WATER POLLUTION CONTROL RESEARCH SERIES
17O1ODJA11/7O
INVESTIGATION
OF
A NEW PHOSPHATE REMOVAL PROCESS
ENVIRONMENTAL PROTECTION AGENCY - WATER QUALITY OFFICE
-------
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, develop-
ment, and demonstration activities in the Water Quality
Office, Environmental Protection Agency, through inhouse
research and 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 Head, Project Reports
System, Office of Research and Development, Water Quality
Office, Environmental Protection Agency, Room 1108,
Washington, B.C. 20242.
-------
INVESTIGATION OF A NEW PHOSPHATE REMOVAL PROCESS
by
Envirogenics Company
A Division of Aerojet-General Corporation
El Monte, California 91734
for the
WATER QUALITY OFFICE
ENVIRONMENTAL PROTECTION AGENCY
Project #17010 DJA
Contract #14-12-487
November, 1970
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price 75 cents
-------
EPA Review Notice
This report has been reviewed by the Water
Quality Office, EPA, 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 recommendation for
use.
-------
ABS TRAC T
A laboratory evaluation was made of the technical and economic feasi-
bility of phosphate removal from secondary effluent by two types of new
sorption resins.
The phosphorus capacities of the first type —which included (1) poly-
hydroxylic esters and amides of various cation-exchange resins (RSO,-
R'-(OH)x and RSO2N(CH3)-R'-(OH)x), and (2) polyhydroxylic ethers of
styrene-divinyl-benzene copolymers (RCJ^O-R'-fOH) — were small,
only 25 to 30% of the commercial polyol resin, Amberlite XE-243.
Various attempts to increase the capacity of these resins by (1) de-
creasing the amount of their residual acid functions, (2) increasing
their porosity, or (3) decreasing their density, had only minor bene-
ficial effects.
The second type, which consisted of multivalent metal derivatives of
sulfonic acid resins, had much higher phosphorus capacities. One such
sorbent, the iron (III) form of Amberlite 200 strong acid cation exchange
resin, exhibited phosphorus capacities up to 9.5 mg P/ml of wet resin.
When this sorbent was evaluated with secondary effluent, a phosphorus
capacity of 2. 75 mg P/ml of wet resin was obtained.
Regeneration of the exhausted resin was readily accomplished with
dilute ferric chloride solutions. On one sample a total of ten exhaus-
tion-regeneration cycles were performed without loss of phosphorus
capacity. Iron and phosphorus were quantitatively removed from spent
regenerant solutions by the addition of a slight excess of lime and re-
moval of the resulting precipitate by filtration. On the basis of the
laboratory results, a phosphorus removal process was designed, and
an order-of-magnitude estimate of the cost of phosphorus removal was
obtained.
This report was submitted in fulfillment of Project Number 17010 DJA,
Contract 14-12-487, under the sponsorship of the Water Quality Office,
Environmental Protection Agency.
111
-------
CONTENTS
Section Page
ABSTRACT . iii
I. CONCLUSIONS AND RECOMMENDATIONS ...... 1
II. INTRODUCTION 3
III. RESULTS AND DISCUSSION 11
Sorbent Synthesis 11
Sorbent Evaluation 26
Economic Analysis ................... 50
IV. EXPERIMENTAL PROCEDURES 61
Sorbent Synthesis ................... 6l
Sorbent Evaluation 66
V. ACKNOWLEDGEMENTS 67
VI. REFERENCES 69
VII. GLOSSARY ....................... 71
VIII. APPENDIX ....................... 73
-------
FIGURES
Figure Page
1. Swelling of RSO-C1 in Various Liquids 16
2. Flow Diagram - Phosphate Removal from Secondary
Effluents Using Cation Exchanger/Metal Salt Sorbent 53
Vll
-------
TABLES
Table Page
1.
2.
3
4
5
6
7
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Comparison of Alternative Phosphorus Removal
Ultimate Contaminant Disposal Methods and Costs . .
Preparation of Resin Sulfonyl Chloride
Preparation of Resin Carbonyl Chloride
Elemental Analysis of Resin Polyol Sulfonates ....
Capacity of XE-243 for Dihydrogenphosphate and
Phosphorus Capacity of Base -Treated Polyol Resins .
Swelling of Various Resins in Water „
Phosphorus Capacity of High-Porosity Polyol Resins .
Phosphorus Capacity of Resin/Metal Sorbents ....
Phosphorus Capacity of A200-Fe Sorbents
Phosphorus Capacities of Sorbents Prepared from
A200 and Various Multivalent Metal Ions
Effect of Method of Preparation on Phosphorus
Capacity of A200-Fe Sorbents
Evaluation of Column-Prepared A200/Fe Sorbents by
Test Procedure 3 . .
4
7
11
1?,
14
15
18
19
21
23
28
29
30
32
34
35
37
39
40
41
4-3
V1U
-------
TABLES - Continued
Table Page
22. Total Phosphorus Capacity of A200/Fe Sorbents ... 44
23. Regeneration of A200/Fe . . . . . . . . ........ 46
24. Regeneration of A200/Fe by the Column Process ... 47
25. Phosphorus Capacity of A200/Fe Sorbents on
Secondary Effluent . „ 48
26. Phosphorus Removal from Secondary Effluent by
A200/Fe Sorbent .................... 49
27. Design Conditions for Phosphate Removal from
Secondary Effluent Using Cation Exchanger/Metal
Salt Sorbent 51
28. Capital Cost Estimate 54
29. Annual Cost Basis ................... 56-57
30. Working Capital 58
31. Cost Summary 59
-------
SECTION I
CONCLUSIONS AND RECOMMENDATIONS
Of the two broad categories of potential phosphate sorbents investigated
in this program, the polyol resins were found to have phosphorus capa-
cities too small to be of practical value; the resin/metal sorbents, on
the other hand, had both the capacity and regenerability characteristics
required for the attainment of a practical phosphorus removal process.
Based on the laboratory data obtained during the course of this pro-
gram, a phosphorus removal process can be envisioned consisting of
the following steps. Amberlite 200 resin is converted to the iron (III)
form by treatment with a 1.0 M ferric chloride solution. Secondary
effluent is processed through a stationary bed of this resin on a sche-
dule requiring daily regeneration. Regeneration is accomplished with
ferric chloride solution, and backwashing and rinsing utilize stored
product water. Waste regenerant is treated with an equivalent amount
of lime and returned to the secondary settling tank for removal of the
precipitated iron and phosphorus. Using the laboratory data on hand,
a cost of 20£/1000 gal. of water is projected.
This cost could be reduced in several ways. Additional laboratory
study should lead to higher phosphorus capacities and reduced rege-
nerant levels. Higher capacities would reduce the capital investment
required, by reducing the amount of resin necessary, while lower
regenerant levels would reduce the chemical cost.
The above process is based upon removal of phosphorus from secon-
dary effluent containing an average phosphorus content of 10 mg/1, with
a product water to contain not more than 0. 32 mg P/l. If the process
is used for polishing effluent from which a portion of the phosphorus
has been removed, a lower cost may result. Such a system could also
take advantage of the low phosphorus content ( 0.07 mg/1) of the pro-
duct from the sorbent over the first 90% of its capacity.
It is recommended that additional data be obtained regarding the econo-
mic aspects of this phosphorus removal proce s. Such data should in-
clude further attempts to improve phosphorus capacitv and regenerant
levels, as well as studies on the effects of bed depth and flow rates. It
is anticipated that such studies would be followed by a pilot scale evalua-
tion to determine the economic feasibility of this promising method for
phosphorus removal.
-------
SECTION II
INTRODUCTION
At the present time the cause(s) of accelerated eutrophication and the
most effective procedures for its control are being debated. Many —
probably the majority — believe that inorganic nutrients (nitrogen and
phosphorus) are responsible, and that the removal of phosphate from
wastewater is the most practical means of controlling nuisance algal
blooms and other growths in aquatic areas. Others maintain that (1)
decomposable organic matter is primarily to blame, (2) carbon dioxide
is the actual limiting nutrient, and (3) attempts to reduce massive algal
blooms by limiting phosphorus in wastewater effluents will not be very
effective unless equal emphasis is placed on removing biodegradable
organic matter (Reference 1).
Although consensus is lacking as to causes and methods of control,
there is virtually no question as to the seriousness of the problem and
the genuine need for its solution. The project described in this report
was undertaken to develop a more efficient phosphorus removal pro-
cess based on new kinds of sorbent resins.
PHOSPHORUS REMOVAL PROCESSES
Although a variety of processes have been applied to phosphorus re-
moval (including ion exchange, electrochemical treatment, electro-
dialysis, and reverse osmosis), the most promising appear to be
certain advanced biological treatment methods, chemical treatment,
or a combination of the two (Table 1). A third approach, which has
been much less studied, is sorption (Reference 2).
Biological Treatment
Biological treatment processes that have been used to remove phos-
phorus include conventional biological treatment (activated sludge and
trickling filter) (Reference 4), modified activated sludge (Reference 5),
and algae harvesting (Reference 6). These processes convert soluble
phosphorus compounds to organically bound forms in biological cell
tissue. Not more than 20 to 40% of the phosphorus normally present
in domestic sewage is removed by conventional biological treatment.
Presumably, the removal of phosphorus by adsorption on activated
sludge floe, rather than by conversion, overcomes this problem. In
this method, the activated sludge process is operated in the dispersed
growth phase with a detention time of about 30 to 60 minutes. The
organisms containing the adsorbed phosphorus are permitted to floe
and are then removed from the effluent and disposed of separately.
A variety of removal methods are currently under investigation.
-------
TABLE 1
COMPARISON OF ALTERNATIVE PHOSPHORUS REMOVAL METHODS
(a)
Process
Conventional Biological
Treatment
Modified Activated
Sludge
Chemical Precipitation
Chemical Precipitation
with Filtration
Sorption
Removal
Efficiency,
%
10-30
60-80
88-95
95-98
90-98
Estimated
Removal Cost
$/MG
30-100
30-100
10-70
70-90(b)
40-70
Type of Wastes
for
Disposal
Sludge
Sludge
Sludge
Liquid and
Sludge
Liquids and
solids
Remarks
Removal efficiency
and cost depend on
chemical used
Cost based on water
treatment costs
(a) Data compiled by Eliassen and Tchobanoglous (Reference 2).
(b) According to "W. D. Hatfield, a more realistic figure would be about $105/MG (Reference 3).
-------
Discrepancies in the reported degree of phosphate removal by activated
sludge apparently result from changes in composition of microbial proto-
plasm with different sludge ages and perhaps differences in chemical
compositions of waste-waters. It is known that microbial protoplasm is
a dynamic, continually changing, material (Reference 7).
Chemical Precipitation
Precipitation of phosphorus in wastewater is usually accomplished by
the addition of coagulants such as alum, lime, or iron salts and poly-
electrolytes. Commercial processes involving the use of metal ions
and polyelectrolytes are also available (Reference 8).
The reactions and removal mechanisms occurring during chemical
precipitation are not completely understood. The precipitates are not
simple aluminum or calcium phosphates. Cecil has recently reviewed
the literature on chemical removal mechanisms (Reference 9).
Chemical precipitation may be carried out as primary or secondary
treatment, or as a separate operation. In the primary stage, most
of the phosphorus is removed in the primary sedimentation tank. The
remaining phosphorus will normally be removed, by conversion, in
the biological treatment phase. When precipitation is to be accom-
plished in the secondary or biological treatment stage, the chemicals
are usually added to the aeration tank. The precipitates formed in the
biological reactor are removed in the secondary settling tank. By a
combination of chemical and biological treatments, the efficiency of
phosphorus removal is 80 to 95%. Depending on the chemical dosage,
incremental removals of BOD, COD, and surfactant are also achieved
(Reference 2).
In some installations, phosphorus is precipitated as a separate operation,
following conventional biological treatment. Using alum or lime as a
coagulant and precipitant, the coagulation/sedimentation process will
remove phosphates to a level of 0. 16 mg P/l or less, provided that
sufficiently large dosages (200-400 mg/1.) are used.
A promising modification of the chemical coagulation process is the
precipitation of phosphorus with a coagulant such as alum on filter or
separation beds (Reference 10). Coagulants, coagulant aids such as
polyelectrolytes, and pH adjusters are added to the activated-sludge-
treatment effluent just before the filter. It is reported that ortho-
phosphate can be reduced to less than 0. 32 mg P/l. using alum doses
of about 100-200 mg/1. In addition, the BOD is consistently reduced
to less than 1 mg/1. , and the COD is reduced by about 50%.
Conventional alum treatment for phosphate removal increases the con-
centration of sulfate ion in solution. A recent report (Reference 11)
describes a sorption process in which activated alumina is used to sorb
phosphates without any increase in effluent sulfate concentration or
-------
change in pH. Regeneration of the alumina is accomplished with small
amounts of caustic and nitric acid. Phosphate removal efficiencies up
to 98% were attained.
ECONOMICS OF PHOSPHORUS REMOVAL
The costs of various phosphorus removal processes are summarized
in Table 1. In order to facilitate the comparison, data on the removal
efficiency, estimated cost of treatment, and the form of the ultimate
wastes for disposal are given. The wide spread in the cost data for the
individual processes reflects the paucity of reliable published informa-
tion. On the basis of Table 1, it would appear that the best possibility
for achieving a high removal efficiency at a minimum cost might be
found in some kind of chemical precipitation or sorption process.,
Besides the cost of phosphorus removal, however, the costs of ultimate
disposal of contaminants must be considered. The quantities of con-
taminants, both liquids and solids, will depend upon the removal method.
For small plants, these contaminants can be handled quite readily in
lagoons, sludge drying beds, or sanitary landfills„ However, as the
treatment plant capacity increases, waste tonnages for ultimate dis-
posal may become a significant problem.
A summary of applicable ultimate disposal methods is given in Table 2.
Three generally acceptable possibilities exist: dumping; conversion and
dumping; and conversion, product recovery, and dumping. Of the three
methods, the last, although the most desirable, is at present the least
attainable. A fourth, subsurface injection, is included for comparative
purposes only, this method not being considered suitable by current
EPA policy.
The cost of ultimate disposal as shown in Table 2 varies over an ex-
tremely wide range depending on the disposal method. Because of this
wide variation, ultimate disposal methods must be considered carefully;
this cost component may be the controlling factor in the choice of re-
moval process or even in the overall feasibility of phosphorus removal
per se.
NEW SORPTION PROCESSES
The investigation described in this report grew out of the demonstration
in our laboratory that the commercial r esin, Amberlite XE-243, has a
phosphorus-sorption capacity of about 1 mg/ml, or about one-third of
that resin's capacity for boron when determined under similar experi-
mental conditions. The composition of XE-243 may be represented as
R-CH2-N(CH3)-CH2(CHOH)4-CH2OH; it is prepared by the condensation
of a chloromethylated styrene-divinylbenzene (SDVB) copolymer with
the polyol, N-methylglucamine. These results suggested that other
polyhydroxylic resins might be prepared with higher phosphorus capa-
city which would be, unlike XE-243, neutral materials with possibly
-------
TABLE 2
ULTIMATE CONTAMINANT DISPOSAL METHODS AND COSTS
(a)
Type of Waste
Method of Disposal
Cosr
Remarks
Liquid
Soil Spreading
Shallow Well Injection
Deep Well Injection
Landfill
Evaporation Pond
Ocean Outfall
$0.001-0.3/1000 G
$13-27/1000 G
$8-64/1000 G
(c)
Provisions must be made to prevent groundwater
contamination
Disposal site should be available (porous strata,
natural or artificial cavities, etc. )
Used as a wetting agent
Provision must be made to prevent groundwater
contamination
Truck, rail hauling or pipeline needed to trans-
port the wastes
Sludge
Lagooning
Landfill
Recovery of Products
Wet Combustion
Incineration
C.C. $4-80/GPD
,<*)
O &M $1.4 to 35/1000 G
C.C. $10-100/GPD
O&M $5- 55/1000 G
(e)
Used as a wetting agent
Depends on waste characteristics
Heat value may be recovered for use
Concentration of sludge is needed. Ash must be
disposed of.
