EPA-600/2-76-015
January 1976
Environmental Protection Technology Series
MOLECULAR SIEVE NOX CONTROL PROCESS IN
NITRIC ACID PLANTS
Industrial Environmental Rssaarch 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 loom 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.
EPA 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-600/2-76-015
MOLECULAR SIEVE
NOx CONTROL PROCESS
IN NITRIC ACID PLANTS
by
Harvey S. Rosenberg
Battelle-Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
Contract No. 68-02-1323, Task 17
ROAPNo. 21ADH-008
Program Element No. 1AB014
EPA Task 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
January 1976
-------
ABSTRACT
An engineering analysis of the applicability of molecular sieve
technology to the control of nitrogen oxide emissions from nitric acid
plants has been conducted. Field test data from a commercial plant using
this technology show that after 6 months of operations the plant is still
controlling the NO emissions to well within the New Source Performance
X
Standard for nitric acid plants (1.5 kg of NO /metric ton of acid which is
X
equivalent to about 200 ppm NO in the tail gas). Field test data from a
Ji
second commercial plant taken 10 months after start up show that NO
2t
emissions are below the New Source Performance Standard even though the sieve
had been damaged by an accidental backup of nitric acid. The process
appears capable of achieving an average effluent NO concentration
Ji
of 50 ppm based on testing at the former plant, but the achievement of
the concentration was not demonstrated at the latter plant during the
performance test because of the prior acid damage to the sieve. The 2-year
sieve life has not yet been demonstrated, although there is no reason to
believe it cannot be achieved. Thus, it appears that molecular sieve
technology is technically feasible.
The economic feasibility of molecular sieve technology for this
application was assessed by comparing the total capitalized cost (including
investment and operating cost) of this technology with those of the catalytic
reduction and extended absorption processes. It should be noted that
molecular sieve technology is potentially able to limit the effluent NO
2t
concentration to 50 ppm, whereas catalytic reduction and extended absorption
usually limit the effluent NO concentration to only 200 ppm. The capitalized
A
cost for the molecular sieve process is higher than for the catalytic reduc-
tion process and lower than for the extended absorption process. However,
considering the shortage of natural gas required for catalytic reduction,
molecular sieve technology is probably economically feasible for most acid
plants. Molecular sieve technology becomes economically competitive with
3
catalytic reduction when the price of natural gas reaches $2.4/1000 ft .
iii
-------
TABLE OF CONTENTS
Page
INTRODUCTION 1
NITRIC ACID MANUFACTURE, AIR EMISSIONS, AND CONTROL ALTERNATIVES ... 2
Nitric Acid Manufacture 3
Sources of Atmospheric Emissions 6
Emissions Control Technology 7
Catalytic Reduction 7
Extended Absorption 11
Wet Scrubbing 11
Adsorption 15
Flaring 17
TECHNICAL EVALUATION OF MOLECULAR SIEVE CONTROL TECHNOLOGY 18
General Description of Molecular Sieves • . . . . 18
Use of Molecular Sieves for NO^- Recovery 18
Description of PuraSiv N Process 20
Performance of the PuraSiv N Process 28
/
Tests by Union Carbide Corporation 28
Tests by U.S. Army 32
Tests by Engineering-Science, Inc 37
Sieve Life 41
ECONOMIC EVALUATION OF MOLECULAR SIEVE CONTROL TECHNOLOGY 42
Basis and Procedures 42
Inves tments 42
Operating and Maintenance Costs 43
Capitalized Costs 45
Investment and Operating Costs 46
Molecular Sieve Adsorption 46
Catalytic Reduction 46
Extended Absorption 48
Cost Equations 48
Capitalized Costs 52
CONCLUSIONS AND RECOMMENDATIONS 54
REFERENCES 56
APPENDIX A
DETAILED COST DATA A-l
iv
-------
LIST OF TABLES
Page
Table 1. Generalized Material Balance for 2- Bed PuraSiv N Process
Applied to 100 Ton/Day Nitric Acid Plant ......... 25
Table 2. Generalized Material Balance for 4 -Bed PuraSiv N Process
Applied to 300 Ton/Day Nitric Acid Plant ......... 29
Table 3. Design Basis for Hercules and Holston PuraSiv N Units. . . 30
Table 4. Performance of Hercules PuraSiv N Unit During Three Day
Performance Run After Start Up .............. 31
Table 5. Performance of U.S. Army - Holston PuraSiv N Unit During
Three Day Performance Run After Start Up ......... 31
Table 6. Three Day Performance Tests of the Holston PuraSiv N Unit
By the U.S. Army ..................... 34
Table 7. Performance Tests of the Hercules PuraSiv N Unit by
Engineering-Science .................... 38
Table 8. Unit Costs Used in Calculating Operating and Maintenace
Costs ........................... 44
Table 9. Costs of PuraSiv N Process (8000 hr/yr operating time) . . 47
Table 10. Costs of Catalytic Reduction Process (8000 hr/yr operating
time) ........................... 49
Table 11. Costs of Extended Absorption Process (8000 hr/yr operating
time) ........................... 50
Table 12. Equations for Investment and Operating Costs for NO
Emission Control Processes at Existing Nitric Acid Plants. 51
LIST OF FIGURES
Figure 1.
Figure 2.
Figure 3.
Flow Diagram of a Typical 120-Ton-Per-Day Nitric Acid
Plant Utilizing the Pressure Process
Typical Nitric Acid Plant Tail Gas Catalytic Reduction
Unit Utilizing Natural Gas
Extended Absorption System on Existing Nitric Acid Plant .
5
9
12
-------
LIST OF FIGURES
(Continued)
Page
Figure 4. Caustic Scrubbing for NO Control at Canadian Chemical
Plant 14
Figure 5. Simplified Flow Diagram for Molecular Sieve Adsorption
of NC- From Nitric Acid Plant Tail Gas 16
a
Figure 6. Two-Bed PuraSiv N Process (Vessel A Under Adsorption
and Vessel B Under Regeneration Heating) 21
Figure 7. Four-Bed PuraSiv N Process (Vessels A and B Under
Adsorption, Vessel C Under Regeneration Cooling, and
Vessel D Under Regeneration Heating) 26
Figure 8. Performance of Hercules Adsorption Beds After About
40 Cycles 33
Figure 9. Inlet NOX Concentration to Adsorption Beds as a Function
of Temperature 36
Figure 10. Total NC) Mass Loading Vs. Average NO Removal 40
* x
Figure 11. Capitalized Costs for NOX Emission Control at Nitric
Acid Plants 53
vi
-------
TOPICAL REPORT
on
ENGINEERING ANALYSIS OF THE PURASIV N
CONTROL PROCESS IN NITRIC ACID PLANTS
to
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA
by
Harvey S. Rosenberg
Task 17, Contract No. 68-02-1323
INTRODUCTION
The main source of atmospheric emissions from the manufacture
of nitric acid is the tail gas from the absorption tower which contains
unabsorbed oxides or nitrogen. These oxides are largely in the form of
nitric oxide (NO) and nitrogen dioxide (N02). NO- is a red-brown acidic
gas with a pungent odor; NO is colorless and odorless but is oxidized by
the oxygen in air to NO-. The toxic effects of NO. are well documented
in the literature. Several control technologies are available to reduce
the nitrogen oxide emissions from nitric acid plants to 1.5 kg per metric
ton of acid (the standard for new plants). Among these is adsorption of
the nitrogen oxides on a unique molecular sieve developed by Union Carbide
Corporation especially for this purpose. The NO removed is recycled back
3C
to the absorption tower as N0? for additional acid production. The developer
and vendor of the molecular sieve process claims that the NO emissions can
be reduced to a guaranteed average level of less than 50 ppmv. The molecular
sieve is guaranteed to have a life of 2 years.
-------
The study described in this report was undertaken for the U. S.
EPA in order to evaluate the technical and economic feasibility of molecular
sieve control systems for nitric acid plant tail gas. The objective of
this study was to determine (1) the technical feasibility of the control
system on the basis of the analysis of data independently obtained from
testing of a molecular sieve adsorption system currently in operation and
from analysis of information made available by the system vendor and (2)
the economic feasibility of the system by comparison with other commercially
available control processes.
NITRIC ACID MANUFACTURE. AIR EMISSIONS.
AND CONTROL ALTERNATIVES
Nitric acid is used in the manufacture of ammonium nitrate and
in numerous other chemical processes. Ammonium nitrate, which is used as
both a fertilizer and in explosives, accounts for about 80 percent of the
nitric acid consumption. Nitric acid is produced by oxidation of ammonia
followed by absorption of the reaction products in dilute acid solution.
Most nitric acid plants in the United States are designed to manufacture
acid with a concentration of 55 to 65 percent, which may subsequently be
dehydrated to produce 99 percent acid.
At the beginning of 1974, 46 companies privately owned and operated
76 nitric acid plants in the contiguous 48 states, in addition to seven
plants operated for the U.S. government by five companies. Nearly all nitric
acid produced in the United States is for captive consumption. Capacities
of U.S. plants range from 40 to 1160 metric tons per day. Privately owned
nitric acid plants are located in 30 states with Texas, Louisiana, Missouri,
and Kansas leading in production capacity.
The reported U.S. capacity for production of nitric acid was 8.4
million metric tons per year at the beginning of 1974, exclusive of production
for the U.S. government. Production from these privately owned plants is
expected to be about 6.6 million metric tons in 1974, 6.9 million in 1976,
-------
and 7.1 million in 1980. Production is currently proceeding at about
78.5 percent of capacity.
Nitric Acid Manufacture
The essential raw materials for the modern manufacture of nitric
acid are anhydrous ammonia, air, water, and platinum- rhodium gauze as
catalyst. The essential reactions for the production of nitric acid by
the oxidation of ammonia may be represented as follows:
4NH3(g) + 502(g) - > 4NO(g) + 6H20(g) AH298 = -216.6 kcal (1)
2NO(g) + 02(g) - > 2N02(g) AH298 = "27'1 kcal
3N02(g) + H20(l) - > 2HN03(aq) + N0(g) A = -32.2 kcal (3)
Several side reactions reduce somewhat the yield of reaction (1) :
NH3(g) + 02(g)-» l/2N20(g) + 3/2H20(g) AH = -65.9 kcal (4)
4NH3(g) + 302(g) - > 2N2(g) + 6H20(g) A^g = -302.7 kcal (5)
4NH3(g) + 6NO(g) - > 5N2(g) + 6H20(g) AH2gg - -302.7 kcal (6)
2N02(g) - >N2°4 AH = -13-9 kcal (7)
Equation (3) is really an absorption phenomenon and is probably
the controlling step in making nitric acid. The rate of reaction (3) can
be increased by employing an absorption tower under pressure and cooling
and by using countercurrent absorption with graded strengths of acid.
Secondary air must be supplied to the absorption tower to provide for the
reoxidation of NO formed and to desorb (bleach) the dissolved nitrogen oxides
which would color the acid. In the "pressure" process which is used for over
90 percent of the nitric acid manufacture in the United States, ' both
the oxidation of ammonia and the absorption of the resulting nitrogen
oxides are carried out at about 80 to 120 psig. High pressure oxidation
has the advantage of larger space velocity at the expense of shorter
catalyst life. Also, pressure oxidation results in a higher-strength
product acid.
-------
A typical arrangement of the pressure process for a 120 ton/day
nitric acid plant, with data on temperatures, pressures, flows, and
compositions, is shown in Figure 1. In this example, 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
converter at a pressure of 112 psig and a temperature of 1,650 F. The
catalyst consists of layers of fine wire gauzes, usually 80-mesh, made
with 0.003-inch-diameter wire composed of approximately 90 percent platinum
and 10 percent rhodium. Contact time with the catalyst is about 1 milli-
second, and conversion of the ammonia to nitric oxide and water is about
95 percent complete at these conditions. The remaining 5 percent is lost
through dissociation and side reactions. The resulting nitric oxide, oxygen,
nitrogen, and water vapor is passed through various arrangements of heat
exchangers. A filter at the exit of the heat exchanger train recovers
catalyst dust.
