- Email: [email protected]
Influence of the mechanical behaviour of different adhesives on an interference-fit cylindrical joint
International Journal of Adhesion & Adhesives 47 (2013) 63–68
Contents lists available at ScienceDirect
International Journal of Adhesion & Adhesives journal homepage: www.elsevier.com/locate/ijadhadh
Influence of the mechanical behaviour of different adhesives on an interference-fit cylindrical joint Giorgio Gallio a,n, Mariangela Lombardi a, Davide Rovarino b, Paolo Fino a, Laura Montanaro a a b
Politecnico di Torino, DISAT Department of Applied Science and Technology, INSTM RU Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy MW Italia S.p.A., Via Pavia 72, Rivoli, Italy
art ic l e i nf o
a b s t r a c t
Article history: Accepted 5 August 2013 Available online 25 September 2013
Hybrid adhesive joining techniques are often used in many industrial sectors to design lightweight structures. A hybrid adhesive joint results from the combination of adhesive bonding with other traditional joining methods such as welding and mechanical fastening, with the aim of combining the advantages of the different techniques and overcoming their drawbacks. This study focuses on the interference fitted/adhesive bonded joining technique. In this application, two cylindrical components are coupled together by inserting one into the other, after having placed an adhesive between them. Generally anaerobic acrylic adhesives, also known as “retaining compound” are used for this application. However the effect of the adhesive nature and of its mechanical and adhesive responses on the performance of the hybrid joint is still unclear. The aim of the present research is to improve the understanding of the behaviour of different adhesives, including rigid epoxies and flexible polyurethanes, in the presence of an interference-fit. Static strength of bonded and unbonded interference fit joints have been compared in order to investigate the role of the different adhesives. & 2013 Elsevier Ltd. All rights reserved.
Keywords: Interference fit Mechanical properties of adhesives Destructive testing Joint design
1. Introduction In recent years, the use of adhesive bonding technology has greatly increased being a suitable technology able to bond dissimilar materials and often to guarantee an uniform stress distribution in the joint area [1–3]. However, the polymeric composition of adhesives implies durability issues, especially in harsh conditions (temperature and humidity). In order to overcome these problems, adhesive bonding can be used in combination with other traditional joining methods, such as mechanical fastening techniques (e.g. rivets or bolts) or welding techniques, generating a hybrid joint. In this way, hybrid adhesive joints should combine the advantages of the different techniques and, if possible, overcome their drawbacks [4]. This study focuses on the interference fitted/adhesive bonded hybrid joint. In this technique cylindrical assemblies are coupled together by inserting one part in another and by adding an adhesive between them. The hub–shaft geometry is usually exploited in technical literature in order to study this hybrid bonding technique [5–8], allowing to evaluate the contribution of both the techniques to the resultant resistance of the hybrid joint. Moreover, it is previously demonstrated that the behaviour of this hybrid joint can be also influenced by the assembling technique (press-fit or shrink fit), the type of sustained load, the fitted position and the interference level [5–11]. n
Corresponding author. Tel.: þ 39 320 6016761; fax: þ 39 011 090 6329. E-mail address: [email protected] (G. Gallio).
0143-7496/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijadhadh.2013.09.021
The acrylic anaerobic adhesives, commercially named “retaining compounds” [12], are usually exploited in the interference fitted/adhesive bonded hybrid joints [5–11], being possible in this configuration to exploit their curing technology based on a one component configuration on active metal surfaces. In this system, in fact, the presence of the interference contributes to guarantee the anaerobic polymerisation, protecting the curing layer of adhesive from the oxygen. Notwithstanding this, it is important to consider that in certain particular industrial applications other curing technologies could be more appropriate, especially when clearance zones are included in the joint design or different mechanical properties have to be matched. To the best of our knowledge, the effect of the adhesive nature and therefore of its mechanical and adhesive responses on the performance of the hybrid joint is still unclear. The adhesive nature and the curing technology could affect the performances of the interference joints. For this reason, in this study the behaviour of the principal families of structural adhesives, from a rigid epoxy to a more flexible polyurethane, will be evaluated in an interference fit joint.
2. Materials and methods 2.1. Design of the samples The role of the different adhesives in the interference fit joint was studied in a hub/shaft geometry. The hollow tubes and the
64
G. Gallio et al. / International Journal of Adhesion & Adhesives 47 (2013) 63–68
shafts, made of C40 steel, were designed according to the requirements imposed by the traction test machines (Fig. 1). The coupling diameter of the samples was set to 30 mm. This represented a versatile value that facilitated a minimum extent of thermal expansion in the case of performing either the press fit or the shrink fit assembling techniques. The threaded extremities of the samples were designed to be correctly hitched to the test machines, furthermore the use of a threaded grabber avoids slipping effects during traction tests. The head of the shaft was designed with a diameter of 40 mm after the threaded tip. This part had the same diameter than the external part of the hub and this allowed the alignment of the component during the press coupling. A coupling length of 10 mm was fixed. A loose clearance fit or a strong interference fit can be obtained by varying the dimensional tolerances of the shafts. In this work it was decided to prepare a series of samples with an interference
Fig. 1. A hub and shaft.
