Citation
Chemiluminescent detection of nitrogen dioxide in the atmosphere

Material Information

Title:
Chemiluminescent detection of nitrogen dioxide in the atmosphere
Creator:
Cofrin, Katie A
Publication Date:
Language:
English
Physical Description:
xi, 47 leaves : ; 28 cm

Thesis/Dissertation Information

Degree:
Master's ( Master of science)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Chemistry, CU Denver
Degree Disciplines:
Chemistry

Subjects

Subjects / Keywords:
Atmospheric nitrogen dioxide -- Measurement ( lcsh )
Chemiluminescence ( lcsh )
Atmospheric nitrogen dioxide -- Measurement ( fast )
Chemiluminescence ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaf 47).
Thesis:
Department of Chemistry
Statement of Responsibility:
by Katie A. Cofrin.

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:
757517449 ( OCLC )
ocn757517449
Classification:
LD1193.L46 2011m C63 ( lcc )

Full Text
CHEMILUMINESCENT DETECTION OF NITROGEN DIOXIDE IN THE ATMOSPHERE
by
Katie A Cofrin
B.S., University of Wisconsin-Stevens Point, 2008
A thesis submitted to the
University of Colorado Denver
In partial fulfillment
Of the requirements for the degree of
Master of Science
Chemistry
2011


The thesis for the Master of Science
degree by
Katie A Cofrin
has been approved
by
Hai Lin
Date


Cofrin, Katie A (M.S., Chemistry)
Chemiluminescent Detection of Nitrogen Dioxide in the Atmosphere
Thesis directed by Professor Larry Anderson
ABSTRACT
A measurement technique for the detection of nitrogen dioxide in air was successfully modified.
Ambient air is sampled by an LMA-3 Scintrex Luminox Analyzer. By taking advantage of the
chemiluminescent reaction between nitrogen dioxide and luminol, one can monitor the
concentration of nitrogen dioxide in ambient air. When nitrogen dioxide comes into contact with
an alkaline solution containing luminol, the oxidized form of luminol (amino phthalate derivative)
is produced. This excited-state molecule emits energy in the form of light when relaxing back
down to a lower-state energy level. A photomultiplier tube within a Scintrex LMA-3 measures the
emission, which occurs at approximately 425 nanometers. A signal is produced that is proportional
to the nitrogen dioxide mixing ratio in the ambient air sample. Working in tandem with a Lab View
Data Acquisition System, the LMA-3 can measure nitrogen dioxide in ambient air samples in parts
per billion. An added scrubbing apparatus for zeroing allows for continuous sampling for up to
several days without maintenance.


This abstract accurately represents the content of the candidates thesis. I recommend its
publication.
Signed
Larry Anderson


DEDICATION
I dedicate this thesis to my parents.


ACKNOWLEDGEMENT
I would like to thank Dr. Larry G. Anderson for his guidance and unwavering support throughout
my research and thesis writing. I also want to express my gratitude for his continued patience
while working with me. Additionally, I would like to thank my colleague, Marc Mayzes, for all of
his help with setting up the Data Acquisition System and programming it to my research needs.


TABLE OF CONTENTS
Figures................................................................................ix
Tables.................................................................................xi
Chapter
1. Introduction......................................................................1
1.1 Background..............................................................1
1.2 Standards and Health Effects............................................2
1.3 Ambient Concentrations..................................................2
1.4 Ambient data from CAMP and Welby........................................3
1.5 Measurement Techniques..................................................8
1.6 Luminol Technique......................................................10
2. Experimental.....................................................................12
2.1 Instrument.............................................................12
2.2 Instrument modifications...............................................13
2.3 Dilution system........................................................15
2.4 Chemicals..............................................................17
2.5 Data Acquisition.......................................................17
3. Optimization.....................................................................19
3.1 Luminol Solution.......................................................19
3.1.1 Additionof sodium hydroxide....................................19
3.1.2 Addition of methanol...........................................20
3.1.3 Addition of sodium sulfite.....................................22
vii


3.2 Luminol flow rate..........................................................25
3.3 Sample volume flow rate....................................................26
3.4 Photomultiplier tube voltage...............................................27
3.5 Optimal parameters.........................................................29
4. Calibration.........................................................................30
4.1 Gas calibration with nitrogen dioxide......................................30
4.2 Gas calibration with nitric oxide and ozone................................31
4.3 Determination of zero......................................................32
5. Data and Results....................................................................33
5.1 Setting collection cycles..................................................33
5.2 Indoor laboratory data.....................................................33
5.3 Outdoor sampling...........................................................36
5.4 Data collection............................................................37
6. Conclusions and Future Work.......................................................44
References.................................................................................47
viii


FIGURES
Figure
1.1 ANNUAL AVERAGE N02 CONCENTRATIONS IN AMBIENT AIR AT TWO
SAMPLING SITES IN COLORADO..............................3
1.2 MONTHLY AVERAGE CONCENTRATIONS OF
HOURLY DATA FOR 1 YEAR..................................4
1.3 DIURNAL TRENDS ILLUSTRATED BY 2 DAILY PEAKS.............6
2.1 TOP VIEW OF SCINTREX LMA-3.............................12
2.2 SCHEMATIC OF VALVE AND SCRUBBER SYSTEM.................14
2.3 SAMPLE AIR FLOW THROUGH INSTRUMENT.....................15
2.4 CALIBRATION CURVE OF GAS FLOW CONTROLLER...............16
2.5 CALIBRATION CURVE OF AIR FLOW CONTROLLER...............17
3.1 NOz SIGNAL ASA FUNCTION OF METHANOL
CONCENTRATION..........................................20
3.2 DEGREDATION OF 0.05% METHANOL IN LUMINOL SOLUTION......21
3.3 DEGREDATION OF 0.10% METHANOL IN LUMINOL SOLUTION......21
3.4 OZONE SIGNAL WITH LUMINOL CONTAINING
SODIUM SULFITE.........................................23
3.5 EFFECT OF CONCENTRATION OF Na2S03 ON SIGNAL............24
3.6 AVERAGE SIGNALS OBTAINED FOR EACH CONCENTRATION
OF Na2S03..............................................24
3.7 FLOW RATE EFFECTS ON SIGNAL............................25
3.8 SLOPE AS A FUNCTION OF LUMINOL FLOW....................26
3.9 OPTIMAL VACUUM FLOW RATE AS DETERMINED BY GREATEST
ATTAINABLE SIGNAL......................................27
3.10 SIGNAL INTENSITY AS A FUNCTION OF PMT VOLTAGE..........28
IX


4.1 CALIBRATION CURVE PRODUCED WITH LUMINOL SOLUTION
MODIFIED WITH METHANOL......................................30
4.2 NO + 03 CALIBRATION CURVE....................................31
4.3 ASCARITE SCRUBBER EFFICIENCY.................................32
5.1 INDOOR LABORATORY DATA AT THE UNIVERSITY OF
COLORADO DENVER OVER A 24-HOUR PERIOD.......................34
5.2 INDOOR LABORATORY DATA AT THE UNIVERSITY OF COLORADO DENVER
OVER A 24-HOUR PERIOD: DAY 2................................35
5.3 MAP OF AURARIA CAMPUS........................................36
5.4 RAW DATA FROM AURARIA CAMPUS: 0800 5/4/2011 0800 5/5/2011..37
5.5 HOURLY AVERAGE DATA FROM AURARIA CAMPUS:
0800 5/4/2011 0800 5/5/2011...............................38
5.6 EXCEL TEMPLATE STEPS FOR RAW DATA ANALYSIS...................39
5.7 RAW OUTDOOR DATA COLLECTED AT AURARIA 5/11/11................40
5.8 HOURLY AVERAGE COMPARISON OF AURARIA, CAMP AND
WELBY DATA..................................................40
5.9 DATA COMPARISON FOR LINEAR RELATIONSHIP BETWEEN
CAMP AND AURARIA............................................41
5.10 RAW DATA FROM AURARIA 5/4/11-5/6/11..........................42
5.11 RAW DATA FROM AURARIA 5/29/11-5/31/11........................42
5.12 DATA FROM AURARIA, CAMP, AND WELBY 5/4/11-5/6/11.............43
X


