SEARCH

SEARCH BY CITATION

Keywords:

  • health risk;
  • high pressure carbon monoxide (HiPco);
  • industrial ecology;
  • Monte Carlo simulation;
  • nanotechnology;
  • uncertainty

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Model Description
  5. Results
  6. Addressing Uncertainty in the Rate of EHS Standards Development
  7. Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window
  8. Conclusions
  9. Acknowledgments
  10. References
  11. About the Authors

Considerable uncertainty exists about occupational risks, future environmental health and safety (EHS) standards, and associated production and compliance costs for single-wall carbon nanotube (SWNT) manufacturing processes. We propose and illustrate the use of risk analysis Monte Carlo (MC) models to assess cost and exposure trade-offs of the high-pressure carbon monoxide (HiPco) SWNT manufacturing process given these uncertainties. Assumptions regarding the timing, frequency, magnitude, and expense of EHS standards are modeled as stochastic events and examined for their impact on the expected values, variances, and probability distributions of total production costs and occupational exposure. With a better understanding of associated risks, these models can help policy makers and manufacturers explore potential EHS benefits, consequences, and trade-offs. For example, results suggest that voluntary implementation of a low level of protection (rather than none at all) can lead to reduced cost and exposure uncertainty with insignificant increases in production costs, as well as lowering total manufacturing and liability costs, depending on the assumptions made. Conversely, slower implementation rates of higher standards produce greater uncertainty in long-term costs and exposure. More generally, the results of this study underscore three important observations: (1) Expected costs alone are insufficient for informed decision making; (2) the best level of standards, overall cost, and optimal voluntary standards are highly dependent on uncertain health effects; and (3) the resultant amount of uncertainty in total costs and exposure can be extreme.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Model Description
  5. Results
  6. Addressing Uncertainty in the Rate of EHS Standards Development
  7. Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window
  8. Conclusions
  9. Acknowledgments
  10. References
  11. About the Authors

As nanotechnology moves from development to commercialization, interest has grown in understanding production costs and occupational health risks associated with various nanomanufacturing processes. Commercialization of nanotechnology is proceeding quickly, with over 600 products identified already in commerce (Project on Emerging Nanotechnology 2006). Global spending in 2006 was cited at more than $12 billion (Maynard 2007). This enormous investment, however, greatly exceeds the funding dedicated to the occupational, consumer, and environmental health and safety (EHS) implications of nanotechnology (Maynard 2006). There are indications that engineered nanomaterials may present potential risks to human health. Types of nanomaterials that are most likely to present health risks include nanoparticles, agglomerates of nanoparticles, and particles of nanostructured materials.

Possible negative properties of these materials include their ability to penetrate dermal barriers, cross cell membranes, breach the gas exchange regions of the lung, travel from the lung throughout the body, and interact at the molecular level (NIOSH 2007). In particular, critical reviews on the toxicity of single-wall carbon nanotubes (SWNTs) generally acknowledge important research showing damage to lung tissue in mice (Donaldson et al. 2006; Lam et al. 2006) but also discuss the large amount of EHS uncertainty and indicate that more work is necessary to determine the full consequences in production environments. A primer on nanotechnology risk assessment issues further cautions against interpretation of preliminary findings (Bell 2007). Related nanomanufacturing EHS research findings are available through a searchable database of information through the International Council on Nanotechnology (ICON 2005).

Several calls for increased attention to environmental health and safety issues related to this emerging technology have been made (Maynard 2006; Maynard et al. 2006), and the U.S. Environmental Protection Agency has announced its intent to develop a roadmap for EHS research needs (EPA 2007). An earlier study undertaken by the Royal Society and the Royal Academy of Engineering (2004) developed a list of recommendations ranging from actions to achieve more sufficient EHS data to suggestions on where regulation should be considered. The National Institute for Occupational Safety and Health (NIOSH) recently published an overview of research in this area undertaken at that agency, summaries of accomplishments, and suggestions for additional research needs (NIOSH 2007). Although NIOSH (2006) also suggested preliminary guidelines for working safely with nanomaterials, research clearly is needed to define risks, provide guidance for safe handling of nanomaterials, and minimize workplace exposure (NIOSH 2005). Various other U.S. government agencies are involved in contributions to the EHS research effort. An overview of the primary EHS research and information needs recently was developed by the National Science and Technology Council (NSTC 2008) of the U.S. National Nanotechnology Initiative to guide the vast effort needed to ensure responsible development of nanotechnology and to inform policy makers.

Completion of this research, however, is expected to take some time. Until then, policy makers and businesses working with nanomaterials are faced with significant uncertainty in how to proceed with commercialization of their products and regulatory safeguards. It may take several years to develop more complete EHS information, let alone consensus on policy to support responsible commercialization of manufactured goods. The first potential exposures will occur in the workplace, which raises issues of occupational safety and health in manufacturing facilities or laboratories. The evaluation of work practices, administrative actions, engineering controls, and personal protective equipment (PPE) in manufacturing environments is likely to lead to improved best practices for reducing exposure to nanomaterials (NSTC 2006). Leading companies are working to develop guidelines (DOE 2007; Medley and Walsh 2007; Ellenbecker et al. 2008), but, again, these results are not yet available or the benefits are unknown.

Life cycle assessment (LCA) and cost modeling techniques have been successfully used to assess the economic and environmental impact of the manufacture of carbon nanotubes (Isaacs et al. 2006). Such models compute the total production, disposal, and environmental cost over the entire life of a product and are becoming an increasingly common materials assessment method (Schmidt 2007). Little is known about the likelihood, cost, and effectiveness of various EHS protections for these materials, however, and results therefore are incomplete.

Given the myriad of uncertainties, predictive models would be advantageous to help explore the potential EHS consequences and trade-offs, to develop better insight on how manufacturing and total costs may change given various assumptions about the imposition and timing of such standards for workers, and to evaluate proactive adoption of higher levels of protection. Monte Carlo (MC) risk analysis models therefore were developed to assess the potential range of production costs and health effects associated with various types and levels of occupational safety requirements (engineering controls, administrative controls, PPE, etc.), representing uncertainties in the level and timing of possible requirements and in the dose-exposure relationship by probability distributions and chance events. Similar models are used in a wide range of other applications in which uncertainty exists, from nuclear power to financial investment decisions (Apostolakis 2004; Pagani et al. 2005; Kimura and Shinohara 2006), and allow exploration of a range of potential scenarios and assumptions.