coupling of 20 mm, according to H7/p6 ISO standard of tolerance, and a series with a clearance coupling of about 40 mm according to a coupling tolerance of H7/f7. The technical drawings of the hollow tubes and shafts are illustrated in Fig. 2. The coupling diameter of each hub and shaft were checked in four points in the first 10 mm (the coupled length) with a Trimos horizontal measuring and calibration instrument. Then an average diameter was calculated for each sample in order to select the best coupling pairs. The roughness of the samples in terms of Ra and Rz was also checked by averaging two 8 mm profiles for each hub and shaft. Then a mean value of Ra and Rz for every coupled assembly was calculated. 2.2. Test method The interference fit joints were assembled through the press fit. The alignment between the shaft and the hub is a crucial parameter to obtain repeatable joints. A not-axial assembling induces far higher coupling and decoupling loads then an axial one. A method of guides, schematically illustrated in Fig. 3, was exploited to guarantee the alignment of the couples. The method was validated by measuring the eccentricity between the two components, after putting in rotation the coupled joints. The pressing was performed by means of an automatic press in order to assure the same pressing rate for each sample. The samples were tested under a static axial push out load, performed with a traction test using a MTS test machine at a crosshead rate of 1.3 mm/min. Considering that for the chosen hybrid joint design no standard tests are available, the crosshead speed was selected on the basis of the low load rate generally suggested in the standard shear tests on adhesive joints. Four different commercial adhesives based on different structural adhesive systems were tested:
A rigid epoxy adhesive Hysols 9492 (will be referred as “R-EP” in the paper).
A flexibilized and toughened epoxy adhesive: Scotch-welds DP490 (will be referred as “FT-EP” in the paper).
A polyurethane-epoxy modified adhesive: Araldites 2029 (will be referred as “PU” in the paper).
Fig. 2. Dimensions and tolerances of the specimen: (a) hollow tube and (b) shaft.
Fig. 3. Schematic representation of the guides system exploited to guarantee the alignment of the joints.
G. Gallio et al. / International Journal of Adhesion & Adhesives 47 (2013) 63–68
An anaerobic acrylic adhesive: Loctites 620. (will be referred as “AC” in the paper). The traction tests were carried out on samples joined with the interference only, samples bonded with the four different adhesives in clearance condition, and hybrid joints interference fitted and bonded with the four different adhesives. For each of the above-described case studies, 4 samples were tested. The same surface preparation was followed for every sample. The surfaces of hubs and shafts were cleaned with acetone, dipped in acetone and put in an ultrasonic bath for 5 min. According to the producer suggestions, the followed curing protocols were 72 h at ambient temperature for R-EP and AC adhesives; 24 h at ambient temperature followed by a postcuring at 80 1C for 1.5 h for FT-EP; 16 h at 40 1C for PU.
65
break up point occurred in correspondence with the maximum of the peak. After the adhesive failure, the load drastically decreased and the interference contribution was then recorded. In particular, in the cases of the two epoxy systems the sudden failure of the adhesive induced the machine to record load values close to zero before the interference contribution resistance comes into play. The decoupling curves of the AC samples presented some differences from the others: they were characterised by a lower adhesive failure peak and an higher interference contribution. It seemed that the cured adhesive enhanced the friction forces in the contact zone heightening the interference contribution and implying a less drastic crash of the adhesive. This particular trend, already observed where anaerobic acrylics are used [8], could be partially imputed to the absence, in these joints, of a cured spew fillet. Indeed in these anaerobic systems, the spew fillets that generally formed on both limits of the bonded area were not able to harden, because the oxygen inhibited the polymerisation.
3. Results 3.1. Hybrid joints decoupling behaviours The results of the traction tests were analysed and for each case study the four decoupling curves were averaged for obtaining a representative, mean curve, also applying a smoothing filter. The mean load/displacement curves of the different hybrid systems and the unbonded interference samples were compared in Fig. 4. The decoupling curves of the “only interference” samples were characterised by a first peak after the elastic region, that corresponds to the point in which the joint forces are broken. After this point, that is considered the resistance of the joint, the shaft starts to move. In these conditions, a quite constant load was recorded, and the curves were characterised by a sinusoidal trend. It is reasonable to assume that the low speed of testing, chosen to evaluate the adhesive behaviour in the hybrid joints, induced a discontinuous decoupling in “only interference” conditions, with consecutive stick–slip states. Grip and friction phenomena took place during the decoupling process, inducing in some case an increase in the maximum load after the joint break. Differences in the mechanical behaviour when different adhesives were employed can be observed from the experimental data in Fig. 4. The curves of the R-EP, FT-EP and PU two-component adhesives were characterised by two main phases: an initial load peak and a following plateau with the stick/slip trend typical of the interference contribution. The initial peak was due to the contribution of the adhesive plus the interference. The adhesive
Fig. 4. Mean load/displacement curves of the bonded and unbonded interference samples.
