CARBON 4 8 ( 2 0 1 0 ) 1 2 5 2 1 2 6 1

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/carbon

Standardization of the Boehm titration. Part I. CO2 expulsion

and endpoint determination

Sarah L. Goertzen, Kim D. Theriault, Alicia M. Oickle, Anthony C. Tarasuk,

Heather A. Andreas *
Department of Chemistry, Dalhousie University, Halifax, Nova Scotia, Canada B3H 4J3

A R T I C L E I N F O A B S T R A C T

Article history:
Received 16 September 2009
Accepted 27 November 2009
Available online 2 December 2009
The various steps required in the Boehm titration (CO2 removal, agitation method, end-
point determination, etc.) are carried out in different ways by different research groups,
making a comparison of the results between these groups difficult. Herein, the methods
of CO2 expulsion and endpoint determination for the Boehm titration were standardized.
Blank samples of the three Boehm reaction bases, NaHCO3, Na2CO3 and NaOH, were exam-
ined for complete CO2 expulsion through sparging with an inert gas (N2 or Ar), heating, or
utilizing a N2-filled glove box. Boehm titrations using NaOH as the reaction base were stud-
ied through direct titration and back-titration. It was found that to minimize errors both
the NaOH titrant and HCl should be standardized prior to titration and that a back-titration
is preferable for all three reaction bases. Additionally, the titration should be performed
immediately after degassing for 2 h with N2 or Ar, and degassing should continue during
the titration. This is found to be particularly true of the NaOH reaction base, where the
effects of dissolved CO2 are the most noticeable and persistent. With sufficient CO2
removal, there is no significant difference between pH electrode or colour indication end-
point determination, and either is satisfactory.
2009 Elsevier Ltd. All rights reserved.

1. Introduction
Traditionally, the so-called Boehm titration has been used
as a chemical method to identify oxygen surface groups on
carbon materials [18]. The Boehm titration works on the
principle that oxygen groups on carbon surfaces have differ-
ent acidities and can be neutralized by bases of different
strengths. Sodium hydroxide (NaOH) is the strongest base
generally used, and is assumed to neutralize all Brnsted
acids (including phenols, lactonic groups and carboxylic
acids), while sodium carbonate (Na2CO3) neutralizes carbox-
ylic and lactonic groups (e.g. lactone and lactol rings) and so-
dium bicarbonate (NaHCO3) neutralizes carboxylic acids [2,5].
The difference between the uptake of the bases can be used to
identify and quantify the types of oxygen surface groups pres-
ent on a carbon sample.
Although the Boehm titration is widely used, it has not
been standardized. Explicit methodology has not been speci-
fied and certain titration steps vary between research groups.
Reference is typically made to Boehms papers [1,8], but his
procedure is often not followed precisely. The lack of stan-
dardization makes it difficult to compare literature results.
Some of the factors needing to be standardized are the ra-
tio of carbon to reaction base, the length of time that samples
are shaken or stirred, the method of CO2 expulsion from the
solutions, and the method of titration endpoint determina-
tion. All these variables introduce error to varying degrees,
and since different methods are being used, it makes compar-

* Corresponding author: Fax: +1 902 494 1310.

E-mail address: heather.andreas@dal.ca (H.A. Andreas).
0008-6223/$ - see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carbon.2009.11.050
 CARBON 4 8 ( 20 1 0 ) 1 2 5 212 6 1
ing results difficult. Therefore, the Boehm titration must be
standardized not only for consistency but to ensure that the
methodology gives the most precise results possible.
In particular, the issues addressed here are the optimum
method for CO2 removal and the best method of endpoint
determination. Factors such as agitation method and dura-
tion, effect of filtering, titrant concentration and error calcula-
tions will be carried out on real carbon samples in Part II of
this research. With respect to CO2 expulsion, in many previ-
ous cases where Boehms titration was performed, no men-
tion is made of how CO2 was expelled from the samples,
although CO2 expulsion may be accomplished through boiling
(or refluxing) [1,2,5,8] or degassing with N2 during reaction
and titration [3,9], conducting the titration in a N2 atmosphere
in a glove box [4], or subtracting the value of a blank titration
(no carbon) from the results [4,9]. Both degassing with N2 and
heating cause the volume of the acidified aliquot to decrease
due to evaporation, a factor that must be taken into account
when titrating. Related to this, the length of time that CO2
is expelled is significant and was also one of the factors
addressed.
In terms of endpoint determination for the titrations, com-
monly used methods include colour indicators (such as phe-
nolphthalein, methyl red and methylene blue), one-point
pH measurements (where the titration is carried out until a
value of 7 is reached) or potentiometric measurement
throughout the titration for the production of a titration
curve. As long as the same method is being used consistently,
the number of acidic groups on different carbons can be com-
pared. However, it is unclear whether comparisons can be
made between the different methodologies for endpoint
determination since different endpoints are determined for
the different methodologies. For example, phenolphthalein
is often used for a basic titrant [1015], and theoretically
changes colour at pH 8.0. Methyl red is often used for an
acidic titrant [11,14,1619], and theoretically changes colour
at pH 6.2. Notice that the equivalence point, which should
be pH 7.0 in a strong acid-strong base titration, cannot be seen
with either of these indicators. Therefore, while the use of
both these indicators is reasonable, they result in an excess
either of base or acid rather than the equivalence point. If re-
sults are compared using the same endpoint determination
method, these excesses will cancel out, but neither colour
indication method will give identical results with the other
nor with a pH 7.0 endpoint-measured titration. Additionally,
it has not yet been shown whether these differences in equiv-
alence point lead to significant differences in the number of
surface functionalities determined by the titration. Colour
indicator endpoint methods are also prone to human subjec-
tivity. Conversely, a potentiometric titration utilizes a pH me-
ter, which can be read to a pH of exactly 7.0, the expected
endpoint, and is less sensitive to human subjectivity. Boehm
[5] suggests using either potentiometric titration or a mixture
of methyl red and methylene blue indicators. The titration
curve of a strong acid with a strong base is close to being ver-
tical near the equivalence point, and so differences in end-
points observed using difference indicators should not be
significant except when using very dilute concentrations.
