Citation
Investigation of ozone interference on the measurement of atmospheric carbonyl compounds

Material Information

Title:
Investigation of ozone interference on the measurement of atmospheric carbonyl compounds
Creator:
Shang, Xiaowei
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
75 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Carbonyl compounds -- Measurement ( lcsh )
Ozone ( lcsh )
Air -- Analysis ( lcsh )
Air -- Pollution -- Measurement ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 74-75).
Thesis:
Department of Chemistry
Statement of Responsibility:
by Xiaowei Shang.

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:
40337073 ( OCLC )
ocm40337073
Classification:
LD1190.L46 1998m .S53 ( lcc )

Downloads

This item has the following downloads:


Full Text
INVESTIGATION OF OZONE INTERFERENCE ON THE MEASUREMENT
OF ATMOSPHERIC CARBONYL COMPOUNDS
Xiaowei Shang
B.S., Peking University, 1987
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Chemistry
by
1997
/


This thesis for the Master of Science
degree by
Xiaowei Shang
has been approved
by
Date


Shang, Xiaowei (M.S., Chemistry)
Investigation of Ozone Interference On The Measurement Of Atmospheric
Carbonyl Compounds
Thesis directed by Professor Larry G. Anderson
ABSTRACT
Laboratory tests conducted by the U. S. Environmental Protection Agency and
others have suggested that ozone present in ambient air interferes with the
measurement of carbonyl compound concentration when using the 2,4-
dinitrophenylhydrazine-coated silica gel cartridge (SGC) technique.
The research was carried out in two parts. The first involved a field study in
which three automated sequential air samplers were used simultaneously to collect
ambient formaldehyde, acetaldehyde and acetone at Marine Street in Boulder. Two
samplers collected sample air after passing it through two different kinds of ozone
scrubbers. The third sampler collected ambient air samples directly. Samples analysis
was performed using HPLC with ultraviolet/visible detection. An ozone deunder
scrubber and KI cartridge scrubber were developed for this study and were tested for


their effectiveness for removal of ozone. Comparison of three sets of field data
showed that when ozone was removed by the ozone denuder scrubber or KI cartridge
scrubber, the carbonyl compound concentrations obtained by SGC technique were not
significantly different than those with no scrubber. This suggests that when the ozone
concentration is low, the ozone interference is small.
Studies of ozone interference and removal were conducted in the laboratory to
evaluate the DNPH cartridge technique. Different concentrations of ozone were
mixed with formaldehyde, generated by a permeation tube filled with
paraformaldehyde. These were sampled using the DNPH silica gel technique. The
results of laboratory tests show a severe negative bias when ozone is present, even at
low ozone concentration.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
IV


ACKNOWLEDGEMENTS
Special thanks to Professor Larry Anderson for his support and guidance
throughout this work. I also wish to express my gratitude to Professor John Lanning
for his advice and assistance, and to Professor Donald Zapien for his encouragement.
Thanks Susan Riggs and Michael Pribil for their help during this study.
/


CONTENTS
CHAPTER
1. INTRODUCTION.......................................................1
1.1 Environmental and Human Impact of Carbonyl Compounds...........1
1.2 Sources of Carbonyl Compounds..................................3
1.3 Measurement Techniques.........................................4
1.4 Derivatization of Carbonyls by 2,4-Dinitrophenylhydrazine......6
1.5 Ozone Interference.............................................7
1.6 Previous Research..............................................8
1.7 Approach in This Work.........................................11
2. EXPERIMENTAL SECTION..............................................13
2.1 Experiment for Field Study....................................13
2.1.1 Field Sampling Site........................................13
2.1.2 Pollutant Measurement......................................13
2.1.3 Sampling Equipment.........................................14
2.1.4 Ozone Denuder Scrubber.....................................16
2.1.5 IQ Cartridge Scrubber......................................17
vi


2.1.6 Sampling Procedures..........................................19
2.2 Laboratory Studies...............................................20
2.2.1 Collection Efficiency of Ozone Scrubbers.....................20
2.2.2 Lifetime Study of KI Cartridge Scrubber......................23
2.2.3 Interference Studies.........................................23
3. DATA ANALYSIS AND RESULTS............................................26
3.1 Field Study......................................................26
3.1.1 July 12-16 Period............................................26
3.1.2 July 16 September 2 Period.................................34
3.2 Results from laboratory interference studies.....................45
4. DISCUSSION...........................................................56
4.1 Field Study......................................................56
4.1.1 Sampler intercomparison period...............................56
4.1.2 Ozone Removal Evaluation.....................................57
4.2 Laboratory interference study....................................61
4.3 Comparison of field and laboratory study.........................66
5. CONCLUSIONS..........................................................75
REFERENCES..............................................................73
Vll


FIGURES
Figure
Figure 1.1 Reaction of carbonyls with DNPH to form 2,4-dinitrophenyl hydrazones. .6
Figure 2.1 Schematic of sequential sampler......................................15
Figure 2.2 Schematic of ozone denuder scrubber..................................17
Figure 2.3 Schematic of KI cartridge scrubber...................................18
Figure 2.4 Schematic diagram of the experimental set up of interference studies.25
Figure 3.1 Average concentration of carbonyl without ozone removal..............28
Figure 3.2 Correlation of formaldehyde, July 12-16..............................31
Figure 3.3 Correlation of acetaldehyde, July 12-16..............................31
Figure 3.4 Correlation of acetone, July 12-16...................................32
Figure 3.5 Ratio of scrubbed to unscrubbed carbonyl concentration versus time
period...............................................................33
Figure 3.6 Average concentration of formaldehyde (Julyl6-Sep2)..................35
Figure 3.7 Average concentration of acetaldehyde (Julyl6-Sep2)..................36
Figure 3.8 Average concentration of acetone (Julyl6-Sep2).......................37
Vlll


Figure 3.9 Correlation of formaldehyde, ozone denuder scrubbed to unscrubbed
concentration.........................................................39
Figure 3.10 Correlation of formaldehyde, KI cartridge scrubbed to unscrubbed
concentration.........................................................39
Figure 3.11 Correlation of acetaldehyde, ozone denuder scrubbed to unscrubbed
concentration.........................................................40
Figure 3.12 Correlation of acetaldehyde, KI cartridge scrubbed to unscrubbed
concentration.........................................................40
Figure 3.13 Correlation of acetone, ozone denuder scrubbed to unscrubbed
concentration.........................................................41
Figure 3.14 Correlation of acetone, KI cartridge scrubbed to unscrubbed
concentration.........................................................41
Figure 3.15 Ratio of scrubbed to unscrubbed formaldehyde concentration versus
ozone concentration...................................................42
Figure 3.16 Ratio of scrubbed to unscrubbed acetaldehyde concentration versus ozone
concentration.........................................................43
Figure 3.17 Ratio of scrubbed to unscrubbed acetone concentration versus ozone
concentration.........................................................44
Figure 3.18 Formaldehyde concentration from experiment set 1 as a function of ozone
added............................................................................47
IX


Figure 3.19 Acetaldehyde concentration from set 1 experiments as a function of
ozone added...................................................48
Figure 3.20 Formaldehyde concentration from set 2 experiments as a function of
ozone added...................................................49
Figure 3.21 Chromtogram of the sample cartridge with no added carbonyls after
exposure to 0 ppb ozone.......................................52
Figure 3.22 Chromatogram of the sample cartridge with no added carbonyls after
exposure to 100 ppb ozone.....................................53
Figure 3.23 Chromatogram of the sample with no added carbonyl after exposure to
190 ppb ozone.................................................54
Figure 3.24 Chromatogram of the sample with no added carbonyl after exposure to
320 ppb ozone.................................................55
Figure 4.1 Chromatogram of the sample with no added carbonyls after exposure
to 230 ppb ozone..............................................64
Figure 4.2 Chromatogram of the sample of the sample with formaldehyde after
exposure to 230 ppb ozone.....................................65
Figure 4.3 Chromatogram of the filed sample #B4L177, measured by sampler #4,
with denuder scrubber.........................................68
Figure 4.4 Chromatogram of the filed sample #B1L177, measured by sampler #1,
with KI cartridge scrubber....................................69
x


Figure 4.5 Chromatogram of the filed sample #B3L177, measured by sampler #3,
without ozone scrubber.......................................70
Figure 4.6 Chromatogram of the field sample #B1L161, measured by sampler# 1, with
a new KI cartridge scrubber..................................73
Figure 4.7 Chromatogram of the laboratory sample, with a KI cartridge scrubber
made a year ago..............................................74
xi


TABLES
Table
Table 2.1 Flow solenoid assignment..............................................14
Table 2.2 Collection efficiency of ozone denuder scrubber.......................21
Table 2.3 Collection efficiency of KI cartridge scrubber........................22
Table 3.1 Carbonyl concentration correlation....................................29
Table 3.2 Average concentrations of formaldehyde................................35
Table 3.3 Average concentration of acetaldehyde.................................36
Table 3.4 Average concentration of acetone......................................37
Table 3.5 Carbonyl concentration correlation during July 16- September 2 sampling
period...............................................................38
Table 3.6 Scrubbed / unscrubbed ratio versus ozone concentration................45
Table 3.7 Laboratory formaldehyde and acetaldehyde concentrations...............46
Table 3.8 Comparison of formaldehyde concentration from two set data............50
Table 3.9 Summary of the area changes with added ozone..........................51
xii


1. INTRODUCTION
1.1 Environmental and Human Impact of Carbonyl Compounds
Carbonyl compounds (aldehydes and ketones) are important intermediates in
combustion of hydrocarbon fuels, and are receiving increasing attention as pollutants
and as key participants in photochemical reactions which influence smog processes
in the atmosphere.
Because of their photolytic reactivities in the solar spectrum of the lower
troposphere, carbonyl compounds play a critically important role in atmospheric free-
radical production.
Formaldehyde photolysis is a major source of hydroperoxyl radicals during
the daylight hours via the reactions of H and HCO with 02:
HCHO + hv - H* + HCO*
H*+02->H02-
HCO + 02 H02* + CO
H202 is formed from the reaction of two hydroperoxyl free radicals:
H02* + H02* > H202 + 02
Photolysis of H202 produces OH radical directly:
H202 + hv -> 20H*
1


Hydroxyl radicals are also formed by the reaction below:
H02+N0->0H- +no2
These free radicals are responsible for the oxidation of hydrocarbons in the
troposphere[l][2].
Carbonyl compounds are also precursors of oxidants including ozone and
peroxyacyl nitrates.
N02+ hv O + NO
O + 02 ^ O3
CH3CHO + OH- -> CH3CO- + H20
ch3co- + o2 -> ch3coo2-
ch3coo2-+no2-> ch3coo2no2
Carbonyl compounds in the atmosphere are also of concern because they have
been associated with adverse biological effects. In particular, formaldehyde has been
associated with a number of health effects in humans. Formaldehyde contributes to
eye, nose and throat irritation. It can also cause bronchial asthma-like symptoms with
some reports of asthma attacks and allergic dermatitis [3]. Based on animal studies,
occupational studies, mutagenic tests and effects on DNA, there is evidence to
indicate that formaldehyde may be a carcinogen in humans [4].
2


1.2 Sources of Carbonyl Compounds
Carbonyl compounds are released into the troposphere from a variety of
biogenic and anthropogenic sources. These compounds are either derived from direct
emissions or secondary formation in the atmosphere.
Direct emissions include exhaust gases of motor vehicles and industrial
machinery in which hydrocarbon fuels are incompletely burned, and by vegetation.
There is potential for increased carbonyl emissions, especially formaldehyde and
acetaldehyde, due to changes in fuel technology such as in the use of methanol,
ethanol, etc., as gasoline additive substitutes.[5][6].
The secondary atmospheric formation of carbonyl compounds in the
troposphere is from the reaction of hydrocarbons with hydroxyl radical. The series of
reactions is initiated by the formation of a carbocentered radical (RCH2*), usually
through reaction of hydroxyl radical (OH*) with a hydrocarbon.
OH* + RCH3-> RCH2* + H20
Carbon centered radicals are very reactive and form peroxy radicals
(RCH202*), via addition of 02.
RCH2* + 02->RCH202*
Peroxy radicals can react with NO to form alkoxy radical (RCH20*).
RCH202* + NO -> RCH20* + N02
Alkoxy radicals can react with 02 to form a carbonyl and H02*
3


RCH20 + 02 -> RCHO + H02'
Ozonolysis of olefins is also a source of carbonyl compounds.
1.3 Measurement T echniques
An understanding and assessment of the role of carbonyls in tropospheric
chemistry requires the accurate and precise measurement of these compounds along
with their parent and product compounds. The concentrations of carbonyl compounds
range from sub-ppb or low-ppb in clean air to higher ppb in urban and polluted air.
Because of these trace concentrations, determination of carbonyl compounds in the
ambient atmosphere poses challenging problems.
Conventional methods of determination of aldehydes are based on
spectrophotometry, which are, in most cases, not sensitive enough for outdoor air
measurements. Spectrometric methods are usually based on functional group
detection. Four different spectroscopic techniques have been used:
(1) Differential Optical Absorption Spectroscopy (DOAS),
(2) Fourier Transform Infrared Absorption (FTIR),
(3) Laser-Induced Fluorescence Spectroscopy,
(4) Tunable Diode Laser Absorption Spectroscopy (TDLAS).
4


Although long-path-length optical techniques are sensitive and have been
used to measure formaldehyde in air at natural levels, their immobility and high cost
make them impractical for most field applications.
In the 1970s, chromatographic techniques, in conjunction with chemical
derivatization methods, paved the way for sensitive and selective determination of
carbonyls in ambient air. The most commonly used chromatographic method is based
on trapping carbonyl compounds by reaction with reagents such as 2,4-
dinitrophenylhydrazine (DNPH), followed by separation and analysis of the
derivatives by high-performance liquid chromatography (HPLC). Liquid impingers or
bubblers containing a DNPH solution are often employed to collect carbonyl
compounds from air. However, they are not well suited for field work. DNPH-coated
solid sorbents are a convenient alternative to impinger sampling and have been
increasingly used. Solid sorbents, including silica gel and C18, which have several
advantages including convenience of use, reproducibility and low blanks [7].
In spite of the widespread use at the current level of development, no
chromatographic method has yet been adapted for continuous sampling and analysis
of carbonyls in the ambient atmosphere. However, such real-time measurement
techniques have been developed for formaldehyde using specific fluorescence
detection [8].
5


1.4 Derivatization of Carbonyls by 2,4-Dinitrophenylhydrazine
Derivatization of DNPH followed by liquid chromatography and u.v.
detection is currently the most popular chromatographic technique used for
determination of formaldehyde, acetaldehyde and acetone at same time.
The acid catalyzed condensation reaction of carbonyl compounds with DNPH
is a well-known reaction for characterizing carbonyl compounds. The reaction
proceeds by nucleophilic addition to the carbonyl followed by 1,2-elimination of
water to form the 2,4-dinitrophenylhydrazone.
Rh
R27
Rh
R27
NO2
C=0 + H2N-NH
:^~^NQ2
OH
C-NH-NH
NO2
:-^r~V-NQ2
NO2
H+ Rl\
5: ^/ONH-NH-^ ^NQ2 + H2O
(1)
(2)
(3)
Figure 1.1 Reaction of carbonyls with DNPH to form 2,4-dinitrophenyl
hydrazones.
Under neutral or mildly acidic conditions, the equilibrium between 2 and 3 is
located more toward 2 due to hydration. However, under more strongly acidic
conditions, dehydration (2>3) is enhanced, and the equilibrium between 2 and 3 is
shifted toward 3.
6


After organic solvent extraction, the hydrazones are separated by high-
pressure liquid chromatography (HPLC) and quantitated with an ultraviolet detector
operated at 360 nm.
DNPH derivatization is used for several reasons:
(1) to allow better chromatography of polar compounds (reducing tailing and
nonlinearity of response),
(2) to improve the resolution of closely related compounds,
(3) to allow analysis of relatively nonvolatile components,
(4) to improve detectability (increasing detection response and selectivity for
specific detectors.
1.5 Ozone Interference
Ozone is one of the most abundant reactive gases in air, hence could
potentially cause sampling artifacts. The possible effect of ozone on DNPH-based
methods for carbonyls are threefold:
(1) formation of carbonyls as artifacts from reaction with sampling substrates,
(2) degradation of 2,4-Dinitrophenylhydrazones,
(3) formation of other interfering compounds. The reagent (DNPH) itself
reacts with ozone. For example, a DNPH solution rapidly became colorless
7


when high concentrations of ozone in air (0.1%) were passed through it, but
the reaction products were not identified [9]. Formation of formaldehyde from
the reaction of ozone with DNPH has not been studied, and cannot be ruled
out.
1.6 Previous Research
In a study by Amts and Tejada [7] the reaction of ozone with HCHO-
DNPHydrazone was identified as a potential problem when DNPH-coated silica
cartridges were used for formaldehyde sampling. In their study synthetic mixtures of
humidified air containing formaldehyde (20-140ppb) and ozone (0-770ppb) were
sampled. The loss of HCHO-DNPHydrazone increased markedly with increase in
ozone concentration, at 25 ppb HCHO and 120 ppb ozone, about 48% of HCHO
derivatives was lost. Also noticed on silica cartridges were concurrent large losses of
DNPH. It was concluded that the silica cartridge exhibited such large reductions in
formaldehyde response because the DNPH derivative, which is largely formed at the
front of the cartridge and immobilized, was being destroyed by ozone. In the case of
impingers, the HCHO-DNPHydrazone is protected by the DNPH, which is always
present in excess and well dispersed. However, a matter of concern in the impinger
technique is that the products formed from the DNPH-ozone reaction can interfere
with resolution of formaldehyde peak in HPLC separation. Recently, this problem
8


was addressed by Smith et al.[10] who used a ternary gradient separation to obtain the
desired chromatographic resolution. In contrast to the silica cartridges, Clg cartridges
exhibited no loss of the HCHO derivative up to 120 ppb ozone [9]. In this case, it was
reasoned that CI8 substrate itself was the site of ozone reaction, thereby preventing
attack on formaldehyde-DNPHydrazone.
The mechanisms of ozone-initiated reactions in the above cases are not clearly
understood. Atkinson and Carter [11] suggested that a chain of free-radical reactions
can be initiated when ozone reacts with hydrazines, either by addition to a nitrogen, or
abstraction of a hydrogen from a weak N-H bond. Amts and Tejada [7] pointed out
that under the acidic conditions of the DNPH reaction, ozone addition to the
protonated nitrogen is restricted and hydrogen abstraction could be the preferred
route. The following pathway has been proposed by Atkinson and Carter for
hydrogen abstraction:
RHNNH2 + 03 -* RNNH2' (or RHNNH) + 02 + OH*
RNNH2 (or RHNNH) + 02 - RH=NH + H02'
RH=NH + 03 (or OH') -> RN=N' + OH* + 02 ( + H20)
RN=N' -> R* +N2
When DNPH-coated C18 substrate is used for sampling, the radicals generated
by ozone attack can be scavenged by the CI8, thus limiting further attack on 2,4-
DNPH or the hydrazones, so the above reactions wouldnt propagate and consume
9


2,4-DNPH and hydrazone. But ozone can react with the C18 to form high molecular
weight carbonyls. These can react with 2,4-DNPH to form artifact hydrazone
derivatives [12].
From the above discussion, it is clear that ozone may be a serious interferent
in methods for sampling carbonyls in ambient air using either C18 or silica-gel
cartridges coated with reagent. Amts and Tejada [7] reported that they obtained
encouraging results in a preliminary study using a potassium iodide-coated copper
tubing inlet to remove ozone prior to collection with a DNPH-coated silica-gel
cartridge. A well known mechanism for the reaction of ozone with KI involves
participation of water and proceeds with the formation of iodine:
03 + 2KI + H20 ^ 02 +12 +2KOH
Hoigne et al. [13] suggested an alternate mechanism involving the formation
of hypoiodite, which does not require water:
03 + KI - 02 + KOI
In the atmosphere, the reaction probably occurs through the iodine pathway
due to air humidity. Grosjean and Parmer [14] and Williams and Grosjean [15]
reported that an annular denuder coated with KI completely removed ozone even at
high flow rates, up to 20-toiin'1 and formaldehyde was not removed by this device.
Recently, Slemr [16] used Teflon tubing packed with crystalline KI to remove ozone
in conjunction with silica-gel cartridge sampling. In laboratory tests, carbonyls were
10


not found to be adsorbed by the KI filter at a relative humidity of 30-35%. However,
the performance of KI scrubbers in moist air remains unclear. As iodide salts are very
hygroscopic, the physical integrity of solid iodide may be affected in moist air. This
may be a problem in the cartridge-based sampling technique for carbonyls. Some
metal oxides, including Mn02 and CuO decompose ozone catalytically. Studies with
gas-phase standards indicate that recovery of carbonyls are not affected by passing
through a CuO cartridge. Gas-phase titration of ozone with nitric oxide (NO), as used
by Tanner et al. [17] in the determination of hydrogen peroxide in the ambient
atmosphere, is also a potential technique to eliminate any ozone interference.
1.7 Approach in This Work
In this study, carbonyl compounds were measured by DNPH silica gel
cartridge technique. The sampled gas streams were either ambient air or synthetic
mixture of carbonyl compounds and ozone.
During the field study, ambient formaldehyde, acetaldehyde and acetone were
measured from July 12 to September 3,1996. An ozone denuder scrubber and KI
cartridge scrubber for removing ozone were evaluated, and ozone unscrubbed ambient
air was also sampled at the same time. Three automated sequential air samplers were
used simultaneously to collect carbonyls at Marine Street in Boulder. Two samplers
collected air after it had passed through the two different kinds of ozone scrubbers.
11


