Loops, chains, sheets, and networks from variable coordination of cu(hfac)2 with a flexibly hinged aminoxyl radical ligand

Article abstract of DOI:10.1021/ic200362c

One pair of reactants, Cu(hfac)₂ = M and the hinge-flexible radical ligand 5-(3-N-tert-butyl-N-aminoxylphenyl)pyrimidine (3PPN = L), yields a diverse set of five coordination complexes: A cyclic loop M₂L ₂ dimer; a 1:1 cocrystal between an M₂L₂ loop and an ML₂ fragment; a 1D chain of M₂L₂ loops linked by M; two 2D M₃L₂ networks of (M-L)n chains crosslinked by M with different repeat length pitches; a 3DM₃L ₂ network ofM2L2 loops cross-linking (M-L)n-type chains with connectivity different from those in the 2D networks. Most of the higher dimensional complexes exhibit reversible, temperature-dependent spin-state conversion of high-temperature paramagnetic states to lower magnetic moment states having antiferromagnetic exchange within Cu-ON bonds upon cooling, with accompanying bond contraction. The 3D complex also exhibited antiferromagnetic exchange between Cu^II ions linked in chains through pyrimidine rings.

Full text of DOI:10.1021/ic200362c

ARTICLE
pubs.acs.org/IC
Loops, Chains, Sheets, and Networks from Variable Coordination
of Cu(hfac) with a Flexibly Hinged Aminoxyl Radical Ligand
2
†
‡
‡
†
§
Martha Baskett, Armando Paduan-Filho, Nei Fernandes Oliveira, Jr., A. Chandrasekaran, Joel T. Mague,
,†
and Paul M. Lahti*
†
Department of Chemistry, University of Massachusetts, Amherst, Massachusetts 01003, United States
‡
Instituto de Fisica, Universidade de S ~a o Paulo, Caixa Postal 66.318, 05315-970 S ~a o Paulo, Brazil
§
Department of Chemistry, Tulane University, New Orleans, Louisiana 70118, United States
S Supporting Information
b
ABSTRACT: One pair of reactants, Cu(hfac) = M and the hinge-flexible
2
radical ligand 5-(3-N-tert-butyl-N-aminoxylphenyl)pyrimidine (3PPN = L),
yields a diverse set of five coordination complexes: a cyclic loop M L dimer; a
2
2
1
:1 cocrystal between an M L loop and an ML fragment; a 1D chain of
2 2 2
M L loops linked by M; two 2D M L networks of (MꢀL) chains cross-
linked by M with different repeat length pitches; a 3D M L network of M L
2 2
2
2
3
2
n
3
2
loops cross-linking (MꢀL) -type chains with connectivity different from
n
those in the 2D networks. Most of the higher dimensional complexes exhibit
reversible, temperature-dependent spin-state conversion of high-temperature
paramagnetic states to lower magnetic moment states having antiferromag-
netic exchange within CuꢀON bonds upon cooling, with accompanying
bond contraction. The 3D complex also exhibited antiferromagnetic ex-
II
change between Cu ions linked in chains through pyrimidine rings.
’
INTRODUCTION
than 4PPN and 4TPN. By varying the reaction conditions using
PPN and Cu(hfac) , we obtained new 1D, 2D, and 3D
coordination solids, which are reported in this article.
3
2
One strategy for the design and synthesis of molecular
magnetic materials is to coordinate paramagnetic transition-
metal ions with organic openꢀshell molecules to make “hybrid”
mixtures of spin sites. Nitronylnitroxides, aminoxyls, and verda-
zyls have all been coordinated into varying structural motifs and
magnetic behaviors, combining the strong magnetic moments of
transition-metal ions with diverse structural units from organic
’
RESULTS
PPN was synthesized by our previous procedure (Scheme 2).
6
3
2,7
Unlike isomer 4PPN, 3PPN decomposes upon extended sto-
rage, so it is best to store precursor 3PPNH and make 3PPN just
before use. Fortunately, the 3PPN solution stability was sufficient
for slow crystallization of layered solutions of it with Cu(hfac)2.
Five new complexes shown in Table 1 were obtained: a M L ML
2
1:1 cocrystal M3L4, a 1D polymer Chain, two different 2D net-
works Net2syn and Net2anti, and a 3D network Net3d. The pro-
perties of the previously reported cyclic M2L2 copper(II) com-
1
chemistry. This allows many possible structural types and
magnetic exchange pathways.
We have described complexation of dicationic transition-
metal salts M(hfac) = M (hfac = hexafluoroacetylacetonate)
with radical ligands (L) 4PPN and 4TPN to give M L cyclic
dimers shown in Scheme 1. For higher reacting ratios of 4PPN to
M(hfac) , ML linear triads were formed. Only Co(hfac) with
2
2
2
3
2
3
2
2
4
2
2
2
6
4
PPN gave an extended, 1D ribbon polymer with stoichiometry
3
plex are also compared in the discussion below. All structures were
5
M L . Higher dimensional coordination is an important goal of
identified by single-crystal X-ray diffraction (XRD) analyses.
M3L4 and Chain were only obtained in small amounts and
not extensively analyzed. Net2syn, Net2anti, and Net3d were
more readily obtained and were subjected to both room tem-
perature and 100 K XRD crystallography, to dc magnetic
susceptibility (χ vs T) studies at 1.8ꢀ300 K, and to magnetiza-
tion versus field experiments (M vs H) at 0.5ꢀ1.4 K. The results
are detailed in various sections below.
2
research to form multidimensional exchange networks in molec-
ular magnetic materials. 1D chain systems are fairly common, 2D
network systems less so, and 3D systems probably the most
challenging among complexes of transition metals with radical-
bearing ligands.
We tested 5-(3-[N-tert-butyl-N-aminoxyl]phenyl)pyrimidine
6
(
3PPN) as a “hinge-flexible” radical ligand and reported its
2
M L -type complexes with M(hfac) (M = Mn, Co, Ni, Cu). The
2
2
hinge flexibility was specifically designed to allow 3PPN more
conformational freedom to form extended conjugation networks
Received: February 21, 2011
Published: April 27, 2011
r 2011 American Chemical Society
5060
dx.doi.org/10.1021/ic200362c
|Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
’
DISCUSSION
possibly extended connectivity was a structurally ill-characterized
10a
product of Mn(hfac) with 4PyrN.
2
2
3
Design Strategy for Extending Radical Ligand Coordina-
tion Networks. In addition to the above-mentioned cyclic M2L2
complexes of various M(hfac) with 4PPN, 4PTN, and 3PPN,
analogous structures have been reported for complexation of
Mn(hfac) and Cu(hfac) with 4PImN, 3PImN, 4PyrN,
In the M L complexes of 4PPN and 4PTN with Mn(hfac)
2
2
2
or Cu(hfac) , one azacycle nitrogen remained uncomplexed.
2
5
3 2 2
2
Only the M L ladder ribbon polymer from Co(hfac) and
4PPN showed extended coordination. A possible hindrance to
8
,9
8,9
9,10
2
2
forming extended networks from 4PPN and 4PTN is their
limited conformational variability. 3PPN can accommodate both
linear and sharply bent extensions across the pyrimidine unit
through anti and syn aminoxyl conformers (Scheme 3). 3PPN is
hinged to accommodate planar chain extension, tight helical
chain extension, and loop formation, among other possibilities.
Yamada et al. have explicitly noted the structural richness
9
,10
11
and 3PyrN (Chart 1). Among these, the only complex with
Scheme 1. Lower Dimensional Complexes of Radical
Ligands 4PPN, 4TPN, and 3PPN with Cu(hfac)2
1
2
possible in magnetic complexes with hinged ligands.
The structural diversity of Table 1 products shows how the
hinged radical ligand 3PNN provided more possibilities for
complexation by comparison to isomer 4PNN. This causes
complications because product formation was sensitive to Cu-
(
hfac) /3PPN reactant ratios and to the solvent types and
2
amounts. Sometimes it was necessary to remove coprecipitated
Cu(hfac) manually from the desired product. Multiple phases
2
occasionally cocrystallized from a reaction mixture but could be
manually separated based on their different appearances. The use
of XRD was important to check the unit cell parameters of
different batches of product and establish identity.
