Structural and Mechanistic Insights into s-Block Bimetallic Catalysis: Sodium Magnesiate-Catalyzed Guanylation of Amines

Article abstract of DOI:10.1002/chem.201602906

To advance the catalytic applications of s-block mixed-metal complexes, sodium magnesiate [NaMg(CH₂SiMe₃)₃] (1) is reported as an efficient precatalyst for the guanylation of a variety of anilines and secondary amines with

Full text of DOI:10.1002/chem.201602906

DOI: 10.1002/chem.201602906
Full Paper
&
Cooperative Effects
Structural and Mechanistic Insights into s-Block Bimetallic
Catalysis: Sodium Magnesiate-Catalyzed Guanylation of Amines
Marco De Tullio,[a] Alberto Hernꢀn-Gꢁmez,[a] Zoe Livingstone,[a] William Clegg,[b]
Alan R. Kennedy,[a] Ross W. Harrington,[b] Antonio AntiÇolo,[c] Antonio Martꢂnez,[c]
Fernando Carrillo-Hermosilla,*[c] and Eva Hevia*[a]
In memory of Roberto Sanchez Delgado
Abstract: To advance the catalytic applications of s-block
mixed-metal complexes, sodium magnesiate
netic studies imply that these guanylation reactions occur
via (tris)amide intermediates that react with carbodiiimides
in insertion steps. The rate law for the guanylation of N,N’-
diisopropylcarbodiimide with 4-tert-butylaniline catalyzed by
1 is first order with respect to [amine], [carbodiimide], and
[catalyst], and the reaction shows a large kinetic isotopic
effect, which is consistent with an amine-assisted rate-deter-
mining carbodiimide insertion transition state. Studies to
assess the effect of sodium in these transformations denote
a secondary role with little involvement in the catalytic
cycle.
[NaMg(CH2SiMe3)3] (1) is reported as an efficient precatalyst
for the guanylation of a variety of anilines and secondary
amines with carbodiimides. First examples of hydrophosphi-
nation of carbodiimides by using a Mg catalyst are also de-
scribed. The catalytic ability of the mixed-metal system is
much greater than that of its homometallic components
[NaCH2SiMe3] and [Mg(CH2SiMe3)2]. Stoichiometric studies
suggest that magnesiate amido and guanidinate complexes
are intermediates in these catalytic routes. Reactivity and ki-
capabilities.[4] In contrast, success with Mg has been more limit-
ed because its smaller radius raises transition-state barriers in
rate-determining alkene insertion steps, which shows similar
patterns to those previously observed for organolanthanide(III)
catalysts.[5] To overturn this trend, we have shown recently that
Mg activated within a sodium magnesiate platform can outper-
form Ca and Ba systems in the hydroamination of isocyanates,
secure higher yields, and show superior substrate scope under
milder conditions.[6] Cooperative effects between the two
metals underpin this catalytic transformation,[7] in which Lewis
acidic Na anchors and activates the isocyanate to enable intra-
molecular attack by the highly nucleophilic Lewis basic tris(am-
ido) magnesiate, which facilitates a synergistic scenario that is
not available in the aforementioned single-metal systems.
To build on these initial findings and on previous studies on
alkali-metal magnesiate chemistry, which have already demon-
strated the unique synergistic properties of these bimetallic
systems (e.g., enhanced reactivity, special regioselectivity, ex-
cellent functional-group tolerance) in several cornerstone stoi-
chiometric transformations,[8] here we assess the ability of
sodium magnesiates to catalyze the guanylation of amines
with carbodiimides (Scheme 1).
Introduction
Over the past decade, alkaline-earth metal catalysis has started
to gain prominence and find applications for the heterofunc-
tionalization of a wide range of unsaturated organic frag-
ments.[1] Seminal contributions from the groups of Hill[2] and
Harder,[3] among others, have added Group 2 metal complexes
to the homogeneous catalytic landscape as low-toxicity, low-
cost alternatives to transition-metal systems. Most initial appli-
cations have involved the hydroamination of unsaturated or-
ganic substrates, such as alkenes and alkynes, in which heavier
Ca or Sr complexes have demonstrated remarkable catalytic
[a] M. De Tullio, Dr. A. Hernꢀn-Gꢁmez, Dr. Z. Livingstone, Dr. A. R. Kennedy,
Prof. E. Hevia
WestCHEM, Department of Pure and Applied Chemistry
University of Strathclyde
295 Cathedral Street, Glasgow, G1 1XL (UK)
E-mail: eva.hevia@strath.ac.uk
[b] Prof. W. Clegg, Dr. R. W. Harrington
School of Chemistry
Newcastle University
Newcastle upon Tyne, NE1 7RU (UK)
[c] Prof. A. AntiÇolo, A. Martꢂnez, Dr. F. Carrillo-Hermosilla
Centro de Innovaciꢁn en Quꢂmica Avanzada (ORFEO-CINQA) Departamento
de Quꢂmica Inorgꢀnica, Orgꢀnica y Bioquꢂmica, Facultad de Ciencias y Tec-
nologꢂas Quꢂmicas, Universidad de Castilla-La Mancha Campus Universitar-
io, 13071 Ciudad Real (Spain)
The synthesis of guanidines has received considerable atten-
tion[9] because these simple nitrogen-containing molecules are
valuable building blocks present in numerous natural products
and pharmaceuticals.[10] Furthermore, they find extensive appli-
cations as precursors of ancillary ligands for numerous transi-
tion-, lanthanoid-, and main-group-metal complexes[11] and
E-mail: Fernando.Carrillo@uclm.es
Supporting information and ORCID(s) from the author(s) for this article is
available on the WWW under http://dx.doi.org/10.1002/chem.201602906.
Chem. Eur. J. 2016, 22, 1 – 12
1
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Scheme 1. Guanylation of primary amines with carbodiimides.
they can also be employed as organocatalysts.[12] Atom-eco-
nomical catalytic addition of amines to carbodiimides (guanyla-
tion reaction, Scheme 1) constitutes one of the most straight-
forward routes to access N-substituted guanidines.[9]
Scheme 2. Catalytic guanylation and hydrophosphination reactions.
Table 1. Optimization of the reaction conditions.
Although certain guanylations can be accomplished catalyst-
free, these processes have high kinetic barriers that require the
use of harsh reaction conditions, which restricts their applica-
tion to activated primary aliphatic amines.[13] Thus, metal catal-
ysis is required when using anilines or secondary amines and
high temperatures are needed even so, with only a select few
catalytic systems able to facilitate these processes at room
temperature.[9]
Entry
Catalyst [2 mol%]
[NaCH2SiMe3]
Solvent
[D8]THF
[D8]THF
Time [h]
Yield [%][a]
The vast majority of these studies have focused on transi-
tion-metal and rare-earth-metal catalysis.[14] Notwithstanding,
some recent studies of lithium[15,16a] or magnesium (and heavi-
er Group 2 elements)[16,17] have already demonstrated the po-
tential of s-block metal complexes to catalyze these reactions.
Related studies that investigated the synthesis of phosphagua-
nidines have revealed the ability of heavier alkaline-earth-
metal amides to catalyze the direct addition of secondary
phosphines to carbodiimides.[18]
0.25
1
0.25
16
0.25
3
72
84
44
99
99
90
1
2
[Mg(CH2SiMe3)2]
3
4
[NaMg(CH2SiMe3)3] (1)
[NaMg(CH2SiMe3)3] (1)
[D8]THF
C6D6
[a] Yields obtained by using spectroscopic 1H NMR integration of signals
for the guanidine product 4h, with addition of ferrocene (10 mol%) as
the internal standard. Reaction conditions: solvent (0.5 mL), 2,6-dimethyl-
aniline (0.55 mmol), DIC (0.5 mmol), and catalyst (0.01 mmol).
With the aim of expanding the scope of s-block cooperative
catalysis, here we report the first catalytic applications of alkali-
metal magnesiates for the synthesis of guanidines. By combin-
ing kinetic experiments with stoichiometric reactivity studies,
we provide informative mechanistic insights into these new
ate-catalyzed transformations.
Mg(CH2SiMe3)2 being significantly less efficient (44% conver-
sion) than the more polar, more reactive NaCH2SiMe3 (72%
conversion; Table 1). This notable influence of the metals con-
trasts with recent studies by Harder et al., who used a naked
{NPh2}ꢀ anion as an organocatalyst.[3a]
Results and Discussion
Subsequently, the catalytic activity of 1 was investigated for
a range of amines and carbodiimides (Table 2, see also the Ex-
perimental Section). Aniline (2a) reacted with DIC (3a), N,N’-di-
cyclohexylcarbodiimide (DCC; 3b), and EtNCNtBu (3c), to give
guanidines 4a–c in high yields (80–96%; Table 2, entries 1–3).
