Mechanistic insights of anionic ligand exchange and fullerene reduction with magnesium(i) compounds

Article abstract of DOI:10.1039/c9dt03976g

Ligand exchange reactions between combinations of the complexes [{(^Arnacnac)Mg}₂], where Ar = 2,6-iPr₂C₆H₃ (Dip), 2,6-Et₂C₆H₃ (Dep), 2,4,6-Me₄C₆H

Full text of DOI:10.1039/c9dt03976g

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Mechanistic insights of anionic ligand exchange
and fullerene reduction with magnesium(I)
compounds†
Cite this: Dalton Trans., 2019, 48,
16936
Samuel R. Lawrence, David B. Cordes,
Andreas Stasch
Alexandra M. Z. Slawin
and
*
Ligand exchange reactions between combinations of the complexes [{(Arnacnac)Mg}2], where Ar = 2,6-
iPr2C6H3 (Dip), 2,6-Et2C6H3 (Dep), 2,4,6-Me4C6H2 (Mes), and 2,6-Me2C6H3 (Xyl), [({Ph2P(NDip)2}Mg)2],
[(Arnacnac)Li], where Ar = Mes or Xyl, and [{Ph2P(NDip)2}Li] were studied in deuterated aromatic and ali-
phatic solvents, and tetrahydrofuran. The reactions afforded product mixtures with asymmetrically substi-
tuted dimagnesium(I) complexes [(Arnacnac)MgMg(Ar’nacnac)], where Ar, Ar’ = Dip, Dep, Mes, Xyl and
[{Ph2P(NDip)2}MgMg(Arnacnac)], where Ar = Mes or Xyl, and suggest that the exchange of anionic ligands
2+
on the Mg2 ion proceeds via an associative mechanism and is strongly dependent on ligand sterics and
ligand shape, and can be very rapid. The activation reaction of fullerene C60 by dimagnesium(I) complexes
[{(Arnacnac)Mg}2] and [({Ph2P(NDip)2}Mg)2] to fulleride complexes is similarly dependent on ligand
sterics and ligand shape, but likely does not involve direct coordination of the fullerene to the Mg centre
Received 9th October 2019,
Accepted 29th October 2019
6−
DOI: 10.1039/c9dt03976g
in dimagnesium(I) compounds prior to its reduction. The new C60 fulleride complex [({Ph2P(NDip)2}
rsc.li/dalton
Mg)6C60] was prepared, and spectroscopically and structurally characterised.
phinate (2),4 and amides.5 In these, the Mg2 ion contains a
2+
Introduction
Mg–Mg σ-bond that shows almost exclusively s-character1 with
a rare non-nuclear attractor at the midpoint of the bond.6 The
2+
Stable coordination complexes of the highly reducing Mg2
ion1 have so far been obtained with a range of suitable steri- Mg–Mg bond in these complexes is relatively long, for example
cally demanding anionic ligands consisting of β-diketiminates ca. 2.85 Å for 1a, and can significantly lengthen upon coordi-
and related ligands, e.g. complexes 1 (Fig. 1),2 a guanidinate,2k nation of donor molecules,2a,i,j e.g. up to ca. 3.20 Å in the
a diazabutadienide and related systems,3 a diiminophos- vicinal bis-(4-dimethylaminopyridine) adduct of 1a.2j
Coordination of some donors such as THF or dioxane has
been found to be facilely reversible. On the other hand, high
pressure can effect a significant shortening of the Mg–Mg
bond in the solid state7 and highlights the ease of stretching
and compressing the Mg–Mg bond. As these complexes are
regularly employed as very strong reducing agents towards a
range of organic, inorganic and organometallic substrates,1
understanding the coordination chemistry of Mg–Mg-bonded
complexes is important for their applications.
Related to Mg–Mg-bonded compounds are coordination
2+
complexes of the Zn2 ion,8 that appear to be less reducing in
Fig. 1 Complexes 1 and 2. Dip = 2,6-diisoproylphenyl, Dep = 2,6-di-
comparison to Mg22+ complexes and show stronger and signifi-
ethylphenyl, Mes = mesityl (2,4,6-trimethylphenyl), Xyl = xylyl (2,6-
cantly shorter metal–metal bonds (by ca. 17%), despite having
dimethylphenyl).
