X-ray photoelectron spectroscopy - An introductionfolk.uio.no/vebjornb/MENA9510/lectures-2016/MENA9510_XPS_2016.pdf · XPS Depth of Analysis. The probability that a photoelectron

MENA 9510 10th October 20161

X-ray photoelectron spectroscopy - An introduction

Spyros DiplasSINTEF Materials & Chemistry, Sector of Materials and Nanotechnology, Department

of Materials Physics-Oslo& Centre of Materials Science and Nanotechnology, Department of Chemistry, UiO

[email protected]@smn.uio.no

mailto:[email protected]

mailto:[email protected]


Material Characterisation Methods


hν

Photoelectron

2p1/2, 2p3/2

2s

1s

Ekin = hν – EB - ω

L23

L1

K

EKL2,3L2,3(Z) = EK(Z) – [EL2,3(Z) + EL2,3(Z + 1)]

Internal transition

(irradiative)

Auger electron

XPS-Basic Principle

valence bandFermi

Vacuum

De-excitationExcitation


Auger electron vs x-ray emission yield

5

B Ne P Ca Mn Zn Br Zr

10 15 20 25 30 35 40 Atomic Number

Elemental Symbol

0

0.2

0.4

0.6

0.8

1.0

Pro

babi

lity

Auger Electron Emission

X-ray Photon Emission

It is easier to detect light elements with Auger than with EDS


XPS spectrum ITOx 104

10

20

30

40

50

60

70

80

CPS

1200 1000 800 600 400 200 0Binding Energy (eV)

In 3dSn 3d

O 1s

In 3pSn 3p

In 3sIn 3s

In MNNSn MNN

O KLL

Auger peaks

Photoelectron peaks

In/Sn 4pIn/Sn 4s

C 1s


Peak width (ΔE)ΔE = (ΔEn

2 + ΔEp2 + ΔEa

2)1/2

• Gaussian broadening:

-Instrumental:

There is no perfectly resolving spectrometer nor a perfectly monochromatic X-ray source.

-Sample

For semiconductor surfaces in particular, variations in the defect density across the surface will lead to variations in the band bending and, thus, the work function will vary from point to point. This variation in surface potential produces a broadening of the XPS

peaks.

-Excitation process such as the shake-up/shake-off processes or vibrational broadening.

• Lorentzian broadening.

The core-hole that the incident photon creates has a particular lifetime (τ) which is dependent on how quickly the hole is filled by an electron from another shell. From Heisenberg’s uncertainty principle, the finite lifetime will produce a broadening of the peak.

Γ=h/τ

Intrinsic width of the same energy level should increase with increasing atomic number

Natural widthX-ray source contribution

Analyser contribution

MENA 9510 10th October 2016 7

Examples of XPS spectrometers


Schematic of an XPS spectrometer


Instrument: Kratos Axis UltraDLD at MiNaLab

Analyser

Monochromator

Sample

Detector

X-ray source

X-ray source

e-

e-


Parallel angle resolved XPS instrument-Theta Probe

10

Spectroscopy Source-defined small area XPS

15 µm to 400 µm

Snapshot spectrum acquisition Up to 112 channels Faster serial mapping Faster profiling

Unique parallel ARXPS with up to 96 channels Large samples (70 mm x 70 mm x 25 mm) Sputter profiles Mapping possible up to full size of sample

holder ISS included

Target applications• Thickness measurements• Surface modification, plasma & chemical• Self assembly• Nanotechnology• Ultra thin film technologies• Shallow interfaces


Sample requirementsHas to withstand high vacuum (≤ 10-7 Torr).

Has to withstand irradiation by X-rays

Sample surface must be clean!

Reasonably sized.


XPS Depth of Analysis

The probability that a photoelectron will escape from the sample without losing energy is regulated by the Beer-Lambert law:

Where λe is the photoelectron inelastic mean free path

Attenuation length (λ) ≈0.9 IMFPIMFP: The average distance an electron with a given energy travels between

successive inelastic collisions


Features of the XPS spectrum Primary structure

- Core level photoelectron peaks (atom excitation)- Valence band spectra - CCC, CCV, CVV Auger peaks (atom de-excitation)

Secondary structure

- X-ray satellites and ghosts- Shake up and shake off satellites- Plasmon loss features- Background (slope)


Quantification Unlike AES, SIMS, EDX, WDX there are little in the way of matrix effects to worry

about in XPS. We can use either theoretical or empirical cross sections, corrected for transmission function of the analyser. In principle the following equation can be used:

I = J ρ σ K λ I is the electron intensity J is the photon flux, ρ is the concentration of the atom or ion in the solid, σ s is the cross-section for photoelectron production (which depends on the element and

energy being considered), K is a term which covers instrumental factors, λ is the electron attenuation length.

