The Reversible Removal of Naturally Occurring Organics Using

Sodium Chloride Regenerated Ion Exchange Resins

By Michael Gottlieb, Published in Watertech, 11/10/93.

soon see, high water retention increases the removal ca-

This is the first in a series of papers on the use of strongly pacity and lowers selectivity.
basic anion exchange resins for the control and removal
of naturally occurring organic substances from water. Lowering the selectivity coefficient of the organics against
the regenerant increases the relative strength of the re-
My specific goals in this presentation are to show the generant. For example, the higher water retention level
mechanisms affecting organics behavior, give preliminary of Type 1 porous gel resins gives them about 20% higher
performance estimates for resin organic interactions, and operating capacity for chlorides then the standard Type 1
to document the major steps in development of the knowl- gel resins in demineralizers regenerated at the same
edge that made this presentation possible. Therefore, a dose level of NaOH, even though the porous type has
good deal of this presentation will be historical in nature. 15% lower total capacity(1). This is because the Cl/OH
selectivity coefficient is about 50% as great in the porous
Naturally occurring organic substances have long been resin, which makes the OH ion a more effective regener-
recognized as being primarily aromatic hydrocarbons with ant against chloride.
carboxylic groups. For many years now it has been
known that these substances can be removed by ion ex- In the case of organic ions, kinetic factors have a much
change resins and that the ion exchange phenomenon greater impact on performance. They are at least as im-
played an important part but not the complete role in their portant as regenerant selectivity and dose level in de-
removal by resins. In an effort to explain the complex termining the operating capacity, by the regeneration
behavior of these substances other mechanisms were process.
assumed to be involved. This was changed in a land-
mark effort by Symons11 in a paper presented at the 1992 Let us review now what we know about "organics" and
International Water Conference in Pittsburgh, PA. He the events leading up to this presentation. Figure 1
showed that ion exchange was the only mechanism in- shows the structure of the colored portion of the humic
volved for the removal of the vast majority of TOC sub- acid molecule according to Gjessing4. Given that it is well
stances from natural sources, except for those with low known that the source of organics in surface waters are
molecular weights. Not only is the concept of an ex- vegetative, from decaying plant life. The sources of plant
clusive ion exchange mechanism new, so is the idea that life include grasses, leaves, decaying tree limbs and soil
the low molecular weight and not the opposite can pass humus. Humus is created from decaying roots and wood
through the resin bed, if the resin has sufficient gel phase stock from plants. It is the "spongy" material in the earth
porosity. and is known to have long lives and very high molecular
weights.
The presence of organic substances in influent waters
generally causes problems in demineralizers such as,
long rinses, poor quality, and reduced anion capacity.
Because of Symons results, we now know that ion ex-
change kinetics and equilibrium relationships are the ar-
COOH
eas to be investigated. This means that resin characteris- OH
O
tics must also be correlated to these relation- HO OH H3CO
O O
COOH C=O C COOH
ships. Today we can say that the mechanism for the re- (C)
O
(C)X (C)X
O

(C)X (C)X (C)X O

X
H3CO
moval of naturally occurring organics by ion exchange OH OH OH OH
OH

OCH3 OCH3 OCH OH

=

3 OH
resins, as measured by the TOC test, is at least under- OH OH O O
O
O

stood, if not well defined. N

The naturally occurring organics are removed as ions.

They have very slow diffusion rates. They are weak acids Figure 1 Proposed structure of colored humus molecule
and therefore their ionic nature is affected by pH, tem-
perature, and concentration. Since they have slower dif-
fusion rates, contact time, especially during regeneration,
is an important factor, equally as important as the
Humus is acidic in nature. In concentrated form, organic
regeneration dose.
acids act as biological preservatives. A few years ago,
bodies were discovered that had been buried in peat
Later I will show that in order to regenerate organic laden
bogs for over 100 years (which are primarily humic mate-
resins sufficient contact time is as important a factor as
rials), they were found to still have their last meals undi-
dose. Also, the higher the relative amount of water in the
gested in their stomachs.
gel phase the less critical the time factor. As we shall
Once these substances dissolve in water, oxygen and
sunlight oxidize them. Oxidation can cleave the carbon
bond in the benzene ring. This reduces the overall size
and molecular weights of the organic molecules.
3-2n

