Citation
The design and control of an atmospheric formaldehyde analyzer

Material Information

Title:
The design and control of an atmospheric formaldehyde analyzer
Creator:
Mayzes, Marc Christopher
Publication Date:
Language:
English
Physical Description:
1 electronic file : ;

Subjects

Subjects / Keywords:
Formaldehyde -- Analysis ( lcsh )
Air -- Pollution -- Measurement ( lcsh )
Air -- Pollution -- Measurement ( fast )
Formaldehyde -- Analysis ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
A fluorometry based gas analyzer was developed for the detection of atmospheric formaldehyde for field sampling. Formaldehyde was collected using a diffusion scrubber containing a permeable Nafion tube inside 1/4 inch Teflon tubing with dilute sulfuric acid as a scrubbing solution. The formaldehyde was mixed with 0.65 M ammonium acetate and 2,4-pentanedione to produce the fluorescent derivative 3,5 diacetyl-1,4 dihydrolutidine (DDL). When excited by 405 nm light, the DDL was quantified by fluorescence detection with peak emission near 510 nm. A custom software solution was developed for instrument control and measurement of fluorescence signals and improvement of signal to noise ratios. The combination of hardware and software components produced and instrument capable of detecting formaldehyde concentrations above 1.22 ppbv in one minute intervals.
Bibliography:
Includes bibliographical references.
Thesis:
Department of Chemistry
Statement of Responsibility:
by Marc Christopher Mayzes.

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:
858280221 ( OCLC )
ocn858280221

Auraria Membership

Aggregations:
Auraria Library
University of Colorado Denver

Downloads

This item has the following downloads:


Full Text
THE DESIGN AND CONTROL OF AN ATMOSPHERIC FORMALDEHYDE
ANALYZER
by
Marc Christopher Mayzes
A.G.S., Arapahoe Community College, 2004
B.S., University of Colorado Denver, 2008
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado Denver in partial fulfillment
of the requirements for the degree of
Master of Science
Chemistry
2012


This thesis for the Master of Science degree by
Marc Christopher Mayzes
has been approved for the
Master of Science
Chemistry
by
Larry G. Anderson, Advisor
John A. Lanning
Hai Lin
2012
Date: January 31,


Mayzes, Marc Christopher (M.S., Chemistry)
The Design and Control of an Atmospheric Formaldehyde Analyzer
Thesis directed by Professor Larry G. Anderson
ABSTRACT
A fluorometry based gas analyzer was developed for the detection of
atmospheric formaldehyde for field sampling. Formaldehyde was collected using
a diffusion scrubber containing a permeable Nation tube inside Va inch
Teflon tubing with dilute sulfuric acid as a scrubbing solution. The
formaldehyde was mixed with 0.65 M ammonium acetate and 2,4-pentanedione
to produce the fluorescent derivative 3,5 diacetyl-1,4 dihydrolutidine (DDL).
When excited by 405 nm light, the DDL was quantified by fluorescence detection
with peak emission near 510 nm. A custom software solution was developed for
instrument control and measurement of fluorescence signals and improvement of
signal to noise ratios. The combination of hardware and software components
produced and instrument capable of detecting formaldehyde concentrations
above 1.22 ppbv in one minute intervals.
The form and content of this abstract are approved. I recommend its
publication.
Approved: Larry G. Anderson


TABLE OF CONTENTS
Tables..................................................................vii
Figures................................................................viii
Chapter
1. Introduction...........................................................1
1.1 History of Formaldehyde...........................................1
1.1.1 Sources of Formaldehyde...........................................1
1.1.2 Health Concerns of Formaldehyde...................................4
1.2 Measurement Techniques for Gaseous Formaldehyde...................6
1.2.1 EPA Method TO-11A.................................................6
1.2.2 Continuous Flow DNPH Methods......................................8
1.2.3 Hantzsch Reaction Derivitization..................................9
2. Experimental..........................................................12
2.1 Field Formaldehyde Analyzer......................................12
2.1.1 Pneumatic System.................................................12
2.1.2 Liquid Handling System.........................................14
2.1.3 Diffusion Scrubber.............................................16
2.1.4 Gilson Model 121 Fluorometer...................................18
IV


2.1.5 Data Acquisition System
19
2.2 Laboratory Formaldehyde Analyzer.....................................21
2.2.1 Pneumatic System..................................................21
2.2.2 Liquid Handling System............................................22
2.2.3 Perkin Elmer LC 240 ..............................................23
2.2.4 Data Acquisition System...........................................23
2.3 Experimental Procedure.............................................24
2.3.1 Reagents..........................................................24
2.3.2 Calibration of Formaldehyde Analyzers.............................24
3. Results and Discussion...................................................26
3.1 Reaction Chamber Optimization........................................26
3.2 Calibration Results..................................................27
3.2.1 Gilson Model 121..................................................27
3.2.2 Limit of Detection................................................30
3.2.3 Tylan FC-260 Mass Flow Controller.................................32
3.3 Collection Efficiency................................................34
3.4 Sampling Results.....................................................38
3.4.1 Indoor Sampling...................................................38
3.4.2 Outdoor Sampling..................................................41
V


4. Conclusion.............................................................52
References................................................................56
VI


LIST OF TABLES
Table
Table 1: Partial List of Carbonyls Quantified With TO-11A....................7
Table 2: Standard HCHO Solutions.............................................25
Table 3: Hantzsch Reaction Completion Data...................................26
Table 4: Average Response voltage and Concentration..........................29
Table 5: Statistics for Zero Standard Solution...............................31
Table 6: Response Per ppbv...................................................32
Table 7: Mass Flow Controller Calibration Data...............................33
Table 8: Statistics for Collection Efficiency................................37
Table 9: Hourly Indoor HCHO Concentrations...................................39
Table 10: Outdoor Hourly Data (1/1/2012)...................................44
Table 11: Outdoor Hourly Data (1/2/2012)...................................46
Table 12: Outdoor Blank Sample Response....................................49
Table 13: Hourly HCHO Concentration (ppbv) Summary..........................51
vii


LIST OF FIGURES
Figure
Figure 1: The Photolysis of Ozone..............................................3
Figure 2: The Production of HCHO From VOCs.....................................4
Figure 3: The Derivitization of HCHO with DNPH.................................7
Figure 4: The Hantzsch Reaction................................................9
Figure 5: Reaction of Formaldehyde With Fluoral-P.............................11
Figure 6: Pneumatic Layout of Field Analyzer..................................13
Figure 7: Sampling Process of Liquid Handling System..........................15
Figure 8: Construction of Reaction Heater.....................................15
Figure 9: Attachment of Nafion to Steel Tubing................................16
Figure 10: Elbow Union Layout.................................................17
Figure 11: Dual Channel Diffusion Scrubber....................................18
Figure 12: Absorption Scan of Emission Filter.................................19
Figure 13: LabView Interface for Field Analyzer...............................20
Figure 14: Laboratory Scrubber Configuration..................................22
Figure 15: Lab Analyzer Liquid Handling System................................22
Figure 16: Optimization of Reaction Heater....................................27
Figure 17: Gilson Model 121 Calibration Data..................................28
Figure 18: Gilson Model 121 Calibration Curve.................................30
viii


34
35
36
38
39
41
42
43
44
45
45
46
47
49
50
Mass Flow Controller Calibration Curve
Summary of Collection Efficiency Data.......
Blank Solution Data (Ten Hours).............
Indoor Air Sampling Results.................
Hourly Average for Indoor Data..............
Gas Flow Rate (12/21/2011)..................
Outdoor Sampling Results (12/24/2011).......
Outdoor Sampling Results (1/1/2012).........
Hourly HCHO Concentrations (1/1/2012).......
Gas Flow Rate (1/1/2012)....................
Outdoor Sampling Results (1/2/2012).........
Hourly HCHO Concentrations (1/2/2012).......
Gas Flow Rate (1/2/2012)....................
HCHO Concentrations (1/7/2012 1/10/2012)
Decreasing Trend of Span Concentration......
IX


1. Introduction
1.1 History of Formaldehyde
Formaldehyde (HCHO) was first discovered in the late nineteenth century
through the work of Butlerow and Hoffman. 1( 2 Since its discovery,
formaldehyde has been a major component of industrial processes. Indeed,
shortly after its initial discovery, formaldehyde was found to be easily produced
and had a wide range of uses from the development of industrial resins to
common use as a disinfectant. In more recent decades, concerns have been
raised regarding the potential hazards of formaldehyde and other chemicals to
human health. Numerous studies have been conducted in the analysis of
formaldehyde toxicity that has lead to the eventual classification of the
compound as a likely carcinogen.3'4'5
1.1.1 Sources of Formaldehyde
The simplest of the aldehydes, formaldehyde is a ubiquitous compound that has
numerous chemical interactions with the environment. It is important to
understand the sources and sinks of formaldehyde and other volatile organic
compounds (VOCs) in order to understand the importance of the chemical in
environmental interactions.
Formaldehyde is widely used within the chemical industry and is used in the
manufacturing processes for resins, foam insulation and fabric processing. In
l


fact, the particular usefulness of this compound has made it one of the most
used chemicals in the manufacturing industry. In the cases of formaldehyde-
based resins used in the manufacturing of items such as particle board and other
pressed wood products, the resulting products have had the greatest impact on
the quality of indoor air. Formaldehyde based resins made with urea, melamine
and phenol have been classified as formaldehyde sources either by the release of
free formaldehyde residues or the decomposition of the formaldehyde based
resin into a gaseous pollutant. 3' 4'5' Recently, in the aftermath of Hurricane
Katrina which devastated the city of New Orleans in 2005, the awareness of poor
air quality due to high concentrations of formaldehyde has increased when
emergency housing was found to have high levels of the compound inside.2
Industrial manufacturing is not the only source due to human activity.
Formaldehyde is a common byproduct for incomplete combustion processes such
as those in automobile engines and power production. 3'5 In the 1980's the
release of the pollutant to the atmosphere from mobile sources was estimated at
666 million pounds per year in the United States. 4 This estimation does not
include the potential contribution by stationary sources such as power plants
fired by coal, natural gas and oil. The combustion of fossil fuels in any case has
had the greatest impact on outdoor air quality as a result of human activity.
2


Natural sources of VOCs also contribute to the overall level of atmospheric
formaldehyde. Seco, et. al, described the characteristics of biogenic sources of
short-chain oxygenated VOCs such as formaldehyde. Plants naturally emit VOCs
into the atmosphere as part of normal metabolic processes 6. Other chemical
species emitted by natural biogenic sources, such as terpenes, also play an
important role in the production of atmospheric formaldehyde through a reaction
with radicals in the air.7
Non-oxygenated VOCs play an important role in the photochemical production of
HCHO in the atmosphere as detailed by Atkinson.7 One of chemical interactions
resulting in the production of atmospheric formaldehyde involves the interaction
of non-oxygenated VOCs with the photochemically produced hydroxyl radical.
One process by which the hydroxyl radical is generated is through the photolysis
of tropospheric ozone.
03 + hu -> 02 + 0( XD) (A < 335nm) (1)
0( XD) + M -> 0( 3P) + M (M = N2, 02) (2)
0( 3P) + 02 + M -> 03 + M (M = air) (3)
0( XD) + H20 -> 20H (4)
FIGURE 1: THE PHOTOLYSIS OF OZONE
3


When ozone is exposed to ultraviolet radiation in the presence of water in the
troposphere, an excited oxygen atom is produced as demonstrated in Figure 1.
Eventually, the combination of other atmospheric species produces the hydroxyl
radical.
0H + CH4 -J h2o + ch3 (5)
M ch3 + 02 -> ch3o2 (6)
CH302 + NO -> ch3o + no2 (7)
ch3o + o2 - -> HCHO + H02 (8)
H02 + NO OH + N02 (9)
FIGURE 2: THE PRODUCTION OF HCHO FROM VOCS
When the resulting hydroxyl species encounters VOCs in the atmosphere,
another series of reactions occur eventually leading to the production of
formaldehyde as well as more hydroxyl radicals. Figure 2 outlines the process of
producing formaldehyde through the oxidation of methane by the hydroxyl
radical and oxides of nitrogen (NOx) that are another important species in
atmospheric pollution.7
1.1.2 Health Concerns of Formaldehyde
Due to its ubiquitous nature, extensive research has been done on the possible
impacts of formaldehyde on human health. A significant number of studies have
been done to explore the effects of formaldehyde based products upon contact
4


with skin. These studies have shown that upon repeated exposure to solutions
containing formaldehyde, the subjects will tend to develop allergic dermatological
reactions leading to rashes on the exposed skin.3'5
Most studies have involved observing the effect of formaldehyde gas on the
respiratory system. Testing on humans has demonstrated that the gas
commonly causes irritation to the eyes, nose and throat of human subjects.
Based on these results, formaldehyde has been classified as an irritant. To
further explore the potential impact of formaldehyde on the human respiratory
system, medical studies using animals have been conducted.3,4,5 These animal
studies have shown that prolonged exposure can lead to carcinomas developing
in the nasal cavities of rats, which has prompted the chemical to be classified as
a likely carcinogen and that exposure levels in both ambient and indoor air are to
be regulated in workplace environments by the Occupational Safety and Hazard
Administration.8
In an effort to limit human exposure to formaldehyde, the ability to monitor and
quantify the concentration of formaldehyde in the air has lead to the
development of several methods for the analysis of air for Hazardous Air
Pollutants.
5


1.2 Measurement Techniques for Gaseous Formaldehyde
In order to fully understand the impact of formaldehyde on the environment and
to explore a correlation of environmental problems to formaldehyde
concentrations, many different approaches have been developed to measure the
compound in the atmosphere. Some designs have utilized derivitization of
formaldehyde followed by instrumental analysis in a laboratory. Such designs
tend to be well designed and easily deployed to the field. However, these
methods occasionally lack the flexibility to be used in studies that require more
rapid response or better time resolution. As a result of this requirement, more
versatile and specific methods have been developed for the near real-time
detection of atmospheric formaldehyde.
1.2.1 EPA Method TO-11A
The United States Environmental Protection Agency (EPA) has produced several
methods for the detection and quantification of hazardous air pollutants (HAPs),
including formaldehyde. Of particular interest, Compendium Method TO-11A
(TO-11 A) is used to detect reactive carbonyl compounds in air samples. This
method utilizes a cartridge consisting of a hollow tube containing a solid silica gel
substrate. The solid material present on the interior of the cartridge is coated
with 2,4-dinitrophenylhydrazine (DNPH) as a derivitizing agent and a known
volume of air is passed through the cartridge. As the air samples pass through
the sampling cartridge, the carbonyl species react with the coating and create a
6


family of hydrazone compounds that are easily detected by High Performance
Liquid Chromatography (HPLC).9 The reaction of interest is outlined in Figure 3.
Ri
R2
d
+
o
0 0 0 t}
II 'iN-,. II "d" -0 0 =N 0
1 .
J.I J NH !/ R1( '"'=-/ n.s
HN' ' o
nh2 r2
+ h2o
(10)
FIGURE 3: THE DERIVITIZATION OF HCHO WITH DNPH
The hydrazones produced through this reaction pathway are later separated
through chromatographic means using a combination of purified water and
purified acetonitrile mobile phases. The sampling cartridges are initially
extracted using a known volume of purified acetonitrile. The DNPH derivatives
are then quantified using reverse phase HPLC and detection with ultraviolet (UV)
absorption from a 360 nm UV source lamp. A list of compounds detectable by
TO-11A is presented in Table 1.
TABLE 1: PARTIAL LIST OF CARBONYLS QUANTIFIED WITH TO-11A
Formaldehyde Isovaleraldehyde Propionaldehyde
p-Tolualdehyde Acetaldehyde Valeraldehyde
Crotonaldehyde Hexanaldehyde o-Tolualdehyde
Butyraldehyde 2,5-Dimethyl benzaldehyde Methyl ethyl ketone
Acetone m-Tolualdehyde Benzaldehyde
7