Ash
Landfill
Soil Conditioner
$0.75-1.0/ton
Depends on waste characteristics
(a) From Eliassen and Tchobanoglous, Reference 12.
(b) Costs vary with plant capacity and land cost; they are based upon volume of waste (i. e. , sludge or concentrate)
handled, not necessarily on volume of sewage treated.
(c) $25/ft for 8" dia pipe.
(d) C. C. = Capital cost
(6) O and M = Operation and Maintenance
-------
increased specificity for phosphate. A program having as its objective
the investigation of the technical and economic feasibility of phosphate
removal from secondary effluents by means of polyol sorption r esins
was funded by the Federal Water Quality Administration.
The scope of this program was to include:
1. Preparation of a number of microreticular and macro-
reticular polyhydroxylic resins.
20 Determination of the phosphate capacity of these resins
and of XE-243 using synthetic ortho phosphate and poly-
phosphate solutions and natural secondary waste effluent.
3. Determination of the regenerability of the candidate resins
(including XE-243) with respect to regenerant levels, re-
covery of sorbed phosphate, and volume of waste solutions
produced.
4. Determination of the stability of the new resins relative
to XE-243.
5. Preparation of an estimate of the economic feasibility
of phosphate removal by sorption on a polyol resin.
The new polyol resins that were desired for evaluation as phosphate
sorbents were esters and amides preparable by the reaction of acylating
cation exchangers with polyhydroxyl alcohols or amines.
RSO2C1 + HO-R'-(OH)5 - >- RSO3-R'-(OH)5
RSO2C1 + HN(CH3)-R'-(OH)5 - »~ RSO2N(CH3) -R1 -(OH)5
RCOC1 + HO-R'-(OH)5 - *.RCO2-R'-(OH)5
RCOC1 + HN(CH3)-R'-(OH)5 - ». RCON(CH3)-R'-(OH)5
Later, polyol ether derivatives from resin quaternary ammonium and
sulfonium halides were also investigated.
RCH2S(Me)2Cl + HO-R'(OH)5 - *~ RCH2O-R' -(OH)5
RCH2N(Me)3Cl + HO-R'-(OH)5 - *. RCH2O-R' -(OH)5
After a number of different polyol resins had been prepared and found
to have disappointingly low capacities, the program was modified to
include the preparation and evaluation of two additional types of sor-
bents: (1) ion-exchange resins containing multi-valent metal ions,
-------
where the metal was iron or aluminum; and (2) resins containing silanol
(Si-OH) groups.
The rationale for proposing the investigation of these materials was as
follows: the removal of phosphate during secondary treatment of waste-
water is accomplished by the addition of lime, iron salts, or aluminum
salts. The mechanism of removal is generally believed to involve
either the formation of metal phosphate or the physical adsorption of
phosphate ion on hydrolyzed metal salts. At the present time, the
former view is possibly more popular.
Phosphate is also known to be relatively immobile in the soil, although
the mechanism or mechanisms involved are subject to debate. Sorption
of iron, calcium, or aluminum phosphates may occur on the silicate
lattice of clay, or phosphate ion may actually replace hydroxyl in the
silicate lattice. In addition, the hydrous oxides of iron and aluminum,
which are a major component of soil, are thought to play a role in the
fixation of phosphate.
Based on the above facts, it seemed possible that materials with high
phosphorus capacity might result if such functions as Fe-OH, Al-OH or
Si-OH could be fixed on an insoluble, stable matrix (as provided by a
resin). To be of practical interest, the phosphate must be removable
from the sorbent, preferably (but not necessarily) without also removing
the moiety responsible for binding the phosphate.
"When sample compounds were prepared as shown below, the silanol
derivative (obtained in 51% conversion) was found, like those of the
RCH2C1 + H2N(CH2)3Si(OEt)3
H20
H20
RSO,Na + Fed, *• +• (RSO,KFe + (RSO,).,FeOH
3 J J 3 3 £
polyol esters and amides, to have a phosphorus capacity much smaller
than that of XE-243; the capacity of the sulfonic acid/iron product, on
the other hand, was much greater than that of XE-243. Accordingly,
the remainder of the original program was spent in studying the sorp-
tion and regeneration characteristics of this iron-containing sorbent.
Based on the promising results obtained with resin/iron sorbents during
the latter part of the original program, a six-months extension was re-
quested and granted for the purpose of studying phosphorus removal by
resin/metal sorbents. The scope of the amended program included es-
sentially the same objectives as the initial program, but was based on
the more effective res in/metal sorbents.
-------
SECTION III
RESULTS AND DISCUSSION
As indicated in Section II, the two major goals of this program were
(1) the preparation of a variety of phosphate-sorption resins, and (2)
the evaluation of their potential utility as sorbents in an economically
feasible phosphorus removal process.
In keeping with the desire to develop ultimately a minimum-cost pro-
cess, the syntheses undertaken were selected, so far as possible, with
a concern for (1) their minimum complexity, (Z) the commercial availa-
bility of the necessary resin substrates, and (3) the relative ease of
preparation of essential intermediates, such as resin sulfonyl chlorides
and carbonyl chlorides.
SORBENT SYNTHESIS
Resin Sulfonyl Chloride
As summarized in Table 3, both commercial sulfonic acid cation ex-
changers and an unsulfonated 1%-crosslinked SDVB copolymer were
readily converted to their sulfonyl chloride derivatives by reaction
with chloro sulfonic acid0 The latter served as both reagent and solvent.
RS03H + C1S03H
RSO2C1
RH + 2C1S03H
RSO2C1
HC1
The volume of chloro sulfonic acid required to maintain sufficient fluidity
varied inversely with the degree of crosslinking in the resin substrate.
TABLE 3
PREPARATION OF RESIN SULFONYL CHLORIDE
Run
36-S
86-S
122-S
Resin, g^'
D50W(H), 23
A200(H), 25
ES-1, 10
C1SO3H, ml
150
100
90
Temp,
oc(b)
80
80
70
Time,
hr
6
6
6
Yield,
%
89.5
88.0
96.6
(a)
(b)
(c)
Dry weight.
Prior to being heated, the mixture stood for 16 hours at room tem-
perature.
Based on Cl content.
11
-------
In preliminary experiments not reported here in detail it was found
that (1) reaction did not occur when chloroform was used as a solvent;
(2) attempted chlorination of the sodium form of the resin resulted in
extensive bead breakage, due perhaps to the precipitation of sodium bi-
sulfate inside the beads; and (3) comparable yields of sulfonyl chloride
were obtained fromIR-120 and D50W, but the latter exhibited only an
insignificant amount of bead breakage, much less than that shown by
IR-120.
Thorough washing of the products with a dilute sodium carbonate solu-
tion (see Section IV) was required toiremove sorbed sulfuric and hydro-
chloric acids. A washed product wasvstable for months, whereas^n-
washed products developed a hydrogen chloride odor in 2 to weeks,
with a consequent loss in chlorosulfonyl content.
Resin Carbonyl Chloride
The conversion of IRC -50, a carboxylic acid cation exchanger (RCOOH),
to the acid chloride (RCOC1) was performed using either thionyl chloride
or phosphorus pentachloride (Table 4). Preliminary experiments with
phosphorus trichloride or phosphorus oxychloride provided only traces
of-acid chloride; reactions with chlorosulfonic acid under a variety of
conditions resulted only in extensive charring.
TABLE 4
PREPARATION OF RESIN CARBONYL CHLORIDE
Run
9 -A
18-B
62-A
Resin, g^ '
IRC -50, 10.9
IRC -50, 0.8
IRC -50, 10.9
Reagent, g
SOC12, 65.0
PC15, 1.8
PC15, 46. 3
Solvent, ml
None
Benzene, 10
Toluene, 90
Temp,
°C
65-76
80
110
Time,
hr
9
4.5
20.5
Cl,
%
9.1
9.1
10.0
(a)
(b)
Dry weight.
A fully chlorinated product contains 32.4% Cl.
The chlorine contents obtained indicated a 30% Conversion to the car-
bonyl chloride. However, it is known that thionyl chloride effectivelv
converts carboxylic acids like maleic, succinic, or phthalic to their
respective anhydrides, (RCO^O (Reference 13). Confirmation of the
presence of significant amounts of anhydride was obtained from the
infrared absorption spectra (Section 4). The mixed acid chloride/
anhydride composition of the products was further confirmed by their
12
-------
reaction with N-methylglucamine (see below), which provided much
larger yields of the amide than could be accounted for on the basis of
the acid chloride content alone.
ES-1 Copolymer
One consideration that was briefly explored in the effort to obtain sor-
bents of higher phosphorus capacity was the possible effect of porosity.
In order to be able to compare a porous but highly crosslinked resin
such as A200 with a porous but low -crosslinked resin, it was first
necessary to synthesize the latter, which was not available commer-
cially. A low -crosslinked intermediate would have the additional ad-
vantage of greater chemical reactivity. It is of interest in these con-
nections to note that XE-243, as well as a number of known anionic and
metal -chelating resins, has a DVB content of only 1 to 2%.
Some typical preparations of ES-1 copolymer beads, made using poly-
vinyl alcohol and magnesium silicate as suspending agents, are sum-
marized in Table 5. The rate of stirring was found to be the most
critical factor in determining bead size.
ES-q Chloromethyl Resin
Reaction of ES-1 copolymer with methyl chloromethyl ether in the
presence of Friedel-Crafts catalysts provided the chloromethyl de-
rivative (Table 6), which was required as an intermediate for the
RH + C1CH2OCH3 catalyst » RCH2C1 + CH3OH
preparation of resin polyol ethers. As the extent of chloromethylation
increased, the bead color changed from white to yellow, and bead hard-
ness increased. The latter is probably indicative of the occurrence of
crosslinking (Reference 14). Zinc chloride was the most effective cata
lyst (Reference 15).
Resin Polyol Sulfonates
Resin polyol esters were prepared by the reaction of resin sulfonyl
chlorides with pyrogallol, mannitol, or sorbitol in the presence of
bases.
RS02C1 + HO-R'-(OH) >>RSO3-R'-(QH)x + HC1
A preliminary study of the behavior of resin sulfonyl chloride in candi-
date reaction media had demonstrated the superior swelling actions of
DMF and DMSO. Water and pyridine produced only slight swelling, and
toluene and triethylamine practically none (Figure 1). DMF was pre-
ferred to DMSO because it showed less tendency to cause bead breakage,
or to decompose or undergo reaction; furthermore, the complete re-
moval of DMSO from the products was difficult.
13
-------
TABLE 5
PREPARATION OF ES-1 COPOLYMER BEADS
Run
141
144
150
222
Water
300
300
2100
2100
Styrene
50
50
350
350
Polyvinyl
Alcohol, g
2
2
14
14
Magnesium
Silicate, g
1
1
7
7
Temp,
°C
80
80
80
80
Time,
hr
6
5.5
6
6
Rate of
Stirring,
rpm
700
800
900
750
Yield of Copolymer, %
20-45
mesh
30
10
9
39
45-100
mesh
20
40
58
43
-------
TABLE 6
PREPARATION OF ES-1 CHLOROMETHYL RESIN
Run
124-C
136-C
145-C
ES-1,
g
4
3
3
Methyl
Chloromethyl
Ether, g
40
30
30
Catalyst,
g
ZnCl2, 3
A1C13, 2
A1C13, 2
Temp,
°C
24
60
24
45
Time,
hr
1
2
16
16
Cl, %
27.2
2.1
9.2
CH2C1,
units/ring
1.15
0. 12
0.53
(J\
-------
60
W
en
I—I
u
INI
o
en
50 —
DMSO
DMF
WATER
PYRIDINE
TOLUENE, TRIETHYLAMINE
DURATION OF SWELLING, Hours
Figure 1. Swelling of RSO,,C1 in Various Liquids
L*
-------
The results of the preparations are summarized in Tables 7 and 8. The
pyrogallol ester (Run 77) was obtained in 58% yield (based on weight gain)
when prepared in DMF, using pyridine as the base. The sulfur content
of the ester, however, corresponded to a much higher yield (Table 8).
Other preparations in pyridine as medium and/or at lower temperatures
were less satisfactory.
The sulfonate esters of mannitol and sorbitol presented more difficulty.
Pyridine was ineffective, and the products obtained using triethylamine
in DMF or DMSO were impure. Thus, a 41% yield of ester resulted in
Run 77, but the product also contained 4, 64% N (Table 8). The absence
of odor and the presence of only trace amounts of DMF (as shown by in-
frared analysis) suggested that the nitrogenous impurity was present as
an amine sulfonate salt or as a sulfonamide.
When the reaction was attempted with ES-1 sulfonyl chloride in a dioxane/
water medium (Run 142), a pure product was obtained as indicated by its
elemental analysis (Table 8)0 Although the yield based on weight gain
was low, the analyses corresponded to a composition that was 45% of
the desired ester and 55% of residual sodium sulfonate.
Pure products were also obtained in reactions with the polyol sodium
salt (see Section 4) in molten polyol. Mannitol was the preferred polyol
RSO9C1 + NaO-R'-(OH) ». RSO^R'-(OH) + NaCl
L* X ~> X
because of its lower melting point. (This method was used in the pre-
paration of resin polyol ethers and is discussed in more detail below.)
As calculated from weight gain, the yields in Runs 202 and 204 were
still low. However, these yields are misleading in that the weight gain
due to esterification was offset by weight loss caused by the high reac-
tion temperatures. The evolution of acid (and a consequent loss in
2RSO3H +. H2S04 + RS02R
weight) was indicated by the neutralization of the excess sodium salt, as
well as by a darkening in color which heating in sulfuric acid produces.
The sulfur and chlorine contents of Run 194 (Table 8) —which was pre-
pared similarly to 202 —were consistent with a 52% yield of ester (and
48% residual sulfonate), considerably greater than Run 202's 12% yield
based on weight gain. Similarly, the elemental analyses for the product
in Run 204 corresponded to a 45% yield of ester (and 55% residual sodium
sulfonate), which was also considerably larger than the weight-gain yield.
17
-------
TABLE 7
PREPARATION OF RESIN POLYOL SULFONATES
00
Sulfonyl
Chloride, 1
[Run g Polyol, g Medium, ml
!
77-SP D50W, 4. 3 Pyrogallol, DMF, 45
3. 8
77-SM D50W, 2. 5 Mannitol, DMF, 20
2. 1
142-SM ES-1, 3.0 Mannitol, Dioxane/ Water ( 10: 1), 44
3. 3
202-SS ES-1, 3.0 Sorbitol, 25 MEE(b^ 8; Toluene, 10
i
204-SMl ES-1, 3.0 Mannitol, 25 ' MEEE(d), 8; Toluene, 10^c)
i i
, i
Base
Pyridine, 2 ml
Triethylamine,
5 ml
Triethylamine,
Sorbitol sodium
salt, 4. 6g(C'
Mannitol sodium
salt, 4. 6g
Temp,
°C
75-80
100
100
130
180
Time, Yield,
hr %(a)
4. 5 58. 0
4 41. 0
4 9. 6
16 11.6
16 14.4 i
1
Expressed as percent of the theoretical weight gain.
