-------
- 152- -
2O
30 .40 50 60
REACTION TIME (hrs)
70
8O
9O
Fig. 5 Effect of pH on chloroform production, settled
water 25°C, 10 mg/1 chlorine dose
The increase of trihalomethane formation rate with
pH was expected because the classical haloform reac-
tion is base catalyzed; however, this explanation is
likely to be an oversimplification where rather complex
humic acid structures are involved. Simple methyl ketones,
models for the haloform reaction, have been shown to
react too slowly to account for trihalomethane formation .
under most drinking water conditions, suggesting a
different reaction mechanism (7). Christman once suggested
a simple "opening up" of the humic acid molecule because
of mutual charge repulsion at high pH increasing the
availability of more reactive sites" on that molecule as
a possible cause of the influence of high pH on reaction
rate (Personal communication)." ;
-------
- 153 -
Characteristics and Concentration of Precursors
In artificial systems, increasing the concentration of
humic acid precursor in the presence of excess chlorine
with otherwise constant reaction conditions causes halo-
form yields to increase in direct proportion to the
humic acid dose {Pig. 6) (7). Prom supply to supply, how-
ever} only crude relationships have been found between
organic carbon concentrations, and trihalomethane yields
(3). Similar effects have been noted upon treatment,
Further, rate curves seem to take on distinctly different
shapes depending on the source of precursor substances.
The work of Rook (8) shows the reaction of fulvic acid
solutions to be characteristic of m-dihydroxyphenyl
moieties (e.g. resorcinol) in that the reaction is near-
ly complete at near neutral pH"in less than two hours
(Pig. 7).
0.3- •
20
30
4O 50 60
TIME (hrs)
7O
80
90
100
Fig. 6 Effect of humic acid concentration on trihalomethane
.o
production, pH 6.7; 25 C; 10 mg/1 chlorine dose
-------
- 154 -
o
| 200-
o
to
O
cc
| 150-
"wlOO-
0
£ (
"a
6 50-
E
^ RESOBCINOL __^< Q
—
0
FULVIC A^D__ ^
«•«• •**** ~~
!J> x*"
r
2
-4O «
'5
i-
-3O O
e
i *•
fc rv
-20 w®
o
X -g
-10 i i
El
10
20 30 4O SO T 11O 120 130
Fig. 7 Reaction of model precusors with aqueous chlorine
so ao
M tO SO 60
REACTION TIME (hrt)
Fig.
j8 Comparison of humic acid, raw water reaction rates
at similar NPTOC concentrations, 10 mg/1 Chlorine dose
-------
- 155 -
Quite a different characteristic curve is observed with
Ohio River water precursor and a different source of
humic acid under similar conditions where the reaction
takes place relatively slowly over a period of many
days (Fig. 8) (7). The probable differences in pre-
cursors at different locations has been further demon-
strated in work at the EPA Cincinnati laboratory where,
as expected, treatment with permanganate at low dosages
was nearly 1OO percent effective in preventing the for-
mation of trihalomethanes on chlorination of resorcinol
and m-dihydroxybenzoic acid solutions, yet permanganate
was only marginally (1O-2O%) effective in reducing the
ability of Ohio River precursors to form trihalome-
thanes upon subsequent chlorination.
Additionally, work at the Cincinnati laboratory has
shown there to be only a slight influence on trihalo-
methane formation rate (or yield) of increasing chlorine
dose (beyond demand) where "precursor" is kept con-
stant (Fig. 9). Both similar and contrary results have
been reported by others while working with different
sources of precursors (6, personal communication).
The above serves only to indicate that although pre-
cursor materials from various supplies may be of large-
ly natural origin, the composition of that material is
likely to be different, depending on the type of supply
involved and the source of precursors in the water shed.
Considerably more work is needed, therefore, to under-
stand the complex mechanisms of trihalomethane forma-
tion during water chlorination and to determine whether .
water treatment strategies for control of THM's could
vary significantly among these various supplies.
-------
- 156 -
0.6
0.5-
§0.4
t>
I 0.3
a
£ 0.2
Chlorine Dose;
6.0 mg/L *
4.0 mg/L .-
3.0 mg/L a
SJO
4.O2
3.O jj
O
•
2
2.O •
10
20
30
4O SO 6O
Tim* (Hrs)
7O
80
90
Fig. 9 Effect of chlorine dose on trihalomethane formation
III. Other identified apparent products of chlorina-
tion/treatment
As mentioned earlier, trihalomethanes represent the most
important group of individual identified halogenated
species from a concentration standpoint identified in
finished drinking water and resulting from chlorina-
tion practice. Although the mechanisms for trihalome-
thane formation are not well understood, many of the
conditions favoring or inhibiting trihalomethane for-
mation have been established. Other non-polar compounds
have been detected in finished water at the ng-ug-/l
level that were not detectable in the source water or
present in lower concentrations. Most of the sources of
these are even less well understood.
At least 19 non-trihalomethane halogenated volatile
compounds have been shown by Rook (8) (Rotterdam Sto-
rage Reservoir)"and more by Stieglitz, et al., (14)
-------
- 157 -
(Rhine River Bank Filtrate) to be formed at low con-
centrations upon chlorination. Rook speculates on a
possible pathway to explain the formation of some of
the observed byproducts in a way related to his pro-
posed mechanism for haloform formation from m-dihydroxy-
phenyl moieties. Stieglitz suggests no mechanism.
Coleman et al. (15) reported the co-presence of chloro-
picrin, chlorobenzene, a chlorotoluene isomer and a
chloroxylene isomer with their respective logical pre-
cursors, nitromethane, benzene, toluene, and m-xylene,
in finished chlorinated tap water. All of the above pre-
cursors but benzene were shown to be reactive with
aqueous chlorine to form the expected products. More
recently, chloroacetonitrile derivatives have been ob-
served in a finished tap water as a result of work at
the Cincinnati laboratory. Milligram per liter concen-
trations of acetonitrile could not be made to react
with chlorine under realistic reaction conditions to
form detectable chlorinated derivatives, however.
Further, even simple aromatic hydrocarbons have been
observed in some studies to be more prevalent or in
higher concentrations in finished tap water than in the
respective raw source water (16,17).
Considerable effort lies ahead to determine mechanisms
for formation of these apparent byproducts of chlorina-
tion (or possibly other treatment in the latter cases)
that in most cases seem to defy straightforward explana-
tions .
-------
- -158 -
IV. Summary
Modern analytical techniques have expanded our knowledge
of the formation,of unwanted by-products during chlori-
nation of drinking water. Of the non-polar fraction, tri-
halomethanes typically are formed in the highest concen-
trations, and much is now known about factors influencing
their formation. Other chlorinated and non-chlorinated
non-polar apparent by-products have been observed, but
little is known about their sources.
-------
-.159 -
(1) ROOK, J.J.
Formation of Haloforms During Chlorination of Natural
Waters
Water Treatm. and Exam. 23 (1974), 2, 234
(2) BELLAR, T.A., LICHTENBERG, J.J., KRONER, R.C.
The Occurrence of. Organohalides in Chlorinated
.Drinking Water
J. AWWA 66 (1974) , 11 , 7O3
(3) SYMONS, J.M., BELLAR, T.A., CARSWELL, J.K.,
DeMARCO, J., KROPP, K.L., ROBECK, G.G., SEEGER, D.R.,
SLOCUM, C.J., SMITH, B.L., STEVENS, A.A.
National Organics Reconnaissance Survey for Halogenated
Organics in Drinking Water
J. AWWA 67. (1975), 11, 634
(4) KUHN, W., SONTHEIMER, H., KURZ, R.
Use of Ozone and Chlorine in Waterworks in the Federal
Republic of Germany '
Ozone/Chlorine Dioxide Oxidation Products of Organic
Materials
Proceedings of Conference, International Ozone Institute,
Cleveland, Ohio (1978), 426-442
(5) -
Report on the Carcinogenesis Bioassay of Chloroform
National Cancer Institute, Div. of Cancer Cause and
Prevention, Washington,'D.C. (1976)
(6) ROOK, J.J.
Haloforms in Drinking Water
J. AWWA 6_8 (1976), 3, 168
(7) STEVENS, A.A., SLOCUM, C.J., SEEGER, D.R., ROBECK, G.G.
Chlorination of Organics in Drinking Water
J. AWWA 68 (1976) ,11,615
(8) ROOK, J.J.
Chlorination Reactions of Fulvic Acids in Natural Waters
Environmental Science and Technology 11 (1977), 5, 478
(9) LOVE, O.T.
Drinking Water Research Div., Municipal Environmental
Research Laboratory, US EPA, Cincinnati, Ohio
Unpublished data
(1O) BUNN, W.W., HAAS, B.B., DEANE, E.R., KLEOPFER, R.D.
Formation of Trihalomethanes by Chlorination of
Surface Water
Environmental Letters IP (1975), 3, 2O5
-------
- 160 -
(11) KLEOPFER, R.D.
Analysis of Drinking Water for Organic Compounds
Identification and Analysis of Organic Pollutants in
Water, Ann Arbor Science, Ann Arbor, Michigan (1976), 399
(12) LANGE, A., KAWCRYNSKI, -E.
Contra Costa County Water District Experience
Presented at the California - Nevada Section- AWWA
Seminar on Organics in Domestic Water Supplies,
Palo Alto, CA, USA, April 12 (1978)
(13) SYMONS, J.M., STEVENS, A.A.
Physical-Chemical Pretreatment for the Removal of
Precursors
Presented at Conference on Oxidation Techniques in
Drinking Water Treatment at Karlsruhe, Federal Republic
of Germany, Sept. 9-13 (1978)
(14) STIEGLITZ, L., ROTH, W. , KtlHN, W., LEGER, W,
The Behavior of Organohalides in the Treatment of
Drinking Water
Vom Wasser £7 (1976), 347
(15) COLEMAN, W.E., LINGG, R.D., MELTON, R.G., KOPFLER, F.C.
The Occurrence of Volatile Organics in Five Drinking
Water Supplies Using Gas Chromatography/Mass Spectro-
metry
Identification and Analysis of Organic Pollutants in
Water, Ann Arbor Science, Ann Arbor, Michigan (1976) ,
305
(16) SEEGER, D.R., SLOCUM, C.J., STEVENS, A.A.
GC/MS Analysis of Purgeable Contaminants in Source and
Finished Drinking Water
Proceedings of 26th Annual Conference on Mass Spetro-
metry and Allied Topics, held in St. ;Louis, Missouri,
May 28 - June 2 (1978)
(17) BRASS, H.J., FEIGE, M.A., HALLORAN, T., MELLO, J.W.,
MUNCH, D., THOMAS, R.F.
The National Organic Monitoring Survey: Samplings and
Analysis for Purgeable Organic Compounds
Drinking Water Quality Enhancement Through Source
Protection, Ann Arbor Science, Ann Arbor, Michigan
(1977), 393
-------
- 161 -
FORMATION AND BEHAVIOUR OF POLAR ORGANIC CHLORINE COMPOUNDS
W. Kiihn and R. Sander
In the previous report, Dr. Stevens reported on non-polar
organic chlorine compounds which can be formed in the chlori-
nation of waters. The resolution of this subject into two
components, i.e. into non-polar and polar chlorinated products,
is not founded on their possible hygienic or toxicological
differences but rather on the very different methods used
for their analytical detection (1,2).
This class of compounds has recently attracted considerable
interest not only because of its toxicity but because its
study was first made possible by modern analytical methods.
The phrase "a water is only as good as analysis permits" is
particularly appropriate to the analysis of organics in water.
For this reason I should like to start with a careful con-
sideration of the analytical treatment of this class of
compounds.
As shown in Table 1 , the treatment of organic chlorine com-
pounds, which as a rule are present in microgram amounts,
can be divided into an enrichment stage, separation into
individual substances, and the actual determination. In
practice, all polar organic chlorine compounds present analy-
tical difficulties. In contrast to the usually low-molecular,
non-polar, volatile compounds, on which the previous speaker
reported, polar compounds are usually high-molecular, less
volatile, and therefore more difficult to deal with. However,
if a group of substances resists individual substance analysis,
it is legitimate and sensible, and not only in water chemistry,
to treat these substances by a general or group method. The
-------
- 162 - •
method developed at the Engler-Bxmte-Institirte for the treat-
ment of all the organic chlorine compounds (TOCli. total
organic chlorine) begins with an adsorptive enrichment,
followed by mineralization in pyrohydrolysis and subse-
quent detection of the now easily analysed chlorine (3-5).
TABLE 1
Treatment of halogen compounds in water
Enrichment;
(preliminary
separation)
Separation:
Determination;
Liquid-liquid extraction
discontinuous
continuous (stages)
Blowing out
static (headspace)
dynamic (concentration)
Adsorption (elution)
batch test
column
Gas chromatography
packed columns
capillary columns .
Mass spectrometer
Conductivity detector
Microcoulometer.
Electron capture detector (BCD)
Plasma detector
Neutron activation
Pyrohydrolysis
(+ chloride determination)
As shown in the next table, the enrichment can be divided
into the following steps:
-------
- 163 -
TABLE 2 Experimental conditions "of the combination
of adsorption and flocculation for TOC1
treatment
ADSORPTION:
FLOCCULATION;
FILTRATION:
WASHING:
FILTRATION:
s-olution
pH
powdered carbon
NaNO3
time
pH
A13+
polymer
time
blue-band
NaNO3
time
: 1-20 litres
: <_ 3
: 100 mg/1 <_ 60 ym
: 0.01 N
: 30 min
: 6.5-7
: 5 mg/1
: 0.4 mg/1
: ca. 5 min
: (pressure filter)
: 0.05 N (200 ml)
: 30 min
blue-band (pressure filter)
With this enrichment process practically all organic consti-
tuents of the water that can be adsorbed and flocculated are
enriched. The water to be analysed is first of all treated
with solid sodium nitrate to obtain a O.01 N nitrate solution.
This reduces the undesirable adsorption of chloride. After
the addition of the active carbon in powder form, the pH is
adjusted to £ 3 with sulphuric acid. If the carbon suspen-
sion is thoroughly stirred,the adsorption process is as a
rule completed in half an hour. Most of the supernatant water
can be easily siphoned off from the powdered carbon if the
carbon is previously flocculated. The flocculating agent is
-------
- 164 -, -
5 mg Al per litre, added in the .form,of; al-uminitmi sulphate.,
For the flooculation the water must be adjusted to pH 6.5-7
with sodium carbonate. To assist the flocculation, 0.4 mg/1
of partly saponified polyacrylamide is added to the water and
is subsequently decanted off.' This process can be repeated
if a more complete treatment is desired. After filtration
of the combined sludges through filter paper, the filter
cake is suspended in 2OO ml of G.O5 N sodium nitrate and
stirred for half an hour. This removes nearly all inorganic
chloride. . .
The filter cake is then subjected to pyrohydrolysis, the
organically bound chlorine being thus converted into inor-
ganic chloride, as is shown in the following diagram.
PIRh-Pt Thermocouple ( PlRh-Pt Thermocouple
Quartz Wool Plug
Pyrohydrolysis Apparatus
Fig.r
This mineralization takes place1 in a current of steam and
oxygen at 90O°C in a tubular furnace. The chloride is then
determined titrimetrically or microcoulometrically.
By means of this pyrohydrolysis technique,the organic chlor-
ine materials can be treated as a group. As preliminary
-------
- 165 -
individual investigations- have shown, the majority of the
polar compounds are high-molecular, chlorinated lignic and
humic acids .
In order to investigate on a laboratory scale all the pro-
cesses taking place during, the ..chlorination of "water,
model waters containing humic acid were treated with
chlorine water, since humic acid can be used as a starting
substance for haloforms. An example of the results of this
work is given in the following figure.
:Gig/0
2500-
2000 J
1500-
1000
500-
TOC1
Chloroform
Free chlorine
-1
40 50 ClJjpg^O
Fig . 2
Chloroform formation, course of TOC1 and final
chlorine content as' a function of the initial
chlorine content at a constant reaction time
of 23 h
-------
- 166 -
Here the concentrations of chloroform and TOC1 are plotted
together with the chlorine content in dependence on the
dose of chlorine? the amount of humic acid weighed in was
2O mg/1 and the pH was 6.9. It is clear that the chlorine
content of the chloroform corresponds only to a small part
of the total organically bound chlorine. Therefore, the
effect of a chlorination cannot be evaluated solely on the
basis of the measured haloform concentrations? the TOC1
must likewise be considered (6).
2500-
2000
1500-
lOOO-i
500-
0 10 20 30 40 50 HS-NaQng/b
Fig. 3 Chloroform formation, course of TOC1, arid
final chlorine ccntent as a function of the
amount of humic acid weighed in, with a
constant reaction.time of 23 h
-------
- 167 -
• -_,:,-, ;..',:' '. ',« •••• • > •"- "• v , ' - " .' •?'••_••>?•• • .•,•.*.••<•;•..•:.• -\ .. :-*-<
As can be seen from Fig. 3, at a constant amount of chlor-
ine and increasing humic acid concentration, similar rela-
tionships are obtained for the formation of organic
chlorine compounds.
To obtain information on further parameters influencing
the formation of chlorine compounds, in addition to vary-
ing the reaction time, we varied the amount of chlorine,
the initial humic acid concentration, and the pH.' The
relationships represented in the following illustration
were obtained.
CHCl, Qjg/Q.
m
250-
200-
150-
100-
10
20 Qj/20 HS~Na
aa/2.o
HS-Nbi
o—o pH 9,2
PH6.3
HS-Ntt
•»•.
4 Chloroform concentrations as a function... of the
reaction time with various starting conditions
(numerical data in mg/1) ,
-------
- 168
In agreement with the relationships reported in the previous
lecture by Dr. Stevens, more chloroform is produced at
pH 9.2 than in the neutral range.'
In contrast to this, the formation of TOC1 is enhanced at
lower pH, as shown in the next figure.
TOO
pH 9. 2
pH 6.9
•2O C12/2O HS-Na
1O C12/2O HS-Na
2O C12/2O HS-Na
10 C12/20 'HS-Na
20 Cl«/1.5 HS- Na
20
30Time CH)
Fig. 5 TOC1 as a function of the reaction time with
various starting conditions (.numerical data in mg/1)
In the experiments performed here, the haloform reaction
ended after about 24 h.
At pH 6.9 there was a clear increase in the amount of TOC1
formed. The influence of the pH on the chlorine consumption
or the TOC1 formation can be explained by the increase in
the electrophilic character of hypochlorous acid in the acid
medium.
-------
- 169 - ,
in addition to these relationships, it was also of interest
how the formation of the organic, ghlorine substances is
influenced by suitable pre-treatment methods (7). Since
ozonization is regarded in waterworks operation as a prac-.
tical possible solution, further, experiments were performed
after a preliminary ozonization..The results on the form-
ation of polar and higher-molecular halogen compounds after
ozonization, measured as TOC1, are shown in the next illus-
tration.
TOCl Qjg/Q
2000-
1500
1000
500
20 Cl^/20 HS-Na
Q.6mgC\/rng C
:' 2.5 '
20 Time(H)
Fig. 6 Course of TOCl as a function of the reaction
time with various ozone consumptions (20 mg/1
chlorine + 2O mg/1 HS-Na)
The diagram demonstrates that even small additions of ozone
noticeably reduce the formation of organic chlorine compounds.
In the present case an ozone addition of 2.5 mg/1 water re-
duces the TOCl formation to one-quarter of that formed in the
absence of ozone.
-------
- 170 -
Similar results were obtained for chloroform formation,
as shown in Pig. 7.
O Q.6mg QeonfmgC
2,5mg
20 Time (h)
Fig. 7 Chloroform formation as a function of the reaction
time with various ozone consumptions (.20 mg/1 of
chlorine + 2O mg/1 HS-Na)
In this respect,however, different authors have obtained
different results.
In practical water-works operation other effects on the
formation and behaviour of the organochlorine compounds are
of interest. Chlorine is used in large quantities especially
for the removal of ammonia in so-called break-point chlor-
ination. For this reason the TOC1 formation in the presence
of ammonia was studied. The results are summarized in Fig. 8.
-------
_ 171 -
TOCl
500n
400-
300-
200-
100i
0
10
20
30 chlorine
addition
Fig. 8 Chloroform and TOCl concentrations as a function
of the initial chlorine concentration after a
reaction time of 2O h
It can be seen clearly that below the break-point, where for
a short time until the formation of chloramines still rela-
tively little free chlorine is present, the formation of
chloroform and of TOCl is very low.
-------
- 172 - "
After the break-point has been reached, however, a small '"-
addition of chlorine is sufficient to form large amounts
of these constituents.
These results are in accord with trials in an experimental ,
plant in Stuttgart municipal works, where good results
were obtained with stepwise chlorination just up to the
break-point. In comparison with the conventional practice,
this reduced the formation of chlorine compounds by a factor
of 10.
The data on the conventional procedure, obtained from the
same experimental plant, are compiled in Table 3.
TABLE 3 Effects of classical treatment for a river water
(Neckar, Stuttgart) with break-point chlorination
and activated carbon filters
Riv«f
water
Dissolved organic 5.0
carbon (DOC) mg/I
UV absofbance 1O.9
at254nm m1
Sum of haiofomis pg/1 0.4
Total organic 33
cHorimtTOCI) pg/l
After breakpoint
chkxinatkxvftoccu
lation.secHman
tatiort and filtration
4.1
8.6
50
524
Aft*r GAC
Carbon
LSS
3.1
5.0
16
364
AfMrGAC
Caftan
F300
1.6
3.6
25
296
-------
- 173--
After 'chlorination with 2O rag of chlorine per litre, both
the haloform concentration and the TOC1 concentration
rise , sharply. The quantity of trihalomethanes accounts
for only 10% of the total chlorine compounds formed. This
large proportion of polar chlorine compounds, which essen-
tially still resist individual analysis, is serious, espec-
ially in view of the fact that, as can be seen from the
last two columns of the table, these compounds are also
adsorbed only with difficulty in the subsequent active
carbon filter (8). A selectivity of the active carbon
filters is evident, since the carbon, which removes the
TOC1 less efficiently, adsorbs the trihalomethanes better
and vice-versa.
These statements are confirmed by measurements made in a
water treatment plant at the Rhine-. The results are collected
in the following table.
TABLE4 Course of the concentration of trihalomethanes
and of the total organically bound chlorines
during the treatment of drinking water in a
Rhine waterworks
River bank filtrate
Raw water after chlor-
ination (2 mg C12/1)
After filtration
After active carbon
filter
x)
CHC13
2,3
7.3
5,1
0,5
CH3rCl2
n.n.
16.5 '
lo.7
1.0
CHSr2Cl
n.n.
15.5
11*1
o.G
CHBr3
n.n.
3.o
2.4
o,2
• I TKM
2.3
42.3
29.3
2.7
TO Cl
35
—
195
55
All data in mg/m
n.n. = no trace
-------
- 174 -
Here too,large amounts of organic chlorine compounds other -
than the haloforms are produced in chlorination with 2 mg
C12/1. However, the variation of the concentration of these
compounds in the course of tha treatment is interesting.
The chloroform, which in comparison with the other trihalo-
methanes is more soluble in water, is also less efficiently
removed by adsorption in the active carbon filter, but in general
the total amount of haloforms is better removed by adsorption
than the polar organic chlorine compounds here grouped as
TOC1. The difficulty that these substances also present
from the point of view of treatment technology, means that
greater attention should be paid to them in water quality
control.
Summing up, it can be said that,in addition to the trihalo-
methanesr far greater amounts of highi-molecular chlorine
compounds can be produced. The compounds that can be deter-
mined by individual substance analysis represent only the
tip of the iceberg. The aim of further research must be to
learn more about the structure of these polar compounds and
about their formation reactions.
-------
- 175
Cl) KtiHNY' W.% SANDER, R.
Vorkojtimen und Bestaromung leichtfliichtiger
Chlorkohlenwasserstoffe
Hydrochem. hydrogeol. Mitt.' 3 (1 978) , -327-340
(2) STIEGLITZ, L. , ROTH, W. , KtiHN, W.,'LEGER, W.
Das Verhalten von Organohalogenverbindungen bei
der Trinkwasseraufbereitung
, Vom Wasser £7_ (1967), 347-377
(3) KfiHN, W., SONTHEIMER, H.
Zur analytischen Erfassung organischer Chlor-
verbindungen mit der temperaturprogrammierten
Pyrohydrolyse
Vom Wasser 43 (1974), 327-341
(4) KUHN, W.
Untersuchungen zur Bestimmung von organischen
Chlorverbindungen auf Aktivkohle
Dissertation, Universitat Karlsruhe (1974)
(5) KtiHN, W. , FUCHS, F. , SONTHEIM1R, H.
Untersuchungen zur Bestimmung des organisch gebundenen
Chlors mit Hilfe eines neuartigen Anreicherungsverfahrens
Z. f. Wasser- und Abwasser-Forschung ^ (1977), 192-194
(6) 'SANDER, R. , KtJHN, W. , SONTHEIMER, H.
Untersuchungen zur Umsetzung von Chlor mit
Huminsubstanzen
Z. f. Wasser- und Abwasser-Forschung j> (1977), 155-16O
(7) KtJHN, W. et al.
Use of ozone and chlorine in water utilities in the
Federal Republic of Germany
J. AWWA, June (1978), 326-331
(8) KtiHN, W. , FUCHS, F.
Untersuchungen zur Bedeutung der organischen Chlor-
verbindungen und ihrer Adsorbierbarkeit
Vom Wasser 45 (1975), 217-232
-------
- 176 -
REDUCTION OF THE'CONTENT OF CHLORINE COMPOUNDS BY A TREATMENT
COMBINING PHYSICO-CHEMICAL AND BIOLOGICAL PROCESSES
J. Chedal and P. Schulhof
INTRODUCTION
The waterworks which supply Paris with drinking water, and
notably the largest one, all treat surface water.
These plants were built about twenty years ago, following the
classical pattern of physico-chemical treatment:
- chemical pre-treatment comprising oxidizing agents,
in this case chlorine and chlorine dioxide, and an
adsorbing agent: powdered activated carbon;
- coagulation;
flocculation, decantation, rapid sand filtration;
final addition of chlorine before injection into the
distribution network.
Several years later a further treatment was added: ozonization
of the filtered water with a residual content of O.4 mg/1
maintained for over 1O minutes. Lastly, treatment aimed at
protecting the distribution network was instituted by means
of a combined treatment with chlorine and chlorine dioxide.
Precautions were taken at each stage of the plant construction
to ensure that the process could be adapted as time went by
with a minimum of labour, to keep pace with technical advances,
degree of river pollution, and quality objectives for the
treated water.
-------
- 177 -
»foday these precautions have proved very valuable., for. many
changes have taken place in the course of the last twenty
years.
In the first place, the pollution of the Seine has greatly
increased. Although the water to be treated is collected
upstream of Paris, the built-up area has spread and there
has been a great deal of urban growth upstream of the
water-collection point. The water has become increasingly
charged with ammonia (see Fig, 1). ,
65 66 67 68 69 7O 71 72 73 74 75 76 77
Fig. 1 Content of ammonia in Seine water
It is above all extremely rich in organic matter. During the
second half of 1977, for example, the mean amount of non-
et!
volatile organic carbon was greater than 5 mg/1, with a
maximum of 9.6 and a minimum of 2.8. On the other hand,
the water contains a relatively small amount of organo-
chlorine compounds and trihalomethanes. In the course of
1977, for example, the amount of chloroform varied between
O and 28 pg/1, with a mean content of 6.5 yg/1.
In parallel with this, during the last twenty years the
requirements on the quality of the treated water have
developed in a way that we all know.
-------
- ,178 -
In connection with this development, one of the first ob-
jectives to reach is discontinuation of the pre-treatment
with chlorine at the breakpoint. This is a process that
certainly eliminates ammonia, but it also leads to the
formation of a substantial amount of haloforms.
Another objective is to obtain treated water containing a
minimum of organic substances.
DESCRIPTION OF THE TESTS
A pilot plant was built to test the efficiency of the various
improvements. It comprises two lines of treatment, each capable
of processing 1O m /h, and in which the stages oif pre-treatment,
flocculation, decantation, filtration, and ozonization are
exactly the same as those in the full-size plant (Figs. 2-4).
Fig. 2
Flocculation and
decantation units
in the pilot plant
-------
_ 179 -
Fig. 3
Filters of the
pilot plant
Fig. 4
Post-ozonization
columns in the
pilot plant
During tiie various tests undertaken one line was equipped for
specific treatments, consisting of modifications or additions
as compared with the reference line. This enabled the effects
of these modifications on the water quality at all the different
stages to be determined.
The aim of the tests, whose results are reported below, was
to examine the effects of two modifications:
- preliminary ozonization of the raw water as an oxidation
process before coagulation and, in conjunction with this,
discontinuation of the chlorine pre-treatment;
-------
- 180 -.
filtration through activated carbon instead of sand,
Several of the filters, with a specific surface of 117 m ,
were filled with activated carbon. The qualities of carbon
and the gradings most suitable for satisfactory filtration
were subsequently tested, and the optimum treatment parameters
were determined. It was found, in this respect, that the life
of the activated carbon filters between two treatments v/as
longer than that of sand filters. We were also able to verify,
as others have done before us, the rapid disappearance of the
adsorptive properties of the carbon, except in matters of
flavour. However, during these tests in the full-scale plant,
s
chlorine-oxidized water was passed through the filters. It
was therefore interesting to judge the effect of a carbon
filter on water oxidized by ozone used in pre-ozonization.
Consequently, the following general, scheme wad adopted in
the pilot tests (see Fig. 5): ....
Pre-ozoni-
WAC
ClONa ClOj
carbon
sand
resx-
. dual' O,,0,4ppm,CI,
MM ELL
io+
. Chemical Flocculation
treatment Post-
FILTRATION ozonization RESERVOI*
DECANTATION
Fig. 5
General scheme of the experimental lines
-------
- 181 - '
- In the control line (ref. No. 1) the treatment corres-
ponds to that in the plant, whose characteristics are
as follows:
1) At the pre-treatment stage: chlorination at the break-
point with Javel water;
/
addition of chlorine dioxide
to eliminate phenols, man-
ganese, and certain flavoursi
coagulation by means of
poly-aluminium chloride.
2) After the O.4 g/m residual post-ozonization treatment:
..-.-, post -chlorination.
- In the experimental line (ref. No, 2), the treatments
differ from the control line as follows:
pre-ozonization carried out with semi-industrial
emulsifier-type equipment. The total contact time
is about 2 min, but the characteristic feature of
the treatment is the time of contact of a bubble
of ozonized air with the water, which is about
2O sec.
Omission of the chlorine pre-treatment.
Filtration through activated carbon instead of filtration
through sand as used in the control line.
The purpose of the experimental scheme for this line was to
favour bacterial development: oxygenation of the raw water,
omission of the chlorine pre-treatment, and suitable filtration
material. In the rest of this report, this will be known as
the biological scheme.
The tests were conducted with a view to determining the
effects of the pre-ozonizatiori conditions on the quality
of the water at different stages of treatment. Consequently,
the following operations were carried out during the essen-
tial phases of the tests:
-------
- 182 -
1. Operation for several months with constant pre-
3
ozonization conditions: treatment dose of O.25 g/m
and an air concentration of 1.5 g of ozone per m ,
or an air/water ratio of 17 %.
2. Operation with pre-ozonization treatment dose vary—
3
ing between O.2O and 4 g/m , bi
air/water ratio of about 16 %.
3
ing between O.2O and 4 g/m , but with a constant
3. Variation of the two characteristics of pre-ozonization,
i.e. the treatment dose and the air/water ratio. This
phase of the trials is currently in progress, and we
can only give the initial results.
During all these experiments the treatments common to both
lines were applied under the same conditions. The treatment
dose of coagulant was determined by the jar test, and the
post-ozonization was chosen with a view to obtaining a resi-
dual virulicidal content of O.4 mg/1 after 1O min of contact.
We shall discuss the quality of the water tested at different
stages of treatment during the various phases outlined above.
EFFECT OF PRE-OZONIZATION ON RAW WATER
As regards the raw water, the results obtained over a period
of three months (trial with pre-ozonization at a constant
strength of O.25 ppm) enable us to draw the following basic
conclusions (see Fig. 6).
The pre-ozonized water is richer in suspended matter, it is
more turbid and more clogging. This development is due to
the decomplexing and coagulating properties of ozone.
-------
- 183 -
Pre-ozonization
t
Before
After
Turbidity
(drops of
mastic)
1)-00
^30
Suspended
matter
(mg/1.)
5*
72
Beaudrey
clogging
power
U-1 )
\
7
9.4
N.V.T.O.C.
(mg/1)
7
7
UV absorp-
tion at
250 nm
(O.D.
. 10-3 cm"1)
1W
. 139
Fig. 6 Quality of the raw water before and after
pre-ozonization
As regards organic matter, the smaller value of UV absorp-
tion in the pre-ozonized line certainly confirms the action
of ozone on organic molecules - the total amount of organic
matter remains unchanged, as indicated by the values of non-
volatile organic carbon.
When the intensity of the pre-ozonization treatment is in-
creased, the difference in UV absorption between the two
treatment lines also increases, as long as the pre-ozonization
treatment dose does not exceed about 1 mg/1. Beyond this
threshold level the difference between the UV absorptions
remains essentially constant (see Fig. 7).
This result shows that pre-ozonization does not eliminate
organic matter but modifies their structure. This action
increases with increasing doses of ozone until about 1 mg/1
is reached.
The results obtained with an air/water ratio of 16 % for the
pre-ozonization seem to be fairly analogous regardless of the
air/water ratio, as shown by the first results of the most
recent tests still in progress.
-------
- 184 - •'
w
30
20
K>
at 250 urn
o
o o
125 2 Pre-ozonizatio
dose (rag
Fig. 7
"Reduction of UV absorption
as a function of the
pre-ozonization dose
The amount of oxygen dissolved in the water increases sig-
nificantly as a result of pre-ozonization. During the last
tests, carried out in August 1978, .pre-ozonization enabled
the amount of dissolved oxygen to be increased from a satur
ation level of 75 to 85 % in relation to the atmosphere.
AMMONIA CONTENT IN THE TWO TREATMENT LINES
Ozone has no direct effect on'the elimination of ammonia.
There is therefore no difference between the ammonia concen-
trations in pre-ozonized raw water and initial raw water.
At the stage where the water is decanted, the experimental
results show that after a period of one month there is a
very marked reduction of the ammonia level in the pre-
ozonized line. The values obtained are of the same order as
those in the control line, in-which the breakpoint pre-
treatment was kept up during the firs,t five months of the
tests (see Fig. 8). .-.,-•
-------
- 185 -•
NH.
ppm
0.5.
raw water
decanted water 1
decanted water 2
filtered water 2
x*i B«c.ll JM.JI Feb. March April
Pig. 8 Ammonia contents
June July
In the pre-ozonized line biological elimination took place
in the layer of sludge formed in the flocculators,
During the subsequent filtration the same test -shows that the
ammonia remaining after decantation is easily eliminated. A
delay in inoculation was, however, necessary for the new
activated carbon introduced into the filter.
Nitrification seems to occur principally in the upper layer
of the filter, as evidenced by the counts of nitrite and
nitrate germs at different depths.
Prom the beginning of June 1978 the treatment at the break-
point was discontinued in the control line. It was ascertained
(see Pig. 8) that nitrification was just as effective in the
decantation unit without pre-ozonization. Under these con-
ditions, this treatment contributed nothing to the bio-
logical elimination of ammonia.
For the amounts of ammonia of the order of 2 ppm that can be
encountered in the river, .the dissolved oxygen will, however,
be insufficient. Pre-ozonization then will provide enough
oxygen to bring about nitrification.
-------
- 186 -
Thus, although pre-ozonization has no direct effect on the
elimination of ammonia, it does seem to be useful in this
respect for bringing about biological nitrification when
the amount of ammonia present exceeds a certain limit.
However, the essential advantage of replacing chlorine by
ozone in the pre-oxidation is that no halogenated compounds
are formed during the pre-treatment.
The content of chloroform in the treated water after the
final protective chlorination treatment was markedly smaller
in the biological line (Fig. 9). This infers a significant
elimination of precursors by a biological route.
A x raw water
treated water,
,.„__ control "*-* —
__._ treated water,
biological line
IM. 77
Jit. 79
Feb.
March
Fig. 9 Chloroform content
ELIMINATION OF ORGANIC MATTER FROM DECANTED WATER
The effect of pre-ozonization in the pre-treatment on
coagulation - flocculation - is very marked. The flocculate
in the treatment line, where pre-ozonization had been used,
was far coarser, as can be seen from" Fig. 1O.
-------
- 187 -
Flocculate in the Flocculate in the
pre-ozonized line non-preozonized line
Fig. 1O
The effects of the pre-ozonization can be noticed from a
dose of about 0.2 ppm onwards. Pre-ozonization is therefore
more effective as an oxidizing pre-treatment for coagulation
than is pre-chlorination. This advantage indicates an im-
proved elimination of the organic matter in the line with
pre-ozonization.
What is the situation with respect to the organic carbon
parameter? The results obtained with UV absorption of the
raw water have shown that it might be useful to proceed
to pre-ozonization with relatively hi'gh doses (about 1 ppm) .
This supposition was confirmed for the decanted water, where
it had been found that the difference between the organic
carbon contents in the two treatment lines increases in
favour of the biological line, with an increasing pre-
ozonization dose of up to about O.6 mg/1 (cf. Fig. 11).
Above this dose the difference between the organic carbon
contents is O.7 ppm.
When the treatment at breakpoint in the control line is
discontinued, it seems that this difference is not altered
to any significant degree. The results obtained so far
indicate a mean difference of 0.6 ppm.
-------
- 188 -
NVfOG
0,5
0.6
Fre-ozonlzation
dose (ing Oyl)
Fig.11 Seduction of the content of non-volatile organic
carbon in decanted water as a function of pre-
ozonization dose
Thus, pre-ozonization is certainly the reason for the better
results obtained with decanted water in the biological line.
ELIMINATION OF ORGANIC MATTER IN THE FILTERED WATER
The first measurements in filtered water showed very low
contents of organic carbon in the biological line. This was
in no way due to the biological activity of the filter, but
to adsorption on the activated carbon, which had only just
been placed in the filter.
At the end of several months of operation at the rate of
4 volumes per hour the findings corroborated earlier ex-
perience and the results reported by many other investi-
gators: the effects of adsorption are then very much reduced
-------
- 189 - V- T •
and essentially constant. Any improvements observed in
the filtered water are thus a result of biological activity.
The measurements taken in the course of the test with vary-
ing pre-ozonization doses show differences between the organic
carbon contents in .the two lines which are much the same for
the filtered water and for the decanted water.
However, the results obtained are certainly not wholly in-
contestable . Thus , the time after every change in the pre-
ozonization conditions, about a week, was perhaps too short
to have allowed the bacterial population of the filter to
adapt itself to each new situation.
In order to evaluate the biological activity of the filter
with respect to organic matter, we also measured the content
of detergents, compounds that are easily biodegradable. The
results obtained show that during filtration through activated
carbon this parameter was only reduced to about 7O % (cf.
Fig. 12) .
100
50
Q
Hit. nc.77 Jn.78 Feb jferfch April May June July'
Fig. 12 Reduction of the amount of detergent by
filtration through carbon (biological line)
-------
- 190 -
It is interesting to compare this finding with those obtained
in a test on filtration through carbon, carried out with
filtered and ozonized water produced in a large plant. It was
noted at the time that the amount of detergents was totally
and constantly reduced, even though the water had been pre-
treated at the breakpoint. However, this total elimination
was obtained only when the flow rate had been reduced to
12 m/hf which corresponded to 4 volumes per hour (see Fig. 13).
V = l2.«/h
2 34 56 789 10 11
Fig. 13 Reduction of the amount of detergents by
secondary filtration after post-ozonization
Subject to the speed of filtration through activated carbon
not being too great, these results lead us to the conclusion
that far more efficient biological activity as regards
organic matter can be obtained on a filter after secondary
ozonization (ozonization after complete clarification) than
on a filter placed in first filtration, even preceded by
pre-ozonization. It is highly probable that the lower bio-
logical yield with respect to organic materials observed
with the latter type of filter is due to the periodical
washing to which the filter is subjected.
-------
- 19-1 .--'-
We are therefore continuing with the pilot tests, studying
the efficiency of filtration after secondary ozonization
to complete the elimination of organic substances already
achieved in the biological line that we have been testing.
The initial results of these trials have been fairly en-
couraging.
EFFECT OF PRE-OZONIZATION ON VIRULICIDAL POST-OZONIZATION
Another very important action of pre-ozonization is the
effect that it has on the behaviour of water during post-
ozonization. This treatment was tried out in a pilot line
with a view to obtaining a residual virulicidal content
of 0.4 g/m in the water for a time of 10 min.
It was found that, in order to arrive at this result, the
treatment doses required should be 40 % to 5O % lower in
the biological line than in the chemical line. This saving
on ozone in post-treatment remains fairly constant for all
pre-ozonizations above 0.25 ppm (cf. Fig. 14).
70
60.
50.
4O
3O
20
10.
4.Reduction of
post-ozonization
rdose %
Pre-ozonization
dose,
0.5
Fig. 14 Reduction of the post-ozonization treatment dose
as a function of the pre-ozonization dose
-------
- 192 -
Since the mean post-ozonizatidn dose in the treatment of
Seine water is 2-. 5 ppm, the application of pre-ozonization
at 1 ppm will not necessitate the installation of supple-
mentary ozone-generation equipment. All that will be re-
quired for pre-ozonization is a different distribution of
the existing production. Thus, pre-ozonization can be
considered to be "gratis".
BEHAVIOUR OP TREATED WATER IN THE DISTRIBUTION NETWORK
To compete the appraisal of the treatment lines tested, the
pilot line was equipped with two mini-networks made of un-
lined cast-iron pipes 10O mm in diameter. The water moved
very slowly in these pipes, to obtain a residence time of
4 days (see Fig. 15). Sampling points were provided to allow
the water to be collected after residence times of 1, 2,
3 and 4 days.
Fig. 15
Mini-network in
the pilot plant
The counts of the banal bacteria present in these samples
enabled the risk of bacterial re-growth in the water
produced to be calculated. The figures obtained for the
water produced in the tests show that in the biological
-------
- 193 .-
line the maximum count reached was about 2OO bacteria/ml,
and in the control line only 15O/ml. This difference is of
little significance, and it can be assumed that bacterial
re-growth will be identical for the two types of water
(cf. Fig. 16).
tN/ml
to1
10'-
eontrol
.* biological
Days
Fig. 16
Banal bacteria counts
in the mini-network
CONCLUSION
The results of the tests indicate that the pre-ozonization
treatment of Seine water before coagulation presents many
advantages over the chlorine pre-oxidation practised so far.
- It enables the pre-chlorination treatment at the breakpoint
to be discontinued, for it constitutes the first link of a
chain in which ammonia is eliminated in a biological way.
-------
- 194 -
This solution to the problem entails a reduction of the amount
of organochlorine compounds in the'water produced, and in
particular of haloforms, this being due to the elimination
of precursors in the course of the biological clarification
process.
— Pre-ozonization leads to an appreciable improvement in
the elimination of organic matter, essentially in the
decantation units.
It does not appear, however, that this improved elimination
occurs by the biological route during filtration through
activated carbon.
The use of a second filtration after secondary ozonization
is undoubtedly necessary to obtain a complementary increase
of elimination of organic substances by a biological route.
- Lastly, it must be pointed out that pre-ozonization does
not entail any supplementary use of ozone, bearing in
mind the economies that such treatment would bring to
post-ozonization.
-------
— 195 - '
PHYSICAL-CHEMICAL PRETREATMENT FOR THE REMOVAL
OP PRECURSORS
J.M. Symons and A.A. Stevens
INTRODUCTION
The reaction of chlorine and precursors to yield
chlorinated organics, as represented by Figure 1,
would indicate three possible ways exist to attack the
trihalomethane problem: rj.) is the removal of the pre-
cursors by some treatment technique; 2) is to replace
chlorine with some alternate disinfectant, or 3} is to
remove the chlorinated organics once they are formed.
This paper will review what is' known about removal of '
precursor as a trihalomethane control procedure.
Discussion of the other two techniques can be found
elsewhere (1,2,3).
CHLORINE + PRECURSORS
Fig. 1
CHLORINATED
ORGANICS
-------
- 196 -
REMOVAL OF PRECURSORS
For removal of precursors, three .possible techniques
have been explored: precipitation, oxidation, and ad-
sorption, precipitation being either during lime soften-
ing or during turbidity and color, removal with a co-
agulant, oxidation either "with ozone, chlorine dioxide,
or potassium permanganate, and adsorption either with
powdered or granular activated carbon.
Precipitation
Pilot plant data would indicate that the amount of
chloroform that is formed during treatment can be
.changed depending on the point of application of the
chlorine (see Figure 2), either ahead of the flocculation
basin, ahead of the filter, or after filtration. This
1 25
RELATIVE
TERMINAL
CHLOROFORM 10
CONCENTRATION
IN FINISHED
WATER AFTFR 75
2 DAYS AT
GIVEN POINT OF
CHLORINATION s
COMPARED TO
CHLOROFORM
FORMATION
POTENTIAL IN
RAW WATFR
RIVER
25
RAW WATER TERM
CHCI3 CONC
FERRIC
}SULFATE
(COAGULANT
E
ALUMINUM
SULFATG |
COAGULANT
t
2 DAYS RAW
vWATER
..STORAGE
O>
© .
"
DISTRIBUTION
SYSTEM
COAGULATION/" FILTRATION
FLOCCULATION.
SETTLING
Fig. 2 Chloroform in finished water relative to
point of chlorination (pilot plant studies)
-------
- 197- -
indicates that precursors are being * removed'during treat-
ment, and the later in treatment that chlorine is
applied, the less chloroform is produced. In this par-
ticular water from the Ohio-River, ferric sulfate was
a more effective coagulant and.sb.the solid part of the
bars in Figure 2 indicates the ferric sulfate data,
the total bar the aluminium sulfate data.
Raw water data are also indicated in Figure 2 for the
terminal trihalomethane concentration and a question
frequently arises as to why the raw water terminal
trihalomethane concentration is higher than the amount
of trihalomethane produced when raw water is chlorinated
in a treatment plant. The explan'atioi is that determi-
nation of terminal trihaIbmethanes is a test that is
done in a bottle in which raw water and chlorine are
mixed and held for some time and the resultant trihalo-
methane measured. In the 'treatment plant the chlorine
is added, and as the water is passing through the co-
agulation and sedimentation basin the precursor and the
chlorine are being separated by the settling process.
Less trihalomethane is produced under these conditions
than would be produced if the raw water was chlorinated
in a bottle with precursors and chlorine in contact
for the'duration of the holding period. Therefore, the
raw water terminal trihalomethane is used to indicate
the potential in the water, but this potential is not
realized even when the"raw water is chlorinated during
treatment.
Precursor removal by coagulation and settling was first
(1975) demonstrated on a full ;scale at Cincinnati,Ohio,
U.S.A. Figure 3 is a diagram of 'the Cincinnati Water
Works. Formerly they added chlorine at point A, where
chlorine and coagulant were both added. Then they moved
the point of chlorination from point A to point B, the
head end of the treatment plant. Because alum was added
before presettling (point A); "some particulate removal
-------
- 198 -
occurred in these b'asins so the water quality at point
B is considerably better than it would be at point A.
The results of this change of chlorination practice
are shown in Figure 4. After the point of application
of chlorine was moved from point A to point B, the
chloroform level dropped very dramatically.
In order to demonstrate that this was really caused by
the moving of the point of application of the chlorine,
and not because of a change in the river quality, the
raw water trihalomethane formation potential was
measured at various times throughout this period to
show that precursors still did remain in the raw water.
Even though a slight downward trend in concentration
may have occurred through the winter months (Figure 4),
it was certainly not equivalent to the dramatic drop
in chloroform concentration that occurred when the
point of chlorination was moved. Note that the bromin-
ated compounds did not change nearly as much in con-
centration on a percentage basis as did the chloroform.
OFF-STREAM STORAGE RESERVOIRS
3 DAYS)
POINT B
WATER
TREATMENT-^
PLANT
FILTERS
POINT C
CLEAR WELL
(TTt
DISTRIBUTION SYSTEM
CHLORINE, ALUM-1 PUMPING
POINT A ( ^STATION-
\ INTAKE
Fig. 3 Schematic of Cincinnati Waterworks
-------
-'199 -.
300r
MOVE CHLORINATION
2600 FROM POINT "A"
TO POINT "B"
-JULY 14, 1975
220
180 -
TRIHALO
METHANE
CONC.
14O -
100 -
60 -
20
NF
A
-CHLOROFORM'
A
KEY
RAW WATER CHLORINATED
AND STORED:
a 3 DAY TERMINAL CHLO
RQFORM CONCENTRATION
©4 DAY
' : . •
A 6 DAY
-BROM ODICH LOROM ETH A NE
,DIBROMOCHLOROMETHANE
Fig. '4 '
Trihalomethanes in
Cincinnati, Ohio
Another trial was conducted in the late summer of 1977.
In this case, the point of application of chlorine was
moved from the rapid mix (hydraulic jump) to a point
following sedimentation and before the rapid sand
filters. Table I shows the trihalomethane concentration
in the distributed water for 1976 during mid-September
and early October and contrasts these data to those
collected during the 1977 study. The generally lower
concentrations during the 9/28/77-1O/7/77 test again
show the advantages of chlorinating as high a quality
water as possible.
Figure 5 is a schematic diagram of the Daytona Beach,
Florida, U.S.A., treatment plant. This utility did a
study in which -they chlorinated at three different'
points, 1) chlorinating the raw water, 2) adding chlorine
at the recarbonation basin, and 3) adding chlorine at the
.clearwell. All of the trihalomethane concentrations were
measured at sample point No. 5, which was several hours
flow time following the clearwell.
-------
- 2OO X'.
TABLE I INFLUENCE OP MOVING POINT OF APPLICATION OF
CHLORINE AT THE CINCINNATI WATERWORKS
Control Year
Point of Application
Bate of Chlorine
9/16/76 Hapid Mix
21
23 "
28 "
29 "
3O
10/5
7 "
12
14
ZTHM in
Distributed
Water
Jif/1
12O.1
117.5
117.8
119.7
108.9 . .
115.8 •
109.7
114.O
92.1
109.0
Experimental Year
Point of Application
Date of Qilorine
9/13/77 Eapid Mix
21 "
28 Settled Water
. , .29- .'
30
,;10/1
2 "
3 "
4
. . .5 - "
6
. • 7
1O ; Rapid Mix
•11
EIHM in
Distributed
Water
jig/I.
no. a
107.4
81.5
1O2.4
76.0
89.2
87.4
116. 0
78.1
77.7
77.7
81.0
112.9
89.9
-------
- 201 - ,
CUMWtU.
Fig. 5 Flow diagram Ralph F. Brennan Water Plant,
Daytona Beach, Florida
3UU
450
400
350
300
iTHM,
U9/l 250
200
150
100
75
50
25
ft
DAYTONA BEACH, FLA.
T
1
1
ti2
RAM
NO
C02
^
'/////////////A
i
L }
f
co;
W/////////A
CI2
SETTLE
2 NO
CO2
^
^
§>
"T
TERM
, IN RAW "
WATER
i
i
D El INST iTHM
f—i^THM FORM,
L— J POTENTIAL
G3«QTERM STHM
CO2 CI2
W/////////A
FILTtRtD
NOCO2
CO2
^ ^
^ ^
$0 V^
^ V
Fig. 6
-------
- 2O2 - . .
The first two bars in' Figure 6 are weekly average data
resulting from chlorinating the raw water; the second pair
from chlorination of the settled water; and the third pair
from chlorination of the filtered water. The cross-hatched
section of the bars shows the instantaneous values that
were measured at sample point 5 and the open section addi-
tional trihalomethane formation potential that would be
exerted out in the distribution system.
The series of studies were conducted both with and without
recarbonation, and recarbonation had some effect on the
instantaneous trihalomethane concentration because of
stripping. In summary, Figure 6 shows two important points:
one, the raw water has essentially the same formation poten-
tial or precursor concentration throughout the study and,
secondly, the benefits of moving the point of application of
chlorine are shown. The removal of precursor through coagu-
lation and sedimentation was again demonstrated. The project
is continuing in Daytona Beach attempting to demonstrate
whether or not this effect can be improved even further
through the use of the addition of polyelectrolytes.
Oxidation
Ozone
Table II shows the effect of ozonation as an oxidant on
pilot plant filtered water in Cincinnati, Ohio, U.S.A.
Ozone alone does not produce any trihalomethanes.
For the first study without ozone and with chlorine,
2O yg/L of trihalomethane was produced. With O.7 mg/L
of ozone plus chlorine, 23 yg/L of trihalomethane was
produced. The effect of ozonation was negligible. The
same occurred with a dose of 18.6 mg/L of ozone. When
a very high dose of ozone was used, 227 mg/L, some
oxidation of precursor occurred. Therefore, ozone can
be effective in certain circumstances, but in the
-------
- 2O3 -
Ohio River water it certainly was not particularly
effective for precursor oxidation at this time.
Table II EFFECT OF OZONATION OF DUAL MEDIA EFFLUENT
Contact Time - 5-6 min.
Applied 0.3
Dose
mg/L
0.7
O
0.7
18.6
O
18.6
O
227
Chlorine
Dose
mg/L
O
8
8
O
8
8
8
8
animation Trihalctnethane
Cone, after 6 days
yg/t-
< O.2
2O
23
< O.2
23
30
123
70
Figure 7 presents the results of another study per-
formed with Ohio River water during a different time
period showing that in this case a big decline in
precursor occurred. This study also included an
evaluation of the effects of ultraviolet radiation
which further enhanced the effects of ozone on re-
moval of precursors. The test system was a 22 liter
batch reactor receiving 2O mg of applied ozone per
-------
.. -,-204 -
minute. Although this study was also conducted on
Ohio River water, the results were quite different
from those shown in Table I.
The variability in the observed effect of ozone on
trihalomethane precursor removed is further shown
by Table III as compiled by Trussel. (4) These data
from the independent work of several investigators
range from negative removal to 9O percent. Several
variables undoubtedly account for the observed wide
range of results: nature of precursor, ozone dose,
contact time, contactor design, including the dis-
persion system, and possibly other water quality
factors.
7-DAY
TRIHALO-
M ETHANE
FORM,
POTENTIAL
(um/1)
2.10
2.00
1.90
1.80
1.70
1.60
1.50
1.40
1.30
1.20
1.10
1.00
.90
.80
.70
.60
.50
.40
3O
»»JW
.20
.10
0
t
I
\
A
!\
d. _^__ __ ~_
all— -*: ^~~-»_OXYGEN ONLY
lj\ / \ .---*x~~~*
\ V %*'-^..-*""" *\
J'^ \UV ONLY
P1
i\\
ll
'^\
* \T*^
\ '^">».
&\ ****
\ Sm--~, OZONE ONLY
\. ""*"*
^V"*
\v.
Tfc:-s='-'~:t" • OZONE
""*=« PLUS UV
0 15 30 60 90 120
TIME (WIN.;
Fig. 7 Ozone and ozone-UV destruction of THM precursors
Ohio river water (2/77)
-------
- 2O5 -
Table III PRECURSOR REDUCTION WITH OZONE
Location
Owens River
Lake Casitas
Columbia River
ii
n
ii
Ohio River
(Louisville)
n
ii
M
Bay Bull's Big Pond
Ohio River
( Cincinnati)
. Mokelumne
Middle River
Rotterdam
Orange County
Caddo Lake, Texas
Ohio River
Dose, mg/L
1 .0
2.0
0.5
1.0
2.0
4.0
1 .O
2.0
4.0
6.O
8.0
1 .0
2.0
3.O
18.6
227
2.0
3.4
4.5
6.O
2.6
2.8
5.5
10
11
2
8
1.0
O-72
O- 109
% THM Precursor
Reduction
78
6
8
14
16
16
6
22
3O
46
46
13
19
27
-31
43
62
59
59
53
-13
- 3
32
: ' 7
22
6O
50
7
O-9O
O-8O
from R. Trussel, 1978 (4)
-------
- 206 -
- Chlorine Dioxide
The data presented in Figure 8 show some removal of
precursor by application of chlorine dioxide. Th**
upper curve is the trihalomethane formation with
chlorine alone, the middle curve is similar data
with chlorine dioxide and the same amount of chlo-
rine. The chlorine dioxide has some effect on the
precursor such that it is changed so that it will
not react with the chlorine to form trihalomethanes.
The lower curve shows that chlorine dioxide alone
does not produce trihalomethanes.
I 0.6
3
Z0.5
lo.4
JE 0,1
F.A. Cl = FREE AVAILABLE CHLORINE
pH = 7.4
T = 24°C
O 0.3 «-~ - m9/' F-A'CI
O
O
mg/l CIO2 + 1-5 mg/l F.A. CI
ClOa ALONE
UJ
< O 10 20 30 40 SO 60 70 80
35 CONTACT TIME, HOURS
DC
Fig. 8 Trihalomethane formation by C1O2 and excess free
available chlorine, ERG Pilot Plant settled
water (6}
-------
- 207 -
Potassium Permanganate
Table IV presents the results of some preliminary
studies investigating the effects of potassium
permanganate on natural precursors from the Ohio
River. In these experiments, the percent removals
of precursor are relatively small, the highest is
19 percent, the lowest is 3 percent. Thus, potassium
permanganate does not have large influence on the
precursors that are in the Ohio River, although it
does do some good. As mentioned in a previous paper
(5), permanganate may have a more dramatic effect
on precursors in some waters where precursors were
observed to react more like m-dihydroxy phenyl
moieties.
Table IV PRELIMINARY DATA ON PRECURSOR REMOVAL FROM
OHIO RIVER WATER BY POTASSIUM PERMANGANATE
KMnO. Reaction
Amount
Added
(mg/L)
O
5
0
5
0
5
0
5
Reaction
Time
(hours)
1.5
1.5
1.5
1.5
1.5
1.5
0.5
0.5
pH
7.1
7.1
9.3
9.3
10.2
10.2
Neutral 7.O
Neutral 7.0
Chlorine
Reaction
Time
(hours)
2
2
2
2
2
2
30
30
ZTHM
% Removal
pg/L
30.
25.
52.
50.
57.
54.
120.
97.
5
7
3
9
2
0
5
6
15.4
2.7
5.6
19.0
-------
- 208 •-
Adsorption
- Powdered Activated Carbon
Figure 9 presents some typical data on the performance
of powdered activated carbon (PAC) for precursor
removal. In this experiment chloroform formation
potential declined some initially, and then decreased
more slowly as a 100 mg/L dose was approached. These
data would tend to indicate that a 5 or 1O mg/L PAC
dose would give an effect that might be favorable if
a finished water had a trihalomethane concentration
just above the proposed regulation, but PAC could not
completely eliminate precursor even with an uneconomical-
ly high dose.
o>
-i 30
P
I 25
O
Q.
I 20
O
u.
5 1(=
CE 1£>
O
u.
O
§10
2
O
O 5
Q
in
2 MINUTES RAPID MIX
5 MINUTES SLOW MIX
30 MINUTES SETTLING
DO 10 20 30 40 50 60 70 80 90 100
N POWDERED ACTIVATED CARBON DOSE, mg/l
Fig. 9 Influence of powdered activated carbon on
chloroform formation potential
-------
- 2O9 -
- Granular Activated Carbon , ',.,.....•
Figure 1O shows data from a treatment plant in
Huntington, West Virginia, U.S.A. The precursor is
nearly completely removed initially and then the
concentration steadily rises in the adsorber effluent,
reaching the proposed maximum contaminant level (MCL)
after about four weeks. Notice that the empty bed
contact time here is only 6-1/2 minutes, which is a
short empty bed contact time. Nevertheless, when the
granular activated carbon is fresh, it will remove pre-
cursor well, but then a slow steady break-through
occurs. Other data with deeper beds (longer empty bed
contact time) tend to show that the time period for
break-through to the proposed MCL is longer, Figure 11.
Figure 12, data frdm the granular activated carbon
treatment of water with a very high concentration of
trihalomethane precursor;again demonstrates the in-
fluence of empty bed contact time.
SUMM,
TRIHALO- O.I SO
METHANE
FORM.
POTENT.
CONC.
me/I
O.14O
0.120
[SAND REPLACEMENT
(ADSORBER""" , PROPOSED [MCL| \
""'NFLUENT f I
SAND REPLACEMENT!
ADSORBER EFFLUENT ;
«PPROX, EMPTY BED CONTACT TIME = 6.5 MINUTES
II 15 20 25 27291 3 5 B 10121517 1922 24 26 29 31
JULY AUGUST
Fig. 1O
Removal of trihalomethane
precursors by granular
activated carbon beds
-------
- 210 -
24 MiN. EBCT
INFLUENT
a
PROPOSED MCL
v. o
b".
-7" "
0 W 20 30 40 50 60 70 60 90 100 110 120 130 "0 1&Q WO -70 iflO
Days ol run
Fig. 11 Performance of terminal THM through Post Filter
Adsorber run 1 (Feb.-Oct.1977) Jefferson Parish, La^
0.80
0.70
0.60
<
0.50
TERM.
SUMM,
TRIHALO °-40
METHANE
ARROW INDICATES PROPOSED
• P MCL EXCEEDED 1
/\ / —INFLUENT j
/
/
! ' P
*
• 1
'* '* '
^ :
*-* " i
5 10 15 20 25 30
TIME IN SERVICE. DAYS
35
Fig. 12 Example of the influence of empty bed contact
on terminal summation trihalomethane removal
-------
- 21 1
BToiogicaily "Active" 'Carbon Fi Iters
A study on the utility of preceding granular activated
carbon by ozone treatment on the removal of trihalo-,
methane precursor was conducted in a pilot plant
treating settled Ohio River water. The data in Figure 13
show that preceding the adsorption system" by an ozone
dose of about 1 mg/L, contact time 2O minutes, did re-
sult in a reduction in trihalomethane precursor in the
adsorber effluent at any given service time when
compared to conventional g'ranular activated carbon.
Both adsorption systems had a 9 to 1O minute empty
bed contact time. The mechanism for this improved
performance is thought to be related to an enhance-
ment of biological activity in the activated carbon
adsorber brought on by the ozone pretreatment.
0,1
0,08
--O.OS
STHMFP
CONC,
ma/l 0.04
0.02
1 1 1 r— 1 r
VALUES COMPUTED FROM MONTHLY AVERAGES OF
WEEKLY DETERMINATIONS 2 DAV, 25°C SUMMATION
TRIHALOMETHANE FORMATION POTENTIAL
4567
TIME IN SERVICE. MONTHS
10
Fig. 13 Influence of ozonation prior to adsorption
on trihalomethane formation potential removal
-------
- 21 2 -
SUMMARY
Precipitation, oxidation, and.adsorption are all
methods of removing trihalomethane precursor. Of
these various techniques, only adsorption on fresh
granular activated carbon is completely effective,
although the capacity of the activated carbon is
finite. The other techniques, alum or iron coagula-
tion, oxidation with ozone, chlorine dioxide or
potassium permanganate, or adsorption with powdered
activated carbon, are only.partially effective but
may still have a role in- controlling trihalomethane
concentrations in some cicumstances.
-------
- 213 -
(1) SYMONS, J.M. -'•"'•' '
Interim Treatment Guide for the Control of Chloroform
and Other Trihalomethanes
U.S. Environmental Protection Agency, Cincinnati, Ohio,
48 pp. plus 4 Appendices, Unpublished (June 1976)
(2) LOVE, O.T., Jr., CARSWELL, J.K., MILTNER, R.J., SYMONS, J.M.
Treatment for the Prevention or Removal of Trihalomethanes
in Drinking Water
J. AWWA (in Press)
(3) SYMONS, J.M., CARSWELL, J.K., CLARK, R.M., DORSEY, P.,
'GELDREICH, E.E., HEFFERNAN, W.P., HOFF, J.C., LOVE, O.T.,
Jr., McCABE, L.J., STEVENS, A.A.
. Ozone, Chlorine Dioxide, and Chloramines as Alternatives
to Chlorine for Disinfection of Drinking Water, State-of-
the-Art
Summary in Proceedings of -Second Conference on Water
Chlorination: Environmental Impact and Health Effects,
Gatlinburg, TN, Oct. 31 - Nov. 4 1977 (1978), 555-56O,
Complete version available from Director, Drinking Water
Research Division, U.S. Environmental Protection Agency,
Cincinnati, OH, 45268
(4) TRUSSELL, R.R.
Factors Influencing the Formation of Trihalomethanes
O'ames M. Montgomery Consulting Engineers, Inc., Pasadena,
CA, presented at the California-Nevada Section AWWA
Seminar on Organics in Domestic Water Supplies, Palo Alto,
CA, April 12, 1978
(5) STEVENS, A.A. et al.
Formation of Non-Polar Organochloro Compounds as By-
products of Chlorination
Presented at Conference on Oxidation Techniques in Drinking
Water Treatment, Sept. 9-13, 1978, at Karlsruhe, Federal
Republic of Germany
(6) MILTNER, J.J.
The Effect of Chlorine Dioxide on Trihalomethanes in
Drinking Water
M.S. Thesis, University of Cincinnati (1976)
-------
- 214 -
UNWANTED BY-PRODUCTS OE CHLORINATION
B. Josefsson
It is a challenge for an analytical chemist to determine
the different organic compounds (in the nanogram per
litre level) which are present in drinking water. For
practical reasons I think it is meaningless to analyse
all of them.
A hierarchic analytical system for the analysis of
chlorinated compounds should begin with the determina-
tion of elemental chlorine for rapid screening of many
samples. In this way it is possible to locate sources
of organic chlorine compounds to the raw water or
during water treatment processes. ;When this procedure
reveals a source, a more laborious analytical determina-
tion of specific compounds can be performed. Dr. W. Kuhn
showed in his lecture that elemental chlorine analyses
can be carried out with the DOC1 (dissolved organic
chlorine measured by pyrohydrolysis) method and DOC1N
(dissolved organic ch^orine-nonpolar, measured with
microcoulometry on a nonpolar solvent extract) (1).
Other methods in this respect include non-destructive
neutron activation analysis (NNAA) on elemental chlorine
and bromine (2). When NNAA techniques are used it is
possible to distinguish between chlorine and bromine,
which cannot be done by coulometric methods, for example.
Liquid-liquid extraction with a nonpolar solvent is
very practical for the determination of lipophilic
organic halogens which can easily bioaccumulate. By
using this method organohalogenated compounds can be
tested for possible persistence by treating the extract
with cone, sulphuric acid, UV light etc. Bioaccumula-
tion and persistence are important in view of chronic
-------
' - 215 -
'' - ' 1 ,- ' ,) - V ; i ." > • 1,T:< , > •
or long-term effects, aspects which have not been dis-
cussed here.
In water treatment processes chlorine is widely used;
howeve , chlorine dioxide has gained popularity especially
when tcs e and odour problems occur upon conventional
chlorinat.-i.on. Chlorine and chlorine dioxide have also •
been extensively used in pulp bleaching processes and
much of the experience gained from these processes may.
be of great value for water chemists. Bleaching of pulp
to eliminate miscoloring substances, e.g. lignin, are
to some extent similar to oxidation of humic substances
in water with different chlorine species. However, there
is a pronounced difference in the concentrations of •
organic materials in the two processes.
Recent studies on by-products formed upon pulp bleach-.
ing with chlorine species have revealed great numbers
of chlorinated organic compounds of both polar and
nonpolar character in effluent water. The total amount
of chlorinated compounds exceeds 1 ppm, expressed as
chlorine in both polar and nonpolar solvent extracts of
the effluent water (3). Only about a tenth of the total
amount of lipophilic chlorine has been characterized,
e.g. chloro-cymenes, chlorinated terpenes etc. (3) (4).
The specific non—volatile compounds which cannot be
separated gaschromatographically are still not•very
well studied. This deficiency may be partly overcome .by-
use of chlorine detectors in liquid chromatography. The'
waste effluents from bleaching exhibit mutagenicity
with the Ames test (5).Through the polar character.of
chlorophenols (from lignin'breakdown)/ these substances
have been found accumulated 'in fishes caught near the
bleaching waste outlet (6).
Chloroform concentrations are about a hundred times
higher in chlorine bleaching effluents,than in drinking
water (7). Hypothetically,' some.of'the other chlorina-
-------
- 216 -
ted compounds found in pulp bleaching effluents may
also be present in drinking water at a two-or three-
order of magnitude lower concentration. If they are
present in drinking water,do they have a long-term
effect?
(1) KUlIN, W., SONTHEIMER, H., STIEGLITZ, L. , MAIER, D.,
KUR8, R.
Use of Ozone and Chlorine in Water Utilities in
the Federal Republic of Germany
J. AWWA TO (1978), 326-331
(2) LUNBE, G. , GET11ER, J., JOSEFSSON, B.
The Sum of Chlorinated and of Brominated Non-Polar
Hydrocarbons in Water
Bull. Environ. Contain. ' Toxicol. 13 (1975), 656-661
(3) EKLUND, G., JOSEFSSON, B., BjGRSETH, A.
Determination of Chlorinated and Brominated
Lipophilic Compounds in Spent Bleach Liquors from
a Sulphite Pulp Mill. Glass Capillary Column Gas-
Chromatography - Mass Spectrometry - Computer
Analysis and Identification
J. Chroraatogr. 1 SO- (1978), 161-168
(4) LINDSTR5M, K. , NORDIN/J.
Identification of Some-neutral Chlorinated Organic
Compounds in Spent Bleach Liquors
Sv. Papperstidn. 81 (1978)', 55
(5) ANDER, P., ERIKSSON, K-E. , KOL7iR, M-C. ,
KRIKGSTAD, K. , RANNUG, U.,- RAMEL, C.
Studies on the Mutagenic Properties' of Bleaching
Effluents '
Sv. Papperstidn. 80 .(1977)., . 454 •
(6) LANDNER, L. , LINDSTR6M,' K. ,'KARLSSON, M. ,
NORDIN, J., SOREMSEN, L..
Bioaccumulation in Fish of Chlorinated Phenols
from Kraft Pulp Mill Bleclchery Effluent
Bull. Environ, "contain. Toxicol. J_8 (1977), 663-673
(7) EKLUND, G. , JOSEFSSON', B. "/ ROOS, C.
Determination of Volatile Halogenated Hydrocarbons
in Tap Water, Seawater and Industrial Effluents by
Glass Capillary Gas "Chromato'graphy and Electron
Capture Detection
J. High Resol.Chrom. J[ (1978), 34-4O
-------
- 217 -
OCCURRENCE OF VOLATILE ORGANOHALOGEN COMPOUNDS IN THE OPER-
ATION OF WATERWORKS WITH VARIOUS TYPES OF WATER AND AMOUNTS
OF CHLORINE • •.
S. Hermann
Since the occurrence of low-boiling organic halogen compounds
in drinking water became known their formation in the water-
works as a function of the operating conditions has been of
particular interest. Detailed investigations on this problem
were also started at. the municipal works in Wiesbaden. The
following contains a brief report on the most important
results. • •
Use of chlorine' in the Wiesbaden water supply
In Wiesbaden chlorine is added at several points:
1) In Wl-Schierstein Rhine water is treated by flocculation,
filtration, and activated carbon filtration and then by
seepage for the purpose of artificial ground water enrichment.
The first measure is to chlorinate the raw water from the
Rhine for the purpose of oxidizing ammonia and for disinfection.
The amount of chlorine added : is'such that after a. few minutes
0.5 mg/1 of free chlorine is still detectable.
2) The drinking water obtained from the artificially en-
riched ground water - containing a considerable proportion of
bank filtrate from the Main' — after a further treatment by
aeration and slow sand filtration, is chlorinated to maintain
hygienic safety to the extent that it leaves the works with
0.4 to 0.5 mg/1 of free chlorine. PAC filtration will be added
to the treatment. -. '
3) A ground water charged with a small amount of anthropo-
genic material from the.Hessisches Ried, transported over
50 km to the supply area of Wiesbaden, must be chlorinated
several times on the way so that bacteriological safety is
-------
- 218 -
ensured for the various smaller consumer communities situated
along the transport route, and on the other hand so that the
limiting chlorine level is not exceeded at any point.
Halomethane formation in the chlorination of raw water and
drinking water
It is worthwhile to compare the action of raw water chlori-
ration and drinking water chlorination in Schierstein as
regards the formation of organic chlorine compounds, and
also to compare the two drinking waters that come from very
different sources.
The three waters were studied gas-chromatographically for at
least 6 months for the content of organohalogen compounds
(after enrichment by pentane extraction and using an electron-
capture detector) and the low-boiling fractions were determined
quantitatively. In addition to this, the DOC and the UV
absorption at 254 run were measured as sum parameters for the
organic loading. The mean values from about 20 samples for
halomethanes, the sum parameters, and the amounts of chlorine
added are listed in Table 1.
As regards the concentrations of halomethanes, it is note-
worthy that these may be affected by an error resulting from
the unreliability of the determination of the partition co-
efficients. With an enrichment factor of 200, a recovery
rate of 10% for chloroform and 33% for the other halomethanes
was calculated.
The interpretation of the results shall thus be restricted
to the relationship of the individual values to one another.
Some interesting statements can then be made:
-------
Table 1 Mean values for halomethane formation in Wiesbaden waters
Halomethane concentration
in pg/1
Chloroform
Carbon tetrachloride
Bromodichloromethane
Dibromochloromethane
Bromoform
Chlorine addition, g/m^
DOC, .g/m3
UV absorption at 254 nm, m
Rhine water
Raw water
not
chlorinated
1.0
0.3
0.06
not
detected
rare
chlor-
inated
3.0
1 .2
. 4.5
1.2
n. detect.
...1.0
-5.5
4.5
9.8
Drinking water I
not
chlorinated
2.5
0.15
1 .2
0.15
not
detected
chlor-
inated
2.5
0.15
5.0
10. 0
5.0
0.75
1 .8
3.3
Ground water from '
Hessisches Ried .
Drinking water II -
Tl
chlorinated
4.0
0.1
4.5
0.3 -
not detected, , .0. 1
1.4
2.0
3.6
I
to
-------
22O -
1. The haloraethane formation during the chlorination of raw
water from the Rhine is only of the same order of mag-
nitude as that during the chlorination of drinking
water. This is surprising, because the organic loading
and also the chlorine consumption in the raw water is
much greater than those in drinking water. It must
therefore be assumed that Rhine water contains relatively
few halomethane precursors.
2. Each compound shows different behaviour in different
waters,
3. The relatively sharp increase in the bromine compounds
is striking in all three waters.
4. Comparison of the two drinking waters clearly shows
that even in the safety chlorination of a"natural,
organically weakly loaded ground water halomethanes are
formed in low concentrations. In respect of the chlorine
compounds listed the two drinking waters show approxi-
mately the same loading. The formation of dibromo-
chloromethane and bromoform in drinking water I can be
explained by the higher bromide contents in the enriched
Schierstein ground water.
Formation of organohalogen compounds in chlorination of raw
water with variation of the amounts of chlorine added.
The effect of various chlorine additions and of various
excesses of chlorine in the chlorination of raw water was
investigated in a series of operational trials at the Rhine
water treatment works lasting several weeks. The chlorine
was added directly before the addition of the flocculation
agent. The effect of the chlorination was studied after
flocculation and filtration and a mean action time of 8 h.
-------
Table 2 Quality parameters in the four chlorination trials
(mean values)
Experimental
period
12.i2.77 to
11 .01.78
17.01.78 to
18.02.78
21.02.78 to
11 .03.78
21.03.78 to
25.04.78
Mean chlorine
addition
g/m
8.0
6,5
4,0
0
Chlorine concen-
tration in f loecu-
lation inflow _
q/mJ
1.1
0.5
0,1
0
Water
temp.
°c
4.4
3.8
5,5
9.6
pH
6.97
7.11
7.20
7.44
NH4+
rag/1
0.62
0.76
O.45
0.11
DOC
mg/1
4.9
5.0
4.8
4.2
CSB
KMn04
rag/1
2U7
21.1
20.3
17.3
OT254
m"1
10.7
10.4
10.1
9.1
to
to
-------
- 222 —
In Table 2 are listed the quality parameters, of the raw
waters as mean values for the experimental period for the
four different additions of chlorine in the range between 0
and 8 g/m . The lower concentration of free chlorine in the
flocculation inflow shows that the bulk of the chlorine is
consumed after only a few minutes. Not until a dose of
8 g/m can appreciable residual concentrations of free
chlorine be measured, which act further during the subsequent
treatment.
The gas-chromatographic measurements were performed at three
temperatures, namely at 60°, 180°, and 230°C. Fig. 1 shows
O.Sn
6 8
2468
y ,
Chlorine addition, q/ni —
CHCIj
g 1.0'
o
Jos-
B
O
c.
6 n\
r
/ C2^
A -/
HQ3
10
C2Ci4
10
CHCl2Br
CHClBr,
Mean values — Maximum values
8 10
Chlorine addition, g/ir,J
Mean value* --- Maximum values
Fig. 1 Formation of low-boiling organohalogen compounds
in the chlorinat.ion of raw water in dependence on
the amount of chlorine added.
-------
- 223
the concentration:increase for'6 different readily volatile
organohalogen compounds with increasing doses of chlorine.
In addition to the mean values, the maximum concentrations
found are also indicated, since they exhibit trends still
more clearly.
The concentration data refer here to values measured in
pentane, recalculated to the amount of water used but not yet
taking into consideration the partition coefficients for the
pentane-water system.
The different courses of the curves show clearly that the
formation of the individual compounds is strongly dependent on
the amount of chlorine: while some compounds are only formed
with larger chlorine amounts, or when the chlorine is in
excess (e.g. trichloroethylene), in the majority of compounds
elevated concentrations can be detected after only small
chlorine additions. The strongest dependence is given for
bromodichloromethane. For the remaining compounds appreciable
concentration increases are only measured when the chlorine
addition exceeds 6 g/m , i.e. in
free chlorine is already present.
addition exceeds 6 g/m , i.e. in the range where an excess of
Most of the substances formed during chlorination are already
present in low concentrations in Rhine water. Only very few
new compounds are formed, and these only at high doses. These
are low-boiling substances. In addition to dichlorobromo-
methane,. seven new compounds were detected at a column temper-
ature of 60°C; at 180°C only two were found in low concen-
trations.
As regards the higher-boiling substances (column temperature
180°C), the concentrations of the first four were increased by
the addition of chlorine - these were probably dichlorobenzene,
hexachloroethane, trichlorobenzene, and hexachlorobutadiene.
-------
- 224'"-
The concentration of the higher-boiling compounds is some-
times decreased by the flocculation (by up to 50%). Even
with high chlorine additions an increase was never observed.
An important factor for the evaluation of the effects of the
concentration increase of organic halogen compounds is the
ease of their removal during the subsequent activated carbon
filtration.
This was also studied, and in addition longer operational
experience is already available in this case. The types of
carbon used (F 300 and P 400) are very effective as regards
the elimination of the DOC and the removal of high-boiling
nonpolar chlorine compounds. The highly-volatile halogen
compounds formed mainly by the chlorination, however, are
only partially.adsorbed even by fresh carbon (e.g. chloroform
and carbon tetrachloride), or pass through the filter already
after a loading of only 10 1/g carbon (e.g. bromodichloro-
methane and dibromochloromethane). With increasing loading
of the carbon these compounds can be found in the filtrate
long after they can no longer be detected in the raw water.
The high-boiling compounds (starting with bromoform) occur
only sporadically in the filtrate, in low concentrations.
Summary
The operational trials on the chlorination of raw water with
various amounts of chlorine, performed on moderately polluted
Rhine water (with average outflow), show clearly as regards
the gas-chromatographic substances that low-boiling organo-
halogen compounds are formed first. The dependence on the
chlorine dose is very different for different compounds.
With low-dosage chlorination of the raw water with doses
o
below 6 g/m - as in the Rhine water treatment works at
Schierstein — the concentrations of readily volatile organo-
halogen compounds formed are kept within permissible limits.
-------
- 225 -
They are approximately in the same range as in the chlorin-
ation of drinking water.
If up to now chloroform formed the centre of interest, in the
f
case of the chlorination of bromide-containing waters the
bromine-containing halogen compounds must receive'more
attention than previously, because these compounds too are
insufficiently eliminated by the activated carbon filtration
and form in appreciable concentrations even under the con-
ditions of low-dosage chlorination of drinking water. In
this connection the bromide content of the- water also gains
in importance.
On the example of a marsh water (Hessisches Ried) it was
demonstrated that the safety chlorination necessary when
water is transported over a long distance - requires relatively
high doses of chlorine, and so can give rise to appreciable
halomethane concentrations.
-------
- 226 -
THM FORMATION IN TWO DIFFERENT WATER TREATMENT SYSTEMS
AT ROTTERDAM
J.J. Rook
1) Rotterdam Waterworks owns two treatment plants, i.e.
the Berenplaat Plant, dating from 1966, with a production
capacity of 12.0OO m /h, and the new Kralingen Plant, ca-
pacity 5.OOO m /h, which started production in July 1977.
Both plants treat the same type of raw water, which is
stored river Maas water. In the Berenplaat Plant conven-
tional chemical treatment is applied; breakpoint chlori-
nation, dosage of PAC, coagulation with Fe (III), sedi-
mentation, rapid filtration, post chlorination.
The new Kralingen,Plant consists of primary coagulation,
floe removal by means of lamella separators, ozonation,
dual media filtration and filtration through granular
activated carbon, post chlorination.
The raw water has been stabilized by self-purification.
Algae control measures have proved to be very effective,
the average chlorophyll content being 5 ppb, with inciden-
tal peaks of 4O ppb. The TOG value averages 5,5 ppm, mainly
caused by natural yellow acids which are identical with
fulvic acids. The raw water is transported through 28 -
3O km pipelines. During the summer period of 1977 the raw
water was pre-chlorinated (dosage circa 5 ppm) before
transport in order to keep pipe walls clean from bacterial
growth. During the winter season no chlorination was app-
lied. Transportation to Berenplaat takes 6 h, to Kralingen
15 h. With the given amount of fulvic acids the pre-chlori-
nation resulted in THM formation:
-------
- 227-
Plant .'" ' -V Contact time Temp.
Berenplaat 6 h
Kralingen 15 h
12-22°C
12-22°C
TTHM at
plant intake
68 ppb
1O3 ppb
During the winter season of 1977-78 pre-chlorination was
interrupted. Consequently, no~ THM formation occurred.
Results of conventional treatment
Table I summarizes the THM contents in the finished water of
the conventional treatment at Berenplaat, both with and without
pre-chlorination.
TABLE I THM in Berenplaat Plant
Conventional: Breakpoint-PAC-coag.-filtr.
o
o
CM
CM
1
CM
t —
O
o
CM
3 —
I
Raw water
pre -chlorinated
(5 ppm 7 h)
Finished water
Raw water
not chlorinated
Finished water
(4 ppm breakpt .chlor . )
CHC1-
3
30 '
40
18
CHCl_Br
2
22
28
O
13
CHClBr,,
2
14
14.5
0
6
CHBr_
3
1 .5
1 .5
0
1
Total
68
84
38
This data shows that post-chlorination (O.8 ppm) caused an
increase of TTHM from 68 ppb to 84 ppb during the summer
season.
The formation potential of the raw water was 16O ppb after
48 h at 2O °C.
-------
- 228 -
TKBUB II THM in Kralingen Plant'
Coag. 2.5 ppm ozone-filtr. (GAC)
o
o
CM
t
i
O
o
T
Raw water
pre-chlorinated
(5 ppm 15 h)
after Floe Separation
after Ozonation
Finished water
(without GAC)
(O.8 ppm 6 h post-chlorin.)
Raw water
not chlorinated
after Coag.+ Ozone
eluted from GAC
(7 -1O months in use)
Finished water
(post-chlorinated)
CHC13
50
32
32
49
' 1
I
1
5 ,
6.5
CHCl2Br
32.5
22
-23'
35 .
'
0
3
6.5
CHClBr2
18
13
12
16.5
O
1 .5
9.5
CHBr3
2.5
1 .5
2
1 .5
0
0
6.5
Total
103
68
69
102
1
9.5
29
TABLE III THM in Kralingen Plant
Influence GAC
after GAC (first month)
after GAC post-chlorin.
(O.8 ppm 6 h)
after GAC (2-5 months
in use)
after GAC post-chlorin .
CHC13
1 .2
3.2
17
18
CHCl2Br
0.8
3.7
11
12.5
CHClBr2
O
5.4
4
13.5
CHBr_
0
5.7
O
8.5
Total
2
18
32
52
-------
- 229 -,
In the winter period when only breakpoint chlorination
with a chlorine dosage of 4 ppm at a contact time of
3O min. was applied, the finished water contained 38 ppb
TTHM.
It is interesting to compare this figure with the results
of the ozone-GAC treatment in the Kralingen plant.
Results of new treatment plant
In the Kralingen plant the -reservoir water is coagulated
with iron, followed by ozonation (3 ppm), dual media fil-
tration and final filtration through activated carbon
(GAG, contact time 12.5 min.).
The plant was started in July 1977, when the raw water was
pre-chlorinated. During 15 ;h transportation TTHM went up
to 103 ppb. In the first month of operation the carbon
filters could not be used, 'because they were in the pro-
cess of being filled.
The effects of this treatment coagulation and ozonation
without carbon are shown in Table II. It is seen that the
THM once formed by pre-chlorination is only partly removed
by the coagulation and ozonation steps. An interesting ob-
servation is that post chlorination of O,8 ppm gave rise
to renewed THM production, i.e. an increase from 69 to
1O2 ppb. The TOC value after .coagulation ozonation was
2.8-3 ppm. '
The advantageous effect of the adsorption of both pre-
cursor and THM onto fresh carbon may be seen from the
data in Table III, upper lines. The TTHM in the raw water
still amounted to 1OO - 1O3 ppb. Adsorption reduced this
to 2 ppb. The observation that post-chlorination (mea-
sured after 6 h) produced 18 ppb TTHM, as compared to
33 ppb without GAG, indicates that precursors remaining
-------
- -23O - '
after ozonation were partly reduce.d-. Ouringr this- first
month of operation the carbon filter treatment., lowered
TOG from 3 ppm to O.9 ppm. In the next three months, how-
ever, the TOC of the effluent gradually rose to 1.5 ppm.
Dufring this period THM breakthrough started to become
significant for CHC13 (Table III, lower case). The more
brominated THM were withheld"longer._
Precursor removal was still performed at a slightly di-
minished level. Post-chlorination now caused a 2O ppb
increase of TTHM (32 to 52 ppb).
During the following winter season the pre-chlorination
was interrupted and consequently THM formation did not
occur. The results of the ozone - GAC - post-chlorination
treatment are given in the lower case of Table II.
Two conclusions may be drawn. First, the TKM orginally
adsorbed was slowly eluted, resulting in average values
over the 4 winter months of 9.5 ppb TTHM. Secondly, pre-
cursor removal appears to remain at about the same level .
as indicated by the 2O ppb increase caused by post chlo-
rination.
However, the effect of lower temperature on the overall
reaction rates have to be taken into account.
Table IV summarizes the observations reported here.
The data clearly show that pre-chlorination of raw water
is deleterious to any further conventional or advanced
treatment. Starting with non-chlorinated raw water j_n the
conventional treatment resulted in 38 ppb TTHM, as compared
to 2O ppb TTHM increase during 6 h post chlorination. In
this case the total effect was negatively influenced by
the elution of 9 ppb TTHM resulting from the preceding
period during which the raw water had been chlorinated.
-------
_ 23-t'-
A final-observation'is'that' o z One ' treatment resulted in a!''
shift towards brominated THM in post chlorination, as shown
by the data of the finished waters of the conventional
treatment (Table I, last line) vs. new treatment Tables II
and III.
TABLE IV
TTHM
PLANT I
Conventional
ppb
Raw water ,
not chlorinated
Finished water 38
Raw water go
prechlorinated
Finished water 84
TTHM
PLANT II
Ozone GAC
Raw water
not chlorinated
Coag . -Qzone-GAC
(6-10 months)
Finished water
Raw water
prechlorinated
Coagulation-Ozone
Coag. -Ozone-GAC
Finished water
(O.8ppm post chlorinated)
ppb
1
9
29
103
69
32
52
-------
- 2 3 2 • - -
CHEMICAL CHANGES AND -REACTION .PRODUCTS.,,IN ,-THE .OZONI^TIQN
OP ORGANIC WATER CONSTITUENTS - -
E. Gilbert
The growing interest in the use of ozone for the treatment
of drinking water and waste water has led not only to
studies on the process techniques for the ozone input (1-10),
the mass transfer from gas- .to water phase (11-13) , and the
behaviour of ozone in water (14-203, but also to a large
number of publications on the chemical effect involved. The
following is- a survey of the reactions of ozone with the
organic constituents of water and of the oxidation products
produced in this way.
