USF Libraries
USF Digital Collections

Hydrogen fluoride method development for the Ogawa Passive Sampling Device

MISSING IMAGE

Material Information

Title:
Hydrogen fluoride method development for the Ogawa Passive Sampling Device
Physical Description:
Book
Creator:
Johansson, Ilsa
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
Phosphate
TEA
Surfer
Aermod
IC
Dissertations, Academic -- Public Health -- Masters -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: This study tested the precision and accuracy of a triethanolamine (TEA) absorbent in the OgawaTM Passive Sampling Device (PSD) for detection of ambient hydrogen fluoride (HF). The project was initiated to develop a method to verify compliance with emissions regulations for Coronet. Field and laboratory trials were conducted. Mixed cellulose ester filters were saturated with 70% TEA and placed in the PSDs. Aermod ISCT3 modeled ambient HF concentrations at Coronet to guide deployment of PSDs at 28 sampling stations, 3 PSDs per station, 500 to 3500 meters from Coronet. After 30 days of sampling, ambient HF concentrations were calculated from ion chromatographic (IC) analysis (NIOSH Method 7906/AS14 column) results to be in the low parts per billion (ppb) range. Concentration increased with proximity to Coronet as predicted by Aermod ISCT3. Average precision for collocated PSDs was less than 5%.Laboratory validation of the method used a HF permeation tube in a Teflon and high density polyethylene (HDPE) sampling train with silica-dried ultra zero air and crushed sodium hydroxide (NaOH) reference samplers. PSD accuracy was a constant 23% average and average precision was 32%, dropping 50% with minor procedural improvements. Validated field results verified compliance with HF emissions regulations for Coronet.
Thesis:
Thesis (M.S.P.H.)--University of South Florida, 2005.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Ilsa Johansson.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 62 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001709524
oclc - 68905221
usfldc doi - E14-SFE0001399
usfldc handle - e14.1399
System ID:
SFS0025719:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001709524
003 fts
005 20060614112158.0
006 m||||e|||d|||||||
007 cr mnu|||uuuu
008 060516s2005 flua sbm s000 0 eng
datafield ind1 8 ind2 024
subfield code a E14-SFE0001399
035
(OCoLC)68905221
SFE0001399
040
FHM
c FHM
049
FHMM
090
RA425 (Online)
1 100
Johansson, Ilsa.
0 245
Hydrogen fluoride method development for the Ogawa Passive Sampling Device
h [electronic resource] /
by Ilsa Johansson.
260
[Tampa, Fla.] :
b University of South Florida,
2005.
502
Thesis (M.S.P.H.)--University of South Florida, 2005.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 62 pages.
520
ABSTRACT: This study tested the precision and accuracy of a triethanolamine (TEA) absorbent in the OgawaTM Passive Sampling Device (PSD) for detection of ambient hydrogen fluoride (HF). The project was initiated to develop a method to verify compliance with emissions regulations for Coronet. Field and laboratory trials were conducted. Mixed cellulose ester filters were saturated with 70% TEA and placed in the PSDs. Aermod ISCT3 modeled ambient HF concentrations at Coronet to guide deployment of PSDs at 28 sampling stations, 3 PSDs per station, 500 to 3500 meters from Coronet. After 30 days of sampling, ambient HF concentrations were calculated from ion chromatographic (IC) analysis (NIOSH Method 7906/AS14 column) results to be in the low parts per billion (ppb) range. Concentration increased with proximity to Coronet as predicted by Aermod ISCT3. Average precision for collocated PSDs was less than 5%.Laboratory validation of the method used a HF permeation tube in a Teflon and high density polyethylene (HDPE) sampling train with silica-dried ultra zero air and crushed sodium hydroxide (NaOH) reference samplers. PSD accuracy was a constant 23% average and average precision was 32%, dropping 50% with minor procedural improvements. Validated field results verified compliance with HF emissions regulations for Coronet.
590
Adviser: Steve Mlynarek.
Co-adviser: Noreen Poor
653
Phosphate.
TEA.
Surfer.
Aermod.
IC.
690
Dissertations, Academic
z USF
x Public Health
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1399



PAGE 1

Hydrogen Fluoride Method Development For The Ogawa TM Passive Sampling Device by Ilsa Johansson A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Public Health College of Public Health University of South Florida Co-Major Professor: Steve Mlynarek, Ph.D. Co-Major Professor: Noreen Poor, Ph.D. Yehia Hammad, Ph.D. Date of Approval: November 18, 2005 Keywords: Phosphate, TEA, Surfer, Aermod, IC Copyright 2005, Ilsa Johansson

PAGE 2

i Table of Contents List of Tables ... iii List of Figures....v Abstractvii Introduction ... 1 Chapter One: Fluoride... 3 Health... 5 Standards.. 7 Chapter Two: Coronet..... 11 Chapter Three: The Sample r....... 13 Analytical Methods.... 15 Chapter Four Hypotheses.... 17 Chapter Five: Methods Sampler Preparation... 18 September 2003 Field Sampling 18 Modeling ........................ 19 March 2004 Field Sampling...21 Laboratory Validation.... 22 Extraction.. 24 Analysis. 24 Chapter Six: Results Field Sampling Results.. 25 Laboratory Validation Results... 27 Chapter Seven: Discussion and Conclusions Discussion.. 29

PAGE 3

ii Conclusions.... 30 References... 32 Appendices Appendix A: Calculations .... 38 Appendix B. Calibration Charts.... 48

PAGE 4

iii List of Tables Table 1. Facilities that produce, process, or use hydrogen fluoride. 4 Table 2. Releases to the environment fro m facilities that produce, process, or use hydrogen fluoride .... 9 Table 3. Common HF analysis methods. 16 Table 4. Laboratory validation tr ial I results...28 Table 5. Laboratory validation tr ial II results. 28 Table A-1. Standard deviation (SD) and limit of detection (LOD) calculations in g/ml.38 Table A-2. September 2003 field sampling trial results analysis Table A-3. March 2004 field sampling trial results analysis.. 40 Table A-4. Hypothetical flow rate and sampling time calculations to obtain detectable HF amount in sampling train.. 44 Table A-5. Laboratory validation trial I results and calculations of fluoride concentration in sampling train 45

PAGE 5

iv Table A-6. Laboratory validation trial II results and calculations of fluoride concentration in sampling train 47 Table A-7 Calculation of UTM coor dinates from lat/long data: March 2004 PSD deployment... 49

PAGE 6

v List of Figures Figure 1. A erial photograph of Corone t and vicinity. 1 Figure 2. PSD components, assembly, a nd prepared sampler clip.. 14 Figure 3. Initial PSD deployment at EPCHC sampling stations 19 Figure 4. Aermod HF concentration ( g/m 3 ) diagram and planned grid PSD deployment................20 Figure 5. PSD deployed in field, March 2004. 21 Figure 6. Flow diagram of laboratory validation sampling train. 23 Figure 7. Results from sampling period (9/19/03 to 10/19/03) 25 Figure 8. Results from sampling period (2/19/04 to 3/21/04).. 26 Figure 9. Surfer contour map of March 2004 sampling results 26 Figure 10. PSD sampling rates compared to HF concentration in sampling train... 27 Figure B-1. September 2003 field sampling da ta analysis calibration chart.... 52

PAGE 7

vi Figure B-2. March 2004 field sampling data analysis calibration chart...52 Figure B-3. Laboratory validation trial I data analysis calibration chart............. 53 Figure B-4. Laboratory validation trial II da ta analysis calibration chart.... 53 Figure B-5. Dual rotameter calib ration chart... 54

PAGE 8

vii Hydrogen Fluoride Method Development for the Ogawa TM Passive Sampling Device Ilsa Johansson ABSTRACT This study tested the precision and accuracy of a triethanolamine (TEA) absorbent in the Ogawa TM Passive Sampling Device (PSD) for detection of ambient hydrogen fluoride (HF). The project was initiated to develop a method to veri fy compliance with emissions regulations for Coronet. Field and laborator y trials were conducted. Mixed cellulose ester filters were saturated w ith 70% TEA and placed in the PSDs. Aermod ISCT3 modeled ambient HF concentrations at Coronet to guide deployment of PSDs at 28 sampling stations, 3 PSDs per station, 500 350 0 meters from Coronet. After 30 days of sampling, ambient HF concentrations were cal culated from ion chromatographic (IC) analysis (NIOSH Method 7906/AS14 column) results to be in the low parts per billion (ppb) range. Concentration increased with proximity to Cor onet as predicted by Aermod ISCT3. Average precision for collocated PSDs was less than 5%. Laboratory validation of the method used a HF permeation tube in a Teflon and high density polyethylene (HDPE) sampling train with sili ca-dried ultra zero air and crushed sodium hydroxide (NaOH) reference samplers. PSD accuracy was a constant 23% average and average precision was 32%, dropping 50% with minor proce dural improvements. Validated field results verified compliance with HF emi ssions regulations for Coronet.

PAGE 9

Introduction The following research was performed to test the accuracy and precision of a triethanolamine (TEA) absorbent in the Ogawa TM Passive Sampling Device (PSD) for detection of ambient concentrations of hydrogen fluoride gas (HF). The project was initiated to develop a method to verify compliance with fluoride emissions regulations for Coronet, a phosphate processing plant in Hillsborough County, Florida (see Figure 1) that recently closed amidst compliance investigations and exhausted mines. 1 Figure 1. Aerial photograph of Coronet and vicinity. (TerraServer.com, February 2004)

PAGE 10

2 Coronet came under investigation for the second tim e in its one hundred years of operation for emissions regulations after an employee reported su spected violations to the county. In 2003 and 2004 the Environmental Protection Commission of Hillsborough County (EPCHC) employed Fourier transform infrared spectroscopy (FTIR) to sample for ambient fluoride concentrations around Coronet to verify compliance with emissions regulations specified in their air permit. The FTIR detection limit for fluoride was 0.5 ppb, howev er, because water and fluoride absorb in the same spectral region, the presence of water can ob scure detection of fluoride at low ppb levels. The EPCHC sought another method to detect and verify the presence of fluorides. The environmental laboratory in the College of Publ ic Health at the University of South Florida housed a set of Ogawa TM passive sampling devices (PSDs) that are commonly used to measure ambient pollutants, though no protocol existed fo r HF. PSD similarity to an existing passive sampler for HF, the Radiello and documented use of the same absorbent in the PSD was enough to initiate development of a HF method for the PSD. When Coronet closed unexpectedly soon after the project was underway, PSDs had to be deployed in the field before the method was verified in the laboratory. One month before operations at Coronet were terminated, PSDs were prepared and deployed around the plant to complete the field sampling segment of the project. Laboratory validation of the method had to follow field sampling.

