 

. A

(“I-tom“

,. .2

. {ﬁnal at“

5
.3
.

53.0.3.0“.

“$.me
as

 

 

I. innv:

:mﬂ. ‘u...,v;.nu.; .

 

This is to certify that the

dissertation entitled

Characterization of Odorous and Hazardous Gaseous
Compounds in Livestock Building Air

presented by

Hyesoon Kim-Yang

has been accepted towards fulﬁllment
of the requirements for

PhD degree in Agricultural Engineering &
Civil & Environmental Engineering

ﬂaw
/$¢4w a W

jor professor

 

 

 

M5 U is an Afﬁrmative Action/Equal Opportunity Institution 0- 12771

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6IO1 c:/CIRC/DateDue.p65-p.15

 

CHARACTERIZATION OF ODOROUS AND HAZARDOUS
GASEOUS COMPOUNDS IN LIVESTOCK BUILDING AIR

By

Hyesoon Kim-Yan g

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Agricultural Engineering
Department of Civil and Environmental Engineering

2002

ABSTRACT
CHARACTERIZATION OF ODOROUS AND HAZARDOUS GASEOUS
COMPOUNDS IN LIVESTOCK BUILDING AIR
By

Hyesoon Kim-Yang

This study was conducted to develop methods for the characterization of
odorous compounds in swine building air and to apply the methods developed to
study the effect of ozone on odors in swine facilities.

Several methods for the determination of odorous compounds in swine
building air were investigated. Conventional air sampling methods, such as sorbent
trapping and impingers, and a Solid Phase Microextraction (SPME) method for the
sampling VOCS in swine building air were studied. Two methods for the
quantiﬁcation of phenolic and indolic compounds (phenol, p-cresol, p-ethylphenol,
indole and Skatole) by SPME were investigated. These methods were based upon
equilibrium theory and diffusion theory.

Without any further sample concentration, solvent desorption sorbent traps
and impinger traps were not able to determine any volatile organics from the swine
nursery housing air using a sample volume of up to 36L. Thermal desorption multi-
bed tubes were able to collect volatile and some of semi-volatile compounds from
swine building air, however the recovery of high boiling point compounds was not
good. SPME air sampling method was the most efﬁcient method for the collection of

odorous volatile compounds in swine building air.

The efﬁcacy of a commercial ozonation system to reduce odors was studied.
In this system ozone was distributed throughout the animal housing facility using a
manifold. In this study the VOCS in the air and the odor of the air were monitored.
The VOCS were measured using thermal desorption tubes and SPME ﬁbers with a
PDMS/CAR coating. Ozone was effective in reducing odor detection threshold, but
did not signiﬁcantly reduce the odor intensity or odor offensiveness in the building.
Sensory testing indicated that the characteristics of the air in the ozonated rooms were
different from those in control room. Monitoring of the VOCS present in the air
showed that ozonation reduced the levels of phenolic and indolic compounds in the
swine building air, however, it did not reduce the level of volatile fatty acids in the

air.

Copyright by
HYESOON KIM-YANG

2002

ACKNOWLEDGEMENT

I would like to express my acknowledgement to my academic advisors, Dr.
R.D. von Bemuth in Agricultural Engineering and Dr. Susan J. Masten in Civil and
Environmental Engineering, for their guidance and advice during my academic years
at Michigan State University. I would like to thank Dr. Simon Davies, Dr. Syed
Hashsham, and Dr. Melvin Yokoyama, for serving on my committee. Also I would
like to express my appreciation to Dr. Davies for his intellectual input and guidance on
my thesis. I also wish to thank the waste management research group at MSU (Eric
Cline, Jeff Hill, Ann Morison, Nazrul Islam, and Dr. Nanda) for their help and support
in my research.

I am very grateful for the love and support of my parents, my brother, and my
sister. I cordially thank my husband supporting and believing me more than I do for
myself. And I would like to express my special thanks to my boys, “Sheeha” and

“Inha” who continued to smile even though I was not reading bed times stories.

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

LIST OF TABLES [x
LIST OF FIGURES XI
PROBLEM STATEMENT AND SIGNIFICANCE 1
CHAPTER 1. LITERATURE REVIEW 4
1.1 ODOR CHARACTERIZATION AND QUANTIFICATION IN SWINE
WASTE AND PRODUCTION FACILITY 4
1.1.1 ODOR ............................................................................................................... 4
1.1.2 ODOROUS COMPOUNDS IN LIVESTOCK SLURRIES AND BUILDINGS .................... 5
1.1.3 THE EFFECT OF DUST ON ODOR ....................................................................... 10
1.2. ODOR MEASUREMENT: QUANTIFICATION AND QUALIFICATION 11
1.2.1. SAMPLING OF VOLATILE ORGANICS IN THE AIR .............................................. 1 1
1.2.1.1. Sorbent trap with solvent desorption ........................................................ 13
1.2.1.2. Impinger traps ........................................................................................... 15
1.2.1.3. Sorbent trap with thermal desorption ....................................................... 15
1.2.1.4. Solid Phase Microextraction ..................................................................... 18
1.2.1 ANALYTICAL METHOD FOR THE CHARACTERIZATION OF THE ODOROUS
COMPOUNDS ............................................................................................... 21
1.2.3 ODOR MEASUREMENT: OLFACTOMETRY AND GC/SENSORY .......................... 21
CHAPTER 2. MATERIALS AND METHODS 23
2.1. SAMPLING METHODS 23
2.1.1. SOLVENT DESORPTION SORBENT TUBES ........................................................ 23
2.1.2. SAMPLE PREPARATION .................................................................................. 25
2. 1. 2. 1. Charcoal ................................................................................................... 25
2.1.2.2. Silica gel .................................................................................................... 25
2.1.2.3. XADZ6
2.1.2.4. Tenax ......................................................................................................... 26
2.1.3 SAMPLING USING A MIDGET 1MPINGER ........................................................... 27
2.1.3. SAMPLING ON THERMAL DESORPTION SORBENT TUBES .................................. 28
2.1.4. SAMPLING FOR THE PARTICULATES ................................................................ 28
2.1.6 SAMPLING USING SOLID PHASE MICROEXTRACTION FIBERS .......................... 30
2.2. ANALYTICAL INSTRUMENTATION 31
2.2.1. THERMAL DESORBER ..................................................................................... 31
2.2.1.1. Desorption of VOCsﬁ'om particulates ..................................................... 31

vi

2.2.2. GAS CHROMATOGRAPHY AND MASS SPECTROMETRY .................................

CHAPTER 3. A COMPARISON OF SAMPLING METHODS FOR THE
CHARACTERIZATION VOLATILE ORGANIC COMPOUNDS IN
LIVESTOCK FACILITIES USING GAS CHROMATOGRAPHY -
MASS SPECTROMETRY-

 

 

3.1. INTRODUCTION
3.2. OBJECTIVES

 

3.3. MATERIALS AND METHODS

 

3.3.1. SWINE HOUSING .........................................................................................
3.3.2. SAMPLINGS ................................................................................................
3.2.3. ANALYTICAL INSTRUMENT ........................................................................

3.4. RESULTS

 

3.4.1. SOLVENT DESORPTION SORBENT TRAPS AND 1MPINGER TRAPS ...................
3.3.2. THERMAL DESORPTION SORBENT TRAPS ....................................................
3.3.3. SOLID PHASE MICROEXTRACTION .............................................................

3.4. CONCLUSION AND DISCUSSIONS

 

CHAPTER 4. QUALITATIVE AND QUANTITATIVE ANALYSIS OF

ODOROUS COMPOUNDS IN LIVESTOCK BUILDING WITH SOLID

 

 

 

 

 

 

PHASE MICROEXTRACTION AIR SAMPLING METHOD
4.1 INTRODUCTION
4.1.1 EQUILIBRIUM THEORY ...............................................................................
4.1.2 NON-EQUILIBRIUM THEORY .......................................................................
4.2 HYPOTHESIS
4.3 OBJECTIVES
4.4 MATERIALS AND METHODS
4.4.1 SPME FIBERS ............................................................................................
4.4.2 TEST MIXTURES .........................................................................................
4.4.3 AIR SAMPLING ...........................................................................................
4.4.4 CALIBRATION CURVE .................................................................................
4.4.5 GAS CHROMATOGRAPHY AND MASS SPECTROMETRY .................................
4.5 RESULTS
4.5.1 ADSORPTION TIME PROFILE FOR PHENOLIC COMPOUNDS ON SPME
CAR/PDMS COATING ...........................................................................
4.5.1.1 Relationship between the amount adsorbed on SPME and the sampling
times. .............................................................................................
4.5.2 CONCENTRATION EFFECTS ON SPME ADSORPTION ..................................

31

33
33
34
34

34
34
36

36

36
38
42

47

49
49

52
53

55
56
56

56
56
57
58
58

60

60

60
73

4.5.3 QUANTIFICATION OF SPME ANALYSES USING THE CALIBRATION FACTORS... 81

4.5.4 SPME QUANTIFICATION METHOD WITH DIFFUSION THEORY ......................

vii

85

4.5.5 ANALYSIS OF PHENOLIC AND INDOLIC COMPOUNDS IN SWINE BUILDING AIR

 

 

 

 

 

 

 

 

 

 

 

WITH SPME ............................................................................................... 97
4. 5. 5. 1 Rapid air sampling with SPME CAR/PDMS ............................................ 98
4.5.5.2 Comparison of SPME fiber coatings. ....................................................... 99
4.6 CONCLUSIONS 103
CHAPTER 5. OZONATION EFFECT ON ODOR AND VOLATILE ORGANIC
COMPOUNDS IN SWINE PRODUCTION FACILITY. 106
5.1 INTRODUCTION 106
5.2 HYPOTHESES 108
5.3 OBJECTIVES 108
5.4 MATERIALS AND METHODS 108
5.4.1 SWINE HOUSING ......................................................................... 108
5.4.2 OZONE DISTRIBUTION .................................................................................. 109
5.4.3 AIR SAMPLING ............................................................................................. 1 12
5.4.4 TEST MIXTURE ............................................................................................. l 13
5.4.5 QUANTIFICATION OF THE COMPOUNDS ......................................................... 1 13
5.4.6 DESORPTION METHODS ................................................................................ 1 13
5.4.7 GAS CHROMATOGRAPHY AND MASS SPECTROMETRY ................................... 114
5.4.8 ODOR MEASUREMENT: OLFACTOMETRY ...................................................... 1 14
5.5 RESULTS 116
5.5.1 AIR VENTILATION DATA ............................................................................... 1 16
5.5.2 TEMPERATURE ............................................................................................. 1 17
5.5.3 OLFACTOMETRY DATA ................................................................................. 1 19
5. 5. 3. I Dilution threshold (DT) .......................................................................... I 19
5.5.3.2 Odor oﬂensiveness and intensity ............................................................ I 19
5. 5. 3. 3 Odor characteristics ............................................................................... 121
5. 5. 3. 4 Olfactometrjy summary ............................................................................ 122
5.5.4 OZONATION EFFECT ON VOLATILE ORGANIC COMPOUNDS IN SWINE
PRODUCTION UNIT. .................................................................................. 124
5. 5. 4. I SPME air sampling ................................................................................. 124
5.5.4.2 VOCs in Thermal Desorption Tubes ....................................................... 135
5.5.4.3 Volatile Organic Compounds in Dust Samples ...................................... 140
5.5.5 COMPARISON OF SAMPLING METHODS FOR DETECTION OF VOCS IN SWINE
BUILDING AIR. .......................................................................................... 140
5.5.6 TEMPERATURE EFFECT ON ODOR AND VOCS ............................................... 141
5.6 CONCLUSIONS 143
CHAPTER 6. CONCLUSIONS AND RECOMMENDATIONS 145
6.1. CONCLUSIONS 145
6.2 RECOMMENDATIONS FOR FUTURE RESEARCH 149
REFERENCES - 150

 

viii

LIST OF TABLES

Hyesoon Kim-Yang ........................................................................................................ i
Table 1-1. Compounds found in air and a slurry from a swine facility. ......................... 7
Table 1-2. The chemicals founds in air and a slurry from a swine facility ..................... 8
Table 1-3. Compounds found in swine wastes and gases ............................................... 9

Table 1-4. Multi-bed sorbent materials and sampling methods used in EPA TO-17. ..18

Table 2-1. Sorbent tubes used for the sampling of VOCS. ........................................... 23

Table 2-2. Sampling ﬂow rates and volumes used for the collection of VOCS on
solvent desorption sorbent tubes. .......................................................................... 24

Table 2-3. SPME Fiber coatings used for the study ..................................................... 30

Table 3-1. Summary of sorbent tubes and desorption solvents chosen for the collection
of various VOCS in livesotk building air. ............................................................. 35

Table 3-2. Liquid trap solutions for impinger air sampling method for the collection of
VOCS in the livestock building air ........................................................................ 35

Table 3-3. Compounds found in swine house using the thermal desorption method
(Tenax TA and Carbosieve SIII) ........................................................................... 40

Table 3-4. Compounds found in swine house using the thermal desorption method ...41

Table 3-5. The compounds found in swine housing air using SPME ﬁber with a

PDMS coating. ...................................................................................................... 44
Table 3-6. The compounds found in swine housing air using SPME ﬁber with a

PDMS/CAR coating .............................................................................................. 44
Table 3-7. The compounds found in swine housing air using SPME ﬁber with a

PDMS/DVB coating. ............................................................................................ 45
Table 3-8. Compounds found in TD tubes and SPME ﬁbers. ...................................... 46

Table 4-1. Stock solution of test mixture used for the SPME air sampling methods.
(Total volume was 25 mL and the solvent was methanol.) .................................. 57

Table 4-2. The amount recovered and Calibration factors for phenolic and indolic
compounds with SPME CAR/PDMS ﬁber for 30 minutes sampling time. .......... 84

ix

Table 4-3. GC response of test mixture, RF, and K' values for the quantiﬁcation of
phenolic compounds with 30 minutes sampling time. .......................................... 85

Table 4-4. Comparison of actual gas phase concentration to the calculated
concentration using RF and K' values for 30 min adsoprtion. ............................ 85

Table 4-5. Estimation of Diffusion Volumes of the constituents of organic molecules
............................................................................................................................... 86

Table 4—6. Molecular diffusivity and molar volume for phenolic and indolic
compounds ............................................................................................................ 87

Table 4-7. Dimensionless average k values for phenolic and indolic compounds for
SPME CAR/PDMS ﬁber coating .......................................................................... 88

Table 4-8. Comparison of actual gas concentration and the calculated gas
concentration. ........................................................................................................ 90

Table 4-9. Amount of VFAS adsorbed on SPME ﬁbers for 1 minute and 10 minute
sampling using PDMS/CAR ﬁber coating. ........................................................... 98

Table 4-10. Quantiﬁcation of the phenolic and indolic compounds in swine building
air with rapid SPME air sampling method ............................................................ 98

Table 4-11. Comparison of the adsorbed amounts of phenolic and indolic compounds
on two different SPME ﬁbers. (Sampling time for both was 10 min. Unit for the

amount adsorbed is ng.) ........................................................................................ 99
Table 4-12. Concentration of VFAS in swine building air sampled on SPME ﬁbers.

(V FA concentrations are in ug as carbon) .......................................................... 100
Table 5-1. Detection Threshold for air samples in swine building. ............................ 119
Table 5-2. Odor characteristics for air samples in swine housing air. ........................ 123

Table 5-3. Volatile organic compounds found in the control room using TD tubes. .138

Table 5-4. Volatile organic compounds found in ozonated room (6&3) using TD tube.
............................................................................................................................. 139

Table 5-5. Temperature, DT, and VOCS in swine building air ................................... 142

LIST OF FIGURES

Figure 2-1. Sampling train for ORBO tubes ................................................................. 24
Figure 2-2. Sampling diagram for the impingers .......................................................... 27
Figure 2-3. Construction of the thermal desorption tube .............................................. 29
Figure 24. Sampling diagram for the thermal desorption tube .................................... 29

Figure 3-1. Chromatograms for air samples which were collected on charcoal tubes. 37

Figure 3-2. Chromatogram for air sample which were sorbed on thermal desorption
tubes. ..................................................................................................................... 39

Figure 3-3. Chromatogram for air sample which were extracted by SPME ﬁbers ....... 43
Figure 4-1. SPME ﬁeld sampler. .................................................................................. 50
Figure 4-2. SPME air sampling system. ....................................................................... 59

Figure 4-3. Adsorption time proﬁle for phenolic compounds using a static adsorption
conditions using SPME CAR/PDMS. (The standard contained 4.3, 4.1, 4.1, 4.1,
4.1 ug/L of phenol, p-cresol, p-ethylphenol, indole, and Skatole, respectively.)
The average temperature was 23°C and pressure was 1 atm. ............................... 62

Figure 4-4. The relationship between the sorbed mass and sorption time on SPME for
Phenol (4.3 ug/L). The upper and lower lines represent the 95% conﬁdence
interval. The average temperature was 24°C and pressure was 1 atm. ................. 63

Figure 4-5. The relationship between the sorbed mass and sorption time on SPME for
p-Cresol (4.1 ug/L). The upper and lower lines represent the 95% conﬁdence
interval. The average temperature was 23°C and pressure was 1 atm. ................. 64

Figure 4-6. The relationship between the sorbed mass and sorption time on SPME for
p-Ethylphenol (4.1 ug/L). The upper and lower lines represent the 95%
conﬁdence interval. The average temperature was 23°C and pressure was 1 atm.
............................................................................................................................... 65

Figure 4-7. The relationship between the sorbed mass and sorption time on SPME for
Indole (4.1 ug/L). The upper and lower lines represent the 95% conﬁdence
interval. The average temperature was 23°C and pressure was 1 atm. ................. 66

xi

Figure 4-8. The relationship between the sorbed mass and sorption time on SPME for
Skatole (4.1 pg/L). The upper and lower lines represent the 95% conﬁdence
interval. The average temperature was 23°C and pressure was 1 atm. ................. 67

Figure 4-9. The relationship between the sorbed mass and sorption time on SPME for
Phenol at lower concentration (0.43 jig/L). The upper and lower lines represent
the 95% conﬁdence interval. The average temperature was 23°C and pressure
was 1 atm. ............................................................................................................. 68

Figure 4-10. The relationship between the sorbed mass and sorption time on SPME for
p-Cresol at lower concentration (0.41 ug/L). The upper and lower lines represent

the 95% conﬁdence interval. The average temperature was 23°C and pressure
was 1 atm. ............................................................................................................. 69

Figure 4-11. The relationship between the sorbed mass and sorption time on SPME for
p-Ethylphenol at lower concentration (0.41 ug/L). The upper and lower lines
represent the 95% conﬁdence interval. The average temperature was 23°C and
pressure was 1 atm. ............................................................................................... 70

Figure 4-12. The relationship between the sorbed mass and sorption time on SPME for
Indole at lower concentration (0.41 jig/L). The upper and lower lines represent
the 95% conﬁdence interval. The average temperature was 23°C and pressure
was 1 atm. ............................................................................................................. 71

Figure 4-13. The relationship between the sorbed mass and sorption time on SPME for
Skatole at lower concentration (0.41 ug/L). The upper and lower lines represent
the 95% conﬁdence interval. The average temperature was 23°C and pressure
was 1 atm. ............................................................................................................. 72

Figure 4-14. Mass injected in the gas sampling bottle vs mass recovered on SPME
ﬁber for Phenol with sampling time of 30 minutes. The upper and lower lines
represent the 95% conﬁdence interval. The average temperature was 23°C and
pressure was 1 atm. ........................ . ....................................................................... 7 6

Figure 4-15. Mass injected in the gas sampling bottle vs mass recovered on SPME
ﬁber for p-Cresol with sampling time of 30 minutes. The upper and lower lines
represent the 95% conﬁdence interval. The average temperature was 23°C and
pressure was 1 atm. ............................................................................................... 77

Figure 4-16. Mass injected in the gas sampling bottle vs mass recovered on SPME
ﬁber for p-Ethylphenol with sampling time of 30 minutes. The upper and lower
lines represent the 95% conﬁdence interval. The average temperature was 23°C
and pressure was 1 atm. ........................................................................................ 78

Figure 4-17. Mass injected in the gas sampling bottle vs mass recovered on SPME
ﬁber for Indole with sampling time of 30 minutes. The upper and lower lines

xii

represent the 95% conﬁdence interval. The average temperature was 23°C and
pressure was 1 atm. ............................................................................................... 79

Figure 4-18. Mass injected in the gas sampling bottle vs mass recovered on SPME
ﬁber for Skatole with sampling time of 30 minutes. The upper and lower lines
represent the 95% conﬁdence interval. The average temperature was 23°C and
pressure was 1 atm. ............................................................................................... 80

Figure 4-19. Effect of molecular weight of chemicals on SPME initial sorption rate.
Initial sorption rates were calculated using 5 minute sampling data. ................... 81

Figure 4-20. The relationship between the sorbed mass and sorption time on SPME for
rapid diffusion based sampling for Phenol. The upper and lower lines represent
the 95% conﬁdence interval. The average temperature was 23°C and pressure
was 1 atm. ............................................................................................................. 91

Figure 4-21. The relationship between the sorbed mass and sorption time on SPME for
rapid diffusion based sampling for p—Cresol. The upper and lower lines represent
the 95% conﬁdence interval. The average temperature was 23°C and pressure
was 1 atm. ............................................................................................................. 92

Figure 4-22. The relationship between the sorbed mass and sorption time on SPME for
rapid diffusion based sampling for p—Ethylphenol. The upper and lower lines
represent the 95% conﬁdence interval. The average temperature was 23°C and
pressure was 1 atm. ............................................................................................... 93

Figure 4-23. The relationship between the sorbed mass and sorption time on SPME for
rapid diffusion based sampling for Indole. The upper and lower lines represent
the 95% conﬁdence interval. The average temperature was 23°C and pressure
was 1 atm. ............................................................................................................. 94

Figure 4-24. The relationship between the sorbed mass and sorption time on SPME for
rapid diffusion based sampling for Skatole. The upper and lower lines represent
the 95% conﬁdence interval. The average temperature was 23°C and pressure
was 1 atm. ............................................................................................................. 95

Figure 4-25. Effect of molecular weight of chemicals on SPME sorption rate.
Sampling times were 5 seconds to 2.5 minutes. The average temperature was
23°C and pressure was 1 atm. ............................................................................... 96

Figure 4-26. Chromatograms Of air samples sampled on CAR/PDMS ﬁber coatings for
1 and 10 minutes in swine building. ................................................................... 101

Figure 4-26. Chromatograms of air samples sampled on two different SPME ﬁber
coatings. (Sampling time was 10 min for both ﬁbers and sampling was done at
the same location.) .............................................................................................. 102

xiii

Figure 5-1. Environmental chambers at Michigan State University ........................... 110
Figure 5-2. Diagram Of one pen in the environmental chamber at MSU. .................. 111

Figure 5-3. Air ventilation rate in the swine building. The data is shown for the east
and west fans. East fan controls the air exchange in rooms 1 and 2 and west fan

controls air exchange in rooms 3 and 4 ............................................................... 116
Figure 5-4. Temperature changes in the environmental chambers. ............................ 117
Figure 5-5. Temperature and ventilation rate in swine building (Room 1&2). .......... 118

Figure 5-6. Effect of ozonation on odor offensiveness in swine building air. Samples
were taken at noon. ............................................................................................. 120

Figure 5-7. Effect of ozonation on odor intensity in swine building air. Samples were
taken at noon. ...................................................................................................... 121

Figure 5- 8. The total VFA concentrations adsorbed on SPME ﬁbers for weekly
samples. ............................................................................................................... 125

Figure 5-9. The average VFA concentration adsorbed on SPME ﬁbers for bi-weekly
samples. ............................................................................................................... 125

Figure 5-10. Average Phenol level in bi-weekly air samples. (Sampling method was
SPME) ................................................................................................................. 127

Figure 5-11. Phenol level in weekly air samples. (Sampling method was SPME) 127

Figure 5-12. Diurnal variations in phenol level. (Sampling was done in July 19‘"

2001) .................................................................................................................. 128
Figure 5-13. Diurnal variations in phenol level. (Sampling was done in August 2"d

2001.) .................................................................................................................. 128
Figure 5-14. Average p-Cresol level in bi-weekly samples. (sampling method was

SPME) ................................................................................................................. 130
Figure 5-15. p-Cresol level in weekly samples. (Sampling method was SPME) ....... 131

Figure 5-16. Diurnal variations in p-Cresol level. (Sampling was done in July 19th
2001) ................................................................................................................... 131

Figure 5-17. Effect of samplings times on p-Cresol level. (Sampling method was
SPME and the a bar represents the averaged p-cresol concentration at the certain
sampling time) .................................................................................................... 132

xiv

Figure 5-18. Effect of light on the average p-Cresol concentration. (Sampling method
was SPME. Concentrations of p-cresol that were measured when the light was on
and Off were averaged.) ....................................................................................... 132

Figure 5-19. Average Indole levels in bi-weekly air samples. (Sampling method was

SPME). ................................................................................................................ 134
Figure 5-20. Average Skatole level in bi-weekly air samples. (Sampling method was

SPME). ................................................................................................................ 134
Figure 5-21. Level Of lndolic compounds in weekly SPME samples. (Sampling

method was SPME) ............................................................................................. 135
Figure 5-22. GC/MS Chromatogram for TD air samples. ........................................... 137

XV

PROBLEM STATEMENT AND SIGNIFICANCE

Modern swine production managements have changed to the larger populations
Of animals on production sites in an effort to improve animal production efﬁciency.
These expansions of concentrated animal feeding operations (CAFO) have increased
awareness by the public and government agencies for the potential impacts of these
facilities on water and air quality. In spite of the increasing demands for Odor and air
quality control in livestock production facilities, researches related to odors have not
been actively conducted in agricultural area.

A reason is that odorous compounds have not been regulated because odor is
deﬁned as a “nuisance”, not as a “contaminant”, in spite of the presence of toxic
compounds in the livestock buildings, such as ammonia, hydrogen sulﬁde and other
sulﬁde gases (Xue et al., 1998).

