Significant influence of surface states on the electroluminescence of CdS nanoparticles

The

_{significant influence}

of

surface states on the

electroluminescence

of

CdS

_{nanoparticles}

Eih-Zhe

_Lianga,

_Ching-Fuh

Ljfl*a, Sheng-Ming Shih", Wei-Fang

Sub

a

_Graduate

_Institute

_of

Electro-Optical Engineering

b

Institute

of

Materials Science and

Engineering

National

Taiwan University, Taipei

1

_06,

_{Taiwan, ROC}

ABSTRACT

The _significance

of

surface states

in

nano-structures

is

_{studied using CdS nanoparticles. Spectral} features like peak

red-shift due

to

organic capping and influence

of

surface states have been observed. The _{pronounced enhancement}

of

emission from surface states

can be

dominant with certain modification

of

CdS _{nanoparticles. Spectral behaviors}

of

electroluminescence

in

different _temperature

are

also studied.

Keywords: CdS nanoparticles, low-dimensional structure, surface states, electroluminescence, exciton.

1. ThTRODUCTION

Nanoparticles formed

_by

chemical methods have _{certain advantages. They have low cost}

of

production and well

control _{physical properties. Compared with epitaxial quantum dots, nanoparticles} can

be

_simply dissolved

in

various solutions _{and applied}

to

_nonspecific substrates. Stimulated _{emission and optical gain had been reported with} CdS quantum dots

by

optical pumping1'2. However,

it

is

challenging

to

employ electrical _pumping

to

realize _{nanoparticle-} based light emitting devices. Electroluminescence from CdS _{nanoparticles reported}

in this

_{work proves}

the

_feasibility

of

using such materials

as

active _{light-emitting media.}

Since nanoparticles are low-dimensional materials

(

5

_{nm), compared} with bulk materials,

it

results

in

_large surface contact area with environment. Surface states formed

_by

_{termination with oxygen}

or

other contaminants are thus

of

great amount. Generally

this

situation

is

avoided

by

using passivation around nanoparticles,

in

our case organic

p-hydroxyl thiophenol group. However, we found that there can

be

significant enhancement

of

light emission from surface states.

It

can be

useful

in

addition

to

_{intrinsic quantum states provided by} low-dimensional structures.

Electroluminescence

of

CdS _{nanoparticles}

in

different environment such

as

_{normal treatment, heat treatment} and

oxygen enrichment

is

achieved. Emission spectrum

of

CdS _{nanoparticles}

is

_{significantly influenced by both process} temperature and oxygen surrounding condition. Radiative recombination due

to

free exciton

in

CdS

is

observed, with spectral peak shift due

to

organic encapping. As surrounding oxygen content level

is

raised, radiative recombination

from _{surface states emerges.} Side effect such

as

coalescence

of

CdS _{nanoparticle into bulk form also presents}

in

_raising process temperature.

At

different temperature,

the

EL spectrum

of

CdS nanoparticle remains quite

the

same. Peak shift

is

_{compared with bulk bandgap shift} and ascribed

as

effect

of

_quantum confinement and _{surface configuration.}

This work also demonstrates

the

electroluminescence

of

_{CdS-nanoparticle} on silicon substrate. The fabrication

of

light emitting active layer

is

simply the spin-coating technique. Carriers

to

achieve light emission

can

be

supplied by quantum tunneling through surrounding barrier into nanoparticles. Using silicon as substrate shows

a

_promising

_way

to

monolithically integrate light emission

of

nanoparticles and conventional electrical circuitry. Moreover, optical functional blocks can

be

built

at

relatively low temperature, following traditional process. This advantage eliminates conflicted _{thermal budget}

of

_epitaxial 111-V materials and silicon _circuitry.