The molecular weight of fresh aquatic humus has been COO

extensively studied by several workers and summarized HUMUS n Fe
by Gjessing4. The molecular weights of humic acids
range from 700 to 80,000+; fulvic acids from 200 to 300 O
for soluble species. Significant amounts of humic sub-
stances also exist as colloids which, when tested by light
scattering and ultra centrifuge techniques, gave molecular
weight values of 30,000 to 80,000. Therefore we know Figure 2 Proposed linkage with humus and iron
that the starting molecular weight of the organic materials
in water can average well into the tens of thousands. Yet
as we will see later, the average molecular weight of oxidation, are two of the common methods used for re-
those substances in water comprising TOC are usually moving this form of iron. Many of the carboxylic groups
well below 10,000. on the humic acid molecules remain ionized, therefore
they are free to interact with and be removed successfully
This may vary depending on climate conditions such as by special purpose ion exchange resins, one of which I
temperature, the amount of sunlight, oxygen concentra- will discuss later. The presence of heme iron is the rea-
tion, water color, pH, and how long it’s been in the water. son why most organically fouled anion resins also contain
iron. They can also have copper, aluminum, zinc, man-
Once the concentrations of the organic acids are reduced ganese, etc. which also form complexes with humic acids.
by dilution, microbiological processes can begin. These
are aided by the availability of oxygen and other nutrients. Symons also proved that nonionic adsorption played little
Microbiological attack is a form of oxidation. It also cre- or no role in the removal of these humic substances ex-
ates various chemical groups such as carboxylics, alde- cept for those of molecular weight less than 1000, which
hydes, ketones, and hydroxyl to name some. Since these coincidentally, were not well removed by the ion ex-
materials started out as plant life, they include other ele- change resins either as they are apparently unionized.
ments besides carbon and oxygen. They also have nitro- Many of these also passed through the adsorbents and
gen, phosphorous and the whole range of elements as- carbon indicating they may be alcohols, ketones or alde-
sociated with life. Figure 1 is just one portion of the hu- hydes.
mic acid group of chemicals found in surface waters.
Symons' experiments were conducted in two stages. The
Surface waters contain several other classes of humus first stage was batch equilibrium studies using R.O. and
derived substances, including fulvic acids, lignins, and U.F. to concentrate and fractionate the organics from
tannins. These are only the major constituents from plant Lake Houston water and dialysis to remove inorganic
sources. All of the various organic classes are similar in ions. His tests were conducted with several different ion
that they are based on multiple benzene rings and have exchangers, which covered a wide variety of types and
significant carboxylic acid functionality. The biggest dif- were manufactured by a single supplier. The chloride
ferences are molecular length and width. They also con- form of the resins were contacted with specific molecular
tain hydroxyls, ketones, aldehydes, sacchorides, etc. weight fractions of the extracted organics for one week
plus a minority of amides, amines and traces of mineral and analyzed. The solutions were tested before and after
elements. The molecular shape is determined by the de- for TOC and chloride. Both styrenic and acrylic polymers
gree of branching in the polymer structure. Calculations in their gel and also their macroporous versions of each
based on Symons11 data show that there are, on aver- were used as well as non ionic macroporous adsorbents.
age, between 1 and 2 carboxylic groups per benzene ring Crosslinking levels in the gel phase was varied in the gel
(Figure 14). Several other studies of these materials resin, and also in the macroporous resins.
have shown that the carboxylic content of humic sub-
stances varies usually between 3 to 12 milliequivalents The results showed that less than 4% of the TOC values
per dry gram, which is similar. These substances behave were removed by the adsorbents. Surprisingly those 4%
like weak acid ion exchangers. Humic acid and lignin are had molecular weights of less than 1,000. This discovery
used commercially as dispersants and chelants. is significant, because it caused us to re-examine our
ideas and beliefs about the nature of "organics".
Humic substances form complexes with metals. Figure 2
shows the structure of the humic acid complex with iron In the past, organics were invariably thought of as being
as proposed by Gjessing4. When iron is present in this of such high molecular weights that they could not fit into
form it is commonly referred to as "heme" iron. The diffi- the gel phase of the resin. It was, until now, believed that
culty in removing this form of iron is due to its non reactiv- naturally occurring organics were removed by a combi-
ity since the complex is only slightly ionized. Flocculation nation of adsorption and ion exchange (6,7,10,12).
and filtration, or total destruction of the organic matter by
Significantly the Symons11 paper proved that the macro- 95% of the overall TOC influent with no indication of
porous resins did not necessarily do better then gel res- breakthrough at over 4,000 BV's of GAC. Of the 5% that
ins. This had also been shown by others [1, 14, 15]. For was unremoved, 70% of it was less than 500 molecular
the fist time we began to appreciate the importance of the weight and thought to be alcohols and similar materials.
gel phase. The lower crosslinked macroporous resins did The resin was regenerated at the sulfate breakthrough
better than higher crosslinked ones and lower crosslinked using a salt/caustic combination.
gel resins did better than higher crosslinked gel resins
and higher crosslinked macroporous resins. Even for the Resin alone removed over 80% of the TOC. Significant
very high molecular weights (above 10,000) the relative organic fouling protection could be accomplished by the
gel phase water retention was the only reliable predictor salt/caustic regenerated anion resin alone without GAC,
of removal performance. The gel phase water retention is since the lower molecular weight organic matter is more
controlled by the crosslinker level. rapidly eluted from and therefore less likely to foul work-
ing resins.
The second stage of Symons' studies were column stud-
ies. These were designed to measure the ability of an For ultra high purity water systems requiring maximum
anion resin regenerated with 2N NaCl and .5 N NaOH to reduction of TOC, a two stage organic removal process
reversibly remove TOC contained in Lake Houston water. with the resin in front of the GAC would be required.
It was found that the sulfate ions offered significant com-
petition for the ion exchange sites. Some of the lower There are four major reasons for the reduction of TOC:
molecular weight organic matter (less than 5,000) was
less preferred than sulfates and were "dumped" off when 1. Bulk color or TOC reduction from water used for
runs extended beyond the sulfate breakthrough and municipal or domestic use.
reached concentrations higher than their inlet concen- 2. Maximum reduction of TOC for ultrapure water ap-
trations. plications for semiconductor manufacturing
and/or for municipal installations to remove THM
Since the organic matter is multivalent, electro selectivity precursors.
will effect their apparent selectivity for the resin in the 3. Prevention of organic fouling of demineralizer res-
same manner as shown in Figure 3 for the sulfate ions. ins.
4. Reduction of organic substances in steam boilers.
10000
Equilibrium relationships are very important to the regen-
eration process and to combat build up which leads to
1000
fouling. The efficiency of regeneration can be affected
APPARENT SELECTIVITY

