Deploying Smart Process Sensors Sensibly

Aug. 14, 2013
Advancing sensor technologies improve choices and reliability
Today’s demanding biopharmaceutical manufacturing operations require careful control of process conditions, whether during cell culture, purification or drug product formulation. Process sensors play a critical role in enabling high-performance manufacturing by providing real-time visibility into how a process attribute is changing as well as the ability to correlate that change with the stage of the process.
“In the pharmaceutical industry, it is extremely valuable to see how an attribute changes with time and correlate that change with part of the process,” says L. Harry Lam, Ph.D., a biopharmaceutical manufacturing industry expert. “It is imperative to have effective equipment that provides reliable measurements.”
With the critical role that process sensors play in bioprocesses, selecting the best sensors for the job is the key to process success and understanding. To successfully deploy sensors, pharma manufacturers should consider following these principles:
• Define objectives, and find a sensor technology that provides the information needed to accomplish your goals.
• The sensible choice depends on the situation; for example, the size and age of the manufacturing operation and relevant regulatory guidelines help determine what is sensible in addition to the particular product being manufactured.
• Successful sensor deployment requires follow-through to implement the system correctly.
MATCHING SENSOR TECHNOLOGY WITH GOALS
The right choice of sensor begins with the particular manufacturing challenge being addressed. Different products require different controls and possibly different sensors. The first principle for successfully deploying sensors is that a manufacturer must clearly define the problem to be solved and identify the sensor technology that provides the data required to solve it.
“You must know what is important about the process, and then you can pick the sensor,” says Tina Larson, a bioprocess industry expert who heads up manufacturing operations and engineering for a major pharmaceutical company.
This principle is especially important to remember for organizations scouting new technologies. Although new technology is exciting and touted for its promise, focus must be maintained on identifying the best suited sensor technology to measure the target critical process parameter. “The key question is,” warns Larson, “do the data give you what you need to solve your problem?” There is no shortage of challenges that drive manufacturing operations to invest in new sensor technology. “Investment in new sensor technology is driven by those manufacturing challenges where there is the greatest amount of uncontrolled variability,” says Larson. “In biopharmaceutical manufacturing, that is the cell culture stage.”  
Accurately measuring cell mass in cell cultures, for example, is an unsolved problem that still eludes effective measurement. Another area of uncontrolled variability is glycosylation in cell cultures; researchers need a better understanding of the biological control of glycosylation to develop proper sensors for this complex problem.

“SUCCESS” DEPENDS ON THE SITUATION
In addition to seeking a sensor technology that provides the data needed to manage a critical process attribute, regulatory and business considerations also impact bioprocess sensor implementation decisions.
The recent U.S. Food and Drug Administration Process Analytical Technology (PAT) initiative is driving a closer look at bioprocess analytics by pharmaceutical manufacturers. The PAT guidelines urge manufacturers to update their sensor technology as needed so that they have data on those process parameters that affect the quality of their products. For example, antibody production is a core process in biopharmaceutical manufacturing. Methods for purifying antibodies and producing the antibody protein are well established, and bioprocess sensors such as pH and DO are standard. These sensors allow operators to manage critical process parameters in real time. Further, having routinely collected the data, they are now available for use in analyzing issues when they do occur, perhaps even allowing  manufacturers to spot success trends not previously recognized.  
The PAT guidelines point to best practices that ensure critical attributes are monitored and that provide insurance against process failures. Other factors to be considered in making a new sensor deployment decision are the size and age of the manufacturing plant, which affect the cost of making the change. “There must be an important business case for going through the change process,” states Larson. 
Process sensors measure a range of attributes, from traditional pH and dissolved oxygen (DO) to cutting-edge attributes whose relevance is still being proven. Whether a new sensor is state-of-the-art, a workflow innovation and/or cutting edge affects the implementation decision.
State-of-the-art sensors are newer versions of well-established sensors. They measure critical attributes that facilitate high-performance plant operation and biopharmaceutical production. “State-of-the-art” might mean more accurate, more stable, or easier to calibrate. Two examples of state-of-the-art sensors are optical DO sensors (see Technology Spotlight 1) and “smart” sensors (see Technology Spotlight 2). 

