Citation
Determination of methanol and ethanol in ambient air by gas chromatography with flame ionization detection

Material Information

Title:
Determination of methanol and ethanol in ambient air by gas chromatography with flame ionization detection
Creator:
Kirschenman, Debra Lee
Publication Date:
Language:
English
Physical Description:
xv, 98 leaves : illustrations ; 29 cm

Subjects

Subjects / Keywords:
Automobiles -- Motors -- Exhaust gas ( lcsh )
Methanol as fuel ( lcsh )
Alcohol as fuel ( lcsh )
Alcohol as fuel ( fast )
Automobiles -- Motors -- Exhaust gas ( fast )
Methanol as fuel ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 97-98).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Science, Chemistry.
Thesis:
Department of Chemistry
Statement of Responsibility:
by Debra Lee Kirschenman.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
34331857 ( OCLC )
ocm34331857
Classification:
LD1190.L46 1995m .K57 ( lcc )

Downloads

This item has the following downloads:


Full Text
DETERMINATION OF
METHANOL AND ETHANOL
IN AMBIENT AIR BY
GAS CHROMATOGRAPHY WITH
FLAME IONIZATION DETECTION
ty
Debra Lee Kirschenman
B. A., University of. Colorado at Denver, 1991
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Chemistry
1995


This thesis for the Master of Science
degree by
Debra Lee Kirschenman
has been approved for the
Graduate School
by
y-za-ys'
Date


Kirschenman, Debra Lee (M.S., Chemistry)
Determination of Methanol and Ethanol in the Ambient Air by Gas
Chromatography with Flame Ionization Detection
Thesis directed by Associate Professors Larry G. Anderson and
John A. Lanning
ABSTRACT
The increased use of methanol and ethanol gasoline blended
fuels is expected to increase the concentrations of these alcohols in
ambient air. Methanol and ethanol are expected to be at low ppbv
concentrations in the presence of a large number of hydrocarbon and
carbonyl species. Currently there is no specific and sensitive technique
available for the quantification of methanol and ethanol in ambient
air.
Our objectives were to develop a method capable of quantifying
methanol and ethanol at the ppbv levels by using a standard gas
chromatograph with a flame ionization detector (GC-FID), and to
determine these alcohols in the ambient air of Denver, Colorado.
in


A standard GC-FID is not sensitive enough to quantify methanol
and ethanol at the ppbv level without a preconcentration procedure.
First we bubbled ambient air through an impinger containing a liquid
solvent capable of absorbing methanol. Experiments indicated that the
impinger preconcentration method was not sensitive enough to
quantify methanol in ambient air, so we tried another method of
preconcentrating the sample. We attempted to trap methanol and
ethanol in ambient air samples onto thermal desorption tubes. The
analytes were thermally desorbed and cryofocused onto the top of a GC
capillary column. Rapid heating of the top of the GC column injects
the analytes. The detection limits of the system were in the ppbv range
for both methanol and ethanol.
Ten ambient air samples of three to twelve liters were collected
at the Auraria Ambient Air Sampling Station on the University of
Colorado Campus in downtown Denver, Colorado/ from November 11
through December 3, 1994. The concentration of methanol in Denver's
air ranged from 1.7 to 12 ppbv with an average of 5.9 3.7 ppbv, and
ethanol ranged from 0.70 to 3.4 ppbv with an average of 2.3 1.2 ppbv.
IV


The form and content of this abstract are approved. We
recommend its publication.
v


DEDICATION
To my parents Elton and Helen
for your generous love and support.
You're the best in the world.


CONTENTS
CHAPTER
L INTRODUCTION....................................1
Using Alternative Fuels to Reduce the United
States' Dependence on Foreign Oil Imports.......2
Using Oxygenated Fuels to Combat the Increasing
Air Pollution in the United States' Major
Metropolitan Areas..............................3
Using Oxygenated Fuels to Reduce Ozone
Concentrations............................4
Using Oxygenated Fuels to Reduce Carbon
Monoxide Concentrations...................6
Measurement Techniques..........................7
Research Objectives for the Impinger
Preconcentration Method.........................9
Research Objectives for the Solid Adsorbent
Preconcentration Method........................11
II. BACKGROUND AND THEORY..........................14
Impinger Preconcentration Method...............14
Solid Adsorbent Preconcentration Method........16
The Sampling Process.....................16
Thermal Desorption and Cryofocusing
Process..................................18
Gas Chromatography and Flame Ionization
Detection................................23
v i i


El. EXPERIMENTAL SECTION...................................24
Impinger Preconcentration Method...................24
Experimental Apparatus for the Impinger
Preconcentration Method......................25
Preparation of Methanol Standards for the
Impinger Preconcentration Method.............25
Solid Adsorbent Preconcentration Method............26
Experimental Sampling Apparatus for the
Solid Adsorbent Preconcentration Method......28
Preparation of Methanol Standards for the
Solid Adsorbent Preconcentration Method......28
IV. RESULTS AND DISCUSSION.................................30
Impinger Preconcentration Method...................30
Solid Adsorbent Preconcentration Method............40
Identification of the Methanol and Ethanol
Peaks........................................40
Reproducibility of the Solid Adsorbent
Preconcentration Method......................45
Retention Times of Methanol and Ethanol in
Ambient Air..................................52
Collection Efficiency Study..................53
Calibration Curve for Methanol and Ethanol...60
Ambient Air Sample Studies...................70
v i i i


IV. CONCLUSIONS..................................86
Recommendations for Future Studies..........91
APPENDIX
A. Estimated Volume of Air Needed to Quantify
Methanol in Ambient Air Using the Impinger
Preconcentration Method.....................94
B. Moles of Water in Ambient Air................96
REFERENCES........................................... 97


FIGURES
Figure
2.1 A block diagram of the apparatus used to load
standards onto thermal desorption tubes...........18
2.2 A block diagram of the apparatus used to load
ambient air samples onto thermal desorption
tubes.............................................18
2.3 A block diagram of the thermal desorption and
cryofocusing apparatus............................19
4.1 A gas chromatogram of a standard solution
containing 0.25 M methanol in benzyl alcohol......31
4.2 A gas chromatogram of a standard solution
containing 0.12 M methanol in benzyl alcohol......32
4.3 A gas chromatogram of a blank solution containing
only benzyl alcohol...............................33
4.4 A gas chromatogram of a standard solution
containing 2.5 mM methanol in benzyl alcohol......35
4.5 A gas chromatogram of a standard solution
containing 1.2 mM methanol in benzyl alcohol......36
4.6 A gas chromatogram of a standard solution
containing 0.25 mM methanol in benzyl alcohol.....37
4.7 A gas chromatogram of a blank solution containing
only benzyl alcohol...............................38
4.8 A gas chromatogram of a blank solution containing
only Millipore Water..............................41
x


4.9 A gas chromatogram of a standard solution
containing 0.97 mM methanol in Millipore Water..42
4.10 A gas chromatogram of a standard solution
containing 0.67 mM ethanol in Millipore Water...43
4.11 A gas chromatogram of a nitrogen blank..........46
4.12 A gas chromatogram of a standard solution
containing 2.4 mM methanol and 1.7 mM ethanol
in organic free water. Run number one...........47
4.13 A gas chromatogram of a standard solution
containing 2.4 mM methanol and 1.7 mM ethanol
in organic free water. Run number two...........48
4.14 A gas chromatogram of a standard solution
containing 2.4 mM methanol and 1.7 mM ethanol
in organic free water. Rim number three and
primary recovery sample.........................49
4.15 A gas chromatogram of an ambient air sample,
11/5/1994.......................................54
4.16 A gas chromatogram of an ambient air sample
spiked with 2.3 x 10"9 moles of methanol and
1.7 x 10'9 moles of ethanol, 11/5/1994..........55
4.17 A gas chromatogram of an ambient air sample
spiked with 4.9 x 10'9 moles of methanol and
3.4 x 10'9 moles of ethanol, 11/5/1994..........56
4.18 A gas chromatogram of the secondary recovery .
sample........................................ 58
4.19 A gas chromatogram of a conditioning blank......59
4.20 A gas chromatogram of standard one containing
2.4 mM methanol and 3.3 mM ethanol.............62
x i


4.21 A gas chromatogram of standard two containing
1.1 mM methanol and 1.7 mM ethanol..............63
4.22 A gas chromatogram of standard three containing
0.72 mM methanol and 1.0 mM ethanol.............64
4.23 A gas chromatogram of standard four containing
0.24 mM methanol and 0.33 mM ethanol............65
4.24 A gas chromatogram of a blank containing only
organic free water..............................66
4.25 Calibration curves for the concentration of
methanol and ethanol versus area................68
4.26 A gas chromatogram of an ambient air sample,
10/14/1994.................................... 71
4.27 A gas chromatogram of an ambient air sample,
10/27/1994......................................73
4.28 A gas chromatogram of an ambient air sample,
10/28/1994.................................... 74
4.29 A gas chromatogram of the first ambient air
sample, 11/2/1994...............................75
4.30 A gas chromatogram of the second ambient air
sample, 11/2/1994............................. 77
4.31 A gas chromatogram of an ambient air sample,
11/19/1994.................................... 78
4.32 A gas chromatogram of the first ambient air
sample, 11/30/1994..............................80
4.33 A gas chromatogram of an ambient air sample,
12/3/1994.......................................82


TABLES
Table
1.1 Reaction rates for selected hydrocarbons with the
hydroxyl radical.....................................5
2.1 Values for the parameters a and b in Equation 2.1
for various compounds of interest including
methanol............................................22
2.2 Vapor Pressure in mm Hg at varying temperatures
for various compounds of interest including
methanol.......................................... 23
3.1 The gas chromatography and recorder operating
parameters for the impinger preconcentration
method..............................................24
3.2 The thermal desorption operating parameters for
the solid adsorbent preconcentration method.........27
3.3 The gas chromatography operating parameters for
the solid adsorbent preconcentration method.........27
4.1 The result of the reproducibility study............50
4.2 The concentration of methanol and ethanol and
the area of the standard solutions analyzed from
10/11/94 through 12/3/94............................51
4.3 The result of the collection efficiency study......57
4.4 The concentration of methanol and ethanol for
each standard with its corresponding area and
retention time on 11/19/94..........................61
XII!


4.5 The concentration of methanol and ethanol for
each standard with its corresponding area and
retention time on 11/30/94............................61
4.6 The concentration of methanol and ethanol for
each standard with its corresponding area and
retention time on 12/3/94.............................67
4.7 The least squares linear regressions and correlation
coefficients for the calibration curves of methanol
and ethanol versus area...............................69
4.8 Summary of the concentration of methanol and
ethanol in ten air samples collected at the Auraria
Ambient Air Sampling Station..........................83
XIV


ACKNOWLEDGMENTS
Special appreciation to Professor Larry G. Anderson for his
support, guidance, and understanding throughout the duration of this
research project. Thanks to Professor John A. Lanning for his advice
and electronics help. Thanks to Professor Doris R. Kimbrough for her
assistance and encouragement.
Special thanks to Dr.. Donald Zapien for the use of his organic
free water. Thanks to Dr. Robert R. Meglen for giving me confidence
in myself and my abilities. Thanks to Dr. Larry L. Jackson for his advice
and reassurance during the course of this research.
I would like to express my appreciation to my parents, Elton and
Helen Kirschenman for their love, understanding, and support.
Finally, I would like to thank Greg and Cris Kirschenman for their
support and encouragement during my studies.
xv


CHAPTER I
INTRODUCTION
The United States' dependence on foreign oil imports and
increasing air pollution problems in its major metropolitan areas are
driving the development of motor vehicle technologies that use
alternative fuels instead of gasoline. The increased use of alternative
fuels might increase the concentrations of the components of these
alternative fuels in ambient air. The quantification of the alternative
fuels in ambient air is necessary in order to study their role in the
increasing air pollution problems.
Two of the alternative fuels currently in use are methanol and
ethanol. The increased use of methanol and ethanol as alcohol
gasoline blended fuels, might increase the concentrations of these
alcohols in ambient air. Several separate processes contribute to
increase in concentration of methanol and ethanol in ambient air:
evaporative losses from the fuel system of motor vehicles, emissions
of unburned fuel by motor vehicles, and evaporative losses during the
fueling process of motor vehicles.
1


Currently there are no reliable methods for determining the
concentrations of methanol and ethanol in ambient air. Previously
developed methods do not have the sensitivity necessary for
quantifying the concentration of methanol and ethanol in ambient air.
In this research we developed a method to quantify the concentration
of methanol and ethanol in ambient air of Denver by using a standard
gas chromatograph equipped with a flame ionization detector (GC-FID).
Using Alternative Fuels to Reduce the
United States/ Dependence on Foreign Oil Imports.
This country's dependence on foreign oil imports drives the
development of motor vehicle technologies that use alternative fuels
instead of gasoline. The United States initiated the Alternative Motor
Fuels Act of 1988 (AMFA) to decrease the country's dependence on
foreign oil imports. The AMFA required the Department of Energy
(DOE) develop new motor vehicle technologies that use alternative
fuels (1). The DOE, through the National Renewable Energy Laboratory
(NREL), supports a number of projects that are studying the use of
methanol and ethanol as alternative motor vehicle fuels (1). The
General Service Administration (GSA) currently has alternative fuels
fleets in seven different areas: Detroit, Atlanta, Denver, Chicago,
Houston, New York, and Washington, DC (2). The Denver Fleet
2


Management Center has 229 light duty vehicles, 131 operating on M-85,
involved in an emission and data collection test program (2). The
Denver Fleet Management Center also has 226 light duty vehicles, 162
operating on M-85, not involved in an emission and data collection
test program (2). M-85 is a methanol gasoline blended fuel composed
of 85% methanol and 15% gasoline. Currently, M-85 is available only
at two Conoco service stations in the Denver, one is at the intersection
of 6th Avenue and Sheridan and the other is at the intersection of
Colfax and Josephine.
Using Oxygenated Fuels to Combat the Increasing
Air Pollution in the United States7 Major Metropolitan Areas
The air pollution problem in some of this country's major
metropolitan areas has driven the development of oxygenated gasoline
programs. Emissions from gasoline powered motor vehicles
contribute to a large portion of the hydrocarbons (HC), nitrogen oxides
(NOx), and carbon monoxide (CO) in ambient air (3). Oxygenated fuels
are being used to reduce ozone concentrations in cities that violate the
national air quality standard for ozone (O3). Oxygenated fuels, such as
alcohol gasoline blended fuels, are also being used in an attempt to
reduce carbon monoxide concentration in cities that violate the
national air quality standard for carbon monoxide.
3


Using Oxygenated Fuels to Reduce Ozone Concentrations
In areas where ozone concentrations exceed the National
Ambient Air Quality Standards for Ozone, there has been an attempt to
reduce ozone levels using oxygenated gasolines. Recently 101 U.S.
cities failed to meet the National Ambient Air Quality Standards for
ozone (4).
The burning of gasoline by automobiles produces hydrocarbons,
carbon monoxide, and nitrogen oxides. Hydrocarbons or carbon
monoxide can react with nitrogen oxides in ambient air to produce
ozone, a suspected player in the photochemical smog problem in
southern California cities. The use of an oxygenated gasoline reduces
the amount of hydrocarbons and increases the amount of oxygenate in
motor vehicle exhaust. The idea is that the oxygenate will react slower
than the hydrocarbons do with nitrogen oxide, there by reducing the
level of ozone in the local ambient air.
The combustion of alcohol gasoline blended fuels or gasoline by
motor vehicles produces alcohol and gasoline emissions in the local
ambient air. Both oxidize to carbon dioxide in the atmosphere. The
first step in the oxidation of a hydrocarbon is the reaction with
hydroxyl radical (OH). A typical alkane reacts about 10 times faster than
methanol and about 3 times faster than ethanol with the hydroxyl
radical. Alkenes and alkynes react even faster than alkanes. Table 1.1
4


gives the reaction rates for selected hydrocarbons with the hydroxyl
radical.
Table 1.1 Reaction rates for selected hydrocarbons with the
hydroxyl radical. (From Finlayson-Pitts & Pitts,
"Atmospheric Chemistry" (3)).
Compound Reaction Rate
Methanol 0.90 x 10"12 (cm^ molecules"! s-!)
Ethanol 2.90 x 10"12 (cm3 molecules"! s-!)
Carbon Monoxide 0.15 x 10"!2 (cm3 molecules"! s"!)
n-Octane 8.72 x 10"!2 (cm3 molecules"! s"!)
Ethene (Slowest Alkene) 8.54 x 10"!2 (cm3 molecules"! s"!)
1-Butyne 8.04 x 10"!2 (cm3 molecules"! s-l)
One of the compounds produced during the oxidation of a hydrocarbon
is the hydroperoxyl radical (HO2), a major player in the photochemical
ozone pollution problem. The reaction of the hydroperoxyl radical
with nitrogen oxide produces ozone through the following set of
chemical reactions.
H02 + NO N02 + HO
NOz + hV NO + O
O + 02 + M O3 + hV
Alcohol fuel combustion emissions react slower with the hydroxyl
radical than gasoline combustion emissions, and in theory are
5


transported farther from their source. The transport effect might
reduce local ozone concentrations near the source.
Using Oxygenated Fuels to Reduce Carbon Monoxide Concentrations
To combat the increasing air pollution problem there has been
an attempt to decrease auto emissions of hydrocarbon and carbon
monoxide from automobiles using oxygenated gasolines. Carbon
monoxide produced by the incomplete combustion of gasoline is a
major pollutant in many cities (3). The idea is that increasing the
concentration of oxygen in the fuel reduces the concentration of carbon
monoxide in the motor vehicle exhaust by burning the fuel completely
to carbon dioxide. Alcohol gasoline mixtures can reduce the level of
carbon monoxide in automobile exhausts by up to 40 to 50 percent (5).
Oxygenated gasoline is being used in an attempt to reduce the
levels of carbon dioxide in areas that violate the National Ambient Air
Quality Standard for Carbon Monoxide. There have been 37 areas that
have implemented oxygenated gasoline programs, as required by the
Clean Air Act Amendments of 1990 (6). The Denver-Boulder area in
Colorado is one of the Oxygenated Gasoline Program Areas in the
United States (6).
The Colorado Oxygenated Gasoline Program started on January
1, 1988, ran through March 1,1988, and has been in effect from
6


November 1st through March 1st every year since its inception. The
most popular oxygenates used in Colorado are methyl tertiary butyl
ether (MTBE) and ethanol.. In the winter of 1992 to 1993, MTBE
accounted for 55% of the market share and ethanol accounted for 45%
of the market share (6). In the winter of 1993 to 1994, MTBE accounted
for 30% of the market share and ethanol accounted for 70% of the
market share (7). Currently, Colorado Regulation No. 13 requires that
in program areas there is a minimum oxygen content of 2.7% per
gallon of gasoline (7). In practical terms this means 10% ethanol per
gallon or 15% MTBE per gallon (7).
Measurement Techniques
The early methods developed for the analysis of methanol were
modifications of methods developed for the determination of
formaldehyde. The methods involved the oxidation of methanol to
formaldehyde followed by spectrophotometric analysis. Bhatt and
Gupta (8) oxidized methanol with acidic potassium permanganate and
removed the excess oxidant before analysis by the colormetric reaction
with oxalydihydrazide and copper II in acetate medium. Verma and
Gupta (9) developed a sensitive method that determined methanol
indirectly by bubbling air through midget impingers containing an
oxidizing solution that oxidized methanol to formaldehyde, before
7


analysis by spectrophotometry in acidic medium with
p-aminoazobenzene and sulfur dioxide.
Maeda, Fujio, Suetaka, and Makoto (10) developed a method for
the determination of trace amount of alcohols based on the formation
of alkyl nitrite when alcohols react with nitrogen dioxide (NO2), before
analysis by gas chromatography with electron-capture detection.
The above methods have the disadvantage of being complicated
and labor intensive. Another disadvantage is that other compounds
interfere with the analysis. For example, glycerols and glycols can be
oxidized to formaldehyde.
Hanst, Wong, and Bragin (11) quantified the concentration of
methanol in Los Angeles smog using a Fourier transforming long path
infrared spectrometry with a 23 m long base path and a 1260 m optical
path. This technique uses expensive and delicate instrumentation not
easily transported to different field sites.
Methods exist for the determination of methanol and ethanol in
automotive vehicle exhaust (12 ,13,14). They bubbled diluted vehicle
exhausts through a series of midget impinger traps containing
deionized water followed by GC-FID analysis. The methods relied on
the fact that alcohols are highly soluble in water, and hydrocarbons are
insoluble in water. The major disadvantage of this method is that it is
not sensitive enough to measure the low part per billion by volume
(ppbv) concentrations of methanol and ethanol expected in ambient
air. Further complicating the matter is the decreased response of the
8