TABLES
3.1 Optimal parameters for ambient sampling
.29
xi


1. Introduction
1.1 Background
Nitrogen dioxide is of interest to atmospheric scientists because it is found in ambient air as a
result of industrial and vehicular combustion processes. In its pure state, nitrogen dioxide is an
orange-brown gas that has a pungent, sweet odor. Nationally, about 58% of nitrogen dioxide
emissions come from on- and off-road vehicles [1]. The second major contributor is industrial
sources, which make up 36% of nitrogen dioxide emissions. In Denver, Colorado, these numbers
vary slightly. About 44% come from large industrial sources such as power plants. Thirty-three
percent of the emissions are from vehicles. The third and last significant contributor is space
heaters, making up 15% of nitrogen dioxide emissions.
Nitrogen dioxide (N02) also exists as a secondary pollutant, resulting from a photochemical
oxidation of nitric oxide (NO) in the atmosphere. The photolysis of N02 leads to the formation of
ozone [2], another pollutant of interest. The atmospheric chemistry resulting from the dissociation
of N02 leads to the oxidation of the reduced gases in the atmosphere including carbon monoxide
(CO), methane (CH4), and non-methane hydrocarbons (NMHCs). These gases are also
atmospheric pollutants. N02 concentrations have the most pronounced effect on the ozone budget
and its height profile in the atmosphere. [3]
The production of nitrate aerosols is also of concern. The formation of nitric acid in the air occurs
when NO and N02 are oxidized by hydroxyl radicals. Once formed, the nitric acid can make its
way to the earths surface by way of precipitation (wet deposition). This is a known contributor to
acid rain. N02 is also removed from air by dry deposition, which causes damage to vegetation.
1


1.2 Standards and health effects
The U.S. Clean Air Act of 1970 called for the establishment of outdoor air standards for air
pollutants that were shown to produce adverse health effects. The EPA established the first NO2
standard in 1971. This was an annual average standard of 0.053 parts per million (ppm), or 53 ppb.
This value set the maximum allowable annual average concentration. Recently, an additional
standard was included; an hourly average standard. On January 22, 2010, the new standard100
ppbwas established. This did not affect the annual average standard. Compliance with the new
standard is evaluated by taking the 3-year average of the 98th percentile of the annual distribution
of daily 1-hour maximum concentrations. The aim of the EPAs transition from annual to 1-hour
standard was to help protect public heath, particularly those affected by asthma, and age groups
with a greater sensitivity, such as the elderly and children.
The health effects associated with elevated exposure to nitrogen dioxide include minor respiratory
distress. Other negative effects reported are degradation of vegetation and clothing. Some of
Denvers visibility problems stem from nitrate aerosols, which also contribute to acid rain and
successively, damage to plant life as the nitric acid falls to the ground via wet and dry deposition.
1.3 Ambient concentrations
In the state of Colorado, there are two sites that monitor N02, CAMP and Welby. CAMP is
located in downtown Denver at Broadway and 21st street. This site reports hourly data for N02 in
addition to several other pollutants. Welby also reports hourly data and is located at 3174 E. 78th
Avenue, a site along the Platte River. Neither site has reported N02 levels that exceed the annual
standard of 0.053 ppm in over thirty years. The Colorado 2009 Air Quality Data Report states that
the N02 levels at the CAMP sampling site exceeded the standard in 1977 and have since been on a
gradual decline.
2


Ambient levels will vary based on where the site is located. For example, since CAMP is located
in downtown Denver, it is expected that measured levels would be higher here due its close
proximity to downtown traffic. Increased vehicular exhaust associated with this location will
contribute to higher concentrations in ambient air. In contrast, Welby is located along a riverbank
in a more rural area, and therefore is expected to contain lower concentrations of NO2. The data
do, in fact, show that this is the case. For annual averages from both sites, refer to Figure 1.1.
0.060
0.050
?
a 0.040
iA
c
S 0.030
2

I 0.020
o
u
0.010
0.000

^

*v "v

dT ^ ^ ^ ^ ^ ^
Figure 1.1 ANNUAL AVERAGE N02 CONCENTRATIONS IN AMBIENT AIR AT TWO
SAMPLING SITES IN COLORADO
1.4 Ambient data from CAMP and Welby
Data collected by the state from CAMP and Welby over a period of one year was analyzed to
determine whether seasonal, diurnal, or other trends exist. Concentrations were reported as an
hourly average. This hourly average was averaged for each month. Figure 1.2 shows the monthly
average concentration collected at CAMP and Welby from June 2010-May 2011.
3


Average monthly concentration (ppb) Average monthly concentration (ppb)
CAMP
a Welby
35
30
25
20
15
10
5
0
May-10 Jul-10 Aug-10 Oct-10 Nov-10 Jan-11 Mar-11 Apr-11 Jun-11
Date (Mon-Yr)


* * 1 1 A
* A
A A A A A
A


CAMP
ss Welby
Jun-10 Aug-10 Oct-10 Dec-10 Feb-11 Apr-11
Date (Mon-Yr)
Figure 1.2 MONTHLY AVERAGE CONCENTRATIONS OF HOURLY DATA FOR 1 YEAR
4


One noticeable trend is that the average monthly concentrations from CAMP were consistently
higher than those measured at Welby. Also, a seasonal trend was present. An increase in average
occurs during the winter months, December 2010-February 2011. As mentioned in the
Background section (1.1), space heaters are a significant contributor to NO2 emissions. A possible
explanation for this increase in measured N02 is an increase in space heater usage due to colder
temperatures during the winter months. In addition, it may suggest that, also due to colder
weather, there is a decrease in foot traffic and vehicular traffic is increased. A variety of
meteorological factors also cause this. None of these concentrations exceed the annual standard;
in fact, the highest average obtained was 32 ppb, which is 21 ppb less than the standard. This
average was measured at CAMP, during February 2011. For Welby, the highest average
concentration was the same during both December 2010 and January 2011, at 25 ppb.
Hourly averages were analyzed for each month separately. In nearly all cases, a diurnal trend did
exist. Two distinct peaks were observed during the sampling day, at around sample hours 8 and
17, corresponding to 8AM and 5PM, respectively. These hours are typically known for high traffic
due to commuting to work, often being referred to as rush hour. This relationship is shown in
Figure 1.3. This figure includes data for one month from each season. When analyzing hourly
data, the highest concentration and corresponding sampling hour was noted. Unexpectedly, the
two highest concentrations occurred during the warmer months, which tended to have an overall
lower average monthly concentration. In April, CAMP measured a nitrogen dioxide level of 88.6
ppb, which exceeds the annual standard (53 ppb), but still remains below the hourly standard of
100 ppb. This measurement was taken during sample hour 21, or between the hours of 8-9PM.
Welby also had a high hourly concentration detected in a warmer month, June, at 79 ppb,
measured at hour 19, or between the hours of 6-7PM. While these maximum hour concentrations
5


were detected during the later hours of the evening, the highest level from each month tended to be
detected around sample hour 9 or 10, which correspond to 8-9AM and 9-10AM, respectively.
This supports the finding that the highest levels of nitrogen dioxide are detected during rush hour,
or between the morning commute hours of 8-10AM.
June 2010
X
a.
n.
s
o
a
a
s
o
u
jj,
3
O
X
V
M

u
V
>
<
35
30
25
20
15
10
5
0

t A
# f t f " A
A A A A 4 A 4 A*.. A
A 4 A A A 4 A 4 A
4 A * A
i ? - i s ?
CAMP
AWelby
10
15
20
25
Hour
Figure 1.3 DIURNAL TRENDS ILLUSTRATED BY 2 DAILY PEAKS
6