PAGE 11

3 Chapter One: Fluoride HF consists of a hydrogen ion and fluoride, the ionic form of the element fluorine, the most electronegative element on the periodic chart. Fl uorine was isolated and introduced to the world of chemistry at the turn of the last century. The utility of fluorine compounds developed since then includes the prominent anti-depressant Prozac, the fluorocarbon Freon, the terminal amino acid labeling protein analysis compound 1-fluoro-2,4-dinitrobenzene (FDNB) (Nelson, 2000), the prominent anti-metabolic chemotherapy drug 5Fluorouracil, the hemoglobin substitute FluosolSA, and the popular non-stick coating Teflon (Hill, 1995), also used extensively in research such as that concerned in this report. These exampl es are just the tip of the iceberg for fluorine compounds in our world today. HF is the most common fluorine compound produced and used as a result of its vast industrial applications. Anhydrous HF is used to make most fluorine compounds, which include refrigerants (fluorocarbons), herbicides, pharm aceuticals, fluxing agents, solvents, industrial coatings, electrical components, and a plethora of plastics (fluoropolymers). HF is also used to produce high-octane gasoline, steel, aluminum, etched silica, ceramics, fluorescent lights, electrical components, and nuclear power. HF is emitted as a byproduct from operations that burn or chemically treat minerals or natural fuels. Power utilities produce the largest quantities of HF as a byproduct and aluminum, brick, and phosphate processing plants are next in line (ATSDR, 2003) The prevalence of HF across the United States as it is used in production of non-fluorinated compounds, used to manufacture fluoride compounds, and emitted simply as a wa ste by product is illustrated in Table 1.

PAGE 12

4 Table 1. Facilities that produce, pr ocess, or use hydrogen fluoride. State a Number of Minimum Maximum Activities and Uses c Facilities Pounds Onsite b Pounds Onsite b AK 2 0 99,999 1, 5, 13 AL 28 0 9,999,999 1, 5, 6, 11, 12, 13 AR 13 0 9,999,999 1, 5, 6, 7, 8, 11, 12, 13 AZ 23 0 999,999 1, 4, 5, 6, 7, 9, 10, 11, 12,13 CA 34 0 49,999,999 1, 3, 4, 5, 6, 7, 9, 10, 11,12, 14 CO 16 0 999,999 1, 5, 7, 9, 11, 12 CT 5 100 99,999 1, 5, 10, 11, 12 DE 3 0 999,999 1, 5, 6 FL 24 0 999,999 1, 2, 3, 5, 6, 7, 9, 10, 11, 12, 13, 14 GA 26 0 999,999 1, 5, 7, 11, 12, 13 IA 15 0 99,999 1, 5, 7, 12 ID 5 100 99,999 1, 5, 10, 11, 12, 13, 14 IL 41 0 9,999,999 1, 2, 3, 4, 5, 6, 7, 9, 10, 11, 12, 13 IN 36 0 999,999 1, 5, 7, 10, 11, 12, 13, 14 KS 14 0 9,999,999 1, 2, 3, 5, 6, 7, 8, 9, 10, 13 KY 29 0 9,999,999 1, 2, 3, 5, 6, 9, 10, 11, 12 LA 19 0 9,999,999 1, 3, 4, 5, 6, 10, 12 MA 11 0 99,999 1, 5, 10, 11, 12 MD 12 0 9,999 1, 5, 11, 13 ME 3 0 9,999 1, 11, 13 MI 25 0 99,999 1, 2, 3, 5, 6, 7, 10, 11, 12, 13 MN 14 0 99,999 1, 5, 6, 10, 11, 12, 13 MO 27 0 99,999 1, 5, 7, 11, 12 MS 9 0 999,999 1, 5, 8, 11, 13 MT 6 0 999,999 1, 5, 10 NC 37 0 999,999 1, 5, 11, 12, 13 ND 8 0 999,999 1, 5, 10, 12, 13 NE 8 0 999,999 1, 3, 5, 9, 13 NH 4 0 99,999 1, 5, 12 NJ 15 0 9,999,999 1, 5, 6, 10, 12 NM 7 0 999,999 1, 5, 10, 11, 13 NV 2 0 999,999 1, 5 NY 29 0 99,999 1, 5, 6, 10, 11, 12 OH 62 0 999,999 1, 5, 6, 7, 8, 9, 10, 11, 12, 13 OK 17 0 999,999 1, 2, 3, 5, 6, 7, 8, 9, 10, 11, 12, 13 OR 20 0 999,999 1, 3, 4, 5, 7, 9, 10, 11, 12 PA 67 0 9,999,999 1, 2, 3, 5, 6, 7, 9, 10, 11, 12, 13 PR 1 100,000 999,999 6 RI 3 100 99,999 6, 10, 11 SC 30 0 999,999 1, 3, 5, 6, 10, 11, 12, 14 SD 1 0 99 1, 13 TN 19 0 999,999 1, 2, 5, 6, 10, 11, 13 TX 82 0 49,999,999 1, 2, 3, 4, 5, 6, 7, 9,10,11, 12, 13, 14 UT 14 0 999,999 1, 5, 6, 10, 11, 12, 13 VA 23 0 99,999 1, 5, 10, 11, 12 VT 2 1,000 99,999 11 WA 18 0 999,999 1, 5, 10, 11, 12 WI 24 0 99,999 1, 5, 7, 10, 11, 12, 13 WV 20 0 999,999 1, 5, 10, 11, 12 WY 9 0 99,999 1, 2, 3, 5, 10, 13 (a) Post office state abbreviations used; (b) Amounts on site reported by facilities in each state; (c) Activities/Uses: 1. Produce, 2. Import, 3. Onsite us e/processing, 4. Sale/Distribution, 5. Byproduct, 6. Impurity, 7. Reactant, 8. Formulation Component, 9. Article Component, 10. Repackaging, 11. Chemical Processing Aid, 12. Manufacturing Aid, 13. Ancillary/Other Uses, 14. Process Impurity. Source: 2001 TRI (ATSDR, 2003).

PAGE 13

5 HF is derived from fluorine containing minerals such as fluorspar or fluorite (CaF 2 ), fluorapatite (Ca 5 (PO 4 ) 3 F) and cryolite (Na 3 AlF 6 ), and phosphate deposits in igneous rock around the globe. HF is produced when fluorine containing minerals are heated or dissolved in acid, as is done at the phosphate plant concerned in this report to free phosphates for fertilizer and feed additive production. HF gas readily dissolves in precipitation or moisture in the air and adheres to particulate matter and aerosols where it can be deposited onto vegetati on, soil, and water. Fluoride reacts readily with all compounds other than oxygen and nitroge n. Resultant fluoride compounds are generally stable and continued reactivity depends upon th e nature of the chemical bonds formed and surrounding conditions. Ambient levels of HF depend u pon proximity to natural and industrial sources, climate factors, and topography. Measurements made in the late 1980s and early 1990s found average ambient fluoride concentrations to vary from 2.43 ppb (1.89 g/m 3 ) in urban areas to 0.21 ppb (0.16 g/m 3 ) in non-urban areas, with 75% existing as HF. Ambient HF levels were found to remain fairly constant throughout the year (ATSDR, 2003). Health Much debate has gone into and still goes on concerni ng what levels of HF or fluoride exposure are safe, as it has become such a common feature not only in the manufacturing industry, but in the health industry as well. At low levels th e antibacterial and bone strengthening properties of fluoride have proven to be beneficial to dental health, while at the same time even smaller quantities have been found to destroy important ve getation. Unregulated fluoride emissions from phosphate processing in west central Florida from at least the 1950s and into the 1970s caused extensive damage to the floral, citrus, and cattle industries, as well as negatively affecting human health in the vicinity of phosphate plants (Dewey, 2001). Determination of what low levels are actually safe to human and environmental health is still being investigated. At higher levels the toxicity to human and animal life is more certain. HF and fluoride toxicity to plant life is 1 to 3 orders of magnitude greater than other common

PAGE 14

6 pollutants, including ozone, chloride, chlorine and sulfur dioxide. Sensitive plant species are affected by concentrations less than 1 ppb (Weinstein and Davison, 2003). Human exposure to HF occurs with skin and eye contact, or via inhalation. HF exposure also occurs indirectly by ingestion of fluoridated wate r, or plants and animals that have incorporated fluoride from the air, soil, or water. Certain plants are effective biaccumulators of fluoride. Exposure to HF can cause mild to severe corrosion of body ti ssues. Sufficient quantities of fluoride in the blood can alter systematic calcium biochemistry, affecting the heart and nervous system. Inhalation of high concentrations of HF can cause pulmonary edema or fluid accumulation in the lungs, possibly leading to death. All of these effects may be delayed up to hours or longer after initial contact, especially the effects of fluorosis. Once in the blood, fluoride is either processed by the kidney and then excreted from the body or incorporated into calcareous tissu es where it has a half-life of several years. Storage of fluoride in skeletal tissue can result in skeletal fluorosis, increased bone density, morphometric alterations and exostoses. Very high exposure to HF and fluor ides usually has to occur to result in skeletal fluorosis, but this is not the case with dental fluorosis. Dental fluorosis generally occurs when children usually under the age of 8, especially between 15 and 30 months, are exposed to high enough levels of fluoride (recommended exposure less than 1 mg/day or 0.06 ppm for an average adult over 8 hours) and developing teeth become marked, mottled, or even pitted as a result (Lewis and Milg rom, 2003). A 69 percent occurrence of dental fluorosis was found from water with 0.7 ppm fluoride and a 98 percent occurrence at 2 ppm fluoride in a study conducted among children 11 to 13 years of age in San Luis Potosi, Mexico (Ortiz-Perez, et al., 2003), though exposure to l ess than 2 mg per day (0.12 ppm) for this age group is considered safe Older individuals exposed to less than 2 ppm fluoride in water may experience beneficial dental effects however, as fluoride aids ionic calcium and phosphate incorporation into tooth enamel and is also itself incorporated into tooth enam el when encountered steadily and at lower concentrations. Fluorapatite or enamel containing fluoride is harder and less soluble to acids produced by oral bacteria (mutans streptococci) and also inhibits bacterial acid production. The

PAGE 15

7 beneficial dental affects from continuous e xposure to low levels of fluoride may have consequences however, as investigation continues. Studies in 2003 showed that exposure to fluor ides could potentially cause teratogenic and mutagenic affects (Goh, 2003 and Ortiz-Perez, 2003), and inhibit reproductive systems in men exposed to concentrations of fluoride from 3 to 27 mg per day (Ortiz-Preza, et al., 2003) or 0.18 to 1.65 ppm with a person breathing 20 m 3 /day in an 8 hour day at normal temperature and pressure (ATSDR, 2003). Most exposure to fluoride occurs from water, beve rages, food, and dental products. Exposure to fluorides from dermal contact with soil depends upon the fluoride compounds present and associated fluoride availability (World Health Organization, 1984). So ils generally contain between 200 ppm and 300 ppm fluoride, though higher levels will occur where phosphate fertilizers are applied or in the vicinity of h azardous waste sites and fluoride-releasing industries such as the phosphate plant of concern in this investigation. Inhalation is unlikely to cause significant exposure, as HF concentrations in ambient air are usually very low, except for areas in the vicinity of HF emitting industrial operations. At normal ambient levels away from emitting industry daily ex posure to HF via inhalation is no more than about 0.01 mg/day (0.60 ppb). Within industrial sources of HF, where airborne concentrations are frequently at the exposure limit of 2.5 mg/m 3 (3 ppm) however, HF exposure from inhalation can reach 16.8 mg/day (1.02 ppm). Standards The Occupational Safety and Health Administra tion (OSHA) regulates worker exposure to 3 ppm (2.5 mg/m 3 ) HF. The National Institute of Occupa tional Safety and Health (NIOSH) has a recommended exposure limit (REL) of 3 ppm, a 15 minute ceiling exposure limit at 6 ppm (5 mg/m 3 ), and 30 ppm is considered immediately da ngerous to life and health (World Health Organization, 1984). The 2005 American Conference of Governmental Industrial Hygienists (ACGIH) Threshold Limit Value (TLV) and Biological Exposure Indices (BEI) Booklet recommends a short term exposure

