TECHNOLOGY

UHF Gen2 operating in the near field benefits yield and performance, even on small, tightly packed liquid-filled vials. Shown is Impinj’s Button item-level tag powered by the company’s Monza tag chip.

When one considers the many conflicting agendas, clashes of confusing disinformation, and the general level of fear, uncertainty, and doubt, it’s no wonder that the California Electronic Pedigree mandate has induced a high state of anxiety. And when it comes to serialization data carriers, the anxiety over options is further compounded. With that in mind, we’ll seek to sort out and simplify the maze of options in a straight­forward, no-nonsense manner to help enable you to make the most appropriate decisions for your organization.

Where serialization data carriers are concerned, there are two fundamental camps: 2-D Data Matrix bar codes and radio-frequency identification (RFID). (Please see Figure 1.) The RFID branch breaks out into two segments: ultrahigh-frequency (UHF) Gen2 and the variety of high-frequency (HF) technologies. Note also that a solution might be composed of a combination of data-carrier strategies. For the manufacturer, the choice determines where serialization occurs, how serialization must be performed, what effects the choice may have on the packaging line, as well as the overall serialization process itself—including who controls the serial (electronic product code [EPC]) numbers. These decisions, though, aren’t made in a vacuum. The serialization strategy the manufacturer ultimately adopts has significant consequences for downstream trading partners, trickling down through the entire supply chain.

As it turns out, where and how serialization will be realized are decisions that carry equal weight. First, such decisions determine who will control the serial numbers that will be applied to your products. Second, they define the extent to which you will have to modify your packaging line to accommodate unit-level serialization.

Mike Celentano, associate director of supply-chain systems at Purdue Pharma, an early adopter of serialization and RFID technology, shares the following perspective. “Purdue Pharma’s work with RFID began several years ago on a UHF Gen1, Class 0 platform. This technology required that the chip be encoded at the chip manufacturing foundry. Therefore, we embraced third-party encoding by default. Now that we have upgraded to Gen2, we have options. We choose to outsource the encoding of our item-level tags to our label converter for several reasons, including minimizing impact to the packaging line. This arrangement is meeting our needs.”

As you sort out your needs, your first consideration is determining control and management of the serial numbers. Will you generate them in-house? Will you rely upon a third party to manage and provide them to you? Once these issues are settled, the next questions are where and how the serial numbers will actually be encoded upon the data carrier(s). As seen in Figure 2, there are many options. There is also some fluidity in how the various schemes might be realized. For example, one approach might entail using preencoded UHF Gen2 RFID tags at the item level and printing 2-D codes in real-time for the cases.

Another approach might serialize 2-D codes in real-time on the packaging line, while the UHF Gen2 tags are bulk-encoded (with the same set of EPC numbers) all at once in the case at the end of the line. This solution maintains one-to-one correlation of the respective 2-D codes and RFID serial numbers and then associates the items with the case into which they are packed (a challenge that RFID solution provider Impinj has solved in a novel way with its Pharma Commissioning Station). This latter scheme might make sense for manufacturers that desire redundant data carriers initially, but wish to eventually phase out the 2-D operation in favor of an RFID-only end-of-line solution. For each set of choices, there are trade-offs.

Let us begin with an overview of the 2-D Data Matrix scheme.

2-D OR NOT 2-D

Figure 1. Data-carrier schemes for serialization. (click image to enlarge)

Let’s get right down to it: 2-D bar codes cost about a nickel; RFID tags, 15 to 20 cents or more. End of discussion? Well, maybe not. The 2-D quick fix may appear at first blush to be penny-wise, but one must look more closely to determine whether it is also pound-foolish. It will depend on how you’ve scoped your objectives. Let’s review a few details. First, 2-D is a “bar code” technology. As such, it is subject to the same limitations of conventional bar codes. For example, 2-D bar codes:

• Require line-of-sight.
• Are not rewritable.
• Are not secure.
• Take up precious label space.
• Cannot be written or read in bulk.
• Are easily counterfeited or cloned.
• Can limit the speed of packaging-line operations.
• Severely limit the speed of inventory operations.
• Are easily obscured, soiled, and damaged, thus precluding reads.

