STUDY OF DIAMOND LIKE CARBON AS TEMPLATE FOR NANOIMPRINT LITHOGRAPHY AND AS A FILLER MATERIAL FOR VERTICALLY ALIGNED CARBON NANOTUBE FORESTS
Seetharaman Ramachandran
Chair : Dr. Lawrence J. Overzet
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
OUTLINE
• • • • • •
Research Objectives Introduction to Diamond like Carbon Experimental Setup Diamond like Carbon and Nanoimprint Lithography PECVD of Carbon-Carbon Composites Summary
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Research Objectives
Diamond like Carbon Films as mold material for Imprint Lithography
a) b)
Establish a process for consistently growing Diamond like Carbon Films Study the change in relevant properties of DLC films for their use a mold material in NIL applications
c)
Demonstrate the applicability of DLC films as mold material for imprint lithography
PECVD Of Carbon-Carbon Composites
a)
Obtain a Carbon nanotube based composite material by subjecting it to a PECVD based process
b)
Understand the relationship between the plasma conditions and nature of infusion process
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Outline
• • • • • •
Research Objectives Introduction to Diamond like Carbon Experimental Apparatus Diamond like Carbon and Nanoimprint Lithography PECVD of Carbon-Carbon Composites Summary
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Diamond Like Carbon is a Synthetic Metastable Form of Carbon
First obtained as part of research efforts on diamond synthesis
Amorphous network consisting of various fractions of Hydrogen, SP2 and SP3 hybridized Carbon
Common synthesis techniques
–
– –
Pulsed Laser Deposition
Ion beam deposition Plasma Enhanced Chemical Vapor Deposition
–
And many other techniques
Ternary Phase Diagram1
Hydrocarbon plasma in the presence of energetic ion bombardment
1. J. Robertson, Materials Science and Engineering R 37 (2002), 129-281
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Energetic Ions, Radical and Atomic Species Created in the Plasma are Critical for DLC Deposition
Precursor In
Showerhead Electrode
1. J. Robertson, Materials Science and Engineering R 37 (2002), 129-281
Substrate
Hydrocarbon Plasma
Insulated rf feedthrough
~
Typical PECVD Reactor
Possible Growth Mechanism1
DC and RF plasmas including microwave, capacitive and inductive plasmas Typical PECVD reactor
–
Reactor geometry, nature of RF coupling results in availability of energetic ions, radical and atomic species
Auxiliary electrode to control the ion-energies
–
Deposition Mechanisms Subplantation model, Cylindrical Spike model, atomic peening model etc
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
DLC Films Have Numerous Potential Applications Due To Their Tunable Properties
DLC film properties are tunable between that of Diamond and Graphite
–
–
–
–
–
–
Protective Coatings Hardness and Wear Resistance Corrosion Barriers (Oxidation barrier) Becoming common for PET bottles Optical Windows Tunable Optical Gap Dielectric Material Controllable Dielectric Properties Biocompatible Coatings Protective Coatings for Artificial Body Parts Field Emission Displays
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
OUTLINE
• • • • • •
Research Objectives Introduction to Diamond like Carbon Experimental Setup Diamond like Carbon and Nanoimprint Lithography PECVD of Carbon-Carbon Composites Summary
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Capacitively Coupled Reactor Setup For DLC Deposition
Matching Network
Basic Reactor Setup Includes
– –
RF
Ar+CH4
1
4
–
– –
Upper showerhead electrode(1) Center grounded electrode(2) Lower biased electrode(3) Teflon/Ceramic Insulators(10) 0-600W RF generator
Ar Purge
5
2
(Ar+CH4) Plasma 7
6 3 Cooling Water
4
Modified to accommodate sample heating
–
–
RF
Matching Network
Pumping Stack
–
Infrared Lamp(5) Thermal Insulator(6) Sample(7)
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Modified Gaseous Electronic Conference Cell Reactor Setup For DLC Deposition
13.56 MHz Source + Matching
Trigger I/P
Trigger O/P
Induction Coil Gas Inlet
Quartz Window
Si Substrate
13.56 MHz Source + Matching
6” Water Cooled Aluminum Chuck To Pump
Modified Gaseous Electronic Conference Cell (MGEC)
– –
–
Inductive mode (ICP) and Capacitive mode (CCP) Flexibility of changing the gap (4-16cm) Coil power (0-1 KW) and Bias power (0-200 W) controlled by output voltage of function generators
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Gas Phase Diagnostics Include FTIR and OES`
Optical Access for Emission Spectroscopy
in-situ Multi-Pass Fourier Transform Infrared Spectrometer (FTIR)
–
–
4–40 pass white-cell arrangement Plasma sampled about 3 to 5 cm above biased electrode Multiple gratings for sampling different range of wavelengths Plasma sampled about 1/4 cm above biased electrode
Broadband Optical Emission Spectrometer (OES)
–
–
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
OUTLINE
• • • • • •
Research Objectives Introduction to Diamond like Carbon Experimental Setup Diamond like Carbon and Nanoimprint Lithography PECVD of Carbon-Carbon Composites Summary
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Technique Entails Pattern Transfer Through Direct Contact with Polymer Coated Substrate
Imprint Mold
Resist
Substrate
Regular NIL
SFIL
Transparent Mold
Resist
Substrate
UV exposure
Imprint Mold
Substrate
Pressure, Thermal Energy
hν
Transparent Mold
Substrate
Imprint Mold
Substrate
Substrate
Plasma Etching
Lithography technique that is a form of contact printing
–
–
Nanoimprint Lithography (NIL) Step and Flash Imprint Lithography (SFIL)
Spincoating, Pressure contacting, Heating and Demolding High throughput, high resolution, low cost
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Potentials of Imprint Lithography is Visible to the Semiconductor Industry
imprint
imprint
imprint
The 2005 ITRS roadmap lists nano imprint lithography as a possible technology in the 32 nm node
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Challenges Facing NIL Motivation For Using DLC as Mold Material
Key Challenges Facing Imprint Lithography Technique
• •
Wear damage to mold material Post-processing of mold surface to meet surface energy requirements
• •
–
– – – – –
Currently using template release coatings like F-SAMs
Mold material with suitable optical properties that could be used for SFIL
Why is DLC a suitable material for serving as the mold?
