- Email: [email protected]
Characteristics of polymer welding by healing process
ELSEVIER
Materials
Chemistry
and Physics 48 ( 1997) 90-93
Characteristics of polymer welding by healing process Mi-Ja Shim a, Sang-Wook Kim bl* a Department of Lge Science, Seoul City University, 90 Jeonnong-Dong, Dongdaemctn-Gu, Seoul 130-743, South Korea b Department of Chemical Engineering, Seoul City University, 90 Jeonuong-Dong, Dongdaemun-Gu, Seoul 130-743, South Korea Received
8 January
1996; revised 3 May
1996
Abstract A self-bondingstrengthdevelopment of ABS duringhot plateweldingwasinvestigated by alap-shear joint method.In thehot platewelding, the flow of moltenABS occursandit is importantfactor of self-bondingstrengthdevelopment. The flow activationenergies,E, obtained from relationof viscosityandtemperature were4-8 kcal mol-’ at differentshearrates.The self-bondingstrengthincreasedwith increasing time andtemperatureandits developmentat temperature rangefrom 180°Cto 260°Cwasfoundto follow a t”4 law basedon the reptation modelappliedto interdiffusionat anamorphous polymer-polymerinterface.In fractureof lapjoint samples, shearfailureandtensilefailure wereobserved.In tensilefailure,the crack-likeflaw at the edgeof bondwasthoughtto act asa stressconcentrator.Crackspropagated from it. The numberanddepthof plasticallydeformedridgesat there-heatedsurfaceby shearwereincreased with weldingtimeandtemperature. Keywords:
Welding;
ABS: Lap-shear
joint method;
Self-bonding
strength
1. Introduction When two samplesof the sameamorphouspolymer are contacted at a temperatureabove the glasstransition temperaturewithout adhesives,the junction surfacegradually develops increasing mechanical strength across a polymerpolymer interface until the mechanicalstrength reachesthat of the bulk polymer [ I] . It is saidto behealed.The molecular model of healing in polymer was explained by the diffusion of polymer chains across the interface of the contacting polymers [ 21. The molecular interpenetration is related to the phenomenon of self-diffusion. To analyze the self-diffusion along the chain and to describethe features of chain motion composed of entangledrandom coil, reptation model [ 31, crack healing theory [43 and minor chain model {5] were alsoproposed. Thesehealing processesprovided a significancefor polymer processing[ 6,7] such assurfacejoining by welding or lamination and preparation of a bulk polymer from polymer pellets. These methods make use of the inherent properties of the materials. Recently, aircraft or automobile applications have called for the ability to satisfy the mechanicalstrength requirementsat a lower weight and thermosetsasmatrix are not able to be recycled and are brittle. So, several thermoplastics with good thermal and mechanicalpropertiescanbe * Corresponding
author.
0254-0584/97/$17.00 PHSO254-0584(96)01858-5
0 1997 Elsevier
Science S.A. All rights reserved
expectedto replacethe thermosets.Thesematerialsare easy to be recycled, molded and able to be applied to high performancesby blending, compoundingand reinforcing [ 81, In this study we investigated the self-bonding strength developmentat varioustemperaturesduring hot platewelding of ABS (acrylonitrile-butadiene-styrene terpolymer) and useda lap-shearjoint method [ 91. The factors affecting the self-bonding strength such as melt flow, welding time and welding temperaturewere discussedandfracture behavior by shearstresswasalso observed.
2. Experimental 2.1. Materials
The ABS was a commercial pellet type ABS, HF-06601 from Cheil Industries Inc. with butadienecontent of 14.4%, acrylonitrile content of 21%. An average molecular weight of free SAN, K, is 95000. 2.2. Method
An appropriateamount of ABS in the shapeof pellet was melted at 16O”C,0 MPa, for 10 min in a preheatedhot press and was molded with increasing pressure to 5.5 MPa for 30 min. Compression-moldedABS was cut and sharpened
M.-J. Shim, S.-W. Kim/Materials
Fig. 1. Form
and dimension
of single lap-shear
Chemistry
and Physics 48 (1997)
90-93
91
joint sample
after slow cooling to a room temperature. The dimension of sample was based on the ASTM D3 163 as shown in Fig. 1. A hot plate was preheated to a temperature from Tg + 50°C 150 to 300°C. Specimens were placed on the hot plate for 10, 20, 30,40,50 and 60 min. Healing was conducted at atmosphere pressure. Two specimens were overlapped and provided about 0.1 MPa of contact pressure for instantaneous wetting. Self-bonding strengths of lap-shear specimens were estimated on the Shimadzu Autograph (Model AGS-1000D) using the stoke-tension mode at a crosshead speed of 0.5 mm min-‘.