It is worth noting that precatalyst 1 was compatible with both
electron-donating and -withdrawing substituents on the
phenyl ring of the amine, such as Me-, tert-butyl-, MeO-, or Cl-
(Table 2, entries 4–7), and gave corresponding substituted gua-
nidines 4d–g in excellent yields (80–98%). Furthermore, 1 also
effectively facilitated the room-temperature addition of hin-
dered anilines with substituents at their ortho-positions (2 f–g)
or even of the secondary aniline N-methylaniline (2h; 86–94%
yield, Table 2, entries 8, 9, and 10) and low activated diphenyl-
amine (2i; 73% yield, Table 2, entry 11). Interestingly, and de-
spite the presence of a pyridyl substituent, which could poten-
tially coordinate to the bimetallic intermediates involved in
this process and inhibit their catalytic activity, the reaction of
3-aminopyridine (2j) with DIC gave guanidine 4l in an 88%
yield (Table 2, entry 12). This versatility and functional-group
tolerance are remarkable when compared with other s-block
Catalytic synthesis of guanidines and phosphaguanidines
We began our studies by testing the efficacy of homoleptic
sodium magnesiate [NaMg(CH2SiMe3)3] (1)[19] in the intermolec-
ular hydroamination reaction of different carbodiimides with
a variety of aromatic, aliphatic, and secondary cyclic amines
(guanylation process). In addition, we tested compound 1 in
the hydrophosphination reaction of the same carbodiimide
substrates with the secondary phosphine Ph2PH (Scheme 2).
First we studied, as a model reaction, the guanylation of 2,6-
dimethylaniline 2 f with N,N’-diisopropylcarbodiimide (DIC, 3a;
Table 1) in C6D6 with 2 mol% of 1. At room temperature, the
reaction gave corresponding guanidine 4h in 90% yield after
3 h. An important solvent effect was noted and when a more
polar ethereal solvent with a greater coordination ability
([D8]THF) was employed, guanidine 4h was obtained in 99%
yield after just 15 min. In contrast, to illustrate the synergic re-
activity of 1, when its single-metal components were tested as
catalysts under the same reaction conditions, lower conver-
sions for guanidine 4h were observed after 15 min, with
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
2
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Table 2. Guanylation and hydrophosphination of carbodiimides.[a]
Table 2. (Continued)
Entry
1
Amine/phosphine
Carbodiimide
Compound/Yield [%][b]
Entry
Amine/phosphine
Carbodiimide
Compound/Yield [%][b]
17
3a
3b
3c
4a (96)
4b (90)
4c (80)
3a
3b
3c
4q (95)[e]
4r (90)[e]
4s (80)
2a
2
3
2o
18
19
4
5
6
7
2b
2c
2d
2e
3a
3a
3a
3a
4d (90)
4e (80)
4 f (98)
4g (96)
[a] Conditions: amine/phosphine (1.00 mmol), carbodiimide (1.00 mmol),
catalyst 1 (2% mol), THF (3 mL), 1 h, 258C. [b] Isolated yields. [c] 1 h,
708C. [d] 24 h, 708C. [e] 30 min, 258C.
catalytic systems in which anilines with large substituents or
coordinating groups give lower yields than nonsubstituted
substrates.[9c,15] Cyclic amines morpholine and piperidine, and
n-butylamine required the use of forcing reaction conditions,
higher temperatures (708C), or a longer reaction time (24 h) to
give guanidines 4m–o in moderate yields (52–65%, Table 2,
entries 13–15). In contrast, no reaction was observed when dii-
sopropylamine (2n) was used, which can be rationalized in
terms of the significant increase in steric bulk in this amine
and its relatively low acidity compared with the other sub-
strates studied (Table 2, entry 16). Although previous studies
have shown the feasibility of homometallic magnesium com-
plexes to catalyze guanylation processes by using unhindered
amines,[16a,b] 1 offers a significant improvement for secondary
amines and substituted anilines,[16a,c] and enables these pro-
cesses to take place at room temperature over short periods of
time. Interestingly, the hydrophosphination of carbodiimides
3a–c with diphenylphosphine (2o) could also be achieved at
room temperature by using catalyst loadings as low as
2 mol%, which afforded phosphoguanidines 4q–s in high
yields (80–95%, Table 2, entries 17–19). To the best of our
knowledge, this represents the first example of a magnesium
complex catalyzing this process, and shows an activity compa-
rable to results with heavier alkaline-earth-metal amides re-
ported by Hill et al., in which efficiency of the catalyst corre-
lates directly with the increase in size of the metal cation.[18]
8
2 f
3a
4h (90)
9
2g
2h
3a
3a
4i (86)
4j (94)
10
11
2i
2j
3b
3a
4k (73)
4l (88)
12
13
2k
3a
4m (15); (52)[c]
14
15
16
2l
3a
3a
4n (19); (65)[c]
4o (0)[c]; (60)[d]
Stoichiometric studies
2m
To gain mechanistic insights into these promising catalytic pro-
cesses, a series of stoichiometric reactions were carried out.
Addition of three molar equivalents of NH2Ar (Ar=2,6-
Me2C6H3; 2 f) to tris(alkyl)magnesiate 1 afforded colorless crys-
2n
3a
4p (0); (0)[d]
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
3
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
lic sodium anilide [{(PMDETA)NaNHPh}2] (mean value
2.42(3) ꢄ).[22] The structure of 5 contrasts with that previously
reported by us for [(THF)2NaMg(NPh2)3] (6). Compound 6 re-
sults from a similar reaction of 1 with three equivalents of di-
phenylamine, and displays a monomeric arrangement in which
the amido groups coordinate terminally to Mg through their N
atoms; whereas the Na center has p interactions with two
phenyl groups in addition to binding to two THF ligands.[6]
Multinuclear NMR spectroscopy characterization of com-
pound 5 was performed in C6D6. 1H NMR spectroscopy (Fig-
ure S5) revealed a complex spectrum with multiple signals in
the aromatic, aliphatic, and NH regions. More informatively, the
13C NMR spectrum showed six different signals (ranging from
d=157.0–152.7 ppm) that can be assigned to the ipso-C atoms
of the 2,6-Me2-C6H4 groups (Figure S6), which suggests a lack
of equivalence between the anilide groups present in 5. This is
consistent with retention in C6D6 of the dimeric structure of 5
found in the solid state, with six nonequivalent anilide frag-
ments derived from four chiral nitrogen atoms and two pro-
chiral nitrogen atoms.[23] Further confirmation of the retention
of the dimeric arrangement of 5 in C6D6 was gained by using
1H DOSY NMR spectroscopy (see the Supporting Information).
Investigation of a solution of 5 (40 mm) in deuterated toluene
with tetramethylsilane (TMS) as an internal reference revealed
Scheme 3. Synthesis of sodium magnesiate 5.
tals of tris(amido)magnesiate [{(THF)3NaMg(NHAr3)}2] (5) in 58%
yield (Scheme 3).
The molecular structure of 5 was determined by using X-ray
crystallography, and was found to be dimeric and comprised
of a tetranuclear Na···Mg···Mg···Na chain arrangement connect-
ed by six anilide bridges (see Figure 1).[21] This gives rise to
three planar four-membered rings composed of two outer
{NaN2Mg} heterometallic rings linked through a central {Mg2N2}
homometallic ring that is orthogonal to the outer rings. Each
Mg atom in 5 is bonded to four amido groups with MgꢀN con-
tacts (mean value 2.08(5) ꢄ) that are similar to those found in
other reported tris(amido) alkali-metal magnesiates.[21] Three
molecules of THF complete the coordination sphere of each
sodium atom, which is also coordinated by two amido groups
with NaꢀN contacts (mean value 2.54(4) ꢄ) that are significant-
ly elongated compared with that reported for the homometal-
D
(diffusion coefficient) values of 5.076eꢀ10 and
2.262eꢀ09 m2 sꢀ1 (Figure S7 in the Supporting Information). By
using the external calibration curve (ECC) for dissipated
spheres and ellipsoids as elaborated by Stalke’s group,[24] the
molecular weight of compound 5 in solution was estimated to
be 1183 gmolꢀ1. This result deviates by only 5% when com-
pared to the dimeric structure observed for 5 in the solid
state. Interestingly, in donor [D8]THF as the solvent, DOSY ex-
periments indicate the formation of solvent-separated ion-pair
species (Scheme 4). In this case, two different diffusion coeffi-
cients were observed for
5.929eꢀ10 m2 sꢀ1). From these values, two molecular weights
5
(D1 =6.864eꢀ10 and D2 =
were calculated (Mr1 =773 gmolꢀ1 and Mr2 =994 gmolꢀ1),
Figure 1. Molecular structure of [{(THF)3NaMg(NAr3)}2] (5) with displacement
ellipsoids drawn at the 30% probability level. Disorder and hydrogen atoms,
except those attached to nitrogen atoms, are omitted for clarity. Selected
bond lengths [ꢄ] and angles [8]: Mg1ꢀN1 2.032(2), Mg1ꢀN2 2.051(2), Mg1ꢀ
N3 2.159(2), Mg1ꢀN5 2.111(2), Mg2ꢀN3 2.078(2), Mg2ꢀN4 2.057(2), Mg2ꢀN5
2.160(2), Mg2ꢀN6 2.028(2), Na1ꢀN1 2.590(2), Na1ꢀN2 2.539(2), Na1ꢀO1
2.40(2), Na1ꢀO 2.362(2), Na1ꢀO3 2.356(2), Na2ꢀN4 2.488(2), Na2ꢀN6
2.562(2); N1-Mg1-N2 105.38(9), N1-Mg1-N3 106.75(9), N1-Mg1-N5 136.73(9),
N2-Mg1-N3 102.89(8), N2-Mg1-N5 109.60(8), N3-Mg1-N5 89.58(7), N3-Mg2-
N4 107.65(9), N3-Mg2-N5 90.45(8), N3-Mg2-N6 137.49(9), N4-Mg2-N5
101.10(9), N4-Mg2-N6 105.43(9), N5-Mg2-N6 108.59(8).