Bis(2-methoxyethyl) ether.
(a)
(b)
Also 2. 2g of sodium iodide.
Bis 2(2-methoxyethoxy) ethyl J ether.
-------
TABLE 8
ELEMENTAL ANALYSIS OF RESIN POLYOL SULFONATES
Sulfonate
77-SP
77-SM
142-SM
194-SSb
204-SM
Cl, %
0.5-0.8a
2.0
0.25
0. 56
0.80
N, %
-
4.6
0.29
0.40
0.08
s, %
Found
10.9
10. 7
11.7
11.3
11.8
Calculated
RS02C1
14.3
13.5
15.8
15.8
15.8
RSO,R'(OH)
3 v 'x
10.2
8.4
9.2
9.2
9.2
(a) The range of chlorine values found in preparations similar to 77-SP.
(b) Prepared similarly to Run 202-SS (Table 7).
-------
Resin Polyol Sulfonamide
The reaction of resin sulfonyl chlorides with N-methylglucamine (NMG)
in DMF or wet dioxane media provided the polyol sulfonamides. Two
equivalents of NMG were used, thereby eliminating the need for a second
base to bind the acid produced.
RS02C1 + 2HN(CH3)-R'-(OH)2 - »*RSO2N(CH3)-R' -(OH)x + NMG-HC1
When the reaction was carried out in DMF, the product appeared to con-
tain dimethylamine (from the degradation of DMF), present either as a
sulfonamide or a sulfonate salt. Reaction in wet dioxane provided higher
yields and a product from which nitrogenous impurities were completely
removed by washing with 2% HC1 solution. The order of increased yields
observed in Runs 97, 105, and 126 of Table 9 parallels the increase in
sulfonyl chloride reactivity (or porosity) in going from the conventional
D50W derivative to the more porous, macroreticular A200 derivative to
the still more porous ES-1 derivative. The sulfonamides1 nitrogen con-
tents corresponded well with the yields obtained.
Resin Polyol Carboxamide
Polyol carboxylic amides were prepared by the reaction of the carbonyl
chloride /anhydride intermediate with NMG in DMF medium.
x
RC(O)C1 + (RCO)?O + 3 HN(CH,)-R'-(OH)
L* J 3
RC(OH)N(CH,)-R'-(OH) + RCOOH + NMG-HC1
~J 5C
Unlike the results experienced with the resin sulfonyl chloride, DMF
was a satisfactory reaction medium, undergoing little or no decompo-
sition. Typical procedures are summarized in Runs 66 and 86 (Table 9).
The nitrogen contents of the products were in satisfactory agreement
with the yields calculated on the basis of weight gain.
Resin Polyol Ethers
The first route investigated, the base -catalyzed reaction of the ES-1
chloromethyl derivative with polyols, was unsatisfactory when attempted
RCH2C1 + R'OH » RCH2OR' + HC1
with either sodium hydroxide or organic amines, such as pyridine, tri-
ethylamine or dimethylaniline. The former was insufficiently reactive;
the latter competed with the polyol for the chloromethyl compound,
forming an organic ammonium chloride (a typical anion exchanger
composition) .
20
-------
TABLE 9
PREPARATION OF RESIN POLYOL AMIDES
Run
97-SN2
105-SN1
126-SN
66-AN
8 6 -AN
A rt J J
Acid
Chloride, g
D50W, 2. 5^
A200, 5.0^
ES-1, 20^
IRC-50, 1. 2^
IRC-50, 3. 0^f^
A T> f f~+
INMLr,
4. 5
8. 5
4. 3
3. 2
9.0
Medium, ml
D/W(d), 25
D/W, 44
D/W, 22
DMF, 30
DMF, 40
Temp.,
°C
100
100
100
85-90
90-105
Time,
hr
3
3
1.5(e>
4
10
Yield,
%(a)
35. 0
47. 2
71. 0
53. 0
59. 2
Cl, %
0.74
1.07
0. 66
(g)
(g)
N,
Found
1. 09
2. 28
2. 84
3. 38
3. 92
%
Theory (b)
3. 54
3.40
3. 79
5.22
5.22
(a)
(b)
(c)
(d)
(e)
Expressed as percent of theoretical weight gain.
For a 100%-converted N-methylglucamide.
Resin sulfonyl chloride.
Dioxane/water (10:1, vol).
Longer reaction times (3 to 5 hours) gave similar results.
Resin carbonyl chloride/anhydride
^' Qualitative tests for chlorine were negative.
-------
RCH2C1 + Z3N >
In the second approach, the quatenary ammonium and dimethylsulfonium
chloride derivatives of the ES-1 copolymer were condensed with polyols
at relatively high temperatures in the presence of caustic (Reference
16).
RNMe,Cl + HO-R'-(OH) NaOH ^ RO.R._(QH) + Me N + NaCl + H?O
3 X X J fc*
RSMe7Cl - HO-R'-(OH) NaOH ^ RQ_R..(QH) + Me?S + NaCl + H?O
LJ X X C* "
The product from the ammonium chloride reaction showed no increase
in weight; 90% of the beads were shattered and the remainder were re-
duced to a fine powder. Although the infrared spectrum contained strong
hydroxyl and ether absorptions, the chlorine and nitrogen contents were
consonant with a residual trimethylammonium content of 30% (22% as
hydroxide and 8% as chloride).
An incompletely reacted product was also obtained from the sulfonium
halide attempt. Although the product again exhibited an ether absorption
in the infrared, there was a loss in weight, and the sulfur analysis indi-
cated a 37% content of residual resin dimethylsulfonium hydroxide. The
prime objection to both products, of course, was that their significant
residual anionic contents would result in misleading and overly large
phosphorus capacity when the products were tested.
The resin polyol ethers were prepared in satisfactory purity and good
yield, however, when the ES-1 chloromethyl derivative was reacted
with the sodium salt of a polyol in molten polyol medium (Table 10).
Sodium iodide was added to increase the reactivity of the chloromethyl
compound through halogen exchange (Reference 17), and small amounts
of high-boiling ethers served to swell the resin and make the mixtures
more fluid. Both the iodide and the ether were necessary for good con-
version; when either was absent, the reaction proceeded to only a small
extent. Toluene effected the azeotropic removal of water, with con-
siderable improvement in yield (compare Runs 198 and 203).
The products exhibited strong ether absorption in the infrared (1040-
1100 cm"1). The higher yield obtained with sorbitol (Runs 203 and 200)
is perhaps due to the lower reaction temperature made possible by the
lower melting point of sorbitol. In order to see if the ether was being
retained by the beads and thus contributing to the gain in weight ob-
served, Runs 200 and Z03 were repeated but without the sodium iodide
and the polyol salt; there was no gain in weight, indicating that the ether
was not being retained.
Just as in the preparations made from the quaternary ammonium and
dimethylsulfonium halides, the yields noted in Table 10 are spuriously
low because of the degradation-caused weight losses that undoubtedly
22
-------
TABLE 10
PREPARATION OF RESIN POLYOL ETHERS
Run
198-CS
200-CM
203-CS
Chloro-
m ethyl
ES-1, g
3. 5
3.3
3. 3
Polyol
Sodium
Salt, g
S, 6. 1
M, 6. 5
S, 6. 5
Sodium
Iodide,
g
3.6
4. 1
4. 1
Polyol,
S, 25
M, 25
S, 25
Ether,
ml
5(0
g(e)
8(0
Toluene,
ml
10
10
Temp,
°C
160
190
130
Time, hr
16
16
16
Yield, %(b)
17. 1
28.7(f)
41. 3
OJ
(a)
(b)
(c)
(d)
M, mannitol; S, sorbitol.
Expressed as percent of theoretical weight gain.
Bis(2-methoxyethyl)ether.
Analysis: 3. 8% Cl and 0. 23% N. (Equivalent to a polyol ether containing 27% RCH Cl and
(f)
Bis 2(2-methoxyethoxy)ethyl ether.
Analysis: 2. 07% Cl and 0. 21% N. (Equivalent to a polyol ether containing 16% RCH2C1 and
RCH2NH3OH).
-------
occurred. This belief is supported by the chlorine analyses, which
correspond to residual chloromethyl resin contents of 27% for Run
198 and 16% for Run 200. In both products the nitrogen content was
insignificant.
Silanol Resins
The preparation of silanol resins was limited to a single attempt at
each of the following reactions:
RCH2C1 + H2N(CH2)3Si(OEt)3
RS02C1 + H2N(CH2)3Si(OEt)3
HO
Overnight stirring of ES-1 chloromethyl resin with excess y-amino-
propyl-triethoxysilane at 90° provided (after thorough washing with
water, rinsing with acetone, and drying) a 51% yield of the aminosilanol
(putative). A similar treatment of the sulfonyl chloride at 75° for four
hours gave only a 10% yield of the sulfonamide (putative).
Res in/ Metal Sorbents
A brief investigation of resin/metal sorbents for phosphate removal was
conducted during the latter part of the initial 12 -months program. Pro-
mising results were obtained with sorbents prepared from cation ex-
changers (Na form) and aqueous solutions of iron or aluminum chloride.
Based on these results, a 6-months extension was granted to study phos-
phate removal by resin/metal sorbents.
In aqueous solution, these metallic cations are hydrated and behave as
acids in the Bronsted sense by donating protons to the solvent water
molecules. Thus, the hydrated iron (III) ion hydrolyzes acidically as
follows:
Fe(H2O)5(OH)2+
Fe(H20)5(OH)2+
These products are also believed to be capable of further reaction to
produce higher molecular weight (and correspondingly less soluble)
polymers (Reference 18).
Between pH 2 and 3, the major species present in iron solutions at a
total iron concentration of 1 x 10~3 M are Fe(H2O)^3+ and Fe(H2O)5(OH)2+
(Reference 18). Assuming a similar situation holds in more concentra-
ted solutions, the reaction between the cation exchanger and aqueous
ferric chloride solution involves mainly the following reactions (water
of hydration is neglected for simplicity).
24
-------
3 RS03~ + Fe3+ >-(RS03~)3Fe3+
2 RSO3" + FeOH2+ *~ (RSC>3~ )2Fe(OH)2+
Presumably the aluminum salt reaction follows a similar course, al-
though the hydrolysis of aluminum is believed to be more complicated
than that of iron (Reference 19).
When 5 g (22.2 meq) of A200(Na) was stirred with 12 g (44.4 mmoles)
of ferric chloride in one liter of water, the vacuum-dried product con-
tained 10.0% Fe, or 1. 79 mmoles Fe/g of product. Assuming all the
sodium ions were replaced by iron, each gram of product contains
0,900/0.2022 or 4.46 meq of resin (where RSO " has an equivalent
weight of 202.2).
Let x = meq of Fe on resin
y = meq of FeOH on resin
then, x + y = 1. 79 (1)
3x + 2y = 4.46 (2)
multiplying (1) by (2) and subtracting
3x + 2y = 4, 46
2x + 2y = 3.58
x =0.88
y = 0.91
The product, if the assumption is valid (which appears reasonable),
contained approximately equal amounts of Fe^+ and FeOH + ions.
Lacking data on the ease of removal of hydrate water from sorbed
iron, it does not appear possible to determine with certainty the nature
of the sorbed iron. The presence of large amounts of Fe(OH)2+ appears
unlikely, however, since this would leave 2.67 meq (vide supra) of Na"*"
on a resin which has been treated with a two-fold excess of Fe ^".
The preparation of resin/metal sorbents has been accomplished by three
different procedures. In the Batch Process (above), the resin was
stirred for several hours with a solution of a multivalent metal salt.
In early experiments the product was washed with water and acetone
and vacuum dried. It was found, however, that the measured weight
gains did not correlate with phosphorus capacity. Furthermore, higher
25
-------
capacities were obtained with samples of resin/iron sorbents which had
not been dried after preparation, and the practice of drying the resin/
metal sorbents was discontinued.
A second method of preparation, the Drip Pr6cess, was used in early
experiments comparing various resin/metal sorbents, which consisted
of dropping the ferric chloride solution onto the resin in the same man-
ner as used in evaluating Procedure 1 (vide infra). After the sorbent
was washed with water on the filter, the phosphorus capacity was deter-
mined by Procedure 1.
The third, and most successful, method for preparing resin/metal sor-
bents was the Column Process. A column of resin was prepared in water
and classified by backwashing. The ferric chloride was added at a flow
rate determined by the diameter of the column; e. g. , 0. 25 ml/min, 2. 1
ml/min, and 3.0 ml/min for 0.9-cm, 2. 5-cm, and 3.0-cm columns,
respectively. The sorbent was washed thoroughly with water, reclassi-
fied if air bubbles had developed, and exhausted in the same column.
The liquid level was maintained above the resin column at all times.
This latter procedure provided sorbents of consistently higher capacity
than those prepared by the other procedures, and proved to be the
method of choice (vide infra).
SORBENT EVALUATION
Evaluation of the phosphorus capacities of the various sorbents required,
initially, rapid screening procedures to compare the relative capacities
of different samples. For optimization of the phosphate removal process,
however, more accurate results were necessary to evaluate the effects
of various changes in the processes involved in sorbent preparation, ex-
haustion, and regeneration. To fill these differing needs three proce-
dures were utilized.
For optimum utilization of resin capacity, a minimum bed depth of 24 in.
is recommended (Reference 20). In wide resin beds, flow rates of 2
gal/min-ft of resin are readily achieved (this is equivalent to 0.27
ml/min-ml of resin). For the present program, a bed depth of 24 in.
presented two disadvantages. A 2. 5-cm I. D. column 60 cm (24 in.)
deep would require nearly 300 ml of resin, while even a 1.0-cm column
would need 38 ml of resin to achieve that bed depth. Furthermore, the
flow rates needed for satisfactory phosphate removal were found to be
drastically reduced in narrow columns. In 0.9 to 1.0-cm columns, 6. 5
cm deep, a flow rate of 0.25 ml/min was necessary for satisfactory per-
formance. At this flow rate a 24-in. deep bed would require over 300
hours to exhaust, ,if the solution contained 8 mg P/l. and the phosphorus
capacity was 1 mg/ml (a low value). Rather than use unrealistically
high phosphorus concentrations to provide reasonable exhaustion times,
resins were evaluated under non-optimum conditions. Rapid compari-
sons between different sorbents were made using a wide, shallow bed
26
-------
(Test Procedures 1 and 2), while a. deeper, narrower bed (0.9 x 6. 5
cm) was used for obtaining more critical data during optimization runs
(Test Procedure 3). In Test Procedure 1 the test solution (generally
8. 16 mg P/l. and 500 mg NaCl/1.) was added at a rate of 1.4 -f 0. 1
ml/min. Effluent was collected in 50-ml increments until the arbi-
trary breakthrough of 0. 326 mg P/l. (equivalent to 1 mg/1. of PO4)
was reached. As applied to XE-243, a 2-g (dry wt) sample was taken;
with the resin/metal sorbents, a 1-g (dry wt) sample was used.
In Test Procedure 2, the test solution was added to 2 g (dry wt) of
sorbent at a rate of 1.0 + 0. 1 ml/min0 Three 25 ml increments of
effluent were collected and analyzed for phosphorus content. This
procedure was used on the polyol resins, which had a relatively low
capacity. Usually, the sample's capacity was exhausted after the
second increment was collected.
