7 июля, 2015 
Malyar Rate of strength development. The effect of temperature on curing rate will vary for different adhesives. In general, low temperatures increase the curing period considerably and many epoxy resin formulations stop curing altogether below 5 °С. A rule of thumb often quoted is that the curing period doubles for every 10 °С fall in temperature below ambient but halves for every 10 °С rise in temperature above ambient.
Fig. 2.11 illustrates how the manufacturer can influence the rate of strength development of a two-part epoxy by appropriate formulation(19). Compare these curves with the more rapid rates of gain of strength attainable with a structural polyester mortar and a general purpose cyanoacrylate ‘super-glue’. For single-part products the rate of curing is controlled closely by the temperature applied. Fig. 2.12 shows a typical cure time/temperature relationship for a toughened epoxy(20).
In general the properties of the adhesive in the hardened state are determined by its internal structure. Strength and elasticity derive
|
Slrengih (N/mm ) Slrength (N/mm )
1———— Г I————————— ————— I—————————————————— 1" 125 10 15 ——————- Days |
ЗО C 20 C 10 C
(a) (fc)
Fig. 2.11. Effect of formulation and cure temperature on flexural strength development of a two-part epoxy (Ref. 19). (a) Normal type, (b) Rapid type.
|
Cure temperature (°С)
Fig. 2.12. Cure temperature/time relationship for a one-part epoxy (Ref. 20). |
from molecular interactions, and a change in formulation to improve one property is likely to be at the expense of another. Further, the value measured is likely to be sensitive to the method of test, for example due to rate of loading or whether the adhesive is tested in bulk or thin-film form. Some important mechanical and physical properties of hardened adhesives are discussed below.
Strength and stress/strain characteristics. For satisfactory bonded joint design the important mechanical properties of the hardened adhesive under short-term loading are tensile and shear strengths, modulus of elasticity, elongation or strain capacity at failure, and fracture toughness (see Chapter 4). One initial experimental problem is the ability to fabricate reliable flaw-free bulk adhesive specimens with viscous cold-curing compounds. Work at Oxford Polytechnic(22) has demonstrated that the centrifuging of mixed adhesive can largely overcome such problems.
For the measurement of tensile properties, dumb-bell specimens of the form shown in Fig. 2.13 are suggested. Tensile modulus, Poisson’s ratio and elongation at failure may be measured with appropriate strain monitoring equipment and a set of stress/strain curves for a typical range of epoxies is given in Fig. 2.14. Similar
|
Specimen thickness = 3 mm Fig. 2.13. Tensile dumb-bell specimen (Ref. 21). |
specimens subjected to wet or moist environments may provide useful information on changes in strength or ductility due to water — induced plasticisation.
One way of monitoring the behaviour of a material in shear is with a torsional test. The method has the advantage that, provided a circular specimen is used, a condition of pure shear can be achieved. However, torsion tests can be relatively difficult to perform unless specimens can be machined accurately to avoid warping effects. The alternative is to test a prismatic specimen in a shear box of the form outlined in Fig. 2.15.
To assess flexural modulus of the hardened adhesive a specimen 200 x 25 x 12 mm deep tested in four point bending may be used. The sample under test is loaded transversely at the third points at a crosshead speed of 1 mm/minute and the central deflection recorded (Fig. 2.16). From the load-deflection curve, the secant modulus at 0.2% strain may be calculated. A lower limit on flexural modulus may be specified to prevent problems due to creep of the adhesive under sustained loads, whereas the upper limit will be to reduce stress concentrations arising from strain incompatibilities, for example at changes in section.
