Applied Surface Science 257 (2011) 41654170

Contents lists available at ScienceDirect

Applied Surface Science

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

Surface treatment of aramid ber by air dielectric barrier discharge plasma at

atmospheric pressure
Caixia Jia a , Ping Chen a,b, , Wei Liu a,c , Bin Li a , Qian Wang a
a
Key Laboratory of Materials Modication by Laser, Ion and Electron Beams (Ministry of Education) & Faculty of Chemical, Environmental and Biological Science and Technology,
Dalian University of Technology, Dalian 116024, China
b
Liaoning Key Laboratory of Advanced Polymer Matrix Composites Manufacturing Technology, Shenyang Aerospace University, Shenyang 110034, China
c
Dalian University of Education, Dalian 116021, China

a r t i c l e

i n f o

Article history:
Received 1 October 2010
Received in revised form
26 November 2010
Accepted 29 November 2010
Available online 9 December 2010
Keywords:
Aramid ber
Air DBD plasma
Treatment time
Surface characterization
Wetting behavior

a b s t r a c t
Aramid ber samples are treated by air dielectric barrier discharge (DBD) plasma at atmospheric pressure;
the plasma treatment time is investigated as the major parameter. The effects of this treatment on the
ber surface physical and chemical properties are studied by using surface characterization techniques.
Scanning electron microscopy (SEM) is performed to determine the surface morphology changes, X-ray
photoelectron spectroscopy (XPS) is analyzed to reveal the surface chemical composition variations and
dynamic contact angle analysis (DCAA) is used to examine the changes of the ber surface wettability. In
addition, the wetting behavior of a kind of thermoplastic resin, poly(phthalazinone ether sulfone ketone)
(PPESK), on aramid ber surface is also observed by SEM photos. The study shows that there seems to be
an optimum treatment condition for surface modication of aramid ber by the air DBD plasma. In this
paper, after the 12 s, 27.6 W/cm3 plasma treatment the aramid ber surface roughness is signicantly
improved, some new oxygen-containing groups such as CO, C O and O CO are generated on the ber
surface and the ber surface wettability is greatly enhanced, which results in the better wetting behavior
of PPESK resin on the plasma-treated aramid ber.
2010 Elsevier B.V. All rights reserved.

1. Introduction
Aramid ber, combining high specic modulus and strength
with excellent thermal stability and chemical inertness, is one of
the best candidates as reinforcement for high-performance composites. However, the employing aramid ber as reinforcement
has been limited by poor ber/matrix interfacial adhesion, because
of the ber having smooth surface and inert chemical structure,
which prevents interfacial bonding with most of the commercially
available resins used in composites [15]. In order to optimize the
aramid-ber/matrix adhesion, a variety of research efforts have
been directed towards the improvement of ber surface properties by various surface treatment techniques [611]. Traditional
chemical treatments are often accompanied by the use of water
and/or organic solvents, thus pose such problems as the disposal
of drained water and recovery of organic solvents, which are
energy-, time- and cost-intensive and involve pollution of water
resources. Therefore, non-thermal plasma technology, which is ef-

Corresponding author at: Dalian University of Technology, Department of Polymer Material, Zhongshan Road 158-42#, Dalian, Liaoning 116012, China.
Tel.: +86 0411 89393866.
E-mail address: chenping 898@126.com (P. Chen).
0169-4332/$ see front matter 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apsusc.2010.11.190

cient and environmentally friendly, has been extensively studied

[12,13].
Plasma techniques may work in very exible conditions. In
comparison with that at low and medium pressure, plasmas at
atmospheric pressure are predominantly studied recently, e.g.
dielectric barrier discharge (DBD) plasma, which can not only separate the ber bundles into individual laments and modify the
ber surface without altering the bulk properties, but also has the
advantages of operating at atmospheric pressure and easy formation of uniform and stable plasmas, avoiding the need of expensive
vacuum equipment and permitting continuous processing of ber
surfaces. This approach has proved to be a promising method and is
increasingly used for polymeric ber or lm surface modications
[1416]. However, until now, only little research [17,18] is done
on the relationship between the air DBD plasma treatment and the
surface properties of aramid ber.
In this paper, the inuence of DBD plasma treatment in air
at atmospheric pressure on aramid ber surface is studied. Various different treatment durations have been explored to optimize
the modication, and the physical and chemical changes on the
treated ber surface induced by the plasma are investigated.
The surface morphology was examined using scanning electron
microscopy (SEM), the surface chemical composition was analyzed
by X-ray photoelectron spectroscopy (XPS), and the surface wet-

4166

C. Jia et al. / Applied Surface Science 257 (2011) 41654170

mental atmosphere according to the post-plasma oxidation effect

[19].
2.3. SEM

Fig. 1. Schematic representation of the air DBD plasma set-up.