Scheme 4. Proposed magnesiate species 5A and 5B observed in [D8]THF.
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
4
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
which are consistent with the presence in solution of mono-
anionic [(THF)3NaMg2(NHAr)6]ꢀ (5A; (Mr =769.66 gmolꢀ1) and
dianionic [Mg2(NHAr)6]2ꢀ (5B; Mr =1008.97 gmolꢀ1) species
(1% error for both species).[25] Furthermore, 2D 1H–1H EX-
SY NMR data (Figures S10 and S11 in the Supporting Informa-
tion) established that a slow exchange takes place between
5A and 5B in [D8]THF.
1H NMR spectroscopic monitoring of the reaction of an equi-
molar mixture of carbodiimide 3a and aniline 2 f in the pres-
ence of 2 mol% of tris(amido) magnesiate 5 indicated the for-
mation of guanidine 4h in a 99% yield after 15 min at room
temperature, which shows an identical efficiency to that found
for tris(alkyl)magnesiate 1 (99%, Table 1, entry 3). This hints at
a possible involvement of (amido)magnesiate 5 as an inter-
mediate in the catalytic cycle (see above). If this is the case,
the higher catalytic activity of 1 in the donor solvent THF com-
pared with benzene (Table 1) can be rationalized in terms of
the different constitutions of the relevant tris(amido) ate spe-
cies in these solvents. THF favors the formation of solvent-sep-
arated ion-pair (SSIP) species that can be anticipated to be
more powerful nucleophiles (with terminal MgꢀN bonds) than
the analogous contacted ion-pair (CIP) ates in which all the li-
gands are bridging between two metals.
contrast, 5 can react effectively with three equivalents of 3a to
give a 1:1 mixture of homometallic magnesium and sodium
guanidinates 8 and 9, which contain the unsymmetrical guani-
dinate ligand [iPrNC(NHiPr)NAr] that results from the formal in-
sertion of the carbodiimide into the NꢀH bonds of the anilide
groups present in 5 (Scheme 6).[26] The formation of [iPrNC-
(NHiPr)NAr] can be rationalized as a result of a proton transfer
from the arylamino nitrogen atom to an isopropylamido nitro-
gen followed by dissociation of the resultant NHiPr group and
formation of a new MꢀNAr bond (M=Mg or Na). This isomeri-
zation not only allows better stabilization of the negative
charge of the ligand (due to the conjugation effect between
the aromatic ring and the C=N bond), but also relief of the
steric hindrance around the metal by replacing one bulky NiPr
arm of the guanidinate ligand with a NAr substituent.[27] Con-
version of 5 into a 1:1 mixture of 8 and 9 occurs quantitatively,
1
as indicated by H NMR spectroscopic monitoring of the reac-
tion. Compound 8 could be crystallized from the reaction mix-
ture in 38% yield. Compound 9 could alternatively be pre-
pared by insertion of 3b in sodium amide [{(THF)2Na(NHAr)}2]
(10; see the experimental details in the Supporting Informa-
tion).
These results suggest that under stoichiometric conditions,
in the case of 5, the insertion of a third equivalent carbodi-
imide into the remaining anilide group induces the dispropor-
tionation of putative magnesiate [{Na(THF)x][Mg{iPrNC(NHiPr)-
NAr}3] into its monometallic guanidinate components 8 and 9.
This process is probably driven by the high steric congestion
around Mg when coordinated by three guanidinate ligands.
Attempts to prepare the relevant products of insertion that
result from the reactions of one and two equivalents of DIC
with 5 gave, in all cases, variable amounts of 8 and 9 (in 1:1
ratio) along with the recovery of unreacted 5. Thus, under the
conditions of this study, it appears that the threefold activation
of the MgꢀNHAr bonds of 5 is significantly favored over a pos-
sible sequential reactivity. In contrast, sodium magnesiate 7
does not react with a further equivalent of carbodiimide even
under forcing reaction conditions (12 h, 608C). This lack of re-
activity can be attributed to the steric congestion around Mg
in 7, which should compromise not only the approach of the
heterocumulene to the magnesiate anion but also the availa-
bility of the remaining NPh2 amido group to act as a nucleo-
phile, with its N atom sheltered by the cyclohexyl scaffolding
of the guanidinate ligands (see Figure 2, right, for a space-fill-
ing model) and also further stabilized by delocalization of its
lone pair across its two Ph substituents (sum of the angles
around N7=359.98; see Figure 2, left).
We next investigated the insertion reactions of tris(amido)
magnesiates 6 and 5 with three molar equivalents of carbodi-
imide 3b and 3a, respectively (Schemes 5 and 6, respectively).
Interestingly, completely different outcomes were observed de-
pending on the amido group present on the magnesiate. Com-
pound 6, with diphenylamido groups, can insert only two mol-
ecules of carbodiimide 3b to give heteroleptic mixed amido/
guanidinate sodium magnesiate 7 in 78% yield (Scheme 5). In
Scheme 5. Insertion reaction of 3b with tris(amido) magnesiate 6.
X-ray crystallographic studies established the molecular
structures of 7, 8, and 9 (Figures 2–4, respectively). Sodium
magnesiate 7 exhibits a SSIP structure comprised of a sodium
cation solvated by THF molecules, balanced by a magnesiate
anion in which the magnesium center is bound by two chelat-
ing guanidinate ligands and a terminal NPh2 group (Figure 2,
left). The five-coordinate Mg center displays a distorted square
planar pyramidal geometry (total bond angles around base=
3388), in which the NPh2 ligand occupies the apical position
(Figure 2, left).
Scheme 6. Insertion reaction of 3a with tris(amido) magnesiate 5.
Chem. Eur. J. 2016, 22, 1 – 12 www.chemeurj.org
5
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Figure 2. Left: Molecular structure of the [Mg{(CyN)2C(NPh2)}2(NPh2)]ꢀ anion present in 7 with displacement ellipsoids drawn at the 30% probability level. Hy-
drogen atoms are omitted for clarity. Right: Space-filling model for this anion. Selected bond lengths [ꢄ] and angles [8]: MgꢀN(1) 2.116(3), MgꢀN2 2.194(3),
MgꢀN4 2.178(3), MgꢀN5 2.130(3), N1ꢀC1 1.326(5), N2ꢀC1 1.299(5), N3ꢀC1 1.440(5), N4ꢀC26 1.304(5), N5ꢀC26 1.309(5), N6ꢀC26 1.469(4); N1-Mg-N7 120.16(15),
N1-Mg-N2 62.4(1), N1-Mg-N5 114.8(1), N2-Mg-N4 160.9(1), N2-Mg-N7 100.0(1), N4-Mg-N5 62.5(1), N4-Mg-N7 99.0(1), N5-Mg-N7 125.1(1).
Figure 3. Molecular structure of [Mg{(iPrNC(NAr)}(HNiPr)}2(THF)] (8; Ar=2,6-
Me2C6H3) with displacement ellipsoids drawn at the 30% probability level.
Hydrogen atoms, except those attached to nitrogen atoms and those on
Hydrogens atoms, except those attached to nitrogen atoms and those on
the CH groups of the isopropyl substituents, are omitted for clarity. Selected
Figure 4. Molecular structure of [Na{(iPrNC(NAr)(HNiPr)}(THF)]2 (9; Ar=2,6-
Me2C6H3) with displacement ellipsoids drawn at the 30% probability level.
the CH groups of the isopropyl substituents, are omitted for clarity. Selected
bond lengths [ꢄ] and angles [8]: Mg1ꢀO1 2.039(2), Mg1ꢀN1 2.158(2), Mg1ꢀ
bond lengths [ꢄ] and angles 8]: Na1ꢀO1 2.277(1), Na1ꢀN1 2.453(1), Na1ꢀN2
2.558(2), N1ꢀC5 1.341(2), N2ꢀC5 1.323(2), N3ꢀC5 1.404(2), N3ꢀH1N 0.9(2);
O1-Na1-N1 122.19(4), O1-Na1-N2 123.79(4), O1-Na1-N1’ 122.27(5), O1-Na1-
N2’ 120.48(4), N1-Na1-N2 53.57(4), N2-Na1-N1’ 89.20(5), N2’-Na1-N1 93.88(5),
N1-C5-N2 116.0(1), N1-C5-N3 120.4(1), N2-C5-N3 123.5(1).
