Agric. Biol. Chem.

, 41 (6), 983•`988, 1977

Specific Surface Area and Specific Pore Volume Distribution of Tobacco•õ

Toshiro SAMEJIMA, Yoshihiko Sox and Toshimasa YANO*

Central Research Institute, The Japan Tobacco & Salt Public Corporation ,
Umegaoka 6-2, Midori-ku, Yokohama 227, Japan
*Department of Agricultural Chemistry
, The University of Tokyo,
Bunkyo-ku, Tokyo 113, Japan
Received December 20, 1976

The specific surface area and the specific pore volume distribution of Japanese tobaccos
were measured by means of the low temperature gas adsorption technique utilizing the B.E.T.
and the Inkley methods. The specific surface area and the specific pore volume of the micro
pores less than 300 A in diameter varied from 6,000 to 17,000cm2/g and from 0.0012 to
0.0036cm3/g, respectively, with types of curing in the ascending order of the sun cured, the flue
cured and the air cured tobaccos. The both specific values were increased by extracting
the tobaccos with water greatly in the case of the flue cured, while slightly in the case of the air
cured tobaccos, suggesting that the effect of the curings on the specific values were due to
differences in the content of low molecular components. Effectiveness of puffing was also
shown. The specific surface area was linearly correlated with the specific volume of the
micropores less than 300.E in diameter, the constant term showing that contribution of the
larger pores more than 300 A in diameter to the specific surface area of tobacco was insignifi
cant.

Many flavor ingredients are added to tobacco The specific pore volume distribution was
to improve the taste of a cigarette, and reten determined on Canadian tobaccos by Parups8)
tion as long as possible of the added flavor employing the Benesi method. However, his
ingredients is a matter of primary concern in results are quite different from ours obtained
the cigarette manufacturing. In order to by the Inkley method.
develope a rational flavoring technique, it will In this paper, the porous structure of Japa
be necessary to analyse the sorption and de nese tobaccos was studied by measuring the
sorption phenomena of the flavor ingredients in specific surface area and the specific pore
tobacco or in the plant leaf tissue, though volume distribution, preceding to quantitative
they have not yet been satisfactorily studied. analyses of the sorption and desorption
The sorption of moisture in tobacco, on the phenomena of moisture and flavor ingredients.
other hand, has been studied by many work The results will also give an important in-
ers,1•`4) since it has significant influences not formation on the filling capacity, the com
only on the handling properties of tobacco bustion efficiency and the autofiltration ef
but also on the quality of the finished cigarette. ficiency of cigarettes.
The sorption of moisture, however, could be

better understood if the porous structure of

the tobaccos had been studied in those papers. MATERIALS AND METHODS

Although the specific surface areas of American Tobaccos. Varieties of tobacco examined are listed
tobaccos were determined by several work in Table I. Since physicochemical properties of the
ers,5•`7) there is no data on Japanese tobaccos. tobaccos greatly differ with types of curing,127 the
samples were selected to cover three types of curing;
•õ Sorption and Desorption Phenomena in Tobacco. a sun cured, a flue cured and two air cured tobaccos.
Part I. A part of this work was presented at the The sun cured tobacco was dried just after the harvest,
Annual Meeting of the Agricultural Chemical Society the flue cured tobacco was dried after yellowing their
of Japan on April, 1976 in Kyoto. leaves in an air conditioned room and the air cured
984 T. SAMEJIMA, Y. SOH and T. YANO

TABLE I. VARIETIES of TOBACCO USED was determined on 2 to 5 g of the tobacco samples with
IN EXPERIMENT the Shibata Surface-Area Analyzer P-700. The samples
were outgassed at 60°C until the pressure reached under
10-4 mmHg. Helium was used for the dead space
determination.

