tTLEAl
WATER POLLUTION CONTROL RESEARCH SERIES
14010 DYG 08/71
      Acid Mine Waste Treatment
       Using Reverse Osmosis
ENVIRONMENTAL PROTECTION AGENCY • WATER QUALITY OFFICE

-------
        WATER POLLUTION  CONTROL  RESEARCII  SERJES
 Tlie Water  Pollution  Control  Research' Reports  describe
 the results  and  progress  in  tlm  control  and abatement
 of pollution in  our  Nation's waters.   They provide  a
 central  source of  information  on the  research,  develop-
 ment,  and  demonstration.activities  in the Environmental
 Protection Agency, Water  Quality Office, through  inhouse
 research and grants  and contracts with Federal, State,
 and local  agencies,  research institutions, and  industrial
 organizations.

 A tripilicate abstract card  sheet is  included in  the
 report to  facilitate information retrieval.   Space  is
 provided on  the  card for  the user's accession number
 and for  additional uniterms.

 Inquiries  pertaining to Water  Pollution .Control.-Research.
 Reports  should be directed to  the Head, Project Reports
 System,  Office of Research and Development, Environmental
"Protection Agency, Water  Quality Of'flee,'-'Washington,  D. C.
 20242,

-------
Acid Mine Waste Treatment Using Reverse Osmosis
                            by
             Gulf Environmental Systems Company
                      P. 0. Box 608
               San Diego, California  92112
                          for the

               ENVIRONMENTAL PROTECTION AGENCY
                    WATER QUALITY OFFICE
                    Program No. 14010 DYG
                   Contract No. 14-12-525
                        August 1971

-------
               EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents neces-
sarily reflect the views and policies of the Environ-
mental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement
or recommendation for use.
                    11

-------
                                 ABSTRACT

The basic objectives of this test program were to demonstrate the applicability
of reverse osmosis to the demineralization of acid mine drainages, both high-
ferrous and high-ferric types, and to reclaim the maximum percentage of such
feedwater in purified form suitable for domestic or industrial purposes,
or as stream discharge.  These goals included the attainment of maximum
water recovery while maintaining the required product water quality and the
determination of pretreatment requirements necessary to maximize water
recovery and membrane life.

Two reverse osmosis test units were operated during the course of these
                                                              2
tests:  a nominal 10,000-gpd unit equipped with eighteen 50-ft  modules
                                                     2
and a nominal 4,000-gpd unit equipped with nine 50-ft  modules.  The modules
used in these units consisted of both high-selectivity and high-flux cellulose
acetate membranes.

The test program was carried out at three different mine drainage sites.
The mine drainage water at the first site, Norton, West Virginia, contained
greater than 98 percent of the iron present in the ferric form; at the other
two sites, Morgantown, West Virginia, and Ebensburg,  Pennsylvania, the
drainage water contained predominantly ferrous iron.  Discharges at the
second site were so concentrated that recoveries were limited to 50 percent;
recoveries of 80 to 90 percent were attained at the first and third sites.
No iron fouling was encountered at any of the three sites.  Specific salt
rejections were >97 percent at all sites.

This report was submitted in fulfillment of Contract No. 14-12-525 between
the Environmental Protection Agency and Gulf Environmental Systems Company.
                                     111

-------
Key Words:  Reverse osmosis, acid mine drainage, demineralization, calcium-
            sulfate solubility, ferrous-ferric iron ratios, water recovery,
            brine treatment.
                                    IV

-------
                                 CONTENTS
Section
            ABSTRACT
   I.       CONCLUSIONS
  II.       RECOMMENDATIONS
 III.       INTRODUCTION
  IV.       DESCRIPTION OF TEST EQUIPMENT
   V.       PERFORMANCE CALCULATION METHODS
  VI.       HIGH-PRESSURE-PUMP PROBLEMS
 VII.       FEEDWATER PRETREATMENT STUDIES
VIII.       TEST PROGRAM
  IX.       OPERATION AND RESULTS AT SITE 1
   X.       OPERATION AND RESULTS AT SITE 2
  XI.       OPERATION AND RESULTS AT SITE 3
 XII.       DISCUSSION OF TEST RESULTS
XIII.       CALCIUM SULFATE SOLUBILITY LIMITATION
 XIV.       LABORATORY NEUTRALIZATION EXPERIMENTS
  XV.       ACKNOWLEDGMENTS
 XVI.       REFERENCES
Page
111
 1
 3
 5
 9
 13
 15
 17
 19
 21
 37
 49
 63
 67
 71
 79
 81

-------
                                 FIGURES
                                                                    Page
1. Piping and instrumentation diagram for 10,000-gpd reverse
   osmosis unit with brine recycle                                   ]Q
2. Schematic of the 10,000-gpd and 4,000-gpd units at
   Norton, West Virginia                                             o?
3. Membrane Performance at Sites 1,  2,  and 3                         66

-------
                                 TABLES
                                                                    Page
 I          Typical Water Analyses of Tested Acid Mine Drainage        g
 II         Proposed Test Plan, Site 1                                20
 III        Module Loading Arrangement for 10,000-gpd Unit at
           Site 1, 2/5/70                                            24
 IV         Operational History of 10,000-gpd Unit at Test
           Site 1, 3/9/70 -  3/13/70                                  25
 V          Mean Operating Values of 10,000-gpd Unit at Site 1,
           3/9/70 - 3/13/70                                          26
 VI         Chemical Analyses  for 10,000-gpd Unit Operation
           at  Site 1, 3/9/70  - 3/13/70                               27
 VII        Mean Operating Values of 10,000-gpd Unit at Site 1,
           3/16/70 - 3/30/70                                         28
 VIII       Chemical Analyses  for 10,000-gpd Unit Operation at
           Site 1, 3/16/70 -  3/30/70                                 29
 IX         Postoperational Performance of Modules from Site 1        31
 X          Spectrochemical Analysis of Backing Material Deposit
           from Module 6/11-7                                        ^
 XI         Spectrochemical Analysis of Membrane Deposit  from
           Module 6/11-7              .                               34
 XII        Module Loading Arrangement for 4,000-gpd Unit at Site 2   38
 XIII       Operational History of 4,000-gpd Unit at Test Site 2,
           5/13/70 - 5/24/70                                         40
 XIV        Comparison of Chemical Analyses  for Operation at
           Site 1 and Site 2                                         41
 XV         Mean Operating Values of 4,000-gpd Unit at Site 2,
           5/13/70 - 5/24/70                                         42
XVI        Chemical Analyses  for 4,000-gpd Unit Operation at
           Site 2, 5/13/70 -  5/24/70                                 43
 XVII       Results of Postoperational Testing of Modules Operated
           at  Site 2                                                 44
 XVIII      Chemical Analysis of Samples from Site 2                  45
 XIX        Module Loading Arrangement for 4,000-gpd Unit at Site 3   50
 XX         Chemical Analyses  for 4,000-gpd Unit Operated at
           Site 3, Run 1                                             52
 XXI        Operational History of 4,000-gpd Unit, Test Site 3, Run 1 53
                                    vii

-------
                            TABLES  (Continued)
XXII


XXIII

XXIV


XXV

XXVI


XXVII


XXVIII


XXIX

XXX
Chemical Analyses  for 4,000-gpd Unit Operated at
Site 3, Run 2
                                                                     Page
55-
Operational History of 4,000-gpd Unit, Test Site 3, Run 2  56
Chemical Analyses  for 4,000-gpd Unit Operated at Site 3,
Run 3                                                      57
Operational History of 4,000-gpd Unit, Test Site 3, Run 3  55.
Chemical Analyses  for 4,000-gpd Unit Operation at Site
3, Run 4
Operational History of 4,000-gpd Unit at Test Site 3,
Run 4
Comparison of Calcium Sulfate Solubility for Different
Reverse Osmosis Brines
Results of Acid Mine Water Neutralization Studies
Results of Bench Neutralization Tests on 10,000-gpd
Unit Brine at 91 Percent Recovery
60


61

68

73

77

-------
                                 SECTION I
                                CONCLUSIONS

The basic objective of the test program, demonstration of the ability of
reverse osmosis to demineralize acid mine drainage water (both high-ferrous
and high-ferric types), was successfully demonstrated at two of the three
experimental test sites.  The feedwater at Site 2 was so polluted that
reverse osmosis demineralization does not seem feasible.  The quality of
the product water produced at Site 1 and Site 3 would be suitable for
domestic or industrial purposes with the exception of pH or dissolved iron
content.  If the pH of the water were adjusted to neutral, the iron would
precipitate and could be easily removed by sand filtration or settling.

Environmental Protection Agency (EPA) personnel carried out neutralization and
decantation operations followed by recycling of the supernatant therefrom
through the reverse osmosis unit at Site 1.  This resulted in effective
98 percent water recovery based on feed volume, with maintenance of excellent
quality in the recovered permeate water.  The details of this work are
reported elsewhere (Ref. 1) by EPA personnel.

The ability to operate with no iron fouling on acid mine discharge waters
similar in type to water with which other investigators (Ref. 2) had experi-
enced fouling was also successfully demonstrated at Site 3.  The method used
to control this type of fouling was pH adjustment of the raw feedwater.

Laboratory studies of the effectiveness of compounds in retarding calcium
sulfate precipitation yielded no successful candidates.  Similar studies
aimed at neutralizing the reverse osmosis brine indicated that lime or MgO
would be satisfactory, with lime being chosen because of its lower cost.

-------
A method of evaluating an acid mine drainage water with regard to the maxi-
mum water recovery possible was developed.  The limiting factor in recovery
is the calcium sulfate concentration.  While the CaSO,  solubility limit can
be exceeded, probably because of the short residence time, low pH, other
ionic interferences, etc., there does seem to be a fairly definite limit of
approximately 300 to 400 percent of saturation.

The standard pretreatment of sand filtration followed by cartridge filters
proved to be adequate with respect to feed clarification, so membrane surface
fouling from suspended materials in the feed stream was not encountered.

The high-pressure feed pumps used on both units during these tests failed
on numerous occasions, resulting in time-consuming delays.  These pumps,
although different types, contained materials that were not compatible with
the acid mine drainage water.

-------
                                 SECTION  II
                              RECOMMENDATIONS

 It  is  recommended  that  to  further  evaluate  the ability  of reverse osmosis
 to  demineralize various acid mine  waters, a larger unit, where no or minimal
 brine  recycle  is required, be utilized.  The larger units can be economically
 equipped with  automatic features that are prohibitive on smaller units.  Cost
 studies of  demineralizing  acid mine water with units of 1 mgpd and  larger
 should be included  in future studies.

 Recent field experience has indicated that  a spiral module should be operated
 at  pressures required to produce 10 to 15 gallons per square foot per day
 of  actual permeate, even though  the membrane is capable of much higher
 productivity.  Preferably  this pressure  should be below 600 psi and in most
 cases  will  be, except for  feeds of abnormally low temperatures or those
 having very high osmotic pressures.

 More work is required on different types of  acid mine drainage waters to
 fully  evaluate reverse osmosis capabilities.  Long-term testing (in excess
 of  six months) is required to accurately predict membrane lifetime.  Further
 testing is  required to examine the calcium  sulfate solubility limitations and
 the validity of adjusting  the pH of ferrous-type feeds.

 It is  further recommended  that all the pumps used in reverse osmosis plants
 be either stainless steel, plastic, or a combination of these two materials.
 Stuffing-box-type packings are not advisable.

 The method  of recycling the treated brine (Ref. 1) back to the feed remains
 to be  evaluated for operation on ferrous-type acid mine drainage.

 Further studies are required to fully ascertain what actually occurs in the
modules that results in flux increases during periods of low recovery opera-
 tion and accidental or planned shutdowns.  The apparent discrepancies in the
 temperature-flux normalization curves used in these tests  need to be resolved.
                                     3

-------
                              SECTION III
                              INTRODUCTION

This  is  the  final report on the reverse osmosis field  testing on acid
mine  water carried out by Gulf Environmental  Systems Company for the
Environmental Protection Agency, Water Quality Office, under Contract
14-12-525.   The work done under this contract followed prior work
performed at the EPA Norton Mine Drainage Treatment Laboratory with
EPA site support under Office of Saline Water Contracts 14-01-0001-1243,
14-01-0001-1836, and 14-01-0001-1836 Amendment No. 3, which has been
previously reported (Refs. 3-5).

Work  done under these referenced contracts had demonstrated that a reverse
                                                      ®
osmosis unit using the Gulf Environmental Systems ROGA  spiral-wound modular
concept could successfully be operated on acid mine drainage at up to 75
percent recovery levels  (and for several limited times, as high as 92 percent)
using either high-flux, lower-selectivity or standard-flux, high-selectivity
modules.  The purpose of the present contract was to determine the
maximum water recovery possible while producing water suitable for
domestic or  industrial purposes or for discharge into streams.  Also to
be investigated were the pretreatment requirements necessary to maximize
water recovery and membrane life.  These tests were carried out at three
different locations where the mine discharges were markedly different:
(1) Norton, West Virginia, Grassy Run Creek;  (2) Morgantown, West Virginia,
mine discharge; and (3) Ebensburg, Pennsylvania, mine discharge.  Typical
analyses of the acid mine discharges are shown in Table I.  Based on past
evaluations  (Ref. 3), the osmotic pressure of these various feeds was assumed
to be 10 psi per 1000-micromho  specific conductance.

The water was taken from Grassy Run Creek (Site 1) about 2 miles downstream
of the mine discharge.   The water in this shallow, cascading creek was
well aerated before it reached the intake lines, and in addition contained
      i\
      Gulf Energy & Environmental Systems, Inc. registered trademark.
                                     5

-------
                                                TABLE I
                          TYPICAL WATER ANALYSES OF TESTED ACID MINE DRAINAGE

Site 1, Grassy Run Creek,
Norton, West Virginia -
Winter
Site 1 - Spring
Site 1 - Summer
Site 1 - Fall
Site 2, Arkwright Mine,
Morgantown, West Virginia
Site 3, Mines 32-33,
Ebensburg, Pennsylvania
pH


2.86
2.98
2.7
2.8

2.3

3.6
Cond
(Vtmhos/cm)


1660
1083
1350
850

7600

1480
Ca as Ca
(mg/1)


114
54
115
72

650

214
Mg as Mg
(mg/1)


33
19
38
25

450

157
Acidity

T
732
3721
6441
3571

5380

360
Total Fe
(mg/1)


146
73
153
74

2960

126
Fe^
(mg/1)


6.6
1.5
<1
<1

1180

101
S04
(mg/D


1040
600
936
610

12285

1638
Al
(mg/1)


32
16
38.5
26

380

35
TDS
(mg/1)


1604
875
_ _
—

—

—
Acidity as CaC03 (mg/1)

-------
the raw domestic wastes discharging from about 40 homes upstream of the
test site.  This water contained 50 to 150 ppm of iron, depending on
climatic conditions, which was present almost totally in the ferric state.

