Selective, green synthesis of six-membered cyclic carbonates by lipase-catalyzed chemospecific transesterification of diols with dimethyl carbonate

Article abstract of DOI:10.1002/adsc.201100822

A facile and green synthesis of six-membered cyclic carbonates, the potential monomers for isocyanate-free polyurethanes and polycarbonates, was achieved by transesterification of diols with dimethyl carbonate catalyzed by immobilized Candida antarctica lipase B, Novozym435, followed by thermal cyclization in a solvent-free medium. The difference in the chemospecificity of the lipase for the primary, secondary and tertiary alcohols as acyl acceptors was utilized to obtain a highly chemoselective synthesis of the cyclic carbonate in high yield. In the lipase-catalyzed reaction with diols, the product contained almost equal proportions of mono- and di-carbonates with 1,3-propanediol having two primary alcohols, a higher proportion of mono-carbonate with 1,3-butanediol having a primary and a secondary alcohol, and mainly mono-carbonate with 3-methyl-1,3-butanediol having a primary and a tertiary alcohol. The chemospecificity of cyclic carbonates formed by thermal treatment at 90 °C was closely related to the proportion of mono-carbonate. The yield of cyclic carbonate was 99.3% with 3-methyl-1,3-butanediol, 85.5% with 1,3-butanediol, and 43.2% with 1,3-propanediol. Copyright

Full text of DOI:10.1002/adsc.201100822

COMMUNICATIONS
DOI: 10.1002/adsc.201100822
Selective, Green Synthesis of Six-Membered Cyclic Carbonates by
Lipase-Catalyzed Chemospecific Transesterification of Diols with
Dimethyl Carbonate
Sang-Hyun Pyoa,* and Rajni Hatti-Kaula
a
Department of Biotechnology, Center for Chemistry and Chemical Engineering, Lund University, Box 124, SE-221 00
Lund, Sweden
Fax : (+46)-46-222-4713; phone: (+46)-46-222-4838; e-mail: Sang-Hyun.Pyo@biotek.lu.se
Received: October 21, 2011; Published online: March 7, 2012
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.201100822.
Abstract: A facile and green synthesis of six-mem-
bered cyclic carbonates, the potential monomers for
isocyanate-free polyurethanes and polycarbonates,
was achieved by transesterification of diols with di-
methyl carbonate catalyzed by immobilized Candi-
da antarctica lipase B, Novozymꢀ435, followed by
thermal cyclization in a solvent-free medium. The
difference in the chemospecificity of the lipase for
the primary, secondary and tertiary alcohols as acyl
acceptors was utilized to obtain a highly chemose-
lective synthesis of the cyclic carbonate in high
yield. In the lipase-catalyzed reaction with diols, the
product contained almost equal proportions of
mono- and di-carbonates with 1,3-propanediol
having two primary alcohols, a higher proportion of
mono-carbonate with 1,3-butanediol having a pri-
mary and a secondary alcohol, and mainly mono-
carbonate with 3-methyl-1,3-butanediol having a pri-
mary and a tertiary alcohol. The chemospecificity
of cyclic carbonates formed by thermal treatment at
908C was closely related to the proportion of
mono-carbonate. The yield of cyclic carbonate was
seals, coatings and high performance adhesives, and
are even expected to find use in biomedicine due to
their features of biocompatibility and low toxicity.[7–10]
Carbon dioxide, dialkyl carbonates, diphenyl carbon-
ate, and phosgene derivatives have usually been used
as sources of carbonate groups in the cyclic carbo-
nates. One of the commonly reported routes for the
synthesis of cyclic carbonates is the transesterification
of polyols with a dialkyl carbonate in a metal-[11] or
enzyme (lipase)-catalyzed reaction.[12,13] The majority
of the studies so far report on the synthesis of five-
membered carbonates,[1–4,14] while the six-membered
cyclic carbonate is the preferred structure for use in
the ring opening polymerization (ROP) process.[11,15,16]
The lipase-catalyzed synthesis of the six-membered
cyclic trimethylene carbonate from 1,3-diol and di-
methyl or diethyl carbonate has been achieved earlier
in a solvent system of acetonitrile and toluene (4:1,
v/v) using a very high concentration (900% w/w of
the diol) of the immobilized Candida antarctica lipase
B (Novozymꢀ435, N435), however with only moder-
ate yield (53%) and low productivity.[13] We have re-
cently reported the synthesis of six-membered cyclic
carbonates with functional groups using the lipase B-
mediated reaction between trimethylolpropane
99.3% with 3-methyl-1,3-butACHTUNTRGNEUNGanediol, 85.5% with
1,3-butanediol, and 43.2% with 1,3-propanediol.