PAGE 16

8 limit (STEL) of 2 ppm HF and an 8-hour time weighted average (TWA) TLV of 0.5 ppm, based on irritation to the skin and eyes, and potentially critical effects to bones and the respiratory tract (ACGIH Worldwide, 2005). The discrepancy between the recommended TLV of 0.5 ppm and the regulated 3 ppm for 8-hour exposure may be a ttributed to the legislative process required to establish regulatory exposure levels. The Safe Drinking Water Act requires that natura lly occurring fluoride levels in community water supplies remain below 4 ppm and consumers are to be notified if levels exceed 2 ppm (ATSDR, 2003). Surface water usually has about 0.2 pp m and well water from 0.02 to 1.5 ppm. Community water is fluoridated to a recommended optimal range of 0.7 ppm to 1.2 ppm if it is not fluoridated naturally (World Health Or ganization, 1984). Dental products have concentrations of fluoride ranging from 230 ppm to 5,000 ppm and are not intended for ingestion. Ambient HF concentrations are controlled by regulating emissions from industrial sources. Industries emitting HF include electric utilities, nuc lear power generators, uranium processing, phosphate processing, ceramic manufacturing, brick manufacturing, chemical manufacturers, refineries, computer manufacturing, pharmaceutical manufacturing, aluminum and steel production, welding operations, and a variety of others. Regulations on HF and water soluble fluoride emissions vary from county to county. The amount of HF released to the environment from industries across the U.S. is illustrated in Table 2, as reported by the 2001 Toxic Release Inventory (TRI).

PAGE 17

9 Table 2. Releases to the environment from facilitie s that produce, process, or use hydrogen fluoride. (Reported amounts released in pounds per year a ) State b Facilities Air c Water Underground Land On-site Offsite AK 2 73,027 No data 0 0 73,027 0 AL 28 3,304,387 0 0 12,000 3,316,387 18,596 AR 13 942,385 0 0 0 942,385 8 AZ 23 694,190 No data 0 3,405 697,595 1,062 CA 35 9,211 5 0 0 9,216 2,886 CO 17 795,831 No data 0 0 795,831 0 CT 5 107,388 0 0 0 107,388 10,732 DE 3 201,815 No data 0 0 201,815 0 FL 24 2,358,937 5 0 7,965 2,366,907 0 GA 28 3,286,300 0 0 0 3,286,300 2,944 IA 16 1,300,253 No data 0 0 1,300,253 0 ID 5 208,065 0 0 0 208,065 255 IL 41 2,255,380 1 0 5 2,255,386 1,510 IN 36 3,888,416 250 0 0 3,888,666 0 KS 16 775,771 0 0 0 775,771 930 KY 29 2,090,877 0 0 0 2,090,877 1,301 LA 19 613,860 250 0 11 614,121 0 MA 11 197,829 No data 0 0 197,829 237 MD 14 1,376,506 No data 0 0 1,376,506 15 ME 3 1,286 0 0 0 1,286 0 MI 27 2,139,731 0 0 0 2,139,731 39,376 MN 15 217,532 0 0 0 217,532 0 MO 29 2,464,416 0 0 158,300 2,622,716 0 MS 9 461,108 197 0 2,287 463,592 0 MT 7 167,298 0 0 0 167,298 0 NC 38 5,160,908 5 0 0 5,160,913 0 ND 8 490,133 0 0 0 490,133 260 NE 8 1,279,219 No data 0 0 1,279,219 0 NH 4 208,550 No data 0 0 208,550 391 NJ 15 249,406 0 0 0 249,406 2 NV 2 439,874 No data 0 0 439,874 0 NY 31 1,137,383 0 0 0 1,137,383 750 OH 64 6,147,565 1,601 4, 400,000 0 10,549,166 34,884 OK 18 892,504 100 0 0 892,604 250 OR 21 67,943 0 0 18,398 86,341 0 PA 70 5,056,848 35 0 5 5,056,888 17,156 PR 1 500 No data 0 0 500 0 RI 3 3,683 No data 0 0 3,683 0 SC 30 2,153,397 0 0 0 2,153,397 0 SD 2 89,000 No data 0 0 89,000 0 TN 19 2,069,004 0 0 0 2,069,004 0 TX 83 3,828,730 10 0 21 3,828,761 1,100 UT 14 467,778 0 0 24,930 492,708 0 VA 23 1,745,413 0 0 0 1,745,413 0 VT 2 4,141 0 0 0 4,141 0 WA 19 323,942 0 0 0 323,942 1,405 WI 24 1,231,556 0 0 0 1,231,556 0 WV 21 3,722,619 19,090 0 0 3,741,709 0 WY 9 337,011 No data 0 52,248 389,259 0 Total 991 67,248,474 21,549 4, 400,000 279,575 71,949,598 136,470 a Data in Toxic Release Inventory (TRI) are maximum amounts released by each facility. b Post office state abbreviations are used. c The sum of fugitive and stack releases are included in releases to air by a given facility. d The sum of all releases of the chemical to ai r, land, water, and underground injection wells. e Total amount of chemical transfe rred off-site, including to publicly owned treatment works (POTW). Source: 2001 TRI (ATSDR, 2003).

PAGE 18

10 The EPA regulates phosphate fertilizers production plants by the type of processes used. The manufacturing of DAP done by the phosphate indus try concerned in this report, Coronet, is regulated to emit no more than 29 grams of total fluorides per ton of phosphate processed (40 CFR 63.623 Appendix BB). Coronet was regulated by the county to emit no more than 23 g per ton. Coronet released from 7,000 to 35,000 pounds of HF to the air per year over the last two decades according to the TRI, accounting for almost all HF released onsite and offsite. The annual emissions reported equate to more than 7,300 to 36,000 mg of HF per produced per minute. According to the Clean Air Act (U.S. Code Title 42, Chapter 85), Coronet is a major source of the hazardous air pollutant (HAP) HF and is not only requ ired to regulate its emissions, but to also to apply maximum achievable current technology (MACT).

PAGE 19

11 Chapter Two: Coronet Located in the Bone Valley region of west centr al Florida, Coronet was in a good location to mine phosphates for fertilizer production. Primarily made up of francolite, (Ca,C) 4 [(P,C)(O,OH,F) 4 ] 3 or [Ca 3 (PO 4 ) 2 ] 3 CaF 2 XCaCO 3 the marine deposits of Bone Valley fueled over a century of phosphate extraction at Coronet since the Lowes method of dissolving ore in sulfuric acid made phosphate available to the worl d from inert rock, rather than the usual sources of fish, bone and guano (Meredith, 1965). P hosphate isolation and fertilizer manufacturing techniques continue to evolve, though the general procedure is about the same. Coronet manufactured the fertilizer diammonium phosphate (DAP), also an animal feed supplement, and the fluoride salt potassium tetraflouroborate (KBF 4 ) that is used to make a common aluminum alloy. HF was initially produced when the phosphate rock mined nearby was dissolved in sulfuric acid to render the phosphate soluble for fertilizer manufacturing. Fluorite for example, as just one of the fluorine containing minerals found in phosphate rock, reacts with sulfuric acid to produce HF according to the following chemical equation: CaF 2 + H 2 SO 4 CaSO 4 + 2HF (1) Within the phosphate plant structure, the primar y sources of HF were from three de-fluorinating kilns and two de-fluorinating fluid bed reactors. The fluid bed reactors shared an exhaust stack with one of the kilns and the other two kilns shar ed a second exhaust stack. Removal of fluoride from kiln emissions was achieved by routing em issions through a dust-settling chamber followed by dual spray towers in series a nd a Tellerett packed scrubber.

PAGE 20

12 Wastewater from the kiln fluoride recovery syst em was re-circulated until the HF concentration reached 5 percent and the water was then used in fluoride salt production. These de-fluorination processes were to maintain fluor ide emission levels to ambient air at or below 5.6 lb/hr (42,335 mg/min), as specified by the EPC Title V air permit issued to Coronet. According to the 2001 TRI, the highest of the annual emissions reports in the last ten years averages to about 4.8 lb/hr (36,345 mg/min). The highest emissions reported were below the regulated limit. The permit required annual m onitoring of HF utilizing EPA Method 13A or 13B, which include determination of stack conditions a nd concentrations of particulate matter. After the complaint, the EPCHC used a different me thod (FTIR) to verify compliance with HF emissions regulations, sampling away form the exhaust stacks in the ambient air of neighborhoods nearby.

PAGE 21

13 Chapter Three: The Passive Sampler Relatively unaffected by temperature and pressure, passive sampling allows measurement of a compound as it collectively diffuses to an absorptive media. Ambient concentrations are calculated from analytical results using the diffusi on constant of the compound and path length of diffusion for the sampler. To measure low ambien t concentrations of gaseous pollutants, passive samplers such as the PSD are advantageous in that many samplers can be deployed simultaneously over a large area. Ogawa TM PSDs are commonly used to measure the distributional occurrence of various pollutant gases such as ozone, ammonia and sulfate. At th e time of this investigation, the only passive sampling method for ambient HF over periods greater than 8 hours was the Radiello Other passive methods exist, but not for measurement of ultra low concentrations that would require sampling over periods lasting up to several weeks. The Radiello sampler is similar in shape and function to the PSD consisting of a small polymer cylinder coated with a triethanolamine (TEA) abso rbent. Dissimilar to the PSD is a gas-selective casing that allows selective sampling HF by excluding particulate matter. The detection range of HF is from 1 1000 mg/m 3 (1.22 1221 ppm) and the sampler can be deployed for up to seven days. Linear sampling is limited to c oncentrations of 10,000 to 50,000,000 mg/m 3 min at a 187 ml/min sampling rate, 25 C, and a relative humidity between 5 and 90 percent. The Radiello claims four months of shelf life before a nd after sampling (Rupprecht & Patashnick, March 2003). The Ogawa TM PSD consists of a 3 cm long, 2 cm diam eter polymer cylinder with depressions in either end that contain a backup pad or disc and o -ring of the same polymer material as the body.