Oxidation of nitric oxide to nitrogen dioxide is a continuing
process once the gas leaves the converter; however, oxidation is favored
by low temperatures, and no significant formation of nitrogen dioxide occurs
until the gas enters the cooler condenser. As the gas becomes cooled,
water condenses and reacts with the newly formed nitrogen dioxide to produce
a weak nitric acid. Cooler condensers are usually designed with sufficient
surface and volume to allow for both cooling of the gas and oxidation of
nearly all nitric oxide to nitrogen dioxide.
The gas stream leaving the cooler condenser is passed through
a cyclone separator before entering the base of the absorber. The separated
condensate, which is 40 to 50 percent nitric acid, enters the column at an
intermediate point. The absorber contains bubble cap plates to provide
intimate countercurrent contact between the aqueous solution and the rising
gas stream. Because the absorption of nitrogen dioxide in this solution
and its reaction with water to form nitric acid are highly exothermic
processes, the tower is provided with internal cooling coils to remove the
heat of reaction. Water is fed to the top of the column for absorption,
and the secondary air enters the bottom of the column to provide oxygen for
the conversion of nitric oxide to nitrogen dioxide in the absorber.
-------
» EFFLUENT
MAIN INLET GAS STREAM
1AIL CAS STREAM
.TACK
COMPRESSOR - EXPANDER
PRODUCT SS-65% ,
HNO,
FLOW, Ib'hr
TEMPERATURE. T
PRESSURE. PS'J
NH, Vol %
NO, Vol %
NO.* Vol %
O. Vol %
H,0. Vol %
HNO Wt %
R.O. WI %
N,. VOl %
CO., Vol %
I
AMMONIA
2.900
170
170
100
_
_
_
O.OS
„
-
2
AIR
51.600
60
0
__
_
_
203
0.9
78J
3
MIX
47,050
450
i 112
100
_
_
18.7
0.8
h 17.5
4
CONVERTER
PRODUCTS
47,050
1.650
112
0
9.3
0
6.3
154
69.0
5
COOLED
GAS
38,350
105
98
0
l.l
6.4
_
0.9
„
91.6
6
CONOENSATE
ACID
8.700
105
98
_
_
\2
_
50-60
_
-
7
SECONDARY
AIR
7,450
450
120
_
_
_
20.8
0.9
__
7as
8
NITRIC ACID
PRODUCT
16,650
135
95
_
_
_
4O
_
-
9
WATER
3350
IOC
125
_
_
—
—
—
100
-
10
TAIL GAS
41.200
es
92
0.10
0.15
30
0.6
—
96.15
FIGURE 1. FLOW DIAGRAM OF A TYPICAL 120-TON-PER-DAY NITRIC ACID PLANT
UTILIZING THE PRESSURE PROCESS
Source: Reference 2.
-------
Unabsorbed gas, principally nitrogen, leaves the absorption tower
at a temperature of about 85 F and is passed through an entrainment
separator. The tail gas is then heated by heat exchange with the hot
process gases. The energy contained in the resultant hot gas is recovered
in a centrifugal expander, which drives the air compressor. The gas
leaving the expander is then discharged to the atmosphere.
Sources of Atmospheric Emissions
Atmospheric emissions from the manufacture of nitric acid occur
from the release of unabsorbed oxides of nitrogen in the tail gases from
the absorption tower. These oxides are largely in the form of nitric oxide
and nitrogen dioxide. The intensity of the reddish-brown color of the gas
depends on the concentration of nitrogen dioxide present. The tail gas
/
contains from 0.08 to 0.30 percent total nitrogen oxides, 2 to 3 percent
oxygen, and the balance water and nitrogen. About 50 pounds of nitrogen
oxides are discharged to the atmosphere for each ton of nitric acid (100
percent basis) made. Small quantities of acid mist usually are also present
in the gases from the absorption tower. This mist is vaporized when the
gas is reheated and usually enters the atmosphere as a gas. The quantity of
acid vapor discharged is relatively insignificant.
Emissions of nitrogen oxides may vary widely with plant operation
and with equipment. A survey of 12 nitric acid plants without tail gas
control equipment showed emission rates ranging from 17 to 111 pounds
(2)
of nitrogen oxides per ton of nitric acid. The main plant-operating
variables that affect tail gas concentrations adversely are insufficient
air supply to the system, low pressure in the system especially in the
absorber, high temperatures in the cooler-condenser and absorber, the
production of an excessively high strength product acid, and operation at
high throughput rates. Faulty equipment includes such items as improperly
operating compressors and pumps, and heat exchangers in which leaks between
rich and lean nitrogen oxide gas streams may occur.
Insufficient air supply may be due to poor compressor design,
malfunctioning of the compressor and power recovery equipment, or leaks
in the air supply system. Oxygen is usually supplied to the system by the
air compressor, which provides air for the oxidation of ammonia and also
-------
for the reaction to form nitric acid in the absorption tower. Lack of
oxygen in the system will hinder the oxidation of nitric oxide to nitrogen
dioxide and thus decrease absorption efficiency.
Recent plant design improvements include injecting a portion of
the air into the cooler-condenser to increase oxidation of nitrogen oxides
at this point. Tower volume and number of plates have an important bearing
on plant efficiency.
Pressure in the absorber is fixed by basic design of the unit,
by compressor capacity, and by the pressure drop in the lines leading to
the absorber. The rate of oxidation of nitric oxide to nitrogen dioxide
increases as the square of the pressure, and a small increase in absorber
pressure provides a substantial increase in oxidation and absorption rates.
High temperatures also cause a decrease in absorption efficiency.
The rate of oxidation of nitric oxide and the rate of absorption of nitrogen
dioxide in water vary inversely with the temperature and are favored by
low temperatures. Absorber temperature is largely a function of entering
gas temperature, ambient air temperature, and temperature and flow of the
cooling water that is circulated through coils on each plate. Throughput
of the cooling water is dictated by pump design, and its temperature is
fixed by the source of the water supply.
Emissions Control Technology
A number of methods are presently available for reducing nitrogen
oxide emissions from nitric acid plants. These methods include catalytic
reduction with certain fuels, extended absorption, wet scrubbing, adsorption,
and flaring. Catalytic reduction is by far the most widely used method
of abatement. About 35 nitric acid plants employ catalytic reduction for tail
gas treatment, but about 25 of these plants practice decolorization rather
than full abatement. About 5 plants employ other types of abatement methods.
Catalytic Reduction
Several types of catalytic reduction units have been used with
varying degrees of success. All employ fuel to reduce nitrogen oxides,
converting N09 to NO and NO to nitrogen. Nonselective reduction uses
-------
8
hydrogen, hydrocarbons, or carbon monoxide as fuel, which reacts with
oxygen in the tail gas, as well as with nitrogen oxides. In selective
reduction, ammonia reacts with NO in preference to oxygen. In all
n
catalytic reductions, catalyst activity decreases somewhat with use.
Nonselective catalytic reduction is particularly suited to the
pressure process for the manufacture of nitric acid in which the absorption
tower tail gas is of uniform composition and flow, is under pressure, and
can be reheated by heat exchange to the necessary reduction system inlet
*
temperature. Removal efficiencies above 90 percent are possible, and in
addition an economic return may be realized through recovery of heat
generated in the catalytic reduction unit if a use for steam is available.
In operation, the tail gases from the absorber are heated to the
necessary ignition temperature, mixed with a fuel such as natural gas, and
passed into the reactor and through the catalyst. There a number of
reactions take place resulting' in the dissociation and decomposition of
nitrogen oxides:
CH4 + 202 > C02 + 2H20 (8)
CH^ + 4N02 > 4NO + C02 + 2H20 (9)
CH, + 4NO > 2N2 + C02 +• 2H20 . (10)
Reactions (8) and (9) proceed rapidly, with the evolution of
considerable heat. Since the NO- is converted to NO in reaction (9), the
gas is now colorless even though there is yet no substantial destruction
of NO. The reaction of further amounts of natural gas with the NO in
accordance with reaction (10), which takes place more slowly, results in
the reduction of NO to nitrogen. The efficiency of abatement is directly
proportional to the degree of completion of reaction (10).
A typical catalytic reduction unit is shown in Figure 2. Temperature
and compositions are shown only for illustrative purposes, since in practice
actual operating conditions are governed by the kind of fuel employed, the
gas composition, and the type of catalyst. Usually the catalyst consists
of 0.5 percent platinum or palladium on a support such as woven Nichrome
ribbon or ceramic material having a pelleted or honeycomb structure.
Reduction from an inlet NO concentration of 2500 ppm to an effluent
concentration of 200 ppm.
-------
800-1000 F
Heated tail gas
0.3-0.5% NOX
3%02
SOpsig
Fuel
Effluent
to stack
Catalytic
reduction
unit
Steam
550 F
Opsig
0.01-0.20% NOX
0-2% CO2
0.5-2% 02
900-I500F
Water
Compressor
Tail gas
expander
800-1000 F
70 psig
Waste heat
boiler
FIGURE 2. TYPICAL NITRIC ACID PLANT TAIL GAS CATALYTIC
REDUCTION UNIT UTILIZING NATURAL GAS
Source: Reference 2.
-------
10
In a single-stage reduction unit, increasing the amount of fuel
raises the temperature of the gas as reactions (8) and (9) take place.
Generally, the upper temperature limit of the catalyst and reactor vessel
of about 1500 F is reached before all of the oxygen has reacted with the
fuel. This is especially true when natural gas is used, since a higher
inlet temperature is then required. Under these conditions, the effluent
is usually a. colorless gas in which nitrogen oxides are largely present
as NO.
When hydrogen or carbon monoxide is used as the reducing fuel,
the inlet temperature may be as low as 285 F, where as with methane the
minimum ignition temperature is about 800 F. Thus, a greater amount of
oxygen may be removed in a single stage when hydrogen is used, and the
subsequent reduction of NO to nitrogen may then take place according to
reaction (10) before the limiting temperature of 1500 F is reached.
/
In most instances natural gas is the most economical fuel.
Where temperature limits the extent of reaction in a single-stage reduction
unit, it may be necessary to treat the gas in two catalytic reactor stages,
with intermediate cooling. By this method, operation can be carried out
with excess fuel and the reduction of total nitrogen oxides to less than
100 ppm obtained.
Selective catalytic reduction must be carried out within the
narrow temperature range of 410 to 520 F. Within these limits, ammonia
will reduce NO- and NO to nitrogen, without simultaneously reacting with
oxygen. As a result, temperature rise in the catalytic reactor is very
small. A ceramic supported (either pellets or honeycomb) platinum catalyst
is used. The reactions are
6N02 - > 7N2 + 12H20 (11)
6NO - > 5N2 + 6H20 . (12)
Below 410 F there is a chance that ammonium nitrate will be formed. Above
520 F, oxides of nitrogen are formed by ammonia oxidation.
A number of older nitric acid plants are equipped with reciprocating
compressors for power recovery. If selective catalytic reduction units
were installed ahead of their expanders, ammonium salts could enter the
lubricating oil system, and an explosion could result. This is not a problem
with turboexpanders .
-------
11
Extended Absorption
In theory, emissions from nitric acid plants can be reduced
to meet regulations by means of more efficient absorption of the NO used
Ji
in acid production. Soclete Chimique de la Grande Paroisse of Paris,
France, has designed and engineered nitric acid plants utilizing extended
absorption technology for many years. Basically, the extended absorption
system forms part of a complete nitric acid process operating at dual
pressure (medium-pressure ammonia burner, high-pressure absorption), but
since most nitric acid plants in the U.S. operate under high pressure, the
extended absorption system lends itself to addition to existing plants.
The first NO abatement facility to meet the 200 ppm limit
A>
(1.5 kg NO as NO. per metric ton of 100 percent acid) utilizing the
X ^
extended absorption technique as an addition to an existing plant came
(3\
on-stream in the U.S. in late 1973. ' A simplified flow diagram of the
plant after addition of the extended absorption system is shown in Figure 3.
