See

discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/223076652

The Dubinin-Radushkevich equation and

the underlying microscopic adsorption
description
Article in Carbon August 2001
DOI: 10.1016/S0008-6223(00)00265-7 Source: OAI

CITATIONS

READS

67

2,393

2 authors, including:
Duong Do
University of Queensland
519 PUBLICATIONS 9,852 CITATIONS
SEE PROFILE

Available from: Duong Do

Retrieved on: 07 November 2016

PERGAMON

Carbon 39 (2001) 13271336

The DubininRadushkevich equation and the underlying

microscopic adsorption description
C. Nguyen, D.D. Do*
Department of Chemical Engineering, The University of Queensland, St. Lucia, Qld 4072, Australia
Received 10 May 2000; accepted 4 September 2000

Abstract
The DubininRadushkevich (DR) equation is widely used for description of adsorption in microporous materials,
especially those of a carbonaceous origin. The equation has a semi-empirical origin and is based on the assumptions of a
change in the potential energy between the gas and adsorbed phases and a characteristic energy of a given solid. This
equation yields a macroscopic behaviour of adsorption loading for a given pressure. In this paper, we apply a theory
developed in our group to investigate the underlying mechanism of adsorption as an alternative to the macroscopic
description using the DR equation. Using this approach, we are able to establish a detailed picture of the adsorption in the
whole range of the micropore system. This is different from the DR equation, which provides an overall description of the
process. 2001 Elsevier Science Ltd. All rights reserved.
Keywords: A. Activated carbon; B. Adsorption; D. Microporosity

1. Introduction
The DubininRadushkevich (DR) equation is widely
used to describe adsorption of sub-critical vapours in
microporous solids such as activated carbons and zeolites.
The utility of the equation lies in the fact that the
temperature dependence is reflected in the adsorption
potential A 5 2 RT ln (x) with x 5 p /p0 , that is, if the
adsorption data at different temperatures are plotted as the
logarithm of the amount adsorbed versus the square of
adsorption potential, all the data points should fall into one
curve. This curve is known in the literature as the
characteristic curve.
The existence of the characteristic curve has been
proven in numerous cases by Dubinin and other workers.
The success of the DR equation in describing adsorption of
many subcritical fluids is undeniable. There are also
successful attempts to modify the equation to make it
applicable to adsorption of supercritical adsorbates [13].
*Corresponding author. Tel.: 161-7-3365-4154; fax: 161-73365-2789.
E-mail address: duongd@cheque.uq.edu.au (D.D. Do).

However, very often the characteristic curve concept is

broadly assumed, and in extreme cases, it fails to describe
the experimental data [4]. Furthermore, there are issues
that are open for discussion, such as the pressure range in
which the equation is applicable, the question of compliance with the Henry law at low pressure, the derived
heat of adsorption at very low or very high pressures, etc.
Originally the DR equation was developed as a semiempirical tool to describe the adsorption isotherms of
microporous materials based on the Polanyi potential
theory of adsorption [5]. It is, however, possible to explain
the equation using some theoretical foundation, e.g., using
statistical thermodynamics as illustrated in the work of
Yang et al. [6].
Recently, we introduced a new method to describe the
adsorption process in microporous carbonaceous materials
with a pore size distribution [7]. It is a structure based
method, where emphasis is placed on the enhancement of
adsorption evaluated by means of molecular potential
calculations. This treatment allows the use of a single
mechanism for adsorption in pores of different sizes. In
this paper, we will use this technique to investigate the
adsorption in carbon micropores in an attempt to shed

0008-6223 / 01 / $ see front matter 2001 Elsevier Science Ltd. All rights reserved.
PII: S0008-6223( 00 )00265-7

C. Nguyen, D.D. Do / Carbon 39 (2001) 1327 1336

1328

Nomenclature
C
affinity coefficient of the BET equation
Cm
micropore capacity
E
potential of gas phase molecule
P
relative pressure
Q
heat of adsorption at zero loading
R
gas constant
T
temperature
W
volume adsorbed
W0
micropore volume
p
pressure
p0
vapour pressure
DH
isosteric heat of adsorption
DHv ap
vaporisation heat
r
pore size (radius or half width)

characteristic energy derived from DR equation

s
liquid surface tension
r
liquid density
gm
liquid molar volume
s
subscript referring to solid slab surface
p
subscript referring to micropore

some light on the working mechanism behind the DR

equation.

