Pfizer Learns by Doing with RFID

June 14, 2006
Pfizer Global Technology packaging expert Tim Marsh takes us through lessons learned on the long road to RFID success.

As it does on many fronts, Pfizer is taking the lead in implementing RFID (radio frequency identification) for product protection and supply chain oversight. It will be some time before the company has amassed the internal expertise on RFID that it desires, says Tim Marsh, technology manager for Pfizer Global Technology.

Even so, Marsh knows quite a bit more now than even six months ago, when the company’s Viagra pilot began. He spoke with Pharmaceutical Manufacturing about the challenges of RFID from a packaging perspective.

P.M.: The impact of packaging on RFID read rates and distances is just now becoming understood. What input/roles have packaging experts had on the Viagra pilot, or on any of your other RFID tests?T.M.: Various inputs from external sources, RF engineers, packaging people with RF experience and our own internal learning from dealing with RFID contributed to our approach for tagging Viagra. You can certainly read a book about RF, but doing it proves to be the more fruitful approach. Theory is fine for pristine applications, but it can fall short in providing knowledge to make appropriate business decisions. Our experiences in dealing with the Wal-Mart mandate gave us early insight into how readability and read rates are affected by packaging and product — and also how to compensate in some cases.P.M.: You’re using HF tags for the pilot. Was the fact that HF is thought to work better with a wider variety of packaging materials a major consideration in this choice? Did you test UHF extensively at the item level, or make a decision based upon talks with vendors and RFID experts?T.M.: We selected HF based on several factors. It’s an established and mature technology. It’s less susceptible to effects from liquids and metals. It is truly constrainable on our packaging equipment, meaning we don’t need to erect and install all sorts of shielding and require SOPs for line technicians on where product with tags can be or not be — it’s stupid-proof, to use an old phrase. The frequency is truly global, which is a huge benefit.

Unbiased third parties familiar with installations of LF, HF and UHF gave a resounding thumbs-up for HF at item level once they understood our requirements. Our own experiences with UHF, and its lack of performance and immaturity, meant we would have needed to pay a much higher price to achieve the same results. Combine that with our partners/integrators not being willing to guarantee the UHF solution. Pfizer still does not consider UHF extensible for pharmaceuticals at item level.

P.M.: Have you experienced difficulties in reading RFID tags due to the packaging materials used with Viagra (whether the bottle, safety seal, paper associated with secondary packaging, or other)? If so, what can you share?T.M.: We are not seeing any unforeseen issues with the HF tags on the bottles. Nothing leaves the packaging line with a dead or inactive tag. We have seen a small number of dead HF tags in the field (at the distributor level) and we’re working with our partners to understand why this might be happening. We expect the answers to be analogous to what we see for UHF failures for products tagged for Wal-Mart and Target — i.e., damaged during shipping.

We have received some complaints about the UHF case tag. This is because of the tags’ placement on top of the case. It’s not the optimal location for full pallet reads, we understand. But it was the only location feasible in the time span of the project.

P.M.: A few experts have suggested that plastics and glasses with certain additives could conceivably interfere with RF signals. Is there anything to suggest this with the Viagra pilot?T.M.: When you use traditional HF tags — that is, tags that are tuned to provide a specific resonating frequency — anything you put a tag on will affect its performance. The trick is to determine how much of an effect. If you have additives in the plastic or glass intended to enhance the capability of the packaging, you need to evaluate performance.

Tags on plastics, corrugated, glass and cartons see only small or marginal effects to resonating frequency — usually too little to make a real difference. The tag manufacturer can compensate for this if it’s evaluated in advance. Tags on metal or in close proximity to metal are a completely different story, however.

P.M.: Have other external materials (such as metals on packaging equipment, or warehouse carts or trays) caused interference or compromised the RF signals?T.M.: We’ve not tested to quantify effects from external metals in the RF environment to any great extent. (See answer to question below on optimizing read rates.)

This is a known issue for UHF and is one of the reasons we stayed away from UHF for item level. Our packaging lines are loaded with metals. I would point people to educate themselves about the realities of UHF in practice.