N2 2.066(2), N1ꢀC10 1.344(3), N2ꢀC10 1.326(3), N3ꢀC10 1.385(3), N3ꢀH1N
0.84(4); O1-Mg1-N1 163.65(6), O1-Mg1-N2 104.94(7), N1-Mg1-N2 64.01(8),
N2-Mg1-N1’ 108.45(8).
Although, as far as we are aware, 7 constitutes the first ex-
ample of an alkali-metal magnesiate with guanidinate ligands
to be structurally defined, the structure of the magnesiate
anion bears a strong resemblance to that found for homome-
tallic magnesium complex 8 (Figure 3), though in this case the
apical position is filled by a molecule of the neutral donor THF.
Structures related to that of 8 have recently been reported by
Kays et al. for magnesium guanidinates, obtained by using an
alternative synthetic approach of MgnBu2 deprotonation of
guanidines with highly sterically demanding groups.[28,29] It
should also be noted that the guanidinate ligands present in 8
are unsymmetrical, with one of the chelating N’ atoms at-
tached to iPr (N1), whereas the remaining N atom (N2) binds
to 2,6-dimethylphenyl (Ar; see above). This lack of symmetry
translates to the formation of noticeably shorter MgꢀNR bonds
if R=iPr (2.066(2) ꢄ) than for the aromatic substituent
(2.158(2) ꢄ).
Sodium guanidinate 9 shows a dimeric arrangement with
the two guanidinate ligands parallel to each other (Figure 4).
Each sodium is coordinated by four N atoms of the Na2N4 core
(contacts range from 2.453(1) to 2.558(2) ꢄ) and a molecule of
THF, with a Na···Na1 vector of length 2.671(2) ꢄ, which lies per-
pendicular to the two guanidinate NCN planes. Although it
contains the same unsymmetrical guanidinate ligand described
above for 8, an opposite trend is observed for the length of
the NaꢀN bonds (NaꢀN1 2.453(1) ꢄ vs. NaꢀN2, 2.558(2) ꢄ) in 9.
The structure of 9 contrasts with that reported for trimeric
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
6
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
guanidinate complex {Na[CyNC(N(SiMe3)2)NCy]}3, which results
one equivalent of amine liberates the guanidine product and
regenerates the active sodium tris(amido) magnesiate.
[30]
from the reaction of DCC with NaN(SiMe3)2.
Protonolysis of guanidinate complexes 7–9 was attempted
by treating them with variable amounts of the relevant amine
(two equivalents of NHPh2 for 7, and two and one equivalents
of NH2Ar for 8 and 9, respectively). In all cases, no reaction was
apparent after 24 h at room temperature. The catalytic ability
of these guanidinate complexes was also investigated. Interest-
ingly, mixed-metal guanidinate 7 was able to catalyze the gua-
nylation of DCC with NHPh2 to give guanidine in almost identi-
cal yields to those found by using sodium magnesiate 1 (73%
vs. 75%; conditions in both cases: 2 mol% catalyst loading, RT,
1 h). However, illustrating the synergistic capabilities in sodium
magnesiate systems, single-metal guanidinates 8 and 9 dis-
played significantly lower efficiencies for the reaction of DIC
and NH2Ar than 1. Thus Mg complex 8 afforded the guanidine
product in a modest 30% yield after 15 min, whereas Na com-
plex 9 gave 72% conversion under the same conditions. Even
when an equimolar mixture of 8 and 9 was employed as a cata-
lyst for this reaction (2 mol% loading, RT, 15 min), the observed
conversions were still lower (76%) than when using preformed
bimetallic precatalyst 1 (99%). Collectively, despite the isola-
tion of single-metal complexes in some of the stoichiometric
studies, these results support the view that these guanylation
reactions are indeed ate-catalyzed transformations and high-
light the limitations of comparing the results of stoichiometric
reactions with the complex equilibria present during catalytic
processes in which variable excesses of reagents are present.
Previous insightful mechanistic studies by Richeson et al.
with LiHMDS as a catalyst have shown that the insertion step
in these processes is initiated by coordination of the carbodi-
imide to the Lewis acidic Li center.[31] By using bimetallic 1 as
a precatalyst, two potential sites are available for coordination
through either the Na or Mg centers. Repeating the guanyla-
tion of DIC (3a) by 2,6-dimethylaniline (2 f) with a 1:1 mixture
of sodium magnesiate 1 and the crown ether 15-crown-5
(which can block Na coordination sites) as a catalyst showed
only a slight decrease in the yield of guanidine 4h (from 99 to
83%). Coupled with the DOSY studies that show preference of
these bimetallic compounds to exist as SSIP structures in THF,
this suggests only a secondary role for sodium in this process,
with pre-coordination of the carbodiimide to Mg. Consistent
with this interpretation of a minor involvement of the alkali
metal, rather than stabilizing the magnesiate anion, using the
lithium derivative [LiMg(CH2SiMe3)3][32] as a precatalyst led to
conversions (97%) that were almost identical to those ob-
served when using sodium-containing 1. Thus, the enhanced
catalytic activity of these bimetallic systems seems to be a case
of anionic activation,[33] in which the formation of magnesiate
anions generates more powerful nucleophilic intermediates
than by using charge-neutral organomagnesium precursors.[16]
To obtain quantitative kinetic data, reactions of N,N’-diiso-
propylcarbodiimide with 4-tert-butylaniline in the presence of
1 as the catalyst were carried out in [D8]THF as the solvent. Re-
action rates of the guanylation were monitored over time by
1
using H NMR spectroscopy and using the integration changes
in the substrate resonances over more than three half-lives. We
first determined the order of the reaction with respect to
amine concentration by keeping the concentration of other
components virtually unaltered. The study was started by
using 2 mol% of catalyst 1 and the carbodiimide/amine molar
ratio was kept at 10:1 to maintain approximately zero-order
conditions for carbodiimide. The plot of ln([A]0/[A]t) versus re-
action time is shown in Figure 5, in which [A]0 is the initial
amine concentration and [A]t is the amine concentration after
a given reaction time. An induction period was not observed,
which indicates that the catalyst was reactive from the begin-
Mechanistic studies
The observations from our stoichiometric studies, together
with knowledge obtained from previous reports on s-block
single-metal catalysts,[9] suggest that these ate-catalyzed gua-
nylation reactions of amines may take place by the mechanism
presented in Scheme 7. Initially, fast protonation of sodium tris-
(alkyl) magnesiate 1 takes place to form a nucleophilic sodium
tris(amido) magnesiate (as seen for 5 and 6), which in turn can
undergo carbodiimide insertion to give a sodium magnesiate
guanidinate complex. Protonolysis of this latter species with
Figure 5. First-order kinetic analysis of the NMR-tube scale reaction of 4-tert-
^
butylaniline (&) or 4-tert-butylaniline-[D2] ( ) and diisopropylcarbodiimide in
Scheme 7. Proposed mechanism for the guanylation of anilines with 1.
[D8]THF with 2 mol% of 1 at RT.
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
7
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
ning of the process. The data confirm a fit consistent with first-
order kinetic behavior with respect to amine concentration.
Next, we determined the order of the reaction with respect
to the concentration of carbodiimide. During this study, we
maintained a relative carbodiimide/amine ratio of 1:10, and
the linearity of the plot of ln([C]0/[C]t), in which [C] is the carbo-
diimide concentration, versus reaction time shows that the re-
action was also first order with respect to this reagent (Fig-
ure S22 in the Supporting Information). Additionally, we carried
out a H/D kinetic isotope effect (KIE) experiment by using N,N’-
diisopropylcarbodiimide and [D2]-4-tert-butylaniline with cata-
lyst 1. This study gave a KIE (kH/kD) value of 5 (Figure 5). The
maximum calculated kinetic isotope effect (KIE) at 258C for
a reaction that involves a NꢀH bond should be approximately
8.5. In the guanylation reaction of tert-butylaniline, the magni-
tude of the measured value was clearly indicative of a primary
KIE[34] and indicates that a NꢀH bond is broken during the
turnover-limiting step. Although this observation would indi-
cate that protonolysis of the starting alkyl compound by the
amine could be the rate-determining step, it seems unlikely
because stoichiometric reactions demonstrate that these pro-
tolytic reactions occur instantaneously at room temperature.
Thus, the more limiting protonolysis of chelate guanidinate in-
termediates could be responsible for this high KIE.
