Group 2 Catalysis for the Atom-Efficient Synthesis of Imidazolidine and Thiazolidine Derivatives

Article abstract of DOI:10.1002/chem.201501328

A wide variety of functionalised imidazolidine-2-ones and -thiones, 2-imino-imidazolidines and thiazolidine-2-thiones have been synthesised under very mild reaction conditions by using simple and cost-effective alkaline earth bis(amide) precatalysts, [Ae{

Full text of DOI:10.1002/chem.201501328

DOI: 10.1002/chem.201501328
Full Paper
&
Synthetic Methods
Group 2 Catalysis for the Atom-Efficient Synthesis of
Imidazolidine and Thiazolidine Derivatives
Merle Arrowsmith, Michael S. Hill,* and Gabriele Kociok-Kçhn[a]
Abstract: A wide variety of functionalised imidazolidine-2-
ones and -thiones, 2-imino-imidazolidines and thiazolidine-2-
thiones have been synthesised under very mild reaction
conditions by using simple and cost-effective alkaline earth
bis(amide) precatalysts, [Ae{N(SiMe3)2}2(THF)2] (Ae=Mg, Ca,
Sr). The reactions ensue with 100% atom efficiency as one-
pot cascades from simple, commercially available terminal
alkyne and heterocumulene reagents. The reactions take
place through the initial assembly of propargylamidines,
which are utilised in subsequent cyclisation reactions
through addition of the isocyanate, isothiocyanate and, in
one case, carbon disulfide reagents. This reactivity is de-
duced to take place through a well-defined sequence of het-
erocumulene hydroacetylenation and alkyne hydroamidation
steps, which are all mediated at the alkaline earth centre.
The rate and regioselectivity of the cyclisation reactions are,
thus, found to be heavily dependent upon the identity of
the catalytic alkaline earth centre employed. Similarly, the
selectivity of the reactions was observed to be profoundly
affected by stereoelectronic variations in the individual
substrates, albeit by a similar Group 2-centred reaction
mechanism in all cases studied.
Introduction
Heterocyclic moieties are constituents of two-thirds of the
top-selling small-molecule pharmaceuticals in the USA.[1] The
imidazolidine-2,4-dione (hydantoin) moiety is a particularly
notable constituent of many biologically active compounds
and numerous hydantoin derivatives have been identified as
anticonvulsant, antiulcer, antiarrhythmic, antimuscarinic,
antiviral and antidiabetic agents.[2] For example, phenytoin is
commonly used in the treatment of epilepsy[3] and aplysinop-
sin, isolated from the sponge Aplysinopsis reticulate, has been
shown to exhibit cytotoxicity against cancer cells.[4] The agonist
of the human-androgen receptor, GLPG04924,[5] and the
hydantoin muscle relaxant, dantrolene,[6] have been widely
employed in a medical context. Many herbicides also contain
hydantoin skeleta as an integral part of their structure, and
they are also applied as intermediates for the synthesis of
enantiomerically pure amino acids by dynamic kinetic
resolution.[7,8]
hydantoin molecules from simple and inexpensive starting ma-
terials is clearly very desirable. In this regard, several transition-
metal-catalysed methods have been reported. A copper-based
amination reaction of esters with di-tert-butyldiaziridinone has
been shown to afford 1,3,5-trisubstituted hydantoins,[9] where-
as a palladium-catalysed carbonylation reaction of aldehydes
with ureas and carbon monoxide furnishes 5-, 3,5-, and 1,3,5-
The development of efficient and preferentially catalytic
methods for the rapid construction of molecularly diverse
[a] Dr. M. Arrowsmith, Prof. M. S. Hill, Dr. G. Kociok-Kçhn
Department of Chemistry, University of Bath
Claverton down, Bath, BA2 7AY (UK)
substituted hydantoins[10] and
a nickel-catalysed protocol
starting from acrylates and isocyanates has recently been
described.[11] Of most relevance to the current work, several
processes have been described, in which the hydantoin core is
assembled from one molecule of phenylacetylene and two
molecules of isocyanate through iron,[12] ruthenium[13] or
manganese catalysis.[14] Although described in some cases as
proceeding as a [2+2+1]cycloaddition, these processes are
otherwise mechanistically uncertain.
E-mail: msh27@bath.ac.uk
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201501328. It contains experimental proce-
dures, full characterisation data, and details of the X-ray analyses of com-
pounds Z-3, E-12, E-25, Z-32, II and V. CCDC1008314, 1008315, 1008316,
1051077, 1051078, and 1051079 contain the supplementary crystallograph-
ic data for this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_re-
quest/cif.