Although no systematic effort was made to determine the
mechanisms of product formation, a tentative proposal is given in
Scheme 4, based on the structural features found in the Table 1
complexes. An initially formed MLN intermediate could aza-
complex to another 3PPN to give ML intermediates. Neither
2
linear nor bent ML complexes were isolated for L = 3PPN, but
2
both structures are part of the complexes in Table 1. M L
2
2
5
systems could form by condensing two MLN or MLO inter-
mediates [presumably azaphilic Cu(II) favors MLN]. All pro-
cesses are shown as equilibria because recrystallizing initially
isolated products occasionally yielded other complexes.
Cu (3PPN) (hfac) Cu(3PPN) (hfac) (M3L4). This cluster
2
2
4
3
2
2
is a 1:1 cocrystal of a Cu (3PPN) (hfac) M L loop with a
2
2
4
2 2
Cu(3PPN) (hfac) = ML fragment. Although the ML frag-
2
2
2
2
ment was never isolated alone, its incorporation into M3L4
suggests that it and M L are simultaneously present in signifi-
2
2
cant amounts during crystallization, consistent with Scheme 4.
Figure 1 shows a structure diagram of M3L4 (ORTEP
13
diagrams are given in the Supporting Information). Table 2
Scheme 2. Synthesis of 3PPN
5
061
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
Table 1. Complexes of Cu(hfac) with 3PPN [M = Cu(hfac) ]
2
2
summarizes its XRD parameters; Figure 2 focuses on its copper(II)
uncoordinated pyrimidine N2 of an M L unit (Figure 1);
2
2
ligand spheres. At all points in this article, “CuꢀON” will refer to a
r(C55 N2) = 3.621(7) Å. A riding model estimate of the
3
3 3
copperꢀaminoxyl bond involving the aminoxyl oxygen. The ML
hydrogen atom on C55 puts it about 2.7 Å from N2. Although
2
fragment has a somewhat distorted octahedral copper(II) coordina-
the small isolated amount of M3L4 precluded magnetic measure-
tion, with 3PPN units in the “bent” ML geometry of Scheme 4.
ments, the short CuꢀON bond lengths in its M L loops are
2
2 2
comparable to those for aminoxyls that are equatorially complexed
The M L loop is centrosymmetric, having both 3PPN units in the
2
2
II
14
syn-aminoxyl conformations required to form a loop. The
Cu1ꢀO5N bond is fairly short and thus equatorial. The equatorial
pyrimidine Cu1ꢀN1 bond in the loop is shorter than the Cu2ꢀ
N31 or Cu2ꢀN34 bonds in the ML unit. The ML and M L
to Cu and exhibit antiferromagnetic (AFM) CuꢀON exchange,
including the low-temperature form of M2L2 discussed below.
Cu (3PPN) (hfac) (M2L2). We reported formation of the
2
2
4
6
2
2
2 2
M L complex of Cu(hfac) with 3PPN previously. Figure 3
2
2
2
fragments are associated in the M3L4 cocrystal by favorable inter-
shows its structure, and Table 2 gives crystallographic details.
action of the C55ꢀH bond of a ML unit with the otherwise
M2L2 is structurally similar to the loops in M3L4 and (as we
2
5
062
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
shall see) Chain and Net3d. M2L2 has two distinct forms in the
lattice that exhibit reversible coordination sphere changes with
the temperature (Figure 4). At room temperature, both forms
have much longer CuꢀON bond lengths than the analogous
loop fragments of M3L4 and Net3d. Cooling to 100 K causes the
axial CuꢀON bonds to contract and change to equatorial, with
corresponding color changes from brownish at room tempera-
ture to green at low temperatures. As we shall see, most of the
network complexes of Table 1 also exhibit temperature-depen-
dent CuꢀON bond length changes with their magnetism. M2L2
is thus a model for behavior in the more complex structures.
Both χ and χT vs T plots for M2L2 steadily decrease from
room temperature to minima at about 100 K, as shown in
Figure 5. If all four possible spins in M2L2 were independent,
for ferromagnetic (FM) exchange in the CuꢀON bonds, the
plateau would be at ∼2.0 emu K/Oe mol. The room tempera-
3
3
ture χT is only about half the free-spin value, with no plateau in
the upper temperature region. So, at least half of the M2L2 lattice
already has some CuꢀON AFM spin pairing at room tempera-
ture, presumably form B in Figure 4 with its shorter CuꢀON
bonds. At lower temperature, both forms of M2L2 lose the
paramagnetic moment with CuꢀON bond contraction. The
magnetization versus field data for M2L2 show much less than
one spin per mole of M2L2 (consistent with the χ and χT
results). We used a 10 000 Oe field to increase the signal for the
susceptibility measurements; the magnetization versus field
dependence is still linear in this range at 1.25 K, so the Figure 5
data are still in the weak-field regime.
χT at high temperatures would plateau at ∼1.5 emu K/Oe mol;
Whether the spins in the M2L2 complex are nearly indepen-
dent or FM-coupled at higher temperatures, the decreases in χ
and χT with the temperature are hallmarks of spin-state
3
3
Chart 1
15
conversion to a low-spin state. The transition occurs over a
fairly broad temperature range rather than a sudden, synergistic
change of molecules in all lattice sites. The conversion from a
paramagnetic state with longer CuꢀON bond lengths to low-
spin AFM exchange at shorter bond lengths is attributable to
changes in the overlap between the magnetic orbitals of octa-
hedral copper(II) and aminoxyl: details have been given else-
1f,6,16
where.
In principle, the magnetic changes could be due
simply to thermal depopulation of paramagnetic excited states in
the CuꢀON bonds, but we feel that correlation of the geometry
changes with the color and magnetic changes supports spin-state
conversion.
Scheme 3. Complexation Geometries Accommodated by
PPN and 3PPN
4
Complex [Cu (3PPN) (hfac) ] (Chain). The Chain complex
3
2
6 n
structure is shown in Figure 6; crystallographic details are given in
Table 1. Chain incorporates M L loops linked by Cu(hfac) to
2
2
2
form 1D chains, such that its 3PNN units are fully coordinated to
give stoichiometry M L . Its M L loops are structurally similar to
3
2
2 2
those in M3L4 and M2L2, with somewhat distorted octahedral
Cu1 ligand spheres. The loop Cu1ꢀO7N bonds are intermediate
in length among those in this study (Figure 7, 1.99 Å).
The chain link Cu2ꢀN3 bonds in Chain are significantly
longer than the Cu1ꢀN bonds of the loop unit or those in the
analogous ML fragment of M3L4: r(Cu2ꢀN3) = 2.27 Å
2
(
Figure 7) versus 2.03ꢀ2.04 Å (Figure 4) in M3L4. The chain
a
Scheme 4. Tentative Mechanistic Scheme for 3PPN Complexation
a
Bracketed structures are putative.
5
063
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
link is bent with —N3ꢀCu2ꢀN(3_2) = 82.9ꢀ. The Cu2ꢀO5-
chain packing are given in the Supporting Information. Unfortu-
nately, insufficient Chain was obtained for reliable magnetic
analysis. However, its short CuꢀON bond lengths are similar to
those of M L form B at 100 K (compare Figures 4 and 7). As we
(
hfac) bonds are contracted and short in the ML unit of Chain,
2
whereas the analogous CuꢀO(hfac) bonds are long and axial in
the ML unit of M3L4.
2
2 2
The chains of Chain pack along the crystallographic c axis with
shall see below, coordinating Cu L loops into larger network
solids seems to encourage bond shortening in CuꢀON of
the loops.
2 2
contact between the hfac CF groups on the periphery of each
3
chain axis, at an interchain distance of 17.5 Å, which should limit
3
PPN and Cu(II) spins to intrachain exchange. Views of the
2D Sheet Complex [Cu (3PPN) (hfac) ] (Net2syn). The
3
2
6 n
2
D sheet complexes Net2syn and Net2anti have the same bond
connectivity but different secondary structure and magnetic
properties. In both, (Cu PyrimPhNO )n chains leave a
3
3 3
3 3 3
pyrimidine nitrogen atom of each 3PPN for Cu(hfac) cross-
2
linking to a nearby chain. However, their chain geometries are
quite different.
