THIN FILMS ON GLASS

Founded 1996

D.G. Ast
N.W. Ashcroft

History
IRG 2 was formed in 1996
Members from Physics, Appl&Eng Physics,
Chem Engineering, Theoretical&Applied
Mechanics, Mat Sci&Eng.
IRG Leaders: D. Ast, N. Ashcroft

D.Ast

Current, Future Applications

Active Flat Panel Displays

Monolithic Fiber Optic Switching Network

Body sized ,digital, X-ray sensor

Large, Biological Sensor Arrays

Basic Issues: Adhesion, Inter-diffusion, Limits of Patterning

Our Goal
To advance the understanding of the
properties of thin film on glass
Surface Characterization
Thin Film Deposition
Initial Atomic Configuration at Interface
Time evolvement of Interface Configuration

Cornell Members
N. Ashcroft, 2 GRAs, Physics (Theory)
D. Ast, 1GRA, MS&E (Thin Film Electronics)
J. Blakely, 1 S. Res. Assoc., MS&E (Surface Science)
R. Dieckmann, 1GRA, MS&E (Diffusion)
H. Hui , GRA, T&AM
S. Baker, GRA, MS&E

} (Adhesion)

J. Silcox, 1 Res. Assoc., A&EP, (Interface Analysis)

Industrial Members *
F. Fehlner (Corning Inc.)
G. Couillard (Corning Inc.)

Industrial Connections
Corning Inc. (Glass-Ceramics, TFTs, MEMs)
Eastman Kodak (Active Matrix OLEDs)
Xerox Rochester (Analysis: TOF-SIMS)

University Connections
Penn State (TFTs), Alfred U (Theory), TU of
Clausthal (Experiment/Theory)
* Those attending bi-weekly IRG 2 meetings

Interactions
Surfaces: Blakely, Umbach, Cooper, Headrick, Fehlner, Couillard
Surface modifications: Blakely, Umbach, Headrick, Ast, Nemchuk,
Fehlner, Couillard.
Thin film deposition: Engstrom, Ast, Couillard, Fehlner, Baker.
Interface Characterization: Silcox, Jiang, Neaton, Ashcroft
Interface Adhesion: Ashcroft, Neaton, Baker, Hui
Interface Diffusion: Dieckmann, Thompson, Hui, Fehlner
Bulk Diffusion: Dieckmann, Hui, Ashcroft
Thin Film Electronics: Ast, Nemchuk, Couillard (Corning), Kalal
(Corning), Fehlner (Corning), Thompson; Malliaras, Ast, Tang (EK),
Williams (EK)

Outline

Research Strategy
Areas of concentration
Accomplishments
Future research plans
Summary

Research Strategy
Concentrate on surface/thin film interaction
Study bulk to the extend required to understand
surface and near surface phenomena.
Investigate both simple and complex glasses.
Couple, wherever possible, theory to experiment.
Develop new instrumentation and techniques to
characterize glass-surfaces and thin films on glass

Surface, Near Surface and Bulk

Surface
Near Surface
Bulk

Cleaning or polishing glass, required in most experiments,

introduces near surface modifications.
Near surface layer has properties different from bulk

Glass Systems Selected by IRG 2

SiO2 (Oxidized Si, fused silica, fused quartz) OUR CONTROL
extremely stable
transport properties sensitive to trace amounts of H2O
lacks glass like intermediate range order, high viscosity
K2O SiO2 THEORY + UHV CLEAVE EXPERIMENTS
simple, binary composition
not stable, reacts with moisture
Corning 1737 METAL & S.C. FILMS
Uniquely stable, reproducible, surface
complex composition

Structure of Oxide Glass

Multi-oxides have meso-scale short range order lowering viscosity

Simple versus Complex Glasses:

1737

K2O x SiO2

SiO2

Smooth, stable Surface

Hydrophilic Surface

High viscosity

Multi-oxide glass

Binary-oxide glass

Single molecule

Technologically useful glasses are complex mixture adjusted

empirically to exhibit both a low viscosity and chemical stability

IRG 2 Accomplishments

Surface Modifications of Glass by Cleaning *

Mass transport mechanisms on glass surfaces
X-ray scattering from glass surfaces (CHESS)
Properties of low temperature oxide films.
Na-22 diffusion in quartz, silica, 1737 glass.