Excellent hardness and wear resistance 30-35 GPa Low energy surface 35-50mJ/m2
Controllable optical gap 1 to 3.5eV High chemical and corrosion resistance
Biocompatible
Can be deposited on both Quartz and Si
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Film Deposition is Performed Under Different Conditions That Change Plasma Properties
All experiments are performed in the reactor MGEC
– –
Vary coil power from 0-500W Vary bias power from 0 to 200W
–
– – –
Vary process pressure from 5mT to 65mT
(Arsccm/CH4sccm) from 22.5% to 87.5% Coil to chuck gap of 6 to 14 cm Duty Cycle [tON/(tON+tOFF)]*100
33% for ICP conditions 100% for CCP conditions
Film characterization
– –
Deposition rates using profilometry Structural properties using Raman spectroscopy
–
–
Optical properties using ellipsometry and n&k analyzer
Surface energy using goniometry
Understand reasons why the film properties are changing the way they do
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Methane Being Broken Up Gives Rise To The Rich Chemistry Required For Film Deposition
•
Incoming Methane molecules primarily lost due to
•
Dissociation and ionization to form species like CH 3, CH2, CH, C, H, CH4+ etc
d ( no )
•
dt
Qn V
o
kd ne no ki ne no k ( jl : m0)n j nl k (0 j : lm)n0 n j
jlm jlm
r
no
Radical and ionic species thus produced are lost to surfaces, including electrode on which sample is placed
•
d (nn ) n k d n e n o kin n e n o k (lm : nj )n l n m k (nj : lm)nn n j klossnn dt jlm jlm
Some critical reactions considered are
• • •
CH3+CH2C2H4+H CH2+CH2C2H2+H2
Behavior of C2H2 and C2H2 with process conditions offer pointers about CH 3 and CH2
•
•
CH4, C2H2 and C2H4 are quantified using FTIR
CH, C2 Dimer and H are primarily characterized using OES
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
FTIR Can Be Used To Quantify Certain Neutral Species Found In The Discharge
C2H2
•
• • •
Molecules vibrating at a certain frequency absorb incoming IR beam at same frequency
Total Absorbance found by adding peak intensities for all rotational (J) levels
Absorbance = (Density of gas) (Path length) (Absorption Cross section of molecule)
Rotational level assignments and Absorption Cross Sections obtained from HITRAN database for corresponding molecules
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
OES Can Be Used To Characterize Species Like CH And Atomic Hydrogen
• • •
Discharge characterization by scanning the emitted light from the plasma for characteristic emission lines Use Actinometry to calculate densities of certain species
Ix I actinometer
k
nx nactinometer
Constant k is independent of plasma parameters if actinometry conditions are satisfied
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Raman Spectroscopy is a Powerful Technique to Probe Structural Properties of Amorphous Forms of Carbon
1000
35
Actual Data Fitted Data
Less at. H% More at. H%
750
Intensity (X 1000) (a.u)
28
Intensity (a.u)
21
500
D-Peak
G-Peak
14
250
7
0 750
1000
1250
1500
1750
-1
2000
2250
0 300
600
900
1200
1500
1800
-1
2100
2400
Raman Shift (cm )
Raman Shift(cm )
514 nm excitation using a Jobin Yvon Labram microRaman spectroscope Collected spectra needs to be deconvoluted to obtain trends in change of film properties
–
–
Positions, widths and intensity ratios of “Graphitic (G)” and “Disordered (D)” peaks G-pk and D-pk moving towards higher wavenumbers sign of increasing Graphitization increasing SP2 fraction
In addition, sloping background of Raman spectra could be used for calculating bonded Hydrogen content
H (at %) log(
m ( m)) I (G )
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Optical Properties are Probed using an n&k analyzer and an Ellipsometer
n&k analyzer
–
–
Si Substrates Measures reflectance (R(E)exp) at normal incidence for λ=190 to 1100nm Calculates theoretical reflectance (R(E) calc) by varying n(E) and k(E) – R(E)calc= [((n(E)-1)2+k(E)2))/((n(E)+1)2+k(E)2))] – Forouhi-Bloomer relationship relates n(E) and k(E) to optical parameters Quartz Substrates Measure both reflectance and transmittance Calculate absorption coefficient [ref] (1-R2 e-2αt) T=(1-R)2 e-αt
Spectroscopic Ellipsometry
–
–
–
–
Light reflected from sample surface undergoes change in polarization Measured in terms of the parameters ψ and Δ, which can also be related to the material’s dielectric function Polarization change is related to dielectric function (ň=n+ik) of material under investigation Tauc-Lorentz relationship relates n and k to optical parameters
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Surface Energy Calculation is Done Using Two-Liquid Method
Goniometry
– –
Measure angle made by liquid droplet on sample surface Knowledge of contact angles for two different liquids (A and B) enables calculation of
total surface energy of sample surface
γ1 Liquid droplet
θ
γ2
Substrate
d ( 2 )1 / 2