2. SIGNifICANCE

OF

SURFACE

STATES

In the

_{unpassivated case, nanoparticle interacts with oxygen as exposed}

to

_atmosphere. Surface states associated

with _oxygen,

in our

_{case Cd-O bond,}

are

formed. The ratio between surface

area

and volume ratio

of

nanoparticle increases

as

individual _{particle size} shrinks into several nanometer _range. Radiative recombination rate with surface * _{Email:cflin@cc.ee.ntu.edu.tw; also with Graduate Institute}

_of

_{Electronics Engineering and Department}

_of

_{Electrical Engineering}

states origin becomes comparable with that

of

CdS core,

or

even dominant. To

be

_{specific, assuming}

n

is the

number

of

atoms along

the

radius

of

nanoparticle,

the

ratio

of

atoms

in the

top surface _layer

to

core atoms

is

_{given by}

(

n3

- (n

-

m),)/

_ _rn)3

where

m is

effective number

of

atoms

of

surrounding surface layer. We can see

in

Fig. 1 that

the

ratio becomes _{unity as}

the

radius shrinks within

n

10.

In

such cases, radiative recombination rate

of

surface states becomes _{comparable to}

intrinsic _{ones, assuming alike oscillation strength.} This situation describes how

the

_{unpassivated CdS nanoparticles} can emit more light with surface state origin than that observed

in

_epitaxial

or

bulk CdS case.

...-—' ..-— —.---—--

.

- . . !• . •_

Fig. 1 Ratio

of

surface atoms

to

_{core atoms with respect to number}

of

_{atoms along} radius.

3.

PREPARATION

OF

MATERIALS

Two kinds

of

CdS _{nanoparticles ready for spin-coating purpose are synthesized}

_by

modification

of

Pietro's

method3. First form

is

CdS _nanoparticle with _{organic capping. Cadmium} acetate _{dihydrate Cd(CH3COO)22H20, 0.80} g, 3 .0 mmole)

is

dissolved

in

20 ml mixed solvent

of

acetonitrile, methanol, and water with volume ratio 1

:

1 :2. Another solution _containing disodium sulfide _{nanohydrate (Na2S9H20,} 0.36

_g,

_{1.5 mmole) and p-hydroxyl thiophenol (0.56 g,} 4.4 mmole)

in

the

same solvent _system

is

_{added into vigorously stirred cadmium acetate} solution. The whole _system was stirred for 18 hours _{without light} illumination. _{After removing} solvent _{and purifying by centrifuge, we} obtained

0.70

_g

_{yellow solid aggregate} ofCdS _{nanoparticles capped by p-hydroxyl thiophenol.}

Second form

of

CdS _nanoparticle

is

_{coated with silica shell. The purpose}

of

preventive use

of

organic component is

to

raise

the

_{thermal budget}

of

whole fabrication _process

and

increase tolerance with low _temperature. The _preparation process

is as

follows. Cadmium acetate dihydrate (Cd(CH3COO)22H20, 1.60 g, 6.0 mmole)

is

dissolved

in

32 ml mixed solvent

of

acetonitrile and water with volume ratio 1

:

1

.

Another solution _containing disodium _{sulfide (Na2SxH2O,}

x7—9, 0.58

g,

—2.4 _mmole) and _{(y-mercaptopropyl) trimethoxysilane (1.41}

_g,

_{7.2 mmole)}

in the

_{same solvent system is} added into _vigorously stirred solution

of

cadmium _{acetate. After being vigorously} stirred

for

18 hours,

the

mixture is basified

to

pH=8.4 with 25% ofNH3 aqueous solution. Additional 64 ml

of

ethanol

is

added

to

the

mixture. The mixture

is stirred

for

48 hours after _{adding 1.89}

_g

of

_{orthotetraethoxysilane (TEOS). Part}

of

the solvent was removed and

precipitation takes place

in

the mixture The _precipitate

is

_centrifuged

for

three times and rmsed with deionized water

2 n

E 0

.1o

'op two surface layers

I

,Top one surface layer

U _{20 40 60 80 100}

The prepared nanoparticle aggregate

is

redispersable

in

ethani

or

other polar organic solvents. After treated by ultrasonic vibration and _percolation, solutions for _spin-coating purpose

are

produced.

To

take

TEM _{image, prepared solution is dipped onto carbon film} coated _{copper plate} and reabsorbed. Sparsely

distributed _{nanoparticles}

are

obtained. The _average diameter

of

the _spherical CdS _{nanoparticles}

is

about 5 nm,

as

shown

in

_Fig.