An alternative method of bulk property determination has been devised at Oxford Polytechnic(22). This involves monitoring the compression of pencil-like specimens from which a large amount of data may be derived (Young’s modulus, Bulk modulus, Poisson’s ratio and Glass Transition Temperature (Tg)).
|
Fig. 2.14. Typical tensile stress/strain curves for a range of epoxy adhesives. |
|
Fig. 2.15. Shear box test on adhesive prism (Ref. 21). |
Table 2.5 summarises typical values of these bulk mechanical properties at around 20 °С for a range of epoxy adhesives. Of particular note is the low strength and stiffness of the epoxy polyamides and polysulphides as compared to those with aliphatic polyamine hardeners. For the design of bonded assemblies, joint tests are often used to determine the relevant mechanical properties. It must be remembered, however, that the results will be highly dependent on the specimen geometry and testing conditions and
|
Table 2.5. Bulk mechanical properties of epoxies at ambient temperature
|
this aspect will be further discussed in Chapter 4. Nevertheless, joint tests can provide valuable comparative data. The stress-strain curves of Fig. 2.17 illustrate the variability in performance between different types of adhesives and between formulations of the same basic generic type as measured using tensile butt-joint tests(23).
Fracture toughness. The traditional approach to design using structural materials has been to compare the average stress or strain acting on the net cross-sectional area of the element with some ultimate stress or strain criteria. For the design of assemblies using brittle materials, which are sensitive to the presence of flaws, the theories of Fracture Mechanics have been introduced to overcome the shortcomings of the traditional approach. These recognise that the stress field around a crack or flaw can be defined uniquely by a parameter termed the stress intensity factor (K) which is directly proportional to the applied load. When К reaches some critical
|
Fig. 2.17. Typical stress-strain curves for adhesives from tensile butt-joint tests (Ref. 23). |
value, Kc — the fracture toughness of the material, a previously stationary or slow moving crack will jump ahead. Since some adhesives tend to be brittle in nature, especially at lower temperatures, the fracture toughness can be an important material design parameter. Kc is not unique but is dependent on factors such as the temperature of testing and the applied strain rate.
To measure bulk mode I fracture toughness, KIC, (tensile load acting normally to the crack surface — Fig. 2.18(a)) single-edge notched (SEN) beam or tension specimens are recommended, as illustrated in Fig. 2.19. These specimens suffer two disadvantages, namely that failure is always catastrophic and the calculation of stress intensity factor requires that the crack length be known. Typical values of KiC are summarised in Table 2.6(a) for a range of epoxy adhesives. Note particularly the increase in KjC achieved by ‘toughening’. Typical values of plain strain fracture toughness for other materials are given in Table 2.6(b) for comparison.
When considering fracture of the bulk adhesive, Mode II (shear) or Mode III (mixed) (Fig. 2.18(b) & (c)), fracture need not be considered since Mode I is the lower energy and therefore critical fracture mode, although in joints mixed mode loading may give rise to a lower critical fracture toughness. It must also be remembered that the fracture toughness of a joint may be controlled by the adhesive/substrate interface rather than that of the bulk adhesive and by the bondline thickness (see Chapter 4). Thus, for the design of bonded connections, measurement of the adhesive joint fracture toughness may be more appropriate.
Temperature resistance. Most synthetic adhesives are based on polymeric materials and as such exhibit properties which are characteristic of polymers. This is particularly so when considering their response to temperature variation. At a certain temperature, known as the Glass Transition Temperature (Tg), polymers change
|
|
Fig. 2.19. Fracture toughness specimens, (a) Single-edge notched beam specimen. (b) compact tension specimen.
|
Table 2.6. Mode I fracture toughness Klc
|
from relatively hard, elastic, glass-like substances to relatively viscous rubbery materials. The transition temperature will vary from one polymer to another. In addition, the transition temperature is dependent on the rate of loading if the measurement process involves mechanical deformation. The relationship between temperature and time effects will be discussed later in this Chapter.
