental error to that obtained at [SImMesHC6F5]0 =
0.036m (Table S2 in the Supporting Information).
Results and Discussion
Preparation of free carbene from SImMesHCCl3: A solution of SIm-
MesHCCl3 (0.056m, 150 mL 8.4 mmol) in [D6]benzene was diluted with
[D5]chlorobenzene (500 mL) in a teflon screw-capped NMR tube (J.
Young) under nitrogen, and the tube was placed to a 608C oil-bath. The
tube was heated slowly and placed under high vacuum every minute.
After 10 minutes at 1008C, 50 mL remained and the tube was cooled to
room temperature and of [D5]chlorobenzene (250 mL) was added. Analy-
sis of the mixture by 1H NMR spectroscopy revealed a 20% conversion
to the carbene SimMes; this was confirmed by charging the tube with an
authentic sample (3 mg, 10 mmol). 1H NMR (400 MHz, C6D5Cl): d=2.14
NHC adducts were prepared by acid-catalyzed diamine/al-
dehyde condensation reaction from readily available starting
materials, (Scheme 1). Three diamines were surveyed: N,N’-
bis(2,4,6-trimethylphenylamino)ethane (1), N,N’-bis(phenyl-
amino)ethane (2), and N,N’-bis(2,6-diisopropylphenylami-
no)ethane (3) with benzaldehyde derivatives functionalized
with electron-withdrawing groups in an effort to tune the
steric and electronic elements of the adduct (Scheme 1). Di-
amines 1 and 2 react with aldehydes to form adducts in
yields of 68 96%. In the presence of a catalytic amount of
acid, hemiacetals also react with 1 and 2 to form NHC ad-
ducts [Eq. (1)]; however, the reaction of 3 with aldehydes
and hemiacetals/acetals did not lead to adduct formation,
presumably due to the steric bulkof 3. The X-ray crystal
structure of 1,3-bis(2,4,6-trimethylphenyl)-2-(pentafluoro-
ꢀ
ꢀ
ꢀ
(s, 6H, p-CH3 Ar), 2.23 (s, 12H, o-CH3 Ar), 3.49 (s, 4H, N CH2),
6.75 ppm (s, 4H, CHarom.).
Preparation of SImMesCS2 from SImMesHCCl3: Carbon disulfide
(10 mL, 168 mmol) was added to a solution of SImMesHCCl3 (0.056m,
300 mL, 16.8 mmol) in [D6]benzene under nitrogen, and the solution was
heated to 608C for 10 min. Red crystals of the zwitterion SImMesCS2
1
ꢀ
precipitated. H NMR (300 MHz, CDCl3): d=2.23 (s, 6H, p-CH3 Ar),
ꢀ
ꢀ
2.54 (s, 12H, o-CH3 Ar), 4.20 (s, 4H, N CH2), 6.87 ppm (s, 4H, CHarom);
MS +ESI: m/z: 403.3 [M+Na]+; elemental analysis calcd (%) for
C22H24N2S2 (380.57): C 69.07, H 6.85, N 7.32; found: C 69.62, H 6.85, N
7.23.
Typical polymerization experiment: In
the glove box, compound 4 (8.1 mg,
17.1 mmol), benzyl alcohol (1.8 mg,
17.1 mmol), and l-lactide (240 mg,
1.7 mmol) were combined in a 20 mL
vial equipped with a stir bar. THF
(10 mL) was then added to the reac-
Chem. Eur. J. 2004, 10, 4073 4079
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4075

J. L. Hedrick, R. M. Waymouth et al.
FULL PAPER
phenyl)imidazolidine (4) is given in Figure 1. While the
quality of the data does not permit a detailed comparison of
the structural features, the structure reveals the expected
dazolidine (6) and 1,3-bisphenyl-2-(pentafluorophenyl)imi-
dazolidine (7) required extended reaction times and higher
temperatures to effect polymerization (Table 1, entries 4 and
5), but with significant loss in control as evidenced by the
generation of some low-molecular-weight oligomers. Lower
polymerization temperatures (entry 3, Table 1) generated
narrowly dispersed products, but with modest conversions.
