Search
Quantify fiber orientation, phase distribution, porosity, and defects in composite and polymer materials from 2D and 3D imaging data
The performance of composite and polymer materials, such as GFRP, CFRP, and CMC, depends on their internal microstructure. Fiber orientation, phase distribution, porosity, filler dispersion, and manufacturing defects all influence mechanical, thermal, and functional properties. Characterizing these features accurately is essential for understanding material behavior, improving manufacturing processes, investigating failures, and developing higher-performing materials.
Thermo Scientific Avizo Software helps researchers transform complex 2D and 3D imaging data into quantitative microstructural insights. Working with data from micro-CT, FIB-SEM, optical microscopy, spectroscopic imaging, and other techniques, Avizo Software combines AI-assisted image enhancement and segmentation with advanced image processing and quantitative analysis, high-end visualization and automated workflows to measure and visualize critical structural features with speed, accuracy, and reproducibility.
By linking quantitative microstructural analysis with material performance, Avizo Software enables researchers and engineers to make more informed decisions throughout materials research, process development, quality control, and failure analysis.
Composite & polymer applications
Enhancing mechanical performance with precise fiber orientation
Compared to cast metals, reinforced plastics dramatically reduce weight while providing fine-tuned control over their material properties. Glass, wood, and carbon can all be embedded within a polymer to optimize the material for specific requirements like tensile strength, heat resistance, or toughness.
Fiber orientation is a crucial aspect of material design, as materials like glass fiber-reinforced composites (GFRP) are strongest when the fibers run parallel to the force placed upon the material. Using micro-CT imaging, GFRP mixtures from EMS Grivory were studied with the XFiber extension in Avizo Software.
This technique segments individual fibers to retrieve per-fiber measures like length, tortuosity, orientation, aspect ratio, etc. Layer-wise statistics are also aggregated for statistical analysis and can be displayed with several visualizations in Avizo Software. The investigation helped confirm fiber orientation parameters within this commercially available GFRP.
Orientation analysis of fibers in glass fiber-reinforced composites (GFRP). Avizo Software with the XFiber extension isolates individual fibers, which informs statistics on a per-region or layer-wise basis. Orientation tendency is represented with tensorial glyphs and a colorized volume. Sample courtesy of EMS Grivory.
Verifying weaving patterns for superior mechanical performance of glass fiber
3D composites show much promise as building materials because they allow tunable performance for specific engineering applications. Compared to bulk materials like steel, fiber-reinforced composites are lighter and can be designed to better suit specific load-bearing applications. However, the complexity of these materials means that they must be carefully studied before they can be safely employed.
Time-series imaging experiments allow researchers to understand failure mechanisms, such as in composites. In this example, fatigue damage was studied in a 3D woven glass fiber-reinforced composite (GFRC). Avizo Software can be used to distinguish the fiber tows by orientation (warp and weft) for visualization, but also to classify damage.
Debonding and cracks were observed as the fatigue study progressed. Since the materials were segmented, the damage could be categorized according to the type of interface, such as warp–resin or binder–binder. This information then feeds back into material performance databases so that the right material is selected for future applications.
Damage evaluation in 3D woven glass fiber-reinforced composites (GFRC). Samples were imaged under tension using time series computed tomography. The fibers were segmented in Avizo Software by orientation (warp and weft), and cracks were isolated by crack type (debonding and transverse cracks). Figure reproduced under CC BY 4.0 from Yu et al. 2015. Data courtesy of Henry Moseley X-ray Imaging Facility, School of Materials, University of Manchester.
Verifying concrete mechanical performance by measuring pores and cracks
Many modern building materials are fundamentally composites, and concrete is one of the most commonly employed. Even though concrete is widely used, the many details of how its components are sourced and prepared give it a range of material properties, which is highly important for ensuring building safety.
