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

3. CASE STUDY: CAPILLIARY OPTICS FOR X-RAY FOCUSING AND COLLIMATING

Accurate, timely analytical information about the chemical composition and crystalline structure of engineering materials is increasingly important for their manufacturability and commercial use. Near-real-time analytical information is also essential for the cost-effective detection of increasingly low concentrations of contaminants in process industries.

In 1992, ATP funded a high-risk technology project to develop capillary optical lenses for hard-to-focus X-rays and neutrons. The successful completion of this project, involving only the focusing and collimating of X-rays, enhanced the analytical performance of laboratory instruments and process control metrology, leading to more effective engineering design practices and improved industrial process controls.

PROJECT HISTORY

More than 50,000 engineering materials are in industrial use, and advanced materials are continually being developed for new product applications. Accurate information about the chemical composition and crystalline structures of engineering materials is critical to their manufacturability and technical performance.

Accurate information about composition and structure can be obtained through X-ray spectroscopy, which measures the absorbed or emitted electromagnetic radiation from material samples. X-ray wavelengths are similar to distances at the atomic scale, leading to ultra-high levels of measurement accuracy.

X-ray diffraction (XRD) and X-ray fluorescence (XRF) are important spectroscopic methods for investigating materials composition and structure. XRD is used for structural analysis, as in degrees of crystallinity, crystal orientation, and material stress. XRD analysis uses collimated or parallel X-rays. XRF is used for chemical composition and thin-film thickness analysis and the detection of trace contaminants. Incident X-rays for XRF analysis are focused.

Prior to the ATP-funded project, high energy X-rays could not be effectively focused or collimated, which limited the measurement accuracy and speed of XRD and XRF instruments.

The physical principles of special optics for guiding X-rays were discovered in the 1960s at the I.V. Kurchatov Institute of Atomic Energy in Moscow, Russia. Despite an understanding of underlying principles, significant technological uncertainties remained before X-ray optics could be effectively used to increase XRD and XRF measurement accuracy and speed in industrial materials analysis and process control.

In 1990, X-Ray Optical Systems, Inc. (XOS), of East Greenbush, NY, was formed to develop commercially viable X-ray optical lenses based on the physical principles discovered at the Kurchatov Institute. Mr. David Gibson, president of XOS, had professional contacts at the Kurchatov Institute and at the Center for X-Ray Optics at the University at Albany, State University of New York, through his father, Dr. Walter Gibson, who was Director of the Center. Dr. Gibson and staff from the Kurchatov Institute served as technical consultants to XOS.

XOS at the time lacked internal financial resources and access to the venture capital community. In 1991, XOS approached the ATP with a proposal to develop methods for the design, fabrication, and testing of X-ray optics. In early 1992, ATP agreed to fund the proposed project, and XOS proceeded with the project in concert with the following collaborators:

• Center for X-Ray Optics at the University at Albany, State University of New York, which conducted radiation damage studies and assessed the impact of surface roughness and waviness on X-ray optical lens transmission efficiencies
• Oak Ridge National Laboratory, which made its micro-focus X-ray fluorescence test site available for testing XOS focusing optics
• INCOM, which assisted in developing part of the early optic manufacturing process.

The ATP-funded project was successfully completed in 1996. XOS began selling optical lenses for collimating and for focusing X-rays on a commercial basis in 1997. The marketing and sales of integrated sensor engines, combining X-ray sources, X-ray optics, detectors, and software, started in 2004.

According to XOS, the ATP provided essential seed capital for the technical development of the high-risk, innovative X-ray optics technology. "Without the ATP, X-ray optics and related sensor engine products could not have been developed, and the resulting benefits to our company and its commercial and industrial customers would not have been realized" (Gibson, 2004).

HOW DOES IT WORK?

Metrology systems using X-rays consist of a point source to generate divergent rays over a large angle (Figure 2). The source is enclosed in shielding with apertures allowing the X-rays to pass through and irradiate a small spot on a material specimen.

While beams of visible light can be collimated and focused with optical lenses, X-rays are a penetrating form of electromagnetic radiation and thus cannot be similarly controlled. This property has been an important source of technical limitation for X-ray spectroscopy.

Without X-ray optics, some degree of beam collimation or focusing can still be achieved by varying distances from the X-ray source to the aperture and the sample. However, as beam intensity or the number of X-ray photons that hit their target varies inversely with the square of the distance from the source, an ever decreasing fraction of photons reaches the sample and remains "useful" as distances are increased.

Figure 2:   X-Ray Point Source without Optics

Though many scientists believed that the laws of physics would prevent the more precise collimating and focusing of high energy X-rays, in the late 1980s the I.V. Kurchatov Institute in Moscow developed a laboratory technique for capturing and channeling X-rays utilizing capillary optics, or arrays of thin-diameter glass tubes with a slight curvature along the longitudinal axis.