Fig. 5. Fracture shaft surfaces of the decoupled joints: (a) typical black layer of the cured FT-EP adhesive on the shaft; (b) typical white layer of the cured AC adhesive on the shaft; (c) grey residues of the cured R-EP adhesive on the shaft and (d) R-EP sample where cured particles are NOT present on the mating surface.
66
G. Gallio et al. / International Journal of Adhesion & Adhesives 47 (2013) 63–68
The mating surfaces of every studied hybrid joints after the decoupling, presented some cured adhesive residues both on the hubs and the shafts, except for the R-EP. In the case of AC, PU and FT-EP samples, the morphology of the cured layer was similar among all the four samples. It is difficult to define if a cohesive type of fracture took place because a very thin layer of adhesive remained inside the junction, as shown in Fig. 5. On the contrary, as illustrated in Fig. 5c and d, different residue quantities of R-EP remained on the shafts and the hubs of the various samples, implying a slightly wider variability in their load–displacement curves. The spreading of the adhesive in an interference fit joints depends on its rheology. Generally a low viscosity is recommended in closer joints, while a high viscosity is preferable when gap filling is required [12]. The FT-EP, AC and PU adhesives tested in this work are thixotropic, so they are “non-sag” pastes that lower their viscosity when subjected to mechanical stresses (as during the press-coupling operation). The thixotropy was not declared from the supplier in the case of the R-EP adhesive, therefore a part of this adhesive could be spewed away during the pressing and could not properly wet the joint surfaces. More studies have to be conducted to deeper analyse the relation between viscosity and the quantity of adhesive that remains in the junction.
interference samples. As an example, the mean load/displacement curves of the 4 samples for each of the 3 different cases (Bonded, Interference, Bondedþinterference) related to the FT-EP system are plotted in Fig. 6. In addition, in Fig. 6, the corresponding mean stress/displacement curves of the same FT-EP samples (Bonded, Interference, Bondedþ interference) are illustrated. The stress was calculated dividing the load by the bonded area, that was considered constant for the entire elastic trait of the load/displacement curve. After the elastic region, when the shaft started to move, we assumed that the length of the bonded area was reducing according to the crosshead displacement from 10 mm to 0. By the comparison of the curves shown in Fig. 6, it is possible to observe that:
In the case of the bonded joints in clearance condition, after the
3.2. Hybrid joints and adhesive bonded joints comparison In order to better understand the behaviour of the hybrid joints, their decoupling curves were compared with those of the corresponding bonded joints in clearance condition and the unbonded
Fig. 6. Average load/displacement (a) and stress/displacement (b) curves of the FT-EP samples (bonded with interference, bonded with clearance) and the interference samples as reference.
breakup of the adhesive a residual load level is recorded. This load is probably due to the friction forces provided by the cured adhesive on the mating surface of the hub and the shaft. This residual load did not behave like the interference, expiring before the complete decoupling of the joint. The interference samples provided a quite constant load resistance to the decoupling (considering an average of the stick–slip trend). In fact, in Fig. 6b it is possible to observe that during decoupling the resultant stress increased, since the bonding area was reducing. In the hybrid samples the contribution of the interference was relevant after the breakup of the adhesive. This aspect was
Fig. 7. First elastic stretch of (a) the bonded with clearance samples and (b) the bonded with interference samples.
G. Gallio et al. / International Journal of Adhesion & Adhesives 47 (2013) 63–68
67
more clearly appreciable in the stress/displacement curve reported in Fig. 6b, where the stress enhancing trend provided by the interference was clearly visible. The presence of the interference can also affect the stiffness of the hybrid joints. The first elastic stretch of the curves of the hybrid systems are compared in Fig. 7 with the ones of the systems bonded with clearance. The differences in the slopes of the curves recorded by testing the bonded with clearance samples were moderated in the hybrid systems by the interference. Thus, it seemed that the interference could be able to control the stiffness of the assembly. 3.3. The effect of the roughness The roughness is generally able to enhance friction effects in the unbonded interference joints, whereas in the hybrid ones the surface valleys can be easily filled by the adhesive. Considering all the studied samples, the roughness values ranged around a mean value of 1.5 μm of Ra and 7 μm of Rz, as shown in Fig. 8. The distribution of the roughness for the tested samples was not able to influence the push out load of the joint. As an example, the relationship between Ra and the push-out load of the hybrid PU samples is illustrated in Fig. 9, in which no particular trend can be noticed.
Fig. 9. Values of Ra and Rz for the hubs, the shafts and the assembly for the 4 PU hybrid samples.