Specific surface area. The specific surface area was

calculated from the nitrogen adsorption isotherm by use
of the following equation that approximates Brunauer,
Emett and Teller's infinite layer adsorption equation
in the range 0.05<P/P0<0.35.10)

whereƒË

: volume of nitrogen adsorbed at a pressure P [ml/g]ƒË

TABLEII. SPECIFICSURFACEAREAAND SPECIFIC m: volume of nitrogen required to form the mono
PORE VOLUME(dp<300A) of VARIOUSTOBACCOS layer on an adsorbent [ml/g]

P: vapor pressure of nitrogen at a time of measure

ment [mmHg]

P0: saturated vapor pressure of nitrogen at a given

temperature [mmHg]

C: constant related to the heat of adsorption [-]

A plot of P/ƒË(P0-P) ƒËs. P/Po yielded a straight line,

the slope of which giving (C-1)/ƒËm. C and the intercept

of the ordinate giving 1/ƒËm•EC. From these two values,

ƒËm was calculated. The specific surface area was then

calculated as following.

S=A x (ƒËm/22400)ƒË6.02•~1023•~10-11 [cm2/g] (2)

where

A: cross-sectional area of a nitrogen molecule [A]

tobaccos were dried after browning the yellowed leaves
By taking the value of A as 16.2 A2,11) Eq. (2) becomes
by a conventional method. The tobacco leaves were
separated into laminae and midribs. Specific surface S=43500•EƒËm [cm2/g] (3)

area and specific pore volume distribution were measur

Specific pore volume distribution. The specific pore
ed on the laminae of all varieties and the midrib of
Bright Yellow -4 (BY-4). volume distribution was computed from the nitrogen

adsorption isotherm. Relationship between the volume

Extraction of tobaccos with water. Each 30g of of pores and the volume of the adsorbed nitrogen is

the shredded flue cured and air cured tobaccos was given by Cranston and Inkley9) as following by com
immersed into 800ml of 60•Ž water for one hour with bining the B.E.T. multilayer adsorption and the capilla

occasional stirring to remove water soluble substances. ry condensation theories.

After repeating this procedure twice, the tobaccos were

washed several times with 60•Ž water and then dried

at 60•Ž.

Puffing. According to the Fredrickson method,14)

the shredded lamina of BY-4 was exposed to the vapor
of freon. After the freon was impregnated enough where

into the tobacco tissue, the shredded lamina was flash V12: specific volume of pores having a radius between

dried. The shredded midrib of BY-4, on the other r1 and r2 [cm3/g]

hand, was exposed to steam, swelled enough, and rl,r2: radii of pores filled by nitrogen through multi

layer adsorption and capillary condensation at

rapidly dried.
pressures P1 and P2, respectively [A]
Nitrogen adsorption isotherm. All the samples were tl,t2: thicknesses of the adsorbed nitrogen layers at

ground, and a fraction of between 32•`45 mesh was pressures P1 and P2, respectively [A]

used in this experiment. The nitrogen adsorption ƒË12: increment of adsorbed nitrogen according to
isotherm at the liquid nitrogen temperature (-195•Ž) change of pressure from P1 to P2 [ml/g]
 Porous Structure of Tobacco 985

rmas: radius of the largest pore [A]

The values of R12, K12
and (r-t12)/2r2 were experimental-
ly obtained as constants by Cranston and Inkley.9)
The specific pore volume was obtained as the sum of
the specific volumes of the pores less than 300 A in
diameter. A DIPS-1 computer was used for the
computation.

RESULTS

1) Nitrogen adsorption isotherms of tobaccos

As shown in Fig. 1, all nitrogen adsorption
isotherms of the tobaccos were of sigmoid
type, corresponding to the type II of the five
types classfied by Brunauer.111 However, the
quantity of adsorbed nitrogen varied with the
types of curing, that is, the air cured tobaccos FIG. 2. Effect of Water-extraction on Nitrogen

(Matsukawa and Burley-21) had the largest Adsorption Isotherm at -195°C.•›•\•›

nitrogen adsorption capacities, the flue cured , intact BY-4; •œ•\•œ, water extracted BY-4;
C-E, intact Matsukawa; •¡•\•¡, water extracted
(BY-4) and the sun cured (Kroumougrad) Matsukawa.
rather smaller capacities and the midrib the
lowest. By extracting the tobaccos with
water, the nitrogen adsorption capacities were
increased extremely in the case of the flue
cured, while slightly in the case of the air
cured as shown in Fig. 2. Effect of puffing is
also shown in Fig. 3.

