You are here

Thin ply composites

News International-French

29 Jul 2011

Large tows of carbon fibre can be spread by a simple, non-intrusive process from which plies down to 1/6 of the conventional 0.12 mm thickness can be obtained. Laminates made from such plies showed remarkable resistance to micro-cracking, delamination and splitting under both static and fatigue loading. With thick-thin ply hybrids, superior performance, lower weight and lower minimum gage can all lead to cost-effective applications not easily achieved with thick plies.

(Published on June-July 2005 – JEC Magazine #18)




This presentation covers some recently obtained results of the processing and property determination of carbon fibre reinforced composite laminates made from thin plies. The premature failure modes by micro-cracking and delamination have imposed severe limitations on the more effective utilization of the fibre strength. For example, traditioinal threshold for micro-cracking is 0.5% strain while most modern fibers have a failure strain about 2%. Thus if micro-cracking can be suppressed or prevented, a composite laminate can be designed to carry load up to 2%, four times the micro-cracking threshold. It will be shown that thin plies in a laminate do not lead to microcracking and delamination. It is therefore possible to load the laminates to a much higher level than what has been accepted thus far.


Spreading process and static behavior


A tow spreading process has been developed by Kawabe of Fukui Technology Center. He blew air across a 12k or 24k tow to create thin plies down to 1/6 of the conventional 0.12mm (5 mil) ply thickness. Because the spreading process does not induce any stress to the fibres so they are not damaged.



When laminates made from these thin plies, all the anticipated ply properties are recovered. When plies were made into cross-ply, quasi-isotopic and other laminates, tests showed that micro cracking and delamination were suppressed up the ultimate strength. The stress-strain curves were different so were the failure modes varied from massive delamination for regular thick ply (T=0.12mm=5.00mil) coupons to a clean break for thin ply (T/6 = 0.02mm = 0.83 mil) coupons.


The dramatic difference in the coupon failure modes can be seen in the photos below.



Close-up photos of the coupon failures in figure 2 are shown in the following two photos, in figure 3. The upper photo shows the failure mode of thick ply (5 mil) quasi-isotropic laminates where extensive micro-cracking and delamination were evident. The lower photo, on the other hand, shows a thin ply (T/6) laminate where a clean failure was seen. The gross fracture was highly localized with no micro-cracking and delamination away from the fracture surface.


When thick and thin ply laminates with open hole were tested, X-ray photos showed massive micro-cracking in thick ply laminates while the thin ply stayed clean. The tensile load was 455 MPa or 90 percent of the ultimate load. The failure modes were very different with the thin ply laminate having a clean fracture with no delamination, consistent with earlier coupon tests. Fatigue tests also showed massive micro-cracking and delamination along all free edges for thick ply laminates while the thin ply laminates stayed clean.


Steve Tsai heads an international team of experts in mechanics of composites which provides timely modeling and validation of composites failure process. Tools such as Super Mic-Mac, presented in JEC Composites No10, is a self-teaching software programme for preliminary design. Progressive failure is being modeled by Element Failure Method; and a health management system, by a strain mapping technique. Thin ply technology is one of the most recent activities.




Fatigue behavior of thin ply laminates


The same quasi-isotropic specimens can be tested under fatigue loading. In the figures below we show the X-ray image of laminates with open hole subjected to tensile-tensile fatigue of 357 MPa (51 ksi) after 100,000 cycles. Edge delamination and micro cracking began to appear after 20,000 cycles in the thick ply laminate on the right, while the thin ply laminate stayed clean. Even up to 100,000 cycles, thin ply laminates stayed amazingly resilient and free of damage. The thick ply laminates had massive damages far beyond the point of practical utility as a structure. For hard laminates where [0] plies are 50% of the laminate, tension-tension fatigue at a stress level of 490 MPa (70 ksi) after 73,000 cycles for thin and thick ply laminates. The extent of damage along the interior and outside edges was drastically more severe for the thick ply.






We therefore realized that thin plies have much to offer. Laminates and structures can be made strong and damage tolerant without high toughness resin, interlaminar resin layers and/or 3D reinforcement. Having lower weight and lower cost simultaneously is possible by using higher design allowable. The fear of premature failure by micro-cracking and delamination may be a thing of the past. Designers now have more options to select the best combinations of materials and processes.


The increased cost of having additional laying-up thin plies can be offset of a new automated lamination machine that can provide biand tri-directional hybrid sublaminates with thick and thin plies. This machine was invented by Mitsuya Company, also located in Fukui, Japan. A picture of this machine is shown in figure 7.



More than one-metre wide sublaminate with continuous length is being made available. Such material can have major impact on how composite structures are made. The lay-up and ply drop can also be simplified in both the design and assembly processes with significant reduction in laminate warpage from curing. While work is continuing in process development, analytical modelling of failure processes and design methodology, we firmly believe that thin ply technology is transfomational because it is simple, cost-effective and opens up to new concepts not feasible with thick plies.




Support of this work by Yasushi Kiyobayashi of Itochu Corporation is gratefully acknowledged. Specimen preparation by Kichiro Isida of Mitsya Company, and testing and evaluation by Ran Kim and Sangwook Sihn of University of Dayton Research Institute are also acknowledged. The contents of this paper were presented at JEC in Paris, April 6, 2005, AIAA in Austin, April 19, 2005, and will be presentes at ICCMin Durban, June 28, 2005, and ICMAT 2005 in Singapore, July 8, 2005.