The system consists essentially of an additional absorption tower which
reduces the NO concentration from 2,000 ppm to 180 ppm. In effect, no
X
additional operating costs are encountered through use of the process.
The largest operating requirement is the additional electric power required
to overcome the pressure drop imposed by the new tower, but the cost of this
and other additional operating requirements is largely offset by the increased
production capacity of the plant.
Wet Scrubbing
N0_ can be absorbed by reaction with a large number of liquids,
including water, hydroxide and carbonate solutions, solutions of complex-
forming salts, organic liquids, and molten alkali carbonates and hydroxides
to form nitrates and/or nitrites in most cases. On the other hand, NO is a
relatively inert gas that reacts with few of the above liquids. However,
N.O-, which is in equilibrium with NO and NO- in aqueous solution according
to the following equation, is more reactive than NO..
NO + N02 ^± N203 (13)
-------
12
ammonia
tail!(.»»
r
T 3k
Mcompressor converter
LJ *ZIs
J T
^
t f '
1
1
t
ex tended r """[process
JT 1 absorber | ^ water
IT i
f^^\ |
absorber ^4~ ~] ^
r T
5Q Power
Y ^ J recovery
fck/T" B fc»/l ft— ht *- -W
^u. — x "" " ^ *L j k
heat condenser \^ ^/ ^V._. '
recovery i| 1
V /T^>i l"" !
f**\ 1 ^^J^weak
product 1 acid pump product ^
pump acid ™
/ . ^
P^
Extended Absorber Performance (360 ton/day plant)
NO inlet tower 2,000 ppm
NO* exit tower 180 ppm
Tower pressure drop 4.5 psi
Cooling water 150 gpm (7 F AT)
Electric power 108 kW*
Catalyst 0
Fuel (additional) 0
Maintenance (additional) 0
Process water (additional) 0
Operations man hours (additional) 0
Additional production 6.1 tons/day
88 kW required to offset 4.5 psi pressure drop.
FIGURE 3. EXTENDED ABSORPTION SYSTEM ON EXISTING NITRIC ACID PLANT
-------
13
(L\
Sherwood and Pigford showed that at an NO concentration of 1 percent,
2C
a gas containing equal molar quantities of NO and N0? was absorbed by a
46 percent NaOH solution as much as ten times faster than N0_ alone. They
found that the rate of absorption of NO + N0« was approximately proportional
to the first power of the NjO^ concentration — indicating the latter to be
the reacting specie.
Tail gas from nitric acid plants usually contains an excess of
N02 over NO, but the molar ratio is much closer to 1 than for stack gas.
In caustic scrubbing, NO in tail gas reacts with sodium hydroxide, sodium
Ji
carbonate, or ammonium hydroxide to form nitrite and nitrate salts according
to the following typical reactions:
2NaOH + 3N02 - > 2NaN03 + NO + H20 (14)
2NaOH + NO + N02— > 2NaN02 + H20 . (15)
Unless the nitrites and nitrates resulting from the absorption of NO are
treated further, the potential for water pollution by these highly soluble
salts is considerable.
A caustic scrubbing system was put into operation at a chemical
plant in Canada ' to treat vent gases containing NO and NO. and coming
from three sources: (1) tail gas from a nitric acid plant, (2) reaction
off-gases, and (3) vent gases from tanks and process equipment. The system
treated from 4,000 to 7,000 cfm of gas containing from 5,200 to 19,400 ppm
NO . The molar ratio of NO to N02 was about 0.5. The liquor system was
3C
designed for batch operation, starting with 3.5 to 4 percent caustic
solution which would be essentially exhausted before the spent liquor was
sent to the sewer and replaced with fresh liquor. The scrubbing system
is shown in Figure 4. The absorber was 8 ft in diameter and contained
41.5 ft of 2- in. random packed porcelain Raschig rings. The NO removal
x
efficiency ranged from 65 to 70 percent. The purge stream dumped to the
sewer contained 5.4 wt percent NaNO. and 1.5 wt percent NaNOg in a 2,570
gal batch which was dumped about every 10 hours .
Nitric acid plant tail gas also can be scrubbed with aqueous urea
solutions. The use of urea to remove NO is interesting because of the
2£
radically different chemical approach taken to the problem. The mechanism
of the urea scrubbing involves dissolution of the NO in slightly acidic
-------
14
Stack
Separator
drain
Low-pressure ^ ^ — |
off-gases ^f
* Blower
drum |
High- pressure S k
-»
off-gases r r
V
^
/
Nitric t
storage
..,, ^\_.*., .. .
r^
X—
Fume
absorber
j
Separator
MM^
\
)
Circulating
tank
Circulating pumps
\
I ,T^ Caustic
Caustic soda
50%
measuring tank
Water
To drain
FIGURE 4. CAUSTIC SCRUBBING FOR NO CONTROL AT CANADIAN CHEMICAL PLANT
-------
15
aqueous urea solution to form nitrous acid which reacts with the urea to
form nitrogen, carbon dioxide, and water. Some ammonium nitrate is also
formed in the process and this must be used as fertilizer.
The following mechanism has been proposed ' for the reaction of
nitrous acid with urea in an acidic solution:
+ CO(NH2)2 - > N2 + HNCO + 2^0 (16)
HNCO + HN0 - > N + C0 + 10 (17)
HNCO + H20 + H+ — > NH4+ + CO . (18)
Laboratory studies^ ' revealed that where the concentration of nitrous acid
is low, reaction (18) predominates over reaction (17), so that the overall
reaction, with nitric acid as the acidic species, is as follows:
HN02 + CO(NH2)2 + HN03 - > N£ + C02 + NH4N03 + H20 . (19)
Adsorption
Molecular sieves can selectively adsorb NO from nitric acid plant
tail gas. NO removal is accomplished in a fixed bed adsorption/catalyst
system, providing recovery and recycle of the nitrogen oxides back to the
nitric acid plant absorption tower. A very simplified flow diagram is shown
in Figure 5. The water saturated nitric acid plant absorption tower overhead
stream is chilled to 45-50 F, the exact temperature level being a function of
the NO concentration in the tail gas stream, and passed through a mist
eliminator to remove entrained water and acid mist. The condensed water,
which absorbs some of the N0» in the tail gas to form a weak acid, is
collected in the mist eliminator and either recycled to the absorption
tower or sent to storage. The tail gas then passes through a molecular
sieve bed where the special properties of the NO removal grade molecular
sieve result in the catalytic conversion of nitric oxide (NO) to nitrogen
dioxide (N0»). This occurs in the presence of the low concentrations of
oxygen typically present in the tail gas stream. Nitrogen dioxide is then
selectively adsorbed. The molecular sieve adsorbent/catalyst provides the
most effective performance and longest life when the tail gas is bone dry.
-------
16
Acid plant
absorber
overhead
Adsorption
Pre-treatment
Recycle to
cascade cooler
Clean taiJ gas
to power .recovery
Furnace
Regeneration
FIGURE 5. SIMPLIFIED FLOW DIAGRAM FOR MOLECULAR SIEVE ADSORPTION
OF NO FROM NITRIC ACID PLANT TAIL GAS
x
-------
17
This is accomplished by drying with a desiccant which exhibits very little
coadsorption of NOX during water removal. The desiccant is located in the same
adsorber vessel as the NOX adsorbent/catalyst in a compound bed arrangement.
Regeneration is accomplished by thermal swinging the adsorbent/
catalyst bed after it completes its adsorption step and contains a high
adsorption loading of NCL. The required regeneration gas is obtained by
using a portion of the treated tail gas stream to desorb the adsorbed
N0_ from the bed for recycle back to the nitric acid plant absorption tower.
This requires that the nitric acid plant absorption tower be able to
accommodate the recycle regeneration stream.
Flaring
Passing the gases that contain nitrogen oxides through a combus-
tion chamber or flame usually will reduce the nitrogen oxides to nitric
oxide and thereby provides a colorless emission. The extent of nitrogen
oxides destruction to nitrogen is influenced by the kind of burner or flare,
the type of fuel, and the initial concentration of NO in the waste gas.
2x
The performance to be realized for any given set of conditions must be
determined experimentally. Flaring is usually too costly for treatment
of nitric acid plant tail gas because of the large volume of gas and low
initial NO concentration.
x
For example, a typical 100 ton/day nitric acid plant generates about
7800 scfm of tail gas containing 0.08 - 0.30 vol. percent NO
x
-------
18
TECHNICAL EVALUATION OF MOLECULAR SIEVE CONTRC
General Description of Molecular Sieves
The term "molecular sieve" was originated by J. W. McBain to
describe fine porous solid materials which have a structure of channels
and voids which permits them to act as sieves on a molecular scale. A
molecular sieve has pores of a very uniform size which are uniquely
determined by the unit structure of the crystal. These pores wij.1
completely exclude molecules which are larger than their diameter.
An important type of molecular sieve is the zeolites, which
are crystalline, hydrated aluminosilicates of the alkali, and alkaline
earth elements. The structure of zeolites consists of a three-dimensional
network of SiO, and A10, tetrahedra linked together by sharing all the
oxygen atoms. There are 34 known species of natural zeolites and about
100 types of synthetic zeolites, although only a few have practical
/0\
significance at the present time. Zeolites must be dehydrated in
order to be used as molecular sieves, and the structure after complete
dehydration must remain intact. Uncontrolled dehydration can collapse
the zeolite structure.
The selectivity of a molecular sieve for a particular adsorbate
depends not only on the dimensions of the interstitial spaces in the
sieve crystal, but also on the polarity of the adsorbate, and, in the
case of hydrocarbons, the extent of carbon-bond saturation of the
adsorbate.
Use of Molecular Sieves for NO Recovery
(9)
In 1955, Addison and Barrerv ' studied the adsorption of
nitric oxide on natural and synthetic zeolites. They found that adsorbed
nitric oxide underwent the following disproportionation reaction:
4 NO^ N20 + N203 . (20)
-------
19
Molecular sieves have shown a high adsorption capacity for nitrogen
dioxide but essentially no capacity for adsorbing nitric oxide by
itself. However, in the presence of oxygen, molecular sieves can
catalyze the oxidation of nitric oxide to nitrogen dioxide which will
subsequently adsorb.
Sundaresan, et al. studied the use of molecular sieves
for the treatment of nitric acid tail gas. The tail gas was simulated
by a nitrogen stream containing 3.5 percent oxygen to which was added
1800 to 2500 ppm of nitric oxide. A synthetic zeolite was used as the
adsorbent. The performance of the adsorbent was tested by cycling it
through periods of adsorption followed by desorption using either hot
air or steam and hot air as a purge. Adsorbed NO was recovered as
j£
enriched NO and nitric acid. The experimental results showed an
initial adsorption capacity of 5 to 6 Ib NO removed per 100 Ib bed
2£
which decreased after 11 cycles to 2 to 3 Ib NO removed per 100 Ib bed.
2£
Several other studies have been carried out on the adsorption
of nitrogen dioxide alone on a variety of molecular sieves. Krasnyy,
et al. examined the effect of SiOj/A^O ratio on the adsorption of
N02. Dealuminated mordenite with an S10 /Al 0_ ratio of 13:1 was found
to have the highest adsorption capacity at low relative pressures of NCL.
More highly dealuminated mordenites and commercial acid-resistant
zeolites had lower capacities. Similar observations were obtained in a
(12)
study by Lewis .
Joithe^ , et al., examined the adsorption of N0/N02 mixtures
from a dry carrier gas. The results showed that molecular sieves have a
high capacity for NO,.. Mixtures of NO and NO^ can also be adsorbed
provided that oxygen is present in the carrier gas to oxidize the NO.
After exposure of the sieves to NO in a carrier gas saturated with
a
water, they can be reactivated without a loss of loading capacity for
adsorbing NO from a dry carrier gas provided the reactivation process
has desorbed all of the water vapor.
