Adhesively Bonded Joints Between Composites
Kevin Potter
Kevin Potter 2012
Advantages of bonded joints
Load distributed over a large area High joint efficiency (Strength/weight ratio) Low part count No holes in the basic laminate Potential for low cost manufacture . Dissimilar materials can be bonded without corrosion problems
Kevin Potter 2012
Advantages of bonded joints
Air and water tight to some degree, although not necessarily perfect Fatigue lives of bonded construction tend to be good due to the flexibility of the adhesives With good design the joints can retain a high level of residual strength after initial cracking A bonded joint is generally aerodynamically better than a rivetted joint
Kevin Potter 2012
Disadvantages of bonded joints
Difficult to inspect non-destructively Very sensitive to peel loading - must be avoided Good bonds require a good fit of parts Usually permanent and cant be disassembled Strength can be affected by temperature / humidity Surface preparation is critical to a good bond Clean room conditions may be called for and the curing process must be closely controlled
Kevin Potter 2012
Disadvantages of bonded joints
Bonding can be expensive due to labour intensive production Most aerospace adhesives require high cure temperatures (>180C) that may damage some joint materials With poor design the joints can fail suddenly Adhesive layer may act as an electrical insulator causing problems when panels must be in electrical connection , e.g. for lightning strike. Difficult to arrive at a realistic strength by analysis
Kevin Potter 2012 5
Fundamentals of adhesion 1
Hydrogen Bond Van der Waals Interatomic Distance Bond Energy Covalent Bond
Ionic Bond
Molecular energy vs. interatomic distances for different types of bond
Kevin Potter 2012 6
Fundamentals of adhesion 2
Repulsion
Energy Resultant Attraction
Interatomic distance
Attractive and repulsive forces and the resultant for a typical molecule.
Kevin Potter 2012 7
Critical surface energies and wetting:
Adherend Material PTFE PVC Polyamide Iron Tungsten 18 40 46 2030 6800
Typical
Typical Adhesive 30 - 47
(mN/m)
Kevin Potter 2012
Adhesive types
Epoxides Primary aerospace adhesives in film and paste form Polyurethanes Can give very high toughness systems Cyanoacrylate Rapid cure but very brittle and low peel strength Anaerobic Generally retention and threadlocking types Reactive acrylic Tough and fast for automotive uses Phenolic (Redux) The first structural adhesive for aircraft, still used Evaporative Solvent based glues, may be used on aircraft interiors VHB Tapes High strength double sided tape Hot melt Higher strength and curing variants are now used
Kevin Potter 2012
Adhesive toughening
Crack-stopping elastomer microspheres in adhesive
Kevin Potter 2012
10
Basic joint types
Single lap Double lap Scarf Bevel Stepped lap Butt strap Double butt strap Butt Peel Grout
Kevin Potter 2012 11
Deformations and shear stress distribution in a single-lap bond
Totally rigid adherends uniform shear stress
Elastic adherends shear stress concentration at ends of joint
Kevin Potter 2012
12
Effect of yield on joint stresses
For most adhesive formulations used in structural applications (usually epoxies) the adhesive yields prior to failure. This yielding leads to a change in the stress distribution shown previously and makes the shear stress more uniform across the joint.
A. Prior to yielding
B. Yielding established
C. Yielding complete
Kevin Potter 2012
13
Transverse (peel) stresses
For single lap joints there is an eccentricity in the load line that leads to distortion of the adherend geometry and very high transverse tensions (or peel stresses) at the ends of the joint
For double lap joints this effect is eliminated, but there are still transverse stresses at the ends of the joint due to induced bending moments
Kevin Potter 2012
14
Adhesive failure modes in composite joints
Cohesive failure, wholly within the adhesive layer, preferred Failure due to induced through thickness stresses in the composite - very common, not ideal Adhesive failure at interface - unpredictable must avoid
Kevin Potter 2012
15
Recommended Fibre Orientations
The fibres in the layer immediately adjacent to the adhesive layer must be aligned in the same direction as the load path.
If the fibres are perpendicular to the load then premature failure WILL occur. If there are multiple potential loading directions use a plain weave cloth as the surface ply
Kevin Potter 2012
16
Design details, dos and donts
Kevin Potter 2012
17
Assembly issues 1. Tolerancing
Recommended bondline thicknesses range from about 0.1mm to 1mm, although thicker bondlines are sometimes used. This can create problems with tolerances
All dimensions need to be tightly controlled
Dimensions of each part can be relaxed
Kevin Potter 2012
18
Assembly issues 2. bondline control
Without bondline control joints can be geometrically distorted or become voidy and defective
1. Use film adhesive, gives a thin and well controlled bondline, so long as the surfaces being bonded are flat and parallel. Pressure needed Use internal spacers in centre of bondline, gives a thicker bondline but some lack of flatness is OK. Use fully tooled bonding jig to directly control bondline and fillet geometry
2.
3.