Of large interest to the glass community

IRG 2 Accomplishments (II)

Na affinity of 1737, OA-2 and NA-35 glass
Properties of poly-Si films deposited on glass,
silica, oxidized Si, LTO, SiNx.
Structure of the Cr/Glass Interface
Substrate - Thin Si Film Interaction
STM imaging and spectroscopy of glass surfaces

Examples of IRG Research

The Cr/Glass Interface
Substrate - Thin Film interactions
STM imaging of Glass Surfaces

CHROME ON GLASS

Chrome on Glass
Cr is known to adhere well to glass
Cr conductivity is low, 12.9 -cm
Al, when sintered at 400 C adheres well to SiO2
Al conductivity 2.7 -cm
Au conductivity 2.35 -cm
Cu conductivity 1.7 -cm
Ag conductivity 1.6 -cm
Cr is frequently used to anchor better conductors, such as Cu
Low line resistance crucial in large displays (gate pulse distortion)

Empiricism of Adhesion
High Heat of Metal Oxide Formation
Ti3O5 = - 586 Kcal/mole TiO2

Good Adhesion
= - 228 Kcal/mol

Al2O3 = - 410 Kcal/mol

Al2O = - 14 Kcal/mole

Cr2O3 = - 276 Kcal/mol

CrO2 = - 142 Kcal/mole

SiO2 = - 210 Kcal/mol

SiO

= - 21 Kcal/mole

Cu 2O = - 40 Kcal/mole CuO = - 37 Kcal/mole

Ag2O = - 8 Kcal/mole

Ag2O2 = - 7 Kcal/mole

Low Heat of Oxide Formation

Poor Adhesion

The Cr-Glass Interface

Oxygen Plasma cleaned
1737 glass
Cr evaporated at room
temperature
Analyzed by STEM
Sample 1 observed 48
hrs. after deposition;
Sample 2 observed 2
weeks after deposition.

Initial
Interface?

Cr film
Glass Substrate
Sample 1
Interdiffusion?

Cr film
Glass Substrate
Sample 2

EELS and Bright

Field Detectors

Electron Energy
Loss Spectrometer

Annular Dark
Field detector

100 keV Incident Beam

DE = 0.7eV
Spatial Drift < 0.3
nm/min
Energy Drift < 0.03
eV/min

Windowless
EDX (X-ray)
Detector

1 atom wide (2.1 ) beam is

scanned across the sample in
the x-y plane to form a 2-D image

Cornells UHV
STEM acquires
both positional,
chemical (EDX)
and bonding
information
(EELS) with
atomic
resolution

Cr/1737 Interface Kinetics

Annular dark field (ADF) image of
Sample 1. Interface is broken up. No
signs of inter-diffusion

TEM bright field image of Sample 2.

A well defined, 5nm wide, inter-diffusion
layer can be seen.

Probing the Inter-Diffusion Layer

Probing beam
located in
Cr,
interface,
1737 glass

Probing the Inter-Diffusion Layer

Diffusion layer

Silicate glass

Silicate glass

Diffusion layer

CrO x

CrOx

O
Cr
Si
~ 5 nm

10

12

14

Distance (arb. unit.)

Annular Dark Field

(ADF) image of
evaporated Cr film on
glass.

Relative composition changes across the diffusion

layer. All data are integrated from EELS spectra.

Probing the Inter-diffusion Layer

Si L-edge of a-SiO

Intensity (arb. unit.)

Cr L -edge

23

Intensity (arb. unit.)

c
b
a
Ba N-edge
90

100

O K-edge due to Cr-O

a
b
c
d
e
O K-edge due to Si-O

110

120

130

Energy-loss (eV)

140

150

528 536 544 552 560 568 576 584

Energy-loss (eV)

EELS spectra of Cr-L, Si-L and O-K edge across interface

Probing the Inter-diffusion Layer

Ba M -edge

Ca L -edge

45

M
5

~ 3eV
L L
3

Arb. Intensity

interface region

near interface

interface region
Arb. Intensity

23

near interface

in bulk glass

760

770

780

790

800

810

820

Energy-loss (eV)

Ba segregates to
Interface

830

in bulk glass
840

320

330

340

350

360

370

380

Energy-loss (eV)

Ca segregates to
Interface

390

400

Proposed Mechanism
Cr2+

Cr film

interface
glass unit
Y3+
O2X2+
Si4+

Y3+ = Al3+ or B3+

X2+ = Mg2+, Ca2+, Sr 2+ or Ba2+

Cr displaces
alkaline earth metal,
which moves to
interface. Atom
arrangements in
glass deduced from
ELNES.