Nevertheless, endpoint determination methods and the effect
1253
of the different endpoints determined were examined in this
study.
2. Experimental
For the majority of this work, blank samples (i.e. no carbon
present) were used in the titrations. This removes the varia-
tion that may arise due to differences in carbon surface func-
tionalities, carbon wettability or carbon provenance. In
addition, it makes the amount of base reacted (and therefore
the amount of carbon surface functionality) expected to be
zero, so that bias can be easily seen. To prevent confusion,
the bases used in the titration to react with the acidic surface
functionalities (NaHCO3, Na2CO3 and NaOH) are denoted as
reaction bases, and the base used for titration is denoted as
the titrator base.
2.1. Boehm titration general procedure
The general titration procedure was carried out based on
Boehms method [5]. A known mass of ca. 1.5 g carbon was
added to 50.00 mL of one of three reaction bases of 0.05 M
concentration: NaHCO3 (Sigma, 99.5%), Na2CO3 (Anachemia,
ACS Reagent) and NaOH (Sigma-Aldrich, 99.998%). In Boehms
early studies [8] he stated that 0.05 N solutions should be
used; however, later on Boehm stipulated that 0.05 M (molar)
solutions should be used, which is particularly significant for
Na2CO3, which is a diprotic base [5]. In the present study,
therefore, 0.05 M solutions were used. The samples were agi-
tated by shaking for 24 h and then filtered to remove the car-
bon, and 10.00 mL aliquots were taken by pipette from the
samples. The study of the possibility that base adsorbed onto
the filter paper will be addressed in a paper in preparation in
the authors lab. The aliquots of the reaction base NaHCO3
were then acidified by the addition of 20.00 mL of 0.05 M
HCl (SigmaAldrich, 99.999%). The aliquots of the reaction
base Na2CO3 were acidified by the addition of 30.00 mL of
0.05 M HCl, to ensure complete neutralization of the base,
which requires two protons, vs. the one proton required by
the NaOH and NaHCO3 reaction bases. The acidified solutions
were then back-titrated with 0.05 M NaOH, the titrator base.
The aliquots of the NaOH reaction base were titrated using
two different methods: (1) the solution was acidified by the
addition of 20.00 mL of 0.05 M HCl and the sample was
back-titrated with 0.05 M NaOH; or (2) the sample was titrated
directly with 0.05 M HCl. All pipettes used should be properly
calibrated. Fresh solutions were used to limit the amount of
CO2 permeating through polyethylene storage bottles and dis-
solving into the reagents. The direct titration of NaOH is de-
noted as NaOHdirect. In all cases, CO2 was removed from
solution immediately before the titration.
The carbon to base ratio was chosen based on the require-
ment that at least ca. 10% of the reaction base should react
with the carbon, which according to Boehm [5] is needed to
ensure reliable results. A carbon black was used: Black Pearls
2000 (Cabot Corporation). Blank samples, containing no car-
bon, were also run with each reaction base. All solutions were
prepared with 18 MX water. All titrations were carried out at
room temperature.
1254 CARBON 4 8 ( 2 0 1 0 ) 1 2 5 2 1 2 6 1
2.2. Endpoint determination
To examine which method of endpoint determination was
most precise, titrations were done using both a pH electrode
and meter and colour indicators separately and simulta-
neously. The potentiometric examinations were done as a
one-point endpoint determination where the titration was
carried out until a pH of 7.0 was measured, at which point
the titration volume was noted. As well, full potentiometric
titrations were carried out and the endpoint was determined
from the first derivative of the pH-volume plot. The indicators
used were methyl red (Fluka), methylene blue (Riedel-de
Haen) and phenolphthalein. Methyl red was used in basic
samples titrated with HCl, and phenolphthalein was used in
acidic samples titrated with NaOH.
2.3. Standardization of solutions
To ensure a correct NaOH concentration was used in the cal-
culation of the carbon surface functionalities, all NaOH solu-
tions were standardized, since solid NaOH is hygroscopic and
solutions made with NaOH pellets are less concentrated than
what is first assumed. Dried potassium hydrogen phthalate
(KHP) (Sigma-Aldrich, 99.95%) (stored in a desiccator) was
the primary standard used to standardize NaOH. A mass of
ca. 0.2 g of KHP, measured to 0.01 mg, was used in each titra-
tion. This mass was chosen as it resulted in the amount of ti-
trant used to be about 80% of the 25-mL burette, as is usually
required in analytical titrations. Since a titration curve
showed that KHP is a weak acid, with an endpoint of ca. 8.6,
phenolphthalein is a good choice of indicator. No more than
2 drops of phenolphthalein per 40 mL were added to avoid a
bias. The standardized NaOH was then used as a titrant in
the standardization titration of the HCl solution to an end-
point of pH 7.0, conducted using a pH electrode and meter.