The remainder of this article is organized as follows. The next section describes the model development, logic, and assumptions. Four sets of assumptions are described that represent the range of possible likelihoods and speeds with which various levels of EHS manufacturing standards could be imposed. The third section illustrates the types of results these models can produce and compares the range of possible costs and exposure levels for the high-pressure carbon monoxide (HiPco) nanomanufacturing process. Given the large amount of uncertainty regarding EHS standards and effectiveness, the fourth and fifth sections expand this approach to address uncertainty in the rate by which increased levels of standards may be required and the unknown exposure effects, taking a Bayesian approach to model these uncertainties and trade-offs.

Model Description

  1. Top of page
  2. Summary
  3. Introduction
  4. Model Description
  5. Results
  6. Addressing Uncertainty in the Rate of EHS Standards Development
  7. Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window
  8. Conclusions
  9. Acknowledgments
  10. References
  11. About the Authors

Although new techniques to tailor the formation of SWNTs are under development (Hata et al. 2004; Zhang et al. 2005; Miller et al. 2006), three established techniques currently exist: arc ablation, chemical vapor deposition (CVD), and HiPco processes. Process-based technical cost models previously were developed to assess the production costs of these processes (Isaacs et al. 2006), but assumptions regarding occupational health risks, EHS costs, and future safety requirements were not included. Because HiPco processes resulted in significantly lower costs ($450 per gram [g] versus $1,830/g for arc ablation and $1,586/g for CVD),1 only HiPco risk models were developed further to explore these assumptions. An earlier MC model explored the utility of this approach (Ok et al. 2007) and has since been expanded to include exposure costs and to investigate trade-offs of various levels of EHS standards.

Four general levels of EHS standards are defined in the model (none, low, medium, and high) to represent the range of rates and probabilities by which nano-EHS standards might be imposed. Examples of each level are shown in table 1 and could correspond to different levels of control banding (NIOSH 2004). Low levels of standards include moderate engineering controls (e.g., 24-hr general exhaust ventilation with a rate of 1,000 cubic feet/meter [cfm]),2 administrative controls (e.g., 1 day of annual training for good industrial hygiene practices and monthly monitoring of nanoparticle exposure in the workplace), and average PPE requirements (e.g., latex gloves and disposable respirators). Medium standards include local exhaust ventilation (e.g., lab fume hoods), the same administrative controls, and nitrile gloves and disposable respirators as additional PPE. High-level standards include full enclosure of processes; annual medical monitoring for workers in addition to annual industrial hygiene training; biweekly nanoparticle monitoring; and tighter PPE requirements, such as full Tyvek suits, gloves, and respirators with high-efficiency particulate air (HEPA) filters.

Table 1.  Summary of assumptions for environmental health and safety (EHS) standards
Type of EHS ControlLevel of EHS standards
None (1)Low (2)Medium (3)High (4)
  1. Note: Assumptions are based on discussion with an industrial hygienist, and costs are estimated from other sources. cfm = cubic feet/meter. One square foot (ft2) ≈ 0.093 square meters (m2, SI).

  2. aBuilding costs are not shown here but are included in the cost model.

  3. bCost of hazardous waste disposal for personal protective equipment at the high level is not shown here but is included in the cost model.

  4. cCosts of medical clearance and fit tests for respirators with high-efficiency particulate air (HEPA) filters are not shown here but are included in the cost model.

Engineering controls
General exhaust—ventilation 24 hr, 1,000 cfm ventilation rate, $10,000 capital cost, $3,000/year operating cost24 hr, 1,000 cfm ventilation rate, $10,000 capital cost, $3,000/year operating cost24 hr, 1,000 cfm ventilation rate, $10,000 capital cost, $3,000/year operating cost
Fume hoods  $4,000 capital cost for 6.25 ft2 equipment and $9,500 for 25 ft2 equipmenta$4,000 capital cost for 6.25 ft2 equipment and $9,500 for 25 ft2 equipmenta
Enclosure of processes   50% decrease in labor productivity, 50% extra equipment cost
 
Administrative controls
Annual worker training 8 hr of training, $560/year instructor cost8 hr of training, $560/year instructor cost8 hr of training, $560/year instructor cost
Air monitoring Monthly monitoring, $20,000/equipment capital costWeekly monitoring, $20,000/equipment capital costBiweekly monitoring, $20,000/equipment capital cost
Medical monitoring   $950/worker/year
 
Personal protective equipmentb
Gloves
 Latex 5 pairs/shift, $0.06/pair  
 Nitrile  5 pairs/shift, $0.09/pair5 pairs/shift, $0.09/pair
Respirators
 Disposable 1/shift, $0.701/shift, $0.70 
 HEPA filters   1 pair/30 hrs, $10/pair
Tyvek suits   1/shift, $4c

The likelihoods of moving from one level to some higher level of EHS standards are defined by transition probabilities, with Year 0 representing the start-up year (with no standards yet required). Notationally, pi,i represents the probability of staying at a current level i of EHS standards, and pi,j represents the probability of transitioning from level i to level j, with levels 1, 2, 3, and 4 corresponding to none, low, medium, and high, respectively. Once a higher level of standards is adopted, it is assumed that subsequently this would not be relaxed to a lower level. For example, if a medium-level EHS standard is enacted in some given year, then from that point forward it is only possible to operate a production facility at this or higher levels of protection.

These implementation rates are dependent on technology, new research on health exposure, and political forces, and are largely unknown and difficult to estimate. Therefore, four separate sets of assumptions were developed to represent the ranges of possibilities, as summarized in table 2. In row 1 of Scenario 1, for example, under these conservative assumptions the year-to-year likelihood that no standards continue to be required is high (p1,1= 0.95), whereas the probabilities of enacting low, medium, or high levels of standards all are low (p1,2= 0.02, p1,3= 0.015, and p1,4= 0.015, respectively). If a low level of standards is imposed in some future year, then under this scenario the probability of remaining at this low level of protection in subsequent years again will be higher (p2,2= 0.4) than the probability of transitioning to higher levels (p2,3= 0.3 and p2,4= 0.3, respectively), as shown in the second row for Scenario 1, although any set of probability assumptions could be used. Once some medium level of standard is introduced, then the probabilities of remaining at this level in the next year or of moving to a higher level of standards both are p3,3=p3,4= 0.50. Similar transition logic is assumed for the other scenarios with the shown probabilities and with this process repeating for a user-specified number of years, with all results based on 10,000 replications of a 10-year analysis window.