only slightly shorter Zn–N bonds (by ca. 2.5%) for comparable
complexes. Low oxidation state zinc chemistry has also been
EaStCHEM School of Chemistry, University of St Andrews, North Haugh, St Andrews,
KY16 9ST, UK. E-mail: as411@st-andrews.ac.uk
2+
extended to complexes of the Zn3 ion and Zn-cluster species,
for example.8,9 The coordination chemistry of Zn–Zn-bonded
complexes has so far allowed a wider variety of ligands as part
of stable isolated complexes, including those of simple dicatio-
†Electronic supplementary information (ESI) available: Experimental, spectro-
scopic, and crystallographic details (PDF). CCDC 1914904. For ESI and crystallo-
graphic data in CIF or other electronic format see DOI: 10.1039/c9dt03976g
16936 | Dalton Trans., 2019, 48, 16936–16942
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
nic species [(Do)3Zn–Zn(Do)3]2+ (Do = neutral donors).10 In the
In order to study the potential rearrangement processes
chemistry of Zn–Zn bonded compounds, stable sources such in LMgMgL complexes, we first reinvestigated ligand
as Cp*ZnZnCp* (Cp* = pentamethylcyclopentadienyl)8,11 have scrambling between different [{(Arnacnac)Mg}2] 1 complexes,
been used to access new complexes with retention of the forming asymmetric complexes [(Arnacnac)MgMg(Ar′nacnac)] 4.
Zn–Zn bond, for example by ligand exchange reactions using Previously, the reaction of [{(Dipnacnac)Mg}2] 1a and
proligands with protic hydrogens,8 or via metathesis reactions [{(Mesnacnac)Mg}2] 1c proceeded extremely slowly at room
involving the exchange of anionic ligands.10a,12 In this work we temperature or at 80 °C over two weeks and with replenish-
report on ligand exchange reactions involving dimeric mag- ment of decomposed 1c, affording predominantly [(Dipnacnac)
nesium(I) compounds and considerations around ligand MgMg(Mesnacnac)] 4ac.2f We studied the ligand influence by
shapes for these reactions, and activation reactions of C60.
mixing the smaller β-diketiminate examples [{(Mesnacnac)
Mg}2] 1c and [{(Xylnacnac)Mg}2] 1d in deuterated benzene or
toluene and found the completed formation of [(Mesnacnac)
MgMg(Xylnacnac)] 4cd alongside 1c and 1d after only a few
Results and discussion
1
minutes as judged by H NMR spectroscopy (Fig. S8 and 9†).
The reactivity of dimeric magnesium(I) compounds has been A statistical distribution of ca. 25% of 1c and 1d, respectively,
found to be typically strongly dependent on ligand and sub- and ca. 50% 4cd is formed due to the essentially identical
strate sterics.1 For the β-diketiminate derivatives 1, for steric and electronic profiles of the Mesnacnac and Xylnacnac
2+
example, it can thus significantly vary with aryl substitution. ligands around the Mg2 unit, affording a ΔG° ≈ 0 for this
For the reduction of (organic) substrates it is believed that the reaction. Similar scrambling reactions can be carried out at
molecules require coordination to one of the Mg centres of room temperature between other [{(Arnacnac)Mg}2] 1 combi-
[{(Arnacnac)Mg}2] 1, or at least a close approach to the Mg2
nations (Scheme 1 and the ESI†). Reactions involving more
2+
unit as a first step. This is supported by the fact that the more sterically demanding examples, e.g. 1a and 1b, are, as
sterically demanding derivatives (1a) can require higher reac- expected, considerably slowed down, for example in the for-
tion temperatures or show no reaction in some cases where mation of [(Dipnacnac)MgMg(Mesnacnac)] 4ac. Complexes
the less sterically demanding ones (1c,d) rapidly react. Also, [(Depnacnac)MgMg(Mesnacnac)] 4bc (Fig. S4 and 5†) and
reactions in neat coordinating solvents such as THF are typi- [(Depnacnac)MgMg(Xylnacnac)] 4bd (Fig. S6 and 7†), were
cally suppressed in comparison to those conducted in hydro- formed in solution reaching a near-equilibrium mixture in
carbons, due to THF coordination and blocking of the Mg under one day at room temperature.