PAGE 22

Figure 2. PSD components, assembly, and prepared sampler in clip ( www.ogawausa.com ). TEA is also used as an absorbent in the Ogawa TM PSD, albeit for ammonia sampling. Available data document TEAs stability applied to mixed ester cellulose filters in the Ogawa TM PSD over periods of ambient sampling up to 30 days and storage periods up to 90 days before and 42 days after sampling (Kirby, 2000). Mixed ester cellulose filters are reported to be resistant to chemical attack as well as relatively hydrophobic (Wight, 1994). Due to the reactivity and solubility of HF, the mixed ester cellulose filter was chosen to be suitable for application in the PSD to collect HF. For the Ogawa TM PSD, twenty-four hours of sampling is suggested to measure nitrogen and sulfur compounds at concentrations as low as 2 ppb (Atkins, February 1986). The Ogawa TM PSD is reported to be capable of sampling for periods of 168 hours or 30 days to measure ultra low concentrations, as is the suspected case for the research reported herein. PSDs have been found to sample independent of temperature and humidity, while sampling rate paralleling wind velocity was mitigated by the rain shelter (Tate, 2002), a half sphere PVC cap that the PSD hangs under. Affects of non-steady state sampling, concentration fluctuation, sorbent saturation, wind velocity, turbulence, temperature, and pressure, among other factors, were found not to produce significant error in well designed passive samplers. Varying 14

PAGE 23

15 concentrations were found not to affect samplers, for instance, if the time of exposure was greater than the time required for a given compound to di ffuse to the sampler. The sampling or uptake rate of the PSD is about 8.3 ml/min, as determined by the following calculation (Tate, 2002). c i = Q/[(DA/L)t] (2) c i = 7.23Q/t c i = ambient concentration of HF being sampled ( g/cm 3 ) Q = mass uptake determined by analysis ( g) D = coefficient of diffusion for HF = *0.2236 cm 2 /s A = cross sectional area of diffusion path = 0.371 cm 2 L = length of diffusion path = 0.6 cm T = time (s) *Grahams Law of Diffusion; rate proportional to 1/( density) Analytical Methods TEA absorbs HF as an acid gas, which can be an alyzed using ion chromatography (IC) or ion selective potentiometry. The nitrogen dioxide protocol for the PSD specifies IC analysis, which NIOSH Method 7906 also specifies for fluoride analysis. Analytical methods for HF include gravimetric, enzyme inhibition, titrimetric, colo rimetric (with coloring agents cerium III alizarin complexone, lanthanum III alizarin complexone, and zirconium eriochrome cyanine), infrared (IR) and ion mobility spectrometry (IMS), ion selective electrode (ISE), and IC analysis. By 1997 IR and IMS analysis were prevalent enough th at a study investigated their agreement with EPA Method 26A and 13B, chemisorption of HF with dilute sulfuric acid and NaOH treated filters, respectively. IR was found to be more sensitive to variation in HF concentration than IMS, though both agreed within 0.5 ppm over time periods exceeding four ho urs. Significant variation existed between all methods below HF concentrations of 3 ppm howev er (Harris and Dunder, 1997). In 1984 NIOSH

PAGE 24

16 published Method 7902 for the analysis of gaseous and particulate fluoride. Ion selective electrolysis (ISE) analysis was specified (NIOSH, 1984). Comparison of IC and ISE analysis methods specified in NIOSH methods 7906 and 7902, respectively, yielded statistically equivalent r esults, though presence of metals did not interfere with IC analysis as much as with ISE analysis (L orberau, 1993). The IC method reports a limit of detection (LOD) of 3 ug per sample in a range of 0.01 0.25 mg fluoride per sample, suitable for application in the method development investigated here. A comparison of the three more common HF analysis techniques is given in Tabl e 2. NIOSH implemented a standard method for fluoride and HF determination employing IC analysis in 1994. Table 3. Common HF analysis methods Method LOD Reagents Ion Chromatography 3.0 ug sodium tetraborate decahydrate Ion Selective Electrode 3.0 ug KCL, sodium acetate, cyclohexylene diamine, tetraacetic acid Alizarin Complexone 0.6 ug alizarin complexone, lanthanic nitrate, sodium acetate trihydrate

PAGE 25

17 Chapter Four: Hypotheses The objective of this research was to test the performance of a TEA absorbent in the PSD to develop a protocol for ambient HF detection us ing the PSD. The three hypotheses were that (1) the proposed method could collect sufficient HF to be measured using IC analysis; (2) that the method could be validated in the laboratory, and (3 ) that levels of HF were in the low ppb range around Coronet so as to be undetectable by the FTIR method used. The first hypothesis will be tested by analyzing PSDs that have been deployed for a sufficient time (detection of ppb nitrogen dioxide requires 30 days) and determining if detectable levels of HF were accumulated in comparison to lab, trip and field blanks. Testing the second hypothesis depends upon successful generation and sampling of HF in the laboratory to obtain the accuracy and precision of th e PSD. The third hypothesis will be tested by applying the accuracy and precision of the method determined to calculate ambient HF from field sampling results to ascertain if HF was at low e nough levels that its de tection was obscured by water peaks in the FTIR method used by the EPCHC.

PAGE 26

18 Chapter Five: Methods Sampler Preparation After the September 2003 field sampling trial, it was realized that stainless steel scrubbed HF at microgram concentrations (Zankel and Miller, 1987) For this reason, the stainless steel screens were not used in the PSDs in the March 2004 field sampling trial and thereafter. For PSD preparation, all components of the PSD were ag itated for 2 minutes in DI, rinsed thoroughly and soaked in DI for 24 hours, rinsed thoroughly in DI again, and then placed in a drier. After removing from the drier, PSD assembly began by placing the backup pads and o-rings into the PSDs with tweezers. Mixed ester cellulose filters of 0.8 um pore size were then prepared for insertion into the PSD. Twenty microliters of a 70 percent solution of TEA in DI was applied the filters using a micropipette. This amount suffici ently saturated the filters. The filters were inserted into the PSDs and an end cap secured on each end. The prepared PSDs were stored in zip-lock bags that were placed in screw cap plas tic containers that were left in the refrigerator until deployment. September 2003 Field Sampling To ascertain that HF was present in measurable amounts, a preliminary sampling event was conducted in coordination with the EPC. In Sept ember 2003, PSDs were prepared and placed at three of the four EPC sampling stations (Figure 3). Two of the EPC stations were on rooftops (Baseball Field and Church) and one on the ground (Fire Station), each greater than 500 meters from Coronet. A fourth site was utilized that di d not coincide with an EPC sampling station, but was less than 500 meters from Coronet and on the ground.

PAGE 27

Figure 3. Initial PSD deployment at EPCHC sampling stations: baseball field, fire station, Cason Road, church ( http://www.terraserver.com 10/28/05). Two samplers were placed at the church and baseball field and three at the fire station and Cason Road. Sampling was done over a 30-day period from September 19 to October 21 in 2003. IC analysis showed presence of HF above blank values at all sites. Modeling Before laboratory validation trials could begin, the phosphate plant ceased operations unexpectedly and a full-scale sampling event was conducted in the month prior to shutdown. An Aermod ISCT3 modeling program published by Lakes Environmental was used to ascertain the general area where HF might be expected to landfall and at what concentrations according to production data for the plant (obtained from the Title V Air Permit for Coronet, issued by EPC) and formatted weather data from 1987 to 1991. Weather data was obtained from the 19

PAGE 28

Environmental Protection Agency (EPA, February 2004) and included wind direction, wind velocity, height, and mixing data as required by the modeling program. As for production data for the plant, the modeling program processed data input on stack locations, inner diameter, height, emission rate, emission temperature, and roughness factors for the surrounding landscape. A rural roughness factor was chosen, as pastureland and single level residential neighborhoods surround Coronet. To model dispersion after accidental releases of large quantities of HF, a study published by Shell Incorporated in 1990 observed that high concentration plumes reacted exothermically with greater than 70% moisture in the air and rose, but with less than 50% moisture in the air the HF polymerized and the plume sank. Aermod ISCT3 predicted lower concentration HF emissions to touch the ground in general agreement with the observations for the more concentrated plumes. Higher HF concentrations were predicted to be roughly 500 to 2000 meters form Coronet for all years modeled. Surfer was used to create a contour map for the 1991 data and to overlay the planned grid deployment pattern for the PSDs (Figure 4). -0 .01 0 0. 0 1 0. 0 2 0. 0 3 0. 0 4 0. 0 5 0. 0 6 0. 0 7 0. 0 8 0. 0 9 0. 1 0. 1 1 0. 12 0. 1 3 0. 1 4 0. 1 5 0m 1000m 2000m 3000m 4000m Figure 4. Aermod HF concentration (g/m 3 ) diagram and planned grid PSD deployment. 20

PAGE 29

March 2004 Field Sampling According to the Aermod ISCT3 model, PSDs were placed from 500 to 3500 meters around Coronet (Figure 4). Ten percent of the samplers deployed were prepared as lab, trip and field blanks. A total of 84 available samplers were prepared and deployed at 28 stations, three samplers per station. PSDs were transported to the field in the plastic storage containers, less one container of three PSDs that was left in the refrigerator to serve as lab blanks. At each sampling station, a PVC dome was hung with an insulated wire about 2 meters from the ground (Figure 5). The height enabling sampling at a general level where exposure may occur to people and that was also accessible (Tate, 2001). The PSDs were removed from their storage bag and placed into plastic clips that were fastened to a long metal screw protruding from the inside of the top of the PVC dome. Three PSDs were placed in each dome. Figure 5. PSD deployed in field, March 2004. A GPS reading was taken at each sampling location. Three PSDs taken into the field served as trip blanks and three PSDs were exposed briefly in the field to serve as field blanks. PSDs were left for 30 days of sampling, February 19 to March 21 of 2004. The samplers were then collected, and stored as before for extraction and analysis back in the lab. 21

PAGE 30

Laboratory Validation To design a sampling train to validate the HF method tested in the field, past experimentation was referenced. A comparative study in 1979 employed a HF permeation tube, water bath, drying components and metered air to determine that relative humidity between 25 and 90 percent did not affect collection efficiency of sodium formate treated mixed cellulose ester filters in a cassette sampler for HF. This study found that a single sodium formate treated mixed ester cellulose filter collected total fluoride equivalent to two 25 ml impingers in series filled with 10 ml each of 0.1N NaOH (Einfeld, 1979). Flow rates for the impingers were not specified. In the mid-1990s temperature-dependent permeation tubes were utilized again in the lab to validate HF samplers and analytical methods, as well as study kinetics of HF absorption by various reagents. The method generated HF concentrations from 0.05 mg/m 3 to 10 mg/m 3 with a relative error of only 8 percent (Ennan et al, 1995). Key components of the gas generating apparatus were carrier gas purification and drying units, and airflow check valves automated in conjunction with pressure regulators and rotameters. Referred to as the -GPSG-HF-001 gas generator, it was approved as a second echelon of calibration reference measuring device for complete verification of HF analyzers (Ennan et al, 1995). These two experiments guided the construction of a high-density polyethylene (HDPE) and Teflon sampling train containing a certified HF permeation tube to validate the proposed HF method for the Ogawa TM PSD. The reference for the HF concentration in the sampling train was initially a set of 0.1N NaOH impingers, which were later replaced with polyethylene tubes containing 0.8 grams of crushed NaOH pellets. Two NaOH tubes in series were placed in parallel to a PSD sampling chamber containing two to three prepared PSDs. Ultra-zero air entered the sampling train through a rotameter, went through a 10 ml drying chamber filled with silica, and then passed over a HF permeation tube in a 50 ml chamber. From the HF permeation tube chamber, the air stream split at a t-valve, one side going to the PSD sampling chamber, the other side going to NaOH reference media, as illustrated in Figure 6. 22