γ12
( 1dA )1 / 2
1A
( 1p )1 / 2 A
1A
( 2p )1 / 2
1 cos A 2 1 cos B 2
( 1dB )1 / 2
1B
d ( 2 )1 / 2
( 1p )1 / 2 B
1B
( 2p )1 / 2
A and B liquids in use θA, θB Contact angles made γp, γd polar of dispersive components of surface energy γp+γd total surface energy (γ)
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
ICP Conditions Result in Higher Degree of Dissociation of Precursor Molecules
60 50 40 30 20 10 0 0 0.5 0.4 0.3 0.2 0.1 0.0 0 10 20 30 40 50 60 70 10 20 30 40 50 60 70
molecules/cm )
3
ICP vs CCP Conditions
–
Degree of Dissociation (%)
CCP
ICP
5.0
atoms/cm )
CCP
ICP
4.0 3.0 2.0 1.0 0.0 0 9.0 7.2 5.4 3.6 1.8 0.0 0 10 20 30 40 50 60 70 10 20 30 40 50 60 70
Degree of dissociation of precursor molecules much higher under ICP conditions Higher density of atomic H for ICP conditions C-C2H4> C-C2H2 but I-C2H4≈I-C2H2
3
[H] (X 10
–
14
–
Pressure (mtorr) ICP
Pressure (mTorr)
C-C2H4 I-C2H4
CCP
C-C2H2 I-C2H2
CH3>CH2 for CCP conditions
–
CH3 ≈ CH2 for ICP conditions
No significant dependence under ICP conditions
Ion Current to chuck
sb
Densities (X 10
b
12
Pressure (mTorr)
Pressure (mTorr)
Slight increase with pressure for CCP conditions
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Deposition Rates for ICP Samples are Much Higher Than CCP Samples
250
Dep Rate (nm/min)
200 150 100 50 0 20
Dektak-Si Dektak-Quartz n&k Ellipsometer
Higher deposition rates for ICP samples
–
More radical species available for deposition
10
20
30 40 50 ICP Pressure (mTorr)
60
70
Dep Rate (nm/min)
Dektak-Si Dektak-Quartz n&k Ellipsometer
–
15 10 5 0
Higher density of atomic H that creates active sites
Lower self-bias (-100V) compared to
–
CCP conditions (-400V) results in reduced sputtering effect
0 10 20 30 40 CCP Pressure (mTorr) 50 60
Deposition on Quartz substrates
–
Quartz substrates float at lower selfbias voltages
Baseline Conditions: ICP200W, 30sccm CH4, 10sccm Ar, 8cm Gap, -100Vsb CCP 0W, 30sccm CH4, 10sccm Ar, 8cm Gap, -400Vsb
–
Under ICP conditions, dep rate on
Quartz is lower than for Si
Shows importance of ion
bombardment in film growth
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Low Pressure Conditions Result in Lower SP2 Contents and Hence Better Mechanical Properties
1575 ICP CCP
1590
Si Substrate
1600
Quartz Substrate
G-Pk Position (cm )
1555
G-Pk Position (cm )
-1
-1
1580
1570
1535
1560 ICP CCP 0 10 20 30 40 50 60 70
1515
1550
0
10
20
30 40 50 Pressure (mTorr)
60
70
Pressure (mtorr)
Si Substrates
–
–
ICP Samples are less graphitic compared to CCP samples Higher atomic Hydrogen densities enhance SP 3 contents by preventing formation of bigger SP2 clusters But Hydrogen content reaches almost 40% at higher pressures indicating more polymeric films Low pressure inductive and capacitive plasmas yield better mechanical properties Both ICP and CCP samples are more graphitic compared to Si substrates Increasing pressure leads to decrease in average ion energies Almost no change in Raman Fit parameters
Quartz substrates
–
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
ICP samples Exhibit Higher Optical Gaps and Lower Surface Energies at All Pressures
3.0 2.5 ICP-FB CCP-FB ICP-TL CCP-TL 35 2.0
(X10 cm )
-1
Quartz Substrate
ICP CCP
Optical Gap (eV)
30 25 20 15 10 5 0 0 10 20 30 40 50 Pressure (mTorr) 60 70 ICP CCP 52 50 48 46 44 42 40 38 36 34 0 0
ICP
1.5 1.0 0.5 52 50 48 46 44 42 40 38 36 34 0 10 20 30 40 50 60 70 Pressure (mTorr)
4
10
CCP
20
30 40 Pressure (mTorr)
50
60
70
Energy (mJ/m )
Surface
Energy (mJ/m )
Surface
2
2
10
20
30
40
50
60
70
Pressure (mTorr)
Optical Bandgap and absorption coefficient
–
CCP samples exhibit lower optical bandgaps than ICP samples, except at the lowest pressure
Consistent with Raman results
–
CCP samples also exhibit smaller absorption coefficients at lower pressures
Surface Energies
– –
Lower pressures lead to higher surface energies in case of both CCP and ICP samples Ion bombardment and availability of atomic Hydrogen dominate this surface property
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Methane Dilution Has Almost No Impact On Important Gas Phase Properties
50 40 30 20 10 0 20 0.5 0.4 0.3 0.2 0.1 0.0 20 40 60 80 100 40 60 80 100
molecules/cm )
3
Degree of Dissociation (%)
atoms/cm )
, ,
CCP ICP
2.0 1.6 1.2 0.8 0.4 0.0 20 7.5 6.0 4.5
, ,
3
, ,
CCP ICP
ICP vs CCP Conditions
–
Degree of dissociation independent on methane dilution
[H](X10
14
CH4 Flow (%) ICP
40
CH4 Flow (%)
, ,
60
80
100
Increasing total flow does affect degree of dissociation
CCP
C-C2H2 I-C2H2
C-C2H4 I-C2H4
– –
Higher density of atomic H for ICP conditions C-C2H4>C-C2H2 and I-C2H4≈I-C2H2
sb
Densities (X 10
12
b
3.0 1.5 0.0 20 40 60 80 100
CH3>CH2 for CCP conditions
CH3 ≈ CH2 for ICP conditions
CH4 Flow %
CH4 Flow (%)
–
Ion current does not show any dependence on methane dilution or total flow