Keywords: 1,3-butanediol; chemoselectivity; 3-
methyl-1,3-butanediol transesterification; 1,3-pro-
panediol; thermal cyclization
(TMP) and dialkyl carbonate in a solvent-free
medium.[17] It was shown that the lipase catalyzes
mainly the formation of linear carbonates and their
conversion to cyclic carbonates is promoted by in-
creased temperature. It was also shown that the use
of hydrophobic solvents in the reaction medium in-
creased the proportion of cyclic carbonates formed.[18]
Cyclic carbonates have attracted attention in recent The equal reactivity of the lipase for the primary alco-
years as potential monomers for the production of hol groups in the diols and polyols investigated so far
polyurethanes and polycarbonates through a non- poses a limitation in terms of the formation of un-
toxic, phosgene-free route.[1–6] These aliphatic poly- wanted side products (e.g., di- and higher carbonates),
mers have a broad range of traditional applications, hence resulting in lower product yield and atom effi-
for example, in engineering, optical devices, seatings, ciency.
Adv. Synth. Catal. 2012, 354, 797 – 802
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
797

Sang-Hyun Pyo and Rajni Hatti-Kaul
COMMUNICATIONS
Reaction kinetics of lipase-catalyzed reactions of
1-, 2- and tert-butanol with dimethyl carbonate
(DMC): To start with, the rate of lipase-catalyzed re-
action between DMC and primary, secondary, and ter-
tiary mono-ols (1-, 2- and tert-butanol), individually
and in mixture, was determined. Figure 1 shows a 1st
order plot of rate for the different reactions. As ex-
pected, the rates of the reactions with individual alco-
hols were in the order: primary>secondary>tertiary
alcohol, and in accordance with the earlier reports on
the transesterification of various alcohols with acyl
donors like fatty acids.[19,20] The rate constant, kd for
1-butanol was 0.29, which was about 4 times higher
than that of 2-butanol (0.075) (Table 1), while no re-
action was observed between tert-butanol and DMC.
In the reaction with a mixture of two alcohols, the re-
action rate for 1-butanol was decreased to 0.168, pos-
sibly due to inhibition by increased total substrate
concentration and the competing reaction, that for 2-
butanol was reduced ten-fold as a result of which the
difference in the reaction rate between the primary
and secondary alcohols was increased to 24 -old. In
the reaction with a mixture of 1- and tert-butanol, the
kd for the primary alcohol was only slightly reduced
Figure 1. First order plot of the rate of the lipase-catalyzed
&
^
reaction of DMC with 1-butanol ( ), 2-butanol ( ) and tert-
butanol ( ), respectively; and of the reaction with 1-butanol
*
~
( ) and 2-butanol (&) in a mixture of 1-/2-butanol; and 1-
*
butanol ( ) in a mixture of 1-/tert-butanol.
In the present study, we demonstrate that the dif- which could be attributed only to the substrate inhibi-
ferential specificity of the lipase for primary, secon- tion effect since there was no competing reaction.
dary and tertiary alcohol groups in the substrate can
Kinetics of the lipase-catalyzed reactions of 1,3-
be used to increase the chemoselectivity of the pro- propanediol (1,3-PDO), 1,3-butanediol (1,3-BDO),
duction and yield of six-membered cyclic carbonates and 3-methyl-1,3-butanediol (3-Me-1,3-BDO) with
from diols and dimethyl carbonate.
DMC: Scheme 1 shows the various intermediates and
Table 1. Kinetics of the reaction of 1-, 2-, and tert-butanol with DMC catalyzed by Novozymꢀ435 at 608C.