Disinfectant effect
Until the nineteen-fifties ozone was used mainly for its
disinfectant effect; it had been used in the treatment of
drinking water as early as the turn of .the century (21), and
therefore the majority of publications have concentrated on
this partial aspect. In addition to more recent results on
its disinfectant action and on viral inactivation by ozone
(22-27), attention is particularly drawn to the proceedings
of the 1976 Chicago conference "Forum on Ozone Disinfection"
(28).
Since the analysis methods are extremely complicated, very
little can be said about the mechanism of disinfection or
the inactivation of viruses, which can often take place in
a matter of seconds with a dose of 1 .mg 0^/1. The first
insight into the chemistry of the disinfection processes was
provided by trials on amino acids, the building blocks 'of
proteins, in pure solutions. Mudd et al.(29), Menzel (3O), and
-------
- 233 -
Shuva'l '•'(3'l)'-r-re'po£ated -that the'; -unsaturated ' ami'no acids such' '-'4 •'
as tryptophan, tyrosine, histidine, and the sulphur-containing
amino acids such as methionine, gluthathione, cystine, and
cysteine are readily attacked by ozone. Shuval (31) found
that viruses react with ozone at much the same rate as the
above-mentioned amino acids (Pig. 1, ozonization of trypto-
phan) . This is an indication that the inactivation of
viruses may be due to an attack of ozone on these amino
acids and that as a result of oxidation the acids lose
their structural characteristics important for biological
activity. In the case of the unsaturated amino acids, such
as tryptophan, the double bond is attacked and the molecular
structure is destroyed (Fig. 2). In the sulphur-containing
amino acids the sulphur is oxidized to sulphoxide, sulphone
or to sulphonic acid, whereby the acids can no longer act
as oxygen carriers in redox processes. It is not known
whether the oxidation products themselves have any disinfec-
tant action, since their identification has been only partial
and qualitative.
O
E
3,
c
a
a.
o
l_
0,5
03=1,8-2.0 /u mole/sample
Sample volumes = 5,0 ml
Temperoture =0 "C
Fig. 1
Reaction of tryptophan with
ozone; dependence on the pH
15 30 60 90 120
Time [ S ]
-------
Ami no acids
Tryptophon
COOH
~*C — CrurCH
* *
H
-•234 -
Oxidation products
Kynurenine
9 ,COOH
-i-C-CH-rCH
OJ 2 N
N-Formylkynurenine
9 DOOM
~ C* H -^"C H
NH
Methionine
CH3- S - CH2-CH2-CHlNH2)COOH
NH
Methionine sulphoxide
CHjS-CHjCHj-CHlNHjlCOOH
6
Cystine
S-CH2-CHiNH2)COOH
Cysteic acid
H03S-CH2-CH(NH2)COOH
Fig. 2 Ozonization of amino acids; oxidation products
Effect, on the sum parameters
Besides the disinfectant effect of ozone which, as already
explained, may be associated with the ready oxidizability of
amino acids, its other beneficial effects have been utilized
for a considerable time. These are .an improvement of the
water's odour and taste (32-36), removal of any colour or a •
reduction of UV-extinction (32, 37-45), as well as a
reduction in the molecular weight of high-molecular sub-
stances (46-50). Fig. 3 shows as an example the effect of
ozonizing humic acid from Hohlohsee (Northern Black Forest)
on the size of its molecule. After a dose of only
1.5 mg 03/mg C, half of the carbon is present in the form
of molecules with molecular weights below 1000 (50).
-------
- 235 -
The DOC and the COD are also decreased by oxidation processes,
It is found that the DOC decreases much more slowly than the
COD and that with - longer ozonization times both approach
asymptotically to a final value. To illustrate this point
the results obtained in the ozonization of 2-nitro-p-cresol
in.water are shown in Fig. 4. An unbuffered 1 mmol solution
at pH 5 was used. The greatest decrease in the two par-
ameters occurs after degradation of the cresol, i.e. the
products formed are present in a higher oxidation state
and - at least in the acid range - react further with
ozone only very slowly.
70
60
— 50
o
f 40
o
g 30
20
10
I
%
I
1
I
M¥
MW
MW
MW
> 30 COO
000- 40 OOO
< 4 000
0 0,31 0,6 1,47 2.1
Ozone coasumpiion,mg/mg starting DOC
Fig. 3 Ozonization of huruic acid (Northern Black Forest)
without the molecular weight fraction > 1OOO
-------
- 236 - ',
60 120 180
Ozonization time (min) dooe 10 mg Oj/mirvl
iq. 4 Ozonization of 2-nitro-p-c.resol (c = 1 mmole/1)
TOG and COD decrease in dependence on the
ozonization time
In a further example, which is representative of many (38,
39, 43-45, 48, 51-69, 120, 134) a paper pulp waste water
was studied (51). As Fig. 5 shows, with a limited ozonization
time a total degradation to C0~ and H-O is not achieved.
The extent of the COD reduction naturally varies from one
substance class to another, and in the case of a given waste
water reacts according to its composition. After the COD
has been decreased by 50-70% further ozonization brings
about only a negligible reduction of this parameter.
According to the present data, a reduction of the COD value
by 1 kg is attained by the consumption of 1-4 kg of ozone
(Table 1), and in extreme cases even 6-10 kg of ozone'may
be necessary.
-------
- 237 - •'
1100
^900
OT
E
CD
CO
o
* 700
Vol. of effluent used' 10 I
Ozone dose 1,63 g/l
246
Ozonization time [h]
Fig. 5
Ozonization of paper mill
effluent; COD decrease in
dependence on Ozonization
time
10
TABLE 1 COD degradation and ozone consumption
Solution used
Ca ligninsulphonate
Pacer mill effluent
Effluent
Industrial effluent
Effluent from a sauerkraut factory
Industrial effluent
Paper mill effluent
Discharge from clarification plant
Discharge from clarification plant
Discharge from clarification plant
Discharge from clarification plant
Humic acid solution
Naphtbol
Phenol
Chlorophenols
Haphthalene-2.7-disulphor.ic acid
p— Toluenesulp!:cnic acid
COD COD
mg 02/1
1130 — » 890
4950 -j. 2500
225 — > 198
3340 — *• 1930
3340 -*. 314
2140 -» 1620
910 -» 150
14000 -» 3000
156 -* 94
1480 -* 1150 -
59 -> 49
59 -> 35
58 -» 50
48 ~> 24
43 -* 42
56 — » 32
80 ~> 16
20— > 10
240 — » 60
26 —» 19
70 — > 24
216 -* 106
265 — > 60
250 — > 60
Dzone consumption
A COD
1
1.8
1.5
0,8
1.8
2.5
5.0
2.C
0.7
1.4
1.4
2.0
1.5
2.5
5.6
1.7
2.5
2.5
2.7
ir
1.6
1.5 - 2.3
2.5
3.7
Ref.
51
38
57
62
66
44
48
60
58
63
52
61
72
54
—
-------
- 238,--
With small ozone doses an increase of the COD has actually
been observed in some cases (43, 61, 64, 68, 69). In the
first phase of the ozonization some substances are made
accessible to the COD analysis by the oxidation, thus pro-
ducing an increase in COD which then decreases again as the
ozonization continues.
The biological degradability is often improved by the attack
of ozone, as has been reported in many recent papers. Data
on the rise of the BOD5/CCD ratio after ozone treatment are
found particularly in (38, 44, 45, 54, 65, 70-72).'
The change in the molecules after ozonization also in-
fluences the chlorination processes in the treatment
of drinking water. Preliminary ozonization reduces the
formation of chloroform during the subsequent chlori-
nation (73-77). Maier (75) observed that as the chloro-
form decreased an equivalent increase of bromodichloro-
methane and dibromochloromethane took place in ozonized
Lake Constance water. Lawrence (77) was able to establish
that in the ozonization of ligninsulphonic acid and aspartic
acid the concentration of "chloroform precursors" increased
at the start of the reaction and only fell again as the
oxidation progressed. Rook (73) found in addition that the
ozonization effect contributing to the reduction of chloro-
form formation becomes the smaller, the longer is the
interval between ozonization and chlorination.
In the large-scale plant at Lengg (Switzerland)(78) increase
in the concentrations of two halogen compounds, dichloro-
butanone and iododichloromethane, could be observed after
ozonization. Block and Buydens found an increase of the
chlorine demand after ozonization of raw water (79, 80).
Maier (75) showed that after ozonization of Lake Constance
water the chlorine consumption proceeded essentially more
-------
- 239 -
slowly than in non-ozonized-water. In addition, several
authors have described an increased repopulation with
bacteria due to ozonization (75, 78).
Finally, the ozonization of raw water or river bank filtrate
can lead to the formation of non-polar chlorine compounds
(81).
The effects mentioned depend of course on the type of the
raw water or on the composition of its constituents. The
latter, however, are so numerous that so far the complex
systems of raw water and waste water can in effect only be
characterized with the aid of sum parameters. Therefore,
the data concentrate on the action of ozone on such measur-
able parameters.
Ozonization of pure compounds
To explain and thereby also to control these phenomena, a
knowledge of the exact compositon of the aqueous solutions
would naturally be necessary. This is the first basic pre-
requisite for informative results. It also seems useful,
as in the case of viral inactivation, to resort first to
model substances and to study the ozone-oxidation on simple
systems. This approach has been described in recent years
by several working groups.
Quinoline
Naimie (82) ozonized 200 ml 'of a 0.25 - 0.5 mM quinoline
solution (pH 6.9), which after 60 min (30 mg O3/min) con-
tained less than 10% of the initial quinoline content.
Cyclohexanol, cyclohexene
Cyclohexanol and cyclohex-l-ene-3-ol (2 g/1) in the presence
of 10.5 g of Ca(OH)2 were eliminated with 4 g/1 or 1.5 g/1
of ozone. Malonic acid, glutaric acid, adipic acid, and
-------
- 240
succinic acid were identified quailita'tiveiy 'amoncj the oxid-
ation products, but amounted only to an estimated 5% of the
total amount of the oxidate (83).
Urea
Eiehelsdo'rfer (84) was able to determine quantitatively the
oxidation products of the ozonization of urea, CC>2 and nitric
acid.
Malonic acid
The ozonization of malonic acid (85, 54) produced tartronic
acid, mesoxalic acid, oxalic acid, £®2' an<^ H2°?' A mass
balance was set up. In this process 192 ppm of ozone are
consumed in 100% elimination of 104 ppm of malonic acid.
Maleic acid
Black (86) ozonized 40 g of maleic acid in 100 ml of water
with the stoichiometric quantity of ozone for the prepar-
ation of glyoxylic acid.
Caffeine
Shapiro (87) found in the ozonization of caffeine (660 ppm)
that 4.2 moles of ozone were necessary to eliminate 1 mole of
caffeine. Dimethylparabanic acid was identified in addition
to several other oxidation products.
Dichlorobenzene
The ozonization of dichlorobenzene in pure aqueous solutions
(88) proceeded most efficiently at pH 8.4. 60 mg of ozone
are needed to oxidize 30 mg of the substance/1.
Glucose
Walter (89) described the reduction in the COD after ozon-
ization of 6 litres of 0.1% glucose solution. After a maxi-
mal dose of 1.6 g of ozone the COD fell by 10-15%.
-------
- 241 -
Ethylenedi.amine,t,etraacetic acid ,
Krause (90) studied the decomposition of ethylenediamine-
tetraacetic acid chelates with ozone. To release 90% of the
metals (Ca, Zn, Ni) from the chelate complexes in 1 litre
of 0.01M solution an ozone amount of 1.7 - 3.4 g 0-,/l was
used. The reaction proceeds from pH 7 in the basic region
twice as rapidly as in the acidic region.
Aliphatic alcohols
The ozonization of methanol (91, 91a, 92), ethanol, butanol,
and octanol (92) proceeds via the aldehydes to carboxylic
acids. 88% of isopropane is transformed into acetone (93).
In the ozonization of ethanol, Gilbert (94) found in addition
H2O2, short-lived peroxides, and formic acid (Table 2).
Alkylbenzenesulphonic acids
Buescher and Ryckman (95) investigated the behaviour of
tert.alkylbenzenesulphonates (ABS) with respect to ozone.
12 mg of ozone were consumed to oxidize 4.5•mg of ABS. Evans
(96) found that the removal of 15.6 mg of ABS in 1 litre of
discharge from a biological clarification plant requires an
ozone dose of 75 mg/1, and that the oxidation products are
utilized better biologically. Kandzas (97) reports that at
pH 11 90 mg of ABS are eliminated by 190 mg of O-j. Small
quantities of formaldehyde and formic acid were found as
oxidation products. Joy (98) studied the ozonization of
nonyl- and decylbenzenesulphonates. 720 mg of CU were
needed at pH 6 to degrade 160 mg of sulphonate and only
240 mg at pH 12. 70% of the sulphonic acid was converted
into sulphate. Formic and oxalic acids were determined
quantitatively and nonylglyoxal qualitatively as organic
oxidation products (Table 3).
-------
- 242 -
TABLE 2 Ozonization of alcohols
Compounds
Methanol
Ethanol
Butanol
Ethanol
Isopropanol
:
Concentration
140 mg/150vm1
210 mg/150 ml
330 mg/150 ml
92 mg/1
192 mg/1
Ozone
consumption
90 mg
360 mg
380 mg
211 mg
amount added
2.6 g
Degrad-
ation
29 %
71 %
64 %
90 %
100 %
pH
5.5 - 5.5
5,4 - 3,8
6.2 - 3.7
6.3 - 3.8
7
Ref.
92
94
93
TABLE 3 Ozonization of alkylbenzenesulphonates
Compounds
Alkylbenzenesulphonate
11
H
Nonyl- and decyl-
benzenesulphonate
Concen-
tration
4.5 mg
15.6 mg
90 mg
160 mg
160 mg
Ozone
consumption
12 mg
75 mg
190 mg.
720 mg
240 mg
Degra-
dation
100 %
100 %
100 %
100 %
100 %
PH
11
6
12
Ref.
82
83
84
85
Polycyclic aromatics
Reichert (99) and Cornelia (100)studied the degradation of
3,4-benzpyrene. 0.7 ppm of ozone are required to remove
1 ppb of benzpyrene from water.
Chlorine-containing pesticides
The Ozonization of pesticides has been studied by several
authors (Table 4). Up to now the toxic dieldrin has been
identified qualitatively as the oxidation product from
aldrin and heptachloroepoxide from heptachlor (101).
Brower (105) established in the fish test that the oxidation
products of aldrin are less, toxic.
-------
243 —
TABLE. ,_4 Ozonization of chlorine-containing pesticides
Pesticide
Lindane
Endosulf&n I
Endosulfan II
Dieldrin
Heptachl oreepaaft
DDT "
Aldrin
Heptachlor
Lindane
Lindane
Lindane
Aldrin
Aldrin
DDT
Concentration
2 ppm
ti
i»
H
II
M
11
"
40-100 ppb
10 ppb
50 ppb
20 ppb
0,005 %
suspension
7 ppp .
Ozone
dose
17 ppm
M
II
It
li
11
It
It
0.4-3 ppm
11 ppm
149 ppm
23.8 ppm
2.85 g
Ozone
con-
sumption
97 ppm
Degra-j
dation
0
0
12 %
26 %
26 %
78 %
86 %
iop %
0
10 f;
100 %.
100 %
100 %
100 %
pH
9.8/6,8
7.5-4,4
'7,9-4.5
Ref .
101
102
103
104
105
112
Phosphorus-containing pesticides
In contrast to the chlorine-containing pesticides, the
phosphorus-containing ones are more readily oxidized.
Richard and Laplanche (106-108) found that after an ozone
dose of 3 ppm 80 ppb of.parathion were oxidized?.toxic para-
oxon is formed among other products,'and is degraded only
after 5 ppm of ozone has been added. 2,4-Dinitrophenol,
picric acid, sulphuric acid were identified qualitatively
and phosphoric acid quantitatively as" the oxidation products
(Table 5).
Waste water constituents
For a better control of the ozonization of waste waters from
photographic processing laboratories, the individual consti-
tuents have also been studied (59,109), such as acetic acid,
glycine, diethylene glycol, benzyl alcohol, and various
colour developers. The concentrations amounted to 1 g/1 and
-------
- 244 -
5 Ozonization of phosphorus-containing pesticides
Pesticide
Malathion
Parathlon
Methyl pa ra th i on
Fenitrothion
Parathion
Fenthion
Concentration
100 ppb
S7. ppb
125 ppb
120 -ppb
2500 ppb
2450 ppb
Ozone
amount
5 ppm
"
it
11
149 ppm
149 ppm
Degra-
dation
88 %
• 83 %
m %
• 93 %
100 %
100 %
PH
8
9,8/6.5
9.8/7.2
Ref.
106-108
11
11
II
104
the ozone dose to 0.5 g O3/h. Ozonization was performed
for 8 h and the COD decrease was measured. Glycine and
acetic acid were the only substances tested that exhibited
no change.
Bauch and Burchard (110) studied a series of organic sub-
stances and determined only the degree of degradation. The
concentrations used were 1-2 g/1 for raethanol, ethanol,
glycerol, ethyl acetate, acetic acid, caprylic acid, sugar,
and glucose and 0.1 - 0.5 g/1 • for- hydrazine, phenol, o-
cresol, hydroquinone, o-salicylic acid, pyridine, benzine,
chloroform, benzene, and toluene.' Acetic acid, ethyl acetate,
caprylic acid, pyridine, and chloroform were shown to be
stable to ozone.
Caulfield (111) established by the Ames test that the muta-
genic action of various policyclic aromatic hydrocarbons,
aromatic amines, and some pesticides can be eliminated by
oxidative degradation with ozone. However, the primary
oxidation products of benzidine had stronger mutagenic
action (111,112), though on further Ozonization they were
degraded to inactive products.
Spanggord (112) studied the Ozonization of 30 organic com-
pounds in high concentrations to obtain large concentrations
-------
- 2 4-5 ' -
of intermediate 'products'.''" "The mutageriicity"of these was
then investigated in the Ames test (113).
The compounds in question were 2,4-dinitrotoluene, diphenyl-
hydrazine, acetic acid, aroclor 1254, glucose, urea, chol-
esterol, benzidine, glycine, cysteine, benzene, thymine,
caffeine, diethylamine, phenol, hydroquinone, glycerol, and
nitrilotriacetic acid in up to 1% solutions. Only the
ozonized solutions of phenol,' 1,1-diphenylhydrazine, nitrilo-
triacetic acid, benzidine, and ethanol were shown to be
mutagenic after ozonization. A selection of the compounds
used is listed in Table .6.
TABLE 6 Ozonization of organic substances
after Spanggo'rd (112) and Simmons (113)
Compounds
2 , 4-dinitrotoluene
Oleic acid
Acetic acid
Glycine
Ethanol
Benzidine dihydro-
chloride
Phenol
Concern
tration
• 80 -pprti
,99 ppm.
1263 ppm
11900. ppm
7759 ppm
4830 ppm
9140 ppm
Ozone
amount
1.7 g
2.55 g
37,2 g
9,8.g
55,8 g
2^25 g
6,97 g
Degra-
dation
55 %
100 %
3,5 %
22 %
74 %
64 %
41 %
Mutagen-
icity
before
03
—
--
—
'--
--
Mutagen—
icity
after Q3
--
--
--
--
+
+
+
Salicylic acid •
Mallevialle (52) established in the ozonization of salicylic
acid (100 - 200 mg/1) with degradation of the starting com-
pound an increase of the extinction at 420 nm, which, after
reaching a maximum fell to zero with the elimination of the
salicylic acid. Thin-layer chromatography revealed the
presence of other phenolic oxidation products not identified
more closely.
-------
- 246 -
Phenols
Hillis (61) investigated the degradation of various substi-
tuted phenols (30 mg/1) at pH 4-10. It was found that not
until pH 10 did the degradation proceed twice as rapidly as
at pH 4 and 7. To achieve a degradation below 0.1 mg/1 the
same ozonization time was necessary in all cases.
Dorc (114) found that during the ozonization of aminophenols,
nitrophenols, and halophenols and of dichlorophenoxyacetic
acid (25 mg/1) UV-active oxidation products were formed in
the first phase of the reaction. The extinctions of the
new absorption bands pass through a maximum and finally fall
to zero after 100% degradation of the starting compound.
Bauch (115) was able to identify qualitatively the following
oxidation products in the ozonization of cresol, chlorophenol,
and xylenol (3.8 g/200 ml)jglyoxylic acid, acetic acid,
propionic acid, maleic acid, glycolic acid, oxalic acid, CO-*
and IIC1. The degradation rate was independent of the pH.
-4
2,4,5-Trichlorophenoxyacetic acid (c=10 M) was oxidized by
ozone to chlor'ide, glycolic acid, oxalic acid, glyoxylic
acid, and dichloromaleic acid -(116) . The oxidation products
were determined quantitatively.
The unidentified oxidation products of nitrophenols and
phenol were shown in the mouse test to be non-toxic (117,
118).
The reactions of phenol with ozone have so far been studied
in the greatest detail. Several working groups have deter-
mined the decrease of the phenol in the course of ozonization
(119-122). The results on the pH-dependence reveal that the
oxidation takes place more rapidly at pH 10-11 and the ozone
consumption is only half of that in the acid region (119)
(Table 7). Eisenhauer (123-125), in the ozonization of
-------
- 247- --
TABLE 7 Ozonization of phenols
Compounds
Salicylic acid
Phenolsulphonic acid
Chlorophenols
Amino phenols
Nitrophenols
Cresol
Xylenol
Naphthol
Amino, nitro, and
halophenols
Cresol, chlorophenol ,
xylenol
Chlorophenols
Dichlorobenzene
Phenol
Phenol
Ozone consumption until the
- phenols fall below the detection
level
3 moles D.,/1 mole acid
4.9 moles O,/mole phenol
(mean value) .
5 moles 0-,/mole phenol
5.5 moles 0,/mole phenol
3-4.5 moles O-,/mole phenol (pH 5)
6 moles O.,/mole benzene
4 moles 0-,/mole phenol (pH 7)
2 moles O3/mole phenol (pH 12)
6 moles O,/mole phenol
4-6 moles O-,/mole phenol
Ref .
52
61
114
115
72
88
119
122
127
phenol solutions (50 mg/1), obtained practically the same
rate of degradation until pH 9. Only at pH 11 was the
degradation rate doubled. Pyrocatechol quantitatively and
o-quinone qualitatively were identified as the first oxi-
dation products.
Reissaus (35) identified oxalic acid qualitatively as one of
the oxidation products.. Casalinl (126) ozonized phenol
solutions with 500, 1400, and 5000 mg/1 in a 3.75-litre
container (ozone doses 23/26 mg O3/min) for 40 h at pH 3-11.
Pyrocatechol and hydroquinone were determined together
-------
- 24-8 '-''- '
quantitatively as'the oxidation .products and maleic acid,
formic acid, glyoxa'l, formaldehyde, and oxalic acid were
found to be present. Gould (127) studied phenol solutions
(90 mg/1, pH 7.7) ozonized for periods of up to 30 min
(ozone dose 72 mg O3/min). Pyrocatechol, hydroquinone,
glyoxylic acid, and oxalic acid were recorded quantitatively
as oxidation products during the o'zonization. On the basis
of the carbon balance these are the only reaction products.
On the other hand, Niki (128) found formic acid to be one of
the main products of the ozonization of phenol. In-this
case 57 mg of phenol in 100 ml of water were treated with
only 5.8 mg 03/min for 180 min. Muconic acid, muconaldehyde,
maleic aldehyde, glyoxylic acid, glyoxal, and oxalic acid
were identified quantitatively as further oxidation products,
making up 100% of the TOC.
Substituted aromatics and their oxidation products
The different results obtained by the various working groups
can be explained by their use of different reaction condi-
tions such as the initial concentration, dose of ozone,
ozonization time, pH, an'd reaction volumes or reaction
vessels. Furthermore, the different rate constants are
responsible for the formation and concentration distribution
of the oxidation products. Fig. 6 shows that the unsatu-
rated aliphatic carboxylic acids are degraded more rapidly
than the substituted aromatics. Therefore, after their
formation, they can be immediately oxidized further in
competitive reactions with the starting compounds.
For example, with low ozone doses and high initial concen-
trations the chances of being able to isolate unsaturated
aliphatic carboxylic acids such as fumaric or muconic as
the oxidation products are higher than in the case of large
ozone doses and low initial concentrations; this can be seen
on the example of the ozonization of phenol under various
conditions (126-128). For a better understanding of the
-------
-249 -,.,-
.fenzoic acid
Glyoxylic acid
Oxalic.arid aesoxalic acifls
\
Muconic acid
\ \Phenol
.Dichlorophenol *v
5 10 15 20 25 30
Oeonization time (min) , dose IQmgOj/mirvl
Fig. 6 Degradation of aliphatic and aromatic compounds in
the ozonization of -aqueous solutions, ozone dose
1O mg
c = concentration at ozonization time t
c = initial concentration 1 mmole/1
o
course of the reaction during ozonization it is necessary
to know the rate constants. Hoigne' has contributed much to
this problem (129,130). It appears that in the acid range
the ozone molecule is directly responsible for the oxidation
reactions, while in the basic range the decomposition product
of ozone formed by OH ions, the OH radical, initiates the
oxidation by addition and abstraction reactions.
In the light of the work presented, the ozonization of sub-
stituted aromatic substances can be represented as follows,
2-6 moles of ozone are necessary for the elimination of
1 mole of the aromatic initially present (Table 7). In the
initial phase of the reaction (up to about 50% degradation)
an increasing coloration (pink or yellow) of the solution
-------
250 -
is observed or a new UV extinction is measured. The colour
intensity and/or the UV extinction reaches a maximum on
elimination of the starting compound and then falls to zero
(52, 114, 116, 119, 128, 133, 134). These observations
indicate that the first step in the ozonization of substi-
r
tuted aromatic substances need not be the spontaneous ring
opening but that the primary oxidation products - it is not
known to what extent they are formed - have an aromatic
character or a six-membered ring structure. These primary
oxidation products could be, for example, hydroxylated or
quinoid compounds, such as have so far only been identified
in the ozonization of phenol (as quinone, hydroquinone, and
pyrocatechol) (123, 126-128) .
If ring cleavage then occurs by further attack of ozone or
OH radicals, muconic acid derivatives can be formed, which
are transformed via maleic or fumaric acid derivatives into
carbonyl and/or carboxyl compounds with one to three carbon
atoms.
The course of these subsequent reactions has been studied on
model compounds such as muconic, maleic, and fumaric acids
and their oxidation products (54, 131, 132). All the com-
pounds (Table 8) could be followed quantitatively, so that
reaction mechanisms may be proposed.
The oxidation products of the aliphatic model substances
were also found in the ozonization of substituted aromatics
and phenols (Fig. 7). This indicates that after ring
cleavage further oxidation takes place via muconic or
fumaric acid derivatives.
After ring cleavage the substituents are converted into their
mineralized form (54, 72). The detection of dichloromaleic
acid (Fig. 7) (116) shows that the hetero-groups do not
-------
- 251 -
TABLE 8 Ozonization of aliphatic carbbnyl and
carboxylic compounds.
Initial concentration 1 mmole/1; ozone dose
10 mg 03/min 1; ozonization time 2O-18O min; pH 3
Compounds
trans, trans-Muconic acid
Formic acid
Glyoxylic acid
Maleic acid
Fumaric acid
Glyoxal
Tartronic acid
Malonic acid
Dihydroxyfumaric acid
Oxaloacetic acid
Oxidation products
Fumaraldehyde, glyoxal, glyoxylic
acid, oxalic acid, formic acid,
C02, H202
C02, H20
Oxalic acid, CO_
Formic acid, glyoxylic acid,
oxalic acid, CO-?
Formic acid, glyoxylic acid,
oxalic acid, ' mesoxalic acid
aldehyde, mesoxalic acid CO2
Glyoxylic acid, oxalic acid,
C02
Mesoxalic acid
Tartronic acid, mosoxalic acid,
C02, H202
Oxalic acid, hydroxytar tar ic
acid, CO2
Formic acid, glyoxylic acid,
oxalic acid, mesoxalic acid,
co2
Ref.
132
131
54
11
It
II
always hydrolyse spontaneously after ring cleavage and are
not always converted completely into the inorganic form.
Depending on the substituent, many further aliphatic oxi-
dation products that contain hetero-groups are possible, but
so far they have not been identified. Even in the case of
p-toluenesulphonic acid (135), 2-nitro-p-cresol (135), and
4-chloro-o-cresol (54) so far only the oxidation products
given in Fig. 7 could sometimes be followed quantitatively
in the course of the ozonization (Table 9).
-------
- 252 -
61
HO
SQjH
CH3
, NH3
^.0
H-C'
H-C*°
ov
K'
o
"H
o
H'
»°
"OH
0
HO'
* 0
c-c' <
OH
0
:H3'C^
OH
H
HO-C-C
H
0
"OH
CH3-
o
c-c:° i
H
.
1 ,0
OH
D i 0 Q i
H' OH HO
°
OH
Cl & 0« OH
0 H H 0 « MHO 0. I I ,0 0 , I 0
*c-c-c-c *c-c-c-c *c-c-c-c* *c-c-c-c*
HO VH HO OH HO OH HO XOH
0 MKMM^O 'Q^HUHH^C
Fig. 7
Oxidation products
identified so far
in the ozonization
of substituted-
aromatic substances
By comparing the TOC calculated from the measured products
with the directly determined TOC (Figs. 8-10), it can be
seen that at the beginning of the ozonization only a small
part of the oxidation products (24-60%) was included. The
maximum of the deficit lies at 90-100% elimination of the
starting compounds. At this time 1.5 - 5 ing of 03 are used
up per mg of the initial C-value. Only after higher ozone
consumption can 60 - 80% of the oxidation products be
detected quantitatively. As mentioned previously, the still
unidentified products after 90% elimination of the starting
compound may be compounds with a 6-membered ring structure.
The first attempts to characterize these products showed
that up to 90% degradation of the initial compounds the
BOD^/COD ratio was 0.1 and lower (134). The ozonization of
aniline may be quoted as an example (Fig. 11). After 90%
degradation the solution has a reddish colour. The BOD5/COD
of 0.64 for the non-ozonized solution falls to 0.01 after
-------
TABLE 9 Ozonization of substituted aromatic substances (pH 5 - 2.5).
Initial concentration 1 mmole/1; ozone dose 24 mg O3 min 1
for the p-toluenesulphonic acid; 10 mg O3/min 1 in all
other cases.
Compounds
p-Toluenesulphonic acid
Oxidation products
Fraction of products in TOG, %,
after 90% degradation
for the starting
] compound
, Methylglyoxal, acetic acid,
| pyruvic acid, formic acid,
.oxalic acid, C0~, H-09 ,
I I!2S04
64%
after 5.7 mg 03/tng C
9.9 mole.03/mole PTA
2-Nitro-p-cresol
Methylglyoxal, glyoxylic j 38%
acid, acetic acid, pyruvic after 2.5 mg 03/mg C
4.4 mole 03/mole NC
acid, formic acid, oxalic
3
acid, C02, H2°2' HN°
after higher
ozone dose
79%
after 8 mg 0.,/mg C
68%
after 6.9 mg 0,/mg C
to
in
w
4-Chloro-o-cresol
Methylglyoxal, acetic acid,
pyruvic acid, formic acid,
oxalic acid, CO,,, HC1,
H2°2
24% •
after 1.3 mg 03/mg C
2.3 mole 03/mole CC
65%
after 3 mg 0-,/mg C
-------
80
50
o
en
E
20
TOC measured
I I 1
50 100 150
Ozonization time [min]
200
Fig. 8 Ozonization of p-toluenesulphonic acid,
carbon balance
c = 1 mmole/1; ozone dose 24 mg O3/min 1; pH 3
80
60
o
01 40
E
20
TOC measured
50
100 150 200
tine [min] IOmgOg/min-1
Fig. 9 Ozonization of 2-nitro-p-cresol, carbon balance
c = 1 mmole/1; ozone dose 1O mg 03/min; pH 5.5
-------
80
o 60
en
E
20
DOC
o
DOC calculated from the
measured products
creso!
20 40 60 80 100 120
ozonotion time [min] IQmgOg/min-l
Fig. 1O Ozonization of 4-chloro-o-cresol, carbon balance
c = 1 mmole/1; ozone dose 10 mg 63/min 1; pH 5.5
10 min of ozonization (Fig. 11). After larger additions of
ozone (6 mg Cu/mg C) the solution again becomes colourless.
In this region the oxidation products formed are also better
degradable biologically.
It can be seen from the example of phenols and substituted
aromatics that some of the reaction mechanisms postulated for
ozonization could only be confirmed by clarifying the oxi-
dation products and particularly by their quantitative deter-
mination. In the light of.the material balances obtained
the existence of primary oxidation products may also be
proved, though up to now little is known about their identity
and their properties. Of the remaining classes of substances
it may be said that so far only the reactions of ozone with
unsaturated aliphatic carboxylic acids and aliphatic alcohols
-------
256 -
2 4 6 >
mg ozone consumption/rug initial DOC
15 25 60 120
ozonation time(min) dose 10mg03/min-l
Fig . 1 1 Ozonization of aniline, biological degradability
of the oxidation products in dependence on the
ozonization time
c = 1 mmole/1; ozone dose 1O mg
1 ; pH 6
and carbonyl compounds have been studied in any detail. As
regards the compounds of other substance classes, so far ther<
are only indications of their possible degradability. Even
if the determination of the oxidation products is problematic
because of the complicated composition of the ozonized sol-
utions, one property should be checked, namely the toxicity.
Little is yet known about this and more detailed work is
required.
-------
- 257--
(1) KURZMAMN1,, G.E. . .. ' .
Das indirnkte Ozondruckverfahren
Wasser-Luft-Betrieb 1_ (1963), 11
(2) BREDTMANN, M. • -..-.'-•.
Ozonausnutzunq durch Anwendung des Beliiftungsprinzips
gwf-Wasser/Abwasser 11 5 (1974), 326
(3) SCIIAEFER, R.J. ' ." "
Ozone sewage treatment apparatus
United States Patent 3, 68O, 7O4, Aug. 1 (1972)
U.S. Cl. 210/218
(4) TREJTNAR, J. '
Ozonizer for treating water
Czechoslovakia:! Patent 139879, C O2'.b. 1/38
(5) BLATTER, M.
Verfahren und Einrichtungen zur Aufbereitung von
Detergentien enthaltendem Abwasser inittels Ozon
Schweizer Patentschrift 521292 C O2 c 5/04, 15. Mai 1972
(6) .BOWERS, A., NETZER,- A., NORMAN, J.D.
Ozonation of waste water — some technical^and
economic aspects
The Can. J. of Chem. Eng. 51 (1973), 332
(7) MASSCKELEIN, VJ. , FRANSELOT , ' G . , GENOT , J.
Techniques for dispersing and dissolving ozone in
water
Water and Sewage Works,Dec. (1975), 57
(8) NEBEL", C., UN'ANGST, P.C., GOTTSCHLING, R.
An evaluation of various mixing devices for dispersing
ozone in water
Water and Sewage Works 12O (1973) , R-6
(9) ROSEN, H.M.
Use of ozone arid oxygen in advanced waste water
treatment
J. WPCE 45 (1973), 2521
(1O) ZAITSEV, V.A., KOLDASHEV, A.N. et al.
Contact unit for ozone treatment of waste water
designed by Nickhimmash
The Soviet Chemical Industry 6_ (1974), 10, 653
-------
- 258 -
(11) HEIST, J.A.
Ozone oxidation of waste water contaminants
A Literature Review
AIChE Symposium Series Water 1973, 70 (1973), 136, 456
(12) MAJUMDAR, S.B., CHECKLER, W.H., SPROUEL, O.I.
A physical and mathematical model of mass transfer
and reaction kinetics of ozonation
AIChE Symposium Series 73 "(1977), 166, 188
(13) RICHARDS, D.A., FLEISCHMANN, M., EBERSOLD, L.P.
Mass transfer coefficients for the ozone-water system
AIChE Symposium Series 73 (1977), 166, 213
(14) HOATHER, R.C.
Some experimental observations on the solubility of
ozone and the ozonation of a variety of waters
J. of the Inst. of Water Engineers London 2, (1948) , 358
(15) STUMM, W.
The solubility of ozone in aqueous solution
Helv. Chim. Acta 3-7 (1954), 773
(16) KILPATRICK, M.L., HERRICK, C.C.
The decomposition of ozone in aqueous solution
J. Am. Chem. Soc. 78 (1956), 1784
(17) CZAPSKI, G., SAMUNI, A., YELIN, R.
The disappearance of ozone in alkaline solution
Israel Journal of Chemistry 6_ (1968) , 969
(18) HEWES, C.G., DAVISON, R.R.
Kinetics of ozone decomposition and reaction with
organics in water
AIChE Journal 1_7 (1971), 141
(19) GORBENKO-GERMANOV, D.S., KOZLOWA, I.E.
Intermediate decomposition products of ozone in
alkaline aqueous media investigated by ESR
Zhur. Piz. Khim. 48 (1974), 1998
(2O) KUO, C.H., WEN, K.Y. et al.
Absorption and decomposition of ozone in aqueous
solutions
AIChE Symposium Series 73 (1977), 166, 230
(21) LEBOUT, H.
Fifty years of ozonation at Nice
Ozone Chemistry and Technology, Advances in Chemistry
Series 21 (1959) , 450
-------
- 259 -
(22) KATZENELSON, E., KELTER, B., SHUVAL, H.I.
Inactlvation of viruses and bacteria in water
by use of ozone
J. AWWA 66 (1974),-725
(23) KATZENELSON, E., BIEDERMANN, N.
Disinfection of .viruses in sewage by ozone
Water Research 1O (1976), 629
(24) KELLER, I.W., MORIN, R.A.
Ozone disinfection, pilot plant studies at
Laconia, H.H. -
J. AWWA 66 (1974), 73O
(25) KINMAN., R.N.
Water and wastewater disinfection with ozone:
A critical review
CRC Critical Reviews in Environm. Control 5_ (1975), 141
(26) PERRIL, J., McCAMMON, J., CRONHOLM, L. et al. •
Ozone disinfection and- oxidation in a model ozone
contacting reactor •
AIChE Symposium Series 73 (1977), 166, 225
(27) HACKER, D.S., LOCKOWITZ, T.
Use of ozone in the disinfection of coliphage T-7 virus
AIChE Symposium Series 73. (1977), 166, 242
(28) FOCHTMANN, E.G., RICE, R.G.-, BROWNING, M.E.
Forum on ozone disinfection
International Ozone Institute Inc., Syracuse, New York (1977)
(29) MUDD, J.R., LEAVITT, R., ONGUN, A., McMANUS, T.T.
Reaction of ozone with amino acids and proteins
Atmosph. Environm. _3 (1969), 669
(3O) ' MENZEL, B.
Oxj-dation of biologically active reducing substances
by ozone
Arch. Environm. Health 23 (1971), 149
(31) SHUVAL, H.I.
Forschungsbericht O2 WA 71, Universitat Jerusalem (1976)
,(32) O1 DONOVAN, D.C.
Treatment with ozone
J. AWWA 57 (1965) , 1167 .
-------
-* 260 T<
(33) HOPF, W.
Zur Wasseraufbereitung mit'Ozon und Aktivkohle
gwf-Wasser/Abwasser HI (197O) ,. 83-92/156-164
(34) EOLLUTA, J.
Ozon in der Wasserchemie
gwf-Wasser/Abwasser 1O4 (1963), 129
(35) REISSAUS, K.f RUMMEL, R.- • •'
Wasseraufbereitung mit Ozon bei der Trinkwasserschomang
Fortschr. in der Wasserchemie 6. (1967), 139
(36) KURZMANN, G.
Ozon in der Wasseraufbereitungstechnik
Wasser-Luft-Betrieb &_ (1959), 2O4
(37) MELNYK, P.B., JUDKINS, b.,-NETZER, A.
An ozone reactor for color "removal from
pulp bleachery wastes
Tappi 60 (1977), 3, 97
(38) BAUMANN, H.D., LUTZ, L.R.
Ozonation of a kraft mill effluent
Tappi 52 (1974), 116
(39) MEIJERS, A.P.
Quality aspects of ozonisation
Water Research 11 (1977), 647
(4O) M5RGELI, B.
Chemisch-physikalische Verfahren zur Reinigung von
Abwassern in der Textilveredelungsindustrie
Melliand Textilberichte 56 (1975), 78
(41) PALIN, A.T. - '
The ozonation of v;ater with special reference to
colour removal
Water and Water Engineering, July (1953), 271
(42) CAMPBELL, R.M.
The use of ozone in the treatment of loch turret water
J. of the Inst. of Water Engineers V7 (1963), 333
(43) SNIDER, E.H., PORTER, J.J.
Ozone treatment of dye waste
J. WPCF 46 (1974), 886
(44) FURGASON, R.R., HARDING, H.L.*,' SMITH, M.A,
Ozone treatment of waste effluent
Research Technical Completion Report OWRR
Project No. A - O37 - IDA - April (1973)
(45) FURGASON, R.R, et al.
Ozone treatment of kraft mill effluent
AIChE Symposium Series70 (1974), 139, 32
-------
- 261> -
(46) SUZUKI, I.
Study on ozone treatment of .water-soluble polymers
I. Ozone degradation of polyethylene glycol in water
J. of Applied Polymer Science 2O_ (1976), 593
(47) MAIER, D., KURZ, R.
Untersuchungcn zur Optimierung der Ozonanwendung
bei der Aufbereitung von Seewasser
Internat. Syrnposium Ozon und Wasser Berlin (1977), 211
(48) SONTHEIMER, H., WOLFEL, P., SARFEPT, E.
Improved biological decomposition of organic substances
contained in biologically-cleansed sewage after ozonation
Proceedings 3rd Symposium of the Internat. Ozone Institute
4-6 May Paris (1977) ' , . •
(49) GIURGIUS, W., COOPER, T., KARRIS, J., UNGAR, A.
Improved performance of activated carbon by pre-ozonation
J. WPCF 50 (1978), 308
(5O) GILBERT, E.
Uber die VJirkung von Ozon auf hochmolekulare Wasser-
inhaltsstoffe . . • -
To be published Vorn Wasser (1978)
(51) KOPPE, P., HEPKELMANN, H., SEBESTA, G.
Die Behandlung des Abwassers einer Zellstoff- und
Papicrfabrik rnittels Ozon u-nd Belebtschlammverfahren
Vom Wasser _4i5 (1976), 221
(52) MALLEVIALLE, J.
Action de 1'ozone dans le degradation des composes
phenoliques simples et polymerises. Application aux
matieres humiques contenues dans les eaux
Techniques et sciences municipales et revue 1'eau
70 (1975),107
(53) MALLEVIALLE , LAVAL , Y . , ' LF.FEBRE , M . , ROUSSEAU , C .
The degradation of humic substances in water by
various oxidation agents
Proceedings Ozone/Chlorine Dioxide Oxidation Products
of Organic Matter
Cincinnati, Ohio, Nov. (1976), 189
(54) GILBERT, E.
Reactions of ozone with organic compounds in dilute
aqueous solution, identification of their oxidation
products
Proceedings Ozone/Chlorine Dioxide Oxidation Products
of Organic Matter - ' .
Cincinnati, Ohio, Nov. (1976) , 227
-------
-262 '.-..•
(55) MEIJERS, A.P. ' '''•-<
Entfernuiig von organischen Wasserinhaitsstoffen,
Farbung sowie Geschmacksstoffen durch Ozon
Internationales Symposium Ozon uncl Wasser Berlin
(1977J , 233
(56) HALFON A. et al.
Organic residue removal from waste waters by
oxidation with ozone
Division of Water, Air and Waste Chemistry, Am. Chem.
Soc., Atlantic City meeting Sept. (1968),~ 32
(57) KWIE, W.W.
Ozone treats waste streams from polymer plant
Water and Sewage Works _11j5 (1969) , 74
(58) GARDINER, D.K., MONTGOMERY ,. H .A.C .
The treatment of sewage effluent with ozone
Water and Waste Treatment .Sept./Oct. (1968), 92
(59) SOBER, T.W., DAGON, T,J.
Ozonation of photographic processing wastes
J. WPCF 47 (1975), 2114
(6O) WACHS, A'., NARKIS, N., SCHNEIDER, M.
Organic matter removal from effluent by lime treatment,
ozonat.ion and biologically extended activated carbon
treatment
3rd Symposium of the Internat. Ozone Inst, 4-6 May
Paris (1977)
(61) HILLIS, M.R.
The treatment of phenolic wastes by ozone
Organic matter removal from effluent by lime treatment,
ozonation and biologically extended activated carbon
treatment
3rd Symposium of the Internat. Ozone Inst. 4-6 May
Paris (1977)
(62) WALTER, R.H., SHERMAN, R.M.
Ozonation of lactic and fermentation effluent
J. WPCF 46 (1974), 18OO
(63) WYNN, C.S., KIRK, B.S., NcNABNEY, R.
Pilot plant for tertiary treatment of waste water
with ozone. Environmental Protection Technology
Series EPA - R2 - 73 - 146 Jan. (1973)
(64) MALLEVIALLE, J.
Ozonation des substances de type humique dans les
eaux
Proceedings of the 2nd Internat. Symp. on Ozone
Technology, Montreal/Canada May (1975), 262
-------
- 263 .--
(65) MELNYK, P.B., NETZER, A.
Reactions of ozone with chromogenic ligniris in pulp
and paper mill waste water
Proceedings of the 2nd Internal. Symp. on Ozone
Technology, Montreal/Canada May (1975), 321
(66) NEBEL, C. , STUBER, L.M.
Ozone decolorization of secondary dye laden effluents
Proceedings of the 2nd Internal. Symp. on Ozone
Technology, Montreal/Canada May (1975), 336
(67) DUFORT, J., JONES, J.P.
Direct physicochemical treatment with ozone
Proceedings of the 2nd Internal. Symp. on Ozone
Technology, Montreal/Canada May (1975), 545'
(68) NETZER, A., BESZEDITS, S., WILKINSON, P., MIYAMOTO, H.K.
Treatment of dye wastes .by .o.zonation
Proceedings of the 2nd Internal. Symp. on Ozone
Technology, Montreal/Canada May (1975), 359
(69) DAVIS, G.M. el al. '''' - .
Ozonation of waste waters from organic chemicals
manufacture
Proceedings o'f the 2nd Internal. Symp. on Ozone
Technology, Montreal/Canada May (1975) , 421 -
(7O) SMITH, M.A., FURGASON, R.R.
Use of ozone in the treatment of kraft pulp mill
liquid wastes
Proceedings of the 2nd Internal. Symp. on Ozone
Technology, Montreal/Canada May (1975), 3O9
(71) MCCARTHY, J.M.
The influence of particle size on oxidation of total,
soluble and particulate municipal waste water
components by ozone
Proceedings"of the 2nd Internal, Symp. on Ozone
Technology, Montreal/Canada May (1975) , 522
(72) GILBERT, E.
Uber den Abbau von organischen Schadstoffen im Wasser
durch Ozon
Vom Wasser 43 (1974), 275
(73) ROOK, J.J.
Haloforms in drinking water
J. AWWA 68 (1976) , 168
-------
- 264 -
(74) SANDER, R. , KttHN, W. , SONTHEIMER', H. ' '
Untersuchungen zur Umsetzung von Chlor mit
Huminsubstanzen
Z. f. Wasser und Abwasser-Forschung 1O (1977), 155
(75) MAIER, D., M&CKLE, H.
Wirkung von Chlor auf natiirliche und ozonte organische
Wasserinhaltsstoffe
Vora Wasser £7 (1976), 379
(76) HUBBS, S.A.
The oxidation of haloforms and haloform precursors
utilizing ozone
Proceedings Ozone/Chlorine Dioxide, Oxidation Products
of Organic Material, Cincinnati, Ohio, Nov. 17-19
(1976) , 200
(77) LAWRENCE, J.
The oxidation of some haloform precursors with ozone
3rd Symp. of the Internat. Ozone In'st. 4-6 May Paris (1977)
(78) SCHALEKAMP, M.
Die Erfahrungen mit Ozon in der Schweiz, speziell hin-
sichtlich der Veranderung von hygienisch bedenklichen
Inhaltsstoffen
Infcernat. Symp. Ozon und Wasser Berlin (1977), 31
(79) BLOCK, J.C., MORLOT, M., FOLIQUET, J.M.
Problemes lies a 1'evolution du caractere d'oxydabilite
de certains corps organiques presents dans 1'eau traitee
par I1ozone
Techniques et Sciences Municipales. L'eau 71 J_ (1976), 29
(8O) BUYDENS, R.
L'ozonation et ses repercussions sur le mode d'epuration
des eaux de riviere
La Tribune du Cebedeau 23 (197O), 286, 319
(81) KtlHN, W., SONTHEIMER, H., STIEGLITZ, L., MAIER, D.,
Kurz, R.
Use of ozone and chlorine in water utilities in the
Federal Republic of Germany
J. AWWA TO (1978), 6, 326-331
(82) NAIMIE, G., AXT, G., SONTHEIMER, H.
Zur katalytischen Beeinflussung "der Oxidation von
organ!schen Wasserinhaltsstoffen
Vom Wasser 37 (1970), 98
(83) REICHERTER, U.F.
Untersuchungen liber die Anwendung von Ozon bei der
Wasser- und Abwasserreinigung
Dissertation, Fakultat fiir Chemieingenieurwesen,
Universitat Karlsruhe (1973)
-------
- 265 - • .
(84) EICHELSDORFER, D., v. HARPE , Th.
Einwirkung von Ozon auf Harns.toff im Hinblick
auf die Badewasseraufbereitung
Vom Wasser 3T (1
C-85) DOBINSON, F.
Ozonisation of malonic acid in aqueous solution
Chemistry and Industry, June 27 (1959) , 853
(86) BLACK, W.T., COOK, G.A. .
Production of glyoxalic acid
I. and EC Product Research and Development 5
(1966) , 4, 351 ~
(87) SHAPIRO R.H. et al . . ' \
Ozonization products from caffeine in aqueous solution
Proceedings Ozone/Chlorine Dioxide Oxidation Products
of Organic Materials, Cincinnati, Ohio (1976), 284
(88) SHARIFOV, R.R., BONDAREVA , ' N . I . , MAMED-YARQVA, L.A.
Oxidation von Dichlorbenzol durch Ozon
Azerbaidzhanskii khimicheskii zhurnal (1973) , 4,
124 (Russ.) :
(89) WALTER, R.H., SHERMAN, R.M.
Reduction in chemical oxygen demand of ozonated
sugar solutions by charcoal
J. of Food Science 43 (1-97,8), 404
(90) KRAUSE, H., HEPP, H., -KLUXER, W,
Versuche zur Zers"t6rung von Komplex- und Chelat-
bildnern in radioaktiven Abwassern durch Oxidation
, Kernforschungszentruin Karlsruhe KFK 287, Sept. 1964
(91) KUO, C.H., WEN, C.P..
Ozonizations of formic acid, formaldehyde and methanol
in aqueous solutions
AIChE Symposium Series 166, 73 (1977), 272
(91 a) KRASNOV, B.P., PAKUL , D.L., KIRILOVA, T.V.
Use of ozone for the treatment of industrial
waste waters
Int, Chem. Engin. 1 4 (1974), 4, 747
(92) PAKUL, D.L., KRASNOV, B.P., SAZHINA, A.M.
Oxidation of alcohols in. 'dilute aqueous solutions
by ozone
Zhurnal Prikladnoi Khimii 47 (1974), 1, 36
-------
266 -
(93) KUO, P.K., CHIAN, E.S.K., CHANG, B.J.
Identification of end products resulting from
ozonation and chlorination of organic compounds
cojmnonly found in water
Environmental Science & Techn. 11 (1977), 1177
(94) GILBERT, E.
Chemischc Vorgange bei der Ozonanwendung
Int. Symp. Ozon und Wasser Berlin 77
Colloquium Verlag Otto H. Hess, Berlin (1977), 277
(95) BUESCHER, C.A., RYCKMAN, D.W.
Reduction of foaming of ABS by ozonation
Proceedings 16th Industrial Waste Conference,
Purdue University, Lafayette, Indiana (1961), 251
(96) EVANS, F.L., RYCKMAN, D.W.
Oijonated treatment of wastes containing ABS
Proceedings 18th Industrial Waste Conference
Purdue University, Lafayette, Indiana (1963), 141
(97) KANDZAS, P.P., MOKINA, A.A.
Vorwendung von Ozon zur Reinigung von Abwassern von
synthetischen oberflachenaktiven Anionen-Substanzen
Ochistka proizvodstvennyk stochnykh vod _4 (1969), 76-86
(Russ.)
(98) JOY, P., GILBERT, E., EBERLE, S.K.
Untersuchungen iiber die Wirkung des Ozons auf Alkyl-
benzolsulfonsaure und Waschmittel in wassriger Losung
"Organische Verunreinigungen in der Umwelt" Hrsg,
K. Aurand, Berlin (1978),164
ISBN 3-5O3-O1713-5
(99) REICHERT, J.
Untersuchungen zur Eliminierung kanzerogener aromatischer
Polyzyklen in der Trinkwasseraufbereitung unter besonderer
Beriicksichtigung des Ozons
gwf-Wasser/Abwasser 110 (1969), 18, 477
(1OO) GOMELLA, C.
Ozone practices in France
J. AWWA £4 (1972) , 39
(1O1) HOFFMANN, I., EICHELSD5RFER,- D.
Zur Ozoneinwirkung auf Pestizide der Chlorkohlen-
wasserstoffgruppe im Wasser
Vom Wasser 38 (1971), 198
-------
- 267 -
(1O2) MORGELI, B. .
Entfernung von Pestiziden aus Trinkwasser
Gas, Wasser, Abwasser 52 (1972), 142
(103) ROBECK, G.G. et al.
Effectiveness of water treatment processes in
pesticides removal
J. AWWA 57 (1965) , 181
(104) ROSS, W.R., v. LEEUWEN, J., GRABOW, W.O.K.
Studies on disinfection and chemical oxidation with
ozone and chlorine in water reclamation
Proceedings 2nd Int. Symp. on Ozone Techn.
Montreal, Canada (1975), 497 '
(105) BRGWER, G.R.
Ozonation reactions of selected pesticides for
water pollution abatement
Dissertation Abstracts B (1967), 28, 722-B
(1O6) RICHARD, Y., BRENER, L.
Oi'ganic materials produced upon ozonization of water
Proceedings Ozone/Chlorine Dioxide Oxidation Products
of Organic Materials, Cincinnati, Ohio, Nov. (1976), 169
(107) LAPLANCHE, A. , MARTIN, G., RICHARD, Y.
Etude de la degradation des pesticides par 1'ozone:
cas du parathion
La Tribune du Cebedeau 27 (1974) , 22
(108) LAPLANCHE, A., MARTIN, G., RICHARD, Y.
Ozonation d'une eau polluee par du parathion
T.S.M. 1'Eau 2 (1974) , 407
(109) DAGON, T.J. .
Photographic processing effluent control
J. of Appl. Photogr. Eng, £ (1978), 2, 62
(11O) BAUCH, H., BURCHARD, H.
Uber Versucne, stark riechende oder schadliche AbwSsser
mit Ozon zu verbessern
Wasser-Luft-Betrieb J_4_ (197O), 134
(111) CAULFIELD, M.J.
Kinetic studies of the ozonation of carc.inogens and
mutagens as monitored by biological assay; Univ.Notre Dame,
Dissertation Abstracts Internat. B 38 (1978) , 1Of 4644-4645
-------
- 268 -
(112) SPANGGORD, R.J., McCLURG, V.J. .,.,.. , .,
Ozone methods and ozone chemistry of selected
Drganics in water
Proceedings Ozone/Chlorine Dioxide Oxidation Products
of Organic Materials, Cincinnati, Ohio (1976), 115
(113) SIMMON, V.F., ECKFORD, S.L., GRIFFIN, A.F.
Ozone, methods and ozone chemistry of selected organics
in water, 2. Mutagenic Assay
Proceedings Ozone/Chlorine Dioxide Oxidation Products
of Organic Materials, Cincinnati, Ohio (1976), 126
(114) DORE, M., LANGLAIS, B., LEGUBE, B.
Ozonation des phenols et des acides phenoxy-acetiques
Water Research 12 (1978), 413
(115) BAUCH, H., BURCHARD, B., ARSOVIC, M.
Ozon als oxidatives Abbaumittel fur Phenole in
wassrigen Losungen
Gesundheits-Ingenieur 91 (197O), 9, 258
(116) STRUIF, B., WEIL, L., QUENTIN, £.E.
Verhalten herbizider Phenoxy-Alkan-Carbonsauren
bei der Wasseraufbereitung mit Ozon
.Z. f. Wasser- und Abwasser-Forschung 11 (1978) ,
3, 4, 118
(117) KOROLEV, A.A., KORENKOV, V.N., ABINDER, A.A.
Problems of.hygienic effectiveness of ozone treatment
of water and sewage containing nitro compounds (Russ.)
Gigiena i sanitarja Moskva 11' (1974) ,13
(118) KOROLEV, A.A., ABINDER, A.A., BOGDANOV, M.V.
Hygienic and toxicologic features of products of phenol
destruction in ozone treatment of water
Gigiena i sanitarja Moskva J3 (1973), 6
(119) NIEGOWSKI, S.J.
Destruction of phenols by oxidation with ozone
Ind. and Eng. Chem. 45 (1953), 3, 632
(120) MATSUOKA, H. et al.
Ozone treatment of waste water
Report Mitsubishi Electric Co. 46 (1972), 552
(121) ANDERSON, G.L.
Ozonation of high levels of phenol in water
AIChE Symp. Series Water 73 (1976), 166, 265
-------
- 269 -
(122) PASYNKIEWICZ, J., GROSSMAN, A.., NAWARA, S.
Application of ozone in purification of drinking
water containing phenols
Przemysl Chemiczny 4^7 (1968) , 4, 231
(123) EISENHAUER, H.E.
Increased rate and efficiency of phenolic waste
ozonation
J. WPCF _43 (1971), 2OO
(124) EISENHAUER, H.R.
The ozonation of phenolic waste
J. WPCF 4O (1968), 1887
(1.25) EISENHAUER, H.R.
Dephenolization by ozonolysis
Water Research 5_ (1971), 467
(126) CASALINI, A., LEONI, A., SALVI, G.
Ozone purification of waste water containing phenol:
Investigations of the oxidation mechanism
La Rivista dei Combustibili 31 (1977) , 3, 92
(127) GOULD, J.P., WEBER, W.J.
Oxidation of phenols by ozone
J. WPCF £8_ (1976) , 47
(128) YAMAMOTO, Y., NIKI, E. et al.
Ozonation of organic compounds, II. Ozonation of
phenol in water
Proceedings Ozone Technology Symp. Internat. Ozone
Inst. May-23-25 (1978), Los Angeles, Cal.
(129) HOIGNE, J., BADER, H.
Ozonation of water: Selectivity and rate of oxidation
of solutes
3rd Congress of the Internat. Ozone Inst. Paris, May 1977
(13O) HOIGNE, J., BADER, H.
Ozone initiated oxidations of solutes in waste water:
A reaction kinetic approach
Prog. Wat. Techn. 1O (1978), 657
(131) GILBERT, E.
tlber die Wirkung von Ozon auf Maleinsaure, Fumarsaure
und deren Oxidationsprodukte in wassriger L5sung
Z. Naturforsch. 32b (1977), 13O8
-------
- 270 -
, _. . • - . . i • / .. *./ . i ' i
(132) GILBERT, E.
Die Ozonung von MuconsSure
To be published
(133) ISHIZAKI, K.f DOBBS, R.A., COHEN, J.M.
Ozonation of hazardous and toxic organic compounds
in aqueous solution
Proceedings Ozone/Chlorine Dioxide Oxidation Products
of Organic Material, Cincinnati, Ohio, Nov. (1976), 21O
(134) GILBERT, E.
Investigations on the changes of biological degradability
of single substances induced by ozonolysis
Ozone Technology Symp. May 1978, Los Angeles, Cal.
(135) GILBERT, E., JOY, P., EBERLE, S.H.
Ozonisierung von p-Toluolsulfonsaure und 2-Nitro-p-
Kresol in wassriger LSsung
To be published
-------
- 271 -
OZONE REQUIREMENT AND OXIDATION COMPETITION VALUES OF
VARIOUS TYPES OF WATER FOR THE OXIDATION OF TRACE
IMPURITIES
J. Hoigne and H. Bader
Many water constituents are oxidized in ozonization pro-
cesses. In the last paper Gilbert showed corresponding
case studies (1) . The following fundamental questions thus
arise: How complete are such oxidation reactions after a
certain ozonization time ? Which materials are concen-
trated in the water as intermediate daughter products ?
The action of an ozonization process on the water consti-
tuents is essentially determined by two types of over-
lapping oxidation-initiating reactions (2-5) :
- Direct reaction of the ozone
- Reactions of secondary oxidants formed in
decomposition of the ozone (OH* radical reaction) .
Following a detailed outline of the parts played by the
two types of reaction in earlier works (see e.g. (5)), the
rest of the paper will be devoted mainly to a method of
characterizing a water type with respect to the action of
secondary oxidants. To this end we introduce the concepts
"oxidation competition value of a water type" and "oxida-
tion competition coefficient of a constituent" .
1. Direct reaction of ozone
Ozone can react with many water constituents in direct
reactions :
Mx
MX +(1/1-1)03— M
oxide
-------
- 272 -
Here M represents a specific water impurity and M an
intermediate daughter product formed from it. Within
the duration of the ozonization M reacts with further
ozone molecules to form the daughter product M ., .
Therefore, for the entire course of the reaction, 1/n
molecules of 03 are used up per converted material, i.e.
n is a yield factor, A[M]/A[0,]. The rate-determining
step of the above reactions is by definition only the
first step. Such direct reactions of ozone are of first
order with respect to the ozone concentration and of first
order with respect to the concentration of M (4-6), and
the rate of elimination of a material M is:
dt
= „
_
O-
[Ml
If the ozonization takes place in a batch-type or a plug-
type reactor, it can be inferred from the above reaction
scheme that the concentration of an individual water con-
stituent decreases with increasing time of ozonization
according to: __*/-r*
r_ -i . r. .i l / *i
with
Nt/[M]0 =
M
T/TM = TI • k0 • [03 ]
(1)
(2)
where t^t-' ^ are tne concentrations of the water
constituent M at time t and at the
beginning of the reaction,
t is the time of ozonization,
TM is the time constant of the reaction with respect
to [M] , and
k~ is the rate constant for the reaction of the ozone,
°3
-------
- 273
t corresponds- to the tame-of ozonization necessary in -.
M c . • • - • _
the presence of an ozone concentration [O33 for an elim-
ination of [M] to 37% (1/e), see Fig. 1. I/TM is a •
reaction rate. In a description of the disinfection pro-
cess this is often denoted by y. If M is an individual
substance, i.e. not a sum parameter, t becomes independent
of the extent of the ozonization reaction, i.e. the log-
arithmic elimination curve shown in Fig. 1 becomes a
straight line.
We measured the rate constants (k ) for about 6O different
substances, e.g. on the basis of the rate .of decrease of
the O_ concentration as a function of the M concentration
(6). With the measurement methods so tested the constants
for further substances are easily determined if required.
Fig, 2 gives some examples of these measurements. The
results show that ozone is also an extremely selective
oxidation agent in water: the k-values of even chemically
similar materials can vary by orders of magnitude.
in [M]t/[M]
"1.
0
-1
-2
-3
[ Nt/[M]0
1 nn
™ - I. UU
0.50
Ontj
.o/
0.10
... . o DR
—
(
TM
i
N,
' J ; ^V •
3*T
III ^-
3 1 2 3 4 t '
ozonation time
Fig. 1
Effect of the "direct O3 reaction" •
Logarithmic decrease of the relative concentration of
a "specified trace impurity M, plotted against the
reaction time at a constant ozone concentration
-------
_ 274 -
(M'W1
to5-
to3-
10-
101
cresol
CH3
xylene
) benzene
OH
[fS phenol
aldehyde
>=< tetrachlorethylene
CH3 Hg+
methyl mercury
oxalic acid ion
-100
-10"
-106
r
sec
Fig. 2 Left scale: Examples of rate constants kg (5,6)
"" Right scale: Time constants T of the reaction,
-5,
calculated according to equation 2 for [O 3] = 1O~"'M
(ca. O.5 g/m3), n = l.O
On the basis of the rate constants measured the time con-
stant TM for the reaction of a constituent M can be calcu-
lated for a given ozone concentration. A corresponding
.time scale is set out in Fig. 2 on the right ordinate. It
— *r
refers to a mean ozone concentration of 10 mole 0,/1
3
{ca. 0.5 g/m ) and to the assumption that the reaction
yield is r\ = 1.0. At an ozone concentration 10 times as
high this time scale would be compressed by a factor of 1O,
The illustration shows that cresol enters the reaction essen-
tially within 10 sec. The concentration of phenol is reduced
within about 100 sec to 37% of its starting value. (Other
compounds too, known for the fact that they are easily
chlorinated in chlorination of water, are readily degraded
by ozone (6)).
-------
- 275.
In contrast to these readily oxidized substances, aldehydes,
for example, which are concentrated in sea water as inter-
mediate ozonization products, require really long ozoniza-
tion times for further oxidation (5, 7, 8). Acids such
as oxalic acids, which are also formed as oxidation pro-
ducts, have no chance of being mineralized further by the
direct ozone reaction, even when the ozonization is con-
tinued for hours. Compounds such as methylmercury and
tetrachloroethylene (.9, 1O) also hardly react by the "direct
ozone reaction".
2. Oxidations via secondarily formed oxidants
In the ozonization of water the decomposition products of
ozone, i.e. secondarily formed oxidants, are also available
for oxidation of any trace - impurities: depending on the
water, a part of the ozone, O., , is decomposed. This part
o , A
increases with increasing pH: at pH 8, depending on the
water constituent, it corresponds to about half of the
available ozone within 1O min (11). The OH" radicals
formed react very rapidly with many substances (for liter-
ature data see ref. 3), initiating known oxidations. The
radicals are significantly more reactive than ozone and
therefore less selective. Correspondingly, they are rapidly
consumed by many of the available water constituents (S).
Thus, the oxidation of specific impurities is strongly
competed against by the presence of other constituents in
the water (1O, 11):
M o.xide
where: O., . is the quantity of ozone decomposed,
•j i A
n' is the yield with which OH" radicals were
formed from decomposed ozone,
-------
- 276 -
n" is the yield with which OH" radicals, consumed
by M, oxidize M,
k'M, k'^ are the rate constants with which OH"
radicals react with M or scavengers S.,
and S^ are scavengers, including M, O-, etc.
The elimination rate for this reaction is:
,rMl dO, A k '[M]
d[M] 3,A , ,„ M
—_ 1_ . r,i.rv" .
dt dt
2.1. Oxidation competition value of a water type (flM)
For a batch or plug-type reactor the relative residual
concentration of M at time t is:
M
t
(3)
AM~"TT~ " «• ** "^->^ PiJ) w
KM ' '
On the basis of equation 3, the relative residual concen-
tration of M plotted on a logarithmic scale against the
quantity of decomposed ozone gives an elimination curve of
slope 1/^M (see Fig. 3). OM is a normalization value de-
pendent on the water composition. It increases with in-
creasing loading of the water with substances S. that con-
sume OH" radicals (see ecruation 4) . If o does not essen-
M
tially alter with ozonization, as happens in a large number
of practical situations (cf. Section 2.4), and if M is an-
individual compound (not a sum parameter), the elimination
curve is a straight line . (see Fig. 3). In this case the'
residual concentration of M is 37% ' (= 1/e) of the initial
value after the amount of the decomposed ozone has reached
the value n.
-------
- .277 -
ln[M]t/[M]0,
01234 g/m3
ozone decomposed ,63,4
Fig. 3 Oxidation initiated by OH* radicals
Logarithmic fall of the relative concentration of
an individual trace impurity M with increasing
amount of ozone decomposed in water. Assumption:
the n-value of the water is independent of ozoni-
zation
CH3HgOH
101
100
50
37
25
M. 9
• 5- TO-' M
o 5-KT8 "
A 5-10'7 '
CH3HgOH
n „
"
n.
1
5
c
X
X
10 15 (g/nv>)
zone added
Fig. 4 Measured logarithmic fall of the relative con-
"~~~~~~~~ centration of methylmercury plotted against the
amount of ozone added (mineralization of the
methylmercury by OH* radical initiated reactions)
Water loaded with 5 g/m^ of octanol. pH 1O.5;
O.05 M phosphate (9)
-------
- 278 -
We suggest the following terms' for the ft-value: '
German : "Oxidationskonkurrier-Wert" (des Wassers)
English: "Oxidations-competition value"
French : "valeur de competition a 1'oxidation"
In principle, it is possible to calculate ft for a specific
compound M and a specific water composition according to
equation 4 on the basis of relative rate constants well
known for OH' reactions (11). However, with such estimations
some uncertainties regarding the yield factors n1 and n"
always remain, and likewise regarding insufficiently charac-
terized water constituents which are generally given only
by sum parameters such as the DOC. It therefore seems
better in practice to determine the ft-value of a water
experimentally with a suitable reference solute (12).
In the following it is shown on some examples that the con-
centration decreases found experimentally can be described
by equation 3 and the ^M-value of equation 4.
Fig. 4 shows the measured oxidation of methylmercury to
inorganic mercury(II) dn.ring ozonization (9) . The water
was "loaded" with a model substance (octanol). The observed
oxidation of the methylmercury cannot be attributed to a
"direct ozonization reaction" (9) (cf. Fig. 2), but the
expected logarithmic decrease of the methylmercury concen-
tration is obtained when the relative residual concentra-
tion is plotted against the quantity of ozone added (and
decomposed) according to equation 3. As expected on the
basis of equation 3, the course of the relative concentra-
tion lines does not change even when the initial concentra-
tion of the trace impurity is increased by some powers of 1O.
This means that tho. reaction is exactly of first order
with respect to the trace impurity concentration. In addi-
tion, the slope of the relative concentration lines over
the measured ozonization region remains constant. This
-------
- 279 -
means that ^M does, nol; alter (significantly with ozoniza-^t ^
tion of the water. We obtained similar,'but for analytical
reasons somewhat more restricted, results for several
other model substances, as well as for trace impurities
detectable in lake water by- gas chromatography (5, 8fs 12) .
The mineralization of methylmercury can be determined
analytically particularly simply, exactly, and over a wide
dynamic range. However, methylmercury forms complexes with
many possible water constituents. Large complex-formation
constants are already known for chloride, carbonate, phos-
phate, etc. However, since the reactivity of a compound is ,
strongly affected by such complex formation (9), for char-
acterizing a given water we recommend the use of inert
reference solutes M, such as benzene or tetrachloroethylene;
Fig. 5 shows a Q-value determination in a lake water that
is (measured under Swiss conditions) strongly loaded. Here
0 only becomes constant after a "spontaneous ozone consump-
tion" of about O.5 g ozone/m . The "spontaneous ozone
consumption" produces•only a parallel displacement of the
elimination lines. Even a preliminary ozonization with
1O g O_/m has only a limited influence on the residual
slope of the elimination line (12).
(In waste water previously purified biologically the resi-
dual slope of. the curves also changes only slightly with
the degree of preliminary ozonization in the region.of.
0-10 g O«/m (1O)). Constant 0-values, i.e. constancy
of the residual slopes of the logarithmic elimination lines,
were observed in all the Swiss lake and ground water samples
tested by us. ". • -
-------
- 28O -
%M
too
50
10-
c) .
tetrachloroethylene
3 ' g/m3
ozone added
Fig. 5 Logarithmic decrease of the relative concentra-
' tion of trace impurities plotted against the
amount of ozone added. Water from Lac de Bret
(DOC 4 g/m3, z C02 -^1.6-M, pH -—8.3). Compounds
added: a) 8O mg/m^ benzene, b) 130 mg/m3 toluene,
c) 5OO mg/m3 tetrachloroethylene (12)
10 15 20
I [Sj] octanol (g/m3)
Fig. 6 Oxidation competition- value 0M plotted against the
DOC model loading OS = octanol). pH 10.5 (O.O5 M
phosphate). Compounds added: [Benzene]o = 80 mg/m3,
[HgCH3OH]Q = 10 mg/m3,_{C2C141 = 15 mg/m3 (9)
-------
- 281 - -
The fi-values of various water -types are compared in
Table 1, benzene and tetrachloroethylene being used as
the reference solutes. In water of Lake Zurich the
ft-value for benzene at pE: 8-was about O.8 g of "ozone
decomposed" (O ) per m of water. This means that O.8 g
j r A , .
of decomposed ozone is sufficient to reduce the concen-
tration of a compound such as benzene by a factor of e.
An elimination to 5% of the initial concentration would
necessitate 3 x n = 2.4 g of "ozone-decomposed" per m
(cf. Fig. 3). At pH 10.5'the ft-value rises to 4 g/m .
For tetrachloroethylene, which reacts exceptionally slowly
with OH" radicals, the corresponding values are about 5
times as large (cf. also Section 2.3).
Oxidation effects of similar extents were measured in
other drinking waters (11, 12).
Table 1 Examples of n-values for various water types
• • • • -•• - • •-' — - • •
Lake Zuric?ib)
c)
Dubendorf
ground water
PH
7,7
10,5a)
7,6
9,0a)
3
-D- (g O3/m )
Benzene
0,6
4,0
1,0
3,0
Tetrachloroethylene
3,0
~26
6
— 12
a) pH increased by addition of NaOH
b) z (C02) = 1.2 rtiM ' ' - 'DOC ="1.2 mg/1
c) z (CO2) = 2.6 mM DOC = 1.6 mg/1
2.2. The oxidation competition coefficient
water -constituents . •
of various
can be treated as composite magnitudes. The contri
bution of each constituent S . can be calculated as the pro-
duct of its concentration and a coefficient to,. . :
-------
- 282 -
'/ '/ *M
or:
ft = ojjs, ] + cu0
Jj [Sj
An experimental example of the calculation of the coeffi-
cients u is given in Fig. 6, which shows the measured in-
crease of ft with increasing octanol content of a model
water. We carried out similar measurements for water con-
taining additions of bicarbonate ions, carbonate ions,
free NH_, and fulvic acids. The linear increase of the
n-values observed with loading of a water corresponds
formally to equation 5. The coefficients a) can be read off
from the slope of the curves.
We propose the following terms for w:
German : "Oxidatior.skonkurrier-Koef f izient"
English: "oxidaticr.-competition coefficient"
French : "coefficient de competition a I1oxidation"
Table 2 gives some examples of measured coefficients co.
It can be seen that the O_. , requirement that decreases
3'A 3
the benzene value to 37% rises by 0.5 g/m per g of octanol
loading. The other values are given in mole/mole units.
For bicarbonate the oxidation competition coefficient is
about 1/10 as large as that for carbonate. That is, an
elevation of the pH without elimination of carbonate leads
to a strong increase of the ozone requirement. Even the
effect of NH., (as an oxidation inhibitor) can be significant
in the case of the oxidation of trace impurities in communal
waste water pre-purified biologically if we work in the pH
region in which ammonium ions are dissociated to free NH.,.
-------
283 -
Table 2. . .Examples of coefficients w for different water.
. •_ "' constituents (values from ref. 12)
s
O.c.tanol
Fulvic materials"
(HCO3~ + CO?"1)
eo/ b)
NH/
NH
PH
8-10
8-9
8,0
10,5
10,5
CA
Benzene
0,5 g/g
0,2 g/g
0,1 (/mole
3,0 g/mole
0,0 g^nole
2 , 0 g/fcole
J
Tetrachloro'ethylene
2,3 g/g
0,6 g/mole
13" g/mole
0,0 g/mole
10 g/roole
• a) Soil fraction, soluble at pH 7 and at pH 1
b) Corrected for the proportion of bicarbonate
It is remarkable that the coefficients o> for NH3 are
higher than would have been expected on the basis of the
low rate constants of the reaction of NH^ with OH* radi-
cals (4). The reason is that the primary daughter products
Mx formed from NH., very rapidly consume further oxidants,
themselves becoming oxidized to nitrate. Only a part of
the oxidation stages in this process is controlled by the
oxygen molecules present; experience shows that the con-
sumption of ozone for the oxidation of NB., to nitrate is
really large (n1 . n" < O.25).
2.3. Ratios of the QM~values for different trace
impurities (M) in a water
As can be seen from equation 4, the flM~values of a given
water for different reference solutes M vary in inverse
proportion to the rate constants, and a correction is still
needed for different yield factors n". For the ratio of
-------
- 284 -
the fl-value of" a reference so-lute A-to that 'of • reference1'•' ''"
solute B, equation 4 gives: ' ' ' ' "i.-^*.v(:. •.
fl-./Q., = k' /k1 x (n "TJ/TI "B) (6)
** Jj £3 i\ ij xi
For the oxidations that are only initiated by secondarily
formed oxidants only the reaction of the OH" radicals is
significant in the systems investigated here by us. There-
fore, in equation 6 only the relative reaction rate of the
OH" radicals need be considered. Our earlier measurements
on other model solutions led to a similar result (3). The
special findings that in the systems tested here the hydro-
peroxide radicals (HOI or O_) do not initiate oxidation of
the reference solutes was specially checked, the OH" radi-
cals being converted into HOI radicals by the addition of
HjQyi in the case of benzene and tetrachloroethylene the
H202 additions caused a positive contribution to the £3
value, which was, however, not yet constant.
Pig. 7 shows the ratios expected on the basis of the con-
stants cited in the literature. The rate constants k , with
which the OH" radicals react with the corresponding trace
impurity M are given on the left ordinate (see ref. 3 for
literature). The scale on the right gives a reference
point for the expected fi -values. This scale is calculated
with the aid of equations 3 and 4 and is normalized for the
case in which the water has a a -value of 1 g/m
benzene
(the waters of Lake Zurich, Lake Constance, ground water
in Zurich, etc. have somewhat lower normalization values
(fl .-^ O.8 g O_ /m )) . In addition, it is assumed that the yield
J , A
factor n" is in all cases the same as for benzene. This
assumption is justified on the basis of our present exper-
ience, provided that only an estimation of the ratios is
required. From Fig. 7 it is not sufficiently clear that for
the usual organic compounds k1 values of comparable magni-
tude are obtained, since substances that display an excep-
tionally low reactivity were used on purpose.
-------
- 285 -
Consultation-, ofmore .detailed tabulated values shows,_> • ,,
however, that the rate constants k' of the majority of
organic compounds of average molecular weight are situ-
ated in a region that extends from tetrachloroethylene to
a little above benzene.
(M"W)
1010
109-
10s-
..-•CH3CH2CH2CH0,
CI-C-C-C1
col"
----- NHo •••• •- ..,,...,.. , ..
'CH3COO"
^HC°3' coo- :...
COQ-
-1 n
- 10
-100
I
(go
Fig. 7. Left-hand scale: Examples of rate constants with
which OH* radicals react with trace im-
purities (for literature see ref. 3).
Note that substances with k < 10 M"1
sec~l are taken to be special exceptions
Right-hand scale: Oxidation competition value QM.
Scale normalized to 0, =1.0
(average value found en enefor Swiss
lake water, pH -~ 8} . It is assumed that
M
= n
benzene
It follows from Pig. 7 that in a type of drinking water
expected for Swiss conditions (provided pH < 8.5) at a
3
decomposition of 1-10 g of ozone per m a reduction of many
of the possible trace impurities by a factor of "e" can be
-------
- 286 -
reckoned with. A reduction to 5% would require about 3
times this fl-value of ozone Ccf. Fig. 3}.
Measurements of the behaviour of the organic trace impur-
ities present in the water of Lake Zurich in concentra-
tions of only yg/m , such as toluene, xylene, chlorobenzene,
etc. show that these eliminations too correspond quantitatively
to the principles outlined here (5,6).
2.4. Phenomenon of constant n-values
In principle the fiM-value is given by the slope of the
products' elimination graphs plotted on a logarithmic scale
according to equation 3. These curves are mostly repre-
sented by a straight line for reference solutes in natural
water and even in waste water that has previously been bio-
logically purified. In some waters a correction must how-
ever be made for an initial spontaneous ozone consumption
manifested by a straight line not passing through zero.
However, the residual slope even in these cases depends only
slightly on the preliminary ozonization, so long as this
occurs with only a small quantity of ozone which is usual in
the treatment of drinking water. The phenomenon that SI is
almost independent of ozonization (the slope of the elimin-
ation curve is constant) simplifies many estimations, but
should not be regarded as a matter of course. The fact
that £2-values frequently remain constant after a small
spontaneous ozone consumption can be explained as follows:
a) An essential proportion of the fi-value is based on
the contribution of bicarbonate and of the content
of carbonate ions in the water (.11) . If the pH
of the water does not change, this contribution
hardly changes during the ozonization.
b) Another essential proportion of the fi-value is based
on the OH"-consuming action of "humic substances"
and other refractory organic compounds, mostly of
-------
- 287 -
higher molecular weight. Oxidation of these
scavengers changes the degree of oxidation and
perhaps also the molecular weight of these
compounds. However, the reactivity of the whole
organic material with respect to the OH" radicals
is scarcely affected by this. Even water in which
a DOC loading is simulated only by the addition
of octanol displays a n value independent of
ozonization. In this case the constancy of fl is
based on the use of a DOC model compound in which
the sum of the oxidation products formed and the
starting substance consume the OH'radicals at
similar rates (for a detailed discussion see
ref. 12) .
c) Even the effect of a substance such as NH_ changes
only slowly with advancing ozonization: the free
NH,. reacts really sluggidily (4) , and, if the pH
is considerably less than 9, the water contains
a large reservoir of unreactive ammonium ions from
which NH_ can always be released in accordance with
the equilibrium conditions.
Many other water constituents, however, lose their radical
scavenger effect as the ozonization progresses. In water
in which such constituents become decisive the ft-value
changes in accordance with the ozonization. Such substances
seem to have only a subordinate significance in our drinking
waters.
In principle, OH* radicals can also be consumed by ozone
molecules themselves (.12) , in which case the n-value would
be expected to change with the nature of the ozone input.
The proportion of ozone in the over-all ft decreases with
increasing instantaneous ozone concentration and increasing
cleanliness of the water. However, an experimental deter-
mination of the u)Q coefficient may prove really difficult.
-------
- 288 -
. . . .
The substance to be measured M also gives a contribution
a>M[M] to the J2 - value. If the sum of the oxidation pro-
ducts formed from M gives a contribution different from
the starting substance, the over-all fl-value of the water
will change with progressing ozonization. This only de-
creases in weight if M is not a true trace but a strong
impurity. In such cases the ozonization effect is more
simply described on the basis of the ozone yield (AM/O-, . ) ,
-j f A O
as is obtained for low conversions (start of the ozoniza-
tion) . The kinetics of such systems have been treated in
earlier papers (9) . The characterization of an ozonization
system with the aid of the n values as suggested in the
present work is, however, generally more informative and
always suitable when the water properties with respect to
indirect oxidation effects of the ozonization are to be
characterized.
In the cases investigated the same coefficients u were
found in model phosphate buffer solutions as in lake
water or ground water. When using a borate buffer (O.O5 M)
however, we observed an increase of all en-values, which at
the moment we can only explain by an alteration of the n '
value caused by the borate .
3. Cooperation of the direct reaction of ozone and
the radical mechanism
Since ozone has a very high substrate selectivity, very few
substances are included by chance in a region in which they
are simultaneously oxidized essentially by the ozone and
by the ozone-decomposition products. In the treatment of
drinking water with a pH of about 8, i.e. a water in which
half of the ozone decomposes within about 1O min, this may
be just the case for xylene (cf. Fig. 2). For a compound
such as benzene the "direct O- reaction" already plays a
part about 1OO times smaller. On the other hand, phenol-
like trace impurities are degraded 1OO times faster directly
by ozone than the ozone can decompose at pH 8.
-------
- 289
There is an important difference between the two types of
reaction: the two types of oxidation lead to different
intermediate products/ i.e. different daughter products,
which can become concentrated in the water, and they are
governed by different ozonization process parameters. The
"direct 0_ reaction" can be improved by increasing the
ozone stability. The indirect reaction, however, the "OH*
radical reaction", is based on a decomposition of ozone,and
in contrast to the "direct 0_ reaction" it is inhibited
by constituents such as bicarbonate or carbonate. This
property of the OH* radical reaction allows the extent of
the reaction to be ascertained experimentally, by deter-
mining the effect of the addition of relatively inert
substances such as bicarbonate (5, 11) or butanol, octanol,
etc.
(1) GILBERT, E.
Chemische Umsetzungen und Reaktionsprodukte bei der
Ozonbehandlung von Trinkwassern
(see paper Gilbert, this publication)
(2) TAUBE, H,, BRAY, C.
Chain Reactions in Aqueous Solutions Containing Ozone,
Hydrogen Peroxide and Acid
J. Am. Chem. Soc. 62_ (194O), 3357-73
(3) HOIGNE J., BADER, H.
The Role of Hydroxyl Radical Reactions
in Ozonation Processes in Aqueous Solutions"
Water Res. 1O (1976), 377-386
(4) HOIGNE, J,, BADER, H.
Ozonation of Water: Kinetics of Oxidation .of Ammonia
by Ozone and Hydroxyl Radicals
Environm. Sci. & Techn. Jj2_ (1978), 79-84
(5) HOIGNE, J,, BADER, H.
Kinetik und Selektivitat der Ozonung organischer Stoffe
in Trinkwasser . . .
Wasser Berlin 77, Tagung der Fachgruppe Wasserchemie
in der GDCh (Colloquium Verlag Berlin 1978), 261-276
-------
- 290 -
(6) HOIGNE, J., BADER, H. • . ,. -, ,,
Rate Constants for Reactions of Ozone with Organic
Solutes in Water
To be published
(7) SCHALEKAMP, M.
Die Erfahrungen mit Ozon in der Schweiz, speziell hin-
sichtlich der Veranderung von hygienisch bedenklichen
Inhaltsstoffen
Wasser Berlin 77, Tagung der Fachgruppe Wasserchemie
in der GDCh (Colloquium Verlag Berlin 1978), 31-69
(8) ZURCHER, F., BADER, H., HOIGNE, J.
To be published
(9) HOIGNE, J., BADER, H.
Ozone and Hydroxyl Radical Initiated Oxidations of
Organic and Organometallic Trace Impurities in Water
175th Am. Chem. Soc. Meeting, Anaheim 1978, ACS Symp.
Series (in press)
(10) HGIGNE, J., BADER, H.
Ozone Initiated Oxidations of Solutes in Wastewater:
A Reaction Kinetic Approach
Progr. Water Technol. 1O (1978), 657-671
(11) HOIGNE, J., BADER, H.
Beeinflussung der Oxidationswirkung von Ozon und
OH"-Radikalen durch Carbonat
Vora Wasser 48 (1977) , 283-304
(12) HOIGNE, J., BADER, H.
To be published
-------
- 291 - ,
TRANSFORMATION OF HUMIC ACIDS BY OZONE
J. Mallevialle
It is no longer necessary to stress the advantage of following
the course of humic-type natural organic materials through
the different stages of water "treatment with a particular
mention of the use of oxidizing agents. The action of
chlorine on the humic and fulvic acids of water^ in particular
with the formation of halomethanes, has been described in
many publications (1, 2). In our laboratory we are working
on the action of ozone, which is often described as being
very effective in decolorizing natural water, but we are not
confining ourselves to strongly coloured water, since we
believe that a large part of the organic material in water
can be compared to humic and fulvic acids - referred to as .
"HA" hereafter (3) .
1. THE DECOMPLEXING ACTION OF OZONE
The structure scheme proposed by Gjessing (Fig. 1) (4) illus-
trates one of the first consequences of the ozonization of
humic acids. These acids have a "core" at the periphery of
which mineral elements may be found in addition to organic
compounds such as pesticides. (Lindane or D.D.T.) . When
ozonization takes place; these various elements are released
into the water;the iron precipitates in the form of ferric
hydroxide, the manganese may be converted into MnOT ions
giving a violet tinge to the water; and the organic compounds
little attacked by ozone may be- found in the water, as happens
for example in the case of Lindane. We have already des-
cribed this type of phenomenon in previous publications
(5,6).