Fig. 10. Comparison between the adhesion strength of all the studied systems. The interference system is named as “Int”, the bonded with clearance systems as “Adh”, the hybrid systems as “Int þAdh”.
3.4. Hybrid joint adhesion strength enhancement
Fig. 8. Statistical distribution of the mean Ra and Rz calculated for every hub–shaft assembly.
The adhesion strength of the joints was calculated by dividing the first load peak by the nominal coupling area, considering also in this case a mean value of the measured data for all the systems studied. The related values are reported in Fig. 10. The hybrid joints prepared with the FT-EP and PU adhesives presented an increase in the adhesion strength in respect to the corresponding “only adhesive” samples. The use of the FT-EP adhesive provided values of maximum decoupling load four times higher than that of the interference alone. In the case of the PU adhesive, the interference decoupling load was doubled in the hybrid joints. The R-EP was the only adhesive that was negatively affected by the interference. This is surely related to the high data dispersion, and was probably due to the quantity of adhesive that remained between the mating surfaces, as previously stated. Particular attention must be paid to the acrylic system. The AC adhesive presented the higher resistance improvement in the presence of the interference fit. Indeed the resistance of the hybrid joint was higher than the sum of the “adhesive” and “interference” contributions. However, it must be taken into account that this heightening could be overestimated, considering the low results obtained with the adhesive alone. The mean value of adhesion strength (equal to 5 MPa), seemed to be too low if compared to
68
G. Gallio et al. / International Journal of Adhesion & Adhesives 47 (2013) 63–68
those of single lap shear tests conducted on acetone cleaned carbon steel ( 12 MPa). Probably the compressed air inside of the hub/shaft joint could partially escape through the clearance of the assembly affecting the curing of the anaerobic adhesive. This poor result did not occur in the hybrid joint, where the interference entrapped the air inside the joint facilitating the curing of the anaerobic adhesive in a correct absence of oxygen. This could confirm the role of an incomplete curing on the low mechanical behaviour of the clearance AC samples. In order to verify this hypothesis, in the future tests a little hole will be opened inside the hub in order to permit an escape route for the compressed air.
4. Conclusions The stiffness of the hybrid joints seemed to be slightly influenced by the different adhesives. On the contrary, the break up point and the following additional resistance depended on the used adhesive system. Both the curing technology and the chemical nature of the adhesive were able to influence the decoupling behaviour. Epoxies and polyurethane adhesives presented similar curves characterised by an high initial peak, identified with the adhesive failure, and a subsequent plateau related to the interference contribution. Anaerobic acrylic samples were characterised by an almost irrelevant adhesive failure peak and an higher interference contribution. The rheology of the adhesive was an important factor to take into account in the press fit joint, influencing the quantity of
adhesive that remains between the mating surfaces and its consequent contribute to the strength of the adhesive joint. References [1] Dillard A. Advances in structural adhesive bonding. New York: Crc Press Inc; 2010. [2] Amancio-Filho S, Santos J. Joining of polymers and polymer–metal hybrid structures: recent developments and trends. Polymer Engineering and Science 2009:1461–76. [3] Petrie EM. Handbook of adhesives and sealants. New York: McGraw-Hill; 2006. [4] da Silva L, Pirondi A, Ochsner A. Hybrid adhesive joints. Heidelberg, Berlin: Springer; 2011. [5] Yoneno M, Sawa T, Shimotakahara K, Motegi Y. Axisymmetric stress analysis and strength of bonded shrink-fitted joints subjected to push-off forces. JSME International Journal Series A 1997;40:362–74. [6] Kawamura H, Sawa T, Yoneno M, Nakamura T. Effect of fitted position on stress distribution and strength of a bonded shrink fitted joint subjected to torsion. International Journal of Adhesion & Adhesives 2003;23:131–40. [7] Sekercioglu T, Gulsoz A, Rende H. The effects of bonding clearance and interference fit on the strength of adhesively bonded cylindrical components. Materials and Design 2005;26:377–81. [8] Croccolo D, Agostinis MD, Vincenzi N. Static and dynamic strength evaluation of interference fit and adhesively bonded cylindrical joints. International Journal of Adhesion & Adhesives 2010;30:359–66. [9] Dragoni E, Mauri P. Intrinsic static strength of friction interfaces augmented with anaerobic adhesives. International Journal of Adhesion and Adhesives 2000;20:315–21. [10] Dragoni E, Mauri P. Cumulative static strength of tightened joints bonded with anaerobic adhesives. Proceedings of the Institution of Mechanical Engineers Part L: Journal of Materials Design and Applications 2002;216:9–15. [11] Kleiner F, Fleischmann W. Technologies of threadlocking and interference-fit adhesive joints. In: Silva LF, Pirondi A, Ochsner A, editors. Hybrid adhesive joints. Heidelberg, Berlin: Springer; 2011. p. 227–55. [12] Henkel Ag & Co. KgaA. Henkel product selector, issue 2; 2011.
