Composites

Abstract

A composite which comprises a first layer of a fibre reinforced polymer and a second layer of a fibre reinforced polymer, between which is an intervening layer comprising an array of thermoplastic islands.

Claims

1. A composite comprising a first layer of fibre reinforced thermoset polymer and a second layer of fibre reinforced thermoset polymer, the first and second layer forming an interface between the two layers, and a thermoplastic polymer distributed at and/or towards the interface between the two layers, the composite being stiffer than an equivalent fibre reinforced thermoset polymer composite absent the thermoplastic polymer.

2. The composite of claim 1 wherein the composite has improved stiffness and/or interlaminar shear strength.

3. The composite according to claim 1, wherein the thermoplastic polymer is discontinuously distributed at and/or towards the interface between the first and second layer.

4. The composite according to claim 1, wherein the thermoplastic polymer is preferentially distributed at or towards the interface such that the relative concentration of thermoplastic polymer is at a peak at or towards the interface.

Description

[0080] In a first set of experiments we investigated the application of various thermoplastic polymers onto a carbon fibre epoxy prepreg.

Examples 7 to 3

[0081] In a first set of experiments we dissolved thermoplastics materials, in various solvents, in the amounts set out in Table 1 below:

TABLE-US-00001 TABLE 1 Parameters of thermoplastic solution and print characteristics Composition of ‘ink’ Diameter of Example Solute wt % Solvent printhead/μm Pattern 1 PMMA 5 DMF* 60 Hexagon 2 PEG 5 Distilled 60 Hexagon water 3 PEG 5 Pure Ethanol 60 Hexagon

[0082] Where: Mn.sub.(PMMA)=20,000; DMF is N,N-Dimethylformamide; Pattern is the shape described by the array of dots.

[0083] The ink jet printer used was a Jetlab 4XL (supplied by MicroFab) and operated in each case according to the parameters as set out in Table 2.

TABLE-US-00002 TABLE 2 Parameters for operation of ink jet printer Example Jet 1 RT 2 RT 3 Polymer

[0084] Where: RT is room temperature and Polymer is Polymer Jet PrintHead, which can be heated to temperatures as high as 240° C.

[0085] In each case the Rise Time, Dwell Time and Dwell Voltage may be altered to ensure that stable droplets are generated.

[0086] The various polymer solutions were each printed or deposited onto a carbon fibre enforced epoxy prepreg, designated as Cycom977-2 (supplied by Cytec Engineered Materials of Wrexham, UK) as a substrate.

[0087] In each case the as-printed substrates were cured according to the following cure regime: [0088] {circle around (1)} Ramp: 20° C..fwdarw.100° C. (rate: 2° C./min) [0089] {circle around (2)} Dwell: 30 min [0090] {circle around (3)} Ramp: 100° C..fwdarw.177° C. (rate: 2° C./min) [0091] {circle around (4)} Dwell: 120 min [0092] {circle around (5)} Ramp: 177° C..fwdarw.20° C. (rate: 2° C./min)

[0093] In order to investigate the behaviour of the islands of thermoplastics due to exposure to elevated temperatures of curing cycle a series of microscopy studies was undertaken:

Example 1A

[0094] Referring to FIG. 1, a series of PMMA islands were applied to a glass slide. The PMMA was doped with fluorescein to enhance optical microscopy and images were taken with an optical microscope (CMO-SLP, Olympus).

[0095] FIG. 1a shows islands prior to heating and FIG. 1b shows the same islands after exposure to the curing cycle. FIGS. 1a′ and 1b′ are images of the same solutions deposited in a hexagonal array. FIG. 1c shows the spots after being exposed to an elevated temperature for 2 hours.

[0096] The results show that the islands reduce in volume (due to solvent evaporation), and demonstrate some ‘coffee staining’ as a result of evaporation of the solvent, but they remain in place and are not subject to translational motion.

Example 1B

[0097] Referring to FIG. 2 a series of PMMA islands were applied to a glass slide. The PMMA was doped with fluorescein to enhance optical microscopy and fluorescein images were taken with an optical microscope (Image Xpress).

[0098] FIG. 2a is before curing, FIG. 2b is after curing and FIG. 2c is after being heated to 100° C. for 2 hours.

