Dynamic Article LinksC<Journal ofMaterials Chemistry

Cite this: J. Mater. Chem., 2012, 22, 11268

www.rsc.org/materials PAPER

Dow

nloa

ded

by L

udw

ig M

axim

ilian

s U

nive

rsita

et M

uenc

hen

on 2

5 Fe

brua

ry 2

013

Publ

ishe

d on

13

Apr

il 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2J

M30

861D

View Article Online / Journal Homepage / Table of Contents for this issue

Triphenyl moieties as building blocks for obtaining molecular glasses withnonlinear optical activity†

Kaspars Traskovskis,*a Igors Mihailovs,ab Andrejs Tokmakovs,b Andrejs Jurgis,b Valdis Kokarsa

and Martins Rutkisb

Received 12th February 2012, Accepted 13th April 2012

DOI: 10.1039/c2jm30861d

The incorporation of trityl and triphenylsilyl groups into low molecular weight molecules allows the

formation of stable molecular glasses. A series of materials based on the N-phenyldiethanolamine core

was synthesized bearing different azobenzenes and benzylydene-1,3-indandione as active

chromophores. Molecular hyperpolarizability of the synthesized compounds was calculated by

a restricted Hartree–Fock method with basis 6-31G(d,p) and measured in solutions by hyper-Rayleigh

scattering. Non-linear optical (NLO) activity of the thin glassy films was confirmed after a corona

poling procedure. Thermal sustainability of the NLO response of up to 85 �C was achieved. Quantum

chemical calculations of the compounds revealed increased steric bulk and conformational freedom of

the triphenylsilyl moiety. While the presence of the triphenylsilyl group results in more stable glasses

and increased material nonlinearity, in the case of trityl groups, measured glass transition temperatures

are higher.

Introduction

During the past decade the research towards development of

organic nonlinear optical (NLO) materials has resulted in

considerable developments, improving prospects for practical

applications relevant in telecommunications, computing and

defense.1 Standard design for such materials requires the pres-

ence of push–pull type chromophores with large molecular

hyperpolarizability (b), high optical transparency, good solu-

bility and chemical and thermal stabilities.1–5 A wide range of

modification possibilities for organic compounds has provided

numerous approaches regarding the placement of NLO active

chromophores, most commonly polymer,6–8 host polymer

matrix9,10 and dendrimeric11,12 systems.

By contrast, small molecular weight amorphous phase forming

materials (molecular glasses), a new emerging class of electro-

optical materials13–15 are less studied. While general principles

linking molecular structure and material, thermal, and amor-

phous phase stability characteristics remain unresolved,16

molecular glasses have several considerable advantages such as

relatively simple synthesis and purification, increased

aRiga Technical University, Faculty of Materials Science and AppliedChemistry, 14/24 Azenes Street, Riga, LV-1048, Latvia. E-mail: kaspars.traskovskis@rtu.lv; Tel: +371 29148070bInstitute of Solid State Physics, University of Latvia, 8 Kengaraga Street,Riga, LV-1063, Latvia

† Electronic Supplementary Information (ESI) available: detailed DSCthermograms, initial atomic coordinates for the quantum calculationsand additional sample microscope pictures. See DOI:10.1039/c2jm30861d

11268 | J. Mater. Chem., 2012, 22, 11268–11276

chromophore density and well defined structure. A widespread

strategy for obtaining molecular compounds capable of forming

stable glasses involves preventing the molecules from interacting

together in a strong and directional fashion. In particular the

presence of arene-rich starburst structural fragments is success-

fully used to obtain such materials,17,18 where the crystallization

and aggregation process is hindered by the steric demands of

conformationally rigid bulky substitutes or insufficient solid

phase packing due to the shape of the molecules. A promising

strategy for obtaining molecular glasses is the modular

approach, where a core molecule not capable of glass formation

is further functionalized with building blocks preventing the

crystallization. Substituents like N,N-diphenylhydrazone19 and

triazines20 are successfully used to obtain amorphous materials.

While minimizing the molecular interactions is one of the key

aspects for designing molecular glasses, it also plays a significant

role regarding the NLO properties of material. Conversion of

high chromophore b values into good macroscopic NLO

efficiencies is often problematic due to strong dipole–dipole

interactions causing aggregation and diminishing acentric order

of the chromophores, resulting in the decay of the electro-optic

coefficient.21 As the solution to this problem, the site isolation

principle proposed by Dalton et al. suggests designing the

molecule in a way so as to minimize dipole–dipole interactions.

Among the applied methods reported, materials involving the

introduction of sterically bulky spacer groups22,23 have shown

a noticeable increase of NLO efficiency.

In this work we present a novel modular approach in the

design of low molar weight molecular glasses for electro-optical

applications. During our research it came to our attention that

This journal is ª The Royal Society of Chemistry 2012

Dow

nloa

ded

by L

udw

ig M

axim

ilian

s U

nive

rsita

et M

uenc

hen

on 2

5 Fe

brua

ry 2

013

Publ

ishe

d on

13

Apr

il 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2J

M30

861D

View Article Online

the presence of triphenylmethyl and triphenylsilyl substituents

noticeably enhances amorphous phase formation of low molec-

ular weight molecules. While the pseudo-spherical shape and

steric demands of these groups prevent the crystallization

process, incorporating them relatively close to the photoactive

chromophore could provide effective shielding of dipole–dipole

interactions. A series of molecular glasses based on triaryl

functionalized N-phenyldiethanolamine was synthesized bearing

different azobenzenes and benzylydene-1,3-indandione as active

chromophores. The influence of the chosen triaryl group on

material properties was closely studied.

Experimental

Synthesis

Azobenzenes used in the study were obtained according to the

given route (Scheme 1) by conventional azo coupling reactions

between N-phenyldiethanolamine and aryldiazonium tetra-

fluoroborate salts followed by functionalization of free hydroxyl

groups with trityl or triphenylsilyl substituents. To better eval-

uate the chosen material design three different azochromophores

were used. In particular, the Disperse Red (DR) dye frag-

ment,24–26 a frequently used chromophore in NLO studies, con-

taining compounds A-1 and A-2 as well as its ortho-brominated

derivative A-3 were synthesized. Attempts to acquire stable

organic glass material using the trityl derivative of compound

A-3 were unsuccessful and this compound will not be discussed

here. Additionally tritylazobenzenes A-5 and A-6 were synthe-

sized to evaluate the influence of the presence of an additional

triaryl group on material characteristics.