greatly by changing the salt concentration. Let us briefly

COEFFICIENT

100
review some of these relationships.
10
Table 1 shows the selectivity coefficients of the various
Selectivity reversal ions for gelular Type 1 and Type 2 anion resins compared
1
to the hydroxide ion. Those substances with the higher
selectivities are more difficult to regenerate off of the resin
0.1
1 10 100 1000 10000 with hydroxide. However, these resins are more easily
TDS, PPM as CaCO3 regenerated with sodium chloride because the relative
affinities of all the substances are lower against the chlo-
Figure 3 APPARENT SELECTIVITY- SULFATE / CHLORIDE ride ion. For the Type 1's, the affinity of the chloride is
from 11 (ResinTech SBG1P) to 25 (ResinTech SBG1)
times higher than the affinities for the hydroxide ion.
Symons work was focused on ways of reducing THM Therefore the selectivity coefficients for the listed sub-
precursors. His study included activated carbon. When stances against the chloride ion is 11 to 25 times lower.
used alone granular activated carbon (GAC) could only This makes the chloride ion significantly more effective as
treat about 1,000 bed volumes removing about 80% of a regenerant than the hydroxide ion. Note the high val-
the TOC before the onset of increased leakage. This is ues for lignosulfonate and benzene sulfonate. These are
believed to be due to its limited capacity and because of similar to the basic structure of the naturally occurring
surface blockage by large sized organics. The use of ion organics. It is believed that they have similar affinities for
exchange ahead of the GAC let the GAC exclusively deal the resin as the humic substances and are therefore use-
with the lower molecular weight unionized material pass- ful as models.
ing through the resin. This combination removed 93% to
 mons11 showed that the same brine could be reused up
Table 1 to 9 times without loss of capacity or quality in TOC re-
ductions. So it is obvious that much longer relative re-
Relative Affinities generation times are needed for organic ions.
(OH=1)
Before discussing the importance of contact time to the
regeneration of the organic acid form anion resin, let us
first review the effect of contact time on the regeneration
Type 1 Type 1 Type 2 of softeners. The common cations come to 80 to 90% of
Standard Porous Standard equilibrium in about 5 minutes(10). It is standard practice
Gel Gel Gel to design softeners with regenerant times of about 30
minutes.
Lignosulfonate 800 400 120
Benzenesulfonate 500 253 75 Figure 5 shows how the change in regenerant contact
Salicylate 450 358 65 time affects operating capacities of softeners. Signifi-
Citrate 220 110 23
cantly the effect on operating capacity is about the same
Iodine 175 97 17
degree as the relative attainment of equilibrium versus
Nitrate 65 42 8
Bromide 50 31.1 6 time.
Chloride 22 11.1 3
Bicarbonate 6 3.6 2.3 Due to their slower diffusion rates, the regeneration of
Hydroxide 1 1 1 organics is considerably more rate sensitive than for com-
mon inorganic ions. It has been shown that the
equilibration times for aquatic organic matter is measured
in hours (12), not minutes. For example, aromatic amines
The separation factor, which is the same as the selectivity (10), like diphenylamine, reached less than 10% of final
coefficient for monovalent exchange, is what determines equilibrium after 4 hours. The same phenomenon has
the efficiency of a regenerant. It can be used to calculate been shown to apply to anion resins (5, 6).
the theoretical regeneration level per unit of regenerant.
In the case of multivalent ions the separation factor is
more closely related to the apparent selectivity coefficient,
•Chloride versus Hydroxide Regenerants
which varies with the total ionic concentration as shown
for the sulfate/hydroxide ion pair in Figure 3. When the •1
•TYPE 1 CHLORIDE K=30 TO 36
regenerant concentration is about 10% the apparent se- •0.9 •Salt regeneration
lectivity and the selectivity coefficient are about the same. •0.8 •TYPE 2 CHLORIDE K=50
So we can use the selectivity coefficient as parameters.
•0.7
•0.6 •TYPE2 OH K=120
The relative separation factor of chloride for the resin,
•0.5 •Hydroxide regeneration
during regeneration of a Type 1 resin with 10% brine ver-
sus 4% NaOH, is about 11 to 25 times higher than hy- •0.4
•TYPE1 Porous Gel OH K=400
droxide. Therefore it is much more effective in removing •0.3
•TYPE 1 Std Gel OH K=800
organics. •0.2
•0.1
•K= Separation factor, organic / regenerant
The chloride cycle regeneration of organics can be seen •0
to be as effective for Type 1 resins as for the Type 2's.
The Type 1's are more chemically stable and can be
made with higher water retentions (low crosslinker levels)
•Figure 4
therefore they can remove higher molecular weight or-
ganics than the Type 2's. This can best be seen in Fig-
ure 4, which shows the theoretical regeneration versus Large molecular sizes make for slower gel phase diffu-
dose levels based on selectivity coefficients for chloride sion rates, so does higher selectivity and valence. Wil-
and hydroxide regenerations of lignosulfonate. son12 studied Type 2 resins that had been fouled, by
equilibrating them with brine/caustic solutions in batch
This gives a good picture of the effect on regeneration equilibrium experiments. He studied the amount of
efficiency of the relative selectivity coefficients. Keep in organics removed from the resin as a function of time. It
mind these curves are based on the assumption that was found that over 16 hours was required for what only
enough contact time has been given for the regenerant appeared to be equilibration. However, it took only 4 to 6
and resin to reach equilibrium. So we see that the porous hours to reach 70% to 80% of the 16 hour amount. Wil-
Type 1 gel resins can remove TOC substances and be son's data is shown in Figure 6. Data such as this could
well regenerated with chloride regenerants. In reality this be used to construct an operating capacity factor curve
is not the case. For example, when brine regeneration similar to that of Figure 5.
was carried out in the same time as for inorganic ions
(using short contact times and no prolonged soak) Sym- A comparison of Figures 5 and 6 shows remarkable simi-
 was very beneficial. The more NaOH that was added the
higher the amount of organic matter regenerated off of
Parameter 100% NaCl Levels (applied @ 10%) the resin.
1.1
FRACTION OF MAXIMUM CAPACITY

