Journal of Colloid and Interface Science 281 (2005) 437–443

www.elsevier.com/locate/jcis

Preparation and characterization of high-specific-surface-area activated

carbons from K2 CO3 -treated waste polyurethane
J. Hayashi a,b,∗ , N. Yamamoto a , T. Horikawa a , K. Muroyama a , V.G. Gomes c
a Department of Chemical Engineering, Kansai University, 3-3-35 Yamate-cho, Suita, Osaka 564-8680, Japan
b High Technology Research Center, Kansai University, 3-3-35 Yamate-cho, Suita, Osaka 564-8680, Japan
c Department of Chemical Engineering, The University of Sydney, NSW 2006, Australia

Received 15 March 2004; accepted 11 August 2004

Available online 12 October 2004

Abstract
An activated carbon with high specific surface area was prepared from polyurethane foam by chemical activation with K2 CO3 and the
influences of carbonization temperature and impregnation ratio on the pore structure of the prepared activated carbon were investigated. It
was found that the specific surface area of the activated carbon was at a maximum value (about 2800 m2 /g) at a carbonization temperature
of 1073 K and at an impregnation ratio of 1.0. It was concluded that the polyurethane foam structure was modified during impregnation
by K2 CO3 , K2 CO3 promoted charring during carbonization, and then the weight loss behavior was changed below 700 and above 1000 K,
carbon in the char was consumed by K2 CO3 reduction, and this led to the high specific surface area. The prepared activated carbon had a
very sharp micropore size distribution, compared with the commercial activated carbon having high specific surface area. The amounts of
three organic vapors (benzene, acetone, and octane) adsorbed on the prepared activated carbons was much larger than those on the traditional
coconut shell AC and the same as those on the commercial activated carbon except for octane. We surmised that the high specific surface
area was due to the modification of the carbonization behavior of polyurethane foam by K2 CO3 .
 2004 Elsevier Inc. All rights reserved.