Scanning electron microscopy (SEM; QUANTA 200, FEI) was

selected to observe the surface morphologies of Twaron ber. And
the wetting behavior of the PPESK resin on the ber laments
was also studied by SEM photos; the samples were prepared by
impregnating the laments with PPESK solution which was made
by dissolving the PPESK resin into N,N-dimethylacetamide (DMAc)
solvent, and then drying the ber/resin samples in an air oven
(120 C/1 h, 175 C/3 h) to remove DMAc completely; it is important
to note that the volume fraction of Twaron ber in every sample was
set to about 60%. The pressure of the chamber for SEM measurement
was less than 60 Pa and the magnication was set at 5000.
2.4. XPS

tability was measured by dynamic contact angle analysis (DCAA).

In addition, the wetting behavior of a kind of thermoplastic resin,
poly(phthalazinone ether sulfone ketone) (PPESK), on aramid ber
surface was also illustrated by SEM photos.
2. Experimental
2.1. Materials
Twaron aramid yarn used in this study was received from
Akzo Nobel Co. Ltd., Arnhem. The ber samples were cleaned with
acetone at room temperature before proceeding air DBD plasma
surface modication in order to remove the surface oily nish, and
then they were dried in an air oven at 110 C for 3 h. The resin of
which the wetting behavior on Twaron ber was observed was
thermoplastic poly(phthalazinone ether sulfone ketone), which
was supplied by Dalian Polymer New Material Co. Ltd., China.
2.2. Air DBD plasma treatment
The schematic diagram of the dielectric barrier discharge apparatus is shown in Fig. 1 and the detailed description about the set-up
was reported in previous work [18]. When a high AC voltage with an
output frequency of 27 kHz was applied continuously, a lamentary
dielectric barrier discharge would take place, which was uniform
and stable. Twaron ber samples were treated by means of going
through the DBD plasma region at a constant speed. During the
treatment process the system was always exposed to atmosphere
through the air hole of the DBD experiment set-up.
The essential parameters about the DBD apparatus are shown
in Table 1. The treatment durations were successively set as 6 s,
12 s and 18 s, which was carried out by the ber entrainment speed
being 1.56 cm/s and the ber going through the discharge region
two, four, and six times respectively with every treatment time
lasting for 3 s. And it is important to note that after the air DBD
plasma treatment the ber samples kept staying in the DBD set-up
for 12 h in order to incorporate more oxygen from the environTable 1
Essential parameters about the DBD treatment in this study.
Parameters

Description

Electrode diameter (material)

Discharge gap
Barrier thickness (material)
Discharge power (power density)
Discharge atmosphere and pressure
Treatment time

4.7 cm (steel)
0.3 cm
0.1 cm (quartz)
143.5 W (27.6 W/cm3 )
Air and at atmosphere pressure
0 s (untreated), 6 s, 12 s, 18 s

The surface chemical composition of Twaron ber was analyzed by X-ray photoelectron spectroscopy (XPS; ESCALAB 250,
Thermo), making use of monochromatic Al K (h = 1486.6 eV)
X-ray source (15 kV, 250 W) radiation from a dual AlMg anode.
The measurements were performed at an operating vacuum lower
than 3.0 109 mbar. Spectra were acquired at a take-off angle of
90 relatively to the sample surface. The pass energy and energy
step set for elemental quantication were 100 eV and 1 eV, respectively, and for C1s peak shape identication purpose the pass
energy and energy step were xed at 20 eV and 0.05 eV, respectively. The non-linear least squares tting (NLLSF) program with a
GaussianLorentzian production function was used for curve tting
of C1s spectra.
2.5. DCAA
The surface free energy and contact angles of Twaron ber were
measured through a dynamic contact angle analysis system (DCAA;
DCA-322, Thermo). The ber sample was cut into about 1 cm in
length and xed indirectly to a wire hook suspended from the
microbalance of the system. The ber was immersed into the testing liquid media by raising the elevating stage at a constant speed of
1 mm/min and then the dynamic contact angles () were obtained
schematically by the measurement. There are two equations (Eqs.
(1) and (2)) as follows, from which the surface free energy can be
calculated:

l (1 + cos ) = 2

s l + 2

sd ld

total = s + sd

(1)
(2)

where  is the dynamic contact angle between ber and testing

liquid which is calculated by the computer program,  l is the surface tension of the testing liquid,  total stands for total surface free
p
energy of the ber, and s and sd are the polar component and
the dispersive component of total surface free energy, respectively
[20]. In our experiment, water (polar solvent; its surface tension
is 72.3 mN/m) and diiodomethane (non-polar solvent; its surface
tension is 50.8 mN/m) were chosen as testing liquids.
3. Results and discussion
3.1. Surface morphology of Twaron ber
SEM was used to investigate the surface morphology changes of
Twaron ber before and after the air DBD plasma treatment. In Fig. 2
is shown a comparison of the SEM images of the untreated ber and

C. Jia et al. / Applied Surface Science 257 (2011) 41654170

4167

Fig. 2. SEM photographs of the untreated and air DBD plasma treated ber surface: (a) 0 s (untreated); (b) plasma treated for 6 s; (c) plasma treated for 12 s; (d) plasma
treated for 18 s.

the ber samples after the plasma treatment for 6 s, 12 s and 18 s

at 27.6 W/cm3 . It can be seen that after 6 s treatment (Fig. 2b) the
surface of the ber is uneven and clearly different from that of the
untreated one (Fig. 2a), which is clean and smooth. When the ber is
treated for 12 s by the air DBD plasma (Fig. 2c), the irregular surface
with apparent bulges and ruts can be observed. However, when the
treatment time increases further to 18 s (Fig. 2d), it is surprising that
there is obvious damage present on the ber involving the surface
peeling off, which maybe implies that the bulk is affected by the
long time plasma treatment.
The results show that surface roughness of Twaron ber
increases with increasing plasma treatment time. The changes
of Twaron ber surface features before and after the treatment
are mainly caused by the etching effects and oxidative reactions
induced by the plasma processing. The results also show that there
may be an optimum treatment time of the air DBD plasma for the
modication of Twaron ber surface physical properties.

3.2. Surface chemical composition of Twaron ber

XPS technique was used to analyze the surface chemical composition of Twaron ber. The results of XPS analysis of the untreated
and plasma-treated ber samples are given in Table 2. It can be
seen that the surface chemical composition of Twaron ber experiences little change after the air DBD plasma treatment for 6 s;
the slight decrease of oxygen (O) concentration and the ratio of
oxygen to carbon atoms (O/C) may be due to the removal of the

Table 2
XPS surface element analysis of Twaron ber before and after treatment.
Samples

0 s (untreated)
Plasma treated for 6 s
Plasma treated for 12 s
Plasma treated for 18 s

Chemical composition [at.%]

Atomic ratio

O/C

N/C

78.9
79.3
75.5
83.1

14.0
13.5
17.6
12.6

7.1
7.2
6.9
4.3

0.177
0.170
0.233
0.152

0.090
0.091
0.091
0.052

small amount of residual oily nish arriving from the commercial

ber surface by plasma etching effects [19]. After the plasma treatment time increases to 12 s, the O concentration increases evidently
from 13.5% to 17.6% and the O/C increases from 0.170 to 0.233;
the nitrogen (N) concentration and the ratio of nitrogen to carbon atoms (N/C) have not shown obvious change. However, when
the treatment time increases from 12 s to 18 s, it is observed that
the C concentration increases and O and N concentrations decrease
sharply with the O/C from 0.233 to 0.152 and N/C from 0.091 to
0.052.
The changes of ber surface chemical composition show that a
suitable processing time (12 s in this study) of air DBD plasma can
introduce much oxygen onto aramid ber surface and increase the
ratio of oxygen to carbon atoms, the shorter treatment time (6 s in
this study) does not have signicant inuence on the ber surface
chemical composition, and a long treatment time (18 s in this study)
might cause the decreases of oxygen and nitrogen concentrations,
which can be explained by the mechanism that surface modica-

