Skip to main content
A run can look acceptable in aggregate but contain internal variation that affects device performance: early-stage drift, mid-run excursions, or layer-to-layer inconsistency. Atomscale catches these by applying the same similarity tools used for run-to-run diagnosis to segments within a single run.

Two Approaches to Uniformity

The same Growth Monitoring and Global Similarity tools used to diagnose why runs differ also work with segments of a single run. Instead of comparing entire runs, you compare phases, layers, or time windows within one run.
Compare specific segments against reference segments with full granularity.Use Growth Monitoring when you want to track how a particular segment (a layer, recipe phase, or deposition cycle) compares to a reference segment from the same or a previous run. This gives you time-resolved detail: the similarity trajectory within the segment, derived metrics over time, and correlated tool state.What you see:
  • Similarity trajectory showing how each segment’s fingerprint evolves relative to the reference segment over time
  • Growth metrics compared between segments as time series, revealing exactly which properties vary
  • Tool state providing context for why a particular segment diverged

Setting Up

1

Create a Growth Monitoring project

On the Project page, create a new Growth Monitoring project.
2

Add reference segments

Select segments that represent your target behavior, such as a known-good layer from mid-run or the first cycle after stabilization. Assign labels or values to each reference.
3

Add segments to compare

Add the segments you want to evaluate as tracked samples. For periodic structures, this could be every cycle in the sequence. For phased recipes, this could be the same phase across different regions of the run.
4

Review the monitoring dashboard

The dashboard shows the similarity trajectory, growth metrics, and tool state for each tracked segment relative to the references. Look for segments that diverge; these are your uniformity problems.

What to Look For

Drift Over Time

Segments that gradually shift away from early-run behavior indicate drift. In Growth Monitoring, this appears as a similarity trajectory that steadily diverges from the reference. In Global Similarity, drifting segments form a gradient on the map rather than a tight cluster.

Outlier Segments

A single segment that diverges sharply from the rest points to a transient disturbance, easiest to spot in Global Similarity where the outlier sits visibly apart.

Phase-Specific Variation

Some recipe phases may be consistently less uniform than others. If segments from one phase always spread more widely, that phase likely needs tighter process control. Growth Monitoring makes this visible through the metrics comparison, showing which derived properties vary most during the problematic phase.

Periodic Structure Consistency

For multilayers and superlattices, compare each cycle against a reference cycle using Growth Monitoring. Consistent cycles produce nearly identical similarity trajectories. If early cycles differ from late cycles, or if specific cycles stand out, the time-resolved view shows exactly where within the cycle the variation occurs.

Connecting Uniformity to Outcomes

Within-run uniformity metrics become most valuable when connected to final device or material properties. Runs with tight internal clustering tend to produce more consistent outcomes. If loosely clustered runs correlate with degraded performance, that gives you a quantitative uniformity threshold to monitor against.

Choosing the Right Approach

ScenarioApproachWhy
Comparing each cycle in a periodic structure to a referenceGrowth MonitoringTime-resolved detail shows where within each cycle variation occurs
Checking whether all segments in a run are consistentGlobal SimilarityImmediately reveals outliers and clustering without pre-selecting references
Identifying which recipe phase has the most variationGlobal SimilaritySegments from different phases naturally separate on the map if they behave differently
Understanding why a specific layer is differentGrowth MonitoringDetailed metrics and tool state pinpoint what changed during that layer
Tracking uniformity improvement across iterationsGlobal SimilarityCompare segment clustering tightness across runs to see if uniformity is improving

Next Steps