Multiple reflections from the smooth inner walls of capillary tubes can be used to guide X-ray photons. Reflections occur at the boundary between media with different refractive indices (air and glass surfaces) when an X-ray strikes the reflecting glass at a grazing angle smaller than an empirically determined critical angle (α).

If the longitudinal curvature of capillary arrays is kept below a certain limit, many divergent X-rays can be reflected off the inside surface of glass tubes at an angle less than the critical angle (α) and effectively bent and transported along capillary channels (Figure 3). When certain X-rays have an angle of incidence (β) greater than the critical angle, these X-rays will be refracted through the surface of the glass tubes.

Figure 3:  (Left) Internal X-Ray Reflection from Glass Surfaces at Less than Critical Angle and
Figure 3:  (Right) Incident X-Ray at Greater than Critical Angle

Capillary arrays consist of thousands of tiny, slightly curved glass tubes held together in a frame. Individual glass tubes have varying cross sections over the length of the tube. By arranging the geometry of thousands of capillary channels, X-rays can be collimated. With different geometrical arrangements, X-rays can be focused (Figure 4).

Laboratory research indicated that capillary arrays could potentially:

• Capture large angle divergent X-rays from a point source
• Form a convergent beam with orders of magnitude higher intensity
• Form a quasi-parallel beam with low divergence
• Suppress transmission of X-rays at energies above a selected cut-off
• Bend X-ray beams
• Change beam cross sections

If these potential capabilities could be achieved in commercial optical products, the existing design constraints of X-ray spectroscopic instruments could be relaxed and their analytical performance could be substantially improved; that is, new capabilities could facilitate higher throughput, better resolution and sensitivity, increased reliability, and potentially lower cost.

As a practical matter, however, it was expected that capillary optical devices would be very difficult to construct within requisite geometric tolerances and would remain laboratory curiosities, unless significant technical uncertainties could be resolved.

The important remaining technical uncertainties included design procedures, manufacturing control issues, and the long-term reliability of capillary arrays in the presence of ionizing radiation.

Figure 4:    Capillary Optics for Focusing and Collimating X-Rays

Source: X-Ray Optical Systems, Inc.

To address these issues, XOS proposed a high-risk technology development program to the ATP and obtained ATP funding. The key technology development goals were:

• Modeling: Establish engineering design modeling and computational capabilities for effectively working through a large number of computationally challenging, alternative design options, including lens geometries, material properties, and capillary surface roughness, and surface waviness.
• Alternative lens materials: Systematically evaluate alternative capillary materials, including borosilicate glass, lead oxide silicate glass, boron oxide glass, and beryllium oxide glass, relative to critical angles for X-ray reflection, transmission efficiencies, glass purity, and use experience with capillary optics.
• Fabrication: Develop, test, and refine fabrication methods for drawing high purity, ultra-thin, hollow glass tubes with variable cross sections. Develop assembly and alignment technologies for assembling hundreds of thousands of glass tubes into closely packed capillary arrays, subject to exacting quality standards.
• Reliability: Develop an understanding of radiation-related reliability issues; that is, determine the radiation flux limits of capillary optical elements and understand radiation-related processes, such as ionization and beam heating, and their effect on X-ray transmission. Explore methods for controlling or minimizing radiation damage and heating effects to enhance the reliability of capillary optics.

The ATP-funded technology project was successfully completed in 1996 and resulted in the following technical accomplishments:

• Modeling: XOS developed a proprietary simulation model to evaluate a wide range of design variables including capillary surface roughness, surface waviness, tapered capillaries, and fiber bending. Models can simulate the performance of complete capillary optical systems and generate detailed fabrication drawings. Actual capillary performance was observed to be closely correlated with model simulation results.
• Alternative lens materials: XOS evaluated alternative optical-quality glasses to optimize capillary performance and developed a relationship with U.S. suppliers that could provide 100 percent of capillary materials.
• Fabrication: XOS successfully developed glass pulling, multi-capillary fiber construction, array assembly, array alignment, testing, and quality control methodologies to produce high purity and high-efficiency arrays. The smallest of these capillary arrays consists of more than 300,000 channels. "Each capillary is so tiny that over 25 would fit in a single human hair. If the capillary channels of the optic were laid end to end, the length would exceed 10 miles" (X-Ray Optical Systems, Inc., website). (See Figure 5.)

Figure 5:   Glass Capillary Optics

• Reliability: Radiation damage experiments were conducted on capillary systems at the National Synchrotron Light Source at Brookhaven National Laboratory. Discoloration and some physical damage were observed, but no loss in lens performance could be determined beyond statistical variations.

These technology advances made it possible for XOS to develop commercially viable lenses for collimating X-rays into tight parallel patterns and for focusing X-rays onto micro-sized spots.