PAGE 31

H H 23 B A D D E G F C Figure 6. Dia g ram of Sampling Train: A. inlet rotameter; B. dry i n g cham b e r; C. perm eatio n tube cham ber; D. NaOH tubes; E. PSD cham ber; F. check valve; G. dual rotameter; H. exhaust ports. The PSD sampling cham ber was initially a volum e of 1 liter (Trial I), but was changed to a 0.25 liter volum e ( T rial II). The per m eation tube w as acqu i red from KI N-TEK and was certified to em it HF at a rate of 674 n g / m in at 25 C and for 7 m onths of contin u ous em ission. The followi ng equation calculates the generation rate of HF at a given air flow rate: (K o 6 74 (n g / m i n)) / diluti on flow rate (ml/m i n), K o = 1.120 (3) A flow rate of 1.5 m l /m in would pro duc e HF at a rate of 50 3 ng /m in or 0. 503 g / m i n. So a sam p ling run of just one m i nute would t h eoretically pr oduce sufficient HF to be detected by I C analy s is. Sam p ling run s consisted of a blank run where zero air flowed thro ugh t h e sy ste m without the HF per m eation tube in place, and then a fluoride run, in which the HF per m e a tion tube was in plac e. Flow rates w e re recorded at the beginning and end of each sa m p lin g run. Average flow rates were used to calculate the volum e of air that flow ed t h roug h the s y stem over the sam p ling period. A Digital Therm o Hy grom eter (Hart Scientif ic 5610 EG&G Model 911S 1, Seri al No. C0804 58, calibrated 7/27/98) was used to m e a s ure the te m p er atu r e and relative hum idity of the air strea m at the beginning and end of each run at the PSD cha m be r and absorption m e dia outlets to m onitor conditions within the s y ste m Air flowi ng from the PSD cham ber and reference media remai n ed

PAGE 32

24 at 22.9C to 23.1C and 2% relative humidity, though the distributor of the zero air cylinder claimed up to 3% humidity was possible. Extraction Extraction of PSD filters was done by placing the filters, together with the backup pads and orings of each PSD into a 50 ml centrifuge tube and adding 3 ml of DI. The contents were then sonicated for 25 minutes, poured into a 10 ml syringe with a 0.2 m 13mm syringe filter, and filtered into a 10 ml Dionex vial that was capped and refrigerated until analysis. Extraction of NaOH reference tubes was done in the same way, except that 45 ml of DI was added to the centrifuge tube containing a NaOH in its entirety prior to sonication. The amount of DI added to the NaOH tubes achieved roughly the same 0.1N NaOH concentration of the impingers initially used, for which IC analysis had proven feasible when testing their reference performance. Analysis Samples were analyzed with a Dionex 600 Ion Chromatograph according to NIOSH Method 7906. The AS4A column of the method was subs tituted with an AS14 column after the AS4A column failed. A 32 mM sodium tetraborate decahydrate eluent was ramped from 3 to 100 percent over the first 12 minutes of the sampling run to achieve resolved fluoride peaks. Five standards and three check standards were prepared (0.05 g/mL to 5 g/mL) from a fluoride and an anion standard. The lowest standard w as analyzed ten times to determine a theoretical 0.16 g/ml limit of detection (LOD) and a middle st andard run ten times to determine analytical precision of the IC at a theoretical 0.0095 g/m 3 Duplicate analyses were performed for each quarter of samples analyzed for precision checks and DI run a few times throughout the analysis to check for incomplete sample elution. Lab, tr ip and field blanks were analyzed, averaged and subtracted from analysis results.

PAGE 33

Chapter Six: Results Field Sampling Results The preliminary sampling event in September of 2003 showed higher fluoride concentrations at sampling stations that were closer to the phosphate plant, Cason Road being closest and the baseball field being the furthest from Coronet (Figures 7 and 9). September 2003 Field Sampling Results0.0000.0200.0400.0600.0800.1000.120Cason RdFire Station ChurchBaseball FieldHF (ppb) Figure 7. Results for September 2003 sampling period (9/19/03 to 10/19/03). The full-scale sampling event that took place in March of 2004 also showed higher concentrations of fluoride at sampling stations closer to the plant, in vicinities predicted by the Aermod ISCT3 modeling program (Figures 8 and 9), though concentrations were generally higher. Relative standard deviation (RSD) among collocated samples was less than 6 for all but four of the stations, for which RSD was greater than 100. 25

PAGE 34

March 2004 Field Sampling Results0.0000.1000.2000.3000.4000.5000.60012345678910111213141516171819202122232425262728HF (ppb) Figure 8. Results for March 2004 sampling period (2/19/04 to 3/21/03). Figure 9. Surfer contour map of March 2004 sampling results. 26

PAGE 35

Laboratory Validation Results The validity of these findings depended upon laboratory validation of the method, carried out after field sampling took place. The sampling rates in the lab averaged 7.9 ml/min, close to the 8.3 ml/min calculated using the HF diffusion coefficient derived from Grahams law and sampler dimensions from literature. Sampling rate may increase systematically as concentration decreases, but the highest fit of a calibration line to the data is not high enough to extrapolate into much lower concentration regions to apply a new sampling rate than that originally calculated to field sampling data. Sampler accuracy for amounts calculated to diffuse from experimental concentrations were high, indicating that the theoretical sampling rate of 8.3 ml/min agrees with PSD performance. Laboratory Trial I and II: PSD Sampling Rates y = 15.965x-0.6705R2 = 0.6727051015202502468HF (ppm)Sampling Rate (ml/ 10 m Figure 10. PSD sampling rates compared to HF concentration in sampling train. When calculating the HF that would be expected to diffuse to the PSD, average accuracy was over 100% with an average precision of 56% as seen in Figures 4 and 5. Between laboratory validation trials I and II the variability in sampling times and flow rates were minimized and the PSD chamber volume was decreased by 75%. These changes resulted in average precision 27

PAGE 36

28 increase between collocated samplers from 24.6% to 13.8%, though average accuracy was less for measurement of HF calculated to diffuse at th e theoretical rate of 8.3 ml/min. Improvements were also observed in average precision in measure ment of total HF by about 50%. The average accuracy for total fluoride measured varied little between trials, remaining at about 23% to validate steady sampling rates for the PSD method. Table 4. Laboratory Validation Trial I Results. Sample Reference PSD1 Measured PSD1 PSD2 RPD Calculated to Measured (ng/ml) (ng/ml) (%) (ug) (ug) (%) Diffuse (ug) (%) 1.0 3.4 0.5 14.7 18.5 17.5 5.6 19.6 94.4 2.0 0.7 0.3 35.7 22.1 15.3 36.4 7.3 302.0 3.0 2.8 0.3 8.9 13.3 9.4 34.7 15.6 85.2 4.0 4.9 0.8 14.3 19.5 14.2 31.5 31.3 62.4 5.0 7.1 0.7 8.5 18.1 13.6 28.7 40.1 45.2 6.0 1.7 1.0 55.9 16.3 14.6 10.5 7.8 208.5 Average 3.4 0.6 23.0 18.0 14.1 24.6 20.3 132.9 SD 2.3 0.3 19.0 3.0 2.7 13.2 13.1 100.7 CV 67.0 47.1 82.4 16.6 19.0 53.6 64.6 75.8 RPD = relative percent deviation, SD = standard deviation, CV = coefficient of variation Table 5. Laboratory Validation Trial II Results. Sample Reference PSD1 Measured PSD1 PSD2 RPD Calculated to Measured (ng/ml) (ng/ml) (%) (ug) (ug) (%) Diffuse (ug) (%) 1.0 3.4 0.9 25.4 6.6 4.7 33.6 4.7 140.4 2.0 4.3 0.8 18.0 6.9 5.5 21.3 13.1 52.5 3.0 5.2 1.1 20.4 20.3 18.0 11.9 18.5 109.6 4.0 4.1 0.9 19.1 16.5 13.6 19.1 25.4 64.9 5.0 3.1 1.0 31.4 21.1 20.6 2.5 31.0 68.2 6.0 2.1 0.4 15.5 16.8 16.1 4.0 26.2 64.1 7.0 4.7 1.0 35.9 10.2 9.7 4.4 13.8 73.6 Average 3.8 0.9 23.7 14.0 12.6 13.8 19.0 81.9 SD 1.0 0.2 7.6 6.1 6.2 11.5 9.2 31.4 CV 27.1 27.9 32.0 43.6 48.8 83.1 48.3 38.4 RPD = relative percent deviation, SD = standard deviation, CV = coefficient of variation Laboratory trials validated that the low ppb HF concentrations measured in the field were reasonably accurate and in the range of tho se expected for outdoor air (Israel, 1973).

PAGE 37

29 Chapter Seven: Discussion and Conclusions Discussion Due to the ability of gases to r eadily diffuse and reach equilibrium in the atmosphere, they present the least difficulty among compounds for sam pling. HF presented some complications however, in that it readily reacts with water, particulates, and most compounds, complicating ambient sampling as well as laboratory validation of the proposed method. The 23% average accuracy obtained from laboratory results for total fluoride measured could be attributed to HF reacting with water molecules in the carrier gas. Sampling periods had to be extended sixteen times the theoretical period to act ually achieve detectable levels of fluoride. Moisture in the air carrying the HF would adsorb to the walls of the PSD chamber and to the PSD itself, whereas the NaOH reference media had no chamber or large sampler body for this to occur to such an extent. It should be noted however, that accuracy did not improve when the sampling chamber was reduced in size by 75 percent. Another explanation for the low accuracy of total fluoride measured could be the polymer material that the PSD is made of, not named in the literature. HF reacts with low density polyethylene (LDPE) or plastics other than HDPE and Teflon. The degree to which the sampler body is absorbing JF could be tested by placing the filter alone or affixed to a Teflon surface in the sampling chamber for comparison. All in all, the accuracy remained constant ove r both sampling trials a nd can be relied upon to estimate ambient HF concentrations in light of the relative agreement between the sampling rates observed in the lab with that calculated theore tically. Precision was observed to improve with minor reduction in procedural variables, indicating further improvements to be feasible

PAGE 38

30 As improvements in the laboratory method are made, a remaining problem common to passive samplers and not touched upon in this study should also be a ddressed. The ability of the PSD to retain fluoride captured over time is unknown. This is a question that would need to be answered to ensure the reliability of the method. Conclusions NIOSH recommends 25 percent variability among 95 pe rcent of the field samplers in a range of half to twice the environmental standard to a ttain acceptability, as well as 10 percent precision in laboratory results. Standard deviation among field samplers was no more than 6 percent for all but 14% of the samplers, which had over 100 times greater values. For laboratory results, precision averaged 8 percent for total HF measure d, but was 32% average at best for measured HF of that calculated to be diffused from the am bient concentration. By NIOSH standards, the HF method proposed in this research may bor der on acceptability, pending improvement in precision. Sampling was conducted around recommended exposure levels of 17 ppb for ambient HF and fluorides (OEHHA, June 20, 2003). By laboratory validation standards, the method appears to measure ambient HF at relativel y predictable levels, pending precision as well. The three hypotheses for this rese arch were that (1) the proposed method could collect sufficient HF to be measured using IC analysis, (2) that the method could be validated in the laboratory, and (3) that levels of HF were in the low ppb range around Coronet so as to be undetectable by the FTIR method used. All three hypotheses were accepted based on the data obtained. For the first hypothesis, the method was able to measure levels of HF even as low as parts per trillion in the field. The second hypothesis was not rejected as the method consistently collected HF in the laboratory with an average accuracy of 23%, confirming the sampling rate at the theoretical value and allowing acceptance of the thir d hypothesis. Levels of HF were determined to be in the low ppb range in consideration of the accuracy and precision determined in laboratory trials. Considering that retention of the HF absorbed by the PSD over time is unknown however, levels may have been higher. The PSD method developed in this study may be of potential use in measurement of ppb and even higher levels of HF and water soluble fluorides to protect public health, as toxicity can occur

