WATER POLLUTION CONTROL RESEARCH SERIES • 17010 EKI 09/71
Phosphate Precipitation
   With Ferrous Iron
                            /
                            /
              r. . /
L^^Jf.
 U.S. ENVIRONMENTAL PROTECTION AGENCY

-------
          WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes the
results and progress in the control and abatement of pollution
in our nation's waters.  They provide a central source of
information on the research,  development and demonstration
activities in the Environmental Protection Agency,  through
inhouse research arid grants and contracts with Federal,  State,
and local agencies, research institutions, and industrial
organizations.

Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Chief, Publications Branch
(Water), Research Information Division, R&M,  Environmental
Protection Agency, Washington, B.C. 20U60.

-------
         PHOSPHATE PRECIPITATION WITH  FERROUS IRON
                              by
                      Masood Ghassemi
                      Howard L. Recht
                    Atomics  International
    A Division of North  American Rockwell  Corporation
                Canoga Park,  California  91304
                            for the

             Office of Research and Monitoring

              ENVIRONMENTAL PROTECTION AGENCY
                     Project  #17010 EKI
                     Contract #14-12-817
                       September 1971
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price 70

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

-------
                          ABSTRACT
Phosphate removal by ferrous iron addition was studied in batch precipi-
tation tests conducted on pure solutions of ortho-, pyro- and tripolyphos-
phate,  orthophosphate solutions containing bicarbonate ion,  and second-
ary effluent.  The effects of pH and reactant concentration on the effi-
ciency of reactant removal were evaluated.

In the absence of dissolved oxygen and for initial conditions  of 1Z mg/1  P
and a reactant equivalence  ratio of  1.0, orthophosphate removal was
maximum (97%) at a pH  of 8.0.  The precipitate formed was identified
as vivianite, Fe3(PO4)2-8HZO. At this pH and at lower pH levels,  Fe(II)
removal nearly paralleled that of orthophosphate removal.   The time for
maximum orthophosphate removal increased with decreasing pH.  The
reaction speed was independent of Fe(II) concentration but decreased at
lower orthophosphate levels.  The system behavior was the  same in
secondary effluent as in pure solutions.  Pyro- and tripolyphosphates
were less efficiently precipitated than orthophosphate.

Dissolved oxygen increased orthophosphate  removal efficiency.  How-
ever, the precipitates obtained in the treatment of oxygen-containing
secondary effluent were usually very  fine and did not settle well.

The  data on Fe(II)-phosphate precipitation are compared with those of
Fe(III)- and Al(III)-phosphate systems.

This report was submitted  in fulfillment of Project No.  17010 EKI,
Contract  14-12-817, under sponsorship of the Environmental  Protection
Agency.
                               111

-------
                       CONTENTS




Section                                                 Page




 I        Conclusions                                     1




 II       Recommendations                               5-




 III       Introduction                                     j




 IV       Experimental                                   9



                                                         PI
 V       Results and Discussion                          ^




 VI       Acknowledgments                              6l




 VII      References                                     63
                            v

-------
                        FIGURES

No.                                                     Page


 1      Schematic Diagram of the Controlled
       Atmosphere Reaction System                       12

 2      Controlled Atmosphere Reaction System            "^

 3      System for Settling Rate Studies                    -^

 4      Sample Removal  from Reaction Vessel              17

 5      Precipitation of Orthophosphate with Fe(II),
       Al(III) and Fe(III) at a  Cation-to-Orthophosphate
       Equivalence Ratio  of  1.0                          pq

 6      Solubility Diagram for Ferrous Phosphate          ^5
                            VI

-------
                        TABLES

No.                                                        Page

 1     Partial Analyses of the Secondary Effluent Samples
      Used in Various  Precipitation Experiments             10

 2     Summary of Experiments Discussed in Text           "

 3     Changes in Solution Turbidity and pH in Orthophos-
      phate Precipitation with Fe(II)                         23

 4     Changes in Solution Turbidity and pH in Orthophos-
      phate Precipitation with Fe(II)                         24

 5     Effect of pH on Orthophosphate Precipitation with
      Fe(II)                                                  26

 6     Orthophosphate Precipitation with Fe(II) at an
      Fe(II)-to-Orthophosphate Equivalence Ratio of 1. 5     32

 7     Effect of Orthophosphate Concentration on Phos-
      phate Precipitation with Fe(II)                         -j.

 8     Effect of Bicarbonate Alkalinity on the Precipitation
      of Orthophosphate with Fe(II)                          35

 9     Orthophosphate Precipitation from Secondary
      Effluent with Fe(II) in the Absence of Dissolved
      Oxygen                                                ^

 10    Precipitation of Pyro- and Tripolyphosphate with
      Fe(II) in the Absence of Dissolved Oxygen             39

 11    Equilibrium Relationships for Orthophosphate-Fe(II)
      System                                                .

 12    Orthophosphate Precipitation with Fe(II) in the
      Presence of Dissolved Oxygen                         47

 13    Orthophosphate Precipitation with Fe(II) from
      Secondary Effluent in the Presence of Dissolved
      Oxygen                                                149
                           VII

-------
                 TABLES (Continued)
No.
14    Formation and Settling Characteristics of Precipi-
      tates in Orthophosphate Precipitation from Secondary
      Effluent with Fe(II) in the Presence of Dissolved
      Oxygen                                                52

15    Orthophosphate Precipitation from Secondary
      Effluent with Fe(II) in the Presence of Dissolved
      Oxygen -  Analysis of the Filtrates                      ^

16    Settling Characteristics of the Precipitates in Ortho-
      phosphate Precipitation from Secondary Effluent
      with Al(III)                                             56

17    Settling Characteristics of the Precipitates in Ortho-
      phosphate Precipitation from Secondary Effluent
      with Fe(III)                                            t-7
                           Vlll

-------
                          SECTION I

                        CONCLUSIONS
Atomics International has conducted a parametric investigation of phos-
phate precipitation using ferrous iron in which the  effects of pH,
reactant concentration and diverse ions on the rate and efficiency of
reactant removal were evaluated. Pure solutions of orthophosphate at
concentrations representative of those in wastewater, orthophosphate
solutions containing bicarbonate alkalinity,  pyro- and tripolyphosphate
solutions and secondary effluent samples from an activated sludge
treatment plant were used in this investigation.  To establish back-
ground data on the  aqueous chemistry of phosphate-Fe(II)  reactions,
the majority of the experiments were conducted in  the absence of dis-
solved oxygen (to prevent oxidation of ferrous iron to ferric iron) in
a special controlled atmosphere reaction system, which also allowed
for continuous monitoring of solution pH and periodic removal of
sample aliquots for analysis of turbidity and determination of residual
reactant concentrations.  Only a limited number of tests were per-
formed on phosphate solutions containing dissolved oxygen.  The pre-
cipitates obtained in various experiments were observed for color
characteristics and some were also  examined by x-ray diffraction for
chemical structure.  Using  a special sampling apparatus, the settling
properties of the precipitates formed in the treatment of secondary
effluent with Fe(II)  in the presence of dissolved oxygen were evalu-
ated and the  results were compared  with similar data obtained using
Al(III) and Fe(III) salts.

The data collected  on orthophosphate precipitation  with Fe(II) in the
absence of dissolved oxygen indicated the following:

(a)  The  efficiency  of orthophosphate removal with Fe(II) is strongly
pH  dependent with the maximum removal obtained in the vicinity  of
pH  8.  For a  reaction time of 5 hr,  an initial orthophosphate concen-
tration of 12 Tcng/l P-, and an Fe(II)-to-orthophosphate equivalence
ratio of 1, the maximum levels of reactant removal were 7, 35,  94,
69,  and 18% for orthophosphate and  9, 39,  97,  ~ 100, and- 100% for
ferrous iron in experiments conducted at initial pH levels of 6,  7, 8,
9, and 10, respectively.

(b)  The  rate of reactant  removal is  generally more rapid at higher
pH  levels. For an orthophosphate concentration of 12 mg/ji P and a
reactant equivalence ratio of 1,  there was no increase in the extent

-------
of reactant removal after 5 mirt at pH1 s of 9 and 10, whereas at pH 8
a reaction time of about 2 hr was required to achieve maximum react-
ant removal.

(c)  The speed of reactant removal appears to be unaffected by  changes
in the Fe(II) concentration but is affected by changes in pH and  ortho-
phosphate concentration. For an orthophosphate concentration of
12 mg/l P and an initial pH of 8,  the patterns of change in pH,  tur-
bidity,  residual reactant concentration,  and precipitate  color were
nearly the same for Fe(II)-to-orthophosphate equivalence  ratios of
1.5, 1.0, and 0.33.  For an Fe(II) concentration of 3. 87 x 10~4 eq/4
and at the optimum phosphate  removal pH, the reaction  was more
rapid at an orthophosphate  level of 12 mg/£ P than at 4 mg/£ P.

(d)  At the optimum precipitation pH and at lower pH levels, not all of
the excess ferrous iron added to the phosphate solution precipitates
with the phosphate or as hydroxide.

(e)  In the vicinity of the optimum  precipitation pH,  bicarbonate
alkalinity at a concentration of 420 mg/.£ NaHCO  did not interfere
with the precipitation of orthophosphate at 12 mg/1 P when a reactant
equivalence ratio of 1. 0 was used.  The  orthophosphate-Fe(II) reaction
in a wastewater (secondary effluent) environment was similar to that
for pure orthophosphate solution and for orthophosphate solution con-
taining  bicarbonate alkalinity.

(f)  On an equivalence basis, pyro- and tripolyphosphates  are less
effectively precipitated with ferrous iron than orthophosphate.  For a
phosphate concentration of  1. 16 x  10   eq/^ , an initial pH of 8. 0 and
a reactant equivalence  ratio of 1,  maximum phosphate and iron re-
movals were, respectively, 94 and 97% for orthophosphate, 87 and
91% for pyrophosphate  and  11  and  12% for tripolyphosphate.

(g)  In all experiments  any  precipitates retained on the membrane
filters were initially light to dark  green  in color.  On exposure to air
all precipitates changed color, turning yellow or blue depending on pH
and reaction time.  The blue precipitate obtained in one orthophosphate
precipitation at the optimum pH was examined by x-ray  diffraction and
identified as vivianite,  Fe  (PO ) • 8H O.
                         J   ^ L*    C*

(h)  Comparison of the  orthophosphate precipitation data for Fe(II) with
the published data for Fe(IH) and Al(III) indicates that the  pH of opti-
mum orthophosphate removal  is higher for Fe(II) than for Fe(III)  or
Al(III) (~ 8. 0 vs. -3.5 for  Fe(III)  and ~  5 for Al(III) ) and that at the
respective pH of optimum orthophosphate removal,  the  equivalence

-------
capacity of Fe(II) for phosphate removal is higher than those for Fe(III)
and Al(HI).

(i)  The pH-dependence of orthophosphate removal efficiency for Fe(II),
Fe(III) and Al(III) and the higher equivalence capacity of Fe(II) for
orthophosphate removal are in reasonable accord with  a phosphate
precipitation model in which hydroxide and orthophosphate ions com-
pete for  reaction with the  cation.

In wastewater treatment with Fe(II) in the presence of dissolved oxygen,
the phosphate precipitation system becomes  complicated by the genera-
tion and  subsequent hydrolysis of Fe(III) and in this regard the system
differs from phosphate precipitation with pure Fe(III) and with Fe(II)
in the absence of dissolved oxygen.  For a given quantity (weight)  of
iron used,  phosphate removal efficiency is higher for Fe(II) than for
Fe(III) from a stock source when  Fe(II)  is completely oxidized in situ
resulting in homogeneous  generation of  Fe(III).

In general, the precipitates formed in the treatment of secondary  efflu-
ent with  Fe(II) in the presence of  dissolved oxygen were very fine  and
did not settle well.  The settling properties of the precipitates formed
with Fe(III) and Al(III) were superior to those obtained  with Fe(II).

A program is recommended to investigate a  process for the removal
of phosphates from wastewater using ferrous iron under oxidizing con-
ditions with the aim of process optimization.   In this process, ferrous
salt,  such  as waste pickle liquor  from steel  mills (an inexpensive
source of Fe(II) in some large metropolitan areas) is added to the raw
or settled sewage with the result  that significant phosphate removal is
effected.  Subsequent aeration of  the wastewater in an activated sludge
tank should result in the oxidation of residual ferrous iron,  further
precipitation of phosphate, partial removal of organics and improve-
ment in the sludge characteristics.  Such a process promises to be
more economical than other precipitation techniques.

-------
                           SECTION II

                     RECOMMENDATIONS
The basic objective of the study described in this report has been to
obtain some fundamental data which heretofore had been lacking on
the orthophosphate-Fe(II) reactions.  Most of the precipitation studies
were  conducted on pure phosphate solutions and in the absence of
dissolved oxygen.  It should be pointed out that the study of cation-
phosphate reaction in pure  systems is an essential prerequisite to
understanding the phosphate removal  process from such complex and
variable systems as domestic  wastewater.  With the  results of this
study as the background it is now possible to extend the work to an
evaluation of the parameters -which influence process efficiency with
the specific objective of developing design criteria for  wastewater
treatment with Fe(II).  The following  areas of research are recom-
mended as a logical extension  of the present  study.  Although pure
solutions of orthophosphate and secondary effluent may be used in a
limited number of tests,  the majority of the experiments should be
conducted on raw and settled sewage under conditions similar to those
encountered in large-scale applications.

(a)  Evaluation of the effects of pH, Fe(II) dosage, dissolved oxygen
level,  and concentration of organic material (e.g., COD or TOC) on
the rate of oxidation of Fe(II) and on the  efficiency of phosphate and
organic removal.  (The chemical treatment of wastewaters for the
purpose of phosphate removal  invariably results in some removal of
dissolved and colloidal organic matter.   The degree of organic
removal associated with phosphate precipitation is an essential con-
sideration in evaluating the process economics.)

(b)  Determination of the nature and extent of particle charge and
their  relations to such precipitate characteristics as  settleability and
filterability as a function of the wastewater parameters mentioned in
(a)  above.  These precipitate properties are of major practical con-
cern in the design and operation of solids-liquid separation units.
Any improvements in these properties without the addition of poly-
electrolytes results in a  saving in the operating cost.   In evaluating
the effect of pH on particle charge and precipitate characteristics,
both sodium hydroxide and lime should be used for pH adjustment and
the results compared.

(c)  Evaluation of the effect of  aeration or addition of  ozone or chlorine
compounds on the rate of in situ oxidation of ferrous  iron and on the
efficiency of phosphate and organic removal.  In a number of aeration

-------
tests,  conditions  should be selected to simulate those existing in the
activated sludge aeration tank and data should be collected on the
settleability of the chemical-biological floes.

(d)  Comparison of the effectiveness of ferrous salts and waste pickle
liquors for wastewater treatment using both sulfuric acid and hydro-
chloric acid pickling wastes.  Waste pickle liquor contains mineral
acidity which may lower the pH of the wastewater thus necessitating
supplementary pH adjustment by addition of base.

(e)  Evaluation of the use of Al(III),  Fe(III), and polyelectrolytes in
conjunction with Fe(II) for improving the characteristics of the pre-
cipitates formed in wastewater treatment with Fe(II). When used
from a stock source,  Al(III) and Fe(IH) hydrolyze in solution to  a con-
siderable extent.  Hydrolysis products of Al(III) and Fe(III), because
of their gelatinous nature,  may serve as flocculant aids thus improv-
ing the precipitate settleability.