Sealed ampoules of a NaOH standard may be used to avoid
the necessity of standardization.
2.4. CO2 expulsion methods
2.4.1. Degasification with N2 or Ar
Studies on the effect of N2 and Ar degasification (sparging) on
CO2 removal were conducted on blank solutions, where no
carbon was used. These solutions were a mixture of
10.00 mL of reaction base and the appropriate volume of HCl
to achieve the volume of a typical Boehm titration aliquot
(as per Section 2.1). Samples were contained in 40 mL glass
vials utilizing polytetrafluoroethylene (PTFE)/silicone septum
lids. The inert gas, either N2 or Ar, was bubbled into the vial
through a needle submerged in the solution, and another nee-
dle was positioned in the vial head space for use as an ex-
haust. Bubble size was ca. 1 mm in diameter. Bubbling rate
was less than 1 mL/min; ca. 23 bubbles/s. Small bubbles were
desired due to the high surface area to volume ratio [20]. The
time of degasification was varied from 0 to 24 h, to test what
is the minimum time required for CO2 removal.
After degasification, the samples were transferred to a bea-
ker that had been purged with the inert gas and covered with
Parafilm. This was done to prevent introduction of environ-
mental CO2 into the sample. Each sample was then titrated
with 0.05 M NaOH. With some samples, the degasification
was continued throughout the titration by bubbling the sam-
ple. During the titration, a seal of Parafilm was maintained
around both the electrode and the burette. Three blank sam-
ples were used for each gas. The mass and pH of the solution
was recorded at each point in the CO2 removal process. The
endpoint was determined using a pH meter (SympHony,
VWR) with a posi-pHIo electrode (VWR).
2.4.2. Use of a N2-filled glove box
Blank titrations, using blank aliquots of each reaction base as
in Section 2.4.1, were performed in a glove box filled with N2,
with and without degasification prior to titration, using
potentiometric endpoint detection. When degasification was
done, it was by bubbling the aliquot with N2 for 1 h in the
glove box.
2.4.3. CO2 removal through heating
To examine the effects of heat on CO2 removal two methods
were used. In the first method, blank aliquots were made in
small sample vials for each reaction base as in Section 2.4.1.
The vials were sealed except for a needle passing through
the cover to act as an exhaust. The samples were heated at
80 C on a hot plate for 30 min and then sealed and left to cool
to room temperature before titrating. Three replicate samples
of each reaction base were weighed before heating and after
cooling to determine the mass loss.
In the second method, the sample preparation was carried
out as above for heating, but the solutions were placed in 50-
mL round bottom flasks. The samples were refluxed using a
hot water bath at 100 C for 30 min and then sealed and left
to cool to room temperature. The samples were weighed be-
fore heating and after cooling to determine the mass loss
due to evaporation, and again after being transferred to a
small sealed beaker for titration. Three replicate samples
were carried out for each reaction base.
2.5. Calculations
The equations used to determine the quantity of surface
groups depends on the titration method: back-titration or di-
rect titration. For a back-titration, amounts of the acidic
groups on carbon were determined by:
 
nHCl Va
HClVHCl NaOHVNaOH BVB  nCSF 1
nB VB
nHCl VB
nCSF BVB  HClVHCl  NaOHVNaOH 2
nB Va
where [B] and VB are the concentration and volume of the
reaction base mixed with the carbon, providing the number
of moles of reaction base that was available to the carbon sur-
face for reaction with the surface functionalities. nCSF denotes
the moles of carbon surface functionalities on the surface of
the carbon that reacted with the base during the mixing step.
Va is the volume of the aliquot taken from the VB, and [HCl]
and VHCl are the concentration and volume of the acid added
to the aliquot taken from the original sample. This gives the
number of moles of acid added to the aliquot, and available
for reaction with the remaining reaction base. The remaining
 CARBON 4 8 ( 20 1 0 ) 1 2 5 212 6 1
moles of acid are then determined through the titration using
[NaOH] and VNaOH, the concentration and volume of the ti-
trant in the back-titration. Thus, through the knowledge of
the remaining moles of acid, leading to the amount of reac-
tion base remaining after reaction, and by difference (know-
ing the total amount of reaction base available initially) the
amount of base reacted with the surface functionalities, the
surface functionalities can be quantified. These equations
are based on the calculations from Chen andWu [21] but
modified to take the molar ratio of acid to base nnHCl
B
into ac-
count to allow for monoprotic vs. diprotic reaction bases.
When NaOH is used in the Boehm titration and is directly
titrated with HCl, the following equation is used:
VB
nCSF BVB  HClVHCl 3 respectively, and when NaOH was titrated with HCl, RSD val-
Va
with a similar description of the reaction calculations as
above.
Volumes used in calculations of nCSF were not the pipette
volumes but were rather determined by the mass and the
density of the solution at room temperature.
The quantity of the different possible surface groups (phe-
nols vs. lactonic groups vs. carboxyl groups) are calculated
through the difference in the calculated amount of surface
functionality reacted (nCSF). NaOH reacts with all surface
groups, (carboxylic acids/carboxyls, lactones/lactols, and phe-
nols), and will therefore have a nCSF that includes all of these
groups. Na2CO3 reacts with carboxyl and lactonic groups and
the difference between the nCSF measured with NaOH and the
nCSF measured with Na2CO3 will denote the number of phe-
nols on the surface. Similarly, since NaHCO3 reacts only with
carboxylic groups, the difference between Na2CO3 and NaH-
CO3 are the number of lactonic groups. The number of car-
boxylic acid groups are found directly from the NaHCO3
reacted. For blank solutions, without carbon, nCSF should be
equal to zero since there is no carbon present and therefore
no carbon surface functionalities to react with the reaction
base.