In Test Procedure 3, a sample of 4.6 ml of sorbent was exhausted in
a classified (backwashed) 0.9x6. 5-cm bed in a 25- or 50-ml buret.
The solution was added at a rate of 0. 25 ml/min. The acquisition of
data was accelerated by the finding that raising the phosphorus con-
centration from 8 to 25 mg P/l. had negligible effect on the phosphorus
capacity or leakage. As expected, the phosphorus capacities were
higher, the leakage lower, and the data more precise with the more
time-consuming Test Procedure 3 than with Test Procedure 1.
Test Procedures 1 and 2 were used for initial evaluation of sorbent
preparation and regeneration procedures, taking care to make com-
parisons only between sorbent data obtained by the same procedure.
Capacity evaluations were also performed with other test solutions,
including a polyphosphate solution and natural secondary effluents.
These results are discussed below.
XE-243
In order to verify Test Procedure 1 and to have a firm basis for com-
parison with other sorbents, a number of capacity evaluations were
performed on XE-243. As shown in Table 11, highly reproducible data
were obtained. In two regeneration experiments (Runs 37 and 44), the
sorbed phosphorus was readily removed by 5% sulfuric acid solution.
27
-------
TABLE 11
PHOSPHORUS CAPACITY OF XE-243
Run
37(a)
38
44(a)
45
48
49
P Removal, mg
1.91
1.90
1.91
1.90
1.94
1.93
B r eakthrough
Concentration, mg P/l
0. 391
0.456
0.652
0.564
0.251
0.440
P Capacity,
mg/ml
wet resin
0.281
0.279
0.281
0.279
0.285
0.284
(a)
Regeneration with 100 ml of 5% ^804 solution removed 1.87 mg P
in Run 37 and 1. 95 mg P in Run 44. In each case, the difference
between P sorbed and P recovered was within the 2% accuracy of
the assay.
The phosphorus capacities of XE-243 for polyphosphate and dihydrogen-
phosphate (H2PO4~) ions were also determined. Test Procedure 1 was
followed, except that the source of phosphorus was different. The poly-
phosphate effluent solutions were boiled for 40 minutes before analysis
in order to depolymerize the phosphate species to the orthophosphate
form. As can be seen in Table 12, the capacities for H2PO4~ and poly-
phosphate were approximately 50 and 100% greater than the capacity for
HPO4 ~ (Table 11). Both species were strongly sorbed by the resin;
removal of these phosphates required larger amounts of sulfuric acid
than was needed for HPO4 . A possible mechanism for the sorption
of orthophosphate ions by polyols is described in the Appendix.
Resin Polyol Amides and Esters
The polyol resins initially synthesized in the program were sulfonamide
and sulfonate ester derivatives of D50W and the carboxamide derivative
of IRC-50. Although the IRC-50 product had a somewhat higher capacity
than the other resins, all the retentions were quite low, being at best no
more than 1% of that of XE-243 (Table 13).
Several possible explanations were considered. First, the relative ex-
tent of conversion of the chloromethyl precursor to the polyol derivative
was estimated at 67% (based on the 3. 74% N content of XE-243) in the
case of XE-243, or only slightly more than the 50 to 60% conversions
28
-------
TABLE 12
CAPACITY OF XE-243 FOR DIHYDROGENPHOSPHATE AND POLYPHOSPHATE
Run
23
25
27
28
Phosphate
Species
O A
T T T"} f~\
o /\
Polyphosphate
Polyphosphate
P Removal,
mg
3. 15
3. 03
3.95
3. 98
Breakthrough
Concentration
mg P/l.
0.365
0. 620
0.358
0. 277
P Capacity,
mg/ml wet
resin
0.463
0.445
0. 580
0. 585
P Recovery by
Regeneration,
mg
3. 20
2.91
3. 98
4. 02
Regenerant, ml
5% H2SO4
200
150
600
—
10% H2SO4
-
-
250
500
From sodium dihydrogenphosphate.
(b)
From "sodium hexametaphosphate" (66. 8-68. 0% P_O )„
-------
TABLE 13
PHOSPHORUS CAPACITY OF POLYOL RESINS
Run
59
81
87-2
66
133
Resin
77-SP
77-SM
87-SN
86-AN
XE-243
Resin Type
D50W Pyrogallyl ester
D50W Mannityl ester
D50W N-methylglucamide
IRC -50 N-methylglucamide
SDVB N-methylglucamide
Retained P,
mg/2g dry resin
0. Oil
0. 023
0. 024
0. 043
0. 591
P Capacity,
mg/ml wet resin
0. 0027
0. 0046
0. 0050
0. 0056
0. 0870
OJ
o
-------
of resin sulfonyl chloride or resin carbonyl chloride to the analogous
amide and ester derivatives (see Tables 7 and 9). These small dif-
ferences in polyol content were not a likely cause for the relatively
large capacity difference observed.
A second possibility was that passage of the test solution through the
polyol products (which are normally washed with acid during workup)
would promote exchange of H* for Na+ at the residual SC>3H functions,
and to a lesser degree at the residual COOH groups in the IRC-50
products. The resulting acidity would tend to counteract phosphorus
uptake. A first test of this hypothesis was made by determining the
phosphorus capacity of two polyol products after a prior treatment
with dilute NaOH solution to convert residual sulfonic acid to the
sodium salt. As shown in Runs 74 and 87-3 of Table 16, there was
some improvement for the D50W resin and essentially no change for
the IRC-50 product (see Runs 87-2 and 66, Table 13).
In a second test of this hypothesis a batch of the D50W N-methyl-
glucamide resin was prepared in the usual way except that the work-up
procedure was modified so as to preclude the transformation of un-
reacted sulfonyl chloride to sulfonic acid. (Instead of wash treatments
with water and dilute hydrochloric acid solutions, the product was
rinsed with DMF and acetone, and thendried. ) Portions of this mate-
rial were reacted (separately) with 10% ammonia solution and with di-
n-propylamine. Elemental analysis of the products yielded N contents
of 5. 38% and 40 98%, respectively, indicating that large amounts of base
had been introduced as amide or salt. (The theoretical value for a fully
converted D50W N-methylglucamide is 3. 54% N. )
Comparison of Runs 87-5 and 87-6 (Table 14) with Run 87-2 (Table 13)
shows the phosphorus capacity had increased, but was still much less
than that of XE-243. In the case of the IRC-50 product, the probable
explanation for the failure of basic treatment (Run 74) to improve the
capacity (see Run 66, Table 13) is that the residual COOH content,
because of its slightly ionized nature, had done little to inhibit phos-
phorus uptake in the first place. It is considered significant that the
sample of highest capacity (Run 87-5) underwent no swelling in water,
whereas the other materials showed volume increases of 60 to 70%.
This is explained on the basis that swelling is essentially due to the
solvation of residual 803?!, and too few of such functions remained in
87-5 after the reaction with the amine. The conclusions from these
experiments were that residual SO3H did have a capacity-lowering
effect and that basic conditioning of the polyol resins before evaluation
was desirable, but that the conditioning treatment per se could not in-
crease the capacity of these resins to a satisfactory level.
A third possibility was that the low phosphate uptake was due to an in-
sufficient swelling of the resin in water, with a consequent inadequate
access of phosphate to the polyol portion. To test this hypothesis, the
relative swelling in water of three polyol resins, four sulfonic acid
31
-------
TABLE 14
PHOSPHORUS CAPACITY OF BASE-TREATED POLYOL RESINS
tv
Run
74
87-3
87-5
87-6
Resin
86-AN
87-SN
87-SN
87-SN
N -Me thy Igluc amide
IRC -50
D50W
D50W
D50W
Treatment
2% NaOH(a)
2% NaOH(b)
Di-n-propylamine
10% ammonia
Retained P,
mg/2g dry resin
0. 041
0. 039
0. 028
0. 036
P Capacity,
mg/ml wet resin
0. 0054
0. 0081
0. 0093
0. 0075
(a)
(b)
(c)
Dropwise addition of 25 ml of 2% NaOH solution, followed by 20 ml of water.
Stirring for one hour, followed by washing with water.
Stirring overnight, followed by washing with acetone.
(d)
Stirring overnight, followed by washing with water.
-------
cation exchangers, two weak-acid cation exchangers, and XE-Z43 was
measured. As shown in Table 15, the degree of swelling of the polyol
resins was greater than that of XE-243, particularly for the two resins
(62 and 86) which, like XE-243, are N-methylglucamine derivatives.
Obviously, the low phosphate sorption of these products cannot be
attributed to a mere lack of swelling.
A particularly interesting aspect of this data, however, was the rela-
tively low density (or high porosity) of XE-243, compared with the other
resins, as indicated by the respective volumes of 1-g samples. This
suggested a fourth possibility - that the low phosphate capacity of the
polyol resins was due to a relatively dense or nonporous nature. This
assumption was supported by the fact that XE-243 has a nominal DVB
content of only 1% (Reference 21), whereas the DVB content of D50W
is 8%.
In order to study the relationship of resin density to phosphorus capa-
city, the preparation of low density or relatively porous sorbents was
undertaken beginning with (1) A200, a commercially available, macro-
reticular sulfonic acid exchanger, and (2) ES-1, an Aerojet-synthesized,
1%-crosslinked SDVB copolymer. Macroreticular resins have been
variously described as having pore sizes from one to several orders of
magnitude greater than that of the conventional gel-type resins (Refe-
rence 22). The 1%-crosslinked copolymer was expected to be less
dense than the materials previously used, and possibly more reactive.
With both resins, the usual synthetic approach through the sulfonyl
chloride intermediate was taken.
The phosphorus capacities of the resulting polyol resins are given in the
first four rows of Table 16. The macroreticular polyol resin (105)
showed much the same capacity as the conventional D50W and IRC-50
polyols previously tested (Table 13). Apparently, the macroreticular
structure with its larger pores did little to enhance phosphorus reten-
tion. In the case of the ES-1 ester and amide resins, however, a much
larger increase in capacity resulted, relative to the higher-cross linked
D50W resins. Although a portion of this improvement can be attributed
to the higher yields obtained in the esterification of the ES-1 sulfonyl
chloride relative to its D50W analogs (Table 9), the major improvement
was probably due to an increased porosity. Nevertheless, the capacities
of the ES-1 polyols were still less than 20% of the value for XE-243
(Table 13),
Resin Polyol Ethers
Because of the general desirability of increased porosity and elimination
of residual acidic functions in the polyol sorbents, the resin polyol ethers
were attractive for investigation. These materials would contain no resi-
dual acidic groups as a result of incomplete conversion during their syn-
thesis and, as ethers, would be inherently more stable to hydrolysis than
the ester or amide derivatives. In addition, they could be prepared in
low-crosslinked form, beginning with the ES-1 chloromethyl derivative.
33
-------
TABLE 15
SWELLING OF VARIOUS RESINS IN WATER
(a)
Resin
6Z-SN
77-SP
8 6 -AN
IR-118
IR-120
IR-124
A200
IRC -50
IRC -84
XE-243
Chemical Nature
D50W N-methylglucamide
D50W Pyrogallyl ester
IRC -50 N-methylglucamide
SDVB Sulfonic acid, < 5% DVB
SDVB Sulfonic acid, 8% DVB
SDVB Sulfonic acid, 12% DVB
SDVB Sulfonic acid, 20% DVB^
Methacrylic-DVB copolymer
(e\
Acrylic acid copolymer
SDVB N-methylglucamide
Volume,
ml(b>
1.3
1.4
1.9
1.3
1.5
1.3
1.8
1.7
1.4
2.6
Volume
Increase,
%
69
43
100
161
80
69
56
65
50
38
(a) Procedure: 3. 5 ml of water was added to 1 g of dry resin in a
calibrated vial and the resin's increase in volume after standing
overnight was noted.
(b) The volume occupied by 1 g of dry resin.
(c) Macroreticular.
(d) Nominal DVB content is not known.
(e) Nature and amount of crosslinking is not known.
34
-------
TABLE 16
PHOSPHORUS CAPACITY OF HIGH-POROSITY POLYOL RESINS
(a)
Run
112
138
204
194
200
203
Resin
105-SN2
127 -SN
204-SM
194-SS
200-CM
203-CS
Resin Type
A200 N-methylglucamide
ES-1 N-methylglucamide
ES-1 Mannityl Sulfonate
ES-1 Sorbityl Sulfonate
ES-1 Mannityl Ether
ES-1 Sorbityl Ether
Retained P,
mg/2g dry resin
0. 031
0. 081
0. 121
0. 129
0. 045
00 044
P Capacity,
mg/ml wet resin
0. 0071
0. 010
0. 014
0. 014
0. 012
0. 012
oo
(a)
Prior to capacity evaluation, the resins were washed with dilute NaOH solution.
-------
The phosphorus capacity data for ES-1 polyol ethers is shown in Table
16 (Runs 200 and 203). On the dry basis, the phosphorus retentions
were considerably lower than those obtained with the analogous polyol
esters and amide (138, 204, 194). Possibly, intermolecular ether for-
mation at the high synthesis temperatures employed (Table 10) de-
creased the number of available hydroxyl groups. On the wet basis,
the capacities were essentially the same as that of the esters and
amides because the ethers underwent only 12% swelling in water in-
stead of the 38 to 100% noted with the other derivatives (Table 15).
The low degree of swelling was probably due to the absence of highly
solvatable acid functions (such as SO^H) and to the hydrophobic ether
function.
Silanol Resins
As noted in the Sorbent Synthesis discussion, the limited effort expended
in the preparation of silanol resins provided an aminosilanol (51% con-
version) and a sulfonamide (10% conversion). The phosphorus capacity
of the amine, as determined by Test Procedure 1, was comparable to
that of the polyol resins; the capacity of the amide was negligible, as
might be expected from the poor conversion obtained.
Resin/Metal Sorbents
Compared to the polyol resins, resin/metal sorbents, prepared by con-
version of cation exchange resins to their iron (III) or aluminum forms,
had very high phosphorus capacities. Table 17 shows phosphorus capa-
city data (determined by Test Procedure 1) for sorbents prepared from
IRC-50 (a weak acid cation exchanger), ES-1S, and A-200 (strong acid
exchangers) by reaction with ferric or aluminum chloride. For com-
parison purposes, the quaternary (strong base) anion exchange resin
IRA-400 is included. It can be seen from Table 17 that, on a weight
basis, five of the sorbents had phosphorus capacities greater than that
of XE-243, while three sorbents exceeded the capacity of the anion ex-
changer, IRA-400. For all three cation exchange resins (IRC-50, ES-1S,
and A-200) the iron form had significantly higher capacity than the alumi-
num form.
It is likely that the lower capacity of the ES-1S product in Run 237 was
due more to the greater swelling this low-crosslinked resin underwent
in water than to an intrinsically lesser capacity for phosphate. In Run
243, the evaluation could not be completed because the flow rate slowed
near the end of the determination; when the test was stopped, the resin
had already taken up as much phosphorus as had Run 242. Larger sor-
bent beads would probably provide better results.
