The mean oxidation number of carbon (MOC ) - A useful concept for describing oxidation processes

Article abstract of DOI:10.1016/S0043-1354(00)00029-4

When characterizing real waste waters and sludges, usually collective parameters like COD and TOC are used. Detailed protocols (USEPA, APHA, DIN) concerned with the analytical determination of these parameters exist. However, some difficulties may arise when interpreting the measured values. Waste waters may contain substances refractory to the COD determination, or high concentrations of easily oxidizable inorganic salts may be mistaken for organic matter.When treating waste waters and sludges using physical, chemical or biological processes, there may be unexpected phenomena apparently contributing to a reduction or even concentration of the dissolved substances: adsorption effects (on reactor walls, catalyst, biomass), stripping of volatile organics, sedimentation, complexation, precipitation, or evaporation of water. When monitoring only COD or TOC, one cannot conclude whether a change of these values is due to a chemical degradation process or to a purely physical elimination like stripping.In this paper the concept of the mean oxidation number of organic carbon (MOC) is presented. By combining TOC (or DOC) and COD measurements, one can obtain valuable additional information concerning the reaction development during waste water treatment processes, especially when oxidative transformations are involved. The applicability of the MOC concept is demonstrated using the wet oxidation of pyridine, stripping of toluene, decarboxylation of malonic acid, and wet oxidation of phenol as well as the wet oxidation of sewage sludge as an example of a real system. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0043-1354(00)00029-4

Wat. Res. Vol. 34, No. 10, pp. 2689±2702, 2000
7 2000 Elsevier Science Ltd. All rights reserved
Printed in Great Britain
PII: S0043-1354(00)00029-4
0043-1354/00/$ - see front matter
www.elsevier.com/locate/watres
THE MEAN OXIDATION NUMBER OF CARBON (MOC)ÐA
USEFUL CONCEPT FOR DESCRIBING OXIDATION
PROCESSES
FREDERIC VOGEL*, JULIEN HARF, ANDREAS HUG and
PHILIPP RUDOLF VON ROHR
Institute of Process Engineering, Swiss Federal Institute of Technology, CH-8092, Zurich, Switzerland
(First received 1 January 1999; accepted in revised form 1 August 1999)
AbstractÐWhen characterizing real waste waters and sludges, usually collective parameters like COD
and TOC are used. Detailed protocols (USEPA, APHA, DIN) concerned with the analytical
determination of these parameters exist. However, some diculties may arise when interpreting the
measured values. Waste waters may contain substances refractory to the COD determination, or high
concentrations of easily oxidizable inorganic salts may be mistaken for organic matter.
When treating waste waters and sludges using physical, chemical or biological processes, there may
be unexpected phenomena apparently contributing to a reduction or even concentration of the
dissolved substances: adsorption e€ects (on reactor walls, catalyst, biomass), stripping of volatile
organics, sedimentation, complexation, precipitation, or evaporation of water. When monitoring only
COD or TOC, one cannot conclude whether a change of these values is due to a chemical degradation
process or to a purely physical elimination like stripping.
In this paper the concept of the mean oxidation number of organic carbon (MOC) is presented. By
combining TOC (or DOC) and COD measurements, one can obtain valuable additional information
concerning the reaction development during waste water treatment processes, especially when oxidative
transformations are involved. The applicability of the MOC concept is demonstrated using the wet
oxidation of pyridine, stripping of toluene, decarboxylation of malonic acid, and wet oxidation of
phenol as well as the wet oxidation of sewage sludge as an example of a real system. 7 2000 Elsevier
Science Ltd. All rights reserved
Key wordsÐoxidation number, waste water, organic carbon, COD and TOC, sewage sludge, global
parameters
NOMENCLATURE
N(C)j
OCi
total number of carbon atoms in molecule j
oxidation number of the organically bound car-
bon atom i
AOX
Ci
COD
adsorbable organic halogens (X=Cl, Br, I)
molar concentration of compound i, mol l
chemical oxygen demand, mg l
1
pO
TOC
ThOD theoretical oxygen demand
partial pressure of oxygen, bar
total organic carbon content, mg l
1
2
1
CODI chemical oxygen demand of the inorganics, mg
l
CODi chemical oxygen demand of substance i, mg l
1
1
1
CODO chemical oxygen demand of the organics, mg l
CODtot chemical oxygen demand of all substances pre-
1
sent, mg l
dissolved organic carbon, mg l
1
DOC
IC
m
MC
1
inorganic carbon, mg l
total number of organic molecules in a solution
INTRODUCTION
1
molar mass of carbon, 12 g mol
molar mass of molecular oxygen, 32 g mol
When characterizing real waste waters and sludges,
usually global parameters like the chemical oxygen
demand (COD) and the total organic carbon (TOC)
are used. Detailed protocols concerned with the
analytical determination of these parameters exist
(APHA, 1995a,b; Keith, 1996; DIN, 1980, 1983).
However, some diculties may arise when interpret-
ing the measured values. Inexperienced users may
not be aware of the factors that can in¯uence TOC
and COD measurements. For instance, waste waters
may contain substances refractory to the COD
1
MO
2
MOC mean oxidation number of the organically bound
carbon (of a single molecule or a waste water)
n
total number of carbon atoms in a speci®c mol-
ecule
*Author to whom all correspondence should be addressed:
Massachusetts Institute of Technology, Energy
Laboratory, E40-486, 1 Amherst Street, Cambridge,
MA 02139, USA. Tel.: +1-617-253-0070; fax: +1-617-
253-8013; e-mail: fvogel@mit.edu
2689

