Synthesis, crystal structures, and solid state quadratic nonlinear optical properties of a series of stilbazolium cations combined with gold cyanide counter-ion

Article abstract of DOI:10.1039/c1jm12105g

Three salts built up from (E)-4′-(dimethylamino)-stilbazolium (DMAS)H⁺, (E)-4′-(diethylamino)-stilbazolium (DEAS)H ⁺, (E)-4′-{2-(methoxymethyl) pyrrolidinyl}-stilbazolium (MPS)H⁺, and gold cyanide as a counter-ion, are rep

Full text of DOI:10.1039/c1jm12105g

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: J. Mater. Chem., 2011, 21, 15940
www.rsc.org/materials
PAPER
Synthesis, crystal structures, and solid state quadratic nonlinear optical
properties of a series of stilbazolium cations combined with gold cyanide
counter-ion†
a
b
c
c
Pascal G. Lacroix, M. Carmen Munoz, Ana Belen Gaspar, Jose Antonio Real, Sebastien Bonhommeau,
d
*
*
ꢀ
ꢀ
ꢀ
d
Vincent Rodriguez and Keitaro Nakatanie
*
Received 12th May 2011, Accepted 8th July 2011
DOI: 10.1039/c1jm12105g
Three salts built up from (E)-40-(dimethylamino)-stilbazolium (DMAS)H+, (E)-40-(diethylamino)-
stilbazolium (DEAS)H+, (E)-40-{2-(methoxymethyl) pyrrolidinyl}-stilbazolium (MPS)H+, and gold
cyanide as a counter-ion, are reported. The crystal structures have been solved for (DEAS)H+ Au
(CN)2ꢀ (Cc space group), and for (MPS)H+ Au(CN)2ꢀ (P1 space group). The semi-empirical (ZINDO)
calculated static hyperpolarizability (b0) of (MPS)H+ is equal to 147 ꢁ 10ꢀ30 cm5 esuꢀ1, in solid state,
which is 25% higher than that of the cation of the well known (E)-40-(dimethylamino)-
methylstilbazolium tosylate (DAST). (MPS)H+ Au(CN)2ꢀ exhibits a unique crystal structure in which
the cations are perfectly aligned. This combination of large hyperpolarizability and strict 1-dimensional
character leads to a giant dzzz quadratic susceptibility (c(2)) tensor component estimated between ꢂ4500
pm Vꢀ1 (ꢂ11 ꢁ 10ꢀ6 esu) and 12 400 pm Vꢀ1 (ꢂ30 ꢁ 10ꢀ6 esu) at 1.064 mm, by a confocal m-SHG
technique. The use of such material in ultra-thin devices is critically evaluated.
component of the light. The polarization (P) of a macroscopic
material is again given by an expression analogous to (1), as
follows:
Introduction
Organic nonlinear optical (NLO) materials have long been
recognized for their potential applications in telecommunica-
tions, all optical data processing and biological imaging.1,2 In
molecular chromophores, the NLO response is ultimately related
to the expression of the molecular polarizability, given by the
following equation:3
P ¼ c(1)E + c(2)E2 + .
(2)
c(2) is the quadratic susceptibility, related to the underlying b and
responsible for the bulk quadratic effects (e.g. second harmonic
generation (SHG) and linear electro-optic behavior). An
important point to note is that, for c(2) to be non zero, the crystal
structure needs to lack centrosymmetry. Efficient NLO materials
are therefore built up in two steps: (i) the design of highly
polarizable species, containing strong donor and acceptor
substituents connected by an extended p-conjugated bridge to
ensure a large b value, and (ii) the engineering of the molecules
into a non-centrosymmetric solid state environment.
m(E) ¼ m0 + aE + bE2 + .
(1)
In this expression, m0 is the permanent dipole moment, a the
linear polarizability and b the first hyperpolarizability (origin of
the quadratic NLO behavior), E being the applied electric field
aLaboratoire de Chimie de Coordination du CNRS, 205 route de Narbonne,
31077 Toulouse, France. E-mail: pascal.lacroix@lcc-toulouse.fr
bDepartament de Fꢀısica Aplicada, Universitat Politeꢀcnica de Valeꢁncia,
Among many potential candidates, the salt (E)-40-(dimethy-
lamino)-N-methyl-stilbazolium
tosylate
(DAMS+
Tsꢀ,
ꢀ
ꢁ
Camı de Vera s/n, 46022 Valencia, Spain
cInstitut de Cieꢁncia Molecular (ICMol), Departament de Quꢀımica
Scheme 1), often called DAST, has become the most promising
system, after the report of an exceptionally large SHG effi-
ciency, up to 1000 times that of urea.4 This material then
reached the step of commercial development in relation to its
quadratic NLO effects (e.g. terahertz wave generation arising
from nonlinear frequency mixing).5 These challenging capabil-
ities have stimulated many synthetic programs for finding
optimized NLO candidates based on the ‘‘push-pull’’ architec-
ture of the benchmark DAST model, and have led to a research
area which remains very active.6
Inorgaꢁnica, Universitat de Valeꢁncia, C/ Catedraꢀtico Jose Beltraꢀn
ꢀ
Martınez, 2, 46980 Paterna, Valencia, Spain. E-mail: jose.a.real@uv.es
dUniversiteꢀ de Bordeaux, Institut des Sciences Moleꢀculaires, CNRS UMR
5255, F-33400 Talence, France. E-mail: v.rodriguez@ism.u-bordeaux1.fr
ePPSM, Institut d’Alembert, Ecole Normale Supeꢀrieure de Cachan,
CNRS, 61 av. du Preꢀsident Wilson, 94230 Cachan, France
† Electronic supplementary information (ESI) available: Details about
the experimental determination of dzzz value. DFT coordinates for the
4 investigated chromophores. CCDC reference numbers 826266 and
826267. For ESI and crystallographic data in CIF or other electronic
format see DOI: 10.1039/c1jm12105g
15940 | J. Mater. Chem., 2011, 21, 15940–15949
This journal is ª The Royal Society of Chemistry 2011

View Article Online
ꢃ
The mixture was heated at 60 C for 2 h. Afterwards 4-(dime-
thylamino)benzaldehyde) or 4-(diethylamino)benzaldehyde) was
added and the mixture was kept at this temperature for 7 h. The
resulting solution was cooled to room temperature and poured
into 150 mL of methanol. Bright yellow precipitates appeared
immediately which was filtered and washed with a mixture of
methanol/cool water (1 : 1) and dried under vacuum. Yield: 55%
(DMAS), 62% (DEAS). 1H NMR (CDCl3) DMAS, d (ppm): 8.49
(H1, 2H, d, J ¼ 4.98 Hz), 7.44 (H2, 2H, d, J ¼ 8.84 Hz), 7.30 (H3,
1H, d, J ¼ 16.12 Hz), 6.80 (H4, 1H, d, J ¼ 16.19 Hz), 7.38 (H5,
2H, d, J ¼ 6.17 Hz), 6.71 (H6, 2H, d, J ¼ 8.84 Hz), 3.02 (H7, 3H,
s). 1H NMR (CDCl3) d (ppm) (DEAS): 8.48 (H1, 2H, d, J ¼ 6.13
Hz), 7.42 (H2, 2H, d, J ¼ 8.89 Hz), 7.35 (H5, 2H, d, J ¼ 6.18 Hz),
7.27 (H3, 1H, d, J ¼ 11.09 Hz), 6.77 (H4, 1H, d, J ¼ 16.17 Hz),
6.67 (H6, 2H, d, J ¼ 8.88 Hz), 3.40 (H7, 2H, q, J ¼ 7.03 Hz), 1.20
(H8, 3H, t, J ¼ 7.04 Hz).