The Union Carbide Corporation has been actively working on
the use of molecular sieves for NO removal from nitric acid plant
J\,
tail gas for over five years. Extensive laboratory screening tests
-------
20
were carried out to select the product best suited for the removal
and recovery of NO from process streams. These included laboratory
JL
adsorption selectivity determinations, equilibrium loading determinations,
catalytic activity measurements, mass transfer rate measurements, and
cyclic life testing. This work led to a pilot scale test program and
thereby to the development of the PuraSiv N process to be considered
here.
*
Description of PuraSiv N Process
I
There are two process flow schemes which can be employed
for the PuraSiv N process. The choice between the two is based
primarily on economic considerations rather than the differences
between the flow schemes. In general, the scheme referred to as the
2-bed process is the economically preferred scheme for nitric acid
plants smaller than about 200 tons/day. Above 200 tons/day, the scheme
referred to as the 4-bed process is the economically preferred scheme.
Figure 6 shows the process flow scheme of the 2-bed system.
Tail gas (Stream 1) from the nitric acid plant absorption tower is
chilled from the absorption tower overhead temperature, typically 85 -
95 F, to 45 - 55 F, the exact temperature level being a function of
the NO concentration, in the feed chiller. The water content is
4V
thus reduced to its saturation level at the lower temperature level.
Brine solution from a chilled water unit is the cooling medium for
this exchanger.
The condensate formed in the feed chiller is removed in the
mist eliminator. The condensate reacts with the nitrogen dioxide
present and forms a weak solution of nitric acid which is sent to
storage (Stream 2). Alternately, this weak acid could be recycled
to an intermediate stage in the nitric acid absorption tower. Entrained
water and nitric acid mist are removed from the chilled gas by a wire
mesh pad in the mist eliminator.
* Based on information obtained from Union Carbide Corporation.
-------
Tail gas from
entrainment
separator
Recovered NOX
recycled back to
absorption tower
Feed
chiller
Recycle gas cooler
Mist
eliminator
-JX-
Adsorbent/
catalyst vessels
B
IXh
A
Regeration
gas cooler
and cooler
- . . .
Treated tail gas
to power recovery
Regeneration
compressor
FIGURE 6. TWO-BED PURASIV N PROCESS (VESSEL A UNDER ADSORPTION AND
VESSEL B UNDER REGENERATION HEATING)
-------
22
The overhead from the mist eliminator (Stream 3) flows downward
through the adsorbent/catalyst vessel, A, which is operating on the
adsorption step of its cycle. The tail gas passes downflow through
the compound bed of desiccant apd NO removal grade adsorbent/catalyst
2C
for water and NO removal, respectively. The particle size of the
3t
desiccant and adsorbent/catalyst is selected to provide high mass
transfer rates, assuring efficient bed utilization and minimum adsorbent
inventory. The adsorbent/catalyst beds are sized to minimize pressure
drop through them while maintaining a velocity high enough for good
flow distribution and mass transfer characteristics.
The effluent stream from the adsorbent/catalyst vessel is
split into two streams. About 80 percent of the treated tail gas
(Stream 4) is returned to the nitric acid plant for power recovery.
The remaining 20 percent of the effluent stream (Stream 5) is used to
regenerate the absorbent/catalyst bed in the vessel which is operating
on the regeneration step of its cycle. The regeneration stream represents
an internal recycle stream between the nitric acid plant absorption tower
and the PuraSiv N unit. Therefore, the flow rate of the treated tail gas
sent to power recovery is equal to that of the tail gas before the addition
of the PuraSiv N unit.
Regeneration of the adsorbent/catalyst bed is effected in a
"thermal pulse11 cycle. The regeneration gas (Stream 5) is first
compressed (Stream 6) in the regeneration compressor. The gas is
then heated in the regeneration heater, to 550 F (Stream 7) and passed
upflow through the adsorbent/catalyst bed until approximately 40 percent
of the bed is at the regeneration temperature. Once this heat zone is
established, the gas is directed through the regeneration gas cooler
and chiller, instead of the regeneration heater. Tower water and brine
solution from a chilled water unit are the cooling media for the
regeneration gas cooler and chiller, respectively. The cold regeneration
gas (Stream 8), 50 F, is also directed upflow through the adsorbent/
catalyst bed pushing the heat zone through the bed in a "thermal pulse",
accomplishing N0~ and then H.O desorption. The cool gas simultaneously
-------
23
cools tlic bed behind the heat pulse in preparation for the next adsorption
step. Complete regeneration is assured by selecting time periods for
the heating and cooling steps that do not allow the cooling front to
overtake the heating front before the heating front passes the bed
exit.
The NCL-rich effluent gas stream (Stream 9) is cooled,
using tower water, to a maximum temperature of 95 F in the recycle gas
cooler, and returned to the nitric acid plant upstream of the absorption
tower. Alternately, if the nitric acid plant cooler-condenser were
adequate, the recycle gas cooler could be eliminated and the recycle
regeneration gas returned to the nitric acid plant upstream of the cooler-
condenser. In either event, the nitric acid plant absorption tower must
be able to accommodate the internal recycle stream required for
regeneration. The effect of this increased load on the performance of
the absorption tower is accounted for in the design of the PuraSiv N unit.
At automatically controlled intervals the function of each
adsorbent/catalyst vessel changes to provide continuous and uninterrupted
NO removal and recovery. The following block diagram illustrates a complete
2£
cycle.
Adsorbent/Catalyst Vessel
Percent of Cycle
50
100
Vessel A
Adsorption
Regeneration
Heating
Cooling
Vessel B
Regeneration
Heating
Cooling
Adsorption
-------
24
A generalized material balance for the 2-bed PuraSiv N
process applied to a 100 ton/day nitric acid plant is shown in Table 1.
Figure 7 shows the process flow scheme of the 4-bed system.
The only difference between this flow scheme and that of the 2-bed
system is that the 2-bed system employs two adsorbent/catalyst vessels
with one vessel on adsorption and the other on "thermal pulse" regeneration
in a flow direction countercurrent to adsorption, while the 4-bed system
employs four adsorbent/catalyst vessels with two vessels on adsorption,
staggered by one-half of the time required for an adsorption step, one
vessel on regeneration heating in a direction countercurrent to adsorption
and one vessel on cooling in a direction cocurrent to adsorption. All
statements applicable to the chilling of the tail gas prior to adsorption,
the adsorption step of the cycle, the properties of the desiccant and
NO removal grade adsorbent/catalyst, the source of the regeneration gas
2L
and the recycle of the regeneration gas to the nitric acid plant for the
2-bed system are also applicable to the 4-bed system.
Referring to Figure 7, the tail gas (Stream 1) from the nitric
acid plant absorption tower is chilled in the feed chiller, and passed
through the mist eliminator for the removal of entrained water and acid
mist. The weak solution of nitric acid (Stream 2) collected in the mist
eliminator is sent to storage or recycled to the nitric acid absorption
tower. The overhead from the mist eliminator (Stream 3) flows downward
through either of the two adsorbent/catalyst vessels, A or 8, which are
operating on the adsorption step of their respective cycles.
The effluent streams from the two vessels operating in parallel
are combined in the adsorption effluent manifold and then split into two
streams. About 85 percent of the treated tail gas (Stream 4) is returned
to the nitric acid plant for power recovery, while the remaining 15 percent
of the effluent stream (Stream 5) is used to regenerate the absorbent/
catalyst beds in the vessels which are operating on the cooling and regenera-
tion heating steps of their respective cycles.
-------
TABLE 1. GENERALIZED MATERIAL BALANCE FOR 2-BED PURASIV N PROCESS
APPLIED TO 100 TON/DAY NITRIC ACID PLANT
Stream Number
Temperature, F
Pressure, psig
Composition, Ib/hr
N2
°2
N02
NO
HN03
H20
Total
Moles /Hr
Tail Gas
From
Absorption Weak
Tower Acid
1 2
85 45
92 91
47, 140
1,683
30
178
23
177 133
49,208 156
1752.6
Absorber
Feed
3
45
90
47,140
1,632
143
93
44
49,052
1743.2
Treated
Tail Gas
4
70
88
38,225
1,283
3
39,511
1405.4
Compressed
Recycle Recycle
Purge Gas Purge Gas
5
70
88.5
8,915
299
1
9,215
327.7
6
135
108
8,915
299
1
9,215
327.7
Heated
Purge Gas
7
500
106.5
8,915
299
1
9,215
327.7
Cooled
Purge Gas
8
50
106.5
8,915
299
1
9,215
327.7
Recycle
Gas to
Acid Plant
9
60-95
105
Ui
8,915
299
1
44
9,541
366.3
(a) Average composition.
-------
Toil gas
from
entrainment
separator
Recycle gas cooler
Feed
chiller,
Recovered NOX
recycled back to
absorption tower
Adsorbent/catalyst vessels
Mist
eliminator
Wear
acid
B
V '' v
'if
\ S «
FIGURE 7.
FOUR-BED PURASIV N PROCESS (VESSELS A AND B UNDER
ADSORPTION, VESSEL C UNDER REGENERATION COOLING,
AND VESSEL D UNDER REGENERATION HEATING)
Regeneration
gas cooler
and chiller
Regenera-
tion heater
Regeneration
compressor
Treated tail gas
to power recovery
-------
27
The regeneration gas (Stream 5) is compressed in the
regeneration compressor, and then cooled to 50 F in the regeneration gas
cooler and chiller. The cool regeneration gas (Stream 8) is employed first
to cool the adsorbent/catalyst bed which is hot from a previous regeneration
heating step. As the cooling gas passes downflow through this hot bed,
it is heated as the bed cools. This heated regeneration gas then passes
through the regeneration heater, where make-up heat is provided. This
gas (Stream 7), now at 550 F, is used to regenerate the adsorbent/catalyst
bed loaded with N02 from a previous adsorption step. The hot gas, passing
upflow through the adsorbent/catalyst vessel, transfers heat to the molecular
sieve, accomplishing thermal desorption of the N02 and H20 from the bed.
Complete regeneration is assured by selecting an adequate heating-cooling
period and regeneration gas rate to allow the heating front to pass through
the bed exit prior to the end of this step of the cycle. The N02-rich
effluent gas stream (Stream 9) is then cooled to a maximum temperature of
95 F in the recycle gas cooler, and returned to the nitric acid plant
absorption tower.
At automatically controlled intervals the function of each
adsorbent/catalyst vessel changes, so that two vessels, staggered by one-
half of the time required for an adsorption step, are on adsorption, one
vessel is on cooling, and one vessel is on heating at all times. The following
block diagram illustrates a complete cycle.
Adsorbent/Catalyst Vessel
Percent of Cycle
25 50
75
100
Vessel A
Adsorption
Heating
Cooling
Vessel B
Cooling
Adsorption
Heating
Vessel c
Heating
Cooling
Adsorption
Vessel D
Adsorption
Heating
Cooling
Adsorption
-------
28
\\
A generalized material balance for the 4-bed PuraSiv N process
applied to a 300 ton/day nitric acid plant is presented in Table 2.
Performance of the PuraSiv N Process
The first commercial PuraSiv N unit was started up in May, 1974a
on a 55 ton/day nitric acid plant of Hercules, Inc. in Bessemer, Alabama.
The unit is designed to remove and recover up to 3300 ppm NO . The tail
3£
gas contains 3 percent oxygen and is water saturated at temperature and
pressure. The maximum design tail gas flow rate for the unit is 5,000 scfm.
The design basis for the unit is given in Table 3.
A second PuraSiv N unit was started up in August, 1974, also on a
55 ton/day nitric acid plant at the U. S. Army's Holston Defense Corporation
in Kingston, Tennessee. This unit is designed to remove and recover up to
3,000 ppm NO from tail gas containing 2.8 percent oxygen and saturated with
2£
water at temperature and pressure. The maximum design tail gas flow rate
for this unit is 4,700 scfm. The design basis for this unit is also given
in Table 3.
Both commercial PuraSiv N units use treated tail gas for the regeneration
of the NO adsorbers. The units contain two NO adsorbers (2-bed process),
x x
operating on a four hour adsorption-four hour regeneration cycle. Both units
are skid mounted and were shop fabricated and assembled. The battery limits
2
plot size for each unit is about 400 ft .