The water at Site 2 was taken directly from a pump discharge of the Arkwright
mine of the Christopher Coal Company.  The iron content of this water
was about 65 percent ferrous, but the total amount was about 30 times that
present in the water at Site 1.

The water at Site 3 also was taken directly from the pump discharge of an
active mine.  This discharge was from Mines 32-33 of the Cambria Division
of the Bethlehem Mines Corporation.  The total iron content of this water
was approximately the same as that at Site 1, but was approximately 80 per-
cent in the ferrous state.

The reverse osmosis equipment, test program, site selection, operation,
and maintenance were supported by and under the direction of the Mine
Drainage Pollution Control Section of the Environmental Protection Agency,
Water Quality Office.  Field test engineering personnel were provided by
Gulf Environmental Systems at the beginning of each set of experiments
and subsequently as required.

-------
                                 SECTION  IV
                       DESCRIPTION OF TEST  EQUIPMENT

 The  reverse  osmosis equipment used in these tests consisted  of a 10,000-gpd
 unit and a 4,000-gpd unit.  The  10,000-gpd  unit was composed of five  4-in.-
 diameter, 10-ft-long Schedule 40 pressure vessels, a high-pressure pump,
 and  instrumentation and controls.  A piping and instrumentation diagram
 of the unit  is  shown in Fig. 1.  The five pressure vessels contained  a
                                                             f\
 total of 15  ROGA® modules, with  each module containing 50 ft of modified
 cellulose acetate membrane.  The pressure vessels were arranged in a  2-2-1
 array in order  to maintain as uniform a brine flow as possible at required
 velocities.  In this arrangement, vessels 1 and 2 are in parallel.  The
 brine from these vessels is combined to be  the feed for vessels 3 and 4,
 which are in parallel.  The brine from vessels 3 and 4 is combined to be
 the feed for vessel 5.  An instrument panel provides for measurement of
 the feed pressure, the feed channel pressure drop, and the product and
 brine flow from each vessel.  The unit is plumbed so that a  portion of the
 concentrated brine can be recycled back to  the feed.  This arrangement
 was required to obtain high recovery levels  while maintaining an adequate
 brine flow of about 3 gpm or greater.  This  brine flow is necessary
 because it minimizes the concentration polarization at the membrane
 surface, which, in turn, minimizes the probability of membrane fouling,
Units of larger size,  i.e., 50,000 gpd or larger,  are capable of high
 recoveries without brine recirculation.   The 4,000-gpd unit had only
 three pressure vessels, but was  similar in controls and operation to
 the 10,000-gpd unit.

 Both high-selectivity, standard-flux modules and high-flux, lower-selectivity
modules were used in these tests.  In addition, for operation at Site 3,
modules manufactured by a different technique were used.   These modules

-------
                   API
                   T1
                  AP2
                   T2
PF1
API
FBI
       VENTURI ORIFICE
       FEED PRESSURE TUBE  1
       PRESSURE  DROP TUBE  1
       BRINE FLOW TUBE 1
                                             AP3
                                             T3
                                    Prt,	**	(FBI,'
                                                                       AP5
                                                                        T5
/\    /\
  \  /
   \/  J
Fig.  1.  Piping and instrumentation  diagram for 10,000-gpd reverse osmosis  unit with brine recycle

-------
combine both the high-flux and high-selectivity properties of the other

modules and are more rugged physically.  A comparison of performance for

typical modules of these three types is as follows:
                              Module Membrane
                             Coefficient,1 AxlO5         Percent  Sodium
                               (g/cm2-sec-atm)        Chlorine  Rejection
      High-selectivity
      module ..........  >1.3<2.0                  >94.5

      High-flux module .....  >2.0<3.0                  >90

      High-flux,  high-
      selectivity module.  .  .  .    1.9-2.2                    2:95
      1            -5     2                      2
       A = 1.0 x 10   g/cm -sec-atm = 8.65 gal/ft /day at 600 psi net
       driving pressure and 77°F.


      2At 600 psi and 2000 ppm feed
 Neutralization of the concentrated brine with hydrated lime at Site 1
 was carried out by EPA personnel using their equipment and facilities
 (Ref.  6).   Because of the fixed nature of this equipment,  no neutraliza-
 tion was possible at the other two test sites,
                                     1.1

-------
                                SECTION V
                      PERFORMANCE CALCULATION METHODS
The product water flow through a semipermeable membrane may be expressed as
                             FW

                           2
where FTT » water flux (g/cm -sec),
       W
                                               2
       A = water permeability coefficient (g/cm -sec-atm) ,
      AP = pressure applied to the membrane (atm) ,
      ATT « osmotic pressure across the membrane (atm) ,
       g = grams.

For these tests, the module or unit water permeability coefficient A (in
    2
g/cm -sec-atm) was calculated by dividing the product water flow, corrected
to 77°F,< by the membrane area and the net pressure.  The net pressure is
the applied pressure minus the osmotic pressure.  The osmotic pressure was
assumed to be directly proportional to the conductivity.  A value of 10 psi
was used for the osmotic pressure of the acid mine water when the specific
conductance of the feed equaled 1000 micromhos/cm.

The percent rejection of the membrane is expressed as
             	Concentration Product	
             (Concentration Feed + Concentration Brine) -f 2 X
In cases where the unit was operating with brine recycle, the blended feed
concentration was used in lieu of the raw feed concentration-.
                                     13

-------
                                 SECTION VI
                        HIGH-PRESSURE-PUMP PROBLEMS

A significant amount of time was lost on this program because of the problems
encountered with the high-pressure pumps on both units.  The 10,000-gpd unit
was equipped with a Moyno SSQ-9P4 progressive-cavity type pump, while the
4000-gpd unit was originally equipped with a Gardner-Denver PQ-2 triplex
positive-displacement pump.*

The Moyno pump failed in several areas.  Both the pumping rotor and the
connecting rod between the rotor and drive shaft were broken several times.
In addition, the original connecting rods were made from carbon steel, which
corroded severely with the resulting corrosion products entering the process
water, which tended to foul the modules.  Attempts at replacing the carbon
steel rods with rods of stainless steel were not successful because the
stainless steel did not have sufficient tensile strength.  A special
precipitation-hardened steel was employed to fabricate several rods which
were installed during the latter part of the last run.  However, not enough
operating hours were accumulated to evaluate this type of steel.  No specific
answer to this material problem is evident.  The Moyno pump also had a
stuffing box seal arrangement which both leaked badly and allowed packing
to work its way into the process water.  Because of these problems, it is
recommended that this type of pump not be used in future reverse osmosis
testing on acid mine waters.

The Gardner-Denver triplex pump originally was fabricated with an aluminum-
bronze fluid end, ceramic plungers, and Monel valves and seats.  The Monel
valve seats corroded within 2 days of operation on brine from the 10,000-gpd
unit.  The Monel valves were replaced with readily available aluminum-bronze
valves, but this set lasted less than 12 hr.   At this point, a new aluminum-
     *Mention of  commercial  products  does  not  imply  endorsement  by  the
 EPA.
                                     15

-------
 bronze fluid head, complete with stainless steel valves, was installed on
 the pump.  After about five days of operation on brine from the 10,000-gpd
 unit, the plunger stuffing boxes were leaking very badly, and therefore
 all testing was suspended until new parts could be obtained.  When the new
 stuffing boxes were installed, the pump was operated only on neutralized
 brine and performed satisfactorily.

Further use of this type of pump seems feasible only if the complete fluid
end is furnished in stainless steel.

-------
                               SECTION VII
                     FEEDWATER PRETREATMENT STUDIES

The feedwater to a spiral-wound reverse osmosis unit requires pretreatment for
three significant reasons.  First, particles in the water capable of plugging
the brine channels must be removed from the water before it reaches the modules.
During preceding test programs at the Norton site, rapid sand filters followed
by 10-micron cartridge filters had proven.satisfactory for this purpose.
Therefore, this method was used at all three sites.

Second, biological growths of algae or molds inside the modules must be
prevented.  Continuous addition of 5 ppm copper sulfate during the warmer
summer months at Site 1 was used for this purpose, while no additions
were required during the winter months.  Because the feedwater at Sites
2 and 3 was taken directly from the mine discharge, and not from a creek
at a considerable distance from the mine discharge, no chemical additions
were deemed necessary, at those sites.  Postoperational examination of
modules supported this assumption.

Third, the precipitation of sparingly soluble salts in the concentrated
brine with subsequent plugging or fouling of the modules must be prevented.
Indeed, it has been demonstrated that the limiting factor in system
recovery is the concentration of these salts.  Previous experience on
other types of waters has shown that calcium sulfate is the most significant
sparingly soluble compound present.  In other tests, where the feed was
concentrated with respect to calcium sulfate, it was found that the water
could be further concentrated in a reverse osmosis system by a factor of two
or more if "threshold inhibitors" (sodium hexametaphosphate, Cyanamer P-35,
etc.) were added to the feed in the range of 5 to 10 mg/1.  To investigate
                                       4*
these and other inhibitors, a synthetic acid mine water was prepared at Gulf
Environmental Systems, based on suggested formulations in the literature
(Ref. 7).  The composition of this synthetic acid mine water was as follows:
                                     17

-------
          Magnesium  (Mg).  .....  	  . 116 mg/1
          Calcium  (Ca)	385 mg/1
          Iron  (Fe"*4""1")	500 mg/1
          Aluminum (Al)	150 mg/1
                        I |
          Manganese  (Mn  )	50 mg/1
          Sulfate  (S04)	3615 mg/1
          Conductivity	3800 micromhos/cm at  25°C
          pH	2.5

This solution was roughly  comparable  to  the concentrate produced  at  75
to 90 percent recovery by  a reverse osmosis unit at the Norton  site.  This
value was deemed to  be close  to the limit of concentration based  on  calcium
sulfate concentration.   The achievement  of further concentration  by
inhibiting CaSO, precipitation through various inhibitors was investigated
and proved to be unsuccessful.  Either the low pH or the high concentration
of iron, or both, prevented the inhibitors from performing normally.  The
inhibitors tried included:

     Calgon C-55                     Magnafloc 837A
     Calgon CL-45                    Polyvinylpyrosilicone K-90
     Cyanamer P-35                    Silspend 180
     Bequest 2000                     Sodium Hexametaphosphate
     Bequest 2010                     Sodium Xylene Sulfonate
     Bequest 2041                     Surfynol 82
     Eltan 382-E                     Tetrasodium Pyrophosphate
     Gantrey AN-119                  Toranil B, 97B
     Hercules Klucel H               Trasitan SPM
                                     Zimmite AM-100

To determine if the  laboratory studies on the synthetic water provided
valid conclusions in the light of field performance, sodium hexametaphosphate
was added at Site 1 during one series of tests.  As in the laboratory
studies, no appreciable effect was observed.
                                    18

-------
                               SECTION VIII
                               TEST PROGRAM

The test program at Site 1 was to involve five phases, as outlined in
Table II.  Operations at this site were to include the use of both the
4,000-gpd and 10,000-gpd reverse osmosis units.  The 10,000-gpd unit
was to be used to concentrate the feedwater by a factor of 4 to 5; then
the 4,000-gpd unit was to be used to further concentrate the brine so
that an overall concentration factor of 10 or more was reached.  The
inability of the particular high-pressure pump (aluminum-bronze fluid end)
on the 4,000-gpd unit to handle the acid mine water necessitated a change
in program, so that all further testing was accomplished in the 10,000-gpd
unit.

Since the previous successful testing on acid mine drainage  (Refs. 3-5)
was on water containing predominately ferric iron, and since other
investigators  (Ref. 2) had encountered difficulty when operating  on high-
ferrous-iron waters, it was decided that the second site should be one where
a large  amount of ferrous iron was present.  Site 2 was selected  on this
basis.   When it became evident that testing was limited at  this site owing
to the high concentration of dissolved salts, which prohibited high recoveries,
a third  site was sought.  The discharge at this third site  contained about
the same amount of dissolved salts as the discharge at Site  1, except that
it contained approximately 80 percent ferrous iron.  At Site 3, the adjust-
ment of  the pH of the feedwater was employed to prevent or minimize the
ferrous-ferric conversion and ferric precipitation.  Although experimental
limitations precluded comparison over a complete range of pH and  recovery
rates, no  significant iron hydrate fouling was encountered.
                                    19

-------
                                               TABLE II
                                      PROPOSED TEST PLAN,  SITE 1
Phase I   - Unit Shakedown and Operational Check
            1.  Run on untreated 10,000-gpd-unit product water for
                approximately 1000 hr.
                a.   Run at various pressures and recoveries  up to
                    600 psi and 80 percent recovery using  brine recirculation.
Phase II  - 90 Percent Recovery Operation
            1.  Run on untreated 10,000-gpd-unit brine to  approximately 90 percent
                total recovery, operating at several combinations of  recovery,  such as:
                a.   10,000 gpd - 60 percent recovery, no brine recirculation
                     4,000 gpd - 75 percent recovery, brine  recirculation
                or
                b.   10,000 gpd - 70 percent recovery, brine  recirculation
                     4,000 gpd - 70 percent recovery, brine  recirculation
            2.  If  90 percent recovery is successful, increase total  recovery
                as  far as possible by increasing the recovery in  the
                4,000-gpd unit.
Phase III - 90 Percent Recovery Operation with Use of Inhibitor
            1.  Run 4,000-gpd unit on treated 10,000-gpd-unit brine,  adding a precipitation
                inhibitor to the brine.
                a.   The minimum recovery would be that which was
                    found to be the maximum under Phase  II and would
                    be increased as far as possible.
                b.   Precipitation inhibitors may be Cyanamer P-35,  sodium
                    hexametaphosphate, Dequest 2000, or  others.
Phase IV  - 90 Percent Recovery Operation, Neutralizing  Brine, Use  of Inhibitor
            1.  Run 4,000-gpd unit on neutralized 10,000-gpd-unit brine, approximately 90 percent
                recovery using precipitation inhibitors.
                a.   Unit conditions would be similar to  those in
                    Phase II.
                b.   The brine from the 10,000-gpd unit would be neutralized
                    to pH 5 using CaO, CaCO-, MgO, or others.
Phase V   - Maximum Recovery
            1.  Run 4,000-gpd unit on treated 10,000-gpd-unit brine to maximum practical
                recovery, optimizing conditions of either  Phase III or Phase IV.
                a.   Determine maximum recovery.
                b.   Determine minimum addition of precipitation inhibitor.