Regulation of the odor and/or odorants from the livestock building has been
challenging, because monitoring the odorous compounds in the air of the livestock
building is complicated for a number of reasons.

First, the concentration of the odorants is very low, and often is below the
detection limits. Second, numerous compounds are present in the air of the livestock
buildings. Many researchers have attempted to characterize the Odor by analyzing the
chemical compositions in the air. More than 30 odorous compounds have been found
in the swine house air. Third, the analysis of the odorous compounds is difﬁcult
because of the wide range of physical and chemical properties of the odorous
compounds found in the livestock wastes and air. Some of the compounds from

livestock production units are gases such as hydrogen sulﬁde, ammonia, and methyl

mercaptan. Other compounds found in the livestock facilities, such as indole and
Skatole, have boiling points more than in excess of 250 °C.

The government agencies such as, NIOSH, OSHA, and EPA have reported
several methods for the determination of toxic organic compounds in ambient air.
These methods were designed for the detection of speciﬁc compounds, such as
aldehydes, phenols, or polycyclic aromatic hydrocarbons (PAHS) in the ambient air. A
single method for the sampling of odorants from livestock facilities is not sufﬁcient
for the monitoring various volatile organic compounds (V OCS) in livestock building
air.

Ozone has been widely used for indoor air control in recreational areas, such
as casinos, offices and hotels. A recent study by Oehrl et a1. (2000) showed that ozone
was eﬁ‘ective in decreasing the low molecular volatile fatty acids on the dust samples
from swine house. The effectiveness of ozone for oxidizing odorous volatile
compounds in swine slurry such as; phenol, p-cresol, p-ethylphenol, indole and
Skatole, in swine slurry have been proven (Wu nad Masten, 1999).

Human health effects from polluted livestock air have been a great issue. Most
of gaseous compounds found in the livestock buildings are “hazardous pollutants”
listed in Clean Air Act (CAA). Finding adequate air sampling and analytical
techniques are the key points for successful air characterization in the livestock
buildings.

This research proposes air sampling methods for odorous compounds in
livestock facilities by testing some air sampling methods used by the US government

agencies (NIOSH, OSHA, and EPA). This research also proposes the adaptation of the

new air sampling method, such as solid phase microextraction (SPME), for the
sampling of odorous compounds in livestock building.
This research will identify suitable air sampling and analytical techniques for

compounds in swine houses.

CHAPTER 1. LITERATURE REVIEW

1.1 Odor characterization and quantification in swine waste
and production facility

1.1.1 Odor

Odor is something that stimulates the olfactory systems or sense of smell. The
perception of Odor is due to the presence of volatile compounds in the inhaled air.
Humans are known to be able to detect over 10,000 odors, despite being able to
identify only a few of them. Human nose can detect and discriminate odors at
concentrations even lower than those detectable by analytical techniques, such as gas
chromatography (e.g. hydrogen sulﬁde) (Mackie et al., 1998). The minimum
concentration required to detect an odor is termed the odor threshold values (OTV).
The lowest toxic value (LTV) of these compounds in air may be at least a factor of
500 higher than OTV. In this case the odorous compound is detected long before it
becomes a health risk (Tamminga, 1992).

There is no universally accepted deﬁnition of an Objectionable odor, nor are
there any legally deﬁned conditions under which livestock producers become required
to reduce Odor emissions emanating from the animal facility (NPPC, 1995). No federal
guidelines exist that regulate and control odors in the environment because of the
technical difﬁculties of deﬁning odor limits and their measurement and evaluation. In
contrast to pollutants for which legal limits are set and enforced by federal agencies,
odor is considered a nuisance and complaints are normally handled by state or local

authorities (Mackie et al., 1998).

1.1.2 Odorous compounds in livestock slurries and buildings

An increase in the number of complaints from the public about odors from
pork production units has stimulated interest in the measurement and reduction of
odors at these facilities. Research has been conducted on qualitative and quantitative
chemical characterization of odorants in livestock wastes, especially in swine Slurries
(Yasuhara et al., 1984; Ritter, 1989; Yu et al., 1991; O’Neill and Phillip, 1992; Hobbs
et al., 1995; Zahn et al., 1997; Wu et al., 1998).

Yasuhara et a1. (1984) identiﬁed approximately 30 organic compounds in fresh
and rotten swine manure using vacuum distillation extraction, followed by the gas
chromatography-mass spectrometry (GC/MS). The study suggested that the Signiﬁcant
odorous compounds in the swine manure were alcohols, 3-hydroxy-2-butanone, fatty
acids, aromatic carboxylic acids, phenols, and indoles.

Hammond et al. (1989) found the most important contributors to the odor in
swine waste lagoons were dimethyl disulﬁde and dimethyl trisulﬁde. Indole, Skatole
and cresol made minor contributions to the odor.

O’Neill and Phillips (1992) identiﬁed 168 chemicals in livestock Slurries and
manures, Of which 30 have odor detection thresholds lower than or equal to 0.001
mg/m3. SulﬁIr containing compounds were the major group of the compounds with
low odor detection threshold.

The research by Hobbs et a1 (1999) studied the average daily and annual
emission rates for several gases and odorous compounds from pig Slurry storage.
Volatile fatty acids (C2 to C7), phenolics (phenol, p-cresol, p-ethylphenol) and indolics
(indole and Skatole) and the major by-product gases of the anaerobic storage; carbon

dioxide, methane, ammonia, and hydrogen sulﬁde, were monitored. The authors

mentioned that there was no relationship between the odor emission rate and that of
H2S, and other major odorants, nor was there any relationship between the Odor
concentration emission rate and the concentration of individual odorants in the slurry
(Hobbs et al., 1999). This research suggests that the odor is not simply correlated to
the individual odorants in the Slurry.

O’Neill and Phillips (1992) compiled a list of the chemical compounds in
livestock air and manure. The results of twelve studies were considered. The most
frequently reported compounds were phenol, p-cresol, ammonia and the volatile fatty
acids. They noted that nearly all workers have used gas chromatography (GC) for the
odor analysis, even though this technique does not guarantee the identiﬁcation of
every compound present. The paper Showed that the reported concentrations in air of
the compounds Showed wide variations. This review paper also Showed the wide
variations in the chemicals identiﬁed in air, depending upon the sampling and
analytical methods used, suggesting that no Single method is suitable for the
identiﬁcation of all chemicals in air.

Hobbs et al. (1995) investigated the odorous compounds in the air and Slurries
from livestock facilities. Silica and carbon based adsorbents were used for the
sampling of air from the pig house prior to analysis by GC/MS. The results showed
that the two different sampling methods (sorbent tube and liquid extraction of the
Slurry) gave different chemical proﬁles. Table 1-1 shows the chemicals that Hobbs et

al. (1995) found in the air and swine Slurries.

Table 1-1. Compounds found in air and a slurry from a swine facility.

 

 

Air Slurry
(Silica sorbent - orbo 52) (Liquid extraction with diethyl ether)
acetic acid 3-methyl butanoic acid

2-methyl butanoic acid

Undecane Phenol
4-methyl phenol 4-methyl phenol
4-ethylphenol 4-ethyl phenol

Indole
Skatole Skatole

2,3—dihydro-4-methyl 1H indole
2,3—dihydro-indole carboxyaldehyde

 

Source: Hobbs et al., 1995.

Zahn et al. (1997) studied the chemical composition of the swine Slurry,
sludge, and air in a swine facility. They used a multi-bed sorbent trap of Tenax TA
and Carbotrap C or Carboxen-569 for the sampling of air. Thermal desorption
followed by GC/MS was used to identify the trapped analytes. Solid Phase Micro
Extraction (SPME) and conventional liquid extraction method were used to analyze
the Slurry samples. The results found for the air and slurry samples were different (see
Table 1.2) and the compounds found were dependent upon the sampling procedures.
This research showed that indole and Skatole were present in high concentrations in
the slurry, but rarely detected in the airborne emissions from the slurry basin. The
study also showed that several compounds, including trimethyl dihydroindene and
butylated hydroxytoluene, were present in air samples, but never detected in slurry
samples. This research suggests that a number of compounds are concentrated at the
air-solution interface (Zahn et al., 1997). Table 1.3 summarizes the chemicals found in

air and swine Slurries based on the literature previously cited.

Table 1-2. The chemicals founds in air and a slurry from a swine facility.

 

 

 

 

 

Air Slurry
(T enax TA + Carbotrap. thermal (liquid extraction, SPME liquid, SPME head
desorption) space)
VFAs VFAs
acetic acid acetic acid
propanoic acid propanoic acid
butanoic acid butanoic acid
pentanoic acid pentanoic acid
hexanoic acid hexanoic acid
octanoic acid octanoic acid
nonanoic acid nonanoic acid
benzoic acid benzoic acid
dodecanoic acid
Phenols Phenols
phenol phenol
4-ethylphenol 4-methylphenol
4-(1 ,1-dimethylpropyl) phenol 4-ethylphenol
4-(1,1-dimethylpropyl) phenol
Others Others
2-butanol 2-butanol
1-decanol 1-decanol
1-dodecanol 1-dodecanol
1-bromodecane benzyl alcohol
butylated hydroxytoluene 1-bromodecane
1 ,1 ,3-trimethyl-dihydroindene butylated hydroxytoluene

hexadecanoic acid methyl ester
2-amino acetophenone
bis (2-ethylhexyl) phthalate

 

1 ,1 ,3-trimethyl-dihyroindene
hexadecanoic acid methyl ester
2-amino acetophenone

bis (2-ethylhexyl) phthalate

 

Source: Zahn et al., 1997

 

Table 1-3. Compounds found in swine wastes and gases

 

 

Carboxylic acids Alcohols Aldehydes & ketones N itrtgen heterochles
formic acid methanol formaldehyde indole

acetic acid ethanol acetaldehyde Skatole

propanoic acid n-propanol propionaldehyde pyridine

n-butyric acid n-butanol acrolein 3-aminopyridine
I-butyric acid 2-butanol butylaldehyde 2-methylpyrazine
n-valeric acid 2-methyl-1-propanol iso-butyraldehyde methylpyrazine
I-valeric acid 3-methylbutanol 2-butenal trimethylpyrazine
2-methylbutanoic acid hex-3-ene-1-ol pentanal tetramethylpyrazine

2-methyl-2-butenoic
acid

2-methy12-pentanol

3-methylbutanal

 

 

 

 

2,3-dihydro-4-methyl-l H-
indole

 

 

 

 

 

 

 

n-caproic acid l-heptanol hexanal
2-methylpentanoic acid I-heptanol 2-heptenal Others
heptanoic acid 2-ethylhexanol octanal methane
octanoic acid 2-methoxyethanol nonanal ammonia
decanoic acid 2-ethoxy-l-propanol 2-nonenal hydrogen sulﬁde
undecanoic acid 2,3-butanediol 2,4-nonadienal sulfur dioxide
dodecanoic acid 2-phenylethanol decanal pentane
tn'decanoic acid l-decanol 2,4-decadienal hexane
tetradecanoic acid l-dodecanol benzaldehyde octane
benzenecarboxylic acid Benzyl alcohol 2-propanone benzene
phenylethanoic acid 2,3-butanedione toluene
phenylpropanoic acid methylethylketone xylene
hydrocinnamic acid Amines 3-hydroxy-2-butanone indane

methylamine diethylketone propione naphthalene
Esters ethylamine cyclopetanone methylnaphthalene
methylfonnate n-propylamine 2-methyl cyclopentanone chloroform
methylacetate iso-propylamine 2-octanone hydrazine
ethylformate pentylamine amylvinylketone 2-methylfuran
ethyl acetate trimethylamine acetophenone 2-pentylfuran
propylacetate triethylamine 2-methylthiophene
I-propylacetate Phenolics 2,4-dimethylthiophene
butylacetate Sulﬁdes Phenol diethyl ether
l-butylacetate carbon disulﬁde p-cresol l-bromodecane
I-propylpropionate carbonylsulﬁde m-cresol butylated hydroxytoluene
hexadecanoic acid 1,1,3-trimethyl-
methyl ester dimethylsulﬁde o—cresol dihydroindene

diethylsulﬁde o-methoxyphenol 2-amino acetophenone
Thlols dimethyldisulﬁde 2.6-dimethylphenol bis(2-cthylhexyl) phthalate
methanethiol dimethyltrisulﬁde 3,4-dimethylphenol benzisothiazole
mropanethiol diethyldisulﬁde p-ethylphenol benzthiazole

p—(l ,l- 2,3-dihydro-indole-

2-propanethiol dipropyldisulﬁde dimethylpropyl)phenol carboxyaldehyde

diphenylsulﬁde

 

1.1.3 The effect of dust on odor

The effect of dust on odors in swine houses has been investigated (Hammond
et a1, 1979). Many swine production facilities use high ventilation rates to remove the
odorants in livestock house resulting in high dust emissions. Dust in livestock
buildings is mainly composed of organic material originating from feed, skin, and
bedding and dried feces (Harry, 1978). Dust can adsorb and concentrate the odorants.
Hammond et el. (1979) showed that odorous compounds in livestock building air were
adsorbed onto particles of dust, as well as being molecularly dispersed. In this report,
nineteen compounds were identiﬁed in the dust collected from air in a swine housing
facility. Aldehydes, acids, phenolics, Skatole, and sulfur compounds were identiﬁed.
The authors also conducted odor tests with a panel Of eight observers by exposing
them to the ﬁltered air and unﬁltered air. The ﬁltered air was odorless, while the
unﬁltered air had an intense odor (Hammond et al., 1979). This result supports the role
of dust in the odor of swine facilities.

Oehrl et al. (2000) conducted an analysis of odorous compounds in swine
house dust using gas chromatography. Dust samples were taken from the exhaust fans
of the swine house and extracted with methanol at 60 °C and analyzed by the GC. Ten
of volatile fatty acids (C2, C3, C4, iso-C4, C5, iso-Cs, C6, C7, C3, and C9) and p-cresol
were found in the dust samples.

Despite efforts to establish a relationship between dust and odor intensity, no
relationship has been determined. Williams (1989) tried to correlate the dust particle
numbers and concentration in terms of mass and surface area of dust particles with

odor concentration, however it was not successful.

10

In a review paper, Carpenter (1986) quoted unpublished data claiming that
particles in 5-20 pm diameter size range are the mainly responsible for transporting
Odors. However, the threshold of dust or particle concentration below which odor
would be effectively eliminated was not indicated. This paper also Showed that the
complexity of particle penetration and deposition in the respiratory tracts make it
difﬁcult to deﬁne the dust size ranges where particles may be trapped in the nasal
passage. No sharp demarcation exists, but it is generally accepted that the particles
with diameter > 10 um are deposited in the nasal passages, particles with a diameter 5
mm um are deposited in the upper respiratory tract and particles with a diameter < 5
pm are deposited in the lungs themselves (Hayter and Besch, 1974).

The concentration of particles and implication of dust on human health has
been reported (Donharn, 1986). The threshold limit value (TLV) for inert mineral total
dust for humans is 10 mg/m3, and for respirable dust, 5 mg/m3 for 8-hour worker
exposure (ACGIH, 1993). According to Maghirang et al. (1997), the overall mean
total dust concentration in an enclosed swine nursery was 0.72 mg/m3, and the dust
concentration ranged from 0.12 to 2.14 mg/m3. The overall mean respirable fraction

was 11% with a range from 2 to 30% (Maghirang et al., 1997).

1.2. Odor measurement: quantification and qualiﬁcation
1.2.1. Sampling of volatile organics in the air

As described above, the chemical composition of the volatiles in livestock
facilities is variable and the results found are highly dependent on the analytical and/or

sampling methods. The analysis of odorous compounds in the air found in the

11

livestock facilities is particularly challenging, as a wide range of volatile compounds
are found in this environment.

The Environmental Protection Agency (EPA) has compiled methods for the
determination of toxic organic compounds in ambient air. There are 17 compendium
methods in the EPA for the determination of toxic organic compounds in ambient air
(Air Toxic Methods, 2001).

Harper (2000) has summarized information for the sorbent trapping of volatile
organic compounds from air. This article has information for the trapping methods
(sorbent trapping and reaction trapping) and desorption methods (solvent desorption,
thermal desorption) for various volatile compounds in air. The sorbents, such as silica
gel, activated charcoal, Anasorb 747, carboxens, porous polymers and carbon
molecular sieves are discussed.

The Solid Phase Microextraction (SPME) method has been Widely used for the
determination of volatile organics in foods and for ﬂavor analysis (Yang and Peppard,
1994; Ng et al., 1996; Clark and Bunch, 1997; Song et al., 1997). SPME methods have
been used mostly for the sampling of volatiles in liquid phases. A recent study showed
that this method was able to detect benzene, toluene, ethylbenzene and xylene (BTEX)
compounds in the air with a relatively short sampling period (Koziel and Pawliszyn,
2000).

Based on a review of the literature, the following sampling methods were
chosen for further evaluation: solvent desorption sorbent tubes, impinger, active multi

bed sorbent tubes and SPME.

12

1.2.1.1. Sorbent trap with solvent desorption

Many NIOSH, OHSA, and EPA methods use the sorbent trapping with solvent
desorption for the determination of the volatile organics from the ambient air. More
than 30 different sorbent tubes are commercially available. Based on a review of the
literature, the following sorbent materials were chosen for evaluation. This choice is
based on whether the sorbent is suitable to adsorb the compounds found in the air of
livestock facility and whether the sorbent material is able to adsorb wide range of
volatile compounds.

1.2.1.1.1. Activated Charcoal

Activated charcoal is commonly used as a sorbent. Coconut shell charcoal and
petroleum charcoals have been most widely used. Coconut shell charcoal is used by
NIOSH for the analysis of acetic acid, aromatic hydrocarbons, ketones, and esters
(NMAM methods 1603, 1501, 1300, and 1450 respectively). It is an inexpensive
adsorbent, however, water uptake is a problem with an activated charcoal. When the
relative humidity is greater than about 50%, water molecules are adsorbed onto the
surface of the charcoal (Harper, 2000). Since the relative humidity in the swine
facilities is often greater than 50%, charcoal may be unsuitable for the air sampling in
these facilities.

1.2.1.1.2. Silica gel

Silica gel has been used for the determination of polar hydrocarbons in air.
Silica gel is used for the detection of aliphatic and aromatic amines (NIOSH, 1994).
Due to the sorption of water, the Silica gel may have a reduced capacity at high

humidity (Harper, 2000).

13

1.2.1.1.3. Carbon Molecular Sieves

Molecular sieves are substances that have discrete pore structures that
discriminate between molecules on the basis of size. Molecules that are small enough
to enter the pores of sieves are retained. Carbon molecular sieves are made from
polymers, such as polyvinylindene and polyacrylonitrile. Molecular sieve materials
such as Carboxieve® 8-111 and Carboxen® 1000 can be used for the trapping of low-
boiling popint compounds (-60 to 80 °) in the C2 through C5 range. Molecular sieves
can be used for the sampling of small volatile molecules, such as methyl ethyl ketone,
Vinyl acetate, methyl chloride, dichrolomethane (NIOSH 1453, 2500 and 2549; Rosell,
1991)

1.2.1.1.4. Porous polymers

Microporous polymers can be used as sorbents. The degree of porosity is
controlled by the amount of cross-linking. Water adsorption on the polymer surface is
not signiﬁcant and the surface is inert. Tenax, Chromosorb 106, Amberlite XAD-4,
and Porapak Q are micoporous polymers. Tenax is useful for the high boiling
compounds (EPA CM TO-l, 1999; Martin et al., 2000; Tsai and Hee, 2000;
Paustenbach et al., 2001). It has been used with thermal desorption, as Tenax is
incompatible with many solvent systems solvent extraction is rarely used for the
recovery of the compounds sorbed on Tenax.

Amberite XADS and Chromosorbs are mesoporous polymers with moderate
surface areas, they are used for the trapping of large, semi volatile molecules such as
polycyclic aromatic hydrocarbons (PAHS) and pesticides. XAD-4 resin is able to

adsorb phenols under acidic conditions, where the molecular phenol species is

14

predominant (Ku and Lee, 2000). XAD-Z adsorbent has been used for the collection
of sulfur compounds in the work place and carbon tetrachloride was used for the
solvent desorption (Suryanayarana et al., 2001). A glass ﬁbre preﬁlter XAD-2
sampling methods was used by Kontsas et al. (1993) for the collection of airborne

chlorophenol.

1.2.1.2. Impinger traps

The midget impinger is another way to trap the VOCS in the ambient air.
NIOSH methods use the midget impingers for the trapping of monomethylamine,
aminoethanol, formaldehyde, hydrazine, and keptone. Before NIOSH changed the
method for phenol and cresol to a method based on sorbent trapping, an impinger trap
with 0.1 N NaOH was used. The major advantage of using impingers for air sampling
is the ease of sample delivery to the analytical instrument. Since the impinger traps the
analytes in solution, the sample can be introduced to the analytical systems, such as a
gas chromatograph, without further sample treatment. A disadvantage is that longer
sampling times are required than for other sampling methods. Normally, a sampling
volume at least 10 times larger than that used sorbent traps is required to obtain

similarsensitivity..

1.2.1.3. Sorbent trap with thermal desorption

A major advantages of the sorbent trap-thermal desorption method over
sorbent trapping-solvent desorption or impinger methods are that:

a) thermal desorber can transfer the entire sample to the analytical instrument,

so lower sampling volume is needed. As the sorption capacity of the

sorbent trap is limited, reduced sampling volumes not only save time, but

15

reduce the chances of interferences due to competitive interactions between
the anlayte and other chemicals present in the air, e. g., water.

b) as the analyte is not dissolved in a solvent in GC the problem of

interference between the solvent and early eluting compounds is avoided.

Multi-bed thermal desorption tubes are useful when the sample contains
compounds with a wide range of boiling points. The ﬁrst bed in the tube is for the
collection of compounds with higher boiling points (semi-volatiles) and the second
bed is for the collection of the more volatile compounds present in the air. In
desorption process, sorbed chemicals are desorbed by heating the tubes. The carrier
gas transfers them to the analytical instrument.

Many types of sorbent materials are used in air analysis. Some of the
commonly used sorbents are discussed below. Tenax TA is a hydrophobic,
macroporous, a semi-crystalline polymer made with diphenyl-p-phenylene oxide. It is
relatively inert, granular and available in a wide range of mesh sizes. Due to the
relatively low surface area (about 15 mz/g), the adsorption capacity is low. However, it
has an ability to have a very low background (less than 1 ng per component) when it is
conditioned and also it is stable at high temperature, so that it allows the recovery of
many semi-volatile compounds (Harper, 2000). It is useful for the collection of the
multiple analytes present at low concentrations. Some of oxidizing gases, such as
ozone and nitrogen oxides, can degrade Tenax.

Chromosorb 106 is very hydrophobic and relatively unaffected by ozone
(Harper, 2000). It has a greater capacity than Tenax but it’s thermal stability is lower

than Tenax, so it is not suitable for analysis of semi—volatile compounds.

16

Carbotrap B and Carbotrap C are graphitized carbons and are made from
carbon black. They do not react with ozone and are relatively hydrophobic and adsorb
little water. Carbotrap B, often known as Carbotrap or Carbopack, has a higher surface
area (100 mzlg) than Carbotrap C (10 m2/g). These sorbents are good for the collection
of low boiling point compounds and are rarely used for a single bed. However they are
commonly used as the last trap in multi-beds.

1.2.1.3.1. M ulti-bed sorbent traps

When the analyst wishes to determine a wide range of analytes, more
than one sorbent materials may be required. Multi-bed thermal desorption tubes
consist of two or three different sorbent materials. The direction of sample ﬂow in the
multi-bed tubes is critical, since the sorbents have different sorption strength. Sample
ﬂow should always be from least sorbing material to material with the maximum
sorbent strength. Recent US government agency methods have used the mixed
sorbents for the air sampling (e.g., EPA Method TO-17, NIOSH Method 2549). The
advantage of a multi-bed tube over the single bed tube is sorbents can be chosen to
collect a broad range of chemical present in the sample. In NIOSH method, a Ms”
stainless steel tube packed with Carbopack Y, Carbopack B, and Carboxene 1003
(Supelco, Inc.) is used. The amount of packing for each sorbents was 90, 115, and 150
mg, respectively. All packing materials Should be 40/60 mesh.

EPA suggests that three different types of multi-bed thermal tubes may be used
for air sampling. The Type 1 tube is packed with 30mm Tenax and 25mm Carbopack
B, the Type 2 tube is packed with 35mm Carbopack B and 10mm of Carbosieve S III

or Carboxene 1000, and the Type 3 tube is packed with 13mm Carbopack C, 25mm

17

Carbopack B, and 13 mm Carbosieve S III or Carboxene 1000. As Shown in Table 1-

4, three types may be used for trapping analytes of different volatility.

Table 1-4. Multi-bed sorbent materials and sampling methods used in EPA TO-l 7.

 

 

 

 

 

Compounds Trap packing Sampling
n-C6 to n-C2o Tenax(30mm) plus Carbopack 2L at any humidity
B(25mm)
n-C3 to n-C.2 Carbopack B(35mm) plus 2L at relative humidity below 65%
Carboxen 1000(10mm) and temperature below 30°C
0.5L at RH above 65%
n-C; to n-C.2 Carbopack B(35mm) plus 2L at relative humidity below 65%
Carbosieve SIII(10mm) and temperature below 30°C
0.5L at RH above 65%
n-C3 to n-Cu, Carbopack C(l3mm) plus 2L at relative humidity below 65%
Carbopack B(25mm) plus and temperature below 30°C
Carboxene 1000(l3mm) 0.5L at RH above 65%

 

1.2.1.4. Solid Phase Microextraction

Solid phase microextraction (SPME) was developed by Pawliszyn’s group
(Arthur and Pawliszyn 1990, Zhang and Pawliszyn 1993). It is a solventless extraction
method from gaseous and liquid phases. This method has been utilized widely for the
food/ﬂavor analysis (Yang and Peppard, 1994; Ng et al., 1996; Clark and Bunch;
1997; Song et al., 1997). It saves time and labor for the sample treatment and also
makes it possible to use aqueous samples for GC/MS without further treatment. The
coated silica ﬁber adsorbs/absorbs the polar and/or non-polar compounds from the
sample and it can be thermally desorbed in the GC injector. The length of the SPME
ﬁber coating is typically 10 mm and the polymer thickness ranges from 7 to 100 um.

Traditional air sampling methods involve drawing air through a sorbent or
impinger trap, followed by solvent or thermal desorption into an analytical instrument.

These methods require the use of expensive equipment and/or the use of toxic solvents

18

for analysis. Since SPME sampling methods does not use solvent, solvent peaks can
be eliminated. Another advantage of the SPME sampling over the conventional air
sampling methods is the re-use of the ﬁber without further cleaning.

Several different types of polymer-coated ﬁbers with different polarities are
commercially available. The following SPME ﬁbers are most popular for air or
headspace sampling applications; poly dimethylsiloxane (PDMS), PDMS/Carboxen
(CAR/PDMS), and PDMS/divinylbenzene (PDMS/DVB).

The partitioning of the analytes of interest between the sample matrix and the
polymer ﬁlm depends on the properties of the SPME coating (Louch et al., 1992).
Each coating material has a different selectivity for a particular compounds. The
selectivity of these coating materials for speciﬁc volatile compounds in the air has not
been well documented, however, suitable coatings may be selected based upon the
general properties of the anlaytes.

The PDMS coating is a nonporous, amorphous polymeric phase, so the analyte
uptake in PDMS is via absorption. The PDMS coating is the one of the most widely
used coatings for extracting volatile analytes from environmental samples (Koziel et
al., 2000). The theory behind the equilibrium and non-equilibrium extraction process
for absorptive PDMS coatings is well described (Wercinski and Pawliszyn, 1999).

PDMS and Divinylbenzene (PDMS/DVB) and PDMS and carboxen
(Carboxen/PDMS) are porous solid-phase ﬁbers. The Pores in the ﬁber have the
ability to adsorb analytes and physically retain them, which results in tighter retention
of the analyte molecules that are small enough to ﬁt into the pores in the sorbent

(Shirey, 1999). Fibers containing porous materials are generally better for trace level

19

analytes (ppt or ppb) with low distribution constants. Therefore, the PDMS/DVB and
PDMS/CAR coatings can extract greater amounts of VOCS than the PDMS coating.
However, competitive adsorption and displacement effects make mass calibration and
quantiﬁcation partially challenging. Koziel et al. (2000) have described a method
based on diffusion controlled extraction for the quantiﬁcation of SPME measurements.
This research also showed that temperature and humidity affect the adsorption
process. High humidity in the ambient air had a negative effect on the adsorption, but
Short sampling times can minimize this effect.

Headspace SPME has been investigated for the analysis of ﬂavor volatiles
using two types of ﬁber coatings (PDMS and Polyacrylate (PA))(Steffen and
Pawliszyn, 1996). This research showed that a PA coated ﬁber extracted polar volatile
compounds more efﬁciently than a PDMS coated ﬁber.

Pan et al. (1997) documented the analysis of amines in gaseous and aqueous
matrices using SPME-derivatization. Three types of the ﬁber coatings were tested;
Carboxen/DVB, Polyacrylate (PA) and PDMS. The results Showed that the amines
had a high affinity for the polar Carboxen/DVB and PA coatings. This research also
showed that hydroxyl groups on a Carboxen/DVB coating interacted more strongly
with amines than is the case with the PA coating. Derivatization coupled with SPME
signiﬁcantly improved the sensitivity over direct SPME. The derivatizing reagents
used were 2,3,4,5,6-pentaﬂuorobenzylaldehyde (PFBAY) and p—nitophenyl
triﬂuoroacetate (N PTFA).

Air sampling and the analysis of volatile organic compounds with SPME were

investigated by Koziel and others (Koziel and Pawliszyn, 2001). This research showed

20

that air sampling with SPME devices is an alternative to NIOSH ﬁeld sampling
methods for VOCS, semi VOCS, and formaldehyde. The paper also showed that the
PDMS coating was more suitable for the sampling of semi-VOCS and that the
PDMS/DVB coating was more suitable for VOCS and was more efﬁcient in retaining

VOCS over a longer period of time.

1.2.1 Analytical method for the characterization of the odorous
compounds

Most researches involving air analyses in livestock houses have used gas
chromatography, however, the analytical conditions and sampling procedures are
varied. The selection of columns, injection, detector and sampling techniques are very
important in GC. Helmig (1999) summarized 17 research papers which used gas
chromatography for air analysis. In this review paper, he summarized the columns,
carrier gases, detectors, and sampling techniques used (Helmig, 1999). Capillary
columns were widely used for air analysis. For mass spectrometric detection, columns

with an inside diameter of 0.32mm or less were preferred.

1.2.3 Odor measurement: Olfactometry and GCISensory

There are four terms (frequency, intensity, duration and offensiveness) that
describe the severity of odor. Frequency and duration depend on wind direction and
the footprint of the odor source, as it affects those in the path. There is no
measurement tool available for these two terms. In general, if several complaints have
occurred within a few years, then there may be an odor problem.

Several methods are used to deﬁne odor intensity. The human nose is the most

accepted standard measurement. Olfactometry is commonly used to measure the odor

21

intensity and Odor offensiveness in air. Olfactometry is a psychophysical method
based upon the responses of individuals snifﬁng diluted odors presented by an
olfactometer. Olfactometry can provide an effective approach to the measurement of
the odor concentration of complex odors.

In the testing procedure, a diluted odorous mixture and an odor-free gas (as a
reference) are presented separately from two snifﬁng ports to a group of panelists. In
comparing the gases emitted from each port, the panelists are asked to report the
presence of odor together with a conﬁdence level, such as guessing, inkling, or
certainty. The gas dilution ratio is then decreased by a factor of two (i.e., chemical
concentration is increased by a factor of two). The panelists are asked to repeat their
judgment. This continues for ﬁve - Six different dilution levels, resulting in a total of
up to 8x6x2 = 96 judgments (snifﬁngs) from eight panelists. Using the panelists
responses over a range of dilution settings, odor concentration, expressed as odor unit
(OU) per cubic meter, can be calculated from the individual threshold estimates. The
OU is used interchangeably with the term dilution to threshold (DT). The dilution to
threshold can be calculated using the following equation.

_VO+VA
V

0

DT

 

DT = dilution to threshold (volume/volume)
V() = volume of odorous air,
VA = volume of fresh air

where V(, and VA are determined at the dilution threshold.
The typical standard for the Olfactometry is l-butanol because of its stability,

low toxicity, and generally agreeable smell.

22

CHAPTER 2. MATERIALS AND METHODS

This chapter describes the sampling and analytical methods that have been

used in the work described in this dissertation.

2.1. Sampling methods
2.1.1. Solvent Desorption Sorbent Tubes

Table 2-1 shows the sorbent tubes used in this research.

Table 2-1. Sorbent tubes used for the sampling of VOCS.

 

 

Packing material Supplier1 Part No. Dimensions Amount of Plug
D , L (mm) packing - material2

material

(mg)
Amberlite XAD-7 Supelco ORBO 615 6, 7O 50/ 100 CW
Anasorb CSC SKC Inc. 226-01 6, 70 50/100 F, GW
Silica gel Supelco ORBO 52 8, 110 150/ 150 GW
Silica gel SKC Inc. 226—10 6, 70 75/150 GW
Tenax SKC Inc. 226-35 6, 70 15/30 F, GW

 

lSupelco, Bellefonte, PA; SKC Inc., Eighty Four, PA

2This is the material used in the plugs that retain the packing in the tubes (see Figure
2.1) F: foam, GW: glass wool

Figure 2-1 shows the sampling train used. A SKC (Aircheck sampler 224-
PCXR8) pump was used for air sampling and the sample ﬂow rate was measured both
before and after sampling using a mass ﬂow meter.

The sampling conditions used are listed in the Table 2-2. The sample ﬂow
rates and the total sample volumes for the tubes chosen were based on the

manufacturer’s recommendations.

23

Table 2-2. Sampling ﬂow rates and volumes used for the collection of VOCS on
solvent desorption sorbent tubes.

 

Sorbent tube Flow rate (ml/min) Total Volume (L)

 

Amberite XAD-7 100 24
Anasorb CSC 200 24
Silica gel (large) 200 36
Silica gel (small) 100 18
XAD 100 24
Tenax 100 6

 

Plug (Foam or glass wool)

 
 
 
 

, Bed 1&2
V4” tubing

 

 

 

ooooooo I
L A A A A A

 

 

 

f Flow re lator Sorbent tube

Pum
p Adapter

Figure 2-1. Sampling train for ORBO tubes

24

2.1.2. Sample Preparation
2.1.2.1. Charcoal

The glass wool and foam plugs were removed from the tubes and the sorbent
materials from the ﬁrst and second beds were transferred to separate 2 mL GC auto-
sampler vials. One mL Of formic acid and CS2 were added to each GC vial to desorb

VOCS from the sorbent. The injection volume to the GC was 5 uL.

2.1.2.2. Silica gel

Two types of Silica gel tubes were used. For the large Silica gel tube, 95%
ethanol was used for the desorption of VOCS. For the small tube, dilute H2804 in 10%
(v/v) aqueous methanol was used for the desorbing solvent.

For the large ORBO 52 tubes, the beds were transferred to separate 2 mL vials
and 1 mL of 95% ethanol was added to each vial. The vial then was capped and placed
in an ultrasonic bath for 1 hour. The vials were allowed to sit for another 30 minutes,
to allow the particles to settle. The supernatant was then transferred to a 2 mL vial.
Five uL of the sample was injected into the GC.

For the SKC silica gel tube, the beds were transferred to separate 2mL vials
and lmL of dilute H2804 in 10% (v/v) aqueous methanol was added to each vial. The
vials were then allowed to sit for 30. A 500 uL aliquot was transferred to a clean 2 mL
vial and 500 pL of 0.3 M KOH was added to the Vial to neutralize the sample. The
samples were then immediately analyzed. Five uL of the sample was injected into the

GC.

25

2.1.2.3. XAD

Three types of solvents were used for the desorption of VOCS from XAD
tubes: methanol, CS2, and methylene chloride. The front and back beds of the sorbent
tubes were transferred to separate 2mL vials and lmL of the desorption solvent was
added to the vials. The vials were then capped and placed in an ultrasonic bath for 30
minutes. The vials then allowed to sit for 30 minutes to allow the particles to settle.
The supernatant in each vial was transferred to 2mL septum Vial. The injection volume

to the GC was 1 uL.

2.1.2.4. Tenax

Three types of solvents were used for the desorption of VOCS from Tenax
tubes: methanol, CS2, and methylene chloride. The glass wool and the foam were
removed from the tube and the sorbent materials in the ﬁrst and second beds were
transferred to separate 2 mL Vials. Five hundred (500) pL of solvent was added to the
vial and it was capped. The vial then was placed in an ultrasonic bath for 30 minutes.
The vials were allowed to sit for 30 minutes to allow the particles to settle. The
supernatant was then transferred to a 2mL septum Vial. An injection volume of 5 uL

was used for GC analysis.

26

2.1.3 Sampling using a midget impinger
Two 25 mL midget impingers (SKC Inc., 225-36-2) were attached to the
sampling pump as shown in Figure 2-2. Fifteen (15) mL of trapping solution was
placed in each impinger. A ﬂow rate of 700 mL/min was used for sampling and total

of 42L of air was collected. The sample solution was injected to GC/F ID or GC/MS

without any further treatment. An injection volume of SuL was used for GC analysis.

 

 

Pump

 

 

 

 

 

 

 

Filter

Flow meter

 

 

 

 

 

 

 

Impinger Diffuser

' Figure 2-2. Sampling diagram for the impingers

27

2.1.3. Sampling on thermal desorption sorbent tubes

Quarter inch diameter ( Mr”) and 7” long stainless steel multi-bed sorbent tubes
were used. The pre-packed tubes were purchased from Supleco. Figure 2-3 shows the
construction of the thermal tube. The tubes contained 30mm of Tenax TA and 30 mm
of Carboxen or 30 mm of Tenax TA and 30mm of Carbosieve $111. The two sorbent
materials were separated by unsilanized glass wool. Prior to sampling, the tubes were
conditioned at 280°C for 30 minutes to remove any contaminants on the sorbent
material. Three liters (3 L) of high purity helium was passed through the tube during
conditioning. The conditioned tube was sealed with a Teﬂon plug and wrapped with
aluminum foil and placed in the opaque clean container and stored in a reﬁigerator at
4°C until it was used.

The sample was taken using a portable and programmable air sampling pump
(SKC Inc.). Figure 2-4 shows the sampling diagram for the thermal desorption tubes.

The sampling time was 30 minutes. A ﬂow rate of 200 mL/min was used for sampling.

2.1.4. Sampling for the particulates
For sampling dust, a 37mm cassette ﬁlter holder (225-3, SKC, Inc.) with a

membrane ﬁlter (pore size of 1.0 pm, SKC Inc. 226-7) and a support pad (Metricel,
Gelman Lab.) was used. The three-piece cassette ﬁlter and ﬁlter holder were
connected to the air sampling pump. A ﬂow rate of 200 mL/min was used for the dust
collection. The total sampling time was 30 minutes. Samples were carried in an ice-

box to the laboratory and stored in a refrigerator (4 °C) until they were analyzed.

28

Tenax TA, 30mm Carboseive SIII, 30mm

f— L :::::::::::::: l

ssssssss
lllllll
........
.......
IIIIIIII
.......
ssssssss
.......

Z ’ ' i
'A” D glass tube\ /

unsilanized glass wool

 

 

 

 

 

 

 

Figure 2-3. Construction of the thermal desorption tube

 

 

 

 

 

ﬂo regulator
lg Sorbent tube ﬁlter
P P

Figure 2-4. Sampling diagram for the thermal desorption tube

29

2.1.6 Sampling using Solid Phase Microextraction ﬁbers

Three different ﬁbers were evaluated to determine which ﬁber has the greatest
efﬁciency for extraction of the odorous volatile compounds from the air of the swine

building. Table 2-3 shows the ﬁbers used.

Table 2-3. SPME Fiber coatings used for the study

 

 

Fiber Phase Frlm thickness
(11m)
PDMS Polydimethylsiloxane 1 OO
PDMS/CAR PDMS and Carboxen (porous carbon) 75
PDMS/DVB PDMS and divinylbenzene 65

 

A Supelco ﬁeld SPME sampler (504831) was used to collect the VOCS. New
ﬁbers were conditioned prior to use according to the manufacturers speciﬁcations, so
as to remove any contaminants on the ﬁber coatings. The temperature and times used
for conditioning were as follows: PDMS, 250°C for 1 hour; PDMS/DVB, 260°C for
30 minutes, and PDMS/CAR, 280°C for 30 minutes. The conditioned samplers were
wrapped in aluminum foil and stored in the refrigerator until they are used.

The grab sampling method was used for SPME air sampling. For the sampling
of VOCS, the ﬁbers were exposed to the ambient air of the swine building for the
required sampling time. After sampling, SPME ﬁbers were carried in an ice-box to the

laboratory and the sample was stored in the reﬁigerator until they were analyzed.

30

2.2. Analytical Instrumentation

2.2.1. Thermal desorber

An Aerotrap 6000 (Tekmar) was used for the thermal desorption of the sorbent
tubes. A cryo-trap in the Aero trap concentrates the volatiles in —165 °C with liquid
nitrogen. The instrument contains a moisture trap to remove water from the sample.

The conditions used for the thermal desorber are as follows; trap cool at -
165°C, sample desorption at 250°C for 20 minutes, trap desorb preheat at 200°C. The
traps weredesorbed for 5 minutes at 250°C. High purity helium was used for the

carrier gas and carrier gas ﬂow rate of 20 mL/min was used.

2.2.1.1. Desorption of VOCS from particulates

An empty %” tube was used for desorption of VOCS from the ﬁlter. A ﬁler was
placed in the empty tube. Other conditions were same as the normal thermal

desorption mode.

2.2.2. Gas chromatography and mass spectrometry
A Varian 3800 gas chromatograph (Varian Chrompack, GC-3800) with a

programmable injector (1079) was used for the analysis of volatile compounds. The
GC was connected to a Saturn 2000 Mass Spectrometer (V arian Chrompack). The GC
also had a ﬂame ionization detector , FID. High purity hydrogen and air were used for
the combustion gases for the FID, and high purity helium was used for the make up
gas for the F ID.

A DB-SMS (60m, 0.25mm ID, 0.25um ﬁlm thickness) capillary column was
used for the separation of the analytes. High purity helium (99.9995%) was used as a

carrier gas. The column ﬂow was held at 1 ml/min. The GC oven temperature program

31

was; 35°C for 5 minutes, ramped to 100°C for 5°C/min and ramped to 220°C for
20°C/min. The operating conditions for the mass spectrometer were: trap temperature
150°C, transfer line temperature 120°C, scan range 40 to 200 m/Z, scanning speed 1

second/scan. A Varian 8200 autosampler was used to inject the liquid samples.

32

CHAPTER 3. A COMPARISON OF SAMPLING
METHODS FOR THE CHARACTERIZATION
VOLATILE ORGANIC COMPOUNDS IN LIVESTOCK
FACILITIES USING GAS CHROMATOGRAPHY - MASS
SPECTROMETRY.

3.1 . Introduction

AS described previously, the characterization of odorous compounds in air in
an around livestock facilities is- difﬁcult for a number of reasons. First, numerous
compounds are present in the air. O’Neill and Phillips (1992) reported that 168
compounds have identiﬁed in livestock wastes or in the air around livestock facilities.
Many other compounds present have not been identiﬁed. Compounds found in
livestock air and livestock wastes include VFAS, phenolics, indolics, amines, and
sulﬁdes. Another problem is that many odorous compounds may be present at very
low concentrations. O’Neill and Phillips (1992) found that at least thirty of the
compounds found in the livestock wastes had odor thresholds lower than or equal to
0.001 mg/m3. Other difﬁculties are posed from the analytical standpoint, due to the
very wide range of physical and chemical properties of the odorous compounds found
in livestock wastes. Some odorous compounds (e.g., hydrogen sulﬁde, methyl
mercaptan and ammonia) are gases, whereas, as others have boiling points in excess of
250°C (e.g., indole , Skatole).

Gas chromatography (GC) has been widely used for the analysis of odorous
compounds in air. Due to the very low detection limits required to identify some
odorous compounds, it is necessary to concentrate the compounds present in the air.

Two concentration methods, sorbent trapping and impinger trapping, have been

33

widely used in standardized methods published by the EPA, NIOSH and OSHA.
Thermal or solvent desorption may be used to transfer the analytes trapped on a
sorbent tube to the analytical instrument. The impinger solution can usually be
transferred directly into the analytical instrument. Recently another sample
concentration method for air analysis, Solid Phase Microextraction (SPME), has been
investigated. Several researchers have investigated the applications of the SPME for
air and headspace analyses (Arthur and Pawlizyn, 1990; Zhang and Pawliszyn, 1993;

Koziel and Pawliszyn, 2001).

3.2. Objectives

The main objective of this research were to determine the most effective air

sampling method for the odorous volatile organics in swine building air.

3.3. Materials and methods
3.3.1. Swine housing

The sampling was conducted in a nursery room in the Michigan Sate
University Swine Teaching and Research Farm. The nursery has 30 pens with eight
pigs in each pen. The temperature in the nursery room was maintained at about 24°C
(76°F) by mechanical ventilation. The sampling was done every day at 4 PM.

Temperature and relative humidity in the room were measured.

3.3.2. Samplings

Four sampling methods were used; solvent desorption sorbent trapping,
thermal desorption sorbent trapping, impinger trapping, and SPME. Table 3-1 shows

the sorbent tubes and desorption solvents that were chosen for the study. Desorption

34

solvents were varied to see the effect of the solvent types on the results. Some of
sorbent tubes were specially designed for the certain solvent, in that case only the

speciﬁed solvent was used.

Table 3-1. Summary of sorbent tubes and desorption solvents chosen for the collection
of various VOCS in livesotk building air.

 

 

ORBO tube Analyte/Sampling Solvent
ORBO 608 Volatile organics Methanol, CS2, methylene chloride.
(Amberlite XAD-2)
ORBO 615 Phenol, p-cresol Methanol
(XAD-7) Other volatile CS2
organics
Tenax Volatile organics Methanol, CS2, methylene chloride
ORBO 328 Acetic acid lmL formic acid
(coconut charcoal) Esters lmL CS2
Ketones lmL CS2
Aromatic lmL CS2
hydrocarbons
ORBO 52S Amines Aromatic: 95% ethanol (1h ultrasonic bath)
(Silica gel) (aliphatic and Aliphatic: 1 mL dilute H2804 in 10% (v/v)
aromatic) aqueous methanol (3 h in ultrasonic)

 

For the experiments used in impingers, the trap solution and conditions were
chosen based on the previous researches. Table 3-2 shows the trapping conditions for

the various organic compounds.

Table 3-2. Liquid trap solutions for impinger air sampling method for the collection of
VOCS in the livestock building air.

 

 

Compounds Solvent Injection
Acids and neutrals (Phenol, 0.1N NaOH Neutralize with 5% sulfuric
p-cresol, VFAS, etc.) acid before inject
Bases (amines, etc.) 5% sulfuric acid Neutralize before inject

 

35

Sampling and conditioning methods for the tubes and SPME ﬁbers are

described in section 2.1.

3.2.3. Analytical instrument

Two identical GC columns, DB-SMS columns, were connected to the
switching valve and TD and GC injector. The switching valve was used to switch the
TD or GC injector to MS.

An Aerotrap 6000 thermal desorber was used for the thermal desorption. The
conditions for the desorber are described in the section 2.2.1. Varian 3800 GC and
2000 MS were used for the analysis of VOCS from various samples. The analytical

conditions used for GC/MS analysez are described in the section 2.2.2.

3.4. Results
3.4.1 . Solvent desorption sorbent traps and Impinger traps

The solvent desorption methods lacked the sensitivity necessary to determine
the odorous compounds. Figure 3-1 shows the chromatograms Obtained with samples
collected on a charcoal trap. As shown in this Figure, only a few small peaks are
observed in the chromatogram.

The impinger traps also lacked the sensitivity necessary to determine the

odorous compounds using the sample volume of up to 72 L.

36

Charcoal bed-l

i 1 1: .j Charcoal bed-2

 

 

' ‘ l

l l 2:
1 .1

i I i ..

.7 i
i l V
I I ll ‘2
1 l .1

I . l
. } I “x
l i i "as.

f 3 3‘
r J A l

\
i i r ‘\-j.\‘
i i ,‘ “‘5':ka
. i .. a-
' ‘ ‘72-}.

j : i bivwwle-‘--»W
‘ I L j i
l 0‘.

i- v v v . V< r V I1 v w v rmVY-v‘v-v v‘V'vﬁ#:‘wv-vvvwv—'V-v'~vvvwwrvvv~ v1 7 vrve vwv—Tv-vv-ow—w—y—vi
L __ __ __LI

 

 

Figure 3-1. Chromatograms for air samples which were collected on charcoal tubes.

37

3.3.2. Thermal desorption sorbent traps

Figure 3-2 shows typical chromatograms for air samples that were trapped on a
thermal desorption tubes. A larger number of volatile organics were found. Figure 3-2
showed that the Tenax TA — Carbosieve SIII tubes were able to collect the volatile
(early eluting) organic compounds more effectively than the Tenax TA — Carboxen
tubes. The recovery Of compounds in the mid-boiling point range was better with the
Tenax TA — Carboxen tubes. The recovery of high boiling point compounds, such as
indole and Skatole, were poor with both types of sorbent tubes. The major compounds
found in swine house air were hydrocarbons and the substituted benzenes (mono, di
and tri substituted alkyl benzenes). Sulfur compounds such as 1,2-ethanthiol, dimethyl
disulﬁde and dimethyl trisulﬁde were also present.

Table 3-3 and 3-4 lists of the compounds which were detected from the swine
nursery air by thermal desorption sorbent tubes and GC/MS. The compounds were
found by matching the spectrum of the compound with the library provided with the
Saturn instrument (N IST/EPA/NIH Mass Spectral Library ‘98). The matching
threshold was set to 700.

Polar biogenic VOCS such as alpha-pinene, beta-pinene, limonene, 3-carene
were also found. These compounds have been identiﬁed and quantiﬁed in air in rural
area (Ciccioli et al., 1993). These compounds were not conﬁrmed by comparing the
retention time and Spectra with those of standards. Isomers were not determined. The

stereo-chemical information for the compounds is not available at this stage.

38

 

 

‘ . Carbosieve

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

i l ‘ z
.. I ‘ . l . i
i l \li 1,!»le if: E. L. 521;-“ I ‘
Carboxen

}

i. «ll.
' [5,521 .— 4 r.‘ , 1.... I” var—rv J
l. _ 75-. __ L A _

Figure 3-2. Chromatogram for air sample which were sorbed on thermal desorption
tubes.

39

Table 3-3. Compounds found in swine house using the thermal desorption method

(Tenax TA and Carbosieve SIII)

 

Tenax TA + Carbosieve SIII

 

 

Retention Time (min) Compounds Molecular Weight
5.427 l-propanol 60
6.098 2-butanol 74
7.945 l-butanol 74
8.871 acetic acid 60

9.38 dimethyl disulﬁde 94
11.294 propanoic acid 74
12.752 Hexanal 100
13.728 butyric acid 88
1 5.528 p-xylene 106
16.485 m-xylene 106

16.78 heptanal 1 14
19.971 unknown
20.23 1 ,2,3-trimethylbenzene 120
20.733 2-ethyl-2-hexanal 126
21.281 1,2,4-trimethylbenzene 120
21.427 lirnonene 136 I
23.562 decane 142
23.865 nonanal 142

 

4o

Table 3-4. Compounds found in swine house using the thermal desorption method

 

 

 

Texax TA + Carboxen
Retention Time (min) Compounds Molecular Weight

5.795 Benzene 78
8.395 acetic acid 60
10.177 dimethyl disulﬁde 94
l 1 .06 1,6-heptadiene-3-yne 92
1 1.984 4-octene l 12
12.329 2,4-dimethylhexane 114
12.511 hexanal 100
12.955 2-octene l 12
13.279 unknown
1 5.066 ethylbenzene 106
15.444 p-xylene 106
1 6.3 l 6 m-xylene 1 06
16.707 heptanal 1 l4
1 7.578 1 -methylethyl benzene 120
17.917 pinene 136
1 8.705 propylbenzene 1 20
1 8 .982 l -ethyl-2-methy1benzene l 20
1 9.23 l ,2,3-trimethylbenzene 1 20
1 9.425 dimethyltrisulﬁde l 26
1 9.685 l-ethyl-3 -methylbenzene 1 20
20. 172 l ,2,4-trimethylbenzene 120
21.386 | limonene 136
22.167 l-methyl-3 -propylbenzene 134
22.398 1 ,2,3,5-tetramethylbenzene 1 34
23.545 2-methyl-1-octanol

 

41

3.3.3. Solid Phase Microextraction

Figure 3-3 Shows chromatograms for air samples which were taken with three
different types of the SPME ﬁbers. The PDMS coating was not effective for the
collection of the volatile organics in the air. The PDMS/CAR coating was the best for
the extraction of the compounds of intermediate volatility, such as acetic acid,
propionic acid, butyric acid, and hexanoic acid. In general the PDMS/DVB coating
permitted the best recovery of the high boiling point compounds, such as phenol, p-
cresol, p-ethylphenol, indole, and Skatole, in swine building air. The recovery of low
boiling point compounds was not good for all three types of ﬁbers.

Tables 3-5, 3-6, and 3-7 Show the compounds found in swine nursery using the
three types of SPME ﬁbers. p-cresol and Skatole were found using all three types of
the ﬁbers. VFAS were found using PDMS/CAR and PDMS/DVB coatings.The
recoveries of the VFAS were higher in PDMS/CAR coating. Phenol was only detected
using the PDMS/DVB coating in the experiment. Indole was not detected usingany of
the ﬁbers, however, in the experiments described in Chapter 5 indole was detected in

swine building air.

42

 

. PDMS

i PDMS/CAR

 

 

’ PDMS/DVB

 

 

 

Figure 3-3. Chromatogram for air sample which were extracted by SPME ﬁbers.

43

Table 3-5. The compounds found in swine housing air using SPME ﬁber with a

 

 

PDMS coating.

R.T. (min) Compound Molecular Weight.
17.411 Amylamine 87
21.119 2-methyl-1-butanol 88
22.978 p—cresol 108
23.908 1 ,8-octanediol 146
25.87 p—ethylphenol 122
27.016 2-decanol 158
29.774 Undecanal 172
31.835 Skatole 131
32.002 Dodecanal 184

 

Table 3-6. The compounds found in swine housing air using SPME ﬁber with a

 

 

PDMS/CAR coating.

Retentlon Time (min) Compound Molecular Weight
6.887 Acetic acid 60
10.253 propionic acid 74
11.88 isobutyric acid 88
13.491 butyric acid 88
15.233 isovaleric acid ~ 102
15.629 p-xylene 106
16.558 m—xylene 106
16.908 hexanoic acid 1 16
19.232 1-methylethyl benzene 120
19.81 heptanoic acid 128
20.126 4-methylvaleric aicd 1 16
20.292 trimethlbenzene 120
21.802 benzylclcohol 108
22.976 p—cresol 108
23.922 nonanal 142
25.87 p-ethylphenol 122
26.987 methyl salicylate 1 52
31 .857 Skatole 131
33.446 diethyl ester 146

 

44

Table 3-7. The compounds found in swine housing air using SPME ﬁber with a

 

 

PDMS/DVB coating.
Retention Time (min) Compound Molecular Weight
6.6 acetic acid 60
9.496 propionic acid 74
11.644 isobutyric acid 88
13.213 butyric acid 88
14.954 isovaleric acid 102
15.37 valeric acid 102
15.636 p-xylene 106
16.583 m-xylene 106
16.75 hexanoic acid 1 16
19.14 ethylmethylbenzene 120
1 9.24 benzaldehyde 106
19.805 phenol 94
20.288 trimethylbenzene 120
20.577 octanal 128
21 .495 limonene 136
21 .772 benzyl alcohol 108
23.019 p—cresol 108
23.914 nonanal 142
25.705 2-nonenal 140
25.872 p-ethylphenol 122
31 .826 Skatole 131
32.228 3-methyl-benzoxazolimine 148

 

45

Table 3-8. Compounds found in TD tubes and SPME ﬁbers.

 

 

 

TD SPME
Retention Molecular Retention Compound Molecular
Time (min) Compound Weight Time (min) Weight
5.427 l-Propanol 60
5.795 Benzene 78
6.098 2-Butanol 74 6.887 Acetic acid 60
7.945 l-Butanol 74
8.871 Acetic acid 60
9.38 Dimethyl disulﬁde 94
11.294 Propanoic acid 74 10.253 Propanoic acid 74
11.984 4-Octene 112 I 1.88 lsobutyric acid 88
12.752 Hexanal 100
13.728 Butyric acid 88 13.491 Butyric acid 88
15.066 Ethylbenzene 106 15.233 lsovaleric acid 102
15.528 Xylene 106 15 .629 Xylene 106
16.78 heptanal 114 16.908 Hexanoic acid 116
17.578 Methylethyl benzene 120
19.81 Heptanoic acid 128
17.917 Pinene 136 19.805 Phenol 94
19.425 Dimethyltrisulﬁde 126 20.126 Methylvaleric acid 116
20.23 Trimethylbenzene 120 20.288 Trirnethylbenzene 120
20.733 Ethyl-Z-hexanal 126 20.577 Octanal 128
21.427 Limonene 136 21 .802 Benzyl alcohol 108
Methyl-3- 22.976 p-Cresol 108
22. 167 propylbenzene l 34
23.562 Decane 142 23.922 Nonanal 142
23.865 Nonanal 142 25.87 p-Ethylphenol 122
31.857 Skatole 13 1
33.446 Diethyl ester 146

 

46

3.4. Conclusion and Discussions

Without any further sample concentration, solvent desorption sorbent tube was
not able to determine any volatile organics from the swine nursery housing air using a
sample volume of up to 36 L. This method also required handling of hazardous
organic solvent for the sample treatment. As sample concentration would be time
consuming and there is a risk that analytes may be lost during this step, this method
was deemed to be less suitable than the thermal desorption or SPME methods.

The impinger trap was also unsuitable for the determination of VOCS from the
swine housing air, even using sample volume as large as 72 L.

Thermal desorption has been widely used for desorption of sorbent tubes. The
major advantage of the thermal desorption method over solvent desorption is that the
entire sample can be introduced into the analytical instrument. Also, as no solvent is
used to desorb the analytes, there is no interference due to the solvent. The principal
disadvantages of thermal desorption are a thermal desorber is required to desorb the
sorbent tubes and the recovery of high boiling point or reactive compounds can be
poor as they may be sorbed to or degraded within the thermal desorption instrument.

This research showed that the multi-bed thermal desorption method can be
used for the sampling of VOCS from livestock building, however the recovery of high
boiling point compounds was not good.

SPME air sampling method showed the good recovery of the high boiling
point compounds, however this method was not good for the recovery of the low

boiling point compounds.

47

Table 3-8 shows the major compounds found in swine nursery using TD tubes
and SPME ﬁbers. Table shows that the TD was able to collect more early eluting
compounds (low boiling point compounds) and SPME was able to collect high boiling
point compounds.

The SPME air sampling methods, has the following advantages over other
sampling techniques:

. Sampling device is lightweight and compact.
- Reagents are not required for desorption.
. Sensitivity is good.
0 Response is linear with concentration.
0 Generally independent of humidity (except at very high humidities).
o Expensive cryotraps or thermal desorbers are not required.
The major drawbacks of SPME includes:
0 Analytes of low volatility do not reach partition equilibrium quickly
0 Samples are not time-integrated

It is apparent from research that no single methods for air sampling appears to
be satisfactory. For the determination of high boiling point compound such as
phenolics and indolics, SPME air sampling method offered the best results. The
combination of SPME and collection on a Tenax TA — Carboxen sorbent tube
followed by thermal desorption seems to offer the best results. For the low to mid
boiling point compounds, recoveries were good using the thermal desorption sorbent
tubes. For the high boiling point compounds, such as phenolics and indolics, SPME

seems to be the most effective method.

48

CHAPTER 4. QUALITATIVE AND QUANTITATIVE
ANALYSIS OF ODOROUS COMPOUNDS IN
LIVESTOCK BUILDING WITH SOLID PHASE
MICROEXTRACTION AIR SAMPLING METHOD

4.1 Introduction

Solid Phase Microextraction (SPME) is a relatively new sampling technique. It
can be used for sampling liquids or air. Due to the simplicity and ease by which it can
be used for sampling air, along with its sensitivity and speciﬁcity, SPME ﬁbers have
been used widely for the determination of volatile organics in food and ﬂavor analysis
(Yang and Peppard, 1994; Ng et al., 1996; Song et al., 1997; Sostaric et al., 2000;
Vianna and Ebeler, 2001).

Supelco Co. (Bellefonte, PA) supplies SPME ﬁeld samplers that are suitable
for ﬁeld use. SPME ﬁeld sampler consists of SPME ﬁber and a ﬁber holder. A septum
in front of the holder protects the ﬁber from contarninations. Figure 4-1 Shows the
SPME ﬁeld sampler. Several different types of polymer-coated ﬁbers with different
polarities are commercially available. The different ﬁbers are used for various needs.
For example, polar ﬁbers are useful for sampling of polar compounds.

Two sampling modes can be used for air sampling using SPME. In grab
sampling, the ﬁber is said to be “exposed” when it is outside of the needle for
sampling period. In this mode, the ﬁber is exposed to the ambient air. The ﬁber is said
to be “retained” when it is inside the needle for sampling period. The second term is

called “time weighted average (TWA)” sampling. (Koziel and Pawliszyn, 2001). In

49

TWA method, the ﬁber is retained inside of the ﬁber holder and VOCS slowly diffuse

onto the ﬁber inside of the ﬁber holder.

 

Portdale field sampler

 

Figure 4-1. SPME ﬁeld sampler.

50

SPME (grab samples) have been widely used for relatively short sampling
periods. Time integrated sampling with SPME has been challenging. Martos and
Pawliszyn (1999) studied time-weighted average (TWA) sampling using a SPME air
sampler in an attempt to enhance personal exposure monitoring to airborne pollutants.
This research used SPME TWA sampling method to collect VOCS in work
environment and compared with the conventional air sampling method (N IOSH
method 2541). Martos and Pawliszyn (1999) showed that the commercially available
SPME device could be successfully used as a TWA diffusive sampler for both
hydrocarbons and formaldehyde.

The major disadvantage for SPME air sampling method is the difﬁculty in
quantifying the analytes. SPME ﬁbers are not uniformly sensitive to all compounds
and therefore, relative GC peak areas for an SPME sample do not properly reﬂect the
true proportions of the components in the sample to which they are exposed.
Furthermore, other environmental/physical factors such as sampling time and
temperature and humidity can affect the uptake of the contaminant on the ﬁber.

Equilibrium theory has been widely used for the SPME quantiﬁcation and has
been described by several researchers (e.g., Martos and Pawliszyn, 1997; Koziel and
Pawliszyn, 2001). However, it is generally believed that equilibrium theory cannot be
used for the quantiﬁcation of the compounds on the mixed porous ﬁbers that are most
widely used for air sampling (Koziel and Pawliszyn, 2001). The diffusion-based
calibration method for rapid air sampling using SPME has been recently studied by
several researchers (Ai, 1997; Koziel and Pawliszyn, 2000; Augusto et al., 2001).

These researchers Showed that SPME with mixed porous ﬁber coatings could be used

51

for the rapid air sampling. Quantiﬁcation methods also have been described by these

researchers.

4.1.1 Equilibrium theory

Equilibrium theory assumes that the concentration of an analyte in the ﬁber

coating is directly proportional to its concentration in the headspace.

where Cﬁbcr is the concentration on ﬁber
C8,, is the concentration in the headspace and K is the equilibrium

constant.

Because the volume of the ﬁber is constant (uniform SPME ﬁbers are now
commercially available), the equation simpliﬁes to the mass of analyte absorbed by
the ﬁber being directly proportional to the headspace concentration (Bartelt, 1997).
The proportionality constant, K' , (referred to as the calibration factor through this
paper) is;

K' = mass on ﬁber /Cair ................... (2)
and K' = K X Vﬁber

Once the proportionality constants, K’ has been deﬁned, the concentration of
the analyte in the air (Cg) can be calculated from the mass (11). The calibration factors,
K' , are different for different analytes and different ﬁber types. In general, high

boiling point compounds tend to have low K' values.

52

The mass on SPME ﬁber coating can be measured by calibrating against direct
standard injection and by calculating the response factor Of the compound. The
response factor can be determined by on-column injection of the analyte. It is
calculated as nanograms of analyte per unit instrumentation response (unit area) by

dividing the injection mass (ng) by area.

Response factor (RF) = mass (ng) /area ............... (3)

The mass of analyte absorbed on the ﬁber, nﬁbc, (ng), is calculated by
multiplying the GC peak area by the measured response factor.

nﬁber = RF factor x GC area .................... (4)

A calibration curve can be constructed using the serially diluted standard
solutions. A working stock solution was made from stock solution by dilution with
methanol. The working stock solution was serially diluted to have ﬁnal concentrations
of 0.001 PPM for each compound. Curves for each compound have been constructed
(see Appendix B). In this research, the mass (n) was measured using a calibration

curve.

4.1.2 Non-equilibrium theory

The quantiﬁcation of analytes on mixed phases coatings, such as PDMS/CAR
and PDMS/DVB has been difficult because of the competition between analytes for
the adsorptive sites available in the ﬁber and the inter-analyte displacement. In a
recent study, Koziel and Pawliszyn (2001) presented an alternate methodology to
overcome this problem. When a ﬁber is exposed to a gaseous sample moving

perpendicularly to the ﬁber axis for a period of time much smaller than the

53

equilibration time, the coating behaves as a perfect sink and all analyte molecules
reaching the ﬁber surface are immediately adsorbed. As a large number of non-
occupied adsorptive sites are available in these conditions, interanalyte competition
and displacement are minimized and can be disregarded. The extracted amount of an
analyte (n) depends on its concentration in the gaseous matrix Cg, its diffusion
coefﬁcient in air Dg; the ﬁber’s length L; radius b; respectively, the thickness of the
effective static boundary layer surrounding the ﬁber 8; and the sampling time t as

shown:

 

Several models are available to estimate diffusion coefﬁcients in air. The
Fuller-Schettler-Giddings model (Schwarzenbach et al., 1993) which is the most
adequate model for a large number of analytes in normal sampling conditions, the

difﬁrsion coefﬁcient, Dg, can be calculated by following equation:

0.001T"75 l + 1
M

 

 

air M vac
(6)

[(ZVairW +(ZVW)%]2 ..................

where Us is the diffusion coefﬁcient (cmz/s)

 

08:
P

T is the absolute temperature (K)

Mai, is the apparent molecular weight of air (g/mol)

(i.e. the weighed average of the molecular weights of the
components of air)

Mvoc is the molecular weight of the analyte (g/mol)

p is the ambient pressure (atm)

54

V8,, is the molar volume of air (cm3/mol)

V,0c is the molar volume of the analyte (cm3/mol)

The thickness, 8, of the effective static boundary layer surrounding the ﬁber

can be calculated from equation 7, where Re is the Reynolds number (Re = 2ub/V; u
is the linear velocity of the air and v is the kinematic Viscocity Of air) and Sc is
Schmidt number (Sc = v/ n, )

6 = 9.52b/Re0'62 Scm ................... (7)

Using these equations, the concentration of analyte can be directly calculated
from the chromatographic peak area, given the sampling conditions (sampling time,
air velocity, temperature, and pressure) and constants (diffusion coefﬁcient and ﬁber
conditions) are known.

In this research, SPME quantiﬁcations with equilibrium theory and rapid non-
equilibrium theory have been demonstrated. CAR/PDMS coating has been used for
this research Since this is the generally used mixed porous ﬁber coating for air

sampling.

4.2 Hypothesis

1. The mass of volatile organic compounds sorbed on SPME ﬁber coating will be
time dependent and will eventually reach an equilibrium.

2. The mass of sorbed on SPME ﬁber coating with CAR/ODMS coating will
depend linearly on the concentration of the analyte in the gas phase at any

particular sampling time.

55

3. The concentrations of odorous compounds in the livestock facility can be

quantiﬁed using SPME CAR/PDMS coating.

4.3 Objectives

The main objectives of this research are
. To demonstrate the adsorption equilibrium time on SPME ﬁber for odorous
compounds in livestock building air using CAR/PDMS coatings.
c To demonstrate the quantitative analysis of SPME air sampling for odorous
compounds using equilibrium theory.
0 To demonstrate the quantitative analysis using fast SPME air sampling

technique for odorous compounds in livestock building air.

4.4 Materials and Methods
4.4.1 SPME ﬁbers

A SPME ﬁeld sampler (Supleco Co.) with 75 um CAR/PDMS coating was
used in this experiment. Conditions for the conditioning the new ﬁbers are given in the

section 2.1.1.4. The dimensions of the ﬁber coatings were: ﬁber length of 1 cm and the

ﬁlm thickness was 75 um.

4.4.2 Test Mixtures

The mixture contains phenol, p-cresol, p-ethylphenol, indole and Skatole. Table
1 lists concentrations of the compounds in the test mixture stock solution. The stock
solution was made with methanol and the total volume was 25 mL. The stock solution
was been stored in the refrigerator and working stock solution was made by dilution of

the stock solution whenever it was needed.

56

Table 4-1. Stock solution of test mixture used for the SPME air sampling methods.
(Total volume was 25 mL and the solvent was methanol.)

 

 

Compound Amount
Phenol 1.07 g
p-Cresol 1.034 g
p-Ethylphenol 1.0247 g
Indole 1.028 g
Skatole 1.0335 g

 

4.4.3 Air sampling

One or two micro liters of stock solution were injected into the IL gas standard
bottle. Because the mixture was made with methanol, the phenolic compounds were
quickly evaporated in the gas standard bottle (1 L, gas sampling bottle with septum
purge). To prevent the sorption of analytes on to the glass sampling bottle, 100 uL of
water was injected to the gas sampling bottle after all samples were evaporated. The
water vapor did not affect sorption of the analytes on the SPME ﬁber, since the
PDMS/CAR ﬁber coating is a non-polar material. Research has shown that humidity
does not greatly affect on SPME sorption, Since the mass loading on the ﬁber
decreased by 10% at relative humidities greater than 90% (Martos and Pawliszyn,
1997)

For sampling, a freshly conditioned ﬁber was passed through the septum into
the gas sampling bottle. Sampling times were varied from 5 seconds to 420 minutes
according to the test objectives. Triplicate samples were taken for the same sampling
period and averaged. All sampling was conducted in the static condition without

disturbing the air ﬂow in the gas sampling bottle. Fibers were not agitated during the

57

sampling. Figure 4-2 illustrates the sampling procedure using SPME ﬁbers and the gas

standard bottle.

4.4.4 Calibration curve

A calibration curve was prepared by direct injection of serial dilutions of the

test mixture. A Varian 8200 autosampler was used for the injection of liquid standards.

4.4.5 Gas chromatography and mass spectrometry

GC/MS was used for the analysis of SPME samples. The SPME ﬁeld sampler
was introduced into the GC injector. The analytical conditions are given in Section

2.2.2.

58

SPMEﬁeld sampler

{—T 0

1L gas sampling bottle

 

Figure 4-2. SPME air sampling system.

59

4.5 Results

4.5.1 Adsorption Time Proﬁle for phenolic compounds on SPME
CAR/PDMS coating

Sorption time proﬁles for phenolic compounds on SPME CAR/PDMS are
shown in Figure 4-3. The ﬁgure shows that the mass of chemical adsorbed on SPME
CAR/PMDS coating increased as sorption time increased. The increase in sorbed mass
of the phenolic compounds can be described by a linear relationship for times up to 30
minutes. When the sorption time exceeded 30 minutes, the mass adsorbed on SPME
coating increased in a non-linear relationship with time. None of chemicals, p—cresol,
p-ethylphenol, indole and Skatole, reached equilibrium within the course of the

experiment (420 minutes).

4.5.1.1 Relationship between the amount adsorbed on SPME and the
sampling times.

Based on the results mentioned in the previous section, a linear line for time
versus adsorbed masses was constructed for sorption times less than 30 minutes.
Figure 4-4 shows the linear relationship between the mass of phenol adsorbed on
SPME coating and sampling times when the sampling time was less than 30 minutes.
Similar results were observed for p-cresol, p-ethylphenol, indole and Skatole (Figures
4-5, 4-6, 4—7, and 4-8).

The effect of concentration on SPME sorption behavior was studied. A
standard mixture containing 0.43, 041,041, 0.41, and 0.41 jig/L of phenol, p-cresol,
p-ethylphenol, indole, and Skatole was used. The standard mixture was injected into
the gas sampling bottle and a SPME ﬁber was used for the sampling of VOCS in the

gas phase. The masses sorbed on the SPME ﬁber were analyzed using GC/MS. All

60

samples were analyzed in triplicate and three data points were averaged. The averaged
data was used for the linear regression. Results showed that the masses sorbed on
SPME for phenolic and indolic compounds increased in curvature as sampling time
increased from 5 minutes to 30 minutes (Figure 4-9, 4-10, 4-11, 4-12, 3-14). A
previous study by Pawliszyn (1997) Showed that the sample concentration has no
affect on the concentration time proﬁle. When the extraction method and distribution
constants between the SPME and sampling system remain constant, the equilibrium
proﬁle and time should be the same for all other concentration. Therefore, the
extraction behavior should be linear for all concentration ranges. However, the results
showed that the extraction behaviors for phenolic compounds on SPME ﬁbers were
not same for the gas phase concentrations at 0.4 and 4 ug/L.

These results suggest the possible systematic experimental errors. It is
possible that the errors resulted from a depletion of available sorption Sites on SPME
ﬁber. In this experiment, it was assumed that the sorption abilities for the SPME ﬁbers
were the same and calibrations for individual SPME ﬁbers were not conducted. When
a small portion of the ﬁber coating is damaged, the sorption ability would be greatly
reduced even though the damage was too minor to be detected. To avoid this problem,
calibrations for all individual SPME ﬁbers would be required. The lifetime of a new
SPME ﬁber is 50 to 100 samplings depending on sampling conditions. As a SPME
ﬁber ages, the sorption ability of the ﬁber is reduced. Therefore, usages and sampling
histories for all SPME ﬁbers should be carefully recorded. Secondly, the error may
result from environmental conditions, such as temperature. Room temperatures

ﬂuctuated from 223°C to 243° during the experiment, and temperature in the

61

sampling bottle was not controlled and the temperature changes were ignored because
the difference was approximately 2 °C. A previous study Showed that the increase in
the sample temperature increased the analyte concentration in the headspace, thereby
providing faster extraction (Wercinski and Pawliszyn, 1999). This error can be
corrected by sampling SPME ﬁbers at several different temperatures. Another error
may result from the standard solution. A new working standard solution was made
when the Old one was older than 7 days. Experimental error may be introduced in
preparing the standard. To avoid this error, the new standard must be calibrated prior

to being used for the gas standard.

 

 

 

 

 

3.5e+5
3.0e+5 a
2.5e+5 *
a
E, 2.0e+5 4
3 1.5e+5 -
m
iii 1.0e+5 2
2
5.0e+4 .
0.0 r
o 100 200 300 400 50°
Time (min)
_.__ phenol
........ 0......“ p-CTCSOI
-_+—— p—ethylphenol
—~-v-—“' indole
_ _. — skatole

Figure 4-3. Adsorption time proﬁle for phenolic compounds using a static adsorption
conditions using SPME CAR/PDMS. (The standard contained 4.3, 4.1, 4.1, 4.1, 4.1

62

ug/L Of phenol, p-cresol, p-ethylphenol, indole, and Skatole, respectively.) The
average temperature was 23°C and pressure was 1 atm.

Phenol

 

 

 

 

 

80000
60000 -
’6:
C
E 40000 .
8
8 20000 4
m
2
o a
‘20000 I I I T l l
0 5 10 15 20 25 30 35
Time (min)
—— Plot1 Regr
— — Plot1 Conf1

Figure 4-4. The relationship between the sorbed mass and sorption time on SPME for
Phenol (4.3 rig/L). The upper and lower lines represent the 95% conﬁdence interval.
The average temperature was 24°C and pressure was 1 atm.

63

p-Cresol

16000

 

14000 - //
12000 -
10000 -
8000 -

6000 -

Mass sorbed (ng)

4000 -:

2000 ‘

 

 

 

 

I I T I I

10 15 20 25 30 35
Time (min)

—— p-Cresol Regr
-—-- p-Cresol Conf1

Figure 4-5. The relationship between the sorbed mass and sorption time on SPME for
p-Cresol (4.1 rig/L). The upper and lower lines represent the 95% conﬁdence interval.
The average temperature was 23°C and pressure was 1 atm.

64

 

p-Ethylphenol

 

 

 

 

4000 /
/
/
/
// '
3000 a /
A // //’
tn
5 / ///
/
E // ///
8 2000 4 / . //
/ /
(I) / /
3 / /
2 /// //
/
/
1000 4 // . /
/
/ e /
/
/
/
0 I/ I I I l I
0 5 10 15 20 25 30 35
Time (min)
— Plot1 Regr
—— Plot1Conf1

Figure 4-6. The relationship between the sorbed mass and sorption time on SPME for
p—Ethylphenol (4.1 pg/L). The upper and lower lines represent the 95% conﬁdence
interval. The average temperature was 23°C and pressure was 1 atm.

65

Indole

8000

 

6000 a

4000 a

2000 a

Mass sorbed (ng)

 

 

 

 

 

-2000 1 TV I I I I
0 5 10 15 20 25 30 35
Time (min)
Indole Regr
— -— Indole Conf1

Figure 4-7. The relationship between the sorbed mass and sorption time on SPME for
Indole (4.1 rig/L). The upper and lower lines represent the 95% conﬁdence interval.
The average temperature was 23°C and pressure was 1 atm.

66

Skatole

2500

 

2000 -

1500 a

1000 4

Mass sorbed (ng)

 

 

 

 

-500 I I T T I l
0 5 10 15 20 25 30 35
Time (min)
— Plot1 Regr
— — Plot 1 Conf1

Figure 4-8. The relationship between the sorbed mass and sorption time on SPME for
Skatole (4.1 rig/L). The upper and lower lines represent the 95% conﬁdence interval.
The average temperature was 23°C and pressure was 1 atm.

67

Phenol

 

 

 

 

 

1000
800 a
a 600 a
5
E 400 -
8
8
(U
2 200 -
o 4
'200 r I I l l I
0 5 10 15 20 25 30 35
Time (min)
— Plot1 Regr
— — Plot 1 Conf1

Figure 4-9. The relationship between the sorbed mass and sorption time on SPME for
Phenol at lower concentration (0.43 ug/L). The upper and lower lines represent the
95% conﬁdence interval. The average temperature was 23°C and pressure was 1 atm.

68

p-Cresol

250

 

200 - /

Mass sorbed (ng)

 

 

 

 

Time (min)

— Plot1 Regr
- — Plot 1 Conf1

Figure 4-10. The relationship between the sorbed mass and sorption time on SPME for
p-Cresol at lower concentration (0.41 pg/L). The upper and lower lines represent the
95% conﬁdence interval. The average temperature was 23°C and pressure was 1 atm.

69

p-Ethylphenol

 

Mass sorbed (ng)

 

 

 

 

o I ﬂ 1 l r I

0 5 10 15 20 25 30 35
Time (min)
—- Plot 1 Regr
— -— Plot 1 Conf1

Figure 4-11. The relationship between the sorbed mass and sorption time on SPME for
p-Ethylphenol at lower concentration (0.41 pg/L). The upper and lower lines represent
the 95% conﬁdence interval. The average temperature was 23°C and pressure was 1
atm.

70

Indole

 

Mass sorbed (ng)

 

 

 

 

Time (min)

— P1011 Regr
— — Plot1 Conf1

Figure 4-12. The relationship between the sorbed mass and sorption time on SPME for
Indole at lower concentration (0.41 pg/L). The upper and lower lines represent the
95% conﬁdence interval. The average temperature was 23°C and pressure was 1 atm.

71

Skatole

 

Mass sorbed (ng)

 

 

 

 

0 5 1O 15 20 25 30 35
Time (min)
-— Plot 1 Regr
-— -— Plot 1 Conf1

Figure 4-13. The relationship between the sorbed mass and sorption time on SPME for
Skatole at lower concentration (0.41 pg/L). The upper and lower lines represent the
95% conﬁdence interval. The average temperature was 23°C and pressure was 1 atm.

72

4.5.2 Concentration Effects on SPME Adsorption

The effect of concentration on the sorbed mass of chemical onto SPME
coatings was determined. Serially diluted test mixtures were used for this experiment.
The diluted standard mixtures were injected into the gas sampling bottle. Two (2) pL
of liquid volume was injected in all cases. SPME sampling from the gas sampling
bottle was conducted when the all liquid sample was evaporated. The sampling times
were 30 minutes. Room temperature varied from 22°C to 24°C and the pressure was 1
atm.

The results showed that the mass of phenolic compounds sorbed on SPME
ﬁber increased as the concentration in the air increased (see Figures 4-14, 4-15, and 4-
16). However, the mass of phenol sorbed on SPME ﬁber did not respond linearly at
95% conﬁdence interval as injection mass increased when the gas phase
concentrations were increased from 0.4pg/L to 4ug/L. Similar results were Shown for
p-cresol, p-ethylphenol, indole and Skatole. Result showed that linearity was observed
for certain concentration ranges for 30 minutes sampling.

The results suggest possible systematic errors. Standard solutions were made
from the high concentrated stock mixture solution which contained all phenolic and
indolic compounds in methanol. The standard solutions were used for SPME gas
sampling. The ﬁfth data points for all ﬁgures differed signiﬁcantly from the linear
regression line ( see Figures 4-14, 4-15, 4-16, 4-17 and 4-18). This result suggests that
there was a mistake on making the standard solution for the data. Calibration of the

standard solution must be conducted to avoid this error.

73

For the indolic compounds, the masses sorbed on SPME coating were less than
that of those of the phenolic compounds (Figures 4-17 and 4-18). This result indicates
that the adsorption rates of indolic compounds on SPME CAR/PDMS coating are
slower than those of phenolic compounds, resulting in less sorption of indolic
compounds.

To see the effect of molecular weight on sorption, a graph for the molecular
weight and sorption rate were constructed (Figure 4-19). The initial sorption rate was
calculated using the mass sorbed after 5 minute sampling. The graph shows that the
initial sorption rates for phenolic compounds decreased exponentially as the molecular
weight increased. The molecular diffusion coefﬁcients decrease as the molecular
weights of the chemicals increase. It appears that molecular diffusivity is the driving
force for the sorption on SPME ﬁber. Larger chemicals move Slowly because the
increased cross-sectional area reduces mean free path, that is, their ability to slip
through a crowd of other molecule (Schwarzenbach et al., 1993). Molecular diffusion
coefficients and the molecular weight of the phenolic compounds are described in
Section 4.5.4.

For a given sampling time, the masses of indolic compounds sorbed on this
SPME ﬁber coating can be increased by increasing sampling time. However,
increasing sampling time may cause problems, such as competition and displacement
of analytes from SPME ﬁber.

The masses of analyte sorbed on SPME ﬁbers were linearly dependent on

sampling time when the sampling times were less than 30 minutes. When the gas

74

phase concentrations were reduced by factor 10, the masses of sorbed on SPME ﬁber
did not respond linearly to the sampling times.

A major disadvantage with this method for the quantiﬁcation is maintaining
constant conditions. The results in this research were obtained under controlled
laboratory conditions where temperature and humidity were relatively constant.
However, the physical and environmental conditions for actual sampling in the
livestock building would be difﬁcult to control.

Humidity was not considered in this research, Since it was shown that humidity
does not affect greatly SPME adsorption (Martos and Pawliszyn, 1997). In addition,
the coating materials used in this research are non-polar and hydrophobic compounds.
Future research is suggested to determine temperature corrections. Temperature
corrections can be made by measuring the amount adsorbed on SPME ﬁber at
different gas phase concentrations and at different temperature.

Several systematic errors may have been introduced during experiments
conducted as part of this research project. The working standard solution was a
possible error source. When a new working standard is made, the standard must be
properly calibrated. Another error source is SPME ﬁber. A SPME ﬁber has a certain
lifetime on it’s usage. As the ﬁber ages, the sensitivity of the ﬁber diminishes. When
SPME ﬁber is used for the quantiﬁcation of the analytes, the ﬁber must be carefully
calibrated.

The gas phase concentrations were calculated using the concentration of the
standard mixture injected into the gas sampling bottle in this research. When sorption

of chemicals onto the gas sampling bottle occurs, the actual gas phase concentrations

\

75

will be less than the calculated concentrations. To avoid this experimental error, the
real gas phase concentration must be measured by injecting the standard gas directly

into GC.

Phenol

 

35000

30000-

25000 -

20000 -

15000 a

Mass sorbed (ng)

 

 

 

in
O
8
.4

 

l I

0 1 2 3 4 5

Gas phase concentration (pg/L)

 

Plot1 Regr
— — Plot 1 Conf1

Figure 4-14. Mass injected in the gas sampling bottle vs mass recovered on SPME
ﬁber for Phenol with sampling time of 30 minutes. The upper and lower lines

represent the 95% conﬁdence interval. The average temperature was 23°C and
pressure was 1 atm.

76

p-Crssol

 

Mass sorbed (ng)

 

 

 

 

0 1 2 3 4 5

Gas phase concentration (pg/L)

—— Plot1 Regr
— — Plot 1 Conf1

Figure 4-15. Mass injected in the gas sampling bottle vs mass recovered on SPME
ﬁber for p-Cresol with sampling time of 30 minutes. The upper and lower lines
represent the 95% conﬁdence interval. The average temperature was 23°C and

pressure was 1 atm.

77

p-Ethylphenol

1400

 

1200 .

1000 a

Mass sorbed (ng)
§

 

 

 

 

400-
2004
01
-200.
o 1 2 3 4 5

Gas phase concentration (pg/L)

Plot 1 Regr
— — Plot 1 Conf1

 

Figure 4-16. Mass injected in the gas sampling bottle vs mass recovered on SPME
ﬁber for p-Ethylphenol with sampling time of 30 minutes. The upper and lower lines
represent the 95% conﬁdence interval. The average temperature was 23°C and
pressure was 1 atm.

78

p-Ethylphenol

 

1600

1400 r

1200 ..

1000 ‘

800 -

600‘

400 -‘

Mass sorbed (ng)

200‘

 

 

 

 

Gas phase concentration (pg/L)

— Plot1 Regr
— — Plot 1 Conf1

Figure 4-17. Mass injected in the gas sampling bottle vs mass recovered on SPME
ﬁber for Indole with sampling time of 30 minutes. The upper and lower lines represent
the 95% conﬁdence interval. The average temperature was 23°C and pressure was 1
atm.

79

Skatole

 

 

 

 

 

400
300 -
a
I:
E 200 ~
O
10
8 100 ~
(0
2
o ..
-100 I I I T
o 1 2 3 4 5

Gas phase concentrati on (pg/L)

 

Plot 1 Regr
— — Plot 1 Conf1

Figure 4-18. Mass injected in the gas sampling bottle vs mass recovered on SPME
ﬁber for Skatole with sampling time of 30 minutes. The upper and lower lines
represent the 95% conﬁdence interval. The average temperature was 23°C and

pressure was 1 atm.

80

 

1800
1600
1400 y ____ 3E+068-0.0815x
1200 R2 = 0.992
1000 -
800
600 .
400 -
200

Initial sorption rate (nglmin)

 

 

 

 

90 100 110 120 130 140
Molecular weight (9)

Figure 4-19. Effect of molecular weight of chemicals on SPME initial sorption rate.
Initial sorption rates were calculated using 5 minute sampling data.

4.5.3 Quantification of SPME analyses using the calibration factors

As described in the introduction, the amount adsorbed on SPME ﬁber
increases linearly as the gas phase concentrations increases when the sampling time is
constant. In this section, quantiﬁcation of SPME was conducted by calculating the
calibration factors. Temperature and pressure during the experiment was 21i2°C and
1 atm, respectively.

Known amounts of a test mixture were injected into the GC/MS and the
response factors for the compounds Were calculated using equation (3). The same
standard solutions were injected into the gas sampling bottle for SPME sampling. The
amount adsorbed on SPME ﬁber was calculated using equation (4), and the calibration
factors for the compounds were calculated using equation (2). Table 4-2 Shows the

amount recovered on SPME and K' (calibration factors) values for phenolic and

81

indolic compounds for a sampling time of 30 minutes. The mass sorbed on SPME
ﬁber (n) was calculated using equation (4) for the response factor, which obtained by
the liquid standard injection, and GC response for SPME. The calibration factors (K’)
were calculated using equation (2) using the known gas phase concentrations and the
mass sorbed on SPME ﬁber (n).

SPME CAR/PDMS coatings were exposed to the volatile compounds for 30
minutes and the adsorbed amounts (n) on SPME coatings were analyzed. With the
adsorbed amounts (n) and calibration factors (K' ), the gas phase concentrations for
the compounds were calculated. Table 4-3 shows the GC responses of the phenolic
and indolic compounds which were adsorbed on SPME ﬁbers, RF and K' values. The
gas phase concentrations were calculated using these values in Table 4-4 and the
calculated gas phase concentrations and the actual amounts injected into the gas
sampling bottle are listed in Table 4-4.

The Standard errors for the calculated gas phase concentrations were 6.23%,
36.6%, 28.2%, 6.27%, and 24.5% for phenol, p-cresol, p-ethylphenol, indole and
Skatole at 95% conﬁdence level, respectively.

For the calculation of the calibration factors, K' , it was assumed that the gas
phase concentrations for the phenolic compounds were same as the amount injected
into the gas sampling bottle. Table 4-4 shows that the calculated gas phase
concentrations were less than the actual amounts injected into the gas sampling bottle
for all compounds. This result indicates that the actual gas phase concentrations may
not to be the same as the amount injected into the gas sampling bottle. This could

happen when the all compounds are not evaporated in the bottle. This problem can be

82

tested by the direct gas injection into GC. Another explanation for this is saturation of
ﬁbers. This reﬂects the situation in which at higher sorbate concentrations, it becomes
more difﬁcult to sorb additional molecules. This can happen when sorption Sites are
ﬁlled or remaining sites are less attractive to the sorbate molecules. This problem can
be tested by using serial steps of standard gases for SPME sorption. When the
chemical is sorbed onto the glass bottle surface, the actual gas phase concentration
would be less than the injected one. This is called “wall effect”. Results for the
concentration effects on SPME sorption Showed that there were non-linearity
relationships between the gas phase concentrations and SPME sorption for phenol at
the concentration ranges from 0.4 to 4 ug/L (Figure 4-14). This result suggests that
wall effect may occur in the sampling bottle. This problem also can be tested by
measuring the gas phase concentration Using direct gas injection into the GC.
Environmental conditions, such as temperature and humidity affect on SPME
sorption. When the temperature in the sampling environment is not constant, this
quantiﬁcation method will not give accurate gas phase concentrations. Since the
humidity did not affect greatly on SPME sorption as discussed previously, humidity

effect may be ignored when non-polar SPME ﬁbers are used.

83

Table 4-2. The amount recovered and Calibration factors for phenolic and indolic
compounds with SPME CAR/PDMS ﬁber for 30 minutes sampling time.

 

 

 

 

 

 

Phenol Cg (pg/L) 11 (ng) K'(ng/pg/L) average K' 95% CI
0.214 0.014 0.066 0097:0053 0.104
0.428 0.028 0.065
0.856 0.041 0.048
1.284 0.118 0.092
1.712 0.204 0.119
2.568 0.495 0.193
p-Cresol 0.207 0.004 0.018 0.031i0.018 0.034
0.414 0.009 0.023
0.827 0.013 0.016
1.241 0.034 0.027
1.654 0.065 0.039
2.482 0.155 0.062
p-Ethylphenol 0.205 0.002 0.009 0.016i0.007 0.014
0.410 0.008 0.019
0.820 0.005 0.007
1.230 0.018 0.015
1.640 0.041 0.025
2.459 0.053 0.021
Indole 0.206 0.006 0.031 0.025i0.010 0.019
0.411 0.016 0.039
0.822 0.011 0.014
1.234 0.023 0.019
1.645 0.045 0.027
2.467 0.043 0.018
Skatole 0.207 0.011 0.052 0.032i0.019 0.037
0.413 0.025 0.059
0.827 0.017 0.020
1.240 0.028 0.023
1.654 0.046 0.028
2.480 0.027 0.01 1

 

Cg: gas phase concentration

n: mass sorbed on SPME )ng)
K' : calibration factor (K’=n/Cg)

CI: conﬁdence interval

84

Table 4-3. GC response of test mixture, RF, and K' values for the quantiﬁcation of
phenolic compounds with 30 minutes sampling time.

 

 

Compounds Response K' RF“

Phenol 548664 0.097 5.58E-08
p-cresol 198826 0.031 9.05 E-08
p-ethylphenol 17131 1 0.016 5.21 E-08
Indole 279680 0.025 5.18E-08
Skatole 347702 0.032 4.92E-08

 

RF’: Response factors were obtained by dividing the amount of liquid standard
injected to GC/MS by instrumental response (area).

K': calibration factor

Table 4-4. Comparison of actual gas phase concentration to the calculated
concentration using RF and K' values for 30 min adsoprtion.

 

Amount on Calculated gas 95% CI Actual gas conc.

 

Compounds Cale/Act.
SPME (pg) conc. (pg/L) Calc. (pg/L)

Phenol 0031:0004 0.32:0.044 0.109 0.856 0.369

p—Cresol 0.018i0.008 0.58:0.259 0.642 0.827 0.704

p-Ethylphenol 0009:0003 0.56:1:0. 199 0.494 0.820 0.681

Indole 0.014i0.001 0.59:t0.047 0.1 17 0.822 0.717

Skatole 0017:0006 0.53:0.173 0.429 0.827 0.644

 

4.5.4 SPME quantification method with diffusion theory

The quantiﬁcation of phenolic and indolic compounds sorbed on SPME ﬁber
was conducted using the diffusion theory. Molecular diffusivities for phenolic and
indolic compounds were calculated using Fuller’s method (1966) as shown in
Equation 6.

There are several ways to estimate the molar volume of the organics. One of
the methods is to estimate Vvoc by dividing the chemical molecular mass by its liquid

density. Since liquid phenol and p-cresol were used in the test mixture, their molar

85

volumes were simply calculated by dividing their molecular mass by their liquid
densities.

Another way to estimate the chemical molar volume is to calculate the “Size”
of the atoms making up the chemical’s structure (Schwarzenbach et al., 1993). Table
4-5 shows molar volumes of various atoms deduced by regression of available

diffusion data.

Table 4-5. Estimation of Diffusion Volumes of the constituents of organic molecules

 

 

Element Volume Contribution (cm3/mol)
C 16.5

H 2.0

O 5.5

N 5.7

CI 19.5

S 17

Rings —20.2

 

Source: “Environmental Organic Chemistry” (Schwarzenbach et al., 1993).

Molar volumes of the p-ethylphenol, indole and Skatole were estimated using
the values in Table 4-5. For example, the estimation of the molar volume of p-
ethylphenol is Shown below. The molecular structure of p-ethylphenol is (3311100 and
it has a ring.

V p—elhylphenol = 8(C) + 10(H) + 1(0) + ring
= 8(16.5) + 10(2.0) + 5.5 — 20.0
= 137.3 cm3/mol

Diffusion coefficients were calculated using equation 6. A sample calculation
for the diffusion coefﬁcient of phenol is shown in Appendix F. Table 4-6 Shows the

molar volumes and diffusion coefﬁcients for phenolic and indolic compounds.

86

Table 4-6. Molecular diffusivity and molar volume for phenolic and indolic

 

 

compounds
Compound ggecctzll: M33211" 7 (cm3/mol) Dg (cmzls)
Phenol C6H60 94 87.96 0.086
p-Cresol C7H30 108 104.5 8 0.079
p-Ethylphenol C3H100 122 13 7.3 0.069
Indole C3H7N 1 17 129.5 0.071
Skatole C9H9N 131 13 1 .8 0.070

 

Gas phase concentrations for phenolic compounds using the amount sorbed on
SPME ﬁbers and sampling time were calculated using equation (5). Physical
characteristics of the SPME ﬁber coatings are: ﬁlm thickness of 75 um, length of 1 cm
and core rod thickness of 0.011 cm. Since commercial SPME ﬁbers have constant
ﬁber length and ﬁlm thickness, equation (5) can be simpliﬁed to equation (8).

Cg = nor/t ............. (8)

Then 0: can be written as equation (9) from equations (5) and (8),

 

Since thickness (b), length (L) are constant, 0: depends on the effective
stationary boundary 8 and molecular diffusivity (Dg). Molecular diffusivity is constant
when the sampling temperature and atmospheric pressure are the same. Then 0:
depends only on effective stationary boundary, 8.

The CI values for phenolic and indolic compounds were calculated using the

known Cg and n at various time t. The dimensionless k values in equation (8) have

87

been calculated using or values. For the calculation of the effective boundary layer 8,

equation (5) has been rearranged as shown below:

”1‘ ......... (10)
2anLt

Since C8 = nor/t (eqn.10), ct = (2anL)k. The k values for the phenolic and

 

When k = In {(b + 5)/b}, equation 5 becomes C8 =

indolic compounds are listed in Table 4-7. A sample calculation for the dimensionless

k value and or value is shown in Appendix G.

Table 4-7. Dimensionless average k values for phenolic and indolic compounds for
SPME CAR/PDMS ﬁber coating.

Coefﬁcient of 95%
Compound D: (cmzls) Average log 01. Average k variance (CV) conﬁdence

 

 

interval
Phenol 0.086 5.1 110027 70.36.354.54 0.0645 5.63
p-Cresol 0.079 5.5910052 193.95i22.7 0.1 172 28.2
p-Ethylphenol 0.069 6.0420061 482.19i62.8 0.1301 62.8
Indole 0.071 5.9710075 420161701 0.1669 87.1
Skatole 0.070 642350088 1 182.53i223 0.1887 277

 

Sampling times for the k values were 30 sec to 150 sec. D3 values were calculated at 20 °C
and 1 atm.

SPME ﬁber coatings are very fragile. When the tip of the ﬁber is worn by
accident, the mass sorbed on the ﬁber coating will be reduced resulting in reporting
less Cg than actual gas concentration in the air. Therefore, k value has to be corrected
with the adjusted ﬁber length to avoid the error.

The response of SPME ﬁber sorption for rapid sampling was tested, and results
are shown in Figures 4-20, 4—21, 4-22, 4-23, and 4-24. The linear regression line for
phenolic and indolic compounds at sampling time from 30 sec to 2.5 minutes are also

shown. The mass sorbed on SPME ﬁber increased linearly as the sampling times

88

increased from 30 seconds to 2.5 minutes for phenolic and indolic compounds at 95%
conﬁdence interval. The masses of indolic compounds sorbed on SPME ﬁber were
less than those of phenolic compounds. This result indicates that the sorption rates for
indolic compounds are less than phenolic compounds resulting in less sorption of
indolic compounds. The mass uptake rates decreased as sampling times increased for
indolic compound, suggesting that the competition with phenolic compounds for
sorption sites on SPME ﬁbers. The effect of molecular weight on sorption rates were
observed in Figure 4-25. Graph shows that the sorption rates decreased exponentially
as the molecular weight of the chemical increased. AS the masses of the chemical
increase, the molecular diffusivities decrease due to the Slower movement to the
sorbate. This result is similar to that previously shown in Figure 4-19.

Using calculated k values (Table 4-7), diffusion coefﬁcient and gas phase
concentrations, the quantiﬁcation of SPME CAR/PDMS ﬁber coatings was conducted.
Five 1 min SPME air samples were taken using SPME CAR/PDMS coatings. Table 4-
8 shows the calculated and measured Cg values. The calculated gas phase
concentration for phenol was 20% less than the actual gas phase concentration. The
calculated gas phase concentrations for p-cresol and p-ethylphenol were 6% and 5%
higher than the actual gas phase concentrations. The calculated gas phase
concentration for indole was 17% higher than the actual gas phase concentration and
the calculated gas phase concentration for the Skatole was 40% higher than the actual
concentration. This error may be induced from the k value. The result suggests that k
values for the indolic compounds are not constant for the time range of 30 seconds to

2.5 minutes. Gas phase concentrations also affect on the k values. In this experiment,

89

relatively high gas phase concentrations were used for the calculation of k values (see
Table 4-8). To adapt this quantiﬁcation method for the ﬁeld samples, k values for the
phenolic and indolic compounds must be re-evaluated using the environmental data

(temperature, atmospheric pressure) and lower concentrations of standard gases.

Table 4-8. Comparison of actual gas concentration and the calculated gas

 

 

concentration.
Compound ;:I::g:gfﬁ CV (3:230 Cg(act)/Cg(cal) 8:33;”
Phenol 1031:0732 0.071 12.84 1.245 0.85
p—Cresol 1312:0572 0.044 12.408 0.946 0.66
p-Ethylphenol 12.97:] .043 0.080 12.296 0.948 1.20
Indole 1440:0666 0.046 12.34 0.856 0.77
Skatole 1730:0928 0.054 12.40 0.717 1.07

 

C8: gas phase concentration (pg/L)
CV: coefﬁcient of variance

The quantiﬁcation method with rapid diffusion based SPME sampling is useful
since the effect of temperature and pressure can be accounted using equations 5 and 6

without making the gas standard calibrations for different sampling environment.

90

Phenol

1 2000

 

10000 4

8000 4

6000*

Mass sorbed (ng)

4000 1

2000*

 

 

 

 

0 I I I I I I

20 4o 60 80 100 120 140 160
Time (seconds)

— Plot1 Regr
— — Plot 1 Conf1

Figure 4-20. The relationship between the sorbed mass and sorption time on SPME for
rapid diffusion based sampling for Phenol. The upper and lower lines represent the
95% conﬁdence interval. The average temperature was 23°C and pressure was 1 atm.

91

p-Cresol

3500

 

3000 ~

2500 -

2000‘

1500 4

Mass sorbed (ng)

1000‘

500 '

 

 

 

 

O I I I I I
20 40 60 80 100 120 140 160

Time (seconds)

— Plot1 Regr
— — Plot 1 Conf1

Figure 4-21. The relationship between the sorbed mass and sorption time on SPME for
rapid diffusion based sampling for p-Cresol. The upper and lower lines represent the
95% conﬁdence interval. The average temperature was 23°C and pressure was 1 atm.

92

Mass sorbed (ng)

1 400

1200 -

1000 *

p-Ethylphenol

 

800*

600 *

400 *

200*

 

 

 

 

o I I I I I

20 4o 60 80 100 120 140 160
Time (seconds)

—— Plot1 Regr
— — Plot 1 Conf1

Figure 4-22. The relationship between the sorbed mass and sorption time on SPME for
rapid diffusion based sampling for p-Ethylphenol. The upper and lower lines represent
the 95% conﬁdence interval. The average temperature was 23°C and pressure was 1

atm.

93

Indole

 

1 600

1400 *

1200 *

1000 *

800 -

600 *

Mass sorbed (ng)

400*

200*

 

 

 

 

0 I I I I I I

20 4o 60 80 100 120 140 160
Time (seconds)

— Plot1 Regr
— — Plot 1 Conf1

Figure 4-23. The relationship between the sorbed mass and sorption time on SPME for
rapid diffusion based sampling for Indole. The upper and lower lines represent the
95% conﬁdence interval. The average temperature was 23°C and pressure was 1 atm.

94

Skatole

 

600

500 ~ //

 

 

 

 

a
5
E
o
in
8
co
2
100 * O / /
/
/
/
/
0 l I I I T I
20 40 60 80 100 120 140 160
Time (seconds)
— Plot1 Regr
— — Plot 1 Conf1

Figure 4-24. The relationship between the sorbed mass and sorption time on SPME for
rapid diffusion based sampling for Skatole. The upper and lower lines represent the
95% conﬁdence interval. The average temperature was 23°C and pressure was 1 atm.

95

 

4500

4000.. O

3500 ~ y = 1E+07e‘°'°8“"

2 _
3000 R - 0.9924

2500 *
2000
1500 4
1000

Initial sorption rate (nglmin)

500 4

 

 

 

 

90 100 110 120 130 140
Molecular weight (g) .

Figure 4-25. Effect of molecular weight of chemicals on SPME sorption rate.
Sampling times were 5 seconds to 2.5 minutes. The average temperature was 23°C
and pressure was 1 atm.

96

4.5.5 Analysis of phenolic and indolic compounds in swine building air
with SPME

Two types of SPME ﬁber coatings were used for the sampling of phenolic and
indolic compounds in swine house. Samples were taken in the MSU Swine Teaching
and Research Farm. Two sampling times, 1 min and 10 min were used for this
experiment. Initial results for the quantiﬁcation of SPME samples with a 30 minute
sampling time showed that the adsorption amounts and gas phase concentrations for
phenolic and indolic compounds did not show the linear relationships. Therefore, a
shorter sampling time, 10 minutes, was used for the sampling of the phenolic and
indolic compounds with CAR/PDMS.

Figure 4-26 shows the Chromatogram of air samples which were taken in the
swine building using SPME CAR/PDMS ﬁbers for 1 and 10 min. The Chromatogram
shows that the analytes adsorbed on SPME coatings were same but the amount
adsorbed on the ﬁber was greater for 10 minute sample. Results showed that the
recoveries of phenolic and indolic compounds from swine building air were small with
SPME CAR/PDMS coatings for both 1 min and 10 min samplings. The majority of
compounds adsorbed on SPME CAR/PDMS coatings in swine building were volatile
fatty acids (V FAs).

Table 4-9 shows the amount of VFAS adsorbed on SPME ﬁbers when it was
sampled for 1 minute and 10 minutes. The amounts adsorbed on SPME ﬁbers for 10

minute sample were about 10 times greater than 1 minute sample.

97

Table 4-9. Amount of VFAs adsorbed on SPME ﬁbers for 1 minute and 10 minute
sampling using PDMS/CAR ﬁber coating.

 

 

Ratio
Compound 10 min 1 min
(lmin/lOmin)
Acetic acid 171.23 24.39 0.14
Propionic Acid 120.77 13.58 0.11
Butyric acid 80.64 9.06 0.11
Valerie acid 27.35 4.47 0.16

 

4.5.5.1 Rapid air sampling with SPME CAR/PDMS

Quantiﬁcation of the phenolic and indolic compounds in the swine building air
was conducted using SPME CAR/PDMS ﬁbers. The dimensionless values, k, for the
phenolic and indolic compounds were used for the calculation of the gas phase
concentrations. Calculated concentrations for phenolic and indolic compounds are
listed in Table 4-10. The calculated concentrations for phenolic and indolic
compounds fell within the concentration ranges found in previous research (O’Neill

and Phillip, 1992).

Table 4-10. Quantiﬁcation of the phenolic and indolic compounds in swine building
air with rapid SPME air sampling method

 

 

Compounds Avg. k t (sec) D, (cmzls) C‘ (mg/m3)
phenol 70.414.54 60 0.086 0.1 64:0.01 l
p-cresol 194122.73 60 0.079 039710.047
p-ethylphenol 482:62.75 60 0.069 043910.064
Indole 420170.13 60 0.071 0.23 710.040
Skatole 1 1801223 60 0.070 021910.041

 

98

4.5.5.2 Comparison of SPME fiber coatings.

To observe sorptive abilities of these compounds on different types of ﬁber
coatings, two types of SPME ﬁber coatings, carboxene/PDMS (CAR/DPMS) and
PDMS/divinylbenzene (PDMS/DVB), were used for the adsorption of phenolic and
indolic compounds in swine building air. Figure 4-27 shows the chromatograms for air
samples that were taken on SPME CAR/PDMS and PDMS/DVB coatings.
Chromatogram shows that VFAS favored CAR/PDMS ﬁbers and phenolic and indolic
compounds favored PDMS/DVB ﬁbers. Table 4-11 shows the amount of phenolic and
indolic compounds which were adsorbed on two SPME ﬁbers. The amounts of VFAs
adsorbed on two types of ﬁbers are listed in Table 4-12.

This result suggests that using these two types of SPME ﬁbers would be useﬁlll
for the sampling of VFAS and phenolic compounds in the livestock building air since
PDMS/CAR favors VFAS and PDMS/DVB favors phenolic compounds. As part of
this research, the quantiﬁcation of gas phase concentration using PDMS/DVB ﬁbers
was not determined.

Table 4-11. Comparison of the adsorbed amounts of phenolic and indolic compounds

on two different SPME ﬁbers. (Sampling time for both was 10 min. Unit for the
amount adsorbed is ng.)

 

 

Compound CARIPDMS PDMSIDVB Ratio
(CARIDVB)
Phenol 165 231 0.72
p-Cresol 219 993 0.22
p-Ethylphenol 35.3 122 0.29
Indole 24.9 51.4 0.49
Skatole 6.75 43.2 0.16

 

99

 

Table 4-12. Concentration of VFAS in swine building air sampled on SPME ﬁbers.
(VFA concentrations are in pg as carbon)

 

 

Compound CAR/PDMS PDMSIDVB Ratio
(CARIDVB)

Acetic acid 68.5 6.30 10.9
Propionic acid 58.8 16.5 3.56
Butyric acid 2.70 1.88 1.43

lso butyric acid 41.3 26.4 1.56
Valerie acid 2.87 3.05 0.94

150 valeric acid 13.2 16.5 0.80
Total VFAs (pg as C) 187 70.6 2.65

 

100

l Plums/okéflorhim

 

 

1 ' ‘ . . ‘.
l ‘ ‘ l " 1 . ‘ '. J :1 ‘
1‘ 7,1, I 1 _‘ .' ‘. _‘,‘ ~ -..' h. “I . \\ ‘ ‘1‘ . VA _ |-J k‘u‘” --. ,_,_lll‘..‘J. .1 ‘L 7’ -,_, VJ(.,J.i:L.- 7 I _ ‘4

PDMS/CAR-1min*

 

 

 

 

Figure 4-26. Chromatograms of air samples sampled on CAR/PDMS ﬁber coatings for
1 and 10 minutes in swine building.

101

T— A, A. ~__ 1__ .__.__._.- _. .— .14. ﬁ __ _ ___._ _#1___‘A.- . 7_ . 7

1
i , PDMS/DVB :

 

 

l
J 3 PDMD/CAR
l
l

 

 

 

 

 

 

Figure 4-26. Chromatograms of air samples sampled on two different SPME ﬁber
coatings. (Sampling time was 10 min for both ﬁbers and sampling was done at the
same location.)

102

4.6 Conclusions

Major odorous compounds in swine building air were sampled with SPME
CAR/PDMS coatings in a laboratory experimental system. The results showed that the
masses of phenolic and indolic compounds sorbed on SPME ﬁber coatings increased
as sorption times were increased, but the sorbed amounts did not reach equilibrium
even when the sampling time was 420 minutes. The amounts of phenolic and indolic
compounds adsorbed on SPME CAR/PDMS coatings are linearly increased as the
sampling time increased up to 30 minutes. When the gas phase concentrations were
reduced by a factor of 10, the masses of sorbed on SPME ﬁber could be described by a
linear relationship as the sampling time increased up to 30 minutes, but the SPME
extraction proﬁle was different from the extraction proﬁle for higher concentration.
Pawliszyn (1997) showed that the sample concentration has no affect on the
concentration time proﬁle. When the extraction method and distribution constants
between the SPME and sampling system remain constant, the equilibrium proﬁle and
time should be the same for all other concentration. Therefore, the extraction behavior
should be linear for all concentration ranges. However, the results showed that the
extraction behaviors for phenolic compounds on SPME ﬁbers were not same for the
gas phase concentrations at 0.4 and 4 ug/L.

These results suggest the possible systematic experimental errors. It is
possible that the errors resulted from a depletion of available sorption sites on SPME
ﬁber. To avoid this problem, calibrations for all individual SPME ﬁbers would be
required. As a SPME ﬁber ages, the sorption ability of the ﬁber is reduced. Therefore,

usages and sampling histories for all SPME ﬁbers should be carefully recorded.

103

Room temperatures ﬂuctuated from 223°C to 243° during the experiment,
and temperature in the sampling bottle was not controlled and the temperature changes
were ignored because the difference was approximately 2 °C. A previous study
showed that the increase in the sample temperature increased the analyte concentration
in the headspace, thereby providing faster extraction (Wercinski and Pawliszyn, 1999).
This error can be corrected by sampling SPME ﬁbers at several different temperatures.

Experimental error may be introduced in preparing the standard. To avoid this
error, the new standard must be calibrated prior to being used for the gas standard.

Previous studies have shown that humidity effect on SPME sampling was
minor. The mass loading on SPME ﬁber was reduced about 10% when the relative
humidity was greater than 95% (Martos and Pawliszyn, 1997). Therefore, humidity
effect on sorption on SPME CAR/PDMS coating was not considered in this research.

Difﬁlsion based rapid air sampling with SPME CAR/PDMS coating was
conducted using the phenolic and indolic compounds. The results showed that the
mass sorbed on SPME ﬁber coatings was linearly dependent on sampling time when
the sampling time was less than 150 seconds. With the data obtained, quantiﬁcation of
the phenolic and indolic compounds in swine building air was conducted. The
calculated concentrations for the phenolic and indolic compounds were 0.042, 0.089,
0.096, 0.49 and 0.040 ppm for phenol, p-cresol, p-ethylphenol, indole and Skatole,
respectively. These concentrations fell within the concentration ranges found in
previous research (O’Neill and Phillip, 1992).

Field samplings with two different SPME ﬁbers showed that CAR/PDMS

coatings favored VFAS and PDMS/DVB coatings favored phenolic and indolic

104

compounds in swine building air. This result suggests that these two ﬁbers can be
useful for analysis of odorous compounds in livestock building air. For the
quantiﬁcation of VFAS in livestock building air, further investigation is needed.

As shown in previous chapters, SPME was the most promising air sampling
method for the detection of odorous VOCs from the livestock building air. In spite of
the difﬁculty on quantiﬁcation of VOCs on SPME, still it has more advantages on air
sampling than other techniques. Experiment with SPME PDMS/CAR ﬁber for air
samplings showed that the mass sorbed on SPME ﬁber can be relatively easily
quantiﬁed using the two theories: equilibrium and non-equilibirum theory. Some
systematic errors were observed during the experiment. This study suggested that
some modiﬁcations are required for the accurate quantiﬁcation of VOCS on SPME
ﬁber. The modiﬁcation includes the standard gas generator, temperature controlled air
sampling systems and individual re-calibration of SPME ﬁbers whenever it is used for
the sampling. This study also showed that the SPME ﬁbers can be used for the
sampling of phenolic compounds in livestock building air using the rapid sampling
method. Once the diffusion coefﬁcient, environmental conditions (temperature and
atmospheric pressure), and the dimensionless calibration factor (k), the analysis of the
chemical sorbed on SPME ﬁber can be calculated using the data without further

calibration steps.

105

CHAPTER 5. OZONATION EFFECT ON ODOR AND
VOLATILE ORGANIC COMPOUNDS IN SWINE
PRODUCTION FACILITY.

5.1 Introduction

Ozone is a powerful oxidizer. It has been widely used for disinfection and
oxidation for domestic wastewater, water and industrial wastewater e. g., (Kruithof and
Masschelein, 1999; Lezcano et al., 1999; Carini et al., 2001; Finch et al., 2001; Turan-
Ertas, 2001). Ozone is able to oxidize odorous volatile compounds such as; phenol, p-
cresol, p-ethylphenol, indole and Skatole, in swine slurry and reduce malodors in
stored swine slurry by oxidation of odorants (Wu, 1999).

In a recent review paper, “Century 21 - Pregnant with Ozone”, Rice (2002)
stated that

“Many odors can be destroyed by ozone — e. g., hydrogen sulfide, odors caused
by cigars and cigarettes, and volatile organic camounds (V OCS), perspiration odors,
odors in animal rearing facilities, etc. ” This paper summarized the previous research
on the use of ozone for odor control and included four conclusions. First, the oxidation
of odorants can occur when ozone is allowed to the odorants for sufﬁcient contact
time. Second, the gas phase oxidations are relatively slow compared to the oxidation
in aqueous phase. Third, residual ozone must be destroyed before it is discharged in to
the atmosphere. And ﬁnally, ozonation must be done when human beings are not in
the facilities.

Ozone has been used widely for the indoor air control in recreational areas,

such as casinos, ofﬁces and hotel. A study by Kilham and Dodd (1999) showed that

106

 

ozone could be used to remove offensive odors and to destroy volatile organic
compounds and smoke. Ozone was introduced in the air conditioning system of the
building. The effect of ozone on a subset of volatile organic compounds (VOCS) found
in tabacco smoke has been examined (Shaughnessy et al., 2001). Shaugnessy et al.
showed that the ozonation was not effective at reducing the concentrations of the
saturated VOCs, however ozone greatly reduced the concentrations of the compounds
with unsaturated carbon bonds. The effect of ozonation on sub-micron particles in
ofﬁce building have been studied (Weschler and Shields, 1999). Terpenes (limonene,
alpha-terpene, or a terpene-based cleaner whose major constituent is alpha-pinene)
were introduced into the ofﬁce and the subsequent particle formation and
redistribution were monitored. This study showed that ozone/terpene reactions were
the signiﬁcant source of sub-micron particles in the ofﬁce building.

Ozone has been used to remove hydrogen sulﬁde in the presence of concurrent
substances, such as toluene, ethanol, and n-butanol (Masuda et al., 2001) with a
combination of an activated carbon system. Masuda et al. (2001) showed that the
ozone was effective to increase the removal efﬁciency for hydrogen sulﬁde.

Ozonation in livestock production units has been studied by several
commercial farms and University researchers (Edie, 2001; Vansickle 2002; Ozone
Solutions Inc.; 2002). A report from Picket Fence Farms Inc., ozone decreased the
number of Ecoli by 75%, reduced by 50% in number of pigs laid on, and increased
pigs weaning weight by the 15% (Ozone Solutions Inc. 2002). According to this
report, the sows consumed up to 20% more feed because of better air quality. A

recent study showed that that ozone reduced the concentration of low molecular

lO7

 

weight volatile fatty acids on the dust samples taken in a swine house (Oehrl et al.,

2000)

5.2 Hypotheses

1) Ozone will reduce the concentrations of major odorous compounds in the
swine building air.

2) Ozone will improve the odor in swine building air.

5.3 Objectives

This research was conducted to observe the air quality changes that resulted
from the ozonation of air in livestock buildings. The most widely reported malodorous
volatile organic compounds associated with livesotock operations i.e., VFAS and
phenolics, were monitored. The experiment was conducted in rooms where the air was

ozonated and in a control room where the air was not ozonated.

5.4 Materials and Methods
5.4.1 Swine Housing

Four environmental rooms, located at Michigan State University Dairy
Teaching and Research Farm, were used for this research. The layout of the facilities
is shown in Figure 5-1. Each room contained one 12’ by 16’ pen. Each pen contained
24 pigs. Mortalities were 0% for control, low, and medium ozone injection. The
mortality in the pen with high ozone injection rate was 8.2%. The mortalities in this
pen were not related to ozone treatment but it was related to a hip injury, inﬂicted by
another pig. A manure handling and storage pan was located directly beneath the

ﬂooring of each pen. The air samples were taken 1m above the ﬂooring.

108

Temperature was measured using a digital thermometer (copper constant
thermocouple) and a data-logger (CSI 23-X). Ventilation airﬂow rates were
determined based on air velocity traverses across exhaust fan outlet cones. The air
velocity measurements were made using a vane anemometer. Two
heating/ventilation/air conditioning (HVAC) systems were installed in the facility and
each HVAC controls air exchange for two chambers. The east HVAC controlled

rooms 1 and 2, and west HVAC controled rooms 3 and 4.

5.4.2 Ozone Distribution

Figure 5-2 shows the ozone distribution in the rooms. Ozone was injected at
the rates of 0, 2, 4, and 6 ft3/hr to provide the ozone dose of approximately 0 (control),
0.01 (low), 0.05 (medium), and 0.1 (high) ppm, respectively. Ozone was distributed
through PVC tubing to the rooms. Every group of 24 pigs was subjected to a ﬁxed
level of ozonation, each group of pigs was moved every two weeks to another
randomly selected pen. The ozone concentration in the ozonated rooms was measured

every day.

109

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 5-1. Environmental chambers at Michigan State University.

110

3
"8 Valve &
g ﬂow
§ meter
. O O . . . .
o 03 . o ()3 03 o
. o ' o .
O
Room 1 Room 2 Room 3 Room 4
High Medium Low Control

 

fan

 

30”

 

 

 

Ventllatlon (1n) .» "i; jam-43:3?

Ventilation (out)

 

Pen size: 12’ x 16’

Figure 5-2. Diagram of one pen in the environmental chamber at MSU.

lll

5.4.3 Air Sampling

The experiment began on July 15’“, 2001 and was terminated on September
309,2001.

Two schedules were used for the air samplings during the 15-week experiment.
Air samples were taken weekly. The objective of this experiment was to monitor air
quality over the 15 weeks experiment. Sampling was commenced on every Sunday at
4 PM. Samples were collected on thermal desorption tubes, Solid Phase
Microextraction (SPME) ﬁbers, and glass ﬁber ﬁlters.

On a biweekly basis, sampling was conducted over a 24 hour period. This
sampling was conducted to monitor diurnal changes in air quality. In this experiment,
thermal desorption tubes and SPME ﬁbers were used for sampling. Sampling was
started at noon and was repeated every four hours and ﬁnished at noon of the next day.

Custom designed thermal desorption tubes were purchased from Supleco Co.
(Bellefonte, PA). Tubes contained Tenax TA plus Carboxen and Tenax TA plus
Carbosieve 8111. The air ﬂow rate was regulated between 150 to 200 mL/min and the
total sample volume did not exceed 5 L. The tubes were conditioned as described in
section 2.1.1.

SPME PDMS/CAR ﬁbers were used to sample the air. Sampling times were
varied according to the research objectives. Fiber conditionings and sampling methods
are described in section 2.1.3.

For the sampling of dust, a 37mm cassette ﬁlter holder (225-3, SKC, Inc.) with
a membrane ﬁlter (pore size of 1.0 pm, SKC Inc. 226-7) and a support pad (Metricel,

Gelman Lab.) were used. The sampling method is described in section 2.1.4.

112

5.4.4 Test mixture

Phenol, p-cresol, p-ethylphenol, indole, and Skatole and VF As (acetic,
propionic, butyric, is-butyric, valeric, and iso-valeric acids) were studied in this
research. A stock solution for phenolic compounds was prepared in methanol.
Analytical grade phenol, p—cresol, p-ethylphenol, indole and Skatole were purchased
from Sigma-Aldrich Co. (St. Louis, MO). The standard VFA mixture was purchased
from Supelco Co. This standard mixture contained 10 mM of acetic acid, butyric acid,
formic acid, heptanoic acid, hexanoic acid (n-Caproic acid), isobutyric acid, isocaproic
acid (4-Methyl-n-valeric acid), isovaleric acid, propionic acid, n-valeric acid, in 100

mL deionized water.

5.4.5 Quantiﬁcation of the compounds.

The column used in the research was DB-SMS (J&W Scientiﬁc) which is non-
polar column. The dimensions of the column are given in Section 2.2.2. The
separation of phenolic compounds with this column was good. However, for the VFA
compounds, some isomers, such as valeric and iso-valeric acids were not well
separated using this column. Therefore, the amount collected on SPME and TD tubes
for VFAs was converted to carbon, and added all VFAS and ﬁnally reported as total

carbon for VFAS.

5.4.6 Desorption methods

Thermal desorption was used for the removal of VOCS from the sorbent tubes.
The conditions used are described in section 2.2.1.
Dust particles on ﬁlters were analyzed using thermal desorption. Since the

conventional solvent desorption method dilutes VOCS during the desorption processes

113

 

from ﬁlters, it lowers the sensitivity of the analysis. Therefore, thermal desroption
method was used for desorption of VOCS from the ﬁlters.

For the analysis of VOCs on a ﬁlter, the ﬁlter was placed on an empty W’
stainless steel TD tube. The desorption processes and conditions were same as

described in section 2.2.1.] .

5.4.7 Gas chromatography and mass spectrometry

GC/MS was used for the separation and identiﬁcation of the compounds in the
air samples. The analytical conditions for TD/GC/MS and SPME/GC/MS are

described in the section 2.2.2.

5.4.8 Odor measurement: Olfactometry

The olfactometric measurements were conducted at Agricultural Air Quality
Laboratory at Purdue University. Air samples for the odor tests were collected in a 1L-
Tedlar bag. Tedlar bags were purchased from SKC Inc. (Eighty Four, PA). All new
bags were purged with high purity air three times prior to the sampling. An air
sampling pump (SKC Inc.) was used for the air sampling in ozonated rooms and
control rooms. Inlet air samples were taken at the outside of the environmental
chambers at the inlets of the two heating/ventilation/air conditioning systems.

The air samples were shipped overnight to Purdue University for analysis by a
trained odor panel. The odor analysis was conducted the day following sampling.

Odor intensity and offensiveness measurements were made at recognition in
which the odor was ﬁrst detected, and at full strength in wich the air samples were
delivered to the olfactometer without mixing with dilution air. Data in the results

section shows the offensiveness and intensities at full strength. Intensity is numerically

114

recorded. A larger number means the odor is more intense. Odor offensiveness was
recorded as negative numbers. Larger negative values mean that sample was more
offensive. Dilution to threshold is the ratio of the dilution air volume to sample
volume when odor is ﬁrst detected by the panelists. A large dilution threshold (DT)
means that odor was detected when large volume of dilution air was used. This means

that the sample has low odor threshold.

115

5.5 Results
5.5.1 Air ventilation data

Air velocity traveling across exhaust fan outlet was measured using a vane
anemometer. The air ventilation rate was calculated using the cross sectional area and
the air velocity. The calculation is shown in the Appendix A. Figure 5-3 shows the air

ventilation rate which were averaged every four hours.

1400

1350 . I'll":

 

1300 1

"‘1."

1250 - I ’
_ . l
1200 L"l ill 'lh‘ 1i”

1150

Ventilation rate (cfm)

 

 

 

1100 1 .
7/23 8/12 9/1 9/21 10/11 10/31

Date
East - - - West

 

Figure 5-3. Air ventilation rate in the swine building. The data is shown for the east
and west fans. East fan controls the air exchange in rooms 1 and 2 and west fan
controls air exchange in rooms 3 and 4.

116

5.5.2 Temperature

Figure 5-4 shows the temperature in the four rooms. Temperature was
measured in two locations in each room and the ﬁgure shows the average temperature
in each room. Temperature readings were recorded every 5 minutes. In this ﬁgure, one
hour average temperature values were used.

Figure 5-5 shows the temperature changes and ventilation rates in room land 2
from July 31St to September 28’“. The graph shows that the ventilation rates were

increased as the room temperature increased.

100

 

95*

90*

85*

80*

75*

Temperature (F)

 

70*

65*

 

 

 

 

60 ., 1 T . .
6/15 7/5 7/25 8/14 9/3 9/23

Figure 5-4. Temperature changes in the environmental chambers.

117

 

 

 

 

 

 

 

86 1350
34
l; 1300
82 , ill
80 *- l 1250 ..
78 __ ' (1 °
g l l l i" 1 ‘1 ~ I
‘5 75 ~ .. l 1 ll , “120° ‘5
3- 74 l l 5
g , . 1 1150 g
’- 72 —1i- g
70 1- -- 1100 >
68 ~
1-1050
66 .. Room 1 —-—Room 2 Vent
64 4 1 1 1 1 1 1000
7/31 8/9 8/17 8/25 9/4 9/12 9/20 9/29

Date

Figure 5-5. Temperature and ventilation rate in swine building (Room 1&2).

118

5.5.3 Olfactometry data

Four measurements were made for the Olfactometry tests: dilution threshold,

offecsiveness, intensity, and characteristics.

5.5.3.1 Dilution threshold (DT)

The data in table 5-1 show that the detection threshold was higher for the
control room and that the DTs decreased as ozone level increased. The effect of ozone
on odor dilution threshold was highly signiﬁcant at p<0.01 for all ozonated rooms.
Odor dilution threshold in the highly ozonated room was signiﬁcantly different from
those in low ozonated room at 0.02 while there were no differences in odor DTs
between the medium dosage and low dosage ozonated room (0.53), and between
highly and medium ozonated rooms (0.15). These results indicate that no signiﬁcant
additional beneﬁt as to the reduction of odor DTs in highly ozone dosage room as

compared to the medium dosage.