Chem. Eur. J. 2015, 21, 10548 – 10557
10548
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
The alkaline earth elements (Mg, Ca, Sr and Ba) comprise an
alternative suite of underexploited but inexpensive and poten-
tially more sustainable and environmentally benign catalytic
elements in their effectively invariant 2+ oxidation state. For
some time, we have been engaged in a programme of re-
search to develop a defined catalytic and reaction chemistry,
derived largely from sequences of metal-centred s-bond meta-
thesis and polarised insertion reactions.[15] Since our initial
report of the calcium-catalysed intramolecular hydroamination
of aminoalkenes (Scheme 1A),[16,17] we and others have applied
alkaline earth pre-catalysts to an ever-growing array of molecu-
lar catalytic reactions.[18–33] Of most relevance to the current
work, we have previously reported that complex bis(hydan-
toins) (II, inset Scheme 1) may be synthesised, albeit in a
stoichiometric sense, through a magnesium-based cascade
reaction between phenylacetylene and an organic isocya-
nate.[33] Subsequent to this discovery, which was rationalised to
take place through a sequence of intramolecular isocyanate
hydroacetylenation and intermolecular CꢀC and C=C hydroa-
midation steps, we have described in preliminary form an ex-
tension of this chemistry to a one-pot catalytic regime.[34] In
this latter case, reactions catalysed by the readily available bis-
hexamethyldisilazides [Ae{N(SiMe3)2}2(THF)2] (Ia, Ae=Mg; Ib,
Ae=Ca; Ic, Ae=Sr) allowed the facile synthesis of a variety of
highly functionalised imidazolidin-2-ones with 100% atom effi-
ciency (Scheme 1B and C). These reactions take place via the
initial catalytic formation of
a propargylamidine (III in
Scheme 1B),[35] whereupon addition of an organic isocyanate
initiates a further cascade of reactions involving isocyanate in-
sertion into the catalytically active Group 2 amidinate (IV in
Scheme 1B and C), intramolecular AeÀN insertion of the alkyn-
yl residue and protonolysis of the cyclised alkaline earth vinyl
intermediate by the remaining propargylamidine (III) to
regenerate IV and provide the imidazolidin-2-one derivative (V,
Scheme 1C). Herein, we provide a full description of the scope
of this reactivity and reveal that this approach is easily
extended to alternative sulfur-based heterocumulenes for the
synthesis of imidazolidin-2-thiones and thiazole derivatives.
Results and Discussion
We have recently reported the use of the alkaline earth bis-
(amido) complexes, [Ae{N(SiMe3)2}2(THF)2] (Ae=Mg Ia, Ca Ib, Sr
Ic) for the synthesis of propargylamidines by hydroacetylena-
tion of carbodiimides.[35] Following this procedure (N,N’-diiso-
propyl)-phenylpropargylamidine (1) was synthesised in situ by
using 5 mol% of Ic. Subsequent addition of one molar equiva-
lent of tert-butylisocyanate resulted in quantitative formation
of the corresponding imidazoli-
din-2-one (3) within the first
point of analysis at room tem-
perature (Scheme 2; Table 1,
entry 3). The resultant 1H NMR
spectrum displayed two new
characteristic benzylidene sin-
glets present in a 22:78 ratio
at d=6.50 and 5.95 ppm, re-
spectively, coupling in the
13C{1H} NMR spectrum to signals
at d=112.4 and 120.9 ppm, re-
spectively. These were assigned
by a NOESY NMR experiment to
the Z and E isomers of 3, respec-
tively (Figure 1).
The catalyst loading could be
lowered to 0.5 mol%, affording
(Z)-3 and (E)-3 in the same ratio
with an 87% yield from an
18 hour reaction at room tem-
perature (Table 1, entry 4). The
reaction of 1 with tert-butyliso-
cyanate by using 5 mol% of the
analogous magnesium and calci-
um precatalysts, Ia and b, re-
vealed the intermediacy of
a linear urea derivative (2) re-
sulting from isocyanate insertion
into the amidine within the first
point of analysis and presenting
a
distinctive downfield NH
Scheme 1. Alkaline earth-catalysed intramolecular hydroamination of aminoalkenes (A), hydroacetylation of
carbodiimides (B) and catalytic synthesis of imidazolidin-2-ones (C).
singlet at d1H =10.48 ppm
Chem. Eur. J. 2015, 21, 10548 – 10557
www.chemeurj.org
10549
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
(Scheme 2). Although the magnesium-catalysed cyclisation to
compound 2 took several days to reach completion at room
temperature (Table 1, entry 1), 3 could not be isolated by
methanol quenching of the catalyst because this also reversed
the formation of 3. The rate of cyclisation dramatically in-
creased with increasing metal cation size, in the order of Sr>
Ca@Mg (Table 1, entries 1–3). A minor dependence on metal-
cation identity was also observed for the Z/E isomer ratio of
the product, varying from approximately 1:2 for magnesium to
1:4 for strontium (Table 1, entries 1–3). Once full conversion to
3 was reached, the isomer ratio did not change upon heating
to 608C, suggesting no interconversion between the two iso-
mers (see below for a rationale for the independent formation
of each isomer). Heating to 808C, however, induced tautomeri-
sation of the benzylidene and isopropylimino moieties to the
corresponding 5-isopropylidenamino-4-benzyl-imidazole-2-one
(4) via a [1,5] sigmatropic H shift. Compound 4 displayed a dis-
Scheme 2.
1
Table 1. Isocyanate scope for the catalytic synthesis of (5-benzylidene-4-
imino)imidazolidin-2-ones from 1 by using precatalysts Ia–c in C6D6 at
room temperature.
tinctive H NMR benzyl singlet (2H) resonance around 3.3 ppm,
which correlated with a 13C{1H} resonance at d=103.2 ppm,
and two distinct isopropylidene methyl singlets (3H each) at
1.6 and 1.8 ppm, which were shown to be mutually coupled
by a COSY NMR experiment.