-------
                                SECTION IX
                      OPERATION AND RESULTS AT SITE 1

Operations at Site 1 began on November 6, 1969.  The mine drainage water
at this location is taken from Grassy Run Creek about two miles downstream
of the mine discharge.  The water contains 50 to 150 ppm of iron, almost
totally ferric.  The overall system flow diagram for operation at this site
is shown in Fig. 2.

Phase I of the test plan, i.e., checking out the pumps, instruments, etc.,
of the 4,000-gpd unit, continued until November 14, 1969.  The high-pressure
pump for this unit was a Gardner-Denver PQ-2 triplex positive displacement
piston pump with an aluminum-bronze fluid head, ceramic pistons, and Monel
valves and seats.  Nothing unusual was experienced during this test phase.

Phase II began on November 14, with the 4,000-gpd unit operating on untreated
brine from the 10,000-gpd unit.  Almost immediately problems were encountered
with the high-pressure Gardner-Denver pump.  These pump problems continued to
plague the test program until this pump was used only for neutralized feeds.
This required modification of the test plans for the remainder of the contract.
                                    21

-------
           RAPID  SAND
           FILTER
                                                    BRINE  RECYCLE
    GRASSY
    RUN
    CREEK
                 10-MICRON
                 CARTRIDGE
                 FILTERS
      FEEDWATER
10,000-GPD REVERSE
OSMOSIS UNIT
BRINE TO 4,000-GPD
REVERSE OSMOSIS UNIT
                                                                      PRODUCT
PO
                            BRINE  RECYCLE
BRINE
DRAIN
            TO
                    PRODUCT
4,000-GPD
REVERSE
OSMOSIS UNIT
     10-MICRON
     CARTRIDGE
     FILTERS
 500-GAL
 HOLDING
 TANK
             Fig. 2.  Schematic of the 10,000-gpd and 4,000-gpd units at Norton, West Virginia

-------
Up to this time, the 10,000-gpd unit had been loaded with modules that were
in the unit during the last experimental contract (Ref. 5).  New modules
were installed as shown in Table III.  Phase IV of the program began and
the 10,000-gpd unit was operated at 91 percent recovery for five days, when
the high-pressure pump on the unit failed.

After the pump was repaired, the 10,000-gpd unit was operated at approximately
91 percent recovery for 100 hr, during which time the brine was collected
and stored.  The operational history of the unit during this run is shown in
Table IV.  Note that during the last half of the run, the AP of tube 5, the
last tube in the array, was increasing steadily.  It was felt that this was
caused by calcium sulfate precipitation.  The unit recovery was lowered to 50
percent, and in about 3 hr the AP was back to its initial value.  Thus, while
CaSO^ precipitation does occur at high recoveries, it apparently can be flushed
easily from the system if caught early enough.  Mean operating values for this
run are shown in Table V, and representative chemical analyses are shown in
Table VI.

The next run consisted of operating the 10,000-gpd unit at 91 percent recovery
during the normal working hours and at 85 percent recovery at all other times.
Small changes in the operating parameters, caused by valve vibration, tempera-
ture,  etc., could cause major changes in recovery due to the small amount
of brine being discharged.  Since the unit was unattended except during
working hours, it was safer to operate at 85 percent than at 90 percent during
these times.   This run continued for 100 hr, at which point the high-pressure
pump failed.   The pump was repaired, and the test continued another 80 hr
before the run was terminated.  Operating values and chemical data for this
run are shown in Tables VII and VIII.

The stored brine was neutralized using hydrated lime and pumped into a
                                              4«
settling tank overnight.  The supernatant was sand-filtered, and the 4,000-
gpd unit operated at approximately 50 percent recovery on this neutralized
brine.  Only enough filtered water for a 2-hr run could be stored, but during
                                      23

-------
                   TABLE III
MODULE LOADING ARRANGEMENT FOR 10,000-GPD UNIT
              AT SITE 1, 2/5/70
                      24

-------
                                                      TABLE IV
                                       OPERATIONAL HISTORY OF 10,000-GPD UNIT
                                          AT TEST SITE 1, 3/9/70 - 3/13/70
Date
3/9/70




3/10/70



3/11/70





3/12/70





3/13/70



Elapsed
Time
(hr)
5.4
9.5
12.6
16.8
20.9
24.9
28.7
33.0
36.6
41.0
44.8
49.4
53.1
56.8
61.0
64.9
68.9
72.7
76.9
81.0
84.8
88.9
92.8
96.8
100.0
Pressure
(psi)
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
Recovery
(percent)
92.2
90.8
91.0
90.9
90.9
91.0
91.5
89.9
90.6
91.0
91.4
91.3
93.1
91.4
90.8
91.3
91.4
91.2
91.2
91.7
90.6
91.2
91.3
90.9
90.9
Product
Flow
(gpm)
5.88
5.80
5.55
5.40
5.26
5.30
5.55
5.26
5.26
5.26
5.00
5.00
6.20
5.88
5.00
5.00
5.00
5.00
5.26
5.26
5.00
5.00
4.76
4.35
4.54
Brine
Flow
(gpm)
0.50
0.59
0.55
0.54
0.53
0.52
0.52
0.59
0.54
0.52
0.47
0.48
0.46
0.55
0.50
0.47
0.47
0.48
0.51
0.47
0.51
0.49
0.46
0.43
0.46
AP,
Tube 5
(psi)
3.4
3.2
3.2
3.2
3.3
3.2
3.1
3.3
3.3
3.3
3.4
3.4
3.1
3.3
3.5
3.6
3.6
3.6
3.6
3.7
4.0
4.1
4.1
4.4
4.4
Temp
(°F)
49
40
37
41
41
45
48
38
36
42
42
47
59
52
40
42
42
46
51
44
41
43
42
43
44
Unit
Water Permeation
Coefficient, AxlO~5
at 77°F
2.15
2.48
2.47
2.26
2.20
2.08
2.06
2.31
2.38
2.17
2.06
1.89
1.88
2.03
2.13
2.06
2.06
1.93
1.85
2.10
2.10
2.03
1.96
1.76
1.81
10
CJ1

-------
                                  TABLE V
            MEAN OPERATING VALUES OF 10,000-GPD UNIT AT SITE 1,
                             3/9-70 - 3/13/70
Pressure  	  600.6 psi
Unit recovery	  91.2 percent
Recovery variation, max.-min  	  93.1-90.6 percent
System recovery 	  91.2 percent
Feed flow	5.75 gpm
Brine flow  .	0.51 gpm
Brine flow,  tube 5	4.12 gpm
Recycle brine flow	3.69 gpm
Temperature	44 °F
Product flow	5.24 gpm
Product flow, 50°F	5.86 gpm
Flux rate, 50°F	11.25 g/f2d
Flux rate, 77°F	18.00 g/f2d
Length of test	  100 hr
                                   26

-------
                                             CHEMICAL ANALYSES  FOR 10
        TABLE VI
,000-GPD UNIT  OPERATION AT
                                                                                                 SITE 1,   3/9/70 - 3/13/70
Sample
Designation
Raw feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Blended feed
Brine
Product
Percent rejection
Raw feed

Blended feed
Brine
Product
Percent rejection
Date
3/9/70




3/11/70




3/13/70




3/9 - 3/13/70
means




Time
1700




2000




1100




—





Percent
Recovery
92.1




91'.4




90.9




91.2





pH
2.7
2.2
2.0
3.3

2.5
2.0
1.7
3.3

2.8
2.4
2.3
3.4

2.7

2.2
2.0
3.4

Conductivity
(ymhos/cm)
.1200
4200
8500
450
92.91
1200
4400
10,000
240
96.67
1200
4500
10,000
210
97.10
1190

4210
9542
248
96.39
Acidity,
pH 7.3
585
2478
5864
140
96.64
533
2098
5397
145
96.13
559
2239
5210
107
97.13
633

2584
5914
116
97.27
Hardness
as CaCOj
390
1936
4296
14
99.55
425
1989
4849
10
99.71
379
1854
3938
8
99.72
405

1920
4385
10
99.68
Ca as Ca
(mg/1)
102
475
1050
4.0
99.48
112
475
1200
2.8
99.66
102
438
1000
2.4
99.67
106

482
1102
3.0
99.62
Mg as Mg
(mg/1)
32
180
395
1.0
99.65
35
192
444
0.7
99.78
30
182
345
0.5
99.81
32

180
401
0.7
99.76
Fe
(mg/1)
100
465
1025
3.2
99.57
114
530
1200
3.0
99.65
110
550
1120
2.5
99.70
110

528
1190
2.8
99.67
Al
(mg/1)
33
151
360
1.2
99.53
37
180
450
1.0
99.68
32
180
380
1.3
99.54
35

172
398
1.1
99.61
S04
(mg/1)
819
4026
9555
22
99.68
874
4232
10,784
19
99.75
628
3549
7644
11
99.81
810

4024
9542
16.8
99.75
ro
                  These analyses were made on-site by EPA personnel.

-------
                     TABLE VII
MEAN OPERATING VALUES OF 10,000-GPD UNIT AT SITE 1,
                 3/16/70 - 3/30/70









Flux rate, 50°F, g/f2d . . . .
Flux rate, 77°F, g/f2d . . . .

91 Percent
Recovery
. . . 600
. . . 91.1
. . . 5.42
. . . 0.48
... 4 07
... 3 62
... 45
... 4 94
... 10.48
... 16 77

85 Percent
Recovery
602 5
84 4
5 63
0 69
4 11
3 44
44
4 94
10 52
1 A ft**
192
                         28

-------
                                                          -               TABLE VIII
                                         CHEMICAL ANALYSES1 FOR 10,000-GPD  UNIT OPERATION AT  SITE  1,  3/16/70 -  3/30/70
Sample
Designation
Raw feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Blended feed
Brine
Product
Percent rejection
Date
Means




Means




Time
—




—




Recovery
90.7




84.8




PH
2.8
2.5
2.3
3.1

2.7
2.5
2 3
3.1

Conductivity
(|imhos/cm)
1100
3600
7000
240
95.47
1200
3000
6000
240
94.67
Acidity
pH 7.3
494
1990
4266
40
98.72
539
1730
3887
37
98.68
Hardness
as CaCOj
407
2010
4423
14
99.56
422
1689
3759
11
99.60
Ca as Ca
(mg/1)
102
500
1050
4.0
99.48
107
412
887
3.2
99.51
Mg as Mg
(mg/1)
36
182
432
1.0
99.67
37
158
370
0.7
99.73
Fe
(mg/1)
80
380
860
2
99.68
80
320
680
2
99.60
Al
(mg/1)
31
160
360
1
99.62
31
126
290
1
99.52
304
(mg/1)
819
2594
8463
14
99.75
846
1775
7098
12
99.73
to
                  tiese analyses were made on-site by EPA personnel.

-------
 this  time,  even with  10 mg/1 sodium hexametaphosphate addition, an increase
 in  the  last vessel AP indicated that calcium sulfate precipitation was
 occurring.  The run was repeated with similar results.  Overall system
 recoveries  of approximately 94 percent were achieved during this short test.
 With  this work, Phase IV of the test plan was completed.  Results of Phase IV
 testing can be summarized as follows:

      1.  Recovery levels of 94 percent or so were attainable during short-
         term operations, but not over a longer time period because of cal-
         cium sulfate precipitation.

      2.  Again, the CaSO^ precipitation inhibitor appeared ineffective.

 After completion of Phase IV testing, discussions with the contracting agency
 were held,  and a decision to try a new approach was made.  This was to blend
 the neutralized brine directly back to the feed of the 10,000-gpd unit.  This
 work wafe done by the EPA and is being reported (Ref. 1) elsewhere by EPA
 personnel.

 The work described above concluded the testing under this contract at Site 1.
 Environmental Protection Agency personnel continued to operate the 10,000-
 gpd unit until May 5, 1970 or until the modules had accumulated 770 hr.
 The modules were then removed from the unit, and a module from each tube was
 returned to San Diego for postoperational inspection while the remainder
were stored in copper sulfate solution.   The results of tests performed using
 the five modules are given in Table IX.

 One module, No. 6/11-7,  was opened and visually inspected and was found to
 contain a large amount of debris,  probably pump packing.  Debris samples
were collected for identification and samples of the backing material were
 taken for evaluation.