1
The pK of the carboxylic groups in humic acids has been
0.9
determined to be between 1 and 6 (Gjessing4). When the
0.8
0.7
0.6
0.5
50 ML SOLN. 1 GR. 18 hr. FOULED TYPE 2
0.4 150
0.3 140
130
0.2 120

PPM HUMIC ACID

0.1 110
100
0 90
0 5 10 15 20 25 30
80
BRINE INJECTION TIME, Minutes 70
Dosage 5 10 15 20 60
50
Figure 5 CAPACITY FACTORS vs BRINE INJECTION TIMES 40
30
20
larity and might be identical except that the X axis for the 10
0
inorganic exchange is measured in minutes compared 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
with hours for the organic exchange. Using time as a pa- NORMALITY OF Cl
rameter, we should be able to project operating capacities Figure 7 NaCl CONC VS HUMIC ACID REMOVED
in a similar manner to the capacity versus the regenerant
level curves used in inorganic ion exchange. The impact
of resin characteristics on diffusion rates is quite signifi- brine is alkaline these groups become fully ionized. In
cant, it will therefore be necessary to establish separate order to keep them ionized and soluble at the high con-
time factor curves for each type of resin. This will be ad- centrations of regeneration, excess hydroxyl groups are
dressed in more detail in later chapters of this work. required. Adding caustic to the brine does both.

Wilson varied the concentration of the brine and found At ResinTech Inc. we have developed a test protocol for
that the amount of organics pushed off of the resin varied fouled anion resins. First we determine total and salt
directly with the concentration of the regenerant his re- splitting capacity in the normal manner. Then the sample
sults are presented in Figure 7. This shows the impor- undergoes a comprehensive organic fouling treatment
tance of equilibrium relationships. process, which includes hot brine and caustic. Then we
retest the resin and make a before and after comparison.
pH not only affects ionization levels in weakly acid mate- Typically the only difference we see is in salt splitting ca-
rial, it can also affect solubility. For example phenol, an pacity. The total capacity usually stays the same while
aromatic carboxylic acid, is soluble to about able 7.5% in the salt splitting capacity increases after treatment.
cold water as the acid, but as the sodium salt its solubility
reaches almost 20%. Wilson also changed the regener- The total capacity test requires the resin to be treated
ant formulas by adding various levels of either NaOH or with hydrochloric acid followed by elution with NaNO3.
hydrochloric acid. The effects of these factors on organic The low pH created by the acid forces the organic acids
elution are presented in graphical form in Figure 8. The to recombine with hydrogen and become unionized but
acidification of the brine regenerant significantly reduced they still remain inside the resin due to either lack of sol-
the regeneration. Adding sodium hydroxide to the brine ubility or insufficient contact time, or both. This allows the
fouled salt splitting site to be converted to its chloride
form, which is then measured by the nitrate elution step.
50 M L 2 N N aC l/1N N a O H 2 G R fld T Y P E
11 0% 2
The organics are not very soluble and diffuse slowly (4, 8,
100 % 9, 10),otherwise the acid treatment would be an easy way
P erce nt H U M IC acid

