Bearing Curve Evaluation Supplies A Clever Way to Classify and Describe the Surface Texture of an Object
The Bearing Curve evaluation integrated in Acquire Automation XT can be utilized for profile and area measurement data. It works by establishing the percentage of data points at each ‘z height’, i.e. the bearing curve is drawn as % (z), where ‘z’ stands for the measured value, e.g. film thickness, topography, or intensity.
The percentage value can be quantified by drawing a line/plane in x(y)-direction at each z value and establishing the number of data points (length or area) above the line/plane. Then the ratio, or percentage, is determined from the amount of valid values above and the total amount of valid data points.
As a result of this evaluation, the graph of the bearing curve is shown, in addition to the outcome values for user-defined Material Ratios z (%) and Material Heights % (z).
The Bearing Curve evaluation can be added to a profile or area measurement by utilizing the drag and drop function in the Compose Task section. The evaluation parameters can be edited after clicking on the evaluation. The results and reports can be observed in the data log section.
Acquire Automation XT. Image Credt: FRT
Sample Classification Function Provides A Sophisticated Method for Classifying Measured Samples
Based on the generated measurement and analysis results, Acquire Automation XT’s Sample Classification function is a sophisticated way of classifying measured samples.
Result values can be sorted to user-defined classes, the sample can be assigned to an overall classification based on the distribution of the individual results to the various classes.
The classic ‘Good/Bad’ classification would be the simplest example. The overall sample result can be assigned with the ‘Good’ or ‘Bad’ label depending on the number of measurements which are counted as ‘Good’ or ‘Bad’.
A classification is defined by a ‘class group’ which is made up of at least one class. The respective class criteria is the information that is employed to assign an evaluation result to a specific class.
For each recipe, the evaluation task, the measuring task, and the output value to classify must be defined individually. The classes may be specified by assigning an expected value or value range in the parameter table.
A recipe may have multiple class groups at the same and also at different levels. One of the used class groups must be chosen as master to produce a final classification result which can be shown in the message box at the end of the process.
Results and reports can be seen in the datalog section, including the information on the sample classification. The appearance of the results/reports varies depending on whether the classification has been applied on element, sample, or task level.
Dual Scan Mode for High-speed Measurement of Features With A Large Discontinuous Step
In order to save time for the measurement of features with a big z-height difference or in instances where two parts of the sample must be measured with different sensor settings the Dual Scan mode is employed. It is used in combination with the FRT field of view sensors CFM, WLI, and CFM DT.
In the Dual Scan mode two single scans are carried out, one of the scans uses the parameters of Single Scan 1, the other uses the scan parameters of Single Scan 2. The raw data of the two scans is subsequently merged to one topography image.
The raw data from the Single Scan which is defined as ‘Master’ is preferred for this. At positions where the topography of the ‘Master’ Single Scan is invalid, values from the other Single Scan only remain in the resulting image.
In the Sensor Settings dialog, the measurement is defined by setting the lamp intensity, scan limits, step size and objective for the two single scans separately. The chosen objective has to be the same for both single measurements for a Dual Scan.
The composite topography measurement is shown when the two scans are finished. The user can use the Post-Processing Dialog to adjust the composite result. For each single scan the topography and intensity limits can be set separately.
Fit Based Defect Inspection to Find Local Deviations of A Measurement in Comparison to A Reference
The fit based defect inspection is an evaluation which should find local deviations of a measurement in comparison to a reference. It can be employed to identify defects on a sample such as digs, cracks, scratches, edged chips, etc.
So, a reference form (the ideal sample form without any defects) must be known and supplied as a mathematical function. Parameters of the mathematical function can be left open to be adapted by a fit algorithm.
Therefore, the reference form can be adapted to the measured sample keeping the basic shape. Deviations between the measurement and adapted reference form are then established, clustered and classified.
The basic steps of the entire procedure are:
- Establish the area of the sample within the measurement and cut out measurement data apart from the sample
- So that the reference form fits best the shape of the measured sample, adapt the parameters of the function that represents the reference form
- Establish deviations between measurement and fit
- Cluster deviations, assess and classify them
The user is able to set the threshold to where deviating measurement point is thought to be a defect point. Deviations which are lower than this threshold are not seen as defects. One might be only interested in defects whose lateral expansion is higher than a set threshold.
This property may be employed to consider only deviations (areas of deviation) whose maximum lateral expansion is higher than the set value. Found defects are classified in terms of their shape, particularly in terms of the aspect ratio between their expansion that is orthogonal to the longest expansion and their longest lateral expansion.
There are two classes available, digs and scratches. Defects with a higher aspect ratio than individually defined by the user are classified as scratches, otherwise as digs.
Individual Execution of Tasks and Evaluations for Recipe Adjustment
Single tasks can be executed during recipe editing in order to ease the creation of recipes in Acquire Automation XT. The chosen task can be measured and the results can be examined right where the scans of a measurement are defined.
Even presupposed tasks can be included in the process, such as alignments or reference measurements. The evaluations can also be adjusted individually and intermediate results can be viewed. The effects of the adjustments to the evaluations can be directly observed.