2.

_{Compared with}

the

_{high-temperature synthesizing method}

_by

_{using trioctylphosphine} oxygen (TOPO), this room-temperature process

is

easier but

the

particle size distribution

is

wider.

By

replacing

part

of

cadmium acetate with _{manganese acetate, we prepared}

Mn

_{doped CdS nano-} particles with

different concentrations

of

_manganese (5%, 10% _{and 20%,}

in

molar _{percentage). Through}

out the

_{experiment, no} significant difference

of

_{doped and undoped CdS nanoparticles}

is

_{found under the spectral resolution}

of

our

monochromator. To synthesize manganese doped CdS nanoparticles, we have used large molar percentage

of

Mn,

5

%, 10

% and 20

_{% respectively}

in the

reaction. However, only trace doping amounts

of

Mn, 0.08%, 0.05% and 1 .10% was

detected _respectively

_by

ICP-Mass _{investigations.} When

the

same _approach was used

to

synthesize MnS nanoparticles,

a

very low yield was obtained. This indicates very low doping content can

be

made and electroluminescence

is

less

affected

_by

Mn content.

4.

SETUP

FOR

ELECTROLUMINESCENCE

The basic idea about setup

of

electroluminescence

is to

use cascade tunneling as injection source. Carriers are supplied

by

tunneling current through potential barrier

of

high bandgap material surrounding nanoparticle,

as

shown in Fig. 3

. In

our case,

the

SiO group

or

organic functional group p-hydroxyl thiophenol serves

as the

goal. The choice

of

substrate

to

_{host CdS nanoparticles}

is

silicon due

to

its

surface _{quality and availability.}

In

addition,

in

_{this early study}

of

light emission, difficulty

to

maintain uniformity

of

film exists

if

nanoparticle solution

of

aqueous solvent system is used.

The Schottky barrier

of

metal-silicon contact has certain prevention from short circuit other than indium-tin-oxide (ITO) glass.

With different type

of

injection carriers, depending

on

n-type

or

p-type silicon substrate, electrons

or

holes are

emitted from silicon and holes

or

electrons are emitted from metal _deposited. Fermi level

of

Si

in

n-type case has

to

be raised

for

electrons

to

_{tunnel through potential barrier}

of

_{p-hydroxyl thiophenol group. Most} carriers tunnel into the

adjacent nanoparticles. Carriers are expected

to

recombine within nanoparticles through intrinsic radiative transition or

surface-state related transition.

Fig. 3 : Schematic ofelectron transport and transition in the device.

A

schematic

of

_{CdS-nanoparticle light emitting stage}

is

shown

in

_Fig.

4.

The fabrication _steps

are as

_{follows. First,}

a

doped silicon wafer (doping

lO15

cm3)

is

used

as

substrate. Acetone, methanol, and

DI

water are used _successively for clean _procedure. Buffered oxide etch

is

_applied

to

remove native oxides. The wafer

is

_{spin-coated with} CdS nanoparticle solutions. Solvents are either removed

by

evacuation

or

heat treatment. The thickness

of

CdS _nanoparticle layer can

be as

large

as

500 nm, verified

by

surface _profile scan.

In

this case, volume density can

be

very high.

Subsequently, both top and bottom metal contacts

are

made

by

thermal _{evaporation. The top semi-transparent} contact layer is lOnm gold, and

the

bottom

is

_{l5Onm gold.} Before

the

_deposition

of

the Au

_layer,

a

3-nm adhesion layer

of

chromium

had

_{been evaporated for both} contacts.

A

CM1 10 monochromator and _{photomultiplier}

is

used

to

record the

electroluminescence spectrum. In every spectrum measurement, the entrance

slit

width 0.6 mm

is

used for maximum detection

and

correct _spectrum.

Samples made

on n- and

p-type Si wafers show different current-voltage curves,

as

shown

in

_Fig.

5.

Both have

rectifying current-voltage (I-V) curves,

but

with opposite polarities. This rectifying effect corresponds to metal-insulator-semiconductor _{tunneling effect as expected.} To

be

_specific,

the

_{thin potential barrier}

of

organic

functional _{group and low substrate doping level} results

in

Schottky-diode-like behavior.