No ROP was observed with the trifluoromethane, benzene,
2,3,4-triflourobenzene, p-nitrobenzene, or (3,5-trifluoro-
methyl)benzene adducts, irrespective of the reaction times
and temperature surveyed.[17]
In recent reports, NHC have been shown to be effective
transesterification catalysts.[18] Transesterification of dimeth-
yl terephthalate (DMT) with excess ethylene glycol (EG) to
give bis(2-hydroxyethyl) terephthalate (BHET), an impor-
tant precursor to poly(ethene terephthalate) (PET), was in-
vestigated with NHC adducts (Scheme 2). The reaction of
excess ethylene glycol with DMT in the presence of 4 and 5
(3.5 mol%) at 658C in THF generated BHET in 75% isolat-
ed yield after 3 h. The transesterification of DMT to BHET
with 7 required higher temperatures (1458C) and was car-
ried out in O-xylene to yield BHET in 70% isolated yield
after 12 h.
Figure 1. Plot of 1,3-bis(2,4,6-trimethylphenyl)-2-(pentafluorophenyl)imi-
dazolidine (4) with thermal ellipsoids drawn at the 50% probability
level.
Adducts 4 and 5 are also convenient synthons for the
preparation N-heterocyclic carbene organometallic com-
plexes.[12] Treatment of 4 and 5 with [(h3-C3H5)PdCl]2 at
808C in toluene yielded [(1,3-bis(2,4,6-trimethylphenyl)imi-
dazolin-2-ylidene)Pd(h3-C3H5)Cl] in >95% isolated yield in
less than 2 h [Eq. (2)]. These reactions can be carried out in
air with no prior solvent purification.
tetrahedral geometry at C1 and average bond lengths for
ꢀ
ꢀ
the C1 N1 (1.47(2) ä) and C1 N2 (1.46(2) ä) bonds that
are longer than those in the corresponding carbene
(1.352(5) and 1.345(5) ä, respectively), as observed for anal-
ogous acetylene or alcohol adducts.[10a,7i,16]
To demonstrate the utility of the NHC adducts as versa-
tile and effective catalyst pre-
cursors for ring-opening poly-
merization, the polymerization
of l-lactide was performed in
the presence of a benzyl alcohol
initiator and 1.5 equivalents of Entry Precatalyst
NHC adduct relative to initia-
Table 1. Characteristics of selected polylactides prepared from NHC adduct precatalysts. All experiments used
benzyl alchohol as the initiator with a target DP of 100. Mn was determined by gel-permeation chromatogra-
phy (GPC) calibrated with polystyrene standards in THF.
[a]
Adduct Polymerization
conditions
% Conv. Mn
[kgmolꢀ1
Mw/Mn
[a]
]
tor in 1 2m THF or toluene lac-
tide solutions, Table 1. The mo-
lecular weights closely tracked
1
4
5
6
658C, THF, 3 h
658C, THF, 3 h
658C, THF, 24 h
80
83
30
9890
9980
4030
1.13
the monomer/initiator (M/I)
ratio (targeted 100), and the
1H NMR
spectrum
clearly
2
3
1.15
shows resonances associated
the benzyl ester a-end group,
consistent with initiation from
benzyl alcohol as well as the
hydroxyl w-chain end. Com-
pound 4 and 1,3-bis(2,4,6-tri-
methylphenyl)-2-(trichlorometh-
yl)imidazolidine (5) effect the
1.10
ring-opening
polymerization
4
5
6
7
1108C, PhCH3, 24 h
66
4700
3200
1.11
1.52
(ROP) of lactide at tempera-
tures as low as 658C in high
yields with narrow polydisper-
sities (Table 1, entries 1 and
2). Conversely, the adducts
1,3-bis(2,4,6-trimethylphenyl)-2-
1448C, o-xylene, 12 h 68
(2,3,5,6 tetrafluorophenyl)imi- [a] From GPC data.