90%
80% to remove fouling. In the salt splitting capacity test the
resin is converted to the chloride form under alkaline con-
re m o ved

70%
60% ditions before the nitrate elution step so the organics re-
50%
40%
main ionically active and fixed to the salt splitting sites
30%
they are attached too. Therefore these sites cannot be
20% converted to the chloride form so they are not measured
10% by the nitrate elution step.
0%
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

T IM E , H O U R S At ResinTech Inc. we have also developed simulation

F ig u re 6 TIM E V S H U M IC A C ID R E M O V E D
technology which is used with the "before" and "after"
treatment test results to project comparative resin per-
formance. The results are useful in that they allow us to factors for each of them. Figure 10 shows the relative
make an economic evaluation of whether or not to carry order of TOC reduction abilities of those resins compared
out a brine rejuvenation, rebed, or do nothing. with their calculated Preuss factors for the 4 molecular

Figure 9 compares the known relative selectivity coeffi-

cients of various anions with the relative organic elution
obtained by Wilson12. He used their sodium and potas-
50 mL 1N Salt, 1 Gram of resin, 18 hr. TYPE 2
sium salts in otherwise identical batch equilibrium ex-
periments. Figure 9 shows that the correlation between

HUMIC ACID ELUTED

Relative Quantities
the relative amount of organics eluted and the selectivity

COEFFICIENT, K
coefficient of the anion portion of the regenerants is quite

SELECTIVITY
good. Figure 9 also confirms the importance of selectiv-
ity in regeneration effectiveness. Indirectly this shows the
importance of proper concentration control in order to en-
hance the regeneration of the multivalent TOC sub-
stances (see Figure 3 also).
Chloride Bromide Iodine Nitrate
Affinities (Type 2) Humic acid elution (NaCl)
TYPE OF SALT
50 ML 1N NaCl 1 GR. 18h. fouled TYPE 2
170 Figure 9 HUMIC ACID REMOVAL & SELECTIVITY COEFFICIENTS
160
150
140
130
120 weight ranges that were studied. Figure 11 contains the
PPM HUMIC ACID

110
100
90
data for all 4 ranges from Figure 10 in a single graph.
80
70
60 The resins Symons11 used included gel and macroporous
50
40 types, styrenic and acrylics. Figures 10 and 11 show
30
20 that from below 1,000 to over 10,000 molecular weight,
10
0 the Preuss factor provides an excellent correlation of in-
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
creasing absorption capabilities (less leakage).The idea
NORMALITY OF ALKALINITY OR ACID
NaOH acid
that the bigger the lake each ion exchange group has
surrounding it, the more effective it was in removing the
Figure 8 ALKALINITY/ ACID VS HUMIC ACID REMOVED various ions holds up well.

It is commonly accepted that organic scavengers become

fouled and performance suffers if they are overrun. There
The Preuss factor, named after Albert F. Preuss who de- can be no doubt that organic substances are considerably
veloped it, by co-worker, Sallie Fisher3, was used to rate larger than the common inorganic ions and can physically
the relative fouling resistance of anion resins in the early fill a resin, even before the ion exchange capacity is con-
50's when researchers were looking for ways to define or- sumed.
ganic fouling and to rate fouling resistance in strong base
anion resins. It was also applied to uranium recovery. If we consider the ratio of gel phase water per site the ion
The Preuss factor is the molar ratio of gel phase water to exchange lake, then when these substances fill that lake,
ion exchange capacity. It was found to be an excellent diffusion becomes hindered and slowed. The bigger this
predictor of performance within resin types; i.e. Type 1 or lake, the more of the physical load it is able to handle.
Type 2.
Figure 12 shows water retention values of products
At first, in attempts to quantify fouling and organic/resin meeting typical supplier specifications versus the ratio of
interactions some researchers tried to equate ion ex- moles of gel phase water to capacity; i.e. Preuss factor.
change intermolecular distances to equivalent gel phase Standard grade gel resins offered today, regardless of the
porosities(10). It was frustrating to do that because of the supplier, have similar capacity and water retention values.
wide variation in the shapes of the organic molecules, The values used in Figure 12 are calculated from mid-
which can vary from the highly branched aromatics to point values for typical gel resins plus ResinTech SIR-
long chain polymers or even simple alkanes. Therefore 22P, which is a special purpose resin intended for organic
the molecular weight does not describe the physical removal. Interestingly, the Type 2 gel resin has the low-
shape. However, the Preuss factor described the relative est and the acrylic resin has the highest Preuss factors of
size of the lake that the organics have to swim in as they the standard gel resins.
move to and from the ion exchange sites, inside the gel
phase of the resin. Therefore you would expect the acrylic resins to be able
to remove the most organics and the Type 2's to remove
Symons11 listed the properties of each of the resins used the least amount of organics of the common gel resins.
in his study. I have used this data to calculate the Preuss
 GEL PHASE WATER RATIO vs RELATIVE
ORGANIC REMOVAL EFFICIENCY GEL Phase Water Ratio Vs
BELOW 1K MW 5-10K MW
Relative Organic removal Efficiency
1 - 10K PLUS MW
ORDER OF INCREASING PERFORMANCE