When ATP-funded X-ray optical lenses are incorporated into analytical instruments for materials analysis and into sensor engines for industrial process control, more photons are delivered to specimens under investigation, beam intensity is improved several orders of magnitude, and measurement sensitivity is enhanced (Beumer, 2004). As a result, laboratory materials analysis that would take 24 hours can be completed in just 1.5 to 2 hours, without loss of measurement sensitivity, and in-line industrial process metrology can be used to detect trace-level contaminants, on a near-real-time basis (Gibson, 2004). Subsequent product development efforts resulted in further improvements in measurement sensitivity.

In industrial laboratory applications, standalone X-ray optics are used as performance-enhancing laboratory instrument components for more accurate and rapid characterization of material composition and crystalline structure.

In industrial process control applications, X-ray optics are incorporated into small-footprint, low-power consumption sensor engines along with an X-ray laser source, detectors, hardware, and software.

BENEFIT ASSESSMENT AND MODELING

U.S. industrial laboratories need increasingly accurate analytical information about the chemical composition and crystalline structure of advanced materials to support the development and ensure the manufacturability of new industrial products. Analytical information must be obtained in a timely and efficient manner.

U.S. process industries need increasingly accurate, near-real-time analytical information about contaminant concentrations to ensure sufficient quality controls and improved process yields.

The ATP-funded X-ray optics technology directly addresses industry needs for increasingly accurate and timely analytical information in three application areas:

• Laboratory materials analysis in a broad range of industries
• Process control metrology in petroleum refining and distribution
• Process control metrology in semiconductor fabrication

LABORATORY MATERIALS ANALYSIS

Laboratory instruments for materials analysis tend to be large-footprint, complex systems offering full analytical range. In the case of elemental composition analysis, full analytical range would include elements from boron to uranium in solids and powders and from sodium to uranium in liquid samples (Uhling et al., 1999).

X-ray diffraction (XRD) and X-ray fluorescence (XRF) are high sensitivity methods for materials analysis:

• XRD systems use highly collimated X-ray beams to study crystalline structure and stress, critical for developing advanced materials for new, high performance industrial applications and for supporting engineering design solutions when utilizing advanced materials.
• XRF systems use highly focused X-ray beams for elemental analysis to detect trace amounts of contaminants and pollutants and to measure thin-film thickness at nano-scales.

Economic Modeling

To quantify cash flow benefits to industry users that can be associated with performance-enhancing X-ray optics for laboratory materials analysis, estimates are developed for average benefits per X-ray optical lens and for retrospective and projected future sales of X-ray optical lenses.

BENEFITS PER X-RAY OPTICAL LENS

For XRD analysis, commercial laboratories tend to charge about $144 per hour as a fully loaded user fee, reflecting such items as equipment depreciation costs, operator salaries, utility costs, equipment maintenance contracts, laboratory consumables, sample preparation, and telecommunications (Morgan, 2001; Pierce, 2004).

For XRF analysis, a fully loaded commercial user fee tends to be around $52.50 per hour, including equipment depreciation costs, operator salaries, and utility costs (Morgan, 2001; Pierce, 2004; Homeny, 2004).

Assuming XOS sales patterns to be 30 percent sales of XRD systems and 70 percent sales of XRF systems, weighted average user fees are computed as $79.95 per hour (Figure 6).

Based on discussions with commercial laboratory operators, it is estimated that 40 percent of the weighted average fees cover such items as equipment depreciation and sample preparation expenses and that 60 percent, or $47.80 per hour, represents the average variable operating cost for the direct use of XRD and XRF analytical instruments.

While multiple orders of magnitude (tenfold to hundredfold) increases in analytical speed were reported from incorporating performance-enhancing X-ray optics, labor, utility, and other variable costs are structurally constrained and only partially variable in commercial operations. In other words, laboratories have fixed salary obligations, and, given less than fully elastic demand for analytical services, laboratories may not always reap economic advantage from increased equipment availability resulting from higher analytical speeds.

To reflect these constraints, we conservatively assume in the base case that tenfold and higher increases in speed will result in only a 25 percent decrease in operating costs. Accordingly, average savings associated with increased analytical speed—at 25 percent of the $47.80 per hour variable costs—yield estimated savings of $11.95 per hour from the use of ATP-funded performance-enhancing X-ray optics.

We further assume that a typical laboratory operates for eight hours per day for 240 days per year. Extended by savings of $11.95 per hour, base-case annual savings associated with each X-ray optic in operation is estimated at $22,944.

Figure 6:   Projected Annual Savings in Laboratory Operating Costs From Use of Standalone X-Ray Optics (2004 Dollars)

For the step-out scenario, hourly savings are increased 10 percent to $13.14, resulting in estimated annual savings of $26,490.

