Commercialisation challenges for printed/flexible/stretchable and functional fabric sensors and sensor-based systems

Roger H. Grace, president, Roger Grace Associates

This article—a sequel to the one entitled Printed/flexible/stretchable sensors: new technologies enable high-volume applications in the November 2015 issue of this publication¹—attempts to address several of the most significant commercialisation issues of these exciting new technologies. The materials contained are the result of over 65 personal interviews I have conducted since embarking on this effort in September 2015.

Whereas the previous article primarily discussed the various technologies and applications of printed/flexible/stretchable (P/F/S) sensors and sensor-based systems, this article addresses the progress made in the approximately 30 months that have elapsed since the previous article’s publication and focuses on several of the most important commercialisation issues of P/F/S and functional fabric (FF) sensors and sensor-based systems.

I find it quite useful to refer back to my Annual MEMS Industry Commercialization Report Card—which I initiated in 1998 and was last published in the September 2015 issue of CMM²—as a means of determining what the critical success factors (CSFs) of these technologies are. The results of the 2016 MEMS Commercialization Report Card are given in figure 1 and provide the grades for the 14 CSFs. The commercialisation topics that I borrow from the study and primarily address in this article include:

• established infrastructure, including manufacturing equipment, packaging/interconnects, design/layout tools and reliability models;
• venture capital (VC) attraction; and
• cluster development.

I propose that there have been several waves of sensor technology over the past approximately 50 plus years (figure 2); starting with electro-mechanical, then MEMS, then P/F/S and then FF. Prof. Steve Walsh, a fellow CMM editorial advisory board (EAB) member, and I conducted a major market research project many years ago to determine the duration of the commercialisation process for several MEMS sensor types. The shocking news was that it took approximately 30 years for full commercialisation to occur³ (figure 3).

Most MEMS technologies have already been commercialised and are now in their mature phase, which means that they are currently exhibiting only incremental improvement over time and are considered commoditised. Constant and widespread research is currently underway into the new sensor platforms—which include paper and plastic, which can be printed, and fabric, which can be woven. This is producing a myriad of new products, which I believe will be the cornerstone for the proliferation of the acceptance of these technologies by design engineers to meet the requirements of applications. It will be interesting to see what the commercialisation process timing will be for various P/F/S and FF sensors.

Motivating factors for adoption

According to my colleague Dr. Janusz Bryzek, in his article Trillion Sensors Initiative⁴, reducing the average cost of a sensor is a requirement for achieving his trillion-sensor goal. This is for the annual volume of sensors to reach the trillion unit level, leveraged by the adoption of ultra-low-cost sensors, many being fabricated from paper, plastic or threads/fabrics. Sensors using low-cost substrates are the key to this. It has been noted by several industry pundits that the cost of paper- and plastic-based sensor materials are approximately two orders of magnitude less expensive on a square area perspective than those made from silicon, the material of choice for today’s popular MEMS devices (figure 4). This alone is a compelling factor for the accelerated adoption of these technologies. Another is their physical shape and their ability to conform to complex three-dimensional structures, for example, the human body, making them ideally suited for wearable applications.

Figure 5 shows the cost and functionality of P/F/S and FF sensors and electronics. As stated previously, I have referred back to my MEMS Industry Commercialisation Report Card to address the CSFs of these technologies. Furthermore, I have spoken to a number of industry cognoscenti to gain their observations and opinions on these CSFs. 

Established infrastructure

I believe that a well-established and broad infrastructure is most significant for the successful commercialisation of a technology. In respect of P/F/S and FF technologies, this encompasses:

• manufacturing equipment and materials/chemicals;
• design, analysis and layout tools;
• formal reliability analysis tools;
• widespread adoption of design for manufacturing and test design for manufacturing and test (DfMt) principles; and
• widespread development and adoption of standards for manufacturing and test.

 At my annual visit to the IDTechEx Show!, which took place in California, US, on November 15–16, 2017, it was quite evident, vis-a-vis the many exhibitors, that the industry has a significant number of players providing capital equipment, including printing equipment, as well as films/substrates and chemicals, for example, specialised inks, to the P/F/S and FF sensors and sensor-based systems providers. Since these tools are widely commercially available, it is contingent on the sensor/electronics designers to use them in their ‘toolbox’ to create sensors and sensor-based systems using P/F/S and FF formats.

