CONTROL OF GAS AND ODOR LEVELS IN SWINE FACILITIES

USING FILTERS WITH ZINC OXIDE NANOPARTICLES

A. C. Alvarado, B. Z. Predicala

ABSTRACT. The effectiveness of zinc oxide (ZnO) nanoparticles as filtering media for controlling the levels of hydrogen
sulfide (H2S), ammonia (NH3), and odor in swine facilities was evaluated in this study. Semi-pilot scale tests were done to
determine basic operational factors, the results of which showed that the fluidized bed air filtration system (FBAFS), loaded
with ZnO nanoparticles at a rate of 0.28 g cm-2 of filter area, and a gas flow rate equivalent to 0.5 m s-1 face velocity
achieved significant reduction in target gas levels. The performance of this filter system was further investigated in a room-
scale environmental chamber representative of normal swine production conditions. When installed as part of the ventilation
air recirculation system of the room, the FBAFS with ZnO nanoparticles achieved about 65% H2S and 42% NH3 reductions
in the human-occupied zones but had no significant impact on pig performance as well as odor levels in the chamber.
Keywords. Ammonia, Face velocity, Filtration, Hydrogen sulfide, Nanoparticles, Odor, Swine, Ventilation, Zinc oxide.

O
ver the past several decades, livestock production of NH3 to the atmosphere has been associated with environ-
systems have shifted toward intensified opera- mental concerns such as the production of fine particles
tions with increasing number of animals raised in (PM2.5) (Wang et al., 2011; Megaritis et al., 2013), which are
large specialized facilities. Swine production has major contributors to local aerosol pollution (Wang et al.,
also followed this trend, a consequence of which is the gen- 2015), formation of acid rain (Gao et al., 2010), and eutroph-
eration of large volumes of manure stored in underfloor pits ication of ground and surface waters through nitrogen en-
in intensive swine confinement units from which significant richment (Hartung and Phillips, 1994; Casey et al., 2006).
amounts of air contaminants are emitted. These contami- The environmental and occupational issues associated
nants are composed largely of gases produced from both with the emission of H2S, NH3, and odor from swine facili-
freshly deposited and stored animal manure (Hartung and ties remain unresolved despite considerable past and on-go-
Phillips, 1994), of which the gases of most concern are am- ing work on mitigation techniques ranging from controlling
monia (NH3) and hydrogen sulfide (H2S) (Watts, 1999). the gas precursors to treating the gases biologically or chem-
Continuous exposure of pigs to 20 mg L-1 of H2S may ically before releasing them into the barn and/or ambient en-
result in loss of appetite, while respiratory diseases may de- vironment. Specific mitigation practices include the use of
velop when pigs are in contact with 50 mg L-1 of NH3 for manure additives (Williams and Schiffman, 1996; Hendriks
prolonged periods (ASABE, 2005). These symptoms can et al., 1997; Smith et al., 2004; Govere et al., 2005; Schnee-
eventually lead to reduced growth and poor performance. gurt et al., 2005; Smith and Nicolai, 2005), diet modification
Bushee et al. (1999) reported that exposure of pigs to high (Kendall et al., 1999, 2000; DeCamp et al., 2001; Godbout
NH3 concentrations could decrease food intake, resulting to et al., 2001), oil sprinkling (Jacobson et al., 1998; Godbout
weight loss and increased susceptibility to disease. Swine et al., 2001; Heber et al., 2004; Kim et al., 2008), and biofil-
workers are likely to experience respiratory tract infections tration (Hartung et al., 2001; Martinec et al., 2001; Nicolai
when exposed to 200 to 300 mg L-1 of H2S or 100 mg L-1 of and Janni, 2001; Chen et al., 2008), to cite a few examples.
NH3 and even death when exposed to more than 500 mg L-1 Recognizing recent advances in nanotechnology, a com-
of H2S or 1000 mg L-1 of NH3 (ASABE, 2005). Several stud- prehensive evaluation of the use of nanoparticles for control-
ies also reported that odors from swine operations caused el- ling emissions of H2S and NH3 from swine manure was com-
evated symptoms of eye, nose, and throat irritation (Schinasi pleted by Alvarado et al. (2015). Alvarado et al. (2015) ini-
et al., 2011) and decreased social interactions and quality of tiated their study by screening nanoparticles in laboratory-
life in exposed communities (Tajik et al., 2008). The release scale experiments to identify particular nanoparticles that
exhibited the highest H2S and NH3 reduction qualities, fol-
lowed by semi-pilot scale tests to establish the operational
Submitted for review in July 2016 as manuscript number PAFS 12010; parameters, particularly nanoparticle-to-slurry ratio, and ap-
approved for publication by the Plant, Animal, & Facility Systems plication procedures. Finally, room-scale tests were con-
Community of ASABE in March 2017. ducted to assess the feasibility of using nanoparticles in an
The authors are Alvin C. Alvarado, Research Assistant - Engineering,
and Bernardo Z. Predicala, ASABE Member, Research Scientist -
actual livestock production setting. Results from their tests
Engineering, Prairie Swine Centre Inc., Saskatoon, Saskatchewan, Canada. proved that the concept of mixing ZnO nanoparticles into
Corresponding author: Bernardo Predicala, P.O. Box 21057, 2105 8th St. manure slurry was effective for controlling emissions of H2S
East, Saskatoon, SK, Canada S7H 5N9; phone: 306-667-7444; e-mail: and NH3 from swine manure. The reduction in gas concen-
[email protected].

Transactions of the ASABE

Vol. 60(3): 943-956 © 2017 American Society of Agricultural and Biological Engineers ISSN 2151-0032 https://doi.org/10.13031/trans.12010 943
trations could be attributed to the chemical adsorption inter- SEMI-PILOT SCALE TESTS
action between the nanoparticles and the target gases The semi-pilot scale tests were divided into two manage-
(Sanchez et al., 2005; Li and King, 2006; Wang et al., 2008). able phases to limit the confounding effects of testing more
Building on the demonstrated effective application of nano- than two variables. Phase 1 was conducted to determine the
particles in air purification (Nonami et al., 2004; Kim et al., most effective filter design and corresponding face velocity.
2006) and the success of using ZnO nanoparticles as a ma- The most effective combination was then used in the second
nure additive to reduce gas emissions from swine manure phase to determine the nanoparticle loading capacity re-
(Alvarado et al., 2015), the present study explored the use of quired to achieve significant reductions in H2S and NH3 con-
ZnO nanoparticles as filtering media in reducing H2S and centrations.
NH3 concentrations in swine barn air. In the first phase, two filtration systems were tested: a
One possible application of using ZnO nanoparticles for packed bed air filtration system (PBAFS), and a fluidized
filtering contaminants in barn air is the installation of a filter bed air filtration system (FBAFS). The tests involved con-
system in either the recirculation or exhaust ducts of the barn structing a filter system with ZnO nanoparticles loaded
ventilation system. However, considering the actual config- evenly into styrene honeycomb structures measuring
uration of ventilation systems in existing barns, installing the 11.4 cm  10.2 cm  3.2 cm thick, with 1.5 cm  1.5 cm pore
filter setup at the exhaust would likely cause adverse impact openings (fig. 1a). Both sides (upstream and downstream) of
on the ventilation system due to the substantial pressure drop the honeycomb structure were covered with a 100% syn-
resulting from the installed filter. A more practical approach thetic fiber pad material (model HPE30621 Electrostatic
would be to install the filter system as part of the barn recir- Hammock filter pad, True Blue Co., LaPorte, Ind.) to con-
culation system. Such systems are already in use in about strain the nanoparticles within the filter. The filter pad mate-
65% of existing swine barns in Canada (Dennis Hodgkinson, rial had a thickness of 0.6 cm and a minimum efficiency re-
DGH Engineering Ltd., personal communication, 5 Dec. porting value (MERV) rating of 7.
2016). It was hypothesized that installation of a filter system Figures 1b and 1c show schematic diagrams of the pre-
with ZnO nanoparticles inside a facility would reduce the pared packed bed and fluidized bed filters. In the PBAFS
concentrations of contaminant gases within the building as (fig. 1b), the nanoparticles were fully compacted into the fil-
well as decrease the corresponding levels emitted into the ter media by slicing the honeycomb structure across its
atmosphere. However, a number of operational factors can cross-section into thinner units to accommodate each load-
significantly influence the effectiveness of the filtration sys- ing rate. Thus, the effective thickness of the filter varied with
tem in reducing the levels of the target gases, including filter the loading capacity of the nanoparticles. Three different
design (packed bed vs. fluidized bed), face velocity, and na- loading capacities were investigated: 30, 60, and 90 g, cor-
noparticle loading capacity. Filter bed design and nanoparti- responding to loading rates of 0.28, 0.56, and 0.84 g cm-2 of
cle loading capacity influence the available surface area for filter area, respectively. The filter with 30 g of ZnO nano-
physical and chemical reactions, while face velocity affects particles measured 11.4 cm long  10.2 cm wide  0.8 cm
the duration of exposure of the target gases to the nanoparti- thick, while the filters with 60 and 90 g of ZnO nanoparticles
cles. had the same length and width but were 1.6 and 2.4 cm thick,
Thus, in this study, semi-pilot scale tests were conducted respectively.
to evaluate the selected operational parameters, and the most Contrary to the packed bed filter design, the pore open-
effective combination of the operational parameters was ings of the honeycomb filter structure of the FBAFS (fig. 1c)
then evaluated in room-scale tests. Specifically, the objec- were not completely filled with nanoparticles, and the filter
tives of this study were to: (1) determine the effects of the thickness was kept constant (3.2 cm) regardless of the nano-
selected operational parameters, i.e., filter design, face ve- particle loading. This was done intentionally to leave air
locity, and nanoparticle loading capacity, on the perfor- space in the filter, which allowed fluid-like behavior of the
mance of a filter with ZnO nanoparticles in semi-pilot scale particle-air mixture when the contaminated airstream passed
experiments; (2) evaluate the effectiveness of a filter with through the filter. The fluidized bed filters measured 11.4 cm
ZnO nanoparticles in reducing H2S, NH3, and odor emis- long  10.2 cm wide  3.2 cm thick and were loaded with
sions in room-scale experiments carried out under conditions 15, 30, or 60 g of ZnO nanoparticles, corresponding to load-
representing typical swine production settings; and (3) as- ing rates of 0.14, 0.28, or 0.56 g cm-2 of filter area, respec-
sess the economic feasibility of applying this technology in tively.
commercial swine production operations. Each filtration system was subjected to three face veloci-
ties of 0.2, 0.5, and 0.8 m s-1, corresponding to volumetric
flow rates of 100, 250, and 400 L min-1, respectively.
MATERIALS AND METHODS Experimental Procedure
The basic approach used in this study was adapted from Figure 2 is a schematic diagram of the general setup for
Alvarado et al. (2015), with some modifications in the setup, each test using the different filtration systems. A centrifugal
parameters, procedures, and equipment. Consequently, Al- fan (model Keho, Edwards Group, Lethbridge, Alberta, Can-
varado et al. (2015) is referenced with respect to procedures ada; 3 hp, 230 V, 60 Hz) was used to draw gaseous emissions
and equipment that are common to both studies. from the headspace above 30 L of liquid swine manure
stored in a sealed 75 L plastic vessel. Fan speed was con-
trolled using an adjustable AC inverter (Leeson Speedmas-

944 TRANSACTIONS OF THE ASABE

 (a)

Filter pad

Nanoparticles
Filter pad and air space
Nanoparticles

Filter pad Filter pad

(b) (c)
Figure 1. (a) Photo of honeycomb structure with 1.5 cm  1.5 cm pore opening and schematic diagrams of (b) packed bed and (c) fluidized bed
filters with both filters loaded with 60 g of ZnO nanoparticles.

Downstream gas sampling port

Test filter

Anemometer Fl

Upstream gas sampling port

Exhaust gas

Manure container
Centrifugal fan

Figure 2. Schematic diagram of setup used for the semi-pilot scale tests.

ter, Mississauga, Ontario, Canada), and air velocity in the ation. The test filter was connected to the centrifugal fan and
duct was monitored and recorded at the start and end of each to the plastic vessel containing stored manure by 10 cm di-
experiment using an anemometer (model 8330 VelociCheck, ameter galvanized sheet metal ducts.
TSI Inc., St. Paul, Minn.; accuracy of 5% of reading) in- Prior to the test, the 75 L plastic vessels were filled with
stalled upstream (about 10 cm) of the test filter under evalu- liquid swine manure collected from one of the underfloor

60(3): 943-956 945

manure pits in a swine grow-finish room at the barn facility stream sampling ports at 1, 3, 5, 10, 15, 20, 25, and 30 min
of Prairie Swine Centre Inc. (PSCI) in Saskatoon, Saskatch- from the start of the test. To account for the effect of the
ewan, Canada, sealed, and then stored at ambient conditions synthetic filter pad material used to hold the nanoparticles
for about three weeks. At the start of each test, the fan was within the filter, a control filtration system using only the
set to the desired speed corresponding to one of the face ve- filter pads (without ZnO nanoparticles) was tested at the
locities to be tested and was re-adjusted after the test filter same face velocity.
was placed in the filter compartment. The manure in the In the second phase of the tests, the performance of the
plastic vessel was agitated for 2 min prior to sampling in or- filters at varying loading capacities of ZnO nanoparticles
der to generate measureable concentrations of the target was assessed based on 50% breakthrough time. The 50%
gases (i.e., H2S, NH3). Manure agitation was done continu- breakthrough time was determined as the period of time
ously using a motorized mechanical mixer. With the opera- from the start of the test until the filter’s performance de-
tion of the centrifugal fan, headspace gas was extracted from creased to 50%, i.e., calculated as the percent reduction in
the plastic vessel, which then passed through the anemome- initial (upstream) concentration of the target gas after pass-
ter, through the filter being tested, and finally discharged ing through the filter. For instance, if the upstream H2S con-
outside the barn through the exhaust outlet of the fan (fig. 2). centration was 100 mg L-1, then the 50% breakthrough time
Gas samples were obtained via two sampling ports, one lo- was the time required for the concentration of H2S in the ma-
cated upstream (about 15 cm) and the other downstream nure gas stream to reach 50 mg L-1 after passing through the
(about 10 cm) of the test filter under evaluation. During the filter. Thus, a filter with a longer breakthrough time indi-
test, the manure gas that passed through the filter was a mix- cated that the filter was generally more effective in removing
ture of the gases emitted from the agitated manure slurry and the target contaminant gases.
the ambient air aspirated into the manure container through
openings in the container lid to make up for the headspace ROOM-SCALE TESTS
gas extracted from the container. In the room-scale tests, the effectiveness of a filter system
In the first phase of the experiments, only one manure ves- installed in an air recirculation system was investigated us-
sel was used to generate the gases. After about 500 L of ma- ing the filter design, rated face velocity, and nanoparticle
nure gas had passed through the filter, gas samples of about loading capacity identified in the semi-pilot scale tests.
5 L were drawn from the upstream and downstream sampling
Test Rooms and Experimental Setup
ports into 10 L Tedlar bags using an apparatus with a sampling
Two identical and fully instrumented environmental
lung setup. The Tedlar bags and sampling lines in the gas sam-
chambers at the PSCI barn facility, each measuring 4.2 m 
pling apparatus were flushed twice with nitrogen prior to sam-
pling to minimize sample loss (adhesion or adsorption) to the 3.6 m  2.7 m high, were used in this study. The two cham-
walls of the tubing and bags. To achieve a relatively constant bers and an adjacent instrumentation room were housed in a
amount (500 L) of gas passing through the filter, the length of building equipped with a heating, ventilation, and air condi-
time of filter exposure to manure gas was varied depending on tioning (HVAC) system described by Alvarado et al. (2015).
the airflow rate, i.e., 5 min for 100 L min-1, 2 min for 250 L For the purpose of this study, a ventilation air recirculation
min-1, and 1.25 min for 400 L min-1. The performance of the system was installed in each chamber (fig. 3). The air recir-
different filter designs at varying face velocities was evaluated culation system was designed to maintain stable air circula-
based on their adsorption capacities, which refers to the max- tion patterns, which resulted in good mixing of the inlet air
imum amount of adsorbate (i.e., H2S and NH3) that can be ad- and room air (Deurloo et al., 1990). Operated using an elec-
sorbed per unit mass of adsorbent (ZnO nanoparticles). Ad- tronic speed controller (model MSC-4, Phason, Winnipeg,
sorption capacity (q) was computed as (Volesky, 2003): Manitoba, Canada), the recirculation fan (Godro Multifan,
Vostermans Ventilation, Bloomington, Ill.) drew in contam-
q
Ci  C f  V (1)
inated air from the manure pit area through a 36 cm diameter
m galvanized duct and then through a 38 cm  38 cm filter. The
fan was set to recirculate air at a rate of 105 L s-1 based on
where the recommended recirculation rate of 7 L s-1 m-2 of floor
q = adsorption capacity (mg mg-1) area for pig housing systems (Darby and Dill, 1988). An an-
Ci = initial concentration of adsorbate (mg L-1) emometer (model 8330 VelociCheck, TSI Inc., St. Paul,
Cf = final concentration of adsorbate (mg L-1) Minn.; accuracy of 5% of reading) was used to measure the
V = volume of air (L) air velocity, which was subsequently used to determine the
m = mass of adsorbent (ZnO nanoparticles) (mg). air recirculation rate. The filter in the air recirculation system
For phase 2, with a constant face velocity, manure gas was sized to meet the required 0.5 m s-1 face velocity deter-
was passed through each test filter for a period of 30 min. To mined from the semi-pilot scale tests while maintaining the
maintain the levels of H2S and NH3 in the manure gas stream recommended output from the recirculation fan. Treated air
passing through the filter, six manure vessels were used con- was then distributed into the room airspace via a 30 cm di-
secutively as gas sources (i.e., one container was used for ameter duct installed on the downstream side of the recircu-
about 5 min and then switched to the next container), with lation fan. The duct had eight equally spaced 3.8 cm diame-
each vessel agitated continuously during the test. Concentra- ter holes that served as outlets for the filtered air to flow back
tions of the target gases (H2S and NH3) were determined into the chamber airspace.
from 5 L gas samples extracted from the upstream and down- In the semi-pilot scale tests, an FBAFS with a loading ca-

946 TRANSACTIONS OF THE ASABE

 2.0 m
Door

0.3 m

Concrete slats

Underfloor manure 1.2 m

collection tub

Water
drinker

Solid floor

2.1 m

Feeder

Air inlet

Air outlet
Recirculation
0.6 m
system

1.5 m 1.5 m 0.6 m

(a)
Air inlet
Recirculation fan
Human level
Air outlet

2.2 m
Test filter

Floor Animal level

0.05 m
1.1 m
Pit level
Underfloor manure
collection tub
(b)
Figure 3. (a) Layout and (b) side view of chamber with ventilation air recirculation system used for room-scale tests (not to scale).

pacity of 30 g performed optimally and was selected for the Experimental Procedure
room-scale tests. However, the 30 g loading capacity, which At the start of each test, eight female grower pigs with
was equivalent to 0.28 g cm-2 of filter area, was adjusted to initial weights of about 28 to 30 kg were housed in each
compensate for the larger filter (38 cm  38 cm) used in the chamber; the average initial weights of pigs in both cham-
room-scale tests. Thus, the test filter installed in one of the bers were within 1 kg of each other. The pigs were fed
chambers (treatment room) was loaded with about 405 g of standard grow-finish diets that were weighed prior to place-
ZnO nanoparticles. In the other chamber (control room), a ment in the feeders. Air temperature in each chamber was set
filtration system using only the filter pads made of 100% to 21°C in the first week and gradually decreased to 18°C by
synthetic material (without ZnO nanoparticles) was in- week 4, following standard temperature guidelines for
stalled. Both filters had a thickness of 3.2 cm and were as- grower pigs (PSCI, 2000).
sembled similar to the FBAFS used in the semi-pilot scale Three replicate trials were conducted, each lasting
tests. 30 days. The first 15 days of the trial served as the manure

60(3): 943-956 947

accumulation period. On days 15, 20, 25, and 30, gas sam- sample collection in accordance with the procedures out-
ples (NH3 and H2S) were collected while the manure was lined by BSI (2003).
being agitated. The samples obtained on day 15 were col- Phases 1 and 2 of the semi-pilot scale tests and the room-
lected prior to installing the filtration system. The Tedlar scale tests were conducted in triplicate. A randomized com-
bags used for storing gas samples, as well as the sampling plete block design with split-plot analysis was used to deter-
lines, were flushed twice with nitrogen prior to sampling mine the most effective filter design and face velocity in
(Alvarado et al., 2015). phase 1 of the semi-pilot scale tests. The test filters (treat-
On each sampling day, NH3 and H2S samples were col- ment) were considered the main plots, while filter designs
lected at three locations: about 5 cm above the manure sur- and face velocity were the sub-plots. In determining the most
face (pit level), at 0.5 m above the pen floor to represent the effective loading capacity in phase 2 of the semi-pilot scale
concentrations at the estimated animal height (animal level), tests as well as in evaluating the effectiveness of an FBAFS
and at 1.6 m above the pen floor based on the estimated av- with ZnO nanoparticles in the room-scale tests, a random-
erage worker height (human level). Similar sampling heights ized complete block design with repeated measures analysis
were used by Moreno et al. (2009) in a previous study in was performed. In all tests, each time-dependent replicated
these same chambers. The NH3 and H2S samples were col- trial was treated as a blocking factor, within which the filtra-
lected 2, 5, and 10 min after manure agitation was initiated. tion system treatments were evaluated randomly.
Odor samples were also collected at the three locations in Prior to conducting the statistical analysis, the data were
each chamber on days 15, 20, and 30; however, only one set checked for normality using the Shapiro-Wilk test. If the
of samples was collected 2 min after manure agitation was data followed a non-normal distribution, a log-transfor-
initiated. mation was performed before conducting the statistical tests.
Following filter installation on day 15, the recirculation Analyses of the main and interaction effects of the independ-
air velocity in each chamber was measured 36 cm before the ent variables (filter design, face velocity, and loading capac-
filter and 36 cm after the fan (along the recirculation duct) ity) on the dependent variables (H2S, NH3, and odor concen-
with an anemometer (model 8330 VelociCheck, TSI Inc., St. trations) were done using the SAS Mixed Model procedure
Paul, Minn.; accuracy of 5% of reading) and was then used with  = 0.05 (SAS Institute Inc., Cary, N.C.). If significant
to set the air recirculation rate. Both parameters (recircula- differences (p < 0.05) were observed between means, the
tion air velocity and flow rate) were rechecked daily from Tukey post-hoc method was used to compare the means and
day 16 until the end of the trial to ensure that the air velocity consequently determine which pairs of means were signifi-
passing through the filter and the flow rate delivered by the cantly different. Tukey’s test was used to minimize the prob-
fan to the recirculation duct were within the desired settings ability of making at least one incorrect decision, classified as
by 0.05 m s-1 and 5 L s-1, respectively. a family-wise type I error (Mendenhall and Sincich, 2007).
Additionally, air temperature and relative humidity were PDMix800 Macro letter groupings (Saxton, 1998) were used
monitored regularly using temperature-humidity sensors to depict significant differences between treatment means.
(Hobo U12-013, Onset Computer Corp., Bourne, Mass.; ac-
curacy of 2.5% of reading), and the ventilation rate through COST ASSESSMENT
each chamber’s exhaust air duct was controlled using IRIS The cost of installing and operating recirculation air fil-
dampers (Continental Fan Manufacturing, Buffalo, N.Y.; ac- ters filled with ZnO nanoparticles in a typical swine opera-
curacy of 5% of reading) to ensure that both rooms were tion was assessed based on the assumption that the filters
subjected to similar environmental conditions throughout were installed in a 100-head grow-finish room (20 to
the trial. The IRIS dampers had airflow taps where the pres- 110 kg). The installation was based on the room-scale exper-
sure drop across the damper was measured. The pressure iment setup in which an FBAFS with ZnO nanoparticles was
drop was then used to determine the exhaust airflow rate added to the ventilation system composed of a horizontal re-
through the damper orifice. circulation duct spanning the length of the room with a ver-
Furthermore, the effects of the two filtration system treat- tical duct connected to one end. The filter, situated in the
ments on average daily gain (ADG) and feed intake (ADFI) vertical duct, had dimensions of 61 cm (L)  61 cm (W) 
of the pigs, water usage, and manure production rate were 8 cm (T), similar to commercial air filters. Thus, all the ex-
also assessed. penses associated with one complete growth cycle in a grow-
finish room were estimated, including material (nanoparti-
GAS ANALYSIS AND STATISTICAL PROCEDURES cles), equipment, labor, and operating costs. A comparison
Levels of ammonia in the Tedlar sampling bags were was made between the estimated cost associated with this
measured using a NH3 analyzer (Chillgard RT, MSA Can- potential nanoparticle gas filtration technique and the overall
ada, Edmonton, Alberta; accuracy of 2 ppm; detection limit cost of production. The goal of this cost analysis was to ob-
of 1 ppm), while hydrogen sulfide was analyzed using an tain an initial indication of the feasibility of the technology
H2S monitor with electrochemical sensor (PAC III, Draeger for application in a commercial swine farm, mainly to serve
Canada Ltd., Mississauga, Ontario; accuracy: 5% of read- as basis for making a reasoned decision about whether a
ing; detection limit of 1 ppm). Odor samples were sent to an more comprehensive evaluation in a full-scale swine opera-
olfactometry laboratory at the University of Saskatchewan, tion would be warranted.
Saskatoon, Saskatchewan, and analyzed within 30 h after

948 TRANSACTIONS OF THE ASABE

RESULTS AND DISCUSSION ture that provided better interaction and more exposed surface
SEMI-PILOT SCALE TESTS area for physical and chemical reactions in the fluidized bed
Phase 1: Effects of Filter Design and Face Velocity filter. Fluidized bed designs promote vigorous agitation of the
adsorbent (solid) in contact with the adsorbate (fluid), leading
The H2S and NH3 adsorption capacities of the different fil-
to excellent contact of the solid particles with the fluidization
ter bed designs at varying face velocities are plotted in fig-
medium (Mazumber, 2010; Garic-Grulovic et al., 2011). In
ure 4. Fluidized bed filters exhibited higher adsorption capac-
addition, the adsorption capacities of the two filter bed designs
ities than packed beds for both H2S and NH3. On average, flu-
appear not to have been affected by the initial concentrations
idized bed filters achieved an adsorption capacity of about
of H2S and NH3. On average, the initial concentrations (before
4.82 mg H2S adsorbed per g ZnO and 2.61 mg NH3 adsorbed
treatment) of H2S and NH3 for the tests with fluidized bed fil-
per g ZnO, while packed bed filters had 2.0 mg H2S adsorbed
ters were 221 and 32 mg L-1, respectively, while the initial
per g ZnO and 1.0 mg NH3 adsorbed per g ZnO. Statistical
concentrations for packed bed filters were 234 and 29 mg L-1,
tests indicated that the differences in adsorption capacities of
respectively.
the two types of filter bed designs were significant (p <
On average, the highest H2S adsorption capacity was at-
0.0001). This implies that, on average, the fluidized bed filters
tained when the gas stream passed through the filters at a face
were better in reducing the levels of H2S and NH3 emitted
velocity of 0.5 m s-1 regardless of the amount of ZnO nano-
from the manure than the packed bed filters, which can be at-
particles loaded into the filters. This corresponds to an average
tributed to the fluid-like behavior of the nanoparticle-gas mix-

12.0
0.2 m/s 0.5 m/s 0.8 m/s

9.0
H2S Adsorbed (mg per g ZnO)

6.0

3.0

0.0
30 g 60 g 90 g 15 g 30 g 60 g
Packed bed Fluidized bed

(a)
12.0

0.2 m/s 0.5 m/s 0.8 m/s

9.0
NH3 Adsorbed (mg per g ZnO)

6.0

3.0

0.0
30 g 60 g 90 g 15 g 30 g 60 g
Packed bed Fluidized bed

(b)
Figure 4. Mean (SD) adsorption capacity of (a) H2S and (b) NH3 for two types of filters with ZnO nanoparticles at varying face velocities during
the semi-pilot scale tests (n = 3).

60(3): 943-956 949

of about 6.92 mg H2S adsorbed per g ZnO when the filters more sites available for surface reaction due to the greater
were subjected to a 0.5 m s-1 face velocity; this adsorption ca- amount of nanoparticles in the filter. Wang et al. (2008) ob-
pacity was not statistically different from 0.2 m s-1 (p = 0.151) served a similar trend when they investigated the removal of
but significantly different from 0.8 m s-1 (p = 0.0165). On av- H2S from gas streams using ZnO nanoparticles. They re-
erage, the adsorption capacity of filters subjected to 0.2 m s-1 ported that the breakthrough time increased from 6 to 39 min
was about 5.58 mg H2S adsorbed per g ZnO, while that of fil- when the ZnO loading amount was increased from 0.50 to
ters subjected to 0.8 m s-1 was about 1.96 mg H2S adsorbed 3.04 wt%. The low initial reduction in H2S achieved with the
per g ZnO. In the fluidized bed filter design, the face velocities 60 g filter was due to the random variation in initial gas lev-
of 0.2 and 0.5 m s-1 caused good fluidization (based on visual els from the source (manure container). On average, the ini-
observations) of the ZnO nanoparticles in the filter bed as well tial H2S concentration for the tests with the 60 g filter was
as longer hydraulic retention time (compared to a face velocity about 453 mg L-1, while the concentrations with the 30 and
of 0.8 m s-1), resulting in longer contact time and consequently 15 g filters were 378 and 351 mg L-1, respectively.
increased interaction between the particles and the target Compared to H2S, all treated filters had lower initial NH3
gases. Other studies (Novochinskii et al., 2004; Li and King, reductions, leading to lower 50% breakthrough times of
2006; Habibi et al., 2010) also reported greater H2S adsorption around 15 min for 30 g, 7 min for 15 g, and 3 min for 60 g
capacity of ZnO at lower face velocity. of ZnO (fig. 5b). Similar to the phase 1 results, this could
Unlike H2S, the NH3 adsorption capacity of fluidized bed possibly be attributed to ZnO nanoparticles not being the
filters was not affected by face velocity. It could be that ZnO best material for NH3 adsorption. It was also observed that
is not the best material for adsorbing NH3; thus, any changes the 60 g filter had the lowest 50% breakthrough time, even
in the conditions that helped with H2S did not have the same though it showed a gradual decrease in concentration reduc-
impact for NH3. Similar results were observed in the study tion with time, mainly because its initial reduction was al-
by Alvarado et al. (2015) in which ZnO nanoparticles were ready close to 50%, the lowest among all nanoparticle-
mixed into swine manure slurry. Other studies (Chung et al., loaded filters.
2005; Awume, 2014) also reported relatively poor perfor- Comparatively, the control filter did not even achieve
mance of ZnO in NH3 adsorption. 50% reduction for either H2S or NH3 at the first minute of
Although the adsorption capacity of fluidized bed filters exposure. This was expected, as the untreated filter did not
when exposed to 0.5 m s-1 face velocity was not significantly contain any ZnO nanoparticles, and it affirms the effective-
different (p = 0.151) from 0.2 m s-1, a face velocity of 0.5 m ness of ZnO nanoparticle filtration systems for purifying pol-
s-1 was selected for the subsequent room-scale tests because, luted air streams.
on average, it resulted in higher gas capture (6.92 vs.
5.58 mg H2S adsorbed per g ZnO) with respect to the differ- ROOM-SCALE TESTS
ent loading capacities. Furthermore, a face velocity of 0.5 m Effect of Filtration Treatment on H2S, NH3,
s-1 would enable a larger volume of air to be treated com- and Odor Concentrations
pared to a face velocity of 0.2 m s-1, and consequently the Figure 6 shows the concentrations of H2S and NH3 meas-
room airspace could be purified in less time. ured at specific locations in each chamber 2 min after ma-
nure agitation was initiated. Prior to the operation of the air
Phase 2: Effect of ZnO Nanoparticle Loading Capacity recirculation system in the control chamber, the H2S concen-
The effects of ZnO nanoparticle loading capacity on the trations of samples collected at the pit level, animal level,
filtration effectiveness of a fluidized bed filter exposed to a and human level on day 15 were 326, 82, and 58 mg L-1,
face velocity of 0.5 m s-1 over an extended period of time are respectively, while the corresponding NH3 concentrations
shown in figure 5. At the different loading capacities, the gas were 22, 12, and 10 mg L-1, respectively. On subsequent
reduction effectiveness of all the filters decreased with time sampling days after the filter was installed (i.e., days 20, 25,
and was significantly greater (p < 0.0001) than the reduction and 30), no significant changes were observed in H2S (p =
by the control filtration system. For H2S (fig. 5a), during the 0.156) and NH3 (p = 0.151) concentrations at each sampling
first minute of gas exposure, 30 g of ZnO nanoparticles de- location in the control chamber. This suggests that the con-
creased H2S to about 72%, which gradually decreased to centrations of the target gases emitted from the manure (pit)
50% reduction after 26 min. The filter with 15 g nanoparti- did not vary significantly between days 20 and 30, and the
cles, on the other hand, achieved about 65% H2S reduction installation of the air recirculation system in the control
initially and showed a sharper decrease thereafter, resulting chamber had no significant impact on human and animal ex-
in a shorter 50% breakthrough time of 10 min. The sharp posure to H2S and NH3 with time.
decrease could be attributed to the lesser amount of nanopar- With the FBAFS treatment, the H2S and NH3 concentra-
ticles in the filter, which likely became fully saturated with tions at the pit level were not significantly different over the
target gas molecules more quickly. Awume (2014) noted duration of the trial (p = 0.593 and p = 0.338, respectively).
that the adsorbent (ZnO nanoparticles) achieved a saturated This was expected, considering that the FBAFS with ZnO
state faster when its quantity was reduced regardless of the nanoparticles in the treatment chamber was intended to treat
upstream H2S concentration. On the other hand, the filter the target gases after being emitted from the source (manure
with 60 g ZnO nanoparticles had the lowest initial H2S re- pit). The air recirculation system with an FBAFS was set up
duction but had the longest 50% breakthrough time (about in such a way that contaminated air from an area close to the
30 min) among all filters. The decrease in percent reduction source was drawn into the inlet of the recirculation duct,
with time was very gradual, which could be attributed to passed through the FBAFS, and then the treated air was ex-

950 TRANSACTIONS OF THE ASABE

 80
0 g (filter pad only) 15 g 30 g 60 g
70

60

H2S Reduction (%)

50

40

30

20

10

0
0 5 10 15 20 25 30

Sampling Time (min)

(a)
80
0 g (filter pad only) 15 g 30 g 60 g
70

60
NH3 Reduction (%)

50

40

30

20

10

0
0 5 10 15 20 25 30
Sampling Time (min)

(b)
Figure 5. Percent reduction of (a) H2S and (b) NH3 in manure gases passed through filters with varying loadings of ZnO nanoparticles (n = 3).
Horizontal dashed line represents concentration reduction for 50% breakthrough time.

pelled back into the room airspace. However, at the animal the ZnO nanoparticle filtration system compared to the con-
level and human level in the FBAFS treatment chamber, sig- trol chamber, where there were mostly increases in the con-
nificant reductions (p = 0.004 and p = 0.002, respectively) in centrations of both gases with time, i.e., relative to the con-
the concentrations of the target gases relative to their initial centrations of the samples obtained on day 15.
concentrations prior to operation of the filtration system Interestingly, the reductions in gas concentrations at the
were observed over the 15-day assessment period. At the an- human level in the FBAFS treatment were higher than those
imal level, initial mean H2S and NH3 concentrations were 94 at the animal level. This could be due to the observed airflow
and 15 mg L-1, respectively, and decreased to 46 and 10 mg pattern in the chamber during sampling. Not all gases emit-
L-1, respectively, on day 30, which was 15 days after the fil- ted from the manure in the pit were completely captured in
ter with ZnO nanoparticles had been installed in the cham- the recirculation duct; a portion may have been dissipated
ber. This translates to reductions of about 46% and 36% in directly upward through the floor slats and most likely cap-
the H2S and NH3 concentrations, respectively, over the 15- tured at the animal level sampling location, which was about
day operation of the ventilation air recirculation system. 0.5 m above the pen floor. On the other hand, the recircula-
Over the same period, mean initial H2S and NH3 concentra- tion duct expelled treated air through outlets directed toward
tions at the human level were 58 and 12 mg L-1, respectively, the human level.
and decreased to 23 and 7 mg L-1, respectively, by day 30, As observed from the results of the semi-pilot scale and
which were equivalent to reductions of about 65% and 42%, room-scale tests, the reduction in H2S levels achieved after
respectively. These reductions can be attributed mainly to passing through the filter with ZnO nanoparticles could

60(3): 943-956 951

Figure 6. Mean (SD) concentrations of H2S and NH3 in gas samples collected from the untreated (control) chamber and the chamber with the
fluidized bed air filtration system (FBAFS) 2 min after manure agitation was initiated (n = 3). Means with the same letters within the same
treatment group are not significantly different (p > 0.05). Means with no letters within the same group and within the same sampling location are
not significantly different (p > 0.05).

probably be due to a chemical adsorption process, or chem- lower capacity to decompose NH3 than ZnO (Park et al.,
isorption, that occurred between H2S and ZnO and produced 2011). Generally, metal oxides such as ZnO are easily con-
insoluble zinc sulfide and water (Sanchez et al., 2005; Li and verted to sulfide form during the adsorption process. An-
King, 2006; Wang et al., 2008), as shown in equation 2: other potential reason is that the oxygen in ZnO would be
consumed during the reaction between ZnO and H2S; if the
ZnO + H2S ↔ ZnS + H2O (2) oxygen is not sufficient, H2O cannot be produced. The H2
from H2S decomposition then hinders the decomposition of
Compared to H2S, the reduction in NH3 achieved by the
NH3 to N2 and H2 due to equilibrium limitation (Jun et al.,
filters with ZnO nanoparticles was relatively low. This could
2003).
be due to the production of metal sulfide (ZnS), which has

952 TRANSACTIONS OF THE ASABE

 The effects of the FBAFS treatment on odor concentra- Table 1. Average daily water usage, manure production rate, feed
intake, and gain of pigs in the FBAFS treatment and control chambers
tions compared to the control are shown in figure 7. Prior to
during room-scale tests. Values are means (SD) of three replicates.
installing the filtration system on day 15, the odor concen- Eight grower pigs were housed in each chamber per replicate.
trations at all sampling locations in both chambers were not FBAFS
significantly different (p = 0.661), with average concentra- Parameter Treatment Control
tions of 2551 and 2536 OU m-3, respectively. In the control Average daily water usage (L d-1 pig-1) 2.4 0.7 2.1 0.4
chamber, the concentrations varied and did not differ signif- Average daily manure production (L d-1 pig-1) 2.09 0.05 2.07 0.12
Average daily feed intake (ADFI, kg d-1 pig-1) 1.70 0.13 1.68 0.20
icantly with time for samples collected at the human level Average daily gain (ADG, kg d-1 pig-1) 0.81 0.06 0.80 0.07
(p = 0.81) and the animal level (p = 0.934) over the 15 days
after the filtration systems were installed. At the pit level,
operations (PIC, 2013). In addition, pigs in both chambers
odor concentration decreased with time, but not significantly
had zero mortalities. Thus, the improvement in indoor air
(p = 0.286).
quality by reducing the concentrations of H2S, NH3, and odor
In the FBAFS treatment chamber, odor concentrations
in the FBAFS treatment did not have a noticeable influence
decreased at the human level from 2087 OU m-3 prior to in-
on pig performance, most likely due to the short study period
stalling the air filtration system on day 15 to 1689 OU m-3
(only 30 days).
on day 20. Odor concentrations decreased at the animal level
from 2888 to 2278 OU m-3 over the same period; however,
COST ANALYSIS OF ZNO NANOPARTICLES
the differences were not significant (p = 0.237). The de-
IN SWINE FACILITIES
creases in odor concentration, especially at the human level,
Table 2 provides an estimate of the individual and total
could be attributed to the combined effects of contaminant
costs associated with the installation and operation of an
gas removal and dilution of contaminant gas levels as a result
FBAFS with ZnO nanoparticles in a recirculation air duct in
of the operation of the FBAFS with ZnO nanoparticles.
a pig barn for a year. One assumption was that the ZnO na-
Effect on Pig Performance noparticles in the FBAFS were replaced four times (i.e.,
During the entire 30-day trial period, pigs raised with the every 30 days throughout the 16-week growth period for
FBAFS treatment consumed more water and feed than those grow-finish pigs), specifically on weeks 0, 4, 8, and 12. This
in the control chamber by about 0.3 L d-1 pig-1 and 0.02 kg assumption was derived from observations made during the
d-1 pig-1, respectively (table 1), but the differences were not room-scale tests. Although the filter pad was coated with
significant (p = 0.882 and p = 0.545, respectively). Further- dust after the FBAFS was operated continuously for 15 days,
more, the average daily gain (ADG) associated with the the color of the impregnated pure white ZnO nanoparticles
FBAFS treatment was not significantly different (p = 0.766) did not change completely to yellowish white. This suggests
from the gain observed in pigs raised in the control chamber. that the nanoparticles were not yet fully saturated and could
In general, the average daily feed intake (ADFI) and ADG possibly be used effectively for another 15 days or longer.
values in this study were within the recommended range Additionally, dust accumulation on the FBAFS filter cover
(ADFI of 1.41 to 1.99 kg d-1 pig-1 and ADG of 0.78 to 0.91 can be minimized by installing a dust filter before the
kg d-1 pig-1) observed in pigs raised in typical commercial FBAFS, thereby extending the life of the treatment filter

4000
Pit Animal Human
Odor Concentration (OU m‐3)

3000

2000

1000

0
Day 15 Day 20 Day 30 Day 15 Day 20 Day 30
FBAFS Control

Figure 7. Mean (SD) odor concentrations at the pit level, animal level and human level in the control and fluidized bed air filtration system
(FBAFS) treatment chambers in the room-scale tests. No significant change (p > 0.05) in odor concentrations was observed after the filters were
installed on day 15.

60(3): 943-956 953

Table 2. Cost breakdown associated with operation of a recirculation With this in mind, the FBAFS operational specifications
air FBAFS in a grow-finish pig barn.[a]
on which the above cost analysis was based could possibly
Specification Value
ZnO nanoparticles application rate 0.28 g cm-2
be modified by extending its equipment life. Presumably, ex-
Number of applications per cycle Four tending the FBAFS life would decrease the replacement fre-
Total amount of ZnO applied per room 4.1 kg quency of the filter. For instance, decreasing the FBAFS re-
ZnO nanoparticles unit price per kg[b] $87.7 kg-1 placement frequency from four times to two times per
ZnO nanoparticles cost per pig $4.0
Number of hours to apply treatment per cycle 5h
growth cycle (i.e., on the 1st and 10th weeks of the cycle)
Labor cost per hour $13.0 h-1 could subsequently reduce the total cost per pig from
Labor cost per pig $0.5 CAD$6.29 to CAD$4.10. Furthermore, because many barns
Total cost for installation and required equipment[c] $5930 in Canada already have air recirculation systems, additional
Estimated equipment life Five years
savings can be achieved by eliminating the cost associated
Total cost of required material per pig $1.6
Capital cost per pig $1.77 with the need to install an air recirculation system.
Estimated energy consumption per year 1481 kWh
Energy cost per kWh[d] $0.10453 kWh-1
Operating cost per pig $0.02
Total cost per pig $6.29 CONCLUSIONS
[a]
All costs are in Canadian dollars (CAD$). Semi-pilot scale tests indicated that the effectiveness of
[b]
Based on the purchase price of ZnO nanoparticles. air filtration systems with ZnO nanoparticles was signifi-
[c]
Includes estimated installation cost ($2000) using contracted labor. cantly influenced by filter design, face velocity, and loading
[d]
SaskPower rate (2015).
capacity of nanoparticles per unit area of the filter. Further-
more, the results showed that an FBAFS at a loading capac-
with ZnO nanoparticles by 30 days or longer.
ity of 30 g (equivalent to 0.28 g cm-2 of filter area) exposed
Furthermore, based on the observed effective application
to a face velocity of 0.5 m s-1 was most effective in reducing
rate of 0.28 g cm-2 of filter area, the total amount of ZnO
H2S and NH3 emissions from agitated liquid pig manure.
nanoparticles needed for one complete growth cycle was es-
In a room-scale study with an FBAFS installed in the re-
timated to be 4.1 kg, amounting to about CAD$4.00 per pig.
circulation air duct, the filtration system was capable of re-
The time required for ZnO filter preparation and installation
ducing the concentration of H2S by 65% at the human level
was estimated to be about 1.5 h, which translates to about
and 46% at the animal level and reducing the concentration
CAD$0.50 labor cost per pig. In addition, the capital costs
of NH3 by 42% at the human level and 36% at the animal
included the cost of the required materials, equipment, and
level, in comparison to a control chamber without the
installation for the recirculation air system, amounting to
FBAFS. However, the reduction in odor concentration at the
CAD$4100, with additional cost for the FBAFS estimated at
human level was minimal at about 19%.
about CAD$1830. Assuming a five-year equipment life, the
With regard to pig performance, the FBAFS treatment ap-
total capital cost was estimated to be about CAD$1.77 per
peared to have no noticeable adverse impact on average
pig.
daily gain of the pigs. The initial cost of installing and oper-
The annual energy consumption of the ventilation air re-
ating an FBAFS with ZnO nanoparticles in a grow-finish pig
circulation system with a FBAFS, calculated based on the
barn was estimated to be about CAD$6.29 per pig.
average pressure loss, airflow rate, total number of operating
Overall, this study more than proved the concept that na-
hours (recirculation system assumed to run continuously
noparticles can be effectively employed for mitigating air
throughout the year), and fan/motor/drive efficiency, was es-
contaminant emissions in swine barns. The initial cost esti-
timated to be about 1481 kWh. Using the local energy billing
mate seems high, suggesting that the merits of this technol-
rate of CAD$0.10453 per kWh, the total operating cost of
ogy need to be explored further (technologically, practically,
the air recirculation system with the FBAFS was estimated
and economically) through a comprehensive evaluation con-
to be CAD$0.02 per pig. Summing up all the costs, the total
ducted at a commercial swine operation.
cost of adding an FBAFS with ZnO nanoparticles to the air
recirculation system of a commercial swine grow-finish
ACKNOWLEDGEMENTS
room was estimated to be about CAD$6.29 per finished pig.
The authors would like to acknowledge the Saskatchewan
The results of the cost analysis suggest that the total costs Agriculture Development Fund and the National Science
associated with the installation of an FBAFS would be about and Engineering Research Council for their financial support
2.7% of the estimated total production cost (CAD$235.08)
of this research. Strategic funding provided by Saskatche-
required to raise a pig in grow-finish operation (MAFRD, wan Agriculture, Saskatchewan Pork Development Board,
2015). It is important to note that this cost estimate was Alberta Pork, Manitoba Pork Council, and Ontario Pork to
based on assumptions (e.g., existing power and labor costs
the research programs at Prairie Swine Centre Inc. (PSCI) is
at the time of the study) and on the findings from the room- gratefully acknowledged. Technical assistance provided by
scale experiments, which were conducted under conditions PSCI barn staff, particularly Sam Gelowitz, Yaomin Jin,
chosen intentionally to produce extremely high levels of the Leila Dominguez, Karu Bandaralage, and John Meier, is
target gases. In other words, because the treatment has been greatly appreciated.
shown to be relatively effective under such extreme condi-
tions, it is presumed that it will perform similarly under typ-
ical commercial barn conditions where lower levels of the
target gases are anticipated.

954 TRANSACTIONS OF THE ASABE

 Hartung, E., Jungbluth, T., & Buscher, W. (2001). Reduction of
REFERENCES ammonia and odor emissions from a piggery with biofilters.
Alvarado, A. C., Predicala, B. Z., & Asis, D. A. (2015). Mixing
Trans. ASAE, 44(1), 113-118.
nanoparticles with swine manure to reduce hydrogen sulfide and
https://doi.org/10.13031/2013.2302
ammonia emissions. Intl. J. Environ. Sci. Tech., 12(3), 893-904.
Hartung, J., & Phillips, V. R. (1994). Control of gaseous emissions
https://doi.org/10.1007/s13762-013-0474-y
from livestock buildings and manure stores. J. Agric. Eng. Res.,
ASABE. (2005). EP470: Manure storage safety. St. Joseph, MI:
57(3), 173-189. http://dx.doi.org/10.1006/jaer.1994.1017
ASABE.
Heber, A. J., Lim, T. T., Tao, P. C., & Ni, J. Q. (2004). Control of
Awume, B. (2014). Control of hydrogen sulphide emissions using
air emissions from swine finishing buildings flushed with
zinc oxide nanoparticles. MS thesis. Saskatoon, Saskatchewan,
recycled lagoon effluent. ASAE Paper No. 044156. St. Joseph,
Canada: University of Saskatchewan, Department of Chemical
MI: ASAE.
and Biological Engineering.
Hendriks, J., Berckmans, D., & Vinckier, C. (1997). Field tests of
BSI. (2003). BS EN 13725: Air quality: Determination of odour
bio-additives to reduce ammonia emission from pig houses. In
concentration by dynamic olfactometry. London, UK: British
Proc. Intl. Symp. Ammonia and Odor Control from Animal
Standards Institution.
Production Facilities (pp. 707-714). Rosmalen, The
Bushee, E. L., Edwards, D. R., Gates, R. S., & Turner, L. W.
Netherlands: NVTL.
(1999). Effects of chemical amendments on the gaseous
Jacobson, L. D., Hetchler, B., Janni, K. A., & Johnston, L. J. (1998).
emissions from swine manure. ASAE Paper No. 994003. St.
Odor and gas reduction from sprinkling soybean oil in a pig
Joseph, MI: ASAE.
nursery. ASAE Paper No. 984125. St. Joseph, MI: ASAE.
Casey, K. D., Bicudo, J. R., Schmidt, D. R., Singh, A., Gay, S. W.,
Jun, H. K., Jung, S. Y., Lee, T. J., Ryu, C. K., & Kim, J. C. (2003).
Gates, R. S., ... Hoff, S. J. (2006). Air quality and emissions
Decomposition of NH3 over Zn-Ti-based desulfurization sorbent
from livestock and poultry production/waste management
promoted with cobalt and nickel. Catal. Today, 87(1-4), 3-10.
systems. In J. M. Rice, D. F. Caldwell, & F. J. Humenik (Ed.),
https://doi.org/10.1016/j.cattod.2003.10.013
Proc. Animal Agriculture and the Environment: National Center
Kendall, D. C., Richert, B. T., Bowers, K. A., DeCamp, S. A., Herr,
for Manure and Animal Waste Management White Papers. St.
C. T., Weber, T. E., ... Powers, W. J. (2000). Effects of dietary
Joseph, MI: ASABE.
manipulation on pig performance, manure composition, aerial
Chen, L., Hoff, S., Cai, L., & Koziel, J. (2008). Odor reduction
ammonia, hydrogen sulfide, and odor levels in swine buildings.
during biofiltration as affected by airflow rate and media
J. Animal Sci, 79(S1), 59.
moisture content. ASABE Paper No. 083859. St. Joseph, MI:
Kendall, D. C., Richert, B. T., Sutton, A. L., Frank, J. W., DeCamp,
ASABE.
S. A., Bowers, K. A., ... Cobb, M. (1999). Effects of fiber
Chung, Y., Lin, Y., & Tseng, C. (2005). Removal of high
addition (10% soybean hulls) to a reduced crude protein diet
concentration of NH3 and coexistent H2S by biological activated
supplemented with synthetic amino acids versus a standard
carbon (BAC) biotrickling filter. Bioresour. Tech., 96(16), 1812-
commercial diet on pig performance, pit composition, odor, and
1820. https://doi.org/10.1016/j.biortech.2005.01.003
ammonia levels in swine buildings. J. Animal Sci., 77(S1), 176.
Darby, D. E., & Dill, G. (1988). Recirculation air systems. In Proc.
Kim, J.-H., Seo, G., Cho, D.-L., Choi, B.-C., Kim, J.-B., Park, H.-J.,
Banff Pork Seminar. Airdrie, Alberta, Canada: Alberta
... Kato, S. (2006). Development of air purification device
Agriculture.
through application of thin-film photocatalyst. Catal. Today,
DeCamp, S. A., Hill, B. E., Hankins, S. L., Bundy, D. C., &
111(3-4), 271-274. https://doi.org/10.1016/j.cattod.2005.10.058
Powers, W. J. (2001). Effects of soybean hulls in commercial
Kim, Y.-K., Ko, H.-J., Kim, H.-T., Kim, Y.-S., Roh, Y.-M., Lee,
diet on pig performance, manure composition, and selected air
C.-M., & Kim, C.-N. (2008). Odor reduction rate in the
quality parameters in swine facilities. J. Animal Sci., 79(S1),
confinement pig building by spraying various additives.
252.
Bioresour. Tech., 99(17), 8464-8469.
Deurloo, J. A., Feddes, J. J., Leonard, J. J., & Darby, D. E. (1990).
https://doi.org/10.1016/j.biortech.2007.12.082
Effect of air recirculation on airspeeds at animal level. Canadian
Li, L., & King, D. L. (2006). H2S removal with ZnO during fuel
Agric. Eng., 32(1), 129-134.
processing for PEM fuel cell applications. Catal. Today, 116(4),
Gao, C., Wang, Z., & Gbaguidi, E.A. (2010). Ammonium
537-541. https://doi.org/10.1016/j.cattod.2006.06.024
variational trends and the ammonia neutralization effect of acid
MAFRD. (2015). Guidelines for estimating swine (26 to 121 kg)
rain over East Asia. Atmos. Oceanic Sci. Lett., 3(2), 120-126.
finishing cost 2016. Winnipeg, Manitoba, Canada: Manitoba
Garic-Grulovic, R., Boskovic-Vragolovic, N., Grbavcic, Z., &
Agriculture, Food and Rural Development. Retrieved from
Pjanovic, R. (2011). Hydrodynamics and mass transfer in
http://www.gov.mb.ca/agriculture/business-and-
heterogeneous systems. In Advanced topics in mass transfer (pp.
economics/financial-management/pubs/cop_swine_finish.pdf
211-228). Rijeka, Croatia: InTech.
Martinec, M., Hartung, E., Jungbluth, T., Schneider, F., & Wieser,
Godbout, S., Lemay, S. P., Joncas, R., Larouche, J. P., Martin, D.
P. H. (2001). Reduction of gas, odor, and dust emissions from
Y., Bernier, J. F., ... Barber, E. M. (2001). Oil sprinkling and
swine operations with biofilters. ASAE Paper No. 014079. St.
dietary manipulation to reduce odor and gas emissions from
Joseph, MI: ASAE.
swine buildings: Laboratory-scale experiment. In Proc. 6th Intl.
Mazumder, A. (2010). Development of a simulation model for
Symp. on Livestock Environment (pp. 671-678). St. Joseph, MI:
fluidized bed mild gasifier. MS thesis. New Orleans, LA:
ASAE.
University of New Orleans, Department of Mechanical
Govere, E. M., Tonegawa, M., Bruns, M. A., Wheeler, E. F.,
Engineering.
Heinemann, P. H., Kephart, K. B., & Dec, J. (2005).
Megaritis, A. G., Fountoukis, C., Charalampidis, P. E., Pilinis, C.,
Deodorization of swine manure using minced horseradish roots
& Pandis, S. N. (2013). Response of fine particulate matter
and peroxides. J. Agric. Food. Chem., 53(12), 4880-4889.
concentrations to changes of emissions and temperature in
https://doi.org/10.1021/jf0404290
Europe. Atmos. Chem. Phys., 13(6), 3423-3443.
Habibi, R., Rashidi, A. M., Daryan, J. T., & Zadeh, A. M. (2010).
https://doi.org/10.5194/acp-13-3423-2013
Study of the rod-like and spherical nano-ZnO morphology on
Mendenhall, W. M., & Sincich, T. L. (2007). Statistics for
H2S removal from natural gas. Appl. Surf. Sci., 257(2), 434-439.
engineering and the sciences (5th ed.). Upper Saddle River, NJ:
https://doi.org/10.1016/j.apsusc.2010.07.007