Industrial laboratory analysis for new product development, using advanced materials, is often constrained by laboratory equipment availability and cost. Substantially increased analytical speeds at reduced cost can therefore be expected to facilitate new product development with advanced materials.

Though this benefit was not quantified at this time, increased analytical speeds and a more efficient development of advanced materials will provide significant competitive advantage to U.S. companies in a broad spectrum of industries—for example, pharmaceuticals, semiconductors, and other industries using advanced materials.

X-RAY OPTICAL LENS SALES

The analytical instruments industry is highly competitive and technologically advanced. The industry's major segments are laboratory instruments, industrial measuring and controlling instruments, and electrical testing and measurement instruments. The United States is the largest producer in each of the three segments and is a net exporter of analytical instruments for industrial process control and for materials research.

The U.S. laboratory instruments segment is estimated to reach $17.9 billion in sales in 2004 and experience average growth rates of 5 percent. Certain sub-segments, including materials analysis for elemental composition and crystalline structure, are expected to grow at annual rates of 8 percent (Frost & Sullivan, 2003; Business Communication Company, 2004).

XOS has actively marketed X-ray collimating and X-ray focusing optics to original equipment manufacturers (OEMs) of analytical instruments since 1997 and the above industry trends provide a context for XOS's ongoing marketing efforts.

Estimated U.S. annual sales over the 1997-2003 period range from 7 optical lenses to 88 lenses. Going forward, projected sales of standalone X-ray optics gradually increase from 95 units in 2004 to 100 units in 2014 (Table 1).

Projected sales levels represent a highly conservative base case in light of expected industry segment growth rates in excess of 5 percent. For a more optimistic step-out scenario, sales levels for each year (over the 2004-2014 period) are increased by 10 percent over base-case levels.

Given seven years of successful commercial experience and only modestly increasing projections of future sales, the probability of realizing future sales projections is estimated to be 90 percent.

Benefit Estimates

Based on discussions with XOS and industry experts, we assume that optical lenses will continue to generate benefits for a period of 10 years after the sale of each lens. Annual cash flow benefit estimates are computed by extending estimated economic benefits per X-ray optical lens by the number of optical lenses projected to be deployed in the United States.

Cash flow time series covering the 1997-2014 period are displayed in column 2 of Tables 3 and 4 in the Benefit-Cost Analysis section of this chapter, where cash flows from different X-ray optics applications are shown side by side.

To illustrate how cash flow benefit estimates in Table 3 are arrived at, consider base-case calculations for fiscal year 2005. As shown in Table 1, 491 lenses are sold as performance-enhancing components for industrial laboratory instruments during the 1997-2005 period. Given expected utilization for 10 years, each of the 491 lenses is presumed to be in continued use and generating $22,944 benefits per year (Table 1 and Figure 6). Expected benefits for 2005 are determined by multiplying 491 lenses by $22,944, resulting in $11,266,000 of total economic benefits. Assuming a 90 percent probability of realizing sales projections and benefit estimates, the expected value of economic benefit cash flows (for 2005) is $10,139,000.

PROCESS CONTROL IN PETROLEUM REFINING AND DISTRIBUTION

In 2001, the U.S. Environmental Protection Agency (EPA) adopted rules requiring drastic reductions in the sulfur content of highway-grade diesel fuel to ultra-low levels, from 500 to 15 ppm (parts per million). While the regulation is to be fully implemented over the 2006-2010 period, most diesel fuel must meet the 15 ppm standard by 2008.

The petroleum refining and distribution industry will be significantly challenged to achieve 15 ppm ultra-low sulfur diesel (ULSD) levels in the 2006-2008 timeframe. Keeping sulfur levels to trace amounts will require innovative new technologies to provide orders of magnitude more accurate and near-real-time metrology (National Petroleum Council, 2000).

Incorporated into small footprint, low power consumption sensor engines, XOS optics will be deployed in petroleum refining and distribution systems to detect EPA-mandated trace amounts of sulfur, on a near-real-time basis. The use of sensor engines will facilitate regulatory compliance, improved quality control, increased product yield, and reduced highway diesel emissions.

Petroleum Refining and Distribution

The U.S. petroleum refining industry operates 16.7 million barrels per day crude distillation capacity (Energy Information Administration, 2001) to produce motor gasoline, diesel fuel, and other petroleum products. There are 146 petroleum refineries currently operating in 32 states across the United States. Of these, 80 percent, or 117 refineries, process crude oil into diesel fuel (National Petroleum Council, 2000).

Diesel fuel and other petroleum products are distributed through a complex system of storage terminals, pipelines, barges, and tanker trucks to retail outlets and end-users. The U.S. distribution system includes approximately 1,300 storage tanks and 95,000 miles of pipelines (Figure 7).

Figure 7:    U.S. Network of Refined Products Pipelines

Various grades of motor gasoline, diesel, and other petroleum products are routinely transported in the same physical pipeline as sequential batches.