Kevin Potter 2012
19
Detail design considerations
Adherend Materials: strength, stiffness, thermal and moisture expansion coefficients, surface treatment. Adhesive properties: strength, ductility and toughness. Static, fatigue, shock and creep loading. Environment and aging. Analysis methods: Hand calcs, computer based methods Variability: bond defects, surface prep, manufacturing tolerances Inspectability Testing
Kevin Potter 2012
20
10
Effect of end detail on strength of steel/CFRP double lap joints Fail load (kN/mm) Steel Cl CRFP Actual 0.93 0.89 0.94 ---3.05 Theory 1.05 1.08 1.10 2.0 3.3
Kevin Potter 2012
21
Reason for improvement in strength from geometry
Max 12.3MPa
Max peel stress 12.3MPa, fails in tension in adhesive
Max 290MPa
Max peel stress for same end load 290MPa, fails in through thickness tension in laminate. In real world adhesive yielding reduces peak stress
Stress in through thickness direction in the CFRP for Linear FEA model
Kevin Potter 2012 22
11
Test methods
14
Kevin Potter 2012
23
Peel tests
Kevin Potter 2012
24
12
Other test methods
15
Napkin Ring Test
18
Boeing Wedge Test
Kevin Potter 2012
25
Some results for the Boeing Wedge Test
300 Gic J/m2 200 5% brine Hydraulic fluid Acetone
This data is for a carbon fibre / epoxy matrix composite. Most composites are not this sensitive to the environment
100
10
20 30 40 50 Hours of exposure
60
While this data is for a composite rather than an adhesive joint as such, the crack propagation is controlled by whichever is weaker of the bond and the laminate
Kevin Potter 2012 26
13
Recommended design practice
Bondline thickness 0.1 to 0.5mm Minimise peel loads & stresses Bond length > 30 mm Ultimate Load/unit width < 1 kN/mm Use stepped-lap joints for thick adherends Use internal end-chamfer & fillets . Use Finite Element Analysis (FEA) computer modelling to refine the geometry if needed
Kevin Potter 2012
27
Surface Preparation
Bonded joints can be very strong, but this strength is critically dependent on surface preparation, and even minor amounts of contaminants such as oil can destroy the bond strength. Surfaces must always be clean and dry prior to bonding. Surfaces are often abraded or grit-blasted prior to bonding The use of peel ply without secondary abrasion may prove to be ineffective Metallic surfaces may be acid etched or subjected to other chemical or physical pre-treatments prior to bonding.
Kevin Potter 2012
28
14
Estimation of joint strength 1
So long as the adhesive can yield a reasonable first estimate of maximum possible joint strength can be got from the area of the joint and the yield stress in shear. This assumes that the joint is fully yielded prior to failure and does not fail from through thickness tension or by adhesion failure, or by tension in the adhesive before complete yielding. This may be OK for short (<30mm) joints but long joints tend to fail by other processes before complete yielding. Doubling the overlap length of the joint is very unlikely to double the strength.
Kevin Potter 2012 29
Estimation of joint strength 2
Reserve of strength to tolerate flaws
Joint strength N/mm of joint width
Strength of perfect adhesive bond
Perfect bond weaker than adherends
Strength of flawed adhesive bond Flawed bond weaker than adherends Adherend strength
Flawed bond stronger than adherends
Adherend thickness Relative strengths of adherends and adhesively bonded joints (with and without flaws)
Kevin Potter 2012 30
15
Estimation of joint strength 3
Achieving a good prediction for the strength of a bonded joint is really rather complicated, requiring detailed stress analysis and the application of a validated failure criterion. This is generally beyond the capability of most organisations, and in any case there is no universally accepted way of carrying out the prediction. Making test joints representative of the real joint and testing them under the same loadings and environmental conditions as the real joints may be the best we can do in many cases then apply a safety factor (but what controls the safety factor?)
Kevin Potter 2012 31
Fatigue effects 1
60 50 Max average shear stress MPa 40 30 20 10 0 0 2 4 Log cycles to failure 6
Fatigue performance for well made composite double lap joints Dotted lines are 95% confidence limits on performance
Kevin Potter 2012 32
16
Fatigue effects 2
To confidently predict fatigue life a good consistent set of fatigue data must be available and .. Any changes in failure mode under fatigue loading at the endurance of interest must be known and understood. The fatigue test environment must be an accurate reflection of the use environment so that there is a direct correspondence between fatigue life and operational life.
Kevin Potter 2012
33
Fatigue effects 3
Mode 4 Mode 3 Mode 2 Mode 1
Load
mode 1, adherend failure, mode 2, cohesive failure, mode 3, peel failure mode 4, adhesive interface failure,
Mode 1 Mode 2 Mode 3 Mode 4
Log cycles to failure
Load/lifetime curves for different failure modes, showing that differing fatigue degradation rates for different failure modes can lead to changes in the expected fatigue failure mode
Kevin Potter 2012
34
17
Environmental effects, temperature and moisture
50 45
Average shear stress MPa
40 35 30 20 15 10 5
DLS, exposed. BU data UD CFRP adherends
Single lap shear 3M data CFRP cloth reinforced adherends Unexposed 90%RH 9 weeks
The single lap shear joints are more affected by low temperatures than the DLS joints due to bending/peel effects on the increasingly brittle adhesive, but are similar at high temperature. The adhesive was 3Ms EC3448 paste. Other adhesives would behave in different ways
-60 -40 -20
20
40
60
80 100
Test temperature Deg C
Kevin Potter 2012
35
Results of testing bonded I beams
3 point bending (see photo) 1st trial, steel support beam yielded 2nd trial, failure at 47Tons 4 point bending: Failure at 63Tons Failure was probably in the adhesive
Kevin Potter 2012
36
18
Conclusion
It is possible to make reliable high strength bonded joints with composite adherends, BUT. Through thickness (peel) failure in the laminate is critical and must be avoided Interface failure must be avoided by good surface preparation The effects of the use environment must be accounted for Simple and fully validated strength prediction methods are not available and some testing will generally be required in support of design
Kevin Potter 2012
37
19