2690
Frederic Vogel et al.
determination or high concentrations of easily oxi- TOC removed from an aqueous solution by gas
dizable inorganics, i.e. salts. stripping under speci®ed conditions.
In this paper the concept of the mean oxidation 6. NPOCÐnonpurgeable organic carbonÐthe frac-
number of organic carbon (MOC) is presented. By
combining TOC (or DOC) and COD measure-
ments, one can obtain valuable additional infor-
mation concerning the reaction evolution during
waste water treatment processes, especially when
oxidative reactions are involved.
tion of TOC not removed by gas stripping.
The de®nition of the particulate organic carbon as
POC may lead to confusion, since POC is also used
for ``purgeable organic carbon''. We therefore rec-
ommend to use SOC (suspended or solid organic
carbon) when referring to particulate organic car-
bon. POC should only be used to designate purge-
able organic carbon, which is consistent with
NPOC for the nonpurgeable organic carbon. These
fractions of carbon are interrelated in the following
way:
GLOBAL PARAMETERS IN WASTE WATER ANALYSIS
Due to the large variety of compounds present in
waste waters and sludges, the quantitative determi-
nation of all or even most organic compounds is
not feasible. For the chemical characterization of
waste waters and sludges, usually the two global
parameters TOC and COD are used. Both are US
standard methods according to the American Public
Health Association (APHA). The US Environmen-
tal Protection Agency (USEPA) also issued a COD
protocol. In Europe, the German DIN (Deutsche
Industrie-Norm) protocol is commonly used. A dis-
cussion of the di€erences between the US and the
German COD protocols can be found in Janicke
(1983).
TC ˆ IC ‡ TOC
TOC ˆ DOC ‡ SOC
TOC ˆ POC ‡ NPOC
POC ˆ VOC
ꢀ1†
ꢀ2†
ꢀ3†
ꢀ4†
In most cases the interest lies in the fate of the or-
ganic carbon only. There are two di€erent standard
methods for the determination of TOC.
Total organic carbon (TOC)
Method A. The sample is acidi®ed to pH 2 or less
using a carbon-free acid, e.g. phosphoric acid, fol-
lowed by purging the acidi®ed sample at room tem-
perature with a carbon-free gas, e.g. O2 or N2,
removing any CO2 formed during acidi®cation. The
TC measurement of the purged sample is reported
as TOC. Unfortunately, sample purging also
removes VOC, so that this organic carbon measure-
ment is actually a NPOC determination.
Method B. Here TC of the original sample is
measured ®rst. Then a second sample is injected
into a chamber containing phosphoric acid which is
continuously purged with gas. The loaded gas
stream from this chamber is routed directly to the
NDIR detector without passing a combustion step.
The CO2 measured by the NDIR at this point is
exclusively from IC. TOC is then calculated by sub-
tracting IC from TC. Although in method B VOC
may be stripped as well, VOC is not detected
because there is not a combustion step, and the
TOC is de®ned as the sum of all organically
bound carbon in dissolved and nondissolved or-
ganic species. To determine the quantity of organi-
cally bound carbon, the organic molecules must be
broken down to single carbon units and converted
to a single molecular form that can be measured
quantitatively. TOC methods utilize heat and oxy-
gen, ultraviolet irradiation, chemical oxidation in
the liquid phase, or combinations of these methods
to convert organic carbon to carbon dioxide (CO2).
The CO2 can be measured directly by a nondisper-
sive infrared (NDIR) analyzer, it may be reduced to
methane and measured with a ¯ame ionization
detector (FID), or CO2 may be titrated chemically
(APHA, 1995a).
The methods and instruments used in measuring
TOC analyze fractions of total carbon (TC). These
fractions of total carbon are de®ned according to
APHA (1995a) as:
1. ICÐinorganic carbonÐthe carbonate, bicarbon- detector does not respond to VOC unless it absorbs
ate, and dissolved CO2. at the same wavelength as CO2. On the other hand
2. TOCÐtotal organic carbonÐall carbon atoms any IC in the form of CN , OCN or SCN will
covalently bonded in organic molecules. not form CO2 in the acidi®cation step and will be
3. DOCÐdissolved organic carbonÐthe fraction of stripped o€ as HCN, HOCN and HSCN, respect-
TOC that passes through a 0.45 mm-pore-diam- ively. Since these compounds are not detected by
eter ®lter. the NDIR, the IC measurement will be too low,
4. POCÐparticulate organic carbonÐalso referred leading to a TOC value which is too high.
to as nondissolved organic carbon; the fraction
Figures 1 and 2 summarize the di€erent methods
and their impact on TOC measurement. If neither
5. VOCÐvolatile organic carbonÐalso referred to VOC nor cyanide, cyanate, and thiocyanate are pre-
of TOC retained by a 0.45 mm ®lter.
as purgeable organic carbon; the fraction of sent in signi®cant quantities, both methods are ap-

MOCÐa concept for describing oxidation processes
2691
Fig. 1. Purging method (method A) for the determination of TOC and its response to various carbon-
aceous species present in waste waters. Shaded compounds highlight inadequacies of the method.
plicable. If the sample contains VOC but no cya- method at pH 4. The DOC of the samples purged
nide, cyanate, and thiocyanate, then only method B at a pH around 2 was lower due to precipitation or
will give correct values for TOC (or DOC). If all of ¯otation of colloidal organic particles.
these species are present, neither method will give
Chemical oxygen demand (COD)
correct results. Therefore it is important to analyze
the sample for these substances prior to a TOC
measurement. This will allow the selection of the
appropriate TOC method assuring that the
measured value correctly corresponds to the TOC.
As an example let us consider a sample of contami-
nated ground water containing trichloroethylene
(TCE), a typical VOC. Using method A to deter-
mine the TOC will result in erroneously low values
because most of the TCE will be stripped o€ during
the sample purging step.
The chemical oxygen demand (COD) is used as a
measure of the oxygen equivalent of the organic
matter content of a sample that is susceptible to
oxidation by the strong chemical oxidant potassium
dichromate (K2Cr2O7). The sample is re¯uxed at
(15022)8C in a strongly acidic H2SO4 solution for
2 h. Silver sulfate (Ag2SO4) serves as a catalyst.
After digestion the remaining unreduced Cr2O72
ions are titrated with ferrous ammonium sulfate to
determine the amount of K2Cr2O7 consumed. A
colorimetric method is also applicable for samples
which do not exhibit any turbidity or recalcitrant
Fukushima et al. (1996) noted considerably lower
concentrations of DOC in natural water samples
determined using the purging method at around parasitic color after digestion. The titrimetric
pH 2 (method A) than those measured by the method must be used if turbidity or parasitic color
di€erential method (method B) or the purging remains after digestion. The oxidizable organic mat-
Fig. 2. Di€erential method (method B) for the determination of TOC and its response to various car-
bonaceous species present in waste waters. Shaded compounds highlight inadequacies of the method.