[0099] The results clearly show that the dots stay as distinct islands.

Example 1C

[0100] Referring to FIG. 3 a series of PMMA dots were applied to Cycom977-2. Images were taken with an optical microscope (Polyvar).

[0101] FIG. 3a is before curing, FIG. 3b is after curing and FIG. 3c is after being heated to 100° C. for 2 hours.

[0102] In this case there is no visible evidence of the thermoplastic islands on the surface after exposure to heat, however, we believe that the thermoplastic remains, as evidenced in the previous images.

Example 1D

[0103] Referring to FIG. 4 a series of fluorescein-doped PMMA dots were applied to Cycom977-2. Fluoroscein images were taken with an optical microscope (Image Xpress).

[0104] FIG. 4a is before curing, FIG. 4b is after curing and FIG. 4c is after being heated to 100° C. for 2 hours.

[0105] In this case it appears that the fluorescein within the islands sprayed out after heating and with increased temperature (there is some evidence of fluorescein in FIG. 4c).

[0106] In order to investigate this behaviour a series of interferometry experiments were undertaken.

Example 1E

[0107] Using a 3D optical microscope (Contour GT supplied by Bruker) interferometry images were taken of PMMA dots applied to a glass slide, before curing (FIG. 5a) and after curing (FIG. 5b).

[0108] The images show that the islands remain in place.

Example 1F

[0109] Using a 3D optical microscope (Contour GT supplied by Bruker) interferometry images were taken of PMMA dots applied to Cycom977-2, before curing (FIG. 6a) and after curing (FIG. 6b).

[0110] The results show that islands are evident before curing but not after curing.

[0111] Whilst not wishing or intending to be bound by any particular theory, we believe that the islands have either splayed out or have penetrated into the surface of the prepreg. However, we believe that there are local peak concentrations of thermoplastics islands at or towards the surface.

Example 2A

[0112] Referring to FIG. 7 a series of fluorescein-doped PEG dots were applied to Cycom977-2. Fluoroscein images were taken with an optical microscope (Image Xpress).

[0113] FIG. 7a is before curing, FIG. 7b is after curing and FIG. 7c is after being heated to 100° C. for 2 hours.

[0114] In this case it appears that the fluorescein within the islands sprayed out after heating and with increased temperature (there is some evidence of fluorescein in FIG. 7c).

[0115] A series of further experiments were undertaken to determine the physical performance of the composites made according to the invention.

Examples 3 to 6—Interlaminar Shear Strength

[0116] In each case a composite test piece was made either using eight prepreg layers (designated virgin), or with eight prepreg layers and the solutions set out in Table 1.

[0117] Determination of apparent interlaminar shear strength by short beam method (SBS) according to BS EN ISO 14130:1998. [0118] Calculation of interlaminar shear stress: τ=3F/4bh & τ.sub.M3F.sub.M/4bh [0119] Where: F is the load, F.sub.M is the maximum load; [0120] b is the width of the test specimen; b=10.0 mm [0121] h is the thickness of the test specimen; h=2.0 mm

Example 3

[0122] The results for repeat interlaminar sheer strength tests (5 runs each) are shown in FIG. 8 for a series of experiments conducted on test pieces which are as cured.

[0123] The results are as shown in Table 3:

TABLE-US-00003 TABLE 3 Results Slope Maximum Load τ.sub.M (×10.sup.3 N/mm) (×10.sup.3 N) (MPa) Avg. SD Avg. SD Avg. SD Virgin 5.340 0.355 2.985 0.228 111.9 8.5 5% PEG + Ethanol 5.781 0.177 2.817 0.107 105.6 4.3 5% PEG + Water 7.093 0.239 3.590 0.105 134.6 3.9 5% PMMA + DMF 7.147 0.082 3.118 0.081 116.9 3.0 *SD: standard deviation

[0124] It can be seen that samples with printed self-healing agent have higher stiffness which is represented by slope of straight part of load versus extension curves than that of virgin samples. And with printed PMMA, samples have the highest stiffness. Moreover, there is no significant difference among virgin and self-healing agent printed samples regarding to average maximum load and average maximum interlaminar shear stress, showing no reduction in the structural integrity of the system due to the deposited self-healing agent. That said, there is a notable reduction in the standard deviation for 5% PMMA printed system. Although we neither wish nor intend to be bound by any particular theory, we believe this to be due to a better damage control by arresting crack propagation through PMMA islands, whilst maintaining the adhesion between the prepreg plies, and hence an increased engineering predictability in the optimised system.