Method TO-11A is capable of quantifying formaldehyde in the low parts per
billion by volume (ppbv) range over a time period of 1 24 hours. The most
common sampling period used with this technique is 24 hours, producing a 24
hour average concentration reading.9
1.2.2 Continuous Flow DNPH Methods
Additional studies have been done on the usefulness of DNPH derivitization of
carbonyls using continuous flow systems and diffusion scrubbers. The
techniques devised utilize the same separation and detection methods as TO-
11A, but work in a continuous manner producing results with better time
resolution. Zhou, et al. explored DNPH derivitization for several carbonyl species
utilizing a glass diffusion coil containing an acidified DNPH solution in 2009. The
solution was pumped through a collection chamber where the carbonyl
compounds would pass through the glass coil and react with DNPH to produce
the desired hydrazone derivatives. The resulting products were then analyzed by
an inline HPLC where the compounds were quantified with a detection limit of
0.036 ppbv. 10 A similar method was studied by Komazaki et al. in 1998 for the
analysis of formaldehyde in automotive exhaust. A porous
polytetrafluoroethylene (PTFE) tube containing the DHPH solution collected
formaldehyde from auto exhaust to be analyzed by an inline HPLC system. 11
8


1.2.3 Hantzsch Reaction Derivitization
In 1953, Nash published a work on the detection of formaldehyde through the
use of the Hantzsch reaction. 12 The Hantzch reaction is an organic synthesis
reaction that was discovered in the late nineteenth century that involves an
aldehyde species, diketone and a source of ammonia or a primary amine.13 In
the case of Nash's research, formaldehyde was combined with 2,4-pentanedione
and ammonium acetate. The net reaction is presented in Figure 4.
H,C_
_.CH,
HjC^O
NH,
O
O
H3C
o
' CH3
CH3
FIGURE 4: THE HANTZSCH REACTION
(11)
The product, 3,5-diacetyl-l,4-dihydrolutidine (DDL) was found to produce a
yellow color and was used for colorimetric detection of formaldehyde. Nash also
noted that elevated temperatures accelerated the formation of DDL from five
hours at room temperature to five minutes when heated near 60 C.12
Other groups have studied the usability of the Hantzsch synthesis of DDL for the
detection of atmospheric formaldehyde. In most cases, a diffusion scrubber is
used as the collection method of formaldehyde gas into an acidic solution.
Dasgupta et al. have studied the use of diffusion scrubbers using porous
9


membranes such as Celgard microporous polypropylene 14,15 and Nafion.16,
17 After collection by the diffusion scrubber, the formaldehyde solution is mixed
with the diketone and ammonium salts before passing through a heated reaction
chamber. Once the DDL generated from the Hantzsch reaction has left the
reaction chamber, it passes through a fluorometer utilizing soft UV radiation as
an excitation source. These experiments have lead to the development of
instruments capable of detecting formaldehyde at levels below 1 ppbv.
More recently, the use of the Fluoral-P reagent has been studied as a method of
detecting carbonyl compounds. Fluoral-P is an intermediate species of the
Hantzsch reaction resulting from the reaction of ammonia with 2,4-pentanedione
that produces DDL when combined with formaldehyde (Figure 5). Compton and
Purdy found that the Fluoral-P reagent is stable and able to be stored as an
analytical reagent. 18 In a continuation of this work, Pinheiro et al. explored the
use of the Fluoral-P reagent in sampling cartridges. In this study, Fluoral-P was
used to coat sampling cartridges and sampled in parallel to DHPH coated
cartridges. Both sample types were analyzed through HPLC, one used UV
detection and the other fluorescence detection. The results of the study show
that the Fluoral-P coated cartridges were able to detect formaldehyde
concentrations that agree with standard methods with a limit of detection of
0.002 mg/L. 19
10


0
H2C = 0 + 2 // \\
6 ch3
HjN'
H3C
o
"CH3
(12)
FIGURE 5: REACTION OF FORMALDEHYDE WITH FLUORAL-P
li


2. Experimental
A combination of in-house components and commercial solutions were used in
the design and implementation of the formaldehyde gas analyzer. Two
detections systems were developed utilizing a Nation based diffusion scrubbers
and fluorescence detectors. One was designed for field sampling and the second
system was used for the determination of the collection efficiency of the primary
diffusion scrubber.
New bottles of ammonium acetate (Fisher Chemical A639-500) were available in
the lab. A new bottle of 37% w/w formaldehyde (37.0 % certified) was ordered
from Acros Organics (AC41073-1000) and new Nation tubing was ordered from
Perma Pure, LLC (TT-030). Open bottles of analytical grade sulfuric acid, acetic
acid (Mallinckrodt 2504-500) and hydrochloric acid were stored in a laboratory
cabinet. An open bottle of 99+% pure 2,4-pentanedione was available, stored in
a lab freezer (Aldrich P775-4).
2.1 Field Formaldehyde Analyzer
The field analyzer consisted of three main systems: A pneumatic system, a
liquid handling system and software for instrument control and data acquisition.
2.1.1 Pneumatic System
The vacuum pump used in the field analyzer was acquired from Thomas
Industries, Inc. The flow of the vacuum pump was regulated by a Tylan FC-260
12


mass flow controller (MFC) capable of 0 3 Standard Liters per Minute (SLPM) of
O2 gas. The flow controller is adjustable by a variable voltage (0 5v) supplied
by a data acquisition interface connected to a Lenovo Thinkpad notebook
computer. The overall pneumatic system was configured to draw air through a 2
micron Teflon particulate filter and then through the back panel sampling inlet.
The sample gas was then passed through the diffusion scrubber, MFC and
vacuum pump before exiting the analyzer's exhaust port. A three-way
pneumatic valve was installed downstream of the sampling inlet that allows for
the production of blank sample by passing air through a DNPH coated sampling
cartridge for collecting sample blanks. The analyzer layout is shown in Figure 6.
All tubing used in the pneumatic system consisted of Va inch PTFE tubing with
polypropylene or brass fittings.
Exhaust Sample Inlet
FIGURE 6: PNEUMATIC LAYOUT OF FIELD ANALYZER
13


2.1.2 Liquid Handling System
The field analyzer utilized a liquid handling system consisting of a Cavro Scientific
Instruments four channel syringe pump (Cavro XL 3000). The pump was
equipped with 0.500 mL syringes and each syringe is equipped with a three port
valve. The syringes share a common drive motor controlled through a serial port
connected to a Lenovo Thinkpad notebook computer. A custom designed
heating chamber was developed in-house consisting of a 40 watt ceramic heating
element and regulated with a West 3100 temperature control unit equipped with
a K-Type thermocouple and a solid state relay. The heater is insulated with
fiberglass insulation and contains 1.0 mL volume of 23 AWG PTFE thin walled
tubing. A solenoid valve was placed in the system to allow for the injection of a
formaldehyde span solution for checking detector response at regular intervals.
The valves were controlled by the data acquisition interface board digital output
ports and solid state relays. The scrubbing solution and a solution containing
ammonium acetate, acetic acid and 2,4-pentanedione (Nash reagent) used in
this system were combined before being pumped into the reaction heater
through a three-way PEEK union. An Omni-fit 006BT bubble trap was installed
downstream of the reaction heater to remove bubbles from the solution before
detection. The diagram in Figure 7 illustrates the liquid handling system of the
field analyzer. Figure 8 outlines the construction of the reaction heater and
direction of sample flow.
14


FIGURE 7: SAMPLING PROCESS OF LIQUID HANDLING SYSTEM
FIGURE 8: CONSTRUCTION OF REACTION HEATER
15


2.1.3 Diffusion Scrubber
The diffusion scrubber for the field analyzer was designed to withstand the
pressure of the syringe pump. Nation was selected as the diffusion membrane
as it is selective for formaldehyde and prevents other carbonyl compounds from
passing into the scrubbing solution. 20 1/16 inch stainless steel tubing used for
HPLC systems was tapered at the ends and a groove was cut behind the tip,
producing a barbed end. The Nation tubing was soaked in pure methanol, as
the membrane increases in size by 88%. 21 The steel tubing was inserted
through the open ends of the expanded membrane tubing and allowed to dry.
The dried membrane tightened around the barbed tip, securing the membrane to
the tubing. The connection between the Nation and the steel tubing is
illustrated in Figure 9.
Steel Tubing
Nation soaked in methanol
Dry Nation
Steel Tubing
FIGURE 9: ATTACHMENT OF NAFION TO STEEL TUBING
16


To improve the collection efficiency of the field analyzer's diffusion scrubber, the
length of the scrubber was extended in comparison to other studies. 19 Two Va
inch polypropylene elbow unions, two Va to 1/8 inch reducing elbows and two 16
inch lengths of PTFE tubing were configured in a dual-channel configuration. A
small hole was drilled in each elbow to be concentric with one opening of the
fitting. Loctite All Plastics super glue was used to bond an autosampler vial
septum over each drilled hole to pneumatically seal the scrubber fittings. A 15
inch length of Nafion tubing was fed through each of the lengths of PTFE
tubing and the attached metal pierced the septa. The Nafion sections were
attached together by a section of metal tubing inserted through two septa
attached to opposite sections of PTFE tubing. A section of 1/8 inch PTFE tubing
was used to connect the sections of Va inch PTFE tubing. Clear polyvinylchloride
(PVC) tubing with an inner diameter of Va inches was used to support the
straight design of the diffusion scrubber. The structure of the diffusion scrubber
is outlined in Figures 10 and 11. The scrubbing solution was configured to run
anti-parallel to the flow of sample gas.
Steel
Tubing
^ PVCTubing
f Nafion
1 PTFE Tubing
FIGURE 10: ELBOW UNION LAYOUT
17


2.1.4 Gilson Model 121 Fluorometer
A Gilson Model 121 filter fluorometer was used to detect DDL produced from the
Hantzsch reaction. The fluorometer is configured with a 9 pL vertical square
flowcell and a nine-stage photomultiplier tube (PMT). The standard source lamp
was replaced with a MR16-UV24-15 lamp (Super Bright LEDs, Inc.) consisting of
24 light emitting diodes (LEDs) rated at a peak wavelength of 405 nm. A 1 mm
pinhole slit was installed on the excitation window of the flowcell chamber to
center the incident light on the flat surface of the flowcell. A cut-on filter was
installed on the emission window of the flowcell chamber to block incident light
from the PMT. The emission filter spectrum is illustrated in Figure 12. The
Gilson 121 has a selectable time constant, ranging from 0.5 s, 1.0 s and 2.0 s as
well as an output voltage range of 10 mV, 100 mV and 1.0 V. The front panel of
the detector has an auto-zero button and a selectable sensitivity range is of 1
0.001.
18


3.0 _
2.5 _
FIGURE 12: ABSORPTION SCAN OF EMISSION FILTER
2.1.5 Data Acquisition System
Custom data acquisition software was developed with the LabView 2010 Student
Edition software development environment. A program was developed with the
capability to sample multiple voltage readings at a defined frequency and
average several values to reduce signal noise. The voltage readings are made
from a Minilab 1008 USB data acquisition interface by Measurement Computing
and acquired on a Lenovo Thinkpad notebook computer. The Minilab 1008 is a
low cost 12-bit data acquisition board that also has the ability to send and read
digital data as well as produce an analog output voltage. Several digital ports on
the board were configured to trigger relays that power pneumatic and liquid
handling valves in other process systems. This system provides the flexibility to
modify the process system as adjustments are required. Fluorescence and flow
19


readings from the MFC are saved to a text file capable of being used in a
spreadsheet program and a new file is generated every 24 hours. A virtual front
panel is displayed on the computer screen that allows adjustments to the data
acquisition process and displays current voltage readings for flow and
fluorescence. Figure 13 presents the LabView program interface as displayed on
the computer screen. The program was also the primary method of
communication to the syringe pump through the serial port of the computer.
a HCHOv2-3.vi Front Panel *
File Edit View £roject Operate Tools Window Help
115pt Application Font '-~]|Sq-1

MFC Slope MFC Intercept
J0.8448 | 'J-0,1238 [
Fluorescence Fluorescence
Slope____ Intercept
l-rtn.

Student Edition I
DAQ
Blank
Span
Fluorescence
0
Flow Output (SLPM)
0
Errors
ITT trrlrr T
>-sni;AQ IlIZlI Stop DAQ I _i_ Exit Program |
1 i i 0 i i i i i
i:: ixu.HEm.zuic.nii.
r
"IUrDICC"
iij ..t
_ij-j-
:rrTn_r^_n'
FIGURE 13: LABVIEW INTERFACE FOR FIELD ANALYZER
20


2.2 Laboratory Formaldehyde Analyzer
The laboratory analyzer consisted of three main systems: A pneumatic system, a
liquid handling system, and software for data acquisition.
2.2.1 Pneumatic System
The pneumatic system of the laboratory formaldehyde analyzer consists of a
Nafion diffusion scrubber and several sections of Va inch PTFE tubing. The lab
apparatus was designed to be used in series with the field analyzer for the
determination of the collection efficiency of the dual-channel diffusion scrubber.
The laboratory diffusion scrubber is simpler than the field design (Figure 13). A
single length of Va inch PTFE tubing 15 inches in length was used as the
scrubber sleeve. 15 inches of Nafion was fed through the jacket and aligned in
a concentric configuration. Two Va inch polypropylene tee unions were used to
seal the ends of the scrubber jacket. Two Va inch tube caps were drilled and
Loctite All Plastics super glue was used to glue autosampler vial septa over the
holes. The Nafion tubing was attached to metal tubing in the same fashion as
the field scrubber (Figure 11) and fed through the septa on the tube caps. The
open ports on the tee unions were connected to a vacuum source as needed.
When in use, the scrubber was connected between the dual-channel diffusion
scrubber and the MFC of the field analyzer. All connections used in the lab
analyzer's pneumatic system were polypropylene fittings.
21


FIGURE 14: LABORATORY SCRUBBER CONFIGURATION
2.2.2 Liquid Handling System
Several components were used in the liquid handling system of the laboratory
formaldehyde analyzer. The pump used in the lab was a Gilson Minipulse2 multi-
head peristaltic pump. Two heads of the pump were used to pump the
scrubbing fluid and the Nash reagent. A three port polypropylene union was
installed downstream of the diffusion scrubber and before a Timberline
Instruments column heater (S/N 1092). The column heater can be refitted with
internal 1/16 inch PTFE of different lengths to change the internal volume. A
second bubble trap was not available for use in the second formaldehyde
analyzer.
FIGURE 15: LAB ANALYZER LIQUID HANDLING SYSTEM
22


2.2.3 Perkin Elmer LC 240
The fluorometer selected for the detection of DDL in the laboratory analyzer was
a Perkin Elmer LC 240. The LC 240 contains dual monochrometers for selecting
excitation and emission wavelengths independently for maximum sensitivity.
The 4.0 pL flowcell is a 1.0 mm square fused-silica flowcell oriented in a
horizontal arrangement. The output range of the detector is 0 1 V and the
time constant is 1.8 s. A Xenon discharge lamp is used as the light source and
produces light in the UV and visible spectrum. An excitation wavelength of 419
nm and an emission wavelength of 516 nm were experimentally determined for
the detection of DDL.
2.2.4 Data Acquisition System
LabView 2010 Student Edition was used to develop a second data acquisition
software solution. The laboratory software was designed to take voltage
readings from a second Minilab 1008 USB data acquisition board. The program
collected data from two separate voltage sources at a specified frequency and
averaged each signal into one minute values. No control programming was
needed for the second formaldehyde analyzer. Signals from both fluorescence
detectors were connected to the second Minilab 1008 board for direct
comparison of DDL detection signals.
23


