Fibre-epoxy composites at low temperatures

Fibre-epoxy composites at low temperatures

The thermal and mechanical properties of carbon, glass and Kevlar fibre reinforced epoxy composites are discussed, with particular reference to the be...

750KB Sizes 14 Downloads 66 Views

The thermal and mechanical properties of carbon, glass and Kevlar fibre reinforced epoxy composites are discussed, with particular reference to the behaviour of these materials at cryogenic temperatures. The effects of production techniques and various fibre arrangements are determined.

Fibre-epoxy composites at low temperatures* G. Hartwig and S. Knaak Keywords: low temperature techniques, composite materials, fibre-epoxy composites



Ultimate compressive stress

cr/crUT Fatigue endurance limit (tensile threshold) E

Young's modulus


Interlaminar shear strength


Fibre volume fraction


Fibre angle


Integral thermal expansion


Room temperature



Relative thickness of angle-ply



Coefficient of thermal expansion



e eUT

Strain Ultimate tensile strain



X /x

Thermal conductivity Poisson's ratio







Ultimate tensile stress



Fibre composites are attractive alternatives to metals because of their high specific strength or stiffness or their excellent fatigue behaviour. They are a necessary supplement to metals because of their low electrical and thermal conductivities, the latter being related to strength or stiffness. Their disadvantage arises from the weak polymeric matrix and results in low interlaminar shear strength and low transverse strength. At low temperatures some properties of the heterogeneous fibre-matrix system are superimposed by such peculiarities as additional thermal resistance at boundaries (Kapitza effect) by which the thermal conductivity is reduced. In the course of cooling the different thermal contractions of fibre and matrix give rise to thermal residual stresses and strains which influence most of the mechanical properties. At low temperatures the majority of currently used matrices are brittle and do not allow relaxation of residual stresses or stress concentrations to take place. In future developments the application of special, low tempera-

ture ductile thermoplastics will improve the mechanical properties. Cryogenic applications of polymeric fibre composites are mainly in superconductivity, space technology and handling of liquefied gases. They include superconducting generators or pulsed magnets for fusion reactors which call for materials resistant to fatigue and electrically resistive to eddy currents. Support elements for liquefied gas containers have to be optimized with regard to low thermal conductivity and mechanical strength. Transport vessels and components used in space technology should be light weight and resistant to fatigue. In several applications low thermal contraction is necessary. The cryogenic, mechanical and thermal properties of important polymeric fibre composites will be presented. The dependence of mechanical properties on production techniques was studied for carbon fibre composites.

*Dedicated to Professor Dr. W. Heinz on the occasion of his 60th birthday,

The fibres used in this work were E-glass fibres, high modulus carbon fibres (M40, Toray), high tensile

0011-2275/84/011639-09 CRYOGENICS. NOVEMBER 1984


$03.00 © 1984 Butterworth ~t Co (Publishers) Ltd. 639

Table 1.

Fibre properties


T 300 Carbon fibre M 4 0 A A S-4 Fibre glass (E-glass) Kevlar 49

oUT II ,






3.5 2.4 2.8 3.0-3.4 2.8

1.5 0.6

230 400 210 70 140

3.1 2.1



EF ± *,

GPa ~ 24 70 ~ 11

* - a t T = 4.2 K carbon fibres (T300. Toray and AS4, Hercules) and Kevlar-49 fibres (Du Pont). The principal fibre properties are shown in Table 1. Material samples were produced by filament winding, for tubes, as well as wet lay-up and prepreg techniques. Three epoxy resin systems were used to form the polymeric matrices. All three are based on bi,phenol-A and manufactured by Ciba Geigy. The system Cy 221/Hy 979 is a flexible laminating resin: the two others, Ly 556/Ht 972 and My 740/Hy 917, are rigid at room temperature and are used in prepreg techniques. Epoxy resins are cross-linked polymers and show brittle behaviour at low temperatures. The tensile fracture strain for flexibilized laminating resins is of the order of BUTM ~ 2% at 4.2 K. The systems used in prepreg techniques are much more brittle and exhibit an even lower fracture strain. Some types of thermoplastics (eg PC, PSU) are ductile in their behaviour even at low temperatures and show a