Run
Substrate
kd
t1/2 [h]
In mixture of 1- and 2-butanol
In mixture of 1- and tert-butanol
kd
t1/2
kd
t1/2
1
2
3
1-butanol
2-butanol
tert-butanol
0.290
0.075
0
2.4
9.3
NA
0.168
0.0071
NA
4.1
98.1
NA
0.243
NA[a]
0
2.9
NA
NA
[a]
NA: not applicable.
Scheme 1. Reaction pathway for production of cyclic carbonates by lipase-catalyzed reaction and thermal cyclization.
798
asc.wiley-vch.de
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 797 – 802

Selective, Green Synthesis of Six-Membered Cyclic Carbonates
(Figure 2 and Table 2). All the reaction rates were
higher than that for the mono-ol, and over 80% con-
version took place within the initial 3 h for all the
substrates. The highest rate of kd =0.550 was obtained
in the reaction with 1,3-PDO, with 99% conversion in
7 h. The reaction profile showed the formation of
over 60% mono-carbonate P3 within 3 h, which de-
creased with time while the proportion of the unde-
sired di-carbonate increased to about 40% at 9 h
(Figure 2, B). The other possible mono-carbonate P4
was not observed in the chromatogram [Figure 3,
(B1)]. In the case of 1,3-BDO, 96% substrate conver-
sion was obtained at 9 h (kd =0.417) and the product
contained only 9.5% of di-carbonate. On the other
hand, the reaction rate was only slightly lower with 3-
Me-1,3-BDO with 95% conversion at 9 h (kd =0.409)
but without any di-carbonate formation. These results
showing slow and no reaction with secondary and ter-
tiary alcohols, respectively, using N435, were in agree-
ment with the reactions with the mono-ols.
Synthesis of cyclic carbonates by lipase-catalyzed
transesterification reaction and thermal cyclization
of mono-carbonate: As in our earlier reports,[17,18] the
lipase-catalyzed reaction was followed by a thermal
treatment step in the presence of molecular sieves
only to promote the cyclization of mono-carbonates.
Table 3 summarizes the results of reactions between
1 g substrates (1,3-PDO, 1,3-BDO and 3-Me-1,3-
BDO, respectively) and 10 mL DMC using 10%
Figure 2. (A) First order plot of the reaction rate of 1,3-
PDO ( ), 1,3-BDO ( ), and 3-Me-1,3-BDO ( ) with DMC
~
&
*
AHCTUNGTREGUN(NN w/w) N435 for 7 h reaction at 608C, and the thermal
catalyzed by N435 at 608C. (B) Proportion of products
reaction at 908C. It is clear from Table 3 and Figure 3
(B) that the resulting content of cyclic carbonate in
the final product was dependent on the proportion of
mono-carbonate formed in the enzymatic reaction.
In the case of the 1,3-PDO, the di-carbonate (P6)
~
~
^
formed, P3+P5 ( ) and P6 ( ) from 1,3-PDO; P3+P5 ( )
&
^
and P6 ( ) from 1,3-BDO; P3+P5 ( ) and P6 (&) from 3-
Me-1,3-BDO in the reactions.
products formed in a reaction between a diol and di- content remained almost unchanged at about 43%
methyl carbonate. The desired cyclic carbonate prod- during the thermal step, while the cyclic carbonate
uct (P5) can be produced only via mono-carbonate (P5) content increased from 5% to 43% after 40 h of
(P3); formation of di-carbonate (P6) prevents the for- thermal treatment [Table 3 (Run 1), Figure 3 (A2)].
mation of P5, and hence should be minimized for pro- In the 1,3-BDO reaction, 85.5% cyclic carbonate was
ducing the cyclic carbonate at high yield. From the re- obtained from the product of the lipase reaction con-
sults obtained with mono-ols, it could be expected taining 86.3% mono- and 3.3% cyclic carbonate
that the selectivity for mono-carbonate formation [Table 3b (Run 2)]. A highly selective synthesis of
from a diol could be enhanced by introducing substi- cyclic carbonate (99.3%, P5) with a small amount of
tutes on one of the two alcohol groups. As a next residual mono-carbonate (<1%, P3) was achieved
step, the reaction rates for N435-catalyszd esterifica- from 3-Me-1,3-BDO having primary and tertiary alco-
tion of the diols 1,3-PDO, 1,3-BDO and 3-Me-1,3- hols through lipase-mediated and thermal reaction
BDO, respectively, with DMC were compared [Table 3 (Run 3), Figure 3 (C2)]. The thermal reac-
tion was focused on an equilibrium shift to cyclic car-
bonate, P5, which has minimum content of ester
AHCTUNGTERNN(GUN CO OCH3), and this was achieved by continuous
Table 2. Kinetics of the reaction of 1,3-diols with DMC cata-
lyzed by Novozymꢀ435 at 608C.