Manufacturing processes

From a manufacturing perspective, P/F/S sensors and electronics can be manufactured using either a sheet/batch mode process, similar to the process in the integrated circuit (IC) industry where each substrate is made in a one-step-at-a-time fashion, or a roll-to-roll (R2R) process, i.e. in a continuous fashion. Each has its pros and cons.

I obtained opinions on both from two highly knowledgeable experts in the printed electronics manufacturing process area, namely Jaye Tyler, president and CEO of Si-Cal, a manufacturer of heat transfer decals and functional printed electronic components, and Wilfried Bair, senior engineering manager at NextFlex, a consortium of academic institutions, companies, governments and non-profits geared towards advancing US manufacturing of flexible hybrid electronics (FHE).

Si-Cal offers both approaches to clients for producing their products, several of these being sensors. Jaye explained: “We are able to provide our customers with a choice of manufacturing methods for creating their sensors; sheet-fed screen printing or roll-to-roll screen printing. We work with them to determine the best solution for their requirements based on product specifications, total cost and production run volumes. 

“We consider sheet-fed screen printing to be the best approach for small volumes (figure 6), which are not cost-effective using R2R screen printing. The sheet process is better for thicker, 3 to >15 mm films due to processing issues with thin films; a minimum thickness of 3 mm is required but ≥4 mm thick films are preferred. R2R can go down to 2 mm thick for 50 μm films and even 1 mm for 25 μm films with multiple layers of functional inks.

“The sheet process can print in a 30+ in. (76.2+ cm) format versus around 20 in. (50.8 cm) for R2R. Sheet colour-to-colour registration is 0.004 in. versus 0.006 in. (0.1016 mm vs. 0.1524 mm) for R2R. The sheet process needs camera alignment and gets slower to match R2R. From a line and spacing perspective, the sheet’s superior performance is 100 μm versus 150 μm for R2R.

“On the other hand, registration is much easier in R2R (figure 7). It has a faster throughput and can be more cost effective due to multiple ink layers being printed continuously in a line. R2R requires far less handling and is easier to keep clean. There is uniform shrinkage and handling of films with R2R. Typical R2R production speeds for printed electronics are 30–40 sq.ft/min (2.79–3.72 sq.m/min), which is substantially faster than sheet.”

He concluded: “There is a great deal of discussion and trade-off necessary for us to provide our clients with the optimum solution for their current production requirements.”

According to Wilfried, a 30 plus-year veteran of the semiconductor/MEMS business: “From a throughput and cost perspective, the common expectation is that R2R processing is less expensive and has higher throughput than sheet-based processing. This may be correct for some printing processes but has not been demonstrated to be the case in assembly/component integration processing.

“Printing speed is determined by the print method and will be faster on R2R for some highly automated processes, for example, screen printing. Common substrate widths for R2R are 6, 8 and 12 in. (15.24, 20.32 and 30.48 cm). Sheet formats are typically 8 x 12 in. or larger. R2R processing should prove to be less costly for very high-volume process of the same device. Equipment cost is higher for R2R and also limits flexibility of a line. With sheet-based processing, I can easily select the process path. R2R has a fixed setup of processing equipment and sequence. Since there no high-volume functional hybrid electronics (FHE) processing at this point, there is no empirical cost data.”

He added: “I run into people every week with the preconceived notion that R2R is an inexpensive way to build volume. When asked, they have no facts to support this assumption. To the contrary, they have typically not even considered the difficulty of making an R2R system work with very thin substrates.”

I also spoke to Prof. Ahmed Busnaina, a professor at Northeastern University and the founding CTO of Nano OPS, a newly launched micro-nano electronics printing equipment company. Nano OPS’ development of this equipment is based on the significant research conducted for a nanoscale sensors and electronics project funded by the National Science Foundation (NSF), a US government agency supporting research and education in non-medical fields of science and engineering. Ahmed said: Our development team, which is well-founded in electronics and sensor design and manufacturing, has been able to come up with a novel piece of equipment which we think will revolutionise the printed electronics and sensors industry (figure 8).

“Our fully automated nanoscale printing system brings together several functionalities into its six-module hub or cluster configuration and provides many of the processes to create a printed electronic function or sensor. It is 10 to 1,000 times less expensive than conventional fabrication processes and it can print circuits 1,000 times faster and smaller than 3D printing.”