Table 2.

fracture strain of BUTM ~ 4-6% at 4.2 IC They are candidates for matrices of future composites for cryogenic applications? For all matrices not only their cryogenic properties are decisive but at RT additional requirements must be fulfilled. This applies mainly to the creep characteristic of composites which should be sufficiently low. However, in most cryogenic magnet applications high loads are applied only at low temperatures, where creep processes are negligible. At RT they have to withstand the intrinsic weight load only (eg structural parts for superconducting magnets and generators). For one- and two-dimensional load conditions the following fibre arrangements were chosen: unidirectional (UD) layers; unidirectional (UDW) made from nearly unidirectional fabrics with a fill of thin glass threads for fixation; angle-ply (0 °, ___ 45 °, 90 °) made from UD plies or 0 ° and 90 ° fabrics and crossply (0 °, 90 °) made from weaves.

Anisotropyof composite properties Most properties of fibre composites are of a tensorial nature. The anisotropy is due to the fibre arrangement, the intrinsic fibre anisotropy (carbon and Kevlar fibres) and the fibre-matrix interfacial bond.

Fibre anisotropy Fibre glass consists of an isotropic material with strong covalent bonding between atoms which results in a high strength in three dimensions. Carbon fibres are anisotropic and have strong covalent bonding only

Properties of matrix and fibre composites OUT, G Pa

•UT, %

E, G Pa

Cuc, G Pa


t'lLS, M Pa

dr/O'uT'b %







4 0 at 7 7 K

140 240 130

0.8-1.3 -

0.32 0.32 -

110 90 140

85 85 -



170 a




38 a









M A T R I X A T 4.2 K Epoxy resin Cy 2 2 1 / H y


FIBRE-EPOXY COMPOSITES (FIBRE CONTENT 6 0 VOL%) A T 4 . 2 K U n i d i r e c t i o n a l II W i t h carbon

2.0 1.5 -

1.5 0.5 -

W i t h fibre glass (E-glass)




W i t h Kevlar 4 9









0.5 < 0.2

1.1 < 0.3

53 65


O.31 0.3


65 65









T 300 M 40A A S4


Unidirectional 1 W i t h carbon


T 300

W i t h Kevlar 4 9


A n g l e - p l y (0 =, 4- 4 5 °, 9 0 °) W i t h carbon

fibre W i t h Kevlar 4 9

T 300 M 40A

a - at 77 K; b - fatigue endurance limit (N > 107 ) at 77 K





KEVLAR- FIBRES -t ..... t N-H O=C




Oo ..0







strong: weak :





l-dimensional I (v°n der Weals) fransversol[y medium weak strong





strong I dlmenslono[ weak : 2 - d i m e n s i o n a l (van der Weals,H-bonds) frcmsversolly weak


