Electro-Optics: The Phenomena

• An electro-optic material (device) permits electrical and optical

signals to “talk” to each other through an “easily perturbed”
electron distribution of a material. A low frequency (DC to 200
GHz) electric field (e.g., a television [analog] or computer
[digital] signal) is used to perturb the electron distribution (e.g.,
p-electrons of an organic chromophore) and that perturbation
alters the speed of light passing through the material as the
electric field component of light (photons) interacts with the
perturbed charge distribution.

• Because the speed of light is altered by the application of a

“control” voltage, electro-optic materials can be described as
materials with a voltage-controlled index of refraction.

Index of refraction = speed of light in vacuum/speed of light in

material
 Electro-Optic Devices: The on-ramps &
interchanges of the information superhighway
• By slowing light down in one arm of the Mach Zehnder device shown below, the
interference of light beams at the output can be controlled. Electrical information
appears as an amplitude modulation on the optical transmission. This works equally
well for analog or digital data. Electro-optic coefficient is the material parameter
that defines how large an effect is observed for a given applied voltage
l = optical wavelength
The Mach Zehnder Interferometer n = index of refraction
r33 = electro-optic coefficient
Vp = ld/(2n3r33LG) L = interaction length
G= modal overlap integral
d = electrode gap
DC bias electrode
ground electrode

Modulated
Light In Light Out
RF electrode

Substrate

Vp is the voltage required to achieve signal transduction

Electro-Optic Behavior Depends on Orbital
Type and Position
Pi-electrons are more easily perturbed (displaced) than
sigma-electrons
 Example of common NLO chromophore design

NC
NC
CN
Electron S S Electron
N
Donor O Acceptor

p-conjugated bridge

 PAS 38, a classical linear charge-transfer chromophore

shows typical chromophore design; an electron donor and acceptor
pair separated by a p-conjugated bridge
Example of an E-O Chromophore: A Charge Transfer
(Dipolar) Molecule
Charge displacement can occur over estended distances using materials
with extended p-conjugation

H H
H H H
C
H C C C
N C C O
H
C H H

The charged separated form H H

will have a larger dipole

moment and thus a stronger H H
H H H
interaction with an applied C
H
electric field N
C
C
C
C
C
O
H
C H H
H H
 NONLINEAR OPTICAL EFFECTS: Microscopic and
Macroscopic Polarization—Power Series Expansion

pi   ij E j   ijk E k E l   ijkl E j E k E l  ...

  is the first nonlinear term, known as molecular first

hyperpolarizability
 For a symmetric molecule even order terms,  and higher, are zero.

2 3
Pi   ij E j   ijk Ek E j   ijkl E j E k E l  ...

 (2)represents material first nonlinear susceptibility

  Oscillators (chromophores) must be aligned acentrically within

the material to realize a nonzero (2)
The coefficient of the second term in the power series
expansion of material Polarization in terms of applied
electric fields, (2), is given by
 ( 2)
zzz N zzz  cos   (const )
3

2
r33  2 zzz /(nz ) 4

r33 = N<cos3>(constant)
Loading Parameter = N<cos3> = (r33/)(constant)
 r33 = electro-optic coefficient
N = chromophore number density (molecules/cc)
 = molecular first hyperpolarizability
N<cos3> = acentric order parameter
*The constant depends on the dielectric properties of the
material lattice
 Optimization of Electro-Optic Activity
Macroscopic Level
Electro-optic activity requires noncentrosymmetric
chromophore symmetry, i.e., <cos3> must be large
Requires optimization of N<cos3>
*Statistical mechanics is the key

Molecular Level
Requires optimization of 
*Quantum mechanics is the key

Let us first focus on the optimization of N<cos3>,

then we will consider the optimization of .
 Use Electric Field Poling to Induce Noncentrosymmetric Order
A DC poling field is applied across the chromophore /
host matrix.

E
Top electrode

CN CN

O NC O
NC
CN
CF3
CF3 O NC
NC NC
CN
CF3
O NC
NC
CN CN
CF3
O NC O
NC NC

F NC
CF3

NC
CN

NC
O

CF3
NC

NC
CN
O

CF3
NC
CF3

i TBDMSO
N
OTBDMS
N
OTBDMS
TBDMSO
N
OTBDMS
TBDMSO
OTBDMS

N N

e
TBDMSO
OTBDMS OTBDMS
TBDMSO TBDMSO
N N
OTBDMS OTBDMS
TBDMSO TBDMSO

l Bottom electrode

d
Ideal case (no intermolecular interactions) : < cos3 θ> = E / 5kT
NONCENTROSYMMETRIC SYMMETRY REQUIRED FOR ELECTRO-
OPTIC ACTIVITY
Translating Microscopic to Macroscopic Electro-Optic Activity
r33 = N<cos3>(constant)

L = Langevin Function; W = Intermolecular Electrostatic Potential; k = Boltzmann

constant; Ep is applied poling field; F is poling field felt by chromophore
EO Activity vs. Concentration: Theory & Experiment