Keywords: Activated carbon; Polyurethane foam; K2 CO3 activation; Porosity; VOC

1. Introduction desired. Several recycle/reuse methods have been reported

Various synthetic polymers have been used in our lives
and in these polymers, polyurethane foam is a popular poly-
mer that finds wide use (500,000 t/y in Japan) [1] in various
applications, such as in vehicles, furniture, and construction
materials. Significant amounts of waste polyurethane foam,
generated globally, must be recycled or reused economically
and in an environmentally acceptable manner. Polyurethane
foam is a thermosetting polymer with a network structure;
thus it does not melt by heating and is solvent insoluble.
Therefore it is difficult to regenerate polyurethane foam.
The development of a proper recycle/reuse method is thus
* Corresponding author. Fax: +81-6-6388-8869.
E-mail address: hayashi7@ipcku.kansai-u.ac.jp (J. Hayashi).
0021-9797/$ – see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcis.2004.08.092
[2–5]. In this study, preparing activated carbon from waste
polyurethane foam was tried as a means of value-added re-
cycling. Activated carbon is a high-porosity material and
it is useful in adsorption of both gases from mixtures and
solutes from liquid solutions. Therefore it has been widely
used for the separation of gases, recovery of solvents, and
removal of organic pollutants from drinking water and as a
catalyst support. As environmental pollution is becoming an
increasingly serious problem, the need for activated carbon
is growing.
Activated carbons are usually prepared from raw mate-
rials such as coal [6,7] and lignocellulose materials [8,9].
These raw materials are inexpensive but there may be prob-
lems with the uniformity of their quality, while the quality of
polyurethane foam is almost uniform as it contains a negligi-
438 J. Hayashi et al. / Journal of Colloid and Interface Science 281 (2005) 437–443
ble amount of ash. Thus polyurethane is a suitable candidate
as a raw material for preparing activated carbon.
There are basically two methods for preparing an acti-
vated carbon: physical and chemical activation. Physical ac-
tivation consists of two steps: the carbonization of the raw
material and the activation of the char using carbon diox-
ide or steam. In chemical activation, both the carbonization
and the activation step proceed simultaneously. Physical ac-
tivation is the more popular of the two methods; however,
recently activated carbons with very high surface area have
been prepared by chemical activation [10–13] and chemi-
cal activation attracts attention. Activated carbon with high
specific surface area is suitable for gas storage, gas purifica-
tion, recovery of solvent vapor, and manufacture of electrical
double-layer capacitors. To get high specific surface area,
KOH and NaOH are used as activating reagents. But alkali
hydroxides such as KOH and NaOH is hazardous chemi-
cal. K2 CO3 has been tried as an activation reagent instead of
KOH and NaOH, because K2 CO3 is not a hazardous chemi-
cal [14–16].
In this study, an activated carbon was prepared from
polyurethane by chemical activation with K2 CO3 and the in-
fluence of carbonization temperature and impregnation ratio
on the pore structure (specific surface area, pore volume, and
pore size distribution) and the adsorption properties of the
prepared activated carbon was investigated.
2. Experimental
2.1. Preparing activated carbon
Polyurethane foam was offered from a certain chemical
company in the shape of a slab. No ash content was not
detected. The polyurethane foam slab was smashed using
a grinder. Polyurethane powder was mixed with saturated
K2 CO3 solution and kneaded. The mixture was then dried
at 383 K and the impregnated sample was prepared. The im-
pregnation ratio was estimated as follows:
(weight of K2 CO3 in solution)
(impregnation ratio) = .
(weight of polyurethane foam)
In this study, the impregnation ratio was varied between 0.1
and 1.0.
The impregnated sample was set on a ceramic boat, which
was located inside a stainless steel tube (I.D. 35 mm). The
impregnated sample was heated up to the carbonization tem-
perature under N2 flow (400 ml/min) at a rate of 10 K/min
and was held for 1 h at the carbonization temperature. The
carbonization temperature was varied from 773 to 1173 K.
After carbonization the sample was cooled under N2 flow
(400 ml/min). The carbonized sample was washed sequen-
tially with hot water several times and finally with cold dis-
tilled water to remove residual chemicals. The washed sam-
ple was dried at 383 K to give the activated carbon product.
2.2. Characterization of pore structure
At 77 K, N2 adsorption and desorption isotherms on the
prepared activated carbon was measured using a volumetric
sorption analyzer (BELSORP 28, Bell Japan Inc.). Before
the adsorption isotherm was measured, the sample was de-
gassed at 473 K for 2 h. The specific surface areas of the pre-
pared activated carbons were estimated by the BET method
using N2 adsorption isotherm data. The micropore volumes
were calculated from the amount of N2 adsorbed at a relative
pressure of 0.1, and the mesopore volumes were calculated
by subtracting the amounts adsorbed at a relative pressure
of 0.1 from those at a relative pressure of 0.95 [17]. The
micropore size distribution was calculated by the Horvath–
Kawazoe method [18].
2.3. Characterization of adsorption property
In order to compare the adsorption properties of prepared
activated carbon with those of commercial activated carbon
(high-specific-surface-area activated carbon and coconut-
shell activated carbon), the adsorption isotherms of a few
typical solvent vapors (acetone, benzene, and octane) were
measured at 298 K using a volumetric sorption analyzer
(BELSORP 18, Bell Japan Inc.). Before the adsorption
isotherm was measured, the sample was degassed at 473 K
for 2 h.
2.4. Thermogravimetric analysis
In order to investigate the interaction between polyure-
thane foam and K2 CO3 during carbonization, the weight loss
behavior of the polyurethane foam impregnated with K2 CO3
was examined using a thermogravimetric analyzer (TGA-50,
Shimadzu). Samples (about 15 mg) were heated to 1273 K
at a heating rate of 10 K/min under N2 flow (200 ml/min).
3. Results and discussion
3.1. Adsorption isotherm of N2
Fig. 1 shows the nitrogen adsorption isotherms on the pre-
pared activated carbons under a range of conditions. The
yield of the activated carbon prepared at 1173 K was not
significant; thus the activated carbon prepared at 1173 K
was not characterized. The amount of nitrogen adsorbed at
P /P0 = 1.0 increased for activated carbons prepared at rel-
atively higher carbonization temperatures; for example, the
amount adsorbed increased by a factor of 10 with increas-
ing temperature from 773 to 1073 K (Fig. 1a). The amount
adsorbed for activated carbons prepared at 973 and 1073 K
rapidly increased over the pressure (P /P0 ) range of 10−4
and 10−3 . For the activated carbon prepared at 1073 K,
a large increase in the amount adsorbed is observed at pres-
sures above 10−2.
 J. Hayashi et al. / Journal of Colloid and Interface Science 281 (2005) 437–443 439