Table 2.  Probabilities of transition to higher environmental health and safety (EHS) standards
ScenarioFrom/toNoneLowMediumHigh
ScenarioNone0.95 0.02 0.0150.015
 1Low0.0000.4  0.3  0.3  
Medium0.0000.0000.5  0.5  
High0.0000.0000.0001.000
ScenarioNone0.4750.4750.0250.025
 2Low0.0000.9  0.05 0.05 
Medium0.0000.0000.5  0.5  
High0.0000.0000.0001.000
ScenarioNone0.3170.3170.3170.049
 3Low0.0000.4640.4640.072
Medium0.0000.0000.8660.134
High0.0000.0000.0001.000
ScenarioNone0.0250.0250.4750.475
 4Low0.0000.0260.4870.487
Medium0.0000.0000.5  0.5  
High0.0000.0000.0001.000

Note that Scenarios 2, 3, and 4 have progressively smaller probabilities for imposing low levels of standards and larger probabilities of enacting medium and high levels. Entropy probabilities—that is, the probabilities of remaining at the current level of EHS standards (p1,1, p2,2, p3,3, p4,4), under each scenario—are shown on the diagonals in table 2 in gray and are summarized in table 3. Because high-level standards are the absorbing states (i.e., after a high level of standards is adopted, the facility will remain in this state indefinitely), the entropy probability for the highest level standards is p4,4= 1 under each scenario. Note that each scenario has its largest entropy probability (before reaching this absorbing state) at a different level of standards, reflecting the degree of aggressiveness in implementing higher standards. For example, under Scenario 1 p1,1= 0.95, under Scenario 2 p2,2= 0.9, under Scenario 3 p3,3= 0.87, and under Scenario 4 p4,4= 1. Because results will be dependent on which set of assumptions is used, we first analyze each scenario separately in the next section and later combine these results in the fourth section by assigning “degree-of-belief” likelihoods to each.

Table 3.  Entropy probabilities of remaining at current environmental health and safety (EHS) standards level
Assumptions about introduction of EHS standardsRequired standards
NoneLowMediumHigh
  1. Note: Shaded valves indicate large entropy values before the absorbing state is reached for each scenario.

Scenario 10.95 0.40 0.50 1.000
Scenario 20.4750.90 0.50 1.000
Scenario 30.3170.4640.8661.000
Scenario 40.0250.0260.50 1.000

To determine baseline per gram manufacturing costs for each EHS level, we reran the HiPco SWNT process-based technical cost model described above using the cost assumptions summarized in table 1 and assuming 10,000 grams per year of production. Figure 1 summarizes the cost results at four levels of EHS standards as the production volume per year increases. With the original base case assumptions of 50% synthesis reaction yield, 90% purification yield, 10,000 g/yr annual production, and no EHS standards, the cost is $450/g. If low, medium, or high levels of standards are adopted, this cost becomes $460/g, $528/g, or $660/g, respectively. These results are shown in figure 1 for the base case production volume.

image

Figure 1. Comparison of high-pressure carbon monoxide (HiPco) production costs for each level of environmental health and safety (EHS) standards.

Download figure to PowerPoint

In actuality, however, standards at any level or in any workplace are likely to have different and largely unknown implementation costs, such as due to PPE change-out or disposal schedules, so these uncertainties are represented with triangular probability distributions according to the parameters listed in table 4—that is, the minimum, most likely, and maximum possible manufacturing costs (here with the minimum and maximum values being set to ±10% of the most likely value). As the MC model executes, for each year in which new EHS standards are implemented, nondecreasing manufacturing costs are generated initially from the corresponding distribution with these parameters, by means of repeated accept–reject logic to ensure that costs increase as a facility implements higher levels of worker protection (e.g., resampling until CostMedium > CostLow). Possible manufacturing cost reductions due to changes in production processes (e.g., technological innovations, change in market conditions) or any exposure extremes in the workplace are not represented in the model.

Table 4.  Assumed cost ranges for each level of environmental health and safety (EHS) standards calculated from cost model
EHS standards levelCost ranges[$/g]
Minimum (−10%)Most likelyMaximum (+10%)
None$405$450$495
Low$414$460$506
Medium$475$528$581
High$594$660$726

Although additional theoretical and experimental studies are needed to evaluate the effectiveness of aerosol control methods for nanoparticles, limited data to date indicate that conventional ventilation, engineering controls, and filtration approaches should be applicable for many potential exposure scenarios (Maynard and Kuempel 2005). Because no data are readily available regarding occupational health exposure for HiPco nanomanufacturing processes, exposure is represented by an arbitrary 0 to 10 scale, as shown in table 5. The lowest and highest possible annual amounts of exposure (0 and 10, respectively) indicate progressively decreasing exposures as a facility moves to higher levels of worker protection. For example, having no standards in place is assumed to produce minimum, most likely, and maximum values of 7, 9, and 10 units of exposure per year, respectively, whereas a low level of standards reduces the worker's exposure to minimum, most likely, and maximum values of 6, 8, and 9 units per year, respectively. Similar to above, nonincreasing units of exposure are randomly generated from triangular distributions with similar consistency logic (e.g., ensuring ExposureMedium < ExposureLow), where the “units” could represent accumulated toxicity or dose levels, currently unknown but with some nonetheless (unknown) cost or consequence. Over the 10-year simulation period, the total and average exposure amounts are calculated as time-based statistics in the usual manner (Law and Kelton 1995) and herein called the annualized exposure.

Table 5.  Exposure assumptions for each level of environmental health and safety (EHS) standards
EHS standards levelExposure units (on scale of 0 to 10)
MinimumMost likelyMaximum
None7910
Low689
Medium346
High012

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Model Description
  5. Results
  6. Addressing Uncertainty in the Rate of EHS Standards Development
  7. Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window
  8. Conclusions
  9. Acknowledgments
  10. References
  11. About the Authors

Cost and Exposure Distributions for HiPco Process

Figure 2 illustrates the probability distributions of per gram manufacturing costs and annual exposure amounts for HiPco processes assuming the standards introduction rates given in Scenario 1, based on 10,000 replications over a 10-year simulation period. Whereas the deterministic baseline manufacturing cost was $450/g, the expected value and standard deviation now are $490/g ± 1.3 and $65/g, respectively, due to the uncertain cost, timing, and levels of EHS standards. In figure 2b, the mean annualized exposure is 7.2 ± 0.05 units, with a standard deviation of 2.4. It is important to note that in this type of analysis the mean corresponds to the most likely one-time outcome, and the standard deviation indicates the amount of uncertainty in this one value (rather than the expected variation in a set of outcomes). Thus, the most likely values of $490/g and 7.2 exposure units do not represent the full range of possible outcomes. The actual manufacturing cost could fall anywhere between roughly $405/g and $726/g (a range of $321), and the actual units of exposure could fall anywhere between 0 and 10 units. These large ranges are due to the amount of uncertainty assumed within the model, with distributions that can be quite skewed and asymmetric about the expected value. For example, the slow standards development rates assumed in Scenario 1 might mean no standards are enacted over the entire 10-year period, resulting in an almost 70% chance that manufacturing costs will be less than $500/g and a 60% probability that annualized average exposure will exceed 7.

image

Figure 2. Probability distributions of high-pressure carbon monoxide (HiPco) (a) manufacturing costs and (b) exposure amounts, under Scenario 1 assumptions. Mean cost is $490/g ± 1.3 with a standard deviation of $65, and mean annualized exposure is 7.2 ± 0.05 with a standard deviation of 2.4 units.

