EPA-600/2-76-048a
March 1976
Environmental Protection Technology Series
MOLECULAR SIEVE TESTS FOR CONTROL OF
NOX EMISSIONS FROM A NITRIC ACID PLANT
Volume I
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
E PA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-048a
March 1976
MOLECULAR SIEVE TESTS FOR
CONTROL OF NOV EMISSIONS FROM A
/\
NITRIC ACID PLANT
VOLUME I
by
John T. Chehaske and Jonathan S. Greenberg
Engineering-Science, Inc.
7903 Westpark Drive
McLean, Virginia 22101
Contract No. 68-02-1406, Task 2
ROAP No. 21AFA-106
Program Element No. 1AB015.
EPA Project Officer: E. J. Wooldridge
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development"
Washington, DC 20460
-------
ABSTRACT
Performance testing for NO emission control was conducted by Engineering-
Ai
Science, Inc. on the Union Carbide PuraSiv N* unit currently controlling
emissions from the tail gas stream of the ammonia oxidation nitric acid
production facility of Hercules Inc. in Bessemer, Alabama.
Measurements of N09/N0 concentrations in the PuraSiv N inlet and outlet
£* Ji
streams were performed during 11 individual four-hour adsorption cycles
using continuous photometric analyzers. The inlet and outlet streams
were sampled simultaneously. NO concentrations were also measured at
X
the test sites using the EPA Method 7 reference procedure in order to
provide comparative data. Total NO mass loading to the sieve was
a
variable from cycle to cycle, ranging from 63,370 to 251,800 grams,
reported as N0_. The average efficiency of the control unit for the
cycles tested ranged from 98.68 to 95.92%. The integrated average
concentrations of NO emitted over the complete cycles ranged from 17
X
to 154 ppm with the average concentration for all eleven tests being
105 ppm.
This report was submitted in fulfillment of Contract Number 68-02-1406,
Task Number 2 by Engineering-Science, Inc. under the sponsorship of the
Environmental Protection Agency. Work was completed as of December 1975.
*Registered Trademark of the Union Carbide Corporation.
iii
-------
CONTENTS
Page
Abstract iii
List of Figures v
List of Tables vi
List of Abbreviations vii
Sections
I Introduction '1
II Summary 3
III Process Description and Operation 6
IV Sampling and Analytical Procedures 14
V Discussion of Results 32
Reference List 60
iv
-------
FIGURES
No. Page
III-l Schematic Diagram of Ammonia Oxidation 7
Process at Hercules Bessemer Plant
III-2 Flow Diagram for PuraSiv-N Molecular Sieve 9
(shown with A adsorbing and B regenerating)
IV-1 Schematic Diagram of Sampling Train Used on 17
Inlet Test Sites HV-12 and HV-23
IV-2 Schematic Diagram of Sampling Train Used on 19
Outlet Test Site HV-13
V-l Instantaneous NO Mass Loading and Control 37
Efficiency Test--2 (1010 to 1408 3/6/75)
V-2 Instantaneous NO Mass Loading and Control 38
Efficiency Test--6 (1400 to 1754 3/14/75)
V-3 Instantaneous NO Mass Loading and Control 39
Efficience Test--ll (0711 to 1114 3/14/75)
V-4 Instantaneous Inlet and Outlet NO Concentrations 41
and Integrated Outlet NO Concentration Test—2
V-5 Instantaneous Inlet and Outlet NO Concentrations 42
and Integrated Outlet NO Concentration Test—6
Ji
V-6 Instantaneous Inlet and Outlet NO Concentrations 43
and Integrated Outlet NO Concentration Test—11
li
V-7 Total Mass Loading Versus Average Control 47
Efficiency for Tests 1-11
-------
TABLES
NO. Page
III-l Nitric Acid Production Rates During Tests 12
IV-1 Process Conditions Monitored 23
IV-2 Calibration Gas Analyses 27
V-l Summary of Calculated NO Mass Flow Rates and 33
Control Efficiencies for Instantaneous Flow and
Concentration Data for Test-11
V-2 Summary of NO Mass Loading, Average Control 44
Efficiency ana Average Concentration of NO
Emitted Test-11 X
V-3 Cumulative Mass Loading, Average Control Efficiency 45
and Average Concentration of NO Emitted
V-4 Summary of Performance of PuraSiv-N Molecular Sieve 46
V-5 Fuel Consumption of Regeneration Heater 49
V-6 Summary of Electrical Power Usage of PuraSiv N System 49
During Tests 1-11
V-7 Comparison of Inlet (HV-12) N0x Concentrations 51
Determined by Photometric Analyzer and EPA
Method No. 7
V-8 Comparison of Inlet (HV-23) N0x Concentrations 53
Determined by Photometric Analyzer and EPA
Method No. 7
V-9 Comparison of Inlet (HV-13) N0x Concentrations 54
Determined by Photometric Analyzer and EPA
Method No. 7
V-10 Results of EPA Method 7 Analysis of Calibration 59
Gases
vi
-------
LIST OF ABBREVIATIONS
Symbol Meaning
C Degrees Centigrade
cfm Cubic Feet per Minute
F Degrees Fahrenheit
3
m /min. Cubic Meters per Minute
2
N/m Newtons per Square Meter
3
Nm /min. Normal Cubic Meters per Minute
(at standard conditions of 15.56°C
(60°F) and 760 milimeters of
mercury pressure)
ppm Parts per Million by Volume
psig Pounds per Square Inch, Gauge
scfm Standard Cubic Feet per Minute
(at standard conditions of
15.56°C(60°F) and 760 milimeters
of mercury pressure)
vii
-------
SECTION I
INTRODUCTION
Engineering-Science, Inc. (ES) was contracted by the Environmental
Protection Agency, Office of Research and Development, Control Systems
Laboratory, for the purpose of testing the proprietary Union Carbide,
PuraSiv N Molecular Sieve Process designed to control NO emissions.
Jx
Tests were performed on a PuraSiv N unit installed at the Hercules,
Inc. nitric acid plant in Bessemer, Alabama. The plant utilizes the
pressure ammonia oxidation process. The overall objectives of this
contract were to determine the effectiveness of the PuraSiv N system in
controlling NO emissions from the nitric acid plant tail gas stream,
Jv
and to determine the unit's utility usage, thus providing a data base
for subsequent technical/economic evaluations of the PuraSiv N system.
A subject of major consideration in this study was whether the unit tested
met Union Carbide claims that the PuraSiv N system limits NO emissions
Jd
to an average of 50 ppm.
The unit selected for evaluation was installed at the plant in May of
1973 and was utilized periodically during the day shift until the Summer
of 1974, at which time 24-hour operation was begun. The unit was not
operated from December, 1974, to February, 1975, due to a temporary
plant shutdown. The unavailability of an operating log for the unit from
the time of installation made it impossible to determine the exact number
of hours the unit had been used.
The tail gas stream from the nitric acid plant absorption tower passes
through the PuraSiv N unit, first undergoing cooling and demisting and
then encountering an adsorption bed which removes the water vapor and NO
Jt
contained in the stream. The unit contains two separate but identical
adsorbers which are utilized alternately. While one bed is operating
in the adsorption mode, the other bed is being regenerated. Each adsorp-
tion/regeneration cycle is timed to last four hours.
- 1 -
-------
The NO sampling sites were located at the inlet and outlet of the
JL
adsorption beds. Two distinct inlet sites were tested at different
times during the study. One inlet site was located before the feed
chiller and demister at the PuraSiv N/plant interface (known as the
"battery limits"). The other inlet site was located between the demister
and the adsorption beds. A total of 11 four-hour cycles were tested
during the period February 5 through 14, 1975. The testing consisted
of simultaneously measuring the inlet and outlet N07/N0 concentrations
" Jt
using continuous photometric analyzers. These continuous measurements
were supplemented by NO determinations using grab samples taken and
a
analyzed in accordance with the EPA Method 7 procedure, "Determination
of Nitrogen Oxide Emissions From Stationary Sources" as located at 40 CFR
60.85 dated July 1, 1974. A total of 117 grab samples were taken during
the 11 cycles tested. All field testing and subsequent laboratory
analyses were performed by personnel of ES and the subcontractor, Common-
wealth Laboratory, Incorporated.
This report is published in two volumes. Volume I contains the body of
the report, while Volume II contains the appendices.
- 2 -
-------
SECTION II
SUMMARY
The overall objectives of the testing program were to measure the
efficiency and utility usage of the proprietary PuraSiv N molecular
sieve system for control and recovery of NO emissions from the tail
X
gas stream of a nitric acid plant. Testing was conducted on a system
installed at the Hercules, Inc. nitric acid plant in Bessemer, Alabama.
Successful completion of the field testing program provided simultaneous
measurements of inlet and outlet N09/N0 concentrations during 11
£» X
separate four-hour adsorption cycles, and measurements of the system's
electrical, water and fuel usage. The analytical technique used to
obtain the NO measurements for evaluation of the molecular sieves per-
a
formance was a continuous photometric method. Measurements were also
made using the EPA Method 7 procedure in order to provide a comparative
data base.
Analysis of the field test data revealed that the highest instantaneous
control efficiency measured during any of the cycles was 99.53%, occurring
at the beginning of an adsorption cycle during a period of slowed pro-
duction. At the time, the inlet mass loading was approximately 25%
of the loading which usually occurs during normal production. It was
during this cycle that the lowest instantaneous outlet concentration was
measured, 6 ppm, occurring after the inlet mass loading had dropped to
about 20% of normal. The lowest instantaneous efficiency measured,
excluding an upset condition, was 91.88%, occurring at the end of an
adsorption cycle when the plant had been operating at 95% capacity.
It was at this time that the highest instantaneous outlet concentration,
291 ppm, was measured.
- 3 -
-------
Average control efficiencies for individual four-hour cycles ranged
from 95.92% to 98.68%. The highest average control efficiency occurred
during the cycle in which acid production and sieve inlet mass loadings
were at their lowest observed levels. The average concentration of
NO emitted during this cycle was 17 ppm. The highest average concen-
X
tration of NO emitted during a cycle was 154 ppm. During the 11 cycles
ji
tested, the average control efficiency was generally inversely proportional
to the total NO mass loading occurring during the cycle. A plot of the
Ai
average control efficiency versus the total mass loading for each test
cycle shows that the lowest efficiencies occurred during the cycles
with the greatest mass loading. The plot and further discussion are
presented in Section V of this volume.
For the 11 tests performed, the integrated average outlet NO concentra-
a
tions for the complete adsorption/regeneration cycles ranged from 17
to 154 ppm. The average for the 11 cycles tested was 105 ppm. The
cycle for which the integrated average outlet concentration was 17 ppm
was the only cycle where the average outlet concentration remained
below 50 ppm, the level claimed by Union Carbide as the maximum integrated
outlet NO concentration which is emitted over a cycle when the unit is
X
operated at design rate.
During all the cycles tested the instantaneous control efficiency
reached a maximum value within the first hour of the cycle and then
gradually decreased during the remainder of the cycle. All cycles
except one exhibited a temporary decrease in instantaneous efficiency
during the first 30 minutes of the adsorption cycle. The exact cause
of the temporary decrease was undetermined.
Electrical power usage by the system was greatly affected by the operation
of the refrigeration unit which supplies cooled brine to the PuraSiv
System. During some of the test cycles the refrigeration unit cycled
on and off, because cooler ambient temperatures placed less load on the
- 4 -
-------
unit. When the refrigeration unit was off, electrical power usage by
the process was reduced by approximately 50%. The average electrical
power usage for the four-hour test cycles in which the refrigeration unit
operated continuously was 236.6 kilowatt hours.
Treated water usage during the cycles average 314.5 liters per minute.