Isotropic and anisotropic bonding structures of fibres

within small g r a p h i t i z e d areas which are m o r e or less c y l i n d r i c a l l y a l i g n e d in the fibre direction ( P A N b a s e d fibres). These areas are weakly b o n d e d by Van d e r W a a l s forces. Therefore, their transverse strength a n d stiffness are rather low. T h e transverse stiffness Ev_L is h i g h e r by o n l y a factor o f two or three t h a n that o f an epoxy resin at 4.2 K (see T a b l e s 1 a n d 2). The a n i s o t r o p y a p p l i e s also to the t h e r m a l properties. T h e t h e r m a l conductivity at RT is m u c h h i g h e r in the fibre direction t h a n p e r p e n d i c u l a r to it. At sufficiently low t e m p e r a t u r e s the intrinsic conductivity t e n s o r o f fibres b e c o m e s m o r e isotropic b e c a u s e only long-wave p h o n o n s are activated which p r o p a g a t e in a s i m i l a r m o d e in the transverse a n d l o n g i t u d i n a l directions o f g r a p h i t i z e d areas. Kevlar fibres are even m o r e anisotropic. T h e y consist o f stretched a r a m i d e m o l e c u l e s with strong covalent b o n d i n g in the fibre directions only. T h e m o l e c u l a r c h a i n s are c o n n e c t e d in two d i m e n s i o n s by weak Van d e r W a a l s or h y d r o g e n bonds. In addition, the Kevlar fibres have a microstructure of fibrillar b u n d l e s with l o n g i t u d i n a l voids between them} Accordingly, the transverse a n d s h e a r strengths a n d stiffnesses are very low. As seen from Fig. 2 the m e a n transverse stiffness o f Kevlar fibres is s i m i l a r to that o f an isotropic p o l y m e r at low temperatures. The anisotropy holds also for the t h e r m a l conductivity t e n s o r at RT. As for c a r b o n fibres the conductivity tensor at low t e m p e r a t u r e s b e c o m e s m o r e isotropic due to long wave phonons. The a n i s o t r o p y of t h e r m a l e x p a n s i o n is strongly m a r k e d for Kevlar fibres. A small negative coefficient


o f t h e r m a l e x p a n s i o n exists in the fibre direction a n d a rather large positive one p e r p e n d i c u l a r to it. This is due to the v i b r a t i o n a l m o d e s a n d the b o n d i n g potentials involved. T h e structures a n d b o n d i n g forces o f the fibres c o n s i d e r e d are shown in Fig. 1. T h e transverse a n d l o n g i t u d i n a l m e c h a n i c a l properties are s u m m a r i z e d in T a b l e 1. T h e m e a n transverse stiffness E v i of c a r b o n

Io ~ ~


E~, ~

• Kevlar-fibre

~UO 8 ~


[L v ~ ' " ~ , ~





" ~




Y_ cD ~





I 0

I 50


I I00


[ ~50


I 200


I z50

I 300

Temperature,K Fig. 2 Transverse stiffness of UD Kevlar fibre composites and stiffness

of the epoxy resin versus temperature


and Kevlar fibres was estimated from data on

Mechanical properties and thermal residual strain

transverse fibre composites and on the epoxy resin applied. Equation (7) was modified for hexagonal cross-sectional fibre arrangement; the transverse modulus is a mean value composed of radial and azimuthal fibre moduli,

Fibre-matrix interracial bond

The bond arises from covalent chemical bonds, adhesion by weak Van der Waals bonds and interfacial friction. The types and reactions responsible for polymer-fibre glass bonds are more or less known and applied in commercial composites, For carbon fibres oxidative surface treatment has been successful but further investigations will be necessary. Possible types of covalent bonds between hydroxyl groups of oxidized carbon fibre surfaces and epoxy resins are shown in Fig. 3.3 The fibre-matrix friction and adhesion depends on compression by shrinkage. For composites cooled down to cryogenic temperatures the matrix and radial fibre contraction are decisive, Fibre glass contracts less than polymers thus increasing the frictional component of the interfacial bond. For Kevlar fibres just the reverse situation holds, The radial or transverse contraction of Van der Waals or hydrogen bonded aramide molecules is stronger than in most isotropic polymers thus reducing adhesion and interfacial friction. For carbon fibreepoxy composites no great temperature dependence is to be expected of interfacial friction. With the present status of fibre surface treatment and coupling agents the highest bond strength is achieved for fibre glass-epoxy composites, followed by carbon fibre-epoxy composites. For Kevlar fibre compositeg only a rather poor bond strength was achieved which reduces even more the transverse and shear strengths. Important aspects of the interfacial bond are reflected in the interlaminar shear strength which is described below.