Probing the Bonding State of Cr

Oxidized Cr layer

Intensity (counts)

5000

Cr diffusion layer

3000

Cr L -edge

4000

Cr L -edge
2

2000

3000
2000

Cr L3 and L2 absorption
peaks are due to the
excitations of the 2p
electrons (2p3/2 and 2p1/2)
to unoccupied bound 3d
states.

1000
1000
0

energy shift ~ 0.6eV

570

575

580

585

Energy-loss (eV)

590

Cr L3,, L2 edge position

shifts to higher energies
in the inter-diffusion
layer.

For the interpretation of this finding visit our poster !

Probing the State of Oxygen

Normalized O K-edge

-1nm
1nm
3nm
5nm
bulk glass

Arb. Intensity

b
a

a Oxygen K-edge due to Cr

b Oxygen K-edge due to Si
Cr induced states

525

530

535

540

Energy-loss (eV)

545

550

Analyzing the State of Oxygen

SiO2 O K-edge (tetrahedrally)
coordinated Si) strongly
resembles 1737s O K-edge
Hybridized Cr d-O p antibonding states (octahedrally
coordinated Cr)
Corner-sharing tetrahedra
having alternately Cr and Si at
the centers
DFT-LDA calculations of of Op-projected
DOS of SiO2, Cr2SiO4, and Cr2O3 by
Neaton and Ashcroft

Substrate Thin Film Interactions

Thin Film Electronic on Glass

Poly-Si directly contacts glass - high potential for interaction

Advanced Thin Film Electronics

Ast, Lam, CNF

Active Matrix: two 32x32 displays

Tang, Fleming, Williams

Finished OLED Display

Preparing Glass for Si Deposition

All glass used for thin
film deposition is
cleaned prior to use
Physics of cleaning is
very little understood
The most demanding
applications use the
RCA Clean to clean
glass

RCA cleaning process:

Base solution of
NH4OH:H2O2:H2O at
75 C [removes organics]
H2O rinse to 6 Mcm
Acid solution of
HCl:H2O2:H2O at 75 C
[removes metals &
particles]
H2O rinse to 16 Mcm

W. Kern and D. A. Poutinen, RCA Rev. 31 187 (1970).

RCA Clean: Better TFTs - Why ?

Source-Drain Current (Amp)

10-5
10-6
10-7

10 min. RCA clean

120 min. RCA clean
10 min. RCA + HF dip
Oxidized Si control

10-8
10-9
-10

10
10-11
10-12

-5

5
Gate Voltage (V)

10

15

TFT has lowest leakage current on 120 min RCA cleaned glass
TFT has highest leakage current on oxidized Si control wafer
Ast, Couillard (CNF)

X-ray Reflectivity of RCA

Cleaned Glass
100
10-1
10-2
10

Data

Modulation at 0.1 -1
is due to interference
from reduced density
surface layer.

Fit to 3-layer model

Surface layer:
60 5 thick
78 3% bulk density

-3

10-4

Surface roughness
unchanged

10-5
10-6
10-7

X-ray Reflectivity: 60 min. RCA Clean

10-8
10-9
0

0.1

0.2

0.3

0.4

0.5

0.6

-1

q ( )
Z

Blakely, Umbach, Headrick CHESS

STEM Image of Cleaned Glass

Silcox, Jiang, STEM Facility

Reduced electron density near surface layer, 60 thick

Properties of Near Surface Layer

X-ray electron density, thickness
(Blakely, Umbach, Headrick - CHESS)
STEM electron density, thickness, SiO2 rich
(Silcox, Jiang - STEM Facility)
TOF SIMS, XPS, ICP SiO2 , B, Al, Sr, Mg, Na, Ca, B
(Fehlner, Couillard, Ast, Xerox)
I-V of TFTs acts like deposited SiO2, must influence Na
transport (Ast, Couillard, Fehlner)

Novel method to evaluate Na - 22 diffusion

across a a very thin layer by analyzing a much
wider profile.