Entry
R
Product Cat. [mol%] t [h] NMR yield [%][a] Z/E [%]
1
2
3
4[b]
5[b]
6[b]
7[b]
8[b]
9[b]
tBu
tBu
tBu
tBu
Ad
Cy
iPr
nPr
Et
3
3
3
3
5
6
7
8
9
10
Ia 5.0
Ib 5.0
Ic 5.0
Ic 0.5
Ic 0.5
Ic 0.5
Ic 0.5
Ic 0.5
Ic 0.5
Ic 0.5
Ic 0.5
Ic 0.5
4d
2
0.1 >99
18
18
3
3
0.5
0.3
53
87
33:67
30:70
22:78
23:77
32:68
40:60
41:59
90:10
83:17
85:15
98:2
Encouraged by these results, the scope of the isocyanate
substrates that could be employed in this reaction was investi-
gated by using propargylamidine 1, with 0.5 mol% of Ic at
room temperature (Scheme 3). Although isocyanate insertion
to form the linear analogue of 2 proved virtually instantaneous
and independent of isocyanate identity, the rate of subsequent
cyclisation to the corresponding imidazolidin-2-one was
strongly influenced by the nature of the isocyanate substitu-
ent. Although the least sterically demanding substrate, ethyl-
isocyanate, gave a 91% conversion, as was determined by
1H NMR spectroscopy within 20 minutes at room temperature
(Table 1, entry 9), the bulkiest substrates, tert-butyl, adamantyl
and 2,6-diisopropylphenylisocyanate, required extended reac-
tion times to reach good conversions (Table 1, entries 4, 5 and
12). When the catalyst loading was increased to 5 mol%, how-
ever, all reactions proceeded to completion within less than
five minutes at room temperature. The Z/E isomer ratio of the
products was found to be governed by both steric and elec-
tronic factors. Although the various arylisocyanates primarily
gave the Z isomer (Table 1, entries 10–12), the selectivity of
isomer formation for reactions with alkylisocyanates was found
to be sensitive to the steric demands of the N-bound organic
substituent. Although isocyanates bearing primary ethyl and n-
propyl substituents provided predominant formation of the Z
isomer (Table 1, entries 8–9), this shifted to a preference for E
87
92
>99
98
97
91
10[b] Ph
0.5 >99
11[b] 2,4,6-Me3C6H2 11
12[b] 2,6-iPr2C6H3
12
6
18
98
78
90:10
[a] Determined by integration of benzylidene protons against an internal
HN(SiMe3)2 standard. [b] By using isolated compound 1.
Figure 1. Identification of (Z)-3 and (E)-3 by NOESY NMR spectroscopy.
Scheme 3.
Chem. Eur. J. 2015, 21, 10548 – 10557
www.chemeurj.org
10550
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
isomer formation for secondary and tertiary alkyl substituents
(Table 1, entries 1–7). The mesityl derivative 11, however, was
formed essentially as a single Z-isomer (Table 1, entries 11),
whereas a maximum E-selectivity of 78% was achieved for the
tert-butyl derivative 3 (Table 1, entries 3). It thus appears that,
for the alkyl derivatives at least, the orientation of the benzyli-
dene phenyl group depends on the relative steric pressure
imposed by the imino-isopropyl and isocyanate alkyl substitu-
ents. We tentatively suggest that this notable preference
toward Z-isomer formation for all aryl derivatives studied possi-
bly reflects a stabilising p interaction between the benzylidene
phenyl substituent and the N-aryl moieties during heterocycle
assembly at the metal centre.
was monitored by H NMR spectroscopy. The rate of insertion
and subsequent cyclisation was found to be highly dependent
on the steric demands of the amidine N-substituents
(Scheme 4). Although di(isopropyl)- and dicyclohexylcarbodi-
The reactions were easily translated to a preparative scale.
The amidine 1 was prepared on a 10 mmol scale by using
2.5 mol% of Ic overnight in hexanes at 808C, and crystallised
in 88% yield. The subsequent reaction of 1 with phenylisocya-
nate on a 0.88 mmol scale by using 2.5 mol% of Ic in 0.5 mL
of n-hexane was highly exothermic, and crystallisation of the
heterocyclic product 10 commenced almost instantaneously.
After one hour at room temperature, the reaction mixture was
filtered, and the colourless solid dried in vacuo to yield 10 in
near quantitative yield (285 mg, 93%). Single-crystal X-ray
diffraction experiments were performed on samples of
compounds 3 and 12, which were obtained from the NMR
scale reactions after quenching with methanol, filtration
and slow evaporation of the solvent at 48C. The results of
these experiments, which have been reported previously in
preliminary form,[34] are displayed in Figure 2 and show (E)-3
(left) and (Z)-12 (right), both of which correspond to the major
isomer observed by 1H NMR spectroscopy (Table 1, entries 3
and 12). In both cases, bond lengths and angles are within the
range expected for these (5-benzylidene-4-imino)-imidazolidin-
2-ones.[33]
Scheme 4.
imide-derived phenylpropargylamidines provided instantane-
ous and quantitative conversion to the respective heterocyclic
products 7 and 13 at room temperature (Table 2, entries 1 and
Table 2. Carbodiimide scope for the synthesis of (5-benzylidene-4-
imino)imidazolidin-2-ones from phenylacetylene and isopropylisocyanate
by using 5 mol% of Ic in C6D6 at room temperature.
Entry
R
R’
Product t [h] NMR yield [%][a] Z/E [%]
1
2
3
4
5
6
iPr
Cy
tBu
Et
iPr
Cy
tBu
tBu
7
0.1 98
0.1 97
41:59
25:75
25:75
34:66
13
14
16
40
75
0.1 99
[c]
Et
(CH2)3NMe2 17[b]
4
4
89
97
–
p-tol p-tol
18[b]
75:25
[a] Determined by integration of benzylidene protons against an internal
HN(SiMe3)2 standard. [b] RN=C=NR’ (2 equiv), 808C. Use of Ia instead of
Ic did not afford better selectivity for the formation of the desired prop-
argylamidine over the double carbodiimide insertion/cyclisation products
17 and 18. [c] Mixture of all four possible regioisomers.