4076
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4073 4079

N-Heterocyclic Carbenes
4073 4079
of pentafluorobenzene from 7 (43%). The large endotherm
observed at 1408C in the MDSC experiment is assigned to
the melting point of 7, and the broad endotherm at 150
2108C is attributed to the elimination of pentafluoroben-
zene. The exothermic peakobserved at 180 8C, is tentatively
assigned to the dimerization of (1,3-diphenyl-2-imidazolidi-
nylidene) to form bis(1,3-diphenyl-2-imidazolidinylidene)
1
(9), as this olefin is clearly observed by H NMR analysis of
1
the sample after thermolysis.[19] (The H NMR spectra of 7
before and after thermolysis are shown in Figure 2.)
Several experiments were carried out to provide evidence
for the reversibility of adduct formation and the intermedia-
Scheme 2.
The data in Table 1 suggest
that not all NHC adducts are
equally efficient precatalysts for
ROP and transesterification
catalysis. The tetrafluoroben-
zene adduct 6 led to lower con-
versions of lactide than the pen-
tafluorobenzene adduct
658C, and the 1,3-diphenyl-sub-
stituted pentafluorobenzene
4 at
adduct 7 required much higher
temperature than the 1,3-mesi-
tyl adduct 4 to reach compara-
ble conversions in the ring-
opening of lactide. The 1,3-
bis(2,4,6-trimethylphenyl)-2-
(trifluoromethyl)imidazolidine
adduct (8) is surprisingly
robust, and did not ring-open
lactide after 12 h at 1008C.[17]
We attribute this behavior to
the differential thermal stability
of the adducts.
Figure 2. Thermogravimetric analysis and nonreversible heat flow of 7 as a function of temperature together
with the H NMR spectra.
1
This hypothesis is supported
by modulated differential scan-
ning calorimetry (MDSC) and
thermal gravimetric analysis (TGA) measurements of com-
pounds 4 7. TGA analysis of powdered samples of 4 and 5
began to show weight loss at 808C, presumably due to loss
of pentafluorobenzene or chloroform, whereas weight loss
in 6 and 7 began at 1308C and 1658C, respectively. The
MDSC and TGA experiments of 7 are shown in Figure 2. In
the TGA experiment, a 45% weight loss was observed at
165 2108C; this correlates with the theoretical% weight loss
cy of the free carbene in solution. Thermolysis of the penta-
fluorobenzene adduct 4 at 658C in the presence of excess
HCCl3 or DCCl3 cleanly yields 5 (or 5D1) and pentafluoro-
benzene [Eq. (3)].[20] Evidence for the intermediacy of the
free carbene was provided by thermolysis of the chloroform
adduct 5 at 1008C under vacuum in [D5]chlorobenzene.
After ten minutes, resonances attributable to free carbene,
that is, (1,3-bis(2,4,6-trimethylphenyl)imidazolin-2-ylidene)
Chem. Eur. J. 2004, 10, 4073 4079
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4077

J. L. Hedrick, R. M. Waymouth et al.
FULL PAPER
The observed first-order rate constants for the disappear-
ance of 4 (kobs =1.85 (ꢂ0.84)î10ꢀ5 sꢀ1 at 39.58C) and 5
(kobs =8.39 (ꢂ2.74)î10ꢀ5 sꢀ1 at 39.38C) imply that the elimi-
nation of chloroform from 5 is faster than elimination of
C6F5H from 4, but analysis of the temperature dependence
of the rate constants reveal activation parameters for 5
(Ea =94.7ꢂ15.4 kJmolꢀ1) and 4 (Ea =103.5ꢂ7.4 kJmolꢀ1)
that are within experimental error.
The observed differences in the rate of elimination of
C6F5H versus CHCl3 from 4 and 5, respectively, at 398C also
reflect the relative stabilities of the adducts in solution.
While the decomposition of 4 in the presence of CS2 in ben-
zene can be observed after ten minutes at room tempera-
ture, 5 is stable under similar conditions. Thus, the penta-
fluorobenzene adduct 4 is superior in solution to the chloro-
form adduct 5 as an in situ carbene source both, as a conse-
quence of its ease of synthesis
(10; 10% conversion), were observed, providing clear evi-
dence that the elimination of chloroform from 5 generates
the free carbene in solution even in the absence of transition
metals [Eq. 4)].
and its room-temperature sta-
bility.
