Finding solid contaminants in food products can feel like a game of hide-and-seek. And some contaminants are very good at hiding.
Many applications are especially challenging for metal detectors, X-ray systems, vision systems and other inline inspection equipment. They can involve contaminants that are oriented in a way that’s difficult to detect or that are surrounded by product or packaging that interferes with the detection method.
The stakes are high, with recalls more traceable than ever and many trade customers emphasizing inline inspection as one of the product safety procedures they expect their suppliers to follow. Some customers require specific detection methods to be used on a product before they’ll consider buying it.
“Whatever you do, somebody wants to do better,” says Steve Gidman, president of Fortress Technology.
Metal is the most common type of contaminant. It’s usually relatively easy to find – unless it’s in a tricky orientation. If a metal contaminant is a flake or any other thin shape, it’s liable to get presented edge-first to a metal detector coil, which means there won’t be enough surface area for it to be detected.
Shavings from rollers, wires, brushes and other metal surfaces are a common source of contamination. Metal detectors are typically calibrated and tested with tiny metal spheres, whose diameters – as little as 1mm – generally form the limits of how small a surface area the metal detector can read. If it can only read the edge of a contaminant, and that edge is narrower than the limit, problems arise.
Using metal detectors in different orientations is one way to prevent metal contaminants from slipping through. Photo: Eriez Manufacturing
“Metal shape and size has a profound impact on its detectability. This is known as the orientation effect,” says Ray Spurgeon, product manager for metal detection at Eriez Manufacturing Co. “This phenomenon is most common with wire or long thin metal foreign objects like a needle.
"Using the needle to illustrate, a needle has two main constituents; its diameter and length, and how it is orientated over the coils, [which] has an impact on its detectability," he continues. "Simply stated, when metal is in a favorable orientation it is easily detected and when it is not, it is less detectable.”
Double feedback
Inspection system suppliers offer several ways to deal with this challenge. One is to use two signals with different electromagnetic frequencies, with an algorithm in the system’s analysis software that can interpret the twin feedback to detect even very thin surfaces. Mettler Toledo offers this with the Profile Advantage metal detector.
“The Profile Advantage metal detectors use various combinations of high and low frequencies simultaneously, along with built-in product signal suppression technology with two stages of discrimination, frequency and phase,” says Camilo Sanchez, Mettler-Toledo’s product line manager for metal detection. “This cancels the information from these combinations of high and low frequencies to effectively remove the product signal or ‘product effect,’ allowing for much smaller metal contaminants to be detected.”
Another method is to use separate signals in horizontal and vertical configurations – in effect, approaching the contaminant from two different angles. This gives much better odds against a contaminant slipping past because it sliced edgeways through a signal.
Fortress has the Interceptor DF, for “divergent fields,” for such applications. The horizontal field, which is the norm for metal detecting, does 70% of the work, Gidman says; the vertical one comes into play for contaminants with hard-to-detect orientations. By using both, he says, “you cover the weakness of each other.”
Other contaminants are more challenging than metal, especially when they’re low in density. These include fragments of glass or bone, or harvest detritus like rocks, says Christopher Young, business development manager at Anritsu Infivis. Detecting low-density contaminants is different enough from detecting metal that customers often have to prioritize.
If a meat producer needed to detect bone fragments, for example, that would be problematic for a thick package, Young says. “The product is just too thick, the density difference of the bone is too small.” Detection would probably be required upstream of packaging, for instance in a pipeline that carries ground meat.
Problematic products
Another difficulty with inspection comes when the product itself is problematic. Especially with metal detection, products that are highly conductive due to moisture, salt content or other factors can be difficult to distinguish from contaminants.
“One challenge in metal detection can be the product itself, which can exhibit 'product effect' if it has a lot of moisture and saline, or if hot, chilled or changing temperature, such as hot baked bread or frozen chicken,” Mettler-Toledo’s Sanchez says. “In these instances, these products are relatively good conductors and thus more likely to produce a signal in the metal detector in the same way as small metal contaminants would. This product effect makes it more difficult for the detector to distinguish between the product and the metal contaminant.”