The influence of amidine N-substitution on the reaction was
evaluated through the synthesis of a number of phenylpropar-
gylamidines from phenylacetylene and commercially available
carbodiimides by using 5 mol% of Ic in C6D6 at 608C. Once full
conversion to the amidine had been achieved, one molar
equivalent of isopropylisocyanate was added, and the reaction
2), the much more sterically hindered di(tert-butyl) derivative
required 40 h at room temperature to reach only 75% conver-
sion to heterocycle 14 (Table 2, entry 3). Although all three
symmetrical (N,N’-dialkyl)phenylpropargylamidines displayed
a bias towards the E-isomer, this was significantly more pro-
nounced for the dicyclohexyl and di(tert-butyl) substituted imi-
dazolidin-2-ones, compounds 13 and 14 (Table 2, entries 1–3).
Notably, the unsymmetrical [1-ethyl-3-(tert-butyl)]phenylpropar-
gylamidine precursor 15 led to exclusive isocyanate insertion
at the less hindered N-ethyl nitrogen atom, yielding a single
heterocyclic product, compound 16. The Z/E isomer ratio of
this latter species was intermediate between that of the di(iso-
propyl) and di(tert-butyl) derivatives (Table 2, entry 4). Carbodi-
imides with less sterically demanding N-substituents, such as
di(p-tolyl)carbodiimide and [1-(N,N’-dimethylaminopropyl)-3-
tert-butyl)carbodiimide, proved unsuitable for simple hydro-
acetylenation in the first step of the reaction. Rather, they
underwent twofold carbodiimide insertion followed by intra-
Figure 2. ORTEP representations of compounds (Z)-3 (left) and (E)-12 (right).
Ellipsoids are drawn at 30% probability. Hydrogen atoms are omitted for
clarity, except for the benzylidene proton H4.
Chem. Eur. J. 2015, 21, 10548 – 10557
www.chemeurj.org
10551
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
molecular hydroamidation/cyclisation to yield the N,N’-[(5-ben-
zylidene-imidazoldin-2,4-ylidene)diamine products 17 and 18
(Scheme 4, Table 2, entries 5 and 6). Sufficiently selective
access to the desired propargylamidines could also not be ach-
ieved through application of the less reactive magnesium pre-
catalyst, Ia, although in this case, double carbodiimide inser-
tion and the rate of subsequent intramolecular hydroamidation
were significantly perturbed. Z-Selectivity of 75% was ob-
served in the N,N’,N’’,N’’’-tetra-p-tolyl product (18) in line with
the Z-selectivity observed for arylisocyanate insertion/cyclisa-
tion (Table 2, entry 6). Unfortunately, unlike the unsymmetrical
carbodiimide-derived product 16, no regioselectivity was ob-
served when employing [1-(N,N’-dimethylaminopropyl)-3-ethyl]
carbodiimide presumably due to the lack of steric discrimina-
tion between the two carbodiimide substituents. In this case,
the formation of all four possible insertion regioisomers, as
well as E/Z-isomerism for each of them, was observed
(Scheme 4; Table 2, entry 5).
argylamidine remained instantaneous and quantitative in all
cases, the subsequent intramolecular cyclisation step proved
to be highly dependent on the substitution pattern of the ace-
tylenic aryl moiety. Ortho-substitution, with substrates, such as
o-tolylacetylene and 1-naphthylacetylene, resulted in signifi-
cantly reduced cyclisation rates compared to the parent phe-
nylacetylene (Table 3, entries 2 and 3). However, para-methyl-
Table 3. Arylacetylene scope for the synthesis of (5-benzylidene-4-imi-
no)imidazolidin-2-ones from di(isopropyl)carbodiimide and isopropyliso-
cyanate by using 5 mol% of Ic in C6D6 at room temperature.
Entry Ar
Product T [8C] t [h] NMR yield [%][a] Z/E [%]
1
2
3
4
5
6
7
8
Ph
o-tol
1-naphthyl
p-tol
3-F-C6H4
4-Cl-C6H4
4-OMe-C6H4
4-CF3-C6H4
4-CF3-C6H4
4-F-C6H4
7
25
25
25
25
25
25
25
25
60
25
25
25
0.1
8
8
0.1 >99
1.5
1
5
24
2
24
98
93
91
41:59
40:60
42:58
42:58
32:68
36:64
55:45
20:80
20:80
40:60
38:62
63:37
21
22
23
24
25
26
27
27
28
98
96
90
89
>99
90
The influence of acetylenic substitution was also
investigated. The strontium-catalysed reaction of isopropyl or
phenylisocyanate with (N,N’-diisopropyl)-n-butylpropargyl-
amidine did not provide the desired heterocycles. Rather, the
reactions stalled after formation of the linear urea molecules,
19 and 20, resulting from isocyanate insertion (Scheme 5). The
latter species did not undergo intramolecular hydroamidation/
9
10
11
12
3-pyridinyl[b] 29
3-thiophenyl 30
0.1 >99
0.1 >99
[a] Determined by integration of benzylidene protons against an internal
HN(SiMe3)2 standard. [b] 3-Pyridinylpropargylamidine was synthesised by
using 5 mol% of Ia at 408C for one day, because Ic resulted in double in-
sertion of the carbodiimide and formation of the imidazolidin-2,4-ylidene
product analogous to 29.