Another sampler drew ambient air directly through the sampling cartridge without
ozone removal. Sample analysis is accomplished using HPLC with ultraviolet/visible
detection.
Ozone deunder scrubber and KI cartridge scrubber were developed for this
study and were tested for effectiveness of removal of ozone.
Ozone interference and removal studies were conducted in the laboratory to
evaluate DNPH silica gel technique. Ozone was mixed with the relative stable
formaldehyde source, which is generated by a permeation device.
The objective of this study was to investigate ozone interference on carbonyl
sampling by 2,4-dinitrophenylhydrazine-HPLC (DNPH-HPLC) method, and use a
field study and laboratory data to evaluate the effect of ozone removal techniques.
12


2. EXPERIMENTAL SECTION
2.1 Experiment for Field Study
2.1.1 Field Sampling Site
Ambient formaldehyde, acetaldehyde and acetone concentrations were
measured in the sampling station, which is located at Marine Street (2320 Marine
Street) in Boulder. It is operated by the Colorado Department of Public Health and
Environment. This sampling site is near downtown Boulder in a residential area near
a well traveled two-lane street.
2.1.2 Pollutant Measurement
Formaldehyde, acetaldehyde and acetone concentrations were measured from
July 12 to September 2, 1996. Data were collected 24 hours per day, seven days per
week. Ozone concentrations were measured simultaneously at the same location using
a Dasibi Model 1003-AH ultraviolet photometric analyzer.
On the first four days (July 12-July 16), the measurements were performed
without adding ozone scrubbers. Four cartridges were used to cover the 4 four-hour
periods, i.e., 4am-8am, 8am-noon, noon-4pm, and 4pm-8am. The fifth cartridge was
13


used to cover 8pm-4am which is an eight-hour period. The flow solenoid assignment
is shown in table 2.1.
Table 2.1 Flow solenoid assignment
Flow Solenoid number Sampling Period
Day 1 0 4am-8am (4-hour)
1 8am-noon (4-hour)
2 noon-4pm (4-hour)
3 4pm-8pm (4-hour)
4 8pm-4am (8-hour)
Day 2 5 4am-8am (4-hour)
6 8am-noon (4-hour)
7 noon-4pm (4-hour)
8 4pm-8pm (4-hour)
9 8pm-4am (8-hour)
2.1.3 Sampling Equipment
Sampling is performed through a sorbent cartridge, by controlling and
measuring the flow through DNPH coated cartridge for a specified time period. A
schematic of the sampler is shown in figure 2.1.
14


pump
Figure 2.1 Schematic of sequential sampler
This is an automated low volume sequential sampler. It can hold tencartridges.
Five cartridges were used to cover the four 4-hour and one 8-hour daily sampling
period for one day and a second five cartridges were used for the second day. As can
be seen in Figure 2.1, the sampler consists of a computerized control system that is
capable of controlling solenoid opening and closing at specific times. An eleventh
cartridge was used as a field blank for the two day sampling period. Eleven cartridges
15


were changed every other day. A diaphragm oil-less vacuum pump was used in the
sampler to draw ambient air through the ozone scrubber and DNPH-coated cartridges.
The total volume of air sampled through each solenoid was measured by a mass flow
controller and the result was saved on the computer. The data was transfered from
computer to field data sheet while changing the sample cartridges.
2.1.4 Ozone Denuder Scrubber
As can be seen from figure 2.2, the ozone denuder scrubber consists of a 1/4-
inch o.d. copper tubing that has been coated internally with a saturated solution of
potassium iodide (KI). The tube is coiled to a diameter of approximately two inches.
The entrance and exit of the coil were fitted with a 1/4-inch brass bulkhead union and
housed in a temperature controlled aluminum chamber. A cord heater was wrapped
around the outside spiral of the coil. The ozone denuder scrubber was maintained at
52 C during sample collection. Heating prevents condensation from occurring in the
tube during sampling. The scrubber is connected to the inlet of the sample collection
system. Sample air as extracted from a sample probe and distribution manifold and
pulled through the scrubber by an oil-less vacuum pump. \i of ozone in the sample
air is converted by the chemical reaction below:
03 + 2KI + H20 - 02 +12 +2KOH
16


The ozone denuder scrubber is reusable. The copper tube should be recoated with a
saturated solution of KI after each six months of use [18 ].
Figure 2.2 Schematic of ozone denuder scrubber
2.1.5 KI Cartridge Scrubber
The ozone cartridge scrubber is a identical in size and shape to the regular
DNPH Silica gel sampling cartridge, filled with approximately 1 gram of the KI
coated silica. The scrubber is positioned at the inlet of the sample collection system.
17


Sample air is extracted from the sample probe and distribution manifold and pulled
through the ozone scrubber by an oil-less vacuum pump. Ozone in the sample air is
converted by the chemical reaction described above.
KI scrubber preparation:
(1) Saturated solution of KI was added to silica gel (70-230 mesh)
(2) Mix well and then vacuum filter the mixture
(3) The solid was then dried in a low temperature oven (70 C)
(4) Approximately lg of the KI coated silica is packed in a sampling
cartridge.
Figure 2.3 Schematic of KI cartridge scrubber
18


2.1.6 Sampling Procedures
The DNPH silica gel cartridges were prepared according to the procedure
similar to that described by Tejada [19]. The dimensions of the body of the cartridges
were about 1.25cm o.d. x 10cm. Each cartridge was packed with approximately 0.6-
0.8 grams silica gel (Chromatographic Specialties, 70-230 mesh). The silica gel
packed cartridges were coated with acidified DNPH coating solution, which was
made by dissolving twice recrystallized 2,4-DNPH in acetonitrile and acidified with
HC1. In the end, the DNPH coated cartridges were dried in a low temperature oven
over night. When not being used for sampling, these cartridges were covered with
plastic caps to prevent interaction with ambient air and kept in a freezer.
Following each sampling period, sampled cartridges were kept in freezer prior
to analysis. The sampled cartridges were extracted with 5ml of acetonitrile. The
eluent was collected in a 5-ml volumetric flask. An HPLC (Varian 5050) was used for
analysis. The DNPH derivatives were effectively separated using a 4.6mm I.D. x
150mm, reverse phase Clg column and using acetonitrile/water as mobile phase. The
eluting components were detected with a UV absorbance detector operating at 360nm.
19


2.2 Laboratory Studies
2.2.1 Collection Efficiency of Ozone Scrubbers
Scrubber collection efficiency for ozone was determined by drawing known
concentration of ozone generated from TECO Model 49 Ozone analyzer/calibrator
through scrubber, and followed by measuring of ozone concentrations with Dasibi
Model 1003-AH ultraviolet photometric analyzer. The background of ozone was
0.035ppm, and the zero offset of the ozone monitor was O.OlOppm.
Collection efficiency was calculated as:
Ci C2
T|%= X 100
Ci
C, is ozone concentration before scrubber
C2 is ozone concentration after scrubber
The collection efficiency was 99.4% 0.6% for the ozone denuder scrubber
and 99.8% 0.2% the KI cartridge scrubber, as shown in Tables 2.2 and 2.3.
20


Table 2.2 Collection efficiency of ozone denuder scrubber
[03 ] Before Scrubber (ppm) [03 ] After Scrubber(ppm) CE
measured[03] corrected[03] measured [03] corrected [03] (%)
0.035 0.000 0.012 0.002 N/A
0.064 0.0029 0.010 0.000 100.0
0.089 0.0054 0.010 0.000 100.0
0.128 0.093 0.011 0.001 98.9
0.147 0.112 0.011 0.001 99.1
0.183 0.148 0.013 0.003 98.0
0.255 0.220 0.011 0.001 99.5
0.300 0.265 0.010 0.000 100
0.335 0.300 0.012 0.002 99.3
0.382 0.347 0.012 0.002 99.4
0.435 0.400 0.013 0.003 99.3
0.494 0.459 0.013 0.003 99.3
0.551 0.516 0.012 0.002 99.6
Average Standard Deviation 99.4 0.6
21


Table 2.3 Collection efficiency of KI cartridge scrubber
[03 ] Before Scrubber (ppm) [03 ] After Scrubber(ppm) CE
measured[03] correeted[03] measured [03] corrected [03] (%)
0.035 0.000 0.010 0.000 100.0
0.083 0.048 0.010 0.000 100.0
0.143 0.108 0.008 -0.002 N/A
0.188 0.153 0.010 0.000 100.0
0.230 0.195 0.010 0.000 100.0
0.278 0.243 0.011 0.001 99.6
0.325 0.290 0.010 0.000 100.0
0.377 0.342 0.011 0.001 99.7
0.424 0.389 0.012 0.002 99.5
0.475 0.440 0.011 0.001 99.8
0.559 0.524 0.011 0.001 99.8
Average Standard Deviation 99.8 0.2
22


2.2.2 Lifetime Study of KI Cartridge Scrubber
The lifetime of the KI cartridge scrubber was also investigated in this study.
The experiment was carried out by drawing 230ppb of ozone through the KI cartridge
scrubber. The ozone concentration was monitored by using Dasibi Model 1003-AH
ultraviolet photometric analyzer and a chart recorder. After 48 hours, the measured
ozone remained zero. This indicates that KI cartridge scrubber still can remove ozone
effectively. Taking an average atmospheric ozone concentration as 40ppb, a KI
cartridge will last at least 11 days (= 230 / 40 x 2).
2.2.3 Interference Studies
Experiments were conducted to investigate the potential for ozone to interfere
when sampling carbonyl compounds using DNPH-coated silica gel cartridges. Ozone
was mixed with gas-phase formaldehyde and acetaldehyde.
Ozone was generated by flowing air through a TECO Model 49 ozone
analyzer/calibrator. The ozone was measured with a Dasibi Model 1003 ozone
analyzer. During the entire experiment, the ozone concentration was monitored at a
manifold continuously.
The gas phase formaldehyde or acetaldehyde was produced using a
permeation device, which consisted of a porous Teflon-walled tube filled with
paraformaldehyde or acetaldehyde. Zero air flowed continuously through the
23


permeation tube generated constant concentration of formaldehyde or acetaldehyde.
Zero air used in this experiment is clean air, it does not contain trace impurities that
could contaminate the experiment. The permeation device was immersed into a
temperature controlled water bath. The temperature was remained at 40C.
Formaldehyde or acetaldehyde gas stream generated by the permeation tube was
diluted with zero air. The concentration of formaldehyde or acetaldehyde was
determined by permeation rate of the tubes at the bath temperature, the flow rate of
gas past the permeation tube, and the flow rate of the diluent zero air. The
experiments were performed with formaldehyde at concentrations typical of relatively
clean air, that is, about 8 ppbv.
Sampling was performed in three sets of experiments. The first set of the
experiments sampled the mixture of formaldehyde, acetaldehyde and ozone. The
second set of experiments sampled the mixture of formaldehyde and ozone. The last
set of experiments sampled ozone only to see how it affected DNPH cartridges. For
each set of these experiments sampling was performed with increments of ozone
concentrations of 0, 20, 50, 100,140,190, 230, and 320 ppb.
Figure 2.4 shows the experimental set up for the interference study.
24


exhaust
t
permeation tube
Figure 2.4 Schematic diagram of the experimental set up of interference studies.
25


3. DATA ANALYSIS AND RESULTS
3.1 Field Study
DNPH silica gel cartridge technique was evaluated under ambient conditions
during the summer sampling period, July 12 through September 2,1996. During this
study, three carbonyl samplers were used, one without any ozone removal device, one
with a ozone denuder scrubber and the third with a KI cartridge scrubber, as shown
below. Ambient carbonyl compounds with and without ozone removal were collected
during the same time periods at Boulder.
July 12-16 July 16-Sep 2
Sampler #1 without scrubber add KI cartridge scrubber
Sampler #3 without scrubber without scrubber
Sampler #4 without scrubber add denuder scrubber
3.1.1 July 12-16 Period
On the first four days, July 12-16, 1996 sampling period, no ozone scrubber
was used. Three samplers were compared during this period. Formaldehyde
concentration measured by sampler #4, which later used an ozone denuder scrubber,
26


range from 0.78 to 8.08 ppbv. Acetaldehyde concentrations ranged from 1.94 to 30.86
ppbv. Acetone concentrations range from 7.04 to 86.57 ppbv. For sampler #1, which
later used a KI cartridge scrubber, formaldehyde concentrations ranged from 0.81 to
4.10 ppbv, acetaldehyde concentrations ranged from 1.38 tol0.21 ppbv, acetone
concentrations range from 2.83 to 12.19. For sampler #3, which never used an ozone
scrubber throughout the sampling period, formaldehyde concentrations ranged from
0.93 to 6.75 ppbv, acetaldehyde concentrations ranged from 1.89 to 17.94 ppbv,
acetone concentrations ranged from 3.97 to 27.18 ppbv. Figure 3.1 shows the sampler
intercomparison result.
27


Average of Formaldehyde Concentration
c
o
c
o
o
0
o
>1
£
0
2

E
o
Time Period
Sampler#4(add denuder later)
Sampler#1 (add KI cartridge later)
Sampler#3(without scrubber)
Average of Acetone Concentration
July12-16
-m- Sampler #4 (add denuder later)
-v- Sampler #1 (add KI cartridge later)
Sampler #3 (without scrubber)
Figure 3.1 Average concentration of carbonyl without ozone removal.
28


In order to find the relationship between the carbonyl concentrations measured
by sampler #4 (with ozone denuder later), and carbonyl concentrations measured by
sampler #3 (without ozone scrubber later), regression analyses were performed to
compare the results of carbonyl measurement data with different samplers. The
sampler #1 (with KI cartridge later) was also compared with sampler #3.
The regression results are summarized in table 3.1.
Table 3.1 Carbonyl concentration correlation
Sampler#4 / Sampler#3 Julyl2-16
Slope RA2 Intercept
Formaldehyde 1.170.34 0.74 0.23 1.13
Acetaldehyde 1.54 0.41 0.78 0.603.01
Acetone 2.820.77 0.77 -4.428.65
Sampler#! / Sampler#3, July 12-16
Slope RA2 Intercept
Formaldehyde 0.48 0.25 0.51 0.990.8 7
Acetaldehyde 0.740.10 0.94 0.600.7 7
Acetone 0.420.12 0.77 1.841.4 2
29


Figure 3.2-4 shows the correlation of carbonyl concentrations measured from
the different samplers.
Correlation of Formaldehyde
July12-July16
Sampler#3 (unscrubbed)
Correlation of Formaldehyde
Julyl 2-16
i
0
Q.
E
CO
w
Sampler#3 (unscrubbed)
30


Figure 3.2 Correlation of formaldehyde, July 12-16
Correlation of Acetaldehyde
July12-July16
Sampler#3 (unscrubbed)
Correlation of Acetaldehyde July12-July16 ?n
^ 15 =6 10 E to cc 5 0 C 5

B


M X.
w

£ Samp 1 )ler#3 (i 3 15 2 mscrubbed) 3 2
Figure 3.3 Correlation of acetaldehyde, July 12-16.
31


Correlation of Acetone
July12-16
Sampler#3 (unscrubbed)
25 , Correlation of Acetone Julyl 2-16
20 1 Cl) lo Cl E 10 CD w 5. 0 J C 3




t

a kn
sri- d

San 1 npler#' 0 1 S (unsc 5 2 :rubbe< D 2 i) 5 3
Figure 3.4 Correlation of acetone, July 12-16.
. Figure 3.5 shows the ratio of the scrubbed carbonyl concentration to the
unscrubbed concentration versus time.
32


Scrubbed/unscrubbed formaldehyde
July 12-16
KI cartridge scrubbed
Scrubbed/unscrubbed acetaldehyde
July12-16
* denuder scrubbed
KI cartridge scrubbed
Scrubbed/unscrubbed acetone
July 12-16
KI cartridge scrubbed
Figure 3.5 Ratio of scrubbed to unscrubbed carbonyl concentration versus
time period.
33


3.1.2 July 16 September 2 Period
During July 16 September 2 period, ozone denuder scrubber and KI cartridge
scrubber were evaluated. Ozone denuder scrubbed formaldehyde concentrations
ranged from 0.77 to 9.39 ppbv. Acetaldehyde concentrations ranged from 1.12 to
27.71 ppbv. Acetone concentrations ranged from 2.22 to 41.35 ppbv. KI cartridge
scrubbed formaldehyde concentrations ranged from 0.54 to 5.48 ppbv. Acetaldehyde
concentrations ranged from 0.87 to 14.38 ppbv. Acetone concentrations range from
1.23 to 20.94 ppbv. Unscrubbed formaldehyde concentrations ranged from 0.84 to
7.23 ppbv. Acetaldehyde concentrations ranged from 0.66 to 21.74 ppbv. Acetone
concentrations ranged from 0.08 to 39.20 ppbv.
Table 3.2, table 3.3 and table 3.4 show the averages of carbonyl
concentrations for the different time periods of the day for the July 16 September 2,
1996 sampling period. Figure 3.6, 3.7, and 3.8 show the diurnal character for three
carbonyl concentrations for the three samplers. Data were collected from five time
periods per day, starting at 4am 8am, followed by 8 am noon, noon 4pm, 4pm -
8pm, 8pm 4am. Formaldehyde, acetaldehyde and acetone concentration reach a
maximum during noon 4pm time period.
34


Table 3.2 Average concentrations of formaldehyde
TIME 04-08 08-12 12-16 16-20 20-04
Denuder 2.15 4.54 6.13 4.67 2.99
KI 1.43 3.03 3.69 3.09 3.35
Unscrubbed 2.17 4.02 4.89 3.89 2.16
Average of Formaldehyde Concentration
c
o
c
o
o
0
-a
>*
o
2
0
E
o
U-
Time Period
Sampler#4(with denuder scrubber)
Sampler#1 (with KI cartridge scrubber)
Sampler#3(without scrubber)
Figure 3.6 Average concentration of formaldehyde (Julyl6-Sep2)
35


Table 3.3 Average concentration of acetaldehyde
TIME 04-08 08-12 12-16 16-20 20-04
Denuder 3.34 8.61 15.26 9.14 3.93
KI 2.85 5.09 10.01 7.05 4.00
unscrubbered 2.13 6.49 11.79 7.11 2.24
Average of Acetaldehyde Concentration
Julyl 6-end
* Sampler#4(with denuder scrubber)
Sampler#1 (with KI cartridge scrubber)
Sampler#3(without scrubber)
Figure 3.7 Average concentration of acetaldehyde (Julyl6-Sep2)
36


Table 3.4 Average concentration of acetone
TIME 04-08 08-12 12-16 16-20 20-04
Denuder 7.02 11.56 22.23 13.52 7.08
KI 5.23 5.89 10.86 9.91 5.91
unscrubbed 4.08 8.72 16.62 10.24 4.04
Average of Acetone Concentration
-m- Sampler#4(with denuder scrubber)
Sampler#1 (with KI cartridge scrubber)
Sampler#3(without scrubber)
Figure 3.8 Average concentration of acetone (Julyl6-Sep2)
In order to detect if there was a significant difference between the scrubbed
cartridge concentrations and unscrubbed cartridge concentrations, scrubbed versus
37


unscrubbed concentration regression analyses were performed to compare the results
of carbonyl measurement data for the three different samplers.
Table 3.5 shows carbonyl compounds concentration correlation results.
Table 3.5 Carbonyl concentration correlation during July 16- September 2
sampling period.
Denuder scrubbed /unscrubbed, Julyl6-Sep 2
Slope RA2 Intercept
Formaldehyde 1.120.11 0.71 0.440.41
Acetaldehyde 1.27 0.13 0.69 1.01 0.95
Acetone 1.150.20 0.45 4.182.01
KI Cartridge scrubbed / unscrubbed, July 16-Sep 2
Slope RA2 Intercept
Formaldehyde 0.390.11 0.23 1.46 0.4 1
Acetaldehyde 0.73 0.05 0.83 1.320.3 8
Acetone 0.480.07 0.54 3.330.7 8
Figure 3.9-3.14 show the correlation of carbonyl concentrations measured
from the sampler with ozone denuder or KI cartridge scrubber to the sampler
without scrubber. These measurements were made from July 16 to September 2.
38


Figure 3.9 Correlation of formaldehyde, ozone denuder scrubbed to
unscrubbed concentration
Figure 3.10 Correlation of formaldehyde, KI cartridge scrubbed to
unscrubbed concentration
39


Figure 3.11 Correlation of acetaldehyde, ozone denuder scrubbed to
unscrubbed concentration.
Correlation of Acetaldehyde
Julyl 6-end
Figure 3.12 Correlation of acetaldehyde, KI cartridge scrubbed to unscrubbed
concentration.
40


Correlation of Acetone Julyl 6-end Denuder/Unscrubbed mn
T3 80 . D

jQ h 60 .
< X >
CO m 40 .