Results of analyses of the deposits on the backing material from module 6/11-7
are shown in Table X.  The major constituent is iron, as would be expected,
 since if iron ions do get through the membrane or enter via leaky areas, the

                                     30

-------
                         TABLE IX
     POSTOPERATIONAL PERFORMANCE  OF MODULES FROM SITE 1
Module
Number
6/17-5
6/11-7
6/13-1
H-6847
H-6853
A
Orig.
2.43
2.04
2.06
1.73
1.48
As Tested
2.17
1.40
1.50
1.65
1.24
M
-0.02
-0.06
-0.05
-0.01
-0.03
B
Orig.
6.26
6.8
6.55
7.31
2.38
As Tested
12.17
7.09
8.62
4.95
3.33
Percent
Rejection
Orig.
93.84
92.2
92.52
95.37
96.06
As Tested
87.52
88.56
87.23
92.84
93.51
Note:  All tests were run on 2000 mg/1 NaCl solution.
       A « g/cm^-sec-atm x 10~^, water flux; A of 1 x 10~5
           at 600 psi net at 77°F.
       B = cm/sec, salt flux.
                                                      8.65 gpd
M = log-log flux decline slope, based on 770-hr pperation.
                             31

-------
               TABLE X
 SPECTROCHEMICAL ANALYSIS OF BACKING
MATERIAL DEPOSIT1 FROM MODULE 6/11-7
                                   2
                      Concentration
Element                (ug/25 cm2)
  Al	    40
  B   	     4
  Ca	    20
  Cu	   100
  Fe	>1000
  Mg	    60
  Mn	   <10
  Si	   100
  Ti	     4
      Sample was leached with HC1 from
a 25-cm2 piece of backing material,
then diluted for analysis.
     2
      Concentration based on
original sample before dilution.
                 32

-------
higher pH of the product water accelerates the precipitation of iron hydrates.
The small amount of copper present is surprising because previous analysis of
the feedwater had not revealed the presence of any copper.  However, at several
times brass fittings were used in the feed lines and were eventually corroded
away, and this could be the source of the copper.

Table XI shows the result of a spectrochemical analysis of material collected
by scraping large areas of the membrane surface of module 6/11-7.  A large per-
centage of iron is present in this sample, with copper, silica, chrome, man-
ganese, aluminum, molybdenum, sodium, and nickel also present.  The source of
the copper is probably as explained above, while pump corrosion and pump packing
could account for most of the remainder, except for the sodium.  The presence
of sodium in this large amount was to be expected, since the modules were tested
on NaCl solution.  While these elements are present in the sample in rather
large quantities, they were present on the membrane surface in only minute
quantities, which resulted in little flux decline as shown in Table IX.  This
material would have an adverse effect on long-term flux values if allowed to
accumulate in the module.  It is believed that all the iron present was a result
of pump corrosion, not precipitation from the feedwater.  Calcium sulfate does
not appear in Table XI because this module did not have CaSO^ precipitation
present, as it was removed from the reverse osmosis unit when recovery levels
were lower than that where CaSO^ precipitation occurs.  Measurements showed
that the water-carrying capacity of the backing material was still within
specifications for new material.

 Two  of the five modules were soaked  in a neutralized BIZ   solution for 4
 days in an attempt  to clean the membrane of  the fouling material previously
 described.   The modules were drained and replaced in the  same solution at  least
 twice daily.   After this soaking,  the modules  were again  tested  on 2000 mg/1
 NaCl solution.   The results were  as  follows:
                                     33

-------
                  TABLE XI
SPECTROCHEMICAL ANALYSIS OF MEMBRANE DEPOSIT
              FROM MODULE 6/11-7
                          Concentration
   Element                	(ppm)	

     Al	     2000

     Ba	       80

     Co	      100

     Cr	>10,000

     Cu	>10,000

     Fe	>10,000

     Mn	      800

     Mg	      400

     Mo	    2,000

     Na	>10,000

     Ni	    6,000

     Pb	    2,000

     Si	>10,000

     Sn	      200

     Ti	      400

     V   	      200
         Loss  of weight  on  ignition =  21%.
                    34

-------

Original test data
After Site 1 testing
After BIZ soak
Module
6853
A
1.48
1.24
1.32
% Rej.
96.1
93.5
93.4
Module
6/17-5
A
2.43
2.17
2.58
% Rej.
93.8
87.5
85.8
The BIZ cleaning evidently removed most of the fouling material present,
which indicates it was probably pump packing or other foreign material,
not precipitated salts.

The effect of the cleaning on Module 6853 is significant:  the flux decline
slope changed from -0.03, calculated after Norton operation, to-0.01,  after
cleaning, with little or no change in rejection.  The results obtained for
Module 6/17-5 seem to indicate a small leak.  Spiral-wound modules exhibit
compaction slopes of -0.01 to -0.03 when operated on reverse osmosis  permeate,

When the modules were opened and samples of membrane were tested, the follow-
ing results were obtained:

Sample No. 1
Sample No. 2
Module
6853
A
1.70
1.46
% Rej.
94.6
95.0
Module
6/17-5
A
2.99
3.07
' % Rej.
86.9
89.2
The membranes from Module 6853 seemed comparable to new membrane, while
that from Module 6/17-5 showed some degradation.  There was no apparent
reason for the difference in the condition of the membranes.  This work
marked the completion of the evaluation of Site 1 modules.
The results of testing at Site 1 indicate that if the module fouling due to
CaSO^ precipitation and pump packing and corrosion products is prevented,
flux decline values approaching those obtained in the laboratory as base-line
values are obtained.  The method of mbnitoring the last tube or tubes in a
                                    35

-------
unit for AP seems to be adequate for determining when excessive CaSO,  precipi-
tation is occurring.  Low recovery flushes are capable of removing the CaSO/
and restoring system performance.

No iron fouling of the membrane as a result of operating on this ferric acid
mine drainage water was apparent during limited testing.
                                     36

-------
                               SECTION X

                     OPERATION AND RESULTS AT SITE 2

 In order to investigate the performance of reverse osmosis on acid mine
drainages containing a high ferrous iron/ferric iron ratio, the 4,000-gpd
unit was moved to Site 2.   The mine discharge at this site came from the
Arkwright mine of the Christopher Coal Company, about ten miles southwest of
Morgantown, West Virginia.  The water was pumped from a flooded mine shaft and
 discharged into a small stream.  A 500-gal tank was placed at the discharge,
and the water was pumped from this tank through a small sand filter, then
through cartridge filters, and from there to the high-pressure pump.  The
unit was placed inside a small portable storage building, and the pump was
left outside.   The pump used for this test was a Moyno 9S2 that had been recon-
 ditioned by the factory.  All the wetted parts were 316 stainless steel or
 hard  rubber.   The unit was loaded with modules that had been  in  the unit
 during previous  testing at Site  1.  The  loading  arrangement is shown  in
 Table XII.

 Operations began on May 13 and were temporarily suspended on May 15 while
 a concrete pad for the Moyno pump, necessitated by excessive vibration, was
 installed.  During this shutdown, the modules were fouled by material,
 identified as CaSO,, which precipitated from the water on standing.  Another
 set of modules was loaded in the unit, and operation was resumed with one
 of the three tubes performing very poorly.  Probing indicated that two of
 the modules were causing the poor performance, so they were replaced.  Later
 the seal of the booster pump, used to deliver the water from the discharge
 to the Moyno pump, was corroded  through and the unit again lost  flow, with
 resulting precipitation and module fouling.  This time, however, much of
 the material was removed with flushing.  A new booster pump was  installed,
 and the unit was then operated at approximately 50 percent recovery over
 the weekend, during which time a brass valve seat corroded causing low
 system flow and subsequent module fouling.
                                      37

-------
                                TABLE XII
         MODULE LOADING ARRANGEMENT FOR 4,000-GPD UNIT AT SITE 2
Feed
                                                                     Tube  1
                                                                     Tube  2
                                                                     Tube  3
                             5/13/70 - 5/15/70
Feed
                                                                    Tube  1
                                                                    Tube  2
                                                                    Tube 3
                             5/22/70  - 5/2A/70
                                    38

-------
Attempts to flush the unit were not successful, so it was mutually agreed
by EPA and GES personnel to terminate testing at Site 2.  The operational
history of the unit for these two short runs is shown in Table XIII.  Note
that the AP increases with operating time and the product flux decreases,
especially during the last run.  Subsequent examination revealed the
presence of substantial calcium sulfate precipitation, which accounts for
each of these occurrences.

While the incidents described above might seem to indicate that the Site
2 test was not successful, this is not the case.  The purpose of this
test was to determine the difference between operating on high-ferrous-
iron water and operating on the high-ferric-iron water at Norton.  The
feedwater at  Site 2 was about 65 percent  ferrous iron, but the  total  iron
concentration was 30  times greater  than at Norton.   The magnesium,  aluminum,
calcium, and  sulfate  concentrations were  also higher by factors of 15, 11,
6,  and  15, respectively.  This means that the second feedwater  concentration
could be compared with the concentrated brine at 92  percent recovery at
Norton  (see Table XIV).  Further concentrating this  type of water is diffi-
cult, primarily because of CaSO, solubility limitation.  The achievement
of  recoveries of 50 percent during the short time the unit operated seemed
possible, with precipitation of CaSO, occurring on the brine samples in
several hours.  This would indicate that  this level  of recovery is the
maximum obtainable.  Mean operating values and chemical analyses for this test
are shown in Tables XV and XVI.

The salt rejections  at this site were somewhat lower than those reported at
Site 1.   The higher  osmotic pressure at Site 2 resulted in less product flow,
whereas the salt flow was comparable to Site 1.

It seems apparent that reverse osmosis would not be feasible  for demineralizing
acid mine waters containing this amount of pollutants.   The product  water would
not meet potable water standards unless treated further.   This  treatment would
entail more than neutralizing and filtration,  as was the case at Site 1.   The
relatively low recovery  would create a serious brine disposal problem.   Not
enough testing was  carried out to successfully predict  membrane and  module
                                    39

-------
                              TABLE XIII
OPERATIONAL HISTORY OF 4,000-GPD UNIT AT TEST SITE 2, 5/13/70 - 5/24/70



Date
5/13/70
5/14/70
5/14/70
5/15/70
5/21/70
5/22/70
5/23/70
5/24/70
5/24/70
5/24/70


Elapsed Time
(hr)
3.6
22.0
27.8


Pressure
(psi)
400
500
600


Recovery
(percent)
39.5
43.5
49.0

Product
Flow
(gpm)
2.30
2.86
3.58

Brine
Flow
(gpm)
3.52
3.70
3.70

AP,
Tube 3
(psi)
6.7
7.2
7.1


Temp
(°F)
68
64
67
Unit
Water Permeation
Coefficient, AxlO~5
at 77°F
1.99
1.97
1.90
Unit down for repairs; modules fouled during this shutdown and were removed.
New load of modules installed.
2.3
17.8
42.5
46.4
600
600
600
600
51.4
50.0
42.5
46.4
3.84
3.45
3.23
3.13
3.64
3.46
3.57
3.57
7.5
7.0
8.0
8.0
68
62
64
70
1.99
1.95
1.76
1.55
Testing terminated.

-------
                                                   TABLE XIV
                      COMPARISON OF CHEMICAL ANALYSES1FOR OPERATION AT SITE 1 AND SITE 2
Sample
Designation
Date
Percent
Recovery
pH
Conductivity
(ymhos/cm) ,
Acidity,
pH 7.3
Ca as Ca
(mg/1)
Mg as Mg
(mg/1)
Fe
(mg/1)
Al
(mg/1)
S04
(mg/1)
                                        Site 1 (Norton,  West Virginia)
Raw feed
Blended feed
Brine
Product
Percent rejection
3/9/70




92.2




2.7
2.2
2.0
3.3

1200
4200
8500
450
92.91
585
2478
5864
140
96.64
102
475
1050
4
99.48
32
180
395
1.0
99.65
100
465
1025
3.2
99.57
33
151
360
1.2
99.53
819
4026
9555
22
99.68
                                      Site 2 (Morgantown, West Virginia)
Raw feed
Brine
Product
Percent rejection
5/14/70



43.5



2.5
2.4
3.1

7250
11,900
520
94.57
5120
8890
180
97.43
625
1030
20
97.58
480
920
15
97.86
1180/
33602
2214/
50882
56/
712
98.32
380
700
12
97.78
12,968
20,475
273
98.38
These analyses were made on-site by EPA personnel.

Ferrous iron/total iron.  Rejection based on total iron.

-------
                              TABLE XV
MEAN OPERATING VALUES OF 4,000-GPD UNIT AT SITE 2, 5/13/70 - 5/24/70

       Pressure	600  psi

       Unit  recovery	49.0 percent

       Feed  flow	6.95 gpm

       Brine flow	3.54 gpm

       Temperature	66°F

       Product flow	3.41 gpm

       Flux  rate, 77°F	13.11 g/f2d

       Tube  3 AP	11.0 psi

       Length of test	46.4 hr
                                 42

-------
                                                     TABLE XVI


                    CHEMICAL ANALYSES  FOR 4,000-GPD UNIT OPERATION AT SITE 2,  5/13/70 - 5/24/70
Sample
Designation
Raw feed

Brine

Product

Percent
rejection-
Percent
Recovery
49.0






pH
2.24


2.00

3.14

Conductivity
(ymhos/cm)
7,040

11,840

423

95.52
Acidity,
pH 7.3
5,230

10,050

148

98.06
Dissolved
Oxygen
4.5

4.5

0

0
Ca as
Ca
526

933

9.6

98.68
Mg as
Mg
420

810

7.6

98.76
Fe2
(mg/1)
1,280/
2,300
2,450/
4,460
29/
39
98.84
Al
(mg/1)
317

598

5.0

98.91
so4
(mg/1)
10,920

20,480

191

98.78
•p.
CO
              These analyses were made on-site by  EPA personnel.
             2
              Ferrous iron/total iron. Rejection based  on total iron.

-------
                              TABLE XVII
   RESULTS OF POSTOPERATIONAL TESTING OF MODULES OPERATED AT SITE 2
Module
6844
6851
68631
68711
68721
6/11-101
6/11-131
6/13-32
6/13-7
6/16-4A2
6/16-8
6/16-10
6/17-31
6/17-7
Original Test Data
A
1.58
1.41
1.50
1.47
1.56
2.08
2.10
2.13
2.51
2.04
2.12
2.01
2.60
2.63
Percent
Rej.
95.6
96.5
95.8
96.7
96.4
95.3
95.1
91.3
91.3
90.4
95.1
93.4
93.1
91.9
6/18/70 Test Data
A
1.42
1.38
1.24
1.40
1.24
1.82
1.87
1.84
1.25
1.91
1.58
1.44
2.44
1.95
Percent
Rej.
95.0
95.6
92.3
91.9
93.8
80.0
81.5
90.9
48.6
92.7
91.1
70.2
85.1
89.1
      This module was operated at Site 1 for approximately 250 hr
before the unit was moved to Site 2.

