The world leader in serving science
Extraction of Depth
Information from ARXPS
Data
John Wolstenholme
Theta Probe Features
X-ray monochromator with spot size
from 15 µm to 400 µm
Real time angle resolved XPS
analysis without sample tilting
2
Theta 300XT
Parallel ARXPS from 300 mm
wafers
3
Microfocusing Monochromator
15 µm – 400 µm spot size
4
Geometry
5
Theta Probe
6
Collection Conditions
Angular Range
• 20°to 80°
Parallel collection
• Up to 96 channels in angle
• Generally, 16 angles are used giving an angular
resolution of 3.75°
• Up to 112 channels in energy
Parallel collection allows rapid ‘snapshot
acquisition’
• Excellent for ARXPS maps
• Thickness maps
• Dose maps
7
Data Acquisition
Binding energy
• Parallel (snapshot) or scanned
Angle
• Always parallel
Video of snapshot acquisition
• SiO2 on Si
Advantages of Parallel ARXPS
•
•
•
•
•
Fast
ARXPS from small features is possible
Mapping possible
Analysis area is constant
ARXPS from large samples
8
Attenuation Length
Each data point represents a
different element or transition
λ ∝~ E
Intensity as a function of depth
• 65% of the signal from <λ
• 85% from <2λ
0.5
• 95% from <3λ
M. P. Seah and W.A. Dench, Surface and Interface Analysis 1 (1979) 2
9
Relative Depth Plot (RDP)
Select range of bulk sensitive angles (Ibulk)
Select range of more surface sensitive angles (Isurface)
Ln (Isurface / I bulk)
Provides qualitative species depth distribution
HfO2
SiO2
Si
Relative Norm. Area
0.6
0.5
O1s
0.4
0.3
Hf4f
0.2
Si2p3/2
Si2p(Ox)
0.1
0.0
30
40
50
60
Angle (°)
10
70
80
RDP – HfO2 Grown by ALD
30 Cycles ALD on thermal SiO2
Hf4f
O1s
Si2pOx
30 Cycles ALD on HF-last Si
Si2p3
Advantages
Hf4f
O1s
Si2pOx
Si2p3
Disadvantages
• Excellent for showing the ordering of
the layers
• Shows gross differences between
samples
• Does not rely on mathematical models
• Does not require a knowledge of the
material properties
• No quantitative data
• No means to interpret the differences
between samples
• Subtle differences not apparent
11
Thickness Measurement
Equations
5
9.0 nm
• Photoemission from a thin
film
4
• Attenuation by a thin film
• I = I0 exp(-d/λcosθ)
• Where
• I0 = Emission from bulk material
• θ = Emission angle
(with respect to sample normal)
• d = Film thickness
• λ = Attenuation length
Silicon dioxide on silicon
Plot:
ln(1+R/R∞ )
• I = I0 [1-exp(-d/λcosθ)]
6.4 nm
3
4.3 nm
2
3.6 nm
2.3 nm
1
1.9 nm
0
0
0.5
• ln[1+R/ R ∞] vs. 1/cos(θ)
Gradient
• d/λ
12
1
1/cos(θ)
1.5
2
Thickness Measurements –Single Layer
Equations
10
• I = I0 [1-exp(-d/λcosθ)]
• Attenuation by a thin film
• I = I0 exp(-d/λcosθ)
• Where
• I0 = Emission from bulk material
• θ = Emission angle (with respect to
sample normal)
• d = Film thickness
• λ = Attenuation length
ARXPS Measurements (nm)
• Photoemission from a thin film
y = 1.077x - 0.914
8
6
4
2
0
0
For thickness calculation
• ln[1+R/ R0] = d/(λA cosθ)
SiO2 on Si
2
4
6
8
Ellipsometry Measurements (nm)
Beware of R0
• For SiO2/Si experimental value is very
different from calculated value
13
10
XPS Measurements of SiO2 Thickness
Thickness (nm)
10
Comparison of ARXPS with fixed
angle XPS
8
• Good agreement except at large
thickness
• Single angle measurement samples
large angular range.