September 2010
a
a
&
c
o
c
o
u
>>
o
X
4)
>
<
40 -.A __ .
3d A * w
3(J " i § A... A . #
Zb "" w A 9 A *
ZU * * * f #r f
lb " ^ 4 A *
1U " A A A A *
b * o - r --r~~~~ 7
CAMP
AWelby
0
5 10
15 20 25
Hour
December 2010
CAMP
a
a
c
o
o
(J
0)
6D

u
4)
>
<
u
3
o
X
35
30
25
20
15
10

A
A

^7
-4-A-
ti
p
10
J
15
~1--
20
AWelby
25
Hour
Figure 1.3 DIURNAL TRENDS ILLUSTRATED BY 2 DAILY PEAKS (CONT)
7


March 2011
.0 a a 40
e 0 35
b 30
a 0 u 25
S O 20
>> "u 15
3 O X 10
0 0x1 03 5
l* CJ > < 0

i * f A
----$-*
4^ .A~.
A I
..A..
-A.A.A1 4.
T*
5
j
10
15
20
Hour
a Welby
*~TT
25
Figure 1.3 DIURNAL TRENDS ILLUSTRATED BY 2 DAILY PEAKS (CONT)
1.5 Measurement Techniques
Various methods have been developed for measuring atmospheric samples for the purpose of
detecting trace amounts of pollutants. Since pollutants in ambient air can be found at a level of sub
parts per billion, sensitivity and selectivity are vital to a successful measuring technique. There
are several published methods by which N02 can be measured. The most commonly used
technique is N0/03 chemiluminescence. This technique relies on the gas-phase chemical reaction
of NO with ozone [4], The instrument actually detects NO. An air sample is reacted with a high
concentration of ozone (03) so that any NO in the sample is able to react and produce light. The
product molecule is an excited-state nitrogen dioxide molecule. This high-energy molecule will
release energy in the form of light to relax to a lower, more stable energy level. See equations (1),
(2) below. An additional sample inlet, or in some cases, an additional instrument, allows for a
8


second ambient sample to be drawn simultaneously. This second sample is introduced to a
reducing agent so that any N02 in the air will be converted to NO, prior to reacting with ozone.
The reducing agent can either be a heated catalyst, usually either molybdenum or gold in the
presence of CO, or a form of light (photolytic reduction). Since there is presumably NO in the
sample being measured along with the reduced N02, the signal is a sum of both NO and N02,
called NOx. The concentration of N02 is calculated by subtracting the NO signal from the NOx
signal. [4]
NO + O3 ^no2* + o2 (1)
NOz* ^N02 + hv (2)
A major disadvantage to this method is that other N02-containing species (i.e. HNO3, PAN) can
also be reduced, thus causing interference with the signal that is identified as the signal from N02.
[5] An alternative method involves the photodissociation of N02 to NO, again followed by the
chemiluminescent reaction with O3. [2] One N02-direct measurement technique uses the
absorption of infrared radiation as the base for detection. [4] The light source employed is a
tunable diode laser. [6]. In this method, an ambient sample flows through a multi-pass optical
cell. The infrared resonant absorption lines are then analyzed. Using second harmonic
techniques, low-pressure absorption lines can be detected with high sensitivity. This approach is
also used at atmospheric pressure, but has decreased sensitivity due to line broadening. [7] This
technique is usually employed in research applications, as it is a fairly expensive method of
analysis.
9


1.6 Luminol technique
In 1980, Maeda, et al., described a novel measuring approach, specific to N02. [8] This method
involved reacting N02 with a solution containing luminol (5-amino-2,3-dihydro-1,4-phthalazine
dione) which produces a chemiluminescence approximately 1000 times stronger than that of the
N0/03 chemiluminescence. See equation (3).
The technique of reacting N02 with luminol has been used for over thirty years. Nitric oxide does
not react with luminol, so it will not interfere with the N02 signal. Since N02 reacts directly with
the luminol in solution, there is no need for reduction of N02 to NO. When luminol is in the
presence of an oxidizer, it is able to react and produce an excited product molecule, an amino
phthalate derivative. An excited molecule is very unstable. Akin to the NO/O3 chemiluminescent
reaction described above, the excited molecule will release energy to relax back down to a lower
energy level. Like N02*, the amino phthalate derivative releases energy in the form of visible
light.
Within a decade of the discovery of luminol, the chemiluminescent reaction with hydrogen
peroxide was widely known. This chemiluminescence is not specific to hydrogen peroxide; rather,
the amino phthalate derivative will produce this emission when luminol is exposed to a number of
oxidizing agents, including N02.
In 1985, an instrument created for the purpose of this measurement was marketed as the Scintrex
LMA-3. It is a highly sensitive, portable, and lightweight instrument capable of continuous
10


measurement of NO2 in air. [3] This instrument surpassed all preceding measurement systems for
NO2. In addition to the 1000-fold increase in signal sensitivity, the response time was greatly
reduced to near real-time.
11


2. Experimental
2.1 Instrument
The instrument used to measure NO2 was a Scintrex LMA-3. This instrument is capable of
continuous ambient measurements with rapid response time. The following are key parts to the
LMA-3 that make it a successful instrument for measuring N02. For an inside view of the
instrument, refer to Figure 2.1.
FIGURE 2.1 TOP VIEW OF SCINTREX LMA-3
The first part is the sample inlet, located on the back of the instrument. Ambient samples are
introduced to the instrument at this location before passing into the luminol chamber. The luminol
chamber is where the chemiluminescent reaction occurs. A peristaltic pump controls the flow of
luminol solution in and out of the reservoir. The flow rate of the luminol solution allows for a
12


constant supply of fresh solution into the chamber. Located inside this chamber is a wick that
becomes saturated with the luminol solution. A photomultiplier tube contained in the chamber
inside the LMA-3 measures the light emitted near the central part of the wick and generates a
signal that is proportional to the mixing ratio of N02 in the sample air. [9] The signal can be
observed on a digital display on the front face of the instrument.
2.2 Instrument modifications
In initial tests, the LMA-3 appeared to be in working order with few minor concerns. First, the
peristaltic pump contained in the LMA-3 ran on a non-programmable six-minute cycle. For three
minutes, the pump would rotate clockwise and then the pump would rotate counter-clockwise for
three minutes. The purpose of this cycle appears to be a zeroing or flushing mechanism.
However, this posed a problem for continuous sampling. In order for the LMA-3 to constantly
flow luminol solution, the pump is required to be able to revolve in one direction for extended
periods of time. A constant flow of luminol solution into the chamber allows continuous sampling
to occur. To accomplish this, the internal pump was replaced with a Masterflex C/L peristaltic
pump, which was placed externally on the top of the LMA-3. Another advantage of the
Masterflex pump is its capability to change revolution speed, which controls the luminol solution
flow rate. Secondly, luminol reservoir and waste containers must meet the volume needs of a
continuous sampling instrument. The containers included in the LMA-3 were modest in size,
having a 500mL holding capacity. These were replaced with 1 -liter Nalgene bottles. The larger
Nalgene bottles allow for longer (2x previous) sampling time without interrupting measurements
to change out containers. Although the reverse cycle on the internal peristaltic pump was present
for blanking, this cycle would last as long as the sampling cycle. This is not ideal for data
collection, since only half of the data collected would be actual sampling. The LMA-3 did not
13


