Aminophenylnitronylnitroxides: Highly networked hydrogen-bond assembly in organic radical materials

Article abstract of DOI:10.1021/cm202303q

2-(Meta-aminophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-1H-imidazole-3-oxide- 1-oxyl (mAPN) and 2-(para-aminophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-1H- imidazole-3-oxide-1-oxyl (pAPN) were synthesized and subjected to magnetostructural analysis. Both form extended hydrogen bonding networks involving both amino NH bonds to radical spin-density bearing nitronylnitroxide NO groups. Their crystallographic assembly motifs and magnetic exchange properties are compared to those of tert-butoxylcarbonyl (BOC) and amide derivatives having only one NH bond. The conversion of pAPN to acid salt derivatives gives a solid that is essentially diamagnetic, although dissolution of the solid shows the radical spin units to be preserved.

Full text of DOI:10.1021/cm202303q

ARTICLE
pubs.acs.org/cm
Aminophenylnitronylnitroxides: Highly Networked Hydrogen-Bond
Assembly in Organic Radical Materials
Safo Aboaku,||,† Armando Paduan-Filho,‡ Valdir Bindilatti,‡ Nei Fernandes Oliveira, Jr.,‡ John A. Schlueter,§
and Paul M. Lahti*,†
†Department of Chemistry, University of Massachusetts, Amherst, Massachusetts 01003, United States
‡Instituto de Física, Universidade de S~ao Paulo, S~ao Paulo, SP 05315-970, Brazil
§Materials Science Division, Argonne National Laboratory, Argonne, Illinois 60439, United States
S Supporting Information
b
ABSTRACT: 2-(Meta-aminophenyl)-4,4,5,5-tetramethyl-4,5-
dihydro-1H-imidazole-3-oxide-1-oxyl (mAPN) and 2-(para-ami-
nophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-1H-imidazole-3-oxide-
1-oxyl (pAPN) were synthesized and subjected to magnetostructural
analysis. Both form extended hydrogen bonding networks invol-
ving both amino NH bonds to radical spin-density bearing
nitronylnitroxide NO groups. Their crystallographic assembly
motifs and magnetic exchange properties are compared to those
of tert-butoxylcarbonyl (BOC) and amide derivatives having only
one NH bond. The conversion of pAPN to acid salt derivatives
gives a solid that is essentially diamagnetic, although dissolution of the solid shows the radical spin units to be preserved.
KEYWORDS: molecular magnetism, hydrogen bonding, nitronylnitroxides, radicals
’ INTRODUCTION
quasi-2D ferromagnetic phase. Matsushita et al. have studied
RSNN, which shows hydrogen-bond assisted spin pairing.4 Taylor
and Lahti investigated5 the effect of sterically inhibiting hydro-
gen bonding by the phenolic OH through tert-butylation in
HOBu2PhNN, and they found that 1-D zigzag OH to ON hydrogen
bonding arrays formed, albeit with changes in hydrogen bonding
geometry relative to 4HOPhNN as a result of steric constraints
caused by the tert-butyl groups. These various studies show
that phenolic hydrogen bonds can assemble organic spin units with
significant intermolecular exchange interactions between radicals.
We aimed to replace the OH group of phenolic nitronylnitr-
oxides with an NH2 group to induce extended hydrogen-
bonding motifs similar to those found in phenolic nitronylnitr-
oxides, but with enhanced networking because NH2 can form
two hydrogen bonds with the major spin bearing aminoxyl
The field of molecule-based magnetism frequently utilizes
molecular assembly to achieve some degree of predictable crystal
packing. Purely organic, metal-free magnetic materials are parti-
cularly dependent on assembly, because interspin exchange
interactions are only significant across short distances for upper-
row elements. Bringing specific portions of spin-bearing mol-
ecules into proximity can frequently be done using hydrogen
bonding interactions. The attachment of various functional
groups to unpaired spin carriers such as aminoxyl, nitronylnitr-
oxide, and verdazyl radicals can induce intermolecular exchange
interactions ranging from simple pairing to three-dimensionally
ordered ferromagnetism in the solid state. Dependable predic-
tion of magnetism remains an elusive challenge for organic
radical-based solids, in part, because sure prediction of crystal
packing also remains elusive. However, the past 10ꢀ15 years of
work by various groups has shown a number of useful correla-
tions between specific intermolecular solid state arrangements
and qualitative magnetic exchange behaviors.1
Phenolic groups have been used with some frequency for
assembling radicals, especially nitronylnitroxides (Chart 1).
Sugawara and co-workers carried out studies of HQNN,2 which
has two allotropes, the α-form of which forms paired hydrogen-
bonding chains and orders as an organic ferromagnet at Tc =
0.5 K. Veciana and co-workers have carried out studies on
2HOPhNN, 3HOPhNN, 4HOPhNN, and NNCatH2,3 which
form various hydrogen bonded motifs including dimers, chains,
and 2-D sheets, at least one of which3d appears to form a
NꢀO groups, whereas OH can only form one (Scheme 1)
3
unless bifurcated interactions occur. Of course, other (albeit
weaker) hydrogen bond like interactions can occur, such
as methyl CH to aminoxyl NꢀO contacts exemplified in
3
Scheme 1 for the phenolic compounds. However, in hopes
of realizing higher dimensionality networks of the stronger
hydrogen bonds involving NH and NH2 groups, we
targeted the syntheses of 2-(meta-aminophenyl)-4,4,5,5-tetra-
methyl-4,5-dihydro-1H-imidazole-3-oxide-1-oxyl (mAPN) and
Received: August 5, 2011
Revised:
September 15, 2011
Published: October 06, 2011
r
2011 American Chemical Society
4844
dx.doi.org/10.1021/cm202303q Chem. Mater. 2011, 23, 4844–4856
|

Chemistry of Materials
ARTICLE
Chart 1
150ꢀ152 ꢀC. FTIR (KBr, cmꢀ1): 3392, 3328, 3232 (NH), 2925
(methyl CꢀH). MS(FAB): found m/z 248.14, C13H18N3O2 requires
m/z 248.30. EPR (9.651 GHz): g = 2.0063, a(N) = 7.35 gauss (2N).
2-(Para-acetamidophenyl)-4,4,5,5-tetramethyl-4,5-di-
hydro-1-H-imidazole-3-oxide-1-oxyl (pAcPN). Blue-black solid,
mp 204ꢀ205 ꢀC. FTIR (KBr, cmꢀ1): 3341 (NH), 2992 (methyl CꢀH),
1707 (CdO). EPR (9.650 GHz): g = 2.0068, a(N) = 7.49 gauss (2N).
Anal. for C15H20N3O3 requires C, 62.05; H, 6.94; N, 14.47. Found: C,
61.90; H, 6.91; N, 14.45.
Scheme 1. Comparing Hydrogen Bonding Possibilities for
OH and NH2 Containing Nitronylnitroxides
2-(Para-(tert-butoxylcarbonyl)aminophenyl)-4,4,5,5-tetra-
methyl-4,5-dihydro-1H-imidazole-3-oxide-1-oxyl (pBocPN).