3D imaging such as micro-CT can be employed to study the empirical distribution of constituent components in composites such as concrete. In the example shown here, helical micro-CT imaging of an anisotropic sample shows the typical composition of coarse aggregate, fine aggregate, cement, and air (porosity).
The deep-learning segmentation within Avizo Software precisely segmented the aggregate and porosity from cement. Cracks were further distinguished from porosity, including difficult-to-resolve cracks at the edges of aggregates. Finally, pores were digitally filtered to retain only the largest 100 pores to understand their spatial correlation with cracks and the spacing of coarse aggregates.
Using helical micro-CT, a concrete sample was digitally segmented to allow bulk composition and spatial correlation to crack propagation. The 3D rendering in Avizo Software shows the aggregate and cement, pores and cracks, and the largest 100 pores in a sequence of 3D renders. Data courtesy of Richard Deschenes at Youngstown State University.
Matching expected specifications with advanced rubber segmentation
Polymer composites include added compounds that change their material properties. Compounds can be added to improve the range of operating temperatures, material stiffness, resistance to chemical degradation, and more. 3D electron microscopy such as FIB-SEM tomography is used to validate the distribution of these compounds.
Polymers can be difficult to investigate within an electron microscope because of their insulating properties, which lead to a charging artifact observed in electron micrographs. If charge mitigation is not available, it may still be possible to retrieve the information you need from less-than-perfect imaging data.
Avizo Software includes deep-learning capabilities that can address uneven imaging contrast such as charging artifacts in FIB-SEM. For example, in the result shown here, there are inclusions within an ABS polymer that are relatively dark in one part of the image and relatively bright in another. A deep learning model trained on a small region can predict inclusions throughout the sample. This allowed the investigation of particle packing density and sizes to continue without needing to send the sample out for an additional imaging session.
FIB-SEM tomography of a polymer shows uneven charge that is typical of this sample type. Using deep learning in Avizo Software, the inclusions within the sample are clearly segmented as shown with a sequence of colors.
Crafting a high-precision functional polymer with target porosity and tortuosity
The electrochemical performance of rechargeable batteries is strongly influenced by the spatial distribution of constituents. Effectively, this is a composite material. Layering the cathode and anode in a series often includes a polymeric separator material that has an engineered permeability to allow electrolyte flow from cathode to anode.
Due to the multi-modal and heterogeneous distribution of components, it is useful to study these materials in 3D with FIB-SEM tomography. By cooling the sample to cryogenic conditions, the true sample morphology can be preserved even when milling through beam-sensitive polymers.
In this example, a commercially available battery separator was imaged in a FIB-SEM at –80°C. In Avizo Software, the pore network was segmented within the polymer and the adjacent ceramic particle coating. The XLab Suite extension was used to simulate molecular diffusivity through the separator and calculate the pore network permeability and bulk tortuosity.
Cryogenic FIB-SEM tomography of a polymer separator used in the cathode of lithium-ion batteries. The top third of the image is ceramic particles with binder atop a large porous polymer separator. The pore network is shown in orange and was segmented with deep learning in Avizo Software. Pore network permeability and tortuosity were studied with Avizo Software and the XLab Suite extension.
Optimizing filter porosity for target applications
Porous filter membranes can be engineered to permit particles of specified sizes to pass while trapping larger particles. To prevent the filters from becoming overwhelmed by large particles, multi-scale filters can capture large particles first, followed by smaller particles later.
Volume imaging such as FIB-SEM tomography or serial block face imaging can help validate that the pore network matches the filtration application. Beam-sensitive polymers can be imaged under low-voltage conditions to preserve the actual structure of the filter without melting, charging artifacts, or deformation.
In this example, Avizo Software was used with a SEM to retrieve an overall porosity of 71.5% in a commercially available PES 0.45 filter membrane. The flow of particles through the membrane was simulated with the XLab and XPoreNetworkModelling extensions. This combination helps validate the performance of this multi-scale porous polymer filter membrane.