Classical methods for measuring Tg include thermal, electrical, optical and dynamic mechanical techniques. For the civil engineer, quasi-static mechanical methods utilising a flexural test on a hardened adhesive prism are more convenient for determining the Heat Distortion Temperature (HDT). However, the results of such methods are dependent on the specimen configuration and the rate of loading selected and as such are only accurate for comparative purposes. A typical ‘Heat Distortion Test’ might utilise the same specimen configuration as for assessing flexural modulus (Fig. 2.16). The sample is placed in a temperature-controlled cabinet and a constant load is applied to achieve a maximum fibre stress of 1.81 N/mm2 in accordance with BS 2782 (24). The central deflection at room temperature (say 20 °С) is then recorded. The HDT of the adhesive is taken as the temperature, measured on a thermocouple attached to the specimen, attained by the sample after undergoing a further 0.25 mm deflection while subject to a surface heating rate of 0.5 °C/minute. The HDTs of a range of cold cure epoxy adhesives measured in this manner are summarised in Table 2.7.
It will be noted that these HDT values lie in the range 34 to 48 °С which for many civil engineering applications may not be much in excess of anticipated maximum service temperatures. For example, on the soffits of concrete bridges temperature extremes in the UK may lie between -20 °С and +38 °С (25). In steel bridges maximum temperature extremes of 60 to 65 °С may occur locally and this is one reason why single part hot cure epoxy products which have higher Tg values of the order of 100 °С and more are preferred in such situations.
The influence of temperature on basic mechanical properties of the hardened adhesive such as bulk flexural modulus and shear strength is illustrated in Figs. 2.20 and 2.21, respectively (26). From these figures it is evident that the response of all five adhesives to temperature variations within the range 15 °C-65 °С is similar. The most noticeable feature of the curves is the rapid deterioration in both stiffness and strength at a temperature close to the measured HDT of the adhesive.
This general correlation between the HDT of the adhesives and the temperature range within which their basic engineering properties undergo significant change, confirms the proposition that the HDT is associated with a temperature at which important changes in molecular structure of the adhesive are occurring and which affect the way in which adhesives are capable of carrying load.
Table 2.7. HDT of cold cure epoxies
Adhesive HDT (°С)
2-part cold cure epoxy polyamide 40
2-part cold cure epoxy polyamine (aliphatic) 41
2-part cold cure epoxy polyamine (aliphatic adduct) 43
2-part cold cure epoxy polyamine (aromatic) 48
2-part cold cure epoxy polysulphide 34
|
|
||
|
|||
|
|||
|
|||
|
|||
|
|||
|
|||
|
|||
|
|||
|
|||
|
|||
 |
One final point worth noting is that, provided the temperature does not rise above 150-200 °С, when a chemical deterioration occurs, the reduction in mechanical properties above the HDT is reversible on cooling. Indeed, stiffness and strength may even be enhanced due to post-cure effects at the elevated temperature. On the debit side, Tg or HDT will be lowered by water absorption into the polymer. Indeed, the Tg should be one of the first criteria in assessing the likely suitability of candidate adhesives.
Moisture resistance. All adhesives are susceptible, to some degree, to the effects of exposure to water or water vapour. Indeed, when cured, the very epoxy groups which give epoxies their adhesive properties also render them hydrophilic(27). This water uptake is accommodated largely by swelling. The effect of exposure to moisture is to alter the adhesive’s properties, often in some undesirable way. Water may enter an adhesive either by diffusion or by capillary action through cracks and crazes. Once inside, the water may alter the properties of the adhesive either in a reversible manner, for example by plasticisation, or in an irreversible manner, for example by hydrolisation, cracking or crazing.