2. Description of the DR equation and our model of

adsorption

2.1. The DR equation

In its basic form, the DR equation can be written as:
W/W0 5 exp[2(A /E)2 ] or W/W0
5 exp[2(RT ln (x) /E)2 ]

(1)

which the characteristic energy and the micropore volume

can be obtained.
As an illustration of the applicability of the DR equation, Fig. 1 shows the characteristic curves of nitrogen and
benzene adsorption onto two activated carbon (AC) samples, Ajax AC and ACF. The adsorption temperature for
nitrogen is 77 K, whereas that for benzene is 303 K. The
equation of the straight lines describing the linear section
of the plots is used to calculate the micropore volume and
the characteristic energy, from which an effective pore size
can be estimated. Results of such calculations are shown in
Table 1. Similar results would be obtained from a direct
fitting of the DR equation to the experimental data using a

with x 5 p /p0 . The characteristic energy E for a given

fluidsolid system can be further expressed using a scaling
factor J as follows:
E 5 b E0

(2)

where E0 is the characteristic energy of a standard

adsorbate with respect to the given solid. For example,
benzene was chosen by Dubinin and his co-workers to be
the standard adsorbate for carbonaceous materials. This is
probably due to the similarity between the benzene ring
and the hexagonal arrangement of carbon atoms in the
graphite basal planes. This parameter is termed the similarity coefficient, and it is a function of the adsorbate
properties only.
The characteristic curve of a system is established by
plotting the logarithm of the amount adsorbed W versus
log 2 (1 /x). This is to test the suitability of the equation
and / or to determine its range of applicability. If the
equation is applicable, the plot would be a straight line
with a slope 2 (RT /E)2 and an intercept log (W0 ), from

Fig. 1. Nitrogen (solid) and benzene (open) characteristic curves

for the Ajax AC (circle) and ACF (diamond) carbons.

C. Nguyen, D.D. Do / Carbon 39 (2001) 1327 1336

1329

Table 1
Parameters calculated from nitrogen and benzene data using the DR equation
Temp. (K)

r (cc / mol)

ACF
Nitrogen
Benzene

77
303

34.68
89.80

Ajax
Nitrogen
Benzene

77
303

34.68
89.80

Cms (mmol / g)

W0 (cc / g)

E (kJ / mol)

r (nm)

6.95
2.53

0.24
0.23

6.9
18.1

0.66
0.66

13.9
5.34

0.48
0.47

7.0
17.8

0.65
0.67

non-linear optimisation routine. As seen in Table 1, the

results calculated from the adsorption data of different
probe molecules are consistent with each other. This is
considered to be another advantage of the DR equation.
Table 1 also indicates that although the mean micropore
sizes are about the same, the micropore volume of Ajax
AC is about two times larger than that of the ACF sample.

2.2. Potential enhancing adsorption model

2.2.1. Adsorption mechanism
In our approach [7], we adopt a common mechanism for
the adsorption process occurring in carbonaceous pores as
a sequence of surface layering followed by condensation
(or pore filling), irrespective of their sizes. The principal
feature in our approach is the allowance for the enhancement in adsorption induced by the overlapping of the
potential fields of opposite pore walls. This enhancement
not only affects the adsorbed phase, but also the gas phase
molecules occluded within the pore. Because of this
enhancement effect on the occluded molecules, the fluid
phase pressure within the pore, pp , is different from the
bulk pressure p. According to the Boltzmann distribution
of molecules, the pore pressure can be estimated as
follows:

2 kE gp (r)l
]]]
pp (r) 5 p exp
RT

coefficient is a function of the difference between the heats

of adsorption. This difference is contributed mainly by the
difference between the adsorption potential energy in a
pore and that on a surface, since other energy contributing
factors such as the vibrational and rotational energies, are
assumed to be similar for those two situations. Even when
there is a difference between the sum of these energies of
the fluid and adsorbed phases, this difference is expected to
be smaller than the change in the potential energy. With
this assumption, we shall use the potential energy at zero
loading as that in the calculation of the difference in heats
of adsorption in Eq. (4) (Q p (r)). The potential energies (in
an absolute sense) at zero loading corresponding to the
well depth and at the center of the pore are plotted in Fig.
2 as a function of pore size for benzene adsorption in
carbon pores. These energies are determined from the
pair-wise summation of a benzene molecule with all atoms
on the walls of the carbon pore. The carbon pore wall is
assumed to have a structure similar to the stack of graphite
layers [8].
Based on the concepts presented above, adsorption in
pores can be pictured as a process whereby gas phase
molecules are drawn into the pore interior. Once inside,
they are further pressed against each other as a result of the
overlapped potential forces. If the enhanced pressure is