Cryogenic properties of composites The anisotropic behaviour of fibre composites arises from fibre and matrix dominated thermal and mechanical components. However, several fibre dominated properties can be determined indirectly by the matrix behaviour. At cryogenic temperatures this is especially true for mechanical properties which are influenced by matrix failures or by propagating microcracks. Their initiation is strongly enhanced by thermal residual stresses or strains, Bond

-~/-/0~ " ] ;-C2H 4 - N H

z +HtC-/CH



Finish eg ominosilone epoxy resin

0 ~

QIoss y/J-O-S;-C~H4-NH -CH 2 -CH -R ~

OH ~ \0/ ~ ~//~/C- 0 - CH z - CH - R ~

oxidotion epoxy resin


Fig. 3 Examplesof chemicalbonds between fibre and matrix 642


{efM} = eUTM -- {~RM}


The brackets comprise longitudinal (z), azimuthal (#) and radial (r) components. Both the latter components are difficult to measure separately. Therefore, approximately only longitudinal (z or II ) and transverse (J_) components, with respect to fibre direction, are considered. For composites with anisotropic carbon and Kevlar fibres a reasonable approximation is to assume that the transverse fibre contraction is equal to that of the matrix. Thus, only the longitudinal component remains. For fibres with a negligible longitudinal contraction (AL/LF~ 0) the thermal residual matrix strain is given by er_M~- AL/LM within the temperature z range considered. For composites with isotropic low contractive fibres, such as fibre glass, matrix shrinkage occurs and a multiaxial stress situation in the matrix must be considered (the equations are listed in reference 5). For lamination epoxy resins the ultimate tensile strain at cryogenic temperatures 4 is: 8UTM (4.2 K) = 2.2%. The integral thermal contraction is typically ,td-,/LM (293-4.2 K) = 1% to 1.2%. The resulting free strain of the matrix yields the following values: - - composites with 60 vol% carbon or Kevlar fibres: elM (4.2 K) = 0.7-0.9% - - composites with 60 vol% fibre glass: efM (4.2 K) = 0.5-0.8%. For a matrix with prepreg resins the values of the free strain are lower by at least a factor of two or three. As seen from Tables l and 2 these values are insufficient for most UD composites considered and matrix failure occurs prior to fibre breakage. The free strain might be little larger than estimated. The accumulation of strain was assumed to start at RT, although some viscoelastic relaxation may occur even down to secondary relaxations. When connecting UD layers as angle ply, additional residual and thermal residual stresses, especially shear stresses, are induced. By proper fibre arrangement the thermal residual stresses in the directions of external load can be minimized. This important optimization of fibre composites has been discussed in a previous paper? However, a rigorous solution of problems arising from residual and thermal residual strains can only be achieved by applying low temperature ductile polymers which exhibit much higher fracture strains (eg PC or PSU as mentioned


Corbon //,I/C-OH -I- H.C,-C H -R ~


The main reasons for matrix failures are brittleness and thermal residual stresses. With decreasing temperature the polymeric matrix becomes stiffer and stronger but also less ductile, and in combination with low contraction of the fibres more and more thermal residual tensile stress and strain {eRM} is accumulated on the matrix. This reduces the effective free stress or strain {EfM} available for external

Elastic moduli, ultimate

stress a n d s t r a i n

According to simplified assumptions the mechanical properties of composites can be modelled


as parallel or serial arrays of springs representing the matrix, fibre and bond. UDfibre composites (C) in thefibre direction ( l[ ). For a parallel array of the fibre (F) and matrix (M) it holds: strain Cell(T) ~ eM (T) -~ EF II


link between the ultimate strengths of ~rF, aM and gB. Experimental results show that (ruTc < (rut M. According to (6) the composite breaks at a strain lower by a factor of (l-F) than the fracture strain of the matrix if the transverse fibre strain is negligible. Then, when fracturing cross-plies the matrix in the transverse layers cracks earlier than the matrix and fibres in the longitudinal, say load-bearing layers.