The structure of Net2syn is shown in Figures 8 and 9, and its
crystallographic parameters are given in Table 3. Its (Cu
3
3 3
PyrimPhNO ) chains are helices wherein 3PPN units have a
3
3 3
n
syn-aminoxyl to pyrimidine relationship, similar to 3PPN in the
M L loop units of M3L4, M2L2, and Chain. If one imagines
2
2
twisting an M L -type loop at a CuꢀON bond to form an “S”-
2
2
shape, that would give the helical repeat unit of the Net2syn
chains. Within the chains, the Cu1ꢀO7N bond length is 1.97 Å
(Figure 9), and the O4ꢀCu1ꢀO1(hfac) bonds are the longest,
Figure 1. Structure diagram of the 1:1 cocrystal molecular complex
M3L4.
Table 2. Crystallographic Data for Molecular and 1D Chain Complexes of Cu(hfac) with 3PPN
2
M3L4
M2L2
Chain
3 36 6 14
38Cu F N O
chemical formula
fw
C
124
H
104Cu
4
F
48
N
18
O
22
C
48
H
36Cu
2
F
24
6
N O
10
C
58
H
3364.41
1439.91
1917.6
298
temperature (K)
space group
a (Å)
293
298 [100]
triclinic, P1
11.4071(1)
18.0922(2)
19.2627(3)
77.4891(4)
80.8945(4)
79.3408(7)
3785.01(8)
1
triclinic, P1
monoclinic, C2/c
32.9455(4)
11.4448(2)
23.3160(3)
90.0
9.6621(2) [9.44230(10)]
16.0241(3) [15.7864(2)]
19.6090(5) [19.5298(4)]
76.5851(8) [76.6127(6)]
81.6545(8) [81.7446(6)]
82.4981(17) [81.6369(8)]
2907.04(11) [2769.31(7)]
2 [2]
b (Å)
c (Å)
R (deg)
β (deg)
119.7464(7)
90.0
γ (deg)
3
V (Å )
7632.96(19)
4
Z
3
cryst dimens (mm )
0.50 ꢁ 0.50 ꢁ 0.20
1.476
1.00 ꢁ 0.75 ꢁ 0.10
1.868
1.00 ꢁ 0.50 ꢁ 0.50
1.669
3
D
calc (Mg/m )
θ range (deg)
F(000)
4.10ꢀ25.08
1698
4.08ꢀ25.09
4.09ꢀ25.01
3804
1440
ꢀ
1
μ (mm
)
0.561
0.978 [0.991]
0.978
data collection method
collected/unique reflns
criterion obsd reflns
ωꢀ2θ scans
7823/3982
I > 2σ(I)
ωꢀ2θ scans
16 661/10 141 [9714/7671]
I > 2σ(I)
ωꢀ2θ scans
12 452/6642
I > 2σ(I)
R
int
0.0208
0.0266 [0.0361]
0.0221
ranges of h, k, l
ꢀ12 f h f þ13
ꢀ11 f h f þ11 [ꢀ11 f h f þ10]
ꢀ19 f k f þ19 [ꢀ18 f k f þ18]
ꢀ23 f l f þ21 [ꢀ22 f l f þ22]
0.978 [0.986]
ꢀ38 f h f þ39
ꢀ12 f k f þ13
ꢀ27 f l f þ27
0.985
ꢀ
21 f k f þ121
22 f l f þ21
ꢀ
0.985
completeness to 2θ
no. of reflns
13 265
10 202 [9714]
6642
no. of restraints
828
0 [0]
0
no. of param
970
799 [799]
552
2
GOF on F
1.034
1.026 [1.029]
1.028
3
ΔFmax, ΔFmin (e/Å )
R1, wR2 [I > 2σ(I)]
R1, wR2 (all)
0.943, ꢀ0.540
0.0753, 0.2102
0.0953, 0.2284
0.930, ꢀ0.852 [1.053, ꢀ0.977]
0.0798, 0.2001 [0.0522, 0.1234]
0.1029, 0.2189 [0.0702, 0.1336]
0.905, ꢀ0.569
0.0869, 0.2362
0.1074, 0.2590
5
064
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
Figure 2. Coordination sphere bonding for M3L4. Bond lengths are in angstroms and angles in degrees. Long bond ligation sites are underlined.
Figure 3. Structure diagram for the M2L2 molecular complex of
Cu(hfac) with 3PPN.
2
Figure 5. χT (left) and χ (right) vs T plots for the M
0 000 Oe.
2 2
L complex at
1
Figure 6. Structure diagram for 1D polymer complex Chain.
longer than the intrahelix Cu1ꢀN1 bonds and, therefore, axial.
The hfac oxygen atoms in the Cu2 coordination sphere are
equatorial with short lengths (Figure 9). Upon cooling of single
crystals to 100 K, the helix repeat distance decreases somewhat
from 12.1 to 11.7 Å, but the ligand spheres of Cu1 and Cu2 do
not change much. The 2D networks in Net2syn stack as roughly
planar slabs, with interslab contacts between the hfac CF groups
on the peripheries above and below, as shown in Scheme 5 (with
details in the Supporting Information).
Figure 4. Ligand-sphere bonding parameters for M2L2 at 299 and 100
K. Bond lengths are in angstroms and angles in degrees. Long bond
ligation sites are underlined.
3
The magnetic susceptibility of Net2syn shows essentially Curieꢀ
Weiss behavior from room temperature to 1.8 K (Figure 10), with
C = 1.04 emu K/Oe mol and θ = (ꢀ)0.5 K. The χΤ data have a
although bent from an ideal trans diaxial angle with —OꢀCuꢀO
3
3
=
169ꢀ. The Cu1ꢀN1 pyrimidine bond length along the chains
high temperature plateau of ∼1.0 emu K/Oe mol down to about
3
3
is 2.03 Å, similar to that in the previously discussed M L
2
2
15 K. The decrease in χΤ at lower temperatures indicates only weak
AFM exchange interactions between spin sites, consistent with the
small Weiss constant.
loop units.
Neighboring helices propagate in alternating directions, cross-
linked across pyrimidine nitrogen atoms of 3PPN by Cu2 to
form 2D networks. The Cu2ꢀN3 cross-links of 2.48 Å are much
Magnetization versus field studies at 0.5 K show a step at
∼1 μ /mol (20 000 Oe) and then an increase to a bit above
B
5
065
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
Figure 7. Ligand-sphere bonding parameters for Chain at room temperature. Bond lengths are in angstroms and angles in degrees. Long bond ligation
sites are underlined.
Table 3. Crystallographic Parameters for 2D Complex
Net2syn
temperature (K)
chemical formula
fw
298
100
C
58
H
38Cu
3
F
36
6
N O
14
C
58
H
38Cu
3 36 6 14
F N O
1917.56
1917.56
cell setting, space group
a (Å)
monoclinic, P2
15.1227(2)
12.0542(1)
20.8771(3)
90.
1
/n
monoclinic, P2
14.9104(2)
11.68780(10)
20.5148(2)
90.
1
/n
b (Å)
c (Å)
R (deg)
Figure 8. Structure diagram for 2D network complex Net2syn.
β (deg)
93.1804(5)
90.
92.6405(5)
90.