FID when large amounts of water are present. Water as a collector has
the added disadvantage that its vapor pressure is too large to use as an
impinger solvent without an ice bath to minimize evaporation. There
is also the possibility of interference from low molecular weight
carbonyls and ketones, which are soluble in water.
Haky and Stickney (15) developed a method for the
determination of residual solvents, such as in bulk pharmaceuticals, by
dissolving samples in benzyl alcohol followed by GC-FID analysis. The
advantage of this method is the elimination of water collected from the
sample, because water is insoluble in benzyl alcohol. The disadvantage
of this method is that the method does not have the required
sensitivity that would enable the quantification of methanol in
ambient air.
Research Objectives for the Impinger Preconcentration Method
The focus of this research was to develop a method to measure
the concentration of methanol by using a modification of the bubbler
method described above. We bubbled ambient air through a midget
impinger containing benzyl alcohol followed by GC-FID analysis. The
method should be simple, selective for methanol, sensitive enough to
measure ppbv concentrations in ambient air, and reproducible.
9


The use of the impinger preconcentration method coupled with
GC-FID approach did not appear to have the sensitivity required to
quantify methanol in ambient air. Preliminary studies and
calculations indicated the volume of an ambient air sample was too
large to quantify methanol, as discussed in further detail later.
The following paragraphs describe the constraints observed
during the development of impinger preconcentration method.
A limit to this method is the sensitivity of the GC-FID to
methanol in benzyl alcohol.
For the method to be useful it should be able to quantify the
concentration of methanol in ambient air in real time, for example
collect an ambient air sample every four hours. Thus, the volume of
air sampled and the volume of benzyl alcohol in the impinger trap
should reflect this restraint.
The flow rate of the air through the impinger trap should be
slow enough to extract all the methanol from air into the benzyl
alcohol.
The method should be easily automated for routine air
monitoring, with a minimal amount of labor.
The above constraints restrict any method using impinger traps
to quantify methanol in ambient air. The major problem of any
method using this approach would be its lack of sensitivity.
We determined that impinger preconcentration method does
not have the sensitivity required to quantify ppbv concentrations of
10


methanol in ambient air. The volume of an individual ambient air
sample was much too large, for real time analysis.
Research Objectives for the
Solid Adsorbent Preconcentration Method
Instrumentation capable of purge and trap analysis is
commercially available. Many of these instruments purge a sample
with helium while the sample is being heated. A cryogenic trap
concentrates analytes purging from the sample. Variations of this
system have been used to quantify volatile polar organic compounds in
ambient air.
Fung (16) developed a method for the measurement of alcohols
and MTBE in ambient air. They cryogenically concentrated samples
before analysis by GC-FID. The method employed a two dimensional
column system that reduced the analysis time to under 10 minutes.
The major disadvantage of this method is the complexity of the two
dimensional GC system.
Kelly, Callahan, Plell, and Evans (17) developed a method for the
determination of polar volatile organic compounds, including
methanol and ethanol, in ambient air. They collected ambient air
samples in evacuated canisters over a 24 hour period and transferred
them to a two-stage solid adsorbent trap followed by thermal
11


desorption, cryogenic concentration, and gas chromatography with ion
trap detection. For us the problem with this method was the
inaccessibility of a gas chromatograph equipped with an ion trap
detector.
The primary focus of our research was to develop a method to
measure the amount of methanol and ethanol in ambient air using the
two-stage trap and purge method of Kelly, Callahan, Plell, and Evans
(17), but utilizing a GC-FID. The method should be simple, selective
for methanol and ethanol, sensitive enough to measure ppbv
concentration in ambient air, and reproducible.
The following paragraphs describe the constraints observed
during the development of the solid adsorbent preconcentration
method.
A limit to this method is the sensitivity of the gas GC-FID to
methanol and ethanol in the presence of large amounts of water,
because trapping air samples on a solid adsorbent also traps the water
in the air. Excessive water could impair the performance of the
column and detector in the GC.
For the method to be useful, it should be able to quantify the
concentration of methanol and ethanol in ambient air with reasonable
time resolution, for example collect an ambient air sample every four
hours. Thus, the volume of air sampled and the time required to
analyze a sample should reflect this restraint.
12


The flow rate of the air through the trap should be slow enough
to trap all the methanol and ethanol in ambient air and not exceed the
efficiency of the trap.
The method should be easily automated for routine air
monitoring, with a minimal amount of labor.
The restriction imposed on the development of the solid
adsorbent preconcentration method would be mandatory to any
method designed to quantify volatile polar organic compounds in
ambient air. Our proposed solid adsorbent preconcentration method
appeared well suited for the routine analysis of alcohols, methanol and
ethanol, in ambient air.
13


CHAPTER II
BACKGROUND AND THEORY
Impinger Preconcentration Method
The initial objective of this research was to develop a technique
capable of detecting methanol in ambient air at ppbv concentrations by
using standard gas chromatography instrumentation. However, a
GC-FID is not sensitive enough to detect methanol at the ppbv
concentrations, thus a method of concentrating the methanol would be
necessary. Bubbling ambient air through an impinger containing a
liquid solvent that would absorb the methanol seemed a practical way
to approach this problem.
The impinger solvent used would have to meet the following
requirements. Methanol should be soluble enough in the solvent to
enable its use as a quantitative technique. The solvent should be
selective for alcohols, that is not dissolve other air pollutants to an
appreciable extent. The vapor pressure of the solvent needs to be high
enough to keep the evaporation of the solvent to minimum. Keeping
14


a solvent with a low vapor pressure cool to minimize evaporation
might be impractical. The solvent should not interfere with the
detection of methanol by the FID.
The first choice of an impinger solvent was water, because
methanol is highly soluble in water. Water would also be somewhat
selective since hydrocarbons are relatively insoluble in it. One
disadvantage of water is that its vapor pressure is too large to use as an
impinger solvent without an ice bath to minimize evaporation.
Another disadvantage is that water adversely affects the performance
of the FID even though it is non-detectable, and could extinguish the
flame.
Benzyl alcohol is an impinger solvent that would meet the
solvent requirements listed above. Methanol is highly soluble in
benzyl alcohol. Water is highly insoluble in benzyl alcohol,
eliminating waters adverse effect on the FID. The vapor pressure of
benzyl alcohol is small enough to keep evaporation to minimum
without cooling. Benzyl alcohol should not interfere with the
detection of methanol, because it should elute after methanol. One
disadvantage of benzyl alcohol is that it is not selective for methanol,
because anything from slightly polar methanol to highly non-polar
hydrocarbons will dissolve in it.
Initial experiments indicated that the impinger method was not
sensitive enough to quantify the concentration of methanol in ambient
air. The volume of an individual ambient air sample was too large.
15


Solid Adsorbent Preconcentration Method
The focus of the solid adsorbent preconcentration method was to
develop a procedure that did not require solvent impinger
preconcentration, but still used standard gas chromatography
instrumentation. Thus, we needed another method of concentrating
methanol and ethanol in ambient air. The use of solid phase
extraction of some sort seemed a practical approach to the
preconcentration needs of this research. Trapping ambient air on a
solid adsorbent followed by thermal desorption, cryofocusing, and
GC-FID analysis seemed to be a practical approach to this problem. The
solid adsorbent preconcentration method divides into 3 separate
processes: sampling, desorption and cryofocusing, and GC-FID analysis.
The Sampling Process
We used solid phase extraction for the sampling process. The
idea of solid phase extraction is to pass liquid or gas samples through a
solid that adsorbs analytes of interest.
Samples trapped on solid desorption tubes are desorbed,
extracted with and an organic solvent, followed by gas chromatography
or high performance liquid chromatography. The solvent extraction
method it is not sensitive enough to quantify the concentration of
16


methanol and ethanol in ambient air similar to the impinger
preconcentration method.
We trapped samples on thermal desorption tubes. A thermal
desorption instrument desorbs the analytes and transfers them to the
GC-FID. This method has the advantage of preconcentrating the
sample with no extraction into a liquid solvent.
In this research we used thermal desorption tubes to trap
methanol and ethanol in ambient air samples. We used thermal
desorption tubes packed with 300 mg of Carbopack B and 300 mg of
Carbosieve S-m. Carbopack B is a graphitized black carbon adsorbent
used for trapping airborne organic compounds, from C4-C5 compounds
to polychlorinated biphenyls. Carbosieve S-in is a spherical carbon
molecular sieve capable of trapping C2 hydrocarbons. Carbopack B and
Carbosieve S-III have similar thermal desorption temperature limits
(400C) and hydrophobic properties.
We loaded thermal desorption tubes by preparing standards to
known concentrations and injected them into a flowing stream of
either helium or nitrogen gas. The helium or nitrogen gas containing
the evaporating methanol and ethanol passed through thermal
desorption tubes. Figure 2.1 illustrates a block diagram of the standard
loading apparatus. We used a vacuum pump to draw ambient air
samples through thermal desorption tubes. Figure 2.2 illustrates a
block diagram of the ambient air sample loading apparatus.
17


Figure 2.1
A block diagram of the apparatus used to load
standards onto thermal desorption tubes.
Injector
] Port
Thermal
Desorption
Tube
Vent To
The Air
Figure 2.2 A block diagram of the apparatus used to load ambient
air samples onto thermal desorption tubes.
Thermal Desorption and Crvofocusing Process
We used a thermal desorption instrument to desorb and
cryofocus the analytes from the thermal desorption tubes. Figure 2.3
illustrates a block diagram of the thermal desorption and cryofocusing
apparatus.
18


Figure 2.3
A block diagram of the thermal desorption and
cryofocusing apparatus.
Heating the thermal desorption tube while carrier gas passes
through the tube, effectively desorbs the trapped analytes. The thermal
desorption temperature was 100C. At 100C methanol and ethanol
will desorb with high efficiency from the solid adsorbent, because
methanol vaporizes at 64.7C and ethanol vaporizes at 78.5C (18).
Compounds with boiling points above 100C would not desorb.
The carrier gas and desorbed analytes from the thermal
desorption tubes then passes through Cryotrap 1. Cryotrap 1
cryofocuses the analytes while venting purge gas and other compounds
to air. Rapid heating of Cryotrap 1 transfers the analytes to Cryotrap 2.
Cryotrap 2, the capillary interface, cryofocuses the analytes onto a
cooled short section of column or pre-column. Rapid heating of the
short section of column or pre-column injects the sample. Pre-column
19


cryofocusing requires attaching an uncoated fused silica tubing to the
column. Cryofocusing occurs on the pre-column. Tekmar introduced
an automated version using pre-column cryofocusing in 1983 (19 ,20).
Pre-column cryofocusing systems do not require controlling the
temperatures outside the GC oven. Pre-column focusing has two
major disadvantages. The first disadvantage is that the pre-column to
column connection must have zero dead volume. The second
disadvantage is the lack of liquid phase, which can aid in trapping
analytes, limits the capacity of the cold trap.
Cryofocusing directly on column simplifies and improves the
focusing process (21). Direct column cryofocusing eliminates the dead
volume problem and leaks downstream because no unions are
necessary. Direct column focusing also has the advantage of having a
stationary phase present, which can improve the trapping efficiency by
acting as an adsorbent (20). The trapping of analytes on column
requires carefully controlling the temperature of all parts of the
column outside the GC oven. The Tekmar 2000 Series Capillary
Interface, which we used in this research, has an advanced cryofocusing
trap mounted in a special housing that allows the sample to be re-
focused with outstanding results (20). The GC column is passed
through an unused injector port to the capillary interface and directly
connected to the transfer line from the thermal desorption instrument.
The cryotrap is mounted on top of the GC in a configuration that
allows the cryotrap to sit directly on top of the injector port, which
20


eliminates any possibility of cold spots outside the GC oven. The liquid
nitrogen used to cool the trap is passed through a 1/4" tube from the
bottom to the top. The trap is heated differentially from the top to the
bottom by a heating wire when power is applied. The differential
heating, heats the inlet of the trap faster than the outlet. This creates a
thermal gradient in the trap such that the inlet is always hotter than
the outlet. The thermal gradient ensures that the sample is always
moving from a hotter to a colder region that compresses compounds
moving through the trap into a tight slug before they entering the GC
oven. With the proper establishment of the gradient heat up rate, the
heat up rate of the cryotrap becomes relatively unimportant. No
differences in chromatographic efficiency were observed for heat up
rates from 100C/min to 2000C/min. (20).
The use of pre-column trapping on an uncoated fused silica
greatly simplifies the design requirements, but the absence of any liquid
phase limits the trapping mechanism to simple condensation. Highly
concentrated or extremely volatile samples often exceed the capacity of
the cold trap, and the resultant breakthrough produces split peaks and
generally poor resolution (20). By trapping on column, breakthrough
occurs only when the capacity of the column is overloaded. With on
column trapping the capacity of the cryotrap is similar to the capacity of
the column.
The solid adsorbent preconcentration method employs two
cryotraps used to concentrate the desorbed sample. The cryotraps were
21


cooled to -125C, to minimize breakthrough of analytes during the two
cryofocusing steps. Breakthrough on the cryotrap is a function of the
cryotrap's temperature and the vapor pressure of the analytes.
Equation 2.1 gives the vapor pressure of a compound at a specific
temperature (18).
logP = (-0.05223 a/T) + b 2.1
Table 2.1 contains values for the parameters a and b in Equation 2.1 for
some typical compounds. Table 2.1 lists the calculated vapor pressure
of various compounds using the values for a and b along with
Equation 2.1.
Table 2.1 Values for the parameters a and b in Equation 2.1 for
compounds of interest including methanol (18).
Compound Formula Temp. Range C a b
Acetaldehyde C2H4O -24.3 to +27.5.1iq 27,707 7.8206
Methane CH4 -174 to -163 liq. 8,516.9 6.8626
Methanol CH3OH -10 to +80 liq. 38,324 8.8017
Propane C3H8 -136 to-40 liq. 19,037 7.217
n-Propanol C3H7OH -45 to -10 liq. 47,247 9.5180
22


Table 2.2
Vapor Pressure in mm Hg at varying temperatures for
compounds of interest including methanol.
Compound -75 C -100 C -125 C -150 C
Acetaldehyde 3.3 0.289 1.1 x 10 "2 1.2x10 -4
Methane 4.1x10 4 2.0x10 4 7.2x10 3 1.8x10 3
Methanol 5.0 x 10 2 1.7xl0-3 1.9 x 10 -5 3.5 x 10 8
Propane 160 30 3.2 0.14
n-Propanol 1.2x10-3 1.8x10-3 7.2xl0-8 3.0 x 10 11
The idea was to find a cryotrapping temperature that would
efficiently trap methanol and ethanol, while minimizing the
consumption of liquid nitrogen used to cool the traps. We used a
cryotrapping temperature of -125C for both cryotraps. At -125C the
breakthrough of methanol and ethanol should be negligible.
Gas Chromatography and Flame Ionization Detection
We used a GC-FID to quantify the analytes. We used DB-1
capillary column from J & W Scientific to separate the analytes. A DB-1
column contains dimethylsilicone, a non polar stationary phase, and is
suitable for the analysis of alcohols. A compound's vapor pressure and
solubility in the stationary- phase influence its retention on a DB-1
column. Thus, polar volatile compounds will elute faster than non-
polar non-volatile compounds from a DB-1 column.
23


CHAPTER III
EXPERIMENTAL SECTION
Impinger Preconcentration Method
The study was accomplished using a Hewlett-Packard 5790 A
Series Gas Chromatograph equipped with a flame ionization detector.
The analytical column was a Hewlett Packard HP-1 (12 m x 0.2 mm x
0.33 um film thickness). The carrier gas was nitrogen. The detector
output was recorded by a Omni Scribe Series D5000 Recorder. Table 3.1
lists the gas chromatography and recorder operating parameters.
Table 3.1 The gas chromatography and recorder operating
parameters for the impinger preconcentration method.
Column Flow Rate 1.4 ml/min
Septum Purge Flow Rate 4.5 ml/min
Split Vent Flow Rate 10.3 ml/min
Injection Volume 10 ul
Injector Temperature 250C
Detector Temperature 275C
24


Table 3.1 (Con't)
Initial Oven Temperature 40C
Initial Oven Hold Time 1 min
Temperature Ramp Rate 20C/min
Final Oven Temperature 250C
FID Hydrogen Flow Rate 40 ml/min
FID Air Flow Rate 180 ml/min
Recorder Full Scale 0.1 volt
Chart Speed 2.5 cm/min
Experimental Apparatus for the Impinger Preconcentration Method
Air samples were collected using a 50 ml impinger containing 25
ml of benzyl alcohol.
The flow rate was measured using a Precision Scientific Wet Test
Meter.
The vacuum pump was a 115 volt, 60 hertz, 1.25 amp pump
manufactured by Thomas Industries.
Preparation of Methanol Standards for the
Impinger Preconcentration Method
Anhydrous methyl alcohol, 95% analytical grade, received from
Mallinckrodt was used to prepare standard solutions. Analytical grade
benzyl alcohol received from Baker was used as the solvent. Methanol
25


standard A 0.25 M was prepared by diluting 100 ul of methanol to 10 ml
with benzyl alcohol. Methanol standard B 0.025 M was prepared by
diluting 100 ul of standard A to 10 ml with benzyl alcohol. Methanol
standard C 2.5 mM was prepared by diluting 100 ul of standard B to 10
ml with benzyl alcohol.
Solid Adsorbent Preconcentration Method
The study was accomplished using a Tekmar 5000 Thermal
Desorber interfaced to a Hewlett-Packard 5790 A Series Gas
Chromatograph equipped with a flame ionization detector. The
analytical column was a J & W Scientific DB-1 (60 m x 0.32 mm x 1 um
film thickness). The carrier gas was helium. The detector output was
amplified using a signal conditioning buffer amplifier with a gain of
approximately 300. An IBM analog to digital board interfaced to a
PC-XT compatible computer was used to cap hare data. Labtech
Notebook and Labtech Chrom software from Laboratory Technologies
Corporation were used for data analysis. Table 3.2 lists the thermal
desorption operating parameters. Table 3.3 lists the gas
chromatography operating parameters.
26


Table 3.2
Table 3.3
The thermal desorption operating parameters for the
solid adsorbent preconcentration method.
Desorb Flow Rate 11 ml/min
Line Heater 200C
Valve Heater 240C
Injector Heater 200C
Furnace READY 40C
START input user
Cryo-1 COOL 1 -125C
Furnace DESORB 190C
DESORB time 10.00 min
CONTINUE input user
Cryo-2 COOL -125C
Cryo-1 TRANS. 200C
TRANS, time 2.00 min
Cryo-2 INJECT 200C
INJECT time 1.50 min
Furnace BAKE 225C
BAKE time 10.00 min
The gas chromatography operating parameters for the
solid adsorbent preconcentration method.
Column Flow Rate 3.3 ml/min
Injector Temperature 250C
Detector Temperature 27C
Initial Oven Temperature -10 C
Initial Oven Hold Time 15 min
Temperature Ramp Rite 20C/min
Final Oven Temperature 150C
FID Make-Up Flow Rate 16 ml/min
FID Hydrogen Flow Rate 20 ml/min
FID Air Flow Rate 240 ml/min
27


Experimental Sampling Apparatus for the
Solid Adsorbent Preconcentration Method
Thermal desorption tubes were 1/4" OD x 7" stainless steel
packed with 300 mg 60/80 Carbopack B and 300 mg Carbosieve S-HI,
purchased from Supelco. Desorption tubes were conditioned at 225C
overnight with nitrogen flowing at 100 ml/min. A Pyrolysis Furnace
manufactured by Dohrmann Envirotech was used to heat the
desorption tubes.
Standards were trapped by injecting 1.0 ul volumes of known
concentrations into a helium or nitrogen stream of gas flowing at 100
ml/min. Samples were trapped using a vacuum pump (115 volt, 60
hertz, 1.23 amp pump) made by Thomas Industries Inc., with an air
flow rate of 100 ml/min. Air flow rates were measured with a Singer
Dry Test Meter from Singer American Meter Division.
Preparation of Methanol and Ethanol Standards
for the Solid Adsorbent Preconcentration Method
Anhydrous methyl alcohol, 95% analytical grade, received from
Mallinckrodt was used to make stock solutions. Absolute ethyl alcohol
received from AAPER Alcohol and Chemical Co. was used to make
stock solutions. All dilutions were made with deionized water that
28


had been filter though a Millipore system, doubly distilled, and
pyrolyzed. A 0.25 M methanol and 0.34 M ethanol stock solution was
prepared by diluting 1.00 ml of methanol and 2.00 ml of ethanol to 100
ml. Standard solutions were prepared by further dilution of stock
solutions.
29


CHAPTER IV
RESULTS AND DISCUSSION
Impinger Preconcentration Method
The first objective was to find the retention time of methanol in
a gas chromatogram of, methanol in benzyl alcohol using the impinger
preconcentration method. To find the retention time of methanol, a
blank and two standards of different concentrations were analyzed.
The first solution contained 0.25 M methanol in benzyl alcohol (see
Figure 4.1). The second solution contained 0.12 M methanol in benzyl
alcohol (see Figure 4.2). The third solution was a blank containing only
benzyl alcohol (see Figure 4.3). The first peak corresponded to
methanol, because the height of this peak changed when the
concentration of methanol in the standard solutions was changed. For
a blank solution the first very small peak at 8 chart divisions had a
peak height of 0.3 cm. For a 0.25 M standard solution the first peak at 8
chart divisions had a peak height of 6.5 cm. For a 0.12 M standard
solution the first peak at 8 chart divisions had a peak height of 14 cm.
30