4168

C. Jia et al. / Applied Surface Science 257 (2011) 41654170

Fig. 3. C1s spectra of the untreated and air DBD plasma treated ber surface: (a) 0 s (untreated); (b) plasma treated for 6 s; (c) plasma treated for 12 s; (d) plasma treated for
18 s.

tion is dominant at beginning and then is overwhelmed by surface

etching at a later stage of the process [21]. These observations are
complemented by the change of the XPS C1s spectrum as a function of treatment time, as shown in Fig. 3. The C1s spectrum of the
ber samples can be tted to four peaks with binding energies of
about 284.6 eV (CC), 286.0 eV (CO), 287.8 eV (C O) and 289.0 eV
(O CO), and the concentrations of these carbon-containing functional groups on aramid ber surface are given in Table 3 which are
calculated from the relevant peak areas in the C1s spectrum.
The peak intensities and peak areas of C1s spectrum experience
distinct variation from the untreated to the plasma treated samples.
Compared to the untreated ber (Fig. 3a), the CC concentration
decreases sharply from 74.1% to 68.5%, while the concentrations of
oxygen functional groups such as CO, C O and O CO increase
from 13.3%, 6.0% and 6.6% to 17.8%, 6.8% and 6.9%, respectively,
after the air DBD plasma treatment for 12 s (Fig. 3c). In addition,
it can be found that the contributions of CO and O CO decline
after the plasma treatment for 6 s (Fig. 3b) owing to the removal
of the residual surface oily nish and the contributions of oxygencontaining groups obviously decrease after the plasma treatment
for 18 s (Fig. 3d) caused by the plasma etching effects, which agrees
well with the discussion above about the variations of surface com-

Table 3
Correlative functional groups of Twaron ber surface.
Samples

0 s (untreated)
Plasma treated for 6 s
Plasma treated for 12 s
Plasma treated for 18 s

Concentrations of correlative functional

groups (%)
CC

CO

C O

O CO

74.1
77.5
68.5
89.3

13.3
9.3
17.8
5.3

6.0
8.5
6.8
3.6

6.6
4.7
6.9
1.8

position. The results indicate that some new oxygen-containing

polar groups (CO, C O and O CO) can be introduced to Twaron
ber surface by the ions, electrons and UV radiation [10] after the
12 s air DBD plasma treatment, which could enhance the adhesion
properties of aramid ber via improving the ber surface chemical
inertness and increasing the ber surface free energy [21].
3.3. Surface wettability of Twaron ber
Surface free energy of aramid ber, which can be derived from
the dynamic contact angle measurements, is used to study the
effects of different air DBD plasma treatment durations on the
ber surface wettability. The advancing contact angles and surface
free energy, consisting of polar and dispersive components, of the
untreated and plasma treated ber samples are given in Table 4.
The advancing contact angles () of water decline by degrees
from 64.3 for the untreated sample to 42.3 for the sample after
a 27.6 W/cm3 and 12 s air DBD plasma treatment, and the polar
component ( p ) of surface free energy and the total surface free
energy ( total ) increase gradually from 11.5 mJ/m2 and 50.6 mJ/m2
to 24.4 mJ/m2 and 62.6 mJ/m2 , respectively. As for the advancing
contact angles of diiodomethane and the dispersive component
( d ) of surface free energy, the changes are not distinct. When the
ber is treated for 18 s by the air DBD plasma, it can be seen that
the total surface free energy increases further to 66.6 mJ/m2 . This
could be explained by the fact that the air DBD plasma increases
the surface roughness of Twaron ber as illustrated by SEM photos and introduces some new polar groups onto the ber surface as
shown by XPS measurements, together improving the ber surface
wettability. After the 18 s air DBD plasma treatment, Twaron ber
surface morphology is changed to a great extent (Fig. 2d), which
might bring the ber higher surface free energy but implies that
the bulk properties may be affected.

C. Jia et al. / Applied Surface Science 257 (2011) 41654170

4169

Table 4
Advancing contact angles and surface free energy of Twaron ber before and after treatment.
Samples

0 s (untreated)
Plasma treated for 6 s
Plasma treated for 12 s
Plasma treated for 18 s
a

Contact angle () SDa ( )

Surface free energy (mJ/m2 )

Water

Diiodomethane

p

d

 total =  p +  d

64.3(2.5)
48.6(2.4)
42.3(0.4)
35.9(2.7)

41.0(2.0)
47.4(1.8)
42.7(2.7)
41.1(2.3)

11.5
21.8
24.4
27.5

39.1
35.7
38.2
39.1

50.6
57.5
62.6
66.6

Standard deviation.