γ (deg)
3
V (Å )
3799.87(8)
2
3571.31(7)
2
Z
3
Dcalc (g/cm )
1.676
1.783
θ range (deg)
F(000)
4.11ꢀ25.13
1902
4.18ꢀ25.03
1902
ꢀ
1
μ (mm
)
0.982
1.045
data collection method
collected/unique reflns
criterion for obsd reflns
ωꢀ2θ scans
12 445/6646
I > 2σ(I)
0.0211
ωꢀ2θ scans
10 546/6208
I > 2σ(I)
0.0241
R
int
ranges of h, k, l
ꢀ17 f h f þ17
ꢀ17 f h f þ17
ꢀ12 f k f þ13
ꢀ24 f l f þ24
0.998
ꢀ
14 f k f þ12
24 f l f þ24
ꢀ
0.981
completeness to 2θ
reflns/restraints/param
6646/0/529
1.030
6208/0/529
1.144
2
GOF on F
3
ΔFmax, ΔFmin (e/Å )
0.717, ꢀ0.482
0.0562, 0.1474
0.0751, 0.1665
1.820, ꢀ0.914
0.0499, 0.1390
0.0561, 0.1422
R1, wR2 [I > 2σ(I)]
R1, wR2 (all)
seemed consistent with AFM pairing of Cu1 and aminoxyl
14
spins, leaving the cross-linker Cu2 ions as the main source of
the magnetic moment at low temperature. To probe the possi-
bility that Net2syn was significantly contaminated with other
Figure 9. Ligand-sphere bonding parameters for Net2syn at 293 and
100 K. Bond lengths are in angstroms and angles in degrees. Long bond
ligated atoms are underlined.
products of the reaction between Cu(hfac) and 3PPN, bulk
2
powder XRD studies were carried out at room temperature. The
experimental wide-angle X-ray scattering pattern is a good match
for that expected from the single-crystal XRD of Net2syn
2
μB/mol at 50 000 Oe and above. These results indicate at least
two distinct spin carriers. The short CuꢀON bonds in Net2syn
5
066
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
Scheme 5. Schematic Representation of a 2D Slab in Net2syn
(
(
above, View of One Slab from above) and Interslab Packing
below, Edge-On View)
a
Figure 11. Structure diagram for 2D network complex Net2anti at
room temperature.
Scheme 6. Schematic of 2D Corrugated Sheet in Net2anti
(
(
above, View of One Slab from above) and Sheet Packing
below, Edge-On View)
a
a
Pyrimidine nitrogen (NN) and aminoxyl oxygen (O) atoms are shown
to indicate alignment of the 3PPN units. Helical chains propagate along
solid bonds; cross links are dashed bonds. The lower figure shows
interslab packing, with fluorine-rich regions shown by F.
a
In the upper view, pyrimidine nitrogen (NN) and aminoxyl oxygen
(O) atoms are shown to indicate alignment of the 3PPN units. Zigzag
chains propagate along the solid bonds; cross-links are dashed bonds.
The lower figure shows interslab packing, with fluorine-rich regions F
and hydrogen-rich regions H.
side and all 3PPN groups syn on the opposite side, giving zigzag
chains with a repeat length of 19.6 Å at room temperature.
Net2syn has hfac and 3PPN groups on alternating sides of its
chains, and its 12.1 Å helix repeat distance is much shorter. In
Net2anti, the aminoxyl groups are anti to the 3PPN pyrimidine
group (not syn, as in Net2syn and other complexes discussed to
this point). However, Net2anti interchain 3PPNꢀCu2ꢀ3PPN
cross-links are quite similar to those in Net2syn. Overall, the
Net2anti networks look like corrugated sheets rather than the
planar slabs of Net2syn. The sheets in Net2anti form close
Figure 10. Paramagnetic susceptibility and magnetization data for
Net2syn. Plot a shows χT data vs T, and plot b shows 1/χ vs T, both
obtained at 100 Oe external field. The solid line in plot b is the
CurieꢀWeiss fit to data <50 K. Plot c shows the magnetization versus
field data at 0.54 K.
(Supporting Information). Net2syn in the bulk seems to be all or
almost only the structure found by multiple single-crystal XRD
analyses in this study. As described in the following section, both
structure and magnetism are quite different in the allotropic
Net2anti samples.
contacts between hfac CF groups on the convex faces of the
3
sheet and phenyl CH groups of the 3PPN units in the concave
faces, as shown in Scheme 6 (see also the Supporting In-
formation). With all hfac units on one side of each chain, rather
than spiraling around as in the helices of Net2syn, the plane-to-
plane intersheet distance in Net2anti is only 8.7 Å at room
temperature, much less than that in Net2syn.
2
D Sheet Complex [Cu (3PPN) (hfac) ] (Net2anti). As
3
2
6 n
mentioned above, Net2anti and Net2syn have the same con-
nectivity but different coordination geometries and radical ligand
conformation. Figure 11 shows the Net2anti structure, in which
the (Cu 3PNN )n chains have all hfac ligands syn on one
3
3 3
3 3 3
5
067
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
Table 4. Crystallographic Parameters for 2D Complex Net2anti
temperature (K)
chemical formula
fw
298
173
100
C
58
H
38Cu
3
F
36
6
N O
14
C
58
H
38Cu
1917.56
monoclinic,
P2 /c
3
F
36
N
6
O
14
C
58
H
38Cu
1917.56
monoclinic,
P2 /c
3 36 6 14
F N O
1917.56
cell setting, space group
orthorhombic, Pbca
1
1
a (Å)
19.6467(13)
17.4889(12)
22.7491(6)
90.0
16.545(3)
22.703(3)
19.720(3)
90.0
16.402(2)
22.618(3)
9.711(2)
90.0
b (Å)
c (Å)
R (deg)
β (deg)
γ (deg)
90.0
94.231(2)
90.0
92.438(2)
90.0
90.0
3
V (Å )
7834.71(16)
4
7387.2(19)
4
7290(1)
4
Z
D
calc
1.634
1.724
1.747
3
(
g/cm )
θ range
2.07ꢀ26.44
2.07ꢀ26.41
1.25ꢀ27.14
(
deg)
F(000)
μ (mm
3804
3804
3804
ꢀ1
)
0.957
1.010
1.024
data collection method
collected/unique reflns
criterion for obsd reflns
ωꢀ2θ scans
116 959/8015
I > 2σ(I)
0.0687
ωꢀ2θ scans
ωꢀ2θ scans
156 975/157 398
I > 2σ(I)
82 396/82 115
I > 2σ(I)
R
int
0.000
ranges of h, k, l
ꢀ24 f h f þ24
ꢀ20 f h f þ20
ꢀ28 f k f þ28
ꢀ24 f l f þ24
157 398/0/1064
0.756
ꢀ20 f h f þ21
ꢀ29 f k f þ29
ꢀ25 f l f þ25
82 396/0/1064
0.883
ꢀ
21 f k f þ21
28 f l f þ28
ꢀ
reflns/restraints/param
8015/74/574
1.041
2
GOF on F
3
ΔFmax, ΔFmin (e/Å )
0.617, ꢀ0.368
0.0759, 0.2445
0.1348, 0.3115
1.653, ꢀ0.633
0.0541, 0.0913
0.1385, 0.1086
1.820, ꢀ0.914
0.0540, 0.0948
0.1060,0.1078
R1, wR2 [I > 2σ(I)]
R1, wR2 (all)
Net2anti has an orthorhombic Pbca space group at room
temperature, but upon cooling, it undergoes twinning with a
upon cooling (attributed to spin-state conversion due to bond
shortening).
phase change to monoclinic P2 /c (see Table 4). The twinned
3D Network Complex [Cu (3PPN) (hfac) ] (Net3d). Com-
1
3
2
6 n
structure units are quite similar (Figure 12 and Supporting
Information) and readily revert to the Pbca phase upon rewarm-
ing. Other examples of reversible phase changes with transforma-
tion from higher to lower crystallographic symmetry accom-
panied by nonmerohedral twinning have been reported. The
bulk phase change in Net2anti correlates with changes in the
Cu1ꢀON ligand sphere (Figure 12). At room temperature, the
Cu1ꢀON bond and trans-related CuꢀO(hfac) bond are long
and axial. Cooling to 100 K shortens both bonds: two other trans-
related CuꢀO(hfac) bonds lengthen to become new axial
substituents.
plex Net3d is structurally the most intricate of the Table 1
complexes, incorporating chains and loops into a highly cross-
linked 3D network, with multiple types of CuꢀON and
Cuꢀpyrimidine coordination. Unlike the behavior in the 2D
networks, the aminoxyl groups in Net3d do not participate in
chain formation but instead only form loops and cross-links.
