E^-600/2-76-014
Japuary 1976
Environmental Protection Technology Series
 MOLECULAR  SIEVE MERCURY  CONTROL  PROCESS
                            IK  CHLOR-ALKALI PLANTS
                                  industrial Environ mental Research
                                       Office of Research and Development
                                      U.S. Environmental Protectm Agency
                                 Research Triangle Park, North Carolina 27711

-------
               RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:

     1.   Environmental Health Effects Research
     2.   Environmental Protection Technology
     3.   Ecological Research
     4.   Environmental Monitoring     '         .  ;   '     ;  ;.  ,'
     5.   Socioeconomic Environmental Studies     '"'           ':

This .report has-been assigned  to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources  of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
                    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-600/2-76-014
         MOLECULAR SIEVE

    MERCURY  CONTROL  PROCESS

     IN  CHLOR-ALKALI  PLANTS
                    by

               M.Y. Anastas

       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

          The applicability of the PuraSiv Hg adsorption process to mercury
removal from the hydrogen byproduct and the end-box ventilation streams
from mercury cell chlor-alkali plants was investigated.  The investigation
included the analysis of data obtained from testing of a system that is
currently in operation and technical information provided by the system
vendor together with that available in the open literature.
          Although the measurements of mercury concentration in the hydrogen
byproduct stream entering the PuraSiv Hg adsorber taken during performance
testing of the control unit appear to be in error, measurements of the
outlet concentration indicate that a concentration less than 60 ppbv may
be achieved.
          The economics of the PuraSiv Hg adsorption process were explored.
Available data indicate that the operating costs by this process vary between
0.58 and 0.33 per ton of chlorine produced for plants with capacities between
100 and 750 tons per day.
          Mercury removal from the hydrogen byproduct stream may also be achieved
by either adsorption over treated activated carbon or by scrubbing with depleted
brine.  Technical and economic data available to the investigator seem to
favor the use of these two processes for mercury control although the data base
thereon is not sufficiently developed to warrant a meaningful comparison.
                                    iii

-------
                            TABLE OF CONTENTS

                                                                     Page

INTRODUCTION	     1

MERCURY-CELL CHLORINE PROCESS MERCURY AIR EMISSIONS AND CONTROL
  ALTERNATIVES . . .	     3

     Description of the Mercury-Cell Process 	     3

     Sources of Mercury Emissions  	     6

     Mercury Control Technology in Chlor-Alkall Plants 	    10

TECHNICAL ANALYSIS OF MERCURY CONTROL BY MOLECULAR SIEVE
,  ADSORPTION	    20

     General Description of Molecular Sieves 	    20

     Use of Molecular Sieves for Mercury Removal from Gaseous
       Streams	    21

     Performance of a PuraSiv Hg Mercury Adsorption System ....    22

     Factors Affecting Performance of Molecular Sieve Mercury
       Control Systems	    31

ECONOMIC ANALYSIS	    33

     Capital Costs	    33

     Operating and Maintenance Costs 	    34

     Total Annual Costs of PuraSiv Hg Adsorption . .	    34

CONCLUSIONS AND RECOMMENDATIONS	    37

REFERENCES	 .    40

                               APPENDIX A

CALCULATION OF AVERAGE MERCURY CONCENTRATIONS AND AVERAGE DAILY
  EMISSIONS IN MOLECULAR SIEVE ADSORBER OUTLET 	    A-l

                               APPENDIX B

CAPITAL, OPERATING AND MAINTENANCE, AND ANNUAL COSTS OF MERCURY
  CONTROL PROCESSES  	    B-l

-------
                             LIST OF TABLES
Table 1.


Table 2.


Table 3.


Table 4.


Table 5.


Table 6.
Estimated Mercury Emissions from a 300 TPD Mercury-
Cell Chlorine Plant 	 .  	
Approximate Material Balance for Molecular Sieve
Adsorption in a 100 TPD Plant 	
Summary of Test Data Obtained from a PuraSiv Hg Mercury
Adsorption Process  	

Mercury Emissions from a 200 TPD Mercury-Cell Chlorine
Plant ... 	

Capital Costs of Cooling/Demisting and PuraSiv Hg
Adsorption  	
O&M Costs of PuraSiv Adsorption
14


23


30


35

36
Table A-l.  Calculation of Average Mercury Concentration for Cycle    A-7
Table B-l.  Capital Costs of Cooling/Demisting and List of
            Equipment  	
Table B-2.  Capital Costs and List of Equipment for PuraSiv Hg
            Adsorption   	 	
Table B-3.  Basis for Computation of Annual O&M Costs
Table B-4.  Utility Requirements for Cooling/Demisting in a 100 TPD
            Plant  	

Table B-5.  Utility Requirements for PuraSiv Hg Adsorption in a
            100 TPD Plant  	
Table B-6.  O&M Cost Components for Cooling/Demisting, $1000
            (Curve A, Figure B-2) 	
Table B-7.  Incremental O&M Costs of Cooling/Demisting Attributed
            to PuraSiv Hg, $1000  	
Table B-8.  O&M Costs of PuraSiv Hg ($1000) Excluding Incremental
            Cooling/Demisting Costs (Curve C, Figure B-2) ....
                                                          B-3


                                                          B-5

                                                          B-7


                                                          B-8


                                                          B-ll


                                                          B-13


                                                          B-14


                                                          B-15

-------
                             LIST OF FIGURES

                                                                     Page

Figure 1.    Basic Flow Diagram for Chlor-Alakli Mercury-Cell
             Operation 	     4
Figure 2.    Mercury Emission Control by Molecular Sieve Adsorption
             from Hydrogen Byproduct Stream	 .     13

Figure 3.    Mercury Control by Activated Carbon Adsorption  ...     16

Figure 4.    Depleted Brine Scrubbing for Mercury Removal from the
             Hydrogen Byproduct Stream 	     19

Figure 5.    Inlet and Outlet Mercury Concentration Profiles for a
             Molecular Sieve Adsorber - Cycle 1  	     24

Figure 6.    Inlet and Outlet Mercury Concentration Profiles for a
             Molecular Sieve Adsorber - Cycle 2  	     25

Figure 7.    Inlet and Outlet Mercury Concentration Profiles for a
             Molecular Sieve Adsorber - Cycle 3  	     26

Figure 8.    Inlet and Outlet Mercury Concentration Profiles for a
             Molecular Sieve Adsorber - Cycle 4  .. 	     27

Figure 9.    Inlet and Outlet Mercury Concentration Profiles for a
             Molecular Sieve Adsorber - Cycle 5	     28

Figure A-l.  Effluent Mercury Concentration from PuraSiv Control
             Unit - Cycle 1	     A-2

Figure A-2.  Effluent Mercury Concentration from PuraSiv Control
             Unit - Cycle 2	     A-3

Figure A-3.  Effluent Mercury Concentration from PuraSiv Control
             Unit - Cycle 3	     A-4

Figure A-4.  Effluent Mercury Concentration from PuraSiv Control
             Unit - Cycle 4	     A-5

Figure A-5.  Effluent Mercury Concentration from PuraSiv Control
             Unit - Cycle 5	     A-6

Figure B-l.  Capital Costs of Control of Mercury in the Hydrogen
             Byproduct Stream	     B-4

Figure B-2.  Operating and Maintenance Costs for Mercury Control
             in Hydrogen Byproduct Stream	     B-10
                                   vi

-------
                              TOPICAL REPORT
                                     on
                  ENGINEERING ANALYSIS OF THE PURASIV HG
                     ADSORPTION FOR MERCURY CONTROL IN
                        MERCURY-CELL CHLORINE PLANTS
                                     to

                INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
                      OFFICE OF RESEARCH & DEVELOPMENT
                      ENVIRONMENTAL PROTECTION AGENCY
                                   from
                                 BATTELLE
                          Columbus Laboratories
                               INTRODUCTION

          Mercury is a hazardous air pollutant that is emitted from three
gaseous streams resulting from the manufacture of chlorine by the mercury-
cell process.  These are the hydrogen byproduct, end-box ventilation and
cell-room ventilation air streams.  A number of control technologies are
commercially available for the control of the pollutant in the hydrogen
byproduct and end-box ventilation streams.  Among these control alternatives
is adsorption of the mercury vapor in the PuraSiv Hg adsorption process
which is designed for recovery of mercury vapor and cyclic regeneration
of the adsorbent.  Union Carbide Corporation (UCC), designer of the PuraSiv
Hg process claims that the level of control of the process is such that the
average concentration of mercury in the effluent gas stream is at or below
60 ppbv.  The life of the adsorbent is also guaranteed for three years.
          In an effort to determine the technical and economic feasibility
of the PuraSiv Hg process in its applicability to the stated purpose, the
U.S. EPA sponsored the study at hand.  It is the objective of this study

-------
to determine (1) the technical feasibility of the control system on the
basis of the analysis of data independently obtained from testing of a .
PuraSiv Hg adsorption system currently in operation and from analysis of
information made available by the system vendor and (2) the economics of
the system.

-------
                   MERCURY-CELL CHLORINE PROCESS MERCURY AIR
                     EMISSIONS AND CONTROL ALTERNATIVES
          Chlorine may be produced by electrolytic methods from fused
chlorides of aqueous solutions of alkali-metal chlorides. (1»2»3»^»5)  In
the electrolysis of aqueous solutions of potassium or sodium chloride,  chlorine
is produced at the anode, while hydrogen and potassium hydroxide or sodium
hydroxide are produced as a result of processes taking place at the cathode.
This requires that the anode and cathode products be kept entirely separate.
Consequently, many ingenious cell designs have been developed and refined; all
of these have been variations either on the diaphragm cell,  or on a cell which
employs mercury metal as an intermediate cathode.  Historically, these two processes
                                        (4)
were developed more or less in parallel.     In the United States, the mercury
process was an early leader, but shrunk to less than 5 percent of the installed
chlorine capacity in 1946.  The use of the mercury cell since, has grown toward
28 percent of the installed U.S. chlorine capacity through 1968.  Since then,
there has been a negative trend in the use of mercury cells.  Chlorine production
in 1974 amounted to 33,000 tons per day    of which 21 to 23 percent were
obtained from mercury cells.     During the first three quarters of 1974,
demand for chlorine was high and plants were operating at 95 percent of
capacity or higher.  Since then the operating rate has dropped to 83
percent.     While the number of mercury-cell chlorine plants have not
increased recently, the introduction of "dimensionally stable" anodes
instead of graphite anodes increased the production of chlorine by mercury
cell processes.