come equipped with any additional programmable device that would blank or zero the instrument
during sampling. A separate, detachable apparatus was added for this purpose. A schematic of
the apparatus (Figure 2.2) illustrates the simple setup. A sample inlet allows air to be drawn
through the tubing at a pump-controlled volume. The air proceeds to a 3-way T, which leads to
the sample inlet on the back of the LMA-3 during regular sampling. When a voltage of 5.0 V is
applied, a valve closes this pathway and air is forced to go through the Ascarite scrubber. The
scrubber is able to remove N02 in the sample prior to passing into the LMA-3 for measurement.
Because this air is free of N02, the signal produced is considered a zero (baseline) for ambient
measurements.
Sample inlet
Figure 2.2 SCHEMATIC OF VALVE AND SCRUBBER SYSTEM
14


Air in
Air out
l
Figure 2.3 SAMPLE AIR FLOW THROUGH INSTRUMENT
Figure 2.3 shows the flow of the air sample through the instrument. When the air leaves the
chamber after the chemiluminescent reaction takes place, it passes through a trap that will catch
any moisture that is exiting the chamber with the air sample. If moisture is able to go through the
air line and exit the LMA-3, the vacuum pump and all tubing will become wet, which can skew
data and incapacitate the pump.
2.3 Dilution System
A cylinder of NO2 (-200 ppm) was initially used for gas calibration. This was a cylinder that had
not been used for several years, so the exact concentration is not known. The listed concentration
was used as a reference for the purpose of generating a reasonable calibration curve. To facilitate
dilutions, a gas flow controller and air flow controller were calibrated prior to use. Flow
15


controllers are devices that can be set to allow a certain volume of air or gas to flow into a
container. To calibrate the flow controllers, first a gas source was connected to the flow controller.
A Bios International DryCal DC-1 Flow Calibrator was connected to tubing at the outlet end of
the flow controller. The DC-1 Flow Calibrator measures gas flow rate in volume per unit of time.
The DC-1 is able to use a near-frictionless piston and photo optic sensing technology to obtain
volumetric flow readings from 1 to 50,000 cm3 per minute. For both the gas and airflow
controllers, a manual dial is used to control the amount of gas (or air) that is allowed through the
controller and into the calibrator. Calibration curves generated to illustrate the relationship
between dial setting and corresponding flow rate for each of the two flow controllers may be seen
in Figures 2.4, 2.5. House air was used as the diluent. Purified air, scrubbed of NO2 and other
contaminants, could not be used in the laboratory due to flow limitations of the purification
apparatus. Largely, the apparatus was unable to provide a sufficient flow rate for calibration.
Figure 2.4 CALIBRATION CURVE OF GAS FLOW CONTROLLER
16


Box setting
FIGURE 2.5 CALIBRATION CURVE OF AIR FLOW CONTROLLER
2.4 Chemicals
Luminol (Reagent grade) was purchased from Fisher Scientific and used as received. Methanol,
also purchased from Fisher, was HPLC grade. Sodium hydroxide and sodium sulfite were
obtained from J. T. Baker and all solutions were prepared in the laboratory. Deionized water was
obtained from passing water through a Millipore Milli-Q system.
2.5 Data acquisition
Data were acquired using a LabView system. The data acquisition system (DAQ) 12-bit hardware
was fixed between the LMA-3 and a laptop computer. When the LMA-3 generated a signal, this
signal was transferred through the LabView and into a file on the computer in the form of a
voltage reading. A voltage range can be set to a small range, IV, up to a very large range, 20V.
When the range increases, the bins (or possible levels) that can be occupied become more spread
17


out. When the bins are smaller (smaller range), the resolution is better than when the bins are
larger. After the voltage readings were taken, the file was then readily available for uploading into
a graphing program, such as Excel. The data collected included Sample Number, Date, Time, and
Signal. Various parameters for the sampling and zeroing cycles were adjustable, including
frequency of measurement and length of time sampling (or zeroing) was performed. Indoor
laboratory tests typically used a frequency of 0.5 Hz, which corresponded to a measurement every
2 seconds, or 0.2 Hz (one measurement every 5 seconds). When testing moved outdoors, the
frequency changed to one measurement every minute.
18


3. Optimization
3.1 Luminol solution
Initially, the luminol solution contained only luminol dissolved in deionized water and sodium
hydroxide. The luminol concentration range tested was between lxlO'4 and lxlO'3 M. This range
was previously reported as providing the most chemiluminescence by Maeda, et al. [8] The
sodium hydroxide was necessary for the chemiluminescent reaction to occur. [10] For calibration
purposes, this solution sufficed. However, to further increase the sensitivity, stability, and
effectiveness of the solution, additional chemicals were added and the resulting solutions were
tested. The effect of concentration on the signal intensity was measured with respect to each
individual component. When a single part was tested, all other components were held constant.
3.1.1 Addition of sodium hydroxide
It was found that NaOH played three important roles with respect to the solution. First, it aided in
the dissolution of the solid luminol. When monitoring the preparation of the luminol solution, it
was observed that solid luminol takes a significant amount of time to dissolve. Maintaining an ice
bath around the luminol solution seemed to assist in dissolving the luminol, however, when NaOH
was added, the time required for complete dissolution was considerably decreased. Secondly, the
literature states that an alkaline environment must be present in order for the chemiluminescence
to take place. [8] Thirdly, as NaOH is added to the luminol solution, a resulting decrease of
interference due to CO2 is observed. [8] This conclusion is based on literature onlyno laboratory
tests were performed to verify this. The concentration of NaOH added was 0.05M. This solution
was added to the luminol/deionized water solution until a pH-11.
19


3.1.2 Addition of methanol
Previous literature noted that adding methanol had a positive effect on the sensitivity of the signal
and also increased the selectivity for N02. [4] A range of concentrations from 0.01-0.2% v/v was
tested. For this experiment, several known concentrations of N02 were flowed into the LMA-3
for each of the luminol solutions. The peristaltic pump tubing was flushed with DI water between
different methanol solutions. The tubing was then filled with the new methanol solution prior to
taking the measurement. In agreement with the results of Maeda, et al., the 0.05% methanol
solution resulted in the highest signal. Refer to Figure 3.1 for the results of changing concentration
of methanol added to luminol solution. Another factor taken into consideration was the stability
of the luminol solution over time. Figures 3.2, 3.3 show the degradation, or decreased signal, as a
result of varied methanol concentration in luminol solution. Figure 3.2 shows data from the
chosen concentration, 0.05% methanol. In Figure 3.3, the data show a larger decrease in signal by
the end of the 1-week testing period with a 0.10% methanol solution.
0.5 1 1.5 2 2.5 3 3.5
N02 concentration (ppm)
0.10%
IB 0.05%
A 0.03%
X0.01%
FIGURE 3.1 N02 SIGNAL AS A FUNCTION OF METHANOL CONCENTRATION
20