ORDER OF INCREASING PERFORMANCE

15 16 17 18 19 20 21 22 15 16 17 18 19 20 21 22
1-5K MW GREATER THAN 10K MW

15 16 17 18 19 20 21 22 15 16 17 18 19 20 21 22 15 16 17 18 19 20 21 22
MOLES GEL PHASE WATER PER EXCHANGE SITE MOLES GEL PHASE WATER PER EXCHANGE SITE
Figure 10
Figure 11

This is quite interesting since it is well known that both of

these products are widely acclaimed for their excellent It is also well known that the Type 2 resins are the most
organic fouling resistance. However, since we are talking efficiently regenerated resins. The acrylic resins, though
about organic fouling resistance rather than organic re- more efficiently regenerated than styrene based Type 1's,
moval, regeneration efficiency is the key factor, not re- are not as efficient as the Type 2's. Since the ability to
moval ability. elute the organics is a key factor in fouling resistance, so
is regeneration efficiency.
The strongly basic acrylic gel resins are indeed known to
remove the most organics of all the common strong base As Type 2 resins age they become partially weakly basic.
gel anion resins. Likewise, it has long been recognized The weak base sites have a lower capacity for the weak
that one reason the Type 2's do not foul is because they organic acids. As I mentioned previously, we have found
let a significant portion of the organics slip by. that when fouling occurs it is usually only the salt splitting
groups that are affected. This explains
why Type 2's tend to foul less as the age than when they lectivities, and higher Preuss factors, so they provide bet-
are brand new. ter overall performance than their lower water retention
level Type 1 brothers. This is a major reason why the
Let us look at inorganic ion regeneration to see the role Type 1 porous gel resins are widely preferred over the
that selectivity can play during regeneration. One exam- standard porosity Type 1 gel resins (ResinTech SBG1P
ple of the importance of the regeneration step is the re- over ResinTech SBG1).
generation of a calcium laden strong acid resin with sulfu-
ric acid. Calcium is the most strongly absorbed ion of the
common cations. Yet the higher the relative calcium con- For Strongly basic Gel anion resins
tent of the water, the lower the resin's operating capacity. 50
This is because calcium is also the hardest ion to elute SIR22P type 1
45 (Styrenic)
Moles H2O/Mole of Capacity

and in this case the operating capacity is determined by

40
the regeneration cycle not the exhaustion cycle.
35

Figure 4, was calculated for the regeneration of lignosul- 30

fonate. It shows how the relative affinity between the re- 25

Acrylic
generant compound and the substance being removed 20 gel
affects regeneration efficiency. Lignosulfonate is believed Type 1 Porous Gel
15 (Styrenic)
Type 1 Gel
to behave in a similar manner to the humic substances. 10 Type 2 Gel (Styrenic)
(Styrenic)
The Type 1 resin has the highest affinity for lignosulfate 5
and, of course, it has the lowest regeneration efficiency. 40 45 50 55 60 65 70 75

The Type 2 resin has the lowest selectivity for lignosulfi- Water retention, %
nate and gives 3 times the regenerable removal rate at
10 pounds of NaOH regeneration than the standard Type Figure 12 Water Retention vs Relative Water
1 resin. This shows why Type 2 resins resist fouling better
than the Type 1's.
When resins become fouled they are usually rejuvenated
The higher water retention Type 1 resins have lower se-
using salt. As you can see in Figure 4 the use of salt,
 Ta b le 2
and enhanced biological activity has not been determined.
M o le c u la r w e ig h t d is tr ib u t io n o f T .O . C There is not enough information to say any more.

L a k e H o u s to n Lake A nna Symons'11 determined the carboxylic acid content of the

Te x a s V ir g i n ia TOC Table 3 and Figure 14 shows the calculated calcium
( S y m o n s e t . a l. ) ( W is e r e t.
a l. ) carbonate to TOC ratio from the Symons data together
+ 11 .5 with the theoretical values for various aromatic carboxylic
10 - 100K 1 0 .1
10 +K 1 0 .9 acids. The higher molecular weight organics have a lower
5 - 10K 2 8 .0 ratio of acidity. This is consistent with the theory of adding
1 - 10K 3 7 .6
1 - 5K 9 .5 functional groups by oxidation and molecular weight re-
1K 5 1 .6 duction that was discussed earlier. The calcium carbon-
500 - < 1K 1 7 .0
150 - 500 2 3 .8
ate equivalency ranged from .55 to 0.75 ppm of CaCO3
per ppm of TOC as the molecular weight lowered from
above 10,000 to 1,000. Values from molecular weights
instead of sodium hydroxide, quadruples the chemical below 1,000 were not determined. However, the trend ap-
efficiency of the regenerant. It is standard practice to use pears to not exceed the limiting value of a dicarboxylic
warm salt and extended contact time during injection, acid of approximately 1.05. Therefore we can use these
values to estimate the equivalent capacity load of TOC
values in the same manner as for inorganic ions.
CLIMATE EFFECTS ?
110 Calculated from the carboxylic content
100 1.1