À
À
removal of the resulting methanol by free evapora-
tion, which prevented the reverse reaction from
taking place. The rate of cyclization decreased in the
order of primary>secondary>tertiary alcohol.
In conclusion, the syntheses of various six-mem-
bered cyclic carbonates were achieved at high yields
Run
Substrate
kd
t1/2 [h]
1
2
3
1,3-propanediol
1,3-butanediol
3-methyl-1,3-butanediol
0.550
0.417
0.409
1.3
1.7
1.7
Adv. Synth. Catal. 2012, 354, 797 – 802
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
799

Sang-Hyun Pyo and Rajni Hatti-Kaul
COMMUNICATIONS
Figure 3. Chromatograms showing products in Table 3 obtained from (1) lipase-catalyzed reaction of (A) 1,3-PDO, (B) 1,3-
BDO, and (C) 3-Me-1,3-BDO, respectively, with DMC and (2) thermal cyclization.
800
asc.wiley-vch.de
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 797 – 802

Selective, Green Synthesis of Six-Membered Cyclic Carbonates
Table 3. Summary of synthesis of cyclic carbonates by a process involving lipase-catalyzed transesterification and thermal
cyclization.
Run
Substrate
Lipase reaction[a]
Product [%]
Thermal reaction at 908C[b]
Time [h] Cyclic carbonate
Structure P5 [%]
Time [h]
Conversion [%]
P3
P5
P6
3
5
7
57.7
55.8
50.7
1.4
6.9
5.2
6.2
34.3
42.9
65.4
97.0
98.8
6
15
40
8.7
24.2
43.2
1
2
3
5
7
75.7
85.5
86.3
1.3
2.4
3.3
2.4
5.5
8.7
79.4
93.4
98.3
6
30
48
6.8
73.3
85.5
3
5
7
78.0
92.2
94.8
1.5
2.6
4
0
0
0
79.5
94.9
98.8
24
48
65
10.8
95.2
99.3
3
[a]
1 g substrates (1,3-PDO, 1,3-BDO and 3-Me-1,3-BDO, respectively) were reacted with 15 mL DMC at 608C.
Thermal reactions were carried out using 0.2 g product obtained from lipase catalysed reactions with each substrate.
[b]
shaking in a thermomixer (MKR 13, HLC Biotech, Germa-
ny) at 700 rpm and 608C. The reactions were started by ad-
dition of 10 mg N435 along with 200 mg molecular sieves.
Aliquots were withdrawn at different time intervals for anal-
ysis of the reaction components.
by a lipase-catalyzed reaction between diols and
DMC combined with thermal cyclization without
using any catalyst. A strategy to be considered for
achieving higher selectivity would be to minimize the
proportion of di-carbonate, by utilizing the differen-
ces in the specificity of the lipase for primary, secon-
dary and tertiary alcohol groups. The highest chemo-
selectivity for the synthesis of a cyclic carbonate was
obtained from reaction between a diol having primary
and tertiary alcohol groups, and DMC followed by
intra-molecular reaction of the carbonate with the hy-
droxy group at high temperature. These results are
a significant improvement over the earlier reports on
the synthesis of six-membered cyclic carbon-
ate.[12,13,17,21] Furthermore, the reaction did not involve
any solvents; the DMC used in excess served also as
the solvent, and is easily recovered by distillation and
recycled.
The rate constant (kd) of first order reaction was calculat-
ed using Eq. (1):
ln([M0]/[M])=kdt
(Eq. 1)
where [M0] and [M] are the concentrations of the alcohol
substrate at time zero and time t (in hours), respectively.