CLD

in
APC

Statistical mechanical calculations permit understanding the

role that chromophore shape has on macroscopic electro-
optic activity
Chromophore Shape Determines EO Activity
Loading Parameter = N<cos3>  r33/(constant)

Region of Enhanced Order

 N

RO O
S
1 RO O
O OR
S
O OR

NC
2 O
NC
CN

With a 2-1 aspect-ratio dipolar head-tail interactions are becoming

predominant over side-side interactions

Undesired Centrosymmetric Order Desired Noncentrosymmetric (EO active) Oder

Towards improved loading: discotic
chromophores
 Progress In Discotic Chromophores
Relative r33 values for a series of modified bi(ProDOT) core in bridges.
The more disc shaped, the more the dipole-dipole interaction are reduced.
Prediction: OLD-4 can go to higher loading and still get improved electro-optic
conversion. The wt% loading is based on core chromophore weight only.
OMe

Bu Bu Me

O O O O
N S N S
O O
Wt % 20% 30% O O
S

NC NC
CN O O
S

NC NC
CN

Bu Bu Me

OLD-1 3.0 --- OMe

OLD-1 OLD-2 OMe
OLD-2 2.0 3.6
Me

OLD-3 1.0 --- N S

N S
O O
F3C
O S O
S

OLD-4 4.0 --- OLD-3

NC NC
CN O

Me
O
NC NC
CN

OMe

OLD-4
 Blends of Organic
Molecular Glass Materials
O

AJC-AJC (50%) HD/FD-AJC(75%) O

O
N

400 AJC-AJC(75%)
O

F
O
F
O
O
S F F
350 O F
F

HD/FD-AJC(50%) NC

O
CF3

F
F

F
NC
F

300 HD/FD-AJC(23%)
CN

HD/FD
r33 (pm/V)

250

200 Si
O
N
O
Si Si
O
N
O
Si

150

100
NC CF3 NC CF2CF2CF3
O O
50 NC
CN
S NC
CN
S

0
0 20 40 60 80 100 AJC146 AJC168
Poling field (V/um)
Multi-Chromophore-Containing Dendrimers Require the Use
of Psuedo-Atomistic Monte Carlo Calculations to Simulate
Poling-Induced Noncentrosymmetric Order

Simulation requires consideration of

Intermolecular electronic (e.g., dipole-dipole) interactions
Intermolecular nuclear repulsive (steric) interactions
Potential functions associate with rotation about covalent bonds
Van der Waals interactions

It is critical to reduce computation time by making logical

approximations
Treat pi-electron segments within the United Atom
Approximation (Pi bonds prevent rotation and segments are thus
rigid)
Treat sigma-electron segments atomistically using correct bond
rotation potentials
 Examples of Monte Carlo Simulations
Consideration of Various Modes of Attachment of Chromphore
NC
NC
CN
N S O N
CN
NC
O O O
O NC
O S CN
CN
O O N
S O O O CN

NC O
O O
NC
O
S O
NC
O O N
O O
O O
S N
CN
O S
CN
NC O

NC CN
N CN

Side-on End-on
D2.PAS.31 AALD-1104
IV V
 D2.pas.41 VIII MC modeling in static DC field (Simulation
of Results for Side-On Attachment Dendrimer)
• Chromophores are the prolate ellipsoids. Solvent is assumed.

Low Density
2% loading
cos   0.16

Simmulations performed by Harry Rommel and Bruce Robinson.

 PAS Side-On Dendrite: High Density

<cos3> greater than 0.3

 Discussion of Results
Results are complicated as might be expected when multiple
forces influence results
Van der Waals interactions are clearly important in defining the
order achieved in the preceding two figures
Thus, results are not as simple as for the case of single
chromophore dendrimers where the competition of electronic and
nuclear intermolecular interactions dominate
Covalent bond potentials do prevent dipole-dipole interactions
from diving centrosymmetric ordering
The two preceding slides clearly demonstrate that covalent
bond potentials and nuclear repulsive (steric) interactions can
result in order parameter increasing with concentration
The next slide shows another possibility: The sum of
interactions favors the noncentrosymmetric ordering potential
component of the electronic intermolecular interactions
 Optimum Monte Carlo Calculated Structure
--Noncentrosymmetric Order By Design--
• Three chromophores (End-on attachment) -- 20 Debye dipole each
• Best of Set of accepted M.C. moves.
• Result: Near perfect order of 3 Chromophores
• This is a limiting case: Very Large Field ( 3000 MV/m) Points on Z (up)

Magic
Angle

Chromophore

Center
 Another Experimental Demonstration: Comparison of
chromophore/APC composite with pure three arm
chromophore dendrimer (D2PASS)

r33 at 1310 nm wavelength (pm/V)