3. Results and discussion In a hot plate weiding, the flow of molten ABS occurred. In order for a polymer melt to flow, the chain segments must be able to move. It was required that the chain segments had to have sufficient thermal energy to overcome energy barriers. At a small amount of flow, the small interpretation of the flowing material resulted in a lower bonding strength. At a large amount of flow, severe deformation of shape occurred during welding. Therefore, the flow pattern was an important factor for good welding and affected the features of the surface deformed by heat welding. Therefore, the activation energy to overcome energy barriers should be considered. The flow activation energy E, can be obtained from the general Arrhenius equation for molten polymers [ lo].
(1)
Fig. 2. Variation
of viscosity
with shear rate:
q , 220°C;
0, 250°C; 0 280°C.
Microscopically, the motion of a chain entangled with many other chains occurs rapidly by thermal fluctuations. In a reptation model for healing by de Gennes [ 31, each polymer chain is considered to have a tube-like region along the length in which it executes a random back-and-forth motion. Crack healing theory proposed by Wool and O’Connor [ 41 is composed with five stages of (A) rearrangement, (B) surface approach, (C) wetting, (D) diffusion and (E) randomization. The potential barrier related to inhomogeneities of the polymer interface disappears at the end of wetting and the polymer chains are free from the moving across the interface by diffusion and randomization stage which is very important on account of the developed strength. In symmetric interfaces, the rate determining steps were wetting and diffusion. Therefore, it can be considered that the self-bonding strength development of ABS was completed by a two-stage process. The stress components acting on the faces of the element are normal stresses, flrr, crYYand LT= and shear stresses r,,,, rrZ and r,,. These quantities are necessary to specify the state of stress at a point of sample. The stresses induced in a lapshear sample loaded in tension are complicated. Assuming forces are transmitted through the y-direction and no relative displacement occurs in either the x or the z direction, only shear stresses in the yz plane (welded area, A) are effective
where C is a constant and 77is the viscosity of a polymer. To calculate the activation energy, Eq. ( 1) is rearranged.
142kec-‘) 271bec-‘)
In 7)=+T+ln
C
720(sec-‘) 77msec-‘) 1600(sec-‘) 1600~sec-‘)
Viscosity as a function of shear rate is obtained from the capilfary rheometer [ 111 and it is shown in Fig. 2. Viscosity decreases with the increasing shear rates and temperatures. A relationship between In 71 and l/T is plotted and it is linear as shown in Fig. 3. From the slope, the activation energy can be obtained. The values of the activation energy, E,, are 4-8 kcal mol- ‘, Over these values, the ABS melt flows and the polymer chains interpenetrate into each other, so welding has happened.
f /
/
/ Ew30hec“)
/ 5.0
’ 1.5
1.7
1.9
l/T
2.1
x
2.3
I 2.5
IO’(K)
Fig. 3. A plot of In r) vs. 1 iT at various
shear rates.
M.-J. Shim, S.-W. Kim/Materials
Chemistry
and Physics 48 (1997)
90-93
Fig. 5 shows a load-displacement curve in fracture of lapshear joint specimens. In fracture of lap joint samples, shear failure and tensile failure were observed. In tensile failure [ 16-181, a distinct flaw appeared at the edge of the bond in which samples slowly bent and cut off after the stress reached a maximum value. These crack-like flaws acted as a stress concentrator and the crack propagated from them.
Fig. 4. Self-bonding strength several constant temperatures: W, 260°C.