The greater phosphorus capacities of the A200 sorbents as compared to
the IRC-50 resins, and the commercial availability of the Amberlite 200
resin (as compared to ES-1S), suggested that our efforts should be con-
centrated on the development of A200/metal sorbents, of which the iron
36
-------
TABLE 17
PHOSPHORUS CAPACITY OF RESIN/METAL SORBENTS
(a)
oo
~o
Run
227
228
237
243
235
242
146
37
Sorbent
IRC-50/A1
IRC-50/Fe
ES-1S/A1
ES-lS/Fe
A200/A1
A200/Fe
IRA-400(d)
XE-243
Wet Volume,
ml
3. 1
2. 7
4. 1
4. 2
2.6
2. 3
3.6
3. 4
P Removal,
mg
0. 763
1. 161
4. 689
>7.4
5. 831
7.409
4.679
1. 914
Breakthrough
Concentration
mgP/1.
0.440
0.326
0.339
-
0.339
0.375
0. 391
0.391
P Capacity,
mg/ml wet
resin
0. 12
0.22
0. 57
-
1. 12
1. 61
0. 65
0. 28
(a)
(b)
(c)
Capacity was determined on a 2-g sorbent sample.
Volume occupied by Ig of dry resin after overnight immersion in water.
The evaluation could not be completed because the flow rate decreased near the end of the determination.
(d)
A quaternary ammonium anion exchanger.
-------
(Ill) form appeared to be most suitable. Several A200/Fe sorbents
were prepared, evaluated by Test Procedure 1, and subjected to
regeneration.
Because the time required to obtain the capacity data in Table 17 proved
excessive (11 hours), the size of resin sample was reduced from 2 to 1
g for subsequent evaluation, all other test parameters being left un-
changed. Using this revised procedure, the data summarized in Table
18 was obtained on A200-Fe sorbents. It is of interest to note the dis-
proportionate effect that halving the sorbent sample size had on the
observed phosphorus capacity. Thus, the 1-g sample of A200-Fe
sorbent (Run 251, Table 18) retained only 21.4% as much P as did the
2-g sample (Run 242, Table 17). This behavior was perhaps due to
the increased channeling that may result in the smaller sorbent sample
(shollower resin bed).
The identical capacity values obtained in Runs 249 and 251 indicated
that the form (H or Na) in which the resin is utilized is unimportant.
Sorbents prepared by the Drip Process and evaluated without prior
drying (Runs 274 and 271 (A-F) had an average capacity value (00 91
mg/ml wet resin) that was about 34% higher than sorbents obtained by
stirring the exchanger in FeCl3 solution and then drying the product
(0.68 mg/ml wet resin). Although drying may reduce the capacity,
later experiments (vide infra) showed that the Drip Process is more
effective than the batch method of preparation.
In addition to the iron sorbents prepared from A200, a number of other
multivalent metal ions were tested as A200 derivatives for phosphorus
sorption capacity. A summary of these experiments is shown in Table
19. It is apparent from the table that, of the ions tested, Fe (III) is by
far the most effective for sorption of phosphorus on a cation exchange
resin.
As the superiority of the phosphorus capacity of A200-Fe over other
resin metal sorbents became increasingly evident, a study of the con-
ditions of its preparation was initiated. To provide greater sensitivity
to changes in phosphorus capacity, Test Procedure 3 (vide supra) was
adopted. For evaluation of process changes, resin samples of 4. 6 ml
(wet volume) were tested in Q. 9 to 1. 0-cm columns, giving a bed depth
of approximately 7 cm. A brief study of flow rates showed that satis-
factory phosphorus removal was accomplished at a flow rate of 0. 25
ml/min. Samples of A200-Fe were prepared by the Batch, Drip, and
Column Processes (vide supra) and evaluated by Test Procedure 1 and
Test Procedure 3. The results, shown in Table 20, show the increased
efficiency of Test Procedure 3, which undoubtedly results from reduced
channeling and improved equilibration between sorbent and solution. It
is also apparent that, while the Drip and Column Processes produced
sorbents with equivalent phosphorus capacities, the capacity of the
Batch-prepared sorbent was definitely inferior. Since the Column
Process appeared (visually) to afford better contact between r esins
38
-------
TABLE 18
PHOSPHORUS CAPACITY OF A200-Fe SORBENTS
(a)
Run
249
251
271A
271B
271C
271D
271E
271F
271G
271H
271 I
274
Sorbent
249
245(C>
249
249
249
249
249
249
249
249
249
274^
P Removal,
mg
1.571
1.581
2.794
2.403
1.963
1.992
1.575
1.999
1.998
1.554
1.989
1.994
Breakthrough
Concentration,
mg P/l.
0.434
0.326
0.381
0. 368
0.489
0.349
0.342
0.346
0.326
0.567
0.385
0.375
P Capacity,
mg/ml
wet resin
0.68
0.69
1.22
1.04
0.85
0.87
0.68
0.87
0.87
0.67
0.86
0.87
(a) Capacity was determined on a 1-g sorbent sample by Test Proce-
dure 1.
(b) Prepared from the H form by the Batch Process.
(c) Prepared from the Na form by the Batch Process.
(d) In Runs 271 (A-I) one resin sample was regenerated nine times by
the Drip Process.
(e) Prepared by the Drip Process and evaluated without prior drying.
39
-------
TABLE 19
PHOSPHORUS CAPACITIES OF SORBENTS PREPARED FROM
A200 AND VARIOUS MULTIVALENT METAL IONS(a)
Breakthrough P Capacity,
Run
263A
263B
372
376
282
384
371
382A
355
385
264
265
Sorbent
A200/A1
A200/Al(b>
A200/A1
A200/A1
A200/Ca
A200/Ca
A200/Fe
A200/Fe
A200/La
A200/Mg
A200/Zr
A200/Zr
Source of
Metal Ion
A1C13
A1C13
Al(OAc)2OH
Al(OAc)2OH
CaCl2
CaCl2
FeCl3(c)
FeCl3(d)
LaCl3
MgCl2
ZrCl4
ZrOCl2
Total P
Sorbed, mg
1. 13
1.59
--
0. 39
--
1. 19
2. 39
--
0. 38
0. 38
Concentration,
mg P/l.
0.49
0. 38
0. 60
0. 32
--
0. 31
0. 34
--
0. 62
0. 68
mg/ml
wet resin
0.49
0. 69
nil
0.17
nil
nil
0. 52
1. 04
nil
nil
0. 16
0. 16
(a)
(b)
(c)
(d)
P Capacity determined on 2. 3-ml samples by Test Procedure 1.
Sorbent 263A regenerated with Aid.,.
1480 ml of 0. 001 M FeCl3> pH 2. 0 (adjusted with concentrated HC1).
100 ml of 0. 1 M FeCl3, pH 3. 0.
40
-------
TABLE 20
EFFECT OF METHOD OF PREPARATION ON PHOSPHORUS
CAPACITY OF A200-Fe SORBENTS
Run
40 7A
40 7B
40 7C
40 8 A
408B
408C
Sorbent*a)
40 7A
40 7B
207C
40 7A
40 7B
40 7C
Preparation
Process
Batch
Drip
Column
Batch
Drip
Column
P Capacity
Test
Procedure
(b)
(b)
(b)
3(c)
3(c)
3(c)
P Capacity,
mg/ml
wet resin
0.52
0.34
0.34
1.30
3.09
2.83
(a) Sorbents were prepared in 46-ml lots.
(b) Determined on 2. 3-ml samples.
(c) Determined on 4. 6-ml samples.
41
-------
and solution than did the Drip Process, sorbents for subsequent runs
were prepared and regenerated by the former.
When the sorbents were prepared or regenerated by the Column Process
in 4. 6-ml lots, still higher capacities resulted. The results listed in
Table 21 show the effects of varying the amount and concentration of
ferric chloride on phosphorus capacity. In Runs 414 and 416, the re-
sults show no difference in capacity between sorbents prepared with
0.2M, 004M, and l.OM ferric chloride solutions, where the amount
of ferric chloride was held constant. In Run 421, the ferric chloride
concentration was held constant while varying the amount of ferric
chloride. The results indicated that reducing the amount of ferric
chloride from 60 meq to 20 meq (per 4.6 ml of wet resin) decreased
the phosphorus capacity by about 15%.
Because column-derived phosphorus capacity values are strongly de-
pendent on the experimental conditions employed, it was of interest to
measure the total capacity of A200-Fe sorbents. The resins were
stirred overnight in phosphate solutions of known concentration, both
in the presence and absence of added sodium chloride. Knowledge of
the phosphorus content of the initial and final solutions provided the
amount of phosphorus sorbed.
As is shown in Table 22, the (dry-basis) capacity varied from 7 to 52
mg/g sorbent, depending on the test solution employed. At very high
phosphorus concentrations (Runs 254 and 286) the capacity was not af-
fected by the presence of sodium chloride. At much lower and relatively
realistic levels (Runs 267; 268, 259, and 260), sodium chloride appears
to increase phosphorus capacity.
The regeneration of A200-Fe sorbents was accomplished by the drop-
wise addition (1.4 ml/min) of 150 ml of 5% sulfuric acid, 200 ml of 2.5%
sulfuric acid, or 300 ml of 1% sulfuric acid solution. However, iron
was also removed, as shown by (1) an 80% loss in phosphorus capacity
on a subsequent exhaustion of the sorbent, and (2) by analysis of the
iron content of the spent regenerant. (With the 1% sulfuric acid solution,
33% of the iron in the sorbent was removed during the phosphate strip-
ping and 21% was removed with the first 50 ml of acid.) Basic rege-
neration was not attempted because preliminary experiments had shown
that treatment of the sorbents with 2% ammonia solution greatly reduced
their capacity.
Regeneration was completely successful, however, when performed by
the dropwise addition of 50 to 200 ml of 0. 1 M ferric chloride solution
(pH 1). The spent regenerant was free of precipitate. A quantitative
determination of the amount of phosphorus removed was not possible
because of interference by the large amount of iron present. The evi-
dence for a successful regeneration can be seen in Runs 249 and 271
(A-l) (Table 18), where a relatively constant phosphorus capacity was
maintained over ten exhaustion/regeneration cycles with a single
42
-------
TABLE 21
EVALUATION OF COLUMN-PREPARED A200/Fe SORBENTS
BY TEST PROCEDURE 3^
Run
41 4A
414B
41 6A
416B
41 6C
421A
421 B
421C
FeCl3 Used,
meq.
60
60
60
60
60
20
40
80
FeCl3 Used,
C one entr ation
0.2 M
0.2 M
0.2 M
0.5 M
1.0 M
0.5 M
0.5 M
0.5 M
P Capacity,
mg/ml wet resin
8.42
9.23
( Q \
I Q )
8.!4
7.33
6.24
9.23
(a) 4. 6-ml Lots of A200/Fe were prepared by the Column Process
in 0. 9 to 1.0-cm (ID) columns and evaluated in the same column
by Test Procedure 3, using a solution containing 25 mg P/l.
(b) The average of these 5 runs is 8.47 + 0.40 mg P/ml. If the high
result of 414B is rejected, the avera"ge is 8. 28 + 0. 14 mg P/ml.
43
-------
TABLE 22
TOTAL PHOSPHORUS CAPACITY OF A200/Fe SORBENTS
Initial Concentration
Run
254
286
£ 259(a)
260(a)
267
268
Sorbent
Weight, g
3
3
1
1
1
1
Solution
Volume, 1
1
1
2.5
2.5
2.5
2.5
mgP/1.
260
260
32.6
32.6
16.3
16.3
mgNaCl/1.
0
16,000
0
2,000
0
1,000
Final
Concentration,
mgP/1.
104. 3
106. 9
29.80
23.54
11.05
5.08
P Capacity
mg/g
dry resin
52.2
51.2
6.9
22. 7
13.1
28.1
mg/ml
wet resin
22.7
22.6
3.0
9.9
5.7
12.2
(a)
These data are perhaps somewhat less reliable because the usual precautions of redrying
the sorbent and rinsing glassware with hot HC1 solution prior to testing were not observed.
-------
sorbent sample. The small variations in observed capacity from cycle
to cycle are probably due to unavoidable variations in the exhaustion
cut-off (breakthrough) point and in the rate of solution addition.
In other early regeneration experiments, the use of weak acids and
weak bases was studied, as well as the use of hydrochloric acid fol-
lowed by ferric chloride treatment. The data in Table 23, obtained
using Test Procedure 1, indicate that regeneration with ferric chloride,
or hydrochloric acid/ferric chloride, is superior to the other regene-
rants. These two regenerant systems were investigated further using
Test Procedure 3. When Batch-prepared A200-Fe was regenerated by
the Column Process, phosphorus capacities equal to those found with
Column -prepared A200/Fe sorbents (Table 24) were obtained. The
data indicate that the added expense of hydrochloric acid treatment
before regeneration with ferric chloride is not warranted.
Table 25 shows the results of a number of exhaustion/regeneration
cycles with A200/Fe sorbents on secondary effluent from the Whittier
Narrows Water Reclamation Plant. The amount of orthophosphate
sorbed averages about 1.7 mg P/10 (Runs 412, 413, 415, 419, 420.)
In Run 419, the total phosphorus capacity was found to be 2. 75 mg P/ml.
Phosphorus removal data for this run are shown in Table 26. Regene-
ration with ferric chloride was successful (Runs 413 and 415), but the
amounts used were the same as for preparation, about 4 mmoles of
ferric chloride per ml of resin (about 45 moles of Fe per mole of P).
Treatment with sodium hydroxide, however, apparently produced a
change in the sorbent that could not be reversed by treatment with
ferric chloride (see Runs 424 and 425).
To achieve a more economical regeneration procedure, two approaches
were investigated. Sorbents 415 (Run 425) and 413 (Run 426) were re-
generated with portions of a used ferric chloride solution collected from
a previous regeneration. Due apparently to the change produced by the
sodium hydroxide (Run 424), sorbent 415 regained little of its previous
phosphorus capacity. Sorbent 413, however, regained 54% of its initial
capacity when regenerated with approximately half the used ferric chlo-
ride solution from Run 419. This would effect a saving of 50% of the
ferric chloride requirement, which was still substantial (2 mmoles/ml
of wet resin, or 22 moles of Fe per mole of P). A more significant re-
duction in the ferric chloride requirement was achieved when sorbent
413 was regenerated with 3 g of ferric chloride (Run 427). While the
amount of ferric chloride used was reduced by a factor of 18, the total
phosphorus capacity of the regenerated sorbent was over one-third of
that obtained with the larger amount of ferric chloride. The require-
ment for ferric chloride, per mole of phosphorus, was thus reduced
by a factor of six (albeit at the expense of a larger resin requirement).
Since the ferric chloride cost proved to be the largest item in the cost
estimate (vide infra), the process using the lesser amount of ferric
chloride was adopted for economic reasons.
45
-------
TABLE 23
REGENERATION OF A200/Fe^
Run
395A
395B
395C
395D
401 A
40 IB
401C
40 ID
404A
404B
40 4C
Sorbent
390A(b)
390B(b)
390C(b)
390D(b)
398A(b)
398B(b)
398C(b)
398D(b)
403A(g>
403B(g)
/ or i
4O ^ f~~*
P Capacity,
1st Exhaustion
mg/ml wet resin Regenerant
0.52 FeCl3(c)
0.52 HCl/FeCl(d)
0.52 HOAc(e)
0.52 NH4OH/ HO Ac ^
0. 70 FeCl3^°^
0.69 HCl/FeCl-,
0. 70 HOAc^6'
0. 70 NH4OH/HOAc^f^
0.70 FeCl3(c)
0.52 FeCl3(h)
0.70 FeCl3(l)
P Capacity, P Capacity,
2nd Exhaustion 3rd Exhaustion,
mg/ml wet resin mg/ml wet resin
0035
0.17
0.17
0.17
0.35 0.35
0.34 0.34
0.35 0.17
0.17 0.17
0.52
0.53
0.34
(a) P Capacities determined by Test Procedure 1 on 2. 3-ml (wet) samples.