The amounts of carbon surface functionalities, nCSF, pre-
sented in this paper are given with error ranges that are equal
to one standard deviation (SD) above and below the stated va-
lue. The standard deviations are based on a minimum of
three data points.
2.6. Experiments with carbon
Although the majority of the titrations discussed in this paper
were carried out with blank samples (i.e. no carbon present),
for some studies carbon was used in the titration. Samples of
carbon black were prepared as in Section 2.1. CO2 was re-
moved by degassing for 2 h with N2 bubbling. Potentiometric
back-titrations were performed while continually degassing
the acidified aliquots with N2.
3. Results and discussion
3.1. Endpoint determination
To ascertain the optimal method for endpoint determination,
colour indicators and a pH electrode and meter were used
1255
with strong acid (HCl)strong base (NaOH) titrations, which
theoretically have an equivalence point at pH 7.0. Boehm sug-
gests using an indicator mixture of methyl red and methylene
blue [5]. The methylene blue is added to make the endpoint
significantly easier to observe as the typical yellow (alkali)
to red (acidic) transition of the methyl red is, with the meth-
ylene blue, a green to purple transition with a grey-coloured
endpoint. However, this indicator mixture is not often taken
advantage of, and therefore the more commonly used indica-
tors were examined in this study: phenolphthalein with a ba-
sic titrant [1015], and methyl red when the titrant was acidic
[11,14,1619]. When HCl was titrated with NaOH, Relative
Standard Deviation (RSD) values of the titrant volume were
0.98% and 0.06% for phenolphthalein and a pH electrode,
ues of 0.26% and 1.3% were obtained using methyl red and a
pH electrode. These low RSD values for all endpoint detection
methods suggest that any method would result in a relatively
high precision, and provides precise enough measurements
to be used in these titrations. Nevertheless, these results do
not indicate the accuracy of the measurements. For instance,
reproducible results are possible based on human observation
of colour changes, but since observation is subjective, results
could change drastically depending on the experimentalist.
Additionally, the endpoints of the colour indicators are not
at a pH of 7.0, as desired for the strong acid-strong base titra-
tions used in the Boehm titration. It must be determined
whether the volume differences which result from these end-
point differences are significant. Conversely, the use of a pH
meter allows for an objective endpoint determination, rather
than subjective, and can result in endpoints at a pH of 7.0.
Calculations were also done to determine whether the col-
our indicators inflated the number of acidic surface groups vs.
potentiometric endpoint of 7. This would have a specific im-
pact on the number of carboxylic groups calculated from
the reaction of carbon with NaHCO3. If the experimentalist
is consistent in the colour at the endpoint for the colour indi-
cators, the amounts of lactonic groups and phenols would be
the same, since the calculation only takes into account differ-
ences of reactions with Na2CO3 and NaOH. If phenolphthalein
is used for endpoint determination, at a pH of 8.9 (the exper-
imental pH endpoint found for phenolphthalein), a difference
in nCSF of 1 lmol is expected compared to a potentiometric
endpoint of pH 7.0. An even smaller difference (0.1 lmol) oc-
curs for a colour indicator endpoint of pH 5.9 (the experimen-
tal pH endpoint found for methyl red) when titrating with HCl
and using methyl red as an indicator. Given that nCSF values of
502000 lmol/g [1,22] are typical for a Boehm titration, this
suggests that the bias introduced by the use of a colour indi-
cator vs. a potentiometric measurement is negligible and any
of these endpoint detection methods are appropriate for the
Boehm titration.
Additionally, the first derivative was examined as a meth-
od to determine where the equivalence point occurred, vs. the
endpoint. The first derivative shows the point that has the
steepest slope. This is more exact than attempting to deter-
mine the equivalence point by observation of a colour indica-
tor, by assuming the equivalence point occurred at a pH of 7.0,
or by inspection of the potentiometric curve. Therefore, for
the most exact results of the equivalence point of a titration,
1256 CARBON 4 8 ( 2 0 1 0 ) 1 2 5 2 1 2 6 1
potentiometric titrations with a first derivative analysis
should be used. However, recording of a titration curve and
the mathematical manipulation to derive the first derivative
is more time consuming than the one-point potentiometric
determination or use of colour indicators. Thus, if exact mea-
surements are not required (e.g. for comparison of carbons
with large changes in surface functionalities), any of these
endpoint detection methods can be used.
The important consequence of the previous discussion is
that all of these endpoint detection methods give the same
results in the Boehm titration since neither the accuracy
nor the precision is greatly influenced by the different meth-
ods. The endpoint detection method should, nevertheless,
still be reported and error ranges should be included in re-
ported values to ease the comparison between relatively close
values.
Included in these error ranges should be a consideration
of the precision of the experimentalist and the burette. This
will be covered in more detail in a paper in preparation from
the authors lab; nevertheless a few points will be high-
lighted here. Sample calculations using Eq. (2) have been
done to show the impact of missing the endpoint has on
the final nCSF (in lmol). Assuming the endpoint is surpassed
by a quarter drop (ca. 0.01 mL), the smallest volume which a
human is typically capable of dispensing, this results in an
increase of nCSF of ca. 2 lmol. A 25-mL burette has an error
of 0.02 mL and would result in a variation in the answer
of 4 lmol. These errors in the nCSF should be taken into ac-
count during the propagation of errors used when reporting
the range in nCSF.