Scheme 1 Synthesis of azobenzene containing molecular glasses: (a)

KBr, NaBO3$4H2O, AcOH, rt; (b) HCl/H2O, NaNO2, 5�C then HBF4;

(c) triphenylcarbinol, AcOH, HCl, reflux; (d) H2SO4/AcOH, NaNO2,

5 �C then HBF4; (e) N-phenyldiethanolamine, DMF, 0 �C; (f) TrCl,

pyridine, Et3N, 60 �C; (g) Ph3SiCl, pyridine, Et3N, rt.

This journal is ª The Royal Society of Chemistry 2012

Based on our previous studies showing the high NLO response

of 1,3-indandione derivatives,27 trityl and triphenylsilyl func-

tionalized benzylydene-1,3-indandiones I-1 and I-2 were

synthesized. The synthesis was realized by the condensation of

a known aldehyde 1028 with 1,3-indandione29 followed by

functionalization with triaryl substituents (Scheme 2).

General procedures and starting materials

The reagents 4-nitroaniline, sodium perborate tetrahydrate,

2-nitroaniline, tetrafluoroboric acid (50 wt% water solution),

triphenylcarbinol, N-phenyldiethanolamine, trityl chloride, tri-

phenylsilyl chloride and triethyl amine were obtained commer-

cially from Alfa Aesar and used as received. Solvents used were

reagent grade, DMF was distilled from P2O5 and pyridine from

CaH2 before use.1H NMR spectra were obtained using a Bruker

Avance 300MHz spectrometer using TMS as the inner standard.

Low-resolution mass spectra were acquired on a Waters EMD

1000 MS detector (ESI+ mode, cone voltage 30 V). The

elemental analysis was carried out with Costech Instruments

ECS 4010 CHNS-O Elemental Combustion System.

4-Nitrobenzenediazonium tetrafluoroborate (3a). Aniline 1

(10.0 g, 72 mmol) was dissolved in HCl (30 mL) and water

(15 mL). The reaction mixture was cooled to 0–5 �C using an ice-

bath and NaNO2 (6.0 g, 87 mmol) in water (20 mL) was added

dropwise. After the reaction mixture turned transparent the

remaining insoluble solid was filtered off and 10 mL of HBF4

(50 wt% water solution) was added. The white precipitate formed

was filtered and washed with water (30mL). After drying at room

temperature, the product was purified by precipitation from

nitromethane with methyl tert-butyl ether, yielding 3a (12.8 g,

76%) as white needle shaped crystals. The product was used in

subsequent reactions without characterization.

2-Bromo-4-nitrobenzenediazonium tetrafluoroborate (3b). The

compound was obtained from aniline 230 following the procedure

given for compound 3a, yielding 3b (69%) as white crystals. The

product was used in subsequent reactions without characterization.

2-Chloro-4-tritylbenzenediazonium tetrafluoroborate (6).

Aniline 531 (10.0 g, 26 mmol) was dissolved in sulfuric acid

(20 mL) and acetic acid (10 mL). The reaction mixture was

cooled to 0–5 �C using an ice-bath and NaNO2 (2.20 g, 32 mmol)

Scheme 2 Synthesis of indane-1,3-dione fragment containing molecular

glasses: (a) indane-1,3-dione, EtOH, piperidine, reflux; (b) TrCl, pyridine,

Et3N, 60 �C; (c) Ph3SiCl, pyridine, Et3N, rt.

J. Mater. Chem., 2012, 22, 11268–11276 | 11269

Dow

nloa

ded

by L

udw

ig M

axim

ilian

s U

nive

rsita

et M

uenc

hen

on 2

5 Fe

brua

ry 2

013

Publ

ishe

d on

13

Apr

il 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2J

M30

861D

View Article Online

in water (10 mL) was added dropwise. After stirring for 1 h, 8 mL

of HBF4 (50 wt% water solution) was added. The slurry

precipitate formed was decanted and dissolved in acetone (5 mL).

After pouring into methyl tert-butyl ether (100 mL), the white

precipitate was filtered and purified by additional precipitation

from nitromethane with methyl tert-butyl ether, yielding 6

(10.1 g, 82%) as white crystals. The product was used in subse-

quent reactions without characterization.

4-((4-Nitrophenyl)diazenyl)-N,N-bis(2-(trityloxy)ethyl)aniline

(A-1)

General procedure A. To a solution of N-phenyldiethanol-

amine (1.00 g, 5.0 mmol) in DMF (5 mL) at 0 �C diazonium salt

3a (1.56 g, 6.6 mmol) in DMF (2 mL) was added dropwise. After

2 h of stirring the mixture was poured into water (50 mL) and the

resulting red precipitate was filtered, washed with additional

amount of water, dried in air and crystallized from isopropanol.

Without further purification the technical product 7 (1.44 g,

4.3 mmol) was dissolved in dry pyridine (5 mL) after which

triethylamine (2.38 mL, 17.0 mmol) and tritylchloride (4.84 g,

17.0 mmol) were added. The mixture was stirred at 60 �C for 2 h

and then poured into isopropanol (50 mL). The resulting red

precipitate was filtered, washed with methanol (3 � 15 mL) and

dried in air. Purification was accomplished by column chroma-

tography over silica gel using DCM/petroleum ether (3 : 1) as

eluent. The resulting red substance was dissolved in DCM (5 mL)

and ethanol (50 mL) was added. After removal of DCM under

reduced pressure and filtration of the formed precipitate,A-1was

obtained as a red powder (3.04 g, 67% over 2 steps). 1H NMR (d,

CDCl3, 400 MHz): 3.31 (4H, t, 3J ¼ 5.50 Hz), 3.60 (4H, t, 3J ¼5.63 Hz), 6.56 (2H, d, 3J ¼ 9.27 Hz), 7.14–7.30 (30H, m), 7.73

(2H, d, 3J¼ 9.14 Hz), 7.86 (2H, d, 3J¼ 9.06 Hz), 8.26 (2H, d, 3J¼9.11 Hz). Elemental analysis (C, H, N): (calc. for C54H46N4O4: C,

82.84, H, 5.92, N, 7.16. Found: C, 79.39, H, 5.59, N, 6.83%).