200
The demonstration of the CF3FTC
advantage of incorporating CF3FTC
D2PASS
a chromophore into a 150 D2PASS
multi-chromophore
dendrimer is shown in the 100
figure to the right. An
electro-optic activity
approximately three times 50
greater than the best value
that can be obtained for a 0
chromophore/polymer 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
composite is achieved. Poling field (MV/cm)

NC NC
TBDMSO CN
S O
N CF3

CF3-FTC
TBDMSO
 Summary of Multi-Chromophore Dendrimer
Results

Electro-optic coefficients (at 1.3 microns wavelength) in the

range 200-430 pm/V have been achieved with multi-
chromophore containing dendrimers

Ultimate electro-optic values (supermolecular structures) have

not yet been achieved so values can be expected to become
larger in the future

Electro-optic activity can also been increased by inserting

chromophores with greater molecular first hyperpolarizability
into optimized supermolecular structures. We now turn our
attention to optimizing chromophore hyperpolarizability
 Comparison of Experimental Results and Theoretical Predictions

O
F 3C
O
Hyperpolarizability values
Fe
NC
CN Fe
NC
CN relative to pNA measured by
NC NC
HRS and calculated by DFT.
1 2
NC
CN
NC
CN Electro-optic coefficients
NC NC
O O determined by simple
S S
F 3C
reflection
Fe Fe

3 4

Cmpd r33(pm/V) Experiment.  DFT Calculations

# 1.3@20% relative to pNA rel
calc
, zzz
1 ---- 3.5 4.6 2.4
2 5 5.7 4.8 3.6
4 /3 25 / ---- 42.2 /33.3 43.5 / 35.5 44 / 11.4
Theoretical Prediction of Variation of Molecular First Hyper-
polarizability and Dipole Moment with Chromophore Structure
(Semi-Empirical Calculation Predictions)
NC NC NC NC NC NC
CN CN CN
O N O NH
N N
CF3
O
1 10
2
µ = 10.1 D µ = 9.9 D µ = 6.3 D
ß1907 nm = 31.6 x 10 -30 esu ß1907 nm = 46.6 x 10 -30 esu ß1907 nm = 62.4 x 10 -30 esu
µß1907 nm = 319.2 x 10 -48 esu µß1907 nm = 461.3 x 10 -48 esu µß1907 nm = 393.1 x 10 -48 esu
l max, ZINDO = 390 nm l max, ZINDO = 403 nm l max, ZINDO = 430 nm

NC CN NC CN
CN
O N
N N N CN
O O
302
301

µ = 11.7 D µ = 13.6 D
ß1907 nm = 60.2 x 10 -30 esu ß1907 nm = 91.7 x 10 -30 esu
µß1907 nm = 704.3 x 10 -48 esu µß1907 nm = 1247.1 x 10 -48 esu
l max, ZINDO = 431 nm l max, ZINDO = 459 nm
Determination of Molecular 1st Hyperpolarizability (relative to CHCl3) by
Femtosecond, Wavelength-Agile Hyper-Rayleigh Scattering (HRS)
--Experiment confirms theoretical prediction--

NC NC NC
NC
CN CN
N O N
N O
O
TCF
TCP2
2285
7582 O
NC NC
CN
N N NC
NC
O
CN
N
TCP1 N
O
8288 O O
TCP3
7943

chloroform = 0.16 x 10-30 esu [Kaatz et al, Opt. Commun. 157 (1998) 177]
chloroform = 0.49 x 10-30 esu [Clays et al, Phys. Rev. Lett. 66 (1991) 2980]
Density Functional Theory (DFT) Calculations Predict
The Same General Trends

Designation Structure HRS/10-30 esu zzz/10-30 esu

CLD (Ref.) 146 293

D1-B2-A2 183 385

D1-B2-A4 223 454

D1-B9-A3 500 990

D1-B9-A4 585 1214

D1-B10-A4 794 1681

 Strategy for Improving NLO Chromophores: Choose the Right
Combination of Donor, Bridge, and Acceptor With Theoretical Guidance
R R

S S Stronger acceptors :
R R  Pyrrolines
 pyrrolizines
Extended bridges:
Stronger donors :

NC NC
R
N CN
R CN
N CN N
N R
N F 3C O CN O
R
NC CN

O
N
P N
Asymmetric bulky 3D O
shaped acceptors: OR

RO

NC NC CN
NC NC NC
CN
CN CN N
O CN
F3C O O
F3C OR
S
RO
R
OR
 Summary
Values of electro-optic activity greater than 300 pm/V (an order
of magnitude greater than the commercial standard lithium
niobate) have been realized for both single and multi-
chromophore-containing dendrimers

Much greater values can clearly be obtained by

Using pseudo-atomistic Monte Carlo calculations to design
dendrimers that lead to even larger values of N<cos3>
Using quantum mechanical calculations to design chromophores
with improved molecular first hyperpolarizability, .

Values of greater than 1000 pm/V are likely possibility. Such

materials would likely have a transformative impact on a number
of technological areas