Blue-black solid, mp 215ꢀ218 ꢀC. FTIR (KBr, cmꢀ1): 3340 (NH),
2980 (methyl CꢀH) 1728 (CdO). EPR (9.652 GHz): g = 2.0068,
a(N) = 7.47 gauss (2N). Anal. for C18H26N3O4 requires C, 62.05; H,
7.52; N, 12.06. Found: C, 62.03; H, 7.55; N, 12.04.
2-(Para-aminophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-
1H-imidazole-3-oxide-1-oxyl (pAPN). Blue-green prisms, mp
147ꢀ150 ꢀC. FTIR (KBr, cmꢀ1): 3446, 3334, 3214 (NH), 2926
(methyl CꢀH). EPR (9.644 GHz): g = 2.0062, a(N) = 7.87 gauss
(2N). Analysis for C13H18N3O2 requires: C, 62.87; H, 7.32; N, 16.92.
Found: C, 63.27; H, 7.41; N, 16.21.
2-(para-aminophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-1H-imi-
dazole-3-oxide-1-oxyl (pAPN). In this article, we report the
preparation, properties, and magnetostructural analyses of
mAPN and pAPN, with comparisons to corresponding aceta-
mide and tert-butoxycarbonyl (BOC) derivatives that have only
one NH unit available for hydrogen bonding. We also describe
preliminary studies of protonated ammonium cations of pAPN
derived by treating the neutral radical with mineral acids.
2-(Para-aminophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-
1H-imidazole-3-oxide-1-oxyl Hydrochloric Acid Salt (pAPN
3
HCl). Neutral radical pAPN (0.20 g, 0.80 mmol) was stirred in 30 mL
of dichloromethane at room temperature under nitrogen, as 2 mL of
1 M HCl in ether was added slowly by syringe. The mixture was stirred
under nitrogen for 20 min. The mixture was filtered, and the solid
residue was washed with dichloromethane to yield 66 mg (29%) of
’ EXPERIMENTAL SECTION
product pAPN HCl as a dark greenish solid. FTIR (KBr, cmꢀ1):
3
General Methods. 2,3-Bis(hydroxylamino)-2,3-dimethylbutane
monosulfate was made by the method of Ovcharenko, Fokin, and
Rey.6 Full details for the syntheses shown in Scheme 2 are given in the
Supporting Information, including intermediates meta-acetamidobenzyl
alcohol,7 meta-acetamidobenzaldehyde,8 and 2-(para-(tert-butoxycarbonyl)-
aminobenzaldehyde.9
2-(Meta-acetamidophenyl)-4,4,5,5-tetramethyl-4,5-dihy-
dro-1H-imidazole-3-oxide-1-oxyl (mAcPN). Dark blue crystals,
mp 197ꢀ199 ꢀC. FTIR (KBr, cmꢀ1): 3262 (NꢀH), 2987 (methyl
CꢀH), 1665 (CdO). MS (FAB): found m/z= 291.0, C15H20N3O3 requires
m/z = 290.3. EPR (9.651 GHz): g = 2.0069, a(N) = 7.42 gauss (2N).
2-(Meta-(tert-butoxylcarbonyl)aminophenyl)-4,4,5,5-tetra-
methyl-4,5-dihydro-1H-imidazole-3-oxide-1-oxyl (mBocPN).
Dark blue crystals, mp 165ꢀ167 ꢀC. FTIR (KBr, cmꢀ1): 3328 (NH), 2980
(methyl CꢀH), 1726 (CdO). MS(FAB): found m/z 348.2, C18H26N3O4
requires m/z 348.4. EPR (9.652 GHz): g = 2.0065, a(N) = 7.41 gauss (2N).
2-(Meta-aminophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-
1H-imidazole-3-oxide-1-oxyl (mAPN). Blue-black crystals, mp
+
2120ꢀ3320 (br, NH). MS(FAB[+]): C13H19N3O2 requires m/z =
249.15, found 250.16 (M+H). The absorption spectra of pAPN HCl are
3
concentration dependent. UVꢀvis (MeOH): (13.8 μM: 320 nm); (55
μM: 325); (220 μM: 354 nm, 418 nm); (878 μM: 440 nm). The EPR of
the neat solid showed only a very weak peak at g ∼ 2. The EPR
spin counts comparing methanol solutions of pAPN and pAPN HCl
3
3
showed ∼30% of spins retained in the salt. Solutions of pAPN HCl
showed loss of color and EPR signal within 24 h. Adding KOH to a
green methanol solution of pAPN HCl regenerates the blue color of
3
neutral pAPN.
2-(Para-aminophenyl)-4,4,5,5-tetramethyl-4,5-dihydro-
1H-imidazole-3-oxide-1-oxyl Hydrofluoroboric Acid Salt
(pAPN HBF4). Neutral radical pAPN (0.100 g, 0.403 mmol) was
3
stirred in 15 mL of methanol at room temperature under nitrogen,
as 400 μL of tetrafluoroboric acid solution (48 wt % HBF4 in water) was
added slowly by syringe. The mixture was stirred under nitrogen for 2 h,
then evaporated under reduced pressure to give a liquid with some
4845
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
precipitate. This residue was further cooled in the freezer. The remaining
liquid was pipetted away, and the remaining solid dried under vacuum
for 4 h. This solid was washed with dichloromethane and dried in air to
corresponding nitronylnitroxides, and hydrolyze the amide
groups to the desired amines. However, various efforts to convert
amidophenyl nitronylnitroxides mAcPN and pAcPN to amino-
phenyl nitronylnitroxides were not satisfactory. The sequences
shown in Scheme 2 gave much better yields of mAPN and
pAPN, as well as providing boc-carbamate-functionalized radi-
cals mBocPN and pBocPN. We have given a preliminary report11
of the syntheses and magnetostructural studies of pAcPN and
pBocPN.
yield 61 mg (45%) of pAPN HBF4 as a brown solid. FTIR (KBr, cmꢀ1):
3
+
2570ꢀ3320 (br, NH), 1056 (BꢀF). MS(FAB[+]): C13H19N3O2
requires m/z = 249.15, found m/z = 250.1 (M + H); [ꢀ]-mode found
m/z = 86.98 (BF4ꢀ). pAPN HBF4 absorption spectra are concentra-
3
tion dependent. UVꢀvis (MeOH): (14 μM: 322 nm, 417 nm);
(55.8 μM: 320 nm, 413 nm); (223 μM: 437 nm); (893 μM:
461 nm). The EPR of the neat solid pAPN HBF4 showed only a very
3
All of the radicals were deep blue solids that appear to be
indefinitely stable in air at room temperature. Single crystals of
each were subjected to X-ray diffraction analysis, the results of
which are detailed in Table 1 and in the Supporting Information.
ORTEP19 diagrams for each are shown in Figure 1. Solutions of
each showed the typical 1:3:5:3:1 pentet electron paramagnetic
resonance (EPR) spectra expected for nitronylnitroxides, arising
from hyperfine coupling of the two equivalent nitrogen atoms.
Crystallographic Packing Motifs that Influence Exchange
between Nitronylnitroxides. Scheme 3 shows some basic
intermolecular parameters that strongly affect exchange between
NꢀO groups. Stronger exchange, in general, is favored by smaller
weak peak at g ∼ 2. EPR spin counts comparing methanol solutions
of pAPN and pAPN HBF4 showed 64% of spins retained in the salt.
3
Solutions of the radical salt showed the same instability in air and color
changes upon addition of base that were observed for pAPN HCl.