Contents lists available at ScienceDirect
International Journal of Adhesion & Adhesives journal homepage: www.elsevier.com/locate/ijadhadh
Influence of the mechanical behaviour of different adhesives on an interference-fit cylindrical joint Giorgio Gallio a,n, Mariangela Lombardi a, Davide Rovarino b, Paolo Fino a, Laura Montanaro a a b
Politecnico di Torino, DISAT Department of Applied Science and Technology, INSTM RU Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy MW Italia S.p.A., Via Pavia 72, Rivoli, Italy
art ic l e i nf o
a b s t r a c t
Article history: Accepted 5 August 2013 Available online 25 September 2013
Hybrid adhesive joining techniques are often used in many industrial sectors to design lightweight structures. A hybrid adhesive joint results from the combination of adhesive bonding with other traditional joining methods such as welding and mechanical fastening, with the aim of combining the advantages of the different techniques and overcoming their drawbacks. This study focuses on the interference fitted/adhesive bonded joining technique. In this application, two cylindrical components are coupled together by inserting one into the other, after having placed an adhesive between them. Generally anaerobic acrylic adhesives, also known as “retaining compound” are used for this application. However the effect of the adhesive nature and of its mechanical and adhesive responses on the performance of the hybrid joint is still unclear. The aim of the present research is to improve the understanding of the behaviour of different adhesives, including rigid epoxies and flexible polyurethanes, in the presence of an interference-fit. Static strength of bonded and unbonded interference fit joints have been compared in order to investigate the role of the different adhesives. & 2013 Elsevier Ltd. All rights reserved.
Keywords: Interference fit Mechanical properties of adhesives Destructive testing Joint design
1. Introduction In recent years, the use of adhesive bonding technology has greatly increased being a suitable technology able to bond dissimilar materials and often to guarantee an uniform stress distribution in the joint area [1–3]. However, the polymeric composition of adhesives implies durability issues, especially in harsh conditions (temperature and humidity). In order to overcome these problems, adhesive bonding can be used in combination with other traditional joining methods, such as mechanical fastening techniques (e.g. rivets or bolts) or welding techniques, generating a hybrid joint. In this way, hybrid adhesive joints should combine the advantages of the different techniques and, if possible, overcome their drawbacks [4]. This study focuses on the interference fitted/adhesive bonded hybrid joint. In this technique cylindrical assemblies are coupled together by inserting one part in another and by adding an adhesive between them. The hub–shaft geometry is usually exploited in technical literature in order to study this hybrid bonding technique [5–8], allowing to evaluate the contribution of both the techniques to the resultant resistance of the hybrid joint. Moreover, it is previously demonstrated that the behaviour of this hybrid joint can be also influenced by the assembling technique (press-fit or shrink fit), the type of sustained load, the fitted position and the interference level [5–11]. n
Corresponding author. Tel.: þ 39 320 6016761; fax: þ 39 011 090 6329. E-mail address: [email protected] (G. Gallio).
0143-7496/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijadhadh.2013.09.021
The acrylic anaerobic adhesives, commercially named “retaining compounds” [12], are usually exploited in the interference fitted/adhesive bonded hybrid joints [5–11], being possible in this configuration to exploit their curing technology based on a one component configuration on active metal surfaces. In this system, in fact, the presence of the interference contributes to guarantee the anaerobic polymerisation, protecting the curing layer of adhesive from the oxygen. Notwithstanding this, it is important to consider that in certain particular industrial applications other curing technologies could be more appropriate, especially when clearance zones are included in the joint design or different mechanical properties have to be matched. To the best of our knowledge, the effect of the adhesive nature and therefore of its mechanical and adhesive responses on the performance of the hybrid joint is still unclear. The adhesive nature and the curing technology could affect the performances of the interference joints. For this reason, in this study the behaviour of the principal families of structural adhesives, from a rigid epoxy to a more flexible polyurethane, will be evaluated in an interference fit joint.
2. Materials and methods 2.1. Design of the samples The role of the different adhesives in the interference fit joint was studied in a hub/shaft geometry. The hollow tubes and the
64
G. Gallio et al. / International Journal of Adhesion & Adhesives 47 (2013) 63–68
shafts, made of C40 steel, were designed according to the requirements imposed by the traction test machines (Fig. 1). The coupling diameter of the samples was set to 30 mm. This represented a versatile value that facilitated a minimum extent of thermal expansion in the case of performing either the press fit or the shrink fit assembling techniques. The threaded extremities of the samples were designed to be correctly hitched to the test machines, furthermore the use of a threaded grabber avoids slipping effects during traction tests. The head of the shaft was designed with a diameter of 40 mm after the threaded tip. This part had the same diameter than the external part of the hub and this allowed the alignment of the component during the press coupling. A coupling length of 10 mm was fixed. A loose clearance fit or a strong interference fit can be obtained by varying the dimensional tolerances of the shafts. In this work it was decided to prepare a series of samples with an interference
Fig. 1. A hub and shaft.