Baseline Conditions: ICP200W, 40mT, 8cm Gap, -100Vsb CCP 0W, 40mT, 8cm Gap, -400Vsb
CCP Flow
Ar/CH4
% CH4
ICP Flow
Ar/CH4
% CH4
40
40 40
30/10
25/15 10/30
25
37.5 75
40
40 40
25/15
15/25 10/30
37.5
62.5 75
40
80
5/35
10/70
87.5
87.5
40
80
5/35
20/60
87.5
75
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
C2 Dimer Species Could Be Playing a Role at Lower Methane Dilutions under ICP Conditions
1580 , ICP Si Substrate , CCP
1600
Quartz Substrate , , ICP CCP
1580
1560
G-pk Position (cm )
-1
1560
1540
1540
1520
1520 20
40
60 CH4 Flow (%)
80
100
1500 20
40
60 CH4 Flow (%)
80
100
Deposition rates do not show any dependence on flow ratios used
– –
ICP conditions result in higher deposition rates Dep rate on Quartz is lower than on Si
ICP Samples are less graphitic compared to CCP samples At reduced CH4 dilution (37.5%) vs (62.5%), films are more SP3 like C2 dimer species is influential under these circumstances
Si Substrates
–
Quartz substrates
–
–
Lower CH4 dilution is more favorable for higher SP 3 content (ICP Only) Higher ion bombardment available for CCP samples less Graphitic compared to ICP samples
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Methane Dilution Affects Surface Energy And Does Not Affect Optical Properties
2.50
16 12 8 4 0 20 50
Energy (mJ/m )
Optical Gap (eV)
2.00
(X10 cm )
Quartz Substrate
1.50
1.00
, , I-FB C-FB , , I-TL C-TL
4
-1
, , 30 40 50 60 CH4 Flow (%) 70 80 , ,
ICP CCP 90 ICP CCP
0.50 20 50
40
60 CH4 Flow (%)
80
100
45
Energy (mJ/m )
2
45
Surface
2
Surface
40
40 , , ICP CCP 40 60 CH4 Flow (%) 80 100
35 20
40
60 CH4 Flow (%)
80
100
35 20
Optical Bandgap and absorption coefficient
– –
Higher hydrogenation of ICP samples explains higher optical bandgaps CH4 dilution does not play a significant role in determining optical bandgaps for Si substrates
Surface Energies
–
–
Higher hydrogenation of ICP samples explains lower surface energies for both Quartz and Si Higher Methane dilution also helps to reduce surface energies
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Ion energies And Particle Densities Cannot Be Controlled Independently Under CCP Conditions
6.0 1.3 1.0
[H] (X 10 atoms/cm )
3
4.5
CCP Conditions
–
0.8 0.5 0.3
–
Bias power increase required to increase self-bias on chuck
3.0
14
–
Plasma generation and ion acceleration
1.5
are coupled together
C2H4 >C2H2
–
0.0
0
0.30
Self-Bias (-V)
200
400
600
0.0 7.5
Densities (X 10 molecules/cm )
3
0
Self-Bias (-V)
200
400
600
CH3>CH2 under all conditions
C2H2 C2H4
Increasing ion currents to chuck an indication of increasing ion flux
0.25
6.0 4.5 3.0 1.5 0.0
0.20
0.15
0.10
0.05
0.00
0
Self-Bias (-V)
200
400
600
12
0
Self-Bias (-V)
200
400
600
Baseline Conditions: CCP 0W, 40mT, 8cm Gap, 30sccm CH4,10sccm Ar
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Ion energies And Plasma Creation Are Decoupled Under ICP Conditions
50 40
3
2.0 1.6
[H] (X 10 atoms/cm )
Degree of Dissociation (%)
ICP Conditions
–
30 20 10 0 0.5 0.4 0.3 0.2 0.1 0.0
1.2 0.8 0.4 0.0 7.5
Bias power increase required to increase
self-bias on chuck
14
–
Plasma generation and ion acceleration are decoupled
Self-bias increase does not produce
additional gas dissociation
0
50 100 150 200
Self-Bias (-V)
0
50 100 150 200
Self-Bias (-V) I-C2H2 I-C2H4
Densities (X 10 molecules/cm )
–
3
C2H2≈C2H4
6.0 4.5 3.0 1.5 0.0
–
CH3 ≈ CH2 under all conditions
P b/Vsb (A)
Ion Current shows no significant dependence on self-bias
0
50 100 150 200
Self-Bias (-V)
12
0
50 100 150 200
Self-Bias (-V)
Baseline Conditions: ICP 200W, 40mT, 8cm Gap, 30sccm CH4,10sccm Ar
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Self-Bias Increase Leads To Completely Opposite Trends For Deposition Rates
400 300 200 100 0
Dep Rate (nm/min)
Dektak-Si Dektak-Quartz n&k Ellipsometer
Contrasting behaviors between CCP
and ICP
–
Direct result of the decoupling of plasma
creation and ion acceleration
0 50 100 ICP Self-Bias (-V) 150 200
ICP conditions increasing sputtering effect with increasing selfbias
30 25 20 15 10 5 0 100 200 300 400 CCP Self Bias (-V) 500
Dektak-Si Dektak-Quartz n&k Ellipsometer
Dep Rate (nm/min)
CCP conditions increased sputtering of sample surface is accompanied by availability of radical species for deposition
600
700
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Conditions Under Which Films With Better Mechanical Properties Are Obtained Are Different For Si and Quartz
1575 ICP CCP 1565
G-pk Position (cm )
-1
Si Substrate
1600 ICP CCP 1590
Quartz Substrate
1555
1580
1545
1570
1535
1560
1525
0
100
200
300
400
500
600
700
1550 -100
100
300 Self-Bias (-V)
500
700
Self-Bias (-V)
Si Substrates
–
Graphitization of films at higher self-bias values for both ICP and CCP conditions Increased ion bombardment of surface causes this -100V is optimum self-bias for Si substrates while depositing using capacitive plasmas
Quartz substrates
– –
–