(3)

with kE gp l, the average potential energy of the occluded

molecules. This average potential is a function of the pore
size, and it will be approximated as the potential energy at
the pore center E gp .
The adsorption layering is also enhanced, and the
statistical thickness of the adsorbed layer can be described
by means of a BET equation. The enhancement is accounted for through the proper use of the BET coefficient
for the pore, Cp , which is related to the BET coefficient for
a flat surface having the same surface chemistry and
structure, as follows:

Q p (r) 2 Q s
Cp (r) 5 Cs exp ]]] .
RT

(4)

Eq. (4) means that the enhanced adsorption affinity

Fig. 2. Potential energy of adsorption used in the DR equation

and our method.

1330

C. Nguyen, D.D. Do / Carbon 39 (2001) 1327 1336

beyond the corresponding vapor pressure, the adsorbed

phase turns into liquid (albeit compressed liquid). It is,
therefore, understood that in some narrow pores, due to the
very large enhanced pressures, the adsorbed phase exists as
liquid even at low bulk pressures. This liquid filling
process progresses to pores larger in size as the bulk
pressure increases. This is a picture of the micropore filling
process in detail, in contrast to the macroscopic description
available from the Dubinin theory [1,2], and to some
extent even the HK approach [9]. Parallel to this, in larger
pores, adsorption occurs as a surface layering process, and
this layer will grow with pressure until a stage is reached
when the enhanced pressure reaches the vapor pressure
value or the conditions for pore filling are satisfied (for
mesopores this is termed capillary condensation).

2.2.2. Single pore isotherms

The statistical thickness of the adsorbed layer t is
evaluated using the BET equation [10], which takes the
following form when variables are allowed to bear dependence on the pore size:
x p (r)Cp (r)
t(r) 5 t m ]]]]]]]]
(1 2 x p (r))(1 2 x(r) 1 x p (r)Cp (r))

(5)

with t m being the monolayer thickness. The threshold half

width r k when the pore filling by pore condensation occurs
as
sm g
r k 5 ]]]
(6)
RT ln (x p (r))

with x p (r) 5 pp (r) /p0 . The liquidsolid contact angle is

assumed to be zero for simplicity. The mathematical
interpretation of the arguments in the previous section can
then be written as follows.
The thickness of the adsorbed layer a
(0 # a # r)
r if pp $ p0

5 t if pp , p0 and a 2 t . r k
r if pp , p0 and a 2 t # r k

(7)

The adsorption isotherm can thus be generated for pores of

any size. Examples of these single pore isotherms are
shown in Fig. 3 for nitrogen adsorption at 77 K in slit-like
carbonaceous pores of different size.
These single pore isotherms are used as the local
isotherms in our technique to calculate the pore size
distribution (PSD) of the subject solid. The optimisation
procedure involved is presented elsewhere [11].

2.2.3. PSD of the ACF sample from nitrogen adsorption

at 77 K
The model is tested against adsorption data of nitrogen
at 77 K for a microporous carbon sample (Takeda ACF).
Results of the model fitting and the extracted PSD of the
sample are shown in Fig. 4. As seen, the fitting is excellent
at both low and high pressures. The extracted PSD is
presented in the form of a histogram. As shown, the
sample is a microporous material with a majority of pores
falling into the micropore range (by IUPAC definition

Fig. 3. Nitrogen single-pore isotherm at 77 K for carbon slit-like pores of half widths of 1.46, 2.54, 3.26, 5.06, 8.66, 17.65 (top) and 0.17,
0.24, 0.38, 0.74, 1.82 nm (bottom); (reading from left to right).

C. Nguyen, D.D. Do / Carbon 39 (2001) 1327 1336

1331

Fig. 4. Nitrogen isotherm fitting and the resulting PSD of sample ACF.

2r , 2 nm). As presented above, the DR equation gave for

this sample a mean pore half width of 0.66 nm (or a pore
width of 1.32 nm).

