# XPS peak fitting online (C 1s, O 1s, N 1s deconvolution)

> Fit an XPS region online: chemical states as pseudo-Voigt GL(30) peaks of one FWHM, binding energy, area % and FWHM ± SE. Free, in your browser.

URL: https://www.labcanvas.com/tools/xps-peak-fitting

Paste a charge-corrected XPS region (binding energy and counts), list the chemical states you expect (for C 1s: C–C/C–H 284.8, C–O 286.3, C=O 287.8, O–C=O 289.0 eV) or let the tool find the peaks, and it fits them together with a linear background as pseudo-Voigt GL(30) peaks of one shared FWHM, the usual constraints for a C 1s region. You get each component’s binding energy, area, % of the total area and FWHM ± SE, drawn as filled components on a reversed binding-energy axis. Atomic ratios between elements also need the relative sensitivity factors.

## Quick answer

- **What it does:** Deconvolutes an XPS core-level region into chemical-state components (pseudo-Voigt, one shared FWHM, linear background), giving binding energies, areas, area percentages and FWHM ± SE.
- **Formula:** I(E) = b₀ + b₁E + Σ Aₖ[η Lₖ(E) + (1 − η) Gₖ(E)], one FWHM for all k, η = 0.3 (GL(30)); % of component k = Aₖ / Σ A
- **Example:** The built-in C 1s region of an oxidised polymer (142 points) fitted with four GL(30) peaks of one FWHM gives C–C/C–H at 284.80 eV (59.2 %), C–O 286.29 eV (22.0 %), C=O 287.79 eV (8.6 %) and O–C=O 289.02 eV (10.2 %), FWHM 1.206 ± 0.003 eV, R² = 0.9997.

Free, in the browser, no account needed. Updated 2026-11-17.

## Features

- Paste binding energy and counts, or drop a CSV, TXT or XY export of one region; the energy axis runs high to low
- Your list of chemical states (position and label) as starting values, or peaks found automatically
- Pseudo-Voigt with a fixed Lorentzian fraction (GL(30) is η = 0.3) or fitted, Gaussian, Lorentzian or Voigt
- One FWHM shared by every component (on by default), as usual for the states of one core level
- Linear, constant or quadratic background fitted with the peaks; fit only a binding-energy window
- Binding energy, height, area, % of area and FWHM ± SE, 95% CIs, R², reduced χ², AIC
- Copy the component table (pastes into Excel) and a methods sentence; PNG at 300 or 600 dpi; CSV

## How to use it

1. **Paste or drop your XPS region**: Two columns, binding energy (eV) and intensity, for one core level, charge-corrected (for example adventitious C 1s at 284.8 eV).
2. **Choose the line shape**: Pseudo-Voigt with η = 0.3 matches the GL(30) shape used in CasaXPS; leave η empty to fit it.
3. **List the chemical states or find peaks**: Keep the C 1s list or type your own positions with labels, one per line. The fit moves each centre; tick “one FWHM” off only with a reason.
4. **Check and report**: Look at the residuals and the component positions, then copy the table of binding energies and area percentages.

## At a glance

| | |
| --- | --- |
| Input | Binding energy and counts of one region (CSV, TXT, XY) |
| Shape | Pseudo-Voigt GL(30) by default; Gaussian, Lorentzian, Voigt |
| Constraints | One FWHM for all components; positions from your list |
| Background | Linear, constant or quadratic, fitted with the peaks (no Shirley) |
| Output | Binding energy, area, % area, FWHM ± SE; 95% CIs; R², AIC |
| Engine | The LabCanvas Graph app’s peak fit (XPS preset), checked against lmfit |

## How it works

$$I(E) = b_0 + b_1 E + \sum_k A_k \left[\eta\, L(E; c_k, w) + (1-\eta)\, G(E; c_k, w)\right]$$

$$\%_k = \frac{A_k}{\sum_j A_j}$$

**Components.** Each chemical state of one core level is a peak at its binding energy. The usual constraints keep a fit physical: one FWHM for all components of a region (they share the instrument resolution and similar lifetimes), a fixed line shape, and positions that start from known chemical shifts. For C 1s in organic material, typical positions are C–C/C–H 284.8, C–O 286.3, C=O 287.8 and O–C=O 289.0 eV; the fit refines each centre.

**Line shape.** The pseudo-Voigt \(\eta L + (1-\eta) G\) with one width approximates the Voigt profile of a core level; η = 0.3 corresponds to the GL(30) shape common in CasaXPS. Asymmetric metal and graphitic sp² peaks need an asymmetric shape this tool does not offer.