2692
Frederic Vogel et al.
ter is calculated in terms of O2 equivalents (APHA, lation, it is possible to correct measured COD
values for the contribution from hydrogen peroxide.
Due to the frequent occurrence of chloride ions
1995b).
The digestion of the sample is carried out either
by using the open re¯ux or the closed re¯ux method. in waste waters and sea water, special attention was
In the closed re¯ux method, small volumes of given to their treatment. Chloride is oxidized by
sample are pipetted into glass vials containing the dichromate and would therefore contribute to the
premeasured reagents, including the silver catalyst COD of
a sample. Adding mercuric sulfate
and the chloride compensator (mercuric sulfate). (HgSO4) to the sample prior to the digestion pre-
The capped vials are placed into holes in an electri- cipitates the chlorides as mercuric salts, preventing
cally heated metal block whose temperature is con- interference by chloride ions in the COD measure-
ment.
trolled. After 2 h the heater is turned o€, and the
vials are allowed to cool down. In the open re¯ux
method, larger volumes of sample (50 ml) and re-
agents are heated in an Erlenmeyer ¯ask on a hot
plate or over an open ¯ame. The ¯ask is ®tted with
a re¯ux condenser which may be water or air
cooled. After 2 h the heat source is turned o€ and
the sample is allowed to cool down. Wol€ (1975)
improved the open re¯ux method to reduce the loss
of volatile organics during sample and reagent ad-
dition. It has to be noted that only certain combi-
nations of digestion method (i.e. open or closed
re¯ux), determination method (i.e. titrimetric or col-
orimetric), and COD concentration ranges are
USEPA approved.
Organic compounds are oxidized almost comple-
tely under the conditions applied. Some classes of
substances cannot be oxidized due to their speci®c
behavior (Baker et al., 1999). Volatile hydrophobic
molecules will evaporate and remain in the gas
phase, where no oxidation takes place. Some inor-
ganic species will also be oxidized, thereby contri-
buting to the COD, while others act as oxidizer
under the conditions applied (e.g. chromium(VI) or
manganese(VII) compounds). Table 1 lists some
examples for the coverage of the APHA COD stan-
dard method. The reasons why a substance is not
accounted for (fully) in the COD determination
may be as follows. (i) Its high volatility (e.g. tolu-
Interpretation of TOC and COD measurements
At this point it should be clear that both COD
and TOC measurements are subject to potential
error. As a consequence some diculties may arise
when interpreting or comparing COD and TOC
measurements. The following two examples are
used to illustrate this problem.
Methods. COD was determined according to
DIN (1980) (closed re¯ux method) using
a
Macherey±Nagel (Germany) R-2 T heating block,
Macherey±Nagel CSB 1500 test kits No. 29, and a
Macherey±Nagel Nanocolor 300D photometer
(measurement of Cr(III) at 620 nm). TOC was
measured according to the di€erential combustion-
IR method using a Dohrmann±Rosemount DC-190
analyzer. Fifty microliters of the samples were
injected manually with a Hamilton 100 ml syringe.
Combustion took place at 8008C in a quartz tube
®lled with a platinum catalyst. Oxygen (Pangas,
Switzerland, 99.95%) was delivered to the analyzer
at a rate of 200 ml/min. The CO2 formed during
combustion was detected using
a Rosemount
BINOS IR detector. Calibration was performed
using potassium hydrogenphthalate (KHP) sol-
utions (Fluka, Switzerland) as a standard. Before
the analyses, the instrument was tuned to achieve
standard deviations of the KHP injections of less
ene, tetrachloroethylene escape from the oxidizing than 2%. All injections were done in triplicate. For
solution into the headspace), (ii) it has already DOC and dissolved COD determinations, the
reached its highest degree of oxidation (e.g. nitrate, samples were ®rst ®ltered using a 0.45 mm Te¯on ®l-
sulfate, carbonate, phosphate), or (iii) its resistance ter (Titan, diameter 25 mm). Although Te¯on is not
to chemical oxidation by dichromate (e.g. am- recommended for aqueous matrices, other ®lter ma-
monium, pyridine, tri¯uoroacetic acid). Acidic pot- terials were found to either leach organic material
assium dichromate solution will also oxidize into the sample or to adsorb organics from the
hydrogen peroxide to molecular oxygen. Kang et al. sample. The samples could be pushed easily through
(1999) have measured the contribution of hydrogen the ®lter disc (secured with a Luer lock) with a 5 ml
peroxide solutions to the COD. Using their corre- glass syringe (SGE, Australia).
Table 1. Selected examples for the coverage of the COD standard method APHA 5220 B, APHA (1995b)
Organic species
Inorganic species
Completely
oxidizable
o-phthalic acid; ethanol; acetic acid; methyl
acrylate; E-caprolactam; NTA; glucose; chitin; 1,4- thiosulfate; hydrogen peroxide
dioxane; cyclohexanol
sul®de; nitrate; sul®te; thiocyanate; chloride; iron(II);
Partly oxidizable
Hardly/not (further) pyridine; dichloromethane; tetrachloroethylene;
toluene
potassium cyanide; dithionite
nitrate; sulfate; carbonate; phosphate; ammonium
oxidizable
trimethylamine; tri¯uoroacetic acid; cyclohexane;
lindane

MOCÐa concept for describing oxidation processes
2693
Fig. 3. Measured COD during the wet oxidation of a pyridine solution with oxygen (2008C, pO =20
bar, pH=1.9, 1.5 mM ferrous sulfate). Error bars represent the 95% con®dence interval of the
analyses.
2
Example 1: wet oxidation of an aqueous pyridine 1178 mg l 1, corresponding to a theoretical COD of
solution. An acidi®ed, aqueous pyridine solution 2625 mg l 1, the experimental COD value is only
ꢀpH ˆ 1:9† was wet oxidized with oxygen 23 mg l 1. Furthermore, during the oxidative treat-
ꢀpO ˆ 20 bar† in an autoclave at 2008C in the pre- ment the COD increases, although pyridine is sup-
2
sence of 1.5 mM ferrous sulfate. Samples were with- posed to be oxidized. What is the explanation for
drawn periodically and analyzed for COD. The this paradoxical behavior?
result is shown in Fig. 3. The ®rst COD value has
Pyridine is refractory to the APHA COD
been corrected for the contribution from the iro- method. Its response is only about 1% of the theor-
n(II). In the other samples all the iron had been etical value. During wet oxidation pyridine is prob-
oxidized to iron(III), and no correction to the COD ably oxidized to hydroxypyridine (Crain et al.,
was thus necessary. 1993). This substance is less refractory to the COD
Although the initial pyridine concentration was determination than pyridine itself, leading to an
Fig. 4. Trend of measured TOC during stripping of an aqueous toluene solution at room temperature
with air. Error bars represent the 95% con®dence interval.