3
Crystallographic Measurements. Room temperature single
crystal X-ray diffraction measurements were carried out for mAcPN,
mBocPN, mAPN and pAPN at the University of Massachusetts Amherst
X-ray Structural Characterization Facility on a Bruker-Nonius Kap-
paCCD instrument (setup supported by NSF CHE-9974648 and the
University of Massachusetts Chemistry Department), using molybde-
num source λ = 0.71073 Å radiation. Structures were solved by direct
methods using SHELXTL.10 The 15 K structure of pAPN was deter-
mined at the ChemMatCARS (Sector 15) beamline at the Advanced
Photon Source. The crystallography for pAcPN and pBocPN has been
previously reported11 by Aboaku and Lahti; various details of these are
used for comparison in this full report. New structure data have been
deposited with the Cambridge Crystallographic Data Centre. Further
details are given in the Supporting Information CIF files.
0
0
distances rO/O and rO/N , which bring sites of strong spin density
into close proximity. However, considering only the distances
between spin sites (dipolar approximation) is inadequate
because π-type orbital overlap and the resulting interaction are
orientation dependent, as in the strong overlap between π-spin
orbitals of an antiparallel stacked arrangement of NO groups
(j < 90ꢀ) that leads to AFM exchange.20 Strong exchange is
favored when j = 45ꢀ90ꢀ, because this forces the NO groups
into closer proximity. Where loci of the three atoms attached to
the nitrogen of an NO radical site form a plane, the parameter ζ
should be about zero to allow the NO spin orbital sites to point
toward one another as needed for AFM exchange. If ζ ∼ 90ꢀ,
then the NO spin units are nearly orthogonal to one another,
a situation that promotes FM exchange. Finally, the pseudodihe-
dral angle ω in Scheme 3 should be about 90ꢀ to minimize
“side-slipping” that decreases spin orbital overlap. The interplay
between these parameters is complex and subtle, but each affects
whether there is strong spin orbital overlap between NO units
that favors AFM exchange, or not. To summarize, in Scheme 3
Magnetic Measurements. Dc magnetic susceptibility (χ) mea-
surements at 1.8ꢀ300 K were carried out at the University of Massachusetts
Amherst Nanomagnetics Facility (NSF CTS-0116498) with a Quantum
Design MPMS-7 magnetometer using crushed, polycrystalline samples
placed in gelatin capsules and held in place by a plug of cotton. Temperature
independent corrections to susceptibility were made by the extrapolation of
the uncorrected high temperature χT data. pAPN was also analyzed at
0.5ꢀ4 K, using a custom built apparatus12 at Universidade de S~ao Paulo.
Computations. Computational modeling of spin densities used
fixed crystallographic geometries of the monoradicals at the UB3LYP13/
EPR-III level of theory with the Gaussian14 program. Estimates of
the exchange interactions were made using computations with crystal-
lographically derived dyad and triad geometries. For intermolecular
exchange modeling, hybrid density functional UB97D computations
using Grimme’s dispersion corrected functional (for better treatment of
nonbonded interactions)15 with the 6-31G+(d) basis set were carried
out using the crystallographic geometries of close-contact dyads: methyl
groups on the radical units were replaced by hydrogen atoms in most
cases. For singletꢀtriplet (S-T) state energy splitting, the singlet
state was modeled by a broken symmetry unrestricted UB3LYP
wave function, which gives a highly spin-contaminated “state” with cor-
responding16 uncertainties. Yamaguchi’s scheme17,18 was used to adjust
for the spin contamination in the state energies by eq 1:
0
the following promote AFM exchange: smaller distances rO/O
0
and rO/N , j = 45ꢀ90ꢀ, ζ ∼ 0ꢀ, and ω ∼ 90ꢀ. In various cases in
the discussion below, we will note to how these parameters
correlate with exchange in Table 2.
Novoa and co-workers21 have used a somewhat larger set
of intermolecular contact definitions (some analogous to those in
Scheme 3) in a sophisticated and compelling study relating
crystallographically identified, pairwise exchange interactions to
the experimental exchange and magnetic behavior of several
nitronylnitroxides. They showed that the subtle interplay among
so many variables renders extremely difficult a confident assign-
ment of overall magnetic behavior to specific exchange interac-
tions coming from specific contacts, unless one of the contacts is
very strongly dominant. Also, at low temperatures, the crystal
lattice volume and intermolecular contact distances (or even
geometries) will change somewhat by comparison to room
temperature structures. The computational dyad exchange inter-
actions given below should therefore be viewed as an effort to
identify qualitative tendencies for exchange: large differences in
exchange behavior between geometric contacts are qualitatively
meaningful, while small differences should be considered too
close to call.
ES ꢀ ET
ΔETꢀS
¼
ð1Þ
hS2HSi ꢀ hS2LSi
where ÆS2æ is the spin-squared expectation value from the triplet (T) and
singlet (S) multiplicity computations and ΔE > 0 for a triplet the ground
state. Zero point energy corrections are not included.
’ RESULTS AND DISCUSSION
Synthesis. Our initial synthetic plan for mAPN and
pAPN was to convert meta- and para-aminobenzyl alcohols to
amides, oxidize to corresponding aldehydes, condense with
2,3-bis(hydroxylamino)-2,3-dimethylbutane, oxidize to the
4846
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
Scheme 2. Syntheses of Nitronylnitroxide Radicals
N-Functionalized Aminophenyl Nitronylnitroxides: Lim-
ited Hydrogen Bonding Cases. Acetyl substituted mAcPN
forms hydrogen-bonded chains along the crystallographic c-axis
involving only the amide units (Figure 2, labeled c), with pendant
radical groups on alternating sides of the chain. Nitronylnitroxide
The 1/χ versus T plot for mBocPN yields a Curie constant of
C = 0.374 emu K/Oe mol and a modest Weiss constant of θ =
3
3
(ꢀ)0.8 K (Figure 5). Computations show nearly zero spin
0
density at the NꢀH group for mBocPN, so the rO/O contact
r[(N9)ꢀO10 O110ꢀ(N7)] (labeled b2 in Figure 4) is pre-
3 3 3
NꢀO OꢀN contacts form chains perpendicular to the
sumably the main contributor to the modest AFM exchange here,
with no exchange coupling across the hydrogen bond. The
modest exchange is in accord with the poor overlap of the
SOMOs in this geometry.
3 3 3
hydrogen-bonded chains, along the crystallographic b-axis
(labeled b in Figure 2), with rO/O contact r(O10 O110) =
0
3 30 3
3.83 Å at 293 K. Because j = —N9ꢀO10 O11 ꢀ(N7) =
3 303
143ꢀ and angle ω = —C8ꢀN9ꢀO10 O11 ꢀ(N7) = 123ꢀ,
Because of our preliminary report of results for pAcPN and
pBocPN,11 we only briefly summarize these results for compar-
ison. Both form cyclic dyads with the amide NꢀH groups
hydrogen-bonded to NꢀO groups, as shown in Figure 6
(contacts a). UB3LYP/EPR-III computations show virtually no
spin density on the amide groups; so, hydrogen bonding should
not provide an electronic pathway for exchange interactions.