3.2. Removal of effect of CO2 through the use of non-
degassed blanks
Often in the literature, the effect of CO2 is removed from non-
degassed samples by the subtraction of the base reacted in
non-degassed blanks [4,12,21,2328]. Thus, the effectiveness
of this subtraction method was also examined. For this meth-
od to be valid, the effect of CO2 in the non-degassed solutions,
as measured by the base reacted, must be consistent there
should be little to no variation in the amount of base reacted
in these samples. Therefore, the nCSF was determined from
three identical samples of non-degassed blanks of the three
reaction bases, in order to determine the reproducibility.
The average nCSF reacted of these non-degassed blanks is seen
in Table 1. The positive values for the non-degassed blanks re-
flect the effect of dissolved CO2 in the Boehm titration (recall
that since these are blanks, no carbon is used, and therefore
the expected value for the nCSF is 0 lmol). From the standard
deviation given for the data of the non-degassed blanks, it can
be seen that there is significant variation in the nCSF with
these identical samples. This suggests that the effect of CO2
is not consistent between samples and would further suggest
that the use of the blank subtraction method to correct results
from non-degassed samples may lead to erroneously high or
low results since there is a significant variation in the effects
of CO2, and therefore it is recommended that blank subtrac-
tion method not be used during quantification using the
Boehm titration.
3.3. Effect of degasification time
It is well known that dissolved gases (e.g. CO2 and O2) can be
removed from solution through sparging with an inert gas,
such as N2 or Ar. However, there is no standardized time for
bubbling of the inert gas, nor any information available for
the length of bubbling time required for effective CO2 removal
from the Boehm titration. Fig. 1 shows the potentiometric
titration curves for sample blanks (no carbon) for each of
the reaction bases with different degasification times. In
Fig. 1ad, pH is plotted vs. the calculated nCSF that reacted
in the sample (which should be zero at the endpoint).
Fig. 1e shows the titration curve for the direct titration of
NaOH (NaOHdirect) vs. the volume of titrant added, to illustrate
the difference in shape of the titration itself compared to the
back-titrations. For ease, a dotted line is shown on each figure
which shows where a pH of 7 falls on the curve. The predicted
Boehm titration curves for both the back-titration, where the
acidified solutions are titrated with NaOH, and direct titra-
tion, where the basic solution is directly titrated with HCl,
are expected to be typical monoprotic acid titration curves,
with an endpoint of pH 7.0. However, as is well known, the
dissolution of CO2 into an aqueous solution, even permeating
through polyethylene storage bottles, causes the pH to shift
[29]. The effect of this in the current titrations is the appear-
ance of a second endpoint, as is evident from the presence
of two inflection points marking two endpoints (at ca. pHs
of 4 and 8), as seen in Fig. 1, with the solutions that have
not been sufficiently degassed. This can be also easily seen
in the first derivative plots of these curves, seen in Fig. 2. As
gaseous CO2 enters solution, it dissolves as in reaction 1:
CO2 g  CO2 aq pK 1:463 1
and the dissolved CO2 reacts with water to form carbonic
acid, which then becomes bicarbonate (Reactions 2 and 3):
CO2 H2 O  H2 CO3 aq pK 2:699 2
H2 CO3 g  HCO
3 H aq pKa 6:352 3
The presence of the two inflection points in the titration
curve may have a significant impact on the calculated nCSF,
particularly if one is using a one-point potentiometric end-
point detection or colour indicators, as a pH of 7.0 is reached
later in the titration resulting in a falsely high calculated
amount of surface functionality; for instance, a measured in-
crease in volume of ca. 0.5 mL corresponds to a false increase
in nCSF of ca. 90 lmol. Clearly it is imperative that the CO2 be
removed before the titration.
The minimum degasification time required for the re-
moval of dissolved CO2 from all the titration samples with
an inert gas was examined, and is shown in Fig. 1. Since there
was no significant difference in degasification time between
N2 and Ar, only the N2 results are shown. As the degasification
time increases, less evidence for dissolved CO2 is seen as the
second endpoint begins to disappear, and the nCSF that sup-
posedly reacted decreases. With NaHCO3 a minimum of
1.5 h of degasification is required to completely remove the
effects of CO2 (see Fig. 1a). Approximately 2 h is required for
the Na2CO3 sample (see Fig. 1b). A back-titration of an NaOH
blank after degassing for 2 h is shown in Fig. 1c. The effects
 CARBON 4 8 ( 20 1 0 ) 1 2 5 212 6 1 1257

Table 1 Comparison of calculated nCSF to determine the effectiveness of each CO2 removal.
Titration nCSF SD (lmol)
NaHCO3 Na2CO3 NaOH NaOHdirect

Non-degassed 110 48 29 20 7 13
N2 Prior (2 h) 5 3 23 23 2 20 35 42
N2 Prior and during 11 11 1 4 2
In N2-filled glove box 13 33 14 17 1 1 10 6
N2 Prior to titration in glove box 8 1 2 1 15 1 12 2
Heating 24 16 13 8 40 26
Reflux 3 17 57 19 8 23

Fig. 1 Titration curves (plotted as amount of carbon surface functionality) for different reaction bases: (a) NaHCO3; (b) Na2CO3;
(c) NaOH back-titration; (d) NaOH direct titration; and (e) titration curve for NaOH direct titration plotted versus titrant volume.