PAGE 39

31 indirectly through ingestion of water or f ood that has accumulated fluorides over time (Groth, 1975) and toxicity to potentially valuable flora can occur below ppb levels. The inability of the method to assure determination of HF exclusive of other fluorides and that from particulate and aerosols is not of great concern, as the EPA uses HF as a surrogate for fluoride measurements in development of emissions regulations for phosphate fertilizer plants (40 CFR Parts 9 and 63). Interference from particulate matter that reacted w ith HF and collected on the filter is possible, though past experimentation has found that HF scrubbed by dust accumulated on sampling cartridge entrances was statistically insignificant with a 90 percent confidence interval (Zankel and Miller, 1987). Had fluoride concentrations de tected been large enough for concern, further characterization would have ensued to veri fy the composition and sources of the fluorides sampled and possible violation of emissions regulations. In conclusion, there is potential for application of the HF method for the PSD developed in this study, as deemed from field and laboratory trials, to measure ppb levels of HF with an average accuracy of 100% percent and average precision of 56%. Future work may address possibly changing the polymer of the sampler body to HDPE or Teflon, and testing the HF retention of the method. A study addressing the amount of HF that is scrubbed from the air by moisture and particulates, and what portion of that remains available for exposure would also be interesting. The public health implications for the HF sampled in the field are good in consideration of the accuracy and precision ranges obtained, as HF levels measured around the Coronet plant were determined to be below 1 ppb and the recomme nded exposure level of about 17 ppb for ambient HF in other states.

PAGE 40

32 References ACGIH Worldwide. (2005) 2005 TLVs and BEIs Based on the Documentation of the Threshold Limit Values for Chemical Subs tances and Physical Agents & Biological Exposure Indices. ACGIH, Cincinnati, OH. Atkins, C.H.F.; Sandalls, J.; Law, D.V.; Hough, A.M.; Stevenson, K. (February 1986) The Measurement of Nitrogen Dioxide in the Outdoor Environment Using Passive Diffusion Tube Samplers. Environmental and Medical Sciences Division, Harwell Laboratory. ATSDR. (2003) Public Health Statement for Fluorides, Hydrogen Fluoride, and Fluorine: September 23, 2003. Available from http:// www.atsdr.cdc.gov/toxprofiles/phs11.html accessed May 2005. Bai, Hsunling; Lu, Chungsying; Chang, Kuan-F oo; Fang, Guor-Cheng. (2003) Sources of Sampling Error for Field Measurement of Nitric Acid Gas by a Denuder System. Atmospheric Environment 37:941-947. Czarnowski, W.; Wielgomas, B.; Krechniak, J. (2002). A Passive Dosimeter for Evaluating Exposure to Hydrogen Fluoride. Fluoride; Research Report. 35:22-27. Douglas, Bodie; McDaniel, Darl H.; Alexander, John J. (1988) Concepts and Models of Inorganic Chemistry, 2 nd Edition. John Wiley & Sons, Inc., New York. Dzerzhko, E.K.; Arabadzhi, V.N.; Ennan, A.A. (May 12, 1988) Miniature Integrated Hydrogen Fluoride Gas Analyzer. Industrial Laboratory. 56(1) pp.22-25.

PAGE 41

33 Einfeld, W. (1977) Investigation of a Dual Filter Sampling Method for the Separate Estimation of Gaseous and Particulate Fluorides in Air. MSPH Thesis, University of Washington. Einfeld, W. and Hoestmann, S.W. (1979) Investigation of a Dual Filter Sampling Method for the Gaseous and Particulate Fluorides. Am. Ind. Hyg. Assoc. J. 40(7) 626. Elrashidi, M.A. and Lindsay, W.L. Effect of fluoride on pH, organic matter and solubility of elements in soils. Environmental Pollution. Volume 47, Issue 2 1987, Pages 123-133. Endemic Fluorosis in San-Luis-Potosi, Mexico: Identification of Risk-Factors Associated with Human Exposure to Fluoride. Environmental Research. Volume 68, Issue 1 January 1995, Pages 25-30. Ennan, A. A.; Sakharov, A.V.; Dzerzhko, E.K.; Leivikova, A.A.; Kostyukova, I.S. (November 9, 1993) Generator for Hydrogen Fluoride Gas Verification Mixtures. Industrial Laboratory. 61(2) pp. 10-12. EPA. (2004) Environmental Protection Agency : Technology Transfer Network Support Center for Regulatory Atmospheric Modeling. Available at http://www.epa.gov/scram001/ accessed in February of 2004. Goh, EH, Neff, EW. Effects of fluoride on Xe nopus embryo development. Food Chemical Toxicology. 2003 Nov;41(11):1501-8. Grandell, J. (1985) Stochastic Models of Air Pollutant Concentration. Springer-Verlag. Groth, Edward III. Fluoride Pollution. Environment, April/May 1975. Harris, Daniel C. (1999) Quantitative Chemical Analysis, 5 th Edition. W.H. Freeman and Company, New York.

PAGE 42

34 Hill, John W.; Kolb, Doris K. (1995) Chemistry For Changing Times, 7 th Edition. Prentice Hall, NJ. Israil Gerhard W. (1974) Evaluation and Comp arison of Three Atmospheric Fluoride Monitors Under Field Conditions. Atmospheric Envi ronment. 8:159-166, Pergamon Press. Kennedy, Eugene R.; Ph. D., Fischbach, Thomas J.; Song, Ruiguang, Ph.D.; Eller, Peter M., Ph. D.; Skulman, Stanley A., Ph.D. A NIOSH Technical Report: Guideline for Air Sampling and Analytical Method Development a nd Evaluation. US Dept. of Health and Human Services May 1995. Kirby, C.; Fox, M.; Waterhouse, J. Reliability of NO 2 Passive Diffusion Tubes for Ambient Measurement: in situ properties of the TEA ab sorbent. J. Environ. Monit. 2000, 2, 307312. Nelson, David L.; Cox, Michael M. (2000) Lehninger Principles of Biochemistry, 3 rd Edition. Worth Publishers, New York, NY. Lewis, C.W.; Milgrom, P. Fluoride. Pediatric Review. 2003 Oct; 24(10):327-36. Lorberau, Charles. (September 1993) Determina tion of Gaseous and Particulate Fluorides by Ion Chromatographic Analysis. Appl. Occup. Environ. Hyg. 8(9) 775-784. McFarlane, K.; Prothero, A.; Puttock, J.S.; Roberts, P.T.; Witlox, H.W.M. (November 1990) Development and Validation of Atmospheri c Dispersion Models for Ideal Gases and Hydrogen Fluoride Part I: Technical Reference Manual. The Industry Cooperative HF Mitigation /Assessment Program Ambient Impact Assessment Subcommittee. Shell Research Limited, Thornton Research Centre, England. Minnesota Pollution Control Agency. (August 2002) Citizens Guide to Air Dispersion Modeling. Air Quality # 1.06. Available at http:// www.pca.state.mn.us/air/modeling.html

PAGE 43

35 NIOSH Manual of Analytical Methods, 4th Edition. Method 7906, Issue 1, August 15, 1994: Fluorides aerosol and gas, by IC. NIOSH Manual of Analytical Methods. Method 7902, Issued 2/15/84, Revision 8/15/90: Fluorides aerosol and gas, ion-specific electrode technique. Ortiz-Perez, D.; Rodriguez-Martinez, M.; Martinez, F.; Borja-Aburto, V.H.; Castelo, J.; Gri maldo, J.I.; De La Cruz, E.; Carrizales, L.; Diaz-Barriga, F. Environmental Research. Volume 93, Issue 1, September 2003, Pages 20-30. OEHHA. (June 20, 2003) California Office of Environmental Heal th Hazard Assessment: Air. Chronic REL Review Scientific Review Panel. Available at http://www.oehha.org/air/chr onicrels/Hy FluoCREL.html accessed November 2005. Pavia, Donald L.; Lampman, Gary M.; Kriz, George S. (1988) Introduction to Organic Laboratory Techniques, 3rd Edition. Saunders College Publishing, Fort Worth, TX. Petrakovskaya, E. A.; Kukhlevskii, O.P.; Pavlov, V. F .; Zeer, E.P. (April 6, 2000) Absorption of Hydrogen Fluoride by Ash Foam Glass. Glass Physics and Chemistry. 27(3) pp. 274278. Poor, N.D. (July 5, 2001) Wind Rose Analysis, Tampa International Airport 1986-1989. USF College of Pu blic Health, Department of Environmental and Occupational Health. Rupprecht & Patashnick. (March 2003) Model 33 10, Radiello Passive Sam pling System, HF. February 2002. Available at http://www.rpco.com/assets/lit/lit 04/amb3310004353310rpfactvoc.pdf accessed in September of 2003. Tate, Paul. (2002) Ammonian Sampling Using OgawaTM Passive Samplers. Thesis, University of South Florida, College of Public Health.

PAGE 44

36 TerraServer.com. (February 2004) The Leader In Online Imagery: Image Search. Available at http://www.terraserver.com accessed in February of 2004. U.S. Department of Health and Human Services. (May 1995) Guidelines for Air Sampling and Analytical Method Development and Evaluation. Weinstein, Leonard H.; and Davison, Alan W. Fluoride-induced disruption of reproductive hormones in men. Environmental Pollution. Volume 125, Issue 1, September 2003, Pages 3-11. Wight, Gregory D. (1994) Fundamentals of Air Sampling. Lewis Publishers. World Health Organization. (1984) International Programme on Chemical Safety, Environmental Health Criteria 36, Fluorine and Fluorides. Zankel, K.L.; McGirr, R.M.; Romm, S.; Campbell, A.; Miller, R. (October 1987) Measurement of Ambient Ground-Level Concentrations of Hydrogen Fluoride. JAPCA. 7(10) 11911196.

