Thermosets and Structural Adhesives

Keywords: hot melts, structural adhesives, thermosets, cure, Tg., cure cycle, gel point RH007

INTRODUCTION

Structural adhesives are usually thermosetting polymers. Generally applied to the surface as low viscosity, reactive liquids, their viscosities increase rapidly as they polymerize ‘ and crosslink in the joint to form a rigid, high strength bond. Therefore, knowing the viscosity as a function of cure time and cure temperature is important for establishing optimum cure conditions. ‘

Thermoset polymers form the matrix in filled plastics and fiber-reinforced composites used in a diversity of products. These range from consumer items and auto body panels to advanced composites for printed circuit boards, aerospace structural components, and expensive, high-performance sports equipment.

THE CROSSLINKING REACTION

In order to build a chemical network, one of the monomers has to be two-functional and one three-functional (figure Figure 2. Measurement of the viscosity of a curing resin in a steady 1). The formation of a crosslinked network can be followed, shear and a dynamic oscillatory test measuring the viscosity as a function of time and temperature as shown in figure 2. As the network structure builds, the The time/cure mode is an efficient way to generate viscosity viscosity increases due to the growing resistance to the flow. profiles of a curing thermoset polymer. The viscosity profile Near the gel point, the steady viscosity increases rapidly and yields the initial viscosity, minimum viscosity, approximate becomes unmeasurable – eventually the stiffening sample gel point, and optimum heating rate of a thermoset during breaks. In contrast an oscillatory measurement of the resin’s curing. Data obtained on a curing epoxy compound are viscosity can be made as the reaction proceeds through the shown in figure 3. This information can then be used to gel point - and beyond- until the resin becomes a stiff solid. develop a production cure cycle.

107 106 Temperature Ramp 5 °C/min

6 10 as]

a] G’ 105 η * [P [P approx. gel point G”

G” η* 105 Viscosity x 10-4 Moduli G’, 104

Minimum Viscosity Comple

103 103 80 100 120 140 160 Temperature T [°C]

Figure 3. Measurement of the minimum viscosity and approximate gel point for a curing epoxy compound

Figure 1. Schematical representation of structural development during thermoset curing for the reaction of a dimer with a trimer

1 RH007 The crossover point of the storage and loss modulus EFFECT OF HEATING RATE curves (Figure 3) is a good estimate of the gel point of a curing thermoset. But it is only an estimate, because, as Figure 6 shows viscosity versus reaction time at several heating the gel network structure forms, the modulus changes. So rates for a typical epoxy structural adhesive. With increasing unless measurements on the gelling system are made fast rate, the viscosity minimum does not only occur earlier, it is enough, the exact gel point will be obscured by changes also lower. This is significant – for example in a lamination in the modulus. The reason is, that the instant of gelation the process, the resin viscosity must be low enough to wet the modulus of the critical gel exhibits power law behavior and embedded fibres uniformly, but not so low to cause excessive tan δ becomes frequency independent for a brief moment bleeding at the laminate edges or in the en-capsulation of (G’ ~ G’’~ aωn). This point can be found by making several integrated circuits by transfer-molding. The viscosity must not simultaneous frequency measurements to obtain tan δ in the be too high to damage the fragile fibers. time scale of the developing gel (figure 4). Because tanδ is 1E + 04 independent of frequency at the gel point, the curves pass Frequency = 10 rad/s through a single point and unambiguously define the instant Initial Temperature = 50 °C the gel forms as shown in figure 5. 1E + 03 3 °C/min 2 °C/min 1 °C/min

6 Equilibrated Stoichiometry 10 as] PDMS r=1 t (data Winter & Chambon) c 1E + 02 105 tc + 6tc + 2tc - 2tc - 6 a] Viscosity η [P 4

[P 10 1E + 01 G” T b 103 G’, T 0.5

t-tc log A 1E + 00 102 -6 +2 0 1000 2000 3000 4000 5000 6000 -2 +1 Moduli b Time t [s] 0 0 101 G’ +2 -1 Figure 6. Measurement of the viscosity profile of an epoxy resin as a G” +6 -2 function of the heat-up rate 100 10-4 10-3 10-2 10-1 100 101 102 103 104 105 106 107 EFFECT OF MOISTURE Frequency A + ωa [rad/s] T Excessive water content in an epoxy resin can cause a Figure 4. Frequency sweeps measuring G’ and G’’ at any point in time plasticizing effect, lubricating the polymer and causing a during cure decrease in viscosity. Figure 7 shows a comparison of the viscosities measured during cure for both a good and a bad performing resin sample during impregnation of a woven 10 glass cloth. The bad resin did not adhere to the cloth but Frequencies instead was flowing through the weave. The results show that 1 rad/s the bad resin’s viscosity peaked at approximately 100 °C but 8 4 rad/s then decreased to a minimum which was below that of the 16 rad/s good resin. 6 Good resin 103 Resin with higher moisture content an δ t

4 as] 102 Precise Gel Point 2 101 Viscosity η* [P x 0 25 30 35 40 100 Comple Reaction Time [min]

Figure 5. The gel point identified by the intersection of tan δ curves 10-1 generated from the simultaneous frequency sweeps 50 100 150 200 Temperature T [°C] Figure 7. Viscosity as a function of temperature for a good and a bad (excess moisture) sample