4-((4-Nitrophenyl)diazenyl)-N,N-bis(2-(triphenylsilyloxy)ethyl)

aniline (A-2)

General procedure B. Compound 7 (0.52 g, 1.6 mmol), obtained

following the procedure described above, was dissolved in dry

pyridine (3mL) after which triethylamine (0.5mL, 3.69mmol) and

triphenylsilyl chloride (1.16 g, 3.69 mmol) were added. After stir-

ring at room temperature for 2 h, pyridine was removed under

reduced pressure and the resulting wet precipitate was washed with

ethanol (50 mL). Purification by column chromatography over

silica gel using DCM/petroleum ether (3 : 1) as eluent gave A-2 as

an orange powder (0.47 g, 21% over 2 steps). 1H NMR (d, CDCl3,

400 MHz): 3.53 (4H, t, 3J ¼ 5.84 Hz), 3.92 (4H, t, 3J ¼ 5.84 Hz),

6.35 (2H, d, 3J¼ 9.04 Hz), 7.30–7.60 (30H, m), 7.68–7.76 (3H, m),

8.19 (1H, dd, 3J ¼ 8.84 Hz, 4J ¼ 2.24 Hz), 8.58 (1H, d, 4J ¼ 2.24

Hz). Elemental analysis (C, H, N): (calc. for C52H46N4O4Si2: C,

73.73, H, 5.47, N, 6.61. Found: C, 73.81, H, 5.48, N, 6.59%). MS

(ESI+) m/z: 847.6 (M+, requires 847.1).

4-((2-Bromo-4-nitrophenyl)diazenyl)-N,N-bis(2-(triphenylsily-

loxy)ethyl)aniline (A-3). Prepared according to general procedure

B described above. Purification by column chromatography over

silica gel using DCM/petroleum ether (3 : 1) as eluent gaveA-3 as

a dark red powder (0.97 g, 42% over 2 steps). 1H NMR

(d, CDCl3, 400 MHz): 3.53 (4H, t, 3J ¼ 6.26 Hz), 3.92 (4H, t,

11270 | J. Mater. Chem., 2012, 22, 11268–11276

3J ¼ 6.26 Hz), 6.37 (2H, d, 3J ¼ 8.99 Hz), 7.30–7.60 (30H, m),

7.68–7.74 (3H, m), 8.20 (1H, dd, 3J ¼ 8.99, 4J ¼ 2.34 Hz) 8.58

(1H, d, 4J ¼ 2.34 Hz). Elemental analysis (C, H, N): (calc. for

C52H45BrN4O4Si2: C, 67.45, H, 4.90, N, 6.05. Found: C, 67.94,

H, 5.03, N, 5.77%). MS (ESI+) m/z: 927.2 (M+, requires 927.0).

4-((2-Chloro-4-tritylphenyl)diazenyl)-N,N-bis(2-(trityloxy)

ethyl)aniline (A-4). Prepared according to general procedure A

described above. Purification by column chromatography over

silica gel using DCM/petroleum ether (2 : 1) as eluent gaveA-4 as

a yellow powder (0.46 g, 49% over 2 steps). 1H NMR (d, CDCl3,

300 MHz): 3.26 (4H, t, 3J ¼ 5.84 Hz), 3.60 (4H, t, 3J ¼ 5.84 Hz),

6.51 (2H, d, 3J ¼ 9.23 Hz), 7.07 (1H, dd, 3J ¼ 8.66 Hz, 4J ¼ 2.07

Hz), 7.10–7.32 (45H, m), 7.36 (1H, d, 4J ¼ 2.07 Hz), 7.70 (2H,

d, 3J ¼ 9.23 Hz). Elemental analysis (C, H, N): (calc. for

C73H61N3O2: C, 83.76, H, 5.78, N, 4.01. Found: C, 83.23, H,

5.72, N, 3.93%). MS (ESI+) m/z: 1046.5 (M+, requires 1046.7).

4-((2-Chloro-4-tritylphenyl)diazenyl)-N,N-bis(2-(triphenylsily-

loxy)ethyl)aniline (A-5). Prepared according to general procedure

A described above. Purification by column chromatography over

silica gel using DCM/petroleum ether (2 : 1) as eluent gaveA-5 as

a yellow powder (0.66 g, 39% over 2 steps). 1H NMR (d, CDCl3,

300 MHz): 3.43 (4H, t, 3J ¼ 5.84 Hz), 3.83 (4H, t, 3J ¼ 5.84 Hz),

6.24 (2H, d, 3J ¼ 9.04 Hz), 7.07 (1H, dd, 3J ¼ 8.47 Hz, 4J ¼ 2.26

Hz), 7.12–7.39 (34H, m), 7.45 (1H, d, 3J ¼ 8.47 Hz), 7.49–7.53

(12H, m), 7.57 (2H, d, 3J ¼ 9.04 Hz). Elemental analysis (C, H,

N): (calc. for C71H61N3O2Si2: C, 79.04, H, 5.61, N, 3.89. Found:

C, 78.89, H, 5.58, N, 3.78%). MS (ESI+) m/z: 1078.5 (M+,

requires 1078.8).

2-(4-(Bis(2-(trityloxy)ethyl)amino)benzylidene)-1H-indene-

1,3(2H)-dione (I-1). Aldehyde 1028 (2.0 g, 9.6 mmol) was dis-

solved in ethanol (120 mL) after which indane-1,3-dione29 and 3

drops of piperidine were added. The dark orange solution was

stirred at room temperature for 3 h and refluxed for an additional

1 h. After cooling, the resulting precipitate was filtered and

recrystallized from isopropanol, yielding 11 as orange crystals

(2.1 g, 65%). Compound 11 (0.74 g, 2.2 mmol) was then dissolved

in dry pyridine (5 mL) after which triethylamine (0.85 mL,

5.0 mmol) and tritylchloride (1.4 g, 5.0 mmol) were added. The

mixture was stirred at 60 �C for 2 h and then poured into iso-

propanol (50 mL). The resulting yellow precipitate was filtered,

washed with methanol (3 � 15 mL) and dried in air. Purification

by column chromatography over silica gel using DCM/petro-

leum ether (2 : 1) as eluent gave I-1 as a yellow powder (1.60 g,

88%). 1H NMR (d, CDCl3, 300 MHz): 3.40 (4H, t, 3J ¼ 6.03 Hz),

3.72 (4H, t, 3J ¼ 6.03 Hz), 6.63 (2H, d, 3J ¼ 9.04 Hz), 7.16–7.37

(m, 30H), 7.74 (m, 2H), 7.78 (s, 1H) 7.95 (m, 2H), 8.44 (2H, d,3J ¼ 8.85 Hz). Elemental analysis (C, H, N): (calc. for

C58H47NO4: C, 84.75, H, 5.76, N, 1.70. Found: C, 83.82, H, 5.68,

N, 1.45%). MS (ESI+) m/z: 822.6 (M+, requires 822.0).

