NIST GCR 05-879 —Photonics Technologies:Applications in Petroleum Refining, Building Controls, Emergency Medicine, and Industrial Materials Analysis

4. CASE STUDY: MEMS-BASED INFRARED MICRO-SENSOR FOR GAS DETECTION

In 1999, the ATP funded a high-risk innovative project to develop photonic crystal sensors to be tuned for the accurate measurement of methane, carbon monoxide, and carbon dioxide concentrations and other common gases. One key outcome was the development of a sensor technology for CO₂ levels, considered to be the best initial target application.

Accurate, reliable, and low-cost detection of trace amounts of CO₂ are needed to meet safety, health, and energy conservation needs in medical and commercial markets.

• Inexpensive catalytic and electrochemical CO₂ detectors in common use only detect the presence of CO₂ without measuring CO₂ levels, are prone to false alarms, and require frequent replacement of sensing elements.
• More expensive infrared (IR) gas sensors—using non-dispersive IR absorption spectroscopy (NDIR)—can provide accurate and reliable measurement of CO₂ levels. Given their high cost, large size, and high power consumption, NDIR systems have not achieved mass-market penetration as CO₂ sensors (Kincaid, 2000).

Simple, standardized, and less expensive sensor technology is needed to lower costs and to bring accurate and reliable CO₂ sensors to mass markets (Johnson, 2004).

The ATP award built upon the results of prior National Science Foundation (NSF)-funded research. Following the successful 2001 completion of the ATP-funded project, Ion Optics, Inc., the recipient of the ATP award, obtained three rounds of venture capital financing ($8.2 million) to commercialize photonics crystals to be tuned to detect CO₂. In the future, Ion Optics may undertake additional technical initiatives to develop photonics crystals tuned to detect methane and other gases.

PROJECT HISTORY

In 1995, Ion Optics, Inc. (IOI), a startup technology company in Waltham, MA, recognized the commercial potential of using photonics crystals for the development of a more accurate, reliable, and inexpensive gas sensor technology.

Prior to receiving an ATP award to develop a photonics crystal technology, IOI obtained several NSF grants totaling $850,000 to evaluate the overall feasibility of miniaturizing NDIR gas sensors with photonic crystals. IOI matched the NSF award with $500,000 and succeeded in demonstrating the feasibility of this approach. In addition, IOI gained improved understanding of how crystal surface structures could be used for tuning crystals to wavelengths of specific gases; that is, matching IR energy emissions and absorptions to the absorption line of the target gas.

To move beyond NSF-funded research, IOI proposed to ATP a project to miniaturize the functionalities of NDIR systems onto an integrated sensor chip, using micro-electromechanical system (MEMS) fabrication technologies. One key technical challenge was to develop an IR radiation source that could both emit radiation in a narrow wavelength band and act as a detector of the reflected radiation. In 1999, the ATP agreed to fund this high-risk technology development project, and in 2001 the project was successfully completed.

At the time of ATP funding, IOI's only source of revenue was contract research, which would have been insufficient to fund the photonics crystal technology project. Accordingly, technical development would have been substantially delayed or, most likely, not completed at all (Johnson, 2004).

After receiving the ATP award and achieving initial technical progress, IOI obtained a first round of venture capital financing, with subsequent rounds bringing total venture capital financing to $8.2 million. Venture capital financing was used to substantially accelerate remaining technical development, to support commercial prototype building, and to facilitate an effective marketing program. Venture capital financing also facilitated an increasingly sharper focus on an initial target gas with the best, near-term commercial prospects, the accurate measurement of CO₂ levels.

Based on technical solutions developed with ATP support, IOI built prototype sensors to measure CO₂ levels in the expired breath of emergency medicine patients. Prototypes were also built for measuring CO₂ levels for air quality control systems of commercial office buildings.

IOI funded extensive independent market research, including bottom-up investigation through discussions with OEMs of CO₂ gas sensors, and extensive top-down market research, to verify aggregate forecasts developed with input from potential customers (Fletcher Spaght, 2001).

In its marketing efforts, IOI is currently targeting OEMs of CO₂ sensors for emergency medicine and OEMs of commercial building controls. More than 30 letters of interest have been received from potential OEM partners, and prototypes have been shipped for customer evaluation.

In 2004, IOI selected Innovative Micro Technology of Santa Barbara, CA, as its MEMS fabrication partner to bring accurate, reliable, and inexpensive sensor chips to market. Initial production runs are planned to start in 2006.

HOW DOES IT WORK?

Non-dispersive infrared absorption spectroscopy (NDIR) is currently the methodology of choice for the accurate and reliable measurement of gas concentrations. A sensor is used to monitor a specific range of the IR spectrum corresponding to the signature wavelength of the target gas. As shown in Figure 14, when the target gas passes between an IR source and the detectors, the absorption spectrum changes and instrument electronics are used to process this information and determine gas concentration (Puscasu, 2003). While accurate and reliable, NDIR sensors remain expensive and are large units with high power consumption.