Fuel oil usage by the regeneration heater averaged 38 liters per cycle.
- 5 -
-------
SECTION III
PROCESS DESCRIPTION AND OPERATION
The Hercules Plant began producing nitric acid by the ammonia oxida-
tion process in 1929. Since then various portions of the plant have
been rebuilt and/or replaced. The last major revisions occurred in
1966 when the acid absorption tower was replaced by a larger unit,
and in 1973 when the PuraSiv N Process was added for NOX removal
and recovery. Current production capacity is 59 metric tons per
day (130,000 pounds).
Figure III-l illustrates the pressure process for the production of
nitric acid which is utilized at the Hercules Plant. In the process,
the PuraSiv N system is located after the absorption tower, between
the entrainment separator and the reheater.
In the process, ammonia (NHo) is first oxidized to nitric oxide (NO)
and then to nitrogen dioxide (N©2). The nitrogen dioxide is absorbed
in water to produce nitric acid (HN03). Oxidation of the ammonia
to nitric oxide is accomplished by passing a preheated mixture of
90 percent air and 10 percent ammonia by volume through a catalytic
reactor at a pressure of about 7.72 x 10^ N/m^ (112 psig) and a
temperature of 899°C (1650°F). The catalyst consists of a fine
platinum/rhodium wire gauze. The reaction is highly exothermic,
and goes to about 95% completion. The compounds leaving the
reactor (nitric oxide, oxygen, nitrogen and water vapor) are passed
through heat exchangers which cool the gases and promote the
further oxidation of nitric oxide to nitrogen dioxide and its dimer
dinitrogen tetroxide (^04). The recovered heat is used to pre-
heat reaction air and reheat the plant's tail gas stream for power
recovery.
- 6 -
-------
Eff1uent
FIGURE III-l, SCHEMATIC DIAGRAM OF AMMONIA OXIDATION PROCESS
AT HERCULES BESSEMER PLANT
Bypass
I
•— Main Inlet Gas Stream
Tail Gas Stream
Recovered NO..
Stack
Ammonia Vapor
Mixer
Ammonia Oxidizer
Catalyst
Preheater
Absorption
Tower
K /'. /« A:
! V V V \!
Reheater
Entrainment
Separator
Water
-A
Cascade
Cooler
Secondary Air
Compressor - Expander
Product Acid
55-65% HML
-------
The product stream passes through a cascade cooler (water con.lprl
exchanger) which cools the stream further before entering the absorp-
tion tower. By this time a portion of the N0£ which has formed has
reacted with condensed water vapor in the stream and produced a
solution of nitric acid and nitric acid mist. The nitric acid is
separated from the product stream by a knock out drum and the gas
stream enters the base of the absorption tower. Water is introduced
at the top of the absorption tower and secondary air at the bottom.
The secondary air provides the additional oxygen needed for the
formation of nitric acid. The absorption tower contains bubble cap
trays which provide countercurrent contact between the falling
aqueous solution and the rising gas stream. The product acid
leaving the bottom of the tower has a concentration in the range of
55 to 65%.
The gas stream leaving the top of the absorption tower passes through
an entrainment separator which removes entrained acid droplets, and
then enters the PuraSiv N molecular sieve control system. A sieve
bypass line is located between the entrainment separator and the
reheater. During the operation of the sieve, the bypass valve is
normally closed, but is set to open automatically when the flow
through the sieve exceeds 161.4 NmJ/min (5700 scfm) . The gas stream
5 2
before entering the sieve has a pressure of about 6.21 x 10 N/m
(90psig) and a temperature of about 32oc (90°F). The approximate
composition of the stream is 3% oxygen, 6000 ppm H90 and 3000 ppm NO ,
^ X
with the balance nitrogen.
Figure III-2 illustrates a simplified flow diagram for the PuraSiv N
system. The major components of the system are the feed chiller,
mist eliminator, two molecular sieve adsorption beds, regeneration
compressor, regeneration furnace, and regeneration coolers.
Upon entering the PuraSiv N System the NOX laden tail gas stream,
saturated with water vapor and acid mist, first passes through the
feed chiller and demister. Here portions of the water and NO
- 8 -
-------
i
VD
FIGURE 111-2, FLOW DIAGRAM FOR PuRASiv-N MOLECULAR SIEVE
(SHOWN WITH A ADSORBING AND B REGENERATING)
3
( ) - Typical Flow Rates Normal m /min
C - Closed Valve
Recovered NO Recycled Back
to Adsorption Tower Through
Cascade Cooler
Feed.Chiller
Tail Gas from
Entrainment
Separator
(161.4)
To + .
Neutralization
Tank
Regeneration
Compressor
V
I
r
!
1
NMisI
Eli mine
nv- \c A
cj
;
i tor
Regeneration J C
Heater
(Furnace)
Adsorber
A
Adsorber
B
.\
Regeneration <
Flow
(31.1)
•oo-
•c-a-
C
-04
|HV-13
J^ Regeneration
Treated Tail Gas to Coolers
Power Recovery
(130.3)
-------
vapors as well as some of the existing acid mist are removed.
Following the mist eliminator, the gas stream enters the top of one
of the two molecular sieve beds and passes through a moisture re-
moving layer. In the bed, residual NO is catalytically converted
to N(>2 in the presence of the 3% oxygen contained in the gas
stream. The N02 and ^0, are then adsorbed by the bed molecular
sieve material.
After passing through the adsorption bed the treated tail gas
stream is split. One portion of the stream is diverted for regen-
eration of the second bed and the remainder is piped back to the
plant for reheating and power recovery before it is discharged.
The two molecular sieve beds alternate between adsorption and
regeneration on a timed cycle. The tail gas is passed through
one bed, while the other bed is being regenerated. At the end of
the cycle the regenerated bed is placed in the adsorption mode and
the first bed begins regeneration. The cycle length is designed
such that the adsorption capacity of the adsorbing bed is not
exceeded before the end of the cycle. If the bed is subjected to
total NOX loadings in excess of the adsorption capacity, breakthrough
will occur prior to the end of the cycle. At the Hercules Plant
the cycle was programmed to last four hours. The design flow rate
for the system was 161.4 Nm^/min (5700 scfm) at an average NOX concen-
tration of 3300 ppm.
Regeneration of the second bed continues for the full four-hour
cycle and consists of two main stages, heating and cooling. Approxi-
mately 31.1Nm°/min (1100 scfm) of the treated tail gas stream
is diverted for the regeneration process and boosted to
8.27 x 105 N/m2 (120 psig) by a compressor to provide sufficient
pressure needed for recycling. During the heating portion of the
cycle the regeneration gas is heated by an oil-fired furnace to
approximately 220°C after which it passes through the bed from the
bottom upward, removing the adsorbed N02 from the sieves. The
- 10 -
-------
regeneration gas containing the desorbed NO- is returned to the acid
plant absorption tower via the cascade cooler for recycling of the
recovered N02. The heating portion of the regeneration cycle lasts
approximately two hours, after which the furnace is shut off and the
regeneration gas is diverted through two heat exchangers where it is
cooled to approximately 7°C (45°F). The stream is passed through the
bed in order to cool it before it is switched back to the adsorption
mode.
Refrigerated brine is used in the feed chiller and the second of the
two regeneration coolers. A closed loop brine system is furnished as
an integral part of the PuraSiv N package. The recirculating brine is
cooled by a refrigeration unit. The refrigeration unit accounts for
nearly half of the total electrical power consumed by the PuraSiv N
system. The balance of the electrical power is consumed by the regener-
ation compressor, the combustion air blower and other miscellaneous
electrical equipment associated with the system.
Treated cooling water is used in the brine refrigeration unit compressor,
the regeneration compressor, and the first regeneration cooler. This
water is not recirculated and comes from an external supply. The water
is essentially at ambient temperature.
Although plant personnel do not record hourly nitric acid production
rates per se, process data is recorded which permits the estimation
of the hourly production rates. Monthly yield checks are performed at
the plant to verify yield estimates. Prior yield checks indicate the
«j
plant produces 0.91 metric tons of nitric acid per day per 2.8 Nm /min
of feed air flow (1 ton per day per 100 scfm). The acid production rate
estimates contained in Table III-l were calculated from feed air flow
rates read from process instrument charts using this correlation. The
production rates estimate the tons of acid produced during each hour of
testing. There were times during the testing when the sieve bypass line
- 11 -
-------
Table III-l. NITRIC ACID PRODUCTION RATES
DURING TESTS
Metric Tons/Hour
TIME
3/5
3/6
3/2
3/11
3/12
3/13
3/14
0700-0800
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
1500-1600
1600-1700
1700-1800
1800-1900
1900-2000
(a) Upset
2.2
2.2
2.2
2.2
2.2
caused
(b) 37.9 m3/min
(c) Upset
(d) 17.0
(e) 17.0
caused
m^/min
m^/min
(f) 37.9 m3/min
(g) Upset caused
(h) 17.0 nT/min
2.3
2.3
2.3
2.3
2.2
2.2
2.2
x(a)
2.0
by failure
compressor
by failure
compressor
compressor
compressor
by failure
compressor
2.
1.
2.
2.
2.
2.
2.
2.
2.
of
put
of
put
put
put
of
put
(b)
2^0 )
1
6(c)
1
1
1
1
1
1
1
1.0 1.0
1.0 1.4(f'8)
1.0 2.2
1.2(e) 2.1
(h)
1.0 1.8 2.3W
(d)
1.5W 1.6 2.2
1.7 1.6 2.3
1.7 1.6 2.3
1.7 1.6 2.3
1.6 2.3
2.4
2.3
2.3
2.3
2.3
XRD at 1709
on line
37.9 m3
on line
on line
on line
XRD at
on line
at 0845
/min compressor at 1000
at 1530
at 1345
at 1130
1133
at 1355
- 12 -
-------
was opened due both to upset conditions and to tail gas flow rates in
excess of the bypass control valve set point. The plant has no means of
measuring the quantity of tail gas bypassing the sieve system. For this
reason the acid production data is not necessarily indicative of the
relative load delivered to the control system.
The acid production rates experienced during the testing program were
quite variable, primarily due to variable air compressor operation.
During normal operation the plant uses four compressors to supply the
reaction air needed for the process. One large compressor, an Ingersoll
Rand PRE rated at 76.2 m /min (2690 cfm), and two smaller compressors,
o
an Ingersoll Rand XLE and a Chicago Pneumatic rated at 37.9 m /min
o
(1340 cfm) and 17.0 m /min (600 cfm) respectively, supply approximately
60% of the reaction air. These three units are powered by electric
motors. After plant start-up, the XRD (reciprocating expander) power
recovery unit is used to supply the remaining 40% of the air requirements.
There were several reasons for the compressors going on and off line.
q
A minor mechanical problem with the 37.9 m /min compressor during test 3
caused a short-term reduction in the production rate. A major mechanical
o
problem with the 37.9 m /min compressor resulted in decreased production
rates during tests 6, 7, 8 and 9. A fluctuating requirement for com-
q
pressed air elsewhere in the plant resulted in the 17.0 m /min unit being
placed on and off line during tests 6, 7, and 9. The net result was
that during the testing program, production rates ranged from 40 to 100%
of nominal plant capacity. Also, during test number 8 problems with the
regeneration heater caused a delay in the regeneration of bed A, result-
ing in the extension of the adsorption cycle for an additional 47 minutes.
- 13 -
-------
SECTION IV
SAMPLING AND ANALYTICAL PROCEDURES
In order to provide the data base needed to evaluate the efficiency of
the Union Carbide PuraSiv N control system, it was first necessary to
select analytical methods for measuring the concentrations of the NO
Ji
species (NO, N09, N90. and acid mist) entering and leaving the sieve.