Figure 7. Arrhenius (left) and Eyring plots (right) for the guanylation reaction
catalyzed by 1.
left) with an Ea value of 20.7 kJmolꢀ1. The activation parame-
ters were quantified by a plot of ln(kobs/T) versus 1/T, which re-
¼
¼
sults in DH =18.1 kJmolꢀ1 and DS =ꢀ25.8 Jmolꢀ1 Kꢀ1
(Figure 7, right).[35] This last value could support the existence
of a concerted transition state. Although the kinetic studies of
the guanylation process are scarce,[16a,35] several authors have
proposed an amine-assisted concerted transition state in com-
paratively analogous alkene hydroamination processes with
Group 2 metal complexes that involve, as in this case, a large
isotopic effect.[2a,36] This amine-assisted state could also explain
the first order observed with respect to amine (and carbodii-
mide). Under catalytic conditions, in which an excess of amine
is present through the main part of the process, we propose
that the magnesium amido complex formed in the first step
could coordinate an amine molecule so that the negatively
charged nitrogen atom of the incoming carbodiimide is stabi-
lized, which favors the attack of an amido ligand on the elec-
trophilic carbon atom (Figure 8).
The dependence of the rate of reaction with respect to cata-
lyst concentration was studied at different catalyst precursor
concentrations ([1]=1–5 mol%) with the carbodiimide/amine
molar ratio fixed at 10:1. A plot of reaction rate versus catalyst
concentration reveals a linear increase in the reaction rate with
catalyst concentration (Figure 6, left).
Figure 8. Proposed carbodiimide insertion transition state.
Figure 6. Plot of reaction rate versus concentration of the catalyst (left) and
van ’t Hoff plot (right).
Conclusion
The first-order rate of the reaction with respect to catalyst
concentration was further confirmed from the van ’t Hoff plot
for the three first concentrations (Figure 6 right). The value of
the slope was determined to be close to 1. Thus, from the
present study, the overall rate law for the guanylation of 4-tert-
butylaniline with N,N’-diisopropylcarbodiimide catalyzed by
1 at low concentrations could be summarized as shown in
Equation (1). A similar rate law has been obtained for trinuclear
zirconium alkyl diamido complexes.[34]
We report herein the first catalytic applications of alkali-metal
magnesiates for guanylation and hydrophosphination reac-
tions. A homoleptic mixed Na/Mg complex, [NaMg(CH2SiMe3)3]
(1), has been found to offer significantly greater catalytic ability
than that of its homometallic components [NaCH2SiMe3] and
[Mg(CH2SiMe3)2], which allows guanylation of a range of substi-
tuted anilines and secondary amines under very mild reaction
conditions (in most cases, at RT). Furthermore, by installing Mg
within this sodium magnesiate platform, it is possible to acti-
vate it towards catalysis of the hydrophosphination of carbo-
diimides at room temperature.
1
1
1
ð1Þ
rate ¼ k½amineꢁ ½carbodiimideꢁ ½catalystꢁ
Stoichiometric investigations have allowed the isolation and
structural elucidation of tris(amido) sodium magnesiate
[{(THF)3NaMg(NHAr3)}2] (5) and mixed amido/guanidinate
Values of kobs at four different temperatures were measured.
These kobs values satisfactorily fit the Arrhenius plot (Figure 7,
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
8
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
sodium magnesiate [Na(THF)5]+[Mg{(CyN)2C(NPh2)}2(NPh2)]ꢀ (7).
These appear to be intermediates in these catalytic transforma-
tions. Reactivity studies in these complexes, coupled with ki-
netic investigations, suggest these guanylation reactions occur
by forming highly nucleophilic (tris)amide intermediates that
can subsequently react with the carbodiiimide in an insertion
step, followed by amine protonolysis of the resultant guanidi-
nate species. Interestingly, all these processes appear to take
place in the coordination sphere of Mg, with Na taking a back-
seat in the catalytic cycle and stabilizing the magnesiate anion
intermediates, which hints that the enhanced catalytic activity
of these systems is due to anionic activation.
collection and processing was performed by using Rigaku and
Bruker software. All structures were refined to convergence on F2
of all independent reflections by using the full-matrix least-squares
method in the SHELXL program.[40] Selected crystallographic and
refinement details are given in Table S1 in the Supporting Informa-
tion.
[{(THF)3NaMg(NHAr)3}2] (5; Ar=2,6-Me2C6H3)
2,6-Dimethylaniline (3 mmol, 0.37 mL) was added to a suspension
of NaMg(CH2SiMe3)3 (1 mmol, 0.309 g) in hexane (10 mL). After stir-
ring for 1 h at RT, THF (2 mL) was added to give a light-brown solu-
tion. The solution was stored at ꢀ208C overnight to give colorless
1
crystals of sodium magnesiate 5 (0.362 g, 58%). H NMR (400 MHz,
The rate law for the guanylation of N,N’-diisopropylcarbodi-
imide with 4-tert-butylaniline catalyzed by 1 was deduced to
be first order with respect to [amine], [carbodiimide], and [cat-
alyst] and shows a large kinetic isotopic effect, which is consis-
tent with the formation of an amine-assisted rate-determining
carbodiimide insertion transition state.
298 K, C6D6): d=1.24 (m, 24H; THF), 1.87, 1.98, 2.07, 2.18, 2.19, 2.22
(36H; CH3, NHAr), 2.52, 2.57, 2.68, 2.76, 2.79, 2.81 (6H; NHAr), 3.19
(m, 24H; THF), 6.3–7.1 ppm (18H; CH, NHAr); 13C NMR (100 MHz,
298 K, C6D6): d=18.8, 19.5, 19.7, 20.0, 20.9, 21.6 (CH3, NHAr), 25.5
(THF), 67.8 (THF), 111.7, 111.8, 112.3, 112.6, 116.6, 121.4, 121.7,
122.0, 124.7, 125.3, 125.9, 128.6, 128.9, 129.0, 129.1, 129.3, 129.5,
129.8 (CH, NHAr), 152.7, 152.8, 156.1, 156.2, 156.6,157 ppm (ipso-C,
NHAr); elemental analysis calcd (%) for C72H108Mg2N6Na2O6: C 69.28,
H 8.72, N 6.73; found: C 69.25, H 8.85, N 7.12.
Experimental Section
[Na(THF)5]+[Mg{(CyN)2C(NPh2)}2(NPh2)]ꢀ (7)
General considerations
All reactions were performed under a protective argon atmosphere
by using standard Schlenk techniques. Hexane, benzene, and THF
were dried by heating to reflux over sodium benzophenone ketyl
and distilled under nitrogen or were passed through a column of
activated alumina (Innovative Tech.), degassed under nitrogen, and
stored over molecular sieves in the glovebox prior to use.
N,N’-Dicyclohexylcarbodiimide (DCC (3a); 0.62 g, 3 mmol) was
added to a solution of sodium magnesiate [(THF)2NaMg(NPh2)3] (6;
0.7 g, 1 mmol) in THF (4 mL). After stirring for 1 h, hexane (4 mL)
was added and the Schlenk tube was stored in the freezer (ꢀ308C)
overnight to allow the formation of colorless crystals of [Na(THF)5]+
[Mg{(CyN)2C(NPh2)}2(NPh2)]ꢀ (7; 1.03 g, 78%). 1H NMR (400 MHz,
298 K, C6D6): d=1.04–1.27, 1.47–1.73 (m, 40H; CH2, CyN), 1.43 (m,
20H; THF), ꢀ3.40–4.48 (m, 4H; CH, CyN), 3.57 (m, 20H; THF), 6.71
(t, J=7.1 Hz, 1H; NPh2), 6.79 (t, J=7.1 Hz, 1H; NPh2), 6.85–6.90 (m,
4H; CH, NPh2, guanidinate), 7.16–7.23, (m, 8H; CH, NPh2, guanidi-
nate), 7.28–7.38 (m, 4H; CH, NPh2), 7.44 (d, J=7.6 Hz, 8H; NPh2,
guanidinate), 7.50 (d, J=7.6 Hz, 2H; NPh2), 7.67 ppm (d, J=7.6 Hz,
2H; NPh2);13C NMR (100 MHz, 298 K, C6D6): d=25.7 (THF), 26.2,
26.5, 26.6, 26.8, 37.3, 37.4, (CH2, CyN) 55, 56.1 (CH, CyN), 67.9 (THF),
121.1, 129.3, 130.2 (CH, NPh2, guanidinate), 145.8 (ipso-C, NPh2, gua-
nidinate), 122, 129.5, 130.3 (CH, NPh2), 146.3 (ipso-C, NPh2),
163.5 ppm (CN3); elemental analysis calcd (%) for C82H114MgN7NaO5:
C 74.32, H 8.67, N 7.40; found: C 74.48, H 8.41, N 8.37.