Different times of degassing or degasification during titration are noted in the legends.
1258 CARBON 4 8 ( 2 0 1 0 ) 1 2 5 2 1 2 6 1
Fig. 2 Sample first derivative of a titration curves (plotted
as amount of carbon surface functionality) for Na2CO3 as in
Fig. 1b for degasification time of 1 h (dots) and 2 h (dashes).
of CO2 were not completely removed, and so the NaOH blank
was degassed for 2 h prior to titration, and the degassing was
maintained during the titration (Fig. 1c). This was carried out
for both the direct titration and the back-titration of NaOH,
the latter shown in Fig. 1d. It was seen that CO2 effects were
easily removed from the titration curve by 2 h of degasifica-
tion with N2 prior to titration, with the addition of continuous
degasification during for the back-titration of NaOH. In the di-
rect titration of the NaOH aliquots, the most often used
Boehm titration for the NaOH reaction base, even 24 h of
degasification does not completely remove the second end-
point (Fig. 1d). This suggests that for the complete removal
of the effects of CO2 a back-titration should be carried out
with the NaOH samples as well, since degasification in the
back-titration takes place in an acidified solution, where the
CO2 exists as dissolved CO2, which can be removed through
degasification.
Thus, the optimum conditions for the removal of CO2 ef-
fects are 2 h of degasification prior to titration followed imme-
diately by a back-titration which is degassed throughout.
Though it seems that degasification during titration is not re-
quired for NaHCO3 and Na2CO3 the same conditions are sug-
gested for consistency.
Endpoints determined from three identical titrations for
each reaction base were used to calculate the nCSF reacted
(Table 1). Since these samples are blank, thus containing no
carbon, it is expected that no base should have reacted, giving
a result of 0 lmol of nCSF. The aliquots that were degassed
with N2 for 2 h before titration but not degassed during titra-
tion are denoted as N2 Prior. The aliquots degassed with N2
prior to (2 h) and during the titration are denoted as N2 Prior
and during.
Although the results from titrations where the solution
was degassed only prior to titration (not during titration)
show little bias, this is because of the large standard devia-
tions in the results. Conversely, the titrations wherein the
degasification continued throughout the titrations evidenced
little or no bias and very small standard deviations. Thus, this
supports the conclusions drawn from Fig. 1 that bubbling the
solution for two hours prior to the titration combined with
bubbling during the titration is the optimum method of CO2
removal using degasification with an inert gas.
3.4. Removal of effect of CO2 through titration in a N2-
filled glove box
Blank back-titrations were performed in an N2-filled glove box
to determine whether titrations done in an inert atmosphere
would remove the effect of CO2 and the results are shown in
Table 1. From this data, it can be seen that doing the titration
in a N2-filled glove box (with no bubbling), denoted as In N2-
filled glove box, results in acceptable results (ca. 0 lmol for
these blank titrations), although the standard deviations are
relatively high. The standard deviations are much lower if
the solution is degassed with N2 in the glove box prior to titra-
tion, but this seems to result in a small negative bias. Never-
theless, this method may be sufficient to remove the effect of
CO2.
Additionally, potentiometric titration curves, such as that
seen in Fig. 3, show that simply leaving the aliquot open un-
der the inert atmosphere (no direct degasification with N2)
was sufficient to remove the effect of CO2 since only one
inflection point is in evidence. This indicates that performing
the Boehm titration in a glove box is a good method; however,
a small negative bias was again seen, and this method is more
complex and more expensive (in terms of the purchase of a
glove box and the volume of inert gas used) than the bubbling
regime proposed earlier (Section 3.3).
3.5. Removal of effect of CO2 through heating
In the majority of his papers, Boehm suggests that heating of
the aliquots is sufficient to remove the CO2 from solution
[1,2,5]. To determine the effectiveness of heating samples to
remove CO2, blanks were heated and cooled, and then titrated
to an endpoint of pH 7.0. The nCSF calculated based on three
titrations of aliquots heated in a vented vial or refluxed under
an open, water cooled condenser are shown in Table 1. In this
table it can be seen that heating of the solution results in a
large positive bias in the NaHCO3 results, and negative biases
in the Na2CO3 and NaOH results, as well as a large standard
Fig. 3 Back-titration in an N2-filled glove box of NaHCO3
blank. The dotted line shows where pH 7.0 occurs.
 CARBON 4 8 ( 20 1 0 ) 1 2 5 212 6 1
deviation with all reaction bases. The large positive bias with
the NaHCO3 reaction base suggests that heating may not be
sufficient to remove the CO2. This is confirmed in Fig. 4a by
the two inflection points in the titration curve and the two
peaks in the first derivative plots, as would be expected for
the diprotic titration of carbonic acid from dissolved CO2. This
residual CO2 may have been trapped in the head space of the
flask during heating and dissolved back into solution during
cooling and the large standard deviation is likely due to the
variability of the quantity of trapped CO2. However, the con-
cern with heating in an open flask is that a significant amount
of the aliquot will be lost due to evaporation. This evaporation
may lead to a negative bias, as HCl evaporates from solution
with the water. When an open flask was heated to 80 C for
2 h, with N2 bubbling through solution, a negative bias of
12 lmol was seen, indicating that HCl was leaving solution
during this heating step. This is likely the cause of the nega-
tive biases in the Na2CO3 and NaOH results when heated, and
the variability in amount of HCl evaporated may be the cause
for the large standard deviations seen with these samples.