centres, despite elongated Mg–Mg bonds in donor adduct
Arenes are often the solvent of choice for these reactions,
complexes.2j In this regard, however, it has recently been and can be involved as neutral ligands in metal complexes, but
found that mono adduct formation of [{(Arnacnac)Mg}2] 1 com- can also form reduced, anionic ligands as part of electron-rich
plexes, e.g. as in [{(Dipnacnac)Mg(Do)}{Mg(Dipnacnac)}] (Do = metal complexes. For example, it has very recently been
4-dimethylaminopyridine, DMAP, or 2,3,4,5-tetramethyl-imid- suggested that some reactions involving the aluminium(I)
azol-2-ylidene), can be used to lower the kinetic barrier for acti- complex [(Dipnacnac)Al] may proceed via a reduced arene
vation reactions and allowed the reductive trimerisation of CO complex,14 and subsequently, structurally characterised Ca/Al
to form magnesium deltate-complexes.2a Also, compounds of and Mg complexes with reduced benzene fragments have been
the type LMgMgL have been concluded to not dissociate into 2 reported, the latter are believed to be formed via radical
LMg• in solution and typically do not react via free radical intermediates.2b,15 To exclude an influence from the aromatic
mechanisms.1,2f,i,k
We have recently reported the facile activation of buckmin-
sterfullerene, C60, with [{(Arnacnac)Mg}2] 1 affording fulleride
complexes [{(Arnacnac)Mg}nC60], n = 6 (3, though not for Ar =
Dip), 4, and 2, that are best described as contact ion systems of
n−
n [(Arnacnac)Mg]+ with C60 ions.13 With respect to its initial
activation, the large and spherical shape of C60 makes is un-
likely in our view that it can easily coordinate directly to a well
protected Mg centre of 1 prior to reduction. As a working
hypothesis we speculated that the initial approach of C60 to
the Mg centres in 1 could be via the relatively open and unpro-
tected β-diketiminate backbone prior to reduction. This raises
questions of how facile ligand rearrangement processes in the
2+
coordination sphere of the Mg2 ion are. For example, pro-
cesses such as partial dissociation and re-coordination of
donor groups of ligands, and/or distortions of the coordi-
nation sphere can inevitably be linked to access of substrates
to the Mg2 core prior to reduction chemistry.
Scheme 1 Ligand scrambling reactions in complexes 1 to mixed-ligand
species 4.
This journal is © The Royal Society of Chemistry 2019
Dalton Trans., 2019, 48, 16936–16942 | 16937

View Article Online
Paper
Dalton Transactions
solvent in the reaction, e.g. via scrambling of possibly short-
lived [(Arnacnac)Mg(C6H5R)Mg(Arnacnac)] (R = H, Me) inter-
mediates with reduced arene ligands, we conducted the reac-
tion between 1c and 1d in deuterated cyclohexane. We again
found the completed formation of 4cd as part of a statistical
mixture at room temperature in this solvent after a few
minutes, similar to the experiments in aromatic solvents
(Fig. S10†). Conducting the reaction between 1c and 1d in neat
deuterated THF, however, showed no scrambling between the
molecules at room temperature even after several days (see
Fig. S11† for a spectrum after a few minutes), although we did
find evidence for slow scrambling at 50 °C after many days,
and formation of a near-equilibrium mixture after about one
month (Fig. S12†). Coordination of THF to complexes 1 has
been found to be reversible2j and elevated temperatures are
expected to support dissociation of THF ligands for entropic
reasons.
Scheme 2 Ligand scrambling reactions between complexes 1 and 5
to 4.
We also studied the reaction of the fulleride complex
[{(Mesnacnac)Mg}6C60] 3c with three equivalents of [{(Xylnacnac)
Mg}2] 1d in benzene at room temperature, which showed new several days. In both cases, the formation of [(Mesnacnac)
1H NMR resonances emerging within minutes at room temp- MgMg(Xylnacnac)] 4cd was observed plus the expected
erature. These support the formation of the scrambling pro- by-product [(Arnacnac)Li] 5 and some decomposition to
ducts [(Mesnacnac)MgMg(Xylnacnac)] 4cd, 1c and 1d in essen- [(Arnacnac)2Mg] complexes and Mg metal (Fig. S17 and 18†).
tially statistical distribution after approximately one to two With further progression, small resonances of the newly
days, plus resonances we assign to a mixture of various formed symmetric [{(Arnacnac)Mg}2] complexes 1d or 1c were
[{(Mesnacnac)Mg}n{(Xylnacnac)Mg}6–nC60] (n = 0–6, but likely also evident and show full ligand substitution of the starting
almost no 0 or 6) complexes (Fig. S14–16†). Although this materials. Although a strictly statistical product distribution
ligand exchange reaction is markedly slower than that between was not reached, a ΔG° ≈ 0 for the overall reaction can still
magnesium(I) species 1c and 1d, it could be possible that full- be assumed. A significantly higher activation barrier com-
eride complexes assist in ligand scrambling reactions between pared to the reaction between 1c and 1d is evidenced by
different [{(Arnacnac)Mg}2] 1 complexes on the surface of the the higher required reaction temperatures and longer reac-
C60 cluster. This could be of interest especially for those with tion times. [(Arnacnac)M] M = Na, K complexes were signifi-
sterically more demanding ligands that require long reaction cantly less effective in these reactions due to their low
times and forcing conditions accompanied by decomposition solubility and the formation of metal from decomposition
reactions. We tested the scrambling reaction of 1a and 1c with reactions.