Z. £ H
The basic approach decided upon was to use an instrumental method capable
of providing continuous NO concentration profiles during sieve operational
Ai
cycles, and to supplement the data with measurements made using EPA
Method 7, "Determination of Nitrogen Oxide Emissions from Stationary
Sources", (PDS procedure). A major consideration in the selection of
the instrumental method to be used was the commercial availability of
reliable instrumentation.
Two instrumental methods, chemiluminescent detection and photometry
received primary consideration during the initial planning stages of the
testing program. Instrumentation was available using both analytical
techniques and either instrument offered the degree of range needed to
measure concentrations as low as 10 to 100 ppm, the expected outlet N02
concentration, and as high as 5000 ppm, the maximum expected inlet concen-
tration.
Process information for the nitric acid plant indicated that the tail
5 2
gas stream would be at approximately 6.21 x 10 N/m (90 psig) pressure,
32°C (90°F), and saturated with nitric acid mist. Because neither the
chemiluminescent analyzer or the UV-visible photometric analyzer were
capable of measuring nitric acid species, it became necessary to consider
another analytical method for measuring nitric acid concentrations.
- 14 -
-------
An investigation of the analytical methods specific for measuring
nitric acid revealed that continuous monitoring instrumentation was
not commercially available, but developmental work had been per-
formed. Miller and Spicer have reported the adaptation of a Mast
microcoulomb meter for use in measuring low concentrations of nitric
acid found in ambient air. ' ' However, adaptation of this method
for analyzing the higher concentrations of nitric acid found in the
tail gas stream was considered to be beyond the scope of this
test program.
The remaining alternative for measuring nitric acid was to collect
and analyze integrated samples of the tail gas stream. The inte-
grated samples would have to be collected isokinetically (due to
the presence of the acid mist) in order to obtain samples which
would have nitric acid concentrations representative of actual
concentrations in the tail gas stream. Because of the high pressure
of the tail gas stream, 6.21 x 105 N/m (90 psig), it would not be
possible to use a conventional sampling technique. A specially-
designed probe would be needed and sampling ports would have to be
installed, necessitating modifications to the plant's existing
piping. Consultation with Hercules concerning this matter made it
apparent that for safety reasons they were reluctant to allow any
modifications to the existing piping. With the approval of EPA it
was decided not to attempt either isokinetic sampling or the
isolation of nitric acid and nitric acid mist from the NO and N02
present in the tail gas stream. The nitric acid and nitric acid
mist concentrations would be included in the total NOX determined
by the PDS method. The decision was made to use photometric
instrumentation rather than chemiluminescent for the analysis of
NO and N0£. This decision was based on reports (unpublished study)
- 15 -
-------
of a corrosion problem that had occurred with a chemiluminescent
analyzer used for monitoring NO in the presence of acid mist.
X
SAMPLING LOCATIONS
Two different inlet sampling sites were tested daring the study.
Initially, during the first eight cycles of testing (Tests 1-8),
the inlet sampling was conducted at HV-12, an existing globe valve
on an 8-inch line located between the mist eliminator and the two
adsorption beds. Because the globe valve was not attached
directly to the 8-inch line, but was connected via a 5-foot
section of uninsulated 1/2-inch pipe which extended vertically
downward, it was necessary to use a knock-out pot to eliminate
acid which collected in the pipe as a result of condensation.
The condensation occurred at night and during the early morning
hours, when the outside temperature was less than the temperature
of the cooled tail gas stream leaving the mist eliminator. A
schematic of the sampling site is shown in Figure IV-1.
During the final three cycles of testing (Tests 9-11), the inlet
sampling was conducted at HV-23, an existing valve on an 8-inch
line upstream of the feed chiller and mist eliminator. The final
work plan for the test did not include testing at this location,
but upon a request by Union Carbide and approval by EPA, HV-23 was
added as an inlet sampling location. The same sampling setup was
used for HV-23 as for HV-12.
Initially, during the planning stages of the testing program, HV-23
had been considered as a possible inlet sampling location, but was
eliminated after the decision was made not to measure the acid mist.
This decision was based on the fact that continuous instrumentation
for monitoring acid mist was not available, and grab samples would
not have been representative without isokinetic sampling. Further-
more, the unique feature of the PuraSiv N system is the use of
molecular sieve adsorption to remove NO and it is primarily because
Ai
of this feature that the system is of interest. The cooler and mist
- 16 -
-------
FIGURE IV-1, SCHEMATIC DIAGRAM OF SAMPLING TRAIN USED ON
INLET TEST SITES HV-12 AND HV-23
Sieve Inlet
8" - Line
Outside
F
Valve HV-12
or HV-23
1/2" Line
Stainless Steel
Throttling Valves
Glass Stopcock
and Sampling
Port for PDS
Samples
By-Pass-
Dekoron Heated
Teflon Line
Stainless Steel
Knock-Out Trap
1/4"
By-Pass]
Line
\S
Inside Trailer
Exhaust
Instrument
Stainless Steel
Regulating Valve
Fl owmeter •*
Dekoron
Teflon Line
Head
Bucket of Water
Three-Way Valve
for Introduction
of Zero and
Calibration Gases
- 17 -
-------
eliminator are standard hardware items with documented performance in
similar applications. Because acid mist is known to have caused serious
sampling difficulties in the past and because only gaseous N0x was Co
be measured, sampling at HV-23 was not considered further until requested
by Union Carbide.
The rationale for finally adding HV-23 as an inlet testing site was that
because the feed chiller and mist eliminator are part of the packaged
system, a more representative system evaluation would be obtained if
samples were taken at HV-23. It was still feared that the acid mist
present at HV-23 might interfere with the sampling instruments, so sampl-
ing at HV-23 was scheduled after the HV-12 sampling was completed.
Fortunately, the knock-out trap and bypass system successfully minimized
acid mist related problems.
The sieve outlet sampling site used during all 11 cycles of testing was
an existing globe valve, HV-13, located on a 9-inch line which carried the
cleaned tail gas stream to the XRD power recovery unit. A schematic of
the sampling train used at this site is shown in Figure IV-2. Because
the cleaned tail gas stream was essentially moisture-free (normal opera-
tion of the sieve bed removes moisture from the tail gas stream) , it
was not necessary to use a knock-out pot. The sampling line was connected
directly to HV-13.
SAMPLING PROCEDURES
During the tests, the photometric analyzers and apparati for evacuating
the sample flasks and recovering the Method 7 grab samples were housed in
a 20-foot trailer. The trailer was heated in order to maintain a stable
temperature environment for the instrumentation. The photometric analyzers
were connected to the sampling ports on the sieve via thermostatically
controlled heated 1/4" Dekoron* teflon line which was maintained at about
32°C.
*Registered Trademark of Samuel Moore & Company -
- 18 -
-------
FIGURE IV-2, SCHEMATIC DIAGRAM OF SAMPLING TRAIN
USED ON OUTLET TEST SITE HV-13
Outsidei Inside Trailer
Stainless Steel
Throttling Valves
Glass Stopcock
and Sampling Port
for PDS Samples
Valve HV-13
/Sieve Outlet
9 " Line
Dekoron Heated
Teflon Line
Bucket of Water
Exhaust
Instruments —
Stainless Steel
Regulating Valve
Dekoron
Teflon Line
Three-Way Valve
for Introduction
of Zero and
Calibration Gases
8" Head
- 19 -
-------
Continuous Photometric Analysis
The NO concentrations at the inlet and outlet test sites were each deter-
a
mined by an individual Dupont 411 Photometric Analyzer and recorded on
strip chart recorders. The measurements were initiated at the beginning
of a four-hour adsorption cycle and when possible continued until the
completion of two full cycles. (Due to upset conditions caused by process
and sieve equipment failure, it was only possible to monitor one complete
four-hour adsorption cycle on March 5 and 11, 1975.)
During the tests, the photometric analyzers were manually switched back
and forth from the N09 to NO (NO + N09) modes of operation. Changing
Z. X fc
to the NO mode was accomplished by sealing the instrument sample cell
j£
and introducing excess oxygen for converting the NO to NOg. The N02
measurements that were made while the analyzers were in the N0_ mode
represented a real time variation of concentration with time, while the
NO measurements made while the sample cells were sealed, represented the
Ji
instantaneous NO + N02 concentration existing at the time the sample cells
were sealed.
During the first three adsorption cycles tested, the analyzers were placed
in the NO mode approximately every 15 minutes. Each of the NO measure-
X X
ments took approximately five minutes to insure complete conversion of NO
to N02. At the end of each N0x determination, the N02 measurements were
resumed. By completion of the third cycle of testing, it had become
evident that NO was not present in the outlet stream. (Normal operation
of the sieve catalyzes the oxidation of NO to NO ) After the third
cycle of testing, NO measurements at the outlet were made approximately
2v
once an hour instead of every 15 minutes.
EPA Method 7
The grab samples for the EPA Method 7 analysis were taken at the same
sampling sites as the samples for the continuous photometric analyzers.
- 20 -
-------
Thus during the first eight cycles (Tests 1-8) the inlet sampling was
conducted at HV-12. The outlet sampling site for all eleven cycles was HV-
HV-13.
The sampling procedures used for taking the grab samples followed the
EPA Method 7 procedures as located at 40 CFR 60.85, however, it was
necessary to deviate slightly from the published method and make the
following exceptions:
1. The evacuation of the sample flask, leak check, and recording
of the initial flask pressure was done inside the trailer.
several minutes before each sample was taken, rather than outside
the trailer at the sampling site.
2. A knock-out pot was incorporated into the inlet sampling train
upstream of the point of sample removal. A heated sample line
was not utilized.
Both variations from the standard sampling procedure were necessary. The
flask evacuation was performed inside the trailer in order to eliminate
the difficulties associated with setting up the evacuation apparatus out-
side, and to reduce the overall equipment requirements. By doing so, it
was possible to use one vacuum pump and one manometer for preparing the
flasks for each site. In addition one person was able to prepare the
flasks and take both samples. The knock-out pot in the sampling train was
required to stop condensation from entering the photometric analyzers.
Because the Method 7 sampling train used was designed to interphase with
the sampling train for the continuous photometric samples, it was not
possible to avoid its use. It is believed, however that the knock-out
pot was more suited for the purpose of eliminating the condensation occurring
in the 1/2 iuch process line than a heated sample line. A heated sample
line would not have been able to evaporate a sudden release of condensation
buildup from the 1/2 inch line.
- 21 -
-------
The grab samples were taken at the inlet and outlet sampling sites approxi-
mately every hour during the tests. (On March 5 and 11, the grab samples
were taken approximately every half hour.) Each sample was taken at a
time corresponding to an NO determination with a photometric analyzer.
X
The inlet and outlet sampling times were staggered by five minutes to
allow time for one person to take both samples. After the grab samples
were collected, the flasks were shaken for 5 minutes, allowed to sit for
16 hours and recovered following Method 7 procedures. The samples were
shipped to the analytical laboratory where the analysis could be carried
out in a nitrate-free environment.
In addition to taking NO samples at the inlet and outlet sampling sites,
Jt
three samples of each calibration gas and one of the zero gas were taken
as a means of checking the accuracy of the Method 7 procedure.
PROCESS DATA
In addition to the NO measurements, various process conditions were
Ji
recorded in a process log once an hour in order to characterize the inlet
and outlet streams with respect to flow rate, temperature, and pressure.
These data were obtained from existing gauges and recorders which were part
of the plant's process instrumentation. Table IV-1 lists the process
conditions measured.
The recorded records of the sieve inlet flow, regeneration flow, pressure
drop, and bed temperature were also made available by Hercules after com-
pletion of the testing.