Figure 14: Conventional NDIR (Cabinet Full of Discrete Components)

More simple, standardized, and less expensive integrated-circuit technology is needed to lower sensor costs and to bring more accurate and reliable gas sensors to mass markets (Johnson, 2004).

The ATP-funded photonic crystal sensor-on-a-chip (SOC) technology addresses the need for improved accuracy, reliability, and lower cost. As shown in Figure 15, SOC incorporates a micro-bolometer, a radiation-sensitive resistance element in one branch of a Wheatstone bridge. The resistance element is both a source of infrared radiation and a sensor detector of the reflected IR radiation with a spherical mirror reflecting the source signal back to the micro-bolometer element.

In the absence of a target gas to be detected, the micro-bolometer (incorporating the gas SOC) reaches radiative equilibrium with its surroundings. When the target gas is present, it absorbs some of the radiative energy and reduces the amount of light reflected back onto the filament, causing the bolometer to cool off. This temperature change is detected by monitoring filament resistance, or voltage with a constant current source.

Gas SOC is built up as a multi-layer stack. The bottom layer is bulk silicon. Photonic crystal, consisting of an array of holes etched into silicon oxide dielectric, is the intermediate layer. A thin metallic film layer, perforated with apertures approximately the size of the emissions wavelength of the targeted gas, is the top layer.

Figure 15: ATP-Funded Gas Sensor on a Chip

The IR radiation, or emission process, begins in the bulk silicon where thermal stimulation produces blackbody-like radiation. The pattern of holes in the silicon oxide (photonic crystal) layer reshapes this spectrum of radiation and centers it around a specific wavelength of resonance defined by the lattice spacing in the crystal. As the photons cannot penetrate through the thin metal film (top layer), they excite surface plasmon waves at the photonic crystal/metal interface and generate a resonant interaction of the incident radiation with the surface plasmons on both the bottom and top surfaces of the metal film. The surface plasmons decay into photons and are emitted from the top metal surface toward the spherical mirror, above the bolometer. Using photonic crystal and surface plasmon interactions, the (top) metallic surface efficiently emits and absorbs in a narrow waveband centered on the signature wavelength of the target gas.

The goal of the ATP-funded photonic crystal project was to resolve high-risk technical and fabrication process challenges associated with commercially viable SOC for gas measurement. Key ATP project objectives included:

• Developing a heated filament IR source to emit radiation in a narrow wavelength band, without the use of optical filters, demonstrating that the filament could act as its own detector (by measuring filament temperature or changes in filament resistance), and operating with a high signal-to-noise ratio
• Fabricating MEMS in standard semiconductor foundries to replace NDIR discrete component instruments "in the same way that integrated circuits replaced large electronic systems made from discrete components" (Johnson, 2004)
• Achieving thermal stability (mechanically holding the device together through thermal cycles during fabrication)
• Achieving operational stability subject to ambient temperature and optical conditions as well as environmental insults from interferent gases and operating with low power consumption and high signal-to-noise ratio
• Reaching high production volumes with a gradually declining cost structure

The ATP-funded project was successfully completed in 2001, resulting in:

• Efficient design tools for photonic crystals and frequency selective crystal surfaces Five-mask MEMS and e-beam lithographic process by the Jet Propulsion Laboratory, acting as subcontractor
• Thermal testing for IR source and for gas sensing and laboratory experiments confirming the predicted effect of interferent gases
• Cost improvement roadmap for reaching production cost targets
• U.S. patent (US 6,528,792 B2) and 24 published reports in scholarly and industry journals

With these results, the ATP-funded project resolved many technical and fabrication challenges for using photonic crystals to measure CO₂ levels accurately, reliably, and inexpensively in a variety of commercial applications.

When tuned to the signature wavelength of other gases, the ATP-funded photonics crystal technology could also have additional future use for the improved detection of other target gases of industrial and commercial interest.

BENEFIT ASSESSMENT AND MODELING

While accurate and cost-effective measurement of CO₂ levels is important in diverse commercial applications, available CO₂ sensors do not currently deliver both accuracy and low cost. Inexpensive chemical sensors, with low accuracy and reliability, comprise 85 percent of the market. Expensive NDIR sensors with high accuracy and reliability make up 15 percent of the market and their market potential is limited by capital cost, large size, and high power consumption (Frost & Sullivan, 2003).

The ATP-funded photonic crystal technology, developed by IOI to detect trace amounts of CO₂ with high accuracy, high reliability, and at low cost, represents a breakthrough technology that will overcome both performance limitations of chemical detectors and the cost, size, and power consumption limitations of NDIR sensors.