(b) Prepared by the Column Process.
(c) 100 ml of 0. 1 M FeCl3.
(d) 100 ml of 5% hydrochloric acid followed by 100 ml of 0. 1 M FeClv
(e) 100 ml of 6% acetic acid.
(f) 100 ml of 1.0 M ammonium hydroxide followed by 100 ml of 6% acetic acid.
(g) Prepared by the Drip Process.
(h) 50 ml of 0. 2 M Fed-,.
(i) 25 ml of 0.4 M
-------
TABLE 24
REGENERATION OF A200/Fe BY THE COLUMN PROCESS
(a)
Run
410A
410B
414C
Sorbent
4lOA(b)
4lOB(b)
409A(b)
P Capacit-y
1st
Exhaustion
1.35
1.48
1. 18
Regenerant
HCl/FeCl3(c)
FeCl^
FeCl3
P Capacity
2nd
Exhaustion
8.68
8.27
9.50
P Capacity
3rd
Exhaustion
8.42
7.06
-
(a) 4. 6-ml samples of sorbent were exhausted in 0.9 to 1.0-cm (ID)
columns by Test Procedure 3. The sorbents were regenerated by
the Column Process and exhausted in the same column. The solu-
tion used contained 25 mg P/l. , except that the 1st exhaustion of
409A was done using 8 mg P/l.
(b) Prepared by the Batch Process.
(c) 100 ml of 5% HC1, then 50 ml of 0. 2 M FeCl3.
(d) 50 ml of 0. 2 M FeCl3.
47
-------
TABLE 25
PHOSPHORUS CAPACITY OF A200/Fe SORBENTS
ON SECONDARY EFFLUENT^
Run
412
413
415
419
420
424
425
426
427
Sorbent
412(b)(c)
4l3(e)(f)
4l5(c)(e)
413(g)
415'C>
415^
415
413(g)
4l3(g)
Regenerant
.
-
-
FeCl3
FeCl3(h)
NaOH^
FeCl3(j)
FeCl3
FeC^1'
P Capacity,
mg P/ml
wet resin
( o r thopho s phate )
1. 74
1.25
1. 70
1.95
1.72
nil
0.38
1.34
0.85
P Capacity,
mg P/ml
wet resin
(total phosphate)
(d)
(d)
(d)
2. 75
(d)
(d)
(d)
1.63
1. 11
(a) 46-ml lots of sorbent were evaluated by Test Procedure 3, using
chlorinated secondary effluent obtained from the Whittier Narrows
Water Reclamation Plant. Lots of effluent were kept for a maxi-
mum of two days, during which the pH varied between an initial 7. 3
to 7. 4 and a final pH of 8. 3 to 8. 4 (no attempt was made to control
pH). The orthophosphate content averaged 6. 2 mg P/l. , and the
total phosphate content averaged 9.8 mg P/l.
(b) Prepared by the Batch Process.
(c) Evaluated in a 3.0-cm (ID) column at 3.0 ml/min.
(d) Not determined.
(e) Prepared by the Column Process.
(f) Evaluated in a 3.6-cm (ID) column at 3.0 ml/min. This ID
apparently resulted in a bed of insufficient depth.
(g) Exhausted in a 2. 5-cm (ID) column at 2. 1 ml/min.
(h) Regenerated in the 3. 0-cm column with 1.01. of 0. 2 M FeClv
(i) Regenerated with 200 ml of 0. 1 N sodium hydroxide.
(j) Regenerated with 200 ml of used FeCl3 solution from Run 420.
(k) Regenerated with 500 ml of used Fed., solution from Run 419.
(1) Regenerated with 30 ml of a 10 wt-% solution of Fed,,- 6H0O at
1.0 ml/min. 3 2
48
-------
TABLE 26
PHOSPHORUS REMOVAL FROM SECONDARY EFFLUENT
BY A200/Fe SORBENT
Fraction
419-1
419-2
419-3
419-4
419-5
419-6
419-7
419-8
419-9
419-10
419-11
Volume,
ml
1000
1295
1000
1800
1000
1660
1000
1610
1000
1000
1680
Total P
Concentration,
Secondary
Effluent
mg/1
9.55
9.55
11.37
11.37
11.37
11.37
11.37
11.95
11.95
11.95
7.89
Total P
Concentration
in Product,
mg/1
0.00
0.00
0.00
0.00
0.01
0.00
0.04
0.07
0.48
0.49
0.82
mg P
Removed
9.55
12.37
11.37
20.47
11.37
18.87
11. 37
19.24
11.95
Total mg P removed = 126.56.
Total P Capacity = 126. 56/46 = 2. 75 mg P/ml wet resin.
49
-------
It was found that spent regenerant solutions could be treated with lime
to remove both the ferric chloride and the eluted phosphate. Aliquots
of spent regenerant (approximately 0.2M in ferric chloride) were
treated with 1, 2, and 3 equivalents of calcium hydroxide. After fil-
tration, analysis of the filtrates by the extraction modification of the
stannous chloride procedure (Reference 23) showed that 99. 9% of the
phosphate was removed, along with virtually all of the iron (K for
Fe(OH)3 is 1. 5 x lO"36 - Reference 24).
On the laboratory scale, then, the present optimum version of this
phosphorus removal process consists of converting A200/Na to its
iron (III) form by treatment with 1. 17 g of ferric chloride (FeCls* 6 H2O)
per ml of wet resin. The resin is exhausted at a flow rate of 0. 5
gal/min/ft^o Regeneration is accomplished with 0.065 g of ferric
chloride per ml of wet resin. Product water is used for backwashing
and rinsing. Iron and phosphorus are removed from spent regenerant
solutions by treatment with lime, with the resulting sludge being trans-
ferred to settling tanks for final disposal.
One aspect of this sorption process is particularly noteworthy. Exami-
nation of the total phosphorus concentration data for the product water
in Run 419 reveals that the first eight fractions, totaling 90% of the re-
ported phosphorus capacity, had product concentrations of 0.07 mg P/l.
or less. Dryden and Stern have reported (Reference 25) that algae growth
is inhibited in effluent containing phosphate concentrations of 0. 16 mg
P/l. or less. Effluent from the A200/Fe sorption process clearly ful-
fills the need for reducing phosphorus content below this level. An es-
timate of the cost of accomplishing phosphorus removal by this method
is discussed in the following section.
ECONOMIC ANALYSIS
Introduction
One objective of this program was the preparation of cost estimates for
methods developed in the laboratory, assuming that such methods will
be used for phosphate removal from the secondary effluent of a one
million gallon per day municipal sewage plant. The method investigated
in the laboratory showed that the cation exchanger/metal salt sorbent —
Amberlite 200/Fe sorbent — has a high phosphate capacity. The econo-
mics of this system, as developed in the laboratory, was determined.
An optimized system, the feasibility of which has not yet been proved,
is also discussed.
Operating and Design Conditions
Table 27 gives the general operating and design conditions for removal
of phosphate from one million gallons per day of secondary effluent. It
is assumed that the total phosphate will be reduced 98.4%, from 10 mg
P/l. to <0. 16 mg/1.
50
-------
TABLE 27
DESIGN CONDITIONS FOR PHOSPHATE REMOVAL FROM SECONDARY
EFFLUENT USING CATION EXCHANGER/METAL SALT SORBENT
(1,000,000 gpd Sewage Treatment Plant)
Resin Capacity, mg P/ml resin ©0.5 gpm/cu ft resin
Secondary Effluent Exhaustion Cycle Rate, gpm
Quantity Amberlite 200 Resin, cu ft
Effluent Treated at Resin Capacity, million gal.
Exhaustion Cycle, hours
Resin Depth, ft
Vessel ID, ft
Vessel Area, Total sq ft
Flow Rate, gpm/sq ft
Pressure Drop, psi
Backwash, @ 8 gpm/sq ft for 10 min, gal.
Regenerant, FeCl3 (as 100%), ton/cycle
Regenerant, gal. H^O/cycle
Lime (90% CaO), Ib/cycle
Regeneration Time, min/cycle @ 0. 5 gpm/cu ft
Rinse Water, gal/cycle @ 50 gal/cu ft
Rinse Time, min/cycle @ 1 gpm/cu ft
Feed Tank, gal. hold-up during clean-up
1.0
760
1, 520
1.0
22
6.1
12.5
245
3.1
1.5
19,600
1.63
19,800
2, 320
-30
75,000
50
168,000
(a)
(b)
(a) Assumes capacity of 2. 75 mg P/ml resin with residue of 1. 75
mg P/ml left on resin after regeneration.
(b) Two tanks required.
51
-------
Design conditions are given for a flow rate of 0. 5 gpm per cu ft Amber-
lite 200 resin. The 0. 5 gpm per cu ft resin rate provided zero phosphate
leakage under the experimental conditions with a resin capacity of 2. 75
mg P/ml resin.
The ferric chloride required for regeneration is still high for an eco-
nomical system. It would be expected, however, that additional experi-
mental work would significantly reduce the ferric chloride requirement.
The cycle, consisting of a long exhaustion period and short regenera- .
tion time, would be ideally suited to continuous countercurrent systems
that are offered in the ion exchange field. A single cell concept, how-
ever, was considered to be more economical wherein the flow through
the bed would be semi-continuous, requiring shutdown of the system
during backwash, regeneration, and rinsing periods. The system is
shown in the Flow Diagram, Figure 2.
The secondary effluent flows from the existing final clarifier to the
secondary effluent feed tank. During the exhaustion cycle, the secon-
dary effluent is pumped from the feed tank through the phosphate ex-
changer cell with the purified effluent being discharged to the product
water storage tank. (The feed tank also serves as an accumulator for
secondary effluent during backwash, regeneration, and rinse periods.)
The overflow from the product water storage tank becomes the essen-
tially phosphate-free sewage plant effluent.
The product water is used to backwash the phosphate exchanger. This
stream is then discharged to the final clarifier where any suspended
solids (removed from the top of the exchanger bed) settle out with the
solids contained in the secondary effluent.
Product water can also be used to make up the regenerant, ferric
chloride solution. This solution is circulated through the exchanger
and removes the accumulated phosphates. A side stream is removed
to the slurry tank where it is reacted batchwise with lime, resulting
in a precipitate of ferric hydroxide and hydroxylapatite. This slurry
is then pumped to a separate sedimentation tank from which the solids
can be removed for disposal.
Product water is also used for rinsing the resin bed. This rinse water,
contaminated with ferric chloride and other residues, is either dis-
charged to the slurry tank or it can be used as make-up water for fresh
ferric chloride solution.
Cost Estimate • --
Cost estimates were prepared for the system described above. The OSW
guideline (Reference 26) provided the basis and technique for these esti-
mates. A rough-order-of-magnitude capital cost estimate was made as
shown in Table 28. This capacity is not utilized in the proposed system,
52
-------
760 GPM
1, 000,000 GPD
PHOSPHATE
EXCHANGER
~ 1,000,000 GPD
SLUDGE
\
-T -
1
F,NAL
CLAR.FIER
TANK
EXHAUSTION
PRODUCT
WATER
STORAGE TANK
SEWAGE
PLANT
EFFLUENT
SLUDGE
A
^
PHOSPHATE
EXCHANGER
T1
19,600 GPD
FINAL
CLARIFIER
PRODUCT
WATER
STORAGE TANK
BACKWASH
19,800 GPD
FERRIC
CHLORIDE
FEED TANK
_ CALCIUM OXIDE
HOPPER AND FEEDER
PHOSPHATE
EXCHANGER
75,000 GPD
PHOSPHATE
EXCHANGER
SEPARATE
SEDIMENTATION
TANK
SLURRY
TANK
PRODUCT
WATER
STORAGE TANK
SLURRY
TANK
PRODUCt
WATER
STORAGE TANK
REGENERATION
RINSE
FIGURE 2. FLOW DIAGRAM - PHOSPHATE REMOVAL FROM SECONDARY
EFFLUENTS USING CATION EXCHANGER/METAL SALT SORBENT
-------
TABLE 28
CAPITAL COST ESTIMATE
Phosphate Exchanger Cell, 12. 5 ft dia x 10 ft high st. sides,
rubber-lined steel with piping, accessories, and
automatic controls (2 cells) $ 60, 000
Regenerant Tank, 30, 000 gal. , 23 ft dia x 10 ft high,
steel with baked phenolic resin lining and 30 HP
agitator 18,000
Pumps, 1000 gpm @ 60 psig, 316 SS, 25 HP motor 9,000
(2 required)
Ferric Chloride Storage Tank, 15,000 gal., steel with
baked phenolic resin lining 8, 000
Ferric Chloride Transfer Pump, 100 gpm @ 60 psig, 3l6 SS 3,000
Lime Hopper and Feeder 8,000
Secondary Effluent Feed Tank, 200, 000 gal. concrete 20,000
Product Water Storage Tank, 120, 000 gal. concrete 14,000
Slurry Tank, 40,000 gal. concrete with 60 HP agitator 20,000
Total Equipment Installed $160,000
Amberlite 200 Resin, 1520 cu ft @ $28/cu ft 42, 500
Total Estimated Fixed Capital „ . „ $202, 500
54
-------
however, since the ferric chloride requirement for total regeneration
is very high. Experimental conditions under which only partial regene-
ration was accomplished were chosen to reduce the ferric chloride cost.
Under these conditions, the spent resin containing 2. 75 mg P/ml is
treated with only enough ferric chloride to remove 40% of the phosphorus
from the resin. Thus, the net capacity of the resin is 1 mg/ml. This
results in a relatively large volume of resin which, however, is lower
in cost than the indicated quantity of ferric chloride needed for com-
plete regeneration.
Table 29 outlines the annual cost basis. It is assumed that the facility
would have a 30-year life and that financing would be done by issuing a
30-year municipal bond. The cost of borrowing money has been rising
rapidly during the last few years. Selection of a fair interest rate is
important, since the rate can be a substantial part of the total annual
operating cost. An average interest rate of 4-1/2% is assigned to this
project which is considered to be a conservative figure during the total
amortization period,, The total carrying charge factor consists of this
interest rate plus the corresponding sinking fund factor required to pay
for the equipment.
The capital costs and amortization of replaceable items and resin are
calculated as two separate items, since they are considered to have
shorter useful lives. Replaceable items are assumed to have a 10-
year life, while resin will have to be replaced every five years. The
appropriate sinking fund factor is applied to each category in addition
to the interest. Only interest is applied to the working capital, since
the working capital will be recovered at the end of the project life.
Derivation of working capital is shown in Table 30.
The estimated costs are summarized in Table 31. It will be noted that
the ferric chloride cost is the highest single cost item, contributing
about 13^/1000 gal. (Kgal) of effluent treated. In this case, the ferric
chloride usage is equivalent to an 8:1 molar ratio of ferric ion to
phosphorus.
It would be expected that additional experimental and development work
would result in cost reductions in an optimized system. Improvement
in resin capacity would reduce both capital cost and resin replacement
cost. Improvements in regeneration would reduce ferric chloride costs.
For example, reducing the molar ratio of ferric ion to phosphorus to
2:1 would reduce costs by more than 9^/1000 gal. effluent treated.
With the various improvements that might be possible, an optimized
system could probably operate at a cost approaching 20^/1000 gal.
effluent.
55
-------
TABLE 29
ANNUAL COST BASIS
Plant Capacity Factor: 90%, 330 days/year; 7920 hours/year.