ppm TOC as C / ppm as CaCO3

90 1
PERCENT OF TOC

80 0.9

70 0.8

60 0.7

50 0.6

40 0.5

30 0.4
over 10 K 1 - 10 K less than 1 K
0.3
MOLECULAR WEIGHT RANGES
SOUTHWEST MID ATLANTIC 0.2
1/2 PHENOL Phenol 1K - 5K
>10K 5K - 10K Dicarboxylic
Figure 13 TOC MOLECULAR WEIGHT DISTRIBUTION
Molecular equivalent sizes
Figure 14 CaCO 3Equivalency to TOC

usually followed by a post soak before rinsing the salt out

of the resin. A common rejuvenation procedure is to use
15 pounds of warm 10% NaCl with 1% NaOH injected Of course, more work is needed to validate these ratios.
over a one to three hour injection time followed up by an However, we can use them now for generalized com-
overnight soak before rinsing. Based on the information putational purposes. While, for more precise estimates,
presented in Figure 6, we might expect that the injection we can determine the TOC values and the organic acids
step achieves 40% to 70% of the organic removal (regen- ratio, or use direct measurements of the organic acids and
eration) and the overnight post soak is responsible for the convert them to calcium carbonate equivalents using 50
other 30% to 60%. mg as CaCO3 per meq of organic acids.

Compared to the normal sodium hydroxide regeneration It has been shown that the lower molecular weight portion
this salt treatment can elute an order of magnitude more of the organic matter breaks through at the same time as
organics from a Type 1 resin. The importance of the post the sulfates (7, 11, 14, 15). After the sulfate breakthrough,
injection contact time has also been noted by others (7, 14, the net organic removal rate dropped by 70% during Sy-
15). This explains why "brine squeezes" for organic fouling
mons' study. The sulfates at the bottom of the column
treatments, when done properly, can be so effective. were displacing some of the previously loaded TOC.

Symons11, and before him Wiser13, determined the mo- If the goal is reduction of the higher molecular weight
lecular weight distribution of the natural organics their re- TOC's, then the service cycle can extend past the sulfate
sults are shown in Table 2 and Figure 13. Figure 13 may break. For example, if you are trying to protect the work-
also show something about the geographical or climatic ing D.I. resin from fouling, you should be able to go well
influences on the molecular weight distribution. It is ap- past the sulfate break. However, this should not be done
parent that the more northern water (the Virginia water) if the resin is protecting a high purity water system and
has a higher molecular weight distribution than the water you want the maximum amount of TOC reduction. In this
in Houston, Texas. Whether this is because of a higher case, the system should have an activated carbon bed
residence time of the organics in the water, or longer after the salt cycle strong base resin bed.
sunlight hours to supply more UV or warmer temperature
 Therefore it is recommended that a minimum temperature
Table 3 of 120o F (not to exceed the design temperature of the
vessel, internals or salt cycle temperature limit of the resin
Ionic Strengths of T.O.C. *
being treated) and a minimum dose level of 15 pounds of
salt per cubic foot be used. The minimum injection time
Equivalency
should be 2 hours followed by a minimum 10-hour soak
meq of COOH per before beginning the displacement rinse.
ppm as CaCO 3
gram of T.O.C.
M.W. per ppm T.O.C.