Near-term, high probability applications of the ATP-funded sensor technology are anticipated for:

• Emergency medicine: More accurate measurement of CO₂ concentrations in respired gases will reduce in-ambulance mortality (en route to the emergency room) and lead to fewer hospital days and reduced treatment costs for surviving patients.
• Internal air quality control: Measuring CO₂ levels in commercial office buildings, to facilitate better control of ventilation systems, will improve internal air quality levels and reduce energy use.

CO₂ SENSORS FOR EMERGENCY MEDICINE

Emergency treatment of trauma and other critically ill patients en route to the emergency room often includes inserting tubes in the trachea to deliver oxygen to the lungs. The insertion process entails considerable risk that the tube will accidentally be misdirected into the esophagus. If the error is not discovered in time, the patient may die.

Endotracheal Intubation

Patients suffering a complex array of injuries and acute medical emergencies, including open wounds from assault and traffic accidents, drug overdoses, strokes, and cardiac arrests, are often brought to the emergency room by ambulance or by air medical evacuation under paramedic care. En route, ambulance paramedics try to stabilize the patient's condition to improve the odds of survival and recovery. However, in trying to stabilize the patient's condition, paramedics must sometimes follow risky procedures.

An important step in stabilizing patients involves ensuring that the airway is open and that the patient is not desaturated (i.e., that blood oxygen levels do not drop below critical levels, which could lead to brain damage and other potentially fatal complications). Paramedics will often have to ventilate a patient (provide appropriate levels of oxygen through mechanical ventilation) in transit to the emergency room by inserting a tube into the trachea (endotracheal intubation (ETI)).

Successful ETI may be difficult to achieve when patients are not adequately relaxed, are combative, or if field conditions are unfavorable, such as with flashing lights, snowstorms, and nighttime (Wang et al., 2001). Under these conditions, endotracheal tubes can be accidentally placed into the esophagus.

Several studies conducted over the 1994-1998 period indicate a high frequency of failed ETIs in pre-emergency room situations, varying from 14 to 22 percent (Lee et al., 1994; Khan et al., 1996; Epstein and Ciubotaru, 1998). A more recent study of an urban emergency service system indicates that "upon arrival in emergency room, up to 25 percent of patients were found to have improperly placed endotracheal tubes" and faced a 56 percent fatality rate (Katz and Falk, 2001). Relying on the former, more general studies, we assumed that for every 100 intubations under emergency ambulance-type conditions, 14 to 22 intubations fail to achieve proper tracheal tube placement, on average resulting in 18 misintubations, and 10.1 deaths (Figure 16).

Figure 16: Rates of Emergency Ambulance Misintubation

Approaches to Detecting Failed ETI

When failed ETI involves esophageal intubation, the result is an absence of expired CO₂. Capnometry is the direct measurement of maximum expired CO₂ concentrations (end-tidal CO₂) during a respiratory cycle and is considered to be an important diagnostic procedure for detecting failed ETI (Grmec, 2002). Two general approaches are currently used for monitoring expired CO₂ levels for the detection and timely correction of failed ETIs under ambulance conditions:

• Colorimetric detectors: Chemically treated paper that changes color when exposed to CO₂ in respired gases is the typical procedure for validating ETI. To verify successful intubations, several breaths are sampled and the paramedic then attempts to match detector color to a color chart. These attempts at visual matching often occur at night under, at best, difficult conditions, leading to high
  rates of false readings. Another source of false reading is gastric acid, which can be misinterpreted as the presence of respired CO₂. Colorimetric disposable detectors are inexpensive, selling at $13 per sensor, but have poor reliability, limited shelf life, and do not provide a quantitative measurement of CO₂ levels (Fletcher Spaght, 2001).
• Infrared sensors: These provide accurate quantitative measurements of end-tidal CO₂ levels. An IR beam is passed through a respired gas sample, and CO₂ molecules absorb specific wavelengths of the IR beam energy. The light emerging from the sample is analyzed, and the ratio of CO₂-affected wavelengths to non-affected wavelengths is reported as the end-tidal CO₂. Currently available IR sensors are part of complex NDIR systems composed of discrete components connected by cables. NDIR systems are expensive, selling at $3,000 to $4,000, require considerable space, and have high power consumption (Novametrics website).

The ATP-funded sensor technology is currently being commercialized as an integrated (rather than composed of discrete components) disposable CO₂ sensor that combines the portability and low cost of colorimetric detectors and the accuracy and reliability of NDIR sensor/monitor systems. In contrast to colorimetric paper detectors, disposable IR sensors have a long shelf life.