However, in pAcPN, the NꢀO groups that are not involved in
hydrogen bonding form antiparallel stacked pairs between hydro-
3 3 3
the overlap of singly occupied molecular orbitals (SOMOs)
here is poor. Magnetic measurements for mAcPN (Figure 3)
confirm this, showing essentially paramagnetic behavior.
From a CurieꢀWeiss plot of 1/χ versus T, the Curie constant
C = 0.368 emu K/Oe mol—in good agreement with the
3
13
expected value for S = /2 radical spin carriers—with a Weiss
constant of only θ = (ꢀ)0.3 K.
By comparison, mBocPN forms hydrogen-bonded chains
between BOC NꢀH donor groups and nitronylnitroxide
NꢀO acceptor groups (Figure 4, labeled b1, dashed lines).
Overall, the nitronylnitroxides are linked by chain contacts along
the crystallographic b-axis (Figure 4, labeled b2). These chain
contacts between NꢀO groups provide a possible AFM ex-
change pathway between sites of large spin density, as shown in
gen-bonded dyads, with rO/O contact r[(N4)ꢀO15 O150ꢀ
0
3 3 3
(N4)] = 4.103 Å, and j = —N4ꢀO15 O150ꢀ(N4) = 73.3ꢀ
3 3 3
across contact b in the upper part of Figure 6. However, the
aminoxyl O15 also is nearly equidistant from the 0neighboring
N40 atom (Table 2, rO/N ) and the attached C11 atom. The
0
O15 N40 contact would typically be expected to favor AFM
3 3 3
Scheme 4, with rO/O contact b2 having r(N9)ꢀO10 O110ꢀ
exchange, but the O15 C110 contact to favor FM exchange.
0
3 3 3
3 3 3
(N) = 3.97 Å and j = —N9ꢀO10 O110ꢀ(N7) = 108.2ꢀ.
This is a case where a small shift of NꢀO groups relative to one
another could influence the exchange significantly, for example,
by lattice contraction during cooling for magnetic measurements.
Radical pAcPN exhibits significant AFM exchange, as shown
by the downturn in its χ versus T susceptibility plot (Figure 7).
3 3 3
The oxygen ends of the NO groups are not well aligned for
SOMOꢀSOMO overlap, with an obtuse angle j and interplane
angle ζ = 43ꢀ, which is much larger than in mAcPN; ζ should be
nearly zero for best overlap.
4847
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
Table 1. Crystallographic Data and Structure Refinement for mAcPN, mBocPN, mAPN, pAcPN, pBocPN, and pAPN
mAcPN
mBocPN
mAPN
temp. (K)
293
293
293
chemical formula
chemical formula wt
wavelength (Å)
crystal system, space group
a, b, c (Å)
C15H20N3O3
290.34
C18H26N3O4
348.42
C13H18N3O2
248.3
0.71073
0.71073
0.71073
orthorhombic, Pna21
orthorhombic, Pbca
monoclinic, P21/c
19.9092(5)
7.7405(1)
9.5638(2)
13.922(5)
11.778(5)
24.558(5)
7.738(5)
19.085(5)
19.085(5)
α, β, γ (deg)
90.0
90.0
90.0
90.0
90.0
90.0
90.0
102.262(5)
90.0
V (Å3)
1473.85(5)
4027(2)
1320.2(12)
Z
4
8
4
Dcalc (Mg/m3)
1.309
1.15
1.249
F(000)
620
1496
532
μ (mmꢀ1
)
0.093
0.082
0.086
Rint
0.0352
0.0840
0.0425
completeness to 2θ
Nref, param.
goodness of fit on F2
ΔFmax, ΔFmin (e ꢀ Åꢀ3
R1, wR2 (I > 2σ(I))
R1, wR2 (all)
1.00
0.992
0.995
2720, 237
3530, 250
1.071
3129, 220
1.051
1.023
)
0.148, ꢀ0.158
0.0363, 0.0924 (2204)
0.0495, 0.0991 (2720)
0.535, ꢀ0.293
0.084, 0.2337 (2530)
0.1065, 0.2643 (3530)
0.175, ꢀ0.153
0.0425, 0.1043 (2109)
0.0716, 0.1193 (3129)
pAcPNa
pBocPNb
pAPN
pAPN
temp. (K)
293
293
293
15
chemical formula
chemical formula wt
wavelength (Å)
cell setting, space group
a, b, c (Å)
C15H20N3O3
290.34
C18H26N3O4
348.42
C13H18N3O2
248.3
C13H18N3O2
248.3
0.71073
0.71073
0.71073
0.44280
monoclinic, P21/c
11.2200(2)
12.4690(2)
11.3440(2)
90.0
monoclinic, P21/n
12.0543(3)
9.9704(4)
16.3448(7)
90.0
orthorhombic, Pbca
13.7796(7)
12.3236(6)
16.0651(4)
90.0
orthorhombic, Pbca
13.7284(6)
12.2673(6)
16.0254(8)
90.0
α, β, γ (deg)
103.289(1)
90.0
105.579(3)
90.0
90.0
90.0
90.0
90.0
V (Å3)
Z
1544.55(5)
4
1892.25(12)
4
2728.1(2)
8
2698.8(2)
8
Dcalc (g/cm3)
1.249
1.223
1.209
1.222
F(000)
620
748
1064
1064
μ (mmꢀ1
)
0.088
0.087
0.083
0.053
Rint
0.0461
0.0427
0.0334
0.0689
completeness to 2θ
Nref, param.
goodness of fit on F2
ΔFmax, ΔFmin (e ꢀ Åꢀ3
R1, wR2 (I > 2σ(I))
0.997
0.993
0.997
0.997
2823, 271
0.986
4305, 330
1.00
3113, 202
1.035
3505, 172
1.033
)
0.224, ꢀ0.202
0.0461, 0.1374 (2256)
0.0578, 0.1521 (2823)
0.147, ꢀ0.283
0.049, 0.1058 (2446)
0.1087, 0.1295 (4305)
0.184, ꢀ0.17
0.0598, 0.1436 (1818)
0.1108, 0.1713 (3113)
0.358, ꢀ0.329
0.0622, 0.1756 (2562)
0.0799, 0.1975 (3505)
Rσ1, wR2 (all)
a From CSD TICXAF, CCDC Deposition No. 619977 and ref 11. b From CSD TICWUY, CCDC Deposition No. 619976 and ref 11.
4848
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
Figure 1. ORTEP diagrams showing thermal ellipsoids at 50% probability. Some hydrogen atoms have been omitted in structures for ease of viewing.