-------
- 292 -
Ft OH
DCNZENECARDDXYLIC ACID
OH
METHOXy-BEH2ENECARBOXYI.li: ACID
Fig. 1 Schematic structure of humlc acids (after Gjessing)
2. STUDY OF THE OZONIZATION BY-PRODUCTS
2.1. Compound-identification tests
If one considers the example of the humic acid core structure
put forward by Schnitzer and Khan (Fig. 2) (7), two important
observations may be made:
a) During the ozonization of natural organic substances
a large quantity of by-products is formed, the number and
nature of which will depend on the type of the water and the
amount of ozone used. Detailed analysis of these by-products
is complicated because of their diversity and the difficulty
of obtaining reference compounds. With small quantities of
ozone we observe depolymerization with liberation of phenolic
or quinonic "monomers"; this is what we observed, for example/
-------
- 293
OH
Fig. 2 Suggested structure of the -humid acid "core"
(after Khan and Schnitzer)
by thin-layer chromatography. With larger quantities of
ozone we observed the opening of benzene rings and the form-
ation of aliphatic aldehydes and acids, as has been shown by
many authors working on the ozonization of phenols (8,9).
b) Humic acids are not well-defined compounds but
complex mixtures of organic compounds in constant evolution,
for which it would be difficult to determine characteristics
other than by over-all techniques.
2.2. Search for parameters permitting the transformation
of the HA by ozone to be studied
Curves 3 . and 4 represent the variation of various "parameters
in the course of ozonization of a strongly coloured water
free of any industrial pollution (little-mineralized water
containing 20-40 mg/1 of dissolved organic carbon) (10).
We preferred to select a real-case rather than a solution of
HA extracted from a given- soil. Fig. 3 is an example of the
results we obtained, confirming' the comments in the preceding
-------
- 294 -
pci/j hA
meq/1 UV
200 H
150n
100-
Fig. 3
Elimination of humic acids
as a function of ozonization
time
120 min
section. The colour disappears very rapidly, while the con-
centration of COOH groups goes through a maximum and dimin-
ishes more or less rapidly. The HA content, measured by the
reduction of phosphotungstic and phosphomolybdic acids
(polyhydroxy aromatics dosage (11)} decreases much
more slowly than the colour.
Over-all parameters of the type of the TOG, COD,
TOD, and
2.2.1.
Fig. 4 shows the variation of several over-all parameters
currently used in this field as a function of the time of
ozonization. Here again it is found that the colour dis-
appears very rapidly. The COD and TOD follow approximately
the same course. For low amounts of ozone we observed an
increase in the BOD,., which goes through a maximum before
falling off markedly. The conclusion from this type of
results is that we have transformed the organic materials
present into more bio-degradable compounds containing more
and more oxygen without reaching total oxidation.
-------
- 295.
200
100
/s-J
/
/
-.......-Color
+ TOD
COD
TOC
0 10 20 30 40 50 60
Ozonation Time (win)
100% . /*- ^,
95
75
50
25
/
/
/~y
/
/
/
,
1 /' /
r' /
/
/
,
/
x
/'
,
^
_/
.
. _.,
X1
X
i ^ 1
X
X
/
— — -*"
^,/
**"*
>
/
/
**"
*
/
'
Color
PolyhycJroxyaromatic
"~~ " corooounas
,_TOD
_.. TOC
0, mg
0 100 200 300 400 500
Fig. 4 Ozonization of natural waters
-------
- 296 -
2.2.2. Other parameters
Since the determinations of BOB ' and TOB are. difficult in the
case of low concentrations, we felt it important'to look for
other characteristic parameters sufficiently sensitive to be
applied in the sector of the production of drinking water.
We have thus been led to use the. absorption at 254 ran
currently applied in certain countries (12), and in parti-
cular the fluorescence intensity in ultra-violet light.
The latter in fact provides two elements of characterization
(Pig. 5), the fluorescence intensity, which is proportional
to the concentration, and the maximum emission wavelength
A which, on the basis of our experience, is determined by
the degree of polymerization or size of the molecule. The
larger is the molecule, the more is A displaced towards
longer wavelengths. For example, Seine water upstream of
Paris has a A of 415 nm, while downstream of Paris its A
is only 405 nm (13).
lionosulphonate
- M.E.iei
^ ticiution: 320 nm
Fig. 5 Fluorescence spectra of a solution of lignosulphonates
and of a solution of humic materials
-------
- 297 - • . , -
While working on a contract for the Ministry of the Environ-
ment we established significant correlations between the
fluorescence intensity and ,the various over-all parameters
described in Section 2.2.1.'- These correlations were based
on 500 water samples div.ided. into different types according
to their nature or origin." By way of example, Table 1 shows
the values obtained for the pair of parameters "fluorescence"
and "organic carbon". The" correlations are significant In
all cases where it is- not desired to carry out an oxidation
treatment by ozone. This result can be easily understood if
one considers the results in Fig. 6, showing the variation
of these parameters as a function of the time of ozonization.
60
60
Fig. 6
Ozonization of
natural waters
-------
TABLE 1 Tests of the correlation fluorescence = f (TOG)
for various waters
Correlation
Coefficient
Equation
No. of Degrees
oE Freedom
Significance
Level
All Classes
0,8 10%
< 25%
Chlorination
0,973
(F)=0,626(C)
+ 1,654
68
«,.,,
oo
I
Correlation
Coefficient
Equation
No. of Degrees
of Freedom
Significance
Level
Mo r sang
0,713
(F)=0,746(C)
+ 0,37
132
<0,1%
Viry
0,363
(F)=0,502(C)
+ 1 ,06
84
<0,1%
Suresnes
0,000
(F)=-0,01 (C)
+ 3,96
50
not
significant
Croissy
0,200
(F)=0,12(C)
+ 3,46
66
> 5%
<10%
Aubergenville
0,225
(F)=0,248(C)
+ 3,3
87
> H
<2,5%
Mare aux Evees
0,543
(F)=0,267(C)
+ 19,2
18
>0,1%
< 1 %
Various
0,417
(F)=0,637(C)
+ 0,868
61
<0,1%
-------
- 299 -
V: Thus/; we have just seen •. that- the' over-a 1-1 'parameters used'
by us vary in a noticeably different manner in the course of
an oxidation treatment. While the TOC gives the absolute
value of the number of atoms of dissolved organic carbon, it
gives no indication of the chemical nature of the organic
compounds present. On the other hand, fluorescence, for
example, indicates the presence of cyclic compounds without
giving their content in absolute value, since we have here
a mixture of compounds. The parameters we have described
are thus in fact complementary, and we believe it is essen-
tial not to calculate the elimination yields of the organic
materials in relation to just one of these parameters, in
view of the risk of grave errors.
2.2.3. Techniques of separation on membranes
As has been demonstrated by certain authors (4), natural
waters containing HA are complicated solutions of organic
compounds with widely different molecular weights. The
technique of ultrafiltration through Diaflo-Amicon membranes
makes it possible to fractionate the organic compounds
according to different ranges of molecular weight. Here
again we believe that it is necessary to use several par-
ameters for characterizing the organic material. (We have
retained the TOC, absorption at 254 nm, and fluorescence).
The ultrafiltration technique has also enabled us to verify
that the HA emission maxima are a function of their degree
of polymerization. Fig. 7 shows an example of this type of
separation with a water described in Section 2.2. After
ozonization that eliminated 80% of the colour, '30% of the
fluorescence and of the 254-nm absorption, and finally a
certain percentage of the organic carbon, we find an increased
homogeneity of the distribution, explained by transformation
of the HA into compounds of lower molecular weight.
-------
- BOO -
ItPC OF
MFMBRih'E
WEIGHT
75 v.-
Mr/. -
"ov.
XM 300 c '. XM5&pM3th * PH10' PM 5,T - •„ ^-IIM 2 JIM OS
300000 SO 000 30 000 IP 000 5000 1000 SOO
. ii)
•
!
i
— _j — ^^^
,
..,„
r*fSt*44+S
I*"'"!
i
[
-------
- 301 -
water through a 5-litre "react9r. ' We -then ran comparative
jar tests on the water with and without ozonization. In the
majority of cases for a given amount of coagulant,differences
of 5-30% were observed in favour of pre-ozonization in the
elimination of the absorption at 254 nm and fluorescence. In
contrast, this improvement was only of the order of
0.1-0.2 mg/1 in the case of organic carbon (raw water
3-4 mg/1). It is probable that in a pilot or industrial
installation the elimination of organic carbon would be
greater due to biological degradation in the sand filter; at
least this is what we are trying to demonstrate at the moment
on a pilot installation (2 m-Vh) .
TABLE 2 Effect of pre-ozonization (1 mg O3/l) on the
efficacy of coagulation of Seine water by
aluminium sulphate with coagulant aid
NON-OZONIZED PAW WATER
A'-IOUNT OF COAGULANT
(ma/1)
riocculation eval. (15 min)
Turbiditv
(drops 0. mastic)
Ontical density at 254 nm
(referred to 1 m)
Fluorescence (v.V)
Total organic carbon
(mg/1)
0
0
48
7,52
O,64
3,3
20
2
25
5,02
O,62
2,9
40
5
16
3,72
0,57
2,3
60
5
18
3,28
0,54
2,2
80
3
23
2,76
0,49
1,9
10O
3
23
2,46
0,47
1,6
OZONIZED RAW WATER (1 mg/1 of O.j)
AMO'JNT OF COAGULANT
(ir.g/1)
Flocculation oval. (15 irirO
Turbidity
(drops of mastic)
Ontical density at 254 nm
(referred to 1 m)
Fluorescence (niV)
Total oraanic carbon
(mg/1)
O
C
51
5,93
o,4i
3,3
20
2
25
3,42
0,37
2,7
40
3
23
2,88
0,32
2,1
60
5
2O
2,32
0,23
2 , I-
80
4
25
1 ,96
0,29
2,2
100
3
29
1 ,86
O,28
1,3
-------
3O2 -
We have also tried to explain the increase in turbidity
observed after ozonization by counting the particles in a
Coulter counter (Figs. 8 and 9).
In most cases we observed an increase in the total particle
count, with a slight reduction in the number of large
particles, as shown-in Fig. 9.
kNO, Of
nartlcles/wl
Raw water (T,n.p.»2 591 50O)
+O3 (2 pran) (T.n'.r».=78 4OO)
T.n.p, * total number of oarticles
Equivalent spher
diameters In
Fig. 8
Cholet water
4 6 810
t. of
pnrticlea/ml
10
_Raw water (T.n.p.-3 133 OOO)
._+03 <2 ppm) (T.n.p,«4 227 OOO)
T.n.p, » total number of particle*
Fig. 9
Cholet water
O 2 4 6 810
-------
- 303 -
' ' '**.'•, i .'< i , -'-.'.- * i".'..'. •'•*''','„ \ <".t' i','r.'"-. '. ,:" "'.. ',' ",'r '", '.''•' ••'.' "; ,
3.2. Effect of ozonization on filtration on granular
carbon .
Fig. 1O shows the flow diagram of a waterworks for treating
Seine water upstream of Paris. Three industrial lines and
a pilot line provide a comparison of the efficacy of
different combinations of ozonization and filtration in a
refining treatment. We shall not list here the results on
the elimination of micropollutants, since this has already
been done and will be dealt with in various publications (14).
We have simply indicated by way of example the variation in
the fluorescence/organic carbon (=F/C) ratio. It is noted
that after each ozonization operation F/C diminishes markedly
and increases again after filtration through granular carbon,
which confirms the results we obtained in the first part of
this study.
3.3. Biological degradation of organic materials
Fig. 11 is a flow diagram of a waterworks for the treatment
of Seine water downstream of Paris. The two parallel
industrial lines (continuous lines) are based on two diff-
erent principles, one (Chabal type) using slow filtration,
the other using standard clarification treatment with the
addition of powdered active carbon. Over several months of
operation these two lines give equivalent results on the
basis of the organic carbon as a parameter. On the other
hand, the absorption at 254 nm and in particular the fluor-
escence are much greater at the outlet of the slow filtration
line.
Furthermore, the maximum emission wavelength of fluorescence
is greater in this last line. Two explanations are possible:
the slow filtration has a smaller eliminating- effect on
humic acids of high molecular weight, or the phenomenon is
caused by the formation of metabolites of relatively high
molecular weight by microorganisms. This is what we intend
to study during the coming months.
-------
- 3O4 - v
Fig. 11 also shows, by means of the broken lines, an experi-
mental line in which preliminary oxonization is carried out
before the slow filtration stage. The results obtained
should enable us to confirm the increased bio-degradability
of organic materials after oxidation with ozone.
Raw Water
Decantation (5)
Deeantation + Active Carbon Decantation (4)
(3) Powder
Sand Filtration (6) Sand Filtration (7) Sand .Filtration
F/C:
0,36
Ozonization
(9)
Ozonization(ll) Granular Carbon Granular Carbon Granular Carbon
Filtration (12) Filtration (13) Filtration (15)
F/C:
0,20
F/C:
0,21
/
*'
Ozonization
(Pilot)l (14)
F
F/C:
0,12
F/C
0,16
F/C:
0,80
Industrial Line Industrial Line Industrial Line Pilot Line
NO-1 No.2 No.3 No.4
Fig. 10 Morsang/Seine Waterworks
Flow diagram of different lines of treatment
-------
Raw water
Pre-
chlorination
Coagulation flocculation
with adjuvant and addi-
tion of active carbon
powder
r _ •* f
I Coagulation floccu-. |
I lation with coagulantj
! aid i
Rapid sand
filtration
r JL___,
I Ranid sand ;
I '
filtration |
L. „. ^, _
F = 0,75
TOC = l,'l
I Ozonization'
V
'Slow filtration
1 _________ i
Bactericidal treatment
with chlorine peroxide
roughing tank
(gravel)
Addition of
active carbon
powder
First sand
filtrati'on
Slow
filtration
R =1
TOC = 1,6
o
ui
Bactericidal treatment
with chlorine peroxide
Fig. 11
-------
- 3O6 - -
4. CONCLUSION
Several observations may be made about the various tests
carried out:
- It is practically impossible to achieve total
degradation of natural humic-type organic materials by means
of ozone. In the majority of cases there is a conversion of
these organic materials into compounds of lower molecular
weight. A practical conclusion may thus be drawn: The time
of contact and the amount of ozone must not be reduced more
than it is necessary, and it will often be advisable to
follow the ozonization with a filtration treatment, e.g. on
granular active carbon.
- When the different behaviour of various waters with
respect to oxidizing agents such as ozone is considered, it
must be realized that each water constitutes a special case
and that a study of the ozonization by-products will be very
difficult in view of their extreme diversity.
- When carrying out oxidation treatments it seems
important to us to have available a selection of parameters
of the type described. Elimination yields calculated just on a
single parameter will.in most cases lead to wrong conclusions.
-------
- 307 -
(1) ROOK, J.J.
Formation of haloforms during chlorination of natural
waters • ...••,.;.•
Water Treatm. Examin. 2_3_ (1974), 234
(2) STEVENS, A.A., SEEGLER, D.R., SLOCUM, C.J.
Products of chlorine dioxide. Treatment of organic
.materials in water
_I.O,I. Workshop Ozone/Chlorine Dioxide Oxidation Products
of 'Organic Materials'. 17-19 Nov. 1976, Cincinnati, USA
(3) SONTHEIMFR, H. ' ......
"Conference at Karlsruhe, Federal Republic of Germany
(1,975)
(4) GJESSING,E.T. ' •
Physical and chemical characteristics of Aquatic Humus
Ann Arbor Sciences (1976) '
(5) BUFFLE, J.Ph., MALLEVIALLE, J.
.Le role des matieres humiques envisagees comme agent
d1 accumulation et vehicule" des" substances toxiques dans
les eaux.' . • ,
Techniques et Sciences Municipales, n°special juin(1974),
331-340 ' • ' ' ••
(6) MALLEVIALLE, J.
Degradation of humic substance in water by different
oxidation agents (ozone, chlorine, chlorine dioxide)
I.O.I.. Workshop Ozone/Chlorine Dioxide Oxidation Products
of Organic Materials, 17-19 Nov.; 1976, Cincinnati, USA
(7) SCHNITZER, M,, KHAN, S.U.
Humic substances in the environment
'Marcel Dekker, New York (1972) ~ . . . .
(8) EISENHAVER, H.R.
The ozonation of phenolic wastes
J. WCPCF 4O (1968), 1887
(9) DORE, M.
Ozonation des phenols et des acides phenoxy-aeetiques
Water Research 12 (1978), 6,
(1O) MALLEVIALLE, J.
Les agents complexants naturels des eaux
Etude des proprietes physico-chimiques des matieres
humiques et de leur transformation par ozonation
These presentee a 1'Universite Paul Sabatier de Toulouse
(Nov. 1974)
-------
- 3O8 -
(11) -
Standard Methods for the Examination of Water and
Waste Water
13 Ed., 346 ' "
(12) SONTHEIMER, H, et al.
The Miilheim Process
J. AWWA TO (1978) , 1, 393-396
(13) MALLEVIALLE, J.
Etude de la signification de la fluorescence dans les
eaux dans ses rapports avec leurs qualites executee
pour le Ministers dfe 1'Environnement (Avril 1978)
(14) RICHARD, Y., FIESSINGER, F.
Emploi complementaire' des traitements ozone et charbon
actif
3©me Congres International de 1'Institut International
de I1Ozone, 4-6 Mai 1977, Paris
-------
OZONE AND HALOGENATED ORGANIC COMPOUNDS
E. de Greef, J. Hrubec and J.C. Morris .
1. Introduction . .
Advances in the analytical determination o± "organic
pounds in water and the discovery of the formation of
halomethanes during chlorination are the most important
factors leading to the initiation of a number of studies
on the influence'of ozoriation on halogenated substances
in drinking water. During the initial period, after the
finding of the formation of trihalomethanes, special
attention was paid to the effects of ozone, which was
proposed as one of the most hopeful alternative oxidants
to chlorine, and to the ramoval of halomethane precur-
sors, as well as to the removal of halomethanes them-
selves. During a more recent period interest has been
focused more upon the general influence of ozonation
on the presence of halogenated compounds.
In the following paper the influence of ozonation on the
occurrence of halogenated organic compounds is discussed.
Special attention is paid to the combination of ozonation
with breakpoint chlorination in relation to observed chan-
ges in the concentrations of haloraethanes. In addition to
this/ possible mechanisms for observed phenomena are pro-
posed.
-------
- 310 -
• - > ' ' ' Y *
2. Formation of Halomethanes by Chlorination Following
Pre-ozonation
The influence of preoxidation by ozone on the formation
of' halomethanes has been studied intensively by Rook (1,2)
and Montiel (3). These studies have been carried out un- •
der conditions similar to, those in practice with regard
to the dose of ozone, pH and temperature of the treated
water.
The use of preozonation to reduce the formation of halo-
methanes was based on the assumption that ozone, being
the stronger oxidant( would destroy or oxidize those sites
in the organic molecule which are suited for the forma-
tion of.halomethanes (2) .
From- Table 1 it can be seen that pre-ozonation as a
part of the, water treatment at two Rotterdam Water-Works.
and the Paris Municipal Water-Works at Orly has no sub-
stantial effect on the reduction of halomethanes when
the water is subsequently chlorinated.
Also, according to the results of studies by the E.P.A.
(6) ozone at a dosage level less than 5 mg/1 will not
decrease the concentration of trihalomethane'precursors.
The described findings are only valid for the application
of pre-ozonation without subsequent adsorption on acti-
vated carbon and without additional biological processes
like slow sand-filtration (7). Also the combination of
ozonation with ultra-violet radiation does not apply be-
cause in these combinations the application of U".V. seems
to provide a positive contribution to the increased re-
moval of both organic matter together with halomethane
precursors (6,9).
-------
- 311 -
TABLE 1 Formation of halomethanes (pg/1) by use of ozone
prior to chlorination (lit. 2,3) '''
' Contact time
after ozonation
(in hours)
0.15
4-5
24
without ozone
Water treatment plants , - ' ,•>-,•
Orly
CHC13
Rotterdam-Kralingen
CHC13'
I
2O
30
22
21
50
-
3O
32
~~__~^__
CHCl2Br
8
11
11 .
18
CHClBr,
8
9
7
11
Rotterdam-Berenplaat
CHC1,
,
CHCl2Br
4 2.3
7
12
12.5
2.5
5.5
5.8
CHClBr2
.-••' • : '. •
2.5
2
• 3
2.5
Furthermore, it should be noted that also some labora-
tory studies have indicated that, when the ozonation is
used prior to"chlorination, the ozone did remove precur-
sors and consequently reduce'the halomethane content to
some extent' (4,5,6). ' : - •• •
It seems that only additional studies directed to,-im-
proving our knowledge of the chemistry of ozonation re-
actions with organic substances, particularly the-mecha-,
nisms and kinetics involved/ can provide a better, insight
into the problem, of preozonation and halomethane forma-
tion.
3. Removal of Halomethanes by Ozonation
The tests which have been carried out to study,the effects
of ozonation on the removal of halomethanes .(3,4) have
shown that ozone, under the conditions of water,treatment
practice and at the level of ozone doses concerned, does..
-------
- 312 -
not'decrease the cbhcehtrati'ons :df" "haiome-thanes already'"
formed prior to ozonation. Only with unrealisticially
high doses of ozone does removal of halomethanes seem to
occur and even then the removal may be a result of vola-
tilization (9).
4. Influence of Ozonation on the Behaviour and Formation
of Halogenated Organic Compounds-
The most extensive data with regard to the influence of
ozonation on halogenated organic compounds have been
published by Stieglitz et al. (1O). These data are based
on the measurement of chlorinated organic compounds at
several German water treatment plants which use ozone
and are situated along the river Rhine. Results on the
oceuriEnce of halogenated organic compounds were obtained
by means of determinations of dissolved organic chlorine
(DOC1), non-polar dissolved organic chlorine (DOC1N) and
gaschromatographically detectable organic chlorine (GOC1),
and further by means of identification of different ha-
logenated compounds by GC-MS. A*s can be seen, 'Table 2 shows
a substantial increase in DOCl-N as well as the gaschro-
matographable organic chlorine after ozonation of a
bankfiltered water. Furthermore, like chloroform, tetra-
chloromethane, trichlorpethylene, tetrachloroethylene,
tetra- and hexachlorobutadiene have increased in measured
concentrations as a result of ozonation.
-------
- 313 -
TABLE 2 Organic chlorine within different•treatment steps
for a Rhine water utility using ozonation (8)
Sample
Rhine water
Sandbank Filtrate
Ozonated Sandbank
Filtrate
Dissolved'
Organic.
Chlorine
99
61
' 45
Chlorine
pg/i
Dissolved
. Organic
Chlorine
Nonpolar
15
4
,0
Gas chroma tographi-1-
• cally
Detectable
Organic Chlorine
9.5
4.8 '
• • .
8-3 i
On the other hand,a decrease in total DOC1 as a result of
ozonation is evident.
Another publication of Stieglitz (1O) (figure 1) gives
the changes of organic chlorine concentrations in water
after different treatment steps. This study demonstrates
a clear increase in GOC1 and "aliphatic chlorine and de-
crease in aromatic chlorine following ozonation.
-------
_ 314 -
5.
Ul
Z
a
S
o
o
X
z
S
^•
o
z
o
u
3_
2.
• GOCL
* AROMATIC Ct
C! ALIPHATIC CL
RIVEB I
WATER *
T 1 1
AFTiR I
OZON *
BANK DRINKING
FILTRATE WATER
Fig. 1
Concentration of organic
chloride after different
treatment processes (13)
According to the authors of these publications the in-
crease in certain chlorinated organic compounds following
ozonation can be explained by a breakdown of chlorinated
organic compounds with high molecular weights to mole-
cules which can be detected and identified by available . .
analytical techniques.
A study similar to those of Stieglitz et al. and Ktihn et al
has been carried dut by the authors at the National Insti-
ture for Water Supply in the Netherlands.
In this study the effect of ozonation under practical
conditions in two treatment plants was observed.
The source of water in both cases is the river Rhine. The
difference in treatment, however, is that the first plant
uses bank filtration, in the second one the water is
stored in an open reservoir with a mean residence time
of 150 days. Before ozonation break point chlorination^
coagulation and filtration are applied.
-------
315 -
The objective of the study was to determine and measure
quantitatively as far as possible the individual organic
compounds. This was done according to the general lay-
out shown in Figure 2.
GC - HC -
IDENTIFICATION
(
Fig. 2 Layout of the analyses of individual organic
compounds and bromide
Details concerning the applied analytical procedures
can be found in the annex.
In Table 3 the results of the" analyses in relation to
the first water treatment plant with bank filtration are
tabulated.
-------
- 316 -
TABLE 3 Organic compounds before and after ozonation
in a water treatment plant using Rhine river
water after bank filtration
Compound
chloroform
dibromoch lor ome thane
bis (2-chloro-isopropyl) ether
1 , 2-dichlorobenzene
1 , 4-dichlorobenzene
1 , 3-dichlorobenzene
trichlorobenzenes
trichloroethene
tetrachloroethene
benzene
. dibuthylphtahalate
i_« _...i —
Concentratic
Before ozone
:
2500
500
200
200
not present
'
30
30
300
Dn (ng/1)
After ozone
100
100
3000
500
200
200
not present
3OO
100
10
3000
A striking feature in these data is the relatively high
amount of bis(2-chloro-isopropyl)ether which is, as to
be expected, not broken down by ozonation.
Furthermore, the observed increases of tetrachloroethene
and trichloroethene should be noted. These increases have
also been found by Stieglitz.
The di-butylphtalate is believed to be an artifact, ori-
ginating from plastic tubes that are used in the treat-
ment plant.
The situation in the second water treatment plant is
shown in Table 4.
-------
- 31
TABLE 4 Organic compounds before and after ozonation
in a water treatment plant using Rhine water
after open storage and chlorination
Compound
chloroform
dich lor obromome thane
ch lor odibromome thane
dich lor oiodome thane
bromoform
tetrachloroethene
bromochloroiodomethane (?)
ch lor o toluene
hexachlorobutadiene
bis (2-chloro-isopropyl) ether
1 , 2-dichlorobenzene
1 , 4-dichlorobenzene
1 , 3-dichlorobenzene
1,2, 3-trichlorobenzene
1,2, 4-trichlorobenzene
heptanol
ethylbenzene
alcohol or ketone
alcohol or ketone
alcohol or ketone
Concentratic
Before ozone
10.000
1 3 . 000
3 . 000
-
300
300
-
30
100
500
1OO
400
200
. 100
10O
-
30
30
300
10O
Dn (ng/1)
After ozone
10.000
12.500
1 0 . 000
1 .000
3.000
1 .000
300
TO
10O
-
1OO
200
200
1OO
10O
30
10
30O
100O
300
-------
318 -
From the results of the analyses it can be ,noted ,that the
use of breakpoint chlorlnation prior to ozonation has pro-
duced much greater concentrations of halogenated organic
compounds than those found in the bank-filtered water..
As expected/ most of the halo:genated compounds are not or
only slightly removed by subsequent treatment with ozone.
Since chloroform and dichlorobromomethane did not Increase
during ozonation, it is not likely that the rather drastic
increases of chlorodibromomethane and bromoform are due to
a prolonged action of the breakpoint chlorination. There-
fore another mechanism must be involved.
What really strikes us from these results are the increased
levels of brominated and possibly iodinated haloform com-
pounds found after ozonation. Possible explanations for
this phenomenon are discussed in section 5.
The oxidation of bromide to the already mentioned hypo-
bromous acid is confirmed in Table 5.
TABLE 5 Results of the analyses of bromide
chlorination before 0
after
without C12 before O
after O
33
•H
33
33
Cl (mg/1)
234
237
177
178
Br (ug/1)
440
"
190
410
2OO
-------
- 319 -
i , . - > * ~ -1
• - . - " • 5 ^ C . f '< " " * " ' '„ _' 7 " . ' , - , ', . • * - • > > • - - - > - -
The bromide-ion concentration prior to ozonation was about
the same in both cases (4OO yg/1) and in both cases a 5O %
reduction occurred as a result of ozonation. .
5. General Considerations on the Mechanisms of Halogenation
During Ozonation
The observed increases in haloforms and other halogenated
compounds after ozonization are generally not to be con-
sidered as a result of halogenation, except for the bro-
minated and iodinated compounds. Rather it must be regar-
ded as an outgrowth of the incomplete analytical techniques
currently available. There is no evidence that increased
chlorination has occurred; rather it appears that non-ob-
served chlorinated compounds have been rendered accessible
to observation.
The increase in brominated and iodinated organic materials,
particularly the halomethanes, is to be attributed directly
to ozonation. Rook (1) has demonstrated formation of HOBr
front Br in ozonation of Rhine water and has shown an in-
creased bromine content in halomethanes formed on chlori-
nation after ozonation. Determinations in the present study
showed a decrease in Br content from O.4 mg/1 to O.2 mg/1
as a result of ozonation. Concurrently with this decrease,
the increases in CHBr^Cl from 3 to 1O yg/1 and in CHBr-,
from O.3 to 3 yg/1 in ozonation of a sample ( Table 3)
are noteworthy. Another possible explanation for these
observations, advanced by Stieglitz et al., is that ozo-
nation splits non-volatile chlorinated organic compounds
into smaller volatile fragments that retain the organic
chlorine originally present. This proposal is attractive
in a general way, but is difficult to adapt specifically
to substances like trichloroethylene and tetrachloroethy-
lene without postulating some unlikely starting materials.
-------
- 320 T-.
A similar but even .more speculative proposal jan'be ad-
vanced with regard to changes in polarity of halogenated
compounds rather than changes in molecular size, as the
properties producing incomplete measurements of organic
chlorine already in water before ozonation.
Chlorinated lower organic acids and amines are examples
of materials that might be responsible for this pheno-
menon. The increase in extractable organic material at
the expense of adsorbable dissolved organic chlorine ob-
served by Stieglitz implies an elimination of water-solu-
bilizing groups like OH, COOH and C=0. Once again, how-
ever, the increase in heavily chlorinated ethenes is not
well explained.
A different type of explanation can be visualized if it
is postulated that organic (or other) material in the
original water can complex or otherwise interact with vo-
latile chlorinated- compounds to decrease their activity
and thus make them less prone to volatilization or extrac-
tion. Then, when the complexing or restraining materials
are broken up or made ineffective by ozonation, the chlo-
rinated materials are released at full activity to give
a greater response in analysis.
The necessary initial assumption may seem improbable, but
so little is known of phenomena at the submicrogram per
litre level that it ought not to be summarily eliminated.
If it is valid, then the observed results with polychlori-
nated ethylenes follow almost automatically.
Finally, it may be noted that, if at any point in the com-
plex sequences of reactions that may follow an initial
ozonation step there occurs a nucleophilic displacement
reaction with transitory formation of carbonium ions> then
the high concentration of chloride in Rhine water provides
-------
- 321 rK V '. -*
an ".excellent .©pportunity for- nueleophilic displacement"
by chloride rather than the normal reagent. In this event
an increase in carbon-chlorine bonds would occur without
intermediate occurrence of elemental chlorine or HOC1.
As an addendum to this mechanism it may be noted that dis-
placement of the carbonbromine bond is normally much
easier than displacement of a carbon-chlorine linkage.
It is thus conceivable that, following -the formation of
tribromomethane according to the first mechanism, the •
reaction
Cl~ + CHBr3 » CHBr2Cl + Br~
occurs subsequently by nueleophilic displacement.
The possible mechanisms are once again summarized in -
Table 6.
TABLE 6 Possible mechanisms for observed increases in
halogenated compounds on ozonation
1. O3 + Br (I~) > HOBr > Br derivatives (I-deriv.)
(Note: Br(.4 mg/1 before) > Br~ (.2 mg/1 after)
2. O-, splits non-volatile chlorinated compounds into
smaller volatile chlorinated fragments
3. O-j attacks polar spots in molecules to eliminate non-
polar (adsorbable) chlorinated fragments
4. Ozonation breaks up complexing or adsorbing materials
(fats, humates) that inhibit volatilization or
extraction
_>. Occurrence of any SN1 nueleophilic processes in ozo-
nation reactions would allow Cl substitution
-------
- 322 -- ' •
6. Conclusions
From the results of this study and previous ones it is
clear that the use of ozone during the preparation of
drinking water can also result in the formation of halo-
genated compounds. Our study of the two water treatment
plants in the Netherlands shows that such a production,
particularly of brominated compounds/ is enhanced in case
chlorination precedes ozonation. The mechanism for this
phenomenon probably involves intermediate production of
bromine, but should be studied further.
A small-scale pilot plant, suitable for studying the
mechanisms of reactions of oxidants in different types
of water is being constructed at the Dordrecht facili-
ties of the N.I.W.S.
The effects of individual oxidants, including U.V.-ra-
diation as well as different combinations of them will
be evaluated by identifying the products formed and by
testing the mutagenicity of the oxidized water using dif-
ferent types of micro-biological screening tests.
-------
- 323 - .
ANNEX 1
Experimental
In order to measure halogenated compounds as completely
as possible, four different.concentration and separation
techniques (head-space analysis, Grob stripping, Junk
XAD-resin adsorption and cyclohexane extraction) were
used on portions of the same samples. Subsequent analyses
were carried out with specialized GC instruments and co-
lumns and also with GC-MS apparatus. In addition, bromide
analyses were performed according to the Fishman-Skougstad
method (11) as modified by Rook (2).
Head-space analyses were used for determination of very
volatile substances, such as chloroform and trichlorethene.
Sample portions of 0.4 1 were mixed with- 0.1 1 N~ in 5OO ml
closed vessels for 1 hour at 30 C. Head-space samples, of
10, 100 and 1000 yl were then injected into the inlet of
a TRACOR - 550 gaschromatograph with a 5O m capillary glass
column coated with UCON and with a Ni-EC detector.
Extraction with cyclohexane-ether was used principally for
estimation of chlorinated aromatic compounds and haloforms
with elimination of background interference of non-extrac-
table substances. One litre portions of sample were ex-
tracted with equal volumes of a 1:1 cyclohexane-ether mix-
ture. After the immiscible liquids had been shaken to-
gether vigorously in a separatory funnel, the liquid layers
were allowed to separate and then portions of the upper
organic layer, without concentration were injected at the
inlet of a VARIAN-180O/20OO gaschromatograph having a ca-
pillary column with 0V 1 equipped with a Grob-splitter with
double parallel detectors, 1 electron capture and flame
ionization (12).
-------
- 324.--
Grob stripping (1.3, 14) was carried out on 5 litre por-
tions of sample that were maintained at 30°C in a closed-
loop apparatus through which N~ was recirculated for 2
hours at 2 litres per minute.
Vapt>rized organic compounds were collected on 1O - 2O yg
of carbon powder. After stripping had been completed, ad-
sorbed materials were eluted from the carbon with three
successive 8 ml.portions of CS_. .The eluates were com-
bined to give a final total volume of about 20 ,yl, which
was stored at -40 C until analyzed.
For analysis, a 1 jil portion of the eluate was. injected
into the part of a CARLO-ERBA 2101 gaschromatograph con-
taining a 50 m glass capillary column coated with UV101;
it was equipped with a flame-ionization detector.
For a more complete survey of higher-boiling and more po-
lar compounds, concentration on XAD resins was performed
according to Junk (15, 16). Fifty litres of a sample
were passed upflow through 3O ml beds of 1:1 mixed XAD-4
and XAD-8 resin "(Serva, Heidelberg) held in a 15 mm glass
tube. The resins had previously been cleaned by repeated
batch extraction with ether, ethyl acetate and ethanol;
they were stored under ethanol. After the adsorption step
was finished, the resins were extracted with successive
portions of ether to a total volume of about 3O ml, which
was subsequently reduced to 5OO jil by evaporation at 0 C
into a pure N~ stream. One pi portions of this concentrate
were analyzed with the same gaschromatographic equipment
described in the section on the Grob method.
GC-MS analyses of the Grob and Junk concentrates were per-
formed with a VARIAN gaschromatograph containing a 5O m,
UV-101 coated, glass capillary column. The chromatograph
-------
-. 325 -
was programmed"for 5'minutes at O ; -then 3 minutes at
3O , followed by continuous increase at 4 per minute
from 3O C to 18O C. The exit was connected to a Finnigan
quadrupole mass-spectrometer with one second scar-time.
fhis was coupled to a W.D.V. data system. Spectra were
compared with the "Eight Peak Index" or, by telephone
hook-upfwith the U.S. E.P.A. data bank.
Bromide determinations were based on the catalysis by
Br , or iodide oxidation by permanganate (11). After
5 minutes of reaction, the-I2 formed was extracted with
tetrachloroethane and measured spectrophotometrically.
(1) ROOK, J.J.
J. AWWA 68 (1976) , 168
(2) ROOK, J.J.
Het ontstaan van trxhalomethanes bij behandeling van
drinkwater met chloor
Dissertation L.H. Wageningen (1978)
(3) MONTIEL, D.
Les halomethanes dans 1'eau. Mecanismes de formation,
evolution, elimination
Dissertation 1'Universite P. et M. Curie, Paris (1978)
(4) CHIAN, R.S.K.
Identification of End Products from Ozonisation of
Compounds Commonly Found in -Water
AWWA, Research Foundation, Water Reuse Report 8
(1976) ,13
(5) - '
EinfluB der Ozonisierung auf organische Wasser-
inhaltsstoffe
Report F.J. Borrus & Cie., Boncourt (1976)
(6) SYMONS, J.M. et al.. ' -'
Ozone, chlorine, dioxide and chloramines as alternatives
to chlorine for disinfection of drinking water
E.P.A. Report (1977) "
-------
- 326 -
(7) EBERHARDT, M. et'al.
gwf-Wasser/Abwasser 116 (1975), 6, 245-247
(8) KOHN, W. et al.
J. AWWA 70 (1978), 326
(9) SYMONS, J.M.
Interim Treatment Guide for the Control of Chloroform
and Other Trihalomethanes in Drinking Water
E.P.A. Report (1976)
(10) STIEGLITZ, L. et al. »
Vom Wasser 47 (1977), 347
(11) FISHMAN, M.J., SKOUGSTAD, M.W.
Anal. Chem. 35 (1963), 146
(12) PIET, G.J. et al.
A fast quantitative analysis of a wide variety of
halogenated compounds in surface-, drinking- and
groundwater
Internat. Symp. on the Analysis of Hydrocarbons and
Halogenated Hydrocarbons in the Aquatic Environment,
Toronto, May 1978 ' ' • •
(13) GROB, K. et al.
J. of Chroin 106 (1 975) , 299-315
(14) GROB, K. et al.
J. of Chroin. 117 (1976) , 285
(15) JUNK, G.A. et al.
J. of Chrom. 99 (1974), 745
(16) JUNK, G.A. et al.
Z. Anal. Chem. 282 (1976), 331
-------
-. 3.27 -
NOTE ON THE HALOFORM FORMATION POTENTIAL OF PRE-OZONIZED
WATER ' -' . -
J. Hoigne
The papers given at the conference about the effect of a
preliminary ozonization on the subsequent chlorination
processes show clearly that the action of ozone can depend on
a wide variety of reaction and water parameters. This corres-
ponds fully to what we have come to expect in the light of the
present-day knowledge of the kinetics of ozonization processes.
We should like to propose, the following hypothesis for dis-
cussion: depending on the performance of the ozonization, the
water constituents are predominantly oxidized by direct
ozonolysis or by secondarily formed OH" radicals [1]. In the
example of the oxidation of an aromatic ring system the two
types of reaction lead to different intermediate products
(Fig. 1). Ring cleavage is mainly expected in ozonolysis,
with the formation of dicarboxylic acids, such as oxalic acid,
in subsequent steps. In these cases enrichment of trihaloform
precursors in the water need hardly be reckoned with. However,
the proportion of ozone that decomposes in water to OH"
radicals oxidizes aromatic ring systems via hydroxylation to
hydroxycyclohexadienyl radicals, which,give hydroxy-substituted
benzenes (phenols, resorcinols etc.). Compounds of this kind
are known to be particularly reactive trihaloform precursors.
If the ozone reaches the reaction centre before it decomposes,
these compounds are rapidly oxidized further by ozone. In
the opposite case, it is to be expected that compounds of
this type will be enriched in the water and on subsequent
chlorination will lead to an increased formation of trihalo-
forms.
If methyl-substituted olefins are present in the humic
substances, it can be expected that they will be oxidized to
-------
- 328 -
Formation and Degradation of
TRiHALOMETHANE PRECURSORS
HA
f-^sfOH
IP
CH3
— *~ THMP — - — -
' krf? ' fast
+ OH* * ^^.--^r^i ^"^
*°3 » m *°3 s
3>C= o s'ow '•
i^H*"*~ +OH**
non THMP
Cr\
— U
1
r — n
XOH
,OH
-c=o
Fig. 1 Dependence of the product formation on the perfor-
mance of the ozonization process [2]
methyl ketones by direct ozonolysis [1]. However, many of
the methyl keinnes are extremely stable to ozone [3] . They can
also be enriched in the water until they are oxidized further
by secondary OH* radicals. This means that the formation
and degradation of methyl ketone-like haloform precursors
proceeds in accordance with laws- quite different from those
governing the formation of phenol- or resorcinol-like tri-
haloform precursors [2]. ...
Whether the ozonization-initiated oxidations proceed via
direct ozonolysis or via previously formed hydroxyl radicals
depends not only on the nature of the substrate molecules
but also on the nature of the ozonization process. Fig. 2
shows that the time during which an ozone molecule is avail-
able for direct reaction is limited by the rate at which it
breaks down into radicals. This rate is known to increase
with rising pH. However, since the. decomposition of ozone
proceeds in addition by a chain reaction in which radicals
-------
- 329 -
Oo-degassing
D-Reaction
**
k"
k
"\
L
/ T-OI TIVI
Jf
^
0
R-Reaction
Fig. 2
Dependence of the type of oxidation on the perfor
mance of the ozonization process [3]
M: Oxidizable water constituents
S: Scavengers
R*: Free radicals
function as chain carriers-, many "water constituents are also
decisive for the decomposition" rate of the ozone [4]. Thus,
certain organic constituents (e.g. benzene in the 0.1 mg/1
range) lead to an accelerated - ozone decomposition. Other
constituents, such as bicarbonate ions or aliphatic alcohols,
inhibit the chain reaction by abstracting radicals from the
process [4] . This gives rise to the situation that the ratio
of the direct ozonolysis to the radical OH" oxidation is
influenced by a wide range of constituents present in the
water. On the basis of the chain reaction accelerating the
ozone decomposition, it can also be understood why the kinds
of reaction can depend on the istantaneous local ozone concen-
tration, i.e. also on the way in which the ozone is introduced,
If the results of ozonization processes are to be comparable,
an ozonization must be described very carefully. It there-
fore seems to us to be very important that the greatest atten-
-------
- 33O
tion be focussed on a careful establishment of 'the process
arrangements and characterization of the water. If this is
not done, the experience cannot be generalized.or extended to
other situations.
(1) HOIGNE, J. , BABER, H.
Ozonbedarf und Oxidationskonkurrierwert verschieclener
Wassertypen beziiglich der Oxidation von Spurenstoffen
(this publication)
(2) HOIGNE, J.
Discussion of paper by M. Dore on "Influence of oxidizing
treatment on the formation and the degradation of
haloform reaction precursors"
Prog. Wat. Tech. 1_O (1978), discussion part, in press
(3) HOIGWE, J., BADER, R.
Kinetik und Seiektivitat der Osonung organischer Stoffe
in Trinkwasser
Wasser Berlin '77, Tagung der Fachgruppe Wasserchemie
in der GDCh, Colloquium Verlag Berlin (1978), 261-276
(4) HOIGNE, J., BADER, H.
Beeinflus;sung der Oxidationswirkung von Ozon und
OH-Radikalen durch Carbonat
Vom Wasser 48 (1977), 283-304
-------
- 331 -
THE CONDITIONS OF OZONIZATION
J.P. Legeron
The number of studies and investigations on ozone and the
number of cases of the practical application of this gas
become greater every day.
However, the results obtained seem to be very variable,which
is perhaps in many cases due to omissions or errors, but
perhaps also to the absence of a common language.
In actual fact the aim of ozonization and the conditions of
the use of ozone could not be interlinked more closely. Ozone
is a chemical substance and as such reacts with other species
present in an aqueous medium: we can therefore speak of oxi-
dation kinetics and thus of conditions for maximum reaction
rates.
The temperature and sometimes also the pH of the water are
difficult to change, but the ozonization parameters are really
easily modified:
This is the situation, for example, with the ozone concen-
tration in the carrier gas, with the dose used for the treat-
ment', with the contact time, with the residual ozone concen-
tration in the gaseous and liquid phases, and with the life-
time of -this residual ozone in water.
The ozone can be added -in one portion or in several portions
and ;at one or more points in the treatment line.
The "contact can be intermittent or continuous. Many studies,
for "example that by Monsieur Chedal at the Berliner Wasser-
-------
332 -
kongress in 1977» have indicated the' influence 'of these con-
ditions.
Let us first of all disregard the case of disinfection with
ozone, which is not-the object of this symposium. Suffice
it to say that in this case the maintenance of a certain
quantity of dissolved residual ozone for some minutes makes
it possible to enhance the organoleptic properties of the
water, a fact that is attested by the readiness of ozone to
react with some substances still present in the medium.
While these reactions proceed relatively slowly, this is not
the case during the oxidation of very many organic and inor-
ganic compounds. Depending on the type of the water, the
ozonization serves a different purpose at the start and some-
times in the middle of the treatment chain.
The case in question may be, for example, the oxidation of
iron and manganese, and here the contact times are short and
the doses depend on pollution of the water.
In other cases we may -be concerned with an improvement of the
flocculation; for this both the contact time and the dose must
normally be small.
Finally, we may be aiming at improving the biological degra-
dation, in this case the dose can be made higher, to introduce
oxygen into the water and to improve the bio-degradability of
the substances present.
Each case has its special features and for this reason rele-
vant trials must be performed either in the laboratory or,
better still, in a pilot plant.
These trials must be as comprehensive as possible before any
qualitative or quantitative conclusions may be drawn.
-------
- 333 - - • •
Without going into the details of the process, which has
already been the subject of several contributions, I should
like to conclude with two specific examples.
The first relates to the oxidation of manganese in a
French river water;
at constant contact time, quadruplication of the ozone
concentration before it is added makes it possible to
reduce the treatment dose and therefore the ozone
production by almost 50%.
- The second and last example relates to the oxidation of
an iron/silica compound in a ground water of African
origin:
.splitting of this complex and the oxidation of the iron
thus released cannot be achieved in a single ozonization
step even with a high dose and a long contact time; the
solution was found in several consecutive small additions
of ozone alternating with aeration phases.
This confirms that the results can be positive or
negative, depending on the ozonization conditions.
-------
- 334 - • *; •
DESCRIPTION OF•REACTIONS;BY GROUP•PARAMETERS
H. Lienhard and H. sontheimer
In this report I should like to discuss in greater detail a
problem that keeps recurring in the study of ozone processes,
namely how the changes caused by ozonization can be detected
and described.
The possibilities existing today for this purpose should be
discussed in the light of the experimental results obtained
in studies on the optimal ozone input, on which
Prof. Sontheimer spoke in his lecture on Monday. Briefly,
such studies are aimed at establishing whether the ozone
should be added in stages, e.g. in a sufficiently large reaction
vessel by fine-bubble aeration, or whether it is more expedient,
as
possibly with a rotor, and then with an initially high concen-
tration to allow the reaction to proceed slowly.
To provide practical data from the results of these studies,
humic acids were used as model substances in the experiments.
This gave rise to the problem that the chemical structure of
humic acids is largely unknown and that during ozonization they
form products not directly accessible to analysis. This means
that the reaction of the model humic acid with ozone cannot be
followed directly, so that the products formed must be charac-
terized by measuring some accessible parameters.
To this end we used the 6 parameters shown in the left-hand
column of Table 1.
We shall now describe how conclusions can be drawn, from the
results obtained with these parameters, about the nature of
the reaction products and about their effect on the treatment.
-------
TABLE 1 Description of reactions by group parameters
Parameter
Measurable effect with
stepwise addition of
ozone as opposed to a
single addition
Interpretation of the
effect
Spectral absorption co-
efficient at 254 nm
Decrease
Degradation of the
double bond system
DOC
No difference
No increased CO,
formation
i Flocculation effectiveness Improved relative
: flocculation activity
More polar molecules
w
w
Molecular weight distri-
bution
Displacement of the
mean molecular weight
towards smaller values
Smaller molecules
Adsorption on
Displacement of the
isotherms to the-left
More polar groups on
the molecule >
Haloform formation
potential
Less trihalomethanes
Degradation of the
electrophilic centres,
fewer double bonds
-------
- 336 -
The middle column shows the effect measured for the parameter
on the left, obtained with fractional addition of the ozone.
The right-hand column lists the conclusions that can be drawn
from this about the change in the organic substances.
From the observed decrease in the spectral absorption co-
efficients at 254 nm we concluded an intensified degradation
of the double bond system with intermittent ozone input. The
fact that the DOC is the same for both modes of ozone input
means that only a different change in the humic substances
occurs, which also happens in practice, since the effective-
ness of flocculation is enhanced by stepwise ozone addition,
as shown in Fig. 1.
9r
O OP
QJ
D-7
U) 4J '
c
QJ Ol
si -H 6
JJ U
•H
c «w ,.
••H »M 5
at
at o
in u ,
m 4
T3 D
M
W
G
o Ruhr humic acid
a Lake Constance
humic acid
0.4 0.6 0.8 1.0
1.4 1.6 1.8 2.0
V 2.4 2.6 2.8
mg O^/mg acid
Fig. 1 Relative improvement of flocculation effectiveness
by stepwise ozone addition for two humic acids
-------
- 337,-
This figure illustrates, the relative improvement of the flocc-
ulation effectiveness due to the intermittent ozone input. It,
can be seen that e.g. for the Ruhr humic acids after stepwise
ozonization by flocculation alone the spectral absorption co-
efficient is reduced by up ,to 7.5% more than in the case of
a single addition. The same applies to the humic acid from
Lake Constance.
To understand this effect the molecular weight distribution
must be measured, as can be seen in Table "I - The distribution
is displaced towards smaller molecular weights. The absorp-
tion experiments additionally performed on CaCO^ confirm an
increase in polar groups compared to the case of the single
addition of ozone. The haloform formation potential is also
reduced (Fig. 2).
° Single ozone addition
f; Stepwise ozone addition
Ruhr humic acid
0 0,2 0,4 0,6 up Ip 1,1 V1 *£ V» V V
.rag O,/mg acid
Pig. 2 Dependence of .the haloform formation potential on
the amounts of ozone for-—two different kinds of
ozone addition
-------
- 338 -
The figure shows the amounts of trihalomethanes formed, during
the chlorination as a function of the ozone input. Prom
the course of the curves it can be seen that stepwise ozone
addition reduces the haloforni formation potential more
strongly, and this indicates that fewer electrophilic centres
and so fewer double bonds are available to the electrophilic
attack of the chlorine when this type of ozonization is used.
If the information yielded by each individual parameter
listed in the right-hand column of Table 1 is compared, it can
be seen that the data provided by each parameter are supported
by the other results.
For example, the results of the flocculation experiments can
be explained both by the change in the molecular weight dis-
tribution and by the increase in the concentration of polar
groups. The two effects run in opposite directions and lead
to an explanation of the different behaviour in the two modes
of ozone addition.
Without going into further detail of all the interesting
relationships, from the results presented it can be concluded
that a combination of a number of experimental parameters can
lead to a better appreciation even of small effects and to a
deeper understanding of the processes taking place during
ozonization.
-------
- 339 -
PRODUCTION OF OZONE FROM OXYGEN
G. Uhlig
The problem appears trivial: a molecule consisting only of
oxygen atoms can likewise only be produced from oxygen.
The production of ozone from air succeeds only because, in
spite of environmental loading, even in regions of accumul-
ation, air always contains 21 vol-% of oxygen.
However, it is also this oxygen concentration that finally
limits the yield and the efficiency of even the best ozone
generators. This applies above all when not only the amount
of ozone produced but also the ozone concentration achieved
is the main criterion, because the dissolution of ozone in
water is subject to the Henry-Dalton law, according to
which the solubility of ozone is proportional to its par-
tial pressure, and thus to its concentration, in the gas
phase.
According to the laws of reactions kinetics, there is no
doubt that the yield and efficiency of-the production of
any material increase with the increasing concentration of
the starting components. Some measurements, here cited only
as examples, for series-connected ozone generators, dimen-
sioned for the production, of; -ozone;.- from air illustrate this
point better than the corresponding reactions and equations:
' .•*•„'•** ^
Fig. 1 shows, on the basis of Masschelein's measurements
(1), the increase in the hourly 'production of ozone
with increasing ' oxygen •"•• concentration at a constant
gas throughput for various' power levels of an ozone gener-
ator. At the same time, as can be seen in Fig. 2, the
energy expenditure on the production of 1 g of ozone is
greatly reduced.
-------
kg Oi/h
20
Ozonour 2
Vol %02
40
60
80
Fig.1 03 production/h in dependence on
the C>2 concentration.
(after W. Masschelein, T.S.M.
L'EAO 71 (1976) 385)
Ozoneur = ozone generator
Wh/g03
15
10
70kWh
50 kWh
Ozoneur 2
20
Vol % 02
60
80
Fig.2 Specific energy requirement
of 03 production in dependence
on the 62 concentration.
(after W. Masschelein, T.S.M.
L'EAU 71 (1976) 385)
to
*»
O
-------
- 341 .- .;•
52-
50-
48-
46 -
44 •
42 •
40 •
38
36
34
32 -
30-
28-
26-
24-
22-4
20-
18-
16-
U-
12 -
100 Vol. V.02
80 VcI.*A02
50 Vol.7.C
Fig. 3
03 concentration produced in
dependence on the gas flow at
various Oz concentrations,
(after M. Bredtmann, Wasser/
Luft und Betrieb 11(1974} 605)
0,2 0,3 0,4
Higher concentrations of ozone are generally achieved by
reducing the gas throughput of the ozone generator. Fig. 3
shows a diagram taken from Bredtmann (2), in which
the concentrations of ozone obtained with a series-
connected tubular ozone generator during the change-over
from air to pure oxygen is demonstrated.
Here too, as can be seen from Fig. 4, the specific energy
expenditure for the production of increasing ozone concen-
trations on going over from air to pure oxygen is essen-
tially reduced.
-------
342 -
V.'h/gOj
20
16-
U-
12-
10-
9-
Fig. 4
50 Vol. %02
BO Vol.%02
100 Vol. % 02
g Ch/m3
15
20
25
30
35
50
55
Specific energy requirement in dependence on the €3
concentration produced at various O2 concentrations.
(after M. Bredtmann, Wasser, Luft und Betrieb 11(1974)
605}
Even when the starting gas contains only 5O to 9O vol-%
of oxygen it is an additional advantage for the dissolution
of ozone in water that the correspondingly smaller propor-
tion of nitrogen still permits a gasification of the water
under pressure, which in the case of ozone production from
air would lead to supersaturation of the water with nitrogen,
However, all these advantages do not cover the costs of the
oxygen consumed if this is not returned to the production
after the ozone has been washed out. Therefore, if the
process is to be economically viable, gas circulation must
be installed. On account of the improved ozone dissolution,
however, only a partial water stream needs to be gasified
with ozone for this purpose, and expensive structures and
installations for the gasification of the whole amount of
raw water become unnecessary.
-------
- 343 -
By means of the production of high ozone concentrations in
the gas phase (5O-9O ozone/m is nowadays a realistic figure)
and by the use of pressure in the gasification of the water
in space—saving scrubbers or columns, ozone concentrations
of 2O-80 g/m of water can be reached. Accordingly, the
dissolution of 2-3 g of ozone per m of raw water.still
requires only the gasification of a partial water stream
amounting to 3-lO% of the total water. The highly concen-
trated partial stream is then mixed with the main stream in
such a way as to ensure a good distribution and a sufficient
reaction time.
The apparatus and the'physico-chemical principles of this
process have already teen given in detail by Axt (3) 1958
in his dissertation "An indirect process for the ozona-
tion of water". This already included calculations for
the design of ozone-introduction columns, calculations of
the energy yield during the production of ozone from oxygen,
and a description of a specialized ozone generator to be
operated at 6-7 bar that provides for gas circulation at
a constant excess pressure. Since such ozone generators
are not produced commercially, the gas circulation must in
practice be operated with two pressure levels: an optimal
pressure for the dissolution of ozone and an optimal pressure
for the production of ozone, which in the majority of cases
is restricted to 1.6 bar with today's ozone generators.
A large-scale realization of the process took place 8,
years later for the first time at the Duisburg AG municipal
works with an ozone plant designed for a throughput of 250O .m
of raw water per hour. This plant was commissioned in 1966
in the Wittlaer waterworks for the treatment of Rhine water
filtrate. Simon and Scheidtmann (4) 1968 have reported in
detail on the extensive preliminary trials on adaptation of
the process and on the technical data of the main plant.
-------
- 344 -* '
The process scheme of this plant is given in Fig. 5: fresh
oxygen is led out of a 5OOO m liquid oxygen tank (bottom
left in the figure) via an evaporator warmed by air and
directed to the gas circulation before the production of
ozone. The ozone production at 1.O5 bar in 4 tubular ozone
generators (only one unit is'shown'in each case) supplies
an ozone concentration of 4O-6O g/irt3. The pressure is then
raised to 1.8 bar and the gas-mixture' 2 is conveyed to the
ozone scrubbers, made in the form'of'packed columns, in
Which the partial stream of: water - "about 1O% of the whole
water treated - trickling in countercurrent through Raschig
rings undergoes gasification.
The unconverted mixture leaving the scrubbers at the top
is directed to a drying plant with a cooled condensate
separator and returned for use in the production of ozone.
Ozone scrubber-
Mixing and reaction vessel
Filter plant
Dryxng plant
Multilayer
Liauio~Qxyqtn tank
Sa fety chlor™ De-acidt fl-
irtation cation
Fig, 5 Diagram of the ozone plant at Wittlaer Waterworks III
-------
- 345 -
The partial stream of water, containing 1O-25 g of ozone/m
and described in the diagram, as "ozone-rich water" is
>
added to the raw water in,two reaction tanks after the
pressure has been raised to-8 bar. The ozonized raw water
coming out at the bottom through a concentric drainage
funnel passes 2O two-stage filters with a total filter area
2
of 2O6 m per stage, consist-ing of a multilayer filtration
for the removal of the flocculated.ozone-oxidized products
and an active carbon stage, connecting safety-chlorination
with chlorine dioxide and deacidification with a solution of
caustic soda. • •
The external appearance of the Duisburg ozone plant is
illustrated in
Fig. 6, which shows the two scrubber columns for the disso-
lution of ozone and the pumps for increasirg the pressure,
Fig. 7, which shows the control panel, the ozone generators
(right), and the reaction tanks (left) in the background,
and
Fig. 8, which shows an external view of the whole plant.
-------
-•: Vff :«wti=^-r -gff?-'.--'-'-• i^** = m¥T'- --1K ' *i«ri!"!!BT.-pJf«?^-* •"" • ,"
•;|
•«,• sHfi^at^' ~ • *
= , 1 ^ 5 ^^Ugi 'i^'r^-tTffil • Sitfis^S^'^r-«?':lr:LJ"-.'«^ ^. .' .;" ; \ •'
6 Ozone scrubber column and pressure
pumps of the ozone plant
I
U)
-------
. ;i: v
.-f >c
Fj-g> 7 Control panel, ozone generators »**-*
and reaction tanks of ozone olant"" *> *
-------
RjV-' ;-*:, i<-yHi*:- '*••'• 0.'*i-.--:,'f If-',/ i"-.. •:i'Vr' •-.- * ift..^ ,U
^SMn^
•;>-^ T^s', T-Tfc.^ »£..;• ;-••"•• t '
-" • •-•* rK-^t '.?;•• ;."> ., .?:. .P. .:
•fe^iSiill-*1
-., ^•riS-^'^r'-"---^.'^
I
w
00
i
Fig. 8 External view of the ozone plant
-------
- 349 -
Control of the ozone production only by the tension in
the ozone generators at constant gas throughput is not
optimal, since the ozone concentration is then diluted
again.
However, the control also of the amount of gas in circul-
ation, with the ozone concentration regulated to a constant
value, already envisaged in this plant, has not been ap-
proved in the proposed form and is the object of further
improvements, as is the design of the columns.
It should be emphasized that in this process the raw water
from the feed pumps of the wells onwards undergoes the
whole treatment already under distribution pressure, and
the pressure need not be reduced for the ozone treatment.
In the selection of the pressure for operating the columns
it must be remembered that the Henry-Dalton law applies not
only to dissolution of the gases — this concerns above all
oxygen in addition to ozone - but also to the release of
gases dissolved in the water. The main component dissolved
is still nitrogen, corresponding to its partial pressure in
the air. However,, in a pure oxygen/ozone mixture the par-
tial pressure of nitrogen is zero, and thus does not corres-
pond to the distribution equilibrium.Consequently, a partial
pressure of nitrogen corresponding to the nitrogen content
of the water is rapidly set up in the gas circulation by
outgassing. The concentration corresponding to this par-
tial pressure depends on the partial pressure of the oxygen
and therefore on the selected total column pressure, which
in turn determines the dissolution of oxygen in the partial
stream of water.
These relationships are presented in Fig. 9 for nitrogen-
saturated water in an easily followed diagram due to
Albrecht (5): when the pressure is raised to 8 bar the
-------
- 350.r
f • % * 5 . i s t *- s f. - < L ,, - v * •, : , « ... . - - ~
nitrogen content in the gas circulation can be reduced to
about 1O%. As shown by measurements carried out by Rosen
(6), and also by Cromwell and Manley (7) , a smaller con-
tent of nitrogen scarcely improves the ozone production
efficiency, and furthermore, increasing the pressure also
reduces the column layout and dimensions. According to
Greiner and Grxinbein (8,9), at 5-7 bar 8-10 theoretical
plates are necessary for almost quantitative washing out of
the ozone.
Oj Nj
mg/l Vol.%
1000
900
000 -
700-
500
400-
300-
200-
100-
100 •
90-
80-
70-
CO-
50-
40-
30-
20-
10-
For ",-saturat-ed water
at 10° C"
j" enrichment
bar
1 2 3 4 5 6 7 8 9 10 It 12 13 «
Fig, 9 Distribution equilibrium of nitrogen (vol-%) in
oxygen and of oxygen (mg/l) in water
(after E. Albrecht, Dechema - Monogr. 75 (1974) 343)
-------
- 351"-
The pressure elevation is,however restricted by the in-
creasing oxygen dissolution, the limit being determined
by the oxygen content and the ozone consumption of the
raw water and also by the oxygen demand of the after-
connected treatment plant. To avoid this situation by
a subsequent aeration of the water to drive out the excess
oxygen necessitates considerable additional expenditure
and is not practicable for a water treatment under distri-
bution pressure in a closed system.
For open plants, on the other hand, even a slight enrich-
ment of the aerial oxygen, without using gas circulation
but with the pressure produced by the greatest possible
depth of immersion of the gasification, can have certain
advantages, as reported by Masschelein (1) and by Rosen
(6) .
For oxygen enrichment of the air Rosen (6) recommends a
"pressure swing oxygen enrichment" with the use of a
molecular sieve in 2 towers which are run alternately
under different pressures.
Summary
1) With the same energy consumption of the ozone
generators, at least twice the amount of ozone per
hour can be obtained, in at least twice as high
concentration, when the ozone is produced from
oxygen and not from air.
2) Corresponding to this increase in concentration,
the ozone dissolution also increases for all types
of gasification of the water, and on an additional
application of pressure in suitable columns the
dissolution is higher by a- factor of 10-30 than in
the case of open gasification of the water with
ozone produced -from air.
-------
— 352 -
3) It is thereby possible to limit the gasification
with ozone to a partial stream of water amounting
to 3-lO% of the raw water to be treated.
4) Since the oxygen is circulated none of the remain-
ing ozone need be wasted. The ozone is almost
completely dissolved or goes to the return gas.
5) At most 1/1O of the amount of gas that must be
transported and dried in the case of ozone produc-
tion from air is conveyed to the gas circulation in
the case of production from oxygen. This reduces the
costs of gas drying and transportation, together
with the energy balance for ozone production and
costs of cooling the ozone generators.
6) In the Duisburg process of ozonization of the drink-
ing water, which is based in its concept on Axt's
fundamental work (3), these advantages were already
realized 12 years ago corresponding to the state of
technology at that time. So far it has been the only
conceivable process for ozone treatment of the water
under distribution pressure in a closed system, and
work on its improvement is continuing.
-------
-. 353> -
(1) MASSCHELEIH, W. .
Perspectives de 1'ozonation de l'e-au au depart d'air
enrichi en oxygene
T.S.M.L'EAU 21 (1976), 385-399
(2) BREDTMANN, M.
Bildung von Ozon und Konseguenzen fur die Konstruktion
von Ozonerzeugern •
Wasser, Luft und Betrieb _1_8 (1974) , 6O5-609
(3) AXT, G,
Uber ein indirektes Verfahren"zur Wasserozonisierung
ur.d dessen apparative und physikalisch-chemische Grund-
lagen
Dissertation Universitat. Karlsruhe (1958)
(4) SIMON, M., SCHEIDTMANN, H.
Die neue Ozonanlage der Stadtwerke Duisburg.
gwf-Wasser/Abwasser 109 (1968), 877-882
(5) ALBRECHT, E.
Die Anwendung von reinem Sauerstoff fur die Wasser-
behandlung
Dechema-Monogr. 75 (1974), 343-356 .
(6) ROSEN, H.M,
Ozone Generation and its Economical Application in
Wastewater Treatment
Water and Sewage Works 119 (1972), 114-12O
(7) CROMWELL, W.E., MANLEY, T.C.
Effect of Gaseous Diluents on Energy Yield of Ozone
Generation from Oxygene • .
Ozone Chemistry and Technology, Washington (1972),
304-312
(8) GREINER, G., GRtlNBEIN, W.
Gewinnung von ProzeBwasser aus Abwasser
Dechema-Monogr.TJS (1974), 399-406
(9) GRUNBEIN, W,
Ozonisierung von Abwasser
Chemie-Ingenieur-Technik 46 (1974), 339
-------
- -354 - -
MEASUREMENT OF THE OZONE-DEMAND
C. Gomella
A - INTRODUCTION
The doses of ozone applied to pre—treated water, i.e.
introduced into the ozonization reactor, which appear
in the bibliography are difficult to compare with one
another, and thoseused in laboratory trials cannot as
a rule be extended to industrial practice.
The doses of ozone giving the same end result can vary
considerably as a function of the experimental conditions
and the experimental or industrial apparatus used. This
is due to the fact that the ozone can be broken down as
follows ;•->••
T -= D + r + A + p
where T is the dose of ozone introduced into the reactor,
D is the amount of ozone actually used up in the
; ; • • •• • '-
various reactions (the ozone-demand),
r is the residual free ozone maintained in the
water within the reactor (entrained by the
water issuing from the reactor),
p is the ozone lost by entrainment with the exhausted
ozonized air leaving the reactor,
and A is the ozone consumed through auto-decomposition.
A technical and economical analysis of an ozonization
process, such as the application of a laboratory or
pilot result, cannot be undertaken without a reasonably
precise knowledge of the elements making up the treat-
ment dose T. The problem is a complicated one, for
-------
• ' - 355 -
these elements are not all independent of one another;
in particular: „ . •
D is a function of the '-quality.,of the water, the time of
the treatment, the value of the O^ residue, and of the,
concentration of the ozonized -air.
r is imposed by the investigator, as a function of the
desired result and also as, a function of its effect on
D and A. ' . '•- . ,
A is dependent on .quality of the water, the duration of
treatment, r, and the concentration of the ozonized air.
p is dependent on the internal-structure of the reactor
and on the concentration of the ozonized air.
The present communication outlines a method used in Fran6e
for fifteen years by the author and his associates,
developed specially to surmount problems of interpretation,
From a pragmatic point of view, this method has proved
very effective and has enabled experimental results to
be transposed to industrial conditions with a good degree
of precision. . „
B. PRINCIPLE '
When the amount of ozone actually introduced into the
water has been determined, D and A are calculated by
measuring r ... at different-times.
Before the calculation can be made, a hypothesis about
the form of the auto-decomposition law must be formu-
lated. The many studies mentioned in the bibliography
give various formulations, notably 2nd or 3/2-power
-------
- 356
laws. In practice, it has been shown that a law of the
form:
In ~ = a(t - tQ) a < 0
where In is the natural logarithm,
r is the residue at time t ,,
r is the residue at time t,
and a is the auto-decomposition coefficient, is quite
sufficient in current practice in the case of slightly
polluted water with the usual pH (6 to 8) . Without wish-
ing to give a rigorous scientific meaning to this law, it
appears that its application in the given case enables
one to arrive at predictions and calculations of industrial
installations .
It should also be noted that the method is not specifi-
cally connected with this, form of the law, and that it
is sufficient to replace the chosen law by one determined
experimentally and better adapted to the case in question:
r = f (a,t), to be able to apply the method in its prin-
ciple by simply modifying the mode of the calculation.
C. THE DIFFERENT STAGES OF, THE METHOD
A quantity of ozonized air of a known concentration and
a quantity of water are introduced into the same flask,
after which they are stirred vigorously and the residual
ozone is measured at successive moments in time.
Two measurements are sufficient to determine the auto-
decomposition coefficient, that is, r, at time t, and r?
at time t0 :
2 r
in = at
{t is taken to be the time or\- :n)
o
-------
- 357 -'
.r
In — = at,,
r2
whence: ,r~
In ^±
a = r
and r = r, e 1 • •
o "1
The ozone demand D can then be determined, provided 'that
the initial concentration of the ozone introduced into
the water (C ) is known: ,
o
D = Co ->ro- ' '
C is easily calculated from the respective volumes v and
O , ;
V of the ozonized air and the treated water, from the
'ozonized air concentration, and" from the distribution co-
efficient S (Henry's law) of the ozone between the air"
and the water at the temperature of the experiment, which
is conducted under atmospheric pressure:
C = c
o
i + v • •
If t is the time of the ozone "treatment in the reactor, '
and r is the residue determined by the investigator, then
the quantity of ozone to be introduced into the water will
be (in the case where the amount of residue is kept con-
stant in the industrial reactor):
D+r+rat=D+r (1+ at)
The treatment dose will depend on the loss of ozone en-
trained by the ozonized air leaving the reactor. It de-
pends essentially on the form of the reactor, the time
of contact of the ozonized air and the water, and the
-------
- 358 -C~- ••
ozone residue. This is why i-t is highly desirable to
conduct the laboratory experiments with a residue similar
to that applied in practice. The loss p can be esti-
mated, to be in the region of lo to 3O%, and under
these conditions:
D * r (1 + at) 4 min.
-------
- 359
•APPENDIX
DETERMINATION OP THE CHARACTERISTICS OP WATER
IN RELATION TO OZONE' AND OF THE INDUSTRIAL ".
OZONIZATION DOSE • • ' : .
1. PRINCIPLE
The aim of the operation is to determine:
- the instantaneous chemical ozone demand;
- the specific auto-decomposition coefficient; '.•
- the industrial ozonization dose.
To this end a known quantity of ozonized air is brought
into intimate contact with a fixed volume of the water
under examination, and the residual ozone dissolved in
this water after contact times of 1 .min and 4 min. is
measured.
2. DESCRIPTION OF THE EQUIPMENT USED '
2.1. TRAILIGAZ LABO 66 laboratory ozonization unit
2.2. Volumetric flask • ••'• •
The ozonization reaction takes place in .a 6OO ml volumetric
flask with a graduation at 50O ml.
A ground-glass stopper fitted'with a tap enables the flask
to be completely filled with the water being examined.
A side tap, placed on the neck of the flask below the
5OO ml mark, enables the contents of the flask to be
drained off down to 50O ml.
-------
- 360 -?
Fig. 1 Ozonization unit - -. . ,
1 = wattmeter; 2 = flow-meter; 3 = manometer;
4 = dry air inlet; 5 = sampling valve;
6 = compressor; 7 = Mohr clip; 8 = free air
500ml -
41
31
21
11
Fig. 2
Measurement of the
ozone demand;
special flasks
-------
- 36V'- -
3. PROCEDURE
3.1. Method of obtaining the ozonized air
Determine flow rate and the power consumption*.
Start the cooling water circulation of the ozoni-
zation unit.
Start the air compressor. •
Regulate the air' flow rate by means of the sampling
valve (the flow rate valve being closed). The
delivery-rate is read off a flow-meter (reading at
the top of the ball).
Regulate the air pressure by means of the pressure
2
valve; the pressure should be set at O.5 kg/cm (the
calibration curves of the ozonization unit were
plotted for this pressure).
Apply power to the ozonization unit.,
The concentration of ozone in the ozonized air produced
depends on two parameters: the air flow rate and the
power consumption. The calibration curves of the ozon-
ization unit enable a flow rate and a power consumption
corresponding to the desired concentration of ozone
to .be determined. The choice of the latter depends
on the nature of the water being studied, the aim being
to obtain a residue of O.4 mg of ozone per litre of
water after a water-ozone contact time of 4 min (see
Section 3.2). In cases where this residue is below a
value of O.3, or above a; value of O.6, it is necessary
to repeat the experiment with a-higher or lower ozone
concentration of the air. In the case where the highest
possible concentration of the ozonized air does not
give a residue of O.3, a special 6 litre flask must
be used, which enables the air/water volume ratio to
be varied from 1/5 to 5/1, this ratio being first made
1/2 and then higher if necessary. By way of an indi-
cation, a content of 10 .mg of • ozone per litre of air
is suitable for only slightly polluted water (such as
well water).
-------
- 362 '- '
Turn the power "hanVMheelclo'ckwise ' to obtain the •'>'*
desired power consumption; this is read off the
wattmeter. ; ,. •
Wait for five minutes to. achieve stable flow rate
and power consumption.
400
3OO
2OO
1OO
20W 30W SOW 60WTOW 10QW
power absorbed in
fl
air: dew point
- 50°C
pressure: O
kg/can'
C3onc.in mg O-,
air
10
15
20
25
Fig. 3 Mean air production curves for the LABO 66 unit
3.2. Performance of,the ozonization
Pill the flask described in Section 1.2 With the water
to be ozonized.
Connect the ozonized air pipe branched to the sampling
valve to the upper arm of, the ground-glass stopper?
the end can be partially closed by, means of a Mohr
clip opening to the atmosphere..
-------
- 363 -
When the Mohr clip has been adjusted so as to produce
a steady flow of ozonized air to the atmosphere
during the following operation, open the side tap of
the flask so that the water level descends to the 5OO
ml mark. The gas volume on top is now occupied by
ozonized air at a concentration corresponding to the
ozonizer setting.
Close the tap on the ground-glass stopper and dis-
connect the ozonized air inflow.
Shake the flask vigorously for 2O sec, timing this with
a stopwatch. (The ozonization reaction and the estab-
lishment of an ozone equilibrium between the air and
the water take place during these 2O sec).
Let the flask stand for 2 min, then tilt it so that the
water clears the side outlet-; the remaining ozonized
air is expelled by blowing air into the end of this
outlet. Determine the residual ozone in the water as
described below.
Repeat the experiment, this time letting the flask
stand for 4 min instead of 2 min.
Repeat as above, this time letting the flask stand
for 8 min.
3.3. Determination of the residual ozone
Immediately after this time (2, 4, or 8 min), place
several crystals of potassium iodide in the flask. The
residual ozone oxidizes- the iodide to iodine. Pour the
5OO ml of the iodine solution into a conical flask,
and 2 ml of 1:2 E^S.O., and titrate with sodium thio-
sulphate.,
-------
- 364 - - ,
Refer to the method of determination of residual ozone
in the water.
4. CALCULATION
4.1. Determination of the auto-decomposition
coefficient (-a)
Find the mean of the three values obtained with the aid
of the equations below* :
, . -1, 1 , r4 1 . r8 1 , r8
-a (mm 5 = ~- In — = T In — = T In —
z r,j 4 r« o r»j
where r,,, r., and r,, are the residual amounts of ozone
after contact times of 2, 4, and 8 min respectively.
4.2. Determination of the immediate ozone demand (D)
By definition, the immediate ozone demand is:
D = C - r
o o
where C is the initial ozone concentration in water,
introduced by means of the air-water contact; it
is calculated as in Section 4.21,
and r is the residual concentration of ozone in the water
o
at time 0, calculated as in Section 4.22.
4.21. Calculation of CQ
C = cv
o
where v is the volume of ozonized air in contact with the
water
V is the volume of water in contact with V-, ,
* If rg is less than O.2, the following value will be retained;
a = * In r4
Z r2
-------
- 365 -
30 -
20 -
10-
c is the con'centration of ozone in the ozonized air,
S is the distribution coefficient of the ozone between
the water and the air? the value of this coefficient,
which is a function of the temperature, is deter-
mined with the aid of Fig. 4 in the Appendix.
Fig. 4
Henry's coefficient (S)
as a function of
temperature 0
S = f(0)
0 0.1 0.2 0.3 0.4 0.5 OB 0.7 s
4,22. Calculation of r
o
r is found from the equations below (take the mean of
the 3 values otbained, or of the first two if r0 is less
o
than O.2);
In
= 2a
_2. = 43
in
8
= 8a
where a is the auto-decomposition coefficient calculated
in Section 4.1 and r«, r., and r0 are the residues of
^4 O
ozone after respective ozone-water contact times of 2, 4,
and 8 min.
-------
366 -
OZONE INPUT
W.J. Masschelein
As far as the mode of action and the passage of ozone into
water are concerned, at least two essentially different
mechanisms must be taken into consideration:
1) contact with bubbles of strongly ozonized air; we were
able to show (cf. our report to the Berliner Wasser-
congress in 1976) that as regards the bactericidal
action of ozone this mechanism plays a very important
if not the main part. Similarly, the literature data
on the use of ozone in the preliminary treatment of
water can be interpreted as favouring micelle formation.
This mode of action is illustrated by the rapid bacteri-
cidal action obtained during the passage of the ozone
from the gaseous into the dissolved phase.
2) continuous, stepwise, and slow action of the residual
dissolved ozone; this mechanism is particularly suitable
for combatting impurities more resistant to oxidation,
for example detergents. This mode of action corresponds
to first-order kinetics with respect to the concentration
of the dissolved ozone and also seems to be of decisive
importance for the viricidal action.
Therefore,during the planning and layout of an ozonization
plant these two basically different action mechanisms must be
borne in mind: passage of the ozone must be ensured, e.g. with
the aid of, turbines, and a relatively short contact time (less
than 1 min) is necessary; the effect is thus achieved by the
contact with bubbles of strongly ozonized air. On the other
hand, the slow mode of action of the residual ozone in the
after-treatnient requires a time of at least 6 min. During
this period a concentration of 0.2 to 0.4 g O^/m must be
maintained.
-------
- 367 -, • . .
OZONIZATI.ON BY-PRODUCTS AND THEIR REMOVAL BY COAGULATION
J.C. Kruithof
1. Introduction
In surface water, taking the river Rhine as an example,
many organic micro-pollutants may be found. One of the
most important groups of pollutants is the fraction of
the aromatics (1-3). In this fraction many alkylated,
hydroxylated, chlorinated and nitrated mono-aromatics
and poly-aromatics can be present. These impurities must
be removed in the process of drinking water treatment.
One of the processes to remove these compounds is ozo-
nization. In this paper the ozonization of phenol,
naphthalene and phenanthrene will be discussed. In
particular, the production and stability of organic
peroxides will be mentioned. Furthermore, the formation
of dialdehydes and their conversion to carboxylic acids
will be discussed. The ozonizations have been carried out
at a pH of about 7, so the initial reaction of ozone will
be an electrophilic substitution and not a reaction of ozone
with hydroxylic ions.
In the second part of the paper some examples are given
for the removal of carboxylic acids produced during ozo-
nization by coagulation with aluminium salts. The removal
of oxalic acid, o-phthalic acid and 1,3,5-benzene tricarbo-
xylic acid will be reported. The models for removal are
based on the hydrolysis of the metal salt used, the disso-
ciation of the carboxylic acid, the complexing of the
metal ion with the carboxylate ion and the solubility of
the metal hydroxide used.
-------
- 368 -
2. Ozonization of some aromatic compounds
2.1 Ozonization of phenol
r
Many authors describe the ozonization of phenol in water.
Bauch' et al. (4) assume the formation of an ozonide which
is converted to smaller molecules, and a direct oxidation
by which the benzene ring is ruptured. Eisenhauer (5-7)
finds catechol and o-benzo quinone as the first reaction
products. Gould (8, 9) states three mechanisms: hydroxyl-
ation, cleavage of the benzene ring and the production of
polymeric compounds. According to our own investigations,
two major reaction paths occur, the first of which, taking
place for about 30 %, is hydroxylation producing'catechol
and hydroquinone as reaction products. The concentration
of these products is given in Figure 1. Both concentrations
reach a maximum after an ozone consumption of 3.6 • 1O
moles (0.4 moles of ozone/mole of phenol originally present).
In addition, there is a rupture of the benzene ring for 70 %,
producing an aldehydic and an acid hydroperoxide. Both per-
oxides prove to be unstable and will either cause an elimi-
nation of H_02 or rearrangement. An elimination of H O will
convert the aldehydic peroxide to glyoxal. When a rearrange-
ment takes place the aldehydic peroxide is converted into
two moles of formic acid, the acid peroxide rearranges to
one mole of formic acid and one mole of CO . The concen-
tration of glyoxal and H_O2 is given in Fig. 2, and that of
formic acid and CO in Fig. 3. The concentrations of glyoxal,
£
H0O0 and formic acid reach a maximum after an ozone consumption
-3
of about 26 • 10 moles (2.9 moles of ozone/mole of phenol).
-------
- 369 -
1O
A
20
,IO"*(")ole$)
Fig. 1 The concentration of catechol and hydroquinone
as a function of the ozone-uptake
(x = catechol; o = hydroquinone)
10.0
8,0
u 6,0
c
O
u
4,0
2.0
O
O 10 20 30 40 50 60 70
fc— A 03" JO'Violes)
Fig. 2 The concentration of glyocal and H202 as a function
of the ozone-uptake
(x = glyoxal? o = H2O2)
-------
- 370 -
c
o
u
50,0
40,0
30,0
20,0
10,0
0
5O.O
o
40,0 £
30,0
20,0
10,0
O
0 10 20 30 40 50 60 70 80
• • A 03 Jo'^rnoles)
O
U
Fig. 3 The concentration of formic acid and the C02-
production as a function of the ozone-uptake
(x = formic acid; o = carbon dioxide)
At this point all phenol has reacted with ozone. In
consecutive reactions glyoxal is converted to glyoxylic
acid, forming oxalic acid. In the absence of H2°2 oxalic
acid proves to be a stable final product. Formic acid is
oxidized to C0~. The concentration of the organic compounds
<&*
formed by secondary reactions is presented in Fig. 4. At
the end of the reaction, at an ozone consumption of
70
10
-3
moles, 75 % of all carbon atoms are found as
C0? and 25 % in the form of oxalic acid.
It can therefore be concluded that when ozonizing phenol,
3O % will be hydroxylated to catechol and hydroquinone,
while 70 % of the phenol will produce unstable peroxides
which eliminate H2O2 or rearrange to glyoxal, formic acid,
or CO,,. CO- and oxalic acid are the only final products.
<*£* ^
The complete ozonization of phenol is shown in scheme 1.
-------
- 371 -
20.0
16,0
12,0
o
u
8.0
4,0
0
0 10 20 30 40 50 60
••i A 03 .lo"*(moles)
70 80
Fig. 4 The concentration of glyoxylic acid and oxalic acid
as a function of the ozone-uptake
(x = glyoxylic acid; . o = oxalic acid)
3 0
35 %
2
H
\ - C
I
HC - 0 -OH
/
HO
2HCOOH ' H,0, + C
4
HO
vc«o
I
HC - 0 - OH
/ 1
HO t
•1 HCOOH +~t
^\J i
°2
I fH 1 °3
j o, cH«o 1 3
' i CO
CC2 | 03 C°2
OH
COT '
c"-o
COOH
| o3
(COOH)-
1 VHZ°2
QH
y
OH
Scheme 1' The ozonization* of phenol
-------
- 372 - ' .
2.2 The ozonization of naphthalene
Many authors describe the ozonization of naphthalene in
organic solvents. Bailey et al.. (.10-12) ozonize naphthalene
in CC1. and methanol. They assume, that there is production
of a mono-ozonide in CCl^, while in methanol they identify
two cyclic aromatic peroxides, which prove to be rather
stable. Sturrock et al. (13), ozonize naphthalene in mixtures
of water and acetone. They assume an.equilibrium between an
open and a cyclic peroxide. .,
From our own experiments we can conclude that an open
hydroxy-hydroperoxide is produced after an ozone consumption
of 2 moles/mole of naphthalene originally present. This open
peroxide proves to be unstable and results in a cyclization
or an elimination of ^2°2' In the first case a cyclic peroxide
(a 1,2-dioxane derivative) is formed.
The concentration of this compound is presented in Fig. 5.
It can be seen that the concentration of the cyclic peroxide
™ "3
reaches a maximum after an ozone consumption of 16 • 1O
moles (2.9 moles of ozone/mole of naphthalene). The highest
concentration found is 2.3 • 1O M, indicating that about
45 % of the total naphthalene originally present will
produce this peroxide. For the remaining 55 % an elimination
of H_O_ takes place, forming o-phthalaldehyde as the organic
^* 4-»
product. Both concentrations are shown in Fig. 6. Again,
the concentrations reach a maximum after an ozone consumption
of 16 • 1O moles. The highest concentrations found are
2.8 • 1O~3 M and 2.7 • 1O~3 M for o-phthalaldehyde and ^2°2'
respectively. Without an additional ozone supply all aro-
matic compounds are converted to o-phthalic acid. The con-
centration of this compound is shown in Fig. 7. It can be
_3
seen that after an ozone consumption of 32 • 1O moles
(5.4 moles/mole of naphthalene originally present) the con-
-3
centration of o-phthalic acid is 6.5 • 1O M. This means
that about 9O % of all naphthalene originally present are
-------
- 373 -
converted to o-phthalic acid, which proves to be rather
stable against further ozonization. After an extensive
ozonization, 7O % of all carbon atoms are measured as
CO- and 3O % as oxalic acid.
Therefore, when naphthalene reacts with ozone, an open
peroxide is produced. 45 % of this peroxide are converted
to a very stable cyclic peroxide, 55 % result in an eli-
mination of H2O2, yielding o-phthalaldehyde. Both organic
compounds form o-phthalic acid with an additional supply
of ozone. The first part of ozonization of naphthalene
is illustrated in scheme 2.
10 15 2O 25
A 03 .io"*(moks)
3O
Fig. 5 The concentration of 3,6-dihydroxy-4,5-benzo-1,2-
dioxane as a function of the ozone-uptake
-------
- 374 -
0
10 15 2O 25
A 03 ,lo"4(fnoles)
Fig. 6 The concentration of o-phthalaldehyde and H2O2 as a
function of the ozone-uptake
(x = o-phthalaldehyde; o = H2C>2)
O
1O 15 2O 25
A 03 . io"'(mo/es)
3O
Fig. 1 The concentration of o-phthalic acid as a function
of the ozone-uptake
-------
- 375
>* 35 Z
H
C- 0
I
C - 0
'OH
XOH
V0-OH
7 V
;" - 0
LO - OH
""OH
OH
"OH
CH-0
OH
Scheme2 The first phase of the ozonization of
naphthalene
2.3 The ozonization of phenanthrene
As for naphthalene, several authors describe the ozonization
of phenanthrene in organic solvents. Schmitt et al. (14)
ozonize phenanthrene in chloroform and find .an.iso ozonide.
Bailey et al. (15, 16) ozonize phenanthrene in methanol and
identify a cyclic peroxide. The same mechanism is given by
Sturrock et al. (17) for the ozonization of phenanthrene
in water-alcohol mixtures.
Our own experiments indicate the production of an open
hydroxy-hydroperoxide after an ozone consumption of
5 * 1O moles of ozone (1 mole/mole of phenanthrene
originally present). In this case the open peroxide
-------
- 376 -
cyclizizes for less than 5 %, and an almost complete
elimination of H^O,^ takes place, forming 2,2 '-diphen-
aldehyde as the organic product. Both concentrations are
given in Fig. 8. In this Figure, a maximum concentration
is shown for diphenaldehyde of 3.6
1O at an ozone
consumption of 5.O • 1O moles. An additional ozone
supply produces a conversion to '.diphenic acid which proves
to be fairly stable against ozonization. After an extensive
ozonization, CO- and oxalic acid are the only final products.
Thus an open peroxide is produced when ozonizing phenanthrene,
This peroxide is almost completely converted to 2, 2 '-diphen-
aldehyde and HpO..,. The aldehyde produces diphenic acid when
reacting with ozone. The first part of the ozonization of
phenanthrene is shown in scheme 3 .