6
4
2
10
0
ARXPS
Instrument 1 Single Angle
Instrument 2 Single Angle
Instrument 2, 2 Angles
Linear (ARXPS)
20
40
60
Measured Thickness
8
80
Maximum Angle (°)
ARXPS measurements
• Effect of angular range upon
measured thickness
• Minimum angle is 23°in all cases
• Highest usable maximum angle
depends upon oxide thickness
6
4
2
0
0
2
4
6
Nominal Thickness (nm)
14
8
10
Multiple Overlayers
1
2
n
Substrate
Choose XPS transitions representing the
composition of each layer
Measure signals as a function of angle
Determine intensity ratios
• Ij(θ)/Ij+1(θ)
• Ij (θ) /Isub (θ)
• Ij (θ) is the XPS signal from layer j as a
function of angle
Fit theoretical ratios to measured ratios
using layer thickness as fitting parameter
Best results are obtained by finding the best
fit to all of the data simultaneously
15
Layer Thickness Calculation
HfO2
SiO2
Layer Thickness (nm)
Si
16
HfO2 Growth on Thermal SiO2
Si
2
SiO2
1
0
0
20
ALD Cycles
40
HfO2 Thickness (ARXPS) (nm)
Thickness (nm)
3
Thickness calculated from
PARXPS data using multilayer thickness calculator
Increasing growth rate
Constant SiO2 thickness
HfO2
HfO2
SiO2
2.5
2
0.15
0.1
0.05
1.5
0
0.00
0.20
0.40
1
0.5
0
0.00
2.00
4.00
Hf Coverage (RBS/1e15)
17
6.00
Thickness Measurement of SAMs Using PARXPS
Measurements from 3 Alkanethiols deposited on gold
• C9H19SH
• C11H23SH
• C16H33SH
C1s
S2p
Au4f
Layer Thickness
2.5
Acquisition Times
• C 1s
• Au 4f
• S 2p
3 min
1 min
15 min
2
1.5
1
0.5
0
0
5
10
15
Number of Carbon Atoms
18
20
Thickness Measured as Function of Time
Conditions
• Angle integrated
• Acquisition time per point = 1.25 min.
1.5
1.2
1.2
Thickness/nm
0.6
Duration of PARXPS
Measurement
0.3
0.9
0.6
0.3
C9H19SAu
0
0
20
2
40
C12H25SAu
0
60
0
1.6
Time/min
Thickness/nm
Thickness/nm
0.9
20
40
60
Time/min
1.2
0.8
0.4
C16H33SAu
0
0
20
40
Time/min
19
60
~20% Decrease
in 70 minutes
Thickness Map
Conditions
•
•
•
•
•
•
Sample C16H33SAu
Angle integrated
Snapshot
20 x 20 pixels
100 µm spot size
Time per point 3 seconds
Clear evidence for X-ray damage
caused during ARXPS
measurement
20
Thickness Measurement
Layer Thickness
2.5
ARXPS can be applied to
delicate samples by mapping the
sample and summing the data.
Only feasible with PARXPS
From Map
2
1.5
1
From Static Point
0.5
0
0
5
10
15
20
Number of Carbon Atoms
21
Generation of Depth Profiles
Summary of the Maximum Entropy method
Start
Generate Trial Profiles
Use Genetic Algorithm
20000 Generations
Maximum
Entropy
Optimisation
Average 5
or 20
profiles
Generate
output
profile
30-point profiles generated
Profiles generated from angular range
<62°
Average 5 cycles of
• 20,000 generations
• Powell optimisation
End
22
Sample
Generate Random
Profile
HfO2
Al2O3
SiO2
Si
100
O1s
Si2p
80
60
Hf4f
40
Si2p(O)
Al2p
20
0
0
0.7
Relative Intensity (%)
Atomic Concentration (%)
Depth Profile Generation
2
4
Depth (nm)
0.6
O1s
0.5
0.4
Si2p
0.3
Hf4f
0.2
Al2p Si2p(O)
0.1
Tj(θ) = exp(-t/λcosθ)
0
20
40
60
Calculate Expected
ARXPS Data (Beer
Lambert Law)
80
Angle (°)
23
6