2-(4-(Bis(2-(triphenylsilyloxy)ethyl)amino)benzylidene)-1H-

indene-1,3(2H)-dione (I-2). Prepared from compound 11

according to general procedure B described above. Purification

by column chromatography over silica gel using DCM/petro-

leum ether (2 : 1) as eluent gave I-2 as yellow powder (0.84 g,

76%). 1H NMR (d, CDCl3, 300 MHz): 3.46 (4H, t, 3J ¼ 6.03 Hz),

This journal is ª The Royal Society of Chemistry 2012

Dow

nloa

ded

by L

udw

ig M

axim

ilian

s U

nive

rsita

et M

uenc

hen

on 2

5 Fe

brua

ry 2

013

Publ

ishe

d on

13

Apr

il 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2J

M30

861D

View Article Online

3.85 (4H, t, 3J ¼ 6.03 Hz), 6.24 (2H, d, 3J ¼ 9.04 Hz), 7.22–7.52

(m, 30H), 7.62–7.68 (m, 3H), 7.82–7.90 (m, 2H), 8.22 (2H, d, 3J¼9.04 Hz). Elemental analysis (C, H, N): (calc. for C56H47NO4Si2:

C, 78.75, H, 5.55, N, 1.64. Found: C, 76.51, H, 5.36, N, 1.49%).

MS (ESI+) m/z: 854.5 (M+, requires 854.2).

Measurements of optical properties of chromophores in CHCl3solution

The extinction coefficient spectra and molecular hyper-

polarizabilities (bHRS) were measured in dilute CHCl3 (Sigma-

Aldrich, analytical grade) solutions (10�6–10�5 mol l�1) of the

investigated compounds. Spectra were obtained using an Ocean

Optics HR4000CG-UV-NIR based spectroscopic system. The

hyper-Rayleigh scattering (HRS) setup was based on a 1064 nm

excitation source (NL640 by EKSPLA) delivering pulses with

10 ns duration at a repetition rate of 1000 Hz. Pulse energy was

adjusted by a half wave plate and a pair of Glan-Taylor polarizers

within a range 0–80 mJ. The incident beam was focused on the

sample solution (f¼ 3 cm) in a fused quartz 10� 2mm fluorimeter

cuvette (Starna Scientific Ltd.). Scattered light was collected by

two plano-convex lenses in 90� and 270� angle geometry with

respect to the incident beam. By means of two custom built gated

photon counting channels, double frequency light intensity was

measured in a time window synchronized with the laser pulse.

Both channels are similar, except for the final second harmonic

(SH) 532 nm interference filter with full width at half maximum

(FWHM) 3 and 10 nm. To eliminate errors caused by laser pulse

energy variation, both obtained SH intensity signals are normal-

ized to a reference signal from the potassium dihydrogen phos-

phate crystal. Such a detection system allows us to separate HRS

and two photon luminescence (TPL) signals by two procedures.

The first of them is based on temporal discrimination of these

phenomena based on the property that the HRS signal is coinci-

dent in time with excitation pulse, but TPL has some decay time.

The efficiency of this procedure drops in the case where the excited

state life time becomes comparable with excitation pulse width.

Another possibility is discrimination based on TPL and HRS

spectral bandwidth characteristics. FWHM of TPL emission is

usually a number of times wider than HRS. Therefore the TPL

and HRS contribution in the measured SH intensity depends on

the spectral bandwidth of the detection system.Without going into

detail, let us consider that our spectral deconvolution of HRS and

TPL signals is based on the assumption that bHRS obtained by the

internal reference method using data acquired from both channels

should be the same independent of detection system bandwidth.

Such a TPL deconvolution procedure applied to our HRS

measurements of DR1 (Sigma-Aldrich) yielded a value

bHRS(532)¼ 780� 10�30 esu. This value, obtained with an internal

reference (CHCl3, b ¼ 0.49 � 10�30 esu), is lower than reported

previously.32 The HRS light intensity obtained after this decon-

volution was used with the external reference (DR1) method to

obtain the bHRS values presented here.

Thin film sample preparation

An appropriate amount of glass forming compound was dis-

solved in analytical grade chloroform with a typical concentra-

tion of 100 mg ml�1. The films from solutions were spin-coated

This journal is ª The Royal Society of Chemistry 2012

with a Laurell WS-400B-6NPP/LITE spin-coater (starting speed

0 rpm, terminal speed 300 rpm, acceleration 200 rpm s�2, spin-

ning time 40 s) on indium tin oxide covered glass substrates. The

obtained good optical quality films were 0.7 to 1.3 mm thick.

Determination of thermal properties

The glass transition temperatures of the investigated organic

glasses were determined by DSC thermograms acquired using

a Mettler Toledo DSC-1/200W at a heating rate of 5 �C min�1,

initially heating the samples above their melting temperatures

(200 �C) and cooling at the rate 50 �C min�1. The temperature

corresponding to the half NLO activity level, TSHI50 was evalu-

ated from temperature scans at 10 �C min�1. Decomposition

temperatures were obtained using a Perkin Elmer STA 6000

thermal analyzer.

Corona poling procedure

To produce NLO active media, a thermo-assisted electrical field

poling procedure was applied to thin film samples on ITO via

a custom built corona triode setup. The corona discharge was

generated by a 9 kV voltage drop over a 15 mm gap between

a tungsten wire needle (diameter 25 mm) and a control grid. At

a distance of 10 mm below the grid the ITO covered slide with

spin coated glassy film was placed on a temperature controlled

heater. The grid-to-ITO layer potential was kept constant at

2.7 kV. The corona discharge started in advance of heating the

sample to TSHI50. At the point when the sample charging current

reached an approximate steady state the poling was considered to

be finished and the sample was cooled to room temperature

under applied corona triode discharge. The samples were poled

through a mask with diameter of 0.8 cm. Some minor crystalli-

zation appeared in the samples during corona poling at

temperatures exceeding Tg values.

Measurements of linear and nonlinear optical properties of thin

films

The absorption spectra of the thin films were measured with an

Ocean Optics HR4000CG-UV-NIR based spectroscopic system.

The thickness of investigated thin films was measured using

a Dektac 150 profilometer and refractive indexes were deter-

mined by a prism coupler Metricon 2010. In some cases thickness

and refractive indices were evaluated by a procedure described

elsewhere33 using interference fringe separation in the sample

reflection spectrum obtained with the same spectroscopic system.