Tests by Union Carbide Corporation
Union Carbide made a three-day performance run at the two commercial
installations of the PuraSiv N process immediately after start up. The
performance records for the Hercules and Holston units are shown in Tables 4
-------
TABLE 2. GENERALIZED MATERIAL BALANCE FOR 4-BED PURASIV N PROCESS
APPLIED TO 300 TON/DAY NITRIC ACID PLANT
Tail Gas
From
Adsorption Weak
Tower Acid
Stream Number
Temperature, F
Pressure, psig
Composition, Ib/hr
N2
°2
N02
NO
HN03
Total
Moles/Hr
1 2
85 45
92 91
125,928
4,493
75
442
62
471 355
131,409 417
4680.4
Absorber
Feed
3
45
90
125,928
4,363
354
231
116
130,992
4655.6
Treated
Tail Gas
4
70
88
106,283
3,578
9
109,870
3907.8
Compressed
Recycle Recycle
Purge Gas Purge Gas
5 6
70 135
88.5 111
19,645 19,645
662 662
2 2
20,309 20,309
722.3 722.3
Heated
Purge Gas
7
500
107
19,645
662
2
20,309
722.3
Cooled
Purge Gas
8
50
109.5
19,645
662
2
20,309
722.3
Recycle
Gas to
Acid Plant
9
60-95
105
19,645
662
2
21,116
743.9
(a) Average composition.
-------
30
TABLE 3. DESIGN BASIS FOR HERCULES AND
HOLSTON PURASIV N UNITS
Hercules
Unit
HoIs ton
Unit
Battery Limits Tail Gas Feed Conditions:
Flow rate (60 F & 14.7 psia), maximum
Temperature, maximum
Pressure, minimum
Composition
°2
NO , maximum
X
HN0
Others
5,000 scfm
90 F
90 psig
4,700 scfm
100 F
92 psig
3.0 percent 2.8 percent
3,300 ppm 3,000 ppm
Saturated at temp. & pressure
Saturated at temp. & pressure
Balance Balance
None None
-------
31
TABLE 4. PERFORMANCE OF HERCULES PURASIV N UNIT
DURING THREE DAY PERFORMANCE RUN AFTER
START UP
May 26 May 27 May 28
NO in Effluent
Ji
Average, ppm 225
Range, ppm 0-6 0-7 0-25
NO in Feed
ji
Average, ppm 2,600 2,400 2,450
Range, ppm 2,000-3,000 2,300-2,500 2,300-2,500
Average Gas Flow
Tail gas flow, scfm 4,850 4,650 4,600
Recycle gas flow, 950 950 950
scfm
Total gas flow, scfm 5,800 5,600 5,550
TABLE 5. PERFORMANCE OF U.S. ARMY - HOLSTON PURASIV N
UNIT DURING THREE DAY PERFORMANCE RUN AFTER
START UP
August 17 August 18 August 19
NO in Effluent
X,
Average, ppm < 1 6 7
Range, ppm 0-2 0-30 0-30
in Feed
Average, ppm 4,100 3,700 3,900
Range, ppm 3,000-5,700 3,500-4,400 2,500-4,700
Average Gas Flow
Tail gas flow, scfm 4,400 4,400 4,400
Recycle gas flow, 1,100 1,100 1,100
scfm
Total gas flow, scfm 5,500 5,500 5,500
-------
32
and 5, respectively. More than 99 percent of the NO present in the tail gas
A
stream was recovered in both units. In each case, the on-stream NO adsorber
is not switched to regeneration until breakthrough of NO in the effluent
has occurred. The Hercules unit provided an average daily effluent NO
concentration of 2 to 5 ppm (0.08 Ib NO /ton acid) under design conditions
of flow and feed concentration. The Holston unit provided an average daily
effluent NO concentration of 1 to 7 ppm (0.1 Ib NO /ton acid) under NO
feed concentrations 35 percent higher than design.
Union Carbide performed a field evaluation of the Hercules unit
0.4)
during the first two weeks of June following start up. After about 40
commercial cycles, the PuraSiv N adsorbers were performing well within
design specifications. NO effluent concentration traces under fairly
X
typical operating conditions are shown in Figure 8. While treating about
6,000 scfm of chilled tail gas containing about 2,600 ppm NO at 50 F and
50 psig, the adsorbers provided 4 ppm NO after four hours of adsorption
time. The design guarantees an average of 50 ppm NO during a four hour
adsorption cycle. One percent breakthrough occurred slightly earlier in
Cycle No. 43 A than in No. 39 B (460 versus about 510 minutes). Based on
the average feed conditions, the reverse would have been expected. This
difference is attributed to fluctuations in feed NO concentrations and
x
inherent differences in the two beds. When a malfunctioning valve caused
hot regeneration gas to enter the adsorbing bed, higher NO effluent
concentrations (up to 45 ppm) were observed after four hours adsorption
time.
Tests by U. S. Army
The U. S. Army ran several three-day performance tests on the
Holston PuraSiv N unit during the period August, 1974 to January, 1975 .
The results of these tests are summarized in Table 6. The test in August,
1974 coincides with the test performed by Union Carbide and reported in
-------
33
30
Q.
Q.
o 20
c
0)
o
o
o
X
o
0)
3
UJ
10
o
Average Feed Conditions
Cycle
43A
Flow Rate, scfm 6300
NO Cone., ppm 2400
Temperature, F 50
Pressure, psig 89
Cycle
39 B
5900
2900
48
91
IOC 200 300
Adsorption Time, minutes
400
500
FIGURE 8. PERFORMANCE OF HERCULES ADSORPTION BEDS AFTER ABOUT 40 CYCLES
-------
34
TABLE 6. THREE DAY PERFORMANCE TESTS OF THE
HOLSTON PURASIV N UNIT BY THE U.S. ARMY
Time Period
August, 1974
October, 1974
November, 1974
December, 1974
January, 1975
Average
Inlet NO
Concentration,
ppm
4,090
3,429
2,760
1,324
1,833
Average
Outlet NOX
Concentration,
ppm
17.1
59.0
33.2
14.2
18.4
Average
NOX Removal,
' percent
99.6
98.3
98.8
98.9
99.0
-------
35
Table 5. The discrepancy in average outlet NO concentration can be
a
attributed to different methods of analysis. The Union Carbide data is
based on "Draeger" tube analyses for NO which were taken every two hours.
X
The U. S. Army data is based on continuous monitoring of NO concentrations
X
using a chemiluminescent analyzer for NO and a photometric ultraviolet
analyzer for N0_. However, the three-day average NO emission of 17.8 ppm
£* X
was well within the required acceptance specifications that NO emissions
X;
average under 50 ppm over an eight hour period. A statistical analysis
of the data resulted in a projected probability that the outlet NO
X.
concentration would exceed an average of 50 ppm over an 8-hour period
less than one percent of the time.
In the three-day performance run in October, 1974, the NO
X,
emissions averaged 59.0 ppm which slightly exceeded the purchase specification
of 50 ppm over an eight-hour period. The abnormal sieve operation was
attributed to improper adsorption in adsorber vessel B. The three-day
performance runs in November and December, 1974, and in January, 1975,
yielded average NO emissions below the 50 ppm specification. However,
Ji
the run in November was terminated after 17 hours because of a maintenance
shut down of the acid plant. During the run in January, adsorber vessel
B continued to exhibit high NO emissions with respect to vessel A. The
X,
tail gas flow rates during the performance runs are not known at the present
time.
An inspection of Table 6 indicates that the inlet NO concentration
X
decreases as ambient temperature decreases (perhaps with the exception of
the run in January). Figure 9 presents the variation of inlet NO concentration
X^
with temperatures of product acid, injection water, and filtered water.
Filtered water cools the absorption tower trays and then is used in the
cascade cooler upstream from the tower. At the present time, it is not
known whether the molecular sieve continued to meet performance specifications
as the NO loading increased with ambient temperature.
The instrumental NO analyses were compared with EPA Method 7
X.
analyses by Midwest Research Institute. Instrumental inlet NO analyses
X
were found to agree within 6.5 percent of Method 7» However, instrumental
-------
36
5000
4000
o.
ex
§ 3000
2
I
O 2000
X
o
.2
£ 1000
o Filtered water
a Product acid
A Injection water
32
0
50
68
86
104
10 20 30
Temperature, C and F
40
FIGURE 9.
INLET NO CONCENTRATION TO ADSORPTION BEDS AS A FUNCTION OF TEMPERATURE
x
-------
37
outlet analyses were found to agree only within +56.1 percent of Method 7
at about 10 ppm NO and -54.8 percent of Method 7 at about 280 ppm NO .
X X
The outlet analyzer apparent error has been attributed to one or more of the
following:
(1) High noise levels and limiting accuracy
(2) Difference between cell pressure and calibration pressure
(3) Presence of nitric acid mist.
Tests by Engineering-Science. Inc.
Engineering-Science, Inc., was given a contract by the EPA Control
Systems Laboratory for the purpose of testing the Hercules PuraSiv N unit.
Simultaneous measurement of NO concentrations in the PuraSiv N inlet and
x
outlet streams were performed during 11 individual four-hour adsorption
cycles in the period March 5 through 14, 1975, when the molecular sieve was
about nine months old, using continuous photometric analyzers. The continu-
ous measurements were supplemented by NO determinations using grab samples
2£
taken and analyzed by EPA Method 7. The complete results of this test pro-
gram are presented in a draft report prepared by Chehaske and Greenberg.
Two different inlet sampling sites were used during the tests.
During the first eight cycles of testing, the inlet sampling was conducted
between the mist eliminator and the two adsorption beds. During the final
three cycles of testing, the inlet sampling was conducted upstream of the
feed chiller and mist eliminator. The latter sampling site was addled at
the request of Union Carbide under the rationale that because the feed
chiller and mist eliminator are part of the packaged system, a more represen-
tative system evaluation would be obtained if samples were taken upstream
from the feed chiller. However, acid mist was not measured because
instrumentation for continuous monitoring was not available and isokinetic
grab sampling would be very difficult because of the high pressure of the
tail gas stream.
The results of the NO emission tests are summarized in Table 7.
Ji.
The average flow rates and average inlet and outlet NO concentrations were
2i
calculated by numerical integration of the data tabulated by Engineering-Science.
-------
TABLE 7. PERFORMANCE TESTS OF THE HERCULES PURASIV N UNIT BY ENGINEERING-SCIENCE
Run
No.
1
2
3
4
5
6
7
8
9
10
11
Average
Tail Gas
4854
4633
4447
4541
4549
3815
3392
4014
4016
5040
4449
Gas Flow Rate, serin C
Recycle Gas
1025
1055
1035
1010
1020
1025
1020
1020
1040
1025
1020
Total
5879
5688
5482
5551
5569
4840
4412
5034
5056
6065
5469
Average
Inlet NOX
oncentration,
ppm
2541
3076
2799
3133
2258
1006
2332
2325
2937
3120
3114
Average Total Mass
Outlet NOX Average Total Inlet of NOX
Concentration, NOX Removal, Adsorption Mass of NOX, Absorbed,
ppm percent Bed grams grains
113
154
140
125
99.0
16.4
62.6
73.2
91.7
137
139
95.6
95.0
95.0
96.0
95.6
98.4
97.3
96.9
96.9
95.6
95.5
B
B
A
B
A
A
B
A
B
A
A
172,780
214,240
198,370
221,610
157,200
59,280
128,150
145,690
220,100
235,580
212,880
165,190
203,540
188,990
212,740
150,330
58,300
124,680
141,110
213,000
225,250
203,410
00
-------
39
The tail gas flow rate was calculated by subtracting the recycle gas flow
rate from the inlet (total) gas flow rate to the sieve. Battelle calculated
the average NO removal as follows:
.. wr, 11 nn A average outlet ppm NOy "\
percent NO removal = 100 (1 a—7—; " *• A
r x \ average inlet ppm
This value corresponds to what is termed the total percent reduction due
to adsorption by Engineering-Science and can also be calculated by dividing
the total mass of NO adsorbed by the total inlet mass of NO . The inclusion
X X
of a percent reduction of NO due to regeneration does not appear to be
2i
relevant.