TD c 20 . k .-s. X 's X
X < £ % -z x
0 Q o: X 7T .-'x X i* X x. X
xxx X x
c c 1 U 0 nsc 1 rub! 5 2 3ed 0 2 5 3
Figure 3.13 Correlation of acetone, ozone denuder scrubbed to unscrubbed
concentration.
Correlation of Acetone
Julyl 6-end KI Cartridge/Unscrubbed
Figure 3.14 Correlation of acetone, KI cartridge scrubbed to unscrubbed
concentration.
41


To evaluate the ozone effect on formaldehyde, acetaldehyde and acetone
measurement, a plot of carbonyl ratios for scrubbed / unscrubbed versus ozone
concentration are shown on figure 3.15-17.
Correlation of HCHO Ratio to Ozone
Denuder Scrubbed/Unscrubbed
"D
0
_Q
JQ
k_
O
(/)
c
0
.o
_Q
D
O
CO
10
8
6
4
2







<- +
VjtT iVrr i- -r
- Tjt- ' ;r ft
0 20 40 60
Ozone Concentration (ppb)
80
Correlation of HCHO Ratio to Ozone
KI Cartridge Scrubbed/Unscrubbed
-o
0
£>
D
o
(0
c
3
o
0
_Q
XI
L
o
CO
10
8
6
4
2
0
0 20 40 60 80
Ozone Concentrion (ppb)








.
r~f^| ^ -r-r
Figure 3.15 Ratio of scrubbed to unscrubbed formaldehyde concentration
versus ozone concentration.
42


Correlation of CH3CHO Ratio to Ozone
Denuder Scrubbed/Unscrubbed
O
0
XJ
.Q
3
i
CJ
w
c
D
T3
0
JD
XI
3
La
o
co
10
8
6
4
2
0
0 20 40 60 80
Ozone Concentration (ppb)






-1. -
f*-
L r-
+ .... J
* T"'
Correlation of CH3CHO Ratio to Ozone
KI Cartridge Scrubbed/Unscrubbed
TJ
0
X
X2
3
i_
O

c
"3
0
XI
xi
3
i_
O
CO
10
8
6
4
2
0
0 20 40 60 80
Ozone concentration (ppb)




-4-


-i- J- .+. ' -4- .*.+
'S- K j.
t i- <*114 -trr T--r
Figure 3.16 Ratio of scrubbed to unscrubbed acetaldehyde concentration
versus ozone concentration.
43


Correlation of Acetone Ratio to Ozone Denuder Scrubbed/Unscrubbed m
T3 0 9 8 . t- 3

D O c c/) b .
4-
c D 4 . 4-

0 .Q § 2 . 4- 4-4- 4-' *-
i. 4 --
b w n . 4-<- ~r- r ; v ~i+-
-r- + ,_i 4- 4
0 Oz 2 one C 3 4 3oncentra1 3 6 ion (ppb) 3 8
Correlation of Acetone ratio to Ozone
KI Cartridge Scrubbed/Unscrubbed
T3
0
JQ
X)
D
L_
o

C
~o
0
-Q
.Q
0
CO
Ozone Concentration (ppb)
Figure 3.17 Ratio of scrubbed to unscrubbed acetone concentration versus
ozone concentration.
The correlation results are shown in table 3.6
44


Table 3.6 Scrubbed / unscrubbed ratio versus ozone concentration.
Denuder scrubbed/unscrubbed ratio vs. ozone
Slope RA2 Intercept
Formaldehyde 0.0043 0.0016 0.0355 1.1 0.3
Acetaldehyde -0.0067 0.0038 0.0177 1.80.7
Acetone -0.01560.0490 0.0006 3.3 9.1
KI Cartridge scrubbed/unscrubbed ratio vs. ozone
Slope RA2 Intercept
Formaldehyde 0.00040.0016 0.0004 0.80.3
Acetaldehyde -0.00880.0028 0.0523 1.50.6
Acetone -0.01900.0232 0.0040 2.24.5
3.2 Results from laboratory interference studies
In our study, a formaldehyde and acetaldehyde generating system was
constructed to deliver formaldehyde and acetaldehyde continuously. Five replicate
experiments were conducted to determine the concentrations of formaldehyde and
acetaldehyde delivered from the system. In these experiments, a DNPH silica gel
cartridge was used to collect carbonyl compounds for 180 min. The volume of air
45


sampled for each cartridge was measured by a dry test meter. The replicate
experimental results are given in table 3.7.
Table 3.7 Laboratory formaldehyde and acetaldehyde concentrations.
Formaldehyde, ppb Acetaldehyde, ppb
Sample 1 8.10 48.38
Sample 2 7.42 47.19
Sample 3 8.49 48.20
Sample 4 7.87 47.95
Mean SD 7.970.44 47.930.52
In order to investigate ozone interference on carbonyl measurements, various
concentrations of ozone were mixed with the gas-phase formaldehyde and
acetaldehyde. Three sets of experiments was carried out. The first set of experiments
sampled the mixture of formaldehyde, acetaldehyde and ozone. For each experiment,
sampling was performed with increments of ozone of 0, 20, 50, 100, 140, 190, 230,
320 ppb.
Figure 3.18 shows the formaldehyde concentration measured in the presence
of additional ozone and that measured after going through ozone denuder or KI
cartridge scrubber as a function of the amount of ozone added. Each data point
46


represents the formaldehyde concentration with either ozone removal device or
without ozone removal.
c
o
*
ro
L_
*
c
CD
O
c
o
O
0
"O
>.
-C
0
p
0
E
o
Formaldehyde Vs. Ozone
Data Set 1
10
8
6
4
2
0
0 50 100 150 200 250 300 350
Ozone Concentration
J
1 4 , 1-.
i Jj 4 1 1 kr fc- " 4. fe
4- 4
^ unscrubbed de-scrubbed
Kl-scrubbed
Figure 3.18 Formaldehyde concentration from experiment set 1 as a function of
ozone added.
Figure 3.19 is a plot of acetaldehyde measured in the presence of additional
ozone, and acetaldehyde measured following an ozone scrubber that removes ozone
as a function of the amount of ozone added. In the presence of 20 ppb ozone, the
chromatograms of ozone unscrubbed samples start to show a little peak with retention
times shorter than formaldehyde. The peak increased in proportion to increasing
ozone concentration. The measured unscrubbed formaldehyde and acetaldehyde
concentrations in the presence of 20ppb of ozone was 7.76 ppb and 48.24ppb,
indicating no carbonyl loss. At 50 ppb ozone between 18%-27% of formaldehyde and
47


19%-23% of acetaldehyde were not measured. At 320 ppb ozone between 86%-88%
of formaldehyde and 83%-84% of acetaldehyde were not measured.
c
o
Acetaldehyde Vs. Ozone
Data Set 1
Ozone Concentration
unscrubbed de-scrubbed
Kl-scrubbed
Figure 3.19 Acetaldehyde concentration from set 1 experiments as a function of
ozone added.
In the laboratory study, the concentration of acetaldehyde in the sample air
was about six times higher concentration than formaldehyde. So ozone may react with
acetaldehyde or its derivatives at higher rate than equal concentration of
formaldehyde and acetaldehyde. This would inhibit the reaction of ozone with
formaldehyde or its derivatives, and make it hard to identify the interference of ozone
on measurement of formaldehyde. In order to detect if there was a different ozone
effect on measurement of formaldehyde due to high quantity of acetaldehyde present,
48


the second set of experiments was conducted. In this set of experiments, only gas
phase formaldehyde and ozone was added in the sample air.
Figure 3.20 is a plot made for this second set of experimental data. It shows
the formaldehyde concentration measured in presence of additional ozone, and
formaldehyde measured after going through an ozone scrubber as a function of ozone
added.
c
o
+-
2
c
CD
O
c
o
O
CD
"O
>.
-C
CD
2
co
E
o
LL
Formaldehyde Vs. Ozone
Data Set 2
^ unscrubbed ^ de-scrubbed
Kl-scrubbed
Figure 3.20 Formaldehyde concentration from set 2 experiments as a function of
ozone added.
From second set of the experiments, the measured unscrubbed formaldehyde
concentrations in the presence of 20 ppb was 8.82 ppb, indicating no formaldehyde
loss. At 50 ppb ozone between 9%-l 8% of formaldehyde is not measured. At 320 ppb
ozone 99% of formaldehyde was not measured.
49


In order to find the formaldehyde differences between two sets of
experiments, regression analyses were made to compare the results. The regression
results are summarized in table 3.8.
Table 3.8 Comparison of formaldehyde concentration from two set data.
Denuder scrubbed formaldehyde
Slope RA2 Intercept
Data set 1 0.0020.002 0.19 7.30.6
Data set 2 0.001 0.002 0.08 8.10.5
KI cartridge scrubbed formaldehyde
Slope RA2 Intercept
Data set 1 -0.0020.001 0.27 8.80.6
Data set 2 0.002 0.003 0.07 8.50.9
Unscrubbed formaldehyde
Data set 1 -0.0200.002 0.97 7.40.5
Data set 2 -0.0240.002 0.96 8.30.6
In order to investigate the products of ozone reaction with DNPH silica gel
cartridge, a third set of experiments was performed. In this set of experiments, DNPH
silica gel cartridges were exposed to different ozone concentrations. Shown in Figure
3.20 23 are the chromatograms of the 2,4-DNPH cartridge exposed to various ozone
50


concentrations when carbonyl compounds are absent. The ozone concentration in
Figure 3.20 through 3.23 is 0,100, 190, and 320 ppb, respectively. It can be seen that
DNPH was destroyed by ozone, and several additional peaks (#1-4) were found
compared with Figure 3.22 ([03] = 0 ppb). The results are summarized in Table 3.9,
where the integrated areas of the four peaks are tabulated at various ozone
concentrations. It clearly shows that the area count of the additional peaks increases
with increasing ozone concentration.
Table 3.9 Summary of the area changes with added ozone.
Integrated Area
[Ozone] ppb Peak 1 pre-HCHO Peak 2 HCHO Peak 3 pre-CH3COCH3 Peak 4 x-DNPH*
0 7.75* 55.09 26.70 0
20 10.89 89.68 73.46 13.47
50 46.99 89.31 155.55 75.37
100 80.83 102.66 209.01 89.10
140 105.03 86.35 389.64 147.57
190 133.96 92.70 361.92 185.34
230 154.96 74.55 423.05 155.18
320 182.49 62.70 287.01 165.32
* x-DNPH is degradation product of DNPH, Its retention time is around 14.5 minute.
f Not properly integrated, should be 0.
51


bab mil i
Analysis data i
Katbod i
Daserlptloa i
Colima i
Carrlar
Data Clla t
Jaapla i
oalT of Colorado Danvar
05/24/1997 12i01i10
Autosaaplar 10 alerolltar
X9LC UV Absorption
150 x 4.C Cll tararsa
Oradlant (3DOff X20
AMD2.ASC (di\paak2\data)
Aurarla Sanplaa 1995
%
*6.4G0oV
64.000mV
iKabir litanefen Atm
0 0.153 12.62
0 1.000 (1.70
0 1.51a 56.19
0 2.250 17.51
0 2.016 9.78
1 3.516 1(766.99 N
0 5.(16 27.79 n
0 5.783 7.78 a
3 6.216 55.09 N
4 7.566 7.(1
0 7.950 5.88
0 6.616
0 10.766 26.70
0 11.(63 16.69
8 12.(16 5.47
0 12.900 10.45
0 13.383 6.83
0 14.133 7.53
0 14.663 6.74
0 15.266 13.56
0 15.750 7.59
Figure 3.21 Chromtogram of the sample cartridge with no added carbonyls
after exposure to 0 ppb ozone.
52
.8 a a


Lab um
Jkzulyili data
Kathod
Daacrlptloa
Colusa
Carrlar
Data file
Saapla
Valr of Colorado Denver
94/01/1997 15i30tl7
Aatoaaaplar 10 alerolltar
i X7LC DV Absorption
i 150 4.5 wm CIS Xtvaria
I Qradlaat ODC* 130
I Antil.ASC (d \peafc2\data)
i Aurarla faaplea 1995
6.400mV
64.000OV
i cm
_ Cooorwnt Vufe r JUtont on Atm
0 0.266 23.03
(iriknom) 0 0.966 92.71
(uAnovi) 0 2.250 1480.14
(viknew) 0 3.100 11724.42 R
(02*dtCOap 2 4.766 60.83 N
(irkno*^) 0 S.300 102.66 R
[WklWA) 0 6.000 4.31
(irknowi) 0 6.783 13.41
4 7.250 37.59
torasldtiy<0 4 7.666 4.98
(u4no*i) 0 1.900 209.01 R
(ifiknotn) 0 10.216 20.67 K
r*Acton* 4 11.150 24.73
Cwknon) 0 12.000 24.20
(wiknotfO 0 12.416 23.23
Itnkrtotn) 0 13.066 4.97
CuUaot) 0 13.583 26.44
(UklWA) 0 14.233 89.10 R
(unknow} 0 15.283 19.76
Figure 3.22 Chromatogram of the sample cartridge with no added carbonyls after
exposure to 100 ppb ozone.
53


Lab OAMa
Analysis date
Matbod
Description
Coluan
Carrier
Data file
Saapla
i Uolv of Colorado Denver
i 06/01/1997 14i32i40
i Autosaapler 10 alerolltar
i HPLC UV Absorption
t 150 x 4.6 n Cl Reverse
i Gradient CH3CH 820
i APB39.A5C (di\peaU\data)
i Aurarla Staples 1995
-6.4Q0mV
64.OOOmV
i H tinknowi) 0 0.233 20.07
; k (irtnovn) 0 0.966 145. a
it (intaotn) 0 1.550 0.62
1 tinkno*0 0 2.450 2807.00
' 1 )
! Cmkno*a>) 0 3.316 S8Z3.2S
- i ~-0~ 0 4.216 42.62
: [mknotf>) 0 4.466 59.80
; i ^ -- tirtnotfi) 0 5.266 133.96
; i S3 0 5.433 92.70
: j (irtknoan) 0 6.743 11.43
: r (ir*no*\) 0 0.016 53.12
§ Curinowi) 0 0.616 7.00
~ V leatalSehyda 5 9.916 361.92
>ra*Aaatona 6 10.066 20.91
; fN >ra*Actona 4 11.300 17.48
Acatona 7 11.683 9.30
>ot*Acatana 0 12.403 10.61
_ I (irtnotft) 0 12.900 33.30
: i (tnknowi) 0 13.400 7.57
L i b (iKknotai) 0 13.050 26.48
(uifcnowij 0 14.433 165.34

i | r Cinknetn) 0 15.316 16.49
Figure 3.23 Chromatogram of the sample with no added carbonyl after exposure
to 190 ppb ozone.
54


Lab oaa*
Analyal* data
Kathod
Daccrlptlon
Coluan
Carriar
Data fila
Sanpla
t Calv o£ Colosado Oaavar
t 05/26/1)97 0l<5Si2G
Autoaasplar 10 alerolltar
i HPLC UV Abaorptloa
s ISO x 4.0 aa CIO Havana
i Cradiant CB30T H20
i APH34.ASC (di\paafc2\data)
i Auraria Saaplai 1)95
*>
-6.400oV
64.OOOmV
c
K>
!r
Cwponcnt (unknown) Hueba 0 r ftatencf 0.200 on Area 12.60
(unknown) 0 1.000 202.99
(unknown) 0 2.266 4335.78
(unknown) 0 3.466 3797.22
*< unknown) 0 3.600 1169.90
(unknown) 0 4.466 40.78
(unknown) 0 4.716 46.31
(unknown) 0 S.650 102.49
<02-acid* 3 6.250 62.70
Formaldehyde 4 7.566 12.64
(unknown) 0 S.666 32.79
(unknown)
(unknown)
(unknown)
ott-Actona
(unknown)
(unknown)
(unknown)
(unknown)
11.400
11.964
12.303
12.633
13.366
13.003
14.283
14.066
207.01
16.90
19.93
16.50
7.07
49.20
9.
16.
165.32
Figure 3.24 Chromatogram of the sample with no added carbonyl after exposure
to 320 ppb ozone.
55
K8


4. DISCUSSION
4.1 Field Study
4.1.1 Sampler intercomparison period
As described above, carbonyl compounds were measured by three different
samplers at same time and same condition. The samplers were calibrated before
putting them in the field, and a leak test was performed every other day, before and
after changing sample cartridges.
On the first three days of the sampling period, no ozone scrubber was used.
The concentrations measured by the three samplers should be the same value. But as
can be been seen from figure 3.1, formaldehyde, acetaldehyde and acetone
concentrations measured by the different samplers are not the same. This indicates
that there was a calibration problem in the system. Since we didnt recalibrate the
samplers right after finishing the field study, the calibration problems remain
uncertain. However, we have some clues about the cause of this problem from the
field data sheets and later laboratory study. It appears that something was not quite
right, about the unstable flowrate reading on each solenoid. Sometimes the flowrate
reading on one of the solenoids dropped a lot. This might be related to the fluctuation
of the voltage.
56


Although carbonyl concentrations from three samplers can not be compared to
each other directly, the ozone effect on carbonyl compounds can be seen from the
ratio of the scrubbed to unscrubbed carbonyl concentration, which are summarized in
Table 3.1 and 3.5.
4.1.2 Ozone Removal Evaluation
Correlation of formaldehyde concentrations with ozone removed to that
without ozone removed.
Figure 3.9 and 3.10 show the formaldehyde concentrations measured using
ozone denuder scrubber and KI cartridge scrubber, plotted against, the corresponding
amounts of unscrubbed formaldehyde data. The slope, intercept and R2 of the best fit
lines are summarized in Table 3.1 and Table 3.5. The uncertainties represent the 95%
confidence levels for the slope and intercept.
As can be seen from the tables, when ozone is removed by denuder scrubber
the slope of the regression line is within the 95% confidence range as compared with
the measurements without scrubber (from 1.170.34 to 1.120.11). The R2is about
the same also, from 0.74 to 0.71. Thus, the degree of correlation remains unchanged
despite the large difference in data collection (4 days versus almost 7 weeks ). The
intercept changes from 0.23 1.13 to 0.440.41, well within the 95% confidence
level. This indicates that assuming the unscrubbed data are without systematic error,
57


formaldehyde concentration is not affected by the ozone denuder scrubber when
sampling with DNPH silica gel cartridges.
When ozone is removed by KI cartridge scrubber the slope of the regression
line is not changed a lot, from 0.480.25 to 0.390.11. The intercept is from
0.990.87 to 1.460.41. These data indicate when sampling with DNPH silica gel
cartridges, formaldehyde concentration didnt change significantly after using the KI
cartridge scrubber. However, the R2 decrease from 0.51 to 0.23 suggests that using the
KI cartridge scrubber caused scatter in the data. There was an extra peak found
between formaldehyde and acetaldehyde from field sample chromatogram which was
shown in figure 4.7. This indicates an interference to DNPH silica gel method exists
when using KI cartridge scrubber.
Correlation of acetaldehyde concentrations with ozone removed to that
without ozone removed.
Figure 3.11 and 3.12 show acetaldehyde concentrations measured using ozone
denuder scrubber and KI cartridge scrubber, plotted against the unscrubbed
acetaldehyde data. The slope, R2 and intercept of the best fit lines are summarized in
Table 3.5. The uncertainties represent the 95% confidence levels for the slope and
intercept.
The table shows that when ozone is removed by denuder scrubber the slope of
the regression line is not significantly changed, from 1.540.41 to 1.270.13. The R2
58


does not change much either, from 0.78 to 0.69. The intercept changes from
0.603.01 to 1.01 0.95, well within the 95% confidence level. These indicate when
sampling with DNPH silica gel cartridges, the acetaldehyde concentration did not
change significantly whether using ozone denuder scrubber or not.
When ozone is removed by KI cartridge scrubber, the slope of the regression
line is almost the same, from 0.740.10 to 0.73 0.05. The R2 does not change much,
from 0.94 to 0.83. The intercept changes from 0.600.77 to 1.320.38, well within
the 95% confidence level. These indicate that when sampling with DNPH silica gel
cartridge, the acetaldehyde concentration did not change significantly whether using a
KI cartridge scrubber or not.
Correlation of acetone concentrations with ozone removed to that without
ozone removed.
Figure 3.13 and 3.14 show the acetone concentrations measured using ozone
denuder scrubber and KI cartridge scrubber, plotted against the unscrubbed acetone
data. The slope, intercept and R2 of the best fit lines are summarized in Table 3.5. The
uncertainties represent the 95% confidence levels for the slope and intercept.
As can be seen from the table, when ozone is removed by denuder scrubber,
the slope of the regression line is decreased significantly, from 2.82 0.77 to
1.150.20. The R2 is decreased from 0.77 to 0.45. The intercept is not changed
significantly, from -4.42 8.65 to 4.18 2.01, within the 95% confidence level.
59