The utility usage of the sieve system was measured in order to provide
data input for subsequent economic analysis. The parameters measured
were the amount of electrical power consumed by the sieve and associated
equipment, the volume of treated water used for cooling, and the fuel
consumption of the regeneration heater. Cooling water usage was determined
by installing a water meter (+ 2% accuracy) in the 3-inch treated water
- 22 -
-------
Table IV-1. PROCESS CONDITIONS MONITORED
PuraSiv N Process
Tail Gas Pressure
Tail Gas Temperature
Sieve Inlet Gas Flow
Sieve Inlet Gas Temperature (after mist eliminator)
Pressure Drop Across Sieve
Sieve Regeneration Gas Flow
Sieve Regeneration Gas Temperature (inlet to bed)
Bed Temperature During Regeneration (top of bed)
Sieve Regeneration Gas Temperature (outlet of bed)
Sieve Regeneration Gas Pressure
Sieve Outlet Gas Flow
Sieve Outlet Gas Temperature
Brine Temperature
PuraSiv N Utility Usage
Electrical Power
Cooling Water Flow
Regeneration Heater Fuel Consumption
Acid Production
Feed Air Flow
Air Receiver Pressure
System Pressure
Back Pressure
XRD RPM
Injection Water Flow
Acid Strength At Trays 4, 16, 26 & 29
Product Strength
- 23 -
-------
feed line. The water meter was not functional until the 7th test cycle
due to a delay in receiving a part need for repairing the meter which had
been damaged in shipment to the test site. Electrical usage was measured
with a recording volt meter and a recording ammeter (+ 3% accuracy) in-
stalled at the main power box serving the sieve system. Fuel consumption
was determined by periodically measuring the level of fuel in the storage
tank with a dip stick. The accuracy of the fuel measurement is unknown.
ANALYTICAL PROCEDURES
Continuous Photometric Analyses
The two DuPont Model 411 Photometric Analyzers used in the study determine
N02 concentration by measuring the absorption of light passing through
a sample at two different wave lengths (436 and 546 nm) . NO concentration
is determined by quantitatively converting it to N09 in the presence of
59
4.14 x 10 N/m (60 psi) of excess oxygen and measuring the absorbance.
For the outlet sampling site, the DuPont analyzer was equipped with the
standard 20-inch cell. The instrument was calibrated for 0-500 ppm full
scale. The inlet instrument was equipped with a special 7-inch cell and
was calibrated for 0-5000 ppm full scale.
The manufacturers specification for the precision of the instruments was
+ 1% of the full-scale range used. Thus for the outlet site, the precision
was + 5 ppm, and for the inlet site + 50 ppm. The accuracy of the measure-
ment is dependent on the accuracy of the calibration standards used. The
calibration gases for the inlet and outlet instruments were specified by
the manufacturer to be within + 2% of the stated concentration and trace-
able to a National Bureau of Standards primary standard.
The measurements of the N02 concentrations in the inlet and outlet streams
were made at atmospheric pressure and at the instruments cell temperature,
approximately 45°C. The reduction in pressure from process conditions of
5 2
6.21 x 10 N/m (90 psig) to atmospheric was accomplished using a series
- 24 -
-------
of regulating and throttling valves located after the sampling ports.
(See Figures IV-1 and IV-2.) The reduction to atmospheric pressure before
measurement was done for several reasons:
1. The reduction to atmospheric pressure shifts the equilibrium
between N02 and its dimer, N20^ to the left.
2N02 t N20A
Because N20, does not absorb light appreciably in the region of
436 nanometers it is not measured by the DuPont 411 Photometric
Analyzer. At the prevaling analytical conditions, atmospheric
pressure and a cell temperature of 43°C, the extent of dimeri-
zation in a sample containing 4000 ppm NO is approximately
0.6%. At lower concentration the extent of dimerization is
even less.
2. Pressure limitations of the instrument cell and valves would not
5 2
permit the introduction of 4.14 x 10 N/m (60 psi) of 02 into the
cell during the NO determination in addition to an initial
CO
cell pressure of 6.21 x 10J N/m (90 psig).
3. Operation of the cells at process conditions would have made it
necessary to monitor cell pressure continuously in order to
correct the instrument calibration for changes in pressure.
Thus, the reduction of the inlet and outlet sampling streams to atmospheric
pressure greatly simplified both sampling and calibration procedures. At
atmospheric pressure, the extent of the dimerization of N02 to N20, was
not considered significant and no corrections were made to account for it.
Neither instrument was equipped with a thermostically-controlled cell for
maintaining constant cell temperature. Heat produced by the electrical
components within the instrument cases resulted in cell temperatures that
- 25 -
-------
were above the ambient temperature. Cell temperatures were monitored by
mercury-in-glass thermometers attached to each cell. During the tests,
the instrument cases remained closed, keeping the sample cells at a stable
elevated temperature. The inlet instrument was originally calibrated at
a cell temperature of 46.1°C (115°F) , and the outlet instrument at a cell
temperature of 42.2°C (108°F). While the tests were being conducted, the
cell temperature varied slightly due to changes in temperature inside the
trailer. The range of variation was + 6°C from the calibration tempera-
tures. The maximum error expected from such temperature changes, based
on ideal gas law calculations, is + 2%. Most of the time, the instrument
cell temperatures were within + 3°C of the calibration temperatures.
Calibration Gases
Two N02 calibration gases were used in the field study. The gases were
obtained from the Rare and Specialty Gas Department of Airco Industrial
Gases, Airco, Inc., Riverton, New Jersey. The gases were supplied in
specially-treated aluminum cylinders to increase concentration stability.
Analyses supplied with the gases indicated the high concentration standard
contained 3980 ppm N02 and the low concentration standard contained 35 ppm
NO,,. In both cases the balance of the gases was nitrogen. Airco's N02
analyses were specified to be within + 2%, and traceable to a National
Bureau of Standards primary standard. Additional information on the method
of preparation and analyses of N09 standards was obtained from two papers
(2) (3)
which were subsequently published.
The N02 field calibration gases were checked in the ES laboratory prior
to the field testing by comparing them with another calibration gas
standard and also by comparing them with the internal calibration filters
contained in the DuPont analyzers. The high concentration standard
checked within 1% of the vendor analysis. The low standard, however,
- 26 -
-------
was found to be 33 ppm, approximately 6% below the manufacturer's stated
concentration. Although the standard was considered suspect, time did
not allow return of the standard for reanalysis prior to the field
testing. Thus it was decided that both gases would be submitted for
reanalysis after completion of the study.
After both gases were submitted for reanalysis, Airco revealed that the
low concentration N02 standard had been initially mislabled and that the
original analysis had been 31 ppm. Table IV-2 lists the concentrations of
the calibration gases reported by Airco before and after the tests.
Table IV-2. CALIBRATION GAS ANALYSES
Outlet Standard (ppm) Inlet Standard (ppm)
Date NO- N0«
1/28/75
4/14/75
31
32
3,980
3,910
Calibration Procedures
Initial instrument calibration and check-out was performed at the ES
McLean, Virginia laboratory. The standard start-up procedure, as outlined
in the instrument manual, was performed using the instrument's built in
test functions. The instruments were found to be within the factory
specifications stated in the manual. The linearity of each instrument's
response was determined over an absorbance range of 0 to 0.95 absorbance
units, (approximately 0 to 3,300 ppm for the inlet instrument and 0 to
950 ppm for the outlet instrument). The observed linearity deviations
were less than 1% of full scale.
Upon arrival at the test site, the instruments were allowed to warm up
overnight and the test functions were checked again. No changes were
observed from the way the instruments had performed in the laboratory.
- 27 -
-------
Using both instruments and the same gain settings as used in the labora-
tory, the calibration of both instruments was checked using the calibration
gases. For the high standard, both instruments yielded values within
+ 1% of the stated concentration. The low standard, when checked on
the outlet instrument measured 32 ppm. Because all indications were that
the reported concentration was incorrect, the decision was made to
consider the gas to be 32 ppm and not adjust the span settings to give
a 35 ppm output. If when reanalyzed the gas proved to be other than 32 ppm,
the measurements made during the study could easily be corrected.
The analysis reported by the vendor for the low concentration N02 standard
verified that the decision not to adjust the instrument span was correct.
No corrections to the measured inlet or outlet concentrations were
necessary.
Each day before initiating the NO measurements, the calibration of the
A-
photometric analyzers was checked using the calibration gases. The
instruments were zeroed before each calibration using zero nitrogen
containing < 0.1 ppm NO . The instruments were also zeroed periodically
JV
during testing by diverting the sample gas flow and introducing zero
gas into the instrument sample cell. Any zero drift which had occurred
was compensated for using the instruments' zeroing potentiometer. All
adjustments that were made were noted on the recorder strip charts and
written down in a field data log kept for each instrument. After com-
pletion of each day's tests, the calibration of each analyzer was verified
again.
EPA Method 7 Analyses
The EPA Method 7 procedure, "Determination of Nitrogen Oxide Emissions
from Stationary Sources", sometimes referred to as the phenoldisulfonic
acid (PDS) procedure, utilizes a wet chemical colorometric method of analy-
sis to determine the nitrate in species present in a grab sample. A sample
of the subject gas is collected in an evacuated 2-liter glass flask.
- 28 -
-------
The flask contains an absorbing reagent which consits of sulfuric acid
and hydrogen peroxide. In the flask NOV species, except nitrous oxide (N?0)
A "
are oxidized to the nitrate over a period of at least 16 hours. After
absorption, the solution is made alkaline with sodium hydroxide and slowly
evaporated to dryness. Phenoldisulfonic acid reagent is added and the
nitrate ion content is determined colorimetrically at 420 nm using a
spectrophotometer. A calibration curve is prepared using standard nitrate
solutions and the phenoldisulfonic acid reagent.
The accuracy and precision of Method 7 have been the subject of a number
of studies. A recent study has estimated the accuracy and pre-
cision of the method from collaborative testing results. The within-
laboratory precision, between-laboratory precision, and laboratory bias
were estimated to be 14.88%, 18.47% and 10.49% of the sample mean, respec-
tively. The method was sited as being accurate at the 95 percent level
of confidence; however, because of the relatively large within-laboratory
and between-laboratory standard deviations (estimates of the method pre-
cision) , there was considerable scatter among the observations. Of the
three different concentration levels tested in the study for determining
the method accuracy, the standard deviations ranged from 7 to 11% of the
sample means. Individual determinations ranged from 76 to 130% of the
true sample concentrations.
DATA REDUCTION
The strip charts containing the recorded outputs of the photometric
analyzers provided a nearly continuous record of the N0» concentrations,
and intermittent measurements of the NO (NO + N09) concentration. A data
Ai £
reduction technique was used which provided instantaneous sieve inlet/
outlet NO concentrations and corresponding flow rates at appropriate
Ji
intervals. The technique utilized linear interpolation and application
of the trapezoidal rule for integration. The following steps summarize
the technique used. A more complete description and examples of the
calculations are contained in Volume II, Appendix A of this report.
- 29 -
-------
1. The nearly continuous records of inlet and outlet N02 concen-
trations were converted to continuous form using linear inter-
polation in order to provide an estimate of the N02 concentra-
tions which existed while the analyzers were being operated in
the NO or zero modes. The technique was applied by drawing a
2i
straight line on the strip chart between the points where the
N02 measurements terminated and started again.
2. Zero drift which occurred during the intervals between zeroing
of the instruments was distributed over the interval using
linear interpolation. This provided an estimate of the instru-
ment zero baseline at any point in time.
3. Points were selected for reading the N02 concentrations from
the recorded charts. The points selected included, but were
not limited to, all points which corresponded to the beginning
of each NO measurement.
J\,
4. The NO 2 concentrations for each point were obtained from the
differences in the recorded outputs and the zero baselines,
using the proper instrument range factors.