Batches of petroleum products are pumped through the system and, where adjacent batches come into contact, some mixing occurs between different products. The mixing at the interface is referred to as the transmix, and the contents of the transmix are generally off spec relative to motor gasoline or diesel specifications. Off-spec batches are downgraded or reprocessed at additional cost.

Petroleum refining and distribution are highly energy intensive. As indicated in Figure 8, 53 percent of the retail price of a gallon of diesel fuel corresponds to the cost of crude oil. In addition, approximately 20 percent of refinery operating costs and product distribution costs correspond to energy costs related to electricity and steam process use (Energy Information Administration, 2004). Accordingly, more than 57 percent of the retail price of diesel fuel (53% + 2.8% + 1.4%) reflects energy costs for crude oil, electricity, and steam. Without considering taxes, more than 77 percent of production and distribution costs are associated with energy inputs.

Given the highly energy-intensive nature of petroleum refining, industry practice is to express operating efficiency improvements as energy or BTU savings. Since refined product losses in the distribution system must also be made up by reprocessing the off-spec transmix, the operating efficiency of storage and pipeline systems tends to be gauged with reference to energy savings as well (Office of Industrial Technologies, 1998).

Figure 8:    Cost Components of Average U.S. Diesel Price

Transition to Ultra-Low Sulfur Diesel

Diesel is the primary fuel for the U.S. commercial transportation sector. In 2000, the transportation sector used 33.1 billion gallons, and more than 94 percent of all freight was moved using diesel power (Diesel Fuel News, 2002). The U.S. refining industry supplies more than 92 percent of domestic diesel consumption (National Petroleum Council, 2000), and the sulfur content in diesel fuel is a major contributor to particulate and sulfur dioxide emissions.

As noted, the 2001 EPA rules require sulfur content reductions in highway grade diesel fuel from 500 to 15 ppm ultra-low sulfur content. "Pipeline operators are expected to require refiners to provide diesel fuel with even lower sulfur content— somewhat below 10 ppm—in order to compensate for possible contamination from higher sulfur products in the distribution system and to provide a tolerance for testing" (Energy Information Administration, 2001). The rule will start to take effect in the fall of 2006, essentially requiring compliance by the end of 2008.

EPA estimates that the cost of reducing sulfur content from 2004 regulated levels to 15 ppm ultra-low levels will be 5.4 cents per gallon, consisting of 4.3 cents in additional refinery costs and 1.1 cents in additional distribution costs (U.S. Environmental Protection Agency, 2000). The American Petroleum Institute projects even higher additional costs, exceeding 15 cents per gallon, consisting of 9 cents in refining costs and 6 cents in distribution costs (Peckhan, 2000). According to the American Trucking Association, the total cost of achieving the ultra-low level is likely to be "someplace in the middle, around 10 cents per gallon" (Thrift, 2003).

Challenges of Achieving Ultra-Low Sulfur Content

The petroleum refining and distribution industry will be significantly challenged to achieve 15 ppm ULSD levels in the 2006-2008 timeframe. ULSD fuels will require not only new processes to extract sulfur from refinery streams but also the utilization of more accurate and near-real-time metrology (National Petroleum Council, 2000).

The optimal blending of crude feedstock with differing sulfur content and the hydro-treating of finished diesel fuel streams will be the two primary methods for achieving ULSD standards in U.S. refineries. Optimal blending and hydro-treating will be technically difficult and high-cost processes (National Petroleum Council, 2000).

Beyond addressing the operational challenges of crude blending and hydro-treating, the lack of precise measurement tools will further exacerbate the difficulty of reliably achieving 15 ppm. Measurement uncertainties of currently available metrology can exceed 16 ppm, "suggesting that without revolutionary improvements in sulfur measurement, refinery blenders will have to produce batches of diesel fuel with zero sulfur content in order to ensure that the refinery is compliant with the 15 ppm standard" (BP America, 2002).

The distribution system for ULSD will also be hampered by limitations in sulfur measurement accuracy and speed. According to the American Petroleum Institute, maintaining product integrity in terminals and pipeline systems and minimizing costly transmix downgrades will be significantly more difficult when transporting 15 ppm ULSD fuel. The American Petroleum Institute projects that interface mixing of fuel batches and sulfur contamination from other fuels will result in up to 17 percent of ULSD fuels being downgraded, requiring costly reprocessing, as compared to less than 1 percent with current higher levels of sulfur content (Peckhan, 2000).

To minimize downgrades and associated economic losses, the flow of refined products should be monitored on a near-real-time basis. According to the National Petroleum Council (2000), "low sulfur fuels are easily contaminated even with a small batch of out-of-spec fuels. Manual sampling takes too long and results in deep interface cuts and significant downgrades. Continuous online monitoring rather than periodic sampling will be essential."

