Run Rietveld Refinement and RIR Quantification Online
InstaNANO Online Rietveld Refinement continues directly from XRD phase identification. Select scientifically plausible matched references, open the refinement popup, and use the method supported by the available structural models and reference-intensity-ratio data. The calculation runs in a browser worker and adds the calculated pattern, difference curve, background, and reflection positions to the graph.
The workflow supports full-pattern fitting, model-dependent phase quantification, lattice refinement, optional crystallite-size and microstrain estimates, preferred-orientation correction, fit statistics, calculated HKL reflections, local-residual diagnostics, and a downloadable PDF report. The available controls change with the included phase models so unsupported combinations cannot be selected.
Scientific note: Refinement tests a defined model against the measured pattern. A low Rwp, GoF, or reduced chi-square does not independently confirm phase identity or guarantee that every refined parameter is physically meaningful. Review the phase chemistry, calculated and difference curves, parameter limits, residual regions, and measurement conditions together.
What the Online Refinement Tool Provides
- Direct continuation from the XRD phase-identification results.
- Automatic recommendation of a method compatible with the included references.
- Selection of one to twelve phases for a refinement run.
- Full Rietveld, hybrid Rietveld–RIR, RIR, profile-supported RIR, and semi-quantitative workflows.
- Observed, calculated, difference, background, and phase-reflection displays.
- Phase composition on the selected quantitative basis when supported by the method.
- Refined unit-cell parameters for compatible complete structural models.
- Calculated HKL reflection families, d-spacings, positions, and relative intensities.
- Optional apparent or instrument-corrected crystallite-size and microstrain estimates.
- Optional preferred-orientation refinement for compatible structural phases.
- Rp, Rwp, and, when statistically defined, Rexp, GoF, and reduced chi-square.
- Largest local fit-difference regions and potential unmatched-peak regions for review.
- Refinement-stage history, parameter warnings, and interpretation notes.
- Structured on-screen results and a downloadable PDF report.
- Editable InstaNANO project saving for reopening the supported refinement state.
Available Rietveld and RIR Methods
The method controls are enabled from the data available for the included references. If the selected phases do not support one method together, that method remains unavailable instead of substituting unsupported information.
| Method | Reference requirement | Interpretation |
|---|---|---|
| Full Rietveld QPA | Complete structural model for every included phase | Full-pattern structural fitting with phase fractions on the selected quantitative basis |
| Hybrid Rietveld–RIR | Compatible mixture of complete structural models and RIR-supported phases | Combines structural-profile fitting with RIR-supported quantification |
| RIR QPA | Compatible RIR data for every included phase | Estimates phase fractions from reference intensity ratios |
| RIR + Profile Fit | Compatible RIR data and supported partial structural profiles | Uses profile fitting to support RIR-based phase quantification |
| Semi-Quantitative Fit | Supported complete or partial profile models | Reports relative phase coefficients; these should not be interpreted as weight fractions |
How to Run an Online XRD Refinement
- Import and inspect the experimental XRD pattern.
Upload a supported XRD or tabular file, confirm the 2θ range and measured intensities, and retain the original experimental series.
- Run XRD phase identification.
Open XRD Match, verify the wavelength, review the automatically selected peaks, adjust them when scientifically necessary, and run Match XRD Pattern.
- Select plausible candidate references.
Compare reference peaks with the complete measured pattern and include only phases consistent with the sample composition and experimental context.
- Open the refinement popup.
Click Run Rietveld Refinement. The popup lists the selected references, their model and RIR availability, and the compatible method recommended for that combination.
- Define the fit and pattern settings.
Confirm the fit range, excluded ranges, background, peak profile, asymmetry, emission spectrum, and weighting. Automatic settings provide a starting point, while explicit settings remain available when the measurement conditions are known.
- Select supported phase and correction controls.
Choose lattice, size, strain, preferred orientation, instrument profile, and peak-position options only when they are scientifically justified. Phase controls allow compatible parameters to be enabled or fixed separately for each included reference.
- Choose the quantification basis and run the calculation.
Use crystalline phases = 100 wt% when the reported fractions should be normalized to the included crystalline phases. Select the internal-standard basis only when a known standard and its measured-mixture weight fraction are available.
- Review the pattern, report, and warnings.
Inspect the observed, calculated, and difference curves; check the phase composition, calculated reflections, fit statistics, refined parameters, local differences, unmatched regions, and interpretation notes before downloading the PDF report.
