When the conversation turns to air bubbles, Werner Halbinger gets very serious. “I’ve declared war on them,” he says. That’s because “they cause all kinds of problems” for him as the image-processing and laboratory engineer at Seidenader Maschinenbau GmbH in Markt Schwaben (near Munich), not to mention the company’s customers. The company, which is part of the Business Area Pharma Systems, manufactures high-tech machines that check syringes, ampoules, and vials of all sizes for the presence of foreign particles and production errors.
Different inspection techniques are used depending on the product and receptacle in question. Most of the products tested are clear liquids in transparent and colorless glass receptacles, which undergo stringent quality assurance screenings. Anything that creates a shadow could or could not be a foreign body. “Unfortunately, air bubbles are also registered as shadows by the detectors,” Halbinger explains. Such bubbles used to present a major challenge to Halbinger, his colleagues, and customers throughout the entire pharmaceutical industry.
It’s often difficult to clearly distinguish between harmless bubbles and problematic particles. Unfortunately, it’s not possible during high-speed inspections to simply wait until the unidentified objects sink to the bottom of a receptacle and thus become clearly identifiable as foreign bodies. For this reason, all products in which image-processing software detected a shadow during screening were rejected. After all, safety is the top priority and the fundamental promise to customers.
These rejects are a problem in the case of some medications and other medical products such as distilled water or vitamin solutions that are manufactured in large quantities. But they are a key cost driver for manufacturers of expensive active ingredients, especially biopharmaceuticals such as cancer medications, vaccines against infectious diseases, and drugs that slow the progress of multiple sclerosis. Substantial savings could be achieved by preventing the destruction of even a small portion of the syringes, ampoules, and vials containing solutions of synthesized molecules that are often difficult and costly to manufacture.
Long before Halbinger discovered a possible approach to solving the bubble or foreign-substance mimicry problem, he and many other engineers looked for ways to minimize the number of rejects. This approach focused on preventing the bubbles from being formed and ensuring that those that were formed were removed. Although the Seidenader engineers played an important role in the steady decline of the reject rates, the problem was still not resolved.
Halbinger faced major challenges when a customer approached Seidenader to request that a new feasibility study be performed. Halbinger took test receptacles into his laboratory and began experimenting with polarization filters that raised contrast levels. He also experimented with color filters and a color camera. He kept on building makeshift apertures out of black cardboard and worked with different distances, filters, and sensitivity settings for image recognition software. “I wanted to take advantage of the fact that the air bubbles behave like optical lenses under certain conditions,” Halbinger, an optical expert himself, explains.
lAir bubbles behave like optical lenses under certain conditions. r
Werner Halbinger, Image-processing and lab engineer at Seidenader
Halbinger understood that his “enemies” were transparent spherical objects, and at some point he noticed that the red-green-blue (RGB) pattern of the background light used for illumination was being reflected in a very big air bubble. “Interesting,” he thought to himself. He then continued his experiments, changed settings, and kept observing how the light pattern in the bubble behaved. And even though he still couldn’t clearly detect air bubbles, especially very small ones, he nevertheless couldn’t get what he had observed out of his mind.
Ultimately, it was a Fresnel lens that enabled him to achieve the breakthrough he had long been waiting for. Fresnel lenses are used to raise the contrast in error detection processes. In the beam path that is focused by the lens, even relatively small particles stand out significantly from the background, which makes them easier to detect and identify. In order to examine the properties of the Fresnel lens, Halbinger placed it in front of his computer screen. He then used a graphics program to draw in the colors red, green, and blue. There they were — regular patterns. “This looks promising; the same thing has got to work with air bubbles too!” he thought — and this turned out to be exactly the case. When illuminated with RGB light through a Fresnel lens, air bubbles always betray their presence with a characteristic pattern — red at the top, green in the middle, and blue at the bottom. Just as importantly, Halbinger also discovered that optically transparent particles — the classic example is glass — displayed either no color pattern, random color patterns, or a pattern opposite that produced by the air bubbles. These effects are caused by differences between the refractive properties of air bubbles and those of particles made of solid transparent materials.
“It took me a second to understand what Werner had done after he showed me — but then I immediately realized how much potential his discovery offered,” says Halbinger’s boss, Andreas Böhme, Manager Vision Engineering at Seidenader. Böhme then began to consider just what could be achieved with such a technology. However, he also understood that they needed to figure out how to install a functioning system in a machine. As Böhme well knew, even the best idea only accounts for 10 to 20 percent of the total development process: “The rest is implementation.”
Böhme, Halbinger, and many other colleagues from Seidenader got down to work. Meanwhile the patent registration was handed in — and Bubble-X is doing well on the market. Indeed, several pharmaceutical companies have already ordered the first machines equipped with this innovative inspection technology. “There’s a huge amount of interest when I present it at conferences,” says Seidenader’s Sales Director, Christian Scherer. “We’ve acquired many new customers over the last few years especially, and one of the main reasons for this is definitely our expertise in technical innovation.”
To ensure that this remains the case, Halbinger has already identified a new “enemy,” which is basically another version of the old one. The issue here involves the fact that Bubble-X reliably detects air bubbles down to a size of 200 micrometers. This figure will gradually decrease to 100 micrometers over the next few years as camera resolutions and image-processing systems improve. After that, the goal will be to move into the realm of micro-bubbles. “That will be difficult, but those micro-bubbles are already on my list,” says Halbinger.
Three light sources (red, green, and blue – RGB illumination) are arranged in such a manner that the green beam in the middle can pass straight through the convex lens (1) and thus always falls on the detector camera (3). The lens refracts the red and blue beams in a way that ensures they pass by the camera very closely when there is no problem. The three most common cases: