Split (1)

train · 91.6k rows

University stringclasses 19 values	Text stringlengths 458 20.7k
Virginia Tech	ramped to about 500°C, and then the chamber is held isothermal for a relatively long period of time to ensure that coal oxidation is complete. The weight change in this phase is thus estimated to be the dry pure coal weight (i.e., non-volatile portion of the coal dust). In the final phase, the calcite is converted to calcium oxide as temperature is ramped to about 750°C, and held isothermal for some time again. The residue (i.e., final weight) at the end of the experiment can be attributed to the calcium oxide and other mineral matter in the coal dust. Taking the weight change of the sample in the in the second and fourth phases of the method as the weight of carbon dioxide released during magnesium carbonate and calcite conversion, a stoichiometric relationship can be used to estimate the weight of these constituents in the sample, and also the weight of their resulting oxides. Finally, the difference between the experiment residue and the estimated oxide weights can be used to estimate the non-carbonate (i.e., inert) mineral weight in the coal. As described, the developed method and determination of dust sample components (i.e., coal, carbonates, and all other mineral matter) does assume that rock dust is comprised completely of carbonates– which is not a practical assumption in many cases. However, the method and calculations can easily be adjusted based on more accurate assumptions. We are currently developing a modified TGA method to allow for determination of magnesium carbonate, which is associated with rock dust products containing significant dolomite. And since rock dust products are often assayed to determine fractions of calcite, dolomite and minor minerals, such information can be used to come up with mine-specific calculations. Regardless, the fundamental TGA work appears very promising in this area. 4. Conclusions and Future Work There are undoubtedly needs for enhanced understanding of respirable dust characteristics in coal mines. TGA provides great potential for very simple determination of coal to total mineral ratios – analogous to proximate analysis of bulk coal samples. Such information could help provide insights into the sources of respirable dusts, and allow an additional basis of comparison between dusts from different mines, different areas of the same mine, or generated under different conditions. As well, TGA may provide opportunities to estimate more specific mass fractions of dust samples – such as the fraction associated with rock dust products. 16
Virginia Tech	The preliminary results and observations presented here reveal that direct-on-filter TGA of samples collected on PVC filter media may provide at least some information about coal vs. mineral fractions of respirable dust. Moreover, such analysis could be easily integrated into current standard methods for gravimetric and silica content analyses. For more accurate estimation of coal and mineral fractions, as well as the rock dust fraction specifically, dust-only TGA appears promising. Continuing research is focused on optimizing the efficiency of dust removal from multiple filter types, and determining detection limits with regards to dust sample weights. Work is also needed to further develop the TGA method(s) appropriate for calcite and dolomite estimation. Complementary analyses to TGA methods for respirable coal mine dusts samples are additionally being considered, such as examining residue from TGA experiments with SEM- EDX to allow identification of specific mineral particles. 5. Acknowledgements We extend our gratitude to the Alpha Foundation for the Improvement of Mine Safety and Health for funding this work. We thank Dr. Charles Potter from TA Instruments for his guidance with the development of a method using TGA. We also thank all mine personnel in Central and Northern Appalachian mines that have provided access and assistance to collect dust samples, and those that have shared their invaluable experience with our research team. 6. References Annual Book of ASTM Standards (ASTM), Vol. 1994, Section 5, American Society of Testing Materials, Philadelphia, 1993, pp. D3172–D3189. Bartley, D. L., and Feldman, R. (1998). NIOSH Manual of Analytical Methods (NMAM), 4th Edition, Particulates Not Otherwise Regulated: Method 0600, Issue 3, 1-6. Castranova, V., and Vallyathan, V. (2000). Silicosis and coal workers' pneumoconiosis. Environmental Health Perspectives, 108(Suppl 4), 675. Centers for Disease Control (CDC) (2006). Advanced Cases of Coal Workers’ Pneumoconiosis- Two Counties, Virginia. MMWR, 55(33). 17
Virginia Tech	Chapter 2. TGA Method for Estimating Coal, Carbonate, and Non-carbonate Mineral Mass Fractions in Respirable Mine Dust Samples Meredith Scaggsa, Emily Sarvera, Cigdem Kelesa aVirginia Tech, Blacksburg, Virginia, USA Abstract Incidence of occupational lung disease amongst coal miners in Central Appalachia has recently increased after decades of decline following the Coal Mine Health and Safety Act of 1969. The causes of this trend have yet to be fully explained – but particular dust characteristics may be a factor. Understanding more about respirable dust components requires analysis beyond what is currently done for demonstrating regulatory compliance. Thermogravimetric analysis (TGA) can provide estimates of coal and non-coal mass fractions of dust samples; and in some cases, the carbonate fraction and hence non-carbonate mineral fraction can also be determined. This paper presents the development and verification of a standard methodology for TGA of respirable coal mine dust samples, including sample preparation by removal of dust from MCE and PVC filter types via sonication. Data corrections related to the TGA instrument drift, sample preparation, and premature carbonate conversion are provided. The method is verified using laboratory-generated dust samples, with known coal, carbonate, and non-carbonate mineral mass fractions. Keywords: Coal Mining, Thermogravimetric Analysis (TGA), Rock Dust, Respirable Dust, Occupational Lung Disease, CWP 1. Introduction Respirable dust has long been recognized as an occupational health hazard for underground coal miners (Seixas et al., 1995). Chronic exposure can result in diseases such as black lung, commonly termed Coal Workers’ Pneumoconiosis (CWP), or silicosis. Between about 1970-2000, regulatory measures coupled with improved technologies and operations management resulted in significant reduction in the incidence of such diseases amongst US coal miners (Suarthana et al., 2013; NIOSH, 1974; WHO, 1999). Between 2000-2010, however, an 21
Virginia Tech	increase in CWP incidence was observed – with incidence rates particularly up central Appalachia (MSHA, 2008; Laney and Attfield, 2010; CDC, 2006; Antao et al., 2005). While the cause for this trend is still not completely clear, changes in dust characteristics have been considered potential factors (e.g., see Pollock et al., 2010; Page and Organiscak, 2000; Bennett et al., 1979; Landen et al., 2011). For instance, mines in the central Appalachian region are tending to mine thinner seams of coal, and thus cut more roof or floor rock (Page and Organiscak, 2002; Laney and Attfield, 2010; CDC, 2006 and 2007; Schatzel, 2009), which may contribute relatively higher amounts of respirable dust constituents of concern (e.g., crystalline silica) (Pollack et al., 2010; Joy, 2012). Some correlation has indeed been shown between increased pneumoconiosis prevalence and increased mass concentrations of respirable silica in mines in this region (vs. other coal mining regions in the US) (Pollock et al., 2010; Joy, 2012) – but additional evidence is needed to draw hard conclusions. Moreover, it is possible that some combination of factors is at play, including those associated with both the dust itself (i.e., what types of particles are contained in the respirable fraction) and with an individual’s exposure patterns (i.e., how frequently is an individual exposed to certain types of dust). It has been noted that central Appalachian mines are often relatively small in size and in personnel numbers (Antao et al., 2005; Laney and Attfield, 2010), which may translate to miners working in different production roles and hence different dust conditions relatively often (Antao et al., 2005; Joy, et al., 2010). In an effort to lower risks of occupational lung disease for US coal miners, the Mine Safety and Health Administration (MSHA) has recently promulgated a new dust rule (effective August 2014). The rule tightens the limit on personal dust exposures from 2.0 to 1.5 mg/m3 on the basis of time-weighted average (TWA) total mass concentration, and stipulates that TWA exposures are now calculated on the basis of an entire work shift rather than 8 hours as was previously the case from Title 30 of the Federal Code of Regulations (2014). It also requires implementation of continuous personal dust monitors (CPDMs) in order to allow miners to keep track of their exposure in (semi) real-time over their work shift, an increase in the frequency of compliance sampling by operators, and bases noncompliance findings on single samples (as opposed to an average of three samples) collected by MSHA (30 CFR Part 75). At the same time that mine operators have been working to further reduce respirable dust and comply with new regulations, they have also been stepping up rock dusting programs (30 22
Virginia Tech	CFR Part 75; Fiscor, 2015). Rock dusting (i.e., coating the roof, ribs, and floor with inert dust such as limestone or dolomite) is required in underground coal mines in order to reduce risk of explosions, and rock dust products are specified on the basis of their particle size and composition. Fine particle size (i.e., less than 200 mesh or 75 µm) and inertness are key for ensuring that the dust functions as desired, and inclusion of crystalline silica must be limited due the respiratory health hazard it poses (30 CFR Part 75). Concerns over the possibility that some products may be out-of-spec for silica content (i.e., as measured on a mass basis) have recently been addressed by Colinet and Listek (2012). They found that, while over 90% of products tested contained less than 4% silica (free or crystalline), nearly 90% contained relatively high proportions of respirable sized dust (i.e., more than 20% of the total particles by volume (MSHA, 2015). With respect to respirable silica, Colinet and Listek concluded that the potential exists, if even remote, for rock dusting to contribute to exposures (2012). Though they did not specifically discuss the contribution of rock dust to the total amount of respirable dust in the mining environment, their results also suggest that this contribution may be sizable in some circumstances. 1.1. Analysis of Respirable Coal Mine Dust At present, routine analysis of respirable dust samples from coal mines can yield two primary metrics: total mass concentration, which is determined gravimetrically, and silica mass fraction, which is determined by infrared absorption spectrophotometry (IR) using the MSHA P7 or NIOSH 7603 method (MSHA, 2014; Schlecht and Key-Schwartz, 2003). These metrics are used for assessing compliance with regulations. However, a more thorough understanding of dust characteristics is clearly needed to shed light on occupational health risks and outcomes, as well as the sources of specific dust constituents. While a variety of analytical techniques might be considered to study coal mine dust, relatively few are compatible with current sampling methods (i.e., collection of dust onto a standard gravimetric or CPDM filter) or requirements (i.e., use of intrinsically safe equipment underground). Thermogravimetric analysis (TGA) is one technique that has potential to be applied to samples gathered using current protocols. TGA involves tracking the weight loss of a sample as it is heated. If specific components in the sample are known to thermally degrade/oxidize at particular temperatures, observations of weight loss can then be related to those components. 23
Virginia Tech	TGA is commonly used for proximate analysis of coal (see ASTM, 1994), which involves heating a sample to remove moisture and subsequently oxidize the organic matter (i.e., coal), and the remaining matter is attributed to ash-forming minerals. Likewise, for dust samples from underground coal mines, TGA can be used to estimate the total coal to total mineral mass fractions, including silica, silicates or other minerals that might be contained within the coal seam or cut from adjacent strata, as well as mineral matter associated with rock dusting activities. Preliminary work has indicated that TGA may allow estimation of the carbonate mass fraction, specifically (see Scaggs et al., 2015). In cases where rock dust products are known to have relatively low non-carbonate fractions and where additional carbonate sources are not prevalent, the carbonate mass fraction may serve as an acceptable surrogate for rock dust. Furthermore, because TGA can be performed after measurement of the total sample mass and determination of the silica fraction requires ashing of the sample anyway, a TGA method could be developed that simply fits between current analytical methods for respirable coal mine dust. Prior work by the authors has focused on identifying critical factors for a successful TGA method and conducting preliminary experiments to observe behavior of expected dust components from coal mines, as well as filter media used for sampling (see Scaggs et al., 2015; Keles et al., 2015). That work has confirmed that the primary dust components of interest (i.e., coal, carbonate minerals, and all other minerals) can be discerned from TGA in an oxidizing atmosphere. Importantly, oxidation of coal and conversion of carbonate minerals to oxides occur in separate temperature regions. However, it was observed that interferences from filter media (PVC and MCE) are problematic for resolving weight losses associated with specific dust components, and so a direct-on-filter TGA method is not advisable. Rather a dust-only approach is needed. Building upon that prior work, the current paper presents a methodology for preparing and analyzing respirable coal mine dust samples thermogravimetrically. The method results in estimations of coal, carbonate, and non-carbonate mineral mass fractions. Verification is presented using laboratory-generated respirable dust samples from pulverized raw coal, rock dust and shale materials. 24
Virginia Tech	2. Method Development The following sections describe experimental procedures used to generate, prepare, and perform TGA on respirable dust samples representative of those that might be collected in coal mines. 2.1. Sample generation and preparation Since TGA is an inherently destructive technique and method development often requires a large number of samples collected under controlled conditions, all dust samples used here were generated in the laboratory. To do this, a sealed enclosure was loaded with dust material, which could be made airborne by use of a small fan or compressed air pulses (Figure 2.1). Escort ELF dust pumps equipped with 10 mm Dorr-Oliver cyclones (i.e., same as those used for compliance sampling in coal mines) were used at a flowrate of 2.0 L/min to sample the respirable-sized particles onto a PVC or MCE filter (5 um pore size) inside a two-piece plastic cassette with filter support pad. At this flow rate, the cyclone cut point is at approximately 10µm. The target sample size was about 500-1,000µg; this was based on a median TWA dust exposure in central Appalachian mines of about 0.70mg/m3 (Suarthana, et al., 2011), which should equate to just under 500µg of dust on 37mm filter for an 8-hour shift. Initial testing of the apparatus indicated that about 100-200µg of sample was collected on a filter for each minute that a pump was running for efficient sample gathering. Dust weights were determined using a microbalance (Sartorius MSE6.6S, Goettengin, Germany). 25
Virginia Tech	Figure 2.1. Schematic of laboratory set-up for respirable dust sample collection The materials used to generate the respirable dust samples were pulverized raw coal, a real rock dust product, and pulverized shale pieces, which were handpicked from a large run-of- mine coal sample collected in central Appalachia (Peerless seam). Due to the fact that none of these are pure materials, it was expected that the coal would contain some small fraction of ash- forming minerals, the rock dust would contain some non-carbonate minerals, and the shale would have some coal content associated with it. For samples used during method development, only the coal and/or rock dust product was collected on each filter, and the total weight of dust collected was determined for all samples. For samples used to verify the method, all three materials were collected on a filter sequentially such that the weight of each material could be determined (i.e., filters were pre-weighed and then weighed after collection of each dust type). As mentioned above, it is necessary to remove dust from the filters prior to TGA. To do this, filters were carefully removed from their cassettes using clean tweezers. Each filter was placed inside a clean 50 mL glass test tube (29 mm diameter). The size of the tube allows for a large enough circumference so that the filter can be slightly rolled, but its surfaces do not overlap obstructing dust removal. The tube was filled with approximately 25 mL of deionized (DI) water. Additional DI water was added to completely submerge the filter if needed. Test tubes were then 26
Virginia Tech	placed in an ultrasonic bath (Cole-Parmer 08895-4, Vernon Hills, Illinois) and sonicated for 1 minute. During preliminary testing, this amount of time was sufficient to recover a significant dust mass, which will be discussed below, without removing excessive filter mass, which may interfere with TGA results. After sonication, the filters were carefully removed from tubes to minimize disturbance of any dust particles in the water. Finally, the test tubes were placed in an oven at 180ºC to evaporate the water, leaving any dust residue in the tube. A volumetric pipette was used to add 200 µL (unless otherwise specified) of reagent grade isopropyl alcohol to each test tube, making an effort to rinse the sides and bottom of the tube to collect as much of the dust residue as possible. The suspension was then removed from the tube using the same pipette tip and deposited into a clean, tared TGA pan. Pans were cleaned prior to each use by exposing to an open flame from a butane torch for at least ten seconds. Pans were tared empty in the TGA instrument (TA Instruments Q500, New Castle, DE) and also weighed using the microbalance. The microbalance data served as a check for TGA instrument measurements. After the alcohol had evaporated (about 30 minutes), each pan was weighed again with the microbalance. 2.2. TGA Routine and Bulk Sample Analysis Typically on the same day that dust samples were deposited into the TGA pans, they were run in the TGA using the routine outlined in Table 2.1. This routine is based on preliminary work by the authors to determine the temperature ranges over which the expected sample components are reactive. Representative thermograms for bulk samples of the pulverized raw coal, rock dust, and pulverized shale are shown in Figure 2.2. These samples contained relatively large particles placed directly into TGA pans rather than collected on dust filters. From these results, it appears that coal oxidation occurs between approximately 360-480°C with the current TGA routine. The weight loss in this region can be used to estimate the coal mass fraction of a sample. The final residue associated with the coal sample shown in Figure 2.2 is non-oxidizable mineral matter, which would be referred to as “ash” in proximate analysis of a bulk coal sample. Considering dust in the respirable size range, particles are expected to be highly liberated such that their mass reports to the correct category. 27
Virginia Tech	Table 2.1. Steps in TGA Routine 1. S elect gas 2, Air 2. R amp 50.00°C/min to 200.00°C 3. Is othermal for 5.00 min 4. R amp 50.00°C/min to 380.00°C 5. Is othermal for 5.00 min 6. R amp 100.00°C/min to 460.00°C 7. Is othermal for 15.00 min 8. R amp 100.00°C/min to 470.00°C 9. Is othermal for 5.00 min 10. R amp 50.00°C/min to 480.00°C 11. Is othermal for 5.00 min 12. R amp 50.00°C/min to 500.00°C 13. Is othermal for 1.00 min 14. R amp 50.00°C/min to 750.00°C 15. Is othermal for 1.00 min 16. C ool in gas 1, Nitrogen for 16 min Figure 2.2. Representative thermograms for the raw coal, rock dust, and shale materials used to develop the TGA method Also evident in Figure 2.2 is the behavior of carbonates in the rock dust sample, which do not begin conversion to oxides until about 480°C – after coal should be completely oxidized. The weight loss between 480°C and the end of the TGA routine can be attributed to carbon dioxide (CO ) release, and accounts for about 40% of the total rock dust sample shown. From this weight 2 loss, the carbonate mass fraction of a sample can be determined stoichiometrically by assuming specific carbonate minerals are present - i.e., the CO molar mass fraction is 44% for pure 2 28
Virginia Tech	calcite, CaCO , and 48% for dolomite, MgCa(CO ) . The residue remaining at 750°C is 3 3 2 attributed to the oxides produced from carbonate conversion, and any other non-oxidizable minerals contained in the rock dust. The non-oxidizable content of the rock dust used here was assumed to be 9% of its total mass, based on X-ray diffraction (XRD) analysis obtained for four different rock dust products used in the central Appalachian region (including some from the same producer as the sample used here). This showed that 91 ± 3% of the product mass is calcite or dolomite, with calcite accounting for roughly 75% of the total mass and dolomite accounting for roughly 17%. Based on the relative proportions of these two minerals from the XRD analysis, CO was assumed to be 45% of the total carbonate mass – and resulting oxides were assumed to 2 be 55%. Using these assumptions, TGA-derived mass fractions for the rock dust sample shown in Figure 2.2 are about 91% carbonate (40.5% CO2 and 50.5% oxides) and 9% non-oxidizable minerals – which are aligned with the XRD-derived fractions. Finally from Figure 2.2, the shale thermogram indicates that this material is largely non- oxidizable – losing only about 10% of its weight. This loss is attributed primarily to impurity of the sample material used here, which likely contained some coal and some carbonate minerals. SEM-EDX data of the respirable sized shale material did confirm that it was primarily alumino- silicates, with some carbonaceous (i.e., coal) and carbonate particles. Because the TGA routine takes over 50 minutes for each sample, the number of samples that can be run consecutively is limited (i.e., because nitrogen and/or air tanks must be replaced). All dust results reported here were obtained from TGA runs that included one empty pan at the beginning of the run, then six consecutive samples, and a final empty pan at the end of the run (i.e., eight total results); empty pan data is necessary for correcting sample results for internal drift of the TGA instrument. After the TGA run, all pans were weighed a final time using the microbalance, as a means of verifying the TGA data. TGA results were analyzed using the Thermal Advantage Universal Analysis software (also from TA Instruments), which allows detailed examination of weight loss and temperature profiles as well as user-defined macros for efficient extraction of specific data (e.g., weight loss values from multiple time ranges). 2.3. Corrections for instrument drift and sample preparation Since respirable dust samples may have very small masses, it is necessary to correct TGA results for the internal drift of the instrument as well as influences of sample preparation. To 29
Virginia Tech	each of these respective temperatures, and the appropriate constants for C should be chosen Drift based on the filter type. M = M - (𝑪 +𝑪 ) (Equation 2.3) Dust 360℃ 𝒅𝒓𝒊𝒇𝒕 𝒇𝒊𝒍𝒕𝒆𝒓 M =(M - M ) –(𝑪 + C ) (Equation 2.4) C 360℃ 480℃ 𝒅𝒓𝒊𝒇𝒕 filter Figure 2.5. Thermograms from respirable coal dust samples collected in the lab on MCE and PVC filters. It can also be seen from Figure 2.5 that the non-oxidizable or non-coal mass fraction of the respirable coal dust sample is similar to that of the bulk coal sample. This suggests that the sample preparation procedure does not bias results toward a particular dust component. Moreover, the fact that significant weight loss is not evident before about 360°C in these samples provides further confirmation that the preparation procedure does not tend to produce excessive residue. Based on Equation 2.2, less than 3 µg of weight loss in the coal region (i.e., 360-480°C) is attributed to the influence of sample preparation for any of these samples. Figure 2.6 shows the TGA results for dust samples generated from rock dust on both filter types. It appears that these samples were slightly contaminated with coal particles (evident from the weight loss behavior that occurs in the coal oxidation region), which is likely due to collection of the samples in the dust enclosure following collection of coal dust. However, the results do confirm that significant weight loss occurs in the temperature range 480-750°C, 33
Virginia Tech	signaling CO release as carbonates are converted to oxides. This weight loss is simply termed 2 the observed CO mass (M ) (Equation 2.5). 2 Obs. CO2 M =M - M (Equation 2.5) Obs. CO2 480℃ 750℃ Figure 2.6. Thermograms from respirable rock dust samples collected in the lab on MCE and PVC filters. The weight loss between 480-750ºC are shown as a percentage of the sample weight at 480ºC. Also from Figure 2.6, it appears that observed CO weight loss is disproportionately 2 correlated to the total sample size (i.e., smaller samples exhibit smaller CO losses than 2 expected). Based on the assumptions that the rock dust product tested contains approximately 91% carbonates by mass, which have a stoichiometric CO content of 45%, the expected CO 2 2 weight loss should be about 40% of the weight value at 480°C. The trend in CO weight loss 2 with sample size suggests that some of the carbonates were converted to oxides prior to TGA analysis – probably during sample preparation, which has been previously been explored (Chou et al., 1989; Pokrovsky, et al., 2005; Plummer et al., 1978; Wollast, 1990). This trend was also observed for the samples containing both coal and rock dust (Figure 2.7). 34
Virginia Tech	Figure 2.7. Thermograms from a mixture of respirable rock dust and coal dust samples collected in the lab on MCE and PVC filters. The weight loss between 480-750°C are shown as a percentage of the sample weight at 480°C. Figure 2.8 shows the correlation between expected and observed carbonate mass for dust samples containing only rock dust and a mixture coal and rock dust (i.e., samples shown in Figures 2.6 and 2.7, respectively); the data fits well to a linear trend line. The observed carbonate mass, (M ), is stoichiometrically determined from the observed CO mass (Equation 2.6). Obs, Carb. 2 The expected carbonate mass, (M ), is determined from the final residue (at 750ºC) that Exp Carb. cannot be attributed to the non-oxidizable fractions of the coal or rock dust product (Equation 2.7). The residue associated with the non-oxidizable fractions is then classified as the non- carbonate mass (M ) for the purposes of predicting the amount of other types of minerals Non-Carb. (e.g. clays, shales, silica) which may be present in the respirable fraction. To predict mass of carbonates in a sample (M ) directly from the TGA data, Equation 2.8 gives the trend line Pre. Carb. shown in Figure 2.8 as a function of observed CO mass. 2 M = MObs. CO2 (Equation 2.6) Obs. Carb. 0.450 M =(0.910M750ºC - 0.056MC ) (Equation 2.7) Exp. Carb. 0.450 M =1.042M +64.483 (Equation 2.8) Pre. Carb. Obs. Carb. 35
Virginia Tech	3000 ) g µ ( 2500 s y = 0.9636x - 67.101 s a M R² = 0.9902 2000 s e t a n 1500 o b r a C 1000 d e v r 500 e s b O 0 0 500 1000 1500 2000 2500 3000 Expected Carbonates Mass (µg) Rock Dust Coal and Rock Dust Figure 2.8. Calibration curve for observed versus expected carbonate mass in samples containing rock dust and coal and rock dust. Figure 2.9 shows the ratio of observed to expected carbonate mass as a function of observed CO weight loss. From Figures 2.8 and 2.9, the trends suggests that when the observed 2 CO weight loss is relatively small (i.e., less than about 30µg), error in the carbonate model 2 prediction may be more significant. From this, the correction for carbonates here is applied to samples with greater than 30µg of CO weight loss in order to contain the variability between 2 predicted and expected carbonates mass fractions. With the CO loss measured from the 2 thermograms, it is possible to determine the minimum amount of carbonates present in the sample, however, this can lead to biasing in the non-carbonates mass fraction. 36
Virginia Tech	3. Method Verification For verification of the above method, a total of 43 composite samples were generated in the lab containing three materials (i.e., coal, rock dust and shale), collected in varying order. The weight of each material on the filter was determined gravimetrically during sample collection, and samples were prepared for TGA using the procedure previously described. Dust recovery is shown in Figure 2.10. Recovery again generally increased with total sample mass on the filter, and was higher for MCE than for PVC media. The sequence of dust collection was not observed to have any real impact on recovery. However, for 12 of the samples, up to 800 µL of isopropyl alcohol (i.e., as opposed to just 200 µL) was used to rinse dust from the test tubes into the TGA pans; and this did seem to increase recovery somewhat. 70% 60% ) 50% % ( y r 40% e v o c e 30% R t s u D 20% 10% 0% 0 500 1000 1500 2000 2500 3000 Total Sample Mass on Filter (µg) PVC Filter -200µL IA MCEFilter -200+µL IA RockDust-Coal-Shale Shale-Coal-Rock Dust PVC Filter -200+µL IA MCEFilter -200µL IA Coal-Rock Dust-Shale Shale-Rock Dust-Coal Figure 2.10. Dust recovery from filters shown by sequence of dust loaded onto the filter, filter type, and amount of isopropyl alcohol used for recovery of dust. 38
Virginia Tech	Figure 2.14 shows the ratio of TGA- to gravimetrically-derived results for coal and non- coal mass fractions for all samples versus the dust mass recovered to the TGA pan from the sample filter. From this plot, it is evident that when only considering coal and non-coal mass fractions, relatively good accuracy can be achieved at even low sample masses (i.e., 50 µg or more). Generally, the TGA-derived mass fractions match the gravimetric values within about ± 25%. However, for determining the mass fractions of carbonates and non-carbonate minerals in addition to coal, there appear to be two key factors affecting the accuracy of the TGA results: observed CO loss and total dust mass recovered to the TGA pan. To illustrate the relationship 2 between these factors, Figure 2.15 shows the ratio of TGA- to gravimetrically-derived mass fractions for coal, carbonates and non-carbonate minerals for all of the lab-generated samples versus the product of observed CO loss and total dust mass. When this product is about 3,500 2 µg2 or more, the accuracy of the TGA results is generally within about ± 25% of the gravimetric values. So from a practical perspective, the carbonate and non-carbonate minerals mass fraction should only be reported for samples exhibiting a sufficiently large product between CO loss and 2 total dust mass. When the product is lower, it is still possible to calculate the carbonate mass fraction, but it is important to note that the computed value represents the minimum amount of carbonates present in the sample. 41
Virginia Tech	4. Conclusions Considering the implications for occupational health, there is a real need to expand the present knowledge about respirable dust in underground coal mines. The dust-only TGA method described here allows for approximation of coal and non-coal dust mass fractions, even in the case where relatively little dust is recovered for analysis. From a health perspective, this method could be a powerful tool for examining the atmosphere in areas of a mine where dust particles differ from conventional categorical notions (i.e., particles are either coal or silica). Indeed, if silica is oftentimes associated with rock, identification of relatively high non-coal mass fractions may provide additional screening capabilities. If sufficient dust can be recovered, the method can also be used to distinguish between carbonate and non-carbonate minerals. With this information, it is possible to help determine the influence of carbonates (i.e. rock dust commonly used to coat the surfaces of the mine) has on the respirable fraction of dust present in the atmosphere. For the safety of the miners, this could help notify operators of rock dust being out of specification for particle size and pose additional risks for accumulation in the respiratory system. As noted, dust recovery for analysis is key for the developed TGA method. While removing more dust from filters may present some challenges related to further premature carbonate loss (i.e., due to the sonication or other factors), additional efforts to recover dust removed from the filters appears promising. MCE filters appear to have some advantage over PVC with respect to dust recovery; although use of PVC filters, which are the standard for respirable quartz mass determination in coal mines, may allow the TGA method to be easily integrated with current analyses. Moreover, the method may well be adaptable for samples collected with CPDMs. Future work should include use of this method to analyze samples collected in the field, and comparison to results from other analysis, such as SEM-EDX. This may help to define how various analytical methods may be used to complement one another. Additionally, such work will establish characteristics specific to regions of mining or localized in different areas of the mines for particular types of dust exposure. With implementation of the new dust rule, testing with CPDM filters should also be completed to understand how TGA may be applied to dust collected on these filters. 43
Virginia Tech	Chapter 3. TGA of Preliminary Set of Field Samples and Recommendations for Future Work 1. Introduction The purpose of the work presented in the preceding chapters was to develop and verify a method for characterization of respirable dust commonly observed in coal mines. Thermogravimetric analysis was implemented in a dust-only approach for approximation of coal and non-coal mass fractions, as well as coal, carbonate, and non-carbonate mass fractions with larger sample sizes. The comprehensive method was developed by means of bulk sample analysis with TGA to identify specific temperature ranges where dusts commonly observed in underground coal mines oxidize or thermally decompose. A dust-only method was applied in order to limit interference from the filter media and gain a more finite understanding of the chemical changes undergone by the dust constituents. During sample preparation, premature loss of CO from carbonates was identified, and a correction was developed to adjust for this loss. 2 Further correction for naturally occurring thermal drift in the TGA instrument was developed. Verification samples were analyzed and the results indicated that this method can provide powerful information about the dust constituents present in the working atmosphere. In this final chapter, results of the TGA on a preliminary set of field samples are presented. 2. Sample Collection A field sampling campaign was undertaken to collect respirable dust samples underground coal mines throughout Appalachia. This impetus was to investigate dust characteristics in regions of Appalachia in relation to increases in the prevalence of occupational lung disease (CWHSP, 2014; CDC 2006 and 2007, dos Santao et al., 2005). Sampling was focused on three regions in Appalachia: northern Appalachia (NA), central Appalachia (CA), and south central Appalachia (SCA), which encompass mines in MSHA Districts 2 and 3, District 4, and District 12, respectively (CMSHD). In the three regions, samples were collected in multiple locations for each mine. The locations were chosen to represent areas of high dust generation and accumulation, or common work areas for miners: the Intake, the Return, the Feeder, and near Production Equipment (e.g., 48
Virginia Tech	Chapter 4. Conclusions and Future Work The purpose of this project was to develop and implement a method with thermogravimetric analysis for further investigation and characterization of respirable coal mine dust. Development of the method involved bulk sample analysis with TGA of materials thought to be commonly found in underground coal mines and determination of temperature ranges where the materials may oxidize or thermally decompose. Preliminary testing and verification of the method was completed with samples generated in the lab in a manner analogous to those collected for compliance purposes. From the method development and verification process, a dust-only approach was determined and a correction for premature decomposition of carbonates was established. A field sampling campaign was established to investigate and identify dust characteristics in distinct regions in Appalachia. Samples were collected in various locations noted for high concentrations of respirable dust along with common work areas for miners. Results from the analysis of a subset of the field samples reveal that while there is no significant difference between mining regions, the location in the mine where the sample was collected is significant and trends may be identified for sampling locations for the concentration of coal and non-coal present in the atmosphere. The collection and analysis of the field samples can provide valuable information and further insight about the characteristics of respirable dust in underground coal mines. Indeed, considerable progress has been accomplished for respirable dust characterization with method development and implementation with field samples, future work is recommended from current results. For the continuation of this work, there several topics that could be further investigated. Possible adjustments could be made to correction for carbonates based on TGA sample size and observed CO weight loss. Analysis of all field samples collected in varying locations in 2 underground mines and by miners should be completed to increase the current understanding of mine location and region dust characteristics. Comparison of mine samples analyzed with TGA along with paired samples analyzed with SEM-EDX should be completed to verify similarities between the methods and trends in the data. Additional exploration of the effects of sonication and TGA with CPDM samples is encouraged with the implementation of this filter for future compliance sampling. After full analysis of the samples gathered in mines and by miner volunteers, mine operators and miners should be informed of high exposure risks found by the 55
Virginia Tech	Abstract Development and Evaluation of a Permeation Plug Release Vessel (PPRV) for the Release of Perfluoromethylcyclohexane (PMCH) in Underground Mine Tracer Gas Studies Edmund Chime Jong The use of sulfur hexafluoride (SF ) as a tracer gas for analyzing underground mine ventilation 6 systems has been practiced for over 30 years. As a result, the methods used to release, sample, and analyze SF are well accepted. Although improvements are still being made to enhance the 6 analysis of this tracer, the overall technique remains largely the same. However, as the complexity and size of underground mine ventilation networks increase, coupled with steadily rising SF background levels, the ability of a single gas to function as a convenient, rapid means 6 of analysis diminishes. The utilization of multiple tracer gases can mitigate these problems by allowing for a more comprehensive evaluation using multi-zone techniques. A well-documented alternative in HVAC studies to SF as a tracer are perfluorocarbon tracers (PFT). Many PFTs 6 exist as volatile liquids at room temperature and pressure. This characteristic prevents a PFT from being released using the same technique as SF . This paper introduces a passive release 6 method for PMCH. Details about the development of the permeation plug release vessel (PPRV) from creating a GC calibration curve for vapor PMCH to the final field evaluation are presented. The following study successfully developed a mine-scale PPRV. The PPRV is designed to passively deploy PMCH vapor at linear. An equation was derived in this study that allows the prediction of the release rate as a function of temperature and plug thickness. Details regarding the development of the PPRV from preliminary laboratory studies to final field evaluations are provided.
Virginia Tech	Chapter 1: Introduction A ventilation systems is an essential element of underground mining. Ventilation provides fresh air to workers, carries harmful gases out of the mine, and prevents the accumulation of hazardous fumes. These systems must be maintained on a regular basis to fulfill these tasks. Ventilation systems are maintained by verifying the operating compliance of air ways, ventilation fans, and ventilation controls. These components ensure that the necessary amount of ventilation is provided to all areas of the mine. Without the cohesive functioning of fans and controls operating at design specifications, the effectiveness of the ventilation system can be adversely affected. Design and placement of new ventilation components as well as the maintenance of existing components are determined from ventilation surveys (Hartman et al. 1997). Mine ventilation surveys involve the collection of air flow data in key areas of the mine. These surveys are designed to check air velocities, air quantities, and pressures. Once complete, the data are used to either create or validate mine ventilation models. These models are used to evaluate the effectiveness of a mine ventilation system, plan for mine expansion, and prepare for future ventilation changes. The model’s degree of accuracy depends on the quality of the survey data. Unfortunately, fully representative data are difficult to achieve due to the intricacies of underground airflow patterns. This problem is especially apparent in areas with complex geometries or in locations that are inaccessible by personnel. For example, when data are gathered in intricate ventilation branches, such as in longwall gobs and bleeders, the sheer number of possible flow paths makes planning a traditional survey very challenging. For surveys of an active mine, the dynamic nature of mines, including geologic conditions, equipment operations, personnel movements, and atmospheric changes, creates sources of error. During an emergency situation, the challenges when using traditional techniques are exponentially increased as accessibility to desired survey areas becomes drastically limited. Surveys assist in the regular maintenance of ventilation systems. Environmental conditions, such as humidity, dust, ground movements, and water influx, stress the integrity of ventilation controls. As metals corrode and barriers degrade, ventilation controls will inevitably fatigue and fail. Visual inspections and regular maintenance are currently the most effective means to 1
Virginia Tech	prevent complete failures. Once a minor fault is discovered, such as a leak, the component can repaired before a failure occurs. However, even with regular inspections, minor leaks and re- circulations can be missed due to the sheer volume of items that must be examined. Tracer gas based surveys provide a viable supplement to traditional means in such cases. Tracer gases have the unique ability to not only traverse inaccessible areas, but to also characterize complex flow patterns and small-scale flows such as leakage. One of the principal requirements of a tracer gas is that it has no significant natural background presence. As a result, if a significant concentration of tracer gas is detected in a location, then the ventilation flow must have a direct path to that location. This property is especially useful in the identification and quantification of leakage. If an emergency situation arises, traditional ventilation surveying techniques are severely limited as information about the ventilation system must be gathered remotely. Tracer gases in this type of situation can not only be readily deployed and analyzed from remote locations, but can also traverse collapsed entries as well as identify damaged ventilation controls. The mining industry has already used tracer gases for this purpose (Grot et al. 1995, Grot and Lagus 1991). Sulfur hexafluoride (SF ) has been used as the mining industry’s standard and sole tracer gas for 6 several decades. SF has been very effective in accomplishing numerous underground tracer gas 6 studies. However, the ability of SF as a lone tracer is beginning to diminish as ventilation 6 systems become more complex. The previous studies were confined to single area, or single zone, deployments of SF . Such deployments are only effective if the area being investigated can 6 be confined to a single flow volume. If information is needed about the interaction between separate, independent flow volumes, then a multi-zone approach must be taken. For example, many single zone releases of SF have analyzed flow paths across gobs. Depending 6 on the release location, these flow paths can only be determined for one direction. A completely separate tracer gas release must be executed in order to determine flow path for the other direction, which is costly and time consuming. The single tracer issue is further compounded by the increasing natural background levels of SF thereby reducing its detection sensitivity. A 6 solution to these problem is the addition of perfluoromethylcyclohexane (PMCH) as a 2
Virginia Tech	Chapter 2: Literature Review 2.1 Tracer Gases Tracer gases have been applied to a large variety of study areas that include commercial heating, ventilation, and air conditioning (HVAC) systems, atmospheric transport models, pollutant dispersion patterns, geological reservoir monitoring, and underground mine ventilation. Tracers are used in these studies to qualitatively or quantitatively characterize various types of suspended, turbulent flows such as atmospheric currents, air pollutant movements, reservoir leakage rates, and ventilation streams in locations where traditional practices are impractical or impossible to implement. In order to perform a flow characterization, a tracer gas, by definition, must be a substance that does not have a significant background presence in the region being investigated. In principle, this definition encompasses a large variety of gaseous compounds such as hydrogen, methane, carbon monoxide, and carbon dioxide. Although these aforementioned substances have been used as tracers in the past, the type and number of desired tracer gas properties have evolved over time. Modern tracers must be gaseous at the desired operating temperature, non-toxic, non-allergenic, chemically inert, odorless, tasteless, non-flammable, non-explosive, easily transported, easily dispersed, easily sampled, quantified with high reliability, measured with repeatability, and economical to deploy (Collins et al. 1965, Grot and Lagus 1991). Sulfur hexafluoride (SF ), 6 halogenated compounds, and perfluorocarbons are examples of such substances, which are also, incidentally, the most commonly used tracer gases. The main advantage of these three tracer groups is afforded by a combination of their trace-level background presence and detectability at extremely low concentrations. The low background presence of these tracers enhances detection sensitivity by limiting interference from background contamination. Therefore, a high signal to noise ratio can be achieved even at trace-level concentrations, such as in the parts per quadrillion (PPQ) by moles. These modern tracers, especially SF and perfluorocarbons, are able to take advantage of this low 6 background presence with enhanced detectability. This characteristic stems from these tracer 4
Virginia Tech	gases’ high molecular electronegativity. Electronegativity allows detection and quantification using a gas chromatograph (GC) equipped with an electron capture detector (ECD). Tracer gases can also be detected with GCs outfitted with a mass spectrometer (MS) or any other detector based on electron capture (EC) such as a negative ionization chemical ionization mass spectrometer (NCI-MS). Although other detection methods are available, GC-ECD and NCI-MS provide the highest sensitivity for electronegative substances. A detailed summary of GC analytical systems with respect to tracer gas analysis is provided in a later section. The GC-ECD is the most common analytical system due to repeatable detection limits in the parts per trillion (PPT) coupled with lower operating costs when compared to the GC-MS. This cost difference is derived from the fundamental operational differences between the two analytical platforms. The GC-ECD system, as opposed to the GC-MS, is designed to be quantitative but not qualitative. If a positive identification of the tracer gas is required, then the GC-MS system is preferred. The added advantage of highly sensitive detectability with modern tracers enables the quantification of small-scale flows such as leakage as well as of large-scale ventilation systems. Additionally, less tracer gas is required to achieve the minimum detection threshold, which in turn facilitates greater usage efficiency and lower operating cost for tracer deployments. Trace- level detectability combined with the desired tracer properties provide these gases with excellent versatility as both a qualitative and quantitative flow analysis tool. This versatility is especially demonstrated by tracer gas investigations of ventilation systems. Although an assortment of tracer gas survey techniques are available for surveying ventilation systems, the overall procedure can be generalized into three main parts: release, sampling, and analysis. The execution of each individual part, however, differ depending on the type of tracer gas study being performed. Two main types of tracer gas studies, continuous or pulse, are commonly performed. The continuous release method will be discussed first. One of the benefits of using many tracer gases is that they exist as a gas throughout the desired operating conditions. Thus, a variety of release techniques may be used to deploy gaseous tracers 5
Virginia Tech	for the continuous release method. However, certain tracers, such as perfluorocarbons, can exist initially as a volatile liquid at normal temperature and pressure (NTP). As such, a specialized release systems must be utilized. Volatile liquid tracer release systems will be covered in detail in later sections. Gaseous tracers, alternatively, can be purchased in convenient, readily available, and standardized compressed gas cylinders. From this medium, many options are available for a controlled release. Three of the more popular release options are the differential weight method, the flow meter method, and the flow controller method. The differential weight method is executed simply by weighing the gas cylinder before and after a release and noting the time allotted for the discharge. Thus, a release weight per unit time for the tracer can be computed. Although crude, this method can be effective if care is taken to provide a constant, continuous flow rate from the cylinder. The flow meter method utilizes an analog or a digital flow meter attached to the gas cylinder. The tracer gas is then released at a pre-determined volumetric flow rate, which can be translated to a mass flow rate if necessary. This method offers significantly more control and flexibility than the differential mass method but requires the acquisition of a suitable flow meter. The third method is the flow controller method. Flow controllers are sophisticated electronic regulators that allows for precise deployment of tracer gases either by mass or by volume. This method provides the greatest accuracy, reproducibility, and control of the three common release approaches. However, a flow controller can be costly depending on the desired features, such as accuracy, data storage, etc. For continuous tracer releases in the field, the manner in which these release methods are utilized can vary. Three main versions of the continuous release method are commonly employed. These versions are the constant injection method, the constant concentration method, and the tracer decay method (McWilliams 2002, Grot and Lagus 1991). All versions of continuous releases, as the name implies, deploy tracer gases based on a continuous flow rate. The fundamental mathematical relationship between tracer gas, concentration, and ventilation flow rate for continuous releases is derived from instantaneous mass balances. For the constant injection method in a single contiguous zone, tracer gas is released into the test area at either a constant mass flow or a constant volumetric flow rate. The tracer is then allowed to equilibrate with the ventilation flow thus producing a mathematical 6
Virginia Tech	relationship between tracer concentration and single zone air flow rate. This relationship is represented in Equation ( 2.1 ). This equation assumes that the additional volumetric flow of the tracer as well as the loss of tracer flow due to leakage is negligible. The air flow is also assumed to be fully turbulent to ensure complete, homogeneous mixing of the tracer at the sampling point. The complete mixing of the tracer is essential for flow quantitation if an advanced ventilation model is not available for comparison. The concentration of the tracer gas by volume is represented in the following manner. 𝑄 𝑀 𝑇 𝑇 𝐶 = = ( 2.1 ) 𝑇 𝑄 𝜌 𝑄 𝐴 𝑇 𝐴 𝐶 Concentration of tracer gas downstream from release point as a fraction by volume 𝑇 𝑄 Volumetric flow of tracer gas (m3/s) 𝑇 𝑄 Volumetric flow of air (m3/s) 𝐴 𝑀 Mass flow of tracer gas (kg/s) 𝑇 𝜌 Density of tracer gas at ambient environmental conditions (kg/m3) 𝑇 Equation ( 2.1 ) can be rearranged as follows to determine the air flow quantity through the zone as a function of tracer gas release rate and concentration. 𝑄 𝑀 𝑇 𝑇 𝑄 = = ( 2.2 ) 𝐴 𝐶 𝜌 𝐶 𝑇 𝑇 𝑇 Equation ( 2.2 ) is valid only if the inflow air quantity to the zone is equal to the outflow air quantity at the sampling point. If the sampling area is located in a discontinuous branch that is not reconciled at the survey zone, then the ventilation characteristics of the individual branches must be known to conduct a proper quantitative analysis. Without prior knowledge of the individual branch ventilation characteristics, flow quantitation is not possible using Equation ( 2.2 ). As a result, the study would be restricted to a qualitative characterization. Qualitative studies, nevertheless, are still useful in their ability to identify flow paths, leakage, and damage to ventilation controls. However, the extent of flows such as short- circuiting cannot be positively identified without quantitation. 7
Virginia Tech	The constant concentration method is the most complex option for the continuous release technique. This method releases tracer gas in a manner that maintains a constant tracer concentration in the target flow volume. In order to achieve this result, the tracer concentration in the zone is continuously monitored. The injection rate of the tracer is then adjusted electronically to maintain a continuous tracer concentration in the ventilation stream. Thus, a mathematical relationship between tracer injection rate and ventilation flow rate is produced. This relationship is given by the following equation. 𝑄 (𝑡) 𝑀 (𝑡) 𝑇 𝑇 𝑄 (𝑡) = = ( 2.3 ) 𝐴 𝐶 𝜌 𝐶 𝑡𝑎𝑟𝑔𝑒𝑡 𝑇 𝑡𝑎𝑟𝑔𝑒𝑡 Target steady state concentration of tracer gas at the monitor point 𝐶 𝑡𝑎𝑟𝑔𝑒𝑡 as a fraction by volume 𝑄 (𝑡) Volumetric flow of tracer gas (m3/s) over time 𝑇 𝑄 (𝑡) Volumetric flow of air (m3/s) over time 𝐴 𝑀 (𝑡) Mass flow of tracer gas (kg/s) over time 𝑇 𝜌 Density of tracer gas at ambient environmental conditions (kg/m3) 𝑇 The tracer decay method is the most widely implemented technique used in building ventilation studies due to its simplicity. Tracer gas is released into a target volume for a finite amount of time to achieve a desired tracer concentration. After the release period has elapsed, the ventilation flow is monitored over time to determine the decay of the tracer concentration. A relationship can thus be established between tracer decay and ventilation flow. This relationship is given in the following equation. 1 𝐶 𝑇 𝐼 = ln( 0) ( 2.4 ) 𝑡 𝐶 𝑇1 𝐼 Air change rate in zone (changes per second) 𝑡 Monitor time (s) 𝐶 Initial tracer concentration as a fraction by volume 𝑇0 𝐶 Final tracer concentration as a fraction by volume 𝑇1 The air change rate is defined as the volume normalized ventilation rate, which is represented by the following equation. 8
Virginia Tech	𝑄 𝐼 = 𝐴 ( 2.5 ) 𝑉 𝐼 Air change rate in zone (changes per second) 𝑄 Air flow rate in the zone (m3/s) 𝐴 𝑉 Volume of the zone The previous equation shows that the air change rate is the number of total air volume replacements per unit time that the zone undergoes. This technique is particularly useful in determining the time necessary to dilute a zone if a contaminant is released. The previous overview of the continuous release technique demonstrates its one great advantage, an overall ease of execution. Planning and post-processing computations are straightforward and procedurally compact thus reducing occurrences of systematic errors. The non-reactive and gaseous nature of the three tracers allow a variety of both passive and active sampling techniques to be used. The most common passive sampling method is the capillary adsorption tubes (CAT). A CAT is a passive sampler used in tracer gas ventilation studies. The most common form of CAT sampler, or CATS, is composed of a glass tube filled with a measured amount of adsorbent, such as activated charcoal (Winberry et al. 1990). Once exposed to atmosphere, the adsorbent will sample the air at a constant rate thus producing self-integrating concentration over the exposure time period. The sensitivity of CATS can be improved if needed with the addition of air pumps and gas filters. Once adsorbed, the air sample will remain in the adsorbent until the CATS is thermally desorbed to a GC (Dietz et al. 1986, D'Ottavio, Goodrich, and Dietz 1986). The most common active sampling methods are gastight syringes, sampling bags, and evacuated containers. The tracer sampling process is further simplified due to the absence of a time dependent sampling interval. This simplification is a result of the fact that the continuous release technique is based on a consistent and homogeneous tracer concentration. Thus, as long as sufficient time is allotted for uniform mixing of the tracer in the ventilation stream, the actual instant and interval at which tracer samples are taken are inconsequential. The primary disadvantage to the continuous release technique is the sheer amount of tracer that is needed for a large-scale ventilation study. Depending on the size of the ventilation system, such a study may become cost prohibitive. Additionally, the elevated size of a large-scale 9
Virginia Tech	continuous release dramatically increases the ambient background of the tracer. As a result, more time will be needed to reduce the tracer’s background presence to an acceptable level if other surveys are desired. These issues are especially prevalent in tracer gas studies of underground mine ventilation systems. As a result, certain situations may warrant the use of an alternative method, such as the pulse release technique. The pulse release technique discharges a known quantity of tracer gas as an “instantaneous” cloud that is allowed to travel through the ventilation system. Consider the analogy of a dye filled balloon submerged in a flowing stream. The balloon is popped at some point in time thus releasing the red dye as a cohesive plume in the stream. Given that the stream has sufficient turbulence, the cloud of dye will uniformly mix across the width of the stream thus producing homogeneous pulse of dye across the cross-section of the stream. This homogeneity across the cross-section of the flow volume is a representation of ideal pulse release tracer behavior. As can be inferred from the aforementioned example, a pulse release requires an integrated mass balance for quantitation. This mass balance is effectively demonstrated with a single zone pulse tracer study. For a single zone pulse release, the downstream concentration of tracer gas must be computed over time. From the continuous release technique, the following relationships were established. 𝑄 𝑔 𝐶 = ( 2.6 ) 𝑄 𝑀 𝑔 𝑄 = ( 2.7 ) 𝑔 𝜌 𝑔 𝐶 Concentration of tracer gas downstream from release point as a fraction by volume 𝑄 Volumetric flow of tracer gas (m3/s) 𝑔 𝑄 Volumetric flow of air (m3/s) 𝑀 Mass flow of tracer gas (kg/s) 𝑔 𝜌 Density of tracer gas at ambient environmental conditions (kg/m3) 𝑔 10
Virginia Tech	Additionally, the total released mass of tracer gas can be represented as follows assuming no losses in the system thus asserting that the mass of the tracer injected into to the zone is equal to the mass of the tracer exiting the zone. 𝑇 𝑀 = ∫𝑀 𝑑𝑡 ( 2.9 ) 𝑔 0 Equation ( 2.8 ) can be substituted into the relationship into Equation ( 2.9 ) give the following. 𝑀 𝐴 = ( 2.10 ) 𝑔 𝜌 𝑄 𝑔 Rearranging Equation ( 2.10 ), the following relationship can be established. 𝑀 𝑄 = ( 2.11 ) 𝜌 𝐴 𝑔 𝑔 Equation ( 2.11 ) can now be used to compute the airflow quantity in the target volume using the recorded concentration over time. Identically to the continuous concentration method, if the sampling locations are located in discontinuous branches, the volumetric flows of the individual branches must be known prior to post-processing so that the sampled concentration of the tracer can be used to establish unexpected ventilation conditions (Hartman et al. 1997, Grot and Lagus 1991, Persily and Axley 1990). The data required to perform this integration is gathered from continuous air sampling of the zone. In order to acquire a sufficient amount of data points, unlike the continuous release technique, the sample must capture the mass change of the pulse over a period of time. As a result, traditional active sampling methods would require multiple individual samples to be taken for the duration of the pulse. Enough data points would need to be sampled to compute an integrated mass. Although the mediums in which the collection of air samples occur are identical to those used in continuous releases, pulse releases require a sampling time and sampling interval to be defined. This added complexity is necessary to capture the mass distribution of the plume 12
Virginia Tech	as it travels through the sampling area. This traditional real-time sampling technique is, however, very cumbersome and difficult to implement. The concentration integral can alternatively be determined using the average tracer gas concentration at the sampling point (Persily and Axley 1990). The average concentration in this sense could be captured through a continuous extraction process. This process would require a container of appropriate size and a means to continually transfer air into the container at a constant rate. The sampling apparatus would be activated before the pulse release and sample until the tracer is completely purged from the flow volume. The concentration integral could then be determined by multiplying the average tracer concentration in the container by the total sampling time. A CATS could also be utilized for the same purpose. As can be seen by the aforementioned overview, the pulse tracer technique has some advantages and disadvantages over the continuous tracer technique. The main advantages are afforded by lower tracer gas quantity requirements and increased repeatability. The nature of the technique allows the deployment of multiple, consecutive pulses to occur over a short period of time thus allowing replication. As a result, an estimate of repeatability and reliability of the study can be calculated (Persily and Axley 1990, Dietz et al. 1986). The summary of the continuous release technique and the pulse release technique given in the previous section presents analysis techniques for simple, single zone applications of a single tracer gas. Tracer gases can, however, be utilized for multi-zone, inter-zone, and multi-path studies by using multiple, simultaneous tracer gas releases, differential equation based flow models, or a combination of the two. Multiple, simultaneous tracer gas studies will be covered in a later section. An in-depth discussion of the associated flow models will not be provided as this subject matter is beyond the scope of this cursory introduction as well as beyond the scope of the study presented in this discussion. 13
Virginia Tech	2.2 Gas Chromatography Gas chromatography (GC) is the primer technique in the trace analysis of volatile compounds. As such, GC is the primary means by which tracer gas samples are evaluated. The overwhelming implementation of this technique in tracer gas research is derived from the ability of GC to separate and analyze multiple volatile compounds in a single sample while simultaneously maintaining exceptional analytical precision. This three-part process occurs in a standalone device called the gas chromatograph (GC). A GC is composed of three primary components, the injector, the column, and the detector. After an air sample is collected, it must be introduced to the GC for analysis. The injector is the portion of the GC where this occurs. This component is responsible for two principal tasks, volatilization and dilution of the sample. As the name implies, a GC can only process samples that are in either a gas or vapor state. In order to ensure that all components of the air sample are in the appropriate state, the injector is heated to a temperature sufficient to completely volatilize all compounds. Once vaporization is complete, the sample is then reduced to a level that allows rapid transfer to the column. The degree of reduction is tailored to the analytical method being utilized at the time. The GC method will change depending on the properties of the analytes. After the injector has processed the sample, the analytes are transferred to the column. The GC column is the medium in which the various constituents of the sample are separated prior to entering the detector. Separation allows the individual analysis of the multiple tracer compounds that may be present in the sample while removing interference from undesired components. Without the column, the analytes would enter the detector as an overlapping collection of compounds thereby nullifying any possible analysis of the individual analytes. In tracer gas studies, a capillary type column is most commonly used. The outside wall of a capillary column is comprised of a cylindrical fused silica shell. The inside wall of the column is coated with a thin film called the stationary phase. The stationary phase is composed of either a high boiling point liquid or a layer of solid adsorbent particles fused to the silica wall. The phase is applied to the interior of the column in a 14
Virginia Tech	manner that allows unobstructed passage of the mobile phase through the center, which essentially makes the column a long, thin hollow tube. The mobile phase in terms of GC is a gas, usually helium, hydrogen, or nitrogen, which carries the sample from the injector, through the column, and into the detector. As the analytes move through the column, they will partition in and out of the stationary phase depending on their individual affinities for the constituents of the phase. The varying rates of partitioning by each individual analyte causes the compounds to separate. Once the analytes have been adequately separated, the compounds then move into the detector (McNair and Miller 2009). A multitude of detectors can be utilized by a GC. The two most popular detectors for tracer gas analysis are the ECD and the MS. The ECD utilizes a Ni63 radioactive source to create an electron cloud inside the detector. This electron cloud is generated when the beta particles emitted by the Ni63 impact a continuous stream of nitrogen gas. Once the electron cloud is produced, a steady electrical response is outputted when a current is applied to the detector. As an electronegative substance pass through the cloud, some of the electrons will be captured by these analytes thus causing a decrease in the electrical signal. The magnitude of the signal attenuation is proportional to the compound’s concentration. The GC is able to quantify this reduction in signal with extreme accuracy (McNair and Miller 2009). The majority of common tracer gas compounds are electronegative in nature. As such, an ECD is one of the preferred detectors in tracer gas studies. Although not as popular, the MS is also utilized to a fair degree due to its ability to positively identify analytes. Unlike an ECD, the MS is able to generate a mass spectrum of each analyte that enters the detector. The actual process by which the spectrum is produced is beyond the purpose of this overview and will not be discussed. The mass spectrum of a pure compound is unique to that analyte. Thus positive identification of fully separated analytes can be achieved, an ability that is generally not required in tracer studies due to the simplistic composition of the sample matrix. A MS also provides an electrical response proportional to the concentration of the analyte thus also allowing for quantitation. The MS additionally has the potential to achieve detection sensitivities beyond the ECD through the use of selected ion monitoring (SIM) and negative ion chemical ionization (NCI). Although more versatile from an analytical standpoint, the significantly higher 15
Virginia Tech	relative operating costs and overall operating difficulty of the MS compared to the ECD prevents the MS from being more significantly utilized. 2.3 Tracer Gases in Underground Mine Ventilation The versatility of tracer gases as ventilation survey tools have resulted in their widespread utilization in a variety industries including underground mine ventilation. Tracer gases have the unique ability to probe areas of a mine that are either difficult to survey using conventional means or inaccessible to personnel. The overwhelming majority of these underground tracer gas studies have implemented SF as the sole tracer compound. 6 SF was originally produced in 1953 as an insulating gas for high voltage transformers and 6 industrial circuit breakers (Turk et al. 1968). The physical properties that allowed SF to serve as 6 an ideal insulator naturally translated to the desired properties of an ideal tracer. SF is a stable, 6 inert, inorganic, colorless, odorless, and non-toxic gas with a boiling point of -64C (-58F) at 100 Pa (Lester and Greenberg 1950, Collins et al. 1965). SF has a vapor pressure of 290 Pa at 6 21.1C and a molecular weight of 146.055 g/mol. SF is an anthropogenic compound and does 6 not naturally exist in significant concentrations in the atmosphere. SF is also readily available in 6 multiple standardized gas cylinders, inexpensive, and readily deployable through various mediums. SF also has a high ECD detection sensitivity resulting from its elevated electron 6 absorption coefficient (Collins et al. 1965). This ECD sensitivity, near 1 PPB by moles, is granted by the high electronegativity of the six fluorine atoms surrounding the sulfur atom in the center (Turk et al. 1968). A graphical representation of SF is displayed in Figure 2.2. 6 16
Virginia Tech	Figure 2.2. Sulfur hexafluoride molecule. The detection sensitivity can be further enhanced using catalytic reaction or cold concentration GC techniques (Turk et al. 1968). Although the release of SF is limited to a maximum of 100 6 PPM by volume, which is set by the American Conference of Governmental Industrial Hygienists (ACGIH), its detection sensitivity is sufficient to maintain its effectiveness as a tracer gas (Timko and Derick 1989). The main disadvantage of SF can be attributed to its high 6 molecular weight. SF is approximately five times heavier than air by molecular weight. This characteristic presents 6 two challenges. These challenges are insufficient homogeneous mixing of SF in non-turbulent 6 conditions and longer background retention of SF in complex geological areas. However, these 6 challenges can be remediated with sufficient advance planning. Despite these issues, the physical properties of SF classify this compound as an ideal tracer gas. As such, SF has seen wide 6 6 utilization in various disciplines. Meteorological tracing and long-distance atmospheric pollutant transport studies have extensively used SF as a tracer gas due to its high stability at altitudes below the mesosphere, 6 the boundary of which is located about 50 km (30 mi) above the earth’s surface (Clemons, Coleman, and Saltzman 1968). As an atmospheric tracer, SF has been used in short-term studies 6 to validate complex atmospheric transport circulation models and in long-term studies to 17
Virginia Tech	investigate atmospheric mixing between the stratosphere, hydrosphere, cryosphere, and troposphere (Maiss et al. 1996). In the field of pollutant transport, SF has been used to track 6 pollutant dispersion from an emission stack at the Keystone Power Plant in Shelocta, PA using a GC continuous analysis system configured for frontal chromatography. Frontal chromatography is a term used to describe a continuous tracer analysis technique that delays the passage of oxygen, nitrogen, and other such constituents to all the separation of SF (Brown, Dietz, and 6 Cote 1975). In related studies SF was used to determine sulfur dioxide (SO ) plume migration 6 2 over the surrounding topography from various power plant stacks found at the Fawley Power station in Southern England. SO plume migration for some of these studies was tracked using a 2 GC-ECD built to apply nitric oxide to the molecular sieve column. This technique was shown to potentially track air mass movement for distances greater than 100 km (Dietz and Cote 1973, Emberlin 1981). In addition, SF has been used to conduct tracer gas studies of various 6 ventilation systems in buildings. Building ventilation investigations have used SF to determine the residence time of 6 contaminants and the exhaust and post-exhaust reentry of contaminants in laboratories equipped with fume hoods (Drivas, Simmonds, and Shair 1972). SF has also been used to evaluate the 6 HVAC circulation profile of a seven-story federal office building (Grot et al. 1991) as well as to characterize ventilation flow rates in animal housing (Leonard, Feddes, and McQuitty 1984). The successful application of SF to atmospheric, pollutant transport, and building ventilation studies 6 demonstrates the flexibility of this substance as a multifaceted tracer gas. As such, this tracer has seen significant implementation in underground mine ventilation. SF has been used to determine the amount of return air recirculation from leakage between the 6 intake and return (Kissell 1982), to quantify ventilation at coal faces (Timko and Thimons 1982), to evaluate auxiliary fans, to quantify low velocity flows, to measure flow rates in large entries, to determine the location of stopping leakage (Thimons, Bielicki, and Kissell 1974, Topuz, Bhamidipati, and Bartkoski 1982), to investigate shaft flows, to establish in-situ fan operating points, to evaluate auxiliary ventilation efficiency, to trace dust transportation paths, and to test escape route integrity with respect to fire contamination (Hardcastle et al. 1992). The unique ability of tracer gases to traverse areas that are either hazardous or inaccessible have simplified 18
Virginia Tech	the survey of gob, bleeder, collapsed areas, and the mine overall during an emergency situation. Studies of this type using SF have included the determination of flow paths in longwall gobs to 6 optimize methane drainage techniques (Diamond et al. 1999), the evaluation of longwall bleeder ventilation and gob gas vent-hole methane control systems (Mucho et al. 2000), the assessment of gob ventilation effectiveness and mine seal integrity (Timko, Kissell, and Thimons 1987), the determination of airflow paths and air velocities through gobs in room and pillar bleeder systems (Urosek and Watkins 1995), the quantification of gob porosity (Timko and Thimons 1982), as well as the identification of the neutral ventilation point in a collapsed tailgate gob (Young, Bonnell, and Genter 2001). The flexibility of SF has also allowed its application to more novel 6 underground tracer gas examinations. Studies of this nature include using SF to quantify the migration of radon from an East German 6 uranium mine in Schlema, to determine the effect of fan-depressurization on radon levels in surface facilities (Raatschen, Grot, and Lobner 1995), as well as to simulate the progression of fire through intake and belt entries in a limited-entry mine (Timko and Derick 1989). SF has 6 also been applied to characterize airflow routes, identify leakage, and determine turbulent diffusion coefficients in the Pongkor cut and fill gold mine in Indonesia in an effort to verify a one dimensional Gaussian distribution ventilation model (Widodo et al. 2008). As can be seen by the aforementioned summary, SF has a variety of uses in surveying mine 6 ventilation systems. However, the increasing complexity and scale of mines has caused a decrease in the capacity of SF to function as a lone tracer. For example, although re-circulation 6 and leakage have been analyzed using SF , the accuracy of such studies can be in question due to 6 unknown discontinuities in flow paths. Along this lines, a single tracer can additionally not be used to fully determine interactions between multiple independent flow paths, a common facet of modern underground mines. The capacity of SF as a tracer in terms of sensitivity is also 6 beginning to diminish due to increasing atmospheric background levels. Tracer gases rely on low background levels for their basic purpose, detectability against ambient background. In order for a tracer to be detectable, its released concentration must be higher than its background for an adequate signal to noise ratio. A higher tracer background presence 19
Virginia Tech	translates to a more concentrated deployment to achieve a detectable concentration. This effect is especially prevalent in high velocity flows or in flows through large openings. High tracer background may also prevent the detection of low quantity flows such as leakage. The problem of increasing background has especially affected SF (Cooke et al. 2001). 6 Since the initial industrial production of SF in 1953, its atmospheric concentration has increased 6 by two orders of magnitude. This increase is a result of not only its popularity as a tracer gas but its widespread application in gas-insulated electrical systems and in molten reactive metal degassing, such as with magnesium and aluminum, as well (Maiss and Brenninkmeijer 1998, Levin et al. 2010). This problem is exacerbated by the fact that SF is only destroyed in the 6 mesosphere thus yielding an atmospheric lifetime of between 800 – 3,200 years depending on environmental conditions (Ravishankara et al. 1993, Geller et al. 1997). In order to increase the accuracy of multi-zone underground ventilation studies while simultaneously remediating the problem of increasing SF background, an alternative tracer gas capable of both independent and 6 concurrent usage with SF can be implemented. One such alternative is presented by 6 perfluoromethylcyclohexane (PMCH). 2.4 Properties of Perfluoromethylcyclohexane (PMCH) PMCH, along with other similar perfluorochemical (PFC) compounds, was originally synthesized as a part of the Manhattan Project during World War II. The purpose of sub-project was to find a compound that was corrosion resistant to uranium hexafluoride, a substance used for uranium enrichment (Lowe 2002). The physical properties of PFC compounds has since allowed their application in other fields. PMCH, specifically, is comprised of seven carbon atoms and fourteen fluorine atoms (C F1 ). Due to its chemical structure, PMCH is classified as a 7 4 cyclic perfluoroalkane. Perfluoroalkanes are molecules that have a hydrocarbon structure in which all the hydrogen atoms are replaced by fluorine atoms. The substitution of fluorine atoms for hydrogen atoms increases the molecular mass. Thus, PMCH is significantly heavier than comparable hydrocarbon oils with a specific gravity about 1.787. As a result, this highly fluorinated molecule has a high molecular mass of 350 g/mol. 20
Virginia Tech	The empirical formula of PMCH reflects the relationship C F . Due to the fact that PMCH is a n 2n+2 cyclic perfluoroalkane, two fluorine atoms are removed from the perfluorinated chain to form the cyclic structure. Therefore, cyclic perfluoroalkanes will have a slightly modified empirical formula of C F (Sandford 2003). A graphical representation of PMCH and its cyclic structure 7 2n can be seen in Figure 2.3. Figure 2.3. Perfluoromethylcyclohexane (PMCH) molecule. As can be seen in the previous figure, PMCH is solely composed of carbon-fluorine single covalent bonds. Chemical bonds of this type are exceptionally strong by nature with typical bond disassociation energies of 452 kJ/mol to 531 kJ/mol. PMCH, as a result, is very stable with an estimated atmospheric lifetime of greater than 2,000 years. In contrast to its robust chemical bonds, PMCH exhibits very low intermolecular forces. This weak attraction between PMCH molecules yields systems with high volatility coupled with low boiling points relative to other compounds with similar molecular weights. The volatility of PMCH, which is reflected by a vapor pressure of 14,100 Pa at 25°C, allows it to rapidly vaporize even at low temperatures (Sandford 2003, Lowe 2002). Once in a vapor state, PMCH will not condense even through cooler temperatures despite a critical temperature of 212.8°C. Although highly volatile, the 76°C (169°F) boiling point of PMCH causes this compound to exist as a liquid at room temperature and pressure (National Institute of Standards and Technology 2011, Cooke et al. 2001). 21
Virginia Tech	PMCH, similarly to other perfluoroalkanes, is inert, colorless, odorless, non-reactive, and non- toxic (Dietz 1986, Sandford 2003). The toxicity of perfluoroalkanes to humans is in fact so low that these compounds have been used in intravascular oxygenation of cells as well as to supply oxygen to the respiratory system through “liquid breathing” apparatuses to infants (Lowe 2002). The previously introduced characteristics of PMCH and of perfluoroalkanes in general have allowed these compounds to be successfully applied in numerous fields of study including biotechnology and medical research. The same properties also ideally suit PMCH for use in ventilation research as a tracer gas. 2.5 PMCH as a Tracer PMCH is a member of the perfluorocarbon tracer (PFT) group of compounds. Other examples of PFT group compounds are, but not limited to, perfluorodimethylcyclobutane (PDCB or PMCB), perfluoromethylcyclopentane (PMCP), perfluorotrimethylcyclohexane (PTCH), and perfluorotrans 1,4 dimethylcyclohexane (ptPDCH). PFT group compounds are used as tracer gases due to their physical properties. PMCH is an effective tracer gas because it is biologically inert, chemically inert, and is exceptionally thermally as well as environmentally stable in vapor form. Additionally, PMCH does not exist naturally in the environment due to its anthropogenic production. Using NCI-GCMS, the background presence of PMCH in the atmosphere was found to be 4.6  0.8 femtoliters per liter (fL/L). This concentration falls in the sub parts per quadrillion (PPQ), 10- 15, range by volume. This level is approximately three orders of magnitude lower than SF and 6 easily classifies PMCH as having no significant natural background presence (Cooke et al. 2001, Dietz et al. 1998, Simmonds et al. 2002, Straume et al. 1998, Watson et al. 2007). As a result, a high signal to noise ratio can be achieved even at exceptionally low concentrations. The resulting low detectability threshold is the greatest advantage afforded by PMCH as a tracer gas. PFC compounds react with free electrons, a property that increases ECD sensitivity. Free electron reactivity generally increases as the number of fluorine atoms in a molecule increases (Lovelock and Ferber 1982). The large grouping of fluorine atoms in PMCH creates molecular 22
Virginia Tech	electronegativity and thus a high affinity for reacting with electrons. As a result, PMCH, similar to other PFTs, have one of the highest detection sensitivities when using a GC-ECD (Dietz 1986, 1991). GC-ECDs, given that the correct analysis parameters have been applied, have the ability to achieve detection limits of 8 to 9 fg. Detection limits with PMCH have even been achieved in the 2 to 3 fg range with a heavily customized GCMS (Begley, Foulger, and Simmonds 1988). The aforementioned properties of PMCH and PFTs have resulted in their widespread utilization as tracer gases for a variety of disciplines. As the sole tracer, PMCH has seen limited application. In this capacity, PMCH been used to identify the location of dielectric fluid leaks in high pressure fluid filled cables (Ghafurian et al. 1999) as well as to detect leaks in commercially buried electrical cables (Hassoun, McBride, and Russell 2000). Due to the extensive atmospheric stability of PMCH, it has been implemented in a variety of atmospheric tracer studies. Tracer gas studies of this nature include the investigation of atmospheric transport patterns of anthropogenic pollutants over the central Alps from the western Po valley in Northern Italy to and from the Swiss Plateau (Ambrosetti et al. 1998, Anfossi et al. 1998) as well as the determination of long-range flow characteristics in the Cross-Appalachian Tracer Experiment (CAPTEX ’83). The CAPTEX ’83 study used PMCH to simulate the dispersion of air pollutants over a 1,000 km area from the northeastern part of the United States to the southeastern part of Canada (Ferber et al. 1986). PMCH was also used as a tracer in the one year long Metropolitan Tracer Experiment (METREX) to study air flow patterns over the Washington D.C. area. The PMCH METREX simultaneously used PMCH to quantify error in meteorological air quality models. PMCH is more commonly combined with other PFTs in multiple tracer gas releases to facilitate complex studies. The additional flexibility afforded by multiple tracer gas release enhances the analytical capabilities for ventilation systems and atmospheric transport projects. Multi-PFT releases have been used to analyze HVAC system performance in residential and commercial buildings, to develop models for contaminant infiltration in residences, and to evaluate the effective of residential weatherization systems (Leaderer, Schaap, and Dietz 1985). PMCH was one of several PFTs used to study the behavior of inert pollutants during nocturnal drainage flows over Brush Creek Valley located in the vicinity of Grand Junction, CO (Clements, Archuleta, and 23
Virginia Tech	Gudiksen 1989) and to improve the understanding of complex flow patterns over Salt Lake City, VT as a part of the Vertical Transport and Mixing (VTMX) field campaign (Fast et al. 2006). PMCH released in conjunction with other PFTs was also used to optimize the placement of wind stations, to determine the dispersion characteristics of airborne contaminants, and to validate atmospheric computer models for New York City, NY as a part of the four year New York City Urban Dispersion Program (UDP) research study (Watson et al. 2006). Due to their desirable tracer characteristic, multiple PFT deployments have also been utilized in research areas outside of air flows. Studies of subsurface containment units have extensively used PFTs to evaluate barrier effectiveness and integrity. PFTs were used in a gas-phase partitioning test to determine the extent of non-aqueous phase liquid (NAPL), water, and air saturations in a chlorinated-solvent contaminated containment area in Tucson, AZ (Simon and Brusseau 2007). Direct subsurface barrier studies have included integrity verifications of a close-coupled polymer grout lined cement containment barrier at the Hanford Geotechnical Development Test Facility near Richland, WA, a closed-coupled containment barrier designed to hold laboratory waste, a colloidal silica grout barrier designed to contain contaminated glassware, and a cover system utilized in post-closure operations at various Department of Energy (DOE) facilities (Heiser and Sullivan 2002). PMCH with other PFTs has also been extensively modeled, evaluated, and deployed as an oil and gas reservoir tracer (Dugstad, Bjørnstad, and Hundere 1993) as well as a carbon capture and sequestration (CCS) tracers (Myers et al. 2013). PMCH-PFT releases have already been used to evaluate water and gas production performance of the Gullfaks field petroleum reservoir in the North Sea as part of a pilot study (Ljosland et al. 1993). Similarly, PFTs are also well suited for short-term CCS studies such as integrity testing of new wells or seal verification of old wells (Watson and Sullivan 2012). Although CCS studies with PFTs are still in their infancy, PFTs have already successfully been used to track CO 2 breakthrough locations, subsurface plume migration patterns, and leakage rates (Myers et al. 2013). Examples of such PMCH-PFT combination studies include the monitoring of new surface leakage of sequestered CO from a horizontal well in Bozeman, MO (Wells et al. 2010), and the 2 24
Virginia Tech	surveying of CO leakage rates in a pilot study for the West Pearl Queen (WPQ) depleted oil 2 formation in New Mexico (Wells et al. 2007). The previous overview of PMCH tracer studies demonstrates the effectiveness and versatility of this substance in HVAC, atmospheric transport over complex terrain, oil and gas reservoirs, as well as subsurface studies (Dietz 1991). The inert, non-reactive, and non-toxic nature of PMCH makes it not only an ideal tracer gas choice but also a suitable alternative to SF . Despite 6 substantial use in other fields, PMCH has not yet been implemented to study underground mine ventilation systems. This absence of use can be attributed to two primary issues, the lack of a suitable analytical method for simultaneous PMCH-SF releases and the lack of a suitable mine- 6 scale PMCH release system. Although recent research has yielded some PMCH-SF analytical 6 methods, this type of tracer gas release has seen far less application than the multi-PFT approach. 2.6 SF and PMCH Simultaneous Tracer Gas Studies 6 The majority of tracer gas studies in commercial buildings and in dwellings before the 1990s were satisfied with the deployment of a single tracer in a single zone. This single zone treatment of ventilation systems is well tested, reliable, and accurate. However, as structures and their associated HVAC systems became more complex, the need for multi-zone tracer analysis techniques began to grow (Sherman 1989). This need is no longer merely isolated to commercial, industrial, and residential buildings, but has expanded to any ventilated space that can be divided into isolatable subsections, such as in underground mine ventilation systems. One of the fundamental principles of multi-zone treatments is the simultaneous release of multiple, independent tracer gases. The use of multiple tracer gases is necessary in order to determine interactions between different zones. This issue can be seen with underground leakage studies. Leakage reduces the amount of air available for diluting contaminants to active areas of mines. In order to compensate for leakage, the air quantity delivered to the mine must be increased, which can result in a significant additional expense in fan power and increased stress on bulkheads. Leakage by nature is already difficult to locate using traditional ventilation survey techniques and even more so when 25
Virginia Tech	conducting a single zone tracer gas release. In order to accurately locate, characterize, and quantify leakage using a single zone tracer gas treatment, detailed knowledge regarding the ventilation flow around the suspected area must be known prior to the release. These details are necessary to correctly interpret the tracer gas concentration values. This difficulty is exacerbated when suspected leakage points are located in generally inaccessible areas such as the gob and the middle section of shafts (Hardcastle, Klinowski, and Mchaina 1993). The use of multiple tracer gases allows not only for more accurate quantitative analyses of multi- zone interactions but can also give rapid qualitative feedback about flow paths. In order to implement a multi-zone analysis in an underground mine, SF , the industry standard tracer, 6 would be released simultaneously with PMCH, the novel alternative tracer. Multiple studies have already utilized PMCH to supplement SF in this manner to characterize complex atmospheric 6 flow patterns over Salt Lake City, UT (Fast et al. 2006), to verify subsurface barrier integrity (Heiser and Sullivan 2002), and in CCS studies to track CO breakthrough, subsurface plume 2 migration patterns, and CO leakage rates (Myers et al. 2013, Galdiga and Greibrokk 1997, 2 Watson and Sullivan 2012). Despite these previous applications, an SF -PMCH simultaneous 6 tracer gas release has not yet been used in an underground mine ventilation system. Before such a study can occur, a suitable mine scale release system must first be developed for PMCH. 2.7 Passive Release Sources Underground mine ventilation tracer gas research has historically used SF as its primary 6 substance. The popularity of SF stems not only because it exhibits all of the ideal tracer 6 characteristic but also due to its ease of acquisition and release. As previously introduced, SF 6 exists as a gas in a wide range of environmental conditions. As such, a large variety of release options are available for both continuous and pulse studies. In contrast, PMCH exists as a volatile liquid. In order to be utilized as a tracer gas, PMCH must first be converted into a vapor and then released in a controlled manner. Several options for PMCH deployment are available. PMCH can be purchased as a vapor in pressurized nitrogen (N ) cylinders, be vaporized in 2 heated N streams, or be released from passive sources. PMCH-N cylinders have been 2 2 26
Virginia Tech	successfully used to release PMCH in the previously introduced UDP study of atmospheric transport over New York City, NY. The turbulent N stream system was used to release PMCH 2 in the CAPTEX ’83 study. This release system was not only designed to vaporize the tracer but also to simultaneously measure the exact concentration of the tracer. This system accomplished these tasks by mixing the volatile liquid with a pressurized stream of N gas that was passed 2 through a 105C oven. Once completely vaporized, the concentration of the tracer stream was then determined with an electronic mass flow meter (Ferber et al. 1986). Both the pressurized cylinder and turbulent stream release systems have not been extensively utilized due to the inherent cost and difficulty associated with these methods. Passive release systems have been the preferred method for PMCH and PFT releases overall due to their cost effectiveness, ease, and reliability. Passive release sources have been developed in two main forms, the permeation tube and the permeation plug. A permeation tube is a polymer tube, such as PTFE, that contains the volatile liquid tracer. The tube is plugged and sealed at both ends. Once sealed, the tube ends are airtight and impenetrable to the tracer. PMCH permeates through the walls of the tube at a constant rate in proportion to temperature and release as a vapor to atmosphere (Boyle 2010b, a). This particular version of the passive release source has seen limited use in tracer studies due to cost and narrow operating constraints. The permeation plug is considerably more common. The property that allows the permeation plug to function is the natural permeability of silicone based rubbers (Van Amerongen 1947). This diffusion through a seemingly impermeable membrane can be achieved due to the chemical composition of silicone rubber. Silicone, also known as polysiloxane, is a name used to define any compound derived from polymerized siloxanes. Polymerized siloxanes are substances whose molecular structure is created by combining monomers into large chains of alternating silicon and oxygen atoms (Encyclopædia Britannica 2012). A siloxane can be further defined as any compound that has alternate silicon (Si) and oxygen (O) atoms (e.g. Si-O) where organic groups or hydrogen atoms are bonded to the Si atom (Merriam-Webster 2012, Polmanteer and Falender 1984). An example of a structural formula for a common type of silicone is displayed in figure that follows. 27
Virginia Tech	Figure 2.4. Chemical structure of polydimethylsiloxane (PDMS), the most common form of silicone rubber. As can be seen in Figure 2.4, the siloxane molecules can rotate freely around the Si-O bond, which allows silicone compounds to be highly flexible (Encyclopædia Britannica 2012). This inherent flexibility afforded by the Si-O bond also permits the existence of free volumes within the compound thus allowing for gas permeability. Incidentally, silicone also has the highest permeability of any polymer, which ideally suits this compound for use in the PMCH permeation plug (Zhang and Cloud 2006). One of the first iterations of the permeation plug was manufactured by impregnating the center of a fluoroelastomer with liquid PMCH using a syringe. The impregnated plug would then be crimped into a cylindrical aluminum shell so that only one end of the plug would be open to atmosphere. The PMCH would then diffuse through the fluoroelastomer and release to atmosphere at a constant rate once equilibrium was reached. This form of the passive release source has been deployed in conjunction with a CATS to measure air infiltration into a home (Dietz and Cote 1982) and to tag illicit drugs and paper currency (Balestrieri and Kaish 1995), as well as to tag pre-detonation blasting caps to facilitate the detection of clandestine bombs at security checkpoints (Senum et al. 1980). This design was followed by the permeation plug release cylinder (PPRC) design. The version of the permeation plug was originally composed of a 32 mm long aluminum cylinder filled with 0.4 mL of PMCH. Both ends of the cylinder were slightly flared at both ends 28
Virginia Tech	to allow the press-fit of two slightly oversized silicone plugs. Once sealed, PMCH would achieve diffusion equilibrium with the plug in approximately two hours (Winberry et al. 1990). The PPRC design was then refined slightly so that one end of the aluminum cylinder was sealed and the other end was left open to atmosphere. The shell was first filled with liquid PMCH and then immediately sealed with an oversized silicone plug pressed flush to the end of the cylinder. Once sealed, the operating principle remained the same. PMCH would diffuse through the plug and release to atmosphere at a constant rate proportional to atmosphere. This final design, for the purposes of this discussion, will be referred to the permeation plug release vessel (PPRV). The PPRV has been utilized in a variety of tracer gas application such as the analysis of air infiltration and air exfiltration in buildings (Dietz et al. 1986, D'Ottavio, Senum, and Dietz 1988), the quantification of air exchange rates (Winberry et al. 1990), and the determination of ventilation rates in Swedish dwellings as a part of a nationwide energy and indoor climate survey (Stymne, Bowman, and Kronvall 1994). The PPRV has also been applied to more unique studies including subsurface barrier integrity studies at Brookhaven National Laboratory’s (BNL) Waldo test site (Heiser and Sullivan 2002), large-scale, continuous CCS integrity verification studies (Watson and Sullivan 2012), and laboratory based research on gas reservoirs located in the North Sea (Galdiga and Greibrokk 1997). Despite the significant use of the PPRV, information regarding how certain design aspects, such as plug thickness, affect the release of PMCH is practically non-existent. This lack of information most likely results from the confined use of the PPRV to HVAC studies thus removing the need for more universal design specification. The PPRV has also not yet been used in tracer gas studies of underground mine ventilation systems. The purpose of the discussion that follows is to develop a PPRV design that is suitable for mine- scale tracer gas studies and well as to present laboratory and field verification studies of the final PPRV design. 29
Virginia Tech	Wells, A. W., B. R. Strazisar, J. R. Diehl, and G. Veloski. 2010. "Atmospheric tracer monitoring and surface plume development at the ZERT pilot test in Bozeman, Montana, USA." Environmental Earth Science no. 2010 (60):299-305. Widodo, N. P., K. Sasaki, R. S. Gautama, and Risono. 2008. "Mine ventilation measurements with tracer gas method and evaluations of turbulent diffusion coefficient." International Journal of Mining, Reclamation and Environment no. 22 (1):60-69. Winberry, W. T., Jr., L. Forehand, N. T. Murphy, A. Ceroli, B. Phinney, and A. Evans. 1990. Compendium of methods for the determination of air pollutants in indoor air. Triangle Park, NC: Atmospheric Research and Exposure Assessment Laboratory. Young, D. A., G. W. Bonnell, and D. G. Genter. 2001. Tracer gas techniques for mapping air and methane migration through a longwall waste in an underground coal mine using tube bundle systems. In SME Annual Meeting. Denver, CO. Zhang, H., and A. Cloud. 2006. The permeability characteristics of silicone rubber. In 38th SAMPE Fall Technical Conference: Global Advances in Materials and Process Engineering. Dallas, TX: Society for the Advancement of Material and Process Engineering. 37
Virginia Tech	Chapter 3: A technique for creating perfluorocarbon tracer (PFT) calibration curves for tracer gas studies ABSTRACT: The use of sulfur hexafluoride (SF ) as a tracer gas for analyzing underground 6 mine ventilation systems has been practiced for many years. As a result, the methods used to release, sample, and analyze SF are well accepted. Although improvements are still being made 6 to enhance the analysis of this tracer gas, the technique remains largely the same. However, as the complexity and size of underground mine ventilation networks increase, the ability of a single gas to function as a convenient and rapid means of analysis diminishes. This problem arises from the need to allow SF to be completely cleared from a ventilation system before 6 another test can be started. The utilization of multiple tracer gases can mitigate this problem by allowing for simultaneous releases in multiple underground locations thus facilitating a more comprehensive evaluation. Additionally, tracer gas tests can be executed in a consecutive manner without the risk of cross-contamination. Although multiple tracer gases have already been extensively used in the fields of heating, ventilation, and air conditioning (HVAC) and large- scale atmospheric monitoring, this technique has not been widely implemented in underground mine ventilation. This lack of use is partially due to the difficulty of releasing additional tracer gases. A well-documented alternative in HVAC studies to SF as tracer gas are perfluorocarbon 6 tracers (PFT). However, many PFTs exist as volatile liquids at room temperature and pressure. This characteristic provides a challenge for quantification using gas chromatography (GC). This paper introduces a method for creating a GC calibration curve for gaseous perfluoromethylcyclohexane (PMCH) from its liquid form and details the experimental parameters used in the evaluation. 38
Virginia Tech	3.1 Introduction The use of sulfur hexafluoride (SF ) as a tracer gas for modeling underground mine ventilation 6 systems has been practiced for many years. As a result, the methods used to release, sample, and analyze SF are well defined and accepted. Although improvements are still being made to 6 enhance the analysis of this tracer gas, the applied techniques remain largely the same across the mining industry. The basic method is initiated by releasing SF and then sampling the mine air at 6 pre-selected locations either on a one-time, instantaneous basis or on a continuous, automated basis. The air sample is then examined to determine the concentration of the tracer gas at the sampling points (Hartman et al. 1997). This data can then be used to model the ventilation flow within the mine. Although this method of SF deployment and analysis is proven, its application 6 may be limited. This difficulty largely stems from the increasing complexity of underground mines. The advancement of mining technologies and techniques has allowed for larger and more intricate mine geometry. The ability of a single gas to function as a convenient, rapid means of ventilation network analysis is diminished by the continual expansion of underground mines. A standard tracer gas examination requires that SF be released either as a constant stream until its 6 concentration is uniform at the release point and the outlet (i.e. conservation of mass) or as a large pulse. Either method requires a significant amount of tracer gas to achieve detectable levels at the sampling points. The background presence of SF caused by such large releases presents 6 some problems. In order to conduct a consecutive release tracer gas studies, sufficient time must be allowed between tests to reduce the background presence of the gas. The reduction of background presence is essential so that subsequent tests of the ventilation system are not affected by the residual concentration of the previous release. At the concentration levels generally captured at sampling locations, which can range from parts per million (PPM) to parts per trillion (PPT), even minute amounts of additional tracer gas can drastically affect the final analysis. Unfortunately, this problem is especially present with SF due to its natural tendency to adhere to 6 surfaces. Additionally, SF is approximately five times heavier than air, which causes the gas to 6 39
Virginia Tech	linger in areas of low ventilation flow. As a result, a substantial amount of time must be allowed to ventilate the excess tracer gas prior to subsequent ventilation surveys (Thimons, Bielicki, and Kissell 1974). The utilization of additional tracer gases has the potential to mitigate these problems in underground mine ventilation. The ability to deploy multiple tracer gases not only allows for consecutive releases with minimal delay and without concern of contamination, but can also facilitate simultaneous, multi-location releases. These added capabilities may allow for a more rapid, comprehensive evaluation of ventilation systems. Multiple tracer gases have already been extensively used in the fields of heating, ventilation, and air conditioning (HVAC), as well as large-scale atmospheric monitoring. For example, a well-documented alternative in HVAC studies to SF as tracer gas 6 are perfluorocarbon tracers (PFT). This classification of tracer gas, more specifically the gas perfluoromethylcyclohexane (PMCH), will be the main topic of discussion in this paper. Despite the use of PFTs in HVAC, there is no evidence in the literature that these tracers have been implemented in underground mines. The lack of use in underground mines is partially due to the difficulty of concurrently analyzing alternative tracer gases with SF Previous work by the 6. authors has shown that both gases can be detected with similar sensitivity on a single column via gas chromatography. However, before field tests can be conducted, an additional complication still exists with PFTs that must be resolved. Many PFTs exist as volatile liquids at room temperature and pressure. This characteristic adds a level of difficulty when quantifying the tracer from air samples. This paper introduces a method for creating a GC calibration curve for gaseous PMCH from its liquid form and details the experimental parameters used in the evaluation. 3.2 Background PMCH is a volatile liquid that is part of the PFT compound group. These compounds are composed of perfluoroalkanes (e.g. PMCH, perfluoromethylcyclobutane (PMCB), perfluoromethylcyclopentane (PMCP), etc.), which are biologically inert, chemically inert, and thermally stable. The inert, non-reactive, and non-toxic nature of PFT compounds makes them 40
Virginia Tech	ideal choices as alternative tracer gases. Although PMCH can potentially be complimented by other PFTs, these other tracers have not yet been successfully deployed in conjunction with SF . 6 The chemical formula of PMCH is C F and has a molecular weight of 350 g/mol with a boiling 7 14 point of 76°C. This compound, as previously stated, exists as a volatile liquid at room temperature. The volatile nature of PMCH allows it to vaporize at relatively low temperatures. Once in a vapor state, PMCH will remain a vapor even through cooler temperatures (National Institute of Standards and Technology 2011). PMCH, similar to other tracer gases, can be used as a tracer gas because it does not adversely affect people or the environment at trace levels. Additionally, the low ambient background of PMCH in the atmosphere, in the parts per quadrillion (PPQ) range, classifies this compound as having no significant natural presence (Watson et al. 2007). Thus, this trace background presence will not significantly impact the concentration of the data collected in tracer gas studies. Perhaps the greatest advantage afforded by PMCH is its ability to be used in concurrent deployments of tracers due to its ability to be separated from SF , the standard tracer gas employed in the mining 6 industry, as well as oxygen. Despite these tracer characteristics, PMCH has not yet been implemented as a tool in underground mine ventilation analysis. However, this tracer along with other PFT group compounds has been deployed extensively in other fields. Building ventilation is an area of study that has seen significant implementation of PFTs. This utilization of PFTs is a result of the inherent limitations present in more conventional ventilation analysis techniques. These limitations impede accurate measurements of low air flows and low leakages, such as those found in HVAC systems. PFTs have been used in extensive studies to investigate air infiltration into single family homes using passive liquid PFT permeation sources coupled with passive adsorption tube samplers (Dietz and Cote 1982, Leaderer, Schaap, and Dietz 1985) as well as to evaluate the performance of multi-zone deployment of passive PFT sources for categorizing air infiltration, air exfiltration, and air exchanges (Dietz et al. 1986). PFTs have also been used to evaluate ventilation rates in the housing stock of Sweden as a nation-wide effort to determine the adequacy of the ventilation systems (Stymne, Bowman, and Kronvall 1994) as well as to characterize down-valley flow, canyon outflow, and interacting 41
Virginia Tech	circulations at the lower slopes of the Wasatch Front (Fast et al. 2006). The versatility of PFTs as a tracer has found some novel deployments in fields outside of ventilation as well. PFTs have been used in the area of explosives for the pre-detonation tagging of electric blasting caps to allow the detection of clandestine bombs in high security areas (Senum et al. 1980). The tracing ability of PFTs has also been used to characterize the complex flow found in the Gullfaks oil fields in the North Sea (Ljosland et al. 1993), to locate dielectric fluid leaks on pipe-type/self- contained cables on the Con Edison transmission system (Ghafurian et al. 1999), and to determine the integrity of subsurface barriers used for the confinement of waste sites (Sullivan et al. 1999). The previously introduced examples of PFT applications demonstrate the versatility of this compound as a tracer. Presently, only PMCH, a PFT group compound, has been successfully separated from SF in preliminary laboratory based mine ventilation research by the authors. 6 This tracer has not yet been deployed in underground field studies due to its initial existence as a volatile liquid at normal temperature and pressure (NTP). Thus, no techniques exist for creating a calibration curve for PMCH in the field of underground mine ventilation. The generation of a calibration curve is a necessary aspect of quantitative GC investigations due to the nature of this technology. Although GCs are manufactured using nearly identical specifications, each device is unique depending on installation parameters (e.g. length of column, composition of the stationary phase, composition of the mobile phase, etc.) and device configurations (e.g. linear velocity of the mobile phase, temperature of the column, etc.). Thus, a calibration curve using known standards must be created for each analysis to serve as a reference for samples with unknown concentrations (McNair and Miller 2009). This paper introduces a method for creating a GC calibration curve for gaseous PMCH from its liquid form to facilitate the quantification of tracer concentration following a tracer release. 42
Virginia Tech	3.3 Experimental Design In order to develop a method for creating a GC calibration curve for gaseous PMCH, three primary tasks were identified: creating the master standard, creating the liquid standards, and creating the gaseous standards. The following sections detail the experimental parameters from preparing the PMCH master standard to final generation of the calibration curve. A section is also provided that discusses the PMCH release sources that can be potentially quantified by the calibration curve. 3.3.1 Equipment The analysis of PMCH was completed using a gas chromatograph (GC) equipped with an electron capture detector (ECD). The ECD was chosen due to its high sensitivity to trace amounts of electronegative compounds (McNair and Miller 2009). The GC was fitted with a 30 m long, 0.25 mm ID porous layer open tube (PLOT) alumina oxide capillary column deactivated with sodium sulfate. The column has a film thickness of 5 µm. In order to minimize error in the sample introduction process, an auto-injector equipped with a 2.5 mL headspace syringe, syringe heater, sample heater, and sample agitator was used. The details of how the auto-sampler was used in conjunction with the GC as well as their analysis settings will be discussed later in this section. 3.3.2 Concentration Range Tracer gas standards are created to calibrate a GC for measuring concentrations of a target substance in gas and liquid samples. Tracer analyses are generally conducted to find volumetric ratios of tracer gas to total volume of gas (i.e. PPM, PPB, and PPT based on volume). Thus standards are created using the same scale. Standards can be generated using a variety of techniques. A simple but popular method is to use a container of known volume that is filled with an inert gas such as nitrogen. A known volume of tracer is then injected into the nitrogen- filled container thus producing a known volumetric ratio of tracer gas to total volume of gas. 43
Virginia Tech	However, this method must be modified for this experiment due to the chemical properties of PMCH. PMCH exists as a volatile liquid at normal temperature and pressure (NTP) but must be deployed as a gas when used as a tracer. As a result, the calibration standards should be made using gaseous PMCH. Although a GC can accept liquid samples, the volume increases as liquid transitions into gas. This expansion causes the concentration to change. As a result, gaseous PMCH must be used in lieu of liquid PMCH to create a suitable calibration curve. The process of creating gaseous PMCH standards from a pure liquid is challenging because errors can be easily propagated through each step of the process. These errors can be minimized by acquiring commercially prepared gaseous PMCH standards. These commercial standards can be further diluted very accurately as a gas to generate calibration standards using various means. However, commercial standards can be expensive, especially if a wide range of concentrations is required. Thus, a technique was developed to produce the calibration standards using readily available equipment. The procedure for creating gaseous standards from pure liquid PMCH for this experiment will now be discussed. The range of concentrations required for the calibration curve was dictated by the expected concentration two PMCH release sources that will be used in a subsequent experiment. These sources are discussed in Section 5. The expected sampling concentrations from the two release sources were computed with the following steps. The sources will be deployed in a straight section of 5.08 cm (2 in) diameter pipe with a constant air flow of 5 m/s (980 fpm). This linear velocity was estimated prior to the experiment to provide a means for predicting PMCH concentrations in the pipe. The cross-sectional area of the pipe was first computed to be 0.00203 m2. The area was then used in Equation ( 3.1 ) to determine the volumetric flow rate in the pipe. 44
Virginia Tech	𝑚3 𝑄 = 𝑉𝐴 ( ) ( 3.1 ) 𝑠 Where: Q = volumetric flow (𝑚3⁄𝑠) V = linear velocity (𝑚⁄𝑠) A = cross-sectional area (𝑚2) Substituting known values into Equation ( 3.1 ) and converting the resulting units, the following calculation can be made. 𝑚3 𝑚3 𝑄 = 5∙0.00203 = 0.01 = 0.61 𝑠 𝑚𝑖𝑛 The volumetric flow was then used in conjunction to the mass flow rate of the release source to compute the expected concentration of PMCH at the sampling point. Due to the similarity of the computation, only the expected concentration calculation for one of the PMCH release sources will be provided. This source is expected to provide a flow rate of 26.4 𝜂𝐿. The computations 𝑚𝑖𝑛 were completed assuming a constant temperature of 21.5°C, a straight pipe, and a constant linear velocity. A liquid density of 1,787 𝑔 is also used for PMCH (National Institute of Standards and 𝐿 Technology 2011). The mass flow of PMCH from the release source was calculated as follows. 26.4 𝜂𝐿 𝐿 1787 𝑔 ∙ ∙ = 4.75∙10−5 𝑔 𝑚𝑖𝑛 109 𝜂𝐿 𝐿 𝑚𝑖𝑛 Using the volumetric flow rate from Equation ( 3.1 ) and the mass flow rate of the release source, the expected concentration of PMCH can determined as follows. 𝑔 4.75∙10−5 𝑃𝑀𝐶𝐻 𝑘𝑔 𝑘𝑔 𝑚𝑖𝑛 ∙ = 7.76∙10−8 𝑃𝑀𝐶𝐻 𝑚3 1000𝑔 𝑚3 0.61 𝑎𝑖𝑟 𝑎𝑖𝑟 𝑚𝑖𝑛 This concentration can also be represented in PPB by molarity by converting the volumetric flow of air and the mass flow of PMCH to molar flow, which is a more convenient means of 45
Virginia Tech	representation in certain data collection scenarios. The molar flow of air is calculated first by determining the mass flow assuming a dry air density of 1.205 𝑔. 𝐿 0.61 𝑚3 1000 𝐿 1.205 𝑔 𝑔 ∙ ∙ = 733 𝑚𝑖𝑛 𝑚3 𝐿 𝑚𝑖𝑛 The mass flow can then be converted to a molar flow using a molecular weight of 28.97 𝑔 for 𝑚𝑜𝑙 dry air. 733 𝑔 𝑚𝑜𝑙 𝑚𝑜𝑙 ∙ = 25.29 𝑚𝑖𝑛 28.97 𝑔 𝑚𝑖𝑛 Please note that the previous calculations for the molar flow of air was completed using the density and molecular weight of dry air. This molar flow does not represent the exact flow in the future quantification experiment due to varying atmospheric conditions, such as humidity and barometric pressure. The chemical properties of dry air were chosen as an initial point or reference from which the tracer concentration could be approximated. The molar flow of PMCH from the release source can now be computed as follows using the previously determined mass flow of 4.75∙10−5 𝑔 and a given molecular weight of 350.05 𝑔 (National Institute of 𝑚𝑖𝑛 𝑚𝑜𝑙 Standards and Technology 2011). 4.75∗10−5 𝑔 𝑚𝑜𝑙 ∙ = 1.35∙10−7 𝑚𝑜𝑙 𝑚𝑖𝑛 350.05 𝑔 𝑚𝑖𝑛 The previously derived values can now be used to find the ratio of the molar flow of PMCH to the total molar flow of PMCH and air. 1.35∙10−7 𝑚𝑜𝑙 𝑚𝑖𝑛 = 5.33∙10−9 1.35∙10−7 𝑚𝑜𝑙 +25.29 𝑚𝑜𝑙 𝑚𝑖𝑛 𝑚𝑖𝑛 This ratio represents the molar percent of PMCH in the total flow. The molar percent can then be multiplied by 109 to convert the ratio into molar PPB. 46
Virginia Tech	5.33∙10−9 ∙109 = 5.33 𝑃𝑃𝐵 The sampling concentration of the other release source was approximated to be 2.23∙10−9 𝑘𝑔 𝑚3 (0.15 PPB by moles). Thus, the calibration curve must at least contain these two points for adequate interpolation. A target concentration range of 1.34∙10−9 𝑘𝑔 to 1.39∙10−7 𝑘𝑔 (0.09 𝑚3 𝑚3 PPB to 9.58. PPB by moles) was selected for the calibration curve to both encompass the two approximated sampling concentrations and allow for any fluctuations in PMCH concentrations. The process by which the standards were created will now be discussed. 3.3.3 Standard Preparation The creation of gaseous PMCH standards from a pure liquid was completed by the three-step process described below. 1. A master liquid standard of diluted PMCH was prepared. 2. Varying volumes of the master standard depending on the desired final concentration were transferred to four separate containers for further dilution. 3. A small, constant volume of each liquid standard was individually transferred to four headspace containers and allowed to vaporize thus creating the gaseous standards. For step one, the master standard was prepared in a 20 mL screw top, airtight, septum-capped vial. Using a volumetric flask, 20 mL of hexanes were added to the sample vial. The vial was then immediately sealed to prevent any significant loss of the solvent. Five µL of pure liquid PMCH was then injected into the hexanes through the septum using a liquid syringe. The 5 µL volume of PMCH was chosen because this amount was the lowest that could be reliably injected with reasonable precision and readily available equipment. The PMCH was injected through the septum to eliminate sample loss from the high volatility of PMCH, which was not a concern for the hexanes. The master standard was then agitated to ensure homogeneous distribution of the solute. The concentration of the master standard was determined using the mass of PMCH per unit volume of hexanes relationship shown below. 47
Virginia Tech	𝐿 1787 𝑔 𝑘𝑔 5 𝜇𝐿 ∙ ∙ 𝑃𝑀𝐶𝐻 ∙ 𝑃𝑀𝐶𝐻 106 𝜇𝐿 𝐿 1000 𝑔 𝐿 𝑚3 20 𝑚𝐿 ∙ ∙ 𝐻𝑒𝑥𝑎𝑛𝑒𝑠 1000 𝑚𝐿 1000 𝐿 𝑘𝑔 𝑃𝑀𝐶𝐻 = 0.45 𝑚3 𝐻𝑒𝑥𝑎𝑛𝑒𝑠 The second step was to prepare the individual liquid standards from which the final gaseous standards would be made. A total of four liquid standards were created with varying concentrations from the master standard. This step was accomplished by injecting varying volumes of the master standard to four separate screw-top, septum-capped vials each filled with 20 mL of hexanes. As a result, the four liquid standards contained varying dilutions of PMCH. The injection volumes from the master standard to each of the sample vials were determined based on how the dilution would affect the final gaseous concentration after the final transfer- vaporization phase in step three. In order to determine the transfer volume in this manner, a computational sheet was created where different transfer volumes could be entered. The resulting gaseous concentration would then be displayed thus allowing for the transfer volumes to be found through a trial-and-error process. A sample calculation showing the computation of the final gaseous concentration will now be explained. The concentration of the master standard was first converted to a molarity, which is defined as moles of solute per liter of solution, to simplify the dilution calculation of the individual dilutions. 0.45 𝑘𝑔 1000 𝑔 𝑚3 𝑚𝑜𝑙 𝑃𝑀𝐶𝐻 𝑃𝑀𝐶𝐻 ∙ ∙ ∙ 𝑚3 𝑘𝑔 1000 𝐿 350.05 𝑔 𝐻𝑒𝑥𝑎𝑛𝑒𝑠 𝑃𝑀𝐶𝐻 𝑚𝑜𝑙 𝑃𝑀𝐶𝐻 = 0.0013 𝐿 𝐻𝑒𝑥𝑎𝑛𝑒𝑠 The molarity of the standard can now be calculated based on the transfer volume based on the following equation. 48
Virginia Tech	𝑀 𝑉 = 𝑀 𝑉 ( 3.2 ) 1 1 2 2 Where: M = molarity (𝑚𝑜𝑙 ⁄𝐿 ) 𝑆𝑜𝑙𝑢𝑡𝑒 𝑆𝑜𝑙𝑢𝑡𝑖𝑜𝑛 V = volume of solution (𝐿) A value of 7 𝜇𝐿 will be used in this demonstration for the volume transferred from the master standard to the separate sample vial filled with 20 mL of hexanes. This transfer volume was used to create the 2.68 PPB by moles calibration point. The following equation is created by rearranging Equation ( 3.2 ). 𝑀 𝑉 𝑚𝑜𝑙 1 1 𝑀 = ( ) ( 3.3 ) 2 𝑉 𝐿 2 Substituting known values into Equation yields the following result for molarity. 0.0013∙(7∙10−6) 𝑀 = = 4.5∙10−7 𝑚𝑜𝑙 2 [0.02+(7∙10−6)] 𝐿 During the final step, 5 𝜇𝐿 of each liquid standard was injected into separate 20 𝑚𝐿 screw-top, septum-sealed headspace vials. Using the molarity of the liquid standard, the mass of PMCH in each of the headspace vials can be calculated by converting molarity to density in units of grams per microliter (𝑔 ). The density of the standard is 1.6∙10−10 𝑔 for the standard in the calculation 𝜇𝐿 𝜇𝐿 previously shown using Equation ( 3.3 ). The final transfer volume of 5 𝜇𝐿 can then be used to determine the mass of PMCH in the final injection. 1.6∙10−10 𝑔 5 𝜇𝐿∙ = 7.8∙10−10 𝑔 𝜇𝐿 Using the mass of PMCH, the gaseous density of PMCH in the final standard can be computed by dividing the mass of PMCH by the volume of the container. This calculation is reasonable because the small amount of liquid PMCH was allowed to vaporize completely inside the sealed headspace vial. Once vaporized, the gaseous PMCH is assumed to be well diffused inside the vial, thus creating a standard with a known mass per unit volume concentration. For this sample 49
Virginia Tech	calculation, a mass of 7.8∙10−10 𝑔 was injected into a 20 mL container giving a gaseous density of 3.9∙10−8 𝑘𝑔. For convenience, this concentration can be converted to a PPB by moles using 𝑚3 the previously introduced process for estimating PMCH concentration of the release sources. The concentration of the gaseous standard in PPB by moles was 2.68 PPB. Using the aforementioned process, four gaseous standards were created for the calibration curve by transferring 0.24 µL, 7 µL, 15 µL, and 25 µL of the 0.0013 molar master standard to four separate vials filled with 20 mL of hexanes. 5 µL from each of the four liquid standards were then injected into their own respective headspace vials to create four gaseous standards thereby defining the calibration curve. The four point quantity was chosen because this is the minimum number of points required to confidently interpolate a linear relationship. The parameters, or method, used by the GC to analyze the PMCH standards will now be discussed. 3.3.4 Gas Chromatograph Method The calibration standards were all injected into the GC using an auto-sampler. Prior to the injection, both the sample vial and the needle were heated to 50°C to encourage vaporization of the PMCH into the headspace and to prevent condensation on the needle, respectively. In addition, the sample vials were agitated at 500 RPM for one minute to ensure even distribution of the gaseous PMCH in the sample vial prior to injection. Once injected, the GC method was initiated. The GC method used for this experiment required 16 min to analyze each injection and consisted of the following static parameters. Table 3.1. GC method static parameters Parameter Description Carrier Gas He Make-Up Gas N 2 Injector Temperature 150°C Split Ratio 30:1 Linear Velocity 30 cm/s Detector Temperature 200°C 50
Virginia Tech	The only dynamic parameter in the method was the temperature of the column. The temperature program used for the column is displayed in the following table. Table 3.2. GC Column Temperature Program Parameter Description Initial Column Temperature 67°C Hold Initial Temperature 2.75 min Temperature Increase 120°C/min at 2.75 min Final Column Temperature Hold 180°C until end Total Program Runtime 15.99 min In order to prevent any contamination from residual compounds present in the GC, the column temperature was raised to 175°C and held for 30 min to thermally dissociate, or bake-off, these compounds. After the baking process, the PMCH analysis method described above was executed without an injection to verify the cleanliness of the column (i.e. column blank). A sample of the laboratory atmosphere was then injected into the GC using the auto-sampler syringe to ensure that PMCH was neither present in the air (i.e. air blank) nor in the syringe (i.e. syringe blank). Once this process was completed, the calibration standards could then be injected. A total of three consecutive injections per calibration standard were made into the GC to quantify relative standard deviations (RSD) between the injections. The column was baked at 175°C for 15 min between injections of different calibration standards to remove any residual contaminants prior to the next set of injections. Another syringe blank was also completed prior to injecting the next calibration standard to re-verify the cleanliness of the syringe. The resulting calibration curve data are presented in the following section. 3.4 Experimental Results The GC analysis of the gaseous PMCH standards produced a chromatogram for each injection. Only select chromatograms that best represent each calibration point will be displayed due to the fact that the inclusion of every chromatogram in this paper would be cumbersome. The overlaid chromatograms for the 0.09 PPB PMCH, 2.68 PPB PMCH, 5.75 PPB PMCH, and 9.78 PPB 51
Virginia Tech	The 0.998 regression value displayed in Figure 3.6 and the low relative standard deviations for the injections displayed in Table 3 show that the methodology used to create the curve is robust. The completed calibration curve can now be used to evaluate the concentrations of two PMCH release sources. These two sources are discussed in detail in the following section. 3.5 Potential PMCH Release Sources The first release method was developed by Brookhaven National Laboratory (BNL) and consists of a hollow aluminum cylinder that is 32 mm (1.25 in) long with an inside diameter of 6.6 mm (0.25 in). One end of the cylinder is open to atmosphere while the other end is closed. Precisely 0.4 mL of liquid PMCH is injected into the aluminum vessel and sealed with an oversized silicone plug pressed flush to the end. The plug is 12.7 mm (0.5 in) long with an outside diameter of 7 mm (0.275 in). Once the vessel is sealed, the PMCH is slowly absorbed by the silicone plug until equilibrium is reached. Once the PMCH equilibrates, vapor PMCH will begin desorbing to atmosphere from the plug at a predictable rate directly proportional to temperature. At equilibrium, the source acquired from BNL will release PMCH at 26.4 ηL/min at 21.5°C (Dietz et al. 1986). This release source is reliable and relatively inexpensive either purchased or manufactured. The main disadvantage of this release source is the exceptionally small release rate thereby requiring the use of many sources in parallel to produce an acceptable concentration on a mine scale. The second release method functions using the same basic principle, but its construction and release rate differs slightly. The second release method is a commercially available permeation tube. The permeation device is a small, inert tube that contains pure PMCH in a two-phase equilibrium, gas and liquid. At a constant temperature, the device releases PMCH by permeation through the walls of a Teflon tube for the entire length between the two impermeable plugs located at the ends. The range of release rates can be modified by varying the length and thickness of the tube. Permeation tubes are available that can release PMCH at rates ranging from 5 ηg/min to 50,000 ηg/min (Valco Instruments Co. Inc. 2011). For the future release experiment, a 2.1 cm permeation tube rated to release PMCH at 645 ηg/min/cm at 40°C was selected. Permeation tubes are readily available and can be produced with exceptionally accurate release rates. However, unlike the aluminum 55
Virginia Tech	vessel, the permeation tube may need to be heated to release any PMCH. This aspect may present a challenge, especially for release in permissible areas of underground coal mines. Additionally, its manufacturing complexity also yields a higher acquisition cost relative to the aluminum vessel. 3.6 Conclusions The study presented in this paper presents a viable method for creating a GC calibration curve for gaseous PMCH from its volatile liquid state. The procedure used to generate the curve can be found in Section 4.3. The final calibration curve is displayed in Section 4.4. The technique introduced in this paper only represents one of many that can be used to achieve the same end. The specific method detailed in Section 4.3 was selected based on the available tools in the laboratory. The 0.998 regression value displayed in Figure 3.6 and the low relative standard deviations for the injections displayed in Table 3.3 show that the methodology used to create the curve is in fact robust, which demonstrates its reproducibility. However, the purpose of this experiment was not only to discover a viable procedure for creating a gaseous PMCH calibration curve but to ultimately quantify the PMCH concentration produced by two potential tracer release sources in future studies. In order to accurately measure the PMCH release, the range of concentrations of the calibration curve was dictated by the expected concentration of PMCH produced by the two potential tracer release media. The first considered release method is manufactured by Brookhaven National Laboratory (BNL) and consists of a hollow aluminum cylinder with one end open to atmosphere and with the other end sealed closed. The second considered release method is a commercially available permeation device. The permeation device is a small, inert tube that contains pure PMCH in a two-phase equilibrium, gas and liquid. At a constant temperature, the tube releases PMCH by permeation through the walls of the Teflon tube for the entire length between the two impermeable plugs located at the ends. The evaluation of these two release sources in the laboratory in preparation for large-scale underground deployment will be the next phase of this research project. 56
Virginia Tech	The aforementioned details presented about the PMCH calibration curve’s characteristics and ultimate purpose show that this technique used for its can be customized to fit many concentration ranges. Only the master standard concentration and injection volumes need to be changed to achieve this end. Thus, the techniques introduced in this paper are universally applicable to similar application of PMCH. Additionally, as more PFTs are identified as viable tracers, this technique may be potentially used to prepare calibration curves for these added volatile liquids. 3.7 Acknowledgements The authors wish to thank Dr. Russell Dietz, Brookhaven National Laboratory, who has been generous with his time and knowledge of perfluorocarbon tracers. This publication was developed under Contract No. 200-2009-31933, awarded by the National Institute for Occupational Safety and Health (NIOSH). The findings and conclusions in this report are those of the authors and do not reflect the official policies of the Department of Health and Human Services; nor does mention of trade names, commercial practices, or organizations imply endorsement by the U.S. Government. 57
Virginia Tech	Chapter 4: A preliminary evaluation of potential perfluoromethylcyclohexane (PMCH) release vessel designs for tracer gas studies in underground mines ABSTRACT: Perfluoromethylcyclohexane (PMCH) is a member of the perfluorocarbon tracer (PFT) group of compounds. PMCH has shown to be a viable alternative to the widely used tracer gas sulfur hexafluoride (SF ). This viability stems from the fact that PMCH can be used 6 concurrently with SF while maintaining adequate chromatographic separation and comparable 6 sensitivity during analysis. However, the release of PMCH in an underground mine ventilation system is challenging due to its physical characteristics. SF exists as a gas at room temperature 6 and pressure and can be accurately released using a variety of means. In contrast, PMCH exists as a volatile liquid at room temperature and pressure, a characteristic that prevents PMCH from being released using traditional means. One of the methods that can be used to release PMCH is the permeation plug release vessel (PPRV). The PPRV allows for a controlled, passive deployment of PMCH as a vapor. This paper evaluates several designs that incorporate varying manufacturing techniques, plug thicknesses, and plug materials for the PPRV so that a source appropriate for the mine scale may be developed in the future. 60
Virginia Tech	4.1 Introduction Sulfur hexafluoride (SF ) has been the predominant compound used in underground mine 6 ventilation tracer gas studies for over 30 years (Thimons, Bielicki, and Kissell 1974). However, the ability of SF to function individually as a tracer is being hindered by the advancement of 6 mines. In order to support such growth, a subsequent increase in the complexity and the size of the ventilation systems must also occur. The sheer quantity of air that is pulled through underground mines and the area that the ventilation system must now service has diminished the effectiveness of SF as a mine ventilation analysis tool. In order to mitigate this issue, recent 6 studies have identified the compound perfluoromethylcyclohexane (PMCH) as a viable supplement for SF . 6 PMCH is classified as a perfluorinated cyclic hydrocarbon. Other compounds in this group classification include perfluoromethylcyclobutane (PMCB) and perfluoromethylcyclopentane (PMCP) (Galdiga and Greibrokk 1997). These compounds are also known as perfluorocarbon tracers (PFT) due to their chemical inertness, low toxicity, and trace level background presence in the environment thus uniquely suiting them to be used as tracer gases (Dietz 1991, Watson et al. 2007). Compounds of this type have already been widely implemented in heating, ventilation, and air conditioning (HVAC) as well as atmospheric monitoring studies (Dietz 1991) but not yet in the field of underground mine ventilation due to deployment complications. In contrast to SF , 6 PFTs generally exist as volatile liquids at room temperature and pressure (National Institute of Standards and Technology 2011). This physical property of PFTs removes one of the great advantages afforded by SF , its existence as a gas. As a gas, SF can be purchased in convenient, 6 6 standardized gas cylinders and released in a controlled manner using a variety of options (Thimons, Bielicki, and Kissell 1974). Two examples of release options are the flow meter method and the flow controller method. The flow meter method utilizes an analog or a digital flow meter attached to the gas cylinder. SF 6 is then deployed at a pre-determined volumetric flow rate over time, which can be computed to a mass flow rate over time if necessary. This method offers a significant amount of control and flexibility. The flow controller method utilizes a sophisticated electronic regulator that allows for 61
Virginia Tech	4.2 Background PMCH is a perfluorinated cyclic hydrocarbon whose chemical structure is composed of perfluoroalkanes (Watson et al. 2007). Compounds of this type are biologically inert, chemically inert, and thermally stable (F2 Chemicals Ltd. 2011). The inert, non-reactive, and non-toxic nature of PMCH makes it an ideal choice as a tracer gases. PMCH is comprised of seven carbon atoms and fourteen fluorine atoms, which gives it a chemical formula of C F . This highly 7 14 fluorinated molecule has a molecular weight of 350 g/mol and a boiling point of 76°C (169°F). Although highly volatile, its molecular weight causes PMCH to exist as a liquid at room temperature and pressure. Simultaneously, the volatile nature of PMCH allows it to vaporize even at low temperatures. Once in a vapor state, PMCH, will remain a vapor even through cooler temperatures (National Institute of Standards and Technology 2011). Another advantage of PMCH is its detectability by GC even at low concentrations. This ability stems from PMCH’s low ambient background in the atmosphere with concentrations in the parts per quadrillion (PPQ) (Cooke et al. 2001, Simmonds et al. 2002, Watson et al. 2007) and high detection sensitivity when using an electron capture detector (ECD) (Simmonds et al. 2002). Perhaps the greatest advantage afforded by PMCH is its ability to be simultaneously analyzed with SF , 6 which is the standard tracer gas employed in the mining industry. Despite these tracer characteristics, PMCH has not yet been implemented in underground mine ventilation. PMCH has, however, seen widespread use in other fields of study in conjunction with other PFT group compounds. Building ventilation is one area that has implemented PFTs. PFTs have been used to investigate air infiltration into single family homes using passive PFT permeation sources coupled with passive capillary adsorption tube samplers (CATS) (Leaderer, Schaap, and Dietz 1985, Dietz and Cote 1982) as well as to evaluate the performance of multi-zone deployments of passive PFT sources for categorizing air infiltration, air exfiltration, and air exchanges (Dietz et al. 1986). PFTs have also been used to evaluate ventilation rates in Swedish housing stock to determine the adequacy of agricultural ventilation systems (Stymne, Bowman, and Kronvall 1994). 63
Virginia Tech	In the field of atmospheric tracing, PFTs have been used to characterize down-valley flow, canyon outflow, and interacting circulations on the lower slopes of the Wasatch Front (Fast et al. 2006). They have also been deployed to evaluate air flow patterns in New York City to improve wind station placements, to supplement knowledge of contaminant flow patterns, and to update atmospheric flow models (Watson et al. 2006). PFTs have additionally been used in long-term, large-scale investigations of the transport and diffusion of gases over the Alpine topography in Switzerland (Ambrosetti et al. 1998) as well as to evaluate the accuracy of meteorological air quality models in Washington, D.C. (Draxler 1967). The aforementioned examples of PMCH applications demonstrate the versatility of this compound as a tracer. Despite PMCH’s use in numerous ventilation studies, its physical properties at room temperature and pressure do present a challenge for translating this technique to a mine environment. In a traditional SF tracer study, a combination of pressure and flow 6 regulators would be used to perform a controlled release. These tools can no longer be used due to the fact that PMCH is a liquid. The PPRV may serve as a viable alternative. The basic concept of the PPRV detailed in this study has been heavily used by Brookhaven National Laboratory (BNL). In order to provide a cursory understanding of how the BNL source operates some background regarding the gas diffusion mechanism must be presented. The controlled release of PMCH is facilitated by the permeability characteristics of silicone rubber. The passage of a gas through rubber-type mediums such as silicone is a well-documented phenomenon that has undergone intensive study for over 50 years (Barbier 1955, Hammon, Ernst, and Newton 1977, Jordan and Koros 1990, Stern, Onorato, and Libove 1977, van Amerongen 1946, Zhang and Cloud 2006). In silicone rubber, similarly to other rubber-type polymers, gas diffusion occurs in three distinct steps: solution of the gas molecules on one side of the silicone membrane, diffusion of the gas molecule through the silicone, and evaporation of the gas from the other side (Barbier 1955, Zhang and Cloud 2006). This diffusion through a seemingly impermeable medium can be achieved due to the chemical composition of silicone rubber. 64
Virginia Tech	Silicone, also known as polysiloxane, is a name used to define any compound derived from polymerized siloxanes. Polymerized siloxanes are substances whose molecular structure is created by combining monomers into large chains of alternating silicon and oxygen atoms (Encyclopædia Britannica 2012). A siloxane can be further defined as any compound that has alternate silicon (Si) and oxygen (O) atoms (e.g. Si-O) where organic groups or hydrogen atoms are bonded to the Si atom (Merriam-Webster 2012). An example of a structural formula for a common type of silicone is displayed in figure that follows. Figure 4.1. Chemical structure of polydimethylsiloxane (PDMS), the most common form of silicone rubber. As can be seen in Figure 4.1, the siloxane molecules can rotate freely around the Si-O bond, which allows silicone compounds to be highly flexible. The flexibility afforded by the Si-O bond permits the existence of free volumes within the compound thus allowing for gas permeability. Incidentally, silicone also has the highest permeability of any polymer, which ideally suits this compound for use in the PMCH release source (Zhang and Cloud 2006). The BNL release source will now be discussed. The original BNL design consists of a hollow aluminum cylinder that is 32 mm (1.25 in) in length with an inside diameter of 6.6 mm (0.25 in). One end of the cylinder is open to atmosphere while the other end is closed. Precisely 0.4 mL of liquid PMCH is injected into the aluminum vessel and sealed with an oversized silicone plug pressed flush to the end. The plug is 12.7 mm (0.5 in) long with an outside diameter of 7 mm (0.275 in). The source deploys PMCH by allowing the state change to occur inside the source. This conversion immediately begins once the vessel is sealed after which vapor PMCH slowly diffuses through the plug (Dietz et al. 1986). 65
Virginia Tech	Once the PMCH equilibrates within the plug (approximately 12 days after the vessel is sealed), vapor PMCH begins desorbing to atmosphere from the plug at a predictable rate directly proportional to temperature. The source releases PMCH at a rate of 4.14∙10−7 g/min at 21.9°C (71.5°F) (Dietz et al. 1986). The release rate changes in direct proportion to temperature. The release source design presented in this paper is based on the overall concept of the BNL source but changes some of the design parameters. A detailed description of the experimental source is presented in Section 3. 4.3 Experimental Design The study presented in this paper will examine several PMCH release source designs. This assessment was completed in two main parts: assembling the PMCH release sources and determining the mass flow rate of each release source design. The following section details the experimental parameters from preparing the PMCH release vessel to final analysis of the flow rate. The PMCH release source components and designs will first be discussed. The main body of the release vessel is comprised of an aluminum cylinder that is 6.35 cm (2.50 in) in length with an outside diameter (OD) of 0.747 cm (0.294 in). The cylinder itself is comprised of a hollow shell with one end sealed and the other end open to atmosphere. The shell has a wall thickness of 0.0356 cm (0.0140 in) that is consistent throughout the body of the cylinder, which gives an inside length of 6.31 cm (2.49 in) and an inside diameter (ID) of 0.711 cm (0.280 in). A schematic of the aluminum shell is displayed in Figure 4.2. 66
Virginia Tech	Two different types of silicone rubber will be used for this experiment to determine the impact of these materials on the release rate of the source. A third material, a silicone coated bromobutyl/chlorobutyl rubber, was also considered in a precursor feasibility study but was immediately eliminated as an option due its highly impermeable nature. The first type of silicone rubber has a temperature rating from -55°C (-67°F) to 204°C (400°F). The second type of silicone rubber is black in color and is significantly softer with a durometer hardness of A50. The black silicone rubber has a maximum rated temperature of 200°C (392°F) with no rated minimum temperature from the manufacturer. The various designs evaluated in this paper will now be discussed. The main modifications of the release vessel’s design were applied to the plug due to the fact that the silicone membrane serves as the means by which PMCH is deployed. Different plug thicknesses (i.e. plug length) and materials were examined in this experiment for their effect on the release rate. Additionally, two in-vessel liquid PMCH filling techniques, pre-filling and post- filling, were compared. The pre-filling technique was completed by injecting liquid PMCH into the vessel prior to capping while the post-filling technique was accomplished by injecting liquid PMCH with a syringe through the plug after capping. The release vessels were all filled with precisely 0.5 mL of liquid PMCH regardless of the design. The size of the aluminum shell itself remained constant throughout the experiment as lengthening the vessel only serves to increase its PMCH storage capacity. The shell manufacturer used for this experiment did not produce shells of varying diameters so the effect of modifying this variable was not examined. A detailed listing of the different designs can be found in Table 4.1 to Table 4.3. 68
Virginia Tech	The sources were created in triplicate for each design to determine its reproducibility and reliability. In addition, triplicate sources served to identify any errors that may have occurred during assembly. Once the sources were completed, their initial masses were recorded. The mass of the sources was then monitored over time at an average temperature of 22.1C (71.8F) to determine the mass flow rate. Once a sufficient number of data points were collected, the data was processed to determine how the release sources were affected. The fully compiled and processed results are presented in Section 4.4. 4.4 Results A full summary of the experimental results including a relative standard deviation (RSD) value for each triplicate can be found in Table 4.4 to Table 4.6. Table 4.4. PMCH post-filling release sources. Change Release Release in Mass Time Rate Source (g) (Days) (g/Day) R2 % RSD G1 0.0504 95 5.44 • 10-4 1.00 G2 0.0409 95 4.76 • 10-4 1.00 8.56 G3 0.0529 95 5.88 • 10-4 1.00 SP1 0.0447 87 5.26 • 10-4 1.00 SP2 0.0432 87 5.22 • 10-4 1.00 6.36 SP3 0.0517 87 5.98 • 10-4 1.00 SS1 0.0476 87 5.86 • 10-4 1.00 SS2 0.0464 87 5.75 • 10-4 1.00 1.42 SS3 0.0455 87 5.66 • 10-4 1.00 70
Virginia Tech	Table 4.5. PMCH silicone plug pre-filling release sources. Change Release Release in Mass Time Rate Source (g) (Days) (g/Day) R2 % RSD S4 0.0715 87 8.44 • 10-4 1.00 S5 0.0722 87 8.56 • 10-4 1.00 0.92 S6 0.0728 87 8.63 • 10-4 1.00 S1 0.0385 87 4.76 • 10-4 1.00 S2 0.0386 87 4.80 • 10-4 1.00 2.50 S3 0.0411 87 5.03 • 10-4 1.00 S7 0.0223 87 3.01 • 10-4 1.00 S8 0.0154 60 3.32 • 10-4 0.99 4.98 S9 0.0168 60 3.33 • 10-4 0.99 Table 4.6. PMCH soft silicone plug pre-filling release sources. Change Release Release in Mass Time Rate Source (g) (Days) (g/Day) R2 % RSD SS7 0.0461 49 9.78 • 10-4 1.00 SS8 0.0496 49 1.05 • 10-3 1.00 3.23 SS9 0.0462 49 9.76 • 10-4 1.00 SS4 0.0465 87 5.80 • 10-4 1.00 SS5 0.0456 87 5.71 • 10-4 1.00 2.28 SS6 0.0439 87 5.49 • 10-4 1.00 SS10 0.0134 49 3.87 • 10-4 1.00 SS11 0.0134 49 3.86 • 10-4 0.99 0.17 SS12 0.0132 49 3.85 • 10-4 0.99 The previous table shows that the techniques used to construct each release vessel design is both robust and repeatable. Additionally, these exceptionally low RSD values demonstrate that the construction of each design triplicate were free of any major errors. The following tables display a summary of how the release rate was affected by the various changes. Table 4.7 shows the percent difference between the average silicone plug and soft silicone plug release rates. 71
Virginia Tech	4.5 Discussion and Conclusions The preliminary study presented in this paper successfully produced multiple PMCH release source designs. The low RSD values between each triplicate set of designs show that the fabrication technique used to create the sources was robust, repeatable, and free of major defects. The experimental results provided in Section 4.4 of this paper clearly present how the various design variables affected the performance of the vessels. The post-filling, silicone plug release vessels will be discussed first. The post-filling vessels were the simplest to construct amongst the other designs. However, the simplicity afforded by being able to inject PMCH through the silicone plug also gave the highest release rate variability. The larger RSD values for the post-filling sources presented in Table 4.1 are immediately apparent in Figure 4.4. The relatively larger separation between the release rates of the post-filling sources is likely the result of damage to the plug caused by the needle during injection. Although the soft silicone material resulted in the lowest RSD values for the post-filling sources, these sources were still subject to injection damage thus placing its long- term reliability in question. Due to the manual nature of the PMCH injection, the plugs were damaged both unpredictably and inconsistently thus causing the higher release rate variability. However, despite the higher RSD values, the individual release rates from the post-filling sources have proven to be exceptionally reliable over the time of this experiment given linear regression values of near one. The pre-filling release vessels were the most difficult to construct but proved to have the least release rate variability. The lower relative RSD values comparatively to the post-filling sources is likely due to the uncompromised integrity of the plug. As a result, the pre-filling technique is preferred over the post-filling technique. Two different materials and three different plug thicknesses were also studied using the pre-filling technique to determine the effect of these variables on the release rate. These variables were not modified in the post-filling sources to a significant degree because the post-filling technique had already been shown to contain the highest possibility for error. 76
Virginia Tech	The relationship between the two plug materials, traditional silicone and soft silicone, is apparent in Table 4.5, Table 4.6, and Table 4.7. The release rates for the soft silicone plugs are consistently higher than the silicone plugs for each plug thickness. The release rates differ consistently across plug materials by plug thickness by about 8%. This result is logical due to the higher flexibility of the soft silicone, which produces a higher permeability according to the physical properties of PDMS materials. Both materials produce PMCH release rates that are equally reliable and reproducible based on the parameters of this preliminary study. A slight advantage is provided by the increased temperature resistance and overall resilience of the traditional silicone over the soft silicone. Figure 4.5 and Figure 4.6 clearly show that the plug thickness has an inverse relationship to the release rate. As the plug thickness increased, the release rate decreased. The amount by which the release rates changed between the different thicknesses by plug material, displayed in Table 4.8, proved to be intriguing. The lengthening of the plug from 0.635 cm (0.25 in) to 1.270 cm (0.50 in) caused an approximate 40% reduction in release rate for both materials. However, the lengthening of the plug from 1.270 cm (0.50 in) to 1.905 cm (0.75 in) caused an approximate 30% reduction in release rate for both materials. This result suggests that the relationship between plug thickness and release rate may not be linear. A full statistical verification of this prospect was beyond the scope of this preliminary study and was not examined. The plug thickness was also found to have a direct relationship with PMCH equilibrium time. Equilibrium time for this study is defined as the time required for PMCH to completely absorb into the plug and achieve a stable release rate to atmosphere. As the plug thickness increased, the equilibrium time, as seen in Figure 4.5 and Figure 4.6, also increased. Thus, the plug thicknesses were found to only have an effect on the PMCH release rate and vessel equilibration time but not on the overall reliability of the sources. This preliminary study successfully determined the relationship between modifying several design variable on the PMCH release sources based on the experimental parameters. The post- filling technique was found to be less robust than the pre-filling technique. Although the pre- filling technique is more difficult to execute, this method is preferred for its increased reliability. The type of material and plug thickness were found to both affect the magnitude of the release 77
Virginia Tech	rate but not the overall reliability of the source. The plug thickness was also found to have an effect on the PMCH equilibration time of each vessel. As can be seen by the aforementioned discussion, the study successfully produced and evaluated several PMCH release source designs. The results of this study will be used to develop a more intricate study to derive an equation for the release rate of the PPRV as a function of environmental conditions and significant design variables. 4.6 Acknowledgements The authors wish to thank Dr. Russell Dietz, Brookhaven National Laboratory, who has been generous with his time and knowledge of perfluorocarbon tracers and Rudy Maurer Jr., Dyno Nobel, who has been instrumental in supplying the aluminum shells used in this experiment. This publication was developed under Contract No. 200-2009-31933, awarded by the National Institute for Occupational Safety and Health (NIOSH). The findings and conclusions in this report are those of the authors and do not reflect the official policies of the Department of Health and Human Services; nor does mention of trade names, commercial practices, or organizations imply endorsement by the U.S. Government. 78
Virginia Tech	Chapter 5: An evaluation of a perfluoromethyl- cyclohexane (PMCH) permeation plug release vessel (PPRV) in a controlled turbulent environment ABSTRACT: The use of sulfur hexafluoride (SF ) as a tracer gas for analyzing underground 6 mine ventilation systems has been practiced for over 30 years. As a result, the methods used to release, sample, and analyze SF are well accepted. Although improvements are still being made 6 to enhance the analysis of this tracer, the overall technique remains largely the same. However, as the complexity and size of underground mine ventilation networks increase, coupled with steadily rising SF background levels, the ability of a single gas to function as a convenient, rapid 6 means of analysis diminishes. The utilization of multiple tracer gases can mitigate these problems by allowing for a more comprehensive evaluation using multi-zone techniques. Although multiple tracer gases have already been extensively used in the fields of heating, ventilation, and air conditioning (HVAC) and large-scale atmospheric monitoring, this technique has not been widely implemented in underground mine ventilation. This lack of use is partially due to the difficulty of releasing the other available tracer. A well-documented alternative in HVAC studies to SF as a tracer are perfluorocarbon tracers (PFT). Many PFTs exist as volatile 6 liquids at room temperature and pressure. This characteristic prevents a PFT from being released using the same technique as SF , which exists as a gas under the same conditions. This paper 6 evaluates a permeation plug release vessel (PPRV) under controlled turbulent conditions. The PPRV used in this study is designed to deploy PMCH. The details of the experimental parameters used in this evaluation are also presented in this paper. 82
Virginia Tech	5.1 Introduction The use of sulfur hexafluoride (SF ) as a tracer gas for modeling underground mine ventilation 6 systems has been practiced for over three decades (Thimons, Bielicki et al. 1974). As a result, the methods used to release, sample, and analyze SF are well defined and accepted. Although 6 improvements are still being made to enhance the analytical tools for detecting this tracer, the overall technique remains largely the same across the mining industry. The basic method consists of the following steps: SF is first released either continuously or in a single pulse depending on 6 the requirements of the study. Mine air is then sampled at pre-selected locations either on an individual, manual basis or on a continuous, automated basis. The air samples are then examined to determine the concentration of the tracer gas at the sampling points (Hartman, Mutmansky et al. 1997). These data can then be used to characterize the ventilation flow within the target area. Although the use of SF is a time proven technique, it is becoming increasingly difficult to 6 implement it in the modern day. This difficulty largely stems from the increasing complexity of underground mines. The advancement of mining technologies and techniques has allowed for larger and more intricate mine designs. As mines grow in this manner, so must the ventilation networks that support them. This escalation has caused the design, maintenance, and evaluation of underground ventilation networks to become increasingly cumbersome. As a result, many traditional ventilation engineering evaluation practices are becoming either inefficient or obsolete. The ability of SF to function as a convenient, rapid means of ventilation network analysis is 6 diminished by the continual expansion of underground mines. This problem arises from a combination of multi-zone analysis difficulties, increasing natural background of SF , and the 6 sheer amount of SF needed to complete a full ventilation investigation. A standard tracer gas 6 examination requires that SF be released either as a constant stream until its concentration is 6 uniform at the release point and the outlet (i.e. conservation of mass) or as a large pulse. This single tracer method is very effective at characterizing a single flow volume but cannot be used to show interaction between independent ventilation branches, which constrains the scope of a tracer gas study. The natural background presence of SF has also been steadily increasing since 6 83
Virginia Tech	1953. High background equates to lower single to noise ratios and thus lower detection sensitivity (Ravishankara, Solomon et al. 1993, Maiss and Brenninkmeijer 1998, Levin, Naegler et al. 2010). In addition, both release methods require a significant amount of tracer to achieve detectable levels in large-scale investigations. The artificially increased background presence of SF resulting from these large releases creates some additional challenges. 6 In order to conduct multiple consecutive release tracer gas studies, sufficient time must be allowed between tests to reduce the background presence of the gas. The reduction of background presence is essential so that subsequent tests of the ventilation system are not affected by the residual concentration of the previous release. As the concentrations generally captured at sampling locations, ranging from parts per million (PPM) to parts per trillion (PPT), even minute amounts of additional tracer gas can drastically affect the final analysis. Unfortunately, this problem is more pronounced with SF due to its natural tendency to adhere to 6 surfaces. SF is also approximately five times heavier than air and causes the gas to linger in 6 areas of low ventilation flow. As a result, a significant amount of time must be allowed to ventilate the excess tracer gas prior to subsequent releases (Thimons, Bielicki et al. 1974). The utilization of additional tracer gases has the potential to mitigate these problems in underground mine ventilation surveys. The ability to deploy multiple tracer gases not only allows for consecutive releases that are free of cross-contamination, but can also facilitate simultaneous, multi-location releases. These added benefits will facilitate a more rapid, comprehensive evaluation of ventilation systems as well as provide insightful information regarding the interaction of independent ventilation streams. Multiple tracer gases have already been extensively used in the fields of heating, ventilation, and air conditioning (HVAC) as well as in large-scale atmospheric monitoring. For example, a well- documented alternative in HVAC studies to SF are perfluorocarbon tracers (PFT) (Dietz and 6 Cote 1982, Dietz 1991, Grot, Lagus et al. 1995, Heiser and Sullivan 2002). This classification of tracer gas, more specifically the gas perfluoromethylcyclohexane (PMCH), will be the main topic of discussion in this paper. 84
Virginia Tech	5.2 Background Perfluoromethylcyclohexane (PMCH) is a volatile liquid that is part of the perfluorocarbon tracer (PFT) compound group. These compounds are composed of perfluoroalkanes which are group of substances that are biologically inert, chemically inert, and thermally stability. The inert, non- reactive, and non-toxic nature of PFT compounds makes them ideal choices as alternative tracer gases to SF . Although PMCH can potentially be complimented by other PFTs, these other 6 tracers have not yet been successfully deployed in conjunction with SF in underground mines 6 (Sandford 2003). The chemical formula of PMCH is C F and has a molecular weight of 350 g/mol with a boiling 7 14 point of 76°C. This compound, as previously stated, exists as a volatile liquid at room temperature. The volatile nature of PMCH allows it to vaporize even at relatively low temperatures. Once in a vapor state, PMCH will remain a vapor even through cooler temperatures (Dietz 1986, National Institute of Standards and Technology 2011). PMCH, similar to other tracer gases, can be used as a tracer because it does not adversely affect people or the environment. Additionally, the low ambient background of PMCH in the atmosphere, with concentrations in the parts per quadrillion (PPQ), classifies this compound as having no significant natural presence (Watson, Wilke et al. 2007). Thus, PMCH’s background presence is effectively zero. Perhaps the greatest advantage afforded by PMCH is its ability to be separated from SF on a single column. Despite these advantages, PMCH has not yet been implemented in 6 underground mine ventilation. However, this tracer, along with other PFT group compounds, has been extensively deployed in other fields. Building ventilation is an area of study that has significantly implemented PFTs. This heavy utilization of PFTs is a result of the inherent limitations present in more conventional ventilation analysis techniques when measuring low air flows and leakages. PFTs have been used in extensive studies to investigate air infiltration into single family homes using passive liquid PFT permeation sources coupled with passive adsorption tube samplers (Dietz and Cote 1982, Leaderer, Schaap et al. 1985) as well as to evaluate the performance of multi-zone deployments of passive PFT sources for categorizing air infiltration, air exfiltration, and air exchanges (Dietz, 86
Virginia Tech	Goodrich et al. 1986). PFTs have also been used to evaluate ventilation rates in Swedish housing stock to determine the adequacy of the ventilation systems (Stymne, Bowman et al. 1994) and to characterize down-valley flow, canyon outflow, and interacting circulations at the lower slopes of the Wasatch Front (Fast, Allwine et al. 2006). The previously introduced examples of PFT applications demonstrate the versatility of this compound as a tracer. Presently, only PMCH, a PFT group compound, has been successfully separated from SF in preliminary mine ventilation research. This tracer has not yet been 6 deployed in underground field studies. This paper seeks to evaluate the predictability and reproducibility of two potential deployment methods designed to release liquid PMCH as a vapor from its liquid state under controlled conditions. 5.3 Experimental Design The study presented in this paper will evaluate the deployment of a PPRV developed in a previous experiment. This experiment was completed in three main parts: sample PMCH released by the PPRV from a controlled turbulent environment, generate the calibration curve, and perform GC analysis. The following sections detail the experimental parameters of this study. 5.3.1 Release Sources The main body of the release vessel is comprised of an aluminum cylinder that is 6.35 cm (2.50 in) in length with an outside diameter (OD) of 0.747 cm (0.294 in). The cylinder itself is a hollow shell with one end sealed and the other end opened to atmosphere. The shell has a wall thickness of 0.0356 cm (0.0140 in) that is consistent throughout the body of the cylinder, which gives an inside length of 6.31 cm (2.49 in) and an inside diameter (ID) of 0.711 cm (0.280 in). A schematic of the aluminum shell is displayed in Figure 5.1. 87
Virginia Tech	Three PPRVs with a total PMCH release rate of 2.97∙10−8 g/s were placed in parallel inside a container designed to create a turbulent, blowing-type airflow environment from which the PMCH will be sampled. 5.3.2 Turbulent Flow Container The vessel consisted of three main parts: the inlet, the PMCH release area, and the outlet. The inlet was built using a nylon 1.27 cm (0.5 in) male NPT to 0.635 cm (0.25 in) female NPT reducing bushing and a 0.635 cm (0.25 in) hose barb. The hose barb accepted airflow through a flexible plastic hose from two air pumps connected in parallel. The combined pumping system was capable of creating flow quantities greater than 100 LPM. The flow from the pump was controlled by an adjustable flow controller placed in-line between the air pump and the turbulence container. The inlet directed airflow to the PMCH release area. A schematic diagram and of the flow induction system can be seen in Figure 5.3 and Figure 5.4. Figure 5.3. Schematic diagram of turbulent flow generation system. 89
Virginia Tech	Figure 5.4. Picture of actual flow generation system used in this study. The release area was constructed using a plastic cylinder with a diameter of 14.60 cm (5.749 in) and a height of 30.27 cm (11.918 in). The bottom of the cylinder was sealed while the top was left opened to accept a removable, airtight cover. The removable cover allowed for the placement and removal of the PPRVs. The outlet was connected to the PMCH release area and opened to the atmosphere. The air stream exhausted into a fume hood to eliminate any contamination resulting from recirculation. The outlet consisted of a plastic pipe with an internal diameter (ID) of 1.58 cm (0.622 in) and a length of 15.54 cm (6.118 in). A sample port was connected perpendicular to the outlet using a T-interface. The sample port, which was 3.782 cm (1.489 in) in length, was attached using the same pipe as the outlet. The port was designed to accept a rubber septum to allow sampling of the exhaust stream using either an evacuated sample vial or a syringe. The entire vessel was designed to be airtight between the inlet and the outlet so that the PMCH stream was neither lost nor contaminated from external sources during sampling. Detailed schematics and a picture of the container can be seen in Figure 5.5 through Figure 5.8. 90
Virginia Tech	The PPRVs were exposed to five different flow quantities representing transitional and turbulent flows. The five flow quantities were 30 LPM, 40, LPM, 50 LPM, 60 LPM, and 70 LPM. Table 5.1 displays each flow quantity paired with their respective Reynold’s number. The Reynold’s numbers were determined at each quantity using a dynamic viscosity of 1.983·10-5 N·s/m2, a hydraulic diameter of 0.0158 m based on the outlet, and a dry air density of 1.225 kg/m3. Table 5.1. Flow quantities with associated Reynold’s numbers. Flow Quantity Reynold's (LPM) Number Flow Type 30 2,759 Transitional 40 3,678 Transitional 50 4,598 Turbulent 60 5,517 Turbulent 70 6,437 Turbulent Each flow quantity was applied to the PPRVs in a random time order and replicated four times to determine release rate variability. Two lab technicians were used to apply the different flow quantities as well as take tracer samples from the turbulence container. In order to account for any variance caused by the technicians, a generalized randomized complete blocking design (GRCBD) was implemented in this study. Each lab assistant was randomly assigned two replicates of each flow quantity. A summary of the GRCBD is provided in Table 5.2 and Table 5.3. Table 5.2. Treatment randomization by investigator. Flow Quantity Investigator 1 Investigator 2 30 6 10 30 9 5 40 3 8 40 10 6 50 4 2 50 5 1 60 7 3 60 1 4 70 8 7 70 2 9 93
Virginia Tech	Table 5.3. Treatment randomization by flow quantity (LPM). Order Investigator 1 Investigator 2 1 60 50 2 70 50 3 40 60 4 50 60 5 50 30 6 30 40 7 60 70 8 70 40 9 30 70 10 40 30 5.4 Gas Chromatography 5.4.1 Equipment The analysis of PMCH was completed using a GC equipped with an electron capture detector (ECD). The ECD was chosen due to its high detection sensitivity to electronegative compounds (McNair and Miller 2009). The GC was fitted with a 30 m long, 0.25 mm ID porous layer open tube (PLOT) alumina chloride capillary column (HP-AL/S). The HP-AL/S column was deactivated with sodium sulfate and had a film thickness of 5 µm. This column was selected based on its ability to separate perfluorinated compounds. 5.4.2 Calibration Curve The generation of a calibration curve is a necessary aspect of quantitative GC investigations due to the equipment-specific nature of GC technology. Although GCs are manufactured using identical principles, each device’s detector response is unique and varies with installation parameters (e.g. length of column, composition of the stationary phase, composition of the mobile phase, etc.) and device configurations (e.g. linear velocity of the mobile phase, temperature of the column, etc.). Thus, a calibration curve using known standards must be created for each analysis to serve as a reference for known sample types with unknown concentrations (McNair and Miller 2009). 94
Virginia Tech	The concentration range of the PMCH calibration curve was defined by the expected concentration of PMCH at the sampling point. As previously stated, the sources were exposed to a range of air quantities from 30 LPM to 70 LPM in 10 LPM increments. The expected range of sampling concentrations can be determined using the low bound and high bound quantities of 30 LPM and 70 LPM respectively. These two points represent the highest and lowest possible concentrations of PMCH at the sampling point thus defining the extremes of the calibration curve. The expected concentration was determined using the following assumption: the turbulence vessel is airtight between the inlet and the outlet thus establishing 𝑄 = 𝑄 . As a 𝑖𝑛 𝑜𝑢𝑡 result, the volumetric flow rate at the inlet, which was set by a flow controller, is the same as the volumetric flow rate at the sampling point within the outlet. The minimum and maximum volumetric flow rates were then used in conjunction with the mass flow rate of the release sources to compute the expected concentration range of PMCH at the sampling point. The additional volumetric flow produced by the three PPRVs in parallel was assumed to have an insignificant impact on the overall volumetric flow of air. The mass flow of PMCH, as previously introduced, was expected to be 1.78∙10−6 g/min. The expected PMCH concentration using an upper bound volumetric flow rate of 70 LPM and a lower bound volumetric flow rate of 30 LPM was determined as follows. 𝑔 𝑘𝑔 1.78∙10−6 𝑃𝑀𝐶𝐻 ∙ 𝑚𝑖𝑛 1000 𝑔 𝑘𝑔 𝑃𝑀𝐶𝐻 = = 2.54∙10−8 𝑃𝑀𝐶𝐻 𝑚𝑖𝑛 𝐿 𝑚3 𝑚3 70 𝑎𝑖𝑟 ∙ 𝑎𝑖𝑟 𝑚𝑖𝑛 1000 𝐿 𝑔 𝑘𝑔 1.78∙10−6 𝑃𝑀𝐶𝐻 ∙ 𝑚𝑖𝑛 1000 𝑔 𝑘𝑔 𝑃𝑀𝐶𝐻 = = 5.93∙10−8 𝑃𝑀𝐶𝐻 𝑚𝑎𝑥 𝐿 𝑚3 𝑚3 30 𝑎𝑖𝑟 ∙ 𝑎𝑖𝑟 𝑚𝑖𝑛 1000 𝐿 The boundaries of the concentration curve must contain a concentration range of at least these calculated values. The PMCH concentrations can also be represented in PPB by moles (PPBM). This conversion can be accomplished by translating the volumetric flow of air and the mass flow 95
Virginia Tech	of PMCH to molar flows. The molar flow of air is determined assuming a dry air density of 1.205 g/L. 70 𝐿 1.205 𝑔 𝑔 ∙ = 84 𝑚𝑖𝑛 𝐿 𝑚𝑖𝑛 30 𝐿 1.205 𝑔 𝑔 ∙ = 36 𝑚𝑖𝑛 𝐿 𝑚𝑖𝑛 The mass flow can then be converted to a molar flow using a molecular weight of 28.97 g/mol for dry air. 84 𝑔 𝑚𝑜𝑙 𝑚𝑜𝑙 ∙ = 2.9 𝑚𝑖𝑛 28.97 𝑔 𝑚𝑖𝑛 36 𝑔 𝑚𝑜𝑙 𝑚𝑜𝑙 ∙ = 1.2 𝑚𝑖𝑛 28.97 𝑔 𝑚𝑖𝑛 The previous calculations for the molar flow of air was completed using the density and molecular weight of dry air. This molar flow does not represent the exact flow in the experiment due to the possibility of varying atmospheric conditions in the laboratory, such as humidity, temperature, and barometric pressure. A more accurate representation of the expected concentrations can be completed using psychometric properties and the idea gas law. However, the purpose of the previous calculation was to provide an initial point or reference from which a calibration curve could be generated. If the calculated boundaries did not provide an adequate range for interpolation, the calibration standards would simply need to be adjusted relative to the existing points. The molar flow of PMCH was computed as follows using the previously determined mass flow of 1.78∙10−6 g/min and a molecular weight of 350.05 g/mol. 96
Virginia Tech	1.78∙10−6 𝑔 𝑚𝑜𝑙 ∙ = 5.08∙10−9 𝑚𝑜𝑙 𝑚𝑖𝑛 350.05 𝑔 𝑚𝑖𝑛 The previously derived values can now be used to find the ratio of the molar flow of PMCH to the total molar flow, M, of PMCH and air. This ratio represents the molar fraction of PMCH in the total flow. The molar fraction can then be multiplied by 109 to convert the ratio into a PPBM. 5.08∙10−9 𝑚𝑜𝑙 𝑃𝑃𝐵𝑀 = 𝑚𝑖𝑛 ∙109 = 1.8 𝑃𝑃𝐵𝑀 𝑚𝑖𝑛 5.08∙10−9 𝑚𝑜𝑙 +2.9 𝑚𝑜𝑙 𝑚𝑖𝑛 𝑚𝑖𝑛 5.08∙10−9 𝑚𝑜𝑙 𝑃𝑃𝐵𝑀 = 𝑚𝑖𝑛 ∙109 = 4.1 𝑃𝑃𝐵𝑀 𝑚𝑎𝑥 5.08∙10−9 𝑚𝑜𝑙 +1.2 𝑚𝑜𝑙 𝑚𝑖𝑛 𝑚𝑖𝑛 Based on the previous computations, a target concentration range of 0.15 PPBM to 23 PPBM was selected to encompass the range of sampling concentrations as well as to allow for any fluctuations in PMCH release due to changing temperatures in the laboratory. 5.4.3 Standard Preparation The creation of gaseous PMCH standards from a pure liquid was completed by following the three step process described below. 1. A master liquid standard was prepared by diluting 90% technical grade PMCH with a solvent. 2. A defined aliquot of the master standard was transferred to another vial for further solvent dilution. 3. A defined volume of the diluted master standard was then transferred to a headspace container and allowed to vaporize to create a gaseous standard. Table 5.4 displays the dilution scheme for the five calibration standards. 97
Virginia Tech	5.5.1 Deployment The PMCH PPRVs were first placed in the release area of the turbulence container. The container was then capped as quickly as possible and sealed with silicone grease to prevent contamination. The PPRVs were allowed to release PMCH into the flow stream for five minutes at the first assigned flow rate. The temperature of the laboratory was monitored to ensure that no significant fluctuations occurred during sampling. After the five minutes had elapsed, a 100 L gas tight syringe was used to extract 30 L of air from the sampling port. The 30 L sample was then immediately injected into the GC for analysis. Two additional consecutive samples were taken and injected in the same manner. All of the replicates for each flow quantity were taken in triplicate (i.e. a total of 60 individual samples). Two minutes were allotted between replicates to allow the PPRVs to equilibrate with each new flow setting. In order to accurately determine the expected concentration of PMCH in the flow stream, the environmental conditions were also recorded. Over the course of the experiment, the turbulence container experienced an average temperature of 295 K and an average absolute barometric pressure (i.e. station pressure) of 94,000 Pa. Using these values and assuming ideal gas behavior, the expected concentration in parts per billion by volume (PPBV) is presented in Table 5.6. Table 5.6. Expected PMCH concentrations at each flow quantity. Air Quantity PMCH (LPM) (PPBV) 30 4.40 40 3.30 50 2.64 60 2.20 70 1.88 100
Virginia Tech	Table 5.13 displays the deviation between the measured PMCH concentration from the sample port and the expected concentrations based on the release rate of the PPRV at the associated flow quantity. Table 5.13. Expected PMCH concentration compared to measured PMCH concentration. Average Expected Measured Difference in Air PMCH PMCH PMCH Quantity Concentration Concentration Concentration % (LPM) (PPB) (PPB) (PPB) Error 30 4.40 5.93 1.54 34.94 40 3.30 3.85 0.55 16.80 50 2.64 3.25 0.61 23.19 60 2.20 2.70 0.50 22.71 70 1.88 2.34 0.46 24.43 5.7 Discussion and Conclusions The study presented in this paper evaluated the potential of a PPRV assembled by the authors for use as an underground mine ventilation tracer deployment medium. The PPRV consists of a hollow aluminum cylinder with one end opened to the atmosphere and with the other end sealed. Liquid PMCH was stored in the vessel and released over time through a silicone plug pressed flush to the end. The PMCH releases to the atmosphere from the plug at a linear rate directly proportional to temperature. Three release sources were placed at a constant temperature and in a controlled turbulent flow stream to determine their performance under these conditions. A range of air flow quantities from 30 LPM to 70 LPM in increments of 10 LPM was induced to test the predictability and repeatability of the PPRVs. These quantities represent both turbulent and transitional type flows to determine their effect, if any. The PMCH concentration in the flow stream was determined by extracting air samples and then processing those samples through a GC. A detailed description of the experimental design can be found in Section 5.3. The PMCH concentration measured across each of the replicate is plotted against flow quantity in Figure 5.10. A line representing the expected concentration of PMCH at each of the flow quantities is also displayed. The tight grouping of the different replicates at each of the flow 106
Virginia Tech	quantities graphically demonstrate high precision at the majority of the sample points. The high sampling precision of both technicians are further demonstrated by the low RSD values. Figure 5.10 also shows that as the flow quantity enters the transitional zone (i.e. 30 LPM and 40 LPM), the precision of the replicates begins to decrease. This behavior is expected as the molecular weight of PMCH would require a fully developed turbulent flow to facilitate homogeneous mixing of the tracer. Transitional flow displays both laminar and turbulent flow characteristics. As a result, the PMCH shifts between being homogeneous and heterogeneous mixed in the stream. The ANOVA results displayed in Table 5.11 show that there is sufficient evidence to suggest that the change in concentration based on peak area was significantly affected by the application of the different air quantities using an alpha value of 0.05. This result also suggests that the release rate of the PPRVs stayed relatively uniform amongst the replicates because a significant difference was only found at the different treatment levels. With a p-value > 0.05, the ANOVA analysis concludes that no significant block-treatment interaction was demonstrated. Thus, any variability in the results caused by the technicians themselves was insignificant and did not affect the results. The expected PMCH concentration are displayed in Table 5.6. These values were determined using the velocity of the volumetric flow of air and the release rate of the sources at 21°C. The release rate of the three PPRVs were recorded in a previous experiment. The difference between the expected concentrations and actual concentrations at each of the flow quantities are shown in Table 5.13. Although the errors are on average about 20%, this value only represents a 0.50 PPB shift in concentration. Additionally, given the consistency in the concentration difference, the parallel response of the PPRVs as compared to the expected response, as well as the high sampling and analysis precision, this error is most likely present in the calibration curve. Even so, a 20% deviation based on a calibration curve derived from multiple PPB level serial dilution standards does not represent a significant error. The low error parallel response, high precision, and highly repeatable results of the analyzed samples demonstrate that the PPRVs did release PMCH at a constant rate. The PPRVs were also 107
Virginia Tech	Chapter 6: Development of a perfluoromethylcyclohexane (PMCH) permeation plug release vessel (PPRV) for tracer gas studies in underground mines ABSTRACT: Perfluoromethylcyclohexane (PMCH) is a member of the perfluorocarbon tracer (PFT) group of compounds. PMCH has shown to be a viable alternative to the widely used tracer gas sulfur hexafluoride (SF ). This viability stems from the fact that PMCH can be used 6 concurrently with SF while maintaining adequate chromatographic separation and high 6 detection sensitivity during gas chromatographic (GC) electron capture (EC) analysis. However, the release of PMCH in an underground mine ventilation system is challenging due to its physical characteristics. SF exists as a gas at room temperature and pressure and can be 6 accurately released using a variety of means. In contrast, PMCH exists as a volatile liquid at room temperature and pressure, a characteristic that prevents PMCH from being deployed using traditional means. This paper presents a design for a permeation plug release vessel (PPRV) for PMCH. The PPRV is designed to passively deploy PMCH vapor at linear rate as a function of temperature and plug thickness. 111
Virginia Tech	6.1 Introduction Sulfur hexafluoride (SF ) has been the predominant tracer gas used in underground mine 6 ventilation studies for over 30 years (Thimons, Bielicki, and Kissell 1974). However, the ability of SF to function as the sole tracer is being hindered by two main issues: the increasing scale of 6 mine ventilation systems and the steadily growing background concentration of SF in the 6 atmosphere. In order to support the advance of underground mines, a sympathetic expansion of their ventilation systems must also occur. The sheer scale and complexity modern mines has diminished the analysis power of a single tracer gas. As a result, the effectiveness of SF as a 6 mine ventilation characterization tool has begun to decrease. This issue is compounded by the growing background concentration of SF in the atmosphere. 6 One of the main characteristics of a tracer gas is low background presence. This property not only reduces the possibility of interference but also allows for the characterization of low velocity flows, such as leakage through a barrier. As the background presence of a tracer increases, the release amount must follow suit in order to maintain an acceptable signal to noise ratio. This increase is the result of two main processes: the popularity of SF as an electrical 6 insulator and the high atmospheric stability of the SF molecule. 6 SF is widely used to insulate dielectric switchgears and transformers due to its inert nature and 6 high electron affinity. Since its initial industrial production in 1953, the atmospheric concentration of SF has increased by two orders of magnitude. This accumulation of SF is the 6 6 result of its stability, which equates to an atmospheric lifetime of 3,200 years (Levin et al. 2010, Geller et al. 1997, Ravishankara et al. 1993, Maiss et al. 1996). In order to mitigate the issues of larger ventilation systems and increased background concentrations, recent studies have identified the compound perfluoromethylcyclohexane (PMCH) as a viable supplement for SF . 6 PMCH is classified as a perfluorinated cyclic hydrocarbon. Other compounds in this group classification include perfluoromethylcyclobutane (PMCB) and perfluoromethylcyclopentane (PMCP) (Galdiga and Greibrokk 1997). These compounds are also known as perfluorocarbon tracers (PFT) due to their chemical inertness, low toxicity, and trace level environmental 112
Virginia Tech	background presence thus uniquely suiting them for use as tracer gases (Dietz 1991, Watson et al. 2007). Compounds of this type have been widely implemented in heating, ventilation, and air conditioning (HVAC) as well as atmospheric monitoring studies (Dietz 1991) but not yet in the field of underground mine ventilation. In contrast to SF , PFTs exist as volatile liquids at room 6 temperature and pressure (National Institute of Standards and Technology 2011). This physical property of PFTs removes one of the great advantages afforded by SF , its existence as a gas. As 6 such, SF can be purchased in convenient, standardized gas cylinders and released in a controlled 6 manner using a variety of options (Thimons, Bielicki, and Kissell 1974). Two examples of release options are the flow meter method and the flow controller method. The flow meter method utilizes an analog or a digital flow meter attached to the gas cylinder. SF 6 is then deployed at a pre-determined volumetric flow rate over time, which can be computed to a mass flow rate over time if necessary. This method offers a significant amount of control and flexibility. The flow controller method utilizes a sophisticated electronic regulator that allows for precise deployment of SF either by mass or by volume depending on the controller. This method 6 provides the greatest accuracy, reproducibility, and control of any readily available release approaches. A controlled tracer release is essential if quantitative data must be generated. However, the established release systems for SF are not applicable to volatile liquids like 6 PMCH. This paper presents a design for a permeation plug release vessel (PPRV) for PMCH. The PPRV passively deploys PMCH vapor at linear rate as a function of temperature and plug thickness. An equation to predict the release rate based on these two variables is provided. Details regarding the PPRV and experimental design are also discussed. 6.2 Background PMCH is a perfluorinated cyclic hydrocarbon whose chemical structure is composed of perfluoroalkanes (Watson et al. 2007). Compounds of this type are biologically inert, chemically inert, and thermally stable (F2 Chemicals Ltd. 2011). The inert, non-reactive, and non-toxic nature of PMCH makes it an ideal choice as a tracer gas. PMCH is comprised of seven carbon atoms and fourteen fluorine atoms, which gives it a chemical formula of C F . The molecular 7 14 structure of PMCH is displayed in Figure 6.1. 113
Virginia Tech	Figure 6.1. Molecular structure of PMCH. The PMCH molecule is composed of two main parts, which are the cyclohexane ring and the methyl group bonded off to the side. This fully fluorinated molecule has a molecular weight of 350 g/mol and a boiling point of 76°C (169°F). PMCH to exist as a liquid at room temperature and pressure due to its molecular weight. The high volatility of PMCH simultaneously allows it to vaporize even at low temperatures. Once in a vapor state, PMCH, will remain a vapor even through cooler temperatures (National Institute of Standards and Technology 2011). Another advantage of PMCH is its detectability by GC even at low concentrations. This ability stems from PMCH’s low ambient background in the atmosphere with concentrations in the low parts per quadrillion (PPQ) (Cooke et al. 2001, Simmonds et al. 2002, Watson et al. 2007) and high electron capture (EC) detection sensitivity (Simmonds et al. 2002). Perhaps the greatest advantage afforded by PMCH is its ability to be simultaneously analyzed with SF , which is the 6 standard tracer gas employed in the mining industry. Despite these tracer characteristics, PMCH has not yet been implemented in underground mine ventilation. PMCH, however, has seen widespread use in other fields of study in conjunction with other PFT group compounds. Building ventilation is one area that has implemented PFTs. PFTs have been used to investigate air infiltration into single family homes using passive PFT permeation sources coupled with passive capillary adsorption tube samplers (CATS) (Leaderer, Schaap, and Dietz 1985, Dietz and Cote 1982) as well as to evaluate the performance of multi-zone deployments of passive PFT sources for categorizing air infiltration, air exfiltration, and air exchanges (Dietz et al. 1986). PFTs have also been used to evaluate ventilation rates in Swedish housing stock as a nation-wide 114
Virginia Tech	effort to determine the adequacy of agricultural ventilation systems (Stymne, Bowman, and Kronvall 1994). In the field of atmospheric tracing, PFTs have been used to characterize down-valley flow, canyon outflow, and interacting circulations on the lower slopes of the Wasatch Front (Fast et al. 2006). They have also been deployed to evaluate air flow patterns in New York City as part of the Urban Dispersion Program to improve wind station placements, to supplement knowledge of contaminant flow patterns, and to update atmospheric flow models (Watson et al. 2006). PFTs have additionally been used in long-term, large-scale investigations of the transport and diffusion of gases over the Alpine topography in Switzerland (Ambrosetti et al. 1998) as well as to evaluate the accuracy of meteorological air quality models under the Metropolitan Tracer Experiment conducted in Washington, D.C. (Draxler 1967). The aforementioned examples of PMCH applications demonstrate the versatility of this compound as a tracer. Thus, PMCH demonstrates great potential for use as an underground mine ventilation tool. In order to successfully deploy PMCH, the release source must not only be able to perform a controlled release but also withstand the environmental conditions in a mine environment, such as dust and water. This paper presents a design for a permeation plug release vessel (PPRV) for PMCH. The PPRV is designed to passively deploy PMCH vapor at linear rate as a function of temperature and plug thickness. Details regarding the design and execution of the development process are also provided. The basic concept of the PPRV detailed in this study was initially introduced by Brookhaven National Laboratory (BNL). In order to provide a cursory understanding of how the PPRV source operates, the gas diffusion mechanism must be discussed. The controlled release of PMCH is facilitated by the permeability characteristics of silicone rubber. The passage of a gas through rubber-type mediums such as silicone is a well-documented phenomenon that has undergone extensive study for over 50 years (Barbier 1955, Hammon, Ernst, and Newton 1977, Jordan and Koros 1990, Stern, Onorato, and Libove 1977, van Amerongen 1946, Zhang and Cloud 2006). In silicone rubber, similarly to other rubber-type polymers, gas diffusion occurs in three distinct steps: solution of the gas molecules on one side of the silicone membrane, diffusion 115
Virginia Tech	of the gas molecule through the silicone, and evaporation of the gas from the other side (Barbier 1955, Zhang and Cloud 2006). This diffusion through a seemingly impermeable medium can be achieved due to the chemical composition of silicone rubber. Silicone, or polysiloxane, is a name used to define any compound derived from polymerized siloxanes. Polymerized siloxanes are substances whose molecular structure is created by combining monomers into large chains of alternating silicon (Si) and oxygen (O) atoms. The alternating atoms (e.g. Si-O) have organic groups or hydrogen atoms bonded to the Si atom (Merriam-Webster 2012, Bondurant, Ernster, and Herdman 1999, Velderrain and Lipps 2011, Van Reeth and Wilson 1994, Encyclopædia Britannica 2012). Silicone generally has two methyl groups attached to each siloxane thus producing a polydimethylsiloxane (PDMS) (Bondurant, Ernster, and Herdman 1999, Van Reeth and Wilson 1994). An example of the structural formula for PDMS is displayed in Figure 6.2. Figure 6.2. Chemical structure of polydimethylsiloxane (PDMS), the basic structure of silicone rubber. Silicone is typically composed of a long chain, from hundreds to thousands, of repeating PDMS groups thus classifying silicone as a linear polymer (Bondurant, Ernster, and Herdman 1999). In order to form silicone rubber, or silicone elastomer, the polymer chains must be cross-linked through vulcanization. In order to produce silicone rubber, the basic PDMS chemical structure must be slightly altered to facilitate cross-linking (Andriot et al. 2007, Parker Hannifin Corp. 2007, Velderrain and Lipps 2011). Methyl vinyl silicone (VMQ) is the most common type of silicone rubber. VMQ has the same overall structure as PDMS with some of the methyl groups replaced with vinyl groups to create 116
Virginia Tech	the reactive double bond necessary for cross-linking. The structure of the VMQ molecules, which is represented in Figure 6.2, allows free rotation around the Si-O bond (Andriot et al. 2007, Parker Hannifin Corp. 2007, Velderrain and Lipps 2011, Van Reeth and Wilson 1994). This chemical property allows the rubber to be highly flexible. This inherent flexibility afforded by the Si-O bond also permits the existence of free volumes within the compound thus allowing for gas permeability. Incidentally, silicone rubber also has the highest permeability of any polymer, which ideally suits this compound for use in the PMCH release source (Zhang and Cloud 2006, Andriot et al. 2007, Parker Hannifin Corp. 2007, Van Reeth and Wilson 1994). The basic PPRV design consists of a hollow aluminum cylinder with one end of the cylinder open to atmosphere and the other end closed. Liquid PMCH is injected into the aluminum vessel and sealed with an oversized silicone plug pressed flush to the end. The source releases PMCH by allowing the phase change to occur internally. Vapor is produced immediately once the PMCH is exposed to atmosphere due to its high vapor pressure (106 torr at 25°C) (Dietz et al. 1986). The high flow resistance caused by the silicone plug produces a pseudo-closed system that allows the PMCH to reach dynamic equilibrium within the source. This equilibrium produces a steady pressure differential equivalent to the vapor pressure at the ambient temperature between the inside and outside of the vessel. This differential causes the vapor PMCH to slowly diffuse steadily through the silicone plug. This steady release rate result from the combination of a constant pressure front coupled with the fact that the silicone plug can be considered as a constant resistance medium. The permeability of the silicone will remain constant as long as the integrity is maintained and the compression of the plug is not excessive enough to significantly reduce the internal free volumes (Jordan and Koros 1990). Once the PMCH equilibrates within the plug, vapor PMCH begins desorbing to atmosphere from the plug at a predictable rate directly proportional to temperature (Dietz et al. 1986). Since vapor pressure is directly proportional to temperature and independent of atmospheric pressure, the release rate will remain constant at a stable temperature. The PPRV presented in this paper is based on the overall concept of the BNL source but includes a manner to predict the release rate using VMQ silicone plug thicknesses. A discussion regarding the expected performance of the PPRV over time is also provided. 117

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.