Figure 4.1 A gas chromatogram of a standard solution containing
0.25 M methanol in benzyl alcohol, 4/23/1993. The
split ratio was 61. (The program on the gas
chromatograph was started late.)
31


Figure 4.2 A gas chromatogram of a standard solution containing
0.12 M methanol in benzyl alcohol, 4/23/1993. The
split ratio was 61.
32


Figure 4.3 A gas chromatogram of a blank solution containing
only benzyl alcohol, 4/23/1993. The split ratio was 61.
33


All three chromatograms had a reagent blank peak at 9 chart divisions
with a peak height of 1.2 cm. It was determined that the first peak in a
gas chromatogram of methanol in benzyl alcohol corresponded to
methanol. The retention time for methanol was at 8 chart divisions,
which corresponded to 1.3 cm and 46 seconds. The blank solution had
a peak at the retention time for methanol, which indicates that the
solvent benzyl alcohol contained a contaminate that co-eluted with
methanol.
The next task was to find the detection limit of chromatography
portion of the impinger preconcentration method. To find the
detection limit of the impinger preconcentration method, standards of
decreasing concentrations were analyzed until the gas chromatograms
of the standard and blank solutions were indistinguishable. A standard
solution containing 2.5 mM methanol in benzyl alcohol produced a
peak height of 4.2 cm for methanol (see Figure 4.4). A standard
solution containing 1.2 mM methanol in benzyl alcohol produced a
peak height of 1.8 cm for methanol (see Figure 4.5). A standard
solution containing 0.25 mM methanol in benzyl alcohol produced a
0.9 cm peak height for methanol (see Figure 4.6). A blank solution
containing only the benzyl alcohol solvent produced a peak height of
1.0 cm where methanol elutes (see Figure 4.7). Figures 4.6 and 4.7
indicate that the gas chromatogram of the blank solution was
indistinguishable from the gas chromatogram of a standard solution
containing 0.25 mM methanol in benzyl alcohol.
34


Figure 4.4 A gas chromatogram of a standard solution containing
2.5 mM methanol in benzyl alcohol, 4/11/1993. The
split ratio was 10.
35


Figure 4.5 A gas chromatogram of a standard solution containing
1.2 mM methanol in benzyl alcohol, 4/11/1993. The
split ratio was 10.
36


Figure 4.6 A gas chromatogram of a standard solution containing
0.25 mM methanol in benzyl alcohol, 4/11/1993. The
split ratio was 10.
37


Figure 4.7
A gas chromatogram of a blank solution containing
only benzyl alcohol, 4/11/1993. The split ratio was 10.
38


Once the detection limit of the impinger preconcentration
method was determined, the estimated minimum volume of air
needed to quantify methanol was calculated. The estimated detection
limit of the impinger preconcentration method was 2.5 x 10"5 mol of
methanol (see Appendix A). The concentration of methanol in
Denver's air was predicted to be in the ppbv concentration range. The
impinger trap contained 25 ml of benzyl alcohol. From this, the
estimated minimum volume of ambient air needed to quantify
methanol was calculated to be 740,000 liters (see Appendix A).
The impinger preconcentration method does not have the
sensitivity required to quantify the concentration of methanol in
ambient air in real time, for example collect an ambient air sample
every four hours. The maximum sampling rate was estimated to be
250 liters per hour, because collection efficiencies might be poor at
larger sampling rates. If we sampled at 250 liters per hour, it would
take 123 days to collect 740,000 liters of ambient air.
39


The Solid Adsorbent Preconcentration Method
Identification of the Methanol and Ethanol Peaks
The first objective was to find the retention times of methanol
and ethanol in a gas chromatogram of methanol and ethanol in water.
To find the retention time of methanol and ethanol, a blank solution, a
methanol standard solution, and a ethanol standard solution were
analyzed using the solid adsorbent preconcentration method. The
standards and the blank were injected directly into an empty thermal
desorption tube packed with a small amount of glass wool. The gas
chromatogram of the blank containing only Millipore Quality Water
(MQW) had three medium peaks at 3.8, 6.4, and 7.2 min and one small
peak at 4.2 min (see Figure 4.8). A gas chromatogram of a standard
solution containing 0.97 mM methanol in MQW had the same three
medium peaks as the blank, but the small peak at 4.2 min had become
larger (see Figure 4.9). The retention time of methanol was at 4.2 min
as indicated by the change in the size of this peak when methanol was
added to the analyzed solution. A gas chromatogram of a standard
solution containing 0.67 mM ethanol in MQW had the same four
peaks as the blank, but there was a new peak at 6.2 min (see Figure
4.10).
40


DETECTOR RESPONSE, Volts
Figure 4.8 A gas chromatogram of a blank solution containing
only Millipore Quality Water, 9/27/1994. The oven
temperature was 0C.
41


DETECTOR RESPONSE, Volts
Figure 4.9
A gas chromatogram of a standard solution containing
0.97 mM methanol in Millipore Quality Water,
9/27/1994. The oven temperature was 0C.
42


Figure 4.10 A gas chromatogram of a standard solution containing
0.67 mM ethanol in Millipore Quality Water,
9/27/1994. The oven temperature was 0C.
43


The retention time of the peak that corresponded to ethanol was at 6.2
min as indicated by the addition of this peak when ethanol was added
to the analyzed solution.
The above data were obtained with a GC oven temperature of
0C, and the data through out the rest of this research were obtained
with a GC oven temperature of -10C. It was determined that
decreasing the oven temperature increased the retention time and
overall resolution of the analytes. However, oven temperatures below
-10C had adverse effects on the chromatography. At -20C the
methanol and ethanol had extreme peak tailing and/or the sample
seemed to plug the column. The result of the sample plugging the
column was that the sample did not elute until the oven temperature
increased during the GC program. At -10C some ambient air samples
had delayed retention times, but all the standards and most ambient air
samples showed no adverse effects.
All standards and samples analyzed throughout the rest of this
research were adsorbed onto thermal desorption tubes as described in
the Background and Theory Section. Standard and blank solutions
were injected into a stream of flowing gas that passed through the
thermal desorption tube. The flow rate was approximately 100
ml/min. A vacuum flow rate of approximately 11/min was used to
load ambient air sample onto the thermal desorption tubes.
The next problem was to determine the retention times of
methanol and ethanol at an oven temperature of -10C. To find the
44


retention times methanol and ethanol, a standard solution containing
2.4 mM methanol and 1.7 mM ethanol in organic free water was
analyzed (see Figure 4.12). The gas chromatogram of a blank
containing only organic free water is displayed in Figure 4.11.
Comparing the gas chromatograms of the standard and blank solution,
it was apparent that the retention times of methanol and ethanol were
at 4.9 min and 8.2 min, respectively (See Table 4.1).
Reproducibility of the Solid Adsorbent Preconcentration Method
The daily reproducibility of the solid adsorbent preconcentration
method was studied by analyzing the same standard three times
consecutively. The gas chromatograms of the blank and the three
standards containing 2.4 mM methanol and 1.7 mM ethanol in organic
free water are displayed in Figures 4.11 through 4.14. The results from
the daily reproducibility study are given in Table 4.1.
45


Figure 4.11 A gas chromatogram of a blank containing only
organic free water, 11/1/1994. The oven temperature
was -10C.
46


Figure 4.12 A gas chromatogram of a standard solution containing
2.4 mM methanol and 1.7 mM ethanol in organic free
water, 11/1/1994.. The oven temperature was -10C.
Run number one.
47


Figure 4.13 A gas chromatogram of a standard solution containing
2.4 mM methanol and 1.7 mM ethanol in organic free
water, 11/1/1994. The oven temperature was -10C.
Rim number two.
48


Figure 4.14 A gas chromatogram of a standard solution containing
2.4 mM methanol and 1.7 mM ethanol in organic free
water, 11/1/1994. The oven temperature was -10C.
Run number three and the primary recovery sample.
49


Table 4.1
The result of the reproducibility study on 11/1/1994.
Run # [MeOH] mol Area Ret (min)
1 2.4 x IO-9 66.4 4.9
2 2.4x10-9 65.9 4.9
3 2.4x10-9 74.4 4.9
Blank 11.1 4.9
Run # [EtOH] mol Area Ret (min)
1 1.7x10-9 53.3 8.2
2 1.7x10-9 55.3 8.2
3 1.7x10-9 51.2 8.2
Blank No Peak 8.0
The average areas for the three analyses of methanol and
ethanol were 68.9 4.8 and 53.3 2.1 area units, respectively. It was
calculated that the relative standard deviations for methanol and
ethanol were 7% and 4%, respectively.
The reproducibility of the solid adsorbent preconcentration
method from October 11 through December 3,1994, was studied by
comparing the area produced when standard solutions were analyzed
during this time period. Table 4.2 contains the concentration of
methanol and ethanol, the area, and the date for each standard.
50


Table 4.2
The concentration of methanol and ethanol and the
area of the standards analyzed from 10/11/1994
through 12/3/1994. The moles of methanol and
ethanol are the moles that were trapped on the
thermal desorption tube.
Date [MeOH] mol Area [EtOH] mol Area
10-11-94 2.4x10-9 79.7 1.7x10-9 27.4
10-28-94 2.4xl0-9 77.1 1.7x10-9 71.1
11-1-94 2.4x10-9 68.9 1.7x10-9 53.3
11-15-94 2.4x10-9 81.6 1.7x10-9 NA
11-19-94 2.4x10-9 80.8 1.7x10-9 74.1
11-30-94 2.4 x 10-9 65.3 1.7x10-9 15.5
12-3-94 2.4x10-9 50.9 1.7x10-9 15.8
NA Data is not available
Table 4.2 indicates the reproducibility of methanol from October
11 through November 19, 1994, was fairly consistent. The calculated
reproducibility was 77.6 5.2 area units with a relative standard
deviation of 6.7% from October 11 through November 19, 1994. From
November 30 through December 3,1994, the reproducibility decreased.
Table 4.2 indicates that the reproducibility of ethanol was worse
than the reproducibility for methanol. The calculated reproducibility
was 56 21 area units with a relative standard deviation of 38% from
October 11 through November 19,1994. The calculated reproducibility
was 66 11 with a relative standard deviation of 17% from October 28
through November 30, 1994. From November 30 through December 3,
1994, the reproducibility decreased.
51


The reproducibility of ethanol was worse then methanol due to
resolution and peak tailing. Both the methanol and ethanol peaks
tailed. Ethanol was never baseline resolved from a co-eluting peak and
the peak tailing was worse for ethanol, causing the reproducibility of
ethanol to be worse then methanol.
It was never determined what caused the response of the solid
adsorbent preconcentration method to vary so much. The variation
might be normal and as more analyses are performed an overall range
might develop. The variation might be caused by a leak somewhere in
the plumbing of the solid adsorbent preconcentration method,
especially in the capillary interface. The variation might be caused by
inconsistent cryofocusing during the transfer of the sample. The
variation might be caused by over use of the thermal desorption tubes.
Compounds in air ambient air samples might irreversibly bind to the
adsorbent in the thermal desorption tube, decreasing collection
efficiencies.
Retention Time of Methanol and Ethanol in Ambient Air Samples
It was necessary to determine if the retention times of methanol
and ethanol in ambient air samples were different than in standard
solutions. Ambient air samples were spiked with a methanol and
ethanol to identify the retention times of methanol and ethanol.
52


Three consecutive ambient air samples were obtained on three separate
thermal desorption tubes. The first desorption tube was not spiked
with methanol and ethanol (see Figure 4.15). The second thermal
desorption tube was spiked with 2.3 x 10'9 moles of methanol and
1.7 x 10"9 moles of ethanol (see Figure 4.16). The third desorption tube
was spiked with 4.9 x 10"9 moles of methanol and 3.4 x 10_9 moles of
ethanol (see Figure 4.17). The size of two peaks at the expected
retention times of methanol and ethanol increased when the sample
was spiked with methanol and ethanol. The three gas chromatograms
indicated that in ambient air samples the retention times for methanol
and ethanol were 5.0 min and 8.4 min, respectively.
Collection Efficiency Study
The next task was to study the collection efficiency of the
thermal desorption tubes. To find the collection efficiency, the
recovery of methanol and ethanol from standard solutions was
investigated. Standard solutions were injected in to a flowing stream
of nitrogen that passed through two thermal desorption tubes in series.
The primary thermal desorption tube was used to capture the analytes
from the standard sample. The secondary thermal desorption tube was
used to capture any analytes that were not adsorb by the primary
thermal desorption tube.
53


Figure 4.15
A gas chromatogram of an ambient air sample,
11/5/1994. Six liters was collected from 11:08-11:14 am.
54


Figure 4.16 A gas chromatogram of an ambient air sample spiked
with 2.3 x 10'9 moles of methanol and 1.7 x 10"9 moles
of ethanol, 11/5/1994. Six liters was collected from
11:15-11:21 am.
55


DETECTOR RESPONSE. Volts
Figure 4.17 A gas chromatogram of an ambient air sample spiked
with 4.9 x 10'9 moles of methanol and 3.4 x 10~9 moles
of ethanol, 11/5/1994. Six liters was collected from
11:22-11:28 am.
56


See Figure 4.14 for the gas chromatogram of the primary recovery
sample. See Figure 4.18 for the gas chromatogram of the secondary
recovery sample. Table 4.3 lists the result of the collection efficiency
study.
Table 4.3 The result, of the collection efficiency study completed
on 11/1/1994. See Figures 4.14.,4.18, and 4.19.
Sample ID [MeOH] mol Area Ret (min)
Primary 2.4 x 10-9 74.4 4.9
Secondary 9.1 4.9
Blank 11.0 4.9
Conditioned Blank 7.8 5.0
Sample ID [EtOH] mol Area Ret (min)
Primary 1.7 x 10-9 51.2 8.2
Secondary 0
The data in Table 4.3 indicate that the recovery of methanol was
78% in the worst case, with no blank correction. However, there was
some evidence to support the theory that the gas chromatogram of the
secondary recovery sample was no different from a gas chromatogram
of a blank. The gas chromatogram of the secondary recovery sample
had a peak area of 9.1 area units for methanol, see Table 4.3. The gas
chromatogram of the blank had a peak area of 7.8 area units for
methanol, see Table 4.3. The blank can have a peak at the retention
time for methanol as large as the peak for methanol in the secondary
recovery sample.
57


Figure 4.18 A gas chromatogram of the secondary recovery sample
for run number 3,11/1/1994.
58


Figure 4.19 A gas chromatogram of a conditioned blank. The
thermal desorption tube was condition over night and
then analyzed, 11/1/1994.
59


Thus, it was possible that the recovery of methanol was near 100%,
because the gas chromatogram of the secondary recovery sample and
the blank were similar.
The data in Table 4.3 indicate that the recovery of ethanol was
100%. The gas chromatogram of the secondary recovery sample did not
have a peak at the retention time for ethanol.
Calibration Curve for Methanol and Ethanol
The next goal was to create a calibration curve for the
quantification of ethanol and methanol in ambient air. Three
calibration curves were obtain by analyzing standard solutions
containing methanol and ethanol on November 19, November 30, and
on December 3,1994 (see Tables 4.4 through 4.6 for the results). Figure
4.25 displays the calibration curves for both methanol and ethanol.
Table 4.7 lists the least squares linear regressions and the correlation
coefficients for the calibration curves.
60


Table 4.4
Table 4.5
The concentration of methanol and ethanol for each
standard with its corresponding area and retention
time. The analysis was performed on 11/19/1994,
(See Figures 4.20 through 4.24).
STD [MeOH] mol Area Ret (min)
1 2.4xl0-9 80.8 4.9
2 . l.lxlO-9 49.6 4.9
3 7.2 x 10-10 35.1 4.9
4 2.4 x 10"10 16.6 5.0
Blank 7.4 5.0
STD [EtOH] mol Area Ret (min)
1 3.3 x 10-9 125.7 8.2
2 1.7x10-9 74.1 8.2
3 1.0 x 10-9 42.4 8.3
4 3.3x 10'10 12.2 8.4
The concentration of methanol and ethanol for each
standard with its corresponding area and retention
time. The analysis was performed on 11/30/1994.
STD [MeOH] mol Area Ret (min)
1 . 2.4x10-9 65.3 4.9
2 1.1 x 10-9 52.9 4.8
3 7.2 x 10"10 31.1 4.9
4 2.4 x 10-10 7.7 5.0
Blank 5.7 5.0
STD [EtOH] mol Area Ret (min)
1 3.3 x 10-9 50.6 8.3
2 1.7x10-9 15.5 8.3
3 1.0 x 10-9 10.3 8.5
4 3.3 x 10-10 2.1 8.6
61


Figure 4.20 A gas chromatogram of standard one containing 2.4
mM methanol and 3.3 mM ethanol, 11/19/1994.
62


Figure 4.21 A gas chromatogram of standard two containing 1.1
mM methanol and 1.7 mM ethanol, 11/19/1994.
63


Figure 4.22 A gas chromatogram of standard three containing 0.72
mM methanol and 1.0 mM ethanol, 11/19/1994.
64


Figure 4.23 A gas chromatogram of standard four containing 0.24
mM methanol and 0.33 mM ethanol, 11/19/1994.
65


Figure 4.24
A gas chromatogram of a blank containing only
organic free water, 11/19/1994.
66


Table 4.6
The concentration of methanol and ethanol for each
standard with its corresponding area and retention
time. The analysis was performed on 12/3/1994.
STD [MeOH] mol Area Ret (min)
1 2.4x10-9 50.9 4.9
2 1.1 x 10-9 40.8 4.8
3 7.2 x 10"10 26.9 4.9
4 2.4 x 10"10 16.3 4.9
STD [EtOH] mol Area Ret (min)
1 3.3 x 10-9 36.7 8.2
2 1.7x10-9 15.8 8.3
3 1.0x10*9 12.3 8.4
The calibration curves in Figure 4.25 indicate that the sensitivity
for methanol decreased with time as indicated by the decreasing slopes.
However, the sensitivity for standards 2, 3, and 4 for methanol were
fairly consistent. The sensitivity from standard 1 for methanol was not
very consistent.
The calibration curves in Figure 4.25 indicate that the sensitivity
for ethanol decreased with time as indicated by the decreasing slopes.
The sensitivity for ethanol on November 19, was much greater than
the sensitivity for ethanol on November 30, 1994, or on December
3,1994. The sensitivity for ethanol on November 30, 1994, was similar
to the sensitivity for ethanol on December 3,1994.
67


Area Area
Figure 4.25 Calibration curves for the concentration of methanol
and ethanol versus area. The analyses were performed
on 11/19/1994,11/30/1994, and 12/3/1994.
68


Table 4.7
The least squares linear regressions and correlation
coefficients for the calibration curves of the
concentration of methanol and ethanol versus area.
Date Least Squares Linear Regressions for MeOH RA2
11/19/94 y = 12.227 + 2.925 x 10 0.993
11/30/94 y = 10.100 + 2.559 x 10 0.868
12/3/94 y = 15.677 + 1.585x10 0.924
Date Least Squares Linear Regressions for EtOH RA2
11/19/94 y = 4.047 + 3.762 x 10 0.989
11/30/94 y = -6.280 + 1.637x1010 0.965
12/3/94 y = -0.231 + 1.093x1010 0.975
The decrease in sensitivity of the solid adsorbent
preconcentration method could have been caused by several different
things. The instability in the flame gas conditions could cause the
decreased sensitivity. It was hard to control the flow rate of hydrogen,
due to a clogging flow restricter. The sensitivity of the system is
dependent on the flow rate of the hydrogen gas. The decreased flow
rate of hydrogen caused the sensitivity to decrease. The make-up gas
and air flow rates could also adversely effect the sensitivity. Overuse of
the thermal desorption tubes could cause the decreased sensitivity.
Compound from previous air samples could irreversibly bind to the
solid adsorbent, decreasing the active sites and the ability to trap
methanol and ethanol. Problems with the electrometer on the gas
chromatograph could have caused the decrease in sensitivity. It was
known that the electrometer was not producing the voltage it should
69


produce. We used an amplifier to increase the voltage of the
electrometer to capture the signal with the computer. However, the
electrometer could have been failing, producing smaller voltages with
time.
The Ambient Air Sample Studies
The reproducibility study above indicated that the solid
adsorbent preconcentration method's response was fairly consistent
from October 11 through November 19, 1994. Thus, estimates of the
concentration of methanol and ethanol in ambient air samples
obtained throughout this time were made using the calibration curve
obtained on November 19, 1994.
An ambient air sample was obtained at the Auraria Ambient Air
Sampling Station on October 14,1994. Nine liters of ambient air was
collected from 8:50 am to 9:00 am (See Figure 4.26). The retention times
in the gas chromatogram were shifted; Thus, the peak positions
relative to the other peaks and the peak shapes were used to determine
which peaks corresponded to methanol and ethanol. Peak 11 at 5.7
min corresponded to methanol in the gas chromatogram. Peak 18 at
9.4 min corresponded to ethanol in the gas chromatogram.
70