A plot of average NOX removal versus total mass loading on the
sieve for runs 1 through 11 is shown in Figure 10. The graph reveals that
the average removal efficiency and total mass loading follow an inverse
relationship. The graph also shows that the average removal efficiences
based on inlet concentrations measured upstream of the feed chiller are
only slightly higher than efficiencies at similar mass loadings based on
concentrations measured after the mist eliminator.
During all the cycles tested the instantaneous control efficiency
reached a maximum value within the first hour of the cycle and then gradually
decreased for 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.
The outlet NO concentrations listed in Table 7 show that the unit
Ji
did not meet the design guarantee of an average of 50 ppm NO during a four-
Ji
hour adsorption cycle. Before the Engineering-Science tests were performed,
the PuraSiv N unit experienced an abnormal shut down which caused a back
flow of nitric acid into the sieve. Although a fail-safe system has
now been installed, the sieve was probably damaged by the acid. However,
it is significant to note that the unit still met the new source performance
standard which is equivalent to about 200 ppm NO in the effluent.
Ji
In comparing the photometric NO analyses with EPA Method 7, the
2v
mean percent difference was -I- 8.2 percent at the inlet site after the mist
eliminator and -1.9 percent at the inlet site before the feed chiller.
The mean percent difference at the outlet test site was 10.7 percent,
-------
40
lOOr
99
98
97
c
Q>
O
O
O
E
0>
cr
X
O
Q>
O»
O
0>
5
96
95
94
93
92
91
90
D
D
o n
O D
O Adsorber A
D Adsorber B
Solid points indicate inlet sampling site
located before feed chiller and
mist eliminator
50 100 150 200 250
Total NOX Mass Loading on Sieve, kg
FIGURE 10. TOTAL NO MASS LOADING yEBSUS AVERAGE NO REMOVAL
X X
-------
41
but the range of difference was larger at the outlet than the inlet,
presumably because of greater relative errors in Method 7 at low concentrations,
It is difficult to explain why there is better agreement between the two
analytical methods at the test site before the feed chiller rather than
after the mist eliminator. Method 7 is sensitive to nitric acid mist and
vapor present in the gas while the photometric analyzer does not detect
these species. More nitric acid mist would be expected ahead of the feed
chiller than after the mist eliminator.
Sieve Life
The performance of the molecular sieve adsorbent used to control
NO in tail gas from nitric acid plants is dependent upon the number of
2C
adsorption/regeneration cycles to which it has been subjected. The PuraSiv
N process design is such that a maximum average effluent NO concentration
2£
of 50 ppm is guaranteed for a period of two years. With the passage of time,
the average concentration of NO in the effluent is expected to gradually
3v
increase from a level which is well below that guaranteed to one which is
close to it. This loss in performance is caused by the gradual but cumulative
destruction of the micropores of the adsorbent as a result of exposure to
the temperature swings between 45 and 500 F for adsorption and regeneration,
respectively, as well as exposure to very low concentrations of nitric
acid mist and vapor present in the tail gas. The effect of nitric acid on
the sieve is demonstrated in the extreme case by the abnormal shut down
that occurred at the Hercules unit prior to the performance tests by
Engineering-Science.^ ' '' The only definitive data available to Battelle
for determining sieve life are the tests by the U. S. Army on the Holston
unit. These data indicate no significant degradation in sieve performance
after six months of operation. However, one adsorber bed has continually
shown better performance than the other bed. Therefore, at the present time
only a six-month sieve life has been demonstrated. However, there appears
to be no evidence to negate Union Carbide's claim of a two-year sieve life.
Union Carbide has tested the PuraSiv N adsorbent through 1,000
adsorption/regeneration cycles in the laboratory. The cycle consisted of
1/2 hour adsorption and 1/2 hour regeneration. At the end of 1,000 cycles,
the adsorption capacity had decreased by only 5 percent. One thousand
-------
42
cycles corresponds to a three-year life for the field units, but the cycle
time is 8 hours rather than 1 hour. Also, it is not known whether any
nitric acid was present in the adsorbate.
ECONOMIC EVALUATION OF MOLECULAR SIEVE CONTROL TECHNOLOGY
In this section the economic feasibility of the PuraSiv N process
for emission control at nitric acid plants will be assessed by comparison
with two of the other control processes described in a previous section of
this report. The processes selected for comparison purposes are nonselective
catalytic reduction with natural gas and extended absorption. Although
catalytic reduction is employed at about two-thirds of the nitric acid plants
practicing full abatement of NO emissions, the current shortage of natural
m
gas may curtail further usage of this process. However, the alternative of
urea scrubbing also is dependent on natural gas supply to the extent that
ammonia required for urea production is produced from natural gas. The
economic comparison of the three processes (molecular sieve adsorption,
nonselective catalytic reduction with natural gas, and extended absorption)
will be based on the concept of capitalized costs.
Basis and Procedures
Investments
The investments (capital costs) for the control alternatives were
obtained from data on units which have been built and other data provided
by system designers, vendors, and operators. The costs reported are mid-1974
estimates as updated by the Marshall and Stevens Index (390 in mid-1974).
All calculations were made for two acid plant sizes, 100 and 1000 tons/day
of 100 percent HNO_. For convenience in further calculations the investments
have been expressed by an equation of the form
Investment = A(tons/day HN03)B (21)
-------
43
where A and B are constants.
The investments include not only on-site facilities but also off-
site facilities associated with utilities. The off-site investments are
based on the following unit costs.
Electricity $360/kw
Cooling water $27.50/gpm
These values were determined in connection with a recent Battelle study.
The off-site investments are an allocated cost applied on a uniform basis
for the three processes under consideration. Whether or not existing
facilities can handle the utilities reuqirements is extremely site specific.
Operating and Maintenance Costs
The operating and maintenance (0 & M) costs were calculated on
an annual basis and include the following
• Operating labor
• Utilities (electricity, cooling water, steam, and fuel)
• Raw materials and chemicals
• Maintenance labor and materials
• Local property taxes and insurance
• By-product credits-..
Table 8 lists the unit costs used in these calculations.
Again, the calculations were made for the two acid plant sizes.
For convenience in further calculations the operating and maintenance costs
have been expressed by an equation of the form
0 & M Cost = C(tons/day HN03>B + D(tons/day HN03)E (22)
where B, C, D, and E are constants. The first term here is the investment-
related 0 & M costs, which includes maintenance and local taxes and insurance,
Since this is taken as a percentage of the investment, the constant B is the
same as the investment equation. The second term above includes all the
other 0 & M costs.
-------
44
TABLE 8. UNIT COSTS USED IN CALCULATING
OPERATING AND MAINTENANCE COSTS
Item Unit Cost
Electricity 1.2c/kWh
Cooling Water 3c/1000 gal
Process water 15c/1000 gal
Steam $1.2/1000 Ib
Fuel oil $12/bbl
Natural gas $1.2/1000 ft3
HN03 credit $25/ton
-------
45
Capitalized Costs
The economic comparison of the control alternatives was done
through the concept of capitalized costs. The basic principle underlying
this method of economic evaluation of alternatives is the indefinite and
perpetual replacement of a given capital investment which has a finite life.
The capitalized cost is defined as the original investment plus the present
value of the perpetuity which will permit the perpetual replacement of the
(19)
item. For an item with an initial cost of I and a salvage value of S
after n years the capitalized cost K (in present dollars) is
K
where i is the fractional annual interest rate or, more precisely, the rate
of return on investment which the investor requires. This is usually of
the order of 15 percent. In this study, the capitalized cost associated
with the investment, which will be designated K_, was calculated from
Equation 23 with the assumption of zero salvage value. That is,
. i (i + i)° . (24)
1 (i + i) - i
The concept of capitalized costs can be extended to include
annual and other periodic costs in addition to capital investments. An
expression for the capitalized cost associated with the annual 0 & M costs
can be obtained from Equation 23 by treating this like an investment with a
life of 1 year but discounting to present dollars, since these are end-of-
year costs. Thus,
_ 0 &M
K0 & M " i ' (25)
The total capitalized cost is the sum of the term for the
initial investment and that for the 0 & M costs.
K
tot - KI + K0 & M <26>
In this study a 15 percent rate of return and a 12-year equipment life were
used.
-------
46
Investment and Operating Costs
Molecular Sieve Adsorption
The on-site investment and operating costs for the PuraSiv N
process were based on data obtained from Union Carbide in February, 1975.
The costs data were given as a function of tail gas flow rate. Based on
data previously presented in this report, the tail gas flow rate was taken
as 78 scfm/tpd. The stated accuracy for the investment costs was + 25 percent.
The data included the annual cost for the adsorbent contract and the requirements
for electrical power, heat, and cooling water. Utilities requirements were
verified by field data obtained by the U. S. Army at the Holston unit
and by Engineering-Science at the Hercules unit.
As a basis for comparison, investments were sought for the two
PuraSiv N units which have been built. For the 55 ton/day Holston unit the
1974 investment was $289,000. This was for a skid-mounted unit and did not
include foundation work and tie-ins. This cost agrees quite well with the
cost based on the Union Carbide data ($294,000 from the cost equation presented
in Table 12). Hercules would not give the investment for their unit but
said that Union Carbide gave them an unusually low price because it was the
first-of-a-kind unit.
The costs were calculated on the basis of 2500 ppm NO in the tail
Jv
gas and a maximum of 50 ppm NO in the effluent. The operating labor
x (20)
requirements for the molecular sieve process are known to be very low
and were taken as $l,000/yr for a 100 ton/day acid plant and $3,000/yr for
a 1,000 ton/day acid plant. Investment and 0 & M costs for the PuraSiv N
process on the two plant sizes are presented in Table 9.
Catalytic Reduction
Battelle personnel developed an equation expressing the investment
for the catalytic reduction process as a function of plant size for an
(21)
existing nitric acid plant as part of the Cost of Clean Air study.
The equation updated to mid-1974 is
,0.6
Investment ($) = 13,000 (tons/day 100 percent
-------
47
TABLE 9. COSTS OF PURASIV N PROCESS
(8000 hr/yr operating time)
Plant Capacity
Item
100 tons/day
1000 tons/day
3 3
Tail gas flow rate, 10 ft /min
at 60 F and 1 atm
Investment, $
Battery limits including
license fee
Off sites
Electricity
Cooling water
Site preparation
TOTAL
Operating and Maintenance Cost, $
Operating labor
Molecular sieve contract
Electricity
Cooling water
Fuel oil
Subtotal
Maintenance (2.5 percent of
inves tment/yr)
Taxes and insurance (2.5 percent of
investment/yr)
TOTAL GROSS
Credit for additional HNO., produced
TOTAL NET
Net Operating and Maintenance Cost
in $/ton
7.8
440,000
57,000
7,500
1.000
505,500
1,000
23,800
13,900
3,900
9.500
52,100
12,600
12,600
77,300
(19,000)
58,300
1.75
78.0
3,500,000
541,000
69,700
10.000
4,120,700
3,000
179,000
131,000
36,500
85.500
435,000
103,000
103,000
641,000
(190,000)
451,000
1.35
-------
48
The utility and catalyst requirements for the catalytic reduction process
/21 22)
were based on the following values: v '
Fuel (natural gas) — 2.10 x 10 Btu/ton acid
3
Steam (credit) — 1.80 x 10 Ib/ton acid
Catalyst (2-year life) -- $0.175/ton acid.
These requirements were taken as directly proportional to the acid plant
capacity. The operating labor requirements are minimal and were taken
as $l,000/yr for a 100 ton/day plant and $3,000/yr for a 1,000 ton/day plant
as in the molecular sieve process. Maintenance cost was given by Searles^ '
as 4 percent of investment per year. Investment and 0 & M costs for the
catalytic reduction process on existing plants of the two sizes are presented
in Table 10 for the case of an NO emission level below 200 ppm.