Angle-ply composites (0 °, 90 °) or (0 °, +_ 45 °, 90°). stress acll(Y) ~ aM(T ) • ( l - F ) + OFII • F


Young's modulus EctI(T) -~ EM(T ) • ( l - F ) +EFII " F


F is the fibre content by volume; polymer and composite properties are indicated in Table 2. For a moderate filling the composite modulus E c II is dominated by the nearly temperature independent fibre modulus E F II The influence of the matrix is less than 10%. Elasticity data are very similar under tension and compression? However, the ultimate compressive strength and strain are always smaller than tensile properties because buckling occurs. 7 This effect is especially serious for transversally weak fibres, such as Kevlar. The ultimate tensile strength ¢rUTc and strain CUTC of U D composites are determined mainly by the strength and content of fibres, irrespective of matrix failure. This advantageous behaviour applies to static uniaxial load in the fibre direction. The situation in the more important case of angle-plies, where cracks from neighbouring plies influence the load bearing layers, is discussed below. However, for several applications such as vacuum tight vessels in cryostats, matrix break is intolerable. •

Among various failure modes, such as delamination, one failure mode will be considered which is typical of brittle polymeric matrices. When (0 °, + 45 °, 90 °) composites are loaded, say in the 0 ° direction, crack initiation and propagation are favoured within 90 ° and - 45° layers. When cracks reach a load bearing layer (0 °) they cut fibres as soon as the stress concentration exceeds their strength. This is true for tranverse weak fibres such as carbon or Kevlar fibres. Fibre glass is rather tough and acts as a crack stopper. In Fig. 4a a fractograph of such a carbon fibre composite shows a plain and smooth fracture area produced at 90 ° and a stepwise fracture area at 45 ° in the load bearing 0 ° layer. By contrast the rough fracture area with fibre pullout of UD composites is shown in Fig. 4b. Although the matrix has negligible load bearing functions, it can reduce the strength of a carbon fibre composite by 30% and more? This fracture mode could be avoided by a matrix which is ductile at low temperatures.

Fatigue characteristics The highest fatigue endurance limit has been found for carbon fibre UD composites; in tensile

UD fibre composites (C) perpendicular to the fibre direction ( l ) . For a serial array of the matrix (M), fibre


(F) and bond (B) it holds:

stress Oc.L(T) ~-- aM(T) --~ OF± ~ o B


strain ecj.(T) ~ eM(T ) • ( l - F ) +eF.L • F


Young's modulus

1 EL(T)

,~ ( l - F ) E~(T)

F + ~ E~,j.(T)


For transversally stiff fibres (fibre glass. E v j_ > > EM(T) ) the temperature dependence is controlled by the matrix. For transversally weak fibres (carbon or Kevlar fibres: EF ± ( T ) ~ EM(T)) an additional temperature dependence arises from the fibres. The temperature dependence of the transverse stiffness E c ± is shown in Fig. 2 for U D Kevlar fibre composites. For comparison the matrix stiffness is plotted in addition. No great difference has been found to exist between both stiffnesses. The ultimate tensile strength ~rUTC and strain CUTC of the composite are determined by the weakest




!~ Fig. 4a - - Fractograph of (0"; :t: 45*; 90") angle-ply composite with M 40A carbon fibres at 77 K. b - - Fractograph of UD carbon fibre composites at 77 K (M 40A-carbon fibres)


Strength : %T and Stiffness: Ell 2.1

The value of the interlaminar shear strength rILS ~ 170 MPa at 77 K supports the assumption that the matrix determines ~'ILSFor UD carbon fibre composites made from untreated fibres one gets a value of the order of VlLS ~ 40 MPa at 77 K. Optimized oxidative treatment and adapted coupling agents improve the ILS up to values of VIES(77 K) ~ 110-140 MPa with H T fibres IT 300 and AS4) and vILS(77 K) ~ 60-80 MPa with HM fibres (M40A)? This behaviour supports the assumption that the bond strength is the limiting factor. For HM fibres the low transverse fibre strength might exert an

Fibre composites

6o vol*/, 2


- -

77 K


i.4 j ~1

~1 b



'~ +-45~

t~ I


t~] ~ 9 0 " )




0 I00


Fatigue endurance limit ( N > I O 7) (tensile threshold)



85%~ lill IIII s~*/, ~

5o _

zs*/, M

65*/, "~ 50"1, NI ~ 4o*/,T | ,.

o E-Gloss


additional influence. For UD Kevlar fibre composites the low value of VlLS(77 K) ~ 38 MPa indicates failure by debonding and probably also by the fibres.