2.3 Experimental Procedure
2.3.1 Reagents
For the continuous flow analyzers, Nash reagent and dilute sulfuric acid were
made. 500 ml. of 0.65 M Nash reagent was made by mixing 24.5 g of
ammonium acetate, 40 |jL glacial acetic acid and 438 |jL of 2,4-pentanedione
with Millipore deionized water in a 500 ml. volumetric flask. 500 ml. of 1.0 x 10'5
M sulfuric acid was made by mixing 100 uL 0.1 M H2SO4 with Millipore deionized
water in a 500 ml. volumetric flask. Each solution was transferred to a separate
500 ml. amber glass bottle capped and labeled. The reagents were allowed to
equilibrate for a day before use at room temperature.
2.3.2 Calibration of Formaldehyde Analyzers
The Gilson Model 121 fluorometer was calibrated with the use of formaldehyde
standard solutions. The solutions were made by dissolving 0.5 ml. of 37% w/w
formaldehyde (37.0 % certified) to 1000 ml. in a volumetric flask. 3.000 ml. of
the diluted solution was transferred to a 250 ml. volumetric flask with a
volumetric pipette. The standard solutions were made by transferring 1 ml. of
1.00 x 10'3 M H2SO4 to a series of 100 ml. volumetric flasks with a volumetric
pipette. Aliquots of the second stock formaldehyde solution in the volumes of 5,
10, 15, 20 and 25 ml. were transferred to the 100 ml. volumetric flasks with
volumetric pipettes. Table 2 summarizes the calculated concentrations of the
formaldehyde standards.
24


TABLE 2: STANDARD HCHO SOLUTIONS
Volume of HCHO Stock Solution (mL) Calculated Standard Concentrations (M) Calculated ppbv Equivalent (ppbv)
5.00 3.84E-06 11.1
10.00 7.69E-06 22.3
15.00 1.15E-05 33.4
20.00 1.54E-05 44.6
25.00 1.92E-05 55.7
0.00 0.00 0.00
25


3. Results and Discussion
3.1 Reaction Chamber Optimization
To maximize the detection capability of the formaldehyde analyzer, the Nash
reagent was mixed with a formaldehyde solution and heated at 60 C to
accelerate the reaction and prevent bubble formation. The solution was held in
the Timberline heater and heated for a measured time. The reaction product
was then flowed into the LC-240 fluorometer and the fluorescence was measured
for each sample. The fluorescence response was measured for heating time
intervals from 1 minute to 45 minutes. The data are summarized in Table 3
below.
TABLE 3: HANTZSCH REACTION COMPLETION DATA
Heating Time (min) Response (V) Heating Time (min) Response (V)
45 0.5824 8 0.5707
1 0.2317 9 0.5697
2 0.3744 10 0.5731
3 0.4774 11 0.5766
4 0.5253 12 0.5795
5 0.5507 13 0.5761
6 0.5648 14 0.5770
7 0.5653 15 0.5751
Plotting the data in Table 3 effectively illustrates the process of the Hantzsch
reaction. The reaction appears to slow significantly after 5 minutes of heating at
60 C (Figure 16). Using the reaction data, the flow of the liquid handling
26


system was optimized so the reagents are heated for at least 5 minutes in the
reaction chamber of the field analyzer.
Optimization of Reaction Heater
Reaction Time (min)
FIGURE 16: OPTIMIZATION OF REACTION HEATER
3.2 Calibration Results
Two components of the field analyzer required calibration before use. The Gilson
Model 121 fluorescence detector and the Tylan FC-260 mass flow controller were
calibrated and the data are summarized below.
3.2.1 Gilson Model 121
The Gilson fluorometer was configured with a 2.0 s time constant, 100 mV
output voltage and a sensitivity setting of 0.1. The Nash reagent and the HCHO
standards were dispensed into the three-way PEEK union, heated with the
27


reaction heater of the field analyzer through the Gilson fluorometer. Each
standard was dispensed for a period of twenty minutes followed by twenty
minutes of fresh 1.0 x 10'5 M H2SO4. The one minute data are summarized in
Figure 17.
Gilson Calibration Data
FIGURE 17: GILSON MODEL 121 CALIBRATION DATA
A calibration curve and regression equation was generated by averaging the
values of each sample peak and performing a linear regression operation of the
data points. A calculated concentration for the ppbv equivalent values was
calculated from the standard solution molarities using equations 13 and 14. The
resulting calibration curve is summarized in Table 4 and Figure 18.
28


Calculation for ppbv Equivalent (1.00 L/min gas flow, 293 K and 84 kPa):
mg , /
ppbv = x (106) x I
Li \
Formula Weight
\ RT
)X~
(13)
3.84 x 106 mol HCHO\ /0.0001 L liquid\ /30.04 g HCHO\
min ) V mol HCHO )
-n
X
/1000 mg\ /1.00 L gas\ 1.15x10 5 mg HCHO
V 9
]\ /l.UU L gas\
) V min )
Li cls
(14)
'1.15 x 10~smg HCHO\
) x 106 x LmThcho) x (
^8.3144 L kPa\
L gas J ~ V30.04 g HCHO) V mol K
(293 K) x ^ ^ = 11-1 ppbv equivalent
(15)
TABLE 4: AVERAGE RESPONSE VOLTAGE AND CONCENTRATION
Average Signal (v) Concentration (ppbv)
0.074315 0.00
0.102111 11.1
0.134859 22.3
0.166064 33.4
0.196655 44.6
0.230576 55.7
29


Calibration Curve
FIGURE 18: GILSON MODEL 121 CALIBRATION CURVE
As illustrated in the figure above, a linear relationship was observed between the
concentration of HCHO and relative voltage output of the fluorometer. The slope
of the line was 355.66 ppbv/V and the intercept was -25.758 V. The correlation
coefficient indicates a strong relationship between the two variables with a R2 of
0.9994. The slope and intercept values were entered into the LabView Virtual
Instrument interface.
3.2.2 Limit of Detection
The limit of detection (LOD) for the Gilson Model 121 fluorometer was calculated
by calculating the mean and standard deviation of the zero formaldehyde
30


standard solution voltage in Microsoft Excel. The zero standard statistics are
summarized in Table 5.
TABLE 5: STATISTICS FOR ZERO STANDARD SOLUTION
Signal Mean (x) Signal Standard Deviation (a) Limit of Detection (LOD)
0.0743 V 0.000508 V 0.0758 V
Equations 15 through 17 were used to calculate the limit of detection by adding
three times the standard deviation of the zero standard to the mean. Using the
regression equation from the calibration curve, the voltage value was translated
into a ppbv equivalent value.
LOD =x + 3a (15)
LOD = 0.0743 V + 3(0.000508 V) = 0.0758 V (16)
LOD = 355.66 ppbv/V *(0.0758 V) 25.758 ppbv = 1.22 ppbv (17)
The voltage for the LOD was determined to be 75.8 mV, corresponding to a
concentration of 1.22 ppbv, assuming a standard gas flow rate of 1.00 SLPM. To
further characterize the calibration data, the concentration of the zero standard
was calculated from the average zero standard voltage. The concentration was
found to be 0.673 ppbv. The regression data was calculated from a single set of
calibration data.
31


The signal voltage response for a concentration change of 1 ppbv was calculated
by taking the difference of the average signal for each calibration peak and the
zero standard. The voltage difference was divided by the concentration of the
respective calibration peak, and the dividends were averaged. An example
calculation is presented in equation 18. A summary of the data are presented in
Table 6.
mV H 0.74315 V 0.102111 V 0.74315 V
r~ = ;-----7 ;: x 1000 =----------- -----------X 1000
ppbv ppbv of solution 11.1
= 2.5 mV/ppbv
(18)
TABLE 6: RESPONSE PER PPBV
Concentration (ppbv) Voltage / Concentration (mV/ppbv)
11.1 2.49
22.3 2.72
33.4 2.74
44.6 2.74
55.7 2.80
Average 2.70
Standard Deviation 0.12
The average response per ppbv was calculated to be 2.70 mV 0.12 mV.
3.2.3 Tylan FC-260 Mass Flow Controller
The MFC installed in the field analyzer was calibrated using a BIOS flow
calibrator and a volt meter connected to the MFC output voltage terminals.
32
The


BIOS flow calibrator was used to average ten flow measurements for each run.
The data are summarized in Table 7 and Figure 19.
TABLE 7: MASS FLOW CONTROLLER CALIBRATION DATA
MFC Flow Output Voltage (Volts) Flow Calibrator Output (SCCPM)
0.151367 32.37
0.336914 175.0
0.664062 437.5
0.996094 709.8
1.32813 982.8
1.66016 1256
1.99219 1540
2.32422 1822
2.66113 2111
2.98340 2408
3.28125 2664
3.28613 2678
33


Tylan FC-260 Calibration Curve
eu
-J
Tfl
5
O.
5
O
o
-fi
U
E
2.5
1.5
2 1
0.5
i i i i i i i i i i i
:y = 0.8448x- 0.1238
: R2 = 0.9996


e:
=*^=
0.5
1.5 2
MFC Output Voltage (V)
2.5
3.5
FIGURE 19: MASS FLOW CONTROLLER CALIBRATION CURVE
A linear regression operation was performed on the calibration data. The
regression equation presented in Figure 18 indicates a strong linear correlation
with a R2 value of 0.9994. The slope of the equation is 0.8448 SLPM per volt,
and the intercept is -0.1238 volts. The slope an intercept values were entered
into the calibration fields of the LabView Virtual Instrument interface.
3.3 Collection Efficiency
The collection efficiency of the primary diffusion scrubber was determined by
measuring breakthrough concentrations of formaldehyde with the laboratory gas
34


analyzer. The laboratory instrument was connected downstream of the field
analyzer sample gas flow. A solenoid valve was installed to redirect the sample
gas flow to the laboratory diffusion scrubber directly from a permeation device
every two hours resulting in alternating high reference signal and sample
breakthrough signals. The day before the collection efficiency experiment was
started, ten hours of a blank sample was sampled by the laboratory analyzer by
installing a DNPH sampling cartridge on the instrument gas sample inlet. The
output from the Perkin Elmer LC-240 is summarized in Figures 20 and 21.
Summary of Collection Efficiency
FIGURE 20: SUMMARY OF COLLECTION EFFICIENCY DATA
35


Collection Efficiency Solution Blank
FIGURE 21: BLANK SOLUTION DATA (TEN HOURS)
The sloping peaks in Figure 20 indicate a pneumatic leak was likely present in
the system during the collection efficiency measurements. All pneumatic fittings
were tightened at the conclusion of the experiment. The collection efficiency
was calculated by comparing the relative response of the high and low signals to
the zero signal collected on the previous day. The standard deviation for the
high reference and breakthrough sample was calculated from all points of the
signal peaks and valleys respectively. The standard deviation of the zero
reference was calculated from the ten hours of data prepared the day before the
collection efficiency experiment was performed. The results for the sample
statistics are summarized in Table 8.
36


TABLE 8: STATISTICS FOR COLLECTION EFFICIENCY
Signal Mean (V) Standard Deviation (V)
High Reference (H) 0.055832 0.002643
Breakthrough Sample (L) 0.021369 0.001062
Zero Reference (B) 0.021385 0.000862
The statistics presented in Table 5 were used to calculate the collection efficiency
of the diffusion scrubber installed in the field analyzer using Equations 19 22.
The collection efficiency was calculated to be 100.05 % 3.97 %.
Calculation for Collection Efficiency:
(L-B)
(H-B)
x 100
(0.0021369 -0.021385)
l1 (0.055832 0.021385)
x 100 = 100.05 percent
(19)
(20)
Calculation for Propagation of Error:
Error =
1
(21)
((1.30965)2(0.002643)2 + (-2902.99)2(0.001062)2
+ (2904.30)2(0.000862)2)z = 3.97 percent error
(22)
37


3.4 Sampling Results
The field analyzer was configured to sample air in both indoor and outdoor
environments. Several days of data were collected in the research laboratory
and in an outdoor research shelter on the Auraria Campus in Denver, Colorado.
3.4.1 Indoor Sampling
The analyzer was set up in the research laboratory with the inlet sampling tubing
open to the ambient laboratory air. The software was programmed to begin
sample collection at 12:00 am on December 21, 2011. A 44.6 ppbv equivalent
span solution was installed for injection through the diffusion scrubber during the
first hour of data acquisition. A full 24 hours of samples were collected on
December 21. The results of the sampling are presented in Figure 22.
12/21/2011 Lab Air Data
FIGURE 22: INDOOR AIR SAMPLING RESULTS
38


The data presented in Figure 22 shows the span solution peaking near 50 ppbv
when the reference value is 44.6 ppbv. Each data peak present in the data
summary represents one hour of sampling, with ten minutes of a zero sample
per hour. The average values of the peaks were calculated and summarized in
Figure 23 and Table 9.
>
-fi
S-5
c
"'S3
-
c
CJ
c
o
U
O
H
U
H
12/21/2011 Hourly Data


...........
* ..................
10 15 20 25
Hour
FIGURE 23: HOURLY AVERAGE FOR INDOOR DATA
TABLE 9: HOURLY INDOOR HCHO CONCENTRATIONS
Hour HCHO (ppbv) Hour HCHO (ppbv) Hour HCHO (ppbv)
0 42.9 Span 8 2.80 16 2.27
1 4.87 9 2.13 17 2.11
2 3.38 10 1.80 18 1.93
3 2.60 11 2.00 19 2.31
4 2.49 12 2.24 20 2.43
5 2.35 13 2.70 21 2.33
6 2.51 14 2.17 22 2.49
7 2.88 15 2.19 23 2.49
39


The indoor HCHO concentration appears to be relatively higher early in the
morning at the 1 am hour and another peak appears during the 7 am 9 am
hours. The concentrations range from 1.80 to 4.87 ppbv inside the laboratory on
December 21, 2011. Ten hours of the 24 hour sampling period have blank
concentrations above the 1.22 ppbv LOD for the instrument. Further evaluation
of the LOD was conducted by taking the standard deviation of the blank
concentration data collected during the indoor sampling period. A total of 75
data points were used as a single data set to calculate the LOD from sampling
data. The 75 points consisted of three minutes from each hourly blank sample in
Figure 22. The standard deviation was calculated to be 0.715 ppbv using the
Microsoft Excel STDEV function. The LOD was calculated using equation 23.
LOD = 3a = 3(0.715) = 2.15 ppbv (23)
40


Gas Flow Rate (12/21/2011)
FIGURE 24: GAS FLOW RATE (12/21/2011)
The gas flow of the analyzer appeared to have drops in the signal near the time
the valve switched to sample zero air and spikes when switching back to sample
air (Figure 24). The fittings on the scrubber and pneumatic valve were tightened
during the 2 pm hour. The signal spikes were significantly reduced in
magnitude. The average flow rate was 1.05 SLPM 0.02 SLPM.
3.4.2 Outdoor Sampling
On December 23, 2011 the field analyzer was relocated to a sheltered air
sampling station located near the intersection of 7th avenue and Lawrence Way
in Denver, Colorado. The roof of the structure was equipped with a glass
sampling inlet manifold eight feet high with a bent end to prevent precipitation
from entering the inlet. The analyzer software was configured to begin sampling
41


from the inlet and collecting data on December 24 at 12 am. Initial results
presented significant interference from bubbles in the fluorometer flow cell as
HCHO concentration readings below 0 ppbv. The 12/24/2011 results are
presented in Figure 25.
Outdoor Sampling Data 12/24/2011
400 600 800 1000 1200 1400 1600
Sample Number (every minute)
FIGURE 25: OUTDOOR SAMPLING RESULTS (12/24/2011)
The bubble trap was disassembled and the internal membrane of the device was
replaced. Further sampling was conducted beginning at 12 am on December 29,
2011. The sensitivity setting of the Gilson fluorometer was checked on
December 31, 2011 and found to be set to a sensitivity of 0.2. The sensitivity
was reset to 0.1. Due to the sensitivity having been set incorrectly, the outdoor
data collected in 2011 was not usable for the quantification of HCHO.
42


On January 1, 2012, the analyzer response was checked. The data collected
during January 1 did not appear to have significant bubble interference, but the
voltage output appeared to have blank signals higher in concentration than
outdoor air. At 11 am on January 1, the carbonyl scrubber was replaced and the
span solution was replaced with a 33.4 ppbv standard. The collected on that day
indicates that the carbonyl scrubber was repaired. On January 2, a fresh batch
of Nash reagent was installed in the analyzer at 12 pm. The data collected on
January 1 and 2 is presented in Figures 26 through 31 and Tables 10 and 11.
Outdoor Sampling Data 1/1/2012
FIGURE 26: OUTDOOR SAMPLING RESULTS (1/1/2012)
43


6
Hourly Averages 1/1/2012
>
-s 5
c
#o
c
c
o
U
O
B
U
a




10
15
20
25
Hour
FIGURE 27: HOURLY HCHO CONCENTRATIONS (1/1/2012)
TABLE 10: OUTDOOR HOURLY DATA (1/1/2012)
Hour HCHO (ppbv) Hour HCHO (ppbv) Hour HCHO (ppbv)
0 2.19 8 3.27 16 1.46
1 1.87 9 3.62 17 2.25
2 2.09 10 3.48 18 2.88
3 2.39 11 2.97 19 2.64
4 2.05 12 1.86 20 3.22
5 2.05 13 1.67 21 3.46
6 2.13 14 1.73 22 2.54
7 2.34 15 1.53 23 37.2 (Span)
44


Gas Flow Rate (1/1/2012)
Sample Number (every minute)
FIGURE 28: GAS FLOW RATE (1/1/2012)
Outdoor Sampling Data 1/2/2012
FIGURE 29: OUTDOOR SAMPLING RESULTS (1/2/2012)
45


Hourly Averages 1/2/2012
I 5 ^
| 4
*c3
fl 3
C
O
U 2
O
H
U 1
M
*
A


0
10
15
20
25
Hour
FIGURE 30: HOURLY HCHO CONCENTRATIONS (1/2/2012)
TABLE 11: OUTDOOR HOURLY DATA (1/2/2012)
Hour HCHO (ppbv) Hour HCHO (ppbv) Hour HCHO (ppbv)
0 3.77 8 4.33 16 1.35
1 5.08 9 3.89 17 2.46
2 4.74 10 2.71 18 2.39
3 3.70 11 2.36 19 2.10
4 3.38 12 1.89 20 1.75
5 3.61 13 1.97 21 1.53
6 3.46 14 2.02 22 1.77
7 4.97 15 1.91 23 31.5 (Span)
46


Gas Flow Rate (1/2/2012)
Sample Number (every minute)
FIGURE 31: GAS FLOW RATE (1/2/2012)
During the course of outdoor sampling, the span injection process shifted in time
from the originally scheduled time due to timing issues with the LabView
program. On the first day of outdoor operation, the second span process
executed an hour early resulting in a span injection at 11 pm instead of 12 am
the following day. After the initial time shift, the span consistently executed at
11 pm on later days. Sampling ended at 11 am on January 3, 2012 due to a
short supply of H2SO4 solution.
On January 1, the HCHO concentrations peaked during the 7 am 9 am hours
and 6 pm 9 pm hours. The HCHO concentrations ranged from 1.46 ppbv to
3.46 ppbv. On January 2, the HCHO concentrations peaked during the 12 am -
47


3 am hours and 7 am 9 am hours. The HCHO concentrations ranged from 1.34
ppbv to 5.08 ppbv. Bubble interference was noted on this day.
On both days of sampling, drops in the gas flow rate were noted at the times
that the gas valve switched to sample a blank. Each of these drops lasted for
only one minute. The average gas flow of the analyzer was calculated to be 1.05
SLPM 0.01 SLPM on each day.
Fresh 1 liter batches of both the Nash reagent and the H2SO4 solution were
installed in the analyzer. A new Omni-fit 006BT bubble trap was installed with a
new membrane upstream of the fluorometer. Formaldehyde concentrations for
January 7 through January 10 were collected and summarized in Figure 32. The
hourly data values were based on absolute response of the fluorometer.
48


Raw Hourly Data Summary
FIGURE 32: HCHO CONCENTRATIONS (1/7/2012 1/10/2012)
Significant drift is apparent in the data spanning the four days of sampling. The
blank response for each day appears to be higher in the early hours of the
morning (7 am 10 am) and then significantly lower in the late afternoon hours
(5 pm 8 pm). The overall response range for the blank samples was -2.27
ppbv on January 10 to 1.69 ppbv on January 7. The daily blank response ranges
are presented in Table 12.
TABLE 12: OUTDOOR BLANK SAMPLE RESPONSE
Date High Blank (ppbv) Hour Low Blank (ppbv) Hour
7-Jan 1.69 9:00 -0.41 20:00
8-Jan 1.21 9:00 -0.78 17:00
9-Jan 0.73 10:00 -1.55 20:00
10-Jan 0.28 7:00 -2.27 18:00
4-day Range 1.69 9:00 -2.27 18:00
49


Negative HCHO concentrations were calculated from the fluorometer for several
hours during the sampling period. The span solution signals were also evaluated
and found to decrease over the course of the four days, indicating a possible
problem with the detector sensitivity. The decreasing span values are
summarized in Figure 33.
^ 35
0.30
a 25
o
HCHO Span Concentrations
x


0.5 1 1.5 2 2.5 3 3.5 4 4.5
Day of Sampling Period
10-Jan 9-Jan A 8-Jan X7-Jan
FIGURE 33: DECREASING TREND OF SPAN CONCENTRATION
The absolute response of the span solution dropped from 29.2 ppbv to 16.7 ppbv
from January 7 to January 10. Baseline correction was applied to the span data
by calculating the difference of the peak and the zero sample before and after
the peak. The corrected span concentrations indicate a drift from 28.3 ppbv to
17.9 ppbv from January 7 to January 10. This decrease was not expected in the
50


sampling results. The sample gas flow during the sampling period was 1.05
SLPM 0.02 SLPM. The hourly HCHO concentrations are summarized in Table
13.
TABLE 13: HOURLY HCHO CONCENTRATION (PPBV) SUMMARY
Hour 7-Jan 8-Jan 9-Jan 10-Jan
0 29.20 (Span) 2.00 1.69 0.19
1 3.18 2.41 1.44 1.09
2 2.30 2.83 1.50 1.70
3 2.71 1.99 1.36 2.32
4 1.77 1.78 1.58 2.32
5 1.90 2.07 2.15 2.36
6 2.02 2.33 3.09 4.76
7 2.82 2.09 3.62 1.95
8 2.96 2.72 3.02 1.95
9 2.49 2.94 2.56 1.60
10 2.08 1.89 2.80 0.85
11 1.61 1.70 2.11 0.35
12 1.69 1.48 1.63 -0.02
13 1.30 1.20 0.86 -0.50
14 0.88 0.75 0.21 -0.60
15 0.91 0.37 -0.18 -0.99
16 0.72 0.34 -0.23 -0.82
17 0.64 0.68 0.51 -0.34
18 0.88 2.44 0.98 0.85
19 0.66 2.70 0.30 0.68
20 0.79 2.24 0.46 -0.16
21 0.73 1.28 0.19 0.01
22 0.86 27.00 (Span) 22.14 (Span) 16.70 (Span)
23 1.31 1.14 0.42 0.53
51


4. Conclusion
A continuous flow formaldehyde analyzer was successfully developed, capable of
quantifying concentrations above 1.22 parts per billion by volume in ambient air.
The Hantzsch reaction was utilized to convert the formaldehyde samples into a
product with fluorescent properties when excited with 405 nm light. A diffusion
scrubber utilizing Nation as a permeable membrane was constructed with a
collection efficiency calculated to be 100.05 % 3.97 %. The Gilson Model 121
fluorometer was calibrated for a 0 55 ppbv range with a signal response of
2.70 mV per ppbv 0.12 mV.
The limit of detection of the analyzer was expected to be below 1 ppbv, but was
calculated to be 1.22 ppbv from the calibration data. The limit of detection was
also calculated from sampling data collected during the indoor sampling
experiment. The detection limit was found to be 2.15 ppbv, a value 76 % higher
than derived from the calibration blank. The increased standard deviation of the
blanks signal during sampling could likely be due to a pneumatic leak near the
carbonyl scrubber that led to a high blank concentration.
One of the primary sources of signal noise in the analyzer is the result of
changing pressure in the liquid handling system. The syringe pump used for
dispensing the reagents through the analyzer uses one minute long injections.
The pressure in the system drops when the syringes stop to refill with fresh
52


reagents, resulting in a decrease in fluorescence in the detector. To reduce the
noise from pressure changes, using a pair of dual syringe pumps configured to
provide a continuous pressure would significantly reduce the noise and lower the
limit of detection. Alternatively, a pulse dampener could be used to equalize the
pressure of the system and an increased electronic sampling rate of the
fluorescence signal would lower signal noise to improve the signal to noise ratio.
The analyzer was used to analyze outdoor ambient air for several days. Several
issues were noted during the sampling process. Leaks in the pneumatic system
of the gas analyzer have proven to be problematic for the accurate collection of a
zero gas sample. A leak near the carbonyl scrubber allowed for indoor air to
enter the sampling system and produce erroneous values. Many of the blank
samples taken during sampling are not likely to truly be zero due to the leaks in
the system. As reported, several blank samples produced signals above the limit
of detection indicating that a leak is highly probable. To improve the system, the
carbonyl scrubber needs to be better sealed in the pneumatic system with the
use of rubber gaskets to prevent air from bypassing the cartridge and should be
replaced periodically. An alternative method that may be useful for reading a
zero concentration of formaldehyde is to stop gas flow through the diffusion
scrubber altogether. Using this method, the background of clean H2SO4 and
Nash reagents may be obtained without introducing formaldehyde contamination
to the system.
53


During several days of sampling outdoor ambient air, an apparent drift in the
signal baseline was noted, ranging from -2.27 ppbv to 1.69 ppbv. The Gilson
model 121 used for detecting DDL was designed as a detector for HPLC in a
controlled laboratory environment. The drift noted may very likely be due to a
variation in outdoor temperature and the photomultiplier tube in the detector,
resulting in lower sensitivity during colder temperatures and higher sensitivity
during warmer temperatures. Due to these changes in sensitivity, the calculated
concentrations of HCHO are likely not accurate, even when using baseline
subtraction. By controlling the temperature of the photomultiplier tube and the
outdoor air monitoring shelter, or by using a photodiode detector, these drifts
could be eliminated.
The use of a span solution allows the user of the analyzer to check the response
of the instrument on a daily basis. However, the use of a liquid standard is
inconvenient and the use of 37% formaldehyde may prove to be a difficult
method of producing a standard solution. Formaldehyde can volatilize and
escape the solution leading to a gradual decrease in concentration while
concentrated solutions may lead to polymerization of the formaldehyde
molecules. An alternative method that may be used is the catalytic conversion of
methanol into formaldehyde. Some studies have been conducted in the
conversion of methane and methanol to formaldehyde with the use of a
54


molybdenum catalyst.22, 23 Using the catalytic conversion method would make
the purchase of gas mixtures containing methane and easy alternative for the
calibration of the gas analyzer.
55


References
1. Walker, F. Early history of acetaldehyde and formaldehyde, A chapter in the
history of orcanic chemistry. J. Chem. Educ. 1933, 10 (9), 546.
2. Salthammer, T. M. S.; Marutzky, R. Formladehyde in the Indoor
Environment. Chem. Rev. 2010, 110 {4), 2536.
3. Gibson, J. E., Ed. Formaldehyde Toxicity; hemisphere Publishing
Corporation: Bristol, 1983.
4. Committe on Toxicology. Formaldehyde An Assessment of Its Health
Effects; Consumer Product Safety Comission: Washingtion, D.C., 1980.
5. U.S. Department of Health and Fluman Services. Toxicological Profile for
Formaldehyde July 1999; Agency for Toxic Substances and Disease Registry:
Atlanta, 1999.
6. Seco, R.; Penuelas, J.; Filella, I. Short-chain oxygenated VOCs: Emission and
uptake by plants and atmospheric sources, sinks and concentrations. Atmos.
Env. 2007, 41 (12), 2477.
7. Atkinston, R. Atmospheric chemistry of VOCs and NOx. Atmos. Env. 2000,
34 (12-14), 2063-2101.
8. United States Department of Labor, Formaldehyde. 1910.1048.
http://www.osha.gov/pls/oshaweb/owadisp. show_document?p_table=STAN
DARDS8ipJd=10075 (accessed January 7, 2012).
56


9. Center for Environmental Research Information, 0. R. D. Compendium
Method TO-11A, Determination of Formaldehyde in Ambient Air Using
Adsorbent Cartridge Followed by High Performance Liquid Chromatography
(HPLC) [Active Sampling Methodology]; Environmental Protection Agency:
Cincinatti, 1999.
10. Zhou, X.; Huang, G.; Civerolo, K.; Schwab, J. Measurement of Atmospheric
Hydroxyacetone, Glycolaldehyde, and Formaldehyde. Environ. Sci. Technoi.
2009, 43 (8), 2753-2759.
11. Komazaki, Y.; Narita, Y.; Tanaka, S. Development of an automated
measurement system using a diffusion scrubber and high-performance liquid
chromatography for the monitoring of formaldehyde and acetaldehyde in
automotive exhaoust gas. Analyst 1998,123 (11), 2343-2349.
12. Nash, T. The Colorimetric Estimation of Formaldehyde by Means of the
Hantzsch Reaction. Biochem. J. 1953, 421 (55), 416.
13. Hantzsch synthesis. http://www.organic-reaction.com/organic-
reaction/multi-component-reactions/hantzsch-synthesis/ (accessed October
15, 2011).
14. Dasgupta, P. K.; Dong, S.; Hwang, H. Diffusion Scrubber-Based Field
Measurements of Atmospheric Formaldehyde and Hydrogen Peroxide.
Aerosol Sci. and Tech. 1990,12(1), 98-104.
57


15. Dasgupta, P. K.; Dong, S.; Hwang, H.; Yang, H.; Genfa, Z. Continuous
Liquid-Phase Fluoromentry Coupled to a Diffusion Scrubber for the Real-
Time Determination of Atmospheric Formaldehyde, Hydrogen Peroxide and
Sulfur Dioxide. Atmos. Env. 1987, 22(5), 949-963.
16. Fan, Q.; Dasgupta, P. K. Continuous Automated Determination of
Atmospheric Formaldehyde at the Parts Per Trillion Level. Anal. Chem.
1994, 66 {4), 551-556.
17. Li, J.; Dasgupta, P. K.; Genfa, Z.; Hutterli, M. A. Measurement of
Atmospheric Formaldehyde With a Diffusion Scrubberand Light-Emitting
Diode-Liquid-Core Waveguide Based Fluorometry. Field Anal. Chem. and
Tech. 2001, 5(1-2), 2-12.
18. Compton, B. J.; Purdy, W. C. Fluoral-P, A Member of a Selective Family of
Reagents for Aldehydes. Anatytica Chi mica Acta 1980, 119(2), 349-357.
19. Pinheiro, H. L. C.; de Andrade, M. V.; de Pereira, P. A.; de Andrade, J. B.
Spectrofluorimentric determinatino of formaldehyde in air after collection
onto silica cartridges coated with Fluoral P. Microchem. J. 2004, 75(1), 45-
20.
58