Instead, Net3d chains come from alternating bent and linear
17
PyrimꢀCuꢀPyrim connections in 1D (Cu pyrimidine)n
3 3 3
chains. Table 5 gives its crystallographic parameters; Figure 14
shows a simplified schematic of the connectivities and 3PNN
conformations in Net3d and the 3D connection of loops and
chains. The 3D lattice has ellipsoidal void channels of about 9 Å
The χT vs T plot for Net2anti (Figure 13) decreases from
(
a axis) by 16 Å (b axis) along the crystallographic c axis
∼
0.6 emu K/Oe mol at room temperature to a minimum of 0.4
3
3
(
Figure 15). Hexane solvent incorporation in these may help
emu K/Oe mol at 200 K. CurieꢀWeiss analysis from the 1/χ vs
3
3
template the surprisingly ready formation of Net3d.
T plot yields C = 0.489 emu K/Oe mol, where T < 50 K, with
3
3
The (Cu pyrimidine) chains have pendant phenylaminox-
1
3 3 3
n
θ = (þ)0.3 K. The Curie constant corresponds to one S = / per
2
yl groups (part of 3PPN molecules) with aminoxyls anti to the
pyrimidine (Figure 14) and radiating around the chain. The
chain repeat length is about 14 Å long. These chain-pendant
aminoxyls cross-link through relatively long [r(Cu1ꢀO1[N]) =
2.195(4) Å] bonds to M L loop units (red and yellow in
mole unit with g = 2.28. The magnetization versus field data at
eff
0
.55 K show fairly rapid saturation at 1.07 μ , consistent with one
B
1
S = / per mole unit, presumably from the Cu2 ions alone. The
2
lack of magnetic contribution from the Cu1ꢀON bond at lower
2
2
temperature occurs as a result of loss of the spin moment
Figure 14) to other chains. The M L loops of Net3d are
2 2
5
068
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
Figure 13. Paramagnetic susceptibility and magnetization data for
Net2anti. Plot a shows χT data vs T, and plot b shows 1/χ vs T both
obtained at 500 Oe external field. The solid line in plot b is the
dC aut ra i ea ꢀt 0W. 5 e5 is Ks .fit to data <50 K. Plot c shows magnetization versus field
no paramagnetic contribution. However, the long Cu1ꢀO1N
bonds at room temperature in the cross-links are consistent with
independent spins or FM exchange; these could contribute an S
=
1 unit to the susceptibility at higher temperature, with Cu(II)
1
ions in the (Cuꢀpyrimidine) chains contributing S = / .
n
2
The large negative Weiss constant of the Curie plot data above
1
50 K is attributed to spin conversion of the Cu1ꢀO1N cross-
Figure 12. Ligand-sphere bonding parameters for Net2anti. Bond
lengths are in angstroms and angles in degrees. Long bond ligated
atoms are underlined.
links, with AFM exchange induced upon cooling. Below 100 K,
II
the Cu1ꢀO1N units are fully spin-paired, leaving only the Cu
1
ions of the chains magnetically active: their S = / contribution is
2
in good agreement with the Curie constant in this region.
Magnetization versus field measurements at 1.4 K for Net3d
structurally similar to loops in M3L4 and Chain, having short
CuꢀON bonds to the syn conformer aminoxyl groups.
saturate at about 1 μ (Figure 17), also consistent with only the
B
1
II
Like M2L2 and Net2anti, Net3d undergoes temperature-
dependent geometry changes (Figure 16). The Net3d loops
already have short CuꢀON bonds at room temperature that
shorten only by 0.01 Å upon cooling to 100 K. However, the
chain-to-loop Cu1ꢀO1N cross-links shorten by 0.2 Å when
cooled, changing from axial to equatorial; this is the presumed
main source of the magnetic spin-state conversion described below.
Like Net2anti, Net3d shows a decrease in χT upon cooling
S = / spin Cu ions of the chains remaining magnetically active
2
at this temperature.
The further decrease in χT of Net3d below 50 K and the
negative Weiss constant for a T < 100 K Curie plot indicate an
additional AFM exchange interaction. Only the Cu3 and Cu4
ions in the (Cuꢀpyrimidine) chains should be magnetically
n
active at this temperature, meta-linked through pyrimidine
rings of the 3PPN. m-Phenylene-type units typically are FM
exchange linkers save for cases of significant geometric or
(
Figure 17), but more gradually from ∼0.5 emu K/Oe mol at
3
3
1
8
3
00 K to ∼0.38 emu K/Oe mol over the range 150ꢀ80 K, then
heteroatom substituent perturbation. However, both FM
and AFM exchange are reported for paramagnetic metal cations
linked through pyrimidine; a summary with leading references
3
3
a drop again to a minimum of 0.2 emu K/Oe mol at 7 K, and
3
3
then a small increase to 2 K. The 1/χ vs T plot of Figure 17 can be
divided into two main regions: T > 150 K, with Curie constant
C = 1.44 emu K/Oe mol and a very large θ = (ꢀ)400 K; T <
1
9
is given by Glaser et al. Computational studies by Mohri
2
0
et al. indicate that a cross-ring σ-overlap interaction favors
AFM exchange (Scheme 7), consistent with the AFM downturn
below 50 K in Net3d. This is unique among the Table 1
3
3
1
00 K, with C = 0.366 emu K/Oe mol and θ = (ꢀ)5.2 K.
3 3
The higher temperature Curie constant is consistent with one
1
II
S = 1 unit plus one S = / unit with g ∼ 2.05 for both. The
complexes because only Net3d forms extended [Cu ꢀ
2
avg
II
short Cu2ꢀO2N bonds in the M L loops of the structure
pyrimidine] chains. Interaction between Cu ions is unlikely
2
2
n
(Figure 16) suggest that their spins are already AFM-paired, with
to involve spin polarization of the pyrimidine ring caused by the
5
069
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
Table 5. Crystallographic Parameters for 3D Complex Net3d
temperature (K)
chemical formula
fw
298
173
100
1
C
58
H
38Cu
3
F
36
N
6
O
14
C
58
H
38Cu
3
F
36
N
6
O
14
C
58
H
38Cu
3
F
36
N
6
O
14
/
2
C
6
H
14
3
1917.56
1917.56
1960.65
cell setting, space group
a (Å)
monoclinic, C2/c
61.5935(3)
15.6399(1)
17.9843(1)
90.0
monoclinic, C2/c
61.1492(3)
15.5084(1)
17.7700(1)
90.0
monoclinic, C2/c
60.8915(4)
15.4268(1)
17.6013(2)
90.0
b (Å)
c (Å)
R (deg)
β (deg)
95.8549(2)
90.0
95.6666(3)
90.0
95.3729(4)
90.0
γ (deg)
3
V (Å )
17234.19(17)
8
16769.41(16)
8
16461.3(2)
Z
8
3
a
Dcalc (Mg/m )
1.478
1.476
1.582
θ range (deg)
4.19ꢀ25.03
7608
4.11ꢀ25.01
7608
4.09ꢀ25.05
7808
F(000)
ꢀ
1
μ (mm
)
0.866
0.890
0.909
data collection method
collected/unique reflns
criterion for observed reflns
ω-2θ scans
27 491/15 042
I > 2σ(I)
0.0253
ω-2θ scans
14 612/11 049
I > 2σ(I)
0.0254
ω-2θ scans
14 368/12 283
I > 2σ(I)
0.0254
R
int
ranges of h, k, l
ꢀ73 f h f þ73
ꢀ72 f h f þ72
ꢀ18 f k f þ18
ꢀ21 f l f þ21
0.988
ꢀ72 f h f þ72
ꢀ18 f k f þ18
ꢀ20 f l f þ20
0.985
ꢀ
18 f k f þ17
21 f l f þ21
ꢀ
0.986
completeness to 2θ
reflns/restraints/param
15 042/0/1020
1.022
14 612/0/1084
1.036
14 368/0/1116
1.035
2
GOF on F
3
ΔFmax, ΔFmin (e/Å )
1.193, ꢀ0.569
0.0910, 0.2513
0.1124, 0.2744
1.287, ꢀ0.696
0.0721, 0.1915
0.0966, 0.2097
1.054, ꢀ0.703
0.0549, 0.1379
0.0639, 0.1454
R1, wR2 [I > 2σ(I)]
R1, wR2 (all)
a
Includes solvation molecules resolved at 100 K.
distant aminoxyl group of the 3PPN unit; that spin polarization
is virtually nil, based on our earlier work. This presumably also
2.0 Å to nearly 2.5 Å. None vary strongly with the temperature
(<0.1 Å).