Mg(CH2SiMe3)2,
NaCH2SiMe3,
[NaMg(CH2SiMe3)3],
and
[(THF)2NaMg(NPh2)3] were prepared according to literature proce-
dures.[6,19,37] LiCH2SiMe3, amines, phosphines, and carbodiimides
were purchased from Sigma–Aldrich and used as received. NMR
spectra were recorded by using a Bruker DPX400 MHz spectrome-
ter operated at 400.13 (1H) or 100.62 MHz (13C), or by using
a Varian FT-400 spectrometer with standard VARIAN-FT software. El-
emental analyses were carried out by using a Perkin–Elmer 2400 el-
emental analyzer.
Preparative scale reaction of the guanidines and phospha-
guanidines
In a glovebox, a solution of compound 1 (2% mol) in THF (3 mL)
was added in a Schlenk tube. Amine (or phosphane; 1.00 mmol)
and carbodiimide (1.00 mmol) were then added to the above reac-
tion mixture. The Schlenk tube was removed from the glovebox
and the reaction was stirred at the desired temperature. After car-
rying out the reaction for the desired time, the solution was con-
centrated under reduced pressure, hexane was added, and the
mixture was placed in a refrigerator at ꢀ308C for 16 h. After filtra-
tion, the products were obtained as white microcrystalline solids
and characterized by comparing their NMR spectra with the litera-
ture data.[14a,e,16a,27,38]
Stoichiometric studies: reaction between 5 and 3a
Sodium magnesiate [{(THF)3NaMg(NHAr3)}2] (5; 0.312 g, 0.25 mmol)
was reacted with N,N’-diisopropylcarbodiimide (1.5 mmol, 0.23 mL)
in THF (2 mL). The reaction mixture was stirred for 1 h, then
hexane (4 mL) was added (if a precipitate formed, it was redis-
solved with gentle heating). The solution was stored at ꢀ158C
overnight to allow the formation of colorless crystals of
[Mg{(iPrN)C(NAr)(HNiPr)}2(THF)] (8; 112 mg, 38%). 1H NMR
(400 MHz, 298 K, [D8]THF): d=0.56 (d, J=6.2 Hz, 12H; CH3, iPr),
0.79 (d, J=6.4 Hz, 12H; CH3, iPr), 1.77 (m, 4H; THF), 2.19 (s, 12H;
CH3, NAr), 3.02–3.15 (m, 4H; CH, iPr), 3.61 (m, 4H; THF) 3.80 (brd,
2H; NHiPr), 6.56 (t, J=7.6 Hz, 2H; para-CH, NAr), 6.80 ppm (d, J=
7.6 Hz, 4H; meta-CH, NAr);13C NMR (100 MHz, 298 K, [D8]THF): d=
19.7 (CH3, NAr), 24.2 (CH3, iPr), 25 (CH3, iPr), 26.2 (THF), 44.7 (CH,
iPr), 45.0 (CH, iPr), 68.0 (THF), 120.1 (para-CH, NAr), 128.1 (meta-CH,
NAr), 132.6 (ortho-C, NAr), 150.3 (ipso-C, NAr), 163.5 ppm (CN3); ele-
mental analysis calcd (%) for C34H56MgN6O: C 69.31, H 9.58, N
14.26; found: C 68.73, H 9.26, N 13.97.
X-ray crystallography
Data for samples 5, 8, and 9 were measured by using Oxford Dif-
fraction diffractometers[39] with MoKa (l=0.71073 ꢄ) or CuKa (l=
1.5418 ꢄ). Data for sample 7 were measured at Beamline I19 of the
Diamond Light Source with l=0.6889 ꢄ radiation and a Crystal
Logics diffractometer with Rigaku Saturn 724+ CCD detector; data
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
9
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[Na{(iPrN)C(NAr)(HNiPr)}(THF)]2 (9; Ar=2,6-Me2C6H3)
Keywords: cooperative effects · guanidines · homogeneous
catalysis · magnesiates · s-block metals
N,N’-Diisopropylcarbodiimide (0.08 mL, 0.5 mmol) was added to
a solution of sodium (2,6-dimethylphenyl)amide [{Na(NHAr)}2] (10;
0.143 g, 0.5 mmol) in hexane/THF (5 mL, 4:1 v/v). The resulting
pale-yellow solution was stored in the freezer (ꢀ308C) overnight
to allow the formation of colorless crystals of [Na{(iPrN)C(NAr)(H-
NiPr)}(THF)]2 (9; 0.108 g, 63%). 1H NMR (400 MHz, 298 K, [D8]THF):
d=0.91 (s, 24H; CH3, iPr), 1.78 (m, 4H; THF), 2.11 (s, 6H; CH3, NAr),
3.30 (brs, 4H+2H; CH, iPr+NHiPr), 3.61 (m, 4H; THF), 6.26 (t, J=
7.2 Hz, 2H; para-CH, NAr), 6.69 ppm (d, J=7.2 Hz, 4H; meta-CH,
NAr).;13C NMR (100 MHz, 298 K, [D8]THF): d=19.8 (CH3, NAr), 24.4
(CH3, iPr), 26.2 (THF), 27.0 (CH3, iPr), 44.6 (CH, iPr), 46.7 (CH, iPr),
115.3 (ortho-C, NAr), 127.7 (meta-CH, NAr), 129.7 (para-CH, NAr),
155.3 (ipso-C, NAr), 160.2 ppm (CN3); elemental analysis calcd (%)
for C34H56N6Na2O (one molecule of THF per dimer was considered,
based on the NMR data): C68.50, H 9.24, N 13.76; found: C 67.14, H
9.22, N 14.51.
[1] For reviews on Group 2 catalysis, see: a) M. S. Hill, D. J. Liptrot, C. Weet-
man, Chem. Soc. Rev. 2016, 45, 972; b) S. Harder, Chem. Rev. 2010, 110,
3852; c) A. G. M. Barrett, M. R. Crimmin, M. S. Hill, P. A. Procopiou, Proc.
R. Soc. London Ser. A 2010, 466, 927.
[2] a) C. Brinkmann, A. G. M. Barrett, M. S. Hill, P. A. Procopiou, J. Am. Chem.
Soc. 2012, 134, 2193; b) C. Weetman, M. D. Anker, M. Arrowsmith, M. S.
Hill, G. Kociok-Kohn, D. J. Liptrot, M. F. Mahon, Chem. Sci. 2016, 7, 628;
c) D. J. Liptrot, M. S. Hill, M. F. Mahon, A. S. S. Wilson, Angew. Chem. Int.
Ed. 2015, 54, 13362; Angew. Chem. 2015, 127, 13560; d) M. D. Anker,
M. S. Hill, J. P. Lowe, M. F. Mahon, Angew. Chem. Int. Ed. 2015, 54, 10009.
[3] a) J. Penafield, L. Maron, S. Harder, Angew. Chem. Int. Ed. 2015, 54, 201;
Angew. Chem. 2015, 127, 203; b) J. Intemann, H. Bauer, J. Pahl, L. Maron,
S. Harder, Chem. Eur. J. 2015, 21, 11452; c) J. Intemann, M. Lutz, S.
Harder, Organometallics 2014, 33, 5722.
[4] B. Liu, T. Roisnel, J. F. Carpentier, Y. Sarazin, Angew. Chem. Int. Ed. 2012,
51, 4943; Angew. Chem. 2012, 124, 5027.
[5] a) J.-S. Ryu, G. Y. Li, T. J. Marks, J. Am. Chem. Soc. 2003, 125, 12584;
b) D. V. Gribkov, K. C. Hultzsch, F. Hampel, J. Am. Chem. Soc. 2006, 128,
3748.
[6] A. Hernꢀn-Gꢁmez, T. D. Bradley, A. R. Kennedy, Z. Livingstone, S. D. Rob-
ertson, E. Hevia, Chem. Commun. 2013, 49, 8659.
General procedure for kinetic experiments
Kinetic experiments were performed by using a Varian FT-400 MHz
spectrometer, and a standard solution of catalyst 1 in deuterated
THF was used. The described kinetic experiments were carried out
on N,N’-diisopropylcarbodiimide 3a and 4-tert-butilaniline 2c to
form the corresponding guanidine. Reactions were carried out in J-
Young NMR tubes and the reaction rates were measured by moni-
toring the disappearance of amine (or carbodiimide) and the for-
mation of guanidine by using 1H NMR spectroscopy at the de-
scribed intervals over more than three half-lives. All data were pro-
cessed by using Varian integral analysis software. Reaction rates
were derived from the plot of Ln[substrate]0/[substrate] vs. time
(by fitting data to the equation Ln[substrate]0/[substrate]=kobst) by
using linear trend lines generated by using Microsoft Excel soft-
ware. To obtain Arrhenius and Eyring plots, kinetic analyses were
conducted at four different temperatures, each separated by ap-
proximately 5–10 K.