Economic Modeling

Use of inexpensive, portable IR sensors will facilitate the verification of proper endotracheal tube placement, essentially eliminating false CO₂ readings. Deaths of misintubated patients in transit to the emergency room and the need for subsequent treatment of misintubated patients who survive will thus be prevented.

These benefits are estimated in two stages:

• Expected benefits from use of portable CO₂ sensors in lieu of colorimetric paper for intubations benefits are computed (per 100 intubations for ease of interpretation)
• A sales forecast for portable CO₂ sensors is extended by expected benefits per sensor

EXPECTED BENEFITS PER 100 INTUBATIONS

Colorimetric detectors are used by ambulance paramedics to verify correct endotracheal tube placement and represent the prevailing defender technology. Several studies conducted over the 1994-1998 period, described above, provide a conservative estimate of the failure rate of colorimetric paper under emergency conditions and thus a basis for comparison with the ATP-funded technology.

Based on the observed failure rate of 14 to 22 percent with colorimetric paper documented in these studies, an average 18 percent failure rate is assumed for the defender technology. In contrast, ATP-funded CO₂ sensors are expected to eliminate false CO₂ readings and will provide correct verification of endotracheal placement 99 percent of the time. A 1 percent failure rate is assumed for operator carelessness.

Thus, for every 100 emergency ambulance intubations, 17 of the 18 expected misintubations are assumed to be immediately detected and corrected with ATP-funded CO₂ sensors. Clinical benefits and avoided treatment costs, resulting from preventing 17 failed intubations, are estimated below.

Clinical Benefits: Drawing upon the 2001 Katz study described above, correcting 17 failed intubations will result in the prevention of 9.5 deaths per 100 intubations during ambulance transport to the emergency room (see Figure 17). The reduced mortality rate of misintubated patients is likely to be the most important benefit from ATP-funded CO₂ sensors.

Prevented in-transit mortality does not constitute lives saved, as one usually understands the phrase. Gunshot wounds, drug overdoses, strokes, and heart attacks still remain to be treated upon arrival in the emergency room and in hospital specialty departments, with uncertain clinical results.

Thus, this study does not attempt to value the benefit of preventing in-transit mortality. Nevertheless, a safer ambulance journey to the emergency room should improve the odds of the patient's recovering with appropriate clinical care.

An additional non-quantified clinical benefit of low-cost, accurate CO₂ sensors is their use in predicting patient survival from resuscitative efforts and additional therapy and in determining when cardiopulmonary resuscitation may reasonably be terminated (Levine et al., 1997). As such, CO₂ monitoring could be used to maximize patient outcomes while reducing futile and costly interventions.

Figure 17: Projected Benefits per 100 Intubations from Improved Detection of Failed ETIs, Using ATP-Funded CO₂ Sensors in Lieu of Colorimetric CO₂ Detectors

Avoided Treatment Costs: Preventing 17 failed intubations will also result in the avoidance of costly treatment (mechanical ventilation) and associated hospital stays for 7.5 patients per 100 intubations (see Figure 17). Risk of pneumonia, infection, and lung injury associated with the use of mechanical ventilation (Meade et al., 2001; Braverman, 2001; Ferrer et al., 2003) will also be avoided.

Conservatively, each patient who survives misintubation during ambulance transport is assumed to avoid two days of treatment and in-hospital days (Ferrer et al., 2003).

For base-case modeling, avoided daily costs are estimated at $2,450, consisting of $1,619 in daily treatment costs and $831 in daily hospital room charges (Cushing, 2004; Epstein and Ciubotaru, 1998). Treatment costs include nursing, respiratory therapist, and mechanical ventilation equipment costs. Hospital costs represent negotiated rates obtained by large health care plans.

For step-out scenario modeling, the avoided daily costs are estimated at $2,695, consisting of $1,719 in daily treatment costs and $976 in daily hospital room charges (American Hospital Association, 2003).

CO₂ SENSOR SALES PROJECTIONS

As was discussed earlier, IOI in 2001 sponsored an independent study to estimate the size of potential markets for disposable CO₂ sensors in emergency medicine applications (Fletcher Spaght, 2001).

The study used a bottom-up approach, relying on extensive direct discussions with nine OEMs of respiratory devices (which collectively represent more than 30 percent of the U.S. market for respiratory devices), emergency room staff at five hospitals, and 10 industry associations and institutions, including the American Ambulance Association, the American Association of Respiratory Care, and the American Hospital Association.

For the near term, extending to 2008, the study estimated the total U.S. market for disposable CO₂ sensors in emergency medicine at 38 million units for 2006, 43 million units for 2007, and 46 million units for 2008 (Table 5, column 1).

Based on direct fact-finding discussions with OEMs, the study estimated accessible levels of sales at 360,000 units in 2006, 4.2 million units in 2007, and 13.3 million units in 2008 (Table 5, column 2), which was considerably less than the total market. Total and accessible unit sales were extended beyond 2008 using a 6 percent long-term growth rate (Fletcher Spaght, 2001).