Download figure to PowerPoint

Figure 3 compares similar probability distributions of HiPco production costs and exposure amounts for all four EHS scenarios, with the corresponding means, standard deviations, and 95% confidence intervals of costs and exposure given in the first row of table 6. Note that Scenarios 2, 3, and 4 exhibit increasingly higher expected production costs ($528/g, $573/g, and $646/g, respectively) and increasingly lower amounts of exposure (5.6, 3.4, and 1.4, respectively). The amount of uncertainty in actual costs and exposure also tends to decrease ($65/g, $53/g, and $31/g for costs and 2.3, 1.5, and 0.6 for exposure) for the higher scenarios (given their more certain imposition of higher standards). For example, under Scenario 3, the actual production cost is most likely to lie between $500/g and $650/g, with less uncertainty than Scenarios 1 and 2 due to some levels of standards almost certainly being implemented after a few years. Accordingly, Scenario 4 (the most aggressive in standards implementation, with the highest expected cost and lowest expected exposure) has the smallest standard deviations, with roughly a 90% chance that the actual manufacturing cost will be between $600/g and $700/g and that exposure will be less than 2, because medium- or high-level standards are likely to be enacted fairly quickly.

image

Figure 3. Probability distributions of high-pressure carbon monoxide (HiPco) (a) manufacturing costs and (b) exposure amounts under four assumption scenarios regarding implementation probabilities of various levels of environmental health and safety (EHS) standards.

Download figure to PowerPoint

Table 6.  High-pressure carbon monoxide (HiPco) production costs and exposure given voluntary initial standards
Initial level of voluntary standards (Year 0)Assumptions about introduction of environmental health and safety (EHS) standards
Scenario 1Scenario 2Scenario 3Scenario 4
Cost $/g SWNT (SD)Exposure units (SD)Cost $/g SWNT (SD)Exposure units (SD)Cost $/g SWNT (SD)Exposure units (SD)Cost $/g SWNT (SD)Exposure units (SD)
  1. Note: Means are given with 95% confidence intervals. SWNT = single-wall carbon nanotube SD = standard deviation. One gram (g) = 10−3 kilograms (kg, SI) ≈ 0.035 ounces (oz).

None$490 ± 1.37.2 ± 0.05$528 ± 1.35.6 ± 0.05$573 ± 1.03.4 ± 0.03$646 ± 0.61.4 ± 0.01
($65)(2.4)($65)(2.3)($53)(1.5)($31)(0.6)
Low$498 ± 1.26.4 ± 0.04$521 ± 1.35.6 ± 0.05$573 ± 1.03.4 ± 0.03$646 ± 0.61.4 ± 0.01
($62)(2.1)($69)(2.3)($53)(1.5)($31)(0.6)
Medium$552 ± 0.93.7 ± 0.02$567 ± 0.93.4 ± 0.02$583 ± 1.02.9 ± 0.02$646 ± 0.61.3 ± 0.01
($44)(1.1)($48)(1.2)($50)(1.2)($32)(0.6)
High$660 ± 0.51.0 ± 0.01$660 ± 0.51.0 ± 0.01$660 ± 0.51.0 ± 0.01$660 ± 0.51.0 ± 0.01
($27)(0.4)($27)(0.4)($27)(0.4)($27)(0.4)

Figure 4 compares the corresponding cumulative probability distributions for HiPco manufacturing costs and exposure units for all four scenarios. Again, the conservative rates assumed in Scenario 1 result in a significantly higher probability (roughly 70%) than the other assumptions that the true cost will be less than $500/g, whereas the aggressive assumptions of Scenario 4 result in a high probability (roughly 70%) that the actual manufacturing cost is above $625/g. Conversely, the low introduction probabilities assumed in Scenario 1 for medium and high levels of standards result in a roughly 70% chance that the actual annualized 10-year exposure exceeds 6, whereas for Scenario 4 there is only a roughly 40% chance that this actual exposure amount will exceed 1. Figure 4 also illustrates that the costs and exposures under each set of assumptions are stochastically ordered; that is, given two scenarios A and B, XA is stochastically larger than XB (written XAsXB) if P(XAd) ≤ P(XBd) for all values of d (Ross 1998), where Xi here represents the actual cost or actual exposure under assumption i. This characteristic may be useful for ranking alternatives in the presence of uncertainty, as it indicates that the probability of exceeding any given level of cost or exposure is less for one of the scenarios (making it preferable). Nevertheless, the stochastically smallest and most preferable scenario with respect to cost (Scenario 1) also is the stochastically largest and least preferable with respect to exposure.

image

Figure 4. Cumulative probability distributions and stochastic orderings of (a) manufacturing costs and (b) exposure amounts for high-pressure carbon monoxide (HiPco), given four assumption scenarios regarding implementation probabilities of various levels of environmental health and safety (EHS) standards.

Download figure to PowerPoint

Voluntary Implementation of EHS Standards

The results thus far all assume that a company does not initially implement any EHS standards for handling nanomaterials (because none are yet required). A natural question, however, is whether some benefit may exist for a company to voluntarily implement a higher level of standards than currently is required. Rows 2, 3, and 4 of table 6 therefore summarize the HiPco expected costs and exposure amounts resulting from a company starting Year 0 instead with low, medium, or high levels of EHS standards, again with standard deviations shown in parentheses. Under each scenario, as the level of initial protection increases, expected costs increase and expected exposure decreases, both with smaller standard deviations. The one minor exception to this trend, with cost decreasing slightly under Scenario 2 for low initial standards, is likely due to cost assumption ranges overlapping for the no-standards and low-level standards categories. As would be expected, under all scenarios, voluntarily implementing the highest EHS standards in Year 0 produces the same mean costs and exposures, the minimum expected exposure, and the minimum uncertainty in cost and exposure (see bottom row of table 6).