                Description of the Mercury-Cell Process

          The use of potassium chloride brines for chlorine production is
entirely analogous to that of sodium chloride.  The latter will serve as
the basis for the following discussion.  The basic process flow sheet for
the production of chlorine and caustic soda is shown in Figure 1.
          Sodium chloride may be obtained from brines or seawater.  Whatever
the source, solid sodium chloride is the starting material fed to the process.
          Purified saturated brine is fed continuously from the main
brine treatment section through the inlet end box to the electrolyzer

-------
BASIC TREATMENT CHEMICALS
(SODA ASH, CAUSTIC LIME,
ACID, CaCL2, ETC.)
  SOLID
NaCL FEED
               CHLORINE
                                                              PRODUCT
                                                              CHLORINE
OTHER
                  I
                 BRINE
             DECHLORINATOR
                               SPENT BRINE
STREAM
RECYCLE
                              TREATED
  MAIN BRINE
  SATURATION,
  PURIFICATION, AND
  FILTRATION
                               BRINE
                            INLET
                            END-BOX-
                   END-BOX
                   VENTILATION SYSTEM'
                    AQUEOUS
                    LAYER
                        STRIPPED
                       - AMALGAM
       END-BOX
       VENTILATION SYSTEM
                       Hg PUMP
               AQUEOUS LAYER
                                                           L
            1
     COOLING,
     DRYING,
     COMPRESSION, AND
     LIQUEFACTION
                                   ELECTROLYZER
                                                 I
— OUTLET END-BOX

_^.END-BOX
     VENTILATION SYSTEM
•**—	' AQUEOUS LAYER
                                              WATER COLLECTION
                                              SYSTEM
                                            END-BOX
                                           ' VENTILATION SYSTEM
           HYDROGEN GAS

             BYPRODUCT '
                                    DECOMPOSER
                                    (DENUDER)
                                Purified^
                                 Water
                                                                      AMALGAM
                                                         CAUSTIC SODA
                                                         SOLUTION
                                                                                    PRODUCT
 PROPRIETARY TREATMENT CHEMICALS INCLUDE PRECIPITATORS,
 FLOCCULANTS, POLYELECTROLYTES, AND SIMILAR MATERIALS
                                                                       TION, AND  JVCAUSTIC
     FIGURE 1.  'BASIC FLOW DIAGRAM FOR CHLOR-ALKALI MERCURY^GELL  OPERATION
                                                                                   (8)

-------
 where it flows  between a stationary  graphite  (or metal  dimensionally
 stable)  anode and a flowing mercury  cathode.   The  following  reactions
 take place in the mercury cell:
           Anodic half-reaction:   2Cl~(aq)   ->•   C12(aq) + 2e

                                  C12(aq)   *  C12(g)
           Cathodic half-reaction:  2Na+ + Hg  + 2e~  •»•  2Na(Hg)
           Overall cell reaction:    2NaCl(  j  + Hg  •*•  Cl2(g) + 2NaHg.

           The inlet box is a receptacle placed on  the inlet and of the electro-
 lyzer to provide convenient connections for the feed brine and the stripped
 mercury returning from the decomposer.   It also serves to keep the incoming
 mercury covered with brine.  The spent  brine  is recycled from the electro-
 lyzer to the main brine treatment section of  the plant; it may also be sent
 through a dechlorination step in which  it is  vacuum  treated and/or air blown.
           The chlorine gas product formed at  the anode leaves the
 electrolyzer for further treatment.   After cooling,  the wet gas is dried
 by scrubbing with concentrated sulfuric acid.  The spent acid from this
 step contains most of the mercury brought along by the wet cell gas.   After
 compression, the dry chlorine gas at elevated pressure either may be used
 directly or may be subjected to a liquefaction step.  It is not known whether
there are emissions of mercury to the atmosphere during cleanup of the chlorine
prior to sale.
          The sodium amalgam flows continuously from the electrolyzer
through the outlet end box to the decomposer  (denuder) where it is made
the anode to a short-circuited graphite or metal cathode (packed bed of
chips) in an electrolyte of sodium hydroxide solution.  The outlet box is
a receptacle placed on the outlet end of the electrolyzer to provide a
convenient means to keep the sodium amalgam covered with spent brine and
at the same time to permit the physical  separation  of the two streams.
Purified water is fed continuously to the decomposer, where it reacts with
the sodium amalgam to produce sodium hydroxide solution as well as byproduct
hydrogen gas according to the following reactions.   The following reactions

-------
 take place in the denuder:
                                        .  '   V '     •'•',:+ f..*,:. •(   '  •
           Anodic (mercury)  half-reaction:   2Na(Hg)  -*•  2Na  + Hg  + 2e
           Cathodic (graphite)  half- reaction:   2H20  + ,2e~ _  + ,20H~ +:1H
           Overall decomposition reaction:          :       •"
                    2Na(Hg)  + 2H20  -»  2NaOH + HZ(  ,  + Hg.

 The high-purity caustic soda generally leaves the  decomposer at the con-
 centration of about 50 weight percent sodium hydroxide.     ,  ,
           The caustic soda solution from the decomposer usually is sent to
 a filtration unit.  The solid waste from the filter, which: contains the
 precoat material, is usually sent to an externally fired retort similar
 to that used for the secondary processing of waste mercury.   The mercury.
 thus recovered is returned to the cell for reuse.           .
          Filtered caustic solution at a concentration of 50 weight percent
may be concentrated further by evaporation  to a 73 weight percent. sodium
 hydroxide  product.   In some  instances,  this material is  heated  to drive off
 the remaining  water  in order  to  produce  anhydrous  pellets or  flakes  of  solid
 sodium hydroxide.                                    : ,  .    -   ,  " •

           The  byproduct hydrogen gas  from  the decomposer may be vented  .to
 the atmosphere,  burned as  fuel,  or  used. as  a feed  material to a process
 after removal  of the  mercury  therein.                    ,.,-...•

                       Sources  of Mercury Emissions    ,

           The  main  sources of  mercury emissions to the atmosphere from
 mercury-cell chlorine plants are
           (1)  The hydrogen byproduct  stream                     .     ;
                    i     '              i                  '             '
           (2)  The cell  end-box  ventilation air stream      .             ,
           (3)  The cell-room ventilation air.        .                 •  . .  .,
           The  hydrogen byproduct stream  leaving the  decomposer  at 180 iF
 and  1 atm  is saturated with mercury vapor  and will carry unknown .

-------
amounts of mercury in the form of fine droplets which may be an order of
magnitude higher than the saturation value.  Most of this mercury may be
removed by cooling in a surface exchanger to 110 F and passing through a
wire-mesh pad mist eliminator.  The stream thus treated is saturated with
mercury vapor at 10 ppmv and may contain 3 to 10 times the saturation value
in the form of mercury droplets most of which are 3 micron in diameter
or less.  At the 10 ppmv level, the mercury losses will be higher than
5 pounds per day for a 100 ton per day (TPD) plant.
          Mercury and mercury compounds leave the cell/decomposer system
in the end-box ventilation air.  The volume of ventilation air depends upon
the age of the plant, the type and specific physical configuration of the
cells and end boxes, and the standard operating procedures employed at a
particular location.     In some cases, the volume flow rate of end-box
ventilation air may approach or exceed that of the hydrogen produced in
the decomposer.  For some new cell modifications currently in operation,
the volume of the end-box ventilation air is usually equal to or less than
that of the hydrogen byproduct stream.  In a certain plant the former is about
80 percent of the latter.  Because of the wide variations in end-box construction,
it is difficult to specify the emissions.  The best current intelligence
indicates that the mercury losses sustained in the untreated end-box
ventilation air range from about 0.1 to 1.2 pounds per 100 tons of chlorine
         (4)
produced.     These values are reasonable in view of the ranges of temperature,
degree of saturation, and ventilation-air volume to be expected.
          The volume of cell-room ventilation air employed varies from
                                                           (45}
approximately 100,000 to 1,000,000 cfm for a 100 TPD plant.  ''  A part
of this air serves to cool the cells and to keep the cell-room temperature
within acceptable limits.  The minimum amount of air required for this
purpose varies seasonally with geographic location, cell design, and the
age of the plant.     Additional ventilation air is usually required to
remove mercury vapor or compounds from the working environment.
Estimates of the mercury losses, under normal conditions of operation, in
the cell room ventilation air range from about 0.5 to 5 pounds per 100 tons
of chlorine produced for plants of 100 to 150 TPD chlorine capacity.  The
mercury emissions from plants larger than 100 to 150 TPD chlorine capacity
apparently are less than would be indicated by a direct proportion based on
increased production capacity.  These estimates are consistent with the

-------
approximations obtained by multiplying the ventilation-air volume by the
                            Q
TLV concentration of 50 mg/m  .  On this basis emissions of 0.45 pounds/day
for 100,000 cfm and 4.5 pounds/day for 1,000,000 cfm of cell room ventilation
airflow may be expected.  Thus, for some plants, the losses of mercury in
the cell room ventilation air may amount to more than the combined losses
in the hydrogen and the end-box ventilation.
          Mercury enters the cell-room atmosphere as a result of a number
of operations or conditions:
          •  End box sampling procedures
          •  Removal of mercury butter from the end boxes
          •  Cell maintenance and rebuilding operations
          •  Other maintenance work which exposes the internal surfaces
             of pipes and equipment
          •  Accidental spills of mercury
          •  Cell and mercury pump leaks
          •  Cell failure, and other unusual circumstances.
The number and variety of sources of mercury in the cell room air indicate
that careful plant operation and good housekeeping are essential in order
to minimize the amount of mercury emitted into the cell room air.  On the
other hand, it is extremely difficult to completely eliminate the contamination
of the cell room atmosphere with mercury.
         Typical emissions of mercury from a 300 TPD mercury-cell chlorine
plant are given in Table 1.

-------
      TABLE 1.  ESTIMATED MERCURY EMISSIONS FROM
                A 300 TPD MERCURY-CELL CHLORINE
                PLANT
                              Mercury
                              Emission        Percent
         Steam              grams per day     of Total
                   a
Hydrogen-Byproduct^ '           8,300           89.63

Cell Room Ventilation^          900            9.72
End Box Ventilation'0^          	60_            0.65

Total                           9,260          100.00
(a)  Saturated with respect to both water and mercury
     at 110 F and 1 atm.  Does not include mercury mist,
     Reference (4).
(b)  Based on a ventilation rate of 1450 cfm per TPD
     chlorine and a TLV of 50 micrograms mercury per
     cubic meter at 70 F and 1 atm.  Reference (5).
(c)  Based on ventilation rate of 8500 scf (60 F and
     1 atm.) air per. ton of chlorine and a mercury
     concentration that is 1/10 of saturation @ 110 F.

-------
                                       10
              Mercury Control Technology in Chlor-Alkali Plants

         The national emission standard for mercury from stationary sources,
 including mercury-cell chlorine plants, limits the daily emission to 2,300
                     (9)
 grams per day  (GMPD).     Recently, EPA proposed amendments to this mercury
 emission standard which would raise the limit to 3,200 GMPD.   '  The 2,300
 GMPD limit is divided among the three sources of mercury air emissions such
 that 1,300 GMPD may be emitted in the cell-room ventilation air and the
 remaining 1,000 GMPD may be emitted in the hydrogen byproduct and end-box
                    (9)
 ventilation streams.
         Mercury emission control technology that is currently being employed
 (including the molecular sieve adsorption process) is apparently capable of
 achieving the 1,000 GMPD limit for both streams in the largest mercury cell
 chlorine plants (about 750 TPD).  As mentioned earlier, primary cooling and
 demisting are a necessary precursor to the other control techniques as it
 is capable of reducing the mercury concentration to about 1 ppmv (as vapor).
Other alternatives are reportedly available for reducing the latter concen-
 tration to below 50 ppbv.  If a daily emission of 500 GMPD is allowed for
 the hydrogen byproduct stream in a 750 TPD plant then an adequate level of
mercury control will allow an effluent mercury concentration of about 250 ppbv.
For a 100 TPD chlorine plant, a concentration of about 1.0 ppmv may be allowed.
When preceded by cooling and demisting the control techniques (molecular
sieve and treated activated carbon adsorption and depleted brine scrubbing)
to be discussed below are capable of reducing the time-average mercury con-
centration to below 50 ppbv.
         The end-box ventilation stream is usually of the same volumetric flow
rate (or less) as the hydrogen byproduct stream.  Furthermore, control schemes
available for the latter are usually also available for the former.
         There are several control processes that are potentially available
for control of mercury emissions from the hydrogen byproduct and end-box
ventilation air streams by reducing the vapor concentration from 1 ppmv to 60 ppbv,
These processes are usually preceded by secondary cooling and demisting.  The
processes are:

-------
                                       11

          •  PuraSiv Hg  adsorption
          •  Sulfur- or  iodine-impregnated activated carbon adsorption
          •  Depleted brine scrubbing
          •  Hypochlorite scrubbing.
          Common to all three processes under consideration are two processing
steps prior to mercury vapor adsorption by molecular sieves or sulfur
impregnated activated carbon or scrubbing with depleted chlorinated brine.
The first step is cooling the hydrogen stream from about 180 F to 110 F using
80 F cooling water followed by mercury mist elimination in a wire-mesh pad mist
eliminator.  In application of the device for collection of acid mist in sulfuric
acid plants, it has been found capable of removing 99 percent  (plus) of particles
of 3 microns in diameter or greater  (> 3y ) and 15 to 30 percent of particules
<3y .      The concentration of mercury in the stream at this point (no F, 1 atm.)
will be the saturation value of 10 ppmv and (probably) an additional 20 to 30 ppmv
(equivalent) in the form of mercury mist.  The second step involves compressing the gas
to 6 psig in a rotary blower followed by cooling to between 55 F and 65 F and passing
it through a tubular type mist eliminator which has a removal efficiency of 99
percent (plus) for particles < 3y  in sulfuric acid-mist removal applications.
                                         (13)
Measured values of mercury concentration     downstream from such a tubular type
mist eliminator strongly indicate that the particulate removal efficiencies
reported for sulfuric acid mist are applicable for mercury mist since mercury
inlet concentrations that are 5 to 10 percent higher than the vapor saturation
value were observed.
           In what  follows  the  first  three control technologies listed
above will be  described.   A flow diagram and approximate material balance
for a 100  TPD  plant will be given.   The  design basis for the cooling/
demisting  step will be  as  follows.
           For  all  processes considered the cooling/demisting step involved
(1) cooling the hydrogen byproduct stream to 110 F  at 14.7 psia in the primary
cooler  from approximately  180  F, (2) mist elimination in a wire-mesh pad
demister,  (3)  compression  in a rotary blower to 20.7 psia,  (4) cooling to
60 F in the secondary cooler, and  (5) mist elimination in a tubular-type
eliminator.  The pollution control alternatives were assumed to reduce the
mercury concentration of about 1 ppmv, obtained after cooling and mist
elimination, to 60 ppbv.

-------
                                        12
 PuraSiv Hg Adsorption

           A process  for  control  of mercury  in  the  hydrogen byproduct  stream
                                                       (13)
 has  been developed by Union  Carbide  Corporation  (UCC).      The mercury vapor
 (about  1 ppm)  is  removed from  these  streams by adsorption  over a proprietary
 adsorbent especially designed  for such use.
           A typical  flow sheet of such a process is given  in Figure 2.  The
 purpose of the two-bed system  shown  is such that at any time one of the beds
 is undergoing  adsorption while the other is undergoing regeneration.  The
 hydrogen byproduct stream is compressed to a pressure between 4 and 8 psig,
 cooled  to a temperature  in the range 50 F to 70 F  and passed through a tubular
 type mist eliminator and steam heated to a temperature between 90 F and 100 F
 before  being introduced  to the PuraSiv Hg adsorber.  Adsorption of the mercury
 occurs  without a  significant rise in temperature.
          The  concentration of mercury in the effluent is guaranteed by Union
Carbide  to be  lower  than 60 ppbv on  the average.   Part of the adsorber effluent
 (10 to 20 percent of the hydrogen from the decomposer) from the bed in which
adsorption occurs is used  to regenerate the other bed.  In a dual-bed system
designed for adsorption over a period of 24 hours,  regeneration of the loaded bed
is carried out by heating  the generation gas to a temperature between 450
to 500 F in an electrical heater and passing it through the bed for about
12 hours.  Thereafter, the heater is shut off and the bed undergoes cooling
for about 12 hours.   The gas from the bed undergoing regeneration is
cooled before mixing with  the hydrogen byproduct  stream from the decomposer.
          The  process  can  be designed for a turndown ratio of 5:1.  The flow
through  the rotary blower  is the sum of the hydrogen byproduct and the
regeneration gas.  Hence for the PuraSiv Hg adsorption process the cooling/
demisting equipment  is 10  to 20 percent larger in capacity.  The life of the
molecular sieve absorbent  is guaranteed by the vendor to be not less than 3
years with effluent discharge of 60 ppbv.  This period was used as the design
life for the adsorbent.  An approximate material  balance appears in Table 2.

-------
                                                       Regeneration

                                                            heoter
            r
              Bypass
                                                                              Purified
                                                                              hydrogen
                                                                              stream
I Note: Adsorber I undergoing
       regeneration  add adsorber 2
       undergoing  adsorption.
                         Adsorber-2
TubuIqr
type mist
eliminator
Secondary
(Brine)
cooler
                                                                         ?C
                                                         ""^^A^egenerat i <
                                                                  gas
                                                                  cooler
              Knock-out
               drum
sealed
rotary
blower
               Primary
               (Water)
Wire-mesh      cooler
pad mist
eliminator and
seal pot
      FIGURE 2.   MERCURY EMISSION  CONTROL BY MOLECULAR SIEVE
                   ADSORPTION  FROM HYDROGEN BY-PRODUCT  STREAM

-------
TABLE 2.  APPROXIMATE MATERIAL BALANCE FOR MOLECULAR SIEVE ADSORPTION IN A 100 TPD PLANT

Stream
Number Description
1
2

3
4
5
6
7

8
9
10
11
12
13
14
(a)
(b)
(c)
Hydrogen Byproduct
from Decomposer
Hydrogen Byproduct
to the First Mist
Eliminator
Mercury and Water
from Separator
Gas to Blower
Mercury and Water
from Knock-out Drum
Gas to Secondary
Cooler
Gas to Tubluar Mist
Eliminator
Liquid Mercury and
Water
Gas to Preheater
Gas to Adsorber
Adsorber Effluent
Purified Gas
Regeneration Gas
Regeneration Gas to
Primary Cooler
Obtained at saturated vapor
At 60 F and 1 atm.
Hydrogen only.
Press
Psia
14.7
14.7

14.7
14.7
14.7
20.7
20.5

14.7
20.4
20.2
19.6
19.6
19.6
14.7
Temperature
F
180
110

110
110
110
110
60

60
60
100
100
100
100
180
conditions only. Extent


Mass
H2, Ibs/hr
234.7
270.0

unknown
270.0
unknown
270.0
270.0

unknown
270.0
270.0
270.0
234.7
35.3
35.3
of entrainment

(a)
Flow Ratev
H20, Ibs/hr Hg,
2207.0
231

unknown
231
unknown
.231
31

unknown
31.0
31.0
28.0
24.4
3.7
8.8 (max.)
not known.


Ibs/day
145.2
7.1

unknown
7.1
unknown
7.1
0.48

unknown
0.48
0.48
0.033
0.028
0.005
0.447


Volumetric
Flow rate,
SCFMOO
1,870
935

--
935
--
935
865

--
865
865
864
751
113
113 C



-------
                                  15
          The advantages and disadvantages of utilizing PuraSiv Hg adsorption
may be summarized as follows:
          Advantages

1.  Guaranteed performance of the ad-
    sorption system.

2.  Relative ease of operation and
    maintenance.

3.  The system does not require special
    materials of construction.  Carbon
    steel is used throughout.

4.  The recovered mercury is continuously
    recycled to the manufacturing
    process.

5.  Applicable for mercury control in end
    box ventilation air as is in the
    hydrogen byproduct stream.
    Disadvantages

Regeneration gas requirements
tend to limit the value of the
turndown ratio of the system
to a maximum value of 5:1.

Mercury adsorption efficiency
is dependent upon the relative
humidity of the adsorber inlet
gas stream.  Control to levels
specified by the vendor is
necessary.
Sulfur-Impregnated Activated Carbon Adsorption
          Mercury in the hydrogen byproduct stream can also be removed by

reaction with sulfur, to form mercury/sulfur compounds, in sulfur-impregnated

activated carbon. *•  »  »  '  The temperature-enhanced chemisorption at 100 F

to 110 F takes place in a one- or two-bed system as shown in Figure 3.  In

a two-bed system the FRP-lined (fiber glass reinforced plastic) absorbers
are in series with provisions for bypass or reversal of the flow through

the beds.  The FRP lining is necessary to prevent vessel corrosion by the

products of reaction that contain sulfur.  This scheme is used to effect a

maximum (100 percent) utilization and reliability of the adsorption system.

However, one-bed systems are currently in use.      Here a small fraction

of the adsorbent remains unloaded in the "mass transfer zone" and cannot be

used since the emissions after breakthrough may exceed allowable limits.

-------
Tubular
type mist
eliminator
Secondary
(Brine)
cooler
Knock-out
 drum
Water
sealed
rotary
blower
               Primary
               (Water)
Wire-mesh      cooler
pad mist
eliminator and
seal pot
Hydrogen
byproduct
stream from
decomposer
               FIGURE 3.   MERCURY CONTROL BY  ACTIVATED CARBON ADSORPTION

-------
                                  17
          Minor departures from the cooling/demisting steps reported
                                   ('14')
earlier are necessary in this case.      Secondary cooling to 50 F and
reheating to 110 F are undertaken.  The temperature of adsorption should
be kept below 150 F since H?S formation may occur.  Calgon, the vendor of
the adsorbent, reports typical outlet concentrations below 50 ppbv that
may approach 1 ppbv.
          A material balance similar to that presented in Table 2 was
performed.  The only difference here is the absence of recycle.  The hydro-
gen byproduct stream is heated to 100 F before entering the adsorber.  A
one-bed design was considered rather than two beds in series since opera-
tional experience indicates that the reliability, length of service, and
                                                           (14)
adsorbent requirements are satisfactory for such a design.  .    Kinetic
considerations seem to dictate the use of an amount of adsorbent that is
sufficient for a 3-year service.  Thereafter, the adsorbent may be shipped
to a processor for retorting of the mercury value therein.  The loading
capacity of the adsorbent is about 20 pounds of mercury per 100 pounds of
adsorbent.  Current designs in operation seem to favor a constant pressure
drop through the bed.  This gives a bed depth of about 4 feet.  A gas
residence time of 8 seconds in the bed is usually employed.
          The projected advantages and disadvantages of utilizing the
treated-activated-carbon adsorption process may be summarized as follows:

          Advantages                                 Disadvantages
1.  Relative simplicity of equipment      1.  A Fiberglas-reinforced-polyester
    design and layout.                        lining of the adsorption vessel
                                              is necessary to prevent corrosion
2.  Mercury adsorption to acceptably          thereof when sulfur-impregnated
    low levels (reportedly 1 ppb)             carbon is used.
    appears feasible.
3.  Ease of equipment operation and
    maintenance.
4.  It is possible that mercury could be
    recovered by retorting the spent
    adsorbent, although the economics of
    such a recovery operation are unknown.

-------
                                  18
          Advantages                            Disadvantages
    Adsorption process applicable to
    mercury from end-box ventilation
    air as well as removal from the
    hydrogen byproduct.
Depleted-Brine Scrubbing

          Another method of mercury removal involves scrubbing the
hydrogen byproduct stream first with depleted brine obtained from the
mercury cell and then with 10 percent caustic soda to remove chlorine
vapors picked up in the depleted brine scrubber (Figure 4).' ' ' '
The cooling/demisting treatment steps are the same as before.  The gas
is not heated prior to scrubbing.  The minimum reported mercury concen-
                                                  /Q \
tration in the scrubber effluent is about 10 ppbv.     In a current
operation where only depleted brine scrubbing is employed and gas cooling
to 80 F is undertaken (in the summertime) scrubber effluent mercury con-
                                                /io \
centrations of about 30 ppbv have been reported.^  '  The scrubbers are
                                                        (2)
usually packed columns with a titanium or rubber lining.
          However, for purposes of this study, the hydrogen stream at 60 F
is contacted with depleted chlorinated brine at 100 to 150 F in the brine
scrubber which is operated at an L/G ratio of about 10 to 12 gal per
       3
1000 ft .  This is followed by scrubbing with 10 percent caustic soda
to remove elemental chlorine that volatilizes into the gas in the depleted
brine scrubber.  A reported L/G ratio in the caustic scrubber is of the
order of 20 gal.  Depleted brine scrubbing is essentially a closed-loop
process.  The mercury is recovered in the brine treatment process and in
the mercury-cell.  Caustic scrubbing appears to consume very little caustic
soda because of the reportedly low chlorine concentrations in the gaseous
                                          /I O\
effluent from the depleted brine scrubber.      in any case, the caustic
scrubber liquid effluent is utilized in brine treatment.

