Alix Cornish of Campden BRI explains the benefits of X-ray computed tomography in measuring food properties and highlights applications for this technology.
Food products often have delicate structures with fine walls and high internal porosity. Measurement of these structures is important for new product development (NPD) trials, product benchmarking and troubleshooting. X-ray computed tomography (CT) is a technique that allows internal food structure to be visualised and measured without any destructive sample preparation.
Conventional imaging techniques generally produce 2D images of the surface or cross-section of a sample. Labour intensive methods, e.g. sectioning to produce thin slices, or chemical fixation to produce contrast, are frequently used to prepare samples for these 2D imaging procedures. However, these processes are usually destructive and the resulting 2D information is often insufficient to draw conclusions regarding the 3D structure. Moreover, these destructive techniques can introduce artefacts which confuse the interpretation of these measurements. X-ray CT can overcome these problems by providing 3D images of foods.
Technical aspects of X-ray CT imaging
Objects are placed within an X-ray beam and rotated; the X-rays are absorbed by the object. Dense regions cause more absorption than low-density regions, and thicker regions of the object cause more attenuation than thinner regions. A series of shadow projections are recorded. The shadow image produced at each angle provides information about the size and density of the object in that orientation. Computer software is used to reconstruct the 3D shape from a series of shadow projection images, recorded over 180° (or more).
The maximum resolution depends on the specific instrument, but standard benchtop systems are capable of imaging down to 1µm (pixel size). There is a trade-off between sample size and resolution, due to the limited number of pixels across the width of the detector, so the highest resolution is only achievable for small samples.
Figure 1 shows X-ray projection images recorded from a twist snack product. Positions on the sample where the X-ray beam has penetrated just one layer show up bright and regions where the X-ray beam has penetrated a thicker region of the product, show up darker. Tiny dark spots are also visible on these images. These are caused by dense salt crystals and seasonings on the surface of the product, which attenuate the X-rays to a greater extent than the snack product itself.
![]()
These, and many more projection images are processed using computer software to generate a 3D model. The 3D model can be viewed and manipulated on the computer to study the internal structure or to make measurements.
Distribution of composition
Contrast in CT imaging is due to atomic composition. Heavier elements are much stronger attenuators of X-rays than lighter elements. At an atomic level, the majority of food is composed of carbon, oxygen and nitrogen. These three elements have very similar masses (they are adjacent to each other in the periodic table) and attenuate X-rays to a similar extent. Therefore, X-ray CT is generally not capable of distinguishing different components, e.g. protein or sugar, to allow the distribution of such components to be mapped within a product. However, there is sufficient contrast to image food ingredients that contain heavier elements, e.g. salt. This technology also works well for aerated products, where good contrast is observed between the solid food component and air within the structure.
Measurement of food properties
Pore size distribution – discrete bubbles
Porous structures exist within many categories of food products e.g. bakery produce, extruded cereals and aerated confectionery and dairy products. The size and distribution of bubbles is important for the size, shape, structure and texture of the overall product. Bubble size distribution heavily influences the texture of products and is therefore an important sensory attribute. Consequently, it is vital to understand the size distribution of bubbles in porous foods. The type of bubble structure is also important. Some products, such as aerated chocolate, contain discrete bubbles entirely surrounded by the product material, in contrast with products, such as bread, that contain an interconnected network of bubbles. Others, such as some types of extruded product, contain a combination of these structures. X-ray CT offers a route to characterise the size and type of bubble distribution without laborious preparation methods, which could damage the structure.
Figure 2 shows a 3D rendered image of a piece of aerated chocolate. The cutaway reveals the internal porous structure. This 3D model can be digitally cut in any orientation to visualise the internal structure.
![]() |
| Figure 2 3D rendered image showing the structure of a piece of aerated chocolate |
In addition to non-destructive imaging, X-ray CT enables quantitative analysis of porous structures. The volume occupied by the whole structure and the volume occupied by each bubble can be measured. Several structural parameters can then be calculated based on this information.
For the piece of aerated chocolate shown in Figure 2, the total volume of space occupied is 5,696mm3 and the total volume of bubbles is 1,348mm3, representing a porosity of 23.7%. A total of 6,848 bubbles were measured, which corresponds to 1.2 bubbles per mm3.
![]() |
| Figure 3 Pore size distribution plot and corresponding 3D models showing a piece of aerated chocolate. The 3D models show the bubbles in 7 size classifications |
Figure 3 shows a bubble size distribution plot for the aerated chocolate piece. The distribution has been plotted using 46 size classification bins, approximately 0.05mm apart. 3D models showing bubbles in seven broader size classifications are displayed around the distribution plot. A 3D model showing all the bubbles (colour coded by size) is shown in Figure 4.
![]() |
| Figure 4 3D model showing the bubbles inside a piece of aerated chocolate colour coded to correspond with size – where purple represents the smallest and red represents the largest bubbles |
These images reveal that the larger bubbles are only located in the centre of the chocolate and that the smaller bubbles are evenly distributed throughout the aerated interior. Some small bubbles are also present in the surrounding chocolate, towards the base of the product.
