Normalization

Normalization is a critical step in XAS data analysis that removes systematic variations and isolates a normalized XANES edge and oscillatory EXAFS signal. EstraPy provides three commands that work together to perform complete XAS normalization:

  1. preedge - Removes the pre-edge background
  2. postedge - Removes the post-edge background (atomic absorption)
  3. normalize - Normalizes the signal to the edge jump

Normalization Workflow

The typical normalization workflow follows these steps in order:

# 1. Remove pre-edge background
preedge .. -50eV

# 2. Remove post-edge atomic absorption
postedge 150eV ..

# 3. Normalize to edge jump
normalize --factor J0

After normalization, you’ll have:

  • mu - Normalized absorption coefficient (\(\mu(E)/\mu_0(E_0)\))
  • chi - EXAFS oscillations (\(\chi(E) = \mu(E)/\mu_0(E_0) - 1\))

What Each Command Does

Pre-edge Correction

The preedge command fits a polynomial to the absorption before the edge and subtracts it. This removes:

  • Instrumental baseline drift
  • Absorption from other edges at lower energies
  • Linear or curved backgrounds

Result: Clean absorption spectrum with zero baseline before the edge.

Post-edge Correction

The postedge command fits a polynomial to the absorption well above the edge and removes the smooth atomic absorption (\(\mu_0(E)\)). This can be done by:

  • Subtraction: \(a(E) - \mu_0(E) + J_0\)
  • Division: \(a(E) / \mu_0(E) \times J_0\)

where \(J_0\) is the edge jump height (stored as metadata variable J0).

Result: Spectrum corrected for atomic absorption, with an edge jump standardized to \(J_0\).

Jump Normalization

The normalize command divides the signal by the edge jump \(J_0\) to create standardized spectra:

  • mu = \(a(E) / J_0\) (normalized absorption)
  • chi = \(a(E) / J_0 - 1\) (EXAFS oscillations)

Result: Dimensionless absorption coefficient and EXAFS signal ready for Fourier analysis.

Note that the a column is not modified by normalize and retains the original absorption signal. The normalized mu and chi columns are created as new columns in the dataset.

Domain and Column Requirements

All normalization commands operate in the reciprocal domain and require:

  • Axes: E (energy), e (relative energy), k (wavevector)
  • Data: a (absorption signal)
  • Metadata: E0 (edge energy)

The edge energy E0 must be set by the edge command for normalization to work properly.

See Also

Next: Choose a specific normalization command to learn more about its options and usage.

Table of contents