-------
Depleted  brine
to treatment
Depleted brine
from mercury  cells
        Brine
        feed pump
                                                 Depleted
                                                 brine
                                                 scrubber
                                              Caustic
                                              scrubber
    T
   Tubular
   type mist
   eliminator
Secondary
(Brine)
cooler
Knock-out
 drum
Water
sealed
rotary
blower
Wire-mesh
pad mist
eliminator and
seal pot
                                                                                Tubular type
                                                                                mist eliminator
                                                   Caustic
                                                   feed pump
                                                                              Purified
                                                                              Hydrogen
                                                                              By-Product
                                                                               Stream
                                                                                Caustic
                                                                                return pump
                                                  Primary
                                                  (Water)
                                                  cooler
                                                   Dilute
                                                   caustic jn.

                                                         "
                                                                                Caustic out
                                                                                to brine
                                                                                treatment
Hydrogen
byproduct
stream from
decomposer
               FIGURE 4.   DEPLETED BRINE SCRUBBING FOR MERCURY  REMOVAL FROM
                            THE HYDROGEN  BY-PRODUCT  STREAM

-------
                                   20
          The projected advantages utilizing depleted brine scrubbing may
be summarized as follows:
          Advantages
1.  Process streams used to scrub          1.
    Mercury and chlorine are ob-
    tained from the chlorine/caustic
    manufacturing process to which         2.
    they are recycled after use.

2.  Mercury recovery is instantaneous.     3.
   Disadvantages


Mercury removal capabilities are
not proved at the present time.

Corrosion resistant materials of
construction are necessary.

Maintenance requirements are
relatively high because of (1)
above and because of corrosive-
fluid pumping requirements.
                  TECHNICAL ANALYSIS OF MERCURY CONTROL
                     BY MOLECULAR SIEVE ADSORPTION
                 General Description of Molecular Sieves
           Zeolites are crystalline hydrated alumino-silicates of the

 alkali and alkaline earth elements.  Structurally, they are framework

 alumino-silicates which are based on an infinitely extending three-

 dimensional network of A10, and SiO, tetrahedra linked to each other by

 sharing all the oxygens.  They may be represented by the general formula:

                        M2/n°'A1203'xSi02'yH2°

 where x is greater or equal to 2 and n is the cation valence.

           The framework contains channels and interconnected voids which
 are occupied by the cation and water molecules.

      . .    The term molecular sieve was originated by McBain to define

 porous solid materials which exhibit the property of acting as sieves on a
 molecular scale.

-------
                                21
           There  are  100  types  of  synthetic  zeolites.   Only a  few have
 practical  significance.  Uncontrolled dehydration  irreversibly alters
 the  framework  structure  and  the positions of  the cations  such that  the
 structure  collapses.   To be  used  as molecular sieves the structure of the
 zeolite  after  complete dehydration must  remain  intact.
           Adsorption takes place  on external  surfaces  as well as the
 internal convoluted  surface  of the dehydrated microporous zeolite crystal.
 Because  of the inherent  characteristics of the crystal structure of the
 zeolite, the adsorption  of a guest molecule depends partially upon  its
 polarity and polarizability.

                       Use of Molecular Sieves for
                  Mercury Removal from Gaseous  Streams

           The  adsorption of  mercury vapor on  zeolite molecular sieves has
 been known for a long time now.   Earlier work in this  area has been
                              (19)
 thoroughly reviewed  by Breck.      He reports that as  early as 1909,
 F. Grandjean found that  heated chabazite  (a Class  I, naturally occurring,
molecular  sieve with a narrow  interstitial channel diameter between 4.89A
and 5.58A)  had mercury adsorption capabilities.   Barrer and Woodhead^  ^'
concluded  that the high  adsorption of mercury vapor in air on this zeolite
was due  to a chemical  adsorption  influenced by the presence of oxygen.  The
adsorption process was found to be irreversible.  However, the adsorption
of mercury under conditions approaching vacuum was reversible and the
capacity (in grams mercury adsorbed per 100 grams chabazite) of the adsorbent
decreased with decreasing temperature.  Adsorption data obtained at elevated
temperatures showed  (after extrapolation) that the mercury adsorption
                                                  /      *2
capacity of chabazite  at 70 F was in the range 10   x 10   gms/100 gms.
               (21)
          Logan     studied the adsorption of mercury vapor in hydrogen gas
over a synthetic molecular sieve  (Union Carbide's 13X) and over chromic
acid-impregnated silica gel.   At a mercury concentration of about 4.0
           3                                                         3
mg Hg per m  and for breakthrough at a concentration of 0.001 mg  Hg/m  the

-------
                                     22

  capacity of the molecular sieve was found to be of the order of 10   to
  10~  gm Hg /100 gm.  The results for the molecular sieve suggest that
  water vapor did not seem to affect the adsorptive capacity although the
  heat of water adsorption caused a 90 F rise in temperature.  Furthermore,
  physical adsorption was seen to occur since the adsorptive capacity
  increased with a decrease in temperature.
            As mentioned earlier, a proprietary  synthetic molecular  sieve,
  manufactured by UCC,  is commercially available for adsorption  of mercury
                                                                      (22)
  vapor from the hydrogen by-product and end box ventilation streams.
  The adsorptive capacity of  this molecular sieve  is orders  of magnitude
  higher than that of  13X.
                   Performance of a PuraSiv Hg Mercury
                             Adsorption System
        The PuraSiv Hg adsorption process is guaranteed by the vendor (UCC)
to control the concentration of mercury in the hydrogen byproduct stream to
                                                     (22)
a maximum average of 60 ppbv for a period of 3 years.      The control
capability claimed for the system was tested by monitoring the mercury con-
centration in the streams influent to and effluent from the adsorption system
                                      (12)
installed on a 200 TPD chlorine plant.      The results of the test, carried
out two years after startup of the adsorption system, are summarized in Table
3.  The inlet and outlet concentrations are plotted  in Figures 5 through 9.
        For purposes of this study, a process that is considered applicable
for controlling the mercury content of the hydrogen byproduct stream should be
capable of reducing the concentration of the pollutant to below a time
average of 60 ppbv.  This limit has been arbitrarily set since (1)  the
vendor of PuraSiv Hg adsorption systems (UCC) claims that such a condition
is met by their system and (2) this limit can probably be achieved  by
alternative control technologies, namely, treated activated carbon  adsorption
and depleted brine scrubbing.

-------
                               TABLE 3.   SUMMARY OF TEST DATA OBTAINED FROM A PURASIV HG MERCURY ADSORPTION PROCESS
                                                                                                                   (a)
                                                   Adsorber Inlet Conditions
                                                                                                                 Adsorber Outlet Conditions


Cycle
1







2







3







4







5







(a) Source:
(b) Obtained
Time, hrs
during
24-hr
cycle
0-2
3.5-5.25
6.5-8.5
9.67-11.67
12.5-14.5
15.75-17.75
18.67-20.67
22-24
0-2
3-5
5.75-7.75
9-11
12-14
15-17
18-7.0
21-23
0-1.75
2-3.25
4.25-6.25
7.25-9.25
10.25-12.25
13.25-15.25
16.25-18.25
19.25-21.25
0-2
3-5
6-8
9-11
12-14
15-17
18-20
21-23
0-2
3-5
6-8
9-11
12-14
15-17
18-20
22-24
Reference 12.

Ave. ,
Pressure,
in., Hg
39.2
39.1
39.1
39.2
39.1
39.1
38.8
38.6
38.6
38.5
38.4
38.5
38.5
38.4
38.4
38.4
38.3
38.6
38.6
38.6
39.3
39.4
39.3
39.3
39.4
39.4
39.5
39.7
39.5
39.5
39.7
39.7
39.4
39.4
39.5
39.5
39.4
39.4
39.2
39.1

for mercury vapor pressure

. Ave . , ,
0 Te0p.,(a>
F
72
72
72
72
72
72
71
71
71
71
71
71
71
71
70
72
71
70
70
71
71
71
70
70
71
71
70
70
70
70
71
71
71
71
71
71
71
71
71
71

in mm Hg between
Measured
Ave., ,
Conc.,U;
ppbV
963
938
876
900
1,146
1,219
958
965
976
954
846
841
840
907
992
668
1,053
837
828
707
806
761
878
949
945
727
873
923
787
830
937
1,067
1,243
1,054
1,267
1,176
1,155
1,231
1,058
1,145

32 and 300
Hg
Cone, at,, \
Saturation, '
ppbv
1,397
1,400
1,400
1,397
1,400
1,400
1,346
1,353
1,353
1,357
1,360
1,357
1,357
1,360
1,297
1,426
1,364
1,291
1,291
1,353
1,329
1,326
1,268
1,268
1,326
1,326
1,261
1,255
1.261
1,261
1,316
1,316
1,326
1,326
1,323
1,323
1,326
1,326
1,333
1,336

F. log.nP - -3212.

Flow .
Rate/c)
scfm
2,431
2,412
2,467
2,513
2,461
2,414
2,383
2,386
2,386
2,377
2,354
2,350
2 , 350
2,350
2,285
2.330
2,395
2,334
2,351
2,403'
2,356
2,323
2.325
2,370
2,378
2,382
2,483
2,567
2,515
?.,!>(•?.
2,483
2,446
2,408
2,406
2,397
2,368
2,373
2,324
2,322
2,296

5/T + 8.025,

Ave'(a)
Temp.,
F
95.5
95.5
95.5
95.5
95.5
95.5
96.0
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.0
95.0
95.0
95.0
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5
95.5

where T is

Ave . , .
Pressure,13'
in. Hg
38.1
38.1
38.1
38.1
38.1
37.9
38.1
38.1
38.1
38.1
38.1
37.9
36.6
37.9
37.9
37.9
38.1.
39. 3 e
39. 3 e
39.6(e)
38.9
38.9
38.9
38.9
39.1
39.1
39.1
38.9
38.9
3P.9
38.9
38.9
39.1
39.1
39.1
39.1
39.1
38.9
38.9
38.9

°K, Kirk & Othmer
Measured
Cone . ,
ppbv
55.4
3.1
<0.1
0.3
<0.1
1.1
95.9
429.5
57.4
53.6
0.4
0.2
0.2
0.4
26.4
38.6
55.0
1.9
0.4
0.1
0.1
0.2
0.2
8.1
4.7
0.2
0.1
0.1
0.1
0.1
3.6
120.1
37.4
0.2
0.2
0.1
0.1
0.1
11.0
425.9

Flow Rate
to Boiler
Vent."7
scfa
1.466
1.457
1.457
1,452
1,452
1,452
1,452
1,454
1,465
1,466
1,466
1,466
1,357
1,221
1,221
1,349
1,457
1,465
1,466
1,466
1,466
1,414
1.191
1,468
1,459
1,465
1,466
1.467
1,465
1,459
1,457
1.457
1,457
1,457
1,457
1,457
1,457
1.457
1,457
1,457

Encyclopedia of
     Chemical Technology, Second Edition, 1967, Volume 13, p 221.
(c)  Obtained from Reference 2 but adjusted to standard conditions at 1 atm and 60 F.
(d)  Obtained from data on chlorine production reported in Reference 12 and by assuming a factor of 180.16 scfm per ton per hour.
(e)  These values are incorrect.