0
y 0
Fig. 5. Load-displacement ABS.
vs. welding 0, 18OT;
1
time to the one-fourth Cl, 2OOT; A, 220°C;
1 2
l
power at ,240”C;
I 3
Displacement(mm) curve in fracture of lap shear joint
samples
of
on the recorded failure load. Hence, the shear stress, rYzacting on the yz plane is given by [ 12,131
‘=A
F (failure load) (welded area)
7rz is a self-bonding expressed as 7= 5-o+ 3-6
(3) strength to be determined
and is (4)
where rO is a self-bonding strength due to the wetting and TV is due to the diffusion of chain penetrating. However, the value of meis so low it can be ignored; therefore r. is removed from Eq. (4). The use of the reptation model yields xa t1’4 and ra x, where x is the interpenetration length at polymerpolymer interface. Eq. (4) is rearranged as { 14,151 7= &I4
(5)
where K is a constant and t is the welding time. Eq. (5) represents the time dependence of self-bonding strength. Fig. 4 shows a self-bonding strength vs. welding time to the one-fourth power at various temperatures. The self-bonding strength development was found to follow a t”4 law based on the reptation model.
Fig. 6. SEM photograph of plastically deformed 30 min; (b) 26O”C, 30 min; (c) 26O”C, 1 h.
welded
area: (a)
18O”C,
M.-J.
Shim, S.-W. Kim/Materials
Chemistry
and Physics 48 (1997)
90-93
93
3. In fracture, shear failure and tensile failure were observed. The latter was caused by a stress concentration effect at the edge of bond. A crack-like flaw at the edge of the bond acted as a stress concentrator. The number and depth of plastically deformed ridges, which were the main features of the welded samples, depended on the welding time and temperature. The fracture surface exhibited stress whitening.
Acknowledgements The authors acknowledge Kyong group in Korea. Fig. 7. SEM photograph
of stress whitening
at deformed
financial
support
from Sun
surface.
References When the lap-shear joint samples were fractured by shear, plastically deformed ridges which were the main features of the welded samples were observed as shown in Fig. 6. The number and depth of ridges increased with welding time and temperature, i.e. the self-bonding strength is related to the number and depth of plastically deformed ridges and flow patterns of molten polymers [ 191. A stress whitening in Fig. 7 was also observed in a deformed surface at the edge of bond.
4. Conclusions 1. In hot plate welding, the flow of molten ABS occurred and it was an important factor in self-bonding strength development. Viscosity decreased with increasing shear rate and temperature. Activation energies obtained from the relation of viscosity and temperature were 4-8 kcal mol-’ at various shear rates. 2. The self-bonding strength development was found to follow a t”4 law based on the reptation model applied to interdiffusion at an amorphous polymer-polymer interface.
[l] S. Prager and M. Tirrell; J. Chem. Phys., 75 (1981) 5194. [2] S.S. Voyutskii, Autoadhesion and Adhesion of High Polymers, WileyInterscience, New York, 1963. [3] P.G. deGennes, J. Chem. Phys., 55 (1971) 572. [4] R.P. Wool and K.M. O’Connor, J. Appl. Phys., 52 (1981) 5953. [5] Y.H. Kim and R.P. Wool, Macromolecules, I6 (1983) 1115. [6] R.J. Crawford and Y. Tam, J. Muter. Sci., 16 (1981) 3275. [7] K. Jud, H.H. Kauschand J.G. Williams,J. Muter. Sci., I6 (1981) 204. [S] IS. Chun, M.J. Shim and S.W. Kim, Proc. IUMRNCA’95, Seoul, 1995, p. 521. [9] D.B. Kline and R.P. Wool, Polym. Eng. Sci., 28 (1988) 52. [ 101 L.E. Nielsen, Polymer Rheology, Marcel Dekker, New York, 1977. [ 1 l] E.B. Bagley, J. Appl. Phys., 28 (1957) 624. [ 121 J.L. Lubkin and E. Reissner, J. Appl. Mech., Trans. ASME, 24 (1957) 255. [ 131 I.N. Sneddon, in D.D. Eley (ed.), Adhesion, Oxford University Press, London, 1961, [ 141 S.F. Edwards, Proc. Phys. SOL, 92 (1967) 9. [ 151 J. Klein, Macromolecules, II ( 1978) 852. [ 161 E.A. DiMarzio, CM. Guttman and J.D. Hoffman, Faraday Discuss., 68 (1976). [ 171 W. Brostow and R.D. Comeliussen, Failure of Plastics, Hanser, New York, 1986. [ 181 Y.S. Don, M.J. Shim and S.W. Kim, J. Korean Ind. Eng. Chem., 6 (1995) 306. [ 191 I.S. Chun, M.J. Shim and S.W. Kim, Proc. ZUMRS-ICA’95, Seozd, 1995, p. 515.