A manufacturing plant’s size and age are key factors in the decision to deploy up-to-date sensor technologies. “A small company building its first or second plant will use state-of-the-art technology so they don’t have to worry about obsolescence,” says Larson. “On the other hand, a manufacturing operation with several plants and a need to align operations globally will only incorporate such upgrades as part of planned obsolescence.”
Workflow innovation sensors do just what their name implies: they improve workflow efficiency. Smart sensors, now state-of-the-art, began as a workflow innovation to reduce human intervention. In place of a potentially wasteful preset maintenance schedule, the smart sensor notifies the operator when its performance is declining, resulting in a better optimized maintenance schedule. “At line” analytics, wireless sensors, and single-use sensors are other workflow innovations that are becoming more common in biopharmaceutical manufacturing.
With at line analytics, also called “QC on the floor,” a test sample is analyzed right on the plant floor3. At line analytics are useful for non-critical parameters or for cases in which a technology is not available in online format yet. The PAT guidelines indicate that critical process attributes should be measured online whenever possible. At line analytics allow manufacturers to test new or non-critical attributes without making breaks in the sterile barrier or impacting the design of the vessel or other capital equipment.   
Wireless technology is a workflow innovation that impacts virtually every aspect of our lives. In 2010, wireless technology reached the biopharmaceutical process sensor industry when Hamilton Company introduced its wireless technology platform. “The direct connectivity eliminates the need to send information through costly transmitters, the signals are more robust and reliable than the low currents produced in traditional measurement systems, and wireless communication to a handheld device enables users to move around the facility and readily monitor data from multiple sensors at a time,” says Hamilton Company’s Jahir Kololli.
“Beyond the advantages of convenience, flexibility, cleanliness and elimination of wiring costs, removing intrusive wiring ensures quality by reducing penetrations into the system,” explains Lindsay Leveen, a bioprocess industry strategist.
Like wireless technology, single-use or disposable technology is well known in the modern world. Certain applications lend themselves well to disposable technology. In biopharmaceutical manufacturing, single-use bioreactors were introduced in 19964, and implementation has become more sophisticated over time5. Single-use sensors are also taking advantage of disposable technology in their designs. Although many still have their limitations, new sensing technology has allowed for innovations in DO and pH measurement specifically, where the expensive electronics, which do not need sterilization, are not single use. 
Cutting-edge bioprocess sensor technology brings novel and unproven approaches to solving problems. This might be a new engineering approach, a new assay, or a new chemical measurement. Solid-state and near-infrared (NIR) sensors are two examples of cutting-edge technology for the biopharmaceutical industry. Solid-state technology is an exciting, promising innovation for bioprocess sensor engineering because of the potential for low-cost manufacturing and low power consumption. Near-infrared data processing, although not a new technology in itself, is experiencing renewed interest, driven by advances in NIR data processing that allow its use in aqueous situations. With these advances, NIR sensors are now being reconsidered for bioprocess monitoring. To what extent each type of cutting-edge technology will be useful in bioprocess sensors remains to be seen. 

FOLLOW-THROUGH 
“It is hard to get new sensors into manufacturing at a large pharmaceutical company,” says Larson. “You need a lot of data to make the case that the new sensor will be beneficial.”
“We test new sensor technology in our pilot plants,” she explains. “Once we have experience with it, and it proves robust and solves the problem that we are trying to solve, then we send the data to manufacturing operations for deployment consideration.” For sensor design engineers wondering how to get started with PAT, Larson advises a “sensible” approach. “The technologist designing a sensor must understand the manufacturing problem.”
Once the decision is made, the deployment itself requires careful attention to ensure proper setup and configuration. Thoroughly planning implementation ensures that the benefits of the new sensor technology are realized, while helping to minimize organizational and validation-related questions. Training, data collection, data analysis, integration into third-party equipment, version management and reporting are just a few of the considerations that require careful planning before implementation begins. Another important question to answer is how data acquisition technology will be used to maintain control of all the new data. 
With the goal of any process improvement being to initiate a new approach, it is important to have support, both internally and externally. Partnerships with sensor vendors extend beyond ensuring successful deployment. “Pharmaceutical manufacturers get involved in sensor design by surveying the landscape for new sensor technology and disposables and partnering with vendors in two basic situations: when they see a technology that they think is worth developing, and when a vendor is close to bringing to market a new technology that is of interest to the company,” says Larson. 
Regardless of the type of innovation being introduced, successful deployment of bioprocess sensors requires that implementation of a new technology follow PAT guidelines, be suited for the manufacturing process, answer the question it is meant to address, and add enough value to justify making a change. The right choice varies with the situation; different products require different controls and possibly different sensors, and the cost-benefit ratio depends on the size and age of the manufacturing plant. Adoption of new technology often results in unexpected improvements in process understanding and performance.  
Acknowledgements
The authors would like to thank the following industry experts: 

Tina Larson is a bioprocess engineer who has spent her career in biochemical manufacturing operations for major pharmaceutical corporations. In her current role, she is the head of technical development operations and engineering for a large biopharmaceutical manufacturing organization.   

L. Harry Lam, Ph.D., is a bioprocess engineer who has spent his career in biochemical manufacturing operations for major pharmaceutical corporations, where he has been responsible for technology implementation and many successful development projects and manufacturing campaigns. Lindsay Leveen, currently a consultant, is a manufacturing process industry strategist and early technology adopter focused on sustainable development. He led manufacturing technology implementation for biopharmaceuticals, where he was an early proponent of single-use methods.

References
1. Tillich D.  Hamilton Visiferm. DO Sensors: Optical Oxygen Measurement with Builtin Electronics in a 120mm Format. CHEManagerEurope May 2007:25.
2. Hamilton Company. October 7 2009. HAMILTON Unveils ARC Sensor Technology.
3. Bakeev, Katherine A.  2010.  Process Analytical Technology: Spectroscopic Tools and  Implementation Second Edition. John Wiley & Sons. 576p.
4. WAVE Biotech
5. Eibl, R, Eibl, D.  2011. Single Use Technology in Biomanufacturing  John Wiley &  Sons.  360p.

About the Author

Amber Ratcliff | Analytical Sensors Market Segment Manager