6
applies to the other complexes.
The variety of products obtained by the reaction of Cu(hfac)
2
General Comparisons. Kinetic and/or thermodynamic fac-
tors apparently favor M2L2 loop formation incorporating the syn
conformer of 3PPN both in clusters and in extended solids. We
obtained no analogous loop-incorporating, higher-dimensional
with 3PPN presented some challenges for isolation and character-
ization. However, complexation of copper(II) with radical ligands
has produced even more challenging product variability in some
2
2
other cases. Ovcharenko and co-workers reported that the
3,4
reaction of Cu(hfac) with 4,4,5,5-tetramethyl-2-(1-methyl-1H-
solids with 4PPN, despite repeated efforts. Chain linking of
M L loops in a possible Chain4PPN may be unfavorable
2
pyrazol-4-yl)imidazoline-3-oxide-1-oxyl (MePyrNN, Chart 2)
yielded 12 distinct complexes, with varying magnetic behavior.
Solvent incorporation was important in forming some of those
phases, analogous to the role of hexane incorporation in Net3d in
our work.
2
2
because it requires significant zigzag extension, as opposed to
the near perfectly linear, trans geometry of cross-linking in Chain
from 3PPN-containing loops (Scheme 8).
The CuꢀON bonds in most of the complexes in this study
shorten during spin-state conversion to AFM exchange as the
temperature drops; a few are already AFM exchange-coupled at
room temperature. The spin-state conversions are gradual as the
temperature drops, even in extended coordination systems like
Net2anti and Net3d. This differs from the relatively sharp spin-
crossover (SCO) transitions seen in complexes that exhibit
The Cu (MePyrNN) (hfac) family of complexes consists of
x
y
2x
clusters and 1D chains as far as bonded connectivity is concerned.
Ovcharenko and co-workers carried out extensive studies while
classifying these and similar coordination complexes as “breath-
23
ing crystals”, which exhibit reversible coordination sphere and
unit cell dimension changes with varying temperature. They note
that such systems show promise for having electronically switch-
able behavior because of the temperature and even photochemi-
cal reversibility of high-temperature weak exchange coupling
21
synergistic interaction between network coordination sites.
The CuꢀN bond lengths with pyrimidine nitrogen atoms vary
significantly among the Table 1 complexes. In loops, they are
between inorganic and organic spin units (M ON exchange)
II
3 3 3
consistently 2.0ꢀ2.1 Å long, where Cu coordinates only one
and strong low-temperature AFM exchange. The complexes in
the present study have similar behavior in many respects and may
have similar promise.
pyridimidine at a time. They are more variable when two
II
pyrimidines are coordinated to Cu , ranging from just under
5
070
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
Ovcharenko and co-workers also pursued extension of the
connective dimensionality of their complexes by using MePyr-
NN biradical ligands composed of monoradicals connected by
flexible linkers, (MePyrNN) (CH ) with n = 4 and 8 (Chart 2).
Cu(hfac) because both linear and highly bent chain geometries
are possible.
2
2
2 n
’
CONCLUSIONS
This approach successfully yielded 2D and 3D networks with
24
macrocyclic and macrohelical secondary structure. The poly-
methylene linkers in the biradical ligands in their work are the
structural cross-links in the overall crystal lattice. In contrast, in
Multiple extended dimensionality complexes form by reacting
the hinge-flexible 3PPN radical ligand with Cu(hfac) . Cyclic
2
M2L2 loops are favored both as isolated clusters and as parts of
extended solids. No isolated ML2 systems were isolated. Both
helical and 1D chains form with connectivity (ꢀCuꢀPyrimꢀ
our work paramagnetic Cu(hfac) provides cross-links between
2
radical ligands, giving more paramagnetic spins per radical unit
with M L stoichoimetry instead of M LꢀL (where LꢀL are the
PhꢀNOꢀ) , but only as part of 2D networks that also have
3
2
2
n
linked radicals in Ovcharenko group’s [MePyrNN) (CH )
PyrimꢀCuꢀPyrim cross-links. A complex 3D network solid,
Net3d, was composed of M2L2 units cross-linked by ꢀCuꢀ
3PPNꢀ chains. The M2L2 loops and the 2D and 3D solids could
be obtained in tens of milligram amounts; the Chain and M3L4
complexes were only obtained in yields of a few milligrams,
presumably as kinetic products.
2
2 4,8
biradical ligands]. Although the 3PPN radical ligand used in the
present study is far more rigid than the LꢀL radicals in
(
MePyrNN) (CH ) , the 3PPN single-bond torsional flexibility
2 2 n
is sufficient to form extended coordination networks of
The magnetic behavior of most of the complexes was domi-
nated by reversible spin-state conversion of ferromagnetically or
poorly exchange-coupled CuꢀON spin units to give antiferro-
magnetically coupled low-spin-state units at lower temperatures.
All of the complexes exhibit conversion over fairly broad temp-
erature ranges. Net3d, the only network with (ꢀCuꢀPyrimꢀ)n
extended chain formation, also showed AFM exchange between
II
Cu ions across pyrimidine units.
This study shows that allowing flexible “sharp-turn” geometries
for coordination between paramagnetic metal ions and radical
ligands assists the formation of loops and chains. As a general
strategy, this offers a wide range of possibilities for making 2D and
3
D network solids with unusual magnetic properties.
’
EXPERIMENTAL SECTION
General Procedures. Diethyl ether and tetrahydrofuran were
distilled under argon from sodium. N,N-Dimethylformamide was dried
over anhydrous magnesium sulfate. Silver(I) oxide was freshly made by
treatment of silver(I) nitrate with sodium hydroxide, washing of the
resultant precipitate with distilled water, and air drying. All melting
points are uncorrected.
Variable-temperature magnetic susceptibility measurements were
carried out on a Quantum Design MPMS-5 SQUID magnetometer at
the Nanomagnetics Characterization Facility at UMass—Amherst.
Samples were placed into gelatin capsules, held in place with a plug of
cotton, and subjected to a helium atmosphere purge before measure-
ment. Paramagnetic susceptibilities were determined by subtraction of
temperature-independent magnetic contributions from the raw data.
Figure 14. (Upper) Schematic of Net3d connectivity; chains 1 and 2
cross-link to other chains to give a 3D net. (Lower) Colorized in 3D to
show different chains (blue and green) and loops (red and yellow), with
omission of the hfac units.
Figure 15. Hexane of solvation (green space-filling) in the void space of the Net3d lattice, from a 100 K crystal structure, viewed down the c axis in the
ab plane.
5
071
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
Figure 17. For Net3d, (a) χT data vs T and (b) 1/χ vs T both at 500 Oe
external field and (c) magnetization versus field at 1.4 K. The solid lines
in plot b are CurieꢀWeiss fits to data <100 and >150 K.
Scheme 7. Proposed Mechanism of Cross-Ring
Pyrimidine-Mediated AFM Exchange between Cu Ions in
II
the Chains of Net3d
Figure 16. Ligand-sphere changes with the temperature for the differ-
ent copper(II) environments in Net3d. Bond lengths are in angstroms
and angles in degrees. Long bond ligated atoms are underlined.
stirred solution of 3PPNH (0.223 g, 0.917 mmol) in benzene (20 mL)
under argon. After 2 h, the reaction was filtered to yield a clear red/
orange solution. This can be used directly for spectroscopic studies and
as a reactant in subsequent reactions. The radical can be isolated as an oil
by solvent removal without heating; it is not stable to prolonged storage
and so is best used as soon as possible after preparation. Electron
Molar magnetization versus field experiments (M vs H) were carried out
25
using a custom-built apparatus at Universidade de S ~a o Paulo.