Using a singletꢀtriplet spin pairing Hamiltonian H = ꢀ2JS1S2
By comparison to pAcPN, pBocPN has no close contacts
between large nitronylnitroxide spin density sites in its crystal
lattice. It also lacks close contacts of its NO groups with other
sites of spin density that are large enough to yield indirect
exchange pathways. The lack of intermolecular exchange path-
ways is consistent with the essentially paramagnetic, isolated spin
behavior of pBocPN. Its 1/χ versus T plot (Figure 8) fitted over
the full temperature range yields a Curie constant of 0.372
emu K/Oe mol, in excellent agreement with the value expected
and eq 2:22
"
#
2Ng2β2
3
χ ¼
ð1 ꢀ FÞ
3
3kðT ꢀ θMFÞ 3 þ expð2J=kTÞ
1
þ ðFÞ 0:375
ð2Þ
3
T
3
3
for S = 1/2 radicals; the Weiss constant derived by linear fit over
the full temperature range is only θ = (ꢀ)0.2 K and that for the
data with T > 100 K only θ = (+)0.5 K. The full temperature
Weiss constant is more consistent with an observed small
downturn in χT versus T (not shown), indicating very small
AFM exchange in the lattice.
where the singletꢀtriplet energy splitting is 2J/k, using a mean
field correction of θMF, where F is a correction for paramagnetic
sites in the lattice, and other terms have the usual meanings, we
found g = 1.998, 2J/k = (ꢀ)5.0 K, and θMF = (ꢀ)0.86 K,
assuming no paramagnetic impurities (F = 0). These are in close
agreement with those from our preliminary report for pAcPN.11
Using the room temperature crystal structure for pAcPN, the
computationally estimated dyad singletꢀtriplet energy splitting
is 2J/k = (ꢀ)2.4 K at the UB97D/6-31+G(d) level and (+)1.2 K
at the UB3LYP/6-31G* level. The qualitatively different com-
puted results possibly arise as a result of the competition of the
FM and AFM contacts at different levels of theory.
Aminophenyl Nitronylnitroxides: Networked Hydrogen
Bonding Cases. As we hoped, when choosing amino groups as
hydrogen bonding units, each mAPN ꢀN18(H2) amino group
hydrogen bonds with nitronylnitroxide N7ꢀO11 groups from
two different molecules, contacts a and b in Figure 9. Overall,
a 3-D network results involving these hydrogen bonds, as well as
4849
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
N9ꢀO10 to nitronylnitroxide C5 methyl group contacts (not
the CurieꢀWeiss plot, the fit is significantly better, with 2J/k =
(ꢀ)62.6 K, θMF = (ꢀ)17.9 K, and F = 6%. The fitted curves are
shown in Figure 10b. With or without the mean-field correction,
the magnetic data show fairly strong AFM interactions, which the
shown). The a hydrogen bond has r(N18(H)2 O110a) =
3 3 3
3.111 Å, with —N18 O110ꢀN70a = 127.8ꢀ; the b hydrogen
3 3 3
bond has r(N18(H)2 O1100a) = 3.168 Å, with —N18
3 3 3
3 3 3
O1100ꢀN700b = 119.1ꢀ. The N9ꢀO10 nitronylnitroxide
groups are not hydrogen bonded with ꢀNH2. Instead,
crystallographic analysis attribute to the antiparallel NꢀO
3 3 3
OꢀN stack. The hydrogen bonding in mAPN thus forms a
scaffold that holds the radical groups in proximity to produce this
interaction. The computationally estimated dyad singletꢀtriplet
they form NꢀO OꢀN antiparallel-stacked dyads related by
3 3 3
0
inversion symmetry (contact c in Figure 9), an rO/O type
contact with r(N9ꢀO10 O100N90) = 3.537 Å and j =
energy splitting across the antiparallel NꢀO OꢀN stack is
0 3 3 3
3 3 3
—N9ꢀO10 O100(N) = 76ꢀ. The acute angle of j with
ΔE(S f T) = 2J/k = (ꢀ)30.5 K at the UB97D/6-31+G(d) level,
and (ꢀ)8.6 K at the UB3LYP/6-31G* level.
3 3 3
0
0
relatively small rO/O and rO/N (Table 2) provides significant
SOMOꢀSOMO overlap between the nitronylnitroxide units,
with significant AFM exchange interaction expected.
The crystallography of radical pAPN has significant differences
from that of mAPN. A private communication to the Cambridge
Structural Database of the crystal structure only for pAPN gives23
essentially the same structure as that found in the present
work. Similarly to mAPN, pAPN forms an intricate network of
hydrogen bonds. However, in pAPN every ꢀNH2 group not
only bridges two nitronylnitroxide NꢀO units by close contacts
but also every nitronylnitroxide group is hydrogen bonded to
two ꢀNH2 groups. Figure 11 shows the contacts schematically.
The hydrogen bonded network forms 2-D corrugated sheets
lying in the crystallographic bc-plane (shown in the Supporting
A plot of 1/χ versus T for mAPN yields a Curie constant of
C = 0.368 emu K/Oe mol and a large, AFM Weiss constant of
3
3
θ = (ꢀ)36.2 K by the linear fitting of the T > 150 K data: the
plot deviates strongly from linearity below 100 K (Figure 10a).
The χ versus T data show a maximum at 40ꢀ50 K (Figure 10b),
characteristic of AFM exchange. At 1.8 K, magnetization versus
field experiments saturate at a small value (Supporting Informa-
tion) attributable to some paramagnetic sites in the lattice—
most magnetic moment is lost at low temperature, consistent
with spin pairing. A nonlinear least squares fit to the χ versus T
data, using the singletꢀtriplet susceptibility eq 1 with fitting of
0
Information). Figure 11 shows the closest rO/O type contact in
pAPN (contact c) across an inversion center with r([N2]ꢀ
1
J/k and S = /2 paramagnetic fraction F (same Hamiltonian as
O2 O20ꢀ[N20]) = 3.60 Å but with a very obtuse j =
3 3 3
eq 1), yields 2J/k = (ꢀ)83.0 K, with F = 6%, using a fixed g = 1.9.
—N2ꢀO2 O20ꢀ(N) = 136ꢀ. These contacts occur along
3 3 3
If the mean field correction of eq 1 is included using g = 1.98 from
Scheme 3. Geometric Close Contact Parameters for Inter-
Radical Exchange
Figure 2. Important hydrogen bonding and NꢀO OꢀN intermo-
3 3 3
lecular close contacts in mAcPN.
Table 2. Experimental and Computer Modeled Exchangea
computed NO ON dyad
exchange, 2J/k UB3LYP, [UB97D]c
3 3 3
exptl exchangeb
ÆS2HSæ,ÆS2LSæd
rO/O (Å)
rO/N (Å)
4.91e, 4.57f 118g, 143h
j (deg)
ζ (deg)
ω (deg)
0
0
mAcPN θ = (ꢀ) 0.3 K
(ꢀ)0.25 K, [(+) 0.35 K]
(not computed)
2.11, 1.11 [2.03, 1.03]
3.83
3.97
3.54
4.10
3.60
10
43
0
48i, 127j
122o, 85p
mBOC
mAPN
pAcPN
pAPN
θ = (ꢀ) 0.8 K
4.08k, 4.53l
108m, 86n
2J/k = (ꢀ) 83.0 K
2J/k = (ꢀ) 5.0 K
θ = (ꢀ) 3.4 K
(ꢀ)8.6 K, [(ꢀ) 30.5 K]
(+)0.60 K, [(ꢀ) 4.8 K]
(ꢀ)0.2 K, [(+) 1.4 K]
2.12, 1.12 [2.03, 1.03]
2.11, 1.11 [2.03, 1.03]
2.11, 1.11 [2.03, 1.03]
3.46
76
74.5
3.93
73
9
77
43
4.62
136
0
a Room temperature NO ON intermolecular dyad contact parameters from definitions in Scheme 3. Double entries have angles calculated starting
3 3 3
b
c
0
0
from the NO group that corresponds to the rO/N distance (from NO to the nearby N ). Negative values for AFM exchange. Computed gaps using
eq 1; UB97D values in brackets, [ ]. d UB97D values in brackets, [ ]. e N9 O11. f N7 O10. g N7ꢀO11 O10. h N9ꢀO10 O11.