coupling of 20 mm, according to H7/p6 ISO standard of tolerance, and a series with a clearance coupling of about 40 mm according to a coupling tolerance of H7/f7. The technical drawings of the hollow tubes and shafts are illustrated in Fig. 2. The coupling diameter of each hub and shaft were checked in four points in the first 10 mm (the coupled length) with a Trimos horizontal measuring and calibration instrument. Then an average diameter was calculated for each sample in order to select the best coupling pairs. The roughness of the samples in terms of Ra and Rz was also checked by averaging two 8 mm profiles for each hub and shaft. Then a mean value of Ra and Rz for every coupled assembly was calculated. 2.2. Test method The interference fit joints were assembled through the press fit. The alignment between the shaft and the hub is a crucial parameter to obtain repeatable joints. A not-axial assembling induces far higher coupling and decoupling loads then an axial one. A method of guides, schematically illustrated in Fig. 3, was exploited to guarantee the alignment of the couples. The method was validated by measuring the eccentricity between the two components, after putting in rotation the coupled joints. The pressing was performed by means of an automatic press in order to assure the same pressing rate for each sample. The samples were tested under a static axial push out load, performed with a traction test using a MTS test machine at a crosshead rate of 1.3 mm/min. Considering that for the chosen hybrid joint design no standard tests are available, the crosshead speed was selected on the basis of the low load rate generally suggested in the standard shear tests on adhesive joints. Four different commercial adhesives based on different structural adhesive systems were tested:
A rigid epoxy adhesive Hysols 9492 (will be referred as “R-EP” in the paper).
A flexibilized and toughened epoxy adhesive: Scotch-welds DP490 (will be referred as “FT-EP” in the paper).
A polyurethane-epoxy modified adhesive: Araldites 2029 (will be referred as “PU” in the paper).
Fig. 2. Dimensions and tolerances of the specimen: (a) hollow tube and (b) shaft.
Fig. 3. Schematic representation of the guides system exploited to guarantee the alignment of the joints.
G. Gallio et al. / International Journal of Adhesion & Adhesives 47 (2013) 63–68
An anaerobic acrylic adhesive: Loctites 620. (will be referred as “AC” in the paper). The traction tests were carried out on samples joined with the interference only, samples bonded with the four different adhesives in clearance condition, and hybrid joints interference fitted and bonded with the four different adhesives. For each of the above-described case studies, 4 samples were tested. The same surface preparation was followed for every sample. The surfaces of hubs and shafts were cleaned with acetone, dipped in acetone and put in an ultrasonic bath for 5 min. According to the producer suggestions, the followed curing protocols were 72 h at ambient temperature for R-EP and AC adhesives; 24 h at ambient temperature followed by a postcuring at 80 1C for 1.5 h for FT-EP; 16 h at 40 1C for PU.
65
break up point occurred in correspondence with the maximum of the peak. After the adhesive failure, the load drastically decreased and the interference contribution was then recorded. In particular, in the cases of the two epoxy systems the sudden failure of the adhesive induced the machine to record load values close to zero before the interference contribution resistance comes into play. The decoupling curves of the AC samples presented some differences from the others: they were characterised by a lower adhesive failure peak and an higher interference contribution. It seemed that the cured adhesive enhanced the friction forces in the contact zone heightening the interference contribution and implying a less drastic crash of the adhesive. This particular trend, already observed where anaerobic acrylics are used [8], could be partially imputed to the absence, in these joints, of a cured spew fillet. Indeed in these anaerobic systems, the spew fillets that generally formed on both limits of the bonded area were not able to harden, because the oxygen inhibited the polymerisation.
3. Results 3.1. Hybrid joints decoupling behaviours The results of the traction tests were analysed and for each case study the four decoupling curves were averaged for obtaining a representative, mean curve, also applying a smoothing filter. The mean load/displacement curves of the different hybrid systems and the unbonded interference samples were compared in Fig. 4. The decoupling curves of the “only interference” samples were characterised by a first peak after the elastic region, that corresponds to the point in which the joint forces are broken. After this point, that is considered the resistance of the joint, the shaft starts to move. In these conditions, a quite constant load was recorded, and the curves were characterised by a sinusoidal trend. It is reasonable to assume that the low speed of testing, chosen to evaluate the adhesive behaviour in the hybrid joints, induced a discontinuous decoupling in “only interference” conditions, with consecutive stick–slip states. Grip and friction phenomena took place during the decoupling process, inducing in some case an increase in the maximum load after the joint break. Differences in the mechanical behaviour when different adhesives were employed can be observed from the experimental data in Fig. 4. The curves of the R-EP, FT-EP and PU two-component adhesives were characterised by two main phases: an initial load peak and a following plateau with the stick/slip trend typical of the interference contribution. The initial peak was due to the contribution of the adhesive plus the interference. The adhesive
Fig. 4. Mean load/displacement curves of the bonded and unbonded interference samples.