-------
 10,000
   1000
X)
a.
o.
o
5
O>
X
                                    Cycle:!

                                    Inlet :D

                                    Out let :O
UCC 3-year guaranteed average

outlet concentration
                                                                                     ei
                                                                                     w
                                                                                     PQ
                                                                                     td
                                                                                     M
                                                                                     C/3
                                                                                     OS
                                                                                     CJ
                                                                                     Cd

                                                                                     i
§
PM

Z
O
                                                                                     W
                                                                                     u
                                                                                     o
                                                                                     CJ
                                                                                     H
                                                                                     3
                                                                                     O

                                                                                     O
                                                                                     2
                                                                                     O
                                8     10     12     14     16

                              Absorption Cycle Time.hrs

-------
 10,000
                                       25
                                   Legend


                                    Cycle: 2

                                    Inlet :D

                                   Outlet : O
             FIGURE 6.  INLET AND  OUTLET MERCURY CONCENTRATION

                        PROFILES FOR  A MOLECULAR SIEVE ADSORBER
   1000
                                                       Ed
                                                       CQ

                                                       Oi

                                                       O
                                                       CO
                                                       Q
                                                                                     Ed


                                                                                     td
    O.I
 I
I
I
I
                               8     10    12    14    16

                                Adsorption  Cycle Time.hrs
                               18    20    22    24
                                                                                     Id



                                                                                     I
    100
.a
a.
Q.
o
c
o
o

o>
X
UCC  3-year guaranteed average

outlet concentration
     10
                                                                                     CO
                                                                                     Id
                                                                                     M

                                                                                     Cn

                                                                                     O
                                                                                      z
                                                                                      O
                                                       u
                                                       §
                                                       O

                                                       I
                                                       I
                                                                                      H
                                                                                      Ed
                                                                                      J
                                                                                      H


                                                                                      O
    1.0
                                                                                      H
                                                                                      Id
                                                                                     z
                                                                                     vD


                                                                                     Cd


                                                                                     3

-------
  10,000 r
    1000
     100
n
CL
o.
o
c
o
o
                               INLET AND OUTLET MERCURY CONCENTRATION

                               PROFILES FOR A MOLECULAR SIEVE ADSORBER
                                         Cycle :  3

                                          Inlet :  D

                                        Outlet :  O
                          UCC 3-year guaranteed average

                          outlet  concentration
ai
w
to
cm
o
CO
a
                                                                                    Cd

                                                                                    w
                                                                                    u
                                                                                    <


                                                                                    o
                                                                                    CO
                                                                                    u
§
a.



§
                                                                                    o
                                                                                    o



                                                                                    I
                                                                                    o
                                                                                    Di
                                                                                    w


                                                                                    H
                                                                                    w

                                                                                    H

                                                                                    O

                                                                                    Q
                                                                                    H
                                                                                    W
                                                                                    r^

                                                                                    w
                                                                                    8
     o.
                                8  .   10 .   12 ,  14   . 16  .

                               Adsorption Cycle Time, hrs

-------
 10,000
                         FIGURE 8.  INLET AND  OUTLET MERCURY CONCENTRATION
                                    PROFILES FOR A MOLECULAR SIEVE  ADSORBER
                                          Cycle
                                          Inlet
                                         Outlet
                       4
                       d
                       o
  1,000
    100
.0
a
o.
u
o
,UCC 3-year guaranteed average
 outlet   concentration
     10
        -A
          \
     i.o
     O.I
                                             I
                                     I
                                            12     14     16    18

                                      Adsorption Cycle  Time,  hrs
                                          20
22
24
26    28

-------
10,000
                  FIGURE 9.
  1000 —
   100
.a
a.
a.
u
c
o
o
CT
X
    	28   	

     INLET AND OUTLET MERCURY CONCENTRATION
     PROFILES FOR A MOLECULAR SIEVE ADSORBER

                Cycle:   5
                Inlet:  D
               Outlet:  O
UCC 3-year guaranteed average
outlet concentration
                                           I
                               8     10    12    14    16
                                 Adsorption Cycle Time, hrs
                                    18
20    22
24

-------
                                  29
        UCC's claim was tested through measurements of mercury inlet and
outlet concentrations in a PuraSiv Hg adsorption system used to control
mercury emissions in a 200 TPD plant.  With the exception of the third cycle
the measurements of concentration were taken over five characteristic 24-hour
cycles of adsorption.  The third cycle was terminated after 22.5 hours
because of an unscheduled plant shutdown.  Mercury outlet cycle - average
concentrations were obtained by integration of the profiles given in Figures
5 through 9.  The procedure is explained in Appendix A.  The amount of
mercury emitted was calculated from this average concentration and the
average hydrogen flow leaving the system.  The results are given in Table 4.
Upon inspection of the values of average concentration given in Table 4, it
may be seen that except for the first two cycles, UCC's guarantee of the
outlet concentration was met after 2 years of service.  The average concen-
tration was 65 percent higher than the guaranteed 60 ppbv for the first cycle
and 25 percent higher for the second cycle.
         The PuraSiv Hg  system which  was  tested  on behalf  of  EPA  as  a
 background to this  analysis  was  operated at mercury  levels beyond those
 normally expected in a  mercury-cell  chlorine  plant of the stated capacity—
 namely 200 TPD.  This situation  arose as a result of complications  caused
 by an anticipated expansion  of the chlorine-producing capacity of the
       (27)      ^
 plant.      As a result,  abnormally  large quantities of hydrogen were being
 recycled.   At the time  the testing was in progress the hydrogen  flow to the
                                                               (13)
 adsorber was about  2400 SCFM.  UCC-recommended  design criteria      pre-
 dicted a flow of about  1700  SCFM.  Furthermore,  secondary cooling to about
 72 F was employed instead of the recommended  cooling to 60 F.  These two
 factors tended to increase the amount of mercury flowing  into the adsorber.
 Since the hydrogen  stream always is  saturated with mercury,  the  increased
 flow through the cooling/demisting equipment  causes  an increase  in  the
 total amount of mercury to the adsorber.  It  is estimated that about 150
 percent more mercury was  carried to  the  adsorber as  a result of  this mode
 of operation.
         The PuraSiv Hg  system that was the subject of EPA testing reportedly
 has been overdesigned with respect to the amount of  adsorbent used  therein.
 Although this has been  confirmed by  the  system  designer-UCC, no  information
 on the system "overdesign" has been  made available.

-------
                    30
TABLE 4.  MERCURY EMISSIONS FROM A 200 TPD
          MERCURY-CELL CHLORINE PLANT

Number of
24-Hour
Test
Period
1
2
3
4
5
24-Hour
Average
Chlorine
Production,
rate
TPD
194
183
190
198
194

24-Hour
Mercury
Emission,
gm
49.6
35.4
16.3
19.1
26.0

24-Hour
Average
Mercury
Concentration,
ppbv
99
75
33
37
51

-------
                                  31
           The results of testing indicate that the guaranteed average mercury
 concentration of 60 ppbv was being achieved by the adsorption system after 2
 years of operation.  However, it was not possible to verify UCC's claim that such
 effluent mercury levels could be achieved after 3 years of system operation.
        The mercury inlet concentrations obtained from Reference 12 are
too low as a result of experimental error.  This is immediately apparent
upon comparison of the data on inlet concentration with the saturation
value at the reported inlet stream temperature and pressure.  The measured
value should be higher than the saturation value.
                      Factors  Affecting Performance  of
                           Molecular Sieve  Mercury
                               Control  Systems
            The performance of  the molecular  sieve  adsorbent  used  in  control
  of mercury in the hydrogen stream is  affected  mainly by  (1)  the  number  of
  adsorption/regeneration cycles  to which it  has been subjected  and  (2) the
  concentrations of mercury and water vapors  in  the influent  stream.
            The adsorption system design is such that a  maximum  average
  effluent mercury concentration  of 60  ppbv is guaranteed  for a  period of
          (22)
  3 years.      With the passage  of time this average concentration of the
  effluent is expected to gradually increase  from a level  which  is well
  below that guaranteed to one  which is close to it.  This loss  in performance
  is directly caused by the gradual, but cumulative,  destruction of the micro-
  pores of the adsorbent as a result of its exposure to  the relatively high
  temperature swings between the  regeneration temperature  of  about 500 F  and
  the adsorption temperature of about 100 F.  Investigators of the Bureau of
  Mines     studied the sulfur  dioxide  adsorption/desorption  characteristics
  of UCC synthetic molecular sieve 13X.  The  adsorption  experiments were
  carried out at a temperature  of 79 F  and a  total  pressure of 26.38  in.  Hg
  for sulfur dioxide concentrations ranging from 1  to 9  percent  by volume.
  The space velocity employed was of the order  of 400 vol/vol/hr.   The
  time for complete saturation  or loading of  adsorbent with sulfur dioxide

-------
                                 32
 was a.bout 36 minut.es.   The observed loading ranged from 27 to 30 Ib S0_
 per 100 Ib adsorbent, being higher  for the  feed richer in  SCL.  Desorption
 .(or regeneration) for  about 2.3 hours was carried out at tmperatures of
 about 700 F while the  space velocity of  nitrogen purge gas was maintained
 at 4 vol/vol/hr.   A 60 percent drop in the capacity of 13X was observed after
 13 cycles of adsorption/regeneration.  Thereafter the capacity of the molecular
 sieve seemed to stabilize at the lower level.   While no independently obtained
 information on the long-term effects of  cyclic heating  (regeneration)  is
 available for the molecular sieve under  consideration,  field  tests carried
 out by UCC indicate a  10 percent loss in loading  capacity  over a  period of
 2 years.   Two types of loss in activity  may occur for the  adsorbent.   The
 first may be a short-term loss taking place during earlier cycles as.observed
 for 13X.   The second may be a long-term  loss as reported by UCC.   Both of these,
 however,  can be anticipated by system design.   Analysis of data on effluent
 mercury concentration from a molecular sieve adsorption unit  indicates that the
 design does anticipate this loss in activity as evidenced by  cycle
 average concentrations that are usually lower than that guaranteed [60 ppbv]
i by UCC.  Field tests indicate that a 2-year sieve life is demonstrated.
          Unspecified concentrations of water vapor in the gas stream
 undergoing adsorption affect the performance of the molecular sieve by
 reducing its selectivity toward mercury.  The  condition of relative humidity
 of the hydrogen stream to the adsorber is considered proprietary  by UCC.
 High water concentrations increase the amount  of  adsorbent needed to remove
 a;given amount of mercury.   While no data on the  effect of relative
 humidity  on adsorption capacity are available,  it has been found  that  a
 design relative humidity of 50 percent does not adversely  affect  the
 selectivity to the point where "uneconomical"  loadings  are produced.

-------
                                 33
                            ECONOMIC ANALYSIS

          In this section an evaluation of the economic feasibility of
mercury control by PuraSiv Hg adsorption is developed and discussed.
While a comparison of the capitalized costs of this system with other
potentially available alternatives (treated activated carbon adsorption
and depleted brine scrubbing) would be useful in assessing their economic
feasibility, the lack of an adequate data base on the cost and perfor-
mance of alternative systems does not permit the development of such a
comparison.