In practice atomic sensitivity factors (F) are often used: [A] atomic % = {(IA/FA)/Σ(I/F)} Various compilations are available.


Chemical shift

ΔE(i) = kΔq + ΔVM – ΔR

Initial state contribution

• Δq: changes in valence charge

• ΔVM : Coulomb interaction between the photoelectron (i) and the surrounding charged atoms.

.

final state contribution

• ΔR: relaxation energy change arising from the response of the atomic environment (local electronic structure) to the screening of the core hole

15

Initial state: Before photoemission-ground stateFinal state: After photoemission


Chemical shift - Growth of ITO on p c-SiIn

tens

ityar

bitra

ry u

nits

Binding Energy (eV)

SiOx

Si In oxide

In

Sn oxide

Sn

3/2 3/2 5/2 5/2

1.5 nm

0.5 nm

BHF 15 sec + 500oC 0.5 nm

1.5 nm

3.0 nm

0.5 nm

1.5 nm

3.0 nm

Si 2p In 3d Sn 3d


Chemical shift


Shake-up satellites in Cu 2p

Shake-up satellites

2p3/2

2p1/2

Cu

CuO

CuSO4

970 960 950 940 930

Binding energy (eV)


Plasmons

Pure elements

Mo-Si-AlCompound

They describe the interaction (inelastic scattering) of the PE with the plasma oscillation of the outer shell (valence band) electrons

Plasmons in their quantum mechanical description are pseudoparticles with energy Ep=hω

ω = (ne2/ε0m)1/2/2π n =valence electron density, e, m electron charge and massε0=dielectric constant of vacuum


Peak asymmetry

Peak asymmetry in metals caused by small energy electron-hole excitations near EF of metal

Arb

itrar

y U

nits

16 14 12 10 8 6 4 2 0 -2Binding Energy (eV)

Arb

itrar

y U

nits

1055 1050 1045 1040 1035 1030 1025 1020 1015 1010Binding Energy (eV)

Zn

ZnO

ZnZnO


Depth profile with ion sputtering

Use of an ion gun to erode the sample surface and re-analyse Enables layered structures to be investigated Investigations of interfaces Depth resolution improved by:

Low beam energiesSmall ion beam sizesSample rotation

SnO2

Sn

Depth500 496 492 488 484 480


Probing different depths by choosing different areas in the XPS spectrum

30

Binding Energy (eV)

Ge3d

Ge(IV)

Ge(0)

4.0 eV

1220

Binding Energy (eV)

Ge2p3/2Ge(IV)

Ge(0)

4.0 eV

1140 1150 1160 1170 1180

Kinetic Energy (eV)

GeLMM

Ge(IV)

Ge(IV)

Ge(0)

Ge(0)8.0 eV

8.2 eV

Ge3d: EK = 1453 eV

λ = 2.8 nm

Ge2p: EK = 264eV

λ =0.8 nm

GeLMM: EK = 1140 -1180 eV


Angle Resolved XPS (ARXPS)for non-destructive depth profile

Substrate

I(d) = Io*exp(-d/λcos θ)θ

Film

I (d) = Io* exp(-d/λ)

λ=attenuation length (λ ≈0.9 IMFP)

λ=538αA/EA2 +0.41αA(αA EA)0.5

(αA3 volume of atom, EA electron energy)

Arb

itra

ry U

nits


OH oxide

bulk

AR

surface

RT


XPS-Check list Depth of analysis ~ 5nm All elements except H and He Readily quantified (limit ca. 0.1 at%) All materials (vacuum compatible) Chemical/electronic state information

-Identification of chemical states-Reflection of electronic changes to the atomic potential

Compositional depth profiling by -ARXPS (ultra thin film <10 nm), -change of the excitation energy -choose of different spectral areas-sputtering

Ultra thin film thickness measurement Analysis area mm2 to 10 micrometres

XPS gives information of what elements exist on the sample surface, how much of each element, at what chemical state, how much of each chemical state.