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
i C5ꢀN7ꢀO11 O10. j C3ꢀN9ꢀO10 O11. k N7 O10. l N9 O11. m N9ꢀO10 O11. n N7ꢀO11 O10. o C3ꢀN9ꢀO10 O11.
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
3 3 3
p C4ꢀN7ꢀO11 O10.
3 3 3
4850
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
Figure 3. Paramagnetic susceptibility (χ) for mAcPN as χT vs T (chart a) and 1/χ vs T (chart b) plots, at 1000 Oe external field.
Scheme 4. Hydrogen Bonding Contacts and Qualitative Spin
Density Distributions in mBocPN
Figure 4. Important hydrogen bonding and NꢀO OꢀN intermo-
3 3 3
lecular close contacts in mBocPN.
the crystallographic a-axis, across boundaries between the hydro-
gen-bonded 2-D sheets. Because of the obtuse j parameter
(Table 2), the NꢀO groups are oriented nearly “head-on” with
very little SOMOꢀSOMO overlap possible, although they are
close enough for some dipoleꢀdipole exchange. This NꢀO
3 3 3
OꢀN contact c also involves close NꢀO HꢀC(aryl) con-
3 3 3
tacts involving the phenyl ring CꢀH that is ortho to the radical
group connection, as shown in Figure 11. With so many atoms
held in close proximity and some of them having large spin
densities, it is not straightforward to analyze qualitatively what
exchange is expected to occur.
The computationally estimated dyad singletꢀtriplet energy
splitting across the pAPN NꢀO OꢀN contact was 2J/k =
3 3 3
(+)1.5 K at the UB97D/6-31+G(d) level; at the UB3LYP/
6-31G* level it is 2J/k = (ꢀ)0.2ꢀ0.3 K. The much smaller
interaction than in the case of mAPN is presumably due to the head-
on orientation of the contact, rather than parallel overlapping
geometry. The small computed interaction suggests that low
temperature changes in crystal contacts could change exchange
behavior relative to higher temperatures. The following analysis of
intermolecular contacts is thus only indicative of qualitative trends
and only if exchange contacts do not change significantly.
Figure 5. Reciprocal paramagnetic susceptibility 1/χ vs T plot for
mBocPN, at 1000 Oe external field.
b contact is longer because H3B is pointed toward an NꢀO
0
SOMO, not toward a lone pair as in a classic hydrogen bond.
The largest exchange component is expected from the rO/O
N2ꢀO2 O20ꢀN20 contact, as a result of the large spin
The a NꢀH OꢀN contact gives near-orthogonality of
3 3 3
3 3 3
the hydrogen s-orbital and the OꢀN SOMO, while the b
densities involved. However, the direct NꢀH OꢀN contacts
3 3 3
NꢀH
OꢀN contact gives good overlap of the analogous
might provide indirect exchange pathways through the small
3 3 3
orbitals; these are shown between molecules Y and Z in
Scheme 5. The two types of hydrogen bonds could give
qualitatively different exchange interactions, as shown in
Scheme 5 using the spin densities computed for an isolated
pAPN molecule (see Supporting Information). Taylor
and Lahti noted5 an analogous difference in comparing
expected exchange in phenolic nitronylnitroxides having direct
spin density on NꢀH. The a N3ꢀH3A O1ꢀN1 contact has
3 3 3
its NꢀH bond pointing almost directly toward an NꢀO lone
pair, and this will not provide close overlap of spin density on the
NH hydrogen with the nitronylnitroxide spin orbital. However,
the b contact N3ꢀH3B O2ꢀN2 has its NꢀH bond pointing
3 3 3
almost directly at the oxygen lobe of an NꢀO SOMO. Notably,
r(N3[H3A] O1) for the a contact is much shorter, 2.936(3)
3 3 3
OꢀH OꢀN contacts with different geometries. Using the
Å versus the b contact of r(N3[H3B] O2) = 3.402(4) Å. The
3 3 3
3 3 3
4851
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
Figure 6. Hydrogen bonding and NꢀO OꢀN intermolecular close contacts in pAcPN (upper) and pBocPN (lower).
3 3 3
Figure 7. Paramagnetic susceptibility χ vs T plot for pAcPN, at 1000 Oe
external field.
Figure 9. Important hydrogen bonding and NꢀO OꢀN intermo-
3 3 3
lecular close contacts in mAPN in the crystallographic ac-plane.
UB3LYP/6-31G* level, 2Ja/k = (+)0.03 K and 2Jb/k = (ꢀ)0.25 K.
The computed numbers are so small that it is hard to depend
strongly qualitative dyad exchange model of the a and b contacts
in Scheme 5. Also, NꢀH bond lengths and positions from the
crystallographic analysis are subject to the uncertainties of estimating
the hydrogen atom positions.
The experimental magnetic behavior for pAPN is shown in
Figure 12. The χ versus T data lie below the paramagnetic
Curie curve, indicating AFM exchange. The higher temperature
region of the 1/χ versus T plot yields a Curie constant of
0.363 emu K/Oe mol and an extrapolated Weiss constant of
θ = (ꢀ)3.4 K but with a change of slope at about 5 K; the 2ꢀ5 K
region extrapolates to θ = (ꢀ)0.3 K. The antiferromagnetic
interactions in higher temperature region are also seen in the χT
versus T plot, which decreases significantly in the range 50ꢀ3 K.
However, the decrease levels out at 2ꢀ3 K, so we also investi-
gated the susceptibility behavior in the range 0.5ꢀ3.6 K.
Figure 8. Reciprocal paramagnetic susceptibility 1/χ vs T plot for
pBocPN, at 1000 Oe external field.
3
3
crystallographic geometries used to formulate the qualitative des-
criptions of Scheme 5, the a NꢀH OꢀN contact is computed
3 3 3
to have 2J/k = (+)0.15 K and the b NꢀH OꢀN contact has
3 3 3
2J/k = (ꢀ)0.03 K, both at the UB97D/6-31+G(d) level. At the
4852
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
Interestingly, χT versus T increases again after minimizing at
about 2 K, indicating competing exchange pathways in pAPN.
A 1/χ versus T plot of the 0.6ꢀ3.6 K data yields a small positive
Weiss constant of (+)0.26 K, with a Curie constant of only about
0.2 emu K/Oe mol. The Curie constant is also consistent with
3
3
the molar magnetization versus field data at 1.8 K, which appear
to approach saturation at about 2/3 the value expected for 1 mol
1
of S = /2 radicals. The behavior suggests multiple exchange
pathways are present, although there is only one crystallo-
graphic form of pAPN in the lattice. As the temperature
decreases, a dominant AFM exchange interaction gives the
downturn of χT versus T, yielding the significant negative
Weiss constant of the higher temperature region of the 1/χ
versus T plot. Presumably, the relatively close AFM contact c,
(N2)ꢀO2
O20ꢀ(N20), between molecules W and X in
3 3 3
Scheme 5, dominates exchange in the higher temperature
region. At lower temperatures, small FM interactions—pos-
sibly through the hydrogen bonding network—cause the χT
versus T plot to rise again.