30 min of heating at ca. 80 C while bubbling with N2 was suf-
ficient to remove the effect of CO2 on the titration curve (not
shown), which is a significantly shorter time than simply
degassing with N2 (Section 3.3). However, a negative bias of
5 lmol was evidenced, likely from a small amount of HCl
evaporation, and a significant volume loss was seen, ca. 30 %
of the aliquot, which makes the pH measurements difficult
as only a very small volume remains after heating when the
typical 10 mL aliquot volume is used. These results show that
merely heating the solutions is not an adequate method for
the removal of CO2 or sufficient care must be taken to ensure
the head space is completely flushed of CO2 prior to cooling.
Additionally, heating may cause a loss of HCl from the acidi-
fied aliquots leading to a negative bias in the results.
Two inflection points are also seen in the titration curve
with the aliquots which were refluxed (see Fig. 4b), showing
that refluxing the samples may not be sufficient to remove
CO2 effects, and this may be the reason for the large positive
bias seen in the Na2CO3 results in Table 1. Although there is
no evident bias in the data for refluxing with NaHCO3 and
NaOH in Table 1, this is because of the extremely large stan-
Fig. 4 Titration curve (solid) and first derivative of titration curve (dotted) of NaHCO3 blank aliquot after removal of CO2
through (a) heating method and (b) refluxing method.
1259
dard deviations which result from this method. It is likely that
the large standard deviations for both the reflux method and
the heating method can be explained by the large mass
changes experienced during the heating steps. Although
some mass loss from evaporation is seen when the CO2 is re-
moved from the aliquot through bubbling, the mass losses
evidenced by heated samples is significantly higher, while a
mass gain (ca. 3 g) was recorded for refluxed samples and this
unexpected mass gain may be attributed to high air humidity
experienced in the geographic region in which these experi-
ments were conducted, which may result in extra water con-
densing from the air as it touches the cool reflux condenser
and adding to the solution contained in the round bottom
flask. This water may introduce CO2 or impurities into the
sample and, since water content in the air varies substantially
by the day and region, it is a difficult parameter to control.
With the positive and negative biases and the large varia-
tion seen from the tabulated SD values, it is suggested that
heating the samples is not a reproducible CO2 removal meth-
od, and should not be used. Additionally, the titration curves
in Fig. 4 showed that the effect of CO2 was not removed by
heating nor by refluxing.
3.6. Experiments with carbon
The Boehm titration was carried out on samples of carbon
black, with CO2 removal by degassing for 2 h with N2 prior
to titration, as well as continuous saturation of the aliquots
with N2, to ensure that this method will work in practice. As
expected, the amounts of carbon surface functionalities on
this carbon (Table 2) are all significantly positive, indicating
that these surface functionalities are present. The standard
deviations shown in Table 2 suggest that this method of CO2
removal is adequate for samples containing carbon and result
in data which are reliable. Although the standard deviations
are larger than the blanks using the same method of CO2 re-
moval, this is due to the heterogeneous nature of the carbon
surface. Titration curves (not shown) performed on these
samples using carbon showed no evidence of CO2, confirming
that this degasification procedure is sufficient to remove the
CO2 from solution.
1260 CARBON 4 8 ( 2 0 1 0 ) 1 2 5 2 1 2 6 1
Table 2 Carbon surface functionalities on Black Pearls 2000
carbon.
Functional groups nCSF (lmol/g)
Phenolic 108 12
Lactonic 24 7
Carboxylic 14 3
4. Conclusions
The widespread use of the Boehm titration for the determina-
tion and quantification of the surface functionalities present
on carbon surfaces necessitates a standardized procedure in
order for comparisons of these results to be possible between
research groups and literature results. The standardization
procedure for the titration endpoint determination and the
expulsion of CO2 from the solution was undertaken here. It
has been shown that either colour indication or potentiomet-
ric endpoint determination is acceptable, with no significant
difference in the results between these endpoint determina-
tion methods. Therefore, either method can be used and com-
parisons between results can be made, although it is
suggested that the method used should be noted since this
may affect the precision of the results.
For the titrations itself, it was found that dissolved CO2 in
the solution has a significant effect on the amount of surface
functionality determined by the Boehm titration and there-
fore it must be completely removed in order to ensure accu-
rate and precise results. Since it was found that it is very
difficult to sufficiently remove CO2 from the solutions used
in a direct titration (as is often used for the NaOH reaction
base), a back-titration should be performed for each of the
reaction bases NaHCO3, Na2CO3, and NaOH. The acidified ali-
quots should be titrated immediately after degassing for 2 h
with N2 or Ar, and degassing should continue during the titra-
tion, to ensure sufficient removal of CO2. A glove box may be
used to remove CO2; however, it is likely to lead to a higher
variance in the results and care must be taken to ensure that
no negative bias is seen. CO2 removal methods including the
use of heat (heating or refluxing) are insufficient to fully re-
move the effect of the dissolved CO2, and will result in extre-
mely large variances. These methods should not be used for
CO2 removal. Based on this work, the suggested method for
CO2 removal in the standardized Boehm titration is degassing
of the solutions for 2 hours using an inert gas, coupled with
continued degasification of the solutions during titration.
Acknowledgement
The authors would like to gratefully acknowledge the Natural
Sciences and Engineering Research Council (NSERC) for sup-
port of this work.
R E F E R E N C E S
[1] Boehm HP. Some aspects of the surface-chemistry of carbon-
blacks and other carbons. Carbon 1994;32(5):75969.
[2] Boehm HP. Carbon surface chemistry.. World Carbon; World
of Carbon 2001;1(Graphite and Precursors):14178.