Large-volume analysis of a multiscale polymer membrane used in filtration. By collecting data on a Thermo Scientific SEM, a large region is imaged in 3D so that the pore network (shown here in purple) can be investigated in Avizo Software.
Predicting material mechanical deformation under stress
When compared to standard materials, composites offer flexibility in cost, performance, and tunability. However, flexibility adds complexity to understanding bulk material performance. Simulations based on assumptions may not match observations of material testing and thus benefit from a feedback loop with real experimental data.
To interrogate the strain distribution within a 3D part, digital volume correlation (DVC) can be applied to a series of 3D images. The method, implemented in Avizo Software, tracks object displacement to sub-pixel resolution. This information informs a globally integrated distribution of uniaxial and biaxial strains. It is possible to map the strain to a mesh that follows important interfaces or features within the part to capture what is otherwise challenging to model in a simulation.
In this study, lightweight carbon fiber sheets were molded into a bulk specimen and subjected to a tensile test. Micro-CT imaging paired with DVC revealed that microcracks formed in areas of overlapping strands and high pore density.
Series of imaging data results showing the complex distribution of displacement and strain within a carbon fiber sheet molding compound through the use of digital volume correlation in Avizo Software. Data courtesy of KU Leuven and the University of Tokyo and published in 10.1016/j.compositesb.2025.112220.
Predicting material properties with CDF
Strong fibers are used to reinforce polymer composites in a way that can tailor stress resistance according to the needs of the part. For example, in a bike frame, carbon fiber-reinforced polymer (CFRP) provides a lighter frame with a carefully tuned balance of frame rigidity and flexibility.
Damage within CFRP can be complex. Voids created during manufacturing can concentrate stress and lead to cracks, delamination, and ultimately part failure. The stress placed on individual fibers can also be heterogeneous and result in an additional mechanism of failure.
In this example, Avizo Software and the XFiber extension were used to segment each carbon fiber within a CFRP imaged with micro-CT. The analysis was extended for modeling by creating a tetrahedral mesh for each fiber using the XWind extension. This provided layer-wise orientation of carbon fibers, a correlation to voids in the polymer, and a mesh model for finite element analysis for further study.
3D rendering of carbon fibers in a carbon fiber-reinforced polymer (CFRP) used in a bike frame. The individual fibers are segmented with XFiber in Avizo Software and individually meshed using XWind. The fiber color corresponds with fiber tortuosity. Data courtesy of Rigaku Corporation.
Why choose Avizo Software for Composites and Polymers?
- 3D visualization of fiber architectures, polymer matrices, fillers and defects
- Quantitative measurement of fiber orientation, volume fraction, dispersion, and porosity
- Segmentation of phases (fibers, matrix, additives, voids) using image processing and machine learning tools
- Analysis of fiber-matrix interfaces and interphase regions
- Characterization of defects such as voids, cracks, and delamination
- Multiscale analysis from microstructure to mesoscale for structure–property relationships
- Evaluation of filler distribution and agglomeration in reinforced polymers
- Determination of the internal correlates of stress buildup within composites using digital volume correlation
- Simulation of permeability for only percolating porosity through porous composites and polymers
- Reproducible workflows for consistent and comparable materials characterization
Use cases: Analyzing composites and polymers using Avizo Software
Webinars
Resources
Customer testimonials
Looking to improve composite and polymer performance using imaging data?
Avizo Software extracts constituent materials from 2D and 3D image data of composite and polymer samples. With this segmentation, it calculates statistics and provides digital twins for characterizing and improving the material performance of composites.
FAQ
Yes. Avizo Software takes 2D or 3D image data, segments features, and provides measures and models based off these data. Our team has been building features for and expertise in studying composites for over two decades. We advise you to request a trial license which will also place you in touch with a team of experts who understand composites applications.