2694
Frederic Vogel et al.
increase in COD response. In this case the COD is mixture. In this paper a systematic approach lead-
not adequate for describing the oxidative degra- ing to the MOC concept (mean oxidation number
dation of pyridine.
Example 2: sparging of an aqueous toluene sol-
of carbon) is presented.
``Oxidation'' in organic chemistry means an
ution with air at ambient conditions. Synthetic air increase of the oxidation number. This can be
from a gas cylinder was bubbled through an aqu- accomplished by two basically di€erent ways:
eous solution of toluene at room temperature. The
TOC was chosen for monitoring the course of treat-
carbon (typically oxygen);
. insertion of an atom more electronegative than
ment. Figure 4 shows the decrease of the TOC
during the sparging. The uncertainty represented by
bon (typically hydrogen).
the 95% con®dence interval is largely due to the
. removal of an atom less electronegative than car-
The mean oxidation number of carbon MOC for a
single organic molecule with n atoms of carbon is
de®ned as
diculty of sampling aqueous solutions of toluene
due to its volatility, and not to the uncertainty in
the TOC determination itself. Of course, the
observed TOC decrease of 87% after 75 min is not
due to a chemical conversion but to the purely
physical stripping of the toluene. A COD plot
would show qualitatively the same behavior. It
should be mentioned, however, that the experimen-
tal COD of toluene is only 40% of the theoretical
oxygen demand. This is due to the volatility of tolu-
ene, which evaporates partly from the re¯uxing sol-
ution.
From a single TOC-vs-time (or COD-vs-time)
plot it is not possible to decide whether a change in
concentration is due to a chemical conversion or a
physical e€ect like stripping, evaporation, or
adsorption. Often these physical phenomena are
unwanted accompanying e€ects of waste water
treatment processes like wet oxidation, aerobic bio-
degradation, or ozonation. Therefore one would
like to check if a desired chemical conversion, e.g.
oxidation, really takes place. In the following sec-
tion we will introduce a concept that will allow one
to di€erentiate between certain e€ects. The basic
idea is to combine both TOC and COD values in a
new parameter.
n
X
OCi
def
ˆ
iˆ1
MOC
ꢀ5†
n
where OCi is the oxidation number of the ith car-
bon atom in the molecule. Classes of substances
with the same MOC value can be formed because
the MOC concept does not di€erentiate between
heteroatoms. All atoms that are more electronega-
tive than carbon contribute in the same way to its
oxidation state, as do all those that are less electro-
negative. The concept of MOC is not dependent on
whether a carbon center is connected to a hydroxy,
an amino, a nitro, or a chloro substituent. All four
groups contribute +1 to the oxidation number of
the carbon center. Thus phenol, aniline, nitroben-
zene, and chlorobenzene all have a MOC value of
2/3. Each additional electronegative substituent
increases the MOC value by 1/3. Hexachloroben-
zene has thus a MOC value of +1.
The mean oxidation number MOCL of m organic
molecules Aj in a solution is de®ned as
m
X
‰AŠj Á MOCj Á NꢀC†
j
THE MOC CONCEPT
jˆ1
def
ˆ
MOCL
ꢀ6†
m
X
Among all atoms occurring in compounds pre-
sent in waste waters, carbon is the most important
one to the environment. Organically bound carbon
is the nutritional base for all animal life. The end
‰AŠj Á NꢀC†
j
jˆ1
where [A]j is the molar concentration of Aj, MOCj
is the mean oxidation number of carbon for Aj, and
N(C)j is the total number of carbon atoms in Aj.
In most real waste waters or sludges, the analyti-
cal determination of the organic matter as single
compounds is not feasible. For such cases COD
and TOC values can be combined to give a MOC
estimate for the waste water or the sludge according
to$
degradation product, CO2, serves as
a carbon
source for the green plants. Waste water treatment
processes try to convert the environmentally harm-
ful carbon species into a less problematic one, often
CO2. It is thus sound to follow the fate of the total
organically bound carbon. A suitable method for
this purpose is to monitor its mean oxidation state,
as proposed by Stumm and Morgan (1981). Man-
tzavinos et al. (1996) used the mean oxidation state
of carbon to follow the oxidative degradation of
polyethylene glycols. They were able to relate the
increase of the mean oxidation state of carbon to
4MC CODO
CODO
TOC
MOCS ˆ 4
Á
ˆ 4 1:5
ꢀ7†
MO
TOC
2
the accumulation of organic acids in the reacting CODO is the organic COD, because the MOC con-
cept applies only to organic substances. While
TOC/IC methods di€erentiate organic from inor-
$A derivation of this equation is given in the appendix..
ganic carbon, COD standard procedures do not do

MOCÐa concept for describing oxidation processes
2695
so directly. In the next section a method to deter-
mine the organic CODO, using the standard COD
protocol will be proposed.
The subscript ``S'' in equation (7) refers to ``sus-
pension'', because TOC and COD also include sus-
pended organic matter. If only the dissolved
substances are of interest, DOC and the CODO of
the ®ltered sample (0.45 mm) are to be used in
equation (7).
For exceptions see Table 1 and Janicke (1983).
Baker et al. (1999) have classi®ed all compounds
listed in Janicke (1983) according to their COD-
to-ThOD ratio.
The MOC value must always lie in the range [ 4,
+4]. Values outside this range are chemically im-
possible and indicate a problem during the analyti-
cal determination of COD or TOC. Table 2 gives
some examples of organic compounds and their
MOC values. Of course, MOC values may also be
non-integer numbers. Phenol, for instance, has a
MOC value of 2/3.
The following assumptions were used to derive
equation (7):
1. The organic substances consist only of C, H, O.
2. H has an oxidation number of +1.
3. O has an oxidation number of 2.
Aliphatic short-chain alcohols like methanol,
ethanol, and propanol are the most reduced organic
species completely soluble in water. At the other
4. The organic COD can be determined (CODO).
5. The measured COD value corresponds to the end of the MOC scale, short-chain mono and dicar-
theoretical total oxygen demand (ThOD).
boxylic acids like formic and oxalic acid are found.
Aqueous solutions usually exhibit values in the
range [ 2, +3]. Urea is an interesting compound as
it is one of the few organic substances (if not the
only one) having a MOC of +4. CO2 and its ions
formed in aqueous solution are put in parentheses,
because these compounds are not organics.
These assumptions will now be discussed in some
detail.
1. In most waste waters the number of organically
bound heteroatoms S, P, and Cl is small com-
pared to C, H, O, and N. In other words: the
frequency of C bound to S, P, or Cl is much
lower than C bound to C, H, O, or N. Therefore
the mean oxidation state of the organic carbon is
determined mostly by C, H, O, and N. The fate
of the organically bound nitrogen deserves
special consideration (see In¯uence of organically
bound nitrogen on the determination of the
MOC value section). The presence of organically
bound nitrogen does not a€ect the applicability
of equation (7) for determining the MOC value
in most cases. In contrast, halogen substituents,
especially Cl, impose a limitation on the applica-
bility of equation (7). This is discussed in the In-
¯uence of chloro substituents on the
determination of the MOC value section.
2. Always correct for organic compounds.
3. Mostly correct for organic compounds. Excep-
tions: organic (hydro)peroxides.
4. Will be discussed in detail in Suggestion for
determining the organic COD (CODO) section.
5. Oxidation of most organic compounds according
to the standard COD method (APHA, 1995b)
yields 95±100% of the theoretical (ThOD) value.
In¯uence of organically bound nitrogen on the deter-
mination of the MOC value
Equation (7) builds on the assumption that only
carbon atoms are oxidized during the COD deter-
mination. If other atoms consume oxygen (or
dichromate, respectively), the value determined with
equation (7) will be lower than the true mean oxi-
dation number of carbon. Likewise, if other atoms
act as oxidizers, lowering the consumption of
dichromate, the value obtained will be higher than
the true mean oxidation number of carbon. Since
organically bound nitrogen exhibits this behavior in
some molecules, a brief discussion of the impact on
the experimentally determined MOC value using
COD and TOC measurements will be given here.
During the COD determination organically
bound nitrogen may be converted to either am-
monium NH+4 , nitrate NO3 , or molecular nitrogen
N2. Nitrogen containing organic compounds fall
into one of the two following categories:
1. the oxidation state of N does not change during
the COD determination,
2. the oxidation state of N changes during the
COD determination.
Table 2. Some organic compounds and their mean oxidation num-
ber of carbon (MOC). Note that most compounds encountered in
aqueous solutions lie in the range from 2 to +3. Of course many
compounds not listed here have fractional MOC values (e.g. phe-
nol has 2/3)
Equation (7) will give accurate results only for com-
pounds adhering to category 1. For these molecules,
the organically bound nitrogen does not take part
in the redox reactions and does not a€ect the oxy-
gen demand of the molecule. The amino group, for
example, is converted to ammonium, so that the
nitrogen stays at the oxidation state of 3. Aniline
thus exhibits the same COD as phenol (on a per
mole basis). On the other hand, the nitro group is
4
3
2
1
Methane
Ethane
Methanol, ethanol, propanol
Acetaldehyde, benzene
0
Formaldehyde, acetic acid, benzoquinone, dichloromethane
Maleic acid
Formic acid
Oxalic acid
urea, (CO2, carbonic acid, carbonate, bicarbonate)
+1
+2
+3
+4