To determine water transport properties, thin film specimens may be immersed in water at a known temperature or stored in an atmosphere of known humidity and temperature. The water uptake of the adhesive is then measured and the fractional uptake plotted against the square root of time per unit thickness to comply with the mathematics of diffusion. Results for 60 x 12 x 2 mm specimens of five cold-cure epoxies immersed in water at 20 °С are shown in Fig. 2.22. The specimen size was selected to enable flexural modulus tests to be conducted. If the plot shows a linear increase followed by an equilibrium plateau then the uptake is termed ‘Fickian’. S — shaped plots are ‘Non-Fickian’ and are thought generally to be typical of glassy polymers. The relatively unmodified polyamide and the aliphatic polyamine absorb more than 5% by weight of water. However, in the latter case independent research (28) using electron microscopy revealed extensive micro-cracking in the resin and pronounced discontinuities at the surface of the relatively large (up to 0.4 mm) silica filler particles. Supplementary dye-penetration experiments with this adhesive confirmed that penetration was rapid in some areas. Conversely, the aromatic polyamine and the adducted, and therefore modified, aliphatic polyamine absorb less than 1% by weight. The polysulphide lies in the middle of the range. All five
|
A——— 2-part epoxy polyamine (aliphatic) В——— 2-part epoxy polyamide C——— 2-part epoxy polyamine (aromatic)
Fig. 2.22. Water uptake plots for a range of epoxy adhesives (Ref. 26). |
adhesives are weakened by water absorption, as measured by bulk shear strength (Fig. 2.23). With the exception of the aromatic poly amine, this is associated with plasticisation of the adhesives, as measured by the flexural modulus of the specimens (Fig. 2.24) (26). However, fracture toughness was increased, at least in the short term.
With regard to adhesive joints, strength loss may be dictated by adhesive plasticisation as discussed above or by displacement of the adhesive from the substrate on water ‘wicking’ along the interface, or both. Such effects on joint properties will be considered in some detail in Chapter 4.
Creep. In general, polymers exhibit a degree of visco-elastic behaviour and thus for full characterisation of such a material a knowledge of its rate dependent response is necessary. To determine the long-term behaviour of a material either stress relaxation or creep tests may be used. The former involves monitoring the time- dependent change in stress which results from the application of a constant strain to a specimen at constant temperature. Conversely,
A———— 2-part epoxy polyamine (aliphatic)
В———— 2-part epoxy polyamide
C———— 2-part epoxy polyamine (aromatic)
 |
Shev^rsnglh о——————— 2-part epoxy polyamine (aliphatic adduct)
A———— 2-part—- epoxy polyamine (aliphatic)
В———— 2-part—- epoxy polyamide
— 2-part epoxy polyamine (aromatic)
D——— 2-part epoxy polyamine (aliphatic adduct)
E——— 2-part epoxy polysulphide
4-
_______________________ A
__________________ , Absorbed
0 і 2 3 ~4 5 6 ’water (%)
Fig. 2.24. Moisture dependence of bulk adhesive flexural modulus (Ref. 26).
creep can be thought of as time-dependent flow under constant load which may lead to fracture or creep rupture.
There are three main parameters affecting creep, namely stress, time and temperature. Moisture can also affect the creep of absorbent materials, such as some of the structural adhesives. During creep experiments the values of stress and temperature are kept constant. As different materials exhibit different creep properties, a method of characterising creep is required. This is usually in terms of its creep modulus (M,) given by where
сt0 = constant applied stress and
e, = total strain at time t.
Alternatively the creep compliance (Ct) can be obtained from
|
|
A conventional creep curve as exhibited by most materials is illustrated in Fig. 2.25 although many engineers present the data using log axes to produce a graph of the form shown in Fig. 2.26. Data from families of strain-time curves at various values of constant stress are used to produce isochronous stress-strain curves (Fig. 2.27). These are obtained by cross-plotting stresses and strains at various times from the commencement of loading. The results of creep tests can also be used to derive constant strain, or isometric, curves of stress versus time, also as illustrated in Fig. 2.27.
Creep tests on structural adhesives can be divided into tests on bulk hardened adhesive specimens and tests on adhesively bonded joints. The former provides information on the mechanical properties of the adhesive rather than the joints made from them. Fig. 2.28 displays the change in creep modulus with time for a range of cold — cure epoxy adhesives(26). These curves were derived from four point bend tests on adhesive prisms loaded in accordance with Fig. 2.16 using extreme fibre stresses ranging from 0.25 to 2.0 N/mm2. The curves represent the stability of the adhesive with time under
 |
Strain
sustained load and in this respect a flatter curve is beneficial. However, the relatively low long-term modulus values for the polyamide and polysulphide epoxies may give some cause for concern as to their potential structural efficiency under sustained load. In general, the more highly cross-linked the hardened adhesive structure and the higher the curing temperature, and hence Tg, the better the creep resistance.