Fig. 5. Fracture shaft surfaces of the decoupled joints: (a) typical black layer of the cured FT-EP adhesive on the shaft; (b) typical white layer of the cured AC adhesive on the shaft; (c) grey residues of the cured R-EP adhesive on the shaft and (d) R-EP sample where cured particles are NOT present on the mating surface.
66
G. Gallio et al. / International Journal of Adhesion & Adhesives 47 (2013) 63–68
The mating surfaces of every studied hybrid joints after the decoupling, presented some cured adhesive residues both on the hubs and the shafts, except for the R-EP. In the case of AC, PU and FT-EP samples, the morphology of the cured layer was similar among all the four samples. It is difficult to define if a cohesive type of fracture took place because a very thin layer of adhesive remained inside the junction, as shown in Fig. 5. On the contrary, as illustrated in Fig. 5c and d, different residue quantities of R-EP remained on the shafts and the hubs of the various samples, implying a slightly wider variability in their load–displacement curves. The spreading of the adhesive in an interference fit joints depends on its rheology. Generally a low viscosity is recommended in closer joints, while a high viscosity is preferable when gap filling is required [12]. The FT-EP, AC and PU adhesives tested in this work are thixotropic, so they are “non-sag” pastes that lower their viscosity when subjected to mechanical stresses (as during the press-coupling operation). The thixotropy was not declared from the supplier in the case of the R-EP adhesive, therefore a part of this adhesive could be spewed away during the pressing and could not properly wet the joint surfaces. More studies have to be conducted to deeper analyse the relation between viscosity and the quantity of adhesive that remains in the junction.
interference samples. As an example, the mean load/displacement curves of the 4 samples for each of the 3 different cases (Bonded, Interference, Bondedþinterference) related to the FT-EP system are plotted in Fig. 6. In addition, in Fig. 6, the corresponding mean stress/displacement curves of the same FT-EP samples (Bonded, Interference, Bondedþ interference) are illustrated. The stress was calculated dividing the load by the bonded area, that was considered constant for the entire elastic trait of the load/displacement curve. After the elastic region, when the shaft started to move, we assumed that the length of the bonded area was reducing according to the crosshead displacement from 10 mm to 0. By the comparison of the curves shown in Fig. 6, it is possible to observe that:
In the case of the bonded joints in clearance condition, after the
3.2. Hybrid joints and adhesive bonded joints comparison In order to better understand the behaviour of the hybrid joints, their decoupling curves were compared with those of the corresponding bonded joints in clearance condition and the unbonded
Fig. 6. Average load/displacement (a) and stress/displacement (b) curves of the FT-EP samples (bonded with interference, bonded with clearance) and the interference samples as reference.
breakup of the adhesive a residual load level is recorded. This load is probably due to the friction forces provided by the cured adhesive on the mating surface of the hub and the shaft. This residual load did not behave like the interference, expiring before the complete decoupling of the joint. The interference samples provided a quite constant load resistance to the decoupling (considering an average of the stick–slip trend). In fact, in Fig. 6b it is possible to observe that during decoupling the resultant stress increased, since the bonding area was reducing. In the hybrid samples the contribution of the interference was relevant after the breakup of the adhesive. This aspect was
Fig. 7. First elastic stretch of (a) the bonded with clearance samples and (b) the bonded with interference samples.
G. Gallio et al. / International Journal of Adhesion & Adhesives 47 (2013) 63–68
67
more clearly appreciable in the stress/displacement curve reported in Fig. 6b, where the stress enhancing trend provided by the interference was clearly visible. The presence of the interference can also affect the stiffness of the hybrid joints. The first elastic stretch of the curves of the hybrid systems are compared in Fig. 7 with the ones of the systems bonded with clearance. The differences in the slopes of the curves recorded by testing the bonded with clearance samples were moderated in the hybrid systems by the interference. Thus, it seemed that the interference could be able to control the stiffness of the assembly. 3.3. The effect of the roughness The roughness is generally able to enhance friction effects in the unbonded interference joints, whereas in the hybrid ones the surface valleys can be easily filled by the adhesive. Considering all the studied samples, the roughness values ranged around a mean value of 1.5 μm of Ra and 7 μm of Rz, as shown in Fig. 8. The distribution of the roughness for the tested samples was not able to influence the push out load of the joint. As an example, the relationship between Ra and the push-out load of the hybrid PU samples is illustrated in Fig. 9, in which no particular trend can be noticed.
Fig. 9. Values of Ra and Rz for the hubs, the shafts and the assembly for the 4 PU hybrid samples.