3. Volume filling of micropore theory versus

sequential adsorption

3.1. Progress of adsorption

Strictly speaking, the fractional loading in the DR
equation defined as u 5 W( p) /W0 accounts for adsorption
in micropores only. This means that the contribution of
larger pores to the uptake must be removed from the
experimental data before they are used with the DR
equation. Furthermore, the DR equation provides only the
macroscopic information. For example, when the change in
the free energy, A 5 RT ln (x), is equal to the characteristic
energy of the system, the fractional loading is 0.37,
meaning 37% of the micropore volume is occupied by the
adsorbate molecules. Since all carbonaceous solids have a
distribution of pore sizes, the DR theory does not provide
information on the individual fractional filling for each
pore size. One would expect small pores to be filled more
than the larger ones. Such information is important in the
better understanding of the variation of adsorption affinity
and heat of adsorption with loading.
Our approach models adsorption as a process occurring
in all pores simultaneously, with the amount adsorbed in a
pore being dependent on the pore size. This mechanism is
schematically illustrated in Fig. 5. As illustrated in the
figure, a fractional loading means that all pores narrower
than a certain size are filled with adsorbate molecules,
while in larger pores, adsorption layer is built up on the
pore walls, and the thickness of the adsorbed layer is a
function of the pore size. This is different from Dubinins
theory, and also from the mechanism assumed in the

traditional HK method, where adsorption in a pore is zero

until a critical pressure, when the pore volume is instantaneously filled by the adsorbate liquid. According to the HK
mechanism, pores are empty or completely filled with the
liquid adsorbate, i.e. there is no half full status for pores.
The overall adsorption on the ACF is a summation of
adsorption in pores of all sizes. The contributions of each
individual pore size, which appear in the PSD shown in
Fig. 4, are presented in Fig. 6. Table 2 summarises the
contributions of adsorption in different pore ranges at three
levels of reduced pressure, 6.3310 25 , 2.98310 22 , and
0.39. In the case of the ACF, we divide the pore spectrum
into 24 sub-ranges, each characterised by a nominal pore
size. The division allows more sub-ranges to be in the
micropore region to reflect the microporous nature of the
ACF. In the table, the half width, pore volume and
adsorption data relating to micropores are shown in italics
for easier reference. Furthermore, to distinguish pores
already filled to their maximum capacity, the amounts
adsorbed in these pores are printed bold. Let us first study
the adsorption at the low relative pressure of 6.3310 25 .
Column 3 of Table 2 shows that adsorption occurs mainly
in the smallest micropores. The first three pore sub-ranges
are filled completely with adsorbate (bold italic) while
larger pores are layered with adsorbate molecules. The
overall capacity at this pressure is 3.04 mmol / g, compared
to 3.06 mmol / g calculated using the DR equation. Here we
see that besides the overall capacity, our approach provides
a detailed dispersion of this adsorption of 3.04 mmol / g. At
a relative pressure of 2.98310 22 , all pores having half
width less than 0.588 nm are filled. This pore filling
progresses to pores of larger size when the pressure
increases. For example, at a reduced pressure of 0.39, the
threshold pore size for complete filling is 1.266 nm, which
is in the mesopore range. The dominance of adsorption in
micropores, especially at low pressures, is obvious from
Table 2. When the relative pressure increases from 6.33
10 25 to 2.98310 22 to 0.39, the contribution of micropore

1332

C. Nguyen, D.D. Do / Carbon 39 (2001) 1327 1336

Fig. 5. Progress of adsorption in pores of three different sizes.

Fig. 6. Nitrogen isotherms in different pore sub-ranges (cf. Table 2) of the ACF sample.

C. Nguyen, D.D. Do / Carbon 39 (2001) 1327 1336

1333

Table 2
Nitrogen adsorption onto ACF at 77 K at three relative pressures (per gram of sample)
x 5 6.3 3 10 25

x 5 2.98 3 10 22

x 5 0.39

Half
width
nm

Fraction
volume
cc

mmol

mmol

0.286
0.320
0.357
0.397
0.439
0.485
0.534
0.588
0.645
0.706
0.772
0.843
0.919
1.001
1.090
1.266
1.683
2.219
2.912
3.805
4.958
6.444
8.361
9.214

0.004
0.047
0.041
0.029
0.009
0.007
0.021
0.030
0.026
0.018
0.006
0.005
0.005
0.006
0.004
0.002
0.002
0.002
0.005
0.006
0.006
0.009
0.014
0.007

