Standard x-ray sources for lab based instruments are Al Kα or Mg Kα which provide photons with 1486.6eV and 1254.6eV energy respectively. Here we demonstrate the use of a monochromatic, Ag Lα x-ray source which uses the second order diffraction of the same quartz crystals as used for Al Kα, allowing it to be easily accommodated on a modern spectrometer.
The photon energy of Ag Lα is 2984.2 eV, approximately twice that of Al Kα (1486.6 eV). This greater energy leads not only to the excitation of additional, higher binding energy, core lines for some elements (Table 1) but also the possibility of analysis of deeper layers due to the decrease in attenuation length with increasing photoelectron energy . For example, the inelastic mean free path (IMFP) of C 1s, binding energy 285 eV, using Al Kα X-rays is 3.14 nm at a kinetic energy (KE) 1201 eV, but for Ag Lα this is increased to 5.89 eV at KE 2699 eV .
Core Line Element
- 1s Al Si P S Cl
- 2s As - Mo
- 2p Br - Ru
- 3s Pr - Re
- 3p* Sm - Tl
- 3d* Tm - Rn
Table 1: Elements with core line binding energies in the range of Ag Lα X-rays but not Al Kα (*at least one half of doublet) 
In the following example we present data from a single crystal MoS2 sample. In figure 1 spectra are presented from the 'as introduced' using the Al Kα x-ray source (red) and Ag Lα (black) monochromatic x-ray sources with the spectra normalised to the Mo 3d peak intensity. It is immediately apparent that the higher Mo 2p and S 1s peaks are excited using the higher photon energy.
Figure 1: Survey spectra from 'as introduced' MoS2 a single crystal sample using Al Kα (red) and Ag Lα (black) x-rays. Inset shows expanded low BE range with the two spectra normalised to Mo 3d peak intensities.
After characterisation of the 'as introduced' sample the surface was etched using the Ar gas cluster ion source using Ar1000+ at 10 keV beam energy to try to minimise surface damage relative to sputtering with monoatomic Ar+ ions. The schematic diagram below shows the interaction of a gas cluster with an ideal surface where the damage induced by the impact is localised at the surface. Also shown on the diagram is the Mo 3d photoelectron escape depth for electrons excited with the Al Kα (red) and Ag Lα x-rays (blue). Due to the greater kinetic energy of the Mo 3d photoelectrons excited using the Ag Lα the sample depth is almost 2 times that of the Al Kα excited spectrum. The Mo 3d spectra shown in figure 2 are both acquired from the same sample after a 20 s etch cycle. It is noted that each of the Mo 3d spin-orbit split components show a doublet, relating to the chemistry of the sample. The curve fits of these envelopes is shown in figure 3.
Figure 2: Schematic diagram of Arn+ ion cluster impact with a surface and relative sampling depths using the two x-ray sources. The spectra show the Mo 3d/S 2s region recored using Al Kα (red) and Ag Lα (black) x-rays.
The Mo 3d envelope has been fitted to two chemical states corresponding to the undisturbed MoS2 where the Mo 3d5/2 peak appears at 229.2 eV binding energy and the reduced metallic-like peak at 228.1 eV binding energy. For the Al Kα excited spectrum the ratio of Mo(IV) to Mo(0) is 0.43 which contrasts to a ratio of 1.25 for the Ag Lα excited spectrum. The difference in the ratio is explained by the increased sampling depth for the Ag Lα such that more of the undisturbed sub-surface is sampled with the higher excitation energy source. Interestingly it is noted that he cross section of the S 2s relative to the Mo 3d is significantly different for the two x-ray sources.
Figure 3: Curve fits for the Mo 3d/S 2s envelope of sputtered MoS2 using the Al Kα and Ag Lα excitation sources.
 K Yates & RH West, Surf. Interface Anal., 1983, 4, 5
 MP Seah & WA Dench, Surf. Interface Anal., 1979, 1, 2
 X-Ray Data Booklet LBNL/PUB-490 Rev.2 Jan 2001