PAGE 45

37 Appendices

PAGE 46

Appendix A: Calculations 38 Table A-1. Standard deviation (SD) and limit of detection (LOD) calculations in g/ml. Standard ICvalue Standard IC value Standard IC value 0.05 0.0499 0.5 0.4743 3 2.9174 0.05 0.0516 0.5 0.4804 3 2.9734 0.05 0.0432 0.5 0.4767 3 2.9644 0.05 0.0411 0.5 0.4645 3 2.9648 0.05 0.0459 0.5 0.4674 3 2.9202 0.05 0.0431 0.5 0.4827 3 2.9363 0.05 0.0457 0.5 0.4905 3 2.9437 0.05 0.0436 0.5 0.4865 3 2.8757 0.05 0.0402 0.5 0.4616 3 2.9543 0.5 0.4809 x = 0.04492 x = 0.4768 x = 2.93891 SD = 0.00381 SD = 0.00956 SD = 0.03079 LOD = 0.01142 LOD = 0.02868 LOD = 0.09237

PAGE 47

Appendix A (Continued) Table A-2. September 2003 field sampling trial results analysis. Sample IC value Less blank average Fluoride Ambient HF* Ambient HF** RSD*** ug/ml F ug/ml F ug mg/m 3 ppb Blank average 0.18 0 Fire Station 0.57 0.39 1.17 1.088E-04 0.133 1 Fire Station 0.52 0.34 1.02 9.487E-05 0.116 Fire Station 0.16 -0.02 -0.06 -5.581E-06 -0.007 Church 0.23 0.05 0.15 1.395E-05 0.017 6 Church 0.15 -0.03 -0.09 -8.371E-06 -0.010 Baseball Field 0.08 0 0 0.000E+00 0.000 2 Baseball Field 0.05 0 0 0.000E+00 0.000 Cason Road 0.61 0.43 1.29 1.200E-04 0.147 1 Cason Road 0.31 0.13 0.39 3.628E-05 0.044 Cason Road 0.48 0.3 0.9 8.371E-05 0.102 *c = 1000*(Q / (DA/L)t) **ug/cm 3 = 1000 mg/m 3 c = ambient HF (ug/cm 3 ) ((mg/m 3 )*24.45)/20.01 = ppm Q = mass uptake (IC value, ug) ppm 1000 = ppb D = coefficient of diffusion for fluoride (0.2236 cm 2 /s); Tate (2002) A = cross sectional area of diffusion path (3.71 cm 2 ); Tate (2002) ***RSD = SD/average L = length of diffusion path (0.6 cm); Tate (2002) t = sampling time (2592000 s) 39

PAGE 48

Appendix A (Continued) Table A-3. March 2004 field sampling trial results analysis. PSD IC Amount. Less blank average Fluoride Ambient HF* Ambient HF** Average RSD*** ug/mL Fug/mL Fug ug/cm 3 ppb ppb blank average 0.008 0.000 1 0.059 0.051 0.154 4.286E-08 0.052 0.054 6.61 1 0.063 0.055 0.164 4.579E-08 0.056 2 0.048 0.040 0.121 3.365E-08 0.041 0.066 0.13 2 0.058 0.050 0.149 4.144E-08 0.051 2 0.048 0.040 0.119 3.307E-08 0.040 3 0.107 0.099 0.296 8.246E-08 0.101 0.071 1.01 3 0.018 0.010 0.029 8.120E-09 0.010 3 0.038 0.030 0.091 2.545E-08 0.031 4 0.083 0.075 0.224 6.237E-08 0.076 0.138 0.69 4 0.045 0.037 0.112 3.122E-08 0.038 4 0.166 0.158 0.473 1.320E-07 0.161 5 0.199 0.191 0.572 1.596E-07 0.195 0.138 1.04 5 0.015 0.007 0.021 5.860E-09 0.007 5 0.079 0.071 0.214 5.969E-08 0.073 6 0.187 0.179 0.538 1.502E-07 0.184 0.479 0.64 6 0.221 0.213 0.639 1.784E-07 0.218 6 0.551 0.543 1.629 4.546E-07 0.556 7 0.178 0.170 0.509 1.420E-07 0.174 0.109 118.35 7 0.052 0.044 0.131 3.642E-08 0.045 8 0.033 0.025 0.074 2.051E-08 0.025 0.117 0.72 8 0.079 0.071 0.214 5.960E-08 0.073 8 0.141 0.133 0.400 1.116E-07 0.136 40

PAGE 49

Appendix A (Continued) Table A-3 (Continued) PSD IC Amount Less blank average Fluoride Ambient HF* Ambient HF** Average RSD*** ug/mL F ug/mL F ug ug/cm 3 ppb ppb 9 0.022 0.014 0.041 1.138E-08 0.014 0.053 0.61 9 0.042 0.034 0.103 2.863E-08 0.035 9 0.064 0.056 0.167 4.646E-08 0.057 10 0.0147 0.0067 0.0201 5.6088E-09 0.007 0.074 1.04 10 0.042 0.034 0.102 2.846E-08 0.035 10 0.112 0.104 0.312 8.715E-08 0.107 11 0.018 0.010 0.030 8.455E-09 0.010 0.042 150.37 11 0.079 0.071 0.214 5.969E-08 0.073 12 0.027 0.019 0.057 1.599E-08 0.020 0.056 0.80 12 0.079 0.071 0.212 5.902E-08 0.072 12 0.028 0.020 0.060 1.683E-08 0.021 13 0.087 0.079 0.238 6.647E-08 0.081 0.073 0.62 13 0.029 0.021 0.063 1.758E-08 0.021 13 0.051 0.043 0.130 3.616E-08 0.044 14 0.037 0.029 0.087 2.428E-08 0.030 0.049 78.15 14 0.074 0.066 0.199 5.542E-08 0.068 15 0.024 0.016 0.047 1.314E-08 0.016 0.028 15 0.031 0.023 0.070 1.959E-08 0.024 0.25 15 0.024 0.016 0.047 1.306E-08 0.016 16 0.133 0.125 0.375 1.046E-07 0.128 0.133 0.76 16 0.019 0.011 0.033 9.292E-09 0.011 16 0.132 0.124 0.371 1.036E-07 0.127 17 0.150 0.142 0.427 1.192E-07 0.146 0.165 0.70 41

PAGE 50

Appendix A (Continued) Table A-3 (Continued) PSD IC Amount Less blank average Fluoride Ambient HF* Ambient HF** Average RSD*** ug/mL F ug/mL F ug ug/cm 3 ppb ppb 17 0.167 0.159 0.476 1.329E-07 0.162 17 0.029 0.021 0.063 1.750E-08 0.021 18 0.065 0.057 0.172 4.805E-08 0.059 0.135 0.65 18 0.162 0.154 0.462 1.288E-07 0.157 18 0.061 0.053 0.159 4.445E-08 0.054 19 0.062 0.054 0.161 4.479E-08 0.055 0.038 89.58 19 0.028 0.020 0.061 1.708E-08 0.021 20 0.111 0.103 0.310 8.656E-08 0.106 0.363 0.96 20 0.506 0.498 1.495 4.171E-07 0.510 20 0.115 0.107 0.322 8.982E-08 0.110 21 0.282 0.274 0.822 2.295E-07 0.281 0.188 1.09 21 0.032 0.024 0.071 1.967E-08 0.024 21 0.078 0.070 0.209 5.818E-08 0.071 22 0.138 0.130 0.391 1.092E-07 0.133 0.133 0.00 23 0.040 0.032 0.095 2.645E-08 0.032 0.054 0.68 23 0.069 0.061 0.182 5.073E-08 0.062 23 0.021 0.013 0.039 1.097E-08 0.013 24 0.022 0.014 0.043 1.189E-08 0.015 0.046 0.56 24 0.037 0.029 0.086 2.411E-08 0.029 24 0.056 0.048 0.143 4.001E-08 0.049 25 0.038 0.030 0.091 2.528E-08 0.031 0.069 0.33 25 0.068 0.060 0.181 5.048E-08 0.062 25 0.052 0.044 0.133 3.717E-08 0.045 42

PAGE 51

Appendix A (Continued) Table A-3 (Continued) PSD IC value Less blank average Fluoride Ambient HF* Ambient HF** Average RSD*** ug/mL F ug/mL F ug ug/cm 3 ppb ppb 26 0.111 0.103 0.310 8.639E-08 0.106 0.128 0.22 26 0.088 0.080 0.241 6.722E-08 0.082 26 0.075 0.067 0.200 5.592E-08 0.068 27 0.085 0.077 0.232 6.471E-08 0.079 0.061 0.83 27 0.022 0.014 0.041 1.138E-08 0.014 27 0.037 0.029 0.087 2.428E-08 0.030 28 0.137 0.129 0.386 1.076E-07 0.132 0.095 77.09 28 0.065 0.057 0.171 4.772E-08 0.058 *c = 1000*(Q / (DA/L)t) **ug/cm 3 = 1000 mg/m 3 c = ambient HF (ug/cm 3 ) ((mg/m 3 )*24.45)/20.01 = ppm Q = mass uptake (IC value, ug) ppm 1000 = ppb D = coefficient of diffusion for fluoride (0.2236 cm 2 /s); Tate (2002) A = cross sectional area of diffusion path (3.71 cm 2 ); Tate (2002) ***RSD = SD/average L = length of diffusion path (0.6 cm); Tate (2002) t = sampling time (2592000 s) 43

PAGE 52

Appendix A (Continued) Table A-4. Hypothetical flow rate and sampling time calculations to obtain detectable HF amount in sampling train. HF generated Q total [HF] Q PSD Time HF PSD Dilution [HF] PSD [HF] PSD [HF] PSD ng/min ml/min g/ml ml/min minutes g ml g/ml mg/m 3 ppm 0.674 10 0.067 5 10 3.370 3.00 1.123 1123.3 1373 0.674 20 0.034 10 10 3.370 3.00 1.123 1123.3 1373 0.674 40 0.017 20 10 3.370 3.00 1.123 1123.3 1373 0.674 60 0.011 30 10 3.370 3.00 1.123 1123.3 1373 0.674 80 0.008 40 10 3.370 3.00 1.123 1123.3 1373 0.674 100 0.007 50 10 3.370 3.00 1.123 1123.3 1373 0.674 120 0.006 60 10 3.370 3.00 1.123 1123.3 1373 0.674 140 0.005 70 10 3.370 3.00 1.123 1123.3 1373 0.674 160 0.004 80 10 3.370 3.00 1.123 1123.3 1373 0.674 180 0.004 90 10 3.370 3.00 1.123 1123.3 1373 *change in flowrate does not change HF concentration when keeping sampling time constant 0.674 20 0.034 10 10 3.370 3.00 1.123 1123.3 1373 0.674 20 0.034 10 20 6.740 3.00 2.247 2246.7 2747 0.674 20 0.034 10 30 10.110 3.00 3.370 3370.0 4120 0.674 20 0.034 10 40 13.480 3.00 4.493 4493.3 5493 0.674 20 0.034 10 50 16.850 3.00 5.617 5616.7 6866 0.674 20 0.034 10 60 20.220 3.00 6.740 6740.0 8240 0.674 20 0.034 10 70 23.590 3.00 7.863 7863.3 9613 0.674 20 0.034 10 80 26.960 3.00 8.987 8986.7 10986 0.674 20 0.034 10 90 30.330 3.00 10.110 10110.0 12359 0.674 20 0.034 10 100 33.700 3.00 11.233 11233.3 13733 *at least 30 minutes acheives detectable HF for IC analysis; can adjust flow rate for impinger performance HF gen. = HF generation rate; Q = flowrate; [HF] = HF concentration 44