40-[2-(methoxymethyl)pyrrolidinyl]-stilbazole
(MPS).
A
mixture of 4-picoline (3.2 mL, 5.5 ꢁ 10ꢀ2 mol) and NaH (1.32 g,
5.5 ꢁ 10ꢀ2 mol) in 10 mL of distilled DMF was heated for 2 h at
60 ꢃC. Then, 4-[2-(methoxymethyl)pyrrolidinyl] benzaldehyde
(7.2 g, 5.5 ꢁ 10ꢀ2 mol) diluted in 10 mL of DMF was added and
the reaction mixture was heated for 7 more hours at 60 ꢃC. After
cooling, the resulting mixture was poured into 300 mL of water,
which led to a red precipitate. The solid was filtered off, washed
with water, and dried under vacuum. The crude product was then
dissolved in hot heptane, filtered hot, evaporated to dryness, and
finally purified by chromatography on SiO2 (60A 70–200 mm)
using heptane-AcOEt (7/3). Yield: 78% of MPS as a bright yellow
solid. 1H NMR (CDCl3) d 8.488 (d, J ¼ 5.8 Hz, 2H), 7.295 (d, J ¼
6.0 Hz, 2H), 7.405 (d,4 J ¼ 8.8 Hz, 2H), 7.222 (d, J ¼ 16.2 Hz,
1H), 6,764 (d, J ¼ 16.2 Hz, 1H), 6.630 (d, J ¼ 8.8 Hz, 2H), 3.918
(m, 1H), 3.42–3.53 (m, 2H), 3.373 (s, 3H), 3.199 (t, J ¼ 8.9 Hz,
2H), 1.72–2.09 (m, 4H).
Scheme 1 Molecular structures of the 4 investigated chromophores
In the present contribution, we report on the synthesis, crystal
structure, theoretical and experimental (second harmonic
generation SHG measurement) NLO studies of new salts built up
from the promising DAMS+ skeleton, combined with [Au
(CN)2]ꢀ anions. This anion was initially chosen for its capability
to provide extended structures with iron(II), with the aim of
making a material with additional magnetic properties.7 The
NLO cations investigated in the present contribution are
(E)-40-(dimethylamino)-stilbazolium (DMAS)H+, (E)-40-(dieth-
ylamino)-stilbazolium (DEAS)H+, and (E)-40-{2-(methox-
ymethyl)pyrrolidinyl}-stilbazolium (MPS)H+ (Scheme 1), and
lead to the three resulting {(DMAS)H+ [Au(CN)2]ꢀ} (1),
{(DEAS)H+ [Au(CN)2]ꢀ} (2), and {(MPS)H+ [Au(CN)2]ꢀ} (3)
materials. Among them, 3 is an intriguing NLO candidate, with
a larger hyperpolarizability than DAST and a perfect alignment
in solid state. Owing to these unique characteristics, additional
NLO investigations were conducted by the non-traditional
confocal m-SHG technique to target the dzzz component of the
c(2) tensor. In the last section, this nearly 1-dimensional system is
critically evaluated for potential applications as advanced
materials in thin films devices.
Synthesis and characterization of the stilbazolium salts. 1, 2,
and 3 were formed accidentally investigating appropriate
synthetic strategies to get iron(II) spin crossover coordination
polymers susceptible of displaying ONL response. The pyridine
nitrogen atom of DEAS and MPS ligands have strong
tendency to protonate, affording the corresponding stilbazo-
lium cation, which stabilizes via hydrogen bonding. The
synthesis was performed by mixing under continuous stirring in
an argon atmosphere a methanolic solution (10 mL) of Fe
(BF4)2$6H2O (0.148 mmol, 50 mg) and a solution of DMAS,
DEAS or MPS ligands (0.30 mmol; 67.5, 75.7 and 88.3 mg,
respectively) in dichloromethane (10 mL). To this mixture,
a water solution (5 mL) of KAu(CN)2 (0.30 mmol, 86.4 mg)
was added drop-wise under stirring. The resulting solution was
allowed to evaporate at room temperature using an argon
stream. The same results have been obtained excluding the iron
(II) salt. Red (1, 2) and pink (3) crystalline samples were formed
in two weeks (yield ca. 60%). Appropriate single crystals were
picked out from these samples. Anal. Calcd for C17H17AuN4
(1): C, 43.05; H, 3.61; N, 11.81; Found. C, 42.68; H, 3.61; N,
11.63. Anal. Calcd for C19H21AuN4 (2): C, 45.43; H, 4.21; N,
11.15; Found. C, 44.93; H, 4.12; N, 11.03. Anal. Calcd for
C21H23AuN4O (3): C, 46.33; H, 4.26; N, 10.29; Found. C,
46.11; H, 4.19; N, 10.13.
Experimental section
Starting materials and equipment
K[Au(CN)2] and all organic reagents and solvents were
1
purchased from commercial sources and used as received. H
NMR spectroscopic measurements were done with an Avance
DRX Bruker 300 MHz spectrometer.
Synthesis
40-dimethyl- and 40-diethylaminostilbazole (DMAS and DEAS).
The synthesis was performed under argon atmosphere and all
glass materials used were dried previously. 4-methylpyridine
(1.78 mL, 20 mmol) and sodium hydride (0.5 g, 21 mmol) were
mixed with 10 mL of distilled N,N0-dimethylformamide (DMF).
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 15940–15949 | 15941

View Article Online
X-ray data collection and structure determination
approach, the mono-excited configuration interaction (MECI)
approximation was employed to describe the excited states. The
lowest 100 energy transitions were chosen to undergo CI mixing.
All calculations were performed using the INDO/1 Hamiltonian
incorporated in the commercially available software package
ZINDO.13 Metrical parameters used for the calculations of
(DEAS)H+ and (MPS)H+, were those obtained from the present
crystal structure determination of 2 and 3, respectively. The
proton linked to the pyridine moieties of the cation, which was
not observed experimentally, was assumed to be located at 0.96
Single crystal X-ray diffraction. Diffraction data for 2, and 3
were collected with a Nonius Kappa-CCD single crystal
ꢂ
diffractometer using Mo Ka radiation (l ¼ 0.71073 A). The
structures were solved by direct methods using SHELXS-97 and
refined by full-matrix least-squares on F2 using SHELXL-97.8 All
non-hydrogen atoms were refined anisotropically. The crystal-
lographic data are summarized in Table 1.