Initial state: Sample 1 contains a well

defined Na-22 distribution, sample 2 with
with a near surface layer (barrier ) is
tracer free.
Experiment: Sample 1 and 2 are clamped
together; annealed

Na-22 Profile

Diffusion of Na-22 tracer slows with increasing thickness of

surface layer from RCA clean

Analysis

Na-22 profiles
for different interfacial
resistances (calculated)

Residual radioactivity for

different resistances
(calculated, integrated
across profile)

Additional complication
To maintain electrical neutrality during the
diffusion of Na+, an other charged species
must move, such as H+ or OH-.

H2O - Na-22 interaction (I)

D*Na in fused quartz, OH- < 5 ppm, is greater than in 1737

D*Na in fused silica, OH- < 1000 ppm, is smaller than in 1737

Na diffusion into dry, wet glass

Residual radioactivity as a function
of depth in 1733 glass in which Na22 was diffused in wet air (80 C
bubbler)
Analysis
Different D values in near surface (ns)
and bulk (bu) regions
Thickness of near surface layer
increases with time as

Correlation with OH- by IR

OH- IR Absorption

Follows parabolic time dependence expected for diffusion controlled process. Estimated
average OH concentration in near surface layer is 4060 ppm, compared to bulk concentration of
330 ppm

Results
treatment of the surface of
Sample 2

log DNa /D
2

log ki/k

(D = 1 cm /s)

(k = 1 cm/s)

untreated

- 9.571 0.014

- 7.354 0.013

10 min RCA-cleaned

- 9.566 0.004

- 7.854 0.093

60 min RCA-cleaned

- 9.547 0.014

- 8.392 0.101

5.2 h annealed in wet air

- 9.564 0.003

- 8.462 0.037

11.4 h annealed in wet air

- 9.539 0.002

- 8.658 0.064

23.6 h annealed in wet air

- 9.572 0.005

- 8.906 0.085

50.1 h annealed in wet air

- 9.565 0.001

- 8.948 0.068

150 nm SiO2 deposited

- 9.559 0.006

- 9.029 0.046

500 nm SiO2 deposited

- 9.563 0.005

- 9.503 0.077

(experiments performed at T = 650 C)

D (650 C) Na, RCA Layer = 3.3 10-15 cm2/s

D (650 C) CVD SiO2 = 1.4 10-14 cm2/s

Resolving the Glass Surface

UHV STM of Si(111)
7x7 reconstruction
[imaged at Cornell]

50

JEOL JSPM 4500 (funded by CCMR)

First in United States

Atomic resolution STM
Atomic resolution non-contact AFM
Ultra-high vacuum
High temperature imaging (up to 500 C)
10 m scan range
Bolts on to existing analysis chamber

UHV Contact AFM

of sputter induced
ripples on SiO2
[imaged at Cornell]

Imaging Glass via STM

Imparting
Transient
Conductivity:
I) Electron beam
deposits charge
in trap states and
conduction band
II) Photoelectron is excited
across metal-glass
Schottky into
conduction band

Future Research Plans

Surface Science of binary glasses - UHV AFM/STM imaging
of cleaved surfaces, coupled to theory of binary glasses. (Major
Capital Investment)
Adhesion of metal films to glass: Quantitative measurements,
coupled with ab initio theory. (Major Personnel Investment)
Interactions of thin films with glass substrates (Materials
used by the thin film electronic industry)

(I) Surface Science

In situ cleaving of glass in the AFM/STM - statistics of
surface atom distribution, (Blakely) - calculation of
electronic states of surface atoms (Ashcroft).
In situ evaporation of sub-mono, mono- and multimono layer metal films in the AFM/STM - adhesion,
nucleation and initial growth of metal films on glass STEM spectroscopy (Blakely, Silcox, Ashcroft).