substitution had no effect on the efficacy of the intramolecular
hydroamidation step (Table 3, entry 4). It is also notable that
the Z/E isomer ratio was unaffected by alkyl or aryl substitution
on the phenyl ring (Table 3, entries 1–4). For para-substituents,
a clear trend of decreasing rate of cyclisation with increasing
substituent electron-withdrawing effect was observed; from
one hour to achieve 96% conversion to the para-chloro deriva-
tive, 25, to 24 h to reach 90% conversion for the para-trifluoro-
methyl and para-fluoro derivatives, 27 and 28 (Table 3, en-
tries 6–10). This trend contrasts with our previous observations,
in which the rate of cyclisation during the intramolecular hy-
droalkoxylation of hydroxyalkynes bearing para-substituted
terminal aryl groups was seen to increase with increasingly
electron-withdrawing substitution in the para position.[16e] We
tentatively ascribe the apparent reversal in reactivity in the
present instance to a more remote effect of stronger electron-
withdrawing para-substitution on the polarisation of the alka-
Scheme 5.
cyclisation even after prolonged heating at 1008C. Similar ob-
servations have been made in the intramolecular hydroamina-
tion of aminoalkenes, in which cyclisation is hindered by the
presence of terminal alkyl substituents on the alkene moiety,
whereas terminal aryl substitution promotes cyclisation due to
an activating electronic effect.[16,17]
In contrast, aryl or heteroaryl substitution on the acetylenic
fragment afforded clean and near-quantitative formation of
the desired imidazolidin-2-ones in stepwise reactions with
di(isopropyl)carbodiimide and isopropylisocyanate by using
5 mol% of Ic (Scheme 6). Although under these conditions, in-
termolecular hydroamidation of the isocyanate with the prop-
Scheme 6.
Chem. Eur. J. 2015, 21, 10548 – 10557
www.chemeurj.org
10552
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
line earth amidinate nitrogen centres and consequent less
facile isocyanate insertion. In mitigation of this hypothesis,
monitoring of these reactions by 1H NMR spectroscopy
revealed that isocyanate insertion into the more electron-poor
amidines is less rapid and provides lower conversion than with
the more electron-rich amidines. In contrast, meta-fluoro-sub-
stitution resulted in only a small decrease in reactivity, yielding
98% of the heterocyclic product, 24, in 1.5 h (Table 3, entry 5).
The nature of the electron-withdrawing substituent also
influences the Z/E selectivity, although no discernible pattern
could be discriminated.
pyl)carbodiimide did provide the desired amidine, allowing
subsequent isocyanate insertion and cyclisation. In this case,
however, three different products were obtained: the expected
Z and E isomers of heterocycle 30, present in a 63:37 ratio, as
well as a new compound, which was identified as the Diels–
Alder adduct of two molecules of this heterocyclic product.
A stoichiometric reaction between Ia, two equivalents of
(N,N’-diisopropyl)phenylpropargylamidine and two equivalents
of 2,6-diisopropy phenylisocyanate did not yield the expected
magnesium insertion complex. Analysis by NMR spectroscopy
instead indicated complete consumption of the substrates to
form the corresponding N-heterocyclic product 3 with reforma-
tion of Ia, suggesting that cyclisation requires the presence of
[HN(SiMe3)2] liberated upon protonolysis of Ia with the
amidine.
It is notable that these reactions were tolerant of chloro sub-
stitution on the phenyl ring, providing the potential for further
catalytic functionalisation of the aryl moiety. After quenching
and extraction with methanol, compound 25 crystallised in
good yield (85%) upon solvent evaporation at room
temperature. Figure 3 shows the results of a single-crystal X-
ray diffraction experiment on the major (E)-isomer. Heteroaryl-
In contrast,
a further stoichiometric reaction utilising
[Mg(CH2Ph)2(THF)2] in place of Ia, and with consequent libera-
tion of the aprotic toluene conjugate acid, gave the desired
homoleptic insertion complex II in quantitative yield
(Scheme 7). An X-ray diffraction experiment, preliminary details
of which were included in our previous communication,[34]
revealed II to be a distorted square pyramidal magnesium 2-
(propargylamidino)imidate complex, with a THF molecule coor-
dinated in the apical position. Coordination in the basal plane
is provided by the oxygen atoms arising from the inserted iso-
cyanate substrates and the imino nitrogen of the amidinate
moieties to form a quasi-planar 6-membered [MgNCNCO] met-
allacycle (Figure 4). The rather short CÀO bond lengths
[1.275(4), 1.274(4) Š], lengthened C1ÀN1 [1.451(4) Š] and C29À
N4 bonds [1.452(4) Š] and planarity of the N1 and N4 nitrogen
atoms within II suggest some degree of delocalisation over the
chelate ring. The short C1ÀN3 [1.289(4) Š] and C29ÀN6
[1.291(4) Š] bond lengths are clearly indicative of pendant
imine functionalities. Isolated samples of complex II also gave
similar catalytic activity to Ia for the formation of 3, suggesting
it is a catalytic intermediate.
Figure 3. ORTEP representation of compound (E)-25. Ellipsoids at 30%
probability. Hydrogen atoms omitted for clarity except for the benzylidene
proton H4.