(f)  Characterization of the sludge produced in the treatment of waste-
water with Fe(II).   This should include water content, compactibility,
and dewaterability.  These sludge characteristics are important
parameters in determining the costs of sludge handling and disposal.

(g)  A preliminary economic assessment of large-scale  wastewater
treatment using Fe(II).

-------
                          SECTION III

                        INTRODUCTION
Even with the advent of phosphate-free detergents, the removal of
phosphates from waste-waters is regarded as essential for the control
of eutrophication and prevention of water quality deterioration in many
receiving streams.  Of a number of methods available for phosphate
removal,  chemical precipitation (often called coagulation) using
aluminum, ferric iron, ferrous iron and lime is considered to be the
                 (12)
most economical.v '  ' The chemistry of phosphate precipitation with
aluminum, ferric iron and lime has been the subject of  several re-
search studies^ '  '  and,  accordingly, the optimum conditions for
phosphate removal with these salts are fairly well defined.

Although ferrous iron has been utilized for the removal of phosphates
in some large-scale applications,    the fundamental aqueous chem-
istry of phosphate-Fe(II)  reactions appears to be almost totally unex-
plored.  No mention of exclusion of dissolved oxygen is made in  all
the reported laboratory evaluation tests with Fe(II).  Since partial or
total oxidation of Fe(II) to Fe(III) is almost inevitable in the presence
of dissolved oxygen,  the reported data cannot be regarded as true
representation of reactions in the Fe(II)-orthophosphate system.
From the standpoint of treatment cost, the use of ferrous iron for
phosphate removal is  economically appealing  since in  certain areas
of the country waste pickle liquor from local steel industries may pro-
vide an inexpensive  source of ferrous iron.

The objective of the present study has been two-fold:  (a) to evaluate
the effects of pH and reactant concentration on the rate and efficiency
of phosphate removal  from synthetic and secondary  effluent waste-
waters and on the nature  of the precipitates formed, and  (b) to com-
pare the data on the Fe(II)-phosphate precipitation system with the
available data for Fe(III)- and Al(III)-phosphate systems.  In order to
obtain for the first time the fundamental data on the orthophosphate-
Fe(II) reaction, most  of the experiments in this study  were conducted
on pure orthophosphate solutions and in the absence of dissolved oxygen
(to prevent oxidation of ferrous iron to ferric iron).  Only a limited
number of tests were  made of the use of ferrous iron for orthophos-
phate removal from wastewaters containing dissolved  oxygen.

-------
                          SECTION IV

                       EXPERIMENTAL
This section describes the materials, apparatus,  and the general pro-
cedures used in this investigation.  Special methods and modifications
are described in the Results section,  when appropriate.

Materials

Reagent grade  sodium monohydrogen (ortho-) phosphate (Na_HPO .)
tetrasodium pyrophosphate (Na P  O • 10 HO), sodium tripolyphosphate
(Na P  O   ), sodium bicarbonate fNaHCO)",  ferrous  ammonium sulfate
[FetNH OSO  )  -6H O], ferrous perchlorate [Fe(ClO ) • 6H O], ferric
ammonium sulfate [FeNH (SO ) •  12H  O],  and aluminum sulfate
[Al (SO  )  • 18H  O]  were used to prepare the test solutions.  All Fe(II),
Fe(ni) and Al(IIl) solutions were prepared  fresh in concentrated form
immediately before use.  Doubly distilled water was  used in the prepara-
tion of all test  and reagent solutions.  The doubly distilled water used in
experiments with Fe(II) was degasified by boiling.  It was then cooled
and purged with either  argon or nitrogen gas prior to use.

The orthophosphate solution used in most experiments contained
12 mg/£  P orthophosphate (1. 16 x 10"3  eq/4  PO~3 or 3. 87 x 10"4 M).
This concentration was selected as representative of that to  be
encountered in a high phosphate  secondary effluent and because the pre-
cipitation of phosphate  at this level with Al(III) and Fe(III) had been
extensively studied in a previous investigation.^  '  An 18 m.g/JL P solu-
tion of tetrasodium pyrophosphate  (1. 16 x  10"  eq/i, P O"  )  and a
21.6 mg/4 P solution of sodium tripolyphosphate (1. 16 x 10    eq/£
P  O, ,T ) were  used in the experiments with condensed phosphates.

The wastewaters (secondary effluent)  used in phosphate precipitation
studies were obtained from the Tapia Park Treatment Plant  of the Las
Virgenes Municipal Water District in  Calabasas,  California.  This
treatment plant is a small (~ 2 mgd) activated sludge plant which serves
a primarily residential community.  The plant is  currently being
expanded and during the months when  samples were collected it was  not
always operating at a high efficiency,  as evidenced by the high concen-
trations of phosphate and turbidity in the plant effluent (see Table 1).
Partial analyses of the wastewater samples used in this study are given
in Table  1.  With the exception of Wastewater #2, which was  not
filtered,  all samples were filtered through Whatman #1 paper prior  to
use in phosphate precipitation  tests.

-------
                              TABLE 1

    PARTIAL ANALYSES OF THE SECONDARY EFFLUENT SAMPLES
          USED IN VARIOUS PRECIPITATION EXPERIMENTS3"
Constituents
pH
Temperature

Conductivity (at
above tempera-
ture) (mmho)
Alkalinity
(mgAe CaCOj
o
Turbidity (JTU)
Orthopho sphate
(mgAe P)
Date Collected
Wastewater
#1
7.25
22.2

1.75
330
4.3
)
15.8
10-27-70
Wastewater
#2
7.35
19.5

1. 75
557
7.5
9.1
12-18-70
Wastewater
#3
7.32
23.0

1.82
530
5.0
18.3
10-19-70
Wastewater
#4
7.51
19.5

1.80
586
15
11.2
12-14-70
Wastewater
#5
7.51
18.0

1.70
553
8
9.0
1-12-71
Data are for the wastewater as received in the laboratory and prior to
filtration through Whatman #1 paper.

Orthopho sphate determinations made on Whatman #42 filtrates.

 Apparatus

 Commercial Items.  Radiometer PHM 26 and PHM 28 pH meters were
 used for pH measurements. All conductivity determinations were made
 with a Radiometer Model CDM2e conductivity meter.  A Radiometer
 automatic titration control unit,  Type TTT 11 was used in conjunction
 with Radiometer PHM 26 pH meter for the constant pH precipitation
 experiments. A Hewlett-Packard/Moseley Model 7100 B two-pen strip
 chart recorder was used for recording pH.  All turbidity measure-
 ments were made with a Hach Laboratory Turbidimeter Model 2100.
 The colorimetric analyses were made with a Beckman DB-G Grating
                                10

-------
Spectrophotometer.   The x-ray diffraction analyses were made on a
Norelco 50 KV diffractometer.  All "jar test" precipitation experi-
ments were conducted using a Phipps and  Bird six-place stirrer.

Controlled Atmosphere Reaction System.   The ferrous iron in an
aqueous solution is subject to oxidation by the dissolved oxygen.  To
minimize the oxidation of ferrous iron during the precipitation experi-
ments, a  special system was constructed  which permitted the experi-
ments to be conducted under an oxygen-free argon atmosphere.  The
schematic diagram and a photograph of the system  are shown in
Figures 1 and 2, respectively.  The heart of the system is a special
thermostatted reaction unit fitted with a #13 rubber stopper containing
openings for the support of pH electrodes,  thermometer,  and gas inlet
tubes, and for the addition of reagents, withdrawal of the samples,  and
gas exit.  This last opening is kept plugged except during reagent addi-
tion or sample withdrawal.  The reaction  vessel which fits into a
6 in.  x 6 in. x 6-3/4  in. Plexiglas water bath is a 2- H Pyrex beaker
with its tip rounded to fit into a 1/4 in. deep  circular groove cut into
the 1/2 in. thick Plexiglas cover.  To provide a tight seal, the  groove
is partially filled with Apiezon Q Sealing Compound.  The  Plexiglas
cover plate also has a 2-1/2 in. diameter hole at its center which
accommodates the #13 rubber stopper containing the pH electrodes,
etc.  Mixing and agitation in the reaction vessel are provided by means
of a magnetic  stirrer.  The reaction vessel can be  purged through
either of the two gas  inlet tubes.  One tube extends  to within 1/2 in. of
the bottom of the beaker; the  second tube extends only to a level just
above the water surface.  The argon gas used for purging passes
through a hot copper  oxygen removal unit,  and a humidifier prior to
entrance into the reaction vessel.

To prevent air oxidation of Fe(II)  to Fe(III) during filtration, a special
vacuum filtration apparatus was constructed  and used which allowed for
operation under an inert gas atmosphere.   The apparatus consisted  of
a standard Millipore  filtration unit with the sample  holder converted to
a covered chamber with two openings:  one for gas  inlet and another
for sample introduction and gas exit.   The collection flask is a  regu-
lar vacuum flask which is connected to a vacuum source.

System for Settling Rate Studies .   Figure 3 is a photograph of the system
used for the measurement of the settling rates of the chemical floes pro-
duced in phosphate removal from wastewater using  Fe(II), Al(III), and
Fe(III) salts.  The design of this system is essentially the same as that
recommended by Cohen.     The system consists of a Phipps and  Bird
six-place stirrer (for mixing and  flocculation for a  short  period after
the addition of coagulants) and a special apparatus which permits
                               11

-------
                                       to
                                       >
                                       CO
                                       o
                                       LU
                                       tr
                                       CL
                                       CO
                                       O
                                       <
                                       Q
                                       o
                                       oc
                                       I-
                                       tr
                                       a
                                       a
                                       o
                                       CJ
                                       CO
                                       UJ
                                       EC
                                       D
                                       (D
12

-------
                                                                          s
                                                                          (U
                                                                         fl
                                                                         o
                                                                         •H
                                                                         -p
                                                                         u

                                                                         C8
                                                                         I)
                                                                         a;


                                                                         
-------

-------
automatic and periodic withdrawal of sample aliquots at any pre-
selected distance below the water surface for the purpose of turbidity
and/or chemical analysis.  The sampling containers are 125-ml
Erlenmeyer flasks,  individually connected to a. vacuum manifold
through small needle valves.   The sampling tubes are 3/32-in.  ID
stainless steel which are connected to the Erlenmeyer flasks by
1/8 in. ID  Tygon tubes.  The "on-off" control for the vacuum to the
system is a plastic "T" which is open to the atmosphere at  one  end
and is connected to the vacuum manifold and a vacuum source at the
other ends.  By closing the open end of  the "T" the Vacuum is diverted
to the manifold and the water from each jar flows into the sampling
containers.  The needle valves were adjusted so that the rates of flow
were nearly the same for all  six sampling units and that a sample of
20 to 25  ml volume could be collected in each flask when the vacuum
was  applied to the manifold for a period of 4 to 5  seconds.  In opera-
tion, following the collection  of each set of samples,  the sampling
flasks are  replaced with empty containers for subsequent sampling.

Analytical  Procedures

All phosphate and iron analyses were performed in accordance with
the procedures described in ASTM Manual on Industrial Water and
Industrial Waste Water. ^ '  The colorimetric amino reduction method
with bismuth modification (ASTM Designation D515-66T) was used for
orthophosphate determination.  The polyphosphates were analyzed by
hydrolyzing them to orthophosphate by boiling with  acid  and then
determining them as  orthophosphates.  Ferrous and total iron analyses
were carried out by the orthophenanthroline colorimetric method
(ASTM Designation D1068-62T, Referee Method A).

Experimental Procedures

Experiments Using the Controlled Atmosphere Reaction System.  The
Controlled Atmosphere Reaction System (described above) was used to
carry out the phosphate precipitation experiments in the absence of
dissolved oxygen.   The procedure in these tests was as follows.
Fifteen hundred ml of the test phosphate solution was placed in the
reaction vessel  and purged  with argon (through the  lower gas inlet
within the solution) for about  45 to 60 min.  Then,  while the phosphate
solution  was  being rapidly mixed with a magnetic stirrer,  5 to 10 ml of
an Fe(II) solution of appropriate concentration was  added from a
pipette to the solution to establish the desired Fe(II)-to-phosphate
equivalence ratio.  After the  addition of Fe(II) and during the course of
the precipitation, the argon was passed to the system either through
the lower gas inlet located in the  solution or through the gas inlet
located above the solution surface.  All experiments were conducted


                              15

-------
at 25. 0 C.  Two methods of pH adjustment were employed.  In a few
experiments,  NaOH (0.1 N)  was added to the test solution prior to the
addition of Fe(IJ).  In other tests,  a Radiometer Automatic Titrator
was  employed to control the pH during and for a short period (see
below) after the addition of Fe(II) through automatic addition of NaOH.

A detailed description of the operating principle of the Radiometer
Titrator is available elsewhere.' ' In short,  the titrator  compares the
meter reading of the pH meter with a current which represents the
desired final pH.  The difference between the two currents (termed the
"error" current) is amplified and applied to a switch circuit which
energizes an electromagnetic valve.  The valve is  used to control the
titrant (NaOH in this study) flow from a burette. The design of the
titrator provides for an  average flow to the sample which  is propor-
tional to the error current within a certain manually selected range
called the proportional band. Within the proportional band the titrant
is added to the sample in increments because the valve has only two
positions,  closed or open.   The  proportional reduction of  the average
flow is accomplished automatically by a continuous reduction of the
increments and  a continuous increase of the interval between two incre-
ments .

In the  experiments where the titrator was used for pH control,  because
of the  operating principle of the  titrator and because of the nature of
the precipitation reactions,  a final pH close to the  desired level could
not be attained by the time all the Fe(II) had been delivered to the reac-
tion  beaker (the addition of Fe(II) generally lasted  about 1 min).
Accordingly,  in the precipitation tests the pH adjustment was allowed
to continue for up to 4. 0 min after the start of Fe(II) addition.  With
this  procedure,  the final pH at the termination of pH adjustment was
usually within 0. 1 pH unit of the desired level.

For  5  hours after the addition of Fe(II), and while  the solution was  con-
tinuously being  stirred,  aliquot  samples of the reaction mixture were
removed from the reaction vessel at selected time intervals and
analyzed for turbidity or soluble residual reactant  concentrations.   The
sampling technique  is demonstrated in Figure 4.  The aliquots  removed
for chemical analysis (40 ml each) were filtered through  100 m^ Milli-
pore membrane filters.   The filtration was performed under nitrogen
or argon in the  special filtration unit  described above under "Apparatus",
(The nitrogen gas used was  not pretreated to remove the  traces of oxy-
gen  which  might have been present. )  To retard the oxidation of Fe(II)
to Fe(III) in the  filtrate, 1 ml 1  N HC1 was placed in the collection
flask.  (In the polyphosphate experiments and in the experiment with
wastewater,  1 ml concentrated HC1 was used in place of 1 ml 1  N HC1
to assure a sufficiently  low  pH. ) In all experiments the test  solution

                                16

-------
Figure 4.   Sample Removal from Reaction Vessel
                    17

-------
was visually observed for turbidity formation.  In a few experiments
turbidity was quantitatively determined on sample aliquots (25 ml)
removed specifically for this purpose.  In all experiments the solution
pH was continuously monitored during the  course of the precipitation.

The precipitates retained on the membrane filters were observed for
color characteristics before and following exposure to air.  In some
cases, the exposed precipitates (retained on the membranes) were
examined by powder x-ray diffraction for crystallinity and chemical
nature.