5,0
4.0
u 3,O
c
o
o
2.0
1.O
O-'
__o
I
I
I
6 8
03 . ic
I
I
10 12 14 16
Fig. 8 The concentration of diphenaldehyde and H2O2 as a
function of the ozone-uptake
(x = diphenaldehyde; o = H202)
-------
- 377,-
Scheme3 The first phase of the ozonization of
phenanthrene , _
OH
2.4 Conclusions from the ozonization experiments
and discussion
From the above described and additional experiments (1)
the following conclusions can be drawn:
1
When ozonizing aromatic compounds H2O2 is always
produced.
zation,
will be removed by exhaustive ozoni-
When ozonizing poly-aromatics , very stable organic
peroxides can be formed if the initial attack of
ozone produces an aldehyde group and a hydroxy- •
hydroperoxide group in ortho position at the same
benzene ring.
In all other cases when ozonizing poly-aromatics
dialdehydes and H2O2 are produced almost completely.
The organic peroxides as well as the dialdehydes are
converted into carboxylic acids such as o-phthalic
acid, diphenic acid, oxalic acid, etc., with an addi-
tional supply of ozone. When ozonizing larger organic
-------
- 378 -
molecules (such as humic acid), benzene polyearboxylie ,
acids (such as 1,3,5-benzene tricarboxylic acids) are
produced. These acids are rather stable against further
ozonization.
5. Formic acid, glyoxalic acid and oxalic acid are the
only aliphatic acids formed when ozonizing phenol,
naphthalene and phenanthrene. Oxalic acid proves to
be a stable end-product in the absence of H0O0.
<£* £
Exhaustive ozonization can never be carried out economically,
so that many carboxyllc acids can be present in the ozonated
effluent. These compounds must be removed in a subsequent
purification step such as activated carbon filtration or a
secondary coagulation step. The second part of the paper'
gives some theoretical and experimental results for the
secondary coagulation step.
3. Removal of carboxylic acids with aluminium salts
In this part of the paper some models will be presented
by which the removal of carboxylic acids produced by ozoni-
zation can be described. For a quantitative approach the
following constants must be known:
- Stability constants for the hydrolysis of the metal
%* **
salt used (*B-i- 84)-
- Dissociation constants for the carboxylic acid (K1-K ).
- Stability constants of the produced metal carboxylate
complexes(P--p ), and in case of soluble complexes:
5*
- The solubility product of the metal hydroxyde ( KqO).
-------
- 379 -
Based on differences in these properties, the following
mechanisms of removal may occur:
1. Formation of a soluble, strong, negative complex. The
removal of this complex takes place by adsor-ption of
the complex on a floe of metal hydroxyde. The pH is a
function of the metal salt concentration and varies
between the pH for optimum complexation and the iso-
electrical point.
2. Formation of a soluble, weak complex. The removal is
caused by adsorption of free carboxylate ions on a
floe of metal hydroxyde. The optimum pH coincides with
the iso-electrical point.
3. Formation of an insoluble complex, which is removed
as such. The optimum pH coincides with the pH for
optimum complexation.
With all carboxylic acids and with aluminium salts,as well
as with ferric salts, the same calculations can be used.
In this paper, the complete calculation will only be given
for the removal of oxalic acid with aluminium salts. The
experimental part will describe the removal of oxalic acid
as well as the removal of o-phthalie acid and 1,3,5-benzene
tricarboxylic acid. -
3.1 The hydrolysis of aluminium salts
Aluminium ions are hydrated in strong acid solutions with
6 coordinatively bound water molecules. When the pH rises
the following steps of hydrolysis take place:
Al (H20)53+
-------
- 380 -
2+
[A1(H20)5OH] "*+ H2Q' ( ' > [A1(H20)4(OH)2] + H30
[A1(H204(OH)2] + + H20 - - >• [A1(H20)3(OH)3] +H30+
[A1(H20)3(OH)3] + H20 - - * [Al (H2O) 2 (OH) 4] ~ +H30+
According to Sillen and Kartell (18,1'9) these reactions
have the following stability constants :
=
3 I '
According to Schwarzenbach (2O) an "Aj (QH) can be defined
This «ai (OH\ gives "the ratio of the sura of the concentra-
tions of all metal ions over the concentration of Al :
aAl(OH)'
IA1 -J [A13+]
(1)
This formula is valid when no polynuclear complexes are
produced. This can be assumed as will be shown in a fol-
lowing paper (21). Introducing the stability constants
gives :
- 1 10~4'89 10~9'37 IP"15'04 10-20,30
a Al(OH) " 1 + £H+, + rH+.j2 + rH+-,3 + rfl+, 4 '
The log «.., /QH\ as a function of the pH is given in
Figure 9.
3.2 The dissociation of oxalic acid
Oxalic acid dissociates in two steps (22)
-------
- 381 -
HOx -
HOx
_.. OX
i
["° I
[HOx ]
_5
As with the hydrolysis of aluminium salts an a , .
can be defined
[Ox1] [H2Ox] + [H0x~] + [Ox ]
Ox(H)
2—
2-
(5)
Introducing the dissociation constants of oxalic acid
gives
[H30+][ [H30+]2
&,
Ox(H)
K,
K1K2
Log ao , . calculated with formula (6) is reprensented
in Figure 1O.
16
12
x
9 8
O
Ol
O
O
Fig. 9
Log a ' -,
of the pH
as a function
O
8
PH
(6)
-------
- 382 -
Fig. 1O
Log aQ
of the pH
as a function
3.3 The complexation of the aluminium-ion with
oxalate-ions
The aluminium-ion forms,with oxalate-ions, the following
complexes (1,23) :
Al(Ox)
3+
A1
A13+ + 30x2~-
- 2
= 107'26
13,0
- 10
[Al(Ox)
3+
- =
2-3
[A1][OX-]
= 10
I6'B
can be de-
(7)
(B)
(9)
For this complexation the following a,-, /Ox\
fined J
[Al1 J [Al3+]+[Al(Ox)+]+[Al(Ox)2^]+[Al(Ox)33~]
aAl(Ox) = ^~ = ^(1
0)
-------
-T 383 - .
Introducing the stability constants 3 in formula, (1O). ,.,
gives :
*Al(Ox)
2-.
-,2
With, the help of formula (5) formula (11) can be written
as :
'2 '3
B2[0x'] B3[0x']
Al (Ox)
aOx(H) aOx(H)
aOx(H)
(12)
So aA-, /o x can be calculated as a function of the pH when
the oxalic acid concentration is known. The aa,/n x for
-3 -4 A.Myx;
oxalic acid concentrations of 1O / 10 and 1O'- M is
given in Figure 11.
aai tr> \ as a function of the pH for three
r\_L \ vJX )
oxalate concentrations;
-4,
b: [Ox']= 10 M; c:[px'j = 1O M; d: [px' ] = 1O M
-------
- 384 -
tot
Besides an a-, can be defined covering all aluminium com-
pounds :
.M. £, u -—- \ — ' in •* can be defined:
«A1(OX)
With formula (1) and (1O) formula (15) can be converted
to :
_1
aAl(OH) aAl(Ox) ' _ aAl
- (16)
Al(Ox) «A1(OH)
Al(OH)
or
log a1 _ ,„„ tot , „
Al(Ox) ~ 10gaAl - l0^ aAl(OH)
-------
- 385 -
Log aA1(o x can be calculated'as a function of the pH
when the total concentration of oxalic acid is known.
For oxalic acid concentrations of 10 ,10 and 10 M
log a' is represented in. Figure 12.,
Al(Ox)
x
o
as a. function of the pH for three
oxalate concentrations;
b: [Qx']= 10 M; c:[0x']= 1O M;
l= 1O~3M
3.4 The precipitation of aluminium hydroxide
When the produced complexes between aluminium and the
carboxylic acids are soluble in water, as is the case
with oxalic acid, the precipitation of aluminium, hydro-
xyde plays an important part. The precipitation of the
hydroxyde can be represented by the following simple
reaction :
-------
- 386 -
A1
3+
3 ' OH"
Al(OH)
3, S
For this reaction the following dissociation constant
is known (18) :
3+
-3
-34
J4t
KQO = [ArH [OH"] J = 10~J* or K
SQ
SO
= 10
8
(18)
Together with formula (13) this can be written as
[Al'j
K^ = "~a
SO T +•, 3 tot
|_H3O J a,.
Formula (19) gives the limiting aluminium concentration
for hydroxyde precipitation as a function of the pH and
the oxalic acid concentration. The results are represen-
ted in Figure 13.
S
10
Fig. 13 Log [Al1] for minimum hydroxyde precipitation as a
function of the pH with the following oxalate
concentrations;
-5,
a: [Ox']= O M; b: [Ox']= 1O M; c: [Ox-']= 10
,-4
-3,
d: [Ox']= 1,0 M
(19)
-------
- 387 -
3.5 Discussion and presentation of the mechanisms for
removal
From paragraphs 3.1 to 3.4 and from Figs, 9 to 13 the
following conclusions can be drawn:
- At a pH higher than 4 aluminium salts hydrolyse to
aluminium hydroxyl complexes.
- At a pH higher than 5,5 oxalic acid is completely
dissociated into oxalate ions.
- Independent of the concentration of oxalic acid
a', ,Q » reaches a maximum at a pH of about 4.5.
So at a pH of 4.5 the complexation dominates over
the hydrolysis as much as possible. The value of
aAlfo ) varies between 2.5 and 7.1 indicating that
aluminium ions produce strong complexes with oxalate
ions.
- The minimum aluminium concentration for precipitation
of Al
acid.
of A1(OH)2 is a function of the concentration of oxalic
- Besides, it is known that aluminium oxalate complexes
are soluble in water.
From these facts it can be concluded that oxalic acid is
removed via mechanism 1. It forms soluble, negatively
charged strong complexes at an optimum pH of 4.5. At this
pH optimum removal takes place when enough aluminium
hydroxyde is precipitated. At low concentrations of alu-
minium this will not be the case, so the optimum pH for
removal will shift to higher values.
-------
- 388 -
Based on the same kind of calculations it can be concluded
that o-phthalic acid is removed by way of mechanism 2.
It gives very weak complexes, which are soluble in water..
The removal takes place by adsorption of free phthalate-
4-ons on a floe of A1(OH)3. 1 ,3,5-benzene tricarboxylic acid
is removed via mechanism 3. It forms rather weak complexes,
which are insoluble in water. The removal will take place
by precipitation of the complex.
The proposed mechanisms for removal will be tested in the
experimental part.
3.6 Experiments
To check the proposed mechanisms for removal, experiments
have been carried out with a constant concentration of
—4
carboxylic acid of about 1O M and a varying pH and alu-
minium dose. The results for oxalic acid are presented in
Fig. 14. From this figure it can be concluded that the opti--
mum pH for the oxalic acid removal is a function of the
aluminium-salt concentration. At an aluminium-salt concen-
-4 "
tration of 1O M the optimum pH is 6.8. This pH coincides
with the iso-electrical point of Al(OH)3. At higher alu-
minium-salt concentrations the optimum pH drops to lower
values, approaching the pH where «,-,,_. > reaches its maximum.
At very high concentrations of aluminium (5 x 1O M) the
oxalic acid removal reaches a maximum at a pH of 4.5, so the
system aims at maximum complexation of the aluminium ion by
oxalic acid. Always an aluminium hydroxide precipitate must
be present to obtain a removal completely in agreement with
the proposed mechanism.
-------
- 389 -
1OO
O
o
O 40 -
Fig. 14 The relative removal of oxalic acid as a function
of the pH and the concentration of aluminium salt;
[Al']= 10 M; 0:[A1']= 2.Ox1O M; x:[Al']= 4.1x10 Mf
:= 6.2x1O~4M
The experiments with o-phthalic acid are presented in
Fig. 15. From this figure it can be concluded that the
optimum pH for removal is about 6.4. This indicated that
o-phthalic acid is removed" by adsorption of free ions on
a flox of aluminium hydroxyde. The complexation does not
play an important part in this case.
Finally, the experiments, with 1,3,5-benzene tricarboxylic
acid are presented in Fig. -16.: This figure shows an opti-
mum removal at a pH of about 4.7, independent of the con-
centration of aluminium salt." This indicates a removal via
an insoluble complex.
Concluding it can be said that dicarboxylic acid, pro-
duced during ozonization, can be removed in a secondary
-------
- 390,-'
coagulation step. The acids are removed via.the proposed
mechanisms. The mechanism for removal can be predicted
when the strength and the solubility in water of the
produced complexes are known.
1001-
Fig. 15 The relative removal of o-phthalic acid as a function
of the pH and the concentration of aluminium salt?
> 1.8x10 4M;
-4,
O:,
x: [Al']= 9.3x10 M;
4.4xlO~4M;
-3,
CAl"If= 1.8x1O M
-------
- 391 -
'1QQ
eo -
O 6O
U'
fy^ 40
-° ai|
ul I—I
»-
ms
20
23456789 10 ;
— pH
Fig. 16 The relative removal of.1,3,5-benzenetricarboxylic
acid as a function of the pH and the concentration
of aluminium salt;
x:
= 1.8x10 M;
= 8.9x1O~4M
0:
.'1= 4.5xlO~4M;
Summary
In this paper the reaction" of ozone with phenol, naphtha-
lene and phenanthrene is reported. Mainly the production
of stable peroxides and carboxylic acids has been investi-
gated. When ozonizing phenol and phenanthrene only hydrogen
peroxide is produced. The ozonization of naphthalene shows
the production of hydrogen peroxide as well as an organic
peroxide. This peroxide (a 1,2-dioxane derivate) is very
stable in the absence of ozone. Additional supply of ozone
to all reaction mixtures will result in the production of
carboxylic acids, such as oxalic acid, phthalic acid,
diphenic acid, etc.
The second part of the paper deals with the removal of
-------
- 392 - •
carboxylic acids by coagulation. Some models are pre-
sented by which the removal can take place. Oxalic acid
is removed via adsorption of aluminium oxalate complexes
on a floe of aluminium hydroxyde; phthalic acid is removed
by adsorption of free phthalate ions on a floe of aluminium
hydroxyde, and benzene tricarboxylic acid is removed as an
insoluble complex.
(1) K.RUITEGF, J.C.
Thesis Delft (1978)
(2) MEYERS, A. P.
Quarterly reports KIWA/RIWA Research (1972-1975), 1-11
(3) v.d. LEER, R.C., v.d. MEENT/W.
Quarterly Reports KIWA/RIWA Research (1975-1976), 12-14
(4) BAUCH, H., BURCHARD, H., A.RSOVIC, H.M.
Gesundh. Ing. 9. (197O), 258
(5) EISENHAUER, H.R.
J. WPCF 4O (1968), 1887
(6) EISENHAUER, H.R.
J. WPCF 43 (1971), 2OO
(7) EISENHAUER, H.R.
Water Res. _5 (1971), 467
(8) GOULD, J.P.
Thesis Michigan (1975)
(9) GOULD, J.P., WEBER, W.J. jr.
J. WPCP 48 (1976) , 47
(10) BAILEY, P.S., GARCIA-SHAPR, P.J.
J. Org. Chem. 22 (1957), 1OO8
(11) BAILEY, P.S., BATH, S.S., DOBINSON, F., GARCIA-SHARP, F.J.,
JOHNSON, C.D.
J. Org. Chem. 29 (1964), 697
-------
- 393 - .
(12) JOHNSON, C.D., BAILEY, P.S".
J. Org. Chem. 29. (1964)-, 7O9
(13) STURROCK, M.G., GRAVY, B.'j'. , WING, V.A. :
Can. J. of Chem. 49 (1971-), 3O4
(14) SCHMITT, W.J., fiORlCONt, E,J., O'CONNOR, W.F.
J. Am. Chem. Soc. 77 (1955), 564O
(15) BAILEY, P.S.
J. Am. Chem. Soc. 78 (1956), 381
(16) BAILEY, P.S., MAINTHIA, S.B.
J. Org. Chem. 21 (1956), 1335
(17) STURROCK, M.G.., CLINE, E.L., ROBINSON, K.R.
J. Qrq. Chem. 2_8 (1963), 2340
(18) SILLEN, L.G., MARTALL, A.E.-
Stability Constants
Special Publication No. 17, Chem. Soc. London (1964)
(19) SILLEN, L.G.,.MARTELL, A.E. '
Stability Constants, Suppl. No. 1, Special Publication
No. 25, Chem. Soc. London (1972)
(2O) SCHWARZENBACH, G., HELLER, J.
Helv. Chim. Acta 3_£- (1 951) , 576
(21) HEERTJES, P.M., KRUITHOF, J.'C,
To be published
(22) WEAST, R.C.
Handbook of Chem. and* Phys., 52th ed., The Chem.
Rubber Co., Cleveland, Ohio (1971/1972)
(23) SECCO, F., VENTURINI, M.
J. Inorg. Chem. 14 (1975), 1978
-------
- 394 -
MICROFLOCCULATION BY OZONE
D. Maier
1. Introduction
Although there is no doubt that ozone is neither a floccu-
lating agent nor a precipitating agent in the conventional
sense, but first and foremost an oxidizing agent, repeated
mention is made in connection with ozonization of so-called
microflocculation. All•observations made up to now, given
in a following chapter, show that the ozonized organic con-
stituents of water acquire a special significance in the
explanation of this phenomenon. For this reason, the known
effects of ozone treatment on natural, organic water con-
stituents will therefore be summarized.
2. Action of ozone on the organic constituents of water
In practically all publications relating to actual prac-
tice/ purely phenomenological reports are found on the
reduction of the water's colour, odour, and taste — an effect
that can be directly perceived by the senses without re-
sorting to measuring apparatus and analytical procedures.
At least the alterations in colour due to the ozone/can be
evaluated quantitatively by a simple photometric process,
and in the case of strongly coloured waters this can be
done directly in the visible spectrum.
By extending the spectrum to the UV-absorption region
further clear changes of the UV-spectra are observed, which
in all cases studied so far show considerably lower ex-
tinction values than for non-ozonized water.
If the dissolved organic carbon, the content of which is
hardly decreased by ozone treatment in the case of waters
charged with natural constituents, is determined at the same
-------
- 395 -
time, the first guiding points become clear: in the first
place ozonization does not bring about a quantitative oxi-
dation of the water constituents to carbon dioxide but rather
a chemical conversion of these constituents.
From a comparison of the chemical oxygen demand/ measured
before and after ozonization of organically loaded waters
it is found in addition that this chemical conversion is
above all a conversion into polar constituents rich in
oxygen. Simultaneous investigations of the molecular-
weight distribution show that in ozonization this chemical
conversion is accompanied by a reduction of the molecular
weight.
Consideration of the reaction mechanisms cited in the liter-
ature (1,2) makes it evident that, depending on the pH,
both electrophilic additions to any multiple bonds presen^
and radical attacks on the molecule are possible. For
example, it can be clearly shown by infra-red spectroscopy
of isolated ozonized water constituents (3) that ozonization
leads to a strong attenuation of the double-bond character
and the aromatic character, while simultaneously an increase
in the contents of hydroxyl, carbonyl, and carboxyl groups
is noticeable by the greater intensity of the corresponding
absorption bands.
Several recently developed methods for the determination
of organic acids (4-6) indicate on the basis of many
experiments with various waters that the carboxyl groups
produced during ozonization hold quantitatively the key
position in the formation of new functional groups.
-------
§
•H
0.3
396 -
-------
- 397 -•
|^ a)Bodensee-Rohwasser aus60m Tiefe
I ] b) Bodensee- Rohwasser aus 60m Tiefe nach Ozonung (0.9 mg/l O3)
(Aktivkohte-Aceton-Extrakte) '
"KiO,
300
. i •
T [C]
200
1 —
10O
40
1
1:10
80 3O
Temperaturprogr.-imm ( 2°C/min)
O
O
Fig. 2 Decarboxylation of organic water constituents
a) raw Lake Constance water.from a depth of 6O m;
b) raw Lake Constance water from a depth of 6O m
after ozonization(O.9 mg 03/1)
(activated carbon-acetate extracts) 1 cm2=1.4yg CC>2
The decarboxylation spectrum reproduced in Fig. 2 shows
in 'the higher temperature range, in which intramolecular
oxidation processes play a part, that more active oxygen
available for reaction is present in the ozonized than in
the non-ozonized material.
This may also be why a shift of the oxidizability to more
readily oxidizable compounds can be seen with ozonization
in the temperature-programmed wet oxidation of ozonized
organic water constituents, as shown in Fig. 3, which is
based on the work of Weindel and Maier (8).
-------
- 39'8 -
45mifl
3J.5tn«
Wet oxidation K2Cr2°7
'take Constance water
Band filtrate without
oEone treatment
Lake Constance water sand
filtrate after ozone
treatment with
0.9 s oyV
Omm
5C-
25
Vol -ppm CO? in $2 carrier gas
gig.3 Oxidation spectra of organic substances in Lake
Constance water. Influence of 63 treatment
3. Selected effects of ozone-induced type alteration
of organic substances on practical treatment of
drinking water
3.1. Toxicity and repopulation with bacteria
All the molecular properties altered by ozonization exhibit
clear effects in the practice of drinking water treatment.
These too should be given.in summarized form.
In the first place it should be mentioned that ozonization
does not produce any material more toxic than those already
present in the water.
-------
- 399 -
The strongest evidence (9) for this is the fact that
sterile water^ treated with various ozone concentrations and
having a content of about 1.2 mg C/l of organic carbonr
tends increasingly towards bacterial regeneration after
the addition of a bacteria-containing water with increasing
ozone concentration. This effect, desirable in the purifi-
cation of waste water and already practised .in some plants,
indicates that ozonization leads to the formation of less
toxic substances better utilizable by the bacteria. It is
now also known that this so-called bacterial repopulation
is only observed in waters with a sufficiently high content
of organic carbon, according to the literature usually in
excess of O.5 mg C/l.
Maier's investigations (1O) on the problem whether humic
acids are made toxic by ozone treatment, with Daphnia
pulex as the test organizm, also show clearly that, at any
rate in the case of the humic acids of Lake Constance water
no toxic substances are produced by ozonization.
3.2. Chlorine consumption and haloform formation
For practice at least equally significant is the clearly
demonstrated fact that the chlorine consumption of water
charged with organic material is essentially slowed down by
preliminary treatment with ozone, and consequently during
the subsequent chlorination a smaller amount of organic
chlorine compounds is formed (11). This can be clearly
observed by investigating the chloroform formation in waters
ozonized to different degrees. According to the work of
Stieglitz et al. (12), some waters after ozonization have an
increased content of particularly the readily volatile
organic chlorine compounds. Since the possibility of oxi-
dation of chloride to chlorine by ozone can be ruled out,
this can only mean that organically bound chlorine is already
present in the material at the start, and that the ozoniza-
-------
- 4OO •- •• -
tion transforms it into compounds of lower molecular weight,
i.e. smaller, more volatile, and capable of determination by
gas chromatography.
4. Observations on microflocculation
4.1. Review of the literature
The described change in the type of the organic water
constituents due to ozone is an essential prerequisite for
understanding the microflocculation associated with ozon-
ization.
For example, Gomella (13) has given a comprehensive report
on the use of ozone in Prance, stating that in certain
special cases the flocculation of colloidal organic sub-
stances occurs, e.g. in the case of strongly coloured, humic-
acid-containing waters.
Gomella and Hallopeau (14) further assume that the floccula-
tion is due to partial destruction of organic macro-protective
molecules of natural origin, and conclude that this effect
could help to reduce the amounts of conventional flocculation
agents used. Taylor (15) also believes that by treatment
with ozone the organic constituents of water are changed in
such a way that during the subsequent flocculation with iron
or aluminium salts the quantity of the flocculating agent
required could be reduced.
In a detailed paper on the removal of iron from ground water,
on the other hand, Cromley and O'Connor (16) describe the
gelling action of ozonized organic substances, which prevents
the flocculation of iron if the dissolved organic substances
have not been largely decomposed.
In the treatment of the water of Lake Constance to supply the
town of St. Gallen with drinking water, Grombach (17) regards
-------
- 401 -
it essential to follow the ozonization by a filtration stage,
since "during oxidation a slight turbidity is produced, due
to the precipitation of dissolved material".
The findings of Campbell and Pescod are also of interest (18);
after ozonization of micro-filtered Scottish lake water, these
authors observed the formation of an organic foam on the
water surface, the amount of this foam increasing with
increasing colour of the water, i.e. with increasing content
of humic acids.
In studies on the removal of bacteria from a severely polluted
stream with ozone, Franz and Gagnaux (19) observed a marked
after-turbidity in" the ozone-gasification chamber, which
exerted an unfavourable effect on the removal of the bacteria.
The authors assume that the bacteria become encapsulated in
the precipitated microflocculated material and so avoid being
attacked further by the ozone.
Rohrer (20) claims that any peptides present in the water are
flocculated by ozone, and that clay minerals are also rendered
flocculatable to a large extent by an ozone treatment.
In a treatment of surface water described in a French Patent
(21), the so-called M.D. process/ after micro-filtration ozon-
ization is used to bring about micelle formation, i.e. the
occurrence of organic turbidity, which is removed by "demi-
cellization", i.e. by flocculation with aluminium sulfate in
a subsequent sand filtration stage.
From a spring water containing humic substances Kopecky (22)
isolated, after treatment with ozone, membrane-filterable
organic compounds which he described as "ozonides" and which
-------
- 402 -
were shown to be bactericidal after thei,r application to
culture media inoculated with bacteria.
Rempel and Summerville (23) also observed increased turbidity
after the ozonization of a surface water.
While this literature survey is essentially limited to the
cases in which increased turbidity of a predominantly organic
nature appeared after ozonization, there are also certain
indications that turbidity present initially in water can be
reduced by treatment with ozone.
Sontheimer (24) was the first to show, in fundamental experi-
ments on the ozonization of Lake Constance water, a clear
reduction of the particle count on the addition of ozone,
which he attributed to an agglomeration of the smallest
particles into micro-flocks.
"Ozone has a precipitating action" claims Kulcsar-Mescery (25),
mentioning at the same time that fine clay suspensions, iron,
manganese, humic substances, and other colloidal impurities
are precipitated.
O1Donovan (26) studied the use of ozone in three Irish lake
waters, and in the analysis of the ozone-treated waters
established that their turbidity is normally lower than that
of raw water.
In investigations on the filtration of suspended matter from
Lake Constance water, Wagner, Keller and Miiller (27) found
that the suspended matter can be more efficiently removed by
preliminary ozonization of the water, independently of the
turbidity of the raw water; in the dosage range studied,
between 0.5 and 2.3 g ozone/m ,the amount of ozone used had no
effect on 'the result. With the simultaneous use of floccu-
lation agents (aluminium sulphate) in these experiments it
-------
,_ 403 -
could also be demonstrated that to achieve the same degree
of removal of the suspended matter -the ozonized water required
less of the flocculation agent than non-ozonized water. The
authors came to the conclusion "that as a result'of ozon-
ization an improvement in the flocculation and filtration
capacity of colloidal substances is obtained". This may be
due to the flocculation effect that occurs in the ozonization
of humic-type organic substances.
Similar findings were reported by Wurster and Werner (28),
who in the treatment of Danube water observed an optimal
removal of very fine suspended particles by means of rapid
filtration with the combined use of secondary flocculation
and ozonization.
The literature evaluation of this problem was conducted as
extensively as possible. It is already becoming evident
that, according to the described experience gained in the
practice of drinking water treatment, more than one
mechanism can be made responsible for the so-called micro-
flocculation, as indicated by the following experimental
results - both our own and those selected from the liter-
ature.
4.2. Selected experimental results on micro-flocpulation
In the ozone treatment of micro-filtered water of Lake
Constance a situation is observed at certain times of the
year - mostly in the summer during intensive production of
algae and a perceptible increase in the organic products of
algal metabolism — that can be characterized purely optically
by an increased foam formation on the water surface in a
reaction tank after the ozonization stage.
This situation, shown in Fig. 4, can be explained by flota-
tion processes of suspended matter newly formed during the
ozonization.
-------
- 404
Fig. 4
Foam formation after
ozonization of
Lake Constance water
TABLE 1 Analysis of the "flotation foam" after ozonization
of Lake Constance water on Nov. 28, 1977
I'arawct or
Organic substance- as loss on ignition
?oo*c/(ioon*c)
Acid insolublu components as SJO,
Calcium
Kjynesiura
Aluminium
Iron
Manganese
Phosphorus as P0^~
Zinc
Total trace eleme-iits Wg, Se, As,
Pb, Cd, Cr, Co, Ki)
Carbonate, calculated from Ca
Sulphate
Nitrate
Chloride
Total
Percentage proportion
of dry, foan !
40* / («)
29
7.9
3 .1
6.5
2.1
0.5
0.6
0.2
0.04
11.9
not detectable
not detectable
not detect^xb \f?
99.8 ' "
-------
- 4O5 -
Analysis of this foam, given in Table 1, shows clearly that
the main component is organic and that the original
supposition of a high proportion of calcium carbonate does
not correspond to the observed facts.
Table 1 shows^ in addition, a considerable proportion of
silica, presumably due to the shells of smaller silicic
algae passing through the microfilter.
An additional check on the water body after every stage of
the treatment of the Lake Constance Water Supply Association
shows, with the aid of Fig. 5, that the freeze-dried filter-
able solids (TR), by the method described by Geller (29),
increase by about 20% after direct contact with ozone, and in
the next downstream intermediate tank are concentrated on the
surface in addition to the enrichment already mentioned.
TR mg/l
0.5-r
0,4--
Ca mg I
0.05
Raw water after after O3- after O3- pure water
micro- aa^hing reaction after sand
filtration chamber container filter
(6.9 g 03/m3)
TR'filterable dry solids.
— ——Ca- calcium fraction of TR
Fig. 5 "Microflocculation" in the treatment of Lake Constance
water
The proportion of calcium,determined from the filterable
solids, corresponds in order, of magnitude to the values
obtained in Table 1; a stoichiometric check shows that this
is not precipitation of calcium carbonate.
-------
- 406 -
What is much more likely is that a small proportion of the
organic acids formed after ozonization is precipitated in the
form of calcium salts.
These first observations on micro-flocculation suggest that
the above situation, which often occurs in lake waters and
has so far been unexplained, is somehow connected with the
phytoplankton content of the water and its metabolic products,
The following figure indicates that the metabolic products
of the phytoplankton probably play the leading role.
May/June 1976
Jan/Feb. 1977
Ozone dosage (g Oo/m )
Fig, jj Dependence of microflocculation in sand-filtered
Lake Constance water from a depth of 6O m on the
type of dissolved organic water constituents
In this experiment both phytoplankton and zooplankton had
been eliminated as a start by intensive sand filtration.
Microscopic examination of the starting water showed an alga-
free filtrate with ^ery good turbidity values of 0.07 form-
-------
- '407 -
azine-turbidity units. Only in water with a high initial
plankton content could larger turbidity increases due to
ozone treatment be observed, while in the water from full
circulation of the lake, with a low initial alga content,
practically no turbidity increase occurred at all ozone con-
centrations. It must be emphasized that the turbidity values
cannot be changed by acidification of the water with nitric
acid or hydrochloric acid, so that here too the influence of
calcium carbonate can be excluded.
It can also be seen from Fig. 6 that in the summer there is
an additional optimum of the ozone concentration, at which '•
these increases in turbidity occur.
If we start out from a water rich in inorganic turbidity but
low in plankton, then, as Fig. 7 shows, the initial turbidity
decreases with increasing dose of the ozone up tq a concen-
tration of about 1.5 mg 0-,/l.
oso-
_p
•H
tf
•H
f,
tt
0.35
Ozone reaction time 1.25 h
Experimental series 25.2.77
Exper iment al
series 1.3.77
Ozone dose (mg 03/1)
1.0
Fig. 7 Change in the raw water turbidity in dependence -on
the ozone concentration
-------
- 408 -
Similar results were obtained by Schalekamp (30} , who reported
on the example of Lake Zurich water that ozone reduced the
turbidity by about 20 - 40%, depending on the amount used.
To gain a better understanding of the 'flocculating action of
ozone, in the case of these Zurich investigations the part-1
icles were counted by size classes before and after the
ozonization.
Rwrtiole
frequency
before ozonizatioa
Total of particles 13. 040/ral
after ozonizatioa
Total of particles 8158/ml
0,7 1,0
Fig. 8
Microflocculation in the
treatment of Lake Zurich
water (after M. Schalekamp)
Rarticle siZB 0
As can be seen from Fig. 8, the smaller particles are reduced
after ozonization with 1 mg/1, while the larger particles
increase. Before the ozonization the total number of all
particles with a mean diameter of 2.35 ym was 13,040 and
after ozonization, with a larger mean particle diameter of
2.97 pm, it was only 8158. In ozonization with 2.5 mg 0^/1
the same phenomenon can be detected, but much more weakly.
-------
- 409 - _ .,,
Still higher ozone concentrations invert the picture, so that
after ozonization more particles are present than before.
These examples/ cited from the practice of drink'ing water
treatment make it clear that two opposing processes are at
work here, which make the interpretation of the phenomenon
of micro-flocculation considerably more difficult.
The input of energy into the water must also be regarded as
a factor influencing the formation of micro-flocculation.
With varied conditions of operation of a container into which
the ozonized water, is introduced for a secondary reaction,
by changing the proportions of the potential energy E = g-h
2
and kinetic energy E = 1/2 m-v , the remaining water prop-
erties remaining constant, a manipulatable change in turbid-
ity can be produced, as shown in Fig. 9.
6-i
E 5-
0-
2.2
-2.0"
C
-1.8*
•1.2-fi
1-1.0
Amount of raw water
required
Throughput through intermediate
container QSt
16 18
2? 2 75
2O ' 22 • 24
i
8 hours
Fig. 9 Microflocculation in ozonized Lake Constance water
-------
• - 410 - -
Since in this example too no precipitate of calcium carbonate
can be detected analytically, which could be easily accounted
for by the different turbulence within the container, it is
certain that the turbidity changes observed are due solely
to the superimposition of two opposing processes and their
dependence on the energy input.
4.3. Results from model experiments
There is no doubt that in the interpretation of experimental
results of this kind in practice which, as already mentioned,
also depend on the season and on the general limnological
state of the waters, a somewhat unsatisfactory feeling almost
of uncertainty is obtained. It therefore is absolutely
necessary to confirm the results relating to actual practice
by aimed model experiments. The main difficulty in an effort
of this kind lies in achieving comparability of the experi-
mental conditions between the practice trial and the model
experiment, which is of special significance in the case of
the organic water constituents.
We now know that it is not possible to characterize the
natural organic constituents of a surface water so precisely
that a synthetic aqueous solution of these could be made up.
Therefore, the preparation of model solutions is always
directed at the enrichment and isolation of these substances
which during the enrichment phase are changed by chemical
reactions with one another or by redox processes in such a
way that the isolated substance is no longer identical with
the substance originally present (7).
This fact alone explains why until now it has only been
possible in a few cases to achieve by means of ozone treat-
ment a perceptible increase in turbidity on a model water
enriched with 30 mg/1 of hymatomelanic acid, and hence a
conversion of the dissolved substances into the colloidal
state, as shown in Fig, 10.
-------
- 411 -
30mg/l UMA
without, CaCU-adc*i't:ioj:l
0,1 0,5
Degree of ozoBissation'fmgOj/mg acidj
Fig. 10 Microflocculation of hymatomelanic acid (HMA) by O.
(after R. Kurz) -
The opposite effect, i.e. a reduction of "the turbidity of a
particle-rich water enriched with huraic acid by ozone treat
ment, has been repeatedly demonstrated, as Kurz's results
with kaolin suspensions (31) show in Fig. 11.
-------
- 412 -
KXJmg/l K*o)in
Smmof/i C»CI2
o 0.2 0.4 o.e as
specific ozone dosage (mg 03/mg HA)
Fig. 11
Destabilization of a
kaolin suspension by
Lake Constance humic
acids (HA) after
ozone treatment
To explain these relationships in the two model trials,
differing from the relationships observed in practice, we
can only assume that in the formation of filterable substances
which increase the turbidity (case 1) from dissolved substances,
in the first place the entire molecule structure plays the
decisive part, while the turbidity-diminishing action (case 2)
depends mainly on the number and type of the functional groups
and less on the basic skeleton of the carrier molecule. To
the first case is added the possibility of the formation of
insoluble alkaline earth metal salts, which is also structure-
dependent. Summarizing case 1, it must be emphasized that
in contrast to true flocculation, these processes should be
regarded rather as precipitation reactions.
Case 2 can be explained by the ozone action on organic water
constituents described in the first chapter: non-ozonized
humic acids are adsorbed on the mineral particles forming the
turbidity and endow them with an increased suspension capac-
ity by a kind of protective-colloid action. If the polarity
of the organic water constituents is increased by the ozon-
ization, which manifests itself in a considerable increase
-------
- 413 -
in the number of carboxyl groups, while on the other hand
the size of the molecule is decreased, the adsorption
capacity on the turbidity particles immediately rises. At
the same time, however, after ozonization substances are
more strongly adsorbed on the turbidity material which, owing
to their greater number of functional groups, now has essen-
tially greater possibilities for cross-linking by bridge-
formation. The fact that this "polyelectrolyte character"
of ozonized organic constituents is noticeable, even with a
relatively small dose of ozone, and in the case of the humic
acids of Lake Constance reaches its optimum with ozone doses
between 0.8 and 1.0 mg/1, indicates that the molecular size
and therefore also steric factors exert an influence that
should not be neglected.
According to Stumm's classification of the agglomeration of
colloidal particles, case 2 comprises true flock formation
in which the destabilization mechanism is designated as floccu-
lation.
5. Summary
Ozonization of the organic constituents of water gives rise
to more polar compounds richer in oxygen and poorer in
double bonds, with increased numbers of hydroxyl, carbonyl,
and carboxyl groups and lower molecular weights. This ozone-
induced change in structure is performed by two simultaneous
and opposing processes.
The first case - the production of precipitating products
from originally dissolved organic compounds - is probably
strongly dependent on the structure of the entire molecule.
v
This includes the precipitation of insoluble alkaline earth
metal salts of organic acids. Particularly in the case of
lake waters there are indications that special metabolic
-------
- 414 -
products of the phytoplankton are made precipitable by
ozonization. This phenomenon is observed in practice as a
turbidity increase after ozonization.
The opposing process - destabilization of turbid suspensions
by ozonized organic water constituents -r should clearly be
regarded as true flocculation, the ozonized organic substances
adsorbed on the turbidity particles contributing by bridge-
formation to cross-linking of the particles as a result of
their polyelectrolyte character. A decisive part is played
here not so much by the entire structure of the organic
constituents as by the nature and quantity of their functional
groups.
In water-treatment practice the two processes presumably
take place in opposing directions and, in addition, are depend-
ent on turbulence. The reported investigations are partly
the results of a research programme on the optimization of
the ozone process, promoted by the Water Economy Board. Some
of the methods of characterizing organic water constituents
were newly developed within the framework of a research
programme, promoted by Dechema, for influencing corrosion
processes with the aid of organic water constituents.
-------
-415-
(1) AUTORENKOLLEKTIV
Organikum, Organisch-Chemisches Grundpraktikum:
Ozonisierung
VEB Deutscher Verlag der Wissenschaften, Berlin (197O),
295-299
(2) REICHERTER, U.-F.
Untersuchungen iiber die Anwendung von Ozon bei der
Wasser- und Abwasserreinigung
Dissertation, Fakultat fur Chemieingenieurwesen,
Uniyersitat Karlsruhe (1973)
(3) MAIER, D.
Wirkung von Ozon auf geloste organische Substanzen
Vom Wasser £3 (1974), 127-16O
(4) EBERLE, S.H., SCHWEER, K.H.
Bestiramung von Huminsaure und Ligninsulfonsaure im
Wasser durch Fliissig-Fliissigextraktion
Vom Wasser 4J_ (1973), 27-44
(5) MAIER, D., FUCHS, F., SONTHEIMER, H.
Bestimmung von organischen Sauren in Wassern und auf
Aktivkohle
gwf-Wasser/Abwasser 117 (1976), 2, 7O-74
(6) WAGNER, I., HOYER, O.
Die Bestimmung von Huminsauren und Ligninsulfonsauren
in Wassern mittels UV-Spektroskopie
Vom Wasser 45 (1975), 2O7-216
(7) MAIER, D.
Temperaturprogrammierte Decarboxylierung organischer
Wasserinhaltsstoffe
Vom Wasser 5O (1978) (in press)
(8) WEINDEL, W., MAIER, D.
Das Oxydationsspektrum, eine analytische Hilfe zur
Beurteilung von Verfahrensschritten der Wasseraufbereitung
Vom Wasser 5O (1978) (in press)
(9) MAIER, D., KURZ, R.
Untersuchungen zur Optimierung der Ozonanwendungibei
der Aufbereitung von Seewasser
Veroffentlichungen "Internationales Symposium Ozon und
Wasser" Wasser Berlin 77 (1977), 211-232
(1O) MAIER, D.
Werden Huminsauren durch Ozonbehandlung toxisch?
Vom Wasser, 5J_ (1979) (in press)
-------
- 416 -
(11) MAIER, D., M&CKLE, H.
Wirkung von Chlor auf natilrliche und ozonte organische
, Wasserinhaltsstoffe
Vom Wasser 47 (1976) , 379-397
(12) STIEGLITZ, L., ROTH, W. , KtlHN, W. , LEGER, W,
Das Verhalten von Organohalogenverbindungen bei der
Trinkwasseraufbereitung
Vom Wasser j47_ (1976), 347-377
(13) GOMELLA, C.
Ozone Practices in France
J. AWWA £4 (1972), 1, 39-45
(14) GOMELLA, C., HALLOPEAU, J.
Progre"s sur la recherche sur les effects de 1'ozone
sur 1'eau pendant le traitement et dans le reseau de
distribution
Publication 9th IWSA Congress New York (1972), 20-24
(15) TAYLOR, E.J.
'Weshalb Philadelphia Ozon nahm
Translation from "The American City Magazine" (1949)
(16) CROMLEY, J.T., O'CONNOR, J.T.
Effect of Ozonation on the Removal of Iron From a
Ground Water
J. AWWA 68 (1976), 315-319
(17) GROMBACH, P.
Die neue Ozonanlage des Bodensee-Trinkwasserwerkes
der Stadt St. Gallen
Monatsbulletin SVGW 4j| (1968), 7, 211-215
(18) CAMPBELL, R.M., PESCOD, M.B.
The Ozonization of Turret and Other Scottish Waters
Water and Sewage Works 113 (1966) , 7, 268-272
(19) FRANZ, J., GAGNAUX, A.
Entkeimung mit Ozon unter besonders hohen Anforderungen
Wasser, Luft und Betrieb 15 (1971), 11, 393-396
(2O) ROHRER, F.
Ozon und seine Verv/endung zur Wasseraufbereitung
Schweizer Brauerei-Rundschau 63 (1952), 135-155
(21) DIAPER, E.W.J.
A new method of treatment for surface water supplies
Water and Sewage Works 117 (197O), 11, 373-378
-------
- 417 •--•'••
(22) KOPECKY, J. '
Erfahrungen mit der Ozonisierung
Fortschritte der Wasserchemie 1_ (1967), 94-10O
(23) SUMMERVILLE, R.C., REMPEL, G.
Ozone for Supplementary Water Treatment
J. AWWA 6_4 (1972) , 377-382
(24) SONTHEIMER, H. .
Erfahrungen beim Einsatz von Ozon in der Trinkwasser-
aufbereitung; 3. Ozon als Oxidations- und Flockungs-
hilfsmittel
Wasserfachliche Aussprachetagung des DVGW/VGW in
Wiesbaden (1971), Sonderdruck DVGW-Brosentire
"Wasseraufbereitung-Wasserzahler"
(25) KULCSAR-MESCERY, J.
'Wasseraufbereitung durch Ozonisierung
Gas, Wasser, .Warms 9_ (1955), 8, 193-195
(26) O1DONOVAN, p.C.
Treatment with Ozone
J. AWWA (1965), 9, 1167-1194
(27) WAGNER, I., KELLER, H., MULLER, R.
Untersuchungen zur Aufbereitung von Bodenseewasser
mittels Ozonung und Flockungsfiltration
gwf-Wasser/Abwasser 118 (1977) , 320-322
(28) WURSTER, E., WERNER, G.
Die Leipheimer Versuche zur Aufbereitung von Donauwasser
gwf-Wasser/Abwasser 112 (1971), 81-9O/193-199
(29) GELLER, W.
Zur gravimetrischen Bestimmung des Suspensagehaltes
von Gewassern
7. Bericht der Arbeitsgemeinschaft Wasserwerke
Bodensee-Rhein (1975), 146-151
(3O) SCHALEKAMP, M.
Die Erfahrungen mit Ozon in der Schweiz, speziell hin-
sichtlich der Veranderung von hygienisch bedenklichen
Inhaltsstoffen
gwf-Wasser/Abwasser 5J7 (1977), 9, 657-673
(31) .KURZ, R.
Untersuchungen zur Wirkung von Ozon auf Flockungs-
vorgange
Dissertation, Fakultat fiir Chemieingenieurwesen,
Universitat Karlsruhe (1977)
-------
- 418 -
PRACTICAL USES OF OZONE IN DRINKING WATER TREATMENT
E.G. Rice, C.M. Robson, G.W. Miller and A.G. Hill
Abstract
Ozone has been used continuously for the treatment of
drinking water since 19O6 in the city of Nice, France,
where it was first installed for disinfection purposes.
Since 19O6, the uses of ozone for water treatment have
grown to include various chemical oxidation processes,
in addition to bacterial disinfection and viral inacti-
vation. Applications for ozonation now include oxida-
tion of inorganic materials (such as sulfides, nitrites,
cyanides, ferrous and manganous ions), organic materials
(such as phenolics, detergents, pesticides, taste and
odor—causing compounds, color-causing organics/ other
soluble organics), turbidity or suspended solids floccu-
lation (by changing the surface characteristics)t micro—
flocculation and recently to promote aerobic biological
processes conducted in filter media. Each of these uses
of ozonation is discussed in terms of their chemistries
and extent of application in the more than 1,OOO drin-
king water treatment plants known to be using ozonation,
most of which are located in Europe.
1. History of Ozone Use in Water Treatment (1)
The earliest experiments on the use of ozone as a germi-
cide were conducted by de Meritens in 1886 in France,
who showed that even dilute ozonized air will effect the
sterilization of polluted water. A few years later (1891),
the bactericidal properties of ozone were reported by
-------
- 419 -- :
Fr5hlich from pilot tests conducted at Martinlkenfeld in
a drinking water treatment plant erected by the German
firm of Siemens & Halske. In 1893, the first drinking
water treatment plant to employ ozone was erected at
Oudshoorn, Holland. Rhine River water was treate'd with
f
ozone, after settling and filtration. Siemens & Halske
next built treatment plants at Wiesbaden (1901) and
Paderborn (1902) in Germany which employed ozone.
A group of French doctors studied the Oudshoorn plant
and its ozonized water and, after pilot testing at
St. Maur (in Paris) and at Lille, a 5 mgd plant was con-
structed at Nice, France (the Bon Voyage plant), which
employed ozone for disinfection. Because ozone has been
used continuously at Nice since the Bon Voyage -plant be-
gan operating in 1906, Nice is referred to as "the birth- '
place of ozonation for drinking water treatment".
Full-scale water treatment plants then were constructed
in several European countries. As of 1916 there were 49
treatment plants in Europe having a total capacity of
84 mgd (2) in operation, and 26 of these were in France.
By 194O the number of drinking water treatment plants
throughout the world using ozone had risen to 119 , and
as of 1977 at least 1O43 plants, mostly in Europe (Table 1)
are known to be using ozone for drinking water treatment
(3). As might be expected, most of the European drinking
water treatment plants using ozone are in France, although
Switzerland and Germany account for most of the remainder
of the European plants.
y
In the United States, only one plant has been using ozone
continuously since the early 1940s (Whiting, Indiana)(4,5).
In 1973 the second U.S. plant to use ozone went on stream
at Strasburg, Pennsylvania' (6) . The remaining ..three plants
became operational in late 1977 or early 1978 (7).
-------
- 42O -
TABLE 1 Operational plants using ozone - 1977
Country
France
Switzerland
Germany
Austria
Canada
England
The Netherlands
Bel gi urn
Poland
Spain
USA
Italy
Japan
Denmark
Russia ...
Norway
Sweden
Al geria
Syria
Bulgaria
Mexico
Finland
Hungary
Corsica
Ireland
Czechoslovakia
Singapore
Portugal
Morocco
Number of Plants
593
150
136
42
23*
18
12
9
6
6
5
5
4
4
4
3
3
2
2
2
2
1
1
1
1
1
1
1
1
Total 1039
Includes expansions. Actual number of operating plants
in Canada = 20, with 3 more under construction.
-------
421 . -
The fact that the U.S. _Environmental Protection Agency
recently has proposed regulations for the larger U.S.
water supply systems regarding control of organic che-
micals in U.S. drinking water supplies has prompted a
keen interest on the part of the North American water
supply industry to know more about the uses of ozone and
its engineering parameters than had been known in late
1974. In response to this interest, the U.S. EPA funded
a survey, conducted by Public Technology Inc., of
Washington, B.C., to assess the .state-of-the-art of the
use of ozone and of chlorine dioxide for the treatment
of drinking water. This assessment involved a significant
questionnaire survey of many European drinking water
treatment plants, in which many of you participated. In
May, 1977, the PTI survey team visited some two dozen
European drinking water treatment plants using ozone
and/or chlorine dioxide. In August 1977, the PTI survey
team visited nine plants in the Province of Quebec, Canada,
The present paper is taken from the results of this EPA-
funded study.
2. Applications of Ozone in Water Treatment
Ozone is a powerful oxidant (Table 2). In acid solution,
the oxidation potential of ozone (2.O7 volts) is second
only to that of elemental fluorine among the commonly
used oxidants for drinking water treatment. Because many
contaminants in raw water supplies are oxidizable, ozone
can be and is being used for many different applications.
The major uses for ozone in modern drinking water treat-
ment processes are listed in Table 3. Although the early
uses for ozone in treating drinking waters were predomi-
nantly for disinfection (bacterial kill and viral inac-
tivation), today oxidative applications account for a
significantly increasing number of installations.
-------
- 422 -
TABLE 2 Oxidation-reduction potentials of water treatment agents
F2
°3
H2°2
MnO.
4
HC1O
Mno4
HOC1
ci2
HOBr
°3 +
cio2
Br2
HOI
cio2
cio-
HO2~
cio2-
OBr~
X2 +
I" 4-
oi- -
°2 +
REACTIONS
+ 2e = 2 F~
•f 2H+ + 2 e =0, + H_O
, * J.
•f 2H -f 2e = 2H2O (acid)
- -f 4H+ + 3e = Mn02 -f 2H2O
_ + 3H+ + 4e = Cl~ + 2H_0
+ 74-
- + 8H +5e = Mn + 4H2O
+ H+-i-2e - Cl~ + H20
+ 2e =2 Cl~
•f H+ + 2e =« Br~ + H2O
H2O + 2e = O2 + 2 OH~
(gas) + e = C1O2~
+ 2e = 2Br~
•f K+ + 2e = I~ + H2O
Caq) + e = C1O2~
+ H2O + 2e = Cl~ + 2OH~
-{- H20 + 2e = 3OH~ (basic)
~ + 2H2O 4- 4e = Cl~ + 4OH~
+ H20 + 2e = Br~ + 2OE~ • •
2a = 2 I~
2e = 3 I~
f- H2O + 2e = I" + 2OH~
2H2O + 4e = 4OH-
POTENTIAL IN VOLTS (Ec)
25 °C
2.87
2.07
1.76
1.68
1.57
1.49
1.49
1.36
1.33
1.24
1.15
1.07
0.99
0.95
0.9
0.87
0.78
0.70
0.54
0.53
0.49
0.40
"Handbook of Chemistry & Physics, 56fch Edition, 1975-76.
Press Inc., Cleveland, Ohio, p. D-141-143.
CRC
-------
- 423 -
TABLE 3 Applications of ozone in drinking water treatment
Bacterial Disinfection
Viral Inactivation
Oxidation of Soluble Iron and/or Manganese
Decomplexing Organically - Bound Manganese (Oxidation)
Color Removal (Oxidation)
Taste & Odor Removal (Oxidation)
Algae Removal (Oxidation)
Oxidation of Organics
- Phenols
- Detergents
-Pesticides
Microf1occulation of Dissolved Organics (Oxidation)
Oxidation of Inorganics
- Cyanides
- Sulfides
- Nitrites
Turbidity or Suspended Solids Removal (Oxidation)
Pretreatment for Biological Processes (Oxidation)
- On Sand
- On Anthracite
- On Granular Activated Carbon
To Make Treated Water Blue
2.1 Bacterial Disinfection
The French have pioneered the use of ozonation for bac-
terial disinfection. Guinvarc'h (8) reported that ozo-
nation at the Paris, France, St. Maur plant produced wa-
ters which never showed the presence of E. coli, although
the raw Marne river waters at the time showed coliform
counts of 15,OOO to 1OO,OOO units/liter, with average
values of 3O,OOO to 5O,OOO. O'Donovan (9) points out
that for bacterial disinfection, the usual ozone dosage
rate in water treatment plants at that time was I.5 to
2 mg/1.
-------
- 424 -
Miller et al. (3) have foundtfiat'the'' current average
ozone dosage rates in drinking water treatment plants
are 1 to 4 mg/1 today. Where preozonation treatment is
non-existent or when loadings of ozone-demanding organic
or inorganic materials are high, such as during pollution
episodes, ozone doses much in excess of 2 mg/1 can be re-
quired to attain the desired degree of bacterial disin-
fection. On the other hand, only O.25 mg/1 of ozone pro-
vided satisfactory disinfection at Boxley, England (1O).
In addition, the bactericidal action of ozone is.little
affected by changes in temperature or pH, and the disin-
fecting action of ozone is virtually instantaneous. Weaker
oxidants require more contact time (and usually higher
concentrations) to provide the same degree of bacterial
disinfection (9).
2.2 Viral Inactivation
In this application the pioneering work was conducted by
French public health officials. Coin and his coworkers
(11,12) used poliomyelitis virus Type I (1964) and Types
II and III (1967) to demonstrate that when an amount of
residual dissolved ozone equivalent to O.4 mg/1 can be
measured after 4 minutes of continuous ozonation, the
degree of viral inactivation surpasses 99.9 %.
Subsequent to this work in the late 1960s, the city of
Paris adopted the O,4 mg/1 residual ozone after 4 minutes
as a standard for the use of ozonation for viral inacti-
vation. Currently this.standard has been adopted through-
out France (13) ,
-------
- 425 -
In actual plant practice, . in. order to,, insure that these
minimum ozonation conditions are met consistently, the
French increase the ozonation time to at least eight mi-
nutes, and sometimes contact times of 12 minutes are.em-
ployed. It is general French practice to use at least
two ozone contacting chambers.. In the first chamber the
initial ozone demand of the water is satisfied and the
residual of O, 4 mg/1 of ozone., is attained; in the second
chamber the O.4 mg/1 of ozone residual is maintained.
Satisfying the ozone demand of the water in the first
contact chamber requires relatively large amounts of
ozone, and some 67 % of the total ozone dosed is applied
here. Lower amounts of ozone (about 33 %) are applied to
the second chamber to maintain the O.4 mg/1 of dissolved
ozone, usually for periods of 4 to .8 additional minutes.
At the present time, France is the only country known to
have formally adopted these ozonation conditions as a
treatment standard for viral inactivation. However, many
plants in countries outside of France have designed ozo-,
nation contacting systems for disinfection which utilize
the same treatment conditions (3). It is important to re-
cognize that under these viral inactivation conditions
of ozonation, bacterial disinfection also is obtained.
Ozonation of organic compounds usually produces oxyge-
nated organic materials which are more readily biodegra-
dable. In addition, ozonation does not readily oxidize
ammonia except at high pH (above 9). For these reasons,
ozonized waters containing these materials generally are
treated with small dosages .of chlorine, chlorine dioxide
or chloramine to prevent bacterial regrowths in the dis-
tribution networks.
-------
- 426 -
In Switzerland, it is common practice in many of the
15O drinking water treatment plants to follow ozonation
with filtration through granular activated carbon in or-
der to deozonize the water, then to add chlorine dioxide.
Deozonation is considered necessary because of reactions
between ozone and ClO^ in aqueous solution (14).
Masschelein (15) points out that ozonation of solutions
of chlorine dioxide or sodium chlorite in the mg/1 range
produces chlorate stoichiometrically:
NaClO2 + 03 > .NaClO3 + O2
The dimerization of C1O2 to C12O4 is considered to be
the first step in the ozonation process (16). The reac-
tion is very fast and is almost controlled by the diffu-
sion rate of ozone (17).
Many European, Canadian and the Strasburg, Pennsylvania
plants employ ozonation as the terminal treatment disin-
fection step, however, without aftergrowth problems.
Miller et al. (3) have concluded that ozonation can be
used as a terminal treatment step only if all of the
following conditions are met simultaneously:
1) The distribution system must be clean and not subject
to leaks,
2) Dissolved organic carbon concentrations must be less
than O.2 mg/1,
3) Ammonia must be absent,
4) Water temperature must be low and
5) Residence time in the distribution must be short
(less than 1 day).
-------
- 427 -
.2.3 Oxidation of Soluble Iron & Manganese
Ferrous iron is oxidized rapidly by ozone to ferric ions,
which then hydrolyze, coagulate and precipitate according
to the following equations:
Fe+2 + 03 —» Fe+3
Fe+3 + H20 > Fe(OH)3
Similarly, manganous ions can be oxidized to manganic
ions, which then form insoluble manganese dioxide:
+ ? +4
Mn + 03 > Mn
Mn+4 + H20 > Mn(OH)4 > MnO2
However, over-ozonation of manganous compounds produces
the very water soluble permanganate ion, which is pink
in color:
+2 +4 -
Mn or Mn + 0_ j> MnO,
Permanganate is toxic and should be prevented from en-
tering water distribution systems. Its presence in dis-
tribution networks can lead to buildup of MnO2 scales.
Permanganate can be produced (or used) in the water treat-
ment plant but kept from entering distribution system net-
works by several procedures, all of which depend upon the
fact that it is a very strong oxidant (see Table 2) with
an oxidation potential of 1.68 volts, close to that of
ozone.
-------
428 -
Normally, ozonation of ferrous and manganous - compounds is'1
conducted early in the water treatment process and the
hydrolyzed and precipitated inorganic hydroxides are fil-
tered. If the filtered water is distinctly pink in color,
the operator is alerted to the fact that he is using too
much ozone, and the ozone dosage should be reduced. Tanks
which provide 15 to 3O minutes of holding time are used
in the Dusseldorf area. This allows the permanganate to
oxidize dissolved organic materials, thereby being reduced
to the insoluble manganese dioxide:
MnO
. + organics > MnO2 + oxidized organics
Alternatively, the pink water can be filtered through gra-
nular activated carbon, where the permanganate is quickly
reduced to MnO- in the first few centimeters of the bed or
column. The insoluble dioxide then is removed during rou-
tine backwashing of the activated carbon medium.
2.4 Decomplexing of Organically-Bound Manganese
When iron and manganese are present as free inorganic
cations, they can be oxidized readily by agents much weaker
than ozone — simple aeration is known to be effective,
for example. However, when manganese is present as orga-
nic complexes, as is the case when decaying vegetation is
present, then even chlorine is not powerful enough to
break down the manganese complexes. In such cases stron-
ger oxidants, such as chlorine dioxide or ozone, are used,
again as a preoxidation step so as to prevent the oxidized
manganese from being passed into the distribution system.
-------
- 429 -
2.5 .Color. Removal • ••, .'.-v ...'•• •• <• •• •' •
Usually the colors in drinking water are derived from the
decomposition of naturally occurring humic materials. Co-
lors usually are caused by the presence of unsaturated
organic moieties conjugated in the compounds (i.e., alter-
nating double and single bonds). Compounds containing
such conjugated groupings are referred to as chrbmophores..
Ozone is particularly reactive with unsaturated groups,
cleaving the carbon-carbon double bonds to produce ketones,
aldehydes or acids, depending upon the other substituents
on the carbon atoms affected, the amount of ozone and con-
tact conditions applied. As soon as the conjugation has
been disrupted by oxidation, the color will disappear. This
does not necessarily mean that all of the color-causing
organic compound has been converted to carbon dioxide and
water, however, but simply that the conjugated unsaturated
groups responsible for the original color have been des-
troyed .
In industrial areas where textile manufacturing or dyeing
is prevalent, organic dyestuffs sometimes are discharged
from these industrial installations and are found in raw
waters entering drinking water treatment plants. These
dyestuffs generally are polycyclic, highly conjugated or-
ganic materials, easily decolorized by a powerful oxi-
dizing agent such as ozone. As before, however, the de-
colorized water will still contain considerable dissolved
organic carbon, which is readily biodegradable.
-------
- 430 -
2.6 Taste and Odor Removal
As a general rule, taste and odor causing compounds are
organic in nature, although many inorganic sulfides also
are encountered which are highly odorous. Many of these.
organic compounds which cause unacceptable tastes and
odors are formed from natural vegetation during anaerobic
decomposition in the ground or in surface waters in which
the dissolved oxygen content may -be too low to support
aerobic colonies. Examples of such materials include ter-
pene derivatives, such as geosmine, and some alicyclic
and/or aromatic alcohols. The latter are classed as phe-
nols.
Odorous compounds containing unsaturation, such as phe-
nols, usually are readily oxidizable by ozone. On the
other hand, saturated organic compounds usually are oxi-
dized only slowly by ozone. Since the amount of ozone re-
quired to remove tastes and odors will vary depending
upon the specific offensive organic compounds present, it
is essential to conduct pilot studies with ozonation to
ascertain the most cost-effective relationship of ozone
dosage and contact time to cope with the specific local
problem. As before, destruction of organic compounds which
cause tastes and odors by oxidation does not necessarily
insure total oxidation of those organic compounds to CO~
and water.
2.7 Algae Removal
During seasonal periods of climate changes and when the
proper nutrient balances are present in the raw waters,
algae growths are promoted. As these plants grow, their
metabolism produces by-products which can cause offen-
sive tastes and odors. In addition, the presence of large
-------
- 431 -
amounts of growing algae in the water treatment plant will
clog filters, requiring more frequent backwashing. Ozona-
tion will disrupt the metabolic processes of many types
of algae by oxidizing the essential organic components.
In treatment plants where ozonation already has been in-
stalled for other purposes (such as disinfection or iron
and manganese oxidation), seasonal blooms of algae in the
raw waters are handled simply by increasing the preozona-
tion dosages until the bloom periods are completed. This
requires installing sufficient ozone generation capacity
initially to cope with this problem.
2.8 Oxidation of cyanides, sulfides and nitrites
Toxic cyanide ions are readily oxidized by ozone to the
much less toxic cyanate ion:
CN~ + O3 > CNO~
At low or high pH, cyanate ion hydrolyzes to produce CO-
and nitrogen:
CNO
Sulfide ion is easily oxidized to sulfur, then to sulfite
and, finally, to sulfate:
S~2 + 03 - > S° — } S03~2 - > S04~2
The degree of 'oxidation attained depends upon the amount
of ozone employed and the contact time. Organic sulfides
will oxidize"to sulfones, sulfoxides and sulfonic acids
upon ozonation at slower rates than the sulfide ion it-
self.
-------
432 -
Nitrite ion is readily oxidized to nitrate ion by ozone
Organic nitrites, nitroso compounds, hydroxylamines and
the like also will be oxidized, first to the correspon-
ding nitro compounds. These then will decompose upon con-
tinued ozonation, liberating nitrate ions and carbonaceous
compounds. As before, the degree of oxidation will depend
upon the specific compounds present, the amount of ozone
employed and the contacting conditions employed.
2.9 Oxidation of Organics
Although there are literally myriads of organic compounds
present in water supplies (at the latest count, the U.S.
EPA has identified over 70O individual organic compounds),
not all are oxidized at the same rates upon ozonation. In-
deed, many highly halogenated organic compounds are not
oxidized at all under ozonation conditions normally encoun-
tered in drinking water treatment plants. Rice & Miller (18)
discuss these aspects in reviewing the nature of organic
oxidation products formed upon reaction with ozone under
aqueous conditions.
During the PTI survey of drinking water plants using ozo-
nation, many plants responding to the questionnaires re-
ported that they are using ozone for "organics removal".
However, in most cases the organics removed are not iden-
tified. Nevertheless, some specific organic materials
known to be readily "removed" (oxidized) by ozone have
been identified, and these include phenols, detergents
and certain pesticides.
-------
- 433 -
2.9. r "Phenols'' ' '
Phenols, especially the non-chlorinated phenols, generally
are readily oxidized iipon ozonation. It is important to
recognize, however, that destruction of the phenolic com-
ponent requires much less ozone than conversion of all of
the phenolic compound to C0? and water.
Eisenhauer (19) ozonized aqueous solutions of phenol for
3O minutes (until phenol was "destroyed" by the analyti-
cal test used) and isolated catechol, hydroquinone, p-
quinone, cis-muconic acid, oxalic acid and fumaric acids
as organic oxidation products. After 4 moles of ozone had
been consumed per mole of phenol, substantially all of the
phenol originally present had disappeared, but very little
CO2 had formed.
In later work, Eisenhauer (2O) showed that as ozonation of
phenol solutions proceeds, no CO2 is formed until after
1.5 moles of ozone/mole of phenol is consumed. However,
after 33 % of the theoretical CO,, had formed, CCU produc-
tion then ceased. He concluded that if destruction of the
aromatic ring is sufficient to solve the particular local
problem, then 98 % of the phenol can be "destroyed" using
5 moles of ozone per mole of phenol present. However, 67 %
of the original carbon present in the phenol still is pre-
sent in the form of other organic compounds which are oxi-
dation products of phenol.
The oxidation of phenol itself proceeds first through
di- and trihydroxyaromatic compounds and quinones.- Con-
tinued ozonation breaks the aromatic ring, forming ali-
phatic acids, the most stable end product usually being
oxalic acid.
-------
- 434 -
Phenols with aliphatic substituents on the ring are first .
oxidized to benzoic acids, then to hydroxy benzoic acids,
after which the aromatic rings are ruptured. As oxygena-
tion proceeds, the oxidized intermediate compounds be-
come more readily biodegradable.
Hillis (21) studied the oxidation of 14 phenols with ozone
over the pH range 4 to 10. Starting with 3O mg/1 concen-
trations of these phenols and ozonizing over 4 to 12 mi-
nutes, phenol concentrations were lowered to O.1O mg/1.
However, the corresponding COD values were lowered by only
5O %. This is further indication that although the speci-
fic phenol is destroyed by ozonation, oxidized organic
products remain in solution.
Nevertheless, the elimination of measurable concentrations
of phenols by ozonation is practised successfully at a
great many drinking water treatment plants (3).
2.9.2 Detergents
Detergents generally fall into two classes, linear alkyl
sulfonates and linear alkylbenzene sulfonates. There are
some aliphatic quaternary amino compounds which are used
as detergents, however, most household laundry detergents
are of the first two types. The more readily biodegradable
detergents will decompose in sewage treatment plants or in
rivers and streams, given sufficient time. However, many
raw water supplies are contaminated from time to time by
detergents or their partial decomposition products.
As discussed earlier, those compounds containing aro-
matic groupings will be more readily oxidized by ozone,
while the aliphatic materials will be less reactive. In
-------
- 435 -
addition,•-••Gilbert (22,23,24) has shown that even aromatic !
compounds containing sulfonic acid groups will be less
readily oxidized by ozonation than the same compounds
without the sulfonic acid groupings.
Gilbert (23) also has shown that in ozonizing pure aqueous
solutions of compounds over the pH range of 3 to 7, 1 kg
of COD can be removed from solution with 1.2 kg of ozone.
On the other hand, in more polluted wastewaters, 2 to 5
kg of ozone are required to remove 1.kg of COD from so-
lution.
In addition, the excellent work of Hoigne (this meeting)
and of Hoigne & Bader (25,26 and prior references cited
therein) on determination of reaction rates of specific
organic compounds with ozone will allow the practising
water treatment plant engineer to ascertain the ease of
oxidation of specific organic compounds that he may find
in his raw waters.
When ozonation already has been installed in a water treat-
ment plant for some other purpose, the presence of signi-
ficant quantities of detergents in the raw water is easily
recognized by the sudden foaming of the waters during pre-
ozonation or in the first contacting chamber when ozone
is used for disinfection. In such instances merely increa-
sing the ozone dosage usually copes with the problem. As
with seasonal algae blooms, it is important that sufficient
ozone generation capacity be available in the plant.
-------
- 436 -
2.9.3 Pesticides
Pesticides cannot be classified so simply as to their
reactivity with ozone. Phosalone and aldrin, for example/
are readily oxidized to destruction with small amounts of
ozone. On the other hand, dieldrin, chlordane, lindane,
DDT, PCBs, PCP and endosulfan are only slightly reactive
with ozone under normal ozonation conditions encountered
at drinking water treatment plants. Little or no removal
of these pesticides can be expected with ozone, or with
other oxidants.
Malathion and parathion represent two unique cases of
pesticides which are oxidized to destruction by ozonation,
but which proceed through intermediates (their correspon-
ding oxons) which are more toxic than are the starting
thions (27).
.PSCH-COQC-H
CHoO
3 £.£.'.
(malathion)
^WKljL/0
(CH3CH2/
(parathion)
°3 ,
«f
t MH
0
CH3OJ| 03
ta. ^ *ocf*i i cnnr* 1 1 ^ v^
Jt 5Ln-LUULpHc ^
pij n PU pnnr* u
un-jw onov/uuortni-
(malaoxon)
°3 2 /=r
*fc» Cm nr \ rt n /
(paraoxon)
-NO,
decomposition
products
\
decomposition
products
Hoffman & Eichelsd6rfer (28) showed that heptachlor is
oxidized "quantitatively to destruction" with ozone but
that heptachlorepoxide is stable to ozonation. This raises
the question as to whether ozonation of heptachlor pro-
duces heptachlorepoxide, which itself is a very toxic ma-
terial.
-------
- 437 -
Cl H Cl
heptachlor
0
heptachlor epoxide
Thus it is incumbent upon a water supply system to iden-
tify the specific pesticides with which it must cope,
then to ascertain the most effective method of removing
them. If ozona'tion is a part of the water treatment pro-
cess, certain pesticides can be removed effectively, but
others will require a different treatment sequence.
2.1O Suspended Solids Removal (Turbidity)
Turbidity is caused by suspended solids, which are small,
colloidal sized particles having surfaces which are highly
charged. The strength of these surface charges, in fact,
keeps the particles in suspension because of the repul-
sive surface forces coupled with the small particle sizes.
Such colloidal particles normally pass through filters
and are not retained.
In some instances addition of a strong oxidant, such as
ozone, will change the nature.and/or extent of these sur-
face charges, thus allowing the charged particles to ag-
glomerate and be more readily"removed by subsequent fil-
tration. If ferrous iron ions are 'present, ozonation will
oxidize these to the ferric state, as discussed earlier.
-------
- 438 -
As these trivalent ions hydrolyze and agglomerate, they
interact with the surfaces of suspended solids, floccu-
lating them and allowing them to be removed from suspen-
sion by filtration.
The use of ozonation alone can accomplish flocculation
without the addition of soluble ions which remain in so-
lution when chemical flocculants are used. A unique exam-
ple of the use of ozone to reduce suspended solids is the
recently commissioned Chino Basin sewage treatment plant,
near Los Angeles, California in the United States. This
secondary treatment plant is required to discharge an
effluent which is very low in suspended solids and in to-
tal dissolved solids, because the water table in the area
is very close to the surface. Simple filtration does not
lower the suspended solids content sufficiently.
Were it not for the stringent TDS requirement,, Chino Basin
could treat its secondary effluent with alum or other
standard chemical flocculating agents. However, such treat-
ment in this case would require an additional step of re-
moving the soluble chemicals after flocculation, coagula-
tion and filtration. The cose of such multiple treatment
is high and the process chosen was the single step of
ozonation. An ozone dose of 1O mg/1 provides the required
coagulation of suspended solids so as to allow their ready
removal by filtration. This 3 mgd plant went on stream in
early 1978 and the process is operating satisfactorily (29)
2.11 Microflocculation
As discussed earlier/ during oxidation of dissolved orga-
nic materials with ozone, oxygen is introduced into many
of the carbonaceous sites in the molecules. Carboxylic
-------
- 439 -
;, - . < •:-'.•//•'••..• - -, . ', ' ••'.-.-••• :•-.--- - ;•..'• -.-r.~- -ji';'r ' i1/.
acids, aldehydes, ketones and alcohols are produced, all
of which are more highly polar than are the non-ozonized
compounds. These polar compounds are capable of hydrogen
bonding, which effectively increases their apparent mole-
cular weights. In addition, the simultaneous presence of
polyvalent metallic cations, such as iron and aluminum,
with these polar organic groupings leads to flocculation
of the oxidized organics. Therefore the ozonation of clear
waters containing dissolved organic compounds can lead to
an increase in turbidity. This process has been termed
"microflocculation" by water chemists, and has been amply
described in the preceding paper by Dr. Maier.
It is because of microflocculation that many water treat-
ment plants insist upon following ozonation with a fil-
tration step. The plant at Langenau, Germany, follows
ozonation with a second addition of ferrous iron com-
pounds to increase the rate of microflocculation, and
produces a very clear water after subsequent filtration.
2.12 Ozone Pretreatment for Biological Processing
It has been pointed out in many examples given above that
the partial oxidation of organic compounds with ozone
renders them biologically degradable at a faster rate
than before oxidation. This fact leads to the conclusion
that if raw waters to be ozonized (or treated with any
other oxidant) contain high amounts of dissolved organic
carbon and those carbonaceous compounds are not readily
removed by flocculation and are not readily oxidized to
CO^ and water by ozone, that ozonation will produce an
aqueous medium high in dissolved oxygen and containing
dissolved organic compounds which are more readily assi-
milated by biological organisms. This reasoning has led
-------
- 440 -
to the concept of incorporating .biological, treatment^
steps following ozonation. Since ozonized aqueous media
are conducive to the development of aerobic bacteria,
not only can the carbonaceous compounds be expected to
be degraded biologically, but ammonia also can be ex-
pected to be converted to nitrate by biological nitri-
fication processes.
Biological processes promoted by preoxidation (or by pre-
aeration or preoxygenation in certain instances) can be
incorporated into the water treatment process by addition
of sand or anthracite filters and/or granular activated
carbon filters. It has been known for many years that
slow sand filters and activated carbon units contain
high degrees of biological activity. Following ozonation
with sand or anthracite filtration (to remove flocculated
materials), then with GAG filtration has been incorporated
into the Rouen-la-Chapelie plant in France and into the
Dohne plant at Miilheim, Germany, specifically to remove
ammonia biologically as well as dissolved organics. This
technique thereby avoids (Rouen) or eliminates (Dohne)
the need for breakpoint chlorination, with its attendent
production of halogenated organic compounds which should
be removed later in the treatment. The process has been
termed Biological Activated Carbon by Rice et al. (3O),
and is being studied at many water treatment plants.
One of the promising secondary benefits of biological
activated carbon is the greatly extended useful life of
the GAG filter media because of the bacterial activity
which is promoted by preozonation (or preaeration or pre-
oxygenation in certain instances). The Rouen plant has
operated satisfactorily since January, 1976 without having
to regenerate its 75 cm deep GAG beds. Later in this mee-
ting, Dr. Jekel will describe the BAG process as employed
-------
- 441 -
at the'Donne'plant, 'which'has not had to regenerate its
4 m deep GAC columns since they were installed in Novem-
ber, 1977. Expectations at Dohne are for at least two
years of GAC operational life before regeneration will
be required.
2.13 To Make Water Blue
The unique property of making finished water blue is one
of the primary reasons given for its use at Langenau,
Germany. This plant is located near the Danube River in
southern Germany, and processes mostly groundwater. Du-
ring dry seasons, however, Danube River water also is
treated by a series of physical chemical processes, in-
cluding two ozonation steps. Preozonation is employed
for suspended solids removal and to dispose of off-gases
from the primary ozonation step (for organics oxidation
and microflocculation). The ozonation process is con-
trolled partly by the shade of blue color which is im-
parted to the finished water. When the water is not suf-
ficiently blue, more ozone is added.
3. Multiple Applications of Ozone
It is important to recognize that even though ozonation
might be installed for a'single purpose, say iron and man-
ganese "oxidation, many other benefits can be derived from
its use. For example, at Rouen, 'Dohne, Wuppertal and
Langenau, preozonation (sometimes with high-speed tur-
bine contactors) aids in the flocculation process. Post-
ozonation at Rouen is for disinfection while the preozo-
nation is for manganese oxidation, organics oxidation and
preparing the following sand and granular activated car-
bon beds for biological conversion of ammonia and removal
of dissolved organics.
-------
- 442 -
At the three Diisseldorf plants (Flehe, Holthausen and.
Am Staad), Duisburg and Wuppertal, ozone's primary func-
tion is iron and manganese oxidation. At the same time
organics are oxidized and disinfection is obtained.
If ozone is applied for, say, color removal, near the
end of the treatment process, a significant amount of
disinfection also will be obtained. The conjunctive use
of contactor off-gases from the primary ozone contacting
chambers can be effective in such multiple ozonation
treatment processes. These off-gases (which contain as
much as 5 to 1O % of ozone)sometimes can be .recycled
economically to an early stage treatment step (as at
Rouen, Miilheim and Langenau) . Alternatively, the ozone
in these off-gases either must be destroyed (thermally,
catalytically, by passing through moist granular acti-
vated carbon) or diluted with air before being discharged
to the atmosphere. If the volumes of contactor off-gases
are not large, recycling them to an early stage oxidation
step in the total water treatment process can be cost-
effective.
4. Summary
Early application of ozonation in drinking water treat-
ment was primarily in France for bacterial disinfection.
In the late 1960s, French scientists defined the ozone
contacting time and dosages required for viral inactiva-
tion, and France since has adopted an ozonation treat-
ment standard for this purpose. Viral inactivation can
be achieved if a residual of O.4 mg/1 of dissolved ozone
can be measured at least 4 minutes after the initial
ozone demand of the water has been satisfied.
-------
- 443 -
Many other applications for ozone have evolved since
ozonation was installed in Nice, France (19O6), most
of which are based upon the high oxidizing power of the
gas. In Table 4 are listed the results of the PTI ques-
tionnaire survey of plants known to be using ozone in
the various countries of the world where ozone is being
used to significant extents.
In Figure 1 is shown a "conventional" drinking water treat-
ment process involving coagulation, sedimentation, fil-
tration and disinfection. The known uses for ozonation
are included at the various points in this conventional
process. Note that ozone is used at different locations
in the drinking water treatment process, depending upon
the purpose or purposes for which it is being used.
It is important to recognize that ozonation of dissolved
organic materials will rarely proceed to completion, e.g.
to produce CO2 and water. In most cases two effects will
be noted after ozonation of dissolved, organic materials:
1) Dissolved organics will be converted to more highly
oxygenated materials which can and do flocculate, re-
sulting in an increase of turbidity. In these instan-
ces, ozonation usually is followed by a filtration step.
2) Dissolved organics will be converted to more highly
oxygenated materials which are more readily biologi-
cally assimilable. Therefore, the higher the residual
DOC after ozonation, the greater will be the chances
for bacterial and slime growths in distribution sys-
tems, if additional disinfectants (such as chlorine, '
chlorine dioxide and chloramine) are not added after
ozonation. Biological treatment processes after ozo-
nation have been incorporated successfully into a few
plants, and the concept is being studied in many more
plants.
-------
TABLE 4 Uses of ozonation by country responding to PTI questionnaires
Ozonation Used For
Country
Great
Britain
Belgium
The
Netherlands
Austria
Switzerland
Germany
France
Canada
USA
No. of
Plants
Using 03
18
9
12
42
150
136
593
20
5
Question-
naires
Returned
6
1
7
5
9
31
64
18
-
Bacterial Viral
Disinfection (a) Inactivation
2 (0) 1
(0) 1
2 (1) 2
5 (4b) 2
6 (1) 5
27 (10) 8
60 (28) 37
13 (3) 9
1 (1)
Fe/Mn T/0 Color
Removal
2 1 6
1 1
Taste 6
Fe-1 Odor 4 6
1
both-4 2
Taste 1
Odor 1
8-both both-11
Fe-1 +
Mn-1 Taste-5 5
5-both both-31 21
fe-2 Taste-5
Mn-1 Odor- 1
both-15 3
taste-3
4-
Organics Turbidity
Oxidation
-
1
3 1
1
3
16 6
23 9
4
(phenols-2)
-
(a) No. of plants known to be using 03 as terminal step or sole disinfectant
(b) More Austrian plants known to be using 03 as terminal step (Dobias & Starz, 1977).
-------
deconplexing organic-Mn
pretreatment for
biological processes
Fc & Mn oxidation
f loccu.lntion
algae removal
cinsstruction of
off-gas ozone
• pretr.eatment for
biological processes
organles oxidation
color removal
tastes & odors
- viral inactivation
- bacterial disinfection
Sand or
Anthracite
Filtration
influent
water
v
ci2, cio2
or C1NH2
for residual
To.
Distribution
To
Distribution
Fig. 1 Typical points of application of ozone in
drinking water processes
-------
TABLE 5 Major advantages and disadvantages of ozone
Advantages
Powerful Qxidant
Powerful disinfectant and virucide
over wide temperature and pll range
Mr preparation, ozone generation «nd
contacting systems are easily automated
— but can be controlled manually
Generated en-site as needed -- operating
costs 2-1 U.S. i/1,000 gallons
Disadvantages
Non-Selective Oxidant
Leaves no residual for protection
of network
Gas/liquid contacting is not a
general practice at water treatment
plant
Capital costs for generation and
contacting are relatively high
Safe -- shutting off electricity
ceases 03 generation
Converts many non-biological ly
liuyratltible organlcs to oxidation
products which are biodegradable
Ones'not produce halogenated organics
will rarely oxidize all organics to CO? *
water. Presence of Biodegradable Organics
requires subsequent biological treatment and/or
residual disinfectant, but at lower levels
Does not oxidize highly halogenated organics
Reduces amount of residual disinfectant
required for network
Adds dissolved oxygen to water
Does not react with ammonia below pH=9
Does not increase total dissolved sol Ids
CPl
I
-------
447 -•
In addition, it is also clear that because the capital
costs for ozonation installations are high and because
ozone reacts with a" wide variety of materials, ozone
should not be used to perform water treatment tasks that
other techniques can do as well (or better) but at lower
cost. For example, since suspended solids are easily re-
moved by chemical coagulation, this process and filtra-
tion should precede ozonation for'most other purposes.
The> major advantages and disadvantages of ozone in trea-
ting drinking water are listed in Table 5.
(1) RIDEAL, E.K,
Ozone
Constable & Co. Ltd., London, England (1920)
(2) VOSMAER, A.
Ozone, Its Manufacture, Properties and Uses
D. van Nostrand Co., New York (1916)
(3) MILLER, G.W., RICE, R.G., ROBSON, C.M., SCULLIN, R.
WOLF , H. , KtJHN , W.
An Assessment of Ozone and Chlorine Dioxide Technologies
for Treatment of Water Supplies
EPA Report No. 60O/2-78-147. U.S. Environmental Protection
Agency, Cincinnati, Ohio (1978)
(4) BARTUSKA, J.F.
Ozonation at Whiting, Indiana
J. AWWA 33 (1941), 11, 2035-2O50
(5) BARTUSKA, J.F.
Ozonation at Whiting (Indiana): 26 Years Later
Public Works, August (1967)
(6) HARRIS, W.
Ozone Disinfection of the Strasburg, Pennsylvania, Water
Supply System
Proc. First Internat. Symp. on Ozone for Water and Waste-
water Treatment. R.G. Rice & M.E. Browning (editors)
Internat. Ozone Inst., Cleveland, Ohio (1975), 186-193
-------
- 448 -
(7) LACY, W., RICE, R.G. ' , V ^ '.
The Current Status of Ozone Treatment Technology in the
United States
Presented at Special Seminar on Water Supply, Internat.
Bank for Reconstruction & Development, Washington, D.C.
Jan 11, 2978- See also Wasser Berlin 1977, AMK Berlin
Germany/Internat. Ozone Inst., Cleveland, Ohio, 1978
(8) GUINVARC'H, P.
Three Years of Ozone Sterilization of Water in Paris
Ozone Chem. & Technol., Adv. in Chem. Series, 21,
Am. Chem. Soc., Washington, D.C. (1959), 416-429
(9) O'DONOVAN, D.C.
Treatment with Ozone
J. AWWA $2 (1965), 9, 1167-1192
(1O) HARDEN, C.H.
The Boxley Works of the Maidstone Waterworks. Company
Trans. Inst. Water Engrs. 48 (1943), 152
(11) COIN.- L., HANNOUN, C., GOMELLA, C.
Inactivation of Poliomyelitis Virus by Ozone in the
Presence of Water
la Presse Med. 72 (1964), 37, 2153-2156
(12) COIN, L.. GOMELLA, C., HANNOUN, C., TRIMOREAU, J.C.
Ozone Inactivation, of Poliomyelitis Virus in Water
la Presse Med. 75f 38, 1883-1884
(13) SCHULHOF, P.
Private Communication, 1977
(14) HOIGNE, J.
Private Communication, 1978
(15) MASSCHELEIN, W.
Chlorine Dioxide (Chemistry and Environmental Impact
of Oxychlorine Compounds)
Ann Arbor Science Publishers, Inc., Ann Arbor,
Michigan (1978)
(16) SCHACK, C.J., CHRISTE, K.O.
Inorg, Chem. JJ5 (1974) , 378
(17) K3LLE, W.
Problem der gemeinsamen Anwendung verschiedener
Oxydationsmittel bei der Wasseraufbereitung
Vom Wasser 35 (1968) , 367-381
-------
_ 449 -
(18) RICE, E.G., MILLER, G.W. '- * '•'•'<"•• :-'
Reaction Products of Organic Materials With Ozone
and With Chlorine Dioxide in Water
Presented at Symp. on Advanced Ozone Technology,
Toronto, Ontario, Nov. 1977. Internat; Ozone Inst.
Cleveland, Ohio (1977)
(19) EISENHAUER, H.R.
The Ozonization of Phenolic Wastes > ' '
J. Water Poll. Control Fed. 4O (1968), 11, 1887-1899
(2O) EISENHAUER, H.R. •
Dephenolization by Ozonolysis
Water Research 5_ (1971), 467-472
(21) HILLIS, M.R.
The Treatment of Phenolic Wastes by Ozone
Presented at Third Internat. Symp, on Ozone Technol.,
Paris, May 1977. Internat. Ozone Inst., Cleveland,
Ohio (1977)
(22) GILBERT, E. •
Ozonolysis of Chlorophenols and Maleic Acid in Aqueous
Solution"
Proc. Sec. Internat. Symp. on Ozone Technology
R.G. Rice, P. Pichet & M.A.- Vincent (editors), Internat,
Ozone Inst., Cleveland, Ohio (1976), 253-261
(23) GILBERT, E.
Chemical Reactions Upon Ozonization
Presented at Internat. Symp. on Ozone and Water,
Wasser Berlin 1977, AMK Berlin/Internat. Ozone Inst.,
Cleveland, Ohio (1977)
(24) GILBERT, E.
Reactions of Ozone with Organic Compounds in Dilute
Aqueous Solution: Identification of Their Oxidation
Products
Ozone/Chlorine Dioxide Oxidation Products of Organic
Materials. R.G. Rice & J.A. Cotruvo (editors), Internat.
Ozone Inst., Cleveland, Ohio (1978), 227-242
(25) HOIGNE, J., BADER, H.
Ozonation of Water: Selectivity and Rate of Oxidation
of Solutes
Presented at Third Internat. Symp. on Ozone Technol.
Paris, May 1977. Internat. Ozone Inst., Cleveland,
Ohio (1977)
-------
- 450 -
(26) HOIGNE J., BADER, H. .. , , . , ,
Rate Constants for Reactions of Ozone With Organic
Pollutants and Ammonia in Water
Presented at Symp. on Advanced Ozone Technology, Toronto,
Ontario, Canada, Nov. 1977., Internat. Ozone Inst.,
Cleveland, Ohio (1977)
(27) RICHARD, Y., BRENER, L.
Organic Materials Produced Upon Ozonization of Water
Ozone/Chlorine Dioxide Oxidation Products of Org. Materials
Internat. Ozone Inst., Cleveland, Ohio (1978), 169-188
(28) HOFFMAN, J., EICHELSDORFER, D.
Zur Ozoneinwirkung auf Pestizide der Chlorkohlenwasserstoff-
gruppe im Wasser
Vom Wasser 3J5 (1971), 197-2O6
(29) NOWAK, T., TATE, C.H., MOUTES, J.G., STONE, B.C.,
TRUSSELL, R.R.