X-Ray Sensor Engines for Sulfur Monitoring in Refinery Processes

XOS is currently marketing an affordable and innovative measurement technology for monitoring ULSD trace concentration; that is, sensor engines equipped with X-ray optics and with limits of detection as low as 0.4 ppm. These sensor engines can be placed in-line with refinery processes for near-real-time monitoring of sulfur content.

XOS recently concluded an agreement with a major petroleum company (among the top three petroleum companies in the United States) to be its exclusive supplier of in­line sulfur detection sensors for measuring ULSD trace concentration. "In-line sulfur monitoring will provide closed-loop process control so that refiners can quickly detect out-of-spec fuel, take corrective action, identify the root causes of fuel quality problems, and avoid unnecessary operating costs" (Beumer and Radley, 2004). These benefits will be particularly valuable for two refinery processes central to the production of ULSD:

• Optimal feedstock blending: Crude supplies vary significantly in sulfur content. Real-time measurement of sulfur content will facilitate the optimal blending of crude oil arriving at distillation columns. Optimal blending will reduce process excursions and downtime due to swings in crude sulfur content. Accordingly, energy savings will be realized as distillation columns are operated at stable sulfur composition levels.
• Avoiding excessive hydro-treatment: Achieving ultra-low sulfur levels will be a very expensive process. It will also result in refinery yield losses. If timely and accurate process information from in-line sulfur analyzers is used to prevent over-treating or extracting too much sulfur, then processing costs and yield losses can be reduced, resulting in energy savings.

X-Ray Sensor Engines for Sulfur Monitoring in Distribution Systems

XOS is also actively marketing affordable and innovative in-line sensor engines for monitoring the sulfur content of ULSD fuels moving through the U.S. storage terminal and pipeline distribution system.

After the deployment of in-line sensor engines, expected to start in 2005, pipeline operators will be able to more quickly detect out-of-spec conditions, take corrective action, and minimize costly interface cuts, or intermix, of different batches of refined petroleum products (Beumer, 2004).

Economic Modeling

To quantify cash flow benefits to industry users and the general public from the deployment of X-ray sensor engines in petroleum refineries and in the refined product distribution system, estimates are developed for average benefits per sensor engine and projected sales of sensor engines.

BENEFITS PER SENSOR ENGINE

Removing enough sulfur to meet regulatory requirements, without over-treating refinery streams, and accurately detecting out-of-spec conditions in storage terminals and pipelines will lead to increased operating efficiencies in refineries and distribution.

As a baseline for expressing efficiency gains in terms of energy savings, the U.S. Department of Energy, Office of Industrial Technologies (OIT), conducted a study of energy-use profiles in the U.S. petroleum industry and developed a computer model to estimate energy savings from specific process improvements (Office of Industrial Technologies, 1998).

The OIT model—available as a web-based interactive Project Evaluation Tool on OIT's website (see References)—was used to estimate energy savings resulting from the use of XOS sensor engines in monitoring sulfur content at U.S. petroleum refineries and in the U.S. distribution system. Based on the OIT model results, annual base-case energy savings were estimated as:

• 20,522 million BTUs per sensor engine used for monitoring sulfur levels in refinery crude blending and hydro-treating processes
• 11,543 million BTUs per sensor engine used for monitoring sulfur levels in storage terminals and pipeline systems.

Figures 9 and 10 indicate how refinery and distribution energy savings can be mapped into cost savings. It should be noted that some refinery processes use process waste streams as energy inputs with economic values lower than the market price of crude oil. Other refinery processes use electricity and steam, where the economic values are higher than the market price of crude oil. Given the substantial variability in refinery processes and energy use patterns, we conservatively assume that each BTU of energy saved from more accurate in-line monitoring of the blending and hydro-treating processes and from avoiding over-treating diesel streams can be valued at the market price of crude oil.

In its 2005 Annual Energy Outlook, the Energy Information Administration of the U.S. Department of Energy (January 2005) projects crude market prices over the 2005-2025 period to decline at least temporarily from current high levels as new deepwater oil fields are brought into production and projects $24.50 per barrel of crude oil (2003 dollars) by 2010 and $30 per barrel in 2025.

Assuming that Energy Information Administration price projections may be somewhat optimistic, we assumed the market price for crude at $35/barrel ($10.50 above 2005 DOE estimates) over the 2005-2014 planning horizon for benefit analysis in the base case. At this price level, annual savings per sensor engine were estimated at $122,822. For a step-out scenario, we assumed crude oil at $40/barrel ($15.50 above 2005 DOE estimates), leading to annual savings of $140,368 per sensor engine (Figure 9).

Figure 9:   Projected Annual Energy Savings from In-Line Sensor Engines Deployed at U.S. Refineries (2004 Dollars)

Should long-term energy prices exceed DOE projections beyond our adjustments, annual public benefits to be realized from X-ray optics technology will be greater than the above base case and step-out scenario estimates.