5. NO concentrations were calculated from each of the recorded
<&
NO determinations using the same procedure as in step 4.
a
6. The ratio of N02 to N0x was computed for each point correspond-
ing to the beginning of an NO measurement.
X
7. For the inlet test sites, all NO- concentrations without corres-
ponding N0x measurements were converted to NO concentrations
using the calculated ratios of NO, to NOV. Linear interpolation
& H
was used to derive the ratios from the ratios determined in
step 6.
- 30 -
-------
8. For the outlet test sites, N09 and NO concentrations were
*• Ji
equal because of the absense of NO in the outlet stream.
9. Sieve inlet and regeneration flow rates (scfm) were read from
the recorded process charts corresponding to all points selected
in step 3. The flow rate discharged from the system to the
stack was calculated for each point by subtracting the regener-
ation flow rate from the inlet flow rate.
10. A computer program was used to calculate instantaneous mass
flow rates and instantaneous control efficiencies occurring
at each of the points during the cycles. The program also
calculated mass loading occurring over the intervals between
the points using the trapezoidal rule, cumulative mass loading
over the cycle, average efficiency during each interval, average
efficiency over each cycle, and average concentration of NO
X.
emitted over the cycle.
- 31 -
-------
SECTION V
DISCUSSION OF RESULTS
During the testing program at Hercules, eleven complete cycles of oper-
ation of the PuraSiv N unit were tested in order to determine the extent
of NO removal occurring during a cycle of operation. The NC^ concen-
trations at the inlet and outlet of the unit were measured continuously
during each cycle using photometric analyzers. The sums of NO and N02
concentrations existing at both the inlet and outlet were determined
periodically during each cycle by the same photometric analyzers. In
addition, 117 NO grab samples were taken for analysis by EPA Method 7.
X
Plant and sieve process data were obtained to permit calculations of
NO mass flow rates, nitric acid production rates, and sieve utility
X
usage. These data also provided a means of characterizing the overall
process conditions existing during each cycle of operation.
Contained in this section are tables and figures summarizing and illus-
trating the performance of the PuraSiv N unit, and comparing the test
results obtained using the photometric analyzer and EPA Method 7.
Table V-l summarizes instantaneous inlet and outlet NO concentrations
X
and sieve flow rates measured during test 11 and the corresponding
calculated instantaneous mass loadings and control efficiencies for
that test. Tables similar to Table V-l have been prepared for tests 1
through 10 and are contained in Volume II of this report. Test 11 was
chosen for illustration here because plant operating conditions during
the test were very stable and were closer to design conditions than
during any of the other tests.
The data presented in Table V-l are based on inlet NO concentration
Ji
measurements made at HV-23. Therefore, the four NO mass flow rates
- 32 -
-------
10
TABLE V-l, SUMMARY OF CALCULATED NOX MASS FLOW RATES AND
CONTROL EFFICIENCIES FOR INSTANTANEOUS FLOW
AND CONCENTRATION DATA FOR TEST—11
HERCULES INC., RUN 11, INLET HV-23, OUTLET HV-13 UNIT A 0711-1114 3/14/75
TIME
INT3
CYCLE
(MINI
0.0
3.0
4.0
6.0
14.0
15.0
19.0
23.0
34.0
39.0
49.0
59.0
64.0
79.0
94.3
97.0
109.0
124.0
139.0
154.3
160.0
169.0
184.0
199.0
214.0
220.0
229.0
243.3
INLET
FLOW
(SCFM)
5500
5500
5500
5500
5475
5475
5450
5450
5475
5475
5475
5475
5475
5450
5475
5475
5475
5475
5475
5475
5450
5425
5450
5475
5475
5475
5475
5475
REGEN.
FLOW
(SCFM)
1020
1020
1015
1015
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
1020
INLET
CONC.
(PPM)
3340
3320
3360
3310
3220
3220
3380
3390
3290
3280
3260
3200
3150
3120
3110
3100
3100
3130
3090
3080
3060
3060
3050
2990
2963
2960
2940
2880
OUTLET
CONC.
(PPM)
190
90
82
78
112
124
101
81
64
61
58
58
61
79
95
99
114
128
141
156
164
178
200
219
230
235
240
246
MASS
FLOW
RATE
IN
(G/MIN)
1010.0
1004.0
1016.1
1000.9
969.3
969.3
1012.8
1015.8
990.4
987.4
981.3
963.3
948.2
934.9
936.2
933.2
933.2
942.2
930.2
927.2
916.9
912.7
913.9
900.1
891.0
891.0
885.0
866.9
MASS
FLOW
RATE
ADS.
(G/MIN)
952.6
976.7
991.3
977.4
935.6
932.0
982.5
991.5
971.1
969.0
963.9
945.8
929.9
911.2
907.6
903.4
898.9
903.7
887.7
880.2
867.8
859.6
854.0
834.1
821.8
820.3
812.8
792.9
MASS
FLOW
RATE
REGEN.
(G/MIN)
10.66
5.05
4.58
4.35
6.28
6.95
5.66
4.54
3.59
3.42
3.25
3.25
3.42
4.43
5.33
5.55
6.39
7.18
7.91
8.75
9.20
9.98
11.22
12.28
12.90
13.18
13.46
13.80
MASS
FLOW
RATE
EMIT.
(G/MIN)
46.80
22.17
20.22
19.23
27.43
30.37
24.60
19.73
15.68
14.94
14.21
14.21
14.94
19.24
23.27
24.25
27.92
31.35
34.54
38.21
39.95
43.11
48.71
53.64
56.34
57.56
58.79
60.4 *
X
REOUC.
OF NOX
DUE TO
AOS.
94.31
97.29
97.56
97.64
9S.52
96.15
97.01
97.61
98.05
98.14
98.22
98.19
98.06
97.47
96.95
96.81
9S.32
95.91
95.44
94.94
94.64
94.18
93.44
92.68
92.23
92.06
91.84
91.46
X
REOUC.
OF NOX
DUE TO
REGEN.
1.05
0.50
0.45
0.43
0.65
0.72
0.56
0.45
0.36
0.35
0.33
0.34
0.36
0.47
0.57
0.59
0.69
0.76
0.85
0.94
1.00
1.09
1.23
1.36
1.45
1.48
1.52
1.59
X
REDUC.
OF
NOX
95.37
97.79
98.01
98.08
97.17
96.87
97.57
98.06
98.42
98.49
98.55
98.53
98.42
97.94
97.51
97.40
97.01
96.67
96.29
95.88
95.64
95.28
94.67
94.04
93.68
93.54
93.36
93.05
-------
given in Table V-l represent instantaneous values of: NO mass flow
2t
rate into the PuraSiv N system; NO mass rate removed by the combined
ji
action of the feed chiller, mist eliminator, and adsorption bed; NO
A
mass flow rate in that portion of the adsorption bed outlet stream that
is diverted for regeneration; and NO mass flow rate in the balance of
Jv
the sieve outlet stream that is discharged from the PuraSiv N system via
the power recovery unit to the plant outlet stack. For tests 1 through 9
where inlet NO concentrations were measured at HV-12, "Mass Flow Rate
X
In" represents the NO., flow rate entering the adsorption bed; and "Mass
X
Flow Rate Ads." represents the NO mass removal rate of the adsorption
J\.
bed only. It should be noted that because plant operation was so variable,
no directly comparable tests were obtained for the two inlet sites. From
the data that was obtained no definitive differences were noted in the
results from the two inlet sampling sites.
The instantaneous values of percent reduction given in Table V-l were
calculated from the instantaneous NO mass flow rates. Thus, it was
2t
possible to separately calculate the NO reduction achieved by the adsorp-
X
tion beds (including feed chiller and mist eliminator for runs 9 through
11) , and the additional NO reduction resulting from the diversion of
li
that portion of the adsorption bed outlet flow that is used for regenera-
tion. The overall effective reduction achieved by the PuraSiv N system
is a combination of these two removal mechanisms. This overall reduction
efficiency, as shown in the last column of Table V-l, was calculated
using the instantaneous NO mass flow rate emitted from the PuraSiv
J±
system to the plant stack and the instantaneous NO mass flow rate into
X*
the PuraSiv N system.
The inlet NO concentrations ranged from 3390 ppm to 2880 ppm during
test 11. The highest inlet flow rate during the test was 155.7 Nm3/min
- 34 -
-------
(a) •*
5500 scfm)v ', the lowest was 153.6 NnT/min (5425 scfm). Outlet NO
~ X
concentrations ranged from 58 ppm to 246 ppm which occurred at the end
of the cycle. The average outlet NO concentration was 139 ppm. For
Jv
tests 1 through 10 the sieve inlet flow rates and concentrations were
somewhat more variable. Test 3 had the greatest range of inlet flow
rates, 76.5 to 166.4 Nm3/min (2700 to 5875 scfm). The flow rate of
o
76.5 Nm /min, the lowest recorded during any test, occurred during a
temporary outage of the XRD power recovery unit. The highest inlet
q
sieve flow rate recorded was 180.5 Nm /min (6375 scfm) and occurred
during test 9. The highest inlet NO concentration measured was 4370
a
ppm, occurring during test 10; the lowest, 680 ppm, occurred during
test 6 when the plant was operating at about 50% of capacity. Outlet
concentrations ranged from a low of 6 ppm occurring near the beginning
of test 6 to a high of 291 ppm occurring at the end of test 2. One
higher outlet concentration, 308 ppm, was observed during test 3. This
was the result of a temporary upset condition and is not considered
representative of normal operation.
During test 11, the instantaneous percent reductions of NO ranged from
2w
a high of 98.55% to a low of 93.05%. During tests 1 through 10, the
maximum instantaneous percent reduction of NO , 99.53%, occurred at the
X
beginning of test 6, when acid production and inlet NO mass flow rate
4fe
were approximately 50% and 25% of the normal operating values, respec-
tively. The minimum percent instantaneous reduction (excluding an
upset period during test 3) was 91.88% which occurred at the end of the
adsorption cycle during test 2. Production during test 2 was about 95%
of plant capacity. During the tests, the contribution of the regeneration
process to the instantaneous percent reduction of NO ranged from a high
X
of 1.88% during test 2 to a low of 0.24% at the beginning of test 6.
(a)
All sieve inlet flow rates are based on standard conditions of
15.56°C (60°F) and 760 mm pressure, the standard conditions for
the calibrated readout on the plants' pro'cess instrumentation.
- 35 -
-------
Figures V-l through V-3 are plots of the instantaneous mass
loading and overall control efficiencies for tests 2, 6 and 11.
Tests 2 and 11 were chosen for illustration because they represent
tests made at each of the two inlet sites at similar operating
conditions. While the operating conditions were not identical,
they were closer for these two tests than any of the other
tests. Test 6 was included to show how the PuraSiv N performed
under highly variable operating conditions.
Examination of these figures reveals that the inlet NO mass
loadings were quite variable during the cycles. The variability
was due in part to the changes in production rate which occurred
when compressors were put on and off line. This is particularly
3
evident in the case of test 6. The 17 m /min compressor and the
3
37.9 m /min unit were both off line at the beginning of test 6.
3
The small 17 m /min compressor was put on line 80 minutes into
the cycle. Variability of NO mass loading was also due to changes
X.
in process conditions which occurred as the plant made a transition
from start up to normal production.
As is apparent in Figures V-l and V-3, the efficiency of the
molecular sieve unit rapidly increases immediately after the
freshly regenerated bed is switched on line. The efficiency then
decreases temporarily, followed by a rise to a maximum value for
the cycle, and then decreases steadily until a new bed is put on
line. The temporary decrease in efficiency occurred in all cases
except test 6. In this case, the molecular sieve was placed on
line after the plant had been shut down for the weekend. The
adsorption be
-------
FIGURE V-l,
c
o
U
3
-o
0)
DC.