Extended Absorption
The investments for the extended absorption process were based
on information supplied by J. F. Pritchard and Company, Kansas City,
(03)
Missouri, the U. S. licensee of the Grande Paroisse process. Operating
(o)
and maintenance costs were based on values given by Boudreaux. Table 11
shows the investments and 0 & M costs for the extended absorption process
on existing 100 and 1,000 ton/day nitric acid plants for the case of an
NO emission level below 200 ppm.
X
Cost Equations
Table 12 presents the equations for total investment and net 0 & M
costs as functions of plant capacity for existing nitric acid plants for the
molecular sieve, catalytic reduction, and extended absorption processes.
The level of control is 50 ppm NO in the effluent gas for the molecular
2£
sieve process, and 200 ppm NO for the latter two processes. The cost savings
Ji
in the molecular sieve process in going from 50 to 200 ppm NO in the effluent
X,
is small and shows up as a reduced cost for the molecular sieve contract.
This cost savings amounts to less than 8 percent of the total gross 0 & M
costs for the molecular sieve process.
-------
49
TABLE 10. COSTS OF CATALYTIC REDUCTION PROCESS
(8000 hr/yr operating time)
Plant Capacity
Item 100 tons/day 1000 tons/day
Investment, $
Battery limits 206,000 820,000
Site preparation 1,000 10,000
TOTAL 207,000 830,000
Operating and Maintenance Cost, $
Operating labor 1,000 3,000
Catalyst 5,800 58,000
Natural gas 84,000 840,000
Steam (72.000) (720.000)
Subtotal 18,800 181,000
Maintenance (4.0 percent of 8,300 33,200
investment/yr)
Taxes and insurance (2.5 percent 5,200 20,800
of investment/yr)
TOTAL 32,300 235,000
Net Operating and Maintenance Cost 0.97 0.70
in $/ton
-------
50
TABLE 11. COSTS OF EXTENDED ABSORPTION PROCESS
(8000 hr/yr operating time)
Plant Capacity
Item 100 tons/day 1000 tons/day
Investment, $
Battery limits 1,200,000 6,000,000
Off sites
Electricity 10,800 108,000
Cooling water 1,100 11,000
Site preparation 2,000 20,000
TOTAL 1,213,900 6,139,000
Operating and Maintenance Cost, $
Operating labor 1,000 3,000
Electricity 2,900 29,000
Cooling water 600 6^000
Subtotal 4,500 38,000
Maintenance (2.5 percent of 30,300 153,000
investment/yr)
Taxes and insurance (2.5 percent 30.300 153,000
of investment/yr)
TOTAL GROSS 65,100 344,000
Credit for additional HN03 produced (14,100) (141,000)
TOTAL NET 51,000 203,000
Net Operating and Maintenance Cost 1.53 0.61
in $/ton
-------
51
TABLE 12. EQUATIONS FOR INVESTMENT AND OPERATING
COSTS FOR NOX EMISSION CONTROL PROCESSES
AT EXISTING NITRIC ACID PLANTS
Process
Total Investment,
$(a)
Net Operating and Maintenance
Cost, $/year
Molecular sieve
(PuraSiv N)
Catalytic reduction
(nonselective
using CH, )
Extended absorption
(Grande Paroisse)
7.670C
0.91(b)
13.100C
0.60
48,600C
0.70
384C0'91 + 602C0'87
852C°-60+206C°-98
2.430C0'70 - 83.7C1'03
(a) Includes off sites.
(b) C equals plant capacity in short tons per day of 100 percent acid.
-------
52
Capitalized Costs
Equations 24, 25, and 26 were used to calculate the capitalized
costs. The investment and 0 & M costs for the three emission control
processes under consideration were calculated for four plant sizes (100,
400, 700, and 1,000 tons/day) using the equations listed in Table 12.
These costs were used in Equations 24 and 25 along with an "interest"
rate, i, of 15 percent and an equipment life, n, of 12 years. Figure 11
shows the capitalized costs for the emission control processes applied to
existing nitric acid plants as a function of plant capacity.
The molecular sieve process is intermediate in cost between
catalytic reduction and extended absorption. While the catalytic reduction
process is the least expensive, the future application of this process is
in doubt because of the curtailment in natural gas supply. The cost of
the extended absorption process is very dependent upon specific plant
conditions, so that in some cases, especially for large plant sizes, extended
absorption may be competitive with the molecular sieve process. It appears
that the economic feasibility of the molecular sieve process depends on the
continued shortage of natural gas. The capitalized costs of the two
processes become equal when the price of natural gas reaches $2.4/1000 ft .
-------
53
lO.O
«*
o
o
O
T3
0)
N
a
o
o
1.0
O.I
100
J I
1000
Nitric Acid Plant Capacity, tons/day
FIGURE 11. CAPITALIZED COSTS FOR NO EMISSION CONTROL AT NITRIC ACID PLANTS
A
-------
54
CONCLUSIONS AND RECOMMENDATIONS
This analysis of the applicability of molecular sieve technology
to the control of NO emissions from nitric acid plants leads totitie
following conclusions:
• Molecular sieve technology can control the NO emissions
to well within the New Source Performance Stahdard for
nitric acid plants (1.5 kg NO /metric ton acid or 3 Ib
NO /ton acid). Union CarbideXfield test data taken at both
the Hercules and Holston units shortly after startup
indicated that the units were limiting emissions to no more
than 0.1 Ib NO /ton acid. Continuous monitoring data taken
by the U. S. Army at the Holston unit for three days out of
each month during the first six months of operation indicate
NO emissions ranged from about 0.2 to 0.6 Ib NO /ton acid.
The data taken by Engineering-Science, Inc. at tfte Hercules
unit 10 months after start up show an average emission
range of 0.24 to 2.3 Ib NO /ton acid, even after the sieve
had been damaged by a back flow of nitric acid.
9 The NO emissions from a molecular sieve control unit
increase with increasing NO mass loading on the sieve.
Inlet NO concentration to che sieve decreases as the
ambient temperature decreases.
0 The assumed 2-year sieve life with NO control at a guaranteed
level of less than 50 ppm (0.75 Ib N0x/ton acid) has not been
demonstrated in the field, although laboratory tests indicate
that this sieve life can be achieved. The Holston unit will
not achieve 2 years of operation until August, 1976, and the
Hercules unit has to have the damaged sieve replaced so that
it will take 2 more years for this unit to demonstrate
acceptable sieve life. However, a six-month sieve life has
been demonstrated on the Holston unit.
• Molecular sieve technology appears to be less costly than
extended absorption but more costly than catalytic reduction.
The former process is capable of reducing effluent NO con-
centrations to less than 50 ppm, while the latter two
processes usually reduce the effluent NO concentrations to
only 200 ppm. The shortage of natural gas probably means
that molecular sieve adsorption is an economically competitive
emission control technology. When the price of natural gas
reaches $2.4/1000 ft^, the cost of catalytic reduction equals
the cost of molecular sieve adsorption. Since the cost of
the extended absorption process is very dependent upon
individual acid plant characteristics, each plant will have
to be evaluated separately.
-------
55
It is recommended that the 2-year sieve life be demonstrated by
continuing to monitor the NO emissions at the Holston unit periodically
2x
until 2 years of operation have elapsed. Also, the cause of improper
adsorption in adsorber vessel B at the Holston unit should be determined.
-------
56
REFERENCES
1. NATO Committee on the Challenge of Modern Society, "Air Pollution:
Control Techniques for Nitrogen Oxide Emissions From Stationary
Sources", PB-240-578, October, 1973.
2. U. S. Public Health Service, "Atmospheric Emissions from Nitric Acid
Manufacturing Processes", PHS Publication 999-AP-27, 1966.
3. Boudreaux, M. L. , "Additional Absorption Capacity Stems Nitric Acid
Plant NOx Pollution11, Nitrogen, (91), 36-38 (September /October, 1974).
4. Sherwood, T. K. and Pigford, R. L. , Absorption and Extraction. Second
Edition, McGraw-Hill, New York (1952).
5. Streight, H. R. L. , "Reduction of Oxides of Nitrogen in Vent Gases",
Can. J. Chem. Eng., 36_ (1), 3-11 (February, 1958).
6. Ringbakken, R. , Lie, 0. H., and Mejdell, G. T., "Urea As An Agent in
the Destruction/Recovery of NOx *n Nitric Acid and Nitrophosphate
Production", paper presented at ISMA Technical Conference, Prague
(September, 1974).
7. Anonymous, "NOx Abatement in Nitric Acid and Nitrophosphate Plants",
Nitrogen, (93), 32-37 (January /February, 1975).
8. Breck, D. W., Zeolite Molecular Sieves, John Wiley & Sons, New York
(1974).
9. Addison, W. E. and Barrer, R. M. , "Sorption and Reactivity of Nitrous
Oxide and Nitric Oxide in Crystalline and Amorphous Siliceous Sorbents",
J. Chem. Soc., 1955. 757-769 (1955).
10. Sundaresan, B. B., Harding, C. 1., May, F. P., and Hendrickson, E. R. ,
"Adsorption of Nitrogen Oxides from Waste Gas", Environ. Sci. Technol.,
I (2), 151-156 (February, 1967).
11. Krasnyy, E. B., Musin, T. G., Piguzova, L. I., and Prokof'eva, E. N. ,
"Nitrogen Oxide Adsorption Capabilities of Zeolites with High Silica
Contents", Sov. Chem. Ind., (9), 52-54 (September, 1969).
12. Lewis, L. C. , "Evaluation of Adsorbents for Purification of Noble
Gases in Dissolver Off -Gas", U. S. Atomic Energy Commission Report
IN- 1402 (1970).
13. Joithe, W., Bell, A. T., and Lynn, S., "Removal and Recovery of
from Nitric Acid Plant Tail Gas by Adsorption on Molecular Sieves",
Ind. Eng. Chem. Process Des. Develop., j.1. (3) 434-439 (1972).
-------
57
REFERENCES
(Continued)
14. Holcombe, T. C., "Field Evaluation of Hercules PuraSiv N. Unit",
PE Memo #39, Union Carbide Corporation, Tarrytown Technical Center,
Tarrytown, N. Y. (October 25, 1974).
15. Forsten, I., Letter to E. Wooldridge, EPA Control Systems Laboratory
from Dept. of the Army, Picatinny Arsenal, Dover, New Jersey (May 30, 1975).
16. Chehaske, J. T. and Greenberg, J. S., "Testing of a Molecular Sieve
Used to Control NOx Emissions from a Nitric Acid Plant", draft report
prepared for EPA by Engineering-Science, Inc., McLean, Virginia
(June, 1975).
17. Casetty, R., Union Carbide Corp., Tarrytown, New York, private
communication to H. So Rosenberg of Battelie-Columbus (June 23, 1975).
18. Study in progress for the American Petroleum Institute.
19. Peters, M. S. and Timmerhaus, K. D., Plant Design and Economics for
Chemical Engineers, Second Edition, McGraw-Hill, New York (1968)
pp 173-175.
20. Clapperton, J. A., PPG Industries, Pittsburgh, Pennsylvania, letter
to M. Y. Anastas of Battelle-Columbus (February 28, 1975).
21. "Cost of Clean Air, 1975", report in progress by Battelle-Columbus,
22. Searles, R. A., "Cleaner Nitric Acid Plant Tail Gases", Nitrogen,
(88), 38-42 (March/April, 1974).
23. "Grande Paroisse Absorption Process for NOX Abatement", J. F. Pritchard
& Company, Kansas City, Missouri (May, 1973).
-------
APPENDIX A
DETAILED COST DATA
-------
APPENDIX A
DETAILED COST DATA
This appendix presents some details related to the cost data
presented in the text. Data are included for the PuraSiv N process,
catalytic reduction, and extended absorption.