1
A variable-temperature H NMR experiment performed on II
in [D8]toluene provided evidence for isocyanate de-insertion at
higher temperatures (Scheme 8). A van’t Hoff analysis of this
equilibrium provided DH° = +88 kJmolÀ1 and DS° =
208 JKÀ1 molÀ1, giving a DG°(298 K) value of +26 kJmolÀ1 for
the de-insertion process. Although a definitive interpretation
of this latter value would require deconvolution of both the
de-insertion and potential dimerisation of the resultant bis(pro-
pargylamidinate) species, III, this positive but low free energy
of activation at 298 K indicates that the potential for this
reversibility is likely to be significant during the course of the
catalysis at ambient or slightly elevated temperatures.
substituted heterocycles could also be obtained in this
manner. N,N’-Diisopropyl-3-pyridinylpropargylamidine was syn-
thesised by using the magnesium precatalyst Ia in preference
to the strontium precursor Ic, which was found to provide
twofold carbodiimide insertion and cyclisation. Under these
conditions, insertion of isopropylisocyanate and cyclisation
were very rapid and effectively quantitative to give the pyri-
dine-substituted heterocycle 29 in excellent yield (Table 3,
entry 11). Attempts to synthesise the 2-pyridinyl analogue,
however, were less discriminating, because even the use of Ia
resulted in competitive carbodii-
mide insertion, with MS (ESI)
analysis detecting molecular
weights indicative of 1:1, 1:2,
2:1, 2:2 and even 3:2 acetylene/
carbodiimide ratios. The stronti-
um-catalysed reaction of 3-thio-
phenylacetylene with di(isopro- Scheme 7.
Chem. Eur. J. 2015, 21, 10548 – 10557
www.chemeurj.org
10553
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 4. ORTEP representation of compound II. Ellipsoids are drawn at 30%
probability. Hydrogen atoms are omitted for clarity.
Based on the structure of II and the selectivity of the reac-
tions toward the formation of the kinetic imidazolidin-2-one
products rather than the thermodynamic 2-iminooxazolidines,
we propose that the N,O-chelate A shown in Figure 5 first iso-
merises to the corresponding N,N-chelate B by decoordination
of the oxygen atom and rotation of the imidate imino residue
to coordinate to the metal centre. The metal-bound nitrogen
atom of the amidino fragment must then necessarily decoordi-
nate and rotate to allow the alkyne to interact with the metal
centre. This may result in either a Cchair or a Cboat conformation,
giving rise upon proton-assisted insertion/cyclisation to the
heterocyclic products (E)-D and (Z)-D, respectively. Taking into
account all these observations, the catalytic cycle presented in
Figure 6 may be envisaged.
Figure 5. Proposed mechanism for the independent formation of E- and Z-
heterocycles.
the resonances. The Z/E isomer distribution of the products,
1
however, quantified by integration of the H NMR singlets of
the benzylidene protons, followed similar patterns to that of
the imidazolidin-2-ones. Aryl substituents and the less sterically
hindering ethyl functionality predominantly yielded the Z
isomer (Table 4, entries 1–3), whereas the more sterically de-
manding isopropyl and cyclohexyl derivatives yielded the E
isomer as the major product (Table 4, entries 4 and 5). The
steric influence of the isothiocyanate alkyl substituent ap-
peared more pronounced, however, than for the correspond-
ing isocyanates, presumably due to the greater steric repulsion
of the exocyclic sulfur versus oxygen atom. In contrast to the
isocyanate-based reactions, no observable isothiocyanate inser-
tion was evident even at high temperature for the larger tert-
butyl and adamantyl derivatives (Table 4, entries 6 and 7). The
isolated imidazolidin-2-thiones proved to be very moisture-sen-
sitive and had to be isolated in a glovebox. An X-ray diffraction
analysis experiment performed on single crystals of the 4-tert-
butylphenyl derivative gave the structure of the predominant
Z-isomer, (Z)-32 (Figure 7) clearly revealing the maintenance of
the thione functionality [C=S 1.654(2) Š].[34]
The insertion/cyclisation reactivity was further extended to
a variety of isothiocyanates, yielding the corresponding imida-
zolidin-2-thiones, compounds 31–35 (Scheme 9). All such com-
pounds presented a characteristic 13C NMR thione resonance
1
around 180 ppm, as well as two distinct H NMR benzylidene
singlets at approximately 6.5 and 5.9 ppm corresponding to
the Z and E isomers, respectively. Reactions using 0.5 mol% of
Ia were significantly slower than for the corresponding isocya-
nates. However, quantitative conversion to the imidazolidin-2-
thiones products was achieved within 30 minutes at room
temperature with 5 mol% of Ic.
Unlike the isocyanate-based reactions, the linear isothio-
cyanate insertion intermediates analogous to 2 were not
Insight into the possible course of reaction during the
synthesis of compounds 31–35 was provided by a reaction
between the homoleptic calci-
1
distinguishable by H NMR spectroscopy due to broadening of
um (N,N’-diisopropyl)phenylpro-
pargylamidinate dimer, IV,[35]
and two molar equivalents of
(para-tert-butyl)phenylisothio-
cyanate in toluene. This process
provided clean access to the
heteroleptic calcium compound
V (Scheme 10), a single-crystal
Scheme 8.
X-ray crystallographic analysis of
Chem. Eur. J. 2015, 21, 10548 – 10557
www.chemeurj.org
10554
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
imidazolidine products bearing
an exocyclic thione moiety are
formed in kinetic preference to
the more thermodynamically
stable thiazole heterocycles.
Although no evidence for the
formation of the alternative thia-
zolidine isomers was observed
during the synthesis of com-
pounds 31–35, a further exten-
sion of this protocol demon-
strated that the thiazolidine ring
system was accessible through
the strontium-catalysed reaction
of carbon disulfide with (N,N’-
diisopropyl)phenylpropargyl-
Figure 6. Mechanism for the alkaline earth-catalysed formation of imidazolidin-2-ones from arylpropargylamidines
and isocyanates.