Fig. 5: I-V curves ofdevices on n-type and p-type Si.

5.

PROPERTIES

OF

ELECTROLUMINESCENCE

We take different post treatments after the _nanoparticle film

is

_{spin-coated onto silicon} substrate. The normal treatment

is

to

remove

the

_{solvent without physical change}

of

nanoparticles. Heat treatment

is

carried

out in

medium

high temperature

to

explore decomposability

of

organic passivation and

test

oxygen cooperation. Oxygen enrichment

raises extent

of

oxygen related surface states. All three conditions _{show prominent spectral features and}

can be

used

to

monitor passivation

of

surface condition.

Origins

of

electroluminescence

of

CdS nanoparticle with bulk CdS are schematically shown

in

Fig. 6. Three

exciton levels _{corresponding}

to

_{bulk CdS bandgap now change their peak positions}

of

quantum states due to modification

of

_{surface configuration. Modified A free exciton level changes from 508nm}

to

526.5nm as

can

be determined

in

normal treatment

of

electroluminescence device. Transition level

of

surface states related

to

Cd-O

termination with peak position

of

571 .Snm can

be

observed. _{This level occurs when passivation}

of

_nanoparticle is removed and

it

has contact with _atmosphere,

as

in

heat

_{treatment and oxygen} enrichment.

with

_capping

E

526.5nni 571 .5nm 508nm ______

E

A' Excitonn level

Fig.

6

Energy diagram

of

CdS nanoparticle.

Fig. 4: Schematic

of

the CdS nanoparticles EL device on Si wafer.

voltage(V) Bulk CdS

E

_________ 'I, EVA X

EB

EB

EC

EC

'Jr Surface states

3.1 _{Normal process}

After _spin-coated,

the

device

is

_placed

in

a

chamber with evacuation _{at room temperature for}

5

minutes

to

remove

ethanol solvents. Both spectra

of

CdS and CdS doped with Mn

are the

same,

as

shown

in

_Fig.

7.

The _spectrum fits into Lorentzian shape with scattering time

of

6

fs. Its FWIIM

is 42

nm. Such broad _spectrum indicates

the

_dispersion

of

particle size and

a

trade-off

of

low temperature synthesis.

This _{spectral peak indicates radiative} recombination

of

free exciton

in

CdS _{nanoparticles} with red-shift due

to

p-hydroxyl thiophenol groups. The green spectral peak

is

at 526.5nm (2.355eV), different from bulk CdS A-exciton

transition _{energy, 2.441eV (508nm)} at room _{temperature. Although quantum confinement} within

the

_{nanoparticles} increases _{exciton energy whenever}

the

_{particle size decreases4'5, organic functional group}

or

silicon dioxide matrix can

modify

the

electron configuration within and change ground states significantly.

It

results

in

energy red-shift

of

86 meV. CdS nanoparticles coated with poly(vinyl alcohol) also show such energy shift

in

absorption spectrum6, where photoluminescence at 2.42eV (10K)

is

observed.

Fig. 7: EL Spectrum

of

CdS with solvent removed. 3.2.

Heat

treatment

The CdS _{nanoparticles}

are

_spin-coated

as

described

in

_previous section. These _samples

are

_{subsequently treated}

by

rapid thermal annealing (RTA) with temperature 425°C for

5

minutes. The annealing process takes place

in

nitrogen

purge and its purpose

is to

remove solvent and both

test

_{decomposition}

of

organic functional group. Electrical property

like I-V curve resembles that

in

Sec. 3.1.

As

shown

in

_Fig.

8 the

emission _spectrum consists

of

two _peaks. One

is

at

513 .7nm and

the

other at 571.Snm. _{The former peak} stands

for

bulk CdS free

A

exciton transition. This _spectral lobe can

be

fitted

_by

Lorentzian _{shape with scattering} time

_of

8

fs and FWHM 40 nm.

1.2 1.0

>

Cl) C

.