PAGE 53

Appendix A (Continued) Table A-5. Laboratory validation trial I results and calculations. Sample Less avg blank Dilution Total F Time Q Volume Ambient Sampling Ambient HF calculated ID (ug/ml) (ml H 2 O) (ug) (min) (ml/min) of air (ml) HF (ug/ml) Rate* (ml/min) HF (ppm) to Diffuse** (ug) Reflb 0.4413 Reflb 0.5233 Lb 0.235 Lb 0.2318 Refb 0.4848 Refb 0.2198 B 0.5025 b 0.3889 11 3.2449 45 146.0205 695 62.27 43277.7 0.0034 7.620 4.16 19.6 12 -0.3297 45 -14.8365 695 62.27 43277.7 0 0.00 1 6.1647 3 18.4941 695 55.68 38697.6 0.0005 0.61 1 5.8388 3 17.5164 695 55.68 38697.6 0.0005 0.61 21 1.1793 45 53.0685 1260 58.15 73269 0.0007 21.213 0.86 7.3 22 -0.5599 45 -25.1955 1260 58.15 73269 0 0.00 2 5.1009 3 15.3027 1260 50.74 63932.4 0.0002 0.24 2 7.3724 3 22.1172 1260 50.74 63932.4 0.0003 0.37 31 2.4021 45 108.0945 650 58 37700 0.0029 6.021 3.55 15.6 32 -0.2216 45 -9.972 650 58 37700 0 0.00 3 3.1275 3 9.3825 650 51 33150 0.0003 0.37 3 4.4392 3 13.3176 650 51 33150 0.0004 0.49 41 2.1957 45 98.8065 770 26.03 20043.1 0.0049 4.471 5.99 31.3 42 -0.3464 45 -15.588 770 26.03 20043.1 0 0.00 4 4.7363 3 14.2089 770 30.15 23215.5 0.0006 0.73 45

PAGE 54

Appendix A (Continued) Table A-5 (Continued) Sample Less avg blank Dilution Total F Time Q Volume Ambient Sampling Ambient HF calculated ID (ug/ml) (ml H 2 O) (ug) (min) (ml/min) of air (ml) HF (ug/ml) Rate* (ml/min) HF (ppm) to Diffuse** (ug) 4 6.5105 3 19.5315 770 30.15 23215.5 0.0008 0.98 51 4.2797 45 192.5865 680 40 27200 0.0071 3.283 8.68 40.1 52 -0.7097 45 -31.9365 680 40 27200 0 0.00 5 6.043 3 18.129 680 37.2 25296 0.0007 0.86 5 4.525 3 13.575 680 37.2 25296 0.0005 0.61 61 0.4258 45 19.161 550 20.8 11440 0.0017 16.525 2.08 7.8 62 -0.6853 45 -30.8385 550 20.8 11440 0 0.00 6 5.4201 3 16.2603 550 29.3 16115 0.001 1.22 6 4.8804 3 14.6412 550 29.3 16115 0.0009 1.10 *Sampling rate = [ug per PSD/((ug/ml)*min)]; **HF calculated to diffuse = [(ug/ml)*(0.138259*seconds)] 46

PAGE 55

Appendix A (Continued) Table A-6. Laboratory validation trial II results and calculation of fluoride concentrations in air of sampling train. Sample IC Less B Dilution Total F Time Q Volume Ambient HF Sampling Ambient HF Calculated to ID (ug/ml) (ml H 2 O)) (ug) minutes ml/min of air (ml) (ug/ml) Rate* (ml/min) ppm Diffuse** (ug) reflb reflb lb lb ref b -0.305 45 -13.725 720 24 17280 -0.0008 -0.97 ref b -0.395 45 -17.775 720 24 17280 -0.001 -1.26 b -0.125 3 -0.375 b -0.155 3 -0.465 11 -0.195 45 -8.775 560 28 15680 -0.0006 -18.284 -0.68 -2.6 12 -0.195 45 -8.775 560 28 15680 -0.0006 -0.68 1 1.595 3 4.785 1 2.225 3 6.675 21 0.975 45 43.875 730 26 18980 0.00231 3.724 2.83 14.0 22 0.805 45 36.225 730 26 18980 0.00191 2.33 2 1.865 3 5.595 2 2.325 3 6.975 31 1.435 45 64.575 720 28 20160 0.0032 8.345 3.92 19.1 32 2.415 45 108.675 720 28 20160 0.00539 6.59 3 6.795 3 20.385 3 6.035 3 18.105 41 1.835 45 82.575 750 26 19500 0.00423 4.770 5.18 26.3 42 -0.295 45 -13.275 750 26 19500 -0.0007 -0.83 4 4.575 3 13.725 47

PAGE 56

Appendix A (Continued) Table A-6 (Continued) Table A-6. Laboratory validation trial II results and calculation of fluoride concentrations in air of sampling train. Sample IC Less B Dilution Total F Time Q Volume Ambient HF Sampling Ambient HF Calculated to ID (ug/ml) (ml H 2 O)) (ug) minutes ml/min of air (ml) (ug/ml) Rate* (ml/min) ppm Diffuse** (ug) 4 5.525 3 16.575 51 2.365 45 106.425 720 28 20160 0.00528 5.521 6.45 31.5 52 1.255 45 56.475 720 28 20160 0.0028 3.42 5 6.905 3 20.715 5 7.085 3 21.255 61 2.115 45 95.175 735 29 21315 0.00447 5.050 5.46 27.2 62 -0.345 45 -15.525 735 29 21315 -0.0007 -0.89 6 5.415 3 16.245 6 5.635 3 16.905 71 0.955 45 42.975 375 24 9000 0.00478 5.613 5.84 14.9 72 -0.485 45 -21.825 375 24 9000 -0.0024 -2.96 7 3.275 3 9.825 7 3.425 3 10.275 *Sampling rate = [ug per PSD/((ug/ml)*min)]; **HF calculated to diffuse = [(ug/ml)*(0.138259*seconds)] 48

PAGE 57

Appendix A (Continued) 49 Table A-7. Calculation of UTM coordinates from lat/long data: March 2004 PSD deployment. Site Direction Degrees M.m d D.d UTM 1 N 28 0.881 0.0147 28.0147 391323.1700 W 82 6.326 0.1054 82.1054 3099151.5300 2 N 28 0.23 0.0038 28.0038 391312.2200 W 82 6.063 0.1011 82.1011 3097940.1800 3 N 27 59.951 0.9992 27.9992 391730.4500 W 82 5.75 0.0958 82.0958 3097425.8900 4 N 27 59.132 0.9855 27.9855 392238.0000 W 82 4.659 0.0777 82.0777 3095892.3300 5 N 27 59.136 0.9856 27.9856 394018.1900 W 82 4.455 0.0743 82.0743 3095500.4600 6 N 27 59.356 0.9893 27.9893 394356.1800 W 82 4.67 0.0778 82.0778 3096313.3900 7 N 27 58.988 0.9831 27.9831 394005.9100 W 82 4.807 0.0801 82.0801 3095628.5400 8 N 27 58.493 0.9749 27.9749 393771.6700 W 82 4.307 0.0718 82.0718 3094712.9400 9 N 27 58.475 0.9746 27.9746 394587.7400 W 82 4.952 0.0825 82.0825 3094688.9900 10 N 27 58.953 0.9826 27.9826 393543.1800 W 82 5.061 0.0844 82.0844 3095576.9000 11 N 27 59.225 0.9871 27.9871 393360.7400 W 82 5.459 0.0910 82.0910 3096081.2000 12 N 27 58.467 0.9745 27.9745 392699.1800

PAGE 58

50 Appendix A (Continued) Table A-7 (Continued) Site Direction Degrees M.m d D.d UTM W 82 5.497 0.0916 82.0916 3094685.8800 13 N 27 58.457 0.9743 27.9743 392639.9700 W 82 6.089 0.1015 82.1015 3094672.4600 14 N 27 58.907 0.9818 27.9818 391673.7200 W 82 5.567 0.0928 82.0928 3095495.6400 15 N 27 58.901 0.9817 27.9817 392529.2800 W 82 6.101 0.1017 82.1017 3095492.4300 16 N 27 59.235 0.9873 27.9873 391659.5500 W 82 6.276 0.1046 82.1046 3096115.3800 17 N 27 59.697 0.9950 27.9950 391382.0700 W 82 6.311 0.1052 82.1052 3096968.9400 18 N 27 59.861 0.9977 27.9977 391325.7700 W 82 4.966 0.0828 82.0828 3097248.3100 19 N 27 59.705 0.9951 27.9951 393525.9500 W 82 4.902 0.0817 82.0817 3096959.3200 20 N 27 59.649 0.9942 27.9942 393633.2400 W 82 4.666 0.0778 82.0778 3096856.2200 21 N 27 59.217 0.9870 27.9870 394009.7200 W 82 5.303 0.0884 82.0884 3096067.8400 22 N 27 59.353 0.9892 27.9892 392969.4300 W 82 4.107 0.0685 82.0685 3096294.2700 23 N 27 58.921 0.9820 27.9820 394919.4900

PAGE 59

51 Appendix A (Continued) Table A-7 (Continued) Site Direction Degrees M.m d D.d UTM W 82 4.075 0.0679 82.0679 3095496.1300 24 N 27 59.783 0.9964 27.9964 394992.4600 W 82 3.831 0.0639 82.0639 3097087.9500 25 N 28 0.089 0.0015 28.0015 395390.7300 W 82 3.807 0.0635 82.0635 3097652.5900 26 N 28 1.182 0.0197 28.0197 395447.6500 W 82 4.293 0.0716 82.0716 3099675.7900 27 N 28 0.48 0.0080 28.0080 394639.9000 W 82 4.562 0.0760 82.0760 3098383.4500 28 N 28 1.215 0.0203 28.0203 394219.2800 W 82 5.816 0.0969 82.0969 3099764.3700 Calculation Sources: http://www.ento.vt.edu/STS/project/manuals/dms.html and http://www.uwgb.edu/dutchs/UsefulData/UTMConversions1.xls

PAGE 60

Appendix B: Calibration Charts September 2003 Field Sampling Data Analysis Calibration Charty = 0.9873x + 0.0425R2 = 0.99990123456012345Standard Value (ug/ml)IC Value (ug/ml) Figure B-1. March 2004 Field Sampling Data Analysis Calibration Charty = 0.9994x + 0.001R2 = 0.99990.0001.0002.0003.0004.0005.0006.000012345Standard Value (ug/ml)IC Value (ug/ml) Figure B-2. 52

PAGE 61

Appendix B (Continued) Laboratory Validation Trial I Data Analysis Calibration Charty = 1.0122x + 0.0052R2 = 0.99290.00000.50001.00001.50002.00002.50003.00003.500000.511.522.53Standard Value (ug/ml)IC Value (ug/ml) Figure B-3. Laboratory Validation Trial II Data Analysis Calibration Charty = 1.0137x + 0.005R2 = 0.99480.0000.5001.0001.5002.0002.5003.0003.50000.511.522.53Standard Value (ug/ml)IC Value (ug/ml) Figure B-4 53

PAGE 62

Appendix B (Continued) Dual Rotameter Calibration Charty = 1.6473x + 7.9119R2 = 0.9958253035404550556065101520253035Rotameter ReadingFlow Rate (ml/min) Figure B-5. 54