Conclusion
In summary, the condensation
of diamines with aldehydes pro-
vides a convenient and general
source of alkane adducts of
saturated N-heterocyclic car-
benes. Unlike adducts derived
from chloroform, the penta-
fluorobenzene-based
adducts
are stable at room temperature.
Thermolysis of these adducts
generates the carbenes in solu-
Thermomolysis of either 4 or 5 in the presence of CS2
cleanly generates the zwitterionic CS2 adduct 11,[21] which
rapidly precipitates from nonpolar solvents. The clean for-
mation of the CS2 adduct 11 and pentafluorobenzene, and
absence of deuterium exchange in the 4/CDCl3 exchange ex-
periment [Eq. (3)] is consistent with a concerted elimination
of HCCl3 or HC6F5, rather than a stepwise ionic mecha-
nism.[10a,22]
tion, which we have shown are effective organic catalysts for
transesterification reactions and ring-opening polymeriza-
tion reactions. These adducts also provide convenient syn-
thons for the generation of transition-metal complexes. The
thermal elimination of the alkanes from the carbene adducts
depends on the substitutents on the carbene and the adduct,
providing a convenient method for tuning the rate of car-
bene generation in situ.
The kinetics of carbene generation were monitored by
1H NMR spectroscopy in [D6]benzene in the presence of
excess CS2 and a small crystal of 11.[23] The disappearance of
the adduct followed first-order kinetics (-ln([A]/[A]0 =kobst).
Experiments at several initial concentrations of the adduct 5
and CS2 confirmed that the rates are first-order with respect
to the adduct and zero-order for CS2.[23] If we adopt the
steady-state assumption for 10 [Eqs. (5) and (6)], then under
conditions in which k2[CS2]@kꢀ1[CHCl3] (k1[5]+kꢀ2[11]),
the observed rate constant kobs =k1, which is the rate con-
stant for the generation of the carbene.
Acknowledgement
The authors acknowledge support through the NSF Center for Polymer
Interfaces and Macromolecular Assemblies (CPIMA: NSF-DMR-
0213618) and an NSF-GOALI grant (NSF-CHE-0313993).
[1] M. Jones, R. A. Moss, Carbenes, Vols. 1 and 2, Wiley, New York,
1973 and 1975.
[2] General reviews; a) D. Bourissou, O. Guerret, F. P. Gabbai, G. Ber-
trand, Chem. Rev. 2000, 100, 39 91; b) H. Tomioka, Acc. Chem.
Res. 1997, 30, 315; c) W. A. Herrmann, C. Kocher, Angew. Chem.
1997, 109, 2256; Angew. Chem. Int. Ed. Engl. 1997, 36, 2162.
[3] This strategy was used by Wanzlickand co-workers to prepare 1,3-
bis(phenyl)-2-(trichlormethyl)imidazolidine see: a) H. W. Wanzlick,
H. J. Kleiner, Angew. Chem. 1961, 73, 493; b) H. W. Wanzlick,
Angew. Chem. 1962, 74, 129; Angew. Chem. Int. Ed. Engl. 1962, 1,
d½5Š
ꢀ
¼ k1½5Šꢀkꢀ1½CHCl3Š½10Š ¼ k1½5Š
dt
ꢀ
ꢁ
ð6Þ
k1½5Š þ kꢀ2½11Š
d½5Š
dt
ꢀkꢀ1½CHCl3Š
ꢀ
ꢁ k1½5Š
kꢀ1½CHCl3Š þ k2½CS2Š
4078
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 4073 4079

N-Heterocyclic Carbenes
4073 4079
75; c) H. W. Wanzlick, T. Esser; H. J. Kleiner, Chem. Ber. 1963, 96,
1208.
Helv. Chim. Acta 1999, 82, 2348; b) H. W. Wanzlick, H. Ahrens,
Chem. Ber. 1966, 99, 1580; c) H. W. Wanzlick, H. J. Kleiner, Chem.
Ber. 1963, 96, 3024; d) D. Enders, K. Breuer, J. Runsink, J. H. Teles,
Liebigs Ann. Chem. 1996, 2019.
