Doctor of Philosophy
Department of Materials Engineering
McGovern, Scott T., The development of embedded sensors to assess the fatigue response of adhesive joints in marine environments, Doctor of Philosophy thesis, Department of Materials Engineering, University of Wollongong, 2007. https://ro.uow.edu.au/theses/748
Historically, the use of adhesives within the marine industry has been limited due to the effects that salt water can have on the adhesive joint in service. Salt water has been shown to have adverse effects on the adhesive bond strength due to corrosion of the adherend surfaces, delamination of the adhesive at the interface and plasticisation of the adhesive material. However, if the progress of such degradation can be properly monitored and better understood then greater confidence may be imparted with the use of adhesives for marine structures, and with adhesive incorporation, the potential benefits of improved weight reduction, speed of manufacture and reduced cost may be realised. The work in this thesis was undertaken to develop miniature sensors that may be embedded within the adhesive bondline so that a continual health monitoring of the adhesive joint in service may be undertaken and a better understanding of the mechanisms of adhesive breakdown in marine environments may be achieved.
Sensor substrates were manufactured from 125 μm polyester insulated platinum wires dip coated with conducting polymer materials. Various conducting polymers were investigated for their ability to sense localised environmental conditions such as the presence of water and pH changes - recognised factors that have been linked to adhesive joint degradation via the instigation of adhesive plasticisation / delamination or the onset of corrosion.
Humidity sensors were manufactured from polyaniline (PAn) based materials due to the high conductivity changes observed with PAn in the presence of water. The best response was obtained from the polymer blend that was manufactured by chemically polymerising aniline monomer using ammonium persulphate in the presence of poly butylacrylate -covinyl acetate (P(BuA-co-VAc)) and camphor sulfonic acid (HCSA) to generate a solid that could be readily dissolved in dichloromethane. A blend of 15 w/w% PAn was developed and a portion of the PAn polymer solid was further de-protonated from the emeraldine salt (ES) to the emeraldine base (EB) form to be applied as a protective barrier layer.
Sensors were constructed by first dip-coating the substrate in 15 w/w% PAn P(BuA-co- VAc) ES polymer followed (after drying) by dip-coating in 15 w/w% PAn P(BuA-co-VAc) EB polymer. The sensors had an exponential response and showed high sensitivities with an 800% resistance change over the entire 95% change in RH. For the largest change in humidity (from 3 to 95% RH) the sensors took approximately 4 to 5 hours to reach 90% of the equilibrium resistance on humidifying and 35 hours was required to reach 90% of the equilibrium resistance upon desiccation.
The best pH sensors were manufactured from a processable polymer blend made from poly 3,4 -diethoxythiophene (PEDT) and poly vinylidene fluoride (PVDF). The blend was able to be cast from an acetone solvent and had an approximate PEDT content of 11w/w%. The sensors showed a linear response to pH changes between 5 and 11 with an approximate 350% increase in resistance when increasing the pH from 1 to 13. However, hysteresis was observed with cycling between high and low pH and significant drift was encountered when cycling over multiple cycles. Further problems were also encountered with the embedding of the sensor in the Araldite 2015 test adhesive, and as such these sensors were abandoned from further study.
To examine the humidity sensor performance in service, Aluminium 5083 Araldite 2015 adhesive joints were constructed in a lap-shear configuration, comparing two different pretreatments of an Optimised FES (FES) and an acetone solvent wipe (solvent wipe). Joints were exposed to a constant temperature salt water spray and the lap-shear tensile strength of the adhesive joints was tested after varied time of exposure for joints with and without embedded sensors and for both pre-treatment types. For the same salt water exposure conditions, joints were also exposed to a constant 1 Hz fatigue cycle with a varied load and S-N curves were generated for joints with and without sensors for both pre-treatments.
It was found that the average dry tensile strength of the solvent wipe and FES joints was 10.3 +/- 1.3 MPa and 16.2 +/- 0.3 MPa respectively. The tensile strength of the solvent wiped joints showed a steady decrease with exposure to salt water spray such that after 21 days exposure time, joint strengths had dropped to 50% of the strength of the unexposed samples. In comparison, the FES samples showed no loss in joint strength even after 23 days exposure to salt water spray. The difference in joint performance between the two pretreatments was attributed to the high degree of surface roughness of the FES treated samples enabling a greater degree of mechanical interlock between the adhesive and substrate.
Solvent wiped joints that were tested in fatigue showed a marked decrease in strength when exposed to marine conditions. It was shown that for the same applied fatigue loads joints that were tested 'dry' lasted approximately 50 times longer than joints that were exposed to a salt water spray. This dramatic change was somewhat uniform across the whole S-N curve even for short exposure times less than 1 day and this accelerated decay of joint strength was attributed to the generation of micro-cracks during the fatigue process that allowed for the faster ingress of water into the joint further accelerating the breakdown seen with static tests.
The greater strength with the FES joints was further highlighted with the fatigue tests. When exposed to marine conditions, the FES joints were seen to last up to 10,000 times longer than solvent wiped joints and no failure was seen with the 'dry' FES joints at allwithin the constraints of the testing apparatus. The 10 million cycle fatigue strength of the FES joints exposed to marine conditions was estimated to be approximately 1.3MPa. It was hoped that the embedded sensors may be used to ascertain a critical water content within each of the adhesive systems that might suggest the onset of adhesive failure, however it was seen that the embedded sensors became unstable between 24 and 48 hours of exposure to salt water. The change in stability of the embedded sensor with static tests was seen after water uptake in the adhesive of 2 w/w%, analogous to an equilibrated environment of approximately 75% relative humidity. It was thought that for each of the tests the sensor breakdown was most likely due to a combination of three main factors: a) swelling of the adhesive with water ingress, promoting delamination of the sensing polymer from the underlying sensor substrate b) further deprotonation of the underlying ES sensing material to that of the nonconducting PAn EB c) the generation of micro cracks and internal strains within the adhesive system (especially within the fatigue tests) that mechanically affected the performance of the embedded sensor response.
It was discovered that with full immersion of Araldite 2015 adhesive films in water, the maximum uptake of water that was achievable was approximately 6 w/w% taking over 6 weeks to reach full stability. Due to the short time frame encountered before sensor breakdown in static tests of 24 - 48 hours it could be concluded that a) is not the most likely cause of sensor failure. Likewise c) is less likely due to the absence of any applied external stress during salt water exposure during static tests.
It may be postulated then, that if b) is the likely cause of sensor breakdown, the sudden change in environmental conditions that could initiate a net increase in resistance (andreduction in the perceived mass uptake of water) by means of deprotonation would be an increase in pH - similar to what is known to occur at the onset of corrosion.
In the current study there were indications that corrosion had been initiated with ingress of water equivalent to 75% RH at the site of the sensor (causing this increase in resistance) and may help explain the significant drop in fatigue strength that was seen with the solvent wiped joints after only 1 day exposure.
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.