0.1256
1.3611
1.1914
0.2410
0.0367
0.0144
0.0241
0.0219
0.0136
0.0069
0.0018
0.0012
0.0010
0.0011
0.0007
0.0003
0.0002
0.0001
0.0003
0.0002
0.0002
0.0002
0.0003
0.0001

4.1
44.7
39.1
7.9
1.2
0.5
0.8
0.7
0.4
0.2
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

1.32
13.10
10.84
2.11
0.31
0.12
0.20
0.18
0.11
0.06
0.01
0.01
0.01
0.01

4.6
46.1
38.1
7.4
1.1
0.4
0.7
0.6
0.4
0.2
0.1
0.0
0.0
0.0

0.1256
1.3611
1.1914
0.8267
0.2653
0.2069
0.6021
0.8551
0.3667
0.2249
0.0674
0.0490
0.0465
0.0516
0.0328
0.0159
0.0085
0.0066
0.0155
0.0144
0.0117
0.0134
0.0161
0.0071

2.0
21.3
18.7
13.0
4.2
3.2
9.4
13.4
5.7
3.5
1.1
0.8
0.7
0.8
0.5
0.2
0.1
0.1
0.2
0.2
0.2
0.2
0.3
0.1

1.32
13.10
10.84
7.24
2.27
1.74
4.99
7.03
3.00
1.83
0.55
0.40
0.38
0.42

2.3
23.3
19.3
12.9
4.0
3.1
8.9
12.5
5.3
3.3
1.0
0.7
0.7
0.7

0.1256
1.3611
1.1914
0.8267
0.2653
0.2069
0.6021
0.8551
0.7624
0.5144
0.1688
0.1340
0.1382
0.1663
0.1141
0.0637
0.0155
0.0118
0.0278
0.0258
0.0208
0.0240
0.0288
0.0126

1.6
17.8
15.5
10.8
3.5
2.7
7.9
11.2
9.9
6.7
2.2
1.7
1.8
2.2
1.5
0.8
0.2
0.2
0.4
0.3
0.3
0.3
0.4
0.2

1.32
13.10
10.84
7.24
2.27
1.74
4.99
7.03
6.23
4.19
1.37
1.09
1.12
1.34

2.0
19.7
16.3
10.9
3.4
2.6
7.5
10.6
9.4
6.3
2.1
1.6
1.7
2.0

Micropore
Overall

0.254
0.308

3.042
3.044

99.9
100

28.39

6.240
6.382

97.8
100

55.09

7.318
7.663

95.5
100

63.87

Loading
mmol

Energy
%

Loading

Energy
%

Loading

Energy
%

3.2. Potential energies of adsorption

adsorption decreases slightly from 99.9 to 97.8 and then

95.5%.
Table 3 presents the results of the nitrogen adsorption
onto the ACF sample at three pressure levels calculated
using the DR and our method. The amount adsorbed
calculated using the DR equation is by and large equal to
the sum of the amounts adsorbed in all pores using our
technique. The difference is more significant at the higher
end of the pressure range, where adsorption in mesopores
occurs in addition to that in micropores. This is because
the DR equation, unable to deal with mesopores, underpredicts the data at high relative pressures, while our technique describes well the data at all pressure levels.

The only parameter concerning the solid properties

appearing in the DR equation is the characteristic energy E
(or E0 ). By investigating a large number of systems, the
dependence of the characteristic energy on the size of the
carbon slit pore is found by Dubinin and co-workers
[12,13], and independently by Stoeckli [14] to take the
following approximate form:
k
r5]
E0

(8)

with k being a constant (12 kJ nm / mol). This equation is

Table 3
Nitrogen adsorption onto ACF sample at three relative pressures calculated using our method and the DR equation
Relative
pressure

Overall capacity
(mmol / g)

Experimental

Our method

DR eq.