Going beyond the Fletcher Spaght market research, our case study developed its own bottom-up sales projections to estimate public benefits from ATP-funded CO₂ sensors following discussions with OEMs and industry associations during the early part of 2004.

The resulting unit sales projections used in our analysis (Table 5, column 3) are significantly lower than the accessible market levels Fletcher Spaght reported in its study.

In our study, base-case projected unit sales start at 10,000 units in 2006, compared to 360,000 units in the accessible market for the same year, and top out at 400,000 units per year, compared to an accessible market of 20 million units.

The probability of achieving unit sales projections (Table 5, column 3) is estimated to be 65 percent, reflecting our conservative assessment of marketplace risks facing a new technology company without current sales but very active in marketing a high performance product.

For a more optimistic step-out scenario, sales levels for each year (over the 2006-2015 period) are increased by 10 percent over base-case levels.

A separate top-down analysis is consistent with Table 5, column 3 sales projections. The primary U.S. target market for portable CO₂ sensors is the 24,000 vehicle ground ambulance fleet. A 2005 market estimate of 32 million capnometry procedures per year corresponds to less than four procedures per ambulance vehicle per day (American Ambulance Association, 2002).

Furthermore, an increased use of CO₂ readings in emergency room ambulance situations is anticipated. With growing concerns about malpractice liability lawsuits and adverse court decisions, the American College of Emergency Physicians recently recommended the significantly expanded utilization of capnometry to monitor intubated patients on their way to hospital emergency rooms, both to enhance patient safety and to reduce economic risk to the hospital. Over time, the expected increase in the frequency of ambulance CO₂ readings further suggests that total and accessible emergency capnometry markets are potentially vast.

IOI's ability to secure second and third rounds ($6.5 million) of venture capital financing is additional evidence for the credibility of conservative sales estimates used in this study.

DEFENDER TECHNOLOGY LOSSES

Portable CO₂ sensors that incorporate ATP-funded technology will displace colorimetric detectors. One CO₂ sensor will substitute for two colorimetric procedures given that approximately two colorimetric procedures are required for an acceptable reading while one CO₂ sensor reading can validate proper intubation.

As the cost of a colorimetric reading is about the same as a reading with a CO₂ sensor, defender technology losses can be estimated as two colorimetric sales minus one CO₂ sensor sale, or a net loss of one colorimetric sale. Firms suffering this loss of business to ATP-funded CO₂ sensors will forgo profit contributions from the sale of colorimetric units. This loss is estimated by extending lost sales revenues of $13 per unit by an average U.S. industry profit margin of 8.3 percent, or $1.08 per unit (Business Week, 2004).

Benefit Estimates

Mortality reductions during ambulance transport are a key benefit even if not quantified in dollar terms. Table 6, column 2 indicates mortality reductions during ambulance transport corresponding to projected sales of CO₂ sensors. On this basis, over the 2006-2015 period, an estimated 112,000 prevented in-ambulance deaths can be associated with the ATP investment.

Beyond prevented mortalities, cash flow benefits are computed by extending U.S. sales of CO₂ sensors by avoided treatment costs per misintubation. Cash flow benefits are computed for a period of 10 years, the estimated relevant lifespan of the ATP funded CO₂ measurement technology (per IOI).

By way of example, consider the following base-case calculation for 2007: The projected sale of 40,000 CO₂ sensors (from Table 6) corresponds to 40,000 intubations. Per Figure 17, each intubation is associated with average economic benefits of $367.50. Multiplying $367.50 by 40,000 intubations results in total estimated economic benefits of $14,700,000. Assuming a 65 percent probability of realizing sales projections, the expected value of economic benefits is $9,555,000. Subtracting defender technology losses of $28,054, corresponding to the lost sale of colorimetric detectors as displaced by more accurate CO₂ sensors, yields net benefits of $9,526,946. Reflecting the relative importance and similar size and timing of both ATP and NSF investments, 50 percent of net benefit cash flows are attributed to the NSF for its research investment and 50 percent or $4,763,473 is attributed to the ATP (for 2005).

Base-case and step-out scenario cash flows covering the 2006-2015 period are displayed in Table 8, column 1, and Table 9, column 1, in the Benefit-Cost Analysis section of this chapter, where cash flows for the different photonic crystal applications are shown side by side.

CO₂ SENSORS FOR OFFICE SPACE INTERNAL AIR QUALITY

High building occupancy increases levels of CO₂, which results in stale air and insufficient humidity. Ventilation, where fresh air is exchanged for polluted, stale air, is the standard approach for improving internal air quality. However, ventilation systems in commercial buildings are often wasteful of energy.