2) Specific surface areas

Sample plots of P/ƒË(P0-P) ƒËs. P/P0 as

FIG. 3. Effect of Puffing on Nitrogen Adsorption

Isotherm at -195°C.•›•\•›

, intact BY-4; •œ•\•œ, freon treated BY-4;

• •\•, intact Midrib; •¡•\•¡, steam treated Midrib.

expressed by Eq. (1) are shown in Fig. 4.

The plots showed straight lines within a low

nitrogen pressure range. From their slopes

and the intercepts of the ordinate, the respec

tive values of ƒÊm were obtained. The specific

FIG. 1. Nitrogen Adsorption Isotherms of Various
surface areas were then calculated from Eq. (3)
Tobaccos at -195°C.• •\•
as shown in Table II. The values ranged from
, Kroumougrad; x-x, Burley-21; •›•\•›, BY-4;
6,000 to 17,000 cm2/g depending on the types
H, Midrib; ƒ¢•\ƒ¢, Matsukawa.
986 T. SAMEJIMA, Y. Sox and T. YANO

FIG, 6. Effect of Water-extraction on Pore Volume

FIG. 4. Plots of P/v (P0-P) vs. P/Pa.• •\•
Distribution Curve.

, Kroumougrad; x•\x, Burley-21; •›•\•›, BY-4;•œ•\•œ 1, intact BY-4; 2, water extracted BY-4; 3, intact
, Midrib; ƒ¢•\ƒ¢, Matsukawa. Matsukawa; 4, water extracted Matsukawa.‡™ƒË/‡™d

: Pore volume differentiated with pore dia

of curing in the ascending order of the midrib, meter.

the sun cured, the flue cured and the air cured
laminae. By extracting with water, the specific
surface area of the flue cured tobacco was
extremely increased, while that of the air
cured tobacco was increased only slightly,
corresponding to the increases in the specific
pore volume that are mentioned next. Effect
of puffing is also shown in Table II.

3) Specificpore volume distributions

The specific pore volume distributions are

FIG. 7. Effect of Puffing on Pore Volume Distri

bution Curve.

1, intact BY-4; 2, freon treated BY-4; 3, intact

Midrib; 4, steam treated Midrib.‡™ƒË/‡™d

: Pore volume differentiated with pore dia

meter.

shown in Figs. 5, 6 and 7. Every pore distri

bution curve showed one or two peaks ar-
FIG. 5. Pore Volume Distribution Curves of Various
round 50 A in diameter with gradual decrease
Tobaccos.
towards the larger and with sharp decrease
1, Kroumougrad; 2, BY-4; 3, Matsukawa; 4, Burley-
towards the smaller pore diameters. The
21; 5, Midrib.‡™ƒË

/‡™d: Pore volume differentiated with pore dia

values of the specific pore volume that are the
meter. integrals of each specific pore volume distri-
 Porous Structure of Tobacco 987

bution curve from zero to 300 A, are shown DISCUSSION

in Table II. They were in the ascending order

The differences in the specific surface areas
of the midrib, the sun cured, the flue cured and
and in the specific pore volumes caused by the
the air cured laminae. Extraction with water
different curing methods may be due to the
caused an extreme increase in the specific pore
loss of some chemical components during the
volume of the flue cured tobacco, while it
curing processes. It is known that starch and
caused only a slight increase in the specific
protein which are prominent in green leaves
pore volume of the air cured tobacco. By are mostly decomposed into such low mole
puffing the lamina of BY-4 with freon, the cular substances as sugars and amino acids
specificpore volume was increased up to about
during the yellowing stage of curing, and these
190%; while by puffing the midrib of BY-4
low molecular products disappear during the
with steam, the specific pore volume was in-
browning stage of curing. Actually the air
creased up to about 140%.
cured tobaccos contained less than 1% of