Materials
Chemistry
and Physics 48 ( 1997) 90-93
Characteristics of polymer welding by healing process Mi-Ja Shim a, Sang-Wook Kim bl* a Department of Lge Science, Seoul City University, 90 Jeonnong-Dong, Dongdaemctn-Gu, Seoul 130-743, South Korea b Department of Chemical Engineering, Seoul City University, 90 Jeonuong-Dong, Dongdaemun-Gu, Seoul 130-743, South Korea Received
8 January
1996; revised 3 May
1996
Abstract A self-bondingstrengthdevelopment of ABS duringhot plateweldingwasinvestigated by alap-shear joint method.In thehot platewelding, the flow of moltenABS occursandit is importantfactor of self-bondingstrengthdevelopment. The flow activationenergies,E, obtained from relationof viscosityandtemperature were4-8 kcal mol-’ at differentshearrates.The self-bondingstrengthincreasedwith increasing time andtemperatureandits developmentat temperature rangefrom 180°Cto 260°Cwasfoundto follow a t”4 law basedon the reptation modelappliedto interdiffusionat anamorphous polymer-polymerinterface.In fractureof lapjoint samples, shearfailureandtensilefailure wereobserved.In tensilefailure,the crack-likeflaw at the edgeof bondwasthoughtto act asa stressconcentrator.Crackspropagated from it. The numberanddepthof plasticallydeformedridgesat there-heatedsurfaceby shearwereincreased with weldingtimeandtemperature. Keywords:
Welding;
ABS: Lap-shear
joint method;
Self-bonding
strength
1. Introduction When two samplesof the sameamorphouspolymer are contacted at a temperatureabove the glasstransition temperaturewithout adhesives,the junction surfacegradually develops increasing mechanical strength across a polymerpolymer interface until the mechanicalstrength reachesthat of the bulk polymer [ I] . It is saidto behealed.The molecular model of healing in polymer was explained by the diffusion of polymer chains across the interface of the contacting polymers [ 21. The molecular interpenetration is related to the phenomenon of self-diffusion. To analyze the self-diffusion along the chain and to describethe features of chain motion composed of entangledrandom coil, reptation model [ 31, crack healing theory [43 and minor chain model {5] were alsoproposed. Thesehealing processesprovided a significancefor polymer processing[ 6,7] such assurfacejoining by welding or lamination and preparation of a bulk polymer from polymer pellets. These methods make use of the inherent properties of the materials. Recently, aircraft or automobile applications have called for the ability to satisfy the mechanicalstrength requirementsat a lower weight and thermosetsasmatrix are not able to be recycled and are brittle. So, several thermoplastics with good thermal and mechanicalpropertiescanbe * Corresponding
author.
0254-0584/97/$17.00 PHSO254-0584(96)01858-5
0 1997 Elsevier
Science S.A. All rights reserved
expectedto replacethe thermosets.Thesematerialsare easy to be recycled, molded and able to be applied to high performancesby blending, compoundingand reinforcing [ 81, In this study we investigated the self-bonding strength developmentat varioustemperaturesduring hot platewelding of ABS (acrylonitrile-butadiene-styrene terpolymer) and useda lap-shearjoint method [ 91. The factors affecting the self-bonding strength such as melt flow, welding time and welding temperaturewere discussedandfracture behavior by shearstresswasalso observed.
2. Experimental 2.1. Materials
The ABS was a commercial pellet type ABS, HF-06601 from Cheil Industries Inc. with butadienecontent of 14.4%, acrylonitrile content of 21%. An average molecular weight of free SAN, K, is 95000. 2.2. Method
An appropriateamount of ABS in the shapeof pellet was melted at 16O”C,0 MPa, for 10 min in a preheatedhot press and was molded with increasing pressure to 5.5 MPa for 30 min. Compression-moldedABS was cut and sharpened
M.-J. Shim, S.-W. Kim/Materials
Fig. 1. Form
and dimension
of single lap-shear
Chemistry
and Physics 48 (1997)
90-93
91
joint sample
after slow cooling to a room temperature. The dimension of sample was based on the ASTM D3 163 as shown in Fig. 1. A hot plate was preheated to a temperature from Tg + 50°C 150 to 300°C. Specimens were placed on the hot plate for 10, 20, 30,40,50 and 60 min. Healing was conducted at atmosphere pressure. Two specimens were overlapped and provided about 0.1 MPa of contact pressure for instantaneous wetting. Self-bonding strengths of lap-shear specimens were estimated on the Shimadzu Autograph (Model AGS-1000D) using the stoke-tension mode at a crosshead speed of 0.5 mm min-‘.