20. Formaldehyde
Permeation
Through
Nation.
http://www.permapure.com/tech-notes/key-concepts/formaldehyde-
permeation-through-nafion/ (accessed October 16, 2011).
21. FAQs. http://www.permapure.com/support/faqs/ (accessed October 16,
2011).
22. Kennedy, M. Selective oxidation of methane to formaldehyde: comparison of
the role of promoters in hydrocarbon rich and lean conditions. Catalysis
Today 1992, 13 {2-3), 447-454.
23. Solomon, S. J.; Custer, T.; Schade, G.; Soares Dias, A. P.; Burrows, J.
Atmospheric methanol measurement sing selective catalytic methonol to
formaldehyde conversion. Atmospheric Chemistry and Physics Discussions
2005, 5, 3533-3559.
59


Full Text

PAGE 1

THE DESIGN AND CONTROL OF AN ATMOSPHERIC FORMALDEHYDE ANALYZER b y Marc Christopher Mayzes A.G.S., Arapahoe Community College, 2004 B .S., University of Colorado Denver, 2008 A thesis submitted to the Faculty o f the Graduate School of the University of Colorado Denver i n partial fulfillment o f the requirements for the degree of Master of Science Chemistry 2012

PAGE 2

This thesis for the Master of Science degree by Marc Christopher Mayzes has been approved for the Master of Science Chemistry by Larry G. Anderson Advisor John A. Lanning Hai Lin Date : January 31, 2012

PAGE 3

Mayzes, Marc Christopher (M.S., Chemistry) The Design and Control of an Atmospheric Formaldehyde Analyzer Thesis directed by Professor Larry G. Anderson ABSTRACT A fluorometry based gas analyzer was developed for the detection of atmospheric formaldehyde for field sampling. Formaldehyde was collected using a diffusion scrubber containing a permeable Nafion tube inside inch Teflon tubing with dilute sulfuric acid as a scrub bing solution. The formaldehyde was mixed with 0.65 M ammonium acetate and 2,4 pentanedione to produce the fluorescent derivative 3,5 diacetyl 1,4 dihydrolutidine (DDL). When excited by 405 nm light, the DDL was quantified by fluorescence detection with peak emission near 510 nm. A custom softwar e solution was developed for instrument control and measurement of fluorescence signals and improvement of signal to noise ratios. The combination of hardware and software components produced and instrument capa ble of detecting formaldehyde concentrations above 1 .2 2 ppbv in one minute intervals. The form and content of this abstract are approved I recommend its publication. Approved: Larry G. Anderson

PAGE 4

iv TABLE OF CONTENTS Tables .......................................................................................................... vii Figures ......................................................................................................... viii Chapter 1. Introduction ............................................................................................... 1 1.1 History of Formaldehyde ........................................................................ 1 1.1.1 Sources of Formaldehyde .................................................................... 1 1. 1.2 Health Concerns of Formaldehyde ....................................................... 4 1.2 Measurement Techniques for Gaseous Formaldehyde .............................. 6 1.2.1 EPA Method TO 11A ........................................................................... 6 1.2.2 Continuous Flow DNPH Methods .......................................................... 8 1.2.3 Hantzsch Reaction Derivitization .......................................................... 9 2. Experimental ........................................................................................... 12 2.1 Field Formaldehyde Analyzer ................................................................ 12 2.1.1 Pneumatic System ............................................................................ 12 2.1.2 Liquid Handling System .................................................................. 14 2.1.3 Diffusion Scrubber ......................................................................... 16 2.1.4 Gilson Model 121 Fluorometer ........................................................ 18

PAGE 5

v 2.1.5 Data Acquisition System ................................................................. 19 2.2 Laboratory Formaldehyde Analyzer ....................................................... 21 2.2.1 Pneumatic System ......................................................................... 21 2.2.2 Liquid Handling System ..................................................................... 22 2.2.3 Perkin Elmer LC 240 ......................................................................... 23 2.2.4 Data Acquisition System ................................................................. 23 2.3 Experimental Procedure ....................................................................... 24 2.3.1 Reagents ....................................................................................... 24 2.3.2 Calibration of Formaldehyde Analyzers ............................................ 24 3. Results and Discussion ............................................................................. 26 3.1 Reaction Chamber Optimization ............................................................ 26 3.2 Calibration Results ............................................................................... 27 3.2.1 Gilson Model 121 ........................................................................... 27 3.2.2 Limit of Detection .......................................................................... 30 3.2.3 Tylan FC 260 Mass Flow Controller ................................................... 32 3.3 Collection Efficiency .............................................................................. 34 3.4 Sampling Results ................................................................................. 38 3.4.1 Indoor Sampling ............................................................................ 38 3.4.2 Outdoor Sampling .......................................................................... 41

PAGE 6

vi 4. Conclusion ............................................................................................... 52 References ................................................................................................... 56

PAGE 7

vii LIST OF TABLES Tables Table Table 1: Partial List of Carbonyls Quantified With TO11A ................................. 7 Table 2: Standard HCHO Solutions ................................................................ 25 Table 3: Hantzsch Reaction Completion Data ................................................. 26 Table 4: Average Response voltage and Concentration ................................... 29 Table 5: Statistics for Zero Standard Solution ................................................ 31 Table 6: Response Per ppbv ......................................................................... 32 Table 7: Mass Flow Controller Calibration Data .............................................. 33 Table 8: Statistics for Collection Efficiency ...................................................... 37 Table 9: Hourly Indoor HCHO Concentrations ................................................ 39 Table 10: Outdoor Hourly Data (1/1/2012) .................................................... 44 Table 11: Outdoor Hourly Data (1/2/2012) .................................................... 46 Table 12: Outdoor Blank Sample Response .................................................... 49 Table 13: Hourly HCHO Concentration (ppbv) Summary ................................. 51

PAGE 8

viii LIST OF FIGURES Figures Figure Figure 1: The Photolysis of Ozone ................................................................... 3 Figure 2: The Produc t i on of HCHO From VOCs ................................................. 4 Figure 3: The Derivitization of HCHO with DNPH .............................................. 7 Figure 4: The Hantzsch Reaction ..................................................................... 9 Figure 5: Reaction of Formaldehyde With Fluoral P ........................................ 11 Figure 6: Pneumatic Layout of Field Analyzer ................................................. 13 Figure 7: Sampling Process of Liquid Handling System ................................... 15 Figure 8: Construction of Reaction Heater ..................................................... 15 Figure 9: Attachment of Nafion to Steel Tubing ............................................. 16 Figure 10: Elbow Union Layout ..................................................................... 17 Figure 11: Dual Channel Diffusion Scrubber ................................................... 18 Figure 12: Absorption Scan of Emission Filter ................................................ 19 Figure 13: LabView Interface for Field Analyzer ............................................. 20 Figure 14: Laboratory Scrubber Configuration ................................................ 22 Figure 15: Lab Analyzer Liquid Handling System ............................................ 22 Figure 16: Optimization of Reaction Heater .................................................... 27 Figure 17: Gilson Model 121 Calibration Data ................................................. 28 Figure 18: Gilson Model 121 Calibration Curve ............................................... 30

PAGE 9

ix Figure 19: Mass Flow Controller Calibration Curve .......................................... 34 Figure 20: Summary of Collection Efficiency Data ........................................... 35 Figure 21: Blank Solution Data (Ten Hours) ................................................... 36 Figure 22: Indoor Air Sampling Results .......................................................... 38 Figure 23: Hourly Average for Indoor Data .................................................... 39 Figure 24: Gas Flow Rate (12/21/2011) ......................................................... 41 Figure 25: Outdoor Sampling Results (12/24/2011) ........................................ 42 Figure 26: Outdoor Sampling Results (1/1/2012) ........................................... 43 Figure 27: Hourly HCHO Concentrations (1/1/2012) ....................................... 44 Figure 28: Gas Flow Rate (1/1/2012) ............................................................ 45 Figure 29: Outdoor Sampling Results (1/2/2012) ........................................... 45 Figure 30: Hourly HCHO Concentrations (1/2/2012) ....................................... 46 Figure 31: Gas Flow Rate (1/2/2012) ............................................................ 47 Figure 32: HCHO Concentrations (1/7/2012 1/10/2012) ............................... 49 Figure 33: Decreasing Trend of Span Concentration ....................................... 50

PAGE 10

1 1. Introduction 1.1 History of Formaldehyde Formaldehyde (HCHO) was first discovered in the late nineteenth century through the work of Butlerow and Hoffman. 1 2 Since its discovery, formaldehyde has been a major component of industrial processes. Indeed, shortly after its initial discovery, formaldehyde was found to be easily produced and had a wide range of uses from the development of industrial resins to com mon use as a disinfectant. In more recent decades, concerns have been raised regarding the potential hazards of formaldehyde and other chemicals to human health. Numerous studies have been conducted in the analysis of formaldehyde toxicity that has lead to the eventual classification of the compound as a likely carcinogen. 3 4 5 1.1.1 Sources of Formaldehyde The simplest of the aldehydes, formaldehyde is a ubiquitous compound that has numerous chemical interactions with the environment. It is important to understand the sources and sinks of formaldehyde and other volatile organic compounds (V OCs) in order to understand the importance of the chemical in environmental interactions. Formaldehyde is widely used within the chemical industry and is used in the manufacturing processes for resins, foam insulation and fabric processing. In

PAGE 11

2 fact, the particular usefulness of this compound has made it one of the most used chemicals in the manufacturing industry. In the cases of formaldehyde based resins used in the manufacturing of items such as particle board and other pressed wood products, the resulting products have had the greatest impact on the quality of indoor air. Formaldehyde based resins made with urea, melamine and phenol have been classified as formaldehyde sources either by the release of free formaldehyde residues or the decomp osition of the formaldehyde based resin into a gaseous pollutant 3 4 5 Recently, in the aftermath of Hurricane Katrina which devastated the city of New Orleans in 2005, the awareness of poor air quality due to high concentrations of formaldehyde has increased when emerg ency housing was found to have high levels of the compound inside 2 Industrial manufacturing is not the only source due to human activity. Formaldehyde is a common byproduct for incomplete combustion processes such as those in automobile engines and power production 3 5 In the 1980s the release of the pollutant to the atmosphere from mobile sources was estimated at 666 million pounds per year in the United States 4 This estimation does not include the potential contribution by stationary sources such as power plants fired by coal, natural gas and oil. The combustion of fossil fuels in any case has had the greatest impact on outdoor air quality as a result of human activity.

PAGE 12

3 Natural sources of VOCs also contribute to the overall level of atmospheric formaldehyde. Seco, et. al, described the characteristics of biogenic sources of short chain oxygenated VOCs such as formaldehyde. Plants naturally emit VOCs into the atmosphere as part of normal metabolic processes 6. Other chemical species emitted by natural biogenic sources, such as terpenes, also play an important role in the production of atmospheric formaldehyde through a reaction with radicals in the air 7 Non oxygenated VOCs play an important role in the photochemical production of HCHO in the atmosphere as detailed by Atkinson. 7 One of chemical interactions resulting in the production of atmospheric formaldehyde involves the interaction of n on oxygenated VOCs with the photochemically produced hydroxyl radical. One process by which the hydroxyl radical is generated is through the photolysis of tropospheric ozone. O + h O + O D ( 335 nm ) (1) O D + M O P + M ( M = N O ) (2) O P + O + M O + M ( M = air ) (3) O D + H O 2 OH (4) FIGURE 1 : THE PHOTOLYSIS OF OZ ONE

PAGE 13

4 When ozone is exposed to ultraviolet radiation in the presence of water in the troposphere, an excited oxygen atom is produced as demonstrated in Figure 1. Eventually, the combination of other atmospheric species produces the hydroxyl radical. OH + CH H O + C H (5) C H + O CH O (6) CH O + NO CH O + NO (7) CH O + O HCHO + HO (8) HO + NO OH + NO (9) FIGURE 2 : THE PRODUCT I ON OF HCHO FROM VOCS When the resulting hydroxyl species encounters VOCs in the atmosphere, another series of reactions occur eventually leading to the production of formaldehyde as well as more hydroxyl radicals Figure 2 outlines the process of producing formaldehyde through the oxidation of methane b y the hydroxyl radical and oxides of nitrogen (NOX) that are another important species in atmospheric pollution 7 1.1.2 Health Concerns of Formaldehyde Due to its ubiquitous nature extensive research has been done on the possible impacts of formaldehyde on human health. A significant number of studies have been done to explore the effects of formaldehyde based products upon contact

PAGE 14

5 with skin. These studies have shown that upon repeated expos ure to solutions containing formaldehyde, the subjects will tend to develop allergic dermatological reactions leading to rashes on the exposed skin 3 5 Most studies have involved observing the effect of formaldehyde gas on the respiratory system. Testing on humans has demonstrated that the gas commonly causes irritation to the eyes, nose and throat of human subjects. Based on these results, formaldeh yde has been classified as an irritant. To furthe r explore the potential impact of formaldehyde on the human respiratory system, medical studies using animals have been conducted 3 4 5 These animal studies have shown that prolonged exposure can lead to carcinomas developing in the nasal cavities of rats, which has prompted the chemical to be classified as a likely carcinogen and that exposure levels in both ambient and indoor air are to be regulated in workplace environments by the Occupational Safety and Hazard Administration. 8 In an effort to limit human exposure to formaldehyde, the ability to monitor and quantify the concentration of formaldehyde in the air has lead to the development of several methods for the analysis of air for Hazardous Air Pollutants.

PAGE 15

6 1.2 Measurement Techniques for Gaseous Formaldehyde In order to fully understand the impact of formaldehyde on the environment and to explore a correlation of environmental problems to formaldehyde concentrations, many different approaches have been developed to measure the compound in the atmosphere. Some designs have utilized derivitization of formaldehyde followed by instrumental analysis in a laboratory. Such designs tend to be well designed and easily deployed to the field. However, these methods occasionally lack the flexibility to be used in studies that require more rapid response or better time resolution As a result of this requirement, more versatile and specific methods have been developed for the near real time detection of atmospheric formaldehyde. 1.2 .1 EPA Method TO 11A The United States Environmental Protection Agency (EPA) has produced several methods for the detection and quantification of hazardous air pollutants ( HAPs ) including formaldehyde. Of particular interest, Compendium M ethod TO 11A (TO 11A) is used to detect reactive carbonyl compounds in air samples This method utilize s a cartridge consisting of a hollow tube containing a solid silica gel substrate. The soli d material present on the inter ior of the cartridge is coated with 2,4 dinitrophenyl hydrazine (DNPH) as a derivitizing agent and a known volume of air is passed thro ugh the cartridge As the air samples p ass through the sampling cartridge, the carbonyl species react with the coating and create a

PAGE 16

7 family of hydrazone compounds that are easily detected by High Performance Liquid Chromatography (HPLC) 9 The reaction of interest is outlined in Figure 3. (10) FIGURE 3 : THE DERIVITIZATION O F HCHO WITH DNPH The hydrazones produced through this reaction pathway are later separated through chromatographic means using a combination of purified water and purified acetonitrile mobile phase s The sampling cartridges are initially extracted using a known volume of purified acetonitrile. The DNPH derivatives are then quantified using reverse phase HPLC and detection with ultraviolet (UV) absorption from a 360 nm UV source lamp. A list of compounds detectable by TO 11A is presented in Table 1. TABLE 1 : PARTIAL LIST OF CARBONYLS QUANTIFIED WITH TO 11A Formaldehyde Isovaleraldehyde Propionaldehyde p Tolualdehyde Acetaldehyde Valeraldehyde Crotonaldehyde Hexanaldehyde o Tolualdehyde Butyraldehyde 2,5 Dimethylbenzaldehyde Methyl ethyl ketone Acetone m Tolualdehyde Benzaldehyde