Full-Scale Tertiary Wastewater Treatment Plant
Presented at Ozone Technol. Symp./ Los Angeles, May 1977
Internat. Ozone Inst., Cleveland, Ohio (1978)
(3O) RICE, R.G., MILLER, G.W., ROBSON, C.M., KtlHN, W.
Biological /Activated Carbon
Carbon Adsorption, P. Cheremisinoff & F. Ellerbusch
(editors), Ann Arbor Science Publishers, Inc.,
Ann Arbor, Michigan (1978)
-------
- 451 -
THE USE OF OZONE IN THE TREATMENT OF DRINKING WATER
J. Chedal
I should like to comment on the present and future use of
ozone in the waterworks supplying the suburbs of Paris with
drinking water by the treatment of surface water.
For 10 years ozone has been used in the after-treatment after
a physical and chemical clarification. The main function of
this treatment is to guarantee a viricidal final sterility.
This aim is fulfilled by maintaining a residual ozone level
of 0.4 g/m in the water for 10 min. To achieve this,
according to the provisions described by R. Rice, the ozone
is added in consecutive chambers.
When the experiments which I described yesterday are completed
we shall convert this chemical plant into a biological plant
and so make use of ozone in the preliminary treatment of the
raw water before each chemical treatment. This new treat-
ment process makes it possible to achieve two different aims:
1. The preliminary break-point treatment with chlorine is
avoided. In the new process the ammonia is eliminated
biologically during the clarification. The advantage
of this process is that the formation of haloforms
during the preliminary treatment is prevented.
2. During the clarification phase the removal of organic
substances is increased. This reduction in their content
largely affects the precursors. Consequently, the final
chlorination for protection of the network gives rise
to only minimal haloform formation.
-------
- 452 -
filtration
sedimentation
chemical
treat men t
~-'^Ts=s*f--r^
1 |1 ..__..,.:...:_,.... K. _
ozone
production
r _\ v — e -r
The additional removal of organic substances achieved by
this method of ozonization is certainly partly due to the
coagulating action of ozonization. The various aspects of
this phenomenon have just been presented by Dr. Maier.
To put this new approach into practice the measures shown in
the adjoining scheme were taken in the largest waterworks
supplying drinking water to the Paris suburbs (capacity
900,000 m^/day). The structures for the preliminary ozon-
ization comprise a container for preliminary chemical treat-
ment, which is currently under construction.
The experiments have shown that a preliminary ozonization
with 1 ppm makes possible an equally large reduction of the
ozone dose in the after-treatment.
-------
- 453 -
Accordingly, no additional ozonization equipment need be
installed. The present ozone generator is adequate for the
after-ozonization plants,.and for .the preliminary ozonization.
The development I have just described shows how careful one
must be in the planning of a w cer-treatment works. It is
always desirable to make provisions for the incorporation of
additional processes in a treatment plant.
As regards the ozone, it is best if its generation and appli-
cation are kept separate. To enhance the quality of the
treatment it may be advisable to apply the ozone at several
points in the treatment line, particularly in"the preliminary
ozonization as in the present case.
-------
- 454 .-.
SYNERGISTIC EFFECT OF OZONE AND CHLORINE•ON BACTERIA
AND VIRUSES IN SECONDARY WASTEWATER EFFLUENTS
Y. Kott
Synergistic effects are prevalent in nature and are most-
ly studied as factors of growth inhibition; synergism
was also observed when gamma radiation and heat were used
together in order to decrease the number of microorganisms
found in wastewater sludge. The current study, supported
by joint German/Israeli Research Projects BMFT and NCRD
Project No. O13-714, was undertaken with a view to fin-
ding whether or not simultaneous application of chlorine
and ozone to secondary wastewater would involve in a sy-
nergistic effect. Various studies have shown that ozone,
chlorine and other disinfectants reduce microorganism
counts in secondary was-tewater effluents. In the present
study, chlorine at concentrations of 5-30 mg/1, applied
to a continuous flow sample in a column, showed reduction
of coliform at range of bacteria 1O /1OO ml by one order
of magnitude up to five,that is 0-2 bacteria in 1OO ml.
4
Poliovirus count (attenuated strain) at range of 1O /5
liter samples decreased by up to two orders of magnitude
at the most.
When ozone was applied at concentrations of 1O-25 mg/1
to secondary wastewater under the same experimental con-
ditions the count of coliform bacteria decreased from
5 4
10 /100 ml to 1O /1OO ml which is one order of magnitude,
4
and that of enteroviruses from 1O /5 liters to 3O which
is three orders of magnitude. Different concentrations
of these oxidation chemicals applied for different con-
tact times showed that chlorine has a much higher kill
efficiency for bacteria than for viruses, while the re-
-------
- 455"-'
verse effect was observed with ozone. In addition
Salmonella typhymurium a bacteria representing the pa-
thogens, and Tj bacteriophage representing other viruses
were examined. The results showed that ozone caused an
80 percent decrease for the bacteria and a 95 percent
decrease for Polio I attenuated viruse particles at equi-
valent experimental conditions. When the synergistic
effect of chlorine and ozone was studied, the chemicals
were applied separately, together and in sequence, all
on the same experimental set-up.
Definite synergistic effect was observed with a better
kill effect on bacteria and viruses. The order of se-
quential application made no difference. Further study
on the synergistic effect "'is underway for economic eva-
luation.
-------
- 456 - ,; .
EXAMPLE OF UV OZONE CONTROL - ELIMINATION OF RESIDUAL OZONE
J. Valenta
The equipment and apparatus available commercially at the
present time enable continuous monitoring of the ozone con-
centrations both in air-ozone or oxygen-ozone mixtures and in-
ozonized water.
As an example of suitable ozone monitoring I should like to
describe briefly — from our own practical experience - the
continuous monitoring of the residual ozone content in the out-
going air of a newly installed ozone eliminator in the lake
waterworks at Lengg.
Here we were able to detect at times certain ozone emissions,
lasting only a short time, particularly during the backwash of
the activated carbon filters. The situation is illustrated
in Fig. 1.
It was therefore decided to supplement the original thermal
ozone elimination by a catalytic decomposition unit, the
earlier plant additionally assuming the function of preheating
the exhaust gas upstream of the catalyst.'
After the installation of the new elimination plant the per-
formance was monitored with the aid of two continually running
ozone detectors based on the principle of UV measurements. One
instrument, with a 2 cm quartz cell,was placed at the entrance
(position 3), the other, with a 50 cm long cell, was placed
downstream of the catalyst (position 6). In addition, the
residual ozone concentrations at seven other points along the
air outlet were measured over a short period. A low-pressure
mercury lamp serves as the light source in these instruments,
providing light with almost 95% of the 254 nm wavelength.
-------
- 457 -':
ELIMINATION OF RESIDUAL OZONE
! Oj-outflow from ACF _
[UJ
1 "0"
ttfV Zun£h. AMSUS
1 1
Fig. 1 Control of the catalytic elimination of
residual ozone in the lake waterworks at Lengg
To obtain the desired flow-through velocity of the measured
gas mixture, a small ozone-resistant membrane pump and a
rotameter were connected; at-the second photometer (6).
The flow velocity was here ^practically a constant., between
0.6 and 0.8 1/min, in the course-of several days.
Fig. 2 shows two recorder charts with the data evaluation
before the elimination and after the catalytic decomposition.
The activated carbon filter .washes-,, during which an earlier
elevated ozone .concentration in the. outgoing air could be
detected, are also marked. At the input the ozone content
was at the maximum 3 mgO/lv corresponding to about 1500 ppm
-------
- 458 -
of ozone, while after the eliminator 0,01 mg of ozone/1
(5 ppm) was never exceeded. The large layer thickness in the
second instrument makes it possible to determine ozone concen-
trations down to'0.01 mg O3 per litre of outgoing air.
This figure demonstrates the very high performance of the
catalytic elimination of ozone and at the same time the fea-
sibility of using the UV method for this purpose.
Copy from photo-
meter charts
... - .
*'^^ -^^—Ji*^~= joo%-
-------
- 459 -
USE OF CHLORINE DIOXIDE FOR THE TREATMENT OF DRINKING WATER '
W.J. Masschelein
1. INTRODUCTION
Though the bactericidal properties of chlorine dioxide have
been known since the beginning of the century, the compound
has only been used in the field of water treatment since the
nineteen-fifties. It was when chlorine, which was in any
case fairly cheap, had been found not to be entirely satis-
factory when used on its own as a disinfecting agent, that
other agents such as chlorine dioxide began to be used.
Many large Water Boards, such as the Paris one (both city
and suburban), the Dusseldorf Board, and many others adopted
chlorine dioxide. The results obtained were and still are
very satisfactory, notwithstanding certain reservations
expressed by some users. Owing to conservatism and ignorance,
frequent attempts were made to use facile pretexts to cast
doubt upon the use of .chlorine dioxide: instability of the
reagent, handling risks, lack of a good method of analysis,
especially for residual quantities, appearance of residual
chlorite for which there was no reliable method of analysis.
It was during this period, and more precisely in 1968, that
the use of chlorine dioxide was adopted by the Compagnie
Intercoinmunale Bruxelloise des Eaux. In its time, this
decision required two years of preliminary research, both
documentary and laboratory investigation with technical
development. At present, as we know, the accumulation of
knowledge concerning the formation of organochlorine compounds
has renewed the interest in chlorine dioxide, as indeed in
other possibilities of water treatment.
-------
- 460 -
Since th,e time'..allotted-,-for.this, paper does not enable me,tp
develop exhaustively the prospects and the limitations of
chlorine dioxide, I shall refer to earlier monographs (1,2).
All the same, I shall try to highlight their most important
aspects, and above all the criticisms alluded to above.
2. PRACTICAL SYNTHESIS OF CHLORINE DIOXIDE
Chlorine dioxide is a gas that must be prepared in situ
where it is to be used. The synthesis reactions used in
treatment stations are as follows:
and
2 NaClO2 + C12 = 2 NaCl + 2 C1O2
5 NaClO2 + 4 HC1 = 4 C1O2 + 5 NaCl + 2
The oxidation of chlorite by persulphate is in most cases
limited to swimming pools:
2 C1O~ + S2Og = 2 C1O2 + 2
Synthesis starting from chlorates (1,2) is hardly ever used.
It is more suitable for production capacities much larger
than those that suffice for water treatment.
2.1. In the synthesis of chlorine dioxide from chlorite and
chlorine, a minimum concentration of chlorine is required to
enable sufficiently rapid formation of the chlorine dioxide
without any concomitant dismutation reaction that would give
chlorate according to the probable mechanism:
6 C102 + 3 H2O = 5' HC1O3 + HC1
-------
- 461 -• "«
In fact, "the" acidity of the-water'-due to" the 'chlorine serves^
mainly to neutralize the alkalinity introduced by the
stabilized chlorite. At initial chlorine concentrations of
the order of 0.4 to 0.5 kg.m" one must work with approxi-
mately equal weights of chlorine and sodium chlorite, namely
an amount of chlorine equal to 200 to 250% of the stoichio-
metric amount. With 1 to 1.5 kg of chlorine per m in the
initial solution the reaction attains a 90% yield, the final
pH of the soluiton being between 4 and 5. At higher
chlorine concentrations, up to 3 g.£ , a stoichiometric
mixture with sodium chlorite gives a yield greater than 95%,
at a final pH lying between 5 and 6. This solution also
contains up to 6 g of C1O2 per litre.
If one operates at these concentrations, the equilibrium
vapour pressure of the dioxide gives rise to a concentration,
in the gaseous phase, of 6 vol-% at 15°C and 10 vol-% at
30°C. After the reaction phase it is thus necessary to
dilute the solution if the reagent is to be preserved for
a certain time before being injected into the water to be
treated. This operation is advantageous mainly in that it
permits metering in proportion to the flow rate of the water
being treated. The dilution will keep the pH between 5 and
7.5, so that both acidic and alkaline dismutations will be
avoided. The final concentration attained is between 0.5
and 1 g of C1O2 per litre.
2.2. For the surplus (see Fig. 1) the following safety
measures should be taken:
automatic cut-off of chlorite in the event of insuffi-
cient chlorine;
- cut-off of gaseous chlorine in the event of
insufficient water
-------
- 462 -
provision of a dilution zone within the reactor;
ventilation of any storage tanks, which will also
avoid any build-up of pressure inside them;
safety vents on the pipelines leading the gases
towards the ClOp elimination plant: graphite plates
for the CIO™, or sodium hydroxide exchanger, depending
on the capacity of the installation;
all the joints and' connections of the reactor and
the plant must be made using flame-proof materials
and installation methods;
free ventilation, or preferably an extractor fan, for
the chlorite storage hopper.
NEUTRALISATION
OU CHIORE
DOSEUR COL0NNE
DE
CHLORE REACTION
RESERVOIRS 06 STOCKAGE
STATION 0£ FILTRATION
(mlcrotamisogt)
|MS $an0
gkslaus
* **«
-6-
9
O *»»«*«*•• » i*****1* ^u P,«4 «»
Sc'qixncts automatiquta - production cJu txaxydc de chEort
L«K Man *?u$t.ira 11
L "6f*5.E_2_J STOCKAGE DU CHLORITE DE SODIUM
P""*"
Ui- *«.»»* <
i-iMiWI* »8'fc, DC ,* »«OOiX:!iO"
Fig. 1 Plant for the preparation of ClO^ from
and Cl. Example: the Lienne station-
-------
- 463 -
:. . , . •• i > ••»..••• . '•' • v • : '•' ~' '
Reactors similar to,the one illustrated here and capable of
producing up to 7.5 kg/h of C1O2 have been operating for
several years at the Tailfer station.
2.3. The direct reaction involving stoichiometrie propor-
tions of chlorite and hydrochloric acid,and producing
M* "I
chlorine dioxide at concentrations of 0.5 to 2 g.£ , only
proceeds with a yield of 60 to 80%. Various secondary
reactions result in the formation of chlorine:
5 C1O~ + 5 H+ = 3 C1O~ + C12 + 3 H+ + E-^O
4 C10~ + 4 H+ - 2 C12 + 3 02 + 2 H2O
Moreover, the chlorous acid appearing as an intermediate in
the reaction can undergo dismutation according to:
4 HC1O2 = 2 C1O2 + HC103 + HC1 + H2O
Yields of up to 100% for C1O2 formation, i.e. the production
of four moles of dioxide from five moles of chlorite, may be
attained provided that the reagents NaClO,-, and HC1 are mixed
in equal weights, in the form of solutions with concentrations
of 7.5 wt-% of NaCl02 and 9 wt-% of HCl. These conditions
imply a final pH of less than 0.5.
This technique means that the commercial chlorite, at 24-
25 wt-% (300 kg.m~ ) must be diluted, an operation best per-
formed with softened or demineralized water so as to avoid
the precipitation of calcium carbonate and calcium silicate
(with magnesium oxide) in the solution used. To avoid these
difficulties, we at C.I.B.E. have developed a high-concen-
tration process based upon the direct use of commercial
300 kg.m sodium chlorite and acid diluted 1+6. Under
. these conditions the reaction is complete within about 10 min
and the d
25-30 g.fc"
and the dioxide is produced at a concentration of some
-1
-------
- 464 -
Pig. 2 General layout of-this type of plant
The yield depends on the design of the zone in which_ the
reagents are mixed within the reactor. A contact enclosure
in the shape of an inverted cone with tangential inlets for
the reagents gives excellent results. The reactor can be
reliably brought into operation by maintaining a reduced
pressure with an ejector that draws off both dissolved and
gaseous dioxide. For safe operation it is essential to
provide automatic shut-down in the event of an insufficiency
of the dilution water, by means of the ejector or by dis- •
continuing the reduced pressure. The other precautions,
such as ventilating the pipelines, fire-proofing safeguards
for the chlorite storage, etc. are similar to those for the
chlorite + chlorine reactors.
-------
- 465 -
2.4. Preparation in the laboratory
If 1.5 to 2 moles of acetic anhydride per NaCIO- are made to
react with an aqueous solution of sodium chlorite, a pure
solution of chlorine dioxide is obtained with a yield of 50%
(3). The remainder of the chlorite is transformed into
chlorate, together with traces of chlorides, while1the
acetic anhydride is converted into the acid and acetate.
Such as it is, this solution is satisfactory for most appli-
cations in the laboratory or on a small scale, in which the
impurities mentioned above do not interfere. If the dioxide
needs to be purified, this can be done by scrubbing with air
and redissolution. This operation is performed in the dark
to prevent any dismutation (4).
3. METHODS OF ANALYSIS
3.1. lodometry is still the best method of determination,
both for the chlorine dioxide and the chlorite, provir^d
that other oxidizing agents that can oxidize iodide to
iodine are absent.
At pH 7 the reaction is:
2 C102 + 2 I~ = I + 2 C10~
At pH 2 the chlorite reacts according to:
C10~ + 4 I~ + 4 H+ = 2 I2 + 2 H20 + Cl~
Under the practical treatment conditions the chlorine is the
major source of interference. Because of this, the iodo-
metric method is mainly suitable for calibration.
3.2. The simultaneous determination of several oxidants,
such as chlorine dioxide, chlorite, and chlorine, often
-------
- 466 -
entails an alkaline dismu'tation phase of the dioxide into
chlorite and chlorate:
2 C1O2 + 2 NaOH = NaClO2 + NaClC>3 + H2
-------
- 467 -
water.- This method involves (6) decolouration of the reagent
in an ammoniacal medium buffered to pH 8.1 to 8.4. This is
measured at a wavelength of 550 nm (a compromise for the
hypsochromic shift of the absorption band).
With the familiar spectrophotometrie techniques the sensi-
tivity and the precision are 0.04 g ClO9.m~ . A value of
-1
0.02 g ClO-.m may be attained by microcell techniques and
by thermostatting the test solutions. The method'can also
be used for the determination of ozone (7). Reference should
be made to that publication for certain operating precautions
additional to those described in the original work. In
practice, interference from ozone need not be feared since,
like the chlorite in solution, the chlorine dioxide is
oxidized to chlorate under the action of ozone.
The use of the ACVK method has been proposed for the determin-
ation of chlorine dioxide in applications relating to the
treatment of waste waters (8). Since the method is colori-
metric, there is a possibility of interferences in this case,
but investigation has established a rigorous parallelism
between the results of the colorimetric method and the
results obtained by analysis of the nuclear paramagnetism.
This latter absolute technique is difficult to apply in
practice, but it has been used to confirm the specificity
and the reliability of the ACVK method.
3.5. 'The formation of residual chlorite by reduction of
chlorine dioxide is an awkward feature in the application of
chlorine dioxide to the treatment of drinking water. It
makes it necessary to control the residual chlorite in the
water, especially since this compound can be toxic when its
concentration is too high.
The only specific and sufficiently sensitive method that we
can suggest is based on a technique involving the differen-
-------
.- 468 -
tial plotting of polarograms obtained by means of pulsed
currents.
The electrode reaction is based on the following over-all
equation (9) :
HC1O2 + 4 e + 3 H
Cl + 2 H20
We recommend the following specimen pre-treatment (10) :
adjust the pH to 4.4-4.5 with a buffer, so as to produce the
final concentrations CH3COONa 0.04 M> CH_COOH 0.06 M, and
)^ SO^ 0.01 M-. If present, chlorine and chlorine
dioxide are removed by bubbling nitrogen through the
measuring cell. (If the presence of heavy metal ions in
concentrations that may cause disturbances is suspected, the
water should be passed through a strong cation-exchange
resin) .
The polarograms in the pulsed mode are recorded and, by
difference, a linear relationship is obtained between the
differential diffusion current and the concentration of
dissolved chlorite. The sensitivity is 0.05 g Ciol-m" (10).
/lA.
1,03
0.11
Pig. 3
Chlorite polarogram
(graphical construktion
by difference, in the
pulsed mode)
0.5
1.0
1.5
-------
- 469 -
3.6. The practical use of chlorine dioxide in the :treatment
of drinking water often entails continuous monitoring of the
residual oxidants. In this connection we have developed an
amperometric analyser for the iodine obtained when iodides
are brought into contact with oxidizing agents.
The analyser is based on the galvanic couple Ag/Au which
gives a zero residual diffusion current. A mixture containing
3 3
0.090 ,m of acetic acid and 800 g of KI per m is added to
the water being analysed, so ,as to obtain a final pH of 5.
Stabilization of the latter is based upon the buffering
effect due to the bicarbonate present in the water.
The direct output current obtained using curved electrodes
2
with an area of 3000 mm , separated by a distance of' about
50 mm, and with a 1000 fi resistance connected in series, is
—1 —3 - •
12.5 mA.Eq .m . The chlorine, the chloramines, and the
chlorine dioxide are shown up in accordance with their
capacity to oxidize iodides .at pH 5. Chlorites do not inter-
fere. The sensitivity its of the. order of 0.01 g ClO^.m"
(11)• . .
4. RISKS INHERENT IN THE USE OF C1C>2
4.1. In the gaseous states'chlorine dioxide is spontaneously
explosive at concentrations higher than 10 vol-% (12). The
reaction is induced by any source of ignition. At 20°C,
this concentration in the gaseous phase corresponds to a
concentration of about 8' g.a~ -of C1O~ in solution. The
acidification process thus generates a dangerous atmosphere.
The maximum allowable concentration in working premises is
0.1 ppm by volume for an 8-h work shift per day, and 0.3 ppm
for brief occupations. Chlorine dioxide is perceptible in
the air at concentrations of 1.4 to 1.7%. At 4.5% it
irritates the respiratory mucosae and .produces severe headaches.
-------
- 470 -
^ \ fr - . ' 1 - ™ ' t '' f. * - ~ | " •' " r ' . _ • 1 H ' . ' ' . . t . , , \L ^ ;
There is no cumulative effect in the event of repeated
exposure (13).
4.2. In drinking water the maximum residual consent that
does not affect the taste or produce a disagreeable odour is
of the order 0.4 to 0.5 g "C1O2 per m . This concentration
is also lower than the maximum allowable concentration from
the point of view of toxicity, since tests on rats have not
shown any serological toxicity at doses as high as 5 g.m
(14). In Belgium the maximum concentration of chlorine
dioxide tolerated in water distributed through the public
mains is 0.25 g.m
4.3. At least some of the dioxide used is likely to be
transformed into chlorite by reaction with organic compounds,
The chlorite is toxic and can cause methaemoglobinaemia.
_ -I
However, the LD^ of sodium chlorite for rats is 140 mg.kg
On the extreme assumption that all of the dioxide could be
transformed into chlorite, this value of LD50 corresponds to
105 mg C102.kg . Consequently, there seems to be no
objection to the use of chlorine dioxide, in the usual
amounts, on these grounds.
5. ADVANTAGES OF CHLORINE DIOXIDE TREATMENT
5.1. It has been found that, the cost of post-disinfection
with chlorine dioxide, compared to that of chlorine, is
1.2 to 1.7 times as high (15,16). However, the relative
cost of treating raw water can change appreciably as a
function of local parameters. Thus, ratios of 1/1 to 1/4
have been found (17). The raw water is treated at the
Tailfer station of the Compagnie Intercommunale Bruxelloise
des Eaux both with chlorine and the dioxide. The average
-3 -3
amounts have been equal to 1.3 g ClO2.m and 6 g Cl2.m ,
corresponding to a respective cost of 1 to 2.4 for the
chlorine compared to the dioxide used.
-------
- 471 -
5.2. The use of chlorine dioxide .constitutes an alternative
to prechlorination. One of its advantages is the smaller
tendency to form organochlorine derivatives, bearing in
mind the reactivity of organic derivatives with this
oxidant (2). Moreover, chlorine dioxide forms fewf if any,
simple chlorinated hydrocarbons of the chloroform type (18).
Another, often decisive, advantage consists in the fact that
chlorine dioxide does not react with dissolved ammonia.
Thus, it is not essential to use amounts greater than the
critical point to obtain satisfactory disinfection. This
point is also important for the treatment of swimming pools.
Like ozone, chlorine dioxide leads to micelle formation in
the coagulation-flocculation pre-treatment. Because of this,
the turbidity of swimming pool waters treated with chlorine
dioxide is lower than that obtained by a comparable treatment
with chlorine (19). Moreover, iron and manganese, which may
be present in the reduced or complexed states, particularly
in association with fulvic or humic acids, are oxidized and
eliminated by subsequent precipitation.
One of the most important aspects of using chlorine dioxide
for the pretreatment of river water is the competition between
the oxidation reactions that produce chlorite and the dis-
mutation into chlorate and chlorites. The respective
reactions are:
C1O2 + 1 e = C1O2
6 C102 + 3 H20 = 5 HC103 + HCl
According to this scheme the C1O~/C1O_ mass concentration
ratio of 1 corresponds to a molar concentration ratio of 1.24,
Thus, if the C1O2/C1O~ mass ratio is observed to be 1, about
7.4 moles of initial dioxide will have reacted for each mole
lost by dismutation.
-------
- 472 -
This dismutation of chlorite is' slow"in the absence of light.'"
The following reaption is of minor importance except in the
event of intense insolation:
3 C1Q~ = 2 C103 + cl~
Experience has given the following"recommendations in relation
to the practical use of chlorine dioxide for the treatment of
raw water:
- A reaction time of 30 to 40 min is sufficient.
- The residual C109 concentration after this time
— 3
should not normally sec eed 0.2 g.m
- The chlorite/chlorate mass ratio after this action
should be between 1 and 2, corresponding to an
effective degree of reaction of 83 to 91%. For a
decantation tank open to the air, if the residual
—1
concentration of C1O~ exceeds 0.2 mg.£ , the loss
of CIO,, by dismutation into chlorides and chlorates
increases.
Moreover, the decomposition of chlorine dioxide by active
carbon, as used in coagulation-flocculation, is appreciably
less rapid than that of chlorine (20). Thus, it becomes
possible to maintain a bactericidal activity in sludges con-
taining flocculated carbon.
5.3. During post-chlorination, chlorine dioxide can main-
tain itself in clean water for up to 48 h. Thus, its
bactericidal efficacy is guaranteed for a longer time than
that of chlorine. The same conclusion is relevant to the
persistence of a disinfecting action in swimming pools (21).
The disinfecting action of chlorine dioxide is recognized to
be at. least equal to that of chlorine, and superior in the
case of waters with a pH higher than neutral. In this
-------
- 473 -
connection we refer to our previous analysis to support this
* .v '.•,.' - . L < '• -, • •'" 'r. ' ' * *f, ' ' ' . -~ k / r , *
conclusion (2). It seems, _ in fact, that the diffusion of
chlorine dioxide depends very little on pH compared to that
of chlorine in its various forms in equilibrium. I shall
not present here, an exhaustive review of the bactericidal
and disinfectant properties of chlorine dioxide, since this
lies outside the scope of .the present paper, and merely refer
to previous reviews of the subject.
5.4. After the reaction a significant proportion of the
chlorine dioxide is reduced-"to chlorite. It is thus
interesting to examine the bactericidal or bacteriostatic
properties of the chlorite ion, in relation to post-
,disinfection.
The disinfectant power of the chlorite ion is minor. Thus,
the philosophy of post-disinfection with chlorine dioxide
rests on the action of chlorine dioxide itself, which is a
strong bactericide, followed by that of the chlorite, which
is weakly bacteriostatic"and weakly bactericidal. The
chlorite, in its own right, is not used directly in dis-
infection. However, we have tried to determine its bacteri-
cidal and bacteriostatic effects.
5.5. The bactericidal effect upon enterobacteria is shown
in Figs. 4 and 5.
In physiological water, type C coli were removed, during the
experiments, in accordance with Chick's kinetic law:
, N , '•••••
log N^ = -k!0 t . ,.
In this equation . . . , •
k!0 = 2^3" = -°-075 day"1.
-------
- 474 -
In other words, under "the conditions of the experiment , the
time required to produce 50% mortality is of the order of
4 days in the absence of any bacteriocide.
In the presence of sodium chlorite a clear increase in the
mortality appears at concentrations above 0,2 g ClO-.m
This degree of mortality no longer corresponds to Chick's
law. To interpret it, we subtracted the mortality in the
absence of chlorite from that observed in its presence.
Moreover, we based ourselves on the theory of multiple site
inactivation, in which it is considered that a certain number
of sites must be reached at least once if the organism is
to be completely deactivated. Since the kinetics are first-
order for each site considered, the probability of deactiv-
ation of m sites is as follows:
N - N
_ ,. -kt,rn »> o .
Pm=il - e J — j-j (fraction of organisms killed)
o
ThUS f- - 1 - Pm - 1 - (1 - a"Kt)m
o
Binomial expansion gives:
-Kt m Cm - 1)
Pm = f1 - 8j ., 1 _ m a
- 1 - m e-kt
Thus, the fraction of surviving organisms is:
o
and loS
-------
- 475 -
-1
.2
Coli C
= ,= -0.35 PAR JOUR
ZJ PEP DAY
PRO TAG
0.2 OPK CL02
PAR RAPPORT A
COMPARED TO
BEZOGEN AUF
O PPM CLOj
T, JOURS/DAYS/TASE
-TO
Fig. 4 Action of chlorite on coli C
,0.25.
-0.25.
-O.b .
-0.75-
l03= 0.25; = 1 6
ENTEROCOOUBS
ENTEROCOCC!
ENTEROKOKKEU
JOURS/DAYS/TACrE
Fig. 5 Action of chlorite on enterococci
-------
- 476
As far as enterococci are concerned (a wild ,s.train Isolated
from the water) , a similar argument gives in ~ 1.8, or about
2. These may have been diplococci.
5.6. Among the ubiquitous microorganisms, Pseudomonas
strains deserve special attention since they show a capacity
for rapid regeneration in treated waters.
Of the pure strains, Pseudomonas cruciviae has no capacity
for surviving longer than a few hours in water. The wild
strains that we were able to obtain in cultures seemed
generally sensitive to chlorite. In contrast, Pseudomonas
putida is more resistant. According to an analysis similar
to the preceding one, the mean number of "sites" was about 6,
while the concentration factor was of the order of 0.3.
PUTIDA - CMerltt
I OB
-------
477 -
5.7, Certain cultures were 'found to constitute exceptions
to the rule. Among these, we isolated, on a King F medium,
a Rodotorula, which is a ubiquitous non-pathogenic species
that can be found among the organisms colonizing the human
body. Whereas in water untreated with chlorite the bacterial
count decreases progressively as a function of time, it seems
that in the presence of chlorite, after a first phase of
decline, the microorganism can adapt and maintain an increased
level of vitality (Fig. 7}. This observation is reported at
face value,, but it still requires detailed checking before
any conclusions can be drawn.
leg N
s -
tEVURE TYPE RODOTQRULE (CHLORITE)
ROBDTORUIA YEAST (CHLORITE)
HEFEBAKTERIUM VOH TYC RODOTORULA
r
0.2
\.
X /
X
^
0.0
HEURES
HOURS
STUKDEN
Fig. 7 Action of chlorite on Rodotorula
-------
- 478 -
">- ',
6. CONCLUSIONS
Though chlorine dioxide has already been in use for several
decades for the .treatment of drinking water, it often
remains an ill-understood reagent. The problems posed by
the formation of organochlorine derivatives by chlorination
have resulted in a renewed interest in the use of C1O2•
This reagent has great potential: chlorination reduced or
eliminated, ability to oxidize organic compounds, higher
bactericidal activity in alkaline media compared with
chlorine (this activity is maintained in the presence of
dissolved ammonia without the need to overchlorinate beyond
the critical point), etc.
In the present paper I have reported specific data concerning
the use of chlorine dioxide, especially in relation to areas
of difficulty:
- Dangers of preparation and handling.
- Toxicity and effects of residual concentrations.
- Analytical methods for the specific determination of
chlorine dioxide and residual chlorite.
- Bactericidal and bacteriostatic effects of the
dioxide used for post-treatment.
It seems that the objections often formulated against chlorine
dioxide do not constitute valid sanitary'grounds for opposing
its use.
-------
- 479 -
(1) MASSCHELEIN, W.
Les oxydes de chlore et le chlorite de sodium
Ed. Dunod (1969)
(2) MASSCHELEIN, W.J.
Chlorine Dioxide (Chemistry and' Environmental Impact
•of Oxychlorine Compounds)
Ed. Ann Arbor Science (in press)
(3) MASSCHELEIN, W.
Ind. Eng. Chem. Prod. Res. Develop. £ (1967), 137
(4) APHA-AWWA
Standard Methods (1975) '
(5) HODGEN, H.W., INGOLS, R.S.
Analytical Chemistry 26 (1954), 1224
(6) MASSCHELEIN, W. ,
Analytical Chemistry 38 (1966), 1839
(7) MASSCHELEIN, W.J., FRANSOLET, G.
J. AWWA 69 (1977), 461
(8) KNECHTEL, J.R., JANZEN, E.G., DAVIS, E.R.
Analytical Chemistry 5_0 (1978), 2O2
(9) HARTLEY, A.M., ADAMS, A.C.
J. Electroanal. Chem., J5 (1963), 46O
(1O) MASSCHELEIN, W.J., DENIS, M. , LEDENT, R.
(in preparation)
(11) MASSCHELEIN, W.J, FRANSOLET, G.
(in preparation)
(12) U.S. National Board of Fire Underwriters,
Report No. 7 (1949)
(13) HALLER, J.F., NORTHGRAVES, W.W.
Tappi 38 (1955), 38
(14) FRIDLYAND, S.A., KAGAN, G.Z.
Gig. Sanit. 36 (1971), 18
(15) BOGNOLESI, U.
Igiene Mod. £6 (1953), 197
-------
- 480 -,
(16) GOMELLA, C.
Techn. Sci. Mimic. 56 (1961), 1.71
(17) DOROLING, L.T.
Water Treatm. Exam. 23- (1974), 19O
(18) VILAGINES,. R., MONTIEL, A., DERREUMAUX A., LAMBERT, M,
96th Annual AWWA Conference Anaheim (1977)
(19) BANDI, E.E.
Mitt. Geb. Lebensmittelunters. Hyg. 58 (1967), 176
(2O) MASSCHELEIN, W., GOOSSENS, R. , MINON, M.
Tijdschrlft BECEWA (in press)
(21) ASTON, R.N., VINCENT, G.P.
Proc. 9th Ann, Course Water Plant Operators
10 (1947) , 54
-------
_ 481.'.V- -
PRACTICAL EXPERIENCE WITH CHLORINE DIOXIDE AND
FORMATION OF BY-PRODUCTS , .
T. Hedberg, B. Josefssort, C. Roos, B. Lindgren and
T, Nemeth
1 . Introduction
The use of chlorine dioxide in water treatment plants
has increased during the past years. In an EPA-report
(1 )• its extensive use in such countries as West Germany,
Switzerland and France are mentioned. Its use in Sweden
is also becoming increasingly common. Chlorine dioxide
was introduced in 1968'at the water treatment plants
in Gothenburg (70 million m per year). In Sweden today
about 15% of the supplied water is treated with chlorine
dioxide. The reason for its use is, as in most countries,
to improve taste and odour, and at the same time in-
crease the disinfection capacity in the distribution -
system.
Although chlorine dioxide has been used during the last
ten years in Sweden, research within this field has only
recently started. The general aim is to investigate the
oxidation efficiency, by-products and usefulness of
different oxidants for the typical surface waters in
Sweden. The Swedish waters contain in contrast to most
continental waters high concentrations of humic sub-
stances and thus they are interesting from a treatment
point of view. In addition the waters are - especially
in the western and northern parts of Sweden - very soft
witn a low content of salts and are also rather acidic
as a result of long distance transport of air pollution.
Since soft waters with"a low buffer capacity are very
aggressive to pipes.it is common to adjust the pH value
to 8-8.5 before distribution. ,
-------
In addition to the use of chlorine dioxide in water treat-
ment it is also employed in the industry for pulp bleach-
ing. This is in some aspects a related process since in
both cases lignin-type compounds are affected. Thus an
account of this process is also presented.
2. Formation and behaviour of volatile halogenated orga-
nic compounds by chlorine dioxide - chlorine treat-
ment of water in Gothenburg water treatment plant
There has been much concern over the formation of halo-
forms during.the chlorination of water.
In search for safer oxidation techniques attention has
been drawn to chlorine dioxide. However, the use of
chlorine dioxide may also create undesirable by-products.
Chlorine dioxide is a strong oxidant and exhibits a
wide spectrum of reactivities towards organic compounds
(2). However, to date no chlorinated by-products from
the treatment of drinking water with chlorine dioxide
have been reported.
Studies of chlorine dioxide bleaching of pulp (section 3)
and laboratory experiments on organic model substances
provide evidence for production of considerable amounts
of organic by-products. On the other hand the concen-
tration of reaction products arising from the use of
chlorine dioxide in drinking water seems to be compara-
tively low. Since no volatile organic compounds have
been identified as by-products, one must assume that
these by-products are dominantly non-volatile. This de-
mands new techniques for analyzing non-volatile com-
pounds in low concentrations.
Presented below are results from analyses of volatile
halogenated compounds at different stages in a water
treatment plant in Gothenburg, Sweden, that utilizes
-------
- 483;-,-
combined chlorine-chlorine dioxide treatment. The
occurrence of these by-products in the distribution
net has also been investigated.
The analyses of volatile organohalogens in the drink-
ing water were carried out with two different methods
during a two-week period a: closed-loop stripping
2
after Grob (3) followed by (GC) -MS and b: pentane
extraction followed by glass capillary column separation
and electron capture detection, (GC)^-ECD (4). The re-
sults are presented in Table 2-1. This work has been
reported elsewhere (5) in detail.
TABLE 2-1 Concentration of organic compounds in drink-
ing water. (The compounds marked with an aste-
risk were enriched by pentane extraction.)
Halogenated hydrocarbons
Dichloromethanex
Trichloromethanex
1,1, 1-trichloroethanex
Tetrachloromethanex
V
Trichloroethene . ' •
Bromodichloromethanex
X.
Dibromochloromethane ,
Tetrachloroethenex
Dichloroiodomethanex
Tribromomethane
Trichloroaniline
Bromonaphthalene
Drinking water, ng/1
< 6O
95OO
60
17
15
22OO
6OO
8
7
16
1
26
Twelve halogenated organic compounds were found of
which some were originally -present in the raw water,
e. g. trichloroaniline..
-------
— 484^--
Fig. 2-1 shows the flow scheme of the plant and the
concentrations of halogenated compounds. -Notable are
the markedly increased'concentrations of carbon tetra-"
chloride after.the different chlorination points. How
ever, this increase' is not observed in the prechlorina-
tion step of raw water with a high concentration of ,
suspended solids. There seems to be an influence from
sedimentation of particulate matter. The increased
concentration of carbon tetra^chloride probably origi-
nates from the, added chiprine gas.which normally con-
tains 25-100 mg carbon tetrachloride/kg. The reaction
of hypochlorous acid with.chloroform by an electrophilic
attack to yield carbon tetrachloride seems to be insigni-
ficant because'of the -slow reaction rate (6).
Seasonal variations in water quality, e.g. fulvic acid
concentration will affect halofprm formation. Samples
taken during the winter (Fig. 2:-1) show lower contents
of haloforms compared with,samples taken during the
spring (Fig. 2-2). ' ; ' ', ••
Concentrations of haloforms in the distribution net
over a distance of 30 Km" in the" Gothenburg region
(population 600 000) are illustrated in Fig. 2-2.
Preliminary evaluation of the. results.shows an increase
of the haloforms with distance .from the plants. This
might be explained by the generally•increasing pH
value -with distance. The 'change^ in pH .value is of course
dependent on the pipe ma-tier ials: used (concrete-and
cement-lined pipes) . - - -, - - - .
Formation of haloforms- from chlorine, chlorine dioxide
and different combinations of these "•'•
Haloforms are produced; during chlorination of drinking
water. It has been demonstrated that they are formed
from degradation product's of humic substances (7) which
are present in all natural waters.
-------
WATER WORKS
1
Raw
CH CI3
CHBrCI2
CHBr2Cl
CHBr3
CCI4
CHCICCI2
C2CI4
I
i Pre chlorination i Pre chlorination
i 0.7 g/m3 0.5 g/m3
0.14
n.d.
h.d.
n:d.
0.006
0.036
0.012
1
r
1
Lake
reservoir
i
3.5
0.5
0.013
n
.d
0.004
0.027
0.090
1.1
0.1
0.003
n.d.
0.004
0.027
0.090
1
8
2
0
n
r
Alkali-
nisation I
1
.9
.1
23
.d
0.116 .
0.018
0.060
8.3
FJoccul,
sediment.
0.6
0.04
n.d.
0.090
0.018
0.090
1
i
Activated
carbon
filter
4.5
0.6
0.06
n;d.
0.050
Q018
0.012
Post chlorination I
chlorine + chlorine dioxide I
0.3 g/m3 + 0.3 g/m3 I
r ,
Alkalin. Or
h * ^
\
silica
i
4.5
0.7
0.06
n
.d.
aoa4
Q027
0,006
7.7
2.0
QSO
0.009
0.108
0.108
0.003
nking I
water
I
J
I
03
Ul
I
Fig.2-1 Halogenated compounds in the LackarebSck Treatment Plant
-------
486' -
25
20
g 15
10
i
_L
10 15 20
DISTANCE , KM
25 30
GOTHENBURG
ALILYCKAN WATER TREATMENT PLANT
LACKARBACK WATER
TREATMENT PLANT
O 2 4 6810km
Fig. 2-2 Concentration of trlchloromethane in the
distribution net :
-------
- 487 -
In a series of experiments on a.; laboratory scale,, the
formation of haloforms has been studied. River water
(pH 6.88), which is the raw water for the water-works
in Gothenburg, was treated with chlorine, chlorine dioxide
and different combinations of these. Analyses of halo-
forms were performed -by pentane extraction followed by
gas chromatography - electron capture detection (4). All
experiments were carried out at room temperature. The
time-dependence of the haloform formation is shown in
Fig. 2-3. Chlorine was added to river water until a con-
centration of 1 mg/1 was reached. The haloforms were
analyzed after different periods. The reaction is completed
after about 1 hour.
Fig. 2-4 shows the results obtained when river water
is treated with different amounts of chlorine. Reaction
time 1 hour. At approximately 2 mg/1 a plateau level
is reached and no additional haloforms are produced
when higher amounts of chlorine are added.
Treating river water with chlorine dioxide results in
only minor formation of haloforms. This is, however, a
result of the small amount of chlorine present in the
chlorine dioxide solution used.
Thus chlorine does and chlorine dioxide does not produce
haloforms. To investigate the effect of a combination
of these, two experiments were carried out, Firstly
1 • ?
chlorine was added to river water at a concentration
of 1 mg/1. After 1 hour different amounts of chlorine
dioxide were added to the samples. After another hour
the samples were analyzed for their contents of. halo-
forms. Secondly different amounts of chlorine dioxide
were added to river water. After 1 hour chlorine was
added to a concentration of 1 mg/1. After an additional
hour the samples were analyzed for their contents of
haloforms. In these experiments pH was kept constant
at 6.97 ± O.O7 by a phosphate buffer. The results are
-------
H9/I
101
- 488 -'
*-2
Fig. 2-3
The time dependence of the
haloform formation
1 . CHC13
2 • CHBrei-
20 40 60 80
Fig. 2-4
Formation of haloforms from
different amounts of chlorine
2 m CHBrCl
2
1 2
1 4 5 mi «2/|
-------
- 489 -. .:-x,v .
shown in Pig. 2-5. Smaller.amounts,of. halof OCTIS, *are .formed
when chlorine:dioxide is added before chlorine. ,It is
clear that the adding sequence,of chlorine and chlorine
dioxide is important with respect to haloform, formation.
One might assume that chlorine dioxide degrades those
compounds which yield haloforms upon chlorination .
1O mg CI
Fig. 2-5
=! /
Formation of haloforms from
combinations of chlorine and
chlorine dioxide
A « CHC1- ? B « CHBrCl2
1 « Chlorine added before
chlorine dioxide
2 = Chlorine dioxide added
before chlorine
3. Chlorine dioxide in wood pulping industry
Chlorine dioxide is used in w.ood pulping industry to .
degrade and dissolve lignin residues in the sulfate.and ,
.sulfite pulps ("bleaching"). Until, now its use has been
due mainly to pulp quality demand. In the future, however,
of minimizing" the discharge of chlorinated
-------
- 490 -
products into lakes; .and 'streams will pjsobably result in
the increased use of chlorine dioxide as well as oxygen
in bleaching in place of chlorine and hypochlorite.
The use of chlorine dioxide in pulp bleaching is very
different from its use in water purification. The
conditions are for example quite different as shown in
Table 3-1. It has, however, one important similarity.
The lignin in pulp is a similar material to the humics
present in many waters. Both are polymeric aromatic
materials.
TABLE 3-1
Comparison between conditions during chlorine
dioxide bleaching and during water purifica-
tion with chlorine dioxide.
Pulp bleaching
Water purification
pH
Temp.
Time
Cone, of added C1O,
Start 6, finish 3
60-90°C
30-180 min
5 m mole/1; or., higher
6-8.5
0-20°C
extensive
<0.03 m mole/1
The most reactive positions in lignin and maybe also
in humics are the phenolic groups, 'The lignin phenolic
groups are oxidized to quinonic groups and muconic acid
groups (Fig. 3-1). Chlorine dioxide is reduced part-
ly to chlorite (1-electron-reduction) and partly to
hypochlorite (3-electron-reduction), •. The hypochlorite
formed then reacts either with chlorite (see below) or
with lignin.
-------
- 491 -
COOCH-
DOOM
Pig. 3-1
The reaction of phenolic
groups of lignin with
chlorine dioxide
The concentrations of inorganic chlorine compounds change
during bleaching as shown in Fig. 3-2. At first, chlorine
dioxide as well as chlorine, which is a companion of
chlorine dioxide in all technical chlorine dioxide pre-
parations, are rapidly consumed. The concentration of
chlorite increases as long as chlorine dioxide is
present but then decreases slowly. Thus chlorite is
actually the dominant bleaching agent. As shown in the
figure, chlorate is also formed. Its formation is a
useless by-product of the bleaching process since it
cannot react with lignin. It is also possible that in
water treatment the reaction of humics with chlorine
dioxide will follow the same pattern described by Fig 3-2,
The reaction between lignin and chlorite involves the
transformation of chlorite into chlorine dioxide which
then reacts with lignin. The hypochlorite formed by
this reaction (see above) oxidizes chlorite into further
amounts of chlorine dioxide, which in their turn react
with lignin. The process is then a chain reaction.
Whether the reaction of chlorite in water with, for
example^ humics is a similar chain reaction is difficult
to say. The low reactant -concentrations are not benefi-
-------
- 492 -
cial for a chain reaction. Chlorite may in water treat-
ment be slowly transformed into chlorine dioxide by an
oxide-reduction. If the reaction in the water is a chain
process/ however, its rate could vary widely depending
on the nature of the organic material, present.
o o
20 40 60
TIME, min
Fig. 3-2 The concentrations of inorganic chlorine compounds
during chlorine dioxide bleaching of a. sulphate pulp
Oxidation of phenols of different types to those pre-
sent in lignin yields guinones, quinols and epoxides
(Fig. 3-3). The epoxides observed are rather unstable
and are not detectable for some hours in water.
-------
-493 -
HO CH3
Fig. 3—3 The reaction of mesitol with chlorine dioxide
Chlorine dioxide reacts rather rapidly with double bonds
activated by conjugation with benzene nucleus. For ex-
ample stilbene is oxidized to stilbene epoxide (Fig. 3-4)
The double bond of other stilbenes examined are split
by chlorine dioxide,
a02+C6%-CH=CH-C6H5 —» CIO(?)+C6H5-CH-CH-C6H5
Y
Fig. 3-4 One of the reactions of stilbene with
chlorine dioxide
An inactivated double bond such as that present in methyl
oleate and cyclohexene reacts slowly with chlorine
dioxide. Oxidation involves formation of a-chloro-ke-
tones, carboxylic acids and traces of epoxides. The
chlorination products of the olefine (di-chloro com-
pounds and chlorohydrins) are formed also by reaction
with chlorine dioxide.
-------
- 494,-
Mixtures of chlorine and chlorine dioxide are often
used in pulp bleaching. The ratio between the two
bleaching agents is then kept at a level which gives
a desired effect. For example, addition of a small
amount of chlorine dioxide to chlorine gives a more
selective attack on the lignin, which spares the
cellulose. Chlorine dioxide works as a radical sca-
venger hindering the radical attack of chlorine on
cellulose but not the non-radical attack on lignin.
Thus, it is also possible that in water purification
good results can be obtained by using mixtures of
chlorine and chlorine dioxide.
4. Chlorine and chlorine dioxide for water supply
disinfection and taste and odour control; experiences
Disinfection
Chlorine is by far the most commonly used chemical for
water disinfection. Chlorine dioxide is used very often
for taste and odour control, but it is seldom used*
for disinfection, even though it is a powerful disinfec-
tant. In fact in certain situations it may be superior
to chlorine. However, more information is still needed
concerning its potential use as a substitute to chlorine.
Problems with the analyses for chlorine arid chlorine
dioxide
There are several methods of analyses for chlorine and
chlorine dioxide which are well suited for water«-works.
In measuring chlorine concentrations one usually deter-
mines the total halogen concentration of the water.
This so-called total chlorine concentration is regarded
-------
495 -
to be the sum of the free and combined chlorine consisting
of chloramines (mono-, di-, and trichloramine). During
analyses interactions with other chlorine compounds as
well as with bromine and iodine compounds can occur.
The total chlorine concentration can thus be written
C1
2 ' total
C10| + jci NH~ + ICl-compounds I +
+ Br,
+ |C13 N
+ BrNH-
+ IBr-compounds| +
+ ]Br3N | +
+ |I-compounds
Usually the bromine and iodine compounds can be neglected.
In reality chlorine in water is in equilibrium with
hypochlorous acid. At pH values exceeding 5, which is
the case in all water treatment processes, the real
chlorine concentration is zero. At a pH of 7.4 about
80% of the total chlorine is in the form of hypochlorous
acid (HC10)i,e. hypochlorite ion (CIO ).
This can be seen from the equilibrium curve. Fig. 4-1 .
It is well known that the hypochlorite ion has a very
low disinfection effect.
Because of the differences in disinfecting power of
each species it is desirable to distinguish them. In
practice, however, the "chlorine residual" is often
measured by the o-toludin method or similar methods
which do not distinguish between the different species.
Recently a method which makes it possible to distinguish
between chlorine and chlorine dioxide has been developed
in Sweden (8).
-------
- 496 *•
2 3 A 5 6 7
pH VALUE
8 9 10
Fig. 4-1
Distribution of chlorine
hypochlorous acid and hypo-
chlorite ion in water at
different pH values
Behaviour of chlorine and chlorine dioxide as disinfec-
disinfectants
In order to prevent precipitation of calcium carbonate
in hard waters or to prevent corrosion in soft waters,
it is necessary to bring about equilibrium in the cal-
cium carbonate - carbonic acid system prior to the
distribution. In a very soft water, which is common in
Sweden this equilibrium cannot be obtained with the
usual techniques. It is actually necessary to increase
the water hardness which is an expensive process. In
order to protect the pipes, and especially the copper
pipes, the pH value is normally kept as high as 8.5.
At this pH value the disinfection capacity of the water
is decreased, which may lead to unsatisfactory quality
in case of a secondary contamination.
The measurements of combined chlorine may not be satis-
factory for control of disinfection. Chloramines are
frequently used for disinfection in the distribution
system. It has been shown (9) that the various combined
chlorine compounds may be different from one another
in capacity to inactivate bacteria or virus.
-------
- 497 -
•a
c
o
u
200
100
60
40
20
10
6
4
2
1
N
^
I
v
1 2 4 6 10 20 6U100 400 1,000 ,
Tirne-spc I
!
Fig. 4-2 Chlorination of Esch coli (37°C, pH 7), using
different forms of combined chlorine (9)
The equivalents of combined chlorine are plotted against
the time for a one log reduction of the concentration
of live bacteria. The solid curve indicates results ob-
tained by the use of chloramine T; the dashed curve shows
results obtained by the use of NH-C1.
5 10 15 20 25
Applied Chlorine —ppm
30
Fig. 4-3
Free residual'chlorination
-------
--498.'—'
The solid curves ' indicate '"the "relatibrfshlp, between-'the1 •''
applied concentration of free chlorine added to a con-
stant concentration of ammonia and the residuals found
of combined chlorine. The squares indicate the first
values at which residuals of free chlorine were also
demonstrable. The dotted curves represent the rate of
reduction of live Esch coli at 37°C and pH,7. The time
in seconds for a one log unit reduction is plotted
against the applied concentration of chlorine. Identical
symbols are used for corresponding curves for each of
the three experiments.
Most experiments with bacteria have been made either
with Escherichia coli or with bacteria described as
belonging to the coliform group. However, Ridenour and
Armbruster (10) carried out extensive experiments with
different bacteria. For all the organisms they observed
similar results for chlorine or chlorine dioxide^ They
concluded that E. coli was more resistant to chlorine
or chlorine dioxide than any of the bacteria on their
list which could be considered pathogenic.
At neutral pH values, parallel experiments" by a number of
researchers have shown chlorine and chlorine dioxide,
at equal concentrations, to be about equally effective
against E. coli (Bernarde, et al, (11), Berndt (12),
and Ridenour and Armbruster,(1O) ). We have obtained
similar results in laboratory experiments on coliform
group bacteria in Gota Ji.lv raw water (Gothenburg). The
effects of variations in pH value were also examined.
The effectiveness of chlorine was observed by all
investigators to diminish with increasing pH value of
the water (eg. Pair and Geyer, (13) ).
With chlorine dioxide, however, there has been no such
agreement among different investigators. Ridenour and
-------
--49S- '•--.-
Armbruster (1O) observed that 'chlorine dioxide was at
least as effective at pH 9.5 as at pH 7. The experiments
were performed with E. coli added to chlorine demand
free water. .
Bernarde et al.(11) found that chlorine dioxide was
more effective at pH 8.5 than at pH 6.5. A different rela-
tionship was observed by Berndt (1965). These experiments
were made using raw water from the water-works at the
city of Lubeck. The result showed that chlorine as well
as chlorine dioxide were more effective at lower pH
values (pH 5) than at higher pH values (pH 9.3).
Experiments with bacterial spores have shown that chlo-
f.
rine dioxide is more effective than chlorine when these
chemicals were compared at equal concentrations (Ridenour
and Armbruster (1O), Berndt (12) ).
Some comparative studies (Gothenburg 1978) on total
counts of bacteria at 22 in drinking waters disinfec-
ted by different techniques were performed. The streak
plate method was used for growth tests on Nutrient Agar
(Oxoid Type CM3), Tryptone Soya Agar (Oxoid Type CM 131)
and Fe Pa medium according to Ferrer, et al. (.14) . The
first results show that on all the media used growth
is more often obtained from samples taken ,from the dis-
tribution system of Uppsala (chlorine treatment) than
from the chlorine-dioxide treated water of Gothenburg.
Considerable data has been publishe'd during the past
twenty years showing the effectiveness of chlorination
against different viruses under different conditions.
In general^ free chlorine in water effectively inacti-
vated the different viruses which were tested. Further—
more,chloramines or other combined chlorine forms were
found to be much less effective. ,
-------
-*. soft --
Chlorine and chlorine dioxide have been compared for
effectiveness against poliovirus (Ridenour and Ingols,
1946). At that time it was not possible to titrate a
virus concentration so 'Conveniently and accurately as
it is today. It appeared that chlorine dioxide added
to a virus suspension at pH 7 was more effective than
an equivalent dosage of hypochlorite.
The effects of changes 'in the pH value have also been
studied. In the range pH 5 to pH 9 chlorination has
been consistently more effective in laboratory testing
at the lower pH values.
On the other hand, several investigators have shown
that chlorine dioxide inactivates viruses faster at
higher pH values than at lower -pH values. Cronier et
al. (15) showed a three-times faster 'inactivation of
poliovirus 1 at pH 9.O than at pH 7.O. Warriner (16)
showed the same effect-on poliovirus 3. This effect was
more marked at higher dosages. Fig.; 4-4, Table 4-1.
When 0.2 mg C12/1 was added together with O.5 mg C1O2/1
no improvement relative to O.5 mg ClO^/l alone was
noticed.
Thus chlorine dioxide appears to be a better alterna-
tive from a microbiological point of view than chlorine
since it is most effective in the pH range in which
water is distributed. - • - -
Except for the generalizations about the effect of pH
noted above, the results of many virus inactivation in-
vestigations have been .contradictory. The practical
application is limited because chlorination according
to the same procedure may give* different results in
different waters (Engelbrecht, et al, (17) ). The diffi-
culty is enhanced because it is not practical to make
routine virological isolation tests in the manner that
routine bacteriological testing can be carried out on
-------
100-
o
z
Ul
ce.
10-
.£*.§£_ . r
0,2 mg/l ClOj
10
30
TIME.min
60
100
0,5 mg/l ClOj
10
15
TIME.min
1,6mg/l C102
I
o
1
Fig.4-4 Remaining virus as a function of contact time at different dosages
of chlorine dioxide at different pH values (16)
-------
T&BLE 4-1 Inactivation of poliovirus with chlorine dioxide at 20 C.
The redox potential is measured against a caldmel electrode.
Parameter
Redox potential, mV
without virus
with virus
Amperometric titration of
chlorine-chlorine dioxide
residual, mg/1
chlorine
chlorine dioxide
chlorite
Cone, of surviving virus
(log ID5Q/0.1 ml)
without chlorine dioxide
with chlorine dioxide,
after 10-20 sec
after 1 min
after 2 min
after 5 min
after 1 5 min
after 60 min
Test No. 1
0.2 mg C102/l
pH 5.7 pH 7.0 pH 8.5
610 440 360
360 330 260
000
0.07 0.03 0.03
0.11 0.11 0.12
3.2 3.8 3.8
2.4 3.4 3.6
2.6 3.5 3.0
_ _- _
_
_ _
2.5 3.2 2.8
Test No. 2
0.5 mg C102/l
pH 5.6 pH 7.2 pH 8.5
670 660 580
630 580 270
0 - 0
0.27 - 0
0.12 - 0.41
2.8 2.5 3.5
1.8 2.2 1.6
1.4 1.6 0.6
1.2 0.8 O.8
0.8 0.5
0.4 0.5 0.2
0.4 0.5 0.4
Test No. 3
1 .6 mg C102/1
pH 5.6 pH 7.1;
•
720 705
700 690 '
•'•
' ~:
0.02 0.03-
1 .40 1 .41;
0.07 0.08J
2.0 2.5:
"
2.0 0.5-
1 .2 1 .3-
0 . 6 -0.4;
<-0.5 - „
— — ":•
4-
O
tsj
-------
- 503 -
!
both raw and treated water., .However, investigations by
Lund (18, 19) have revealed a relationship between oxi-
dation potential and 'the rate of virus inactivation.
This work has-' opened the possibility of regulating ,
chlorination, including chlorine dioxide treatment, for
the purpose of inactivating virus.
Figure 4-5 shows results obtained for the inactivation
of poliovirus using chloramines (9). It shows the range
of potential measurements (with a calomel reference elec-
trode) which are likely to be of interest in water ;
treatment. Among 'the practical questions raised concerning
this procedure are the difficulties encountered in
measuring potentials and the effectiveness in con-
trolling the bacteria.
The oxidation potentials are plotted against the time
in minutes i "»ded for a reduction of one log unit of
active virus.
700
600
500
100
:Fig.' 4-5
Oxidation potentials obtained
by chlorination of poliovirus
(37°C, pH 7) (9)
30
60
Time—nun
120
-------
- 504 -
The common denominator 'for1 disinfection chemicals' ds • • >v
thus their ability to break through the redox buffer
capacity and thereafter change -the redox potential.
The redox potential of a chlorine -solution is strongly
dependent on the pH value/as can be seen from Fig. 4-6.
Th"e redox potential decreases with increasing pH and
the possibility to break the redox buffer capacity is
thus reduced. In soft waters, in which it is necessary
from a corrosion point of view to have a high pH value,
a. five-times higher chlorine concentration as in hard
waters is required and still the disinfection effeciency
cannot be guaranteed.
goo.
7,0
,CI2
:edox=((pH)
0,5 1,0 £
CHLORINE RESIDUAL nig/l
8,0 9,0
pH VALUE
10,0
Fig. 4-6 The redox potential as a function of the chlorine
residual and as a function of the pH value for
chlorine dioxide (—-), chlorine (-8-) and
chloramines ( - - -)
-------
-' 505. ,-..
Transformation, of£ ,chl
-------
- ,506
with powdered activated carbon, ozone ,and, chlorine
dioxide, showed a satisfactory result by using a
mixture of chlorine and chlorine dioxide, Fig. 5-1.
uj lo-
SS
r.VW.-.
Fig. 5-1
Results from taste control
investigations
LOT SAHD
FILTER
ACTIVATED CARBON
3-5 9/»3 6-8 <
EHim BEFORE TREATMENT
[V^.v] AFTER TREATMENT
OZON
1.7 i
Subsequently the raw water source - the river Gota
Hiv,- was contaminated by wastes and this, of course,
affected the taste and odour. The raw, water intake was
rebuilt and the river water was.pumped to a natural
lake acting as a reservoir..Together with increased
river pollution control, the situation was markedly
improved. An additional problem then arose occasionally
- during spring and autumn a mass production of algae in
the lake resulted in taste and odour in the water. On
such occasions chlorine dioxide treatment alone was
insufficient and therefore granular activated carbon
was introduced and has been used for nearly 10 years.
-------
- 5O7 -
In the taste" and odour investigations in Gothenburg
it was noted in agreement with others(1) that it is
possible to have a higher "chlorine residual" with
chlorine dioxide than with chlorine without producing
a chlorinous taste (Fig. 5-2).
Fig. 5-2
Taste intensity as a function
of the "chlorine residual"
from the different methods of
of treatment
0.05 0.010 0.015 0.020
CHLORINE RESIDUAL, mgj\
We conclude the following: chlorine dioxide can im-
prove the taste and odour and does not form haloforms.
Furthermore, ;chlorine dioxide is an excellent disin-
fectant in the distribution system and this facili-
tates the/supply of a safe and potable water.
-------
- 508 -
(1) U.S. Environmental Protection Agency
Ozone, Chlorine Dioxide and Chioramdnes as Alternatives
to Chlorine for Disinfection of Drinking Water
State-Of-The-Art, Nov. 1977
(2) GORDON, G., KIEFFER, R.G., ROSENBLATT, D.H.
The Chemistry of Chlorine Dioxide
Progress in inorganic Chemistry, Wiley-Interscience,
New York 15 (1972), 259-286
(3) GROB, K.r ZtiRCHER, F.
Stripping of trace organic substances from water,
equipment and procedure
J. of Chromatography 112 (1976) , 285
(4) EKLUND, G., JOSEFSSON, B., ROOS, C.
Determination of Volatile Halogenated Hydrocarbons in
Tap Water, Seawater and Industrial Effluents by Glass
Capillary Gas Chromatography and Electron Capture
Detection
J. High Res. Chrom. 1_ (1978), 34-4O
(5) EKLUND, G., JOSEFSSON, B., ROOS, C.
Trace Analysis of Volatile Organic Substances in
Goteborg Municipal Drinking Water
Vatten 31 (1978), No. 3
(6) MORRIS, J.C.
The Chemistry of Aqueous Chlorine in Relation to
Water Chlorination
Water Chlorination, Environmental Impact and Health
Effects
Ann Arbor Science, Ann Arbor _1_ (1978) , 21-35
(7) ROOK, J.
Chlorination Reactions of Fulvic Acids in Natural Waters
Environ. Sci. Technol. 1_1_ (1977), 5, 478-482
(8) ISACSSON, U., WETTERMARK, G.
Selective analysis of chlorine (hypochlorous acid) and
chlorine dioxide using chemiluminescence
Anal. Lett. r\_ (1978), 13-25
(9) KJELLANDER, I., LUND, E.
Sensitivity of Esch Coli and Poliovirus to Different
Forms of Combined Chlorine
J. AWWA 57 (1965),893
-------
- 509 -
(TO) RIDENOUR, G.M., ARMBRUSTER, E.H,
Bactericidal Effect of Chlorine Dioxide
J. AWWA 41_ (1949) , 537
(11) BERNARDE, M.A., ISRAEL, B.M., OLIVIERI, V.P.,
GRANSTROM, M.L.
Efficiency of Chlorine Dioxide as a Bactericide
Applied Microbiology 13 (1965), 776
(12) BERNDT, H.
Untersuchungen zur Wasseraufbereitung und Wasser-
desinfektion mit Chloridoxyd, insbesondere zu Fragen
der pH-Wert-Abhanaigkeit und der Chloritriickbildung
Arch. Hyg. 145 (1965), 1O
(13) FAIR, G.M., GEYER, J.C., OKUN, D.A.
Water and Wastewater Engineering, Wiley 2^ (1968), 31-18
(14) FERRER, E.B., STAPERT, E.M., SOKOLSKI, W.T.
A Medium for Improved Recovery of Bacteria from
Water
Can. J. Microbiol. 9^ (1963), 42O-422
(15) CRONIER, S., SCARPING, P.V., ZINK, M.L., HOFF, J.C.
Destruction by Chlorine Dioxide of Viruses and
Bacteria in Water
Abstr. Ann. Meeting American Soc. for Microbiol. (1977)
(16) WARRINER, R.
Inactivation of Poliovirus with Chlorine Dioxide
Vatten, Arg 23_ (1967), 284-29O
(17) ENGELBRECHT, R.S., WEBER, M.J., SCHMIDT, C.A., SALTER, B.-L.
Virus Sensitivity to Chlorine Disinfection of Water Samples
EPA-600/2-78-123
(18) LUND, E.
Inactivation of Poliomyelitis Virus by Chlorination at
Different Oxidation Potentials
Arch. Ges. Virusforsch. 11 (1961), 33O
(19) LUND, E.
Effect of pH on the Oxidative Inactivation of
Poliovirus
Arch. Ges, Virusforsch. 12 (1963, 1
(2O) Investigation of Different Methods for Taste and Odour
Improvemen t
Chalmers Univ. of Tech., Dept. of Water Supply and
Sewerage, Series B 64 (1964) , 1
-------
- 510 T- , :.-
REMOVAL OF ORGANIC MATTER FROM WATER BY UV AND
HYDROGEN PEROXIDE
L. Berglind, E. Gjessing and E. Skipperud Johansen
1. Introduction
In 1966 Armstrong and coworkers (1) introduced a method
for the removal of organic matter from samples of water
to be analyzed on total phorphorus. Samples of alcohols,
carbohydrates, organic acids and humic acid in aqueous
solutions were exposed to ultraviolet (UV) irradiation
in the presence of oxygen. It was found that the content
of organic compounds was reduced below detectable levels.
The only final product which was registered was carbon
dioxide, CO2.
It has been noticed that sunlight may induce both fading
of the colour and reduction of organic matter in water (2,3),
Elimination of organic pesticides has also been carried out
by photo-induced oxidation (4,5) as well as photo-induced
decomposition of formic acid and dodecyl benzene sulfonate
in aqueous solutions (6,7) . Photo-oxidation of refractory
organic compounds in municipal wastewater has been given
increased attention during the past years (8,9). The process
of oxidation is usually very complex. Some highly reactive
intermediates are involved. One of these is the hydroxyl
radical OH*, which is formed during light absorption (7,1O,
11). The rate of oxidation is primarily limited to the yield
of OH*. For the purpose of water purification, the method of
photo-oxidation requires an excess of OH.' One way to obtain this
is by adding hydrogen peroxide (H2O2) to the system. It
is known that UV irradiation may induce breakdown of H2O2
into hydroxyl radicals and thereby accelerate the process
of oxidation (1,11,12).
-------
- 51 1 '-
2. Earlier Results •
In the late sixties we suggested that some 20 % of the
observed reduction of humus colour in lakes and rivers
may partly be due to natural UV radiation and that this
principle should be considered as an alternative method
for the purification of surface water based on artificial
UV (13). Laboratory experiments showed that humus in
water, which is persistent soil originating organics,
were completely mineralized by a combination of UV radi-
ation and hydrogen peroxide. Several other oxydants were
also tested together with the UV, such as O3, O- and air.
However, H-O- appears to be outstanding with respect to
these effects. Some results 'are summarized in Table 1.
TABLE 1; Effect of UV radiation of aquatic humus.
Per-cent reduction of colour (or increase(+))
and organic carbon by different "doses" of
UV radiation under various conditions.
"Chemical"
added
Air
N2
Air
Air
H2°2
Cont
Ba02
(PH 7)
(pH 3)
.aeration
+ air
Col
Minutes of
1
+ 2
+ 1
+ 12
+ 5
45
6
11
5
+ 6
+ 2
+ 1O +
4
97
O
8
our
Radiation
20
13
11
8
14
98
21
10
60
41
16
25
35
99
. -
-
' Organic
Minutes
1
6
O
6
0
43
2
9
5
13
5
9
3
73
2
1 1
Carbon
of Radiation
2O
23
11
16
14
92
10
16
6O
41
26
19
15
100
-
—
-------
- 512
Essentially the same procedure as that used for humus was
also used on an aqueous solution of the polyaromatic
hydrocarbon, 3.4-Benzopyrene (B(a)P) : 99 % removal from a
100 yg/1 solution resulted.
It is strongly believed that the combination of UV and
H000, being so effective as to mineralize humus (and
£* «&>
B(a)P), might be a potential method for the removal of
organics from water. However, is it practical and what
will it cost? The purpose of the present work is to gather
more information around these two questions.
3, Recent Results
In a set-up as illustrated in Figure 1 different kinds of
organic matter were recirculated for several hours. Samples
were collected periodically. In Table 2 the tested organics
and chemicals are listed.
Wedeco
E/10-2
uv—
uv
210W
Pump-
^jfJA.
xT^xxxxvxxxx/v
,»
100 liter
Cooler
Fig. 1 UV-radiation plant. The radiation unit consists of a
quartz tube surrounded by 6x3O W U¥ lamps (254 nm).
The hydrogen peroxide is added to the water in the
tank immediately before start. The pumping rate was
O.8 I/sec.
-------
- 513 -
TABLE 2; The water samples and chemicals tested in the
UV pilot plant
1 . Humus water
2 . Humus water
3 . 3 , 4-Benzopyrene
4. 2-Methylisoborneol (MIB)
5 . Chlorof orm/bromo-
dichloromethan (BDCM)
Colour
mg P-t/1
109
27
-
-
-
TOG
mg C/l
14,5
4,2
-
-
-
Cond.
yS/cm
31
32
-
-
-
pH
4,9
6,2
-
-
-
Cone.
yg/1
_
-
65
1
1 OO/1OO
Humus
Eight 50-litre aliquotes of humus sample No. 1 were circu-
lated for four hours (O.8 I/sec) in the system illustrated
in Fig. 1, with different amounts of 35 % H202 added,
ranging between O and 5O ml/1. Water samples were collected
after 3O, 6O, 12O, 18O and 24O minutes and analyzed on
colour and TOG. The relationship between colour- and TOG
removal and the amount of H-O., added is illustrated in
^ £
Fig. 2. 12O minutes radiation is considered.
Humus sample No. 2 showed the same H-0., response with regard
^ £
to humus removal, namely: maximum effect with O.2 ml of
35 % H20? per litre. It is interesting and important to note
that higher concentrations of H?O2 did not result in a better
removal of organic matter and that the amount of H2O? appears
to be independent of the amount of humus.
-------
- '514 -
WO
10
HOC
Radiation time 120 min
Q = 0.81/Sec,
Energy used:420Wh
Volume of sample : 501
\COLOUR
'.0
ml 35% H20,/I water
20
Fig. 2 The relationship between TOC colour removal and the
amount of hydrogen peroxide (H7O7) added at a fixed
11UV dose"
In Figure 3 the "UV dose"-response relationship regarding
colour removal with a fixed ^2°2 dose ^°*2 ml 35 % H2°2/1
is plotted. Fig. 4 shows the corresponding reduction of
total organic carbon.
100-
Fig. 3 .
Reduction of humus colour
with different "UV doses";
O.2. ml 35 % H2O2/1 added.
The low-coloured water is
also run with low flux
(O.2 I/sec). It was noticed
that the colour of the samp-
les collected during the run
was reduced during 1-2 days
of storage
•5 30 60
Minutes ol UV- Radiation
120
-------
-'515 -
Fig.
Reduction of total organic carbon
(TOG) in two different humus samp-
les -with different "UV doses" and
O.2 ml 35 % H2Q2/1 added. The TOC
analyses are to some extent un-
certain, probably due to some
interfering substances in the
samples.
30 SO 120 180
M mules of UV-Radralion
Figure 5 illustrates the percent reduction (with time) of
colour and TOC.
Fig. .5
Percent reduction of colour and
organic carbon in two different
humus samples with different "UV
doses" and with O.2 ml 35 % H2O2/1
added.
30 60 120
Minutes of UV-Radiolion
180
-------
- 516 T- •---:. ,
In the brief discussion below>the carbemi results,are omitted
because at present we are uncertain as to whether the routine
method used (wet oxydation in glass ampulla and IR deter-
mination of the resulting CC^) is appropriate for these
samples. Considering Figure 3, it appears that the same
"UV-dose" (and amount of ^02) removes, in the beginning of
the process, twice as much colour from the highly-coloured
water (1O9 mg Pt/1) as from that of low colour (27 mg Pt/1),
The results suggest that the apparent linear colour reduction
is levelling off when the residual colour is in the range of
10 mg Pt/1.
Figure 3 shows that there is only need for twice as much
energy to reduce the "1O9-coloured" water by 95 % as that for
"27-coloured" water. All this suggests that, under the con-
ditions outlined, the method appears to be relatively
most effective on highly coloured waters.
Organic Chemicals in Water
1OO litres of distilled water, containing the concentrations
and chemicals listed in Table 2, were recirculated for 1-4
hours. Samples were taken at different time intervals and
analyzed on the chemical substance concerned. The results are
given in Table 3. ...
The results are also illustrated in Fig. 6. In general,
these results show more than 9O % reduction of the chemicals
treated with UV radiation for 2 hours. MIBr which is one of
the chemicals responsible for a characteristic soil smell and
-taste, which may appear in drinking water reservoirs, is
removed below detectable limits within 3O minutes.
-------
- 517 -
TABLE 3; Removal of organic chemicals from water with
UV radiation and hydrogen peroxide
without H~0~
£. Z
1 O —
>-H X
P 4J
«J ft
g 05 ra
0 65(0)
30
,60
120 42(29)
24O
yg/1
e
IH -
0 "R
o a
-H 0
fi Q
U ffl
1OO(O) 1OO(O)
- -
92(8) 91(9)
42(58) 75(25)
- -
with O.1 ml 35 % H2O/1
1 a
^ o
^j Hi
^x O
-------
- 518 •-'- ' '- '"
4. Cost Estimates
Except for MIB and the low humus-coloured water, it appears
from the results that,-, with a H2O2 dose of O.1 - O.2 ml
35 %/l and at'12O minutes of radiation, a 9O - 95 %
removal is achieved.
Considering 5O % removal and the following price estimates:
Electric energy
H202 (35 %)
UV-lamps
Pilot plant
Lifetime of UV-lamps
1O.OOO hours
1O % interest of
Plant cost
Nkr. 2.2OO.
Nkr. 15.OOO.
Nkr/kWh O.15 (O.O3)
Nkr/1 3.85 (0.70)
Nkr/h
Nkr/h
0.22 (0.04)
0.17 (0.03)
In Table 4 the resulting costs are shown.
TABLE 4; Estimated cost of water with 5O % removal of the
organic matter. All figures are in Nkr (or $)
per m3.
Sample
1
2
3
4
5
6
Humus 1
Humus 2
B(a)P
MIB
Chloroform
BDCM
Mean %
H2°2
.8
.8
.4
.4
.4
.4
14
El.
Energy
.2
.4
.2
.1
.4
.2
6
Lamp Plant
1.5 1.1
2.7 2.1
1 .5 1 .1
.6 0.4
2.5 1 .9
1.7 1.3
46 34
Total
Nkr. $
3.6
6 .O
3.4
1 .5
5.2
3.6
1OO
.65
1 .10
.62
.27
.94
.65
1OO
-------
-SIB -,,,' ...
It appears from the figures given in Table 4 that the main
costs(8O %) are connected with the equipment, whereas the
energy and hydrogen peroxyde expenses are in the range of
1 Nkr/m (* $-18/m ). For a 95 % removal, the costs are from
4O- % to several hundred per cent higher. ;
5. General Discussion
Laboratory experiments have suggested that organic matter
in water is quantitatively converted to C07 and H9O by
*-n £*
UV radiation and hydrogen peroxide. For drinking water
purposes this seems ideal: no foreign chemicals added and
no residual-component problems I However, the present pre-
liminary experiments suggest that the costs are high.
Nevertheless, the method should be further investigated.
It is important to emphasize that the mechanisms behind
this mineralization of aquatic organics are not clearly
understood and that an increased efficiency may be
achieved by a combined use of theoretical consideration
and recent research and practical experiments.
We find it reasonable that the work is continued along
the following lines, experimenting on:
1. Action, of cataysts ; •
2. Combination of chemicals
3. The most effective wave length
4. Increased efficiency of DV-lamp
5. Most effective way of dosing the active .oxygen
6. Identification and biological evaluation of
residuals , ' • "
-------
- 520 - '-. 7
It has been stated that some reactive intermediates are
involved which are formed by the action of light. Hydroxyl
radicals are, as mentioned in the introduction, suggested
as such an intermediate (7,10,11).
Zepp and coworkers (14) are .using the term "singlet oxygen",
stating that this may be generated in water by light energy
and oxidize a variety of organic substances. They suggest,
as Kautsky, that the most likely mechanism for the oxygena-
tion in the environment is that light energy adsorbed by a
sensitizer is transferred to ground-state oxygen to form
singlet oxygen, which in turn reacts with the organic sub-
stance or "acceptor" to form a peroxide. In their experiments
they demonstrate the importance o'f the presence of "singlet
oxygen sensitizer". The results outlined above indicate a
catalyzing or sensitizing action of some constituents
in the humus sample, and a combination of chemicals may
accelerate the mineralization processes.
In the work of Zepp and coworkers (14) the 366-nm line
is used. In the work presented above, the UV source is a
lamp which is claimed to have a sharp maximum at 254 nm.
It is reasonable to assume that the wave length of the
radiation will greatly influence the results. The UV-
lamps commercially available have a relatively low effect.
Those used in the present work are reported to give less
than 25 % of the energy used..Hpweyer, according to the
manufacturer, this will be improved in the future.
The hydrogen peroxide is added to the test water immedi-
ately before the start of the experiment. It is interesting
to note that the most effective H2O2-dose appears to be
independent of the concentration of the organic matter
present. The fate of the I^O^ added to the polluted water
during the 2-4 hours lasting experiment is, however, not
-------
- 521 --• .*;••• -
completely known. It is possible that the mineralizing
effect may be improved by applying the active oxygen con-
tinuously or stepwise during-the run .
The "ghost" behind drinking water and drinking water
treatment today is the organic residuals. The engagement
with regard to chlorination of water containing organics
in general and humus in particular is considerable through-
out the whole world. This 'Concern about the residuals and
their potential toxicity should also be applied to both
natural and artificial UV radiation. In the present work
some analyses on the residuals .after UV-H2O2 treatment
are in progress. At any rate, .this type of research is
essential in future, work with this principle of drinking
water treatment. ••.-..•'
6. Conclusion . . ;
There is need for a method by which it is possible to
eliminate organic matter from 'drinking water. Combined
use of UV radiation and active oxygen appears in prin-
ciple to be a promising method in this respect.
Experimental results from a relatively small UV radiation
plant (necessary residence time 5-2O h/m ) suggest that
the price of this treatment Is" high. The mechanisms involved
are, however, not completely understood and the efficiency
of the method may be improved by altering the conditions,
studying the effect
- at different wave lengths"
of a combination of chemicals
of different ways of applying the H-O, of modern
UV-lamps and.of different catalysts.
-------
- 522 .-
(1) ARMSTRONG, F.A.J., WILLIAMS, P.M., STRICKLAND J.D.H.
Photooxidation of Organic Matter in Sea Water by
Ultraviolet Radiation, Analytical and Other Application
Nature 211 (1966), 5O48, 481-483
(2) BEATTIE, J., BRICKER, C., GARVIN, D.
Photolytic Determination of Trace Amounts of Organic
Materials in Water
Anal. Chem. 33 (1961), 189O-92
(3) GJESSING, E.T., SAMDAL, J.E.
Humic Substances in Water and the Effects' of .
Impoundment
J. AWWA j5O (1968), 4, 451-454
(4) HENDERSEN, G.L., CROSBY, D.G.
Photodecoraposition of Dieldrin Residuals in Water
Bull. Environ. Contain. Toxicol. _3 (1968), 131-134
(5) BOLLER, C.D., EDGERLY, E. jr.
Photochemical Degradation of Refractory Organic
Compounds
J. WPCP 40 (1968), 546-556 ;
(6) MATSUURA, T., SMITH, J.M./.'
Photodecomposition Kinetics of Formic Acid in
Aqueous Solution
AIChE J. 16 (1970), 1064-1071
(7) MATSUURA, T., SMITH, J.M.
Kinetics of Photodecomposition of Dodecyl Benzene
Sulfonate
Ind. Eng. Chem.Fundam. 9_ (197O), 252-26O
(8) SCHORR, V., BOVAL, B., HANCIL, ¥., SMITH, J.M.
Photooxidation Kinetics of Organic Pollutants in
Municipal Waste Water
Ind. Eng. Chem. Fundam. 10 (1971), 4, 5O9-515
(9) MAUK, C.E., PRENGLE, W. .
Ozone with Ultraviolet Light Provides Improved Chemical
Oxidation of Refractory Crganics
Pollut. Engineer. 8^ (1976), 42-43
(1O) FOOTE, C.S.
Photosensitized Oxidation and Singlet Oxygen Conse-
quences in Biological Systems
Free Radicals in Biology II, -Academic Press, New York
(1976), 85-133
-------
523 -
(11) KOUBEK, E.
Oxidation of Refractory Organics in Aqueous Waste
Streams by Hydrogen Peroxide and Ultraviolet Light
US Patent Nr. 4, 012, 321 (1977)
(12) SCHENK, G.O.
Institut fur Strahlenchemie, Max Planck-Institut fur
Kohlenforschung, Mxilheim, personal communication (1977)
(13) GJESSING, E.T.
Influence of Ultra-Violet Radiation on Aquatic Humus
Vatten 2_ (197O), 144-145
(14) ZEPP, R.G., WOLFE, N.L., BAUGHMAN, G.L., HOLLIS, G.L.
Singlet Oxygen in Natural Waters
Nature 267 (1977), 421-423
-------
- 524.-
i-
ANODIC OXIDATION AS A PROCESS STEP IN THE TREATMENT ,OF
BACTERIALLY CONTAMINATED WATER
N, Kirmaier
Extensive scientific studies in recent years have been, able
to demonstrate the effectiveness of the bacterial decontam-
ination of water by anodic oxidation [1-3].
The original hopes of being able to use anodic oxidation
specifically in electrosynthetic processes were disappointed
in most cases on account of the formation of undesirable by-
products. Nevertheless, there was a positive outcome for the
treatment of water: microbial constituents and oxidizable
chemical substances present in the water are oxidized unspeci-
fically. In this way microorganisms are inactivated and the
chemical loading of water is degraded.
The electrochemical action of anodic oxidation in water treat-
ment can be arranged under two group headings:
a) the actual step-through reaction at the phase
boundary between the electrode and the electrolyte,,
b) the secondary reaction @.ue to reaction products)
within the solution.
Since as far as we know cathodic reactions do not cause
bacterial inactivation, and chemical reactions cannot.be
solely responsible for the inactivation performance of the
anodic oxidation reaction cells, the actual oxidation of the
bacterial substrate of microbes must take place on the
electrode surface or on the boundary region of the anode. The
basic course of the reaction can thus be described by the
general equation:
-------
- 525--
Ox (I) + Red (II) ^ > Red (I) +.Ox (II)
where Red (II) stands for an active bacterium and-Red (I)
for an inactivated bacterium [4].
If the bacterium were inactivated by active oxygen, the
reaction would depend on processes of diffusion through the
bacterial cell wall and on further partial steps of the
reaction in the cytoplasm and the cell nucleus up to the
. • ' >
oxidative attack on the DNA or on messenger RNA [4].
Studies were carried out with waters (mains water, surface
water) contaminated with viruses and bacteria. Fig. 1 shows
the inactivation dynamics for the polio virus in Munich mains
water, and Fig. 2 the inactivation dynamics for the ECBO
virus [5]. Only a relatively slight reduction of the infec-
tiousness could be achieved at low current densities, while
2
with current densities of over 5 mA/cm a complete inactiv-
ation of the test viruses used is possible [5].
Other problem microorganisms (in concentrations of up to
10 cells/ml) '(gram-negative bacilli, aerobic spore-formers,
yeasts, mycelial fungi, and gram-positive cocci), can likewise
,be fully inactivated.
-------
- 526.--
* Water 400 pS
fc Water 800 pS
zoo
300
Current, mA
Fig. 1 Inactivation dynamics of the polio virus in Munich
water, sterile filtration? through-flow 7.5 ml/sec
at 1O-12 C; conductivity adjusted with sodium chloride
H
<
o
tn
O
r-l
C
fl)
4J
C
O
o
M
•H
* Water 400 pS
O Water 800 pS
Current,
Fig. 2 Inactivation dynamics for the ECBO virus (5)
-------
- 527' -'-
.(1.) Reis, A.
Die Anodisclxe Oxidation als Inaktivator pathogener
Substanzen und prozesse , . ' .. • "
Klin. Wschr. .11951), 484
(2) REIS, A., KIRMAIER, N.
Anoden-Warturiscrubber zur Luftreinigxing
,G-I~T. Fachz. Lab. 22 (1978)., 197
(3) REIS, A.
Ingenieuraufgaben in der Medizin
Biomed. Technik, Oldenbourg-Verlag, Mtinchen/Wien
2. Aufl. (1977)
(4) EIBL, V.
Dissertation, TU Miinchen (in preparation)
(5) MA.HNEL, H.
Virusinaktivierung. in Wasser durch Anodische
Oxidation • - :. . . . .
Zbl. Bakt. Hyg., Orig. B 166 (1978), 542
-------
- 528- - ,
EXPERIENCE WITH POTASSIUM PERMANGANATE
K. K6tter
1. Preliminary remarks
When a conference considers the problem of oxidation processes
and in so doing gives emphasis to chlorine and ozone,potassium
permanganate should not be omitted from the discussion. This
oxidizing agent undoubtedly does not rank among the primary
agents used in water treatment, but it does have some inter-
esting properties. Reference should therefore also be made
to these, even though there are already detailed publications
on potassium permanganate, on its chemical reactions, and on
i-ts possible uses in water treatment [1,2, etc.].
This short experience report will be limited to the use of
potassium permanganate in recent years on Gelsen water.
Before this, however, a brief mention of the commercial situ-
ation of potassium permanganate: a major manufacturer of this
compound estimates that potassium permanganate is currently
in use at more,than 250 waterworks in North America and at
more than 30 in Europe. For three years there has been a
German Industrial Standard for this product in West Germany
[3].
2. Use of potassium permanganate to prevent the growth
of filamentous algae in infiltration reservoirs
The artificial enrichment of ground water with slow sand
filtration with the use of large infiltration reservoirs is
regarded as the principal process step in the Ruhr valley
and in other large water treatment plants. Mass growth of
algae and its adverse effects on this method of obtaining
water and on the water quality have been known for a long
time. In the spring the microscopic plankton algae predom-
inate, in the summer it is the macroscopic filamentous algae
-------
*
Fig. 1 Mass growth of algae in infiltration reservoirs
-------
Ul
U)
O
Fig. 2 Swimming filamentous algae as in Fig. 1, photographed from a
closer distance
-------
- 531 -
(Figs. 1 and 2). In the infiltration reservoirs the algae
find conditions favourable for their proliferation,- namely a
slowing down of the flow rate, favourable illumination
conditions, and a steady supply of food. Since the algae
constantly release harmful metabolic products and on dying
give off harmful incorporated substances, it is better to
prevent the algal growth than to try and destroy it later,
especially as relatively few chemical controlling agents are
permissible in the treatment of drinking water.
The action of potassium permanganate on the growth of fila-
mentous algae was studied at Gelsenwasser AG in large-scale
trials lasting several years, performed in varioiis water-
treatment plants on the Ruhr and in reservoirs [4,5]. An
addition of 1 to 1.5 g/m of,potassium permanganate 'to the
inlets into the infiltration reservoirs largely prevents the
growth of filamentous algae, while in a parallel trial with
untreated tanks dense .mats of Hydrodictyon were produced.