Sensors with ATP-funded CO₂ sensor chips, in combination with variable air volume (VAV) systems, can optimize heating, cooling, and ventilation and contribute to improved levels of air quality in commercial buildings at lower energy costs.

According to a U.S. Department of Energy (DOE) survey, there are approximately 50 billion square feet of enclosed and conditioned (heated and/or cooled) commercial space in the United States (U.S. Department of Energy, 1999) with approximately 1.8 billion square feet added annually through new construction (Fletcher Spaght, 2001). This includes 10.2 billion square feet of office space.

Internal Air Quality

Poor internal air quality in commercial office space affects occupant comfort, productivity, and health. The chief sources of poor air quality include CO₂ production from people breathing and combustion contaminants (furnaces and space heaters). Other sources include biological contaminants (from wet or damp materials, filters, and insects and rodents), volatile organic compounds (from paints, waxes, cleansers, sealants, copy machines, and printers), formaldehyde (from furniture, carpeting, and fabrics), disinfectants, rodenticides, printing, and paper handling.

Component factors of poor air quality, individually or in combination, can cause discomfort from lack of air movement and from low humidity, including eye, nose, and throat irritations. Acute and potentially chronic conditions can also develop, including respiratory problems resulting from a buildup of chemical and biological pollutants.

Loss of worker productivity, absenteeism, unhappy tenants, and the threat of litigation can result from poor air quality and can have a negative impact on building owners' return on investment as well as the occupants' health. EPA suggests that average office worker productivity losses from poor indoor air quality can range from 2 to 4 percent, while noting the need for additional studies (U.S. Environmental Protection Agency, website).

Approaches to Building Ventilation and Energy Efficiency

During times of high building occupancy, increased levels of CO₂ can result in stale air and insufficient humidity. Ventilation, or the exchange of polluted, stale air with fresh air from the outside, is considered one of the more effective approaches for adequately controlling CO₂ levels and for providing acceptable overall levels of internal air quality (U.S. Environmental Protection Agency, 1990; Persily and Gorfain, 2004).

Generally, providing air movement through ventilation (using electric motors to run large fans to bring in fresh outside air and to remove stale inside air) is very energy intensive. When ventilation is operated at high constant levels (running large electric fans when air exchange is not needed), unnecessary energy costs will be incurred.

Over-ventilation can also lead to potential buildup of undesirable moisture levels, resulting in the formation of molds and other biological contaminants.

The energy efficiency of ventilation systems is, in large measure, constrained by installed HVAC (heating, ventilation, and air conditioning) system design as well as the system's ability to accurately measure CO₂ levels. In this regard, HVAC systems fall into two broad categories: constant air volume and variable air volume systems.

• Constant air volume systems: Many commercial buildings have HVAC systems that operate on the basis of predetermined design occupancy levels for daytime, nighttime, and weekends and do not have the ability to vary air volumes in response to changing levels of occupancy and CO₂.
• Variable air volume (VAV) systems: These systems are designed to vary building ventilation rates by tracking detected CO₂ levels. According to a DOE survey, 23 percent, or 2.3 billion square feet, of U.S. commercial office space is equipped with VAV systems that are able to follow CO₂ load and adjust ventilation rates as appropriate for changing levels of occupancy (U.S. Department of Energy, 1999).

To reach and maintain desirable internal air quality levels, without wasteful use of energy, buildings need to have accurate CO₂ sensors as well as VAV systems that can vary ventilation rates in response to detected CO₂ levels.

Commercially available CO₂ sensors fall into two broad categories: chemical detectors, which tend to be small and inexpensive but also inaccurate and unreliable, and discrete IR sensors, which are accurate and reliable but expensive and large. Discrete IR sensors are composed of many separate components including monitors, calibration kits, digital displays, and duct sampling kits, with systems priced at many thousands of dollars per building.

In contrast, the ATP-funded sensor chip, when integrated with an OEM's building control systems, will have comparable accuracy to discrete IR sensors but will be sold at a substantially lower price point, expected to be about $25 per sensor, or $250 for a typical building control system (Fletcher Spaght, 2001).

The EPA estimates that accurate and cost-effective CO₂ sensors, in combination with VAV ventilation systems, can reduce commercial office space energy consumption by 10-20 percent through avoided heating, cooling, and fan operations (U.S. Environmental Protection Agency, 1990; Emmerlich and Persily, 2001):

• It is costly to heat cold winter air and to cool hot summer air. Unnecessary air exchange from bringing in excess outside fresh air will increase the volume of air that must be heated and cooled at additional expense.
• Electric power is used to operate ventilation fans, and unnecessary ventilation will result in wasteful electricity consumption.