Fig. 5 Tensile strength GUT, stiffness E II and fatigue endurance limit a / % T of UD and (0"; :t: 45*, 90*) fibre composites at 77 K

Thermalconductivity A survey of thermal conductivity of UD fibre composites is given in Fig. 6. Fibre-glass and Kevlar fibre composites do not show a great temperature dependence of thermal conductivities) H3 their values being five to ten times higher than that of the epoxy matrix. By contrast, carbon fibre composites show a very large temperature dependence due to freezing out of electron thermal fibre conductivity at low temperatures. In Fig. 7 the thermal conductivity of carbon fibre composites is shown with different types and arrangements of fibres. At RT all UD composites with HM fibres (M 40A) show a higher conductivity

threshold tests at 77 K the value is 85%. The values of other UD and angle ply composites are shown in Fig, 5. Tests under compressive fatigue loading yield lower values because of microbuckling, especially for fibres which are weak in the transverse direction. The data are very scarce. Some information on fibre glass composites has been published? One important question is the degradation of properties in the course of fatigue cycling. For fibre glassandcarbonfibrecross-pliesitwasshownthat stiffness does not degrade by more than 15% under uniaxial tensile fatigue conditons at 77 K until fracture occurs. Under biaxial fatigue cycling with combined torsion and tension a much higher and cumulative degradation has been found. 1°

(so vol %)

For shear loading of composites the decisive limiting parameter is the ILS. If this value is exceeded, delamination of composite layers occurs. Fibre breakage occurs at loads defined by bending strength. The ILS measurements are performed in short beam tests. The values of ILS are determined mainly by failures of the matrix, interfacial bond or fibres. The question of the upper limit of ILS arises and another question is which components constitute the limitations of the composites considered. The low temperature shear strength of epoxy resins (and most polymers) lies between 150 and 200 MPa at T ~ 77 K and constitutes the upper limit of composites containing epoxy resins. For fibre glass composites the high, nearly isotropic fibre strength and very likely the bond strengths do not constitute the limiting factors,



/ - ~ /

Io 1~






poxy . . . ~ . - -

x,, __

'l[]/_....-~-I ......_~_ .......-~





~ io-I I



__ -- --

T 300

Kevlor/epoxy Gloss/epoxy Epoxy






Fig. 6










zoo Temperoture,K

I ~ ~ i i

i ~ i



Survey of thermal conductivities of UD fibre composites versus




1 984

Carbon fibres/epoxy 60 vol %

150 M40A II



xlO -4




///,"" / f ///~"

~ / f


T 3 0 0 II TSO0 0", +-45",90 ° T:500±45°


Kevlar fibre-composite Fibre content 60 vol%


~-E.~ox~yr -

oo i 50


- ~'~,~

~,,~ransver se


m ~0

~" "~ ~" "... "-. I










~///~'/// ~ / /

I0- I





Epoxy resin CY221 HY979


xtO"5 -I00


6 4 5

16 2

Fig. 7

I I I I II I II I 20 40 60 80 IO0









7", K Thermal conductivity of different carbon fibre composites versus


compared to those with HT fibres. The electric conductivity of HM fibres at RT is about 2.5 times higher than that of HT fibres. This is due to larger graphitized areas of HM fibres, At very low temperatures this difference vanishes because wavelengths which are characteristic of that temperature range (dominant phonon wavelength) are too large tbr resolving the microstructures of fibres. A similar consideration holds for the longitudinal and transverse components of the fibre conductivity tensor, At RT it is rather anisotropic and gets more isotropic at low temperatures due to the increase of the dominant phonon wavelengths. In addition the carbon fibre conductivity decreases with decreasing temperatures and gets comparable to or lower than that of the epoxy matrix. As seen from Fig. 7 the different types of carbon fibre composites considered vary by a factor of 50 at RT. At 7 K this factor reduces to 1.5. For the thermal conductivity transverse to the fibre direction a thermal boundary resistance by phonon mismatch (Kapitza resistance) may be relevant at very low temperatures. This means that at low temperatures the thermal conductivity of carbon fibre composites is rather small and nearly independent of fibre type and arrangement. Thus thermal conductivity imposes no major restrictions if one wants to optimize other composite properties at low temperatures. Thermal