Day 1
SI Day 2
A Day 3
XDay 4
XDay 7
Figure 3.2 DEGREDATION OF 0.05% METHANOL IN LUMINOL SOLUTION
Day 1
HDay 2
A Day 3
XDay 4
XDay 7
N02 concentration (ppm)
Figure 3.3 DEGREDATION OF 0.10% METHANOL IN LUMINOL SOLUTION
21


3.1.3 Addition of sodium sulfite
Earlier literature suggested a considerable interference in N02 signal due to the presence of ozone
(O3). [8] Ambient O3 is found with N02 in comparable concentrations and will also cause
chemiluminescence when contact is made with luminol. Therefore, 03 must be removed to ensure
it is not producing any signal in addition to the N02 signal. An experiment was attempted to
determine what effect ozone has on the signal as detected by the LMA-3. An ozone generator was
set to produce a low concentration (4ppb) of ozone. The resulting air sample at this setting was
fed directly into the back of the LMA-3 for analysis. A signal was detected by the LMA-3,
indicating N02 was present in the sample. However, this air sample containing generated ozone
does not likely contain N02, thus the signal was a false indicator of N02. This would imply that
sampling ambient air that presumably contains both N02 and ozone would result in a higher
signala mix of both ozone and N02 unrepresentative of N02 exclusively. To minimize this
interference, sodium sulfite was incorporated into the solution. Sodium sulfite is used for 03
removal in water treatment [11] and appeared to be an effective additive to the luminol solution.
In a laboratory study, the ozone air sample was reacted with a luminol solution untreated with
Na2SC>3. The resulting signal was quite high, peaking at 1.45. This same concentration of ozone
was reacted with each of several other luminol solutions containing various amounts of Na2SC>3to
observe which solution had the greatest reducing effect on ozone signal, detected as an N02 signal
by the LMA-3. Data taken over a five-minute period was averaged and reported. The results may
be seen in Figure 3.4. While the data appeared to lack a linear relationship, it was obvious that
Na2S03 was successfully able to reduce the ozone signal. An attempt to reproduce this finding at
a later date was unsuccessful. The ozone analyzer did not appear to be reporting the correct ozone
concentration, and did not respond linearly when the ozone generator was increased linearly at
22


small intervals. Consequently, it is possible that the sample stream containing what was thought
to be 4 ppb actually had a higher concentration. These results, therefore, can only be considered
qualitatively. In addition, an ozone concentration closer to hourly average ambient concentrations
(30-50 ppb) should be used for this study in order to substantiate the efficacy of the sodium sulfite
solution. A qualitative conclusion that can be drawn from this data is that it supports the literature,
which indicates including sodium sulfite into the luminol solution will reduce the ozone
interference in N02 signal.
Concentration of Na2S03 (mols/L)
Figure 3.4 OZONE SIGNAL WITH LUMINOL CONTAINING SODIUM SULFITE
Another laboratory study involving sodium sulfite was performed with the goal of determining
what effect Na2S03 concentration had on N02 signal intensity. A range of 0M to 0.05M Na2S03
was developed to include concentrations noted in literature. An air sample containing a constant
concentration of N02 was reacted with theses different solutions. Tests concluded which
23


concentration of sodium sulfite produced the greatest signal. The 0.01M Na2S03 solution appeared
to be the most effective at obtaining the highest average signal when this luminol solution was
allowed to react with N02-containing air. See Figures 3.5, 3.6. The results from this experiment
also lacked a linear relationship between concentration and signal, however, the 0.01M solution
proved to be successful at both decreasing ozone interference and increasing N02 signal intensity.
Therefore, this concentration was chosen for the optimized luminol solution.
150 200 250 300 350
Time (s)
Figure 3.5 EFFECT OF CONCENTRATION OF Na2S03 ON SIGNAL
0 0.005 0.01 0.02 0.05
Concentration Na2S03 (mols/L)
Figure 3.6 AVERAGE SIGNALS OBTAINED FOR EACH CONCENTRATION OF Na2S03
24


3.2 Luminol flow rate
The rate at which the luminol solution flows through the LMA-3 is important for two reasons.
First, the operator must set the flow correctly to ensure a fresh luminol solution is flowing through
the chamber at all times, providing an opportunity for N02 to react with new solution. Second, it
is critical to know what effect, if any, the flow rate has on the signal intensity. The external
Masterflex peristaltic pump has a manual dial that allows for a range of low flow rates to be
achieved. Flow rates from 0.08 mL/min to 0.59 mL/min were tested to determine whether or not
there is a relationship between flow rate and signal intensity. It was concluded that the signal
intensity was independent of flow rate. Figure 3.7 verifies this conclusion. An additional graph
plotting flow rate versus slope of regression line may be seen in Figure 3.8. This graph further
supports the conclusion that the flow rate had virtually no effect on signal intensity. Therefore, in
order to minimize liquid waste, the lowest flow rate attainable was used, 0.08 mL/min.
0.08 mL/min
S3 0.12 mL/min
A 0.17 mL/min
X0.27 mL/min
X0.34 mL/min
0.44 mL/min
0.50 mL/min
0.59 mL/min
Figure 3.7 FLOW RATE EFFECTS ON SIGNAL
25


IS
14
13
w 12
fS
%
Z. 11
1 .0
id
C
c
J2
A
c
&
01 0.2 0.3 0.4 0.5 0.6 0.7
luminol flow (mL/min)
Figure 3.8 SLOPE AS A FUNCTION OF LUMINOL FLOW
3.3 Sample volume flow rate
The air sample flow rate was altered to investigate effects on signal intensity. A vacuum pump is
connected to the back of the LMA-3 via the air outlet. This pump works by drawing air in a
controlled volume through the instrument. The flow rate used was in units of standard cubic feet
per hour (scfh). One scfh is equal to 0.48 liters per minute. While earlier literature [4] stated an
optimal vacuum flow rate of 1.5 L/min, it was found that the greatest signal was obtained at a
volume flow rate of 1.05 scfh, or 0.5 L/min. The results of this study are shown in Figure 3.9.
26


1.1
1
0.9
n
6 0.8
1/*
0.7
0.6
O.S






T
O.S 1 IS
Air sample flow rate (scfti)
2
Figure 3.9 OPTIMAL VACUUM FLOW RATE AS DETERMINED BY GREATEST
ATTAINABLE SIGNAL
3.4 Photomultiplier tube voltage
Some studies suggested changing the photomultiplier tube (PMT) voltage depending on what
concentration range was being tested [3]. However, this study only used one working PMT
voltage, which was determined by observing signal change as a function of PMT voltage change.
The NO2 concentration remained constant throughout this test. The result was an exponential
increase in signal until the maximum PMT voltage achievable by the LMA-3 was reached, at -
810.8V. Refer to Figure 3.10 for this graph.
27


Signal
PMT voltage
Figure 3.10 SIGNAL INTENSITY AS A FUNCTION OF PMT VOLTAGE
28


3.5 Optimal parameters
The following table shows each of the optimized parameters.
Table 3.1 Optimal parameters for ambient sampling
Instrumental Luminol solution
PMT voltage -810.8 V Reagent Concentration
luminol flow rate 0.08 mL/min luminol 8.33 x 10-4 M
ambient sample flow rate 0.5 L/min NaOH 0.05 M until pH~10
DAQ voltage range 2V methanol 0.05% (v/v)
Sampling frequency 1/60 Hz Na2S03 0.01 M
29