PAGE 17

8 Method TO 11A is capable of quantifying formaldehyde in the low parts per billion by volume (ppbv) range over a time period of 1 24 hours. The most common sampling period used with this technique is 24 hours, producing a 24 hour average concentration reading 9 1.2.2 Continuous Flow DNPH Methods Additional studies have been done on the usefulness of DNPH derivitization of carbonyls using continuous flow systems and diffusion scrubbers. The techniques devised utilize the same separation and detection methods as TO 11A, but work in a continuous manner producing results with better time resolution. Zhou, et al. explored DNPH derivitization for several carbonyl species utilizing a glass diffusion coil containing an acidified DNPH solution in 2009. The solution was pumped through a collection chamber where the carbonyl compounds would pass through the glass coil and react with DNPH to produce the desired hydrazone derivatives. The resulting products were then analyzed by an inline HPLC where the compounds were quantified with a detection limit of 0.036 ppbv 10 A similar method was studied by Komazaki et al. in 1998 for the analysis of formaldehyde in automotive exhaust. A porous polytet ra fluoroethylene (PTFE) tube containing the DHPH solution collected formaldehyde from auto exhaust to be analyzed by an inline HPLC system 11

PAGE 18

9 1.2.3 Hantz s ch Reaction Derivitization In 1953, Nash published a work on the detection of formaldehyde through the use of the Hantzsch reaction. 12 The Hantzch reaction is an organic synthesis reaction that was discovered in the late nineteenth century that involves an aldehyde species, diketone and a source of ammonia or a primary amine 13 In the case of Nashs research, formaldehyde was combined with 2,4 pentanedione and ammonium acetate. The net re action is presented in Figure 4. (11) FIGURE 4 : THE HANTZSCH REACTIO N The product, 3,5 diacetyl 1,4 dihydrolutidine (DDL) was found to produce a yellow color and was used for colorimetric detection of formaldehyde Nash also noted that elevated temperatures accelerated the formation of DDL from five hours at room temperature to five minutes when heated near 60 C 12 Other groups have studied the usability of the Hantzsch synthesis of DDL for the detection of atmospheric formaldehyde. In most cases, a diffusion scrubber is used as the collection method of formaldehyde gas into an acidic solution. Dasgupta et al. have studied the use of diffusion scrubbers using porous

PAGE 19

10 membranes such as Celgard microporous polypropylene 14, 15 and Nafion 16, 17 After collection by the diffusion scrubber, the formaldehyde solution is mixed with the diketone and ammonium salts before passing through a heated reaction chamber. Once the DDL generated from the Hantzsch reaction has left the reaction chamber, it pas ses through a fluorometer utilizing soft UV radiation as an excitation source. These experiments have lead to the development of instruments capable of detecting formaldehyde at levels below 1 ppbv. More recently, the use of the Fluoral P reagent has been studied as a method of detecting carbonyl compounds. Fluoral P is an intermediate species of the Hantzsch reaction resulting from the reaction of ammonia with 2,4 pentanedione that produces DDL when combi ned with formaldehyde (Figure 5 ) Compton and Purdy found that the Fluoral P reagent is stable and able to be stored as an analytical reagent 18 In a continuation of this work, Pinheiro et al. explored the use of the Fluoral P rea gent in sampling cartridges. In this study, Fluoral P was used to coat sampling cartridges and sampled in parallel to DHPH coated cartridges. B oth sample types were analyzed through HPLC, one used UV detection and the other fluorescence detection. The results of the study show that the Fluoral P coated cartridges were able to detect formaldehyde concentrations that agree with st andard methods with a limit of detection of 0. 0 02 mg/L 19

PAGE 20

11 (12) FIGURE 5 : REACTION OF FORMALDEHYDE WITH FLUORAL P

PAGE 21

12 2. Experimental A combination of inhouse components and commercial solutions were used in the design and implementation of the formaldehyde gas analyzer. Two detections systems were developed utilizing a Nafion based diffusion scrubbers and fluorescence detectors. One was designed for field sampling and the second system was used for the determination of the collection efficiency of the primary diffusion scrubber. New bottles of ammonium acetate (Fishe r Chemical A639 500) were available in the lab. A new bottle of 37 % w/w formaldehyde (37.0 % certified) was ordered from Acros Organics (AC41073 1000) and new Nafion tubing was ordered from Perma Pure, LLC (TT 030). Open bottles of analytical grade sulfuric acid, acetic acid (Mallinckrodt 2504 500) and hydrochloric aci d were stored in a laboratory cabinet. An open bottle of 99+% pure 2,4 pentanedione was available, stored in a lab freezer (Aldrich P775 4). 2.1 Field Formaldehyde Analyzer The field analyzer consisted of three main systems: A pneumatic system, a liquid handling system and software for instrument control and data acquisition. 2. 1 .1 Pneumatic S ystem The vacuum pump used in the field analyzer was acquired from Thomas Industries, Inc. The flow of the vacuum pump was regulated by a Tylan FC 260

PAGE 22

13 mass flow controller (MFC) capable of 0 3 Sta ndard Liters p er Minute (SLPM) of O2 gas. The flow controller is adjustable by a variable voltage (0 5v) supplied by a data acquisition interfac e connected to a Lenovo Thinkpad notebook computer. The overall pneumatic system was configured to draw air through a 2 micron Teflon particulate filter and then through the back panel sampling inlet. The sample gas was then passed th rough the diffusion scrubber, MFC and vacuum pump before exiting the analyzers exhaust port. A three way pneumatic valve was installed downstream of the sampling inlet that allows for the production of blank sample by passing air through a DNPH coated sampling cartridge for collecting sample blanks. The analyzer layout is shown in Figure 6 All tubing used in the pneumatic system consisted of inch PTFE tubing with polypropylene or brass fittings. FIGURE 6 : PNEUMATIC LAYOUT OF FIELD ANALYZER

PAGE 23

14 2. 1 .2 Liquid Handling System The field analyzer utilized a liquid handling system consisting of a Cavro Scientific Instruments four channel syringe pump (Cavro XL 3000). The pump was equipped with 0.500 mL syringes and each syringe is equipped with a three port valve The syringes share a common drive motor controlled through a serial port connected to a Lenovo Thinkpad notebook computer. A custom designed heating chamber was developed inhouse consisting of a 40 watt ceramic heating element and regulated with a West 3100 temperature control unit equipped with a K Type thermocouple and a solid state relay. The heater is insulated with fiberglass insulation and contains 1.0 mL volume of 23 AWG PTFE thin walled tubing. A solenoid valve was pla ced in the system to allow for the injection of a formaldehyde span solution for checking detector response at regular intervals The valves were controlled by the data acquisition interface board digital output ports and solid state relays The scrubbing solution and a solution containing ammonium acetate, acetic acid and 2,4 pentanedione (Nash reagent) used in this system were combined before being pumped into the reaction heater through a three way PEEK union. An Omni fit 006BT bubble trap was installed downstream of the reaction heater to remove bubbles from the solution before detection. The diagram in Figure 7 illustrate s the liquid handling system of the field analyzer. Figure 8 outlines the construction of the reaction heater and dir ection of sample flow

PAGE 24

15 FIGURE 7 : SAMPLING PROCESS OF LIQUID HANDLING SYST EM FIGURE 8 : CONSTRUCTION OF REAC TION HEATER

PAGE 25

16 2. 1 .3 Diffusion Scrubber The diffusion scrubber for the field analyzer was designed to withstand the pressure of the syringe pump. Nafion was selected as the diffusion membrane as it is selective for formaldehyde and prevents other carbonyl compounds from passing into the scrubb ing solution. 20 1/16 inch stainless steel tubing used for HPLC systems was tapered at the ends and a groove was cut behind the tip, producing a barbed end. The Nafion tubing was soaked in pure methanol, as the membrane increases in size by 88% 21 The steel tubing was inserted through the open ends of the expanded membrane tubing and allowed to dry. The dried membr ane tightened around the barbed tip securing the membrane to the tub ing. The connection between the Nafion and the steel tubing is illustrated in Figure 9 FIGURE 9 : ATTACHMENT OF NAFION TO STEEL TUBING

PAGE 26

17 To improve the collection efficiency of the field analyzers diffusion scrubber, the length of the scrubber was extended in comparison to other studies 19 Two inch polypropylene elbow unions two to 1/8 inch reducing elbows and two 16 inch lengths of PTFE tubing were configured in a dual channel configuration. A small hole was drilled in each elbow to be concentric with one opening of the fitting. Loctite All Plastics super glue was used to bond an autosampl er vial septum over each drilled hole to pneumatically seal the scrubber fittings A 15 inch length of Nafion tubing was fed through each of the lengths of PTFE tubing and the attached metal pierced the septa. The Nafion sections were attached together by a section of metal tubing inserted through two septa attached to opposite sections of PTFE tubing. A section of 1/8 inch PTFE tubing was used to connect the sections of inch PTFE tubing. Clear polyvinylchloride (PVC) tubing with an inner diameter o f inches was used to support the straight design of the diffusion scrubber. The structure of the diffusion scrubber is outlined in Figure s 10 and 11. The scrubbing solution was configured to run anti parallel to the flow of sample gas. FIGURE 10: ELBOW UNION LAYOUT

PAGE 27

18 FIGURE 11: DUAL CHANNEL DIFFUSI ON SCRUBBER 2. 1 .4 Gilson Model 121 Fluorometer A Gilson Model 121 filter fluorometer was used to detect DDL produced from the Hantzsch reaction. The fluorometer is configured with a 9 L vertical square flowcell and a nine stage photomultiplier tube (PMT). The standard source lamp was replaced with a MR16UV2415 lamp (Super Bright LEDs, Inc.) consisting of 24 light emitting diodes (LEDs) rated at a peak wavelength of 405 nm. A 1 mm pinhole slit was installed on the excitation window of the flowcell chamber to center the incident light on the flat surface of the flowcell. A cut on filter was i nstalled on the emission window of the flowcell chamber to block incident light from the PMT. The emission filter spectrum is illustrated in Figure 12. The Gilson 121 has a selectable time constant, ranging from 0.5 s, 1.0 s and 2.0 s as well as an output voltage range of 10 mV, 100 mV and 1.0 V. The front panel of the detector has an auto zero button and a selectable sensitivity range is of 1 0.001.

PAGE 28

19 FIGURE 12: ABSORPTION SCAN O F EMISSION FILTER 2.1.5 Data Acquisition System Custom data acquisition software was developed with the LabView 2010 Student Edition software development environment. A program was developed with the capability to sample multiple voltage readings at a defined frequency and average s everal values to reduce signal noise. The voltage readings are made from a Minilab 1008 USB data acquisition interface by Measurement Computing and acquired on a Lenovo Thinkpad notebook computer. The Minilab 1008 is a low cost 12 bit data acquisition bo ard that also has the ability to send and read digital data as well as produce an analog output voltage Several digital ports on the board were configured to trigger relays that power pneumatic and liquid handling valves in other process systems. This s ystem provides the flexibility to modify the process system as adjustments are required. Fluorescence and flow

PAGE 29

20 readings from the MFC are saved to a text file capable of being used in a spreadsheet program and a new file is generated every 24 hours. A vir tual front panel is displayed on the computer screen that allows adjustments to the data acquisition process and displays current voltage readings for flow and fluorescence. Figure 13 presents the LabView program interface as displayed on the computer scr een. The program was also the primary method of communication to the syringe pump through the serial port of the computer. FIGURE 13: LABVIEW INTERFACE FOR FIELD ANALYZER

PAGE 30

21 2. 2 Laboratory Formaldehyde Analyzer The laboratory analyzer consisted of three main systems: A pneumatic system, a liquid handling system, and software for data acquisition. 2. 2 .1 Pneumatic System The pneumatic system of the laboratory formaldehyde analyzer consists of a Nafion diffusion scrubber and several sections of inch PTFE tubing. The lab apparatus was designed to be used in series with the field analyzer for the determination of the collection efficiency of the dual channel diffusion scrubber. The laboratory diffusion scrubber is simpler than the field design (Figure 13) A single length of inch PTFE tubing 15 inches in length was used as the scrubber sleeve. 15 inches of Nafion was fed through the jacket and aligned in a concentric configuration. Two inch polypropylene tee unions were used to seal the ends of the scrubber jacket. Two inch tube caps were drilled and Loctite All Plastics super glue was used to glue autosampler vial septa over the holes. The Nafion tubing was attached to metal tubi ng in the same fashion as the field scrubber (Figure 11) and fed through the septa on the tube caps. The open ports on the tee unions were connected to a vacuum source as needed. When in use, the scrubber was connected between the dual channel diffusion scrubber and the MFC of the field analyzer. All connections used in the lab analyzers pneumatic system were polypropylene fittings.

PAGE 31

22 FIGURE 14: LABORATORY SCRUBB ER CONFIGURATION 2. 2 .2 Liquid Handling System Several components were used in the liquid handling system of the laboratory formaldehyde analyzer. The pump used in the lab was a Gilson Minipulse2 multi head peristaltic pump. Two heads of the pump were used to pump the scrubbing fluid and the Nash reagent. A three port polypropylene union was installed downstream of the diffusion scrubber and before a Timberline Instruments column heater (S/N 1092). The column heater can be refitted with internal 1/16 inch PTFE of different lengths to change the internal vol ume. A second bubble trap was not available for use in the second formaldehyde analyzer. FIGURE 15: LAB ANALYZER LIQU ID H ANDLING SYSTEM

PAGE 32

23 2. 2 .3 Perkin Elmer LC 240 The fluorometer selected for the detection of DDL in the laboratory analyzer was a P erkin Elmer LC 240. The LC 240 contains dual monochrometer s for selecting excitation and emission wavelengths independently for maximum sensitivity. The 4.0 L flowcell is a 1.0 mm squa re fused silica flowcell oriented in a horizontal arrangement. The output range of the detector is 0 1 V and the time constant is 1.8 s. A Xenon discharge lamp is used as the light source and produces light in the UV and visible spectrum. An excitatio n wavelength of 419 nm and an emission wavelength of 516 nm were experimentally determined for the detection of DDL. 2. 2 .4 Data Acquisition S ystem LabView 2010 Student Edition was used to develop a second data acquisition software solution. The laboratory software was designed to take voltage readings from a second Minilab 1008 USB data acquisition board. The program collected data from two separate volta ge sources at a specified frequency and average d each signal into one minute values. No control programming was needed for the second formaldehyde analyzer. Signals from both fluorescence detectors were connected to the second Minilab 1008 board for direct comparison of DDL detection signals.

PAGE 33

24 2.3 Experimental Procedure 2.3.1 Reagents For the continuous flow analyzers, Nash reagent and dilute sulfuric acid were made. 500 mL of 0.65 M Nash reagent was made by mixing 24.5 g of ammonium acetate, 40 L glacial acetic acid and 438 L of 2,4 pentanedione with Millipore deionized water in a 500 mL volumetric flask. 500 mL of 1.0 x 105 M sulfuric acid was made by mixing 100 uL 0.1 M H2SO4 with Millipore deionized water in a 500 mL volumetric flask. Each s olution was transferred to a separate 500 mL amber glass bottle capped and labeled. The reagents were allowed to equilibrate for a day before use at room temperature. 2.3.2 Calibration of Formaldehyde Analyzers The Gilson Model 121 fluorometer was calibrated with the use of formaldehyde standard solutions. The solutions were made by dissolving 0.5 mL of 37% w/w formaldehyde (37.0 % certified) to 1000 mL in a volumetric flask. 3 .000 mL of the diluted solution was transferred to a 250 mL volumetric flask with a volumetric pipette. The standard solutions were made by transferring 1 mL of 1.00 x 103 M H2S O4 to a series of 100 mL volumetric flasks with a volumetric pipet te Aliquots of the second stock formaldehyde solution in the volumes of 5, 10, 15, 20 and 25 mL were transferred to the 100 mL volumetric flasks with volumetric pipettes Table 2 summarizes the calculated concentrations of the formaldehyde standards.