Example 4

[0125] The results for repeat interlaminar sheer strength tests (5 runs each) are shown in FIG. 9 for a series of experiments conducted on test pieces which are as cured and subsequently ‘healed’ by exposing to the cure regime as set out above.

[0126] The results are as shown in Table 4:

TABLE-US-00004 TABLE 4 Results Slope Maximum Load (×10.sup.3 N/mm) (×10.sup.3 N) τ.sub.M (MPa) Avg. SD Avg. SD Avg. SD Virgin 5.900 0.368 3.170 0.269 118.9 10.1 5% PEG + 6.780 0.044 3.088 0.128 115.8 4.8 Ethanol 5% PEG + 6.906 0.145 3.363 0.088 126.1 3.3 Water 5% PMMA + 7.205 0.406 3.177 0.066 119.1 2.5 DMF

[0127] It can be seen that samples with printed self-healing agent have higher stiffness than that of virgin samples. And with printed PMMA, samples have highest stiffness. Moreover, there is no significant difference between virgin samples and self-healing agent printed samples regarding to average maximum load and average maximum interlaminar shear stress values, showing no reduction in the structural integrity of the system due to the deposited self-healing agent.

[0128] Comparing the results from Examples 3 and 4, it can be seen that the average maximum interlaminar shear stress (τ.sub.M), average maximum load and average stiffness of almost four groups are slightly enhanced after the healing cycle, it could be either caused by the printed self-healing agent or by the post curing of epoxy in pre-preg itself or both.

Example 5

[0129] The results for repeat interlaminar sheer strength tests (5 runs each) are shown in FIG. 10 for a series of experiments conducted on test pieces which are as cured and subsequently ‘damaged’ by exposing the test piece to a tensometer and stopping loading just after the maximum load displaced is seen on the monitor.

[0130] The results are as shown in Table 5:

TABLE-US-00005 TABLE 5 Results Slope Maximum Load τ.sub.M (×10.sup.3 N/mm) (×10.sup.3 N ) (MPa) Avg. SD Avg. SD Avg. SD Virgin 4.268 0.172 2.463 0.079 92.4 3.0 5% PEG + Ethanol 5.620 0.283 2.376 0.235 89.1 8.8 5% PEG + Water 6.061 0.537 3.054 0.318 114.6 11.9 5% PMMA + DMF 5.770 0.393 2.593 0.073 97.3 2.7

[0131] We can see (by comparing the results from Example 3 and those in Table 5) that the mechanical properties of four groups are reduced after damage, as expected. Samples with printed self-healing agent are stiffener than the virgin ones. Moreover, no significant difference is observed regarding to average maximum load and average maximum interlaminar shear stress among four groups, showing no reduction in the structural integrity of the system due to the deposited self-healing agent. Since the damage process cannot guarantee introducing the same amount of damage to every test specimen, the results of some groups have a variation which does not obey the common rules.

Example 6

[0132] The results for repeat interlaminar sheer strength tests (5 runs each) are shown in FIG. 11 for a series of experiments conducted on test pieces which are as cured and subsequently ‘damaged’ by exposing the test piece to a tensometer and stopping loading just after the maximum load displaced is seen on the monitor and subsequently ‘healed’ by exposing to the cure regime as set out above.

[0133] The results are set out in Table 6.

TABLE-US-00006 TABLE 6 Results Slope Maximum Load τ.sub.M (×10.sup.3 N/mm) (×10.sup.3 N) (MPa) Avg. SD Avg. SD Avg. SD Virgin 5.046 1.011 2.520 0.406 94.5 15.2 5% PEG + Ethanol 5.224 0.464 2.486 0.300 93.2 11.3 5% PEG + Water 6.334 0.502 2.881 0.339 108.1 12.7 5% PMMA + DMF 5.756 0.316 2.674 0.174 100.3 6.5

[0134] It can be seen that samples with printed self-healing agent are slightly stiffener than virgins.