[4] M. K. Denk, A. Thadani, K. Hatano, A. J. Lough, Angew. Chem.
1997, 109, 2719; Angew. Chem. Int. Ed. Engl. 1997, 36, 2607.
[5] a) A. J. Arduengo, R. L. Dais, R. Harlow, M. Kline, J. Am. Chem.
Soc. 1991, 113, 361; b) A. J. Arduengo, R. L. Dais, R. Harlow, M.
Kline, J. Am. Chem. Soc. 1992, 114, 5530.
[11] a) B. Cetinkaya, E. Cetinkaya, J. A. Chamizo; P. B. Hitchcock; H. A.
Jasim, H. K¸c¸kbay, M. F. Lappert, J. Chem. Soc. Faraday Trans. 1
1998, 2047 2054; b) H.-W. Wanzlick, F. Esser, H.-J. Kleiner, Chem.
Ber. 1963, 96, 1208 1212.
[12] a) T. M. Trnka, J. P. Morgan, M. S. Sanford, T. E. Wilhelm, M.
Scholl, T.-L. Choi, S. Ding, M. W. Day, R. H. Grubbs, J. Am. Chem.
Soc. 2003, 125, 2546; b) M. Scholl, S. Ding, C. W. Lee, R. H. Grubbs,
Org. Lett. 1999, 1, 953; c) M. S. Viciu, F. K. Zinn, E. D. Stevens, S. P.
Nolan, Organometallics 2003, 22, 3175.
[13] D. Enders, K. Breuer, G. Raabe, J. Runsink, J. H. Teles, J. P. Melder,
K. Ebel, S. Brode, Angew. Chem. 1995, 107, 1119; Angew. Chem.
Int. Ed. Engl. 1995, 34, 1021.
[14] a) K. Denk, P. Sirsch, W. A. Herrmann, J. Organomet. Chem. 2002,
649, 219 224; b) S. Randl, S. Gessler, H. Wakamatsu, S. Blechert,
Synlett 2001, 3, 430 432.
[15] A. J. Arduengo, J. C. Calabrese, F. Davidson, H. V. R. Dias, J. R.
Goerlich, R. Krafczyk, W. J. Marshall, M. Tamm, R. Schmutzler,
Helv. Chim. Acta 1999, 82, 2348.
[16] CCDC-242948 contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge via
www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge
CB21EZ, UK; fax: (+44)1223-336-033; or deposit@ccdc.cam.uk).
[17] See Supporting Information.
[18] a) G. W. Nyce, J. A. Lamboy, E. F. Connor, R. M. Waymouth, J. L.
Hedrick, Org. Lett. 2002, 4, 3587; b) G. A. Grasa, R. M. Kissling,
S. P. Nolan, Org. Lett. 2002, 4, 3583; c) R. Singh, R. M. Kissling,
M. A. Letellier, S. P. Nolan, J. Org. Chem. 2004, 69, 209.
[19] M. K. Denk, K. Hatano, M. Ma, Tetrahedron Lett. 1999, 40, 2057.
[20] In contrast, heating 5 in excess pentaflurobenzene does not yield 4,
but rather a mixture of unidentified products. Control experiments
reveal that the carbene 1,3-bis(2,4,6-trimethylphenyl)-4,5-dihydro-
imidzaol-2-ylidene leads to a similar mixture of products with no
evidence for the formation of 4.
[6] a) A. J. Arduengo, Acc. Chem. Res. 1999, 32, 913; b) A. J. Arduengo,
R. Krfczyk, Chem.-Ztg. 1998, 32, 6.
[7] For example: a) D. Bourissou, O. Guerret, F. P. Gabbai, G. Bertrand,
Chem. Rev. 2000, 100, 39; b) W. Herrmann, C. Kˆcher, Angew.
Chem. 1997, 109, 2256; Angew. Chem. Int. Ed. Engl. 1997, 36, 2162;
c) M. Scholl, T. M. Trnka, J. P. Morgan, R. H. Grubbs, Tetrahedron
Lett. 1999, 40, 2247; d) J. Huang, E. D. Stevens, S. P. Nolan, J. L. Pe-
tersen, J. Am. Chem. Soc. 1999, 121, 2674; e) T. Weskamp, W. C.