These indicate adding ozone denuder scrubber caused low acetone concentration
reading and scattering data. Therefore ozone denuder interferes with acetone
measurement when using DNPH silica gel method.
When using KI cartridge scrubber to remove ozone, the slope of the regression
line does not significantly change, from 0.42 0.12 to 0.48 0.07. The intercept
increases from 1.841.42 to 3.33 0.78. But this change is not significant, still within
the 95% confidence level. These indicate when ozone was removed by KI cartridge
scrubber, the acetone concentration measured by DNPH silica gel method did not
change significantly. But the R2 decreased from 0.77 to 0.54, as can be observed the
data scatters more in Figure 3.14. This means using KI cartridge scrubber causes an
interference in measurement of acetone by DNPH silica gel technique.
Ozone concentration dependence of the ratio of the scrubbed to
unscrubbed carbonyl concentration. Figure 3.15-17 show the ratio of scrubbed to
unscrubbed carbonyl concentration as a function of ozone concentration. The best fit
lines are essentially parallel to the x-axis, indicating scrubbed/unscrubbed ratio does
not have dependence on the ozone concentration. As shown in Tables 2.2 and 2.3
which will be discussed in detail in chapter 4.2, the denuder scrubber and KI cartridge
scrubber efficiently remove ozone. The ratio of scrubbed to unscrubbed results would
reveal any interaction between the ozone and carbonyl compounds. That is if ozone
interference cause measured carbonyl concentration to decrease, the scrubbed /
60


unscrubbed ratio should increase as ozone concentration increases. The lack of
correlation between scrubbed/unscrubbed ratio and the ozone concentration suggests
that ozone does not produce significant interference to carbonyl compounds over in
these measurements.
4.2 Laboratory interference study
As can be seen from Figure 3.24, the chromatogram of sample with no added
carbonyl clearly shows that ozone (320ppb) reacts with 2,4-DNPH on the surface of
silica gel. By comparison with Figure 3.21, several extra peaks are seen in the
chromatogram. Peak #1,2,3 and 4 are the results of decomposition products from the
reaction of ozone with 2,4-DNPH. Peak #2 is identified as formaldehyde, other peaks
are unknown. In general, these peaks increased proportional to increasing ozone
concentration, as shown by Figure 3.21, 3.22, 3.23 and 3.24. The DNPH reagent
loading in the extract was found to be significantly lower after exposure to ozone.
In first and second sets of the experiments, carbonyl compounds were added
to the gas stream. The measured formaldehyde and acetaldehyde concentrations did
not significantly change in presence of 20 ppb ozone, but the concentrations
decreased as ozone increased in the range from 50 ppb to 320 ppb of ozone, when no
ozone removal device was used. This indicates a negative ozone interference exists,
which means the measured carbonyl concentration is lower than the actual value.
61


Ozone might react with the formaldehyde and acetaldehyde derivatives on the surface
of silica gel. The chromatogram of sample with or without carbonyl added are very
similar, only the size of the carbonyl peaks are different, can be seen from figure 4.1
and 4.2. This indicates formaldehyde derivative react with ozone in the same manner
as DNPH, they yield same products. The rate constant for the reaction of ozone with
formaldehyde or acetaldehyde 2,4-DNPHydrazone is unknown, but the gas phase
ozone reaction for the formaldehyde hydrazone (CH2=NNH2) is 2.5 X 10'17 cm3
molecule'1 s'1 [9]. The rate constant for ozone reaction with formaldehyde and
acetaldehyde are very slow, with rate constants of <10'20cm3 molecule'1 s'1 at 298K
[11]. So these reactions are negligible under atmospheric conditions.
Table 3.8 shows the comparison of formaldehyde data from set 1 and 2. As
can be seen from the slope, intercept of the best fit line of formaldehyde concentration
versus ozone, there is no significant change between two sets of data. This means that
high fractions of acetaldehyde ( [acetaldehyde] : [formaldehyde] 6 ) did not affect
the ozone effect on formaldehyde measurement.
As can be seen in figure 3.18 to 3.20, the ozone denuder scrubbed or KI
cartridge scrubbed carbonyl concentrations do not change as ozone concentration
increases. This indicates the ozone denuder and KI cartridge scrubber can be used to
remove ozone effectively. As a result, the measured concentration of carbonyl
62


Lab nama
Analyaia data
Mathod
Daacription
Column
Carrlar
Data flla
Saopla
i Unlr of Colorado Danvar
> 06/01/1997 1402iSl
i Aueoaamplar 10 mlerolltar
: HPLC CV Abaorpeioa
: 150 x 4.6 mm C18 Ravarsa
i Sradlant CH3CN H2G
< APH38.ASC (dt \paak2\data)
< Aurarla Samplaa 1995
6.400mV
64.OOOmV
Nuibar latent I on Arta
0 0.233 18.05
0.966
1.533
154.97 N
11.57 N
0 2.450 3139.15 N
0 3.166 7250.76 N
4.266
5.050
5.600
6.466
6.650
7.283
7.733
8.666
8.916
9.600
10.700
11.200
11.616
12.216
12.816
13.316
\i-.m
14.400
a u-.m
38.92 H
154.96 N
74.55 M
7.30 *
7.45 M
16.14 *
44.13 N
10.16
6.51
423.05
14.04
28.00
8.42
7.73
54.94
13.70
155.18
Figure 4.1 Chromatogram of the sample with no added carbonyls after
exposure to 230 ppb ozone
63


Lab nama
Analyaia data
Hathod
daacription
Coluan
Carriar
Data film
Saapla
i Univ of Colorado Danvar
06/09/1997 13:08:31
: Autoaasrplar 10 aicrolitar
: HPLC UV Abaorption
: ISO x 4.6 am C18 Xavaraa
: Qradiant CH3CH H20
> APN43.ASC (d:\paak3\data)
: Aurarla Saaplaa 199S
-6.400mV
64.OOOmV
8?5t,on 36.45
0.600
0.983
1.630
3.083
3.700
3.066
3.883
4.830
5.383
6.916
7.566
7.950
9.433
10.300
10.533
11.166
11.616
13.350
13.933
13.516
13.883
14.566
1S.500
17.33
109.39
15.08
147.93 II
1386.46 N
3454.93 N
113.41 K
69.07 N
160.83 M
12.85 N
28.09 N
5.67 N
276.55 N
5.46 N
19.16 N
33.04 M
17.00
6.69
63.16
13.39
10.96
71.67
9.83
Figure 4.2 Chromatogram of the sample of the sample with formaldehyde
after exposure to 230 ppb ozone.
64


compounds after using denuder scrubber or KI cartridge scrubber exhibit no
dependence on the added ozone concentration. Furthermore, use of ozone scrubbers
does not introduce any contamination to carbonyl compounds, since noadditional
new peaks are observed in the chromatogram in the samples collected after the
scrubber.
4.3 Comparison of field and laboratory study
The laboratory study shows that a negative interference exists at an ozone
concentration of 50 ppb or above. However, from field study we have found no ozone
interference on measurement of carbonyl compounds when using DNPH silica gel
method. As we learned from laboratory study, the magnitude of ozone interference
depend on the concentration of ozone sampled. Ambient ozone concentrations ranged
from 6.25 to 74.25 ppb with significant diurnal variation during the field sampling
period, it reaches a maximum from noon to 4pm. The average ozone concentration
measured in air during noon to 4pm is ~60 ppb. According the laboratory study, at
this level of ozone, we should see 18%-27% of formaldehyde and 19%-23% of
acetaldehyde losses. Figure 4.3-5 are three chromatograms of three samples,
chromatogram #B4L177 with a ozone denuder scrubber, #B1L177 with a IQ cartridge
scrubber and #B3L177 without ozone removal device. These DNPH cartridges were
sampled at same time period, from noon to 4pm, that ozone reaches a maximum. As
65


Lab oaaa
Analysis dacs
Katbod
Dasezipeion
Coluan
Cacriar
Data flla
Saapla
t UblT of Colorado Danvar
i 07/24/1996 16i02il9
i Autosaaplar 10 aierolitar
i HPLC UV Absorption .
i ISO x 4.6 aa CIS Kavarsa
i Oradiant CH30T H20
: L22KU107.ASC (di Vpaak2\data)
i Auraria Saaplas 199S
-6.400mV
64.000mV
0 0.116
9.26
0 1.03 7.92
0 1.450 10.67
0 2.100 12.29
0 3.150 4337.13
0 4.283
15.55
0 5.616 247.94
0 7.683 437.78
Icctilcttiyd* 5 9.516 11.98
0 10.666 476.28
7 11.600 53.99
0 13.700 8.91
0 14.700 76.73
Figure 4.3 Chromatogram of the filed sample #B4L177, measured by sampler
#4, with denuder scrubber.
66


Lab nan#
Analyaia data
Katbod
Daacrlptlon
Coluan
Carriar
Data £11#
Saapla
i OnIt o£ Colorado Danvar
i 07/32/1996 23i18i12
i Autoaaaplar 10 alerolltar
i HPLC OV Abaorptloa
i 150 x 4.S aa C18 KaTeraa
i Oradiant CH3CH'- S20
> L22R035.ASC (d<\pak2\data)
> Aurarla Saaplaa 1995
-6.400mV
64.000mV
>
>
Caaperant Nu#ar Mttntlon ArtM
(irtnom) 0 0.350 45.41
(infcnow) 0 1.050 6.62
(u*now) 0 1.433 10.39
(irtzwMi) a 2.116 13.51
[iriraui) 0 3.200 4287.04
(uifcnowi) 0 4.350 14.88
Cuanowi) 0 5.183 5.81
(tnknatn) 0 5.766 146.34
Foraaldcfcycfe 4 7.083 27.01
(irfcnoMi) 0 8.133 384.74
s 9.816 28.81
Pr-Aetone 6 10.900 200.08
UlCOM 7 11.866 52.41
Cunknowi) 0 13.883 5.38
Cinfcnaui) 0 14.616 26.39
CU*nowi) 0 14.933 16.12
Figure 4.4 Chromatogram of the filed sample #B1L177, measured by sampler
#1, with KI cartridge scrubber.


Lab naaa i
Analyaia data i
Katbod t
Daacriptlon i
Coluan i
Casriar i
Data £ila i
Saapla i
Unir o£ Colorado Danvar
07/23/1996 19104 t2S
Autoaaaplar 10 aierolitar
HPLC DV Abaorption
ISO x 4.6 aa CIS Ravaraa
Oradiant CH3CJT H20
L22R071.ASC (di \paaU\data)
Auraria Saaplaa 1995
6.400mV
64.000mV
Ceapomnc itiafem- (atentfon Am
(unknown) 0 0.200 79.47
(tnknoun) 0 1.016 13.52
(uifcnow) 0 1.466 14.81
(tnknow) 0 2.100 93.75
(infcnew) 0 3.166 3090.70
(inknotn) 0 4.066 30.25
Cinknowi) 0 9.100 14.09
CirJcnow) 0 S.6S0 183.82
(infcnowi) 0 7.916 382.35
Cinfcnow) 0 9.500 42.88
(inknotn) 0 10.700 157.30
lector* 7 11.633 47.56
>ott-Ac*ton* 6 12.316 7.04
(wfcnowi) 0 13.700 10.52
(mknswn} 0 14.663 73.89
Figure 4.5 Chromatogram of the filed sample #B3L177, measured by sampler
#3, without ozone scrubber.
68


can be seen from chromatogram #B3L117, the size of the peak in front of
formaldehyde is very small. This peak is produced by ozonation of 2,4-DNPH. It
means only small amount of ozone reached the DNPH cartridge, it is not big enough
to cause interference. Therefore, ozone interference was not observed. Using long
polyethylene tubing and connector might remove some of ozone, since ozone can add
to the double bond in polyethylene. In laboratory study only Teflon tube and
connector were used to limit ozone loss in the system.
The results from field study shows that using KI cartridge scrubber caused
interference on measurement of formaldehyde concentration, but the KI cartridge
worked well in laboratory study. This disagreement could be due to differing KI
cartridges used. Figure 4.6 shows a chromatogram of KI cartridge scrubbed sample
from field study. As can be seen clearly, there is an unidentified peak between
formaldehyde and acetaldehyde after adding KI cartridge. It gets smaller with
continuous use. This unidentified peak did not show up in the chromatograms of
laboratory study. In order to ascertain this change is not due to differing experimental
conditions, an experiment was conducted to sample formaldehyde using KI cartridge
scrubber made for field study, ozone concentration added in the gas stream was
50ppb, which is similar to atmospheric condition. As can be seen from figure 4.7
(chromatogram #30), the peak between formaldehyde and acetaldehyde appeared on
the chromatogram again. The way we prepare KI cartridge for laboratory use and field
69


use is exactly the same. The cause of the difference could be due to an impurity or
contamination of KI reagent and silica gel used for field study.
Since acetone source for laboratory study is unavailable, the measured acetone
results from field study can not be compared.
70


Lab naaa i
Analyaia data i
Katbod i
Daacrlptioa i
Coluan t
Carriar i
Data £ila i
Saapla i
Unir of Colorado Danvar
07/22/1996 21il3127
Autoaaaplar 10 alerolltar
H9LC OV Abaorptlon
ISO x 4.6 aa C18 Kararaa
Oradiant CH30T- B20
L22RU29.ASC (d>\paak2\data)
Aurarla Saaplaa 1995
-6.400mV
64.000mV
Caapontnt Outer tatantien Atm
0 0.563 67.19
0 1.066 0.52
0 1.(66 0.71
0 2.116 13.66
0 3.200 2777.69
6.303
6.603
13.(0
6.50
5.200 13.06
5.766 127.75
6 7.066 95.07
0 0.116 216.59
7 11.050 30.23
13.766
16.066
15.(66
15.916
12.79
50.67
5.60
7.10
Figure 4.6 Chromatogram of the field sample #B1L161, measured by
sampler#!, with a new KI cartridge scrubber.
71


Lab sum
Analyaia data
Katbod
Oaaeription
Coluan
Carriar
Data flla
Saapla
i Onlv of Colorado Danvar
i 06/05/1997 08i31>06
i Autoaaaplar 10 alerolltar
t HPLC OV Abaorptlon
i ISO x 4.6 aa CIS Havana
i Oradiant CH3QT H20
< APH35.ASC (dt\paaU\data)
i Aurarla Saaplaa 199S
-6.400mV
64.000mv
attention
0.200
1.033
1.483
2.216
2.750
3.300
5.133
5.666
*r
39.16
20.19
140.54
7.59
5.66
11361.12 N
43.64 a
464.08 N
0 6.783 107.64 N
4 7.716 1945.29 N
9.216
9.563
18:%*
10.550
11.500
12.116
12.433
13.150
13.750
14.366
15.350
12.16 a
19.17 a
*2:*! H
72.93 a
12.53
9.98
6.96
6.60
24.19
159.16
12.93
Figure 4.7 Chromatogram of the laboratory sample, with a KI cartridge scrubber
made a year ago.
72


5. CONCLUSIONS
In the laboratory study, we found a severe negative bias on formaldehyde and
acetaldehyde concentrations when ozone is present, even as low as -50 ppb. The
ozone denuder scrubber and KI cartridge scrubber can be used to remove ozone
without interfering with measurement of carbonyl compounds by DNPH silica gel
method. As we expected, the ozone interfering peaks increased as a function of ozone
concentration. The magnitude of the ozone interference depended on the amount
ozone sampled.
In field study, comparison of three sets of field data showed that when ozone
was removed by ozone denuder scrubber or KI cartridge scrubber, the carbonyl
compound concentrations obtained by DNPH silica gel technique were not
significantly different. No correlation was found between ozone concentration and
carbonyl losses in unscrubbed samples. KI cartridge scrubber used for field study
interferes with the measurement of formaldehyde concentration. Likely some other
mechanisms exists, that can remove low level of ozone in this sampling system.
Consequently the final result shows that when ozone concentration is low, the ozone
interference is small.
73


REFERENCES
1. Finlayson-Pitts, B. J.; Pitts, J. N., Jr. Atmospheric Chemistry: Fundamentals
and Experimental Techniques', Wiley and Sons: New York, 1986.
2. Pitts, J. N., Jr,; Wan, J. K. S. In The Chemistry of the Carbonyl Group', Patai,
S., Ed.; Interscience Publishers: New York, 1966.
3. Grant L. D. (1989) and Tikton B. E. Health effects of major combustion ad
transformation products of alcohol motor vehicle fuels. Presentation at
Workshop on the Effects of Atmospheric Aldehydes on Health and
Environment, Sau Paulo, Brazil, 12-13 April.
4. EPA Health assessment document for actaldehyde. EPA 6008-86-015A,
external review draft, Environmental Criteria and Assessment Office. Reserch
Triangle Park, NC.
5. Tanner R. L., Miguel A. H. Atmospheric chemistry of aldehydes: enhanced
peroxyacetyl nitrate formation from ethanol-fueled vehicular emissions. Envir.
Sci. Technol. 22, 1026-1034.
6. Williams R. L., Lipari F. and Potter R. A. Formaldehyde, methanol and
hydrocarbon emissions from methanol-fueled cars. J. Air Waste Man. Ass. 40,
747-756.
7 Amts R. R. and Tejada S. B. 2,4-Dinitrophenylhydrazine-coated silica gel
cartridge method for determination of formaldehyde in air: identification of an
ozone interference. Envir. Sci. Technol. 23,1428-1430
8. Dasgupta P. K., Dong S. Continuous liquid-phase fluorometry coupled to a
diffusion scrubber for the real-time determination of atmospheric
formaldehyde, hydrogen peroxide and sulfur dioxide. Atmospheric
Environment 22, 949-963.
9. Tuazon, E.; Carter, W, P. L. Reactions of hydrazines with ozone under
simulated conditions. Environ. Sci. Technol. 1981,15, 823-828.
10. Smith D. F., Kleindienst T. E. and Hudgens E. E. Improved high-performance
liquid chromatographic method for artifct-free measurements of aldehydes in
74


the presence of ozone using 2,4-dinitrophenylhydrazine. J. Chromat. 6,425-
444.
11. Atkinson R. and Carter W. P. L. Kinetics and mechanisms of the gas-phase
reactions of ozone with organic compounds under atmospheric conditions.
Chem. Rev. 84, 437-470.
12. Vairavamurthy A. (1993) Sampling of atmospheric carbonyl compounds for
determination by liquid chromatography after 2,4-dinitrophenylhydrazine
labelling. In Sampling and Analysis of Airborne Pollutants, (edited by
Winegar E' and Keith L. H.), Lewis, Chelsea, MI.
13. Hoigne J., Bader H., Haag W. R. and Staehelin J. Rate constants of reaction
of ozone with organic compounds in water-III. Wat. Res. 19, 993-1004.
14. Grosjean D. and Parmar S. S. Laboratory tests of KI and alkaline annular
denuders. Atmospheric environment 24A,2695-2698
15. Williams E. L. II and Grosjean D. Removal of atmospheric oxidants with
annular deunders. Envir. Sci, Technol. 24, 811-814.
16. Slemr J. Determination of volatile carbonyl compounds in clean air. Fres. J.
Anal. Chem. 340,672-677.
17. Tanner R. L., Markovits G. Y., Ferreri E. M. and Kelly T. J. Sampling and
analysis of gas-phase hydrogen peroxide following removal of ozone by gas-
phase reaction with nitric oxide. Analyt. Chem. 58,1857-1865.
18 Purdue, L. J., Dayton, D. P., Rice, J., and Bursey, J. (1991). Assistance
Document For Sampling And Analysis Of Ozone Precursors. US
environmental Protection agency, Research triangle park, NC, EPA/ 600/
8-91/215.
19 Tejada S. B. (1986) Evaluation of silica gel cartridges coated in situ with
acidified 2,4-dinitrophenylhydrazine for sampling aldehydes and ketones in
air. Int. J. envir. Analyt. Chem. 26, 167-185.
75


Full Text

PAGE 1

INVESTIGATION OF OZONE INTERFERENCE ON THE l\ffiASUREMENT OF ATMOSPHERIC CARBONYL COMPOUNDS by Xiaowei Shang B.S., Peking University, 1987 A thesis submitted to the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Master of Science Chemistry 1997 [A] /

PAGE 2

This thesis for the Master of Science degree by Xiaowei Shang has been approved by /

PAGE 3

Shang, Xiaowei (M.S., Chemistry) Investigation of Ozone Interference On The Measurement Of Atmospheric Carbonyl Compounds Thesis directed by Professor Larry G. Anderson ABSTRACT Laboratory tests conducted by the U.S. Environmental Protection Agency and others have suggested that ozone present in ambient air interferes with the measurement of carbonyl compound concentration when using the 2,4dinitrophenylhydrazine-coated silica gel cartridge (SGC) technique. The research was carried out in two parts. The first involved a field study in which three automated sequential air samplers were used simultaneously to collect ambient formaldehyde, acetaldehyde and acetone at Marine Street in Boulder Two samplers collected sample air after passing it through two different kinds of ozone scrubbers. The third sampler collected ambient air samples directly. Samples analysis was performed using HPLC with ultraviolet/visible detection An ozone deunder scrubber and Kl cartridge scrubber were developed for this study and were tested for 111

PAGE 4

their effectiveness for removal of ozone. Comparison of three sets of field data showed that when ozone was removed by the ozone denuder scrubber or KI cartridge scrubber, the carbonyl compound concentrations obtained by SGC technique were not significantly different than those with no scrubber. This suggests that when the ozone concentration is low, the ozone interference is small. Studies of ozone interference and removal were conducted in the laboratory to evaluate the DNPH cartridge technique. Different concentrations of ozone were mixed with formaldehyde, generated by a permeation tube filled with paraformaldehyde. These were sampled using the DNPH silica gel technique. The results of laboratory tests show a severe negative bias when ozone is present, even at low ozone concentration. This abstract accurately represents the content of the candidate's thesis. I recommend its publication. lV

PAGE 5

ACKNOWLEDGEMENTS Special thanks to Professor Larry Anderson for his support and guidance throughout this work. I also wish to express my gratitude to Professor John Lanning for his advice and assistance, and to Professor Donald Zapien for his encouragement. Thanks Susan Riggs and Michael Pribil for their help during this study /

PAGE 6

CONTENTS CHAPTER 1. INTRODUCTION ..................................................................................................... 1 1.1 Enviromnental and Human Impact of Carbonyl Compounds .............................. I 1.2 Sources of Carbonyl Compounds ....................................................................... .3 1.3 Measurement Techniques .................................................................................... 4 1.4 Derivatization ofCarbonyls by 2,4-Dinitrophenylhydrazine .............................. 6 1.5 Ozone Interference ............................................................................................... 7 1.6 Previous Research ................................................................................................ 8 1. 7 Approach in This Work ..................................................................................... 11 2. EXPERIMENTAL SECTION ................................................................................. 13 2.1 Experiment for Field Study ................................................................................ l3 2.1.1 Field Sampling Site ..................................................................................... 13 2.1.2 Pollutant Measurement ............................................................................... 13 2.1.3 Sampling Equipment. .................................................................................. l4 2.1.4 Ozone Denuder Scrubber ............................................................................ 16 2.1.5 KI Cartridge Scrubber ................................................................................. 17 Vl