To provide a feel for the distribution of potential costs and exposures, figure 5 summarizes the 2.5th and 97.5th percentiles under each initial level of voluntary standards, indicating the inner 95% probability interval within which the actual long-term cost and exposure will fall. For all scenarios, roughly the same probability intervals exist for cost and exposure if one starts with no or low standards, whereas medium and high levels of standards show higher costs, lower exposures, and tighter ranges for both. Although for no, low, and medium standards, selection of the initial standards level under any scenario becomes fairly arbitrary and irrelevant with respect to cost (under any scenario with a slightly higher expected cost under medium standards but roughly the same probability interval), the exposure probability intervals are narrower and lower under higher medium and high initial standards. As one can see by also including exposure risk in the analysis, it therefore may be best to adopt voluntarily medium to high levels of standards regardless of external requirements.

image

Figure 5. Expected values and 2.5th and 97.5th percentiles of high-pressure carbon monoxide (HiPco) (a) production costs and (b) exposure amounts under different environmental health and safety (EHS) standards assumptions. (Center points indicate expected values.)

Download figure to PowerPoint

Under the assumptions of Scenario 1, if a company initially does not adopt any standards, the resultant 10-year manufacturing cost per gram may fall anywhere between $418/g and $646/g (with 0.95 probability). If the company were to voluntarily adopt higher levels of standards (i.e., low, medium, or high), the expected cost would increase, because these standards otherwise would be required later (or never) within the 10-year analysis horizon. However, the upper 97.5th percentiles under the first three initial options (i.e., no, low, and medium level of standards) are very close; that is, the “worst case” cost per gram is roughly the same regardless of what level of standards the company voluntarily implements. The exception is implementing a high level of standards, which results in this (probabilistic) upper bound being roughly $60/g higher. The lower bounds on manufacturing costs under Scenario 1 assumptions are very close for no standards and the low level of standards, and increase as the company elects to implement an increasing level of standards (i.e., medium and high level of standards).

Initially implementing no standards under Scenario 1 results in the highest expected exposure (7.2 units) and a wide probability interval ranging from 1.5 units to 9.6 units, whereas proactively adopting a low level of standards results in a lower expected exposure of 6.4 units and a 95% probability interval ranging from 1.5 to 8.7 units. Voluntarily implementing a medium level of worker protection decreases the upper 97.5th exposure percentile; that is, the worst case exposure is significantly lower (i.e., 5.5 units) than the first two scenario options.

Under the assumptions of Scenario 4, given high probabilities to transition toward high standards, the manufacturing costs are roughly the same under the first three voluntary implementation options, all having an expectation of $646/g and a probability interval ranging from roughly $584/g to $706/g. Expected exposure and probability intervals also are roughly the same, 1.4 units and 0.3 to 2.8 units, respectively, under the first three voluntary implementation options. Adopting high standards in Year 0 results in the lowest expected exposure (1.0 units) and a probability interval of 0.2 units to 1.8 units. Similar comparative conclusions of this type can be made for other assumption scenarios.

If SWNT manufacturing eventually were found to cause serious injury, such as similar to that in asbestos manufacturing, exposure consequences could be substantial. Damage costs might include medical expenses, lost income, lost earnings, lost profit, and pain and suffering. For example, in a 1995 Pennsylvania case, a jury awarded $150,000 to $275,000 per worker for past and projected future medical expenses and noneconomic damages for asbestos-related disease (Giordano v. A.C. & S. Inc 1995). In a 1999 Florida case, $5 million was awarded for past and future pain and suffering in asbestos liability litigation (Owens-Corning Fiberglas Corp. v. McKenna 1999), with upward of $35 million awarded across the United States against asbestos manufacturers for occupational health injuries. For HiPco processes, for a facility with 25 workers (as assumed in the HiPco technical cost model), a conservative payout of $150,000 per worker would result in $3.75 million in projected medical expenses or could be as high as $35 million in liabilities against the manufacturer (Dennison and Freedman 2003). Break-even probabilities using potential liability damages therefore might be identified under which each level of EHS investment offsets liability costs.

Addressing Uncertainty in the Rate of EHS Standards Development

  1. Top of page
  2. Summary
  3. Introduction
  4. Model Description
  5. Results
  6. Addressing Uncertainty in the Rate of EHS Standards Development
  7. Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window
  8. Conclusions
  9. Acknowledgments
  10. References
  11. About the Authors

Because the best business strategy clearly depends on assumptions about EHS standards, we conducted a Bayesian type of analysis by placing two sets of likelihoods (“degrees of belief”), shown in table 7, on each of the four scenarios for introducing occupational health standards. The first case assumes that the conservative assumptions of Scenario 1 are the most likely (with 0.70 likelihood) and that the other three scenarios are increasingly unlikely (with respective likelihoods 0.15, 0.10, and 0.05). The second case assumes somewhat the opposite, with more likelihood placed on the more aggressive standards scenarios (with respective weights 0.20, 0.40, 0.30, and 0.10).

Table 7.  Degree of belief about environmental health and safety (EHS) standards implementation assumptions
Likelihood of implementation scenariosAssumptions about introduction of EHS standards
Scenario 1Scenario 2Scenario 3Scenario 4
Case 1 (conservative)0.700.150.100.05
Case 2 (aggressive)0.200.400.300.10

The probability distributions of manufacturing cost and exposure under each of the above cases are shown in figures 6 and 7, under the assumption that a company starts Year 0 with no voluntary standards. The expected cost is $511/g ± 1.5 for Case 1 and $545/g ± 1.5 for Case 2 ($34/g higher), with standard deviations of $74 for both cases. Corresponding expected annualized exposures are 6.3 ± 0.05 (with a standard deviation of 2.7) and 4.8 ± 0.05 (with a standard deviation of 2.6), respectively. The 95% probability intervals for costs differ slightly, ranging from $421/g to $668/g for Case 1 and from $431/g to $679/g for Case 2. Similarly, the 95% probability intervals for exposures differ slightly in the upper tail: 1.0 to 9.6 for Case 1, and 0.8 to 9.2 for Case 2. Either of these pairs of distributions therefore might be taken as fairly representative of the range of possible long-term manufacturing costs per gram and occupational exposure, given the current level of uncertainty about environmental health risks and subsequent potential standards.

image

Figure 6. Probability distributions of high-pressure carbon monoxide (HiPco) (a) manufacturing costs and (b) exposure amounts for Case 1 (conservative) assumptions. Mean cost is $511/g ± 1.5 with a standard deviation of $74, and mean annualized exposure is 6.3 ± 0.05 with a standard deviation of 2.7 units.

Download figure to PowerPoint

image

Figure 7. Probability distributions of high-pressure carbon monoxide (HiPco) (a) manufacturing costs and (b) exposure amounts for Case 2 (aggressive) assumptions. Mean cost is $545/g ± 1.5 with a standard deviation of $74, and mean annualized exposure is 4.8 ± 0.05 with a standard deviation of 2.6 units.