Scheme 5. Hydrogen Bonding Contacts and Qualitative Spin
Density Distributions in pAPN
Figure 10. Paramagnetic susceptibility data plots for mAPN, as 1/χ vs T
(a), with linear fitted line from data where T > 150 K, and χ vs T (b), with
a broken line for the fitted curve to eq 1 without mean field and a solid
line for the curve with the mean field included. All data obtained using
1000 Oe external field.
Figure 11. Important hydrogen bonding and NꢀO OꢀN intermolecular close contacts in pAPN in the crystallographic bc-plane.
3 3 3
4853
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
Figure 12. For pAPN: (a) paramagnetic susceptibility χ vs T plot showing a noninteracting spins Curie line; (b) CurieꢀWeiss 1/χ vs T plot showing
linear fits to data at higher (solid line) and lower (broken line) temperature regions; (c) χT vs T plot; and (d) low temperature 1/χ vs T plot showing
linear fit to data. Data obtained at T g 1.8 K at 1000 Oe are shown as open triangles in chart a; data obtained at 0.6 K < T < 4 K at zero field, by
ac-susceptibility (modulation field 5 Oe, modulation frequency 155 Hz) are shown as open circles.
in Table 1, the crystallographic axes contract by only 0.3ꢀ0.5%;
the largest absolute change is 0.056 Å along the b-axis. The unit
cell volume contracts by only 1.1%. A crystal packing similarity
analysis24 of the structures at both temperatures gives a root
mean square (rms) difference of only 0.03. The high resistance to
contraction for pAPN upon cooling is attributable to the highly
networked array of hydrogen bonds, which rigidify the lattice
packing by comparison to what would be expected for more
localized interactions. It seems unlikely that cooling below 15 K
would lead to additional major changes to the crystal lattice.
The complexity of the contact network in Figure 11 makes it
challenging to identify any structure feature as the cause of the
lower temperature magnetic behavior. As Novoa and co-workers
have pointed out,21 it is very risky to attribute magnetic behavior
to any one or limited number of intermolecular contacts in an
organic molecular solid, unless the contact creates an exchange
interaction that greatly dominates. That is not the case here. As
a result, one must be satisfied to note that competing, small
exchange effects seem to be present in pAPN.
Aminophenyl Nitronylnitroxide Acid Salts. The Coulombic
forces in ionic solids can produce crystallization motifs and
exchange effects in acid salts that are quite different from those
in corresponding neutral solids. Therefore, pAPN was treated
with HBF4/methanol or HCl in diethyl ether to yield salts
Figure 13. UVꢀvis spectra in methanol for (a) pAPN HCl at 878 μM
pAPN HBF4 and pAPN HCl. Interestingly, the salts showed
3
3
3
(i), 220 μM (ii), 55 μM (iii), and 13.8 μM (iv); (b) pAPN at 805 μM (v),
only very weak EPR spectra as neat solids, but strong 1:3:5:3:1
pentet spectra with aN = 7.8 (HBF4 salt) and 8.2 gauss (HCl salt)
when dissolved in methanol. Unfortunately, solutions of the
radical salts decomposed completely within 24ꢀ48 h; so, we
could not obtain diffraction quality crystals of either. Ullman
proposed that nitronylnitroxides disproportionate to nonradical
forms upon exposure to acid;25 such a process may decompose
pAPN HBF4 and pAPN HCl in solution. The salts were much
201 μM (vi), and 12.6 μM (vii). The letters in the chart a inset box at
long wavelength show the descending order of the lines at that point.
We considered the possibility that the network of contacts in
pAPN changes as temperature decreases, changing the exchange
behavior. However, the pAPN crystal structure at 15 K remains
remarkably similar to the room temperature structure. As shown
3
3
4854
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
more stable as solids, producing strong EPR spectra when
dissolved after months of storage under ambient conditions.
Green-colored solutions of either salt in methanol turn azure
blue when treated with KOH. The green color is regenerated
upon addition of acid to basic solutions (see Supporting In-
formation). The color changes can be cycled multiple times, so
long as the salt solution does not stand for more than a few hours.
Consistent with their lack of strong EPR signals as neat solids,
pAPN HBF4 and pAPN HCl samples exhibited little or no
paramagnetic susceptibility. The EPR and colorimetric experi-
ments show that the nitronylnitroxide groups remain present
in the solids. Apparently there is very strong AFM exchange
between radical sites, J/k , (ꢀ)300 K. FTIR spectra show
strong, broad bands over 2100ꢀ3200 cmꢀ1, consistent with the
strong hydrogen bonding of NꢀH groups in an ionic salt.
Without the crystal packing information, one can only speculate
about what structural features give the strong spin pairing.
Hydrogen bonding must bring the radical groups into unusually
close proximity to one another, in a geometry having strong
overlap of their spin orbitals for strong AFM exchange.
computational results reported in this article. This material is
available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*lahti@chem.umass.edu.
Present Addresses
Department of Chemistry, University of Mississippi, University,
Mississippi 38677.
3
3
’ ACKNOWLEDGMENT
This work was supported by the U.S. National Science
Foundation (CHE-0415716 [S.A. and P.M.L.], CHE-0809791
[P.M.L.]; synthesis, characterization of all molecules). We thank
Dr. A. Chandrasekaran and Dr. P. Khalifah for assistance with
crystallographic analyses carried out at the University of Massa-
chusetts. A.P.F., V.B., and N.F.O. thank the Fundac-~ao de Ampara
ꢀa Pesquisa do Estado de S~ao Paulo, Brazil (FAPESP-07/50968-
0) for support to carry out the low temperature magnetic study of
pAPN. Use of the Advanced Photon Source was supported by the
U.S. Department of Energy, Office of Science, Office of Basic
Energy Sciences, under Contract No. DE-AC02-06CH11357.
ChemMatCARS Sector 15 is principally supported by the
National Science Foundation/Department of Energy under
Grant No. NSF/CHE-0822838.
Although we lack crystallographic data for pAPN HBF4 and
3
pAPN HCl, their solution UVꢀvis spectra clearly show signifi-
3
cant inter-radical aggregation. At about 10 μM concentrations in
methanol, the spectra of the free amine and the salts are quite
similar, save that the salt has a much-decreased molar absorptivity
of its long wavelength nitronylnitroxide absorption at 650 nm.
This band may be due to the small amount of free amine in
equilibrium with the ammonium salt. At concentrations >200
μM, the ammonium salts exhibit a new, shoulder band at 420 nm,
not seen at any concentration of the free amine. The comparative
’ REFERENCES
behavior of pAPN and pAPN HCl is shown in Figure 13, and it
3
indicates significant aggregation of the aminium salt.
(1) (a) Miller, J. S., Gatteschi, D., Eds.; Chem. Soc. Rev., 2011, 40,
3053ꢀ3368. (b) Magnetism: Molecules to Materials; Miller, J. S., Drillon,
M., Eds.; Wiley-VCH: Weinheim, Germany, 2001ꢀ2005; Vols. 1ꢀ5.