Ion bombardment is critical for obtaining good quality films on Quartz -300V optimum for CCP samples Optimal levels of ion bombardment Vsb >-200V helps promote SP3 bonding under ICP conditions
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Optical Gap And Surface Energy Are Inversely Related to Self-Bias
3.75 3.25
Optical Gap (eV)
I-FB C-FB
I-TL C-TL
20
Quartz Substrate
2.75
16
(X10 cm )
-1 4
2.25 1.75 1.25 0.75 0 48 46 100 200 300 400 500 Self-Bias (-V) 600 700
12 8 4 0 0 46 44 200 400 Self-Bias (-V) 600 ICP CCP
Surface 2 Energy (mJ/m )
Surface 2 Energy (mJ/m )
44 42 40 38 36 0 100 200 300 400 500 600 Self-Bias (-V) ICP CCP 700
42 40 38 36 0 200 400 600 Self-Bias (-V) ICP CCP
Optical Properties and Absorption Coefficient
– – –
Increasing SP2 content with increasing bias is reflected in reducing optical gaps Possible densification of film deposited is evident through the increasing absorption coefficients for Quartz substrates Considering optical properties in conjunction with Raman fit parameters, self -bias needs to be limited to values that provide good mechanical and optical properties Increased ion bombardment reduces surface energy Hydrogen being sputtered away Simple damage to film surface
Surface Energies
–
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Increasing Coil Powers Above 150W Results In Excessive Dissociation Of Precursor Molecules And Its Byproducts
3.5 3.0
atoms/cm )
00
80 2.5 2.0 1.5 1.0 0.5 0 0 200 400 600 0.0 7.5
molecules/cm )
3 3
ICP Conditions
–
60
–
–
20
[H] (X 10
40
0
1.0
Coil Power (W)
Coil Power (W)
200
400
600
C2H2
C2H4
0.8
6.0
–
4.5 3.0 1.5 0.0
0.6
Coil power increases ne increase Precursor molecules dissociated to higher degrees Both C2H2 and C2H4 decreasing above 150W Steeper decrease for C2H4 Increase in loss of CH 3 to ionization and dissociation is attributed as the reason Near linear increase in ion currents to chuck
0.4
0.2
Densities (X 10
12
14
Power (W)
0
Coil Power (W)
(Kine+Kdne+Kloss)-C2H4
5.85E+04 1.15E+05
(Kine+Kdne+Kloss)-C2H2
2.88E+04 5.55E+04
0.0
0
Coil Power (W)
200
400
600
200
400
600
0 150
Baseline Conditions: ICP 40mT, 8cm Gap, 30sccm CH4,10sccm Ar, -100Vsb
200
300
1.72E+05
2.28E+05
8.22E+04
1.09E+05
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Another Example Of Fact That Deposition On Si And Quartz Do Not Proceed In Similar Fashion
300 250 200 150 100 50 0
–
Dektak-Si Dektak-Quartz n&k Ellipsometer
Si and Quartz Substrates
–
Dep Rate (nm/min)
Precursor molecules dissociated to higher degrees Deposition rate on Si more dependent on
radical fluxes compared to Quartz
substrates
Si substrates floating at higher selfbias values
0
100
200
300
400
500
Coil Power (W)
Higher sputtering of film expected on Si compared to Quartz
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Depletion Of Radical Species Like CH3 and CH2 Is Detrimental To Mechanical Properties
1560 1555
1590
G-Pk Position (cm )
-1
Si Substrate
1595
Quartz Substrate
G-Pk Position (cm )
1550 1545 1540 1535 1530 1525
1585
1580
1575
1570
0
100
200
300
400
500
0
100
200
300
400
500
Coil Power (W)
Coil Power (W)
Si Substrates
–
–
ICP conditions produce graphitic films compared to CCP conditions Depletion of radical species like CH3 and CH2 at higher coil powers leads to increased graphitization of ICP samples
Quartz substrates
–
Higher coil powers no increase in ion energies more polymeric films on Quartz substrates
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Depletion Of Radical Species Like CH3 and CH2 Is Also Detrimental To Optical Properties
2.5 FB TL
10
Optical Gap (eV)
Quartz Substrate
2.0
8
(X10 cm )
-1 4
6 4 2 0 0 46 44 100 200 300 Coil Power (W) 400 500
1.5
1.0 0 46 100 200 300 Coil Power (W) 400 500
Surface 2 Energy (mJ/m )
44
42 40 38 36 0 100 200 300 Coil Power (W) 400 500
Surface 2 Energy (mJ/m )
42 40 38 36 0 100 200 300 400 500 Coil Power (W)
Optical Properties and Absorption Coefficient
– –
Depletion of CH3 and CH2 above 150W correlates to drop in optical gaps
Considering absorption coefficients for films on Quartz even a coil power of 150W is detrimental
Surface Energies
–
Higher Coil Powers Higher Hydrogen concentration lower surface energies
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Changing Source To Chuck Gap Changes Ion Currents To Chuck Without Changing Radical Densities
50 40 30 20 10 0
atoms/cm )
3
2.0
1.6 1.2 0.8 0.4
ICP Conditions
–
[H] (X 10
–
–
6 8
Gap (cm)
10
12
14
0.0 9
Densities (X 1e molecules/cm )
6
8
0.8
Gap (cm) C2H2
10
12
14
C2H4
–
No change in degree of dissociation Plasma creation still in region closer to top coil Almost no change in densities of C2H2 and C2H4 precursor molecules dissociated to higher degrees Ion Flux significantly coupled to distance between source and chuck Loss rates at wall surfaces explains this behaviour
Plasma Production Zone
14
8 6 5 3
0.6
3
n
0.4
12
nradicals(x) ni(x)