PAGE 7

2.1.6 Sampling Procedures ........................................ ......................................... 19 2.2 Laboratory Studies .................................................. .... ........... .............. ............ 20 2.2.1 Collection Efficiency of Ozone Scrubbers .................................................. 20 2.2.2 Lifetime Study ofK.I Cartridge Scrubber ................................................... 23 2.2.3 Interference Studies ..................................................................................... 23 3 DATA ANALYSIS AND RESULTS ................ ..... ... .................................. .... . ... 26 3.1 Field Study .......................... ................................... . ........................................ 26 3.1.1 July 12-16 Period ........... ............................................................................ 26 3.1.2 July 16-September 2 Period ...................................................................... 34 3.2 Results from laboratory interference studies ....................................... ......... ... .45 4. DISCUSSION .......................................................................................................... 56 4.1 Field Study ......................................................................................................... 56. 4.1.1 Sampler intercomparison period ........................................................... ..... 56 4.1.2 Ozone Removal Evaluation .................................................................... ... 57 4.2 Laboratory interference study ....................................................................... .... 61 4.3 Comparison of field and laboratory study ......................................................... 66 5. CONCLUSIONS ........................ ......................................................................... .... 75 REFERENCES ............................................................................................................ 73 Vll

PAGE 8

FIGURES Figure Figure 1.1 Reaction of carbonyls with DNPH to form 2,4-dinitrophenyl hydrazones .. 6 Figure 2.1 Schematic of sequential sampler ................................................................ 15 Figure 2.2 Schematic of ozone denuder scrubber ........................................................ 17 Figure 2.3 Schematic of KI cartridge scrubber ........................................................... 18 Figure 2.4 Schematic diagram of the experimental set up of interference studies ...... 25 Figure 3.1 Average concentration of carbonyl without ozone removal.. ................... 28 Figure 3.2 Correlation offormaldehyde, July 12-16 ................................................... 31 Figure 3.3 Correlation of acetaldehyde, July 12-16 ..................................................... 31 Figure 3.4 Correlation of acetone, July 12-16 ............................................................. 32 Figure 3.5 Ratio of scrubbed to unscrubbed carbonyl concentration versus time period ......................................................................................................... 33 Figure 3.6 Average concentration of formaldehyde (July16-Sep2) ............................. 35 Figure 3.7 Average concentration of acetaldehyde (July16-Sep2) ............................. 36 Figure 3.8 Average concentration of acetone (July16-Sep2) ....................................... 37 V111

PAGE 9

3.9 Correlation of formaldehyde, ozone denuder scrubbed to unscrubbed concentration ... ............ ............ ... ..... ... .... .... ....... .... . ............ ...... . .......... 39 Figure 3.10 Correlation of formaldehyde, KI cartridge scrubbed to unscrubbed concentration .............. .... ................................. . ........................................ 3 9 Figure 3.11 Correlation of acetaldehyde, ozone denuder scrubbed to unscrubbed concentration . ... ... ......... ..... ... . ... .... .... ..... . .... .... . ..... .... ...... ..... ......... 40 Figure 3.12 Correlation of acetaldehyde, KI cartridge scrubbed to unscrubbed concentration .............. ......................................................................... ..... 40 Figure 3.13 Correlation of acetone, ozone denuder scrubbed to unscrubbed concentration .. ..... ........... ............. ................................................ ... ...... 41 Figure 3.14 Correlation of acetone KI cartridge scrubbed to unscrubbed concentration ...................................................... ................................... ... 41 Figure 3.15 Ratio of scrubbed to unscrubbed formaldehyde concentration versus ozone concentration ..... ....... ........................... ............................. .... ...... 42 Figure 3.16 Ratio of scrubbed to unscrubbed acetaldehyde concentration versus ozone concentration ................ . ........................... . . ... ................................. ....... 43 Figure 3.17 Ratio of scrubbed to unscrubbed acetone concentration versus ozone concentration ............ . ... ....................................................................... ..... 44 Figure 3.18 Formaldehyde concentration from experiment set 1 as a function of ozone added ............................................................ . ... ............................... ......... 47 IX

PAGE 10

Figure 3.19 Acetaldehyde concentration from set 1 experiments as a function of ozone added ............................................................................................... 48 Figure 3.20 Formaldehyde concentration from set 2 experiments as a function of ozone added ................................ . ......................... . ........ . ........ ...... .... ... 49 Figure 3.21 Chromtogram of the sample cartridge with no added carbonyls after exposure to 0 ppb ozone ............................................................................ 52 Figure 3.22 Chromatogram of the sample cartridge with no added carbonyls after exposure to 100 ppb ozone .................................................. ..... ................. 53 Figure 3.23 Chromatogram ofthe sample with no added carbonyl after exposure to 190 ppb ozone .................... ......... ... ............... ...................... ................... ... 54 Figure 3.24 Chromatogram of the sample with no added carbonyl after exposure to 320 ppb ozone ...... ....... ............................................................................. 55 Figure 4.1 Chromatogram of the sample with no added carbonyls after exposure to 230 ppb ozone ........................................................................................ 64 Figure 4.2 Chromatogram ofthe sample of the sample with formaldehyde after exposure to 230 ppb ozone .................................................. ....... .... .......... 65 Figure 4.3 Chromatogram ofthe filed sample #B4L177, measured by sampler #4, with denuder scrubber ................................................................................ 68 Figure 4.4 Chromatogram of the filed sample #B1L177, measured by sampler #1, with KI cartridge scrubber ........................................................................ 69 X

PAGE 11

Figure 4.5 Chromatogram ofthe filed sample #B3L177, measured by sampler #3, without ozone scrubber .............................................................................. 70 Figure 4.6 Chromatogram .ofthe field sample #B1L161, measured by sampler#l, with a new KI cartridge scrubber ....................................................................... 73 Figure 4.7 Chromatogram ofthe laboratory sample, with aKI cartridge scrubber made a year ago .......................................................................................... 74 Xl

PAGE 12

TABLES Table Table 2.1 Flow solenoid assignment.. .............................. ........................................... l4 Table 2.2 Collection efficiency of ozone denuder scrubber ........................................ 21 Table 2.3 Collection efficiency ofKI cartridge scrubber .... ....................................... 22 Table 3.1 Carbonyl concentration correlation .................. ... ......................................... 29 Table 3.2 Average concentrations of formaldehyde ........ ........................................... 35 Table 3.3 Average concentration of acetaldehyde ...... ... ...... ....................................... 36 Table 3.4 Average concentration ofacetone ............................................................... 37 Table 3.5 Carbonyl concentration correlation during July 16September 2 sampling period .............................................................. ..... ...................................... 38 Table 3.6 Scrubbed I unscrubbed ratio versus ozone concentration ........................... .45 Table 3.7 Laboratory formaldehyde and acetaldehyde concentrations ....................... .46 Table 3.8 Comparison of formaldehyde concentration from two set data ................... 50 Table 3.9 Summary of the area changes with added ozone ........................................ 51 Xll

PAGE 13

1. INTRODUCTION 1.1 Environmental and Human Impact of Carbonyl Compounds Carbonyl compounds (aldehydes and ketones) are important intermediates in combustion of hydrocarbon fuels, and are receiving increasing attention as pollutants and as key participants in photochemical reactions which influence smog processes in the atmosphere. Because of their photolytic reactivities in the solar spectrum of the lower troposphere, carbonyl compounds play a critically important role in atmospheric free radical production. Formaldehyde photolysis is a major source ofhydroperoxyl radicals during the daylight hours via the reactions ofH and HCO with 02 : HCHO + hv H + HCO H+ 02 H02 HCO + 02 H02 + CO H202 is formed from the reaction of two hydroperoxyl free radicals: H02 + H02 H202 + 02 Photolysis ofH202 produces OH radical directly: H202 + hv 20H 1

PAGE 14

Hydroxyl radicals are also formed by the reaction below: H02 +NO OH + N02 These free radicals are responsible for the oxidation of hydrocarbons in the troposphere[1][2]. Carbonyl compounds are also precursors of oxidants including ozone and peroxyacyl nitrates. O+NO 0+02 03 CH 3 CHO + OH CH 3 CO + H 2 0 CH3CO + 0 2 CH3C002 CH3C002 + CH3C002N02 Carbonyl compounds in the atmosphere are also of concern because they have been associated with adverse biological effects. In particular, formaldehyde has been associated with a number of health effects in humans. Formaldehyde contributes to eye, nose and throat irritation. It can also cause bronchial asthma-like symptoms with some reports of asthma attacks and allergic dermatitis [3]. Based on animal studies, occupational studies, mutagenic tests and effects on DNA, there is evidence to indicate that formaldehyde may be a carcinogen in humans [4]. 2

PAGE 15

1.2 Sources of Carbonyl Compounds Carbonyl compounds are released into the troposphere from a variety of biogenic and anthropogenic sources. These compounds are either derived from direct emissions or secondary formation in the atmosphere. Direct emissions include exhaust gases of motor vehicles and industrial machinery in which hydrocarbon fuels are incompletely burned, and by vegetation. There is potential for increased carbonyl emissions, especially formaldehyde and acetaldehyde, due to changes in fuel technology such as in the use of methanol, ethanol, etc., as gasoline additive substitutes.[5][6]. The secondary atmospheric formation of carbonyl compounds in the troposphere is from the reaction of hydrocarbons with hydroxyl radical. The series of reactions is initiated by the formation of a carbo centered radical (RCH2 ), usually through reaction ofhydroxyl radical (OH) with a hydrocarbon. OH + RCH2 + H20 Carbon centered radicals are very reactive and form peroxy radicals (RCH202), via addition of02 RCH2 + 02 RCH202 Peroxy radicals can react with NO to form alkoxy radical (RCH20 ). RCH202 +NO RCH20 + N02 Alkoxy radicals can react with 02 to form a carbonyl and H02 3

PAGE 16

RCH20 + 02 RCHO + H02 Ozonolysis of olefins is also a source of carbonyl compounds. 1.3 Measurement Techniques An understanding and assessment of the role of carbonyls in tropospheric chemistry requires the accurate and precise measurement of these compounds along with their parent and product compounds. The concentrations of carbonyl compounds range from sub-ppb or low-ppb in clean air to higher ppb in urban and polluted air. Because of these trace concentrations, determination of carbonyl compounds in the ambient atmosphere poses challenging problems. Conventional methods of determination of aldehydes are based on spectrophotometry, which are, in most cases, not sensitive enough for outdoor air measurements. Spectrometric methods are usually based on functional group detection. Four different spectroscopic techniques have been used: (1) Differential Optical Absorption Spectroscopy (DO AS), (2) Fourier Transform Infrared Absorption (FTIR), (3) Laser-Induced Fluorescence Spectroscopy, (4) Tunable Diode Laser Absorption Spectroscopy (TDLAS). 4

PAGE 17

Although long-path-length optical techniques are sensitive and have been used to measure formaldehyde in air at natural levels, their immobility and high cost make them impractical for most field applications. In the 1970s, chromatographic techniques, in conjunction with chemical derivatization methods paved the way for sensitive and selective determination of carbonyls in ambient air The most commonly used chromatographic method is based on trapping carbonyl compounds by reaction with reagents such as 2,4dinitrophenylhydrazine (DNPH), followed by separation and analysis of the derivatives by high-performance liquid chromatography (HPLC). Liquid impingers or bubblers containing a DNPH solution are often employed to collect carbonyl compounds from air. However, they are not well suited for field work. DNPH-coated solid sorbents are a convenient alternative to impinger sampling and have been increasingly used. Solid sorbents, including silica gel and C1 8 which have several advantages including convenience of use, reproducibility and low blanks [7]. In spite of the widespread use at the current level of development, no chromatographic method has yet been adapted for continuous sampling and analysis of carbonyls in the ambient atmosphere. However, such real-time measurement techniques have been developed for formaldehyde using specific fluorescence detection [8]. 5

PAGE 18

1.4 Derivatization of Carbonyls by 2,4-Dinitrophenylhydrazine Derivatization ofDNPH followed by liquid chromatography and u.v. detection is currently the most popular chromatographic technique used for determination of formaldehyde, acetaldehyde and acetone at same time. The acid catalyzed condensation reaction of carbonyl compounds with DNPH is a well-known reaction for characterizing carbonyl compounds. The reaction proceeds by nucleophilic addition to the carbonyl followed by 1 ,2-elimination of water to form the 2,4-dinitrophenylhydrazone. N02 R/C=O + N02 (1) OH N02 N02 (2) N02 N02 + H20 (3) Figure 1.1 Reaction of carbonyls with DNPH to form 2,4-dinitrophenyl hydrazones. Under neutral or mildly acidic conditions the equilibrium between 2 and 3 is located more toward 2 due to hydration. However, under more strongly acidic conditions dehydration is enhanced and the equilibrium between 2 and 3 is shifted toward 3. 6

PAGE 19

After organic solvent extraction, the hydrazones are separated by high pressure liquid chromatography (HPLC) and quantitated with an ultraviolet detector operated at 360 nm. DNPH derivatization is used for several reasons: (1) to allow better chromatography of polar compounds (reducing tailing and nonlinearity of response), (2) to improve the resolution of closely related compounds, (3) to allow analysis of relatively nonvolatile components, (4) to improve detectability (increasing detection response and selectivity for specific detectors. 1.5 Ozone Interference Ozone is one of the most abundant reactive gases in air, hence could potentially cause sampling artifacts. The possible effect of ozone on DNPH-based methods for carbonyls are threefold: (1) formation of carbonyls as artifacts from reaction with sampling substrates, (2) degradation of 2,4Dinitrophenylhydrazones, (3) formation of other interfering compounds. The reagent (DNPH) itself reacts with ozone. For example, a DNPH solution rapidly became colorless 7

PAGE 20

when high concentrations of ozone in air (0.1 %) were passed through it, but the reaction products were not identified [9]. Formation of formaldehyde from the reaction of ozone with DNPH has not been studied, and cannot be ruled out. 1.6 Previous Research In a study by Amts and Tejada [7] the reaction of ozone with HCHO DNPHydrazone was identified as a potential problem when DNPH-coated silica cartridges were used for formaldehyde sampling. In their study synthetic mixtures of humidified air containing formaldehyde (20-140ppb) and ozone (0-770ppb) were sampled. The loss ofHCHO-DNPHydrazone increased markedly with increase in ozone concentration, at 25 ppb HCHO and 120 ppb ozone, about 48% ofHCHO derivatives was lost. Also noticed on silica cartridges were concurrent large losses of DNPH. It was concluded that the silica cartridge exhibited such large reductions in formaldehyde response because the DNPH derivative, which is largely formed at the front of the cartridge and immobilized, was being destroyed by uzone. In the case of impingers, the HCHO-DNPHydrazone is protected by the DNPH, which is always present in excess and well dispersed. However, a matter of concern in the impinger technique is that the products formed from the DNPH-ozone reaction can interfere with resolution of formaldehyde peak in HPLC separation. Recently, this problem 8

PAGE 21

was addressed by Smith et al.[10] who used a ternary gradient separation to obtain the desired chromatographic resolution. In contrast to the silica cartridges, C1s cartridges exhibited no loss of the HCHO derivative up to 120 ppb ozone [9]. In this case, it was reasoned that C18 substrate itself was the site of ozone reaction, thereby preventing attack on formaldehyde-DNPHydrazone. The mechanisms of ozone-initiated reactions in the above cases are not clearly understood. Atkinson and Carter [11] suggested that a chain of free-radical reactions can be initiated when ozone reacts with hydrazines, either by addition to a nitrogen, or abstraction of a hydrogen from a weak N-H bond. Amts and Tejada [7] pointed out that under the acidic conditions of the DNPH reaction, ozone addition to the protonated nitrogen is restricted and hydrogen abstraction could be the preferred route. The following pathway has been proposed by Atkinson and Carter for hydrogen abstraction: RHNNH2 + 03 RNNH2 (or RHNNH) + 02 + OH RNNH2 (or RHNNH) + 02 RH=NH + H02 RH=NH + 03 (or OH) RN=N + OH + 02 ( + H20) RN=N +N2 When DNPH-coated C1s substrate is used for sampling, the radicals generated by ozone attack can be scavenged by the C1s thus limiting further attack on 2,4DNPH or the hydrazones, so the above reactions wouldn't propagate and consume 9

PAGE 22

2,4-DNPH and hydrazone But ozone can react with the C18 to form high molecular weight carbonyls. These can react with 2,4-DNPH to form artifact hydrazone derivatives [12]. From the above discussion, it is clear that ozone may be a serious interferent in methods for sampling carbonyls in ambient air using either C18 or silica-gel cartridges coated with reagent. Arnts and Tejada [7] reported that they obtained encouraging results in a preliminary study using a potassium iodide-coated copper tubing inlet to remove ozone prior to collection with a DNPH-coated silica-gel cartridge. A well known mechanism for the reaction of ozone with KI involves participation ofwater and proceeds with the formation of iodine: 03 + 2KI + H20 02 + 12 +2KOH Hoigne et al. [13] suggested an alternate mechanism involving the formation ofhypoiodite, which does not require water: 03 + KI 02 + KOI In the atmosphere, the reaction probably occurs through the iodine pathway due to air humidity. Grosjean and Parmer [14] and Williams and Grosjean [15] reported that an annular denuder coated with KI completely removed ozone even at high flow rates, up to 20Rmin1 and formaldehyde was not removed by this device. Recently, Slemr [16] used Teflon tubing packed with crystalline KI to remove ozone in conjunction with silica-gel cartridge sampling. In laboratory tests, carbonyls were 10

PAGE 23

not found to be adsorbed by the KI filter at a relative humidity of 30-35%. However, the performance ofKI scrubbers in moist air remains unclear. As iodide salts are very hygroscopic, the physical integrity of solid iodide may be affected in moist air. This may be a problem in the cartridge-based sampling technique for carbonyls. Some metal oxides, including Mn02 and CuO decompose ozone catalytically. Studies with gas-phase standards indicate that recovery of carbonyls are not affected by passing through a CuO cartridge. Gas-phase titration of ozone with nitric oxide (NO), as used by Tanner et al. [17] in the determination ofhydrogen peroxide in the ambient atmosphere, is also a potential technique to eliminate any ozone interference. 1.7 Approach in This Work In this study, carbonyl compounds were measured by DNPH silica gel cartridge technique. The sampled gas streams were either ambient air or synthetic mixture of carbonyl compounds and ozone. During the field study, ambient formaldehyde, acetaldehyde and acetone were measured from July 12 to September 3, 1996. An ozone denuder scrubber and KI cartridge scrubber for removing ozone were evaluated, and ozone unscrubbed ambient air was also sampled at the same time. Three automated sequential air samplers were used simultaneously to collect carbonyls at Marine Street in Boulder. Two samplers collected air after it had passed through the two different kinds of ozone scrubbers. 11

PAGE 24

Another sampler drew ambient air directly through the sampling cartridge without ozone removal. Sample analysis is accomplished using HPLC with ultraviolet/visible detection. Ozone deunder scrubber and KI cartridge scrubber were developed for this study and were tested for effectiveness of removal of ozone. Ozone interference and removal studies were conducted in the laboratory to evaluate DNPH silica gel technique. Ozone was mixed with the relative stable formaldehyde source, which is generated by a permeation device. The objective of this study was to investigate ozone interference on carbonyl sampling by 2,4-dinitrophenylhydrazine-HPLC ( DNPH-HPLC) method, and use a field study and laboratory data to evaluate the effect of ozone removal techniques. 12

PAGE 25

2. EXPERIMENTAL SECTION 2.1 Experiment for Field Study 2.1.1 Field Sampling Site Ambient formaldehyde, acetaldehyde and acetone concentrations were measured in the sampling station, which is located at Marine Street (2320 Marine Street) in Boulder. It is operated by the Colorado Department of Public Health and Environment. This sampling site is near downtown Boulder in a residential area near a well traveled two-lane street. 2.1.2 Pollutant Measurement Formaldehyde, acetaldehyde and acetone concentrations were measured from July 12 to September 2, 1996. Data were collected 24 hours per day, seven days per week. Ozone concentrations were measured simultaneously at the same location using a Dasibi Model 1 003-AH ultraviolet photometric analyzer. On the first four days (July12-July 16), the measurements were performed without adding ozone scrubbers. Four cartridges were used to cover the 4 four-hour periods, i.e., 4am-8am, 8am-noon, noon-4pm, and 4pm-8am. The fifth cartridge was 13

PAGE 26

used to cover 8pm-4am which is an eight-hour period. The flow solenoid assignment is shown in table 2.1. Table 2.1 Flow solenoid assignment Flow Solenoid number Sampling Period 0 4am-8am (4-hour) 1 8am-noon (4-hour) Day 1 2 noon-4pm (4-hour) 3 4pm-8pm (4-hour) 4 8pm-4am (8-hour) 5 4am-8am (4-hour) 6 8am-noon ( 4-hour) Day2 7 noon-4pm (4-hour) 8 4pm-8pm (4-hour) 9 8pm-4am (8-hour) 2.1.3 Sampling Equipment Sampling is performed through a sorbent cartridge, by controlling and measuring the flow through DNPH coated cartridge for a specified time period. A schematic of the sampler is shown in figure 2.1. 14