Precipitation Experiments with Fe(II) in the Presence of Dissolved
Oxygen .  The effect of dissolved oxygen on phosphate removal with
Fe(II) was evaluated in three  experiments  in which no measures were
taken to remove oxygen from the phosphate test  solution.  A 12 mg/f P
orthophosphate  solution containing 420 mg/jl NaHCO was used in one
of the experiments conducted at 25. 0 C in the thermostatted reaction
vessel shown in Figure  1.  An Fe(II)-to-orthophosphate equivalence
ratio of 1. 0 was used in this test.  The other two experiments were
conducted on an unfiltered wastewater (Wastewater #2,  Table 1) using
Fe(II)-to-orthophosphate equivalence ratios of 0.5 and 1.5.   These
tests were performed at room temperature,(25 4- 2°C) with mixing and
flocculation provided by a Phipps and Bird six-place stirrer.  The mix-
ing and flocculation consisted of 2 min rapid mixing at 90 rpm followed
by one hour of slow mixing at 20  rpm.  After 1 hr of quiescent settling,
aliquots were filtered through 100 m|_i Millipore membrane filters and
the filtrates analyzed for phosphate and total iron.   Other experiments
conducted with Fe(II) in the presence of dissolved oxygen are described
in the next section where the  procedures used for the settling rate
studies are described.

Settling Rate Studies .  The system for settling  rate studies (described
earlier under "Apparatus" section) was used for the evaluation of the
effect of Fe(II) dose (actually Fe(II)-to-orthophosphate equivalence
ratio) on the formation and settling characteristics of the precipitates
formed in the treatment of secondary effluent with Fe(II).  The experi-
ments were  conducted on 1.5-A samples of a filtered secondary efflu-
ent (Wastewater #3, Table 1) using Fe(II)-to-orthophosphate equiva-
lence ratios of  0.4, 0.8,  1.2,  1.5, and 2.0.  Following mixing and
flocculation (2 min rapid mixing at 90 rpm,  10 min  slow mixing at
20 rpm),  the precipitates were allowed to settle. The automatic
sampler was used to obtain 20  to 25 ml sample aliquots from a posi-
tion about 2. 5 in. below the water surface after  2 min,  5 min,  15 min,
1 hr,  and 2 hr of  quiescent settling.  The  aliquots were thoroughly
mixed and their turbidities determined immediately.  The 2-hr ali-
quots were subsequently analyzed for orthophosphate content.
                                18

-------
For comparison with the results of Fe(II) experiments, settling rate
studies were also made of phosphate precipitates obtained with Fe(III)
and Al(III).  Wastewaters #4 and #5  (Table  1) were used in the experi-
ments with Al(III) and Fe(III), respectively.  In these experiments, at
the end of 2 hr of settling, an aliquot from  each solution was filtered
through Whatman #42 paper.  The filtrates were analyzed for ortho-
phosphate content.
                                 19

-------
                          SECTION V

                 RESULTS AND DISCUSSION
Table 2 is a  summary of the conditions used for 38  representative
experiments.  The results of these experiments are presented in
Tables 3 through 17 and are discussed in detail in the text.

Precipitation Experiments with Fe(II) in the Absence of Dissolved Oxygen

Experiments with Pure Orthophosphate Solutions
(Experiments 1-13)

Effect of pH  on Turbidity Formation (Experiments 1-5).  In the initial
precipitation experiments with Fe(II),  only the solution pH and turbidity
were monitored.  Table 3 gives the results of three experiments (Experi-
ments 1-3) conducted with orthophosphate solutions initially at pH levels
of 5. 0,  7. 0 and 9. 0,  with no subsequent pH adjustments during the addi-
tion of Fe(II) or after.  As indicated by the data  in this table,  the solu-
tion pH  in these experiments dropped immediately upon the addition of
Fe(II), reaching values of 4. 87,  6. 77 and 7. 32 within 2 min,  respectively.
In Experiment 1, it may be noted that no significant amount of turbidity
was formed and no further change in pH occurred during the 6. 0 hr
observation period.  In Experiment 2,  the pH remained at  6. 77 for the
first 1 to  1-1/2 hr  and then gradually dropped,  reaching  6. 50  at the end
of 2 hr and an apparently constant value of 6.4 at the end of 2-1/2 hr.
Turbidity formation was  significant only after 1  to 1-1/2 hr.   In Experi-
ment 3, the initial  pH remained constant for a shorter period  of time
(approximately 12 min) before it gradually decreased to a final value of
6.4.  Also, the increase in turbidity over its initial value was more
rapid in this  test (30 min vs 1-1/2 hr in Exp. 2).

From the results of Experiments 2 and 3,  it became evident that a
study of the Fe(II)-orthophosphate reaction at a near constant  high pH
would require  supplementary addition of a base during and subsequent
to the addition of Fe(II).  Table 4 shows the data obtained in two experi-
ments (Experiments  4 and 5) in which the pH was maintained constant
at 8.0 and 9.0,  respectively, during the addition of Fe(II).  In Experi-
ment 4, the pH remained at 8. 0 for about 2 min  after Fe(II) addition and
discontinuation of pH control.  It then gradually  decreased,  reaching
7. 60 after 2 hr and 7.51 after  5 hr.  The turbidity content of the product
solution increased  with the drop in pH,  reaching a maximum value of
48 JTU  after 1-1/2 hr,  and remained constant thereafter.  At pH 9.0
(Experiment  5) the precipitate formation was more  rapid and the
                               21

-------
                                                       TABLE  2

                                 SUMMARY OF EXPERIMENTS DISCUSSED IN TEXT

I. Experiments with Fe(II) in the
absence of dissolved oxygen
A. Tests with pure ortho-
phosphate solutions
1. Effect of pH on precipi-
tate formation
2. Effect of pH on speed
and efficiency of
reaction
3. Effect of reactant con-
centration on speed and
efficiency of reaction

B. Test with orthophosphate
solution containing
420 mg/t NaHCOj
C. Test with secondary
effluent (Wastewater #1,
Table 1)
D. Test with pyro- and tri-
polyphosphate

n. Experiments with Fe(n) in the
presence of dissolved oxygen
A. Test with orthophosphate
solution containing
420 mg// NaHCOj
B, Test with secondary
effluent (Wastewater HZ,
Table 1)
C. Settling r?te studies with
secondary effluent (Waste-
water #3, Table 1)
HI. Settling rate studies with
Fe(in) and Al(IH) on secondary
effluent
A. Experiments with Al(UI)
(Wastewater #4, Table 1)
B. Experiments with Fe(IH)
(Wastewater #3, Table 1)
Exp.
No.




1-5

6-10


11-13



14


15



16, 17



18


19, 20


21-26





27-32

33-38

Initial
PH




5.0-
9.0
6.0-
10.0

7.9



7.9


7.4



7.9



7.2


7.5,
7.4

-





-

-

Initial Phosphate
Concentration
(mg/ 1 P)




12

12


12, 4



12


15.8



18(pyro-)
21.6(tripoly-)


12


9.1


18.3





11.2

9.0

Cation-to-
Phosphate
Equivalence
Ratio




1.0

1.0


1.0, 1.5,
0.33


1.0


1.5



1.0



1.0


0.5, 1.0


0.4-2.0





0.4-2.0

0.4-2.0

Parameters
Measured or
Observed*




pH, TF

pH, RRC,
PC, PCN

pH, PC,
RRC
(Exp. 11
only)
pH, RRC,
PC

pH, RRC,
PC


pH, RRC,
PC


pH, RRC,
PC

pH, RRC,
PC

pH, TF,
SR, RRC




pH. TF,
SR, RRC
pH, TF
SR, RRC
Table( s)
Containing
Results




3,4

5


6,7



8


9



10



12


13


14, 15





16

17

Abbreviations: TF = Turbidity Formation, RRC = Residual Reactant Concentration, PC ^ Precipitate Color,
PCN = Precipitate Chemical Nature,  SR = Settling Rate
                                                           22

-------






















W
J
ffl
2
H















W
H
t~M

«
P-i

o
ffi
P^
0
ffi
H
BJ
Q

<2
n
0.

(^
 03s
CD 11
a ffi
'£"' (i
*"• ^_^
_»
B4 -H
W '£
HH
"-^^



O
^ .
-u Is-
§ II
c S!
'S a
rti i— H
J|

o"
<— H
4J un
d
1 1
QJ ' 1
S a:
^
|;|


f-< MH
CD O
-1-1
N
• r^ *"*^
'd LD
*H LJ
J_, l—^
rj ^..^^
H



ffi
V-M




>•>
"tn ^^
T3 ^i
£ EH
d ^2-
EH

ffi
(X
£•_
'O 1— '
**"* F~H
|S

ffi
a.




§
CD
h




r^ r™H (\J ^J 0^ ^J ^J ^] '




^3 (\J nH (\J QQ ^^ QQ f^ ,_( ^^
oopofMoomin^^Ttn^

^ ^






^ OOfMrOoOOOir)i4ff~
o-i-i-H^r-cooooo


r^~ r^" ts^ t^ oo fM c^ v o c^ ^^
r^~ r** ^~ Is* r*** t^* LO "^ ^^ ^*


^J ^3 (N^ ^J cO oO ^J ^1 ^J cO

1 oooooooooo

OOOOOOOOOOOOOOOOOO



^,
g g | -S 5
ss^i^lllj^^1^^1
" " ID O |A5rCrClrClri3
(vj^^nrO^Hr-icMro^invO
















, — .
t— i
t-l

h
MH
O
0
• H
-4-1
•H
rt
CD
"rt
.

bO
Vi
d
T)
-t-i
fl
CD
a
+->
•H,
•^j
ni
ffi
O

nj
CD
cxJ

CQ
O
Q,


"5
0
i
o
1
Cj"
H- 1
u
tn
• »s
iX
-5
bfi

fXl
r— 1
II
C!
O
.^
ntrat:
CD
o
fl
o
o
CD
+>
rt
0.
CO
o
r£
P,
^
£H
O
i— I
• r-l
1— |




























O
•
o
£

• l-t
C71
CD























CD
_J
(w
a
CD
CD
CQ
4->
d
erne
iH
measi
o
d
CO
CD

"rt
o
•rH

1
u
23

-------
       w
       H
w
J
ffl
H  ^
H  fc
n  °
§H
       ^ U
       o w
       Q cd
        W
        O
        K
        U
* *i-H ^••s
_i j ^\ r^ *"")
fl M 'o (H
(U II rQ ~
d Hrl rt "^
fi K 3 •—
'§ -? ^
Hrtj h"1
2 K
M ft
^^ ?**
• •!— 1 *•"*••»
-i-> CO T3 ID
g n 3 H
g tr H 3
C 4-, jj - —
•fj ft H
CD f-J
&'H
X -P M
H -g ffi
**^*
CU O
Time Af1
Addition
Fe(II)

CO PO rO


O O O O
QN Q\ Q\ ON

U O O vO
1 c\] CM f\}


o oo in \o
o 0s 0s co
oo r- r~ r^-

ti ft
.H -S 'H 'S
C\j in —I ro

CM (M rH C\J CO in


o o o o o o
• ( • * * •

CM oo oo r^ oo oo


CM in O CO ro r-H
t~- vo xo m m m
f^. f^* f^» [*^. f^« ^^.

•— i r— < CM ro ^ m
m
t) __!
0
d
o
•H
ig
nJ
bJD
•S
'^
T3

fi w
£ «*
f4
O r-^H
CJ CJ
H~I IT*
ft £
a
ft
CO
O
ft
O
O
1
o
1
g
"oT
fc
t"
be
a
CM
r— I
II
fi
0
•H
"S
fll °
o -1
§ »
0 O
•H
CO -4->
£ a
ni >-i
"r^ ^
0 PI
r* tU
r*-H p^.)

o »>
•S '3
M ^H
O 0)
,0














(U
nJ
a
CQ

emen
^<
3
meas
o
ti

CO
0)
-t->
nJ
CJ
•r-t
S
•r-J
1
CJ

-------
precipitates were larger in size than in Experiment 4.  No change in
solution pH occurred subsequent to addition of the Fe(II) in this experi-
ment and turbidity remained constant after the initial reactions.

As will be  seen later, the  observed changes in pH and turbidity of the
product solutions correspond to changes in the concentration of residual
soluble reactants.  If the overall precipitation reaction is represented
by 3Fe+2 + 2H  PO" (or 2HPO ~2) = Fe (PO  )  + 4H+ (or 2H+), the
release of protons from such a reaction can account for the observed
decrease in the pH.  The data in Tables 3 and 4 indicate that the rate of
precipitation of orthophosphate with Fe(II) in the pH 5 to 8 region is  pH
dependent,  increasing with a  rise in pH. As will be discussed later,
the formation of turbidity at pH 9. 0 is due to the precipitation of the
excess iron as hydroxide in addition to precipitation of ferrous  phos-
phate.

Effect of pH on Rate and Efficiency of Reactant Removal (Experi-
ments^)- 10).  Table 5 contains  data on  changes  in solution pH and resid-
ual reactant concentrations for a reaction time of 5 hr for five experi-
ments (Experiments 6 through 10) conducted at initial pH levels of 6. 0,
7. 0,   7.9,  8.9, and 10. 0.  In  Experiments 6, 1, and 9 the turbidity of
the product solution was also measured at the termination of the experi-
ments (after 5 hr).  The following is a discussion of results for the
individual experiments.

In Experiment 6 (initial pH =  6. 0) the changes in pH and reactant  con-
centrations were fairly  small.  The total drop in pH and orthophosphate
and ferrous iron concentrations were 0. 2 pH unit,  and  7 and 9%,
respectively.  The turbidity of the reacted mixture  after 5 hr was only
4 JTU in this experiment.

In Experiment 7,  the solution pH remained constant at  7. 0 for about
20 sec following the termination of pH adjustment and then dropped to
the various levels indicated in Table 5.   Most of the change in solution
pH occurred during the  second half of the first hour.  This decrease in
the pH was accompanied by a significant decrease in residual reactant
concentrations.   The orthophosphate and ferrous iron  removals re-
mained essentially constant after the second hour with maximum
removal levels of 35  and 39%, respectively.  In this experiment,  the
turbidity of the  solution mixture at the end of the 5 hr was 22 JTU.