ꢂ
ꢂ
A, according to the report of a N–H bond refined at 0.96(2) A in
related dimethylamino-stilbazolium crystal structure.14
Theoretical methods
a
The all-valence INDO (intermediate neglect of differential
overlap) method,9,10 in connection with the sum over state (SOS)
formalism was employed for the calculation of the electronic
spectra and the molecular hyperpolarizabilities of the organic
cations in the solid state.11 Details for the computationally effi-
cient INDO-SOS-based method for describing molecular optical
nonlinearities have been reported elsewhere.12 In the present
However, we have checked that a modification of this distance in
ꢂ
the 0.91–1.01
A
range affects only slightly the hyper-
polarizability, with static b0 values decreasing from 147.7 to
146.3 ꢁ 10ꢀ30 cm5 esuꢀ1. The calculation performed on DAMS+
for comparison was based on the structure of DAST, previously
reported in the literature.15
Additionally, the properties in MeCN have been computed
after full optimization of the molecular structures, using the
Gaussian09 program package16 within the framework of the
DFT at the B3LYP/6-31G** level,17 the polarizable continuum
model (PCM) being used to account for solvent effect (acetoni-
trile). The ZINDO spectra were calculated from the first 100
excited states.
Table 1 Crystal data for 2, and 3a
2
3
Crystal Data
Empirical formula
Molecular weight
Crystal size/mm
Crystal system
Space group
C19H21AuN4
501.36
0.04 ꢁ 0.06 ꢁ 0.06
Monoclinic
Cc
25.731(2)
8.0720(7)
18.411(2)
90
106.441(3)
90
C21H23AuN4O
544.40
0.01 ꢁ 0.02 ꢁ 0.02
Triclinic
P1
6.615(1)
7.6610(3)
10.2670(6)
83.806(2)
83.545(2)
77.497(3)
502.87(4)
1
NLO measurements
The nonlinear optical properties were evaluated in the solid state
as the efficiency in second harmonic generation (SHG) expressed
versus that of a reference (urea). The measurements were per-
formed by the Kurtz-Perry powder technique,18 using a nano-
second Nd-YAG pulsed (10 Hz) laser operating at l ¼ 1.064 mm.
Due to a residual absorption of stilbazole-based chromophores
at the second harmonic (l ¼ 532 nm), the fundamental incident
wavelength used for the experiments was that of the outgoing
Stokes-shifted radiation (l ¼ 1.907 mm) generated by Raman
effect in a hydrogen cell (1 m long, 50 atm.). The SHG signal was
selected through a suitable interference filter, detected by a pho-
tomultiplier, and recorded on an ultrafast Tektronic TDS 620 B
oscilloscope. Samples were powders obtained by grinding, and
pressed between two glass plates. They were calibrated in the
50–80 mm range, by use of standard sieves. The thickness of the
samples was 200 mm.
ꢂ
a/A
ꢂ
b/A
ꢂ
c/A
a/ꢃ
b/ꢃ
g/ꢃ
V/A
Z
rc/Mg mꢀ3
3
ꢂ
3667.6(5)
8
1.816
1.798
Data Collection
T/K
2q range
Radiation (Mo Ka)/A
F(000)
Scan mode
No. of reflections
Measured
293
293
3.17 < 2q < 27.54
0.71073
264
q–2q
3.19 < 2q < 24.37
0.71073
1928
ꢂ
q–2q
5288
3520
8.029
Empirical
0.7463
0.8924
3270
3104
7.331
Empirical
0.8672
0.9303
Unique
Abs. coef. (m)
Absorption corrections
Tmin
Owing to the strong 1-dimensional character of 3, in solid
state, the value of dzzz (only non-vanishing tensor component of
c(2)) was determined. For this investigation, the technique of
confocal m-SHG was employed using a modified m-Raman
spectrometer (Horiba HR800) so that the backward waves are
analyzed.19 This technique of high depth selectivity allows the
investigation of the NLO response within a volume restricted to
the vicinity of the surface (depth of focus), thus avoiding
coherent effects and most of the absorption effect occurring in an
experiment conducted by transmission.20 In the present case, the
thickness of the crystals of 3 were equal to ca. 2 mm that mini-
mizes absorption losses at the harmonic waves. Hence, no
correction for absorption loss has been considered. The source
was a 1064 nm diode pumped picosecond laser (EKSPLA
PL2200: pulse duration 65 ps, repetition rate 2 kHz) focused at
Tmax
Refinement
Refined on
No. of variable
H-atom treatment
R (>2s(I))
F2
434
F2
249
Mixed
0.0522
0.0911
0.774
ꢀ0.771
0.07(4)
1.046
Mixed
0.0360
0.0948
0.631
ꢀ0.844
0.023(17)
1.017
wR
Drmax (e A
ꢀ1
ꢂ
)
ꢀ1
ꢂ
Drmin (e A
)
Flack parameter
GOF
a
R1 ¼ SkFo| ꢀ |Fck/S|Fo|; wR ¼ [S[w(Fo2 ꢀ Fc2)2]/S [w(Fo2)2]]1/2. w ¼ 1/
[s2(Fo2) + (m P)2 + n P] where P ¼ (Fo2 + 2Fc2)/3; m ¼ 0.0291 (2), and
0.0631 (3); n ¼ 29.5342 (2), and 0.5090 (3).