(c) Kahn, O. Molecular Magnetism; VCH: New York, 1993. (c)
Molecular Magnetism: New Magnetic Materials; Itoh, K, Kinoshita, M,
Eds.; Gordon and Breach, Kodansha Ltd.: Tokyo, Japan, 2000. (d)
Magnetic Properties of Organic Materials; Lahti, P. M., Ed.; Marcel-
Dekker: New York, 1999. (e) Carbon-Based Magnetism: An Overview
of the Magnetism of Metal-Free Carbon-Based Compounds and Materials;
Makarova, T., Palacio, F., Eds.; Elsevier: Amsterdam, Netherlands, 2006.
(f) Fovira, C., Veciana, J., Eds.; Crystal. Eng. Comm. 2009, 11, 2017ꢀ
2212.
(2) (a) Sugawara, T.; Matsushita, M. M.; Izuoka, A.; Wada, N.;
Takeda, N.; Ishikawa, M. Chem. Commun. 1994, 1723–4. (b) Matsushita,
M. M.; Izuoka, A.; Sugawara, T. Mol. Cryst. Liq. Cryst. Sci. Technol., Sect. A
1996, 279, 139–144. (c) Matsushita, M. M.; Izuoka, A.; Sugawara, T.;
Kobayashi, T.; Wada, N.; Takeda, N.; Ishikawa, M. J. Am. Chem. Soc. 1997,
119, 4369–4379.
(3) (a) Garcia-Munoz, J. L.; Cirujeda, J.; Veciana, J.; Cox, S. F. J.
Chem. Phys. Lett. 1998, 293, 160–166. (b) Veciana, J.; Cirujeda, J.;
Rovira, C.; Vidal-Gancedo, J. Adv. Mater. 1995, 7, 221–225. (c) Cirujeda,
J.; Mas, M.; Molins, E.; Lanfranc de Panthou, F.; Laugier, J.; Park, J. G.;
Paulsen, C.; Rey, P.; Rovira, C.; Veciana, J. Chem. Commun. 1995,
709–710. (d) Cirujeda, J.; Hernandez-Gasio, E.; Rovira, C.; Stanger,
J.-L.; Turek, P.; Veciana, J. J. Mater. Chem. 1995, 5, 243–52. (e) Heise,
H.; Koehler, F. H.; Mota, F.; Novoa, J. J.; Veciana, J. J. Am. Chem. Soc.
1999, 121, 9659–9667. (f) Cirujeda, J.; Vidal-Gancedo, J.; Juergens, O.;
Mota, F.; Novoa, J. J.; Rovira, C.; Veciana, J. J. Am. Chem. Soc. 2000,
122, 11393–1140. (g) Cirujeda, J.; Ochando, L. E.; Amigo, J. M.; Rovira,
C.; Rius, J.; Veciana, J. Angew. Chem., Int. Ed. Engl. 1995, 34, 55–57.
(4) Matsushita, M. M.; Izuoka, A.; Sugawara, T.; Kobayashi, T.;
Wada, N.; Takeda, N.; Ishikawa, M. J. Am. Chem. Soc. 1997, 119,
4369–4379.
Conclusion. Aminophenyl nitronylnitroxides, having two
NꢀH donor sites for hydrogen bonds, form more extensive
hydrogen bonding in their NꢀH crystal lattices than similar
compounds having only one NꢀH or OꢀH donor site available.
For most of the systems, the strongest magnetic exchange
interactions are attributable to direct intermolecular contacts
having good spin orbital overlap between high spin density
nitronylnitroxide sites, typically involving NꢀO units. Only
pAPN shows indications that NꢀH hydrogen bonding to the
radical units may play a direct electronic exchange role in the
magnetic behavior. The use of versatile amino groups in these
radicals offers promise for further structureꢀproperty modification,
as exemplified by conversion of pAPN to ammonium salts.
However, it seems likely that strategies to raise the critical magnetic
ordering temperatures of pure organic radicals (especially for those
based solely on elements of the first two rows of the periodic table)
will require designs in which directional molecular assembly
moieties such as hydrogen bonds can also bear significant spin
density on the hydrogen bond donors and acceptors. Doing this
would improve prospects for multidimensional exchange inter-
actions to be propagated throughout the molecular material.
’ ASSOCIATED CONTENT
S
Supporting Information. Crystallographic CIF sum-
b
maries; figure of 2-D sheet packing in pAPN; Mecury24 format
files of extended hydrogen bonding in mAPN and pAPN; pH-
variable color picture of pAPN solutions; infrared spectra of all
compounds and the ammonium salts; magnetization versus field
plots for all radicals in this study; archive information for
(5) Taylor, P.; Lahti, P. M. Chem. Commun. 2004, 2686–2688.
4855
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856

Chemistry of Materials
ARTICLE
(6) Ovcharenko, V.; Fokin, S.; Rey, P. Mol. Cryst. Liq. Cryst. Sci.
Technol., Sect. A 1999, 334, 109.
(7) Grice, R.; Owen, L. N. J. Chem. Soc. 1963, 1947–1954.
(8) Friedlaender, P.; Fritsch, R. Monatsh. Chem. 1903, 24, 1–12.
(9) Rai, R.; Katzellenbogen, J. J. Med. Chem. 1992, 35, 4150–4159.
(10) Sheldrick, G. M. SHELXTL97, Program for the Refinement of
Crystal Structures; University of G€ottingen: Germany.
(11) Aboaku, S.; Lahti, P. M. Polyhedron 2007, 26, 1959–1964.
(12) Oliveira, N. F., Jr.; Paduan-Filho, A.; Salinas, S. R.; Becerra,
C. C. Phys. Rev. B 1978, 18, 6165.
(13) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (b)
Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (c) Stephens, P. J.; Devlin,
F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623.
(14) Frisch, M. J.; et al. Gaussian 03, B03 ed.; Gaussian, Inc.:
Pittsburgh, PA, 2003. The full citiation is given in the Supporting
Information.
(15) Grimme, S. J. Comput. Chem. 2006, 27, 1787–1799.
(16) Davidson, E. R. Int. J. Quantum Chem. 1998, 69, 241–245.
(17) Yamaguchi, K.; Fukui, H.; Fueno, T. Chem. Lett. 1986,
625–628.
(18) Yamaguchi, K. Chem. Phys. Lett. 1988, 149, 537–542.
(19) Program ORTEP3:Farrugia, L. J. J. Appl. Crystallogr. 1997,
30, 565.
(20) Cf. Yamaguchi, K.; Kawakami, T.; Oda, A.; Yoshioka, Y. In
Magnetic Properties of Organic Materials; Lahti, P. M., Ed.; Marcel
Dekker: New York, NY, 1999; pp 403ꢀ425.
(21) Deumal, M.; Ciujeda, J.; Veciana, J.; Novoa, J. J. Chem.—Eur. J.
1999, 5, 1631–1542.
(22) Bleaney, B.; Bowers, K. D. Proc. R. Soc. London, Ser. A 1952, 214.
(23) Wagner, B.; Gompper, R.; Polborn, K. CCDC Deposition
#192763, CSD Code MUKKUY, as a private communication; 2002.
(24) Similarity analysis was carried out using the program Mercury,
version 2.3, build RC4, from the 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.
(25) Ullman, E. F.; Osiecki, J. H.; Boocock, D. G. B.; Darcy, R. J. Am.
Chem. Soc. 1972, 94, 7049–7059.
4856
dx.doi.org/10.1021/cm202303q |Chem. Mater. 2011, 23, 4844–4856