iPowderAI-powered XRD Analysis
AI for XRD

Bring AI into
your XRD workflow.

Understand the goal, call scientific tools, check intermediate results and decide what to do next.

iPowder brings language models, XRD tools, databases and knowledge resources into one environment. Start with a scientific question and let the AI Agent connect the analysis steps.

See how AI analyzes data ↓
Start with a question
“Analyze these data and tell me
which phases are present.”
  1. 01
    Ask a question Describe the sample and your goal
  2. 02
    AI interprets the request Use data and project context
  3. 03
    Call scientific tools Peak finding, search, refinement and result checks
  4. 04
    Produce analysis results Explain the evidence and suggest next steps

Connect natural language with scientific computation.

01 / How to use

Say what you want to analyze.

Start with a scientific question. AI selects tools and organizes the workflow, while the graphical interface remains available.

Questions you can ask
“Identify the phases in these data.”
“Continue with Rietveld refinement.”
“Why is this peak not fitted well?”
“Check whether a phase is missing.”
“Compare unit-cell parameters across these samples.”
Beginners

Learn while analyzing your data

You do not need to master every menu and parameter first. Ask what to do next, what parameters mean and whether a result is reasonable as you learn XRD methods.

Experienced users

Spend less time setting up and more time interpreting

Let AI organize repetitive steps, use scientific tools directly, or return to the graphical interface to control parameters. Focus on samples, structures and scientific questions.

Scientific capability with a simpler way to work.

02 / Scientific foundations

Understand questions and analyze data
in a dedicated XRD environment.

Language models interpret requests; domain knowledge, databases, algorithms and your data support the analysis.

DATABASE

Crystal structures & Data

  • Crystal structure databases
  • CIF structure data
  • Current experiments and analysis history
KNOWLEDGE

XRD knowledge & Rules

  • XRD knowledge and phase-analysis rules
  • Rietveld refinement knowledge
  • Peak profiles, peak finding and residual stress methods
TOOLS

Scientific algorithms

  • Peak finding, search and multiphase analysis
  • Refinement, unit-cell and stress calculations
  • 2D integration and crystal structure calculations
Language model / AI Agent

Interpret & Organize

Interpret requests, select methods, plan workflows, call tools, inspect intermediate results and explain conclusions.

Scientific algorithms / Databases

Execute & Calculate

Automatic peak finding, search–match, multiphase identification, Rietveld refinement, unit-cell calculations, residual stress, 2D integration and crystal structure calculations.

General-purpose language model XRD domain knowledge Crystal structure databases Scientific algorithms
03 / Agent workflow

Six connected stages of analysis.

From understanding the goal to explaining results, AI adjusts subsequent steps using intermediate findings.

  1. 01

    Understand the question

    Use experimental data and project context to identify the sample, analysis goal and scientific question.

  2. 02

    Plan the workflow

    Arrange data checks, preprocessing, peak finding, search and candidate evaluation; continue to refinement where needed.

  3. 03

    Call scientific tools

    Select CNN, DoubleQ, DecisionTree, IP-Match, crystal databases and Rietveld tools.

  4. 04

    Check intermediate results

    Inspect unexplained peaks, candidate phases and unusual residuals to decide whether to change the model or continue.

  5. 05

    Iterate

    Search again, change candidates, add structure models, refine further or inspect the difference pattern.

  6. 06

    Explain results

    Explain the evidence, identify uncertainties and recommend further analysis.

04 / Typical application

AI + Rietveld
Make refinement easier to understand.

Combine refinement knowledge with current results to organize the workflow, inspect the fit and decide what to do next.

Ask about your current results
“Why has Rwp stopped decreasing?”
“Which parameter should I refine next?”
“What might cause this difference peak?”

Parameter release, fit quality, difference peaks and missing phases can each guide the next iteration.

What can AI help with?

  • Suggest a parameter-release sequence and run refinement
  • Track Rwp, Rp and χ²
  • Inspect difference patterns and possible missing phases
  • Check abnormal parameters
  • Suggest the next refinement step

AI organizes refinement; scientific refinement algorithms perform the calculations.

05 / Results & Review

Results supported by computation.
A process you can inspect.

Scientific values come from algorithms, databases and computational modules. AI interprets, plans, calls tools, checks and explains.

Values from scientific computation

Peak positions, cell parameters, phase matches, refined parameters, phase fractions, residual stress and 2D integration results come from their respective scientific modules.

AI selects the methods; scientific algorithms calculate the results.

Traceable analysis

  • Which tools and steps were used
  • Which databases and data were used
  • Which intermediate results were obtained
  • Why the workflow continued and how conclusions were reached

Review input data, parameter settings, databases and fit diagnostics to support traceability, verification and reproducibility.

06 / Analysis example

Connect scientific tools into
one complete analysis path.

The user sets the goal
“Analyze this sample and check for missing phases.”

Starting with phase identification, AI connects structure models, whole-pattern refinement and difference checks, then evaluates the next step.

Workflow illustration: tools and their order depend on the data and analysis goal.
  1. 01Check data quality
  2. 02Find peaks automatically
  3. 03Multiphase search–match
  4. 04Select candidate crystal structures
  5. 05Build a Rietveld model
  6. 06Automatic refinement
  7. 07Inspect the difference pattern
  8. 08Check for unexplained peaks
  9. 09Report results and recommend next steps
Learning & Efficiency

Beginners learn during analysis.
Experts delegate repetitive work to AI.

For beginners

Ask → Analyze → Understand

Ask about where to start, parameter meanings, next steps or the reliability of results. iPowder helps you learn XRD through your own data.

For experienced users

Batch processing → Automatic checks → Comparison

Let AI organize pattern processing, peak search, refinement, checks, parameter extraction, comparisons and summaries so you can focus on scientific interpretation.

07 / Autonomous laboratories

From AI analysis
to the next experiment.

Feed analysis results into instruments and workflows to build a measurement–analysis–decision loop.

XRD instrument Automatic sample changer Robot Experiment scheduler Data management system AI Agent
  1. MeasureMeasure
  2. AnalyzeAnalyze
  3. EvaluateEvaluate
  4. DecideDecide
  5. Measure AgainMeasure again

Develop iPowder from AI for XRD into an intelligent XRD analysis and decision hub within autonomous laboratories.

Explore autonomous laboratories →
iPowder AI

Language models interpret; scientific algorithms calculate.
AI Agents connect the complete XRD workflow.

Professional XRD analysis,
simpler than ever.

Download iPowder Explore analysis features →