C
O)
100
90
£ 1200
0)
4->
(O
800
700
600
500
400
300
200
100
0
INSTANTANEOUS NOX MASS LOADING AND
CONTROL EFFICIENCY TEST—2
(1010 TO 1108 3/6/75)
Total NOX Mass Loading 228,990 gms
Total NOX Emitted 9,340gms
Average Efficiency 95.92%
Outlet HV-13
50
100
150
200
250
Elapsed Time,(mins.)
- 37 -
-------
FIGURE V-2,
c
o
•r»
•M
0
T3
O)
IX.
0)
o
OJ
0.
400
375
350
325
300
275
250
225
200
175
150
\.
25
0
INSTANTANEOUS NOX MASS LOADING AND
CONTROL EFFICIENCY TEST—6
(1400 TO 1754 3/11/75)
Total NO Mass Loading 63,370 gms
A
Total NO Emitted 840 gms
rt
Average Efficiency 98.67%
Inlet HV-12
i
50
100
Outlet HV-13
150
200
250
Elapsed Time, (mins.)
- 38 -
-------
FIGURE V-3,
O
•a
O)
ac
0)
u
s_
to
on
o
in
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
INSTANTANEOUS NOX MASS LOADING AND
CONTROL EFFICIENCY TEST—11
(0711 TO 1114 3/W75)
Inlet HV-23
Total NO Mass Loading 227,550 gms
Total N0y Emitted 8240 gms
A
Average Efficiency 96.37%
Outlet HV-13
50
100
150
200
250
Elapsed Time,'(mins.)
- 39 -
-------
Figures V-4 through V-6 are plots of the inlet and outlet NO concentra-
.X
tion profiles and integrated outlet NO concentrations for test 2, 6 and
X
11. Examination of these figures reveals that the variation of inlet N0x
concentration was the major factor affecting the shape of the inlet mass
loading profiles shown in Figures V-l through V-3. The integrated outlet
NO concentrations for the complete adsorption/regeneration cycles were
X
154, 17 and 139 ppm for tests 2, 6 and 11 respectively. The integrated
outlet NO concentrations for the remaining cycles ranged from 63 ppm
X
for test 7, to 137 ppm for test 10. Test 6 was the only test for which
the integrated outlet NO concentration remained below 50 ppm, the
2i
level claimed by Union Carbide as the maximum integrated outlet NO
H
concentration which is emitted over a cycle when the unit is operated
at design rate.
Table V-2 summarizes mass loading, control efficiency, and average ppm
of NO emitted for intervals during test 11. Table V-3 summarizes
Ji
cumulative mass loading, average control efficiency and average ppm of
NO emitted during the test. Similar tables for the other ten tests
X
are contained in Volume II. The average control efficiencies for tests
1 through 11 ranged from 98.68 to 95.92% and are summarized in Table V-4.
The average efficiencies were calculated from the ratio of total mass
of NO emitted from the PuraSiv system to total mass of NO entering the
x x
system during each cycle. As was expected, the highest average efficien-
cy, 98.67% occurred during test 6 when the inlet mass loading was sub-
stantially lower than the loading for the other tests, due to a lower
production rate. The lowest efficiency, 95.92% occurred during test 2
when the plant was operating at approximately 95% of design capacity-
Figure V-7 is a plot of average control efficiency versus total mass
loading for tests 1 through 11. The graph reveals that average control
efficiency and total mass loading follow an inverse relationship. The
graph also shows that the average control efficiencies during tests 9,
10 and 11 which are based on inlet concentration measurements made
- 40 -
-------
FIGURE V-4
FIGURE V-4, INSTANTANEOUS INLET AND OUTLET NOX
CONCENTRATIONS AND INTEGRATED OUTLET
NOX CONCENTRATION TEST - 2
(1010 TO 1408 3/6/75)
3500
3000
2500
a.
a.
c
o
(O
4J
C
0)
u
c
o
o
—I 1 1—
Instantaneous Inlet HV-12
500
Instantaneous Outlet HV-13
200
100
0
Integrated Outlet HV-13
50 100 150
Elapsed Time, (mins.)
200
250
- 41 -
-------
FIGURE V-5, INSTANTANEOUS INLET AND OUTLET NOX
CONCENTRATIONS AND INTEGRATED OUTLET
Q.
Q.
-M
-------
FIGURE V-6, INSTANTANEOUS INLET AND OUTLET NOX
CONCENTRATIONS AND INTEGRATED OUTLET
NOX CONCENTRATION TEST - 11
(0711 TO 111/1 3/W75)
3500
0)
(J
C
O
O
3000
2500
500
0
200
100
0
Instantaneous Inlet HV-23
Instantaneous Outlet HV-13
Integrated Outlet HV-13
50 100 150 200
Elapsed Time, (mins.)
250
- 43 -
-------
TABLE V-2,
SUMMARY OF NOX MASS LOADING, AVERAGE
CONTROL EFFICIENCY AND AVERAGE CONCENTRATION
OF NOX EMITTED FOR INTERVALS DURING TEST~11
HERCULES INC.* RUN Hi INLET HV-23, OUTLET HV-13 UNIT A 0711-1114 3/14/75
INTERVAL
OF
CYCLE
(MINI
a.o- 3.0
3.0- 4.0
4.0- 6.0
6.0- 14.0
14.0- 15.0
15.0- 19.0
19.0- 23.0
23.0- 34.3
34.0- 39.0
39.0- 49.0
49.0- 59.0
59.0- 64.0
64.0- 79.0
79.0- 94.0
94.0- 97.0
97.0-109.0
109.0-124.0
124.3-139.1)
139.0-154.0
154.0-160.0
160.0-169.0
169.0-184.0
184.0-199.0
199.0-214.0
214.0-220.3
220.0-229.0
229.0-243.0
NOX
IN
DURING
INTERVAL
I GRAMS i
3021.0
1010.0
2017.0
7880.9
969.3
3964.2
4057.2
11034.0
4944.3
9843.5
9723.1
4778.7
14123.5
14033.2
2804.0
11198.1
14065.3
14042.7
13929.9
5532.2
8233.4
13699.9
13604.9
13433.2
5346.2
7992.2
12263.7
NOX
ADSORBED
DURING
INTERVAL
(GRAMS)
2894.0
984.0
1968.6
7651.7
933.8
3829.0
3948.2
10794.5
4850.2
9664.4
9548.5
4689.2
13808.2
1 3641 . 1
2716.4
10813.4
13519.0
13435.4
13259.3
5243.9
7773.3
12852.2
12661.0
12419.5
4926.3
7348.7
11239.6
NOX
RE GEN.
DURING
INTERVAL
(GRAMS)
23.55
4.81
8.93
42.54
6.62
25.24
20.41
44.72
17.53
33.37
32.53
16.68
58.89
73.19
16.32
71.67
101.79
113.14
124.92
53.84
86.31
158.99
176.23
188.85
78.23
119.87
190.79
NOX
EMITTED
DURING
INTERVAL
I GRANS 1
103.45
21.19
39.45
186.67
28.90
109.95
88.66
194.73
76.54
145.74
142.07
72.87
256.38
318.84
71.28
313.04
444.57
494.17
545.61
234.47
373.75
688.68
767.67
824.84
341.69
523.56
833.29
PERCENT
OF NOX
ADSORBED
DURING
INTERVAL
95.80
97.43
97.69
97.09
96.34
96.59
97.31
97.83
98.10
98.18
98.20
98.13
97.77
97.21
96.88
96.56
96.12
95.68
95.19
94.79
94.41
93.81
93.06
92.45
92.15
91.95
9' .65
PERCENT
NOX FOR
REGEN.
DURING
INTERVAL
0.78
0.48
0.44
0.54
0.68
0.64
0.50
0.41
0.35
0.34
0.33
0.35
0.42
0.52
0.58
0.64
0.72
0.81
0.90
0.97
1.05
1.16
1.30
1.41
1.46
1.50
1.56
PERCENT
REDUC.
OF NOX
DURING
INTERVAL
96.58
97.90
98.04
97.63
97.02
97.23
97.81
98.24
98.45
98.52
98.54
98.48
98.18
97.73
97.46
97.20
96.84
96.48
96.08
95.76
95.46
94.97
94.36
93.86
93.61
93.45
93.21
AVERAGE
PPM OF
NOX EMIT.
DURING
INTERVAL
140
86
80
95
118
113
91
72
62
60
58
59
70
87
97
106
121
134
148
160
171
189
210
224
232
238
243
-------
TABLE V-3, CUMULATIVE NOX MASS LOADING, AVERAGE CONTROL
EFFICIENCY AND AVERAGE CONCENTRATION OF NOX
EMITTED DURING TEST—11
HERCULES INC.t RUN 11, INLET HV-23, OUTLET HV-13 UNIT A 0711-1114 3/14/75
Ul
I
TIHE
INTO
CYCLE
(MINI
0
3.0
4.0
b.3
U.O
15.0
19.0
23.0
34.0
39.0
49.0
59.0
64.0
79.0
94.0
97.0
109.0
124.0
139.0
154.0
1*0.0
169.0
184.0
199.0
214.0
220.0
229.0
243.0
T3TAL
NOX
IN
(GRAMS)
3021.3
4031.0
6048.0
13928.9
14898.2
18862.4
22919.7
33953.6
38897.9
48741.4
584*4.5
63243.2
7736*. 7
91399.8
94203.9
105401.9
119467.3
133510.0
147439.9
152972.1
161205.5
174905.4
188510.3
201943.5
207289.6
215281.8
227545.5
TOTAL
NOX
ADS.
1 GRAMS!
2894.0
3878.0
5846.6
13498.3
14432.1
18261.1
22209.3
33003.8
37854.0
47518.4
57066.8
61756.0
75564.2
89205.4
91921.8
102735.2
116254.2
129689.6
142948.9
148192.8
155966.2
168818.4
181479.4
193898.9
198825.1
206173.9
217413.5
NOX
USED FOR
REGEN.
(GRANS)
23.6
28.4
37.3
79.8
86.4
111.7
132.1
176.8
194.3
227.7
260.2
276.9
335.8
409.0
425.3
497.0
598.8
711.9
836.8
890.7
977.0
1136.0
1312.2
1501.1
1579.3
1699.2
1890.0
TOTAL
NOX
EMIT.
(GRAMS)
103.5
124.6
164.1
350.8
379.7
489.6
578,3
773.0
849.6
995.3
1137.4
1210.2
1466.6
1785.4
1856.7
2169.8
2614.3
3108.5
3654.1
3888.6
4262.3
4951.0
5718.7
6543.5
6885.2
7408.8
8242.1
X
NOX
REDUC.
DUE TO
AOS.
95.80
96.20
96.67
96.91
96.87
96.81
96.90
97.20
97.32
97.49
97.61
97.65
97.67
97.60
97.58
97.47
' 97.31
97.14
96.95
96.88
96.75
96.52
96.27
96.02
95.92
95.77
95.55
X
NOX
REOUC.
DUE TO
REGEN.
0.78
0.70
0.62
0.57
0.58
0.59
0.58
0.52
0.50
0.47
0.45
0.44
0.43
0.45
0.45
0.47
0.50
0.53
0.57
0.58
0.61
0.65
0.70
0.74
0.76
0.79
0.83
AVG.
X
REDUC.
OF
NOX
96.58
96.91
97.29
97.48
97.45
97.40
97.48
97.72
97.82
97.96
98.05
98.09
98.10
98.05
98.03
97.94
97.81
97.67
97.52
97.46
97.36
97.17
96.97
96.76
96.68
96.56
96.38
AVG.