Some application examples


Interfacial studies of Al2O3 deposited on 4H-SiC(0001)Avice, Diplas, Thøgersen, Christensen, Grossner, Svensson, Nilsen, Fjellvåg, WattsAppl. Physics Letters, 2007;91, 52907, Surface & Interface Analysis,2008;40,822

• d=λSi cosθ ln(1+R/R∞)

• d: SiOx film thickness

• λSi:inelastic mean free path for Si,

• Θ: the angle of emission,

• R: the Si 2p intensity ratios ISiox/ISiC,

• R∞ the Si 2p intensity ratios I∞Siox/I∞SiC where I∞ is the intensity from an infinitely thick substrate.

• R∞=(σSi,SiO2 . λSi,SiO2) / (σSi,Si . λSi,Si)

• where σSi,SiO2 and λSi,SiO2 are the number of Si atoms per SiO2 unit volume and the inelastic mean free path respectively

• The σSi,SiO2 / σSi,Si ratio is given by

• σSi,SiO2 / (σSi,Si = (DSiO2 . FSi) / DSi . FSiO2

• where D is the density of the material and F the formula weight.

• For the calculations we also assumed that the Si 2p photoelectrons from both SiC and Si oxide film will be attenuated by the same amount as they travel through the Al2O3 film therefore, their intensity ratio will reflect the attenuation of the Si 2p electrons coming from the SiCthrough the Si oxide film.

From XPS

d= 1nm at RT, d=3nm at 1273 K

Si0

Si4+Si3+

SiC/Si2+ Al 2p plasmon contribution

Si+

SiC

Al2O3

SiOx

as-grown

N2/H2 873 K 15 mins

N2/H2 873 K 30 mins

Ar 1273 K 15 mins,

Ar 1273 K 30 mins

Ar 1273 K 60 mins


XPS on ITO e-beam deposited prior and after annealing (SINTEF internal)

Sn 3d In 3d

O 1s Valence band

e-beam deposited

e-beam deposited

Air annealed 300 C

Air annealed 300 C


Nitrogen in ZnO

Arb

itra

ry U

nits

410 408 406 404 402 400 398 396 394 392 390 388Binding Energy (eV)

AG

900 oC

200 sec sputtering

200 sec sputtering

outermost

outermost

Arb

itra

ry U

nit

s

414 412 410 408 406 404 402 400 398 396 394Binding Energy (eV)

N

N-C=Oor

N-CNO2


Band bending in ZnOR. Schifano, E. V. Monakhov, B. G. Svensson, and S. Diplas, 2009, Appl. Phys. Lett. 94, 132101

Arb

itrar

y U

nits


Arb

itrar

y U

nits

980 982 984 986 988 990 992 994 996 998 1000Kinetic Energy (eV)

Zn 2p - O 1s energy difference

490

490,2

490,4

490,6

490,8

491

491,2

491,4

491,6

0 20 40 60 80 100

Etching time (sec)

Zn 2

p - O

1s

ener

gy d

iffer

ence

(eV)

LR4 Zn2p-O1sLR3 Zn2p-O1s

Zn 2p-LM45M45 Auger parameter comparison

2008,5

2009

2009,5

2010

2010,5

2011

0 20 40 60 80 100

Etching time (sec)

Zn 2

p-LM

45M

45 A

uger

par

amet

er (e

V)

LR4 2p-LM45M45LR3 2p-LM45M45


3SEM of a Cu(In,Ga)Se2 solar cell (cross-section) and its mode of operation

CIGS solar cell• Energy/environmental application

• Solar cells based on Cu(In, Ga)Se2 (CIGS) • Thin-film stack on glass• Mo and Zn oxide layer form electrical contacts• p-type CIGS film (sunlight absorber) and n-type

CdS film form p-n junction• Excellent efficiency• Low cost compared to thicker silicon-based solar cells

• Practical problem• Controlling film composition and interfacial chemistry

between layers (affects electrical properties)• XPS solution

• XPS sputter depth profiling Elemental and composition information as a function of depth Identify chemical gradients within layers Investigate chemistry at layer interfaces

Acknowledgement: Thermo Electron Corporation

Presenter

Presentation Notes

Solar cells based on (CIGS) thin films have demonstrated excellent efficiencies and offer a low-cost potential as compared with bulk-silicon-based solar cells, which are about 200 times thicker. CIGS solar cells consist of a thin-film stack on a substrate (typically glass) as shown in the SEM image to the left of the slide. The molybdenum layer and the zinc oxide layer form the electrical contacts. The p-type CIGS film acts as the sunlight absorber layer, with a thin n-type CdS layer forming a p-n junction. The most common manufacturing methods are simultaneous or sequential evaporation or sputtering of copper, indium, and gallium. Vaporised selenium reacts with the metals in order to establish the final film composition. One of the major challenges in producing these thin film solar cells is to control the film composition. Reproducibility of the required layer design in commercial volumes has proven to be problematic. This is critical as the electrical properties of the cell depend on the exact composition of the layers. XPS depth profiling can be used to determine both the composition through the device, and the interfacial chemistry.