Refinement Controls and Their Purpose
| Control | Purpose | Important consideration |
|---|---|---|
| Fit range and excluded ranges | Define which measured observations contribute to the fit | Exclude regions only for a documented experimental reason |
| Background and polynomial degree | Model the slowly varying non-peak contribution | An unnecessarily flexible background can absorb real diffraction intensity |
| Peak profile and asymmetry | Represent symmetric broadening and optional low-angle axial-divergence behavior | Select a model consistent with the instrument geometry and observed peak shapes |
| Emission spectrum | Use Kα1 alone or the Kα1 + Kα2 doublet | The selection should match the measured radiation treatment |
| Lattice parameters | Adjust compatible unit-cell dimensions and angles | Large departures from the reference cell require scientific review |
| Crystallite size and microstrain | Represent sample-related contributions to peak broadening | Without a calibrated instrument profile, reported values are apparent estimates |
| Preferred orientation | Apply a phase-specific correction about an automatic or entered hkl axis | Use only when the intensity distortion is physically plausible for the specimen |
| Instrument profile | Use automatic profile refinement or a wavelength-compatible calibrated InstaNANO profile | A calibrated profile separates supported instrumental and sample broadening contributions |
| Peak-position correction | Apply no shift, a constant zero shift, or Bragg–Brentano sample displacement | Choose the correction that corresponds to the measurement geometry and likely error source |
| Weighting | Use Poisson weighting for suitable raw counts or uniform weighting for arbitrary intensities | Rexp, GoF, and reduced chi-square are statistically interpretable only when Poisson weighting is appropriate |
| Evaluation target and tolerance | Control the nonlinear search budget and convergence threshold for each stage | A stage reaching its full target can be rerun with a larger target after the model is reviewed |
| Quantification basis | Normalize included crystalline phases or use a known internal standard | Unmodelled or amorphous content requires a valid internal-standard workflow |
Understanding the Refinement Report
| Report section | What it shows | How to interpret it |
|---|---|---|
| Phase composition | Reference, phase, composition, approximate uncertainty when estimable, and refined or fixed unit cell | The result applies to the included models and selected quantification basis |
| Calculated reflections | Unique calculated hkl families at or above 1% of the strongest calculated reflection for each phase | These are model assignments; overlapping reflections may prevent a unique experimental-peak assignment |
| Rp and Rwp | Profile residual measures for the observed and calculated patterns | Lower values indicate closer profile agreement but do not alone validate the phase model |
| Rexp, GoF, and reduced chi-square | Counting-statistical fit measures when Poisson weighting is selected | Interpret statistically only when the measured intensities are appropriate raw counts |
| Refined parameters and stages | Profile, cell, size, strain, orientation, correction, and convergence information when applicable | Review warning messages and parameters that settle at an allowed limit |
| Largest local fit differences | Regions with the largest observed-minus-calculated differences | Use them to inspect local mismatch; they do not independently prove another phase |
| Potential unmatched peaks | Positive residual regions without a nearby modeled reflection | Review for missing phases, model limitations, artifacts, or measurement effects |
Instrument Profiles, Crystallite Size, and Microstrain
Observed peak breadth can contain both instrumental and sample contributions. When the instrument profile is left on Automatic, the tool can fit the measured peak shape, but crystallite-size and microstrain outputs are labelled as apparent because instrumental broadening has not been independently fixed.
When a valid wavelength-compatible .instanano-profile file is loaded, its calibrated profile parameters are held as the instrumental contribution and the compatible sample-broadening terms can be estimated separately. This improves the physical basis of size and strain interpretation, but it does not remove the need to check specimen preparation, anisotropy, peak overlap, and model suitability.
Weighting and Chi-Square Values
Poisson (raw counts) weighting uses the expected counting variance and permits the tool to report Rexp, GoF, and reduced chi-square. These statistics are meaningful only when the imported intensities can reasonably be treated as raw counts and the uncertainty model is appropriate.
Uniform (arbitrary intensity) weighting gives every fitted observation equal weight. Rp and Rwp remain available, but Rexp, GoF, and reduced chi-square are reported as unavailable because uniform weighting does not supply counting-statistical uncertainties.
A GoF near one is not a target that should be reached by adding parameters without physical justification. If the fit is unsatisfactory, first review the phase selection, wavelength, fit range, background, peak profile, instrument calibration, specimen-displacement correction, preferred orientation, and unmodelled peaks. Optional parameters should be introduced only when supported by the experiment and the residual pattern.
Credit Access and Saved Projects
- Refinement uses one credit when the same XRD sample already has an active complete matching unlock.
- Starting refinement directly from a locked single-sample match uses two credits because matching and refinement access are opened together.
- Repeating the refinement for the same unchanged sample during its active access period does not use another credit.
- Access remains available for up to 30 days from the original sample unlock, subject to the applicable account and plan validity.
- Experimental data and selected matching peaks remain locked together after unlock so a different sample or peak set is treated as a new analysis.
- Saving the
.instananoproject preserves the supported graph, matching, entitlement, settings, and refinement-result state for later continuation.
Preparing Data for Responsible Refinement
- Confirm that 2θ values increase and that the intensity column represents the intended sample.
- Select the experimental wavelength used for the measurement.