3.4. The wetting behavior of PPESK resin on Twaron ber

As for the purpose of this paper, the modication of air DBD
plasma on Twaron ber surface is to make the treated ber have
practical application as reinforcement agents in resin matrix composites. We all know that the ber/matrix interface governs the
overall performance of composites, and from a practical point of
view, good wetting behavior of resin on ber surface plays an
important role in the formation of interface with excellent performance. In this paper, the inuence of air DBD plasma treatment
times on the wetting behavior of a kind of thermoplastic resin,
poly(phthalazinone ether sulfone ketone), on Twaron ber was also
studied by SEM measurements. The surface morphologies of the
untreated and plasma treated Twaron laments impregnated with
PPESK resin are shown in Fig. 4.
The four samples all show that there is more or less PPESK
resin adhering to Twaron ber surface before and after the air DBD
plasma treatment. For the untreated sample (Fig. 4a), only a lit-

tle resin is on the ber surface and some smooth ber surface is
visible. After the ber is treated for 6 s (Fig. 4b), there is a little
more PPESK resin being observed and we can also see some undulations on the sample. When the ber is treated for 12 s (Fig. 4c),
the ber is covered with more resin evenly and it can be found
that the resin is present among the ber laments as well. After
the treatment time lasts for 18 s (Fig. 4d), there is a mass of PPESK
resin on the ber surface, however, the resin has a nonuniform
distribution.
The results indicate that the PPESK resin could show a better wetting behavior on Twaron ber surface when the ber is
treated for 12 s by the air DBD plasma at 27.6 W/cm3 , which agrees
with the results of SEM, XPS and DCAA measurements shown in
above sections, showing that suitable treatment of the plasma can
increase the surface roughness of Twaron ber and improve the
surface chemical inertness, enhancing the ber surface wettability and leading to good interfacial adhesion between the ber and
matrix.

Fig. 4. The wetting behavior of PPESK resin on Twaron ber surface: (a) 0 s (untreated); (b) plasma treated for 6 s; (c) plasma treated for 12 s; (d) plasma treated for 18 s.

4170

C. Jia et al. / Applied Surface Science 257 (2011) 41654170

4. Conclusions
The experimental results presented in this study show that the
air DBD plasma at atmospheric pressure can effectively modify
Twaron ber surface adhesion properties, the plasma treatment
time has great inuence on ber surface physical features and
chemical compositions and there might be an optimum treatment
condition existing for the ber modication. The results indicate
that after a 12 s, 27.6 W/cm3 plasma treatment, the aramid ber
surface roughness is evidently improved as illustrated by SEM,
the concentrations of oxygen and oxygen-containing functional
groups on the ber surface increase signicantly as determined
with XPS, and the ber surface wettability derived from DCAA is
greatly enhanced. The improvements of the physical and chemical
properties induced by the air DBD plasma could account for the
better wetting behavior of PPESK resin on plasma-treated Twaron
aramid ber.
Acknowledgements
This work was funded by the National Natural Science Foundation of China (No. 50743012), National Defense 11th 5-year
Program Foundational Research Program (No. A3520060215), and
Liaoning Province Innovation Organization (No. LT2010083). In
addition, the authors are indebted to Mr. Xinglin Li for XPS analysis
and Dr. Litao Shi for his skillful experimental assistance.
References
[1] J.S. Lin, Effect of surface modication by bromination and metalation on Kevlar
bre-epoxy adhesion, Eur. Polym. J. 38 (2002) 7986.
[2] L. Liu, Y.D. Huang, Z.Q. Zhang, Z.X. Jiang, L.N. Wu, Ultrasonic treatment of aramid
ber surface and its effect on the interface of aramid/epoxy composites, Appl.
Surf. Sci. 254 (2008) 25942599.
[3] P.J. de Lange, E. Mder, K. Mai, R.J. Young, I. Ahmad, Characterization and
micromechanical testing of the interphase of aramid-reinforced epoxy composites, Compos. Part A 32 (2001) 331342.
[4] S.P. Lee, K.S. Chian, C.Y. Yue, H.C. Looi, Effects of bromination and hydrolysis
treatments on the morphology and tensile properties of Kevlar-29 bres, J.
Mater. Sci. Lett. 13 (1994) 305309.