catalytic amounts of C60 both at room temperature and at elev-
To gain further information on the influence of the ligand-
ated temperatures. Not surprisingly, we found no significant type in the dimagnesium(I) compounds for these activation
acceleration of these scrambling reactions to form [(Dipnacnac) reactions, we compared reactions between various [{(Arnacnac)
MgMg(Mesnacnac)] 4ac when catalytic quantities of C60 were Mg}2] compounds 1 (Scheme 1), to those involving the non-
added, though noticed a significant suppression of by-product diketiminate derivative [({Ph2P(NDip)2}Mg)2] 2.4 The overall
formation via the decomposition of [(Arnacnac)MgMg size of the diiminophosphinate ligand in the latter has pre-
(Arnacnac)] species into [(Arnacnac)2Mg] and Mg metal (see viously been suggested to be less than that of the Dipnacnac-
Fig. S2 and S3†). We also used this method successfully to ligand, approximately the same as the Mesnacnac-ligand.4 It
generate [(Dipnacnac)MgMg(Depnacnac)] 4ab (Fig. S1†) as part forms a four-membered chelating ring system with magnesium
of a reaction mixture from dimagnesium(I) compounds 1a and (Fig. 1), and contains a very different and sterically more
1b bearing the most sterically demanding aryl substituents in demanding profile in its backbone region. The scrambling
the series (Scheme 1). The role of the fullerene in the suppres- reaction of [({Ph2P(NDip)2}Mg)2] 2 with [{(Mesnacnac)Mg}2] 1c
sion of the decomposition is not known yet and could involve in deuterated benzene at room temperature proceeded very
the stabilisation of reactive intermediates as part of fulleride slowly and afforded resonances for [{Ph2P(NDip)2}MgMg
complexes.
(
Mesnacnac)] 7c as part of the expected product mixture
Next, we studied the reactions of 2 [(Mesnacnac)Li] 5c (Fig. S19–22†). An approximate statistical distribution of
with [{(Xylnacnac)Mg}2] 1d and 2 [(Xylnacnac)Li] 5d with {Ph2P(NDip)2} and (Mesnacnac) ligands in the products sup-
[{(Mesnacnac)Mg}2] 1c, respectively, Scheme 2. Here, complex 5 ports an overall energy neutral reaction for this system. The
provides a source of anionic ligands together with an inert reaction required approximately ten days and is thus signifi-
s-block metal ion. An elevated temperature of 50 °C was cantly slower than the formation of 4bc from 1b and 1c which
required for the reactions to proceed at a significant rate over took less than one day.
16938 | Dalton Trans., 2019, 48, 16936–16942
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
2+
exchange reactions of anionic ligands on the Mg2 ion via an
associative mechanism (“nucleophilic substitution”) and does
not involve breaking of the Mg–Mg bond, the latter having
already been suggested previously.2f Flexible ionic metal–
ligand interactions with typically fast ligand exchange rates17
are the norm for ionic s-block metal cations and this also
2+
extends to the Mg2
ion. Nucleophilic attack on a
β-diketiminate-coordinated Mg centre is more facile than on a
comparably sized diiminophosphinate-coordinated Mg centre
likely due to the different ligand shape with more open back-
bone fragment of the former. In this respect, the nucleophilic
γ-CH fragment of the β-diketiminate ligands can furthermore
act as another coordination site as demonstrated in the mole-
cular structure of a crowded heteroleptic (Mesnacnac)Mg-
complex, [{{(Mesnacnac)Mg}2{(NCPh)2}}2], where two Mg are
N,N′-chelated by one Mesnacnac-ligand and show a further κ1-C
coordination to the γ-CH fragment of a second Mesnacnac-
ligand.18 Such a coordination interaction may be involved in
an associative intermediate in the anionic ligand exchange
and a similar interaction is not possible for the diiminophos-
phinate derivative 2. The nucleophilic character of this γ-CH
fragment in [{(Arnacnac)Mg}2] 1 is also demonstrated by the
Scheme 3 Ligand scrambling reactions between complexes 2 and 5c,
and 1c and 6.