Download figure to PowerPoint

Similar analyses were performed under the assumption that a company voluntarily starts production with something higher than no occupational health protections, much as in results section, to help make decisions about initial plant design, worker protection, and financial investments. Figure 8 compares the corresponding cumulative probability distributions for HiPco production costs and exposure units for each level of initial voluntary EHS standards, under Case 1 assumptions. If a company does not adopt any nanomaterials EHS standards, the probability that the production cost will exceed $500/g is less than 0.50, whereas if a company voluntarily implements high EHS standards, the probability that the manufacturing cost will be above $625/g is roughly 0.90. Conversely, if a company does not voluntarily adopt any level of standards, there is more than a 50% chance that the exposure will exceed 7 and an 85% chance that it will be lower than 1 if a company voluntarily implements high EHS standards.

image

Figure 8. Cumulative probability distributions and stochastic ordering of high-pressure carbon monoxide (HiPco) (a) manufacturing costs and (b) exposure amounts for different initial voluntary environmental health and safety (EHS) standards under Case 1 (conservative) assumptions.

Download figure to PowerPoint

Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window

  1. Top of page
  2. Summary
  3. Introduction
  4. Model Description
  5. Results
  6. Addressing Uncertainty in the Rate of EHS Standards Development
  7. Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window
  8. Conclusions
  9. Acknowledgments
  10. References
  11. About the Authors

As is intuitive and evident in the above results, inherent trade-offs exist between production costs and health risks. Figure 9 illustrates these trade-offs for each level of initial voluntary standards, in all cases indicating a fairly equal trade-off rate between expected manufacturing cost and exposure. Because the total cost, T, of nanomanufacturing ultimately includes both the production cost and the cost of exposure, these two performance measures might be combined into some type of expected total cost function, such as simply T=C+kEa, where C is the per gram production cost, E is the amount of exposure, and k and a define the linear and non-linear per gram exposure costs. For example, these exposure costs could include long-term health effects and associated liabilities.

image

Figure 9. Trade-offs (a) between mean production costs and mean occupational exposure and (b) between standard deviation of production costs and standard deviation of occupational exposure. Each point is labeled descriptively (e.g., S2-M represents Scenario 2 with a Medium voluntarily implemented level of standards).

Download figure to PowerPoint

In this example, values of a > 1 or a < 1 result in increasing or decreasing exposure costs as exposure increases, respectively. Although these parameters are difficult to estimate and most likely will remain unknown for some time, figure 10 illustrates that situations exist for which each initial voluntary level of standards would be optimal (i.e., the level resulting in the smallest expected total cost for the above conservative Case 1). For example, if a= 0.1 and k= 100, then not voluntarily adopting any level of standards is optimal for a company, whereas with a= 0.5 and k= 75, a low level of voluntary EHS standards is optimal. In the linear case a= 1 and k= 50, voluntarily adopting a high level of standards is optimal, and for a > 1 where a= 3 and k= 1, the expected total cost becomes smallest as a company implements medium levels of standards (figure 10A). In keeping with earlier results, the minimum uncertainty in the total cost always occurs when a company voluntarily implements the highest possible standards (figure 10B).

image

Figure 10. Examples of the expected value and uncertainty of total cost (production plus exposure costs), for the case where T=C+kEa, where T is the total cost, C is the per gram production cost, E is the amount of exposure, and k and a define the linear and non-linear per gram exposure costs. The four cases correspond to each level of voluntary initial standards being optimal.

Download figure to PowerPoint

All results discussed to this point are based on a 10-year analysis window. Alternately, table 8 summarizes HiPco expected costs and exposure amounts for 5-year, 10-year, 15-year, 20-year, and 25-year analysis periods, assuming no initial standards in Year 0, with the corresponding probability distributions shown in figure 11. Expected values are listed with 95% confidence intervals, and standard deviations are shown in parentheses. Note that all four scenarios exhibit higher expected annual production costs and lower annualized exposure as the analysis window increases. The standard deviation increases as the analysis timeline increases only under Scenario 1, however, which has the greater uncertainty.

Table 8.  High-pressure carbon monoxide (HiPco) production costs and exposure amounts for different environmental health and safety (EHS) standards assumptions of simulation periods
Analysis periodAssumptions about introduction of EHS standards
Scenario 1Scenario 2Scenario 3Scenario 4
Cost $/g SWNT (SD)Exposure units (SD)Cost $/g SWNT (SD)Exposure units (SD)Cost $/g SWNT (SD)Exposure units (SD)Cost $/g SWNT (SD)Exposure units (SD)
  1. Note: Means are given with 95% confidence intervals. In all cases, it is assumed that no voluntary standards are implemented at Year 0. SWNT = single-wall carbon nanotube. SD = Standard deviation.

5-year$470 ± 1.07.9 ± 0.04$498 ± 1.16.6 ± 0.04$541 ± 1.14.5 ± 0.03$632 ± 0.81.7 ± 0.02
($51)(1.9)($56)(2.0)($55)(1.7)($43)(1.0)
10-year$490 ± 1.37.1 ± 0.05$528 ± 1.35.5 ± 0.05$573 ± 1.03.4 ± 0.03$646 ± 0.61.4 ± 0.01
($65)(2.4)($66)(2.3)($53)(1.5)($31)(0.6)
15-year$509 ± 1.46.5 ± 0.05$552 ± 1.34.7 ± 0.05$593 ± 1.02.9 ± 0.03$651 ± 0.61.2 ± 0.01
($73)(2.7)($66)(2.3)($49)(1.3)($29)(0.5)
20-year$523 ± 1.55.9 ± 0.05$569 ± 1.24.1 ± 0.04$608 ± 0.92.4 ± 0.02$653 ± 0.51.2 ± 0.01
($76)(2.7)($64)(2.2)($45)(1.1)($28)(0.5)
25-year$538 ± 1.55.4 ± 0.05$583 ± 1.23.6 ± 0.04$617 ± 0.82.2 ± 0.02$654 ± 0.51.1 ± 0.01
($76)(2.8)($60)(2.0)($41)(1.0)($27)(0.4)
image

Figure 11. Probability distributions of high-pressure carbon monoxide (HiPco) manufacturing costs and exposure amounts given Scenario 1 assumptions regarding implementation probabilities of various levels of environmental health and safety (EHS) standards for 5-year, 10-year, 15-year, 20-year, and 25-year simulation timelines.