AI methods are relatively new and can work for noisy or non-ideal images that traditional image processing methods may not handle well. Avizo Software includes both AI methods and image processing methods. Sometimes, the combination of these techniques enables otherwise difficult work to be done more efficiently. More information on AI capabilities in Avizo Software.
Yes. Avizo 3D Pro Software includes the AI Assisted Selection tool which is a pre-trained AI model. This works well for composites applications. Other pre-trained models are available for download using a built-in Xtras Installer. These tools can be browsed in the Xtra Library.
Viewing pore network models in Avizo Software requires the XPNM extension. Creating a pore network model also requires Avizo 3D Pro Software. Learn more about Avizo Software packages.
To calculate fiber length, tortuosity, orientation, and similar statistics, you must digitally isolate individual fibers from 3D data. The XFiber extension for Avizo Software uses a unique method to achieve this that is not found in other software. The advantages of this approach include its accuracy and flexibility. As input, you can use any 3D data such as micro-CT or FIB-SEM tomography. Once fibers are traced, you can retrieve statistics on a per-fiber basis or aggregate the statistics for regionalized statistics. This can be useful for layered materials. More information on Avizo Software packages.
Composite materials are interesting because they have complex material properties. Unlike a bulk sample composed of one material, the response of a composite to a force like heating, tensile stress, or compression may be non-uniform within the specimen and marked by cracking, delamination, and collapse mechanics not found in mono-material samples.
The Digital Volume Correlation method extends the Digital Image Correlation premise into the sample in 3D. Using 3D image data, the texture within the sample (such as pores or fibers) is subtly tracked over a sequence of images collected while the sample is experiencing a force upon it. Avizo Software with the XDVC extension will digitally register the volumes, retrieve an overlapping region between time points, track the 3D distribution of sample movement (displacement) with subpixel precision, and return a conformal mesh that can be compared to finite element analysis of your idealized samples. By comparing empirical results to simulations, you can update your understanding of the way the composite part responds to the force that it was placed under.
Our team is uniquely specialized for the DVC technique, so we encourage you to contact us early in your investigation, as DVC experiments benefit from having the image analysis considerations inform the sample loading and imaging experiment.
Avizo Software primarily works with image data. Traditional composites imaging includes micro-CT imaging, synchrotron tomography, 2D SEM, and FIB-SEM tomography. Less common imaging methods can be opened, viewed, and analyzed in Avizo Software, such as 3D Mass Spectroscopy, TEM tomography, 3D EBSD, 3D Raman Spectroscopy, 3D XPS, etc. If you want to test Avizo Software, we advise that you request a trial license or contact us to validate whether your data are suitable for analysis in the software.
Yes. Avizo 3D Pro Software offers an extension called XLab Suite that runs Lattice Boltzmann Method fluid dynamics simulations of flow. This is advantageous because the simulation is matched to the observed pore phase in your sample. It is then helpful to also do preprocessing and segmentation, such as with the Avizo 3D Pro Software edition. Avizo Software offers powerful and dynamic 3D visualizations of the simulated fluid flow to aid in identifying how tortuous pathways are and what is the effective permeability of a sample. For more detail, you are encouraged to review Avizo Software packages.
Yes. Avizo Software includes a dedicated Segmentation+ Workroom. The workroom provides complementary tools for labeling, such as a lasso, brush tool, watershed, and magic wand. This workroom allows users to generate ground truth for customized Deep Learning models that can be reused and transferred between Avizo Software installations.
It is recommended to have an NVIDIA GPU with CUDA support. The XFiber method requires NVIDIA GPU with CUDA support. There are other considerations shared with the rest of Avizo Software. As these requirements may change, we list full hardware requirements in unified location here.
Thermo Fisher Scientific offers training sessions, video tutorials, webinars, and technical support to ensure you are using Avizo Software to its fullest capacity. It is recommended to begin with the Learning Center or the User’s Guide packaged with the software.
Related products
For Research Use Only. Not for use in diagnostic procedures.