The  solution pH in Experiment 8 was 7.90 at the termination of pH
adjustments and 7. 89 at the end of 5 min.  After 5 min,  the solution
pH decreased slightly,  reaching a minimum value of 7.82 by the end of

-------
m

W
     a

,
2
H

g
E
n
O
W
ti
&
w
      w
      O
      E
      cd
      O
      ffi
      cu
      H
      U
      W
0°
~* 0
4J v™H
5 it
6 »T«
K
•c °-
ID *— *
IS
w •?
S
o-5:
-M 00
g „
Effi
•c ^
fr-3
w -s
£1
co5:
•a1"
5 ti
E a
•fi *
0) TJ
8-3
w -g
£
<-°
.4J f-
V II
E ffi
•n *
w •-<
fi-3
W '3
3
^ °
4J vO
£3
V II
EE
•fi *
rt
*I
&
i
fiSI
S ri >
K « S
£ s S
PH fvi f/

u
E
p.
JJ 4J f-l
« )S 2
5 rt 5*
2S P
« S §
^ erf cd
u
ffi
ft
.u •+•> i— i
d C «)
«) rt >
y o 8
S •* 6
« V 0
w Cd CJ
u
ffi
P4
• 1 4J i-l
^ 5 2
S$ 2
S^ 6
» (1) 
* o §
« •« s
« o (U
(^ ci oi
U
ffi
O.
V
6
H
0
fu

04



V
Cu

P*


U
fe
0.


u
h
(X


V
fe
0*


oooooooo
oooooooo
I I I I I I I I
t^cOOOrOcOOOO

o^ oo oo oo oo oo in
O^ O^ (Xs O** O^ O^ O^
0^ O^ O^* O^ O^ O^ 0s
o o o o o o o
, o o o o o o o
S & « ft a A a
1 CT^ 0s 0s 0s 0s ^ ^
sO \O vO vO ^^ vO NO
00 CO CO 00 00 00 00
CO 00 00 00 00 00 00
00. CO 00 00 00 CO 00
00 O^ ^1* vO vO t^ t>-
i i*- r- o^ 0^ o> o^ 0^
1 «Nl CO M CO •<* Tf •<*<
vO >O 00 ON ON ON ON
ONQO•>
CM
CM


D
H
H»
^
Final
Turbidity
II
o

rt
M
O
P!
cu
r— i
n)
•M

CD
_|J
n)
CU
CO
O
!•
rg
o
t
o
1
g
h
07
"Sb
E
fM
r— «
II
PI
0
• r-l
"rt
|
P!
O
u
cu
rt
•§,
CO
O
O
O

rt
•43
•0
jS

•T?
X
CU
o

1
CU
-d
pi
Pi
o
• H
4J
^
T)
nt
g
"w"
o
M
n)
4J
CO
0)
E
O

**H
cu
E
•H
H
JO
*rt
C
•H
E
rf
O

s-
o
VI
PI
a
+3
*g
•*
1
h
u
PI
o
u
g
rt
f-t
0
•g
-J
o
*4J
•r-l
T)
tJ
n)
^
£
'a
• rH
4J
• H
1)
n)
jj *-.^'
rt «
CU |r|
rj
£ 0
(0

• r-\ O
"S £
(0 .r-4
M "^
cu m
u






















v
-d
n)
E
cu
ta
V
it
CO
"S
cu
E
V
to
CO
rt
E
o
Pi
CO
cu
JJ
(4
U
•d
•S
1
•o
                                             26

-------
30 min.  In the second half of the first hour the solution pH increased a
total of 0. 17 pH unit and then gradually decreased reaching its initial
value of 7. 89 by the end of 4 hr.  To assure that the observed increase
in solution pH was real and not an artifact, Experiment 8 was repeated
several times.  The pattern of change in pH was the same in all cases.
The increase in solution pH was unique to Experiment 8 and was not
observed in experiments  conducted at other pH levels.

As  with Experiment 7,  a  significant portion of the total reactant re-
moval in Experiment 8 occurred in the second half of the  first hour.
The observed pattern of change in solution pH and the variation in the
rates of orthophosphate and iron removal together with the data collec-
ted on precipitate characteristics (see discussion on precipitate charac-
teristics) suggest that in  the vicinity of pH 8. 0 the mechanism by which
reactants are removed from solution and the composition of the products
formed vary with  time.  Although the exact mechanism(s) which results
in a rise in solution pH is not known,  the following types of reaction
could account for  the observed phenomenon:
     Fe+2 + HO = Fe(OH)+ + H+

     H+ + OH" = HO
                  L*

     Fe+2 + 2 OH" = Fe(OH) (s)
Reactions during initial
pH adjustment
     3Fe(OH)+ + 2HPO"2 -  Fe (PO )  (s) + 2H O + OH"
                      T:       J    4 ^       Lt

     3Fe(OH) + 2HPO~2 -  Fe (PO )  (s) + 2H O + 4 OH"
             £        4t       j    ^t £*       c*
                  Subsequent
                  reactions
During the initial pH adjustment,  some of the iron may form soluble
hydroxy complexes or precipitates as hydroxide.  Subsequent reaction
of orthophosphate with ferrous hydroxide or ferrous hydroxy complexes
results in the release of OH  ion and thus a rise in pH (the last reac-
tions) .

In conjunction with Experiment 8, three side experiments were con-
ducted to assess (a.) the extent of oxidation of  ferrous to ferric iron
during the  5-hr precipitation test, (b) the  effect of prolonged reaction
time on the efficiency of phosphate removal, and (c) any release of
phosphate from the precipitate when the reacted sample is exposed to
air.   Three 40-ml aliquots of the  reaction mixture (designated as
Samples A, B and C) were removed from the reaction beaker at the end
                               27

-------
of 5 hours.  Sample A was acidified with 1 ml N_ HC1 and analyzed for
both ferrous and total (ferrous plus ferric) iron content.  The concen-
tration of ferric iron in this sample was found to be less than 1 rag/JL
which indicated that under the experimental conditions used the extent
of oxidation of ferrous iron to ferric form was small (less than 3%).
Samples B and C were not acidified.  Sample B was thoroughly purged
with argon and kept under an argon atmosphere.  Sample C was not
purged with argon and was left exposed to the laboratory atmosphere.
On standing overnight the precipitate in flask C developed a blue color
due apparently to partial  oxidation of ferrous iron in the precipitate
(see discussion on precipitate characteristics below).   The precipi-
tate in flask B retained its initial light green color.  Both samples
were filtered  through 100 mjj, membrane filters and the filtrates were
analyzed for orthophosphate and residual iron content.  No ferrous
iron could be  detected in  either sample.  The orthophosphate concen-
tration was  0. 81  mg/A  P for sample B and 1. 09 mg/JL P for sample C
(the orthophosphate residual measured at the end of 5 hr in Experi-
ment 8 had been 0. 71 mg/l P).   These results indicate that, for the
conditions of Experiment 8,  the partial oxidation of  the ferrous iron
in the precipitate to the ferric form only results in a small release of
orthophosphate to solution, and that reaction times in excess of 5 hr
will not result in an increase in orthophosphate removal. (Indeed,  in
both Experiments 7 and 8, most of the change in the residual reactant
concentrations occurred within the first 2 hours.)

As indicated by the data in Table  5, in Experiment 9 there was no
measurable change in solution pH or residual reactant concentrations
with time.  Large green precipitates (presumably containing Fe(OH) )
were formed immediately after the addition of Fe(U).  No iron could
be detected  in any of the filtrates.  The maximum orthophosphate
removal and the final turbidity of the product solution were 69% and
35 JTU,  respectively.  Except for less removal of the phosphate, the
results of Experiment  10 were very similar to those of Experiment 9.

In summary,  both the speed and efficiency of orthophosphate removal
with Fe(II) are strongly pH dependent.  The maximum orthophosphate
removal efficiencies obtained in Experiments 6-10 are plotted in
Figure 5  as a function of  pH. (This figure also includes precipitation
data for Fe(III) and Al(III) from Reference 4, which  will be  discussed
later. ) The pH for maximum orthophosphate removal is  in the vicinity
of 8. 0 for Fe(II).  In this study,  no data were collected on orthophos-
phate precipitation at intermediate pH levels in the pH  7 to  8 and  pH 8
to 9  regions to determine the exact pH of maximum orthophosphate
removal. As will be discussed later,  the observed pH dependency of
                               28

-------
 en

 z
 01
 O
 Z
 O
 u
 uj
 Q-
 t/3
 O
 I
 0-
 O
 EC
 O
 oi
 CC
      12
      11
 _    10
 CL


^



 I     9
                 Fe(lll)
                                                                                                       10
20
                                                                                                       30
40     c
                                                                                                       50
60
70
                                                                                                       80
                                                                                                       90
                                                                                                       100
o_
(/)
O
I
Q_
O
I

cr
O
                                                                                     10
                                              SOLUTION pH
                Figure 5.  Precipitation of Orthophosphate with Fed I), A 1(1 1 1), and Fed 1 1) at a Cation-to-

                         Orthophosphate Equivalence Ratio of 1 0
                         Initial Orthophosphate Concentration = 1 2 mg/j2  P N


                         Residual Orthophosphate Determined on  100 mu.  Membrane Filtrates
                                                   29

-------
orthophosphate removal with Fe(II) can be explained in terms of a com-
petition between hydroxide ion and orthophosphate ion to react with
ferrous ion.

Precipitate Characteristics.  The precipitates obtained in Experi-
ments 6-10 were observed for color  characteristics and some were
also examined by the powder x-ray diffraction technique.  In all cases
the precipitates retained on 100 mji membrane filters were initially
light green at lower pH levels  and darker green at higher pH levels.
On exposure  to air,  however,  all precipitates changed color.   The
precipitates from 15 min and 30  min samples in Experiments 7 and 8
(initial pH 7. 0 and 7.9,  respectively) and all precipitates from Experi-
ments 6,  9 and 10 (initial pH 6.0, 8.9 and 10.0, respectively) turned
yellow.  The precipitates from the 1  to 5 hr samples in Experiment 7
developed a light blue-green color and those  in Experiment 8  turned
intensely blue.  The yellow color was most intense in the precipitates
from Experiments 9 and 10 and was barely visible in samples from
Experiment 6. The 5 hr sample from Experiment 8 was initially blue
but changed to green on vacuum drying (see below).

The development of the yellow color  in the precipitates is believed to
be due to the oxidation of ferrous iron by the atmospheric oxygen.  In
Experiment 9, following the termination  of the experiment under argon
atmosphere,  the  solution containing the precipitates was allowed to
stand exposed to  laboratory atmosphere.  It,  too,  developed an intense
yellow color  on exposure to air.

The pattern of color change from green to blue-green or to intense blue
has been  reported to be a characteristic  of the mineral vivianite,
Fe (PO )  • 8H O.* '  When freshly mined, vivianite is almost colorless
(actually  a very pale green) but develops a blue color on exposure to
air.  The change in  color is attributed to the partial oxidation of Fe(II)
to Fe(III). In Experiment 8, following the termination of the  experi-
ment,  the solution containing the precipitate  was exposed to air.  On
overnight standing, this precipitate also  developed an intense blue
color.  The blue precipitate was collected by centrifugation and dried
in a desiccator under vacuum at room temperature.  The vacuum dry-
ing of the precipitate resulted in an irreversible change in the precipi-
tate color from blue to green.

The vacuum  dried precipitate and the materials from the 15 min and
5 hr samples dried in the laboratory atmosphere (also from Experiment
8) were examined by the powder x-ray diffraction technique.  The 15 min
sample was amorphous  to x-rays but the 5 hr sample and the vacuum
dried precipitates gave  sharp diffraction patterns corresponding to that
for vivianite.  Since x-ray analysis was conducted under a  vacuum
                              30

-------
environment,  the 5 hr precipitate, initially blue in color, developed a
green color which did not revert to blue on subsequent exposure to air.
No explanation can be offered for this color reversion.

The x-ray diffraction analysis results for the precipitates in Experi-
ment 8, and the observed variation in the color of the precipitates in
various experiments following exposure to air indicate that composi-
tion of the precipitate is affected by both the pH of the precipitation
and by the reaction time.

Effect of Reactant Concentration on Reaction Speed and Efficiency
of Reactant Removal (Experiments 11-13).  Experiments  1 through 10
(Tables 3, 4 and 5) were all performed on 12 mg/X P orthophosphate
solutions  using an Fe(II)-to-orthophosphate equivalence ratio of 1. 0
(a ferrous iron dosage of 32.4 mg/&).  To assess the effect of reactant
concentrations on the rate and efficiency of reactant removal,  three
other experiments (Experiments 11,  12 and 13) were conducted.  Ex-
periments 11  and 12  were  conducted  on  12 m.g/1 P orthophosphate solu-
tions using Fe(II)-to-orthophosphate  equivalence  ratios of 1.5  and 0.33,
respectively (ferrous iron dosages of 48. 6 mg/4 and 10. 8 mg/^).  The
ferrous iron dosage in Experiment 13 was the same as that in  Experi-
ment  12,  but the orthophosphate concentration was reduced to
4. 0 mg/l  P (Fe(II)-to-orthophosphate equivalence ratio =  1.0).

The experimental results for Experiment 11 are  presented in Table 6.
The solution pH decreased  slightly in the first half hour and increased
slightly in the second half hour.  Visual observation of the test solu-
tion during this  experiment indicated a substantial increase in solution
turbidity during the second half of the first hour (i. e. concurrent •with
the  slight rise in solution pH).  As indicated in Table 6,  the percent
total orthophosphate  removal during  this 1/2 hour period increased
from 86% to 99%.   This  additional removal of orthophosphate,  how-
ever,  was not accompanied by a proportionate rise in the extent of
ferrous iron precipitation.   No significant change in the concentration
of either reactant was observed after the first hour.  For a reaction
time of 5  hr,  maximum  removals of  orthophosphate and ferrous iron
were 99 and 76%, respectively.  The results thus indicate that at a pH
close to 8, not all the excess ferrous iron added  to the phosphate solu-
tion precipitates with the phosphate or as hydroxide.  Turbidity measure-
ments made on an aliquot removed at the end of the 5 hr test period
yielded a  value of 63 JTU.  All precipitates retained on 100 mjj, mem-
brane filtrates were  initially light green in color.  On exposure to air,
the precipitates from 15 min and 30 min samples turned yellow whereas
the precipitates from 1 to 5 hr samples  developed a dark  bluish green
color.
                               31

-------
                     TABLE 6

ORTHOPHOSPHATE PRECIPITATION WITH Fe(H)
      AT AN Fe(n)-TO-ORTHOPHOSPHATE
         EQUIVALENCE RATIO OF 1.5a

                (Experiment  11)
Time
0 min
5 min
15 min
30 min
45 min
1 hr
1-1/4 hr
2 hr
3 hr
4 hr
5 hr
PHC

7.88
7.81
7.81
7.90
7.95
7.90
7. 78
7. 71
7.68
-
Percent Reactant
Removal
P
0
_d
79.5
86.3
99.0
Fe
0
-
67.9
71.4
75. 1
'
99.0
99.3
99.4
99.2
76.2
75.5
76.4
74. 5
 Initial orthophosphate concentration =  12 mg/4 P

 Time from the start of Fe(II) addition  and beginning
 of pH adjustment

Q
 pH adjustment made initially by addition of 0. 1 N
 NaOH concurrent with and for about 4  min after the
 addition of Fe(II)
  - indicates no measurements were made

-------
Comparison of the results of Experiment 11 with those for Experi-
ment 8 (Table 5) indicates that in terms of the patterns of change in
pH,  reactant removal and precipitate color,  orthophosphate precipi-
tation at an Fe(II)-to-orthophosphate equivalence ratio of  1. 5 is very
similar to that for a reactant ratio of 1.0.

The data on formation and characteristics of the precipitates and on
changes in solution pH for Experiments  12 and 13 are presented in
Table 7.  (In these experiments   no data were collected on the change
in residual reactant concentration with  time.)  Comparison of the
results of Experiment 12 with the results of Experiments  11 (Table 6),
8 (Table 5)  and 4 (Table 4) indicates that for an initial orthophosphate
concentration of 12 mg/A P, the patterns of change in solution pH  and
turbidity and in the color of the precipitates are essentially unaffected
by the change in Fe(II) concentration.  In Experiment 13 there were
essentially no changes in solution pH and no visible formation of tur-
bidity during the first 2  hr of the test.   After  2 hr and 20 min,  how-
ever, precipitation occurred accompanied by  a rapid drop in solution
pH.   Only the 3 to  5 hr precipitates in this experiment developed the
characteristic bluish color of vivianite.  Comparison of the results of
Experiments 12 and 13 thus  indicates that the rate of orthophosphate
precipitation with Fe(II) is lower at lower orthophosphate  concentra-
tions.

Effect of Anions on Orthophosphate Precipitation with Fe(II).   The
effect of anions on orthophosphate precipitation with Fe(II) was evalu-
ated in two  experiments.  Ferrous ammonium sulfate had been used as
source of Fe(II) in Experiments  1-13.   To compare the effect of sulfate
ion with an anion such as perchlorate which has a lower tendency for
complexing cations,  an experiment was performed with ferrous per-
chlorate (orthophosphate 12  mg/H P, reactant equivalence ratio  1.  0,
initial pH 7.98).   The pattern of turbidity development and change in
precipitate  color in this experiment were similar to those in the ex-
periments using ferrous ammonium sulfate.