2696
Frederic Vogel et al.
transformed to nitrate, which corresponds to an addition of mercuric sulfate. In order to apply the
oxidation of the nitrogen atom from +3 to +5. concept of the mean oxidation number of carbon,
Nitrobenzene, for example, exhibits a higher COD CODO must be known. The following is a method
than phenol and aniline (on a per mole basis), for determining CODO.
because additional oxygen is required to oxidize the
1. Determine the total CODtot according to APHA
nitrogen in the nitro group to nitrate. Another
5220 (or an equivalent protocol).
example is the azo group, which is mainly converted
2. Check if the sample contains VOC.
to molecular nitrogen, corresponding to an oxi-
dation of the nitrogen atoms from 1 to 0. As a
test strips) for cyanide, sul®de, cyanate, thiocya-
3. If YES c check (e.g. semiquantitatively using
result, using COD values to determine the MOC
value according to equation (7) will give MOC
sparge with gas (removal of cyanide as HCN,
nate. Go to step 4. If NO c acidify sample and
values that are too low when signi®cant concen-
sul®de as H2S, cyanate as HOCN, and thiocya-
trations of nitro or azo compounds are present.
nate as HSCN).
4. Check (e.g. semiquantitatively using test strips)
for sul®te, nitrite, thiosulfate, dithionite, Fe2+
Mn2+, H2O2.
5. Measure quantitatively the substances found in
,
In¯uence of chloro substituents on the determination
of the MOC value
The same basic considerations discussed in the
preceding section for nitrogen containing com-
pounds have to be made for chloro substituents.
Since the chlorine atoms are oxidized during the
COD determination, the COD value will be too
high and the MOC value too low. We would like to
compare the MOC value determined by equation
(7) with the true MOC value of chlorophenol. The
complete oxidation of chlorophenol can be written
as:
signi®cant amounts in step 3 and/or 4 c Ci in
1
.
mol l
6. CODI ˆ Si CODiÁCi
7. CODO=CODtot CODI
CODI is the inorganic contribution to the total
COD, and CODi is the COD of the individual com-
pounds determined in step 5 in g O2 per mole sub-
stance. Some of these values are reported in Table 3.
Uncertainty of MOC values
C6H5OCl ‡ 6:75 O2 46 CO2 ‡ 2:5 H2O ‡ 0:5 Cl2
ꢀ8†
Using Gauss' law of error propagation, the
uncertainty of the MOC value obtained by equation
(7) can be determined. The mean square error is
and the theoretical, or total, oxygen demand is
1
1681 mg O2 g chlorophenol. The TOC is 560 mg
s
ꢀ
ꢁ
ꢀ
ꢁ
2
2
CODO
TOC
DCODO
CODO
DTOC
TOC
1
C g chlorophenol. The corresponding MOC value
DMOC ˆ 1:5
‡
ꢀ9†
is then 0.5, compared to the true MOC value of
chlorophenol 0.3.
The absolute error depends not only on the relative
errors of the CODO and TOC but also on the
MOC value itself. Relative errors for MOC values
are meaningless since the MOC value can be zero
leading to in®nite relative errors. Even MOC values
di€erent from but close to zero would exhibit very
large relative errors, which lacks any signi®cance.
For solids-free samples maximum relative errors for
the COD and the TOC of 3 and 2%, respectively,
have been attained with our equipment (see
Methods). Harf (1998) has also determined MOC
values for samples containing suspended solids.
The con®dence interval at the 95% signi®cance
level is [MOC 1.96DMOC, MOC+1.96DMOC],
and has been included in Figs 6±8.
Using COD values to determine the MOC value
according to equation (7) will give MOC values
that are too low when signi®cant concentrations of
halogenated compounds (AOX) are present.
An interesting observation is the fact that the
presence of mercuric sulfate does not hinder the for-
mation of molecular chlorine during the COD de-
termination of chloro-compounds (Janicke, 1983).
This means that the chloro substituents do not
form chloride ®rst, which would then be oxidized to
chlorine gas, since the mercuric sulfate would
rapidly precipitate any chloride formed. We assume
that the chloro substituents form chlorine free rad-
icals directly which recombine rapidly to molecular
chlorine gas.
Applications of the MOC concept
When treating waste waters and sludges with
physical, chemical or biological processes, there
may be unexpected phenomena (apparently) contri-
buting to an elimination or even concentration of
the dissolved substances:
Suggestion for determining the organic COD
(CODO)
COD does not di€erentiate between organic and
inorganic carbon. All oxidizable substances contrib-
ute to the COD measurement with some exceptions,
as noted in Table 1. Interference can be avoided in . adsorption on reactor walls, ba‚es, stirrers, etc.
the case of the otherwise oxidizable chloride ions by (fouling, scaling);