Carrying Charges
Lifetime, years
Interest, %
Sinking Fund Factor, %
Carrying Charge Multiplier, %
Recurring Costs
Taxes: 2% fixed capital
Insurance: 1% fixed capital
Labor
Plant Replaceable Items Resin
30 10 5
5.0 0.25 0.25
1.505 10.00 20.00
6.505 10.25 20.25
$ 4,050/year
$ 2,030
Operating: 8 hr/cycle @ $4. 00/hr
Supervisory: 15% of operating labor
Maintenance: 2. 5% of fixed capital (-resin)
Total ..........
Supplies and Maintenance Materials
2% of fixed costs (-resin)
General and Administrative
Payroll Burden, 20% of total labor
Plant Overhead, 50% of total labor, maintenance
and supplies
$ 8,000/year
1,200/year
4,000
$13,200
$ 3, 200/year
$ 2,640/year
$ 8,200/year
56
-------
TABLE 29 (Continued)
Chemicals
Ferric Chloride (as 100%), 538 tons @ $80/ton
Lime, 383 tons @ $18/ton
Resin 200, 304 cu ft @ $28/cu ft
Utilities
Power: 265,000 kwh @ 10 mills
Replaceable Items (10-year life)
Pumps
Agitators
Feeder
Automatic Controls
$12,000
9,000
3,000
5,000
$43,040/year
$ 6,890/year
$ 8,510/year
$ 2,650/year
Total $29, 000
or
$ 2,900/year
57
-------
TABLE 30
WORKING CAPITAL
Dollars
Chemical Inventory, 2 months $ 9, 750
Labor, 3 months 3, 300
Payroll Burden, 3 months 660
Plant Overhead, 4 months 2, 730
Fixed Cost, 0. 5% of fixed capital 1, 010
Spare Parts and Miscellaneous, 1% of fixed capital 2, 020
Total ....„.„ $19,470
58
-------
TABLE 31
COST SUMMARY
Phosphate Removal System for 1, 000, 000 gpd Secondary Effluent
Capacity Factor: 0.90 Output: 330, 000 K gal/yr
Interest Rate: 4-1/2% Plant Life: 30 years
CAPITAL COSTS
1. Fixed Capital (30 -yr bond)
2. Replaceable Items (10 yr)
3. Resin (5 yr)
4. Working Capital
Total Capital Costs
RECURRING COSTS
5. Taxes
Capital Cost
Dollars
131,000
29,000
42,500
19,470
221,970
6. Insurance
Total Annual Fixed Costs
OPERATING AND MAINTENANCE COSTS
7. Operating Labor
Carrying
Charge
0.0614
0.0814
0.1828
0.0450
%
2
1
8. Supervisory Labor
9. Maintenance Labor
10. General and Administrative
11. Supplies and Maintenance Materials
12. Ferric Chloride
13. Lime
14. Resin
1 5. Pov/er
16. Other
Total Operation and Maintenance Costs
TOTAL FIXED PLUS O&M COSTS
Annual
Cost, $
8,040
2, 360
7, 770
880
4,050
2,030
25, 130
8,000
1,200
4,000
10,840
3, 200
43,040
6,890
8,510
2,650
-
88,330
113,460
.
Treating Cost
£/K gal
2.44
0. 72
2.35
0.26
1.22
0.62
7.61
2.42
0.36
1.21
3.29
0.97
13.04
2.09
2.58
0.81
26. 77
34. 38 ]
%
7,1
2.1
6.8
0.8
3.5
1.8
22.1
7.1
1.0
3.5
9.6
2.8
37.9
6.1
7.5
2.4
77.9
00.0
59
-------
SEC TION IV
EXPERIMENTAL PROCEDURES
SORBENT SYNTHESIS
Resin Sulfonyl Chloride
Dowex 50W-X8 (H) (50 g wet weight, 46% solids) was stirred with 200
ml of 50% (vol) H2SC>4 solution at room temperature for 1 hr. The acid
was removed with a filter stick and the procedure was repeated succes-
sively with 80%, 95%, and 100% H2SO4. The beads were placed in a
200-ml 3-neck flask fitted with a paddle stirrer, thermometer with
Therm-o-Watch control (I R Company, Cheltenham, Pennsylvania),
and drying tube. CISC^H (150 ml) was added (some gas evolution) and
the mixture was stirred for 1 hr. The acid was replaced with fresh
CISC^H and the mixture was stirred overnight at room temperature
and then at 80° for 6 hr. The contents were cooled and the liquid re-
moved with a filter stick. The beads were washed on the filter (no
suction) with the following: chloroform; dioxane (and overnight standing);
acetone (and 1-hr standing); ice water; and 5% Na2CC>3 (0°). The beads
were then soaked at 0° for 1 hr, first with 5% Na2CC>3 (the final pH was
alkaline), and then with water (neutral pH). After standing for 1 hr in
acetone, the beads were air-dried, and then warmed for 3 hr at 40°
(3 mm). The dark brown, unfractured beads remained odor-free and
unchanged in appearance during 3 weeks of surveillance, at which time
a faint HC1 odor was detected.
The extent of conversion of RSC^H to RSC^Cl was determined by ele-
mental analysis for chlorine and by direct titration of RSCuCl. The
former gave 14.6% Cl (theory, 16. 3), corresponding to an o9. 5% con-
version. The latter (performed by immersion of the sample in excess
0. 1 N NaOH for 3 days, followed by back titration with 0. 1 N HC1) indi-
cated a chlorine content of 16.0%, or a 98. 3% conversion. The smaller
(elemental analysis) value was preferred because of the known impre-
cisions of the titration procedure.
Resin Carbonyl Chloride
In a 300-ml 3-neck flask fitted with a paddle stirrer, thermometer, and
reflux condenser topped by a drying tube were placed 10. 9 g (0. 12 eq)
of dry IRC-50 resin and 90 ml of dry toluene. The mixture was stirred
at 90-95° for 1. 2 hr, cooled to room temperature, and 46. 3 g (0. 22
mole) of PClr was added. The contents were stirred at 110° for 20 hr
(HC1 evolution was noted). The beads were washed repeatedly with
toluene and then with methylene chloride (stirring, soaking for 1 hr,
and decantation). After removal of solvent in vacuo and drying for
4 hr at 50° (3 mm), 11.2 g of intact, light tan beads were obtained,
which did not develop an odor of HC1 on standing.
61
-------
The resin contained 10.0% Cl (theory, 32.4), and the presence of
strong absorptions at 1820 and 1760 cm in the infrared indicated
a significant anhydride content. The apparent discrepancy between
the slight weight increase and the significant chlorine content is
probably due to a weight loss (such as organic sloughage) occurring
during the reaction.
ES-1 Copolymer
In a 500-ml resin kettle fitted with a stainless steel stirrer (3-in.
propeller), reflux condenser, nitrogen inlet tube extending below
the liquid, and oil bath, were placed 300 ml of water, 2.0 g of poly-
vinyl alcohol (Matheson, Coleman and Bell — 99% hydrolyzed; vis-
cosity of 4% solution at 20°, 28-32 cps), and 1.0 g of magnesium
silicate. The mixture was stirred at 85° (oil bath temperature,
Matheson Lab-Stat temperature control) for 30 minutes while nitrogen
was bubbled through at a fast rate. The heat was removed, and when
the temperature had dropped to 60°, the nitrogen inlet tube was re-
placed with static nitrogen pressure applied to the top of the condenser.
A solution of 0. 5 g of benzoyl peroxide in 50 ml of distilled styrene and
0.8 g of technical DVB (62% DVB, 38% ethylstyrene) was added, the
stirrer speed adjusted to 800 rpm, and the suspension stirred at 85°
(oil bath temperature) for 6 hr. Stirring was continued while the mix-
ture cooled. After the decantation of some floating granules, the beads
were strained through a 100-mesh sieve, washed three times with water
and three times with methanol, and left to stand in methanol overnight.
The methanol was removed by suction, and the beads were dried in air,
and then for 3 hr at 45° (3 mm). The yield of hard, almost transparent,
beads was 25 g, 80% of which were of 45-100 mesh and 20% of 20-45
mesh.
ES-1 Chloromethyl Resin
In a 100-ml 3-neck flask fitted with paddle stirrer, reflux condenser,
and drying tube, were placed 4 g of ES-1 copolymer (45-100 mesh),
40 g of chloromethyl methyl ether, and 3 g of ground, anhydrous zinc
chloride (previously melted and cooled in a desiccator). After being
stirred for 1 hr at room temperature and for 2 hr at reflux, the beads
were filtered off, washed with methylene chloride, and left overnight
in that solvent. They were then rinsed with cold water and finally with
acetone. After drying in air and for 3 hr at 45° (3 mm), 5. 24 g of in-
tact, light yellow beads were obtained. The chlorine content was 27.2%,
indicating the substitution of 1.15 CH->C1 groups per aromatic ring.
Sorbitol Sodium Salt
A 500-ml, three-neck flask fitted with mechanical stirrer and drying
tube was flamed out, flushed with nitrogen, and cooled with a dry ice-
acetone bath. Into the flask were introduced 250 ml of liquid ammonia,
and after replacing the drying tube with a static nitrogen pressure, 25
g (0. 14 mole) of sorbitol and 3. 2 g (0. 14 g-atom) of sodium. The dark
62
-------
blue mixture soon decolorized. After 15 min stirring, the cooling bath
was removed, and most of the ammonia was driven off by warming the
stirred contents of the flask (hot-air gun) for 3. 5 hr. At this time, a
Nessler test was strongly positive (orange color and precipitate), al-
though ammonia was not detectable by odor or by reaction with HC1
vapor. The white powder was heated at 120° (3 mm) for 3 hr. The
fractured plates (28.0 g) obtained from the molten layer of sorbitol
sodium salt gave a weakly positive Nessler test (faint yellow color).
Studies by Chablay (Reference 27) indicate that the reaction yields
the mono sodium salt of the polyol.
Resin Pyrogallyl Sulfonate
In a 250-ml flask equipped with paddle stirrer, condenser, drying tube,
and thermometer were placed 4. 34 g (19. 5 mmoles) of D50W sulfonyl
chloride (15.8% Cl, 97.1% converted), 3.78 g (30 mmoles) of pyro-
gallol, 2 ml (25 mmoles) of pyridine, and 45 ml of dry DMF. The
mixture was stirred for 1 hr at room temperature and then heated to
75° over a period of 30 min. Stirring was continued for 4. 5 hr at 75
to 80°. After standing overnight, the contents were added to an ice/
dilute HC1 mixture. The beads were washed by repeated swirling and
decantation (30 min) until no odors of pyridine or DMF could be detected.
The beads were soaked and rinsed with several portions of acetone (40
min) to remove water. After removal of most of the acetone at the
water pump, the product was dried at 45 to 50° (3 mm) for 3 hr. The
weight gain was 1.01 g, 58. 1% of theoretical. The product contained
10. 90% S; the calculated value, for a 58. 1% converted product was
11.95%. Qualitative tests for chloride were weakly positive. (Later
preparations of the pyrogallyl sulfonate, similarly performed, con-
tained only 0.5 to 0.8% Cl). Strong infrared absorptions at 1200, 1630,
1540 and 1485 cm"-'- indicated the presence of a catechol derivative.
Resin Mannityl Sulfonate (DMF/Triethylamine Method)
In a 50-ml Erlenmeyer flask fitted with magnetic stirrer, condenser,
and drying tube were placed 2. 5 g (10.5 meq) of D50W sulfonyl chloride
and 20 ml of dry DMF. After stirring at room temperature for 1 hr,
2. 1 g (11.5 meq) of mannitol and 5 ml triethylamine were added. The
mixture was then stirred at 100° for 4 hrs. The warm solution was
decanted; the beads were washed with water, transferred to a column,
and further washed by the passage of 200 ml of water (2 hr) and 300 ml
of acetone (2 hr). The product was dried in air and then at 40 for 3 hr
(3 mm). The beads weighed 3.1 g, and contained 10. 7% S, 2.0% Cl, and
4. 6% N. The infrared spectrum had strong OH and CH2 absorptions
(mannitol) and a trace of carbonyl (DMF). When the beads were rinsed
with a 2% NaOH solution, a strong amine odor was detected. Although
the weight of the product corresponded to a 41% conversion to the manni-
tyl ester, the actual conversion was somewhat less, because of the pres-
ence of the amine, as sulfonate salt or sulfonamide, and the small
amount of DMF.
63
-------
Resin Mannityl Sulfonate (Mannitol Sodium Salt Method)
In a 100-ml, 3-neck flask fitted with paddle stirrer and nitrogen inlet
were placed 25 g of mannitol, 3 g (15 meq) of ES-1 sulfonyl chloride,
4. 6 g (22.5 mmoles) of mannitol sodium salt, 2. 2 g (15 mmoles) of
sodium iodide, 8 ml of bis [2-(2-methoxyethoxy)ethyl] ether, and 10
ml of toluene. The reaction mixture was stirred at 180° overnight
(Matheson Lab-Stat control) under a stream of nitrogen. The mixture
was then cooled to 100°, the polyol dissolved in hot water, the beads
separated, washed with water, stirred in water for 1 hr, washed with
acetone, air dried, then vacuum dried for 3 hr at 45° (3 mm). The
yield of dark-brown, intact beads was 3. 31 g, or 14.4% of theory based
on the weight increase.
Resin N-Methylglucamyl Sulfonamide
In a 100-ml 3-neck flask fitted with paddle stirrer, reflux condenser
and drying tube were placed 5 g (19.8 meq) of A200 sulfonyl chloride,
8. 5 g (43. 5 meq) of NMG, 40 ml of dioxane, and 4 ml of water. After
stirring at reflux temperature for 3 hr, the beads were thoroughly
washed with water, stirred in 2% HC1 solution for 1 hr, rinsed with
water, immersed in water for 1 hr, and rinsed with acetone. After
standing in acetone for 1 hr, the beads were dried in air and then for
3 hr at 45° (3 mm). The grey beads weighed 6.48 g; the gain in weight
(1.48 g) was 47.2% of theory. Analysis indicated 1.07% Cl and 2.28%
N; the theoretical nitrogen content for 100% conversion was 3.40%.
Resin N-Methylglucamyl Carboxamide
In a 250-ml flask equipped with paddle stirrer, condenser, drying tube,
and thermometer were placed 3.0 g of IRC-50 chloride/anhydride, 9.0
g of NMG, and 40 ml of DMF. The mixture was stirred at room tem-
perature for 1.5 hr, permitted to stand overnight, and then stirred for
10 hr at 90-105°. The supernatant was decanted and the product was
stirred with several portions of 2% HC1 solution (25 min). The acid
was removed by repeated washing with water. The beads were soaked
in several portions of acetone to remove water, dried at the water pump,
and then warmed at 50° (3 mm) for 3 hr. The weight gain was 2. 57g,
or 59. 2% of theoretical. Analysis indicated 3. 92% N (theory, for a
completely reacted product, 5. 22% N). No residual chlorine was
present.
Resin Sorbityl Ether
In a 100-ml 3-neck flask fitted with paddle stirrer and nitrogen inlet
were placed 25 g of sorbitol, 3. 3 g (25 meq) or chloromethylated ES-1
copolymer, 8 ml of bis(2-methoxyethyl)ether, 10 ml of toluene, 4. 1 g
(27. 5 mmoles) of sodium iodide, and 6. 5 g (32 mmoles) of sorbitol
sodium salt. The reaction mixture was stirred under a stream of
nitrogen at 130° for 16 hr. The mixture was cooled to 100° and the
64
-------
solids were added to hot water. The beads were successively washed
with water, stirred in water for 1 hr, washed with acetone, and air
dried. After further drying for 3 hr at 45° (3 mm), the yield of intact
yellow beads was 5.0 g, or 41. 3% based on the weight increase. The
infrared spectrum contained strong absorptions in the ether region
(1040-1100 cm'1).