The experimental set-up for second harmonic generation is

described elsewhere.29 To avoid an electric field-induced second

harmonic generation (EFISHG) signal from charges trapped on

the film surface, the nonlinear coefficients were usually measured

2 days after poling. The second harmonic intensities were

recorded as functions of the fundamental light incidence angle

and polarization (Maker fringe technique). For corona poled

films the CNn symmetry was assumed and the material could be

characterized by three nonzero NLO coefficients – d33, d31 and

d15. As it is usually done for poled polymer films, we assume that

d31 ¼ d15,34 according to Kleinman symmetry. The NLO coeffi-

cients were obtained by a least squares fit of the experimental

curves to the theoretical approximation. The theoretical value of

J. Mater. Chem., 2012, 22, 11268–11276 | 11271

Fig. 1 Trityl and triphenylsilyl group phenyl ring rotation potential

energy wells according to RHF 6-31G(p,d) simulation. Deflection angle is

CPh2–CPh1–C–O or CPh2–CPh1–Si–O torsion angle detune from equilib-

rium position.

Dow

nloa

ded

by L

udw

ig M

axim

ilian

s U

nive

rsita

et M

uenc

hen

on 2

5 Fe

brua

ry 2

013

Publ

ishe

d on

13

Apr

il 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2J

M30

861D

View Article Online

second harmonic intensity (SHI) was calculated using the Her-

man–Hayden17 approach, taking the absorption of the film into

account. The fitting was carried out in two steps: the value of d31was evaluated from experimental s-p polarized SHI, then the d33was calculated from the p-p SHI. An x-cut quartz crystal was

used as reference (d11 ¼ 0.3 pm V�1) to calibrate the instrument

response function.35

Quantum chemical modeling

All molecular modeling calculations were done by Gaussian 09W

software package.36 Molecular geometries of the investigated

amorphous phase forming chromophores were optimized by

restricted Hartree–Fock (RHF) ab initio calculations with basis set

6-31G(d,p). Static and frequency-dependent (for 532 nm SH)

molecular hyperpolarizability tensor components are obtained by

G-09W taking into account solvent (CHCl3) effect (keywords

SCRF¼ IEFPCM andPolar¼DCSHG). Two hyperpolarizability

quantities of experimental interest could be calculated from these

tensor components. The vector component along dipole moment

bm37 is important for EFISHG and external electrical field poling

experiments. Another value, bHRS,38 is related to HRS measure-

ments of molecular hyperpolarizability in the solutions.

Results and discussion

Materials

It should be noted that the trityl and triphenylsilyl substituents

used in the study as the amorphous phase formation enhancers

are widely used in organic chemistry as easily cleavable pro-

tecting groups for alcohols. This restricts the use of materials

because during both the processing and practical application

several precautions need to be taken. In particular, these

compounds are labile towards acidic hydrolysis (both trityloxy

and triphenylsilyloxy), basic hydrolysis and fluorine ion presence

(triphenylsilyloxy). The chloroform used in our spectral

measurements and spin-coating process needed to be analytical

grade and chemically stabilised to avoid the presence of trace

amounts of HCl. No decomposition of the compounds was

observed when storing the compounds in solid form. It is worth

mentioning that in order to improve the materials’ chemical

stability structurally new molecular glasses have been obtained

where the trityl moiety is C–C bonded to a chromophore core.

The results regarding those materials will be reported elsewhere.

Quantum chemical modeling

Optimization of molecular structures revealed noticeable differ-

ences between the shapes of molecules depending on whether

trityl or triphenylsilyl groups were present. A fan-shaped triaryl

pseudo-sphere is sterically bulkier when Si is present in the core

due to the Si–CPh bond (1.884 �A) being significantly longer than

the C–CPh bond (1.544 �A). At the same time, according to our

modelling by means of RHF ab initio calculations (Fig. 1), such

an elongated bond allows more rotational freedom for phenyl

groups. The combination of increased size and less defined

conformation of the substituent allows us to conclude that the

triphenylsilyl group would be a more suitable substituent for

amorphous phase formation enhancement. Further results

11272 | J. Mater. Chem., 2012, 22, 11268–11276

revealed that this was the case. The results of quantum chemical

(QC) modeling of synthesized compounds and reference chro-

mophores (DR1, DR19 and DMABI (N,N-dimethylamino-

benzylidene 1,3-indandione)) in CHCl3 solution are presented

(Table 1). According to our quantum chemical calculations the

trans configuration for all azo dyes is energetically preferable, so

the given values correspond to this conformer.

Introduction of trityl and triphenylsilyl groups just slightly

changes the chromophore NLO properties. If compounds A-1

and A-2 are compared with DR19, the dipole moment as well as

hyperpolarizability values of experimental interest are within less

than 3% variation. Replacement of the nitro group with an

additional trityl moiety in the acceptor part (A-4 and A-5)

decreases conjugation effects in the chromophore. As a result of

the weakened acceptor, these molecules have a lower dipole

moment and molecular hyperpolarizability (bHRS) as well as

hypsochromically shifted CT absorption band maxima (Table 2).

Additionally, the presence of chlorine in the ortho position for

A-4 and A-5 causes a spatial mismatch of hyperpolarizability

with the dipole moment—the angle between them is 65 and

78 degrees, respectively. Due to this the external poling field

vector component along the dipole moment bm is additionally

diminished. Obviously, in such a situation one could not expect

to achieve highly NLO active films by external electrical field

poling as it is demonstrated later in this paper. As a comple-

mentary example, in the case of A-3, ortho substitution with

bromine in the presence of a strong acceptor group (NO2) at the

para position has no negative impact on the spatial alignment of

dipole moment and hyperpolarizability. It is worth mentioning

that at the same time bromine enhances overall acceptor

strength, and therefore a higher hyperpolarizability and a bath-

ochromic shift of the charge transfer (CT) band takes place for

A-3 in comparison with A-1 and A-2.