In combination with VAV systems, sensors with ATP-funded CO₂ sensor chips will contribute to reaching improved levels of internal air quality in commercial buildings, in addition to minimizing unnecessary heating, cooling, and ventilation costs. Better indoor air quality leads to:

• Increased worker productivity—U.S. EPA estimates performance losses from poor indoor air quality at 2 to 4 percent for all buildings (U.S. Environmental Protection Agency, website)
• Reduced incidence of respiratory diseases as well as diseases from other pathogens that can be diluted and removed through properly modulated ventilation (Aries, see website)

Economic Modeling

Energy savings are estimated and projected as quantitative cash flow benefits (Figure 18). Owing to the substantially greater complexity and uncertainty of productivity and health benefits, these benefits are expressed in non-monetary terms at this time.

To arrive at energy savings cash flow estimates, the following assumptions are used for each CO₂ sensor and the projected sales of CO₂ sensors.

EXPECTED BENEFITS PER CO₂ SENSOR

Based on electric and gas utility experience (Madison Gas & Electric, 2005), average energy consumption of commercial office space is estimated at 0.520 therms (520,000 BTUs) per square foot per year for space heating alone and 3.14 Kwh per square foot per year for cooling and ventilation, but excluding interior lighting and office equipment power consumption.

Consistent with IOI's sponsored independent market study (Fletcher Spaght, 2001), it is projected that one CO₂ sensor will be installed per each 1,500 square feet of office space. Extending baseline average energy consumption per square foot by 1,500 square feet, corresponding to one sensor, results in energy savings of 780 therms, or 4,710 Kwhs.

VAV ventilation systems in combination with CO₂ sensors can be expected to reduce HVAC costs by 10-20 percent according to EPA estimates (U.S. Environmental Protection Agency, website).

The base-case analysis assumes annual energy savings of 14 percent of baseline energy use, just below the midpoint of the EPA-estimated range (i.e., 109.2 therms and 659.5 Kwh per sensor). Extended by 2005 retail utility prices (76 cents per therm and 6 cents per Kwh (Madison Gas & Electric, 2005)), annual energy cost savings from ATP-funded CO₂ sensors are estimated at $122.60 per sensor.

Figure 18: Projected Annual Energy Savings from ATP-Funded CO₂ Sensors in Commercial Office Space Building Controls

The step-out scenario assumes annual energy savings of 16 percent of baseline energy use, just above the midpoint of the EPA-estimated range (i.e., 124.8 therms, or 753.6 Kwh per sensor). Extended by 2005 retail utility prices (Madison Gas & Electric, 2005), annual energy cost savings are estimated at $140 per sensor.

CO₂ SENSOR SALES PROJECTIONS

In 2001 IOI sponsored an independent study to estimate the potential U.S. market for CO₂ sensors in commercial buildings (Fletcher Spaght, 2001). Market research used a bottom-up approach, relying on direct discussions with 13 OEMs of building control systems and several industry associations for commercial buildings and building controls.

For the 2006-2007 period, the market study estimated the total U.S. market for CO₂ detectors at 5.2 million units per year. Four million units would be sold each year to retrofit existing commercial buildings and an additional 1.2 million units would be sold for new construction. Going forward, the size of the total market was assumed to remain flat at 5.2 million units (Table 7, column 1).

Based on fact-finding discussions with OEMs, the Fletcher Spaght study estimated accessible levels of sales at 600,000 units in 2006, increasing to 1 million units by 2010 and 1.2 million units by 2015 (Table 7, column 2).

Going beyond the Fletcher Spaght market study, our case study developed its own sales projections to estimate public benefits from ATP-funded CO₂ sensors through discussions with OEMs and industry associations during the early part of 2004. The resulting sales projections, indicated in Table 7, column 3, are significantly lower than accessible market levels, reflecting our conservatism concerning financial, manufacturing, and marketing limitations of new technology companies.

Our base-case sales projections start at 1,000 units in 2006, compared to 600,000 units in the accessible market for the same year, and top out at 292,000 units per year, compared to an accessible market of 1.2 million units.

The probability of achieving base-case sales projections shown in Table 7, column 3, is deemed to be 60 percent, reflecting our conservative assessment of marketplace risks facing a new technology company without current sales but very active in marketing a high performance product. IOI's ability to secure $6.5 million in two consecutive rounds of venture capital funding on the basis of their higher sales projection is additional evidence of the credibility of projected base-case unit sales used in this study.

For a more optimistic step-out scenario, sales levels for each year (over the 2006-2014 period) are increased by 10 percent over base-case levels.

Benefit Estimates

Total cash flow benefits each year are computed by extending U.S. sales of CO₂ sensors by economic benefits from energy savings.