PAGE 27

pump Figure 2.1 flow controller Schematic of sequential sampler This is an automated low volume sequential sampler. It can hold tencartridges. Five cartridges were used to cover the four 4-hour and one 8-hour daily sampling period for one day and a second five cartridges were used for the second day. As can be seen in Figure 2.1, the sampler consists of a computerized control system that is capable of controlling solenoid opening and closing at specific times An eleventh cartridge was used as a field blank for the two day sampling period. Eleven cartridges 15

PAGE 28

were changed every other day. A diaphragm oil-less vacuum pump was used in the sampler to draw ambient air through the ozone scrubber and DNPH-coated cartridges. The total volume of air sampled through each solenoid was measured by a mass flow controller and the result was saved on the computer. The data was transfered from computer to field data sheet while changing the sample cartridges. 2.1.4 Ozone Denuder Scrubber As can be seen from figure 2.2, the ozone denuder scrubber consists of a 1/4inch o.d. copper tubing that has been coated internally with a saturated solution of potassium iodide (KI). The tube is coiled to a diameter of approximately two inches. The entrance and exit of the coil were fitted with a 1/4-inch brass bulkhead union and housed in a temperature controlled aluminum chamber. A cord heater was wrapped around the outside spiral ofthe coil. The ozone denuder scrubber was maintained at 52 o C during sample collection. Heating prevents condensation from occurring in the tube during sampling. The scrubber is connected to the inlet of the sample collection system. Sample air as extracted from a sample probe and distribution manifold and pulled through the scrubber by an oil-less vacuum pump. I.e of ozone in the sample air is converted by the chemical reaction below: 03 + 2KI + H20 02 + 12 +2KOH 16

PAGE 29

The ozone denuder scrubber is reusable. The copper tube should be recoated with a saturated solution ofKI after each six months of use [18]. tape Thermometer Figure 2.2 Schematic of ozone denuder scrubber 2.1.5 KI Cartridge Scrubber The ozone cartridge scrubber is a identical in size and shape to the regular DNPH Silica gel sampling cartridge, filled with approximately 1 gram of the KI coated silica. The scrubber is positioned at the inlet of the sample collection system. 17

PAGE 30

Sample air is extracted from the sample probe and distribution manifold and pulled through the ozone scrubber by an oil-less vacuum pump. Ozone in the sample air is converted by the chemical reaction described above. KI scrubber preparation: (1) Saturated solution ofKI was added to silica gel (70-230 mesh) (2) Mix well and then vacuum filter the mixture (3) The solid was then dried in a low temperature oven (70 a C) (4) Approximately 1g ofthe KI coated silica is packed in a sampling cartridge. Figure 2.3 Schematic ofKI cartridge scrubber 18

PAGE 31

2.1.6 Sampling Procedures The DNPH silica gel cartridges were prepared according to the procedure similar to that described by Tejada [19]. The dimensions of the body of the cartridges were about 1.25cm o.d. x 1 Ocm. Each cartridge was packed with approximately 0.60.8 grams silica gel (Chromatographic Specialties, 70-230 mesh). The silica gel packed cartridges were coated with acidified DNPH coating solution, which was made by dissolving twice recrystallized 2,4-DNPH in acetonitrile and acidified with HCL In the end, the DNPH coated cartridges were dried in a low temperature oven over night. When not being used for sampling, these cartridges were covered with plastic caps to prevent interaction with ambient air and kept in a freezer. Following each sampling period, sampled cartridges were kept in freezer prior to analysis. The sampled cartridges were extracted with 5ml of acetonitrile. The eluent was collected in a 5-ml volumetric flask. An HPLC (Varian 5050) was used for analysis. The DNPH derivatives were effectively separated using a 4.6mm I.D. x 150mm, reverse phase C18 column and using acetonitrile/water as mobile phase. The eluting components were detected with a UV absorbance detector operating at 360nm. 19

PAGE 32

2.2 Laboratory Studies 2.2.1 Collection Efficiency of Ozone Scrubbers Scrubber collection efficiency for ozone was determined by drawing known concentration of ozone generated from TECO Model 49 Ozone analyzer/calibrator through scrubber, and followed by measuring of ozone concentrations with Dasibi Modell 003-AH ultraviolet photometric analyzer. The background of ozone was 0.035ppm, and the zero offset of the ozone monitor was O.OlOppm. Collection efficiency was calculated as: C1 is ozone concentration before scrubber C2 is ozone concentration after scrubber The collection efficiency was 99.4% 0.6% for the ozone denuder scrubber and 99.8% 0.2% the KI cartridge scrubber, as shown in Tables 2.2 and 2 3. 20

PAGE 33

Table 2.2 Collection efficiency of ozone denuder scrubber [03 ] Before Scrubber (ppm) [03 ] After Scrubber(ppm) CE measured[03 ] corrected[ 03 ] measured [03 ] corrected [03 ] (%) 0.035 0.000 0.012 0.002 N/A 0.064 0.0029 0.010 0.000 100.0 0.089 0.0054 0.010 0.000 100.0 0.128 0.093 0.011 0.001 98.9 0.147 0.112 0.011 0.001 99.1 0.183 0.148 0.013 0.003 98.0 0.255 0.220 0.011 0.001 99.5 0.300 0.265 0.010 0.000 100 0.335 0.300 0.012 0.002 99.3 0.382 0.347 0.012 0.002 99.4 0.435 0.400 0.013 0.003 99.3 0.494 0.459 0.013 0.003 99.3 0.551 0.516 0.012 0.002 99.6 Average Standard Deviation 99.4 0.6 21

PAGE 34

Table 2.3 Collection efficiency ofKI cartridge scrubber [03 ] Before Scrubber (ppm) [03 ] After Scrubber(ppm) CE measured[03 ] corrected[ 03 ] measured [03 ] corrected [03 ] (%) 0.035 0.000 0.010 0.000 100 0 0.083 0.048 0.010 0.000 100.0 0.143 0.108 0.008 -0.002 N/A 0.188 0.153 0.010 0.000 100.0 0.230 0.195 0.010 0.000 100.0 0.278 0.243 0.011 0.001 99.6 0.325 0.290 0.010 0.000 100.0 0.377 0.342 0.011 0.001 99.7 0.424 0.389 0.012 0.002 99.5 0.475 0.440 0.011 0.001 99.8 0.559 0.524 0.011 0.001 99.8 Average Standard Deviation 99.8 0.2 22

PAGE 35

2.2.2 Lifetime Study ofKI Cartridge Scrubber The lifetime of the KI cartridge scrubber was also investigated in this study. The experiment was carried out by drawing 230ppb of ozone through the KI cartridge scrubber. The ozone concentration was monitored by using Dasibi Modell 003-AH ultraviolet photometric analyzer and a chart recorder After 48 hours, the measured ozone remained zero. This indicates that KI cartridge scrubber still can remove ozone effectively. Taking an average atmospheric ozone concentration as 40ppb, a KI cartridge will last at least 11 days(= 230 I 40 x 2). 2.2.3 Interference Studies Experiments were conducted to investigate the potential for ozone to interfere when sampling carbonyl compounds using DNPH-coated silica gel cartridges. Ozone was mixed with gas-phase formaldehyde and acetaldehyde. Ozone was generated by flowing air through a TECO Model 49 ozone analyzer/calibrator. The ozone was measured with a Dasibi Modell 003 ozone analyzer During the entire experiment, the ozone concentration was monitored at a manifold continuously. The gas phase formaldehyde or acetaldehyde was produced using a permeation device, which consisted of a porous Teflon-walled tube filled with paraformaldehyde or acetaldehyde. Zero air flowed continuously through the 23

PAGE 36

permeation tube generated constant concentration of or acetaldehyde. Zero air used in this experiment is "clean" air, it does not contain trace impurities that could contaminate the experiment. The permeation device was immersed into a temperature controlled water bath. The temperature was remained at 40C. Formaldehyde or acetaldehyde gas stream generated by the permeation tube was diluted with zero air. The concentration of formaldehyde or acetaldehyde was determined by permeation rate of the tubes at the bath temperature, the flow rate of gas past the permeation tube, and the flow rate of the diluent zero air. The experiments were performed with formaldehyde at concentrations typical of relatively clean air, that is, about 8 ppbv. Sampling was performed in three sets of experiments. The first set of the experiments sampled the mixture of formaldehyde, acetaldehyde and ozone. The second set of experiments sampled the mixture of formaldehyde and ozone. The last set of experiments sampled ozone only to see how it affected DNPH cartridges. For each set of these experiments sampling was performed with increments of ozone concentrations ofO, 20, 50, 100, 140,190, 230, and 320 ppb. Figure 2.4 shows the experimental set up for the interference study. 24

PAGE 37

pump flow meter zero au DNPH cartridge paraformaldehyde permeation tube exhaust t mass flow controller ozone generator L-----' Figure 2.4 Schematic diagram of the experimental set up of interference studies. 25

PAGE 38

3. DATA ANALYSIS AND RESULTS 3.1 Field Study DNPH silica gel cartridge technique was evaluated under ambient conditions during the summer sampling period, July 12 through September 2, 1996. During this study, three carbonyl samplers were used, one without any ozone removal device, one with a ozone denuder scrubber and the third with a KI cartridge scrubber, as shown below. Ambient carbonyl compounds with and without ozone removal were collected during the same time periods at Boulder. July 12-16 July 16-Sep 2 Sampler#1 without scrubber add KI cartridge scrubber Sampler#3 without scrubber without scrubber Sampler#4 without scrubber add denuder scrubber 3 .1.1 July 12-16 Period On the first four days, July 12-16, 1996 sampling period, no ozone scrubber was used. Three samplers were compared during this period. Formaldehyde concentration measured by sampler #4, which later used an ozone denuder scrubber, 26

PAGE 39

range from 0.78 to 8.08 ppbv. Acetaldehyde concentrations ranged from 1.94 to 30.86 ppbv. Acetone concentrations range from 7.04 to 86.57 ppbv. For sampler #1, which later used a KI cartridge scrubber, formaldehyde concentrations ranged from 0.81 to 4.10 ppbv, acetaldehyde concentrations ranged from 1.38 to10.21 ppbv, acetone concentrations range from 2.83 to 12.19. For sampler #3, which never used an ozone scrubber throughout the sampling period, formaldehyde concentrations ranged from 0.93 to 6.75 ppbv, acetaldehyde concentrations ranged from 1.89 to 17.94 ppbv, acetone concentrations ranged from 3.97 to 27.18 ppbv. Figure 3.1 shows the sampler intercomparison result. 27

PAGE 40

Average of Formaldehyde Concentration Cll 7 6 5 July12-16 c.---"' ....... -r-.... ....... / r-.... / ........ -g, 4 ..c:: Cll "'C ca E 0 u.. 1: IJ 1: 0 0 Cll "'C >-..c:: Cll "'C iii Cll u < 3 2 LL ----......... "=:::,.. / / ....... / / 04-08 08-12 12-16 16-20 20-04 Time Period ...... Sampler#4(add denuder later) .....,_ Sampler#1 (add Kl cartridge later) ...... Sampler#3(without scrubber) 20 15 / .'\. / '\. 10 / A '\. / / '"' 5 / 0 04-08 08-12 12-16 16-20 20-04 Time Period ...... Sampler#4(add denuder later) ...,_ Sampler#1 (add Kl cartridge later) ...... Sampler#3(without scrubber) Average of Acetone Concentration July12-16 1: u 1: 0 0 Cll 1: j 35 30 25 20 15 10 5 0 08-12 12-16 16-20 20-04 Time period ...,_Sampler #4 (add denuder later) ........ Sampler #1 (add Kl cartridge later) ...... Sampler #3 (without scrubber) Figure 3.1 Average concentration of carbonyl without ozone removal. 28

PAGE 41

In order to find the relationship between the carbonyl concentrations measured by sampler #4 (with ozone denuder later), and carbonyl concentrations measured by sampler #3 ( without ozone scrubber later ), regression analyses were performed to compare the results of carbonyl measurement data with different samplers. The sampler #1 (with Kl cartridge later) was also compared with sampler #3. The regression results are summarized in table 3 .1. Table 3.1 Carbonyl concentration correlation Sampler#4 I Sampler#3, July12-16 Slope R"2 Intercept Formaldehyde 1.17.34 0.74 0.23.13 Acetaldehyde 1.54 .41 0.78 0.60.01 Acetone 2.82.77 0.77 -4.42.65 Sampler#1 I Sampler#3, July 12-16 Slope R"2 Intercept Formaldehyde 0.48.25 0 .51 0 .99.8 7 Acetaldehyde 0 74.10 0.94 0.60. 7 7 Acetone 0.42.12 0.77 1.84 1.4 2 29

PAGE 42

Figure 3.2-4 shows the correlation of carbonyl concentrations measured from the different samplers. Correlation of Formaldehyde July12-July16 10 8 6 4 2 0 v Q) 0.. E ro C/) 0 1 v ..,. .,.... .. .,........-ti .., 2 3 4 5 6 Sampler#3 (unscrubbed) Correlation of Formaldehyde July12-16 6 5 4 Q) Q. 3 2 Cl) 1 0 ........ IlL .,..,. ,...,. !--""' .. 1- 1- 0 1 2 3 4 5 6 Sampler#3 (unscrubbed) 30 :,.... 7 8 7 8

PAGE 43

Figure 3.2 Correlation of formaldehyde, July 12-16 (],) 0.. E ctl (f) ....... 1!: (],) 0.. E ctl (f) Correlation of Acetaldehyde July12-July16 50 40 30 20 10 0 _.,....F' --_....... !!! ......... __ .. __.... ... 0 5 10 15 20 Sampler#3 (unscrubbed) Correlation of Acetaldehyde July12-July16 20 15 .........'-10 / 5 .2 ........... y 0 0 5 10 15 20 25 Sampler#3 (unscrubbed) Figure 3.3 Correlation of acetaldehyde, July 12-16. 31

PAGE 44

Q) c. E co Cf) 1!: Q) c.. E co Cf) Correlation of Acetone July12-16 100 80 60 40 20 v 25 20 15 10 5 0 ..,.. """ i/ v ... ./ 0 ... 0 5 1 0 15 20 25 30 Sampler#3 (unscrubbed) Correlation of Acetone July12-16 --i'"""' ..r.: 'Iii" _.... 0 5 1 0 15 20 25 30 Sampler#3 (unscrubbed) Figure 3.4 Correlation of acetone, July 12-16. Figure 3.5 shows the ratio of the scrubbed carbonyl concentration to the unscrubbed concentration versus time. 32

PAGE 45

Scrubbed/unscrubbed formaldehyde July 12-16 0 :;:::; 1.8 1.6 1.4 1.2 0.8 0.6 0.4 0 4-0 0-04 6-2 2-1 time oeriod denuder scn.Jooea Kl cartridge scrubbed Scrubbed/unscrubbed acetaldehyde JuiY12-16 0 3 2 5 2 1.5 :: r-...., F0.5 0 4-0 0-0 6-2 12-1 time period denuder scrubbed Kl cartridge scrubbed Scrubbedlunscrubbed acetone July 12-16 2.5 2 1. 5 0 5 0 08-1 04-0 0 16-2( timeBeriod denuder scr ooed Kl cartridge scrubbed 8-1 4-0 8 8-1 12-1E Figure 3.5 Ratio of scrubbed to unscrubbed carbonyl concentration versus time period. 33

PAGE 46

3.1.2 July 16-September 2 Period During July 16-September 2 period, ozone denuder scrubber and KI cartridge scrubber were evaluated. Ozone denuder scrubbed formaldehyde concentrations ranged from 0.77 to 9.39 ppbv. Acetaldehyde concentrations ranged from 1.12 to 27.71 ppbv. Acetone concentrations ranged from 2.22 to 41.35 ppbv. KI cartridge scrubbed formaldehyde concentrations ranged from 0.54 to 5.48 ppbv. Acetaldehyde concentrations ranged from 0.87 to 14.38 ppbv. Acetone concentrations range from 1.23 to 20.94 ppbv. Unscrubbed formaldehyde concentrations ranged from 0.84 to 7.23 ppbv. Acetaldehyde concentrations ranged from 0.66 to 21.74 ppbv. Acetone concentrations ranged from 0.08 to 39.20 ppbv. Table 3.2, table 3.3 and table 3.4 show the averages of carbonyl concentrations for the different time periods of the day for the July 16-September 2, 1996 sampling period. Figure 3.6, 3.7, and 3.8 show the diurnal character for three carbonyl concentrations for the three samplers. Data were collected from five time periods per day, starting at 4am 8am, followed by 8 am noon, noon 4pm, 4pm 8pm, 8pm 4am. Formaldehyde, acetaldehyde and acetone concentration reach a maximum during noon 4pm time period. 34

PAGE 47

Table 3.2 Average concentrations of formaldehyde TIME 04-08 08-12 12-16 16-20 Denuder 2.15 4.54 6.13 4.67 KI 1.43 3.03 3.69 3.09 Unscrubbed 2.17 4.02 4.89 3.89 Average of Formaldehyde Concentration c 0 c 0 7 6 / // July16-end / / / ..,.. ........... .... ........... ..,.. ""' .......... 0 Cl) "C >. ..c Cl) 5 4 3 2 1 // _..---...... "C ns E .... 0 u.. r / / / "' 04-08 08-12 12-16 16-20 20-04 Time Period ---Sampler#4(with denuder scrubber) .._..._ Sampler#1 (with Kl cartridge scrubber) Sampler#3(without scrubber) 20-04 2.99 3.35 2.16 Figure 3.6 Average concentration of formaldehyde (Julyl6-Sep2) 35

PAGE 48

Table 3.3 Average concentration of acetaldehyde TIME 04-08 08-12 12-16 16-20 Denuder 3.34 8.61 15.26 9.14 KI 2.85 5.09 10.01 7.05 unscrubbered 2.13 6.49 11.79 7.11 Average of Acetaldehyde Concentration July16-end r:::: 0 :;:::: f.! ... r:::: Cl) (.) r::: 0 0 ... Cl) (.) <( 16 14 12 10 / / // / / / / / / / // // // / / "' ' ' ........ "' ........... ...................... 1"-"""' i""' ............................... 8 6 4 2 / /...,.,-'"" """" v "' 04-08 08-12 12-16 16-20 20-04 Time Period Sampler#4(with denuder scrubber) ....... Sampler#1 (with Kl cartridge scrubber) Sampler#3(without scrubber) Figure 3.7 Average concentration of acetaldehyde (July16-Sep2) 36 20-04 3.93 4.00 2.24

PAGE 49

Table 3.4 Average concentration of acetone TIME 04-08 08-12 12-16 16-20 20-04 Denuder 7.02 11.56 22.23 13.52 7.08 KI 5.23 5.89 10.86 9.91 5.91 unscrubbed 4.08 8.72 16.62 10.24 4.04 Average of Acetone Concentration c: 25 July16-end 0 -.... 20 J cu ... / "' .... c: / / "' Q) 15 (.) // c: 0 10 ./ / "" u .// / Q) / v """" c: 5 0 t' .... Q) 0 (.) <( 04-08 08-12 12-16 16-20 20-04 Time Period -11-Sampler#4(with denuder scrubber) ..,._ Sampler#1 (with Kl cartridge scrubber) Sampler#3(without scrubber) Figure 3.8 Average concentration of acetone (July16-Sep2) In order to detect if there was a significant difference between the scrubbed cartridge concentrations and unscrubbed cartridge concentrations, scrubbed versus 37

PAGE 50

unscrubbed concentration regression analyses were performed to compare the results of carbonyl measurement data for the three different samplers. Table 3.5 shows carbonyl compounds concentration correlation results. Table 3.5 Carbonyl concentration correlation during July 16September 2 sampling period. Denuder scrubbed /unscrubbed, July16-Sep 2 Slope R"2 Intercept Formaldehyde 1.12.11 0.71 0.44.41 Acetaldehyde 1.27.13 0.69 1.01 .95 Acetone 1.15.20 0.45 4.18.01 KI Cartridge scrubbed I unscrubbed, July 16-Sep 2 Slope R"2 Intercept Formaldehyde 0.39.11 0.23 1.46.4 1 Acetaldehyde 0.73.05 0.83 1.32.3 8 Acetone 0.48.07 0.54 3.33.7 8 Figure 3.9-3.14 show the correlation of carbonyl concentrations measured from the sampler with ozone denuder or KI cartridge scrubber to the sampler without scrubber. These measurements were made from July 16 to September 2. 38

PAGE 51

"0 (]) .c .c ::J ..... () Cf) ..... (]) "0 ::J c (]) 0 Correlation of Formaldehyde July16-end 0 0 1 2 3 4 5 6 7 8 Unscrubbed Figure 3.9 Correlation of formaldehyde, ozone denuder scrubbed to unscrubbed concentration "0 (]) ..c ..c :::::1 L.. (.) (f) Q) C) "0 c t co (,) Correlation of formaldehyde July12-end Kl Cartridge/Unscrubbed 6 5 4 "':' 'x )< _;.,_ k---. 0 1 2 3 4 5 6 7 8 Unscrubbed Figure 3 .10 Correlation of formaldehyde, KI cartridge scrubbed to unscrubbed concentration 39

PAGE 52

"'0 Q) .c .c :::J .... (.') C/) .... Q) "'0 :::J c Q) 0 Correlation of Acetaldehyde July16-end 50 40 30 20 10 0 0 X X X X. X X '><0> < l x :._.;:;>w-x'x X X X ">X .x. 5 10 Unscrubbed 15 20 Figure 3.11 Correlation of acetaldehyde ozone denuder scrubbed to unscrubbed concentration .... Q) .c .c :::J .... (.) en :::.:::: Correlation of Acetaldehyde July16-end 20 15 y "' >-y x: 10 ,:;z X, x_z? <": 'X >< jx'X% p<#' 5 0 0 5 10 15 Unscrubbed ./ v X 20 25 Figure 3.12 Correlation of acetaldehyde, KI cartridge scrub bed to unscrub bed concentration. 40