15942 | J. Mater. Chem., 2011, 21, 15940–15949
This journal is ª The Royal Society of Chemistry 2011

View Article Online
the surface of the sample with a 50ꢁ NIR objective (numerical
aperture—NA—equal to 0.42), and using an optimal pinhole of
100 mm. The depth of focus was estimated to be ca. 3 mm. The
energy per pulse was always less than 100 nJ and adjusted by
a power unit composed of a rotating half-wave plate followed by
a Glan-Taylor prism. The incident power pulse was monitored
by a fast InGaAs photodiode for quantitative power calibration
purpose. A transparent polarized glass (formula: (1 ꢀ x)NaPO3
+ xNb2O5, x z 0.47) where a NLO layer of ca 3 mm has been
implemented at the anode during a poling process,21 and in which
the d tensor components were previously determined (d33 ¼ 2.1
pm Vꢀ1 ¼ 5.0 ꢁ 10ꢀ9 esu, d31 ¼ d33/3) was used as a reference for
the calibration of the d value. Owing to the size of the crystal
(ꢂ75 ꢁ 50 ꢁ 2 mm3), no preferred orientation was possible except
that, using a rotating stage, the crystal plate was oriented in-
plane so that the reflected m-SHG signal was maximum. In
addition, it was necessary to use a calibrated density filter to
attenuate the excitation line (OD ¼ 3 at 1064 nm) for 3 because of
the huge efficiency of its response with regards to the reference
sample. Finally, for both the reference and the crystal, polarized
SHG intensities have been recorded as a function of the incident
power so as to obtain a quadratic dependence of the harmonic
NLO response. Assuming that the probed thickness is defined by
the depth of focus (DOF) of the focused beam, the SHG intensity
(I2u) is related to the NLO susceptibility deff as follows:
of the pyridine groups of the corresponding (DEAS)H+, respec-
ꢂ
tively. The N7/N2 distance is 2.85(3) A (x + 1/2, y ꢀ 1/2, z + 1/
ꢂ
2) while the N4/N5 distance is 2.74(3) A (x ꢀ 1/2, ꢀy + 1/2,
z ꢀ 1/2). Furthermore, C/C intermolecular lateral contacts
ꢂ
shorter than the sum of the van der Waals radii (3.7 A) indicate
the presence of strong p–p interactions. The Au atoms are
organized in pairs suggesting the occurrence of aurophilic
ꢂ
interactions, however, the Au/Au distance is 3.6326(12) A
(x + 1/2, y ꢀ 1/2, z), and so higher than the sum of the van der
ꢂ
Waals distance (3.4 A). These intermolecular interactions define
layers of 2 units which stack along c–a direction. The planarity of
the two DEAS ligands indicates that the possibility for amine to
pyridinium charge transfers, and hence for large molecular NLO
response, will be nearly optimized in the cations. Nevertheless,
the relative orientation of the two crystallographically indepen-
dent units rotated each other by 175.8ꢃ, which shows that most of
the contribution of one of the component is cancelled by that of
the other one in the solid state. Therefore, it may be anticipated
that crystals of 2 should exhibit an almost vanishing SHG
efficiency.
Compound 3 crystallizes in the non-centrosymmetric triclinic
P1 space group. The molecular structure is shown in Fig. 2, with
the atomic numbering scheme. The p-conjugated skeleton
defined by the fifteen N(1–2)C(1–13) atoms is roughly planar
ꢂ
with a largest deviation of 0.152 A from the mean plane observed
at C(2), which suggests an efficient charge delocalization along
the stilbazolium entity, and again a large molecular NLO
response. The crystal packing is shown in Fig. 3. As in the
precedent case, strong hydrogen bonding is established between
the N1 atom of the pyridine moiety and the N3 atom of the quasi
(I2u) f DOF2 ꢁ (deff)2 ꢁ (Iu)2
(3)
Results and discussions
linear [Au(CN)2]ꢀ anion [N1/N3 distance ¼ 2.74(2) A (x + 2,
ꢂ
Description of structures
y ꢀ 1, z); C–Au–C angle ¼ 178.2(11)ꢃ]. What is immediately
striking in the overall structure of this material is the fact that,
due to the presence of a single molecule in the asymmetric unit in
conjunction with the lack of symmetry of the P1 space group, all
the chromophores are necessarily perfectly aligned in the solid.
{(MPS)H [Au(CN)2]} should therefore be one of the very rare
examples of nearly perfect 1-dimensional NLO crystal.
Compound 2 crystallizes in the non-centrosymmetric monoclinic
Cc space group. The asymmetric unit consists of two non-
equivalent {(DEAS)H [Au(CN)2]} entities, the molecular struc-
tures of which are shown in Fig. 1, with the atomic numbering
scheme. The two p-conjugated skeletons of the cations defined
by the fifteen N(4–5)C(4–16) and N(6–7)C(21–33) atoms are
Although there is no significant interaction between (MPS)H+
entities in the crystal, the question of the origin of the overall and
unique 1-dimensional chromophores alignment is naturally
addressed in the crystals of 3. In a previous paper, Nicoud et al.
have discussed the solid state engineering in a series of 33 crystals
built up from the 40-donor substituted stilbazolium fragment
ꢂ
almost perfectly planar with largest deviations of 0.045 A and
ꢂ
0.080 A from mean planes observed at C(13) and C(33),
respectively. The [Au(CN)2]ꢀ anions have quasi-linear structure
with C–Au–C angles in the interval 174.7–177.7ꢃ. The
[Au1(CN)2]ꢀ and [Au2(CN)2]ꢀ groups interact via strong
hydrogen bonds through N2 and N4 with the N7 and N5 atoms
Fig. 1 ORTEP representation of the asymmetric unit of {(DEAS)H [Au(CN)2]}. Thermal ellipsoids are represented at 50% probability. Red broken
lines represent strong hydrogen bonds.
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 15940–15949 | 15943

View Article Online
Fig. 2 ORTEP representation of the asymmetric unit of {(MPS)H [Au(CN)2]}. Thermal ellipsoids are represented at 50% probability. Hydrogen bond
interaction has been represented by a red broken line.
incorporated in zwitterionic building units.22 Their investigation
points out how important intermolecular interactions (e.g.
H-bond networks) are in the crystalline cohesion, and hence in
the bulk NLO properties. In the case of 3, the role of the Au
(CN)2ꢀ anion is anticipated to be a key parameter to account for
the stabilization of the molecular alignment. To verify this idea,
the main short contacts have been summarized in Table 2, and
are shown in Fig. 4. Clearly two types of interactions are evi-
spectra are dominated by a single and intense band centred in the
450–500 nm domain (lmax ¼ 477 nm, 3 ¼ 41 700 L molꢀ1 cmꢀ1 for
3), with a tendency for red shift observed on changing the amine
from dimethylamino, to pyrrolidinyl and finally diethylamino.
This effect should slightly reinforce the b value of the stilbazo-
lium based derivatives, with respect to the DAST reference
materials (vide infra). Interestingly, these tendencies are observed
at both experimental and theoretical level, making the use of the
ZINDO computed electronic transitions relevant in the
description of the NLO response of these salts in the next section.
ꢀ
denced: (i) within a layer, where Au(CN)2 plays a major role;
and (ii) through a set of interlayer interactions where both (MPS)
H+/[Au(CN)2]ꢀ and (MPS)H+/(MPS)H+ contacts take place.
Finally, one half of the contacts reported in Table 2 involves the
anion, which demonstrates its important role to account for the
cohesion of the crystal of 3.
NLO studies
The SHG efficiencies of the three [Au(CN)2]ꢀ salts, recorded on
calibrated powdered samples, are shown in Table 4. While 1 is
SHG silent, probably in relation with a centrosymmetric crystal
environment, 2 exhibits a residual efficiency, which is consistent
with the pseudo-centrosymmetric packing of the (DEAS)H+
cations in the acentric Cc space group, as previously discussed.