Linear and nonlinear optical properties in solutions

The CT band maxima wavelength and extinction coefficients of

diluted solutions of the compounds (10�6–10�5 mol l�1) in CHCl3

This journal is ª The Royal Society of Chemistry 2012

Table 1 Results of RHF 6-31G(p,d) quantum chemical calculations in CHCl3

Compound m, D cos<bm>bm(0)� 1030 esu

bHRS(0)� 1030 esu

bm(532)� 1030 esu

bHRS(532)� 1030 esu

DR1 10.2 0.99 56.4 41.0 81.9 58.6DR19 9.3 1.00 55.0 39.4 79.4 55.9A-1 9.2 1.00 56.5 40.6 83.4 58.9A-2 9.5 1.00 56.9 40.5 83.4 58.7A-3 9.4 0.99 64.7 46.6 97.5 69.4A-4 2.6 0.43 13.5 22.9 20.7 33.6A-5 2.1 0.21 6.0 22.1 9.6 31.6DMABI 4.3 0.99 38.9 28.3 54.5 39.5I-1 2.8 0.95 39.9 29.9 57.1 42.9I-2 3.1 0.97 41.4 30.0 59.2 43.0

Table 2 Results of UV-vis spectral and HRS measurements of investi-gated compounds in CHCl3

Compoundlmax,nm 3max

bHRS(532)� 1030 esu

bHRS(0)� 1030 esu

DR1 481 35700 780 113A-1 486 32900 710 93A-2 485 30900 870 116A-3 506 33100 870 64A-4 433 32300 320 90A-5 431 32900 290 83DMABI 482 70800 950 135I-1 485 70900 1750 224I-2 485 72200 1890 253

Dow

nloa

ded

by L

udw

ig M

axim

ilian

s U

nive

rsita

et M

uenc

hen

on 2

5 Fe

brua

ry 2

013

Publ

ishe

d on

13

Apr

il 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2J

M30

861D

View Article Online

are given (Table 2). In each case a single wide absorption peak

characteristic of the given chromophore classes is observed.

Comparison of triaryl substituted compounds (A-1 and A-2, I-1

and I-2) with the corresponding reference chromophores DR1

and DMABI shows no noticeable differences in the absorption

bands (Fig. 2), indicating that no interactions occur between the

chromophore core and modifying groups.

The molecular hyperpolarizabilities of the azo compounds

obtained by HRS are in good correlation (0.97) with quantum

chemical calculation results. At the same time, probably as

a result of underestimation of resonance effects by RHF calcu-

lations, a scaling factor (�12) between QC-calculated and HRS-

measured values takes place. In the case where indanedione

derivatives were included in the comparison of measured and

QC-calculated values this correlation breaks down. Values

measured by HRS for indanedione derivatives are much higher

than one would expect from QC calculations. This disparity

could be explained by the high proportion of TPL signal in the

measured 532 nm light intensity. Numeric simulations showed

that when TPL has an equal or larger intensity than HRS

intensity, our deconvolution procedure fails, resulting in an

overestimation of bHRS values. Clarification of bHRS values for

indanedione derivatives will be reported in the future after more

detailed spectroscopic HRS investigations.

Fig. 2 UV-vis absorption bands of triaryl functionalized chromophores

and their non-modified counterparts. No changes are observed in

absorption properties.

Linear and nonlinear optical properties of films

The UV-vis absorption spectra of thin films are presented in

Fig. 3. Given the fact that the chromophores under investigation

have high dipole moment values, under normal conditions

This journal is ª The Royal Society of Chemistry 2012

head-to-tail stacking (J-aggregation) was expected to dominate,

producing a bathochromic shift. The lack of such shifts in the

spectra leads us to assume that sterically bulky triaryl substitu-

ents do not allow the preferred solid phase packing of chromo-

phores. Furthermore, final packing in the solid phase determined

by p–p stacking of aromatic rings causes the apparent presence

of hypsochromically shifted shoulder peaks, characteristic of the

formation of H-aggregates where linear chromophores are

stacked face-to-face with each other.

The absorbance maxima wavelengths, corresponding absorp-

tion coefficient, as well as refractive indices and NLO coefficients

J. Mater. Chem., 2012, 22, 11268–11276 | 11273

Fig. 3 UV-vis absorption coefficient spectra of several thin films.

Hypsochromic shift is observed in all cases.

Table 4 Decomposition, glass transition and TSHI50 temperatures ofmaterials

Compound Tdec,�C Tg,

�C TSHI50,�C

A-1 292 75 49A-2 295 51 52A-3 315 53 53A-4 300 —a 84A-5 330 71 56I-1 330 88 85I-2 340 62 56

a No Tg peak was observed in DSC thermograms.

Dow

nloa

ded

by L

udw

ig M

axim

ilian

s U

nive

rsita

et M

uenc

hen

on 2

5 Fe

brua

ry 2

013

Publ

ishe

d on

13

Apr

il 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2J

M30

861D

View Article Online

for the investigated compounds are presented in Table 3. This

table presents the experimental d33 values corrected to zero

frequency according to the two level model,39 which is usually

used for chromophore comparison and benchmarking. The

compound A-1 is an exception, as this sample underwent fast

crystallization after the corona poling procedure at temperatures

close to Tg.

The obtained NLO coefficients are within expectations

based on QC calculations. The highest nonlinearity d33(532) ¼300 pm V�1 was achieved in the case of A-3. The high value may

be explained by more pronounced resonance effects due to the

proximity of the chromophore absorbance band maximum to the

SH frequency. In any case, value d33(0) ¼ 19.6 pm V�1 corrected

to zero frequency is still the highest observed among all investi-

gated structures. The weakest nonlinearity among the azochro-

mophore based glasses is observed for compounds A-4 and A-5.

As discussed above, this is caused by the lost acceptor strength

and the spatial mismatch of hyperpolarizability with the dipole

moment. Just a slightly higher nonlinearity was achieved for

indanedione based films. In comparison with A-2, smaller bm and

dipole moment value characteristics for I-2 result in an almost

five times lower d33(0) value under similar poling conditions.

Based on QC calculation results one could expect quite similar

behaviour for I-1, but that is not the case. For these films, the

lowest ratio of d33/d31 ¼ 1.76 was observed indicating incomplete

poling. For the structurally similar chromophore pairs A-4, A-5

and I-1, I-2, under the applied poling conditions a higher

nonlinearity was observed for the triphenylsilyl compounds. This

can be expected from the higher conformational flexibility

predicted by QC modelling. During chromophore poling this

enhanced flexibility allows easier alignment of molecular dipoles

Table 3 Linear and nonlinear optical properties of amorphous phase

Compoundlmax,nm

amax

� 10�4, cm�1 RI532

A-1 490 3.23 —A-2 487 3.41 1.93A-3 509 4.85 2.34A-4 437 2.71 1.84A-5 438 2.09 1.79I-1 490 5.03 1.90I-2 489 5.56 1.94

11274 | J. Mater. Chem., 2012, 22, 11268–11276

along the external electrical field direction. It is probable that, for

trityl moiety containing materials, a longer poling time is

necessary to achieve higher polar order.