3. Results and discussion In a hot plate weiding, the flow of molten ABS occurred. In order for a polymer melt to flow, the chain segments must be able to move. It was required that the chain segments had to have sufficient thermal energy to overcome energy barriers. At a small amount of flow, the small interpretation of the flowing material resulted in a lower bonding strength. At a large amount of flow, severe deformation of shape occurred during welding. Therefore, the flow pattern was an important factor for good welding and affected the features of the surface deformed by heat welding. Therefore, the activation energy to overcome energy barriers should be considered. The flow activation energy E, can be obtained from the general Arrhenius equation for molten polymers [ lo].
(1)
Fig. 2. Variation
of viscosity
with shear rate:
q , 220°C;
0, 250°C; 0 280°C.
Microscopically, the motion of a chain entangled with many other chains occurs rapidly by thermal fluctuations. In a reptation model for healing by de Gennes [ 31, each polymer chain is considered to have a tube-like region along the length in which it executes a random back-and-forth motion. Crack healing theory proposed by Wool and O’Connor [ 41 is composed with five stages of (A) rearrangement, (B) surface approach, (C) wetting, (D) diffusion and (E) randomization. The potential barrier related to inhomogeneities of the polymer interface disappears at the end of wetting and the polymer chains are free from the moving across the interface by diffusion and randomization stage which is very important on account of the developed strength. In symmetric interfaces, the rate determining steps were wetting and diffusion. Therefore, it can be considered that the self-bonding strength development of ABS was completed by a two-stage process. The stress components acting on the faces of the element are normal stresses, flrr, crYYand LT= and shear stresses r,,,, rrZ and r,,. These quantities are necessary to specify the state of stress at a point of sample. The stresses induced in a lapshear sample loaded in tension are complicated. Assuming forces are transmitted through the y-direction and no relative displacement occurs in either the x or the z direction, only shear stresses in the yz plane (welded area, A) are effective
where C is a constant and 77is the viscosity of a polymer. To calculate the activation energy, Eq. ( 1) is rearranged.
142kec-‘) 271bec-‘)
In 7)=+T+ln
C
720(sec-‘) 77msec-‘) 1600(sec-‘) 1600~sec-‘)
Viscosity as a function of shear rate is obtained from the capilfary rheometer [ 111 and it is shown in Fig. 2. Viscosity decreases with the increasing shear rates and temperatures. A relationship between In 71 and l/T is plotted and it is linear as shown in Fig. 3. From the slope, the activation energy can be obtained. The values of the activation energy, E,, are 4-8 kcal mol- ‘, Over these values, the ABS melt flows and the polymer chains interpenetrate into each other, so welding has happened.
f /
/
/ Ew30hec“)
/ 5.0
’ 1.5
1.7
1.9
l/T
2.1
x
2.3
I 2.5
IO’(K)
Fig. 3. A plot of In r) vs. 1 iT at various
shear rates.
M.-J. Shim, S.-W. Kim/Materials
Chemistry
and Physics 48 (1997)
90-93
Fig. 5 shows a load-displacement curve in fracture of lapshear joint specimens. In fracture of lap joint samples, shear failure and tensile failure were observed. In tensile failure [ 16-181, a distinct flaw appeared at the edge of the bond in which samples slowly bent and cut off after the stress reached a maximum value. These crack-like flaws acted as a stress concentrator and the crack propagated from them.
Fig. 4. Self-bonding strength several constant temperatures: W, 260°C.