Scheme 9.
Table 4. Isothiocyanate scope for the catalytic synthesis of (5-benzyli-
dene-4-imino)imidazolidin-2-thiones from 1 by using 5 mol% of Ic in
C6D6.
Entry
R
Product
T [8C]
Z/E [%]
1
2
3
4
5
6
7
Ph
31
32
33
34
35
–
25
25
25
25
25
80
80
97:3
95:5
88:12
5:95
15:85
–
4-tBu-C6H4
Et
iPr
Cy
tBu
Ad
Figure 7. ORTEP representation of compound (Z)-32. Ellipsoids are drawn at
30% probability. Hydrogen atoms are omitted for clarity, except for that on
the benzylidene moiety.
–
–
which revealed it to be a dimeric calcium species bearing ter-
minal amidinate ligands and bridging (2-propargylamidino)imi-
dothioate ligands (Figure 8). The latter units form a six-mem-
bered N,S-chelate, with the sulfur atom bridging between both
calcium centres, and coordination to the second alkaline earth
centre through the pendant N-
amidine. At 608C, CS2 insertion followed by intramolecular hy-
drothiolation, proceeded smoothly to give the (4-benzylidene-
5-imino)thiazolidin-2-thione product, 36, which over a period
of 48 h fully tautomerised to the corresponding 5-benzyl-1,3-
thiazole-2-thione (37; Scheme 11).
(4-tert-butylphenyl)imino nitro-
gen.
The
sulfur-containing
ligand is effectively analogous
to the isocyanate-derived mag-
nesium chelate within the ho-
moleptic species, compound II.
Thus, we suggest the formation
of the catalytic production of
the imidazolidin-2-thione deriva-
tives also follows a similar mech-
anistic pathway, in which the
Scheme 10.
Chem. Eur. J. 2015, 21, 10548 – 10557
www.chemeurj.org
10555
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Keywords: alkaline earths metals
cycloaddition · synthetic methods
·
cascade reactions ·
[1] J. A. Joule, K. Mills, Heterocycles in Medicine, in Heterocyclic Chemistry
5th ed., Wiley-Blackwell, Hoboken, NJ, 2010.
[2] M. Meusel, M. Gutschow, Org. Prep. Proc. Int. 2004, 36, 391–443.
[3] K. K. Borowicz, M. Banach, Pharmacol. Rep. 2014, 66, 545–551.
[4] J. M. Chezal, G. Delmas, S. Mavel, H. Elakmaoui, J. Metin, A. Diez, Y.
Blache, A. Gueiffier, M. Rubiralta, J. C. Teulade, O. Chavignon, J. Org.
Chem. 1997, 62, 4085–4087.
[5] F. Nique, S. Hebbe, N. Triballeau, C. Peixoto, J.-M. Lefrancois, H. Jary, L.
Alvey, M. Manioc, C. Housseman, H. Klaassen, K. Van Beeck, D. Guedin, F.
Namour, D. Minet, E. Van der Aar, J. Feyen, S. Fletcher, R. Blanque, C.
Robin-Jagerschmidt, P. Deprez, J. Med. Chem. 2012, 55, 8236–8247.
[6] T. Krause, M. U. Gerbershagen, M. Fiege, R. Weisshorn, F. Wappler, Anaes-
thesia 2004, 59, 364–373.
[7] M. Shiozaki, Carbohydr. Res. 2002, 337, 2077–2088.
[8] S. G. Burton, R. A. Dorrington, Tetrahedron: Asymmetry 2004, 15, 2737–
2741.
[9] B. Zhao, H. Du, Y. Shi, J. Am. Chem. Soc. 2008, 130, 7220–7221.
[10] M. Beller, M. Eckert, W. A. Moradi, H. Neumann, Angew. Chem. Int. Ed.
1999, 38, 1454–1457; Angew. Chem. 1999, 111, 1562–1565.
[11] T. Miura, Y. Mikano, M. Murakami, Org. Lett. 2011, 13, 3560–3563.
[12] Y. Ohshiro, K. Kinugasa, T. Minami, T. Agawa, J. Org. Chem. 1970, 35,
2136–2140.
Figure 8. ORTEP representation of compound V. Ellipsoids are drawn at 30%
probability. Hydrogen atoms and isopropyl methyl groups are omitted for
clarity.
[13] G. Süss-Fink, G. F. Schmidt, G. Herrmann, Chem. Ber. 1987, 120, 1451–
1453.
[14] Y. Kuninobu, K. Kikuchi, K. Takai,
Chem. Lett. 2008, 37, 740–741.
[15] a) A. G. M. Barrett, M. R. Crimmin,
M. S. Hill, P. A. Procopiou, Proc. R.
Soc. London Ser. A 2010, 466, 927–
963; b) S. Harder, Chem. Rev. 2010,
110, 3852–3876; c) M. Arrowsmith,
M. S. Hill, Alkaline earth Chemistry:
Applications in Catalysis in Compre-
Scheme 11.
hensive Inorganic Chemistry II, Vol. 1
(Ed.: T. Chivers), Elsevier, Amster-
dam,
2013,
pp. 1189–1216;
d) M. R. Crimmin, M. S. Hill, Alkaline Earth Metal Compounds: Oddities
and Applications, Topics in Organometallic Chemistry, Vol. 45, Springer-
Verlag, Berlin Heidelberg, 2013, pp. 191–241.