Experiment 14 was an evaluation of the  effect of bicarbonate ion
(420 mg/£  NaHCO ) on orthophosphate precipitation with Fe(II).  The
results are presented in Table 8.  The  solution pH in this experiment
decreased slightly during the first hour and increased thereafter.   The
rise in solution pH is probably due to the expulsion of some CO  from
the system due to purging with argon.  (This was verified in one experi-
ment in which a solution of 210 mg/JL NaHCO   was treated with argon in
the same manner as the phosphate test  solution; the pH rose from an
initial value of 8. 27 to 8. 57  by the end of 5 hr. )  The data in Table 8
                               33

-------
H
HH
W

g
   O


   o
^ r
W H

IS ^
O !z
U O

WH
H<:

rt
CD
a
JH
CD
8-









CD
"^
M «!
r5 °
U CD
CD 3
M W
II
•3 w
CD
h
PH
-3
K
a
CD
£ h
h <
r5 0
O "*"*
tj CD
CD 3
-P CQ
$ o
US4
•3 w
CD

PH


TJ
ffi
(X

rt
CD
a
H



£
o
r-t
i-H
CD
•s '
• H
rt
h


CO -1
0s* O^
r- r-





£
o
'iJ CD
!>-< 1





r-- oo
°^ i>
[•*•- [s^

rt rt
•r-l -H
a a
in in
r-H (M



> ^ ^
o o o
I— ( r-t i— 1
,-H i— i ,—1 m
CD CD CD
-p -P -P 1
rt rt rt
• H •!-! «r-t
rt rt rt


•— i oo r~- r—
O 00 00 00
r- r- r- t-



§
CD
bO
CD CD CD
3 3 3 i





o^ o o in
vO m m ^
r^ r^ r^ r*™*
T)
f S S
•H rt •!-<
ad
!-( fn }-t H
O O
rO •— i CM p-3 (VJ
^
o
I— 1
r-H
CD
1
CQ
•H
3
r~H
X!
CD
i— I
rt
PH


CM
^h
^






CD
i— I
CQ





m
•*
^


h
PO
^
o
t— 1
I — 1
CD
1
CQ
3
CD
i-H
rt


O CM
CM CM
t- t-






CD CD
S 3





m oo
TJH ^
r-^ c^


h h
•^ m

CD
a

CO
rt
ffi
0.
O
bO

•H
rt
rt
'bO
CD
rO
T)
rt
rt
rt
o
'•4J
•H
ti
rt
§
h
"o
•j^
rt
4J
CQ
CD
a
o
£

a
•H
H
rt
CO
CO
o
II
o
•r-l
-P
rt
CD
U
rt
i— i
rt

*j
&
CD
«
rt
CO
O

o,
O

-£
O
i
O
-p
—•
Cij

.^
bfl
a
i— i
n
PH
•.
="?
bO
a
00
o
r—J

II
^^
a
IT
PL,
ro
o
i— i
H
0
•H
CD
O
fl
CD
r—t
rt

•1-1
0*
CD
CD
n!
CU
CO
O
rC
Pi
O
•^
^1
o
1
o
1
§
CD


i"
bO
a
•<*<
PH
• *>
^
bO
a
oo
0
i-H

ti
ii^^s
1-H
tn
u
1
rt
-u

fl
CD

t
3
U
fl
0
u
HI
O
rt
^^

z
r— H

O
-
o

•43 J»
^a
'Ti '( 1

rt u
•^.2
1 1
• rl rt
•rt CD
•S^
CD ^
'S ^
S»j 1
rt
•s.s
CD d
a «
CQ
d 0
•?-J -p

rt 03
ffi rS
n. cj
TJ

















CD
O
CD
•— i
,— i
O
O
-p
o
CO
rt
£
CD
i— t
%
rt
CO
rt
CO
CD
•+-1
rt
o

rt
i
CD













T)
CD
J>
>H
CQ
O
CQ
rt
*

IT
TJ
•r-t
rQ
3
•S
CD
W
al increa
•rH
"rt
rt
CO
cH
m

-------
                     TABLE  8

EFFECT OF BICARBONATE ALKALINITY ON THE
     PRECIPITATION OF ORTHOPHOSPHATE
                   WITH Fe(H)a
                 (Experiment 14)
Time
0 min
5 min
15 min
30 min
1 hr
2 hr
4 hr
5 hr
PH°

7.88
7.88
7.85
7.90
8. 10
8. 18
8.28
Percent Reactant
Removal
P
0
56.1
59.8
62.4
79.0
86.8
d
90.3
Fe
0
75.0
81.5
88.9
90.4
96.3
98.4
o
 Initial orthophosphate concentration 12 mg/4 P,
 initial bicarbonate alkalinity = 420 rag/A NaHCO ,
 Fe(II)-to-orthophosphate equivalence ratio = 1

 Time from start of Fe(II) addition
Q
 No supplementary pH adjustment made with NaOH;
 initial pH after addition of Fe(II) (time = 20 sec) = 7.88
 - indicates no measurements made
                        35

-------
indicate that a  substantial portion of orthophosphate was precipitated
in the first 5 min.  In this experiment,  as with the experiments with
pure orthophosphate solutions,  there was a  significant increase in
orthophosphate removal in the second half of the first hour; that ortho-
phosphate and ferrous iron removals were essentially complete in
about 2 hr.  For the reaction time of 5 hr, maximum orthophosphate
and ferrous iron removals were 90 and 98%, respectively.  The corres-
ponding data for  Experiment 8 (Table 5) conducted in the absence  of
bicarbonate alkalinity were 94% P,  97% Fe.  Considering the small
difference between the solution pH in the two experiments and allowing
for a certain degree of experimental error, the results  would thus
indicate that for  the conditions tested the removal efficiencies are not
appreciably affected by the presence of bicarbonate ion.

Orthophosphate Precipitation with  Fe(II) from Actual Wastewater
(Experiment 15).  The data obtained in the phosphate precipitation from
secondary effluent (Wastewater #1, Table 1) are presented in Table 9.
The addition of Fe(II) to the wastewater resulted in a small immediate
(within about 20 sec) drop in pH followed by a subsequent gradual  fur-
ther decrease  which lasted for about 1 hr.   After  1 hr,  the solution
pH gradually increased, probably due to the expulsion of some CO
from the system due to purging with argon.   As indicated by the data
in Table 9,  a substantial portion of the orthophosphate was precipi-
tated in the  first 15 min and the increase in the percent  reactant re-
moval was small after the second hour.  For the reaction time of 5 hr,
maximum removals of orthophosphate and iron were 98  and 66%,
respectively.  The removal efficiencies indicate a near  1:1 Fe(II)-
orthophosphate reaction stoichiometry.  As with the experiment with
pure orthophosphate solution (Experiment 11),  not all the excess
ferrous iron added to the sample precipitated with the orthophosphate
or as hydroxide.

Visual observation of the wastewater sample during the  precipitation
experiment  indicated that little turbidity was formed during the first
15 min after the addition of Fe(II).  A substantial  amount of turbidity,
however,  was  formed after 15 min.  All precipitates retained on
100 mjj, membrane filters were initially light green in color.  On
exposure to air,  the precipitate from the 15 min sample remained
light green.  The 30 min sample developed  a darker green  color.  All
1 to 5  hr samples developed a blue-green color.  As was discussed
earlier,  the pattern of color  change from light green to  blue  is reported
to be a characteristic of vivianite,  Fe (PO  )  • 8H O.
                                     j   4 LI    c*
                               36

-------
                               TABLE 9

ORTHOPHOSPHATE PRECIPITATION FROM SECONDARY EFFLUENT
       WITH Fe(II) IN THE ABSENCE OF DISSOLVED OXYGENa

                           (Experiment 15)
Time
0 min
5 min
15 min
30 min
1 hr
2 hr
3 hr
4 hr
5 hr
PH°
7.80
7.30
7.21
7.02
7.03
7.25
7.42
-
7.77
Percent Reactant
Removal
P
0
_d
48.2
^71.9
85.3
94. 1
95.2
95.8
98. 0
Fe
0
33. 5
47.3
54.3
59.1
61.7
60.9
65.9
   Wastewater #1 (Table 1);  Fe(II)-to-orthophosphate equivalence
   ratio = 1.5

   Time  after start of Fe(II)  addition; Fe(II) added after the sample
   had been purged with argon for about 1 hr
  c
   No supplementary pH adjustment made with NaOH; initial pH
   after addition of Fe(II) (time =*  20 sec) =  7.40

   - indicates no measurements made
                                  37

-------
In terms of the rate of orthophosphate removal, the general pattern of
turbidity development and the change in precipitate color on exposure
to air, the precipitation of orthophosphate with Fe(II) from actual
wastewater appears to be very similar to that for orthophosphate pre-
cipitation from pure orthophosphate solutions and from orthophosphate
solutions  containing NaHCO  .

Precipitation of Pyro- and Tripolyphosphates with Fe(II)
(Experiments 16 and 17).  Solutions containing orthophosphate were used
in Experiments  1 through  14.  In some wastewaters, however,  condensed
phosphates  such as pyro- and tripolyphosphates account for a significant
portion of the total phosphate content.  The  effectiveness of ferrous iron
for the precipitation of condensed phosphates was examined in two ex-
periments (Experiments 16 and 17),  conducted on 1. 16 meq/1  solutions
of pyrophosphate (18 mg/A P) and tripolyphosphate (21.6 mg/£ P) using
an equivalent reactant ratio of 1.0.  The results are presented in
Table 10.

In the pyrophosphate experiment, the solution pH decreased slightly at
first  reaching a minimum value  of 7.81 by the end of 45 min.   After
45 min, the pH  increased  slightly reaching a maximum value of 7. 98
by the end of the first hour and decreased gradually thereafter.  A
substantial amount of turbidity was formed during the last quarter of
the first hour.  In the tripolyphosphate experiment,  there was a small
decrease in pH,  but the solution remained clear during the 5 hr obser-
vation period.

In both Experiments  16 and 17,  the precipitates retained on 100 m^
membrane filters were initially  light green.   On exposure to air,  all
precipitates in the tripolyphosphate test  and the 15 min precipitate in
the pyrophosphate test  turned yellow and all 1 to 5 hr  precipitates  in
the pyrophosphate test  developed a yellow-green color.  As indicated
in Table 10, maximum phosphate and iron removals obtained during
the 5 hr test period were 87% P  and 91% Fe in Experiment 17.   The
data for orthophosphate precipitation at  12 mg/£ P level (also  1. 16 meq/^)
for Experiment 8 (Table 5),  indicate maximum orthophosphate and iron
removals of 94% and 97%, respectively.  Thus, on an equivalent basis
and at the pH levels examined,  orthophosphate is more efficiently pre-
cipitated  than pyrophosphate which in turn is more efficiently precipi-
tated than tripolyphosphate.  The data on the patterns of change in pH
and on precipitate formation indicate a more rapid rate of precipitation
for orthophosphate than for pyrophosphate.

-------
                              TABLE 10

    PRECIPITATION OF PYRO- AND TRIPOLYPHOSPHATE WITH
          Fe(II) IN THE ABSENCE OF DISSOLVED OXYGENa
Time
0 min
5 min
15 min
30 min
45 min
1 hr
2 hr
3 hr
4 hr
5 hr
Experiment 16
Pvropho sphate
PHC

7.94
7.89
7. 85
7.81
7.98
7.96
-
-
7.59
Percent Reactant
Removal
P
0
_d
20. 7
-
-
76. 7
85. 7
86.6
86.4
84.5
Fe
0
-
27.3
-
-
79. 2
89-3
90. 5
91.5
88.3
Experiment 17
Tripolypho sphate
PHC

7.91
7.89
7.87
7.85
7.83
.7.80
7. 77
7.72
7.71
Percent Reactant
Removal
P
0
-
10.4
10.6
-
10. 1
-
9.5
-
-
Fe
0
-
10.3
11.6
-
9.6
-
9.0
-
-
 Phosphate initially present at  1. 16 meq/.£; 18 mg/j£ P pyrophosphate
 and 21.6 mg/l P tripolyphosphate; Fe(II)-to-phosphate equivalence
 ratio =1.0

 Time from start of Fe(II) addition

"pH adjustment made initially be addition of 0. 1 N NaOH; pH at the
 temination of pH adjustment (time ~ 4 min) = 8. 00,  and 7. 98 in the
 experiments with pyro- and tripolyphosphate, respectively

 - indicates no measurements made
                                 39

-------
Comparison of Phosphate Precipitation Using Fe(II) with Phosphate
Precipitation  Using Fe(III) and Al(III) Salts

Effect of pH on Orthophosphate Precipitation.  The precipitation of
orthophosphate with Fe(III) and Al(III) was recently investigated in some
detail.  '  Figure 5 compares the orthophosphate precipitation results
for Fe(II), Fe(III) and Al(III)  salts at an equivalence reactant ratio of
1. 0.  From the data in this figure it may be seen that the pH of maxi-
mum orthophosphate removal is in the vicinity of 8 for Fe(II),  close to
3. 5 for Fe(III) and about 5 for Al(III).  Also, for the conditions tested,
minimum residual orthophosphate concentrations of 0. 7,  1.8 and
3. 5 mg/1 P were obtained with Fe(II), Fe(III), and Al(III), respectively.
Thus,  at the respective pH of maximum effectiveness and on an equiv-
alent basis, ferrous iron is more effective than either Al(III) or Fe(III)
for the precipitation of orthophosphate.  The data  in Figure 5 indicate
that,  compared to Fe(III) and Al(IlI), the  effectiveness of Fe(II) for
phosphate precipitation extends over a relatively narrow pH range. It
should, however, be noted that the results for experiments with Fe(II)
are for a reaction time of 5 hr.   While in  the vicinity of pH 8 and at
higher pH levels, Fe(II)-orthophosphate reactions may be essentially
complete in 5 hr, at lower pH levels,  a longer time may be necessary
to achieve equilibrium and attain maximum removal of phosphates.

The pH of most wastewaters is in the 7 to 8 range. To reach the opti-
mum precipitation pH with Fe(III) and Al(III), a substantial  amount of
acid or a considerable excess of  coagulant must be added to the waste-
water to overcome its natural buffer capacity.  When Fe(II) is used
for phosphate precipitation in the absence of dissolved oxygen,  such an
extensive pH adjustment would not be necessary.   Any minor pH adjust-
ments can be accomplished by addition of  a small  quantity of lime. In
the present study no data were collected on the effect of pH on ortho-
phosphate precipitation with  Fe(II) from actual wastewater.  In water
and wastewater treatment with Al(III) and  Fe(III) salts,  the  treatment
efficiency-pH curve has been shown to be  affected by the presence of
diverse ions.   It is possible  that  the phosphate precipitation-pH  curve
for wastewater is somewhat  different (perhaps broader) than that shown
in Figure 5 for a pure phosphate  solution.

The reaction  of orthophosphate with both Al(III) and Fe(III)  salts has
been shown to be very  rapid  and  completed in less than 1 sec.      At
and near the pH of maximum orthophosphate removals, readily settle-
able floes are formed with both cations and any excess  Al(III) and
Fe(III)  added  to the phosphate solution precipitates with the phosphate
or as hydroxide.'  '  The reaction of ferrous iron  with orthophosphate

-------
in the absence of dissolved oxygen, however, is relatively slow and
results in the formation of a fine and poorly settleable precipitate.
Also, with ferrous iron not all the excess  cation added to the phos-
phate solution precipitates with the phosphate or as hydroxide.