Schattenmann, M. Spiegler, W. A. Herrmann, Angew. Chem. 1998,
110, 2631; Angew. Chem. Int. Ed. 1998, 37, 2490; f) W. A. Herrmann,
M. Elison, J. Fisher, C. Kˆcher, G. J. Artus, Angew. Chem. 1995, 107,
2602; Angew. Chem. Int. Ed. Engl. 1995, 34, 2371; g) W. A. Herr-
mann, L. Goossen, C. Kˆcher, G. J. Artus, Angew. Chem. 1996, 108,
2980; Angew. Chem. Int. Ed. Engl. 1996, 35, 2805; h) M. J. Lappert,
J. Organomet. Chem. 1988, 358, 185; i) D. Enders, K. Breuer, G.
Raabe, J. Runsink, T. Henrique, J.-P. Melder, K. Ebel, S. Brode,
Angew. Chem. 1995, 107, 2602; Angew. Chem. Int. Ed. Engl. 1995,
34, 1021.
[8] For example: a) W. A. Herrmann, Angew. Chem. 2002, 114, 1342;
Angew. Chem. Int. Ed. 2002, 41, 1290; b) M. S. Sanford, M. Ulman,
R. H. Grubbs, J. Am. Chem. Soc. 2001, 123, 749; c) W. A. Herrmann,
L. J. Goossen, M. Spiegler, Organometallics 1998, 17, 2162; d) T.-L.
Choi, R. H. Grubbs, Chem. Commun. 2001, 2648; e) J. Haung, H.-J.
Schanz, E. D. Stevens, S. P. Nolan, Organometallics 1998, 18, 2370.
[9] a) R. Breslow, J. Am. Chem. Soc. 1958, 80, 3719; b) H. Stetter,
Angew. Chem. 1976, 88, 695; Angew. Chem. Int. Ed. Engl. 1976, 15,
639; c) D. Enders, J. H. Breuer, Helv. Chim. Acta 1996, 79, 1217;
d) J. Teles, J. Melder, K. Ebel, D. Enders, J. H. Breuer, Helv. Chim.
Acta 1996, 79, 61; e) M. S. Kerr, J. Read De Alaniz, T. Rovis, J. Am.
Chem. Soc. 2002, 124, 10298; f) D. Enders, U. Kallfass, Angew.
Chem. 2002, 114, 1822; Angew. Chem. Int. Ed. 2002, 41, 1743;
g) E. F. Connor, G. W. Nyce, M. Myers, A. Mˆck, J. L. Hedrick, J.
Am. Chem. Soc. 2002, 124, 914; h) G. W. Nyce, T. Glauser, E. F.
Connor, A. Mˆck, R. M. Waymouth, J. L. Hedrick, J. Am. Chem.
Soc. 2003, 125, 3046; i) G. A. Grassa, T. Guveli, R. Singh, S. P.
Nolan, J. Org. Chem. 2003, 68, 2812;.j) J. L. Hedrick, T. Magbitang,
E. F. Connor, T. Glauser, W. Volksen, C. J. Hawker, V. Y. Lee, R. D.
Miller, Chem. Eur. J. 2002, 8, 3308.
[21] N. Kuhn, H. Bohnen, G. Henkel, Z. Naturforsch. B 1994, 49, 1473.
[22] M. K. Denk, S. Gupta, J. Brownie, S. Tajammul, A. J. Lough, Chem.
Eur. J. 2001, 7, 4477.
[23] A small amount (2.6 mmol) of crystalline 11 was added to keep the
solution concentration of [11] constant during the kinetic runs. In
the absence of added 11, the buildup and rapid crystallization of 11
interfered with the kinetic analysis.
[10] a) A. J. Arduengo, J. C. Calabrese, F. Davidson, H. V. R. Dias, J. R.
Goerlich, R. Krafczyk, W. J. Marshall, M. Tamm, R. Schmutzler,
Received: February 27, 2004
Published online: July 19, 2004
Chem. Eur. J. 2004, 10, 4073 4079
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4079