Fig. 10. Comparison between the adhesion strength of all the studied systems. The interference system is named as “Int”, the bonded with clearance systems as “Adh”, the hybrid systems as “Int þAdh”.
3.4. Hybrid joint adhesion strength enhancement
Fig. 8. Statistical distribution of the mean Ra and Rz calculated for every hub–shaft assembly.
The adhesion strength of the joints was calculated by dividing the first load peak by the nominal coupling area, considering also in this case a mean value of the measured data for all the systems studied. The related values are reported in Fig. 10. The hybrid joints prepared with the FT-EP and PU adhesives presented an increase in the adhesion strength in respect to the corresponding “only adhesive” samples. The use of the FT-EP adhesive provided values of maximum decoupling load four times higher than that of the interference alone. In the case of the PU adhesive, the interference decoupling load was doubled in the hybrid joints. The R-EP was the only adhesive that was negatively affected by the interference. This is surely related to the high data dispersion, and was probably due to the quantity of adhesive that remained between the mating surfaces, as previously stated. Particular attention must be paid to the acrylic system. The AC adhesive presented the higher resistance improvement in the presence of the interference fit. Indeed the resistance of the hybrid joint was higher than the sum of the “adhesive” and “interference” contributions. However, it must be taken into account that this heightening could be overestimated, considering the low results obtained with the adhesive alone. The mean value of adhesion strength (equal to 5 MPa), seemed to be too low if compared to
68
G. Gallio et al. / International Journal of Adhesion & Adhesives 47 (2013) 63–68
those of single lap shear tests conducted on acetone cleaned carbon steel ( 12 MPa). Probably the compressed air inside of the hub/shaft joint could partially escape through the clearance of the assembly affecting the curing of the anaerobic adhesive. This poor result did not occur in the hybrid joint, where the interference entrapped the air inside the joint facilitating the curing of the anaerobic adhesive in a correct absence of oxygen. This could confirm the role of an incomplete curing on the low mechanical behaviour of the clearance AC samples. In order to verify this hypothesis, in the future tests a little hole will be opened inside the hub in order to permit an escape route for the compressed air.
4. Conclusions The stiffness of the hybrid joints seemed to be slightly influenced by the different adhesives. On the contrary, the break up point and the following additional resistance depended on the used adhesive system. Both the curing technology and the chemical nature of the adhesive were able to influence the decoupling behaviour. Epoxies and polyurethane adhesives presented similar curves characterised by an high initial peak, identified with the adhesive failure, and a subsequent plateau related to the interference contribution. Anaerobic acrylic samples were characterised by an almost irrelevant adhesive failure peak and an higher interference contribution. The rheology of the adhesive was an important factor to take into account in the press fit joint, influencing the quantity of
adhesive that remains between the mating surfaces and its consequent contribute to the strength of the adhesive joint. References [1] Dillard A. Advances in structural adhesive bonding. New York: Crc Press Inc; 2010. [2] Amancio-Filho S, Santos J. Joining of polymers and polymer–metal hybrid structures: recent developments and trends. Polymer Engineering and Science 2009:1461–76. [3] Petrie EM. Handbook of adhesives and sealants. New York: McGraw-Hill; 2006. [4] da Silva L, Pirondi A, Ochsner A. Hybrid adhesive joints. Heidelberg, Berlin: Springer; 2011. [5] Yoneno M, Sawa T, Shimotakahara K, Motegi Y. Axisymmetric stress analysis and strength of bonded shrink-fitted joints subjected to push-off forces. JSME International Journal Series A 1997;40:362–74. [6] Kawamura H, Sawa T, Yoneno M, Nakamura T. Effect of fitted position on stress distribution and strength of a bonded shrink fitted joint subjected to torsion. International Journal of Adhesion & Adhesives 2003;23:131–40. [7] Sekercioglu T, Gulsoz A, Rende H. The effects of bonding clearance and interference fit on the strength of adhesively bonded cylindrical components. Materials and Design 2005;26:377–81. [8] Croccolo D, Agostinis MD, Vincenzi N. Static and dynamic strength evaluation of interference fit and adhesively bonded cylindrical joints. International Journal of Adhesion & Adhesives 2010;30:359–66. [9] Dragoni E, Mauri P. Intrinsic static strength of friction interfaces augmented with anaerobic adhesives. International Journal of Adhesion and Adhesives 2000;20:315–21. [10] Dragoni E, Mauri P. Cumulative static strength of tightened joints bonded with anaerobic adhesives. Proceedings of the Institution of Mechanical Engineers Part L: Journal of Materials Design and Applications 2002;216:9–15. [11] Kleiner F, Fleischmann W. Technologies of threadlocking and interference-fit adhesive joints. In: Silva LF, Pirondi A, Ochsner A, editors. Hybrid adhesive joints. Heidelberg, Berlin: Springer; 2011. p. 227–55. [12] Henkel Ag & Co. KgaA. Henkel product selector, issue 2; 2011.