The action mechanisms of the potassium permanganate remain
largely unexplained. They could be based on its oxidizing
properties or on other properties toxic to the algae.
The considerable reduction of .the water's transparency to
light undoubtedly plays -an important part. Certain wave-
lengths are absorbed to the extent of up to 99%. Therefore
the light conditions at the bottom of the reservoir are. no
longer adequate for the development' of the algae.
To stop the growth of filamentous algae the KMnO^ addition
must be commenced promptly in the spring. When already
formed, algae swimming,to the surface are little affected and
can reproduce on the well illuminated surface of the water.
In the light of the positive experience, potassium perman-
ganate dosage plants have in the -meantime been installed in
all the Ruhr waterworks of Gelsenwasser AG, and since the
-------
- 532 -
spring of 1977, with; .the. except ion of ..some,, remotely situated >
reservoirs, all the artificially Infiltrating water in these
Ruhr waterworks has been first, treated with potassium per-
manganate.
3. Action of potassium permanganate on plankton algae
in infiltration reservoirs • . .
The favourable action of potassium permanganate unfortunately
extends only to the filamentous algae, and not to the plankton
algae. This may be partly due to the fact that these algal
forms develop in free water bulks, ~ where because of the
small layer thickness the light absorption of the permanganate
has less effect than at the bottom of a reservoir.
In the summer of 1973 an attempt was made at Halingen water-
works to prevent the growth of plankton algae by increasing
3
the amount of potassium permanganate added to 4 g/m . The
treated water - in some reservoirs "also receiving 0.3 to
3
0.6 g/m of copper sulphate - assumed an intense violet
coloration. Interestingly enough, these additions had no
clearly positive or negative effects, either on the plankton
algae or on the biological purification performance- of the
slow sand filter. Nor could any harm to the zooplankton be
established.
On the over-all evaluation of potassium permanganate as an
anti-algal agent in water treatment it must be said that it
is sadly less effective against the plankton algae, i.e.
against the algal species that at times impair the seepage.
However, the filamentous algae, whbs'e biomass production far
outweighs that of the plankton algae' and that can give rise
to a considerable impairment of odour, can be controlled by
potassium permanganate and so far no disadvantages of any
kind have been discovered. This on the whole positive balance
has lead to the decision to add potassium permanganate contin-
uously to the raw water between March and September.
-------
- 533 - -
4. '• • ' '• Potassium permanganate* as" 'an agent-for the oxidation
of dissolved manganese in an oxygen-containing water
The following partial report concerns the water treatment of
the Stiepel waterworks of Wasserbeschaffung Mittlere Ruhr
GmbH (WMR). The waterworks are situated in the Ruhr valley
and are operated with Gelsen water. Up to 2500 m /h of true
ground water and Ruhr baiiK filtrate are used. Tnere is no
artificial ground water enrichment via infiltration reservoirs
or absorption wells. The manganese contents in the drinking
water rose in the 'sixties to 5 - 8 mmole/m (0.25 - 0.40 mg/1)
of Mn> although the oxygen content in the natural ground water
and ground water enriched by bank filtration, i.e. in the
mixture of raw waters of different origins, never fell below
3 mg C>2/1. By shutting down the most severely affected wells
a temporary slight improvement of the mixed water quality
could be obtained, but this measure was not regarded as a
satisfactory long-term solution. The only promising counter-
measure was the erection and setting in operation of a deman.—
ganization plant.
In collaboration with drinking water treatment companies and
with the Engler-Bunte Institute of Karlsruhe University,
various treatment trials were carried out with the particular
aim of clarifying whether a biological or autocatalytic
manganese oxidation takes place in water containing sufficient
oxygen or whether the use of oxidizing agents would be necess-
ary. At first biological oxidation was favoured, because this
had proved to be effective and reliable in the reservoirs
used at the Haltern waterworks (up to 50 m/h filtration ve- •
locity and 10,000 m/h throughput). At the Stiepel waterworks,
on the other hand, a filtration preceded by chemical oxidation
gave better results. Surprisingly, potassium permanganate
proved to be a more effective oxidizing agent than chlorine.
For this reason, and because without preliminary chlorination
the formation of organic chlorine compounds does not come
-------
T 534 -
into consider at ion",™!" "plant" "for" "long-term potassium permanganate^
addition was set up and put into operation. The construction
and the operating results are presented in Fig. 3 and in the
following list:
KMnO. addition
: a constant 0.5 g/m , i.e. referred
to the fluctuating manganese concen-
tration in the filter inflow 0.9 to
2.1 times the stoichiometric amount.
Filter bed
70 cm of pumice, 2-3 mm, bulk
density 350 kg/m3
70 cm of activated, carbon, 1.5-2 mm,
bulk density 700 kg/m3
70 cm of gravel, 0.7 - 1.2 mm, bulk
density 1600 ,kg/m
Filter velocity
25 m/h max.
Running times
: about 7 days
Throughput/running
time
Manganese content
pH~ (corrected by
additions of NaOH)
about 2500 m3/m2 :
before the filter 2-6 mmole/m
Mn = 0.1 - 0.3 mg/1 Mn
after the filter 0.0 - 0.6 mmole/nf
Mn = 0.00-0.03 mg/1 Mn
ground water 6-9
filter inflow 7.4-'
waterworks outlet '7.8
-------
- 535 -,
Control
!Chlorine : ^
l —T^pH ' I
| • Buflrtt jaravel I I
J L J C J , , , ,
~^i r-1i Y—-i i—« T~—i r—1 - : r~. Chlorine
_J I I! _ I I U_j U> - ...- - to cons
Fig. 3 Stiepel waterworks of -the WMR, scheme of the
treatment, metering of potassium permanganate
into the draw-wells upstream of. the filters
5.
Oxidation of manganese in oxygen-free water
In contrast to the Ruhr bank filtrate, the Rhine bank filtrate
coming from the gasworks and waterworks on the Lower Rhine is
practically free from oxygen. The result is that in spite of
its relatively high content of neutral salts it is only
slightly corrosive. This, situation would'be changed if the
freedom from oxygen were lost during- the treatment. For this
reason NGW did not make use of a-biological or autocatalytic
manganese oxidation in the process-technical design of the'
treatment plan for waterworks 1, and selected potassium per-
manganate as an oxidizing- agent. The process scheme (see
Fig. 4) has already, been described by Heymann [6] at one of
the 'earlier Karlsruhe conferences.
-------
- 536 '«
Ctuitic todn
\
\ T
j Preliminary filter * "&
0. 3 m Activated carbon, 3-5 sas 1
1.O m Anthracite, 1.6-2.5 inn
?. O.S m Filter asnd, o.5-t em J
f" O.2 » Tilter gravel, 1,5-2.5 las V
j
( Sflconaary filter
3
cti
container
^
.-44
^t
^•-T x-?-
?ere vator f j f
s -»
u
-.•i
/
5/w^
J_jv- - ._*^-,,, 1 i
> > Ha
I — \ 21 IMno. and *1 iol«li
\\ !? I • 1
s
1
I '
1 1
K «
n
1
^i »
Vv
v\
Braakoolnt cttlorin
Sodiun
hypochlorite
f — )j
1
t
n v
1
.M-TOlUti
n
1
cp
€H "-€!)—'
i.-J
•^peculation ng*nt
Fig. 4 Waterworks 1 of the NGW, scheme of the treatment,
metering of pptassium permanganate into the draw
flow upstream of the filters
The potassitun permanganate is added in proportional
amounts to the raw water in the .form of a 1% solution. The /•
3
specific amount of potassium permanganate (in g/m ) required
2+
is 2.1 times the weight concentration of the Mn , i.e. the
oxidizing agent is added in a .slight excess over the sto.ichio-
.»*
metric amount. The flocculating action of the manganese
hydroxide formed on organic water constituents is increased
•i
by the addition of aluminium chloride. After a residence
time of 10 to 20 min the flocks produced are removed,
following stabilization with a polymeric flocculation aid,
through a three-layer filter. For the reduction of the
excess permanganate ion the upper layer consists of activated
carbon. This process removes iron and manganese ions to below
the analytical detection levels.
-------
- 537
Filtration of the water through a fibre-glass filter has been
shown to be a simple check on the potassium permanganate
dosage. The amount added is- sufficient when a weak red
coloration is discernible at a layer thickness of about 5 cm.
Fully automatic regulation of the potassium permanganate
•
addition is possible by continual measurements of the excess
KMnO,, but this is not necessary with the small changes in
the concentrations of iron and manganese in the Rhine bank
filtrate.
6. Further experiments on manganese oxidation
The Witten waterworks of Gelsenwasser AG are situated in the
Ruhr valley about 8 km upstream of the Stiepel waterworks.
The so-called pure water (up to 5000 m /h) obtained from
below ground consists of small proportions of bank filtrate
and true ground water, but mostly of artificially enriched
ground water. It contains 1-2 mmole/m = 0.05 - 0.10 mg/1
of Mn. Because of the'various origins d~ the manganese, only
a partial load would be treated, by demanganization of the raw
water to be infiltrated.
Experimental treatments were performed following the prototype
method of the Stiepel waterworks. These gave the surprising
result that potassium permanganate, in spite of the physical
proximity of the two waterworks and at first glance similar
raw water situation, did not here give the same good results.
Therefore, in the more recent large-scale plant the stronger
oxidizing agent ozone is used. This example shows that
potassium permanganate, despite its advantages from the
economic and other points 'of view, is restricted in its uses
and the limitations must be"investigated in each individual
cape. '•*..'•
-------
- 538 -
7. Disinfection of newly laid pipes with potassium
permanganate
The disinfection of newly laid pipes is normally carried out
o
with chlorine water in a concentration of about 50 g/m C1-.
In recent years various waterworks found that a single dis-
infection does not give the desired freedom from bacteria
even with 24-h duration of action. This is particularly true
for pipes with a high wall alkalinity (concrete pipes, pipes
covered in cement mortar, pipes of asbestos cement, etc.),
presumably because of the reduction in the bactericidal
action of chlorine at higher pH.
Repeated high chlorination is a possible means — though
troublesome — of avoiding this disadvantage. The DVGW
pamphlet W 291 "Disinfection of Water Supply Plants" [7], in
the light of the. good experience particularly at the Hamburg
waterworks [8], which were the first to practise this process
on a large scale, contains the proposal to use potassium
3
permanganate in a concentration of 5 - 10 g/m . At Gelsen-
wasser AG sodium hypochlorite and potassium permanganate have
been used together for years to disinfect newly laid pipes.
The disinfectant solution normally contains 45 g/m of C^
and 12 g/m3 of KMn04.
Fig. 5 shows the dependence of the redox potential of various
disinfectant solutions on the pH. Similar results were
obtained earlier by Gras and Konrad [9]. If we start from
the assumption that the redox potential is a measure of the
bactericidal action of disinfectants not only at various con-
centrations of a particular disinfectant but also in a com-
parison of different disinfectants, the greater effectiveness
of pure potassium permanganate in comparison with chlorine
is not quite comprehensible, but that of the disinfectant
mixture is easily interpreted.
-------
539
400
Fig. 5
Redox potential of
various solutions of
permanganate and/or
chlorine, measured
against an SCE in
dependence on the pH
PH
Under working conditions, the disinfectant mixture seems to
lead to the desired freedom from bacteria more' rapidly than
pure chlorine. The only disadvantage is the greater expend-
iture- in the preparation of the solution, but' against this
is the advantage of the coloration of the .disinfectant sol-
ution in addition to the enhanced effectiveness already
mentioned. During the addition flushing can take place at
the hydrant until the red-violet solution appears, and with
more recent rinsing analytical aids can be omitted. A visual
check on disappearance of the disinfectant is sufficient.
-------
- 540 - • • . .
Obviously it must first be checked whether a canalization or
drainage is available, into which the solution of disinfectant
containing the potassium permanganate can be harmlessly
introduced.
8. Transportation and metering of potassium permanganate
Potassium permanganate is usually supplied in the solid form
as a granulate in 50 kg drums. At the place of its use
either a solution of constant concentration is prepared and
metered out in liquid form (see, for example, Fig. 6) or a
dry dosage is selected, regulated directly in accordance with
the fluctuating volume flow of the water. All processes can
be automated, and it is advantageous if the potassium per-
manganate granulate is "free running". A small addition of
silicate base (according to the manufacturer's information)
clearly reduces caking of the material, even at an elevated
relative humidity.
When the potassium permanganate is used in larger quantities
the necessary opening, emptying, and subsequent removal of
the sheet-metal drums is a certain inconvenience. On the
suggestion from Gelsenwasser one of the larger potassium
Free-running KHnO4
to injection point
7
\T
Charging hopper
Dry metering unit \^ Moving membrane ! !
jVvl ' '*! i- Metering coil i !
:- ~ '"z* ', ' ~ ^*s.f ^|LN- f'i^-
--^J.'..'.-£-£'.^-> ' ".\ ,-l Level controller >^k' '^L,' Metering raump
*
/ ' \
Dissolution chamber Residence chamber Supply chamber
Fig. 6 Plant for the treatment and metering of potassium
permanganate solution of constant concentration,
schematic; after (2)
-------
- 541 -
permanganate suppliers obtained in 1975 special permission
from the Federal Transport Minister to transport the potassium
permanganate on roads in stackable cubic containers (tank
containers) of up to 1050-litre capacity. At the same time
a filling installation for the tank containers was built at
f . •
the manufacturers. In parallel with this, five metering
stations of the.kind shown in Fig. 7 were constructed at
Gelsenwasser. By using refined alloy steel tank containers
- shown in diagrammatic form - exchanged back and forth
between the suppliers and the consumer, the use of potassium
permanganate available in the solid form is substantially
simplified.
Y-
r"-- ''—"—..'^Dissolu-
i |" i"i '" \ 'jtion
-| j.i;J. [; '•> |water f
Seenaae reservoirs
Rpsprvnir inflow
^%^i i i i
\
Fig. 7 Potassium permanganate metering station with
delivery in containers, schematic
9. Summary
The present report of experience concerns the use of potassium
permanganate in the Gelsenwasser group. While potassium per-
manganate is being replaced more and more in the laboratory
-------
- 542 -
as an analytical aid for the determination of oxidizability
by stronger oxidizing agents, the technical product is used
increasingly in water treatment. Potassium permanganate
exhibits some characteristics, although sometimes in a weaker
form, of properties and possibilities of use typical of the
usual oxidizing, disinfectant, algicidal, and flocculating
agents.
- As an oxidizing agent, its action on organic
substances is weaker - but by the same token more
sparing — than that of ozone and, unlike chlorine,
it does not form reaction products that remain in
the water.
- As a disinfectant it is suitable for the dis-
infection of plants but not for disinfecting drinking
water at the end of the treatment.
- As an algicide it is effective within the described
limits and has the particular advantage over other
algicides of being permissible for the drinking
water treatment of raw water.
- It may not be regarded as a flocculation agent in
the true sense of the term, although small floes
with adsorbent and coagulating action are formed.
- Like some true flocculating agents, potassium
permanganate is delivered in the solid form as a
granulate and so requires a separate dissolution
step. As with other flocculation agents,, the
floes then formed must be removed in yet another
process step.
Potassium permanganate is a chemical additive for water
treatment processes which is used not in terms of individual
strong actions but rather as a combination of several weaker
actions.
-------
- 543 -
(]} REIDIES, A.H., MACK, E.
Die yerwendung von Kaliumperxnanganat in der Wasser-
aufbereitung in den USA
Vom Wasser 3O (1963), 81-1O3
(2).MACK, E.
Die Verwendung von Kaliumpermanganat in der
Trinkwasseraufbereitung . " •-
Informationssschrift der Pirma Th. Goldschmidt AG
6/73 (1973), 21, 2-23, siehe dort weitere Literatur
(3) DIN 19619 Kaliumpermanganat zur Wasseraufbereitung
Berlin, Beuth Verlag (1975)
(4) SCHMIDT, W.-D.
Probleme und Praxis bei der Algenbekampfung in
Infiltrationsbecken
Wasserfaehl. Aussprachetagung 1972 Dortmund
Tagungsheft Frankfurt, ZfGW-Verlag (1972), 28-35
(5) PlTSCH, B.
MaBnalimen zur Verhinderung tibermaBigen Algenwachsturns
in den Wassergewinnungsanlagen der Gelsenwasser AG
Bericht "Ruhrwassergiite 1976" des Ruhrverbandes/
Essen (1977)
(6) HEYMANN, E.
Erfahrungen bei der.Anwendung eirier Flockungsfiltration
vor den Aktivkohlefiltern
Veroffentl. des Bereichs und des Lehrstuhls ftir Wasser-
chemie, Engler-Bunte-Institut der Universitat Karlsruhe
(1975), 9, 107-118
(7) DVGW-Arbeitsblatt W 291 Desinfektion von Wasserver-
sorgungsanlagen, Frankfurt, ZfGW-Verlag (1973)
(8) GOLDA, W.
Desinfektion von Rohrleitungen und Trinkwasser-
aufbereitungsanlagen mit Kaliumpermanganat
Informationsschrift der Firraa Th. Goldschmidt AG
6/73 (1973), 27, 34-35
(9) GRAS, KONRAD
Klarung der Oxidationskraft des Kaliumpermanganats
in Abhangigkeit vom pH-Wert
Unverdffentl. Laborbericht 1972
-------
- .544 -
THE USE OF HYDROGEN PEROXIDE^IN- WATER,-TREATMENT ,
H. Overath
1. Introduction
In many respects hydrogen peroxide invites application as a
disinfectant and oxidizing agent in water technology:
- Hydrogen peroxide can be easily and exactly metered
out in the liquid state.
- Hydrogen peroxide solutions between 35 - 50% are com-
pletely safe when handled correctly.
- Excess hydrogen peroxide can be removed quantitatively
by activated carbon or by chlorine.
It is thus not surprising that hydrogen peroxide is from time
to time considered as an alternative to chlorine and ozone.
The following brief report attempts to clarify why hydrogen
peroxide has not yet established itself, and what are the
chances of its use in the future.
2. The most important chemical properties of hydrogen
peroxide
Hydrogen peroxide has three important chemical properties:
a) It is an, oxidizing agent
H_00 + 2H+ + 2e~ > 2 HO E = + 1.77 V
£ £» £ O
H,00 + 2e~ > 2 OH~ E = + O.88 V
£• £, (-)
b) It is a reducing agent
H9O_ > 2 H* + O- + 2e~ E = + O.68 V
t- £. £• O
-------
-. 545 -
c) It can decompose"Into oxygen'and water in. an inter-
molecular redox reaction
2 H202 » 2 H20 + O2 +/46.9 kcal
3. Examples of the application of hydrogen peroxide in ,
the used water and waste water sectors,
For some time now hydrogen peroxide has been used for the
treatment of industrial waste waters whose components are
toxic to activated sludge and which can be readily oxidized
by it. Thus cyanide is oxidized to cyanate, nitrite to
nitrate, formaldehyde to formic acid, and .phenol - particu-
2+ 3+
larly in the 'presence of Fe and Fe ions - to the less
toxic pyrocatechol, hydroquinone, and quinone (Table 1).
Since hydrogen peroxide can oxidize H-S to elementary sulphur
and thiols to dialkyldithiols, it has already been used for
odour and corrosion control in the industrial and communal
waste water sector (Table 1).
In highly charged waste waters hydrogen peroxide is able, as
an oxygen donor, to maintain O7 concentrations such as cannot
be achieved on the basis of the entry values for atmospheric
oxygen given by Pick's law. The same applies to the nitrifi-
cation of discharges of two-stage clarification plants by
filtration (Table 1).
While large doses of hydrogen peroxide can in many cases be
used in industrial waste water treatment, only concentrations
of 1C mg/1 at the most are justified in the treatment of
drinking water for hygienic and economic reasons. Since at
this concentration it is second only to ozone in oxidizing
strength, and it is known that on catalytic decomposition
hydrogen peroxide can form very reactive intermediate products
-------
Table 1 •« Use of "
o^
t*ie waste water sector
Function of the
Process
Reaction
as an oxidizing agent
as an oxygen source
-•Detoxification e.g. of
.- cyanide
".- nitrite
.- formaldehyde
.- phenol
-.Corrosion and odour control
.Oxidation of -SH
--Inflated sludge control
--Degradation of biologically oxidizable
•material (BOD)
-•Nitrification in sand filters
CN
HCHO
CNO
N03~
HCOOH
(Fe2*K
>H
H2S ~
RSH
RSSR
unknown
aerobic biological (not chemical)
oxidation
aerobic biological oxidation of NH_ i
by Nitrosomonas and Nitrobacter "
I
Ul
-------
- 547 -'
of sufficient lifetime, decades of effort have been devoted
to finding a suitable catalyst. -The stimulus is always the
hope that one's own work will be rewarded with the Discovery
of >the "correct" catalyst. The magical intermediate products
are called: - . . . •
OH radicals and singlet oxygen.
These can both be formed from hydrogen peroxide and are
extremely "fierce". •
4. Formation of the OH radical from hydrogen peroxide ,
The decomposition of hydrogen peroxide is strongly ;exothermic,
with an enthalpy change of AH = -22.62 kcal/mole. Water and,
oxygen are formed as the final products:
(H2°2>diesolved F=* {H2°2^ liq. + X/2 (O2^gas j
It is assumed that the non-catalysed decomposition of hydrogen
peroxide always proceeds _via hydroxyl and perhydroxyl radicals
in a non-chain reaction:
(1) H202 » 20H- ;. • ' , <
(2) OH- + H2O2 *'H2° + H0°*
(3a) HOO- + HOO * H2O2 + °2
(3b) HOO- + OH-
The initial reaction (1) is rate-determining. Since the diss-
ociation energy of the oxygen bond D(HO - OH) is 48.5 kcal/mole,
it is not surprising that pure hydrogen peroxide decomposes at
an immeasurably rapid rate.'
-------
- 548 -
The .thermodynamic instability of the hydrogen peroxide in
relation to its decomposition products H2O and O~, however,is
manifested in the feet that its decomposition rate is acceler-
ated catalytically by small quantities of a very large number
of substances in the dissolved and solid state. Practically
all types of dust and dirt, in particular certain heavy metal
ions occurring in several valence states, and further more or
less all vessel surfaces have this effect, various reaction
mechanisms having been established for the homogeneous and
heterogeneous catalysis.
If hydrogen peroxide is added to Rhine water, the natural
concentration of potential catalysts is not sufficient for a
rapid decomposition since, according to our measurements, the
half-life amounts to a few days. An additional catalyst is
required.
2
4.1. Production of OH radicals by catalysis with Fe
... •, 2+
In most cases the catalyst recognized as the best is the Fe
ion. In combination with hydrogen peroxide, it is known as
Fenton's reagent. The first hypothesis about the reaction
mechanism was put forward in 1934 by F. Haber and J.Weiss [1]:
START (a) Fe2+ + E2°2 *F-e3+ +,OH~,+ OH*
Unfortunately, in addition 'to the desired reaction with
organic molecules, the OH radical formed can enter, some other
undesirable subsidiary and radical-capture reactions (Fig. 1),
-------
549 -
desired
OXIDATION
REACTIONS
Fe2*- OXIDATION
MOO- O2 FORMATION
(radical chain
mechanism)
further RADICAL-
CAPTURE REACTIONS
+
Fig. 1 Reactions of Fe -catalysed
1.
It can (again according to Haber and Weiss) function
as a catalyst-independent chain carrier leading to
O- formation:
CHAIN
(b) OH
(c) HOO- +
H00
+ OH-
The new OH radical is .formed at the expense of two
hydrogen peroxide.molecules. Although the inter-
mediate perhydroxy radical can also react with organic
molecules, it is essentially more critical in this
reaction than the OH radical.
2.
2+
The OH radical can in addition oxidize Fe ions:
(d) Fe
2+
OH
3+
Fe 4- OH
-------
- 55u -
This route is not only; undesirable as a radical-
2+
capture reaction but it also removes Fe ions
which are required for the initial reaction (a).
Although, according to H.H. Baxendale [2], the
2+ 3+ •
Fe can be reformed' from Pe ions
(e) Fe3+ + HOO-
this reaction is undesirable since the OH radical
formation in accordance with reaction (c) is no
longer possible. Moreover, the rate of this
reaction is substantially smaller at pH 7 than that
3+
of the hydrolysis of the .Fe ions.
Finally there is • '
3. A number of other conceivable reactions, such as
hydration,rcharge-transfer-complex formation, etc.
that deactivate the OH radical and so remove'it from
the desired oxidative reaction.
In other words, the yields of OH radicals available
for the oxidation are related to the hydrogen per-
oxide used, i.e. the [OH*]/[HQO0] ratio, considerably
= £t £t
smaller than unity. ' • •
2+
Fig. 2 shows the Fe -catalysed reaction of hydrogen peroxide
with ethanol, one of the closer-studied reactions with an
organic compound. The H-abstraction by the, OH radical takes
place predominantly at the a-carbon. The oxidation to acetal-
dehyde and, further to acetic acid takes place preferentially
in the absence of atmospheric oxygen. If oxygen is present,
the re-formation of the alcohol shown below plays a major role
with the intermediate organic radical functioning as an
2+ ' " 3"+
electron carrier in the conversion of Fe into Fe
-------
- 551 -
Fe*
CH3-CH=0
OXIDATION
/ y-H20»HO.
L-Y
H202
CH3-CH2OH J^CH3-CHOH ^ CH3-CHOH
H3CHOH ^r13-UnUn
DIMERIZATION
HO.
[CH3-CHOH]
CH3-CH2OH
RE-FORMATION
Fig. 2 Reactions of the OH radical with ethanol
It is therefore understandable that concentrated and.acidified
mixtures of hydrogen peroxide and FeSO'4 are used; successfully
for the quantitative ashing of foodstuffs. However, in the
elimination of bacteria from drinking water the intensifi-
cation of the bactericidal action of hydrogen peroxide by
2+ ' • '
catalytic amounts of Fe ions is less spectacular. Results
from the Water Research Centre at Medmenham, England, sub-
stantially confirmed by studies in our laboratories in
Wiesbaden in collaboration with the Degussa Company, show that
even 120 mg'of H^O,,/! in 300 min destroy only 99% of E.coli.
2+ • •
If in addition 2.8 mg Fe . /I and 0.05 mM of EDTA are added,
the contact time necessary for'this is reduced to :110 min
[3,4]. The bactericidal action, on the mesophilic and thermO-
philic bacteria in Rhine water was even worse (Fig. 3). ;
The preliminary results.of our experiments on the use of
hydrogen peroxide'as an oxidizing, agent in Rhine water are
equally sketchy. Evaluation of the gas chromatogram with an
ECD shows only relatively slight changes in the fingerprint.
-------
- 552 -
H202 (mg/l)
50 110
300 500 1000 5000 10000
CONTACT TIME (min)
Fig. 3 H2O2 demand for 99% bacterial elimination (E.coli)
Only a few compounds, probably as a result of hydroxylation,
had their boiling points raised. .These phenomena should be
investigated more fully in the future, extended to IR-spectro-
metric analyses, and compared; .with the action of ozone.
4.2.
Production of OH radicals by photolysis
The production of OH radicals by photons in accordance with
the reaction:
hv
H2°2
2 OH-
2+
has the advantage over the Pe - -.generated OH radical synthesis
Pe
2+
OH- + OH + Fe
3+
that two OH radicals are formed per hydrogen peroxide
molecule. However the difficulty8lies in the fact that,
because of the large bond-dissociation energy of the O-O bond,
only the shortwave UV-C region with wavelengths of 200-280 nm
leads to usable radical yields. Therefore, at least in the
raw water sector, all the absorption and control losses
-------
- 553 -
-. f ''
caused by pollution of the .light-admitting surfaces will
4 ~ *" . ,
occur. However, in flocculated and filtered water the •
combined use of hydrogen peroxide with ultraviolet for the :
oxidation of organic .constituents deserves more thorough
study, all the more so since completely new antimony-doped ;
mercury lamps emit ten times"as much in the UV-C region as
conventional mercury lamps and thus guarantee a high radical
yield.
5. The formation of singlet oxygen from hydrogen peroxide
In add'ition to the OH radical singlet oxygen is a second very
reactive particle that can be formed from hydrogen peroxide
in a stoichiometric reaction with hypochlorite:
NaOCl
The singlet oxygen is richer in energy by 22 kcal/mole than
oxygen in its triplet ground state and its chemical behaviour
is basically like that of an electrophi"1 ic olefin. It reacts
a) as a highly reactive Diels-Alder component with
conjugated diene systems,
b) in a (2 + 2) cycloaddition with electron-rich
olefins,
and
c) in accordance with the general reaction type of the
En reactions with dlef ins" having an allyl hydrogen.
From these reaction types it can be seen that singlet
oxygen - similarly to ozone - reacts essentially more
selectively than the OH radical (Fig. 4).
-------
1,
2.
3,
HC*^~ C
* \ \
HC—C—S
R
ROV,H
r ir
- 554 .-
R R
^.6=0 wr^C=0
no " w
HCX * HCX
C-R C=0
M I
I R
0
hv
(2.2J-CVCIOADOIT10N
CH3
H
A,
9H3
EN-REACTION
CH3-6-OQH
CH
Fig. 4 Reaction types of singlet oxygen ( O2)
6.
Summary and outlook
Although hydrogen peroxide is highly unlikely to win
a place as a primary disinfection agent, it promises
in conjunction with low concentrations of other
2+
additives, such as NH2C1 .and Cu ions, a good
bacteriostatic action .on water en route to the
consumer.
The oxidizing action exerted on organic water
2+
components by Fe -catalysed hydrogen peroxide
cannot compete with that of ozone. Nevertheless,
there are hopes for its use, combined with UV, in
water free from turbidity. Results similar to those
in ozonization with subsequent activated carbon
filters are conceivable in-this sector.
-------
- 555 -
The replacement of potassium permanganate by hydrogen
peroxide for -demanganization has already been con-
sidered in connection with improving the operating
time of the downstream filters.
Hydrogen peroxide has up to now been used in drinking
water technology but not as a supplier of oxygen.
However, in the artificial enrichment of ground
water it could serve as a source of oxygen and help
to prevent the adverse consequences of oxygen Con-
sumption in the ground.
In England a case is reported where hydrogen peroxide
in backwashing water considerably prolonged the
filter running times.
(3) SABER, F., WEISS, J.
Proc. Roy. Soc. A. 147 (1934), 332-351
(2) BAXENDALE, J.H.
Advances in Catalysis' _4 (1952), 31
(3) -
An Investigation of Hydrogen Peroxide as a
Disinfectant of -Potable Waters
Water Research Centre, Medmenham, England
Den. (1974)
(4) OVERATH, H. .
Versuche mit H^O,-, und UV-Strahlen
ESWE-Bericht Nr. F 7/78
-------
- 556 •-
SOME ASPECTS OF THE USE OF CHLORINE OR CHLORINE DTOXIDE
IN WATER TREATMENT
J. Valenta
The drinking water in the Zurich water supply obtained from
Lake Zurich has been treated with chlorine for over 25 years.
Chlorine is still used as the active agent in the preliminary
chlorination, above all to protect the raw water plants against
algae and DPP (Dreissena polymorpha Pallas) larvae.
Since 1971 chlorine dioxide has been used successfully as a
mains-protecting agent. In the Zurich water supply system
seven such C102 plants are at present in operation.
Although the concentrations of trihalomethanes as a possible
result of chlorinating drinking water should give no cause
for concern, and although we still know relatively little
about the possible reaction products associated with the use
of chlorine dioxide, we are investigating the idea of replac-
ing chlorine in the preliminary chlorination by chlorine
dioxide. In connection with this, well-substantiated infor-
mation on the effectiveness of chlorine dioxide on "Wander-
muschel" larvae, and on the corresponding optimal dosage would
naturally be important.
At Hardhof ground waterworks the bank filtrate is also to be
treated with chlorine dioxide instead of chlorine before the
inflow into the enrichment reservoirs. To this end the
algicidal action of chlorine dioxide is currently being
studied by the limnological division of the Zurich water
supply.
Two further brief remarks may be made on the two oxidizing
agents:
-------
- 557 i- • >' •
The first concerns the 'so-called- "stabilized -chlorine d-ioxide
solutions." The subject was discussed 2 years ago at a
conference in Zurich, and not much has changed in these 2
years. Such products are still offered to the drinking
water works under various tradenames as a highly effective
concentrate of chlorine dioxide which can be stored for at
least one year without losing its effectiveness.
On the basis of several analyses in a number of Swiss and
German laboratories, it can be stated that this product is
a sodium chlorite solution, which must be regarded as im-
permissible for the treatment of drinking water. In order
to determine the C102 concentration, the suppliers prescribe
only the usual iodometric analysis. In this method, the
chlorite ion present is acidified and converted into chlorine
dioxide, and then determined as such.
Finally, a remark on the analysis of the two oxidizing agents:
In spite of the ever-increasing number.of instruments for
checking the residual amounts of.disinfectant by colorimetric,
amperometric, or UV-photometric measurements, in many places
simple manual methods are used for the same purpose with the
aid of various comparators or simple portable colorimeters.
For the determination of chlorine dioxide the colour disc or
the chlorine scale is used practically without -exception, and
the value found is converted to C1O2 with the aid of a factor.
The problem now lies in the fact that, depending on which
instructions are followed, this factor "= 2 is sometimes used
for multiplication and sometimes for division.
Our own measurements have confirmed the practically only 50%
colouring action of chlorine dioxide in comparison with
chlorine both with OTO and with DPD. We found the mean value
for this factor to be 1.8, and the -chlorine value must be
multiplied by it.
-------
- 558'-:
Consider three figures on this subject: '•'" ••'••-
Fig. 1 shows the simple absorption curves with OTO and DPD,
with a maximum at 436 nm for OTO and two typical maxima at
511 nm and around 553 nm (about 7% higher) for DPD. In
addition to this the figure shows the exactly defined maxima
of the holmiura calibration filter.
Fig. 2 shows the peak for chlorine, about twice as high as
that for chlorine dioxide at the same concentrations.
The last figure shows the colour development curves for three "
commercial reagents between the first and the 26th minute.
This slide also reveals the high instability of the colour
with one of the two DPD reagents.
Such facts, perhaps banal at first sight can, however, have
unpleasant and unexpected results when put into practice in
various situations.
1
too ~
9O-
7r
i
„„
Tl
431
t
jL
1
T)
»rms
« — Ho
._..
446nir
, ,
. Uft 4f
._
1 453.4nm i - - — , j
^—~1 553nm -f-
Stinm jr, = ;
r\ i
600
Fig. 1
Absorption curves of chlorine
dioxide with OTO ( ) and
DPD (-•-•-) and of the holmium
calibration filter
-------
- 559-- -
d»2cm -I \0.49mgC!/l
500 nm 550 600 " 400 nm 450
\
Fig. 2 Absorption curves of chlorine and chlorine dioxide
in equal concentrations with OTO and DPD
Colour development
CI02 -mit: DPD
DPD «B»
OTO
—
_.
—
••-
"•-
.— — — •
_.
J_ 1
j '
-.Lj-4-
_j g_
4 / ' _
it;: .
44 - --:
.....
._.
-
i"
...
....
._
—
..:..
I
"• -1
_.
...
I J i 4 4 J
"1 LU...L
1 1 1 i !
i 1 | ill
1 ; : • ! •
• 26% '
i
•
>—
,-
I 1 . I J
-1 J
i
i 1
_1
—
! j«T,8%
.._
._l-.'
1
-O'TC
_
....
— i
.3% • — ;- ^«0.6% _l_^:?A»_
' ' ; , ill L...
-10.4% ", i ~ • *';—• i~*—
i ! ! i ! ! I !
tf
~JL..
.^_:
'±:
j
3 5 10 15 2O 25
(VZunch. Au8«. 1978 Tittle, Klin
Fig. 3 Colour development curves of chlorine dioxide with
various reagents
-------
- 560 -
BIOLOGICAL METHODS FOR THE TREATMENT OF GROUND'WATER
Y. Richard
I. INTRODUCTION
In recent years renewed interest has focussed upon ground
water. For example, in the United States the production of
water of subterranean origin has doubled in 20 years.
These waters are sought after for their organoleptic qual-
ities and the theoretical absence of organic pollution.
Nevertheless, their use sometimes entails considerable diffi-
culties: bacterial proliferation in the distribution system,
precipitation of iron or manganese compounds, and sulphurous
or unpleasant odour - and these are only the most direct
manifestations.
The cause of these various disadvantages must be sought in
the very nature of ground waters, wh_ich are characterized by
the absence of oxygen and which constitute a reducing medium.
In waters of this kind one may find, depending on the compo-
sition of the surrounding soil, the various mineral species
in their reduced form: divalent iron, ammonia, manganous
ions, and hydrogen sulphide. All these elements are involved
in well known biological cycles.
Ground water contains no organic carbon and does not permit
the development of heterotrophic bacteria. In the reducing
medium of ground water, in contrast, one finds a very small
number of specific bacteria in a state of latent development.
These are generally autotrophic with iron, manganese, or
ammonia. It suffices for' the medium to be very slightly
aerated and set in motion, to create conditions favouring
bacterial development and leading to a profound change in the
quality of the water.
-------
- 561 - - •
Biological treatment is nothing more than the understanding
and control of this natural biological process, and the
development of plant required to accelerate it and to deliver
a water that will preserve .all its organoleptic qualities.
Apart from the diversity of the mineral species involved
(Fe, Mn, NHO , we must also define the principal character-
istics of biological treatment, while indicating the advan-
tages and the disadvantages of such treatments.
It should be pointed out that a different form of pollution
in ground water has been developing for some years. The form
in question is nitrate pollution. This ion can be eliminated
by biological means, using heterotrophic bacteria. However,
this technique is rather similar to a reduction of nitrate
nitrogen N03 to nitrogen N, and we shall not deal with this
technique.
II. THE BIOLOGICAL PROCESS
In general, all biological treatments make use of autotrophic,
or sometimes facultatively heterotrophic bacteria that use
CO2 and, for example, ammonia for their cellular development.
These gain the required energy from redox reactions.
II. 1. BIOLOGICAL DEFERRIZATION
This relies upon f errobacteria, which have been studied in
particular by Hasselbarth and Liidemann (1) . Many stations
in Germany make use of the principle of biological deferr-
ization.
In the list of bacteria drawn up by Starkay (2) (see Table 1)
we find the following species: Leptothrix, Sphaerotilus ,
Crenothrix, and Galionella. When these bacteria undergo an
autotrophic metabolism, they draw their energy from the exo-
thermic reaction:
4Fe(HC03)2 + 02 + H20 - > 4Fe(OH)3
-------
- 562 -
TABLE 1 Various types of ferrobacteria ,..,..-.--,•.>_
(after Starkay, 1945, in the classification
of Prevot, 1961}
Position in the classification
Species
Remarks
1. Sub-branch : Eubaeteria
1.1. Class : Asporulales
1.1.1. Order : Bacteriales
1.1.1,1. Paiily : Protobaeteriaoeae
Genus : fhiotaoillus
*l. Sub-branch ; Algolj&cteria
f.l. Class : SiderobEcterialea
2.1.1. Order s Chlamydobactiriales
?.1.1.1. Faaily : Chlanydobacteriaceae
Genus : Sphaerotilus
Leptothrix
2.1.1.2. Faaily : Crenothrieaeeae
Genus : Crenothrix
Clonothrix
£.1.1.3. Faaily : Siderooapsaceae
Geaur : Siderocapsa
Ferrobacillus
Sideremonas
2.1,2. Order : Caulobacteriales
2.1.2.1. Faaily : Galionallaceae
GBDUE : Galionella
Th. ferro-oxydans
S. natans
S. dichotoma
S. diooophorus
L. ochracea
L. cr&ssa
Or. polyspora
Cl. fsrruginea
01. fusca
S. tretiMi
S. major
F. ferro-oxydans
S. eottferwnm
G. ferrtiginea
G. aajor
filanentary, coated
heterotrophio (problematic
antotrophs)
(or Cladothrix diohotona)
(or Leptothrix discophora)
nedia poor in organic matter
polluted media : heterotrophic
facultative autotrophic
heterotrophic
(problematic autotrophs)
pedicular
strictly autotrophic
nedia poor in organic matter
-------
- 563 -
II.2. DEMANGANIZATION
Organisms that oxidize manganese have been studied by
Schweisfurth (3) and by Mulder and Van Veen (4) . "They are
listed in Table 2.
II.3. BIOLOGICAL NITRIFICATION
Nitrification reactions that oxidize ammonia to nitrate rely
upon autotrophic bacteria. The nitrification process is
divided into two stages, which bring into play two types of
bacterial strains. This balance is described empirically by
the following equations:
Nitrite formation
During this phase the ammonia is oxidized to nitrite:
55 NH4+ + 76 02 + 109 HCO3~ Nitrosomona»
C5H?NO2 + 54 NO2 + 57 HjO + 1O4
Nitrate formation
The phase in which nitrites are oxidized to nitrates:
4 N02~ + NH4+ + 4 H2C03 + HCO3~ """" *
C5H7N02 * 3 H2O + 4OO NO3 + 195
where C-H-NO- represents the chemical formula of the bacter-
ial cell.
-------
564 •-••t
TABLE 2 Microorganisms capable of oxidizing manganese
(after Mulder and Van Veen, 1963, and
Schweisfurth, 1972)
A. BACTERIA
1.
1.1.
1.1.1.
Sub-branch Eubacteria
Class : Asporulales
Order : Bacteriales
1.1.1.1. Family : Pseudomonadaceae
Genuses : Pseudomonas » :.
Metallogenium
Ps. manganoxydans
M. personatura
M. symbioticum
2.
2.1.
2.1.1. Order
2.1.1.1. Family
Genuses :
Sub-branch Algobacteria
Class : Siderobacteriales
Chlamydobacteriales
Chlamydobacteriaceae
Leptothrix :
Sphaerotilus
2.2. Class :
2.2.1. Order ;
2.2.1.1. Family
Genus i
L. echinata
L. lopholes
S. discophorus
(or Leptothrix discophora)
Thiabacteriales
Hyphomicrobiales
Hyphomicrobiaceae
Hyphomicrobium : H. vulgare
3. Sub-branch Mycobacteria
3.1. Class : Actinomyceta'les : '•
Certain unidentified Actinomycetes,
B. PROTOZOA
Flagellae branch
Family : Monadaceae
• Anthophysa vegetans
(ace.. , to Pringsheim, 1966)
C. ALGAE
' An alga of the Diatomea class (ace. to Peklo, 1909)
-------
- 565.-
TABLE._3 Microorganisms capable of oxidizing ammonia
Microorganisms responsible for nitritation:
- Nitrosomonas europea and monocella
- Nitrococcus -
- Nitrosospira , .
- Nitroscystro
N.B. Nitrosomonas is the most abundant, and its activity
the.greatest.
Microorganisms responsible for nitratation:
- Nitrobacter winograd_skyi
- Nitrobacter agile-
- Nitrocystro
- Bactodenna
- Microdenna
III. DEVELOPMENT CONDITIONS
Bacteria will only survive in a medium that conforms to
certain conditions of pH, salinity, redox potential. The
redox potential ranges vary appreciably according to the
bacteria. . ,.:••-,.
All the bacteria that affect the oxidation phenomena of
ground water (iron, manganese, ammonia) develop in aerobic
media and require the water to be-aerated. Pig. 1 shows that
the introduction of very-small quantities of oxygen brings
about a rapid increase in the redox potential. The intro-
duction of 0.2 mg.Jl"1 produces a change of more than 250 mV
in the water. However,'variations in E^ are not always
favourable to the development of all bacteria.
-------
- 566 '-..I
KV/H- ELECTRODE
300
280
260
240
/
/° Redox potential
DISSOLVES OXYGEN (MG/O
0 05 1 1.1 2 2,5
For each type of bacteria we shall state the optimum redox
potential conditions that favour their development.
III.l. DEFERRIZ&TIQN
Biological deferrization by ferfobactefia has been studied
in particular by Hasselbarth and Ludemann, 1971 and 1973,
who quote very high passage velocities, and retention capacities
for filters working on this principle; many plants of this
type are now in existence. _••;•:
The pilot plant upon which we conducted tests is described in
Pig. 2. The aeration system is designed to intrcduce only a
limited amount of oxygen into the water (this being one of
the main differences compared to the physico-chemical process).
The results were as follows:
- Biological deferrization developed under the following
conditions:
-------
- - 5,67 .<--.'.
Fig. 2 '• ..;.
Biological pilot plant
ANTHRACITE
1,5-2,4 HM
SAND
1,0m
dissolved oxygen concentration: 0.2 to 2 g-m
pH: identical to that of the untreated water
(6.3 - 6.4)
redox potential: 40 to- 200 mV on average
rH: . 14 to 20; if rH < 14,, some Fe' is left in
, the water.. 'For a pH of 6.3 - 6.4 and a potential
of 100. to 200 mV, biological def err ization takes
place at the boundary between the..Fe and _Fe (OH) ->
zones (see Fig. 3). Any excess oxygen will result
in an ,i,ncr,ease of "the redox potential. =
If, on the. other hand, xH > 20 the filters axe
likely-to become "rapidly clogged'since physico-
chemical1 deferrization1 (with precipitation of
amorphous floccular material) then competes with
the activity, of the ferrobacteria.
There was no formation of a silica-iron .complex. .
-------
- 568 -
Fig. 3
Potential pH diagram
for iron
1 -2 3 4 S S 7 8 9 10 11 1! 13-14pH
III. 2. NITRIFICATION
The curves of the redox potential as a function of the pH
(see Fig. 4) , show that nitrification takes place in poten-
tial regions higher than deferrization, between 300 and
500 mV. Nitrification is manifested in a fairly wide range
of pH between 5 and 10, but the optimum pH range is about 8.
The nitrification yields fall rapidly below pH 7.
Nitrification requires more oxygen "than deferrization. 4.53 mg
of oxygen are needed to transform 1 mg of annnoniacal nitrogen
into nitrogen in the form of nitrate.
-------
- 569 *-,_:•
Fig. 4
Potential pH diagram for
nitrification
III.3. DEMANGANIZATION
The results for manganese are less clear. Manganese can only
precipitate in an alkaline medium" (pH >_ 9). Nevertheless,
the redox diagram does not exclude chemical precipitation of
manganese between pH 8.0 - 7,5 and at a redox potential
between +400 and +500 mv". However, the reaction should be
very slow. ' .
The manganese bacteria can develop above pH 5.5 and the ground
water must be aerated so as to give a redox potential between
+100 and +200 mV if any biological oxidation of manganese is
to be achieved.
III. 4. SIMULTANEOUS ELIMINATION OF SEVERAL ELEMENTS
Manganese, iron, and ammonia,are often found together in
ground water at various concentrations! the concentration of
the manganese is often the lowest. Are there specific con-
ditions of redox potential and pH which permit elimination
-------
- 570 -
of all three elements, and if- so, •-under--'what conditions?
Several situations may be encountered.
a) Large amount of iron in the presence of ammonia and
manganese
This case occurs frequently in ground waters. The intro-
duction of very small amounts of oxygen, monitored -by
measuring the redox potential, enables biological elimination
of the iron by filtration. The quantity of oxygen is-always
sufficient to eliminate traces of ammonia of the order of
0.2 to 0.5 mg-2, NH^ either by nitrification or by bacterial
assimilation.
Hasselbarth has observed- bacterial elimination of manganese
at the deferrization filter. This would be made - possible by
a very small addition of oxygen, to give the correct redox
potential. Nevertheless, true manganese bacteria are not
found in the filter (5). • . , -
In many cases iron is eliminated chemically in a first filter,
and the biological elimination of manganese begins at the
bottom of this filter and continues in a second filter (see
Table 4).
b) Large amount of ammonia in the presence of iron and
manganese
If the quantity of ammonia pre.sent in the untreated water is
greater than 2.5 mg NHL per litre, the ammonia can be elimin-
ated by means of an immersed filter_ packed with pozzolana,
into which air is blown to satisfy the oxygen demand required
for nitrification. The most favourable air/water ratios lie
between 0.6 - 1.2. The ammonia is eliminated at the pozzolana
filter. The redox potential of ,the. "water increases consider-
ably with aeration, and deferrization is effected by chemical
means.
-------
- 571 -
TABLE 4 Deferrization in the presence of manganese
and ammonia
Raw water
Water F
Water F
Rate
m/h
10
10
Iron
mg/1
5-15
0.1-1
0.01
Manganese
mg/1
0.3 - 2
0.1 - 0.5
O.O1
mg/1
0.1 - 0.3
0.05
O.05
- Aeration by spraying
- Reaction tank: 1 h
Filter I: sand depth 2 m
Filter II: sand depth 1.5 m
Duration of filtration cycles: Fn
II
48 h
1 week
The precipitation of ferric hydroxide leads to a reduction in
the number of filtration cycles and entails frequent cleaning
of the sand filter placed behind the pozzolana filter.
IV.
KINETICS OF TH£ BIOLOGICAL PROCESSES
When the pH and redox potential conditions become favourable,
the biological process is observed to develop very rapidly
after an incubation phase that may be fairly long (from 3
days to 1 month).
-------
- 572
The growth of the bacteria -can be>• expressed by the'following"
relationship:
i = „-.
in which x = concentration of microorganisms (mg-.£ ),
y = specific reproduction rate of the bacteria
(day"1)
cr = mortality rate of the bacteria (day ) .
The specific reproduction rate of the bacteria depends on the
concentration of the substrate S to be eliminated, in accord-
ance with Monod's equation. In our case iron, manganese, and
ammonia are designated by the letter S.
where u = maximum specific reproduction rate (day" ) ,
S = concentration of the mineral substrate to be
~. i
eliminated (mg? £ ) ,
K = concentration of S when p = ~jp.
Moreover, the elimination of the substrate is proportional to
the concentration of bacteria present i\n the medium. We have:
ds - Y ££ (3)
~ dt ~ * dt
where Y is the cell yield.
By substituting eqs. 1 and 2 into 3, we obtain the substrate
elimination rate:
_
dt ~ Y (K + S) Y
S
_ £ x (4)
-------
- 573 -
If we put —— = q and ?- -=• K ,'an 4endogenesis term,
equation 4 becomes:
The importance of equation 5 is related to the kinetics of
mineral substrate elimination, which govern the dimensioning
of the water-treatment plant to be brought into operation.
In effect, the reaction order may vary according to the con-
centration of S to be eliminated. The*K„ values of all the
elements considered are small; for example, for ammonia K
-1 '
is less than 0.1 mg-i of ammoniacal nitrogen and the value
of K may be,neglected in comparison with S. We shall assume
that for the filtration rates used during the treatment of
surface waters the distribution of the biomass is homogeneous
in the filter. If S » K , -equation 5 becomes:
^ _
ctt (3,
The elimination of S is independent of the concentration of
S and the reaction order is zero with respect to S.
Integrating eq. 6 and putting t = yj, where U is the mean
filtration rate, we obtain
So- S = (q - Kd) X (7)
Thus the S-elimination profile is linear (zero order)
If S <_ K , eq. 5 becomes:
dS _ ,* S K . x
dt tq K d;
-------
The term K,K can be neglected, and integration gives:
- 3_ . H
S = S e K U
o s
-------
- ^575 - '
where K
H
U
8
- 10 °C)
depends on (q - Kj)
filtration material,
(11)
X and on the size of the
depth of the bed (in metres) ,
filtration rate (mrh ),
a value between 1 and 1.3.
This equation expresses the maximum quantity AS of substrate
that can be eliminated by a filter of depth H, packed with
a given filtration material and operating at a given rate.
The ratio H/U is the apparent residence time of the water.
Since the temperature of well waters is constant, and close
to 12° Cr it does not constitute a determinant factor.
In this case the elimination profiles of both ammonia and
iron are linear (see Figs. 5 and 6).
WATIR THROUSHPWT » 4173/1717
0, DISSOLVED H6/L
2 i
DEFTHCn)
Fig. 5 Ammonia elimination profile
-------
— 57'6 J-
FILTER CHARACTERISTICS!
- ANTHRACITE 1.6-2,3 MM
- BED HEI6HT! 1.25 H
- FH.TRATIOH RATE! 10 H/M
-rH 6.3
- DISSOLVED OXYGEN! 1,8 PPM
Fig. 6
Biological iron elimination
profile
FILT1H BEPTH {«)
0.5
IS
V.I.
INFLUENCE OP THE FILTRATION RATE
Pig. 7 shows the influence of the water throughput on the
quantity of ammonia eliminated by a filter 4 m in depth. The
quantity eliminated decreases with increasing throughput.
While one can eliminate 12 mg.£ of NH. at 4 m«h , the
quantity of ammonia that can be eliminated falls to 3 mg«-£
to 20 m-h .
-1
V.2.
DISTRIBUTION OF THE BIOMASS
Equation 11 also shows that if one wants to eliminate large
quantities of substrate,the residence time must be increased,
either by increasing the depth of the filter bed or by
lowering the water throughput.
-------
--577J-
15
10
ELIMINATED
\
BED DEPTH 4ft
0
WATCR THROUGHPUT H9/N2H
10
15
20
25
Fig. 7 Quantity of ammonia eliminated by an immersed filter,
as a function of the water throughput
Experience shows that it is not advisable to reduce the water
throughput too much, since the biomass will then be poorly
distributed. ATP measurements along the filter show that
its distribution along the filter follows the equation
-KH
X
X e
o
(see Fig. 8), which results in a non-linear
profile.
In this case, to eliminate the same quantity of the substrate
AS, one must achieve a longer residence time and this again
entails an increase in the bed depth, in contradiction to
the objective sought initially.
-------
- 578 -
Log(ATPou?1H*4)
101
5
D
0 25
. ATP Pg/ml/cm
aHH4 mg/l
87
140
8° C
I mVm2.h
Fig. 8
ATP and ammonia profiles
along a submerged bed
V.3.
INFLUENCE OF THE LOADING S.
Equation 11 also shows that a biological filter eliminates
a given quantity AS, and that if the loading S at the input
to the filter is increased, the filter will still not elim-
inate any more than the amount AS in equation 11. Fig. 9
shows that when the NH, loading at the input increases from
—I
2.5 to 6.5 mg» fc , the filter continues to eliminate.a
constant amount of NH. (3 mg-2. ) , so that the amount of NH4
_i
present in the effluent is increased (0 to 3.5 mg*& ).
One must therefore take into account the nature of the bore-
holes and the future course of pollution when calculating
the depth of the filters, A safety margin must be allowed
for in case the loading S increases.
-------
- 579 -
_._ NITR1CATOR INLET
4 10
AUGUST . OCTOBER
Fig. 9 Nitrification cycle
(pilot trial pnder pressure? bed packing depth 2 m)
The case when S is small:
Reaction 10 is applicable and shows that the elimination is
of first order. This case corresponds to the elimination of
trace amounts of NH.,
It should be noted that, all other things being equal, a
filter 4 m in depth will enable the ammonia content to change
from 15 to 0.5 mg«£ i to change from 0.5 to 0.05 mg«£ one
would then have to add another 1.5 m of depth (giving a total
depth of 5.5 m) - see Fig. 5.
V.4. INFLUENCE OF THE MATERIAL SIZE
The size of the materials used influences:
- the biological growth
- the filtration.
-------
- 580 ->u- -
a)
Biological growth'
trt'f -j T
The biomass forms a biological film which grows around the
material and develops an exchange area that will be the
greater, the greater is the specific surface area expressed
2 -3
in m ».m . The finer the material,•.• the higher the activity.
Pig. 10 shows the quantities of ammonia" eliminated by two
grades of pozzolana, one with a size between 3 and 5- mm and
the other between 5 and 10 mm-. „.•.„•
TEST TEMPERATURE: 11°C '
— POZZOUNA (^-£ MM)
POZZOLANA (5-10 MM) —*
m'
" T"— "
Fig. 10
Influence of material
particle size on the
quantity of ammonia
eliminated in an
immersed'filter '
b)
Filtration
Can one, however, use finer and finer materials to. improve
the efficiency of the biological treatment?
In the case of classical filtration proceeding from the top
downwards, and if we accept the Kozeny-Carman model, then
for a filter of depth H packed with a material of specific
-------
- 581 ^r .
surface area a the specific pressure drop may be written
AP _, „. 2 (1 — e)
. K v a 3— . .
£
where n = viscosity of the water,
V - filtration rate,
a = specific surfa.ce area of the material,
e = porosity,
H = depth of the filter bed.
We see that an increase in the specific surface area results
in a very rapid increase in the initial specific pressure
drop across the filter, and leads either to a reduction in
the duration of the cycles, or to the use of more powerful
pumps i
This formula also shows that the pressure drop further depends
on the porosity of the material, according to the factor:
(1 -e)2 , ,
During the filtration cycle, the elimination of the Fe, Mn,
or ammoniacal N is accompanied,by a.development of the bio-
mass around the filtration material and the porosity
decreases with time,' causing, the tdtal pressure drop across
the filter to become greater, - Thus, the evolution of the
pressure drop and the frequency of cleaning out the filter
will be governed by the -grain size of the material and the
bacterial reproduction rate..
Influence of the reproduction rate on the frequency of cleaning
out:
-------
- 582 -
' ' - • • > • -. , '» • . . - ., , ) I . . ] .. „ . , . .'- , ' , ' . . :'. '
The reproduction rate y can easily be related to the cell-
division time, also known as the doubling time (tj):
= 1 clX = d"Log S _ 'O,693
v X dt dt ' t
In the case of a filter operating correctly, where the con-
centration of Fe, Mn, or ammoniacal nitrogen at the filter
inlet is constant and greater than K , the biomass grows
o
exponentially according to:
X = Xo eyt
The exponential growth of the biomass within a filter rapidly
reduces the porosity and increases the pressure drop across
the filter.
The bacterial reproduction rate var-ies from one species to
another, and one finds different cleaning sequences in
operation, depending on whether deferrization or nitrification
is taking place.
- Nitrification: ......
Thus, according to the basic equations, the elimination of
20 mg. £ of ammoniacal N will produce 3 mg•& of Nitro-
_ i
somonas (10) and 0.5 mg.& of Nitrobacter. .The elimination
of 5 mg of ammoniacal nitrogen at 10 m»h will produce
_3
8 g-m of sludge.
- Deferrization:
Our tests of biological deferrization enabled us to eliminate
2 —1
an average-of 4000 g of iron per m over 30 h at 10 m-h ,
_2
or 133 g-m «h of iron. This iron accumulates within the
filter in the form of a ferric sludge and the filter must be
cleaned on average every 30 h.
-------
- 583 -
A rapid calculation shows that the nitrification filter will
in theory attain the same max:m, pernissible pressure drop
in 20 days. .
In fact, in the Paris region we have been using immersed
filters for nitrification for 15 years, packed with pozzolana
in the size range of 7 to 15 mm. The use of a coarser material
limits filter cleansing to once a month. In the case-of
nitrification under pressure (1.2 bars), the filters- needed
no washing for a' whole year.
V.5. INFLUENCE OF THE METHOD OF WASHING
For a biological filter to work well, the biomass must be in
its exponential growth phase. In the event of major cleaning
the biological activity of the filter is reduced.
The order of the reaction, which is 0 or i for optimum oper-
ation during most of the cycle, becomes 1 at the start of the
cycle.
Excessive washing will remove a very large fraction of the
biomass fixed to the filtration material, leaving behind a
total initial biomass that is too small, and that must be
regenerated at the start of the cycle. ;
In the case of biological deferrization, washing has a funda-
mental influence upon the filter-maturation time.
The best results have been obtained with:
Blowing the sand with air only: 5 to 10 sec.
This blowing allows .elimination-of old trichomes.
-------
- 584 - ••••
- Rinsing with water only, 10'0'm-fr - 3 min.
This rinsing causes a slight expansion of the material without
impoverishing the biomass within the filter.
VI.6. CONCLUSION
An understanding of the biological' factors that govern the
biological processes makes it possible to eliminate undesir-
able elements such as iron, manganese, and ammonia by bio-
logical filtration.
The plant operation, packing depths, and the choice of the
particle size of the filtration 'materials depend closely
upon the element being removed and on the growth Qf the
bacteria. These treatments can be relatively difficult to
use unless the pH and redox potential conditions required
of the bacteria are provided. The loading variations from
different boreholes can give rise to poor elimination of the
undesirable element, and it is necessary to include an
aeration zone and a filtration zone.
-------
- 585
(1) ESSSELBARTH, U. , ''LUDEMANN, -D. , . • f
Enteisenung und Entmanganung
Vom Wasser 38- (1971), 233-253
(2) STARKAY, R.L. , > • •. . .; , .
Transformation of Iron by Bacteria .
J. AWWA 37 (1945) , 963-984
(3) SCHWEISFURTH, R.
Manganoxydierende, Mikrporganismen in Trinkwasser-
versorgungsanlagen
gwf-Wasser/Abwasser '113 '(1972) , 12 , 562-572 :
(4) MULDER, van VEEN
Investigations on the Sphaerotilus Leptothrix'group'
J. Ant van Leenwenhook 29 (1963), 121-153
(5) H&SSELBARTH , U . ,' LfJDEMANN ,";D . ; '
Removal of. Iron and- ..Manganese from Ground Waters by
Micro-Organisms '
Water Treatment' arid'' Examination ;22_ (1973) , 1, 62-77
. (6) DEVTLLERS , ....... \ ....
Nitrification ;de' i''ea'u: 'e'limination de 1'ammoniaque
des eaux d-'alimentation -.•; • • ,: . - ' • . •
T.S.M. L'EAU^Oct. 1965..
{7} BRENER,t L.. , RICHARD^- ,Y..-, • .MARTIN ,• G.. ' . •
Elimination de 1'aramoniague des eaux de surface -
Conference prononcee au 58e "cbngres de 1'A.G.H.T.M.
Bordeaux (1978)
(8) MOUCHET, P., MAGNIN;, J.
Un cas complexe de deferrisation d'une eau souterraine
58e congres de 1'A.G.H.T.M., Bordeaux (1978)
(9) RICHARD, Y., BRENER, L., MARTIN, G., LEBLANC, C.
Study of the nitrification of surface water
I.A.W.P.R. 9th International Conference
(10) EAUG, R., MCCARTY, P.L.
J. WPCF 44 (1972), 11, 2O86
-------
- 586 -
BIOLOGICAL REMOVAL OF AMMONIA
J.B. Goodall
Introduction
Pilot scale studies to evaluate cost and efficiency of
ammonia removal from River Thames water by biological
sedimentation (fluidised bed), biological filtration,
and air stripping, were carried out between 1971.and
1975 at the Medmenham Laboratory of the Water Research
Centre (1).
The early work quickly revealed that biological sedi-
mentation was both technically and economically the best
of the three processes and so the work was extended
with the specific aim of being able to advise on the
design of full scale plant (2).
This paper describes the biological sedimentation pro-
cess and gives brief comparative details of the other
two processes.
All the work was financed"-by-the Directorate General,
Water Engineering, of the Department of the Environ-
ment.
The ammonia problem *
When ammonia is found in surface waters it is always
as a result of pollution, usually from the excretions
of wild and farm animals and sometimes from untreated
or imperfectly treated sewage or leakage from sewers,
cesspits, and septic tanks.
Ammonia is occasionally found in. groundwaters as a
result of biological denitrification or from the break-
down of protein by saprophytic bacteria and fungi.
-------
587 -
It is believed that ammonia in drinking water does not
present any direct health hazard and it is interesting
to note that from many samples of water from distribu-
tion mains analysed at Water Research Centre there has been
little or no evidence of oxidation of ammonia to nitrite
or nitrate. Work in the Netherlands has shown, however,
that levels of ammonia in excess of 0.3 mg/1 as N can
lead to aftergrowths in the distribution system and
associated taste problems. Ammonia is toxic to fish,and
ammonia removal may become an important aspect of fish
farming, particularly where water is recycled.
There is no doubt that ammonia interferes with chlorina-
tion although the hazards of poor disinfection have
been exaggerated in the past. The real disadvantage is
the cost of the large chlorine dose required to oxidise
ammonia and the difficulty of controlling the dose when
the ammonia concentration varies. Increased chlorine
dose can also presumably increase trihalomethane .forma-
tion.
World Health Organization (WHO) European standards re-
commended a level of not more than 0.05 mg/litre in
supply. .f
Non-biological ammonia removal processes
Air stripping " . •,
In this method the nitrogenous material is removed from
the water - not merely converted into a different chemi-
cal state. This advantage must be offset against the
possibility that ammonia gas discharged to atmosphere
may cause air pollution or even re-dissolve in open
water.
-------
- 588 -
As ammonia is extremely soluble a high air to water ra-
tio is required for efficient stripping, pH is also im-
portant, high values giving best results because only
gaseous ammonia can be removed by air stripping and it
is necessary to hold the pH between 1O and 11 to get
9O% of dissolved ammonia in the gaseous form.
Experimental work
The experimental work was carried out in two rectangu-
2
lar mild steel towers each of Q.58 m area and containing
2.1 m depth of 1O x 5O mm serrated wooden slats (Fig. 1).
Water entered the towers through distributors at the
top and air was supplied from centrifugal fans at the
base.
DETAIl OF PACKING CONSTRUCTION
Fig. 1
Air stripping tower
PUN VIEW OF
OtStWBWCft
-------
- 589--
The best .result consistently obtained was 9O% removal
at pH 11 with a water to air ratio in the range O.12
to O.16 by weight.
Operating problems
There are potential operating problems in the air strip-
ping of ammonia.
1. Ice formation from evaporative cooling in cold
weather.
2. Calcium salt precipitation from CO2 stripping. •
3. Temperature effect - efficiency reduces with fall-
ing-water temperature (Fig. 2) and reduced air tempe-
rature also affects the process adversely.
100-1 "-I ii;11 _
BIOLOGICAL
SEDIMENTATION
is unaffsctsd by
temperaturas of 5 ta 20 °C
C
n
f
H
•H
Fig. 2 ,
Effects of
water temperature
20
-------
590
The cost of air stripping will always be much higher
than that of biological processes (Fig. 3). The major
cost is running the air blower but the cost of pH ad-
justment may vary widely depending on the buffering
capacity of the raw water. Air and water temperature
has a big effect on cost - it has been claimed, for ex-
ample, that both capital and operating costs of air
stripping would be three times higher in Sweden than
in California (3) .
Cost
BIOLOGICAL
SEDIMENTATION
is the chtapsstby far-
even for ammonia
concentrations as low as
01 to 0-2mg«/l«tre
6-
4-
2-
A1R STRIPfMNG
Fig. 3
Comparative costs
AMMONIA CdNCElwnUTfm (hi,/I)
Breakpoint chlorination
Ammonia is converted to chloramines. Conversion to tri-
chloramine requires a chlorine dose about 8 1/2 times
the ammonia concentration in addition to the dose re-
quired for disinfection. The method has the disadvan-
tage of increasing the chloride content of the treated
-------
- 591---'
water and control can be difficult if the ammonia con- ' '
centrations vary widely. High chlorine dos'es could pre-
sumably increase trihalomethane formation.
Breakpoint chlorination is worth considering at very
low ammonia concentrations or following'biological
sedimentation when the dissolved oxygen content of
the raw water limits.biological removal. (See later). •
Breakpoint chlorination can be used under-these circum-
stances to remove any small excess passing through the
biological sedimentation unit.
Electrochemical removal (electro-oxidation)
%
It has been demonstrated that it is possible to convert
ammonia in sewage directly to nitrogen gas at the anode
of an electrolytic cellf but high efficiency is only
possible if very expensive platinum electrodes are used (4)
Fouling and corrosion would probably have a considerable
adverse effect on cost.
Biological ammonia removal processes
Biological filtration
Pilot plant experiments were carried out using two O.75 m
diameter spun concrete pipes containing 2 metres depth
of 20 to 4O mm graded washed shingle (Fig. 4).
The best ammonia removal achieved was 02.5% at a super-
ficial velocity of O.8 m/hr but increasing the treatment
rate gave lower removal efficiencies with only 4.7% re-
moval at 2.8 m/hr (Fig. 5). Temperature had a big effect
on removal efficiency which was approximately linear
between 6 and 12 C. At - a 'nominal feedwater concentration
of 3 mg/1 N, 75% removal at 12°C declined to 51% removal
at 6°C.
Analysis showed that ammonia.was essentially converted
co nitrate.
-------
r- 592 -
• 0.71m
irlc*
Fig. 4
Biological filter
construction
PLAN VIEW OF
DISTRIBUTOR
nts.
Nominal raw water ammonia concentr. = 3 rog/1 N
Eaw water temperature range ,11 to 19°C
ts
Flowrate
zo
2S
Fig. 5 Effect of flowrate on biological filter performance
-------
- 593 -
Biological sedimentation
The experimental work was done in two mild steel, square
section tanks 4 metres high and of O.58 m area. Raw
water entered through a pipe directed downward into-the
tapered bottom and flowed upwards to be removed by
launders at the top (Fig. 6). The initial ammonia concen-
tration in the raw water was 3 mg/1. ,
Samph'ng
taps
Fig. 6
Biological sedimentation
tank construction
Bottom
drain valve
n
In tne early experiments no seeding of the biological
sedimentation tanks was carried out and even after eight
months running at low upflow rates there was no floe
build up in the units. Ammonia removal was poor, gene-
rally between O and 2O%. Only when sufficient fine sand
-------
- 594 -
(5O-15O micron) had been* -added '-to-1' .form* a">detectabld "" (-r;'•'
fluidised bed did the ammonia removal increase .above 2O%,
Increased removal started 1O days after the addition
of sand and reached a peak of 85% after 28 days. There-
after it stabilised at around 6O%.
Sixty percent removal was maintained for 35 days' and
then the ammonia dose was reduced from 3 mg/1 to 2 mg/1.
The two filters immediately began to remove 100% of the
ammonia and this performance was maintained for the
duration of the test.
The biological sedimentation units were evidently sensi-
tive to raw water anmonia concentration but as the raw
water contained about 10 mg/1 of dissolved oxygen and
theoretically this is capable of converting 2.2 mg/1 of
ammonia nitrogen to nitrate, the units were clearly
working at full capacity - availability of oxygen being
the limiting factor. (4.57 mg/0? is needed to convert
1 mg of ammonia nitrogen to nitrate but it was found
that the loss of dissolved oxygen in the biological
sedimentation units was slightly more than the theoreti-
cal figure by about 1 mg/1. It did appear, however, that
oxidation of ammonia took precedence over other oxygen
demands and no substance used oxygen preferentially to
ammonia. A brief experiment during which pure oxygen was
injected into the feed gave improved ammonia removal.)
Effect"of temperature '•
A fall in raw water temperature from 21 C to 4 C did not
affect performance although the performance of the bio-
logical filters deteriorated significantly under the
same conditions and over the same period. The higher
solubility of oxygen at lower temperatures is presum-
ably a contributing factor in offsetting decreased bio-
logical activity at lower temperatures (Fig. 2).
-------
- 595 -
Concentration of fluidised solids < -t • v_ • • >• - • - ,
At one stage in the work the ammonia removal efficiency
dropped from 10O% to 6O%. This was shown to be due to
a reduction in the concentration of fluidised solids.
(Fluidised solids concentration was measured as the
percentage solids volume after four minutes settling -
very little further settlement occurs after four minu-
tes.) Ammonia removal was found to be practically
zero for fluidised solids concentrations less than 15.%.
Low water temperature, which increases viscosity, will
reduce fluidised solids concentration by expanding the
bed further than the same flow velocity at higher tempe-
rature . This effect can be eliminated by keeping flui-
dised solids concentration above 35% as above this figure
changes do not effect efficiency.
Conclusions .
Biological sedimentation was 'found to be the cheapest and
most effective method for removal: of ammonia from river
water in the concentration range O.1 to 2 mg/1 ammonia
Nitrogen. Unlike biological, filtration and air strip-
ping it was not affected by declining water temperature
in the range 21°C'to 4°C.
Effective operation on Thames river water required a
fluidised solids concentration (measured as percent vo-
lume settled after four minutes) greater than 35% and:
an upflow velocity not greater than 25 m/h.
In the biological sedimentation process ammonia is pre-
ferentially oxidised before all other oxygen demands but
removal efficiency is limited by the dissolved oxygen
content of the raw water and the process is unlikely
to be able to remove much more than 2 mg/1 ammonia un-
less oxygen is injected or the effluent from biological
sedimentation is 'polished' by' chlorination.
-------
- 596 -
(1) SHORT, C'.s.:.vw. -A-A .':{--. .-; , :.:••. . .-• .'..-.• • '. o/.iv ;> -:
Removal of ammonia from river water
Technical Publication TP 1O1 , Water Research Assoc. (1973)
(2) SHORT, C.S.
Removal of ammonia from river water
2. Technical Report, TR 3, Water Research Centre (1975)
(3) NERETNIEKS, I. et al.
Removal of ammonia from wastewater.by countercurrent
stripping with air - an economic study
Vatten 2£ (1973), 3, 269-8O
(4) IONICS, Inc.
The electro-oxidation of ammonia in sewage to nitrogen
Project No. 17O1O EED, Report to Water Quality Office
OSEPA, July (197O)
-------
- 597 -'. •
PURIFYING ACTION OF THE GROUND IN THE TREATMENT. OF DRINKING'
WATER '':.,''-'
H. KuBmaul
There are good reasons for u'sing ground water as the pre-
ferred source for drinking wa,ter supply; because of the
strong purifying action of the ground it is generally
hygienically unimpeachable and balanced, at least when it
comes from large-scale, porous ground-water systems with
intact covering strata. No mention will be made here of the
difficulties in recovering the ground water from fissured
rock formations.
The purifying action of the ground is due above all to the
following factors. When water penetrates underground it
undergoes a filtration process in which the solid particles
that can adsorb the water constituents are removed. Harm-
ful material can then be eliminated in the ground e.g. by
adsorption and ion exchange on the mineral particles and on
the humus substance by precipitation in the form of sparingly
soluble or insoluble compounds, possibly by co-precipitation
with other substances, and organic material in particular
can be degraded by bacteria.
The effectiveness of the underground purification processes
is presented in this report on the example of ground water
enrichment with river water, in the light of experience
based on years of investigations at the river filtrate water-
works on the Lower Rhine. The process for the supplement-
ation of the ground-water reserves by river filtration is
used extensively and successfully in West Germany and in
some neighbouring countries. In this process the'infil-
tration of river water into the ground is promoted by
-------
- 598 -
1 . V .. ,r . ,.,,,-.
the formation of descending funnels through feed' Wells "in" :'" ""' "'
the immediate vicinity of the bank (1). In the waterworks
studied the feed wells were only about 50 m from the river
bank (Fig. 1). Between the wells and the river are
situated other observation wells, which were likewise
included in the study [6/III - 6/1]. The residence times
of the river filtrate underground are on average three weeks,
but fluctuations between 3 and 46 days are possible, depending
on the water level of the river; The mean proportion of
river water in the raw water passed on for treatment is
about 80%, fluctuations of 30-100% being observed depending
on the water level of the river (2).
In the discussion of the behaviour of the water constituents
underground a distinction should be drawn between the
organic loading, since biological degradation predominates
in the former case and physico-chemical processes in the
latter.
Forderbrunnen^Feed weiis/Puits
. . . '. . Bra6/I.
oBr18/m
•Rhine-
20 40 60 m
Fig. 1 Position of waterworks I on the Rhine
-------
- 599
Organic substances .•",,, ..,---., ,.,.,,, .
The dissolved organically bound carbon is a measure
of the organic substance content of a water (DOC = dissolved
organic carbon). The mean values of two waterworks are
shown in Fig. 2, from which it can be seen that up to the
first observation well [6/III] in the immediate vicinity of
the bank of works 1 the DOC content has halved, and that no
further degradation takes place during the following under-
ground passage.
DOC
ImgA)
RhE HI R>G (WHO
1 11
20
10
A
COO
^
J?>
-
N»»
-
— ,
n
1
I
XI
^
&.
S1.
'2
^
•n
\l
Rhllin I feO RhtoG
I 11
50
w
30
20
10-
0
KMn,O4 -
kn^I)
>
x
^J
>
Vvrbr.
Consurnt-lr
m
o
s
f
^
Th
HI
mo
S.O.
°2
ImgA)
t.
I
0
-:
'£•
\f.
^"
^
^
T
?
1
n
K
(*mil I teG RhRoG
II
RhlD 0 1 RoG RhfoS
I It
Fig._ 2 Variation of the DOC, COD, KMnOj^ consumption, and free
oxygen during underground passage (bank filtration) in
two Rhine waterworks
(Rh = Rhine; III,II,I = bank filtration wells 6/III-I?
Ro = raw water of the feed gallery; G = ground water)
-------
- 6OO -'
Indications on the content of degradable organic substances
in a water can be obtained from the chemical oxygen demand
(COD) and the KMnO^ consumptiqn. It'is found that the mean
COD content of the waterworks in the 'immediate vicinity of
the bank has also decreased by 50%, but the substances in
question undergo an additional decrease during their under-
ground passage. The KMnO. consumption falls by 68% in the
immediate vicinity of the bank and undergoes no further
change(Fig. 2).
The microbial degradation of organic substances is associated
with a consumption of oxygen dissolved in the water. Accord-
ingly, the lowest amount of oxygen in the water is found
immediately after the loaded water has penetrated under-
ground, and the O2 content then rises again (Fig. 2). If
the dissolved oxygen is insufficient for biological oxi-
dation of the degradable organic substances, oxygen is
released from inorganic compounds, e.g. from nitrates, which
can lead to reducing anaerobic conditions in the ground
water, with all their adverse, effects on the nature of the
water.
The elimination rates for the underground passage and the
subsequent treatment of the drink'ing water with ozonization
and active carbon filtration, calculated with an allowance
for the mixing ratios of river water to ground water and the
respective contents, are shown in Fig.' 3. Further details
of the calculations can be obtained from (3).
In Fig. 3 it can be seen that in- comparison with the under-
ground passage the drinking Water•treatment only reduces the
DOC content further by 8%, the COD -content by 20%, and the
KMnO* consumption by 11%. •.••,..
-------
- 6O1 -
DOC
COD
KMml'V
n
y*
i
4
NOT
2
fci«--
j
3—
P0*
Na*
a"
K*
f-
P
CoO
MgO
G«sH.
KarhH.
l KXJ
i! Y//A
I J%%? '
| &^'//il
f 1 *S/tTL " '
i i vr[/44/./
ii i / rxx/xxxxx
I D^x x^X'/XXX/XXXXl
HKX/VXXXXX/xV^
1 1 '(//////////^/////fr'///
H L
I 1 "
t e
ii H
i E
II
" .C
» ! ^
f 1 $^//*&/y/^
i
H
i
n
I
D
I
1!
•100 «0 69 XO 29 C
Knderunf laai Unter-
grurtdpaBiiage
J ground passage
Kitd&rimg feel TriUsk-
waaseraufbereittmg
\f//4 drinking watstr
Variation lora dii
traitement <3e 1'eau
potable
2
a
[]
l3'^ '
-.M
*20 Ahdtfung'{*/»if
Review of the concentration changes of some water
constituents during bank filtration and drinking
water treatment . • . ' .
(KMnOi^-V = KMnOi^ consumption j Ges.H. = total hardness;
Karb.H. = carbonate hardness)
In addition to these sum parameters for the organic loading
only one other group of: substances- should be characterized
in closer detail, this be ing > the, organochlorine compounds.
Since organochlorines do not normally occur in nature, they
indicate an anthropogenic o-rigin and can therefore serve as
a pollution indicator. Only about 50% of the readily vol-
atile organochlorine compounds is. . eliminated during the
underground passage, and after the drinking water treatment
a -residual content of about 20% is still present. ; Somewhat
more favourable relationships are found for the difficultly
volatile organochlorine compounds; here about 70% of the
-------
- 602 -
initial content,.'is.••.removed during the underground, passage
and a further 20% or so during the treatment of drinking
water (Fig. 4) (4f5).
Organochlorvtrbindungtn
Orqanochlorine compounds
Composes organo-chlores
leichtfluchtig
schwerfluchtig
1
50
40
30
20-
10
n
easily volatile t
pgA tr3s volatlls
77,
^
' /
'''*.'
. /
'' ''•
N
'•/'.
' 'X
VX
___.
20-
10-
n
difficultly vola
pg/l ncu volatlls
r^n
', S
'''^
''\
/ f
'''/',
\
J
— , ,
"I ".
Rh Ro T
Rh Ro T
Fig. 4 Changes in the contents of easily volatile and
difficultly volatile extractable organochlorine
compounds during underground passage and drinking
water treatment
(Rh = Rhine; Ro = raw water; T = drinking water)
The behaviour of other groups of compounds in the ground
was also determined. This is described in greater detail
elsewhere [6,7]. However, it was found almost invariably
that the greatest purifying action took place in the
immediate vicinity of the bank and that the purifying effect
of the ground is very high in comparison with the subsequent
treatment of the water. ... -
-------
- 603 -
;.- ••'However, if the individual substances are examined it is
found that the retention capacity can fluctuate considerably
with respect to the various components; thus, many components,
even of similar composition, are retained very well, while
others can pass through the ground, and often even through
the drinking water treatment, suffering practically no
decrease (5-7).
Inorganic materials
The concentration changes of the usual inorganic water
parameters during bank filtration and drinking water treat-
ment are described in detail in (3). Here only a summary
is given (Fig. 3).
Although nitrate in the vicinity of the bank is/reduced to
a quarter of its original concentration, in connection with
biological oxidation of the organic substances, the calcu-
lated elimination rate for the underground passage is only
26% because of the admixture of nitrate-rich ground water.
After the drinking water treatment the nitrate returns
approximately to the original content in the river water.
This is not the case with ammonia and nitrite, which are
almost completely eliminated during the underground passage
and drinking water treatment. The orthophosphate content
of the infiltrated river water is reduced to a residual
content of,a few percent in the direct vicinity of the bank.
The calculated elimination rate is 95% and is only very
slightly altered during the drinking water treatm'ent.
In the case of sodium and chloride no changes take place
during the underground passage and drinking water treatment.
Potassium and borate show a slight.decrease of 15%, and in
the case of fluoride the elimination is about 33%. The
hardness, given as CaO, MgO, total, and carbonate hardness,
increases during the underground passage by 5-20% as the
water dissolves some mineral constituents.
-------
- 604 -,
The elimination rates calculated for some 'heavy metals
shown in Fig. 5. It can be seen that zinc, chromium, and
iron are retained almost quantitatively during the under-
ground passage. Elimination rates between 50 and 71% are
obtained for lead, cadmium, and mercury. Copper has an
elimination rate of 11%, while an increase of 17% is
established for nickel, and manganese is increased by a
factor of 3. This increase in the manganese content is due
to dissolution processes in the ground as a result of the
weakly reducing conditions. With the exception of mercury,
manganese, and nickel, no changes occur during the drinking
water treatment. Manganese is almost wholly precipitated
under the oxidizing conditions now present, while nickel
increases by a further 85%, presumably on account of disso-
lution processes in the conveyor and treatment plants.
Pb
Zn
Cd
Cr
Cu
Hg
Ni
Mn
Ft
I—ESS
I NXVSX
pr.^icTUru* tjoi l'p.teroriins!i-.issnoe
Ohar.o^ durl.nn urulnrarounci nssanni-
Variation lr*rs du P.ISR.VJO Jans IP sol
\m!i*rv:rv* bcl Trlc.kwflispraiifl'i'roltur.a
Variation lor* i!u traitp-i'nt i r 1'oflu
Ovl
-100
so
so
«100
(Thannr
Variation
Fig. 5 Changes in heavy metal and trace metal contents
during bank filtration and drinking water treatment
-------
- 6O5- - •
The numerical elimination rates given here are only valid
for the conditions on the Lower Rhine. They cannot be
generalized since, as can be seen in Table 1, the heavy
metal concentrations in the raw water (second column) are
very similar to the natural background (fifth column); they
therefore correspond to the concentrations of the individual
heavy metals in solution equilibria in the ground.
TABLE 1 Average 1972 and 1973 values for the dissolved shares
of some heavy metals in the river water and raw water
of waterworks I and II(yg/1) as well as background
values
Rhine
Pb 28
Zn 295
Cd 1.2
Cr 44
Cu 21
Hg 0.8
Ni 11
Mn 164
Fe 1132
Raw water
(= bank filtr.)
11
10
0.4
0.8
16
0.2
11
534
35
Change (%)
during under-
ground passage
- 52
- 96
- 60
- 98
- 11
- 71
+ 17
+ 301
- 96
"Background"
3
10
0.2
1
7
0.1
3
7
50
Summarizing, as regards the removal of heavy metals, it can
be stated that the elimination action of the ground under
aerobic conditions is so high that only those concentrations
are still present in the river filtrate that correspond to
the natural background values; the retention capacity
increases numerically with increasing loading of the surface
water with heavy metals.
-------
- 606
The elimination of heavy metals is, however, partly revers-
ible. If a change takes place in the redox potential in
the ground, or if larger amounts of chelate formers are
present, there is a danger that the heavy metals will re-
dissolve. This danger is increased by the fact that con-
siderable amounts of heavy metals are bound in the sediments
and the suspended matter of the river, and these can also
pass into solution under anaerobic conditions. This is
especially dangerous for river filtrate waterworks, since
the current water-treatment processes are not particularly
suited to the removal of heavy metals.
Summary
The results of the investigations indicate that the inorganic
and organic loading of river water is reduced immediately
after penetration into the ground by filtration, adsorption,
precipitation, or microbial degradation. However, some of
the material is not retained during the underground passage
or the drinking water treatment. The man-made foreign
substances are particularly important in the evaluation of
the safety of drinking water supplies.,
The penetration of highly charged river water underground
during ground water enrichment by bank filtration is
admittedly an extreme case. However, taking into account
the low loading of harmful materials in precipitation water,
the present results also allow conclusions to be drawn on
the possible endangering of ground water in natural ground-
water formation.
-------
- 607 -
.(1). BMI-FachausschuB "Wasserversorgung und Uferfiltrat"
Uferfiltration, Bonn 1975
(2) KUSSMAUL, H. , MUHLHAUSEN, D., BEHRENS, H.
Hydrol. und hydrochem. Untersuchungen zur Uferfiltration
Tell I: Chlorid, Borat und Uranin im Flufiwasser als
Leitsubstanzen zur Ermittlung von Verweilzeiten und
•Mischungsverhaltnissen bei der Trinkwassergewinnung
durch Uferfiltration
gwf-Wasser/Abwasser 118 (1977), 521-524
(3) KUSSMAUL, H., MtJHLHAUSEN, D.
Verweilzeiten, Mischungsverhaltnisse und Veranderungen
der Wasserbeschaffenheit bei der Uferfiltration am
Niederrhein
WaBoLu-Bericht 42 (1977)
(4) KUSSMAUL, H.
Behaviour of Persistent Organic Compounds in Bank-
filtrated Rhine Water
Aquatic Pollutants - Transformation and Biological
Effects, London (1978), 265-274
(5) KUSSMAUL, H., FRITSCHI, U., 'FRITSCHI, G., SCHINZ, V.
Leichtfliichtige Halogenkohlenwasserstoffe im Rheinwasser,
"Uferfiltrat und Trinkv/asser.
WaBoLu-Bericht 3^ (1978) , 75-85
(6) HEGAZI, M.
Analytik und Verhalten von Phenylharnstoff-Herbiziden
und deren Metaboliten bei Uferfiltration, Trinkwasser-
aufbereitung und Bodenpassage
Diss. Bonn 1977
(7) FRITSCHI, G., KUSSMAUL, H., SONNENBURG, J.
Cholinesterase-hemmende Stoffe im Qberflachenwasser,
Uferfiltrat und Trinkwasser
WaBoLu-Bericht 4O (1976)
-------
- 608 -
BEHAVIOUR OF MICROPOLLUTANTS IN RIVER WATER DURING
BANK FILTRATION
G.J. Piet and C.F. Morra
1. Abstract .
The analysis of individual organic chemicals in water
of the river Rhine and of this.water after bankfiltra-
tion gives an idea of the behaviour of organic com-
pounds during passage through the soil.
Chemicals which pass into the drinking water supply are
selected, their maximum concentrations are listed and
a notation is made when they belong to a suspected group
of chemicals. Means are indicated to prevent drinking
water pollution by industrial compounds. Recommendations
are given for future research on substances which impair
the quality of drinking water derived from bankfiltered
riverwater.
2. Introduction
Bankfiltered water can become important for the drinking
water supply in The Netherlands. In the year 1976 drinking
water was abstracted from groundwater (700 mlll.m ) and
from surface water (4OO mill.m ) with the expected increase
of drinking water consumption more and more surface
water has to be used because of the limited availability
of local groundwater (1). In The Netherlands a use of
250 million m bankfiltrated water for the drinking
water supply is foreseen which quantity could be extended
to 100O million m /year after the introduction of revised
structural plans. The advantages of bankfiltration such
as a reduction in the concentration of organic pollutants
-------
- 609 -
disinfection without the application o'f chemical oxi-
dation/ reduced influence of -calamities in the river on
the quality of the bankfiltered water when residence
times in the soil are sufficiently long, the fact that
the water is stored in a protected place where organisms
which can introduce unwanted metabolites cannot affect
the water quality, make it a good and non-expensive
water treatment system.