0
y 0
Fig. 5. Load-displacement ABS.
vs. welding 0, 18OT;
1
time to the one-fourth Cl, 2OOT; A, 220°C;
1 2
l
power at ,240”C;
I 3
Displacement(mm) curve in fracture of lap shear joint
samples
of
on the recorded failure load. Hence, the shear stress, rYzacting on the yz plane is given by [ 12,131
‘=A
F (failure load) (welded area)
7rz is a self-bonding expressed as 7= 5-o+ 3-6
(3) strength to be determined
and is (4)
where rO is a self-bonding strength due to the wetting and TV is due to the diffusion of chain penetrating. However, the value of meis so low it can be ignored; therefore r. is removed from Eq. (4). The use of the reptation model yields xa t1’4 and ra x, where x is the interpenetration length at polymerpolymer interface. Eq. (4) is rearranged as { 14,151 7= &I4
(5)
where K is a constant and t is the welding time. Eq. (5) represents the time dependence of self-bonding strength. Fig. 4 shows a self-bonding strength vs. welding time to the one-fourth power at various temperatures. The self-bonding strength development was found to follow a t”4 law based on the reptation model.
Fig. 6. SEM photograph of plastically deformed 30 min; (b) 26O”C, 30 min; (c) 26O”C, 1 h.
welded
area: (a)
18O”C,
M.-J.
Shim, S.-W. Kim/Materials
Chemistry
and Physics 48 (1997)
90-93
93
3. In fracture, shear failure and tensile failure were observed. The latter was caused by a stress concentration effect at the edge of bond. A crack-like flaw at the edge of the bond acted as a stress concentrator. The number and depth of plastically deformed ridges, which were the main features of the welded samples, depended on the welding time and temperature. The fracture surface exhibited stress whitening.
Acknowledgements The authors acknowledge Kyong group in Korea. Fig. 7. SEM photograph
of stress whitening
at deformed
financial
support
from Sun
surface.
References When the lap-shear joint samples were fractured by shear, plastically deformed ridges which were the main features of the welded samples were observed as shown in Fig. 6. The number and depth of ridges increased with welding time and temperature, i.e. the self-bonding strength is related to the number and depth of plastically deformed ridges and flow patterns of molten polymers [ 191. A stress whitening in Fig. 7 was also observed in a deformed surface at the edge of bond.
4. Conclusions 1. In hot plate welding, the flow of molten ABS occurred and it was an important factor in self-bonding strength development. Viscosity decreased with increasing shear rate and temperature. Activation energies obtained from the relation of viscosity and temperature were 4-8 kcal mol-’ at various shear rates. 2. The self-bonding strength development was found to follow a t”4 law based on the reptation model applied to interdiffusion at an amorphous polymer-polymer interface.
[l] S. Prager and M. Tirrell; J. Chem. Phys., 75 (1981) 5194. [2] S.S. Voyutskii, Autoadhesion and Adhesion of High Polymers, WileyInterscience, New York, 1963. [3] P.G. deGennes, J. Chem. Phys., 55 (1971) 572. [4] R.P. Wool and K.M. O’Connor, J. Appl. Phys., 52 (1981) 5953. [5] Y.H. Kim and R.P. Wool, Macromolecules, I6 (1983) 1115. [6] R.J. Crawford and Y. Tam, J. Muter. Sci., 16 (1981) 3275. [7] K. Jud, H.H. Kauschand J.G. Williams,J. Muter. Sci., I6 (1981) 204. [S] IS. Chun, M.J. Shim and S.W. Kim, Proc. IUMRNCA’95, Seoul, 1995, p. 521. [9] D.B. Kline and R.P. Wool, Polym. Eng. Sci., 28 (1988) 52. [ 101 L.E. Nielsen, Polymer Rheology, Marcel Dekker, New York, 1977. [ 1 l] E.B. Bagley, J. Appl. Phys., 28 (1957) 624. [ 121 J.L. Lubkin and E. Reissner, J. Appl. Mech., Trans. ASME, 24 (1957) 255. [ 131 I.N. Sneddon, in D.D. Eley (ed.), Adhesion, Oxford University Press, London, 1961, [ 141 S.F. Edwards, Proc. Phys. SOL, 92 (1967) 9. [ 151 J. Klein, Macromolecules, II ( 1978) 852. [ 161 E.A. DiMarzio, CM. Guttman and J.D. Hoffman, Faraday Discuss., 68 (1976). [ 171 W. Brostow and R.D. Comeliussen, Failure of Plastics, Hanser, New York, 1986. [ 181 Y.S. Don, M.J. Shim and S.W. Kim, J. Korean Ind. Eng. Chem., 6 (1995) 306. [ 191 I.S. Chun, M.J. Shim and S.W. Kim, Proc. ZUMRS-ICA’95, Seozd, 1995, p. 515.