Application of Solubility Data to the Analysis of Phosphate Precipi-
tation Results . The pH dependence of orthophosphate removal
efficiency with hydrolyzing cations  can be  explained in terms of the
solubility of the precipitates formed and the competition between
hydroxide and orthophosphate  ions to react with the cation.  As was
pointed out  before, orthophosphate precipitation with Fe(II) results
in the formaHc.-. of ferrous phosphate.  Table 11 contains some of the
equilibrium relationships  pertinent to the solubility of ferrous phos-
phate in water. Except for a value of 1. 3 x 10     recently reported
by Singer,'"'  no data have been reported on the solubility constant of
ferrous phosphate.  The data compiled by  Sillen and Martell'^' indi-
cate that the values reported for th= equilibrium constant for  reac-
tion 4 (Table  11)  varies over several orders of magnitude.  However,
since the majority of the values listed are close to 10 , this value was
selected as a  representative value for use in the present analysis. On
the basis of the data in Table 11 and the  assumptions that (a) Fe(OH)
and Fe  (PO )   are the only solid phases formed,  and  (b) the formation
of soluble Fe(ll)-phosphate complexes can be neglected,  a mathemati-
cal expression can be derived for the equilibrium total soluble ortho-
phosphate concentration as a function of  pH.

Equilibria 1-3 (Table  11) define the fraction a  of the total orthophos-
phate ion (P' ) which exists in solutions as PO ~  ions at any pH level.
                _       _
a =    r   -    = - - 7-7 - TF       (a)
                                              3  + DK ]
             rp  -j                       +            7
             L  Tj
where brackets represent concentrations in moles/A  and K ,  K  and K
are the  equilibrium constants for reactions 1,  2 and 3, respectively.
From equilibria 4, 5 and 6,  a similar expression can be obtained for the
fraction |3  of the soluble Fe(II)  species which exists as Fe+^ ion.
            [Fe+2]             [H+]3
     ^ =  fre(m--"1 =  [H+]3 + K K  [H+]2 + K K.K 3
                      L  J     4 w         5  6 w


-------
W
H-}
CQ
<
H
s
H
H
S
CO
(1)
u/
fa
W
SPHAT
o
ffi
PH
O
ffi
H
Qj
0
cd
O
fa

CO
&
LATIONSHI
W
2
pq
a
H
CD
CJ
fl
!H
CD
CD
eJ
fi .p
.2 S —
£ 42 o
X3 to O
•H pi m
IH O CM
CT1 '
H











.librium
•I—I
0^
W


0

















cu
rC
p T
CO
O
rC
fr
rQ
4->
h
O
<+H
0
fl
O
• 1-1
15
N
• 1-1
ri
i— i
i— i



rH i— 1 r— 1
i — 1 i — 1 i — 1


CO 00 CO
1 1 r-l
O O 1
•— I i— 1 O
X X -
i— 1 O^
— 1 O 00
• • •
Is- t^ T^










•4- "T"
hn IT1 _i_
^_ -f- "
1 CM +
g o""^
CM & O
ffi ffi &(
II 11 11
^ 1 ^ CM
2 S o"
co CM fL(
K ffi ffi


* ™ CO


O CM
r-t I-H


in
rH
O
1— 1
X
m
t*- CM
o







"aT
*— T^
i i— i
ffi 0
2. ^
to CJ fa
1^ II 1
*2 ' o
>^ O CM
ffi 4- 4-
ti CM CM
04- 4-
•H CD CD
nJ fa fa
U
l-H
\F m


CO
i— i



vO
o
I— 1
X
CO
00






1 CO
ffi
	 ^
o
fa
II
1
ffi
o
't?
/vl
Fe(OH),
i


0


o



o
CO
1
o
1— 1
X
CO







CO
| |t
0
OH
CM

>, ^
4-J 4-
^j flj
•i—( [T i
1 "
-o "
CO ^_fM
CD *^f
ta 0
rd P ]
A «-^
CO CO
0 V
r(~] ["TI
OH
I-H
t^
                                                      142

-------
where K ,  K  and K, are equilibrium constants for reactions 4, 5 and
6 (Table 11), respectively, and K   is the ionization constant for water.
From equilibrium 7
 or
lp(Fe(II)T]
         L
                                  =  K,
(d)
When Fe  (PO )  is the only source of orthophosphate and ferrous iron
in water,  dissolution of one mole of ferrous phosphate would result in
the release of 3 moles of Fe(II)   and 2 moles of PT> i. e.,
        .5[PT] =
 Substituting equation (e) in equation (d):

                    3
 or
JP(1-5)[PT]


FP  1 = 8.26
L TJ
                                                                      (f)
                                                                      (g)
When values of a and p from equations (a) and (b) are substituted in
equation (g),  a relationship is obtained which expresses the concen-
tration of total soluble orthophosphate which is in equilibrium with the
solid phase at various pH levels.
                (8.26 x 10 ?)
                               [H+]3
                              K1K2K3
                [H
                   +
                                 [H
                                            + KCK,K
                                               5  6  w
                                                         -0.4
                                                                -0.6
                                                                      (h)

-------
Curve A in Figure 6 is a plot of equation (h) based on the values listed
in Table 11 for the various constants.  This curve indicates a pattern
of pH dependency similar to that observed in the orthophosphate pre-
cipitation experiments with Fe(II) (curve C).   In contrast to the experi-
mental  results,  however,  equation (h) predicts a higher level of resid-
ual orthophosphate.  Four possible explanations may be offered for this
discrepancy.  First,  the actual solubility constant for ferrous phos-
phate may be lower than value of  1. 3 x 10    used in the computation
of Curve A.  Second,  in the phosphate precipitation experiments some
of the soluble phosphate is  removed by adsorption on the precipitating
ferrous phosphate or hydroxide.  Third,  equilibria other than those
listed in Table 11 may also be involved in the precipitation system.
Fourth, the actual  equilibrium constants for  Fe(II) hydrolysis reac-
tions (reactions 4-6,  Table 11)  may be lower than those listed in
Table 11.  In the calculation of Curve A (Figure 6),  a value of K = 10
had been assumed for the first hydrolysis reaction of Fe(II) (reaction 4,
Table 11). Curve B (Figure 6)  is based on the data in Table 11, but
using a value of 5 x 10  (reported by Leussing and Kolthofr  ') for
reaction 4.  As  indicated in Figure 6, Curve C indicates, lower  levels
of residual orthophosphate  and the values in the vicinity of the pH of
minimum solubility are in fair agreement with the experimental
results.  Although neither Curve A nor Curve B coincides with  the
experimental results (perhaps largely due to the uncertainty regarding
the true values of the various equilibrium constants),  in terms  of the
general effect of pH on the  solubility of ferrous phosphate, the  experi-
mental  results are in reasonable  accord with theoretical considera-
tion.

Orthophosphate  precipitations with Fe(III) and Al(III)  are not amenable
to simple mathematical treatment due to the  complex and poorly under-
stood chemistries of the hydrolysis reactions of Al(III) and Fe(III) and
the uncertainty regarding the exact chemical nature of the precipitates
formed at different pH levels.  Although A1PO  and FePO  have been
suggested as the insoluble reaction products, the formation of pre-
cipitates of such composition is not consistent with the observed reac-
tion stoichiometry which indicates that even  under optimum pH condi-
tions hydrolysis products of Al(III)  and Fe(III) and not the simple Fe
and Al   ions are involved in orthophosphate precipitation^ ' (i. e., the
precipitates formed are hydroxy cation phosphates and not A1PO  or
FePO.).  Despite these complexities,  considerations  similar to those
presented above for the Fe(II)-orthophosphate  system may be used to
explain the pH-dependency of orthophosphate removal with both Fe(m)
and Al(III).  Thus with these cations,  as with Fe(II),  the optimum
orthophosphate  removal pH and the extent of orthophosphate removal

-------
         28
         24
         20
16
     "Si



     g

     h-
     111
     o
     z
     o
     o

     01

     <
     I
     O-
     c/)
     O
     X
     O_
     o
     I

     DC
     O
     <

     O
     2   8
     cr
     GO
12
                                    8 ,
                                                10
                                                             12
              FIGURE 6. SOLUBILITY DIAGRAM FOR FERROUS PHOSPHATE
CURVES A & B ARE CALCULATED DATA BASED ON EQUILIBRIUM DATA SHOWN IN TABLE

CURVE A IS FOR l<4= 1 x 107; CURVE B IS FOR l<4= 5 x 105

CURVE C PRESENTS THE RESULTS OF ACTUAL PRECIPITATION TESTS FROM FIGURE 5.

-------
at various pH levels are determined by the solubility product constants
for the insoluble reaction products and by the equilibrium constants
for orthophosphate ionization reactions and cation hydrolysis.  Solu-
bility product values of 1. 25 x lO'1^ 1Q-3&, and 10"32 which have been
reported for Fe(OH)  , Fe(OH)  and A1(OH)  , respectively, (12> 14) indi-
cate that ferrous hydroxide precipitates at a higher pH than ferric and
aluminum hydroxides.  For a 10~  M cation  concentration, for example,
the pH at which cation hydroxide precipitation commences are 8.0, 2.0
and 3.3 for  Fe(II),  Fe(IH) and Al(III), respectively.  The observed
higher pH of orthophosphate precipitation with Fe(II)  can  be explained
in terms of  a combination of two factors,  namely,  that Fe(II) hydro-
lyzes at a higher pH than Al(III) and  Fe(III) and that the phosphate pre-
cipitates formed with Fe(III) and Al(III) have lower solubilities than
ferrous phosphate.

Al(III) and Fe(IH) hydrolyze in solution to a considerable  extent.
Accordingly,  even under optimum pH conditions, their potential
capacity for phosphate removal is not fully realized.   Based on the
data in Figure 5, the maximum orthophosphate removal capacity for
Fe(III) and Al(III) are 0. 85 and 0. 71  moles per mole of cation pre-
cipitated, respectively.  Compared to Al(III) and Fe(III),  Fe(II) has a
considerably lower tendency to hydrolyze and on an equivalent basis is
more effective in precipitating orthophosphates (see Figure 5).

Orthophosphate Precipitation with Fe(II) in the Presence  of
Dissolved Oxygen

Wastewater treatment with ferrous iron under conditions where dis-
solved oxygen is totally  excluded from the system is  not practical
from the  standpoint of large scale application.  In the presence of
even a small amount of dissolved oxygen, oxidation of some of the
ferrous iron to ferric iron is inevitable.  To simulate actual condi-
tions in wastewater treatment, a number of experiments  were con-
ducted in which no precautions were taken to exclude dissolved oxygen
from the  system.  A discussion of the results of the experiments will
follow.

Effect of Dissolved Oxygen on the Rate  and Efficiency of
Orthophosphate Removal (Experiments  18-20).  Table 12 presents the
results  of an orthophosphate precipitation test (Experiment 18) with
ferrous iron in the presence of dissolved oxygen.   Comparison of the
results of this experiment with the results of Experiment 14 (Table 8)
conducted in the absence of dissolved oxygen indicates that the pres-
ence of dissolved oxygen results in a drop in pH, rapid removal of
orthophosphate and near quantitative precipitation of iron.  Compari-
son of the maximum orthophosphate  removal efficiencies in the two

-------
                     TABLE 12

ORTHOPHOSPHATE PRECIPITATION WITH Fe(II)
  IN THE PRESENCE OF DISSOLVED OXYGENa
                 (Experiment 18)
Time
0 min
5 min
15 min
30 min
1 hr
2 hr
3 hr
4 hr
5 hr
pH=

7.23
7.30
7.32
7.50
7. 78
7.95
8.08
8. 18
Percent Reactant
Removal
P
0
71.8
71.8
71.8
71.4
71.4
_d
-
69.9
Fe
0
-=100
-100
-
-
-100
-
-
 Initial orthophosphate concentration 12 mg/£  P;
 initial bicarbonate alkalinity 420 mg/A NaHCO •
 Fe(II)-to-orthophosphate equivalence ratio =1.0

 Time from start of Fe(II) addition
Q
 No supplementary pH adjustment made with NaOH;
 initial pH after addition of Fe(II) (time -  20 sec)  =  7.23
 - indicates no measurements made

-------
experiments (72% vs 90%) also indicates that under the conditions
tested the efficiency of orthophosphate removal with Fe(II) is reduced
in the presence of dissolved oxygen.   (It should be noted,  however,
that for reaction  times less than about 1 hr, the phosphate removal
efficiency is higher in the presence than in the absence of dissolved
oxygen.)

Table 13 contains data for two experiments (Experiments 19 and 20)
conducted on a secondary effluent (wastewater #2, Table 1) using
Fe(II)-to-orthophosphate equivalence  ratios of 0.5 and 1.5.  In both
experiments the  solution pH decreased gradually in the first hour and
then increased slightly during the second hour.  Comparison of the
data for Experiments 19 and 20 with those for Experiment 18 (Table  12)
indicates that while the initial pH of the reacted mixtures was not sig-
nificantly different in the three  experiments, the reactant removal was
slower in the  experiment with the waste-water.  It is interesting to note
that in Experiment 20  nearly all the excess ferrous iron added to the
wastewater precipitated after 2 hr and this is in contrast to the re-
sults  of Experiment 15 (Table 9,  conducted in the absence of dissolved
oxygen) which indicates an incomplete precipitation of the excess iron.

Some Fundamental Considerations Regarding Oxidation of Ferrous
Iron by Dissolved Oxygen and Its Effect on Phosphate Precipitation.
Dissolved oxygen in water oxidizes ferrous iron to ferric iron.   The
rate of oxidation  of ferrous iron by oxygen in a bicarbonate buffer sys-
tem has been  shown to follow the expression
where brackets denote concentrations in mole/^.      The value of k
at 20. 5°C in the bicarbonate buffer system is about 1 x lO1^ liter3 mole~3
min  .     From this equation it may be seen that the rate of oxidation
of ferrous iron is strongly dependent on solution pH and also on Fe(II)
and dissolved oxygen concentrations.   Thus in the presence of 0. 5 mg/jj
dissolved oxygen, the half-life of ferrous iron would be  7. 5 hr at
pH = 7.0 and 4. 5 min at pH 8.0. The  corresponding half -life values for
a dissolved oxygen concentration of 5 mg/Ji  would be  45 min and
0. 5 min, respectively.   The rate of oxidation of ferrous iron by dis-
solved oxygen has also been found to be affected by the presence of
diverse ions.  For example,  substantial increases in the oxidation rate
have been observed in the presence of orthophosphate,  pyrophosphate
and a number of other anions.^   ' In the aeration treatment of ground
waters containing ferrous iron,  the presence of dissolved organics has
been found to lower the efficiency of iron removal. ' •*•"'  No data are
available on the rate of oxidation of ferrous iron in the waste-water

-------
                              TABLE 13
ORTHOPHOSPHATE PRECIPITATION WITH Fe(H) FROM SECONDARY
      EFFLUENT IN THE PRESENCE OF DISSOLVED OXYGENa
Time
0 min
1 min
5 min
10 min
15 min
30 min
1 hr
1-1/2 hr
2hrh
Q
Experiment 19
PH6
7. 60
7.48
7.40
7.35
7.31
7. 30
7.30
7.35
7. 35
Percent Reactant
Removal
P
_g
-
-
-
53
-
-
58
Total Fe
-
-
-
-
94
-
-
-100
j
Experiment 20
PHe
7.60
7.41
7. 15
7. 10
7.09
7.08
7.06
Percent Reactant
Removal
P
_
-
-
-
96
-

7. 10 :
7.10 i -100
i
Total Fe
-
-
-
-
78
-
-
-100
 Wastewater #2,  Table 1


 Time after Fe(II) addition

Q
 Fe(II)-to-orthophosphate  equivalence ratio =0.5


 Fe(II)-to-orthophosphate  equivalence ratio =1.5

Q
 No  supplementary pH adjustment made with NaOH


 Because of the very fine nature of the precipitates formed, it took

 about 30 min for each aliquot to filter

or
 - indicates no measurements made


 The turbidity of the reacted mixture at the end of 2 hr was 35 JTU

 in Experiment 19 and 60 JTU in Experiment 20

-------
environment.  In wastewater treatment with ferrous iron in the pres-
ence of dissolved oxygen, the phosphate precipitation system becomes
complicated by the generation and subsequent hydrolysis of Fe(III) and
in this regard the system differs from phosphate precipitation  with pure
Fe(III) and with Fe(II) in the absence of dissolved oxygen.