PAGE 53

"'C Q) ..c ..c 2 (.) (/') 'Q) "'C ::J c Q) 0 Correlation of Acetone July16-end Denuder/Unscrubbed 100 80 60 40 X X' X 0 5 10 15 20 25 30 Unscrubbed Figure 3.13 Correlation of acetone, ozone denuder scrubbed to unscrubbed concentration. "'C Q) ..c ..c ::J '-(.) (/') Q) 0) "'C ;:: -e ro (.) Correlation of Acetone July16-end Kl Cartridge/Unscrubbed 25 20 15 10 5 0 X '*-X X X XX X --" x, X ,)
PAGE 54

To evaluate the ozone effect on formaldehyde acetaldehyde and acetone measurement, a plot of carbonyl ratios for scrubbed I unscrubbed versus ozone concentration are shown on figure 3.15 17. Correlation of HCHO Ratio to Ozone Denuder Scrubbed/Unscrubbed "C Q) .0 .0 .... u (/) c: :::> -"C Q) .0 .0 :::J .... u Cl) "C Q) .0 .0 2 u (/) c: :::> =c Q) .0 .0 2 u Cl) 8 6 4 2 + 0 -1-" .;. .. -+ + ..... 0 20 40 60 Ozone Concentration (ppb) Correlation of HCHO Ratio to Ozone Kl Cartridge Scrubbed/Unscrubbed 80 8 6 4 2 .,..f .. t-..... -ft.;--+0 +-... "'"" _' ._,___ -l. _;.;r __ ..---1 0 20 40 60 80 Ozone Concentrion (ppb) Figure 3.15 Ratio of scrubbed to unscrubbed formaldehyde concentration versus ozone concentration. 42

PAGE 55

Correlation of CH3CHO Ratio to Ozone Denuder Scrubbed/Unscrubbed "'0 1 0 -r--r--r-----r--r--..---.--...-----, Q) .0 8 .0 :l ..... C) 6 en r:: ::::.> -4 "'0 Q) .0 .0 2 :l ..... C) C1) 0 r '' r t-o- tf-+ .,. ... -t-. ::+;: ; . ; .._ 0 20 40 60 Ozone Concentration (ppb) Correlation of CH3CHO Ratio to Ozone Kl Cartridge Scrubbed/Unscrubbed i-80 10 "'0 Q) .0 8 .0 :l ..... C) 6 en r:: ::::.> -4 "'0 Q) .0 .0 2 :l + ..... C) C1) 0 0 20 40 60 80 Ozone concentration (ppb) Figure 3.16 Ratio of scrubbed to unscrubbed acetaldehyde concentration versus ozone concentration. 43

PAGE 56

Correlation of Acetone Ratio to Ozone Denuder Scrubbed/Unscrubbed "C Q) .c .c 2 (.) en c:: ::J "C Q) .c .c :::J ,_ (.) Cf) 8 6 4 2 .. S$+ ;:t=J + .......... -........ -::-_.,_ . + .. ,.. --r ;: + + 0 0 20 40 60 80 Ozone Concentration (ppb) Correlation of Acetone ratio to Ozone Kl Cartridge Scrubbed/Unscrubbed "C 1 0 ...---..---.---,-----.---.---.---..........----. Q) .c 8 .c :::J ,_ (.) 6 en + c:: ::J -4 "C Q) .c .c 2 :::J + + + + + ,_ (.) Cf) 0 -;-1-; 0 20 40 60 80 Ozone Concentration (ppb) Figure 3.17 Ratio of scrubbed to unscrubbed acetone concentration versus ozone concentration. The correlation results are shown in table 3.6 44

PAGE 57

Table 3.6 Scrubbed I unscrubbed ratio versus ozone concentration. Denuder scrubbed/unscrubbed ratio vs. ozone Slope R/\2 Intercept Formaldehyde 0.0043 .0016 0.0355 1.1.3 Acetaldehyde -0.0067 0.0038 0.0177 1.8.7 Acetone -0.0156.0490 0.0006 3.3.1 KI Cartridge scrubbed/unscrubbed ratio vs. ozone Slope R/\2 Intercept Formaldehyde 0.0004.0016 0.0004 0.8.3 Acetaldehyde -0.0088.0028 0.0523 1.5.6 Acetone -0.0190.0232 0.0040 2.2.5 3.2 Results from laboratory interference studies In our study, a formaldehyde and acetaldehyde generating system was constructed to deliver formaldehyde and acetaldehyde continuously. Five replicate experiments were conducted to determine the concentrations of formaldehyde and acetaldehyde delivered from the system. In these experiments, a DNPH silica gel cartridge was used to collect carbonyl compounds for 180 min. The volume of air 45

PAGE 58

sampled for each cartridge was measured by a dry test meter. The replicate experimental results are given in table 3.7. Table 3.7 Laboratory formaldehyde and acetaldehyde concentrations. Formaldehyde, ppb Acetaldehyde, ppb Sample 1 8.10 48.38 Sample 2 7.42 47.19 Sample3 8.49 48.20 Sample 4 7.87 47.95 MeanSD 7.97.44 47.93.52 In order to investigate ozone interference on carbonyl measurements, various concentrations of ozone were mixed with the gas-phase formaldehyde and acetaldehyde. Three sets of experiments was carried out. The first set of experiments sampled the mixture of formaldehyde, acetaldehyde and ozone. For each experiment, sampling was performed with increments of ozone ofO, 20, 50, 100, 140, 190, 230, 320 ppb. Figure 3.18 shows the formaldehyde concentration measured in the presence of additional ozone and that measured after going through ozone denuder or K.I cartridge scrubber as a function of the amount of ozone added. Each data point 46

PAGE 59

represents the formaldehyde concentration with either ozone removal device or without ozone removal. c 0 :;::; c Q) (.) c 0 (.) Q) Q) "'0 ro E ..... 0 u. Formaldehyde Vs. Ozone Data Set 1 10 8 6 4 2 0 Ia n - ... """ .... ... ,. 0 50 1 00 150 200 ... ill' 250 Ozone Concentration unscrubbed e Kl-scrubbed de-scrubbed 1 1-1-300 350 Figure 3.18 Formaldehyde concentration from experiment set 1 as a function of ozone added. Figure 3.19 is a plot of acetaldehyde measured in the presence of additional ozone, and acetaldehyde measured following an ozone scrubber that removes ozone as a function of the amount of ozone added. In the presence of 20 ppb ozone, the chromatograms of ozone unscrubbed samples start to show a little peak with retention times shorter than formaldehyde. The peak increased in proportion to increasing ozone concentration. The measured unscrubbed formaldehyde and acetaldehyde concentrations in the presence of20ppb of ozone was 7.76 ppb and 48.24ppb, indicating no carbonyl loss. At 50 ppb ozone between 18%-27% of formaldehyde and 47

PAGE 60

19%-23% of acetaldehyde were not measured. At 320 ppb ozone between 86%-88% of formaldehyde and 83%-84% of acetaldehyde were not measured. c: 0 :.;::::; ...... c: Q) (.) c: 0 0 Q) .c: Q) '"0 -rn ...... Q) Acetaldehyde Vs. Ozone Data Set 1 60 50 40 30 20 10 0 .. .. !I!! ,.. .. -... .. I"": Ia. .... 0 50 100 150 200 250 300 350 Ozone Concentration A unscrubbed e Kl-scrubbed de-scrubbed Figure 3.19 Acetaldehyde concentration from set 1 experiments as a function of ozone added. In the laboratory study, the concentration of acetaldehyde in the sample air was about six times higher concentration than formaldehyde. So ozone may react with acetaldehyde or its derivatives at higher rate than equal concentration of formaldehyde and acetaldehyde. This would inhibit the reaction of ozone with formaldehyde or its derivatives, and make it hard to identify the interference of ozone on measurement of formaldehyde. In order to detect if there was a different ozone effect on measurement of formaldehyde due to high quantity of acetaldehyde present, 48

PAGE 61

the second set of experiments was conducted. In this set of experiments, only gas phase formaldehyde and ozone was added in the sample air. Figure 3.20 is a plot made for this second set of experimental data. It shows the formaldehyde concentration measured in presence of additional ozone, and formaldehyde measured after going through an ozone scrubber as a function of ozone added. c: 0 -c: Q) (.J c: 0 () Q) ..c: Q) "C ro E 0 u.. Formaldehyde Vs. Ozone Data Set 2 10 8 6 4 2 0 u .. -- .... -... E -.. 0 50 100 150 200 250 300 350 Ozone Concentration A unscrubbed e Kl-scrubbed de-scrubbed Figure 3.20 Formaldehyde concentration from set 2 experiments as a function of ozone added. From second set ofthe experiments, the measured unscrubbed formaldehyde concentrations in the presence of20 ppb was 8.82 ppb, indicating no formaldehyde loss. At 50 ppb ozone between 9%-18% of formaldehyde is not measured. At 320 ppb ozone 99% of formaldehyde was not measured. 49

PAGE 62

In order to find the formaldehyde differences between two sets of experiments regression analyses were made to compare the results. The regression results are summarized in table 3.8 Table 3.8 Comparison of formaldehyde concentration from two set data. Denuder scrubbed formaldehyde Slope R"2 Intercept Data set 1 0.002.002 0.19 7.3.6 Data set2 0.001 .002 0.08 8.1.5 KI cartridge scrubbed formaldehyde Slope R"2 Intercept Data set 1 -0.002.001 0 27 8.8.6 Data set 2 0.002.003 0.07 8.5.9 Unscrubbed formaldehyde Data set 1 -0.020.002 0.97 7.4.5 Data set 2 -0.024.002 0.96 8.3.6 In order to investigate the products of ozone reaction with DNPH silica gel cartridge, a third set of experiments was performed. In this set of experiments, DNPH silica gel cartridges were exposed to different ozone concentrations. Shown in Figure 3.2023 are the chromatograms of the 2,4-DNPH cartridge exposed to various ozone 50

PAGE 63

concentrations when carbonyl compounds are absent. The ozone concentration in Figure 3.20 through 3.23 is 0, 100, 190, and 320 ppb, respectively. It can be seen that DNPH was destroyed by ozone, and several additional peaks (#1-4) were found compared with Figure 3.22 ([03 ] = 0 ppb). The results are summarized in Table 3.9, where the integrated areas of the four peaks are tabulated at various ozone concentrations. It clearly shows that the area count of the additional peaks increases with increasing ozone concentration. Table 3.9 Summary of the area changes with added ozone. Integrated Area [Ozone] Peak 1 Peak2 Peak3 Peak4 ppb pre-HCHO HCHO pre-CH3COCH3 x-DNPH* 0 7.75t 55.09 26.70 0 20 10.89 89.68 73.46 13.47 50 46.99 89.31 155.55 75.37 100 80.83 102.66 209.01 89.10 140 105.03 86.35 389.64 147.57 190 133.96 92.70 361.92 185.34 230 154.96 74.55 423.05 155.18 320 182.49 62.70 287.01 165.32 x-DNPH is degradation product ofDNPH, Its retention time is around 14.5 minute. t Not properly integrated, should be 0. 51

PAGE 64

f I L l r l I Lab ..... 1 1711.1.,. ol! Colorado .... ., Aa&lyJ. date 1 05/24/1JJ7 12101110 X.tbod 1 Autoa.plea: 10 aica:oliter Dec:a:iptioa I ur.c: IN .Aborptioa 1 150 X 4.C .. C11 .......... C&ZTiea: I laa:&c!iUit CIDOI 1120 Date f11e 1 Aftl2,ASC (dl\peak2\date) la.ple 1 Auraa:ia Sa.ple 1JJ5 .400mV u.ooomv ..,.,_,, ..,...._,, ..,...._,, ..,.,_,, ..,...._,, .,. lluliblt let:lnt I on ...... 0 0.133 IZ.I2 I.DCO 41.711 1.516 sa.tt z.zso 17.51 Z.l16 t.71 3.516 14766.ft. 5.416 27.711. 5.713 7.71. 3 6.ZI6 ''D:9. 7.566 7.41 7.'50 s.ea 0 1.616 ;;:,,, 10.766 26.711 11.40 16.69 IZ.416 5.47 IZ 900 10.45 u .su 1.13 14.133 7.3l 14.683 6.74 a IS.Z66 13.56 a IS.no 7.59 Figure 3 .21 Chromtogram of the sample cartridge with no added carbonyls after exposure to 0 ppb ozone. 52

PAGE 65

r i i Lib ,.... 1 lhl of Colozado lluYU Aaalyt dat 1 01/01/1117 1S120111 X.t!l.od 1 Aato .. 10 llico:oliter u .. c:npttoa 1 ur.c: vv .NIorpUoa Col-1 150 & 4.1 Cll a.YV c.J:rier 1 Gradbat CIDCir 1120 U.ta fila 1 .UXU.ASC: (41\paak2\datal lupl 1 AIIE'uia lupl ... 1US .400mV u.ooomv . '-' _, ...,_,, ...,_,, ...,_,, cite .... .._,, ..-,, ...,_,, _,.,.,. ........ ,. ...,_,, rtActfone ...,_,, ...,_,, ...,_,, _, _, ...,_,, l\lllber ltttlftcton AI-H 0 O.Z66 ZS.OJ 0.1'66 R.71 z.zso 1480.14 3,100 117Z4.4Z o ,766 110,113 S.lOO 10Z.66 6.000 6.31 6.713 1l.61 7.ZSO 17. 59 7.666 1.900 Z09.01 0 10.ZI6 Z0.67. 6 11.150 z,. n 0 1Z.OOO Z4.ZO 0 1Z.616 ZJ.ZJ 0 1l.066 6 97 0 ll.SIU Z6." 0 "m 119.10. 0 15.m 19.76 Figure 3.22 Chromatogram of the sample cartridge with no added carbonyls after exposure to 100 ppb ozone 53

PAGE 66

:-Lab a.aa 1 lrllJ.Y of Colorado U...Yu AAalye1e date 1 05/01/1117 14132140 ICel:!lod 1 Autoea.p1eo: 10 aL=oUtar Deec:..-1pU01l 1 ULC UV AbeorpUaa Coluaa 1 150 z '' .. C11 a.Yeo:ee Ca=Lar 1 QraclL&Ilt CIDar H20 Data f11 1 .UIDI .UC ld1\peak2\d&UI Saaple 1 Auruia Sa.plee 1115 .4CCmV N 64 000mV c_, '"*'-'l '"*'-'l '"*'-'l '"*'-'l '"*'-'l '"*'-'l '"*'-'l '"*'-'l '"'-'l '"'-'l '"'-'l reAcetont reAntOftl C41:0fte oetActtane '"*'-'l ...,_,, ...,_,, ....,_,l ...,_,, lulblr .. t.,tlon Art I a a.m Z0.07 a a.9116 14S. :rz 1.SSO 8.62 Z.450 2807.00 3.316 118Zl.Z5 4.216 42. 62 4 466 S9.80 5.2 .. 133. 96 S.ll3 9Z.l'Cl 6 713 11.43 8.016 Sl.12 8.616 7.00 9,916 l61 .92 1a.l66 Z0.91 11.lGG 17.41 7 11.613 9.30 12.613 10.61 1Z.90G 33.lG 13.400 7.S7 13.1150 Z6.41 14.433 1115.:14 15.316 16.49 Figure 3.23 Chromatogram of the sample with no added carbonyl after exposure to 190 ppb ozone 54

PAGE 67

Lab .,._e 17G1Y of Colo.:ado DeY..: Analyaia date OS/2,/1117 OloSS120 Katbod Autoampla.: 10 a1c.:ol1ta.: Daac.:iptioa I &ILC av Abao.:ptioo ColUDA 150 x 4 .. Cll Rever Ca.:.:ia.: OracSimc CBlar 820 Data Ula r .UIQ4.ASC (dl\paak2\datal Sample 1 Aura.:ia Seaple lt'S .400mV N ,., 54.000mV c-c o.Nr-.1 '"'-I o.Nr-.J '"'-I '"'-I o.Nr-.J ..,.,_,, ..,.,_,, azide ...... lddlyde ..,.,_,, ..,.,_,, ...,_,I ...,_,I ottAcecane ..,.,_,, ...,_,I -I ..,.,_,, ...,_,I ltYIDir lecenc fan a a.zaa Aru IZ.68 t.aaa Z0Z.99 a Z.Z66 43:15.71 a J.U6 Jm.ZZ a l.IOO 1169.110 4,446 4G.71 4.716 46.31 5.650 I&Z.49 6.ZSO 6Z.70 7.566 12.64 1.666 lZ.79 ta.7QQ Zl7.01 11.4GG 16.110 0 11.966 19.91 IZ.lll 16.SI 0 IZ.m 7.07 13.166 49,21 0 13.113 9.1Z 0 14.ZI3 16. 36 14.166 165.JZ Figure 3.24 Chromatogram of the sample with no added carbonyl after exposure to 3 20 ppb ozone. 55

PAGE 68

4. DISCUSSION 4.1 Field Study 4.1.1 Sampler intercomparison period As described above, carbonyl compounds were measured by three different samplers at same time and same condition. The samplers were calibrated before putting them in the field, and a leak test was performed every other day, before and after changing sample cartridges. On the first three days of the sampling period, no ozone scrubber was used. The concentrations measured by the three samplers should be the same value. But as can be been seen from figure 3.1, formaldehyde, acetaldehyde and acetone concentrations measured by the different samplers are not the same. This indicates that there was a calibration problem in the system. Since we didn't recalibrate the samplers right after finishing the field study, the calibration problems remain uncertain. However, we have some clues about the cause of this problem from the field data sheets and later laboratory study. It appears that something was not quite right, about the unstable flowrate reading on each solenoid. Sometimes the flowrate reading on one of the solenoids dropped a lot. This might be related to the fluctuation of the voltage. 56

PAGE 69

Although carbonyl concentrations from three samplers can not be compared to each other directly, the ozone effect on carbonyl compounds can be seen from the ratio ofthe scrubbed to unscrubbed carbonyl concentration, which are summarized in Table 3.1 and 3.5. 4.1.2 Ozone Removal Evaluation Correlation of formaldehyde concentrations with ozone removed to that without ozone removed. Figure 3.9 and 3.10 show the formaldehyde concentrations measured using ozone denuder scrubber and KI cartridge scrubber, plotted against the corresponding amounts ofunscrubbed formaldehyde data. The slope, intercept and R2 of the best fit lines are summarized in Table 3.1 and Table 3.5. The uncertainties represent the 95% confidence levels for the slope and intercept. As can be seen from the tables, when ozone is removed by denuder scrubber the slope of the regression line is within the 95% confidence range as compared with the measurements without scrubber (from 1.17.34 to 1.12.11). The R2is about the same also, from 0. 7 4 to 0. 71. Thus, the degree of correlation remains unchanged despite the large difference in data collection ( 4 days versus almost 7 weeks ). The intercept changes from 0.23.13 to 0.44.41, well within the 95% confidence level. This indicates that assuming the unscrubbed data are without systematic error, 57

PAGE 70

formaldehyde concentration is not affected by the ozone denuder scrubber when sampling with DNPH silica gel cartridges. When ozone is removed by KI cartridge scrubber the slope of the regression c line is not changed a lot, from 0.48.25 to 0.39.11. The intercept is from 0.99.87 to 1.46.41. These data indicate when sampling with DNPH silica gel cartridges, formaldehyde concentration didn't change significantly after using the KI cartridge scrubber. However, the R2 decrease from 0.51 to 0.23 suggests that using the KI cartridge scrubber caused scatter in the data. There was an extra peak found between formaldehyde and acetaldehyde from field sample chromatogram which was shown in figure 4. 7. This indicates an interference to DNPH silica gel method exists when using KI cartridge scrubber. Correlation of acetaldehyde concentrations with ozone removed to that without ozone removed. Figure 3.11 and 3 .12 show acetaldehyde concentrations measured using ozone denuder scrubber and KI cartridge scrubber, plotted against the unscrubbed acetaldehyde data. The slope, R2 and intercept of the best fit lines are summarized in Table 3.5. The uncertainties represent the 95% confidence levels for the slope and intercept. The table shows that when ozone is removed by denuder scrubber the slope of the regression line is not significantly changed, from 1.54.41 to 1.27.13. The R2 58

PAGE 71

does not change much either, from 0.78 to 0.69. The intercept changes from 0.60.01 to 1.01 .95, well within the 95% confidence level. These indicate when sampling with DNPH silica gel cartridges, the acetaldehyde concentration did not change significantly whether using ozone denuder scrubber or not. When ozone is removed by KI cartridge scrubber, the slope of the regression line is almost the same, from 0.74.10 to 0.73.05. The R2 does not change much, from 0.94 to 0.83. The intercept changes from 0.60.77 to 1.32.38, well within the 95% confidence level. These indicate that when sampling with DNPH silica gel cartridge, the acetaldehyde concentration did not change significantly whether using a KI cartridge scrubber or not. Correlation of acetone concentrations with ozone removed to that without ozone removed. Figure 3.13 and 3.14 show the acetone concentrations measured using ozone denuder scrubber and KI cartridge scrubber, plotted against the unscrubbed acetone data. The slope, intercept and R2 of the best fit lines are summarized in Table 3.5. The uncertainties represent the 95% confidence levels for the slope and intercept. As can be seen from the table, when ozone is removed by denuder scrubber, the slope ofthe regression line is decreased significantly, from 2.82.77 to 1.15.20. The R2 is decreased from 0.77 to 0.45. The intercept is not changed significantly, from -4.42 .65 to 4.18.01, within the 95% confidence level. 59