4. Calibration
4.1 Gas calibration with N02
A gas cylinder containing N02at a concentration of -200 parts per million (ppm) was fitted with a
regulator and length of tubing. The tubing was connected to the back of a gas dilution system
located in a rack. House air was also connected to the back of this box. Both the gas and air flow
controllers, described previously in section 2.3, were calibrated before use. The two gases (air and
200 ppm N02) were allowed to mix at a manually controlled ratio and left the system as a mixture
into the gas dilution manifold. Tubing was inserted above the diluted inlet to catch the mixture
(at a set flow rate) and that sample was flowed into the back of the LMA-3. Knowing the
approximate initial concentration of the standard gas cylinder and the rate at which the gas flowed
along with the rate of the air flow, a standard calibration curve could be generated. See Figure 4.1.
This initial calibration curve was produced using an un-optimized luminol solution. The luminol
solution contained luminol dissolved in deionized water, 0.05% MeOH and 0.05 M NaOH added
until the pH was between 10 and 11. The sodium sulfite was not incorporated into the solution
until a later date.
Figure 4.1 CALIBRATION CURVE PRODUCED WITH LUMINOL SOLUTION MODIFIED
WITH METHANOL
30


4.2 Gas calibration with NO and 03 reaction
To create very low precise concentrations of N02for calibration, an ozone generator was used in
conjunction with a cylinder of NO. The reaction between NO and 03 results in N02 being
produced. (This reaction may be observed in equations (1) and (2) in the Section 1.4
Measurement Techniques.) Ozone is generated in low concentrations, single parts per billion
(ppb). An ozone generator is used in tandem with an ozone analyzer to accomplish this. The
ozone analyzer displays the concentration of ozone that was generated in the sample. When
creating the low concentration calibration curve for N02, a set amount of ozone is allowed to mix
with an excess of NO. Reacting a small, known amount of O3 with a large excess of NO ensures
that all ozone has reacted to produce N02. By setting the initial concentration of ozone produced
and assuming that all ozone has been used up by the excess NO, one can determine a
concentration of N02 by analysis of the signal acquired. Figure 4.2 illustrates this calibration on
the ppb level. The luminol solution used in this study was not optimized; it did, however, contain
the optimal amounts of NaOH, luminol, and methanol. Sodium sulfite was not incorporated into
the solution at this point.
Figure 4.2 NO + 03 CALIBRATION CURVE
31


4.3 Determination of zero
Recording a signal from air that does not contain N02 creates a baseline (also considered a zero or
blank). To make certain that the zeroing air is actually N02-free, a scrubber is used between the
sample collection inlet and the LMA-3. The scrubber is fashioned out of a length of PVC pipe
filled with Ascarite, which is able to successfully remove N02 from air. Ascarite is a type of
sodium hydroxide coated silica, more commonly known as a C02 scrubber. The scrubbed air then
proceeds through tubing attached to the other end of the scrubber and enters the inlet on the LMA-
3 for analysis. The signal produced from the N02-free air is considered zero (although the actual
reading may not be 0) and this number is subtracted from the signal obtained from the N02-
containing air sample. This number is adjusted with respect to the zero and is reported as the
signal on a graph. The following figure (4.4) illustrates the effectiveness of the Ascarite scrubber.
Decreased signal troughs are zero (baseline) measurements of N02-containing air that has
passed through the scrubber, while a higher, non-zero signal is obtained from air allowed to
bypass the Ascarite scrubber. The zero may change slightly if data collection parameters are
changed (i.e. voltage range), but remains at or near the 0 line.
0 02
0.015
0.01
0 005
C
t
55 0
-0.005
-0.01
-0.015
. r&BL
.


\md

1000 1100 1200 1300 1400 1500 1600 1700
Time (sec)
Figure 4.3 ASCARITE SCRUBBER EFFICIENCY
32


5. Data and Results
5.1 Setting collection cycles
Measurements were collected on a schedule. One cycle consisted of a set amount of sampling
time and zeroing time. At the two Colorado testing sites, sampling occurs for twenty-three of the
twenty-four hours in each day. The zeroing hour is different for the two sites and is either from
1:00-2:00 AM (CAMP) or 2:00-3:00 AM (Welby). One drawback to this schedule is that no data
are ever collected for that hour of the day at that particular site, and therefore no data exists for
that site during that early morning hour. For this research, the DAQ is set to zero the instrument
for 12 minutes followed by a sampling period of 48 minutes. Setting the DAQ to sample for 48 of
the 60 minutes in an hour secures 48-minutes of sampling in an hour, while continuing to zero
throughout the day. With this collection arrangement, there is no need to suspend sampling for an
hour for the purpose of zeroing.
5.2 Indoor laboratory data
The first trial of air sampling and analysis was conducted inside a research laboratory on the
University of Colorado Denvers downtown campus. The data are shown in Figure 5.1. Sampling
began on February 15,2011, at 2:00 PM and ran for 24 hours. This graph contains raw data, as
the zero is visible throughout the measurement period. Measurements were taken at a frequency
of 0.2 Hz. The cycle time was 12 minutes: a 2-minute zero followed by a 10-minute sampling
period.
33


Time (s)
Figure 5.1 INDOOR LABORATORY DATA AT THE UNIVERSITY OF COLORADO
DENVER OVER A 24-HOUR PERIOD
The second trial of indoor air sampling started on February 26, 2011, at 11:58 AM and proceeded
for a 24-hour sampling period. Data are shown in Figure 5.2. The cycle time was 12 minutes: a 2-
minute zero followed by a 10-minute sampling period. The sampling frequency was 0.2 Hz. From
12:00 PM (measuring began after 2 minute zero) until 4:00 PM, no observable signal was attained.
A strong increase in signal was observed at approximately 4:00 PM on a Friday afternoon. This
signal remained above zero until approximately 5:30 AM the next morning.
At this point, the signal decreased back down to zero and remained until the completion of the
sampling. The reason behind this observation is not fully known, however, during an ambient
study, the peristaltic pump ceased rotation. A similar dramatic decrease in a signal and continued
reading of zero was observed in that particular ambient study.
34


Signal
1400
1200
10 00
800
600
400
200
000
-2.00

I jl
JL .


. 1* 7 C ,***.. _**_** *
^15000^
Time (s)
Figure 5.2 INDOOR LABORATORY DATA AT THE UNIVERSITY OF COLORADO
DENVER OVER A 24-HOUR PERIOD: DAY 2
35


5.3 Outdoor sampling
The LMA-3 was taken outdoors to test ambient air at the on-campus air monitoring trailer. This
sampling site is located on the comer of a campus parking lot on the Auraria Campus, located in
downtown Denver, Colorado.
Figure 5.3 MAP OF AURARIA CAMPUS


5.4 Data collection
For ambient sampling, the LMA-3 was allowed to run continuously for 3 days. The goal was to
determine whether a diurnal relationship exists. This parking lot typically fills in the morning
hours between approximately 7-9AM. As mentioned earlier, CAMP is located in downtown
Denver. Although the Auraria Campus collects samples several blocks from CAMP, a correlation
between measurements from the two downtown Denver sites should exist. Figures 5.4, 5.5 contain
data collected from the first of several outdoor tests. Beginning at 8 AM on May 4, 2011,
sampling continued for 24 hours. Raw data are plotted in Figure 5.4. In Figure 5.5, the average
signal obtained in a 1-hour period is plotted against the hour of sample collection, (i.e. From 8-
9AM (hour 1), the average signal was about 1.75).
Time (min)
Figure 5.4 RAW DATA FROM AURARIA CAMPUS: 0800 5/4/2011 0800 5/5/2011
37