[0135] And comparing the results of Table 6 with those of Table 5, almost every parameter of each group has a slight enhancement except τ.sub.M of 5% PEG+Water group which, we believe may well be a variation of experiment process. Because the damage process cannot guarantee all damaged samples were introduced the same amount of damage, that means the results of Table 6 samples in have higher damage level than the samples set out in Table 5, the values of all parameters of the Example 6 should lower than that of Example 5, in which case, we will not be able to see the self-healing efficiency if the self-healing efficiency is not apparent enough.

[0136] The summary of the results is set out in Table 7 below:

TABLE-US-00007 TABLE 7 Results Summary of SBS tests Average value n = 5 Slope Load.sub.MAX τ.sub.M Damage Heal (×10.sup.3 N/mm) SD* (×10.sup.3 N) SD* (MPa) SD* Virgin x x 5.340 0.355 2.985 0.228 111.9 8.5 x ✓ 5.900 0.368 3.170 0.269 118.9 10.1 ✓ x 4.268 0.172 2.463 0.079 92.4 3.0 ✓ ✓ 5.046 1.011 2.520 0.406 94.5 15.2 5% PEG + x x 7.093 0.239 3.590 0.105 134.6 3.9 Water x ✓ 6.906 0.145 3.363 0.088 126.1 3.3 ✓ x 6.061 0.537 3.054 0.318 114.6 11.9 ✓ ✓ 6.334 0.502 2.881 0.339 108.1 12.7 5% PEG + x x 5.781 0.177 2.817 0.107 105.6 4.3 Ethanol x ✓ 6.780 0.044 3.088 0.127 115.8 4.8 ✓ x 5.620 0.283 2.376 0.235 89.1 8.8 ✓ ✓ 5.224 0.464 2.486 0.300 93.2 11.3 5% PMMA + x x 7.147 0.082 3.118 0.081 116.9 3.0 DMF x ✓ 7.205 0.406 3.177 0.066 119.1 2.5 ✓ x 5.770 0.393 2.593 0.073 97.3 2.7 ✓ ✓ 5.756 0.316 2.674 0.174 100.3 6.5

[0137] From SBS test, it can be concluded as following: [0138] a. Printed polymeric agent of only 0.02% weight addition can significantly increase the shear stiffness of the composite before and after damage in shear, where the composite is most vulnerable and such improvements can significantly improve the damage tolerance of the overall system. [0139] b. Even though this experiment was conducted to evaluate structural soundness of the system alone, there is evidence of self-healing efficiency as a result of the damage process, showing no reduction in the structural integrity of the system due to the deposited self-healing agent.

Examples 7-10

[0140] A further series of experiments were undertaken to determine the mode I interlaminar facture toughness, G.sub.IC for unidirectionally reinforced materials, according to BS ISO 15024:2001. Each test sample, whether a ‘virgin’ or test sample, comprised twelve layers of prepreg material, the difference between ‘virgin’ and ‘test’ samples being the provision of thermoplastic polymer at the interface between successive layers in the ‘test’ samples.

[0141] According to the standard, there are several important G.sub.Ic values of particular points which are shown in FIG. 25. These points are defined as follows: [0142] 1) NL point—non linear point of deviation from linearity on the load versus extension trace [0143] 2) 5%/MAX point—the point which occurs first on the loading the specimen between: [0144] a) The point of 5% increase in compliance (C.sub.5%) from its initial value (C.sub.0); [0145] b) The maximum load point. [0146] 3) PROP points—propagation value of fracture toughness, consisting of points of discrete delamination length increments beyond the tip of the insert or starter crack tip marked on the load-extension trace, points where the crack has been arrested being excluded The important points are shown graphically in FIG. 25, which is a load-displacement curve for a DCB test showing initiation from the resulting mode I precrack followed by crack propagation and unloading.

Comparative Example 7

[0147] In order to determine a base line, ‘virgin’ samples were tested. The tested samples were then subjected to the ‘heal’ procedure and tested again to determine if any self-healing occurred.