[5] S.J. Park, M.K. Seo, T.J. Ma, D.R. Lee, Effect of chemical treatment of Kevlar bers
on mechanical interfacial properties of composites, J. Colloid Interface Sci. 252
(2002) 249255.
[6] H. Watanabe, M. Furukawa, T. Takata, M. Yamamoto, Surface improvements of aramid bers by physical treatments, Macromol. Symp. 159 (2000)
131141.
[7] T.K. Lin, S.J. Wu, J.G. Lai, S.S. Shyu, The Effect of chemical treatment on reinforcement/matrix interaction in Kevlar-ber/bismaleimide composites, Compos.
Sci. Technol. 60 (2000) 18731878.
[8] P.A. Tarantili, A.G. Andreopoulos, Mechanical properties of epoxies reinforced with chloride-treated aramid bers, J. Appl. Polym. Sci. 65 (1997)
267276.
[9] R.J. Day, K.D. Hewson, P.A. Lovell, Surface modication and its effect on the
interfacial properties of model aramid-bre/epoxy composites, Compos. Sci.
Technol. 62 (2002) 153166.
[10] S. Mukhopadhyay, R. Fangueiro, Physical modication of natural bers and
thermoplastic lms for compositesa review, J. Thermoplast. Compos. Mater.
22 (2009) 135162.
[11] Y. Katoh, T. Nozawa, L.L. Snead, T. Hinoki, Effect of neutron irradiation on tensile
properties of unidirectional silicon carbide composites, J. Nucl. Mater. 367370
(2007) 774779.
[12] R.Z. Li, L. Ye, Y.W. Mai, Application of plasma technologies in bre-reinforced
polymer composites: a review of recent developments, Compos. Part A 28A
(1997) 7386.
[13] N. De Geyter, R. Morent, C. Leys, Surface modication of a polyester non-woven
with a dielectric barrier discharge in air at medium pressure, Surf. Coat. Technol.
201 (2006) 24602466.
[14] K.L. Wang, W.C. Wang, D.Z. Yang, Y. Huo, D.Z. Wang, Surface modication of
polypropylene non-woven fabric using atmospheric nitrogen dielectric barrier
discharge plasma, Appl. Surf. Sci. 256 (2010) 68596864.
[15] G. Borcia, C.A. Anderson, N.M.D. Brown, Surface treatment of natural and synthetic textiles using a dielectric barrier discharge, Surf. Coat. Technol. 201
(2006) 30743081.
[16] G. Borcia, C.A. Anderson, N.M.D. Brown, Using a nitrogen dielectric barrier discharge for surface treatment, Plasma Sources Sci. Technol. 14 (2005) 259267.
[17] M. Xi, Y.L. Li, S.Y. Shang, D.H. Li, Y.X. Yin, X.Y. Dai, Surface modication of aramid
ber by air DBD plasma at atmospheric pressure with continuous on-line processing, Surf. Coat. Technol. 202 (2008) 60296033.
[18] C.X. Jia, P. Chen, B. Li, Q. Wang, C. Lu, Q. Yu, Effect of Twaron ber surface
treatment by air dielectric barrier discharge plasma on the interfacial adhesion
in ber reinforced composites, Surf. Coat. Technol. 204 (2010) 36683675.
[19] L. Tusek, M. Nitschke, C. Werner, K.S. Kleinschek, V. Ribitsch, Surface characterisation of NH3 plasma treated polyamide 6 foils, Colloids Surf. A: Physicochem.
Eng. Aspects 195 (2001) 8195.
[20] P. Chen, C. Lu, Q. Yu, Y. Gao, J.F. Li, X.L. Li, Inuence of ber wettability on the
interfacial adhesion of continuous ber-reinforced PPESK composite, J. Appl.
Polym. Sci. 102 (2006) 25442551.
[21] C.-M. Chan, T.-M. Ko, H. Hiraoka, Polymer surface modication by plasmas and
photons, Surf. Sci. Rep. 24 (1996) 154.