Apr 25, 2003

Chlorine Residual Boosting in Distribution Water: Problems with Chlorine Application and Disinfection Byproducts - Part 2


Low Chloramine

The town of Driscoll (population: 690, 30 miles south of the
plant) receives water from the Corpus Christi (C.C.) plant with a total
chlorine of 0.3-0.5 mg/L and NH3-N: 0.6 mg/L. (The plant discharge is 3.5-4.0
mg/L total chlorine, total ammonia 1.2-1.5 NH3-N.) The water also has a slight
increase of nitrate (0.7 mg/L) (Figure 1). The BP chlorination with Driscoll
intake water (4/30/01) occurs with a typical BP curve. However, the curve has a
lower maximum level, and reaches the BP at a lower chlorine dose, as expected
from water with a low total ammonia to start.

Table 2 (Driscoll) shows that there is a significant
dissipation after BP. The boosted chlorine, 0.83 mg/L (checked 60 minutes after
BP chlorination) dissipates in 24 hours to an unacceptable 0.19 mg/L.

Table 2 (Driscoll) suggests that an application of a higher
concentration (6.0 mg/L) after BP may be appropriate and sufficient to yield a
2.4 mg/L free chlorine residual at 24 hours.

Driscoll is injecting chlorine gas with a target for free
chlorine of 5.0-6.0 mg/L in the storage tank (170,000 gallon, retention time of
24 hours). This level drops to about 2.5 mg/L free chlorine residual after 24
hours. The water then moves to the elevated tank (150,000 gallon, 8 hour
retention time) where it discharges the water (approximate 2.0 mg/L) into the
distribution system. The distribution water sampled was found to be stable for
several days in the laboratory at 25° C, and in the field, maintaining a
1.5-1.7 mg/L free chlorine residual at the terminal distribution sites.

Case of Depleted Chloramine

Kingsville, the largest user of C.C. water, receives the
intake water 40 miles south of the plant. The residual is practically zero
(0.05-0.10 mg/L) at the intake with the total ammonia not detected (<0.1
mg/L). Zero chlorine residual also was found at the southeast end of
distribution system (40 miles) at Padre Island.

It seems that the additional 10 miles makes a significant
reduction in the chlorine residual. This means that the retention time in the
last 10 miles plays a critical role in the chlorine dissipation in the pipe.
The enhanced level of the nitrate (1.0 mg/L) in both Kingsville and Padre
Island may suggest that nitrification is partially responsible for the zero
chlorine level in intake waters.

With ammonia depletion in the intake water, chlorination
with Kingsville intake water results into a straight line instead of a
breakpoint curve (Figure 1). However, supplement this water with ammonia
sulfate (1.0 mg/L NH3-N) and a typical BP curve is formed. This indicates that
there is no inhibitor formation to block the breakpoint chlorination in the
Kingsville's intake water, instead the depletion of ammonia (free or combined)
would be responsible for incomplete formation of the BP curve.

Significant dissipation also was found in the newly formed
chlorine after chlorination in Kingsville's intake water. It was not observed
with an added preformed monochloramine at the concentration (2.0-7.0 mg/L).
This difference in susceptibility to the free and combined chlorine suggests
that intake water a long distance from the C.C. plant water still carries the
chlorine demand, but is not reactive with the chloramine. Therefore, BP is not
a prerequisite for chlorine dissipation. Water in the chemical receiving basin
(prior to chlorine and ammonia sulfate injection) in the plant was found to
contain such demand in considerable amounts.

This intake water, with practically zero chlorine residual,
can be neutralized with the free chlorine and, as a result, the residual free
chlorine becomes stabilized for several days at room temperature. However, the
addition of straight intake water (50 percent) to the previously stabilized
water (50 percent) could result in destabilization, with the stabilized free
chlorine residual dissipating again.

The rapid decline of chlorinated water follows a hyperbolic
line with time, regardless of the sampling sites (plant discharge or
distributions). Data show that 24 hours of rapid decline is followed by a slow
dissipation (Table 2). Therefore, the chlorination dosing at 3.0-4.0 mg/L is
sufficient enough to maintain the free chlorine residual level, 1.5-2.0 mg/L
after the first 24 hours (50 percent reduction) judging from the laboratory
data at 25° C. However, the operator indicated that the free chlorine
residual (3.5-4.5 mg/L) dissipated as low as 0.5-1.0 mg/L after 24 hours inside
the storage tank. A second boost was necessary to increase the free chlorine to
a level of 2.0 mg/L. This double boosted water is stable without dissipation in
the free chlorine residual for several days, and it is blended with the
chlorinated groundwater (80 percent) for the distribution discharge.