Each of these attributes carries a price tag: line-of-sight operations are manually intensive; encoding on the packaging line limits line efficiency; as an anticounterfeiting measure (the point of the California mandate), 2-D bar codes are easily defeated and quite susceptible to diversion schemes. Another aspect to consider is what it takes to deploy and maintain a 2-D-bar code–based serialization scheme on a packaging line. For starters, according to numerous manufacturers, each packaging line must be taken out of operation for 6–12 weeks while it is outfitted with the required 2-D serialization, printing, and vision systems equipment. And once it is back up and running, the packaging line efficiency could be affected in myriad ways:

• 2-D image quality and reliability necessarily limits line speed (to generally less than 200 imprints per minute), lest the 2-D images smear or distort or are otherwise unable to maintain a B grade and above in print quality.
• Unlike RFID, 2-D bar codes are not printed and verified in the same step—they require downstream validation, adding to line and operational complexity
• Troubleshooting is complicated:
  • Is the error due to encoding?
  • Is it an image production or detection error?
  • Is the laser or camera out of spec or calibration?

In the maintenance arena, each new label type may require a different camera positioning, and, potentially, a different line speed. Consequently, 2-D bar codes require a great deal of manual attention, monitoring, and intervention during line operations.

A variety of UHF Gen2 tag designs make it possible to identify the best tag for a given application, taking into account the product form factor, package material, and dose form. Shown is Impinj’s PaperClip tag powered by Monza tag silicon.

When the product advances to case-pack operations, things don’t get much better. The integrity of the item-to-case association created at the point of aggregation is vital, particularly for downstream trading partners that are dependent upon such data hierarchy (let alone the accuracy of the electronic pedigree). Case-level aggregation of 2-D-serialized items is typically a manual process, which means it is also an error-prone process. And errors are costly—particularly when you’ve lost visibility of the individual products’ serial numbers. If the packaging operation also includes bundles or inner-packs, unit-level visibility of the serial numbers is lost even before the product gets put into the case. It doesn’t take much imagination to see what this means to rework operations or efforts to reconcile discrepant cases. As such, exception handling with 2-D-based data carriers becomes unwieldy.

For these and other reasons, Purdue’s Celentano says, “We never seriously contemplated 2-D as the primary serialization data carrier. Guidance from FDA and major trading partners was pointing toward RFID. Additionally, our serialization, data-collection, and data-aggregation strategy was always designed with a mind-set toward minimum disruption and maximum transparency to our packaging operations. To us, this meant RFID, given its unique capabilities for transcending the limitations imposed by direct line-of-sight or packaging orientation-sensitive alternatives. This has enabled us to effectively maintain our pre-RFID packaging line speeds. We are only considering 2-D as a redundancy feature at this time.”

Figure 2. Encoding options for serialization. (click image to enlarge)

So while the use of 2-D bar codes as a serialization data carrier, technically speaking, does satisfy the California serialization mandate, the larger question is, after weighing its shortcomings and hidden costs, does it serve the objectives at hand? What about the total cost of ownership, a cost that extends to all downstream trading partners? This is particularly important to wholesalers, whose operating efficiencies are severely affected by the logistical requirements of line-of-sight technologies.

McKesson’s Kevan MacKenzie, senior solutions analyst and a Six Sigma black belt, explains that the McKesson distribution centers ship and receive thousands of cases of pharmaceutical product daily. “The advent of pedigree and serialization has added to the complexity of both shipping and receiving,” he says. “But the additional impact of 2-D bar codes at the item level will have a large detrimental effect to productivity, forcing wholesalers to use inference. When a pallet of product is received into the distribution center, it is impossible for the wholesalers to break down every pallet to scan the serialized identifier on every case.