Compound 3, in which the chromophores are all perfectly
aligned exhibits a very large SHG efficiency up to 190 times that
of urea. In order to analyse these results and the full capabilities
of this 1-dimensional material, the present section is divided in
Optical properties
The UV-visible electronic spectra for 1, 2, and 3 were recorded in
acetonitrile. Experimental and ZINDO computed absorption
maxima are listed in Table 3, while the experimental spectrum of
the relevant salt 3 is presented in Fig. 5. The geometries used for
the computations were both those obtained by DFT optimiza-
tions and those available from X-ray data. In any case, the
Fig. 3 Crystal packing of {(MPS)H [Au(CN)2]}. Blue arrows have been added to the terminal atoms of the [Au(CN)2]ꢀ group to highlight the relative
orientation of the molecules.
15944 | J. Mater. Chem., 2011, 21, 15940–15949
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Table 2 Short contact in the crystal structure of 3
Intralayer contacts
into two components, so-called two-level (b2level) and three-level
(b3level) terms.11 In the present case, it clearly appears that b2level
dominates the nonlinearity, b3level being reduced and directed in
the opposite direction to that of b (Table 5). Therefore, it can
readily be assumed that b is qualitatively described by the two-
level term, which relates the NLO response to the optical tran-
sitions according to the following relation:23
N1/N3 (hydrogen bond)
0
ꢂ
ꢂ
C14/C20 ¼ 3.68(2) A (x, y, z + 1)
0
C18/C20 ¼ 3.70(2) A (x, y, z + 1)
Interlayer contacts
-
2
E i4
X
3e hfiDmi
2mEi3
0
ꢂ
ꢂ
ꢂ
ꢂ
ꢀ
In this equation, fi, Dmi, and Ei are the oscillator strength, the
ꢁꢀ
ꢁ
C1/C12 ¼ 3.56(2) A (x + 1, y ꢀ 1, z)
b2Level
¼
ꢁ
(4)
2
2
-
-
0
E 2i ꢀ ð2huÞ
E 2i ꢀ ðhuÞ
C2/C13 ¼ 3.62(2) A (x + 1, y, z)
i
0
C5/C20 ¼ 3.68(3) A (x + 1, y ꢀ 1, z)
0
C5/C21 ¼ 3.50(3) A (x + 1, y, z)
0
ꢂ
C6/C20 ¼ 3.59(2) A (x + 1, y, z)
0
difference between ground and excited state dipole moment and
-
ꢂ
C9/C17 ¼ 3.695(13) A (x + 1, y, z)
C10/C19 ¼ 3.37(2) A (x, y ꢀ 1, z)
the energy of the ith transition, respectively (hu being the energy
0
0
ꢂ
ꢂ
C11/C19 ¼ 3.65(2) A (x, y ꢀ 1, z)
of the incident laser beam). The main transitions with significant
contributions to b2level are described in Table 6. If one considers
that the data for (DEAS)H+ are somewhat close to the average
calculated for the two crystallographically independent mole-
cules present in the crystal of 2, the overall behavior is very
similar for any chromophores, with a single HOMO/LUMO-
based electronic transition, centered at 470–490 nm (intensity f ¼
1.20–1.30) contributing around 95% to b2level, which is usually
the case in NLO chromophores having a strong ‘‘push-pull’’
character. This transition possesses a strong push-pull character
with a large Dm value between 17 and 18 Debye. The electronic
charge transfers drawn in Fig. 6 for (MPS)H+ and DAMS+
clearly indicate that the origin of the NLO response is qualita-
tively the same in these systems, and are only slightly modulated
by the different nature of the substituents.
two parts to discuss the properties (i) at the molecular level, and
(ii) at the solid state level.
(i) Molecular properties. The solid state NLO response of
a molecular material (eqn (2)) is ultimately related to the
hyperpolarizability (b) of discrete entities present in the crystal
and exhibiting intramolecular charge transfers (eqn (1)). It is
therefore important to use the solid state geometries for the
computations. The calculated b values of (DEAS)H+, and (MPS)
H+ are gathered in Table 5, and compared to that of DAMS+, the
related chromophore of DAST, reference system regarded as one
of the most efficient NLO materials reported to date.4 No crystal
structure being available for 1, no solid state calculation was
performed on (DMAS)H+. The values suggest that the present
stilbazolium ions have a hyperpolarizability enhanced by about
25–30% versus that of the well known DAMS+ cation, with static
hyperpolarizabilities (b0 at zero laser frequency) equal to 144.0 ꢁ
10ꢀ30 cm5 esuꢀ1 (average value of the btNotal hyperpolarizabilities
reported in Table 5), 147.0 ꢁ 10ꢀ30 cm5 esuꢀ1, and 117.8 ꢁ 10ꢀ30
cm5 esuꢀ1, for (DEAS)H+, (MPS)H+, and DAMS+, respectively.
To analyse these differences more precisely, it has to be
reminded that, within the sum over state (SOS) approach, b is
related to all excited states of a molecule and can be partitioned
Interestingly, the different values computed between the two
chromophores present in 2 are of the same order of magnitude as
those observed between DAMS+ and (MPS)H+. This encouraged
us to check the intrinsic molecular properties by a study carried
out on DFT optimized molecular structures as well. The data are
gathered in Table 7, where the spectroscopic features are pre-
sented in relation to the concept of bond-length alternation
(BLA).24 A reduced BLA indicates a better charge delocalization
through the p-conjugated bridge and therefore a possibility for
enhanced NLO response. The data show a reduced BLA from
+
+
ꢂ
0.071 to 0.066 A on going from DAMS (DAST) to (MPS)H
Fig. 4 Short contacts in {(MPS)H [Au(CN)2]}.
This journal is ª The Royal Society of Chemistry 2011 J. Mater. Chem., 2011, 21, 15940–15949 | 15945

View Article Online
Table 3 Experimental (in MeCN) and ZINDO computed absorption
maxima (lmax in nm) of 1, 2, and 3. The data for DAST are provided for
comparison
Table 5 ZINDO calculated hyperpolarizabilities (b in 10ꢀ30 cm5 esuꢀ1) of
(MPS)H+ and (DEAS)H+ at l ¼ N (b0) and 1.907 mm. The values for
DAMS+ are provided as references
ZINDO computed values
Angle between
b2level and b3level
a
l/mm
btotal
b2level
b3level
Experimental
values
In MeCN
(DFT geometry)
In solid state
(X-ray geometry)
(DEAS)H+
mol 1b
N
1.907
N
1.907
N
1.907
N
1.907
152.4
241.2
135.6
206.0
147.0
227.5
117.8
174.8
248.3
361.8
209.8
297.4
231.8
332.6
197.0
271.2
96.1
120.6
74.7
92.0
85.1
105.5
79.4
96.6
177.8ꢃ
—
1
2
470
483
477
471
473
478
477
470
n/a
497a
494
471
mol 2c
174.8ꢃ
—
3
DAST
(MPS)H+
DAMS+
176.0ꢃ
—
176.6ꢃ
—
a
Averaged value from the two independent (DEAS)H+ present in the
crystal cell.
a
b
btotal ¼ b2level + b3level
.