                              Capital Costs

          The capital costs of control by PuraSiv Hg adsorption were
generated from data available on actual operating systems and from data
provided by UCC.  Where actual economic data were lacking, conceptual
designs based on known process operating characteristics were performed.
                                                                       (24 25)
In these conceptual designs the module technique recommended by Guthrie   '
was used.  Briefly, this approach to capital cost estimation involves the
use of "experience" multipliers to obtain the installed costs of major
components of a processing system from known purchased equipment costs.
As explained in Appendix B, the installed equipment costs obtained in
this manner include the costs of foundations, piping, instrumentation,
insulation, and paint.  The estimates of capital cost reported would
normally be correct to within a ±30 percent range that is characteristic
                  (26)
of such estimates.      The costs reported are end-of-1974 estimates for
a value of the Marshall and Stevens Index of 398.  Where conceptual
designs were necessary (as in the case of cooling/demisting) the capital
costs were estimated for plants of 100 and 750 TPD capacity, this range
representing 87 percent of the mercury cell-chlorine plants.
          A conceptual design for the cooling/demisting step was performed
in order to obtain the incremental capital and operating
costs incurred by recycling the regeneration gas in PuraSiv Hg adsorption.
The cost estimates obtained from the conceptual design were about 50

-------
                                   34

                                         (13)
percent lower than those supplied by UCC.      The actual cost for a
PuraSiv Hg adsorption system was found to be 70 percent higher than the
estimates reported by UCC for a 750 TPD plant.      The estimated
capital costs for cooling and demisting and PuraSiv Hg adsorption are
given in Table 5.  Appendix B outlines how they were obtained.
After the addition of the incremental (15 percent) costs of higher
cooling and demisting capacity which is necessary in the case of
PuraSiv Hg adsorption, it was found that the capital costs were in
the range of 853 and 313 dollars per TPD for plants with chlorine
capacities ranging from 100 to 750 TPD.

                     Operating and Maintenance Costs
          The operating and maintenance (O&M) costs  include those incurred
annually for (1) operating labor, (2) cooling water, electricity, and
steam, (3) raw materials and chemicals, (4) local property taxes and
insurance, and (5) maintenance labor and materials.  Details  of the esti-
mated O&M costs for cooling/demisting and PuraSiv Hg adsorption and the
method by which they were obtained are given in Appendix B.
          The PuraSiv Hg adsorption process is a mercury recovery
process.  Therefore, a credit for mercury recovery was included.  Since
primary and secondary cooling are necessary precursors to the adsorption
processes, a mercury credit was not taken for the former.  The O&M
costs of PuraSiv Hg adsorption (including costs of additional cooling/
demisting capacity) were found to vary between 0.24 and  0.14 dollars
per ton of Cl_ for plants in the chlorine capacity range of 100 to 750
TPD.  A summary of O&M costs is given in Table 6.               .

                Total Annual Costs of PuraSiv Hg Adsorption

          The total annual costs of control by PuraSiv Hg adsorption
consist of the sum of the operating and maintenance costs, adsorbent
replacement costs, and costs incurred for depreciation of the capital
investment over its life.  Furthermore, 15 percent of the depreciation

-------
                                           35
                     TABLE 5.   CAPITAL COSTS OF COOLING/DEMISTING
                               AND PURASIV HG ADSORPTION
Capital Costs, $1000
Cl2 Capacity
TPD
100
250
500
600
750
Thousand
Tons per
Year
34.7
86.7
173.4
208.0
260.0
Cooling/
Demisting
143.0
269.6
435.6
494.1
576.6
Incremental Cooling/
Demisting
Costs Attributed
to PuraSiv HgOO
14.3
26.7
43.6
49.4
57.7
PuraSiv Hg
Adsorption
71.0
107.6
147.4
160.1
177.2
Total
PuraSiv Hg
and Incremental
Cooling/Demisting(b)
85.3
134.3
191.0
209.5
234.9
(a)   Obtained by computing the difference  in capital  costs  between a plant of the
     stated capacity and another that  is 15  percent larger.   Thus  the incremental
     cost at 100 TPD is  5907(100)°-692[ (1.15)0'692- 1].

(b)   Total includes  PuraSiv Hg adsorption  and incremental cooling/demisting at
     stated capacity.  Depreciation costs  computed from  numbers  in this  column
     are given in Table  6.

-------
                              TABLE 6.  O&M COSTS OF PURASIV ADSORPTION

Cl Capacity
Thousand
Tons per year
34.7
86.7
173.4
208.0
260.0

Depreciation
$1000)
20.3
37.6
62.9
72.7
86.2

Costs of
PuraSiv Hg
Adsorption
Per Ton
of C12, $
0.58
0.43
0.36
0.35
0.33
(a)   Depreciation computed using capital  costs  in  Column 6  of  Table 5  assuming
     10 year period.

(b)   Includes depreciation.
                                                                                                         OJ

-------
                                   37

cost of cooling/demisting needs to be added for reasons previously
mentioned.
          The capital investment for all processes was depreciated over
a period of 10 years.  This is an average equipment life in the chemical
                   ( 26)
process industries.      The length of the adsorbent guarantee of 3
years was taken as the life of the adsorbent.   Table 6 also gives the variation
with capacity of the annual costs and the costs per ton of chlorine
produced.  In the capacity range of 100 to 750 TPD it is seen that the
costs of PuraSiv Hg adsorption per ton of Cl. will vary between $0.58
and  $0.33.

                     CONCLUSIONS AND RECOMMENDATIONS

          The data generated from EPA-sponsored tests together with
other available information seem to indicate that PuraSiv Hg adsorption
is effective in the removal of mercury vapor present in concentrations
of the order of 1500 ppbv from the hydrogen by-product stream obtained
in the manufacture of chlorine by the mercury-cell process.  Removal
to levels below 60 ppbv was observed in the above-mentioned tests.
There is evidence that sulfur-impregnated activated carbon absorption
and depleted brine scrubbing may be effective in removal of the pollu-
tant from the hydrogen by-product stream.  However, testing of the other
systems for performance is believed to be necessary at this time in order
to clearly define those technologies and to establish their capabilities
to decrease mercury concentrations to desirable levels (below 60 ppbv).
Operators and vendors of treated activated carbon control systems claim
that such levels of mercury concentration are easily achievable.  Like-
wise the operator of a depleted-brine scrubbing system (in-house design)
claims maximum effluent mercury emissions of 0.03 Ib per day from the
hydrogen generated in a 150 TPD chlorine plant (about 30 ppbv).  Apparently
this is being achieved with one-step cooling of the stream to 75 to 80 F.
It also is claimed that chlorine concentration in the gaseous effluent
from the depleted brine scrubber is so. low that caustic scrubbing has
been deemed unnecessary.
          When the operator of a mercury-cell chlorine process is faced
with the problem of choosing a mercury control process, the factors which

-------
                                  38
are most likely to influence his decision as to which control process is
most attractive are
           (1)  Performance
           (2)  Reliability
           (3)  Economic feasibility relative to other processes
           (4)  Ease of operation and flexibility of design.
          As to performance, PuraSiv Hg adsorption does seem to achieve
average mercury effluent concentrations lower than 60 ppbv after two
years of operating such an adsorption system.  This has been observed in
the testing of a unit designed to control mercury vapor in a 200 TPD
chlorine plant.  This has been found to be true from data obtained by
the monitoring of adsorber effluent mercury concentration in three out of
five adsorption cycles.  In the two cycles where the guaranteed level of
60 ppbv was not achieved, the departure amounted to 60 percent in the first
and 25 percent in the second.  This departure may be attributed to either
operating the adsorption unit at levels above those recommended by the
design value with respect to hydrogen flow rate and to insufficient
cooling of the hydrogen stream prior to mercury adsorption at the time of
testing.  Cooling to 71 F was employed instead of the recommended 60 F.
Evidence accumulated from the operators and/or vendors of depleted brine
and treated activated carbon adsorption systems indicates that these
processes are potentially capable of achieving the same level of control
as PuraSiv Hg adsorption.
          From available information on the PuraSiv adsorption system, an
availability greater than 95 percent may be inferred.  The process involves
a relatively simple equipment layout and the materials of construction are
not highly specialized.  Carbon steel is used throughout.  The adsorption
system is easy to operate and can be designed in such a manner that a high
degree of automation may be achieved.
          In summary, it is concluded that
           (1)  PuraSiv Hg adsorption is a technically viable
               and feasible mercury control alternative of
               the end-box and hydrogen by-product streams in
               mercury-cell chlorine plants
           (2)  Depleted brine scrubbing and treated activated
               carbon adsorption are potentially feasible mercury
               control processes in mercury-cell chlorine plants

-------
                                    39
          (3)  The economics of PuraSiv Hg adsorption are such
               that the total annual costs are in the range of
               0.58 to 0.33 dollars per ton of Cl, for plant
               capacities in the range of 100 to 750 TPD with
               respect to chlorine production.  These cost
               estimates are based on an adsorbent life of three
               years as guaranteed by UCC.  The vendors guarantee
               of adsorbent life could not be independently
               verified.  However, in EPA testing of a PuraSiv Hg
               system, the adsorbent was found effective after
               two years of system operation.

On the basis of this study it is recommended that an engineering analysis

similar to the study at hand be performed for depleted brine scrubbing
and treated activated carbon adsorption.

-------
                                    40
                                REFERENCES
 (1)  Shreve, R. N., "Chemical Process Industries", Third Edition, McGraw-
      Hill, New York (1967).

 (2)  Anonymous, "Atmospheric Emissions from Chlor-Alkali Manufacture", A
      cooperative study project MCA & PHS, Publ. No. AP 344, EPA
      (January, 1971).

 (3)  Anonymous, "Development Document for Effluent Limitations Guidelines
      and New Source Performance Standards for the Major Inorganic Products
      Segment of the Inorganic Chemicals Manufacturing Point Source Category",
      Reported prepared for Office of Air and Water Programs, U.S. EPA,
      Publ. No. EPA-440/l-74-007-a (March, 1974).

 (4)  Stambaugh, E. P.  and Hall, E. H., "Topical Report on Basis for
      National Emissions Standards on Mercury", to Office of Air Programs,
      U.S. EPA, Contract No. EHSD 71-33 (June 15, 1971).

 (5)  Anonymous, "The Cost of Clean Air, 1974", Final Report to U.S. EPA
      by Battelle's Columbus Laboratories under Contract No. 69-01-1538
      (January 15,  1974).

 (6)  Anonymous, "Can Chloride Cutbacks Hold the Price Line?", Chemical
      Week (March 12, 1975).

 (7)  Laubusch, E., The Chlorine Institute,  personal communication
      (January and February, 1975).

 (8)  Anonymous, "Control  Techniques for Mercury Emissions from Extraction
      and Chlor-Alkali  Plants", U.S.  EPA, Office of Air Programs,  Research
      Triangle Park,  North Carolina,  Publ. No.  AP-118 (February, 1973).

 (9)  Federal Register, _38 (66), 8831 (April 6, 1973).

(10)  Anonymous, American  Metal Market (October 29, 1974).

(11)  Anonymous, "Engineering Analysis of Emissions Control Technology for
      Sulfuric Acid Manufacturing Processes",  Final Report to PHS, US DHEW,
      Contract CPA22-69-81 (March, 1970).

(12)  Chehaske, J.  T. and  Cline, J. R.,  "Testing of a Molecular Sieve Used
      to Control Mercury Emissions from a Chlor-Alkali Plant", Report to
      EPA, Control  Systems Laboratory, Research Triangle Park, North
      Carolina, Contract No. 68-02-1406,  Task Nc. 3 (January, 1975).

(13)  Based on information made available to Battelle's Columbus Laboratories
      by Union Carbide  Corporation in January and February,  1975.