[0148] The results are shown in FIG. 12 and set out below in Table 8:

TABLE-US-00008 TABLE 8 Results for Comparative Example 7 G.sub.IC (kJ/m.sup.2) G.sub.IC (kJ/m.sup.2) Before healing cycle After healing cycle NL 5%/MAX Avg. PROP NL 5%/MAX Avg. PROP point point points point point points 1 0.12473 0.14723 0.1825 0.03625 0.05311 0.07353 2 0.15226 0.17972 0.17082 0.06807 0.08122 0.07775 3 0.12558 0.12749 0.12761 0.06182 0.06327 0.062 4 0.15013 0.16851 0.16379 0.07914 0.08282 0.07273 5 0.11611 0.12309 0.12177 0.0873 0.09164 0.08688 Average 0.13376 0.14921 0.1533 0.06652 0.07441 0.07458 SD 0.01636 0.02481 0.02704 0.01957 0.01575 0.009

Example 8

[0149] An experiment was conducted using a composite formed from Cycom977-2 and provided with dots applied in accordance with Example 2 (as set out in Table 1 and 2), i.e. 5% PEG in water. The samples were tested. The tested samples were then subjected to the ‘heal’ procedure and tested again to determine if any healing occurred.

[0150] The results are shown in FIG. 13 and set out in Table 9

TABLE-US-00009 TABLE 9 Results of Example 8 G.sub.IC (kJ/m.sup.2) G.sub.IC (kJ/m.sup.2) Before healing cycle After healing cycle NL 5%/MAX Avg. PROP NL 5%/MAX Avg. PROP point point points point point points 1 0.21025 0.22645 0.25724 — — — 2 0.18471 0.19142 0.22501 0.1845 0.19109 0.19153 3 0.17013 0.19376 0.21784 0.13057 0.1328 0.13037 4 0.1546 0.17097 0.20639 0.11506 0.11948 0.12052 5 0.19575 0.2283 0.234 0.1435 0.18661 0.18015 Average 0.18226 0.20218 0.22798 0.14341 0.1575 0.15564 SD* 0.02307 0.02466 0.01913 0.02976 0.03666 0.03541

Example 9

[0151] An experiment was conducted using a composite formed from Cycom977-2 and provided with dots applied in accordance with Example 3 (as set out in Table 1 and 2), i.e. 5% PEG in ethanol.

[0152] The samples were tested. The tested samples were then subjected to the ‘heal’ procedure and tested again to determine if any healing occurred.

[0153] The results are shown in FIG. 14 and set out in Table 10

TABLE-US-00010 TABLE 10 Results of Example 9 G.sub.IC (kJ/m.sup.2) G.sub.IC (kJ/m.sup.2) Before healing cycle After healing cycle NL 5%/MAX Avg. PROP NL 5%/MAX Avg. PROP point point points point point points 1 0.14259 0.15983 0.17721 0.11871 0.12171 0.11021 2 0.18581 0.18884 0.19879 0.06709 0.06928 0.05951 3 0.11545 0.13142 0.13464 0.09012 0.09193 0.08407 4 0.12066 0.13389 0.13222 0.08993 0.09331 0.08488 5 0.16128 0.17188 0.17279 0.11456 0.11487 0.09212 Average 0.14516 0.15717 0.16313 0.09608 0.09822 0.08616 0.02916 0.02465 0.02885 0.02102 0.0208 0.01824

Example 10

[0154] An experiment was conducted using a composite formed from Cycom977-2 and provided with dots applied in accordance with. Example 1 (as set out in Table 1 and 2), i.e. 5% PMMA in DMF.

[0155] The samples were tested. The tested samples were then subjected to the ‘heal’ procedure and tested again to determine if any healing occurred.