10 K 11.03 .55
Multiple treatments can be used as a preventative meas-
1-10K 13.13 .66 ure when performed on a regular basis. A 5 pound mini-
1- 5K 15.05 .75 mum dose level with 1 hour injection time and a combined
< 1K ND displacement rinse and soak time of 1 hour is recom-
*Calculated from Symons' data of carboxylic content of mended. The frequency of this treatment depends on the
T.O.C. values in Lake Houston Water percent of TOC of the anion load. It is suggested that for
standard resins the TOC loading factor be kept to a small
We have found that for maximum reduction of organics it fraction, probably less than 25% of the total resin capacity.
is best to limit the service cycle to the sulfate breaks, and We intend to provide more specific information in the fu-
to polish with activated carbon to remove non-ionics. We ture.
compute the resins' throughput capacity based on the sul-
fate plus the TOC loading. We recommend a minimum
one hour brine injection time and use a rating factor of SUMMARY and FUTURE WORK
45% or less of the total capacity of the resin depending on
the resin. The regenerant solution should be 10 pounds of I have shown, through historical developments, how we
salt in a 10% solution with an additional 2% of sodium hy- come to know that naturally occurring organic substances
droxide applied at a temperature of 120o F to a pre- are best dealt with by ion exchange resins. In this paper
warmed bed, not less than 70o F. these substances were examined only as ions. We have
reviewed recent and historical work to demonstrate that
If only partial reduction or maximum throughput is desired, equilibrium and kinetic factors, especially time, are as im-
the service cycle can be run past the sulfate break. Once portant in defining regeneration (desorption) effectiveness
sulfate breakthrough occurs, the resin will continue to re- as the dose level of the regenerant. It has been shown
move the higher molecular weight organics. However, all that the naturally occurring organic substances have affini-
or most of the lower molecular weight organics, those with ties for the resin similar in degree to sulfate.
lower selectivities than sulfate, will be discharged as they
are displaced by sulfates and the high molecular weight From the information presented in this paper, we can de-
TOC's. During the later stage of the exhaustion cycle, the scribe the properties of what would make an efficient or-
overall removal of organics may drop to as low as 30%, ganic scavenger:
depending on the molecular weight distribution. After the A. It must have sufficient gel phase water to physically
sulfate breakthrough, TOC levels will rise and the concen- accommodate the organic matter.
tration of the lower molecular weight TOC's may rise B. It must have sufficient ion exchange functionality to
above their influent levels. This style of operation could provide acceptable capacity.
be useful for using a scavenger resin ahead of or as a
separate layer in the same vessel on top of a cation resin C. It must have excellent osmotic shock characteris-
in a softener to remove or reduce color. However, unless tics to withstand the internal stresses during load-
the water is presoftened, the regenerant should not in- ing and regeneration.
clude sodium hydroxide, in order to avoid hardness pre-
cipitation. The operating capacity will be less than 20% of ResinTech SIR-22P is an example of such a product. It
the total resin capacity. has a gel phase water retention of approximately 75% and
a total capacity over 10 Kgrs. per cubic foot. ResinTech
Let us look at how the information presented here can be SIR-22P has a Preuss factor of over two and one half
used to reclaim fouled resins. An organically fouled resin times the standard resins, including the acrylic types. In
by definition is an anion resin that is no longer working an ongoing trial it has already demonstrated its ability to
well because it is loaded with too many organics for it to run to complete organic breakthrough without fouling and
continue to be operated as an exchanger of inorganic ions provide consistently high TOC reduction while operated in
in an efficient manner. The goal of the treatment is to re- the chloride cycle. Its detailed performance characteris-
move enough organics to return the resin to successful tics will be the topic in a future chapter of this work.
operation. This requires sufficient dose, temperature,
concentration, and contact time. This then concludes the first portion of this endeavor.
ResinTech Inc. is continuing the study of ways to further
Time is the most important of these parameters, but all quantify the kinetic and equilibrium factors discussed here.
must be properly controlled. If the regenerant and resin Our studies include an ongoing full-scale trial of SIR-22P.
are too cold it will decrease the kinetics to the point where
the necessary contact time will increase significantly.
Bibliography 9. Kunin, Robert, The Role of Organic Matter in Water
Treatment - A Universal Theory, Amber-hi-Lites,
1. Anderson, Robert E., Estimation of Ion Exchange 179 (Spring, 1986)
Process Limits by Selectivity Calculations, Jour-
nal of Chromatography, 201 (1980) 10. Kunin, Robert, Ion Exchange Resin, Robert Krieger
Publishing Co. (1972)
2. Abrams, I.M., Removal of Organics from Water by
Synthetic Resinous Adsobents, Chemical Engi- 11. Symons, J.M., Fu, P.L-K, Kim, P. H-S, The Use of
neering Progress Symposium Series, 106-112 (1969) Anion Exchange Resins for the Removal of Natu-
ral Organic Mater from Municipal Water, Interana-
3. Fisher, S., personal communication, August, 1993 tion Water Conference Proceedings Book, 92-120
(Pittsburgh, PA, October, 1992)
4. Gjessing, E.T., Gel- and Ultramembrane Filtration of
Aquatic Humus; A Comparison of the Two Meth- 12. Wilson, A.L., Organic Fouling of Strongly Basic
ods, Schw. Z. Hydrol, 286-293, (1973) Anion Exchange Resin, J. Applied Chemicstry, Vol.
9, 352-359 (1959)
5. Kressman, T.R., and J.A. Kitchener, J. Chem. Soc.,
1201 (1949) 13. Wiser, S.L., Lee, Y.H., Stroh, C.R., O'Brien, Mr. Re-
moval of Organics from Secondary Makeup Water
6. Kunin, R., and R.J. Myers, Discussions Faradya -- Case Study, International Water Conference Pro-
Soc., 114 (1949) ceedings Book, 1-17 (Pittsburgh, PA, November,
1985)
7. Kunin, R.F., Suffet, I., Activated Carbon Adsorption
of Organics from the Aqueous Phase, Ann Arbor 14. Wolff, J.J., Strong Base Anion Exchangers and
Science: Ann Arbor, Vol. 2, 425. (1980) Organic Matters, Dia-prosium Bulletin, 68.C.2
(1968)
8. Kunin, Robert, The Role of Organics in Water
Treatment, Amber-hi-lites, 168 (Autum, 1968) 15. Wolff, J.J., Organic Matters Scavengers, Dia-
prosium Bulletin, 68.C.3.A (1968)