Fig. 2. Influence of carbonization temperature on specific surface area.

Fig. 1. Nitrogen adsorption isotherms for prepared activated carbons with

(a) variable temperature (ratio = 1.0) and (b) variable impregnation ratio
(Tcarb = 1073 K).
At pressures (P /P0 ) below 10−3 , the difference in the
isotherms for samples with various impregnation ratios was
small. However, the amount adsorbed for the activated car-
bons prepared with ratios in the range of 0.1–0.5 increased
with increased impregnation ratio at pressures above 10−3 .
The amount of N2 adsorbed increased by a factor of about
2.5 times when the impregnation ratio increased from 0.1
to 0.5 (Fig. 1b).
3.2. Pore structure of the prepared activated carbon
Specific surface area, micropore volume, and mesopore
volume of the prepared activated carbons and commer-
cial activated carbons are summarized in Table 1. The ac-
tivated carbons prepared in this work are described by
PU–AC–impregnation ratio–carbonization temperature. The
commercial activated carbon, HS-AC, has a high specific
surface area and the commercial activated carbon, CS-AC,
is prepared from coconut shell.
3.2.1. Influence of carbonization temperature on
pore structure
Fig. 2 shows the influence of carbonization temperature
(Tcarb ) on specific surface area. The specific surface area
increased with an increase of carbonization temperature. Es-
pecially between Tcarb = 873 and 973 K, it increased rapidly
and the prepared activated carbon had significantly high spe-
cific surface areas (>2000 m2 /g).
The micropore volume increased rapidly between Tcarb =
873 and 973 K, as well as the specific surface area, as shown
in Table 1. Change of mesopore volume is not significant
below Tcarb = 973 K; however, it increases rapidly above
Tcarb = 973 K.
As described above, the specific surface area and the mi-
cropore volume increased rapidly beyond Tcarb = 873 K.
This fact shows that K2 CO3 works effectively as an activa-
tion reagent above 873 K.
Fig. 3 shows the influence of carbonization temperature
on micropore size distribution. At Tcarb of 773 and 873 K,
broad and small peaks are found; however, a significant
change in the distribution is found between Tcarb = 873 and
973 K. At Tcarb of 973 K, a sharp peak is observed and the
peak shifts to a larger diameter range at 1073 K. This figure
shows that a significant part of the micropores have diameter
ranging from 0.6 to 0.7 nm.
3.2.2. Influence of impregnation ratio on pore structure
Fig. 4 shows the influence of the impregnation ratio on
the specific surface area. At an impregnation ratio of 0.2,
the specific surface area of the product is about 2000 m2 /g.
Thus, it is highly significant that activated carbon with high
specific surface area can be prepared with a low impregna-
tion ratio. A large increase of mesopore volume is observed
for impregnation ratios between 0.2 and 0.5 as shown in
Table 1. The difference in the pore structure (specific sur-
face area and pore volume) for activated carbons prepared