Conclusion
[16] a) M. Crimmin, I. Casely, M. Hill, J. Am. Chem. Soc. 2005, 127, 2042–
2043; b) J. Spielmann, F. Buch, S. Harder, Angew. Chem. Int. Ed. 2008, 47,
9434–9438; Angew. Chem. 2008, 120, 9576–9580; c) J. Spielmann, S.
Harder, Eur. J. Inorg. Chem. 2008, 1480–1486; d) F. Buch, J. Brettar, S.
Harder, Angew. Chem. Int. Ed. 2006, 45, 2741–2745; Angew. Chem.
2006, 118, 2807–2811; e) C. Brinkmann, A. G. M. Barrett, M. S. Hill, P. A.
Procopiou, S. Reidt, Organometallics 2012, 31, 7287–7297.
[17] a) M. R. Crimmin, M. Arrowsmith, A. G. M. Barrett, I. J. Casely, M. S. Hill,
P. A. Procopiou, J. Am. Chem. Soc. 2009, 131, 9670–9685; b) S. Datta,
P. W. Roesky, S. Blechert, Organometallics 2007, 26, 4392–4394; c) S.
Datta, M. T. Gamer, P. W. Roesky, Organometallics 2008, 27, 1207–1213;
d) F. Buch, S. Harder, Z. Naturforsch B J. Chem. Sci. 2008, 63, 169–177;
e) P. Horrillo-Martínez, K. C. Hultzsch, Tetrahedron Lett. 2009, 50, 2054–
2056.
[18] B. M. Day, W. Knowelden, M. P. Coles, Dalton Trans. 2012, 41, 10930–
10933.
[19] J. F. Dunne, S. R. Neal, J. Engelkemier, A. Ellern, A. D. Sadow, J. Am.
Chem. Soc. 2011, 133, 16782–16785.
[20] V. Leich, T. P. Spaniol, L. Maron, J. Okuda, Chem. Commun. 2014, 50,
2311–2314.
[21] B. Liu, J.-F. Carpentier, Y. Sarazin, Chem. Eur. J. 2012, 18, 13259–13264.
[22] B. Liu, T. Roisnel, J.-F. Carpentier, Y. Sarazin, Angew. Chem. Int. Ed. 2012,
51, 4943–4946; Angew. Chem. 2012, 124, 5027–5030.
We have demonstrated the applicability of readily available
and sustainable alkaline earth bis(amide) precatalysts to the
facile one-pot, 100% atom-efficient, stepwise synthesis of
a wide variety of highly functionalised imidazolidin-2-ones,
imidazolidin-2-thiones, N,N’-[(5-benzylidene-imidazoldin-2,4-yli-
dene)diamines and 1,3-thiazolidin-2-thiones from the simplest
commercially available building blocks. Although the current
investigation is relatively broad in scope, we suggest that the
formation of compound 37 indicates that even broader mani-
festations of this protocol may be achieved through the incor-
poration of further alternative heterocumulene substrates. Cat-
alytic access to a variety of valuable heterocyclic derivatives
may thus be readily achieved in a cost-efficient manner and
with complete atom efficiency. We are continuing to explore
these possibilities and will describe our investigations in future
publications.
[23] B. Liu, T. Roisnel, J.-F. Carpentier, Y. Sarazin, Chem. Eur. J. 2013, 19,
13445–13462.
Acknowledgements
[24] B. Liu, T. Roisnel, J.-F. Carpentier, Y. Sarazin, Chem. Eur. J. 2013, 19,
2784–2802.
We thank the EPSRC (UK) for generous funding.
Chem. Eur. J. 2015, 21, 10548 – 10557
www.chemeurj.org
10556
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[25] D. Mukherjee, A. Ellern, A. D. Sadow, Chem. Sci. 2014, 5, 959–964.
[26] S. R. Neal, A. Ellern, A. D. Sadow, J. Organomet. Chem. 2011, 696, 228–
234.
[27] T. D. Nixon, B. D. Ward, Chem. Commun. 2012, 48, 11790–11792.
[28] N. Romero, S.-C. Rosca, Y. Sarazin, J.-F. Carpentier, L. Vendier, S. Mallet-
Ladeira, C. Dinoi, M. Etienne, Chem. Eur. J. 2015, 21, 4115–4125.
[29] R. J. Schwamm, M. P. Coles, Organometallics 2013, 32, 5277–5280.
[30] R. J. Schwamm, B. M. Day, N. E. Mansfield, W. Knowelden, P. B. Hitchcock,
M. P. Coles, Dalton Trans. 2014, 43, 14302–14314.
[33] M. S. Hill, D. J. Liptrot, M. F. Mahon, Angew. Chem. Int. Ed. 2013, 52,
5364–5367; Angew. Chem. 2013, 125, 5472–5475.
[34] M. Arrowsmith, W. M. S. Shepherd, M. S. Hill, G. Kociok-Kçhn, Chem.
Commun. 2014, 50, 12676–12679.
[35] M. Arrowsmith, M. R. Crimmin, M. S. Hill, S. L. Lomas, M. S. Heng, P. B.
Hitchcock, G. Kociok-Kçhn, Dalton Trans. 2014, 43, 14249–14256.
Received: April 4, 2015
[31] J. S. Wixey, B. D. Ward, Dalton Trans. 2011, 40, 7693–7696.
[32] J. S. Wixey, B. D. Ward, Chem. Commun. 2011, 47, 5449–5451.
Published online on June 11, 2015
Chem. Eur. J. 2015, 21, 10548 – 10557
www.chemeurj.org
10557
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim