EPA-650/2-73-010
October 1973
Environmental Protection
Technology
Series
jmmmammammmmmmssm§iiiiiMi»filillllliii

-------
EPA-650/2-73-010
DEVELOPMENT OF INSTRUMENTATION
FOR MEASUREMENT OF
STATIONARY SOURCE ALDEHYDE,
ORGANIC ACID,
AND AMINE EMISSIONS
by
J . Daniel Bode
Bendix Research Laboratories
Bendix Center
Southfield, Michigan 48076
Report No. 6635
Contract No. 68-02-0551
Program Element No. 1A1010
EPA Project Officer: Fredric C. Jaye
Chemistry and Physics Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
October 1973

-------
This report has been revic -ved by the Environmental Protection Agency and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Agency, nor does
mention of trade names or commercial products constitute endorsement
or recommendation for use.
ii

-------
TABLE OF CONTENTS
Page
PHASE I - PRELIMINARY INVESTIGATIONS	1
SECTION 1 - INTRODUCTION AND SUMMARY	3
SECTION 2 - SAMPLING AND ANALYSIS	7
2.1	Collection Techniques	7
2.2	Analysis Techniques	12
SECTION 3 - ANALYTICAL METHODS FOR FIELD MONITORING	25
3.1	Requirements	25
3.2	Selected Analytical Methods	25
PHASE II - PROTOTYPE DEVELOPMENTS	31
SECTION 4 - INTRODUCTION	33
SECTION 5 - ALDEHYDE MONITOR	35
5.1	Analytical Method	35
5.2	Tape Monitor Design	36
5.3	Laboratory Tests	40
SECTION 6 - AMINE MONITOR	49
6.1	Analytical Method	49
6.2	Tape Monitor Design	49
6.3	Laboratory Tests	50
SECTION 7 - ORGANIC SULFUR MONITOR	55
7.1	Analytical Method	55
7.2	Tape Monitor Design	56
7.3	Laboratory Tests	62
SECTION 8 - SUMMARY	67
PHASE III - FIELD TESTING	69
SECTION 9 - INTRODUCTION	71
SECTION 10 - PROBLEMS AND MODIFICATIONS	73
i

-------
Page
SECTION 11 - ODOR PANEL COMPARISON	79
SECTION 12 - SUMMARY AND CONCLUSIONS	81
SECTION 13 - REFERENCES	ci
ii

-------
ire
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
9
10
11
13
14
17
19
37
38
40
41
42
43
44
45
46
47
48
50
51
52
52
57
58
59
60
61
63
63
65
68
74
75
77
LIST OF ILLUSTRATIONS
Title
Cold Trap Collection Tubes
Complete Collection Train
Activated Charcoal Collector
Thermal Partitioning Apparatus
Partition Cold Trap
Sniffing Port on Gas Chromatograph
Press-Vent Sample, Apiezon-L Chromatogram
Front Panel Components
Reagent Addition Heads
Aldehyde-MBTH Spot Geometry
Amplifier and Automatic Zero
Interval Timer Circuit
Internal Analysis Components
Analysis Components with Mounted Syringe Pump
Aldehyde Monitor in Case
Aldehyde Analytical Cycle
Recorder Plots During Sampling and Analysis
Variation in Aldehyde Reading and Spot Diameters
Amine Analytical Cycle
Front Panel of Amine Monitor
Amine-Ninhydrin Spot Geometry
Amine Monitor - Depletion of Amine from 1000 ppm
Propylamine Aqueous Solution
Organic Sulfur Monitor - Sampling System
Organic Sulfur Monitor - Internal Analysis
Components
Analysis System Components
Organic Sulfur Monitor
Organic Sulfur Monitor
Blank Development in Sulfur Tape
Organic Sulfur Recordings
Oxidizing Furnace Blank
Aldehyde, Amine, and Organic Sulfur Monitors
Ready for Field Testing
Aldehyde Monitor - Plant Test Comparisons
Exhaust Scrubber Samplings
Amine Monitor - Plant Test Comparisons
Organic Sulfur Monitor - Plant Test Comparisons

-------
LIST OF TABLES
Table No.	Title	Page
1	Compounds Found in Rendering Emissions	4
2	Cold-Trap Collection Trains	8
3	Gas Chromatographic Columns	15
4	Retention Time Deviation	21
5	Class Analysis Methods	29
6	Evaluation of Analytical Methods	34
7	Comparison of Odor Levels and Monitor Readings	79
iv

-------
PHASE I
PRELIMINARY INVESTIGATIONS
1

-------
SECTION 1
INTRODUCTION AND SUMMARY
The ultimate goal of this development program was analytical instru-
mentation to field-monitor the major odorous compounds emitted from animal
rendering plants. The program consisted of three phases: Phase I -
Preliminary Investigation to determine practical methods for measuring
the individual odorous compounds and for monitoring the principal classes
of odorous compounds, Phase II - Prototype Development to produce field
usable instruments for class compound monitoring, and Phase III - Field
Testing of the Instruments in a rendering plant. Earlier studies had
identified aldehydes, carboxylic acids, and amines as among the most
odorous compounds emitted in the cooking, rendering, and decay of meat
and offal. The program plan, therefore, called for developing monitor-
ing instruments for these three classes of organic compounds or for a
representative member of each class.
The work in Phase I was carried out with two objectives: first,
an analysis technique to determine the major odorous compounds in render-
ing plant emissions and, second, selection of chemical analysis techniques
suitable for development for field monitors for aldehydes, organic acids,
and amines. The wide variety and low concentration of compounds present
in rendering plant emissions made gas chromatographic separation the most
practical way to analyze for the odorous constituents. Analysis was
based on cross-correlating retention times on different chromatographic
columns using a sniffing port to detect the odorous components. Table 1
lists all of the odorous compounds found in this study as well as those
reported in two other studies. The presence of a large number of organic
thiols and sulfides and their more obnoxious odors compared to the organic
acids present led to the replacement of the organic acid monitor by an
organic sulfur monitor.
The obnoxiously odorous compounds actually represent only a small
fraction of the vapor emissions from rendering plant operations. Of
these, the thiols and sulfides have the most intense and objectionable
odors even at extremely low concentrations. The normal aldehydes, acids,
and amines rank next, requiring higher concentrations, in that order, to
cause an objectionable odor. The aldehydes found highest in concentra-
tion were 2-methylbutanal, 3-methylbutanal, and 2-methylpropanal, which
probably have the least objectionable odor of the low molecular weight
aldehydes. Ethanal (acetaldehyde), methanal (formaldehyde), propanal,
and hexanal ranked intermediate in concentration but much higher in
objectionable odor. The other aldehydes, the acids, and the amines
ranked more than an order of magnitude lower in concentration along with
the thiols and sulfides. Non-odorous hydrocarbons (alkanes), faintly
odorous ethanol and methanol, and propanone (acetone), 2-butanone, and
2-hexanone made up the largest part, 50-70%, excluding water vapor, of
many of the samples.
3

-------
Table 1 - Compounds Found in Rendering Emissions
Compound Class
Compound
Found By - X
+
ft
+ + +


BRL
IITRI
ELW
Aldehydes
Methanal (Formaldehyde)
-
-
X

Ethanal (Acetaldehyde)
X
X
X

Propanal
X
X
X

n-Butanal
X
X
-

n-Pentanal
X
X
-

n-Hexanal
X
X
X

n-Heptanal
X
X
-

n-Oc tanal
X
X
-

n-Nonanal
X
X
-

2-Methylpropanal
X
X
X

2-Methylbutanal
X
X
-

3-Methylbutanal
X
-
X
Acids
Methanoic (Formic)
(X)
_
X

Kthanoic (Acetic)
(X)
X
X

Propanoic
X
X
X

Butanoic
(X)
X
-

Hexanoic
(X)
-
-

2-Methylpropanoic
(X)
-
X

3-Methylbutanoic
-
X
-
Amines
Methylamine
(X)
NR*
X

Dimethylamine
(X)
NR*
-

Ethylamine
(X)
NR*
-

Propylamine
X
NR*
-

Butylamine
X
NR*
-

Hexylamine
X
NR*
-
Thiols
Methanethiol
X
_
X

Ethanethiol
X
-
X

Propanethiol
X
-
-

Butanethiol
X
-
-

Pentanethiol
X
-
-

Nonanethiol
X
-
-

l-Propene-3-thiol (Allylmercaptan)
X
-
-
Sulfides
Hydrogen Sulfide
X
-
X

Dimethyl Sulfide
(X)
-
X

Diethyl Sulfide
X
-
-
Alcohols
Methanol
X
_
X

Ethanol
X
X
X

Propanol
X
X
-

Butanol
X
X
-

Pentanol
X
X
-

Hexanol
X
-
-

Isopropanol
X
X
-

l-Penten-3-ol
X
X
-
Ketones
Propanone (Acetone)
X
X
X

2-Butanone
X
X
X

3-Pentanone
X
-
-

2,3-Butanedione
X
X
-
Alkenes
1-Octene
X
_


1-Decene
X
-
-
Aromatics
Benzene
X
X
X

Toluene
X
X
•

p-Dichlorobenzene
X
X
-
Others
Acrylonitrile
(X)
X
_

2-Methyltetrahydrofuran
X
X
-
^Bendix Research Laboratories
^Illinois Institute of Technology Research Institute1*
t++E. L. Wick11*
*
Not Reported
(V\
v Found on only one type of chromatographic column
4

-------
The objectionable odor of rendering plant emissions, it must be
concluded, arise from a mixture of many highly odorous compounds (thiols,
sulfides, aldehydes, acids, amines, and some higher alcohols) present in
very low concentrations. Class analysis, therefore, is the best approach
to monitoring the odorous emissions.
5

-------
SECTION 2
SAMPLING AND ANALYSIS
2.1 COLLECTION TECHNIQUES
Samples were collected at four different sites in a local render-
ing plant over a period of four months. Three of the sites, two differ-
ent press vents and the exhaust line from the condenser, provided high
enough concentrations of the emission compounds to make separation and
identification practical without recourse to mass spectrometry or tedious
microchemistry. The fourth site, the exhaust fan room, gave concentra-
tions typical of the ambient air inside the plant.
Most of the samples were collected in cold-traps consisting of two
or three glass tube fingers (straight or Allihn bubble type) connected
in series and cooled with either dry ice or liquid nitrogen. Freezing
of water in the traps posed the major and essentially unsolvable problem
with low temperature cold-trap collection. Both 3A-Molecular Sieve and
calcium sulfate drying filters proved impractical for removing the water.
The molecular sieve had too low a capacity for a practical size filter,
and the calcium sulfate adsorbed the organic compounds.
Five different cold trap arrangements, listed in Table 2, gave
different collection efficiencies and collection times before freezing
shut. The initial sampling rate was always 5 Jt/min (10 cfh) but decreased
with ice formation in the traps. Complete plugging usually occurred
within two hours even starting with the cold fingers only about one-third
immersed in the coolants. The longest collection time, 2-3/4 hr, was
achieved with the triple-D arrangement starting with the cold fingers
about one-third immersed. All three traps filled completely with ice and
sample in this time without fully plugging.
Figure 1 shows the triple series traps on the metal frame used to
hold them near the sampling source at the rendering plant. Tubing con-
nections between the traps were Teflon and the sampling line was 3 mm
ID Teflon tubing with 1.5-ram wall thickness wrapped with electrical heating
tape. A 28-cm long section of stainless steel tubing (4 mm ID, 1-mm wall)
served as a probe at the end of the sample line to take samples from the
press vents. The condenser exhaust line had a petcock sample port suit-
able for direct connection to the sampling line. A separate rack held
the sampling pump with particle pre-filters and flow meters, power out-
lets and controls for the electrical heating tapes, and a thermocouple
meter and selector switch to measure sampling line temperatures. Fig-
ure 2 shows the complete sampling unit. The inlet to the sample pump
was fitted with an "X" connection so the pump could pull sample gas
through two different collectors at the same time with a gage to monitor
the partial vacuum.
7

-------
Table 2 - Cold-Trap Collection Trains
Series
Trap //
Coolant
Typo uf Cold-Finger Tube

I
Dry ice/2-methoxyethano 1*
Straight
Double




2
Liquid N2
Straight

1
Dry ice/2-methoxyethanol*
Straight
Triple-A
2
Liquid N£
Allihn bubble

3
Liquid
Allihn bubble

1
Liquid
Straight
Triple-B
2
Liquid
Centrifugal inlet

3
Liquid N.>
Allihn bubble

1
Ice/water
Straight
Triple-C
2
Liquid
Centrifugal Inlet

3
Liquid
Straight

1
Dry ice/2-methoxyethanol*
Straight
Triple-D
2
Liquid
Centrifugal inlet

3
Liquid
Straight
*
Methyl Cellosolve
Samples were also collected on activated charcoal, on 13X Molecular
Sieve, and in 1% sodium bisulfite aqueous solution. The charcoal and
Molecular Sieve were packed in stainless steel tubes, 1.2 cm ID and 30
or 50 cm long, with sample gas flows of 3 £/min with the shorter tubes
and 4 £/min with the longer tubes. Activated charcoal at ambient temper-
ature collects the least amount of water and offers the only practical
way to make sample collections over a long period of time from moisture-
saturated atmosphere. However, if the sample cannot be eluted and analyzed
immediately after collection, the packed tubes should be refrigerated at
-10°C or colder to retard chemical interactions on the charcoal surface.
The Molecular Sieve saturates with water and gives much of it up when
the sample is eluted. Figure 3 shows the activated charcoal collector
with three parallel collection tubes that was used to sample the ambient
exhaust fan room air at the rendering plant. Samples were collected for
13.5 hr at 3 £/min flow during the overnight working hours at the plant.
The bisulfite solution collects and preserves aldehydes by converting
them to the bisulfite addition compounds. These solutions were checked
for formaldehyde and pyruvic aldehyde using chromotropic acid.
8

-------

-------
Figure 2 - Complete Collection Train

-------

-------
2.2 ANALYSIS TECHNIQUES
Separation and correlation of retention times on different chroma-
tographic columns served as the principal analytical method for identi-
fying the odorous compounds in the collected samples. All the samples
collected in cold traps were kept at liquid nitrogen temperature (-196°C)
until analyzed. Samples collected on activated charcoal and the Molecu-
larlar Sieve were stored in the freezing compartment of a refrigerator
until analyzed. The large amount of water present in the samples made
the concentration of odorous compounds too small for reliable, direct
analysis. Thermal partitioning of the samples before chromatographic
separation eliminated most of the water. The partitioning apparatus,
shown in Figure 4, consisted of three glass cold traps made with 14/35
standard taper glass joints with the dimensions shown in Figure 5. The
first trap, cooled in an ice/salt water bath had a 12 cm long cold finger
to collect the water transferred during partitioning. The others had
7.5 cm long cold fingers. The second trap was cooled in a dry ice/2-
methoxyethanol (Methyl Cellosolve) bath and the third in liquid nitrogen
or liquid argon. The partitioning consisted of passing pure nitrogen
gas at 0.05 £/min through the sample trap, initially at liquid nitrogen
temperature, then through the three series cold traps. The sample trap,
suspended from a ring stand in the ambient air, gradually warmed to room
temperature with the nitrogen flowing. Finally, immersing the trap in
successive beakers of hot water (90-95°C) forced out the less volatile
compounds with a minimum amount of water vapor being transferred into
the first partition trap. The partitioning was continued for two to
three hours depending on the sample size.
After partitioning was complete, the traps were separated, immedi-
ately stoppered by forcing lightly greased (Apiezon 101) serum bottle
stoppers (6 mm top diameter) into the hemispherical joint openings,
then stored in dry ice pellets or chips. Most of the rendering plant
samples contained only enough organic material to give a thin liquid-
film condensate on the partition trap walls. This was transferred to
the gas chromatograph by inserting a 5 cm (2 in.) long syringe needle
on a 30 ml syringe through the septum portion of the serum stopper, then
drawing out the vapor sample as the partition trap was warmed over a hot
plate or in hot water. Vapor samples were injected immediately into the
gas chromatograph. For the few samples that gave enough liquid conden-
sate in the partition trap for direct liquid sampling, the traps were
separated at the standard taper joint and sealed with a large size
(16 mm top diameter) serum bottle stopper. Samples were then transferred
with a 10 syringe.
Separations were made on four different gas chromatographic columns
under the conditions listed in Table 3. The injection port and the out-
let lines were kept around 250°C. The effluent from the column was split
with half going to a flame ionization detector (FID) and the other half
out through a sniffing port. The stainless steel tubing leading to the
sniffing port was heated to around 200°C and carried the gases up through
12

-------
Figure 4 - Thermal Partitioning Apparatus

-------
5mmlD
7.5cm
or
12cm
14/35 Ts

*
Q.
Figure 5 - Partition Cold Trap
14

-------
Table 3 - Gas Chromatographic Columns
Column
Number
Column
Length, m*
Stationary
Phase
Support
Temperature
Program, °C
Heating Rate,
°C/min
Carrier
Gas
Carrier
Flow, ml/min
5
5
Carbowax 2QM
(20%)
Chromosorb P-AW
(80—100 mesh)
60-250
4
He
25
9
6.6
Apiezon-L
(10%)
Chromosorb P-AW
(80-100 mesh) DMCS
60-250
3
He
34
12
3
DEGA + H3PO4
(25% + 2%)
Chromosorb W-AW
(80-100 mesh) DMCS
100-220
4
He
25
13
3
Pennwalt 223
(28%)
Gas Chrom R
(80—100 mesh)
75-180
3
He
34
Column tubing - stainless steel (316), 2mm ID, 3mm (1/8 in.) OD.

-------
the center of a Teflon cylinder 5 cm in diameter and 5 cm high. An
annular space between the stainless steel tube and the surrounding
Teflon cylinder let cool, humidified air mix with the hot effluent
gases in a shallow cup-shaped depression in the top of the cylinder.
Figure 6 shows the sniffing port on the Bendix 2500 Gas Chromatograph
used for all the chromatographic separations. Odor notations were made
directly on the chromatogram as the fractions eluted.
Samples collected on activated charcoal and Molecular Sieve were
partitioned by eluting the vapors into the thermal partitioner. Pure
nitrogen gas was passed through the collector tube at 0.05 £/min and
into the partition traps while the collector tube was slowly heated
in a tubular furnace to 350°C. After partitioning, these samples were
stored and chromatographed as described above for the samples collected
in cold traps.
The difference in molecular properties of the different classes
of compounds found in rendering emissions makes complete separation of
all the compounds on one gas chromatographic column impossible. This
means many of the chromatographic fractions from a given column will
contain two or possibly more compounds. Collection and subsequent sepa-
ration of individual fractions on a different column is not practical
for the analysis of many samples because the volume of each fraction
is so small and the technique too time consuming. Figure 7, for example,
shows the chromatogram from an Apiezon-L column separation of 13 ml of
sample vapor from the third, liquid argon cooled, trap of the thermal
partitioner. In this sample, two fractions make up over 50% of the
volume while the remaining 5-6 ml are distributed among 43 other fractions.
Different samples analyzed on different columns, however, provide
a simple, more rapid method of identifying fractions by cross-correlating
retention times for specific compounds. If the different columns vary
sufficiently in stationary phase polarity, the retention times for dif-
ferent classes of compounds will vary greatly between them, Zarazir,
Chovin, and Guiochon1^ recommend three different columns; SE30 (a non-
polar silicone), Carbowax 20M (polyethyleneglycol), and PGS (polyethylene
glycol succinate). Four different columns were used in this study;
Apiezon-L (non-polar hydrocarbon); Carbowax 20M; DEGA + H3PO4 (diethyle-
neglycol adipate polymer); and Pennwalt 223 + NaOH (a Pennwalt Corp.
material). Apiezon-L and DEGA + H3PO4 would not resolve the amines
(excluding them from these chromatograms), and the Apiezon-L and Pennwalt
223 + NaOH would not resolve the organic acids (excluding them from their
chromatograms). Acids and amines, therefore, could be cross-correlated
on only two columns. Additional correlations would have required using
more than four different columns or forming derivatives of these compounds
in the samples. Time limitations on the program, however, would not per-
mit this extra development work.
The columns were characterized with homologous series of n-alkanes,
n-alkenes, n-aldehydes, branched aldehydes, n-acids, n-amines, n-alcohols,
n-thiols, n-sulfides, and 2-ketones.
16

-------
Figure 6 - Sniffing Port on Gas Chromatograph

-------
PAGE NOT
AVAILABLE
DIGITALLY

-------
Retention times were also measured for several iso-acids, iso-
alcohols, dienes, allylsulfides and thiols, disulfides, aromatics, and
specific compounds such as 2-methyltetrahydrofuran, t-2-hexenal, acrylo-
nitrile, l-penten-3-ol, etc. Standard deviations for repeated retention
time measurements made during the course of the program are listed in
Table 4 for the principal compound classes. The greater deviations for
both acids and amines reflect the difficulty of separating these com-
pounds directly on a gas chromatographic column unless it is specifically
prepared for them to the detriment of other separations.
Table 4 - Retention Time Deviation
Compound Class	Average Standard Deviation) min
Acids
1.2
Alcohols
0.8
Aldehydes
0.5
Amines
2.8
Thiols
0.6
Because of the large number of samples analyzed, the complexity of
the chromatograms, and the wide variation in concentration among differ-
ent samples, direct comparison of chromatograms from a given column was
not practical. To facilitate cross-correlation, therefore, retention
times of odorous fractions were plotted on a vertical time scale on
large graph paper with different sample separations placed side by side
along the horizontal axis. Matching retention times for all the samples
separated on a given column were then noted and possible compounds cor-
responding to these retention times were listed. Comparison of these
lists of compounds for all the samples analyzed on the four different
columns along with the odor notations on individual chromatograms showed
which compounds were common to two or more of the columns and which
fractions from a given column contained several compounds. Table 1 lists
the compounds found by this cross-correlation of 72 gas chromatographic
separations. With the exception of the acids and amines, which could
only be separated on two different pairs of columns, all the compounds
listed were found common to either three or all four of the different
column separations.
Odor notations for the chromatographic fractions confirmed many
of the correlations between different columns but could not furnish
positive identification because of overlapping fractions and persistence
of thiol and sulfide compound odors. Interference from these intensely
odorous sulfur compounds was not anticipated. The report from IITRI^
made no mention of specific sulfur compounds, giving only a total sulfur
by flame emission analysis; E. L. Wick reported only hydrogen sulfide,
21

-------
methanethiol, and ethanethiol, all highly volatile compounds. Both
odor and retention time correlations confirmed the presence of metha-
nethiol through pentanethiol as well as allylmercaptan, allyl methyl
sulfide, and diethyl sulfide in the samples analyzed for this program.
Although they were not considered part of this program as initially
planned, the extremely intense and disagreeable odor of these compounds
forced consideration of their measurement in any attempt to monitor
rendering plant emissions.
Concentrations of the odorous compounds varied widely from sample
to sample, even from the same sampling site in the rendering plant. How-
ever, they fit roughly into three concentration groups. A few compounds
in different combinations made up 50-70%, exclusive of water, of many of
the samples. The compounds most often found in these predominant groups
were:
2-methylbutanal
3-methylbutanal
2-methylpropanal
pentane
propanone
ethanal
ethanol
2-butanone
in that general order. A few other compounds were predominant in one
or two chromatographic separations. These included the aldehydes n—
propanal, butanal, pentanal, and hexanal; the normal alcohols methanol,
propanol, and butanol; 2-hexanone; and 1-pentene.
A larger number of compounds fell into an intermediate concentra-
tion range, making up from 10-50% of the sample - again exclusive of
water. Compounds in this group included
n-aldehydes: methanal through nonanal
amines: dimethyl and trimethyl
acids: 2-methylpropionic (isobutyric)
thiols: methane-, ethane-, allyl-
n-alcohols: methanol through hexanol
sulfides: hydrogen sulfide and diethyl sulfide
ketones: 2-pentanone
aromatics: toluene
alkenes: 1-octene
22

-------
The remainder of the sample was made up of a large number of compounds
present in very small concentrations, many of which were not identified.
Based on the samples analyzed, therefore, the characteristic odor from
rendering operations is not constant and results primarily from the
interaction of a large number of different compounds present in moderate
or low concentrations.
23

-------
SECTION 3
ANALYTICAL METHODS FOR FIELD MONITORING
3.1	REQUIREMENTS
Analytical methods suitable for field monitoring instruments must
be adaptable to simple mechanical operations and must have as well the
requisite sensitivity, accuracy, specificity, and chemical stability.
For this reason, sophisticated instrumental methods and complex chemical
procedures were not considered. In particular, methods were sought that
involve direct colorimetric, spectrophotometry, reflectometric, fluoro-
metric, or electroconductimetric principles that could be used in a
simple tape-monitoring type instrument. Fortunately, the most sensitive
chemical analytical methods with measurable concentration limits in the
parts-per-billion (ppb) or parts-per-million (ppm) range are fluorometric
or colorimetric. The only major disadvantages to these types of methods
for automatic field monitoring applications are the complex manual manipu-
lations (filtrations, extractions, etc.) required for many of the proce-
dures, the instability or incompatibility of some of the reagents under
field conditions or on paper tapes, and interferences from related
compounds.
The methods screened included the standard procedures as published
in ASTM,1 AOAC,2 Welcher,13 Siggia,10 and similar volumes; photometric
methods covered in Snell,11 Sawicki,9 and related books; methods specific
for gaseous and air pollutants reported in books such as Stern,12 Ruch,7
Leithe,6 and the recent Intersociety Committee5 volume; as well as
methods reported in cumulative and recent yearly indexes of Chemical
Abstracts. The initial screening led to more than 60 possible methods
for aldehydes, only 14 for organic acids, and 15 for amines. Review of
original papers and simple laboratory tests reduced this list to the
five aldehyde methods and three acid and three amine methods discussed
in the following section.
3.2	SELECTED ANALYTICAL METHODS
The five methods selected as best suited for development for a field
monitor instrument for aldehydes are
(1)	Reduction of silver ion to silver metal (Tollen's method)
(2)	Reduction of Nessler's reagent to mercury
(3)	MBTH colorimetric
(4)	Bisulfite-hydrazine colorimetric
(5)	Aminobenzaldehyde and methylamine hydrochloride colorimetric
25

-------
The first two have the advantage of a simple direct reaction on an
impregnated paper tape which essentially parallels the widely used lead
acetate tape procedure for hydrogen sulfide. The transmittance or re-
flectance of the metal deposit on the tape could be measured with the
photometric equipment on conventional tape sampler-monitors. Laboratory
experiments using tapes impregnated with ammoniacal silver nitrate
treated with aqueous acetaldehyde and acetaldehyde bisulfite solutions
showed that the reactions proceeded on the tape in a few minutes time
with gentle heating. The less-than-ppm sensitivity claimed for the
Tollen's reagent was not achieved in this test; however, optimum condi-
tions were not established in these tests and more sensitive modifica-
tions of the reagent were not tried. The major drawback for these
metal ion reduction methods is the interference by other reducing agents,
including any present in the paper tape itself. Sulfur compounds could
possibly be removed with neutral or slightly acid silver nitrate or
mercuric acetate which will not oxidize aldehydes. The ease of adapt-
ing these methods to tape sampler monitoring made them worth investi-
gating despite their drawbacks.
The MBTH (3-methyl-2-benzothiazolone hydrazone hydrochloride)
method has the advantages of extreme sensitivity, good specificity
(especially for the water soluble, lower molecular weight aldehydes),
simplicity, and widespread application for aldehyde determinations.
Aldehydes react with MBTH so the reagent can collect the sample directly
with only the need for subsequent addition of an oxidizing agent (ferric
chloride in sulfamic acid solution) to develop the color. The major
drawback to this method is the slowness of the reaction between the
MBTH and aldehydes. After sample collection, the MBTH-aldehyde mixture
is aged one hour to assure complete reaction. This is not a serious
disadvantage for long time sampling but could interfere if short time
(less than one hour) sampling is desired. Otherwise this is the pre-
ferred colorimetric method.
The bisulfite-hydrazine method also has the advantages of high
sensitivity and good selectivity (again for water soluble aldehydes)
and permits direct collection of the aldehydes by sodium bisulfite.
2,4-Dinitrophenyl-hydrazine, the most widely used phenylhydrazine for
aldehyde determination, is reportedly stable indefinitely in phospheric
acid - ethanol solution. For maximum sensitivity, however, the proce-
dure is more complex than that with MBTH, requiring reagent addition
after collection, heating for 30 min, then reaction with alcoholic KOH
to develop the color. Its proved reliability in aldehyde determinations,
however, made it a good back-up choice for the MBTH method.
A third possible colorimetric method which has the advantage of
being specific for aliphatic aldehydes is the reaction with o-aminobenz-
aldehyde and methylamine. The amine condenses onto the aldehyde group
of the o-aminobenzaldehyde and the collected aldehyde onto the amine
group. Aliphatic aldehydes can then close a second ring forming a
26

-------
1-hydro-quinazolium compound. When methylamine hydrochloride is used
(because of its stability), the reagents must be added sequentially to
the collected aldehyde. First the methylamine must be freed from its
hydrochloride by adding sodium pyrophosphate; then the o-aminobenzaldehyde
is added. The color develops in about 10 min at 25°C. The specificity
to aliphatic aldehydes and more rapid reaction rate at room temperature
give this method an advantage over the MBTH and bisulfitehydrazine
methods. Poorer reagent stability, more mechanical steps, and poorer
sensitivity, however, make it a third choice.
Fluorometric methods were considered for aldehydes but unfortunately
they required either heating in concentrated sulfuric acid or manipula-
tions too complicated to adapt for a simple field monitor.
Of the few available methods for the quantitative measurement of
organic acids, only three colorimetric ones offered much hope for adapta-
tion to an automatic field monitor. These were
(1)	pH color or fluorescent indicator
(2)	Phosphomolybdic acid colorimetric
(3)	Ferric-5-nitrosalicylate decoloriation
The pH color or fluorescent indicators are the oldest analytical compounds
used for acid determinations either titrimetrically or photometrically.
They are very sensitive over a narrow acid concentration range which
makes them suitable for monitoring low concentrations of the weak organic
carboxylic acids. They are not selective, however, and will react with
all types of acids as well as bases. A prefilter (a non-volatile acid
on a solid support) would be necessary to eliminate interference from
the basic amines present in rendering emissions. Volatile inorganic
acids and acid gases are not ordinarily encountered in rendering plant
operations, so these would not cause a problem except in monitoring at
remote distances from the plant exhaust. Laboratory spot tests on paper
tape showed that both methyl orange and bromophenol blue, sensitive in
the pH 3.0-4.6 range, could measure acetic and propionic acids down to
10 ppm just with visual observation. Sodium fluorescein gave much
brighter spots under UV light, but had correspondingly brighter blanks.
The acid quenched the fluorescence so that lower concentrations were
more difficult to distinguish by eye. Rhodamine 6G, an ester form of
the rhodamine dye, is reportedly much more sensitive to organic acids,
measuring as little as 0.6 yg of stearic acid. These color or fluorescent
indicators, therefore, offered the best possibility for adapting to a
tape monitor system.
Phosphomolybdic acid in ethanol is commonly used to develop fatty
acid spots on thin layer chromatographic plates. It should work equally
well on paper tape since, if necessary, silica gel impregnated tape can
be used. It is sensitive to 1 ug or less of saturated fatty acids but
the color formed is not very intense and its quantitative characteristics
must be investigated.
27

-------
A third method is the bleaching of the color of ferric-5-nitrosali-
cylate by water soluble organic acids. The reaction is independent of
pH in the range 2.5-3.0, and shows no temperature effect between 10°
and 50°C. Moreover, the reagent is stable for over a week. The method,
however, has only been tested with dicarboxylic acids; the effect of
reducing compounds, such as aldehydes and thiols, was not reported.
Mineral acids do not cause bleaching, so this method is potentially spe-
cific for organic acids if it is sensitive enough to monocarboxylic acids.
From the methods available for determining amines, two long-standing
well-proven methods and a more recently developed one offered the sensi-
tivity and simplicity needed for a field monitor. These were:
(1)	NBD (4-chloro-7-nitrobenzofurazan) fluorescence
(2)	Dansyl chloride fluorescence
(3)	Ninhydrin colorimetric
The NBD chloride method was developed four years ago to determine
amino acids and peptides. The compound itself is not fluorescent but
its amine derivatives are:
The compound is quite stable which gives it an advantage over dansyl-
chloride. It does not have the background of widespread use that dansyl
chloride and ninhydrin have, however, so more work would be necessary
to establish the best reaction conditions and eliminate interferences.
Dansyl chloride (l-dimethylamino-naphthalene-5-sulfonyl chloride)
is a standard reagent for amines, amino acids, and other amino organic
compounds. It has a blue-green fluorescence itself but the amine deriva-
tives fluoresce yellow-green, permitting a clear differentiation although
the blank readings are high. Measurement of as little as 50-100 ng of
amine compounds have been claimed with the dansyl reaction. It has been
used directly on paper chromatograms for the quantitative measurement
of separated amino compounds, so it should be directly applicable to
tape sampler monitoring. Poor stability of the reagent is its greatest
drawback.
28

-------
Ninhydrin is the classic reagent for the colorimetric determina-
tion of amino acids and is commonly used in commercial automatic amino
acid analyzers. It is extremely sensitive and is used to develop amino
acid spots on paper chromatograms, giving a quantitative measure of the
amino acid concentration. If the proper reducing conditions can be
achieved to make it react on paper tape with the simple amines, it would
be the best colorimetric method for the amines.
The recommended methods for field monitor development are summar-
ized in Table 5.
Compound Class
Aldehydes:
Acid:
Amine:
Table 5 - Class Analysis Methods
Recommended Analytical Methods
Reduction of silver ion to silver
Reduction of Nessler's reagent to mercury
MBTH colorimetric
Bisulfite-hydrazine colorimetric
(1)
(2)
(3)
(4)
(5)
o-Aminobenzaldehyde and methylamine hydrochloride
colorimetric
(1)	pH color or fluorescent indicator with amine
prefliter
(2)	Phosphomolybdic acid colorimetric
(3)	Ferric-5-nitrosalicylate - decolorizing
(1)	NBD (4-chloro-7-nitrobenzofurazan) fluorescence
(2)	Dansyl chloride fluorescence
(3)	Ninhydrin colorimetric
29

-------
PHASE II
PROTOTYPE DEVELOPMENTS
31

-------
SECTION 4
INTRODUCTION
The prototype development in Phase II was based on modification of
standard commercial tape samplers. The ambient air hydrogen sulfide
monitor served as the basic instrument for modification to measure each
of the three types of odorous compounds found in rendering plant emis-
sions. Originally the program plan called for developing an aldehyde,
amine, and an organic acid monitor. These were reportedly the three
most odorous types of compounds resulting from meat rendering. The
analytical work in Phase I of the project, however, disclosed that more
organic sulfur compounds than organic acids appeared in the rendering
plant emissions. This, coupled with the much more repulsive nature of
the odor of volatile thiols and sulfides, led to the decision to replace
the organic acid monitor with an organic sulfur monitor.
Laboratory tests with different Whatman papers available in tape
form led to the selection of Whatman No. 1 paper for the aldehyde and
amine monitors and No. 54SFC for the organic sulfur monitor. Of the
papers tested, numbers 3MM, 40, 120, and 541 had very high aldehyde
blanks (greater than 10,000 ppm aldehyde), while numbers 1, 20, and 540
had acceptable but still moderately high blanks. Amine blanks for these
papers were all low with number 1 the lowest, making it the best choice
for both aldehydes and amines. While number 54SFC gave the best reagent
blank for sulfur, all the papers developed a very high blank for the
sulfur determination when they were stored for a few hours at room temp-
erature after being impregnated with reagent.
Spot tests on the paper tapes with the different reagents selected
for each compound class eliminated the methods least suited to tape
monitoring analysis. Table 6 summarizes the results. The sample solu-
tions for these tests ranged from 0.01 to 1000 ppm in concentration.
The results clearly indicated the superiority of MBTH for aldehydes and
ninhydrin for amines, and the greater practicality of determining the
sulfur compounds as SO2.
Selection of these analytical methods meant that the basic tape
samplers for aldehydes and amines required the addition of reagent heads,
blank (background) reading heads, reagent pumps, and interval timers for
the automatic analysis. The organic sulfur monitor, in which the sulfur
compounds are first thermally oxidized to SO2 which then reacts with
reagent added to the tape, required a reagent addition system, oxidizer
furnace, and a bubbler to provide water vapor for the reaction. In
their final design, all three instruments performed repetitive sampling
and analysis automatically with continuous strip chart recording readout.
33

-------
Table 6 - Evaluation of Analytical Methods
Compound Class
Aldehydes
Amine
Sulfur
Method*
Silver ion
Nesslers
MBTH
Bisulfite
Benzaldehyde
NBD
Dansyl
Ninhydrin
h2s
SO,
Tape Spot Test Results
Reagent unstable on tape
Reagent unstable on tape
Moderate blank, stable,
sens.it ive
Not sensitive enough
Unstable
Unstable
High blank on tape
Low blank, high sensitivity
Requires complete reduction
of air sample
Requires oxidation of organic
compounds
See Table 5
34

-------
SECTION 5
ALDEHYDE MONITOR
5.1 ANALYTICAL METHOD
In the aldehyde monitor, 3-methyl-2-benzothiazolone hydrazone (MBTH)
in the paper tape selectively collected the aldehydes from the sample air.
Addition of acid ferric chloride solution developed a characteristic
blue color related in intensity to the concentration of aldehyde. The
reaction is well established for analytical use and is the method selected
by the Intersociety Committee for determining formaldehyde.5 The aldehyde
reacts with MBTH by a condensation reaction:
The ferric chloride oxidizes excess reagent which couples with the alde-
hyde product giving the dark blue compound.
Formaldehyde produces the most intense color with the MBTH method,
but other volatile aldehydes also react producing the same blue color.
The method, therefore, can serve for class analysis of volatile aldehydes.
Heating speeds the color development on the tape, 100°C giving
maximum color in 2 min, but also increases the blank. A high blank re-
duces the sensitivity of the color intensity measurement by excluding
too much light from the measuring photocells, so the monitor procedure
was based on room temperature color development.
The impregnating solution for the paper tape consisted of l.Og
S-methyl-Z-benzothiazolonehydrazone'HCi, dissolved in 1000 ml distilled
water and filtered through No. 40 Whatman filter paper. The color-
developing acid ferric chloride solution contained 1.6g sulfamic acid
(analytical grade) and l.Og FeCl3 dissolved in 100 ml distilled water
and filtered through No. 40 Whatman filter paper.
CH
3
CH
3
35

-------
5.2 TAPE MONITOR DESIGN
The additional reading and reagent addition heads fit in-line on
the front panel of the basic tape sampler after moving the tape storage
canister, sampling head, and the attendant tape guide spindles to the
left. Figure 8 outlines the position of these components on the front
of the instrument. The panel width of this particular sampler permitted
matching the spacing between the heads to the tape advance distance.
This eliminated any need to modify the tape advance system in the
instrument.
The first design for the aldehyde monitor required only addition
of the extra heads, a reagent addition metering pump, and rewiring of
the timer circuit. It relied on metered addition of reagent from a
center feed reagent head [Figure 9(a)]. Capillary action would draw
the acid ferric chloride rapidly onto the tape as soon as the solution
column touched the bottom of the tape. This would drain all the solution
from the center feed plug and horizontal connecting channel in the head.
A peristalic pump, triggered by the sample timer, would refill the rea-
gent head channels with solution during the first tape advance. Attempts
to synchronize the filling rate with the tape transport proved futile,
however, and this simple design had to be replaced with one using a
separate interval timer to control the analytical steps.
In the second design, the additional interval timer, allowed full
initial tape advance before starting the reagent addition pump. It
also permitted individual adjustment of both pump operating time and
total reagent addition interval time before the second tape advance.
This eliminated the impractical synchronization requirement of the first
design but confirmed two other problems in reagent addition. The first
was the radial transport of developed aldehyde color by the flow of
reagent from the center feed point. A colored ring formed, instead of
a color spot, leaving a largely clear center area. The high light trans-
mission of the colored ring reduced the measurement sensitivity. More-
over, even slight spreading of the ring outside the light area of the
reading head caused a large error in the opacity reading. The second
problem was variation in the amount of reagent drawn from the reagent
head by the capillary action in the paper tape. The tape would some-
times pull reagent from the connecting tubing to the head, so that the
amount of reagent added by the pump during the next analytical cycle
was not enough to fill the head and reach the tape.
A small cup reservoir in the center of the reagent head, Figure 9(b)
did not eliminate the variations in reagent addition. Capillary action
in the centerfeed channel still carried reagent from the connecting tubing
to the cup as reagent flowed onto the tape. Changing to an annular reser-
voir, Figure 9(c), improved the spot formation by introducing bi-
directional reagent flow in the tape. The reagent, however, usually
wet the tape first just above the feed channel causing an off-center
spot. Chamferring the inner edge of the annular reservoir [Figure 9(d)]
36

-------
FLOW
REGULATOR
STOCK
SPOOL
SAMPLING
TIMER
FLOW
METER
PRESSURE
BLOCKS
RECORDER
SAMPLING
HEAD
REAGENT
HEAD
(SI
MEASURING
HEAD
(S)
REAGENT
HEAD
(R)
MEASURING
HEAD
(R)
MANUAL TAPE
ADVANCE
u>
"vl
Figure 8 - Front Panel Components

-------
Figure 9 - Reagent Addition Heads

-------
reduced this tendency somewhat but did not eliminate it. Attempts to
fill the annular reservoir through two feed channels were not successful
with either a completely beveled [Figure 9(e)] or chamferred [Figure 9(f)]
inner edge. The tape would usually wet first above one or the other
feed channel leading to an off-center spot. Packing the reservoir with
fine glass wool gave uniform wetting around the reservoir as long as the
glass wool stayed wet with reagent. When the reagent dried in the reser-
voir during sampling, however, the next reagent addition was always
off-center.
Adding the reagent as a drop from above spread the spot into a ring
extending beyond the transmission area of the reading head. About one-
third to one-half a drop (0.02 - 0.03 ml) gave the best spot size. For
the final design, therefore, it was necessary to return to the center
feed head. Introducing a Teflon center plug with a funnel-shaped cup
[Figure 9(g)] decreased the capillary action in the feed channel making
the volume of reagent added more uniform. Occasional variations in flow
volume still occurred leading to a missing spot in the next analysis
cycle. Placing a Teflon "washer" into a cup-shaped reservoir [Figure 9(h)]
reduced the possibility of missing additions still further by providing
three or four small pulse additions of reagent as the pump added reagent
to the cup. This type of reagent head was used in the final prototype
design.
Individual pumping to each reagent head proved necessary to elimi-
nate preferential flow into the head which first wet the tape. A peri-
staltic pump with two feed tubes was satisfactory except for the periodic
low volume addition when a pump roller lifted from the tubing. A syringe
pump (Sage Model 341) modified to hold two 10 ml syringes eliminated this
problem. Plastic, disposable syringes proved unsatisfactory, however,
because their sliding rubber seals showed stick-slip action along the
barrel. This caused delayed addition of reagent with the tape wetting,
at times, during or after its second advance. Glass syringes gave smooth
uniform addition of reagent to the head. Thorough greasing with a hydro-
carbon or silicone grease prevented "freezing" of the syringe through
drying of reagent solution between the barrel and plunger.
The ring nature of the developed aldehyde spot, illustrated in
Figure 10, led to two additional changes in the reading heads. Block-
ing the center portion of the opening to the photocell chamber below
the tape eliminated a large part of the excess light transmitted by the
colorless, eluted center area of the spot. The circular area blocked
was 14 mm in diameter, leaving a 4 mm wide annular reading zone for
light from the upper chamber to pass through the tape and into the
lower photocell chamber. The photoconductive cell was then replaced
with one of larger diameter located at the bottom of the chamber to
intercept more of the light from the annular reading zone.
The electronic circuit of the tape sampler was not modified. The
sample photoconductor cell controlled the inverse input to an opera-
tional amplifier as shown in Figure 11 so the output reading increased
39

-------
ELUTED
ZONE
V
2 CM
SAMPLE
SPOT
I	I
READING
ZONE
CN
Figure 10 - Aldehyde-MBTH Spot Geometry
with increasing opacity of the sample spot* The electrical wiring was
changed to accommodate the interval timer as shown in Figure 12 along
with the syringe pump and extra reading heads. Detailed wiring diagrams
appear in the Operating Manuals provided with the instruments.
The final design for the front panel of the monitor is shown in
Figure 8. The reagent heads were polycarbonate plastic with blocks
of the same material pressing the tape against their delivery surface.
The sampling head collected a 1.5 cm diameter spot while the reading
heads measured a 2.2 cm diameter area. Figure 13 shows the position
of the analysis components inside the monitor. The photograph in Fig-
ure 14 shows these components in place with the syringe pump mounted
on top of the sampling pump. Figure 15 shows the monitor in its case
but with the front cover removed.
5.3 LABORATORY TESTS
Both ethanal (acetaldehyde) and methanal (formaldehyde) were used
in preliminary tests with MBTH-impregnated tape. Once spot formation
and color development was established, aqueous methanal solutions proved
most convenient for evaluating design modifications in the monitor.
Solutions of less than 1% methanal in water have the same vapor and
liquid composition,7 so that parts per million (ppm) solutions in a
bubbler provide a convenient source of methanal at fixed concentration.
Sampling times of 3 to 4 min were adequate for repetitive tests with a
350 ppm methanal solution in water. A standard 25 ml midget impinger
containing 10-15 ml of the solution was connected to the input of the
monitor sampling head with a short length of polyvinyl chloride tubing
(Tygon tubing). With the monitor flow control set to draw room air
through the bubbler and tape at 0.5 or 1 fc/min (1 or 2 CFH), an increase
in spot color intensity was clearly visible by eye after 3 min sampling.
The room air alone did not produce a visible change in spot color.
40

-------
REFERENCE
PHOTOCONDUCTOR
SAMPLE
PHOTOCONDUCTOR
+ 15V
O -15V
Amplifier and Automatic Zero

-------
115V
AC
(AC)
O ^ TAPE
I 4f | ADVANCE
' 2 J Relay
%
o
o
o
/
pNO
'
Q NC
TT
SAMPLING
PUMP
	Q	
POWER
@ SWITCHING
^ BOARD

C
0
NO
o
NC

c
o
NO
o
NC
TT
SYRINGE
PUMP
_Q_
c
)c

c
J C

o
O N0

O NO
[ MOTOR

c
} NC

c
^ NC

o
©
A/Z CONTROL
BOARD
Figure 12 - Interval Timer Circuit

-------
o
kriiiiiifPf.i'K'iVil
s\
/ N
1 1
V }


/ \
\ 1
SAMPLING
TIMER
w
SYRINGE
PUMP
INTERVAL
TIMER
FRONT
PANEL
I
Figure 13 - Internal Analysis Components
.o
u>

-------
Figure 14 - Analysis Components with Mounted Syringe Pump

-------

Figure 15 - Aldehyde Monitor in Case (front cover removed)


-------
The analytical cycle shown in Figure 16 was used in operational
tests of the monitor. The sampling timer triggered the initial tape
advance at the end of the sampling period. This activated the interval
timer which automatically controlled the analytical cycle. The sampling
pump turned off immediately as the tape heads opened for the first tape
advance to the reagent addition position. The tape spot that was in the
first (sample) reading head moved to the right hand reagent head to
become the blank spot for the reference head. For most tests, the rea-
gent syringe pump stayed on 9 s with a delivery speed of 0.19 ml/min.
This would transfer 0.03 ml of acid ferric chloride solution to the
reagent head and tape. The tape then advanced to the reading position
where it stayed for 87 s for recording of the color development. The
final tape advance brought fresh tape into the sample and reading heads;
at the same time, the interval timer activated the automatic zero circuit.
The recorder plotted continuously giving a record of blank tape
opacity during the sampling period and the difference in blank and sample
spot color development after reagent addition. The tape transport and
reagent addition steps, being of short duration compared to sampling and
reading, gave scattered points on the recorder chart. Sample recordings
are illustrated in Figure 17. The straight, continuous vertical lines
represent the sampling period preceded by the automatic zero adjust.
The horizontal dotted lines, converging toward a continuum, show the spot
color development. If the sample and reference spots developed the same
geometrically, the recorded points converged toward a higher opacity
reading as the sample spot became darker than the reference spot. A
slight difference in the development of the spots caused the opacity
readings to converge toward a lower reading as occurred in the last run
SAMPLING PUMP
TAPE ADVANCE
REAGENT ADDITION
READOUT
AUTO ZERO
ON --
OFF
I r
mm ^
Ml







:
1




E-





0 11 31 42
ELAPSED TIME - SECONDS
129 140 S
Figure 16 - Aldehyde Analytical Cycle
46

-------
TIME
MIN.
30-
40-
50-
60-
70-

-T"
10
i
20
RECORDER READING
-	SAMPLING PERIOD
-	TAPE TRANSPORT, REAGENT ADDITION
	<
AUTO ZERO
COLOR DEVELOPMENT READOUT
TAPE TRANSPORT
ANALYTICAL SEQUENCE
Figure 17 - Recorder Plots During Sampling and Analysis
shown in Figure 17. A large difference in spot diameter or elution pat-
tern caused erratic or off-scale opacity readings.
The experimental work with different reagent addition heads led
to a design that produced reasonably uniform spot diameters. It was
then possible to compare recorder opacity readings with reference and
sample spot diameters. Test results, such as those plotted in Figure 18
showed that with differences of 0.5 cm or less in sample and reference
spot sizes, the spot size did not control the recorder reading. As the
individual spot diameters plotted at the top show, none of the spots
exceeded the 2.2-cm diameter reading zone. The pattern of the recorder
readings, however, showed no correlation with either the sample or refer-
ence spot diameters nor with their difference. The only other variable
in the color spots was the elution pattern which formed as the reagent
flowed onto the tape from the center feed in the head.
Blocking the light passing through the center, mostly colorless
portion of the spots so that only a 4 mm wide annular zone affected the
light transmission proved to be an effective, short-term solution to
the extreme variations in opacity readings. With this center stop in
place, none of the readings in 37 consecutive runs exceeded 26. While
these annular zone readings were more sensitive to spot diameter, they
eliminated the very high opacity readings (for low aldehyde concentra-
tion) caused by the greater transmission of the center, colorless part
of the reference spot compared to that of the sample spot. This made
47

-------
2.2-
2.0-
DIAMETER
CM
1.8-
1.6-

A
/
\
/
/ c
\


u
u
1 /




\.
i
i
i\
*
t/
Yt
\

7


\ 1
\l







\
\
i
/
/
f



A
/ *
U
A
>
TAPE SPOT DIAMETERS
CV— SAMPLE SPOT
REFERENCE SPOT
A-DIAMETER
CM
SPOT DIAMETER DIFFERENCES
SAMPLE - REFERENCE
RECORDER
READING
I 2 3 4 5 6 7 8 9 10 II 12 .13
RUN NUMBER
RECORDER READING
35 PPM HCHO-AIR
SAMPLED 3 MIN. AT 0.5 8 /MIN.
Figure 18 - Variation in Aldehyde Reading and Spot Diameters
it possible, then, to record the presence of high aldehyde concentra-
tions in a sequence of runs without serious interference from wayward
readings. Measurements at a rendering plant, compared to laboratory
runs discussed in Section 6, confirmed this.
48

-------
SECTION 6
AMINE MONITOR
6.1	ANALYTICAL METHOD
The classic ninhydrin colorimetric method for amino acids proved
best in preliminary experiments for amine analysis on tape strips.
Reagent stability and color development temperature posed the two major
problems of immediate concern in adapting the method for a tape monitor.
Impregnating the tape with only the sodium acetate/acetic acid buffer
gave a very stable tape with enough acidity (pH 5.5) to collect amine
compounds. The ninhydrin/stannous chloride color-developing solution
made in 2-methoxyethanol (Methyl Cellosolve) was stable for over three
days at room temperature and under ambient light.
The solutions finally used in the tape monitor method consisted
of a tape-impregnating buffer containing 270g of sodium acetate trihydrate
and about 50 ml of glacial acetic acid in 500 ml of distilled water
(final pH adjusted to 5.5 + 0.1), and a color reagent of 2.0g ninhydrin
and 0.5g stannous chloride in 100 ml of 2-methoxyethanol (Methyl Cello-
solve) . This combination gives a very sensitive measure of amine com-
pounds as a class. Color development is slow at room temperature, but
experiments with heating the tape to speed the reaction showed an unde-
sirable increase in the blank color above 40°C. To avoid the increased
blank and the undesirable complexities of controlled heating for the
tape, the final procedure was based on ambient temperature color develop-
ment with a four-minute color development period.
6.2	TAPE MONITOR DESIGN
The amine tape monitor required the same analytical steps as the
aldehyde monitor, so its physical design was essentially the same. The
solvent-penetrating action of 2-methoxyethanol made it incompatible with
polycarbonate and most other plastics useful as reagent head materials.
In addition, its low surface tension caused it to creep over the surface
of ordinary structural plastics. Only Teflon met both the physical and
chemical requirements for delivering the 2-methoxyethanol reagent solution.
The combined low surface tension and low density of the reagent
solution simplified the choice of reagent head design. Its weaker capil-
lary action made its flow rate in the tape much slower than that of the
aqueous aldehyde reagent. As a result, the elution effect was much
smaller and the color spots were more uniform. The straight centerfeed
channel [Figure 9(a)] still gave erratic feeding, however, by occasionally
carrying some reagent out of the delivery tube during reagent addition.
This meant that there was not enough reagent to completely fill the reagent
head channels during the next addition period, thus resulting in a missed
49

-------
spot. The funnel-shaped cup design [Figure 9(g)] finally proved satis-
factory so the reagent heads were made of solid Teflon with a delivery
channel of this shape.
The sampling, reagent addition, and reading heads had the same
positions as on the aldehyde monitor (Figure 8) with a 1.5 cm sampling
spot diameter and 2.2 cm diameter measuring areas. The uniformity of
the spots eliminated any need to block out their center portion for
transmission measurement. The reagent syringe pump and the interval timer
occupied the same positions shown in Figures 13 and 14. The amine analy-
sis cycle required different time intervals and these are shown in Fig-
ure 19. Again, the electronic circuit was not changed; only the electri-
cal connections were changed to accommodate the pump, interval timer, and
additional reading heads as shown in Figure 12. The front panel of the
monitor, with its cover removed, is shown in Figure 20.
6.3 LABORATORY TESTS
Operation of the prototype model was tested in the laboratory with
both monomethylamine and propylamine. The acetate buffer tape collected
the amines readily at air sample flows of 0.5 and 1 £/min. Experiments
with different reagent addition head designs showed that using excess
reagent to completely flood the sample spot produced more reproducible
opacity measurements than precisely metered reagent addition. The flow
of reagent in the tape paper is irregular and forms unsymmetrical spots.
Flooding with reagent gives a large color spot exceeding the reading
head area, but the collected amines react in place. Since the eluting
SAMPLING PUMP
OFF
TAPE ADVANCE

0
REAGENT ADDITION
READOUT
AUTO ZERO
m
0 11 43 54
289 300 g
ELAPSED TIME - SECONDS
Figure 19 - Amine Analytical Cycle
50

-------
tn
Figure 20 - Front Panel of Amine Monitor (cover removed)

-------

T
1.7 cm	-2.5 cm
SPOT i wteX'te&d:-: I spot
~~l	~/.T: ¦
SAMPLE	REAGENT
$
i	r
l	i
I	i
M-2.2 CM-H
READING
ZONE
1
Figure 21 - Amine-Ninhydrin Spot Geometry























J
i




















I


















—<
1
\

(CC
)NSI
	 Ill —i
r=
nvt
SA
MPL
m
CO
_ i
4 IV
IN.
SAM
_iJ
MG -
-1 &
/Ml
M.)
























yj.

-<
h





















V
«
\

L


/

L





R
F
ECO
1EA
RDE
DIN<
:r









/


[



%







\
/



\
t
1
r~


















V







1








i



0 1 2 3 4 5 6	7 8 9 10 11 12 13 14 15 16 17 18 19
|	| RUN NUMBER
HIGH	HIGH
BLANK	BLANK
Figure 22 - Amine Monitor - Depletion of Amine from
1000 ppm Propylamine Aqueous Solution
52

-------
action of the flowing reagent is small, the amines form a darker sample
spot inside the large reagent spot, thus giving a more reproducible
background color. This is illustrated in Figure 21.
The chart recording of the sampling and analysis runs was similar
to that for the aldehyde monitor (Figure 17) except for the longer mea-
surement time. Recorded measurements converging toward lower opacity
readings were more common with the amine analyzer, and the readings
sometimes reached a constant value near the end of the measurement period.
Reproducibility of readings was very good for a tape spot measurement.
A series of blank runs, made for comparison with test runs at a render-
ing plant, gave recorder readings ranging from 1 to 7.5 with an average
of 4 and a standard deviation of 1.8. Sequential readings on air
bubbled through an aqueous solution initially containing 1000 ppm
propylamine clearly recorded the loss of amine from the solution. Fig-
ure 22 shows the plot of these readings. High blanks, that is, blank
spots more dense than the sample spots gave below zero readings for
samples 5 and 8. These results indicated that the instrument could
monitor amine emissions for either sudden changes or long-term drifts
in concentration.
53

-------
SECTION 7
ORGANIC SULFUR MONITOR
7.1 ANALYTICAL METHOD
The presence of several different types of malodorous organic
sulfur compounds in the rendering plant emissions dictated the choice
of a total sulfur analysis instead of analysis for a single class or
a specific compound. The total sulfur analysis would also include any
sulfur dioxide or hydrogen sulfide, although these are not common to
rendering emissions. Oxidation of the organic compounds in the emissions
sample followed by complete reduction of all the sulfur compounds to
H2S would provide a sample for direct measurement on existing H2S tape
monitors. The reduction, however, would require continuous reaction of
all the oxygen and oxidized compounds in the flowing sample gas with
excess reducing agent. The need for a continuous supply of reducing
gas, such as hydrogen, would effectively make the monitor a fixed posi-
tion instrument, instead of portable, and would add considerably to
its cost. Modifying a tape monitor to collect and measure SO2 provided
a simpler solution to the problem.
Common methods of analysis for SO2 involve oxidation-reduction
reactions measured titrimetrically (starch-iodide) or coulometrically
(iodide or bromide) and were not suitable for adaptation to a tape
monitor. Colorimetric methods based on fuschin or rosaniline (the
West-Gaeke method, for example) are too complicated and technique-
sensitive for automation on a tape monitor. One of the most direct and
simple colorimetric methods for SO2 involves reaction with ferric ammonium
sulfate to form ferrous ions which give a deep pink-orange color with
1,10-phenanthroline. This is also one of the most sensitive methods,
giving a measurable color with less than 100 ng of S02. The two com-
pounds are stable and compatible so they can be mixed and added to the
tape as one reagent which develops a color as SO2 is absorbed. This
method was chosen, therefore, to measure the total sulfur as SO2 after
thermal oxidation of the sample.
Copper oxide wire pellets packed in a quartz tube and heated to
600°C in a small resistance-wire furnace served as the oxidizer. The
absorbing-measuring solution contained 0.2g FeNH4(S04)2*12 H2O, 0.65g
1,10-phenanthroline, and 40 ml glycerol in 65 ml of 1:1 methanol-water.
The glycerol acted as a hygroscopic agent to help provide water for the
reduction reaction. SO2 absorption and reaction take place at the same
time in tape wetted with the reagent solution.
55

-------
7.2 TAPE MONITOR DESIGN
The combined absorption-reaction property of the SO2 reagent per-
mitted modifying the basic tape monitor for continuous recording of
the color development as the oxidized sample gas passed through the
tape. The reagent addition head was placed just before the combined
sampling measuring head. Gravity feed was adequate in delivering enough
reagent during the tape advance to flood the tape. This assured that
the portion of the tape held in the sampling-measuring head during the
collection period was saturated with reagent. Using mixed methanol-
water instead of just water for the reagent solution avoided weakening
the paper tape so much that the air flow would tear it in the sampling
head.
The need to assure adequate water for the SO2 oxidation in the tape
made an auxiliary moist air supply necessary. This consisted of a bub-
bler with a fritted-glass gas dispersion tube immersed in distilled water
in a polyethylene bottle. Filtered exhaust air from the sampling pump
passed through the bubbler and joined the sample gas stream at the en-
trance to the sampling head. A valve and flowmeter in this recirculating
air line gave the necessary control and flow reading so the sample flow
could be set as the difference between the total flow and the recircu-
lated flow. Figure 23 shows diagrammatically the complete analytical
train including the reagent addition system. Connecting lines were all
Teflon tubing with appropriate Teflon or polyethylene fittings.
The mixed reagent, stored in a vented-cap polyethylene bottle
clamped near the top of the front panel inside the instrument, flowed
by gravity down through an all-Teflon solenoid actuated valve (Angar
Scientific Corp. 190P12) then through a glass-Teflon needle valve and
into the reagent head. The switching circuit that automatically opened
the tape head during tape advance also opened the reagent valve to let
reagent flood the tape as it advanced. A manual push button switch
allowed opening the reagent valve for preliminary flow adjustment or to
drain the reservoir.
Figure 24 shows the components inside the instrument with an iden-
tifying diagram in Figure 25. A transformer mounted on the hinged chassi
cover provided 55 V AC for the heating elements in the oxidation furnace 8
Operating the furance under steady state conditions eliminated the need
for temperature-controlling circuitry; using 55 V on the heating elements
rated for 120 V, assured extended life under continuous operating condi— '
tions. The front panel components, shown in Figure 26 and identified in
Figure 27 were essentially the same as those used for the aldehyde and
amine monitors. The reagent head was the simple centerfeed design, Fig-
ure 9(a), made from polycarbonate plastic. A polished plastic block
clamped under gentle spring tension held the tape against the reagent
to assure contact and even flow of the reagent. With the bubbler flow
set for 1400 ml/min and the total flow for 2000-2600 ml/min, giving
sample flows of 600-1200 ml/min, the monitor would automatically collect
56

-------
SAMPLES
OXIDATION

GAS
FURNACE

REAGENT
RESERVOIR
SOLENOID
VALVE
TAPE
SAMPLING-
REFERENCE
HEAD
MANUAL
VALVE
t
i	
READING
HEAD
FLOW
METER
I	
VALVE **
f


TOTAL

FLOW

METER


Figure 23 - Organic Sulfur Monitor - Sampling System

-------
Figure 24 - Organic Sulfur Monitor - Sampling System

-------
FURANCE
TRANSFORMER
REAGENT
STORAGE
TOTAL
FLOW
METER
FILLING
PORT
Ln
v©
Figure 25 - Analysis System Components

-------
Figure 26 - Organic Sulfur Monitor

-------
SAMPLING	TOTAL FLOW PRESSURE BUBBLER FLOW
TIMER	METER	BLOCK	METER	RECORDER
HEAD	ADVANCE
Figure 27 - Organic Sulfur Monitor

-------
and record continuously the SO2 in the oxidized sample air. Increasing
or decreasing the total sampling time then permitted covering a concen-
tration range from 10 ppm to over 10,000 ppm SO 2 if necessary.
7.3 LABORATORY TESTS
Initial experimental laboratory tests were made with standard
S02-Air mixtures to establish the collection and measuring technique.
Tests with ethanethiol solutions, in a heated bubbler to assure complete
volatilization of the sample, verified the operation of the oxidizing
furnace. The original plan to use pre-impregnated tape proved impracti-
cal because of the inherent reducing nature of the paper tape. Even
acid-washed tapes retained enough reducing capacity to cause an intoler-
ably rapid increase in background (blank) color development. Tape im-
pregnated with the ferric ammonium sulfate and ortho-phenanthroline mixture
remained stable if kept refrigerated below -10°C. At room temperature
(24°C) , however, the impregnated tape developed is full background color
in 6 hours. Figure 28 compares blanks from tapes stored in a refrigerator
and at room temperature.
Adding the reagent to the tape just before starting collection and
measurement provided the best general solution to the background color
problem. In addition to eliminating storage problems, the reagent addi-
tion technique permitted the use of standard, commercial paper tape as
received. This technique also supplied some of the water to the tape,
needed in the collection-measuring reaction, as part of the reagent
solvent. With all the reagents present in the tape during sampling,
the problems of irreproducible spot development and color elution were
circumvented; as long as the tape was uniformly wetted by the reagent,
the spots formed completely and reproducibly. Because of this, the
sulfur monitor is potentially the most precise and accurate of the tape
samplers.
Without the intermediate step of reagent addition, the recorded
chart of sulfur monitoring was simplified as shown in Figure 29. Only
the tape advance introduced stray points; the remaining points all
represented opacity measurements on a continuing basis as SO2 from the
oxidized sample gas collected and reacted on the tape. The small
tungsten lamp in the uppermost portion of the sampling-measuring head
illuminated the tape during collection; the photoconductive cell beneath
the tape recorded the opacity increase as the reaction took place on the
tape. Any marked change in sulfur concentration in the sample gas during
sampling changed the slope of the opacity curve. Collecting too long
or using too high a sulfur concentration with a low span setting (low
sensitivity) caused the opacity curve to stop rising as the tape reagent
became exhausted. A 10 min collection time at 1.0 to 1.5 £/min sample
gas flow was selected for routing tests in comparing laboratory and
plant tests.
62

-------
HOURS STORED
Figure 28 - Blank Development in Sulfur Tape
TIME
MIN.
10-
20-
30-
x!

\
SAMPLE READING
• TAPE ADVANCE
TAPE ADVANCE
I
10
I
20
Figure 29 - Organic Sulfur Recordings
63

-------
Figure 30 compares recorder readings for consecutive samplings
of 90 ppm SO2 drawn through the complete analytical train first with
the oxidizing furance off (cold) and then at temperature (600-650°C).
The first set of data, including all readings, averaged 15.8 with a
standard deviation of 4.5. The second set, again including all the
readings, averaged 19.8 with a standard deviation of 5.1. Within these
deviation limits, the copper oxide in the oxidizing furnace had no sig-
nificant effect on the sample readings. Discarding the extreme read-
ings, it had no effect at all. A series of 30 consecutive samples taken
unattended in the laboratory showed a standard deviation of 5.3 with
only one extreme reading.
64

-------
90 PPM S02 THROUGH COMPLETE ANALYTICAL TRAIN
RUN NUMBER	RUN NUMBER
Figure 30 - Oxidizing Furnace Blank

-------
SECTION 8
SUMMARY
All three tape samplers in their final design modifications per-
formed well enough in laboratory tests to warrant field testing at a
rendering plant. The aldehyde monitor showed the greatest deviation in
readings, including the greatest number of extreme readings because of
the inherent elution of the developing color compound by the flow of
reagent in the tape. The reaction was sensitive enough, however, so that
a change in the average over consecutive readings would show any large
change in aldehyde concentration. Both the amine and sulfur monitors
showed enough precision in their measurements to serve as continuous
monitors of concentration drifts, as well as large changes, if the oc-
casional extreme readings were ignored or if a running average was
taken over consecutive samples.
The three monitors in their cases ready for field testing are shown
in Figure 31. The flowmeters shown on the instruments were replaced with
metric unit flowmeters to make recording of data in metric units easier.
At the conclusion of the field tests, these instruments were delivered
to the Project Officer.
67

-------
Figure 31 - Aldehyde, Amine, and Organic Sulfur Monitors
Ready for Field Testing

-------
PHASE III
FIELD TESTING

-------
SECTION 9
INTRODUCTION
Field tests were carried out at a local rendering plant to uncover
any problems in the analysis method or instrument operation that environ-
mental conditions in a plant might introduce. Sampling the vent for the
press offered the most extreme environmental conditions in terms of am-
bient temperature, vibration, exposure to organic vapors, and assault
from flying crackles, etc. A sampling line from the roof vent of a
scrubber offered more general ambient operating conditions. Both these
positions were used for continuous operating tests to check initial per-
formance and the effect of subsequent modifications. The testing periods
covered the first operating shift at the plant starting in late after-
noon. This subjected the monitors to the greatest stress since plant
temperatures and operating loads generally reached their peak between
5:00 p.m. and 8:00 p.m. during the test months (June and July). The
sampling time was set at 10 min for all the plant tests.
Through the cooperation of Dr. Peter 0. Warner at the local county
air pollution control office, several concurrent odor samples, evaluated
by an odor panel, were compared to monitor readings.
71

-------
SECTION 10
PROBLEMS AND MODIFICATIONS
An immediate problem appeared during the first attempts to sample
exhaust gas from the cooker condenser. Condensed liquid, primarily
water, occasionally carried into the condenser exhaust and then flooded
the sampling system in the monitor. A simple cold-finger trap, oper-
ated at ambient temperature, placed in the sampling line acted as a
convenient drop-out for the condensed water. This trap was used for
collection at the press vent as well, but was not needed for sampling
from the exhaust scrubber vent.
Ambient temperatures caused the other problems encountered in the
plant tests. With both the aldehyde and amine samplers, evaporation of
reagent from the reagent head increased as the ambient temperature in-
creased, leading to missed spots or underdeveloped spots. This problem
could be solved without any change in the instruments by increasing the
delivery rate of the reagent syringe pump. The rate steps on the pump
delivery control switch were small enough to conveniently compensate
for the evaporation. Finer control was available through the adjustment
on the interval timer cam but was not needed.
Overheating of the electronic components was the most serious
problem uncovered by the field tests. When the monitors were operated
beside the press vents in an area of poor air circulation, the ambient
temperature reached 45°-50°C (115-122°F). Because the electronic com-
ponents generate heat internally, they approached, and in the organic
sulfur monitor sometimes exceeded, their 70°C (158°F) maximum rated
operating temperature. This produced erratic response in the zeroing
and measuring portions of the operating cycle. The rectifier bridges
in the regulated power supplies were replaced with higher power units
to avoid overloading at higher temperatures. The organic sulfur monitor,
because of the presence of the oxidation furnace and its transformer,
suffered most readily from overheating. Therefore, a small fan with a
fiberglass filter was installed to circulate air across the electronic
printed circuit boards.
No attempt was made to optimize the span, sampling rates, and
analysis cycle for each instrument and sampling point to obtain optimum
comparable data. Graphic comparison of some of the measurements, how-
ever, indicates the potential usefulness of the instruments for odor
emission monitoring. The aldehyde measurements in Figure 32, despite
their relatively large deviations, show a definite concentration dif-
ference between the laboratory samples and the plant samples. Figure 33
shows a copy of some of the recorded traces for these measurements along
with a description of the plotting sequence. The numerical values with
73

-------
100
90
80
70
60
50
40
30
20
10
0
4 MM ANNULAR READING ZONE IN TAPE HEADS
(10 MINUTE SAMPLING TIMES)













































LAB
ORA
1
TOF
350
3 Ml
?Y IV
PPIV
N. @
1EAS
HC
11/
>UR
HO
MIN
EME
NTS












































i










)





1
1
1
1
I
\
i

r/

J

(
\

s


1
1
I
\
I
I ,

*
» i

T 1 1
T













R
XH/
END
VUS1
ERI
SC
MG
RUB
»LAf
BER
UT
EX



E
T
































1
1



/





1
1
1









1
1
1








V
1
1
1




I
J
' '










2 3 4 5 6 7 8 9 10 11 12 13 14
RUN NUMBER














REN
FRO
DEF
NT
UNC
PRE
PL;
SS V
XNT
ENT
















I





r






7


!