Table 1
Pore structure of prepared activated carbons and commercial activated carbons
Impregnation Carbonization Specific surface Micropore Mesopore Yield
ratio [–] temperature [K] area [m2 /g] volume [ml/g] volume [ml/g] [–]
PU-AC-1.0-773 1.0 773 264.9 0.104 0.048 0.127
PU-AC-1.0-873 1.0 873 755.2 0.296 0.055 0.122
PU-AC-1.0-973 1.0 973 2183 0.852 0.126 0.085
PU-AC-1.0-1073 1.0 1073 2772 1.088 0.371 0.049
PU-AC-0.1-1073 0.1 1073 1566 0.622 0.044 0.085
PU-AC-0.2-1073 0.2 1073 2000 0.792 0.097 0.056
PU-AC-0.5-1073 0.5 1073 2618 1.027 0.430 0.045
HS-AC – – 2767 1.084 0.566 –
CS-AC – – 1131 0.446 0.048 –
440 J. Hayashi et al. / Journal of Colloid and Interface Science 281 (2005) 437–443
Fig. 3. Influence of carbonization temperature on micropore size distribution.
Fig. 4. Influence of impregnation ratio on specific surface area.
with impregnation ratios between 0.5 and 1.0 is negligible.
It was found that an impregnation ratio of 0.5 is sufficient to
achieve high specific surface area
Fig. 5 shows the influence of impregnation ratio on mi-
cropore size distribution. The influence of impregnation ratio
on the distribution is smaller than that of carbonization tem-
perature. It is observed that although the impregnation ratio
was increased, the shape of the distribution curve remains
sharp.
3.3. Weight loss behavior during carbonization
The weight loss behavior (TG and DTG) of polyurethane
foam and K2 CO3 , measured independently, is shown in
Figs. 6a and 6b as a function of temperature. Fig. 6a shows
that the weight loss of K2 CO3 is not observed below 1200 K.
The melting point of K2 CO3 is 1164 K and at higher tem-
peratures, K2 CO3 decomposes and then a small weight loss
was observed above 1200 K. Fig. 6b shows that there are two
peaks in the weight loss rate curve of polyurethane foam dur-
ing carbonization.
In Figs. 6c and 6d, the calculated weight loss behavior
for an impregnation ratio of 1.0 was obtained by assuming
that the weight loss for the polyurethane impregnated with
K2 CO3 was a summation of the values for polyurethane and
K2 CO3 measured independently. In other words, the cal-
culated weight loss behavior is the behavior assuming that
there is no interaction between polyurethane and K2 CO3 .
For the impregnated polyurethane, differences between
experimental and calculated rates were observed within the
two temperature ranges shown in Figs. 6c and 6d. One tem-
perature range is below about 700 K and the other is above
about 1000 K.
3.3.1. Weight loss behavior below 700 K
The experimental weight loss begins at about 400 K,
while the calculated weight loss begins at about 500 K, as
shown in Fig. 6c. By impregnation with K2 CO3 , the decom-
position temperature becomes lower. As shown in Fig. 6d,
the peaks of experimental weight loss rate are observed at
about 430, 520, and 630 K, while the peaks of calculated
weight loss rate are observed at about 570 and 650 K. Thus,
the peaks of weight loss rate shift to a lower temperature
range and a new peak appears at about 430 K by impregna-
tion with K2 CO3 .
It was found that the weight loss behavior was changed
by impregnation with K2 CO3 . Polyurethane foam consists
of two components, a hard segment and a soft segment.
The weight loss behavior depends on the segment length
and segment concentration [19]. Therefore, it is concluded
that the structure of polyurethane foam was modified by im-
pregnated K2 CO3 during impregnation step and the weight
loss behavior was changed at the beginning of carboniza-
tion; however, the detail of the modification mechanism is
not obvious.
 J. Hayashi et al. / Journal of Colloid and Interface Science 281 (2005) 437–443 441

Fig. 5. Influence of impregnation ratio on micropore size distributions.