reaction of
1 with diphenylketene which converts their
β-diketiminate ligands via a nucleophilic γ-CH “attack” on the
diphenylketene carbonyl centre to form Mg–Mg bonded com-
plexes with new tripodal ligands in high yields, leaving the
We also tested the reaction of two equivalents of [(Arnacnac) Mg–Mg unit surprisingly unaffected.2g
Li], Ar = Mes 5c, Xyl 5d, with [({Ph2P(NDip)2}Mg)2] 2 and found
To gain possible insight on the influence of the ligand
that, at 50–60 °C over several weeks, slow substitution of a shape of dimagnesium(I) complexes in reduction reactions; i.e.
{Ph2P(NDip)2} ligand to form [{Ph2P(NDip)2}Li] 6 (31P{1H} if the unique shape of β-diketiminate ligands in [{(Arnacnac)
NMR: −6.7 ppm)16 and [{Ph2P(NDip)2}MgMg(Arnacnac)], Ar = Mg}2] 1 with its open backbone geometry may similarly facili-
Mes 7c, Xyl 7d (31P{1H} NMR: 5.8–5.9 ppm) was achieved tate reactions, we tested reactions of three equivalents of
based on NMR spectroscopy (Scheme 3), although at a severely [({Ph2P(NDip)2}Mg)2] 2 with C60 in deuterated aromatic sol-
suppressed rate compared to those of the related vents at room temperature for comparison. These reactions do
β-diketiminate reactions in Scheme 2 (Fig. S27–36†). The reac- proceed, although significantly more slowly than those of the
tion is accompanied by some decomposition reactions due to β-diketiminate derivatives 1, and eventually form the fulleride
the long reaction times at elevated temperature. In contrast, complex [({Ph2P(NDip)2}Mg)6C60] 8, with full conversion of 2
the reaction of [{Ph2P(NDip)2}Li] 6 with [{(Mesnacnac)Mg}2] 1c only after approximately 2–3 weeks in benzene or six days in
proceeded significantly faster at 50 °C within a few days and at toluene (Scheme 4). As with the formation of complexes
an approximately comparable rate to that of [(Xylnacnac)Li] 5d [{(Arnacnac)Mg}6C60] 3b–d, this solvent dependence is believed
with [{(Mesnacnac)Mg}2] 1c indicating that nucleophilic attack to be due to the lower solubility of C60 in benzene.13
2+
on a β-diketiminate-coordinated Mg2 ion is more facile than
The complex [({Ph2P(NDip)2}Mg)6C60]·4C5H12 8′ could be
that on the diiminophosphinate-coordinated one (Scheme 3) structurally characterised (see Fig. 2) and crystallised from
with comparable overall size (Fig. S23–26†). Although ΔG° may n-pentane with one sixth of the molecule in the asymmetric
deviate from zero for this system, we propose that nucleophilic unit in the trigonal crystal system showing well-ordered
attack of [(Mesnacnac)Li] 5c on [({Ph2P(NDip)2}Mg)2] 2 is kineti- [{Ph2P(NDip)2}Mg]+ units and a severely disordered C606− frame-
cally more hindered (higher ΔG‡) and that the slow rate of work (Fig. 2). In the molecular structure of 8′, six [{Ph2P(NDip)2}
6−
reaction is not caused by an overall thermodynamic penalty for Mg]+ cations sit comfortably around the C60 framework in an
this reaction path. This kinetic hindrance is likely due because approximately octahedral manner and thus the overall diimino-
the {Ph2P(NDip)2} ligand shows a smaller chelating ring size phosphinate ligand size is broadly comparable to that of the
with small (Ar)N⋯N(Ar) ligand “bite”, a higher steric demand smaller β-diketiminate examples (e.g. 3c,d with Ar = Mes, Xyl)13
near the Mg centre (Dip groups) and a more protected back- when the different ligand shape and chelating ring system are
bone region compared to the (Mesnacnac) ligand, despite an disregarded. For comparison, complex 3b (Ar = Dep) is sterically
overall comparable size for the two ligands.