sugars, while the flue cured tobacco contained

4) Relationship between specific surface areas
about 20% of sugars. The slight and the
and speck pore volumes
great increments both in the specific surface
As shown in Fig. 8, the specific surface areas area and in the specific pore volume caused by
of tobaccos were highly correlated to the
the extraction of water soluble substances
specific pore volumes. By employing the
from the air cured and the flue cured tobaccos,
linear regression analysis, the relation between respectively, will also support the above ex
them was given as following.
planation.
In the correlation between the specific sur-

face area and the specific pore volume shown

by Eq. (5), the constant term 511.9 [cm2/g]

gives the specific surface area of the larger

pores more than 300 A in diameter. Since it

was below 8% of the total specific surface area,

it can be concluded that the specific surface

area of tobacco is mostly due to the micropores

smaller than 300 A in diameter. By assuming

that the pores are connected each other to make

a uniform long cylinder, the specific surface

area S [cm2/g] and the specific pore volume

Vg [cm3/g] can be given as following.

S= ƒÎ•Ed•EL x 10- 16 [cm2/g] (6)

FIG. 8. Relationship between Specific Surface Area
and Specific Volume of Pores Smaller than 300 A in Vg=ƒÎ•E(d/2)2 • L x 10-24 [cm3/g] (7)
Diameter of Tobaccos.
where
S, specific surface area [cm2/g]; Vg, specific pore
d: average pore diameter [A]
volume [cm3/g]; **, significant at I % level.
L: length of cylinder [A/g]
S=4609000Vg+511.9 (5)
By eliminating L from Eqs. (6) and (7),
where
S: specific surface area [cm2/g] S=4/d• Vg x 108 [cm2/g](8)

Vg: specific pore volume [cm3/g] By comparing the corresponding coefficients

The correlation coefficient was 0.977, showing of Vg in Eqs. (5) and (8), d=87 [A]. This

that the above relation was significant at 1% value may be slightly larger than that expected

level. from Figs. 5, 6 and 7. However, if the long

988 T. SAMEJIMA, Y. Sox and T. YANO
cylinder of pores is crossing over at places, the 2)
above calculation gives a larger value of d
3)
than the actual diameter of the cylinder. 4)
Therefore, the calculated value seems to be in
good agreement with the experimental results. 5)
The specific volume of the larger pores more
than 300 A in diameter could not be measured 6)
by the Inkley method. However, the contribu 7)
tion of these larger pores to the sorption and
desorption phenomena in tobacco will not be 8)
significant because their specific surface area 9)
was less than 8% of the total specific surface
10)
area. The adsorption or the capillary conden
sation of moisture and flavor ingredients will 11)
take place significantly on the micropores
smaller than 300 A in diameter. 12)
13)
REFERENCES
1) M. Samfield, Tob. Sci., 18, 21 (1974). 14)
E. E. Locklair, L. G. Veasey and M. Samfield,
Agr. Food Chem., 5, 294 (1957).
L. Hubschen, Tabak Forsch., 3, 25 (1965).
S. Kamei and S. Suzuki, Nippon Kogyokagaku
Kaishi, 44, 351 (1941).
C. S. Chang and W. H. Johnson, Tob. Sci., 17,
115 (1973).
C. H. Keith and C. F. De Laet, ibid., 10, 67
(1966).
B. R. Warner, Arch. Biochem. Biophys., 40, 143
(1952).
E. V. Parups, Tob. Sci., 2, 64 (1958).
R. W. Cranston and F. A. Inkley, Adv. Canal., 9,
143 (1957).
T. Keii, "Kyuchaku," Kyoritsu Publishing Co.,
Tokyo, 1973, p. 95.
H. K. Livingston, J. Am. Chem. Soc., 66, 569
(1944).
W. G. Frankenburg, Adv. in Enzymol., 6, 309
(1946).
S, Brunauer, P. H. Emmett and E. Teller, J, Am.
Chem. Soc., 60, 309 (1938).
J. D. Fredrickson, U. S. Patent, 3524451 (1970).