Consider the following base-case calculation for 2007: The projected sale of 48,000 CO₂ sensors (from Table 7) is associated with average economic benefits of $122.60. Multiplying $122.60 by 48,000 unit sales results in total estimated economic benefits of $5,884,800. Assuming a 60 percent probability of realizing sales projections, the expected value of economic benefits is $3,530,880. Reflecting the relative importance and similar size of both ATP and NSF investments, 50 percent of net benefit cash flows are attributed to the NSF for its prior research investment and 50 percent or $1,765,440 is attributed to the ATP (for 2007).

Base-case and step-out scenario cash flows, covering the 2006-2015 period are displayed in Table 8, column 2, and Table 9, column 2, in the Benefit-Cost Analysis section of this chapter, where cash flows for two photonic crystal applications are shown side by side.

BENEFIT-COST ANALYSIS

Prevented deaths in transit to the emergency room represent a significant societal and human benefit from the commercial use of the ATP-funded CO₂ sensor technology.
In addition, the technology will also yield significant economic benefits (Figure 19) in the form of:

• Treatment cost savings in emergency medical departments
• Reduced energy costs in commercial office buildings

This section presents performance metrics that capture the economic benefits from avoided hospital days, treatment costs, and energy savings as compared to ATP's investment in the high-risk photonic crystal sensor technology.

Figure 19: Flow of Benefits from ATP-Funded CO₂ Sensor Technology

ATP, NSF, AND INDUSTRIAL PARTNER INVESTMENTS

During the 1999-2001 period, ATP invested $753,000 toward project direct costs and industry partner, IOI, invested $626,000 in the development of photonic crystal CO₂ sensor technology. Several phases of prior and concurrent NSF grants ($850,000) supported prior research and were matched with $500,000 from IOI.

For purposes of cash flow analysis and computation of performance metrics, the ATP investment was normalized to 2005 dollars (as $827,000) and included as one lump sum investment in 2000, the midpoint of the ATP investment period.

PERFORMANCE METRICS

Our quantitative analysis of benefits was limited to public benefits that could be meaningfully quantified and excluded benefits and costs to IOI. Benefits attributable to ATP were determined to be 50 percent of total public benefits identified because NSF and ATP funding were approximately the same magnitude and both were deemed essential to realizing these benefits. We estimated benefit cash flows for a conservative base case and for a more optimistic step-out scenario. We compared these benefits to the cash flow representing ATP investment costs.

This comparison resulted in three sets of economic performance measures: net present value, benefit-to-cost ratio, and internal rate of return.

BASE-CASE ANALYSIS

As indicated in Table 8, the public return on ATP's investment in the CO₂ sensor technology over the 1999-2015 period can be expressed as a net present value of $143 million. Public benefits attributable to ATP are $174 for every dollar invested and the internal rate of return is estimated at 75 percent.

STEP-OUT SCENARIO ANALYSIS

A step-out scenario analysis investigated the sensitivity of base-case performance metrics to a more optimistic assumption about the projected future benefits of each

CO₂ sensor in emergency medicine and office building air quality applications, in combination with a 10 percent increase in projected sales.

As indicated in Table 9, the public return on ATP's investment in the CO₂ sensor technology is associated with a net present value of $175 million. Public benefits attributable to ATP are $212 for every dollar invested and the internal rate of return is estimated at 79 percent.

PRIVATE BENEFITS TO ATP INDUSTRY PARTNERS

ATP's industry partner's continued motivation to refine and commercially market the ATP-funded technology is a necessary pre-condition for commercial- and industrial-scale impact. Only with continued investment will the general public come to enjoy the health care and economic benefits expected to result from ATP's investment in the photonic crystals CO₂ sensor technology.

IOI's commercial progress to date and its continued motivation to actively market and improve CO₂ sensors is indicated by its progress toward commercialization. It has closed three rounds of venture capital financing of $8.2 million subsequent to receiving ATP funding. It has received more than 30 letters of interest, spelling out detailed sensor specifications from OEMs of medical respiration equipment and OEMs of building controls systems.

IOI has shipped customer evaluation units to these and other OEMs, has put in place a MEMS fabrication agreement with a leading MEMS foundry, and plans initial production runs for 2006.

SUMMARY

Performance metrics presented above point to exceptional performance from ATP's investment in the development of photonics crystals for CO₂ detection, as demonstrated by quantified medical benefits, high public rates of return, and important qualitative benefits.

Return to Table of Contents or go to next section of report.

Date created: July 12, 2006
Last updated: September 14, 2006

ATP website comments: webmaster-atp@nist.gov  / Technical ATP inquiries: InfoCoord.ATP@nist.gov.

NIST is an agency of the U.S. Commerce Department
Privacy policy / Security Notice / Accessibility Statement / Disclaimer / Freedom of Information Act (FOIA) /
No Fear Act Policy / NIST Information Quallity Standards / ExpectMore.gov (performance of federal programs)