_C 0.8 ci) N a) E 0

z

0.6 0.4 0.2 450 500 550 \veIength (nm) 600

Fig. 8: The EL Spectrum of _{CdS particles after heat} treatment.

500 550

wavelength (nm)

The 57 1 _{.5nm peak results from}

the

_{trapped carriers}

in

surface states related

to

_{oxygen, Cd-O termination7.} In

medium _{high temperature treatment, decomposition}

of

_{p-hydroxyl thiophenol group causes oxygen termination} with

cadmium

to

occur.

It

_{proves p-hydroxyl thiophenol group}

to be

effective overcoat

of

CdS _{nanoparticles against oxygen}

influence. The surface states induce radiative transition

as

well. The _{peak magnitude at}

the

spectral lobe

is

smaller than

the magnitude

at

5

13

.7

nm, indicating emission from surface states

is

weaker than that resulted from CdS _{nanoparticles.}

However, light power from this sample

is

generally stronger than that

in

Sec.

3 .

1

.

This _phenomenon

is

due

to

the

participation

of

surface level luminescence, leading

to

increase

of

total light output.

3.3 Effect

of

surrounding oxides

To further study the surface states related

to

_oxygen, we immersed CdS _{nanoparticles} into high oxygen content environment. Two ways

are

proceeded. First,

the

nanoparticle solutions were mixed with spin-on-glass (50G Filmtronics

3

_{15FX), and the second way,} mixed with 5102 nanoparticles (average diameter

of

12 nm, dissolved in isopropanol). The cleaned, oxide-free silicon substrate

is

spin-coated and treated

by

425°C

to

_{sinter with 5i02 glass.} The

similar EL _spectrum

is

found

in

mixture

of

CdS _{nanoparticles} with SOG and 5i02 nanoparticles. The peak

at

513.7nm (2.414eV) resembles

A

free exciton signal

of

bulk CdS at temperature 65°C, and

the

peak

at

57l.5nm (2.414eV) corresponds

to

radiative transition due

to

surface states. The magnitude

of

total light emission

in

current setups

is

ten

times stronger than that

of

unheated _samples

in

Section

3 .

1

,

under _{the same carrier inj ection} condition.

Co Cl)

ci

Fig. 9: EL spectrum ofCdS nanoparticles with oxygen enrichment

EL

_{spectrum shown}

in

_{Fig.9 indicates two} mechanisms

as

shown

_{by the}

_{energy diagram}

in

Fig.

6.

First, the

coalescence

of

CdS nanoparticles into bulk form results

in

_{less broadening spectrum} around

5

13 .7 nm. Since the

potential barrier

of

p-hydroxyl thiophenol group disappears due

to

decomposition, carriers

in

bulk powders stay for

enough time (about

ins

transition _lifetime)

to

recombine _{radiatively between each tunneling process. Second, relative}

magnitude

of

surface states luminescence

is

much _stronger than that

in

Sec. _{3.2. Highly increased} concentration

of

surface state _{levels, which are supplied}

_by

_{surrounding oxygen termination, contributes}

to

the enhancement

of

internal

quantum efficiency. The magnitude difference between mixture with SOG

and

Si02 nanoparticles comes from excess

dangling Si-O bond

of

the

latter case. With the _{same sintering time,}

the

latter mixture makes more extent

of

oxygen enrichment.

3.4 Temperature effect

To examine _{EL property}

of

CdS _nanoparticle with varied _{temperature, original organic functional group is}

substituted with _inorganic silica _passivation shell,

in

order

to

prevent instability

of

organic composition at low temperature.

EL

spectra

of

silica-passivated CdS nanoparticles are shown

in

Fig. 10. The resembling peak

of

surface

states _{represents Cd-O} termination, now introduced

_by

_{silica passivation}

at

surrounding surface probes

to

origin

of

light emission.