Sampling hour (1-8AM)
Figure 5.5 HOURLY AVERAGE DATA FROM AURARIA CAMPUS: 0800 5/4/11- 0800 5/5/11
The graph shown above indicates a signal, or average signal, but do not specify a concentration.
This is due to the fact that no calibration curve exists for the final optimized luminol solution as it
was used in these outdoor studies. The NO2 and NO cylinders used previously were completely
used up and the only N02 available was in a cylinder of unknown concentration. The cylinder was
too concentrated, so N02 was diluted with N2 into a new cylinder via a gas dilution procedure,
which included evacuation of a new cylinder, introduction of N02, and pressurization with N2.
While the signal should be proportional to the N02 mixing ratio in the sample, one may not
calculate an exact N02 concentration based upon this information or by the use of previously
generated calibration curves. The signal may be used as a reference point to compare other
measurements upon. An Excel spreadsheet template was created that analyzed the raw data into
averages. For a breakdown of the template, refer to Figure 5.6.
First, 24 hours of raw data is pasted into the template. The first picture (top left) shows this step.
The template contains formulas for averaging both the ambient samples and the zero samples.
This is shown in the second picture (top right). The zero average for each cycle is subtracted from
38


the signal average for each cycle to produce an adjusted signal for that cycle. Outdoor cycle times
were 48-minute sample period followed by a 12-minute zero period. Finally, the averages were
plotted in a graph and were ready for comparison to state data.
0 A B
Sample Nu Dale
2 1 5/4/2011
3 2 54/2011
4 3 5*72011
EM 4 54/2011
6 5 5*4/2011
7 8 5*4/2011
8 7 5*4/2011
9 8 5*4/2011
XO 9 5/4/2011
IX 10 54/2011
12 11 5*4/2011
13 12 5/4/2011
14 13 54/2011
IS 14 54/2011
16 15 54/2311
17 16 54/2011
18 17 54/2011
19 18 54/2011
20 19 54/2311
21 20 54/2311
22 21 54/2011
23 22 54/2311
24 23 5*4/2011
c 6
Tme $
7.57.32 Ak 0 292969
7.57.39 AK 0.195312
7.58.39 Ak 0 585937
7.5939 AA 0927734
800:39 Ak 1416016
8:01.39 Ak 1.123047
8 02.39 Ak 1269531
8 93.39 Ak 1611326
604.39 Ak 1464944
8.05.39 Ak 1.318359
8.06 39 Ak 1 5625
8.37:39 Ak 1606641
8 08.39 Ak 1611328
8.09.39 Ak 2.001953
8.10.39 Ak 1611328
8 11 39 Ak 1708964
6.12.39 Ak 2146437
6.13.39 Ak 166C156
6 14.39 AK 1 706964
615.39 Ak 1513872
8:18.39 Ak 1904297
817.39 Ak 2 246094
618.39 Ak 2CSC781
O A B c 0
1 Sample Nu Date T me
Hi 1 54/2011 7.57.32 A*! "0552559
3 2 54/2011 7:57:39 Ak 0195312
4 3 54/2011 7.58:39 Ak 0 585937
s 4 54/2011 7:59:39 Ak 0.927734
6 5 54/2011 8 00.39 Ak 1.416016
7 6 54.2011 8:01.39 Ak 1 123347
8 7 54/2011 8.02:39 Ak 1 269S31
9 8 54/2011 8:03.39 Ak 1 611328
10 9 54/2011 8.34.39 Ak 1 464844
11 10 54.2011 8.05.39 Ak 1 318359
12 11 54/2011 8.36.39 Ak 1 5825
13 12 54/2011 8:07.39 Ak 1.806641
B-9AM 9-10 AM
175665 2 201335
r 1.233927 r 1 312256
r -0.42725 r *0.52897
14 15
9-10PM KM1PM
1681173 1 841227
J[1 340739 ^0 941976
l.frumb 'lO-HAM 1TT2PM
1 723228 129598
r 1 075236 r 2 C38574
r -C.5127 r -C.37028
16 17
11-12AM 12-1 AM
1 587931 2 408854
r 1 647132'2 ^ 11616 1 340739 C 94 W6 9 C 957235 0 9W321 1.243637 1.45674 1 342773 1 232929 9 1362101 1 C96596 1 15153
' -C 38952 -v 06952 -0.36249 ' -0354 r -C.46794' 43 43131 -0 50049 -0.512?' -043131 -C.55746' 4352897 r -0.49235' -C467M
10C 2.00 3 90 4 CO 500 6 00 7.0C 800 90S 10.CC 1100 12.X 13.3C
8- 9AU 9-10AM 16-11AM 11-12PM 12-1 PM 1-2PM 2-3PM 34PM 4-5PM 5-6 PM 6-7PM 7-6PM I-9PM
1 75665 2231335 1,723226 1.28566 1 425171 1393636 1 746125 1971435 1 774368 165C382 1891072 1588949 161&*
r 1233827 1 312256 r 1.075236 f 2 036574'2 636753*' 1 567566 9 1 423136 9 1 741536 9 2 346636 r 2 252197 9 2 091471
9 -C 42725 r -0 52867 9 -05127 9 -047028' 4.43132' -0 48235' -C.476C?' 4.54525' -0.43132' -0.53304 -0 43501
14 15 16 1 7 16 19 20 21 22 22 2*
9- 10PM 16-11PM 11-12AM 12-1AM 1-2 AM 2-3AM 3-4AM 4-5AM S-6AM 6-7AM 7-6AM
1661173 1 641227 1 567931 2 406654 3 070068 2 C58936 199921 2296764 2 760151 2 765237 2 526465
35 (Date] Auraria Campue



I 3 * *

05

C5C 5X O.CO 5.X XC0 Santpt* Hwr
Figure 5.6 EXCEL TEMPLATE STEPS FOR RAW DATA ANALYSIS
The next outdoor study was modified slightly in order to better relate the data to known data
collected at a site located in downtown Denver. The data obtained at the air monitoring trailer on
the Auraria Campus was compared to data collected at CAMP and Welby, a N02 two sampling
39


sites described previously. Instead of starting the sampling (and subsequent graphs) at 8AM, the
data from 12AM to 11:59PM on a single day is reported. On May 11, 2011, the first full day of
data was collected and compared to data collected at the state sampling sites. Figure 5.7 contains
raw data from 5/11/11. On this day, one is able to notice two peaks in the data. In Figure 5.8, the
hourly average signal is plotted against time. Due to the low signal obtained, a factor of 3.33 was
applied to the signal to magnify it in order to be on scale with the hourly average reported on that
date from CAMP and Welby.
3 t
m
c
o>



I

+i-l<1
i /
-1.5 J
0 200 400 600 800 1000 1200 1400
Time (min)
Figure 5.7 RAW OUTDOOR DATA COLLECTED AT AURARIA 5/11/11
40


Auraria
Sample Hour
Figure 5.8 HOURLY AVERAGE COMPARISON OF AURARIA, CAMP AND WELBY DATA
One way to verily data as being reasonable is to compare two sets data collected on the same day
from similar collection sites. A linear correlation should exist. As shown in Figure 5.9, there is a
very little linear correlation between CAMP and Auraria. The linear coefficient is 0.28, which is
drastically lower than the goal of 1.0.
CAMP N02 concentration (ppb)
Figure 5.9 DATA COMPARISON FOR LINEAR RELATIONSHIP BETWEEN CAMP AND
AURARIA
41