     2
      This module was operated at Site 2 for less than 24 hr.  It
was removed from unit because of poor performance.


NOTE:  All tests were run on 2000 mg/1 NaCl solution.
       A = g/cm2-sec-atm x 10~5; A of 1 x 10~5 =8.65 gpd at
           600 psi net at 77°F.
                                 44

-------
lifetimes.  Fouling caused by the ferrous iron present in the acid mine water
was not apparent, but due to the short exposure times and massive calcium
sulfate fouling, no definitive statement regarding iron fouling can be made.

The EPA has prepared an in-house report (Ref. 9) in which the performance
(A values and AP) of the individual tubes  is  reported for the testing at
this site.

Since the recovery level was limited at Site 2, it was decided to select
a third site, where the water would have a high ferrous iron/ferric iron
ratio but a total iron content similar to that at Site 1.

Fourteen of the modules used in this test were returned to San Diego for
postoperational evaluation and cleaning.  The modules were tested for '
rejection and water flux on 2000 mg/1 NaCl solution.  The results of
these tests are shown in Table XVII, along with the original quality
control test data.  All modules experienced a decline in "A" values and
also in rejection, with a few undergoing greater declines than others.

Module 6/13-7 is an extreme case, but this module lost the brine seal,
probably during loading, and was virtually full of precipitated material.
Samples of this material were collected from the pressure vessel when the
module was removed and, along with a sample of the precipitate in a brine
sample at 50 percent recovery, were analyzed.   Results of these analyses
are shown in Table XVIII.  In both cases,  the material is almost pure gypsum
(CaSO^*2H20), which again points out that  the limiting substance governing
recovery was the calcium sulfate, rather than iron hydr\ate.

Two of the returned modules (6/11-10 and 6/13-7) were run on acidified (pH
3 to 4) San Diego tap water for two days.   These modules were known to be
loaded with calcium sulfate, with module 6/13-7 being much worse in this
respect than module 6/11-10.  The modules  were then tested on 2000 mg/1
NaCl solution.  Results obtained were as follows:
                                     45

-------
                        TABLE XVIII
        CHEMICAL ANALYSIS OF SAMPLES FROM SITE 2
    Solid Material Collected from Tube No. 3, 5/17/70
                                    Concentration
                                      (percent)
                Fe	<0.3
                Mg	   1.6
                SO^	60.8
                Ca	21.7
Precipitate from Brine Sample, 5/24/70, ^50 Percent Recovery
                                    Concentration
                                       (percent)
                C03	<0.5
                Fe	   0.70
                Mg	  <0.5
                SO^1	63.0
                Ca1	 .  25.0
                Al	<0.3
                Clay and silica  ...   1.5
        Present in stoichiometric amounts found in CaSO,.
                            46

-------

Original test
After Site 2
After low-pH
data
operation
flush
Module
6/11-10
A
2.08
1.82
1.92
Percent
Rej .
95.3
80.0
80.7
Module
6/13-7
A
'2.51
1.25
2.15
Percent
Rej.
91.3
48.6
87.4
Module 6/13-7 had been severely packed and weighed two or three times
more than a new module.  As evidenced by the te.st results, much of this
material was removed by the flushing.  The low rejection value after
Site 2 operation was caused by the extremely high boundary layer concen-
tration in the brine channels, and as this concentration was reduced, the
rejection improved.  Further flushing might have increased the water flux
and rejection even more, but no further flushing was performed.

Modules 6/11-10 and 6/13-7 were opened, and the remaining CaSO^ was removed
by rinsing with distilled water.  Membrane samples were taken for testing in
the membrane test cells using 2000 mg/1 NaCl solution.  The results of these
tests were as follows:

Sample No. 1
Sample No. 2
Module
6/11-10
A
2.77
2.47
Percent
Rej.
85.66
87.42
Module
6/13-7
A
2.30
Percent
Rej.
91.45
These results indicate that the CaSO,  present in the modules,  even after
flushing, affects rejection values due to the boundary layer concentration.
The membrane from Module 6/13-7 has rejection values equivalent to the
original module test data.  Module 6/11-10 results indicate a slight deteriora-
tion of the membrane as compared with the original test data.   This module had
been operated at Site 1 for 250 hr and was in the unit when it was moved to
Site 2,  No reason for this decrease in performance is apparent.

No further work was performed on any of the remaining modules.
                                      47

-------
                               SECTION XI
                      OPERATION AND RESULTS AT SITE 3

The site selected for the third series of tests was about four miles east of
Ebensburg, Pennsylvania.  The water was discharged from active mines of
the Cambria Division of the Bethlehem Mines Corporation.  The discharge
at this location was over 4 mgd and contained about 130 mg/1 of iron,
about 80 percent of which was ferrous.

The 4,000-gpd unit was moved to this site and set up, with the high-
pressure pump, 500-gal. feed tank, filters, etc., in an arrangement
similar to that used at Site 2.  Modules were loaded as shown in Table XIX.
Provisions were made to add chemicals to the feed to the reverse osmosis unit.
In reviewing data from past successful operations at various locations, it was
noted that in all cases, the pH of the feed was below 3, whereas at Site 3,
the mine discharge pH was 3.6.  No iron precipitation problems had been
encountered in any Gulf Environmental Systems previous runs, but other in-
vestigations (Ref. 2) had encountered precipitation, and it was observed that
this appeared to occur when operating on feeds with a pH above 3.  The reaction
which results in iron precipitation is

        JFe(H?0)>.    + H«0 - IFe(OH) (H»0)_    + H

                               Successive steps
                               Fe(OH)3


The pH at which  there  is 50 percent conversion of the   Fe(H-O),     to
 P           12+                                      L       J
 jFe(OH)(H_0)5     is 2.8, which is  the pK value.  Below pH  2.8,  there  is pre-
dominantly  the trivalent ion, which is highly soluble;  above 2.8j  there  is more
of the other species,  which results in the  favoring of the  precipitation of
                                    49

-------
                                TABLE XIX
         MODULE LOADING ARRANGEMENT FOR 4,000-GPD UNIT AT SITE 3
Peed



<
4338

4350

43531




4337

4349

4354




4319

4346

4355




                                                                     Tube 1
                                                                     Tube 2
                                                                     Tube 3
4353 replaced by 4320 on 8/7/70.
4320 replaced by 4327 on 8/27/70.
4327 replaced by 4328 on 9/4/70.
                                   50

-------
Fe(OH).,.  Also, the conversion rate of ferrous (soluble) to ferric iron
increases rapidly with increasing pH.  It was felt that since this
particular discharge was over pH 3, precipitation of iron with the
resulting fouling of the modules was a distinct possibility.
In order to evaluate the effects of pH, four test runs were conducted at
this site.  The principal parameters of these runs were:

     Run 1     Sulfuric acid added to feed to lower pH.  Unit operated at
               ^85 percent recovery with brine recycle .

     Run 2     Nothing added.
               Unit operated at ^85 percent recovery with brine recycle.

     Run 3     Caustic added to feed to increase pH.  Unit operated at
                   percent recovery with brine recycle.
     Run 4     Nothing added.
               Unit operated at ^50 percent recovery with no brine recycle.

In run 1, sulfuric acid was initially added to lower the blended feed pH
to 2.5; acid addition was slowly decreased until the blended feed pH was
2.8.  The unit was operated at approximately 85 percent recovery for 200 hr.
Results of chemical analysis for this run are shown in Table XX and the
operational history is shown in Table XXI.

There was an increase of AP in the last pressure vessel, accompanied by a
decrease in product flow, but since this condition could be corrected by
operating at 50 percent recovery for ^3 hr as shown in Table XXI, calcium
sulfate precipitation, not iron precipitation, was thought to be the cause.

An inhouse report has been prepared by the EPA (Ref . 10) in which the per-
formance ("A" values and AP) of the individual tubes is reported for the
testing at this site.
                                    51

-------
                                                                          TABLE XX

                                               CHEMICAL ANALYSES  FOR 4,000-GPD UNIT OPERATED AT SITE 3, RUN 1

Sample
Designation
Raw feed
Acidified feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Acidified feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Acidified feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Acidified feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Acidified feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Acidified feed
Blended feed
Brine
Product
Percent rejection


Date
7/30/70





7/31/70





8/3/70





8/3/70





8/6/70





8/7/70





Elapsed
Time
(hr)
4.0





23.5





96.0





100.0





168.0





191.0







PH
3.7
2.7
2.4
2.2
3.3

3.6
2.7
2.4
2.2
3.5

3.6
2.75
2.4
2.3
3.7

3.6
3.1
2.8
2.6
4.2

3.6
3.1
xrt
2.6
4.3

3.6
3.1
2.9
2.7
4.5


Conductivity
Qiiihos/cm)
1,600
2,300
6,400
9,700
205
97.45
1,500
2,000
5,000
8,000
148
97.72
1,600
2,150
4,900
7,300
102
98.33
1,480
1,750
4,200
6,800
63
98.85
1,460
1,650
4,000
6,200
60
98.82
1,440
1,600
3,900
5,900
57
98.84

Acidity,
pH 7.3
410
560
1,680

Ca as Ca
(mg/1)
170
168
600
3,280 1,008
70 1.7
97.18
390
540
1,540
3,020
70
96.93
No value
reported




390
400
1,160
2,150
50
96.98
380
420
1,200
2,130
50
97.00
390
430
1,210
2,010
50
96.89
99.79
168
160
600
1,020
1.6
99.80
200
196
600
900
1.2
99.84
200
196
600
900
1
99.87
200
200
610
870
1.1
99.85
190
190
600
880
1
99.86

Mg as Mg
(mg/1)
48
48
190
336
0.6
99.77
49
50
185
324
0.6
99.76
59
57
188
340
0.3
99.89
57
57
186
340
0.3
99.89
55
54
208
360
0.3
99.89
55
54
204
304
0.3
99.88

Al
(mg/1)
28
27
101
180
0.6
99.57
29
28
100
168
0.6
99.55
34
33
106
192
1
99.33
34
35
104
188
1
99.32
33
33
104
180
1
99.30
33
32
103
168
1
99.26
Total
Fe
(mg/1)
124
124
450
840
0.8
99.88
128
124
440
900
0.8
99.87
156
188
450
750
2
99.67
136
132
420
700
2
99.64
136
126
425
700
2
99.64
140
130
440
700
2
99.65
Ferrous
Iron
(mg/1)
101
101
354
624
99.8
101
101
325
574
<1
>99.8
101
101
292
522
<2
99.26
101
101
281
511
<2
99.24
101
101
320
523
<2
99.30
101
101
308
510
<2
98.53

S04
(mg/1)
1,636
1,633
6,552
12,012
32
99.66
1,665
1,583
6,006
11,193
25
99.71
1,638
1,638
4,914
8,463
6
99.91
1,638
1,747
4,914
8,463
12
99.82
1,638
1,638
4,914
8,463
12
99.82
1,638
1,638
4,914
8,327
11
99.84
Dissolved
Oxygen
(mg/1)
2

















2


2
2

2


2
2

2


2
2


Percent
Recovery
85.5





84.8





82.4





82.5





84.2





84.7





ui
ro
                These analyses vere made by EPA personnel.

-------
                                                         TABLE XXI
                                 OPERATIONAL HISTORY OF 4,000-GPD UNIT,  TEST SITE 3,  RUN 1

Date
7/30/70



7/31/70

8/1/70
8/3/70
8/3/70
8/4/70

8/5/70


8/6/70


8/7/70


Elapsed Time
(hf)
0.25
1.5
4.0
7.5
23.5
28.0
57.7
96.0

Pressure
(psi)
400
420
410
410
395
390
390
390

Recovery
(percent)
48.3
81.3
85.5
85.7
84.8
84.9
84.3
82.4

Product Flow
(gpnt)
3.33
3.45
3.17
3.22
3.13
3.13
3.06
2.63

Brine Flow
(gpm)
3.57
0.80
0.54
0.54
0.56
0.56
0.55
0.56

AP,
Tube 3
(psi)
8.8
10.8
10.8
11.2
10.6
10.4
11.1
12.5

Temp
(°F)
59.5
62.5
64.5
64.5
63.5
66.5
65.0
62.5
Unit
Water Permeation
Coefficient, AxlO
at 77'F
2.76
2.76
2.68
2.69
2.70
2.58
2.60
2.32

Raw
Feed
pH
3.5
3.6
3.7
3.7
3.6
3.6
3.7
3.6

Blendec
Feed
PH
2.5
2.5
2.4
2.4
2.4
2.5
2.4
2.4
Ran for 2-3/4 hr at 50 percent recovery.
100.0
120.0
125.0
144.0
147.0
150.0
168.0
171.0
174.0
191.0
400
375
405
400
400
400
400
400
400
400
82.5
81.2
83.4
83.7
84.0
84.0
84.2
83.4
83.6
74.0
3.12
2.86
3.12
2.86
2.94
2.94
2.70
2.70
2.79
2.50
0.66
0.66
0.67
0.56
0.56
0.56
0.55
0.56
0.56
0.55
10.3
9.2
10.4
10.0
10.2
10.3
11.1
11.5
11.8
12.8
66.0
59.0
66.5
59.5
67.5
67.5
60.5
62.0
66.0
61.0
2.45
2.67
2.38
2.46
2.27
2.27
2.32
2.27
2.20
2.12
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
2.8
2.8
2.7
3.1
2.7
2.6
2.8
2.7
2.7
2.9
Unit down, last module of Tube 3 removed and replaced with a new one.
Calcium sulfate present in Tube 3. System flushed with acidified permeate.
en

-------
One module, No. 4353, was removed at the end of 200 hr.  An increase in the
AP of  the  last pressure vessel had occurred, and it was assumed that this
module would have precipitated material present.  When the module was opened,
it was indeed full of white, flaky precipitate.  Analysis of this material
showed it  to be 23.7 percent calcium  and 55.1 percent sulfate, with a 22.6
percent loss of weight at 600°C.  The theoretical amount of CaSO* in gypsum
(CaSO^-2H20) is 79.1 percent.  Thus, this material was almost pure gypsum.
The precipitate could easily be washed from the membrane, which is what
takes  place in the unit when the recovery level is lowered.  No presence of
iron precipitate was noted.