Depth profile of CIFS film stack

CIGS solar cell• Depth profile of CIGS film stack

• Demonstrates standardless quantification of XPS• Excellent quantification agreement between XPS and

Rutherford BackScattering (RBS)• Both techniques show cross-over of In and Ga close to

1.6µm depth• XPS tool is able to analyze product solar cell device

Zn

O

Se

Cu

Ga

In

Mo

CdS

C

Sputter depth profile of CIGS film stack Rutherford BackScatter profile of CIGS film stack

Acknowledgement: Thermo Electron Corporation

Presenter

Presentation Notes

The XPS depth profile gives a standardless quantification of the CIGS film stack as a function of depth. A comparison of the XPS data with Rutherford Backscattering data from a similar test sample verifies the accuracy of that quantification. (In the case of the RBS, it was not possible to analyze a real product solar cell film, as in the XPS example, but a similar test film deposited onto a wafer was tested.) In both the XPS and RBS data, we can see the composition gradient of the gallium and indium in the CIGS layer, where the two elements cross profiles close to 1.6 microns.


Interfaces in Solar cells


Interface between p-Si/ZnO: Si HF with and without Ar etched (SINTEF SEP 09)

SiO2

Si Zn mixed oxide

Si

RT depos + 300 C annealingRT deposDepos at 500 C Depos at 500 C + Ar etching Depos at RT + Ar etching

No Ar etchingAr etching


Elemental distribution and oxygen deficiency of magnetron sputtered ITO filmsA. Thøgersen, M.Rein, E. Monakhov, J. Mayandi, S. Diplas

JOURNAL OF APPLIED PHYSICS 109, 113532 (2011)

34

XPS depth profiling reveals metallic In and Sn at the interface

High resolution TEM showing metallic In at the interface


Initial stages of ITO/Si interface formation studied with XPS and DFTO.M. Løvvik, S. Diplas, A Romanyuk, A. Ulyashin, JOURNAL OF APPLIED PHYSICS 115, 083705 (2014)

35

• Presence of pure In and Sn, as well as Si bonded to oxygen at the ITO/Si interface wereobserved. • The experimental observations were compared with several atomistic models of ITO/SiInterfaces. • These results support and provide an explanation for the creation of metallic In and Sn alongwith the growth of SiOx at the ITO/Si interface.


Theoretical efficiency of Cu2O/ZnO based solar cell:[1-3]

~10-20 %

Highest experimental efficiency reported: ~ 2 % without buffer layer [2]

~ 1-4% with buffer layer [4,5]

[1] S. Jeong et al. , Electrochim. Acta (2008), 53, 2226 .[2] A. Mittiga et al. Appl. Phys. Lett. (2006), 88, 163502 .[3] T. Gershon , et al., Sol. Energy Mater. Sol., C. (2012), 96,48[4] Z. Duan et al. Solar Energy Materials & Solar Cells 96 (2012)[5]T. Minami et al., Appl. Phys. Exp.4, 062301(2011)

Cu2O as a Potential Solar Cell MaterialNFR project "Heterosolar" (SINTEF-UiO)


Samples Investigated with XPS

Top film nominal thicknesses: 1, 3, 5, 10 nm


AZO/Cu2O AZO/CuO

XPS on ultrathin AZO deposited on Cu2O. The satellite indicates formation of CuO at the interface in agreement with TEM (see next slide). Variations in the satellite intensity as the interface is approached indicate that ZnO influence the electronic structure of the interfacial CuO


STEM-HAADF imaging of Cu2O/ZnO cross-section interface. (A) The overview STEM HAADF image of the Cu2O film grown on O-polar ZnO substrate. (B) The HRSTEM HAADF image of the boxed region indicated

in (A). The different layers in the cross-section are marked by false coloring. (C) and (D) The FFT patterns from the top and the interfacial

layer in (B) respectively were identified as Cu2O and CuO phases.

STEM HHADF and FFT

Documents

X-ray photoelectron spectroscopy - An introductionfolk.uio.no/vebjornb/MENA9510/lectures-2016/MENA9510_XPS_2016.pdf · XPS Depth of Analysis. The probability that a photoelectron