Creep curves obtained from experiments using bulk hardened specimens do not necessarily compare with those obtained from joints under similar stress conditions due to the nature of adherend restraint. A further contributing factor to this difference is the reduction in stress concentrations which will occur in joints during creep.
|
Stress Strain Stress
Stress |
|
|
Fig. 2.27. Isometric and isochronous curves taken from a set of creep data. (a) Isometric stress v. log time, (b) Creep curves, (c) Isochronous stress v. strain.
Creep rate varies with stress level — generally the higher the stress, the greater the creep rate. It has also been suggested (29) that when the sustained stress is lower than some equilibrium value then indefinite creep will not occur. When stressed above this value the material will creep to failure. What happens to adhesive joints at the lower stress levels is perhaps more important in civil engineering than rapid creep to failure at high stress levels. This is so because structural adhesive joints tend to be designed to withstand low mean stresses, for example 10% of ultimate, but which have to be sustained for many years. Creep rate also varies with temperature. An increase above room temperature results in an increase in creep rate until the Tg is reached, when there is a marked further increase in creep rate.
It has already been noted that the Tg is sensitive to rate of loading and this has led to the development of time/temperature superposition techniques being used to characterise the response of polymer
|
Creep modulus M(t) (kN/mm2)
log time (s) Fig. 2.28. Change in creep modulus with time for a range of epoxy adhesives (Ref. 26). |
systems. If a family of property curves (e. g. strain v. time or log time) is plotted at a series of temperatures all the curves can be shifted parallel to the time axis until superposition produces a master curve at some reference temperature. This curve can then be used to predict the sample behaviour that would be obtained at the reference temperature if the property concerned was measured directly over the wide time scale which results. The technique permits the prediction of long-term creep behaviour at different temperatures and load conditions from limited short-term data.
The superposition approach can be used to produce a constitutive equation which expresses the creep compliance (Ca) of the adhesive in terms of a reference creep compliance (Cr) and shift factors for stress (flCT), temperature (aT) and resin content (av) such that Ca = Cr x aa x At x av x f". The method has been used by Dharmarajan et al. (30) to characterise the creep behaviour of epoxy, polyester and acrylic mortars in the form of prism specimens under 3 point loading. From relatively short-term tests, strain v. time curves such
|
Creep strain x m3 4°C) ofMN/m2) v |
|
Fig. 2.29. Strain v. time curves for polyester and epoxy concretes (Ref. 30). (a) Polyester concrete, (b) Epoxy concrete. |
as those shown in Fig. 2.29 have been used to produce reference curves as illustrated in Fig. 2.30. Data have shown the time exponent m to be independent of the polymer matrix but to have a value of 0.6 in flexural loading and about 0.2 in tension or compression. The transition from stable creep (m < 1) to creep rupture (m > 1) appears to occur at an equilibrium stress of between 45% and 55% of the short-term ultimate strength of the system.
Fatigue. The fatigue performance of an adhesive under cyclic loading will be influenced by the visco-elastic nature of the material and its resistance to crack propagation, or fracture toughness. At low frequencies and high temperatures, visco-elastic effects will predominate in a similar manner to that experienced with creep. At higher frequencies and lower temperatures fracture due to crack propagation either within the adhesive layer or at the adhesive/substr — ate interface will tend to control the number of load cycles that can be sustained prior to failure. Because the fatigue performance of an adhesive in a structural joint is closely linked to the joint
|
In C (C in m2/MN)
In t (f in hours) Fig. 2 .30. Reference curves for flexural creep (Ref. 30). |
Concluding remarks
configuration and stress distribution within the joint further discussion on fatigue will be deferred to Chapter 4.
Опубликовано в рубрике 