Download figure to PowerPoint

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Model Description
  5. Results
  6. Addressing Uncertainty in the Rate of EHS Standards Development
  7. Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window
  8. Conclusions
  9. Acknowledgments
  10. References
  11. About the Authors

Assessing the trade-offs between nanomanufacturing costs and occupational health consequences is especially difficult given the dearth of data on the health effects of nanoparticles. Previous work using LCA methodologies to assess alternate SWNT processes resulted in limited conclusions due to the lack of EHS data. Few LCA studies have been undertaken in nanotechnology (Lekas 2005), and a recent report (Klöpffer et al. 2007) indicates that although LCA can have major benefits and produce useful information, its application and use are restricted due to limited toxicity data. Given that it will be some time until toxicological data become available, risk assessment methods can be used to provide insight to policy makers and other decision makers. This article illustrated the use of MC risk models to study the impact of these uncertainties on long-term manufacturing costs, exposure risks, and inherent trade-offs.

Although some authors recommend no changes in existing policies for nanomaterials (FDA 2007), others believe that the existing regulatory structure is inadequate and that new frameworks should be developed to control these materials (Rejeski 2004). Business strategies for commercialization given the current uncertain regulatory environment remain unclear. Results from the SWNT HiPco MC model suggest that voluntary implementation of higher standards than initially required may be optimal in the sense of minimizing uncertainty of total expected production and liability costs, especially if workplace exposure is expected to induce high-profile long-term health effects. This strategy may be especially advantageous for small start-up companies to avoid future liabilities and damages, including medical expenses, lost income, and lost profit. Further, slower implementation rates of higher standards produce greater uncertainty in long-term costs. These high uncertainties are important to consider in production and technology decision making. More generally, the results of this study underscore three important issues:

  • 1
    Expected costs alone are insufficient for informed decision making.
  • 2
    The best level of standards, overall cost, and optimal voluntary standards are highly dependent on uncertain health effects.
  • 3
    The resultant amount of uncertainty in total costs and exposure can be extreme.

For example, across the four assumption scenarios examined, the possible error between point estimates and probability bounds on the true cost per gram ranged from $51/g to $156/g, or 8% to 32% of the best guess (expected value) estimated cost.

In future work, experimental design methods also could be applied to the MC model to identify which unknown inputs or model parameters translate into the greatest cost and exposure uncertainties, which could provide very useful and important information to influence research and funding foci. Until research progresses on the health risks and safe practices for SWNT production process, these types of models and analyses can provide useful information for private and regulatory decision makers. Other helpful methods to study this type of problem might include multicriteria decision analysis (MCDA), probabilistic analysis, goal programming, desirability functions, and other multiobjective trade-off and uncertainty assessment methods. MCDA recently was discussed as an approach to nanomaterial management of the societal benefits and potential health risks (Linkov et al. 2007), but the discussion did not incorporate uncertainties in the above sense.

Regulatory workplace standards for nanoparticles will affect the pace of innovation and commercial production (Bosso et al. 2006). It is important to note, however, that many industries routinely work with toxic materials safely and that if nanomaterials are found to be hazardous to environmental health or safety, then procedures most likely can be devised to mitigate the risk of undesired exposure. Failure to do so will have effects on public acceptance of any technology, especially one with such revolutionary properties. New strategies for managing the health, safety, and environmental issues related to nanotechnology are starting to be discussed (Greenwood 2007; NSTC 2008), but until more insight can be drawn from toxicological studies, nanotech companies can only follow prescribed best industrial hygiene practices. Results from models of the type presented here allow a better understanding of potential economic and health consequences of different occupational health protection strategies.

Acknowledgments

  1. Top of page
  2. Summary
  3. Introduction
  4. Model Description
  5. Results
  6. Addressing Uncertainty in the Rate of EHS Standards Development
  7. Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window
  8. Conclusions
  9. Acknowledgments
  10. References
  11. About the Authors

This work was supported by National Science Foundation Grants SES-0404114 and EEC-0425826 through the Nanoscale Science and Engineering Center for High-Rate Nanomanufacturing. The authors would like to thank Dr. Michael Ellenbecker, Director of the Toxics Use Reduction Institute and Professor of Work Environment at the University of Massachusetts Lowell, for his input on the EHS assumptions of the model.

Notes
  • 1

    One gram (g) = 10−3 kilograms (kg, SI) ≈ 0.035 ounces (oz).