Packaging/interconnects

Another major element in successful commercialisation of these technologies is the issue of packaging/interconnects strategies. There is much to be learned from the integrated circuit (IC) as well as the MEMS industry on this topic. 

The future success of P/F/S and FF electronics (and their embedded sensors) resides in the ability of suppliers to judiciously select the required functionalities of the individual circuit functions and determine the optimum format for their realisation in specific applications and their integration into an optimum solution for those applications. This encompasses:

• thorough assessment of monolithic versus heterogeneous versus hybrid integration strategies;
• adopting a systems solutions approach⁵ where the sensor is an integral and important part of an assemblage of various electronic functionalities and packaging/interconnects which provide an entire solution to the client’s application (figure 9);
• connectivity of individual devices to each other and to the substrate;
• packaging/encapsulating of the total solution;
• creation of models and software tools for the design/layout and reliability determination of these technologies; and
• determination of the manufacturing/integration/assembly method to minimise cost and maximise functionality and reliability.

Packaging

Dr. Leland (Chip) Spangler is founder and CEO of Aspen Microsystems—a provider of MEMS and microelectronics product design and consultancy services—and another 30- plus-year veteran of the MEMS business. He said: “Sensors face many of the same issues regardless of the technologies that are used to make them. The fidelity of a sensor’s signal and its reliability can be impacted by stress, temperature and electromagnetic fields, so a practical sensor must be impervious to these stresses while still providing the sensitivity to the desired factor to be sensed. 

“MEMS and silicon-based sensors benefit from the huge infrastructure and deep scientific understanding of materials and processes from the IC industry. These devices often have an explicit package that can be optimised specifically for the application. P/F/S and FF sensors represent a fundamentally new sensing paradigm and thus cannot leverage an existing infrastructure of knowledge. These types of devices will require new materials, manufacturing methods and scientific knowledge, which will require a great deal of investment. Since P/F/S and FF sensors have no explicit package, the materials and structures that perform the sensing function must also perform the packaging function by providing robustness over long periods of time while being used by consumers in wide ranging conditions, all without degrading the desired sensor signal. This will be a huge challenge to overcome.”

Reliability

For P/F/S and FF-based sensors and electronics to find their way into current and future products, reliability analysis is critical. Allyson Hartzell is managing engineer at Veryst Engineering—an engineering consultancy and services provider for product manufacture—co-author of the book MEMS Reliability and a 25-year veteran of the MEMS industry. She said: “Knowledge of known failure mechanisms, material properties and assembly methods can result in reducing cycle times through designing for manufacturing, yield and reliability. With new technologies, there will always be new failure mechanisms that are unknown, but if a company in the early stages can develop a product that doesn’t succumb to already known failure mechanisms, the product has a higher probability of success. 

“For instance, new developments in interconnect technology for P/F/S and FF sensors is of high priority. The same can be said of multi-chip packaging, which allows sensors and microprocessors to be packaged in a very tiny form factor.  Coupling the sensor packaging and, for instance, a flexible interconnect, you will still have the need for low-contact resistance over time with a low-loss interconnect that provides the proper impedance and conductivity without early failure.”

Software tools

Finally, designers and developers of these technologies require the software tools to design and develop the devices/systems. Once again, lessons learned from the IC and MEMS industries play a vital role in this learning process.    Dr. Mary Ann Maher, the founder and CEO of SoftMEMS—a provider of these software tools—and a 20-plus-year veteran of the MEMS business, shared her views: “The MEMS market has matured to the point that there is now a strong supplier ecosystem for design for manufacturing and test. This ecosystem is still developing for P/F/S and FF sensors. The flexibility of these sensors adds complexity with regards to their modelling, testing, reliability and performance specification.

“Some similarities between the two technologies are that there are a variety of manufacturing methods and materials used to create devices and products and few standard device structures and models, i.e. intellectual property (IP) as in the IC industry. One of the current promising technology approaches is the combination of mature MEMS devices and flexible substrates.”

Funding opportunities

I believe that the major route to the success of startup P/F/S and FF companies is to create a collaboration with various members of the value chain/ecosystem. Materials and chemistry play a significant role versus previous scenarios where the mechanical and electrical engineers dictate the direction.

It is encouraging that Brewer Science—a manufacturer of materials for the fabrication of microdevices—had the foresight to develop P/F/S sensors based on its internal materials/chemicals IP. I believe that organisations specialising in materials/chemicals that do not wish to and/or do not have the competencies to perform in-house development in P/F/S and FF sensors and electronics should consider acquiring the IP from startup developers and buy these companies.  