With the relatively small reverse osmosis unit, a large amount of the brine
is recycled in order to achieve the high recovery.  This recycled brine is
of a lower pH than the raw feed, so the blended feed is still lower in pH
than the raw feed.  During run 2, the unit was operated in this manner with
no feed additives for 86 hr.

Chemical analyses and operational history for this run are shown in Tables
XXII and XXIII.  Note that although there was an increase in the AP of Tube
3, the unit water flux remained relatively constant.  In fact, after each
shutdown due to pump failure, the "A" value increased.  This phenomenon has
been noted many times in the operation of reverse osmosis units containing
spiral-wound modules and is not readily explained.  No iron precipitation
was noted during this run.

In run 3 an attempt was made to bring the blended feed pH up to the raw
feed pH by adding sodium hydroxide to the feed stream.  This resulted in
mass precipitation of iron in the sand filter, which plugged rapidly.  The
chemical analyses are shown in Table XXIV and the operational history is shown
in Table XXV.  Because of the massive amounts of iron precipitation throughout
the system, the only conclusion reached was that if the pH of this feed is
increased by caustic, iron precipitation will surely occur.
                                     54

-------
                                                                             TABLE XXII

                                                    CHEMICAL ANALYSES  FOR 4,000-GPD UNIT OPERATED AT SITE 3, RUN 2
on
en

Sample
Designation
Raw feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Blended feed
Brine
Product
Percent rejection
Raw feed
Blended feed
Brine
Product
Percent rejection

Date
8/11/70




8/17/70




8/18/70




8/19/70




8/20/70




8/21/70




Elapsed
Time
(hr)
3.0




2.8




24.7




48.5




66.0




86.3





pH
3.7
3.1
2.9
4.6

3.6
3.1
2.9
4.5

3.6
3.1
2.9
4.4

3.7
3.3
3.1
4.7

3.6
3.1
3.0
4.7

3.6
3.1
3.0
4.6


Conductivity
(ymhos /cm)
1460
3900
7200
78
98.59
1480
4200
6800
108
98.03
1490
4100
7600
91
98.44
1460
4000
6400
81
97.62
1460
4000
6600
84
97.60
1460
3800
6000
71
98.55

Acidity
pH 7.3
390
1200
2160
50
97.02
400
1160
2280
60
96.51
400
1220
2200
60
96.49
380
1090
2020
60
96. t4
370
1150
2210
60
96.43
370
1140
2010
60
96.19

Ca as Ca
(mg/1)
185
650
1200
1.5
99.84
220
680
1260
2.4
99.25
230
600
1100
1.8
99.79
210
630
1170
1.6
99.82
210
640
1180
1.8
99.80
200
690
1280
1.6
99.84

Mg as Mg
46
148
300
0.6
99.73
69
200
425
1.6
99.49
69
222
425
1.2
99.63
67
216
390
1.0
99.67
66
204
380
1.2
99.59
66
222
390
1
99.67

Al
(mg/1)
27
85
180
1
99.24
40
124
255
1
99.47
38
114
200
1
99.36
38
110
225
1
99.40
40
120
230
1.2
99.31
40
116
220
1.2
99.29
Total
Fe
(mg/1)
114
390
690
2
99.63
140
460
900
3
99.56
154
450
890
2
99.70
150
470
900
1.7
99.75
170
490
900
1.8
99.74
168
500
950
1.1
99.85
Ferrous
Iron
(mg/1)
101
286
545
>2
98.56
101
292
570
>2
98.61
101
314
600
>2
98.69
96
292
567
<2
98.60
96
292
534
<2
98.55
96
275
500
<1
98.45

(mg/1)
1583
4505
9282
21
99.70
1693
6006
10238
27.3
99.66
1747
4914
10238
1
99.72
1693
5324
10238
23
99.70
1693
4778
10101
25
99.66
1747
5870
10238
22
99.73
Dissolved
Oxygen
2

2
2

2

2
2

2

2
2

2

2
2

2

2
2







Percent
Recovery
84.3




84.5




84.4




82.8




84.3




84.4




                       hese analyses were made by EPA personnel.

-------
                                                                TABLE XXIII

                                        OPERATIONAL HISTORY OF 4,000-GPD UNIT TEST SITE 3, RUN 2
Date
8/11/70



8/17/70



8/18/70



8/19/70



8/20/70





8/21/70

Elapsed
Time
(hr)
0.5
3.0
6.5
Pressure
(psi)
400
400
400
Recovery
(percent)
84.7
84.3
86.0
Product Flow
(gpm)
3.33
3.33
3.33
Unit down; main high-pressure-pump failure.
0.7
2.8
5.9
400
400
400
84.5
84.5
84.0
3.33
3.33
3.23
Unit down; main high-pressure-pump failure.
22,9
24.7
27.8
400
400
400
84.0
84.4
84.0
3.22
3.22
3.13
Unit down; main high-pressure-pump failure.
46.5
48.5
51.0
400
400
400 .
83.9
82.8
83.8
3.33
3.13
3.18
Unit down; main high-pressure-pump failure.
63.5
66 .'0
68.5
400
400
400
83.9
84.3
85.2
3.13
3.13
3.13
Unit down; main high-pressure-pump failure.
70.4
72.5
86.3
400
400
400
83.5
83.4
84.4
3.22
3.22
3.03
Brine Flow
(gpm)
0.60
0.60
0.55

0.62
0.62
0.62

0.62
0.60
0.60

0.64
0.65
0.62

0.61
0.59
0.55

0.64
0.65
0.57
AP,
Tube 3
(psi)
9.2
9.2
9.2

8.7
8.8
8.8

8.9
8.9
8.9

8.9
8.9
8.9

8.9
9.1
9.2

9.2
9.3
9.8
Temp
C'F)
59.5
67.0
68.0

66.5
67.5
67.0

64.0
67.5
67.0

64.5
65.0
68 .-0

62.5
64.5
67.0

64.0
65.0
60.0
Unit Water
Permeation
Coefficient, AxlO"5
at 77°F
2.85
2.57
2.57

2.56
2.55
2.49

2.67
2.49
2.44

2.63
2.48
2.41

2.58
2.51
2.41

2.60
2.56
2.60
Raw
Feed
PH
3.6
3.7
3.6

3.6
3.6
3.6

3.6
3.6
3.7

3.7
3.7
3.6

3.6
3.6
	

	
	
	
Blended
Feed
pH
3.1
3.1
3.1

3.1
3.1
3.1

2.5
3.1
3.2

3.3
3.3
3.3

3.2
3.1
	

	
	
	
Unit recovery lowered to 50 percent; unit then flushed with acidified permeate and shut down.
01
CTk

-------
                                          CHDttCAL ANALYSES  FOR 4,
                                                                  TABLE XXIV
                                                                  000-GPD TOUT OPERATED AT SITE 3, RUN 3
Sample
Designation
Raw feed
Blended feed
Brine
Product
Percent rejection
Saw feed
Caustic feed
Blended feed
Brine
Product
Percent rejection
Date
8/25/70




8/26/70





Elapsed
Time
(hr)
16.4




29.6





PH
3.8
3.3
3.1
4.7

3.6
4.1
3.6
3.4
4.6

Conductivity
(pmhos/cm)
1460
3400
5300
90
97.93
1460
1500
4000
6400
105
97.98
Acidity
pH 7.3
350
990
1790
40
97.12
390
360
870
1700
60
95.33
Ca as Ca
(mg/1)
225
660
1188
2
99.78
200
184
600
960
2.2
99.72
Mg as Mg
(mg/1)
62
180
350
1
99.62
58
59
195
380
1.2
99.58
Al
(mg/1)
38
104
200
1.4
99.08
32
38
78
145
1.4
98.74
Total
Fe
(mg/1)
165
400
740
1
99.82
136
150
425
765
1.2
99.80
Ferrous
Iron
(mg/1)
96
253
478
<1
99.18
107
101
281
551
<1
98.56
S04
(mg/1)
1583
3822
7644
27.3
99.52
1474
1529
3549
7098
27
99.49
Dissolved
Oxygen
(mg/1)
2

2
2

2


3.5
4.5

Percent
Recovery
82.6




83.0


„


01
            These analyses were made by EPA personnel.

-------
                                                              TABLE XXV

                                        OPERATIONAL HISTORY OF 4,000-GPD UNIT, TEST  SITE  3,  RUN  3
Date
8/25/70

8/26/70



Elapsed
Time
(hr)
1.5
16.4
17.6
24.1
27.7
29.6
Pressure
(psi)
400
400
Recovery
(percent)
79.9
82.6
Product
Flow
(gpm)
3.22
3.13
Brine
Flow
(gpm)
0.90
0.66
AP,
Tube 3
(psi)
8.5
8.5
Temp
(°F)
64.0
59.0
Unit Water
Permeation
Coefficient, AxlO~5
at 77°F
2.59
2.68
Raw
Feed
PH
—
3.6
Blended
Feed
PH
—
3.3
Started adding caustic.
400
400
400
83.5
83.4
83.0
3.33
3.22
3.13
0.67
0.65
0.64
8.1
8.1
8.4
65.0
65.5
64.0
2.61
2.52
2.52
3.7
3.7
3.6
3.3
3.6
3.6
Run terminated; system full of iron and calcium sulfate precipitate. Module removed from Tube 3
and replaced with new one. System flushed with acidified permeate.
tn
CD

-------
The Moyno high-pressure pump had failed just prior to this run and was
repaired using a precipitation-hardening stainless steel in place of the
316 stainless steel previously used.  It was felt that the 316 stainless
steel did not possess adequate tensile strength for this use.

The final test run was conducted with no recycled brine and no chemical
addition, so that the natural raw feed pH was not changed.  The system
performed very satisfactorily, as expected at the low recovery level, with
no plugging or fouling observed.  This is different from the results of
runs 1 and 2, where fouling is quite evident.  The results of chemical analy-
ses for this run are shown in Table XXVI, and the operational history is
jiven in Table XXVII.  The conductivity rejection averaged 98 percent or
better throughout the tests performed at Site 3, and the salt rejections
were greater than 99 percent except for aluminum.

The presence of 2 mg/1 of dissolved oxygen in the feedwater had no apparent
effect on the reverse osmosis operation at this test site.  The dissolved
oxygen was present in both the brine and product in amounts equal to the
feed, which would indicate no rejection for this module.

Module 4327 was removed at the end of this run and returned to San Diego for
testing.   During shipment, the shipping container was broken and the module
was fairly dry upon arrival.  The module was tested for rejection and water
flux, with the following results:

Original test data
After Site 3 operation
Module No. 4327
A
2.05
2.19
Percent Rej . on
2000 mg/1 NaCl Solution
95.48
92.58
After this test the module was opened, and inspection revealed a very small
amount of foreign material present on the membrane which in no way affected
performance.  Not enough of the material could be collected for analysis,  but
it was determined that iron was present.  However, it is not clear why traces

                                    59

-------
                             CHEMICAL ANALYSES  FOR
    TABLE XXVI
4,000-GPD-UNIT OPERATION AT SITE 3, RUN 4
Sample
Designation
Raw feed
Brine
Product
Percent rejection
Raw feed
Brine
Product
Percent rejection
Raw feed
Brine
Product
Percent rejection
Raw feed
Brine
Product
Percent rejection
Raw feed
Brine
Product
Percent rejection
Date
8/31/70



9/1/70



9/2/70



9/3/70



9/4/70



Elapsed
Time

-------
                                                       TABLE XXVII
                               OPERATIONAL HISTORY OF  4,000-GPD  UNIT AT  TEST  SITE  3,  RUN 4
Date
8/31/70


9/1/70



9/2/70


9/3/70


9/4/70
Elapsed
Time
(hr)
0.5
2.5
7.9
21.4
24.4
27.4
30.3
45.3
48.1
52.0
69.4
71.9
75.2
92.0
Pressure
(psi)
405
405
405
400
400
400
400
400
400
400
400
400
400
400
Recovery
(percent)
54.8
54.8
53.8
53.8
53.2
52.9
52.5
52.9
52.3
52.3
52.3
53.9
52.3
52.3
Product
Flow
(gpm)
3.46
3.46
3.33
3.33
3.33
3.33
3.33
3.33
3.33
3.33
3.33
3.39
3.39
3.33
•Brine
Flow
(gpm)
2.86
2.86
2.86
2.86
2.94
2.99
3.03
2.99
3.03
3.03
3.03
3.03
3.03
3.03
AP,
Tube 3
(psi)
6.9
6.8
6.9
6.9
7.3
7.4
7.4
7.4
7.4
7.4
7.4
7.4
7.4
7.4
Temp
(°F)
62,0
63.0
61.5
54.0
62.0
63.0
62
55
57
63
56
62
63
60
Unit Water
Permeation
Coefficient, AxlO"5
at 77°F
2.56
2.52
2.51
2.88
2.52
2.48
2.52
2.84
2.75
2.47
2.79
2.56
2.51
2.61
Raw
Feed
pH
3.8
3.9
3.6
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.6
3.6
3.6
3.6
crt
                   Run terminated.  Module removed from Tube
                   Unit flushed with acidified permeate.
3 and replaced with new one.

-------
 of  iron occurred.  A certain amount of the mine drainage water was presumably
 present in  the  feed channels of the module when it was shipped.  Dissolved
 iron could  have been present then, but it may have been deposited when the
 container broke and the module drained.  Since the module was allowed to dry
 out somewhat, no definite conclusion can.be drawn, but the amount of iron
 was negligible.

 Testing at  Site 3 concluded the field testing portion of the contract.  The
 4,000-gpd unit, pumps, tanks, etc., were then moved back to the Norton, West
 Virginia, site where they will be operated as desired by the EPA.

 The test program demonstrated the ability to operate a reverse osmosis
 system without iron fouling on predominately ferrous iron feed by observing
 proper pH control and calcium sulfate limitations.  Product water quality was
 good and would meet potable water standards with lime neutralization and
 filtration.

The high pressure pump performed adequately during this last run, but not
 enough hours were accumulated to fully evaluate the precipitation hardening
steel used for the connecting rod.   In any event,  this type of pump does not
seem adequate for reverse osmosis operation when the feed is acid mine water.
                                     62

-------
                                SECTION XII
                        DISCUSSION OF TEST RESULTS

If the system at Site 1 had been kept free of CaSO^ precipitation by operating
at a recovery level lower than that at which precipitation occurs, and if the
system had been equipped with a high-pressure pump suitable for use on acid
mine drainage that did not corrode or allow packing to work its way into the
process stream, the flux decline values would be similar to those obtained
in the laboratory  for base-line compaction.  For this type of feedwater,
module lifetimes and costs can be extrapolated.  Reference 5 contains a
parametric cost study based on these conditions.