Molecule 1 corresponds to that having a C]C
c
central ethenyl bond length of 1.34 A. Molecule 2 corresponds to that
ꢂ
ꢂ
having a C]C central ethenyl bond length of 1.35 A.
reported.4 In solid state, the hyperpolarizability tensor (compo-
nent bijk in the molecular framework) is related to the corre-
sponding crystalline first-order nonlinearity tensor c(2)
(component dIJK in the crystalline framework).25 Assuming a 1-
dimensional character of the molecular tensor (which is totally
justified in the present case), b has only one large coefficient
along the charge transfer axis x of the molecule (namely bxxx).
The relations between bxxx and dIJK have been established for any
space groups. For instance, in the case of the monoclinic Cc
group (e.g. DAST), these expressions lead to the following
relations:26,27
Fig. 5 Optical spectra of (MPS)H+ [Au(CN)2]ꢀ, in acetonitrile.
(3), which is indeed correlated to a better charge transfer, a red
shift, and finally to enhanced NLO capabilities. Although the
differences are modest, all data available lead to a set of struc-
tural (BLA), experimental (UV-visible spectra) and computed
(solid state and DFT geometries) features which all suggest that
the new 1-dimensional compound 3 exhibits better NLO capa-
bilities than the promising DAST, at the molecular level.
dZXX ¼ Nbxxxcosqsin2q
dZZZ ¼ Nbxxxcos3q
(5)
(6)
where the Lorentz local-field factors have been assumed to be
equal to 1. All other components of the d tensor are negligible (q
is defined as the angle between b and the OY glide mirror of the
crystal, N ¼ Z/V is the number of molecule per unit volume). For
(ii) Solid state property. The fact that a chromophore crys-
tallizes in a non-centrosymmetric space group does not guarantee
that the NLO response is optimized. In the case of 2, the
examination of the crystal structures reveals that the asymmetric
unit cell is built up from two molecules (mol1 and mol2 in Table
5), which are embedded in a nearly centro-symmetric environ-
ment. Indeed, the angle between the two calculated b is equal to
176.5ꢃ, thus cancelling most of the NLO effect with a modest
resulting efficiency reduced to 0.5 time that of urea.
3
ꢃ
4
ꢂ
DAST, Z ¼ 4 and V ¼ 2098.2 A , and q ¼ 20 , which, at the
1.907 mm wavelength, leads to:
dZXX ¼ 3.66 ꢁ 10ꢀ8 cm2 esuꢀ1 (15.33 pm Vꢀ1
)
(7)
(8)
dZZZ ¼ 27.65 ꢁ 10ꢀ8 cm2 esuꢀ1 (115.8 pm Vꢀ1
)
In the case of 3 (P1 space group) and in the absence of any
crystal symmetry, all the molecules are strictly aligned in the
solid. Therefore, the d tensor is equal to its b molecular coun-
The situation encountered in compound 3 is much more
appealing and can be compared to that of DAST, for which
a record SHG efficiency of 1000 times that of urea has been
3
ꢂ
terpart, scaled by N ¼ 1/V (V ¼ 502.9 A ). This implies:
dZZZ ¼ N bxxx ¼ 47.96 ꢁ 10ꢀ8 cm2 esuꢀ1 (200.8 pm Vꢀ1
)
(9)
Table 4 Powdered SHG efficiencies at 1.907 mm for 1, 2, and 3
Since the SHG intensity (I2u) is proportional to the square of
the d tensor, crystals of 3 are roughly expected to exhibit
a potential SHG efficiency 3 times (200.82/115.82) higher than
that of DAST.
SHG efficienciesa
Space groups
Urea
1b
0
0.5
190
P421m
n/a
Cc
1
2
3
At first, it seems therefore surprising that the Kurtz-Perry
powder test indicates for 3 a SHG efficiency 5 times lower than
that of DAST. However, two important parameters must be
taken into account in order to express I2u in the powder test: (i)
P1
a
I2u/Iu. b Used as a reference.
15946 | J. Mater. Chem., 2011, 21, 15940–15949
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Table 6 Energy (lmax in nm), oscillator strength (f), dipole moment change (Dm in D) between ground and excited states, contribution (%) to b2level, and
composition of the dominant excited state involved in the NLO response of (MPS)H+ and (DEAS)H+. Data are provided for DAMS+ as references
Compound
Transition
lmax
f
Dm
state %a
90.5
Composition of CI expansion
(MPS)H+
(DEAS)H+
mol 1
mol 2
DAMS+
1 / 2
494
1.30
17.0
0.948cHOMO / LUMO
1 / 2
1 / 2
1 / 2
507
487
471
1.02
1.46
1.20
20.9
14.3
17.9
88.1
89.6
87.5
0.950cHOMO / LUMO
0.940cHOMO / LUMO
0.939cHOMO / LUMO
a
Contribution of the transition to b2level calculated at l ¼ N.
Like any highly efficient NLO material, DAST is phase
matchable, which means that the SHG efficiency is roughly
proportional to the particle size at least in the 0–100 mm
domain.18a The giant efficiency of 1000 times that of urea
reported by Marder was measured on non-calibrated samples,4
which suggests that tightly grinded (z 10 mm) samples would
likely lead to much reduced efficiencies.
By contrast, 3 is not phase matchable (because of its strong
1-dimensional character), which implies an inverse relation
between efficiency and particle size, when the sizes are larger than
the coherence length ($10–20 mm).18a We may therefore assume
that experiments performed on much smaller (z 10 mm) crystals
would provide efficiencies larger than the actual 190ꢁ urea value,
recorded on a particle range of 50–80 mm. This point leads to the
conclusion that 3 might be one of the most efficient NLO
materials at the micron scale. To verify this possibility, the value
of the dZZZ tensor component has been investigated, and deter-
mined by m-SHG.
Fig. 6 ZINDO-computed charge transfers in (MPS)H+ (top) compared
to those of DAMS+ (bottom). The surface of the circles are proportional
to the electron densities (white and black contributions indicate increases
and decreases of electron densities, respectively).
ꢂ
Table 7 ZINDO computed data in relation with BLA (in A) calculated
on molecular structures optimized by DFT
Fig. 7 shows the quadratic dependence of the polarized SHG
signals (I// and It) obtained upon exciting a thin microcrystal of
3 (thickness around 2 mm) in which, to a large extent, the
deleterious effect of light absorption may be neglected. The
corresponding quadratic laws obeyed by I// and It can be
easily deduced from a least-squares fitting procedure. We find I//
Bond-lengths
(reference sample) ¼ 45.5Iu2, I//(3) ¼ 152.0 ꢁ 106Iu2 and It(3) ¼
2
.