Biased Electrode
Quartz Window
0.2
2 0
IonCurrent
0.0
6
8
Gap (cm)
10
12
14
6
8
Gap (cm)
10
12
14
Pc
Cext
Pb
Baseline Conditions: ICP 200W, 40mT, 30sccm CH4,10sccm Ar, -100Vsb
x
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Deposition Rate Trend With Gap Illustrates Importance Of Ion Flux
350 300 250 200 150 100 50 0
350
Dektak-Si Dektak-Quartz n&k Ellipsometer
300 250
Dektak-Si Dektak-Q
Dep Rate (nm/min)
Dep Rate (nm/min)
200 150 100 50 0
6
8
10 Gap (cm)
12
14
0
4
8 12 15 16 -1 -1 Ion Flux (X10 cm s )
20
24
Deposition Rates on Quartz and Si
–
–
Changing gap ion flux changes without change in radical densities
Explicit dependence of deposition rates on both Si and Quartz substrates to ion flux
Optical and Structural Properties
–
Larger gaps result in polymeric films that are hard to characterize using Raman spectroscopy
Decreasing ion flux less number of ions bombarding surface more radical dominated deposition softer and polymeric films
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Summary of Observations
Deposition does not proceed similarly on Quartz and Si substrates
–
–
Different window of parameters for achieving good quality films Primary reason is insulating nature of Quartz substrates Better mechanical and optical properties but higher surface energies at low pressures C2 dimer species is important at low Methane dilutions under ICP conditions Lower Ar dilution is preferred for lower surface energies Contrasting results for ICP and CCP conditions for deposition rates Above an optimal self-bias value, mechanical and optical properties degrade with increasing ion bombardment Excessive dissociation of precursor molecules at high coil powers Limits maximum coil power for better optical and mechanical properties of films Increasing gaps results in reduction in ion flux but radical densities are unchanging Higher gaps are not favorable for obtaining good quality DLC films Limited ion flux availability is primary reason for this behavior
Effect of Pressure
–
Effect of Flow
– –
Effect of Self-Bias
–
–
Coil Power
– –
Gap
–
–
–
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Hardness and Wear Resistance Of DLC Films Are Much Better Than Bare Si And Quartz
Tribological measurements performed on DLC films grown on Silicon and Quartz Samples were characterized for their Hardness and wear resistance Silicon and Quartz substrates coated with DLC film show significant improvement in both hardness and wear resistance
30
Nanoindentation performed with a 75nm Diamond tip and 100 μN load
26
H (GPa)
22
Nano-Indentation 300 nm DLC on Si
Si + DLC
Quartz + DLC
Bare Si
18
Bare Quartz
14
10 0 0.1 0.2 0.3 0.4 0.5
ht (um)
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Sequence Of Steps For Mold Fabrication
Master mold Master Mold
ICP etch conditions
contact angle (˚) surface energy* (=d+p) (mJ/m2)
PMMA DLC Film Si/quartz
H2O
Ethylene glycol
p
d
48.6 38.0 30.7 42.0 43.9
Mold Release
DLC Film Si/quartz ICP etch using CF4
Before etch 57.0 52.5 45.8 2.9 CF4, 20 mTorr 89.5 55.0 1.4 36.6 CF4, 60 mTorr 90.0 61.0 2.7 28.0 CF4+5% O2, 20 mTorr 59.0 43.0 32.4 9.7 CF4+10% O2, 20 mTorr 57.0 41.5 34.8 9.1 *Surface energy is the sum of d (dispersion component) and p(polar component)
Si/quartz
Strip Si/quartz PMMA Process Sequence for mold fabrication
– –
Transfer required feature from master mold to substrate with DLC film ICP etch with CF4 or CF4/O2 chemistry typical etch rates of 75 nm/min with selectivity to PMMA of about 5:1
Effect of ICP etch on key properties of mold
–
– –
CF4 etch does not affect roughness of DLC films
Pure CF4 etch increased water contact angle on DLC 5% O2 addition does not affect contact angles in any way
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
DLC Mold Structures Created And Successfully Used For Imprinting
AFM pictures of DLC mold (left) and imprinted pattern (right)*. A
1μm wide, 70nm deep trench
was successfully imprinted onto PMMA resist layer
AFM pictures of DLC mold (left) and imprinted pattern (right)*. A 40nm line and space feature successfully imprinted onto
a SU8 resist layer is shown here
*Mold Fabrication and Characterization Were Done By Li Tao and Caleb Nelson
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
OUTLINE
• • • • • •
Research Objectives Introduction to Diamond like Carbon Experimental Setup Diamond like Carbon and Nanoimprint Lithography PECVD of Carbon-Carbon Composites Summary
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Nanotubes & Composites
• •
•
Electrical Conductivity- Max. current density = 109 A/cm2 (copper = 106 A/cm2)
Mechanical- Tensile yield strength = 63 GPa (1.2 GPa for steel), Young’s modulus = 1200 GPa ( 200 GPa for steel), Density = 1.4 g/cm3 (2.7 g/cm3 for Al, 4.5 g/cm3 for Ti) Thermal- Approximately 2000 W/m K
Nanotube based Composites
Handling
1. 2. 3. 4. 5.