[0156] The results are shown in FIG. 15 and set out in Table 11

TABLE-US-00011 TABLE 11 Results for Example 10 G.sub.IC (kJ/m.sup.2) G.sub.IC (kJ/m.sup.2) Before healing cycle After healing cycle NL 5%/MAX Avg. PROP NL 5%/MAX Avg. PROP point point points point point points 1 0.32722 0.32753 0.24347 0.20649 0.21411 0.15388 2 0.23818 0.28664 0.25612 0.11919 0.12132 0.11574 3 0.36688 0.37538 0.27223 0.19209 0.19501 0.17671 4 0.25826 0.26611 0.19555 0.13601 0.15049 0.13204 5 0.25664 0.25858 0.18249 0.1542 0.1555 0.11857 Average 0.28944 0.30285 0.22997 0.1616 0.16729 0.13939 SD* 0.055 0.04859 0.03902 0.03692 0.03706 0.02574

[0157] Referring now to FIG. 16, we can see the average G.sub.Ic values of NL, 5%/MAX and avg. PROP points of PMMA printed specimens are higher than virgin and PEG printed specimens both before and after healing cycles, which means the printed PMMA enhances the interlaminar fracture toughness of all specimens. Whilst the PEG+ethanol results are only marginally better than the virgin results there is a slight improvement for this system as well.

[0158] Advantageously, the use of ink jet printing allows for easy application of the thermoplastic materials. Moreover, using direct application techniques (or indeed removal techniques) it is possible to introduce functionally graded improvements in fracture toughness, i.e. to have a resistance to fracture which changes across the width, or length of a part. FIG. 17 provides a schematic of two proposed parts with different regions bearing the thermoplastic polymer.

Example 11

[0159] The results, shown in FIG. 18, show results for crack resistance in a part in which regions that were partly virgin and partly printed with polymer patterns. The graph shows how G.sub.Ic swaps over and increases once the fracture reaches the patterned region.

Example 12

[0160] In FIG. 19, G.sub.IC is seen to improve as the printed polymer concentration is increased (in each dot). The fracture toughness of the system printed with 20% polymer concentration (upper line) is significantly higher compared to 10% system.

Example 13

[0161] In a further set of experiments the stiffness of samples (before and after heating) were determined. The results are shown in FIGS. 20 and 21. The samples were virgin (v), as Example 1 (a), as Example 2 (b) and as Example 3 (c). It was demonstrated that samples containing the printed thermoplastic polymer have up to a 33.8% increase in stiffness compared to the virgin samples.

Example 14

[0162] In a further set of experiments Cycom977-2 was printed with higher concentration of polymer (20% PMMA in DMF) in dots applied in accordance with Example 10.

[0163] The samples were tested. The tested samples were then subjected to the ‘heal’ procedure at slightly lower temperature of 160° C. for 40 minutes to avoid the reduction in toughness seen in the previous data due to the secondary cross-linking that occurs at 177° C., and tested again to determine if any healing occurred.

[0164] The results are shown in FIG. 22. The graphs show non-printed control samples (NP) and samples printed with an array of dots comprising 20 wt/wt % of PMMA. The fracture toughness was significantly increased both before and after self-healing treatment (i.e. heating to 160° C. for 40 mins), and the standard deviation was significantly reduced for the system with 20% concentration PMMA printed patterns.

Example 15

[0165] Synchrotron X-ray computed tomography was carried out on the samples at Joint Engineering, Environmental and. Processing (JEEP) beam line at Diamond Light Source. 1800 projections were recorded with the exposure time of 0.1 second over a 180-degree rotation; the beam energy was 53 keV. The distance between the sample and the detector was short therefore no phase contrast was observable. The detector was 2560×1373 pixels with a resolution of 3.24 μm per voxel. The width of the specimens were bigger than the field of view and region of interest tomography was performed which due to low attenuation of carbon did not require a ROI correction during reconstruction. A back projection code with limited ring-artifact suppression was used to reconstruct the data. The evidence of self-healing is presented in FIG. 23, showing the SBS tested samples before (left) and after (right) thermal treatment with the layers being fused in the self-healed sample.

Example 16

[0166] A fracture toughness test Was carried out to evaluate the difference between interleaf (film) area between the plies, commonly used to increase the toughness in composites at the expense of other mechanical properties due to the loss of adhesion between the plies, and the here presented printed patterns Which allow epoxy surfaces in CFRP to remain in contact. The systems are presented in FIG. 24 (a-c), as a) 20% concentration filth between the plies, b) 20% printed dots in hexagonal patterns (equivalent to 10% film deposited volume fraction of PMMA) and c) summary of the results for three systems; showing a notable increase in fracture toughness for 20% PMMA printed dots in CFRP compared to 10% printed full area or film in the same system.