Figure 4.26 A gas chromatogram of an ambient air sample,
10/14/1994. Nine liters was collected from
8:50-9:00 am.
71


The estimated concentrations of methanol and ethanol were 3.6 ppbv
and 0.7 ppbv, respectively.
An ambient air sample was obtained at the Auraria Ambient Air
Sampling Station on October 27, 1994. Twelve liters of ambient air was
collected from 11:32 am to 11:44 am (See Figure 4.27). Peak 13 at 5.0 min
corresponded to methanol in the gas chromatogram. Peak 22 at 8.2
min corresponded to ethanol in the gas chromatogram. The estimated
concentrations of methanol and ethanol were 3.0 ppbv and 2.3 ppbv,
respectively.
An ambient air sample was obtained at the Auraria Ambient Air
Sampling Station on October 28,1994. Twelve liters of ambient air was
collected from 12:19 pm to 12:31 pm (See Figure 4.28). Peak 11 at 5.0
min corresponded to methanol in the gas chromatogram. Peak 18 at
8.2 min corresponded to ethanol in the gas chromatogram. The
estimated concentrations of methanol and ethanol were 2.3 ppbv and
3.4 ppbv, respectively.
Two ambient air samples were obtained at the Auraria Ambient
Air Sampling Station on November 2, 1994. For the first sample, six
liters of ambient air was collected from 12:07 pm to 12:14 pm (See
Figure 4.29). Peak 10 at 4.9 min corresponded to methanol. The
program did not integrate the ethanol peak, so an estimate of the peak
area was made by comparing the ethanol peak to another peak. Peak 21
was similar to the ethanol peak and its area was used to estimate the
concentration of ethanol.
72


DETECTOR RESPONSE. Volts
Figure 4.27 A gas chromatogram of an ambient air sample,
10/27/1994. Twelve liters was collected from
11:32-11:44 am.
73


Figure 4.28
A gas chromatogram of an ambient air sample,
10/28/1994. Twelve liters was collected from
12:19-12:31 pm.
74


Figure 4.29
A gas chromatogram of the first ambient air sample,
11/2/1994. Six liters was collected from 12:07-12:14 pm.
75


The estimated concentrations of methanol and ethanol in the first
sample were 1.7 ppbv and 3.4 ppbv, respectively. For the second
sample, six liters of ambient air was collected from 12:15 pm to 12:21
pm (See Figure 4.30). The retention times in the gas chromatogram
were shifted. Thus, the peak positions relative to the other peaks and
the peak shapes were used to determine which peaks corresponded to
methanol and ethanol. Peak 9 at 7.2 min was the peak that
corresponded to methanol in the gas chromatogram. There was not a
peak in the gas chromatogram that corresponded to ethanol. The
estimated concentration of methanol in the second sample was 2.9
ppbv.
An ambient air sample was obtained at the Auraria Ambient Air
Sampling Station on November 5,1994. Six liters of ambient air was
collected from 11:08 am to 11:14 am (See Figure 4.15). Peak 12 at 4.9 min
corresponded to methanol in the gas chromatogram. Peak 19 at 8.3
min corresponded to ethanol in the gas chromatogram. The estimated
concentrations of methanol and ethanol were 9.1 and 3.1 ppbv,
respectively.
The concentrations of methanol and ethanol for following
ambient air samples were calculated using calibration curves obtained
on the same day.
An ambient air sample was obtained at the Auraria Ambient Air
Sampling Station on November 19, 1994. Six liters of ambient air was
collected from 2:55 pm to 3:01 pm (See Figure 4.31).
76


Figure 4.30 A gas chromatogram of the second ambient air sample,
11/2/1994. Six liters was collected from 12:15-12:21 pm.
77


Figure 4.31 A gas chromatogram of an ambient air sample,
11/19/1994. Six liters was collected from 2:55-3:01 pm.
78


The retention times in the gas chromatogram were shifted. Thus, the
peak positions relative to the other peaks and the peak shapes were
used to determine which peaks corresponded to methanol and ethanol.
Peak 9 at 6.2 min corresponded to methanol in the gas chromatogram.
There was not a peak that corresponded to ethanol in the gas
chromatogram. The concentration of methanol was calculated to be 5.9
ppbv.
Two ambient air samples were obtained at the Auraria Ambient
Air Sampling Station on November 30, 1994. For the first sample, nine
liters of ambient air was collected from 1:41 pm to 1:50 pm (See Figure
4.32). The retention times in the gas chromatogram were shifted.
Thus, the peak positions relative to the other peaks and the peak
shapes were used to determine which peaks corresponded to methanol
and ethanol. Peak 8 at 8.8 min corresponded to methanol in the gas
chromatogram. Peak 8 was off scale. Thus, the area of peak 8 was
underestimated. Peak 14 at 11.5 min corresponded to ethanol in the gas
chromatogram. The concentrations of methanol and ethanol in the
first sample were calculated to be 12 ppbv and 1.0 ppbv, respectively.
For the second sample, three liters of ambient air was collected from
4:12 pm to 4:15 pm. There is not a chromatogram available for this
sample. There was not a peak in the gas chromatogram that
corresponded to ethanol. The concentration of methanol in the second
sample was calculated to be 9.7 ppbv.
79


Figure 4.32 A gas chromatogram of the first ambient air sample,
11/30/1994. Nine liters was collected from
1:41-1:50 pm.
80


An ambient air sample was obtained at the Auraria Ambient Air
Sampling Station on December 3,1994. Six liters was collected from
2:15 pm to 2:21 pm (See Figure 4.33). Peak 11 at 4.9 min corresponded to
methanol in the gas chromatogram. There was not a peak in the gas
chromatogram that corresponded to ethanol. The concentration of
methanol was calculated to be 9.2 ppbv.
One of the problems with some of the ambient air sample data is
the delayed retention times of the peaks in the chromatograms. The
delayed retention times were probably caused by the large amount of
water in the ambient air samples. The concentration of water was
estimated to be 0.73% at 0C and 5.0% at 30C, assuming a relative
humidity of 100% (see Appendix B). It was thought that this amount of
water might be capable of freezing in the column and plugging it
partially or completely.
A summary of the concentration of methanol and ethanol in
ppbv for the ten ambient air samples listed above is given in Table 4.8.
Sample volumes range from three to twelve liters with a sampling rate
of 1 liter per hour.
81


Figure 4.33
A gas chromatogram of an ambient air sample,
12/3/1994. Six liters was collected from 2:15-2:21 pm.
82


Table 4.8 Summary of the concentration of methanol and
ethanol in ten ambient air samples collected at the
Auraria Ambient Air Sampling Station.
Air Sample Date [MeOH] ppbv [EtOH] ppbv
10-14-94 @ 8:50 am 3.6 0.7
10-27-94 @ 11:32 am 3.0 2.3
10-28-94 @ 12:19 pm 2.4 3.4
11-2-94 @ 12:07 pm 1.7 3.4
11-2-94 @ 12:15 pm 2.9 *
11-5-94 @11:08 am 9.1 3.1
11-19-94 @ 2:55 pm 5.9 *
11-30-94 @ 1:41 pm 12 1.0
11-30-94 @ 4.21 pm 9.7 *
12-3-94 @ 12:15 pm 9.2 *
Average 5.9 3.7 2.3 1.2
* The results were below the minimum quantifiable
limit for ethanol.
The results in Table 4.8 indicate that methanol is present in the
ambient air in Denver at ppbv concentrations. No reliable historical
data exists on the concentration of methanol in Denver's ambient air,
for comparison with our data. Kelly, Callahan, Plell, and Evans found
from 22 samples that the concentration of methanol in Boston was 7.2
to 47 ppbv with and average of 17 ppbv, and in Houston was 5.6 to 31
ppbv with an average of 17 ppbv (17). In Los Angeles Hanst, Wong,
and Bragin found, from one infrared study, that the concentration of
methanol was 10 ppbv (11). In Denver we found, from 10 ambient air
samples, that the concentration of methanol was 1.7 to 12 ppbv with
an average of 5.9 3.7 ppbv. The results of our research indicate that
83


the ambient air in Denver contains less methanol than the ambient air
in Boston, Houston, or Los Angeles.
The least concentrated standard was used as the minimum
quantifiable limit for methanol. Thus, the minimum quantifiable
limit for methanol was 0.24 ppbv. The results of this experiment were
more than three times the minimum quantifiable limit for methanol.
The results in Table 4.8 indicate that ethanol is also present in
the ambient air in Denver at ppbv concentrations. No reliable
historical data exists on the concentration of ethanol in Denver's
ambient air, for comparison with our data. Kelly, Callahan, Plell, and
Evans found, from 15 samples, that the concentration of ethanol in
Boston was <1.0 to 38 ppbv with an average of 4.9 ppbv, and in
Houston was <1.0 to 22 ppbv with and average of 2.9 ppbv (17). In
Denver we found, from 6 ambient air samples, that the concentration
of ethanol was 0.7 to 3.4 ppbv with an average of 2.3 1.2 ppbv. The
results of our research were in fair agreement with the results from the
other two studies.
The least concentrated standard was used as the minimum
quantifiable limit for ethanol. From October 14 through November 19,
1994, the minimum quantifiable limit for ethanol was 0.33 ppbv. Due
to the decreasing response to standard solutions from November 30
through December 3, 1994, the minimum quantifiable limit for ethanol
was 1.0 ppbv. From October 14 through November 19,1994, the results
of this experiment were more than three times the minimum
84


quantifiable limit for ethanol for all but one ambient air sample
obtained of October 14,1994. The sample obtained on October 14,1994,
was more than two times the minimum quantifiable limit for ethanol.
Due to the decreasing response of the solid adsorbent preconcentration
method, from November 30 through December 3, 1994, the results of
this experiment were near or below the minimum quantifiable limit.
The decreasing response of the solid adsorbent preconcentration
method effected the analysis of ethanol more than methanol.
85


Full Text

PAGE 1

,. . -:-: .. DETERMINATION OF METHANOL AND ETHANOL IN AMBIENT AIR BY GAS CHROMATOGRAPHY WITH FLAME IONIZATION DETECTION by Debra Lee Kirschenman B.A., University of. Colorado at Denver, 1991 A thesis submitted to the Faculty of the Graduate School of the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Master of Science Chemistry 1995

PAGE 2

This thesis for the Master of Science degree by Debra Lee Kirschenman has been approved for the Graduate School by f-F;) -75 Date

PAGE 3

Kirschenman, Debra Lee (M.S., Chemistry) Determination of Methanol and Ethanol in the Ambient Air by Gas Chromatography with Flame Ionization Detection Thesis directed by Associate Professors Larry G. Anderson and John A. Lanning ABSTRACT The increased use of methanol and ethanol gasoline blended fuels is expected to increase the concentrations of these alcohols in ambient air. Methanol and ethanol are expected to be at low ppbv concentrations in the presence of a large number of hydrocarbon and carbonyl spedes. Currently there is no specific and sensitive technique available for the quantification of methanol and ethanol in ambient air. Our objectives were to develop a method capable of quantifying methanol and ethanol at the ppbv levels by using a standard gas chromatograph with a flame ionization detector (GC-FID), and to determine these alcohols in the ambient air of Denver, Colorado. iii

PAGE 4

A standard GC-FID is not sensitive enough to quantify methanol and ethanol at the ppbv level without a preconcentration procedure. First we bubbled ambient air through an impinger containing a liquid solvent capable of absorbing methanol. Experiments indicated that the impinger preconcentration method was not sensitive enough to quantify methanol in ambient air, so we tried another method of preconcentrating the sample. We attempted to trap methanol and ethanol in ambient air samples onto thermal desorption tubes. The analytes were thermally desorbed and cryofocused onto the top of a GC capillary column. Rapid heating of the top of the GC column injects the analytes. The detection limits of the system were in the ppbv range for both methanol and ethanol. Ten ambient air samples of three to twelve liters were collected at the Auraria Ambient Air Sampling Station on the University of Colorado Campusin Denver, Colorado, from November 11 through December 3, 1994. The concentration of methanol in Denver's air ranged from 1.7 to 12 ppbv with an average of 5.9 3.7 ppbv, and ethanol ranged from 0.70 to 3.4 ppbv with an average of 2.3 1.2 ppbv. iv

PAGE 5

The form and content of this abstract are approved. We recommend its publication. Signed v

PAGE 6

DEDICATION To my parents Elton and Helen for your .generous love and support. You're the best in the world.

PAGE 7

CONTENTS CHAPTER L INlRODUCfiON .................................................................... 1 Using Alternative Fuels to Reduce the United States' Dependence on Foreign Oil Imports ............ ; .......... 2 Using Oxygenated Fuels to Combat the Increasing Air Pollution in the United States' Major Metropolitan Areas ................................................................. 3 Using Oxygenated Fuels to Reduce Ozone Concentrations ............................................................. 4 Using Oxygenated Fuels to Reduce Carbon Monoxide Concentrations ......................................... 6 Measurement Technigues ................................................. 7 Research Objectives for the Impinger Preconcentration Method ...................................................... 9 Research Objectives for the Solid Adsorbent Preconcentration Method ...................................................... 11 II. BACKGROUND AND THEORY ........................................ 14 Impinger Preconcentration Method .................................... 14 Solid Adsorbent Preconcentration Method ....................... 16 The Sampling Process .................................................. 16 Thermal Desorption and Cryofocusing Process ............................................................................ 18 Gas Chromatography and Flame Ionization Detection ........................................................................ 23 vii

PAGE 8

lli. EXPERIMENTAL SECTION ................................................. 24 Impinger Preconcentration Method ................................... 24 Experimental Apparatus for the Impinger Preconcentration Method .......................................... 25 Preparation of Methanol Standards for the Impinger Preconcentration Method ........................ 25 Solid Adsorbent Preconcentration Method ....................... 26 Experimental Sampling Apparatus for the Solid Adsorbent Preconcentration Method ........... 28 Preparation of Methanol Standards for the Solid Adsorbent Preconcentration Method ........... 28 IV. RESULTS AND DISCUSSION ............................................ 30 Impinger Preconcentration Method .................................... 30 Solid Adsorbent Preconcentration Method ...................... .40 Identification of the Methanol and Ethanol Peaks ............................................................................... 40 Reproducibility of the Solid Adsorbent Preconcentration Method .......................................... 45 Retention Times of Methanol and Ethanol in Ambient Air .................................................................. 52 Collection Efficiency Study ........................................ 53 Calibration Curve for Methanol and Ethanol. ...... 60 Ambient Air Sample Studies .................................... 70 viii

PAGE 9

IV. CONCLUSIONS ........................................................................ 86 Recommendations for Future Studies ................................ 91 APPENDIX A. Estimated Volume of Air Needed to Quantify Methanol in Ambient Air Using the Impinger Preconcentration Method ...................................................... 94 B. Moles of Water in Ambient Air ........................................... 96 REFERENCES ............................................. ................................................... 97 ix

PAGE 10

FIGURES Figure 2.1 A block diagram of the apparatus used to load standards onto thermal desorption tubes ........................... 18 2.2 A block diagram of the apparatus used to load ambient air samples onto thermal desorption tubes ............................................................................................ 18 2.3 A block diagram of the thermal desorption and cryofocusing apparatus ........................................................... 19 4.1 A gas chromatogram of a standard solution containing 0.25 M methanol in penzyl alcohol.. ............... 31 4.2 A gas chromatogram of a standard solution containing 0.12 M methanol in benzyl .............. 32 4.3 A gas chromatogram of a blank solution containing only benzyl alcohol .................................................................. 33 4.4 A gas chromatogram of a standard solution containing 2.5 mM methanol in benzyl alcohol.. ............. 35 4.5 A gas chromatogram of a standard solution containing 1.2 mM methanol in benzyl alcohol.. ............. 36 4.6 A gas chromatogram of a standard solution containing 0.25 mM methanol in benzyl alcohol.. ........... 37 4.7 A gas chromatogram of a blank solution containing only benzyl alcohol. ................................................................. 38 4.8 A gas chromatogram of a blank solution containing only Milli pore Water ............................................................. 41 X

PAGE 11

4.9 A gas chromatogram of a standard solution containing 0.97 mM methanol in Millipore Water ........ .42 4.10 A gas chromatogra,m of a standard solution containing mM ethanol in Millipore Water ............ .43 4.11 A gas chromatogram of a nitrogen blank .......................... .46 4.12 A gas chromatogram of a standard solution containing 2.4 mM methanol and 1.7 mM ethanol in organic free water. Run number one ........................... .47 4.13 A gas chromatogram of a standard solution containing 2.4 mM methanol and 1.7 mM ethanol in organic free water. Run number two ......................... .48 4.14 A gas chromatogram of a standard solution containing 2.4 mM methanol and 1.7 mM ethanol in organic free water. Run number three and primary recovery sample ...................................................... 49 4.15 A gas chromatogram of an ambient air sample, 11/5/1994 .................................................................................... 54 4.16 A gas chromatogram of an ambient air sample spiked with 2.3 x 1 o-9 moles of methanol and 1.7 x 1o-9 moles of ethanol, 11/5/1994 ............................ 55 4.17 A gas chromatogram of an ambient air sample spiked with 4.9 x 1o-9 moles of methanol and 3.4 x 1o-9 moles ofethanol, 11/5/1994 ............................. 56 4.18 A gas chromatogram of the secondary recovery sample .......................... ............................................................. 58 4.19 A gas chromatogram of a conditioning blank. .................. 59 4.20 A gas chromatogram of standard one containing 2.4 mM methanol and 3.3 mM ethanol. ............................. 62 xi

PAGE 12

4.21 A gas chromatogram of standard two containing 1.1 mM methanol and 1.7 mM ethanol. ............................. 63 4.22 A gas chromatogram of standard three containing 0.72 mM and 1.0 mM ethanol.. .......................... 64 4.23 A gas chromatogram of standard four containing 0.24 mM methanol and 0.33 mM ethanol.. ........................ 65 4.24 A gas chromatogram of a blank containing only organic water ..................................................................... 66 4.25 Calibration curves for the concentration of methanol and ethanol versus area ...................................... 68 4.26 A gas chromatogram of an ambient air sample, 10/14/1994 .................................................................................. 71 4.27 A gas chromatogram of an ambient air sample, 10/27 /1994 .................................................................................. 73 4.28 A gas chromatogram of an ambient air sample, 10/28/1994 ................................. ................................................. 74 4.29 A gas chromatogram of the first ambient air sample, 11/2/1994 .................................................................... 75 4.30 A gas chromatogram of the second ambient air sample,_ 11 /2/1994 ..... .............................................................. 77 4.31 A gas chromatogram of an ambient air sample, 11/19 /1994 .................................................................................. 78 4.32 A gas chroinatograp1 of the first ambient air sample, 11 I 30 I 1994 .................................................................. 80 4.33 A gas chromatogram of an ambient air sample, 12/3/1994 .................................................................................... 82 xii

PAGE 13

TABLES Table 1.1 Reaction rates for selected hydrocarbons with the hydroxyl radical ........................................................................ 5 2.1 Values for the parameters a and bin Equation 2.1 for various compounds of interest including methanol .................................................................................. 22 2.2 Vapor Press.ure in mm Hg at varying temperatures for various compounds of interest including methanol ................................................................................... 23 3.1 The gas chromatography and recorder operating parameters for the impinger preconcentration method ....................................................................................... 24 3.2 The. thermal desorption operating parameters for the solid ads<:>rbent preconcentration method ................... 27 3.3 The gas chromatography operating parameters for the solid adsorbent preconcentration method .................. 27 4.1 The result of the reproducibility study ................................ SO 4.2 The concentration of methanol and ethanol and the area of the standard solutions analyzed from 10/11/94 through 12/3/94 ..................................................... .51 4.3 The result of the collection efficiency study ....................... 57 4.4 The concentration of methanol and ethanol for each standard with its corresponding area and retention time on 11/19/94 .................................................... 61 xiii.