As

_{temperature increases,} the

EL

_intensity decreases. _{Such reduced emission efficiency comes from} increased

nonradiative mechanism with _{increased temperature.} Surface states

and

_{carrier-phonon scattering both play roles.} Ten

times _{stronger magnitude}

of

_{light emission intensity}

at

low _temperature

of

15

K

_compared with room _{temperature case} reveals that nonradiative mechanism

is

a

_{crucial factor influencing} emission _efficiency

of

the emitter. The reason

for

the

similar spectra is due

to

two causes. _{First, the shift}

of

_{quantized energy} levels due

to

small _{nanoparticles (<5nm)}

contributes

to the

_spectrum around 520nm. Second, transition

of

surface _{states may have multiple levels corresponding}

to

broad spectrum at room temperature.

C >s C,) cD U) Co ci 700

Fig. 10: EL spectrum

of

CdS nanoparticles with silica shell at varied temperature.

Spectral peak at 520nm

in

Fig. 10

is

attributed

to the

same origin

as

peak

at

526.5nm

in

organic-capped CdS nanoparticles, with additional effect

of

quantum confinement and surface configuration.

As

a

result, peak shift with

varied temperature

is

_compared with bulk _bandgap shift

in

_Fig. 11 and ascribed

as

characteristics

of

_{as-synthesized} CdS

nanoparticle. Ten times increase

of

emission intensity

is

observed

at

cryogenic temperature (15K), indicating strong

influence

of

surface _traps due

to

the

large surface area. Reduced thermal _scattering also contributes

to

enhancement

of

light emission

at

low temperature. The relative magnitude

of

recombination rate

of

surface states

to

exciton level

changes slightly and may result from

its

nearness

to

conduction band

2.5

2.4

0 100 200 300

Terrperature (K

Fig. 11: Spectral peak variation with temperature compared with bulk exciton energy.

With energy parameters

of

bulk CdS given above6, red-shift due

to

p-hydroxyl thiophenol group

is

determined as 86meV. Also, transition level

of

surface states

is

found

to be

273 meV below bulk _bandgap.

In

addition

to

observation

of

such peak

at

571 .Snm, similar phenomena

at

_{spectral range,} SSOnm—600nm, as

an

indication

of

imperfect CdS crystal

or

nanoparticles

had

been reported elsewhere. Okamoto8 also ascribed

their

broad peak

at

6SOnm (1 .9eV)

to

surface

state emission. _Hong7 also demonstrated _{broad peak}

of

592nm _(2.0944eV) due

to

_S-vacancy

_(in

this case termination

of

oxygen).

3.5

Raman

_spectrum

Previously observed transition

of

manganese ion

in

CdS nanoparticle at S8Snm (2. 1 _19eV)9

is not

_clearly observed

in

our _{samples, mainly due}

to

insufficient spectral resolution and trace content

of

Mn

_incorporated

to

CdS core. To

550 600

v.eveIength(nm)

•

Free Aexaton (exp.)

•

Free A exaton in

caJ4

determine

the

factor

on

firm _ground, Raman _spectroscopy

is

taken as shown

in

_Fig. 12.

Raman shift

at

305cm1 corresponds

to

the

_{longitudinal optical phonons LO) mode}

of

bulk CdS'°.

No

wave number shift

is

observed

for

_{CdSIMn nanoparticles,} therefore Mn

is not in the

CdS core lattice. The other Raman shift

peaks

are

from

the

vibrations

of

the _{organic molecules. Consequently,}

the

_{EL peaks}

of

CdS _{nanoparticles prepared} with

or

without addition

ofMn

are

the

same

as

shown

in

_Figure 7.

We have mentioned reaction rate

of

MnS synthesis

is

slower than that

of

CdS synthesis. Therefore, the

position

of

Cd

in the

lattice

of

CdS can barely

be

replaced by Mn.

It

is

trapped

by the

hydroxyl group

of

p-hydroxyl thiophenol capping on

the

surface

of

CdS _{nanoparticles.}

5. CONCLUSIONS

Significant surface influence

is

observed for _{nanoparticles. Chemical preparation}

of

CdS _{nanoparticles} ready for

spin-coating and LEDs made

of

CdS and CdS:Mn nanoparticles on Si substrates

are

described

in

detail.