“Additionally, it is extremely impractical for the wholesalers to open every case that they receive to scan the serialized identifiers on every item, as this would have an even further impact to productivity. McKesson has customers that purchase product only in full-case quantities and contractually demand that their cases have not been opened. If wholesalers are required to open cases to capture serialized identifiers on eaches marked with 2-D bar codes, we believe it will introduce additional risk to the supply chain.”

MacKenzie adds that when product is picked in the McKesson distribution centers, it is usually at the item level. “The use of 2-D bar codes at the item level requires wholesalers to manually scan every item using a line-of-sight optical reader/camera. The manual scan process must be repeated for all items in totes prior to it being shipped to our customers. McKesson must ensure that all items in the tote have the appropriate electronic pedigree associated with the serialized identi­fier.” The same goes for the pharmacies that receive and check-in the product.

Figure 3. UHF Gen2 is the RFID “superset” for RFID applications, addressing items, cases, and pallets, and all dose form and package types. It exhibits throughput rates much higher than those of HF.

McKesson further believes that the use of 2-D bar codes will slow the eventual adoption RFID and will increase costs throughout the supply chain. “All of our customers and suppliers have a common need to ex­change critical information flowing into their enterprises,” MacKenzie says. “Data collection is a very expensive and time-consuming process. Business has learned that when manual processes are involved, these processes can be error-prone. Accuracy and precision of data capture by using RFID ensures that the manual processes in operations are minimized. RFID introduces speed and accuracy into the capture of pedigree data that enables patient safety.”

It is easy to see how the stated 2 to 3-cent cost of 2-D codes can rapidly escalate when introduced into the supply chain. An RFID-based solution, on the other hand, is simple, robust, reliable, and easy to deploy, automate, and maintain. But what kind of RFID solution?

RFID AND THE GREAT FREQUENCY DEBATE

In December 2004, EPCglobal ratified the UHF Gen2 Protocol, creating the first worldwide standard for RFID. Since then, myriad products built to UHF Gen2 have proliferated, proving their functionality in applications ranging from items to cases to pallets. They can be used with objects in the near field and far field, and on liquids, metals, tightly stacked and packed items, and more. And the performance levels and innovative application of these products continue.

Three years later, the HF camp is yet to answer the UHF Gen2 achievement with a ratified standard of its own. Instead, the latest HF candidate specification has left even its own authors disillusioned. According to Ken Laing, the editor of the proposed HF “V2” (the HF version of UHF Gen2), their work thus far offers little if any improvement over existing standards and commercially available products. “Companies encoding an EPC on a Gen2 HF tag,” he says, “will get an improvement over encoding EPC in ISO 15693 [the current prevailing HF standard]. Will it be a significant improvement? I suggest not. You’ll get something marginally better than [HF] products on the market now.”

Perhaps the larger point is that even if ratified, the first qualified V2 products will follow by a year or more. That’s a long time to wait for something that, even if available today, will fall far short of UHF Gen2’s current capability. And given the pace of UHF Gen2 adoption and deployment innovation, imagine the advances that will occur in that time frame. But let’s step back and take the longer view of the frequency debate as it relates to the pharma space.

First, consider a few basic facts: UHF Gen2 covers all supply-chain applications, near and far, worldwide. Furthermore, UHF Gen2 covers all dose forms (liquids, gel tabs, powders, etc.) and package types (blister packs, bottles, vials, etc.). As such, HF RFID technology is actually redundant to UHF Gen2 in the following ways:

• There is nothing that HF can do that UHF Gen2 can’t do.
• There are a great many things that HF can’t do that UHF Gen2 can.
• HF addresses only a small subset of the UHF RFID universe (see Figure 3).

A properly deployed near-field UHF Gen2 system operates equally well on items large and small, liquid and metallic, cases and pallets. Such capability erases any item-level advantage HF enjoyed prior to the advent of near-field UHF.