For several applications in cryogenic temperature technology low contractive materials are of advantage, Due to the anisotropic tensor of thermal expansion this is not possible for all fibre composites in all directions. Fibre glass has a low coefficient of thermal expansion (atr ~ 4.8 10-n K-1) which is nearly isotropic in all directions. Thus, the contraction of those composites simultaneously can be made small in all three directions, Carbon and Kevlar fibres exhibit a very small or even negative coefficient of thermal expansion only in





I ~ I00




I I 200




I 300

Tempereture, K Fig. 8 Longitudinal and transverse integral thermal expansion for UD Kevlar fibre composites versus temperature

the fibre directions due to stretching vibrations in covalent bonding potentials. The negative expansion of Kevlar fibres is assumed to arise from bending vibrations perpendicular to the aligned molecular chains of fibres. In contrast their transverse expansion is large and larger than that of most epoxy resins. This is, due to vibrations in the weak asymmetric Van der Waals or hydrogen bond potentials. Thus, a low thermal expansion is only available in the fibre direction. This is demonstrated in Fig. 8 for the thermal expansion in the longitudinal and transverse directions of Kevlar fibre UD composites./3 For carbon fibre composites similar results apply. In angle-plies a medium low thermal contraction can be achieved in plane by means of residual stresses between plies. However a large thermal contraction exists in the direction perpendicular to the laminates. The coefficient of thermal expansion of UD composites atc II can be calculated 5 from those of the fibres o~F II and the matrix o~M by C~M --~rll Sell = erll +

(8) 1 + 1.1 F/(1-1.1/7) (Evlp/EM)

where F is the fibre content per volume: E M and EF II are the moduli of matrix and fibre, respectively. For calculating thermal expansion of angle-plies see a previous publication? The thermal expansion can be varied by the fibre content and arrangement or by different thicknesses of angle-plies. Experimental values of integral thermal expansion 4.2

AL/L =


e c dT

293 on angle-ply fibre glass composites are shown in Fig. 9. The parameter of the curves is the fibre angle __. to related to the direction of measurement) 4


Fibre gloss/epoxy 70 -',c5_ vol%

Fibre gloss/epoxy 70 -+ 3 vol %

I~'~x = 90 °

Directionof measurement




Direction of +~_~_+~ measurement


= 60 =

300 (

Thicknesses of plies ,


Is=tso./t 1












oJ= 450 ~

0)=30 °`


I00 --

o I

t I00

t 200


Temperature, K

Fig. 9 Integral thermal expansion of balanced angle-ply fibre glass composites versus temperature. The parameter is the fibre angles ± to I

In Fig. 10 the values of composites with three fibre directions (___to; 90 °) are shown. Additional parameters are the relative thicknesses of angle-plies. Is Both figures are examples demonstrating the great v a r i a b i l i t y o f fibre c o m p o s i t e s . I n m a n y c a s e s t h e s a m e contraction behaviour can be achieved by different fibre arrangements. This opens up the possibility of optimizing additional requirements, eg thermal or mechanical properties.