[7] a) T. Klatt, J. T. Markiewicz, C. Saemann, P. Knochel, J. Org. Chem. 2014,
79, 4253; b) A. Harrison-Marchand, F. Mongin, Chem. Rev. 2013, 113,
7470; c) F. Mongin, A. Harrison-Marchand, Chem. Rev. 2013, 113, 7563;
d) R. E. Mulvey, S. D. Robertson, Top Organomet Chem 2014, 45, 103.
[8] For selected references on alkali-metal magnesiates, see: a) N. Boudet,
J. R. Lachs, P. Knochel, Org. Lett. 2007, 9, 5525; b) L. Shi, Y. Chu, P. Kno-
chel, H. Mayr, Angew. Chem. Int. Ed. 2008, 47, 202; Angew. Chem. 2008,
120, 208; c) A. Krasovskiy, V. Krasovskaya, P. Knochel, Angew. Chem. Int.
Ed. 2006, 45, 2958; Angew. Chem. 2006, 118, 3024; d) S. H. Wunderlich,
C. J. Rohbogner, A. Unsinn, P. Knochel, Org. Process Res. Dev. 2010, 14,
339; e) D. Tilly, F. Chevallier, F. Mongin, P. C. Gros, Chem. Rev. 2014, 114,
1207; f) C. Y. Liu, P. Knochel, Org. Lett. 2005, 7, 2543; g) R. J. Kloetzing, A.
Krasovskiy, P. Knochel, Chem. Eur. J. 2006, 12, 215; h) F. F. Fleming, V. Gu-
dipati, O. W. Steward, Org. Lett. 2002, 4, 659; i) J. Kondo, A. Inoue, H.
Shinokubo, K. Oshima, Angew. Chem. Int. Ed. 2001, 40, 2085; Angew.
Chem. 2001, 113, 2146; j) P. C. Andrikopoulos, D. R. Armstrong, D. V.
Graham, E. Hevia, A. R. Kennedy, R. E. Mulvey, C. T. O’Hara, C. Talmard,
Angew. Chem. Int. Ed. 2005, 44, 3459; Angew. Chem. 2005, 117, 3525;
k) E. Hevia, G. W. Honeyman, A. R. Kennedy, R. E. Mulvey, D. C. Sherring-
ton, Angew. Chem. Int. Ed. 2005, 44, 68; Angew. Chem. 2005, 117, 70;
l) B. Conway, E. Hevia, A. R. Kennedy, R. E. Mulvey, Chem. Commun.
2007, 2864; m) A. J. Martinez-Martinez, D. R. Armstrong, B. Conway, B. J.
Fleming, J. Klett, A. R. Kennedy, R. E. Mulvey, S. D. Robertson, C. T.
O’Hara, Chem. Sci. 2014, 5, 771; n) A. J. Martꢂnez-Martꢂnez, A. R. Kennedy,
R. E. Mulvey, C. T. O’Hara, Science 2014, 346, 834.
Determination of reaction order with respect to aniline 2c
and carbodiimide 3a
The order of the reaction with respect to amine concentration was
determined by keeping the concentration of the other compo-
nents virtually unaltered, using a 1.9 mm solution of catalyst 1, and
using a carbodiimide/amine molar ratio greater than 10:1. The
excess carbodiimide concentration maintains approximately zero-
order conditions. The order of the reaction with respect to the con-
centration of carbodiimide was studied at an amine/carbodiimide
molar ratio higher than 10:1.
[9] For reviews in the area, see: a) C. Alonso-Moreno, A. AntiÇolo, F. Carril-
lo-Hermosilla, A. Otero, Chem. Soc. Rev. 2014, 43, 3406; b) W.-X. Zhang,
L. Xu, Z. Xi, Chem. Commun. 2015, 51, 254; c) L. Xu, W.-X. Zhang, F. Xi,
Organometallics 2015, 34, 1787.
[10] See, for example: a) R. G. S. Berlinck, A. E. Trindade-Silva, M. F. C. Santos,
Nat. Prod. Rep. 2012, 29, 1382; b) D. Castagnolo, S. Schenone, M. Botta,
Chem. Rev. 2011, 111, 5247; c) S. Ekelund, P. Nygren, R. Larsson, Biochem.
Pharmacol. 2001, 61, 1183; d) P. P. Toth, Postgrad. Med. 2014, 126, 7;
e) D. T. Nash, J. Clin. Pharmacol. 1973, 13, 416; f) N. Scheinfeld, Dermatol.
Online J. 2003, 9, 4; g) M. von Itzstein, Nat. Rev. Drug Discovery 2007, 6,
967; h) Y. Hirata, I. Yanagisawa, Y. Ishii, S. Tsukamoto, N. Ito, Y. Isomura,
M. Takeda, U. S. Patent 1981, 4.283.408; i) E. Buchdunger, J. Zimmer-
mann, H. Mett, T. Meyer, M. Mueller, B. J. Druker, N. B. Lydon, Cancer Res.
1996, 56, 100; j) L. Troncone, V. Rufini, Anticancer Res. 1997, 17, 1823;
k) C. Loesberg, H. Van Rooij, J. C. Romijm, L. A. Smets, Biochem. Pharma-
col. 1991, 42, 793.
Acknowledgements
We thank the University of Strathclyde (studentship to Z.L.)
and the European Research Council (ERC-STG grant to E.H.) for
generous sponsorship of this research, and the Spanish Minis-
terio de Economꢂa y Competitividad (MINECO) for support
under project CTQ2014-51912-REDC. We also thank Diamond
Light Source for access to beamline I19. A.M., F.C.-H., and A. A.
gratefully acknowledge financial support from the Junta de Co-
munidades de Castilla-La Mancha (project PEII-2014-041-P).
[11] See, for example: a) P. J. Bailey, S. Pace, Coord. Chem. Rev. 2001, 214, 91;
b) W. E. Piers, D. J. H. Emslie, Coord. Chem. Rev. 2002, 233, 131; c) M. P.
Coles, Dalton Trans. 2006, 985; d) P. C. Junk, M. L. Cole, Chem. Commun.
2007, 1579; e) F. T. Edelmann, Adv. Organomet. Chem. 2008, 57, 183;
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
10
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
f) F. T. Edelmann, Adv. Organomet. Chem. 2013, 61, 55; g) J. Barker, M.
Kilner, Coord. Chem. Rev. 1994, 133, 219; h) F. T. Edelmann, Coord. Chem.
Rev. 1994, 137, 403; i) F. T. Edelmann, Chem. Soc. Rev. 2009, 38, 2253;
j) F. T. Edelmann, Chem. Soc. Rev. 2012, 41, 7657; k) C. Jones, Coord.
Chem. Rev. 2010, 254, 1273; l) A. A. Trifonov, Coord. Chem. Rev. 2010,
254, 1327.
cos, B. J. Fleming, P. Garcꢂa-ꢅlvarez, A. R. Kennedy, K. Reilly, G. M. Robert-
son, S. D. Robertson, C. T. O’Hara, Dalton Trans. 2014, 43, 14424.
[22] D. Barr, W. Clegg, L. Cowton, L. Horsburgh, F. M. Mackenziec, R. E.
Mulvey, J. Chem. Soc. Chem. Commun. 1995, 891.
[23] X-ray analysis shows only three nonequivalent anilide fragments with R,
S, R, S configuration for N4, N6, N2, and N1. However, during the iden-
tification of the hydrogen atoms on the Fourier map, a possible disor-
der was noticed that precludes the accurate establishment of the real
configuration. Thus, two different possibilities can be found; a diastereo-
mer with a configuration of the type R, R, R, S, which would show six
inequivalent anilide fragments, or a mixture of diastereomers, such as
R, S, R, S and R, R, S, S, which similarly would display a total of six sig-
nals for the anilide fragments.
[12] See, for example: a) T. Ishikawa, T. Isobe, Chem. Eur. J. 2002, 8, 552; b) T.
Ishikawa, T. Kumamoto, Synthesis 2006, 737; c) T. Ishikawa, Superbases
for Organic Synthesis: Guanidines, Amidines, Phosphazenes and Organo-
catalysts, Wiley, New York, 2009; d) J. E. Taylor, S. D. Bull, J. M. J. Williams,
Chem. Soc. Rev. 2012, 41, 2109; e) D. Leow, C.-H. Tan, Chem. Asian J.
2009, 4, 488; f) P. Selig, Synthesis 2013, 703; g) M. P. Coles, Chem.
Commun. 2009, 3659; h) X. Fu, C.-H. Tan, Chem. Commun. 2011, 47,
8210; i) A. Avila, R. Chinchilla, E. Gꢁmez-Bengoa, C. Nꢀjera, Eur. J. Org.
Chem. 2013, 5085; j) C. Thomasa, B. Bibal, Green Chem. 2014, 16, 1687.
[13] a) M. K. T. Tin, N. Thirupathi, G. P. A. Yap, D. S. Richeson, J. Chem. Soc.