A most interesting story broke during my writing of this article. BeBop Sensors, in Berkeley, California, received US$10 mn in a Series A funding round led by local venture capital (VC) firm Bullpen Capital for the development of its fabric-based bend sensing technology. I hope that other VCs see the merit of this and follow suit.

Cluster development

The US Department of Defense (DoD) recently funded two significant organisations to take the lead in the development of manufacturing approaches for the commercialisation of P/F/S and FF technologies. NextFlex, in San Jose, California, focuses on P/F/S sensors and electronics, and Advanced Functional Fabrics of America (AFFOA), in Cambridge, Massachusetts, focuses on fabrics. 

Uniforms, tents and parachutes are just a few of the items for which the DoD is considering the integration of P/F/S and FF technologies. My recent visits to AFFOA, UMass, a public research university institute in Lowell, Massachusetts, and the US Army Natick Soldier Research and Development Engineering Center (NSRDEC) in Natick, Massachusetts, have been truly eye-opening in terms of the many military applications that these technologies are being targeted to support.

Based on my comments in the article Technology Clusters and Their Role in the Successful Commercialisation of Micro and Nano Technologies, published in the March 2012 issue of CMM⁶, I believe that the following research facilities provide a cornerstone for future spin-out companies which, in turn, are likely to form vibrant organisations for the commercialisation of these technologies.

AFFOA

Founded as part of a MIT concept under the direction of Prof. Yoel Fink in 2016, AFFOA has built its Fabric Discovery Center headquarters in Cambridge, Massachusetts, in close proximity to MIT, for the purpose of enabling the manufacture of sensors and electronics that will be embedded in threads and/or fabrics. An initial US$75 mn grant from the US DoD has been matched by approximately US$250 mn from academia, governments and industry through its 100 plus members.  

AFFOA’s strategy is to create a national network of advanced fabric startup incubators and connect them with market facing companies to enable existing product ideas to emerge across the US. It acts as a programme manager for the evaluation of proposals and dispersal of grants to member organisations for advanced development of manufacturing strategies for this technology.

NextFlex

Founded in 2015, NextFlex opened a 1,394 sq.m. facility in San Jose, California, for the manufacture of sensors and electronics on P/F/S carriers. It also received US$75 mn funding from the US DoD. 

Like AFFOA, NextFlex is a collaboration of academia, governments and industry and acts as a programme manager for the dispersal of funds to its member organisations for advanced manufacturing project initiatives.

Fabric Discovery Center, UMass Lowell

I visited the Fabric Discovery Center in early April 2018 to see a splendid work in progress. Both AFFOA and NextFlex have invested in this facility. Under the careful supervision of its technical programme manager, Cheryl Gomes, this facility will provide testing and evaluation of sensors and electronics from both printed and woven technologies. The facility is expected to be in full operation in Q4 2018.

It is interesting to note that Lowell, Massachusetts, was the world’s leader in the weaving of fabrics from its beginnings in the 1820s until the early 1900s when business moved to the Southern US because of the lower cost of labour. The Commonwealth of Massachusetts has invested significant resources in the Fabric Discovery Center and AFFOA in the hope of reviving the ‘smart’ fabric industry in Massachusetts.

CSFs summary

The following is a summary of what I consider to be the most significant CSFs for the commercialisation of P/F/S and FF sensors and sensor-based systems:

• ability of P/F/S and FF sensors and electronics to demonstrate value-add to critical applications;
• ability of P/F/S and FF sensors and electronics to demonstrate equal-to-better performance versus non-P/F/S and FF embodiments (my market research to date shows this is already happening);
• the lack of certainty in the growth of the wearables market;
• ability of non-flexible/non-organic functional solutions to accommodate application requirements;
• entry of additional players (with big R&D budgets) to help commercialise the technology;
• adoption of standard processes based on the broad availability of materials;
• continuation of robust funding of research activities in P/F/S and FF sensors and electronics by governments;
• ability of existing sensor companies to adopt a paradigm shift from an electromechanical technology mindset to a materials one; and
• continued over-hyping of flexible electronics opportunities.

The bottom line is I believe that there is a great deal of uncertainty at this time. The outcome of P/F/S and F/F electronics commercialisation will be a determining factor for the expected outcome of P/F/S and F/F sensors.