Elemental analyses were carried out by Dr. G. Dabkowski of the
UMass—Amherst Microanalysis Laboratory. High-resolution mass
spectra were obtained by Dr. Stephen Eyles of the UMass—Amherst
Mass Spectrometry and Molecular Weight Facility.
pramagnetic resonance (benzene, room temperature, 9.649 GHz): a =
N
12.28 G, a = 2.03, 1.92, 1.82, 0.83 G.
H
Cu (3PPN) (hfac) (M3L4). 3PPNH (0.060 g, 0.247 mmol) in 2 mL
3
4
6
Preparation of 5-[3-(N-tert-Butyl-N-aminoxyl)phenyl]
pyrimidine (3PPN). Compound 3PPNH was made as described
previously. Silver(I) oxide (0.531 g, 2.29 mmol) was then added to a
of dichloromethane under argon was stirred over silver(I) oxide (0.114 g,
0.493 mmol) to give a solution of radical 3PPN and then was filtered
through Celite. Cu(hfac) xH O (0.122 g, 0.247 mmol) was dissolved in a
6
2
3
2
5
072
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
Scheme 8. “Chaining” of M L Loops in 3PPN by Linear Extension, Compared to 4PPN
2
2
Chart 2
114ꢀ119 ꢀC. Single-crystal XRD analysis was carried out. The crystal
and structural parameters in CIF format were deposited at the Cam-
bridge Crystallographic Data Centre.
Cu (3PPN) (hfac) (Net2anti). Compound 3PPNH (0.060 g,
3
2
6
0
0
.247 mmol) was stirred for 2 h over silver(I) oxide (0.114 g,
.493 mmol) in dichloromethane (1.5 mL) under argon to give a red
solution of 3PPN, which was filtered through Celite into a glass vial.
Cu(hfac) xH O (0.122 g, 0.247 mmol) was dissolved in a mixture of
2
3
2
distilled diethyl ether (2.0 mL), dichloromethane (2.0 mL), and distilled
hexanes (5.0 mL), which was then carefully layered atop the 3PPN
solution: then 5.0 mL of distilled hexanes was layered on top of this.
Crystals formed after 5 days. Mp: 107ꢀ108.5 ꢀC. Single-crystal XRD
analysis was carried out. The final parameters in CIF format were
deposited at the Cambridge Crystallographic Data Centre as CCDC.
solution of 0.5 mL of diethyl ether and 1 mL of dichloromethane. The
PPN solution was layered atop the Cu(hfac) solution and then covered
3
2
with a layer of distilled hexanes. After 4 days, a mixture of green and yellow
crystals precipitated; a few milligrams of the green crystals were separated
manually. Single-crystal XRD analysis was carried out. The crystal and
structural parameters in CIF format were deposited at the Cambridge
Crystallographic Data Centre as CCDC.
Cu (3PPN) (hfac) (Net3d). Compound 3PPNH (0.153 g,
3
2
6
0.627 mmol) was stirred with silver(I) oxide (0.182 g, 0.784 mmol) in
dichloromethane (2 mL) under argon to give a red solution of 3PPN,
which was filtered through Celite into a 5 mL test tube. Cu-
Cu (3PPN) (hfac) (M2L2). The complex was synthesized as
2
2
4
6
previously described to give a brownish powder (which reversibly turns
green upon cooling in liquid nitrogen), which yields yellow-green needles
when recrystallized by slow evaporation from toluene. Mp: 104ꢀ106 ꢀC.
The crystal and structural parameters in CIF format have been deposited
at the Cambridge Crystallographic Data Centre (PAXBAS, PAXBAS02).
2 2
(hfac) xH O (0.466 g, 0.941 mmol) was dissolved in a mixture of
3
dichloromethane (1 mL), diethyl ether (1 mL), and distilled hexane
(5 mL), which was carefully layered atop the solution of 3PPN. After 5
days, large dark-green-yellow crystals formed. Mp: 123.5ꢀ124.5 ꢀC.
Single-crystal XRD analysis was carried out. The crystal and structural
parameters in CIF format were deposited at the Cambridge Crystal-
lographic Data Centre as CCDC.
Cu (3PPN) (hfac) (Chain). Compound 3PPNH (7 mg,
3
2
6
0
.03 mmol) was oxidized in benzene by the procedure described
above, filtered, and evaporated, and the resulting 3PPN was dissolved
2
in dichloromethane (0.1 mL) in a 5 mL test tube. Cu(hfac) xH O
2
3
(
(
3 mg, 7.0 micromol) was dissolved in a mixture of dichloromethane
0.2 mL) and hexanes (1.0 mL), which was carefully layered atop the
’
ASSOCIATED CONTENT
red radical solution. After 4 days, green and yellow crystals appeared in
the test tube. The green crystals were manually separated for analysis
and single-crystal XRD work. The crystal and structural parameters in
CIF format were deposited at the Cambridge Crystallographic Data
Centre as CCDC.
S
Supporting Information. General synthetic procedures,
b
ORTEP diagrams, and CIF format crystallographic summaries
for all complexes and for 3PPNH, chain packing diagrams for
Chain, Net2syn, and Net2anti, 100 K crystallographic analysis
procedure for Net2anti and a comparison to the structure at
room temperature, diagrams of void space solvation in Net3d,
and powder XRD for Net2syn compared to simulations from
other single-crystal XRD. This material is available free of charge
via the Internet at http://pubs.acs.org.
Cu (3PPN) (hfac) (Net2syn). 3PPNH (0.223 g, 0.917 mmol)
3
2
6
was oxidized as described above, filtered, and evaporated, and the
product 3PPN was dissolved in dichloromethane (6.5 mL) and pipetted
2 2
into a 25 mL graduated cylinder. Cu(hfac) xH O (0.454 g, 0.917 mmol)
3
was dissolved in dichloromethane (6.0 mL) and hexanes (14.0 mL) and
layered on top of the radical solution. Crystals formed after 1 week. Mp:
5
073
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074

Inorganic Chemistry
ARTICLE
’
AUTHOR INFORMATION
(15) See a similar case in: Lanfranc de Panthou, F.; Belorizky, E.;
Calemczuk, R.; Luneau, D.; Marcenat, C.; Ressouche, E.; Turek, P.; Rey,
P. J. Am. Chem. Soc. 1995, 117, 11247.
Corresponding Author
*E-mail: lahti@chem.umass.edu.
(16) (a) Luneau, D.; Rey, P.; Laugier, J.; Fries, P.; Caneschi, A.;
Gatteschi, D.; Sessoli, R. J. Am. Chem. Soc. 1991, 113, 124.
(
b) Ressouche, E.; Boucherle, J.-X.; Gillon, B.; Rey, P.; Schweizer, J. J.
’
ACKNOWLEDGMENT
J. Am. Chem. Soc. 1993, 115, 3610. (c) Boucherle, J.-X.; Ressouche, E.;
Schweizer, J.; Gillon, B.; Rey, P. Z. Naturforsch. 1993, 48, 120.
M.B. and P.M.L. thank the U.S. National Science Foundation
(
17) (a) Choe, W.; Pecharsky, V. K.; Pecharsky, A. O.; Gschneider,
K. A.; Young, V. G., Jr.; Miller, G. J. Phys. Rev. Lett. 2000, 84, 4717.
b) Colombo, D. G.; Young, V. G., Jr.; Gladfelter, W. L. Inorg. Chem.
001, 39, 4621. (c) Jazdzewski, B. A.; Holland, P. L.; Pink, M.; Young,
for support through Grants CHE-0109094 and CHE-0415716.
A.P.-F. and N.F.O. thank Funda c- ~a o de Amparo ꢀa Pesquisa do
Estado de S ~a o Paulo (FAPESP) in Brazil for support. J.T.M.
thanks the Louisiana Board of Regents through the Louisiana
Educational Quality Support Fund (Grant 2003-ENH-TR-67)
for the purchase of a diffractometer and the Tulane University
Chemistry Department for support of the X-ray diffraction
facility.