MOCÐa concept for describing oxidation processes
2697
Table 3. ThOD and measured COD values for selected inorganic species. The COD values of the ®rst ®ve species have been determined
by the authors. Data for hydrogen peroxide were taken from Kang et al. (1999)
1
ThOD (g O2 mol substance)
1
COD (g O2 mol substance)
Species
Formula
Nitrite
Sul®te
Thiosulfate
Dithionite
Iron(II)
NO2
SO23
16
16
64
48
8
15.5
14.8
72.6
35.2
8.0
13.5±20.4
S2O23
S2O24
Fe2+
H2O2
Hydrogen peroxide
16
. adsorption on suspended particles (catalyst, bio- the COD determination, and (ii) oxidation of the
organics (pyridine and its partial oxidation pro-
ducts).
mass);
. stripping or evaporation of volatile organics;
. sedimentation of solids;
. precipitation due to complexation or change in
pH;
If the determination of the MOC value for a
waste water sample yields values outside the range
[ 2, +3], one should check with Table 5 for abnor-
mal COD or TOC responses. In this case, the MOC
value can be used to verify the consistency of COD/
TOC values of the same sample. Three more
examples will now be presented how the MOC con-
cept can be applied for monitoring chemical conver-
sions.
. concentration of substances due to evaporation
of water.
When monitoring only COD or TOC, one cannot
decide whether a change of these values is due to a
chemical degradation process or to a physical elim-
ination like stripping. Table 4 shows some of these
phenomena and their e€ect on COD, TOC, and
Example 3: decarboxylation of malonic acid in an
1
aqueous solution. A solution containing 2 g l of
MOC values. The down arrow symbolizes
a
malonic acid was heated up to 1408C without oxy-
gen in an autoclave. When the desired temperature
had been reached, samples were withdrawn period-
ically and analyzed for COD and TOC. Figure 5
shows the trend of the measured COD and TOC
values. Note that the initial portion of the graph
(between 0 and 5 min) is distorted due to the heat
up time. The value at t ˆ 0 corresponds to the in-
itial solution before heating. The second sample
decrease, the up arrow an increase of the corre-
sponding value, Ð indicates that the value is not
a€ected. The quantity CODO is the chemical oxy-
gen demand of the organics. A method for its deter-
mination is presented in the section Suggestion for
determining the organic COD (CODO). MOCL
refers to the MOC value of solutions and has to be
replaced by MOCS in case of suspensions (e.g.
sludges). Of course, some of these processes may was taken about 5 min after the temperature in the
autoclave had stabilized, or about 40 min after
turning on the heaters.
also take place simultaneously. It should be noted
that decarboxylation (i.e. splitting o€ of CO2) and
oxidation (typically incorporation of an oxygen
While the COD remains constant within the ex-
perimental uncertainty, the TOC decreases steadily.
This can be explained by considering the reaction
taking place. Malonic acid decarboxylates according
to
atom by forming
a carbon±oxygen bond) are
handled as two separate processes. As noted earlier,
the COD and TOC determinations themselves may
be a source of unexpected e€ects. In Table 5, ®ve
important cases are listed likely to occur in waste
water analysis. This table can be used to detect pro-
blems in COD and TOC determination which may
lead to erroneous interpretation of COD or TOC
data. The e€ects in Table 4 can occur together with
phenomena in Table 5, as observed in Example 1:
HOOC-CH2-COOH 4 CH3-COOH ‡ CO2
ꢀ10†
A simple calculation may prove that during such a
decomposition no change in the COD occurs,
because only carbon in its highest oxidation state
(i) presence of a substance (pyridine) refractory to (+4) is eliminated from the solution. It is therefore
Table 4. Physico-chemical processes typically encountered in waste water treatment and their in¯uences on CODO, TOC and MOC
Phenomenon
CODO
TOC
MOC
Physical removala of organics with MOC > MOCL
Physical removal of organics with MOC < MOCL
Physical removal of organics with MOC=MOCL
Evaporation of water
Oxidation of the organics
Reduction of the organics
Decarboxylation
q
q
q
Q
q
Q
±
q
q
q
Q
±
±
q
q
Q
±
±
Q
q
q
aComprises stripping, adsorption, precipitation, evaporation, sedimentation, without change of chemical oxidation state.

2698
Frederic Vogel et al.
Table 5. In¯uence of abnormal COD responses and inadequate TOC method selection on MOC values
Cause
CODO
TOC
MOC
Example
Waste water contains substances refractory to COD determination
Waste water contains substances with COD > 100% of ThOD
CODI > 0
High concentrations of nitro or azo compounds present
TOC determined using purging method when VOC are present
too lowa
±
±
±
too high
too low
too low
too lowb
too low
Pyridine
Furan
Nitrite (cf. Table 3)
Nitrobenzene, Orange II
Tetrachloroethylene, toluene
too higha
too higha
±
±
±
too low
aWhen assuming that CODO=CODtot
.
bContribution of organically bound nitrogen to the CODO value is consistent with the de®nition of CODO. Equation (7) builds on the
assumption that only carbon atoms are oxidized. Because the nitrogen in nitro and azo compounds is oxidized as well, MOC values
determined with equation (7) will be lower than the true mean oxidation number of carbon in these molecules.
not very helpful to follow the course of this reaction be expected in the ®nal solution. In fact, the MOC
values seem to approach 0 asymptotically. The pre-
sence of acetic acid was con®rmed by HPLC.
Example 4: wet oxidation of an aqueous phenol
using COD values. Although the decrease of the
TOC accounts for the formation of CO2, it does
not provide any further information on the fate of
the remaining dissolved carbon.
1
solution. A solution containing 2 g l
of phenol
was wet oxidized in an autoclave at 1408C using
pure oxygen and an initiator. These investigations
are described in more detail in Vogel et al. (1996);
Vogel (1997). After the reaction had been started,
samples were withdrawn periodically and analyzed
for TOC, COD, and individual substances using
HPLC. Figure 7 shows the calculated and measured
MOC values for this experiment.
Combining both valuesÐCOD and TOCÐand
calculating the MOC value according to equation
(7) leads to the representation shown in Fig. 6.
According to Table 4 there are a priori three di€er-
ent phenomena able to explain the observed MOC
decrease.
1. Stripping of organics with a MOC > MOC sol-
ution: not possible, because there was no gas
¯ow during the experiment.
According to Table 4 there are a priori two
phenomena able to explain the observed MOC
increase during the ®rst 15 min.
2. Reduction of the organics: not possible, because
no reducing agent was present.
1. Stripping of organics with a MOC < MOC of
the solution: basically possible considering the
continuous gas ¯ow maintained during the exper-
iment. The reactor, however, was equipped with
a re¯ux condenser, and no organics could be
detected in the o€-gas. Furthermore all identi®ed
substances in the reacting solution have a MOC
valuer 0.33 (hydroquinone, catechol, p-benzo-
quinone, formic acid, acetic acid, maleic acid,
3. Decarboxylation: possible and probable regard-
ing the properties of malonic acid (saturated
dicarboxylic acid); Streitwieser and Heathcock
(1985), p. 865, con®rm reaction 10
The initial MOC value of the solution is 1.34. This
®gure compares well to the theoretical value of
malonic acid of 1.33. After complete conversion of
the malonic acid only acetic acid ꢀMOC ˆ 0† is to
Fig. 5. COD (*) and TOC (q) evolution during the thermal treatment of an aqueous solution of
malonic acid at 1408C. Error bars represent the 95% con®dence interval of the analyses.