Re sin/Metal Sorbents (Batch Process)
To 11.5 ml (wet volume, 22 meq) of A200/Na was added 500 ml of 0. 1
M aqueous ferris chloride solution. The suspension was stirred gently
for 2 hours at room temperature, and the solution was decanted. The
beads were washed several times with water, by decantation, and were
stored wet. Analysis of one such preparation (after drying) showed that
the beads contained 0. 20% Cl and 10. 0% Fe.
Resin/Metal Sorbents (Drip Process)
A 2. 3-ml portion (wet volume, 4. 4 meq) of A200/Na was placed in a
20 x 45-mm glass funnel equipped with a coarse fritted disc. The resin
was washed several times with deionized water and covered with a disc
of filter paper. A fresh 0.2 M solution of ferric chloride was prepared
by dissolving 2. 7 g (0.01 mole) of FeCl3- 6 HzO in 50 ml of water. The
ferric chloride solution was added from a dropping funnel at 1. 4 + 0. 1
ml/min. The sorbent was washed thoroughly with water on the filter
and stored wet until used.
Resin/Metal Sorbents (Column Process)
A 3. 0-cm (ID) chromatographic column equipped with a teflon stopcock
was plugged with a small amount of glass wool. Enough 1/8-in. beads
were added to reach a depth of 1 -in. , and water was added to a depth of
6- to 8-in. A slurry of 46 ml (wet volume, 88 meq) of A200/Na in water
was poured in, and most of the water was drained off. The resin beads
were classified by backwashing with water, at a rate which produced at
least a 50% expansion of the resin bed. After draining the water to a
depth of 2 to 5 cm above the resin, a plug of glass wool was placed in
the column approximately 1 cm from the top of the resin bed. A drop-
ping funnel was connected to the column by means of plastic tubing and
a rubber stopper. A freshly prepared solution of ferric chloride (1.0 1. ,
0. 2 M) was added from the dropping funnel at a rate of 3.0 + 0.2 ml/min
by adjusting the column stopcock to that rate with the dropping funnel
stopcock in the open position. After addition of ferric chloride, the
sorbent was washed with 500 ml of water at the same rate, followed by
2 to 3 1. of water over a period of 30 min. The sorbent was stored
under water until analyzed.
65
-------
SORBENT EVALUATION
Phosphorus Capacity (Test Procedure 1) — (Used for the evaluation of
XE-243 and resin/metal sorbents)
All glassware was rinsed twice with warm 5% HC1 solution and with
distilled water prior to use. An amount of wet sample corresponding
to Z g of dry resin (for XE-243) or 1 g of dry resin (for resin/metal
sorbents) was placed in a 20 x 45-mm Pyrex funnel equipped with a
coarse fritted disc. After repeated washings with distilled water, the
resin was covered with a disc of filter paper and standard Na2HPC>4
test solution (8.16 mg P/l. ; 500 mg NaCl/1. ) was added dropwise from
a calibrated dropping funnel at 1.4 + 0. 1 ml/min. Effluent was collec-
ted in 50-ml aliquots until the breakthrough concentration of 0. 326 mg
P/l. was reached. Phosphorus analyses were made with the standard
stannous chloride-ammonium molybdate procedure (Reference 23). De-
termination of the phosphorus concentrations in the effluents permitted
the calculation of the amount of phosphorus retained in the sorbent.
Regeneration was accomplished by first rinsing the resin bed to remove
adhering test solution and then adding the regenerant (100 ml of 5%
H2SO4 solution for XE-243 samples and 200 ml of 0. 1 M FeCl3 solution
in the case of resin/metal sorbents) at the dropwise rate of 1.4 + 0. 1
ml/min. The stripped sorbents were rinsed with water prior toTurther
cycling.
Phosphorus Capacity (Test Procedure 2)
For the evaluation of polyol sorbents, the above procedure was modified
as follows: (a) 2 g (dry weight) of sample was taken; (b) the rate of addi-
tion of test solution was 1.0 + 0.1 ml/min; and (c) 25-ml aliquots of ef-
fluent were collected. Regeneration studies were not made because of
the low capacity of these sorbents.
Phosphorus Capacity (Procedure 3)
Sorbent columns were prepared and classified as described in the Col-
umn Process for sorbent preparation. For 4. 6-ml samples, burets
having an ID of 0.9 to 1.0 cm and equipped with Teflon stopcocks were
used. Columns of 2.5 and 3.0-cm ID were used for larger scale work.
Phosphate solution, 8. 16 or 25 mg P/l. , was added from a reservoir
connected to the column by plastic tubing. Effluent was released from
the column at 0. 25 ml/min for the burets, 2. 1 ml/min for the 2. 5-cm
column, and 3.0 ml/min for the 3.0-cm column. Effluents were col-
lected in suitable aliquots for phosphorus analysis.
66
-------
SECTION V
ACKNOWLEDGMENTS
This program was performed by the Envirogenics Company, a Division
of Aerojet-General Corporation, El Monte, California, under the direc-
tion of Dr. Richard A. Dobbs, Project Officer, Water Quality Office.
Envirogenics Company personnel participating in the program were
Dr. L. M. Soffer, Program Manager; Dr. B. I. Loran; Mr. J. R.
Lowell, Jr.; and Mr. L. E. Gressingh. The advice and guidance
offered by Mr. W. P. Knight and Dr. B0 J. Mechalas of the Enviro-
genics Water Quality Department throughout the program is gratefully
appreciated and acknowledged.
67
-------
SECTION VI
REFERENCES
1. Kuentzel, L. E. , J. Water Pollut. Contr. Fed. , 41 (10), 1737
(1969).
2. Eliassen, R. , and Tchobanoglous, G. , "Removal of Nitrogen and
Phosphorus," Paper presented at the 23rd Purdue Industrial
Waste Conference, Purdue University, Lafayette, Indiana,
May 8, 1968.
3. Hatfield, W. D. , Letter to the Editor, Environ. Sci. Technol. ,
2_ (6), 392 (1968).
4. Oswald, "W. J. , Crosby, D. G. , and Golueke, C. G. , "Removal
of Pesticides and Algal Growth Potential from San Joaquin Valley
Drainage Waters (A Feasibility Study)." Report to the San Joa-
quin District, California State Department of Water Resources
(1964).
5. Nesbitt, J. B. , "Removal of Phosphorus from Municipal Sewage
Plant Effluents," Engineering Research Bulletin B-93, College
of Engineering, Pennsylvania State University, February 1966.
6. Bureau of Sanitary Engineering, Waste Water Reclamation,
California State Department of Public Health, Berkeley, Cali-
fornia, November 1967.
7. Albertson, O. E. , and Sherwood, R. J. , Ind. Water Eng. , 4(11),
30 (1967). ~
8. Anon., Environ. Sci. Technol. , 2_ (3), 182 (1968).
9. Cecil, L. K. , "Progress and Practices of Phosphate Removal in
Water Reuse, " Chem. Eng. Progr. Symp. Ser. , 63(78), 159
(1967).
10. Gulp, R. L. , J. Water Pollut. Contr. Fed. , 35^(6), 799(1963).
11. Yee, W. C., J. Am. Water Works Assoc., 5_8 (2), 239-247(1966).
12. Eliassen, R. , and Tchobanoglous, G. , J. Water Pollut. Contr.
Fed., 40_, R176 (1968).
13. Kyrides, L. P., J. Am. Chem. Soc. , 59, 206 (1937).
14. Helfferich, F. , "Ion Exchange, " McGraw-Hill Book Co. , Inc.,
New York, 1962, p. 53.
69
-------
15. Small, H. , Ind. Eng. Chem. , Prod. Res. Develop. , 6 (3), 147
(1967). ~
16. Hatch, M. J. , U.S., Patent 3, 280, 046, October 18, 1966;
M. J. Hatch, U.S. Patent 3, 300,416, January 24, 1967.
17. Rogozhin, S. V., etal. , Vysokomol. soyed. A10 (6), 1277
(1968).
18. Singley, J. E. , and Black, A. P., J. Am. Water Works Assoc. ,
59(12), 1549-1564 (1967); Singley, J. E. , and Sullivan, J. H. , Jr.
TFid., 6M4), 190-192(1969).
19. Stumm, W. , and Morgan, J. J. , ibid., 54(8), 971-992(1962);
Sullivan, J. H. , Jr., and Singley, J. E.~ibid. , 60 (11), 1280-
1287 (1968).
20. Amberlite 200 Technical Bulletin, Rohm and Haas Co., Phila-
delphia, Pa., June 1970.
21. Lyman, W. R. , and Preuss, A. F. , U.S. Patent 2, 813, 838,
November 19, 1957.
22. Kun, K. A., and Kunin, R. , J. Polymer Sci. Cl6 1457 (1967);
Werotte, L. E. , and Grammont, P. D. , U. S. Talent 3, 418, 262,
December 24, 1968.
23. Standard Methods for the Examination of Water and Wastewater,
12th edition, American Public Health Association, Inc., New
York, 1965, p. 234.
24. Pierce, W. C. , and Haenisch, E. L. , "Quantitative Analysis, "
3rd edition, John Wiley and Sons, Inc., New York, 1956, p. 487.
25. Dryden, F. D. , and Stern, G. , Environ. Sci. Technol. , 2, 268
(1968). ~
26. Office of Saline Water, "Guideline for Uniform Presentation of
Desalting Cost Estimates," R&D Program Report No. 264,
July 1967.
27. Chablay, M. E. , Ann, de Chim. Phys. , 9 (8), 165 (1917).
70
-------
SECTION VII
GLOSS ARY*
A200 Amberlite 200, sulfonic acid resin, 20% DVB
(Rohm and Haas Co.)
D50W Dowex 50W-X8, sulfonic acid resin, 8% DVB
(Dow Chemical Co.)
DMF N, N-Dimethylformamide
DMSO Dimethyl sulfoxide
DVB Divinylbenzene
ES-1 Styrene-Divinylbenzene copolymer, 1% crosslinked
ES-1S ES-1 sulfonic acid resin
IR-118 Amberlite IR-118, sulfonic acid resin, 5% DVB
(Rohm and Haas Co.)
IR-120 Amberlite IR-120, sulfonic acid resin, 8% DVB
(Rohm and Haas Co.)
IR-124 Amberlite IR-124, sulfonic acid resin, 12% DVB
(Rohm and Haas Co.)
IRA-400 Amberlite IRA-400, trimethylammonium chloride resin,
8% DVB (Rohm and Haas Co.)
IRC-50 Amberlite IRC-50, carboxylic acid resin (Rohm and Haas Co.)
IRC-84 Amberlite IRC-84, carboxylic acid resin (Rohm and Haas Co.)
NMG N-Methylglucamine
SDVB Styrene-divinylbenzene copolymer
XE-243 Amberlite XE-243, N-methylglucamide resin, 1% DVB
(Rohm and Haas Co.)
'•'Mention of commercial products does not imply endorsement by the
Environmental Protection Agency.
71
-------
SECTION VIII
APPENDIX
INTERACTION OF PHOSPHATE WITH ORGANIC POLYOLS
The following mechanism has been advanced for the complexation of
borate or boric acid with polyhydroxylic compounds such as mannitol,
catechol, or pentaerythritol (Reference 1):
B40?" + 7H2O N 4H3B03 + 2OH
H,BO, + H9O N B(OH) ~ + H+ (Reference 2)
3 j £ ^
^v
1 I \
B(OH) ~ + HO-R-OH N R XB~ + 2H-O
/ \
O OH
. O. .OH „ , O
I \ / K2 v ' \
R XB + HO-R-OH N R XB" R + 2
• O OH I O
-O ;OH , O
\ / .H-1- ^ R \
•^ , B-OH + H90
\ X 2
o' OH I cr
Thus, the tetrahydroxyborate species, B(OH). , which exists in dilute
solutions ( 0.25 Molar) of borates, may form as many as two different
anionic complexes and one molecular complex with suitable polyhydroxy-
lic compounds, under the proper conditions.
The above mechanism may be applied to the dihydrogenphosphate
(H?PO,~) and monohydrogenphosphate (HPO4 ) species as shown below.
73
-------
R
-OH
-OH
H0
\
HO'
O
-O
R
O
R
-OH
-OH
HO
O
O
I —
°
OH O'
O
O
Its application to polyphosphate ions and to the tribasic species (PO. )
is not so easily rationalized, although it is possible that the latter,
which is known to undergo sorption on XE-243, does so as the dibasic
species with which it exists in equilibrium.
HOH
HPO
2-
+ OH
R-CH7-N{CH,)-C,HQ(OH)[. + HPO,
LJ j o o D 4
2-
R-CH2-N(CH3)-C6Hg(OH)4(OP03)
2-
74
-------
REFERENCES
1. Steinberg, H. , "Organoboron Chemistry, " Volume 1, Inter-
science Publishers, New York, 1964, pp 675*696.
2. Cotton, F. A. , and Wilkinson, G. , "Advanced Inorganic
Chemistry," Interscience Publishers, New York, 1966,
pp 262-264.
75
-------
1
Accession Number
w
5
2
Subject Field & Group
O5D
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
Envirogenics Co., Division of Aerojet-General Corporation, El Monte, California
Title
Investigation of a New Phosphate Removal Process
10
Authors)
Soffer, L. M.
Lowell, J. R., Jr.
Lor an, B. I.
16
Project Designation
17010 DJA
21
Note
22
Citation
23
Descriptors (Starred First)
*Phosphates, *Sorption, * Tertiary treatment, Cation exchange, Economic
feasibility, Eutrophication, Resins, Sewage treatment, Water reuse
2*) Identifiers (Starred First)
27
Abstract
A laboratory evaluation was made of the technical and economic feasibility of
phosphate removal from secondary effluent by two types of new sorption resins.
Multivalent metal derivatives of sulfonic acid resins were found to have good
phosphorus capacities. The iron (III) form of Amberlite 200 strong acid cation
exchange resin, for example, exhibited capacities up to 9. 5 mg P/ml of wet
resin, and, in tests with secondary effluent, a capacity of 2. 7 mg P/ml was
obtained. Regeneration of the exhausted resin was readily accomplished with
dilute ferric chloride solutions; ten exhaustion-regeneration cycles were per-
formed on one sample without loss of phosphorus capacity. Iron and phosphorus
were quantitatively removed from spent regenerant solutions by the addition of
lime and removal of the resulting precipitate by filtration. With a second type
of resin, phosphorus capacities only 25 to 30% of the commercial polyol Amberlite
XE-243 were obtained. These resins included (1) polyhydroxylic esters and amides
of various cation exchange resins, and (2) polyhydroxylic ethers of styrene-divinyl
benzene copolymers. On the basis of the laboratory results, a phosphorus removal
process was designed and a cost estimate prepared. (Wilson - Envirogenics)
Abstractor
E.
M.
Wilson
Institution
Envirogenics
Co.,
El
Monte,
California
WR:102 (REV JULY 1969)
WRSIC
SEND, WITH COPY OF DOCUMENT
TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON, D. C. 20240
* GPOl 1970 — 389-930
------- |