To estimate the economic value of distribution system BTU savings, we used the same assumptions about the market price of crude oil as for refinery energy savings. Over the 2004-2014 planning horizon, we assumed a base-case market price for crude oil at $35/barrel, leading to estimated annual savings of $69,085 per sensor engine. For a step-out scenario, we assumed crude oil at $40/barrel, leading to estimated annual savings of $78,954 per sensor engine (Figure 10).
Beyond quantitative cash flow benefits from energy savings, the use of in-line sensor engines is expected to result in additional economic benefits, including:

• Reduced hydrogen use in refinery hydro-treating processes from the avoided over-treatment of diesel streams. Approximate value of hydrogen savings is $1,000 per standard cubic foot (Beumer, 2004).
• Reduced environmental pollution from the use of ULSD fuel in commercial transportation. Achieving effective compliance with ULSD regulations, SO2 and particulate emissions will be reduced. The Office of Management and Budget (2003) estimates that the economic value of avoided SO2 emissions is $7,800 per ton.

REFINERY SENSOR ENGINE SALES

Based on input from XOS and discussions with process control OEMs, refinery operators, regulators, and refinery industry associations, we developed the following sales estimates for modeling economic benefits of the ATP-funded X-ray optics project.

For a conservative base case, it is assumed that XOS will deploy 6 sensor engines at each of the 23 refineries that produce diesel fuel over the 2004-2008 period, representing 20 percent of the U.S. diesel market (see Figure 11). The timing of sensor engine sales is projected at 6 engines in 2004, 12 in 2005, 30 in 2006, 42 in 2007, and 48 in 2008, for a total of 138 sensor engines over the 2004-2008 period (Table 2).

Figure 10: Projected Annual Energy Savings from In-Line Sensor Engines Deployed in U.S. Petroleum Distribution System (2004 Dollars)

Figure 11: Projected Deployment of XOS In-Line Sensor Engines in U.S. Petroleum Refineries over 2004-2008 Period

For a more optimistic step-out scenario, sales levels for each year (over the 2004-2008 period) are increased by 10 percent over base-case levels.

Given XOS's agreement with a major U.S. petroleum refining company to be the exclusive supplier of in-line refinery metrology for ULSD, the probability of achieving these plans is estimated to be 90 percent.

SENSOR ENGINE SALES TO TERMINAL AND PIPELINE OPERATORS

For a conservative base case, one sensor engine will be deployed at each of 520 terminals and one sensor engine for each 600 miles of 38,000 miles of pipeline (63 engines), representing 40 percent of the U.S. petroleum product distribution market. The timing of in-line sensor engine sales is projected at 117 engines in 2005, 146 engines in 2006, 160 engines in 2007, and 160 engines in 2008, for a total of 583 sensor engines over the 2005-2008 period (Figure 12 and Table 2).

Figure 12: Projected Deployment of XOS In-Line Sensor Engines in U.S. Petroleum Distribution System over 2005-2008 Period

For a more optimistic step-out scenario, sales levels for each year (over the 2005-2008 period) are increased by 10 percent over base-case levels.

The probability of achieving these plans is estimated to be 75 percent.

Benefit Estimates

Based on discussions with XOS and industry experts, we assume that sensor engines will continue to generate benefits for 10 years after the sale and deployment of each engine. Annual cash flow benefit estimates are computed by extending estimated economic benefits per X-ray sensor engine by the number of sensor engines projected to be deployed in the United States.

Annual cash flow estimates of the economic value of projected benefits from using high accuracy in-line sensor engines, and computed on the basis of the above assumptions, are displayed in columns 3 and 4 of Tables 3 and 4 in the Benefit-Cost Analysis section of this chapter, where cash flows for the different X-ray optics applications are shown side by side.

PROCESS CONTROL IN SEMICONDUCTOR FABRICATION

Metrology tools of substantially higher sensitivity are needed to support the implementation of technology advances in semiconductor fabrication to:

• Replace aluminum with copper interconnect metal, in combination with low dielectric constant (low k) insulation
• Facilitate transition from 200 mm to 300 mm wafers

Metrology, including materials characterization in semiconductor fabrication, facilitates the introduction and manufacture of new materials, facilitates miniaturization, enables tool improvements, and can reduce manufacturing costs and time-to-market for new products, through better characterization of processes (International Technology Roadmap for Semiconductors, 2003).