PAGE 34

25 TABLE 2 : STANDARD HCHO SOLUTI ONS Volume of HCHO Stock Solution (mL) Calculated Standard Concentrations (M) Calculated ppbv Equivalent (ppbv) 5.00 3.84E 06 11.1 10.00 7.69E 06 22. 3 15.00 1.15E 05 33.4 20.00 1.54E 05 44. 6 25.00 1.92E 05 55.7 0.00 0.00 0.00

PAGE 35

26 3 Results and Discussion 3.1 Reaction Chamber Optimization To maximize the detection capability of the formaldehyde analyzer, the Nash reagent was mixed with a formaldehyde solution and heated at 60 C to accelerate the reaction and prevent bubble formation. The solution was held in the Timberline heater and heated for a measured time. The reaction product was then flowed into the LC 240 fluorometer and the fluorescence was measured for each sample. The fluorescence response was measured for heating time intervals from 1 minute to 4 5 minute s The data are summarized in Table 3 below. TABLE 3 : HANTZSCH REACTION COMPLETION DATA Heating Time (min) Response ( V ) Heating Time (min) Response (V ) 45 0.5824 8 0.5707 1 0.2317 9 0.5697 2 0.3744 10 0.5731 3 0.4774 11 0.5766 4 0.5253 12 0.5795 5 0.5507 13 0.5761 6 0.5648 14 0.577 0 7 0.5653 15 0.5751 Plotting the data in Table 3 effectively illustrates the process of the Hantzsch reaction. The reaction appears to slow significantly after 5 minutes of heating at 60 C (Figure 16). Using the reaction data, the flow of the liquid handling

PAGE 36

27 system was optimized so the reagents are heated for at least 5 minutes in the reaction chamber of the field analyzer. FIGURE 16: OPTIMIZATION OF REAC TIO N HEATER 3. 2 Calibration Results Two components of the field analyzer required calibration before use. The Gilson Model 121 fluorescence detector and the Tylan FC 260 mass flow controller were calibrated and the data are summarized below 3.2 .1 Gilson Model 121 The Gilson fluorometer was configured with a 2.0 s time constant, 100 mV output voltage and a sensitivity setting of 0.1. The Nash reagent and the HCHO stand ard s were dispensed into the threeway PEEK union, heated with the 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 10 20 30 40 50 Response (v) Reaction Time (min) Optimization of Reaction Heater

PAGE 37

28 reaction heater of the field analyzer through the Gilson fluorometer. Each standard was dispensed for a period of twenty minutes followed by twenty minutes of fresh 1.0 x 105 M H2SO4. The one minute data are summarized in Figure 17. FIGURE 17: GILSON MODEL 121 CALIBRATION DATA A calibration curve and regression equation was genera ted by averaging the values of each sample peak and performing a linear regression operation of the data points A calculated concentration for the ppbv equivalent values was calculated from the standard solution molarities using equations 13 and 14 The resulting calibration curve is summarized in Table 4 and Figure 1 8 0 0.05 0.1 0.15 0.2 0.25 0 50 100 150 200 Signal (Volts) Time (min) Gilson Calibration Data

PAGE 38

29 Ca lculation for ppbv Equivalent (1.00 L/min gas flow, 293 K and 84 kPa): = ( 10 ) 1 (13) 3 84 10 0 0001 30 04 1000 1 00 = 1 15 10 (14 ) 1 15 10 10 30 04 8 3144 ( 293 ) 1 84 = 11 1 (15 ) TABLE 4 : AVERAGE RESPONSE VOL TAGE AND CONCENTRATI ON Average Signal (v) Concentration (ppbv) 0.074315 0 .00 0.102111 11.1 0.134859 22. 3 0.166064 33.4 0.196655 44. 6 0.230576 55.7

PAGE 39

30 FIGURE 18: GILSON MODEL 121 CALIBRATION CURVE As illustrated in the figure above, a linear relationship was observed between the concentration of HCHO and relative voltage output of the fluorometer. The slope of the line was 355.66 ppbv/V and the intercept was 25.758 V. The correlation coefficient indicates a strong relationship between the two variables with a R2 of 0. 9994. The slope and intercept values were entered into the LabView Virtual Instrument interface. 3.2 .2 Limit of Detection The limit of detection (L OD) for the Gilson Model 121 fluorometer was calculated by calculating the mean and standard deviation of the zero formaldehyde y = 355.66x 25.758 R = 0.9994 0 10 20 30 40 50 60 0 0.05 0.1 0.15 0.2 0.25 Concentration (ppbv) Signal (V) Calibration Curve

PAGE 40

31 standard solution voltage in Microsoft Excel The zero standard statistics are summarized in Table 5 TABLE 5 : STATISTICS FOR ZERO STANDARD SOLUTION Signal Mean ( ) Limit of Detection (LOD) 0.0743 V 0.000508 V 0.0758 V Equations 15 through 17 were used to calculate the limit of detection by adding three times the standard deviation of the zero standard to the mean. Using the regression equation from the calibration curve, the voltage value was translated into a ppbv equ ivalent value. = + 3 (15) LOD = 0.0743 V + 3( 0.0005 08 V) = 0.0758 V (16) LOD = 355.66 ppbv/V *( 0.0758 V) 25.758 ppbv = 1.22 ppbv (17) The voltage for the LOD was determined to be 7 5.8 mV corresponding to a concentration of 1.22 ppbv assuming a standard gas flow rate of 1.00 SLPM. To further characterize the calibration data, the concentration of the zero standard was calculated from the average zero standard voltage. The concentration was found to be 0.673 ppbv. The regression dat a was calculated from a single set of calibration data.

PAGE 41

32 The signal voltage response for a concentration change of 1 ppbv was calculated by taking the difference of the average signal for each calibration p eak and the zero standard. The voltage difference was divided by the concentration of the respective calibration peak, and the dividends were averaged. An example calculation is presented in equation 18. A summary of the data are presented in Table 6. = 0 74315 1000 = 0 102111 0 74315 11 1 1000 = 2 5 / (18) TABLE 6 : R ESPONSE PER PPBV Concentration (ppbv) Voltage / Concentration (mV/ppbv) 11.1 2 49 22. 3 2 7 2 33.4 2 74 44. 6 2 74 55.7 2 80 Average 2 70 Standard Deviation 0 12 The average response per ppbv was calculated to be 2.70 mV 0.12 mV. 3. 2 .3 Tylan FC 260 Mass Flow Controller The MFC installed in the field analyzer was calibrated using a BIOS flow calibrator and a volt meter connected to the MFC output voltage terminals. The

PAGE 42

33 BIOS flow cali brator was used to average ten flow measurements for each run The data are summarized in Table 7 and Figure 1 9 TABLE 7 : MASS FLOW CONTROL LER CALIBRATION D ATA MFC Flow Output Voltage (Volts) Flow Calibrator Output (SCCPM) 0.151367 32.37 0.336914 175.0 0.664062 437.5 0.996094 709.8 1.32813 982.8 1.66016 1256 1.99219 1540 2.32422 1822 2.66113 2111 2.98340 2408 3.28125 2664 3.28613 2678

PAGE 43

34 FIGURE 19: MASS FLOW CONTROL LER CALIBRATION CURV E A linear regression operation was performed on the calibration data. The regression equation presented in Figure 18 indicates a strong linear correlation with a R2 value of 0.9994. The slope of the equation is 0.8448 SLPM per volt, and the intercept is 0.1238 volts. The slope an intercept values were entered into the calibration fields of the LabView Virtual Instrument interface. 3.3 Collection Efficiency The co llection efficiency of the primary diffusion scrubber was determined by measuring breakthrough concentrations of formaldehyde with the laboratory gas y = 0.8448x 0.1238 R = 0.9996 0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3 3.5 Flow Calibrator Output (SLPM) MFC Output Voltage (V) Tylan FC -260 Calibration Curve

PAGE 44

35 analyzer. The laboratory instrument was connected downstream of the field analyzer sample gas flow. A so lenoid valve was installed to redirect the sample gas flow to the laboratory diffusion scrubber directly from a permeation device every two hours resulting in alternating high reference signal and sample breakthrough signals The day before the collection efficiency experiment was started, ten hours of a blank sample was sampled by the laboratory analyzer by installing a DNPH sampling cartridge on the instrument gas sample inlet The output from the Perkin Elmer LC 240 is summarized in Figure s 20 and 21. FIGURE 20: SUMMARY OF COLLECTION EFFICIENCY DATA 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0 200 400 600 800 1000 1200 1400 1600 Signal Response (V) Sample Number (every minute) Summary of Collection Efficiency

PAGE 45

36 FIGURE 21: BLANK SOLUTION DA TA (TEN HOURS) The sloping peaks in Figure 20 indicate a pneumatic leak was likely present in the system during the collection efficiency measurements. All pneumatic fittings were tightened at the conclusion of the experiment. The collection efficiency was calculated by comparing the relative respo nse of the high and low signals to the zero signal collected on the previous day. The standard deviation for the high reference and breakthrough sample was calculated from all points of the signal peaks and valleys respectively. The standard deviation of the zero reference was calculated from the ten hours of data prepared the day before the collection efficiency experiment was performed. The results for the sample statistics are summarized in Table 8 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0 100 200 300 400 500 600 Signal Response (V) Sample Number (every minute) Collection Efficiency Solution Blank

PAGE 46

37 TABLE 8 : STATISTICS FOR COLLECTION EFFICIENCY Signal Mean (V) Standard Deviation (V) High Reference (H) 0.055832 0.002643 Breakthrough Sample (L) 0.021369 0.001062 Zero Reference (B) 0.021385 0.000862 The statistics presented in Table 5 were used to calculate the collection efficiency of the diffusion scrubber installed in the field analyzer using Equations 1 9 2 2 The collection efficiency was calculated to be 100.05 % 3.97 %. Calculation for Collection Efficiency: 1 ( ) ( ) 100 (1 9 ) 1 ( 0 0021369 0 021385 ) ( 0 055832 0 021385 ) 100 = 100 05 (20 ) Calculation for Propagation of Error: = + + (2 1 ) ( ( 1 30965 ) ( 0 002643 ) + ( 2902 99 ) ( 0 001062 ) + ( 2904 30 ) ( 0 000862 ) ) = 3 97 (2 2 )

PAGE 47

38 3.4 Sampling Results The field analyzer was configured to sample air in both indoor and outdoor environments. Several days of data were collected in the research laboratory and in an outdoor research shelter on the Auraria Campus in Denver, Colorado. 3.4.1 Indoor Sampling The analyzer was set up in the research laboratory with the inlet sampling tubing open to the ambient laboratory air. The software was programmed to begin sample collection at 12:00 am on December 21, 2011. A 44.6 ppbv equivalent span solution was installed for injection through the diffusion scrubber during the first hour of data acquisition. A full 24 hours of samples were collected on December 21. The results of the sampling are presented in Figure 22. FIGURE 22: INDOOR AIR SAMPLI NG RESULTS 0 5 10 15 20 25 30 35 40 45 50 0 200 400 600 800 1000 1200 1400 1600 Formaldehyde Concentration (ppbv) Sample Number (Minute) 12/21/2011 Lab Air Data

PAGE 48

39 The data presented in Figure 22 shows the span solution peaking near 50 ppbv when the reference value is 44.6 ppbv. Each data peak present in the data summary represents one hour of sampling, with t en minutes of a zero sample per hour. The average values of the peaks were calculated and summarized in Figure 23 and Table 9 FIGURE 23: HOURLY AVERAGE FO R INDOOR DATA TABLE 9 : HOURLY INDOOR HCH O CONCENTRATIONS Hour HCHO (ppbv) Hour HCHO (ppbv) Hour HCHO (ppbv) 0 42.9 Span 8 2.80 16 2.27 1 4.87 9 2.13 17 2.11 2 3.38 10 1.80 18 1.93 3 2.60 11 2.00 19 2.31 4 2.49 12 2.24 20 2.43 5 2.35 13 2.70 21 2.33 6 2.51 14 2.17 22 2.49 7 2.88 15 2.19 23 2.49 0 1 2 3 4 5 6 0 5 10 15 20 25 HCHO Concentration (ppbv) Hour 12/21/2011 Hourly Data

PAGE 49

40 The indoor HCHO concentration appears to be relatively higher early in the morning at the 1 am hour and another peak appears during the 7 am 9 am hours. The concentrations range from 1.80 to 4.87 ppbv inside the laboratory on December 21, 2011. Ten hours of the 24 hour sampling period have blank concentrations above the 1.2 2 ppbv LOD for the instrument. Further evaluation of the LOD was conducted by taking the standard deviation of the blank concentration data collected during the indoor sampli ng period. A total of 75 data points were used as a single data set to calc ulate the LOD from sampling data. The 75 points consisted of three minutes from each hourly blank sample in Figure 22. The standard deviation was calculated to be 0.715 ppbv using the Microsoft Excel STDEV function. The LOD was calculated using equation 23. = 3 = 3 ( 0 715 ) = 2 15 (23)

PAGE 50

41 FIGURE 24: GAS FLOW RATE (12 /21/2011) The gas flow of the analyzer appeared to have drops in the signal near the time the valve switched to sample zero air and spikes when switching back to sample air (Figure 24) The fittings on the scrubber and pneumatic valve were tightened during the 2 pm hour. The signal spikes were significantly reduced in magnitude. The average flow rate was 1.05 SLPM 0.02 SLPM. 3.4.2 Outdoor Sampling On December 23, 2011 the field analyzer was relocated to a sheltered air sampling station located near the intersection of 7th avenue and Lawrence Way in Denver, Colorado. The roof of the structure was equipped with a glass sampling inlet manifold eight feet high with a bent end to prevent precipitation from entering the inlet. The analyzer software was configured to begin sampling 0.1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 160 40 240 440 640 840 1040 1240 1440 Flow (SLPM) Sample Number (every minute) Gas Flow Rate (12/21/2011)

PAGE 51

42 from the inlet and collecting data on Dece mber 24 at 12 am. Initial results presented significant interference from bubbles in the fluorometer flow cell as HCHO concentration readings below 0 ppbv The 12/24/2011 results are presented in Figure 2 5 FIGURE 25: OUTDOOR SAMPLING RESULTS (12 /24/2011) The bubble trap was disassembled and the internal membrane of the device was replaced. Further sampling was conducted beginning at 12 am on December 29, 2011. The sensitivity setting of the Gilson fluorometer w as checked on December 31, 2011 and found to be set to a sensitivity of 0.2. The sensitivity was reset to 0.1. Due to the sensitivity having been set incorrectly, the outdoor data collected in 2011 was not usable for the quantification of HCHO. 20 10 0 10 20 30 40 50 60 70 80 0 200 400 600 800 1000 1200 1400 1600 HCHO Concentration (ppbv) Sample Number (every minute) Outdoor Sampling Data 12/24/2011

PAGE 52

43 On January 1, 2012, the analyzer response was checked. The data collected during January 1 did not appear to have significant bubble interference, but the voltage output appeared to have blank signals higher in concentration than outdoor air. At 11 am on January 1, the ca rbonyl scrubber was replaced and the span solution was replaced with a 33.4 ppbv standard The collected on that day indicates that the carbonyl scrubber was repaired. On January 2, a fresh batch of Nash reagent was installed in the analyzer at 12 pm. The data collected on January 1 and 2 is presented in Figure s 2 6 through 31 and Tables 10 and 1 1 FIGURE 26: OUTDOOR SAMPLING RESULTS (1/1/2012) 0 5 10 15 20 25 30 35 40 45 0 200 400 600 800 1000 1200 1400 1600 HCHO Concentration (ppbv) Sample Number (every minute) Outdoor Sampling Data 1/1/2012