The following figure (5.10) shows data collected over a 3-day period, containing the data shown
above and continuing for the two days following. Sampling began on May 4, 2011, and continued
through May 6, 2011. The next figure (5.11) shows inconsistencies with the data when compared
to Figure 5.10. First, the zero is drifting. Also noticeable is the very sharp decrease in N02 signal
around 2200 minutes. The drifting zero is an indicator that the Ascarite in the scrubber is in need
of replacement. The decrease in signal and continued zero reading indicates a peristaltic pump
issue. This problem was mentioned earlier; the peristaltic pump stopped rotating, and therefore,
no luminol solution was able to flow into the chamber to react with the air sample. This signal is
essentially a non-scrubbed zero.
5,5
-100 400 900 1400 1900 2400 2900 3400 3900 4400
Time (min)
42


Concentration (ppb)
Figure 5.10 RAW DATA FROM AURARIA 5/4/11-5/6/11
Figure 5.11 RAW DATA FROM AURARIA 5/29/11-5/31/11
Figure 5.12 compares the hourly averages reported for CAMP and Welby measurement sites over
a 3-day period in early May. A correlation is visible for most of the sampling during the three
days. CAMP measurements are typically higher, but not by a large margin.
60 CAMP $ A Welby
bU " 9 it £ £
4U "
iU * 0 9
ZU a* 1. ...*.**
1U ~ o - ''a 9..,/ 9 A.*** /sa * A' A -,niinf l ? r- S i M.. n
0 10 20 30 40 50 60 70
Sample hour
Figure 5.12 DATA FROM CAMP AND WELBY 5/4/11-5/6/11
43


6. Conclusions and future work
A reasonable method for the measurement of nitrogen dioxide in ambient air was described. This
method involves the luminol chemiluminescent reaction between nitrogen dioxide and 5-amino-
2,3-dihydro- 1,4-phthalazine dione (luminol), which produces a chemiluminescence lOOOx
stronger than that of the currently used method of NO/O3 chemiluminescence, according to the
literature. Additionally, this method does not require reduction of N02 species prior to analysis, as
it is a N02-specific measurement technique. The ambient samples can be analyzed in near-real
time.
Several effective modifications were added to the existing instrument, the Scintrex LMA-3
Analyzer. First, a Masterflex C/L peristaltic pump capable of continuous liquid solution flow was
added externally to the LMA-3 to replace the internal, non-programmable pump. Larger luminol
solution and waster containers were added to increase the volume capacity, and thus reduce
maintenance to the solution and apparatus during sampling. Lab View, a data acquisition system,
was added. The hardware included a voltage-controlled valve that opened and closed on a
schedule set by the operator. This addition also meant that a zero baseline could be obtained
throughout the sampling period. State sampling sites collect data for 23 of the 24 hours in a day,
and use one whole hour of each day to zero. The modified hardware, which employs an Ascarite
scrubber, allows a zero to be collected throughout each hour, eliminating the need to refrain from
sample collection for any extended period of time. The luminol solution was optimized by adding
or varying components of the solution including methanol, sodium sulfite, and sodium hydroxide.
The optimized solution allowed for the greatest sensitivity and selectivity toward nitrogen dioxide.
While this method appears to be a suitable for continuous nitrogen dioxide measurements, there
are several facets of both the method and analysis that need to be addressed. It is clear that
44


although this method is useful, the comparisons between this data and state-collected data are
insignificant until an updated calibration curve with a standard N02 cylinder and the modified
solution is created. Because there is no additional testing method running concurrently with the
luminol method at Auraria, there is no way to verify the validity of the data collected with the
LMA-3. Likewise, a further investigation might include testing the luminol method at one or both
of the state sites, CAMP and Welby, to determine reliability of data collected with the LMA-3
analyzer. The sodium sulfite effect on ozone signal experiment should be redone to gauge the
significance of the sodium sulfite addition to the luminol solution.
One recurring issue with the apparatus was the laptop computer, which appeared to recognize the
software differently than the desktop that the data collection originated on. This issue caused
other problems with data collection, since the laptop did not report the value displayed on the
LMA-3. As a result, parameters within the software had to be changed, including voltage range
and the addition of a multiplier to correct the output.
The external peristaltic pump, on several occasions, stopped rotating, and therefore stopped fresh
luminol flow into the chamber within the LMA-3. When this occurred, no useful data was being
collected. Further work might include replacing the peristaltic pump, changing the type of tubing,
or using a different make of pump.
Taking the known optimal parameters and running tests to verify their reliability would be a
reasonable step throughout the experiment. This would ensure that every aspect was running
consistently, or would notify the operator of any discrepancies.
Finally, an interesting experiment would be to analyze the exhaust from different vehicles.
Comparing N02 levels from vehicles that bum different fuels, or different models of vehicles from
the same make are examples of experiments that could be ran. This, of course, would occur after
45


validation of LM A-3 data. There are many applications that this method of analysis would be
useful for. This technique is selective, the instrument is portable and user-friendly, and as a fairly
inexpensive system, this could be operated nearly anywhere.
46


REFERENCES
1. 2009 Colorado Air Quality Data Report
2. Gaffney, J.S., R.M. Bomick, Y.-H Chen, and N. A. Marley. Capillary Gas
Chromatographic Analysis of Nitrogen Dioxide and PANS with Luminol
Chemiluminescent Detection. Atmospheric Environment. 32 (1998): 1445-1454.
3. Schiff, H. I.; Mackay, G. I.; Castledine, C.; Harris, G. W.; Tran, Q. (1986). A sensitive
direct measurement N02 instrument. In Proceedings of the 1986 EPA/APCA symposium
on measurement of toxic air pollutants; April; Raleigh, NC (pp. 834-844). Research
Triangle Park, NC:: Environmental Protection Agency, Environmental Monitoring
Systems Laboratory. 023025.
4. Wendel, Gregory J., Stedman, Donald H., Damrauer, Lenore, and Christopher A.
Cantrell. Luminol-Based Nitrogen Dioxide Detector. Anal. Chem. 55 (1983): 937-940.
5. Sigsby, J.E., Jr., F. M. Black, T.A. Bellar, D. L. Klosterman. Chemiluminescent Method
for Analysis of Nitrogen Compounds in Mobile Source Emissions (NO, NO2, and NH3).
Environ. Sci. Technol. 7 (1973): 51
6. Winer, A. M., J. W. Peter, J. P. Smith, J. N. Pitts, Jr. Response of Commercial
Chemiluminscent Nitric Oxide-Nitrogen Dioxide Analyzers to Other Nitrogen-
Containing Compounds. Environ. Sci. Technol. 7 (1974): 1118-1121.
7. Cassidy, D. T. and J. Reid. Atmospheric pressure monitoring of trace gases using
tunable diode lasers. Applied Optics. 21 (1982): 1185-1190.
8. Maeda,Y., Aoki, K., Munemori, M. Chemiluminescence Method for the Determination
ofNitrogen Dioxide. Anal. Chem. 52(1980): 307-311.
9. NOx Box Instructions and Troubleshooting Tips. Billow, T., E. Davidson, S. Hall, D.
Herman, J. Moen.
http://pangea.stanford.edu/research/matsonlab/Protocol/NOxBoxProtocol831QO.pdf
10. Stedman, D. H., Wendel, G. J., Cantrell, C.A., and L. Damrauer. A Novel Nitrogen
Dioxide Detector. Proceedings National Symposium On Recent Advances In Pollutant
Monitoring Of Ambient Air And Stationary Sources. (1984): 14-22.
11. Powell, Sheppard T. Laboratory and Plant: The Design and Operation of Ozone Water
Purification Systems. J. Ind. Eng. Chem. 8 (1916): 632-636.
47