PuraSiv N Process
This section contains data obtained from Union Carbide on the
investment and operating costs for the PuraSiv N process. The following
data are included:
Figure A-l. PuraSiv N Cost vs Tail Gas Flow Rate
Table A-l. PuraSiv N Adsorbent Price Estimates Through 1975
Table A-2. PuraSiv N Utility Estimates.
These data were used in the cost estimates presented in this
report, based on a tail gas flow rate of 78 scfm/tpd. The cost of the
molecular sieve was taken as an 0 & M cost based on a yearly service
contract, since this is the manner in which Union Carbide quotes this
cost.
Catalytic Reduction
The capital cost for the catalytic reduction process as a function
(21)
of plant size was taken from the Cost of Clean Air study. Tables A-3,
(22)
A-4, and A-5, originally presented by Searles, provide backup data on
the investment and operating costs.
-------
A-2
5000
4000
3000
2000
1000
800
600
500
400
300
200:
Accuracy +/- 25 %
NOX = 2500 ppm/v
Multiple beds
or trains
4 beds
2 beds
Summer 1974
M 8 S = 386
CE = 165
Multiple
skids
i i i i
Field
assembly
I
I i i i i
4 5 6 8 10 20 30 40 50 60 80 100
Tail Gas Flow, M SCFM
FIGURE A-l. PURASIV N COST VERSUS TAIL GAS FLOW RATE
-------
A-3
TABLE A-l. PURASIV N ADSORBENT PRICE ESTIMATES THROUGH 1975
F = TAIL GAS FLOW RATE FROM CUSTOMER'S ABSORBER, MSCFM @ 60°F, 14.7 psia
A.
CUSTOMER NEEDS 250 OR HIGHER ppmv GUARANTEED NO IN EFFLUENT
A
ppmv NO
A
in Tail Gas
1500
2500
3500
Outright Sale Price Due
Upon Unit Acceptance
3-Year Pro-Rated Life Guarantee
2-Bed Type
$4,000 F
$5,900 F
$7,200 F
4-Bed Type
Yearly Service Contract, First
Payment Due Upon Unit Acceptance
Contract Runs for 4 Years
2-Bed Type
$3,100 F $1,460 F/Yr
$4,300 F $2,130 F/Yr
$5,200 F $2,680 F/Yr
4-Bed Type
$1,120 F/Yr
$1,600 F/Yr
$1,900 F/Yr
B. CUSTOMER NEEDS 50 ppmv GUARANTEED NO IN EFFLUENT
ppmv NO
X
in Tail Gas
1500
2500
3500
Outright Sale Price Due
Upon Unit Acceptance
2-Year Pro-Rated Life Guarantee
2-Bed Type 1 4-Bed Type
$4,000 F i $3,100 F
$5,900 F i $4,300 F
$7,200 F | $5,200 F
Yearly Service Contract, First
Payment Due Upon Unit Acceptance
Contract Runs for 4 Years
2-Bed Type
$2,090 F/Yr
$3,050 F/Yr
$3,830 F/Yr
4-Bed Type
$1,600 F/Yr
$2,300 F/Yr
$2,700 F/Yr
2-BED/4-BED FIRST PASS SELECTION GUIDELINE
ppmv NO
in Tail Gas
Below 2500
2500 - 3000
More than 3000
F Less Than
7.5
2-Bed
2-Bed
4-Bed
F Between 7.5
and 20
2-Bed
Toss-Up
4-Bed
F Greater Than
20
4-Bed
4-Bed
4-Bed
-------
A-4
TABLE A-2. PURASIV N UTILITY ESTIMATES*
F = MSCFM TAIL GAS @ 60°F, 14.7 psia
**
AVERAGE USAGE
2 Bed Unit
4 Bed Unit
Power
KW
18.5 F
17.5 F
Cooling H20
@ 15°F Rise,
GPM
35 F
32.5 F
Heat Absorbed @
100% Furnace Eff.
MM BTU/DAY
1.5 F
1.35 F
Tail Gas at 90 F and 90 psig used as Base Point. Lower
temperatures and higher pressures will reduce the utility
requirements.
**
Peak (demand) usage figures are, of course, somewhat
higher. The customer uses the average figures above for
his operating cost economic studies, but he needs to know
the demand figures to size utility lines, etc. Here are
some guidelines:
Demand Power = 1.10 x Avg. Power
Demand Cooling ^0 = 1.0 x Avg. Cooling H20
Demand Absorbed Heat = 2 x Avg. Absorbed Heat
-------
A-5
TABLE A-3. COST OF PRODUCING NITRIC ACID
(300 tons/day plant built in 1962)
(without NO abatement)
Unit
Cost/
Unit
Quantity $
Cost/Ton
Acid
$
A. Raw material
1. Ammonia
B. Utility Consumption
1. Cooling water
2. Steam
3. Electricity
C. Other Cost
1. Maintenance
2. Depreciation
3. Operating labor
Basic
1.
2.
3.
4.
5.
Investment
Production
Maintenance
Depreciation
Operating labor
Tons
M. Gal.
Import M Ib.
kWh
$
$
$
0.2998
34.000
0.610
6.000
35.00
0.02
0.92
0.01
10.49
0.68
0.56
0.06
0.71
1.18
0.86
$14.54
$1,800,000
102,000 tons/yr
$72,000
15 years
8,760 man-hours/yr at $10.00/hr
Source: Reference 22
-------
A-6
TABLE A-4. COST OF PRODUCING NITRIC ACID
(300 ton/day Plant Built in 1962)
(NO Abatement Added 1972)
Unit
Cost/ Cost/Ton
Quantity Unit, $ Acid, $
A.
B.
C.
Raw Material & Catalyst
1. Ammonia
2. Abatement catalyst
Utility Consumption
1. Cooling water
2. Steam
3. Electricity
4. Fuel
Other Cost
1. Maintenance
2. Depreciation
3. Operating labor
Tons
$
M gal
Export M
kWh
MM Btu
$
$
$
0.2998 35
34.000 0.02
Ib 1.090 0.92
6.000 0.01
1.75 0.65
10.49
0.15
0.68
(1.00)
0.06
1.14
0.88
1.47
14.75
Basic
1. Investment
2. Production
3. Maintenance
4. Depreciation
5. Operating labor
$2,200,000
100,000 ton/yr
$88,000
15 years
8,760 man-hours /yr at
$10/hr
Source: Reference 22.
-------
A-7
TABLE A-5. COST OF PRODUCING NITRIC ACID
(300 ton/day Plant Built in 1972)
(NO Abatement Included)
A.
B.
C.
Raw Material & Catalyst
1. Ammonia
2. Abatement catalyst
Utility Consumption
1. Cooling water
2. Steam
3. Electricity
4. Fuel
Other Cost
1. Maintenance
2. Depreciation
3. Operating labor
Unit
Tons
$
M gal
Export M Ib
kWh
mm Btu
$
$
$
Quantity
0.290
20.5
2.430
3.00
1.75
Cost/
Unit, $
35
0.02
0.92
0.01
0.65
Cost/Ton
Acid, $
10.15
0.15
0.41
(2.24)
0.03
1.14
0.49
1.31
0.20
$11.64
Basic
1. Investment
2. Production
3. Maintenance
4. Depreciation
5. Operating labor
$2,000,000
102,000 ton/yr
$50,000
15 years
2,000 man-hours/yr at $10/hr
Source: Reference 22.
-------
A-8
The cost of natural gas at which molecular sieve adsorption
becomes economically competitive with catalytic reduction was calculated
by the following method:
100 ton/day plant
K , (PuraSiv N) = K . (Catalytic Reduction)
total total
Catalytic Reduction:
K0+M = Ktotal " KI
K^v, - $1,013,600 - 255,300
OrM
KOfM=$758»30°
0 + M = K^/6.67
0 + M = $113,700
0 + M - cost of natural gas = $32,300 - 84,000 = -$51,700
cost of natural gas = $113,700 + 51,700 = $165,400
$1.2 $165.400 = $2.4
1000 ft3 X $ 84'000 " 1000 ft3
A similar analysis for the 1,000 ton/ day plant also yielded a
3
natural gas cost of $2.4/1000 ft .
Extended Absorption
The data included in Figure A-2 and Table A-6 were provided by
J. F. Pritchard and Company. However, the investment costs are out of
date and updated costs were supplied by Mr. L. Boudreaux, Senior Process
Engineer, J. F. Pritchard and Company.
-------
A-9
2.or
in
JJ
"5
1.0
0.8
0.6
0)
I 0.3
'o.
o
O
0.2
O.I
100 200 300 400
Existing Plant Capacity (100 Percent HN03), short tons per day
500
FIGURE A-2. CAPITAL INVESTMENT VERSUS EXISTING PLANT CAPACITY
GRANDE PAROISSE EXTENDED ABSORPTION PROCESS FOR
NO ABATEMENT
x
-------
A-10
TABLE A-6. ECONOMICS OF TYPICAL PROCESS APPLICATIONS GRANDE
PAROISSE ABSORPTION PROCESS FOR NO ABATEMENT
x
I.
II.
Ill
IV.
PROCESS CONDITIONS
Existing Capacity, Tons/Day
N(0)x Emission, ppm (v)
Tail Gas Pressure, PSIG
Tail Gas Temperature, °F
Residual Oxygen, Mol . %
Design Emission Level, ppm(v) N(0)x
UTILITIES
Cooling Water, GPM
Electric Power, KW
Instrument Air, SCFM
Pressure Loss, PSI
.INVESTMENT
Order-of-Magni tude
(Including Royalty)
YEARLY OPERATING COST *
1 . Operating Labor
2. Maintenance Material & Labor
@ 1.5/& of Investment
3. Depreciation, Taxes, Insurance
@ 12.5% of Investment
4. Electricity @ 10 tnils/Kwh
5. Cooling Water (Circ.)
(? 2
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-015
2.
4. TITLE ANDSUBTITLE
Molecular Sieve NQx Control Process in Nitric
Acid Plants
3. RECIPIENT'S ACCESSION-NO.
5. REPORT DATE
January 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
Harvey S. Rosenberg
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Battelle-Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
10. PROGRAM ELEMENT NO.
1AB014; ROAP 21ADH-008
11. CONTRACT/GRANT NO.
68-02-1323, Task 17
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; 10/74-11/75
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT The repor(; gives results of an engineering analysis of the applicability of
molecular sieve technology to the control of NOx emissions from nitric acid plants.
Field test data from a plant using this technology show that, after 6 months of oper-
ation, the plant still controls NOx emissions to well within the New Source Perfor-
mance Standard (NSPS) (1. 5 kg of NOx/metric ton of acid; equivalent to about 200 ppm
NOx in the tail gas). Field test data from a second plant, 10 months after start-up,
show that NOx emissions are below the NSPS, even though the sieve had been acci-
dentally damaged. The process appears able to achieve an average effluent NOx con-
centration of 50 ppm, based on tests at the former plant; however, this concentration
was not achieved during the tests at the latter plant because of the damaged sieve.
Although a 2-year sieve life has not been demonstrated, there is no reason to be-
lieve it cannot be achieved, and it appears that molecular sieve technology is tech-
nically feasible. The economic feasibility of molecular sieve technology for this
application was assessed by comparing this technology with the catalytic reduction
and extended absorption processes, both of which usually limit effluent NOx concen-
tration to only about 200 ppm. The capitalized cost for the molecular sieve process
is higher than for catalytic reduction and lower than for extended absorption.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Air Pollution
Absorbers (Materials)
Nitrogen Oxides
Industrial Plants
Nitric Acid
Catalysis
Cost Analysis
Evaluation
3. DISTRIBUTION STATEMENT
Unlimited
b.lDENTIFIERS'OPEN ENDED TERMS
Air Pollution Control
Stationary Sources
Molecular Sieves
Nitric Acid Plants
Catalytic Reduction
Extended Absorption
19. SECURITY CLASS (This Report,
Unclassified
20. SECURITY CLASS (This page}
Unclassified
14A
c. cos AT I Field/Group
13B~
11G
07B
07D
21. NO. OF PAGES
_ J75
22. PR~lCE~
EPA Form 2220-1 (9-73)
------- |