2 RH007 At the boiling point of water, some of the excess moisture 106 evaporated, leaving voids in the material and causing the resin’s viscosity to rise. But as the temperature increased further, the remaining excess water acts briefly as a plasticizer as] and the viscosity decreased to a lower minimum than that of the good resin, which was drier. When pressure was applied to the resin during impregnation of the glass cloth, the low viscosity resulted in resin migration and poor adhesion of the 105 At mold temperature:

resin to the cloth. Viscosity η * [P Sample A x Sample A + 1% Znst. Moisture, even in low concentration, causes large changes in Sample B the rheological and chemical behavior of adhesives. Figure At barrel temperature: 7 shows the change in the reaction profile due to moisture. Comple Sample A This data can be used to predict, for example, bond layer Sample A + 1% Znst. Sample B thickness and cycle times in printed circuit board laminations. 104 0 2 4 6 8 10 12 EFFECT OF CURE TEMPERATURE Cure Time t[min]

Beyond the gel point, the modulus increases to give the Figure 9. Effect of Zn stearate on the viscosity profile during cure at adhesive its ultimate strength. Adhesive strength is evaluated barrel and mold temperature from dynamic modulus measurements as a function of Simulations of the viscosity under various cure conditions frequency and temperature. Figure 8 shows the dynamic provides the necessary data in process design: barrel elastic modulus G’ and tan versus temperature. For three δ temperature, nozzle size and effect of shear heating. different cure temperatures of 93, 135 and 350 °C, the Tg of the cured epoxy resin shifted from 130 °C to over 250 °C. EFFECTS OF VARIOUS CURING AGENTS ON EPOXY RESIN 1E+11 1E+01 Epoxy resin used in printed circuit boards (PCB’s) requires sufficient rigidity during exposure to high end- 93 °C 1E+10 135 °C 1E-00 use temperatures (approximately 140 °C), and soldering temperatures (approximately 250 °C). When exposed to a]

[P 350 °C high operating temperatures, the PCB’s may became t an δ 1E+09 1E-01 dimensionally unstable, resulting in a defective product and material waste when not formulated correctly. 1E11 100 Modulus G’ 1E+08 1E-02 Frequency 1Hz

1E10 1.0 a]

1E+07 1E-03 [P 200 250 -150 -100 -50 0 50 100 150 1 Temperature T [°C] 1E9 an δ Figure 8. The extent of the cure of a thermoset as a function of the t 0.1 curing temperature shows in Tg 1E8 age Modulus G’ or

St Azelaic Acid & BOMA 0.01 BOMA EFFECT OF ADDITIVES 1E7 DDM BTDA The ideal injection molding thermoset (RIM) needs to be 1E-3 highly stable at a low viscosity level at the screw barrel -150 -100 -50 0 50 100 150 200 250 300 350 temperature, yet cure rapidly upon entering the mold. Temperature T [°C]

Figure 9 shows a simulation of the cure behavior of two Figure 10. Effect of different curing agents on the mechanical phenolic compounds at two different temperatures, the barrel properties of the final product and the mold temperature. Compound A showed fast curing Using different curing agent, an increase of the maximum in the mold and made a good final part, but had too high a continuous use temperature by increasing the glass transition viscosity at the barrel temperature. The compound B injected (Tg) was obtained under the same curing conditions. well with low shear heating, but cured slower. The addition of DMA measurements as a function of temperature (ASTM 1% of Zn stearate to compound A lowered the viscosity at the D 4065), shown in Figure 10, were performed on the final barrel temperature compared to coumpound B, but without samples. Sample 4 (BTDA) was found to be the most efficient, significantly affecting the cure time. increasing the Tg to over 270 °C, higher than both the end- use temperature and the soldering temperature. The peak in tan δ or glass transition temperature (Tg) can be evaluated

3 RH007 to determine the degree of cure. The storage modulus, G’ 200 (amount of energy stored in a material), is a measure of Temperature 103 180 each sample’s stiffness. Curing agents not only affect the Neat Resin Prepreg 160 rate of change of the polymer’s viscosity, but also it’s physical as] 102 properties by increasing the crosslinking density and as such 140

the performance of the final product. T [°C]

101 120 e

CURE BEHAVIOR OF COMPOSITE PREPREGS 100 atur Viscosity η * [P 0 x 10 80

The resin in many structural adhesives is also used as matrix emper T material in fiber reinforced composites. Dynamic mechanical 10-1 60 testing is a valuable characterization technique for composite Comple matrix resins. Rheological analysis of the neat resin to predict 40 10-2 Test Frequency 10 rad/s processing characteristics of glass and graphite reinforced 20 composites has become a standard procedure in product 0 50 development of advanced composites. Cure Time t [min] Figure 11. Viscosity profile of composite (prepreg) and the neat resin Figure 11 compares results for the cure of a graphite prepreg during cure and neat resin using parallel plate fixtures in a oscillatory rotational rheometer. Note that, although there are large differences in minimum viscosity, cure times are identical.

CONCLUSIONS

Dynamic mechanical measurements of thermosetting adhesives can give valuable information about their curing behavior, including minimum viscosity and gel point, glass transition temperature and degree of crosslinking.

A more comprehensive discussion of this topic is available from TAI in “Understanding Rheological Testing—Thermosets.”

ACKNOWLEDGEMENT

Revised by A. Franck, TA Instruments

For more information or to place an order, go to http://www.tainstruments.com/ to locate your local sales office information.

*This note was previously published as AAN003. 4 RH007