-------
                                    41


                                REFERENCES
                                (Continued)
(14)   Istas,  L.  J.,  Manager,  Engineering, Pennwalt Corporation, Calvert
      City,  Kentucky, letter to M.  Y.  Anastas (February 5, 1975).

(15)   Cunniff,  F.  T., Manager - Product Development, Calgon Corporation,
      Pittsburgh,  Pennsylvania, letter to M.  Y.  Anastas (March 26, 1975).

(16)   Dreibelbis,  J. A.  and Joyce,  R.  S., "Method of Removing Mercury
      Vapor  from Gases", U.S. Patent 3,194,629 (July 13, 1965).

(17)   Clapperton,  J. A., Manager of Manufacturing Industrial Chemicals,
      PPG Industries, Pittsburgh, Pennsylvania,  letter to M. Y. Anastas
      (February 28,  1975).

(18)   Spomer, M.,  BASF Wyandotte, Port Edwards,  Wisconsin, communication
      with M. Y. Anastas (April 3,  1974).

(19)   Breck,  D.  W.,  "Zeolite Molecular Sieves -  Structure Chemistry and
      Use",  John Wiley,  New York (1974).

(20)   Barrer, R. M.  and Woodhead, M.,  transactions of the Faraday Society,
      .44, p  1001 (1948).

(21)   Logan,  W.  R.,  "Mercury Removal from Hydrogen Gas Streams, Journal
      of Applied Chemistry, 1.6_, p 285  (October,  1966).

(22)   Collins,  J.  J., Miller, W. C., and Philcox, J. E., "The PuraSiv Hg
      Removal and Recovery",  paper presented at  the 65th Annual Meeting of
      the Air Pollution Control Association,  Miami Beach, Florida
      (June,  1972).

(23)   Martin, D. A.  and Brantley, F. E., "Selective Adsorption and Recovery
      of Sulfur Dioxide from Industrial Gases by Using Synthetic Zeolites",
      U.S. DOI,  Bureau of Mines, RI 6321 (1963).

(24)   Guthrie,  K.  M., "Capital Cost Estimating", Chemical Engineering
      (March 24, 1969).

(25)   Guthrie,  K.  M., Process Plant Estimating Evaluation and Control.
      Craftsman Book Company of America, Solana  Beach, California (1974).

(26)   Peters, M. S.  and Timmerhaus, K. D., Plant Design and Economics for
      Chemical Engineers, 2nd Edition, McGraw-Hill, New York (1968).

-------
              APPENDIX A
CALCULATION OF AVERAGE MERCURY CONCENTRATIONS
AND AVERAGE DAILY EMISSIONS IN MOLECULAR SIEVE
            ADSORBER OUTLET

-------
                              APPENDIX A

             CALCULATION OF AVERAGE MERCURY CONCENTRATIONS
             AND AVERAGE DAILY EMISSIONS IN MOLECULAR SIEVE
                            ADSORBER OUTLET
                                                                        3
          Data on the 2-hour average mercury concentration (in rag per Nm )
in the adsorber outlet obtained from Reference 12 was plotted versus
cycle time on semi-log graph paper.  The average concentration as reported
was considered to represent the average time in the interval.  A smooth
curve was drawn through the points and extrapolated to both ends of the
cycle, namely, 0 and 24 hours.  The extrapolation was such that the cycle
end points of the outlet concentration profile did not exceed the inlet
concentration.  Plots of the outlet concentration profiles for cycles
1 through 5 are given in Figures A-l to A-5.
          The average concentration was obtained by integration of the
outlet concentration profile according to the expression
                            f&2
          H    = _1	         H(e)d0           (A-l)  .
                  Vei.    ^
Where H is the 2-hour average concentration and Q is the cycle time in hours.
Linear portions of the semi-log plot were fitted to the equation

          H(e)=E1010                           (A.2)   f
Integration of this function gives
                  ft
          I =    f 2  H(e)d0 = 2 1;031   (10192 - 1019l)     (A-3)  . ,
               el               '            .
          Where the curve is not  linear it was divided into smaller linear
segments to which the equations given above were applied.  The cycle average
concentration was obtained by summing the I's and dividing by the cycle time
of 24-hours.  An example calculation is given for cycle 2 in Table A-l.  The
                                             3
average concentration obtained was 0.62 mg/Nm „

-------
                                         A-2
o
o
c.

0
o
o
    O.I —
    0.01 —
   0.001
                                8     10    12     14     16
                                 Hours in Cycle  (24 hrs total)
                                        18
20
22
24
          FIGURE A-l.
EFFLUENT MERCURY CONG. FROM  PURASIV CONTROL UNIT
   Date:   24-25, 1975
   Cycle:  1
   Plant:  Sobin - Orrlngton, Maine

-------
                                       A-3
o

8
I
E
6
§
o
 0.001
                                                   Ave inlet cone. - between
                                                   21 and 23 hours
   0.01 —
                               8     10    12    14     16
                              Hours in Cycle (24 hrs total)

           FIGURE A-2.  EFFLUENT MERCURY CONC. FROM PURASIV CONTROL UNIT

                          Date:   25-26, 1975
                          Cycle:  2 - Adsorber 2
                          Plant:  Sobin - Orrington, Maine

-------
                                            A-4
O'
E
i_
o
c
o>

E

d
c.
O
o
 0.001
                                             -   extrapolation
  O.Ol  —
                               8"    10   " 12    14     16

                              Hours in Cycle (24 hrs total)

                FIGURE A-3.   EFFLUENT MERCURY CONG.  FROM PURASTV CONTROL UNIT


                               Date:    26-27, 1975

                               Cycle:  3

                               Plant:  Sobin  - Orrington,  Maine

-------
      10
                                        A-5
                                 Note:	extrapolation
      1.0
 o
 E
 t_
 o
10
 c'
 o
 o
     O.I

     0.01
    0.001
J—i—ua
                                8     10     12     14     16


                              Hours .in Cycle (24 hrs total)



            FIGURE A-4.  EFFLUENT MERCURY CONG. FROM PURASIV CONTROL UNIT


                           Date:    September 27-28, 1974

                           Cycle:   4

                           Plant:   Sobin - Orrington, Maine

-------
                                       A-6
     10
     1.0
o
E
o
o>
u
c
o
O
    0.01
   0.001
o
              2     4     6     8     10    12     14     16
                               Hours in  Cycle  (24 hrs  total)

          FIGURE A-5.   EFFLUENT MERCURY CONG.  FROM PURASIV CONTROL UNIT
                          Date:    September 28-29,  1974
                          Cycle:   5
                          Plant:   Sobin  - Orrington, Maine

-------
                                   A-7
              TABLE A-l.  CALCULATION OF AVERAGE MERCURY
                          CONCENTRATION FOR CYCLE
Segment
1
2
3
4
5
6
7
8
61
0
4
6
8
16
18
20
22
92
4
6
8
16
18
20
22
24
Hl
0.48
0.48
0.006
0.002
0.002
0.06
0.65
3.4
H2
0.48
0.0062
0.0022
0.002
0.06
0.65
3.4
5.5
I
1.920
0.218
0.008
0.016
0.034
0.495
3.320
8.848
El = 14.86

Have = 14.86 =0.62 mg/Nm3
         24

-------
                                   A-8
          The daily mercury emission was obtained by multiplication of the
average cycle concentration with the total hydrogen gas flow leaving the
process computed from the cycle average chlorine production in tons per
hour.  Thus for cycle 2:

          Nm  [   70 F and 1 atm] per day
              =  7492.0 X (TPD) average
              =  7492.0 X 7.64
              =  57,240
                            3
grams Hg per day =H    xNm  per day
                    El Vc
                 = 0.62 *. 57.240
                       1000
                 = 35.4
ppbv - Hg        =120.65 XHave
                 = 75
For the other cycles the results are given in Table 3.

-------
                APPENDIX  B
CAPITAL. OPERATING AND MAINTENANCE. AND ANNUAL
      COSTS OF MERCURY CONTROL PROCESSES

-------
                               APPENDIX B
             CAPITAL. OPERATING AND MAINTENANCE..AND.ANNUAL
                   COSTS OF MERCURY CONTROL PROCESSES
                              Capital Costs

          Where actual economic data were lacking, conceptual designs
based on schematic flow sheets were performed for the mercury removal
processes considered.  The capital costs were obtained from sizing
process equipment for a 100 TPD plant and obtaining the Installed costs
                                                     (12)
of major equipment using Guthrle's modular technique.  '    The
equipment cost for a 750 TPD plant was obtained in a similar manner.
The capital costs for the 100 and 750 TPD cases were fitted to the
equation:
                              C = A-L (TPD)n                           (B-l)
The conceptual designs were carried out for cooling/demisting (without
recycle) and molecular sieve and treated activated carbon adsorption only.
The capital costs for depleted brine scrubbing were obtained from published
data(3) and data supplied by PPG.^

Cooling/Demi sting

          This step involves the use of five major items of equipment.
The first was cooling of the hydrogen by-product stream from the decomposer
at 180 F to 110 F at a pressure of 1 atm, using 80 F cooling water, in a
shell and tube heat exchanger.  The second is a wire-mesh pad mist
separator which in sulfuric acid mist removal applications is capable of
removing nearly 100 percent of mist particles 3 microns in diameter or
larger but removes only 15 to 30 percent of particles less than 3 microns
in diameter.  The third item of equipment is a water-sealed rotary blower
followed by a knock-out drum in which pressure of the gas is increased to
about 20.7 psia.  The gas is then cooled to 60 F in a shell and tube
exchanger using a chilled brine as coolant.  A tubular-type mist eliminator

-------
                                   B-2
which reportedly removes 99+ percent of particles smaller than 3 microns
follows the secondary cooler.
          A list of equipment and associated costs appears in Table B-l.
All costs were updated to a Marshall and Stevens installed equipment
index of 398.  Unless indicated otherwise all costs were obtained from
Reference 25 in the text.  The capital costs obtained for cooling/ demisting
were found to follow relation:
                          CCD = 5907 (TPD)0'692                       (B-2)
where C   is the cost in dollars and TPD is the daily chlorine capacity
       UD
in tons.  This capital cost/capacity relation is graphically displayed
in Figure B-l.

PuraSiv Hg Adsorption

          The PuraSiv Hg adsorption system consists of (1) a steam heated
preheater in which the gas to the adsorber is heated to approximately 100 F and
(2)  two vessels in which at any given time adsorption of mercury is taking place
in one and regeneration of the adsorbent is taking place in the other.
A portion (15 percent assumed) of the gas leaving the adsorber is heated
to a temperature above 500 F in an electrical heater and routed to the
vessel undergoing regeneration.  The gas effluent from this vessel is cooled
to more than 100 F in a shell and tube heat exchanger using 80 F cooling
water.  A list of equipment and estimated costs by the Guthrie method
appears in Table B-2.
          The costs supplied by UCC rather than those generated through the
conceptual design were used.  A linear log-log relation was assumed between
the 100 and 750 TPD cases.  This led to the expression:
                                          0
                          CL-, = 8775 (TPD)U*   .                      (B-3)
                           Mb                                '
          Because of the recycle of regeneration gas through the
cooling/ demisting module, a 15 percent increase in the capacity of that
unit will be required.  Therefore, the capital costs of PuraSiv Hg
adsorption at any plant capacity are those given by Equation B-3 in

-------
                               B-3
TABLE B-l.  CAPITAL COSTS OF COOLING/DEMISTING AND LIST OF EQUIPMENT

Primary Cooler
Wire Mesh Pad
Demister
Rotary Blower
Secondary Cooler
Refrigerator
Tubular Type
Demister
Bare Module Cost
Total Module Cost
(18 percent
contingency)
100 TPD
Cost,
Size $1,000
400 ft2(a) 16
900 cfm 18(b)
900 cfm 12(c)
100 ft2(a) 4
25 ton 45
530 cfm 26(b)
121
143

750 TPD
Cost,
Size $1,000
3,000 ft 62
6,750 cfm 65
6,750 cfm 67(C)
750 ft2 16
187 ton 192
4,000 cfm 85
487
575

 (a)  Reference 5.