Thermal and phase behaviour properties

The thermal properties of the synthesized compounds were

explored using differential scanning calorimetry (DSC) and

complementary data showing temperature dependent decay of

second harmonic generation intensity TSHI50. The results of these

measurements are given in Table 4.

The DSC thermograms analyzed (Fig. 4) were taken during

the second heating of the samples that underwent initial heating

above their melting temperature, followed by rapid cooling

(50 �C min�1). Although compounds form glass during slow

cooling of melt (10 �Cmin�1) fast cooling was necessary to obtain

distinct Tg peaks. Even with fast cooling and repeated

measurements no Tg signal was obtained in the case of

compound A-4 (see Supplementary Information† for full heat-

ing–cooling–heating cycles). In the all cases, except for

compound I-1, the absence of any additional peaks associated

with the crystallization or melting processes in the temperature

range up to 200 �C indicates the formation of a thermodynami-

cally stable amorphous state. Compound I-1 shows an

exothermic crystallization peak during both heating and slow

(10 �C min�1) cooling at 170 �C. This peak disappears during

a fast cooling (50 �C min�1).

With the DSC measured glass transition temperature values,

Tg well above ambient, the materials are considered stable at

room temperature. They are in the good agreement with the half

NLO activity temperatures (TSHI50), evidently linking the loss of

acentric molecular order and increased conformational freedom.

Considering the chemical structures of the compounds, it is

apparent that the Tg values are greatly influenced by the chosen

RI1064

d31(532),pm V�1

d33(532),pm V�1

d33(0),pm V�1

— — — —1.61 31.9 83.5 10.71.69 135.0 300.0 19.61.65 3.2 9.5 2.571.62 3.7 11.5 3.081.60 9.3 16.4 1.961.59 12.6 34.1 4.17

This journal is ª The Royal Society of Chemistry 2012

Fig. 4 DSC thermograms showing Tg peaks.

Fig. 6 Optical microscope images of corona poled samples for

compounds I-1 (a) and I-2 (b) after a month of storage. Sample (a)

contains numerous growing crystals while sample (b) reveals no visible

signs of crystallization.

Dow

nloa

ded

by L

udw

ig M

axim

ilian

s U

nive

rsita

et M

uenc

hen

on 2

5 Fe

brua

ry 2

013

Publ

ishe

d on

13

Apr

il 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2J

M30

861D

View Article Online

triaryl group. Structurally similar chromophore pairs A-4, A-5

and I-1, I-2 in both cases show increased material thermal

stability by up to 30 �C when a trityl group is present as

demonstrated in Fig. 5.

This observation once more confirms our QC calculation

results, showing that the C–CPh bond allows less rotational

freedom for phenyl groups than the Si–CPh bond, thus increasing

energy requirements for phase transition processes. While

rigidness of the molecule may seem a counter-intuitive property

for obtaining amorphous material, the steric demands of trityl

group seem sufficient to inhibit aggregation processes under

supercooled liquid conditions. Nevertheless, the general

assumption, that a stand-alone triphenylsilyl group is better

suited as an amorphous phase formation enhancer, is still valid

according to our observations. Quality evaluation of the

prepared samples with an optical microscope after a month of

storage showed signs of crystallization processes in compounds

containing the trityl group. Comparative illustrations of the

samples made from compounds I-1 and I-2 are given in Fig. 6.

Moreover, in the case of compound A-1, complete crystallization

was observed in a matter of days after the samples underwent

poling close to Tg. While the trityl group can generally be

considered less suitable for stable glass formation, it is interesting

to note that the examination of samples made from the

compound containing an additional trityl moiety in the acceptor

Fig. 5 Polar order thermal stability of structurally comparable chro-

mophore pairs. Trityl group containing compounds show significantly

higher thermal sustainability.

This journal is ª The Royal Society of Chemistry 2012

part, A-4, revealed no signs of crystallization, indicating

increased amorphous phase stability in the presence of bulky

substituents at both ends of the linear chromophore.

Conclusions

In this paper we have presented a new structural design of low

molecular organic glasses suitable for photonics studies. The

synthesized triaryl functionalized N-phenyldiethanolamine

derivatives have shown excellent solubility in non-polar organic

solvents and are able to form good optical quality glassy films

without mixing in a polymer matrix. The stability of the films was

sufficient to undergo corona discharge poling at elevated

temperatures, making the materials NLO active.

Triphenyl moieties as crystallization preventing structural

elements have shown no effect on electronic properties of push–

pull NLO chromophores. QC calculations have revealed

noticeable differences in the chosen triphenyl moiety structure

regarding the conformational freedom and occupied volume.

Due to these dissimilarities the replacement of carbon with

silicon in the triphenyl core has a conflicting impact on the

properties of the investigated NLO active organic glasses. On the

one hand, the amorphous phase formation favours the presence

of a bulkier and conformationally less defined triphenylsilyl

group which was most evident in the case of azochromophore

containing compounds. At the same time this structural element

reduces the thermal stability of polar order in corona poled films.

For structurally comparable pairs (A-4 and A-5, I-1 and I-2) the

temperature corresponding to the half NLO activity, TSHI50, is

lowered by �30 �C. It is noteworthy that the presence of an

additional trityl group in the acceptor part of azochromophores

enhanced the amorphous phase stability and the polar order

thermal stability of corona poled films. At the same time the

increased rigidness of triphenyl groups resulted in lowered

nonlinearity due to obstructed alignment of molecular dipoles

along the externally applied electrical field direction during

corona discharge poling.

Acknowledgements

This work has been supported by the European Social Fund

within the project ‘‘Support for the implementation of doctoral

studies at Riga Technical University’’ and European Recon-

struction and Development Fund project LIAA Nr.2010/0308/

2DP/2.1.1.1.0/10/APIA/VIAA/051; ‘‘Development of polymer

based electro-optical modulator prototype’’.

J. Mater. Chem., 2012, 22, 11268–11276 | 11275

Dow

nloa

ded

by L

udw

ig M

axim

ilian

s U

nive

rsita

et M

uenc

hen

on 2

5 Fe

brua

ry 2

013

Publ

ishe

d on

13

Apr

il 20

12 o

n ht

tp://

pubs

.rsc

.org

| do

i:10.