Notwithstanding, let’s dig a little deeper and examine the practical implications of deploying HF-based RFID systems. First, HF cannot perform in the far field; that means it cannot be used on cases and pallets, where longer-range-read capability is required by warehouse and distribution center logistics. Consequently, enterprises that opt for HF for item-level tagging must also implement UHF Gen2 for case and pallet operations. Now, once you cross the border into the land of multiple data carrier infrastructures, cost, complexity, efficiency, and maintenance—to say nothing of data logistics—take major hits.

UHF Gen2 tags will always be less expensive than HF tags. In fact, two to three times less expensive. Why? Because UHF tags are so much easier to manufacture. Unlike HF tags, UHF Gen2 tags lend themselves to simple, high-speed manufacturing techniques that scale exceptionally well. By virtue of UHF Gen2’s uncomplicated antenna geometries that can be manufactured using inexpensive conductive ink processes, UHF is a far more practical, economical band on which to base a standard.

Impinj’s collaborations with the pharmaceutical packaging industry produce innovative solutions, such as this Rexam RFID-enabled pill bottle with a UHF Gen2 tag embedded in the package itself.

Continuing advances in the application of UHF Gen2 further widen the cost-performance-capability gulf that already exists between it and HF technologies—a gap that HF will never be able to close. This is a fundamental point, as the economics of UHF Gen2 actually benefit from UHF physics.

In short, the UHF band is 60 times more efficient for RFID operations than the HF band (per Faraday’s law). Therefore, if the objective is to couple communications between a tag and a reader, the UHF band is vastly superior to the comparatively weaker HF scheme. And that translates to UHF Gen2’s greater speed, reliability, and flexibility of operation. It’s also why Blue Vector CEO Nancy Anderson concludes, “We’re not doing a lot with HF right now; it’s just not as flexible as UHF.”

For a more direct case in point, Cardinal Health’s Julie Kuhn, director of pedigree, explains, “You can’t read an HF RFID tag as fast as you can read a UHF tag. That means our conveyor can’t run any faster than the slowest read rates at the reader.” And this is a limitation that could actually affect distributors’ order throughput. “Right now,” she continues, “we take orders up to 8 p.m. and orders are delivered as early as 5:30 a.m. in the morning. This complicated [HF/UHF] infrastructure could inhibit our ability to maintain those cutoff times.”

This raises the issue of multiprotocol infrastructures generally. Unfortunately, addressing these problems with multiprotocol RFID readers—devices that read both HF and UHF tags—only creates more problems. These include higher reader complexity and cost, and, more critically, lower read rates and less overall read reliability. That is because read opportunities are always compromised when an interrogator must cycle through multiple protocols. And when multiple data carrier technologies are in use, the supply chain suffers.

Kuhn continues: “We are concerned about how we’re going to make all of these technologies work in a single, highly automated environment to get the pedigree information at the unit and case level and maintain the high level of throughput that we have now.” The use of HF in the pharma supply chain not only contributes no value over UHF Gen2, it actually hinders efficiency.

Pharma is not the only industry to grapple with this issue and arrive at the same conclusion. Consider, for example, a major item-level deployment by the apparel manufacturer, Lemmi. Ashley Stephenson of Reva, one of Lemmi’s RFID solution providers, had this to say: “To date, Lemmi’s deployment has used HF technology, but [given] the higher speed, better flexibility, decreasing cost, and improved performance of UHF, the company is swapping out HF for UHF. If you were trying to read 100 garments, UHF could read them all 10 times faster than HF.

“The cost of UHF tags has come down markedly, a trend that is only expected to continue. UHF vendors like Impinj have invested in and advocated a flavor of UHF technology called ‘near field,’ thereby expanding the range of UHF infrastructure options. Fashion and retail RFID tagging did start with HF, but now everything is migrating to the UHF Gen2 standard in the retail and fashion world.”