In the absence of cation hydrolysis reactions, oxidation of ferrous iron
to ferric iron should result in an increase in the efficiency of phos-
phate removal, since on a molar basis the capacity of ferric iron for
phosphate removal should be 50% higher than that of ferrous iron.  How-
ever, because ferric iron hydrolyzes in solution to a greater extent
than ferrous  iron,  oxidation of ferrous iron may or may not result in
an increase in orthophosphate removal efficiency.   If no oxidation of
ferrous iron  had  occurred in Experiment 19, a maximum orthophos-
phate removal efficiency of only 50% would have been expected.  The
observed orthophosphate removal in this experiment, however, was
higher (58%)  thus suggesting a partial or total oxidation of ferrous iron.
In the absence of hydrolysis reactions, a phosphate removal efficiency
of 75% would have been expected from the quantitative oxidation of fer-
rous iron.  The observed phosphate  removal efficiency,  however, is
lower than this theoretical amount (58 vs 75%), thus suggesting either
partial oxidation  of ferrous iron with or without subsequent hydrolysis
of ferric iron,  or total oxidation of ferrous iron with subsequent
hydrolysis of ferric iron.  Even if total oxidation of ferrous iron is
assumed,  a phosphate removal  efficiency of  58% is higher than what
can be obtained at a pH of close to 7. 35 with an Fe(III)-to-phosphate
molar ratio of 0. 75 when ferric iron from a  stock solution is used for
phosphate precipitation.  (According to the data in Figure  5, the phos-
phate removal efficiency is 63% at pH 7. 35 for an Fe(III)-to-
orthophosphate molar ratio of 1.0.)  The observed increased efficiency
may be attributed to the in situ  generation of ferric iron which, accord-
ing to Singer and Stumm, ^ i7)  promotes   a more effective contact
between the iron  and the phosphate.

Singer and Stumm'   ' studied and compared the precipitation of ortho-
phosphate (at a 10    M level) from a bicarbonate solution by direct
addition of Fe(III) and by homogeneous generation  of Fe(III) through
in situ oxidation  of previously added Fe(II).  The experiments  were
conducted at pH values of 4. 5,  5. 0,   5. 5, 6. 0,  6. 5, and 7. 0 using iron
concentrations ranging from 10"  to 10    M.  At lower pH levels
ozone was continuously bubbled through the solution to assure  quanti-
tative oxidation of ferrous iron. With the exception of the results at
pH 7. 0, the degree of removal  was  generally enhanced by the utilization
of the homogeneous precipitation technique.  The results  of the experi-
ments at pH  7. 0, however, are questionable since a possibility exists
                               50

-------
that the precipitates formed at this pH with homogeneously generated
Fe(III) were too fine to be quantitatively retained by the 1. 2 p, mem-
brane filters employed for sample filtration.  When Fe(III) is generated
homogeneously within the solution,  it is uniformly distributed through-
out the  system thereby promoting a direct contact between the iron and
the phosphate.  On the other hand, when ferric  iron is added from an
external source, due to the formation of "short-lived" regions of high
Fe(III) concentration,  some of the Fe(III) may be coprecipitated
through hydrolysis.  The precipitation of phosphates from wastewater
by homogeneously generated ferric iron has not been investigated in
detail and the parameters which influence process efficiency are not
defined.

Formation and Settling  Characteristics of the Precipitates in
Wastewater Treatment with Fe(II) in the Presence of Dissolved
Oxygen (Experiments  21-26).  In  any application of the chemical pre-
cipitation to the treatment of wastewater for phosphate removal,  the
properties of the precipitates  formed are of major engineering con-
cern.  Table 14 presents data on the effect of Fe(II) dosage on the
settleability of the precipitates formed  in the treatment of  secondary
effluent with Fe(II) in  the presence of dissolved oxygen.  Although the
actual concentration of dissolved oxygen in these experiments was not
determined,  with the type of sample pretreatment and mixing employed,
it would be  reasonable to assume that the test solutions in all cases
were nearly saturated with dissolved oxygen (~8 mg/£ dissolved oxygen).
Immediately upon addition of Fe(II),  all samples developed a milky-
yellow color.   By the  end of 2 min of rapid mixing,  floe formation was
noted in all samples and the samples were all yellow in color. By the
end of 30 min of settling, a significant amount of a yellowish precipi-
tate had settled to the bottom in each jar.  After 1-1/4 hr of settling,
the solution in Experiment 21  was yellowish gray and the solutions in
Experiments 22-26 were reddish-brown,  with the intensity of the  color
increasing with the increase in Fe(II) dosage.

The turbidity data presented in Table 14 indicate that under the condi-
tions  of the experiments (a) the higher the Fe(II) dosage, the greater is
the amount of turbidity formed, (b) with the exception of the sample
with an Fe(II)-to-orthophosphate  equivalence ratio of 0.4, there was an
increase in turbidity after  15 min, thus  suggesting the formation of
additional precipitate,  and (c) the precipitates formed with Fe(II)  do
not settle well and,  even after standing overnight,  all samples con-
tained substantial amounts of turbidity.

At the end of 2 hr of settling,  a portion of the wastewater from each
experiment was filtered through Whatman #42 filter paper.  The filtrates
from  Experiments  21  and 22 appeared colorless, whereas those from

-------
CO  ^
W  W
H  ^


5  fa  W
&  H  O
^  Q  Q
fa  O  >

^wo
y  co  co
•si*
1-1
TABLE
05 cd fa
H fa O
J CHARACl
PITATION
'RESENCE

3s After Indicated Periods
lingc (JTU)
P-> "CD
a w
o
•H
T3
•iH
H
o
d
CM
^
fl
a
^
«
a

«
•M
a
CM
^
OO CD
fa ^ .2
1— 1 •'-' rv>
R ^ K
faW
•
W !Z
ro r- CM t*- m r-
CM r— o oo CM CM

CM o in r«- o
vO OO vO 00 CM 1
1— 1 r— 1 r— 1 <\]
m m o o o
\o CM m [~~ i o
i— I i— I i— I (M
o o r^ m o
Is- CM CM oo i in

m f>- CM o m
o •— i oo oo oo i
i — i i — i i— t i — i i — t

o in o m o
CM Is™ CT^ f^" O^ rO
i — 1 i — I i — I i — I i — i i
o oo in vo oo o
rO O 00 Is- sO O

-vt1 oo o CM m o
O O ^H i-H ^ CM
I~H CM oo "^ in ^0
CM CM CM CM CM CM









^

cd
a
CO
• r-<
to
_— 1
Turbidity ana.
CJ














water surface

cd
a
co
*•<
>

CO

fl
0)
a
^
CO
cd
0)
a
- indicates no
Td

-------
samples 23-26 were reddish-brown.  On standing overnight, all fil-
trates developed turbidity which could be only partially removed by
filtration through Whatman #42 paper.  The amount of  turbidity was
especially  significant in samples 24-26. These filtrates were refil-
tered through Whatman #42 filter papers.  A portion  of each new fil-
trate was subsequently filtered through  a 100 mp, membrane filter.
The new Whatman #42 filtrates and the  100 mp,  membrane filtrates
were analyzed for turbidity, phosphate  and total iron content.  The
results are presented in Table 15.

The data in Table 15 indicate that some of the precipitates formed in
the treatment of  secondary effluent with Fe(II) are very fine and cannot
be totally removed even by very fine filter papers (compare,  for
example, the turbidity data for 100 mp  and Whatman #42 filtrates  in
Table 15).   Also, at the pH levels tested and for the Fe(II) concentra-
tions used,  iron  precipitation  appears to be slow and incomplete
(especially  at low pH and high Fe(II) concentration).  In the present
study, no data were collected  on the effect of pH and  level of dissolved
oxygen on the rate of formation and settling properties of the pre-
cipitates.  Since the rate of oxidation and hence precipitation of fer-
rous iron would depend on both pH and dissolved oxygen concentration,
addition of a base and/or extended aeration of the wastewater would be
necessary for the removal of excess ferrous iron.  Thus, in second-
ary treatment plants employing the activated sludge process, best
results should be obtained when ferrous iron is  added to the waste-
water  ahead of the primary or activated sludge tanks.  Aeration of
the wastewater in the activated sludge tank should result in effective
oxidation of ferrous iron.  Also, the activated sludge floes  should  aid
in the  agglomeration and subsequent removal of the phosphate pre-
cipitates by settling.   When only primary treatment is  employed, the
addition of a base would be necessary for the precipitation of excess
ferrous iron.  These conclusions are consistent with the results
obtained at Mentor,  Ohio,  and at Texas City,  Texas,  wastewater treat-
ment plants where acid pickling wastes  containing 6 to 9% iron (mainly
ferrous but probably also  containing some ferric iron) were used for
the treatment of  raw sewage. *• '

In the  Texas City operation where an iron dosage of 19 mg/^ was used
for the treatment of wastewater containing 5. 8 mg/A P phosphate
(an iron-to-phosphate  molar ratio of 1.  75),  the  phosphate removal
with iron in primary treatment was less than 1%.   The bulk of the
phosphate was removed in the aerator and final  settler,  yielding an
overall phosphate removal of 81% with an effluent phosphate concen-
tration of 1. 1 mg/l  P.  Although it was not reported, the low phos-
phate removal efficiency in primary treatment may have been due  to
                              53

-------
m
rH


w

HH

ffl
   EH



   H


   r^  '



   to W
     Hp"^

     O
oS1^
U rH W
w o H
w w <;

22e
0Q3
Cd h rH


^ H W
   OH Cd W

   o^S

   SES
   CG i-Ln -f(
   OnEng

   w a <
   EH 1~l
   < rH~
   E •—
   0, r^
   CO rH
   O M
   o
   ffi
   H
o
CO
-j \
cd
•P
i— i
•rH
to
CO
cd
rQ

CO
)5>

a
o
o
rH
'H
O
CO
CO

rH
cd
fl
^
CO
cu
-p
cd
to
-p
rH
•rH
to
CM
^3
nJ
£
cd
rS
^
rt
o
o
co
*o
CO
•r-l
CO

rH
cd
fl
^
oo
1 ,
o
£l
£7
t— (
(p


[
fl ^
O r
to to
i— I
3 ~M
EH J.


co
-p
cd
UST
CO "
o .,
a*]?
^S .£•
to
0


4_>

""0 £D
3 EH
H 1-2
EH ~~"
C! 'M"
O f
to "
*~* ^
rH -«-I^
cd bo
£ 5

CO

"cd
rC _
Q-i A
CO MH
!!
rH ~~"
O

-p
• r-l >, 	 „

•H r-^
rQ H

P1 . 	 .
EH
CO

§
"~d T-*
CU -+J
.!H -^
^J I-M
D^
H
^o
^ £

in in m m m ro
i — 1 < — 1 i — 1 rH 1 — I OO
o o o o o o
V V V V V






r*1- 0s1 oo "^ r** o^
oo o o o m oo
• ••«••
\O rH rH CM rH O







r~- CM to PO vo CM
O O 0 O O —1



m oo CM ^0 oo "^
rH OO ^* ^t^ OO rH

O O O O O OO
V





in CM rH o oo 0s
CM O OO rH -^ f~






vO ^o m \o rH r^*
• •••••
co ^f in r^ ^H oo





•sh oo o CM m o
O O rH rH i— 1 CM



rH CM OO Tf m vD
CM CM CM CM CM CM

































rH
co
I— I
rQ
cd
CM"
=fe
to
CD
Wastewat
cd
-t-1
CO
6
CO
£
i— i
-p
bo

ing overn:


-p
CD
o
o
fl
o
•rH
"3
"o
•J— *
T)
3
CO
Id
CU
CO
o

CU
o
£
to
O
IT
• T-l
i-Q
j^
5
rH
O
^H
rO
CD
N
cd
cd
CO
CD
-p
rd
-P
rH
•rH
co
ti
-p

a
OS
rH
a
cd
CU

nalyzed for turbidi
cd
T)
cd
CO
CO
-p
rH
irH
CO
ti
cd
^
1
CD
a
a"

o
o
rH
bo
P
O
to
-p
CO
rH
CO
-p
,J—j
•H
co

co
^
CO
co
"cd
to
rH
i^
CM
•*
=*=
Whatman
rrt

O
O
CO
CO
CD
EH
u


























+J
pj
CO
o
0
pi
0
to
•H
i— H
cd
-P
o
-p
cd
CD
-p
cd
rS

orthophos


-------
any, a combination,  or all of the following factors: (a) formation of
fine and non-settleable precipitates,  (b) inadequate reaction time, and
(c)  formation of soluble organic-iron complexes.  At Mentor,  Ohio,
for an average influent phosphate concentration of 15.7 mg/^ P a
phosphate removal efficiency of 46% was obtained in the primary  treat-
ment with an iron dosage of 50 mg/l  (an iron-to-phosphate molar ratio
of 1. 71).  When lime (55 mg/jg) was used in conjunction with iron
(43 mg/ji), the phosphate removal efficiency was increased to 76%.
The primary removal efficiency was  raised to 83. 5% by using 0. 4 mg/A
of anionic polymer in conjunction with 40 mg/1 of iron and 66 mg/j£ of
lime.  The function of base (lime or sodium hydroxide) in improving
the removal  efficiency is not clearly  understood, although it is asserted
to be one of floe conditioning and not  that of pH effect per  se.^  '  It is
possible that the improvement  in flocculation is  a result of a change in
particle surface charge, presumably brought about by the addition of
base.  The effect  of pH on  particle charge and the role of particle
charge in the agglomeration of the floes have not been investigated
for phosphate removal by addition of  Fe(II).

Settling Properties of the Precipitates Formed in Phosphate
Precipitation Using Al(III)  and  Fe(III)

To  compare  the settling properties of the precipitates formed in the
treatment of secondary effluent with Fe(II) with those of the precipi-
tates formed with Fe(III) and Al(III) salts, experiments were conducted
to evaluate the settling rates for the Fe(III)- and Al(III)-orthophosphate
systems.   The results are presented in Tables 16 and 17.