PPM
OF
NOX
EMIT.
140
126
111
102
103
105
103
93
89
83
79
77
76
78
78
81
86
91
97
99
103
110
118
125
128
132
139
-------
Table V-4. SUMMARY OF PERFORMANCE OF PURASIV-N MOLECULAR SIEVE
Test
Number
1
2
3
4
5
6
7
8
9
10
11
Date
3/05/75
3/06/75
3/06/75
3/07/75
3/07/75
3/11/75
3/12/75
3/12/75
3/13/75
3/13/75
3/14/75
Time of Cycle
1412-1800
1010-1408
1408-1812
0850-1255
1255-1658
1400-1754
1002-1404
1404-1806
1017-1503
1503-1905
0711-1114
Bed
on
Absorb .
8
B
A
B
A
A
B
A
B
A
A
Inlet
Sample
Location
HV-12
HV-12
HV-12
HV-12
HV-12
HV-12
HV-12
HV-12
HV-23
HV-23
HV-23
Total Mass
of NOX into
Sieve over Cycle
(grams)
184,680
228,990
212,030
236,870
168,030
63,370
136,970
155,720
235,260
251,800
227,550
Total Mass
of NOX
Emitted over
Cycle (grams)
6,670
9,340
8,110
7,790
6,000
840
2,850
3,900
6,060
9,170
8,240
Average
Concentration of
NOX Emitted over
Cycle (ppm)
112
154
136
127
99
17
63
73
98
137
139
Average
Control
Efficiency
(%)
96.39
95.92
96.18
96.71
96.43
98.68
97.92
97.49
97.42
96.36
96.38
Average
4 hr. 4 min.
178,716
5,865
96.72
NOTE: Tests 9, 10, and 11 are based on inlet test site HV-23.
-------
FIGURE V-7. TOTAL MASS LOADING VERSUS AVERAGE
CONTROL EFFICIENCY FOR TESTS 1-11
IUU
99
fr 98
^
0)
o
•r-
t
UJ
97
e
o
o
96
95
0
(
I I I 1 1
0
D
0
o
° D Oo
O
D
o = Adsorber A
o = Adsorber B
Dashed Points = efficiency based on HV-23 inlet"
test site located before mist
eliminator/feed chiller.
i i i i i
D 50 100 150 200 250
Total Mass Loading (kg)
- 47 -
-------
upstream of the feed chiller/mist eliminator are only marginally higher
at best than efficiencies based on similar mass loading derived frdm
concentrations measured after mist elimination. The existing scatter of
the points makes definitive evaluation difficult.
The utility usage of the PuraSiv N unit is summarized in Tables V-5
and V-6. Table V-5 summarizes the regeneration heater fuel consumption.
The average fuel consumption per regeneration cycle was 37.65 liters
(9.95 gal.) of number 2 fuel oil.
Table V-6 summarizes the electrical power consumption of the PuraSiv N
unit for each of the 11 cycles tested. The power consumed per cycle
was not constant due to slight differences in the lengths of each cycle,
and a variable demand by the brine refrigeration unit. (Cycle length
varied slightly due to inconsistencies of the control timers.) In the
case of test 9, the cycle was extended due to a malfunction of the
regeneration heater. The major factor affecting power consumption
during the cycles was the operation of the refrigeration unit. During
tests 1, 3, 5, 7, 8 and 10 the refrigeration unit remained on throughout
the entire cycle. During tests 2, 4, 6 and 9, the refrigeration unit
cycled on and off for part of the cycle and during test 11, the refrig-
eration unit was cycling on and off throughout the whole cycle. When
the refrigeration unit was off, electrical power consumption was cut
nearly in half. Cycles 2, 4, 6, 9 and 11 were morning cycles when the
outside ambient temperature was cooler and there was less load on the
refrigeration unit. It was noted that the refrigeration unit began to
cycle when ambient temperatures dropped below 10°C. Based on those
cycles in which the unit operated continuously, the average power consump-
tion during a four-hour cycle was 236.6 kilowatt hours.
- 48 -
-------
Table V-5. FUEL CONSUMPTION OF REGENERATION HEATER
Liters of
No. 2 Fuel
Period Oil Used
1300-3/5 to
1730-3/11 to
TOTAL
1700-3/7 363.4
1000-3/14 690.8
1054.2
Number Liters
of Regen. per
Cycles Cycle
12 30.28
16 43.18
28 Avg 37.65
103Btu
per
Cycle
1088
1551
1353
Table V-6. SUMMARY OF ELECTRICAL POWER
USAGE OF
PURASIV N SYSTEM
DURING TESTS 1-11
Test
1
2
3
4
5
6
7
8
9
10
11
Cycle Length
(Minutes)
228
238
244
245
243
234
242
242
286
242
243
Kilowatt Hours
224.9(b)
235.7(c)
241.7
215.2(c)
226.6(b>
223.5(c)
237.8(b)
241.3(c)
283.8^a^c^
248.0(b)
210. 6
(a)
(b)
(c)
(d)
Extended cycle due to malfunction of the regeneration heater.
Refrigeration unit operating continuously during the cycle.
Refrigeration unit cycling on and off during part of the cycle.
Refrigeration unit cycling on and off during the entire cycle.
- 49 -
-------
Treated water consumption measured from 0900 on 3/12/75 to 1100 on -3/14/75
was 943,771 liters, an average flow of 314.5 liters per minute. The
flow rates computed for various time intervals within the 50 hour measur-
ing period ranged from 199.1 to 334.2 liters per minute.
Tables V-7 through V-9 compare the NO concentrations measured using
Ji
EPA Method 7 with the NO concentrations determined by the photometric
Ji
analyzers. The EPA Method 7 NO samples were taken strictly for the
X
purpose of providing a comparison between the two analytical methods,
and were not intended to be used as a means of evaluating the efficiency
of the molecular sieve process.
It is generally recognized that there are problems associated with EPA
Reference Method 7 which include poor precision and an undemonstrated
accuracy. A more recent study has sited the method to be accurate
at the 95 percent level of confidence.
It must be recognized that both analytical methods do not measure exactly
the same NO species. EPA Method 7 measures oxides of nitrogen species
Ji
(NO, N203, N02, N204 and HNOg) which are collected in an evacuated flask
and oxidized to nitrate in the presence of hydrogen peroxide. The DuPont
411 Photometric Analyzer operated in the NO mode does not measure nitrates,
X
but measures nitrogen dioxide and nitric oxide which is oxidized to
nitrogen dioxide by oxygen added to the cell. Because EPA Method 7 is
sensitive to nitrates, any nitric acid mist or nitric acid vapor present
in the gas stream would be included in the analysis. The gas streams
for both inlet test sites HV-12 and HV-23 contained moisture and were
expected to contain nitric acid mist and vapor. More acid mist was
expected to be present at site HV-23 which was located before the feed
chiller and mist eliminator. The amount of nitric acid vapor in the
inlet stream is dependent on the equilibrium vapor pressure of the nitric
acid. The amount of nitric acid mist is now known.
- 50 -
-------
Table V-7. COMPARISON OF INLET (HV-12) NOX CONCENTRATIONS DETERMINED BY
PHOTOMETRIC ANALYZER AND EPA METHOD NO. 7
Ul
H*
Test
1
1
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
Date
3/05/75
3/05/75
3/05/75
3/05/75
3/05/75
3/05/75
3/06/75
3/06/75
3/06/75
3/06/75
3/06/75
3/06/75
3/06/75
3/06/75
3/07/75
3/07/75
3/07/75
3/07/75
3/07/75
3/07/75
3/07/75
3/07/75
Time
1415
1447
1516
1545
1645
1745
1015
1115
1215
1316
1415
1515
1616
1715
0915
1015
1115
1215
1316
1415
1517
1615
X
Concentration
by Photometric
Analysis (ppm)
2080
2010
2550
2650
2930
2680
2410
3100
3100
3450
2910
3350
3340
2420
2800
1610
4150
3910
2390
2050
2250
2200
Y
Concentration
by Method No. 7
(ppm)
2490
2270
3090
3040
3090
2160
2070
2840
3270
3370
2640
3080
2680
2630
2140
1800
4130
4140
2820
2760
2710
2300
A
Difference
410
260
540
390
160
-520
-340
-260
170
-080
-270
-270
-660
210
-660
190
-020
230
430
710
460
100
%A
Percentage
Difference
19.7
12.9
21.2
14.7
5.5
-19.4
-14.1
- 8.4
5.5
- 2.3
- 9.3
- 8.1
-19.8
- 8.7
-23.6
11.8
- 0.5
5.9
18.0
34.6
20.4
4.5
-------
Table V-7. COMPARISON OF INLET (HV-12) NOV CONCENTRATIONS DETERMINED BY
X
PHOTOMETRIC ANALYZER AND EPA METHOD NO. 7 (cont'd)
01
N>
Test
5
6
6
6
6
6
6
6
6
7
7
7
7
8
8
8
8
Date
3/07/75
3/11/75
3/11/75
3/11/75
3/11/75
3/11/75
3/11/75
3/11/75
3/11/75
3/12/75
3/12/75
3/12/75
3/12/75
3/12/75
3/12/75
3/12/75
3/12/75
Time
1658
1415
1445
1515
1545
1620
1645
1715
1745
1023
1117
1215
1316
1430
1516
1615
1745
X
Concentration
by Photometric
Analysis (ppm)
2310
1060
710
710
970
1260
1150
1140
1130
2090
2590
2290
2280
2350
2280
2520
2590
Y
Concentration
by Method No. 7
(ppm)
2860
1080
720
1130
1860
840
1090
1580
1430
1440
2870
2850
2690
2400
2840
2800
2420
A
Difference
550
020
10
420
890
-420
-060
440
300
-650
280
560
410
50
560
280
-170
%A
Percentage
Difference
23.8
1.9
1.4
59.2
91.8
-33.3
- 5.2
38.6
26.5
-31.1
10.8
24.4
18.0
2.1
24.6
11.1
- 6.6
%A - + 8.2
Y = 434 + 0.863 X
Where
Standard Deviation = 23.6
Y = Concentration by EPA Method No. 7
X = Concentration by Photometric Analyzer
Correlation Index = 0.788
-------
Table V-8. COMPARISON OF INLET (HV-23) NO,, CONCENTRATIONS DETERMINED BY
JV
PHOTOMETRIC ANALYZER AND EPA METHOD NO. 7
I
m
Test
9
9
9
9
10
10
10
10
11
11
11
11
Date
3/13/75
3/13/75
3/13/75
3/13/75
3/13/75
3/13/75
3/13/75
3/13/75
3/14/75
3/14/75
3/14/75
3/14/75
Time
1045
1130
1217
1315
1530
1630
1730
1845
0745
0845
0945
1045
X
Concentration
by Photometric
Analysis (ppm)
1800
1850
3020
3940
2980
2990
3050
3580
3290
3110
3080
2960
Y
Concentration
by Method No. 7
(ppm)
2060
1580
3130
3800
2510
2950
3480
3380
3040
3670
2530
2840
A
Difference
260
-270
110
-140
-470
- 40
430
-200
-250
560
-550
-120
ZA
Percentage
Difference
14.4
-17.1
3.6
- 3.6
-15.8
- 1.3
14.1
- 5.6
- 7.6
18.0
-17.9
- 4.1
ZA = - 1.9 Standard Deviation = 12.3
Y = 151 + .739 X
Where Y = Concentration by EPA Method No. 7
X = Concentration by Photometric Analyzer
e.