PAGE 72

These indicate adding ozone denuder scrubber caused low acetone concentration reading and scattering data. Therefore ozone denuder interferes with acetone measurement when using DNPH silica gel method. When using KI cartridge scrubber to remove ozone, the slope of the regression line does not significantly change, from 0.42.12 to 0.48.07. The intercept increases from 1.84 1.42 to 3.33 .78. But this change is not significant, still within the 95% confidence leveL These indicate when ozone was removed by KI cartridge scrubber, the acetone concentration measured by DNPH silica gel method did not change significantly. But the R2 decreased from 0.77 to 0.54, as can be observed the data scatters more in Figure 3.14. This means using KI cartridge scrubber causes an interference in measurement of acetone by DNPH silica gel technique. Ozone concentration dependence of the ratio of the scrubbed to unscrubbed carbonyl concentration. Figure 3.15-17 show the ratio of scrubbed to unscrubbed carbonyl concentration as a function of ozone concentration. The best fit lines are essentially parallel to the x-axis, indicating scrubbedlunscrubbed ratio does not have dependence on the ozone concentration. As shown in Tables 2.2 and 2.3 which will be discussed in detail in chapter 4.2, the denuder scrubber and KI cartridge scrubber efficiently remove ozone. The ratio of scrubbed to unscrubbed results would reveal any interaction between the ozone and carbonyl compounds. That is if ozone interference cause measured carbonyl concentration to decrease, the scrubbed I 60

PAGE 73

unscrubbed ratio should increase as ozone concentration increases. The lack of correlation between scrubbed/unscrubbed ratio and the ozone concentration suggests that ozone does not produce significant interference to carbonyl compounds over in these measurements. 4.2 Laboratory interference study As can be seen from Figure 3.24, the chromatogram of sample with no added carbonyl clearly shows that ozone (320ppb) reacts with 2,4-DNPH on the surface of silica gel. By comparison with Figure 3.21, several extra peaks are seen in the chromatogram. Peak #1,2,3 and 4 are the results of decomposition products from the reaction of ozone with 2,4-DNPH. Peak #2 is identified as formaldehyde, other peaks are unknown. In general, these peaks increased proportional to increasing ozone concentration, as shown by Figure 3.21, 3.22, 3.23 and 3.24. The DNPH reagent loading in the extract was found to be significantly lower after exposure to ozone. In first and second sets of the experiments, carbonyl compounds were added to the gas stream. The measured formaldehyde and acetaldehyde concentrations did not significantly change in presence of 20 ppb ozone, but the concentrations decreased as ozone increased in the range from 50 ppb to 320 ppb of ozone, when no ozone removal device was used. This indicates a negative ozone interference exists, which means the measured carbonyl concentration is lower than the actual value. 61

PAGE 74

Ozone might react with the formaldehyde and acetaldehyde derivatives on the surface of silica gel. The chromatogram of sample with or without carbonyl added are very similar, only the size ofthe carbonyl peaks are different, can be seen from figure 4.1 and 4.2. This indicates formaldehyde derivative react with ozone in the same manner as DNPH, they yield same products. The rate constant for the reaction of ozone with formaldehyde or acetaldehyde 2,4-DNPHydrazone is unknown, but the gas phase ozone reaction for the formaldehyde hydrazone (CH2=NNH2 ) is 2.5 X 10"17 cm3 molecule1 s1 [9]. The rate constant for ozone reaction with formaldehyde and acetaldehyde are very slow, with rate constants of <1 o-20 cm3 molecule1 s1 at 298K [11]. So these reactions are negligible under atmospheric conditions. Table 3.8 shows the comparison of formaldehyde data from set 1 and 2. As can be seen from the slope, intercept of the best fit line of formaldehyde concentration versus ozone, there is no significant change between two sets of data. This means that high fractions of acetaldehyde ( [acetaldehyde] : [formaldehyde] ::.::! 6 ) did not affect the ozone effect on formaldehyde measurement. As can be seen in figure 3.18 to 3.20, the ozone denuder scrubbed or K1 cartridge scrubbed carbonyl concentrations do not change as ozone concentration increases. This indicates the ozone denuder and KI cartridge scrubber can be used to remove ozone effectively. As a result, the measured concentration of carbonyl 62

PAGE 75

Lab aam Univ of Colorado Deuvr Aaalyaia d&t 1 06/01/1tt7 14102151 Mthod 1 Autoa.-p1er 10 aicrolitr Dacriptioa 1 HPLC UV Abaorptioa Column 1 150 x 4,6 .. Cl8 Rverae Carrier 1 GracUeat CIOOI H2v Data fil 1 APH38 .ASC (41\pealtl\d&ta) Sample Auraria Sampla 1ttS -6.400mV 64.000mV CQII!)Onel'lt Muobet' htont I an Aree ..w-nJ 0 o.m 0 0.966 0 1.533 '------------------!< 0 2.450 ............,, 0 3.166 ............,, 0 4,266 IW:I'IOY\) 0 5.050 lril>aton) 0 5.600 2ezlde l 6.466 ezlde l 6.650 7.283 4 7.733 0 1 .666 0 8.916 5 9.600 0 10.700 6 11.200 7 11.616 a 12.216 \6lknown) 0 12.816 lril>aton) 0 13.316 8 U:ti! ............,, 0 14.400 l:fWISIII) 8 U:!U Figure 4 1 Chromatogram of the sample with no added carbonyls after exposure to 23 0 ppb ozone 63 18.05 154.97" 11.57 II 3\39.15 IC n50.76" 3a.92 154.96 II 74.55 II 7.30 II 7.45" 16.14 II 44.\l II 10.16 6.51 423.05 14.04 21.00 8.42 7.73 54.94 13.70 155.18

PAGE 76

Lab D- 1 'CI'IIiv of Colorado Davr Azl&ly-.1 d&t I 06/0J/1!97 13 I 08: ll ICathoci 1 Autouzpler 10 aJ.c:roliter 1 BPLC uv Aborptioa Column 150 x ' .. C18 Rever C&rrier 1 Qrac1ieat calor -JUO Data file 1 ARN.l.ASC (41\peakl\d&ta) Sample 1 AU:aria Sa.pl 1995 -6. 400mV 64.000mV ....,_,, ...a-.) ...a-.) ....,_,, ....,_,, ....,_,, ....,_,, on. tdellyde ....,_,, reAcetone .pi" I on 36.4! a a.600 17.23 0 O.Nl 109.39 0 1.650 15.01 a z.oa:s 147.9l II a Z.TOCI 12116.46 II a 1.066 l454.9l II 0 l .&al 112.41 II z 4.asa 69.a7 II a S.lal 160.&l II a 6.916 1z.as 11 4 7.566 25.09 II 0 7.95a 5 .6711 0 9.4]3 276.55 II 0 1a.z()() 5.46 II 0 10.533 19.16 II 6 11.166 ll.04 II 7 11.616 17.()() a 12.350 6.69 0 12.933 63.16 a 13.516 tZ.l9 0 ll.&al 10.96 0 14.566 71.67 0 15.5()() 9.az Figure 4.2 Chromatogram of the sample of the sample with formaldehyde after exposure to 23 0 ppb ozone. 64

PAGE 77

compounds after using denuder scrubber or KI cartridge scrubber exhibit no dependence on the added ozone concentration. Furthermore, use of ozone scrubbers does not introduce any contamination to carbonyl compounds, since noadditional new peaks are observed in the chromatogram in the samples collected after the scrubber. 4.3 Comparison of field and laboratory study The laboratory study shows that a negative interference exists at an ozone concentration of 50 ppb or above. However, from field study we have found no ozone interference on measurement of carbonyl compounds when using DNPH silica gel method. As we learned from laboratory study, the magnitude of ozone interference depend on the concentration of ozone sampled Ambient ozone concentrations ranged from 6.25 to 74.25 ppb with significant diurnal variation during the field sampling period, it reaches a maximum from noon to 4pm. The average ozone concentration measured in air during noon to 4pm is -60 ppb. According the laboratory study, at this level of ozone, we should see 18%-27% of formaldehyde and 19%-23% of acetaldehyde losses. Figure 4.3-5 are three chromatograms of three samples, chromatogram #B4L177 with a ozone denuder scrubber, #B1L177 with a KI cartridge scrubber and #B3Ll77 without ozone removal device. These DNPH cartridges were sampled at same time period, from noon to 4pm, that ozone reaches a maximum. As 65

PAGE 78

Lab aaae AD.&1:yd date 1Cet:llo4 Dac:riptiOD C:o1uaa Data file Saaple -6.400mV I t I it-1 OIU.Y C:oloracSo D&AV&r I 07/24/1996 16102119 Autoaap1er 10 a1cro1itar ur.c: ov Aborptioa 150 X 4.6 .. C:18 R&Y&re QracU&At: CIDCII" 1120 1 L22Ril'l07 .ASC: (cll\peu.hdat:a) 1 Auraria Saap1e 1J95 64. OOOIIIV ""'-'1 ""'-'1 1\..n:r-,1 '"-I '"'-I '"-I 11\aer Rt.,..tlon Ana 0 0.116 9.26 0 1.033 7.92 0 1.450 10.67 0 z.1oo tZ.ZP 0 3.150 4337.13 0 4.Z8l tS.SS 0 5.616 Z47.94 0 7.aa3 437.71 5 9.516 11.98 0 10.666 476.24 7 11.600 53.99 0 13.700 1.91 0 14.700 76.73 Figure 4.3 Chromatogram ofthe filed sample #B4L177, measured by sampler #4, with denuder scrubber. 66

PAGE 79

Lab aa.e 1 U'lli"r of De11ve: Aa&lya:l.a date 1 07/22/1996 23118112 X.thocS 1 Autoaaaplu- 10 llic:olite: Deac::l.ptioa 1 IUr.<: UV Coluan 1 150 X .. C18 ca:::tu-1 o:acS:I.e11t aoar 1120 Data file 1 L22aU35.ASC (cSI\peak2\cSata) Saaple 1 Au:aria Saaplea 1!95 .400mV 64.000mV ec.parw,t &rinawt) &rinawt) &rinawt) lriniM!) &rinawt) &rinawt) &rinawt) &rinawt) -lcMII,..S. &rinawt) ceca I eMil \'de lllaler letlntton ..,... 0 0.350 45.41 0 1.050 6.62 0 t.433 10.39 0 2.tt6 1l.5t 0 .s.zoo 4287.04 a 4.350 14.1111 0 5.183 s.a1 0 5.766 146.34 4 7.oil 27.01 a a.1JJ 384.74 5 9.816 za.a1 6 ta.900 zoo.oa 1 tt.866 0 t.S.Sa3 a 14.616 0 t4.9ll 52.41 5.38 Z6.l9 t6.12 Figure 4.4 Chromatogram of the filed sample #BlL177, measured by sampler #1, with KI cartridge scrubber. 67

PAGE 80

Lab 1 'C'niY of Colorado Analyaia d&te 1 07/23/1996 1JI04t2S X.tbod 1 Autoaa.ple:r: 10 Deac:r:iptiOA 1 ULC UV .Abao:r:ptiOD Colu.D 1 150 X 4.6 .. Cll ca:r::r:ie:r: 1 ooa H2o Data file 1 L22R0"71 .ASC (4t\p .. k2\d&tal la.ple 1 la.plea lJJS -6.400mV r.,_ 64.000111V ....,_,, ....,_,, '"'-"> '"'-"> lrincM!l lrincM!l lrincM!l ....,_,, ....,_,, t-lrincMI) lrincMI) letentfan Ar 0 0.200 79.47 0 1.016 1l.5Z 0 1.466 14.11 0 z.too 93.75 0 ].166 3090.7'0 0 4.066 lO.ZS 0 5.100 14.09 0 5.650 113.1Z 0 7.916 liZ.lS a 9.500 I.Z.II 0 10. 7'00 157.30 7 11.633 47.56 a IZ.l16 7.04 0 13.7'00 IO.SZ 0 14.643 73.89 Figure 4.5 Chromatogram of the filed sample #B3Ll77, measured by sampler #3, without ozone scrubber. 68

PAGE 81

can be seen from chromatogram #B3L117, the size of the peak in front of formaldehyde is very small. This peak is produced by ozonation of2,4-DNPH. It means only small amount of ozone reached the DNPH cartridge, it is not big enough to cause interference. Therefore, ozone interference was not observed. Using long polyethylene tubing and connector might remove some of ozone, since ozone can add to the double bond in polyethylene. In laboratory study only Teflon tube and connector were used to limit ozone loss in the system. The results from field study shows that using Kl cartridge scrubber caused interference on measurement of formaldehyde concentration, but the Kl cartridge worked well in laboratory study. This disagreement could be due to differing KI cartridges used. Figure 4.6 shows a chromatogram ofKI cartridge scrubbed sample from field study. As can be seen clearly, there is an unidentified peak between formaldehyde and acetaldehyde after adding Kl cartridge. It gets smaller with continuous use. This unidentified peak did not show up in the chromatograms of laboratory study. In order to ascertain this change is not due to differing experimental conditions, an experiment was conducted to sample formaldehyde using KI cartridge scrubber made for field study, ozone concentration added in the gas stream was 50ppb, which is similar to atmospheric condition. As can be seen from figure 4.7 (chromatogram #30), the peak between formaldehyde and acetaldehyde appeared on the chromatogram again. The way we prepare KI cartridge for laboratory use and field 69

PAGE 82

use is exactly the same. The cause of the difference could be due to an impurity or contamination ofKI reagent and silica gel used for field study. Since acetone source for laboratory study is unavailable, the measured acetone results from field study can not be compared. 70

PAGE 83

Lab ca.. date X.tbocl IleacriptiCD COliDID Carrier Date fila S.-p1a -6-400111V UlliY of Co1orac!o Dezlver 07/22/1JJ6 21a1la27 a Autoa.-plar 10 .tcro11tar 1 DLC O'V .Abaozpt1CD 150 x 4,5 .. C18 CkacUallt CBJor-1120 L22RU2J .ASC (cia \peak2\data) Aur&ria Sa.plea lJJS 64 .000mV :". f \ Cooiporwlt llulbw .. tencfon Aree I ... i i 0 j I I n,_,, n,_,, ...tncMI n,_,J n,_,l n,_,, .....,_,, cetal doll )'de reAcetcne cetcne ....,_.., a a.S43 a 1.066 a 1.466 a 2.116 a 3.200 a 4.343 a 4.683 a 5.200 a $,766 7.066 a 1.116 5 9.766 6 10.a5a T 11.850 a 13.766 a 14.866 a 15.466 0 15.916 Figure 4.6 Chromatogram ofthe field sample #B1L161, measured by sampler#!, with a new KI cartridge scrubber. 71 67.19 1.52 1.71 13.44 2m.69 13.48 6.5a 13.16 127.75 9S.a7 216.59 50.81 16-24 31.23 12.79 50.47 5.61 7.10

PAGE 84

L&b a&ae 1 OQ1v of An&ly date 1 OC/OS/1337 081ll10C Method. 1 Auto ... 10 aiCJ:"Olitar Dec:dptioa 1 IDILC f1V Coluan 1 150 x 4.6 .. Cll C&rrier I CJ:r&dient CHJCR' R20 Deta file a UMJS.ASC (d1\peak2\datal Sample 1 Aura:ia Sa.ple 1995 -6.40011\V i I ; f ! I u.ooomv Caoponent ......,_,, l.ri:novl) lri:nown) lri:nown) lri:nown) lri:nown) Ul'llr:nclw1 ) l.ri:novl) ......,_,, ...... I lri:nown) lfiRstllj l.ri:novl) cc-octAct-oecAcet-lri:nown) '"'-'> lri:nown) l.ri:novl) 1\aber htenefon 0 0.200 Area 39.16 0 1.033 20.19 0 1.443 140.54 0 2.216 7.$9 0 2.7$0 5.64 0 3.300 11l61.1Z II 0 s 133 43.64" 0 5.666 484.08 0 6.783 107.64" 7.716 1945.29" 0 9.216 12.16" 5 9.5a3 19.17" 8 18:9ti U:!l 0 10.5$0 n.n" 7 11.500 1Z.Sl a 1Z. 116 9.98 a 12.433 6.96 0 13.150 6.10 0 13. 7$0 24.19 0 14.366 159.16 0 15.3$0 12.93 Figure 4 7 Chromatogram of the laboratory sample, with a KI cartridge scrubber made a year ago. 72

PAGE 85

5. CONCLUSIONS In the laboratory study, we found a severe negative bias on formaldehyde and acetaldehyde concentrations when ozone is present, even as low as -50 ppb. The ozone denuder scrubber and KI cartridge scrubber can be used to remove ozone without interfering with measurement of carbonyl compounds by DNPH silica gel method. As we expected the ozone interfering peaks increased as a function of ozone concentration. The magnitude of the ozone interference depended on the amount ozone sampled. In field study, comparison ofthree sets of field data showed that when ozone was removed by ozone denuder scrubber or KI cartridge scrubber, the carbonyl compound concentrations obtained by DNPH silica gel technique were not significantly different. No correlation was found between ozone concentration and carbonyl losses in unscrubbed samples. KI cartridge scrubber used for field study interferes with the measurement of formaldehyde concentration. Likely some other mechanisms exists, that can remove low level of ozone in this sampling system. Consequently the final result shows that when ozone concentration is low, the ozone interference is small. 73

PAGE 86

REFERENCES 1. Finlayson-Pitts, B. J.; Pitts, J. N., Jr. Atmospheric Chemistry: Fundamentals and Experimental Techniques; Wiley and Sons: New York, 1986. 2. Pitts, J. N., Jr,; Wan, J. K. S. In The Chemistry of the Carbonyl Group; Patai, S., Ed.; Interscience Publishers: New York, 1966. 3. Grant L. D. (1989) and Tikton B. E. Health effects of major combustion ad transformation products of alcohol motor vehicle fuels. Presentation at Workshop on the Effects of Atmospheric Aldehydes on Health and Environment, Sau Paulo, Brazil, 12-13 April. 4. EPA Health assessment document for actaldehyde. EPA 6008-86-0 15A, external review draft, Environmental Criteria and Assessment Office. Reserch Triangle Park, NC. 5. Tanner R. L., Miguel A. H. Atmospheric chemistry of aldehydes: enhanced peroxyacetyl nitrate formation from ethanol-fueled vehicular emissions. Envir. Sci. Techno/. 22, 1026-1034. 6. Williams R. L., Lipari F. and Potter R. A. Formaldehyde, methanol and hydrocarbon emissions from methanol-fueled cars. J Air Waste Man. Ass. 40, 747-756. 7 Arnts R. R. and Tejada S. B. 2,4-Dinitrophenylhydrazine-coated silica gel cartridge method for determination of formaldehyde in air: identification of an ozone interference. Envir. Sci. Techno/. 23,1428-1430 8. Dasgupta P. K., Dong S. Continuous liquid-phase fluorometry coupled to a diffusion scrubber for the real-time determination of atmospheric fonnaldehyde, hydrogen peroxide and sulfur dioxide. Atmospheric Environment 22, 949-963. 9. Tuazon, E.; Carter, W, P. L. Reactions ofhydrazines with ozone under simulated conditions. Environ. Sci. Techno/. 1981, 15, 823-828. 10. Smith D. F., Kleindienst T. E. and Hudgens E. E. Improved high-performance liquid chromatographic method for artifct-free measurements of aldehydes in 74

PAGE 87

the presence of ozone using 2,4-dinitrophenylhydrazine. J. Chromat. 6, 425444. 11. Atkinson R. and Carter W. P. L. Kinetics and mechanisms ofthe gas-phase reactions of ozone with organic compounds under atmospheric conditions. Chem. Rev. 84, 437-470. 12. Vairavamurthy A. (1993) Sampling of atmospheric carbonyl compounds for determination by liquid chromatography after 2,4-dinitrophenylhydrazine labelling. In Sampling and Analysis of Airborne Pollutants, (edited by Winegar K and Keith L. H.), Lewis, Chelsea, MI. 13. Hoigne J., Bader H., Haag W. R. and Staehelin J. Rate constants of reaction of ozone with organic compounds in water-III. Wat. Res. 19, 993-1004. 14. Grosjean D. and Parmar S. S. Laboratory tests ofKI and alkaline annular denuders. Atmospheric environment 24A,2695-2698 15. Williams E. L. II and Grosjean D. Removal of atmospheric oxidants with annular deunders. Envir. Sci, Techno/. 24, 811-814. 16. Slemr J. Determination ofvolatile carbonyl compounds in clean air. Fres. J. Anal. Chem. 340,672-677. 17. Tanner R. L., Markovits G. Y., Ferreri E. M. and Kelly T. J. Sampling and analysis of gas-phase hydrogen peroxide following removal of ozone by gas phase reaction with nitric oxide. Analyt. Chem. 58, 1857-1865. 18 Purdue, L. J., Dayton, D.P., Rice, J., and Bursey, J. (1991). Assistance Document For Sampling And Analysis Of Ozone Precursors. US environmental Protection agency, Research triangle park, NC, EPA/ 600/ 8-91/215 19 Tejada S. B. (1986) Evaluation of silica gel cartridges coated in situ with acidified 2,4-dinitrophenylhydrazine for sampling aldehydes and ketones in air. Int J. envir. Analyt. Chem. 26, 167-185. 75