- Retain the original measured pattern when applying baseline correction or smoothing.
- Avoid smoothing that changes peak positions, widths, or relative intensities.
- Inspect the entire fitted range rather than only the strongest reflections.
- Include only phases supported by the chemistry, synthesis history, and diffraction pattern.
- Use excluded ranges only for known artifacts or regions outside the intended model.
- Use raw-count Poisson weighting only when the intensity scale supports that assumption.
- Use a compatible calibrated instrument profile for stronger crystallite-size or microstrain interpretation.
- Record the included phases, fit range, weighting, background, profile, corrections, and refined parameters.
Scientific Scope and Limitations
Results can be affected by missing phases, amorphous material, peak overlap, preferred orientation, anisotropic broadening, fluorescence, specimen displacement, imperfect background modelling, instrumental broadening, and reference-model limitations. Quantitative values describe the selected model under the stated assumptions; they should be checked against the residual pattern and independent knowledge of the specimen.
Scientific References for Method Interpretation
- GSAS-II laboratory X-ray refinement tutorial for staged background, scale, lattice, displacement, size, and microstrain refinement.
- GSAS-II constant-wavelength instrument-profile tutorial for separating calibrated instrumental broadening from sample broadening.
- NIST Special Publication 846 for reference-intensity-ratio methods and quantitative powder-diffraction analysis.
Frequently Asked Questions
Can I run Rietveld refinement online without installing software?
Yes. The workflow runs in a modern desktop browser and continues from the InstaNANO XRD matching results. No separate desktop refinement package is required for the supported methods and controls.
Do I need to identify phases before refinement?
Yes. Refinement requires one or more selected candidate references. Run XRD matching first, verify the candidate phases against the sample chemistry and measured pattern, and then open the refinement popup.
How many phases can be included?
One to twelve phases can be included in a refinement run. Additional selected references remain available in the matched-reference list but cannot be included beyond this limit.
Why are some refinement methods or controls disabled?
Availability depends on the selected references and method. Complete structural models are required for full Rietveld lattice and preferred-orientation controls, while RIR methods require compatible reference-intensity-ratio data. Unsupported combinations remain disabled.
Why does uniform weighting show N/A for Rexp, GoF, and reduced chi-square?
Uniform weighting does not define counting-statistical uncertainties. Use it for arbitrary intensity units and interpret Rp and Rwp. Poisson weighting reports the additional statistics, but they are meaningful only when the intensities are suitable raw counts.
Does a reduced chi-square near one prove that the model is correct?
No. It indicates agreement relative to the assumed statistical weights. Phase identity, systematic error, model completeness, parameter correlations, and the local difference curve still require scientific review.
Can the tool estimate crystallite size and microstrain?
Yes, when enabled for compatible phases. Without a calibrated instrument profile, the report labels these values as apparent estimates. A compatible calibrated profile provides a stronger basis for separating instrumental and sample broadening.
Can I estimate amorphous or unmodelled content?
The internal-standard basis can estimate an amorphous or unmodelled fraction when a valid internal-standard phase and its known weight percentage in the measured mixture are supplied. Normalizing crystalline phases to 100 wt% does not determine amorphous content.
Can I download a Rietveld refinement report?
Yes. Completed results can be reviewed in the popup and downloaded as a PDF containing the method settings, phase composition, calculated reflections, fit statistics, refined parameters, stage history, residual diagnostics, warnings, and interpretation guidance.
Can the results support publication preparation?
The report documents the supported refinement settings and outputs for review and reporting. Researchers should verify the model, disclose the relevant experimental and refinement conditions, retain the original data, and follow the requirements of the target journal. No single fit statistic establishes publication quality.
Online Rietveld Refinement Tool at a Glance
- Tool: InstaNANO Online Rietveld Refinement and RIR Phase Quantification
- Workflow: XRD plotting, candidate-phase matching, compatible method selection, refinement, interpretation, and PDF reporting
- Methods: Full Rietveld QPA, Hybrid Rietveld–RIR, RIR QPA, RIR + Profile Fit, and Semi-Quantitative Fit
- Included phases: One to twelve references per run
- Fit controls: Range, exclusions, background, profile, asymmetry, emission, corrections, weighting, evaluation target, and tolerance
- Phase controls: Lattice, crystallite size, microstrain, and preferred orientation when supported
- Quantification: Included crystalline phases normalized to 100 wt% or an internal-standard basis
- Output: Calculated pattern, difference curve, phase composition, calculated reflections, fit statistics, residual diagnostics, and PDF report
- Project format: Editable
.instananoproject for supported saved state - Platform: Modern desktop browsers on Windows, macOS, and Linux
Start with the experimental XRD pattern above. Identify plausible candidate phases, include the compatible references, review the recommended method and settings, and interpret the full calculated and difference pattern before reporting the result.