Correlation Index - 0.738
-------
Table V-9. COMPARISON OF OUTLET (HV-13) NO CONCENTRATIONS DETERMINED BY
X
PHOTOMETRIC ANALYZER AND EPA METHOD NO. 7
Test
1
1
1
1
1
, 1
£ 2
' 2
2
2
3
3
3
4
4
4
4
5
5
5
5
5
Date
3/05/75
3/05/75
3/05/75
3/05/75
3/05/75
3/05/75
3/06/75
3/06/75
3/06/75
3/06/75
3/06/75
3/06/75
3/06/75
3/07/75
3/07/75
3/07/75
3/07/75
3/07/75
3/07/75
3/07/75
3/07/75
3/07/75
Time
1420
1454
1520
1550
1650
1750
1021
1120
1220
1320
1420
1520
1621
0921
1021
1121
1220
1320
1420
1524
1620
1654
X
Concentration
by Photometric
Analysis (ppm)
80
64
81
91
135
179
88
71
145
268
186
92
93
64
56
150
232
96
64
95
125
139
Y
Concentration
by Method No. 7
(ppm)
67
81
71
87
143
157
101
49
89
131
161
82
99
35
29
140
156
58
81
62
130
94
A
Difference
- 13
17
- 10
- 4
8
- 22
13
- 22
- 56
-137
- 25
- 10
6
- 29
- 27
- 10
- 76
- 38
17
- 33
5
- 45
Percentage
Difference
-16.3
26.6
-12.3
- 4.4
5.3
-12.3
14.7
-31.0
-38.6
-51.1
-13.4
-10.9
6.5
-45.3
-48.2
- 6.7
-32.8
-39.6
. 26.5
-34.7
4.0
-32.4
-------
Table V-9. COMPARISON OF OUTLET (HV-13) NOV CONCENTRATIONS DETERMINED BY
X
PHOTOMETRIC ANALYZER AND EPA METHOD NO. 7 (cont'd)
Oi
Ui
Test
6
6
6
6
6
6
6
7
7
7
8
8
8
8
9
9
9
9
10
10
10
10
Date
3/11/75
3/11/75
3/11/75
3/11/75
3/11/75
3/11/75
3/11/75
3/12/75
3/12/75
3/12/75
3/12/75
3/12/75
3/12/75
3/12/75
3/13/75
3/13/75
3/13/75
3/13/75
3/13/75
3/13/75
3/13/75
3/13/75
Time
1420
1450
1520
1550
, 1645
1721
1750
1028
1120
1220
1430
1520
1620
1720
1049
1135
1200
1325
1535
1635
1735
1850
X
Concentration
by Photometric
Analysis (ppm)
10
6
10
14
21
29
29
66
61
59
58
51
72
91
45
30
44
124
56
85
146
261
Y
Concentration
by Method No. 7
(ppm)
40
8
11
96
23
23
26
54
49
65
46
49
86
101
42
36
64
112
58
101
46
92
A
Difference
30
2
1
82
2
- 6
- 3
- 12
- 12
6
- 12
- 2
14
10
- 3
6
20
- 12
2
16
-100
-169
%A
Percentage
Difference
300.0
33.3
10.0
585.0
8.7
- 20.7
- 10.3
- 18.2
- 19.7
10.2
- 20.7
- 3.9
19.4
11.0
6.7
20.0
45.5
- 9.6
3.6
18.8
- 68.5
- 64.7
-------
Table V^9. COMPARISON OF OUTLET (HV-13) NO CONCENTRATIONS DETERMINED BY
A
PHOTOMETRIC ANALYZER AND EPA METHOD NO. 7 (cont'd)
Ul
ON
Test
11
11
11
11
Date
3/14/75
3/14/75
3/14/75
3/14/75
Time
0750
0850
0951
1050
X
Concentration
by Photometric
Analysis (ppm)
61
101
164
235
Y
Concentration
by Method No. 7
(ppm)
67
92
204
228
A
Difference
6
- 9
40
- 7
V
Percentage
Difference
9.8
- 8.9
24.4
- 3.0
%A - 10.7
Y = 27.4 + 0.576 X
Standard Deviation
98.91
Where Y = Concentration by EPA Method No. 7
X = Concentration by Photometric Analysis
Correlation Index = 0.601
Excluding the values with deviations of 300 and 585% (during test 6)
%A = -8.0 Standard Deviation = 25.93
Y = 22.97 + 0.607X
Correlation Index = 0.6311
-------
Because the EPA Method 7 samples cannot be taken isokinetically it is
impossible to estimate the contribution of nitric acid mist to the meas-
ured N0x concentrations. At the outlet site, both methods would be
expected to be measures of the same thing, since N02 is the only expected
species.
The tables comparing the results determined by the two methods are cate-
gorized according to the molecular sieve test sites, HV-12, HV-23 and
HV-13. In all three tables the difference and the percent difference
between the concentrations measured by EPA Method 7 and the photometric
analyzer have been computed.
Also, for each of the test sites, the mean and standard deviation of the
percent difference between the determinations have been calculated. A
linear relationship was assumed between the concentrations determined by
both methods and regression analysis was performed using the method of
least squares. The multiple correlation coefficient for the methods was
computed for each test site.
For the inlet test site, HV-12, the percent differences ranged from -23.6%
to +91.8%. The mean percent difference was +8.2% and the standard devia-
tion was 23.6%. For inlet test site HV-23 the range of percent difference
-17.9% to +18%, was narrower than for inlet test site HV-12. The mean
percent difference was -1.9% and the standard deviation was 12.3%. The
outlet test site, HV-13 had the largest range of percent difference be-
tween the results of the two methods, -68.5% to +585%. The mean percent
difference was +10.7% and the standard deviation was 98.91%. It is
apparent however, that the two Method 7 results with deviations of 300
and 585% are outliers. By eliminating these two values, the range becomes
-68.5% to +45.5%, the mean percent difference becomes -8.0% and the
standard deviation becomes 25.93%. This larger range is at least partially
due to the fact that there are greater relative errors in the Method 7
analytical procedures at low concentrations.
- 57 -
-------
EPA Method 7 results correlated more highly with the photometric results
at the inlet sites than the outlet site. The correlation indexes for
the inlet sites were 0.788 and 0.738 for HV-12 and HV-13 respectively.
For the outlet site, the correlation index was 0.601.
Table V-10 lists the results of three determinations of each of the N02
calibration gases by EPA Method 7.
For the low concentration calibration gas standard, EPA Method 7 yielded
an average result which was 26% higher than the vendors determination.
The Method 7 analysis for the higher concentration standard, excluding
determination number 1, averaged only 3% higher than the vendors deter-
mination .
In summary there was a better agreement between the concentrations
measured by the DuPont 411 Photometric Analyzer and EPA Method 7 for the
inlet test sites than for the outlet site. The EPA Method 7 results for
the inlet test sites in most cases collaborated with measurements made
using the DuPont analyzer. For inlet site HV-12, 67% of the Method 7
results were within + 20% of the DuPont analyzer, and 92% were within
+ 35%. For inlet site HV-23, 100% of the Method 7 results were within
+ 20% of the DuPont analyzer measurements. Analysis of the inlet instru-
ment calibration gas by EPA Method 7 further verified the agreement
between the two methods. The average of two determinations of the inlet
standard using Method 7 was only 3% higher than the vendors analysis.
(The result of a first determination of the inlet standard by EPA Method
7 was eliminated because it was suspected that the sampling line was
improperly purged.) It is believed that the differences between the
results of the two methods of measurement are due primarily to the lack of
precision of EPA Method 7. The within-lab coefficient of variation
estimate for replicate single determinations by Method 7 has been sited as
being 14.88% . Because the test procedures used in the study at Hercules
- 58 -
-------
Table V-10. RESULTS OF EPA METHOD 7 ANALYSIS
OF CALIBRATION GASES
Calibration Gas
Concentration
(ppm N02 + NO)
32.5
3970
Analysis
No.
1
2
3
Avg.
1
2
3
Avg. of
No. 2 & 3
Concentration
EPA Method No
(ppm NO )
H
40
39
A3
41
2700
4167
4030
4098
by
. 7
did not involve the taking of replicate measurements using Method 7, no
within-lab coefficient of variation estimate for these tests has been
calculated. Estimation of the within-lab coefficient by any other means
is considered beyond the scope of this study.
The EPA Method 7 results for the outlet test site corroborated with the
DuPont analyzer results, but did so in fewer cases than for the inlet
sites. When the two outlier values with deviations of 300 and 585% were
excluded, 61% of the outlet Method 7 measurements were within + 20% of
the DuPont analyzer results and 83% were within + 35%. Again it is
believed that the differences between the results of the two methods
are due primarily to the lack of precision of EPA Method 7.
- 59 -
-------
REFERENCE LIST
(1) "A Continuous Analyzer for Detecting Nitric Acid", David F. Miller
and Chester W. Spicer, presented at the 67th Annual Air Pollution
Control Association Meeting, June, 1974.
(2) "Preparation of Stable Pollution Gas Standards Using Treated Aluminum
Cylinders", Stephen G. Wechter, presented at the ASTM Calibration
Symposium at Boulder, Colorado on August 5, 1975.
(3) "Gas Standards, How Standard Are They?", Stephen G. Wechter and
Henry A. Grieco, presented at the ASTM Calibration Symposium at
Boulder, Colorado on August 7, 1975
(4) Collaborative Study of Method For The Determination of Nitrogen
Oxide Emissions From Stationary Sources. EPA Report 6501/4-74-025,
1973.
(5) An Improved Manual Method for NO Emission Measurement. EPA Report
R2-72-067, 1972. X
(6) Collaborative Study of Method For The Determination of Nitrogen
Oxide Emissions From Stationary Sources. EPA Report 650/4-74-028,
1974.
- 60 -
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. REPORT NO.
EPA-600/2-76-048a
2.
3. RECIPIENT'S ACCESSIOf»NO.
. TITLE ANDSUBTITLE
Molecular Sieve Tests for Control of NOx Emissions
from a Nitric Acid Plant; Volume I
5. REPORT DATE
March 1976
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
John T. Chehaske and Jonathan S . Greenberg
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Engineering-Science, Inc.
7903 Westpark Drive
McLean, Virginia 22101
10. PROGRAM ELEMENT NO.
1AB015; ROAP 21AFA-106
11. CONTRACT/GRANT NO.
68-02-1406, Task 2
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task Final; 11/74-12/75
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
project officer for ^fe report is E. J. Wooldridge , Ext 2547.
16. ABSTRACT
_, -, ,, .... ^^^ . .
The report gives results of performance testing for NOx emission control
on Union Carbide's PuraSiv N unit, now controlling emissions from the tail gas
stream of the ammonia oxidation nitric acid production facility of Hercules, Inc. in
Bessemer, Alabama. Simultaneous measurements of NO2/NOx concentrations were
performed in the PuraSiv N inlet and outlet streams during 11 individual 4-hour
adsorption cycles, using continuous photometric analyzers. NOx concentrations were
also measured at the test sites, using the EPA Method 7 reference procedure, to
provide comparative data. Total NOx mass loading to the sieve was variable from
cycle to cycle, ranging from 63,370 to 251,800 grams, reported as NO2. Average
efficiency of the control unit for the cycles tested ranged from 98. 68 to 95. 92%. The
integrated average concentrations of NOx emitted over the complete cycles ranged
from 17 to 154 ppm.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Air Pollution
Nitric Acid
Chemical Plants
Absorbers (Materials)
Nitrogen Oxides
Adsorption
Air Pollution Control
Stationary Sources
Molecular Sieves
Ammonia Oxidation
Tail Gas
PuraSiv N
13B
07B
07A
11G
13. DISTRIBUTION STATEMENT
TTnlimitpd
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
68
20. SECURITY CLASS (This raeel
22. PRICE
Unclassified
EPA Form 2220-1 (9-73)
- 61 -
------- |