The information on bankfiltration is also of importance
to other infiltration techniques such as dune infiltra-
tion and groundwater recharge in combination with slow
sandfiltration. A close examination of the processes
which take place in the soil is of great value to select
additional techniques such as activated carbon treatment
necessary to produce wholesome and agreeable drinking
water when heavily polluted river water is used as raw
water source (2, 3).
The quality of drinking water derived from bankfiltered
Rhine water is affected by organic compounds, particular-
ly industrial chemicals which pass into the drinking
water supply.
Suspected chemicals or groups of chemicals have to be
considered in relation to the long-term effect on human
health and to the odour and taste of finished water. It
is important to have knowledge of the processes which
play a role in the reduction of organic chemicals during
passage in the soil. Chemicals which endanger the quali-
ty of drinking water from bankfiltered Rhine water have
to be selected.
3. Organic chemicals in water of the river Rhine
which affect the drinking water quality
In tapwater from bankfiltered sources more than 2OO
organic substances have been identified and subsequent-
ly quantitatively measured by analytical procedures based
-------
- 610 -
on a gas—stripping technique and a XAD adsorption
technique (4). The analysis of Rhine water is carried
out by an extraction technique with cyclohexane-diethyl-
ether and by means of a static head-space method (5, 6).
A capillary GC-MS system is used for identification.
Special attention is given to1 those chemicals which are
present in river water and pass during bankfiltration.
H thorough chemical analysis of tapwater from three
water production plants using bankfiltered Rhine water,
was made to select suspected chemicals.
The selection of compounds was made on the basis of the
following criteria
- Experimental evide.nee of toxicity for man or animals „
including carcinogenity, mutagenicity and teratogeni-
city (7)
- Identified in drinking water at relatively high con-
centrations
- Molecular structure closely related to other toxic
or odour-intensive compounds
- A known odour threshold concentration in water which
is 1% or.less of the actual concentration (8).
..In the investigated water plants only aeration and rapid
sand filtration is applied after bankfiltration. The time
of residence in the soil is at least 1OO days in all
cases.
4, Organic chemicals in drinking water from bankfiltered
Rhine water after passage of the' soil
The selected organic chemicals analysed in finished
water from bankfiltered Rhine water in The. Netherlands
are listed in Table I. Though the concentrations of these
substances differ from place to place due to different
conditions of the soil passage, usually the same type
of compounds appears at different plants.
-------
- 611 -
Table I • • " ""
Selected organic chemicals in tapwater from bankfiltered
Rhine water in The Netherlands
(maximum concentration in ng/litre)
Component
chloro-ethers
Max, cone.
ng/litre
bis (2-chloroethyl)ether 3O
bis (2-chloroisopropyl) 3OOO
ether
chloro-benzenes
chloro-benzene 3O
o-dichloro benzene • 10O )
m-dichloro benzene 1OO )
p-dichloro benzene 3OO )
e-trichloro benzenes. 3OO
chloro-methyl benzene 5
chloro-alkanes and
alkenes
chloroform 3OO
tetrachloromethane 1OO
1,2-dichlorethane 5OO
tetrachlorethene 5O
1,2-dichlorpropane 3OO
trichlorethene 5.OO
other chloro compounds
chloro aniline (m,p) 1OOO
5-chloro-o-toluidine 3OO
tri(2-chloroethyl) - 1OO
phosphate . . ••."'-..• •
suspected animal carcinogen
'suspected animal carcinogen
suspected animal carcinogen
(continued)
-------
- 612 -
component
aromatic chemicals
Max, cone.
ng/litre
benzene .
toluene
'ethylbenzene
o-xylene
m/p-xylene
Co-benzenes
c^-benzenes
others
naphtalene
divinylbenzene
2-methylnaphtalene
acenaphtene
biphenyl
anthracene
pyrene
alkanols, aldehydes
C[--alkanols
Cg-alkanols
geosmin
dimethylbenzaldehyde •
cinnamaldehyde
others
present
1OO
30
30
1OO
1 OOO
10O
1OO
•' 30
30
30
10O
• 30
. ' 30
•.300
. • 10Q
•10
• . - , 1 ooo •
30O
oxygen-containing components
1,1-dimethoxy propane 1OO
1,1-dimethoxy isobutane , , 30O
bis(3-methoxy-ethyl)ether 3OO
bis(2-methoxy-2-ethoxy)ether 1OO
methylisobutyrate 1OO
dimethyl acetophenon 10O
(continued)
-------
- 6V3- -
component . /Max. cone. ••'•". ' -.'."'"''"'
ng/litre
.triethyl phosphate 1000 ., ,
tributyl phosphate ,30
3 unknown compounds 300 - 1000
It is evident that several industrial chemicals are
not fully e'liminated even during long residence times
in the soil.
Some lower esters and ethers, which improve the odour
quality of water are present after groundpassage.
It is not fully understood whether bis(2-chloroethyl)—
ether is formed during groundpassage or not.
Some aromates have rather low odour threshold concen-
trations (such as naphthalene, mesitylene etc.). Toluene,
xylenes, and ethylbenzene are substances which may
affect water quality too.
Polynuclear aromatic hydrocarbons are mainly presented
by fluoranthene and,always in concentrations below
5O ng/litre.
The concentration of alkanes, with exception of C,. -
alkanes was always below the 1OO ng/litre level.
An important reduction of several chemicals can be
reached if the residence times are sufficiently long.
For this reason a consideration of the analysis of
the river water of the Rhine is of interest.
5, Organic chemicals in the river Rhine
The investigation of the water of the river Rhine is
performed by the same instrumental analysis as the bank-
filtered water,and it'became evident that the following
major compounds in the Rhine .water are almost com-
*.
pletely removed during groundpassage (Table II).
-------
- 61.4 -
Table II
Organic chemicals at concentrations > 1 jjg/litre 'in
Rhine water which were not detected in related tap-
water (9)
Component
nitrobenzene
o-nitrotoluene
m-nitrotoluene
p-nitrotoluene
di-nitrotoluene
p-nitro aniline
N-ethyl aniline
N.N.-diethyl aniline
amino-nitro toluene
diphenyl amine
m/p-chloro nitro benzene
o-chloro nitro benzene
m/p-chloro toluene
o-chloro toluene
methyl-tert.butylphenol
diethylene glycol diethylether
2,6-di(t)butyl-1,4-benzoguinone
2(methyl-thio)benzo thiazol
Concentration
(ug/litre)
1 - 10
10
1
3
1
1
T
1
3
I
1
1
1
1
1
1
2
1
It is of interest that some of these chemicals seem
to be rather persistent to microbial decomposition in
river water, they are eliminated, however,during passage
of the soil.
The chlorinated butadienes, chlorinated phenols and
cresols are not found as major contaminants in The
Netherlands. The same is true,for o-phenyl-phenol, p-
nitro-phenol, phenol, m-cresol and 2,4-dimethyl phenol (1O)
-------
- 615 -
Though a great diversity of industrial pollutants is
present in Rhine water, only a limited number passes
into the drinking water in The Netherlands at concen-
trations > 10 ng/litre.
Several processes involving the reduction of organic
chemicals during bankfiltration are similar to those
during slow sand.filtration where mainly microbial
decomposition modifies organic compounds. In 2 water
treatment plants using slow sand filtration techniques
after dune infiltration of river water, the following
compounds were present in concentrations> 10 ng/litre
(see Table III).
TABLE III Organic chemicals in drinking water after dune
infiltration and slow sand filtration of polluted
river water (concentrations > 10 ng/litre)
alkanes
alkyl benzenes
naphthalenes
alkanols
aldehydes
ketones
ethers
esters
phtalates
organo-halogens
nitrogen-containing
compounds
sulfur-containing
compounds
CQ-CI. alkanes
toluene, C^-benzenes, C.-benzenes,
ethyl styrenes
naphthalene
C5-C_ alkanols
dimethyl benzaldehyde, cinnamal-
dehyde
acetophenone
1,1-dimethoxy isobutane
ethyl acetate, butyl acetate
dimethyl-, diethyl-, dibutylphthalate
tetrachloromethane, trichloroethene,
tetrachloroethene,hexachlorobuta-
diene, chloro benzene, o-dichloro-
benzens, p-dichloro-benzene, bis-
(2-chloroisopropyl) ether
: N-ethyl aniline
. : benzo thiazole
-------
- 616 -
A correlation' 'between' 's'ome 'dhemicaIs which are; not'''- •'•"•'•'-'•"•'
completely removed by slow-sand 'filtration and bank
filtration is evident.
6. Processes which play a role 'during bankfiltration
Filtration of particulate matter with adsorbed organic
chemicals such as PAH's, PCB's, pesticides and in-
secticides, high-boiling hydrocarbons leads to the reduc-
tion of the concentration of several organic contami-
nants during passage of the soil. Adsorption and ion
exchange will also take place. .
In the case of precipitation of e.g.carbonates and sul-
fides co-precipitation of organics•can occur. This is,
however, a slow process where the equilibrium between
dissolved and precipitated substances is time-dependent (11),
Chemical reaction such as hydrolysis can occur also.
Microbial decomposition,under aerobic and an-aerobic
conditions, the presence or nitrates for the additional sup-
ply of oxygen affect for a g'reat deal the decomposition
effect of organic chemicals in the soil. The degradability
of compounds which decreases with the increasing number of
halogen atoms and depending on the type of halogen atoms
oftentimes determins the fate of organics in the soil (11).
Several organo-chlorine compounds which are not readily
adsorbed seem to survive. For this reason after-treatment of
bankfiltered water should be- performed with strong adsorption
procedures such as activated carbon treatment. But even when
this procedure is applied after ozonation of bankf iltered water
substances such as chloroaniline,- .chloroform, dichloro-
benzene, tetrachloromethane, 1, 2-dichJLoroethane,dibutyl-
phthalate,C., benzenes, C-o alkanes, triethylphosphate
and some unidentified substances were.not completely
removed during the processes- in a drinking water plant.
Water sanitation programs have to be installed to im-
prove drinking water quality while structural plans to
-------
- €17 - ,; . .
increase residence time in the gr,o,und, .are also necessary.
Recharge of groundwater .leading to a certain dilution
of pollutants is of assistance to provide a good quality
drinking water from polluted surface water.
Pretreatment processes/ such as .f locculation which re-
moves organic material but more so ozonation which im-
proves the biodegradability. of organic substances and
makes these better adsorbable, are recommended to improve
the quality of bankfiltered water.
In the case of possible mobilization of chemicals or a
certain degree of "saturation" of the ground after pro-
longed use of bankfiltration, model studies as will be
performed at the National Institite for Water Supply (11)
are of assistance. The fate of chemicals during passage
of the soil can be better predicted when more insight in-
to the processes taking place in the soil is obtained.
7. Conclusions and recommendations
- Bankfiltration is a safe, inexpensive and reliable
technique of water treatment
- Suspected chemicals which "break through" have to be
monitored and proposed for surface water sanitation
programs
- Structural plans have to be developed to attain optimal
residence in the ground if possible
- Some fundamental studies.of.mobilization of chemicals
have to be set up
- Additional treatment systems, "necessary to produce a
wholesome and agreeable type- of drinking water are
necessary when bankfiltered Rhine water is used for
the drinking water supply
- The introduction of pre-treatment of bankfiltered water
by flocculation and/or ozonation can be of great impor-
tance.
-------
- 618 -
(1) S50ETEMAN, B.C.'J. ...'..'. ,. , • , ,, ,
Kwalitatieve beschikbaarheid van grond- en
oppervlaktewater " * ,
H2O _8 (1975) , 4, 75-81
(2) SONTHEIMER H., MISSING, W.
Xnderung der Wasserbeschaffenheit bei der Bodenpassage
unter besonderer Berticksichtigung der Uferfiltration
am Niederrhein - •
Gas-Wasser-Abwasser 57 (1977) , 9, 639-645
(3) ROBERTS, P.V. et al.
Ground-water recharge by injections of reclaimed water
in Palo Alto '••••.
Civil Engineering Technical Report No. 225, Dept. of
Civil Eng., Stanford University, Stanford, Ca. (1978)
(4) ZOETEMAN, B.C.J.
Sensory assessment and chemical composition of
drinking water
Thesis, National Inst. for Water Supply, Leidschendam,
The Netherlands
(5) PIET, G.J. et al.
A fast quantitative analysis of a wide variety of
halogenated compounds in surface- drinking- and
groundwater
Internat. Symp. on the Analysis of Hydrocarbons and
Halogenated Hydrocarbons in the- Aquatic Environment,
Hamilton, Ontario, Canada, May 23-25 (1978)
(6) PIET, G.J. et al.
Determination of very volatile organic compounds in
water by means of direct head • space' • analysis
Analytical Letters, A11. (1978),~5, 437-448
(7) Drinking Water and Health
At the request of and funded by the U.S. Environmental
Protection Agency, National Academy of Sciences,
Washington, D.C. (1977)
(8) Van GEMERT, L.J., NETTENBREIJER, A.H.
Compilation of odour threshold values in air and water
National Inst. for Water-Supply; Central Inst. for
Nutrition and Food Research, TNO, Zeist, The Netherlands
(1977)
-------
- 619 -
(9) MORRA, C.F.
Internal communic., National Inst. for Water Supply,
The Netherlands (1978)
(10) SMIT, Z.
Determination of some phenolic compounds by concen-
tration on XAD resins and T.L.C.
Internal communication, National Inst. for Water
Supply, The Netherlands (1978)
(11) HARMSEN, K.
Column and adsorption experiments to be performed for
the examination of the quality of groundwater in
The Netherlands (in Dutch)
Internal communication, National Institute for
Water Supply, The Netherlands (1978)
-------
- 620
EXPERIENCE WITH THE REMOVAL OF MICROIMPURITIES IN SLOW SAND
FILTERS
K. Schmidt
The behaviour of an impurity in water and the degree of its
elimination in a biologically active filter are determined
primarily by the following factors:
1. Chemical and physical properties of the impurity
material.
2. Concentration, loading duration, and origin of the
impurity.
3. Biological assimilability or persistence of the
material.
4. Concentration and nature of the animate and inanimate
organic charge of the filter material.
5. Possible reactions of the unwanted substance with
other substances present in the system.
6. Functions of the active agent of the unwanted substance.
A number of other material properties and mechanisms relevant to
elimination could be added to this list, but these are
mostly of secondary importance.
As regards its construction and the reactions taking place in
it, a slow sand filter has a very complicated structure, and
the individual factors are connected to one another by
numerous interrelationships and regulation circuits. It is
therefore difficult to trace the exact course of the elimin-
ation of an unwanted substance in detail and to interpret it
satisfactorily from the scientific point of view. The
empirically established performance of a system is difficult
-------
- 621 -
to fit into known patterns and its applicability to other
times or to other places is limited to at least an order of
magnitude.
The principal factors in 'the elimination cf impurities from
water, given in the preceding list, will now be considered
in closer detail.
1. Chemical and physical properties of the impurity
material
The chemical and physical material properties that have a
bearing on elimination in biological filters are as follows:
1. Hydrolysability.
2. Solubility in water.
3. ' Polarity.
4. Volatility.
5. Molecular structure.
6. Molecular size.
7. Electrical charge.
It is generally known that' polarity, solubility in water and
molecular size are factors decisive for the adsorbability of
a given compound. The molecular structure determines the
compound's degradability, and the electrical charge and the
charge distribution determine the behaviour with respect to
other dissolved and undissolved constituents. Different
materials vary strongly in these properties, giving rise to
correspondingly different patterns of behaviour.
2. Concentration, loading duration, and origin of the
impurity
The significance of the loading duration and concentration
for "the elimination of an unwanted substance in biological
filters is presented in .Table 1. Normally the unwanted
-------
- 622 -
materials are present in surface waters as a permanent more
or less uniform loading and usually in low concentrations.
If the substances are degradable, they are removed effi-
ciently by the biological action of a slow sand filter.
However, many persistent substances break through, usually
after a shorter or longer delay. In the long run only that
fraction is removed which is removed from the system with
the topmost sand layer when the filter is cleaned.
TABLE 1 Significance of the duration and concentration
in loading with undesirable substances for
their elimination in biological filters
Type of loading
Continuous loading
Relatively low
concentration
(e.g. waste
water pipes)
Sudden loading
Relatively high
concentration
(e.g. transport
accident)
1
Material properties
readily degradable
difficultly degradable
persistent , adsorbable
persistent, non-
adsorbable
readily degradable
difficultly degradable
persistent, adsorbable
persistent, non-
adsorbable
Behaviour in
the filter
good elimination
good elimination
restricted elim-
ination
no elimination
anaerobiosis
inital break-
through
good elimination
no elimination
-------
- 623 -
In the case of sudden loadings, e.g. after accidents, the
unwanted materials are present for only a short time but
usually in relatively high concentrations. Readily
degradable substances cause an intensified oxygen consumption,
which can give rise to anaerobic conditions in the system.
Sudden and heavy loadings with impurities that can only be
degraded after an adaptation pass through the filter in so far
as they are not very readily adsorbable.
Short-term increases in the concentration of persistent
adsorbable substances are well intercepted and often
pass through the filter with a long time delay, the concen-
trations being frequently reduced by two to three powers of
ten.
Table 2 shows the behaviour of biogenic impurities in bio-
logical filters. Many organic substances are formed by the
microorganisms in the water and in the slow sand filters.
Even in main canals loaded with sewage more than half of
the organic carbon can be of biogenic origin. It is known
that many of the substances responsible for undesirable
taste and smell of the water are released by algae and
actinomycetes. Some excreta are highly stable as metabolic
end products in the given system. They can also pass through
biologically active filters and cause serious problems in
the production of drinking water.
3. Biological assimilability or persistence of the
material
Various types of biological degradation in filters are shown
in Table 3. Most organic materials are bio-degradable. If
they can be assimilated by many bacterial species, their
elimination will be spontaneous and far-reaching. Sometimes
a necessary adjustment of the intracellular metabolism is
responsible for a certain delay in the degradation (enzymatic
adaptation). A longer "running-in" time is needed
-------
- 624 - >
if the substance concerned can only be utilized ,by, a sjnall
number of specialized organisms whose proliferation is
initiated by the impurity substance itself.
TABLE 2 Behaviour of biogenic impurities in biological filters
Origin Material properties
,
Normal plankton
content
.
Destruction of
a mass
population
readily degradable
difficultly degradable
persistent, adsorbable
metabolic end-products
persistent, non-
adsorbable metabolic
end-products
readily degradable
difficultly degradable
persistent, adsorbable
metabolic end-products
persistent, non-
adsorbable metabolic
end-products
Behaviour in the
filter
good elimination
good elimination
restricted
elimination
no elimination
anaerobiosis
initial break-
through
good elimination
no elimination
-------
- 625 -
TABLE '3' Varfous "types of biological
degradation in filters
Spontaneous degradation without adaptation
by several species
Slightly delayed degradation after enzymatic
adaptation
Delayed degradation, only by "specialist"
bacteria
Slow degradation as a secondary metabolic
reaction
Final degradation to inorganic end-products
Incomplete degradation only as far as org-
anic "refractory" end-products
No degradation of persistent organic materials
Many harmful substances can only be transformed in secondary
metabolic processes, the microorganisms involved then
requiring other organic materials for their basic nutrition.
Their proliferation therefore cannot be controlled by the
harmful substance. . '
Not all degradable substances can be mineralized in bio-
logical filters to inorganic end-products. Under certain
conditions biologically stable metabolic products are formed,
which can only again become assimilable after far-reaching
changes in the medium (e.g. a shift of the redox potential).
4. Concentration and nature of the animate and inanimate
organic charge of the filter material
Both the biological degradation and the adsorption processes
take place predominantly in the top few centimetres of the
filter bed, in which layer the suspended detritus is fixed
and the autochthonous filter fauna and flora develop most
intensively. Both are absent in new filters and must be
-------
- 626 -
established in the course of a fairly long running-in time.
In filters exposed to the light algae pl'ay an essential part
in the elimination of undesirable materials.
Pig. 1 shows schematically the phases involved in the
substance conversion on the example of the degradation of a
readily soluble compound: the granular skeleton of the
filter material, the accumulated clay minerals, the bio-
logical population of the filter material consisting of
algae, bacteria, and fungi, the embedded suspension material,
some of which consists of particulate organic carbon, the
water phase, and the harmful material degraded by bacteria
with the aid of nutrients, a harmless degradation product
being formed in this case. The arrows indicate the inten-
sity of the material transport. It should be emphasized
that biological filters are self-regulating in response to
Tine
|j~] Mater IfC^vy-lAlgae, bacteria [p
^TJ^ Suspended natter Clay Minerals A Unwanted aubatence'
Nutrient Jj^ Degradation product
Fig. 1
Degradation of a readily
soluble substance
Filter material
-------
- 627 -
increases in the concentration of degradable materials in
the raw water: the microorganisms which perform the
degradation proliferate and thus provide additional
adsorption sites and increased degradation capacity.
•'Fig. 2 shows the material transport for the case of a sudden
loading with a non-degradable, readily adsorbable substance.
This is fixed mainly on the organic material in the topmost
layer of the filter bed. A small proportion not fixed here
can also be fixed deeper in.
Fig. 2
Sorption after short-
lasting loading of
unwanted material:
persistent, well-
adsorbable
Water
Suspended
matter
Algae,
,T "bacteria
unwanted
material
KiO<
Filter
material
Clay
minerals
If the loading with persistent harmful materials continues
for a longer time, the conditions in the upper layers
approach a state of equilibrium with a corresponding concen-
tration of the material (Fig. 3) . The region of particu-
larly intensive fixation of harmful material is gradually
displaced in the direction of greater depth. When layers
-------
- 628
with little or no biological growth have been reached, for
many materials the adsorption capacity of the solid phase
decreases and the passage through the filter bed is con-
siderably accelerated.
Fiq. 3
Long-lasting loading
with a persistent,
readily adsorbable
substance
Water
Suspended
matter
-i Algae,
^ bacteria
Unwanted
material
Filter
material
Clay
minerals
If the contamination of the raw water decreases the material
equilibrium between the water and the solid phase is per-
turbed and the material is desorbed until a new equilibrium
has become established (Pig. 4) . Desorption processes in
the case of fluctuating water quality sometimes result in
higher concentrations of unwanted materials at the filter
outlet than in the raw water itself.
If the biological growth suddenly dies off and decomposes,
the adsorbed or accumulated harmful substances are once more
released (Fig. 5). This can happen when, e.g. iron bacteria
die as a result of an oxygen deficiency. The best example,
however, is the release of harmful materials from decomposing
algae, with high enrichment factors for many substances.
-------
- 629 r
[• Water
I Suspended
I matter
rrrn Algae, ::
'•'':'' fcapteria
•-Unwanted •
material
»• Filter
-! material
S" Clay
1 minerals
Fig. 4
Desorption after
temporary loading
with a persistent,
well ad'sorbable
material
Fig. 5
Release of unwanted
materials on destruction
of the biological growth
due to a change in the
medium
rrjTi Suspended
H^J matter
bacteria i'""'1 material
Unwanted ~ ^^ Clay
material minerals
-------
- 630 -
The topmost zone of the filter,which is particularly bio-
logically active, plays a major part in all these reactions.
5. Possible reactions of the unwanted substance, with
other substances present in the system
By reacting with other water constituents an unwanted
substance can be removed from adsorption equilibria primarily
set up and be fixed more or less finally. Examples of this
are the sulphide precipitates of some heavy metals under
anaerobic conditions and the precipitation of manganese as
manganese dioxide or of iron as the hydroxide.
Another manner in which a substance can be removed from an
equilibrium between the liquid and the solid phase is by
complexing with natural or anthropogenic complex formers.
The active accumulation of harmful substances within the
cells of microorganisms can also be viewed in this way.
6. Functions of the active agent
Many constituents of water exert an inhibitory or promoting
action on reactions of other substances. Phosphate promotes
the proliferation of bacteria without increasing the loading
of the organic matter (1). The degradation of higher hydro-
carbon loadings can be accelerated by the addition of bound
nitrogen (2). The addition of growth substances, e.g.
vitamins, can also enhance the degradation. Many substances
stimulate bacterial activity at low concentrations and
inhibit it when their loading is higher.
As a biological system, a slow sand filter is highly sus-
ceptible to certain influencing factors, yet on the other
hand it can be very adaptable. The physical filtration
processes taking place within it provide an additional
-------
- 631 -
security. Thus, in the case of a sudden loading with'"a
bacterial substance, its concentration can be reduced to
such an extent in the upper filtration layer that the lower-
lying regions of the filter can still fulfil their degrad-
action function. This is the case, for example, in the
addition of chlorine to combat algae.
Ruhr water
inflow
prel. mter
1,50 m
0,50m
2.00 tn
0,8Drn
t-?.tti°AS«>>.o.sM
%*°'V!*&S' c'
;«*«ravel ,?°j,|B^ c2
Fig. 6 Preliminary filter - main filter system
(experimental ground-water plant)
It is not feasible to illustrate systematically all the
above-mentioned processes by quoting relevant experimental
results. The publications of the Dortmund Water Research
Institute over the last 15 years should be consulted for
this purpose. Some examples,- most of which were obtained in
the experimental plant shown in Pig. 6, can, however, be
given to clarify the above statements.
-------
- 632 -'
In Fig. 7 it can be seen how after a trial with additions
of cadmium strongly increased cadmium concentrations were
produced in the wash-out phase by sudden destruction of the
algal population right up to deep regions of the f,ilter bed
(3). This demonstrates the high rate of fixation in algae
during the introduction of cadmium.and remobilization of the
t
fixed cadmium when the algae die off, as well as the release
of complex formers from the algal material, which prevent
renewed fixation of the cadmium in subsequent passage through
the filter.
ppbcd in the water
Amount added
Death of algae
in the filter
•X
,_,,_- °~T, Ubers'qu
f *—0.1mTiefe
x—0,25m Ti«fe •~3,OmTiefe
•— I.OmTiefe
f— Fig. .7
Remobilization of
cadmium on death of
an algal population
Oberstau = upper water; Tiefe = depth; Tage = days
-------
- 633 -
The preferred fixation of microimpurities in the organic
material of the upper filter layers is demonstrated in
Figs. 8 and 9.
2O 3O 41) SOppn
Fig. 8 Heavy metal enrichment in the filter sand in
long-term addition experiments
0 cm
PCB
10 20 30 40 50 SO 70 80 90 100—•/.
Sandfilter
2400 ppb £ 100 %
Discharge
Fig* 9 Percentage decrease of CLOPHSN A 30 against filter
depth at the end of the trial? experimental ground-
water plant
-------
- 634 -
Fig. 8 shows the enrichment of lead, cQpper, zinc, and
chromium in the upper sand layers after heavy metals have
been introduced over a period of several months (4).
Fig. .9 illustrates the PCS content of the filter sand in a
dosage trial after a three-year wash-out phase (5). Even
after this long time the greater part of the fixed PCB is
still present in the top few centimetres of the filter bed.
It is also clear that when the slow sand filters are cleaned
a considerable fraction of the harmful substances introduced
is removed from the system.
The illustrated adsorption processes can also delay the break-
through and even out the sudden loadings occurring in the raw
water to a considerable extent, even over relatively short
sections of the filter bed. Fig. 1O represents, on the
example of hexachlorobenzene,a break-through delayed by
4 weeks with a reduction of the concentration by two powers
of ten (6).
Hexachlorobenzene ng/l
10000
1000
100
10
Fr-A
? / -,-
/
0*0
i
A\
A
/\
*.
/ •«
/ i
'
\
Cl% C.
Cl \ ff~C{
cf ~ci
*
1 "^V i
1
l
1
--— v\
~-\
0
•V
£= Beginning of addition
E= End of addition
Preliminary filter inlet
Preliminary filter outlet
Main filter outlet
!
i
«>-0
*»0
S*t:
,: -^-- r-v.,.* v-._
i :
limit
defection Z
•
J
/
i » *-.
ng / 1 % ._ —
\ \
-•*_.
"V.
10
20
30
50
Days
Fig. 1O Delayed breakthrough of hexachlorobenzene
in a preliminary filter/main filter system
-------
- 635 -
Fig. 11 shows the course of the concentration as a function
of time, at various depths in the filter, after a sudden
copper addition (4). It is clear that the concentration
decreases with increasing depth and the loading duration
becomes longer.
PPb Cu in the.water
fDosage)
0,10 >»
0,25
\ \ \
0,40 10
\ \ J**~ " . •' ' ' "-^
\ \s ••" •-
\ /-•"•-• •: —-
\ f ••--.::•- -:- • : - - -• :• i:"::"
> ... .*-...
rtl
\
$
\ \ \ \ \
\ \ \ \ \ \ \
1,00 '«
\\v\\ \
3,00
\ \ A—;
\! \ \
..N \ \
10 20 30 40 SO
11 Behaviour of a sudden loading of copper in slow
sand filtration
-------
- 636 - -. \,-v
In Table 4 are listed some of the organic substances that
have been shown to occur in anaerobic breakdown of algae,
their occurrence, and their behaviour in various trials.
As the third column shows, most of these materials could
also be found in surface water of the Ruhr. The elimination
rates in column 4 were different in the slow sand filter
model.
The formation of palmitic acid, stearic acid, and methylamine
in the upper water of slow sand filters with algal growth
could be detected. The last column of the table shows the
results of an experiment in which an algal suspension was
added to the filter. Palmitic, hexadecatetraenoic, and
stearic acids, tyramine, and methylamine were released in
such quantities that their concentration in the filter dis-
charge was even higher than in the upper water (7).
On the addition of many heavy metals a clear stimulation of
the bacterial activity could be observed. Fig. 12 shows
the variation of the colony counts for two experiments in
which mercury was added in different concentrations (8). In
the case of the lower mercury loading the increase of the
bacterial count in the upper water of the filter can be seen
in the right-hand half of the figure, together with the
increased bacterial content in the discharge from the prelim-
inary filter. The discharge from the main filter shows no
reaction, but a reaction can be clearly seen in the left-
hand half of Fig. 12, relating to 10-times greater mercury
loading. In the upper water'of the preliminary filter the
bactericidal action of the mercury is here clearly evident,
at least in the initial stages.
-------
TABLE 4 Occurrence and behaviour of fatty acids and amines from, algae
in slow sand filtration
Palmitic acid
Ilexadecatetraenoic acid
Stearic acid
Linolenlc acid
Cadaver ine
Putrescine
Tyramine
Methylamine
Release on algal
degradation
in vitro
(anaerobic)
4
44
4
44
44
44
++
•1-
Detection
in the
Ruhr water
+~<1 ug/1
44A.1-20 ug/1
44
+
4
4+
NN
( + )
NN
44
Elimination on
•slow sand
filter model
Model expt.
with high
loading
3O - 60 %
0 %
30 - 5O %
O %
99,5 %
99,0 %
O' %
50, 0 %
Concentration
increase in
upper water
of filter due
fco algal
excretion
(+>
-
+
-
-
-
_
+
Concentration
increase in
filter due to
degradation in
filter dis- -
charge in the
case of loading
with algal
cells
+
.4+
4 '
-
-
-
44
+4
-------
- 638 -
——— Preliminary filter outlet
Upper water of main filter
Main level outlet
5 ( 10
ABOUnt „ 1 50 ppb Hg
15 "Days
Bacteria/ml
105-
10'
uoooo Inlet
Upper w&ter of
~-—— preliminary filter
——-—• Preliminary- filter outlet
—— Upper water of main filter
• Main level outlet
Amount isppb HQ'
Rlter A
Fitter C
Fig. 12 Effect of loading with mercury on the bacterial
count in the preliminary gravel filter and the
main sand filter system
Fig. 13 shows that small additions of indole and skatole are
eliminated in the preliminary filter after a running-in time
of a few days (9). At the same time a fall of the oxygen
content in the preliminary filter discharge is observed,
which cannot be explained by degradation of the brganic
substances added. Their complete oxidation would have
required less than 0.5 mg O2/l. A stimulating action on
other mineralization processes must therefore be assumed.
Polarographic determination of the heavy metals with mainten-
ance of definite pH permits a differentiation of heavy
metals bound stably and weakly in complexes (Fig. 14). As
the pH decreases, increasing amounts of heavy metals are
detectable in the Ruhr water (10).
-------
- 639 -
a) Indole
[Ha/iJ
»•
20-
10-
b) Skatole-
iHS/lj
w-
5-
•
c) 02
[ma/1]
5-
djNO,
, .15
[mg/1] -
10
5-
0,
t) P04
[mfl/tj
l(H
0,5
0.
Fig. 13
fW\T
A y
r*
jfs V • • Haw water
l|i>^V x -x Preliminary filter outlet
* * Main filter outlet
r A /•
A v .
uv
\ / \
\ /
\ 7 • / '
\ A /
X / \ /
N. / \ / Tank
^v / \f containing
* - * '„.. .,_
-------
- 640.-
K/uAl
8-
0,7-
Q6-
0,5-
03-
0,2-
1 Untreated Ruhr water
(pH65
2 Acetate-buffered Ruhr water {pH 7}
3 Acidified Ruhr water (pH 1!
-0,7 -0,6 -0,5 -Q4 -Q3 -Q2 -0,1 E(V)
Fig. 14 Determination of labile (Curve 2) and total
{Curve 3) Cd-Pb-Cu by means of DPASV.
Electrolysis time 18O sec.
According to Table 5, when the loadings of the Ruhr are
relatively low, lead is found to be 100% stably bound and
cadmium to be 100% labile. Both elements were completely
eliminated in a 70 cm long experimental biological column.
Copper and zinc are present in higher concentrations and in
both cases only part of the metal is labile. As expected,
the labile copper is completely removed in' the filtration.
Only 50% of the stably complexed copper could be removed.
-------
- 641 .-
TABLE 5 Proportions of stably and weakly bound heavy metals
in river water and filtrate
Raw water
Filtrate
Elimination
PPB
»
PPB
«
%
LEI
stably
bound
1,3
1OO.O
<0,2
-
10O,O
vb
labile
<0,2
<0,2
-
CO!
itably
bound
11,O
8O,O
7,O
1OO,O
50, 0
'PER
labile
3,5
2O,O
< 2,0
-
1OO,O
Ztt
stably
bound
60,5
55,0
<2,0
-
1OO,O
1C
labile
49,5
45,0
5,5
1OO,O
9O,O
CAD^
stably
bound
<0,1
<0,1
-
ilUM
labile
0,6
1OO,O
<0,1
-
1OO,O
A different situation is observed with zinc. In this case
the stably bound fraction is completely removed and the
labile fraction only to the extent of 90%. From this it is
evident that the stability of the heavy metal complexes
does not allow any general rules to be formulated about
their elimination. The individual elements behave very
differently, according to their type and their concentration
(10).
In conclusion, some further indications may Ije given on the
efficiency of continually loaded plants with respect to
microimpurities. The investigated water-production plant
consists of a preliminary gravel filter, a main sand filter,
and a 50-m long soil passage, which is completed in 1-2 days
(Fig. 15).
The DOC shown in Fig. 16, measured with Maihak apparatus,
is reduced on average by about 50% (11). However, the
figure shows clearly that the elimination rate of the organic
carbon, contrary to expectations, is better during the winter
months than in the summer. This can be explained by loading .
with biogenic substances from algae.
-------
- 642 -
-1 ! ? i>rw.'-.*~^5 r"-"'.";l"*"'V
Ruhr gravel and shingle
L-oc- '" •' '« - » '" * ~.-J
6 • .x
' «- « " 9^
».i-:.'
.j-^*
Fig. 15 Artificial ground-water enrichment scheme
DOC
m)l\
10
S
I
Q5
o Huhr water (1 )•
n Preliminary filter outlet (H)
T Enriched ground water (5)
I fcfc J MarcJi I April | Mqy [June I Jul^ I *ug I S«pt | OC1 | No» | Dee
"
Fig. 16 Elimination of the DOC in artificial
ground-water enrichment
-------
- 643 -
Fig. 17 gives the measured elimination of the Ruhr zinc
content, which is on average 150 yg/1. The measurements were
made in the same system. The mean elimination rate is 92%
(12).
Inlet
Zn
Elimination rate 90% 80% 50%
Remobi1izatioh
<10 10
50 100
500 1000/jg/IZn
Outlet
Fig. 17 Preliminary gravel filter, slow sand filter
and soil passage(DFG 1/5)
Fig. 18 illustrates the situation for copper, the mean con-
tent of which in the raw water was 30 yg/1. In this case the
mean elimination was 65%. However, some analytical data
clearly lie in the right-hand ,field. As the copper concen-
tration in the ground water was higher than that in the raw
water, the possibility of remobilization processes must be
considered.
-------
- 644 -
Inlet
Elimination: 90V. 80'/. 50%
Remobilization
<2,5
Pig. 18
Preliminary gravel
filter, slow sand
filter and soil
passage(DFG 1/5)
5 10
50 100yug/lCu
Outlet
Fig. 19 shows the same situation as a function of time. It
can be seen clearly that the poor or even negative results
again occur in the summer months, from which it can be
deduced with high probability that algal excretion products
are involved in these processes (12).
Cu
jugl(
100
50
10
5
<2.5
-°~ Ruhr water (l)
0 Preliminary filter outlet (4) __
~r~ Enriched ground water (5)
Jon. | Feb | Mdrz | April | Mai | Juni | Juli | Aug. | Sept. | Ok«. | Nov. | Dez
1977
Fig. 19 Elimination of copper in the artificial enrichment
of ground water (DFG 1/4/5)
Whereas we have been able to extend considerably our knowledge
of the activity of slow sand filters with respect to micro-
impurities in recent years, this has given rise to many new
problems that require further elucidation.
-------
- 645—.
.(•!)• SCHMIDT, K.H.
Die Abbauleistungen.der Bakterienflora bei der Langsam-
sandfiltration und ihre Beeinflussung durch die Rohwasser-
qualitat und andere Umwelteinf liisse. Biologische Studien
zur kunstlichen Grundwas-seranreicherung
Veroffentl. d. Hydrol. Forschungsabt. der Dortmunder
Stadtwerke AG (1963), Nr. 5"
(2) STUHLMANN, F. '
Studien zum Verhalten stickstoffhaltiger Wasserinhalts-
stoffe bei der Langsamsandfiltration
Verof f entl. des Instituts' fiir Wasserforschung GmbH
Dortmund und der Hydrol. Abt. der Dortmunder Stadtwerke
(1972), Nr. 10
(3) SCHOTTLER, U.
Schwermetalle und Reinigungsleistung von Langsamsand-
filtern
Fachtagung der Deutschen .Sektion fur Limnologie in der
Internationalen Vereinigung fiir Limnologie, Siegburg
6-10 Okt. 1975
(4) SCHOTTLER, U.
Das Verhalten von Schwermetallen bei der Langsamsand-
filtration .'•-
Zeitschrift der Deutschen Geologischen Gesellschaft
126 (1975). , 373-384- -
(5) ZULLEI, N.
Polychlorbiphenyle '- Literatur - Analytik- Langsam-
sandfiltration
Verof f entl. des Instituts fiir Wasserforschung GmbH
Dortmund und der Hvdrol. Abt. der Dortmunder Stadtwerke AG
(1977), Nr. 24 ' '
(6) BAUER, U. '':'.-•
tiber das Verhalten von Bioziden bei der Wasseraufbereitung
- unter besonderer Beriicksichtigung der Langsamsand-
filtration - ", . '
Verof f entl. des Instituts fiir Wasserforschung GmbH Dortmund
und der Hydrol. Abt. der Dortmunder Stadtwerke AG (1972)
Nr. 15
(7) KLEIN, G. ,'
Studie zum Verhalten von Algenabbauprodukten bei der
Langsamsandfiltration •
(in preparation)
-------
- 646 -
(8) SCHC5TTLER, U. .
Naturliche Filtrationssysteme und Schwermetalle,
aufgezeigt am Beispiel von Quecksilber und Cadmium
J. Vom Wasser 49 (1977), 295-313
(9) KLEIN, G.
Die Bildung' von Tryptophanmetaboliten beim anaeroben
Abbau von Algen und deren Bedeutung fur Gewasser und
Wassergiitewirtschaft
Z. f. Wasser- und Abwasserforschung 9_ (1976), 2, 55-59
(1O) N&HLE, C.
Zwischenbericht 1978 zum KfW-Forschungsvorhaben
C.OO-4.O1/78 "Untersuchungen iiber den EinfluB biogener
und anthropogener Komplexbildner auf natiirliche
Filtrationssysteme" (1978)
(11) SCHOTTLER, U.
AbschluBbericht zum DFG-Forschungsvorhaben Schm 294-12
"Das Verhalten von Spurenmetallen bei der Wasserauf-
bereitung unter besonderer Beriicksichtigung der
kiinstlichen Grundwasseranreicherung"
(12) SCHQTTLER, U.
Die Aufbereitung von schwermetalibelastetern Ober-
flachenwasser durch Langsamsandfliter
11. Essener Tagung, 8-1O Marz 1978 (in press)
-------
- 647 -
REMOVAL OF TRACE CONTAMINANTS FROM, RECLAIMED WATER
DURING AQUIFER PASSAGE -
P.V. Roberts
In the arid Western regions of the U.S.A., reclamation
and reuse of wastewater is assuming an increasingly im-
portant role in planning to meet future water supply
needs. Potable reuse of highly treated reclaimed waters
is among the alternatives being considered. Potential .
health risks posed by trace organic and inorganic micro-
pollutants as well as pathogens are difficult to evalu-
ate.
Because of these risk factors, public health officials
in California require that reclaimed water may not be
used directly as a potable supply but rather only by
the indirect route of groundwater recharge. Water for
potable reuse must be treated by granular activated car-
bon with an empty bed contact time of at least 3O mi-
nutes. The treated water must then be introduced into
the groundwater either by percolation from the surface
through the vadose zone or by direct injection into a
confined aquifer. It must be demonstrated that the re-
claimed water has resided in the groundwater zone for
a minimum of one year prior to extraction for potable
reuse.
In view of the public health risks entailed by potable
reuse, the authorities certainly are justified in their
cautious stand that requires groundwater passage prior
to potable reuse. Water quality benefits anticipated to
result from passage through an aquifer .include :
-------
- 648
(1) reducing the concentrations of contaminants, and
(2) damping concentration fluctuations so as to decrease
the frequency of extreme high values. Unfortunately, da-
ta are lacking to help us evaluate what degree of quali-
ty assurance is provided. Particularly for the organic
and inorganic micropollutants that are suspected of ha-
ving chronic toxic or carcinogenic .effects, little is
known regarding elimination or retention in aquifers.
This research was undertaken to answer questions con-
cerning the contribution of .aquifer passage to improv-
ing the reliability of a water reclamation system. In-
vestigations of the transformations and fates of trace
contaminants, especially organic micropollutants, are
emphasized. Specific objectives include the following:
1. To determine the extent to which trace conta-
minants are removed.during aquifer passage.
2. To identify the processes responsible for re-
moval, e.g. biodegradation, adsorption, ion
exchange, chemical oxidation, or precipitation.
3. To quantify the rate of transport of trace con-
taminants relative to the rate of movement of
injected water.
4. To estimate the field retention capacity of
the aquifer with respect to individual trace
contaminants, where processes such as adsorp-
tion or ion exchange are believed responsible
for removal.
Experimental Methods
Reclamation Facility
The data presented herein are based in large part on
field studies carried out at the Palo Alto Reclamation
-------
- 649 „-
Facility. This facility consists of a water reclamation
plant and a well field for groiindwater recharge (1). The
hydraulic capacity is O.09 m /s.
The water reclamation plant is an advanced treatment fa-
cility based on the concept of physical-chemical treat-
ment. The unit operations include high lime addition,
coagulation, and sedimentation; ammonia removal by sur-
face aeration; ozonation; granular activated-carbon
treatment; mixed-media filtration; and disinfection
with chlorine. The anticipated quality of influent and
effluent is summarized in Table 1.
During the observation period from which the data in
this paper are drawn, the reclamation plant was operated
in a start-up mode. Treatment consisted of lime treat-
ment at pH 9, air stripping, and'recarbonation, followed
by ozonation and sand filtration. The ozone dose was
approximately 50 mg/1. Owing "to the absence of activated-
carbon treatment, the organic* quality of the effluent
was not as good as anticipated for the full treatment
sequence. The average COD concentration in the effluent
was 20 mg/1.
The full-scale injection/extraction well field is composed
of a network of 9 injection wells, 9 extraction wells, and
62 monitoring wells in an area approximately 3 kilometers
long by 1 kilometer wide. The wells are intended for re-
charge and removal of water in an aquifer approximately
12 to 15 m below the ground surface and 1 to 3 m thick.
Pilot Experiment
A pilot experiment was conducted at a test well from
August to November 1977 to gain experience with direct
-------
- 65O -
TABLE 1
Projected quality of reclamation plant effluent
based on present Palo Alto secondary effluent
average characteristics and results of treatment
studies (2)
COD, mg/1
MBAS, mg/1
Nitrogen, mg/1 as N
NH3
N0~
NO3
Organic
Total
Heavy metals, yg/1
Cd
Zn
Cu
; Pe
i
! Mineral characteristics,
; rag/1
Sodium
Potassium
Calcium
Magnesium
Chloride
;
| Sulfate
Present -Palo
Alto Secondary
Effluent Average
Characteristics
• 53
0.12
24
O.3
0.4
"3."o
27.7
. - , ,2.5 ,
- 5
49
, 3
2.3
.". 1,62
11
43. , .
. -'15".
204 ' .-
85
Projected
Characteristics
of Reclamation
Plant Effluent
6
O.-O5
2-
0.3
0.4
0.5
2.7
-------
- 651 -
injection of reclaimed water prior to beginning injec-
tion and extraction at full scale. The test site for the*
pilot experiment is shown schematically in Figure 1. Re-
claimed water was injected into the recharge aquifer
through Well 12. Wells P1 through P4 are sampling piezo-
meters completed in the same aquifer, located approxima-
tely 8 m from the injection well. From core samples and
drawdown test results, it was concluded prior to the
commencement of injection that the injected water would
•flow preferentially in the direction of Well P4. A
drawdown test at Well 12 indicated an average trans-
miss ivity of .7 x 1O m /m.s based on the assumption of
radial flow, but the transmissivity value in the direc-.
tion-of P4 is believed to be substantially greater.. ..The .
porosity of a core sample from Well P4 was 0.22. i
PI
S3
P4,
P2
X
: :0.3m
>XP3
I.Om
2.1m
.5m
0.3m 12 ±O.3m
Fig. 1 .Injection and, observation well arrangement.for
' 'pilot study. 12 is the injection well; P1, P2,
P3, P4, and S3 are observation wells
-------
- 652
Analytical Methods
Highly volatile organic substances were analyzed using
a head space analysis technique (3). Compounds deter-
mined by this method included halogenated aliphatics
containing one and two carbon atoms.
Moderately volatile compounds were determined using a
closed-loop stripping procedure (4). Concentration of
the organic solutes was achieved by circulating air
for a period of two hours through a loop that passed
through the 5OO-ml sample and a 1-mg activated-carbon
adsorption trap. The adsorbed organic substance was
then eluted with carbon disulfide. The extracts were
separated by gas chromatography on a 20 m UCON HB510O
glass capillary column using a Carlo Erba gas chroma-
tograph with flame ionization detector. Quantification
was based on comparison with 1-Cl-Cg/ l-Cl-C.,-, and
1-Cl-C^g internal standard peaks. Identification of
peaks was achieved by means of a Finnigan Model 4000
gas chromatograph-mass spectrometer.
Trace metals were determined by flameless atomic absorp-
tion spectrophotometry (Perkin Elmer 403) following APDC
chelation and methyl isobutyl ketone extraction of un-
filtered, acid-preserved samples.
Interpreting Field Data
The experiment described is equivalent in principle to
imposing a step change in the composition of fluid at
the injection point. Prior to injection, the water in
the aquifer was relatively homogeneous. As injection
-------
- 653 -
proceeded, the formation groundwater was displaced out-
ward from the injection point toward and past the ob-
servation wells, resulting in concentration changes at
the sampling points.
Estimating the Rate of Transport and the Field Retention
Capacity
The rate of transport and field retention capacity for
a specific pollutant can be estimated from observations
of the concentration history at an observation point
following a step change in concentration at the injection
point (5). The approach is premised on analogy to chemi-
cal reactor analysis (6). The aquifer is treated as a
reactor of arbitrary shape and volume. The effective
pore volume of the aquifer element is evaluated from the
integral
V = /°°(1 - fTT7) dVTr7 ...
p o IW IW (1 )
where V is the effective pore volume in m , fiw the
fractional breakthrough of injection water as measured
by a conservative tracer, and VTW is the volume of water
3
injected in m . The field retention capacity of the aqui-
fer with respect to pollutant i is given by
( dviw
aq o
where r. is the specific retention capacity in g of com-
1 3
ponent i retained per m aquifer;(C ). is the average
3
concentration of i in the injected water, g/m ; f. is
the fractional breakthrough of component i , a dimension-
-------
o' <1 - fl
o--" <1 - fi
w' dviw
> dviw
- 654 -
less function of the injected volume; and e is the
aq
effective porosity of the aquifer element. The ratio of
the average transport velocity of a pollutant to that
of water is derived from a mass balance for the aquifer
element:
ui =
T] "°° I 1 — f \ JTT ( 3 )
H2°
where UH „ is the average velocity across the aquifer
element Boundary in m/s ; and U. is the corresponding
velocity of pollutant i.
Identifying Transformation Processes from Field Data
The concentration response at an observation well can
be interpreted to identify the processes affecting the
transport of a pollutant. Several types of concentration
responses are illustrated in Figure 2, A conservative
tracer that does not interact within the aquifer will
appear as a sharp concentration front in the absence
of dispersion, or as an S-shaped wave if dispersion is
significant (6,7). A solute which is adsorbed in the
aquifer will be delayed compared to the conservative
tracer, and the length of the concentration front will
be extended. The concentrations of conservative as
well as sorbing solutes eventually reach the upper limit
imposed by the concentration in the injection water, at
which point breakthrough is complete. If a solute is
biodegradable under the conditions in the aquifer, its
concentration may first rise until a population of mi-
croorganisms has developed that is capable of metaboliz-
ing it after this acclimatization period; the concentra-
tion may then decline to a steady-state level, as shown
in Figure 2.
-------
- 655 -
1.0
EXPECTED RESPONSES TO A
STEP CHANGE IN CONCENTRATION
C.
Co
0
PlugFbw'/DisPersion
Adsorption
and Dispersion
J^qdegradation
and DisF>ersTo7T
0123
TIME RELATIVE TO MEAN RESIDENCE
TIME OF WATER, t/tH2o
Fig. 2 Forms of response to a step-change concentration
stimulus
Hence, processes such as adsorption or biodegradation
that may transform a solute or attenuate its movement
can be identified by comparing the concentration res-
ponse to that of a conservative tracer and to the ty-
pical forms such as those in Figure 2.
Results
Breakthrough of Injected Water
The rapid appearance of water of low salinity at Ob-
servation Well P4 proved the existence of a good hy-
draulic connection between this well and the injection
well, confirming the expected preferential flow in the
direction of Well P4 compared to other directions. Hence,
monitoring of trace contaminants was limited to Well P4
to the exclusion of the other observation wells.
-------
656 -
The breakthrough of injected water was calculated from
conductivity measurements. The correlation coefficient
between conductivity and chloride was O.967 for 17 paired,
values, which is significant at the 99.99-percent level.
The breakthrough at the observation well commenced with-
in the first 1O m injected volume, reached a fractional
value of O.5O at 35 m injected volume, and was virtu-
ally complete after 20O m had been injected (Figure 3).
The effective pore volume of the aquifer element defined
by Well P4 is 45 m , evaluated from the integral in Fi-
gure 3. The average residence time in the aquifer ele-
ment is approximately 12 hours, at an average injection
rate of one liter per second.
FRACTIONAL BREAKTHROUGH OF
INJECTED WATER
cr
u.
-tr
2OO
f NAREA=/(l-fIW) dVIW = 45m3
O
0 50 (00 150
CUMULATIVE VOLUME INJECTED, VIW, m
200
3
Fig. 3 Breakthrough of injected water at observation well
P4
-------
- 657 -
Concentrations of Organic Micropollutants
The injection water contained measurable amounts of re-
sidual organic micropollutants that were not removed in
treatment. An overview of the compounds regularly quan-
tified in the injection water at concentrations exceeding
1OO ng/1 is presented in Table 2. Their large number may
be explained by the fact that the water did not receive
activated-carbon treatment during this period of obser-
vation..Subsequent observations after commencement of
activated-carbon treatment have shown that substantially
lower concentrations of organic micropollutants are at-
tained compared to the values reported here.
TABLE 2
COMPOUNDS REGULARLY QUANTIFIABLE IN INJECTION WATER
CRITERION: CIW>IOOng/.|
CLOSED LOOP STRIPPING ANALYSIS
HEAD SPACE ANALYSIS
CHLORINATED AROMATIC COMPOUNDS
Cl
"Cl
CMLOROBENZENE
Cl Cl
1,2 1,3 1.4 1,3,4
D1CHLOROBENZENE TRICMUOROBENZENE
ISOMERS
AROMATIC HYDROCARBONS
CH=CH2
STYRENE
NAPHTHALENE
ARYL AND ALKYL CYANIDES
CN
ALSO: €5 TO Cg ALKYL CYANIDES
C.H,
• CN
BENZONITRILE
TRIHALOMETHANE COMPOUNDS
CHCI, CHLOROFORM
CMCIg Br BROMOOICHLOfiOMETHAME
CHCIBr2 DIBROMOCHLOROMETHANE
CHBrj BROMOFORM
OTHER CHLORINATED ALIPHATIC COMPOUNDS
ClgC'CHCI TRKHLOROETHYLENE
Cl2C=eCI2 TETRACHLOROETHYLENE
CljCCHj 1,1,1-TRICHLOROETHANE
-------
- ^658 -
Aromatic Compounds
Data for aromatic and substituted aromatic hydrocar-
bons are summarized in Table 3. From gas chromatogra-
phic results it is apparent that a small number of
compounds predominate in the injection water (Figure 4).
The sources of chlorinated benzene compounds are be-
lieved to be industrial solvent wastes, while benzo-
nitrile is thought to be present in electroplating
wastes. Heptaldehyde and styrene may be products of
ozonation and chlorination in the reclamation plant (8),
The concentrations of organic micropollutants analyzed
by closed-loop stripping analysis were near or below
their respective detection limits at the observation
well shortly after injection began (Table 3). The con-
centrations of some compounds rose appreciably during
the course of the experiment, as exemplified by chloro-
benzene, styrene, and benzonitrile. During the final
TABLE 3
CONCENTRATIONS OF AROMATIC AND SUBSTITUTED AROMATIC MICROPOLLUTANTS
CHLOROBENZENE
1, 3-DICHLOROBENZENE
1,4-DICHLOROBENZENE
1, 2it>ICHLOROBENZENE
1, 2, 4-TRICHLOROBENZENE
NAPHTHALENE
STYRENE
HEPTALDEHYDE
BENZONITRILE
CONCENTRATIONS, ng/Jl
INJECTED WATER
ENTIRE PERIOD;
V = OT045OOm3
LOG
MEAN
4,130
630
95% CI FOR MEAN
(n = 9)
1,480 TO 1 I.5OO
255 TO 1,550
530 ' 265 TO 1 ,060
1,940
150
910
1,000
1 1,700
5,500
1,160 TO 3.25O
38 TO 590
320 TO 2,540
500 TO 2,000
9,500 TO I4.5OO
3,300 TO 8,900
OBSERVATION WELL
Vny = 0 TO 5OOm3
-------
659 -
third of the observations* the concentrations of chlo-
robenzene and styrene "were not"significantly different
from their respective -concentrations in the injected
water, as demonstrated by the overlap of the 95 % con-
fidence intervals for the mean (95 % CI in Tpble 3).
This is taken as an operating definition of "complete
breakthrough."
A
16
12 9/15/77
2000 NG/L IS
15 II ' 8
1 ? '? f Is?
P4 INITIAL
B
15,
C
1
ra
—^~L^^
50 40 '
170° ISO" !3O°
200 NG/L IS
15 - - . -
P4 9/15/77 ', " '•'•
10CO NG/L IS. ' •
8" •
7-'
f
5
4A
(-80
-70
• 60
••50
-40
-30
• 20
- 10
H-Ut-x^ium. , Q
- ' ' 60
• 50
- 40
- 30
• 20
.- 60
5 . 8 '-
n •' -
.if . f llr»9 •
30 . 20
IIO° 9O° 70°
7
— «_
10
6
1
5
43o
ILjmcj
• 50
- 40
• 3C
. • 20
• 1C
^ .- 0
• 5 TIME ICT
50° TEMP ;°
LjJ
_ J
t )
_1
f
fc
a
E:
i
in)
C)
Fig. 4 Gas-chromatographic analyses of injected water and
observation well samples
Substances identified are: (1) toluene + tetrachloroethylene;
(2), ethylbenzene; (3) p-xylene; (4) m-xylene; (5) chlorobenzene
with trace of o-xylene; (6) unknown y; (7) sturene + unknown;
(8) ClC8 {Internal Standard)? (9) 1,3-dichlorobenzene;
(1O) 1,4-dichlorobenzene; (11) 1,2-dichlorobenzene; (12) benzo-
nitrilej (13) 1,2,4-trichlorobenzene; (14) naphthalene;
nS) C1-G12 (Internal-Standard) ;. (16) C1-C116 (Internal Standard)
-------
- 660 -
The concentration response of chlorobenzene is plotted
in Figure 5 in a form suitable for estimating the
field retention capacity and the relative transport
velocity. According to Eq. 2, the field retention ca-
pacity for chlorobenzene is
Co0/ro
dviwi °-0041
m
158o[m3]
1
"aq o
- fiw) dviw.
0.22
45 [m3]
= O.032 g chlorobenzene re-
tained per m aquifer
•and the ratio of the transport velocity to that of wa-
ter according to Eq. 3 is ' . '
u
•C^H5C1
ssL C\ Id*.ii i I i i i i I i i i i I i i'i i I t i i i I i i i i 1 i i i 'I i r i i i i i i i i
h 0 1000 2000 3000 4000
tr
CUMULATIVE -INJECTION VOLUME, m3
Fig. 5 Response of chlorobenzene 'concentration to a
step-change stimulus
-------
- 661 -
Halogehated Aliphatic 'Compounds
Data for the components analyzed by head space analysis
are shown in Table 4. During the early portion of the
experiment (VTW < 2OO m ) the concentrations at the ob-
servation point were reduced to less than the detection
limit for all components. Breakthrough was observed at
a point corresponding to 1OOO m injected volume or less.
For all compounds, the 95 % confidence limits for con-
centrations at the observation well during the period
3 3
embracing injected volume between 10OO m and 15OG m
overlap those for the injected water. Unfortunately the
confidence limits are quite broad, owing to the varia-
bility of the analyses and the small number of samples.
Only, the data for chloroform are sufficiently reliable
to permit interpretation in the form of a breakthrough
response (Figure 6). It appears that the midpoint of
the chloroform breakthrough is reached at approximately
2OO m injected volume. The breakthrough appears to be
virtually complete by 1000 m injected volume. The rate
TABLE 4
Concentrations of halogenated aliphatic
micropollutants •
CHLOROFORM
1,1,1-TRICHLOROETHANE
TRICHLORETHYLENE
BROMODICHLOROMETHANE
TETRACHLOROETHYLENE
DIBROMOCHLOROMETHANE
BROMOFORM
CONCENTRATIONS, fj.q/jl
INJECTED WATER
ViW=O-TO I50Om3
LOG'
MEAN
3.3
2.7
9.9
I.I
O.5
5.1
3.3
95% CI
FOR MEAN
'1.9 TO' 5.8
0.63 TO 11.5
4.i TO, 24 ,
O.3 TO 4.3
O.I2TO 2
2 TO 13
0.33 TO 33
OBSERVATION WELL
V1W=OTO 2OOm3
LOG
MEAN
<0.l
< 0. 1
< 0. 1
-------
- 662 -
of transport of chloroform can be approximated only'
very roughly. The 'best estimate of the ratio of the
chloroform transport velocity to the velocity of water
movement is
u
CHC1.
45
uwater 45 + 2O° 5
evaluated from Eq. 3. The field retention capacity, for
chloroform is
x 200 [m3]
0722 X 45 fm ^
=' O.OO32
g chloroform retained
m aquifer
BREAKTHROUGH-OF CHLOROFORM
. o
AREA= /(fiw-fcHCI J dVlW = 200 m3
0 ' • ' .
0 INJECTED WATER
o CHCI3
.5 I.O I.5 2.0
CUMULATIVE VOLUME INJECTED, I03m3
2.5
Fig. 6 Response of chloroform to a step-change stimulus
-------
- -663 -'
Biodegradation of Individual Organic Compounds -•: M
The behavior of naphthalene exhibited strong evidence .
of the influence of biodegradation (Figure 7) ., The conj:
centration rose significantly above the background value
during the period following the breakthrough of injected
water. The concentration exceeded 1OO ng/1 in the range
between 40 m and 15OO m Injected volume, exceeding 10
percent of the average injected concentration through-
out that range. A peak concentration of 90O ng/1/ ap-
proximately equal to the average injected concentration,
' 3
was reached at an injected volume.. of 680 m . After
15OO m injected volume, the concentration decreased
below 1O percent-of the average injected concentration
for the remainder of'the observations. This decrease
cannot be explained by a secular decrease in the con-
centration in the injected water.
NAPHTHALENE RESPONSE
o --
— INJECTED WATER
---NAPHTHALENE
1234
CUMULATIVE VOLUME INJECTED, IO3m
Fig. 7 Response of -naphthalene
-------
- 664 - =
A plausible explanation for the decline in naphthalene
concentration at the observation well is that the rate
of biodegradation was enhanced by the development of a
population of microorganisms capable of metabolizing
styrene. It is hypothesized that acclimation occured
during the initial breakthrough of naphthalene. The
onset of biodegradation .after an initial lag period
is commonly observed in degradation studies. Further-
more, it has been postulated that substrates will be de-
graded in natural systems-'to the point at which a low,
steady-state concentration is reached (9). McCarty (9)
estimated the steady-state concentrations of acetate
and glucose after an infinite residence time in an
— 8
aerobic environment to be 25 x 1O mol per liter and
_ o
13 x 10 mol/liter, respectively. These conditions cor-
respond to the open ocean, or -large oligothrophic lakes.
The steady-state concentration of naphthalene observed
in the groundwater environment in this work is 0.47 x
-9 -9
1O + O.17 x 1O mol per liter. Hence, the value for
naphthalene is two to three orders of magnitude lower
than expected for more readily degradable substrates
such as acetate and glucose. From this comparison it
can be hypothesized that the groundwater zone consti-
tutes an environment especially amenable to biodegra-
dation. '
It is possible that biodegradation also was responsible
for the removal of heptaldehyde. However, the field data
provide no confirming evidence in the form of a concen-
tration peak followed by a decline to a low, steady-
state concentration. Since heptaldehyde is believed to
be readily degradable, it is conceivable that the acclim-
atization of microorganisms occured within the time
frame of the breakthrough of injected water.
-------
- 665 -
It .might be suspected that,the time of travel between
the injection and observation points, i.e. approximately
12 hours, was too short for biodegradation of poorly de-
gradable substrates to occur. Concentrations of organic
micropollutants were determined before and after a 25-
day "rest period" during which no water was injected in-
to the aquifer (Table 5). Since hydrogeologic studies
showed no evidence of regional flow in the aquifer, it
can.be assumed that essentially the same water was sam-
pled at the beginning and end of. .the period. Of the
eight compounds analyzed, the. concentrations of six
appeared to decrease, while two-others increased. Be-
cause only a single pair of measurements was made, no
statistical conclusions can be. reached. It is believed
that only changes by a factor- greater, than two can be
considered significant. Only the .decrease in the con-
centration of styrene, which amounts to a tenfold change,
is thought to be significant. Biodegradation is a plau-
sible explanation for the disappearance of styrene (1O).
TABLE 5
CONCENTRATION CHANGES OF ORGANIC MICROPOLLUTANTS
DURING A 25-DAY RESIDENCE PERIOD
CHLOROFORM
TRICHLOROETHANE
TRICHLOROETHYLENE
TETRACHLOROETHYLENE
CHLOROBENZENE
1, 2-DICHLOROBENZENE
STYRENE
NAPHTHALENE
CONCENTRATIONS, /ig/l
INITIAL
2.4
2.3
3.3
1. 8
2.4
0.38
2. t ,
0.035
AFTER
25 DAYS
2.7
2.2
3.I
I.I
• I-7
0.18
0.15
0.05
CHANGE,
PERCENT
+ 12
- 4
- 6
-40
-30
-47
-93
+ 40
-------
- 666 -
Chloroform, trichloroethylene, tetrachloroethylene, chlo-
robenzene, 1,2-dichlorobenzene, and naphthalene were not
degraded to a significant extent under the conditions of
this experiment. However, it must be borne in mind that
the conditions were anoxia, less than O.5 mg/1 dissol-
ved oxygen.
Moreover, the concentration of naphthalene at the samp-
ling point was very low, corresponding to the steady-
state level that presumably represents the lower limit
attainable by biodegradatipn. Hence, it may not be in-
ferred from these data that,, the1 substances in question
would not be degraded in other,situations in which their
concentrations were higher-and sufficient oxygen were
present. _ '-.•••• ;
Dissolved Oxygen and Organic Collective Parameters
Dissolved oxygen, COD, and „TOG were, determined in paired
samples from the injection well head and Well P4 (Table 6),
There was a consistent decrease in dissolved oxygen con-
centration between the injection well and Well P4. The in-
jected water was saturated with oxygen, having an average
concentration of 8 mg/1. Samples from Well P4 consistently
showed an oxygen concentration less than O.5 mg/1. The
disappearance of dissolved oxygen coincided with a de-
crease in the COD concentration of the same magnitude
(Table 6) . '
From this agreement it can be inferred that aerobic de-
gradation of organic substances occurred during the pe-
riod of residence in the aquifer. The concentration of
total organic carbon indicates that the biodegradation
must have been relatively complete. The ratio of the COD
-------
- 667 -
TABLE 6 DECREASE IN CONCENTRATIONS OF DISSOLVED OXYGEN AND
COLLECTIVE ORGANIC PARAMETERS DURING AQUIFER PASSAGE
DISSOLVED OXYGEN, mq/Ji 02
COD, mg/J Og
TOO, mg/J C
(ACOD)/(ATOC)
DECREASE BETWEEN
INJECTION AND
OBSERVATION WELLS
MEAN
8.5
8.8
3.4
2.6
STD, DEV,
5.5
2.1
NUMBER
OF PAIRED
OBSERVATIONS
16
13
decrease to that for TOC corresponds to the range of
2.5 to 3.5 expected when organic substances present in
wastewater are completely oxidized to end-products such
as CO2 and H2O (11). However, the data are insufficiently
precise to permit a firm conclusion in this regard.
It is unlikely that the observed reduction in the con-
centrations of COD and TOC resulted from the removal of
particulate organic material. The turbidity of the in-
jected water was less than 2 FTU.
Removal of Trace Metals
Concentrations of trace metals are summarized in Table 7.
Ag and Cu were removed during aquifer passage throughout
the observation period. Cd and Pb were removed during the
early part of the period, but concentrations at the ob-
servation well were not significantly different from
those in the injected water at the end of the period.
The concentration of As was higher at the observation
well than at injection point throughout the experiment, "
possibly owing to dissolution or desorption of As from
aquifer minerals. •
-------
- 668 -
TABLE 7
REMOVAL OF TRACE METALS DURING AQUIFER PASSAGE
Ag
As
Cd
Cu
Pb
CONCENTRATIONS./zg/l
INJECTED WATER,
MEAN ± STD. DEV
FOR ENTIRE PERIOD
2.0 ± 1.6
l.2± 0.7
2.0± 0.8
102 ± 24
1.3 ± 1.0
5 SAMPLES DURING
EARLY PART
OF PERIOD
V = OTO 500m3
INJECTED
MEAN
< 0.5
15
1.0
4.4
0.6
STD. DEV.
5.4
0.9
3.7
0.2
5 SAMPLES DURING
LATTER PART
OF PERIOD
V=3000T045OOm3
INJECTED
MEAN
< 0.5
12
1.6
28
1.6
STD. DEV.
—
2.5
1.7
3.2
0.7
The breakthrough of Cd and Cu are compared to that for
the injected water in Figure 8. The fractional break-
through of Cu reached a value of approximately O.3 at
the end of the experiment after 45OO m had been injec-
ted. Hence the rate of transport for Cu is estimated to
be less than one-hundredth as great as the rate of move-
ment of the water through the aquifer. The breakthrough
of Cd is more rapid, the midpoint being reached within
an injected volume of 15OO m . The rate of Cd trans-
port is estimated to be approximately one-fortieth as
rapid as that of water movement.
Adsorption is the probable mechanism for trace metal re-
moval in the aquifer. The field retention capacity for
Cd is calculated to be 17 mg Cd retained per m3 aquifer.
Since the breakthrough of Cu was incomplete, we can only
evaluate a lower limit for the field capacity, approxi-
mately 2 g Cu per m aquifer.
-------
- 669 -
RESPONSE OF TRACE METALS
l.y
X
o
o
K
H 1 O
5 ''U
<
Ld
CC
CD
XAREA=J(l-fCd)dVIW = l600m3
Ix O
^s o ___ja— o- — — '
VQ°'^ i i i i
1234
CUMULATIVE VOLUME INJECTED, I03m3
Fig. 8 Responses of Cd and Cu
Conclusions
Much can be learned about the behavior of trace
contaminants in the groundwater environment by
conducting controlled stimulus-response experi-
ments under realistic field conditions.
From the response at an aquifer sampling point
following a step change in concentration at the
injection point, insights into water quality
changes can be obtained. The processes respon-
sible for removal can be identified tentatively.
The transport rate relative to water and the
effective field capacity can be estimated for
compounds for which complete breakthrough is
observed.
-------
- 670 -
3. Evidence of degradation in the aquifer environ- "•'
ment was seen for naphthalene and possibly
styrene. Naphthalene was degraded to a steady-
— Q
state concentration of O.5 x 1O mol per liter
following initial breakthrough.
4. Chlorinated aliphatic and aromatic compounds are
retained effectively by adsorption during aquifer
passage. Their concentrations initially are re-
duced to less than O.1 yg per liter. However, the
adsorption capacity ultimately is satured and
breakthrough occurs.
5. The specific field retention capacity for chlo-
robenzene was estimated to be O.032 g chloroben-
zene per m aquifer, at an average injected con-
centration of 4 yg per liter. The specific re-
tention capacity for chloroform was a factor of
ten smaller, O.O032 g chloroform per m aquifer
at approximately the same injected concentration.
6. Organic micropollutants vary widely in the rate
at which they are transported through an aquifer.
The chlorinated aliphatic compounds, exemplified
by chloroform, are transported most rapidly among
the substances studied in this work. Chloroform
was transported one-fifth as rapidly as the water
with which it was injected. Chlorinated aromatic
compounds are transported much less rapidly.
Chlorobenzene moved 36 times more slowly than
the injected water, for example. Dichloroben-
zene and trichlorobenzene isomers are transpor-
ted much more slowly.
7. Trace metals are removed in the aquifer by ad-
sorption. Cd was transported most rapidly among
the trace metals studied. Cd travelled with a ve-
locity one-fortieth that of the injected water.
The specific field retention capacity for Cd
3
was O.O17 g Cd per m aquifer.
-------
- 671 -
Acknowledgments
This work is funded by the U.S. Environmental Protection
Agency, Research Grant No. R-804431. Additional finan-
cial support was received from the Department of Water
Resources and the Water Resources Control Board, State
of California. The study is a cooperative venture with
the Santa Clara Valley Water District, who operate the
reclamation facility. The author is indebted to Dr.
Martin Reinhard and Joan Schreiner for the organic cha-
racterization data and to D. M. Mackay and G. D. Hopkins
for the trace metal determinations.
-------
- 672 -
(1) ROBERTS, P.V., MCCARTY, P.L., REINHARD, M., SCHREINER,
Direct Injection of Reclaimed Water into an Aquifer
J. Environ. Div. ASCE (1978) (in press)
(2) MCCARTY, P.L., SCHERTENLEIB, R., NIKU, s.
f Preproject Water Quality Evaluation for the Palo Alto
Water Reclamation Facility
-Technical Report No. 2O6, Civil Engineering Department,
Stanford University, Stanford, CA (1976)
(3) BELLAR, T.A., LICHTENBERG, J.J.
Determination of Volatile Organics at the pg/1 Level
in Water by Gas Chromatography
J. AWWA 6^7 (1974) , 634
(4) GROB, K., ZURCHER, F.
Stripping of Trace Organic Substances from Water
J. Chromatography 117 (1976), 285
(5) ROBERTS, P.V., McCARTY, P.L., REINHARD, M.
Direct Injection of Reclaimed Water: Attenuation of
Organic Contaminant Movement
J. WPCF, submitted for publication (1978)
(6) LEVENSPIEL, O.
Chemical Reactor Engineering
Wiley, New York (1962)
(7) FRIED, J.J.
Groundwater Pollution
Elsevier Scientific Publishing Co.-, Amsterdam (1975)
(8) SIEVERS, R.E. et al.
Generation of Volatile Organic Compounds from Non-
volatile Precursors in Water by Treatment with Chlorine
or Ozone
Water Chlorination: Environmental Impact and Health
Effects, Ann Arbor Science Publ. Inc. Ann Arbor, Mich.
£ (1978) , 615-624
(9) MCCARTY, P.L.
Energetics of Organic Matter Degradation
Water Pollution Microbiology, Wiley Interscience, Inc.,
New York (1972), 5
(10) SIELICKI, M., FOCHT, D.P., MARTIN, J.P.
Microbial Transformations of Styrene and C Styrene
in Soil and Enrichment Cultures
Appl. and Environm. Microbiology 35 (1978), 1, 124
(11) WAGNER, R.
Abbaubarkeit und Persistenz
Vom Wasser 4O (1973), 335
-------
- 673 -
BIOLOGICAL "PROCESSES FOR THE TREATMENT OF DRINKING WATER
J. Chedal
Following the lectures of this morning I should like to des-
cribe our experience with the biological elimination of
ammonia.
We tested the sludge-bed clarification plant described
by Dr.J.B.Goodall and found that this process has two dis-
advantages when applied to the raw waters used by us:
fluctuating ammonia elimination rates
when starting from an empty plant, a very long time
until an active sludge layer is formed (about 1 month).
The use of sand particles as in the Fluorapid plant definitely
makes it possible to eliminate the latter disadvantage to a
certain extent. It should be added by way of explanation
that our experiments were performed with water containing
less than 2 ppm of ammonia.
It has also been established that the above-mentioned draw-
backs can be partially eliminated by increasing the ammonia
content in the v?ater. However, we consider it ridiculous to
add ammonia to the water only to remove it again. For this
reason we do not plan to incorporate a process of this kind.
As regards the biological filtration, the experiments so far
have shown that this plant operates much more satisfactorily
and delivers constant deposition of about 80% even in winter.
The main features of the plant are:
increased water flow (maximum velocity about 18 m/h)
-------
-• 674
- air flow
carrier layer enriched with trace impurities of the type
of Biodamine.
As is well known, the filter is of the over-damming design,
in contrast to the bacteria beds in clarification plants.
In conclusion, reference may be made to the biological
activity of the rarely-mentioned storage process.
Our experiments have shown that with a residence time of 2
days this process makes it possible to eliminate about 50%
of the ammonia.
-------
- 6'75 -
NITROGEN REMOVAL IN BIOLOGICAL REACTORS AT LOW
TEMPERATURES
G. Halm0, K. Eimhjellen and T. Thorsen
Laboratory tests have been performed with continuous
biological reactors treating sewage and organic chemiT
cals at low temperatures. Three steps were operated in
series: Activated sludge, nitrification and denitrifi-
cation. Most effort was done to investigate the deni-
trification step and its performance.
The tests
Both activated-sludge and•denitrification steps have
shown better performance using psycrophilic or psycro-
trophic sludge if water temperatures are low. The low-
temperature bacteria have.significant effectiveness
right down to O C, optimum at .about 15 C and maximum
at about 20 C. Previous tests have shown that a deve-
loped low-temperature sludge has greater denitrification
velocity than high-temperature (mesophilic) sludge
over a greater part of the low-temperature range. It
has been realized, however, that around 17°C, an irre-
versible change destroys the good performance of low-
temperature operation.
Tests in a continuous laboratory plant,involving all
three steps, have been done at 5 ± 1°C. Due to high
COD reduction in the act, sludge step, methanol was
added to the denitrification step to ensure enough
feed for the bacteria. Between each change in operating
conditions, 1-4 weeks was allowed to attain pseudo-
equilibrium. •
-------
- 676; -
Results ... . . , . ... ...„.,,,,
Because of unpredicted destruction of the Nitrobacter-
types in the nitrification step, this step only
oxidized ammonia to nitrite during the later part of
the tests. However, this did not influence the effec-
tiveness of the denitrification step. Mean values
from several tests are shown in Table 1.
The results show that biological N-removal is easily
obtained at low temperatures. With the mentioned resi-
dence times approx. 9O% of total N is removed in sewage
at 5 C. The nitrification/denitrification steps remove
approx. 95% of N from the act-sludge step. Denitrifica-
tion alone is more than 98% effective. Previous tests
indicate that about equal performance can be expected
at least down to 3°C. With increasing temperature up to
about 17°C, the same performance may be attained with
shorter residence times.
TABLE 1 Mean values of test results for nitrogen (ppm)
Influent
NH/
43
48
NO2+NO3
0.6
0.5
Res . time -*•
After acti-
vated sludge
NH4 +
27
34
N02+N03
2.2
0.6
1 .5 hours
After nitri-
fication
NH4 +
2.5
9
NO~+NO3
25
32
9 hours
After de-
nitrification
NH4
1 .5
8
NO2+N03
0.4
0.4
4 hours
NB: First line: Mainly NO., produced in nitrification step;
Second line: Mainly NO^ produced in nitrification step.
-------
- 677 -
Sedimentation of psycrophilic sludges is significantly" '
better than corresponding mesophilic sludges. No bulking
problems were observed contrary to experiences' with
mesophilic sludges in the-same plant.
On-ly methanol was used as electron-donor and source of
carbon. Approx. 3 mg methanol is needed for each mg of N.
Microbiological investigations show that none of the
dominant bacteria are obligate methylotropes.
The results show that psycrophilic/psycrotrophic slud-
ges can be used with advantage for removing N at low
temperatures, particularly below 5 C. It seems that ni-
trification is the most sensible step, and further in-
vestigations should be performed here.
Even though these experiments-,were done with sewag'e and
chemicals, the results in general should be useful to
drinking water treatment. , . -
-------
- 678 -
MICROBIOLOGICAL STUDIES ON ACTIVATED CARBON FILTRATION
P. Werner, M.Klotz and R. Schweisfurth
1. Introduction
The results relate to the processing of Rhine water into
drinking water by means of activated carbon.
In regards to the method, the determination of the microorganism
count in the water poses no real problems if suitable nutrient
media, incubation temperature, and incubation time are used
after appropriate preliminary treatment. Determination of the
colony counts in water according to the recommendations of the
Deutsche Einheitsverfahren is unsatisfactory, .since this method
only reflects a very small and non-representative part of the
microflora. The reported colony counts were obtained by a
special method and are substantially higher than ones obtained
according to the Deutsche Einheitsverfahren. Therefore, in
the light of these values, no statements on the hygienic and
bacteriological state of the water car be made in accordance
with the current Drinking Water Decree (Trinkwasserverordnung)[1]
2. Quantitative determination of microorganism populations
When activated carbon is used in the treatment of drinking
water a bacterial population in the filtrate always occurs.
The colony counts in the characteristic treatment stages are
shown in Fig. 1.
As a rule colony counts of 2*10 /ml are found in the raw
water. By -means of the preliminary chemical treatment of the
water the colony counts can be reduced to about 10 /ml before
the input to the activated carbon 'filter. If this is followed
by a high-dose chlorination, nearly all the bacteria are
destroyed. In the activated carbon filter an increase of the
-------
679 -
bacteria to values of about 7?10 /ml again occurs. The acti-
%• x
vated carbon filter outflow is normally hygienically and
bacteriologically unimpeachable in accordance with the current
Drinking Water Decree,
10D-
105-
104-
103~
102~
10* J
BAf
(KC
CTERIEN
WNIEZAHLlmlWASSBR) a)
•^n?-rt~" ."^T
1
PtlsgfSpl
¥ft§8
•
/ \
ROH-
WASSER
b)
""I'*5''I^^J"J^
^'
AKTIVKOF
E1NLAUF
OHhifMlf
CHLOR
HI
X*X* XvMvi
ti***t"r vXvX*.
ft***i*c **•*****%••>
Ulill
^iilll
r'x% ;¥:-X:K-'.-
•r-X'X'y.v:^:-:
Hpyiii
:?SS::xSr;
;:tv:::vi::::: •:'!'
\-^ ^.
HLEFILTER
lAUSLAUF
c)
\
Fig. 1 Colony counts on SPC Agar at characteristic treatment
stages in a Rhine water treatment plant
Key; a) = Bacteria (colony count/ml water)
b) = Raw water
c) = Activated carbon filter
inlet outlet
with/without
chlorine
-------
- 680 -
Approximately a thousand times more bacteria per unit volume
can be found on the carbon than in, the water by colony count
determinations.
Since culture methods can never include all the live bacteria,
the living cell count is determined by enzymatic methods.
The total cell count (all living and dead bacteria) was deter-
mined microscopically after enrichment on membrane filters
The colony count determination reflects up to 20% of all
living bacteria and up to 5% of the total cell count, i.e.
one including both living and dead bacteria.
3. Qualitative determination of microorganism populations
3.1. Comparison of the populations in raw water and in the
carbon filter
A comparison of populations with the methods of numerical
taxonomy should make it clear whether the bacterial flora
changes in the course of the water treatment,, while the
colony count remains practically unchanged. The morphological
and biochemical properties of the bacterial strains were deter-
mined and later compared with the aid of a computer.
It was found that:
- the adaptability of the bacteria in the activated carbon
filter is lower than that of the bacteria in the raw
water,
- there are proportionally more bacteria of the Pseudomonas
genus in the activated carbon filter than in the raw
j
water ,
the populations differ clearly in respect of the utili-
zation of substrates, in particular harmful substances.
-------
- 681 -
A. special microorganism population thus develops in the
activated carbon filter, which is different from that of the
raw water, although both populations have almost the same
colony counts.
3.2. The. species composition of a microorganism population
3.2.1.. Bacteria • . . ' ,. • :
A total of 26 species were isolated, belonging to 11 genera
(Table 1). The majority of the bacteria belonged to the
genus Pseudomonas. From the point of view of species, the
genera Bacillus and Azomorias were also well represented.
The bacteria found are 'not'pathogenic and are normally
present in water.
Table 1 Bacterial species in the water of an activated
carbon filter
Pseudomonas alcaligenes
Pseudomonas cepacia
Pseudomonas facilis
Pseudomonas flourescens
Pseudomonas lemoignei
Pseudomonas mendocina
Pseudomonas ruhlandii
Pseudomonas stutzeri
Pseudomonas spec..
Gluconobacter oxidans
Azoiaonas agilis
Azomonas insignis
Azomonas macrocytogenes
Chromobacterium violaceum
Neisseria sicca
Acinetobacter calcoaceticum
Micrococcus luteus
Staphylococcus saprophyticus
Bacillus cereus
Bacillus circulans
Bacillus 'iicheniformis
Bacillus megaterium
Bacillus pumulis
Bacillus thuringensis
Corynebacterium spec.
Micromonospora spec.
-------
- 682 -
3.2.2. Fungi
Moulds and yeasts are found seldom and irregularly in the
activated carbon filtrate. They therefore play a'"sub-
> r • , ,
ordinate role in the treatment of water.
Table 2 gives a summary of the moulds and yeasts which are..
found in the water of the activated carbon filters. The fungi
present are non-pathogenic.
Table 2 Fungal species in the water of an
activated carbon filter
Phialophora hoffmannii
Phialophora mutabilis
Taphrina spec.
Rhodotorula minuta var. texensis
Cryptococcus uniguttulatus
Candida guilliermondii'var. guilliermondii
Hansenula anomala var. anomala
3.3. Bacterial capacity
Activated carbon adsorbs organic substances that can serve as
a substrate for the bacteria, but it also adsorbs bacteria,
this adsorption following Freundlich's isotherm (Fig. 2).
Because of the large difference in s-izes, the bacteria and
the substrate are separated spatially by the pore structure
of the activated carbon during the adsorption. This has a
secondarily adverse effect on the bacterial metabolism. In
the absence of activated carbon the bacteria and the substrate
are distributed uniformly or enriched at the same points.
-------
- 683 -
Fig. 3 shows the metabolic activity of the bacteria as a
function of time. The oxygen consumption serves as a measure
of metabolic activity. Without activated carbon the degra-
dation of the substrate proceeds essentially more rapidly.
Loading
(logarithmic scale)
[log (colony count/g activated carbon])
10 —
c __,
HhJ-
.JQO Loading
80 -
60 -
40 -
20 -
•10
10
Adsorbed concentration
(log [colony count/200 ml])
o — o Q=32-c"~ a o 0=10-c3--a
Fig. 2 Studies on the adsorption of bacteria on'activated
carbon. Isotherms
-------
- 684--
SAUERSTQFFVERBRAUCH
In*) .
a)
I
,-p©
10
c)
20
30
d)
50
e)
80
BAKTERIEN
1.2 10*/ml
1
2
S.UBSTRAT
PHENOL Ot1g/l
KOHLE
0,5g 1 -1.25mm
ZEIT (STUNDEN)
b)
Pig. 3
Studies on the effect of activated carbon on the
metabolic activity of bacteria
Utilization of phenol
Key:
a)
b)
c)
d)
= Oxygen consumption (ml)
= Time (h)
= Bacteria
e) =
Substrate
Phenol, 0.1 g/1
Carbon,
0.5 g, 1-1.25 mm
Activated carbon also has a beneficial effect on the meta-
bolism of the bacteria-. It enriches organic substances and
increases their residence time in the filter; a buffering
of the system also occurs when toxic substances are present,
-------
- 685" -
Fig, 4 represents the outcome of degradation experiments
with four different concentrations of phenol.
15-
10-
5-
SAUERSTOFFVERBRAUCH
1 a)
i—pi
50
r—j—r
150
c)
200
ZEIT (STUNDENi
b)
BAKTEREN
35 10s /ml
SUBSTRAT PHENOL KOHLE 0.5g
.2.5 a/I i...l.g/i 0.3g/l |0,1 g/l 1 -1.25 mm
Fig. 4 Studies on the effect of. activated carbon on the
metabolic activity of bacteria
Utilization of various phenol concentrations.
Key: a) = Oxygen consumption (ml)
b) = Time (h)
c) = Bacteria
d) = Substrate
Phenol, o.l g/1
e) = Carbon 0,5 g,1-1.25 mm
-------
- 686 r , ,
At high phenol concentrations no degradation was found in the
absence of activated carbon, due to the toxicity of this
substance. The following was found when activated carbon was
present: the carbon adsorbs the phenol, renders it "harmless",
and releases it continuously in the manner of a slow-flowing
carbon source in non—toxic concentrations.
3.3.1. Contribution of the bacteria to the water treatment
The following values apply to the drinking water treatment of.
Rhine water with high chlorination before the activated carbon
filtration. At the time of the studies the activated carbon
filter had an efficiency of about 80%. The bacteria partici-
pated as follows in this treatment:
- reduction of the amount of dissolved organic
substances: 5%,
- reduction of the amount of readily degradable
organic substances (BOD-): about 70%,
- reduction of the amount of difficultly degradable
organic substances (BOD^r*): about 17%
- oxygen consumption: about 60%,
— carbon dioxide production: about 60%.
It should be noted that during the subsequent inevitable
decrease in the adsorption capacity of the activated carbon
the bacterial fraction increases strongly in activity. The
already low degradability of the organic substances in this
water is additionally reduced by the high chlorination. The
high degree of the decrease of readily degradable organic
substances is significant as regards the repopulation with
bacteria. These substances, causing the repopulation, are
partly removed by the biologically active activated carbon
filter and thus the tendency' towards bacterial population of
water in the supply network is reduced.
-------
- 687 -
In addition, it must be mentioned that the bacterial activity
effects a continuous partial regeneration of the carbon and
so prolongs its running time.