very crowded around the fulleride fragment and not stable in
These series of ligand scrambling experiments suggest that solution. Complex 3a (Ar = Dip) could not be formed due to the
the reactions between dimeric magnesium(I) complexes are large size of the Dipnacnac-ligand and [{(Dipnacnac)Mg}4C60]
This journal is © The Royal Society of Chemistry 2019
Dalton Trans., 2019, 48, 16936–16942 | 16939

View Article Online
Paper
Dalton Transactions
formed instead. Complex 8 shows resonances in 1H and 13C{1H}
NMR spectra for one ligand environment though signals for iso-
propyl methyl groups are extremely broadened at room tempera-
ture (Fig. S37–39†). These gradually sharpen with increased
temperature and show one relatively sharp doublet at 100 °C in
the 1H NMR spectrum (Fig. S40 and 41†). As expected, one
sharp 13C{1H} NMR resonance for the C60
framework is
6−
observed at 154.9 ppm (benzene-d6) or 155.3 ppm (toluene-d8)
at room temperature. Note that these chemical shifts are slightly
2−
closer to β-diketiminate C60 complexes (ca. 156 ppm) than
C606− complexes 3 (ca. 153 ppm).13
That steric effects are important in the activation of C60 by
dimagnesium(I) compounds was demonstrated by the severely
suppressed rate of formation for the diiminophosphinate
species [({Ph2P(NDip)2}Mg)6C60] 8 in comparison to related
β-diketiminate derivatives, and by a competition reaction of a
freshly prepared mixture of three equivalents of [{(Dipnacnac)
Mg}2] 1a and three equivalents of [{(Mesnacnac)Mg}2] 1c with
one equivalent of C60 in an arene solvent. Here, only reso-
nances for the fulleride complex [{(Mesnacnac)Mg}6C60] 3c were
observed soon after addition of the fullerene.13 For reactions
in coordinating solvents, we found that the facile and rapid
ligand exchange reaction between 1c and 1d to 4cd observed
in hydrocarbon solvents is completely suppressed in neat THF
at room temperature, vide supra, due to blocking of the Mg
centres from THF coordination. We previously noted that the
addition of THF to an arene solution of [{(Mesnacnac)Mg}6C60]
3c leads to the precipitation of insoluble, likely ionic
fulleride complexes.13 For context, previously the reaction of
[{(Mesnacnac)Mg}6C60] 3c with six equivalents of [(Xylnacnac)Li]
very rapidly formed the magnesium(II) complex [(Mesnacnac)
Mg(Xylnacnac)] and supports the very facile breaking of the
ionic {(Mesnacnac)Mg}+⋯C606− coordination bonds.13 When we
tested the reaction of three equivalents of [{(Mesnacnac)Mg}2]
1c with C60 in deuterated THF, however, we found that a reac-
tion did occur to an insoluble precipitate within several hours
to one day. This rate is not much slower than that for a reac-
tion in benzene, and especially notable when the poor solubi-
lity of C60 in THF is considered.19 Although we were unable to
identify the product composition as yet, the fact that essen-
tially all the THF-adduct of [{(Mesnacnac)Mg}2] 1c had reacted
and that the small soluble proportion of C60 experience an
excess of the magnesium(I) complex in solution during the
Scheme 4 Synthesis of complex 8.
6−
reaction, led us to believe that a product of the C60 ion is
likely. This reactivity implies that the Mg centres of dimagne-
sium(I) compounds do not require direct coordination to C60
as part of the rate-determining activation step. This may
instead hint at a different mechanism as part of the initial acti-
vation, e.g. electron transfer processes involving a radical inter-
mediate when the species are in close proximity, possibly
transferred via the ligand backbone.
Fig. 2 Molecular structure of 8’ (25% thermal ellipsoids); C60 shown as
spheres. Hydrogen atoms not shown. Only one of twelve sets of dis-
ordered C60 positions is shown. Due to the disorder of the C60 unit, no
Mg–C or C–C interactions are given. Selected bond lengths and angles:
Mg(1)–N(1) 2.0271(19), Mg(1)–N(2) 2.0197(19), P(1)–N(1) 1.6166(18), P(1)–
N(2) 1.6236(17), P(1)–C(13) 1.810(2), P(1)–C(19) 1.819(2), P(1)⋯Mg(1)
2.6391(9); N(2)–Mg(1)–N(1) 75.30(7), N(1)–P(1)–N(2) 99.44(9), C(13)–
P(1)–C(19) 101.87(10).
Conclusions
Exchange reactions of anionic ligands between magnesium(I)
dimers with different supporting ligands likely occur via an
16940 | Dalton Trans., 2019, 48, 16936–16942
This journal is © The Royal Society of Chemistry 2019

View Article Online
Dalton Transactions
Paper
associative process and are strongly dependent on sterics.