[3] Salame II, Bandosz TJ. Surface chemistry of activated
carbons: combining the results of temperature-programmed
desorption, Boehm, and potentiometric titrations. J Colloid
Interface Sci 2001;240(1):2528.
[4] Guo Y, Rockstraw DA. Activated carbons prepared from rice
hull by one-step phosphoric acid activation. Micropor
Mesopor Mater 2007;100(13):129.
[5] Boehm H. Surface chemical characterization of carbons from
adsorption studies. In: Bottani Eduardo J, Tascon Juan MD,
editors. Adsorption by carbons. Amsterdam: Elsevier; 2008.
p. 30127.
[6] Seredych M, Hulicova-Jurcakova D, Lu GQ, Bandosz TJ.
Surface functional groups of carbons and the effects of their
chemical character, density and accessibility to ions on
electrochemical performance. Carbon 2008;46(11):147588.
[7] Boehm HP, Heck W, Sappok R, Diehl E. Surface oxides of
carbon. Angew Chem Int Ed 1964;3(10):66977.
[8] Boehm HP. Chemical identification of surface groups. In: Eley
DD, Herman Pines, Paul B Weisz, editors. Adv
Catal. Academic Press; 1966. p. 179274.
[9] Barton SS, Dacey JR, Evans MJB. Surface oxides on porous
carbon. Colloid Polym Sci 1982;260(7):72631.
[10] Zhang N, Wang L, Liu H, Cai Q. Nitric acid oxidation on
carbon dispersion and suspension stability. Surf Interface
Anal 2008;40(8):11904.
[11] Kato Y, Machida M, Tatsumoto H. Inhibition of nitrobenzene
adsorption by water cluster formation at acidic oxygen
functional groups on activated carbon. J Colloid Interf Sci
2008;322(2):3948.
[12] Meng G, Li A, Yang W, Liu F, Yang X, Zhang Q. Mechanism of
oxidative reaction in the post crosslinking of hyper-
crosslinked polymers. Eur Polymer J 2007;43(6):27327.
[13] Hermans S, Diverchy C, Demoulin O, Dubois V, Gaigneaux
EM, Devillers M. Nanostructured Pd/C catalysts prepared by
grafting of model carboxylate complexes onto functionalized
carbon. J Catal 2006;243(2):23951.
[14] Machida M, Mochimaru T, Tatsumoto H. Lead(II) adsorption
onto the graphene layer of carbonaceous materials in
aqueous solution. Carbon 2006;44(13):26818.
[15] El-Sheikh AH, Newman AP, Al-Daffaee HK, Phull S, Cresswell
N. Characterization of activated carbon prepared from a
single cultivar of Jordanian olive stones by chemical and
physicochemical techniques. J Anal Appl Pyrol
2004;71(1):15164.
[16] Kalpakli YK, Koyuncu I. Characterization of activated carbon
and application of copper removal from drinking water. Ann
Chim (Rome Italy) 2007;97(1112):1291302.
[17] Otowa T, Nojima Y, Miyazaki T. Development of KOH
activated high surface area carbon and its application to
drinking water purification. Carbon 1997;35(9):13159.
[18] Beker UG, Malik DJ, Strelko Jr V, Streat M. Influence of
activated carbon oxidation treatments on the selective
removal of copper and lead. Chem Eng Commun 2003;190(5-
8):61029.
[19] Strelko Jr V, Malik DJ. Characterization and metal sorptive
properties of oxidized active carbon. J Colloid Interf Sci
2002;250(1):21320.
[20] Burns SE, Yiacoumi S, Tsouris C. Microbubble generation for
environmental and industrial separations. Sep Purif Technol
1997;11(3):22132.
[21] Chen JP, Wu S. Acid/Base-treated activated carbons:
characterization of functional groups and metal adsorptive
properties. Langmuir 2004;20(6):223342.
[22] Salame II, Bagreev A, Bandosz TJ. Revisiting the effect of
surface chemistry on adsorption of water on activated
carbons. J Phys Chem B 1999;103(19):387784.
 CARBON 4 8 ( 20 1 0 ) 1 2 5 212 6 1
[23] Duman G, Onal Y, Okutucu C, Onenc S, Yanik J. Production of
activated carbon from pine cone and evaluation of its
physical, chemical, and adsorption properties. Energy Fuels
2009;23(4):2197204.
[24] Tennant MF, Mazyck DW. The role of surface acidity and pore
size distribution in the adsorption of 2-methylisoborneol via
powdered activated carbon. Carbon 2007;45(4):85864.
[25] de Mesquita JP, Martelli PB, Gorgulho HdF. Characterization of
copper adsorption on oxidized activated carbon. J Braz Chem
Soc 2006;17(6):113343.
1261
[26] Zhao N, Wei N, Li J, Qiao Z, Cui J, He F. Surface properties of
chemically modified activated carbons for adsorption rate of
Cr(VI). Chem Eng. J (Amsterdam,Neth) 2005;115(12):1338.
[27] Dastgheib SA, Rockstraw DA. Pecan shell activated carbon:
synthesis, characterization, and application for the removal
of copper from aqueous solution. Carbon 2001;39(12):184955.
[28] Barton SS. The nature of the reaction between aqueous
potassium hydroxide and the acid surface oxides on porous
carbons. Colloid Polym Sci 1986;264(2):17681.
[29] HarrisDC. Quantitative Chemical Analysis. 6th Ed., 2003; 928.