Electronics Field emission displays Supercapacitors Structural Applications Heat resistant shields
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Motivation CNT Composite Materials
Materials
PMMA & MWNT Parylene & MWNT Copper & MWNT SiO2 & MWNT
Technique
Spin Coating Thermal CVD Sintering CVD
Purpose
Electrical, Mechanical Mechanical Mechanical Thermal Conductivity
Reference
E. Lahiff et al. Huai-Yuan Chu et al. Kyung Tae Kim et al. Jinwei Ning et al.
Ceramic & MWNT DLC & SWNT
Sintering PLD
Mechanical & Electrical Wear Resistance
Guo- Dong Zhan et al. H. Schittenhelm et al.
Issues
• • • •
Uniform Dispersion Loss of Alignment Damage to tubes
Weak Interaction between filler and nanotube
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Why DLC as a Filler Material?
DLC
Nanotubes
Better control of filler properties in case of PECVD Stronger covalent bonds at the interface between the filler material and the nanotubes Symbiotic structure
–
Stress relief for DLC and stability for CNTs
Room temperature deposition
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Typical DLC Deposition Conditions Always Yielded CNT Forest Covered With Film Only At The Top
Typical DLC deposition conditions always lead to film growth only at the top few 100’s of nanometers SEM picture corresponds to deposition done at 100mT, 100W, 20sccm of Ar and 40 sccm of CH4(Reactor CAPPY)
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Time Evolution Experiments To Study CNT Filling Process
1500 CNT Forest t=20mins 1200
Intensity (a.u)
t=5mins t=60mins
900 600 300 0 750
Id/Ig=1.22
Time (minutes)
Id/Ig=1.17
Film thickness (nm)
80 280 900
5 20
Id/Ig=1.9
60
Id/Ig=0.81
1000
1250 1500 -1 Raman Shift (cm )
1750
2000
Identical CNT samples subjected to different deposition times
–
Increase in disorder initially signs of increase in Graphitization with time
SEM pictures for these samples indicate increase in film thickness around the nanotubes only at the top
few 100’s of nanometers
Infusion experiments performed with a-Si: H deposition and DLC deposition at elevated temperatures pointed towards needs for testing effects of high density plasma
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Characterize Depth of Infusion Into CNT Forest Under Various Plasma Conditions
All experiments are performed in the reactor MGEC
Vary coil power from 0-500W – Vary bias power from 0 to 200W – Vary process pressure from 5mT to 65mT – (Arsccm/CH4sccm) from 22.5% to 87.5% – Coil to chuck gap of 6 to 14 cm – Duty Cycle [tON/(tON+tOFF)]*100 33% for ICP conditions 100% for CCP conditions Film characterization – Depth of infusion using Scanning Electron Microscopy
–
Deposited Film
Depth of Infusion
Vertically Aligned CNT Forest
Si Substrate
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
1600
Depth Of Infusion Trends Differently Under ICP And CCP Conditions While Changing Pressure
ICP CCP
Depth of Infusion (nm)
1200
Depth of infusion increases with increasing pressures
–
800
Increase near monotonic for CCP samples Only upto 40mT in case of ICP samples
400
–
0 0 10 20 30 40 50 60 70
Pressure (mT)
12.5 1.5
Behavior correlates well with ion current under CCP conditions Under ICP conditions
–
(X 10 ions/cm .s)
Depth of Infusion (m)
2
10.0 7.5 5.0 2.5 0.0 10 20 30 40 50 60
1.2 0.9 0.6 0.3 0.0 70 0.8 0.6 0.4 0.2 0.0 0 10 20 30 40 50 60
Ion Flux
Species that promote faster deposition increase in gas phase (like CH and H) Accelerate deposition at top at higher
15
–
ICP Pressure (mTorr)
0.20 0.16
pressures
Depth of Infusion (m)
–
Leads to Faster capping of CNT forest at the top
Pb/Vsb(A)
0.12 0.08 0.04 0.00
CCP-Pressure (mTorr)
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Increasing Self-Bias Results In Contrasting Trends For Depth of Infusion Under ICP And CCP Conditions
0.30 0.25 0.20 0.9 0.15 0.6 0.10 0.05 0.00 100 200 300 400 500 600 0.3 1.5
15
1.8 1.5 1.2
Ion Flux(X10 ions/cm .s)
Depth ofInfusion(m)
Depth ofInfusion(m)
2
1.2
12
Pb/Vsb(A)
9 0.9 6 0.6 3 0.3 0.0 0 50 100 150 200
0.0 700
15
0
CCP Self-Bias (-V)
ICP-Self-Bias (-V)
CCP conditions
–
–
Increasing self-bias, achieved by increasing bias power also increases radical densities More ions available with higher energies that can penetrate deeper into the nanotube forest creating active sites Behavior not similar to CCP conditions reiterates the fundamental difference in the way these discharges operate
ICP conditions
–
Increasing average energies of ions increases deposition rate at the top Widening of sheath that increases angular distribution of ions before they reach CNT surface
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Ion Flux Explains Trends Exhibited By Depth Of Infusion To A Large Extent
40.0 1.5
25.0
1.8 1.5
Depth of Infusion (m)
Ion Flux (X10 ions/cm .s)
Ion Flux (X10 ions/cm .s)
32.0
1.2
20.0
Depth of Infusion (m)
2
1.2 15.0 0.9 10.0 0.6 5.0 0.3 0.0
24.0
0.9
16.0
0.6
8.0
0.3
15
0.0 0 100 200 300 400 500
0.0
0.0
6
8
Coil Power (W)
10 Gap (cm)
12
14
Importance of ion-flux reiterated
–
–
Gap-dependence easily illustrates this fact Gap increases ion flux decreases radical densities are unchanging Coil Power dependence also illustrates ion-flux dependence But ion flux alone does not explain complete behavior Could radical density trends help explain this?