(
2
V. G., Jr.; Spencer, D. J. E.; Tolman, W. B. Inorg. Chem. 2001, 40, 6097.
(18) (a) Borden, W. T., Ed. Diradicals; Wiley: New York, 1982.
(b) Borden, W. T.; Iwamura, H.; Berson, J. A. Acc. Chem. Res. 1994,
27, 109. (c) Fang, S.; Lee, M.-S.; Hrovat, D.; Borden, W. T. J. Am. Chem.
Soc. 1995, 117, 6727.
(
19) Glaser, T.; L €u gger, T.; Fr €o hlich, R. Eur. J. Inorg. Chem.
004, 394 and references cited therein.
20) Mohri, F.; Yoshizawa, K.; Yamabe, T.; Ishida, T.; Nogami, T.
2
’
REFERENCES
(
(
1) (a) Kahn, O. Molecular Magnetism; VCH: New York, 1993.
b) Iwamura, H.; Inoue, K.; Hayamizu, T. Pure Appl. Chem. 1996,
8, 243. (c) Caneschi, A.; Gatteschi, D.; Sessoli, R.; Rey, P. Acc. Chem.
Mol. Eng. 1999, 8, 357.
(
6
(21) (a) G €u tlich, P.; Garcia, Y.; Goodwin, H. A. Chem. Soc. Rev.
2000, 29, 419. (b) Niel, V.; Thompson, A. L.; Mu ~n oz, M. C.; Galet, A.;
Goeta, A. E.; Real, J. A. Angew. Chem., Int. Ed. 2003, 42, 3760.
(c) Bousseksou, A.; Molnar, G.; Matouzenko, G. Eur. J. Inorg. Chem.
2004, 22, 4353.(d) G €u tlich, P., Goodwin, H. A., Eds. Topics in Current
Chemistry; Springer: Berlin, 2004; Vols. 232ꢀ234.
Res. 1989, 22, 392.(d) Caneschi, A.; Gatteschi, D.; Sessoli, R. In Magnetic
Molecular Materials; Gatteschi, D., Kahn, O., Miller, J. S., Palacio, F.,
Eds.; Kluwer: Dordrecht, The Netherlands, 1991; p 215. (e) Gatteschi,
D.; Rey, P. In Magnetic Properties of Organic Materials; Lahti, P. M., Ed.;
Marcel Dekker: New York, 1999. (f) Caneschi, A.; Gatteschi, D.; Rey, P.
In Progress in Inorganic Chemistry; Lippard, S. J., Ed.; Wiley: New York,
(22) Fokin, S.; Ovcharenko, V.; Romanenko, G.; Ikorskii, V. Inorg.
Chem. 2004, 43, 969.
1
1
991; Vol. 39, p 331. (g) Ouahab, L. Coord. Chem. Rev. 1998,
78ꢀ180, 1501. (h) Hicks, R. G. Aust. J. Chem. 2001, 54, 597.
(23) (a) Ovcharenko, V. In Stable Radicals: Fundamentals and
Applied Aspects of Odd-Electron Compounds; Hicks, R. G., Ed.; John
Wiley & Sons, Ltd.: Chichester, U.K., 2010; Chapter 13. (b) Fedin,
M. V.; Veber, S. L.; Maryunina, K. Yu.; Romanenko, G. V.; Suturina,
E. A.; Nina, P.; Gritsan, N. P.; Sagdeev, R. Z.; Ovcharenko, V. I.;
Bagryanskaya, E. G. J. Am. Chem. Soc. 2010, 132, 13886. (c) Veber, S. L.;
Fedin, M. V.; Maryunina, K. Yu.; Romanenko, G. V.; Renad, Z.; Sagdeev,
R. Z. Inorg. Chim. Acta 2008, 361, 4148. (d) Fedin, M. V.; Veber, S. L.;
Romanenko, G. V.; Ovcharenko, V. I.; Sagdeev, R. Z.; Klihm, G.;
Reijerse, E.; Lubitz, W.; Bagryanskaya, E. G. Phys. Chem. Chem. Phys.
2009, 11, 6654. (e) Fedin, M.; Ovcharenko, V.; Sagdeev, R.; Reijerse, E.;
Lubitz, W.; Bagryanskaya, E. Angew. Chem., Int. Ed. 2008, 47, 6897.
(24) Treyakov, E.; Fokin, S.; Romanenko, G.; Ikorskii, V.; Vasilevsky,
S.; Ovcharenko, V. Inorg. Chem. 2006, 45, 3671.
(i) Lemaire, M. T. Pure Appl. Chem. 2004, 76, 277.
(2) Field, L. M.; Lahti, P. M.; Palacio, F.; Paduan-Filho, A. J. Am.
Chem. Soc. 2003, 125, 10110.
(
(
3) Field, L. M.; Lahti, P. M. Inorg. Chem. 2003, 42, 7447.
4) Lahti, P. M.; Baskett, M.; Field, L. M.; Mor ꢁo n, M. C.; Palacio, F.;
Paduan-Filho, A.; Oliveira, N. F., Jr. Inorg. Chim. Acta 2008, 361,
697–3709.
5) Field, L. M.; Mor ꢁo n, M. C.; Lahti, P. M.; Palacio, F.; Paduan-
Filho, A.; Oliveira, N. F., Jr. Inorg. Chem. 2006, 45, 2562.
6) Baskett, M.; Lahti, P. M.; Paduan-Filho, A.; Oliveira, N. F., Jr.
Inorg. Chem. 2005, 44, 6725.
3
(
(
(7) Field, L. M.; Lahti, P. M. Polyhedron 2005, 24, 2639.
(8) Ishimura, Y.; Inoue, K.; Koga, N.; Iwamura, H. Chem. Lett.
(25) Oliveira, N. F., Jr.; Paduan-Filho, A.; Salinas, S. R.; Becerra,
C. C. Phys. Rev. B 1978, 18, 6165.
1
994, 1693.
9) Ishimaru, Y.; Kitano, M.; Kumada, H.; Koga, N.; Iwamura, H.
Inorg. Chem. 1998, 37, 2273.
10) (a) Kitano, M.; Ishimaru, Y.; Inoue, K.; Koga, N.; Iwamura, H.
(
(
Inorg. Chem. 1994, 33, 6012. (b) Rabu, P.; Drillon, M.; Iwamura, H.;
Gorlitz, G.; Itoh, T.; Matsuda, K.; Koga, N.; Inoue, K. Eur. J. Inorg. Chem.
2
000, 211.
11) See also a summary in: Iwamura, H.; Koga, N. Mol. Cryst. Liq.
Cryst. 1999, 334, 437.
12) Yamada, K.; Yagishita, S.; Tanaka, H.; Tohyama, K.; Adachi, K.;
(
(
Kaizaki, S.; Kumagai, H.; Inoue, K.; Kitaura, R.; Chang, H.-C.; Kitagawa,
S.; Kawata, S. Chemistry 2004, 10, 2647.
(13) Software for crystallographic pictures includes the following.
(a) Mercury, version 2.4, Cambridge Crystallographic Data Centre:
Macrae, C. F.; Bruno, I. J.; Chisholm, J. A.; Edgington, P. R.; McCabe,
P.; Pidcock, E.; Rodriguez-Monge, L.; Taylor, R.; van de Streek, J.;
Wood, P. A. J. Appl. Crystallogr. 2008, 41, 466–470. (b) ORTEP-3 for
Windows: Farrugia, L. J. J. Appl. Crystallogr. 1997, 30, 565.
(14) (a) Caneschi, A.; Gatteschi, D.; Sessoli, R.; Rey, P. Acc. Chem.
Res. 1989, 22, 392.(b) Caneschi, A.; Gatteschi, D.; Sessoli, R. In Magnetic
Molecular Materials; Gatteschi, D., Kahn, O., Miller, J. S., Palacio, F.,
Eds.; Kluwer: Dordrecht, Netherlands, 1991; p 215. (c) Gatteschi, D.;
Rey, P. In Magnetic Properties of Organic Materials; Lahti, P. M., Ed.;
Marcel Dekker: New York, 1999.
5
074
dx.doi.org/10.1021/ic200362c |Inorg. Chem. 2011, 50, 5060–5074