These can be very rapid at room temperature as demonstrated
by exchange reactions between [{(Mesnacnac)Mg}2] 1c and
[{(Xylnacnac)Mg}2] 1d to form [(Mesnacnac)MgMg(Xylnacnac)]
4cd. In the chemistry of Zn–Zn-bonded complexes, the
exchange of anionic ligands as part of metathesis chemistry
has been used for the synthesis of new dizinc(I) complexes.8,10a,12
Similarly, anionic ligands can be exchanged between mag-
nesium(I) dimers and lithium complexes of suitable ligands;
the latter process required more forcing reaction conditions in
comparison. These reactions can be used to generate new mag-
nesium(I) dimers bearing two different anionic ligands for
example. The studies furthermore suggest that in addition to
the overall ligand size, the ligand shape is important for these
reactions. We suggest that the open backbone region of the
β-diketiminate ligand facilitates this exchange possibly invol-
ving the “vacant coordination site” of the nucleophilic γ-CH
unit of β-diketiminates. Ligand exchange reactions between
different magnesium(I) dimers catalysed by C60 did not show
significant acceleration though we found a significant suppres-
sion of decomposition reactions from disproportionation reac-
tions. While anionic ligands are exchanged in the uncatalysed
reactions, scrambling of LMg units could in addition be poss-
ible when fullerides are involved.
(c) A. Stasch and C. Jones, Dalton Trans., 2011, 40, 5659–
5672.
2 (a) K. Yuvaraj, I. Douair, A. Paparo, L. Maron and C. Jones,
J. Am. Chem. Soc., 2019, 141, 8764–8768; (b) T. X. Gentner,
B. Rösch, G. Ballmann, J. Langer, H. Elsen and
S. Harder, Angew. Chem., Int. Ed., 2019, 58, 607–611;
(c) S. J. Bonyhady, D. Collis, N. Holzmann, A. J. Edwards,
R. O. Piltz, G. Frenking, A. Stasch and C. Jones, Nat.
Commun., 2018, 9, 3079; (d) J. Li, M. Luo, X. Sheng,
H. Hua, W. Yao, S. A. Pullarkat, L. Xua and M. Ma, Org.
Chem. Front., 2018, 5, 3538–3547; (e) I. Pernik,
B. J. Maitland, A. Stasch and C. Jones, Can. J. Chem., 2018,
96, 513–521; (f) C. Bakewell, A. J. P. White and
M. R. Crimmin, J. Am. Chem. Soc., 2016, 138, 12763–12766;
(g) A. J. Boutland, I. Pernik, A. Stasch and C. Jones, Chem.
– Eur. J., 2015, 21, 15749–15758; (h) R. Lalrempuia,
C. E. Kefalidis, S. J. Bonyhady, B. Schwarze, L. Maron,
A. Stasch and C. Jones, J. Am. Chem. Soc., 2015, 137,
8944–8947; (i) S. J. Bonyhady, C. Jones, S. Nembenna,
A. Stasch, A. J. Edwards and G. J. McIntyre, Chem. – Eur. J.,
2010, 16, 938–955; ( j) S. P. Green, C. Jones and
A. Stasch, Angew. Chem., Int. Ed., 2008, 47, 9079–9083;
(k) S. P. Green, C. Jones and A. Stasch, Science, 2007, 318,
1754–1757.
We also propose that the reduction of C60 with [{(Arnacnac)
Mg}2] 1 is facilitated by the unique shape of β-diketiminate
ligands. In contrast, reactions of C60 with the diiminophosphi-
nate complex [({Ph2P(NDip)2}Mg)2] 2 with a more protected
ligand backbone region are significantly slowed down, and
form the expected structurally characterised fulleride complex
[({Ph2P(NDip)2}Mg)6C60] 8. The fact that six cationic diimino-
phosphinate magnesium complex fragments arrange comfor-
tably around the fulleride is further support that its overall
size is more comparable to that of [{(Mesnacnac)Mg]+ cation.
Coordination of C60 to a Mg centre of a magnesium(I) dimer is
likely not involved in the rate-determining activation step
because a reaction between [{(Mesnacnac)Mg}2] 1c, and C60 still
occurs in THF at room temperature. Steric factors do play a
role in the activation and a close approach of the molecules
without initial Mg⋯C60 coordination may be involved.