One way to deal with such applications is to use X-ray systems, which don’t depend on electromagnetic detection. But when metal detection is the only option, they can overcome the challenge by combining low- and high-frequency signals.
Mettler-Toledo’s Profile Advantage uses these along with “Product Signal Suppression” software that interprets the signals’ frequency and phase. “This cancels the information from these combinations of high and low frequencies to effectively remove the product signal or ‘product effect,’ allowing for much smaller metal contaminants to be detected,” Sanchez says.
Product also can be problematic just because there happens to be a lot of it surrounding the contaminant. Large packages, especially bulk quantities, are always a challenge simply because they have to go through a bigger aperture; signals decrease in readability as the aperture increases.
Large packages are especially challenging for contaminant detection. Photo: Bunting Magnetics
Suppliers like Bunting Magnetics make equipment designed and calibrated to work with large packages. “Our new Big Bag meTRON 07 CI coil with Bulk Sense Kit metal detector was designed specifically to detect metals in large bags of dry bulk products,” says Barry Voorhees, Bunting metal detection product manager. “The meTRON 07 can detect 316 stainless steel as small as 2.0mm and ferrous/non-ferrous metal particles as small as 1.5mm when inspecting bags of dry product between 50 to 100 lbs.”
Incompatible packaging
Packaging can be another difficulty. If it’s metal or even has metal components, using a metal detector can be difficult if not impossible.
“Regarding packaging, if metalized film is used, the film will trip the detector and needs to be ignored," says Eriez’ Spurgeon. “Therefore the detector needs to be de-rated, resulting in a less sensitive detector.”
Non-metal packaging can present its own difficulties, especially if it’s made of similar material as the contaminants that are being searched for (or is in fact the same material – packaging fragments are a major source of contamination).
X-ray inspection is the preferred method for such applications. Similar to dual-wave metal detectors, X-ray systems are available with dual beams for challenging situations.
“Depending whether the goal is removing glass in glass or metal in cans, a single or dual beam X-ray system can provide exceptional levels of detection with minimal false rejections,” says Geri Foley, Mettler-Toledo’s product line manager for X-ray. “Mettler-Toledo's advanced imaging software with adaptive filtering enables the blind spots to be masked, so the contaminant is easily found and automatically rejected.”
Similarly, Anritsu’s DualX system uses software that interprets the X-ray signal at differing energy levels, creating two images that can be compared and contrasted to lessen false rejects and other problems. “It’s a higher level of X-ray detection to specifically look for low-density [contaminants] like bone, stone or glass in images that are more difficult, like a bag of french fries,” Young says.
Contaminants being hard to find because they resemble the surrounding good product is a problem that can occur before packaging. Bulk product flows, especially of raw produce, often include contaminants that must be separated before further processing. When there’s little difference in shape or color between the contaminant and the good product, it can present a challenge for vision-based separation systems.
“For example, pieces of nut shell often look like a good piece of nut meat,” says Marco Azzaretti, director of marketing at Key Technology. “Similarly, a piece of wood might look virtually identical to a mushroom slice. When processing french fries, a clump of batter can resemble a battered potato strip.”
Vision systems use multiple sensors and other sophisticated approaches to find contaminants that “hide” in a product stream. Photo: Key Technology
Key attacks this problem in two ways. Similar to dual-source metal detection and X-rays, Key’s Pixel Fusion system combines information from multiple sensor types at the image pixel level to maximize the contrast between good product and contaminants.
Key’s BioPrint is an even more sophisticated approach. BioPrint identifies and distinguishes objects based on their chemical composition and material structure. This lets it differentiate a piece of good product from a contaminant even when the objects appear visually identical.
“Knowing the material compositions of different objects, the sorter doesn’t need to rely on external appearance to tell them apart,” Azzaretti says.
Certain physical contaminants will always be hard to find, due to their nature or the nature of what surrounds them. Advanced inspection technology, matched to the application, can help reduce the odds of an embarrassing (or worse) contamination incident.