MOCÐa concept for describing oxidation processes
2699
Fig. 6. MOC evolution during the thermal treatment of an aqueous solution of malonic acid at 1408C.
Error bars represent the 95% con®dence interval.
oxalic acid, pyruvic acid).
fact, for this experiment up to 25% of the DOC
could not be accounted for by the carbon in indi-
vidual compounds analyzed by HPLC.
2. Oxidation of the organics: likely due to the reac-
tion conditions and the presence of an oxidizing
agent (oxygen). Detection of oxidized organics
(quinones, carboxylic acids) con®rmed this hy-
pothesis.
The increase in MOC during the ®rst 10 min is
due to the disappearance of phenol (MOC 0.67)
and the formation of carboxylic acids (mainly for-
mic, pyruvic, oxalic) with high mean oxidation
states of carbon (+2, +0.67, +3, respectively).
Pyruvic and oxalic acids start to decompose quickly
after 10 and 20 min, respectively, leveling-o€ the
MOC-vs-time pro®le after about 40 min. At this
point, the incorporation of oxygen into the mol-
During the ®rst 10 min of the reaction, the MOC
values determined with equation (7) agree well with
the calculated MOC values from HPLC concen-
trations of individual compounds using equation
(6). The di€erence at later times indicates that one
or more compounds are not detected by HPLC. In ecules (increasing the MOC value) and the decar-
Fig. 7. Evolution of dissolved MOC during the wet oxidation of an aqueous phenol solution. Dots (*)
represent the experimental values calculated using equation (7), squares (q) are the MOC values calcu-
lated from the concentration of individual compounds identi®ed in the mixture (equation (6)). Error
bars represent the 95% con®dence interval.

2700
Frederic Vogel et al.
Fig. 8. Evolution of MOC during the wet oxidation of sewage sludge. Dots (*) represent the exper-
imental values calculated with equation (7) using DOC and COD from the ®ltered samples (0.45 mm),
squares (q) are the MOC values calculated with equation (7) using TOC and COD from the un®ltered
samples. Error bars represent the 95% con®dence interval. Straight bar ends correspond to the squares,
curved bar ends to the black dots.
boxylation reactions (decreasing the MOC value) the oxidation of phenol, where mainly formic acid
was observed (see example 4).
seem to balance each other. The DOC in the reac-
tion solution after 100 min consists of 47% formic
acid, 23% acetic acid, 7% other carboxylic acids,
and 23% of one or more unidenti®ed compound(s).
Since the measured and calculated MOC values do
not di€er much at 100 min, the unidenti®ed com-
pound(s) must have a MOC value close to 1.5. This
important piece of information was used to ®nd
possible compounds that form the unaccounted
23% of carbon (Vogel, 1997). This example demon-
strates how the concept of MOC can provide valu-
able information which global parameters alone do
not.
The MOC of dried sewage sludge can be esti-
mated using elemental analysis. Because sewage
sludge usually contains signi®cant amounts of orga-
nically bound nitrogen, this atom has to be con-
sidered in the calculation of the MOC from an
elemental analysis formula. Fortunately, in many
organic molecules the nitrogen has an oxidation
number of
3. Exceptions are hydroxylamines
( 1), nitro (+3), nitroso (+1) and azo ( 1) com-
pounds, but which are not expected to be present in
large quantities in sewage sludge. We therefore
assume that the organic nitrogen present in the
dried sewage sludge has an oxidation state of 3.
The elemental analysis for the untreated sludge
Example 5: wet oxidation of sewage sludge. A sus-
pension of sewage sludge (dry matter ˆ 1 wt%) was
wet oxidized at 2008C with pure oxygen at acidic
conditions (Harf, 1998). Again, samples were taken
periodically and analyzed for TOC, DOC, total
COD and dissolved COD, yielding two sets of
MOC values, one for the total organic content, and
one for the dissolved organics. Figure 8 shows the
evolution of the sewage sludge wet oxidation using
these MOC values. Interestingly, the MOC-vs-time
pro®le qualitatively resembles the one measured for
the phenol oxidation (Example 4), but the absolute
change in MOC is lower (from 0.15 to 0.95 vs 0.3
to 1.5 for the phenol experiment). This is consistent
with the observation that the wet oxidation of sew-
age sludge catalyzed by ferrous ions (Harf, 1998)
forms mainly acetic acid (MOC 0) with only small
used in example
C
5
199H382O76N27P6S (Harf, 1998). We will neglect
yields the formula
the S and P atoms in our MOC calculations
because of their low occurrence. This sum formula
yields a MOC of 0.75, which compares to about
0.1 for the suspension using COD and TOC
values (see Fig. 8 at t ˆ 0).
One explanation for this di€erence is that during
the drying phase at 1058C preceding the elemental
analysis some volatile substances present in the
unoxidized sludge, e.g. acetic and formic acid, have
been evaporated and thus do not appear in the el-
emental analysis. Because their MOC values are 0
and +2, respectively, their loss through evaporation
decreases the MOC of the dry matter compared to
the suspension. This e€ect corresponds to the physi-
amounts of formic acid (MOC +2), as opposed to cal removal (evaporation) of compounds with a