4. Discussion and outlook
The occurrence of microorganisms and their proliferation in
the activated carbon filter was felt to be undesirable in the
past - and sometimes still today - without any exact know-'
ledge of the microbiological relationships and their signifi-
cance.
Accordingly, processes for disinfection of the activated
carbon filters were developed, which proved to be of no use
on a large industrial scale.
Bacteria on the activated carbon should not be combatted. On
the contrary, their activity should be promoted, i.e. their
contribution, manifested ultimately in the conversion and
mineralization of organic substances, should be optimized by
suitable measures. An example of this is offered by experi-
ments in which high chlorination - which among its other
adverse effects impairs the degradability of the water consti-
tuents - is replaced e.g. by ozonization to keep the readily
degradable compounds in the water.
-------
-' 688 -
(1) KLOTZ, M., WERNER, P., SCHWEISFURTH,.R.
Mikrobiologische Untersuchungen der Aktivkohle-
filtration zur Trinkwasseraufbereitung
Forschung und Entwicklung in der Wasserwerkspraxis
Wissenschaftl. Ber.ichte iiber Untersuchungen und
Planungen der Stadtwerke Wiesbaden AG 3_ (1976),
75-82
(2) KLOTZ, M,, WERNER, P., SCHWEISFURTH, R.
Untersuchungen zur Mikrobiologie der Aktiv-
kohlefilter
7. Vortragsreihe mit Erfahrungsaustausch iiber
spez. Fragen der Wassertechnologie
Veroffentl. des Bereichs und des Lehrstuhls f.
Wasserchemie der Universitat Karlsruhe (1975)
9, 27O-282; English translation; EPA 6OO/9-76-O3O
Dec. (1976)
-------
- 689 -
PROCESSES DURING BIOLOGICAL OXIDATION IN FILTERS
D. van der Kooij
Introduction • ; •
Biological processes, which result in the oxidation
of organic and some inorganic compounds as well as
in the removal of bacteria of hygienic significance
from water have always been important in drinking
water preparation. However, the increased pollution
of water resources by pathogenic micro-organisms and
non-biodegradable compounds forced the waterworks to
extend their water treatment systems by addition of
physicochemical processes including oxidation, ad-
sorption and disinfection. These techniques may have
direct effects on biological and physicochemical pro-
cesses, therefore selection and sequence of both bio-
logical and physicochemical processes are of major
importance to obtain optimal treatment efficiency.
Some recently introduced combinations of physioche-
mical and biological processes are: ozonation followed
by filtration and the use of granular activated carbon
(GAC) filters in which adsorption and biological oxi-
dation occur next to each other. This paper focusses
on some interactions between adsorption and biological
activity in GAC-filters applied to prepare drinking
water.
GAC-filtration proved to be very useful for the remo-
val of dissolved toxic or taste and odour affecting
organic substances originating from domestic and in-
dustrial water pollution. However, in GAC-filtrates
-------
- 690 -
frequently increased colony counts have been observed. .
(1/2,3,4,5'). This, increase of colony counts is in con-
trast with the reduction of bacterial numbers as usu-
ally observed in slow sand filters. Today it seems
widely accepted that the observed growth of micro-
organisms ,in GAC-columns .is enhanced by substrate en-
richment in the filter bed resulting from the adsorp-
tion of organic compounds by the carbon (6,7,8,5,9,
10,11).
Moreover, some investigators concluded that an in-
creased contact-time between organisms and substrate
(adsorbed) is allowing the organisms to adapt to the
less readily biodegradable organic substances. The
consumption of adsorbed compounds from the carbon has
been called "biological regeneration" of the carbon (12)
The application of GAC-filters in drinking water pre-
paration in the Netherlands made it necessary to evalu-
ate the microbiological phenomena occurring in these
filters. Results of some experiments performed earlier
by the author (13) revealed that die-off rates of some
types of micro-organisms on GAG and granular-non-acti-
vated carbon from filters supplied with river, water
during one year did not differ. These results sugges-
ted that -the micro-organisms were unable to utilize
organic substances adsorbed on AC.
Experiments
The influence of different filter materials on the
presence of micro-organisms in filter beds and fil-
trates was investigated using small experimental fil-
ter-columns containing GAG (Norit ROW 0.85), a non-
activated carbon (GNAC) with granules of similar di-
mensions and sand (diameter O.8-1.O mm). These filter-
-------
- 691 -
columns (diameter 6 cm) were supplied with tap water
(DOC: 3 ppm/ water temperature 14-18°C) without chlo-^
rine. The flow rate was 3.5 m/h and the apparent con-
tact time was 3 minutes. During a one-year period,
filter materials/ filtrates and the influent were sam-
pled and colony counts (c.f.u./ml) were estimated.
For this purpose the surface spread technique on
8-fold diluted Lab-Lemco broth (Oxord CM 15.) agar
plates was used. The samples of filter materials were
treated ultrasonically for 3 minutes in the dilution
water to detach micro-organisms. The plates were in-
cubated at 25°C for 1O days.
Moreover/ the removal of organic substances was mea-
sured by ultra-violet light absorbance at 275 nm in
5 cm cuvettes.
Results and discussion
The colonization of the filter: materials and the co-
lony counts of the tap water and the filtrates are
presented in Figure 1. Maximum colony counts on the
filter materials and in the filtrates were reached
after a filtration period of 20 to 30 days. The ultra-
violet absorbences revealed that both the GNAC filter
and the Sand filter had an immediate total break-
through of UV-absorbing compounds. The GAC-filter
reached a breakthrough level of 8O % after 3O days.
After 5O days only 1O % of the UV-absorbing compounds
were removed by the AC-fliter.
A comparison of the colony counts observed on the
filter materials is presented in Figure 2. This fi-
gure reveals that the colony counts of GAC were usu-
ally larger than those of GNAC and Sand and reached
-------
- 692 -
A AC *—tfiltermaterial A---A water (effluent)
A NAC A—A . A—A
o Sand o—o . o—o
—• dnnkmgwater (influent)
2O 4O 6O BO, 1OO 120 14O 16O 18O 2OO 220 24O 26O 28O 3OO 32O 34
Fig. 1 Colony counts of the influent, the filter materials
and filtrates during a period of 34O days
a maximum level of about 7x1O c.f.u./ml. The colony
counts of GNAC and sand were similar to each other.
The colony counts of the filtrates did not differ
from each other (cf. Figure 3).
From the observed similarities in microbial behaviour
in the presence of the three different filter materi-
als it is concluded that adsorption of organic com-
pounds by the activated carbon is not the cause of
the high colony counts as usually observed in GAG-fil-
ters . This conclusion is supported by the observation
that the majority of micro-organisms isolated from
GAG was only able to grow on simple non-adsorbing
compounds like acetate, pyruvate and lactate, whereas
adsorbing substances like aromatic compounds were not
utilized.
-------
- 693 -
10
10
co
L_
o>
•*-•
CO
E
e 6
•"-. 10
3
H-^
cj
10
0 20 40 60 80
—*" cum % of samples
100
Fig. 2 Comparison of the colony counts of the filter
materials GAG, GNAC and sand
Observations on sand samples from the effluent side
of the slow sand filters of -the Hague (flow rate :
O.3 m/hr) revealed that the bacterial content of
4
this sand was about 2-3x1O -c.f.u, per ml, whereas
colony counts in the filtrate usually were below
10O c.f.u. per ml. Similar observations were repor-
ted by Schmidt (14). Comparison of these colony
counts with those observed in the experiment (cf.
Figures 2 and 3) suggested that a relationship may
exist between the flow.-rate-.and the number of- micro-
-------
- 694 -
c
CD
3
-------
- 695 -
c f u /tnl GAC
10'
10*
105
O
20
r i
40
60
cum. % of samples
80
100
gig. 4 The colony counts of GAC present on the slow
sand filters of The Hague
The results presented in Figure 2 revealed that the
colony counts of GAC were usually larger than those
of GNAC and Sand although the colony counts in the
filtrates did not differ. This phenomenon is probably
due to the relatively large surface area per. unit vo-
lume of the GAC, on which micro-organisms utilizing
substrates from the passing water can attach. For this
reason GAC may be a favourable material for biological
filtration processes.
-------
- 696
Effect of micro-organisms on adsorption
In the previous part of this paper it is showji that
colony counts on GAG, in filter beds may reach a maxi-
mum level of about 7x1O c.f.u. per ml GAG (i.e.
P
2'. 1O c.f.u, /gram GAG). Large numbers of micro-orga-
nisms on GAG in filters were also observed by Klotz
et al. (3). These large numbers of micro-organisms on
the GAG might hinder the, transport of organic com-
pounds to the adsorbing surface of the carbon. Some
laboratory experiments were conducted to investigate
the influence of bacterial ce,ll attached on GAG on the
adsorption of some compounds.. In these experiments,
adsorption isotherms and .adsorption rates of 4-nitro-
phenol and 4-hydroxybenzoate on GAG (Norit ROW 0.-8
Supra) were estimated in the presence and in the ab-
sence of a large number of bacteria on the carbon.
The used bacteria were a P s eudomonas fluorescens
strain and a Pseudomonas alcaligenes strain. The
fluorescent pseudomonad could grow on 4-hydroxyben-
zoate but did not utilize 4-nitrophenol, whereas
P. alcaligenes could not use either of these com-
pounds. Portions of O.1 gram GAG were sterilized in
bottles containing 2OO ml demineralized water with
KH2PO4 2.7 mg/1; K2HPO4-3H20 .5.3. mg/1 and Na2HP04.
12H2O 8.O mg/1. Final pH after .sterilization: 7.3,
In order to cultivate bacteria on . the carbon ammo-
nium-acetate was added to the sterilized bottles
from a separately sterilized solution in a final con-
centration of 10 mg acetate-carbon per liter. Further,
the bottles were inoceulated with either the fluore-
scent pseudomonad or the P^^alca^lig^enes, isolate and
incubated for 3 days at 25° _in a rotary shaker
(12O rev/min.). The bacteria developed to the maxi-
mum level of 4x1O c.f.u. per ml of medium and the
"A R
GAG contained about 2.1O c.f.u. per ml (6x1O /gr) of"
-------
- 697 -r-. -
carbon. In these conditions 4-nitrophenol was added
to the bottles with P. flucrescens and 4-hydroxyben-
zoate was added to the bottles with P.alcaligenes
from sterilized solutions in final concentrations
of 1OO mg/1; 5O mg/1;. 25 mg/1 and 10 mg/1. These com-
pounds were also added to bottles containing steri-
lized 'GAG without bacteria. The bottles were re-
placed in the rotary incubator -at 25°C and the con-
centrations of either 4-hydr<5xybenzoate or 4-nitro-
phenol were measured by UV-absorbance (at 245 and
269 nm respectively) in membrane filtered samples of
5 ml after 24,48 and 144 hours of incubation. UV-
absorbances were measured every-hour during the first
eight hours after addition in bottles with an initial
adsorbate concentration of 1OO'mg/l. All experiments
were performed in duplicate.
Results and discussion
With both compounds, the adsorption-equilibrium was
reached within 48-144 hours. The adsorption-isotherms
of 4-nitrophenol and 4-hydroxybenzoate on the carbon in
the presence and _in the abscence of bacteria were cal-
culated from measured concentrations and are presented
in Figure 5. The disappearance of the adsorbates from
the solution initially containing 1OO mg of adsorbate
per liter is shown in Figures 6 and 7. Figures 5, 6
and 7 show that 4-nitrophenol 'is better adsorbed than
4-hydroxybenzoate. The results presented in Figure 5
reveal that the adsorption isotherms were not affected
by the described presence -of -:the bacteria. Moreover,
the adsorption of the adsorbates in the 100 ppm-bottles
was not affected by the presence" df bacteria on the
carbon. The adsorption rate of- 4-hydroxybenzoate in
-------
- 698 -
A AC without bactenx
J AC with
02 04 06 08 tO 12 1.4 16 18
-02
Fig. 5 The adsorption isotherms (144 hours) of 4-nitrophenol
and 4-hydroxybenzoate on GAC (Norit ROW 0.8 Supra)
in the presence and in the absence of bacteria on the
carbon
30
so
10
&) AC witliuul bk«m
I) AC with bjcleiu
Fig. 6
The disappearance of
4-nitrophenol and
4-hydroxybenzoate from a
solution as a result of
adsorption on GAC with
and without bacteria
TinwChoura)
-------
- 699 -
£) AC without tucteru
• AC with bacteria
Fig. 7
the presence of bacteria could not be calculated be-
cause both bottles became infected by adsorbate-consu-
ming micro-organisms. This compound is, in contrast
to 4-nitrophenol,easily biodegradable.
The experiments described in this paper showed that
the number of micro-organisms on a filter material
depends on the flow rate of the water through the fil-
ter bed. This phenomenon is probably due to limita-
tion of substrate transport to the micro-organisms.
Moreover, it is shown that adsorption of compounds
was not inhibited by the presence of a large number
of bacteria on the carbon. GAG in filter beds sampled
by the author always had lower colony counts than the
level applied in the experiments described in this
paper. Therefore it is concluded that in most GAG—fil-
ters used to prepare drinking water, hinderance to
-------
- 7OO -
the adsorption of organic compounds by micro-organisms
is very unlikely. However, it.still may occur when the
water is containing relatively large amounts of bio-
degradable compounds in which situations extremely - .
large numbers of micro-organisms develop on the carbon.
Moreover, adsorption may be affected by the pollution
of GAC by colloidal and suspended matter. These effects
were not investigated.
(1) FORD, D.B.
The use of granular carbon filtration for taste and
odour control
Proc. Wat. Res. Ass. Conf. Reading (1973), 263-278
(2) KNOPPERT, P.L., ROOK, J.J.
Treatment of river Rhine water with activated carbon
Proc. Wat. Res. Ass. Conf. Reading (1973), 1O9-125
(3) KLOTZ, M., WERNER, P., SCHWEISFURTH, R.
Investigations concerning the microbiology of activated
carbon f i1te rs
Trans. Rep. Special Problems Water Techno1. EPA-6OO
19-76-030 (1976), 312-33O
(4) MELBOURNE, J.D., MILLER, D.G.
The treatment of river Trent water using granular activated
carbon beds
Proc. Wat. Res. Ass. Conf. Reading (1973), 73-1O8
(5) WALLIS, G., STAGG, C.H., MELNICK, J.L.
The hazards of incorporating charcoal filters into
domestic water systems
Water Research 8_ (1974), 111-113
(6) WEBER, W.J., FRIEDMAN, L.D., BLOOM, R.
Biologically extended physicochemical treatment
Adv. Wat, Poll. Res., Proc. 6th Internat. Conf.
Jerusalem (1972), 641-656
(7) KOLLE, W., SONTHEIMER, H.
Experience with activated carbon in West Germany
Proc. Wat. Res. Ass. Conf. Reading (1973), 347-367
(8) LOVE, O.T., ROBECK, G.G., SYMONS, J.M., BUELOW, R.W.
Proc. Wat. Res. Ass. Conf. Reading (1973), 279-312
-------
- 701
(9) EBERHARDT, M.
Untersuehungen zur optimalen Kombination von Adsorption,
Filtration und biologischer Reinigung
Publication Wasserchem., Engler-Bunte-Inst., Univ.
Karlsruhe _5 (1971) , 358-279
(1O) GUIRGUIS, W., COOPER, T., HARRIS, J./UNGAR, A.
Improved performance of activated carbon by pre-
ozonation
J. Wat. Poll. Contr. Fed. 5O (1978), 3O8-32O
(11) SYMONS, J.M.
Interim treatment guide for controlling organic contami-
nants in drinking water using granular activated carbon
US EPA, Cincinnati (1978)
(12) EBERHARDT, M.
Experience with the use of biologically effective acti-
vated carbon
Trans. Rep. Special Problems Wat. Technol.
EPA-6OO/9-76-03O (1976). 331-347
(13) Van der KOOIJ, D.
Some investigations into the presence and behaviour
of bacteria in activated carbon filters
Transl. Rep. Special Problems Wat. Technol. EPA-6OO/9-
76-030 (1976) , 348-354
(14) SCHMIDT, K.
Die Abbauleistungen der Bakterienflora bei der Langsam-
sandfiltration und ihre Beeinflussung durch die Rohwasser-
qualitat und andere Urawelteinflilsse
Publication Hydrol. Fcrsch. Dortmund, Dissertationsschr.
(1963) , 5
-------
- 702 -
PROCESS ENGINEERING ASPECTS IN THE COMBINATION OF CHEMICAL
AND BIOLOGICAL OXIDATION
H. Sontheimer
Biological processes for removing dissolved organic substances
from water can only be applied successfully in the treatment
of drinking water if the substances to be removed are bio-
logically degradable in the time set by the selected process
and under the prevailing ecological conditions. Furthermore,
the water must not contain materials that are harmful to
microorganisms. As a result of these restrictions, but some-
times also for purely technical reasons, other purification
measures must first be introduced before the biological treat-
ment process can be run reliably and effectively. ;
Among the most important of these are the processes that
allow a far-reaching removal of solids and colloids, i.e.
flocculation, sedimentation, and filtration. These are per-
formed mainly to relieve the loading on the filter surfaces
in the biological filters and. to prevent them from blocking
up. Measures of this kind are especially .necessary where
the access to the infiltration surfaces for the purposes of
cleaning is limited. For this reason particularly high
demands are made in the case of underground seepage galleries
and absorption wells. A preliminary treatment is here almost
invariably necessary.
If a flocculating /agent is added in such treatment processes
preceding t$.he biological purification, in many waters an
''i '
additional removal of the dissolved organic substances takes
place, so genuinely, relieving the biological stage. Fig. 1
shows, in the light of the results obtained at Gelsenwasser
AG, that a preliminary purification of this kind also has a
beneficial effect on the water quality after the soil pass-
age.
-------
- 7O3 -
Frequency suras in %
— ui ih '~t in ID
M eg ca ra ea - e S
1 I 111! 1 1
Ground water from 7
flocculated infil- /
trate f
*a 13
o o ,a
/Ground / /
^ater from /
o _.' w' /a
/ unf locou- /
ul a ted J «/
infUtPPte /
Cl/ , O
o/" X°
~V
£9 t3 Cl y/ O
n/ / Q/Ruhr water/ Ruhr water
T f /"after flo-/ before floccu-
/ / /cculation lation
2 3 4 5 6 7 8 9 10 .13
UV-absorption coefficient (254 ran) m
Fig. 1 Improvement of the quality of ground water by_
infiltration of flocculated water at the Witten
waterworks
This figure shows the frequency distributions for organic
substances determined by W extinction before and after
flocculation. The results indicate that the decrease in the
organic loading caused by precipitation with the flocculating
agent still persists after the soil"passage.
A very similar result as regards the final outcome can be
achieved with a process of a very different kind, namely with
the aid of chemical'oxidation upstream of the biological
purification stage. Although chemical oxidation on its own
is generally insufficient to reduce the total concentration
of dissolved organic carbon, but only the chemical oxygen
demand, this effect shows, just like the clear reduction of
the UV extinction when ozone is used, that when a chemical
pre-oxidation of this kind is performed a conversion of the
-------
- 7O4 -
organic water constituents takes place. In addition to the
reduction of the mean molecular weight already mentioned in
other papers, particularly with ozonization, this pre-
oxidation also causes a rise in polarity by increasing the
proportion of the carboxyl groups and thereby a simultaneous
partial conversion of biologically resistant organic sub-
stances into biologically degradable ones (1,2).
An example of this effect, which is representative for many
similar observations, is shown in Fig. 2.
14
12
-10
la
Inflow
Without
ozone
With
ozone
Ozone consumption:3"*"
ni iii i i
Inflow
[Without
ozone
With
orone
8 10 12 14 16 18 20
weeks
Fig. 2 Influence of ozonization on the change in the DOC
and COD in the slow filtration of a biologically
pre-purified waste-water
-------
- 705' -
Here are shown the results for the further purification of a
Berlin waste water that has already been very effectively
biologically purified (3). It can be seen that by the treat-
ment with ozone in combination with biological oxidation in
a sand filter a clear improvement in both the'DOC and the COD
may be attained.
In connection with the theme of the present paper there is a
finding - also shown in Fig. 2 - that is particularly
important. It is that the favourable effects of preliminary
ozonization can be easily detected when only 5.4 mg/1 of
ozone has been used. No further detectable improvements are
obtained by the use of higher doses. 'All the experience
gained so far with very different waters shows that on the
whole only about 0.5 g/g DOC is necessary for such prelimin-
ary oxidizing treatment with ozone, and that the ozone
requirement never exceeds 1 g/g DOC. This applies to crude
water of normal composition, such as is used to obtain
drinking water. ...
When using slow filters or soil infiltration large surfaces
or suitable underground conditions are normally necessary.
Since these are not always available, much time has been
spent on the possibility of performing biological purifi-
cations in filters as well, and in this content biologically
and adsorptively working activated carbon filters have
aquired particular significance in recent years. -The bio-
logical processes that occur here have long been observed
but, owing to the high bacterial counts found in many cases,
they have often been regarded as merely disadvantageous.
The increased interest in chemical and biological oxidation
with the use of activated carbon filters in recent years is
partly due to the results of studies carried out initially
at the municipal works in Bremen (4,5) and of subsequent
-------
- 706 -'
studies both at an experimental plant and at a main'plant 'in
Mulheim (6). Details of the experiments and experience in
Mulheim will be given in the next lecture. Right now, how-
ever, using fairly simple arguments and model considerations,
we shall attempt to arrive at some conclusions about what
can be achieved with a process of this kind, about its limit-
ations, and how the observations made at various plants can
be explained.
c
Co
t
t
Fig.
25 30
t in weeks—••
10
15
20 25 30
t in weeks —»
_3 Filter breakthrough curves with (-«-•-) and
without ( ) biological reactivation
We start out from the premise that for a given water and a
given activated carbon filter the loading will increase
steadily with time in the case of pure adsorption, as can be
seen from curve 1 in the lower half of the figure. If the
organic substances are removed by adsorption processes alone,
then, depending on the nature of the organic materials and
on the quality of the carbon, to each given loading in the
-------
- 707
filter 'there will correspond a certain outflow concentration
or a certain filter efficiency. The higher the loading, the.
greater is the outflow concentration, as can be seen on the
example of curve 1 in the top half of the figure.
It has been shown repeatedly that biological processes can
lead to a reduction of the loading, since by means of these
the adsorbed materials are then.oxidized to CO- and H20. If
a certain biological activity is assumed in the filter, we
obtain curve 2 for the time variation of the loading in
curve 1, i.e. with a biologically operating filter the load-
ing increases more slowly than with pure activated carbon.
On the strongly simplifying assumption that the relationship
between the loading and -the filter efficiency remains cons-
tant, the breakthrough curve 2 can be easily obtained from
the two loading curves and the filter activity curve 1, as
is shown here as an example, if an additional biological
regeneration of the activated carbon takes place. The
better this regeneration, the,longer is the service life of
the filter until a certain residual concentration has been
reached. .In the simple example of a filter with a short
residence time shown in Fig. -3 the running time is prolonged
by about 50%. In practice the corresponding values are
usually considerably higher. A prolongation of the running
time by a factor of 3 - 5 is frequently observed, this
factor being decisively dependent on the residence time of
the water in the filter and on the proportion of biologically
degradable substances.
As shown in the next figure, it is also of significance for
prolonging the running time whether an appreciable part of
the adsorbed substances is biologically degradable. In the
case of a high proportion of resistant materials these con-
centrate on the carbon in the course of time. The biological
-------
- 708 -
20 25 30
t in weeks —^
without biological
reactivation
small proportion of
difficultly degra-
dable substances
large prooortinn of
difficultly degra-
tdable
IS
20 25 30
t in weeks—*
Pig. 4 Breakthrough and loading courses for biological
— activated carbon reactivation
activity must therefore gradually decrease, since smaller
amounts of degradable materials are present on the carbon,
giving rise to a considerably earlier breakthrough
represented by curve 3 - in place of the improved break-
through curve 2.
Among other reasons, this loading with biologically resistant
substances becomes evident from the fact that the oxygen con-
sumption for the biological; oxidation decreases with increas-
ing loading.
This is shown in the following figure, based on results
obtained by Engels at the Dlisseldorf municipal works.
-------
- 7O9 -
o
Cr>
220
200
180
160
140
120
100
80
, 60
40
20
0
2000 40OO 6000 8000 10000 12000 1400O
Water throughput in m3/m3 GAC
Fig. 5 Oxygen consumption ,in -GAG. filters, as a function
of the water throughput
This unfavourable effect of the proportion of difficultly
degradable substances explains why much time and effort is
spent on reducing this proportion by treatment with ozone.
A preliminary oxidation of this -kind is also carried out in
Dusseldorf, but it is apparently- insufficient to prevent the
decreased oxygen consumption.
An example of the effects that can in principle be expected
from a preliminary oxidation in the breakthrough curves is
shown in Fig. 6. , , - , •
According to this, when a small amount of ozone has been
added a clearly more favourable breakthrough curve is obtained,
since a part of the resistant organic material becomes bio-
logically degradable as a result of ozonization, so that the
previously mentioned enrichment of the difficultly degradable
materials on carbon does not occur to the same extent as in
the absence of such preliminary treatment.
-------
- 710 -
3mg/l03
\6mg/I 0
Fig. 6 Example of the changes in filter breakthrough curves
with different ozone doses
Nevertheless, ozone oxidation, particularly when carried out
with excessive ozone doses, can also modify the adsorb-
ability of the organic substances. Thus, the more strongly
polar substances produced by ozonization are often adsorbed
less well. In extreme cases, as can be seen from the curve
in Fig. 5 for an addition of 6 mg 03/1, this effect can
cause the initial breakthrough to occur at the same time,or
under unfavourable conditions even earlier, than is the case
when no ozone has been added at all. Only with very long
running times and higher residual concentrations does this
method become advantageous again with high ozone doses.
Readily degradable organic compounds are often produced, when
high doses are used in this way, such as acetic and oxalic
acid (7). These are immediately biologically oxidized on
the outer surface of the•activated carbon and are thereby
rapidly and extensively removed. Under such conditions the
breakthrough curves obtained basically follow the curve for
-------
-71.1 -
1 mg/1 of ozone, and later pass into one of the. curv.es for a
higher ozone dose. Although the filter activity is enhanced
under such conditions for all ozone doses, excessive doses
of ozone are undesirable from the economic point of view.
This model explanation and clarification of the possible
relationships shows why in practice very different effects
are observed in dependence on toe water quality, the filter
dimensions, the properties of the carbon, and the amount of
ozone used. For this reason the discussed combination of
chemical and biological oxidation is not a treatment that
can be performed universally arid on all "waters with equal
effect and always following the same procedures. Moreover,
since the possible dependences can seldom be obtained from
simple laboratory tests, the performance of semi-technical
trials is nearly always necessary.
If a summary of the essential criteria that 'can be deduced^
from the existing experience is nevertheless to be attempted,
despite the above' mentioned complexity of the situation in
"this oxidation process, 'the values given in the following
table are obtained.
TABLE 1; Process parameters for biologically adsorptive
treatment of drinking water . ;•:,••.-
Ozone dose
Biological oxidation
b~ demand
(for DOC oxidation)
Residence time in filter
presumed empty
O..5 - 1 .O g O3/g DOC
3
100 g -DOC/m"
20O g' 02/m3
15-30 min
d
-------
- 712 -
Accordingly, 0.5 - 1 g of ozone is generally required per g
DOC. Larger amounts are almost invariably disadvantageous if
a different effect is to be avoided with the ozone. In
nearly all of the filters studied so far an increase in the
amount of CO^ was found in the filter outflow under these
conditions, corresponding to a DOC of 100 g per m of filter
volume and per day. As a rule, 200 g of oxygen per m of
the filter volume per day are necessary for this oxidation
of organic substances. If NH, ox'idation is taking place in
the carbon filter at the same time the oxygen consumption
becomes considerably higher, but the C02 production is there-
by usually only slightly influenced.
The length of the residence time in the filter normally
determines the time between two regenerations. If it is so
long that, calculated on the carbon, the incoming organic
substance corresponds in amount to the biologically oxidized
substance, a biological filter can theoretically remain in
use indefinitely, needing only a back-wash from time to time.
However, in practice the effectiveness of biological purifi-
cation nearly always deteriorates as a result of the accumu-
lation of resistant materials on the carbon surface. Toxic
water constituents and their accumulation can also have an
adverse effect. For this reason the running time of bio-
logical activated carbon filters is nearly always only some
5-10 times as long as that in the case of pure adsorption.
The prolongation factor decreases with decreasing residence
time.
Since the amount of CO^ produced can be measured just as
easily as the DOC decrease and the oxygen consumption, the
most expedient operating parameters can be obtained for
every individual case on the basis of the results with an
experimental filter, the establishment of the optimal ozone
-------
- 713 -
dose being often the main difficulty. In an evaluation of
the experimental results from biological activated carbon
filters it is therefore advisable to observe and check the
relationships mentioned above.
If, after an ozone oxidation, the measurements and operation
of biologically adsorptive activated carbon filters are
carried out in this manner, the combined•process can prove'
to be a valuable aid in producing unimpeachable, naturally
pure drinking water.
-------
- 714 -
(1) MAIER, D.
Wirkung von O2on auf die gelttsten organischen
Substanzen im Eodenseewasser
Vom Wasser _43_ (1974), 127
(2) KURZ, R.
Untersuchungen zur Wirkung von Ozon auf
Flockungsvorgange
Dissertation, Universitat'Karlsruhe (1977)
(3) SONTIIEIMER, H., W5LFEL, P.
Amelioration de la degradation biologique des eaux
residuaires par un traitement a 1'ozone
Tagungsbericht Internat. Ozon-Kongress, Paris
4-6 Mai (1977)
(4) EBERHARD, M. , MADSEN, S.'r SON-THEIMER, H.
Untersuchungen zur Verv/endung biologisch arbeitender
Aktivkohlefilter bei der Trinkwasseraufbereitung
Engler-Bunte-Institut der Universitat Karlsruhe,
Veroff. des Bereichs und.des Lehrstuhls fur
Wasserchemie (1974), 7
(5) EBERHARD, M.
Erfahrungen bei der Anv/endung biologisch wirksamer
Aktivkohlefilter • •
Engler-Bunte-Institut der Universitat Karlsruhe,
Veroff. des Bereichs und des Lehrstuhls fur
Wasserchemie (1975), 9, 283; English translation:
EPA 60O/9-76-030 (1976)
(6) SONTHEIMER, H., HEILKER, E., JEKEL, M.R., NOLTE, H.,
VOLLMER, F.-H.
The Mulheim Process
J. AIWA (1978) , 393
(7) GILBERT, E.
Reaction of ozone with substituted aromatics and
with their oxidation product^
Proc. of workshop on ozone'and chlorine dioxide,
Cincinnati, Ohio, Nov. (1976);
-------
-715-
EXPERIENCE WITH BIOLOGICAL ACTIVATED CARBON FILTERS
M. Jekel
Introduction
The results presented here and the experience with the use
of biological activated carbon filters originate from a
common BMFT research project of the Rheinisch^-Westfalische
Wasserwerksgesellschaft (RWW) in Miilheim/Ruhr and of the
Engler-Bunte-Institut in Karlsruhe. The object of the in-
vestigations was to replace the classical process used for
the treatment of water of the river Ruhr - breakpoint .
chlorination, flocculation, and sedimentation, gravel and
activated carbon filtration and sand bank filtration - ,
by a direct process in technical plants, so that the high
degree of chlorination with its disadvantageous effects
on the water quality could be avoided. On the basis of the
experience in waterworks at the lower Rhine (1,2) and of
investigations at the municipal works in Bremen (3) the
solution to the problem was seen in a combination o.f chemical
oxidation with ozone and biological-adsorptive treatment in
activated carbon filters (4).
Results from the pilot plants
Practical experience with this treatment was obtained both at
an experimental plant at Dohne waterworks of the RWW, which
had been running for two years, and at the main Dohne plant
itself, which had been converted to the new process 15 months
ago (5) .
-------
- 716 -
In the pilot plant Ruhr water was treated by flocculation,
sedimentation, ozonization, and rapid filtration, and was
passed through several activated carbon filters, various
types of carbon being tested. In a smaller pilot plant,
working in parallel, highly chlorinated flocculate discharge
from the main plant was similarly treated with ozonization,
rapid filtration, and activated carbon filtration. The results
of this parallel trial are shown in Fig. 1.
without breakooint chlorination
with breakpoint chlorination
8
10
12
14
16
18
20
of water/1 activated carbon
Fig. 1
Influence of breakpoint chlorination on the
effectiveness of water treatment with flocculation
and sedimentation, ozonization, filtration and
activated carbon treatment (2.5 m at 1O m/h)
This figure gives the percentage reduction of the UV ex-
tinction at 254 run, as a measure of the removal of organic
substances in the whole treatment process, against the
throughput in m per litre of activated carbon. The early
breakthrough of the GAG filter with preceding breakpoint
chlorination is a consequence of the high loading of the
activated carbon with organic chlorine compounds, which
hinders the development of an effective biological activity
in the filter. Without breakpoint chlorination, however, an
-------
- 717 -
essentially better water quality was obtained, caused by
the effective microbial processes in the activated carbon
filter. These processes also ensured that, over a period
of about 14 months, the effectiveness of the treatment re-
mained essentially constant at 75 %, i.e. the upper curve
in the illustration can be extended to 4O m /I activated
carbon (5).
The biological oxidations in the GAC filter can be balanced
over the chemical parameters DOC, inorganic carbon, ammonia,
and dissolved oxygen. The following table shows as,an example
the variation of these four parameters in the GAC filters of
the pilot plant for the winter months January to March 1977
with low water temperatures and a relatively high ammonia
loading.
TABLE 1 Biological activity in activated carbon filters
at low temperatures
Activated
carbon
LSS 2,5 m
ROW 2,5m
NK 12 2,5m
F 400 2,5m
BKA 2,5m
LSS 5,0m
ROW 5,0m
-A DOC
ppm
1,1
1,2
1,0
1,3
1,2
1,6
1,7
+ AanC
ppm
1,0
1,1
1,2
1,2
0,9
1,2
1,3
-ANH4
ppm
1/3
1,41
1/5
1/3
1/4
1/6
1/7
-AQ2
ppm
7,2
7,2
7,1
7,1
7J
7,7
7,7
Moan valuos: Jaa-Mdrz 1977
Mean water temn. : 6,8°C
Inflow (gravel 2'6 m9/1 DOC
filtrate): 1,53mg/lNH4
i - .. — — , — . . — .
-------
718 -
In the first 2.5 m of the filters, which operate at 1O m/h,
the amount of DOC removed is in general only a little greater
than the production of inorganic C, i.e. during this period
most of the organic substances removed were biologically
mineralized. At the same time an almost complete nitrifi-
cation of the ammonia took place, which was also responsible
for the high oxygen demand. In the second half of the 5 m
filters, however, the organic material was still predominantly
removed by adsorption, as was detected by the differences in
A DOC and A inorg. C.
It is worth noting the really small differences in the bio-
logical activity of the activated carbons tested, which,
however, showed considerable differences in their adsorption
capacity. Moreover, further observation of the activated
carbon filters at the pilot plant showed that the nitrifi-
cation had virtually no effect on the biological oxidation
of the organic compounds. Furthermore, no significant in-
fluence of temperature on the biological activity could be
observed in the pilot p-lant.
The breakthrough behaviour of biological activated carbon
filters is influenced, among other things, by the adsorbed
biologically resistant substances. Both the removal perfor-
3
mances of the pilot plant filters in g DOC/m activated
carbon per day and the total loadings of the activated
carbons for the first 18 months of operation are given in
the following figure.
In. addition, the removal performance was divided into a bio-
logical oxidation component and an adsorption componentr
calculated from the carbon loading. As shown in Fig. 3, an
average of about 75 % of the organic substances removed is
biologically oxidized, while the remainder, predominantly
resistant compounds, is adsorbed. Owing to the enrichment
-------
- 719 - " '
*
u.
%
•a
$ 175
*>
•H 150
•U >i
o *
10 -o 125
G) fl -
JS g
-u "--100
u
M-l O
0 ° 75
O &
O
i 5 so
.2 i 2s
**-* *Q
M N
*s
-
-
-
_
^-..-um--
—
___
""""""
i—
1
-Adsorption
-biol. Oxidation
§
CJ *» <
W *~ "• ^
W z °
p-^-^^
i
1
^
, ,
1st half-year ! 2nd half-year 3rd half-year
Fig. 2 Performance and loading of biological activated
carbon filters
of these compounds, recognizable by the increasing loading
of the activated carbon, the performance of the filters
clearly falls off, particularly in the third half-year
period. Simultaneously, a clear improvement in the quality
of the raw water occurred, so that the DOC in the influx
was decreased from about 2.5 to 1.8 ppm. This certainly
also intensifies the decrease in performance. The measure-
ment results indicate that the degradable substances are
at first adsorbed and only then biologically mineralized.
The relatively high proportion of about 75 % of biologically
oxidized substances furthermore permits the conclusion that
the operating time of biological activated carbon filters
in the treatment of Ruhr water is prolonged by a factor
of approximately 4 in comparison with pure adsorption.
-------
- 720
Experience with the treatment at the Dohne waterworks
On the basis of the results with the pilot plant, the Dohne
^waterworks were converted to the new treatment process in
the spring of 1977, as shown in the following flow diagram
(4):'
Mixing Tank PUHMOK Gasification Double- Activated
Tank layer Carbon
Ozonized Air Filter Filter
Pumps
1
j
•%
„• /
s
\,^.,_, ...
L
/
^
v
A?-
Pumps
Ozonized
Generator Air
Safety chlorination
Collecting Percolation Withdrawal
Well Well Well
Filter
Fig. 3 Flow diagram of the Dohne works
The Ruhr water is directed into a small mixing chamber .for
preliminary oxidation with about 1 ppm of ozone. Poly-
aluminium chloride is added at the same time for flocculation.
Die preliminary ozonization with an immersion gas tube :results
in clear water after the solids have been removed in a pul-
sator. On average about 2 ppm of ozone are introduced into
the gasification chamber. If necessary, additional pure oxygen
can be mixed in before the filters to cover the consumption in
the subsequent rapid and activated carbon filters. The bed
-------
- 721 -
length of the activated carbon filters was increased from
2 to 4 m to maintain a sufficient residence time at a velocity
of 22 m/h. These were set into operation in November of
last year with virgin activated carbon.
The following table shows how the conversion to the new
treatment affected the quality of the drinking water.
TABLE 2 Effectiveness of classical and new_treatment
at Dohne waterworks, RWW, Mlilheim/Ruhr
New treatment, without
' UV
m
Ruhr
after
after
after
after
f locculation
ozonization
filtration
soil passage
8
4
3
3
2
activated carbon filter (July-Oct . 1 977 )
Ext. DOC NH£ Colony E.coli/
1 mg/1 mg/1 count/ml 100 ml
.6
.9
.2
. 1
.3
4
3
3
2
1
.O O.79 963O
.3 0"._43. 3510
.1 0.43
.7 O.O3 144O
.2 O.O1 19
1550
16
-
1
<1
Classical treatment - - .
(1975)
Ruhr
after
after
after
after
f locculation
ozonization
activ. carbon
soil passage
6
4
4
4
3
.8
.5
.4
.O
.1
4
3
3
3
1
.O - '• • •
.2
.2
.O
.8
The mean values of some parameters in the new treatment
without activated carbon filters are compared here with
the values of the earlier classical process in 1975, when
-------
- 722 -
the quality of the untreated water was similar.'The oxi-
dation effect of the ozone is noticeable, by means of which
resistant material is converted to biologically degradable
substances. The improved purification effect of the soil
passage is also due to this factor. The fact that the rapid
gravel filters are also biologically active is obvious from
the extensive reduction of ammonia and the re-population
of the filtrate with bacteria. The numbers of colonies
decrease in the soil passage to two-digit numbers, since
the degradable substances have been practically completely
removed.
With the inclusion of the enlarged activated carbon filters
a large proportion of the purification effect of the soil
passage was shifted to the plant. The behaviour of the newly
installed large filters is shown in the following figure, in
which the mean values of the removed DOC and the inorganic
carbon formed in the whole activated carbon filtrate are
plotted against the time of operation.
c^
E
M
0
C „. _
•H 0.6
^3
X °
Q
^ 0.2
_
~
-
Dec.
1977
Jan.
'Feb.
T Ainorg. C
1
I
I
, , A IW*
1 I 1 H
March Aoril ' May ' June ' July'
1978
Fig. 4 Behaviour of the newly installed biological
activated carbon filter at Donne waterworks
-------
'-• 72-3 - ^
In the first three months with .the water temperature below
8 C the activated carbon filters showed practically no
biological activity. The A DOC values indicated a really
early breakthrough of the organic substances, since the
carbon was already becoming increasingly loaded. With the
higher concentrations of adsorbed degradable substances
and with rapidly rising temperatures in the spring the
biological oxidation set in, which from May clearly covered
the substance already adsorbed. The activated carbon was
regenerated biologically and from this month on displayed
an improved purification effect.,
Exhaustive investigations on the"bacteriological nature of
the activated carbon 'filtrate were carried out in the large
plant, because this aspect plays an important role in a
direct input into the .-distribution network. Since -four, types
of activated.carbons-with-different adsorption capacities
were used in the large plant, the filtrates showed sometimes
considerable differences inVthe.DOC values. The effect of
this on'the colony'counts determined in-parallel (incubated
for 48 h at 22 C on nutrient gelatine) is shown in the
following figure. ••••."".•.•.
o
(N
(N
~~»
JC
CO
•3-
o
o
140
120
100
80
60
40
20
* X
O Dec,77
a Jan.78
a Feb.78
• March 78
- A April 78
• May/June 78 a
x July 78
I
0.5
1O
1.5
DOC in .mgll
2£>
2.5
3.0
Fig. 5
Influence of the DOC in.the filtrate of biological
activated carbon .filters .on their bacterial
population ' ...•••.
-------
- 724 -
The geometric mean monthly coloriy counts are here correlated
with the corresponding mean DOC values of the individual
carbon filtrates. While in the first month of operation,
December 1977, with still' scarcely any developed biological
activity, no relation between bacterial population and the DOC
can yet be distinguished, in the following months with
increasing biological activity of the active carbon filters
a dependence of the bacterial population on the organic
loading of the filtrate is established. Remarkably, greater
deviations appear in May and June, when a different compos-
ition of the organic substance during the growth of algae was
clearly present. From these results it may be concluded that
a larger proportion of biologically resistant compounds
occurred at this time. • "
From these results it follows that a good adsorbing activated
carbon, which also removes degradable substances more effi-
ciently, produces lower colony counts in the filtrate. The
mean values themselves show predominantly two-digit and three-
digit colony counts. However, if the ozone dosage is increased,
the experience with the Dohne:waterworks indicates that the
bacterial counts become considerably higher.
A natural far-reaching elimination-of bacteria from the
activated carbon filtrate can still be achieved, according
to trials with a rapidly operating slow sand filter, at a
velocity of 2 m/h, the colony counts being then reduced on
average to single-digit values. (6),
With the development of an effective microbiological activity
in the spring and summer of this year, a mass development of
nematodes was observed in the rapid and active carbon filters.
The reasons for this are to be found mainly in the long running
times of the gravel filters, which, because of the very good
-------
- 725 -
quality of the clear water after floeculation, were one week.
On the other hand, for technical reasons the gravel filter
in the Donne plant could be backwashed only with J±ie rela-
tively low speed of 2O m/h. ,
According to running experience, this problem can be solved
by frequent washing of the> filter, so that the running time
should not be more than four days, the reproduction cycle of
nematodes. This applies above all to the time of increased
water temperatures in the summer, when a continuous obser-
vation of the backwash -water is recommended to prevent a
mass development by frequent and vigorous backwashes.
However, the back-flushing speed should not be so high
that the microbiological activity of the filters is impaired,.
Summary • • • • - -
The experience with biologically working activated.carbon
filters in the treatment 'of Ruhr water can be essentially
summarized in the following "points:
1) The high biological activity of the activated carbon
filters prolongs the running•time about fourfold in
comparison with pure adsorption.
2) The bacteriological nature of the activated carbon
filtrate will be satisfactory when a far-reaching- -
removal of biologically degradable substances is
achieved. .,-•,*- • ,
3) Reliable operation of biological filters can be
achieved by means of a suitable backwashing technique
-------
- 726 -
(1) HOPF , W . -...••
Versuche mit Aktivkohlen zur Aufbereitung des Diisseldorfer
Trinkwassers
gwf-Wasser/Abwasser 1O1 (196O), 14, 33O-336
(2) HOPF, W.
Zur Aufbereitung rnit Ozon und AktiVkohle
gwf-Wasser/Abwasser V1_1 (1970), 13, 156-164
(3) EBERHA'RDT, M. , MADSEN, S., SONTHEIMER, H.
Untersuchungen zur Verwendung biologisch arbeitender
AktivkohlefiIter bei der Trinkwasseraufbereitung
Publication Wasserchemie, Engler-Bunte-Institut,
Univ. Karlsruhe 1_ (1976); gwf-Wasser/Abwasser 116
(1975), 6, 245-247
(4) SONTHEIMER, H., HEILKER, E., JEKEL, M.R., NOLTE, H.,
VOLLMER, F.H.
The Mulheim Process
J. AVJWA 10_ (1978) , 393-396
(5) HEILKER, E., JEKEL, M., SONTHEIMER/' H.
Biologisch-adsorptive Trinkwasseraufbereitung in
Aktivkohlefiltern
DVGW-Schriftenreihe Wasser, Nr. 1O1, zum BMFT-DVGW-
Statusseminar "Heue Technologien, Hannover, Jan. 1978
To be published
(6) SCHALEKAMP, M.
Die V7irksamkeit yon schnell betriebenen Landsamsand-
f iltern . .•
Publ. Wasserchemie, Enaleir-Bunte-Institut, Univ. Karlsruhe,
5 (1971), 3, 49-66
-------
- 727 -_
THE USE OF COMBINED CHEMICAL AND BIOLOGICAL OXIDATION
PROCESSES . .
P. Schulhof
Before I talk about- the processes combining 'Chemical" and bio-
logical oxidation I should like to dwell, briefly on what I
would say .about the -.papers given in- the, course of this " , ,
symposium if I were one of the men responsible for the con-
struction or modernization of the drinking-water-treatment <
plants in France.
I would first have remarked on the incredible advances in '•
analytical methods in recent years. I-would, have been- parti"
cularly surprised by the absolutely thorough knowledge
attained within a very, short .time with the aid, of these. t •_
methods on the subject of the oxidation of polluted water;
However, my enthusiasm would have -been somewhat dampened ;:by
the fact that the majority of cases dealt with studies , . •
on isolated products or'-substances, while many of the speakers
have made reference, to the competition situations that can. ,
arise in the course of oxidation'reactions in mixtures. Let us
not forget that raw river water-is-a particularly complex
mixture 1
However, the agreement of the presented results and in parti-
cular of the questions that have been put would have disturbed
me too. My concern would have "'been intensified by the fact
that the sanitary and hygienic significance of all the by-
products mentioned is still far from known. After several
years we are still debating about the simplest of them,
chloroform and, according' to your remarks this represents
only the tip of the iceberg.
-------
- 728 •-
It must be added that,the industry evolves new compounds from
year to year, indeed at least as many as those that have been
mentioned during the last three days.
On the other hand, my concern would have been somewhat
abated by the consideration that the studies are often per-
formed under extreme conditions: very severely polluted
water such as Rhine water or oxidation with very massive
doses as in the United States. The average reality in
moderate France is perhaps less gloomy.
Finally, I would now perhaps understand better the wisdom of
the German standards cited by you, recommending the use of
waters low in man-made substances. • Within the wider framework
of the EEC I would also understand the reasons for the caution
of the guidelines established 'in Brussels for the quality of
the raw water about to undergo treatment.
To return to everyday matters in France, however, the con-
struction of treatment plants must be developed further and
the plants must be run with the water available. The plants
for treating the severely polluted waters in the Paris area
must be modernized. What must be done here?
From these days of study some rules have emerged clearly that
normally should be a matter of plain common sense.
For example, the fact that in the case of a polluted water
the first treatment process to be put into action should not .
be an aggressive oxidation. Therefore I would strive towards
a gentler oxidation, i.e. biological, and one that is as
natural as possible.
I concede that I do not have available a site suitable for a
soil passage. These do exist in France, but by no means
everywhere.
-------
- 729 -
First of all I would investigate whether a storage reservoir
can be installed for the raw water. This biological reactor
hardly mentioned this morning ••- is at once a clarifying tank
and a safety stage, thus offering a whole series of simul-
taneous advantages.
If I had no room for a reservoir of this kind I would look
for a different biological reactor, but which one? This
would pose considerable problems.
Let us continue. A physical and chemical clarification line
is put in. For the pre-oxidat.ion. I would use ozone in a
moderate dose. After allr we have seen that preliminary
ozonization is a logical supplement to flocculation.
Then there is the viricidal ozonization, still indispensable
for the treatment of polluted water. Tts logical supplemen-
tation was described this morning. This involves a second bio-
logical reactor that,followed by a weak final chlorination,
should protect the network.
Perhaps I would have had certain difficulties in designing
the ozonization with a minimal viricidal dose. I have main^
tained that the addition should be continuous, neither too
strong nor too weak and extended over a really long period.
All this is very rough, and in my special case I would defi-
nitely construct a pilot,plant, which would present no undue
difficulties.
On the other hand, I would find it much more difficult to
determine the biological reactor to be installed. For what
is to be done when the problem is not one of modernizing an
old plant that already makes use of slow sand filters or
activated carbon filters?
-------
- 73O -
It seems highly unlikely that when a new plant is to be built
the slow sand filter technology designed almost a century ago
in a completely different context or the activated carbon
filter technology developed 30 years ago for the purposes of
f
dechlorination and the prevention of undesirable taste are
ideal'for a biological reactor.
This is the real problem that I wished to come to with these
lengthy preliminary remarks. (fhe technology of chemical
oxidation has apparently reached the age of maturity. The
most recent of these techniques, ozonization, has grown up
in Prance over the last 15 years. In my view the same cannot
be said of the technology of the biological reactors, which
is still in its infancy.
In the years to come we shall need at least two different
types of such biological reactors, one for raw water and one
for clarified water.
I believe this is the challenge that progress in analysis
places before the technologists. Only when these developments
succeed will the studies presented here make possible spec-
tacular advances in the treatment of water.
The technology of the biological"reactors could, and this is
a thoroughly classical conclusion towards the end of a sym-
posium, become the very likely arid universally useful theme
of another symposium. " '
-------
- 731 -
A. -Bousoulengas (Greece)
1. Most of Greece's small cities and villages use ground
water for drinking purposes. In bigger cities surface
water, or both ground and surface water is used.
2. Since 1958 chlorination has been established by Sani-
tary Regulation as the disinfection means for drinking , •
water supplies. There should always be 0,2 mg/1 free C12
present at any point in the distribution network.
This is checked regularly by taking samples from a nujn- -
ber of points, depending on the size of the network .in
question.
3. Sanitary Regulation of 1968 "On the Quality of Drin-
king .Water" determines the physical, chemical and micro-
biological characteristics that drinking water should, • •
have. It determines ma'ximum allowable concentrations -for,
various inorganic substances (Ag, As, Cd, Pb, ,F etc.) . •
and organic substances as well. ~ . . ;•-•••
Regarding organic substances and their products (NH,,. NO!,
NO2 etc.) there is a recommendation for their guanti-: "
tative determination. ' ' ' ,-•••.-
4. There are no regulations determining the treatment
process. However, the above-mentioned sanitary regula-
tion prohibits the use of'a water source if - it ,is found •
to contain a series of 'substances in concentrations •:
higher than those determined by the regulation. These
substances include sulphates, carbonates and other inor-
ganic substances and organic ones such as detergents , ,
and phenolic compounds. . ..
-------
- 73-2'.V •
5. Research is carried out by various state institutes'"''1
and university laboratories on mainly applied problems
concerning drinking water.
6. If a water source is found suitable for .drinking
use, restrictions are applied by- regulation or decree
prohibiting the discharge of any effluents or the dis-
posal of any materials into the source or into the wa-
terways feeding this source - lake etc.
-------
- 73.3'.-
L. Coin' • (France) -
QUALITY OF RAW WATER:
France is trying to use raw water of adequate quality.
However, as a result of modern developments and changes
in the habits of the consumer in our industrial society -
despite the fierce battle being fought against environmental
pollution - certain correcting measures seem inevitable.
They would be:
- either on a. quality level, generally by choosing
different locations as raw water sources, or by making
new water reserves accessible;
- or on a treatment level, by reconstructing existing
plants and by introducing the most recent adsorption
agents, or by modified oxidation processes.
We are therefore awaiting with interest the realization of.
joint regulations on the quality grade of raw water permitted
to be used for drinking-water treatment.
REGULATIONS ON THE CONTROL OF DISINFECTION
There are two forms of control:
1. performed on-site by the waterworks;
2. spot-checks carried out by the Department of Public Health,
. Dates and guidelines are laid down by the C.S.H.P.F.
(Conseil Superieur.d1Hygiene Publique de France).
-------
- 734 -
s - f *
On account of the special guidelines on maximum concen-
trations of micro-pollutants and similar substances, drawn
up in Brussels but not in force yet, France has modified
the regulations. These modifications have partly been
legalized already/ for instance concerning those regu-
lations which were only slightly changed, e.g. for heavy
metals. Organic micro-pollutants, not mentioned previously,
are now included in the programme of treatment plants in•
densely populated towns.
NECESSITY AND VALUE OF SAFETY CHLORINATION:
This is a controversial question. In smaller residential
areas drinking water is supplied to the consumer without
previous treatment, taking into account water protection
areas and issueing certain restrictions. In densely popu-
lated areas chlorine is added to the water. The real problem
is subsequent chlorination after ozonization. In Paris, for
instance, the excessive chlorine is neutralized again before
distribution. Our ideas on raicrobial control will certainly
have to be revised; Concerning fecal germs, the present
strict control measures should be retained. Improvements
seem necessary, however, in connection with the significance
of the fluctuation of the saprophyt numbers. A chlorine con-
tent of O.O5 ppm to O.I ppm at the most, as it is laid down
in Prance, covers short-period risks,; it should, however, be
increased in times of epidemics.
BREAKPOINT CHLORINATION - yes(or no?
After having been in use for many years, breakpoint chlori-
nation is about to be abolished in France - and with good
reason.
-------
- 735 -
REGULATIONS''ON"METHODS: OF TREATMENT:'^
This is a complicated question. The Offices of Public Health
decide on the type of treatment used for producing drinking
water from raw water. However, this decree does not come into
effect in those cases where the size of the project requires
the decision of the C.S.H.P.F. This point in particular will
be treated in the ruling still under consideration, in complete
agreement with all parties concerned.
RULES FOR APPLICATION AND FUTURE RESEARCH:
All the various regulations have been taken very seriously
in France, especially as concerns the treatment of raw water
Questions concerning the distribution network are governed
at present by hygiene regulations of the Departements. This
presents a serious shortcoming, which will surely be over-
come in the future.
As far as research is concerned, the various methods of
treatment, their development and improvement remain very
important. Epidemology will forcibly include the effects of
the distribution system on the water quality at the con-
sumer ' s tap, and here the responsibility also lies with the
consumer himself, because he uses the water supplied to him
and thereby changes the quality himself. This is a decisive
point, and the quality of the material used for domestic
plumbing is also gaining importance in this connection.
-------
- 736 -
T.A. Dick (United Kingdom)
1. Sources and Treatment of Water
In the United Kingdom, one third of the potable water is
derived from underground supplies, one third from upland
surface supplies (lakes, reservoirs and streams) and one
third from lowland rivers. The proportions vary conside-
rably in different parts of the country and in the drier
eastern areas there is little choice but to draw water
from rivers which may contain a substantial proportion
of waste water, that is, water which has passed through
a sewage treatment plant. Sewage treatment plant effluents
are not chlorinated before being discharged to UK rivers.
Underground water supplies and upland surface water sup-
plies are generally of high quality and require only the
conventional treatments. For surface waters from lowland
rivers it is recommended practice in the UK to store the
water in reservoirs, as a safeguard against fluctuating
river flows, possible accidental contamination and to ob-
tain self-purification benefit. The water then receives
treatment by slow sand filtration, or coagulation and ra-
pid gravity filtration or both. The water then passes to
a closed service reservoir before distribution or may go
directly into supply. Pre-chlorination treatment of the
raw water before or during treatment is fairly common.
With few exceptions, all public water supplies in the
United Kingdom are chlorinated after treatment by conven-
tional methods. As a general policy, the degree of final
chlorination is such that O.1 to O.2 mg/1 of free chlorine
remains at the end of 3O minutes contact.
-------
- -737 -
2. Quality of Water . •', , <. '. •••. LtVi ;; •><•', ,A '•'
There are, at present, no mandatory standards in the UK for-
the quality of water put into supply other than that the
water shal-1 be 'wholesome', which is generally accepted
as meaning "clear, palatable and safe". In practice, the
quality of 'water put into supply follows, the recommenda-
tions of the. 1970 World Health Organisation European Stan-
dards, and for bacteriological quality, the recommendations
given in a UK Government document Report 7.1, "The Bacterio-
logical Examination of Water Supplies". -There are n'o- man-
datory regulations for plate counts for drinking waters in
the UK. However, plate, counts are carried .out by many of
the water suppliers to assess changes in the general qua-
lity of water in.the treatment plant and supply system,
particularly wheri the raw water has been derived from sur-
face sources. The 37 plate count is generally used as
pollution indicator but the total plate .counts at 22. were
of. particular value during the 1976 drought year to" ensure
that there was no in-leakage of. contaminated water into
-distribution mains.
3. Present UK Policy
The United Kingdom is well aware of the work .discussed at
the present Conference but its approach to the establish-
ment of water regulations and policy is at present direc-
ted more to establish whether a health risk exists, rather
than making any fundamental change in treatment policy in-
to unknown areas of risk. The UK is well situated for such
research. It has a wide variety of situations where there
is reuse of water and where the Water Research Centre -and
the 1O Regional Water Authorities have accumulated exten-
sive data on water quality. We have excellent .health data
-------
- 73'8 - "
aecumulated'-'over'-many ye*ars and • we have an intensive .pro- .
gramme 'on micropollutant identification and mutagenic
screening of waters and concentrated water samples. De-
tailed statistical examination of the data is actively in
progress. There is also a national programme for the de-
termination of trihalomethanes in drinking water.
We believe this type of work is of fundamental importance
In establishing whether our policy on reuse of water and
on chlorination is sound. It is encouraging to note, for
example, that the national data for the incidence of death
from cancer of the stomach and the bladder have fallen
significantly between 1970 and 1975 for adults between 45
and 64 years. I should also- add that the UK has another
drinking water problem concerning reduction of lead, pri-
marily in the older houses with lead pipes in soft water
areas, and at present, we regard this as our number one
priority for action.
4. Catchment Control
As a final point, the UK believes that there is consi-
derable advantage in reducing the input to raw water sour-
ces of organic materials rather than trying to remove
them from the water during the treatment process. Some of
the UK Water Authorities are now intensively monitoring
the use of organic materials in factories, to see what
is used in the process and what can be lost from the pro-
cess during manufacture. At the same time, the effluents
leaving the plant and the sewage works are analysed and
also the river itself to establish a 'balance1 on the
'missing' organic material. The amounts can be small but
very significant. Each type of industry is being examined,
and attention will also be given to commercial bodies
-------
•' --739 - ,
(e, gV "laun'drie's, dry-cleaners etc)' and to the--domestic
use of chemicals. The system, known as catchment control,
requires a great deal of patient work but can be very '._
successful in reducing, un.de s.ir able micropollutants from
factory processes, particularly_where less harmful ma-
terials can be used in- the process.
-------
- 740 -
1.." Heinohen .(Finland^
The water supply of the largest towns in Finland is de-
pendent on surface water sources.- In treatment the water
passes through coagulation with aluminium sulphate at a
suitable pH («6), clarification and rapid sand filtra-
tion. Chlorine is the chemical,-used for disinfection and
generally also for the .oxidation,of substances affecting
taste and odor.
The ammonia content in: river waters, used as raw water
supplies and carrying a pollution load of domestic waste-
waters, rises in the winter to such values that break-
point chlorination of, NH, is necessary. Also spring run-
off from fields during the flo.od period causes taste and
odor disturbances that are not: removed by treatment pro-
cess. They are reduced ,by .superchlorination and the use
of powdered activated carbon. Blooms of algae'and micro-
organisms in lake basins especially in the summer often
cause taste and odor disturbances in normally treated
drinking water. Chlorine dioxide has been used to elimi-
nate these and experiments have even been made with KMnO,
instead of, or in addition to chlorine oxidation. The
most effective way of combating algae in watercourses and
at water intakes has proved .to be treatment with copper
sulphate. . ,
The Finnish National Medical Board has given standards
for drinking water; these were last revised in 1971. They
are principally based on the international drinking water
standards of WHO dating from ,1963 with a few points taken
-------
741 -
from the WHO European area standard Of 19"70.^Gft special
interest with re-gard to the conditions prevalent in Fin-
land are the norms for KMnO.-consumption in order to li-
mit organic matter in the water (lower limit 20 mg/1 and
upper limit 40 mg/1), forj-NH., as. an indicator of pollution
(O.2 and O.5 mg/1), as well'as for Al-residue, characte-
rizing the treatment process CO.5 - 1.O mg/1). Of organic
compounds limits have only been set for phenols and an-
ionactive detergents,- bait--not for example for biocides
or any of the general parameters (e.g.. CCE, TOO...-
The bacteriological standards give limit values to fecal
streptococci in addition to .conforms and thermotolerant
coli bacteria* -" ,-•;-.• . .
The most common difficulties in surface water .treatment
are due to the water's fairly high humus content, some-
times also to similar e'ffects -caused by wastewaters from
pulp production in the-.wood-processing industry. The re-
moval of organic matter is incomplete in Al-coagulation
(TOC residue >4 mg/1) and a fair amount of aluminium re-
mains in the water (>0.3 mg/1); the high chlorine demand
and its corresponding use result in the formation of con-
siderable amounts of organic-chlorine compounds obser-
vable both as taste and.odor disturbances as well as the,
incidence of CHC13 (>5O ug/1),:observed in analyses.
Taste and odor problems are intensified'during the - winter
time with break-point chlorination to destroy the ammonia.
Similar problems are experienced sporadically also during
periods of spring flood and warm summers.
The measures used -at the major waterworks in Finland to
solve problems arising in water- treatment are .principally
the following:
-------
- 742 -
1, Improving the quality of the raw water sources in use
Water conservation, dilution with water transferred from
neighboring watercourses, and impoundment before taking
the water into treatment plants- have been the methods
most commonly used. All of the -above-mentioned have been
used especially by 'Helsinki City -Waterworks.
2. Transferring water intake to sources of a better quality
The numerous lakes in Finland form" alternative water supplies
of good quality and abundant"quantity to the polluted river
or lake waters presently in use.- The Helsinki metropolitan
area on the southern coast as well as several communities
north of it are presently engaged in -a- project of construc-
ting a rock tunnel 12O km long from the' southern end of
Lake Paijanne, a large lake basin -in Central Finland. The
tunnel will be completed in 1982. Several other smaller wa-
ter transfer projects have been completed or are in the
planning stage for the water supply of other major Finnish
cities.
3. Improving present treatment methods
The methods mentioned above for the improvement of raw
water quality already aim at diminishing the substances
causing taste and odor as well as other disturbances in
drinking water. The treatment process proper has been im-
proved both by increasing the effectiveness of chemical
treatment (increased use of activated carbon and chlorine
dioxide) and by making changes in the order of the various
phases of the process so that the formation of disturbing
substances has been diminished (e.g. decrease in the for-
mation of chlorinated hydrocarbons by transferring the
main chlorination point to after coagulation and filtra-
tion) .
-------
- 743/r
4. Developing new .treatment methods
The trend is to substitute chlorine oxidation with ozo-
nization. Helsinki and its neighboring municipalities
will begin ozone treatment already in 1979.' Several other
towns are at-present carrying•out study programs on its'
use or are contemplating the starting of such programs in
the near future.
Activated carbon filtration after chlorine oxidation has
been studied with large-scale pilot plant test runs in
Helsinki and its introduction is at the moment studied at •
Turku, one of the largest cities. next to Helsinki, to relieve
local water treatment problems.
The test runs in Helsinki showed as drawbacks for the me-
thod the rapid decrease in - adsorption capacity (during a
run of 3-5 months) showed-by measurements, and a corres-
ponding rapid deterioration as well as large-fluctuations
in taste and odor characteristics. This led to the decision
to abandon this method of quality improvement and to adopt '
ozonization, which had consistently given good results in
the test runs, as an additional method to make the treat-
ment process more effective.' ' ' '•
-------
- 744 -
Victor J. Kimm (U.S.A.')
Legislative control of contagious or infectious disease
first began in the United States with the National Quarantine
Act of 1878. Regulations developed under this Act were inten-
ded primarily to maintain sanitary; drinking water aboard vessels
However, in 1912, due to severe outbreaks of intestional disease
among passengers aboard steamships on the Great Lakes, the
Treasury Department issued rules which required State certi-
fication that drinking water aboard interstate vehicles would
not cause disease. This concept beicame the basis for the 1914
Treasury Standards and the subsequent modifications knows as
the U.S. Public Health Service Drinking Water Standards.
Early in 1968, a series of bills"were introduced in Congress
for the general protection of drinking water. Eventually the
Safe Drinking Water Act of 1974 was passed, and it applies to
all public water systems serving 25 persons or more on a
regular basis, for a total of about 2OO,OOO,OOQ Americans.
This Act mandates the U.S. Envirorimental Protection Agency
to develop primary drinking water !regulations containing maxi-
mum contaminant levels for microbiological, chemical and radio-
logical constituents that affect the public health.
NATIONAL INTERIM PRIMARY DRINKING WATER REGULATIONS
The Interim Regulations establish Maximum Contaminant Levels
(MCL) for coliform bacteria, turbidity,:ten inorganic chemicals,
six pesticides and radiological contaminants (Table 1). As of
June 24, 1977, water utilities are required to conduct periodic
monitoring at a prescribed frequency to ensure compliance with
-------
- 745 - ,
TABLE T The maximum contaminant levels for constituents
in the National Interim Primary Regulations
Constituent
Level
mg/1 unless specified
Biological Parameters
Coliform bacteria
Turbidity
Inorganic Chemicals
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Nitrate (as N)
Selenium
Silver
Fluoride
Organic Chemicals
Endrin
Lindane
Methoxychlor
Toxaphene
2,4D
2,4,5TP Silvex
Radionuclides
Radium 226 and 228 (combined)
Gross alpha particle activi.ty
Gross beta particle activity
1 per 1OO mi (mean)
1 NTU (Waiver to 5 NTU
possible)
0.05
1
O.O1O
,0.05
0.05
0.002
10
O.O1
0.05
1 .4-2.4"
•0.0002
0.004
0.1
0.005
0.1
O.O1
5 pCi/1
15 pCi/1
4 millirem/year
Based on annual average air "temperature
-------
- 746 -' . ,
the regulations and '-to notify the public i,f , the ^standards, ,are , ( .
exceeded. The Safe Drinking Water Act also provides that States
will develop their own regulations and some States have more
stringent requirements than those in the Interim Primary
Regulations.
The Maximum Contaminant Levels for coliform bacteria and
turbidity are essentially those contained in the 1962 Public
Health Service Standards (Table 2). It is generally agreed
that these MCLs are adequate to protect the public from
infectious disease transmitted by recent fecal contamination;
however, there is data to suggest that the coliform testing
procedures may not be entirely effective as an indicator of
certain diseases of viral and .protozoan etiologies. The
apparent problem with the indicator procedures is,not whether
those particular disease-producing organisms are found in
water contaminated with fecal material, but rather that a
number of viruses and protozoans can survive in water for
longer periods than the coliform indicators. Some viruses
and protozoans also seem more resistant to disinfection than
coliforms.
At the present time there are no national disinfection stan-
dards in the United States. Drinking water disinfection is
hot mandatory nor are there restrictions on the quantity and
type of disinfectant that can be used. The MCL of a monthly
mean of 1 coliform/1OO ml is the sole measure of microbio-
logical quality, but the means of achieving acceptable coliform
levels are not prescribed. Monitoring frequencies based upon
the size of the population served range from a minimum of
1/month for populations of less than 1OOO to SOO/month for
populations larger than 4,690,OOO. There is, however, a
provision in the Interim Primary regulations that allows
-------
- 747' -
TABLE 2 Provisions" In-the'National-' IntWrim Primary
• Drinking ..Water Regulations affecting the -. -
presence of microorganisms
-1 . Coliform -bacteria - 1 coliform/10O ml of water
.. a) A supply may substitute chlorine residual
determinations for not more than 75 % of the
required number of coliform analyses if
0.2 mg/1 free chlorine is maintained.
2. Turbidity - 1 nephelometric turbidity unit (NTU).
The State may allow up to 5 NTU if water does not
cause a risk to health and the Increased turbidity
does not: . . - . •
a) interfere with disinfection; •
b) prevent maintenance of an effective disinfectant
agent throughout the distribution system;
c) interfere with microbiological determinations.
States to permit the substitution of chlorine residual analyses
for up to 75 % of the monthly coliform sampling requirements.
If that option is selected, a free chlorine residual of at
least O.2 mg/1 must be maintained. Should the chlorine resi-
dual fall below O.2 mg/1, immediate sampling for coliform
bacteria is required.
The value of disinfectants in controlling pathogens in water
is unassailable, but, in one sense, disinfectants and their
by-products, and particularly chlorine and chloramlnes, are
-------
- 746 -
contaminants of increasing public health concern. They are
the most widely used synthetic chemicals in water treatment
in the US and they are used at relatively high concentrations.
EPA finds itself in a difficult position in prescribing con-
trols because there is very little .information currently
available on the human toxicology of chronic exposure to
disinfectant chemicals, their degradation products, or their
reaction by-products with other contaminant chemicals in
water. That is an appalling fact considering some of these
disinfectants have been in use for over 60 years.
The issue concerning the misuse of disinfectants has become
particularly acute in the past three years with the identifi-
cation of chloroform and other trihalomethanes in chlorinated
drinking water. But trihalomethanes are not the entire problem
by far; it is only the relative ease of identification and
quantification that has caused interest to concentrate on them.
The trihalomethanes should be considered indicators of the
existence of a host of undefined and perhaps undefinable
oxidized and halogenated chemicals that are introduced as a
result of chlorination. The same questions can be raised about
any disinfectant under consideration, be it chlorine, chlora-
mine, ozone, chlorine dioxide, or iodine.
The Environmental Protection Agency has recently proposed
regulations to limit trihalomethane.concentration aimed at
minimizing risks from unnecessary exposure to the by-products
of the disinfection process. High doses of disinfectants
should not be used to provide chemical oxidation treatment
unless it is part of another process which would control the
chemical by-products produced. In high organic water, water
treatment processes should be applied that involve purifi-
cation to reduce precursor levels before application of a
disinfectant.
-------
- 749 - -.-"••
These proposed 'regulations further conclude that it is ne'cessary
to take steps to limit and minimize' trihalomethanes in drinking
water by means that would not interfere with the maintenance of
pathogen control. The current proposed regulations include an
initial limit on total trihalomethanes (the sum of chloroform,
bromodichloromethane, chlorodibromomethane and bromoform) of
0.10 mg/1 for approximately 400 public water systems serving
populations greater than 75,OOO persons. Only monitoring is
required in smaller systems (10,000 to 75,000), and systems '
with less than 1O,OOO are not covered. The initial standard
was selected based on current feasibility and is not to be
construed as a "safe" level. This standard will be reduced
and population coverage increased as experience is gained.
To assure that any steps taken to reduce THM concentrations
in drinking water will not increase the possibility of
microbial contamination, additional microbiological moni-
toring is proposed for•a water system that is modifiying
existing treatment practice. Standard Plate Count (SPG)
determinations must be made at least daily both at the treat-
ment plant and in the distribution system, for one month
before and six months subsequent to the treatment change to
assure that no degradation of water quality occurs. Analyses
prior to any treatment change are intended to provide a base-
line to which subsequent effects can be compared. The appro-
priate number and sampling locations of SPG analysis should
be determined by the State or EPA depending on local con-
ditions, and significant deviations from the "normal" range
must be reported to the State or EPA and corrective actions
taken immediately. The proposed regulations also limit the
use of chlorine dioxide as a primary disinfectant to not
exceed 1 mg/1 because of possible adverse effects of by-
products and would not permit the use of chloramines as primary
disinfectants but only for maintenance of a distribution system
residual.
-------
- 750 -
EPA has also proposed that water systems subject to significant
raw water contamination from pollutions-related synthetic organic
chemicals initiate a program to study, design and construct
facilities utilizing granular activated carbon or equivalent
technology^ to minimize human exposure to those contaminants.
In conclusion, I think the United States has come a long way
in providing safe drinking water to her population. The Interim
Primary Regulations and the proposed trihalomethane regulations
are an important beginning. Our Revised Regulations, presently
under consideration, will supplement our present efforts and
begin to deal with contaminants where long-term impacts are
beginning to be understood (especially carcinogens). However,
in the future I see our attention being focused on the estab-
lishment of regulations that require specific water treatment
for supplies that are shown to be at risk from specific
contaminants. These regulations could specify disinfection,
maximum contaminant levels for individual contaminants,
coagulation and filtration, or other treatments when warranted.
Such an approach is not only cost-effective, but in reality
a sensible way to provide the public health protection needed.
I want to thank you for the opportunity to speak with you today
and I look forward to working with you in the future as we
together, in cooperation, provide safe drinking water to the
citizens of our respective countries.
-------
- 751 -
Y. Kott (Israel)
The surplus of water in the northern area of the country
and severe scarcity in the south, the utilization of over
95 percent of the water potential in the country and the
utilization of about 85 percent of the water for agricul-
tural purposes, caused an intervention of the government
to pass a law which nationalized all water resources from
individual property to the public.
This law enabled the national water company to pump win-
ter surplus water from lake Kinnereth and recharge it
into sandy aquifer in the coastal plane through a dual-
purpose well. During storage, water quality changed and
the number of coliform bacteria rose up to 1O /1OO ml;
turbidity was very high. Water quality improved at a long
storage or due to continuous pumping. The various quality
problems that have arisen by these operations caused the
Ministry of Health to request a committee that was wor-
king on a new law for water quality to establish a qua-
lity criterion for such water. Indeed, the Israel Drinking
Water Standard published in 1974 states at clause 14
that when drinking water is pumped from a dual-purpose
well it should be examined and found to be free of fecal
coliforms, streptococci and salmonella bacteria. In addi-
tion no more than two coliforms in 1OO ml would be allowed
and the water must be disinfected. As much as we .know, no
other country has as yet established this criterion in
its water standard.
As already mentioned above, most of the water is used for
agricultural purposes. The need for water has created a
-------
- -752 - .
situation in which the farmers are ready to utilize va- '<•
rious qualities of wastewater. It is estimated that re-
use of wastewater will reach in the near future to over
120 million cubic meters. The importance of quality ver-
sus type of crop growth has brought to create new cri-
teria for purified wastewater quality.
The irrigation with treated wastewater will have four
quality levels in which the unlimited one will have to
be 80 percentite of coliform, equal or less than 12, fe-
cal coliform, fifty percentile 2.2/100 ml. Residual chlo-
rine of 0.5 mg/1. It is mandatory that the water will
have to pass sand filtration. On the other hand, group A
which will allow irrigation of industrial crop like cotton,
dry fodder, seeds etc. will allow BOD of 6O mg/1, suspen-
ded solids up to 5O mg/1.
The mechanism of utilizing most of the water resources
and dividing the quality for each type of crop will main-
tain the true balance between needs and capability. It is
thought that combination of utilisation of all water re-
sources for the various needs will postpone the need for
water desalination from big plants which are much more
expensive.
-------
- 753 -
G. Muller- (P.'R.G.)
In Germany the addition of chlorine to drinking water is
regulated by two laws, i.e. the Foodstuffs'and Necessities
Law and the Federal Epidemics Law. On the basis of these two
laws statutory decrees have been laid down, one of these
being a decree concerning the treatment of drinking water
specifying the maximum amount of chlorine that can be added
to drinking water. The amount in question-is 0.3 mg Cl-/!.
In times of emergency or disaster this value can be increased
to 0.6 mg C12/1. On the other hand, the Drinking Water Decree
on the authorization of the Federal Epidemics Law lays down
the minimum amount of chlorine when this is to be used as a
disinfectant in drinking water. The chlorine content in
drinking water leaving the works must be 0.1 mg Cl^/l- In
combination with this, chlorinated water ex works must not exceed
a standard colony count value of 20 ml. The purpose of this
combination is to guarantee the actual effectiveness of the
chlorination.
The minimum and maximum \alues prescribed in the decrees for
chlorine in drinking water show that in the Federal Republic
the chlorine level permitted by law lies within very narrow
limits.
Experience to date has shown that this chlorine content may
be just sufficient to destroy pathogens entering the drinking
water network by penetration of waste water or river water,
especially since the organic substance flowing in at the same
time consumes a large proportion of the chlorine before the
disinfection occurs. This has been demonstrated time and
again in recent decades in cases of typhoid and paratyphoid-B
epidemics or epidemics due to other Salmonella species. The
so-called safety chlorination normally makes it possible to
reduce elevated colony counts, but is rarely sufficient for
-------
- 75.4 -
a reliable destruction of the pathogens that penetrate the ..
supply network via short circuits, cross connections, and
sucking back.
In Germany there are many public water supplies working on
ground water that cannot be faulted from a bacteriological
point of view, where the water is not chlorinated on leaving
the works because bacterial proliferation, detectable by
elevated colony counts, is simply not observed during the
distribution or storage of the water.
-------
- 755.-
J.A. Myhrstad (Norway)
The drinking water regulations in Norway are based on
the Health Act, which is administered by the Ministry
of Social Affairs.
According to the regulations, potable water should be
hygienic, safe, the water source should be protected
from microbiological and chemical pollution, and the wa-
ter treated in a proper way.
The regulations also state that waterworks supplying
between 100 and 1000 persons are subject to approval by
the local Boards of Health, and the greater works are
subject to approval by the National Institute of Public
Health. Since there is no detailed information on how to
proceed in order to achieve a wholesome potable water,
it is the responsibility of these authorities to consi-
der the necessary actions to be taken.
The waterworks can be given the right to expropriate
land in the catchment areas and put restrictions on the
existing and future activities, according to the Water
Resources Act, in order to protect the water sources from
pollution.
The Water Pollution Act is an important law which covers
all forms of water pollution, and it contains a general
prohibition, unless permission is granted, against most
activities liable to cause water pollution. This law is
administered by the Ministry of Environment.
According to the drinking water regulations, the water-
works are responsible for the water quality control. This
control is very often carried out in close cooperation
-------
- 756 -
* -..,.. . , . , . i , -i , ,-,
with the local 'Board of Health and National Institute of
Public Health.
In addition to the analysis,of raw water, treated water
and tap water samples, plant control and weekly reports
concerning communicable diseases in each community are
integrated in the control'procedure.
Surface water is by far the most important water source
in Norway, at least for waterworks supplying more than
100 persons. About 95 % of the population supplied by
these waterworks use surface water. The most important
quality problems are related to the. low pH-values of the
surface water (down to a pH-value of 3.4), the very soft
water (total hardness usually below 5-1O mg CaO/1) and
the coloured water (e.g. colours in the range of 3O-6O
mg Pt/1).
Treatment problems are mainly related to low water tem-
peratures in the winter time, which influence the coagu-
lation process, and the coloured water being extensively
used without adequate treatment, interfering with the
chlorination process.
Chlorine is still the most important disinfectant used,
but breakpoint chlorination is never achieved owing to
the very small concentrations of ammonia (usually well
below O.1 ppm ammonia - nitrogen). There is a tendency
to replace chlorine by ultraviolet irradiation. Chlorine
dioxide will not be used until the health effect has been
evaluated.
Examinations carried out have revealed high concentra-
tions of trihalomethanes caused by chlorination of co-
loured water.
-------
- - 757 -
Other problems are related to the fact that many water-
works do not treat the water in the proper way, and that
waterworks equipment, usually imported from abroad, is
produced in countries with water qualities differing very
much from the Norwegian water quality. One example is
the asbestos-cement pipes without internal coating. Ca
is extracted from the pipe material resulting in pH-va-
lues as high as 11.5. These pH-values increase the dis-
solution of heavy metals from taps made of brass alloys
and solders. The asbestos-fibre content in the water
will also be influenced.
We have no specific or detailed regulations concerning
water quality, but we have recommendations. Some orga-
nic micro-pollutants are included in the recommenda-
tions; however, trihalomethanes are not included for
the time being.
The present drinking water regulations will be revised
in the near future.
An expert group has recently been appointed by the Royal
Norwegian Council for Scientific and Industrial Research
to evaluate the need for research in the field of drin-
king water supply.
-------
758 -
G „
-------
- 759 -
city screening tests are already studied at. this moment
and will be a subject of thorough research in the near
future.
On safety ground it is recommended that water should ge-
nerally be treated by physical and biological processes
to the greatest possible extent and that chemical oxidants
should not be used more often than strictly necessary.
However, the use of chlorine as a disinfectant will be re-
placed by other means only after their effects have been
thoroughly studied and have proved to be fully acceptable.
An important subject which will be studied is the question of
whether the risks of chemical oxydation are great enough
to outweigh.large claims on surface areas which are used
at this moment, e.g. in dune infiltration. A deeper insight
is necessary to evaluate the long-term risks to human
health. Epidemiological studies are carried out at this
moment for this purpose.
From the viewpoint of risks to human health standards or
criteria for organic compounds in tapwater or raw water
destined for the drinking water supply can only be set
as interim regulations until more insight is obtained.
Other aspects which are considered are future structural
plans to obtain a reliable quality of potable water. The
protection of raw water sources can be performed by water
sanitation programs and protection ,of the underground
against industrial pollution. Storage of pretreated water
should be in places where aging of the water can occur,
with avoiding further pollution.
Criteria for organics in drinking water will possibly
be proposed not only for individual compounds but also
for group- and sum parameters and possibly by a series
of mutagenicity screening tests.
-------
_, 760 -
T. Stenstrorn (Sweden)
Public water supplies in Sweden should, according to the
law, be investigated bacteriologically with a frequency
of at least once a week for large water supplies to once
every third month for ground water supplies serving less
than 4OOO people. Private water supplies, usually hotels
etc., should be checked in the'same way. The chemical
quality of the water is tested once a year. The samples
are taken at the water-works.
Water- is classified as not suitable if the total number
of coliforms (MPN-method, 35°G, 48" h) is more than 1O/1OO
ml. If this number is 1-9 or i'f the total plate count
(22 C, 48 h) is more than 1OO the water is classified as
less suitable. The total plate" count at 22°C is a valuable
indicator of general contamination and of the efficiency
of the water treatment and disinfection.
Private wells are also tested for fecal coliforms (MPN
44 C, 48 h). Such water is classified as unsuitable if the
number of fecal coliforms is-more than 1O or if the total
number of coliforms is more' than 5OO.
The chemical quality control is'relatively similar to that
recommended by WHO. Heavy metals and organic micropollutants
are not analysed routinely. The only parameter used routi-
nely for organic substances is permanganate consumption. A
large survey for trihalometanes has, however, shown that
most water supplies have concentrations below 1O pg/1 of
the trihalomethanes at the consumer * s tap.
In only two cases out of 144 the concentration was above
1OO yg/1.
-------
-• 761 --
The Swedish regulations from 1967 and 1970 should-=be'revi-
sed in a few years. Recommendations for chemical and bac-
teriological monitoring of the water quality in distribu-
tion systems and at the consumer's tap will be published
early in 1979. The purpose of this is
1) to extend the monitoring of parameters of importance
for possible health effects to the consumer's tap, and
2) to provide data for making operational decisions in
order to avoid problems like .taste, odor or colour cau-
sed by corrosion o,r. bacteriological activity in the net-
work. For the bacteriological monitoring the membrane
method will probably be accepted. It is not so today.
Of the water from public water supplies in Sweden 53% is
surface water. 21% is infiltrated water usually from sandy
ridges deposited during the glacial period. 26% is ground
water without infiltration. About 15% of the population is
not connected to public wa,ter-supplies. They generally use
ground water. Restrictions on the use of polluted surface
waters as water supplies are not .considered necessary in
Sweden.
8O% of the surface water is treated with chemical floccula-
tion and filtration. Half of this water is also treated by
slow sand filtration, for example by the water-works in
Stockholm. Only rapid or slow sand filtration is used for
6 and 9% of the water, respectively.
The most common disinfectant i-s chlorine although chlorine
dioxide is used in some water-works such as Gothenburg.
Breakpoint chlorination as an oxidation step is not used.
With very few exceptions the ammonia concentrations are
below O.5 mg/1 and do not cause any problems except that
nitrite up to a few tenth of a mg/1 may be produced in the
distribution system. Many water-works, e.g. Stockholm, add
-------
- 762 -
ammonia to get chloramines. Ozon is used only at two
water-works.
Much of the Swedish research on drinking water problems is
now focused on the deterioration of water quality during
distribution. New treatment methods such as activated
carbon filtration will also be studied. The philosophy is
that water at the consumer's tap should be not only whole-
some but of a high quality with regard to taste, odor and
appearance.
Another problem, which should be studied, is the treatment
at small water-works of water containing e>g~. high concen-
trations of iron and manganese,.
Except for some local areas, Sweden has no shortage of raw
waters of a high quality. This is to a large extent due to
strict regulations on the discharge of domestic and indu-
strial sewage water. About 70% of the Swedish population
living in cities and villages with more than 2OO persons
are now connected to waste-water-works with biological as
well as chemical treatment. Another 26% has mechanical and
biological treatment (usually active sludge). Only 3% has
only mechanical treatment which is not accepted according
to the law. An increasing number of waste-water-works is
also supplied with a final filtration.
-------
- 763 -
P. Toft (Canada)
Canada has a federal system of government and consists
of 10 provinces and two territories. The distribution
of powers between the provinces and the federal govern-
ment is stated in the British North America Act of 1867.
Although this Act does not deal specifically with water.
resources, and therefore drinking water supply, judicial
interpretation over the years has resulted in a situ-
ation where there is a shared federal-provincial res-
ponsibility.
Under the Act, the ownership of natural resources, in-
cluding water, is vested in the provinces. Thus the pro-
vinces have the primary authority to legislate in the
area of municipal water supply.
Under their respective Health Acts the provincial de-
partments of health also have the power to control con-
ditions which constitute a threat to human health.
At the federal level, responsibility is assigned to the
Department of National Health and Welfare to investigate
and conduct programmes related to public health. For ex-
ample, research and investigation into the health aspects
of drinking water are undertaken with a view to public
health protection. In carrying out.such activities the
Department is required to co-ordinate its efforts with
those of the appropriate provincial authorities.
In Canada, drinking water is defined as a food and there-
fore in theory it could be regulated under the Food and
Drugs Act. Although this has been done for bottled (mi-
neral and spring) waters, standards have not been ex-
-------
- 764 -
tended to i'tap ' wa'ter-' owing "to" the major role traditionally
assumed by the provinces.
Generally the provinces play the lead role in ensuring
an adequate and safe supply of drinking water whereas the
Federal government provides leadership in ensuring ade-
quate standards for drinking water quality especially to
protect human health. •'•...
There are a few specific cases where the Federal govern-
ment is solely responsible for drinking water. These in-
<•
elude administering potable water regulations for all
common carriers (transportation crossing Canadian Inter-
provincial and International borders), and on Canadian
coastal shipping vessels, and the provision of potable
water in the Territories, Indian, reservations and mili-
tary bases.
Standards
With the exception of a few provinces which specify that
municipal water supplies must contain a certain minimal
level of chlorine, there are no legally enforceable stan-
dards for drinking water supplies in Canada. Rather,
there are guidelines for potable water quality and strict
control over the design and operation of treatment plants,
The federal government plays a lead role in establishing
drinking water guidelines.
Revision of the Canadian Drinking Water Guidelines
It was in 1968 that the first Canadian Drinking Water
Document was brought out by a joint committee comprising
-------
-.765 -
the Canadian Public Hearth Association Drinking <.Water>
Standards Committee and a federally-convened Advisory
Committee on Public Health Engineering. The latter in-
cluded representatives of the Federal and Provincial
Departments of Health, several universities and Canadian
water pollution and water resource control agencies.
Provincial governments, in turn,, .have adapted these
"guidelines" to suit their own particular situation.
A joint federal-provincial group led by the Department
of National Health and Welfare.is, currently revising
the drinking water standards published in 1968.
* U.S. GOVERNMENT TOWING OFHCE: 1979-281-147/122
-------
-------