Table 5-1. Detection Threshold for air samples in swine building.

 

 

Date Inlet High Ozone Medium Ozone Low Ozone Control NB
7/6/01 46 327 386 621 1090 1510
8/3/01 48 382 683 683 808 1580

8/31/01 63 676 866 799 1110 1430
9/28/01 39 317 622 734 944 1030

 

*NB: normal butanol

5.5.3.2 Odor offensiveness and intensity

Figure 5-6 shows the odor offensiveness for the air samples. Ozone did not

signiﬁcantly reduce the odor offensiveness in swine building air at P>0.5 for all

119

treatments. A study by Hargesheimer (1996) supports this ﬁnding. Hargesheimer
(1996) showed that ozone did not remove bad odors in drinking water while it altered
the ﬁsh odor to an undesirable “plastic-like” odor.
The result suggests that the VOCs that affect offensiveness are still present
after ozone treatment or the by-products by the ozonation have an offensive odor.
Figure 5-7 shows the odor intensity in swine building air in the control and

ozonated rooms. Again, ozonation had no signiﬁcant effect on odor intensity.

 

  
     
     
     
     
     

  

 

 

-10
-9
-8
3 -7
a ,.
'17: ‘5 ”:1
5 4 {:1
a: -3 91111
o 1:11
-2 Fig
.111:
0 i113
Inlet High ozone Medium ozone Low ozone Control
a e
E] 7/6/2001 [3 8/3/2001 I 8/31/2001 I 9/28/2001

Figure 5-6. Effect of ozonation on odor offensiveness in swine building air. Samples
were taken at noon.

120

 

Odor intensity

 

 

 

Inlet High ozone Medium ozone Low ozone Control
Date
El 7/6/2001 El 8/3/2001 I 8/31/2001 I 9/28/2001

Figure 5-7. Effect of ozonation on odor intensity in swine building air. Samples were
taken at noon.

5.5.3.3 Odor characteristics

Table 5-2 shows the odor characteristics of the air samples from the control
and ozonated rooms. Both ozonated and control room had odor characteristics of
manure and urine. The highlighted data indicates the unique smells in ozonated rooms
which are different from the odor characteristics in control room. The data indicate
that the ozonated rooms had slightly different odor characteristics from those for the
control room. Air in the ozonated room was reported to have a greasy, soapy, sour,
mossy or musky odor, whereas the control room did not show such odor
characteristics. Soapy smell was reported in all ozonated rooms. Acidiﬁed smells,
such as sour, sauerkraut, vomit smells were reported in low, medium, and high ozone

level room, respectively. As ozone level increased the intensity of the acidiﬁed smell

121

increased. This ﬁnding suggests that ozone may produce the acidiﬁed smells in swine

building air.

5.5.3.4 Olfactometry summary

Olfactometry data showed that the ozone was effective in the reducing odor
dilution threshold, however ozone did not signiﬁcantly reduce the odor offensiveness
and intensity. Odor characteristics of the ozonated air were different from those of the
air in the control room. These results suggest that some odorous compounds were
destroyed by ozone but the residual compounds still exhibited a strong and
offensiveness odor. Another possibility is that the by-products formed during
ozonation had an offensive odor. A previous study for ozonation on odor and taste in
drinking water has shown that ozone altered odor characteristics in drinking water but
it did not reduce odor intensity (Hargesheimer and Watson, 1996). A study by Esswein
and Boeniger (1994) showed that ozone was not effective in reducing formaldehyde
where formaldehyde is the main odor producer. A publication from US EPA also
showed that the ozone was not effective in reducing odor causing chemicals at

concentrations below the public health concern (IOA Publications, 2002).

122

Table 5-2. Odor characteristics for air samples in swine housing air.

 

 

 

 

 

 

 

  

 

 

 

 

 

Date Inlet 6cft 4cft 2cft 0cft
7/6/01 Musty Mildew Mildew Sewage Manure
Fresh air Mold Mold Lightly bitter Rancid
Salty Foul Foul ari Salty Foul
Plastic Earthy Greasy Fecal Fecal
Corn Greasy Fecal Burnt oil Salty
Dust Swamp Sauerkraut Corn Blood
Musty Urea Com Dust Urea
Fetid Manure Dust Cadaverous Manure
Smoke Smoke Smoke slip Septic
$11151 Urine Spicy
Vomit Pig feces gemmm
8/03/01 Chemical Sewer Cat urine Cat urine Cat urine
Septic Pungent Fecal Fecal Dirt
Spicy Decayed Dirt Foul Foul
Foul Dirt Foul Earthy Blood
Rotten Spicy Spicy Pungent Fecal
Solvent Foul Fecal Fecal Salty
Dust Rotten meat Decayed Rotten egg Rotten egg
Phenol Urine Rotten egg Salty Sewer
Cork H28 Decayed Garbage Swamp
Hay Fecal Leaves Sewer Ammonia
Smoked Smoke Musty Raw meat Rotten leaves
leaves Cigarﬁ Féitf , Rotten egg
8/31/01 Grass Mold Grass Grass Sewer
Dirt Rotten Dirt Dirt Moldy
Grain Urine Grain Grain Sour
Mold Moldy Moldy Sour Grain
Chemical Pungent Dust Ferment Rotten eg
New plastic Rotten egg Spicy Foul Foul
H28 Feces Pungent Pungent Ammonia
Feces Salty Urine Spicy Salty
Smoke Garbage Ashes Ashes Feces
Hay wood Rotten meat Spicy Feces Smoke
Animal Acidic Foul Rotten leaves
Farmhouse Garlic
. - . -:, Rotten meat
9/28/01 Fresh Foul Rancid Rancid
Plastic Pungent Fecal Bitter
Foul Musty Pungent Pungent Fecal
Septic Stale Urine Urine Urine
Wood Septic Stale Septic Foul
H28 Sauerkraut Cat urine Grassy Pungent
Earthy Cat urine Fresh cow Musty Sewer
Garbage Fresh cow manure manure Cat urine Anesthetic
Musk Fresh cow manure Cat urine
Fecal Strong fecal Fresh cow
manure

 

 

 

 

 

 

123

 

5.5.4 Ozonation effect on Volatile Organic Compounds in Swine
Production Unit.

As described in the method section, three air sampling methods (SPME, TD,
and ﬁlter) were used. For each sampling method, two groups of compounds were

analyzed; phenolic compounds and VFAS.

5.5.4.1 SPME air sampling
5. 5. 4. 1.1 Volatile Fatty Acids

Short chain VFAS (C2, C3, iso-C4, C4, iso-CS, and C5) were determined.
VFAs adsorbed on SPME ﬁber coatings were analyzed and the individual VFA
concentrations were converted to a carbon basis, added and reported as total carbon.
The data shown in the Figure 5-8 shows the total VFAs recovered from SPME ﬁber.
Ozone appeared to have no effect on total VFA concentrations in the swine building
air. T tests for control and ozonated rooms showed the ozonation effect was not
signiﬁcant at the 95% level. Figure 5-8 and 5-9 show the ozonation effect on VFAS for
the weekly and bi-weekly air samples, respectively. The VFA concentrations appeared
to decrease over time. This result suggests that generation of VFAs decreases, as pigs
get older. However, the effect of pig ages on VFAs generation has not been well
documented.

There was no co-relationship between VFAS and room temperature.

124

 

 
       

 

 

 

 

 

 

 

 

 

 

 

E 120
f!
5’ 100
.9
g 80’
C
0
1% 60~
a 40* I I
s 111 "11 u 111
> 2° II I | i "'
79 I I?!
Imo- o“ 111 x.
7-8 7-15 7-24 7-29 8—12 8-19 8-26 9-2 9-9 9-16
Date
El Control 1:] O-2cft I O-4cft I O-6cft

Figure 5-8. The total VFA concentrations adsorbed on SPME ﬁbers for weekly

 

 

 

 

 

 

 

 

 

 

 

samples.
5 120
3100.
.2
E 80~ i
C
3 60
8 1:
g 40. 1 1
< .
LL if}:
2 201
6 O _ 7 ’T— . ... -. ‘.‘.".- I 3|
'- 7-3 7-19 8—2 8-16 8-29 9-11 9-16

Date
D Control El O-2cft I O-4cft I O-6cft

Figure 5-9. The average VFA concentration adsorbed on SPME ﬁbers for bi-weekly
samples.

125

5. 5.4. 1.2 Phenolic compounds in swine building air.

Figure 5-10 shows the average phenol levels for the bi-weekly air samples.
Seven phenol levels for a day samples were averaged. Results showed that ozonation
signiﬁcantly reduced the phenol level in bi-weekly samples at 95% conﬁdence level
(<0.001). Phenol level was lower in the highest ozone treatmed room than in the other
rooms (<0.02) but there were no difference in phenol level between the low and
medium ozone treated rooms at 0.53. Figure 5-11 shows the GC response for phenol
for weekly samples. This result was obtained from the weekly samples that were taken
at the same time (4:00 PM) on every Sunday. Result showed that ozonation did not
change the phenol level in weekly samples (>0.4). Ozone was effective at reducing
overall phenol levels in ozone treated rooms but it did not reduce the phenol level in
the weekly samples. This result implies that there were diurnal effects on phenol level
in bi-weekly samples.

Figure 5-12 and 5-13 show the diurnal variations in phenol level in swine
building air in July 19‘h and August 2"d 2001. Figures show the phenol level changes
over time, however there were no signiﬁcant ozonation effect on phenol levels.

The average mass of phenol adsorbed on SPME ﬁber for the whole test period
were 125:15, 109i2l, 1071-10, and 98:8 ug for control room, ozonated rooms with
low, medium, and high ozone dose respectively. The average percent reduction of
phenol in ozonated rooms were 12, 14, and 21% for the rooms with ozone treatment of

low, medium and high ozone dose, respectively.

126

 

180

160
140 i
120

100
80
60
40
20

 

Mass (n9)

      

 

 

 

 

 

 

 

1 ‘ 1 4
7/5 7/19 8/2 8/16 8/29 9/11 9/26
Date
[3 Control 121 Low Ozone I Medium Ozone I High Ozone

Figure 5-10. Average Phenol level in bi-weekly air samples. (Sampling method was
SPME)

160
140

 

Mass (n9)

 

 

 

 

7/8 7/15 7/24 7/29 8/12 8/19 8/26 9/ 9/9 9/16
Date
El Control El Low Ozone I Medium Ozone I High Ozone

Figure 5-1 1. Phenol level in weekly air samples. (Sampling method was SPME)

' 127

160

 

140

1 00
80

Mass (ng)

60
40
20

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12 16 20 0 4 8
Time (hr)

El Control El Low ozone I Medium Ozone I High ozone

Figure 5-12. Diurnal variations in phenol level. (Sampling was done in July 19th
2001.)

 

Mass (ng)

 

 

 

 

 

 

 

 

 

 

 

Time (hr)

El Control 1:] Low ozone I Medium Ozone I High ozone

Figure 5-13. Diurnal variations it; phenol level. (Sampling was done in August
2“ 2001.)

128

Figure 5-14 shows the averaged p-cresol concentration in the air of the swine
building. T test results showed that ozone did not signiﬁcantly reduce p-cresol level in
the low ozone treated room (0.14), but the ozonation signiﬁcantly reduced the p-cresol
level in medium and high ozone treated rooms at P values of 0.002 and 0.0002 for
medium and high ozone treated rooms, respectively. Figure 5-15 shows the p-cresol
level in weekly samples. The results showed that ozone did not reduce p-cresol in
swine building air. Figure 5-16 shows the diurnal variations on p-cresol levels in
swine building air sampled on July 19th 2001 for 24 hour sampling period. Ozone
reduced the p-cresol level in the highest ozone treated room (0.002) but there no
ozonation effects on low and medium ozone treated rooms. The averaged data showed
that p-cresol concentrations for the whole experiment period were decreased in

ozonated rooms. The averaged mass of p-cresols were 287i98, 247:103, 217i72, and

198191 for control room, ozonated room with low, medium, and high ozone dose,
respectively. The percent reductions in the average p—cresol concentrations in ozonated
rooms for the whole experiment period were 14%, 28%, and 41% for the ozonated
rooms with low, medium, and high ozone dose, respectively.

To see the effect of sampling time on VOC concentration, p-cresol levels for
the same sampling times during the whole experiment were averaged. Figure 5-17
shows the averaged p-cresol concentrations at different sampling times. T test results
(at 95% CI) showed that sampling time of the day did not signiﬁcantly affect on p-
cresol levels in both ozonated rooms and control room. Similar results were shown in

Figure 5-12 and 5-13. Ozone was effective at reducing the average phenolic

129

concentrations in ozonated rooms, however the sampling time of the day did not affect
on phenolic concentration.

All lighting in the swine rooms was from artiﬁcial lights. They were either on
or off. Pig activity is higher when the lights are on, and pig activity causes gases that
can react with ozone to be released. Figure 5-18 shows relationship between lights
being on and p-cresol concentration in the air. Lighting did not signiﬁcantly affect on

p-cresol level in swine building air.

 

fl.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7/5 7/19 8/2 8/16 8/29 9/1 1 9/26
Date

(:1 Control 1:] Low Ozone I Medium Ozone I High Ozone

Figure 5-14. Average p-Cresol level in bi-weekly samples. (sampling method was
SPME)

130

 

Mass (ng)

 

 

 

 

Date
El Control El Low Ozone I Medium Ozone I High Ozone

Figure 5-15. p-Cresol level in weekly samples. (Sampling method was SPME)

 

Mass (ng)

 

 

 

Time (hr)

CIControl 1] Low ozone I Medium Ozone I High ozone

Figure 5-16. Diurnal variations in p-Cresol level. (Sampling was done in July 19th
2001)

131

450
400
350
300
250
200
150
100

50

Mass (ng)

 

  

 

0:00 4:00 8:00 12:00
Time
El Control L'l Low Ozone I Medium Ozone I High Ozone

Figure 5-17. Effect of samplings times on p-Cresol level. (Sampling method was SPME
and the a bar represents the averaged p-cresol concentration at the certain sampling time)

 

350 D on
I off

Mass (ng)

 

 

 

 

 

Control Low Ozone Medium Ozone High Ozone
Treatment

Figure 5-18. Effect of light on the average p-Cresol concentration. (Sampling method was
SPME. Concentrations of p-cresol that were measured when the light was on and off
were averaged.)

132

5. 5. 4. 1.3 Ozonation effect on indolic compounds

Small quantities of indolic compounds (indole and skatole) were recovered
from the swine building air using a SPME CAR/PDMS coating. Figure 5-19 shows the
GC response of indole for ozonated and control rooms. Ozone reduced indole and
skatole level in the ozonated rooms at a level that was signiﬁcant (0.0001). Indole was
not found in the ozonated room with highest ozone level. The highest indole level was
found in the control room. Figure 5-20 shows the average skatole adsorbed on SPME
ﬁber coatings from bi-weekly samples. The amount of skatole adsorbed on SPME
ﬁber coatings was reduced as ozone level in the room as increased. No skatole was
found in ozonated room with highest ozone dose. This suggests that ozone may reduce
the level of indolic compounds in the swine building air. Previous research (Wu et al.,
1999) has shown that ozone reduced indolic compounds in swine slurry.

Weekly air samples also showed similar results. Figure 5-21 shows the indolic
compounds collected in swine building air in weekly samples. Skatole level was
highest in the control room and no skatole was found in the ozonated room with
medium and high ozone dose. Small quantities of indole were detected in the control

room and there were no detectable indole in the ozone treated rooms.

133

12

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10
.. a ’ '
a - .
I; 6 ”1' J ..
a
a "1
5 4 ,, ._ ..
7/5 7/19 8/2 8/16 8/29 9/1 1 9/26
Date
El Control El Low Ozone I Medium Ozone

Figure 5-19. Average Indole levels in bi-weekly air samples. (Sampling method was

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SPME).
30
25 7
a 20 -I- P
5
g; 15 ~-
's'
10 -
5 i
0 . . . , .
7/5 7/19 8/2 8/16 8/29 9/1 1 9/26
Date 1
1:] Control Low Ozone I Med. Ozone 1

Figure 5-20. Average skatole level in bi-weekly air samples. (Sampling method was
SPME).

134

14

12 1
10. I I

 

Mass (n9)
J—‘l
1——'T—‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.* , 1 1, 1.1 11 111 111,11"

7l8 7/15 7/24 7/29 8/12 8/19 8/26 9/2 9/9 9/16
Date

 

El Indole-control I Indole-low ozone D Skatole-control El Skatole-low ozone

Figure 5-21. Level of lndolic compounds in weekly SPME samples. (Sampling
method was SPME).

5.5.4.2 VOCs in Thermal Desorption Tubes

Figure 5-22 shows a typical GC Chromatogram for air sample taken with a TD
tubes. As shown in Table 5-3 and 5-4, the major volatile organic compounds detected
were hydrocarbons and substituted benzenes (benzene, toluene, and xylene). '

Quantiﬁcation analysis for VOCs in TD samples was not conducted in this
research. Most of the compounds found in both ozonated and control room were
aromatic hydrocarbons. A few notable differences in air samples between the control

and ozonated rooms were detected. First, benzyl alcohol and methylphenol were found

135

in the control room and ozonated room with the low ozone dose but these were not
found in the medium and high ozone treated rooms.

A few VFAS compounds were found with this method, but recoveries of VF As
compounds were poor for the following reasons; ﬁrstly most of VFAS compounds had
same retention times with other alcohols and substituted benzene compounds,
resulting in poor separation in GC, and secondly the amount of VFAs adsorbed on
sorbent tubes were very small. This method was not sensitive for the analysis of
phenolic compounds in the swine building air with the total air volume of 5L. Only
small quantity of p-cresol was detected with this method.

The difﬁculties in collection of VFAS compounds in swine building air with
this air sampling method are likely due to the polarity of the sorbent materials in TD
tubes. The ﬁrst bedding sorbent, Tenax TA, is made of non-polar compounds which
favor non-polar compounds such as benzene and xylene. Due to the high uptake of
these non-polar compounds on the sorbent materials, other polar odorous compounds
were not able to adsorb onto the sorbent materials resulting in poor recoveries. Results
suggest that more polar sorbent materials in the TD tubes are required to collect semi

to high boiling polar odorous compounds in the livestock building air.

136

 

1 Control

 

 

“ l 1 1 i
1 11 1

‘ 1; 11 ll 7
1 .111 1‘1 "111 1 g

_ ,. ,- . -.. . 9L ,, ,, l ,,’l 1.- f‘ﬁ..- ..- We...) L-‘L ,3 in ‘1‘ l. . A .. . . . Aswan .,.'- {‘- >~ ,, 7,... . .1. I. ‘
Low Ozone 5

: .1 A ii 1
_ 1____‘,__/--"\______ __‘_,_,4-"_J\‘_’L_._

HUB. _,

 

 

 

Medium Ozone

 

 

 

5 1
_ i 4
1
. 1
High Ozone
. . 1
.1 1 1 -
_ i. 1 1,11.» . ’1' 1: z - ;\ A -:

 

 

 

 

Figure 5-22. GC/MS chromatogram for TD air samples.

137

 

Table 5-3. Volatile organic compounds found in the control room using TD tubes.

 

 

RT Compound Name Match

6.994 Benzene 885
8.337 Methyl isobutyl ketone (MIBK) 729
1 0.148 Bromomethane 834
1 1.022 Toluene 903
14.937 4-Chlorotoluene 746
16.236 Toluene 887
16.41 8 p-Xylene 798
1 7.459 lsopropylbenzene 91 8
1 7.797 m-Xylene 722
18.594 n-Propylbenzene 93 l
18.88 1,2,4, trimethylbenzene 943
19.475 4-Chlorotoluene 714
20.673 sec-Butylbenzene 850
21 .059 4—Isopropyltoluene 82 1
22.66 Benzyl alcohol 874
23.729 4-Methylphenol 785

 

RT: GC retention time (min), Match: MS spectrum library match (threshold:
700)

NIST MS library (provided by Varian) was used for the identiﬁcation of the
compounds. The search method used was “reverse search”. The program looked same
spectrum with the reference spectrum from the sample chromatogram and when the
matching value exceeded 700, the compound was identiﬁed. When the sample
spectrum matches perfectly with the reference spectrum, then match value become

1000.

138

Table 5-4. Volatile organic compounds found in ozonated room (6ﬁ3) using TD tube.

 

 

RT Compound Name Match
8.27 Methyl isobutyl ketone (MIBK) 753
l 0. 1 1 Bromomethane 73 5
l 1 Toluene 939
12.812 Tetrachloroethene 890
l 5 .008 Ethylbenzene 941
1 5.325 p-Xylene 944
1 7.539 lsopropylbenzene 928
1 8.672 n-Propylbenzene 937
19.642 1 ,2,4-Trimethylbenzene 94 1
20.129 1,3,5-Trimethylbenzene 958
20.74 sec-Butylbenzene 949
2 1 .108 4-lsopropyltoluene 948

 

139

 

5.5.4.3 Volatile Organic Compounds in Dust Samples

Due to the relatively clean air in the environmental chambers, dust levels in the
chambers were very low, so the level of VOC mass on the ﬁlters was a below the GC
detection limits. Concentrations of the dust in swine building air were not measured in

this research.

5.5.5 Comparison of sampling methods for detection of VOCs in swine
building air.

The results showed that SPME air sampling method was more selective than
sorbent tube method for both phenolic compounds and VF As. No phenolic compounds
and few VFAS were found using the sorbent tubes. For VFA compounds, peak
broading was observed in GC chromatogram, however, this is not related to the
sampling process but it is related to the polarity of the column material.

One disadvantage of using SPME ﬁber with a CAR/PDMS coating was the
poor recovery of phenolic compounds. The recovery of phenolic compounds can be
improved by using different types of SPME ﬁbers. Another difficulty for SPME air
sampling is quantiﬁcation. A recent study by Koziel et al. (2001) showed the
possibilities of quantiﬁcation of SPME with rapid air sampling methods. This method
is also useful for the quantiﬁcation of SPME samples at different temperatures and
atmospheric pressure. Koziel showed a rapid diffusive sampling method that will
make quantiﬁcation relatively simple. The sue of this methods for quantifying SPME

results is discussed in Chapter 4.

140

Overall, SPME air sampling was the most effective method for the collection
of phenolic and VFA compounds in swine building air. The sorbent tube method was

more effective for collecting substituted benzenes.

5.5.6 Temperature effect on odor and VOCs

Table 5-5 shows the temperature, odor detection threshold, and average GC
responses of phenolic compounds. Phenolic and indolic levels in control and ozonated
rooms were compared with temperature and odor DT. Since four olfactometry
measurements were made, only four sets of data were used. T tests for temperature
effect on VOCs and odor DT found that temperature did not affect signiﬁcantly on

odor DT and VOC emissions in swine air.

141

Table 5-5. Temperature, DT, and VOCs in swine building air.

 

Phenolic and indolic compounds sorbed on SPME

 

 

 

 

 

Ozone Temp. ﬁber (ng)
Date dose (°F) DT Phenol p-Cresol p-Ethylphenol Indole Skatole
7/6/01 0 70.38 1086 107 117 2.07 0 0
2 71.94 621 105 115 2.15 0.59 0
4 71.98 386 99.5 96.8 1.94 0 0
6 73.55 327 93.6 98.4 1.98 0 0
8/3/01 0 74.44 808 144 353 2.93 7.26 19.85
2 77.02 683 122 296 2.99 3.36 5.92
4 75.80 683 118 297 2.76 1.34 1.40
6 73.39 382 102 164 2.08 0 0
8/31/01 0 80.19 1113 128 308 2.28 6.10 13.33
2 79.74 799 147 430 2.69 3.04 9.22
4 76.92 866 100 224 1.92 0 5.9
6 75.87 676 110 366 2.33 0 0
9/28/01 0 78.47 944 138 408 2.28 4.33 14.69
2 75.47 734 89 229 1.17 0 1.71
4 77.23 622 1 1 1 260 1.90 0.82 0.95
6 78.82 317 89 192 1.71 0 0

 

142

Units for the phenolic and indolic compounds: ng

5.6 Conclusions

This research showed that ozonation reduced the odor detection threshold, but
it did not reduce odor offensiveness or odor intensity. Odor characteristics of the
ozonated air were different from those of the air in the control room. These results
suggest that some odorous compounds were destroyed by ozone but the residual
compounds still exhibited a strong and offensiveness odor. Another possibility is that
the by-products formed during ozonation had an offensive odor. A previous study for
ozonation on odor and taste in drinking water has shown that ozone altered odor
characteristics in drinking water but it did not reduce odor intensity (Hargesheimer
and Watson, 1996). According to US EPA publications, ozone is not effective at
removing many odor-causing chemicals at concentrations that do not exceed public
health standards (IAQ Publications, 2002).

Ozone was not effective in reducing VFAs but ozone reduced phenolic and
indolic compounds in the swine building air. Previous research (Wu et al., 1999) has
shown that ozone reduced phenolic and indolic compounds in swine slurry.

The study showed that SPME air sampling methods were suitable to collect
both VFAs and phenolic compounds in the swine building air. Analysis for VFAs and
phenolic compounds on thermal desorption tubes with Tenax TA plus Carbosieve was
difficult because the high mass loading of non-polar hydrocarbon compounds on the
tubes.

SPME air sampling appears to be an effective method of sampling for odorous

compounds in livestock building air. It is an easy and solvent free method and also

143

does not require sample pretreatment steps. However, quantiﬁcation of SPME air

samplings is difﬁcult because the recovery is sensitive to environmental conditions.

144

CHAPTER 6. CONCLUSIONS AND
RECOMMENDATIONS

6.1 . Conclusions

The main purposes of the research were to ﬁnd suitable air sampling methods
for the odorous VOCs, to develop a new sampling technique for VOCs in air based
upon SPME, and to use the methods developed to study the effect of ozonation on
levels of odorants and other VOCs in the livestock building air.