Comparison  of the turbidity data presented in Tables 16  and 17 with
the turbidity data  for the experiments using Fe(II) (Table 14),  indicates
several interesting points.  First,  in all experiments with Al(III) and
Fe(III) the turbidity decreased  with the  increase in the settling time.
This is in sharp contrast to the Fe(II) experiments where there was
generally an increase in turbidity after  15 min.   Second,  in general
the higher the Al(III) and Fe(III) dosages the lower was the settled solu-
tion turbidity.  This  is  also in contrast to the Fe(II) case where the
higher the Fe(II) dosage, the greater was the turbidity of the settled
solution.  Third,  at all ratios of Al(III)- and Fe(lII)-to-orthophosphate
equivalence ratios tested the settling was nearly  complete after 30  min.
With Fe(II),  on the other hand,  all samples contained substantial
amounts of turbidity  even after standing overnight.  The orthophosphate
removal data for Al(III) and Fe(III)  indicate that  although the removal
by settling was higher for  samples with lower residual turbidity,  the
phosphate concentration in an aliquot of settled sample was not directly
                                 55

-------
W

pq
       W
       w
       ft  a
       H
       cd
       co J";
       W i2
       H W
       I—I [tj

       u w
H |


h8
O co
CO ^
U
           O
           oJ
       cd

       EH

       EH
       H
       H
       CO
T£
"S CD QJO
!"^ 'II £4
CQ ^ S
O i— I -l->
,£5 Cd CD
A > CO
0 ° m
5 a ^
!H CD s.
0 Cd ^
(M
CQ
T)
O
•H
CD
ft
T)
CD
•4-1
cd
O
rt S^


§\n
a w
eg "8


O
• 1-1
3
H

ffi
cx

^ ^
B-"
H 2
.tered
•r-i
fa


Ti
CD
LJ
Unfilte:


(^
T^l
^



M
i-^H
i— I

fl
a
o
a
a

m
i— i
C!
a
LO





D
H
_j
TJ
cj



CM o ro o m m







CO t^" O^ ^JH i-H i — 1
CM ^ m vD 00 O



o "^ "vO r^- co ' — '
m ro co co CM CM



i— H so o^ r^ \o co
•
m co co co CM CM


•- CM o r^ r--
m co ^t1 ^ CM CM

o^ ^ r^ t^- CM oo
• *•••«
^O ^4^ N^ ^^ CO CO



o ^ co CM oo m

OO CO OO O^ O^ O
^ CM i— I O O O^
t^- r^- c — t^ t^- \o


•^ oo o CM m o
O O i— I i— I '— < (M

t^ CO CJ^ O ' — ( CM
CM CM CM CO CO CO

                                                                            0)
                                                                           1—I

                                                                           rQ
                                                                            rt

                                                                           H
                                                                    =*=
                                                                     t-l


                                                                     "cd

                                                                     CD

                                                                     CO
                                                                     cd


                                                                   f
                                                                            bfl
                                                                            fl
                                                                           CM


                                                                            CO
                                                                                   CQ
                                                                                   rd
                                                                                   CD
                                                                                   a
                                                                                          CD
                                                                                          O
                                                                                          cd
                                                                                          CO


                                                                                          CD
                                                                                   CD
                                                                                  rQ



                                                                                  •S

                                                                                  un

                                                                                  CM



                                                                                   O

                                                                                   rd
 O

 a
 CD


 CQ
4->
 O

 cr
 CD
r—I
 O-

 a
 cd
 co

 fl
 O

 CD
TJ


 a
 CQ

'CQ
                                                                                          cd
                                                        56

-------
r-



W


CQ
       H
       PH
       CO

       Ort
       o ir


       cd W
       co Ej
       W ^
       H W
U  W

W  SH
w  I


^8

O  co
co  5

y  o
H  cd

M  "

W  §
H  |ri
          W


          D.
       H

       W
       CO


5?


rt >
Bg/
fro
H Z
Ti
o



Tt1 oo o CM m o
O O i-H i— I •— 1 CM
co ^t* in so r^* oo
CO co CO CO CO CO
                                                                   (U
                                                                  r—<


                                                                   nJ
                                                                  m
                                                                  =*=


                                                                   
                                                                                        U
                                                                                        ni
                                                                                 nJ



                                                                                 (U





                                                                                 O
                                                                                i—i


                                                                                43



                                                                                 a
                                                                                •H

                                                                                m


                                                                                CM
                                                                                       O

                                                                                       43
                                                                                       rt
                                                                                       O
                                                                                       a
                                                                                       03
                                                                                       4->
                                                                                       O


                                                                                       cr
                                                                                       a.

                                                                                       a
                                                                                       nJ
                                                                                       CO

                                                                                       ti
                                                                                       O
 nJ

 a

 to
-H
 to
 cd




•H

Tj


43

-------
proportional to its turbidity content.  In all cases the degree of phos-
phate removal achieved by plain sedimentation and by filtration in-
creased with the Al(III) and Fe(III) dosages used.  For the lowest and
highest Al(III) doses, phosphate removals were 32 and 95% for the fil-
trates and 23 and 91% for the unfiltered samples.   The corresponding
phosphate  removal values for Fe(III) were 28 and  93% for the filtrates
and 16 and 90% for unfiltered samples, respectively.

In summary,  in the precipitation of phosphates from secondary efflu-
ent,  the settling properties of the precipitates formed with Al(III) and
Fe(III) are superior to those formed with Fe(II).   Thus,  while plain
sedimentation is generally adequate for the removal of the Fe(III)-
and Al(III)-phosphate precipitates, adjustment  of pH, addition of poly-
electrolyte or process modification would be required for effective
removal of the precipitates formed with Fe(II).  However, as was dis-
cussed before, the poor settling behavior of the precipitates formed
with Fe(II) should not be a problem when ferrous iron is added to the
wastewater ahead of activated sludge.

Advantages of the Use of  Ferrous Iron for Phosphate Precipitation

The use of ferrous iron for phosphate precipitation from wastewater
has potentially two major advantages over other precipitation tech-
niques. First, because of the existence of oxidative conditions in con-
ventional biological wastewater treatment, ferrous iron added to  the
wastewater is oxidized in situ resulting in the homogeneous  generation
of ferric iron. Homogeneously generated ferric iron is a. more effi-
cient phosphate precipitant than ferric iron added from an external
source.  Second,  in some metropolitan areas (e.g.,  those in the
Great Lakes region) waste pickle liquor from local steel industries
may provide an inexpensive source of ferrous iron.  The cost of acid
pickle liquor is generally determined by the transportation cost which
is in turn dependent on the hauling distance. The chemical costs of
wastewater treatment with pickle liquor at Texas  City,  Texas,  and at
Mentor, Ohio, have been estimated at 1. 5 and  2^/lb of iron (0. 18<£  and
0. 25$ /gm-mole of Fe), respectively.^ '  In comparison, the cost of
alum is about 24£ /lb of Al  or 1. 43^/gm-mole of Al.(2)  Although  in
recent years  a number of large-scale experiments have been conducted
to assess the effectiveness of Fe(II) for phosphate precipitation, because
of the lack of a basic understanding of the  chemistry involved, a trial-
and-error approach has been used in these evaluation tests.  Accord-
ingly,  the  results of these  experiments are not believed to represent
the optimum utilization of ferrous iron, although they nevertheless
point out the economic  attractiveness of wastewater treatment with

-------
Fe(II).  For example, at Grayling and Lake Odessa, Michigan,  treat-
ment plants where phosphate was precipitated by chemical addition in
the primary (34 to 57 mg/£ Fed ,  12 to 14 mg/4  NaOH,  0.3 to
0. 5 mg/,6 anionic  polymer),  totaioperating costs are estimated to
range from a possible low of $0. 01/1000 gal.  to a high of $0. 05/
1000  gal.^ '  The  chemical costs for Mentor,  Ohio,  utilizing waste
pickle liquor,  are estimated to be $0.023/1000 gal.  with an  additional
cost of $0.01/1000 gal.  for handling the excess  sludge produced by
chemical precipitation.'^/ In comparison, for a wastewater containing
10 mg/A  P phosphate, the total costs for 90% phosphate removal by
addition  of aluminum in the primary stage  are estimated at $0.036/
1000  gal.(2)

Section II of this report contains recommendations for a study which
will seek to define the optimum conditions  for effective utilization of
ferrous iron in wastewater treatment. Unless such a  study  is under-
taken, the economics of the  phosphate precipitation by addition of fer-
rous  iron cannot be  accurately defined.

-------
                         SECTION VI

                     ACKNOWLEDGMENTS
The support of this project by the Water Quality Office, Environ-
mental Protection Agency, is acknowledged with gratitude.  Sincere
thanks are due to Dr. Sidney A. Hannah,  the Project Officer,  and
Mr. Jesse M. Cohen,  Chief, Physical and Chemical Treatment
Research,  both of the Cincinnati Advanced Waste Treatment Research
Laboratory, for the interest,  expertise and guidance provided during
the course of this  investigation.

The help provided by the director and staff of the Las Virgenes
Municipal Water District, Calabasas, California, in providing waste-
water samples is gratefully acknowleded.

At Atomics International, Drs.  Masood Ghassemi and Howard L.
Recht were, respectively,  the principal investigator and program
manager for this project.  Most of the laboratory work was performed
by Miss Nancy M. Trahey.  Special thanks are due to Drs. Donald E.
McKenzie, Eugene V.  Klebar and Samuel J.  Yosim for their  critical
review of the manuscript and for their numerous helpful  suggestions
during the  course  of this work.  Finally,  special thanks are due to
Mrs.  Evelyn Cress for typing the final manuscript and for her numer-
ous other secretarial assistance.
                               ol

-------
                           SECTION  VII

                          REFERENCES
1.    Nesbitt,  J.  B. ,  "Removal of Phosphorus from Municipal Sewage
      Plant Effluents, " Engineering Research Bulletin B-93,  Pennsyl-
      vania State University,  University Park,  Pa. (1966).

2.    Convery, J. J. , "The Use of Physical-Chemical Treatment Tech-
      niques for the Removal of Phosphorus from Municipal Waste-
      waters, " Advanced Waste Treatment Seminar,  jointly sponsored
      by California State  Water Resources Control Board and Pacific
      Southwest Regional Office of FWQA,  San Francisco, Calif.,
      Oct.  28  and 29,  1970.

3.    Stumm,  W.,  Discussion in  "Advances in Water Pollution Control
      Research," Proc.  1st Intl.  Conf.  Water Poll. Res.,  Pergamon
      Press Ltd., London, England, Vol. 2, 216(1964).

4.    Recht, H. L.  and Ghassemi,  M. , "Kinetics and Mechanism of
      Precipitation and Nature of the Precipitate Obtained in Phosphate
      Removal from. Wastewater Using Aluminum(III) and Iron(III) Salts, "
      Water Pollution Control Research Series 17010EKI 04/70, U.  S.
      Dept. of the Interior, FWQA (1970)

5.    Cohen, J. M. ,  "Improved Jar Test Procedure,"  J. Am.  Water
      Works Assoc., 49.  p 1425(1966).

6.    Manual on Industrial Water and Industrial Waste Water,  2nd Ed.,
      ASTM Special Publication No.  148-1 (1966).

7.    Instruction Manual  for Titrator Type TTT11, Radiometer Co.,
      Copenhagen,  Denmark.

8.    Hush, N. S. , "Intervalence-Transfer Absorption.  Part 2.   Theo-
      retical Considerations and Spectroscopic Data, " in Progress in
      Organic Chemistry, F. A.  Cotton, Ed.,  Vol.  8,  Interscience
      Publishers, New York (1967)

9.    Singer,  P.  C. , "Anaerobic Control of Phosphates by Ferrous  Iron,"
      Paper presented at the 43rd Annual Conference, Water  Pollution
      Control  Federation, Boston, Mass., Oct.  1970.

-------
10.   Sillen,  L. G. , and Martell, A. E. , "Stability Constants of Metal-
      Ion Complexes," Spec. Publ. No.  17,  The Chemical Society,
      London, Burlington House (1964)

11.   Van Wazer,  J. R.,  "Phosphorus and Its Compounds, " Interscience
      Publishers,  New York (1958).

12.   Leussing, D. L. , and Kolthoff, I.  M., "The Solubility Product of
      Ferrous Hydroxide  and lonization of the Aquo-Ferrous Ion, "
      J.  Am. Chem. Soc. ,  75, p 2376 (1953).

13.   Gayer, K. H. , and  Woentner, L., " The Solubility of Ferrous
      Hydroxide and Ferric Hydroxide in Acidic and Basic Media at  25 , "
      J.  Phys.  Chem.,  60, p 1569 (1956).

14.   Galal-Gorchev, H., and Stumm, W. , "The Reaction of Ferric Iron
      with Orthophosphate, " Inorg. Nucl. Chem.,  25, p 567 (1963).

15.   Stumm,  W.,  and Lee, F.,  "Oxygenation of Ferrous Iron," Ind.  and
      Eng. Chem. , 53,  p  143(1961).

16.   Robinson, L. R. , "The Effect of Organic Materials on Iron
      Removal from Ground Water, " Water and Sewage Works,  114,
      p 337(1967)

17.   Singer, P.  C. , and Stumm, W., "Oxygenation of Ferrous Iron, "
      Water Pollution Control Research  Series DAST-28,  U.  S. Dept.  of
      the Interior,  FWQA(1969).

18.   Wukasch, R. D. ,  "The Dow Process for Phosphorus Removal, "
      FWQA Phosphorus Removal Symposium,  Chicago,  111.,  June 1968.

-------
 SELECTED WA TER
 RESOURCES ABSTRACTS
 INPUT TRANSACTION FORM
                                                                  3. Accession No.
                                                                    w
   4. Title

    PHOSPHATE PRECIPITATION  WITH FERROUS IRON
 7. Author^)
          Ghassemi,  M.,  and Recht, H. L.
   9. Organization
          Atomics International
          A Division  of North American Rockwell Corporation
          Canoga Park, California

  12. Sponsoring Organization                                   :*',

  15. Supplementary Notes
                                                                 4.
                                                                 e-
                                                                 t.
                                                                   10. Project No.

                                                                      17010 EKI
                                                                 11. Contract!Grant No.
                                                                 fa>•
16. Abstract
     Phosphate  removal by ferrous iron addition was  studied in batch precipitation
tests conducted on pure solutions of ortho-, pyro- and tripolyphosphate, orthophosphate
solutions containing bicarbonate ion, and secondary  effluent.  The effects  of pH and
reactant concentration on the efficiency of reactant removal were evaluated.
     In the absence  of dissolved oxygen and for initial conditions of 12 mg/1 P and a
reactant equivalence ratio of 1.0, orthophosphate removal was maximum (97%) at a pH of
8.0.  The precipitate formed was identified as vivianite, Fe3(P04)2- 8H20.  At this
pH and lower pH levels, Fe(II) removal nearly paralleled that of orthophosphate removal.
The time for maximum orthophosphate removal increased with decreasing pH.   The reaction
speed was independent of Fe(II)  concentration but decreased at lower orthophosphate
levels.  The system  behavior was the same in secondary effluent as in pure  solutions.
Pyro- and tripolyphosphates were less efficiently precipitated than orthophosphate.
Dissolved oxygen increased orthophosphate removal efficiency.  However, the precipitates
obtained in the treatment of oxygen-containing secondary effluent were usually very
fine and did not settle well.  The data on Fe(II)-phosphate precipitation are compared
with those of Fe(III)- and Al(III)-phosphate systems.
17a. Descriptors
*Phosphates,  *Chemical precipitation, *Waste water treatment, Eutrophication,
Water pollution  control
17b. Identifiers

*Ferrous phosphate,  *Phosphate precipitation,  Ferrous salts, Reaction rate
17c. COWRR Field & Group
  18. Availability
                        19,  Security Class.
                           (Report)

                        30.  Security Class.
Abstractor
21. No. of
   Pages
Send To:
                                              I>f-lV«
                                              ^ i ittW
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C. 20240
                                       Institution
WRSIC 102(REV JUNE1971)
                                                                                GPO 913. 2g I

-------