PAGE 53

44 FIGURE 27: HOURLY HCHO CONCENTRATIONS (1/1/2012) TABLE 10: OUTDOOR HOURLY DA TA (1/1/2012) Hour HCHO (ppbv) Hour HCHO (ppbv) Hour HCHO (ppbv) 0 2.19 8 3.27 16 1.46 1 1.87 9 3.62 17 2.25 2 2.09 10 3.48 18 2.88 3 2.39 11 2.97 19 2.64 4 2.05 12 1.86 20 3.22 5 2.05 13 1.67 21 3.46 6 2.13 14 1.73 22 2.54 7 2.34 15 1.53 23 37.2 (Span) 0 1 2 3 4 5 6 0 5 10 15 20 25 HCHO Concentration (ppbv) Hour Hourly Averages 1/1/2012

PAGE 54

45 FIGURE 28: GAS FLOW RATE (1/ 1/2012) FIGURE 29: OUTDOOR SAMPLING RESULTS (1/2/2012) 0 0.2 0.4 0.6 0.8 1 1.2 1.4 160 40 240 440 640 840 1040 1240 1440 Flow (SLPM) Sample Number (every minute) Gas Flow Rate (1/1/2012) 10 0 10 20 30 40 50 60 70 0 200 400 600 800 1000 1200 1400 1600 HCHO Concentration (ppbv) Sample Number (every minute) Outdoor Sampling Data 1/2/2012

PAGE 55

46 FIGURE 30: HOURLY HCHO CONCENTRATIONS (1/2/ 2012) TABLE 11: OUTDOOR HOURLY DA TA (1/2/2012) Hour HCHO (ppbv) Hour HCHO (ppbv) Hour HCHO (ppbv) 0 3.77 8 4.33 16 1.35 1 5.08 9 3.89 17 2.46 2 4.74 10 2.71 18 2.39 3 3.70 11 2.36 19 2.10 4 3.38 12 1.89 20 1.75 5 3.61 13 1.97 21 1.53 6 3.46 14 2.02 22 1.77 7 4.97 15 1.91 23 31.5 (Span) 0 1 2 3 4 5 6 0 5 10 15 20 25 HCHO Concentration (ppbv) Hour Hourly Averages 1/2/2012

PAGE 56

47 FIGURE 31: GAS FLOW R A TE (1/2/2012) During the course of outdoor sampling, the span injection process shifted in time from the originally scheduled time due to timing issues with the LabView program On the first day of outdoor operation, the second span process executed an hour early resulting in a span injection at 11 pm instead of 12 am the following day. After the initial ti me shift, the span consistently executed at 11 pm on later days. Sampling ended at 11 am on January 3, 2012 due to a short supply of H2SO4 solution. On January 1, the HCHO concentrations peaked during the 7 am 9 am hours and 6 pm 9 pm hours. The HCH O concentrations ranged from 1.46 ppbv to 3.46 ppbv. On January 2, the HCHO concentrations peaked during the 12 am 0 0.2 0.4 0.6 0.8 1 1.2 1.4 160 40 240 440 640 840 1040 1240 1440 Flow (SLPM) Sample Number (every minute) Gas Flow Rate (1/2/2012)

PAGE 57

48 3 am hours and 7 am 9 am hours. The HCHO concentrations ranged from 1.34 ppbv to 5.08 ppbv. Bubble interference was noted on this day. On both days of sampling, drops in the gas flow rate were noted at the times that the gas valve switched to sample a blank. Each of these drops lasted for only one minute. The average gas flow of the analyzer was calculated to be 1.05 SLPM 0.01 SLPM on each day. Fresh 1 liter batches of both the Nash reagent and the H2SO4 solution were installed in the analyzer. A new Omni fit 006BT bubble trap was installed with a new membrane upstream of the fluoromete r. Formaldehyde concentrations for January 7 through January 10 were collected and summarized in Figure 32 The hourly data values were based on absolute response of the fluorometer.

PAGE 58

49 FIGURE 32: HCHO CONCENTRATIO NS (1/7/2012 1/10/2012) Significant drift is apparent in the data spanning the four days of sampling. The blank response for each day appear s to be higher in the early hours of the morning (7 am 10 am) and then significantly lower in the late afternoon hours (5 pm 8 pm) Th e overall response range for the blank samples was 2.27 ppbv on January 10 to 1.69 ppbv on January 7. The daily blank response ranges are presented in Table 12. TABLE 12: OUTDOOR BLANK SAM PLE RESPONSE Date High Blank (ppbv) Hour Low Blank (ppbv) Hour 7 Jan 1.69 9:00 0.41 20:00 8 Jan 1.2 1 9:00 0.7 8 17:00 9 Jan 0.73 10:00 1.5 5 20:00 10Jan 0.2 8 7:00 2.27 18:00 4 day Range 1.69 9:00 2.27 18:00 1 0 1 2 3 4 5 0 5 10 15 20 25 HCHO Concentration (ppbv) Hour Raw Hourly Data Summary 10 Jan 9 Jan 8 Jan 7 Jan

PAGE 59

50 Negative HCHO concentrations were calculated from the fluorometer for several hours during the sampling period. The span solution signals were also evaluated and found to decrease over the course of the four days, indicating a possible problem with the detector sensitivity. The decreasing span val ues are summarized in Figure 33. FIGURE 33: DECREASING TREND OF SPAN CONCENTRATIO N The absolute response of the span solution dropped from 2 9.2 ppbv to 16.7 ppbv from January 7 to January 10. Baseline correction was applied to the span data by calculating the difference of the peak and the zero sample before and after the peak. The corrected span concentrations indicate a drift from 28.3 ppbv to 17.9 ppbv from January 7 to January 10. This de crease was not expected in the 0 5 10 15 20 25 30 35 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 HCHO Concentration (ppbv) Day of Sampling Period HCHO Span Concentrations 10Jan 9 Jan 8 Jan 7 Jan

PAGE 60

51 sampling results. The sample gas flow during the sampling period was 1.05 SLPM 0.02 SLPM. The hourly HCHO concentrations are summ arized in Table 13. TABLE 13: HOURLY HCHO CONCENTRATION ( PPBV) SUMMARY Hour 7 Jan 8 Jan 9 Jan 10Jan 0 29.20 (Span) 2.00 1.69 0.19 1 3.18 2.41 1.44 1.09 2 2.30 2.83 1.50 1.70 3 2.71 1.99 1.36 2.32 4 1.77 1.78 1.58 2.32 5 1.90 2.07 2.15 2.36 6 2.02 2.33 3.09 4.76 7 2.82 2.09 3.62 1.95 8 2.96 2.72 3.02 1.95 9 2.49 2.94 2.56 1.60 10 2.08 1.89 2.80 0.85 11 1.61 1.70 2.11 0.35 12 1.69 1.48 1.63 0.02 13 1.30 1.20 0.86 0.50 14 0.88 0.75 0.21 0.60 15 0.91 0.37 0.18 0.99 16 0.72 0.34 0.23 0.82 17 0.64 0.68 0.51 0.34 18 0.88 2.44 0.98 0.85 19 0.66 2.70 0.30 0.68 20 0.79 2.24 0.46 0.16 21 0.73 1.28 0.19 0.01 22 0.86 27.00 (Span) 22.14 (Span) 16.70 (Span) 23 1.31 1.14 0.42 0.53

PAGE 61

52 4 Conclusion A continuous flow formaldehyde analyzer was successfully developed, capable of quantifying concentrations above 1.22 parts per billion by volume in ambient air. The Hantzsch reaction was utilized to convert the formaldehyde samples into a product with fluorescent properties when excited with 405 nm light. A diffusion scrubber utilizing Nafion as a permeable membrane was constructed with a collection efficiency calculated to be 100.05 % 3.97 %. The Gilson Model 121 fluorometer was calibrated for a 0 55 ppbv range with a signal response of 2.7 0 mV per ppbv 0.12 mV. The limit of detection of the analyzer was expected to be below 1 ppbv, but was calculated to be 1.22 ppbv from the calibration data The limit of detection was also calculated from sampling data collected during the indoor sampling experiment. The detection limit was found to be 2.15 ppbv, a value 76 % higher than derived from the calibration blank. The increased standard deviation of the blanks signal during sampling could likely be due to a pneumatic leak near the carbonyl scrubber that led to a high blank concentration. One of the primary sources of signal noise in the analyzer is the result of changing pressure in the liquid handl ing system. The syringe pump used for dispensing the reagents through the analyzer uses one minute long injections. The pressure in the system drops w hen the syringes stop to refill with fresh

PAGE 62

53 reagents, resulting in a decrease in fluorescence in the dete ctor. To reduce the noise from pressure changes, using a pair of dual syringe pumps configured to provide a continuous pressure would significantly reduce the noise and lower the limit of detection. Alternatively, a pulse dampener could be used to equali ze the pressure of the system and an increased electronic sampling rate of the fluorescence signal would lower signal noise to improve the signal to noise ratio The analyzer was used to analyze outdoor ambient air for several days. Several issues were noted during the sampling process. Leaks in the pneumatic system of the gas analyzer have proven to be problematic for the accurate collection of a zero gas sample. A leak near the carbonyl scrubber allow ed for indoor air to enter the sampling system and produce erroneous values. Many of the blank samples taken during sampling are not likely to truly be zero due to the leaks in the system. As reported, several bl ank samples produced signals above the limit of detection indicating that a leak is highly probable. To improve the system, the carbonyl scrubber needs to be better sealed in the pneumatic system with the use of rubber gaskets to prevent air from bypass ing the cartridge and should be replaced periodically An alternative method that may be useful for reading a zero concentration of formaldehyde is to stop gas flow through the diffusion scrubber altogether. Using this method, the background of clean H2SO4 and Nash reagents may be obtained without introducing formaldehyde contamination to the system.

PAGE 63

54 During several days of sampling outdoor ambient air, an apparent drift in the signal baseline was noted ranging from 2.27 ppbv to 1.69 ppbv The Gilson model 121 used for detecting DDL was designed as a detector for HPLC in a controlled laboratory environment. The drift noted may very likely be due to a variation in outdoor temperature and the photomultiplier tube in the detector, resulting in lower sensiti vity during colder temperatures and higher sensitivity during warmer temperatures. Due to these changes in sensitivity, the calculated concentrations of HCHO are likely not accurate even when using baseline subtraction By controlling the temperature of the photomultiplier tube and the outdoor air monitoring shelter, or by using a photodiode detector, these drifts could be eliminated. The use of a span solution allows the user of the analyzer to check the response of the instrument on a daily basis. Ho wever, the use of a liquid standard is inconvenient and the use of 37 % formaldehyde may prove to be a difficult method of producing a standard solution. Formaldehyde can volatilize and escape the solution leading to a gradual decrease in concentration whi le concentrated solutions may lead to polymerization of the formaldehyde molecules An alternative method that may be used is the catalytic conversion of methanol into formaldehyde. Some studies have been conducted in the conversion of methane and methanol to formaldehyde with the use of a

PAGE 64

55 molybdenum catalyst. 22, 23 Using the catalytic conversion method would make the purchase of gas mixtures containing methane and easy alternative for the calibration of the gas analyzer.

PAGE 65

56 R eferences 1. Walker, F. Early history of acetaldehyde and formaldehyde, A chapter in the history of orcanic chemistry. J. Chem. Educ. 1933, 10 (9), 546. 2. Salthammer, T. M. S.; Marutzky, R. Formladehyde in the Indoor Environment. Chem. Rev. 2010, 110 (4), 2536. 3. Gibson, J. E., Ed. Formaldehyde Toxicity; Hemisphere Publishing Corporation: Bristol, 1983. 4. Committe on Toxicology. Formaldehyde An Assessment of Its Health Effects; Consumer Product Safety Comission: Washingtion, D.C., 1980. 5. U.S. Department of Health and Human Services. Toxicological Profile for Formaldehyde July 1999; Agency for Toxic Substances and Disease Registry: Atlanta, 1999. 6. Seco, R.; Penuelas, J.; Filella, I. Short chain oxygenated VOCs: Emission and uptake by plants and atmospheric sources, sinks and concentrations. Atmos. Env. 2007, 41 (12), 2477. 7. Atk inston, R. Atmospheric chemistry of VOCs and NOx. Atmos. Env. 2000, 34 (12 14), 20632101. 8. United States Department of Labor, Formaldehyde. 1910.1048. http://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=STAN DARDS&p_id=10075 (accessed Januar y 7, 2012).

PAGE 66

57 9. Center for Environmen tal Research Information, O. R. D. Compendium Method TO 11A, Determination of Formaldehyde in Ambient Air Using Adsorbent Cartridge Followed by High Performance Liquid Chromatography (HPLC) [Active Sampling Methodology]; Environmental Protection Agency: Cincinatti, 1999. 10. Zhou, X.; Huang, G.; Civerolo, K.; Schwab, J. Measurement of Atmospheric Hydroxyacetone, Glycolaldehyde, and Formaldehyde. Environ. Sci. Technol. 2009, 43 (8), 27532759. 11. Komazaki, Y.; Narita, Y.; Tanaka, S. Development of an automated measurement system using a diffusion scrubber and highperformance liquid chromatography for the monitoring of formaldehyde and acetaldehyde in automotive exhaoust gas. Analyst 1998, 123 (11) 23432349. 12. Nash, T. The Colorimetric Estimation of Formaldehyde by Means of the Hantzsch Reaction. Biochem. J. 1953, 421 (55), 416. 13. Hantzsch synthesis. http://www.organic reaction.com/organic reaction/multi component reactions/hantzschsynthesi s/ (accessed October 15, 2011). 14. Dasgupta, P. K.; Dong, S.; Hwang, H. Diffusion Scrubber Based Field Measurements of Atmospheric Formaldehyde and Hydrogen Peroxide. Aerosol Sci. and Tech. 1990, 12 (1), 98 104.

PAGE 67

58 15. Dasgupta, P. K.; Dong, S.; Hwang, H .; Yang, H.; Genfa, Z. Continuous Liquid Phase Fluoromentry Coupled to a Diffusion Scrubber for the Real Time Determination of Atmospheric Formaldehyde, Hydrogen Peroxide and Sulfur Dioxide. Atmos. Env. 1987, 22 (5), 949 963. 16. Fan, Q.; Dasgupta, P. K. Continuous Automated Determination of Atmospheric Formaldehyde at the Parts Per Trillion Level. Anal. Chem. 1994, 66 (4), 551 556. 17. Li, J.; Dasgupta, P. K.; Genfa, Z.; Hutterli, M. A. Measurement of Atmospheric Formaldehyde Wit h a Diffusion Scrubberand Light Emitting Diode Liquid Core Waveguide Based Fluorometry. Field Anal. Chem. and Tech. 2001, 5 (1 2), 2 12. 18. Compton, B. J.; Purdy, W. C. Fluoral P, A Member of a Selective Family of Reagents for Aldehydes. Analytica Chimi ca Acta 1980, 119 (2), 349 357. 19. Pinheiro, H. L. C.; de Andrade, M. V.; de Pereira, P. A.; de Andrade, J. B. Spectrofluorimentric determinatino of formaldehyde in air after collection onto silica cartridges coated with Fluoral P. Microchem. J. 2004, 78 (1), 45 20.

PAGE 68

59 20. Formaldehyde Permeation Through Nafion. http://www.permapure.com/tech notes/key concepts/formaldehyde permeation through nafion/ (accessed October 16, 2011). 21. FAQs. http://www.permapure.com/support/faqs/ (accessed October 16, 2011). 22. Kennedy, M. Selective oxidation of methane to formald ehyde: comparison of the role of promoters in hydrocarbon rich and lean conditions. Catalysis Today 1992, 13 (2 3), 447454. 23. Solomon, S. J.; Custer, T.; Schade, G.; Soares Dias, A. P.; Burrows J. Atmospheric methanol measurement sing selective catalytic methonol to formaldehyde conversion. Atmospheric Chemistry and Physics Discussions 2005, 5, 35333559.