Fig. 6. Weight loss behavior of polyurethane foam and K2 CO3 and impregnated polyurethane foam with impregnation ratio of 1.0.

When polyurethane foam was carbonized solely, the char
could not be obtained at 700 K, as shown in Fig. 6a. Howev-
er, the char (activated carbon) can be obtained from the im-
pregnated polyurethane foam. This fact shows that K2 CO3
promotes the charring during carbonization.
3.3.2. Weight loss behavior above 1000 K
In Fig. 6c, large differences between the experimental and
the calculated rates are observed above about 1000 K. At
1000 K, the generation of volatiles is almost finished by car-
bonization and it is concluded that this weight loss is due to
the reaction between the polyurethane char and K2 CO3 . Mc-
Kee reported that K2 CO3 was reduced by carbon in the char
and metal K and CO were formed [20]:
2C + K2 CO3 → 2K + 3CO.
In this case, it is concluded that the carbon in polyurethane
char is consumed by the reaction with K2 CO3 and then the
specific surface area increased rapidly, as shown in Fig. 2.
3.4. Adsorption property of prepared activated carbon
Figs. 7a–7c show the micropore size distributions for PU-
AC-1.0-1073, HS-AC, and CS-AC, respectively, and Fig. 7d
shows the cumulative pore size distributions of these acti-
442 J. Hayashi et al. / Journal of Colloid and Interface Science 281 (2005) 437–443
Fig. 7. Micropore size distributions of (a) PU-AC-1.0-1073, (b) HS-AC, and (c) CS-AC and (d) cumulative micropore size distribution of these activated
carbons.
Fig. 8. Adsorption isotherms on prepared activated carbon and commercial
activated carbons for (a) benzene, (b) acetone, and (c) octane vapor.
vated carbons. Figs. 7a–7c show that the distribution of PU-
AC-1.0-1073 is very sharp compared with the commercial
activated carbons (HS-AC and CS-AC). Fig. 8c shows that
pore volume of HS-AC is larger than that of PU-AC-1.0-
1073 above about 0.7 nm. Figs. 8a–8c show the adsorption
isotherms for benzene, acetone, and octane on the PU-AC-
1.0-1073 and the commercial activated carbons HS-AC and
CS-AC. The amount of benzene (nonpolar) and acetone (po-
lar) vapor adsorbed on PU-AC-1.0-1073 is comparable to the
amount adsorbed on HS-AC and is much greater (by a factor
of 4) than that adsorbed on CS-AC. Thus, the chemical char-
acterization of PU-AC-1.0-1073 is similar to that of HS-AC.
The amount of octane vapor adsorbed on PU-AC-1.0-1073
is three times the amount adsorbed on CS-AC. However,
the amount of octane vapor adsorbed on PU-AC-1.0-1073
is slightly less than on HS-AC. The chemical characteriza-
tion of HS-AC is similar to that of PU-AC-1.0-1073 and it is
speculated that this difference is due to the physical charac-
terization. Then the difference between pore size distribution
of HS-AC and that of PU-AC-1.0-1073 was investigated.
The micropore size distribution of PU-AC-1.0-1073 has a
very sharp distribution at about 0.6 nm (in Fig. 7a), but the
pore volume (>0.7 nm) of HS-AC is larger than that of
PU-AC-1.0-1073 (Fig. 7d). The exact suitable pore size was
unknown, but if the suitable pore size for octane adsorption
is above 0.7 nm, the amount adsorbed on HS-AC is larger
than that on PU-AC-1.0-1073.
4. Conclusion