Operations at Site 3 were of too short a duration to make those extrapolations,
while the results of the operation at Site 2 cannot be considered because of
the CaSO, precipitation and pump failures.  Treatment of acid mine water by
reverse osmosis is clearly feasible at Site 1, probably feasible at Site 3,
and not feasible at Site 2.

No difference in reverse osmosis operation was noted when operating on pre-
dominantly ferrous or ferric iron acid mine waters.  Iron fouling of the
membranes as a result of iron precipitation from the acid mine water was not
a critical factor during the short-term runs at the three test sites.
Figure 3 is a log-log plot of the water permeation coefficient, "A," for the
10,000-gpd unit at Site 1 and the 4,000-gpd unit at Sites 2 and 3.  Much of
the scatter of these data is caused by apparent discrepancies in the tempera-
ture - water flux normalization values used in these tests.

Membrane fouling and compaction usually appear as -a straight line when plotted
in this manner.  The fact that three of the lines plotted tail off rather
sharply suggests that module plugging and/or mass fouling was taking place.
Pressure drop readings taken at these times indicate plugging was indeed occur-
ring.  Modules removed at the end of these runs at Site 2 and Site 3 had cal-
cium sulfate precipitation present.
                                     63

-------
en
J>
o

u_  
-------
Calcium sulfate precipitation was a distinct problem at all three test sites,
with the recovery level being the controlling parameter.  These levels are
discussed in Section XIII, and further long-term testing would be required
to establish the validity of the assumptions made there.

The product water at Sites 1 and 3 would meet potable water standards after
being neutralized with lime and filtered.  The product water at Site 2 would
require even further treatment to meet the potable standards.

The presence of dissolved oxygen at Sites 2 and 3 had no effect on the reverse
osmosis operation.
                                   65

-------
                               SECTION XIII
                   CALCIUM SULFATE SOLUBILITY LIMITATION

It was assumed that iron concentration in the .brine with resulting
precipitation and membrane fouling would be the limiting factor in
the total water recovery of a reverse osmosis system operating on acid
mine drainage.  This has not been the case at locations where the spiral-
          CD
wound ROGA  modules have been operated.  Another assumption has been
                                                   I [
that it is more difficult to operate on ferrous (Fe  ) iron water than
             I | [
on ferric (Fe   ) iron water.  Early tests in Pennsylvania (Ref. 8) did
not support this contention, and no iron fouling was observed.  Indeed,
the one limiting factor seems to be the calcium sulfate concentration.
While tests have shown that the CaSO, solubility limit can be exceeded
in the concentrated brine, there are indications that a fairly definite
limit exists.

Table XXVIII compares results of tests conducted on three very different
acid mine drainage waters and a calculated recovery for one of them.
Operations at Norton were conducted with no major fouling or plugging of the
modules as long as the recovery level was kept below 90 percent.  However, if
the recovery was allowed to increase to 92.3 percent through only a slight
change in operating parameters on that particular reverse osmosis unit
(equal to an increase in feed concentration from 10 to 13), plugging of
the downstream tube of modules due to CaSO, deposition was observed.
Tests at the second test site near Morgantown, West Virginia, at the
Arkwright Mine,  where the discharge was approximately ten times more
concentrated than the feed at Norton, resulted in several modules being
completely plugged with calcium sulfate when the unit was operated at
50 percent recovery.  Tests at the third site near Ebensburg, Pennsylvania,
reinforced these previous results.
                                   67

-------
                    i                         TABLE XXVIII

                  COMPARISON OF CALCIUM SULFATE SOLUBILITY FOR DIFFERENT REVERSE OSMOSIS BRINES
00



Brine
Concentration
Ca, mg/1
S04, mg/1
Ca, mol/1
SO. , mol/1
M-
P
me
JPmc/KBP

Norton,
80
Percent
Recovery
551
4,771
13.78xlO~3
49.69xlO~3
6.85xlO"4
1.78

Norton,
90
Percent
Recovery
1,102
9,542
27.55xlO~3
99. 40xlO~ 3
27.38xlO~4
3.56
Calculated
Norton,
92.3
Percent
Recovery
1,433
12,405
35.82xlO~3
129.21xlO~3
46.28xlO~4
4.63

Arkwright ,
50
Percent
Recovery
933
20,480
23.32xlO~3
213.33xlO~3
49.75xlO~4
4.80

Ebensburg ,
84.7
Percent
Recovery
870
8,463
21.75xlO~3
88.15xiQ~3
19.17xlO~4
2.98
                    Note:   P


                           K
                            sp
                        me'  SP
product of molar concentration = (mol/1 Ca)(mol/1 SO,)


theoretical solubility of product, assuming 2000 mg/1 CaSO,

                    -4
saturation = 2.16x10


ratio of P   to K0
          me     SP

-------
From these test results, a relationship between the product of the molar
concentration (P  ) of calcium and sulfate in the brine and the theoretical
                me
saturation (K  ) for CaSO, (neglecting pH, other ions, etc.)' has been
             sp          H
established.  The solubility of CaSO, in distilled water is approximately
                                                     -4
2000 mg/1.  The solubility product, K  , is 2.16 x 10  , based on this
                                                                   2+
                                     sp
solubility value.  Thus, if supersaturation is attained, and the Ca
       2_
and SO,    concentrations are known, the percent of saturation, neglecting
the presence and strength of other ions, may be calculated as follows:
               Percent Saturation = 100
                                          P  /K
                                           me  sp
when P   is the product of the molar concentration of calcium and sulfate
      me
ions.  For reverse osmosis purposes, in determining feasible water recovery
rates (feed concentration factors), the calcium and sulfate ion levels in
the concentrate stream are used for this calculation.  The ability to
operate at saturation levels above 100 percent is affected by the combination
of short residence time, low pH and other ionic strengths, and temperature.

Experience with acid mine water has yielded the following values for CaSO,
in the recovery range where precipitation was first detected:

Site 1, Norton, West Virginia
Site 2, Morgantown, West Virginia
Site 3, Ebensburg, Pennsylvania
Percent
Recovery
90
50
85
Percent
Saturation
356
480
298
Thus, acid mine drainage experience to date indicates that a limit on
recovery parameter is CaSO, concentration in the range of 300 to 400 percent
of saturation when calculated as above.  A feedwater can then be analyzed
for calcium and sulfate, and a maximum recovery value can be calculated.
                                     69

-------
                                 SECTION XIV
                   LABORATORY NEUTRALIZATION EXPERIMENTS

A series of laboratory experiments, both at Gulf Environmental Systems
and at the Norton test site, were conducted to determine which alkalies
would produce a sludge with the best settling and dewatering qualities
and a supernatant with the best properties for further concentration by
reverse osmosis.

Alkalies of practical interest and the amounts and costs required to
neutralize 100 Ib of sulfuric acid are listed below:
Material
CaO
MgO
NH3
NaOH
Weight
(Ib)
56
40
34
106
80
Cost
($)
0.56
1.20
1.60
2.12
2.40
Other factors that must be considered in the choice of alkalies are the
density and settling rate of the sludge, clarity of supernatant, and
concentration of sparingly soluble compounds.  The actual amounts of
alkali required would probably differ from those listed above because
of the desirability of neutralizing to the lowest feasible pH rather
than to neutrality or beyond.

Bench neutralization experiments were carried out using a typical
synthetic mine drainage feed (Ref. 7) whose analysis is quite similar
to the brine obtained from the primary 10,000-gpd unit at Norton.  This
synthetic acid mine water has the following composition and characteristics:
                                    71

-------
          Magnesium  (Mg)             116 mg/1
          Calcium  (Ca)               385 mg/1
          Iron  (Fe"1"1"1")               500 mg/1
          Aluminum (Al)              150 mg/1
                        I |
          Manganese  (Mn  )           50 mg/1
          Sulfate  (S04)              3,615 mg/1
          Conductivity               3,800 ymhos/cm at 25°C
          pH                         2.5

Ferric sulfate was used instead of ferrous sulfate since at Norton,
West Virginia, almost all of the iron in the feedwater is present as
the ferric ion.

In addition to the factors mentioned above, particular attention was given
to the iron concentration in the neutralized and clarified mine water.
Previous experience indicated that even a fraction of a milligram per
liter of particulate ferric hydrate caused significant flux declines
in a reverse osmosis unit because it coated the membrane surface.  For-
tunately, such deposits can be removed by fairly simple and inexpensive
cleaning procedures.

If the iron content of the supernatant at pH 4 or 5 is only a few mg/1
and consists primarily of ferrous iron, it may be possible to further
process the water without additional problems.  If the iron is largely present
as ferric iron, further processing is possible, such as dropping the pH
by using raw acid mine drainage or other acid, or by tfye use of reducing
agents or iron complexing agents.

The laboratory neutralization studies of the acid mine water thus centered
around, neutralizing only to pH 4 or 5.  The principal problem remaining
was the calcium sulfate concentration, regardless of the method used,
the supernatant was saturated or near-saturated with calcium sulfate
and further concentration could result in module plugging.  Table XXIX
shows the results obtained with the previously listed neutralizing
                                    72

-------
                      TABLE XXIX
   RESULTS OF ACID MINE WATER NEUTRALIZATION STUDIES
Neutralizing
Agentl
CaO

CaC03

MgO

NH^OH

NaOH

PH
4.0
5.4
4.0
5.1
4.0
5.2
3.8
4.9
3.9
5.0
Amount of
Agent Used
(ml)
90
145
140
260
60
110
90
147
100
152
Amount of
Fe in
Supernatant
(mg/1)
9.5
0
>15
0.95
>15
5.1
>15
0
2
0
Amount of
Floe After 30 Min
(ml/1)
190
535
40
70
30
48
200
600
180
720
Solutions or suspensions;  10,000 mg/1 used,
                         73

-------
 agents.   Some  tests were  run with  soda  ash  as well, but  they  showed no
 particular advantage  in cost or improvement in  results  over those  listed.
 Both  calcium carbonate  and magnesium  oxide  produce very  dense sludges by
 comparison with  the other agents.  The  calcium  carbonate used was  a very fine
 powder equivalent  to  chalk or whiting,  and  ordinary limestone would not serve
 as well unless some abrading mechanism  was  available during neutralization.
 The cost  of powdered  calcium carbonate  is roughly the same  as the  cost of
 lime  but  much more calcium carbonate  is required to neutralize  the same water.

 Magnesium oxide  appeared  to be  the most promising agent  with  respect to
 cost, sludge density, and reduction of  iron concentration (in the  super-
 natant) to below 5 ppm  at a final pH  of 5.2.  At this pH, and at the
 prevailing low iron concentration,  it was  thought that  the calcium salt
 precipitation inhibitors would  operate normally.  The iron  concentration
 was low enough to make  the use  of chelating or reducing  agents feasible
 (from a cost point of view) to  prevent  iron hydrate precipitation when
 the supernatant was further concentrated.

 Since ammonia has been  reported to be only moderately rejected at  about
 80 to 90 percent by reverse osmosis,  it was desirable to determine to what
 extent ammonia would be rejected when present essentially as  an ammonium
 sulfate solution at a pH of about 5.  By using an ammonium sulfate solution
 containing 500 mg/1 as  ammonia  at pH  levels between 5 and 6,  product water
 concentrations of 2 mg/1  or rejections  of 99.5 percent were attained.  This
                          2
 work  was  done with 3.5-ft  spiral-wound modules of the same construction as
                  2
 the standard 50-ft  modules.

Based on the previous information, bench studies were then carried out on
the actual acid mine drainage at the Norton test site.   A large sample of
brine was taken from the 10,000-gpd unit and chemically analyzed.  At the
time of sampling, the unit was  operating at 600 psi and  72.4 percent recovery.
Results of the chemical analyses were as follows:
                                    74

-------
pH	2.3
Conductivity. .  . 4000 ymhos
Hot acidity  . .  . 1865 mg/1 as CaC03
Calcium	510 mg/1 as Ca
Magnesium .... 217 mg/1 as Mg
Total hardness. . . 2142 mg/1 as CaCO.,
Cold acidity.
.  2728 mg/1 as CaCO,
Aluminum	150 mg/1
Iron	470 mg/1
Sulfate	3013 mg/1
 In the  previous  studies using  the synthetic acid  mine water brine,
 magnesium oxide  appeared to  be the most  favorable neutralizing agent
 in terms  of  cost,  settling time and density of  sludge, and  amount of
 iron removed from  the solution.  It was  decided to further  examine
 magnesium oxide  by testing it  on the sample of  reverse osmosis brine
 described above.   A one-liter  sample of  brine was stepwise  neutralized
 using a 10,000 mg/liter suspension of magnesium oxide.  An  analysis of
 the supernatant  from this neutralization gave the following results:
pH	5.8
Conductivity. .   . 3000 ymhos
Hot acidity  . .   .4.0 mg/1 as CaC03
Calcium	375 mg/1 as Ca
Magnesium .... 547 mg/1 as Mg
Total hardness.
  3126  mg/1 as CaCO,
Aluminum	0.9 mg/1
Iron. .	1.4 mg/1
Sulfate1	1325 (?) mg/1
      Value  in  question  in  analysis  does  not  balance.

 The  amount of iron  remaining  in  the  supernatant  and  the  amount  of  floe
 formed  using the  synthetic  and actual  acid mine  water  were  as follows:


Water
Synthetic
10,000-gpd
unit brine


pH
5.2
5.6

Amount of
Agent Used
(ml)
110
105
Amount of
Iron in
Supernatant
(mg/1)
5.1
1.4

Amount of Floe after
30 Min. Settling
/(tal/1)
48
30
  The fact  that  the  calcium and sulfate  ions decreased  after neutralization
  indicated  that precipitation of  calcium  sulfate did take place.   This,
  in turn,  indicated that  the supernatant  was  saturated in calcium  sulfate.
                                     75

-------
Further concentration by reverse osmosis would be difficult using this
supernatant.  While the use of MgO results in considerably less sludge that
must ultimately be disposed of, the cost advantage of lime and the better-
quality effluent obtained using lime may outweigh the less-sludge advantage
of MgO.  Therefore, additional tests were conducted using brine from the
10,000-gpd unit to determine whether MgO or hydrated lime should be used for
neutralization.  The results of these tests are shown in Table XXX.  The
amount of sulfate remaining after neutralization with MgO cannot be explained
readily.