Our method

3.03
6.36
7.63

3.04
6.38
7.66

3.12
6.25
6.9

3.04
6.24
7.31

25

6.3310
2.98310 22
0.39

Micropore capacity
(mmol / g)

C. Nguyen, D.D. Do / Carbon 39 (2001) 1327 1336

1334

used to estimate the average pore size of a porous carbon

once the characteristic energy E (or E0 ) is known from the
adsorption isotherm data. More complex equations are also
introduced to describe the relationship between the characteristic energy and the pore size [15].
The plot of the rE0 relationship for benzene is presented in Fig. 2 for comparison with the potential energies
used in our approach. The potential energy curves used in
our approach represent two extreme energy levels for a
molecule residing inside a pore, i.e. the potential energy at
the pore centre and the minimal potential energy a
molecule may have within the pore interior. It is known
that those two potential levels coalesce for very narrow
pores. If a plot representing the average adsorption potential energy is drawn, it would lie between those two
curves, and would follow the two curves when they unite
at smaller pore sizes. Here we argue that this is the case of
the DR characteristic energy plot shown in Fig. 2, which,
as discussed above, is an approximation of the average
adsorption potential energy. There is a big difference
between our and the DR curves at the lower end of the
pore size range. The DR characteristic curve, which is
essentially a hyperbola gives an infinite potential energy
for pore size approaching zero, which is not the case. Our
curves, on the other hand, show a maximum potential
energy (in an absolute sense) for a pore size equal to the
solidfluid collision diameter. The potential for adsorption
decreases as the pore size is lowered beyond that threshold
dimension. The difference between the potential curves is
the main cause for our and the DR equation to give
different average pore sizes and isosteric heats of adsorption, as will be discussed further below.

3.3. Heat of adsorption

Another advantage of the DR equation is that it allows
for the isosteric heat of adsorption to be calculated
analytically. By using Clapeyrons equation, the equation
for the isosteric heat is given below:

S D

1
2 DH 5 DHvap 1 E ln ]
u

1/2

Ed T
1]
2

Sln ]u1 D

21 / 2

(9)

where d is the coefficient of expansion of liquid adsorbate,

which is very small [16]. Fig. 7 shows the plot of the
isosteric heat of adsorption of nitrogen at 77 K onto the
ACF sample. Despite the convenience and simplicity, Eq.
(9) suffers from the fact that the isosteric heat is infinite at
zero loading.
As discussed above, the decrease in the potential energy
of adsorbate molecules is the principal source of heat
released during the adsorption process. One can expect
therefore that the changing pattern of the heat of adsorption with loading would mimic the variation of this
quantity. For adsorption in micropores, where the solid
fluid interaction is much more significant compared to the

Fig. 7. Isosteric heat versus fractional loading for nitrogen

adsorption onto ACF.

fluidfluid interaction, the potential change, which is

calculated using the Lennard-Jones potential energy equation, can be taken as the first approximation for the heat of
adsorption. The heat released from individual micropores
at any stage of adsorption can thus be estimated. Energies
generated by nitrogen adsorption in pores of different sizes
of the ACF sample are presented in Table 2. It is obvious
that the energy released is larger in smaller pores and is
proportional to the pore volume. Since the ACF sample
has a distribution of pores mostly having half widths less
than 0.40 nm, the bulk of the actual energy is generated by
adsorption occurring in that pore range. Knowing that the
majority of adsorption takes place in micropores, and that
adsorption in micropores produces more heat than in the
larger pores, we get an estimation of the total heat released
at any overall loading (Cm ) by adding up the differential
heats. This is the cumulative heat produced when the
amount of Cm is loaded into a clean sample. The cumulative heats for the three relative pressure levels are calculated using the above method and the results are presented
in Table 2. For example, at a relative pressure of 2.983
10 22 , an amount of 6.38 mmol of nitrogen is loaded on 1 g
of the ACF sample, and the cumulative heat is 56.24 J (an
average of |9 kJ / mol). The isosteric heat of adsorption
(kJ / mol) is the heat released at a given loading Cm , i.e. it
is the incremental heat produced per mole when the
loading is incrementally increased from Cm to Cm 1 DCm
(DCm < Cm ). The results are shown in Fig. 7. The
maximum possible isosteric heat for nitrogen is |16.2
kJ / mol if all pores have widths equal to the nitrogen
collision diameter (i.e. 0.38 nm). However, due to a
distribution of pore sizes in the ACF sample, the isosteric
heat will be less than this theoretical maximum value. The
isosteric heats calculated using the DR equation and our
technique are quite different. As discussed before, the
difference in potential energy curves could be a reason for
the difference in isosteric heats of adsorption. A com-

C. Nguyen, D.D. Do / Carbon 39 (2001) 1327 1336

parison between the methods and / or a justification of the

suitability can be made based on experimental measurements of the isosteric heats, but these results are not
available at this stage.