60(3): 943-956 955

 Prentice Hall. Evaluating biostimulant effects in swine production facility
Moreno, L., Predicala, B., & Nemati, M. (2010). Laboratory, semi- wastewater. In Proc. State of the Science of Animal Manure and
pilot, and room scale study of nitrite and molybdate mediated Waste Mgmt. Symp. Raleigh, NC: National Center for Manure
control of H2S emission from swine manure. Bioresour. Tech., and Animal Waste Management.
101(7), 2141-2151. Smith, D. R., Moore, P. A., Haggard, B. E., Maxwell, C. V., Daniel,
https://doi.org/10.1016/j.biortech.2009.11.011 T. C., VanDevander, K., & Davis, M. E. (2004). Effect of
Nicolai, R. E., & Janni, K. A. (2001). Determining pressure drop aluminum chloride and dietary phytase on relative ammonia
through compost-wood chip biofilter media. ASAE Paper No. losses from swine manure. J. Animal Sci., 82(2), 605-611.
014080. St. Joseph, MI: ASAE. https://doi.org/10.2527/2004.822605x
Nonami, T., Hase, H., & Funakoshi, K. (2004). Apatite-coated Smith, S., & Nicolai, R. (2005). Hydrogen sulfide reduction of
titanium dioxide photocatalyst for air purification. Catal. Today, swine manure using potassium permanganate and hydrogen
96(3), 113-118. https://doi.org/10.1016/j.cattod.2004.06.112 peroxide. ASAE Paper No. SD05801. St. Joseph, MI: ASAE.
Novochinskii, I. I., Song, C., Ma, X., Liu, X., Shore, L., Lampert, J., Tajik, M., Muhammad, N., Lowman, A., Thu, K., Wing, S., &
& Farrauto, R. J. (2004). Low-temperature H2S removal from Grant, G. (2008). Impact of odor from industrial hog operations
steam-containing gas mixtures with ZnO for fuel cell on daily living activities. New Solutions, 18(2), 193-205.
application: 1. ZnO particles and extrudates. Energy Fuels, https://doi.org/10.2190/NS.18.2.i
18(2), 576-583. https://doi.org/10.1021/ef030137l Volesky, B. (2003). Sorption and biosorption. Montreal, Quebec,
Park, J. J., Park, C. G., Jung, S. Y., Lee, S. R., & Kim, J. C. (2011). Canada: BV Sorbex.
A study on Zn-based catal-sorbents for the simultaneous Wang, S., Nan, J., Shi, C., Fu, Q., Gao, S., Wang, D., ... Zhou, B.
removal of hydrogen sulfide and ammonia at high temperature. (2015). Atmospheric ammonia and its impacts on regional air
Res. Chem. Intermed., 37(9), 1193. quality over the megacity of Shanghai, China. Sci. Rep., 5,
https://doi.org/10.1007/s11164-011-0384-x 15842-15855.
PIC. (2013). Wean to finish manual. Hendersonville, TN: PIC Wang, S., Xing, J., Jang, C., Zhu, Y., Fu, J. S., & Hao, J. (2011).
North America. Retrieved from http://www.pic.com Impact assessment of ammonia emissions on inorganic aerosols
PSCI. (2000). Pork production reference guide (1st Ed.). Saskatoon, in east China using response surface modeling technique.
Saskatchewan, Canada: Prairie Swine Centre Inc. Environ. Sci. Tech., 45(21), 9293-9300.
Sanchez, J. M., Ruiz, E., & Otero, J. (2005). Selective removal of https://doi.org/10.1021/es2022347
hydrogen sulfide from gaseous streams using a zinc-based Wang, X., Sun, T., Yang, J., Zhao, L., & Jia, J. (2008). Low-
sorbent. Ind. Eng. Chem. Res., 44(2), 241-249. temperature H2S removal from gas streams with SBA-15
https://doi.org/10.1021/ie0497902 supported ZnO nanoparticles. Chem. Eng. J., 142(1), 48-55.
Saxton, A. M. (1998). A macro for converting mean separation https://doi.org/10.1016/j.cej.2007.11.013
output to letter groupings. In Proc. 23rd SAS Users Group Intl. Watts, P. J. (1999). Development of a pig effluent emissions
(pp. 1243-1246). Cary, NC: SAS Institute. database and analysis of promising control strategies. Kingston,
Schinasi, L., Horton, R. A., Guidry, V. T., Wing, S., Marshall, S. ACT, Australia: Pig Research and Development Corporation.
W., & Morland, K. B. (2011). Air pollution, lung function, and Williams, C. M., & Schiffman, S. S. (1996). Effect of liquid swine
physical symptoms in communities near concentrated swine manure additives on odor parameters. In Proc. Intl. Conf. on Air
feeding operations. Epidemiol., 22(2), 208-215. Pollution from Agricultural Operations (pp. 409-412). Ames,
Schneegurt, M. A., Weber, D. L., Ewing, S., & Schur, H. B. (2005). Iowa: MidWest Plan Service.

956 TRANSACTIONS OF THE ASABE