(II) Adhesion of Thin Films to Glass

Develop new fracture mechanical experimental
(Baker) and theoretical (Hui) methods to measure
the work of adhesion
Link these measurements to ab initio calculations
of the adhesion of metal films (Ashcroft)

Calculated Adhesion
SiO2

SiO2

Metal

Metal

Ashcroft, Neaton

Calculated Adhesion of Cu, Cr

Ashcroft and Neaton,

50 Atom Supercell

Subtracting Plastic Work

The calculated work of adhesion
W 1.6 J / m ,
2

ad

Film curls due to residue stresses

[ CuSiO2 O terminated ]

Is very small compared to fracture toughness

Active Plastic Zone

int erface

Reason:
Plastic Deformation
Correct Subtraction of
this term will be a
major challenge
Hui, Baker, Wang

(III) Thin Film-Glass Interactions

Lateral diffusion on glass surfaces (Dieckmann,
Ast, Couillard)
Impurity transport across diffusion barrier layers
(Ast, Couillard, Dieckmann, Silcox, Revesz)
Properties of electron and hole injection layers
deposited on glass substrates (Malliaras, Ast)

Long Range Future Plans

Our planned area of expansion is:
Fundamental limits of patterning glass
(surface relaxation, meso-scale order,
interaction with ion beams)
Wide variety of potential applications, from microfluidic to glass based biology chips

IRG2 Summary
A three year old IRG, concentrating understanding the
interactions of thin films with glass.
Major expansion into developing STM of insulators
and the quantitative, ab initio investigation of adhesion
of thin films to glass.
Long range interest in limits of patterning glass.
Interacting within Cornell, with other Universities and
with nearby industry.

Back-up Slides

Poster Guide (I)

Glass Substrate/Thin Film Electronics Interaction
(D. Ast, N. Nemchuk, G. Couillard, S. Krasula
Failure Mechanism of Thin Metal Films on Glass
Substrates (X. Wang, C-Y Hui, S. Baker)
Atomic level study of Cr/oxide glass interfaces
(N. Jiang and J. Silcox, J. Neaton and N. W.
Ashcroft)

Poster Guide (II)

Nanometer-Scale Structure and MicroMorphology Development of Glass Surfaces
(K. Umbach, J. Blakely, F.Fehlner, G.
Couillard, M. Onyiriuka, B. Cooper,
R.Headrick
Theory of IRG 2 (A. Koenig, J.B. Neaton, W.
Ashcroft)
PACKING AND EFFECTIVE FORCES IN GLASS-FORMING MIXTURES
AN AB INITIO STUDY OF THIN METAL FILMS ON GLASS

Poster Guide (III)

Na-22 Tracer Diffusion in 1737 Glass and
Exchange of Na-22 between 1737 Glass
Substrates ( L. Tian, C-Y Hui, J.G. Couillard,
M. Thompson, R. Dieckmann)

Glass Surface

Mass Transport on Surfaces

AFM image of 4m
grating as fabricated

AFM image of same

area after 2 hours
at 827 C in air

20 m

1500

Patterned with optical

lithography and HF
etch

Normalized Amplitude

Wavelength Dependence

6 m
4 m
3 m
1.0
5.0
Annealing time (hours)

Viscous flow dominates mass transport

on 1737 glass at m length scales

Microfabricated
gratings with
different wavelengths
on same substrate

OLED Pixel and TFT

OLED - TFT Interaction

1.E-03
24 TFTs with OLED

ID (amp)

1.E-04
1.E-05
1.E-06

1 TFT with OLED

1.E-07
1.E-08

1 TFT without OLED

1.E-09
-5

5
10
VG (volt)

15

20

TFT Lifetime

Lifetime of stressed TFTs fabricated on three different glasses

and oxidized Si, RCA cleaned and/or covered with CVC SiO2
barrier layers. Normalized to the lifetime on 100 nm covered
1737

Expansion: Glass-Ceramics

Formed like glass

Transparent like glass
CTE matches Si
High fracture toughness
960oC processing
Intriguing diffusion

TFT made on Glass-Ceramic

CNF

Fine Ceramics from glass

Established Production Methods

Flexible Physical Properties

Thin Film Electronic on LGA

Current Research Themes