Dalton Trans. 1999, 2947; b) M. K. T. Tin, G. P. A. Yap, D. S. Richeson,
Inorg. Chem. 1998, 37, 6728; c) E. W. Thomas, E. E. Nishizawa, D. C. Zim-
mermann, D. J. Williams, J. Med. Chem. 1989, 32, 228.
[14] a) T.-G. Ong, G. P. A. Yap, D. S. Richeson, J. Am. Chem. Soc. 2003, 125,
8100; b) A. Grirrane, H. Garcia, E. ꢅlvarez, Beilstein J. Org. Chem. 2013, 9,
1455; c) S. Zhou, S. Wang, G. Yang, Q. Li, L. Zhang, Z. Yao, Z. Zhou, H.-B.
Song, Organometallics 2007, 26, 3755; d) F. Han, Q. Teng, Y. Zhang, Y.
Wang, Q. Shen, Inorg. Chem. 2011, 50, 2634; e) L. Xu, Z. Wang, W.-X.
Zhan, Z. Xi, Inorg. Chem. 2012, 51, 11941; f) J. Romero-Fernꢀndez, F. Car-
rillo-Hermosilla, A. AntiÇolo, C. Alonso-Moreno, A. M. Rodrꢂguez, I.
Lꢁpez-Solera, A. Otero, Dalton Trans. 2010, 39, 6419; g) D. Elorriaga, F.
Carrillo-Hermosilla, A. AntiÇolo, F. J. Suꢀrez, I. Lꢁpez-Solera, R. Fernꢀn-
dez-Galꢀn, E. VillaseÇor, Dalton Trans. 2013, 42, 8223.
[24] R. Neufeld, D. Stalke, Chem. Sci. 2015, 6, 3354.
[25] For examples of sodium magnesiates that display a SSIP structure in
which sodium is solvated by THF molecules, see: a) D. R. Armstrong, W.
Clegg, A. Hernan-Gomez, A. R. Kennedy, Z. Livingstone, S. D. Robertson,
L. Russo, E. Hevia, Dalton Trans. 2014, 43, 4361; b) V. L. Blair, W. Clegg,
A. R. Kennedy, Z. Livingstone, L. Russo, E. Hevia, Angew. Chem. Int. Ed.
2011, 50, 9857; Angew. Chem. 2011, 123, 10031.
[26] See the Supporting Information for synthesis and spectroscopic charac-
terization.
[27] For other examples of structurally defined guanidinates where a similar
rearrangement has taken place, see references [14g], [16a], [16d], and
W. X. Zhang, M. Nishiura, Z. Hou, Chem. Eur. J. 2007, 13, 4037.
[28] G. J. Moxey, A. J. Blake, W. Lewis, D. L. Kays, Eur. J. Inorg. Chem. 2015,
5892.
[29] A similar motif has been also reported by Chang for the reaction of
Mg[N(iPr)2]2 with two equivalents of DIC, see: B. Srinivas, C.-C. Chang,
C.-H. Chen, M. Y. Chiang, I.-T. Chen, Y. Wang, G.-H. Lee, J. Chem. Soc.
Dalton Trans. 1997, 957.
[30] G. R. Giesbrecht, A. Shafir, J. Arnold, J. Chem. Soc. Dalton Trans. 1999,
3601.
[31] C. N. Rowley, T.-G. Ong, J. Priem, T. K. Woo, D. S. Richeson, Inorg. Chem.
2008, 47, 9660.
[32] S. E. Baillie, W. Clegg, P. Garcıa-ꢅlvarez, E. Hevia, A. R. Kennedy, J. Klett,
L. Russo, Organometallics 2012, 31, 5131.
[33] For an example of anionic activation of magnesium reagents by form-
ing mixed ammonium-magnesiate salts, see: C. Vidal, J. Garcia-Alvarez,
A. Hernan-Gomez, A. R. Kennedy, E. Hevia, Angew. Chem. Int. Ed. 2014,
53, 5969; Angew. Chem. 2014, 126, 6079.
[34] A. Mukherjee, T. K. Sen, S. K. Mandal, B. Maity, D. Koley, RSC Adv. 2013,
3, 1255.
[35] J. H. Espenson, Chemical kinetics and reaction mechanisms, 2nd ed.;
McGraw-Hill: New York, 1995.
[15] T.-G. Ong, J. S. O’Brien, I. Korobkov, D. S. Richeson, Organometallics
2006, 25, 4728.
[16] a) C. Alonso-Moreno, F. Carrillo-Hermosilla, A. Garcꢆs, A. Otero, I. Lꢁpez-
Solera, A. M. Rodrꢂguez, A. AntiÇolo, Organometallics 2010, 29, 2789;
b) D. Elorriaga, F. Carrillo-Hermosilla, A. AntiÇolo, I. Lꢁpez-Solera, R.
Fernꢀndez-Galꢀn, A. Serrano, E. VillaseÇor, Eur. J. Inorg. Chem. 2013,
2940; c) A. Baishya, M. Kr. Barman, T. Peddarao, S. Nembenna, J. Organo-
met. Chem. 2014, 769, 112; d) J. R. Lachs, A. G. M. Barrett, M. R. Crimmin,
G. Kociok-Kçhn, M. S. Hill, M. F. Mahon, P. A. Procopiou, Eur. J. Inorg.
Chem. 2008, 4173.
[17] To best of our knowledge, only three examples of magnesium com-
plexes have been described so far in the literature as catalysts for gua-
nylation reactions: MgnBu2, [MgBz2(THF)2], and an N-heterocyclic car-
bene (NHC)-supported magnesium amide, see references [16a], [16b],
and [16c], respectively.
[18] M. R. Crimmin, A. G. M. Barrett, M. S. Hill, P. B. Hitchcock, P. A. Procopiou,
Organometallics 2008, 27, 497.
[19] S. E. Baillie, W. Clegg, P. Garcia-Alvarez, E. Hevia, A. R. Kennedy, J. Klett,
L. Russo, Chem. Commun. 2011, 47, 388.
[36] J. F. Dunne, D. B. Fulton, A. Ellern, A. D. Sadow, J. Am. Chem. Soc. 2010,
132, 17680.
[37] J. Klett, W. Clegg, B. Conway, A. R. Kennedy, R. E. Mulvey, L. Russo, Eur. J.
Inorg. Chem. 2011, 721.
[20] a) X. Zhang, C. Wang, C. Qian, F. Han, F. Xu, Q. Shen, Tetrahedron 2011,
67, 8790; b) X. Zhu, Z. Du, F. Xu, Q. Shen, J. Org. Chem. 2009, 74, 6347.
[21] For other examples of structurally defined tris(amido)magnesiates, see:
a) E. Hevia, F. R. Kenley, A. R. Kennedy, R. E. Mulvey, R. B. Rowlings, Eur. J.
Inorg. Chem. 2003, 3347; b) G. C. Forbes, A. R. Kennedy, R. E. Mulvey,
P. J. A. Rodger, R. B. Rowlings, J. Chem. Soc. Dalton Trans. 2001, 1477;
c) W. Clegg, K. W. Henderson, R. E. Mulvey, P. A. O’Neil, J. Chem. Soc.
Dalton Trans. 1994, 769; d) B. J. Fleming, P. Garcꢂa-ꢅlvarez, E. Keating,
A. R. Kennedy, C. T. O’Hara, Inorg. Chim. Acta 2012, 384, 154; e) J. Fran-
[38] a) Q. Li, S. Wang, S. Zhou, G. Yang, X. Zhu, Y. Liu, J. Org. Chem. 2007, 72,
6763; b) Z. Du, W. Li, X. Zhu, F. Xu, Q. Shen, J. Org. Chem. 2008, 73,
8966; c) M. P. Coles, P. B. Hitchcock, Chem. Commun. 2002, 2794.
[39] CrysAlisPro; Oxford Diffraction: Oxford, UK 2008.
[40] G. Sheldrick, Acta Crystallogr. Sect. C 2015, 71, 3–8.
Received: June 17, 2016
Published online on && &&, 0000
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
11
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
FULL PAPER
Magnesium activATES: Showing en-
hanced catalytic ability compared with
their homometallic components (see
figure), mixed Na/Mg complexes effec-
tively catalyze the guanylation of
a range of amines under mild reaction
conditions, the key for which is the
anionic activation of magnesium after
installation within a sodium magnesiate
platform.
&
Cooperative Effects
M. De Tullio, A. Hernꢀn-Gꢁmez,
Z. Livingstone, W. Clegg, A. R. Kennedy,
R. W. Harrington, A. AntiÇolo, A. Martꢂnez,
F. Carrillo-Hermosilla,* E. Hevia*
&& – &&
Structural and Mechanistic Insights
into s-Block Bimetallic Catalysis:
Sodium Magnesiate-Catalyzed
Guanylation of Amines
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
12
ꢃ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!