The activated carbons were prepared from polyurethane
foam by chemical activation with K2 CO3 . For the acti-
vated carbons, the specific surface area increased with an
increase of carbonization temperature in the range from 773
to 1073 K and with an increase of the impregnation ratio in
the range from 0 to 1.0. The activated carbons prepared at the
carbonization temperature of 1073 K and impregnation ratio
of 1.0 had a maximum surface area of 2772 m2 /g. It was
concluded that polyurethane foam was modified by K2 CO3
during impregnation, K2 CO3 promoted charring during car-
bonization, and the carbonization behavior changed below
 J. Hayashi et al. / Journal of Colloid and Interface Science 281 (2005) 437–443
700 K and above 1000 K, the carbon in polyurethane char
was consumed by K2 CO3 reduction, and the specific surface
area increased significantly.
This activated carbon had a very sharp micropore size dis-
tribution, compared with commercial activated carbon. The
adsorption properties are about the same as those of the com-
mercial activated carbon having high specific surface area
(HS-AC) and significantly improved compared to those of
commercial coconut-shell-based activated carbon (CS-AC).
Thus waste polyurethane foam was found to be an attrac-
tive source of raw material for recycling as a value-added
activated carbon via chemical activation with K2 CO3 .
References
[1] K. Iwasaki, in: Y. Oyanagi (Ed.), Foundations and Applications of
Plastic Recycling, CMC Publishing, Tokyo, 2003, pp. 159–165.
[2] S. Chuayjuljit, C. Norakankorn, V. Pimpan, J. Met. Mater. Miner. 12
(2002) 19.
[3] K. Troev, A. Tsekova, R. Tsevi, J. Appl. Polym. Sci. 78 (2000) 2565.
[4] F. Shutov, in: Proceedings of the 15th International Conference on
Solid Waste Technology and Management, 1999, pp. 3D/1–3D/6.
[5] T. Taniguchi, N. Sato, M. Matsushita, H. Takahashi, T. Suzuki,
Y. Hoshino, N. Abe, Kobunshi Ronbunshu 56 (1999) 709.
443
[6] M.A. Lillo-Rodenas, D. Lozano-Castello, D. Cazorla-Amoros, A. Li-
nares-Solano, Carbon 39 (2001) 751.
[7] B. Buczek, A. Swiatkowski, S. Zietek, B.J. Trznadel, Fuel 79 (2000)
1247.
[8] O.N. Baklanova, G.V. Plaksin, V.A. Drozdov, V.K. Duplyakin, N.V.
Chesnokov, B.N. Kuznetsov, Carbon 41 (2003) 1793.
[9] W.E. Marshall, M. Ahmedna, R.M. Rao, M.M. Johns, Int. Sugar J. 102
(2000) 147.
[10] V. Verheyen, R. Rathbone, M. Jagtoyen, F. Derbyshire, Carbon 33
(1995) 763.
[11] G. Amarasekera, M.J. Scarlett, D.E. Mainwaring, Carbon 36 (1998)
1071.
[12] D. Lozano-Castelló, M.A. Lillo-Ródenas, D. Cazorla-Amorós, A. Li-
nares-Solano, Carbon 39 (2001) 741.
[13] T. Otowa, Kagaku to Kogyo (Osaka, Japan) 64 (1990) 331.
[14] J. Hayashi, A. Kazehaya, K. Muroyama, A.P. Watkinson, Carbon 38
(2000) 1873.
[15] J. Hayashi, T. Horikawa, I. Takeda, K. Muroyama, F.N. Ani, Carbon 40
(2002) 2381.
[16] J. Hayashi, T. Horikawa, K. Muroyama, V.G. Gomes, Micropor. Meso-
por. Mater. 55 (2002) 63.
[17] F. Rodriguez-Reinoso, J.D. Lopez-Gonzalez, C. Berenguer, Carbon 20
(1982) 513.
[18] G. Horvath, K. Kawazoe, J. Chem. Eng. Jpn. 16 (1983) 470.
[19] Z.S. Petrovic, Z. Zavargo, J.H. Flynn, W.J. Macknight, J. Appl. Polym.
Sci. 51 (1994) 1087.
[20] D.W. McKee, Fuel 62 (1983) 170.