PAGE 14

4.5 The concentration of methanol and ethanol for each standard with its corresponding area and retention time on 11 I 30 I 94 .................................................... 61 4.6 The concentration of methanol and ethanol for each standard with its corresponding area and retention time on 12/3 /94 ...................................................... 67 4.7 The least squares linear regressions and correlation coefficients for the calibration curves of methanol and ethanol yersus area ......................................................... 69 4.8 Summary of the concentration of methanol and ethanol in ten air samples collected at the Auraria Ambient Air Sampling Station ............................................ 83 xiv

PAGE 15

ACKNOWLEDGMENTS Special appreciation to Professor Larry G. Anderson for his support, guidance, and understanding throughout the duration of this research project. Thanks to Professor John A. Lanning for his advice and electronics help. Thanks to Professor Doris R. Kimbrough for her assistance and encouragement. Special thanks to Dr .. Donald Zapien for the use of his organic free Thanks to Dr." Robert R. Meglen for giving me confidence in myself and my abilities. Thanks to Dr. Larry L. Jackson for his advice and reassurance during the course of this research. I would like to express my appreciation to my parents, Elton and Helen Kirschenman .for their love, understanding, and support. Finally, I would like to thank Greg and Cris Kirschenman for their support and encouragement during my studies. XV

PAGE 16

CHAPTER I INTRODUCTION The United States' dependence on foreign oil imports and increasing air pollution problems in its major metropolitan areas are driving the development of motor vehicle technologies that use alternative fuels instead of gasoline. The increased use of alternative fuels might increase the concentrations of the components of these alternative fuels in ambient air. The quantification of the alternative fuels in ambient air is necessary in order to study their role in the increasing air pollution: problems. Two of the alternative fuels currently in use are methanol and ethanol. The increased use of methanol and ethanol as alcohol gasoline blended fuels, might increase the concentrations of these alcohols in ambient air. Several separate processes contribute to increase in concentration of methanol and ethanol in ambient air: evaporative losses from the fuel system of motor vehicles, emissions of unburned fuel by motor vehicles, and evaporative losses during the fueling process of motor vehicles. 1

PAGE 17

Currently there are no reliable methods for determining the concentrations of methanol and ethanol in ambient air. Previously developed methods do not have the sensitivity necessary for quantifying the concentration of methanol and ethanol in ambient air. In this research we developed a method to quantify the concentration of methanol and ethanol in ambient air of Denver by using a standard gas chromatograph equipped with a flame ionization detector (GC-FID). Using Alternative Fuels to Reduce the United States' Dependence on Foreign Oil Imports. This country's dependence on foreign oil imports drives the development of motor vehicle technologies that use alternative fuels instead of gasoline. The United States initiated the Alternative Motor Fuels Act of 1988 (AMFA) to decrease the country's dependence on foreign oil imports. The AMFA required the Department of Energy (DOE) develop new motor vehicle .technologies that use alternative fuels (1). The DOE, through the National Renewable Energy Laboratory (NREL), supports a number of projects that are studying the use of methanol and ethanol as alternative motor vehicle fuels (1). The General Service Administration (GSA) currently has alternative fuels fleets in seven different areas: Detroit, Atlanta, Denver, Chicago, Houston, New York, and Washington, DC (2). The Denver Fleet 2

PAGE 18

Management Center has 229 light duty vehicles, 131 operating on M-85, involved in an emission and data collection test program (2). The Denver Fleet Management Center also has 226 light duty vehicles, 162 operating on M-85, not involved in an emission and data collection test program (2). M-85 is a methanol gasoline blended fuel composed of 85% methanol and 15% gasoline. Currently, M-85 is available only at two Conoco service stations in the Denver, one is at the intersection of 6th Avenue and Sheridan and the other is at the intersection of Colfax and Josephine. Using Oxygenated Fuels to Combat the Increasing Air Pollution in the United States' Major Metropolitan Areas The air pollution problem in some of this country's major metropolitan areas has driven the development of oxygenated gasoline programs. Emissions from gasoline powered motor vehicles contribute to a large portion of the hydrocarbons (HC), nitrogen oxides (NOx), and carbon monoxide (CO) in ambient air (3). Oxygenated fuels are being used to reduce ozone concentrations in cities that violate the national air quality standard for ozone (03). Oxygenated fuels, such as alcohol gasoline blended fuels, are also being used in an attempt to reduce carbon monoxide concentration in cities that violate the national air quality standard for carbon monoxide. 3

PAGE 19

Using Oxygenated Fuels to Reduce Ozone Concentrations In areas where ozone concentrations exceed the National Ambient Air Quality Standards for Ozone, there has been an attempt to reduce ozone levels using oxygenated gasolines. Recently 101 U.S. cities failed to meet the National Ambient Air Quality Standards for ozone (4). The burning of gasoline by automobiles produces hydrocarbons, carbon monoxide, and nitrogen oxides. Hydrocarbons or carbon monoxide can react with nitrogen oxides in ambient air to produce ozone, a suspected player in the photochemical smog problem in southern California cities. The use of an oxygenated gasoline reduces the amount of hydrocarbons and increases the amount of oxygenate in motor vehicle exhaust. The idea is that the oxygenate will react slower than the hydrocarbons do with nitrogen oxide, there by reducing the level of ozone in the local ambient air. The combustion of alcohol gasoline blended fuels or gasoline by motor vehicles produces alcohol and gasoline emissions in the local ambient air. Both oxidize to carbon dioxide in the atmosphere. The first step in the oxidation of a hydrocarbon is the reaction with hydroxyl radical (OH). A typical alkane reacts about 10 times faster than methanol and about 3 times faster than ethanol with the hydroxyl radical. Alkenes and alkynes react even faster than alkanes. Table 1.1 4

PAGE 20

gives the reaction rates for selected hydrocarbons with the hydroxyl radicaL Table 1.1 Compound Methanol Ethanol Reaction rates for selected hydrocarbons with the hydroxyl radical. (From Finlayson-Pitts & Pitts, "Atmospheric Chemistry" (3)). Reaction Rate 0.90 x 10-12 (cm3 molecules-1 s-1) 2.90 x 10-12 (cm3 moletules-1 s-1) Carbon Monoxide 0.15 x 10-12 (cm3 molecules-1 s-1) n-Octane 8.72 x 1o-12 (cm3 molecules-1 s-1) Ethene (Slowest Alkene) 8.54 x 1o-12 (cm3 molecules-1 s-1) 1-Butyne 8.04 x 1o-12 (cm3molecules-1 s-1) One of the co:rnpounds produced during the oxidation of a hydrocarbon is the hydroperoxyl radical (H02), a major player in the photochemical ozone pollution problem. The reaction of the hydroperoxyl radical with nitrogen oxide produces ozone through the following set of chemical reactions. H02 +.NO N02 + HO N02 + h-.J NO + 0 0 + 02 + M 0 + 3 h-.J Alcohol fuel combustion emissions react slower with the hydroxyl radical than gasoline combustion emissions, and in theory are 5

PAGE 21

transported farther from their source. The transport effect might reduce local ozone concentrations near the source. Using Oxygenated Fuels to Reduce Carbon Monoxide Concentrations To combat the increasing air pollution problem there has been an attempt to decrease auto emissions of hydrocarbon and carbon monoxide from automobiles using oxygenated gasolines. Carbon monoxide produced by the incomplete combustion of gasoline is a major pollutant in many cities (3). -The idea is that increasing the concentration of oxygen in the fuel reduces the concentration of carbon monoxide in the motor vehicle eXhaust by burning the fuel completely to carbon dioxide. Alcohol gasoline mixtures can reduce the level of carbon monoxide in automobile exhausts by up to 40 to 50 percent (5). Oxygenated gasoline is being used in an attempt to reduce the levels of carbon dioxide in areas that violate the National Ambient Air Quality Standard for Carbon Monoxide. There have been 37 areas that have implemented oxygenated gasoline programs, as required by the Clean Air Act Amendments of 1990 (6). The Denver-Boulder area in Colorado is one of the Oxygenated Gasoline Program Areas in the United States (6). The Colorado Oxygenated Gasoline Program started on January 1, 1988, ran through March 1, 1988, and has been in effect from 6

PAGE 22

November 1st through March 1st every year since its inception. The most popular oxygenates used in Colorado are methyl tertiary butyl ether (MTBE) and ethanol.. In the winter of 1992 to 1993, MTBE accounted for 55% of the market share and ethanol accounted for 45% of the market share (6). In the winter of 1993 to 1994, MTBE accounted for 30% of the market share and ethanol accounted for 70% of the market share (7). Currently, Colorado Regulation No. 13 requires that in program areas there is a minimum oxygen content of 2.7% per gallon of gasoline (7). In practical terms this means 10% ethanol per gallon or 15% MTBE per gallon (7). Measurement Techniques The early methods developed for the analysis of methanol were modifications of methods developed for the determination of formaldehyde. The methods involved the oxidation of methanol to formaldehyde followed br spectrophotometric analysis. Bhatt and Gupta (8) oxidized methanol with acidic potassium permanganate and removed the excess oxidant before analysis by the colormetric reaction with oxalydihydrazide and copper II in acetate medium. Verma and Gupta (9) developed a sensitive method that determined methanol indirectly by bubbling air through midget impingers containing an oxidizing solution that oxidized methanol to formaldehyde, before 7

PAGE 23

analysis by spectrophotometry in acidic medium with p-aminoazobenzene and sulfur dioxide. Maeda, Fujio, Suetaka, and Makoto (10) developed a method for the determination of trace amount of alcohols based on the formation of alkyl nitrite when alcohols react with nitrogen dioxide (N02), before analysis by gas chromatography with electron-capture detection. The above methods have the disadvantage of being complicated and labor intensive. Another disadvantage is that other compounds interfere with the analysis. For example, glycerols and glycols can be oxidized to formaldehyde. Hanst, Wong, and Bragin (11) quantified the concentration of methanol in Los Angeles smog using a Fourier transforming long path infrared spectrometry wjth a 23m long base path and a 1260 m optical path. This technique uses expensive and delicate instrumentation not easily transported to different field. sites. Methods exist for the determination of methanol and ethanol in automotive vehicle exhaust (12 ,13, 14). They bubbled diluted vehicle exhausts through a series of midget impinger traps containing deionized water followed by GC-FID analysis. The methods relied on the fact that alcohols are highly soluble in water, and hydrocarbons are insoluble in water. The major disadvantage of this method is that it is not sensitive enough to measure the low part per billion by volume (ppbv) concentrations of methanol and ethanol expected in ambient air. Further complicating the matter is the decreased response of the 8

PAGE 24

FID when large amounts of water are present. Water as a collector has the added disadvantage that its vapor pressure is too large to use as an impinger solvent without an ice bath to minimize evaporation. There is also the possibility of interference from low molecular weight carbonyls and ketones, which are soluble in water. Haky and Stickney (15) developed a method for the determination of residual solvents, such as in bulk pharmaceuticals, by dissolving samples in benzyl alcohol followed by GC-FID analysis. The advantage of this method is the elimination of water collected from the sample, because water is insoluble in benzyl alcohol. The disadvantage of this method is that the method does not have the required sensitivity that would enable the quantification of methanol in ambient air. Research Objectives for the Impinger P.reconcentration Method The focus of this research was to develop a method to measure the concentration of methanol by using a modification of the bubbler method described above. We bubbled ambient air through a midget impinger containing benzyl alcohol followed by GC-FID analysis. The method should be simple, selective for methanol, sensitive enough to measure ppbv concentrations in ambient air, and reproducible. 9

PAGE 25

The use of the impinger preconcentration method coupled with GC-FID approach did not appear to have the sensitivity required to quantify methanol in ambient air. Preliminary studies and calculations indicated the volume of an ambient air sample was too large to quantify methanol, as discussed in further detail later. The following paragraphs describe the constraints observed during the development of impinger preconcentration method. A limit to this method is the sensitivity of the GC-FID to methanol in benzyl alcohol. For the method to be useful it should be able to quantify the concentration of methanol in ambient air in real time, for example collect an ambient air sample every four hours. Thus, the volume of air sampled and the volume of benzyl alcohol in the impinger trap should reflect this restraint. The flow rate of the air through the impinger trap should be slow enough to extract all the methanol from air into the benzyl alcohol. The method should be easily automated for routine air monitoring, with a minimal amount of labor. The above constraints restrict any method using impinger traps to quantify methanol in ambient air. The major problem of any method using this approach would be its lack of sensitivity. We determined that impinger preconcentration method does not have the sensitivity required to quantify ppbv concentrations of 10

PAGE 26

methanol in ambient air. The volume of an individual ambient air sample was much too large, for real time analysis. Research Objectives for the .Solid Adso:r.bent Preconcentration Method Instrumentation capable of purge and trap analysis is commercially available. Many of these instruments purge a sample with helium while the sample is being heated. A cryogenic trap concentrates analytes purging from the sample. Variations of this system have been used to quantify volatile polar organic compounds in ambient air. Fung (16) developed a method for the measurement of alcohols and MTBE in ambient air. They cryogenically concentrated samples before analysis by GC-FID. The method employed a two dimensional column system that reduced the analysis time to under 10 minutes. The major disadvantage of this method is the complexity of the two dimensional GC system. Kelly, Callahan, Plell, and Evans (17) developed a method for the determination of polar volatile organic compounds, including methanol and ethanol, in ambient air. They collected ambient air samples in evacuated canisters over a 24 hour period and transferred them to a two-stage solid adsorbent trap followed by thermal 11

PAGE 27

desorption, cryogenic concentration, and gas chromatography with ion trap detection. For us the problem with this method was the inaccessibility of a gas chromatograph equipped with an ion trap detector. The primary focus of our research was to develop a to measure the amount of methanol and ethanol in ambient air using the two-stage trap and purge method of Kelly, Callahan, Plell, and Evans (17), but utilizing a GC-FID. The method should be simple, selective for methanol and ethanol, sensitive enough to. measure ppbv concentration in ambient air, and reproducible.The following paragraphs describe the constraints observed during the development of the solid adsorbent preconcentration method. A limit to this method is the sensitivity of the gas GC-FID to methanol and ethanol in the presence of large amounts of water, because tiapping air samples on a solid adsorbent also traps the water in the air. Excessive water could impair the performance of the column and detector in the GC. For the method to be useful, it should be able to quantify the concentration of methanol and ethanol in ambient air with reasonable time resolution, for example collect an ambient air sample every four hours. Thus, the volume of air sampled and the time required to analyze a sample should reflect this restraint. 12

PAGE 28

The flow rate of the air through the trap should be slow enough to trap all the methanol and ethanol in ambient air and not exceed the efficiency of the trap. The method should be easily automated for routine air monitoring, with a minimal amount of labor. The restriction imposed on the development of the solid adsorbent preconcentration method would be mandatory to any method designed to quantify volatile polar organic compounds in ambient air. Our proposed solid adsorbent preconcentration method appeared well suited for the routine analysis of alcohols, methanol and ethanol, in ambient air. 13

PAGE 29

CHAPTER II BACKGROUND AND THEORY Im:pinger Preconcentration Method The initial objective of this research was to develop a technique capable of detecting methanol in ambient air at ppbv concentrations by using standard gas chromatography instrumentation. However, a GC-FID is not sensitive enough to detect methanol at the ppbv concentrations, thus a method of concentrating the methanol would be necessary. Bubbling ambient air through an impinger containing a liquid solvent that would absorb the methanol seemed a practical way to approach this problem. The impinger solvent used would have to meet the following Methanol should be soluble enough in the solvent to enable its use as a quantitative technique. The solvent should be selective for alcohols, that is not dissolve other air pollutants to an appreciable extent. The vapor pressure of the solvent needs to be high enough to keep the evaporation of the solvent to minimum. Keeping 14

PAGE 30

a solvent with a low vapor pressure cool to minimize evaporation might be impractical. The solvent should not interfere with the detection of methanol by the FID. The first choice of an impinger solvent was water, because methanol is highly soluble in water. Water would also be somewhat selective since hydrocarbons are relatively insoluble in it. One disadvantage of water is that its vapor pressure is too large to use as an impinger solvent without an ice bath to minimize evaporation. Another disadvantage is that water adversely affects the performance of the FID even though it is non-detectable, and could extinguish the flame. Benzyl alcohol is an impinger solvent that would meet the solvent requirements listed above. Methanol is highly soluble in benzyl alcohol. Water. is highly insoluble in berizyl alcohol, eliminating waters adverse effect on the FID. The vapor pressure of benzyl alcohol is small enough to keep evaporation to minimum without cooling. Benzyl alcohol should not Interfere with the detection of methanol, because it should elute after methanol. One disadvantage of benzyl alcohol is that it is not selective for methanol, because anything from slightly polar methanol to highly non-polar hydrocarbons will dissolve in it. Initial experiments indicated that the impinger method was not sensitive enough to quantify the concentration of methanol in ambient air. The volume of an individual ambient air sample was too large. 15

PAGE 31

Solid Adsorbent Preconcentration Method The focus of the solid adsorbent preconcentration method was to develop a procedure that did not require solvent impinger preconcentration, but still used standard gas chromatography instrumentation. Thus, we needed another method of concentrating methanol and ethanol in ambient air. The use of solid phase extraction of some sort seemed a practical approach to the preconcentration needs of this research. Trapping ambient air on a solid adsorbent followed by thermal desorption, ciyofocusing, and GC-FID analysis seemed to be a practical approach to this problem. The solid adsorbent preconcentration method divides into 3 separate processes: sampling, desorption and cryofocusing, and GC-FID analysis. The Sampling Process We used solid phase e?Ctraction for the sampling process. The idea of solid phase extraction is to pass liquid or gas samples through a solid that adsorbs analytes of interest. Samples trapped on solid desorption tubes are desorbed, extracted with and an organic solvent, followed by gas chromatography or high performance liquid chromatography. The solvent extraction method it is not sensitive enough to quantify the concentration of 16

PAGE 32

methanol and ethanol in .ambient air similar to the impinger preconcentration method. We trapped samples on thermal desorption tubes. A thermal desorption instrument desorbs the analytes and transfers them to the GC-FID. This method hasthe advantage of preconcentrating the sample with no extraction into a liquid solvent. In this research we used thermal desorption tubes to trap methanol and ethanol in ambient air samples. We used thermal desorption tubes packed with 300 mg of Carbopack B and 300 mg of Carbosieve 5-III. Carbopack B is a graphitized black adsorbent used for trapping airborne organic compounds, from C4-CS compounds to polychlorinated biphenyls. Carbosieve S-Ill is a spherical carbon molecular sieve capable of trapping C2 hydrocarbons. Carbopack Band Carbosieve 5-III have similar thermal desorption temperature limits (400C) and hydrophobic properties. We loaded thermal desorption tubes by preparing standards to known concentrations and injected them into a flowing stream of either helium or nitrogen gas. The helium or nitrogengas containing the evaporating methanol and ethanol passed through thermal desorption tubes .. Figure 2.1 illustrates a block diagram of the standard loading apparatus. We used a vacuum pump to draw ambient air samples through thermal desorption tubes. Figure 2.2 illustrates a block diagram of the ambient air sample loading apparatus. 17

PAGE 33

Figure 2.1 Figure 2.2 Air Sampling Station Manifold A block diagram of the apparatus used to load standards onto thermal desorption tubes. ,... I I .._ Helium Gas Thermal Tank .._ plnjector fDesorption Port Tube -II o.....-.:-Vent To The Air A block diagram of the apparatus used to load ambient air samples onto thermal desorption tubes. Thermal Air Vacuum Desorption Sample f---___... Pump ____.. Tube Volume Meter Thermal Desorption and Cryofocusing Process We used a thermal desorption instrument to desorb and cryofocus the analytes from the thermal desorption tubes. Figure 2.3 illustrates a block diagram of the thermal desorption and cryofocusing apparatus. 18

PAGE 34

Figure 2.3 A block diagram of the thermal desorption and cryofocusing apparatus. II I I I Cryotrap 1 I I Cryotrap 2 I II Thermal Desorption Gas Chromatograph Tube Furnace With A Flame Ionization Detector Tekmar Thermal Desorption Instrument Heating the thermal desorption tube while carrier gas passes through the tube, effectively desorbs the trapped analytes. The thermal desorption temperature was 100C. At 100C methanol and ethanol will desorb with high efficiency from the solid adsorbent, because methanol vaporizes at 64.7C and ethanol vaporizes at 78.5C (18). Compounds with boiling points above 100C would not desorb. The carrier gas and desorbed analytes from the thermal desorption tubes then passes through Cryotrap 1. Cryotrap 1 cryofocuses the analytes while venting purge gas and other compounds to.air. Rapid heating of Cryotrap 1 transfers the analytes to Cryotrap 2. Cryotrap 2, the capillary cryofocuses the analytes onto a cooled short section of column or pre-column. Rapid heating of the short section of column or pre-column injects the sample. Pre-colwnn 19

PAGE 35

cryofocusing requires attaching an uncoated fused silica tubing to the column. Cryofocusing occurs on the pre-column. Tekmar introduced an automated version usirig pre-column cryofocusing in 1983 (19 ,20). Pre-column cryofocusing systems do not require controlling the temperatures outside the GC oven. Pre-column focusing has two major disadvantages. The first disadvantage is that the pre-column to column connection must have zero dead volume. The second disadvantage is the lack of liquid phase, which can aid in trapping analytes, limits the capacity of the cold trap. Cryofocusing directly on column simplifies and improves the focusing process (21). Direct column cryofocusing eliminates the dead volume problem and leaks downstream because no unions are necessary. Direct column focusing also has the advantage of having a stationary phase present, which can improve the trapping efficiency by acting as an adsorbent (20). The trapping of analytes on column requires carefully controlling the temperature of all :parts of the column outside the GC oven. The Tekmar 2000 Series Capillary Interface, which we used in this research, has an advanced cryofocusing trap mounted in a special housing that allows the sample to be re focused with outstanding results (20). The GC column is passed through an unused injector port to the capillary interface and directly connected to the transfer line from the thermal desorption instrument. The cryotrap is mounted on top of the GC in a configuration that allows the cryotrap to sit directly on top of the injector port, which 20