Depth Profile Generation (2)
Determine error between observed and
calculated data:
Relative Intensity (%)
0.7
0.6
Calculate the entropy associated with
a particular profile (the probability of
finding the sample in that particular
state)
c 
S = ∑ ∑ c j,i − c 0j,i − c j,i log 0j,i 
c 
j
i
 j,i 
O1s
0.5
0.4
0.3
Si2p
0.2
Hf4f
Al2p Si2p(O)
0.1
cj,i is the concentration of element i in
layer j
Maximise the joint probability function
0
20
40
χ2 = ∑
k
(I
60
Angle (°)
calc
k
−I
σ k2
Q = α S − 0. 5 χ 2
80
)
obs 2
k
Repeat process to obtain most likely
profile
24
Effect of Alpha on Generated, Unconstrained Profile
α=0
α=3x10-8
Si0
80
O
Si4+
60
Hf
40
20
Atomic Concentration (%)
100
80
60
40
20
0
0
0
1
2
0
3
1
α=3x10-7
80
60
40
20
0
0
1
2
3
α=10-4
100
Atomic Concentration (%)
100
2
Depth (nm)
Depth (nm)
Atomic Concentration (%)
Si
Atomic Concentration (%)
100
HfO2
SiO2
3
Depth (nm)
80
60
40
20
0
0
1
2
Depth (nm)
25
3
Add Constraints and Automate
Need to apply constraints to prevent chemically unreasonable solutions
(analogy to spectrum peak fitting)
• We know the composition of the substrate
• Assume mixtures of stoichiometric components (“Fit Units”)
• Examples
HfO2 and SiO2
SiOxNy = (SiO2 ) a + (Si3N4) b
• Non-stoichiometric components / “Free” oxygen
Examine data to obtain the optimum value for ‘α’ (automated)
Choose most appropriate angular range (automated)
• Depends upon layer thickness
26
Use Fit Units
Si
Atomic Concentration (%)
HfO2
SiO2
100
Si0
80
O
60
40
Si4+
Hf
20
0
0
0.5
1
1.5
Depth (nm)
27
2
Effect of Interlayer on 30 Cycle ALD Layer
Grown on thin SiO2 layer
Atomic Concentration (%)
100
80
O1s Si2pOxSi2p3
Hf4f
Atomic Concentration (%)
Hf4f
Grown on HF last surface
100
Hf4f
O1s
Si2pO
Si2p
60
40
60
40
20
0
0
0
2
Depth (nm)
28
Hf4f
O1s
Si2pO
Si2p
80
20
0
O1s Si2pOxSi2p3
1
2
Depth (nm)
3
SiON Chemical State Profile
N
N 1s
(Low BE)
N
(High BE)
405
402
399
396
Binding Energy (eV)
N
Si4+
O
393
Atomic Concentration (%)
100
80
O
N
Si0
(Low BE)
60
Si4+
40
N
20
(High BE)
(Low BE)
0
0
N
1
2
Depth (nm)
(High BE)
Si0
29
3
Ti/W Alloy
100
Atomic Concentration (%)
W
WO3
WO2
C1s
W4f
O1s
80
60
W4f
W4f
(WO3)
(WO2)
Ti2p
(TiO2)
40
20
Ti2p
W 5p3/2
0
41
39
37
35
33
31
29
27
0
Binding Energy (eV)
1
2
Depth (nm)
30
3
4
Profiles from SAMs
1-Mercapto-11-undecyl-tri(ethylene glycol)
Nonanethiol
C 1s
100
Au 4f
80
60
40
S 2p
20
Au 4f
Concentration/%
Concentration/%
100
0
80
C 1s
C-O 1s
60
O 1s
40
20
S 2p
0
0
0.5
1
1.5
0
Depth/nm
1
2
Depth/nm
31
3
Nitrogen Dose and Thickness
300 mm wafer
Single measurement
49-point maps
Thickness
Dose
32
Conclusions
Angle Resolved XPS is a powerful tool for the
characterisation of ultra-thin films
• Layer ordering
• Multiple layer thickness determination
• Depth profiles
Parallel ARXPS extends the capabilities of the technique
•
•
•
•
Small features
Large samples
Mapping
Larger number of angles
33
Acknowledgements
Dr K Bonroy, IMEC, Belgium
Dr T Conard, IMEC, Belgium
Dr D Graham, Asemblon, USA
Financial support of the EU-CUHKO project
34