Bubble size distribution heavily influences the texture of products and is therefore an important sensory attribute.
Distribution of open networks
It is also possible to measure the size distribution for open porous structures, such as bread and cake. These porous networks typically only contain one air void, with a complex interconnected shape. In order to obtain meaningful information about the size of the internal network, a parameter called ‘structure separation’ is often calculated. This mathematical procedure involves measuring the maximum size of sphere that can fit at every location within the interconnected porous network. The distribution of these maximum sphere sizes is then measured. An analogous procedure can be performed on the solid phase of the product in order to calculate the distribution of wall thicknesses. It is also possible to measure the change in structure as a function of height through a product and the anisotropy (directionality) of the pores.
Frozen products
It is important that samples remain static during conventional CT scans to avoid artefacts related to sample movement. For soft samples that flow over time, such as dough, this can cause problems. A route to overcome this is to freeze samples prior to scanning and to keep the samples frozen, and hence static, during the scan.
Food industry applications
Measurement of food structure is important during NPD trials, especially where a complex or novel structure is being developed. Another application for CT imaging is during product benchmarking, where it is important to characterise the products as completely as possible, in order to identify specific attributes that need to be changed to better match the target product. Differences in structure can result in very different mouthfeel and even taste. CT imaging can also be used as a troubleshooting tool if problems arise.
The detailed structural measurements provided by X-ray CT are complementary to other instrumental measurements commonly used in the food industry, such as texture analysis, and can be used to help interpret data from these tests. For example, texture analysis on a snack product may reveal that it is firmer than a competitor product. CT imaging can be used to understand which aspect of the structure is different between the two samples, for example, it could be because the firmer product has a thicker wall structure.
Imaging dynamic processes
It is possible to image products during dynamic processes by taking a series of images for processes with medium time scales. For example, bread can be imaged during proving and baking to generate movies that show how features, such as crusts, form. Imaging cakes during the baking process can help identify how defects, such as dense streaks, are formed. Over longer time scales, non-destructive imaging of products, such as mousses, can be made on multiple occasions over shelf-life to assess stability of structures. Faster processes, such as the evolution of bubbles in a sparkling drink, cannot be imaged using this technology.
Imaging of dynamic processes with medium time scales, such as baking, is not possible using standard benchtop laboratory instruments, but can be performed on larger instruments, such as clinical CT scanners, that are more commonly used by hospitals for examining patients!
Measurement of food structure is important during NPD trials, especially where a complex or novel structure is being developed.
Modelling of food structure
X-ray CT scanning can be used to provide input for finite element modelling (FEM). The measurements can be binarised to give a 3D representation of the boundaries between the solid and gas phases, which can be exported to FEM software to define the geometry for modelling. It is also possible to export 3D images from CT instruments for use in 3D printing.
Structural mechanics models are used in many mechanical engineering applications to assess the loads on components and to optimise their design. For processing equipment, this can include assessment of the torque required from a motor, or the thickness of a process vessel to withstand the expected load. Finite element modelling (FEM) is also useful for design of food packaging to ensure that it provides sufficient mechanical protection for the product while remaining practical to open and while using the minimum amount of material to save cost and to minimise environmental impact. Modelling of the mechanical properties of food materials themselves is also useful to understand effects on food texture and to model interaction with processing equipment and packaging, for example to ensure that loads during conveying and transport will be sufficiently low to avoid damage to the product.
Food products and processes are becoming increasingly complex. However, computational power is also increasing. X-ray CT data combined with FEM provides a route to model complex food products, reducing the need for lengthy pilot trials.
Other applications
X-ray CT is not limited to food structure research. There are several other applications relevant to food, for example foreign body investigations. The traditional approach for these involves removing the object from the food product for destructive analysis. CT techniques can be used to identify how the object is situated inside the food product without destroying it or the surrounding food product. It is therefore possible to gain insights into how the object got there, and in particular, at what stage the foreign body entered the product. For example, if a foreign body was introduced into a bakery product before baking, the cell structure surrounding the foreign body would be intact. If the foreign body was forced into the product at a later stage, tears and holes would be visible, marking the path of entry.
This technique can also be used to investigate packaging defects. For example, leaking bottle caps can be investigated for faults, without the need to open the bottles.
Conclusions
X-ray CT imaging offers solutions for a diverse range of problems where there is a requirement to see inside products in 3D, without destroying them. It can be used to measure the pore size and wall thickness distribution in a wide range of products, such as bread, cakes and chocolate. The bubble size distribution in bread dough can also be measured. It is even possible to image products dynamically during baking to generate movies that show how features, such as crusts, and defects, such as dense streaks in cakes, are formed.
Alix Cornish, Senior Scientist
Campden BRI, Station Road, Chipping Campden, Gloucestershire, GL55 6LD
Email alix.cornish@ campdenbri.co.uk
Telephone +44(0)1386 842054