The half-life, t1/2, that is, the time for the substrate concen-
tration to reach half of the initial concentration, was calcu-
lated using Eq. (2):
t1/2 =ln(2)/kd
(Eq. 2)
For studying the thermal cyclization, 1 g each of 1,3-PDO,
1,3-BDO and 3-Me-1,3-BDO, respectively, were reacted
with 10 mL DMC using 2 g molecular sieves and 10% (w/w)
N435 for 7 h at 608C, followed by filtration to remove solids
including the immobilized lipase and molecular sieves, and
evaporation. The product samples (0.2 g) were then heated
with shaking at 908C under atmospheric pressure.
Experimental Section
Quantitative Analyses and Structure Elucidation
Quantitative analyses of reaction components were per-
formed using gas chromatography (GC, Varian 430-GC,
Lipase-Catalyzed Reaction and Thermal Cyclization
Novozymꢀ435 (N435) was a generous gift from Novozymes
A/S (Bagsvaerd, Denmark). Dimethyl carbonate (97%,
DMC), 1,3-propanediol, (Æ)-1,3-butanediol (>99%), 3-
methyl-1,3-butanediol (>97%), 1-butanol, 2-butanol, tert-
butanol, chloroform-d (CDCl3) and molecular sieves (4 ꢂ)
were purchased from Sigma–Aldrich. HPLC grade acetoni-
trile was purchased from Merck (Germany). All chemicals
were used without further treatment. To investigate the re-
activity of lipase with different alcohol groups, 1.5 mmol
substrates, for example, 111 mg mono-ols (1-, 2- and tert-bu-
tanol, respectively), and diols (114.2 mg 1,3-PDO, 135.2 mg
1,3-BDO, and 156.23 mg 3-Me-1,3-BDO, respectively) were
dissolved in 1.5 mL (17.8 mmol) DMC in a 5 mL vial by
Varian, USA) equipped with a FactorFour Capillary
column, VF-1 ms (Varian, 15 mꢃ0.25 mm) and a flame ioni-
zation detector. The initial column oven temperature was in-
creased from 50 to 2508C at a rate of 208CminÀ1. The sam-
ples, diluted with acetonitrile, to a concentration of 0.1–
0.5 mgmLÀ1, were injected in the split injection mode of
10% at 2758C. The conversion of the substrates and ratio of
products formed were calculated by comparison of peak
areas on the gas chromatograms.
The structures of the products were then determined by
1H NMR using a 400 MHz NMR spectrometer (Bruker, Ul-
traShield Plus 400, Germany), from which it was seen that
high proportions of mono-carbonates were obtained in the
Adv. Synth. Catal. 2012, 354, 797 – 802
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
801

Sang-Hyun Pyo and Rajni Hatti-Kaul
COMMUNICATIONS
initial stages of the lipase-catalyzed reaction, while more di-
carbonates were obtained by performing the reaction for
a longer time. Cyclic carbonates were prepared from mono-
carbonate by thermal treatment at 908C.
[2] Z.-Z. Yang, L.-N. He, C.-X. Mioa, S. Chanfreau, Adv.
Synth. Catal. 2010, 352, 2233.
[3] J. Song, Z. Zhang, S. Hu, T. Wu, T. Jiang, B. Han,
Green Chem. 2009, 11, 1031.
Products of 1,3-PDO Reaction (see Supporting Informa-
tion, Figure S1): 1H NMR (400 MHz, CDCl3). Mono-car-
bonate (3-methoxycarbonyloxypropan-1-ol, P3): d=1.861
(2H, m), 2.535 (1H, t), 3.669 (2H, t), 3.733 (3H, s), 4.237
(2H, t). Cyclic carbonate (trimethylene carbonate, P5): d=
2.068 (2H, m), 4.250 (4H, t). Di-carbonate [1,3-bis(methoxy-
carbonyloxy)propane, P6]: d=2.058 (2H, m), 3.792 (3H, s),
3.797 (3H, s), 4.253 (4H, t).