MOCÐa concept for describing oxidation processes
2701
(CSB) im Bereich uber 15 mg/l. Deutsches Institut fur
È
Normung.
MOC value higher than the MOC of the suspension
(cf. Table 4).
DIN. (1983) DIN 38 409 Teil 3, Deutsche Einheitsverfah-
ren zur Wasser-, Abwasser- und Schlammuntersuchung,
Summarische Wirkungs- und Sto€kenngrossen (Gruppe
H), Bestimmung des gesamten organisch gebundenen Koh-
lensto€s (TOC). Deutsches Institut fur Normung.
Fukushima T., Imai A., Matsushige K., Aizaki M. and
Otsuki A. (1996) Freshwater DOC measurements by
high-temperature combustion: Comparison of di€eren-
tial (DTC DIC) and DIC purging methods. Wat. Res.
30(11), 2717±2722.
Harf J. (1998) Simultaneous Wet Oxidation of Waste
Water and Sewage Sludge with Oxygen (Simultane Nas-
soxidation von Abwasser und Klarschlamm mit Sauerst-
È
o€). Ph.D thesis, Swiss Federal Institute of Technology
ETH, Zurich, Switzerland, Diss No. 12,677, published
1999 as VDI-Fortschrittsbericht Reihe 3, Nr. 576, ISBN
3-18-357603-1.
CONCLUSIONS
. Whenever feasible, measure both COD and TOC
(or DOC).
. Obtain an estimate of the inorganics that contrib-
ute to the COD by checking for some common
inorganic substances.
. The MOC concept allows one to discriminate
between oxidative/reductive chemical, and physi-
cal processes.
. MOC values reveal inaccuracies in TOC (DOC)
and especially COD measurements.
Janicke W. (1983) Chemical Oxidizability of Organic
Water Constituents (Chemische Oxidierbarkeit Orga-
nischer Wasserinhaltssto€e). WaBoLu-Beriche, Dietrich
Reimer Verlag, Berlin.
Kang Y. W., Cho M. J. and Hwang K. Y. (1999) Correc-
tion of hydrogen peroxide interference on standard
chemical oxygen demand test. Water Research 33(5),
1247±1251.
Keith L. H. (ed.) (1996) Compilation of EPA's Sampling
and Analysis Methods, 2nd ed. CRC Press, Boca Raton.
Mantzavinos D., Livingston A. G., Hellenbrand R. and
Metcalfe I. S. (1996) Wet air oxidation of polyethylene
glycols; mechanisms, intermediates and implications for
integrated chemical-biological wastewater treatment.
Chemical Engineering Science 51(18), 4219±4235.
Streitwieser Jr A. and Heathcock C. H. (1985) Introduction
to Organic Chemistry, 3rd ed. Macmillan, New York.
Stumm W. and Morgan J. J. (1981) Aquatic Chemistry,
2nd ed. John Wiley & Sons.
Vogel F. (1997) Wet Oxidation of Phenol with Oxygen at
Mild Conditions (Nassoxidation von Phenol mit Sauerst-
o€ bei milden Bedingungen). Ph.D thesis, Swiss Federal
Institute of Technology ETH, Zurich, Switzerland, Diss.
No. 12,267, published 1998 as VDI-Fortschrittsbericht
Reihe 3, Nr. 533, ISBN 3-18-353303-0.
. When the composition of the waste water is
known, i.e. when investigating model substances,
MOC values furnish valuable information on
reaction development and formation of inter-
mediates.
. The MOC concept is applicable to sludges as
well. In such instances special care must be taken
with TOC and COD analyses in order to achieve
reproducible measurements with acceptable scat-
ter.
. MOC values are a valuable tool for validating
sets of corresponding TOC/COD measurements
(consistency check).
. MOC values ®ll a gap between single substance
analyses yielding a high degree of chemical infor-
mation and global parameters with a low degree
of chemical information.
AcknowledgementsÐThe authors would like to thank
Kommission fur Technologie und Innovation (KTI),
Kanton Basellandschaft, and Bertrams AG, Switzerland,
for the ®nancial support of this work under grant no.
2617.2. Special thanks to Joanna DiNaro (MIT) who criti-
cally read the manuscript.
Vogel F., Harf J., Hug A. and Rudolf von Rohr P. (1996)
Oxidative degradation of phenol using molecular oxy-
gen. In Third International Conference on Advanced Oxi-
dation Technologies for Water and Air Remediation
AOTs-3, Cincinnati, OH, Science & Technology Inte-
gration.
Wol€ C. J. M. (1975) COD determination of volatile com-
pounds. Water Research 9(11), 1015.
REFERENCES
APHA. (1995a) Standard Methods for the Examination of
Water and Wastewater. American Public Health Associ-
ation, Washington DC, 19th Edition, Method 5310
Total Organic Carbon (TOC).
APPENDIX. DERIVATION OF EQUATION (7)
APHA. (1995b) Standard Methods for the Examination of
Water and Wastewater. American Public Health Associ-
ation, Washington DC, 19th Edition, Method 5220
Chemical Oxygen Demand (COD).
Baker J. R., Milke M. W. and Mihelcic J. R. (1999) Re-
lationship between chemical and theoretical oxygen
demand for speci®c classes of organic chemicals. Water
Research 33(2), 327±334.
We consider an organic compound A with a sum for-
mula CnHbOcNd. The organically bound nitrogen is
assumed to have an oxidation state of 3. The oxidation
of the compound A under the conditions of the COD de-
termination (no NOx formation) can be written as:
CnHbOcNd ‡ e O2 4n CO2 ‡ f H2O ‡ d NH3
ꢀ11†
Crain N., Tebbal S., Li L. and Gloyna E. F. (1993) Kin- Note that the ammonia is actually present in its proto-
etics and reaction pathways of pyridine oxidation in nated form as ammonium. For the sake of simplicity the
supercritical water. Ind. Engng Chem. Res. 32(10), 2259± protonation of ammonia has not been included here, for
2268. this does not in¯uence the following derivation. Important
DIN. (1980) DIN 38 409 Teil 41, Deutsche Einheitsverfah- is only the fact that the nitrogen does not change its oxi-
ren zur Wasser-, Abwasser- und Schlammuntersuchung, dation state.
Summarische Wirkungs- und Sto€kenngrossen (Gruppe The stoichiometric coecients for water, f, and for oxy-
H), Bestimmung des Chemischen Sauersto€bedarfs gen, e, are:

2702
Frederic Vogel et al.
b
3d
The theoretical oxygen demand ThOD can be expressed in
f ˆ
ꢀ12†
2
terms of the molar concentration of the compound, [A],
and the stoichiometric coecient for its oxidation, e,
according to equation (11)
2n ‡ f
c
2n
c
b
3d
e ˆ
ˆ
‡
ꢀ13†
2
2
4
ThOD ˆ e‰AŠMO
ꢀ19†
2
The balance for the oxidation numbers of the neutral mol-
ecule A can be written as
The same can be written for the total organic carbon con-
tent:
nMOC ‡ b 2c 3d ˆ 0
ꢀ14†
assuming oxidation numbers of MOC for carbon, +1 for
hydrogen, 2 for oxygen, and 3 for nitrogen. Rearrang-
ing gives
TOC ˆ n‰AŠMC
The ratio ThOD/TOC gives
ꢀ20†
b
3d ˆ 2c nMOC
ꢀ15†
This expression can be substituted into equation (13), giv-
ing
ThOD
TOC
eMO
n 4 MOC MO
2
2
ˆ
ˆ
ꢀ21†
nMC
n
4
MC
2n
c
2c nMOC
e ˆ
‡
ꢀ16†
ꢀ17†
ꢀ18†
2
4
Solving for MOC gives
c
c
nMOC
MC ThOD
e ˆ n
‡
MOC ˆ 4
4
ꢀ22†
2
2
4
MO TOC
2
4
MOC
corresponding to equation (7) using the assumption
CODO ˆ ThOD:
e ˆ n
4