EL

properties

are

_{investigated. Spectral} shift

of

free exciton transition due

to

passivation

of

p-hydroxyl thiophenol group around nanoparticles

is

discovered. Process modifications such

as

heat treatment and oxygen enrichment

are

influential

to

intrinsic _{green emission. P-hydroxyl thiophenol group}

is

shown

to

have _protection from diffusion

of

contaminants into nanoparticles,

but

cannot resist temperature deterioration above 400°C.

Radiative recombination

of

_{carriers trapped}

in

surface states _{present and magnifies}

itself as

_{long as extent}

of

surface states increases. Ten times increase

of

emission _intensity

is

observed

at

cryogenic temperature (1 5K), indicating

strong influence

of

surface _{traps due}

to the

_{large surface area. Reduced thermal scattering} also contributes to

enhancement

of

light emission

at

low temperature.

At

varied temperature,

the EL

spectrum

of

CdS nanoparticle remains

quite

the

same. Peak shift

is

compared with bulk bandgap shift and ascribed as effect

of

quantum confinement and surface configuration.

ACKNOWLEDGEMENTS

The authors acknowledge

the

_support from National Science Council, ROC under

the

contract number NSC 90-2215-E-002 and NSC 90-2622-L002.

REFERENCES

1. V.1. Klimov, A.A. _{Milkhailovsky, Su Xu,} A. _{Malko, J.A. Hollingsworth,} C.A. _{Leatherdale, H.-J. Eisler,}

and

M.G. Bawendi, "Optical Gain

and

Stimulated Emission

in

_{Nanocrystal Quantum Dots}

,"

Science _{290, pp.3 14-3 17, 2000.}

2.

J.

Butty, Y.Z. Hu, N. Peyghambarian, Y.H. Kao, and J.D.Mackenzie, "Quasicontinuous gain

in

sol-gel derived CdS

quantum dots

,"

App!. Phys. Lett. 67, pp.2672-2674, 1995.

3 .

J. G. C. Veinot,

M.

Ginzburg, and W.

J.

Pietro, "Surface Functionalization

of

_{Cadmium Sulfide Quantum-Confined} Nanoclusters.

3.

Formation and Derivatives

of

a

Surface _{Phenolic Quantum} Dot ,"Chem. Mater.

9,

pp.2117-2122, 1997.

4.

B.G. _Potter, Jr. and J.H. _{Simmons, "Quantum} size effects

in

_{optical properties}

of

CdS-glass composites," Phys. Rev. B

6661

'6-I

dd

'Z9Z

1

'SJOOflpUOO!WOS OU!jjt2Si(IOOUU SpDJO 1doosoipods uuw>j,,

'ioqo

MS

pu

qo

cs

'nqs

NS

'us

MS

'puN

rioi

sc

T

UON

•i4j

sp!1os1

'gjq,jq

'99-Lc9dd

661

Jo sUJodoJd

pu

sisoquX

ssooid

p-jos

U! SJSfljO Jo2onpuoolluos,, H

pui

'oid.iy

j

'joqudS

T

6 866T '1I-Ldd 'SOT UflWWOJ 31flJp//Og UOIW!OXO OAUOIOS Xq sj sj(jzougu

Sp

s

'1piu1A

oouoosouiwnjooq,,

wag

poddio-ooijns

pui 'iqso)pJnJAJ

,j

'1M1)joso]j H

'nswu,I

A

'o3owjO

s

8 000Z '9Z-6Udd '81Z tI10JD

71'3

T

'pOqw

uoU1Wjqns

/q

UMOJ

spjj°

soiodoid

jiodo

'uiqg TA

U1

'UOOA

f3

'UOOf 'UOH I TX L OOO 't'117-OWdd 'SIz/Iz

'1°D

l"'J

'1)jUJ

1 '!O

A

'OOWflS1JAI soiiodoid podopunjo

pui

podopuys4

pj

sps(ioouiu

U!

'wXjod

T

TAT 9 1661 'ZZE8-6O8dd 'S6

/d

1!1D T 'SJOSflO JOonpuOo!woS U S1OOJJO UOUIOUTJUOO 'JOUSOTJd

)J

pui

'UtjSU)J

1U1I

A1AT c 8861

'clioT-8Esordd

'