The major odorants found in the livestock building air were volatile fatty acids
(VF As) and phenolic (phenol, p-cresol, p-ethylphenol) and indolic (indole and skatole)
compounds. These compounds are formed by bacteria under anaerobic conditions.

Without any further sample concentration, the solvent desorption sorbent tube
methods studied were not able to detect any volatile organic compounds in the swine
nursery housing air, using a sample volume of up to 36 L. This method also required
handling of hazardous organic solvents for sample preparation. As sample
concentration is time consuming and there is a risk that the analytes may be lost during
this step, this method was deemed to be less suitable than the thermal desorption or
SPME methods. Using the impinger trap method the presence volatile organic
compounds in the swine nursery housing air, was not detect, using a sample volume as
large as 72L, so it was also unsuitable.

Thermal desorption has been widely used for the desorption of sorbent tubes.
The major advantage of the thermal desorption method over solvent desorption is that
the entire sample can be introduced into the analytical instrument. Also, as no solvent

is used to desorb the analytes, there is no interference due to the solvent. This research

145

 

showed that multi-bed thermal desorption tubes can be used for the sampling of
odorous compounds from livestock buildings. However the recovery of high boiling
point compounds was not good using thermal desorption.

The SPME air sampling method deve10ped showed good recovery of the high
boiling point compounds, however this method was not good for the recovery of the
low boiling point compounds. The SPME air sampling method has advantages over
other sampling techniques: the major advantages of SPME are; good sensitivity, ease
of sampling, and minimal sample preparation is required.

It is apparent from research that no single method for air sampling appears to
be satisfactory for sampling of VOCs in swine building air. For the determination of
high boiling point compounds, such as phenolics and indolics, the SPME air sampling
method offered the best results. Sampling using both SPME and a Tenax TA —
Carboxen sorbent tube followed by thermal desorption seems to offer the best results.
For the low to mid boiling point compounds, recoveries were good using the thermal
desorption sorbent tubes. For the high boiling point compounds, such as phenolics and
indolics, SPME seems to be the most effective method.

A major drawback of the SPME air sampling technique is that quantiﬁcation of
the results is difﬁcult. SPME ﬁbers are not uniformly sensitive to all compounds and
therefore, the relative GC peak area for an SPME sample does not reﬂect the true
proportions of the components in the sample. Furthermore, other
environmental/physical factors can affect the uptake of the contaminant on the ﬁber.
Two methods can be used for the quantiﬁcation of SPME samples; these are based on

equilibrium theory and diffusion theory. This research compared the quantiﬁcation of

146

phenolic compounds using these two methods. The results showed that for phenolic
compounds the mass sorbed on PDMS/CAR ﬁber coating, at a gas phase
concentration of approximately 4 ug/L increased linearly with sampling times of up to
30 minutes. However, the mass sorbed on the SPME ﬁber did not increase linearly
with time if the gas phase concentrations decreased by a factor of 10. Rapid sampling
with SPME showed that at the higher concentration the mass sorbed on SPME ﬁber
increased linearly with time with sampling time of up to 150 seconds.

Several systematic experimental errors were observed in the SPME
experiments. For the quantiﬁcation of VOCs sorbed on a SPME ﬁber, gas standards
were prepared by injecting a known amount of the analyte dissolved in methanol into
the gas sampling bottle. It was assumed that all VOCs injected in the gas sampling
bottle were evaporated, and that the sorption of the analyte to the walls at the container
was minimal. When sorption occurs onto the surface of the gas sampling bottle
(termed a wall effect), the actual gas concentration is less than that calculated. Wall
effects may not be negligible, and where this is the case the concentration of the
analyte in the gas phase should be determined by direct gas injection in GC or other
suitable means. Temperature is also another possible source of error. Temperature was
not controlled during the sampling. Rate of uptake of the analyte on the SPME ﬁber
increases as the room temperature increases due to the increases in molecular
diffusivity. In this work temperature changes in the laboratory during the experiment
were i2°C. Calibration of the SPME ﬁbers should be conducted under temperature

controlled conditions.

147

The studies conducted showed that VFAs were preferentially sorbed onto
CAR/PDMS ﬁbers and that extent of sorption of phenolic and indolic compounds was
greater using PDMS/DVB ﬁbers. This result suggests that using two types of ﬁbers
may be desirable for the sampling of FVAs and phenolic compounds in the livestock
building air.

The experiments conducted on the ozonation of swine building air showed that
ozonation reduced the odor detection threshold in the swine building air, but it did not
reduce odor offensiveness or odor intensity. The odor characteristics in ozonated
room were different from those in the control room. These results suggest that some
odorous compounds were destroyed by ozone, but the residual compounds still
exhibited a strong and offensiveness odor. Another possibility is that the by—products
formed during ozonation had an offensive odor. Studies on the composition of the air
in the swine building air showed that ozonation did not signiﬁcantly reduce the levels
of VFAs in the building. However, the concentration of phenolic and indolic
compounds in swine building air was reduced in pens where the air was ozonated.
Previous research has shown that ozone is very reactive with these compounds and
that ozonation of a swine slurry reduced the level of phenolic and indolic compounds
(Wu et al., 1999). Research recently published by the US EPA support the
conclusions made here. It was found that at concentrations that did not exceed public
health standard ozone was not effective in removing odor-causing chemicals in indoor

air (IAQ Publications, 2002).

148

6.2 Recommendations for future research

This study has compared several air sampling methods for the odorous
compounds in the livestock building air and also studied the efﬁciency of ozone to
control odor. The followings recommendations are made for future research:

1. SPME air sampling was the best method for the detection of VFAs and
phenolics in swine building air. Two types of SPME ﬁbers are recommended:
PDMS/CAR for VFAs and PDMS/DVB for phenolics. When SPME air sampling
methods are chosen, accurate gas standards are required for the quantiﬁcation of the
analytes on SPME. This study showed that the technique of preparing gas standards by
dilution in a gas sampling bottle could not be used to produce accurate gas standards.
Temperature control during the SPME calibration is also recommended. The sorptive
abilities of commercial SPME ﬁbers decrease with the usage, so frequent recalibration
of SPME ﬁbers is recommended.

2. Ozonation reduces the level of phenolic compounds in swine building air,
however, it was not effective for the control of odor offensiveness, odor intensity or
VFAs in these buildings. Given the cost of the systems, ozone seems not to be a very
cost effective method for the reduction of odors in swine buildings. Other

technologies for odor reduction should be considered in further research.

149

REFERENCES

Abalos M. and J .M. Bayona. 2000. Application of gas chromatography coupled to
chemical ionization mass spectrometry following headspace solid-phase
microextraction for the determination of free volatile fatty acids in aqueous
samples. J. Chromatogr. A. 891: 287-294.

ACHIH. 1993. Threshold limit values for chemical substances, physical agents and
biological exposure indices. Cincinnati, Ohio: Am. Conf. Gov. Ind. Hygienists.

Ai J. 1997. Headspace solid phase microextraction. Dynamics and quantitative
analysis before reaching a partition equilibrium. Anal. Chem. 69 (16): 3260-
3266.

Air Toxic Methods. 2001. http://www.epa.gov/ttn/amtic/airtox.htrnl. US EPA.

APHA. 1998. Section 2170. Flavor Proﬁle Analysis, Standard Methods for the
Examination of Water and Wastewater. 20th Edition.

Arthur CL and J. Pawliszyn 1990. Solid-phase micrpextraction with thermal
desorption using fused silica optical ﬁbers. Anal. Chem. 62: 2145-2148.

Augusto F., J. Koziel, and J. Pawliszyn. 2001. Design and Validation of Portable
SPME Devices for Rapid Field Air Sampling and Diffusion-Based Calibration.
Anal. Chem. 73: 481-486.

Bartelt R.J. and B.W. zilkowski. 1999. Nonequilibrium Quantitation of Volatiles in
Air Streams by Solid-Phase Microextraction. Anal. Chem. 71: 91-101.

Bataller M., E. Veliz, R. Perez-Rey, L.A. Fernandez, M. Gutierrez, A. Marquez, 2001,
Ozone swimming pool water treatment under tropical conditions, Ozone
Science and Engineering. 22:677-682.

Beltran F.J., J .F. Garcia-Araya, J. Rivas, P.M. Alvarez, and E. Rodriguez, 2000.
Ozone remediation of some phenol compounds present in food processing
wastewater. J. Environ. Sci. Health. Part A. 35 (5): 681-699.

Carini D, von Gunten U., Dunn IJ, and Morbidelli M., 2001. Ozonation as Pre-
Treatment Step for the Biological Degradation of Industrial Wastewater
Containing 3-Methyl-Pridine. Ozone Science and Engineering, 23:189-193.

Carpenter G.A. 1986. Dust in livestock buildings - review of some aspects. J. Agric.
Eng. Res. 33:227-241.

Ciccioli P, E. Brancaleoni, A. Cecinato, R. Sparapani, M. Franttoni, 1993.
Identiﬁcation and determination of biogenic and anthropogenic volatile

150

organic compounds in forest areas of northan and southern Europe and a
remote site of the himalaya region by high resolution gas chromatography
mass spectrometry. J. of Chromatography. 643 (1-2): 55-69.

Clark T.J. and J .E. Bunch. 1997. Qualitative and quantitative analysis of ﬂavor
additives on tobacco products using SPME-GC-Mass Spectrometry. J. Agric.
Food Chem. 45: 844-849.

Dombi A., I. Ilisz, Z. Laszlo, and G. Wittmann. 2002. Comparison of Ozone-based
and other (V UV and TiOleV) Radical Generation Methods in Phenol
Decomposition. Ozone Science and Engineering. 24: 49-54.

Donham K.J. 1988. Hazarous agents in agricultural dusts and methods of evaluation.
American aerial dust particles in swine ﬁnishing buildings. Trans. ASAE. 31:
882-887.

Doong R. and S. Chang, 2000. Determination of Distribution Coefﬁcients of Priority
Polycyclic Aromatic Hydrocarbons Using Solid-Phase Microextraction. Anal.
Chem. 72: 3647-3652.

Edie weekly summaries. 2001. First cheap pollution-free pig farm in Indiana.
http://www.edie.net/newsr/Archive/4437.cfm.

EPA Compendium Method TO-l. 1999. Method for the determination of volatile
organic compounds in ambient air using tenax adsorption and has
chromatography/mass spectrometer. Compendium of methods for toxic
organic air pollutants. US EPA.

Esplugas, S. J. Gimenez, S. Contreras, E. Pascual, and M. Rodriguez. 2002.
Comparison of different advanced oxidation processes for phenol degradation.
Water. Res. 36 (4): 1034-1042.

Esswein BI. and M.F. Boeniger. 1994. Effects of ozone-generating air purifying
device on reducing concentrations of formaldehyde in air. Applied Occup.
Environ. Hygiene. 9(2): 139-146.

Fakhoury K.J., A.J. Heber, P. Shao, and J .Q. Ni. 2000. Correlation of odor detection
thresholds with concentrations of hydrogen sulﬁde, ammonia and trace gases
emitted from swine manure. 2000 ASAE Annual international meeting,
Midwest Express Center, Milwaukee, Wisconsin, July 9-12, 2000.

Finch G.R., C.N. Haas, J .A. Oppenheimer, G. Gordon, R.R. Trussell, 2001. Design
Criteria for Inactivation of Cryptosporidium by Ozone in Drinking Water.
Ozone Science and Engineering. 23: 259-284.

Hammond, E.G., C. Fedler, and G. Junk. 1979. Identiﬁcation of dust-borne swine
house conﬁnement facilities. Trans. of the ASAE. 22(5): 1186-1192.

151

Hammond, E.G., C. F edler, and R.J. Smith. 1981. Analysis of particle-borne swine
house odours. Agriculture and Environment. 6:395-401.

Hammond, E.G., C. Heppner, and R. Smith. 1989. Odors of swine waste lagoons.
Agriculture, Ecosystems and Environment, 25:103-110.

Hargesheimer BE. and SB. Watson. 1996. Drinking-water treatment options for taste
and odor control. Water. Res. 30 (6): 1423-1430.

Harper M. 2000. Sorbent trapping of volatile organic compounds from air. J. of
Chromatography A. 885: 129-151.

Hartung J. and HG. Hilliger. 1980. Odour characterization in animal houses by gas
chromatographic anlaysis on the basis of low temperature sorption. In
“Effluents from Livestock” (Gasser J .K.R., ed.) London, Applied Science
Publishers, 561-578.

Hayter, RB. and EL. Besch. 1974. Airborne particle deposition in the respiratory tract
of chickens. Poultry Science. 53: 1507-1512.

Helmig D. 1999. Air analysis by gas chromatography. J. of Chromatography A. 843:
129-146.

Hines EL. and J .W. Gardner J .W., Artiﬁcial neural emulator for an odour sensor
array. Sens Actuators B, 19 661-664, 1994.

Hobbs P.J., Misselbrrok T.H., Pain, B.F., Assessment of odours from livestock wastes
by a photoinization detector, an electronic nose, olfactometry and gas
chromatography-mass spectrometer, J. agric. Engng Res. 60, 137-144, 1995.

Hobbs P.J., Misselbrook T.M., Cumby T.R., Production and emission of odours and
gases from aging pig waste, J. Agric. Engng. Res., 72, 291-298, 1999.

Harry, E.G. 1978. Air pollution in farm buildings and methods of control: a review.
Avian Pathology. 7: 441-454. ‘

IAQ Publications. 2002. Ozone generators that are sold as air cleaners: An assessment
of effectiveness and health consequences. US EPA

Karlik JF. and AM Winer, 2001. Measured isoprene emission rates of plants in
California landscapeszcomparison to estimates from taxonomic relationships.
Atmospheric environment. 35 (6): 1123-1 131.

Kilham, LB. and RM Dodd. 1999. The Application of Ozone for Air Treatment
(Case Study of a Bingo Hall HVAC System), in Proc. 14th Ozone World
Congress, Dearbom, MI Vol 2. pp.49-56.

152

Kontsas H., C. Rosenberg, P. Jappinen, and ML. Riekkola. 1993. Glass ﬁbre preﬁlter-
XAD-2 sampling and gas chromatographic determination of airborne
chlorophenols. J. Chromatogr. 636(2): 255-261.

Koziel J. A., M.Y. J ia and J. Pawliszyn. 2000. Air sampling with porous solid-phase
microextraction ﬁbers. Anal. Chem. 72 (21): 5178-5186.

Koziel J .A. and J. Pawliszyn. 2001. Air sampling and anlaysis of volatile organic
compounds with solid phase microextraction. J. Air and Waste Manage. Assoc.
5 1 : 173-1 84.

Kruithof J .C. and W.J. Masschelein. 1999. State-of-the-Art of the Application of
Ozonation in BENELUX Drinking Water Treatment. Ozone Science and
Engineering. 21 :139-152.

Ku Y. and KC. Lee. 2000. Removal of phenols from aqueous solution by XAD-4
resin. J Hazard. Mater. 80 (1-3): 59-68.

Lezcano 1., RP. Rey, Ch. Baluja. E. Sanchez. 1999. Ozone Inactivation of
Psedomonas aeruginosa, Esterichia coli, Shigella sonnei, and Salmonella
typhimurium in Water. Ozone Science and Engineering. 21 :293-300.

Liu Y., Y. Shen, M.L. Lee, 1997. Porous Layer Solid Phase Microextraction Using
Silica Bonded Phases. Anal Chem. 69: 190-195.

Louch D., S. Motlagh, J. Pawliszyn. 1992. Dynamics of organic compound extraction
from water using liquid-coated fused silica ﬁbers. Anal. Chem. 64: 1187-1199.

Mackie, R.I., G. P. Stroot, and VB. Varel, 1998, Biochemicalidentiﬁcation and
biological origin of key odor components in livestock waste. J. Anuim. Sci.
76:1331-1342.

Martin V.F.V., P.L. Mahia., and SM. Lorenzo. 2000. Development of a thermal
desorption-gas chromatography-mass spectrometry method for determination
of styrene in air. Application to workplace air. ANALUSIS 28 (8): 737-742.

Martos PA. and J. Pawliszyn. 1997. Calibration of Solid Phase Microextraction for
Air Analyses Based on Physical Chemical Properties of the Coating. Anal.
Chem. 69: 206-215.

Martos P. and J. Pawliszyn, 1999. Time-Weighted Average Sampling with Solid-
Phase Microextraction Device: Implications for Enhanced Personal Exposure
Monitoring to Airborne Pollutants. Anal. Chem. 71: 1513-1520.

Masuda J ., Fukuyama J. and F ujii S., 2001. Ozone injection into an activated carbon
bed to remove hydrogen sulﬁde in the presence of concurrent substrates. J. Air
Waste Manage. Assoc. 51 (5): 750-755.

153

Mighirang, R.G., M.C. Puma, Y. Lui, and P. Clark, 1997. Dust concentrations and
particle size distribution in an enclosed swine nursery. Trans. ASAE. 40(3):
49-754.

Ng L.K., M. Hupe. J. Harnois. D. Moccia. 1996. Characterization of commercial
Vodkas by Solid-phase Microextraction and Gas Chromatography/Mass
Spectrometry Analysis. J. Sci. Food Agric. 70, 380-388.

NPPC. 1995. A review of the literature on the nature and control of odors from pork
production facilities. National Pork Producers Council, Des Moines, IA.

Oehrl, L.L., K.M. Keener, R.W. Bottcher, and KM. Connelly, 2000. Characterization
of odor components from swine housing dust using gas chromatography. Proc.
of ASAE meeting. 2000 ASAE Annual International Meeting, Milwaukee, WI,
July 9-12, 2000.

O’Neill DH. and V.R. Phillips. 1992. A review of the control of odor nuisance from
livestock buildings. 3. Properties of the odorous substances which have
identiﬁed in livestock wastes or in the air around them. I. Agric. Eng. Res., v.

53 (1): 23-50.

Ozone Solutions Inc. 2002. Ozone and swine operations manual.
http://www.mtcnet.net/~jdhogg/ozone/oznmanual.htrnl

Pain B.F., V.R. Phillips, C.R. Clarkson, T.M. Misselbrook, T.J. Rees, J .W. Farrent.
1990. Odour and ammonia emissions following spreading of aerobically-
treated pig slurry on grassland, Biological Wastes, 34, 149-160.

Pan L., J. Chong, and J. Pawliszyn. 1997. Determination of amines in air and water
using derivatization combined with solid-phase microextraction. J.
Chromatography A. 773: 249-260.

Pan L., M. Adams, J. Pawliszyn, 1995. Determination of Fatty Acids Using Solid-
Phase Microextraction. Anal. Chem. 67: 4396-4403.

Paustenbach D, M.L. Burke, and M. Shum. 2001. Airborne concentrations of ethyl
cyanoacrylate in the workplace. AIHAJ 62 (1): 70-79.

Rice R.G. 2002, Century 21 — Pregnancy with Ozone. Ozone Science and
Engineering. 24: 1-1 5.

Ritter WF. 1989. Odor control of livestock wastes-state-of-the-art in North-America.
J. of Agric. Eng. Res. 42 (1): 51-62.

Schaefer J. 1977. Sampling, characterization and analysis of malodors, Agriculture
and Environment, 3: 121-127.

154

Schwarzenbach, R.P., P.M. Gschwend, and D.M. Imboden. 1993. Environmental
Organic Chemistry, John Wiely & Sons, Inc. pp. 194-200.

Shaughnessy, R.J., T.J. McDaniels, and CJ. Weschler. 2001. Indoor chemistry: Ozone
and volatile organic compounds found in tabacco smoke. Environ. Sci.
Technol. 35 (13): 2758-2764.

Shirey R. E. 1999. SPME Fibers and Selection for Speciﬁc Applications. In: In: Solid
Phase Misroextraction A Practical Guide. Marcel Dekker, Inc. pp.59-110.

Song J ., B.D. Gardner. J .F . Holland. and RM. Beaudry. 1997. Rapid Analysis of
Volatile Flavor Compounds in Apple Fruit using SPME and GC/Time-of-ﬂight
Mass Spectrometry. J. Agric. Food Chem. 45, 1801-1807.

Sostaric T., M.C. Boyce and BE. Spickett. 2000. Analysis of the Volatile Components
in Vanilla Extracts and F alvorings by Solid-Phase Microextraction and Gas
Chromatography. J. Agric. Food Chem. 48: 5802-5807.

Steffen A. and J. Pawliszyn. 1996. Analysis of ﬂavor volatiles using headspace solid-
phase microextraction. J. Agr. Food Chem. 44 (8): 2187-2193.

Sukola K, J. Koziel, F. Augusto, and J. Pawliszyn, 2001. Diffusion-Based Calibration
for SPME Analysis of Aqueous Samples. Anal Chem. 73: 13-18.

Sun. B., M. Sato, and J .S., Clements. 2000. Oxidative processes occurring when
pulsed high voltage discharges degrade phenol in aqueous solution. Environ.
Sci. Technol. 34(3): 509-513.

Suryanarayana M.V.S., R.K. Shrivastava, and D. Pandey. 2001. Simple time weighted
average level air-monitoring method for sulfur mustard in work places. J
Chromatograph A 907 (1-2): 229-234.

Tamminga S. 1992. Gaseous pollutants by farm animal enterprises. In: Farm animals
and the environment. pp 345-357. C. Phillips and D. Piggins (Ed.), CAB
International, Wallingford, U.K.

Thurman E.M., and MS. Mills. 1998. Chapter 8.10.4. Vitamin B12 in Urine or
Aqueous Solution. In: Chemical Analysis: A series of monographs on
Analytical Chemistry and Its Applications: Volume 147, Solid Phase
Extraction: Principles and Practice. A Wiley-Interscience Publication, John
Wiley & sons, Inc., pp. 213- 214.

Tsai SW. and S.S.Q. Hee. 2000. A new passive sampler for regulated workplace
ketone. AIHAJ 61 (6): 808-814.

Turan-Ertas T. 2001. Biological and Physical-Chemical Treatment of Textile Dyeing
Wastewater for Color and COD Removal. Ozone Science and Engineering. 23:
1 99-206.

155

Vansickle J. 1999. Ozone holds promise for odor control. National Hog Farmer. 44
(7): 1-6. http://www.ozoneww.com/Newl .htm.

Vianna E. and S. E. Ebeler. 2001. Monitoring Ester Formation in Graph Juice
Fermentations Using Solid Phase Microextraction Coupled with Gas
Chromatography-Mass Spectrometry. J. Agric. Food Chem. 49: 589-595.

Watkins, B.D., S.M. Hengemuehle, H.L. Person, M.T. Yokoyama, and SJ. Masten,
1997. Ozonation of swine manure wastes to control odors and reduce the
concentrations of pathogens and toxic fermentation metabolites. Ozone
Science and Engineering. 19: 425-437. -‘

 

Wercinski S.A.S and J. Pawliszyn. 1999. Solid Phase Microextraction Theory. In:
Solid Phase Misroextraction A Preactical Guide. Marcel Dekker, Inc. pp.4-26.

Weschler OJ. and HG Shields. 1999. Indoor ozone/terpene reactions as a source of ‘
indoor particles. Atmos. Environ. 33 (15): 2301 -23 12.

Williams AG. 1989. Dust and odour relationships in broiler house air. J. Agric. Eng.
Res. 44:175-190.

Wu, J .J ., S. Park, S.M. Hengemuehle, M.T. Yokoyama, H.L. Person, J .B. Gerrish, and
SJ. Masten, 1999. The use of ozone to reduce the concentration of
malodorous metabolites in swine manure slurry. J. Agric. Eng. Res. 72: 317-
327.

Wu J.J., S.H. Park, S.M. Hengemuehle, M.T. Yokoyama, H.L. Person, and SJ.
Masten, 1998. The effect of storage and ozonation on the physical, chemical,
and biological characteristics of swine manure Slurries. Ozone Science and
Engineering. 20: 5-50.

Wu, J .N., K. Rudy, and J. Spark. 2000. Oxidation of aqueous phenol by ozone and
peroxidase. ADV Environ. Res. 4(4): 339-346.

Xue, S.K., S. Chen, and RE. Hermanson. 1998. Measuring ammonia and hydrogen
sulﬁde emitted from manure storage facilities. Trans. ASAE. 41(4), July/Aug,
1125-1140.

Yang X. and T. Peppard. 1994. Solid-phase microextraction for ﬂavor analysis, J.
Agric. Food Chem. 42, 1925-1930.

Yasuhara A., K. Fuwa, and M. J irnbu. 1984. Identiﬁcation of odourous compounds in
fresh and rotten swine manure, Agric. Bio. Chem., 48 (12), 3001-3010.

Yasuhara A., 1980. Relation between odor and odorous components in solid swine
manure. Chemosphere 9: 587-592.

156

 

Yu J.C., C.E. Isaac, R.N. Coleman, J. Jr. Feddes, and BS. West. 1991. Odorous
compounds from treated pig manure. Canadian Agric. Engng. 33 (1): 131-136.

Zahn J.A., J .L. Harﬁeld, Y.S. Do, A.A. Dispirito, D.A. Laird, R.L. Pfeiffer. 1997.
Characterization of volatile organic emissions and wastes from a swine
production facility, J. Environmental Quality, 26, 1687-1696.

Zhang Z. and J. Pawliszyn. 1993. Headspace Solid-phase microextraction. Anal.
Chem. 65: 1843-1852.

Zhou X., D.A. Reckhow and J.E. Tobiason. 1992. Formation and Removal of
Aldehyde in Drinking Water Treatment Processes. Water Quality Technology
Conference Proceedings. Toronto. Ontario. pp.29l-315.

Zhu J ., G.L. Riskowski, and M. Torremorell. 1999. Volatile fatty acids as odor
indicators in swine manure- a critical review, Trans. ASAE, 42(1): 175-182.

Zimmerman R. and D. Richard, 1990. Understanding the effects of ozonation on a
combined municipal/industrial secondary efﬂuent. Ozone Science and
Engineering. 12: 107-114.

157

11111111111111