4.0 ꢁ 106Iu
a þ c
BLA ¼
ꢀ b
Chromophores lmax (nm) f
a
b
c
2
A straight comparison between the polarized intensities I// of
the reference and investigated sample (using eqn (3)) leads to deff
(DMAS)H+
(DEAS)H+
(MPS)H+
DAMS+
473
478
478
470
1.24 1.4352 1.3677 1.4368 0.068
1.35 1.4342 1.3684 1.4384 0.066
1.25 1.4346 1.3682 1.4360 0.067
1.26 1.4368 1.3664 1.4383 0.071
the expression of the angular averaged d2 tensor, and (ii) the
efficiency–particle size relationship. Indeed, in powdered sample,
I2u becomes proportional to the angular averaged d2 value, the
expression of which has previously been treated:18a,28
X
X
ꢂ
2
d2〉 ¼ ð19=105Þ
ðdiiiÞ þ ð13=105Þ
ðdiiiÞðdijjÞ þ ð44=105Þ
i
isj
X
X
2
2
ꢁ
ðdiijÞ þ ð13=105Þ
ðdiijÞðdjkkÞ þ ð5=7ÞðdijkÞ
isj
ijk
(10)
This expression clearly shows that a strictly 1-dimensional NLO
crystal with a single non-zero d tensor component (e.g. 3) may
have a more reduced powder efficiency than a multi-components
material (e.g. DAST). More importantly, the efficiency is
strongly dependent on the particle size.
Fig. 7 Quadratic dependence of the polarized (I//) SHG signal of the
reference poled glass sample (squares) and of the crystal of 3 (I//: circles;
It: full circles). Note the difference of intensity scales between the two
materials.
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 15940–15949 | 15947

View Article Online
z 3660 pm Vꢀ1 (ꢂ8.7 ꢁ 10ꢀ6 esu) which is actually a huge
dynamical (at 1064 nm) value as expected.
thickness of only few microns.31 This could provide an easy
access to these intriguing materials, in which the absorbance, and
related optical losses and chromophores damage, should be very
limited even near resonance.
To estimate the quadratic susceptibility dzzz ¼ Nb*xxx of 3,
where b*xxx is the molecular component bxxx corrected from local
field effects and N is the number density of molecules in the
crystal. A more accurate interpretation requires the assumption
that, due to its strong 1-dimensional character with a dominating
bxxx component, the dipolar molecule lies on a vertical mirror
plane (pseudo-local symmetry) (Fig. 8). In this case, the x axis
corresponds to the direction of light propagation and the inci-
dent polarization is along the z direction. Hence, I// is associated
with the IZZZ term and It to the IYZZ term in the (x, y, z)
laboratory frame. The above mentioned assumption is supported
by the fact that neither I// nor It signals could be detected when
rotating at 90ꢃ the plate on top of which the sample was oriented
in-plane to maximize its reflected m-SHG signal. From this
information, we can indeed deduce that the SHG tensor is now
described by four components, namely dZZZ, dXZZ, dZXX and
dXXX. Finally, we find that the dynamical value measured at
1.064 mm dzzz ¼ Nbx*xx value ranges between ꢂ4500 pm Vꢀ1
(ꢂ10.8 ꢁ 10ꢀ6 esu) and 12400 pm Vꢀ1 (ꢂ29.6 ꢁ 10ꢀ6 esu). The
static value extrapolated from the 2-level model (eqn (4)) is
estimated between ꢂ705 pm Vꢀ1 (ꢂ1.7 ꢁ 10ꢀ6 esu) and 1940 pm
Vꢀ1 (ꢂ4.6 ꢁ 10ꢀ6 esu). Details about the analytical calculations
are provided in supplementary information.†
Conclusion
Three salts derived from the well known DAST NLO materials
ꢀ
have been characterized. Among them, (MPS)H+ Au(CN)2
possesses the double intriguing feature of a very large molecular
hyperpolarizability and a perfect 1-dimensional alignment in
crystal state. This leads to a giant dzzz tensor component
(between 10.8 ꢁ 10ꢀ6 esu and 29.6 ꢁ 10ꢀ6 esu at 1.064 mm),
a unique feature which would deserve more attention in micron-
thick devices operating near resonance, where no absorption and
velocity dispersion take place even with ultra-short (sub-pico)
pulses of light.
Acknowledgements
ꢀ ꢀ
We thank CALMIP (calcul intensif en Midi-Pyrenees, Toulouse,
France) for computing facilities. V. R. thanks F. Adamietz for
ꢀ
technical assistance and the Region Aquitaine for financial
support in optical, laser, and computer equipment. This work
has been supported by the University of Bordeaux 1 and the
CNRS through the excellence chair of S.B. Financial support is
acknowledged from the Spanish Ministerio de Ciencia e
Critical evaluation of 1-dimensional devices with large
hyperpolarizabilities
ꢀ
Innovacion (MICINN) (CTQ2010-18414-FEDER).
Recent years have witnessed the development of lasers with ultra-
short (sub-pico) pulses of light (e.g. the Ti : sapphire laser) in
which the radiation cannot be regarded as strictly mono-
chromatic. These advanced technologies can reinforce the
potential of nonlinear optics in the future photonics,29 which
should greatly benefit from the use of micron-thick materials, to
avoid velocity dispersion during the traversal of the devices, the
overall NLO efficiency being ensured by the possibility of giant
values of the hyperpolarizabilities. The typical dzzz range of
standard NLO crystals present on the market place, such as
LiNbO3 and KTiOPO4 (KTP), is around 20 pm Vꢀ1 (ꢂ4.8 ꢁ 10ꢀ8
esu) at 1.064 mm,30 about three orders of magnitude below that of
the present 3 material. This leads to the conclusion that a 1 mm-
thick lens could be replaced by a micron-thick crystal of 3, in
which any velocity dispersion problem is avoided. The technique
of microtomy, which is traditionally employed in biology, allows
to cut single crystals embedded in polymer matrices at the
References
1 (a) Nonlinear Optical Properties of Organic Molecules and Crystals,
ed. D. S. Chemla, J. Zyss, Academic Press, New York, 1987; (b)
P. N. Prasad, D. J. Williams, Introduction to Nonlinear Optical
Effects in Molecules and Polymers, John Wiley & Sons, New York,
1991; (c) Molecular Nonlinear Optics: Materials, Physics, and
Devices, ed. J. Zyss, Academic Press, New York, 1994; (d)
Nonlinear Optics of Organic Molecules and Polymers, ed. H. S.
Nalwa, S. Miyama, CRC Press, Boca Raton, 1997.
2 (a) S. R. Marder, Chem. Commun., 2006, 131; (b) M. G. Kuzyk, J.
Mater. Chem., 2009, 19, 7444.
3 D. J. William, Angew. Chem., Int. Ed. Engl., 1984, 23, 690.
4 S. R. Marder, J. W. Perry and W. P. Schaefer, Science, 1989, 245, 626.
€
5 For recent examples: (a) M. Jazbinsek, L. Mutter and P. Gunter,
IEEE J. Sel. Top. Quantum Electron., 2008, 14, 1298; (b) Z. Yang,
L. Mutter, M. Stillhart, B. Ruiz, S. Aravazhi, M. Jazbinsek,
€
A. Schneider, V. Gramlich and P. Gunter, Adv. Funct. Mater.,
2007, 17, 2018; (c) J. McEntee, Chemistry World, March 2007, p. 52.