3.4. PSD and the applicability of the DR equation

In this section we will try to address the question
regarding the application range of the DR equation. To
increase the smoothness (resolution) of the simulated
isotherm, we sub-divide the pore spectrum (half widths
from 0.286 to 9.214 nm) of the ACF sample into 44
sub-ranges. Fig. 8 shows the isotherms of nitrogen adsorption in these sub-ranges. Here the isotherms are shown
for a relative pressure interval ranging from 1310 24 to
1310 21 , which is the pressure interval where the DR
equation is normally investigated. As seen, none of the
single pore isotherms can be considered to be a straight
line. However, by adding up the individual isotherms, we
get the overall isotherm shown as the dotted line in Fig. 8,
which is practically a straight line. This line is seen to fit
the experimental data very well. A linear approximation of
the overall isotherm plotted in this mode is an explanation
for the existence of a DR characteristic curve.
It is then understood that the DR equation is not
applicable for adsorbents having a single pore size, and
that a distribution of pore sizes is a prerequisite for the DR
equation. It is also possible to prove that samples having
very skewed PSDs may not comply with the DR equation.
Fig. 9 shows the PSD together with the individual and the

1335

overall nitrogen isotherms for such a hypothetical carbon

sample. It is clear that the overall isotherm of this sample
does not support the concept of an adsorption characteristic
curve. Thus, for the DR equation to be applicable, the
porous solid must have a distribution of micropores, and
the distribution must not be very skewed. This explains
why the DR equation is not applicable to all microporous
materials.
A more general equation of the volume filling of
micropores theory, the DubininAstakhov equation [17],
can also be explained similarly by considering the isotherms presented in the form of a log (Cm ) versus log n
(1 /x) plot.

4. Conclusions
Several techniques for analysis of adsorption in micropores are presented. The DR equation as well as our
technique are capable of describing the adsorption data.
The former is by and large a means to fit the experimental
data, while the later can be used to extract information
regarding the PSD. It is found that the overall adsorption
behaviour as expressed by the DR equation can be derived
by combining processes occurring in individual pore
ranges using our method of analysis. The method also
allows the assessment of details such as how much of the
adsorption capacity and the heat released is contributed by
individual pore ranges.

Fig. 8. Nitrogen isotherms plotted in a log (Cm ) versus log 2 (1 /x) mode (single pore isotherms, solid lines; overall isotherm, dotted line;
experimental, solid circle).

1336

C. Nguyen, D.D. Do / Carbon 39 (2001) 1327 1336

Fig. 9. An arbitrary PSD (left) and the nitrogen isotherms in pore sub-ranges (right). The overall isotherm (dotted line) does not comply
with the DR equation principle.

Acknowledgements
Support from the Australian Research Council is gratefully acknowledged.

References
[1]
[2]
[3]
[4]

Dubinin MM. Chem Rev 1960;60:235.

Dubinin MM. Carbon 1978;25:593.
Kaneko K, Murata K. Adsorption 1997;3:197.
Rouquerol F, Rouquerol J, Sing K. Adsorption by powders
and porous solids, London: Academic Press, 1999.
[5] Polanyi M. Trans Faraday Soc 1932;28:316.
[6] Chen SG, Yang R. DR equation from statistical thermodynamic Langmuir 1994;10:4244.

[7] Nguyen C, Do DD. Langmuir 1999;15:3608.

[8] Marsh H. Introduction to carbon science, Cornwall: Butterworths, 1989.
[9] Horvath G, Kawazoe K. J Chem Eng Japan 1983;16:470.
[10] Greg SJ, Sing KSW. Adsorption, surface area and porosity,
London: Academic Press, 1982.
[11] Nguyen C, Do DD. Langmuir 2000;16:1319.
[12] Dubinin MM, Plavnik GM. Carbon 1966;6:183.
[13] Dubinin MM, Plavnik GM, Zaverina ED. Carbon
1964;2:261.
[14] Stoeckli HF. Chimia 1974;28:728.
[15] Dubinin MM, Stoeckli HF. J Colloid Interface Sci
1980;75:34.
[16] Wang K, Do DD. Langmuir 1997;13:6226.
[17] Dubinin MM, Astakhov BA. Izv. Akad. Nauk SSSR, Ser
Khim. 1971; 5.