**Background.** The background here is a polynomial (usually a straight line) fitted with the peaks. The Shirley background (Shirley 1972), which rises in proportion to the integrated intensity at higher kinetic energy, is the standard for most regions and is not offered; over a narrow C 1s window with similar levels on both sides a linear background gives similar areas, but over wide regions or metal peaks use Shirley in your XPS software.

**From areas to composition.** Area percentages within one region are fractions of that element’s atoms in each state. Atomic ratios between elements need each area divided by its relative sensitivity factor (photoionisation cross-section, Scofield 1976, with the analyser transmission).

**Same numbers as the Graph app.** The page runs the LabCanvas Graph app's peak-fit analysis (its XPS preset), which is checked against lmfit; the hand-off opens the same table and fit there.

Sources: Shirley DA. *Phys Rev B* 1972;5:4709, [doi:10.1103/PhysRevB.5.4709](https://doi.org/10.1103/PhysRevB.5.4709). Scofield JH. *J Electron Spectrosc Relat Phenom* 1976;8:129, [doi:10.1016/0368-2048(76)80015-1](https://doi.org/10.1016/0368-2048(76)80015-1). Newville M et al. LMFIT: non-linear least-squares minimization and curve-fitting for Python. Zenodo, [doi:10.5281/zenodo.11813](https://doi.org/10.5281/zenodo.11813) (line shapes; area = amplitude). Olivero JJ, Longbothum RL. Empirical fits to the Voigt line width. *J Quant Spectrosc Radiat Transf* 1977;17:233–236, [doi:10.1016/0022-4073(77)90161-3](https://doi.org/10.1016/0022-4073(77)90161-3).

## Worked example: C 1s of an oxidised polymer surface

The built-in example is a simulated C 1s region with noise, not a measurement: four components (C–C/C–H 284.8, C–O 286.3, C=O 287.8, O–C=O 289.0 eV) as GL(30) peaks of one FWHM on a linear background, 142 points.

| Component | Binding energy (eV) | Area | % of area |
| --- | --- | --- | --- |
| C–C/C–H | 284.798 ± 0.001 | 7367 ± 17 | 59.2 |
| C–O | 286.289 ± 0.004 | 2738 ± 13 | 22.0 |
| C=O | 287.794 ± 0.012 | 1066 ± 14 | 8.6 |
| O–C=O | 289.017 ± 0.009 | 1263 ± 15 | 10.2 |

- **Shared width.** All four share FWHM = 1.206 ± 0.003 eV, so the fit has 4 centres, 4 areas, one width and two background terms: 11 parameters for 142 points.
- **Chemical shifts.** Relative to C–C at 284.80 eV, C–O is +1.49 eV, C=O +2.99 eV and O–C=O +4.22 eV.
- **Composition.** 59.2 % of the carbon is unoxidised; 22.0 + 8.6 + 10.2 = 40.8 % is bound to oxygen.
- **Fit quality.** R² = 0.9997 with residuals scattered around zero.



## Questions

### How do I fit an XPS C 1s spectrum?

Charge-correct the energy scale, subtract a background, and fit components at the expected chemical states (C–C/C–H 284.8, C–O about 286.3, C=O about 287.8, O–C=O about 289.0 eV) with one shared FWHM and a GL(30) line shape. Check that the residuals are flat and the positions are chemically sensible.

### What is GL(30)?

A product or sum of Gaussian and Lorentzian line shapes with 30 % Lorentzian character, the default in CasaXPS. Here it is the pseudo-Voigt sum with η = 0.3.

### Should all components have the same FWHM?

Usually yes for the chemical states of one core level measured in one region: they share the instrument resolution. Free widths let the fit trade width for area and give unphysical results; relax the constraint only with a reason, such as a differently charged phase.

### Which background should I use for XPS?

Shirley for most core levels, Tougaard for quantitative work over wide ranges, linear for narrow regions with flat levels on both sides. This tool fits a linear (or quadratic) background with the peaks; for a Shirley background use your XPS software.

### How do I get atomic percentages from XPS peak areas?

Within one region, the area percentages are the fractions of that element in each chemical state. Between elements, divide each region’s area by its relative sensitivity factor, then normalise the results to 100 %.

### How do I charge-correct XPS data?

Shift the whole binding-energy scale so a reference peak sits at its accepted value, most often adventitious carbon C–C/C–H at 284.8 eV, and apply the same shift to every region of that sample.