[0167] As the amount (%) of PMMA is increased in film printed samples, the standard deviation is also increased, whereas the standard deviation is decreased for 20% PMMA dots due to a better crack arrest and higher degree of engineering predictability. 20% film pattern is only shown for the standard deviation purposes, considering that it would require close to 50% concentration of PMMA in hexagonal patterns to provide a comparative result for the discrete system, which is difficult to achieve using the existing inject printing system. The 20% film composite would also significantly increase the weight.

[0168] These results demonstrated the improved G.sub.IC for patterned surfaces and reduced standard deviation of the system at only 0.02% addition of discrete PMMA islands, compared to the film printed ‘interleaf’ method in the same system, if a comparable volume fraction of PMMA is used.

In Conclusion

[0169] From the optical, fluorescein and interferometry images of PMMA (and other) dots on glass slides we can see the printed PMMA dots can stay as droplets after curing cycle. But from fluorescein and interferometry images of PMMA dots on pre-preg, we cannot see any dots after curing cycle, which means the printed PMMA is likely to react with epoxy and form localised bonds whilst fluorescein may depart due to the high temperature.

[0170] In the experiments set out above, in order to investigate the self-healing efficiency, a damage process has been employed to introduce an appropriate amount of damage into specimens for self-healing agency to heal. Double cantilever beam (DCB) and short beam shear (SBS) tests have been adopted to evaluate the self-healing efficiency, due to the damage occurring between the composite plies during the tests. Fluorescein was added to ink in order to investigate behaviours of the printed polymer dots before and after heating to different temperatures. From optical, fluorescein and interferometry images of PMMA dots on glass slides, it can be seen the printed PMMA dots stay as droplets after curing cycle. The fluorescein sprayed out after curing cycle both in PEG and PMMA cases, which can be spotted from fluorescein images of polymer dots on pre-preg. From interferometry images of PMMA dots on pre-prep, no dots are visible after curing cycle.

[0171] From the SBS test results, it is apparent to see that the printed self-healing agent can stiffen composite materials both before and after self-healing, and that printed PMMA samples have the highest stiffness among the four groups. Besides, we can see the values of slope, maximum load and τ.sub.M of samples after healing are higher slightly than that of before healing both in undamaged groups and damaged groups. This may be either because of the post curing of the epoxy in pre-preg or the printed self-healing agent but we prefer the latter explanation. Since the damage method cannot guarantee introducing the same amount of damages into samples, some results did not obey the common rules, which caused some variations.

[0172] From DCB test results, we can see almost all average values of initial (non-linear), 5% of the maximum load and average propagation values of fracture toughness with printed PMMA are higher than that of virgin and PEG printed specimens both before and after healing cycle, which indicated that printing PMMA between composite plies significantly enhanced the interlaminar fracture toughness, both before and after the thermally treated damage (self-healing process).

[0173] It is clear that printed self-healing agent can stiffen the composite, which can be concluded from SBS test results. Samples with printed PMMA have the highest stiffness among the four groups. However, no apparent self-healing efficiency has been observed among virgin samples and samples with self-healing agent through SBS test. It can be seen that specimens with printed PMMA have the highest mode I interlaminar fracture toughness (G.sub.Ic) both before and after healing cycles among the four groups, indicating a substantial recovery of the material after introduced damage, and a significant increase in fracture toughness of the system before the damage has been introduced. This also implies that the material is likely to sustain more load, due to its improved capacity to resist crack initiation and propagation. Hence, the overall service life of the material will improve and also impart self-healing property, further extending its durability and lowering maintenance costs. This simultaneous increase in a number of thermo-mechanical properties both before and after damage, is achieved at less than 0.1% weight increase, rendering this as a unique system, capable of imparting properties into the original material that could not be achieved in any other way.

[0174] In respect of improving the self-healing properties, we believe, although we do not wish or intend to be bound by any particular theory, that increasing the molecular weight of the polymer will lead to improved self-healing, We also believe, and our experiments bear out, that increasing the amount of polymer will lead to increased robustness of the dots during cure and will lead to markedly improved self-healing properties. This, we believe, is due to increased viscosity of the polymer in the solution and/or islands of higher concentration.