1039

/C2J

M30

861D

View Article Online

References

1 L. R. Dalton, P. A. Sullivan and D. H. Bale, Chem. Rev., 2010, 110,25–55.

2 X. Zhang, I. Aoki, X. Piao, S. Inoue, H. Tazawa, S. Yokoyama andA. Otomo, Tetrahedron Lett., 2010, 51, 5873–5876.

3 J. Hao, M.-J. Han, K. Guo, Y. Zhao, L. Qiu, Y. Shen and X. Meng,Mater. Lett., 2008, 62, 973–976.

4 S. Wang, L. Zhao, Z. Xu, C.Wu and S. Cheng,Mater. Lett., 2002, 56,1035–1038.

5 F. Borbone, A. Carella, L. Ricciotti, A. Tuzi, A. Roviello andA. Barsella, Dyes Pigm., 2011, 88, 290–295.

6 U. Caruso, M. Casalboni, A. Fort, M. Fusco, B. Panunzi, A. Quatela,A. Roviello and F. Sarcinelli, Opt. Mater., 2005, 27, 1800–1810.

7 S. M. Tambe, A. A. Kittur, S. R. Inamdar, G. R. Mitchell andM. Y. Kariduraganavar, Opt. Mater., 2009, 31, 817–825.

8 S. Fusco, R. Centore, P. Riccio, A. Quatela, G. Stracci, G. Archettiand H.-G. Kuball, Polymer, 2008, 49, 186–191.

9 S. Lee, M. Cho, G. Kim, W. Jun, J. Jin, L. Dalton and D. Choi, Opt.Mater., 2007, 29, 451–456.

10 R. R. Barto, C. W. Frank, P. V. Bedworth, R. E. Taylor,W. W. Anderson, S. Ermer, A. K.-Y. Jen, J. Luo, H. Ma,H.-Z. Tang, M. Lee and A. S. Ren, Macromolecules, 2006, 39,7566–7577.

11 Z. Li, G. Yu,W.Wu, Y. Liu, C. Ye, J. Qin and Z. Li,Macromolecules,2009, 42, 3864–3868.

12 J. J. Choi, K. M. Kim, J. S. Lim, C. Lee and D. W. Kim, Macromol.Res., 2007, 15, 59–64.

13 C.-S. Choi, I. K. Moon and N. Kim, Appl. Phys. Lett., 2009, 94,053302.

14 T.-D. Kim, J.-W. Kang, J. Luo, S.-H. Jang, J.-W. Ka, N. Tucker,J. B. Benedict, L. R. Dalton, T. Gray, R. M. Overney, D. H. Park,W. N. Herman and A. K.-Y. Jen, J. Am. Chem. Soc., 2007, 129,488–489.

15 K. Kreger, P. Wolfer, H. Audorff, L. Kador, N. Stingelin-Stutzmann,P. Smith and H.-W. Schmidt, J. Am. Chem. Soc., 2010, 132, 509–516.

16 J. D. Wuest and O. Lebel, Tetrahedron, 2009, 65, 7393–7402.17 Y. He, X. Gu, M. Guo and X. Wang, Opt. Mater., 2008, 31, 18–27.18 H. Nakano, T. Takahashi, T. Tanino and Y. Shirota, Dyes Pigm.,

2010, 84, 102–107.

11276 | J. Mater. Chem., 2012, 22, 11268–11276

19 R. Lygaitis, V. Getautis and J. V. Grazulevicius, Chem. Soc. Rev.,2008, 37, 770–788.

20 A. Meunier and O. Lebel, Org. Lett., 2010, 12, 1896–1899.21 Y. V. Pereverzev, O. V. Prezdho and L. R. Dalton, Chem. Phys. Lett.,

2001, 340, 328–335.22 Z. Li, P. Li, S. Dong, Z. Zhu, Q. Li, Q. Zeng, Z. Li, C. Ye and J. Qin,

Polymer, 2007, 48, 3650–3657.23 Z. Li, G. Yu, S. Dong, W. Wu, Y. Liu, C. Ye, J. Qin and Z. Li,

Polymer, 2009, 50, 2806–2814.24 H. Taunaumang, Herman, M. O. Tjia and M. Samoc, Opt. Mater.,

2003, 22, 289–294.25 A. L. P�erez-Mart�ınez, T. Ogawa, T. Aoyama and T. Wada, Opt.

Mater., 2009, 31, 912–918.26 S. Proutiere, P. Ferruti, R. Ugo, A. Abbotto, R. Bozio, M. Cozzuol,

C. Dragonetti, E. Emilitri, D. Locatelli, D. Marinotto, G. Pagani,D. Pedron and D. Roberto, Mater. Sci. Eng., B, 2008, 147, 293–297.

27 M. A. Rutkis, A. Vembris, V. Zauls, A. Tokmakovs, E. Fonavs,A. Jurgis and V. Kampars, Proc. SPIE, 2006, 6192, 61922Q.

28 T. L. Bouder, P. Massiot and H. L. Bozec, Tetrahedron Lett., 1998,39, 6869–6872.

29 G. Vanags, R. Silaraja and G. Glemitis, Chem. Abstr., 1963, 58, 483d.30 D. Roche, K. Prasad, O. Repic and T. J. Blacklock, Tetrahedron Lett.,

2000, 41, 2083–2085.31 R. A. Benkeser and R. B. Gosnell, J. Am. Chem. Soc., 1956, 78, 4914–

4916.32 J. Campo, F. Desmet, W.Wenseleers and E. Goovaerts,Opt. Express,

2009, 17, 4587–4604.33 E. Nitiss and M. Rutkis, Proc. SPIE, 2012, 8430, 8430–49.34 W. N. Herman and L. M. Hayden, J. Opt. Soc. Am. B, 1995, 12, 416–

427.35 J. Jerphagnon and S. K. Kurtz, Phys. Rev. B: Solid State, 1970, 1,

1739–1744.36 M. J. Frisch et al., Gaussian 09, Revision A.1, Gaussian, Inc.,

Wallingford CT, 2009.37 H. A. Kurtz, J. J. P. Stewart and K. M. Dieter, J. Comput. Chem.,

1990, 11, 82–87.38 T. Verbiest, K. Clays and V. Rodriguez, Second-Order Nonlinear

Optical Characterization Techniques, Taylor & Francis, 2009.39 J. L. Oudar and D. S. Chemla, J. Chem. Phys., 1977, 66, 2664–

2668.

This journal is ª The Royal Society of Chemistry 2012