PAGE 36

eliminates any possibility of cold spots outside the GC oven. The liquid nitrogen used to cool the trap is passed through a 1 I 4" tube from the bottom to the. top. The trap is heated differentially from the top to the bottom by a heating wire when power is applied. The differential heating, heats the inlet of the trap faster than the outlet. This creates a thermal gradient in the trap such that the inlet is always hotter than the outlet. The thermal gradient ensures that the sample is always moving from a hotter to a colder region that compresses compounds moving through the trap into a tight slug before they entering the GC oven. With the proper establishment of the gradient heat up rate, the heat up rate of the cryotrap becomes relatively unimportant. No differences In chromatographic efficiency were observed for heat up rates from 100C/ptin to 2000C/min. (20). The use of pre-column trapping on an uncoated fused silica greatly simplifies the design requirements, but the absence of any liquid phase limits the trapping mechanism to simple condensation. Highly concentrated or extremely volatile samples often exceed the capacity of the cold trap, and the resultant breakthrough produces split peaks and generally poor resolution (20). By trapping on column, breakthrough occurs only when the capacity of the column is overloaded. With on column trapping the capacity of the cryotrap is similar to the capacity of the column. The solid adsorbent preconcentration method employs two cryotraps used to concentrate the desorbed sample. The cryotraps were 21

PAGE 37

cooled to -125C, to minimize breakthrough of analytes during the two cryofocusing steps. Breakthrough on the cryotrap is a function of the cryotrap's temperature and the vapor pressure of the analytes. Equation 2.1 gives the vapor pressure of a compound at a specific temperature ,(18). log P = (-0.05223 a I T) + b 2.1 Table 2.1 contains values for the parameters a and bin Equation 2.1 for some typical compounds. Table 2.1 lists the calculated vapor pressure of various compounds using the values for a and b along with Equation 2.1. Table 2.1 Compound Acetaldehyde Methane Methanol Propane n-Propanol Values for the parameters a and bin Equation 2.1 for compounds of interest including methanol (18). Formula Temp. Range C a b C2H40 -24.3 to +27.5.liq 27,707 7.8206 CH4 -174 to ;..163liq. 8,516.9 6.8626 CH30H -10 to +80 liq. 38,324 8.8017 C3Hg -136 to liq. 19,037 7.217 C3H70H -45 to -10 liq .. 47,247 9.5180 22

PAGE 38

Table 2.2 Vapor Pressure in mm Hg at varying temperatures for compounds of interest including methanol. Compound -75C -100 oc -125 oc -150 oc Acetaldehyde 3.3 0.289 l.lxlo-2 1.2 X 10-4 Methane 4.1 X JQ4 2.0 X JQ4 7.2x 10 3 1.8 X 10 3 Methanol 5.0 X 10-2 1.7 X 10-3 1.9 X 10-5 3.5 X 10-8 Propane 160 30 3.2 0.14 n-Propanol 1.2x lo-3 1.8 x lQ-5 7.2x 10-8 3.0 X 10-11 The idea was to find a cryotrapping temperature that would efficiently trap methanol and ethanol, while minimizing the consumption of liquid nitrogen used to cool the traps. We used a cryotra:pping temperature of -125C for both cryotraps. At -125C the breakthrough of methanol and ethanol should be negligible. Gas Chromatography arid Flame Ionization Detection We used a GC-FID to quantify the analytes. We used DB-1 capillary column from J & W Scientific to separate the analytes. A DB-1 column contains dimethylsilicone, a non polar stationary phase, and is suitable for the analysis. of alcohols. A compound's vapor pressure and . solubility in the stationary phase influence its retention on a DB-1 column. Thus, polar volatile compounds. will elute faster than nonpolar non-volatile compounds from a DB-lcolumn. 23

PAGE 39

CHAPTER III EXPERIMENTAL SECTION Impinger Preconcentration Method The study was accomplished using a Hewlett-Packard 5790 A Series Gas Chromatograph equipped with a flame ionization detector. The analytical column was a Hewlett Packard HP-1 (12 m x 0.2 mm x 0.33 urn film thickness). The carrier gas was nitrogen. The detector .output was recorded by a Omni Scribe Series DSOOO Recorder. Table 3.1 lists the gas chromatography and recorder operating parameters. Table 3.1 The gas chromatography and recorder operating parameters for the impinger preconcentration method. Column Flow Rate 1.4 ml/min Septum Purge Flow Rate 4.5 ml/min Split Vent Flow Rate 10.3 ml/min Injection Volume. 10 ul Injector Temperature 250C Detector Temperature 275C 24.

PAGE 40

Table 3.1 (Con't) Initial Oven Temperature 40C Initial Oven Hold Time 1 min Temperature Ramp Rate 20C/min Final Oven Temperature 250C FID Hydrogen Flow Rate 40 ml/min FID Air Flow Rate 180 ml/min Recorder Full Scale 0.1 volt Chart Speed 2.5 em/min Experimental Apparatus for the Impinger Preconcentration Method Air samples were collected using a 50 ml impinger containing 25 ml of benzyl alcohol. The flow rate was measured using a Precision Scientific Wet Test Meter. The vacuum pump was a 115 volt, 60 hertz, 1.25 amp pump manufactured by Thomas Industries. Preparation of Methanol Standards for the Impinger Preconcentration Method Anhydrous methyl alcohol, 95% analytical grade, received from Mallinckrodt was used to prepare standard solutions. Analytical grade benzyl alcohol received from Baker was used as the solvent. Methanol 25

PAGE 41

standard A 0.25 M was prepared by diluting 100 ul of methanol to 10 ml with benzyl alcohol. Methanol standard B .025 M was prepared by diluting 100 ul of standard A to 10 ml with benzyl alcohol. Methanol standard C 2.5 mM was prepared by diluting 100 ul of standard B to 10 ml with benzyl alcohol. Solid Adsorbent Preconcentration Method The study was accomplished using a Tekmar 5000 Thermal Desorber interfaced to a Hewlett-Packard 5790 A Series Gas Chromatograph equipped with a flame ionization detector. The analytical column was a J & W Scientific DB-1 (60 m x 0.32 mm x 1 urn film thickness). The carrier gas was helium. The detector output was amplified using a signal conditioning buffer amplifier with a gain of approximately 300. An IBM analog to digital board interfaced to a PC-XT compatible computer was used to capture data. Labtech Notebook and Labtech software from Laboratory Technologies Corporation were used for data analysis. Table 3.2 lists the thermal desorption operating parameters. Table 3.3 lists the gas chromatography operating parameters. 26

PAGE 42

Table 3.2 Table 3.3 The thermal desorption operating parameters for the solid adsorbent preconcentration method. Desorb. Flow Rate 11 ml/min Line Heater 200C Valve Heater 240C Injector Heater 200C Furnace READY 40C START input user Cryo-1 COOL 1 -125C Furnace DESORB 190C DESORB time 10.00 min CONTINUE input user Cryo-2COOL -125C Cryo-1 TRANS. 200C TRANS. time 2.00 min Cxyo-2 INJECT 200C INJECT time 1.50 min Furnace BAKE 225C BAKE time 10.00 min The gas chromatography operating parameters for the solid adsorbent preconcentration method. Column Flow Rate 3.3 ml/min Injector Temperature 250C Detector Temperature 27C Initial Oven Temperature -10 oc Initial Oven Hold Time 15 min Temperature Ramp Rate 20C/min Final Oven Temperature 150C FID Make-Up Flow Rate 16 ml/min FID Hydrogen Flow Rate 20 ml/min FID Air Flow Rate 240 ml/min 27

PAGE 43

Experimental Sampling AJ?paratus for the Solid Adsorbent Preconcentration Method Thermal desorption tubes were 1 I 4" OD x 7" stainless steel packed with 300 mg 60/80 Carbopack Band 300 mg Carbosieve s-m, purchased from Supelco. Desorption tubes were conditioned at 225C overnight with nitrogen flowing at 100 ml/min. A Pyrolysis Furnace manufactured by Dohrmann Envirotech was used to heat the desorption tubes. Standards were trapped by injecting 1.0 ul volumes of known concentrations into a helium or nitrogen stream of gas flowing at 100 ml/min. Samples were trapped using a vacuum pump (115 volt, 60 hertz, 1.23 amp pump) made by Thomas Industries Inc., with an air flow rate of 100 ml/min. Air flow rates were measured with a Singer Dry Test Meter from Singer American Meter Division. Preparation of Methanol and Ethanol Standards for the Solid Adsorbent Preconcentration Method Anhydrous methyl alcohol, 95% grade, received from Mallinckrodt was used to make stock solutions. Absolute ethyl alcohol received from AAPER Alcohol and Chemical Co. was used to make stock solutions. All dilutions were made with deionized water that 28

PAGE 44

had been filter though a Millipore system, doubly distilled, and pyrolyzed. A 0.25 M methanol and 0.34 M ethanol stock solution was prepared by diluting 1.00 ml of methanol and.00 ml of ethanol to 100 mi. Standard solutions were prepared by further dilution of stock solutions. 29

PAGE 45

CHAPTER IV RESULTS AND DISCUSSION Impinger Preconcentration Method The first objective was to find the retention time of methanol in a gas chromatogram of methanol in benzyl alcohol using the impinger preconcentration. method. To find the retention time of methanol, a blank and two standards of different concentrations were analyzed. The first solution contained 0.25 M methanol in benzyl alcohol (see Figure 4.1). The second solution contained 0.12 M methanol in benzyl alcohol (see Figure 4.2). The third solution was a blank containing only benzyl alcohol (see Figure 4.3). The first peak corresponded to methanol, because the height of this peak changed when the concentration of methanol in the standard solutions was changed. For a blank solution the first very small peak at 8 chart divisions had a peak height of 0.3 em. For a 0.25 M standard solution the first peak at 8. chart divisions had a peak height of 6.5 em. Fora 0.12 M standard solution the first peak at 8 chart divisions had a peak height of 14 em. 30

PAGE 46

Figure 4.1 A gas chromatogram of a standard solution containing 0.25 M methanol in benzyl alcohol, 4/23/1993. The split ratio was 61. (The program on the gas chromatograph was started late;) 31

PAGE 47

Figure 4.2 i A gas chromatogram of a standard solution containing 0.12 M methanol in benzyl alcohol, 4/23/1993. The split ratio was 61. I : ; .. ____ ., -,-; I. ; r-I i i i i ; . : . L, ... . 'I '' i; i: i riL: :, U.Lt !J.: :.: II j; !l Y : : : : .:.:... ... I I' 'i I ::. : I L!_;.: I' 1:; . :! I I I II: ; ' o I ... ._!.-". 32 ; . I .-,, .1: ; : : j < ; :. -: i -1 1., -_: _I I I .. I '1 : :I I ; I I j __ -, . I : I

PAGE 48

Figure 4.3 ; A gas chromatogram of a blank solution containing only benzyl alcohol, 4/23/1993. The split ratio was 61. I t I .:T : -.. i 33 1' I 'I

PAGE 49

All three chromatograms had a reagent blank peak at 9 chart divisions with a peak height of 1.2 em. It was determined that the first peak in a gas chromatogram of methanol in benzyl alcohol corresponded to methanol. The retention time for methanol was at 8 chart divisions, which corresponded to 1.3 em and 46 seconds. The blank solution had a peak at the retention time for methanol, which indicates that the solvent benzyl alcohol contained a contaminate that co-eluted with methanol. The next task was to find the detection limit of chromatography portion of the impinger preconcentration method. To find the detection limit of the impinger preconcentration method, standards of decreasing concentrations were analyzed until the gas chromatograms of the standard and blank solutions were indistinguishable. A standard solution containing 2.5 mM methanol in benzyl alcohol produced a peak height of 4.2 em for methanol (see Figure 4.4). A standard solution containing 1.2 mM methanol in benzyl alcohol produced a peak height of).B em for methanol (see Figure 4.5). A standard solution containing 0.25 mM methanol in benzyl alcohol produced a 0.9 em peak height for methanol (see Figure 4.6). A blank solution containing only the benzyl alcohol solvent produced a peak height of 1.0 em where methanol elutes (see Figure 4.7). Figures 4.6 and 4.7 indicate that the gas chromatogram of the blank solution was indistinguishable from the gas chromatogram of a standard solution containing 0.25 mM methanol in benzyl alcohol. 34

PAGE 50

Figure 4.4 1 I i ,. I I i I I A gas chromatogram of a standard solution containing 2.5 mM methanol in benzyl alcohol, 4/11/1993. The split ratio was 10. ljlllt I . -L_ ,. l -:r-!i-.: I-! I j I II ., I I : : I ; : I ' .... : . i : t--1 '!'< i i ff, 1 :!;: 1' [ I ;;.; I i ; ... +1 1(1 d 1 I rm! I : ' : 1 -i I ; t l ji! ! r:..t J" II. I I I 1 !1iill)' t -11 1: i :I+::'. ;( I ... 11 :II I I 1 ; : ,:1 T1"' ... It _; ___ --' . : ! +--; -I : i i i i I I i 1 35

PAGE 51

Figure 4.5 A gas chromatogram of a standard solution containing 1.2 mM methanol in benzyl alcohol, 4/11/1993. The split ratio was 10. 1-T. ; -1T: 1 -_--r ':'':; T ,-: i--, mn .. ,1,:_ :!! ','. 1.l t:i.yl.f--j-i ... __ L_ .... "., .,. ; r;-!-+++H+-1+ I I ; ITi I l -, I I I : ' ' I I I ' f+ ti '-h--H I -H_ : -rt:r-;: 1 -H-h I rhi 'j' Hr. M n t+ tt H :; ; i : -++t-r H ... q. -! H] tf--1:j:,+t jq-H,. HH -, T rt+H ith.-1+H nl! ;. iti-I+H+H+H-1-1-H-1 36

PAGE 52

Figure 4.6 :! I .. I I : .. : ; A gas chromatogram of a standard solution containing 0.25 mM methanol in benzyl alcohol, 4/11/1993. The split ratio was 10. : . I 37 I I I , l ; 0 : I' I i:' . i. .. I'. .; .. I I-. l i: :' :: '.: I .i i

PAGE 53

Figure 4.7 A gas chromatogram of a blank solution containing only benzyl alcohol, 4/11/1993. The split ratio was 10. 38

PAGE 54

Once the detection limit of the impinger preconcentration method was determined, the estimated minimum volume of air needed to quantify methanol was calculated. The estimated detection limit of the impinger preconcentration method was 2.5 x lo-S mol of methanol (see Appendix A). The concentration of methanol in Denver's air was predicted to be in the ppbv concentration range. The impinger trap contained 25 ml of benzyl alcohol. From this, the estimated minimum volume of ambient air needed to quantify methanol was calculated to be 740,000 liters (see Appendix A). The impinger preconcentration method does not have the sensitivity required to quantify the concentration of methanol in ambient air in real time, for example collect an ambient air sample every four hours. The maximum sampling rate was estimated to be 250 liters per hour, because collection efficiencies might be poor at larger sampling rates. If we sampled at 250 liters per hour, it would take 123 days to collect 740,000 liters of ambient air. 39

PAGE 55

The Solid Adsorbent Preconcentration Method Identification of the Methanol and Ethanol Peaks The first objective was to find the retention times of methanol and ethanol in a gas chromatogram of methanol and ethanol in water. To find the retention time of methanol and ethanol, a blank solution, a methanol standard solution, and a ethanol standard solution were analyzed using the solid aqsorbent preconcentration method. The standards and the blank were injected directly into an empty thermal desorption tube packed with a small amount of glass wooL The gas chromatogram of the blank containing only Millipore Quality Water (MQW) had three medium peaks at 3.8, 6.4, and 7.2 min and one small peak at 4.2 min (see Figure 4.8). A gas chromatogram of a standard solution containing 0.97 mM methanol in MQW had the same three medium peaks as the blank, but the small peak at 4.2 min had become larger (see Figure 4.9). The retention time of methanol was at 4.2 min as indicated by the change in the size of this peak when methanol was added to the analyzed solution. A gas chromatogram of a standard solution containing 0.67 in MQW had the same four peaks as the blank, but there was a new peak at 6.2 min (see Figure 4.10). 40

PAGE 56

Figure 4.8 3 2.5 ctl ..., !J 2 r4 (/) z 0 1.5 6 1 0 0.5 A gas chromatogram of a bl.ank solution containing only Millipore Quality Water, 9/27/1994. The oven temperature was 0C. 4 11 8 5 0 2 4 6 8 10 12 14 16 REfENTION TIME. min. 41

PAGE 57

Figure 4.9 A gas chromatogram of a standard solution containing 0.97 mM methanol in Millipore Quality Water, 9/27/1994. The oven temperature was 0C. 0 2 4 6 8 10 12 14 16 REiENTION TIME, min. 42

PAGE 58

Figure 4.10 A gas chromatogram of a standard solution containing 0.67 mM ethanol in Millipore Quality Water, 9/27/1994. The oven temperature was ooc. 16 6 5 1 7 ................ .............. .......................................................................................... 0 ............... 0 2 4 6 8 10 12 14 16 RETENTION TIME, min 43

PAGE 59

The retention time of the peak that corresponded to ethanol was at 6.2 min as indicated by the addition of this peak when ethanol was added to the analyzed solution. The above data were obtained with a GC oven temperature of 0C, and the data through out the rest of this research were obtained with a GC oven temperature of -10C. It was determined that decreasing the oven temperature increased the retention time and overall resolution of the analytes. However, oven temperatures below -10C had adverse effects on the chromatography. At -20C the methanol and ethanol had extreme peak tailing and/ or the sample seemed to plug the column. The result of the sample' plugging the column was that the sample did not elute until the oven temperature increased during the GC program. At -10C some ambient air samples had delayed retention times, but all the standards and most ambient air samples showed no adverse effects. All standards and samples analyzed throughout the rest of this research were adsorbed onto thermal desorption tubes as described in the Background and Theory Section. Standard and blank solutions were injected into a stream of flowing gas that passed through the thermal desorption tube. The flow rate was approximately 100 ml/min. A vacuum flow rate of approximately 11/min was used to load ambient air sample onto the thermal desorption tubes. The next problem was to determine the retention times of methanol and ethanol at oven temperature of -10C. To find the 44

PAGE 60

retention times methanol and ethanol, a standard solution containing 2.4 mM methanol and 1.7 mM ethanol in organic free water was analyzed (see Figure 4.12). The gas chromatogram of a blank containing only organic free water is displayed in Figure 4.11. Comparing the gas chromatograms of the standard and blank solution, it was apparent that the retention times of methanol and ethanol were at 4.9 min and 8.2 min, respectively (See Table 4.1). Reproducibility of the Solid Adsorbent Preconcentration Method The daily reproducibility of the solid adsorbent preconcentration method was studied by analyzing the same standard three times consecutively. The gas chromatograms of the blank and the three standards containing 2.4 mM methanol and 1.7 mM ethanol in organic free water are displayed in Figures 4.11 through 4.14. The results from the daily reproducibility study are given in Table 4.1. 45

PAGE 61

Figure 4.11 A gas chromatogram of a blank containing only organic free water, 11/1/1994. The oven temperature was -10C. 9 10 7 CIJ 8 14 17 23 0 ............................................ 0 2 4 6 8 10 12 14 16 RETENTION TIME. min. 46

PAGE 62

Figure 4.12 A gas chromatogram of a standard solution containing 2.4 mM methanol and 1.7 mM ethanol in organic free water, 11/1/1994 .. The oven temperature was -l0C. Run number one. 12 8 10 7 9 17 CfJ 8 13 2 2342 ------0 -0 2' 4 6 8 10 12 14 16 RETENTION TIME, min 47

PAGE 63

Figure 4.13 12 10 1 A 2 A gas chromatogram of a standard solution containing 2.4 mM methanol and 1.7 mM ethanol in organic free water, 11/1/1994. The oven temperature was -10C. Run number two. 6 4 7 14 1 1 18 g 17 3 ) 11f6 9 2021 li \ 11 \ J\, .... ' 1.-t. 0 ---........ ooUOoOOOooOO ............. -2 I I I I I I I 0 2 4 6 8 10 12 14 16 Rm:NTION TIME, min 48

PAGE 64

Figure 4.14 12 10 1 A gas chromatogram of a standard solution containing 2.4 mM methanol and 1.7 mM ethanol in organic free water, 11/1/1994. The oven temperature was -l0C. Run number three and the primary recovery sample. 7 15 4 8 12 19 2 ). \ \_ 18 an 3 !11 \ 22 f"t 1'1 ,I \ 0 ............... .................................. u u ............................................................................ -2 I 0 2 I I I I I 4 6 8 10 12 REI'ENTION TIME, min. 49 I 14 16