An enhanced decline in the storage tank results not only
from the post breakpoint dissipation, but also from the bacterial activity in
the biofilmed layer on the wall of the storage tank. The biofilm problem does
not exist in laboratory tests using the very clean acid-washed glassware. In
addition, much higher temperatures in the storage tank during the summer months
in south Texas can lead to dissipation.

Normal Level of Chloramine

BP chlorination cannot be applied to the distribution water
of chloramine if it is in a normal range. In this article, the addition of free
chlorine to the C.C. water was conducted for the purpose of testing and for a
comparison with a low level of chloraminated water. BP chlorination (1/31/02)
with distribution water in the city (20 miles from plant) yields the same BP as
other samples (2/4/02), showing the breakpoint with increasing dose of chlorine
(Figure 2), and followed by instability of the formed free chlorine after BP
(Table 2 C.C.). The BP curve has a chlorine residual reaching the maximum point
and the breakpoint in lower chlorine dosing because of the preexisting normal level
of combined chlorine in the sampled water. Chlorine residual in the BP curve
does not start at the zero level, but at the residual level of the distribution
water (Figure 2). The plant discharge (P1 sample shown in Figure 2) does not
have the maximum peak of BP curve as seen in C.C. water. This suggests that
this plant water sample has been treated with a high ratio of chlorine to
ammonia (e.g., 5:1 or slightly higher).

Regardless of sampling sites, the C.C. water is potentially
reactive to free chlorine after BP chlorination. Its reactive strength may be
different depending on the season or weather.

When forming the monochloramine in the plant, the ratio
(Cl:NH3-N) is the most critical factor. A simple formula ranging from 3:1 to
5:1 encourages stability and the formation of combined chlorine. There is no
definite formula when boosting low levels of chloramines with free chlorine
because of the complexity involved in the boosting process as well as plant and
distributional influences. Besides BP dissipation, the monitoring breakpoint
chlorine concentration (dosing chlorine concentration to reach BP) would change
depending on the total ammonia concentration in the field. However, both
utilities have made an empirical application to acquire stable chlorinated
water for their distribution systems, without information as to BP and chlorine

Boosting Check

Before boosting the chloaminated water with BP chlorination,
routine water quality information (e.g., pH, alkalinity, free and total
chlorine, etc.,), the total ammonia and, if possible, the ammonia specification
(free and combined ammonia) is needed to determine the water's quality. A BP
curve at 60 minutes after dosing free chlorine at different concentrations as
well as a stability test for the newly formed free chlorine also is necessary.
While the results obtained in the laboratory may not represent the reaction in
the storage tank, these results still will provide a basic line (dosing and
residual chlorine at BP) in order to prepare for boosting in the field.

Small utilities with limited lab equipment can use a pocket
chlorine photometer for the breakpoint test. The free chlorine for dosing can
be prepared by diluting an approved commercial bleach (5.25 percent) for water

If the BP curve at 60 minutes after chlorination is
deviating from a typical BP figure (i.e., without a sharp drop between the
maximum and breakpoint), the water to be boosted must be contaminated with
intrusions of organic nitrogen or some other chlorine demand.

BP Boosting and THMs

The THM rule has made many utilities switch from
chlorination to chloramination as a disinfectant byproduct mitigation strategy.
Breakpoint chlorination changes the disinfection mode from monochloramine to
free chlorine. This process induces instability in the newly formed free
chlorine in C.C. water, and may have formed THMs as well. Breakpoint
chlorination can initiate THMs formation as chloramines decline and free
chlorine starts to increase.3

The Kingsville's intake water, though practically zero in
the chlorine residual, still contains 44 mg/L TTHM. This is almost the same
level discharged from the C.C. plant. This level increases to 220 mg/L with 2.2
mg/L total chlorine (August 2002) after the first boosting. While the data on
DBPs is too scarce to draw any conclusions, it does suggest that the THMs are
formed during retention in the storage tank. It may be that THMs are formed
more efficiently in the treated water after BP chlorination than in raw water
sources by chlorination.

Kingsville maintains a low level of THMs in the distribution
system (annual average less than 15 mg/L) through dilution. The distribution
water is blended with 20 percent boosted water and 80 percent chlorinated
groundwater. If utilities can obtain a high level of chloramine in the
distribution, problems such as dissipation and THMs can be reduced.


The city of Corpus Christi is in the coastal bend of Texas,
where hurricanes and tropical storms occur. Hurricane Bret2 which caused $20
million in damage to south Texas, touched down just 70 miles south of Corpus
Christi. The west side of Driscoll was flooded with three feet of water for a
few days. Therefore, all types of boosting methods must be prepared for such an
emergency in order to avoid intrusion by pathogens into the distribution

About the author

Shin-ichi Tokuno retired as a microbiologist in December 2002 from the Water Utilities Laboratory at the City of Corpus Christi, Texas.