1 /


/






/
if







1
1
1

f !




* ,
V

|
1
I
1








1
1 1 1



1234 56789 10 11
RUN NUMBER
1 2 3 4 5 6 7 8 9 10 11 12
RUN NUMBER
Figure 32 - Aldehyde Monitor - Plant Test Comparisons

-------
TIME
MIN.
10-
20-
30
40-
50-
60-
70-
80
90-
100-
1	
V
• • _
• • •
¦r...:..
.
• • •
• •
r.:
?•
I	h
10 20 30 40 50
RECORD READING
60
RUN 1 - 4.5
RUN 2 - <0
RUN 3-45
RUN 4-69
RUN 5-59
RUN 6-29
H	1
70 80
EXHAUST SCRUBBER
SAMPLING
SAMPLING PERIOD
TAPE TRANSPORT, REAGENT ADDITION
					 COLOR DEVELOPMENT READOUT
AUTO ZERO		TAPE TRANSPORT
ANALYTICAL SEQUENCE	1
Figure 33 - Exhaust Scrubber Samplings
75

-------
each run number designation give the recorder reading for that run,
taken at the end of the color development readout. Visual observation
of the spots on the tape confirmed the high aldehyde concentrations.
The measurements plotted in Figure 34 from some of the amine
analysis tests clearly show the low concentration of amines in the
exhaust scrubber exit gas, the consecutive readings averaging about the
same as the laboratory blanks. The amine concentration in the press
vent air varied both during a shift and from night to night as the mea-
surements from two different nights show. The organic sulfur measure-
ments showed less variation and were generally lower than the 90 ppm SO2
readings obtained in laboratory measurements. Figure 35 shows some ex-
haust scrubber and press vent sulfur measurements compared to a series
of laboratory measurements.
The ability of the tape monitors to produce consistent measurements
under the most adverse plant conditions (the press vent measurements in
these tests) indicates that they are potentially suitable for direct
plant odor emission monitoring.
76

-------
RECORDER
READING

-------
RECORDER
READING

-------
SECTION 11
ODOR PANEL COMPARISON
As a trial comparison, four odor samples taken concurrently during
the field testing were evaluated by an odor panel. With only one monitor
at each of the two sampling points, intercomparison of the different
classes of compounds or synergistic interactions could not be considered.
This, of course, would require a full project in itself.
Dr. Peter 0. Warner of the local county air pollution control
office took the odor samples using 0.25 I gas pipets. An odor panel,
assembled according to guidelines of Benforado, Rotella, and Horton,
evaluated the samples using the ASTM dilution method D1391-57. Table 7
gives the results of the comparison. The only immediate conclusions to
be drawn from the comparison is that the sulfur compounds did not appear
to interfere with odor detection of other odorous compounds and that the
monitors could record concentrations at or below the Federal guideline
concentration of 5.6 OU/m^ (200 OU/ft^).
Table 7 - Comparison of Odor Levels and Monitor Readings
Sampling Point Monitor Reading Odor Level, OU/m'
3
Press Vent
Amine
20
3.14 x 10
7
Scrubber
Amine
0.5
4.25
Press Vent
Sulfur
8.5
4.42 x 10
7
Scrubber
Sulfur
8.0
3.03
79

-------
SECTION 12
SUMMARY AND CONCLUSIONS
Collection and gas chromatographic analysis of emission samples
from various process stages in a rendering plant revealed the presence
of aldehydes, amines, thiols, organic sulfides, and organic acids as the
major odorous compounds. Table 1 lists specific compounds found in the
process emissions. The larger number of organic sulfur compounds and
their particularly obnoxious odors gave them priority over the organic
acids originally considered for monitor development in the program.
Laboratory experiments showed that the MBTH aldehyde method, the nin-
hydrin measurement for amines, and the ferric ammonium sulfate-ortho-
phenanthroline method for organic sulfur compounds (oxidized to SO2)
offered the best promise of adaptation for tape monitoring instruments.
Prototype monitors, based on modification of standard, commercial
H2S tape monitors, proved acceptable in laboratory tests for sensitivity
and reproducibility. The aldehyde monitor showed the greatest devia-
tions in readings because of inherent reagent-flow elution effects on
color spot formation. 2-Methoxyethanol solvent for the amine-ninhydrin
reagent gave acceptably uniform spot formation for amine monitoring.
Adding both reagents for the sulfur determination to the tape just be-
fore sampling avoided any spot formation problem in the sulfur monitor.
Final testing of the monitors in a rendering plant led to minor
modifications, primarily to prevent the electronic components from over-
heating under extreme conditions in the plant. All three monitors left
unattended sampled and analyzed automatically in the plant tests. Com-
parison of several amine and sulfur monitor readings with odor panel
evaluation of concurrent samples showed these instruments could record
concentrations at or below the Federal guideline concentration of 5.6
0U/m3 (200 OU/ft3). The results showed that, despite the simplicity
of the technique and inherent limitations of color spot methods, the
instruments could serve as emission warning monitors for undesirable
increases in odorous compound concentrations. The precision of the
amine and organic sulfur monitors indicated that they could be considered
for process control monitoring as well as for odorous emission concen-
tration warning.
More extensive and systematic field tests are clearly indicated as
the next step in developing this type of instrument for automatic, low-
cost odor monitoring. Correlations between odor panel evaluations and
the monitor analysis measurements or some combination of the amine, alde-
hyde, and organic sulfur measurements must be determined to establish
the practical application of the instruments for plant emission monitoring.
Once their practicality is established, redesign of the instruments,
81

-------
electronically and mechanically, should be undertaken to increase their
endurance and performance ability under the extreme ambient conditions
often encountered in plant operation. The redesign should also aim at
reducing manufacturing costs so the final design will offer low cost
instruments that can serve equally as on-line process control indicators
and plant emission control, alarm, and compliance monitors.
82

-------
SECTION 13
REFERENCES
1.	"Part 23: Water; Atmospheric Analysis," Annual Book of ASTM
Standards, American Society for Testing and Materials, 1916 Race
Street, Philadelphia, Pennsylvania 19103, 1970.
2.	Official Methods of Analysis of the AOAC, 11th ed., Association of
Official Analytical Chemists, P. 0. Box 540, Benjamin Franklin
Station, Washington, D.C. 20044, 1970.
3.	D. M. Benforado, W. J. Rotella, and D. L. Horton, "Development of
an Odor Panel for Evaluation of Odor Control Equipment," J. Air
Poll. Control Assoc., 19, 101 (1969).
4.	T. A. Burgwald, Identification of Chemical Constituents in Render-
ing Industry Odor Emissions, Final Report, Project No. C8172, IIT
Research Institute, 10 West 35 Street, Chicago, Illinois 60616,
January 26, 1971.
5.	Methods of Air Sampling and Analysis, Intersociety Committee,
American Public Health Assoc., 1015 - 18th Street, N.W., Washington,
D.C., 1972.
6.	W. Leithe (translated by R. Kondor), The Analysis of Air Pollutants,
Ann Arbor-Humphrey Science Publisher, Ann Arbor, Michigan, 1970.
7.	W. E. Ruch, Chemical Detection of Gaseous Pollutants and Quantita-
tive Analysis of Gaseous Pollutants, Ann Arbor-Humphrey Science
Publishers, Ann Arbor, Michigan, 1970.
8.	E. L. Piret and M. W. Hall, "Distillation Principles of Formaldehyde"
Solutions, Ind. Engr. Chem., 40, 661 (1948).
9.	E. Sawicki, "Photometric Organic Analysis" in Vol. 31 of Chemical
Analysis, Wiley-Interscience, New York, 1970.
10.	S. Siggia, Instrumental Methods of Organic Functional Group Analysis,
Wiley-Interscience, New York, 1972, and Quantitative Organic Analysis
via Functional Group, 3rd ed., Wiley-Interscience, New York, 1963.
11.	F. D. Snell and C. A. Snell, Colorimetric Methods of Analysis,
Vol. Ill, Organic-I, D. Van Nostrand Co., Inc., Princeton, New
Jersey, 1962, and F. D. Snell, C. T. Snell, and C. A. Snell,
Colorimetric Methods of Analysis, Vol. III-A, D. Van Nostrand Co.,
Inc., Princeton, New Jersey, 1967.
83

-------
12. A. C. Stern, Air Pollution, Vols. I-III, 2nd ed., Academic Press,
New York, 1968.
13.	F. J. Welcher, ed., Standard Methods of Chemical Analysis. 6th ed.,
Vols. 2 and 3, D. Van Nostrand Co., Inc., Princeton, New Jersey,
1966.
14.	E. L. Wick, "Volatile Components of Irradiated Beef," Chapter 2,
pp. 5-18 in Exploration in Future Food-Processing Techniques,
S. A. Goldblith, ed., MIT Press, 1963.
15.	Zarazir, Chovin, and Guiochen, Chromatographia, _3, 180 (1970).
84

-------