XOS is currently developing metrology products for semiconductor fabrication. In­line sensor engines with X-ray optics will be sold to OEMs of process diagnostic tools and are expected to "add value as metrology sensors for wafer etching, vapor deposition, ion implantation, and chemical mechanical planarization" (Gibson, 2004). In-line sensors will be used to:

• Measure thin-film (copper) thickness and detect iron contamination in silicon wafers
• Characterize wafer crystal lattice orientation and texture, monitor epitaxial growth, and measure thin-film residual stress levels

The use of sensor engines is expected to provide a number of benefits in semiconductor fabrication, including:

• Reduced wafer handling by eliminating the need to transport wafers from a process tool to a laboratory metrology system for critical measurements (Dance et al., 1996).
• Accelerated product development. New chip fabrication facilities, starting with 80 percent test wafers, end up with fewer than 20 percent test wafers after 3-5 years (Gibson, 2004). If ATP-funded X-ray optical technology can accelerate semiconductor fabrication facilities toward a steady state 80 percent yield, waste and manufacturing costs will be reduced.
• Improved quality control with faster detection of out-of-spec process excursions, improved operating yields, and lower cost of quality.

The timing of in-line sensor sales to the semiconductor industry is expected to lag sensor sales for petroleum refinery and distribution applications by several years (Gibson, 2004).

Given the longer time frame and associated commercial uncertainties of benefits of the ATP-funded X-ray optics project to the semiconductor industry, these benefits are not quantified at this time.

BENEFIT-COST ANALYSIS

The commercial utilization of ATP-funded X-ray optics has resulted in actual, realized economic benefits. Even greater future benefits are anticipated from the continued use of standalone optics and from the introduction of new sensor engines incorporating X-ray optics. The flow of these benefits, via four distinct pathways in multiple industries, is summarized in Figure 13.

Figure 13: Flow of Benefits from X-Ray Optics Technology

ATP AND INDUSTRIAL PARTNER INVESTMENTS

During the 1992-1996 period, ATP invested $1.95 million (toward project direct costs) and its industry partner, XOS, invested $371,000 in the development of the X-ray optics technology.

For purposes of cash flow analysis, the ATP investment was normalized to 2004 dollars ($2.496 million) and included as one lump sum investment made in 1994, the midpoint of the four-year investment period.

PERFORMANCE METRICS

Our quantitative analysis was limited to public benefits (to downstream industry, end-users, and society) that could be meaningfully quantified and excluded private benefits to XOS. 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—that compute directly the impact of the ATP investment.

In this case, all benefits to downstream industrial users and the general public are attributed to the ATP project because it is highly unlikely these benefits would have occurred without ATP.

BASE-CASE ANALYSIS

Laboratory materials analysis savings result from the commercial sale of X-ray optics. In contrast, refinery and distribution savings—associated with the use of sensor engines—are driven by EPA's ULSD regulations, to be implemented by the 2008 regulatory deadline.

As indicated in Table 3, for the base case, the public return (combined retrospective and prospective) on ATP's investment in the X-ray optics technology over the 1994-2014 period can be expressed as a net present value of $184 million. Public benefits are $75 for every dollar invested, and the internal rate of return is estimated at 49 percent.

A purely retrospective analysis of the return on ATP's investment over the 1994-2003 period indicates a realized net present value of $7.4 million, a benefit-to-cost ratio of $4 of public benefits for every dollar invested, and an internal rate of return of 30 percent.

STEP-OUT SCENARIO ANALYSIS

A step-out scenario analysis was conducted to investigate the sensitivity of performance metrics to more optimistic assumptions about the projected future benefits of X-ray optics installed into laboratory analytical instruments and in-line sensor engines, in combination with a 10 percent increase in projected sales.

As indicated in Table 4, for the step-out scenario, the public return on ATP's investment in X-ray optics technology is associated with a net present value of $233 million. Expected public benefits are $94 for every dollar invested, and the internal rate of return is estimated at 53 percent.

PRIVATE BENEFITS TO ATP INDUSTRY PARTNERS

ATP's industry partners' motivation to continue product development and marketing of the ATP-funded technology is a necessary pre-condition for industrial-scale impact. Only then will the general public come to enjoy the economic and environmental benefits expected to result from ATP's investment in X-ray optics.

XOS's commercial success and continued motivation to actively market and improve in-line sensor engine products are reflected by its growth from 2 employees in 1991 to 35 employees in early 2003. It continues to develop and market new commercial products based on the ATP-funded technology and to enter into joint ventures with instrumentation and process control OEMs. The company is privately held and had positive net income for five years.

SUMMARY

Base-case and step-out scenario performance metrics point to an exceptional performance of ATP's investment in the X-ray optics project, demonstrated by high public rates of returns and important qualitative benefits.

__________________
Note: A 1994 base year and an OMB-designated 7 percent discount rate were used for analysis. Performance metrics were computed from time series assuming ATP investment in 1994 (project midpoint) and cash flow benefits from 1997 to 2014. Positive cash flows represent public benefits attributable to ATP; negative cash flows represent ATP investment costs (assumed to occur at project midpoint).

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)