P.M.: Some companies/vendors have experienced increased RF signal strength due to the presence of certain metals in the vicinity. Can you comment on this?T.M.: Not really. You can’t get more power from a reader than what it emits by placing metal in the field — that would defy conservation of energy. Placing a metal wave guide, for instance, might channel RF in a manner that allows more of the available power to get to the tags you really want to see. I’ve seen wave guides for UHF in use and they work up to a point. Incidentally you don’t need a wave guide with HF because it’s not a propagating wave. With HF you control the shape of the magnetic field with antenna design and power adjustment.P.M.: Ambient moisture and humidity are also thought to impact signal strength and thus read rates. While pharmaceutical production and warehousing areas tend to be humidity-controlled, have you had any issues along these lines?T.M.: This is a fact and I believe there are even some published test reports available. We’ve not had issues to date and not likely to within our four walls. Pharmaceuticals need to be kept within controlled temperatures and humidity throughout the entire supply chain. It’s a bit more environmentally friendly than consumer goods sitting in warehouses with no temp or humidity control.P.M.: Have you been testing RFID (either HF or UHF) on blister packs and other types of packages, and if so, have you experienced readability challenges?T.M.: We’ve done some exploratory testing with HF on blister packs and even Rolaids (roll of tablets in a foil liner). In general, we would seek to minimize changes to packaging wherever possible and balance the needs of business and RFID performance.

The challenges will be to economically get the tag on the package in a location conducive to managing the effects of metal content. You can try and fight the fact that there is metal present. However, a better approach is to take into account the metal content and properly tune the tag. This is an old trick used on some EAS theft-deterrent labels. A problem can occur, however, where the tag will then not work properly in the absence of metal — which makes scalability an issue. Metal is more of an issue with UHF because even though you can tune UHF tags, metal still reflects the propagating wave component of near-field UHF. Thus you get RF in places you don’t want it.

P.M.: How much have you learned about optimizing RFID read rates through trial and error? Can you share an example or two of such a lesson learned?T.M.: Yes, there are numerous examples. Here are two worth mentioning:

We’ve taken an approach with the HF item-level tags that enables Pfizer to evaluate the tags on the packaging line in the same manner as the tag manufacturer and the label converter. So we learned that if all parties look at the tag with similar hardware, hardware setups (position, distance, etc.) and power levels, we can achieve a better process. We simply are better able to prevent weak or poor performing tags from getting through the process. This is necessary because we use a flexible inlay. It’s bent and contorted throughout its path to getting on the bottle. This can kill tags or make tags perform less than optimally.

We also discovered during live line trials that the UHF hardware was affecting the HF hardware. At one point on the line, an HF tunnel is directly next to a UHF antenna. The UHF was having a negative effect on the read reliability of tags going through the tunnel. We had to experiment with different settings and antenna orientation to compensate. We also added some logic into the process, which controls when the two different readers can be operating. This learning would be similarly applied to an installation where there happens to be a good bit of RF noise in the environment from motors, lights or other equipment.

P.M.: While it’s well known that water and liquid products can interfere with different RF signals, it’s generally thought that solids such as tablets have little or no significant impact (unless they have a significant moisture content). Has this been your experience? Are you pretty sure that the Viagra tablets are not impacting signals?T.M.: Yes, very little impact at the bottle level from SODs. However, when you read a full pallet of empty bottles with UHF tags on the cases, and then put tablets in the bottles, you get a lower read percentage. Even tablets have some absorptive RF characteristics, though minimal. I venture to guess it’s a cumulative effect of the mass of tablets absorbing RF and lowering the available RF to power tags to communicate. We’ve tested for this with our pilots on both liquids and solids to confirm that there was no effect to product efficacy or safety.P.M.: We would imagine that you’re relying heavily upon the expertise of your RFID equipment vendor partners and outside consultants right now. How long will it take to develop your in-house expertise to handle equipment/design and other challenges as they arise?T.M.: Yes, our partners were instrumental in the success of the pilot. It will be some time, perhaps years by current adoption rates, before Pfizer has colleagues at the packaging line level who are experienced with RFID in the same manner that they are experts with barcoding systems, vision systems and packaging equipment. Development of robust standards focused on the industry’s needs, like UHF Gen 2 tags and Gen 2 readers for consumer goods, is what the industry needs for HF. Then, colleagues can be less “RF experts” and more application specific.P.M.: These are all very technical issues. How much are you working with other manufacturers, academics, FDA on these issues? What significant steps still need to happen to gain a better understanding of how packaging can impact the success of RFID applications?T.M.: We collaborate to a high degree with the FDA, our customers and those active in groups like GS1 and EPCglobal. This is a patient safety initiative and we simply don’t feel it’s good for the industry if we keep all of our learning to ourselves.

We need alignment. We need more adoption. We need more than just a few piloting companies in a position to educate and inform the FDA on how well or not well the technology is suited for pharmaceutical safety and security. I don’t just mean manufacturers. We’ve got the attention of the major wholesalers, but we need more piloting to occur at the retail level. Then we’ll start getting some real-world examples of what packaging does to reliability. Lab work is good and necessary for making initial business decisions. But it’s how well it works in practice that will shape future direction.

About the Author

Paul Thomas | Managing Editor