Dependence on production technique and fibre


The data summarized in Table 2 are valid for a wet lay-up technique and for an exact fibre alignment of each ply. Composites made from fabrics yield values of fracture strength lower by more than 30%. This arises from the wave shaped misalignment and the transverse compression exerted by fill and warp. The latter is especially of influence for transverse brittle fibres such as carbon and Kevlar fibres, The interlaminar shear strength tiES is sensitive to the quality of lamination. Two examples are given for UD composites with carbon T 300 fibres tiES(77 K) ~ 110-140 MPa (wet lay-up technique) tiES(77 K ) ~

80-95 MPa (prepreg technique)

An appropriate compromise is a matter of material quality and economy,



I I00


I 200


IkL 300

Temperoture,K Fig. 10 Integral thermal expansion of balanced angle-ply fibre glass composites versus temperature. The parameters are the fibre angles ± to

and the relative thicknesses t"a of the 90* ply

Conclusions An important advantage of polymeric fibre composites is their great variability in matching mechanical and thermal properties. In cryogenic applications the brittleness of the epoxy resin matrix reduces the mechanical strength, especially under multiaxial load conditions. The study of degradations of composite properties under fatigue cycling is a subject of further investigations. For future developments a substitution of brittle epoxy resins by special low-temperature ductile thermoplastics would help to reduce thermal residual stresses and stress concentrations within fibre composites. This requirement becomes more stringent with the improvement of fibre properties. Carbon fibres with ultimate tensile strains of 1.8% will be available in the near future. An additional advantage is that thermoplastics can be welded by solvent agents or by heat. This improves the quality of vacuum tight connections of members made from thermoplastic fibre composites. After curing


epoxy resins can only be glued. This is m a i n l y a n a d h e s i o n by weak Van der Waals forces, Each type of fibre composite has its specific merits. A n appropriate c o m b i n a t i o n of different fibre types in hybride composites opens up a large field for o p t i m i z i n g composite properties.

Authors The authors are from the K e r n f o r s c h u n g s z e n t r u m Karlsruhe, Institut ftir Reaktorbauelemente, D7500, Karlsruhe, F R G . Paper received J u n e 1984. Most of the samples were prepared by Messerschmit-B61kow-Blohm, M u n i c h . T h e c o l l a b o r a t i o n of B. Vogeley a n d Dr. Weiss is greatly acknowledged.

References 1 2

Hartwig,G. Proc lnt Cryogenic Materials Conf, Eds K. Tachikawa and A. Clark. Butterworths (1982) 495 Preston, J. Aramid Fibres Encyl. Chem. Technol. 3, Wiley, New York, 3rd Edn. (1978)



Fitzer, E., Weiss, R. Processing and uses of carbon fibre reinforced plastics, VDI-Verlag, Diisseldorf(1981) 45 Hartwig, G., Jaeger, H., Knaak, S. Nonmetallic Materials and Composites at Low Temperature II1; Eds. G. Hartwig, D. Evans: Plenum Press New York (1985)(in print) 4 Hartwig,G. Nonmetallic Materials and Composites at Low Temperatures I: Eds. A, Clark, ILP. Reed; G. Hartwig; Plenum Press, New York, (1978) 33-50 5 Schneider,W. Kunststoffe 61 (1971) 273 6 l~sen, M.B. Adv Cryo Eng (Materials) 28 (1982) 165-177 7 Chamis,C.C. Composite Materials 5 Ed. L.J. Broutmann, Academic Press, New York (1974) 121 8 Hartwig,G. Adv Cryo Eng (Materials) 28 (1982) 179-189 9 Nishijima,S., Keta, S., Ol~da, T. Nonmetallic Materials and Composites at Low Temperatures; Eds: G. Hartwig, D. Evans; Plenum Press New York, (1982) 139-149 10 Wang,S.S., Chim, E.S,M. Adv C~'o Eng (Materiala9 28 (1982) 201-210 11 Rosenberg,H. Nonmetallic Materials and Composites at Low Temperatures I1: Eds: G. Hartwig~ D. Evans: Plenum Press New York, (1982) 181-196 12 Hust, H.G. Cryogenics 15 (1975) 126 and Kasen, M.B. Cryogenics 21 (1981) 323 13 Hartwig,G. Tieftemperaturtechnologie, Ed. F.X. Eder, VDI-Verlag, Diisseldorf, (1981) 177 14 Hartwig,G., Puck, A., Weiss, W. Kunststoffe 64 (1974) 32 15 Bauder,P., Hartwig, G,, Sehindler, R. J Mat Sci 22 (1976) 225