PAGE 65

Table 4.1 The result of the reproducibility study on 11/1/1994. Run# [MeOH]mol Area Ret (min) 1 2.4 x 1o-9 66.4 4.9 2 2.4x 1o-9 65.9 4.9 3 2.4 x 1o-9 74.4 4.9 Blank 11.1 4.9 Run # [EtOH]mol Area Ret (min) 1 1.7 x 1o-9 53.3 8.2 2 1.7x 1o-9 55.3 8.2 3 1.7x 1o-9 51.2 8.2 Blank No Peak 8.0 The average areas for the three analyses of methanol and ethanol were 68.9 4.8 and 53.3 2.1 area units, respectively. It was calculated that the relative standard deviations for methanol and ethanol were 7% and 4%, respectively. The reproducibility of the solid adsorbent preconcentration method from October 11 through December 3, 1994, was studied by comparing the area produced when standard solutions were analyzed during this time period. Table 4.2 contains the concentration of methanol and ethanol, the area, and the date for each standard. 50

PAGE 66

Table 4.2 Date 10-11-94 10-28-94 11-1-94 11-15-94 11-19-94 11-30-94 12.:.3-94 The concentration of methanol and ethanol and the area of the standards analyzed from 10/11/1994 through 12/3/1994. The moles of methanol and ethanol are the moles that were trapped on the thermal desorption tube. [MeOH]mol Area [EtOH]mol Area 2.4x 1o-9 79.7 1.7x 1o-9 27.4 2.4x 1o-9 77.1 1.7 x 1o-9 71.1 2.4x 1o-9 68.9 1.7x 1o-9 53.3 2.4x 1o-9 81.6 1.7 x 1o-9 NA 2.4 x 1o-9 80.8 1.7x 1o-9 74.1 2.4 x Jo-9 65.3 1.7x 1o-9 15.5 2.4 x 1o-9 50.9 1.7x 1o-9 15.8 NA Data 1s not available Table 4.2 indicates the reproducibility of methanol from October 11 through November 19, 1994, was fairly consistent. The calculated reproducibility was 77.6 5.2 area units with a relative standard deviation of 6.7% from October 11 through November 19, 1994. From November 30 through December 3, 1994, the reproducibility decreased. Table 4.2 indicates that the reproducibility of ethanol was worse than the reproducibility for methanol. The calculated reproducibility was 56 21 area units with a relative standard deviation of 38% from October 11 through November 19, 1994. The calculated reproducibility was 66 11 with a relative standard deviation of 17% from October 28 through November 30, 1994. From November 30 through December 3, 1994, the reproducibility decreased. 51

PAGE 67

The reproducibility of ethanol was worse then methanol due to resolution and peak tailing. Both the methanol and ethanol peaks tailed. Ethanol was never baseline resolved from a co-eluting peak and the peak tailing was worse for ethanol, causing the reproducibility of ethanol to be worse then methanol. It was never determined what caused the response of the solid adsorbent preconcentration method to vary so much. The variation might be normal and as more analyses are performed an overall range might develop. The variation might be caused by a leak somewhere in the plumbing of the solid adsorbent preconcentration method, especially in the capillary interface. The variation might be caused by inconsistent cryofocusing during the transfer of the sample. The variation might be caused by over use of the thermal desorption tubes. Compounds in air ambient air samples might irreversibly bind to the adsorbent in the thermal desorption tube, decreasing collection efficiencies. Retention Time of Methanol and Ethanol in Ambient Air Samples It was necessary to determine if the retention times of methanol and ethanol in ambient air samples were different than in standard solutions. Ambient air samples were spiked with a methanol and ethanol to identify the retention times of methanol and ethanol. 52

PAGE 68

Three consecutive ambient air samples were obtained on three separate thermal desorption tubes. The first desorption tube was not spiked with methanol and ethanol (see Figure 4.15). The second thermal desorption tube was spiked with 2.3 x 1Q-9 moles of methanol and 1.7 x 1o-9 moles of ethanol (see Figure 4.16). The third desorption tube was spiked with 4.9 x 1o-9 moles of methanol and 3.4 x 1Q-9 moles of ethanol (see Figure 4.17). The size of two peaks at the expected retention times of methanol and ethanol increased when the sample was spiked with methanol and ethanol. The three gas chromatograms indicated that in ambient air samples the retention times for methanol and ethanol were 5.0 min and 8.4 min, respectively. Collection Efficiency Study The next task was to study the collection efficiency of the thermal desorption tubes. To find the collection efficiency, the recovery of methanol and ethanol from standard solutions was investigated. Standard solutions were injected in to a flowing stream of nitrogen that passed through two thermal desorption tubes in series. The primary thermal desorption tube was used to capture the analytes from the standard sample. The secondary thermal desorption tube was used to capture any analytes that were not adsorb by the primary thermal desorption tube. 53

PAGE 69

Figure 4.15 A gas chromatogram of an ambient air sample, 11/5/1994. Six liters was collected from 11:08-11:14 am. 12 3 10 4 12 9 1 2E 0 .---. ---. ---. 0 2 4 6 8 10 12 14 16 RRI'ENTION TIME. min. 54

PAGE 70

Figure 4.16 12 10 en 8 1 0 --2 0 A gas chromatogram of an ambient air sample spiked with 2.3 x lQ-9 moles of methanol and 1.7 x 1o-9 moles of ethanol, 11/5/1994. Six liters was collected from 11:15-11:21 am. 11 4 10 9 17 5 1 f 7 14 20 8 as &v ?3 \ 15 \lg \ 27 2lB .N .. -.... ................................................................................................................................. I 2 I I I l -, 4 6 8 10 12 RETENTION TIME. mm. 55 I 14 16

PAGE 71

Figure 4.17 A gas chromatogram of an ambient air sample spiked with 4.9 x 1o-9 moles of methanol and 3.4 x 1Q-9 moles of ethanol, 11/5/1994. Six liters was collected from 11:22-11:28 am. 11' 10 10 8 6 5 8 4 18 2 0 13 ,:;, 19 1 \ J_h \ ftf'... -2 0 2 4 6 8 10 12 14 16 RETENTION TIME, min. 56

PAGE 72

See Figure 4.14 for the gas chromatogram of the primary recovery sample. See Figure 4.18 for the gas chromatogram of the secondary recovery sample. Table 4.3 lists the result of the collection efficiency study. Table4.3 The result. of the collection efficiency study completed on 11/1/1994. See Figures 4.14.,4.18, and 4.19. SampleiD [MeOH]mol Area Ret (min) Primary 2.4 x Jo-9 74.4 4.9 Secondary 9.1 4.9 Blank 11.0 4.9 Conditioned Blank 7.8 5.0 SampleiD [EtOH]mol Area Ret (min) Primary 1.7 x Jo-9 51.2 8.2 Secondary 0 The data in Table 4.3 indicate that the recovery of methanol was 78% in the worst case, with no blank correction. However, there was some evidence to support the theory that the gas chromatogram of the secondary recovery sample was no different from a gas chromatogram of a blank. The gas chromatogram of the secondary recovery sample had a peak area of 9.1 area units for methanol, see Table 4.3. The gas chromatogram of the blank had a peak area of 7.8 area units for methanol, see Table 4.3. The blank can have a peak at the retention time for methanol as large as the peak for methanol in the secondary recovery sample. 57

PAGE 73

Figure 4.18 12 10 {f.J 8 1 2 A gas chromatogram of the secondary recovery sample for run number 3, 11/1/1994. 7 5 16 8 6 20 34 17 1\. AI rf 2:l22324 0 ............................................ .. I I I I I I I 0 2 4 6 8 10 12 14 16 REiENTION TIME. min. 58

PAGE 74

Figure 4.19 A gas chromatogram of a conditioned blank. The thermal desorption tube was condition over night and then analyzed, 11/1/1994. 10 en 8 ...,; ..-( 0 > [Ij 6 (/) z 0 4 o:; o:; 2 Cl 0 -2 4 2122 0 2 4 6 8 10 12 14 16 RETENTION TIME. mm. 59

PAGE 75

Thus, it was possible that the recovery of methanol was near 100%, because the gas chromatogram of the secondary recovery sample and the blank were similar. The data in Table 4.3 indicate that the recovery of ethanol was 100%. The gas of the secondary recovery sample did not have a peak at the retention time for ethanol. Calibration Curve for Methanol and Ethanol The next goal was to create a calibration curve for the quantification of ethanol and methanol in ambient air. Three calibration curves were obtain by analyzing standard solutions containing methanol and ethanol on November 19, November 30, and on December 3, 1994 (see Tables 4.4 through 4.6 for the results). Figure 4.25 displays the calibratio_n curves for both methanol and ethanol. Table 4.71ists the least squares linear regressions and the correlation 'coefficients for the calibration curves. 60

PAGE 76

Table 4.4 Table 4.5 The concentration of methanol and ethanol for each standard with its corresponding area and retention time. The analysis was performed on 11/19/1994, (See Figures 4.20 through 4.24). SID [MeOH]mol Area Ret (min) 1 2.4x 1o-9 80.8 4.9 2 1.1 x 1o-9 49.6 4.9 3 7.2 x 1o-1o 35.1 4.9 4 2.4 x 1o-1o 16.6 5.0 Blank 7.4 5.0 SID [EtOH]mol Area Ret (min) 1 3.3 x 1o-9 125.7 8.2 2 1.7x 1o-9 74.1 8.2 3 1.0 x lo-9 42.4 8.3 4 3.3x 1o-10 12.2 8.4 The concentration of methanol and ethanol for each standard with its corresponding area and retention time. The analysis was performed on 11/30/1994. SID [MeOH] mol Area Ret (min) 1 2.4x 1o-9 65.3 4.9 2 1.1 X 10-9 52.9 4.8 3 7.2 x 1o-1o 31.1 4.9 4 2.4 x 1o-1o 7.7 5.0 Blank 5.7 5.0 STD [EtOH]mol Area Ret (min) 1 3.3 x 1o-9 50.6 8.3 2 1.7 x 1o-9 15.5 8.3 3 1.0 x 1o-9 10.3 8.5 4 3.3 x 1o-1o 2.1 8.6 61

PAGE 77

Figure 4.20 A gas chromatogram of standard one containing 2.4 mM methanol and 3.3 mM ethanol, 11/19/1994. 8 6 10 5 1 14 1\ 1\ 0 . 0 2 4 6 8 10 12 14 16 REI'ENTION TIME. min. 62

PAGE 78

Figure 4.21 A gas chromatogram of standard two containing 1.1 mM methanol and 1.7 mM ethanol, 11/19/1994. 6 8 7 7 8 5 0 4 3 2 Q 2 11 5 1 1 0 .............................................................................................................................. -1 0 2 4 6 8 10 12 14 16 RETENTION TIME, min. 63

PAGE 79

Figure 4.22 A gas chromatogram of standard three containing 0.72 mM methanol and 1.0 mM ethanol, 11/19/1994. 8 7 7 6 Ul +l 8 5 r4 (I) z 4 0 11 11. 3 u 2 15 Q 1 0 -1 0 2 4 6 8 10 12 14 16 REI'ENTION TIME, min. 64

PAGE 80

Figure 4.23 12 10 12 3 A gas chromatogram of standard four containing 0.24 mM methanol and 0.33 mM ethanol, 11/19/1994. 20 8 7 9 22 16 56 21 \ \_ 2 B 0 ................................................. ........................................................... -2 I I I I I I I 0 2 4 6 8 10 12 14 16 REIENTION TIME. min. 65

PAGE 81

Figure 4.24 A gas chromatogram of a blank containing only organic free water, 11/19/1994. 5 8 7 16 6 1 19 2b 0 ...... ..... ....... .. .......... ................................................. ..... . . ........ . . . ........................................ ....... ........ . ..... ................ -1 .---, .---. 0 2 4 6 8 10 12 14 16 RETENTION TIME. min. 66

PAGE 82

. Table 4.6 The concentration of methanol and ethanol for each standard with its corresponding area and retention time. The analysis was performed on 12/3/1994. SID [MeOH]mol Area Ret (min) 1 2.4x 1o-9 50.9 4.9 2 1.1 x 1o-9 40.8 4.8 3 7.2 x 10-10 26.9 4.9 4 2.4 x 10-10 16.3 4.9 SID [EtOH]mol Area Ret (min) 1 3.3 x 10-9 36.7 8.2 2 1.7x 1o-9 -15.8 8.3 3 1.0 x 10-9 12.3 8.4 The calibration curves in Figure 4.25 indicate that the sensitivity for methanol decreased with time as indicated by the decreasing slopes. However, the sensitivity for standards 2, 3, and 4 for methanol were fairly consistent. The sensitivity from standard 1 for methanol was not very consistent. The calibration curves in Figure 4.25 indicate that the sensitivity for ethanol decreased with time as indicated by the decreasing slopes. The sensitivity for ethanol on November 19, was much greater than the sensitivity for ethanol on November 30, 1994, or on December 3,1994. The sensitivity for ethanol on November 30, 1994, was similar to the sensitivity for ethanol on December 3,1994 .. 67

PAGE 83

Figure 4.25 90 75 60 lU 45 < 30 15 0 140 120 100 lU 80 < 60 40 20 0 Calibration curves for the concentration of methanol and ethanol versus area. The analyses were performed on 11/19/1994,11/30/1994, and 12/3/1994. 5.0e-10 1.0e-9 15e-9 2.0e-9 [MeOH] moles injected 1!1 a 1!1 11-19-94 11-30-94 a 12-3-94 25e-9 11-19-94 11-30-94 12-3-94 3.0e-9 5.0e-10 l.Oe-9 1.5e-9 20e-9 25e-9 3.0e-9 3.5e-9 4.0e-9 [EtOH] moles injected 68

PAGE 84

Table 4.7 Date 11/19/94 11/30/94 12/3/94 Date 11/19/94 11/30/94 12/3/94 The least squares linear regressions and correlation coefficients for the calibration curves of the concentration of methanol and ethanol versus area. Least Squares Linear Regressions for MeOH R"2 y = 12.227 + 2.925 x 1010 0.993 y = 10.100 + 2.559 x 1010 0.868 y = 15.677 + 1.585 x 1010 0.924 Least Squares Linear Regressions for EtOH R"2 y = 4.047 + 3.762 x 1o1o 0.989 y = -6.280 + 1.637 x 1o1o 0.965 y = -0.231 + 1.093 x 1010 0.975 The decrease in sensitivity of the solid adsorbent preconcentration method could have been caused by several different things. The instability in the flame gas conditions could cause the decreased sensitivity. It hard to control the flowrate of hydrogen, due to a clogging flow restricter. The sensitivity of the system is dependent on the flow rate of the hydrogen gas. The decreased flow rate of hydrogen caused the sensitivity to decrease .. The make-up gas and air flow rates could also adversely effect the sensitivity. Overuse of the thermal desorption tubes could cause the decreased sensitivity. Compound from previous air samples could irreversibly bind to the solid adsorbent, decreasing the active sites and the ability to trap methanol and ethanol. Problems with the electrometer on the gas chromatograph could have caused the decrease in sensitivity. It was known that the electrometer was not producing the voltage it should 69

PAGE 85

produce. We used an amplifier to increase the voltage of the electrometer to capture the signal with the computer. However, the electrometer could have been failing, producing smaller voltages with time. The Ambient Air Sample Studies The reproducibility study above indicated that the solid adsorbent preconcentration method's response was fairly consistent from October 11 through November 19, 1994. Thus, estimates of the concentration of methanol and ethanol in ambient air samples obtained throughout this time were made using the calibration curve obtained on November 19, 1994. An ambient air sample was obtained at the Auraria Ambient Air Sampling Station on October 14, 1994. Nine liters of ambient air was collected from 8:50am to 9:00am (See Figure 4.26). The retention times in the gas chromatogram were shifted. Thus, the peak positions relative to the other peaks and the peak shapes were used to determine which peaks corresponded to methanol and ethanol. Peak 11 at 5.7 min corresponded to methanol in the gas chromatogram. Peak 18 at 9.4 min corresponded to ethanol in the gas chromatogram. 70

PAGE 86

Figure 4.26 A gas chromatogram of an ambient air sample, 10/14/1994. Nine liters was collected from 8:50-9:00 am. 12 710 13 21 26 10 Cll 2 +J 8 -fir 11 2 z 6 0 f2 g 0:: 0:: 4 25 a _[ 416 23 u 2 .B455 Q lle A 1 \ Hll \I\ 4 I.. 0 ........................................................................................................................................................................................ 0 2 4 6 8 10 12 14 16 RETENTION TIME, min. 71

PAGE 87

The estimated concentrations of methanol and ethanol were 3.6 ppbv and 0.7 ppbv, respectively. An ambient air sample was obtained at the Auraria Ambient Air Sampling Station on October 27, 1994. Twelve liters of ambient air was collected from 11:32 am to 11:44 am (See Figure 4.27). Peak 13 at 5.0 min corresponded to methanol in the gas chromatogram. Peak 22 at 8.2 min corresponded to ethanol in the gas chromatogram. The estimated concentrations of methanol and ethanol were 3.0 ppbv and 2.3 ppbv, respectively. An ambient air sample was obtained at the Auraria Ambient Air Sampling Station on October 28, 1994. Twelve liters of ambient air was collected from 12:19 pm to 12:31 pm (See Figure 4.28). Peak 11 at 5.0 min corresponded to methanol in the gas chromatogram. Peak 18 at 8.2 min corresponded to ethanol in the gas chromatogram. The estimated concentrations of methanol and ethanol were 2.3 ppbv and 3.4 ppbv, respectively. Two ambient air samples were obtained at the Auraria Ambient Air Sampling Station on November 2, 1994. For the first sample, six liters of ambient air was collected from 12:07 pm to 12:14 pm (See Figure 4.29). Peak 10 at 4.9 min corresponded to methanol. The program did not integrate the ethanol peak, so an estimate of the peak area was made by comparing the ethanol peak to another peak. Peak 21 was similar to the ethanol peak and its area was used to estimate the concentration of ethanol. 72

PAGE 88

Figure 4.27 12 10 en ...; 8 ...... 0 > w U1 6 z 0 0... ij} [:j 0:: 4 u w f-. ) w .::.., 0 0 1 -2 0 A gas chromatogram of an ambient air sample, 10/27/1994. Twelve liters was collected from 11:32-11:44 am. 14 25 30 36 I 32 12 29 4 22 26 6 34 5 3 18 31 2 24 u \A 33 .. ... 1 .... 03 7 ..A . .\ ... r-1/\ I I I I I I T 2 4 6 8 10 12 14 REfENTION TIME. min. 73 ... 16

PAGE 89

Figure 4.28 12 10 rn 8 ,._) ...... C:4 6 [fJ z 0 4 2 A gas chromatogram of an ambient air sample, 10/28/1994. Twelve liters was collected from 12:19-12:31 pm. ,...,. !J 10 21 26 9 2 11 18 25 28 22 32 30 4 16 7 \: \, \..A ....!', 1 lA..LC 0 0 -.. --------4 0 I 2 I I I I I I 4 6 8 10 12 14 16 RETENTION TIME. min. 74

PAGE 90

Figure 4.29 A gas chromatogram of the first ambient air sample, 11/2/1994. Six liters was collected from 12:07-12:14 pm. g 25 10 12 en 8 7 > 2 1 27 30 28 13 0 0 2 4 6 8 10 12 14 16 RITENTION TIME. min 75

PAGE 91

The estimated concentrations of methanol and ethanol in the first sample were 1.7 ppbv and 3.4 ppbv, respectively. For the second sample, six liters of ambient air was collected from 12:15 pm to 12:21 pm (See Figure 4.30). The retention times in the gas chromatogram were shifted. Thus, the peak positions relative to the other peaks and the peak shapes were used to determine which peaks corresponded to methanol and ethanol. 9 at 7.2 min was the peak that corresponded to methanol in the gas chromatogram. There was not a peak in the gas chromatogram that corresponded to ethanol. The estimated concentration of methanol in the second sample was 2.9 ppbv. An ambient air sample was obtained at the Auraria Ambient Air Sampling Station on November 5, 1994. Six liters of ambient air was collected from 11:08 am to 11:14 am (See Figure 4.15). Peak 12 at 4.9 min corresponded to methanol in the gas chromatogram. Peak 19 at 8.3 min corresponded to ethanol in the gas chromatogram. The estimated concentrations of methanol and ethanol were 9.1 and 3.1 ppbv, respectively. The concentrations of methanol and ethanol for following ambient air samples were calculated using calibration curves obtained on the same day. An ambient air sample was obtained at the Auraria Ambient Air Sampling Station on November 19, 1994. Six liters of ambient air was collected from 2:55pm to 3:01pm (See Figure 4.31). 76