  • 2

    One cubic foot (ft3) ≈ 0.0283 cubic meters (m3, SI). One meter (m, SI) ≈ 3.28 feet (ft).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Model Description
  5. Results
  6. Addressing Uncertainty in the Rate of EHS Standards Development
  7. Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window
  8. Conclusions
  9. Acknowledgments
  10. References
  11. About the Authors
  • Apostolakis, G. E. 2004. How useful is quantitative risk assessment? Risk Analysis 24(3): 515520.
  • Bell, T. E. 2007. Understanding risk assessment of nanotechnology. http://www.nano.gov/Understanding_Risk_Assessment.pdf. Accessed February 2008.
  • Bosso, C. J., J. A. Isaacs, W. D. Kay, and R. L. Sandler. 2006. Leaving the laboratory: Regulatory and societal issues confronting nanotechnology commercialization. In Nanomanufacturing handbook, edited by A.Busnaina. Boca Raton , FL : CRC Press.
  • Dennison, M. S. and W. Freedman. 2003. Handling toxic tort litigation. American Jurisprudence Trials 57: 395.
  • DOE (Department of Energy). 2007. Nanoscale science research centers approach to nanomaterial ESH. Washington , DC : DOE.
  • Donaldson, K., R. Aitken, L. Tran, V. Stone, R. Duffin, G. Forrest, and A. Alexander. 2006. Carbon nanotubes: A review of their properties in relation to pulmonary toxicology and workplace safety. Toxicological Sciences 92(1): 522.
  • Ellenbecker, M., S.-J. Tsai, and J. A. Isaacs. 2008. Guidelines for the safe handling of nanoparticles in university research laboratories. Submitted for publication to the Journal of Science and Technology for Advanced Materials.
  • EPA (Environmental Protection Agency). 2007. Nanotechnology white paper. 100/B-07/001. Washington , DC : U.S. EPA, EPA Science Policy Council.
  • FDA (Food and Drug Administration). 2007. Nanotechnology task force report. Rockville , MD : FDA.
  • Giordano v. A.C. & S. Inc. 1995. In 666 A2d 710: Pa Super.
  • Greenwood, M. 2007. Thinking big about things small: Creating an effective oversight system for nanotechnology. PEN7. Washington , DC : Project on Emerging Nanotechnology, Woodrow Wilson International Center for Scholars.
  • Hata, K., D. Futaba, K. Mizuno, T. Namai, M. Yumura, and S. Iijima. 2004. Water-assisted highly efficient synthesis of impurity free SWNT. Science 306(5700): 13621364.
  • ICON (International Council on Nanotechnology). 2005. International Council on Nanotechnology. http://icon.rice.edu. Accessed February 2008.
  • Isaacs, J. A., M. L. Healy, and A. Tanwani. 2006. Economic assessment of SWNT production. Paper presented at 2006 IEEE International Symposium on Electronics and the Environment, 8–11 May 2006, San Francisco , CA .
  • Kimura, T. and T. Shinohara. 2006. Monte Carlo analysis of convertible bonds with reset clauses. European Journal of Operational Research 168(2): 301310.
  • Klöpffer, W., M. A. Curran, P. Frankl, R. Heijungs, K. Annette, and S. I. Olsen. 2007. Nanotechnology and life cycle assessment: A systems approach to nanotechnology and the environment. Synthesis of results obtained at a workshop in Washington, DC, 2–3 October 2006. Washington , DC : Project on Emerging Nanotechnologies, Woodrow Wilson International Center for Scholars and the European Commission.
  • Lam, C.-W., J. T. James, R. McCluskey, S. Arepalli, and R. L. Hunter. 2006. A review of carbon nanotube toxicity and assessment of potential occupational and environmental health risks. Critical Reviews in Toxicology 36(3): 189217.
  • Law, A. M. and W. D. Kelton. 1995. Simulation modeling and analysis. New York : McGraw-Hill.
  • Lekas, D. 2005. Analysis of nanotechnology from an industrial ecology perspective part I: Inventory and evaluation of life cycle assessments of nanotechnologies. Revised draft. New Haven , CT : Yale School of Forestry & Environmental Studies .
  • Linkov, I., F. K. Satterstrom, J. Steevens, E. Ferguson, and R. C. Pleus. 2007. Multi-criteria decision analysis and environmental risk assessment for nanomaterials. Journal of Nanoparticle Research 9(4): 543554.
  • Maynard, A. D. 2006. Nanotechnology: A research strategy for addressing risk. PEN3. Washington , DC : Project on Emerging Nanotechnology, Woodrow Wilson International Center for Scholars.
  • Maynard, A. D. 2007. Weighing nanotechnology's risks. International Herald Tribune: 28 March 2007 section Opinion.
  • Maynard, A. D. and E. D. Kuempel. 2005. Airborne nanostructured particles and occupational health. Journal of Nanoparticle Research 7: 587614.
  • Maynard, A. D., R. J. Aitken, T. Butz, V. Colvin, K. Donaldson, G. Oberdörster, M. A. Philbert, J. Ryan, A. Seaton, V. Stone, S. S. Tinkle, L. Tran, N. J. Walker, and D. B. Warheit. 2006. Safe handling of nanotechnology. Nature 444(7117): 267269.
  • Medley, T. and S. Walsh, eds. 2007. Nano risk framework draft document. New York : Environmental Defense–DuPont Nano Partnership.
  • Miller, G. P., S. Okana, and D. Tománek. 2006. Toward uniform nanotubular compounds: Synthetic approach and ab initio calculations. Journal of Chemical Physics 124(12): 121102-1121102-5.
  • NIOSH (National Institute for Occupational Safety and Health). 2004. Safety and health topic: Control banding. http://www.cdc.gov/niosh/topics/ctrlbanding. Accessed February 2008.
  • NIOSH. 2005. Strategic plan for NIOSH nanotechnology research: Filling the knowledge gaps. Draft document. Washington , DC : U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, NIOSH .
  • NIOSH. 2006. Approaches to safe nanotechnology: An information exchange with NIOSH. Version 1.1. Washington , DC : U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, NIOSH .
  • NIOSH. 2007. Progress toward safe nanotechnology in the workplace. Publication No. 2007-123. Washington , DC : U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, NIOSH .
  • NSTC (National Science and Technology Council). 2006. Environmental, health, and safety research needs for engineered nanoscale materials. Washington , DC : National Nanotechnology Initiative, NSTC, Executive Office of the President of the United States.
  • NSTC (National Science and Technology Council). 2008. Strategy for nanotechnology-related environmental, health, and safety research. Washington , DC : National Nanotechnology Initiative, NSTC, Executive Office of the President of the United States.
  • Ok, Z. D., J. C. Benneyan, and J. A. Isaacs. 2007. Modeling production costs for SWNT manufacturing given uncertain health and safety standards. Paper presented at 2007 IEEE International Symposium on Electronics and the Environment, 7–10 May 2007, Orlando , FL .
  • Owens-Corning Fiberglas Corp. v. McKenna. 1999. In 726 So. 2d 361: Fla. Dist. Ct. App. 3d Dist.
  • Pagani, L. P., G. E. Apostolakis, and P. Hejzlar. 2005. The impact of uncertainties on the performance of passive systems. Nuclear Technology 149(2): 129140.
  • Project on Emerging Nanotechnology, Woodrow Wilson Center for Scholars. 2006. Inventory of nanotechnology consumer products. http://www.nanotechproject.org/consumerproducts. Accessed February 2008.
  • Rejeski, D. 2004. The next small thing. Environmental Forum 21(2): 4249.
  • Ross, S. 1998. A first course in probability. 5th ed. Upper Saddle River , NJ : Prentice Hall.
  • Royal Society and Royal Academy of Engineering. 2004. Nanoscience and nanotechnologies: Opportunities and uncertainties. London : Science Policy Section, The Royal Society.
  • Schmidt, K. F. 2007. Green nanotechnology: It's easier than you think. PEN8. Washington , DC : Project on Emerging Nanotechnology, Woodrow Wilson International Center for Scholars.
  • Zhang, G., D. Mann, L. Zhang, A. Javey, Y. Li, E. Yenilmez, Q. Wang, J. P. Mcvittie, Y. Nishi, J. Gibbons, and H. Dai. 2005. Ultra high yield growth of vertical single walled carbon nanotube; Hidden roles of hydrogen and oxygen. PNAS 102(45): 1614116145.

About the Authors

  1. Top of page
  2. Summary
  3. Introduction
  4. Model Description
  5. Results
  6. Addressing Uncertainty in the Rate of EHS Standards Development
  7. Trade-Offs Between Production Costs and Occupational Health Risks and Decision Analysis Window
  8. Conclusions
  9. Acknowledgments
  10. References
  11. About the Authors

Zeynep D. Ok is a doctoral candidate in the Department of Mechanical & Industrial Engineering at Northeastern University in Boston, Massachusetts. James C. Benneyan is the Director of Quality and a senior fellow of the Institute for Health Care Improvement and an associate professor in the Department of Mechanical & Industrial Engineering at Northeastern University. Jacqueline A. Isaacs is an associate director for the Center for High-Rate Nanomanufacturing, in Boston, as well as its Research Thrust Leader for the Societal Implications of Nanotechnology, and an associate professor in the Department of Mechanical & Industrial Engineering at Northeastern University.