6 For recent examples: (a) B. J. Coe, J. Fielden, S. P. Foxon,
J. A. Harris, M. Helliwell, S. Bruce, B. S. Brunschwig,
I. Asselberghs, K. Clays, J. Garin and J. Orduna, J. Am. Chem.
Soc., 2010, 132, 10498; (b) X. J. Feng, P. L. Wu, F. Bolze,
H. W. C. Leung, K. F. Li, N. K. Mak, D. W. J. Kwong,
J. F. Nicoud, K. W. Cheah and M. S. Wong, Org. Lett., 2010, 12,
21194; (c) F. Hao, X. Zhang, Y. Tian, H. Zhou, L. Li, J. Wu,
S. Zhang, J. Yang, B. Jin, X. Tao, G. Zhou and M. Jiang, J. Mater.
Chem., 2009, 19, 9163; (d) F. Dumur, C. R. Mayer, K. Hoang-Thi,
I. Ledoux-Rak, F. Miomandre, G. Clavier, E. Dumas, R. Mallet-
ꢀ
Renault, M. Frigoli, J. Zyss and F. Secheresse, Inorg. Chem., 2009,
48, 8120; (e) K. Akiyama, S. Okada, Y. Goto and H. Hachiro
Nakanishi, J. Cryst. Growth, 2009, 311, 953.
7 G. Agusti, A. B. Gaspar, M. C. Munoz, P. G. Lacroix and J. A. Real,
Aust. J. Chem., 2009, 62, 1155.
8 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008,
64, 112.
9 J. A. Pople, D. L. Beveridge and P. A. Dobosh, J. Chem. Phys., 1967,
47, 2026.
Fig. 8 Image of the crystal of 3 in the frame of the laboratory (left). The
1-dimensional molecule is located in the mirror plane (XZ) with an
unknown azimuthal angle q (right).
15948 | J. Mater. Chem., 2011, 21, 15940–15949
This journal is ª The Royal Society of Chemistry 2011

View Article Online
10 (a) M. Zerner, G. Loew, R. Kirchner and U. Mueller-Westerhoff, J.
Am. Chem. Soc., 1980, 102, 589; (b) W. P. Anderson, D. Edwards
and M. C. Zerner, Inorg. Chem., 1986, 25, 2728.
11 J. F. Ward, Rev. Mod. Phys., 1965, 37, 1.
12 D. R. Kanis, M. A. Ratner and T. J. Marks, Chem. Rev., 1994, 94,
195.
M. Lahaye, E. Fargin, M. Bellec, A. Royon, L. Canioni,
M. Dussauze, F. Adamietz and V. Rodriguez, Appl. Phys. Lett.,
2010, 96, 091908; (c) M. Dussauze, V. Rodriguez, A. Lipovskii,
M. Petrov, C. Smith, K. Richardson, T. Cardinal, E. Fargin and
E. I. Kamitsos, J. Phys. Chem. C, 2010, 114, 12754.
20 T. Verbiest, K. Clays, V. Rodriguez, Second-order nonlinear optical
characterizations techniques: An Introduction, Chapter 5, CRC
Press, New York, 2010.
13 Zindo, release 96.0, Molecular Simulations Inc., Cambridge, UK,
1996.
14 P. G. Lacroix, J. C. Daran and K. Nakatani, Chem. Mater., 1998, 10,
1109.
21 M. Dussauze, E. Fargin, M. Lahaye, V. Rodriguez and F. Adamietz,
Opt. Express, 2005, 13, 4064.
15 S. R. Marder, J. W. Perry and C. P. Yakymyshyn, Chem. Mater.,
1994, 6, 1137.
22 Ch. Serbutoviez, J.-F. Nicoud, J. Fischer, I. Ledoux and J. Zyss,
Chem. Mater., 1994, 6, 1358.
16 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci,
G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada,
M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida,
T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark,
J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov,
R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell,
J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega,
N. J. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken,
C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas,
J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox,
GAUSSIAN 09 (Revision A.1), Gaussian, Inc., Wallingford, CT, 2009.
17 (a) A. D. Becke, J. Chem. Phys., 1993, 98, 564; (b) J. P. Perdew and
Y. Wang, Phys. Rev. B: Condens. Matter, 1992, 45, 13244.
18 (a) S. K. Kurtz and T. T. Perry, J. Appl. Phys., 1968, 39, 3798; (b)
J. P. Dougherty and S. K. Kurtz, J. Appl. Crystallogr., 1976, 9, 145.
19 (a) V. Rodriguez, D. Talaga, F. Adamietz, J. L. Bruneel and
M. Couzi, Chem. Phys. Lett., 2006, 431, 190; (b) A. Delestre,
23 (a) J. L. Oudar, J. Chem. Phys., 1977, 67, 446; (b) J. L. Oudar and
J. Chemla, J. Chem. Phys., 1977, 66, 2664.
ꢀ
24 (a) G. Bourhill, J.-L. Bredas, L. T. Cheng, S. R. Marder, F. Meyers,
J. W. Perry and B. G. Tiemann, J. Am. Chem. Soc., 1994, 116, 2619;
(b) S. R. Marder, D. N. Beratan and L. T. Cheng, Science, 1991, 252,
103.
25 J. Zyss and J. L. Oudar, Phys. Rev. A, 1982, 26, 2028.
26 J. Zyss, J. F. Nicoud and M. Coquillay, J. Chem. Phys., 1984, 81,
4160.
27 S. R. Marder, D. N. Beratan and L. T. Cheng, Science, 1991, 252, 103.
28 S. J. Cyvin, J. E. Rauch and J. C. Decus, J. Chem. Phys., 1965, 43,
4083.
29 (a) M. Pawlicki, H. A. Collins, R. G. Denning and H. L. Anderson,
Angew. Chem., Int. Ed., 2009, 48, 3244; (b) J. J. Bloembergen, J.
Chem. Educ., 1998, 75, 555.
30 (a) Handbook of Laser Science Technology, Vol. 111, Part 1, ed. M. J.
Weber, CRC Press, Boca Raton, FL, 1986; (b) V. G. Dmitriev,
G. G. Gurzadyan, D. Nikogosyan, Handbook of Nonlinear Optical
Crystals, Springer-Verlag, Berlin, 1991.
31 For reviews on microtomy and its application in ultra-thin crystalline
materials, see: (a) H. K. Plummer, Microscopy Microanal., 1997, 3,
239; (b) G. McMahon and T. Malis, Microsc. Res. Tech., 1995, 31,
267.
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 15940–15949 | 15949