Summary and conclusions

I have listed a number of the CSFs that I deem important for the commercialisation of P/F/S and FF sensors and sensor-based systems. Below I have summarised the results of my market research findings.

1. P/F/S and FF technologies possess qualities that make them unique and ideal for many sensor and sensor-based applications, specifically they are:

• low-cost due to large-format batch mode and scalable R2R manufacturing processes; and
• small, lightweight, thin and complex shape conforming.

1. There are a limited number of:

• success stories to date;
• commercial companies in production; and
• parameters currently being measured, specifically pressure/force/touch, gas analyses, temperature and humidity.

1. Rapid growth is foreseen by market research groups. Lux Research anticipates a compound annual growth rate (CAGR) of 35 percent from 2014 to 2024, and IDTechEx a total market value of US$8 bn by 2025. This will be driven by:

• a replacement of existing solutions; and
• the creation of new solutions based on the enabling nature of P/F/S and FF technologies.

1. Commercialisation success will require a significant marketing/application pull versus technology push.
2. A systems solution/hybrid approach will be required.
3. Wearables for sport and fitness, disposables, package tracking, e-healthcare and environmental monitoring are likely to propel the market.
4. The existing and projected future low cost of capital equipment and therefore low capital expenditure (CapEx) as well as the existence of organisations that manufacture these devices makes rapid commercialisation by many organisations possible and means the barrier to entry is low.
5. The success of P/F/S and FF sensors and sensor-based systems will weigh heavily on the success of P/F electronics.

As a final summary note, as the MEMS industry must learn from the IC industry, I believe that the P/F/S and FF industry must learn from the IC and MEMS industries. To quote George Santayana: “Those who forget the past are condemned to relive it⁷.”

Learn more

Roger Grace has organised and will chair an all-day Printed/flexible/stretchable and functional fabric sensors and sensor-based systems preconference symposia on June 26, 2018, at the Sensors Expo & Conference 2018, in San Jose, California, US. For more information—including agenda, abstracts, presenter biographies and registration information—please visit www.sensorsexpo.com. Use the code ‘Rgrace100’ to receive a registration discount.

In addition, Roger Grace is organising and will chair a Printed/flexible/stretchable and functional fabric sensors and sensor-based systems technical session at the MANCEF Commercialisation of Micro, Nano and Emerging Technologies Conference (COMS 2018), in Montreux, Switzerland, on September 24–26, 2018. For registration information, please visit www.mancef.org.

Acknowledgements

The author would like to acknowledge the following individuals for their valuable inputs to this article: Wilfried Bair, senior engineering manager, NextFlex; Prof. Ahmed Busnaina, founding director, NSF Nanoscale Science and Engineering Centre for High-Rate Nano-Manufacturing, Northeastern University, director of the NSF Centre for Microcontamination Control, Northeastern University, W.L. Smith chair professor, Northeastern University, and founding CTO, Nano OPS; Allyson Hartzell, managing engineer, Veryst Engineering; Dr. Mary Ann Maher, founder and CEO, SoftMEMS; Dr. Leland (Chip) Spangler, founder and CEO, Aspen Microsystems; and Jaye Tyler, president and CEO, Si-Cal.

Roger Grace Associates

www.rgrace.com

References

¹Grace, R. (2015). Printed/flexible/stretchable sensors: new technologies enable high-volume applications. CMM; volume 8, issue 7. Available at: https://flickread.com/edition/html/56530db7033c0#3

²Grace, R. (2015). Barriers to the commercialisation of MEMS: the 2014 MEMS Industry Commercialization Report Card. CMM; volume 8, issue 5. Available at: https://flickread.com/edition/html/55fc2dc48505e#41

³Grace, R. and Walsh, S. (2002). International MEMS, Microsystems, Top Down Nano Roadmap; p. 53. Micro, Nano and Emerging Technologies Commercialisation Education Foundation (MANCEF).

⁴Bryzek, J. and Grace, R. (2014). Trillion Sensors Initiative. CMM; volume 7, issue 2.

⁵Grace, R. (2011). Think outside the chip: MEMS-based systems solutions. MEMS Technical Review; volume 1, issue 7.

⁶Grace, R. (2012). Technology clusters and their role in the successful commercialisation of micro and nano technologies. CMM; volume 5, issue 2.

⁷Santayana, G. (1905). Life of Reason.