Products of 1,3-BDO reaction (see Supporting Informa-
tion, Figure S2); 1H NMR (400 MHz, CDCl3). Mono-car-
bonate (3-methoxycarbonyloxybutan-1-ol, P3): d=1.246
(3H, d), 1.760 (1Ha, m), 1.848 (1Hb, m), 1.996 (1H, d),
3.955 (1H, m), 3.797 (3H, s), 4.257 (1Ha, m), 4.387 (1Hb,
m). Cyclic carbonate (1-methyl-trimethylene carbonate, P5):
d=1.346 (3H, d), 1.957 (2H, m), 4.221 (2H, b), 4.899 (1H,
b). Di-carbonate [1,3-bis(methoxycarbonyloxy)butane, P6]:
d=1.341 (3H, d), 1.968 (2H, m), 3.784 (3H, s), 3.789 (3H,
s), 4.229 (2H, t), 4.907 (1H, m).
[4] N. Della Caꢄ, B. Gabriele, G. Ruffolo, L. Veltri, T. Za-
netta, M. Costa, Adv. Synth. Catal. 2011, 353, 133.
[5] L. Liao, C. Zhang, S. Gong, Eur. Polym. J. 2007, 43,
4289.
[6] J. Fahlꢅn. R. Martinsson, B. Midelf, Eur. Coating J.
2007, 6, 28.
[7] S. H. Pyo, P. Persson, M. A. Mollaahmad, K. Sçrensen,
S. Lundmark, R. Hatti-Kaul, Pure Appl. Chem. 2012, in
press.
[8] Z. Zhang, R. Kuijer, S. K. Bulstra, D. W. Grijpma, J.
Feijen, Biomaterials 2006, 27, 1741.
[9] Y. Yamamoto, S. Kaihara, K. Toshima, S. Matsumura,
Macromol. Biosci. 2009, 9, 968.
[10] S. A. Guelcher, Tissue Eng. Part B 2008, 14, 3.
[11] A.-C. Albertsson, M. J. Sjoeling, Macromol. Sci. Pure
Appl. Chem. A 1992, 29, 43.
[12] S. C. Kim, Y. H. Kim, H. Lee, D. Y. Yoon, B. K. Song,
J. Mol. Catal. B: Enzym. 2007, 49, 75.
[13] H. Tasaki, K. Toshima, S. Matsumura, Macromol.
Biosci. 2003, 3, 436.
Products of 3-Me-1,3-BDO reaction (see Supporting In-
1
formation, Figure S3); H NMR (400 MHz, CDCl3). Mono-
carbonate (3-methoxycarbonyloxy-3-methyl-butan-1-ol, P3):
d=1.267 (6H, s), 1.835 (1H, s), 1.873 (2H, t), 3.776 (3H, s),
4.319 (2H, t). Cyclic carbonate (1,1-dimethyl-trimethylene
carbonate, P5): d=1.507 (6H, s), 2.022 (2H, t), 4.451 (2H,
t).
[14] D. J. Darensbourg, A. Horn Jr, A. I. Moncada, Green
Chem. 2010, 12, 1376.
[15] S. Kꢅki, J. Tçrçk, G. Deꢆk, M. Zsuga, Macromolecules
2001, 34, 6850.
[16] H. R. Kricheldorf, B. Weegen-Schulz, Polymer 1995, 36,
4997.
[17] S. H. Pyo, P. Persson, S. Lundmark, R. Hatti-Kaul,
Green Chem. 2011, 13, 976.
[18] S. H. Pyo, K. Nuszkiewicz, P. Persson, S. Lundmark, R.
Hatti-Kaul, J. Mol. Catal. B: Enzym. 2011, 73, 67.
[19] T. Nishio, M. Kamimura, M. Murata, Y. Terao, K.
Achiwa, J. Biochem. 1998, 104, 681.
Acknowledgements
The work was supported by the Swedish Foundation for Stra-
tegic Environmental Research (Mistra) and Perstorp AB.
[20] S. J. Kuo, K. L. Parkin, J. Am. Oil Chem. Soc. 1993, 70,
393.
[21] G. Rokicki, T. Kowalczyk, M. Glinski, Polym. J. 2000,
32, 381.
References
[1] B. Barkakaty, K. Morino, A. Sudo, T. Endo, Green
Chem. 2010, 12, 42.
802
asc.wiley-vch.de
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 797 – 802