3 (a) M. Ma, H. Wang, J. Wang, L. Shen, Y. Zhao, W.-H. Xu,
B. Wu and X.-J. Yang, Dalton Trans., 2019, 48, 2295–2299;
(b) Y. Liu, S. Li, X.-J. Yang, P. Yang and B. Wu, J. Am. Chem.
Soc., 2009, 131, 4210–4211.
4 A. Stasch, Angew. Chem., Int. Ed., 2014, 53, 10200–10203.
5 A. J. Boutland, D. Dange, A. Stasch, L. Maron and C. Jones,
Angew. Chem., Int. Ed., 2016, 55, 9239–9243.
6 (a) J. A. Platts, J. Overgaard, C. Jones, A. Stasch and
B. B. Iversen, J. Phys. Chem. A, 2011, 115, 194–200;
(b) J. Overgaard, C. Jones, A. Stasch and B. B. Iversen, J. Am.
Chem. Soc., 2009, 131, 4208–4209.
7 L.-C. Wu, C. Jones, A. Stasch, J. A. Platts and J. Overgaard,
Eur. J. Inorg. Chem., 2014, 32, 5536–5540.
8 (a) T. Li, S. Schulz and P. W. Roesky, Chem. Soc. Rev., 2012,
41, 3759–3771; (b) E. Carmona and A. Galindo, Angew.
Chem., Int. Ed., 2008, 47, 6526–6536.
9 (a) C. Bakewell, B. J. Ward, A. J. P. White and
M. R. Crimmin, Chem. Sci., 2018, 9, 2348–2356;
(b) H. Banh, K. Dilchert, C. Schulz, C. Gemel, R. W. Seidel,
R. Gautier, S. Kahlal, J. Y. Saillard and R. A. Fischer, Angew.
Chem., Int. Ed., 2016, 55, 3285–3289; (c) J. Hicks,
E. J. Underhill, C. E. Kefalidis, L. Maron and C. Jones,
Angew. Chem., Int. Ed., 2015, 54, 10000–10004.
10 (a) H. Banh, C. Gemel, R. W. Seidel and R. A. Fischer,
Chem. Commun., 2015, 51, 2170–2172; (b) S. Schulz,
D. Schuchmann, I. Krossing, D. Himmel, D. Bläser and
R. Boese, Angew. Chem., Int. Ed., 2009, 48, 5748–5751.
11 (a) A. Grirrane, I. Resa, A. Rodriguez, E. Carmona,
E. Alvarez, E. Gutierrez-Puebla, A. Monge, A. Galindo,
D. del Rio and R. A. Andersen, J. Am. Chem. Soc., 2007, 129,
693–703; (b) I. Resa, E. Carmona, E. Gutierrez-Puebla and
A. Monge, Science, 2004, 305, 1136–1138.
Conflicts of interest
The authors declare no conflict of interest.
Acknowledgements
This work was supported by the University of St Andrews and
the EPSRC (PhD studentship for SRL; EP/N509759/1).
References
1 (a) C. Jones, Nat. Rev. Chem., 2017, 1, 0059; (b) C. Jones and
A. Stasch, Top. Organomet. Chem., 2013, 45, 73–101;
This journal is © The Royal Society of Chemistry 2019
Dalton Trans., 2019, 48, 16936–16942 | 16941

View Article Online
Paper
Dalton Transactions
12 S. Gondzik, D. Bläser, C. Wölper and S. Schulz, Chem. – 16 A. L. Hawley and A. Stasch, Eur. J. Inorg. Chem., 2015, 258–
Eur. J., 2010, 16, 13599–13602. 270.
13 S. R. Lawrence, C. A. Ohlin, D. B. Cordes, A. M. Z. Slawin 17 L. Helm and A. E. Merbach, Coord. Chem. Rev., 1991, 187,
and A. Stasch, Chem. Sci., 2019, DOI: 10.1039/C9SC03857D. 151–181.
14 S. Jain and K. Vanka, Chem. – Eur. J., 2017, 23, 13957–13963. 18 M. Ma, A. Stasch and C. Jones, Chem. – Eur. J., 2012, 18,
15 S. Brand, H. Elsen, J. Langer, W. Donaubauer, F. Hampel 10669–10676.
and S. Harder, Angew. Chem., Int. Ed., 2018, 57, 19 R. S. Ruoff, D. S. Tse, R. Malhotra and D. C. Lorents,
14169–14173.
J. Phys. Chem., 1993, 97, 3379–3383.
16942 | Dalton Trans., 2019, 48, 16936–16942
This journal is © The Royal Society of Chemistry 2019