Purdue’s Celentano concurs. “We have seen major improvements to our program on every level with our conversion to UHF Gen2 technology. Reliability, performance, tag availa­bility, multisourcing, interoperability amongst vendors—it’s across the board. These are the very baseline requirements we needed to evidence before making the decision to tag our entire domestic OxyContin production. UHF Gen2 has enabled us to do our case- and item-level tagging on a single technology platform. All of our major downstream trading partners that have issued guidance on package serialization have indicated a preference for, or at minimum, an acceptance of the UHF Gen2 platform. We see this as a strong affirmation of the technology.”

HOW ABOUT A LITTLE OF BOTH?

Let’s assume you’ve settled on UHF Gen2 as the data carrier of choice for RFID-based serialization. However, for certain business reasons, you’ve also determined that your needs will be best met by a “hybrid” 2-D/RFID approach—in other words, redundant data carriers, using 2-D codes as backup. To this end, there are two strategies for the coexistence of 2-D codes and UHF Gen2 RFID—one that maintains the 2-D and RFID serialization operations on the packaging line permanently, and an alternative off-line solution.

No matter the level of deployment, because it must be encoded serially, 2-D codes necessarily alter the packaging line (unless, of course, labels are preencoded by a third party). Packaging line configurations vary, so there is no standard, off-the-shelf solution. And it is entirely possible, if not likely, that the 2-D serialization installation must be tailored for every line. Obviously, this does not scale well for manufacturers with multiple packaging lines, and consequently, it is a major impediment to adoption. That said, Impinj does support such installations, having deployed several major examples that combine 2-D and RFID operations on the packaging line, and as a result understands the challenges well. Experience with these deployments also suggests that the packaging line is not the best place to perform serialization, for reasons that include the following:

• The multiple data carrier packaging line deployment is fragile; it requires a great deal of fine-tuning and hand holding.
• Supporting several data carriers multiplies costs and complexities for all trading partners, while also imposing operating inefficiencies.
• Synchronization of serial numbers between multiple data carriers adds complexity and overhead.
• The use of multiple data carriers introduces multiple points of failure.
  • RFID and 2-D both must work.
  • Rework or waste increases when one or the other fails.
  • Line efficiencies are negatively affected.
• These systems are monumentally difficult to scale, and they certainly cannot be scaled sufficiently to meet the California mandate—even with a delay—with any critical mass of serialized product.

If viewed in the traditional way, that is, performing serialization in-line with 2-D—with or without RFID—the task is both enormous and costly. Further, the resources for integrating such complex in-line systems are constrained by the limited number of solution providers working in this area. This could be why so many pharma companies are struggling with electronic pedigree requirements and consequently seeking to delay the mandate.

Impinj’s RFID-based Commissioning Station was designed to ease these problems and make it possible to quickly encode every product mandated by the California law. The idea here is to perform the RFID serialization operation at the end of the line, encoding the items all at once after they’ve been case-packed—and even if a 2-D operation is performed in-line. This is not because RFID is any more difficult to encode in-line—in fact, it is both easier and faster than encoding 2-D.

But performing the RFID operation off-line provides long-term strategic benefits. As such, you can think of 2-D as a bridge to a far more efficient RFID-only serialization solution, and one that provides a means of seamless transition. This arrangement also serves piloting efforts, where the two data carrier approaches can be evaluated simultaneously, but with minimal impact to the line operations. In this scenario, the Commissioning Station keeps track of the EPC numbers encoded onto the item labels (pre­encoded or encoded in-line) and uses certain data captured from the line operations to write those same EPC numbers into the appropriate RFID tags also carried by the item labels. In other words, even though the RFID tags are written in bulk in the case at the end of the line, for each item, the serial numbers written into both the 2-D label and the RFID tag match.

John Schroeter is marketing director for RFID, Health & Life Sciences, at Impinj, a Seattle-based provider of UHF Gen2 products and system solutions. Prior to joining Impinj, Schroeter held marketing management posts at UTMC, Seattle Silicon, and Fairchild Semiconductor’s memory and high-speed logic division. He is the author of the Prentice Hall book on integrated circuit design, Surviving the ASIC Experience.