The supported measurement tasks include:
- Fine Alignment (if available)
- Area Alignment (if available)
- Circular Point Scan
- Point Scan
- Full Sample Inspection Map
- Line Scan
- Area Scan
Using Single Task Execution in Acquire Automation XT is simple. A recipe must be loaded in the Recipes view. The scans of each task can be defined in the Edit Task(s) view after composing the tasks of a recipe.
A command button with the label ‘Execute Task’ exists for tasks supporting Single Task Execution. The Single Task Execution is launched by pressing the command button.
Additional Evaluation Adjustment can be performed on an executed task, if the task has at least one Evaluation. The purpose of the dialog is to edit evaluations and filters of a task easily, and instantly see step by step what happens to the data. An input data set and the result or the output data are shown for each scan and evaluation in order to ensure this.
Additionally generated output data is also displayed if possible. At any time, the user can see all of the data sets that are available for further processing. If the evaluation is a filter which may alter the data, you can see the data in the input before filtering and in the output after filtering.
Area Alignment Task - Automatic Determination and Compensation of the Displacement of A Measurement Area in Relation to the Die Layout
For many use cases, the alignment of a sample is vital. Mostly Fine Alignment is utilized to align the sample. For samples with a die/element layout it might also be helpful to utilize Site Alignment.
Prior to processing a die/element, Site Alignment will be carried out for this die/element only. The die/element alignment is a superposition of the Fine Alignment and Site Alignment.
Fine Alignment and Site Alignment are based on pattern recognition. The pattern recognition must be taught during recipe creation by grabbing a system camera image and defining a part of the image as pattern.
An image processing interface will then produce a pattern using the image data. Another system camera image is taken at the same position as the teaching position during Fine or Site Alignment.
The image recognition algorithm will identify the pattern within the image data and detect the offset to the teaching position in x/y-direction and the rotation of the pattern. This information will be passed on to every following process step.
The Area Alignment task is an independent task which has been introduced as an addition to the less flexible Fine and Site Alignment. The biggest differences are outlined below:
- Pattern recognition can be based on sensor camera images, system camera images, or measurement data of any sensor.
- As an independent task it can be placed anywhere in the task composition.
- To increase detection quality and improve offset detection, it can be performed multiple times.
- Several result values can be activated that will be added to the task results. It can also be utilized to only detect the position of a feature on a sample without using the alignment information in the following process steps.
- Artificial patterns can be imported as image files.
- The optional result image that will be saved alongside the other result values clearly marks the recognized pattern area in the grabbed image data or measurement data for further investigation.
Setting up the Area Alignment task is a two step process. First the general task settings – scan settings, camera/sensor selection, behavior on error – must be set. Next, the pattern itself must be taught.
One Area Scan must be produced after setting up the fundamental task settings. Once the Area Scan has been established, the pattern teaching dialog can be employed to teach the pattern needed for pattern recognition.
Area Alignment tasks are tasks comparable to Area Scan tasks, so they will appear as every other task in the Process view. They will be performed for every die/element (if a layout is utilized) and the alignment will only be kept for the current die/element.
Area Alignment tasks provide preview images during the process, in the same style as that of the Fine Alignment and Site Alignment. It will show the header label Area Alignment found if a pattern has been detected, in addition to the detected quality/ score and the detected pattern rotation.
The expected absolute position is shown if the detection failed. Result values can also be included in the wafer result list if a remote interface such as SECS/GEM is utilized.
Access the Full Capability of Frt Multi-sensor Tools by Using Hybrid Task and Hybrid Evaluation
Acquire Automation XT’s hybrid functions provide a sophisticated method of combining the measurement or evaluation results from different tasks and evaluations to produce new data buffers or results. Hybrid functions in Acquire Automation XT are a vital key component to access the full capability of their multi-sensor tools.
The program features a hybrid measuring task and two hybrid analysis functions. The two available hybrid analysis functions are Hybrid Evaluation (Value Pairs) and Hybrid Evaluation (Arithmetic).
The Hybrid Evaluation can be added to a recipe by using the ‘drag & drop’ function in the recipe section, like any other evaluation technique. Typically, a hybrid analysis requires at least two individual results as an input to produce a new (combined) result.
The Hybrid (Arithmetic) evaluation enables simple arithmetic operations which are applied to multiple results in order to create a combined result. The general settings can be accessed by clicking on the evaluation function.
The variables employed for the evaluation can be defined in the Edit Parameter dialog. The variables available are all individual results from the preceding evaluations. The Hybrid (Value Pairs) evaluation enables simple arithmetic operations for pairs of values.
The list of value pairs to be evaluated are shown in the Edit Parameter dialog. The user defines the name of the result value and the expression, i.e. the arithmetic operation which is applied to the variables to produce the output value.
The Hybrid Task is a function which can be employed to create a new data set from two previously measured data sets. Different to the Hybrid evaluations, the Hybrid Task only writes new raw data. An evaluation must be added to the task in order to get a result.
The general settings can be accessed by clicking on the task function. The two original measuring tasks (Task A and Task B) can be selected in the top section of the Edit Parameter dialog. The operation which is employed to create the new data set can be defined under “Method”.
The process view will show the raw data and the filtered data of the previous measuring tasks when performing the measurement. The result report can be seen after completing the measurement successfully. The result report contains the results of the individual analysis tasks and also the ones of the hybrid analysis.
This information has been sourced, reviewed and adapted from materials provided by FRT.
For more information on this source, please visit FRT.