PAGE 92

Figure 4.30 12 10 UJ ._) 8 -0 :> w CfJ 6 z 0 ill 0:: 4 0:: 2 [i] Cl A gas chromatogram of the second ambient air sample, 11/2/1994. Six liters was collected from 12:15-12:21 pm. 8 16 21 10 2 7 g 14 "' I 1 1 5 2324 t ? 1 lA k.J gA A .A -II 0 -----. ... .... ................................. -2 I I I I I I I 0 2 4 6 8 10 12 14 16 RETENTION TIME. min. 77

PAGE 93

Figure 4.31 A gas chromatogram of an ambient air sample, 11/19/1994. Six liters was collected from 2:55-3:01 pm. 18 23 10 2 9 28 -2 0 2 4 6 8 10 12 REI'ENTION TIME. min. 78 14 16

PAGE 94

The retention times in the gas chromatogram were shifted. Thus, the peak positions relative to the other peaks and the peak shapes were used to determine which peaks corresponded to methanol and ethanol. Peak 9 at 6.2 min corresponded to methanol in the gas chromatogram. There was not a peak that 'corresponded to ethanol in the gas chromatogram. The concentration of methanol was calculated to be 5.9 ppbv. Two ambient air samples were obtained at the Auraria Ambient Air Sampling Station on November 30, 1994. For the first sample, nine liters of ambient air was collected from 1:41pm to 1:50pm (See Figure 4.32). The retention times in the gas chromatogram were shifted. Thus, the peak positions relative to the other peaks and the peak shapes were used to determine which peaks corresponded to methanol and ethanol. Peak 8 at 8.8 min corresponded to methanol in the gas chromatogram. Peak 8 was off scale. Thus, the area of peak 8 was underestimated. Peak 14 at 11.5 min corresponded to ethanol in the gas chromatogram. The concentrations of methanol and ethanol in the first sample were calculated to be 12 ppbv and 1.0 ppbv, respectively. For the second sample, three liters of ambient air was cpllected from 4:12pm to 4:15pm. There is not a chromatogram available for this sample. There was not a peak in the gas chromatogram that corresponded to ethanoL The concentration of methanol in the second sample was calculated to be 9.7 ppbv. 79

PAGE 95

Figure 4.32 12 10 (IJ 8 .-l -w 6 [fJ z 0 4 f2 2 Q 0 ................... -2 -4 0 A gas chromatogram of the first ambient air sample, 11/30/1994. Nine liters was collected from 1:41-1:50 pm. 89 17 291 2B 7 19 1 5 0 14 12 1 : 8 v '-, ................................ ................ ------I I I I I I I 2 4 6 8 10 12 14 16 RETENTION TIME. mm. 80

PAGE 96

An ambient air sample was obtained at the Auraria Ambient Air Sampling Station on December 3, 1994. Six liters was collected from 2:15 pm to 2:21 pm (See Figure 4.33). Peak 11 at 4.9 min corresponded to methanol in the gas chromatogram. There was not a peak in the gas chromatogram that corresponded to ethanol. The concentration of methanol was calculated to be 9.2 ppbv. One of the problems with some of the ambient air sample data is the delayed retention times of the peaks in the chromatograms. The delayed retention times were probably caused by the large amount of water in the ambient air samples. The concentration of water was estimated to be 0.73% at 0C and 5.0% at 30C, assuming a relative humidity of 100% (see Appendix B). It was thought that this amount of water might be capable of freezing in the column and plugging it partially or completely. A summary of the concentration of methanol and ethanol in ppbv for the ten ambient air samples listed above is given in Table 4.8. Sample volumes range from three to twelve liters with a sampling rate of 1 liter per hour. 81

PAGE 97

Figure 4.33 A gas chromatogram of an ambient air sample, Six liters was collected from 2:15-2:21 pm. 10 12 18 24 7 2123 28 1 0 0 2 4 6 8 10 12 14 16 REI'ENTION TIME. min. 82

PAGE 98

Table 4.8 Summary of the concentration of methanol and ethanol in ten ambient air samples collected at the Auraria Ambient Air Sampling Station. Air Sample Date [MeOH]ppbv [EtOH]ppbv 10-14-94 @ 8:50 am 3.6 0.7 10-27-94 @ 11:32 am 3.0 2.3 10-28-94 @ 12:19 pm 2.4 3.4 11-2-94@ 12:07pm 1.7 3.4 11-2-94 @ 12:15 pm 2.9 11-5-94@ 11:08 am 9.1 3.1 11-19-94@ 2:55pm 5.9 11-30-94@ 1:41 pm 12 1.0 11-30-94 @ 4.21 pm 9.7 12-3-94@ 12:15 pm 9.2 Average 5.9.7 2.3 1.2 . .. -The results were below the mmtmum quantifiable limit for ethanol. The results in Table 4.8 indicate that methanol is present in the ambient air _in Denver at ppbv concentrations. No reliable historical data exists on the concentration of methanol in Denver's ambient air, for comparison with our data. Kelly, Callahan, Plell, and Evans found from 22 samples that the concentration of methanol in Boston was 7.2 to 47 ppbv with and average of 17 ppbv, and in.Houston was 5.6 to 31 ppbv with an average of 17 ppbv (17). In Los Angeles Hanst, W<;mg, and Bragin found, from one infrared study, that the concentration of methanol was 10 ppbv (11). In Denver we found, from 10 ambient air samples, that the concentration of methanol was 1.7 to 12 ppbv with an average of 5.9 3.7 ppbv. The results of our research indicate that 83

PAGE 99

the ambient air in Denver contains less methanol than the ambient air in Boston, Houston, or Los Angeles. The least concentrated standard was used as the minimum quantifiable limit for methanol. Thus, the minimum quantifiable limit for methanol was ppbv. The results of this experiment were more than three times the minimum quantifiable limit for methanol. The results in Table 4.8 indicate that ethanol is also present in the ambient air in Denver at ppbv concentrations. No reliable historical data exists on the concentration of ethanol in Denver's ambient air, for comparison with our data. Kelly, Callahan, Plell, and Evans found, from 15 samples, that the concentration of ethanol in Boston was <1.0 to 38 ppbv with an average of 4.9 ppbv, and in Houston was <1.0 to 22 ppbv with and average of 2.9 ppbv (17). In Denver we found, from 6 ambient air samples, that the concentration of ethanol was 0.7 to 3.4 ppbv with an average of 2.3 1.2 ppbv. The results of our research in fair agreement with the results from the other_ two studies. The least concentrated standard was used as the minimum quantifiable limit for ethanol. From October 14 through November 19, 1994, the minimum quantifiable limit for ethanol was 0.33 ppbv. Due to the decreasing response to standard solutions from November 30 through December 3, 1994, the minimum quantifiable limit for ethanol was 1.0 ppbv. From October 14 through November 19, 1994, the results of this experiment were more than three times the minimum 84

PAGE 100

quantifiable limit for ethanol for all but one ambient air sample obtained of October 14, 1994. The sample obtained on October 14, 1994, was more than two times the minimum quantifiable limit for ethanol. Due to the decreasing response of the solid adsorbent preconcentration method, from November 30 through December 3, 1994, the results of this experiment were near or below the minimum quantifiable limit. The decreasing response of the solid adsorbent preconcentration method effected the analysis of ethanol more than methanol. 85

PAGE 101

CHAPTER V CONCLUSION We developed a method of quantifying methanol and ethanol in ambient air during this research project. Methanol and ethanol adsorbed on to thermal desorption tubes were thermally desorbed, cryofocused, and analyzed by GC-FID. This report the results from ten ambient air samples collected at the Auraria Ambient Air Station from October 14 through December 12, 1994. In Denver we found the concentration of methanol was 1.7 to 12 ppbv with an average of 5.9 3.7 ppbv, and ethanol was 0.7 to 3.4 ppbv with an average of 2.3 1.2 ppbv. The results of this study are in fair agreement with previous studies by Kelly, Callahan, Plell, and Evans (17). We concluded the method described by the solid adsorbent preconcentration method is capable of quantifying methanol and ethanol in ambient air. However, we need to collect more data to gain confidence in the solid adsorbent preconcentration method and the results produced by it. Our objective was to find a method of quantifying methanol and ethanol in ambient air by using a standard gas chromatography with 86

PAGE 102

flame ionization detection. At present, it is necessary to cryofocus the ambient air samples before GC analysis. The solid adsorbent preconcentration method and the Kelly, Callahan, Plell, and Evans method, both used thermal desorption instruments to cryofocus analytes (17). The Fung method used a 6-port valve as a freeze out trap to cryofocus analytes before GC analysis (16). The Fung method does not allow for the cryofocusing of analytes directly onto the GC column. Cryofocusing directly on column eliminates dead volume problems between the cryofocusing trap and GC column. The solid adsorbent preconcentration method is in the early stages of development and still needs optimization. However, the information presented in this report provides a basis for future studies. We focused on obtaining results assuming we could optimize the method later. Several problems that developed during this research will need addressing to optimize the solid absorbent preconcentration method . One problem dealt with the water in ambient air. Water interferes with most analyses of methanol and ethanol and this research was no different. We estimated the concentration of water was 0.73% at 0C and 5.0% at 30C in the ambient air of Denver, assuming 100% humidity (see Appendix B). We thought the large amount of water in the ambient air caused the shifted retention times in the gas chromatograms. The idea was that there was enough water to freeze into a plug inside the capillary column partially or completely 87

PAGE 103

blocking the flow of carrier gas. Removing the water before GC analysis or using a packed column might solve this problem. Kelly, Callahan, Plell, and Evans used a 13 min dry purge cycle in which dry nitrogen flowed through the trap at 100 cc/min at room temperature to remove approximately 90% of the water (17). We did not attempt to reinove the water before analysis. We focused on using a capillary column fqr the separation of methanol and ethanol. We did not investigate use_ of a packed column for the separation of methanol and ethanol. A packed column might \ provide better results than a capillary column .. The large amounts of water would not plug a packed column, because it has a larger sample capacity. A packed column also allows larger carrier gas flow rates, which permits more optimal flame ionization conditions. The second problem deals the conditioning of the thermal desorption tubes. The helium and nitrogen gas used to load the standards produced gas chromatograms that contained a peak at the retention time for methanol and a peak near the retention time for ethanol. The purity of helium was 99.995% and nitrogen was 99.998%. The concentrations of imp:urities in the helium was 50 ppm and nitrogen was 20 ppm. Because the solid adsorbent preconcentration method is capable of detecting ppb concentrations of methanol and ethanol, the ppm concentrations of impurities cause significant interference. The thermal desorption tubes contain carbon trap type packing. Thus, in future studies, the use a carbon trap might remove 88

PAGE 104

some of the impurities from the helium or nitrogen. Alternatively, a heated copper trap could pyrolyze the hydrocarbon impurities in the helium or nitrogen to carbon dioxide. The carbon dioxide would not interfere with the analysis, because it is not detectable by the FID. The third problem was the inability to predict the sample volume before collection. The solid adsorbent preconcentration method allowed for ambient air sample from 3 to 121iters. Currently there is no accurate way of predicting the sample size necessary for the analysis. If volume of ambient air sampled was too small, then the concentrations of methanol and ethanol would be below the minimum quantifiable limits of the solid adsorbent preconcentration method. Largesample volumes might trap large amounts of water, methanol, and/or ethanol. We thought the large amount of trapped water froze in the GC column partially or completely blocking the flow of carrier gas. Trapping large amount of methanol and/ or ethanol might exceed the capacity of the IBM analog to digital interface board. With large sample volumes the water from the sample plugged the GC column and/ or the methanol and/ or ethanol peaks were off scale. One way to deal with this problem would be to collect 2 or 3 samples simultaneously of different sample volumes. Another to deal with this problem would be to collect ambient air in an air canister. Analysis of aliquots from the canister would then be possible. SUMMA process canisters containing six liters are commercially available. A bellows pump could be used to force twenty to thirty liters 89

PAGE 105

of ambient air into a SUM:MA canister allowing replicate analyses on the same sample. The fourth problem was the poor resolution of methanol and ethanol. We never accomplished complete baseline resolution of ethanol. The resolution might improve by changing to a different capillary column, or by using two capillary columns in series. We investigate only one type of capillary column during this research. A different capillary column might provide better resolution than the DB-1 column used in this research. The DB-1 column has a non-polar stationary phase, and J & W Scientific recommends the very polar DBWAX capillary column as the best stationary phase for the analysis of alcohols. Changing to a DB-WAX column or some other capillary column might improve resolution and decrease tailing of the methanol and ethanol peaks. The use of two colu.mns in series might improve resolution. A column switching apparatus could allow the use of two capillary columns in series. The carrier gas from the first column could be foreflushed to another column until the last compound of interest elutes and then vented to air. The Fung method used a 6-port valve for column switching (16). Fung's method has the added capability of being able to backflush both columns shortening the cycle time. Fung obtained a detection limit of 1 ppbv for methanol and a complete cycle time under ten minutes. Using two columns with a column switching 90

PAGE 106

apparatus allows increased resolution and the reduction in analysis time. The fifth problem is that the solid adsorbent preconcentration method consumed large quantities of liquid nitrogen. The solid adsorbent preconcentration method uses liquid nitrogen to cool both cryotraps and to cool the GC oven. At present, using the equipment available, one chromatography run takes approximately 4 liters of liquid nitrogen. The cost of liquid nitrogen is $175 for 196 liters. The cost of the liquid nitrogen for one analysis is approximately $4. We did not attempt to optimize the cryotrapping temperature. Cooling the cryotraps to -100C instead of -125C would consume less liquid nitrogen, and might not decrease trapping efficiencies. The use of a different capillary column or two capillary columns in series might allow GC separation at a warmer temperature. A higher GC oven temperature will reduce the liquid nitrogen consumed by the GC. Recommendations for Future Studies The gas chromatograms from conditioned thermal desorption tubes indicated the method for conditioning thermal desorption tubes needs improving. We conditioned thermal desorption tubes by flowing helium or nitrogen at 230C overnight. We determined that the peaks in the conditioned thermal desorption tube chromatograms 91

PAGE 107

came from the ppm contaminates in helium or nitrogen. A gas purifier trap of some sort might clean up the helium or nitrogen. Commercially available traps containing activated charcoal, or a heated copper trap might eliminate contaminates. We never optimized the chromatography parameters for the solid adsorbent preconcentration method. Part of the problem was that there was no way of obtaining two or more identical ambient air samples. A sampling apparatus designed to sample through two or more thermal desorption tubes simultaneously could provide replicate ambient air samples. Replicate ambient air samples might allow optimization of the chromatography parameters for the solid adsorbent preconcentration. A column switching device of some sort might improve the resolution of the solid adsorbent preconcentration method. The idea is to foreflush analytes trapped on the first GC column until the last compounds of interest elutes to the second GC column. Then switch the valve permitting foreflushing to vent or backflushing the first GC column, while the second GC column resolves the analytes. The use of column switching would keep late eluting species like water from entering the second GC column shortening the analysis time. We assumed that the water eluted from the GC column at approximately eighteen minutes, because.the flame in the FID was extinguished at this time. A column switching device could prevent water from 92

PAGE 108

entering the second GC column, eliminating waters adverse effects on the FID. An improved data acquisition system capable collecting a larger range of output signals from the FID might eliminate some of the sample volume problems mentioned above. H larger output signals were quantifiable, then larger sample volumes are possible without the signals for methanol and ethanol exceeding the range of the data collection system. Sampling larger volumes of ambient air decreases the chance that the concentration of methanol and ethanol will be below the minimum quantifiable limits. However, larger volumes of ambient air contain more water that might need removal before thermal desorption when using capillary columns. Removal of the water on the thermal desorption tube is possible by flowing helium or nitrogen through the thermal desorption tube before analysis. With a drying procedure the loss of methanol and ethanol during the procedure needs investigation . The use of different types of capillary columns and/or glass packed columns needs investigation to see which combination provides the best resolution in the shortest amount of. time. 93

PAGE 109

APPENDIX A ESTIMATED VOLUME OF AIR NEEDED TO QUANTIFY METHANOL IN THE AMBIENT AIR USING THE IMPINGER PRECONCENTRATION METHOD The detection limits of the instrument are 1.00 mM methanol in benzyl alcohol. H the impinger contains 25 ml of benzyl alcohol, then it takes 2.50 x 1o-5 moles of methanol to reach the detection limit of the impinger preconcentration method. (1.00 x 10 moll-1)(0.0251) = 2.50 x 1o-5 mol The atmospheric pressure in Denver is approximately 630 mm Hg. To convert to atmospheres you multiply by 1.316 x 1o-3 atm/mm Hg(CRC). (630 mm Hg)(1.316 x 1o-3 atm mm Hg-1) = 0.829 atm The Ideal Gas Law was used to calculate the moles of gas in one liter of air. We calculated that there is approximately 0.0337 moles of gas per liter at 300 K. 94

PAGE 110

PV = nRT n = PV RT n = (0.829 atm)(1 1) (0.0821 atm K-1 mol-1)(300K) n = 0.0337 mol H the concentration of methanol in Denver's atmosphere is one part per billion, there are 3.37 x 10-11 moles of methanol in one liter of air. (0.0337 moll-1)(1 x 1o-9 ppmv) = 3.37 x 10-11 moll-1 It takes 2.50 x 1 o-5 moles of methanol in 25 ml of benzyl alcohol to detect methanol with the instrument. There are 3.37 x 10-11 moles per liter of methanol in Denver's air. Thus, the system requires that 9.2x 10-51iters of air is bubbled through the impinger before the system can detect and quantify the methanol. (2.50 X 10-5 mol methanol) = 7.4 X 1051 (3.37 x 1o-ll moles 1-1) 95

PAGE 111

APPENDIX B MOLES OF WATER IN THE AMBIENT AIR The moles of water in the ambient air were estimated from the vapor pressure of water at 0C and 30C. The atmospheric pressure in Denver is approximately 630 mm Hg. The vapor pressure of water at ooc is 4.6 mm Hg, and at 30C is 31.8 mm Hg (18). If the relative humidity was 100% and the ambient temperature was 0C, the concentration of water was calculated to be 0.73%. If the relative humidity was 100% and the ambient air temperature was 30C, the concentration of water was calculated to be 5.0%. 96

PAGE 112

REFERENCES (1) Public taws 100-494 [A.1518]; October 14, 1988. Alternative Motor Fuels Act of 1988. (2) "Denver Fleet Management Center'' Personal Communication with the General Service Administration in Denver, Colorado. (3) Finlayson-Pitts, B.J.; Pitts, J.N., Jr. "Atmospheric Chemistry: Fundamentals and Experimental Techniques", John Wiley & Sons, 1986. (4) U.S. Environmental Protection Agency. Reported in "EPA Adds 37 Additional Cities to Ozone Non-Attainment List". Air/Water Pollution Report, 1989, 245. (5) Bata, R.M.; Elrod, A. C.; Rice, R.W. Journal of Engineering for Gas Turbines and Power 1989,111.425. (6) Miron, W.L., "Air Pollution Control Division Report to the Colorado Air Quality Control Commission", 1993. (7) Livo, K.B., Miron, W.; Hollman, T. "Air Pollution Control Division Report to the Colorado Air Quality Control Commission", 1994. (8) Bhatt, A.; Gupta, V. K. Ind. J. Environ. Health 1980, QQ., 588. (9) P; Gupta, V.K. Talanta 1984, 394. (10) Maeda, Y; Fujio, Y.; Suetaka, T.; Makoto, M. Analyst 1988,113, 189. (11) Hanst, P.L.; Wong, N.W.; Bragin, J. Atmospheric Environment, 1982, 1Q, 969. (12) Gabele, P.; Ray, W.; Duncan, J.; Burto.n, C. "Characterization of Exhaust Emissions from Methanol and Gasoline Fueled Automobiles", EPA 460/3-82-004, August 1982. 97

PAGE 113

(13) Gabele, P.A.; Ray, W.D. "Preliminary Evaluation of a Method Using an FID for Measurement of Methanol in Auto Emissions", EPA/600/3-87 /035, September 1987. (14) Schuetzle, D.; Jensen T.E.; Nagy, D.; Prostak, A.; Hochhauser, A. Anal. Chern. 1991,63 (23), 1149A. (15) Haky, J.E.; Stickney, T.M. 1 Chrom., 1985, 321, 137. (16) Fung, K. 'Troceedings of the 1991 U.S. EPA/ A&WMA International Symposium: Measurement of Toxic and Related Air Pollutants" 1991, Vol. 2, 770. (17) Kelly, T.J; Callahan, P.J.; Plell, J.; Evans, G.F. Environ. Sci. Technol., 1993, 'l:Z, 1146. (18) ''Handbook of Chemistry and Physics" 71st Edition, 1990-1991. (19) "Capillary Column Use in Purge and Trap Gas Chromatography IT: Use of the Model 1000 Capillary Interface," Tek/Data B021684, Tekmar Company, Cincinnati, Ohio. (20) Kirshen, N., Am. Lab., 16 ,1984,Jb 60. (21) 'Terformance of a Third Generation Cryofocusing Trap for Purge and Trap Gas Chromatography'' Tekmar Company, Cincinnati, Ohio. 98