Based on the work described above, it was decided to use lime for the actual
                                                                     i
neutralization at the Norton site.  The brine from the 10,000-gpd unit
was neutralized, settled, and filtered before being fed to the 4,000-gpd
unit.  After two short-term runs, the new process of reintroducing the
neutralized supernatant back to the feed of the 10,000-gpd unit was begun.
All further neutralization was then done using lime.
                                    76

-------
                                           TABLE XXX
       RESULTS OF  BENCH NEUTRALIZATION TESTS  ON 10,000-GPD-UNIT BRINE AT 91 PERCENT RECOVERY]
Agent
—
MgO
MgO
MgO
Ca(OH)2
Ca(OH)2
Ca(OH)2
Amount
(8/1)
—
1.869
2.000
2.280
3.000
3.575
3.614
Initial
pH
2.1
5.0
5.5
6.0
5.0
5.5
6.0
Final
pH
—
4.3
4.5
6.7
4.8
5.8
7.2
Mg
(mg/1)
444
1383
1608
1899
444
414
414
Ca
(mg/1)
980
162
165
142
682
710
730
Acidity,
pH 7.3
5650
267
185
7
102
14
0
S04
(mg/1)
6416
7917
7371
7098
2320
2184
2084
Fe
(mg/1)
1560
1.6
1.3
0.8
2.0
1.0
0.8
Al
(mg/1)
310
68
48
2.1
16
2.8
0.4
Sludge
(ml/1)
__
40
60
60
125
130
185
      These analyses were made on-site by EPA personnel.

Visual Observations

     1.  Settling times were about equal.

     2.  After 24 hr, the effluent from the lime neutralization was clearer and contained less
         suspended solids.

     3.  The lime sludge appeared denser.

-------
                               SECTION XV
                             ACKNOWLEDGMENTS

The experiments described in this report could not have been accomplished
without the aid of EPA personnel at Norton, West Virginia.  Project
Officer Robert Scott, .Roger Wilmoth, James Kennedy, and Alvin Irons are
extended special thanks for their cooperation in providing the study sites
and facilities, operation and maintenance of equipment, and chemical
analyses.  Their encouragement, suggestions, and assistance are greatly
appreciated.

Gulf Environmental Systems also wishes to thank the Christopher Coal
Company and the Bethlehem Mines Corporation for their cooperation in
allowing the use of their facilities.

The work was performed by James H. Sleigh and S. S. Kremen; counsel and
assistance were given by I. Nusbaum, A. Riedinger, C. Mungle, and R. Truby
of Gulf Environmental Systems.
                                    79

-------
                                SECTION XVI
                                REFERENCES

1.  Hill, Ronald D., Wilmoth, Roger C., and Scott, Robert B.,' Norton
    Mine Drainage Field Site, EPA, Norton, West Virginia, "Neutrolosis Treat-
    ment of Acid Mine Drainage," paper presented at 26th Annual Purdue
    Industrial Waste Conference, Lafayette, Indiana, May 1971.
2.  Mason, Donald G., "Treatment of Acid Mine Drainage by Reverse Osmosis,"
    Third Symposium on Coal Mine Drainage Research, Pittsburgh, Pennsylvania,
    May 19, 1970.
3.  Rusnak, A., and Nusbaum, I., "Reverse Osmosis Field Testing on Acid
    Mine Waters at Norton, West Virginia, Final Report," Office of Saline
    Water Report GA-8796, Gulf General Atomic Incorporated (August 6, 1968).
4.  Riedinger, A. B., "Reverse Osmosis Field Testing on Acid Mine Waters
    at Norton, West Virginia, Final Report," Office of Saline Water Report
    GA-9181, Gulf General Atomic Incorporated (January 17, 1969).
5.  Kremen, S. S., Riedinger, A. B., Sleigh, J. H., and Truby, R. L.,
    "Reverse Osmosis Field Testing on Acid Mine Waters at Norton, West
    Virginia, Final Report," Office of Saline Water Report GA-9921, Gulf
    General Atomic Incorporated (February 6, 1970).
6.  Wilmoth, R. C., and Scott, R. B., "Neutralization of High Ferric Iron
    Acid Mine Drainage," Third Symposium on Coal Mine Drainage Research,
    Pittsburgh, Pennsylvania, May 19, 1970.
7.  Pollio, Frank, and Kuniri, Robert, "Ion Exchange Processes for the
    Reclamation of Acid Mine Drainage Waters," Environmental Science and
    Technology, 1, pp. 235-241 (March 1967).
8.  Riedinger, A., and Schultz, J., "Acid Mine Water Reverse Osmosis Test
    at Kittanning, Pennsylvania," Office of Saline Water Report OSW-No.-217,
    General Dynamics, General Atomic Division (March 31, 1966).
9.  Wilmoth, R. C., and Scott, R.  B., Mine Drainage Field Site, EPA,  Norton,
    West Virginia, "Reverse Osmosis Treatment of Concentrated Ferrous Iron
    Acid Mine Drainage," EPA Inhouse Report,  August 1970.
                                    81

-------
10.   Wilmoth, R.  C., and Scott, R. B.,  Mine Drainage Field Site, EPA,
     Norton, West Virginia, "Reverse Osmosis Treatment of Ferrous Iron
     Acid Mine Drainage," EPA Inhouse Report, January 1971.
                                    82

-------
 BIBLIOGRAPHIC!   Gulf  Environmental  Systems
 Company.
 Acid Mine Waste  Treatment Using Reverse Osmosis,
 Final Report FWQA Contract 14-12-525
     The basic objectives of  this test program were
 to demonstrate the applicability of reverse osmosis
 to the demoralization of acid mine drainages, both
 high-ferrous and high-ferric  types, and to reclaim
 the maximum percentage of such feedwater in purified
 form suitable for domestic or industrial purposes,
 or as stream discharge.  These goals included the
 attainment of maximum water recovery while maintain-
 ing the required product water quality and the deter-
 mination of pretreatment requirements necessary to
 maximize water recovery and membrane life.
     Two reverse osmosis test units were operated
 during the course of  these tests:  a nominal
 10,000-gpd unit  equipped with eighteen 50-ft2
 modules and a nominal 4,000-gpd un't equipped with
 nine 50-ft2 modules.  The modules used in these
 units consisted  of both high-selectivity and high-
 flux cellulose acetate membranes.

 BIBLIOGRAPHIC:   Gulf  Environmental Systems
 Company.
 Acid Mine Waste Treatment Using Reverse Osmosis,
 Final Report FWQA Contract 14-12-525
     The basic objectives of  this test program were
 to demonstrate the applicability of reverse osmosis
 to the deminerallzatlon of acid mine drainages, both
 high-ferrous and high-ferric  types, and to reclaim
 the maximum percentage of such feedwater in purified
 form suitable for domestic or Industrial purposes,
 or as stream discharge.  These goals included the
 attainment of maximum water recovery while maintain-
 ing the required product water quality and the deter-
 mination of pretreatraent requirements necessary to
 maximize water recovery and membrane life.

     Two reverse osmosis test units were operated
 during the course of these tests:  a nominal
 10,000-gpd unit equipped with eighteen 50-ft2
 modules and a nominal 4,000-gpd unit equipped with
 nine 50-ft2 modules.  The modules used in these
 units consisted of both high-selectivity and high-
 flux cellulose acetate membranes.

 BIBLIOGRAPHIC:  Gulf Environmental Systems
 Company.
Acid Mine Haste Treatment Using Reverse Osmosis,
 Final Report FWQA Contract 14-12-525
     The basic objectives of this test program were
 to demonstrate the applicability of reverse osmosis
 to the demlneralization of acid mine drainages, both
high-ferrous and high-ferric types, and to reclaim
 the maximum percentage of such feedwater in purified
 form suitable for domestic or industrial purposes,
or as stream discharge.  These'goals included the
attainment of maximum water recovery while maintain-
 ing the required product water quality and the deter-
mination of pretreatment requirements necessary to
maximize water recovery and membrane life.
     Two  reverse osmosis test units were operated
during the course of- these tests:  a nominal
10,000-gpd unit equipped with eighteen 50-ft4
modules and a nominal 4,000-gpd unit equipped with
nine 50-ft2 modules.  The modules used In these
units consisted of both high-selectivity and high-
flux cellulose acetate membranes.
ACCESSION NO.

KEY WORDS:

Reverse Osmosis
Acid Mine Drainage
Deminerallzatlon
Calctum-Sulfate
  Solubility
Ferrous-Ferric Iron
  Ratios
Water Recovery
Brine Treatment
ACCESSION NO.

KEY WORDS:

Reverse Osmosis
Acid Mine Drainage
Deminerallzation
Calcium-Sulfate
  Solubility
Ferrous-Ferric Iron
  Ratios
Water Recovery
Brine Treatment
ACCESSION NO.

KEY WORDS:

Reverse Osmosis
Acid Mine Drainage
Demineralizatlon
Calcium-Sulfate
  Solubility
Ferrous-Ferric Iron
  Ratios
Water Recovery
Brine Treatment
                      83

-------
      The test program was carried  out  at  three  different
 mine drainage sites.   The mine  drainage water at  the  first
 site,  Norton, West  Virginia,  contained greater  than 98
 percent of  the iron present  in  the ferric  form; at the
 other  two sites,  Morgantown,  West  Virginia,  and Ebensburg,
 Pennsylvania, the drainage water contained predominantly
 ferrous Iron.  Discharges at  the second site were so
 concentrated  that recoveries  were  limited  to 50 percent;
 recoveries  of 80  to 90 percent  were attained at the
 first  and third sites.   No iron fouling was  encountered
 at  any of the three sites.   Specific salt  rejections
 were >97 percent  at all  sites.

     This report  was  submitted  in  fulfillment of  Contract
 No.  14-12-525 between the Environmental Protection
 Agency and  Gulf Environmental Systems  Company.
     The  test program was  carried out at  three different
mine drainage sites.  The  mine drainage water at  the  first
site, Norton, West Virginia, contained greater than 98
percent of the iron present in the ferric form; at the
other two sites, Morgantown, West Virginia, and Ebensburg,
Pennsylvania, the drainage water contained predominantly
ferrous iron.  Discharges  at the second site were so
concentrated that recoveries were limited to 50 percent;
recoveries of 80 to 90 percent were attained at the
first and third sites.  No iron fouling was encountered
at any of the three sites.  Specific salt rejections
were >97 percent at all sites.

     This report was submitted in fulfillment of Contract
No. 14-12-525 between the  Environmental Protection
Agency and Gulf Environmental Systems Company.
     The test program was carried out at three different
mine drainage sites.  The mine drainage water at the first
site, Norton, West Virginia, contained greater than 98
percent of the iron present in the ferric form; at the
other two 'sites, Morgantown, West Virginia, and Ebensburg,
Pennsylvania, the drainage water contained predominantly
ferrous iron.  Discharges at the second site were so
concentrated that recoveries were limited to 50 percent;
recoveries of 80 to 90 percent were attained at the
first and third sites.  No iron fouling was encountered
at any of the three sites.  Specific salt rejections
were >97 percent at all sites.

     This report was submitted in fulfillment of Contract
No. -14-12-525 between the Environmental Protection
Agency and Gulf Environmental Systems Company.
                                                    84

-------
     Accession Number
                            Subject field & Group
                                 05D
                                           SELECTED WATER  RESOURCES ABSTRACTS
                                                  INPUT TRANSACTION FORM
     Organization

     Environmental Protection Agency, Water Quality Office
     Title
     Acid Mine Waste Treatment Using Reverse Osmosis
10

Authors)
James H. Sleigh
S. S. Kremen
16

21
Project Designation
Program 14010 DYG
Note •
 22
Citation
     Water Pollution Control Research Series 14010 DYG 11/70,
     Environmental Protection Agency, Water Quality Office,
    Washlnafi
             n  r
 23 I
Descriptors (Starred First)

Reverse  osmosis,  acid mine drainage, demineralization, calcium-sulfate solubility,
ferrous-ferric  iron  ratios, water recovery, brine treatment
 25 I Identifiers (Starred first)

    West Virginia,  Pennsylvania, water recovery, brine treatment
 27
Abstract
    The basic  objectives of this test program were to demonstrate the applicability of
Reverse  osmosis  to the demineralization of acid mine drainages, both high-ferrous and
high-ferric  types,  and to reclaim the maximum percentage of such feedwater in purified
form suitable  for domestic or industrial purposes, or as stream discharge.  These goals
included  the attainment of maximum water recovery while maintaining the required product
*ater quality  and the determination of pretreatment requirements necessary to maximize
water recovery and  membrane life.
    Two reverse  osmosis test units were operated during the course of these tests:  a
n°minal 10,000-gpd  unit equipped with eighteen 50-ft2 modules and a nominal 4,OOU-gpd
jwit equipped  with  nine 50-ft* modules.  The modules used in these units consisted of
b°th high-selectivity and high-flux cellulose acetate membranes.
    The test program  was carried out at three different mine drainage sites.  The mine
^ainage water at the first site, Norton, West Virginia, contained greater than 98 per-
Jent of the  lron present in the ferric form; at the other two sites, Morgantown, West
yirginia, and  Ebensburg, Pennsylvania, the drainage water contained predominantly ferrous
*r°n.  Discharges at  the second site were so concentrated that recoveries ««e limltjd
-J° 50 percent; recoveries of 80 to 90 percent were attained at the first and third sites.
1° iron fouling was encountered at any of the three sites.  Specific salt rejections were
percent at all sites.
    J.
            leigh
                          institution  Gulf  Environmental Systems Company
      |REV'
               198B)
                                              SEND TO:
                                                      WASHINGTON, D. C. 20240
                                                                              * SCO! 18«g-35S-3Se

-------