 (b)  References 4 and 8 in the text.



 (c)  Reference 26 in the text.

-------
                                       B-4
     1000
  o
  o
  o
  o

  "5

  'd
   o
  o
      100
       10
         10
                                             Cooling/demisting
PuraSiv Hg adsorption

with additional cooling/

demisting included
                                      PuraSiv Hg adsorption only
                     i      I     i   J_  l  _L l  i
                                  j	i    i    i   i  i  i  i
                      100


             Chlorine Capacity, TPD
1000
FIGURE B-l.  CAPITAL COSTS OF  CONTROL  OF MERCURY IN THE HYDROGEN BYPRODUCT STREAM

-------
                              B-5
        TABLE B-2.  CAPITAL COSTS AND LIST OF EQUIPMENT
                    FOR PURASIV HG ADSORPTION


Major Equipment
Preheater
Adsorber vessels
100

Size
10 ft2
4'd> x 11'
TPD
Cost,
$1000
1
35
750 TPD

Size
75ft2
10"d> x 11'

Cost,
$1000
5
80
  (2)

Regeneration gas
  heater

Regeneration gas
  cooler

Bare module

Total module 18
  percent con-
  tingency

Cost given by UCC
 26KW      Nominal      200KW         5


28 ft2        1        213 ft2       10

             37                     100



             44                     118

             69                     175

-------
                                   B-6
addition to the capital costs additional for cooling/demisting.  The
cost/capacity relation is displayed in Figure B-l.
          Adsorbent requirements for the process are such that a 3-year
replacement is necessary at a cost/capacity variation given by

                              C   = 108 TPD.                          (B-4)
                               am

                     Operating and Maintenance Costs

          Operating and maintenance costs include  (1) operating labor;
(2) maintenance;  (3) utilities including water, electricity, and low
pressure steam; and (4) taxes and insurance.  An operating rate of 95
percent (or 8,322 hours per year) was assumed.  The values of  the above
items used in computing the O&M costs for the various control  processes
are summarized in Table B-3.
          All processes considered recover mercury.  Therefore, a credit
for mercury recovery was included.  For PuraSiv Hg adsorption  the mercury
is returned immediately and as such can be treated as an annual credit
in this case.  For  purposes of  this  study, primary and  secondary
cooling are necessary precursors to the adsorption process.  The costs
of cooling/demisting, however, affect the adsorption process as it
requires a higher capacity for cooling/demisting because of the recycle
of the regeneration gas stream.  Therefore, taking a mercury credit for
secondary cooling/demisting is not feasible.

Cooling/Demisting

          Operating labor costs that may be attributed to the mercury
processes  for this module were assumed negligible since the operation
of this unit may be considered as an integral part of operating the
manufacturing process.   Maintenance requirements were assumed to be 3
percent of capital per year.  Taxes and insurance usually amount to
2.5 percent of capital per year.
          The utility requirements of this module are detailed in Table B-4
for a 100  TPD plant.

-------
                          B-7
 TABLE B- 3.  BASIS FOR COMPUTATION OF ANNUAL O&M COSTS
                                 Cooling/     PuraSiv Hg
                                 Demisting    Adsorption


Utilities
  Steam. (Ib/hr per TPD) @ 50
-------
                           B-8
  TABLE B-4.  UTILITY REQUIREMENTS FOR COOLING/DEMISTING
              IN A 100 TPD PLANT
  Equipment          Duty         Utility        Rate

Primary boiler    2 MMBtu/hr    80 F, cooling
                                  water         200 gpm
                        (a\
Rotary blower     37.5hpv '     Electricity    37.5hphr
                                                  hr

Refrigerator      62.5hp        Electricity    62.5hphr
                                                  hr
(a)  Reference 5.

-------
                                   B-9
          At  cooling water  costs of  3C/1000 gal, the annual costs of
 cooling water are  given by:
          _$__ _ 200 gpm    _    60  min    g „  hrs    $0.03   m 3Q
          7ea7  100 TPD  X  TPD X   hr   X  *'J^ yr  X 1000 gal   JU
          The annual costs  of electricity  at 1.2
-------
                                      B-10
  100
o
o
o
o
o
    10
                              Cooling/demisting
                PuraSiv Hg adsorption -

                incremental  cooling/

                demisting cost included
                                        PuraSiv Hg adsorption only
                            I   i   i  i  i  i
                    l
i   i  i  i  i
     10
        100


Chlorine Capacity, TPD
         1000
             FIGURE B-2.  OPERATING AND MAINTENANCE COST FOR MERCURY

                          CONTROL IN HYDROGEN BYPRODUCT STREAM

-------
                                 B-ll
             Table B-5.   UTILITY REQUIREMENTS FOR PURASIV HG ADSORPTION
                         IN A 100 TPD PLANT
Equipment
Preheater
Regeneration Gas
Heater
Regeneration Gas
Cooler
Duty
32,000 Btu/hr
90,000 Btu/hr
75,000 Btu/hr
Utility
LP Stream
Electricity
80 F Cooling
Water
Utility Rat(
32 Ib/hr
26 kw
7.5 gpm
(a)   On for only 14  hours  out  of  every  24.

-------
                                   B-12
           The annual mercury  credit  amounts  to
           0.48 Ib per  day   .,,  _. day.          $200  =  4  4 TPD
               100 TPD      X 346'75 yr   X TPD X  76  Ib    4'4 TPD*
Therefore,  total O&M costs are
           O&M = 483 (TPD)0'454 + 13.2 TPD +  1000.                      (B-9)
The  above  relationship is  graphically displayed in Figure B-2.  Components
of the O&M costs shown in this figure are detailed in Tables B-6, B-7, and
B-8 for various C10 capacities.
                           Total Annual Costs

          The total annual costs of PuraSiv Hg adsorption are the sum
of the O&M, depreciation, and adsorbent replacement costs.

-------
                                       B-13
             TABLE B-6.   O&M COST COMPONENTS FOR COOLING/DEMISTING
                         $1000 (CURVE A,  FIGURE B-2)
ci2.





Capacity
TPD
100
250
500
600
750
Labor (a)
0
0
0
0
0
Utilities
12.3
30.7
61.5
73.8
92.1
Maintenance
4.3
8.1
13.1
14.8
17.3
Taxes and
Insurance
3.6
6.7
10.9
12.3
14.4
Total
20.2
45.5
85.5
100.9
123.8
(a)   Operating labor for cooling/demisting was assumed to be part of the
     manufacturing costs.

-------
                          B-14
TABLE B-7.  INCREMENTAL O&M COSTS OF COOLING/DEMISTING
            ATTRIBUTED TO PURASIV HG, $1000
TPD
100
250
500
600
750
Labor
0
0
0
0
0
Utilities
1.8
4.6
9.2
11.1
13.8
Maintenance
0.4
0.8
1.3
1.5
1.7
Taxes and
Insurance
0.4
0.7
1.1
1.2
1.4
Total
2.6
6.1
11.6
13.8
16.9

-------
                       B-15
TABLE B-8.  O&M COSTS OF PURASIV HG ($1000) EXCLUDING
            INCREMENTAL COOLING/DEMISTING COSTS
            (CURVE C, FIGURE B-2)
TPD
100
250
500
600
750
Labor
1.0
1.0
1.0
1.0
1.0
Utilities
1.8
4.4
8.8
10.8
13.2
Mercury
Credit
0.4
1.1
2.2
2.6
3.3
Maintenance
1.4
2.1
2.9
3.2
3.5
Taxes and
Insurance
1.8
2.7
3.7
4.0
4.4
Total
5.6
9.0
14.2
16.4
18.8

-------
                                   B-16
                               REFERENCES
(1)   Guthrie, K. M.,  "Capital Cost Estimating", Chemical Engineering
      (March 24, 1969).

(2)   Guthrie, K. M.,  Process Plant Estimating Evaluation and Control.
      Craftsman Book Company of America, Solana Beach, California (1974).

(3)   Anonymous, "Control Techniques for Mercury Emissions from Extraction
      and Chlor-Alkali Plants", U.S. EPA, Office of Air Programs, Research
      Triangle Park, North Carolina, Publ. No. AP-118 (February, 1973).

(4)   Clapperton, J. A., Manager of Manufacturing Industrial Chemicals,
      PPG Industries,  Pittsburgh, Pennsylvania, letter to M. Y. Anastas
      (February 28, 1975).

(5)   DeAngelis, P., Sobin Chemicals, Orrington, Maine, communication to
      M.  Y.  Anastas (March, 1975).

(6)   Information made available to Battelle's Columbus Laboratories by
      Union  Carbide Corporation in January and February,  1975.

-------
                                 TECHNICAL REPORT DATA
                          (Please read Instructions on the reverse before completing)
 1. REPORT NO.
  EPA-600/2-76-014
                            2.
                               3. RECIPIENT'S ACCESSION-NO.
 4. TITLE AND SUBTITLE
 Molecular Sieve Mercury Control Process in
    Chlor-Alkali Plants
                               5. REPORT DATE
                               January 1976
                               6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)

 M.Y. Anastas
                                                       8. PERFORMING ORGANIZATION REPORT NO
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Battelie-Columbus Laboratories
 505 King Avenue
 Columbus, Ohio  43201
                               10. PROGRAM ELEMENT NO.
                               1AB014; 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
  . ABSTRACT The y^pf gjves results of an investigation of the use of the PuraSiv Hg
 adsorption process to remove mercury from the hydrogen byproduct stream and the
 end-box ventilation stream from mercury cell chlor-alkali plants.  The investigation
 included the analysis of data obtained from testing of a system that is currently in
 operation  and technical information provided by the system vendor together with that
 available in the open literature.  Although the measurements of mercury concentra-
 tion in the hydrogen byproduct  stream entering the PuraSiv Hg adsorber, taken during
 performence testing of the control unit, appear to be in error, measurements  of the
 outlet concentration indicate that a concentration less than 60 ppbv may be achieved.
 The economics of the PuraSiv Hg adsorption process were explored.  Available data
 indicate that the operating costs by this process vary between $0. 58 and $0. 33 per
 ton of chlorine produced for plants with capacities between 100 and 750 tons per day.
 Mercury may also be removed from the hydrogen byproduct stream either by brine
 adsorption over treated activated carbon or by scrubbing with depleted brine.  Tech-
 nical and economic data available to the investigator seem to favor the use of these
 two processes for  mercury control, although the data base thereon is not sufficiently
 developed to warrant a meaningful comparison.
 7.
                             KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                          b.IDENTIFIERS/OPEN ENDED TERMS
                                           c.  COSATI Field/Group
 Air Pollution
 Mercury (Metal)
 Absorbers  (Materials)
 Chemical Plants
 Chlorine
 Sodium Hydroxide
Activated Carbon
Brines
Air Pollution Control
Stationary Sources
Molecular Sieves
Chlor-Alkali Plants
PuraSiv  Hg Unit
13B
07B
11G
07A
 B. DISTRIBUTION STATEMENT
 Unlimited
                                          19. SECURITY CLASS (This Report)
                                           Unclassified	
                                            21. NO. OF PAGES
                                                72
                                          20. SECURITY CLASS (Thispage)
                                           Unclassified
                                                                   22. PRICE
EPA Form 2220-1 (9-73)

-------