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Behavior Of Radical Densities Has To Be Taken Into Account To Completely Explain Effect of Coil Power
Ion Flux (X10 ions/cm .s)
40 32 24 16 8 0 3.0 0 100 200 300 400 500 Coil Power (W) 1.5 1.2 0.9 0.6 0.3 0.0 1.5 1.2 0.9 0.6 C2H2 C2H4 0 100 200 300 400 500 Coil Power (W) 0.3 0.0
Depth of Infusion (m)
2
By changing coil powers
–
15
Radical fluxes are decreasing at coil powers >150W
–
–
Unlike increasing gap conditions
Depletion of less reactive radical species
Drives deposition towards one dominated by ions Increases deposition rate at the top leading to capping of CNT forest
Depth of Infusion (m)
Densities(X10 cm )
2.4 1.8 1.2 0.6 0.0
Results in saturation of depth of infusion
12
-3
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Different Locations Of CNT Forest Exhibit Contrasting Infusion Levels
A
Ar+CH4 Plasma E-, ions, radical
A’
Γion E
Sheath Edge
Γradicals
Aluminum Chuck
~ RF
Bias
Different regions of CNT forests exhibit different levels of infusion
–
–
–
Same CNT forest examined at two locations Edge of CNT forest shows conformal deposition along length of CNT forest (Гradicals>>Гions) at this location Ions reaching this location have lost most of their energy to collisions Center of CNT forest shows a depth of infusion of about 1.5µm (Гradicals>Гions) at this location Ions reaching this location are directed and possess high kinetic energies
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Examples of Complete vs Incomplete Fill
16cm, -60V
5cm, -60V
6µm CNT Forest
6µm
30µm CNT Forest
5cm, -60V
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
Summary of Results-CNT Infusion Using a PECVD Process
Depth of infusion of a PECVD film inside a CNT forest has been studied under different plasma conditions Effect of Pressure
– –
For CCP conditions, infusion depth correlates well with ion current to chuck and increases with pressure For ICP conditions, infusion depth decreases at higher pressures
Effect of Flow
–
Methane dilution has no significant impact on the infusion depth CCP conditions-infusion depth increases monotonically with self-bias Ion current and ion-energy are increasing simultaneously ICP conditions-infusion depth is decreasing slightly with self-bias Ion-energy is increasing at a constant ion current Changing ion-current by either changing gap or coil power results in contrasting behavior for infusion depth Coil power increase also alters radical densities significantly Changing gap does not change radical densities
Effect of Self-Bias
–
–
Coil Power and Gap
–
While ion flux controls depth of infusion to a large extent it is also shown that relative contribution of radical and ion-flux determine ultimate depth of infusion into CNT forest
Plasma Science and Applications Laboratories
Seetharaman Ramachandran
ACKNOWLEDGEMENTS
•
•
My advisor Prof. Lawrence J. Overzet – for his thoughtful ideas, motivation through some tough times and unstinted patience in guiding me through this dissertation
Committee members – Prof. Gil Lee, Prof. Matthew Goeckner, Prof J.B.Lee, Prof. Walter Hu and Dr. Slade Gardner Post Docs – Dr. Jeong Soo Lee, Dr. Bong Mo Park , Dr. Mike Kozlov, Dr. Amandeep Sra and Dr. Byeong Jun Lee for their cooperation
• • •
•
Graduate Students –Eric Joseph, Baosuo Zhou, Yonghua Liu, Anand Chandrasekhar, Sanket Sant, Ashish Jindal, Caleb Nelson, Daisuke Ogawa, Monali Mandra, Iqbal Saraf, Li Tao, Nandini Sundaram, Frank Yeh, Chang Soo Kim, Kyung Hwan Lee and Fabian Reyes Special thanks to Keith Slinker for Lockheed Martin for providing the CNT samples for our experiments
Clean Room Staff - Dr. Gordon Pollack, John Maynard, Keith Bradshaw, Tom Chaddick and John Goodnight
• •
University of Texas at Dallas and Lockheed Martin for financial support during my research
Friends and family for support and encouragement and my wife Harini in particular for standing by me through some tough and uncertain times
Plasma Science and Applications Laboratories
Seetharaman Ramachandran