Contact Us
→Site Map



Synchrotron-based and laboratory-scale hard X-ray photoelectron spectroscopy (HAXPES) of deep-layer interfaces

X-ray photoelectron spectroscopy (XPS) is a surface analysis technique that is commonly used to measure the chemistry and morphology of sample surfaces. While laboratory-scale XPS (Al K-alpha E = 1.5 keV) can measure the elemental composition of the surface (top 3 nm or so) of the material being analyzed, laboratory-scale HAXPES (Ag L-alpha E = 3 keV) and synchrotron-based HAXPES (SPring-8: 8 keV) have the ability to measure deeper layers of multilayered samples and actual devices in their near as-received states.


■ The higher the excitation energy,
the higher the kinetic energy of photoelecrtrons.
⇒The detection depth increases.


HArd X-ray PhotoElectron Spectroscopy

HAXPES comment
HAXPES graph Lab-scale HAXPES graph
SPring-8-Based HAXPES Instrument
Laboratory-Scale HAXPES Instrument
Laboratory-Scale XPS Instrument
SPring-8-Based HAXPES Instrument
Laboratory-Scale HAXPES Instrument (Ag Lα)
Laboratory-Scale XPS Instrument (Al Kα)
Courtesy: Japan Synchrotron Radiation Research Institute (JASRI) (SPring-8 HAXPES Instrument)


HAXPES is X-ray photoelectron spectroscope (XPS) that uses high-energy X rays.

HAXPES has deeper detection limit than laboratory-scale XPS, and thus can measure photoelectrons ejected from deeper energy levels. TNA utilizes SPring-8, a third-generation synchrotron radiation facility, to perform depth profiling that requires a high-intensity X-ray source.

The detection limit of XPS is a function of the energy of the X-ray photons used. Laboratory-scale HAXPES with an argon (Ag) L-alpha X-ray source and 8-keV synchrotron-based HAXPES provide the chemical state of elements approximately twice and four times deeper below the sample surface respectively than laboratory-scale XPS with an aluminum (Al) K-alpha X-ray source. Additionally, HAXPES simplifies chemical-state analysis since Auger electron spectra from HAXPES do not overlap.

Generally, the use of high-energy X-rays reduces the excitation efficiency. However, SPring-8-based HAXPES is free from this problem because it relies on high-intensity synchrotron radiation as an X-ray source, and laboratory-scale HAXPES with an Ag L-alpha X-ray source improves the yield rate of photoelecrtrons by using an electron detector with a magnetic lens.
Furthermore, since HAXPES can excite inner-shell orbitals (e.g., from Si2p to Si1s), its detection sensitivity is not a concern for certain elements.
System comparisons

Comparisons between laboratory-scale XPS and HAXPES

• HAXPES — Si1S photoelecrtrons (Ek: up to 6.1 keV) are measured with 8-keV excitation energy
• Laboratory-scale XPS — Si2p photoelecrtrons (Ek: up to 1.4 keV) are measured using Al K-alpha X-rays with 1.5-keV excitation energy
Comparison between laboratory-scale XPS and HAXPES: HAXPES allows detection of a buried interface and is less affected by the surface layer.
1) IMPS: Inelastic Mean Free Path

Since it is necessary to reserve the use of SPring-8, it takes a while to perform XPS analysis using 8-keV HAXPES.
Although the detection depth supported by laboratory-scale HAXPES (4-keV Ar L-alpha) is less than 8-keV synchrotron-based HAXPES, laboratory-scale HAXPES allows timely analysis of the top few nm of the sample surface, such as a layer beneath a cap layer. Therefore, laboratory-scale HAXPES is routinely used prior to more detailed analysis using synchrotron-based HAXPES.

Example: Measurement of a multilayered sample that consists of metal and insulation layers (HAXPES)

• Because radiation from HAXPES reaches deep below the sample surface, it has the ability to see aluminum (Al) oxides through the top cap layer (with a thickness of approx. 10 nm).
• This is difficult for laboratory-scale XPS to achieve since it cannot excite deep electron energy levels of the atom (e.g., Al1s). Additionally, HAXPES provides support for wider cross-sections (i.e., sensitivity), and it can measure even thin films and trace elements in a short time.
• As HAXPES can see through the cap layer, oxidizable and deliquescent materials can be measured without being affected by exposure to the atmosphere.
(Multilayered samples can be measured in their as-received states.)

System comparisons

HAXPES (SPring-8)

Laboratory-Scale XPS
Laboratory-Scale XPS

Example: Analysis of Al distribution in a CMOS gate stack using HAXPES

NiSi (Al) (a few tens of nm) — HfSiON (a few nm) — Si substrate
• The sample is obliquely etched, and regions with uneven NiSi film thicknesses are measured using HAXPES.(Chemical-state analysis reveals the presence or absence of aluminum in an NiSi layer over a region with different thicknesses of the remaining NiSi layer)

• The chemical-state analysis shows aluminum distribution at the atomic level.
Al distribution profiling example (HAXPES)
The aluminum distribution at layer interfaces can be estimated as a result of Al1s spectrum analysis.
■ Al2O3>: Distributed in the HfSiON layer
■ Al metal: Distributed in the NiSi layer
■ AlOx (X < 1.5): Distributed at the NiSi-HfSiON interface
Al distribution profiling example (HAXPES)
Al distribution profiling example (HAXPES)
This data was printed in IEDM (2008).

Evaluation of band bending in p-GaN and p+-GaN (HAXPES)

—Bending depth profiling


Bending depth profiling
Yoshiki et al., Journal of the Japanese Society for Synchrotron Radiation Research, 22,20(2009).  
The above figure shows that band bending differs between p-GaN and p+-GaN.
The magnitude of contact resistance can be compared by evaluating the band alignment with the formed metal contacts.

Example: Spectral of Co2p and Fe2p in a CoFe sample (Comparison between laboratory-scale HAXPES and XPS)

— Avoiding an overlap of Auger peaks


Bending depth profiling
Laboratory-Scale XPS
Laboratory-Scale HAXPES

Bending depth profiling
Laboratory-Scale XPS
Laboratory-Scale HAXPES
For accurate chemical-state analysis, X-ray energy can be adjusted to shift the Auger peak to avoid overlapping with the photoelectron peak.


  • Bulk analysis of samples with thick films (rechargeable batteries, etc.)
  • Analysis of interfaces in multilayered films such as laminated films (gate stacks, GMR, MRAM)
  • Analysis of post-oxidizable and deliquescent layers beneath the cap layer
  • Analysis of the in-gap state of transparent electrodes in solar cells and transparent semiconductors (ZnO, IGZO, etc.)
  • Analysis of the band alignment and bending of electrodes and GaN structures
  • Analysis of the overlapping of the 2p and 2s orbitals, and samples with overlapping Auger peaks
  • High-throughput depth profile analysis using HAXPES under total reflection condition
  • Screening analysis (using laboratory-scale HAXPES) prior to an analysis using SPring-8

[Last updated: February 12, 2015]

Page top